Page i
Advances in Giardia Research edited by: Peter M. Wallis Kananaskis Centre for Environmental Research Univer...
204 downloads
1195 Views
5MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Page i
Advances in Giardia Research edited by: Peter M. Wallis Kananaskis Centre for Environmental Research University of Calgary, Calgary, AB CANADA T2N 1N4 and Brian R. Hammond Alberta Environment, 10405 Jasper Ave., Edmonton, AB CANADA T5K 3N4
The University of Calgary Press
Page ii
Disclaimer: This book contains characters with diacritics. When the characters can be represented using the ISO 88591 character set (http://www.w3.org/TR/images/latin1.gif), netLibrary will represent them as they appear in the original text, and most computers will be able to show the full characters correctly. In order to keep the text searchable and readable on most computers, characters with diacritics that are not part of the ISO 88591 list will be represented without their diacritical marks. © 1988 Kananaskis Centre for Environmental Research, University of Calgary. All rights reserved. ISBN 0919813860 The University of Calgary Press, 2500 University Drive NW, Calgary, AB CANADA T2N 1N4 Canadian Cataloguing in Publication Data Main entry under title: Advances in Giardia research Papers from the Calgary Giardia Conference held Feb. 2325, 1987. Includes index. ISBN 0919813860 1. Giardia lamblia—Congresses. 2. Giardiasis—Congresses. I. Wallis, Peter Malcolm. II. Hammond, Brian R., 1934 III. Calgary Giardia Conference (1987 : Calgary, Alta.) QR201.G45A38 1989 616'.016 C880916370 No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission from the Publisher. Printed in Canada
Page iii
PREFACE The papers in this book were prepared from the Calgary Giardia Conference which was intended to provide a forum for the reporting and summarizing of the results of recent research and development in the study of this important, worldwide parasite. All of the papers were reviewed by the Editors and the Chairperson of the session in which they were presented. The editors have made the format of the papers as uniform as possible but have not attempted to standardize spelling in recognition of the international nature of the papers contained in this volume. The Conference was attended by 150 scientists, engineers, and public health officials from Canada, the United States, Central America, the United Kingdom, Australia, and Europe. Their enthusiastic participation was directly responsible for making the Conference a success. The editors would like to express their deep appreciation to the following agencies and individuals for their efforts and support. • The Calgary Giardia Conference was funded by the Alberta Environmental Research Trust, Alberta Environment, Health and Welfare Canada, the Alberta Heritage Foundation for Medical Research, the Alberta Environmental Centre and the University of Calgary. Without their support, the Conference and this volume would not have been possible. • The editors wish to thank all of the authors of scientific papers for their excellent presentations and patience throughout the lengthy process of publication. • The Conference and this volume would not have been possible without the organizational and word processing efforts of Grace Lebel and the public relations work of Janice Crowther. • The editors are especially indebted to Jane Lancaster who mastered desktop publishing in record time, to Anne Hannan for her accurate and patient formatting of many papers and to Dave Savage for his imaginative programming which saved us all an enormous amount of work. • We also wish to acknowledge the assistance of Terry Zenith, Section Head of the Alberta Environment Drafting Pool for his help in overhauling some of our graphics. PETER WALLIS BRIAN HAMMOND
Page iv
This book is dedicated to my wife Marcia without whose understanding and support this project would never have been completed
Page v
ADVANCES IN GIARDIA RESEARCH iii
Preface Epidemiology, Pathogenesis and Drug Sensitivity
Drug Resistance and the Treatment of Giardiasis P.F.L. Boreham, N.C. Smith and R.W. Shepherd Cell Injury in Giardia lamblia Detected by Forward Light Scatter B. Kinosian, R.H. Gilman, J. Ordonez, J. O'Hare, S. Wahl, F. Koster and W. Spira
913
The Importance of Nonwaterborne Modes of Transmission for Giardiasis, A Case Study S. Harley
1519
A New Miniculture Technique for determining In Vitro Antimicrobial Agent Sensitivity of Axenically Cultivated Strains of Giardia lamblia S.M. Wahl, R.H. Gilman, J.P. O'Hare, D.B. Keister and W.M. Spira
2124
Ultrastructural Study of a Bacterial Symbiont of Giardia lamblia S. Radulescu, E.A. Meyer, B. Burghelea and T. Meitert
2528
Morphology of Giardia Encystation In Vitro D.G. Schupp, M.M. Januschka and S.L. Erlandsen
2932
Cytopathogenicity of Giardia lamblia in HeLa and Vero Cell Monolayers A. Jyothisri and U.K. Baveja
3337
Studies on the Prevalence of Giardiasis in Czechoslovakia M. Giboda
3941
Immunology
Immunology of Giardia Infections M.F. Heyworth
4548
The Secretory Immune Response in Rats Infected with Rodent Giardia duodenalis Isolates and Evidence for Passive Protection with Immune Bile G. Mayrhofer and A. Waight Sharma
4954
Biological Differences in Giardia lamblia T.E. Nash and A. Aggarwal
5558
Animal Models and CrossInfection
37
Prevalence of Giardia sp. in Dogs from Alberta P.D. Lewis, Jr.
6164
Location of Giardia Trophozoites in the Small Intestine of Naturally Infected Dogs in San Diego H. Douglas, D.S. Reiner, M.J. Gault and F.D. Gillin
6569
Seasonal Increase in the Incidence of Giardia lamblia in Arkansas J.J. Daly, M.A. Gross, D. McCullough, T. McChesney, S.K. Tank, E.B. Daly and C.L. Puskarich
7174
Page vi
Infection of Mongolian Gerbils (Meriones unguiculatus) with Giardia from Human and Animal Sources K.D. Swabby, C.P. Hibler and J.G. Wegrzyn
7577
Transmission of Giardia duodenalis from Human and Animal Sources in Wild Mice P.D. Roach and P.M. Wallis
7982
Water Treatment
Water Treatment and the Giardia Cyst A. van Roodselaar
8586
Removal of Giardia Through Slow Sand Filtration 100 Mile House, British Columbia J.M.G. Bryck, B.L. Walker and D.W. Hendricks
8793
Comparison of Some Filtration Processes Appropriate for Giardia Cyst Removal G.S. Logson Monitoring as a Tool in Waterborne Giardiasis Prevention J.L. Sykora, W.D. Bancroft, A.H. Brunwasser, S.J. States, M.A. Shapiro, S.N. Boutros and L.F. Conley
103 106
The Efficiency of Point of Use Devices for the Exclusion of Giardia muris cysts from a Model Water Supply System D.R. Cullimore and H. Jacobsen
107 112
Diatomite Filtration: Why it Removes Giardia from Water H.G. Walton
113 116
Small Water System Improvements for Giardia Removal A Case Study M.R. Alberi, S.J. Quail and R.A. Kruse
117 124
Inactivation of Giardia lamblia Cysts from a Surface Water by Oxidation with Ozone C. Nebel, A. Lally, T. Bosher, J.W. Hmurciak, L. Hmurciak and D.A. Breen
125 128
A Regulatory Agency's Experience with Giardia S. McClure and I.B. Mackenzie
129 131
Effects of Chlorine on the Ultrastructure of Giardia Cysts M. Neuwirth, P.D. Roach, J.M. BuchananMappin and P.M. Wallis
133 135
Removal and Inactivation of Giardia Cysts in a Mobile Water Treatment Plant Under Field Conditions: Preliminary Results P.M. Wallis, J.S. Davies, R. Nutbrown, J.M. BuchananMappin, P.D. Roach and A. van Roodselaar
137 144
Differentiation of Giardia Isolates
95102
The Genome of Giardia intestinalis P. Upcroft, P.F.L. Boreham and J.A. Upcroft
147 152
The Partial Characterization of an Immunodominant Antigen of Giardia intestinalis J.A. Upcroft, A.G. Capon, A. DharmkrongAt, P. Upcroft, and P.F.L. Boreham
153 157
Immunofluorescence Differentiation Between Various Animal and Human Source Giardia Cysts Using Monoclonal Antibodies H.H. Stibbs, E.T. Riley, J. Stockard, J.L. Riggs, P.M. Wallis and J.Issac Renton
159 163
Comparison of Giardia Isolates by DNADNA Hybridization A. Uji, P.M. Wallis and W.M. Wenman
165 167
Page vii
Differentiation of Giardia duodenalis from Giardia muris by Immobilization in Various Sera D.L. Lehmann and P.M. Wallis
169 172
Conserved Sequences of the HSP Gene Family in Giardia lamblia A. Aggarwal, P. Romans, V.F. de la Cruz and T.E. Nash
173 175
The Response of Humans to Antigens of Giardia lamblia M.G. OrtegaPierres, R. Lascurain, R.A. Garcia, R.C. Vazquez, G. Acosta and J.I. Santos
177 180
Properties of Giardia lamblia RNAs. C. Montanez, L. Cervantes, C. Ovando and M.G. OrtegaPierres
Enzyme Activites of Giardia lamblia and Giardia muris Trophozoites and Cysts D.G. Lindmark, and J.J. Miller
187 189
Studies on Giardia lamblia Trophozoite Antigens Using Sephacryl S300 Column Chromatography, Polyacrylamide Gel Electrophoresis and Enzymelinked Immunosorbent Assay P.P. Chaudhuri, S. Pal, S.C. Pal, and P. Das
191 194
Detection of Giardia Cysts
181 185
An Overview of the Techniques Used for Detection of Giardia Cysts in Surface Water C.P. Hibler
197 204
Methods for the Recovery of Giardia and Cryptosporidium from Environmental Waters and their Comparative Occurrence J.B. Rose, D. Kayed, M.S. Madore, C.P. Gerba, M.J. Arrowood, C.R. Sterling and J.L. Riggs
205 209
Comparison of Five Procedures for the Sedimentation of Giardia lamblia and Other Protozoan Cysts D.R. Pennell, J.F. Stoebig, D.E. Sampson, and R.F. Schell
211 213
Comparison of the Modified "Reference Method" and the Indirect Fluorescent Antibody Technique for Detection of Giardia Cysts in Water B.E. Quinones, C.P. Hibler and C.M. Hancock.
215 217
Giardia Detection using Monoclonal Antibodies Recognizing Determinants of In Vitro Derived Cysts C.R. Sterling, R.M. Kutob, M.J. Gizinski, M. Verastegui, and L. Stetzenbach
219 222
Routine Monitoring of Watersheds for Giardia Cysts in Northeastern Pennsylvania S.A.M. McFarlane
223 225
Waterborne Giardiasis: Sources of Giardia Cysts and Evidence Pertaining to their Implication in Human Infection S.L. Erlandsen and W.J. Bemrick
227 236
Analysis of Municipal Water Samples for Cysts of Giardia C.P. Hibler
237 245
Page viii
Viability Testing A Review of Methods that are used to Determine Giardia Cysts Viability F.W. Schaefer, III
249 254
Fluorescent Dye Exclusion as a Method for Determining Giardia Cyst Viability S.J. Hudson, J.F. Sauch and D.G. Lindmark
255 259
A New Method for Excystation of Giardia J.F. Sauch
261 264
Assessing Giardia Cysts Viability with Fluorogenic Dyes: Comparisons to Animal Infectivity and Cyst Morphology by Light and Electron Microscopy D.G. Schupp, M.M. Januschka and S.L. Erlandsen
265 269
Panel Discussions Excystation and Encystation F.D. Gillin, E.A. Meyer, S. Erlandsen, C. Sterling
273
The Implications of Regulatory Changes for Water Treatment in the United States S. Regli, A. Amirtharajah, B. Borup, C. Hibler, J. Hoff, and R. Tobin
275 286
Taxonomy of the Genus Giardia S.L. Erlandsen, E.A. Meyer, T.E. Nash
287 289
Methods of Handling Giardia in the Laboratory W. Jakubowski, E.A. Meyer, T.E. Nash, C.P. Hibler
291 294
Index
295 302
Page 1
EPIDEMIOLOGY, PATHOGENESIS AND DRUG SENSITIVITY
Page 3
Drug Resistance and the Treatment of Giardiasis P.F.L. Boreham*, N.C. Smith and R.W. Shepherd Queensland Institute of Medical Research, Bramston Terrace, Herston, Brisbane, Qld. 4006, Australia The possible existence of drug resistant Giardia intestinalis has been investigated, as a possible explanation for treatment failures in patients. Analysis of 15 isolates of G. intestinalis has demonstrated major differences in sensitivities to metronidazole, tinidazole, furazolidone and quinacrine. Each isolate is heterogeneous and is composed of populations of parasites with differing drug sensitivities and doubling times. Cross resistance between the 5nitroimidazoles has been demonstrated in an in vitro test. Clinical and laboratory data provide strong evidence for drug resistance in G. intestinalis. Investigations of the molecular basis of drug resistance suggest that different mechanisms occur with the nitrofurans and the 5nitroimidazoles, with the former being related to the glutathione cycling enzymes, glutathione perioxidase and gluthathione reductase and the latter to pyruvate: ferredoxin oxidoreductase activity.
Introduction Management of patients with giardiasis often proves to be difficult due to problems related to accurate diagnosis, lack of knowledge concerning pathophysiology and a lack of fully effective chemotherapeutic agents (8,10,11). Research into the treatment of giardiasis has been limited by the lack of suitable laboratory models. All the existing drugs were developed for other infectious diseases and subsequently found empirically to be active against Giardia intestinalis. In this paper we briefly review the existing drugs and their deficiencies and discuss some of our current research which is designed to effect improvements in the therapy of infected humans, particularly children. The Current Armamentarium Four drugs, metronidazole, tinidazole, furazolidone and quinacrine are commonly used to treat giardiasis, but the choice is largely dependent upon the personal preference of the prescribing physician, drug availability and to a degree, the occurrence of untoward effects. The two 5nitroimidazoles commonly used are metronidazole and tinidazole. These drugs may cause nausea, gastrointestinal discomfort, lassitude, skin rashes, drowsiness, disulfiramlike reactions with alcohol and occasionally transient leucopenia and peripheral neuropathy. Nitroimidazoles have been shown to be carcinogenic in rodents and mutagenic in bacteria. Single and multidose regimens have been evaluated but there is not general concensus on the most appropriate course of treatment (10,11). A third nitroimidazole, ornidazole, has been evaluated and appears to be equipotent to tinidazole (22). A new member of this group, satranidazole, is currently undergoing clinical trials in India and the preliminary data look most promising (20). Many physicians consider furazolidone, a 2nitrofuran, to be the drug of choice for the treatment of giardiasis in young children. However, a wide range of mild side effects can occur, including headache, nausea, vomiting, skin rashes, diarrhea and malaise. More severe side effects, such as agranulocytosis and hemolytic anemia, in patients with glucose6phosphate dehydrogenase deficiency, may occur. Quinacrine is commonly used in North America. This antimalarial compound may cause dizziness, headache and gastrointestinal disturbances. The fact that quinacrine causes toxic psychoses in 12% of patients, together with occasional cases of exfoliative dermatitis and aplastic anemia, has resulted in this drug not being used by some physicians. Paromomycin sulfate has been recommended for the treatment of giardiasis in pregnancy, mainly because, as an aminoglycoside, it is not absorbed from the gut (14). However, controlled trials have not yet been conducted and it should be used with caution. Other drugs which have been recommended for the treatment of giardiasis include amodiaquine (21), berberine sulfate (13), sulfasalazine (1) and erythromycin (19) but again none of these drugs have been exposed to controlled clinical trials. A major problem with the current drugs is that treatment failures are known to occur with all of them. It is very difficult to assign accurate figures to these failure rates since every study uses different assessment criteria for cure, ranging from a single stool examination to multiple examinations over several months together with a small intestinal biopsy. Based on 11 published reports metronidazole has a cure rate of 4695%, tinidazole 88100%, furazolidone 5892% and quinacrine 60100%. It is generally accepted that treatment failures do occur with all four drugs and that this poses a serious problem to physicians. Many reasons can be postulated to explain these treatment failures including: patient noncompliance with the prescribed drug regimen. This is certainly an important consideration and recent studies have demonstrated major problems in this area (3) the possible reinfection of the patient. At present there is no way to effectively monitor this by typing isolates changes in the pharmacokinetics of the drug * Corresponding author.
Page 4
possible escape of organisms to priviliged sites where antigiardial drugs are unable to reach deficiencies in the host's immune system the inactivation of the drug by concommitant bacterial infections (12) the existence of drug resistant strains. Resistance to metronidazole has been well documented in the Trichomonads and also in some anaerobic bacteria. Screening Tests Against Giardia intestinalis 1. In vitro Research on drugs for the treatment of giardiasis has been severely hindered by the lack of appropriate screening tests. Most work has either involved testing drugs directly on man, where standardization and assessment of cure have been problems, or by using Giardia muris in the mouse as a model (2,11). In order to investigate the possible existence of resistant strains of G. intestinalis we first developed appropriate standardized drug screening tests. Culture in microtitre trays was achieved by incubating the plates in an atmosphere of nitrogen in sealed containers and a test to measure reproductive viability, based on the uptake of [3H]thymidine into the nuclei of the organisms, was developed (4). This assay proved to be considerably more sensitive than using either flagellar movement or dye exclusion as an index of viability. Development of this in vitro test has allowed the drug sensitivities of different isolates of human G. intestinalis to be compared and compounds to be screened for their activity against G. intestinalis (5). Analysis of 15 isolates cultured from patients attending the Royal Children's Hospital, Brisbane, has shown that there is a tenfold difference in sensitivity between these isolates for metronidazole and furazolidone, a threefold difference for tinidazole and a twentyfold difference for quinacrine (8,15 and unpublished data). In addition, by examining the drug sensitivities of cloned lines derived from a single stock it has been shown that each isolate of G. intestinalis is not homogeneous, but is composed of different populations of organisms having differing drug sensitivities (7). Doubling times of the stocks also vary considerably, ranging from 12.5 to 44.5 hours when grown in axenic culture in the absence of bile from the medium (15). 2. In vivo An in vivo test for drug sensitivity has also been developed using a neonatal mouse model (6). Litters of mice less than 5 days of age are infected with 3 × 104 trophozoites via an intragastric tube and after 6 days half of each litter are treated with an appropriate concentration of the drug under study in 50µL of vehicle also by the intragastric route. The other half of the litter act as controls and are treated with the vehicle alone. A further 2 days later the mice are killed, the small intestine removed, opened longitudinally and placed in cold buffer to allow the trophozoites to detach. The total number of parasites present can then be counted and expressed as a percentage of the untreated controls. Using this technique a number of compounds have been assayed and it has been shown that there is a direct correlation between in vitro and in vivo activity for twelve 5nitroimidazoles (P<0.05) (6). Experimental results have also shown that the observed activity in man for the commonly used drugs is reflected in this model system. For example, tinidazole is more active than metronidazole but roughly equipotent with ornidazole. Furazolidone is slightly less active than tinidazole but more active than quinacrine. It is interesting to note that these in vivo studies have indicated that berberine sulfate, sulfasalazine and erythromycin are inactive. Paromomycin sulfate is active when given by the oral route to neonatal mice but is inactive when injected subcutaneously. These results indicate that isolates of G. intestinalis exist which have differing sensitivities for the four drugs commonly used to treat giardiasis and that a single stock is heterogeneous, being composed of populations of organisms having different drug sensitivities. From studies such as these it is not possible to define what constitutes resistant or sensitive organisms. Such information can only be obtained from appropriate studies on patients. By analogy with the early work on malaria drug resistance, it is necessary to compare in vitro and in vivo responses of a large number of isolates with the clinical response and effect on parasite numbers following treatment. Only then will it be possible to nominate drug concentrations where, if the organisms are not killed in vitro, drug resistance can be suspected. Preliminary results to investigate this question are given below. Evidence for Drug Resistance in Human Patients To date we have studied in detail ten children treated with furazolidone (8mg/kg/day) alone (15). Clinical data were collected before and after therapy and studies conducted on the organisms taken from these patients and established in axenic culture. Seven of the ten patients had prompt resolution of symptoms but in three patients the symptoms persisted. Clinical cure only was used in these studied because of the well known difficulties associated with accurately documenting parasitological cure. Two of the three isolates from patients with persistent symptoms were the least sensitive of the ten tested to furazolidone in vitro (Figure 1). These two patients did not respond clinically
Figure 1. Furazolidone sensitivity of trophozoites isolated from patients showing either a clinical response or no clinical response to therapy. ID50 the concentration of drug required to inhibit the uptake of [3H] thymidine in vitro by 50%. *patient treated for 5 days only.
Page 5
to a second ten day course of treatment. The isolates from these two patients fell within the normal range for metronidazole and tinidazole sensitivity. Both patients responded satisfactorily when treated with a nitroimidazole. The third patient in whom furazolidone therapy failed, had organisms with in vitro sensitivity in the range where therapy would be expected to be satisfactory. This patient however, received only 5 days treatment instead of the recommended 10 days (18) and when given a complete 10 day course of furazolidone responded satisfactorily. This study indicates that in two patients either the organisms were resistant to furazolidone or that differences in the pharmacokinetics of the drug occurred. Since little absorption of furazolidone takes place from the gut the second explanation would seem unlikely. In addition, these results illustrate that it is possible to design specific therapy regimens for those patients who proved to be refractory to standard treatments. Selection of Drug Resistance in the Laboratory Experimental selection of a drug resistant line of G. intestinalis has been achieved in the laboratory by growing stock BRIS/83/HEPU/106 in medium containing twice the dose of metronidazole required to kill 10% of the organisms (2ID10). Over a period of twelve months the concentration of drug required to kill 50% of trophozoites (ID50) increased approximately tenfold when compared to the stock grown in the absence of the drug. This selected line has been used to examine the possiblity of the existence of cross resistance between the nitroimidazoles and also to investigate the mechanisms of resistance.
Figure 2. The decrease in sensitivity of eleven 5nitroimidazoles when tested against a line of stock BRIS/83/HEPU/106 selected for resistance by growing the organisms in a sublethal concentration of metronidazole for one year, compared with the stock grown in the absence of drug: NIMnimorazole; S75S75 0400A [1methyl2(1dimethylaminomethyleniminophenoxymethyl) 5nitroimidazole hydrochloride]; FEXfexinidazole; ORNornidazole; PANpanidazole; SATsatranidazole; FLUflunidazole; RONronidazole; TINtinidazole; SECsecnidazole; METmetronidazole.
Figure 3. Diagramatic representation of the mode of action of metronidazole based on Muller (17).
Cross Resistance Between the Nitroimidazoles A total of eleven 5nitroimidazole compounds have been tested against the laboratory selected line of G. intestinalis and all showed less susceptibility to the drug when compared with the control stock (Figure 2). The reduction in sensitivity varied between two and six fold. Little or no resistance was found with unrelated drugs, including furazolidone and quinacrine. In view of the considerable interest shown recently by pharmaceutical companies in compounds of the 5nitroimidazole series for the treatment of protozoal infections, this finding is of importance for designing treatment regimens for patients who fail to respond effectively to metronidazole or tinidazole. Mechanisms of Drug Resistance Although no studies on the mode of action of metronidazole have been conducted with Giardia there is much information on its mode of action and on possible mechanisms of resistance with other microorganisms. The mode of action of metronidazole has been reviewed by Muller (17) and is shown diagramatically in Figure 3. It is known that metronidazole is taken up into cells and reduced to reactive intermediates with the formation of associated free radicals which are responsible for the drug's activity by causing strand breakage and cross linking of DNA. Drug resistance could result in three ways: decreased uptake of metronidazole by the resistant cells accumulation of free metronidazole within the cell resulting from the failure of the formation of the reactive intermediates by the enzyme pyruvate:ferredoxin oxidoreductase failure of the reactive metabolites of metronidazole to act on DNA.
Page 6
The mode of action of furazolidone has not been studied in the same depth but it seems probable that it also produces free radicals within the cell which cause damage to DNA. There is evidence that the nitroimidazoles and nitrofurans act via slightly different mechanisms (16). Anaerobic reduction of metronidazole is dependent upon the hydrogenosomal enzyme pyruvate:ferredoxin oxidoreductase but the nitrofurans may be reduced by cytosolic enzymes such as NADH and NADPH oxidase as well as pyruvate:ferredoxin oxidoreductase. This is because nitrofurans have a higher positive reduction potential than the 5nitroimidazoles. To date we have investigated aspects of the first two mechanisms of resistance. 1. Uptake of Metronidazole into the Cell The amount of [14C]metronidazole taken up into the cell is less in the line selected for resistance than in the same stock grown in the absence of metronidazole. This suggests accumulation of a free intracellular pool of metronidazole resulting in a decreased uptake across the concentration gradient or alternatively, a defective uptake mechanism. Since the uptake of the nitroimidazoles is believed to be by a passive process the former explanation would seem to be most likely. 2. Enzyme Studies The glutathione redox cycling enzymes, glutathione peroxidase and glutathione reductase are present in most cells and are responsible for scavenging free radicals, especially peroxides. They thus act as a protective mechanism for the cell. Glutathione reductase ensures the supply of reduced glutathione for this reaction as well as converting NADPH to NADP, the first step in the pentosephosphate pathway. We have measured the concentration of glutathione peroxidase and glutathione reductase in 15 stocks of G. intestinalis and correlated the results with drug sensitivities to metronidazole, tinidazole, furazolidone and quinacrine determined by the in vitro [3H]thymidine incorporation assay. Furazolidone sensitivity of these stocks correlated with the activity of the glutathione redox cycling enzymes but this was not true for the 5nitroimidazoles. This indicates that furazolidone resistance is probably related to free radical scavenging by glutathione peroxidase, but the same does not appear to be true for metronidazole or tinidazole. A significant negative correlation between scavenging enzyme concentrations and quinacrine resistance was found in these stocks but the interpretation of this result is unclear. Pyruvate:ferredoxin oxidoreductase appears to be primarily responsible for the metabolism and subsequent toxicity of 5nitroimidazoles (16) and deficiencies in this enzyme have been reported to account for metronidazole resistance in Tritrichomonas foetus (9). Preliminary data indicates that the same mechanism may be true for G. intestinalis since the metronidazole drug selected line has only onesixth the amount of pyruvate:ferrodoxin oxidoreductase activity compared to the parent isolate. Thus the primary defect in the line resistant to metronidazole appears to be a decrease in the pyruvate:ferredoxin oxidoreductase enzyme resulting in a decreased rate of reduction of the drug. This will lower the drug concentration gradient across the cell and hence reduce drug uptake. The absence of superoxide dismutase activity in any of the isolates of G. intestinalis so far examined would favour an anaerobic interpretation of drug resistance. Thus the hydrogenosomal ferredoxins appear to play a critical role in nitroimidazole toxicity, whereas flavines may be responsible for the reduction of nitrofurans (16). Conclusions There is an urgent need to develop new drugs and improved strategies for treatment of giardiasis to overcome the difficulties of drug failures and adverse reactions inherent in the existing drugs. A rational approach to the development of new chemotherapeutic agents should include a study of parasitespecific metabolic pathways and development of an appropriate inhibitor. Currently we know very little of the metabolic pathways of G. intestinalis. Any new potential compounds would need to be tested in the laboratory both in vitro and in an in vivo model prior to clinical trials. Until recently this approach has not been possible because of the lack of suitable models, but the work described here is the first step towards such a rational design of antigiardial compounds. The recognition of the likelihood that drug resistant organisms exist is an important consideration. Although drug resistance appears to be a significant clinical problem with only a few of the parasitic diseases, its existence with at least two flagellates, Trypanosoma sp. and Trichomonas sp., means that careful surveillance of the situation must be maintained. Much more information on the extent and significance of resistance in giardiasis is required as our knowledge of this area is very limited. One important implication of these studies is that it is now possible to develop specific therapy regimens for those patients refractory to standard treatments, provided that it is possible to establish axenic cultures of their parasitic organisms. Acknowledgements The original research described in this paper has been supported by generous grants from the National Health and Medical Research Council of Australia. We thank Professor C. Bryant for helpful discussions and Mr. R.E. Phillips and Mrs. K. Anderson for excellent technical assistance. We also thank G.D. Searle and Co. for their generous gift for [14C]metronidazole and the following companies for gifts of drugs: Boehringer Ingelheim Pty Ltd.; Ciba Geigy Research Centre, India; Farmitalia Carlo Erba Ltd.; Hoechst AG; May & Baker Ltd.; Merck, Sharpe & Dohme (Aust) Pty Ltd.; Norwich Eaton Pty Ltd.; Parke Davis Pty Ltd.; Pfizer Pty Ltd.; Roche Products Pty Ltd. and Specia. Literature Cited 1. Azizi, E., J. Karpuchas, and T. Rosenberg. 1983. Salicylazosulfapyridine therapy in a child with chronic lambliasis due to acquired hypogammaglobulinemia. Helv. Paediatr. Acta 38:8790.
Page 7
2. Bemrick, W.J. 1963. A comparison of seven compounds for giardiacidal activity in Mus musculus. J. Parasitol. 49:819823. 3. Boreham, P.F.L., S. Benrimoj, M. Ong, M. Craig, R.W. Sheperd, D. Hill, and B. Singleton. 1986. A compliance study in pediatric patients receiving treatment for giardiasis. Aust. J. Hosp. Pharm. 16:138142. 4. Boreham, P.F.L., R.E. Phillips, and R.W. Shepherd. 1984. The sensitivity of Giardia intestinalis to drugs in vitro. J. Antimicrob. Chemother. 14:449461. 5. Boreham, P.F.L., R.E. Phillips, and R.W. Shepherd. 1985. A comparison of the in vitro activity of some 5nitroimidazoles and other compounds against Giardia intestinalis. J. Antimicrob. Chemother. 16:589595. 6. Boreham, P.F.L., R.E. Phillips, and R.W. Shepherd. 1986. The activity of drugs against Giardia intestinalis in neonatal mice. J. Antimicrob. Chemother. 18:393 398. 7. Boreham, P.F.L., R.E. Phillips, and R.W. Shepherd. 1987. Heterogeneity in the responses of clones of Giardia intestinalis to antigiardial drugs. Trans. R. Soc. Trop. Med. Hyg. 81:406407. 8. Boreham, P.F.L., and R.W. Shepherd. 1985. The treatment of giardiasis, pp. 320326, In S. Tzipori (ed.), Infectious diarrhoea in the young. Elsevier, Amsterdam. 9. Cerkasovova, A., J. Cerkosov, and J. Kulda. 1984. Metabolic differences between metronidazole resistant and susceptible strains of Tritrichomonas foetus. Mol. Biochem. Parasitol. 11:105118. 10. Craft, J.C. 1982. Giardia and giardiasis in childhood. Pediatr. Infect. Dis. 1:196211. 11. Davidson, R.A. 1984. Issues in clinical parasitology: the treatment of giardiasis. Am. J. Gastroenterol. 79:256261. 12. Edwards, D.I., E.J. Thompson, J. Tomusange, and D. Shanson. 1979. Inactivation of metronidazole by aerobic organisms. J. Antimicrob. Chemother. 5:315 316. 13. Gupte, S. 1975. Use of berberine in treatment of giardiasis. Am. J. Dis. Child. 129:866. 14. Kreutner, A.K., V.E. Del Bene, and M.S. Amstey. 1981. Giardiasis in pregnancy. Am. J. Obstet. Gynecol. 140:895901. 15. McIntyre, P., P.F.L. Boreham, R.E. Phillips, and R.W. Shepherd. 1986. Chemotherapy of giardiasis: clinical responses and in vitro drug sensitivity of human isolates in axenic culture. J. Pediatr. 108:10051010. 16. Moreno, S.N.J., R.P. Mason, and R. Docampo. 1984. Distinct reduction of nitrofurans and metronidazole to free radical metabolites by Tritrichomonas foetus hydrogenosomal and cytosolic enzymes. J. Biol. Chem. 259:82528259. 17. Muller, M. 1983. Mode of action of metronidazole on anaerobic bacteria and protozoa. Surgery 93:165171. 18. Murphy, T.V., and J.D. Nelson. 1983. Five vs ten days' therapy with furazolidone for giardiasis. Am. J. Dis. Child. 137:267270. 19. Nash, P.H. 1976. Giardiasis. Can. Med. Assoc. J. 115:1819. 20. Poltera, A.A., R. Figueroa, H. Reyes, T. Koura, and R. Degen. 1986. Successful low dose therapy in Giardia lamblia patients with CibemideR a new 5nitro imidazole. Abstr. 516, Handbook, Sixth Int. Congr. Parasitol. M.J. Howell (ed.). Australian Academy of Sciences, Canberra. 21. Rosenberg, J., and E. Neumann. 1957. Efficacy of amodiaquin hydrochloride (Camoquin hydrochloride) against giardiasis. Am. J. Trop. Med. Hyg. 6:679680. 22. Sabchareon, A., T. Chongsuphajaisiddhi, and P. Attanath. 1980. Treatment of giardiasis in children with quinacrine, metronidazole, tinidazole and ornidazole. Southeast Asian J. Trop. Med. Public Health 11:280284.
Page 9
Cell Injury in Giardia lamblia Detected by Forward Light Scatter B. Kinosian*, R.H. Gilman, J. Ordonez, J. O'Hare, S. Wahl, F. Koster, and W. Spira Dept. of Medicine, Francis Scott Key Medical Center, The John Hopkins University School of Medicine, Baltimore, Maryland, 21224 Current methods for determining antimicrobial susceptibility of Giardia lamblia rely on tedious microscopic observation of trophozoite motility or time consuming assays of clonal growth. We axenically cultured Giardia trophozoites in the presence of various concentrations of metronidazole (.1610 µg/mL) and furazolidone (0.82.5 µg/mL). Viability was assessed at 24 hours with 1) fluorescein diacetate (FDA), a metabolism dependent dye, 2) microscopic observation of motility and 3) mean forward light scatter (FLS), measured by passing the cells through a fluorescence activated cell sorter. Cell injury, indicated by a significant increase in FLS, occurred at 0.64 µg/mL of metronidazole (MTZ) and 0.08 µg/mL of furazolidone (FZD). Loss of fluorescence (0.64 µg/mL MTZ) and loss of motility (1.25 µg/mL MTZ and 0.32 µg/mL FZD) occurred at the same or higher drug concentrations. Increase in FLS correlated with decrease in FDA fluorescence, a marker of cell viability (r= 0.94, p<.001). When injury was assessed after various durations of drug exposure, FLS increased 8 hours after exposure to 1.25 µg/mL MTZ and 12 hours after exposure to 0.64 µg/mL MTZ. At effective doses (>0.64 mcg/mL), increase in FLS was correlated with increasing time of exposure, up to 24 hours (r>0.94 for all drug concentrations). Analysis of light scatter provides a rapid, sensitive, and accurate method of detecting early injury in G. lamblia.
Introduction Current methods for determining the antimicrobial susceptibility of Giardia lamblia rely on microscopic observation of trophozoite motility (6), or assay of clonal growth (2). The former method is tedious, and highly observerdependent. The latter method is time consuming, requiring up to 3 days for final culture results. Flow cytometry provides a means of examining structural and functional properties of large numbers of individual cells rapidly, with high precision. Applications of cytometry in microbial studies have included rapid identification of organisms in environmental and human samples (9,12,15), determination of cell cycle parameters (15), growth characteristics (18) and antibiotic sensitivities of bacteria (5,8,19). The primary method used has been quantitative fluorescence. The nonfluorescent parameter of light scatter has been relatively neglected (3,4,10,11). This is the first study that uses light scatter to detect cell injury in viable, nonfixed organisms. Our hypothesis is that early cell injury will be reflected by alterations in surface structure, such as swelling, which will increase the amount of light scatter (3). We examined whether antimicrobials would affect the amount of light scattering by Giardia to a detectable and reproducible degree. First, we compared the parameter of light scatter to two independent measures of cell viability, microscopic observation of motility (6), and fluorescence of a methabolismdependent dye, fluorescein diacetate (FDA) (13). Organisms were exposed to varying concentrations of two different antimicrobials, and had viability assessed by all three methods. Our criteria to accept light scatter as a marker for early cell injury was demonstration of both a threshold and a doseresponse relationship. Next, we compared changes in light scatter to increasing doses of antimicrobial by 1) increasing antimicrobial concentration and 2) increasing the duration of exposure. We report that flow cytometric analysis of light scatter provides a rapid, sensitive, and accurate method of determining cell injury in G. lamblia. Materials and Methods Clinical Specimens Previously described G. lamblia strains WB, LT, and RS were obtained from O. Keister (Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases) (16). Medium Organisms were axenically cultured in filter sterilized TY1S33 medium modified by the addition of bile salts, as described by Keister (7). AntimicrobialAgents Metronidazole (Searle) and furazolidone (NorwichEaton) dissolved in TYIS33 were used at the specified concentrations. Fluorescing Agent Fluorescein diacetate (FDA) stock was prepared by dissolving 5 mg FDA (US Biochemical Corporation) in 1 mL of acetone. The working solution was prepared immediately prior to experimentation by adding 10 mL PBS to 0.05 mL of stock solution. The working solution was kept on ice, in a foilwrapped container to preserve stability. In the indicated experiments, FDA was added precisely 3 minutes prior to analysis, as time was a critical factor in quantitating fluorescence (13). FDA must be metabolized by the cells in order to fluoresce. Therefore, the intensity of the fluorescence is proportional to the uptake of the dye and the metabolic activity of the cell (13). * Corresponding author. Present address: Department of Medicine, University of Maryland, 22 Greene St., Baltimore, MD 21201, U.S.A..
Page 10 TABLE 1. Forward light scatter 24 hour metronidazole exposure dose (µg/mL) Organism
N
0
0.16
0.32
0.64
1.25
2.5
G. lamblia WB
9
Mean scatter ± 50
119 ± 11.9
125 ± 7.4
135 ± 7.2
157 ± 5a
Fluorescence (% control)
100
95 ± 2
84 ± 10
62 ± 3a
182 ± 9a a 49 ± 9
Motility (%)
83 ± 3
80 ± 4
73 ± 6
58 ± 4
23 ± 7
0.32
0.64
180 ± 9.2 24 ± 16 16 ± 6
24 hour furazolidone exposure dose (µg/mL) Organism
N
G. lamblia
9
0
0.08
0.16
2.5
Mean scatter
75 ± 2.5
95 ± 14.5b
125 ± 10
131 ± 14.5
126 ± 11
97 ± 12
Motility (%)
84 ± 4
88 ± 3
82 ± 8
64 ± 3
36 ± 7
14 ± 2
ap < .01 b p < .05
Twotailed ttest
Susceptibility Tests A 0.1 mL inoculum of 105/mL logphase organisms per tube, with 15 mL of medium at the specified drug concentration, was used. The tubes were incubated at 37° C, for 24 hours. In the time series study, however, the specified times are the hours prior to analysis that the tubes were inoculated. The cells were harvested by cooling the culture tubes at 4°C for 20 minutes to release adherent trophozoites from the glass. Immediately prior to analysis, cells were filtered through a 40 µm wire mesh to remove clumps. Microscopy In selected experiments, each tube was inspected at a magnification of 450X at the time of cytometry, to assess cell motility qualitatively. At least four fields of 50 cells were observed per tube. We quantitatively assessed viability for G. lamblia at 24 hours by calculating ''percent motility". In this method, two volumes of cells cultured under the specified conditions, were mixed with one volume of a 0.1% solution of trypan blue and the resulting solution was loaded in a hemacytometer. One hundred cells were examined at 450X magnification. The cells that maintained motility of their ventral flagella and excluded trypan blue were reported as "percent motile" (Wahl, et al., Conference Proceedings). Flow Cytometry Analysis of the prepared cells was performed with a fluorescenceactivated cell sorter (Becton Dickson FACS IV). In this instrument, the organisms, carried by a flow of PBS, pass singly through a viewing chamber where they are illuminated by an argon laser beam. The laser light both excites the metabolized FDA to fluoresce, and is scattered by the cell. The FDA was excited with 300 mw of the 488 nm line from a SpectraPhysics 16445 argon laser. The fluorescent pulses emitted from the organisms were transformed into electrical impulses, and stored by a multichannel pulseheight analyzer (MCA). Fluorescence intensity is expressed as the percentage of fluorescence from exposed cells to fluorescence from cells without exposure to the indicated antimicrobial. The amount of excitation light scattered by each cell is detected through a second filter and stored in the same MCA. We examined forward light scatter, that is, light collected from 8° above and below the horizontal plane (3). Light scatter results are expressed as mean scatter, which are obtained by integrating cells over channels (ranging from the lowest value in channel 1 to the highest value in channel 255) and taking the mean. Cells overrange are stored in channel 255, resulting in a slight underestimate of the true population mean. Statistical Analysis Population means were compared with a tstatistic, accepting a significance level of 0.05. The significance level was corrected for multiple comparisons when appropriate. Results G. lamblia cells exposed to metronidazole and harvested at 24 hours showed a significant increase in mean forward light scatter at a metronidazole concentration of 0.64 µg/mL. At this concentration, there were significant
Figure 1. Mean forward light scatter and mean fluorescence of G. lamblia trophozoites exposed to varying concentrations of metronidazole. Mean fluorescence for metronidazoleexposed trophozoites is fluorescence (trophozoites)/fluorescence (control). Mean fluorescence for control is the fluorescence of unexposed trophozoites of the ith run/lowest fluorescence of unexposed trophozoites (N=9).
Page 11
decreases in fluorescence (62±3%) and motility (58±4%), compared to control (100% and 83±3%, respectively) (Table 1). There is a continued increase in mean light scatter with increasing drug concentration to a maximum of 2.5 µg/mL, after which mean scatter declines. With increasing drug concentration, up to 2.5 µg/mL of metronidazole, there is a continued decline in fluorescence (to 24%) and motility (to 16%) paralleling increasing forward light scatter. Over the dose range zero to 2.5 µg/mL of metronidazole, fluorescence and mean light scatter were significantly, negatively correlated (r= 0.94, p<.001 Figure 1). A similar correlation between mean light scatter and motility was demonstrated for G. lamblia exposed to furazolidone. While the maximal degree of light scatter and initial loss of motility occurred at 0.32 µg/mL of furazolidone, the initial increase in forward light scatter occurred at a concentration of 0.08 µg/mL (Table 1). The initial increase in mean forward light scatter is depicted in serial histograms of G. lamblia exposed to increasing concentrations of furazolidone (Figure 2). These histograms are derived by plotting the number of
Figure 2. Histograms of G. lamblia trophozoites exposed to increasing concentrations of furazolidone. The vertical axis represents the number of cellcounts, while the horizontal axis represents the channel number from 1 (at origin) to 255 (on the right). Cells with greater forward light scatter are placed in the highernumbered channels. The top curve shows the scatter histogram of the control culture (a). Subsequent curves show cultures exposed to 0.1 µg/mL (b), 1 µg/mL (c), and 10 µg/mL (d) of furazolidone, respectively. Increased counts in the very lowscatter channels reflect cellular debris (c & d).
Figure 3. Mean forward light scatter of G. lamblia trophozoites after incubation with different concentrations of metronidazole, for various time periods. Different symbols are used to indicate the difference concentrations of metronidazole. Time (of exposure) is the number of hours prior to analysis that the culture was inoculated.
cells (vertical axis) against their channel number (horizontal axis). With increasing concentrations of antimicrobial, there is an initial shift in the histogram to higher channels, signifying increasing forward light scatter. At higher concentrations (1 µg/mL), the curve becomes bimodal, reflecting low scatter debris and swollen, intact cells. The mean scatter declines, due to the debris in lowscatter channels. At very high concentrations (>10 µg/mL) only low scatter debris is present. Visual inspection of cells corroborates the relationship between light scatter and antimicrobial effect. At 1.25 µg/mL, the metronidazole concentration demonstrating the maximal effect on light scatter, 23±7% of the test cells were motile. At this concentration, there as marked clumping and notable swelling. At 10 µg/mL, only debris and rare ghosts were seen. Antimicrobial susceptibility of G. lamblia is dependent upon both concentration of drug and time of drug exposure (6). We assessed injury after various durations of exposure to concentrations of metronidazole ranging from 0.16 µg/mL to 2.25 µg/mL (Figure 3). At 8 hours, there is a significant increase in mean light scatter at a concentration of 1.25 µg/mL or greater. Mean light scatter is significantly increased after 12 hours of exposure to 0.64 µg/mL of metronidazole. For those concentrations of metronidazole having a significant effect on fluorescence and motility at 24 hours (>0.64 µg/mL), mean forward light scatter is significantly correlated with increasing time of exposure. At a concentration of 2.25 µg/mL, the correlation of light scatter and time of exposure is 0.975. In the control cells, at a concentration of 0.16 µg/mL, the correlation is 0.002.
Page 12
Discussion The close association of changes in forward light scatter with other markers of injury and viability suggest that forward light scatter can be used to detect cell injury in G. lamblia. Compared to the other widely used methods of motility and clonal growth, it is rapid, sensitive, and standardized between observers. Forward light scatter correlated well with two standard measures of viability: metabolismdependent fluorescence and motility. The effective concentrations of metronidazole defined by forward light scatter are also in accord with previous work. Jokipii and Jokipii (6) reported 1060% motility in G. lamblia after 24 hours of exposure to a metronidazole concentration of 0.08 mcg/mL Gillin and Diamond (2), using a clonal growth assay, reported a minimum lethal concentration of 2 µg/mL after 24 hours. In both of those studies, qualitative changes in motility, reflecting altered cell function, occurred prior to the complete cessation of motility (2,6). Forward light scatter provides a simple method to quantify these effects of antimicrobials, early in the process of cell injury. The concentrations of antimicrobials at which a significant change in light scatter was detected after 24 hours of exposure correspond to the minimum lethal concentration after 72 hours of exposure (2). This suggests that the minimum concentration to increase the mean forward light scatter might perform well as a measure of antimicrobial susceptibility. To test this would require comparing forward light scatter with the gold standard of clonal growth. Previous studies have examined changed in DNA as measured by fluorescent dyes (ethidium bromide, mithramycin, propidium iodide) to evaluate susceptibility of bacteria and fungi (5,8,19). None of these studies have examined protozoa. However, fluorescent dyes are expensive, require strict control of lighting conditions, frequently require fixation, and may require precise timing of the dyeexposure. Other investigators have noted changes in light scatter, though only with cells of higher organisms, or with fixed, nonliving microbes. Nash, et al., Used light scatter to separate viable, rodshaped rat ventricular muscle cells from damaged, round cells (10). Huttern and Eipel, in reporting on double fluorescent staining to assess viability of ethanol fixed yeast (Saccharomyces cerevisiae) presented data demonstrating a strong correlation between fluorescence defined viability and light scatter pattern without comment (5). However, ethanol fixation has been shown to alter the light scatter pattern (1). Under appropriate conditions, forward light scatter performs well as a singleparameter to detect cell injury. It is a simple, sensitive, and rapid method to determine the effect of antimicrobials and other agents on protozoa. Forward light scatter can be used as a marker for cell injury in nonfixed cells in conjunction with fluorescent probes of cellular function. Use of light scatter and a single fluorescent probe is less complex than double fluorescence. Further, unlike most fluorescent markers of viability, fixation is not required, permitting examination of cellular physiology. For example, calcium flux in early cell injury and subsequent death could be evaluated with Quin II (14). As quantitative fluorescence is perhaps the most power technique in dual parameter cytometry, the use of forward light scatter as a viability parameter would facilitate more elegant, in vivo evaluation of Giardia physiology. Literature Cited 1. Braylan, R.C., N.A. Benson, U. Nourse, and H.S. Kroth. 1982. Correlated analysis of cellular DNA, membrane antigens, and light scatter of human lymphoid cells. Cytometry 2:337343. 2. Gillin, F., and L. Diamond. 1981. Inhibition of clonal growth of Giardia lamblia and Entamoeba histolytica by metronidazole, quinacrine, and other antimicrobial agents. Antimicrob. Chemother. 8:k305316. 3. Hansen, W.P., R.A. Hoffman, S.H. Ip, and K.W. Healey. 1982. Light scatter as an adjunct to cellular immunofluorescence in flow cytometric systems. J. Clin. Immunol. 2(Suppl.):32s41s. 4. Hoffman, R.A., P.C. Kung, W.P. Hansen, and G. Goldstein. 1980. Simple and rapid measurement of human Tlymphocytes and their subclasses in peripheral blood. Proc. Natl. Acad. Sci. (USA) 77:49144917. 5. Huttern, J.J., and H.E. Eipel. 1978. Advances in determination of cell viability. J. Gen. Microbiol. 107:165167. 6. Jokipii, K., and A.M.M. Jokipii. 1980. In vitro susceptibility of Giardia lamblia trophozoites to metronidazole and tinidazole. J. Infect.Dis. 1412:317325. 7. Keister, D. 1983. Axenic culture of Giardia lamblia in TY1S33 medium supplemented with bile. Trans. Roy. Soc. Trop.Med. Hyg. 77:487488. 8. Martinez, O., H.G. Gratzner, T.I.O. Malinin, and M. Ingram. 1982. The effect of some betalactam antibiotics on Escherichia coli studied by flow cytometry. Cytometry 3:129133. 9. Muldrow, L.L., R.L. Tyndall, and C.B. Fliermans. 1982. Aapplication of flow cytometry to studies of pathogenic free living amoebae. Appl. Environ. Microbiol. 44:12581269. 10. Nash, G., P.E.R. Tatham, T. Powell, V.W. Twist, R.D. Speller, and L.T. Loverock. 1979. Size measurements on isolated rat heart cells using Coulter analysis and light scatter flow cytometry. Biochimica. et Biophysica. Acta 587:99111. 11. Salzmann, G.C., P.F. Mullaney, and B.J. Price. 1979. Light scattering approaches to cell characterization, pp. 105124. In: Melamed, M.R., R.F. Mullaney, and M.L. Medelsohn (eds.), Flow Cytometry and Sorting, John Wiley and Sons, New York. 12. Saul, A., P. Myler, T. Mangan, and C. Kidson. 1982. Plasmodium falciparum: automated assay of erythrocyte invasion using flow cytofluoromety. Exp. Parasitol. 54:6471. 13. Serentz, M. 1973. Fluorescein diacetate. In: Thaer A.A. and M. Serentz (eds.), Fluorescent Techniques in Cell Biology, SpringerVerlag, Heidelberg. 243. 14. Shev, S.S., V.K. Sharma, and S.P. Banergee. 1984. Measurement of cytosolic free calcium concentration in isolated rat ventricular myocytes with Quinn II. Circulation Res. 55:830834. 15. Skarstad, K., H.B. Steen, and E. Boyle. 1983. Cell cycle parameters of slowly growing Escherichia coli B/r studied by flow cytometry. J. Bacteriol. 154:656 662. 16. Smith, P., F. Gillin, N. Kawhal, and T. Nash. 1982. Antigenic analysis of Giardia lamblia from Afghanistan, Puerto Rico, Ecuador, and Oregon. Infect. Immun. 36:714719.
Page 13
17. Steen, H.B. 1980. Further developments of a microscopebased flow cytometer: light scatter detection and excitation intensity compensation. Cytometry 1:2631. 18. Steen, H.B., and E. Boyle. 1980. Escherichia coli growth studied by dualparameter flow cytometry. J. Bacteriol. 145:10911094. 19. Steen, H.B., E. Boyle, K. Skarstad, B. Bloom, T. Godal, and S. Mustafa. 1982. Applications of flow cytometry on bacteria: cell cycle kinetics, drug effects, and antibody binding. Cytometry 2:249257. 20. van Dilla, M.A., R.G. Langlois, D. Pinkel, D. Yajko, and W.K. Hadley. 1983. Bacterial characterization by flow cytometry. Science 220:620622.
Page 15
The Importance of Nonwaterborne Modes of Transmission for Giardiasis A Case Study V. Susan Harley Sturgeon Health Unit, Box 174, 23 Sir Winston Churchill Avenue, St. Albert, Alberta T8N 1N3, Canada It is a common assumption that outbreaks of giardiasis are most frequently associated with the waterborne mode of transmission. Recent studies on outbreaks of giardiasis have shown that other modes of transmission are equally as important. Fecaloral transmission whether foodborne or person to person is as frequently implicated as the water route. This was shown during a recent investigation of a giardiasis outbreak at an Alberta Hutterite colony. The infection rate at the colony was 37% of the population (76 members). The field investigation resulted in waterborne transmission being eliminated as the most probable route. Fecaloral transmission via person to person contact appeared to be the most probable mode of transmission. In addition, deficiencies in food service and laundry facilities at the colony may have also contributed to the outbreak. Control procedures included antiprotozoal treatment of infected individuals, as well as education in food handling practices and enteric disease transmission. Within a three month period a 90% recovery rate was achieved.
Introduction The field investigation of an outbreak of giardiasis on an Alberta Hutterite colony revealed that while the initial source of the infection could not be isolated, several modes of transmission were found which facilitated the spread of the infection. The nature of the water supply precluded waterborne transmission. The fecaloral route via person to person and foodborne transmission is believed to have been most important. The purpose of this paper is to review the modes of transmission of giardiasis and to examine the relative importance of nonwaterborne modes within the context of this field investigation. In particular, parallels between the communal lifestyle of a Hutterite colony and institutional settings within our North American society are examined to highlight the need for additional research into prevention and control of this increasingly important enteric infection. Modes of Transmission Benenson (1) describes three modes of transmission for giardiasis, namely waterborne, foodborne and person to person. 1. Waterborne Transmission occurs when water contaminated by Giardia cysts is consumed. Two risk factors associated with waterborne transmission are the source of the water and the method of treatment used. Illnesses have occurred where unfiltered water from surface sources (3,10) has been consumed. In British Columbia, sixtynine cases of giardiasis were traced to consumption of a municipal water supply which derived its raw water from a creek (3). The water supply was unfiltered. In Alberta, two localized outbreaks have been linked to municipal water supplies. In Banff, the link between the water reservoir and the surface water source is believed to have been responsible for the outbreak in 1982 (Epidemiologic Notes and Reports, Alberta Social Services and Community Health, 1983). In Edmonton, deficiencies in the municipal water treatment facilities are believed to be the probable cause of an outbreak in 1983 (Logsdon, G.S. Evaluation of Monitoring and Treatment for Protozoan Pathogens, Edmonton Water Supply Evaluation, Edmonton, 1986). 2. Foodborne Transmission occurs when fecally contaminated food is consumed. While Giardia cysts do not replicate in food, they may survive on moist foods which are not cooked prior to eating. Osterholm (5) traces twentynine cases of giardiasis resulting from the ingestion of canned salmon. He hypothesizes that conditions within the food's environment may enhance the infectivity of Giardia cysts. 3. Person to Person Transmission occurs when the cysts from the feces of an infected individual are directly or indirectly transferred to another person. Sartwell and Last (7) state that the three most obvious risk factors in person to person transmission are the population density, hygienic conditions and the proportion of susceptible individuals in the population. These risk factors interact to facilitate the rapid spread of infection throughout institutional and household settings. In child care and residential institutions, the frequency of interpersonal contact within a susceptible population coupled with poor environmental and hygienic conditions contribute to the spread of disease (9). Keystone's (4) report on person to person transmission of giardiasis in day care nurseries highlights the susceptible population as young children who are mobile but not yet toilettrained or educated in personal hygiene. In Alberta in 1984 one third of the cases of giardiasis were under five years of age with 31% between one and four years (Epidemiologic Notes and Reports, Alberta Social Services and Community Health, 1985:271273). Household contacts are a
Page 16
highly susceptible portion of the population for person to person transmission. Polis (6) suggests in his report of the spread of giardiasis from a day care to the community that "47% of the children in the study group transmitted the disease to at least one household contact". In Alberta in 1984, almost 30% of giardiasis cases had at least one other case within the family and household contacts. Case Study Introduction The Hutterite colony is located in a rural farming area northwest of Edmonton. At the period of the investigation (June to September, 1985) there were fifteen families living on the colony with a total population of seventysix people ranging in ages from four months to eightytwo years. The facilities of the colony will be discussed as they existed at the time of the investigation. Changes made since that time will be discussed in the postscript. While each family maintains separate living quarters, many activities are conducted communally. The food and laundry facilities are located in one central building. A kindergarten which is run by the mothers of the colony cares for children three years to school age. A school supported by the local school district has programs for children up to age fifteen. The colony livelihood is based upon a mixed farming operation with hogs, cattle, poultry, vegetables and grain. Sales of poultry from an onsite poultry abbatoir and vegetable sales at local farmer's markets provide additional income. The earliest documented case of giardiasis at the colony occurred in June, 1983. A second case, the wife of the first case, was diagnosed in October, 1984. The second case was the head cook for the colony. She required two courses of treatment with metronidazole before negative stool samples were obtained. Although stool samples from family members were requested, there is no record of samples being submitted to the Provincial Laboratory of Public Health. On June 24, 1985 the Sturgeon Health Unit received notification of a case of giardiasis in a ten month old male. During the communicable disease interview with the child's mother on June 25, 1985 it became apparent that a number of individuals on the colony were experiencing symptoms of abdominal cramps, bloating and diarrhea. Methods To assess the extent of the outbreak, it was decided to collect stool specimens from all colony members, both asymptomatic and symptomatic. Stool kits containing formalin solution were distributed to each family. A family tree supplied by the preacher was used to develop a case record and all members were listed according to family group and birthdate. Because of some duplication in Christian names plus the use of a common last name, everyone was requested to supply birthdates on the stool requisition form. Although it is normal procedure to request three stool samples per person, it was decided by the Provincial Epidemiologist that one sample per person would facilitate both the collection on the colony and identification by the Provincial Laboratory of Public Health. Initially, considerable resistance to the submission of samples was encountered. As the scope of the outbreak became apparent, most people complied with the request. Of the seventysix members, seventy people supplied at least one sample. The onsite investigation focussed on the water supply, sewage disposal system, food and laundry service and person to person transmission as the possible sources of infection and modes of transmission. Results The colony water is supplied from an unfiltered, untreated deep well located north of the building site. Routine water samples have been submitted from various locations throughout the colony since 1984. All twelve samples submitted in the seventeen month period to the end of August, 1985 had nil fecal and total coliform counts. Water from deep wells is not considered to be as likely a source of giardiasis as water from surface water supplies. This would suggest that the possibility of contamination of the water by Giardia cysts was low, however, laboratory verification of Giardia cysts was not feasible at that time. Every dwelling unit plus the kindergarten and school is equipped with water closets and lavatories. The poultry abbatoir and kitchen have full hot and cold water service. The final disposal of all effluent is into a large natural lagoon located west of the building site and approximately seventyfive feet below the abbatoir. Apparently, there have not been any problems with sewage blockages or backups in the dwelling units. There have been recurring problems with the grey water from the kitchen backing up into the kitchen sinks. To rectify this problem, a new septic tank was installed in early summer, 1985 to handle kitchen effluent prior to final discharge into the lagoon. The effluent disposal from the abbatoir was upgraded during the winter of 1985. Previously, the effluent from the slaughtering operation, including blood, feathers and viscera drained directly down the slope into the lagoon. The upgrading included the installation of covered floor gutters and a settling tank. Because of the mixed farming operation, there is constant contact with animal manure, especially by the men. There are two duck ponds, one to the south of the animal barns and one to the north at a lower elevation from the building site. Additionally, several sheep graze among the dwelling units. The sewage disposal does not differ greatly from many similar large mixed farming enterprises. The possibility of contact with either human or animal manure is high. The food and laundry facilities share the area known as the kitchen. While each dwelling unit has a small kitchen area with sinks, water and a small hot plate, most food service is conducted communally in the kitchen. The food is prepared and served by the women of the colony according to a weekly duty roster. Three daily meals are served "family style" in the adjoining room. The young children are served in their homes prior to the adult meal time. The kitchen is primarily the food preparation area with a large commercial grill and pressure cooker, as well as preparation tables. The food storage area is in a separate building located north of the kitchen. This building contains a large freezer room where most products are
Page 17
stored. Milk and dairy products are kept in a large commercial refrigerator in the dining hall. During the investigation of the food service, deficiencies were noted in cold food storage, dishwashing method and hygiene. Perishable products were often stored at room temperature. Dishwashing was being performed manually in two sinks located back to back. No sanitizers were used in the rinse water. There were no separate handwashing facilities available either for food handlers, or the colony members who arrived for meals directly from other areas of the colony. The second case of giardiasis as previously stated was the head cook. At the time of that investigation, it was suggested that she refrain from food handling. She was not able to conform with this suggestion. During the outbreak investigation, it was suggested again that all symptomatic or positive asymptomatic persons be excluded from food handling. There is a cultural barrier which does not permit compliance with this suggestion. The laundry facilities are located at the west end of the kitchen. There are two large commercial front loading washing machines and a water extractor. Drying is accomplished by outdoor clotheslines. During observations made during several laundry sessions, two problems became obvious. The washing machines leaked or were interrupted by opening the door during operation. This caused wash water to run all over the kitchen floor. The second potentially more serious problem was the use of the large dishwashing sinks as laundry sinks. Clothing was rinsed in the sink prior to being placed in the extractor. The suggestion was made to cease usage of the dishwashing sink for laundry rinsing until improved arrangements could be installed. The potential for transmission of the disease through contamination of the food service area from soiled water and clothing would appear to be very high. Several recommendations were directed at these problems. Interpersonal contact is very frequent among the members of the colony and with people from outside of the colony. Within the colony, the communal food service, laundry facilities, kindergarten and general farming operations provide frequent daily contact with all members. The men work and eat together, while the women and children spend a considerable portion of their day in shared duties or communal care. The colony members frequently visit their relatives and friends in other colonies throughout Alberta and Montana. Social events such as weddings are occasions for long visits. During June, 1985 one wedding was held on the colony and three others were attended by many colony members on colonies throughout Alberta. Often only the adults attend these weddings while their children are cared for by another family. During conversation with various individuals, they mentioned that similar symptoms were being experienced on other colonies that they frequently visit, or whose members frequently visit. There were reports of giardiasis on other Alberta Hutterite colonies during this time. The possibility of the disease being introduced into the colony by visitors or being spread to other colonies by members of this colony appear high. TABLE 1. Case distribution by age and sex Age
Male
Female
011 months
2
1
Total 3
14 years
7
3
10
59 years
4
1
5
1014 years
0
1
1
1519 years
0
0
0
2024 years
1
1
2
2529 years
3
0
3
3039 years
1
1
2
4059 years
1
1
2
60+ years
0
0
0
TOTAL
19
9
28
Of the stool specimens submitted by seventy colony members, twentyeight were positive for Giardia lamblia. The case distribution is shown in Table 1. Sixtyeight percent of cases occurred in the under fifteen age category with 37% in the under five age group. Males accounted for 68% of cases in the under fifteen age group. This is understandable in light of the age and sex distribution of the colony population. In the under fifteen age category, males account for 70% percent of the group. Overall, 43% of the population is under fifteen years with only 10% over fortyfive years. The age fifteen was chosen because that is the age at which children leave school and begin to participate in the colony duties. The clustering of cases in the lower age groups corresponds to previously reported infection patterns. The percentage of symptomatic cases was 61% overall with 39% reporting no significant symptoms. Because of the difficulties interviewing the men, who are normally working away from the main area, as well as the mildness of symptoms experienced by many people, the information relating to symptoms and the rate of symptomatic/asymptomatic cases is unreliable. Of the people reporting symptoms, the most common were bloating, diarrhea, abdominal cramps and fatigue. The most severe symptoms were experienced by the children. The control procedures included treatment of all cases with metronidazole and education in several facets of communicable disease control. No side effects to the medication were reported although several children encountered problems swallowing the pills. None of the female cases were pregnant, therefore treatment problems in the adult population were not encountered. Stool samples were requested after the completion of the medication. Three cases did not respond to the initial treatment and were prescribed a second course of medication. Two of these three involved children who had difficulty in swallowing the pills and may not have received the full amount required. The medication was distributed by the colony representative when she received positive results from the public health inspector. There was a time lag of approximately 10 days between the submission of stool samples and receipt of results.
Page 18
Discussion The outbreak of giardiasis at the Alberta Hutterite colony was identified in late June, 1985. The investigation, control and treatment procedures were conducted over the summer months with all twentyeight cases treated by mid September, 1985. The most probable source of the outbreak was not identified. The extended time period which had elapsed since the previous cases on the colony coupled with the fact that no one in that family was identified as a case or carrier in this investigation, would suggest that they were not involved as a source of infection. The most probable mode of transmission would appear to be person to person contact between both colony members and visitors from other colonies. The deficiencies noted in the food service area may also have been implicated in the disease transmission. The attack rate of 37% is significantly higher than would be expected. Because most persons infected were identified as a part of the investigation, the information relating to the onset and intensity of symptoms was not reported reliably. The symptoms reported were similar to those normally experienced. The attack rate and severity of symptoms were higher in the below fifteen age category. This finding is consistent with the results of other outbreaks where children cared for in a day care or extended family environment appear to be at risk in acquiring the infection. The outbreak required approximately three months to bring under control with all positive cases being identified and treated within that time frame. Concern was expressed that the colony members' selfadministration of the medication may have prolonged the period of communicability. The results of the followup stool samples showed a 90% recovery rate with two of the three positive cases having encountered treatment difficulties. Because of the risk of a reoccurrence of communicable disease at the colony, recommendations were made in those areas where deficiencies were noted and where it was felt that improvements would be most effective. It was recommended that the laundry facilities be relocated away from the kitchen and that the machines be repaired in the interim. Food handling recommendations included the use of a sanitizer in the rinse water and the adoption of a policy to excuse ill individuals from food handling duties. It was also recommended that a separate handwash area be established in the kitchen for the use of both food handlers and colony members. An educational seminar on the prevention and control of communicable disease was proposed for all members of the colony. At the health unit level, the recommendation was made to develop an educational program suitable for adults on the prevention and control of communicable disease. Continued surveillance of the prevalence of giardiasis on Alberta Hutterite colonies was also recommended. Postscript One year later, it is encouraging to report that several recommendations have been implemented. A new laundry facility has been constructed well away from the kitchen. At this writing, one further case of giardiasis has been reported from the colony; no family contacts were positive during screening. Conclusions Can the experience gained from the investigation of an outbreak of an enteric infection in the unique environment of an Alberta Hutterite colony be generalized to mainstream North American society? If we consider a colony to be a large extended family, we can appreciate the increased risk of disease to its members. There are numerous opportunities for disease transmission through the close interpersonal contact and communal food, laundry and child care service. In this context, we should consider again the risk factors previously discussed for person to person transmission. The frequency of interpersonal contact, the proportion of susceptible people in a population where 24% percent are under the age of five years, and the combination of lower levels of environmental sanitation and understanding of personal hygiene all interact to create a high risk environment for the spread of enteric illnesses. These risk factors are present in child care and residential institutions. To appreciate the magnitude of the problem, recent estimates by Alberta Social Services suggest that 10% of preschool children are being cared for in formal childcare care facilities (Alberta Social Services Day Care Branch, Personal Communication, Edmonton, 1986). This percentage rises sharply when those children being cared for in informal babysitting arrangements are included. Controlling the incidence and transmission of enteric as well as other communicable diseases in these facilities has emerged as a public health priority. In recent reports of outbreaks of enteric infections in day care centres, several methods of control have been examined. Black's (2) study on handwashing suggests that a supervised handwashing program could reduce the incidence of enteric infections in day care centres. A study on the control of shigellosis in day cares strongly supports the isolation of cases within the facility as opposed to exclusion or facility closure (8). If a facility is closed or a child excluded, parents may be forced into making alternate day care arrangements through which the disease may be introduced and spread to the community. Does the increase in incidence of communicable disease in child care institutions warrant raising their standards to those of a small hospital with trained nursing staff, isolation rooms and an infection control committee? While such meaures may be uneconomical, the trend towards preschool child care, particularly infant and toddler care, coupled with the ease of disease transmission from the institutional environment to the household and community, demands public health intervention. The delivery of educational programs designed to teach the nature, prevention and control of communicable diseases as well as contingency plans for handling such situations are within the financial and personnel resources of existing public health programs. What is required is a complete understanding of the mechanisms of disease
Page 19
transmission. The case study of the Alberta Hutterite colony suggests several possible modes of transmission of which person to person transmission was the most important. Research into giardiasis which focuses on the importance of nonwaterborne modes of transmission would be of great assistance to field epidemiologists and health education program designers. Literature Cited 1. Benenson, Abram S. 1985. In: Control of Communicable Diseases in Man, 14th ed. The American Public Health Association, Washington, D.C. 2. Black, R.E., A.C. Dykes, K.E. Anderson, J.G. Wells, S.P. Sinclair, G.W. Gary, M.H. Hatch, and E.J. Gangarosa. 1982. Handwashing to prevent diarrhea in day care centres. Am. J. Epidemiol. 113:445451. 3. Fogel, D.A. 1982. Waterborne giardiasis and the role of the health inspector: A case history. Environ. Health Review, 2629. 4. Keystone, J.S., S. Krajden, and M.R. Warrne. 1978. Person to person transmission of Giardia lamblia in day care nurseries. Can. Med. J. 119:241248. 5. Osterholm, M.T., J.C. Forfang, T.L. Ristinen, A.G. Dean, J.W. Washburn, J.R. Godes, R.A. Rude, and J.G. McCullough. 1981. An outbreak of foodborne giardiasis. The New England J. Med. 304:2428. 6. Polis, M.A., C.U. Tuazon, D.W. Alling, and E. Talmanis. 1986. Transmission of Giardia lamblia from a day care centre to the community. Am. J. Public Health 76(9):11421144. 7. Sartwell, P.E., J.M. Last. 1980. Epidemiology, p. 70. In: Last J.M., (ed.), MaxcyRosenau Public Health and Preventative Medicine, 11th ed. Appleton CenturyCrofts, New York. 8. Tauxe, R.V., K.E. Johnson, J.C. Boase, S.D. Helgerson, and P.A. Blake 1986. Control of day care shigellosis: A trial of convalescent day care in isolation. Am. J. Public Health 76(6):627630. 9. Thacker, S.B., S. Simpson, T.J. Gordon, M. Wolfe, and A.M. Kimball. 1979. Parasitic disease control in a residental facility for the retarded. Am. J. Public Health 69(12):12791281. 10. Weniger, B.G., M.J. Blaser, J. Gedrose, E.C. Lippy, and D.D. Juranek. 1983. An outbreak of waterborne giardiasis associated with heavy water runoff due to warm weather and volcanic ashfall. Am. J. Public Health 73(8):868871.
Page 21
A New Miniculture Technique for determining In Vitro Antimicrobial Agent Sensitivity of Axenically Cultivated Strains of Giardia lamblia Stephen M. Wahl, Robert H. Gilman*, Jane P. O'Hare, David B. Keister, and William M. Spira Division of Geographic Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland We developed an assay in which the in vitro sensitivities of Giardia lamblia to antimicrobial agents can be determined. We used a miniculture technique, in which adherence of trophozoites to the culture plate was the criterion for viability. Measurement of adherence permits one to determine the adhesive activity of aggregates of G. lamblia rather than confining determination of viability to individual trophozoites. The adherence assay was validated by comparing its results to those of the standard assay for in which ventral flagella motility of G. lamblia is the criterion for viability. Five difference strains of G. lamblia tested had the same sensitivity to metronidazole, while one (the CAT strain) was slightly more resistant to quinacrine than the other four. An evaluation of sensitivity to various antimicrobial agents demonstrated that 80S ribosomal inhibitors (such as cycloheximide and anisomycin) and furazolidone and nalidixic acid (which like metronidazole and quinacrine affect DNA synthesis) are effective in vitro against G. lamblia.
Introduction Recent developments in techniques for the in vitro cultivation of Giardia lamblia (1) have studied the parasite's susceptibility to antimicrobial agents. Earlier methods for investigating G. lamblia susceptibility were tedious or time consuming. In the clonal method of Gillin (2), susceptibility is measured by exposing cells to an antimicrobial agent for 72 hours, incubating them for an additional 72 hours in nutrient agar, then counting the number of visible colonies. In the standard motility assay by Jokipii and Jokipii (3), the effects of antimicrobial agents on G. lamblia trophozoites are determined by direct microscopic examination of individual trophozoites: after each 24 hours of exposure, 100 randomly selected G. lamblia are examined for ventral flagella motility, the criterion for viability. In developing a new in vitro miniculture technique, our goal was to avoid the tedium of the Jokipii and Jokipii method and the long waiting period of the Gillin method while keeping the sensitivity of these procedures. In the current study, we have validated a microculture adherence assay for viability (a modified version of the Jokipii and Jokipii method), and used it to examine the susceptibility of G. lamblia to different antimicrobial agents and to examine variations between different G. lamblia strains to the same antimicrobial agents. Materials and Methods Giardia lamblia Trophozoites Five previously characterized strains (4,5) of G. lamblia (WB, PO, LT, RS, and CAT) were maintained in axenic culture. Media All cultures were grown in a modified Diamond's TYIS33 medium described by Keister (4). After the medium, including serum, was heated in at 56°C water bath for 30 minutes, penicillin and streptomycin were added. The medium was then filtersterilized. All medium was prepared one day prior to use to avoid erratic results due to differences in the oxidation of cysteine. Gillin has demonstrated that the inclusion of penicillin and streptomycin in the medium does not effect antimicrobial agent susceptibility testing (2). Antimicrobial Agents The antimicrobial agents tested are listed in Table 1. All drugs were reconstituted in sterile medium. Culture of G. lamblia Trophozoites and Preparation of Trophozoites for Assays The trophozoite cultures were grown standing upright for 72 hours at 37°C in 15 mL (11 × 150 mm) screwtop borosilicate tubes which were filled with 14 mL of medium (type mentioned above). After 72 to 96 hours of incubation, the culture was examined microscopically at 100x magnification to verify that the trophozoites had grown successfully. Successful growth was defined as a layer of trophozoites completely or almost completely covering the inner wall of the culture tube. Trophozoites were released from the walls of the tube by incubating the tube in a 4°C water bath for 15 minutes and then gently agitating the tube to evenly suspend the trophozoites. Adherence Assay Each well of a 24 well (1.7 × 16 cm) flat bottom tissue culture plate (LinbroT M) was filled with 2 mL of medium containing varying concentrations of antimicrobial agent and inoculated at the same time with 0.1 mL of medium containing between 70,000 and 80,000 log phase G. lamblia trophozoites. Each plate included a negative control of G. lamblia trophozoites and medium alone and a positive control containing trophozoites incubated in medium containing 30 µg/mL of metronidazole. Plates were incubated anaerobically (Gas PakR System; BBL) at 37°C and examined at 24, 48, and 72 hours. Prior to examination, the plates were covered with Parafilm to avoid leakage from well to well. The plates were gently agitated in order to suspend nonadherent (i.e. dead) trophozoites into the medium. The percentage of adherence, an estimate of the percent of the total surface area of the bottom of each well that was covered with adhering G. lamblia trophozoites, was then recorded. Wells were coded and read blindly for percentage of adherance at 100x magnification. Each test was run with at least six replicates of each drug concentration. * Corresponding author.
Page 22
A replicate set of plates was made to be read at each of 24, 48, and 72 hours. After each reading, the set of replicate plates was discarded, and a new set read at the next period. Standard Viability Assay In order to validate the adherence assay, we compared its results with those obtained using a modified version of Jokipii and Jokipii's motility assay (3). Trophozoites were considered viable if they fulfilled the following two criteria: the ventral flagella were motile and Trypan Blue was excluded from entering the trophozoite. Validation studies were performed using the same plates as used for the adherence assay. After the percentage of adherence was recorded for each well, the plates were cooled at 4°C for 3040 minutes (or until the trophozoites released themselves from the plate walls) and then gently agitated to produce homogeneous suspension of trophozoites. An aliquot of suspension was taken from each well and mixed in a ratio of 2 volumes of cell suspension to 1 volume of 0.1% solution of Trypan Blue. A portion of this mixture was loaded on a hemacytometer in order to count the total number of trophozoites and the total number of viable trophozoites at 450x magnification. Results of the adherence and viability assays were compared after 24, 48, and 72 hours of incubation at 37°C. Preliminary studies were done to prove that trophozoites (nonadherant) floating in wells containing antibiotic media were not able to reproduce again under standard culture conditions. Conversely, we also demonstrated that adherent cells were able to reproduce normally. Plates incubated at 72 hours with varying concentrations of either metronidazole or quinicrine, or with medium alone were used. The plates were gently agitated and 0.3 mL of medium was aspirated from each well (avoiding contact with the bottom or side of the wells) and inoculated into 3.5 mL of fresh medium. The concentration of any antibiotic remaining in the medium was thus diluted at least tenfold. The vials were then incubated in the upright position for 72 hours at 37°C after which they were examined microscopically for evidence of regrowth. Regrowth was considered present if over 75% of the inner wall of the vial was covered with a complete or almost complete layer of trophozoites. The viability of adherent trophozoites was determined after the medium containing loose cells was aspirated from the plates used in the above test. Fresh medium was introduced and the plates were cooled to 4°C and held for 3045 minutes in order to release the adherent trophozoites. The medium was then aspirated and cultured as described above. In all cases, adherent trophozoites proved viable by culture and floating trophozoites in medium containing greater than an IC50 of an antimicrobial were nonviable. Floating trophozoites in medium control wells or in wells with low concentration of antimicrobial agents were often culturable and probably reflected sloughing due to crowding of the adherent population. Results Validation of the Adherence Assay The plots of the metronidazole dose response curve determined by the adherence assay and by the standard viability assay are shown in Figures 1 and 2, respectively. In order to compare the two assays, the results of each time period were fitted to a linear regression. No significant differences were seen (by analysis of variance) between the dose response curves determined: 1) at different times for the same assay; or 2) between the different assays for the same time period. The results of the two assays were highly correlated (r=0.94, p<.002 at 48 hours) as shown in Figure 3. Unless noted, all further results will be expressed as 48 hour readings. The reproducibility of the adherence assay was very good; the variation between replicate wells was less than 7.3% and the variation between end points between tests was less than 12.6%.
Figure 1. Susceptibility of G. lamblia (WB) trophozoites to varying concentrations of metronidazole at 24, 48, 72 hours using the standard viability assay.
Figure 2. Susceptibility of G. lamblia (WB) trophozoites to varying concentrations of metronidazole at 24, 48, 72 hours using the adherence assay.
Figure 3. Susceptibility of G. lamblia (WB) trophozoites to varying concentrations of metronidazole at 48 hours: Comparison of results in adherence assay to the results in the standard viability assay.
Page 23 TABLE 1. Comparison of antimicrobial agents as to their effectiveness on G. lamblia trophozoites. The parameter used to compare difference antimicrobial agents as to their effectiveness on G. lamblia was the 50% Inhibitory Concentration or IC50. (µg/mL). High Susceptibility (< 5 µg/mL)
Intermediate Susceptibility (5 to 50 µg/mL)
Low Susceptibility (50 to 500 µg/mL)
Negligable Susceptibility (> 500 µg/mL)
cycloheximide < 0.0625
actinomycin D 8 10
minocycline 50 60
erythromycin > 800
quinacrine HCl 0.09 0.12
nalidixic acid 15 20
paromomycin 90
azosulphamide > 800
anisomycin 0.125 0.25
tetracycline 90 100
bacitracin > 800
metronidazole 0.3 0.75
chlorotetracycline 90 100
neomycin > 1000
furazolizone 0.75 1.5
doxycycline hyclate 100 150
bycozomycin > 1000
rifampicin* > 125
clindamycin > 100
chloramphenicol 137 250
TMP/SMX** 75/375
polymyxin B sulfate 400 500
streptozoticin 400
* The highest concentration tested. ** Trimethoprim/sulphamethoxazole were combined in the ratio 1/5.
Effectiveness of Antimicrobial Agents The parameter used to compare different anitmicrobial agents as to their effectiveness on G. lamblia was the 50% inhibitory concentration, also called IC50. IC50 is defined as the concentration of antimicrobial agent at which the percentage of adherence reading is 50% of the percentage of adherence reading of the negative control. All IC50s were calculated from the percentage of adherence seen in G. lamblia after 48 hours of exposure. As shown in Table 1, G. lamblia was highly susceptible to quinacrine HCl, metronidazole, and furazolidone. In addition to these common antigiardial drugs, cycloheximide and anisomycin showed inhibitory effects at the same or even lower concentrations. G. lamblia also showed intermediate susceptibility to actinomycin D and nalidixic acid. Strain Differences Five G. lamblia strains were tested for differences in sensitivity to metronidazole and quinacrine HCl (Figure 4). For metronidazole, no meaningful differences in IC50. were observed, but the CAT strain was significantly less susceptible to quinacrine (IC50. = 0.185 µg/mL) compared to the others (IC50.=0.10 µg/mL).
Figure 4. Comparison of susceptibilities of G. lamblia strains to metronidazole and quinacrine at 24, 48, 72 hours: 50% Inhibitory Concentration (IC50.) in the adherence assay. Sample variances were negligible for all strains (n = 18).
Discussion The in vitro miniculture adherence assay for determining G. lamblia susceptibility to antimicrobial agents showed a high correlation with the standard method over a number of agents, strains and test conditions. The adherence assay appears to be as sensitive as Gillin's clonal method or Jokipii's motility assay yet offers results in onethird of the time of the former (48 hours to 144 hours) respectively, with much less tedium than the latter. A prominent characteristic of living G. lamblia trophozoites is adhesion to surfaces. Adhesion is probably caused by the flow generated by the continuing activity of the ventral flagella which produces a suction pressure under the ventral disc 3 (7,8). Adherence thus is synonymous with ventral flagella motility, the criteria which Jokipii used to judge viability in his assay. This, of course, means it is not very common for dead trophozoites to adhere passively to culture plates. The use of adherance permits one to examine viability by measuring the adhesive activity of aggregates of G. lamblia rather than confining determination of viability to individual trophozoites. We used the adherence assay to determine G. lamblia susceptibility to several different antimicrobial agents. We examined geographicallydistinct axenicallycultured strains for their sensitivity to antimicrobial agents. No differences were seen in the sensitivity to metronidazole between the five strains. After 48 hours exposure, there was seen a significant decrease in the sensitivity of the CAT strains to quinacrine HCl compared to the other four strains, but this difference was less than 2fold. In contrast, Jokipii using nonaxenically cultured strains found up to a 4fold variation in antimicrobial agent sensitivity patterns at 48 hours. This difference may be caused by differences in our techniques for testing antimicrobial agent sensitivity as well as the possible differences in strains tested. We used well established axenic culture strains whereas the Finnish group used newly isolated strains.
Page 24
G. lamblia was susceptible to anisomycin and cycloheximide, both of which are thought to act by binding to the 80S ribosome and interfering with the translation of peptidyl +RNA (9,10). Emetine, which is also thought to bind to the 80S ribosome, has also been reported to be an effective antigiardial agent when tested by the clonal method (2). G. lamblia was also susceptible to antimicrobial agents which are believed to affect the synthesis or structure of DNA. These include quinacrine HCl, metronidazole, furazolidone, actinomycin D, and nalidixic acid. Of these, metronidazole, quinacrine HCl, and furazolidone are the current agents for use against giardiasis since they are comparatively nontoxic. Nalidixic acid was shown in this study to be a potentially useful antigiardial agent. Due to its pharmokinetics, however, its clinical use is unlikely. Drugs such as minocycline, tetracycline, chlorotetracycline, and doxycycline (all of which are believed to inhibit protein synthesis on the 70S ribosome along with some interaction on the 80S) showed only small effects on G. lamblia. We were not able to confirm Gillin's findings (2) that bacitracin was an effective antigiardial agent since it demonstrated no inhibition of growth in our experiments even at concentrations as high as 1000 µg/mL. Our data suggest that, in addition to the traditional antigiardial drugs (the DNA inhibitors: metronidazole and quinacrine), G. lamblia is especially sensitive to drugs which bind solely to the 80S ribosome, and is not affected by drugs which bind solely to the 70S ribosome. Current drugs used for the therapy of G. lamblia are expensive and have significant effects. Better tolerated agents would indeed be welcome. One approach to this problem may be to search for antimicrobial agents which bind to the 80S ribosome that would affect only protozoal eucaryotic cells. Literature Cited 1. Keister, D.B. 1983. Axenic culture of Giardia lamblia in TYIS33 medium supplemented with bile. Trans. Roy. Soc. Trop. Med. and Hyg. 77:487488. 2. Gillin, F.D., and L.S. Diamond. 1982. Inhibition of clonal growth of Giardia lamblia and Entamoeba histolytica by metronidazole, quinacrine and other antimicrobial agents. J. Antimicrob. Chemother. 8:305316. 3. Jokipii, L. and A.M. Jokipii. 1980. In vitro susceptibility of Giardia lamblia trophozoites to metronidazole and tinidazole. J. Inf. Dis. 141:317325. 4. Meyer, E.A. 1970. Isolation and axenic cultivation of Giardia trophozoites from the rabbit, chinchilla, and cat. Exp. Parasitol. 27:179183. 5. Smith, P.D., F.D. Gillin, N.A. Kaushal, and T.E. Nash. 1982. Antigenic analysis of Giardia lamblia from Afghanistan, Puerto Rico, Ecuador, and Oregon. Infect. Immun. 36:714719. 6. Zar, J. 1974. Biostatistical Analysis. Prentice Hall, Engelwood Cliffs, N.J. 228 pp. 7. Holberton, D.V. 1973. Fine structure of the ventral disk apparatus and the mechanism of attachment in the flagellate Giardia muris. J. Cell. Sci. 13:1141. 8. Holberton, D.V. 1974. Attachment of Giardia A hydrodynamic model based on flagellar activity. J. Exp. Bio. 60:207221. 9. Franklin, T.J., and G.A. Snow. 1978. Biochemistry of Antimicrobial Action. 2nd Ed. Chapman and Hall, London. 10. Gilman, A.G., L.S. Goodman, and A. Gilman. 1980. The Pharmacological Basis of Therapeutics. 6th Ed. Macmillan Publishing Co., Inc. New York, N.Y.
Page 25
Ultrastructural Study of a Bacterial Symbiont of Giardia lamblia S. Radulescu, E.A. Meyer*, B. Burghelea, and T. Meitert Cantacuzino Institute, Bucharest, Romania We report here the presence of bacterial endosymbionts in Giardia cysts from humans. Subjects of this study were 10 preschool children with Giardia infections. Five of these children received a live Shigella vaccine, produced in Romania. The other 5 children were not vaccinated. TEM examinations of the cysts of the control group failed to reveal endosymbionts. Cysts from the vaccinated children revealed bacterialike endosymbionts. These endosymbionts were similar to those described by Nemanic et al. (12) in G. muris, and are apparently bounded by two unit membranes. The possible significance of this association between Giardia and Shigella are discussed. We believe that this is the first report of a bacterial symbiont in Giardia lamblia.
Introduction Giardia, perhaps the earliest described intestinal protozoan, has recently been the subject of renewed research interest with the recognition of (a) its pathogenicity and (b) its involvement in epidemics of waterborne diarrheal disease. One area of interest has been the study of the ultrastructure of these protozoa; it is to be hoped that an understanding of the cell biology of Giardia may include an understanding of an effective defense strategy against these parasites. During these ultrastructural studies of parasite organelles, the presence of symbiotic organisms was noted; the role of these symbionts remains unclear. Most of the reports to date describe endosymbionts in Giardia from mice. Thus, Nemanic et al. (12), described endosymbionts in trophozoites and cysts from the G. muris employed in the mouse model originally established by RobertsThomson et al. (17); the original source of these organisms was the golden hamster. Other workers earlier had reported similar structures. Boeck, for example, noted in 1917 (5) rodlike bacteria covering the trophozoites of Giardia from meadow mice; some occurred as inclusions within the trophozoite as well. Wenrich in 1940 (23) observed intracytoplasmic rods in Giardia trophozoites and cysts from rats and mice. Brug (6) described cytoplasmic inclusions from the Giardia of white mice, which Ball (2) later concluded could have been chytrid fungi. Soloviev and Chentsov in 1970 (21) illustrated by transmission electron microscopy the presence of rounded structures, of unknown nature, surrounded by a double membrane in the cytoplasm of binucleate cysts of the G. muris type. In an unpublished observation in 1978, Radulescu found endosymbionts in Giardia isolated from naturally infected hamsters. Radulescu et al. (15) noted in 1982 the presence of rodshaped structures in Giardia cysts from children who had received a live Shigella vaccine; similar rod shaped inclusions were noted intracytoplasmically in the trophozoites of G. muris from mice which had been intubated with S. flexneri. These bodies were not observed in the protozoa collected before the administration of the vaccine. Sogayar and Gregorio (20) reported finding two types of cytoplasmic inclusions, in trophozoites of G. muris from hamsters, and in trophozoites of the G. duodenalis type from domestic rats. Both were limited by a double membrane. More recently, Feely (Abstr. Annu. Meet. Am. Soc. Parasitol. 1986,; 97, p.56) described the use of a DNA specific fluorescent dye, Hoechst 33258, to detect symbionts inside G. microti trophozoites. More than threequarters of the parasites examined had symbionts, some containing more than 150 in the cytoplasm. It seems likely that the occurrence of symbionts within Giardia is not a rare event, but a relatively common one that has not been more frequently reported because it was not searched for. In the experiments described here, we decided to experimentally reproduce the conditions present in chronic giardiasis associated with Shigella infection. To do this, we administered a live oral Shigella flexneri vaccine to a group of children naturally infected with Giardia. The vaccine regimen, which involves repeated doses of large numbers of bacteria, provided the opportunity for the bacterialprotozoal association. Material and Methods These experiments were conducted on a group of 10 preschool children with Giardia infection, from a child care facility with a history of intestinal disease. The children were vaccinated orally with live dysentery vaccine (VADIZEN) produced by the Cantacuzino Institute, Bucharest. The vaccine was administered in 5 graded doses ranging from 0.3 to 1.2 mL of live Shigella flexneri 2a T32 Instratesuspension (1011 * Corresponding author. Address: Dept. of Microbiology & Immunology, Oregon Health Sciences University, 3181 Sam Jackson Park Rd., Portland, Oregon, 97201, U.S.A..
Page 26
organisms/mL), at three day intervals. Before vaccination, and three days after the last vaccine administration, feces were collected, and Giardia cysts were concentrated according to the filtration and sucrose flotation method of Bingham et al. (3). Preparation of the Material for Electron Microscope Study The pellet of cysts was fixed in 2.5% glutaraldehyde in 0.1M phosphate buffer (pH 7.2) for a hour at 4°C. After successive washing in 0.15M phosphate buffer (pH 7.2) the pellet was postfixed in 1% osmium tetroxide in 0.1M phosphate buffer for 1 hour at 4°C, then the pellet was embedded in agar. The resulting fragments were dehydrated in increasing acetone concentrations and finally embedded in Vestosol. Thin sections were cut on a Porter Blum MT1 microtome and stained with saturated uranyl acetate in 50% ethanol, followed by lead citrate. The sections were examined in a Hitachi HU11 electron microscope. The control cysts, obtained from the feces of the five children with giardiasis who had not received the vaccine, were concentrated as described above. Results The electron microscope study of the Giardia cysts obtained from the children who had received the Shigella vaccine revealed the following: The cysts are elliptically shaped and range in length from 6 to 10 µm. The cyst wall is often contiguous with the cell membrane, but in some places a space, containing flagellar axonemes, can be observed. A network of vacuoles and mictrotubules exists in the peripheral cytoplasm. Other identifiable organelles in the cytoplasm include nuclei, flagellar axonemes, rough endoplasmic reticulum, median bodies, ventral disc fragments, and ribosomes.
Figure 1. TEM of a G. lamblia cyst illustrating the fibrillar cyst wall (CW), profiles of fragmented portions of the ventral disc (VD), flagellar axonemes (FA), and endosymbiont (E). x27,000.
Figure 2. TEM of part of G. lamblia cyst. Microtubules of the median body (MB) endosymbiont suggesting binary fission (E). x37,000.
In addition to parasite organelles, we observed other structures, apparent bacilluslike endosymbionts, whose size ranged from 250350 nm in width and 5001200 nm in length (Figure 1). The endosymbionts were bounded by twounit membranes, often with a clear space between the membranes. These endosymbionts have the appearance of bacteria, with dense peripheral cytoplasm and central bright areas containing strands of fibrillar material resembling DNA. Some endosymbionts resembled bacteria undergoing binary fission. None of the endosymbionts appeared to be in the process of degeneration or intracellular digestion (Figures 2 and 3). The Giardia cysts from the control group of children presented all of the same characteristics as those of the vaccinated group, except that in none of them were endosymbionts observed. Discussion This study indicates that bacteria can associate endosymbiotically with Giardia from humans as well as lower animals. We believe that this is the first report of such endosymbionts in Giardia isolated from humans. The endosymbionts in Giardia from humans were essentially the same as those described in G. muris by Nemanic et al. (12), except that in the latter study larger
Page 27
Figure 3. TEM of part of a G. lamblia cyst illustrating the endosymbiont (E), nucleus (N), and ventral (VD). x37,000.
numbers of symbionts were seen. The detection of these endosymbionts raises a number of questions which require further study. One question unaddressed thus far concerns the means by which the endosymbiont becomes internalized within the parasite. Giardia are not known to consume particles the size of bacteria. Bockman and Winborn (4) have demonstrated that ferritin can be taken up from the small intestinal lumen and localized in vacuoles beneath the plasmalemma of Giardia trophozoites. There is no evidence to suggest that these vacuoles are involved in the uptake of particles as large as bacteria. We suggest the possibility that the bacteria may actively invade the trophozoite in a manner similar to that described in some amoebae (2). Once inside the trophozoite, the bacterium is surrounded by membranes which fuse with lysosomes as in the case of phagolysosomes. The strategy, in the evolution of the bacteriaprotozoan relationship toward endosymbiosis and away from digestion, consists in the capacity to hamper this fusion with lysosomes. In the case of amoeba, the membranes of symbiontcontaining vesicles have a 200kD polypeptide on the cytoplasmic side, which is suspected to be involved in the prevention of lysosomal fusion (16). These membranes play a key role in the interaction. They are specifically formed around the symbiont as a result of cellcell recognition events, they have protective properties, and they can probably control metabolic changes. Nonsymbiotic bacteria are digested in 48 hours, while symbiotic ones survive and establish the association. We consider it likely that the intracellular bacteria we observed in this study are symbiotic because of their intact morphological appearance and the absence of any sign of intracellular digestion. In nature, a given protozoan parasite is exposed to a limited range of potential bacterial partners; of these, even fewer apparently become associated in a specific symbiotic relationship. In the present situation, bacterial endosymbionts were observed in Giardia cysts in children who had received live Shigella bacteria; endosymbionts were absent in those not exposed to the bacteria. It seems likely that this relationship was the result of the parasite being overwhelmed by large numbers of bacteria capable of participating in the relationship. The above observations do not permit conclusions regarding the significance of the relationship between Giardia and endosymbiotic bacteria. It is generally considered that a symbiotic association results in the acquisition of new qualitatively superior properties by each participant. These properties become of medical significance when they involve such characteristics as pathogenicity, antigenicity, drug resistance, and environmental survival. Giardia and Shigella often occur together in a chronic diarrheal syndrome. A number of workers have reported that the frequency and duration of diarrheal episodes is higher in the double infection than in infection caused by either organism (8,22). No one, however, has suggested the presence of Shigella endosymbionts as contributing to prolongation of diarrhea. Another aspect of this relationship yet to be studied is the determination whether a plasmid is involved in the ability of Shigella to invade Giardia; an S. flexneri plasmid is presently known to be involved in its invasion of HeLa cells (10). Other examples are known of protozoanbacterial symbiosis, in which one partner or the other is involved in human disease. They have been described in Acanthamoeba sp., for example (7,14), Naegleria fowleri (13), and Dientamoeba fragilis (19). Rowbotham (18) has suggested that Acanthamoeba and Naegleria are possible natural hosts for Legionella; he suggests even more importantly that infected amoebae or vesicles containing bacteria could be the infective particle for Legionella infection of humans. Trypanosomatids also harbor bacterialike endosymbionts. While this group of flagellated protozoa includes members which are intracellular parasites of a variety of hosts, some species are in turn hosts to intracellular symbionts. One of these organisms was the subject of a recent study reported by Krylov et al. (9). The organism studied, Crithidia oncopelti, was considered a composite species, made up of cells containing bacterial endosymbionts and symbiontfree cells. When separated, the two differed significantly in a number of respects including size, flagellum length, colony morphology, velocity of movement, malate dehydrogenase isoenzymes, growth at 35°C,
Page 28
oxygen requirement, and sensitivity to antibiotics. The authors propose that the two isolates be given separate species names. The presence of bacterial endosymbionts in Giardia clearly could have any of a variety of ramifications. On the one hand, it may result in protecting, and facilitating the transfer of, a bacterial endosymbiont pathogenic to humans. On the other hand, endosymbionts represent extrachromosomal carriers of genetic information, the consequences of which may result in morphologic, physiologic, and pathogenic changes in the protozoan. Bacterial endosymbionts can affect the results of endonuclease restriction, isozyme, and antigen analyses that presently are being conducted in efforts to clarify the taxonomy of this complex group of protozoa. One approach to determining the significance of the GiardiaShigella relationship is its study in vitro. Literature Cited 1. Alfieri, S.C., and E.P. Camargo. 1982. Trypanosomatidae: Isoleucine requirement and threonine deaminase in species with and without endosymbionts. Exp. Parasitol. 53:371380. 2. Ball, G.H. 1969. Organisms living on and in protozoa. In: T.T. Chen (ed.), Research in protozoology, Vol. 3, Pergamon Press, Oxford, p 565718. 3. Bingham, A.K., E.L. Jarroll, E.A. Meyer, and S. Radulescu. 1979. Introduction of Giardia excystation and the effect of temperature on cyst viability as compared by eosinexclusion and in vitro excystation. In: Jakubowski, W. and J.C. Hoff (eds.), Waterborne Transmission of Giardiasis, Environmental Protection Agency 600/979 001, p 217229. 4. Bockman, A.K,. and W.B. Winborn. 1968. Electron microscopic localization of exogenous ferritin within vacuoles of Giardia muris. J. Protozool. 15:2630. 5. Boeck, W.C. 1917. Mitosis in Giardia microti. Univ. Calif. Publ. Zool. 18:126. 6. Brug, S.I. 1942. Eigentumliche Einschluss in Lamblia muris Zentralbl. Bakteriol. (Orig. A) 148:166168. 7. Hall, J., and H. Voelz. 1985. Bacterial endosymbionts of Acanthamoeba sp. J. Parasitol. 71:8995. 8. Ingram, V., F.L. Rights, K. Hashimi, and K. Asari. 1966. Diarrhea in children of West Pakistan: occurrence of bacterial and parasitic agents. Am. J. Trop. Med. Hyg. 15:743750. 9. Krylov, M.V., S.A. Pidlipaev, A.S. Khaetskii, L.M. Belove, A.O. Frolov, and B.Y. Niyazbekova. 1985. Is only 1 species present in a culture of Crithidia oncopelti (Kinetoplastmonada, Trypanosomatidae) Zool. Zh.64:165171. 10. Maurelli, A.T., B. Baudry, H. D'Hauteville, T.L. Hale, and P.M. Sansonetti. 1985. Cloning of plasmid DNA sequences involved in invasion of HeLa cells by Shigella flexneri. Infect. Immun. 49:164171. 11. McGhee, R.B., and W.B. Cosgrove. 1980. Biology and physiology of the lower Trypanosomatida. Microbiol. Rev. 44:140173. 12. Nemanic, P.C., R.L. Owen, D.P. Stevens, and J.C. Mueller. 1979. Ultrastructural observations on giardiasis in a mouse model. II. Endosymbiosis and organelle distribution in Giardia muris and Giardia lamblia. J. Infect. Dis. 140:222228. 13. Phillips, B.P. 1974. Naegleria: Another pathogenic amoeba. Studies in germfree guinea pigs. Am. J. Trop. Med. Hyg. 23:850855. 14. ProcaCiobanu, M., G.H. Lupascu, A.L. Petrovici, and M.D. Ionescu. 1975. Electron microscopic study of pathogenic Acanthamoeba castellani strain: the presence of bacterial endosymbionts. Intern. J. for Parasitol. 549556. 15. Radulescu, S., M. Smolinschi, C. Rau, and E.A. Meyer. 1982. Giardia spp. cysts: possible vectors of enterobacteria. Mol. Biochem. Parasitol. Suppl. p 578. 16. Reisser, W., R. Meier, H.D. Gortz, and K.W. Jeon. 1985. Establishment, maintenance, and integration mechanisms of endosymbionts in protozoa. J. Protozool. 32:383390. 17. RobertsThomson, I.C., D.P. Stevens, A.A.F. Mahmoud, and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterology 71:5761. 18. Rowbotham, R.J. 1980. Preliminary report on the pathogenicity of Legionella pneumophila for freshwater and soil amoebae. J. Clin. Pathol. 33:11791183. 19. Silard, R., and B. Burghelea. 1986. Endosymbionts in Dientamoeba fragilis trophozoites resistant to antiprotozoal drugs. Arch. Roum. Pathol. Expl. Microbiol. 45:6574. 20. Sogayar, M.L., and E.A. Gregorio. 1966. Electron microscopic localization of exogenous ferritin within vacuoles of Giardia muris. J. Protozool. 80:4952. 21. Soloviev, M.M., and J.S. Chentsov. 1970. Ultrastructure of cysts of Lamblia muris. Parasitologia 4:510514. 22. Wanner, R.G., F.O. Atchley, and M.A. Wasley. 1963. Association of diarrhea with Giardia lamblia in families observed weekly for occurrence of enteric infections. Am. J. Trop. Med. Hyg. 12:851853. 23. Wenrich, D.H. 1940. Observations on parasites and inclusion bodies in certain intestinal protozoa. Science 92:416417.
Page 29
Morphology of Giardia Encystation In Vitro Daniel G. Schupp, Mary M. Januschka, and Stanley L. Erlandsen* Department of Cell Biology and Neuroanatomy, School of Medicine, University of Minnesota, Minneapolis, Minnesota 55455, U.S.A. Development of an in vitro encystation model for Giardia together with current excystation models would permit in vitro regulation of the entire life cycle of this protozoan. To pursue this goal we monitored several axenically grown strains of Giardia and detected the production of cysts in vitro. These cysts displayed a light microscopic morphology identical to control Giardia cysts isolated from feces. This morphology was based on the characteristic size and shape of Giardia cysts, and the presence of 24 nuclei. Giardia cysts formed in vitro were shown to have the same positive immunoreactivity for the cyst wall as control Giardia cysts. The viability of cysts formed in vitro was measured by incorporation of the fluorogenic dye, fluorescein diacetate, and was variable, ranging up to 50%. Ultrastructural analysis, using transmission electron microscopy, revealed an outer cyst wall with a fibrillar organization, the presence of cyst wall membranes, a peritrophic space, and nuclei. Other intracellular organelles observed included elements of the disassembled ventral adhesive disc, peripheral vacuoles, flagellar axonemes, and basal bodies. Examination of cultures with the scanning electron microscope showed that newly formed cysts had the characteristic size and shape of control cysts and were interspersed among trophozoites. These results demonstrated that Giardia encystation occurred in axenic cultures of trophozoites. Cysts formed in vitro were both morphologically and immunologically similar to Giardia cysts formed in vivo and also were viable, as demonstrated by the uptake of fluorogenic dyes.
Introduction Giardia encystation in vivo has been described by Perroncito in 1888 (13) and by Lavier in 1942 (8), who microscopically studied trophozoites flushed from the intestines of amphibians, rats, and the diarrheic stools of humans. No description of Giardia encystation in vitro has been reported even though axenic cultures of Giardia were developed almost a decade ago (11). The inability of Giardia trophozoites to undergo encystation in vitro has raised questions as to whether host related factors, such as gutassociated microorganisms or nutritive factors and stimuli from the small intestine, might play a role in this process. The development of an in vitro model for encystation of Giardia would enable the complete life cycle of this protozoan to be manipulated outside of the host. This model would be beneficial in many aspects of Giardia research, including biochemical analysis of the cyst and the cyst wall, development of giardicidal agents involved in blocking cyst wall formation, the testing of Kochs' postulates in regard to infectivity of cysts, and the investigation of questions related to nuclear division and cyst wall production. Here we report morphological evidence for the formation of Giardia cysts in axenic cultures of trophozoites and present preliminary findings on their viability, as determined by the incorporation of fluorogenic dyes (15). Materials and Methods Axenic cultures of Giardia derived from man (WB strain, American Type Collection, Bethesda, MD), beaver (IP 0482:1 and IP0583:1 from Dr. Louis Diamond; PB1 and B5 from Drs. Peter Wallis and Henry Stibbs), and muskrat (MR4 from Drs. Peter Wallis and Henry Stibbs) were grown on TYI33 medium with adult bovine serum as described by Keister (5). Encystation in vitro was stimulated by addition of bile to the medium (17). The cultures were examined daily by inverted phase microscopy and, after three days of growth, the contents of the culture were pelleted by centrifugation for 10 minutes at 600 × g. A slide was made directly from this pellet for light microscopic observation, and examined using either a Zeiss photomicroscope II or an Olympus BH2 light microscope equipped with phase, DIC and UV epiillumination. The viability of cysts formed in vitro was determined by the incorporation of the fluorogenic dyes, fluorescein diacetate (FDA) and propidium iodide (PI) (15). For scanning electron microscopy (SEM), the pelleted cultures were washed twice in 0.9% saline, centrifuged as above between washes, then resuspended in saline and placed on glass chips that had been coated with poly1lysine (Sigma Chemical Co., St. Louis, Mo.). After allowing the cultured cells to adhere to the glass chips for 1 hour, the samples were fixed for three to four hours with 2.5 % glutaraldehyde buffered with 0.1 M sodium cacodylateHCl, pH 7.4. This was followed by postfixation in 1% osmium tetroxide in the same buffer, for one to two hours at 4°C. The glass chips with attached Giardia were then dehydrated in an ascending ethanol series and critical point dried using the technique of Anderson (1). The specimens were sputter coated with goldpaladium and mounted on stubs with silver paint and copper tape. All samples were examined at 20 kV in a Hitachi S450 scanning electron microscope and micrographs were recorded on Polaroid type 55 P/N film. For transmission electron microscopy (TEM) the pelleted cultures were fixed as described above with the exception of adding a few drops of albumin to the cultured material between the fixative steps. The fixed albumin block containing the cell suspension was then dehydrated and embedded in epoxy as described by Luft (10). Sections were cut on a LKB Huxley * Corresponding author.
Page 30
ultramicrotome, stained with uranyl acetate and lead citrate, then examined with a JEOL 100 CX electron microscope. Results Axenic cultures of Giardia were stimulated to undergo encystation in vitro by addition of bile to the medium (17). After one day, examination of the cultures by
Figures 1 and 2. Differential interference contrast (DIC) micrographs of a Giardia culture three days after induction of encystation with bile. A cluster of viablelooking Giardia cysts (arrowheads) formed in vitro are seen between trophozoites (T). In Figure 2, the viable cysts possess a welldefined cyst wall (CW) and peritrophic space (PS). The cytoplasm has a hyaline appearance, but faint outlines of some organelles including nuclei (N), axonemes of flagella (AX), and portions of the adhesive disc (AD) can be seen. Bar equals 5 microns.
Figures 3 and 4. DIC and fluorescence micrograph of two Giardia cysts after incorporation of the fluorogenic dyes, FDA and PI. The cyst formed in vitro (arrowhead) shown in Figure 3 had the morphological characteristics used to determine viability in G. muris cysts (16). This cyst incorporated the fluorogenic dye, FDA, and emitted a green fluorescence when viewed with an excitation wavelength of 450490 nm. The cyst with nonviable morphology (N) seen in Figure 3 was stained with the fluorogenic dye, PI, as shown in Figure 4, and fluoresced orange. Bar equals 5 microns for Figures 3 and 4.
Figure 5. TEM of Giardia cyst formed in vitro. This cyst has the morphological characteristics of viable cysts by DIC, namely a clearly delineated cyst wall (CW) of uniform thickness and a peritrophic space (PS) just interior to the cyst wall. Other organelles within the cyst include flagellar axonemes (AX), peripheral vacuoles (V), and portions of the disassembled adhesive disc (AD). Bar equals 1 micron.
Figures 6 and 7. SEMs of a Giardia culture after exposure to bile for three days. Giardia cysts (arrowheads) formed in vitro were seen scattered between flagellated trophozoites, which were attached to the substratum. At higher magnification, seen in Figure 7, the contrast between the flagellated trophozoite and the in vitro formed cyst was readily apparent. The cysts were oval to round in shape and possessed a smooth surface, whereas the pearshaped trophozoites were flattened in a dorsalventral plane and attached to the substratum by their ventral adhesive disc. Bar equals 5 microns for Figure 6 and 1 micron for Figure 7.
inverted microscopy revealed small foci or clusters of trophozoites. By day three, these clusters had increased in size and included trophozoites as well as cysts. Examination of the cultures on day three by DIC microscopy revealed the presence of Giardia cysts interspersed amongst trophozoites (Figures 1 and 2). The cysts formed in vitro were determined to be of two types, using morphological criteria. One type appeared viable based on previously published criteria (16) and possessed the typical morphological features of cysts formed in vivo. These viable cysts, shown in Figure 2, were recognized by their characteristic size, shape, and the presence of distinguishing features, such as cyst wall, peritrophic space, nuclei, flagellar axonemes, and curved portions of the adhesive disc. The other type of cyst (not shown) was of similar size, shape, and had a cyst wall, but appeared nonviable since the cyst usually lacked the presence of typical organelles or contained two to four nuclei within a shrunken cytoplasmic mass. The viability of Giardia cysts formed in vitro was tested using the incorporation of fluorogenic dyes. Morphologically intact cysts were often seen to incorporate
Page 31
FDA, whereas those cysts lacking recognizable contents or having a shrunken cytoplasmic mass stained with PI (Figures 3 and 4). A wide variation in cyst viability was detected using the incorporation of FDA, ranging from 4050% to less than 1%, in Giardia cysts formed in vitro. A Giardia cyst formed in vitro is illustrated by TEM in Figure 5. This cyst, based on its' morphological appearance, is similar to the viable cysts seen by DIC (Figure 2). Typical organelles characteristic for Giardia are seen within the cytoplasm, including a cyst wall, peritrophic space, flagellar axonemes, and curved portions of the adhesive disc, all of which are indistinguishable from those of control cysts formed in vivo. Examination of axenic cultures producing cysts, using SEM, also revealed the presence of cysts interspersed among trophozoites (Figures 6 and 7). The flattened trophozoites were pearshaped, often attached to the substratum by their adhesive discs, and were easily recognized by the presence of flagella. The cysts formed in vitro were round to oval in shape, ranging from 711 microns in width and 1114 microns in length, and had a cyst wall that, examined by SEM, appeared to be smooth. Discussion Our results, using both light and electron microscopy, have demonstrated that Giardia cysts formed in vitro have a morphological appearance indistinguishable from that of murine or human cysts formed in vivo (3,4,7,9,12,18). The morphological appearance of the cysts formed in vivo, as seen by DIC, closely resembled that of G. muris cysts that had been shown to be viable due both to their ability to incorporate fluorogenic dyes and infect animals (16). The ability to differentiate viable from nonviable cysts was based on the distinct appearance of the cyst wall, the presence of a peritrophic space between the cyst wall and cytoplasm, and the distinct definition of the intracellular organelles observed. These morphological features were also prominent in cysts formed in vitro (Figure 2) and suggested that they too were viable. The striking ultrastructural similarity between cysts formed in vitro and viable Giardia cysts corroborated our prior assumption that the cysts formed in culture were indeed viable. The viability of Giardia cysts can be measured by either the incorporation of fluorogenic dyes (15), excystation (2) or by producing an infection in an animal model (14). A brief report has indicated that small intestinal stimuli can induce Giardia cyst formation in vitro, but no information was provided on either the ultrastructural appearance of these cysts, or their viability (Gillin et al., 35th Annual Meeting of the Am. Soc. Trop. Med. Hyg., abstract no. 37, 1986). The use of fluorogenic dyes to determine the viability of G. muris cysts has been correlated with both their morphology and their ability to produce infection in an animal model (15). As presented here, the incorporation of FDA by Giardia cysts formed in vitro has provided the first physiological evidence for the viability of cysts produced in culture. We have also recently demonstrated that Giardia cysts, formed in vitro, were capable of undergoing excystation and were able to produce infection in an animal model (17). Therefore, Giardia cysts formed in vitro have fulfilled all of the criteria currently being used to measure cyst viability. Our studies have been the first to demonstrate the viability of Giardia cysts formed in vitro. The physiological evidence obtained using the incorporation of the fluorogenic dye, FDA, together with the excellent morphological appearance of the cysts produced, in vitro, has demonstrated that the life cycle of Giardia could now be completed outside of the animal host. The development of this in vitro model of the life cycle of Giardia should facilitate the testing of Kochs' postulates, since it will now be possible not only to isolate trophozoites from an animal and grow them in culture, but also to produce cysts in vitro, which could be used to reinfect the same (or possibly a different) type of host. Acknowledgements The authors wish to thank Dr. W.J. Bemrick for his review of the manuscript and Ms. LeeAnn Sherlock for her excellent technical assistance. Although the research described in this article has been funded in part by the U.S. Environmental Protection Agency through cooperative agreement #CR811834 to the University of Minnesota, it necessarily does not reflect the view of the Agency, and so official endorsement should not be inferred. Mention of trade names or material products does not constitute endorsement or recommendation for use. Literature Cited 1. Anderson, T.F. 1951. Techniques for the preservation of three dimensional structure in preparing specimens for the electron microscope. Trans. N.Y. Acad. Sci. 13:130134.3. 2. Bingham, A.K., Jarroll, E.L., Meyer, E.A, and S. Radulescu. 1979. Giardia sp.: physical factors of excystation in vitro, and excystation vs. eosin exclusion as determinants of viability. Exp. Parasitol. 47:281291. 3. Coggins, J.R. and F.W. Schaefer. 1986. Giardia muris: ultrastructural analysis of the in vitro excystation. Exper. Parasit. 61:219228. 4. Feely, D.E., Erlandsen, S.L., and D.G. Chase. 1984 Structure of the trophozoite and cyst. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology (Erlandsen, S.L., and Meyer, E.A., eds.) Plenum Press, New York, NY, pp. 331. 5. Keister, D.B. 1983. Axenic culture of Giardia lamblia in TYIS33 medium supplemented with bile. Trans. R. Soc. Trop. Med. Hyg. 77:487488. 6. Knight, D.P. 1977. Cytological staining methods in electron microscopy. In: Lewis, P.R., and Knight, D.P., eds., Staining Methods for Sectioned Materials North Holland, Amsterdam, pp. 2976. 7. Kulda, J., and E. Nohynkova. 1978 Flagellates of the human intestine and of intestines of other species. In: Parasitic Protozoa (Kreier, J.P., ed.) Academic Press, New York, NY, pp. 69104. 8. Lavier, G. 1942. Les Modalites de l'enkystement chez les flagelles du genre Giardia. Soc. Biol. Paris 136: 67 70. 9. Luchtel, D.L., Lawrence, W.P., and F.B. DeWalle. 1980. Electron microscopy of Giardia lamblia cysts. Appl. Environ. Micro. 40:821832.
Page 32
10. Luft, J.H., 1961. Improvements in epoxy resin embedding methods. J. Biophysical and Biochemical Cytology 9:409413. 11. Meyer, E.A. 1976. Giardia lamblia: isolation and axenic culture. Exp. Parasitol. 2:187196. 12. Owen, R.L. 1980 The ultrastructural basis of Giardia function. Trans. R. Soc. Trop. Med. Hyg. 74:429433. 13. Perroncito, E. 1888. Note sur l'enkystement du Megastoma intestinale. Bull. Soc. Zool. Fr. 13:1618. 14. RobertsThomson, I.C., Stevens, D.P., Mahmoud, A.A.F., and K.S. Warren. 1976 Giardiasis in the mouse: an animal model. Gastroenterology 71:5761. 15. Schupp, D.G. and S.L. Erlandsen. 1987. A new method to determine Giardia cyst viability: correlation between fluorescein diacetate/propidium iodide staining and animal infectivity. Appl. Envir. Micro. 53:704707. 16. Schupp, D.G. and S.L. Erlandsen. 1987. Determination of Giardia muris cyst viability by differential interference contrast, phase, or bright field microscopy. J. Parasit. 73:723729. 17. Schupp, D.G., Januschka, M.M., Sherlock, L.A.F., Stibbs, H.H., Meyer, E.A., Bemrick, W.J. and S.L. Erlandsen. 1988. Production of viable Giardia cysts in vitro: Determination by fluorogenic dye staining, excystation, and animal infectivity in the mouse and Mongolian gerbil. Gastroenterology, In press. 18. Sheffield, H. 1979. The ultrastructural aspects of Giardia. In: Jakubowski, W. and Hoff, J.C., eds, Waterborne Transmission of Giardia. U.S. Environmental Protection Agency, Cincinnati, OH, EPA 600/979001 pp.921.
Page 33
Cytopathogenicity of Giardia lamblia in HeLa and Vero Cell Monolayers A. Jyothisri* and Usha K. Baveja Department of Microbiology, Maulana Azad Medical College, New Delhi 110001, India The behaviour and pathogenic effects of four human strains of Giardia lamblia (three locally designated ND1, ND2, ND3 and Portland 1), were studied in HeLa and Vero cell culture systems using light and scanning electron microscopy. All the strains tested have shown the same degree of initial attachment to the target cell and have produced cytopathogenic effects in both cell culture systems. After inoculation of parasites morphological changes appeared 4 hours later in HeLa cells and 6 hours later in Vero cells. Changes in cytoplasmic granulation, vacuolation and nuclear margination were apparent by 24 hours in HeLa and Vero cells. The number of cells showing these changes increased progressively with time. By 96 hours, the monolayer of HeLa cells and by 87 hours, the monolayer of Vero cells was completely destroyed and replaced by G. lamblia trophozoites. Significant changes in the surface morphology of cells after interaction with trophozoites were examined by scanning electron microscopy. Cells that were not in direct contact with the trophozoites showed little damage compared to those cells in direct contact with trophozoites. Cell free extracts and culture supernatants of G. lamblia produced no morphological changes in either HeLa or Vero cells in vitro. These observations suggest that direct cell to parasite contact may play an important role in damaging the host cells.
Introduction Giardia lamblia is an established human pathogen causing a wide spectrum of clinical illness varying from an asymptomatic carrier state to acute diarrhea and malabsorption (1,15). A number of different mechanisms have been proposed to explain the intestinal dysfunction in giardiasis. These include the presence of a barrier to absorption, epithelial cell damage, inflammatory reaction to parasites, bile salt deconjugation and mucosal invasion by trophozoites (2,10,13). None of these mechanisms precisely explains how G. lamblia produces symptoms in only a fraction of infected individuals. The axenic cultivation of G. lamblia has made it possible to carry out in vitro studies to understand the mechanism by which the parasite may be producing its deleterious effects and disease. So far, two such studies have been conducted. One reported the significance of bile salt deconjugation by the parasite (12) and the other the cytopathogenic effects of Giardia on cell monolayers, in vitro (11). The present study was undertaken to find out the probable mechanism by which the different strains of G. lamblia damage the cells (epithelial cells and fibroblasts) in vitro and to assess whether similar mechanisms operate in human diarrhea and malabsorption. This study reports the morphological changes induced by four different human strains of G. lamblia on HeLa and Vero cells. Materials and Methods Giardia lamblia Culture Four human strains of G. lamblia ND1, ND2, ND3 (3) and Portland 1 ACTC No. 30888 (9) were maintained in Diamond's TPS1 (4) complete filter sterilized medium according to the method of Visvesvara (14). Preparation of Different Inocula 1). Each of the four strains of G. lamblia trophozoites were prepared by chilling tubes with 48 h old cultures in an ice water bath and centrifuging at 600 g for 10 minutes at 4°C. The trophozoites were resuspended in TPS1 medium, so as to get counts of 1 × 104 to 1 × 107 trophozoites/mL. 2). Cell free culture supernatant was collected from 48 h old cultures after chilling and centrifugation as described above. 3). Cell free extracts were prepared from actively multiplying trophozoites at concentrations ranging from 1 × 104 to 1 × 107 trophozoites according to the method of Lushbaugh et al. (8). After washing with phosphate buffered saline, trophozoites were ruptured by a freeze/thaw cycle and then centrifuged (105 × g for 1 h) at 4°C. The supernatants (cell free extract) were used to assay the cytotoxicity. 4). Fresh TPS1 complete medium without the parasite and minimal essential medium (MEM) with 5% of calf serum were used as controls. Cultivation and Preparation of HeLa and Vero Cell Monolayers HeLa and Vero cells were maintained in Dulbecco's MEM supplemented with 5% adult calf serum and antibiotics (Penicillin G 100 units/mL and Streptomycin 100 µg/mL). Trypsinized monolayers (0.01% Trypsin and 0.05% EDTA) were harvested and adjusted to a concentration of 1.5 × 105 cells/mL to initiate the cultures for cellparasite interactions. Selection of Inoculum Size Before starting the main experimental series, it was necessary to determine the optimal number of G. lamblia trophozoites that would produce optimal results for a proper and meaningful analysis after interactions between the parasite and cell culture. For this purpose, effects of inocula of varying number of trophozoites ranging from 1 × 104 to 1 × 107 on HeLa and Vero cells at different intervals (0, 5, 18, 24, 48, 72 and 96 h) after incubation at 37°C were studied. Methods of Study of CellParasite Interactions 1). Preparation of Cell Monolayers on Coverslips Tissue culture cells (HeLa and Vero) at a density of 1.5 × 105 cells/mL • Corresponding author.
Page 34
were pipetted into Leighton tubes containing coverslips closely applied to the flat surface of the tubes (7). The cells formed monolayers over the coverslip after incubation at 37°C for 48 h. The MEM was gently decanted before the start of the experiment and replaced with one of the experimental or control culture systems described above according to the method of Radulescu (11). All the inoculated tubes were incubated at 37°C in a slanted position. These were examined daily for attachment and detachment of parasites to the surface of cells, and other events occurring up to a period of 96 h were noticed. In all the series, the experiments were terminated at different intervals (0 5, 18, 24, 48, 72 and 96 h) after incubation and the coverslips (2 for each interval in each series for the two cell systems studied), were removed from the tubes and processed for light and electron microscopic studies as described below. 2). Light Microscopy After the respective interaction period the used medium was decanted off and coverslips were washed gently with warm (37°C) PBS in situ. After fixing with methanol and staining with Giemsa, cover slips were removed from the tubes, airdried, placed monolayer sidedown on a drop of mounting medium DPX on a glass microslide. These coverslips were examined with a Zeiss microscope (W. Germany HBO 100 w/2) equipped with a 35 mm camera (winder M, 470 799901 model). At least 10 representative fields on each coverslip were studied to compute the results. Representative photographs were taken at different intervals after interaction. 3). Scanning Electron Microscopy (SEM) To prepare the specimens for SEM, coverslips with the cell monolayer after interactions were fixed with 0.2M sodium cacodylate buffer. The monolayers were dehydrated through a graded series of ethanol. The exposure time for each concentration of ethanol was 5 minutes. The specimens were dried at the critical point in a Polaron apparatus with liquid CO2. Finally, the coverslips were mounted on aluminum stubs and shadowed with silver in an argon atmosphere using Polarondriode sputter coater E 5000. The specimens coated with silver were observed in a Scanning Electron Microscope (Phillips 500B) operating at 10 to 20 kV and photographs were taken with 120 mm camera (Super lollex, Holland). Results Selection of Inoculum An inoculum consisting of 1 × 105 trophozoites was found to be ideal for a proper evaluation of progression of morphological changes with convenient time limits. The changes occurred too rapidly with a higher inoculum (1 × 107 trophozoites) and too slowly with a lower inoculum (1 × 104 trophozoites). Hence, an inoculum size of 1 × 105 trophozoites was used in all the subsequent experiments. Light Microscopic Studies 1). Attachment of G. lamblia to the Monolayers Many trophozoites were found to be closely attached to the external surfaces of the cultured cells as early as 15 minutes after incubation at 37°C. The number of attached trophozoites increased with time during the first three hours and by this time almost half the culture cells were in direct contact with the trophozoites. All the four strains of G. lamblia showed a similar tendency to attach to the surfaces of HeLa and Vero cells. Attachment did not seem to be prolonged or permanent, as the trophozoites attached and detached from the target cell in a fraction of a second. This was observed in the wet preparations during examination by phase contrast microscope. Some trophozoites were swimming freely in the medium and showed active movement and division.
Figure 1. Light micrographs showing epithelial cells (HeLa cell line) after incubation with G. lamblia. A). Control: 24 h incubation of HeLa cell monolayer without G. lamblia in TPS1 medium showing normal morphology. Giemsa × 400. The term h (hours) indicates the length of exposure of culture cells with or without G. lamblia trophozoites. B). 24 h epithelial cells showing vacuolation in the cytoplasm and attached parasites. Giemsa × 400. C). 48 h cytoplasm of cells showing extensive vacuolation and the parasites filled in the cell denuded areas. Giemsa × 1000. D). 72 h cytoplasm of epithelial cells showing vacuolation and nuclear abnormalities after progressive increase in the parasite number. Giemsa × 400.
2). Interactions Between Parasites and HeLa Cells Morphological changes in HeLa cells during the initial hours of interaction (4 to 18 h) consisted of cytoplasmic vacuolation. These were apparent by 24 h after incubation (Figure 1B). With further incubation, cell free areas appeared in the monolayers as a result of detachment of the parasite damaged cells. At 48 h after interaction the majority of cells had a highly vacuolated cytoplasm. The number of trophozoites increased steadily as the interaction period proceeded. Trophozoites filled the cell free areas (Figure 1C). By 72 h after incubation almost all the cells showed cytoplasmic vacuolation. Few cells showed pyknotic nuclei and few had multiple nuclei (Figure 1D). At the end of the interaction period (96 h) the cell monolayer was replaced completely by the parasite monolayer. Identical morphological changes of the same intensity were induced by all the 4 strains viz. ND1, ND2, ND3 and Portland 1 on interacted HeLa cells. The morphological changes exhibited by Vero cells upon interaction with G. lamblia trophozoites are shown (Figure 2AF). These changes consisted of retraction of cytoplasm in some cells (6 h, Figure 2B) loss of cell to cell contact with slight vacuolation of cytoplasm (24 h,
Page 35
Figure 2. Photomicrographs of fibroblast cells (Vero) after incubation with G. lamblia. A). 6 h control Vero cell monolayer showing the normal morphology. Giemsa × 400. B). 6 h showing early signs of changes in appearance of cell free areas. Attached G. lamblia are seen. Giemsa × 400. C). 24 h fibroblast cells showing slight vacuolation, increase in cell free areas. Giemsa × 400. D). 48 h control fibroblast monolayer. Giemsa × 100. E). 48 h fibroblast cells showing loss of intercellular contacts, multiple nuclei and abnormal cytoplasm. Giemsa × 100. F). 72 h fibroblast cells showing excessive vaculoation in the cytoplasm, two nuclear, and attached G. lamblia. Giemsa × 400.
Figure 2C) which gradually increased so that by 48 h the fibroblasts were narrow, elongated and branched (Figure 2E). Extensive vacuolation of cytoplasm, nuclear pyknosis and multiple nuclei were observed 72 h after interaction (Figure 2F). The trophozoites multiplied and increased in number progressively until the whole cell monolayer was replaced by a G. lamblia monolayer (87 h). Morphological changes in Vero cell monolayers of the four human strains of G. lamblia were similar. 3). Interaction Between Vero and HeLa Cells and the Culture Supernatant No morphological changes were seen in either HeLa or Vero cells from each of the four strains upon incubation with the culture supernatant up to a period of 96 h compared to the matched controls. 4). Interaction Between Vero and HeLa Cells and the Cell Free Extract Cell free G. lamblia extracts apparently had no deleterious effects and did not induce any morphological changes compared with TPS1 and MEM control culture cells in either HeLa or Vero cells till 96 h after interaction. Scanning Electron Microscopy (SEM) The surface changes seen in fibroblast cells (Vero cells) at different intervals (0, 6, 24, 48, 72 h) after interaction with G. lamblia are shown in Figure 3 (AE). The first interaction was the adhesion of trophozoites to the cell surface followed by progressive destructive changes in the cell. A number of trophozoites were attached to the cell surfaces by 6 h after interaction. In these trophozoites the dorsal surface and four pairs of flagella were noticeable (Figure 3B). The ventral surface was in close contact with the cell surface. On further interaction fibroblasts lost cell to cell contact and ruffling appeared on the surface (24 h, Figure 3C). Some of the trophozoites were seen to be attached to the coverslip. The lateral shield along the periphery of the trophozoites and the ventrolateral flange could be seen in direct contact with both the cell surface in some and the coverglass in others. Filamentous processes extending from the cell borders and increased ruffling of surfaces was seen 48 h after interaction (Figure 3D). Trophozoites were oriented at random and were attached to any part of the cell surface. These changes progressively increased as the interactions progressed (Figure 3E) until the degenerated and damaged cells detached from the surface of the coverslip. Discussion The predominance of morphological changes in those Vero and HeLa cells that were in contact with the trophozoites emphasizes the direct damaging effects of the parasite. Cell free extracts and culture supernatants from Giardia culture had no deleterious effect on either Vero or HeLa cells. Similar findings have been reported earlier also (11). All the four strains of G. lamblia induced similar morphological changes in Vero and HeLa cells. This may be because all four strains were isolated from cases of symptomatic giardiasis (3,9). Hence, the capacity of a strain to produce CPE in tissue culture system may relate to its pathogenicity in vitro. The morphological changes observed in parasite interacted cells were similar to those reported by Radulescu et al. (11). Scanning electron microscopy revealed the ventral surface, lateral shield and ventrolateral flange of the parasite to be in close contact with the surface of cells. Erlandsen et al. (5) reported a similar pattern of attachment of G. lamblia to the microvillous border of the epithelial cells. Fibroblasts showed ruffling of the surface upon interaction with trophozoites. However, the cell membrane of the cells surrounded by parasites was continous indicating that the parasite did not penetrate the membrane and produced the cell damage just by attaching to the surface of the cells. G. lamblia may be damaging the cultured cells either by mechanical or chemical injury during close contact with the cell or by activation and release of factors/toxins after contact with the cell resulting in cell damage. The results of the present study indicate that G. lamblia most probably does not produce toxins as reported in other protozoans like E. histolytica (8) and T. vaginalis (6). We believe that parasite to cell contact alone is responsible for morphological changes and damage in epithelial and fibroblast cells. However, these data are not sufficient to differentiate between the damage caused by chemical and mechanical effects of the parasite. Future studies on the localization of the parasite enzymes on the plasma membrane and the molecular mechanisms of the
Page 36
Figure 3. Scanning electron micrographs of fibroblast cells after addition of G. lamblia trophozoites. A). 6 h scanning electron micrograph of control fibroblast cell monolayer (Vero cell line) having well spread normal morphology (original magnification × 1250). B). 6 h adherence of G. lamblia trophozoites to the fibroblasts (slightly out of focus). Initial contact between ventral surface of the parasite and external surface of the fibroblast cell (original magnification × 2500). C). 24 h cells showing more ruffles on the cell surface than the normal cells and loss of contact with adjacent cells. The trophozoites attached to the fibroblast cells as well as the coverglasses are seen (original magnification × 1250). D). 48 h fibroblast cells showing extensive ruffling on the surface and cytoplasmic processes that are indicative of further damage caused to fibroblasts by G. lamblia trophozoites. Trophozoites are in direct contact with the surface of fibroblasts (original magnification × 1250). E). 72 h fibroblasts showing further damage due to G. lamblia trophozoites. Branched filamentous processes and rufflings on the surface can be seen clearly (original magnification × 1250) with the surface of fibroblasts (original magnification × 1250).
specific adherence of the ventral disc of the parasite to the cell membrane will be important in pinpointing the exact mechanisms of pathogenesis. Studies with strains isolated from asymptomatic patients will also be valuable in distinguishing the differences between virulent and avirulent strains. Our results indicated that both HeLa and Vero cell lines are susceptible to G. lamblia trophozoites. This leads to the fact that one can use a variety of cell culture systems to study the host parasite interactions in case of G. lamblia. This simple in vitro assay system is rapid and sensitive compared with in vivo systems for the study of pathogenesis. Acknowledgements This work was supported by a research grant (70/288/85ECD II) from the Indian Council of Medical Research. Literature cited 1. Ament, M.E., and C.E. Rubin. 1972. Relation of giardiasis to abnormal intestinal structure and function in immunodeficiency syndromes. Gastroenterol. 62:216 226. 2. Anand, B.S., R. Chaudhary, A.S. Jyothi, R.S. Yadav, and U.K. Baveja. 1985. Experimental examination of the direct damaging effects of Giardia lamblia on intestinal mucosal scrapings of mice. Trans. R. Soc. Trop. Med. Hyg. 79:613617. 3. Baveja, U.K., A.S. Jyothi, M. Kaur, D.S. Agarwal, B.S. Anand, and R. Nanda. 1986. Isoenzyme studies of Giardia lamblia isolated from symptomatic cases. Aust. J. Exp. Biol. Med. Sci. 64:119126. 4. Diamond, L.S. 1968. Techniques of axenic cultivation of Entamoeba histolytica Schaudinn, 1903 and E. histolyticalike amoeba. J. Parasitol. 54:10471056. 5. Erlandsen, S.L., and D.E. Feely. 1984. Trophozoite motility and the mechanisms of attachment, pp. 3361, In: Erlandsen, S.L. and E.A. Meyer (eds.), Giardia and giardiasis: Biology, pathogenesis and epidemiology. Plenum, New York. 6. Farris, V., and B.M. Honigberg. 1970. Behaviour and cytopathogenicity of Trichomonas vaginalis Donn liver cultures. J. Parasitol. 56:849882.
Page 37
7. Koester, S.K., and P.G. Engelkirk. 1984. Glass coverslip technique for studying in vitro interactions between Giardia trophozoites and host leukocytes by TEM, SEM, and light microscopy. 70:443445. 8. Lushbaugh, W.B., A.F. Hofbauer, A.B. Kairalla, J.R. Cantey, and F.E. Pittman. 1984. Relationship of cytotoxins of axenically cultivated Entamoeba histolytica to virulence. Gastroenterol. 86:14881495. 9. Meyer, E.A. 1976. Giardia lamblia: isolation and axenic cultivation. Exp. Parasitol. 39:101105. 10. Meyer, E.A., and S. Radulescu. 1979. Giardia and giardiasis. Adv. Parasitol. 17:147. 11. Radulescu, S., C. Rau, D. Petrasincy, N. Gaicu, and E.A. Meyer. 1980. Behaviour and cytopathogenicity of Giardia lamblia in cell cultures. Arch. Roum. Path. Exp. Microbiol. 39:163170. 12. Smith, P.O., C.R. Horsburgh, Jr., and W.R. Brown. 1981. in vitro studies on bile acid deconjugation and lipolysis inhibition by Giardia lamblia. Dig. Dis. Sci. 26:700704. 13. Smith, P.D. 1985. Pathophysiology and immunology of giardiasis. Ann. Rev. Med. 36:295307. 14. Visvesvara, G.S. 1980. Axenic growth of Giardia lamblia in Diamond's TPS1 medium. Trans. R. Soc. Trop. Med. Hyg. 74:213215. 15. Wolfe, M.S. 1978. Giardiasis. N. Engl. J. Med. 298:319321.
Page 39
Studies on the Prevalence of Giardiasis in Czechoslovakia Michal Giboda Institute of Parasitology, Czechoslovak Academy of Sciences, 370 05 Ceske Budejovice Branišovska 31, Czechoslovakia. The prevalence rate of giardiasis within a particular age group is not uniform. Among children of preschool age, the highest prevalence of giardiasis was found in children's homes, followed by children who attend daycare centres. Giardiasis is least commonly found in children who remain at home. The permanent contact between the children in the children's home creates better conditions for Giardia transmission than in daycare centers and kindergartens where close contact occurs during the period of the parents' working day. The validity of the theory of familial occurrence of giardiasis was tested by carrying out studies in different geographical regions in Czechoslovakia. The results showed that in families with one reported case of giardiasis, the risk of other family members acquiring the infection was between 5.7 and 8.3 times higher than that experienced by the rest of the general population. The author suggests that the theory of familial occurrence of giardiasis is valid universally.
Introduction Giardia lamblia is historically important in Czechoslovakia. In 1859 the Czech physician Dušan Lambl described for the first time a parasite found in the stools of a child suffering from diarrhea. He named it Cercomonas intestinalis. By the end of the 1960's and during the 1970's Czechoslovak parasitologists paid more attention to the prevalence and epidemiology of giardiasis. The prevalence of giardiasis in Czechoslovakia has been reported from a number of sources (1,2,3,7,8,9,11,12,13,14,15,16) and the data are reported in Table 1. The prevalence rate within a particular age group is not always uniform. The prevalence rate of giardiasis is low among children of preschool age who remain at home. It is greater among children who attend daycare centres and the highest prevalence of giardiasis is usually found in children's homes (institutions). It is interesting to note that a special section of the Czechoslovak population is represented by gypsies. Many authors have studied giardiasis in gypsies (3,4,5,9) and they have found that the rate of giardiasis is 4 to 6 times higher among gypsies than in similar age groups within the same region. Discussion Why does the frequency of giardiasis vary within the same age groups of the population who live in different social and epidemiological conditions? Pazdiora and Palicka (11) state that in daycare centers (for children up to 3 years) the distribution curve of giardiasis reaches its peak by the third year of life. Fortynine percent of children who attend daycare centers for less than 1 year are reported to be infected compared with only 12.2% of children who have attended for more than two years. There does not appear to be such a striking difference in the prevalence of Giardia infection between children 3 to 5 years of age who attend preschool facilities and those attending kindergarten (7.3% vs. 10.1% on average). TABLE 1. Prevalence of Giardia lamblia in Czechoslovakia from surveys 1955 1984 (%).
0 2 years
Authors Family Cerva (1962)
Ditrich et al. (1984)
6.6
Giboda (1978)
11.3
Kvasz (1979, 1980)
Moravec (1980)
Daycare Center Children'sHomes 20.0
Giboda (1971)
7.3
3 5 years
43.0
Family
55.2
21.9
Pazdiora (1972)
7.6
Vošta (1955) Zitek, Palicka (1979) Average
17.8
3.0
6.7
13.2
32.8
4.5
3.2
7.3
10.1
4.0 23.7
19.0
7.3
4.1
10.0
3.2
3.5
2.0
6.6
15.2
2.7
14.3
11.2
12.2
>15 years
Children's homes
8.2
6.8
28.1
5.4
Family school
12.7
10 15 years
Orphan school
12.1 29.1
11.3 1.8
7.5
24.1
5.2
6.5 16.7
15.7
35.6
9.1
5.5
Volna, Ašmera (1968)
11.6
5.2
8.3
Pazdiora, Palicka (1971)
Škracikova et al. (1981)
Family Kindergartens Children'shomes school
27.3 15.0
6 9 years
3.7
6.8
2.5 10.0
3.2
Page 40 TABLE 2. Effectiveness of the search for further infection with Giardia among family members in the microfoci of giardiasis (Palicka 1973). Frequency in % Giardia positive families examined micro foci
Examined members of families
107
394
Family members with secondary without primary Giardia giardiasis infections 59
with primary Giardia infections (n=107)
15
prevalence of giardiasis in population (%)
Index prevalence in microfoci prevalence in population
42
1.8
8.3
Children in institutions represent a special case. Every author who studied the prevalence of Giardia in these institutions found that children of every age category suffered from giardiasis more frequently than those at daycare centres (Table 1). The permanent contact between the children in these homes creates better conditions for Giardia transmission than in daycare centers and kindergartens where close contact is restricted to the period of the parents' working day. This is also the case with other intestinal parasites such Pentatrichomonas hominis, Chilomastix mesnili and Hymenolepis nana which are commonly transmitted by the fecal oral mode of transmission (3). In their study of Giardia transmission in preschool facilities, Pazdiora and Palicka (1971) examined the staff of these establishments for intestinal parasites. The prevalence rate of giardiasis corresponded to the average rate among the adult population in the region studied. This suggests that the staff of preschool facilities do not play an active part in spreading Giardia among the children although they can influence transmission through food handling, changing diapers etc. Kvasz (6) examined all the members of families of children who were infected with Giardia, both from preschool facilities and hospitalized patients. He examined 790 members of 142 families and found that the average prevalence rate in those families was 23.07%. In another region Pazdiora (12) reported a Giardia frequency of 9.77% in families of Giardia positive children. In families with noninfected children the frequency of Giardia stood at only 3.21%. The effectiveness of antiepidemic measures among infected families was studied by Palicka (10) whose results are reproduced in Table 2. He examined 394 members of 107 families (micro foci) who were found to have at least one infected member. Additional cases of giardiasis were discovered in 15% of their family members. When 107 original infections were added, the frequency of giardiasis increased in such families to 42%. Since the average frequency of giardiasis in the rest of the population was 1.8%, this represented a 23 times higher frequency. The risk of becoming infected with Giardia in families with even a single case of giardiasis was computed to be 8.3 times higher (15/1.8%) than that experienced by the general population. Giboda (4) tested the validity of the theory of familial occurrence of giardiasis. The study was carried out in different geographical and epidemiological conditions from those of the authors mentioned above. The data are reported in Table 3. Of 157 members of 44 families in which one child was infected with giardiasis, new giardiasis was discovered in 31 persons from 25 families (56.8%). This means that the prevalence rate among the members of an infected family was 19.74%. Among newly discovered infections adults were more commonly infected than children (20.8% vs. 17.6%). In a control group (families without primary giardiasis) of 151 members of 35 families, new giardiasis was detected in three individuals only who belonged to three families. The prevalence rate in the noninfected families was therefore only 1.91%. A comparison of the average prevalence rate of Giardia in children up to 15 years of age in the study region (9.8%) with the prevalence of Giardia in children from infected families (primary plus secondary infections 55.8%) resulted in an index of 5.7 (55.8/9.8 = 5.7). This index was lower than the one found in Palicka's study (8.3). The frequency of giardiasis in both studies among infected families was similar (Palicka 33.1%; Giboda 37.3%) as were the overall prevalence rates of giardiasis in both regions of Czechoslovakia. These results have already had an influence on the practical activity of the Public Health Service in Czechoslovakia. In the event of Giardia and geohelminths diagnosis, all members of the family are parasitologically examined as well. This antiepidemic measurement is highly effective, especially in regions with low prevalence of intestinal parasites. TABLE 3. Giardia infections in family members of Giardia positive and Giardia negative children. Infected child in house
Giardia () adult
children
Giardia (+) total
adult
children
total examined total
adult
total
84
42
126
22
9
31
106
51
157*
no
77
71
148
2
1
3
79
72
151**
* Represents 44 households 25 (57%) of households had a second case attributed to index case ** Represents 35 households
children
yes
Page 41
Literature Cited 1. Cerva, L. 1962. Occurrence of intestinal parasites in the population of Central Bohemia. (In Czech, English summary). Cs. Parasitol. IX:135141. 2. Ditrich, O., J. Šterba, J. Prokopic, K. Kadlcik, and I. Maleckova. 1984. Intestinal parasitoses in South Bohemia farmer. (In Czech). IV. Prowazkovy dny, Komarno, 4.5.10.1984. 3. Giboda, M. 1971. The problem of intestinal parasites especial Protozoa in East Slovakia. (In Slovak). Thesis to B.Sc. 4. Giboda, M. 1978. Conditions of occurence of Giardiasis and ascariasis in children population of East Slovakia. (In Slovak). Thesis to Ph.D. 5. Jecny, V. 1965. The results of examination on intestinal parasites in some groups of population in district Most. (In Czech, English summary). Cs. Parasitol. XII:185195. 6. Kvasz, L. 1972. Contribution to lambliasis in Slovakia. (In Slovak). Thesis to Ph.D. 7. Kvasz, L. 1979. Accumulation of Giardiasis in families and closed collectives. (In Slovak, English summary). Bratisl. Lek. Listy 72:597600. 8. Kvasz, L., B. Petranska, M. Pavlina, A. Halasova, and J. Vodrazka. 1986. Screening of parasites of gastrointestinal tract in employs of large scale animal farms and the meat concern in the Nitra district. (In Slovak, English summary). Cs. Epidem. 35:5054. 9. Moravec, P. 1980. Prevalence of intestinal parasites in population of district Opava. (In Czech). Cas. Slez. muz. Opava (A) 29:5764. 10. Palicka, P. 1973. Effectivity of epidemiological work in the foci of intestinal parasitosis. (In Czech, English summary). Cs. Epidem. 22:3944. 11. Pazdiora, E., and P. Palicka. 1971. Notes to epidemiology of some intestinal parasitoses. (In Czech, English summary). Cs. Epidem. 20:216220. 12. Pazdiora, E. 1972. Some epidemiological aspects of occurrence of lambliasis in creche. (In Czech, English summary). Cs. Epidem. 21:271276. 13. Škracikova, J., S. Straka, E. Galikova, and G. Klimentova. 1981. Familial incidence of Giardiasis. Bratisl. Lek. Listy 76:369373. 14. Volna, L., and J. Ašmera. 1968. The occurrence of parasites at the population of the Ostrava region. (In Czech, English summary). Prirodoved. Sborn. (Ostrava):179183. 15. Vošta, J. 1955. Intestinal parasites of children in the surroundings of Tabor (In Czech). Cs. Parasitol. II:177180. 16. Zitek, K., and P. Palicka. 1979. Incidence of intestinal parasites in the population of a community and scope for influencing it. (In Czech, English summary). Cas. Lek. ces. 118:447450.
Page 43
IMMUNOLOGY
Page 45
Immunology of Giardia Infections Martin F. Heyworth Intestinal Immunology Research Center, Cell Biology Section, Veterans Administration Medical Center, San Francisco, California 94121 and Department of Medicine, University of California, San Francisco. Many studies have shown that human subjects infected with Giardia lamblia, and mice infected with Giardia muris, develop antibody responses to Giardia trophozoites. The present author has shown that immunocompetent BALB/c mice produce intestinal IgA and IgG antitrophozoite antibodies during G. muris infection. Such mice eliminate the infection. In contrast, athymic (nude) mice do not clear G. muris infection, and show little evidence of an intestinal antibody response to Giardia trophozoites. These observations suggest that antibodies play an important part in clearance of G. muris infection. Experiments in which BALB/c mice were selectively depleted of either helper/inducer (Th/i) or cytotoxic/suppressor (Tc/s) T lymphocytes by treatment with monoclonal antibodies, have shown that Th/i lymphocytes are necessary for clearance of G. muris by infected mice. Tc/s lymphocytes and natural killer cells are not required for elimination of this parasite from the mouse intestine. Important areas of future study include the following: (i) to determine whether intestinal antibodies are cytotoxic to Giardia trophozoites, and (ii) to identify and characterize trophozoite antigens which are major targets for mouse and human intestinal antiGiardia antibodies.
Introduction Human subjects become infected with the intestinal protozoan parasite Giardia lamblia by ingesting Giardia cysts. These can be acquired by drinking cyst contaminated water (11,42,49), or by fecal/oral contact, as in infant daycare centers (4,38). G. lamblia is an important cause of diarrhea in immunologically normal individuals (51), and patients with immunodeficiency diseases (particularly common variable hypogammaglobulinemia and Xlinked immunoglobulin deficiency) show increased susceptibility to giardiasis (18,31). Such immunodeficiency diseases predispose to chronic giardiasis, which can lead to severe, persistent diarrhea and malabsorption (18,31). The association of chronic giardiasis with immunodeficiency diseases suggests that immunological processes are responsible for clearing G. lamblia infections in immunologically normal individuals. This suggestion is strengthened by the demonstration of antiGiardia antibodies in immunologically normal human subjects (16,48). These antibodies include IgG antitrophozoite antibodies which occur in human sera (43), IgM antitrophozoite antibodies which are present in patients' sera during G. lamblia infection (17), and IgA antitrophozoite antibodies found in human milk (33). The functional significance of these antibodies is, however, unknown. Furthermore, very little is known about human immunological responses to G. lamblia at the site of the infection, namely in the gastrointestinal tract, although IgA has been demonstrated on Giardia trophozoites present on the epithelial surface of human jejunal biopsy specimens (6). Study of the pathophysiology of giardiasis, and of the immunological response to Giardia trophozoites, has been facilitated by the development of a mouse model of giardiasis. In this model system, mice are infected with the intestinal parasite Giardia muris (5,40). By analogy with G. lamblia infection in human subjects, G. muris trophozoites colonize the mouse small intestine (5,15,37). In immunocompetent mice, G. muris infection lasts for several weeks and the parasites are then cleared from the gastrointestinal tract (5,40). Mice with various types of immunodeficiency have an impaired ability to clear G. muris infection. The infection is chronic in athymic (nude) mice (39,46), in mice treated from birth with an antiserum directed against mouse IgM (44; such mice are deficient in IgM, IgA, and IgG), and in mice depleted of helper/inducer T lymphocytes (20). Such observations indicate that immunological events play an important part in the clearance of G. muris infection. Production and Role of AntiGiardia Antibodies Immunocompetent mice which are infected with G. muris produce antibodies directed against Giardia trophozoites. Such antibodies have been demonstrated in the serum, milk, and intestinal secretions of G. murisinfected mice (1,2,24,44,45). There is evidence that antibodies directed against G. muris trophozoites play a part in the clearance of G. muris infection. Thus, in mice treated with rabbit antiserum directed against mouse IgM, G. muris infection is persistent, and antibody production against G. muris trophozoites is impaired as judged by titers of trophozoitespecific antibody in the serum and intestinal secretions of treated mice (44). The present author has shown that IgA and IgG become bound to G. muris trophozoites in the intestinal lumen of Giardiainfected immunocompetent BALB/c mice, from day 10 of G. muris infection onwards (19). There is little evidence of IgA or IgG on trophozoites harvested from immunocompetent mice less than 10 days after the start of Giardia infection, suggesting that immunoglobulins detected
Page 46
on trophozoites later in the infection are Giardiaspecific antibody molecules, rather than immunoglobulins that are nonspecifically adsorbed to the trophozoites. There is little if any IgA or IgG on trophozoites harvested from the intestine of nude mice, at any time after the start of Giardia infection, suggesting that these mice have an impaired antibody response to Giardia trophozoites in the intestinal lumen (19). This impairment of antibody production may explain why nude mice are unable to clear G. muris infection. It has been shown that introduction of G. lamblia trophozoites into the duodenal lumen of rats leads to the appearance of trophozoitespecific IgA in rat bile (30). Although the source of this IgA is unknown, it may include IgA transported from serum to bile via the rat liver (35). It is likely that much of the intestinal trophozoitespecific IgA produced in rodents with Giardia infection arises from plasma cells in the intestinal mucosa. Numerous plasma cells are present in the lamina propria of the intestine in various mammalian species (9,10), and these cells are believed to originate from B lymphocytes in Peyer's patches (3). Carlson et al (7) have shown that the number of IgA+ cells in Peyer's patches of immunocompetent mice with G. muris infection increases before the infection is cleared. It is justifiable to speculate that the Peyer's patch IgA+ cells which increase in number during G. muris infection are precursors of intestinal mucosal plasma cells that secrete trophozoitespecific IgA. Although it is probable that trophozoitespecific antibody contributes to the clearance of Giardia infections, little is known about the mechanisms by which antibodies may eliminate trophozoites from the intestine. One theoretical possibility is that antibodies might inhibit adherence of Giardia trophozoites to the luminal surface of intestinal epithelial cells. There is recent evidence that Giardia trophozoites have a lectinlike surface molecule by which they bind to carbohydrate residues on mammalian cell membranes (14,29). This lectin may facilitate trophozoite attachment to the luminal surface of intestinal epithelial cells in vivo. If antibodies are directed against the lectin, they might inhibit this attachment. Similarly, antibodies directed against components of the trophozoite adhesive disk might impair attachment of the parasites to the intestinal epithelium (12,30,47). It has been shown that rabbit serum and mouse milk which contain antibodies directed against G. muris trophozoites inhibit adherence of trophozoites to mouse intestinal villi (23). Another possibility, that warrants investigation, is that antitrophozoite antibody may actually kill Giardia trophozoites in the intestinal lumen. It has been shown that monoclonal antibodies directed against a trophozoite surface antigen are able to kill G. lamblia trophozoites in vitro (34). Important areas of future study include the following: (i) to determine whether intestinal secretions and serum, from immunocompetent mice which have recently cleared G. muris infection, contain antibodies that are cytotoxic to G. muris trophozoites in vitro, and (ii) to determine whether antibodycoated G. muris trophozoites harvested from the intestinal lumen of immunocompetent mice (19) are viable or nonviable. Roles of T Lymphocytes, Natural Killer Cells, and Macrophages As noted above, nude mice lack the ability to clear G. muris infection, and become chronically infected with Giardia trophozoites (39,46). This observation indicates that T lymphocytes play an important part in clearance of G. muris infection from the gastrointestinal tract of immunocompetent mice, but does not identify the Tcell subset that is involved in clearance. Either helper/inducer (Th/i) or cytotoxic/suppressor (Tc/s) T lymphocytes, or conceivably both of these subsets, might be important. To identify the Tcell subpopulation that plays a major part in elimination of G. muris infection, the present author treated immunocompetent BALB/c mice with monoclonal antibody directed against either the mouse helper/inducer Tcell antigen L3T4 (13) or the cytotoxic/suppressor Tcell antigen Ly2 (27). This maneuver depletes L3T4+ or Ly2+ lymphocytes respectively (28,50). The Th/idepleted and Tc/sdepleted mice were then infected with G. muris cysts, and the timecourse of the infection was compared in these two groups of animals. It was found that Tc/sdepleted mice cleared the infection at the same rate as immunologically normal mice that were treated with phosphatebuffered saline. In contrast, mice depleted of Th/i lymphocytes became chronically infected, and continued to excrete large numbers of G. muris cysts for the duration of the study (20). These data indicate that helper/inducer T lymphocytes play a major role in the clearance of G. muris infection, and that cytotoxic/suppressor T cells are of little importance for the elimination of this infection. It is likely that the impaired ability of nude mice to mount an intestinal antibody response against G. muris trophozoites is the result of helper/inducer (L3T4+) Tcell deficiency. Nude mice are known to have a more profound deficiency of helper/inducer T lymphocytes than of cytotoxic/suppressor (Ly2+) T lymphocytes (8,32). Preliminary studies by the present author suggest that immunocompetent BALB/c mice which have been depleted of Th/i lymphocytes, by treatment with antiL3T4 monoclonal antibody, have an impaired ability to mount an intestinal antibody response against G. muris trophozoites (M.F. Heyworth, unpublished data). Natural killer (NK) cells are present in the intestinal mucosa of immunocompetent mice. However, little is known about the role of NK cells in the gastrointestinal tract. To determine whether NK cells play a part in the elimination of G. muris infection, the timecourse of this infection has been studied in beige mice, which are deficient in NK cells (41). The results of this work show that beige mice eliminate G. muris infection at the same rate as mice with normal NKcell activity (21). This finding strongly suggests that NK cells are not involved in the clearance of G. muris infection from the mouse intestine.
Page 47
In an attempt to determine whether macrophages are important for the clearance of G. muris infection, the present author harvested leukocytes from the intestinal lumen of Giardiainfected immunocompetent BALB/c mice and nude mice. Leukocyte subsets were then identified by immunofluorescent staining with monoclonal antibodies directed against leukocyte surface antigens, including the macrophage surface antigen Mac1 (22). Leukocytes bearing this antigen were quantified by fluorescence microscopy. This work showed that there was no appreciable difference between the number of Mac1+ cells harvested from the intestinal lumen of BALB/c mice or nude mice, and that only small numbers of cells bearing the Mac1 antigen were present in cell suspensions harvested from the mouse intestinal lumen (25 × 103 Mac1+ cells per mouse; 22). These observations suggest that intraluminal macrophages do not play an important effector role in the elimination of G. muris infection. Transmission electron microscopy of Peyer's patch sections from Giardiainfected mice has shown that macrophages in the patches are able to phagocytose Giardia trophozoites (36). Because Peyer's patches are known to be important sites for the initiation of intestinal immune responses (25,26), it is likely that ingestion of trophozoites by Peyer's patch macrophages is followed by presentation of trophozoite antigens to local helper/inducer T cells and B cells, with subsequent production of Giardiaspecific antibodies in normal immunocompetent mice. Conclusions There is extensive evidence that human subjects infected with Giardia lamblia, and mice infected with G. muris, develop antibody responses to Giardia trophozoites. The ability of mice to mount an antibody response to Giardia trophozoites correlates with the ability of these animals to eliminate G. muris infection. The observation that nude mice are unable to clear G. muris infection indicates that T lymphocytes play an important part in antiGiardia immunity. Studies in which immunocompetent BALB/c mice were selectively depleted of either helper/inducer or cytotoxic/suppressor T lymphocytes, by treatment with monoclonal antibody, have shown that clearance of G. muris infection is dependent on helper/inducer T cells. Cytotoxic/suppressor T cells and natural killer cells play little, if any, part in elimination of G. muris infection from the mouse intestine. Important areas of future study include: (a) to determine whether Giardia trophozoites are killed by trophozoitespecific antibodies present in mouse or human intestinal secretions, and (b) to characterize trophozoite antigens which are major targets for intestinal anti Giardia antibodies. Acknowledgements Grant support from the National Institutes of Health (grants AM33930 and AM33004) and from the Academic Senate Committee on Research of the University of California, San Francisco, is gratefully acknowledged. Literature Cited 1. Anders, R.F., I.C. RobertsThomson, and G.F. Mitchell. 1982. Giardiasis in mice: analysis of humoral and cellular immune responses to Giardia muris. Parasite Immunol. 4: 4757. 2. Andrews, J.S.,Jr., and E.L. Hewlett. 1981. Protection against infection with Giardia muris by milk containing antibody to Giardia. J. Infect. Dis. 143: 242246. 3. Bienenstock, J., and A.D. Befus. 1980. Mucosal immunology. Immunology 41: 249270. 4. Black, R.E., A.C. Dykes, S.P. Sinclair, and J.G. Wells. 1977. Giardiasis in daycare centers: evidence of person to person transmission. Pediatrics 60: 486491. 5. Brett, S.J., and F.E.G. Cox. 1982. Immunological aspects of Giardia muris and Spironucleus muris infections in inbred and outbred strains of laboratory mice: a comparative study. Parasitology 85: 8599. 6. Briaud, M., M. MorichauBeauchant, C. Matuchansky, G. Touchard, and P. Babin. 1981. Intestinal immune response in giardiasis. Lancet ii: 358. 7. Carlson, J.R., M.F. Heyworth, and R.L. Owen. 1986. Response of Peyer's patch lymphocyte subsets to Giardia muris infection in BALB/c mice. II. Bcell subsets: enteric antigen exposure is associated with immunoglobulin isotype switching by Peyer's patch B cells. Cell. Immunol. 97: 5158. 8. Carlson, J.R., M.F. Heyworth, and R.L. Owen. 1987. Tlymphocyte subsets in nude mice with Giardia muris infection. Thymus 9:189196. 9. Crabb'e, P.A., D.R. Nash, H. Bazin, H. Eyssen, and J.F. Heremans. 1969. Antibodies of the IgA type in intestinal plasma cells of germfree mice after oral or parenteral immunization with ferritin. J. Exp. Med. 130: 723744. 10. Crago, S.S., W.H. Kutteh, I. Moro, M.R. Allansmith, J. Radl, J.J. Haaijman, and J. Mestecky. 1984. Distribution of IgA1, IgA2, and J chaincontaining cells in human tissues. J. Immunol. 132: 1618. 11. Craun, G.F. 1979. Waterborne giardiasis in the United States: a review. Am. J. Public Health 69: 817819. 12. Crossley, R., and D. Holberton. 1985. Assembly of 2.5 nm filaments from giardin, a protein associated with cytoskeletal microtubules in Giardia. J. Cell Sci. 78: 205231. 13. Dialynas, D.P., Z.S. Quan, K.A. Wall, A. Pierres, J. Quintans, M.R. Loken, M. Pierres, and F.W. Fitch. 1983. Characterization of the murine T cell surface molecule, designated L3T4, identified by monoclonal antibody GK1.5: similarity of L3T4 to the human Leu3/T4 molecule. J. Immunol. 131: 24452451. 14. Farthing, M.J.G., M.E.A. Pereira, and G.T. Keusch. 1986. Description and characterization of a surface lectin from Giardia lamblia. Infect. Immun. 51: 661 667. 15. Fleck, S.L., S.E. Hames, and D.C. Warhurst. 1985. Detection of Giardia in human jejunum by the immunoperoxidase method. Specific and nonspecific results. Trans. R. Soc. Trop. Med. Hyg. 79: 110113. 16. Gilman, R.H., K.H. Brown, G.S. Visvesvara, G. Mondal, B. Greenberg, R.B. Sack, F. Brandt, and M.U. Khan. 1985. Epidemiology and serology of Giardia lamblia in a developing country: Bangladesh. Trans. R. Soc. Trop. Med. Hyg. 79: 469473. 17. Goka, A.K.J., D.D.K. Rolston, V.I. Mathan, and M.J.G. Farthing. 1986. Diagnosis of giardiasis by specific IgM antibody enzymelinked immunosorbent assay. Lancet ii: 184186. 18. Hermans, P.E., J.A. DiazBuxo, and J.D. Stobo. 1976. Idiopathic lateonset immunoglobulin deficiency: clinical observations in 50 patients. Am. J. Med. 61: 221237. 19. Heyworth, M.F. 1986. Antibody response to Giardia muris trophozoites in mouse intestine. Infect. Immun. 52: 568571.
Page 48
20. Heyworth, M.F., J.R. Carlson, and T.H. Ermak. 1987. Clearance of Giardia muris infection requires helper/inducer T lymphocytes. J. Exp. Med. 165: 1743 1748. 21. Heyworth, M.F., J.E. Kung, and E.C. Eriksson. 1986. Clearance of Giardia muris infection in mice deficient in natural killer cells. Infect. Immun. 54: 903904. 22. Heyworth, M.F., R.L. Owen, and A.L. Jones. 1985. Comparison of leukocytes obtained from the intestinal lumen of Giardiainfected immunocompetent mice and nude mice. Gastroenterology 89: 13601365. 23. Kaplan, B., and D. Altmanshofer. 1985. Giardia muris adherence to intestinal epithelium the role of specific antiGiardia antibodies. Microecology and Therapy 15: 133140. 24. Kaplan, B.S., S. Uni, M. Aikawa, and A.A.F. Mahmoud. 1985. Effector mechanism of host resistance in murine giardiasis: specific IgG and IgA cellmediated toxicity. J. Immunol. 134: 19751981. 25. Keren, D.F., P.S. Holt, H.H. Collins, P. Gemski, and S.B. Formal. 1978. The role of Peyer's patches in the local immune response of rabbit ileum to live bacteria. J. Immunol. 120: 18921896. 26. Kiyono, H., J.R. McGhee, M.J. Wannemuehler, M.V. Frangakis, D.M. Spalding, S.M. Michalek, and W.J. Koopman. 1982. In vitro immune responses to a T celldependent antigen by cultures of disassociated murine Peyer's patch. Proc. Natl. Acad. Sci. USA 79: 596600. 27. Ledbetter, J.A., and L.A. Herzenberg. 1979. Xenogeneic monoclonal antibodies to mouse lymphoid differentiation antigens. Immunol. Rev. 47: 6390. 28. Ledbetter, J.A., and W.E. Seaman. 1982. The Lyt2, Lyt3 macromolecules: structural and functional studies. Immunol. Rev. 68: 197218. 29. Lev, B., H. Ward, G.T. Keusch, and M.E.A. Pereira. 1986. Lectin activation in Giardia lamblia by host protease: a novel hostparasite interaction. Science 232: 7173. 30. Loftness, T.J., S.L. Erlandsen, I.D. Wilson, and E.A. Meyer. 1984. Occurrence of specific secretory immunoglobulin A in bile after inoculation of Giardia lamblia trophozoites into rat duodenum. Gastroenterology 87: 10221029. 31. LoGalbo, P.R., H.A. Sampson, and R.H. Buckley. 1982. Symptomatic giardiasis in three patients with Xlinked agammaglobulinemia. J. Pediatr. 101: 7880. 32. MacDonald, H.R., C. Blanc, R.K. Lees, and B. Sordat. 1986. Abnormal distribution of T cell subsets in athymic mice. J. Immunol. 136: 43374339. 33. Miotti, P.G., R.H. Gilman, L.K. Pickering, G. RuizPalacios, H.S. Park, and R.H. Yolken. 1985. Prevalence of serum and milk antibodies to Giardia lamblia in different populations of lactating women. J. Infect. Dis. 152: 10251031. 34. Nash, T.E., and A. Aggarwal. 1986. Cytotoxicity of monoclonal antibodies to a subset of Giardia isolates. J. Immunol. 136: 26282632. 35. Orlans, E., J.V. Peppard, A.W.R. Payne, B.M. Fitzharris, B.M. Mullock, R.H. Hinton, and J.G. Hall. 1983. Comparative aspects of the hepatobiliary transport of IgA. Ann. N. Y. Acad. Sci. 409: 411427. 36. Owen, R.L., C.L. Allen, and D.P. Stevens. 1981. Phagocytosis of Giardia muris by macrophages in Peyer's patch epithelium in mice. Infect. Immun. 33: 591 601. 37. Owen, R.L., P.C. Nemanic, and D.P. Stevens. 1979. Ultrastructural observations on giardiasis in a murine model. I. Intestinal distribution, attachment, and relationship to the immune system of Giardia muris. Gastroenterology 76: 757769. 38. Pickering, L.K., W.E. Woodward, H.L. DuPont, and P. Sullivan. 1984. Occurrence of Giardia lamblia in children in day care centers. J. Pediatr. 104: 522 526. 39. RobertsThomson, I.C., and G.F. Mitchell. 1978. Giardiasis in mice. I. Prolonged infections in certain mouse strains and hypothymic (nude) mice. Gastroenterology 75: 4246. 40. RobertsThomson, I.C., D.P. Stevens, A.A.F. Mahmoud, and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterology 71: 5761. 41. Roder, J.C. 1979. The beige mutation in the mouse. I. A stem cell predetermined impairment in natural killer cell function. J. Immunol. 123: 21682173. 42. Shaw, P.K., R.E. Brodsky, D.O. Lyman, B.T. Wood, C.P. Hibler, G.R. Healy, K.I.E. MacLeod, W. Stahl, and M.G. Schultz. 1977. A communitywide outbreak of giardiasis with evidence of transmission by a municipal water supply. Ann. Intern. Med. 87: 426432. 43. Smith, P.D., F.D. Gillin, W.R. Brown, and T.E. Nash. 1981. IgG antibody to Giardia lamblia detected by enzymelinked immunosorbent assay. Gastroenterology 80: 14761480. 44. Snider, D.P., J. Gordon, M.R. McDermott, and B.J. Underdown. 1985. Chronic Giardia muris infection in antiIgMtreated mice. I. Analysis of immunoglobulin and parasitespecific antibody in normal and immunoglobulindeficient animals. J. Immunol. 134: 41534162. 45. Snider, D.P., and B.J. Underdown. 1986. Quantitative and temporal analyses of murine antibody response in serum and gut secretions to infection with Giardia muris. Infect. Immun. 52: 271278. 46. Stevens, D.P., D.M. Frank, and A.A.F. Mahmoud. 1978. Thymus dependency of host resistance to Giardia muris infection: studies in nude mice. J. Immunol. 120: 680682. 47. Torian, B.E., R.C. Barnes, R.S. Stephens, and H.H. Stibbs. 1984. Tubulin and highmolecularweight polypeptides as Giardia lamblia antigens. Infect. Immun. 46: 152158. 48. Visvesvara, G.S., P.D. Smith, G.R. Healy, and W.R. Brown. 1980. An immunofluorescence test to detect serum antibodies to Giardia lamblia. Ann. Intern. Med. 93: 802805. 49. Vogt, R.L., A.A. Little, K.C. Spitalny, and G. Visvesvara. 1984. Investigation of a waterborne outbreak of giardiasis using serologic testing by IFA. Am. J. Public Health 74: 272. 50. Wofsy, D., D.C. Mayes, J. Woodcock, and W.E. Seaman. 1985. Inhibition of humoral immunity in vivo by monoclonal antibody to L3T4: studies with soluble antigens in intact mice. J. Immunol. 135: 16981701. 51. Wolfe, M.S. 1978. Current concepts in parasitology. Giardiasis. N. Engl. J. Med. 298: 319321.
Page 49
The Secretory Immune Response in Rats Infected with Rodent Giardia duodenalis Isolates and Evidence for Passive Protection with Immune Bile Graham Mayrhofer* and Agnes Waight Sharma Department of Microbiology and Immunology, The University of Adelaide, Box 498, G.P.O., Adelaide, South Australia, 5001, Australia. Two isolates of Giardia, one from mice and one from rats, have been identified as Giardia duodenalis by morphological criteria. They are identical with each other and with a human and a feline isolate by isoenzyme analysis. Nevertheless, the rat isolate produced a chronic infection in all seven inbred rat strains tested, while the mouse isolate in each case produced an acute infection. Infections with both isolates were chronic in congenitally hypothymic nude rats. Natural or metronidazoleinduced termination of primary infections with either organism was followed by a high level of immunity to reinfection. IgM and IgA antibody responses to homologous trophozoite antigens have been measured in serum by enzymelinked immunoabsorbent assays during primary and secondary infections with both isolates. However, only IgA antibodies were detected in bile from infected rats. Immune bile, but not normal bile, led to a substantial fall in faecal cyst excretion when infused into the duodena of conscious rats infected with the mouse isolate. The findings suggest that secretory antibodies are protective and that comparisons between the immune responses against these closely related rodent isolates may help define protective antigens.
Introduction Most of the Giardia that infect mammals have similar morphology and they have been grouped into a single species, Giardia duodenalis (5). By this classification, the organisms responsible for giardiasis in man are referred to as G. duodenalis (lamblia). G. duodenalis has been divided into races on the basis of supposed host specificity and by morphometry (20). However, the relationships between the tentative species are only now being explored by more sophisticated methods such as isoenzyme analysis (3,6), restriction endonuclease analysis of DNA and genome probing with cloned fragments of Giardia DNA (9). Even within isolates from humans, these methods have revealed evidence of genetic diversity, while the usefulness of host specificity is now recognized to be limited. Most of the experimental work on the immunology of giardiasis has been carried out in mice, using isolates presumed to be Giardia muris. G. muris can be separated easily from G. duodenalis on morphological grounds and is therefore a distinct species. This may affect its value as an experimental model for human giardiasis, especially in work aimed at defining antigens responsible for inducing protective immunity. Although antigens are known to be shared between G. muris and G. duodenalis (lamblia), the extent of sharing is not known (13). There may therefore be value in using animal isolates of G. duodenalis as models for human disease. This would allow studies of infection in the natural host with organisms that are potentially more closely related to G. duodenalis (lamblia) than is G. muris. The most readily available parasitehost combination is G. duodenalis (simoni) in the rat. The choice of the rat has the added advantage that secretory immunity is well understood in this species. In particular, secretory antibody responses to intestinal infections can be measured in bile because most of the IgA that enters the intestine in rats is transported there from the blood via the liver (8). There is evidence that the rodent isolates of G. duodenalis used in this study are in fact very similar to G. duodenalis (lamblia). This raises the possibility that protective antigens identified in this model may be more closely similar to those of G. duodenalis (lamblia) than would be the case for G. muris. Materials and Methods Animals The specific pathogen free (SPF) rats used to study the kinetics of cyst excretion after infection with Giardia isolates were 810 week old females obtained from the Animal Resources Centre, Western Australia. No evidence of any parasitic infection was found in these animals. The strains used were Fischer 344 (F344, RT1lvl), WAG (RT1u), Lou/M (RT1u), Wistar Furth (WF, RT1u), Brown Norway, (BN, RT1n), PVG/c (RT1c), and DA (RT1avl). Hypothymic nude rats (CBHmu/mu) came from the same source. Female 810 week old rats used to study antibody responses were obtained from the Gilles Plains Animal Resource Centre, South Australia. These were believed to be SPF but they were subseqently found to excrete cysts of Entamoeba spp. They were free from all other parasites including Giardia spp. During experiments, animals were housed in a clean conventional room on sterilized litter and had free access to sterilized water and food. Giardia Isolates and Parasitological Techniques Two isolates of G. duodenalis have been studied. The mouse Giardia was isolated from randombred albino mice in a nonlaboratory colony and the rat Giardia was isolated from a conventional laboratory colony of inbred Ginger Hooded rats. The isolates have been maintained by serial passage through nude rats at 34 month intervals. For cyst counts, faeces collected from individual
Page 50
rats over a 2 hour period were emulsified in 0.01% Tween 20 in distilled water and the cysts concentrated by centrifugation over 1 M sucrose (14). The cysts were then counted in a haemocytometer by phase contrast microscopy and the counts are expressed as the log10 mean cysts per gram of faeces for the animals in each group. To infect animals, cysts were concentrated from faecal suspensions as described above. After counting, the concentration was adjusted to allow intragastric intubation of 5000 cysts in 0.05 mL of distilled water. To completely eradicate the primary infection, animals were treated with 50 mg of metronidazole by intragastric intubation on three consecutive days. Morphological Studies Trophozoites were obtained by excystation from purified cysts, as described below. Small amounts of suspension were partially airdried on slides and fixed in Schaudin's fixative. The smears were then stained with Trichrome (19) to identify the median bodies. Isoenzyme Studies Trophozoites were excysted from cysts purified from rat faeces by initial concentration over 1 M sucrose (see above), followed by repeated sedimentation at unit gravity through Percoll gradients (15). The cysts were held in distilled water containing penicillin (200 µg/ml), gentamicin (200 µg/ml) and amphotericin B (2 µg/ml) for 3 days and were shown to be bacteriologically sterile. Excystation was performed essentially as described by Schaefer, Rice and Hoff (16). Aliquots containing 5 × 107 trophozoites were snapfrozen in dry iceacetone and transported on dry ice to the Evolutionary Biology Unit of the South Australian Museum. Enzyme analysis was performed by electrophoresis on cellulose acetate gels using a sonicate of the organisms, essentially as described elsewhere (11). Surgical Procedures and Specimen Collection At the time of bile duct cannulation, blood was collected from the tailtip under ether anaesthesia. The bile duct was then approached by a midline incision and cannulated as near the porta hepatis as possible with a polythene cannula. Approximately 2 mL of bile was collected immediately into icechilled tubes from the conscious animals held in Bollman metabolic cages. Bile and serum samples were frozen in dry iceacetone and held at 100°C until assayed for antibodies. Further bile was then collected into icechilled containers for 24 days. This material was also frozen and stored for use in passive transfer experiments. Intraduodenal infusion of bile was achieved through a cannula connected to a peristaltic pump delivering 0.5 mL per hour. The cannula consisted of medical grade polyethylene tubing (0.4 mm internal diameter, 0.8 mm external diameter), tipped with 2.5 cm of soft silicone rubber tubing, (Silastic, DowCorning; 0.012 in. internal diameter, 0.025 in. external diameter). An anterior midline abdominal incision was made under ether anaesthesia. The cannula was passed through the posterior abdominal wall on the left side and into the stomach through a puncture in the antral region. The cannula was fed through the pylorus and anchored with a purse string suture as it entered the stomach. The end of the silicone tubing was adjusted to lie in proximity to the entry of the common bile duct. The animals were held unanaesthetized and with free access to food and water in Bollman metabolic cages. At the conclusion of the period of bile infusion, cannulae were removed by quick traction and the animals returned to individual holding cages. Antigens Antigens for coating ELISA plates were prepared from cysts purified by unit gravity sedimentation through Percoll gradients (see above). The purified cysts were stored at 4°C in distilled water containing antibiotics (see above) and used within 7 days. Trophozoites were excysted, suspended in phosphate buffered saline (PBS, 400 mOsm/L), adjusted to 2×106 organisms per mL and sonicated. Preliminary experiments showed that sonicates prepared in this way optimally sensitized ELISA trays for antiGiardia antibody estimations. Immunological Reagents Rabbit antirat IgA was raised against IgA purified from rat thoracic duct lymph and it was absorbed by passage through 2 Sepharose 4B columns, one coated with normal rat serum proteins and the other with purified rat IgG (all classes). Pure antiIgA antibody was prepared by adsorption to and elution from an IgA Sepharose 4B column. The antibody was conjugated with alkaline phosphatase (Calf Intestine VIIS, Sigma Chemical Co., Missouri, USA) by the one step glutaraldehyde procedure (1). The conjugate of rabbit antirat IgA with alkaline phosphatase was prepared as above, using immunopurified antibody kindly provided by Dr. D.W. Mason (Oxford). The antiIgA and antiIgM conjugates were isotypespecific when tested by ELISA in wells coated with optimal amounts of purified rat IgA, IgG, IgM or IgE. Enzyme Linked Immunoabsorbent Assay (ELISA) Assays were carried out using 96 well roundbottomed vinyl microtitre plates (Costar, Data Packaging Corporation, Cambridge, Mass.) coated with antigen by incubation with 100 µL of Giardia sonicate for 1 hour at 37°C and then overnight at 4°C. After washing with normal PBS containing 0.05% Tween 20 and 0.05% sodium azide (PBSTween 20 buffer), free binding sites on the plates were blocked by incubation with 1% bovine serum albumen (BSA Fraction V, Flow Laboratories, NSW, Australia) in PBSTween 20 for 68 hours at 4°C. To assay antibodies in serum or bile, 2fold serial dilutions in PBSTween 20 containing 1% BSA were incubated at 4°C overnight in wells coated with the homologous trophozoite antigen preparation. After washing, bound antibody was detected by a further incubation overnight at 4°C with predetermined dilutions of alkaline phosphataseantibody conjugates. Substrate (pnitrophenylphosphate disodium, Sigma Chemical Co., in 10% diethanolamine buffer) was added in 100 µL aliquots to the washed wells and further incubated for 4 hours at 37°C. Optical densities of wells were read at 405 nm using a Titertek Multiscan automated spectrophotometer adjusted to zero on a substrate blank. A positive antibody titre in any sample was defined as the reciprocal of the dilution which produced a mean optical density of 0.150, representing twice the mean OD405 produced by conjugates reacting in antigencoated wells without added rat antibodies. Results Characterization of the Two Giardia Isolates Trophozoites from the two isolates are shown in Figure 1. Morphometry has not been performed, but they appear very similar in shape and size. In particular, the median bodies in both isolates have the ''claw hammer" appearance and this places them in the G. duodenalis group. Evidence supporting a close relationship between the isolates came from comparison of the electrophoretic mobilities of the 27 enzymes shown in Table 1. Differences in the electrophoretic mobilities of enzymes between individuals or between species (i.e. isoenzymes)
Figure 1. Examples of trophozoites excysted in vitro from the rodent isolates. a) Mouse isolate. b) Rat isolate. Both have median bodies characteristic of G. duodenalis and they have similar general morphological features.
Page 51 TABLE 1. Enzymes identified in the G. duodenalis isolates. Enzyme
E.C. No.
Enzyme
E.C. No.
Aconitase
4.2.1.3
Glutathione Reductase
Acid Phosphatase
3.1.3.2
Hexokinase
2.7.1.1
Adenosine Deaminase
3.5.4.4
Malate Dehydrogenase
1.1.1.27
Alcohol Dehydrogenase
1.1.1.1
Malic Enzyme
1.1.1.40
Aldolase
4.1.2.13
MannosePhosphate Isomerase
5.3.1.8
Enolase
4.2.1.11
Nucleoside Phosphorylase
2.4.2.1
FructoseDiphosphatase
3.1.3.11
Peptidase (Valineleucine)
3.4.11or13
Glyceraldehyde3Phosphate Dehydrogenase
1.2.1.12
Phosphoglycerate Mutase
2.7.5.3
Glutamate Dehydrogenase
1.4.1.3
6Phosphogluconate Dehydrogenase
1.1.1.44
GlutamateOxaloacetate Transaminase
2.6.1.1
Phosphoglycerate Kinase
2.7.2.3
Glucose6Phosphate Dehydrogenase
1.1.1.49
Phosphoglucomutase
2.7.5.1
Glycerophosphate Dehydrogenase
1.1.1.8
Sorbitol Dehydrogenase
1.1.1.14
5.3.1.9
TriosePhosphate Isomerase
5.3.1.1
Uridine Monophosphate Kinase
2.7.4.?
GlucosePhosphate Isomerase
1.6.4.2
reflect corresponding structural differences between the genes that encode those enzymes. No differences were noted between the rat and mouse isolates for any of the 27 enzymes examined and identity at this number of loci suggests strongly that both organisms belong to the same species. Furthermore, there were no differences at any of these loci when the two rodent isolates were compared with a human isolate (Adelaide1) or with the feline isolate Portland1 (E.A. Meyer, Portland). The rodent isolates of G. duodenalis therefore appear to be closely related to G. duodenalis (lamblia) and to a G. duodenalis from cats. Infections in Normal Rats The duration of infection with G. muris varies, depending on which strains of mice are studied (2,10,12). Primary infections with the mouse (Figure 2) and rat (Figure 3) isolates of G. duodenalis have been studied in the strains of inbred rats that were available. The strain had little influence on either the magnitude or duration of cyst excretion by rats during infections with either isolate. However, when the infections produced by the two isolates were compared, there were striking differences. With the mouse isolate, cyst excretion reached a peak quickly and within a week commenced to decline, until finally it ceased after approximately 56 weeks. In contrast, all rat strains became infected chronically with the rat isolate and showed no evidence of expelling the organisms. The infection has been observed to persist at a similar level for at least 6 months in F344 strain rats (data not shown). Infections with both isolates were treated with metronidazole after 10 weeks. When challenged with the homologous organism 2 weeks later, the animals were found to be highly resistant to reinfection. Some animals in most strains failed to excrete cysts at detectable levels and where excretion did occur, it was in small numbers and for only a few days. In more recent experiments (not shown), where 4 weeks have been allowed between metronidazole treatment and reinfection, resistance after a primary infection has been complete.
Figure 2. The course of infection with the mouse isolate of G. duodenalis in 7 normal inbred rat strains and in congenitally hypothymic nude mu/mu rats. Primary and secondary infections were both initiated with 5000 cysts. Rats were treated with metronidazole 10 weeks after commencement of the primary infection and were challenged with a second dose of cysts 2 weeks later. (•) course of primary infection. course of secondary infection. Limits of detection of cysts in faeces are indicated by horizontal dotted lines. Points are mean for 5 animals, ± standard error.
Page 52
Infections in Hypothymic Nude Rats Nude rats of the CBH mu/mu strain have been used to maintain stocks of the two G. duodenalis isolates. As shown in Figures 2 and 3, these rats continued to excrete high levels of cysts for long periods of time after infection with either organism. In particular, the infection produced by the mouse isolate was dramatically different when compared in thymusdeficient and immunocompetent rat strains. Serum and Biliary Antibody Responses Titres of IgM and IgA antibodies were estimated by isotype specific ELISA in the serum and bile of rats infected with the mouse and rat isolates (Figures 4 and 5). During primary infections with either organism there was an early rise in IgM antiGiardia antibodies in serum, which declined between 10 and 15 days after infection. This was followed by a rise in serum IgA antibodies, which was sustained until the time of metronidazole treatment 6 weeks after infection. Secondary infections produced serum IgM antibody responses similar to those following primary infections. However, the IgA antibody responses in serum were secondary in character (i.e. large and more rapid). IgM antibodies were not detected in bile at any time during primary or secondary infections with either isolate. IgA antibody levels in bile rose in parallel with the levels in serum. A secondary IgA antibody response was detected in the bile of rats undergoing secondary infections with either isolate. Prior to infection, the rats used in these experiments had significant levels of IgM antibody in serum and of IgA antibody in bile. These low levels of specific or cross reactive antibody appear related to the source of animals. SPF rats of the same strain from a different source (Animal Resources Centre, Western Australia) were found to have levels of serum and biliary antibodies near to the backgrounds of the ELISA assays (data not shown). Passive Protection With Immune Bile In this experiment, 8 rats were infected with the mouse isolate of G. duodenalis. Cannulae were inserted into the duodena of the animals on day 8 after infection. Four rats received approximately 12 mL of bile per day from a pool collected from the animals with the primary infections in the previous experiment. The ELISA titre of IgA antibody in this pool was 2048. The remaining 4 rats received the same volume of bile, but from a pool of uninfected donors. The rate of bile infusion approximated the normal daily bile output. Infusion was continued for a period of 84 hours, after which the cannulae were removed and the animals were returned to holding cages. Faecal cyst excretion during the course of infection in the two groups is shown in Figure 6. Infusion of control bile from uninfected donors had no effect on the rate of cyst excretion. In contrast, in animals receiving immune bile there was a very significant decline in cyst excretion, commencing towards the end of the infusion period. Animals that had received immune bile continued to excrete fewer cysts until the end of the infection, when cyst excretion declined in both groups.
Figure 3. The course of infection with the rat isolate of G. duodenalis in 6 normal inbred rat strains and in congenitally hypothymic nude CBH mu/mu rats. All details are as described in the legend to Figure 2. (•) course of primary infection. course of secondary infection.
Figure 4. Specific antibody titres, against homologous trophozoite antigens, in the sera and bile of female DA rats infected with the mouse isolate of G. duodenalis. a) Primary infection. b) Secondary infection. The primary infection with 5000 cysts was terminated after 6 weeks with metronidazole and the animals were reinfected with the same dose of cysts 4 weeks later. Antibody titres were measured using isotypespecific enzymelinked immunoabsorbent assays. (Dotted lines) measurements on sera, (Solid lines) measurements on bile, ( ) IgA antibodies, (•) IgM antibodies. Points are mean ± standard error for 5 animals.
Page 53
Figure 5. Specific antibody titres, against homologous trophozoite antigens, in the sera and bile of female DA rats infected with the rat isolate of G. duodenalis. a) Primary infection. b) Secondary infection. All other details as described in legend to Figure 4. (Dotted lines) measurements on sera, (Solid lines) measurements on bile, ( ) IgA antibodies, (•) IgM antibodies. Points are mean ± standard error for 5 animals.
Figure 6. The effect of passive intraintestinal immunization with immune bile on the course of infection with the mouse isolate in female DA rats. The rats in each group were infected with 5000 cysts and 7 days later a cannula was introduced into the duodenum. Immune bile (Dotted line), or control normal bile (Solid line), was infused at the rate of 12 mL per day for the period indicated by the horizontal bar. The cannulae were then withdrawn and cyst excretion was monitored for the period indicated. Horizontal dotted line indicates limits of cyst detection. Points are mean ± standard error for 4 animals.
Discussion Important findings with regard to speciation of G. duodenalis have emerged from this study. Although related by morphology and at all enzyme loci tested, the two rodent isolates cause very different infections in rats. The infection produced by the mouse isolate was acute. This isolate also produces an acute infection in BALB/c mice (unpublished results), similar to that described for G. muris (12). A further similarity to G. muris was that the mouse isolate produced a chronic infection in C3H/HeJ mice that was only cleared after approximately 15 weeks (unpublished results). In contrast, the rat isolate produced a chronic infection in rats, with no evidence of resolution after many weeks. In BALB/c mice, this isolate produced an acute infection similar to that caused by the mouse isolate, although in C3H/HeJ mice the infection with the rat isolate was less chronic, resolving in approximately 12 weeks (unpublished results). In general terms, each organism appears to produce a more chronic infection in its host of origin, but this is more pronounced with the rat isolate. No evidence was obtained for large differences in susceptibility to infection between rat strains, in contrast to the findings in mice (2,10,12). Comparative studies between these organisms may therefore provide important clues to the nature of virulence determinants in Giardia. The findings suggest that parasite factors, as well as host factors, influence the chronicity of giardiasis. They raise the possibility that similar parasiterelated factors may be responsible for the chronicity of infection in some cases of human giardiasis. Furthermore, it is clear that differences between isolates producing acute or chronic infections in man could be quite subtle and they could be missed by techniques such as isoenzyme analysis. A potential advantage of this model of giardiasis is that comparisons can be made of the immune responses to infection with the two closely related organisms in the same strain of rat. This may allow identification of the antigens against which the protective immune response is directed. A similar strategy has been attempted with G. muris, when the immune response was compared between strains of mice that differed in their susceptibilities to infection with the one organism (4). Previous isoenzyme studies have suggested a close relationship between some human isolates and the feline Portland1 strain (3). In the present study, examining a greater number of enzymes, Portland1 was found to be identical to a human isolate (Adelaide1) and to the two rodent isolates of G. duodenalis. It is therefore likely that some human isolates are closely related to Giardia that infect other mammalian species. The similarity of the rodent G. duodenalis isolates to G. duodenalis (lamblia) may therefore assist in defining the protective antigens in human giardiasis. The mechanism of protective immunity in giardiasis is still poorly understood. The chronic infections in hypothymic rats suggest a role for T lymphocytes, as has been suggested from studies on murine giardiasis (12,18). However, further studies, including antibody measurements in sera and secretions, will be necessary to identify the precise role of T cells. Evidence of the role of antibody in protection has been strengthened by recent studies in immunoglobulindeficient mice (17). Levels of IgM and IgA antibodies have therefore been measured in the sera and bile from rats infected with each rodent isolate. The serological findings indicate that both isolates are comparably immunogenic and that both induce secretory antibody responses which could be important in immunity to this essentially lumendwelling organism. Studies are in progress to examine whether qualitative differences exist between the antibody responses to the two organisms. These studies may reveal why the immune response is effective against the mouse isolate but not against the rat
Page 54
isolate in primary infections. However, the processes that frustrate the immune response to the rat isolate may be subtle, because animals primed against this organism and cured by treatment with metronidazole are subsequently highly resistant to reinfection. The latter phenomenon will provide an interesting area for future investigation. Finally, direct evidence has been sought for the role of secretory antibody in immunity against Giardia by passive transfer of bile, which can be obtained easily and is a major source of intestinal IgA in rats (7,8). Immune bile, containing IgA against the homologous isolate, was infused into the duodena of animals infected with the mouse isolate. Infusion of immune bile, but not of bile from uninfected rats, led to a marked decrease in cyst excretion. This sort of study is limited by the amounts of bile available and the time for which animals can be restrained. Nevertheless, the evidence is compelling that immune bile delivered in physiological amounts reduced the parasite load. This effect is attributed to the content of IgA antibody in the bile. Further studies are in progress to examine the effects of immune bile in preventing establishment of infections with the homologous isolates and in affecting the course of infection with the rat isolate. These studies suggest that infections in the rat with rodent G. duodenalis isolates may be useful as a model for investigating immunity in giardiasis. Although the importance of bile as a route of secretion of IgA antibodies into the intestine differs between rats and man, the convenience of bile in rats for measurement and collection of IgA antibodies is obvious. The mode of action of secretory antibodies in reducing the trophozoite population in the duodenum is unknown, but they are likely to interfere with attachment to the mucosa. In vitro studies suggest that serum antibodies can prevent attachment of trophozoites to artificial substrates by immobilization of flagellae (Mayrhofer, unpublished results). Acknowledgements This work was supported by grants from the National Health and Medical Research Council of Australia and the Channel 10 Children's Medical Research Foundation, Adelaide. The authors thank Dr. P. Ey for his advice on immunoassays and Mrs. Glenys King and Mrs. Rosie Thomas for their help in preparing the manuscript. The isoenzyme analysis was kindly performed by Dr. R. Andrews at the Evolutionary Biology Unit of the South Australian Museum. Literature Cited 1. Avrameas, S. 1969. Coupling of enzymes to proteins with glutaraldehyde. Use of the conjugates for the detection of antigens and antibodies. Immunochemistry 6:4352. 2. Belosevic, M., G.M. Faubert, E. Skamene, and J.D. MacLean. 1984. Susceptibility and resistance of inbred mice to Giardia muris. Infect. Immun. 44:282286. 3. Bertram, M.A., E.A. Meyer, J.D. Lile, and S.A. Morse. 1983. A comparison of isozymes of five axenic Giardia isolates. J. Parasitol. 69:793801. 4. Erlich, J.H., R.F. Anders, I.C. RobertsThomson, J.W. Schrader, and G.F. Mitchell. 1983. An examination of differences in serum antibody specificities and hypersensitivity reactions as contributing factors to chronic infection with the intestinal protozoan parasite, Giardia muris, in mice. Aust. J. Exp. Biol. Med. Sci. 61:599615. 5. Filice, F.P. 1952. Studies on the cytology and life history of Giardia from the laboratory rat. Univ. Calif. Publ. Zool. 57:53146. 6. Korman, S.H., S.M. LeBlancq, D.T. Spira, J. El On, R.M. Reifen, and R.J. Deckelbaum. 1986. Giardia lamblia: identification of different strains from man. Z. Parasitenkd. 72:173180. 7. LemaîtreCoelho, I., G.D.F. Jackson, and J.P. Vaerman. 1977. Rat bile as a convenient source of secretory IgA and free secretory component. Eur. J. Immunol. 7:588590. 8. LemaîtreCoelho, I., G.D.F. Jackson, and J.P. Vaerman. 1978. Relevance of biliary IgA antibodies in rat intestinal immunity. Scand. J. Immunol. 8:459463. 9. Nash, T.E., J. McCutchan, D. Keister, J.D. Dame, J.D. Conrad, and F.D. Gillin. 1985. Restrictionendonuclease analysis of DNA from 15 Giardia isolates obtained from humans and animals. J. Infect.Dis. 152:6473. 10. Olveda, R., J.S. Andrews, and E.L. Hewlett. 1982. Murine giardiasis: localization of trophozoites and small bowel histopathology during the course of infection. Am. J. Trop. Med. Hyg. 31:6066. 11. Richardson, B.J., P.R. Baverstock, and M. Adams. 1986. Allozyme electrophoresis: a handbook for systematic and population studies. Academic Press, Sydney. 12. RobertsThomson, I.C., and G.F. Mitchell. 1978. Giardiasis in mice. I. Prolonged infections in certain mouse strains and hypothymic (nude) mice. Gastroenterology 75:4246. 13. RobertsThomson, I.C., and R.E. Anders. 1981. Serum antibodies in adults with giardiasis. Gastroenterology 80:1262. 14. RobertsThomson, I.C., D.P. Stevens, A.A.F. Mahmoud, and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterology 71:5761. 15. Sauch, J.F. 1984. Purification of Giardia muris cysts by velocity sedimentation. Appl. Environ. Microbiol. 48:454455. 16. Schaefer III, F.W., E.W. Rice, and J.C. Hoff. 1984. Factors promoting in vitro excystation of Giardia muris cysts. Trans. Roy. Soc. Trop. Med. Hyg. 78:795 800. 17. Snider, D.P., J. Gordon, M.R. McDermott and B.J. Underdown. 1985. Chronic Giardia muris infection in antiIgMtreated mice. I. Analysis of immunoglobulin and parasitespecific antibody in normal and immunoglobulindeficient animals. J. Immunol. 134:41534163. 18. Stevens, D.P., D.M. Frank, and A.A.F. Mahmoud. 1978. Thymus dependency of host resistance to Giardia muris infection: studies in nude mice. J. Immunol. 120:680682. 19. Wheatley, W.B. 1951. A rapid staining procedure for intestinal amoebae and flagellates. Am. J. Clin. Pathol. 21:990991. 20. Woo, P.K. 1984. Evidence for animal reservoirs and transmission of Giardia infection between animal species. p. 341364. In: S.L. Erlandsen and E.A. Meyer (eds.), Giardia and Giardiasis. Plenum Press, N.Y.
Page 55
Biological Differences in Giardia lamblia T.E. Nash* and A. Aggarwal National Institutes of Health, NIAIB, LPD, Building 5, Rm 118 Bethesda, Maryland 20205, U.S.A.. Although isolates of Giardia differ biochemically it is not known if these or other differences result in altered biological behavior. Gerbils were infected with two unique human isolates, GS/E and WB. All WB inoculated gerbils became infected and were able to selfcure by day 28. In contrast GS/E infected gerbils tended to remain infected. After treatment, previously WB infected gerbils resisted infection after rechallenge with WB and GS/E, while GS/E infected gerbils partially resisted rechallenge with GS/E. All were infected with WB. Resistance to infection correlated with the development of cytotoxic antibodies which reacted with the surface of the Giardia. Experimental infections in humans confirmed that Giardia isolates differed biologically. In the first study, five volunteers were enterally inoculated with 50,000 trophozoites of isolates GS/M or Isr, followed serially and treated on day 15. In the second study, two of the previously inoculated, infected and treated volunteers were rechallenged along with five new volunteers as controls. All 10 of the GS/M inoculated volunteers became infected compared to none of the 5 inoculated with Isr. Both rechallenged individuals became infected although one only transiently shed cysts. Of the 10 infected, 5 or 50% became ill, 4 with diarrhea and typical symptoms of giardiasis. Humoral immune responses to Giardia occurred in all infected volunteers. Therefore, these experiments give credence to the idea that Giardia are not only biochemically different, but also differ in their biological behavior.
Introduction Morphologically, Giardia lamblia isolates appear to be similar (4); however, both casual observations and detailed analysis reveal differences among isolates. Some isolates, after axenization, adhere mostly to the surface of the tubes while others swim vigorously in the medium and tend not to adhere to surfaces (personal observation). Other isolates are long and slender while some are plumper and fuller in shape (personal observation). Some are easily axenized while others grow axenically with difficulty or not at all (personal observation). Biochemical differences have also been noted (6,8,9,10). When the DNAs of various human Giardia isolates were digested with endonuclease restriction enzymes and hybridized to Giardia specific probes, the number and position of the bands differed in a majority of the isolates (9). In addition, the surface antigens varied in most isolates as demonstrated by surface labeling (8) and the binding of monoclonal antibodies (McAb) to the surface of isolates possessing a 170 kd antigen (5). Other studies suggested most of the heterogeneity resided on the surface (12). Since many interactions among cells occur at the surface, it is not unreasonable to expect that varying surface antigens and other differences would result in alterations in behavior. Others have noted differences in cyst shedding in mice infected with human infective Giardia cysts (1), and we have noted marked differences in the ability of cysts obtained from infected humans to infect infant mice (personal observation). To more fully study whether isolates varied in their biological behavior, gerbils (3) and humans were infected with characterized isolates and the course of infection followed (2). The goals of the gerbil infections were to answer the following questions: (i) Do gerbils infected with different isolates become infected and undergo a similar course of infection? (ii) Does resistance to reinfection develop? (iii) If resistance develops, is it the same to homologous and heterologous isolates? (iv) What are the immune responses to infection and do any correlate with the course of infection or the development of resistance? Results Sixweek old Mongolian gerbils (Meriones unguiculatus) were inoculated with 2 million trophozoites by gavage and the number of Giardia in the intestines determined over time. The two compared isolates differed dramatically. Both originated from symptomatic infected humans. WB was isolated from a patient infected in Afghanistan (9,11) and GS/E from a scientist infected while fishing and camping in Alaska (9). The isolates differed in all parameters tested including endonuclease restriction patterns (9), surface antigens (8), and excretorysecretory products (8). Some gerbils were also infected with isolate Isr. This isolate closely resembles WB; in fact, WB and Isr appear indistinguishable biochemically (8,9). Both the course of infection and ability to induce resistance to reinfection differed (2). Although all the gerbils inoculated with either WB or GS/E became infected, * Corresponding author.
Page 56 TABLE 1. Percent Giardia trophozoites killed by cytotoxic antibodies induced by WB or GSE infection in gerbils. Infection WB
GSE
Day post inoculation
Isolate Used as Target
7
14
21
28
WB
12.5 ± 0.5*
27.5 ± 5.7
48.3 ± 4.3
48.5 ± 0.5
GSE
10.3 ± 5.5
25.0 ± 3.0
47.0 ± 5.0
48.0 ± 5.0
WB
3.6 ± 2.3
3.1 ± 3.1
10.5 ± 5.0
18.7 ± 2.0
GSE
20.4 ± 4.0
29.0 ± 5.0
32.3 ± 3.0
38.5 ± 15.0
Cytotoxicity was determined as previously described (2). Briefly, surviving Giardia were subcultured in TYIS33 after 12 h exposure to gerbil serum. The number of viable Giardia after 24 h was proportional to the original inoculum. Controls consisted of Giardia exposed to medium alone. * Each time point represents the mean ± S.D. percent cytotoxicity of three experiments. Pooled sera from 35 animals were used for each time point per experiment.
those infected with WB selfcured (Figure 1, Panels A and D) by day 35. In contrast, most GS/E infected gerbils were still infected even on day 42 although the number of trophozoites in the intestines had decreased. Gerbils infected with either isolate were treated with metronidazole on day 28 and challenged with either the homologous or heterologous isolate 7 days later. Gerbils previously infected with WB (Figure 1, Panels B and E), resisted infection with either isolate; however, gerbils previously infected with GS/E were partially resistant to challenge with the homologous isolate but all animals challenged with the WB isolate became infected although there were fewer organisms in the intestines (Figure 1, Panels C. and F). The pattern of infection with Isr was similar to that with WB. Complement independent cytotoxic antibodies developed in the sera of gerbils during infection and the degree of cytotoxicity was dependent on the infecting Giardia isolates, the test isolate employed, and the duration of the infection (Table 1). The development of cytotoxic antibodies correlated with the ability to resist infection. WBinfected gerbils developed appreciable cytotoxicity for both isolates and resisted infection to both isolates. On the other hand, GS/E infected gerbils developed higher levels of antibodies to GS/E than to WB. These animals were partially resistant to both isolates but more so to the homologous GS/E Giardia. These studies conclusively show that Giardia isolates not only differ biochemically but in their behaviour in vitro. In addition, the ability of each isolate to induce varying amounts of cytotoxic antibodies against these two isolates suggest surface antigens may be important as target antigens in protecting gerbils from reinfection. Humans were also infected with two different Giardia isolates (7). The goals of these experiments were necessarily different than those using gerbils mainly because much less is known about experimental Giardia infections in humans. The goals of these experiments were to answer some of the following questions: (i) Can axenized trophozoites infect humans? (ii) Are different isolates equally capable of infecting humans? (iii) Do these isolates cause disease?
Figure 1. The lower half of the graph shows the number of trophozoites in the small intestines of gerbils. The upper half shows the percent of animals infected on different days after inoculation with WB or GSE isolates. Each point represents the log of mean trophozoite counts from 1015 gerbils. WB,GSE Panels A & D, infection with the WB or GSE isolate. WB/WB Panels B & E, primary WB infection challenged with WB. WB/GSE Panels B & E, primary WB infection challenged with GSE. GSE/WB Panels C & F, primary GSE infection challenged with WB. GSE/GSE Panels C & F., primary GSE infection challenged with GSE.
(iv) What is the course of infection and disease? (v) Does resistance to infection develop, and is it equal in homologous and heterologous isolates? (vi) What are the immunological responses to infection in humans? Two isolates were used, GS/M and Isr. GS/M is the same isolate as GS/E, except the trophozoites used for axenization were obtained from neonatally infected mice and not after in vitro excystment of purified cysts. Isr was isolated from an infant (from Bethesda, MD) with diarrhea (8,9). As mentioned above, this isolate resembled WB. WB could not be used because the isolate originated from a person who was resistant to standard courses of chemotherapy.
Page 57
Extensive studies were done to exclude infection with Giardia or other agents and to exclude other underlying or complicating conditions. On day 0 volunteers were inoculated with 50,000 viable trophozoites enterally via a polyvinyl catheter positioned in the jejunum. Each day they were asked about the presence or absence of particular symptoms and examined if necessary. Stools were collected daily, and the consistency and presence of Giardia noted. Jejunal aspirates were obtained on days 0, 14, and 19. Stools were initially examined without concentration, and if negative, were reexamined following concentration. Volunteers were treated on day 15. Twelve weeks after inoculation and infection, 3 GS/M infected and treated patients were challenged with the same isolate as before. An additional 5 volunteers were inoculated at the same time, as controls. In the first experiment, all 5 of the GS/M inoculated volunteers became infected and none of the Isr inoculated volunteers. In the second experiment all 5 of the controls became infected. Therefore, 10 of 10 GS/M inoculated volunteers became infected while 0 of 5 Isr inoculated were infected (p < .004, twotailed Fishers exact test). Of the 10 GS/M infected volunteers, 4 developed diarrhea which was defined as stools which took the shape of the container (Grade 3). One person developed fever and headaches but without diarrhea, and although evaluated extensively, no other cause for the fever and headache was found. Altogether there were 17 episodes of diarrhea in the infected group and 1 in the 5 nonIsrinfected group. Diarrhea occurred significantly more often in the infected volunteers than in the noninfected volunteers (p < .035, Yates mean score). Although these symptoms were more frequent in the infected group, they were not significantly more frequent at the p < .05 level. The course of infection in two symptomatic volunteers in the first group is shown in Figure 2. Cysts were first noted in their stools on day 6. In patient 2, a single episode of diarrhea occurred on day 3 followed by numerous episodes beginning on day 7, 1 day after the onset of cyst shedding. Other symptoms including vomiting, cramping, anorexia, and flatulence also began at about this time. Patient 1 showed a less symptomatic course. Diarrhea began on day 8. Cramps and flatus were his main complaints. The mean prepatent period for all 10 infect volunteers was 7.5 ± .97 (mean ± SD)days and diarrhea began 7.25 ± 2.99 days after inoculation. Of the challenged volunteers, one was excreting cysts at the time of challenge, and the other two became reinfected. None were symptomatic. Giardiaspecific IgM antibodies developed in all infected patients and in none of the Isr inoculated patients. Specific IgG and IgA responses occurred in 70% and 60% of the infected volunteers, respectively. Intestinal IgA responses occurred in 60% of patients. Cytotoxic antibodies were not detected; failure to detect this response may be due to the short duration of infection prior to treatment. These human studies are important in a number of ways although not always comparable to the experimental gerbil infections. First, a safe and reliable method of infecting and studying humans was developed which will
Figure 2. Horizontal bars denote the presence of the symptom or finding noted on the left of the graph. The bottom most bar refers to stool grade: grade 1, formed; grade 2, soft; and grade 3, liquid. The number of grade 3 stools in a particular day is denoted below the horizontal bar.
allow further experiments. Second, for the first time Koch's postulates were fulfilled and Giardia formally shown as a cause of diarrhea. Third, isolates differed in their ability to initiate infections, confirming the importance of differences among isolates. Fourth, the immune responses were better defined, particularly the universal IgM response in presumably firsttime infections. Many of the questions answered and raised by experimental infections in gerbils were not answered in human infections. There were many obvious differences between the hosts and between experimental protocols; however, biological differences among isolates were clearly demonstrated in both series of experiments. Literature Cited 1. Aggarwal, A., A. Bhatia, S.R. Naik, and V.K. Vinayak. 1981. Variable virulence of isolates of Giardia lamblia in mice. Ann. Trop. Med. Parasitol. 77:163167. 2. Aggarwal, A., and T.W. Nash. 1987. Comparison of two antigenically distinct Giardia lamblia isolates in gerbils. Am. J. Trop. Med. Hyg. 36:325332. 3. Belosevic, M., G.M. Faubert, J.D. MacLean, C. Law, and N.A. Croll. 1983. Giardia lamblia infections in Mongolian gerbils: An animal model. J. Inf. Dis. 147:222226. 4. Bertram, M.A., Meyer E.A., Anderson D.L., and C.T. Jones. 1984. A morphometric comparison of five axenic Giardia isolates. J. Parasit. 70:530535. 5. Nash, T.E., and A. Aggarwal. 1986. Cytotoxicity of monoclonal antibodies to a subset of Giardia isolates. J. Immunol. 136:26282632. 6. Nash, T.E., F.D. Gillin, and P.D. Smith. 1983. Excretorysecretory products of Giardia lamblia. J. Immunol. 131:20042010. 7. Nash, T.E., D.A. Herrington, G.A. Losonsky, and M.M. Levine. Experimental human infections with Giardia lamblia. J. Inf. Dis. 156:974984. 8. Nash, T.E., and D.B. Keister. 1985. Differences in excretorysecretory products and surface antigens among 19 isolates of Giardia. J. Inf. Dis. 152:11661171.
Page 58
9. Nash, T.E., T. McCutchan, D. Keister, J.B. Dame, and F.D. Gillin. 1985. Endonuclease restriction analysis of DNA from 15 Giardia isolates obtained from man and animals. J. Inf. Dis. 152:6473. 10. Smith, P.D., F.D. Gillin, N.A. Kaushal, and T.E. Nash. 1982. Antigenic analysis of Giardia lamblia. J. Immunol. 6:714719. 11. Smith, P.D., F.D. Gillin, W.M. Spira, and T.E. Nash. 1982. Chronic giardiasis: studies on drug sensitivity, toxin production, and host immune response. Gastroenterology 3:797803. 12. Ungar, B.L.P., and T.E. Nash. 1987. Crossreactivity among different Giardia lamblia isolates using immunofluorescent antibody and enzyme immunoassay techniques. Am. J. Trop. Hyg. 37:283289.
Page 59
ANIMAL MODELS AND CROSSINFECTION
Page 61
Prevalence of Giardia sp. in Dogs from Alberta Paul D. Lewis, Jr. Department of Biological Sciences, University of Lethbridge, Lethbridge, Alberta, T1K 3M4, Canada. A total of 101 of 1,005 stray dogs from four cities in Alberta (10%) were infected with Giardia sp. as detected by single stool examinations. Prevalences reported are 33.3% of 81 dogs in Lethbridge, 11.9% of 310 dogs in Calgary, 6.9% of 246 dogs in Red Deer, and 5.4% of 368 dogs in Edmonton. In Calgary and Edmonton, where separate animal shelters are operated by the municipalities and by the Society for the Prevention of Cruelty to Animals (SPCA), prevalences of infection were higher in dogs from SPCA shelters than municipal shelters (8.7% vs 3.2% in Edmonton, 14.3% vs 10.7% in Calgary). At the peak of an outbreak of human giardiasis in Edmonton, during February 1983, dogs from the municipal animal shelter showed a prevalence 5 times higher than dogs from the same shelter the previous summer (17.2% vs 3.2%), suggesting that both dogs and humans were exposed to the same source of infection. Surveillance of prevalence of Giardia sp. in dogs may serve as a monitor for incipient waterborne outbreaks in the human population.
Introduction Epidemic outbreaks of giardiasis are well known in communities along the eastern slopes of the Rocky Mountains. Although reports first centred in the United States, particularly in Colorado (22,23), outbreaks have been reported more recently in Canada from British Columbia, Alberta and Saskatchewan as well (7,8,9,28). Giardia lamblia from humans can develop in a variety of nonhuman host species (14), and beavers have been implicated as reservoir hosts (15). Specific pathogen free beagles have been used successfully to test for infective cysts of Giardia sp. from public water supplies (25), and mongrel dogs have been infected with both human source cysts and cultured trophozoites (16). As well, the Mongolian gerbil now seems established as a reliable laboratory model for human giardiasis (4,19,20). Thus, the prevalence of Giardia in dogs and/or other potential reservoir hosts may constitute useful epidemiological information with respect to human outbreaks. This paper reports the results of a coproanalytic survey to determine the prevalence of Giardia sp. in stray dogs in Lethbridge, Calgary, Red Deer and Edmonton (Alberta) during the summer of 1982. Additionally, the occurrence of an epidemic outbreak of human giardiasis in Edmonton, peaking in February, 1983 (9), provided an unusual opportunity to sample stray dogs at that time. Methods and Materials Cities Surveyed Stool samples were obtained between midMay and midAugust, 1982, from stray and unwanted dogs in (2) municipally operated animal shelters in the four largest cities in Alberta: Edmonton (53° 33' N, 113° 28' W, population 560,000), Red Deer (52° 16' N, 113° 48' W, population 50,000), Calgary (51° 03' N, 114° 05' W, population 620,000), and Lethbridge (49° 42' N, 112° 49' W, population 58,000); and (2) shelters operated by the Society for the Prevention of Cruelty to Animals (SPCA) in Calgary and Edmonton (Table 1). Collection Methodology At each participating shelter except for Lethbridge, where samples were collected weekly by the author, shelter personnel were trained to collect single stool samples from incarcerated dogs. Each sample was enclosed in a numbered bag, and its bag number, the shelter's animal ID number, and the location of capture recorded on a data sheet. Except for Lethbridge, collected samples and their data sheets were transported to the laboratory weekly by courier for analysis (usually 24 to 36 h transit time); samples were refrigerated (4°C) upon arrival until they were processed (up to 96 h). Coproanalyses In the laboratory, samples were examined once by each of (1) MIF (merthiolateiodineformalin) centrifugation, and (2) ZnSO4 flotation procedures (27). Reference material for each stool was retained as permanent smears stained with Wheatley's Stain (29) and as aliquots preserved in vials of SAF (sodium acetate, acetic acid, formalin) preservative (30). Analysis of Techniques Employed (1) Technique Sensitivity: A comparison of the MIF centrifugation and ZnSO4 procedures was made by determining the percent of all positive records obtained by each method. (2) Examination of Fresh vs. SAFPreserved Stools by ZnSO4 Flotation: Six fresh stools from Lethbridge dogs found to be heavily infected with Giardia sp. were examined as follows: for each of the 6 stools, two equal portions weighing 0.5 g ± 0.02 g were taken; the first was examined immediately by ZnSO4 flotation, and the second was examined by the same procedure after preservation for 18 h in SAF. For each, 6 scans at 400X were made of 22 mm coverglass preparations, and the numbers and appearances of trophozoites and cysts were recorded. Analysis of Collections During the Edmonton Outbreak An outbreak of human giardiasis occurred in Edmonton between December, 1982, and April, 1983; it reached a peak in February (9). Stool samples from a total of 100 dogs from the Edmonton municipal animal shelter were collected in vials of SAF by shelter personnel who had been trained to ensure that proper collection methodology was used. Upon transport to the laboratory, samples were examined using the ZnSO4 flotation technique. TABLE 1. Summary of sources of canine stool samples
Locality
Number of Samples Collected Municipal Shelters
S.P.C.A. Shelters
Edmonton
218
150
Red Deer
246
Calgary
205
105
Lethbridge
81
All Centres
750
255
Page 62 TABLE 2. Prevalence of Giardia sp. in Alberta dogs*
CITY
MAYJUNE
JULYAUGUST
TOTAL
Positive/Examined
% Prevalence
Positive/Examined
% Prevalence
Positive/Examined
% Prevalence
Edmonton
12/196
6.1
8/172
4.7
20/368
5.4
Red Deer
8/117
6.8
9/129
7.0
17/246
6.9
Calgary
10/114
8.8
27/196
13.8
37/310
11.9
Lethbridge
17/51
33.3
10/30
33.3
27/81
33.3
All Cities
47/478
9.8
54/527
10.2
101/1005
10.0
* Based upon dual analyses (MIF centrifugation plus ZnSO4 flotation) of single stool samples from stray dogs incarcerated in animal shelters.
TABLE 3. Prevalence of Giardia sp. in dogs from municipal vs. S.P.C.A. animal shelters in Calgary and Edmonton, 1982*
CITY
MUNICIPAL SHELTERS
SPCA SHELTERS
TOTAL
Positive/Examined
% Prevalence
Positive/Examined
% Prevalence
Positive/Examined
Edmonton
7/218
3.2
13/150
8.7
20/368
% Prevalence 5.4
Calgary
22/205
10.7
15/105
14.3
37/310
11.9
Both Cities
29/423
6.9
28/255
11.0
57/678
8.4
* Single samples were collected from incarcerated stray dogs from midMay through midAugust and were examined by both MIF centrifugation and ZnSO4 flotation techniques.
Results Coproanalyses In all, 10% of the stools of stray dogs from Alberta cities were positive for Giardia sp. during the summer of 1982 (Table 2). Morphologically, the parasites appeared to be congruent with the duodenalis type as evidenced by cyst appearance and dimensions, and by the presence of a claw hammer median body in those trophozoites which were observed. Prevalences varied only slightly between the MayJune and JulyAugust periods from Red Deer and Lethbridge. However, for Calgary the JulyAugust prevalence increased by 57% compared to MayJune, and for Edmonton the JulyAugust prevalence decreased by 23% compared to MayJune. For the entire period, as well as for JulyAugust, prevalences tended to decrease from north (Edmonton) to south (Lethbridge), but this was less apparent for MayJune. When the results for Edmonton, Calgary, and both cities combined are compared by shelter source (municipal vs SPCA), prevalences of infection with Giardia sp. in dogs were higher at shelters operated by the SPCA than at shelters operated by the municipality in each city and for both cities combined (Table 3). Prevalence of infection for dogs at the SPCA shelters exceeded that for dogs from the corresponding municipal shelters by nearly one half in Calgary, by nearly 3 fold in Edmonton, and by just over one half for both cities combined. Analysis of Techniques The sensitivities of the two coproanalytic techniques employed in this study were compared. Of the 101 positive records obtained, 67 (66%) were detected using the MIF centrifugation technique, and 94 (93%) using the ZnSO4 flotation technique. A study was made of the effectiveness of the ZnSO4 flotation procedure in detecting Giardia sp. in paired (same stool) fresh and SAF preserved samples from dogs known to be heavily infected. The flotation procedure revealed 1 to 89 (average 34) cysts per horizontal scan for fresh stools, whereas it revealed 4 to 47 (average 21) cysts per horizontal scan for the paired SAF preserved samples. In the ZnSO4 coverslip mounts of some SAF preserved samples, up to 1/3 of the parasites observed were trophozoites, whereas trophozoites rarely were detected using this technique with the paired fresh stools. Moreover, the characteristic morphology of Giardia sp. cysts in stools preserved in SAF remained observable for up to several hours in ZnSO4 flotation coverslip preparations, whereas cysts from fresh stools quickly became unrecognizable in such preparations because of collapse of the organism against the inner surface of the cyst wall. Prevalence of Giardia sp. in Dogs During a Human Outbreak SAF preserved stool samples from 100 dogs in the municipal animal shelter, Edmonton, were obtained during the peak of the outbreak of human giardiasis in February, 1983. These were examined by the ZnSO4 flotation method. One sample was found to be a duplicate, and therefore was not counted. The prevalence of Giardia sp. in Edmonton dogs in February, 1983, was 17.2% (Table 4), a value 5 times greater than for dogs from the same (municipal) shelter, and 3 times greater than for all Edmonton dogs, the previous summer.
Page 63 TABLE 4. Prevalence of Giardia sp. in feces of Edmonton dogs during and prior to an outbreak of human giardiasis* Stool Sources:
Municipal Shelter
Municipal Shelter
SPCA Shelter
Municipal + SPCA Shelters, 1982
Date of Samples:
February 1983§
Summer 1982#
Summer 1982#
Summer 1982#
Pos/Exam.
17/99
7/218
13/159
20/368
% Prevalence:
17.2
3.2
8.7
5.4
* The peak of the human outbreak occurred in February, 1983; the preoutbreak data are from 1982. § Prevalence of infection in dogs in February, 1983, was determined using single ZnSO4 flotation examinations of single stool samples preserved in SAF. # Preoutbreak prevalences were determined using dual MIF centrifugation plus ZnSO4 flotation techniques on fresh single stool samples from stray dogs held in the municipal and SPCA shelters in Edmonton.
Discussion The prevalences reported here for Giardia sp. in Alberta dogs fall within the range of prevalences reported elsewhere in the world (Table 5). However, comparisons between stool analysis surveys are difficult to make because the reported results can vary for a variety of reasons which are quite additional to differences in endemicity which may occur through time or from area to area. The increase in prevalences from north to south for the entire sampling period appears to be an artifact created by the higher prevalence of Giardia sp. in Calgary dogs during JulyAugust, and by the consistently high prevalence in Lethbridge dogs throughout the summer. The JulyAugust prevalence for Calgary dogs is nearly 50% higher than for MayJune (Table 2), and was observed in animals from both the municipal and the SPCA shelters. The reason for the high prevalence in Lethbridge dogs is not known. The Lethbridge shelter appeared to be somewhat less sanitary than shelters in the other cities; but the fact that stool samples were obtained from Lethbridge dogs within 1 to 6 days following capture, a time span less than the usual prepatent period for Giardia sp. infections in dogs (16), suggests that the Lethbridge dogs already were infected at capture, and that the prevalence in Lethbridge dogs indeed is higher than for dogs elsewhere in the Province. The higher prevalences of Giardia sp. observed in dogs from SPCA shelters over municipal shelters in Calgary and Edmonton (Table 3) is puzzling. Both municipal and SPCA shelters were clean and well run. It is possible that the SPCA shelters tended to hold dogs for longer periods of time, thus increasing the opportunity for dogtodog transmission. At least in this laboratory, the ZnSO4 flotation technique appears to be more efficient than the MIF centrifugation technique in detecting infections with Giardia sp. TABLE 5. Prevalences of Giardia sp. from dogs. Reference
Pos./Exam.
% Prevalence
Locality
Hewlett et al. 1982
25/37
68
Ohio
Burrows & Lillis 1967
99/273
36
New Jersey
Cotteleer & Fameree 1980
25/94
27
Belgium
Swan & Thompson 1986
70/333
21
Australia
Yang & Scholten 1977
96/495
19
Ontario
Pfeiffer & Supperer 1976
13/70
19
Austria
Catcott 1946
20/113
18
Ohio
de Carnieri et al. 1964
3/31
10
Italy
14/160
9
California
158/2063
8
Minnesota
Levine & Ivens 1965
7/175
4
Illinois
Agresti et al. 1977
11/300
4
Italy
Jungmann et al. 1986
5/141
4
GDR
Hoskins et al. 1982
34/4752
1
Louisiana
PRESENT STUDY
101/1005
10
Alberta
Craige 1948 Bemrick 1961
in dogs. The value of SAF as a preservative has been noted previously (30), but its value in enabling both cysts and trophozoites of Giardia spp. to resist the distorting effects of ZnSO4 is a new observation. It may be that the apparent absence of trophozoites in unpreserved stools is explained by their fragility, which is reduced by the preservative. An intermittent pattern of cyst release is observed in natural infections with Giardia sp. in dogs (3,16), and in experimental infections with Giardia lamblia in dogs (16), and gerbils (4,20). This suggests that surveys such as the present one, which are based upon single stool collections, will fail to identify all active infections. Multiple stool examinations will reveal additional infections; the use of a series of 6 consecutive stool examinations resulted in a prevalence rate of 68% for Giardia sp. in a population of 37 stray dogs in Ohio (16). Furthermore, the cortisone induced recrudescence of infection with Giardia lamblia in apparently self cured gerbils as much as 7 months after their initial infection suggests that the actual numbers of infected individuals in a population may be higher than that determined by stool examinations for cysts (20). Thus, prevalences reported in the present study undoubtedly are understated. There is no published study of the human outbreak of giardiasis in Edmonton, but summaries indicate that the outbreak extended from late 1982 to April, 1983 with a peak in February, 1983 (9,28). A total of 895 human cases were reported (28), but the actual number of infections undoubtedly was much higher (9). The cause of the outbreak could not be determined by the investigators (9). However, a number of communitywide outbreaks in North America have been well studied, and, where it has been demonstrated, the mode of infection invariably has been waterborne (15,23,25). No previous studies have compared the prevalence of infection in dogs or other potential reservoir hosts before and during a communitywide outbreak of human giardiasis. At the peak of the human outbreak in Edmonton (February, 1983) the average number of cases reported per
Page 64
week (about 60) was 4 times higher than the average number of cases reported per week (about 15) for the preoutbreak month of October, 1982 (9). This compares well with the 5fold increase in the prevalence of Giardia sp. in dogs from the Edmonton municipal shelter between the summer of 1982 and February, 1983. In view of the failure of investigators to implicate pets as a cause of human illness (9), it thus appears that both humans and dogs in Edmonton were exposed to the same source, probably the municipal water supply. Acknowledgements This research was funded by a grant from the Alberta Environmental Research Trust. I gratefully acknowledge the cooperation and assistance of the participating animal shelter personnel: Mr. Lorne Creor (Lethbridge), Ms. Trudy DeBecker (Calgary, municipal), Ms. R. Falco (Calgary, SPCA), Ms. K. McLaren (Red Deer), Mr. R. Wilson (Edmonton, municipal), and Mr. Bill Hart and Ms. Lori Stingley (Edmonton, SPCA). Indefatigable assistance in the laboratory was provided by Ms. Lesley Curthoys, Ms. Paula Harper and Mr. Joe Harrison. Literature Cited 1. Agresti, A., G. d'Ambrosio, and E. Gravino. 1977. La giardiasi nei cane: indagine epizoologica ed osservazioni cliniche. Acta Medica Veterinaria. Napoli 23:175. 2. Allison, D.J. 1984. Giardiasis a recent investigation. Can. J. Public Health 75:318320. 3. Barlough, J.E. 1979. Canine giardiasis: a review. J. Small Animal Practice 20:613623. 4. Belosevic, M., G.M. Faubert, J.D. MacLean, C. Law, and N.A. Croll. 1983. Giardia lamblia infections in Mongolian gerbils: An animal model. J. Infect. Dis. 147:222226. 5. Bemrick, W.J. 1961. A note on the incidence of three species of Giardia in Minnesota. J. Parasit. 47:8789. 6. Burrows, R.B., and W.G. Lillis. 1967. Intestinal protozoan infections in dogs. J. Am. Vet. Med. Assoc. 150:880883. 7. Canada Diseases Weekly Report. 1982a. Waterborne giardiasis outbreak Alberta. 8:9798. 8. Canada Diseases Weekly Report. 1982b. Giardia lamblia in a community water supply British Columbia. 8:98100. 9. Canada Diseases Weekly Report. 1983. Giardiasis Edmonton, Alberta. 9:189192. 10. de Carnieri, I., and S. Castellino. 1964. Entamoeba canibuccalis, Trichomonas canistomae, Giardia canis nei cani a Milano. La Clinica Veterinaria, Milano 87:193196. 11. Catcott, E.F. 1946. The incidence of intestinal protozoa in the dog. J. Am. Vet. Med. Assoc. 108:3436. 12. Cotteleer, C., and L. Famere'e. 1980. Helminthes et protozoaires intestinaux parasites du chien en Belgique. Cas particulier des Eucoccidia. Schweizer Archiv fur Tierheilkunde 122:519526. 13. Craige, J.E. 1948. Differential diagnosis and specific therapy of dysenteries in dogs. J. Am. Vet. Med. Assoc. 113:343347. 14. Davies, R.B., and C.P. Hibler. 1979. Animal reservoirs and crossspecies transmission of Giardia, pp. 104126. In: W. Jakubowski, and J.C. Hoff (eds.). Waterborne Transmission of Giardiasis. EPA Publ. No. 600/979001. 15. Dykes, A.C., D.D. Juranek, R.A. Lorenz, S. Sinclair, W. Jakubowski, and R. Davies. 1980. Municipal waterborne giardiasis: an epidemiologic investigation. Ann. Int. Med. 92:165170. 16. Hewlett, E.L., J.S. Andrews Jr., J. Ruffier, and F.W. Schaefer III. 1982. Experimental infection of mongrel dogs with Giardia lamblia cysts and cultured trophozoites. J. Inf. Dis. 145:8993. 17. Hoskins, J.D., J.B. Malone, P.H. Smith, and S.A. Uhl. 1982. Prevalence of parasitism diagnosed by fecal examination in Louisiana dogs. Am. J. Vet. Res. 43:11061109. 18. Jungmann, R., T. Hiepe, and C. Scheffler. 1986. Zur parasitaren Intestinalfauna bei Hund und Katze mit einem speziellen Beitrag zur GiardiaInfektion. Monatshefte fur Veterinarmedizin 41:309311. 19. Kirkpatrick, C.E., and J.P. Farrell. 1984. Feline giardiasis: observations on natural and experimental infections. Am. J. Vet. Res. 45:21822188. 20. Lewis, P.D. Jr., M. Belosevic, G.M. Faubert, L.C. Curthoys, and J.D. MacLean. 1987. Cortisone induced recrudescence of Giardia lamblia infections in gerbils. Am. J. Trop. Med. and Hyg. 36:3542. 21. Levine, N.D., and V. Ivens. 1965. Prevalence of nematodes, Giardia and Demodex in Illinois dogs. Illinois Vet. 8:19,2123. 22. Meyer, W.T. 1973. Epidemic giardiasis. A continued elusive entity. Rocky Mtn. Med. J. 70:4849. 23. Moore, G.T., W.M. Cross, D. McGuire, C.S. Mollohan, N.N. Gleason, G.R. Healy, and L.H. Newton. 1969. Epidemic giardiasis at a ski resort. New England J. Med. 281:402407. 24. Pfeiffer, H., and R. Supperer. 1976. Uber den Giardiabefall der Hunde und sein Auftretung in Osterreich. Wiener Tierarztliche Monatsschrift 63:16. 25. Shaw, P.K., R.E. Brodsky, D.O. Lyman, B.T. Wood, C.P. Hibler, G.R. Healy, K.I.E. MacLeod, W. Stahl, and M.G. Schultz. 1977. A communitywide outbreak of giardiasis with evidence of transmission by a municipal water supply. Ann. Int. Med. 87:426432. 26. Swan, J.M., and R.C.A. Thompson. 1980. The prevalence of Giardia in dogs and cats in Perth, Western Australia. Aust. Vet. J. 63:110112. 27. United States Naval Medical School. 1965. Medical Protozoology and Helminthology. U.S. Government Printing Office, Washington. 238 pp. 28. Weekly Epidemiological Record. 1984. Giardiasis surveillance. No. 15:114115. 29. Wheatley, W.B. 1951. A rapid staining procedure for intestinal amoebae and flagellates. Am. J. Clin. Path. 21:990991. 30. Yang, J., and T. Scholten. 1977. A fixative for intestinal parasites permitting the use of concentration and permanent staining procedures. Am. J. Clin. Path. 67:300304.
Page 65
Location of Giardia Trophozoites in the Small Intestine of Naturally Infected Dogs in San Diego Herndon Douglas, David S. Reiner, Michael J. Gault and Frances D. Gillin* Department of Pathology H811F, University of California, San Diego Medical Center, 225 Dickinson Street, San Diego, CA 92103, U.S.A.. Since little is known of the anatomical locus of trophozoites in naturally infected humans or canines, we ascertained the presence and distribution of Giardia in the small intestine (SI) of 29 apparently healthy adult dogs which were euthanized for other studies. Of 13 dogs sampled by aspirating duodenaljejunal fluid, only 3 (23%) were positive. In contrast, Giardia was found in 12 of 16 dogs (75%) when the entire SI was examined. The trophozoite distribution was determined by ligating the SI at 12 inch intervals and enumerating Giardia in both saline washes and mucosal scrapings. Three patterns were observed: (1) Broad distribution, with trophozoites in > 90% of the intestinal segments; (2) Intermediate, with trophozoites in ~ 50% of the segments; (3) Restricted, with trophozoites in only ~ 20% of the SI. In the latter class, Giardia was found in the upper first and second segments or the lowest two segments. Thus, in naturally infected dogs, Giardia can colonize anywhere in the SI. The total trophozoite burden was also extremely variable: 3 × 105 to 6.9 × 108 per dog. Few cysts were observed in the cecal contents. This resembles a form of human giardiasis in which cysts are not readily detectable in feces, but trophozoites may be found by aspiration.
Introduction Despite the prevalence of giardiasis in the United States and less developed countries, (5,19) little is known about the mechanism or frequency of colonization of the human small intestine by Giardia lamblia. Duodenal and upper jejunal fluid aspiration, ''string tests", and biopsies are frequently used in diagnosis (reviewed in 19) when fecal cysts are not detected. These methods are based on the assumption that trophozoites colonize the upper small intestine, but there is no detailed information on the distribution of trophozoites in the human small intestine. The present study was undertaken to elucidate the location of Giardia trophozoites in the intestinal tracts of naturally infected adult dogs, as a model of the distribution in people. Canine giardiasis may be a useful model of human infection for several reasons: (i) Experimental infection of dogs with Giardia from humans has been reported (6,12). (ii) Although the host specificities and species definitions of Giardia from humans and dogs are not clear (11,13, see also 14), both are of the duodenalis, or intestinalis morphologic type, with "claw hammer" median bodies (13). (iii) Both symptomatic and asymptomatic giardiasis are well documented in dogs (15) and humans (19). (iv) Both the numbers and frequency of cysts excreted in feces can be extremely variable in dogs, as in humans. Often trophozoites can be recovered from the small intestine when cysts are not detected in the stool (3,15). (v) It has been proposed (12) that dogs may be a reservoir for dissemination of Giardia cysts to humans since fecal contamination is common both near the home and in wilderness areas (See also 4,20,21 for critical discussions of crossspecies transmission). Moreover, recent studies have demonstrated high incidences (38% and 68%) of giardiasis in apparently healthy dogs from pet shops (17) and an animal shelter (12). In initial studies, we attempted to identify infected dogs by examination of fresh stools. Very few cysts were observed either in smears or purified preparations (see methods). Therefore, we ascertained the presence and distribution of Giardia trophozoites in the small intestines of 29 unselected, consecutive, apparently healthy adult dogs sacrificed for other studies which involved neither the intestine nor drugs that might affect the parasites. Both the number and pattern of colonization of trophozoites were extremely variable. Giardia were observed in fluid withdrawn by needle aspiration from 23% of anesthetized dogs. In contrast, Giardia were found in 75% of animals in which the entire small intestine was examined. Therefore, the diagnosis may be missed by sampling only the upper small intestine. Materials and Methods All animals used were acquired from the San Diego Pound for other studies and kept in the vivarium for three to five days before use. During this isolation period, no medication or special precautions were used to alter the intestinal flora. Dogs were fed Iams Chunks dog food and watered according to University policy. Food was withheld the night before experiments. We obtained all dogs immediately after the other experiments were completed. * Corresponding author.
Page 66
Thirteen dogs were acquired from a medical student teaching lab (UCSD) during a study of cardiac drug reactions. Giardia were sought by inserting an 18 gauge needle into the duodenum of an anesthetized dog, withdrawing the small volume of fluid, and examining it microscopically in a hemacytometer chamber for Giardia. Trophozoites were observed in fluid from only three dogs. The entire intestines of two dogs were processed by method 1 below (dogs 1 and 2, Table 1). An additional sixteen dogs were either donors for liver transplants by the Surgery Department or used by the Pulmonary Division for cardiac output and blood gas studies. Immediately after sacrifice, the animals were sampled by two methods. Method one consisted of ligating the small intestine with "O" silk (Ethicon) into approximately twelve inch sections, starting immediately below the pylorus and ending just below the cecum. After ligation was completed, the gut was removed en bloc. Each section was filled with 30 mL of sterile pyrogen free saline (Travenol), rinsed thoroughly, and the fluid emptied into a 50 mL tube. The sections were then opened lengthwise and scraped gently with wooden tongue depressors to recover the mucus. Ten mL of cold normal saline was added to all mucus preparations. Giardia in the fluid and mucus preparations were counted in hemacytometer chambers (five counts were averaged) and also analyzed by indirect immunofluorescence (IIF). This analysis was used for 2 dogs identified as positive by aspiration and 8 dogs which were not prescreened (dogs 110, Table 1). Method two consisted of separating the entire small intestine into four (dogs 1117) or five (dog 18) equal segments which were then processed and quantified as above, again isolating the cecum as a unit before removal. In some cases, Giardia cysts and trophozoites were purified from samples of stool or cecal contents by suspending up to 20 g of feces in cold saline solution, filtering through cheese cloth, passing over a Sephadex G50 column (7) and washing 3 times at low speed. Cysts and trophozoites were quantified by hemacytometer counts as above or IIF as below. Indirect Immunofluorescence IIF was used to detect and count Giardia in the viscous mucus layer since they were not readily visible otherwise. The lumenal phase was analyzed in the same way, so that the numbers of parasites in each compartment could be compared directly. Samples (20 µL) of intestinal fluid or mucus were spotted on 8well slides, airdried and fixed for IIF with 1% Formalin in distilled water for 10 minutes at room temperature. After 3 distilled water washes, the cells were treated with acetone for 10 minutes at room temperature and dried. Rabbit antiserum to purified human cysts (10) or sonicated cultured strain WB trophozoites, diluted in PBS containing 1% bovine serum albumin (BSA) and 1% (v/v) Tween 20, was added to each well (20 µL) and the slides were incubated at 37 °C in a humidified chamber (20 minutes). After extensive washing with distilled water and air drying, 20 µL of FITClabelled goat antirabbit IgG (Sigma) were added to each well. This conjugate was diluted 1:800 in PBS BSATween 20 with ethidium bromide (2 µg/mL) as a counterstain. After further extensive washing with distilled water and air drying, the slides were mounted in pH 9 glycerol. The number of parasites/field was the mean of five fields. Results and Discussion Incidence of Giardiasis in Dogs In previous studies (see reference 1 for review of literature up to 1979), the incidence of giardiasis in dogs ranged from 0.6% to 67% (12). These differences may reflect actual variation in frequency of infection but our data suggest that they may be due, at least in part, to the methods of determining the presence of infection. We found very few cysts in feces of healthy adult dogs, but did not rely on this as a screening method. The incidences of infection we observed were greatly affected by the sampling method. We first assayed for the presence of giardiasis in 13 dogs by examining upper intestinal fluid aspirates and found only three positive. The distribution of trophozoites along the small intestine was determined in two of these dogs and found to be quite broad, with trophozoites present both in the lower and upper small intestine. Since we observed trophozoites in the lower intestine, we proposed that infections might not always be manifest in the upper small intestine and would not be detected by aspiration at that site. Therefore, we examined the entire small intestine in the next 16 unselected dogs and found 12 positive (p < 0.01, Fisher exact probability test). We did not observe loose stool in the cecum of any dog. Two methods of enumeration were used, as described in Methods. Direct hemacytometer counts of trophozoites in the fluid phase wash yield an estimate of the number of parasites per segment, since the fluid volume is known. However, the fluid phase is actively being moved "downstream" and may not represent the long term pattern of colonization. Any postmortem changes in distribution of trophozoites would also affect the fluid phase. In contrast, Giardia trophozoites associated with the mucus blanket probably reflect stable associations. Since it is difficult to count the trophozoites in the viscous mucus preparation, we used antibodies against trophozoites to visualize them by IIF. This method does not yield absolute numbers of parasites and cannot be compared directly with the hemacytometer counts. Accordingly, we also used IIF analysis of the trophozoites in the fluid phase for comparison (mean cells/field) between fluid and mucus phases and from animal to animal. The Numbers and Distribution of Giardia Trophozoites in the Small Intestine The small intestines of ten dogs were cut into segments of approximately 12 inches and the numbers of parasites determined as described in Methods. Since the number of segments varied from dog to dog, the parasite densities were plotted in terms of the percent of distance along the small intestine with the pylorus as 0% and the cecum as 100% for comparisons of the distributions from animal to animal. Both the number of parasites per dog and their distribution were extremely variable. The number of trophozoites in the fluid phase ranged from 0.3 × 106 to 692.0 × 106 per dog. The proportion of the intestine colonized varied from 24% to 100% (Table 1, last column). Interestingly, of the two dogs with trophozoites in each segment (Table 1), dog #1 had the greatest total number of trophozoites (> 6.9 × 108), while dog #17 had relatively few (< 2 × 106). These studies indicate that the parasite burden does not correlate with the distribution of Giardia trophozoites. In addition, the location of the section with the greatest number of parasites varied greatly. The "peak" section was in the upper third of the small intestine in three dogs (dogs 1,2,3) while it was in the lowest segment in four dogs (dogs 7,15,17,18). Illustrative distributions of Giardia trophozoites are shown in Figures 1 to 4. Dog #1 had the highest parasite burden in this study and trophozoites were recovered from each segment of the small intestine (Figure 1). The distribution was somewhat bimodal with peaks in the upper and
Page 67 TABLE 1. Total and peak parasite burdens in intestinal segments a
Dog #
Total Trophozoites (× 106)
Number and Percentage of Trophozoites in Peak Segment × 106 (%)
Terminus of Peak Segment (% distance # Positive Segments / from pylorus to Total d secum)
1
692.0
273.8
(40) c
33
9/9
2
38.9
9.4
(24)
30
8/10
3
7.6
5.6
(74)
8
4/12
4
33.8
10.4
(31)
50
10/14
5
15.7
4.6
6
negative b
7
9.1
5.5
8
2.6
9
(29)
62
12/13
negative
0/9
(60)
100
3/8
0.9
(35)
56
4/9
19.8
10.6
(53)
45
7/11
10
19.2
8.4
78
5/9
11
negative
negative
negative
0/4
12
negative
negative
negative
0/4
13
negative
negative
negative
0/4
14
0.3
0.3
(100)
75
1/4
15
36.9
27.0
(73)
100
3/4
16
1.3
0.9
(69)
100
2/4
17
1.5
0.6
(40)
100
4/4
18
137.6
78.6
(57)
60
3/5
negative
(44)
#a All data are based on the mean of 5 hemacytometer counts of parasites in the fluid phase of segments of the small intestine (pylorus to cecum). #b A negative segment has < 3.3 × 104 trophozoites (the limit of detection of this assay). A negative dog has no positive segment. #c Numbers in parentheses are the numbers of trophozoites in the peak segment divided by the total number of trophozoites (× 100). d A positive segment has 105 trophozoites.
lower jejunum. Dogs #3 and #7 illustrate the other extreme, a very narrow distribution of a Giardia trophozoites. In dog #3 (Figure 2), the peak parasite burden was in the uppermost segment (duodenum). In contrast, in dog #7, the peak segment was the lowest (ileumcecum) and no parasites were observed in the upper 60% of the small intestine (Figure 3). The pattern of trophozoite distribution was intermediate in dog #9 (Figure 4). The distribution patterns determined by the three analytic methods were roughly parallel except that no trophozoites were detected by IIF in the lumenal contents of dog
Figure 1. Broad distribution of Giardia trophozoites: dog #1.
#7 (Figure 3). Giardia were observed, however, in the mucus layer by IIF, as well as in the lumenal phase by hemacytometer counts. This discrepancy was also observed in the lowest segment of dog #1. Hemacytometer counts of this segment were positive. A possible reason for the negative IIF, when trophozoites are detectable by hemacytometer, is that bacteria or their products in the lower small intestine may digest antigens detected by IIF. In general, the distribution of Giardia in the mucus phase was similar to that in the lumen, suggesting that trophozoites in the more mobile fluid phase usually reflect
Figure 2. Restricted distribution of Giardia trophozoites, mainly anterior: dog #3.
Page 68
Figure 3. Restricted distribution of Giardia trophozoites, posterior: dog #7.
Figure 4. Intermediate distribution of trophozoites: dog #9.
colonization of the mucus layer. We have reported that intestinal mucus promotes attachment (21) and growth (9) of G. lamblia trophozoites in vitro. Many factors may influence both the incidence of giardiasis and the intestinal distribution of trophozoites. In a survey of > 2,000 dogs, 7.66% of zinc sulfate concentrates of stool samples had Giardia cysts (2). It was striking that 70% of the positive dogs were younger than 6 months and ~ 90% less than one year (2). Similarly, in endemic areas, giardiasis is more prevalent in human children than in adults. Differing loci of Giardia colonization in dogs have been reported previously. Faust (8) reported that in an unspecified number of naturally infected dogs, "the primary seat of the organism is the cecum and appendix" and occasionally in the terminal ileum. Similar trophozoite distributions were found in 11 dogs infected rectally with trophozoites. Only three dogs had trophozoites in the anterior ileum or above (8). Tsuchiya (18) infected four puppies orally with Giardia cysts from dogs. Most trophozoites were found 10 to 20 inches below the pylorus in the two puppies fed a high carbohydrate diet. Trophozoite concentrations were very much lower in the two puppies fed a high protein diet and the peak incidence was 30 to 35 inches (intestine length ~55 inches to cecum). Cysts were observed in the lower half of the lower intestine as well as in the cecum and colon (18). These observations are congruent with our studies of a larger group of unselected dogs of unknown age (apparently fullgrown), diet (at least prior to arrival at the pound), and duration of infection, in which trophozoites colonized anywhere in the small intestine. Our detailed studies of giardiasis in apparently healthy adult pound dogs in San Diego reveal: (i) a high incidence of infection, (ii) few cysts, (iii) highly variable patterns of trophozoite colonization. Infection in these dogs resembles giardiasis in two groups of humans in whom cysts are not found in the feces. In the first group, trophozoites are found by intestinal aspiration, biopsy, or string test (19). In the second group, intestinal aspiration is negative, but symptoms disappear in response to antigiardial therapy (19). Therapeutic response to broadspectrum antiprotozoan compounds does not prove the diagnosis of giardiasis however, and the true etiology of infection in these patients usually remains problematic. Our observation that Giardia trophozoites were restricted to the lower small intestine in certain dogs suggests that some of these people may be infected with Giardia trophozoites only in the lower intestine. Acknowledgements This work was supported by NIH AI 19863 and EPA #811950010. We are grateful to C. Davis, L. Corbeil, J. Sauch, and W. Jakubowski for critical reading of the manuscript, to S. McFarlin for excellent typing, and to key UCSD personnel for help in obtaining dogs. This document has been reviewed in accordance with the U.S. Environmental Protection Agency policy through assistance agreement number 811950010 to University of California at San Diego and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. Literature Cited 1. Barlough, J.E.. 1979. Canine giardiasis: A review. J. Small Anim. Pract. 20:613623. 2. Bemrick, W.J.. 1961. A note on the incidence of three species of Giardia in Minnesota. J. Parasitol. 47:8789. 3. Bemrick, W.J.. 1963. Observations on dogs infected with Giardia. J. Parasitol. 49:10311032. 4. Bemrick, W.J.. 1984. Some perspectives on the transmission of giardiasis. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. Erlandsen S.L. and E.A. Meyer, eds. Plenum Press, New York. pp. 379400. 5. Craun, G.F.. 1984. Waterborne outbreaks of giardiasis. In: Giardia and Giardiasis. Erlandsen S.L. and E.A. Meyer. Plenum Press, New York. pp. 243261. 6. Davies, R.B. and C.P. Hibler. 1979. Animal reservoirs and crossspecies transmission of Giardia. In: Waterborne Transmission of Giardiasis. Jakubowski W. and J.C. Hoff, eds.. U.S. Environmental Protection Agency, Cincinnati, Ohio. pp. 104126.
Page 69
7. Douglas, H., Reiner D.S. and F.D. Gillin. 1987. A new method for purification of Giardia cysts. Trans. Roy. Soc. Trop. Med. Hyg.. 81:315316. 8. Faust, E.C.. 1931. Habitat of Giardia in the intestine. Proc. Soc. Exp. Biol. Med. 528:621623. 9. Gault, M.J., Gillin, F.D. and A.J. Zenian. 1987. Giardia lamblia: Stimulation of growth by human intestinal mucus and epithelial cells in serum free medium. Exp. Parasitol.. 64:2937. 10. Gillin, F.D., Reiner, D.S., Gault, M.J., Douglas, H., Das, S., Wunderlich, A. and J. Sauch. 1987. Encystation and expression of cyst antigens by Giardia lamblia in vitro. Science. 235:10401043. 11. Hegner, R.W.. 1922. A comparative study of the Giardias living in man, rabbit and dog. Am. J. Hyg. 2:422454. 12. Hewlett, E.L., Andrews, J.S., Ruffier, J. and F.W. Schaefer III. 1981. Experimental infection of mongrel dogs with Giardia lamblia cysts and cultured trophozoites. J. Infec. Dis. 145:8993. 13. Levine, N.D.. 1979. Giardia lamblia: Classification, structure, identification. In: Waterborne Transmission of Giardiasis. Jakubowski W. and J.C. Hoff, eds.. U.S. Environmental Protection Agency, Cincinnati, Ohio. pp. 28. 14. Nash, T.E., McCutchan, T., Keister, D., Dame, J.B., Conrad J.D. and F.D. Gillin. 1985. Restrictionendonuclease analysis of DNA from fifteen Giardia isolates obtained from humans and animals. J. Infec. Dis. 152:6473. 15. Pitts, R.P., Twedt, D.C. and K.A. Mallie. 1983. Comparison of duodenal aspiration with fecal flotation for diagnosis of giardiasis in dogs. J. Am. Vet. Med. Assoc. 182:12101211. 16. Sogayar, M.I.L., Curi, P.R., and E.F. da Silva. 1987. Giardia canis Hegner, 1922. Localizacao no tubo digestiva de caes naturalmente infectados. Arq. Bras. Med. Vet. Zool. 39:265272. 17. StehrGreen, J.K., Murray, G., Schantz, P.M. and E. RuizTiben. 1985. Intestinal parasites in pet store puppies. Abstract # 90, Amer. Soc. Trop. Med. Hyg. 34th Meeting, Miami, Florida, U.S.A. Nov. 3, 1985. 18. Tsuchiya, H.. 1931. The localization of Giardia canis (Hegner, 1922) as affected by diet. Am. J. Hyg. 15:232246. 19. Wolfe, M.S.. 1984. Symptomatology, diagnosis and treatment. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. Erlandsen S.L. and E.A. Meyer, eds.. Plenum Press, New York. pp. 147162. 20. Woo P.K., 1984. Evidence for animal reservoirs and transmission of Giardia infection between animal species. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. Erlandsen S.L. and E.A. Meyer, eds.. Plenum Press, New York. pp. 341364. 21. Woo, P.K. and W.B. Patterson. 1986. Giardia lamblia in day care centres in southern Ontario, Canada, and susceptibility of animals to Giardia lamblia. Trans. Roy. Soc. Trop. Med. Hyg. 80:5659. 22. Zenian, A. and F.D. Gillin. 1985. Interactions of Giardia lamblia with human intestinal mucus: Enhancement of trophozoite attachment to glass. J. Protozool. 32:664668.
Page 71
Seasonal Increase in the Incidence of Giardia lamblia in Arkansas James J. Daly, Mark A. Gross*, David McCullough, Thomas McChesney, Suzanne K. Tank, Eleanora B. Daly, and Cheryl L. Puskarich. Department of Electronics and Instrumentation, University of Arkansas at Little Rock, Graduate Institute of Technology, Little Rock, AR 72204, U.S.A.. A seasonal increase for the presence of Giardia lamblia in stool examinations was found by three clinical laboratories in central Arkansas. Data from the records of the Arkansas State Health Department (19831985), Arkansas Children's Hospital (19831985), and St. Vincent Infirmary (19801986) show that stool specimens positive for G. lamblia begin to increase in midsummer, peak in September, and decrease thereafter. Geographic origins of the patients whose stools were examined at these institutions are primarily from the Little Rock North Little Rock metropolitan areas and the mountainous regions of the state. The seasonal epidemictype increase is, in part, dependent on the total number of stools examined per month but the reason for the increase in the percentage of positive stools is not known. Ages of patients were not available from the Arkansas State Health Department and St. Vincent Infirmary but data from Arkansas Children's Hospital show that the seasonal increase occurs in the pediatric population.
Introduction A review of the records of the parasitology service of the Arkansas State Health Department in 1985 revealed a noticeable increase in the number of stools positive for Giardia lamblia during the months of August, September, and October. Further examination of these records showed that this seasonal increase occurred every year from 19831985. Since such a seasonal increase could have both epidemiologic and health care significance, it was decided to investigate this finding further by examining the records of two other laboratories in central Arkansas. Selection of these laboratories, Arkansas Children's Hospital and St. Vincent Infirmary, was based on the different socioeconomic characteristics of the patients seen at each institution and the availability of laboratory data. Endemicity of Giardia lamblia in Arkansas Giardia lamblia is currently the most commonly reported enteric protozoan parasite in Arkansas. Its presence in the state has been known for many years. An average of 37 positive stools per year were found by the old State Hygiene Laboratory between the years 1954 and 1963 (2). In 1978, 67 G. lambliapositive stools were found by the State Health Department Laboratory, which represented 4.1% of the total stools examined by the State. The midsouth regional percent figure for G. lamblia positive stools at that time was 5.7 and the national figure was 3.8% (8). Since then an average of about 5% has been found for all stools examined by the State Laboratory. Giardia had also been found in a potential animal reservoir in Arkansas. In 1985 a prevalence rate of 11.5% was reported in 78 beavers collected from diverse areas of Arkansas. Animals that were found positive for Giardia had been trapped from seven different counties (6). Institutional Profiles The parasitology service of the Arkansas State Health Department (AHSD) performs approximately 2200 fecal examinations a year. The stools examined do not represent a fair sampling of the state's population and the test results cannot be used as an accurate assessment of a parasite's incidence in Arkansas. Because of the central location of ASHD, many of the stools submitted are from the Little Rock North Little Rock area. Until 1986 the University Hospital of the University of Arkansas for Medical Sciences utilized the parasitology service of the ASHD and this represented a considerable proportion of the stools received from the central Arkansas area. Outside of central Arkansas the service is primarily used by physicians who do not have ready access to a good parasitology laboratory. Stools are infrequently seen from patients living in or near many of the larger cities in Arkansas. The percent distribution by geographic region of stools done by ASHD in a typical year are: Little Rock North Little Rock, 48%; Ozark Mountains, 25%; Mississippi Plains, 16%; Ouachita Mountains, 7%; and Gulf Coast Plains, 4%. Arkansas Children's Hospital (ACH) is a nonprofit pediatric hospital associated with the University of Arkansas for Medical Sciences. The parasitology service of ACH performs from 1000 to 1200 stool examinations per year. Because of its pediatric emphasis the relative number of stools positive for parasites is higher than at the other two institutions in this study. Geographic data on patients was similar to ASHD and St. Vincent Infirmary with heavy usage by central Arkansas residents (55%). St. Vincent Infirmary is one of the largest (600 bed) private hospitals located in central Arkansas. Patients tend to be higher in economic status than patients utilizing the other two facilities. The parasitology service of the clinical laboratory performs from 1500 to 1800 examinations per year. Due to the type of patients at St. Vincent Infirmary the percent positive stools for parasites is very low, relative to the other two institutions. * Corresponding author.
Page 72
Figure 1. Monthly incidence of Giardia positive stool specimens found by the Arkansas State Health Department and Arkansas Children's Hospital Laboratories for the years 19831985.
Methods Data were obtained from the records of the parasitology services of the clinical laboratories of ASHD, ACH, and St. Vincent Infirmary. All three laboratories utilize the PVA fixationtrichrome staining and formalinethyl acetate concentrationiodine staining procedures for examination of stool specimens. Cases positive for G. lamblia were recorded, with repeat or duplicate positives on the same individual being ignored, if occurring within a six week period. Patient's clinical histories were not consulted, therefore positive stools in this study only reflect incidence of the parasite and not the clinical status of the giardiasis. Statistical methods for correlation and the use of the Wilcoxon Two Sample Test (a nonparametric analog of analysis of variance) are from Sokol and Rohlf (12). Results Analysis of data from the records of ASHD and ACH for a three year period show a yearly cycle in which an increase in the number of G. lamblia positive stools began in midsummer, peaked in or near September, and decreased thereafter (Figure 1). This effect is even more evident if the data for each month for the three year period are averaged (Figure 2). In Figure 3 are data from St. Vincent Infirmary in which the data for each month are totaled for a six year period (19801986). The relatively fewer numbers of positive stools seen at St. Vincent precluded an individual monthly average as was done for ASHD and ACH. Data from all three institutions clearly show that the cyclical increase is not peculiar to one laboratory nor to one set of patients.
Figure 2. Average monthly percent of Giardia positive stool specimens found at the Arkansas State Health Department and Arkansas Children's Hospital for the years 19831985.
Page 73 TABLE 1. Monthly incidence of G. lamblia in stools examined by the Arkansas State Health Department for the years 1983 to 1985 Average No. of Specimens Submitted
Average No. of Positive Specimens
Percent Positive
January
190.0 ± 37.8*
6.3 ± 5.5*
3.3
February
175.3 ± 53.g
6.7 ± 4.9
3.8
March
212.7 ± 75.1
6.0 ± 6.0
3.1
April
151.3 ± 17.6
6.0 ± 3.0
4.0
May
175.3 ± 29.3
6.0 ± 3.6
3.4
June
171.3 ± 41.0
4.3 ± 4.1
2.5
July
208 ± 35.4
11.3 ± 3.2
5.4
August
245.7 ± 71.5
13.7 ± 2.9
5.6
September
215.3 ± 60.g
17.7 ± 2.1
8.3
October
232.3 ± 64.8
13.0 ± 3.5
5.6
November
186.7 ± 38.1
5.0 ± 2.0
2.7
December
116.7 ± 14.2
3.0 ± 1.0
2.6
Month
*Mean ± S.D.
Table 1 shows the average number of stool examinations done per month and the average number of positive stools for each month for a three year period at the ASHD. The Wilcoxon Two Sample Test was used to compare the mean number of positives in the months comprising the seasonal increase (July, August, September, and October, n = 12) with the means of the other months (n = 36). A significant difference is found between the number of positive stools seen during the peak months and the other months of the year (p = < 0.001). It was considered that the increase in the total number of stools during the peak of the cyclical period might be a factor in the increase in the number of positive stools found. Such a relationship would indicate an enhancement in the probability of finding G. lamblia. The number of stool examinations done per month was highly variable, as indicated by the large standard deviations, but there was a significant correlation between the number of positive tests and the total number of stools submitted (r = 0.61, p = < 0.01).
Figure 3. The total number of positive stool specimens found at St. Vincent Infirmary for each month from 19801986.
This indicates that the increased number of positive Giardia stool specimens may, in part, be a result of the increase in the total number of stools submitted during the seasonal period of peak incidence. The correlation, however, cannot address the question of why more positive stools cluster around the month of September. Evaluation of patient data from ASHD based on region and the presence of G. lamblia revealed the following percentage of positive stools by physiographic region for a three year period (19831985): Little Rock North Little Rock, 5.8%; Ozark Mountain, 5.3%; Ouachita, 4.3%; Mississippi Plain, 4.0%; and Gulf Coast Plain, 6.8%. From these values there appears to be no obvious regional concentration of G. lamblia. This data also confirms an earlier, less specific report (9) showing statewide distribution of the parasite. Discussion Epidemics of giardiasis in the United States have been primarily associated with contaminated water supplies (4), and day care centers (13). In the present study epidemiclike curves are reported for the presence of G. lamblia that show a yearly cyclical occurrence. Recently the Division of Public Health, State of Delaware (1), reported a similar seasonal occurrence in Delaware, which peaked during the late summer and early fall period for both 1985 and 1986. This increase could not be explained by day care center involvement, age distribution, or geographic location. In Arkansas, specimens from day care centers were not numerous enough to alter the refined data and, although the data was skewed toward central Arkansas, geographic locations did not seem to be a factor. Age distribution was unavailable for two of the institutions in the present study but data from ACH showed that the seasonal increase is found in the pediatric population. The data from Arkansas and from Delaware suggest that there may be a "Giardia season" in certain areas of the United States. A major factor in the increase of positive stools during mid and late summer in Arkansas appears to be partly due
Page 74
to the increase in the total number of stools submitted. However, other factors are probably involved as indicated by the greaterthanexpected percent increase in the number of positive stools during certain months of the seasonal increase. The reasons for this increase are not known. The cyclical occurrence in Arkansas coincides with the greatest period of outdoor activity and water resource usage. It has been noted that outbreaks of waterborne diseases associated with water supplies have peaked during the summer months in the U.S. (5). The most obvious possibility to explain this yearly seasonal increase in Arkansas would be the accidental or purposeful use of untreated surface water for drinking. This is supported by the known endemic human and animal reservoirs in the state which would provide a source of water contamination as previously noted by epidemiologic studies done in the western United States (12, 11). The mountainous, western sections of Arkansas resemble the western states in having fastmoving streams with deceptively clear water and heavy recreational use. However, there are no substantive data from Arkansas to support this conclusion. Other factors may be involved, including the exotic possibility of a circannual rhythm (10) such as reported with gastric ulcers (7). Because of the public health implications, further epidemiologic studies should be done to clarify the seasonal increase of Giardia lamblia in states where this phenomenon is found. Acknowledgements The authors want to thank Mr. John Clarke, Department of Physiology, University of Arkansas for Medical Sciences and Paul T. Archer, University of Arkansas at Little Rock, Graduate Institute of Technology, for production of the figures. Literature Cited 1. Anonymous. 1986. Frequently reported disease series: Giardiasis. Division of Public Health. State of Delaware. Delaware monthly surveillance report 86:No.12. 2. Anonymous. 1964. Public health at a glance; Intestinal parasites. J. Ark. Med. Soc. 61:8485. 3. Barbour, A.G., Nichols, C.R. and T. Fukaskima. 1976. An outbreak of giardiasis in a group of campers. Am. J. Trop. Med. Hyg. 25:384389. 4. Craun, G.F. 1979. Waterborne giardiasis in the United States: A Review. A. J. P. H. 69:817819. 5. Dykes, A.C., Juranek D.D., Lorenz, R.A., Sinclair, S., Jakubowski, W. and R. Davies. 1980. Municipal waterborne giardiasis: An epidemiologic investigation. Ann. Intern. Med. 92:165170. 6. Heidt, G.A., Nichols, A.H. and J.J. Daly. 1985. Incidence of Giardia in Arkansas beaver. Ark. Acad. Sci. Proc. 34:137. 7. Horwitz, M.A., Hughes, J.M. and G.F. Craun. 1974. Outbreaks of waterborne disease in the United States. J. Infect. Dis. 133:588592. 8. Howell, R.T. and B.S. Waldron. 1978. Intestinal parasites in Arkansas. J. Ark. Med. Soc. 75:212214. 9. Ivy, C.S. and J.E. Steed. 1971. Intestinal parasites of Arkansas, J. Ark. Med. Soc. 67:329331. 10. Markiewicz, A., Koszyk, T. and J. Reising. 1986. Seasonal variation of inflammation and ulcer incidence as assessed by upper digestive tract endoscopy. Chronobiologia 1986: 117121. 11. Pasley, J.N.. 1987. University of Arkansas For Medical Sciences, Personal Communication. 12. Rohlf, J.F. and R.R. Sokol. 1981. Statistical Tables, 1st ed., W.H. Freeman and Co., San Francisco. 13. Sealy, D.P. and S.H. Schuman. 1983. Endemic giardiasis and day care. Pediatrics 72:154158.
Page 75
Infection of Mongolian gerbils (Meriones unguiculatus) with Giardia from Human and Animal Sources K. Diane Swabby*, Charles P. Hibler and John G. Wegrzyn Department of Pathology, Colorado State University, Fort Collins, Colorado 80523, U.S.A.. For the past four years a specificpathogenfree breeding colony of Mongolian Gerbils, (Meriones unguiculatus), has been maintained for the purpose of providing human source cysts of Giardia duodenalis for our research and service commitments. When cases of giardiasis in humans and other animals are presented, and gerbils are not committed to other projects, crosstransmission studies have been attempted. The standard procedure is to use 5 to 7 weekold gerbils and expose each to 5000 cysts by gavage. Five uninoculated gerbils serve as negative controls. Twentyfour human source transmissions have been attempted and 14 (58%) were successful. The prepatent period varies between 5 and 8 days. For other animals the results are: 4/4 beaver, prepatent period 5 to 8 days; 2/2 chinchilla, prepatent period 5 to 6 days; 2/2 domestic cats, prepatent period 5 to 7 days; 6/6 muskrats, prepatent period 5 to 7 days; 1/2 domestic cattle, prepatent period 7 days; and 1/1 horses, prepatent period 6 days. Numerous attempts to crosstransmit from dogs have failed. One domestic pig and 5 gerbils were infected with one human source, but cysts from this pig failed to infect two different groups of gerbils.
Introduction Research on giardiasis in humans and other animals has increased considerably over the past decade, initially because of the discovery that Giardia is responsible for most of the epidemics of waterborne disease (3) and subsequently because of the frequency the disease is diagnosed in children in daycare centers, kindergartens and nurseries and in other situations (travelers, hikers, etc.). Much of the research necessitates use of cysts and/or trophozoites from Giardia duodenalis of human origin. Often, as in the evaluation of pilot filters designed to test efficiency of filtration techniques or to develop C.t values for inactivation by ozone, chlorine, etc., millions of viable cysts are necessary. While effective in vitro excystation procedures have been developed for some of these needs, in vitro encystation has not been perfected. Human donors rarely are available and human stools obtained from hospitals, etc. frequently contain cysts that are too old, moreover, they are seldom available when needed by the investigator. When Davies and Hibler (1979) (4) performed some of the early crosstransmission experiments and noticed that the gerbil (Gerbillus gerbillus) was a reasonably good host, this prompted Faubert (1,3) to evaluate the Mongolian gerbil (Meriones unguiculatus) as an animal model for Giardia. Their success prompted us to use this animal for our many research and service commitments. Our laboratory has supplied human source Giardia cysts for use in engineering studies, as quality controls for waterborne diagnostic procedures, and many other research efforts over the past two years. Currently we supply from 20 to 100 × 106 cysts/week to various researchers. The purpose of this report is to share with you some of the problems, failures and successes associated with infecting and maintaining infection in Mongolian gerbils with Giardia cysts from various sources. Methods and Materials The Gerbils The original breeding pairs were purchased from Tumblebrook Farms, Inc., West Brookfield, Massachusetts. They were isolated and treated daily for 5 days with 0.6 mg of Flagyl (metronidazole) by gavage. If the offspring (weanlings) from any of these pairs were infected with Trichomonas sp. or Endamoeba sp., the breeding pairs were again treated. If adults were infected after a second treatment they were eliminated from the colony. Presence or absence of infection in the offspring was determined at postmortem by scrapings and/or washings from the intestine and cecum. When the colony was determined to be free of these commensals, F2 generation offspring were used for experimental studies. Thorough records were kept for age, breeding success, and subsequent matings. Eventually, gerbils were selected for breeding pairs based on the success of infecting their siblings with cysts of Giardia from human sources. While the success of this approach would be difficult to analyze, we continue to select breeding pairs on this basis. Infection of Gerbils Fecal material containing Giardia cysts was diluted with distilled water, filtered through several layers of gauze (if necessary), and then through a 40 mesh (40 openings/inch) screen and the cysts/mL of suspension determined by direct count with a calibrated micropipette. If not used immediately the suspension was stored at 5°C. If samples were to be stored for several days the water to the sediment was replaced daily and the sample gently agitated. The routine dosage/gerbil used in an attempt to establish infection was 5 × 103 cysts administered by gavage. Five 5 to 7 weekold gerbils were used in all experimental crosstransmission trials and 5 uninoculated gerbils served as negative controls. Collection of Cysts from Gerbils Beginning 4 days postinoculation and continuing through 810 days postinoculation, gerbils were placed as a group in a standard clear plastic mouse cage containing a wire mesh floor raised about 25 mm above the bottom. The bottom of the cage was flooded with distilled water. Gerbils remained in the cage for 1 hour and were then returned to their original cage to be fed sunflower seeds before repeating * Corresponding author.
Page 76
the procedure. Often it was necessary to collect gerbils individually and, sometimes, euthanasia was necessary depending on the experiment. Pellets collected were immediately macerated through an 80 mesh screen and distilled water added at a ration of 1/50 suspension/distilled water and stored at 5°C. Initial examination of the suspension was by direct microscopy. If no cysts were observed, the sample was evaluated by ZnSO4 centrifugal floatation. If cysts were being produced, an arbitrary scale of cyst numbers/100× by direct microscopy was used: rare (15); occasional ~ (550); + (50100); ++ (100500); +++ (500 1000); and ++++ (1000 to too numerous to count). Results and Discussion General Early attempts to infect Mongolian gerbils with cysts of Giardia from various sources often produced variable results. Frequently these were dog sources and/or human sources obtained from children infected in daycare centers. Sources from beaver and/or muskrat were always successful, but our interest was primarily to establish human sources. Unfortunately during these early attempts (almost a year) good records of success versus failure were not maintained. Eventually, through experience, we improved on our ability to evaluate the quality of the cysts from the source animals, improve fecal collection and evaluation procedures (postmortem) for the gerbils, and keep better records regarding the source and quality of the cysts used for inoculation. The only attempts at crosstransmission included in this report are those attempts where the cysts were fresh and microscopic evaluation indicated they were viable. The results are summarized in Table 1. The 14/24 (58%) human sources that established in gerbils originated from individuals ranging in age from 4 months to 39 years. Generally cysts obtained from individuals (supposedly) infected while backpacking, hiking, etc. would establish an infection while cysts obtained from young children either did not establish infection or was poorly adapted to gerbils and could not be further passaged. In most of the infections cyst passage began between days 5 and 7. One source began cyst passage on day 8. Those animals not passing cysts by day 8 were euthanised and examined postmortem for trophozoites. Trophozoites were not found in animals inoculated with the 10 sources that did not infect gerbils. Four of the 14 successful attempts began a heavy (++++) cyst passage on day 5 postinoculation and 9598% of the cysts were morphologically excellent. Most of the infections resulted in low to moderate cyst passage by days 6 to 7. Subsequent passage of these cysts into gerbils often, but not always, resulted in greater cyst production and better quality cysts than the initial infection; however, it was not unusual for the source to fail after two or three passages if cyst production was initially poor. Generally if cysts could be passaged 5 to 6 times into gerbils at weekly intervals, the source could be maintained for at least 15 passages before the source failed. One source was passaged 27 times before failing and at that time the cysts appeared viable. At attempt to reestablish the infection with cysts from passage 25 was successful to passage 27 when they again failed. Currently we maintain two sources that have been TABLE 1. Susceptibility of Mongolian gerbils to human and animal Giardia cyst sources Cyst Source
Month
Prepatent Period (Days)
HUMAN
January
5
++++
January
7
+
January
not infectious
February
7
~
February
not infectious
February
not infectious
February
7
+
February
6
+++
February
7
+++
March
5
++
April
not infectious
April
not infectious
May
not infectious
June
5
++
July
not infectious
July
5
++
July
not infectious
July
not infectious
July
8
~ to +
August
6
++++
September
6
++
November
5
++
December
5
~ to +
December
not infectious
Cyst Production in Gerbils
ANIMAL
Beaver 4
Chinchilla 2
5
Dom. Cat 2
57
2/2 ++ to +++
Muskrat 6
6
6/6 ++ to ++++
Dom. Cattle 2
7
1/2 +
Horse 1
6
1/1 ++++
Pig 1
6
not infectious
Dog numerous
not infectious
58
4/4 ++ to ++++ 2/2 +++
passaged weekly 21 to 50 times, respectively. Some human sources that can be passaged regularly through gerbils never seem to adapt sufficiently so that they can be used for other experimental purposes (e.g. chlorine studies). Cyst production is often variable, beginning as late as day 6 or 7, and cyst quality ranges between 85 90% viable as determined by microscopic evaluation. Cyst production by gerbils infected with a human source that quickly adapts and results in an abundance of viable cysts will usually continue cyst production through days 8 to 10 post inoculation and then cyst production either ceases or becomes rare and sporadic. Generally we collect cysts from days 5 through 8 and then terminate the trial. One human source we currently use often produces considerable numbers of cysts 15 days post inoculation. If human source Giardia is to be maintained in Mongolian gerbils as a source of cysts for experimental purposes, the human source must be welladapted to gerbils and produce 5 × 105 to 8 × 105 cysts/gerbil/day (8 hour collection) during peak production (days 5 through 8) and the cysts must be 9598% viable. If cysts from a
Page 77
source that will not readily adapt to gerbils are used, the results of experiments using the cysts (e.g. establishing C.t values for inactivation of cysts by chlorine, etc.) can be misleading and often disastrous. Even when using sources welladapted to gerbils the cysts must be collected properly, maintained properly, and passaged regularly through the animals. Continuous monitoring of the source in use is necessary to achieve reliable results. Most of the other crosstransmission experiments (Table 1) were never passaged more than 2 or 3 times in gerbils before the source was discarded, primarily due to the cost and a fear of contamination of the human sources maintained for other experiments. We did note that while beaver, muskrat and chinchilla sources readily established in gerbils, some of the beaver and/or muskrat sources were no better adapted, based on cyst production and quality of the cysts, than some of the human sources. The horse source and one of the two cat sources apparently were welladapted for gerbils, producing large numbers of viable cysts between days 5 and 7. The cattle source was only passaged two times and cyst production by the gerbils was always poor. In one experiment a weanling domestic pig was maintained and examined for Giardia cyst production for two weeks. It was then exposed by intubation to a human source and 5 gerbils also were exposed. Both species began passing cysts on day 6. Cysts from the pig were of excellent quality, but infection could not be established in two different groups of 5 gerbils/group with cysts collected on days 7 and 8 from the pig. Literature Cited 1. Belosevic, M., G.M. Faubert, J.D. MacLean, C. Law, and N.A. Croll. 1983. Giardia lamblia infections in Mongolian gerbils: animal model. J. Inf. Dis. 147:222 226. 2. Belosevic, M., G.M. Faubert, T.S. Walker, E. Meerovitch. 1983. Comparative studies on the pattern of infection with Giardia spp. in Mongolian gerbils. J. Parasitol. 802805. 3. Craun, S.F. 1986. Waterborne diseases in the United States. CRC Press, Inc., 2000 Corporate Blvd., N.W., Boca Raton, FL. 295 pp. 4. Davies, R.B., and C.P. Hibler. 1979. Animal reservoirs and crossspecies transmission of Giardia In: Waterborne Transmission of Giardiasis. Jakubowski, W. and J.C. Hoff, (eds.). United States Environmental Protection Agency. pp 104126. 5. Wegrzyn, John G. 1987. Giardia in Colorado muskrats; PhD. dissertation; Colorado State University, Fort Collins, Colorado.
Page 79
Transmission of Giardia Duodenalis from Human and Animal Sources in Wild Mice P.D. Roach and P.M. Wallis* Kananaskis Centre for Environmental Research, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada, T2N 1N4 Mature Deer Mice (Peromyscus maniculatus) were trapped at random in the vicinity of the Kananaskis Field Station. Absence of Giardia infection was confirmed by three faecal analyses both before and after treatment with three consecutive, daily doses of 10 mg of metronidazole. Cysts of seven strains of G. duodenalis were produced by infecting gerbils (Meriones unguiculatus) with trophozoites from cultured stocks. Cysts of two additional strains that had been maintained by serial passage through Swiss Webster mice were also used to infect gerbils. Large numbers of viable cysts were used to infect wild Deer Mice by gavage and the results were determined by duodenal biopsy. Wild Deer Mice became infected with cysts of G. duodenalis strains that originally came from sheep, muskrat, cattle, and some human strains. No infection resulted from G. duodenalis from beaver and a dog. An inverse relationship was found between length of time a strain had been kept in culture and success of infection.
Introduction Historically, parasitologists have assumed that different hosts harboured separate species of Giardia and the literature is replete with species names derived from host animals. Filice (4) simplified the taxonomy of this genus by reorganizing it into three species; G. agilis (a morphologically distinct parasite of amphibians), G. muris (found in mice and rats), and G. duodenalis (found in man and other mammals including some rodents). This scheme is now generally accepted by most workers. G. muris is very similar in morphology to G. duodenalis but no evidence is available to suggest that it is infective to humans. It is sometimes possible to distinguish G. muris from G. duodenalis (4,5) by morphology although the degree of overlap between the ranges normally encountered for the relevant morphometric measurements often make identification of individual specimens impossible (1). Giardiasis is a notifiable disease in Alberta and approximately 150 cases are reported each month. Veterinarians commonly see Giardia in animal stools. The frequency of these reports suggest that Giardiasis is endemic among both animals and humans. Transmission between humans probably accounts for many of the reported cases but infection from animals may also occur. Beavers (Castor canadensis) have been implicated in zoonotic infections (3, 7, 8, 9, 16) but the role of other wild and domestic animals in the transmission of Giardia is unknown. Giardia spp. have been detected in a number of animals in southern Alberta including voles (Clethrionomys gapperi, Microtus pennsylvanicus), mice (Peromyscus maniculatus), dogs (Canis familiaris), wood rats (Neotoma cinerea), muskrat (Ondatra zibethicus), cattle, (Bos bovis), sheep (Ovis aries) and beaver (13, 15). Evidence supporting crossinfection of strains of G. duodenalis from animals other than beavers is accumulating. Davies and Hibler (2) have shown that human Giardia cysts are infective to a variety of animals including rats (Rattus norvegicus), gerbils (Gerbillus gerbillus) guinea pigs (Cavia porcellus), beavers, racoons (Procyon lotor), dogs, and bighorn X mouflon sheep (Ovis canadensis X O. musimon). Hewlett et al. (6) successfully infected SPF beagles with Giardia cysts from a human donor. Woo (17) criticized the results of many of these crossinfection experiments but concluded that dogs, and probably beaver, could act as a reservoir for human infective Giardia. We believe therefore, that Giardia found in humans and some animals are nonspecific and zoonotic transmission is probable. The purpose of this study was to determine if deer mice (Peromyscus maniculatus) could harbour G. duodenalis from different animal sources. The deer mouse was chosen as an animal model because it is common throughout North America. Its natural rate of infection in southwestern Alberta is approximately 7% (15), a prevalence rate similar to that in human populations (12). This study was not designed to demonstrate that Peromyscus maniculatus is an important reservoir of human infective Giardia but that transmission of G. duodenalis to an abnormal host can occur under the right conditions. Materials and Methods Animal Model Mature P. maniculatus were trapped at random in the vicinity of the Kananaskis Field Station in southwestern Alberta, Canada. Several of these were found to be pregnant when captured and the young were reared in the laboratory and used for infection experiments. All animals trapped in the field were assumed to be adults and the age of younger animals born and reared in the laboratory was recorded. All deer mice were housed in shoe box cages in the animal rooms at the Field Station. Three consecutive, daily, faecal samples were collected using wire mesh grids supported over a small * Corresponding author.
Page 80 TABLE 1. Sources of strains of G. duodenalis. Strain
Original Host
Date
Location
H8
Human
850826
Canmore
H7
Human
850803
Calgary
CH3
Human
860715
Fort Collins, Colorado (from C. Hibler, his strain #H3)
Sheep (domestic)
850702
farm near Strathmore 45 km east of Calgary
MR4
Muskrat
850920
Sibbald Meadows Pond Kananaskis Country 65 km west of Calgary
MR7
Muskrat
860529
farm near Strathmore 45 km east of Calgary
D3
Dog
840816
Calgary Animal Shelter
B5
Beaver
850619
Sibbald Meadows Pond Kananaskis Country 65 km west of Calgary
Cow
860529
farm near Strathmore 45 km east of Calgary
S1
CW2
amount of water. Faecal samples were examined for cysts using the sucrose gradient method of RobertsThomson et al. (11). All deer mice were then treated with 3 consecutive, daily doses of 7 mg of metronidazole by gavage. Higher doses were abandoned because of unacceptable mortality rates. After completing metronidazole treatment, the mice were again placed on grids for three consecutive days and the collected faeces were examined for the presence of cysts. If they were found to be free of Giardia, the mice were placed in fresh shoe box cages and supplied with autoclaved bedding, food and water (ad lib). Strains of G. Duodenalis Deer mice were challenged with cysts of nine strains of G. duodenalis (Table 1). These included 3 from humans, 1 from a domestic sheep, 2 from muskrat, 1 from a dog, 1 from a beaver, and 1 from a cow. Procedures for the isolation and culturing of strains were given in Wallis and Wallis (14). Cyst Production Cysts of all strains except CW2 and MR7 were produced by infecting gerbils (Meriones unguiculatus) with trophozoites from cultured stocks. This was accomplished by cooling a mature culture tube of trophozoites to 5°C for 30 minutes, centrifuging, and resuspending the trophozoites in 2 mL of phosphate buffered saline. Half a mL containing about 106 suspended trophozoites was used to inoculate gerbils treated as described in Wallis and Wallis (14). We have not been able to culture CW2 and MR7 and they are maintained in vivo in outbred Crl:CFW(SW)BR Swiss Webster mice. Cysts were concentrated from the faeces of these animals and approximately 105 were used to infect gerbils. In order to enhance cyst production, the gerbils were immunocompromised by adding dexamethasone (Schering) to their drinking water at a concentration of 40 µg/mL. Dexamethasone treatment began on the day of inoculation and was maintained until the animal was sacrificed. Cysts recovered from gerbil faeces were counted using a haemocytometer and their viability was determined by in vitro excystation after the method of Rice and Schaefer (10). Infection of Peromyscus Cysts recovered from gerbils were washed in tap water containing Triton X100 and used to inoculate from 6 to 11 mature Peromyscus for each strain of the 9 strains of Giardia tested. A total of 81 deer mice were inoculated. Infected animals were held for 5 days to allow any possible infection to establish itself, and then sacrificed by cervical dislocation after being anaesthetized with diethyl ether. Upon necropsy, biopsy samples of the duodenum, lumen, and caecum were taken and examined for protozoa using phase microscopy. Experimental deer mice were not treated with dexamethasone or deliberately immunocompromised in any way. The experiment was complicated by the presence of Trichomonas sp. in deer mice trapped in the field and in the gerbils used for production of cysts. The presence of Trichomonas is difficult to detect by faecal analysis and the organism is not very responsive to metronidazole at the dosages used (higher dosages were often lethal to the mouse). For this reason, the presence of Trichomonas sp. was recorded whenever it was detected. Results The results of all faecal analyses for deer mice before and after treatment with metronidazole were negative. We therefore assumed that the previous exposure of these deer mice to Giardia was minimal or non existent. The average numbers and viabilities of cysts inoculated are given in Table 2 along with the percentage of positive Giardia infections that resulted. The numbers of cysts inoculated ranged from 16,000 to 640,000 and viabilities ranged from 12 to 97%. Strains CH3 (9/11), S1 (1/6), MR4 (4/11) MR7 (6/10) and CW2 (5/11) all resulted in infection. In all cases large numbers of trophozoites were found. No infections were observed with strains H7, H8, D3, and B5 despite the inoculation of large numbers of viable cysts. The complete results of gut necropsies are listed in Table 3. A total of 25 deer mice became infected with Giardia originally obtained from human, sheep, muskrat and bovine sources. Trichomonas sp. was found in 57 of the deer mice but its presence was not well correlated with Giardia infection. Of the 25 positive Giardia infections, 16 (64%) were also positive for Trichomonas sp. Discussion The results in Table 3 show that it is possible for deer mice to act as a reservoir for G. duodenalis from a variety of hosts. Although some of the mice used were reared in the lab and were therefore less than 5 weeks old when used, these animals did not show a higher infection rate than older adults. No attempt was made to follow the course of the infection in positive cases but the numbers of trophozoites observed were always high when infection occurred. Our experience with mice in general is that they either become infected with large numbers of trophozoites that persist for TABLE 2. Numbers of cysts, % viability and results of infection in deer mice.
Strain
Average Inoculum
% Viability
% Giardia Positive
H8
50,000
12
0
H7
50,000
27
0
CH3
288,000
75
82
S1
150,000
35
17
MR4
500,000
25
36
MR7
250,000
95
60
D3
17,000
16
0
B5
640,000
37
0
CW2
16,000
97
45
Page 81 TABLE 3. Giardia (G) and Trichomonas (T) detected in Deer Mice infected with various strains of G. duodenalis.
Total
T
T+
T
T+
Strain
n
G+
G+
G+
G
G
H8
9
0
0
0
3
6
H7
6
0
0
0
3
3
CH3
11
9
5
4
0
2
S1
6
1
1
0
0
5
MR4
11
4
3
1
0
7
MR7
10
6
2
4
0
4
D3
7
0
0
0
2
5
B5
10
0
0
0
0
10
CW2
11
5
5
0
0
6
Total
81
24
16
9
8
48
at least 2 weeks or they do not become infected at all. Further work will be required to fully describe the course of infection with G. duodenalis in deer mice but it is probable that it will be analogous with G. duodenalis in gerbils. It is possible that the previous history of each strain of Giardia had an influence on the success of infection. The strains available to us had been kept in vitro or in vivo for varying periods of time. For example, CH3 had been passaged through gerbils 20 times before it was cultured and the trophozoites used for infecting gerbils had only been held in vitro for 2 months. CH3 was therefore a relatively new strain at the time of use. The other two human strains (H7 and H8) had been propagated in vitro for over a year before they were used in these experiments. Examination of the data in Tables 1 and 2 shows that none of the strains that had been held in culture from before July, 1985 were capable of infecting deer mice (although all of them infected gerbils). Further inspection of the data reveals that the % positive infection is inversely correlated with the length of time between isolation and use, regardless of the strain. We conclude that although it may be possible to infect young gerbils with older strains, mature deer mice trapped in the wild are more resistant to challenge. It is also interesting to note that two of the strains used, MR7 and CW2, proved refractory to culture despite repeated attempts. Both were obtained from the same farm on the same date and it is possible that they are identical. A previous isolate obtained from a sheep held at a different location on the same farm one year previously was successfully cultured and was also infectious to deer mice. The ability of deer mice to act as a host for potentially human infective Giardia could not have been predicted from some previous studies. It is apparent from these experiments, however, that these animals are capable of harbouring Giardia duodenalis for at least a short time. The purpose of these experiments was not to show that deer mice are an important reservoir for what must surely be an unnatural parasite but rather to demonstrate that crosstransmission can occur under the right circumstances if strains are sufficiently virulent. In other words the ability of G. duodenalis to crossinfect depends as much on the virulence of the strain, the quality of the inoculum, and the immunological status of the host as it does on the identity of the source host. Acknowledgements This research was supported by grants and contracts from the Alberta Environmental Research Trust and the Research Management Division of Alberta Environment. Literature Cited 1. Bertram, M.A., Meyer, E.A., Anderson, D.L. and C.T. Jones. 1984. A morphometric comparison of five axenic Giardia isolates. J. Parasit. 70:530535. 2. Davies, R.B. and C.P. Hibler. 1979. Animal reservoirs and crossspecies transmission of Giardia. In: Waterborne Transmission of Giardiasis. Jakubowski, W. and J.C. Hoff (eds.). U.S. Environmental Protection Agency, Cincinnati, Ohio, EPA600/979001, pp. 104126. 3. Dykes, A.C., Juranek, D.D., Lorenz, R.A., Sinclair, S., Jakubowski, W., and R. Davies. 1980. Municipal waterborne Giardiasis: an epidemiological investigation. Ann. Int. Med. 92:165170. 4. Filice, F.P. 1952. Studies on the cytology and life history of a Giardia from the laboratory rat. Univ. Calif. Pub. Zool. 57:53146. 5. Grant, D.R. and P.T.K. Woo. 1978. Comparative studies of Giardia spp. in small mammals in southern Ontario. I. Prevalence and identity of the parasites with a taxonomic discussion of the genus. Can. J. Zool. 56:13481359. 6. Hewlett, E.L., Andrews, J.S., Ruffier, J., and F.W. Schaefer III. 1982. Experimental infection of mongrel dogs with Giardia lamblia cysts and cultured trophozoites. J. Infect. Dis. 145:8993. 7. Lippy, E.C. 1981. Waterborne disease: occurrence is on the upswing. J. Am. Wat. Wks. Assoc. 73:5762. 8. Lopez, C.E., Dykes, A.C., Juranek, D.D. et al. 1980. Waterborne giardiasis: A communitywide outbreak of disease and a high rate of asymptomatic infection. Am. J. Epidemiol. 112:495507. 9. Morbid. Mortal. Wkly. Rpt. 1977. Waterborne giardiasis outbreaks. 26:269275. 10. Rice, E.W. and F.W. Schaefer III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709710. 11. RobertsThomson, I.C., D.P. Stevens, A.A.F. Mahmoud and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterol. 71:5761. 12. Schmidt, G.D. and L.S. Roberts. 1981. Foundations of Parasitology. 2nd ed., p. 89, C.V. Mosby Co., St. Louis, p. 795. 13. Wallis, P.M., BuchananMappin, J.M., Faubert, G.M. and M. Belosevic. 1984. Reservoirs of Giardia spp. in southwestern Alberta. J. Wildlife Dis. 20:279283. 14. Wallis, P.M. and H.M. Wallis. 1986. Excystation and culturing of human and animal Giardia spp. by using gerbils and TYIS33 medium. Appl. Environ. Microbiol. 51:647651. 15. Wallis, P.M., Zammuto, R.M. and J.M. BuchananMappin. 1986. Cysts of Giardia spp. in mammals and surface waters in southwestern Alberta. J. Wildlife Dis. 22:115118. 16. Wilson, H.P.S., Stauffer, S.J. and T.S. Walker. 1982. Waterborne giardiasis outbreak Alberta. Can. Dis. Wkly. Rpt. 820:9798. 17. Woo, P.T.K. 1984. Evidence for animal reservoirs and
Page 82
transmission of Giardia infection between animal species. In: Giardia and Giardiasis. Erlandsen, S.L. and E.A. Meyer (eds.), Plenum, New York, pp. 341364.
Page 83
WATER TREATMENT
Page 85
Water Treatment and the Giardia Cyst Albert van Roodselaar Alberta Environmental Centre, Bag 4000, Vegreville, Alberta, Canada TOB 4L0. The concern of this session is with Giardia as it pertains to Potable Water Treatment. The history of Potable Water Treatment reflects different concerns at various stages of societal development. At the time of the Romans, as described in the preserved writings of the engineer, Sextus Julius Frontinus, the primary emphasis was on transport of the optimum attainable aesthetic quality water from point of availability to the point of need. For this purpose the Romans built an impressive system of aqueducts. Even at this time, the special needs of drinking water were recognized and the waters of highest purity were reserved for this application, ''Marcia, so charming in its purity and and coldness, should serve wholly for drinking purposes". The difference between then and now is the practice of treatment to improve quality and the development of comprehensive criteria by which to judge "purity". This is important in our assessment of current water treatment processes such as clarification, filtration and disinfection since these are evaluated by their effect on various physical, chemical and microbiological parameters. Consequently, in determining the relevance of a given treatment process with respect to it's effectiveness on Giardia cysts, we may simply be considering another parameter which behaves in parallel with those already available in the literature or we may be considering a parameter which must be dealt with in an independent manner from those which have traditionally been of concern. Herein lies an important question, whether a simple, easily measurable parameter will serve as an adequeate surrogate of Giardia removal and/or inactivation. Giardia cysts have been detected in raw waters varying from contaminated lakes and highly turbid, coloured rivers to mountain streams which are crystal clear. The wide ranging presence is achieved by the variety of hosts available for the transmission of this organism. Traditionally, communities located on water sources of high quality, as determined by turbidity, colour and coliform count, have had nominal treatment facilities, often relying solely on chlorine disinfection to ensure microbiological quality suitable for potable use. This is in keeping with the Guidelines for Canadian Drinking Water Quality, 1978, which states, "All supplies derived from surface water sources should receive disinfection as a minimum treatment." (p.33) and "The total coliform group is therefore preferred as an indicator of treatment adequacy in drinking water supply systems. (p.25). However, outbreaks of giardiasis in communities served by systems which have in the past been considered adequate for microbiological protection have thrown into question the basis of this conclusion. Comparing the response of Giardia cysts and other microbiological organisms to disinfectants and other water treatment processes became necessary to understand how the goal of microbiologically safe drinking water could be attained inclusive of the recognition of this additional risk. Concerns relevant to drinking water treatment have thus come full circle. It is important to reiterate that the prime consideration for drinking water has been and should continue to be the production of a water of superior microbiological quality. It is of interest to examine the aspect of public health as it pertains to microbiological quality of drinking water. Table 1 demonstrates the high prevalence of gastroenteric difficulties among waterborne diseases. Waterborne diseases are either increasing or are increasingly reported in America (Craun 1977). Sharpened public awareness will place an ever increasing pressure on the water treatment industry to improve product quality and safety. Microbiological quality has been thrown into doubt and joins the other parameters currently in the public conciousness such as organics (carcinogens), lead, mercury, and taste and odour. Table 2 shows the response of a few selected organisms to chlorine disinfection (as hypochlorous acid). This data indicates that even prior to the generation of specific information on Giardia cysts, the resistance of another cyst, E. histolytica, relative to E. coli and polio virus, showed that the cyst structure was capable of being two or three orders off magnitude more resistant to chlorine than certain bacteria and viruses. Table 3, using data available in 1967, indicates that the condition under which chlorine TABLE 1. Distribution of etiology by waterborne disease outbreaks from 1946 to 1974* Outbreaks for water systems Community Disease
Other
Number
Percentage
Number
Percentage
Gastroenteritis (unknown etiology)
71
52.2
153
47.6
Infectious hepatitis
22
16.2
44
13.7
Shigellosis
13
9.6
33
10.3
Chemical poisoning
8
5.9
13
4.0
Giardiasis
7
5.1
8
2.5
Typhoid
6
4.4
51
15.9
Salmonellosis
6
4.4
9
2.8
Amebiasis
1
0.7
4
1.2
Poliomyelitis
1
0.7
Enteropathogenic E. coli
4
1.2
Tularemia
2
0.6
Leptospirosis
1
0.7
136
321
Total outbreaks
* Sources: Craun and McCabe (1973) and Craun et al. (1976).
Page 86 TABLE 2. Dosages of chlorine required to achieve 99% inactivation. Concentration Contact time c.ta (mg/L) (min)
Test Microorganism
Chlorine Species
pH
Temp °C References
E. coli
Hypochlorous Acid (HOCl)
0.1
0.4
0.04
6.0
5
Scarpino, et al., 1974
Poliovirous 1
Hypochlorous acid (HOCl)
1.0
1.0
1.0
6.0
0
Weidenkopf, 1958
0.5
2.1
1.05
6.0
5
Englebrecht, et al., 1978
1.0
2.1
2.1
6.0
5
Scarpino, et al., 1974
E. histolytica Cystsb
Hypochlorous acid (HOCl)
5.0
18.0
90
6.0
5
Snow, 1956
a Concentration of compound multiplied by contact time (mg/L)(min). b Extrapolated data.
TABLE 3. Lethality coefficients for different microorganisms based on treatment with free and combined chlorine at 5°C. Oxidant
Enteric bacteria
Protozoan cysts
Virus
Bacterial spores
HOCl
20
0.05
1.0
0.05
OCL
0.2
0.0005
0.02
0.0005
NH2Cl
0.1
0.02
0.005
0.001
From Morris, 1967.
disinfectant is applied (pH and concentration of NH4 in the water being treated) has a large impact on the effectiveness of the disinfectant dose. It should therefore not come as a major suprise that Giardia is playing havoc with disinfectant based treatment systems. The emphasis should be on establishing those conditions which provide adequate levels of protection against Giardia cysts, determining the current status of operating plants and implementing improvements in those plants which are at an unacceptable high level of risk. To do this, an understanding of treatment processes, both conventional and otherwise is required. Effectiveness of more than one process in reducing Giardia cyst risk in the treated water, either through removal or inactivation, results in a multibarrier screen. Some plants may depend on only one such barrier with a resultant Giardia penetration if this barrier fails or if the effectiveness of this barrier is overestimated. Consequently, detailed information on each barrier and the net impact of multiple barrier operation is necessary for the treatment engineer to be able to assess if his plant is at risk.
Page 87
Removal of Giardia through Slow Sand Filtration 100 Mile House, British Columbia Jack M.G. Bryck*, Brian L. Walker and David W. Hendricks Dayton & Knight Ltd., P.O. Box 91247, West Vancouver, B.C. V7V 3N9, Canada. In the fall of 1981, about 60 cases of waterborne giardiasis were confirmed at the village of 100 Mile House, Canada. Contamination of the village's water system was suspected. The service area population was about 2,000 persons. Beavers and muskrats, subsequently confirmed positive for the Giardia cyst, were trapped upstream of the village's surface water intake. At the time of the incident the surface water treatment included only chlorination with minimal contact time. After a pilot filtration program to monitor the effort required to operate several filter types in a manner identified by the U.S. EPA that would remove the Giardia cyst, the slow sand filtration process was chosen for incorporation into the village's new water treatment plant. Construction of the water treatment plant was completed December 15, 1984 and operation commenced November 15, 1985. The total construction cost was $780,000 CDN or $0.11 per litre of installed peak capacity. With funding from Health and Welfare Canada, an evaluation of the treatment plant was undertaken between October 1985 and October 1986. Included was monthly high volume sampling in the raw and filtered water for the Giardia cyst and other microorganisms and an evaluation of the microorganisms and biological processes present on the sand media. The removal through the slow sand filters of Giardia and other microorganisms in the raw water was completed.
Introduction In the fall of 1981 at least sixty cases of giardiasis were confirmed at 100 Mile House, British Columbia. The outbreak affected people living in a wide geographical area around 100 Mile House and no other source explained the findings as well as waterborne contamination of the municipal water system. The source was suspected to be beavers and muskrats located upstream of the village's surface water intake, which were subsequently confirmed positive for Giardia cysts. At the time of the incident the water supply was surface water from Bridge Creek, located within the uncontrolled Horse Lake watershed, and to a minor extent from a groundwater well. The addition of chlorine solution, with minimal contact time provided in the distribution system, was the only treatment practised. The surface water quality at the time of the incident is not known. However, subsequent testing over a calendar year determined that water quality generally exceeded the recommended objective levels set by the British Columbia Ministry of Health. This included turbidity which was less that 2 NTU over a calendar year and was generally less than 1 NTU over several months. The further development of the Bridge Creek surface water source, including a treatment plant incorporating filtration, was chosen by village council as the preferred method of meeting the village's water demands. After an onsite pilot testing of four types of filters between June and October 1983, including gravity and pressurized rapid rate multimedia, slow sand and diatomaceous earth filtration, slow sand filtration was chosen. The plant went into production November 15, 1985. Between November 15, 1985 and November 15, 1986 research was conducted to ascertain removal efficiencies of Giardia cysts, of coliform bacteria and of any other measurable indicator particles by the operating and pilot slow sand filters. Also, determination of the biological activity and enumeration of invertebrates within the filter media was made. Slow Sand Filtration The main component of a slow sand filter is the filter sand supported by graded gravel through which drains are placed to collect the filtered water. The effective sand size typically ranges from 0.15 to 0.35 mm with a uniformity coefficient less that 2. The depth of the sand is typically .75 to 1.2 mm. The system is contained by a filter box of sufficient height to permit about 1.8 m of headwater above the sand bed. The surface area of a slow sand filter is quite large when compared with rapid rate filtration, as the hydraulic loading rate for slow sand filtration is 0.04 to 0.4 m/h compared to 5.0 to 2.4 m/h for rapid rate filtration, a factor of 50 to 100 times slower. The operation of the filter is "passive", which means that the physical/biological processes within the filter bed require no active intervention by an operator other than adjustment of the filter production rate and scraping of the filter surface at terminal head loss. The initial headloss of the filter is typically 0.2 m increasing to about 1.3 m at the end of the run, depending on the headwater depth. At terminal headloss the filter bed is partially dewatered and the surface deposit, termed the schmutzdecke, is removed by scraping. Usually about 5 to 10 mm of sand is removed. After repeated scrapings over 5 to 10 years the depth of the bed is reduced to a lower limit of about .30 m, at which time it must be rebuilt to its original depth of 1.0 to 1.3 m. * Corresponding author.
Page 88
Operation A simplified process of transport, attachment and purification has been developed to explain removal of particulates (4). The screening process traps and retains particles too large to pass through the voids in the sand media. This typically occurs at the filter surface. With an effective grain size of 150 µm those particles greater than 20 µm should be removed by screening. The schmutzdecke or buildup of particulates at the filter will increase the effectiveness of the screening mechanism. The sedimentation process involves settling of particulates within media voids. For a water temperature of 10°C and organic matter with a density slightly greater than water the complete removal of particles greater than 4 µm by sedimentation can be expected. Particles less than 1 µm in diameter will not be removed by sedimentation. The principal mechanism that holds the particles in the raw water in place once contact has been made with the media is adhesion. The adhesion mechanism occurs at the surface of the slow sand filter, where bacteria and other microorganisms using the organic matter trapped on the filter surface as an energy source, produce a slimy mass termed zoogloea. The zoogloea, active bacteria, their wastes and dead cells and organic matter, forms a gelatinous film to which particulates in the raw water adhere. The trapped organic particles are assimilated into the film while the inert matter is held until the filter is scraped. Bacteria and microorganisms multiply using the deposited organic matter as an energy source for their metabolism and growth. As the bacterial population is dependent on the food source available the bacterial growth is accompanied by an equivalent bacterial dieoff which provides a simpler organic food source to lower depths in the sand filter. Through a number of steps the organic matter is gradually broken down into carbon dioxide, water, sulfates, nitrates and phosphates. Microorganism and bacterial activity is most pronounced in the upper part of the slow sand filter, gradually decreasing with filter depth as food levels decrease. As their food source decreases they starve, particularly during warm water temperatures, when metabolic rates are high. The slow sand filtration efficiency will be reduced at lower temperatures as the rate of metabolism of bacteria and other microorganisms such as bacteria consuming protozoa and nematodes will decrease. The result is an increase in the survival rate of intestinal bacteria carried through the bed. Recent work by Bellamy et al. (1) has shown that absorption and then metabolism by microbial films on the sand grains constitute a major removal mechanism. Pilot plant work has shown 99.9% removals of coliform bacteria spikes and 100% removals of Giardia cysts when the biofilm is sufficiently developed. Removal Efficiency An evaluation (4,5) of the effectiveness of slow sand filtration and the role of operating conditions looked at the effect of cyst removal on hydraulic loading rate, temperature, cyst concentration, age of schmutzdecke, and age of filter bed (biological maturity). The hydraulic loading rates included 0.04 m/h, 0.12 m/h and 0.40 m/h, the raw water cyst concentrations ranged from 50 to 5,075 cysts/litre, and the raw water temperatures were 15°C and 5°C. Conclusions were that the cyst removal can be expected to be greater than 98% when the filter is establishing the biopopulation and is virtually complete once established. Within the limitations of the research, temperature, cyst concentration and hydraulic loading rate did not affect cyst removal. The most important variable in the performance of the filter media was established to be the biopopulation. Logsdon (6) reported coliform removal of 99% for the months of March to November and 94% to 97% for November to January at a fullscale slow sand filter. The applied cyst removal was 99.9% at a variety of operating conditions but the author noted that at low temperatures Giardia cyst removal efficiency may decrease. The United states Public Health Service reported virus removals at 22 to 96%. The Metropolitan Board of Health in London, England, found reduction of 99.9% in viruses through a laboratory size slow sand filter at a temperature of 1112°C and 99.8% at a temperature of 6°C (7). McConnel (8) demonstrated excellent removal of reovirus through slow sand filtration even at filter startup in the absence of filter maturation. E. coli removal was 99.97% or greater at an average loading of 7.5 × 107 E. coli per pilot filter, but when the loading was increased by 3 log units breakthrough occurred. Biological activity was present through the entire filter depth. Treatment Plant Description The water treatment plant, with a nominal and peak capacity of 3.63 ML/d and 7.26 ML/d respectively, includes a surface water intake, raw water pump station, 3 slow sand filters, chlorination equipment and chlorine contact tank, a clearwell, treated water pumps and a control building. The 3 slow sand filters are each 43 m long by 6 m wide for a filter area of 258 m2. The total filter area is 774 m2. The nominal hydraulic loading rate is 0.19 m3/m2/h with a peak hydraulic loading rate of 0.40 m3/m2/h. The filter media has a depth of 1,050 mm, an effective size of 0.20 to 0.30 mm and coefficient of uniformity of 3.30 to 3.80. The filters are covered to minimize the impact of low winter temperatures ( 20 to 40°C for 24 weeks) and to prevent algae growth in the filter headwaters. Total construction cost was $780,000, $0.11/L at peak capacity. The tenders for construction of the water treatment plant closed May 29, 1984, the contract awarded June 4, 1984 and the substantial completion was December 11, 1984. Due to problems with delivery of media the plant startup using only two filters was November 1, 1985 with treated water delivery to the village distribution system starting November 15, 1985. The third slow sand filter was put into production May 29, 1986. Three pilot slow sand filters were constructed of 300 mm diameter Class 200 PVC pipe and included a wrapping of 50 mm thick foamglass cellular glass insulation to minimize the water temperature change through the filter. The vertical dimensions and the media and gravel support of each pilot filter were identical to the village's operating filters. A 25 mm diameter PVC pilot header pipe was connected to the plant's raw water header and to each pilot filter. From the base of each filter was a 17 mm PVC pipe
Page 89
to a treatment plant drain. On the pilot discharge piping was a Gilmont rotometer (10 1,000 mL/min) to regulate the pilot plant treatment rate and a filter housing to hold the sampling cartridge. Methods Water Quality The raw and filtered water samples were from the respective treatment plant pipe headers. Giardia sampling was accomplished using one micron pore size, polypropylene cartridge filter housed in clear plexiglass Cuno Filter Model IMI4459101. From 3,600 to 4,500 L was sampled over a 3 to 5 hour period for a sampling rate of 15 to 20 L/m. The cartridge filters were wrapped in plastic bags, placed on ice in an ice cooler, transported via overnight courier to Department of Pathology, Colorado State University, Fort Collins, Colorado, U.S.A. The elapsed time between completion of sampling at 100 Mile House and arrival at Colorado State University was typically 3660 hours. Analysis of the grab samples for turbidity was done using a HACH 2100A turbidimeter. Water samples for total coliform (Multiple Tube Fermentation Technique) and standard plate count were taken as grab samples in sterile 250 mL plastic bottles with the results reported as MPN/100 mL and colonies/mL for total coliform and total plate count, respectively. Water samples were analyzed for solids (total, suspended, dissolved), nitrogen (ammonia, nitrate, TKN), phosphorus (total, ortho), total organic carbon (TOC), alkalinity, hardness and pH, with procedures as per Standard Methods (9). These water samples were taken as grab samples in sterile 1 L plastic bottles with duplicate samples taken at each location. One of the duplicate samples was preserved with sulfuric acid. The samples were placed on ice and analyzed by CanTest Ltd., Vancouver. Filter Media Filter media samples at 0 to 5 mm, 300 mm and 600 mm were taken from filter 2 immediately prior to scraping. Each sample, weighing about 0.5 kg (1 lb), was removed by shovel, placed in a plastic bag, stored on ice in an ice cooler and transported via overnight courier to EVS Consultants Ltd. of North Vancouver. Sand samples for invertebrates were preserved in isopropanol containing phloxineB (a histological stain used to facilitate sorting), returned to the laboratory, washed through a 250 µm mesh sieve, and sorted into major taxa using a dissecting microscope. Sand and water samples for phytoplankton analysis were preserved with isopropanol and phloxineB or Lugol's iodine, respectively. Utermohl chamber counts of 50 mL subsamples were undertaken with 20 field counted for each subsample as per Standard Methods (9). The activity of heterotrophic bacteria at various depths in the sand filter was assessed by a modification of the resazurin reduction technique (9) to measure microbial dehydrogenase enzymes. A wet weight sample (100 g) of sand was incubated in 100 mL of plate count broth (PCB) supplemented with resazurin. The activity of heterotrophs (bacteria utilizing organic matter as a source of energy) was assessed by incubating the sandbroth mixture at 20°C and monitoring the reduction of resazurin at 610 nm using a Nova Spec spectrophotometer. Duplicate control samples contained 5 mg/L HgCl as a bacteriocide to indicate the contribution to resazurin reduction by chemical constituents. The difference between total and chemical reduction tests was expressed as µmole/g/hour of resazurin reduction to provide a relative estimate of biological activity among the different samples. Results Temperature A temperature profile is shown in Figure 1. From November 1985 to February 28, 1986 the water temperature in the raw water was between 1°C and 3°C. From March 1, 1986 to October 31, 1986 the water temperature in the raw water ranged from 3°C to a peak of 19°C in August and decreased to a minimum of 3°C. Turbidity A turbidity profile is shown on Figure 2. For the one year period November 15, 1985 to November 15, 1986 the maximum raw water turbidity value was 1.8 NTU in May 1986 while the minimum value was 0.5 NTU in December 1985. The raw water turbidity was reduced 20 to 50% between November 1985 and April 1986 and was reduced 40 to 80% in the period May 1986 to November 1986. The filtered water turbidity ranged between 0.15 NTU and 0.8 NTU and the profile with time tracked the raw water profile with time. Giardia Cysts and Other Microorganisms Between November 1985 and November 1986 the raw water to the treatment plant was sampled 21 times. In 52% of the samples taken (11 samples), cysts were detected in numbers that varied from 14 to > 300 cysts (Figure 3). In most cases the cysts were identified as "excellent and probably infectious". The presence of the cysts varied not only throughout the year but also between two samples taken 1624 hours apart. For example, on August 26, 1986 a sample taken between 0900 and 1329 hours found 110 cysts but no cysts were detected in a sample taken the next day between 0850 and 1400 hours.
Figure 1. Air and filtered water temperatures.
Figure 2. Average raw and filtered water turbidity, = raw water, + = filtered water.
Page 90
No cysts were detected in the 22 samples of treated water from the slow sand filters. This was even true of samples taken in November and December when the filters had been in operation a total of only 21 and 40 days, respectively. In November over 300 cysts were detected in the raw water while in December 14 cysts were detected. At the time the raw water temperature was 1 to 2°C and at this water temperature it is unlikely any biological activity was present in the media. Pilot filter B was spiked with Giardia cysts and coliform on November 27, 1985 and March 20, 1986. The pilot filter had been in continuous operation for 12 days at the time of the spiking in November and 125 days at the time of the spiking in March. The water temperature to November was continuously at 1°C while it varied between 1°C and 4°C from November 27, 1985 to March 20, 1986. On November 27, 1985 an estimated 2,000,000 cysts were supplied but only 675,000 cysts were detected in two1 litre grab samples from the filter headwater. It is unknown if there was a dieoff of cysts between Fort Collins and 100 Mile House that resulted in a smaller amount recovered, or if the cyst distribution was not equal throughout the headwater. Ten cysts were recovered in the filtered water for a removal efficiency of at least 99.99%. It is unknown if cysts moved through the media after completion of the 31 hours of sampling. On March 29, 1986 the estimated number of cysts supplied was 4,400,000. The two 1litre filter headwater grab samples were spoiled in transit. There were no cysts detected in the filtered water. The variation in algae levels is identified in Figure 4. The algae level in the raw water peaked in the July and August period when the total count on two occasions was 59 and 460 cells/mL. There were generally low algae levels in the filtered water with the count on two occasions being 21, 22 and 43 cells/mL. The dominant genera of algae in the raw water in March 1986 were Synedra sp., Tabellaria sp. and Fragellaria sp., while Fragellaria crotonensis, F. virescens and Achnanthes sp. were found in the June 1986 raw water sample. For the raw water
Figure 3. Number of Giardia cysts in raw water to village's water treatment plant
Figure 4. Relative algae levels in raw and filtered water
sample taken in August 1986 the dominant genera were Cocconeis placentula and Achnanthes minutissima. Synedra and Ceratium contribute to taste and odour. Dinobryon, Synedra, and Fragilaria contribute to filter clogging. The village has had comments from users of the system that the taste associated with the raw water has disappeared with startup of the water treatment plant. It is suspected that the removal of Synedra and Ceratium algae was the reason. Coccidia is a mammalian parasite not infectious to man. Coccidia were detected in the raw water fourteen times out of 21 samples taken. The type was from beaver, rodent or mammal. The occurrence was throughout the year, and whenever Giardia cysts were detected Coccidia was also detected. The Coccidia was not detected in the filtered water. Plant debris in the context of this research refers to rodent (beaver, muskrat, mouse) fecal debris. Plant debris was detected in every sample of raw water taken in amounts that varied between occasional to a small amount, but was in moderate amounts in one sample. Plant debris was not detected in the filtered water except in a rare amount in November 1985 and an extremely rare amount on January 20, 1986 when the filters were still in the ripening stage. Complete removals were also found for crustaceans/eggs, pollen, ciliates and flagellates. General Water Quality Parameters A major purification mechanism in the slow sand filter is biological oxidation of organic matter. The nitrogen, phosphorus and total organic carbon concentrations taken on five occasions are presented on Table 1. The main biodegradable substances, ammonium and organic matter, have the potential to be oxidized by bacteria in the sand media. Organic removal is achieved with a high growth rate of heterotrophic bacteria while nitrification is achieved by autotrophic bacteria. Important parameters include media size, oxygen levels, solids retention time and shear stress. Biodegradation of organic matter as measured by the reduction of TOC seemed to have occurred in June but not September while nitrification of the ammonia occurred in June but not the March or November/85 samples.
Page 91 TABLE 1. Nitrogen and phosphorus concentrations (mg/L)
Ammonia
Nitrate
TKN
Total
Ortho
Nov. 21/85
Raw
.014
.018
.32
< .02
< .02
< 1.0
Filtered
.014
.016
.34
< .02
< .02
< 1.0
March 21/86
Raw
.029
< .01
.44
.10
.031
< 1.0
Filtered
.040
< .01
.46
.10
.044
< 1.0
June 6/86
Raw
.026
.019
.31
.10
< .02
14
Filtered
< .01,
.010
.35,
.075,
< .02,
4.6,
< .01
.031
.32
.068
< .02
5.8
0.19
.036
.27
.035
< .01
4.8
Sept. 29,30/86
Raw
< .01
< .01
.14
.12
< .02
3.5,3.0
Filtered
< .01
< .01
.089
.14
< .01
4.0,3.0
Nov. 4/86
Raw
< .01
.044
.21
.021
< .02
Filtered
< .01
.033
.12
< .02
< .02
Nitrogen
Phosphorus
TOC
TABLE 2. Summary of coliform and standard plate count
OPERATING FILTERS
Coliform (MPN/100mL)
Standard Plate Count (MPN/100mL)
Raw
Filtered
Raw
Filtered
13, 7.8
2.0
2400, < 1.0
30,000
January 21/86
< 2.0
49, 22
March 20/86
< 2.0, 33
< 2.0
1, 110, 67
40
April 28/86
< 2.9, 240
< 2.0, < 2.0
750, 10,000
100, 320
< 2.0
< 2.0, < 2.0
500
70, 30, 100
November 15/85
June 5/86
September 29/86
< 2.0, < 2.0 < 2.0, < 2.0
< 2.0, < 2.0
16,000 < 1, 5200
5, < 1
Coliform A summary of the total coliform bacteria and standard plate count is shown on Table 2. The total coliform was typically below the detectable limit in the filtered water except in November/85 and January/86 samples. In the latter cases media biological activity was likely minimal due to low water temperatures and low organic matter. The standard plate count in the filtered water was also very high in the November/85 sample but was generally reduced over the raw water level except on one sample taken in June/86. The schmutzdecke was only about 12 hours old at the time of sampling. Pilot Filter B was spiked with total coliform on November 27 and March 20, 1986 with water samples taken approximately every 1 to 2 hours for up to 36 hours after spiking. The water samples were analysed for total coliform. On November 22, 1986 the number of coliform equivalent units supplied was 1 × 108 or about 75,100 MPN/100 mL. Table 3 is a summary of the coliform count in the raw and filtered water samples with time. There is no explanation for the results presented. Not only was the peak coliform count in the raw water detected 11 hours after spiking but the peak treated water coliform counts occurred at 6, 13 and 30 hours after spiking and were higher than the raw water sample. On March 20, 1986 the number of coliform equivalent units supplied was 1 × 108 or 75,000 MPN/100 mL. Table 4 is a summary of the total coliform count in the raw and treated water samples with time. The peak raw water coliform count was 1,600,000 MPN/100 mL in the interval 0 to 1:15 h after start of spiking, and 9,200 MPN/100 mL in the water sample taken 4:45 hours after spiking. The reduction was in the order of about 99.5%. Filter Media On April 3, June 4 and August 21, 1986 samples were taken of the schmutzdecke and of the filter media at 300 mm and 600 mm depths. On the latter two dates samples were taken immediately prior to filter scraping. The analyses included bacterial activity and enumeration of phytoplankton and zooplankton. The heterotrophic bacterial activity of sand samples from three depths are presented on Table 5. The highest levels of bacterial activity were at the filter surface or the schmutzdecke. On June 3, 1986 and August 21, 1986 the schmutzdeckes were about to be removed as the filter surface was sealed off. The bacterial activity was low in the March sample probably due to the continuous low raw water temperature of 1 to 3°C since filter startup. Bacterial activity increased in the June and August samples, likely reflecting a higher water temperature and more favourable climate for bacterial activity.
Page 92 TABLE 3. Coliform spiking November 27, 1986 Elapsed Time From Start of Spiking (h:mins)
Total Coliform Count (MPN/100 mL) Raw Water
Filtered Water
0
49,000; 79,000
0:40
23
1:40
540,000
170
3:00
13,000
540,000
6:00
3,500
1,600,000
8:00
13,000
540,000
11:00
1,600,000
49,000
13:00
23,000
1,600,000
15:00
110,000
920,000
18:00
23,000
540,000
19:30
3,300
350,000
30:00
1,600,000
30:30
79,000
2,300
TABLE 4. Coliform spiking March 20, 1986 Elapsed Time From Start of Spiking (h:mins)
Total Coliform Count (MPN/100 mL) Raw Water
Filtered Water
0
3,300
4.5
1:15
1,600,000
2.0
2:45
350,000
3,500
4:45
350,000
9,200
8:45
17,000
2,400
11:45
11,000
540
14:15
9,200
540
17:15
3,500
920
23:55
17,000
540
25:55
16,000
540
34:10
79,000
920
TABLE 5. Heterotrophic bacterial activity filter media Rate of Resazurin Reduction (µmole/g/hour) Depth (mm)
March 17
June 3
August 21
17
0
0.32
0.060
300
0.12
0.016
600
0
0.016
TABLE 6. Summary of invertebrates Total Number of Specimens per 400g of Sample Depth (mm)
Total Number of Taxa per 400g of Sample
March
June
August
March
June
August
17
13
61
2264
2
8
6
300
2
852
2
5
600
0
337
0
2
A summary of invertebrate numbers found in the media samples is summarized on Table 6. Counts of invertebrates within the sand filter demonstrated an increasing trend from March to August at all depths, with the majority of the organisms being located in the top layer. In March, organisms were found only in the top layer. They were few in number and were likely entrapped from the incoming water. Over time, colonization by organisms more suited to the environment occurred, where in June the taxa Chironomidae (midge fly) predominated. In August the taxa Oligochaeta (worm) was dominant. Oligochaetes are true aquatic animals which feed on bacteria, and their increasing presence reflects the maturation of the biological ecosystem within the filter (i.e. the development of indigenous species) and indicates an effective filtration process. Summary and Conclusions The removal of Giardia cysts through the slow sand filter was essentially complete even in the absence of schumutzdecke. A similar removal was found over fifteen categories of other microorganisms. When the pilot filter was spiked with 5,000 cysts/L, an extremely high concentration compared to cysts detected in the raw water, the removal efficiency was 99.99%. The removal was in spite of the fact that the heterotrophic bacterial activity and invertebrate population were not measurable until the summer period. This was due to the cold water temperatures and the small amount of biodegradable organic matter in the raw water during the winter of 1985 and spring of 1986. A mature biological ecosystem was detected only in the August sampling period when the water temperature peaked at 19°C and the algae was heavy. The increase in heterotrophic bacterial activity was also likely responsible for the reduction in the ammonia and TOC concentrations in the summer period and the absence of same in the winter of 1985 and spring of 1986. The impact of the slower forming nitrates was noted in the summer of 1986 when the nitrate levels in the filtered water were higher than in the raw water. The TKN was also observed to decrease between raw and filtered water but only in the fall of 1986. Total phosphorus was reduced in the summer of 1986, possibly because of an increase in the activity of heterotrophs. The reduction in turbidity was about 2050% between November/85 and April/86 while it was 4080% in the period May/86 to November/86. The filtered water, with
Page 93
one exception, was less than the B.C. Ministry of Health objective of 1 NTU. The increased efficiency in the summer and fall of 1986 was likely due to the effectiveness of the schmutzdecke in trapping and holding the particles. This activity was negligible in the winter of 1985 and spring of 1986. The total coliform and standard plate count reduction increased with the increase in the heterotrophic activity and invertebrate population. On one occasion in the summer of 1986 the filtered water standard plate count was high. This was some 12 hours after scraping when due to village water demand, the filter was operating at a high filter rate. The trend to higher removal efficiency with increasing filter age and biological activity was noted with the spiked pilot filters. Acknowledgements This research was completely funded by Health & Welfare Canada through Supply and Services Canada. The authors thank Dr. R.S. Tobin of the Department of National Health and Welfare Canada, Ottawa. Thanks also to the village council and staff, especially Mr. G. Mills and Mr. R. Hume. Literature Cited 1. Bellamy, W.D., Silverman, G.P., Hendricks, D.W. and G.S. Logsdon. 1985. Removing Giardia cysts with slow sand filtration. J. AWWA 77(1):52. 2. Bellamy, W.D., Silverman, G.P., and D.W. Hendricks. 1984. Filtration of Giardia cysts and other substances. Volume 2: Slow sand filtration. Colorado State University Environmental Engineering Technical Report 5847843 for U.S. EPA under contract CR880865002. 3. Bellamy, W.D., Silverman, G.P. and D.W. Hendricks. 1983. Giardia lamblia removal by slow sand filtration. In: Proceedings of Sunday Seminar on Innovative Filtration Techniques. American Water Works Association Annual Conference. Las Vegas, Nevada. 4. Huisman L. and W.E. Wood. 1974. Slow sand filtration. World Health Organization, Geneva, Switzerland. 122 p.. 5. Liu, D. and W.M.J. Strachan. 1980. Characterization of microbial activity in sediment by resasurin reduction. Arch. Hydrobiol. Berh. 12:2431. 6. Logsdon, G.S., Hendricks, D.W. and G.R. Pyper. 1983. Control of Giardia cysts by filtration. The laboratory's role. In: Proceedings of American Water Works Association Water Quality Technology Conference, Norfolk, Virginia. AWWA. 7. Logsdon, G.S. and E.C. Lippy. 1983. The role of filtration in preventing waterborne disease. J. AWWA 74(12):649. 8. McConnel, L.K., Sims, R.C. and B.B. Barnett. 1984. Reovirus removal and inactivation by slow rate sand filtration. J. Appl. Env. Microbiol. 48(4):818. 9. Standard Methods for the Examination of Water and Wastewater. 1985. 16th Edition, APHA, AWWA and WPCF, Washington, D.C..
Page 95
Comparison of Some Filtration Processes Appropriate for Giardia Cyst Removal Gary S. Logsdon Drinking Water Research Division, Water Engineering Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio 45268, U.S.A.. Slow sand filtration, diatomaceous earth (DE) filtration, and coagulationfiltration (including conventional treatment, direct filtration and inline filtration), have been evaluated for Giardia cyst removal at pilot plant and/or field scale. Properly designed and operated, the above processes can attain 99% cyst reductions, or higher. This paper discusses relative advantages and disadvantages of the processes, and factors that may result in success or failure of treatment. Slow sand filtration is the least complicated from the operator's perspective. It may be the most appropriate for small systems if the raw water is treatable. It very effectively removes viruses, bacteria and cysts; but is not very effective for removal of THM precursor or organic chemicals. It gives the operator the least ability to change treatment in response to changes in raw water. DE filtration is very effective for cyst removal, but removal of very small particles requires use of fine grades of DE or chemical preconditioning of DE. Process modifications can yield iron and manganese removal. THM precursor removal is small. Operator skills required are mostly of a mechanical nature. Coagulationfiltration has the greatest flexibility, and can remove 30 to 50% of THM precursor; also turbidity, microorganisms, and metals that can be precipitated before filtration. Many factors influence process performance so a good understanding of coagulation chemistry is needed for most effective operation regardless of plant size. This requires the greatest level of operator ability for continued dependable performance. Process variations include conventional treatment, direct filtration and inline filtration.
Introduction Waterborne giardiasis outbreaks have been occurring in the USA for the past two decades, and continue to occur. This suggests a need for better water treatment. Disinfection provides a barrier for waterborne transmission of Giardia cysts. Craun (8) reported that 19,770 cases of waterborne giardiasis were related to deficiencies in treatment of surface water sources by community water systems from 1965 through 1984. Of these, 61% were related to failures to adequately disinfect in systems having disinfection as the only treatment. Another barrier is effective filtration. This paper reviews filtration studies at pilot scale or full scale, or both, and compares performance capabilities and advantages of slow sand filtration, diatomaceous earth (DE) filtration, and coagulationfiltration. The latter category includes conventional filtration (coagulant feed and rapid mix, flocculation, sedimentation, and filtration), direct filtration (coagulant feed and rapid mix, flocculation, and filtration), and inline filtration (coagulant feed and rapid mix, followed by filtration). All of the above filtration processes, if they are properly designed and operated, and if they are treating a source water of suitable quality, can reduce the concentration of Giardia cysts by 99% or more. Filtration failures can occur because of improper design or operation, or because a given process is not appropriate for the raw water being treated. Of the 19,770 cases of giardiasis mentioned above, 38% occurred because of failures in filtration. Aspects of filter plant design and operation are discussed in subsequent sections of this paper. The relative costs of the processes are not discussed because these would be influenced by conditions that are site specific; thus general comparisons would be of limited usefulness. Slow Sand Filtration Slow sand filtration studies have been supported in recent years by the U.S. Environmental Protection Agency, the American Water Works Association Research Foundation, and the State of Utah, among others. Some parameters in EPA funded studies are given in Table 1. Filters have been evaluated for ability to remove Giardia cysts, bacteria, turbidity, particles, and trihalomethane (THM) precursor. Slow sand filters have been shown capable of removing 99 to 99.99% of the Giardia cysts in raw water (3,4,21). Using pilot filters Bellamy et al. (3) found that cyst removal did not deteriorate after filter scraping. Pyper (21) observed that at 7.5°C to 21°C, cyst removal was 99.98% to 99.99%. At 0.5°C to 0.75°C, removal ranged from 99.36% to 99.91%; however, at 0.5°C, cyst removal deteriorated to 93.7% when both Giardia cysts and primary unchlorinated sewage effluent were added to the raw water simultaneously. In this situation, the loading of organisms in the influent water may have been greater than the established biological population of the slow sand filter could cope with. Total coliform removal was found to be adversely influenced by increases in filtration rate from 0.04 to 0.4 m/h (3), by decreases in filter bed depth from 0.97 m to
Page 96 TABLE 1. Parameters in slow sand filter research Filter Design Sand Size mm
Raw Water Quality
Reference Other
Uniformity Coefficient
Filtration Rate m/h
Bed Depth m
Temperature °C
Turbidity NTU
Total Coliform per 100 mL
0.17
2.1
0.12
0.76
~ 25 room temp.
< 1 10
10 10,000
(11)
0.32
1.4
0.12
0.94
2 28
< 1 to > 30
40 10,000
0.2 143 mg/m3 chlorophylla
(7)
0.33
2.8
0.08
1.07
0 25
0.2 59
1 to 8,700
(2.1 26) × 106 Giardia cysts spiked [35 425 cysts/L if diluted over filter uniformly]
(21)
0.13 to 0.62
1.5 to 1.6
0.04 to 0.40
0.48 to 0.97
2 17
2.7 11
0 209,000
50 5,075 Giardia cysts/L spiked
(3,4)
0.48 m (4), by increases in sand size from 0.13 mm to 0.61 mm (4), and by decreases in temperature from 17°C to 2°C (4). Of these parameters, the 0.61 mm sand size would be greater than sizes typically used and might have accentuated the adverse impact of that variable. The use of 0.61 mm sand resulted in average total coliform removal of 96% vs. 99.4% for 0.13 mm sand. Temperature decreases from 17°C to 5°C or 2°C resulted in deterioration in coliform removal from the 99% level to about 90% for the colder waters. Cleasby et al. (7) found that total coliform removal was lower during the first two days after scraping than during the remainder of the run. In some instances, differences in the two time periods were slight, but 5 of 9 runs exhibited coliform removals ranging from 82% to 95% during the first two days. During the remainder of the runs, removals ranged from 97% to 100%. Letterman and Cullen (14) in most cases did not observe any effects of scraping (a ripening period) in total coliform data collected in a study of seven operating slow sand filter plants in the State of New York. Virus removal has been reported (Taylor, No Date) to be influenced by temperature and filtration rate. At 0.20 m/h and 11 to 12°C, removal was 99.9999% vs. 99.8% for 0.40 m/h and 6°C. In another set of experiments, Taylor reported 99.8% removal at 0.20 m/h but only 91% at 0.40 m/h. Researchers have observed variation in the ability of slow sand filters to reduce turbidity to the 1 Nephelometric Turbidity Unit (NTU) Maximum Contaminant Level (MCL) specified in the U.S. Environmental Protection Agency's Drinking Water Regulations. Fox et al. (11) found that when water from a gravel pit in southwestern Ohio was filtered at 0.12 m/h, after an initial ripening period had allowed the biopopulation to become established on new sand, the 1 NTU MCL was always met. Raw water turbidity ranged from 0.2 to 10 NTU. Cleasby et al. (6) reported that after the first two runs, typical effluent turbidity was 0.1 NTU except during the first two days after scraping. Water for that research came from a gravel pit in central Iowa, with turbidity ranging from < 1 to 30 NTU. Pyper observed slow sand filtered water turbidity of 0.1 NTU or less for 50% of the time, and 1.0 NTU or less for 99% of the time in McIndoe Falls, Vt. The source of water was Coburn Pond, a body of water with an open water surface area of about 4 ha, plus about 20 ha of wetland. Raw water turbidity ranged from 0.4 to 4.6 NTU and color averaged 24 C.U. In contrast with these results, when Horsetooth Reservoir was treated (3,4), the filtered water turbidity ranged from 3 NTU to 5 NTU, and the 1 NTU MCL was not met. Raw water turbidity of Horsetooth Reservoir generally was 6 NTU to 8 NTU. Slezak and Sims (23) reported that about 15% of 27 plants surveyed produced filtered water with an average turbidity of 1.0 NTU or higher, whereas turbidity averaged 0.4 NTU or lower at half of the plants. The different degrees of turbidity reduction in some cases may be attributed to the nutrient condition of the filters. Water collected high in the Rocky Mountains and transported to Horsetooth Reservoir would not be expected to be high in nutrients for growth of biopopulation in filters. Bellamy et al. (4) reported adding sterile nutrient (BOD about 4 mg/L) to one test filter, which should have increased the biopopulation in the filter. Under parallel operation, turbidity reduction averaged 52% from this filter vs. 15% from the filter treating unaltered Horsetooth Reservoir water. Pavoni et al. (20) reported that exocellular polymers produced by bacteria in an activated sludge culture were capable of flocculating kaolinite suspensions and promoting settling. It appears possible that the biological population of a slow sand filter may produce exocellular polymers that enhance the ''stickiness" of filter media and inorganic particles in the slow sand filter, thus improving the filter's capability to remove such particles. The surface waters tested in Iowa and Ohio contained sufficient nutrient to support algae during the summer, and the water in Vermont would be expected to be high in
Page 97
nutrients resulting from decaying vegetation in the wetlands. Thus, we would infer that those waters had higher nutrient levels than the Horsetooth Reservoir water. Slow sand filters should not be expected to remove large amounts of THM precursor unless something has been done to chemically alter the precursor before filtration. Humic materials, although in contact with microorganisms in nature, seem to persist in the environment. The biopopulation in the Vermont slow sand filter removed about 10% of the trihalomethane formation potential (THMFP) that was between 100 µg/L and 200 µg/L in raw water. Fox et al. (11) reported TOC removal of 19% and THMFP removal of 18% when treating southwestern Ohio gravel pit water. In the research at Iowa State (7), algae were encountered and evaluated for removal and influence on filter efficiency. Chlorophylla measurements were less than 5 µg/L during the winter and spring of 19811982, until midApril, increasing to nearly 60 µg/L in late April. Chlorophylla declined in May and June but appeared to peak near 140 µg/L in July. Algal blooms occurred, and these influenced run length. Four runs ranged in length from 10 to 22 days when mean chlorophylla values were 8 to 138 µg/L. Runs of 34 to 123 days were associated with chlorophylla values of 1 to 4 µg/L. Algae removal, as measured by chlorophylla reductions, was quite high and similar to removal of other particulate matter (generally approaching 99%). Raw water quality limits for slow sand filters are stringent because particulate matter tends to be removed at the top of the filter and because slow sand filters have limited capability to remove inorganic contaminants and synthetic organic chemicals. Cleasby et al. (6) reported that enumeration of algae or performing a surrogate measure of algal population was necessary to judge the suitability of raw water for slow sand filtration. Fox et al. (11) reported that treatment of Ohio River water (0.423 NTU) resulted in progressively poorer filtered water quality over 250 days of operation, with effluent turbidity exceeding 1 NTU during the last 20 days of operation, and time to terminal head loss (0.4 m) decreasing from 98 days to 6 days. During the first 230 days, mean influent turbidity ranged from 2.4 to 7.6 NTU, levels that do not seem excessively high. Average raw water turbidity was 10 NTU or lower at 90% of the operating plants surveyed by Slezak and Sims (1984). Experience thus far suggests that the most reliable way to determine treatability of water by slow sand filtration is to conduct an extended pilot plant study. Slow sand filters are simple to operate and maintain, when raw water quality is appropriate and when the plants are small enough that complicated equipment is not needed for filter scraping. Daily duties at a small installation (10,000 to 1,000,000 L/day) would include reading and recording head loss, flow rates or totals, chlorine residual, raw and filtered water turbidity, and adjusting flow. Letterman and Cullen (14) studied filter scraping at seven slow sand filtration plants in New York. Average flows ranged from 1 to 23 million L/day. Scraping, or removal of a thin layer of sand when terminal head loss is reached, required an average of 5 hours per 100 m2 of filter surface. The thickness of the layer removed was typically 2 to 3 cm. The frequency of scraping would be determined by run length, which would be influenced by the turbidity and algae in the raw water. After a sand filter has been scraped a number of times, the full depth of the bed is restored in an operation called resanding. Letterman and Cullen estimated that resanding a depth of 15 to 30 cm would require 4859 hours of labor per 100 m2. The advantages of slow sand filters are related mainly to the simplicity inherent in the process. Small plants are simply to construct. Simple, manually controlled valves can serve to control flow. Head loss can be measured by a piezometer. Because changes in head loss occur slowly, recording equipment is not needed. Coagulant chemicals are not used in slow sand filtration, so operators do not need to understand coagulation chemistry. Chemical feed pumps would not be needed for coagulant chemicals, so fewer pumps would be used, lowering mechanical maintenance work. Operator skills do not need to be as high as for plants using coagulation. Another advantage associated with absence of coagulation is a minimum of waste disposal problems. Scraped sand is essentially the only waste, and often it is washed and reused. Many of the disadvantages of slow sand filtration are also related to the absence of coagulation. Without pretreatment, limitations exist on the quality of water that is suitable for slow sand filtration. These were explained earlier. Because modifying a slow sand filter plant to treat a difficult water might be costly, or not possible, pilot studies should be performed to verify treatability. In addition, a study should be conducted to establish that the raw water source is not likely to change or deteriorate in quality to such a degree that the water would become untreatable in the future. This may not always be possible to ascertain, but an effort should be made to predict what sort of human activities or development might happen in the foreseeable future. This would at least alert authorities to possible need for changes in treatment if raw water quality deteriorated. Because pretreatment is minimal or nonexistent at slow sand filter plants, little capability generally exists to remove synthetic organic chemicals, trihalomethane precursors, and dissolved inorganic substances such as heavy metals. In addition, very fine clays or glacial flour may not be readily removed. Finally, slow sand filters may not be appropriate for medium to large installations in the USA, because of operating labor costs and land costs. The trend for large systems is to automate and use mechanical equipment where possible, but cleaning enclosed slow sand filters by mechanical means is very difficult. Thus, they seem most appropriate for small systems located on very high quality source waters. Diatomaceous Earth Filtration Diatomaceous earth (DE) filters have been studied for removal of a variety of contaminants. They have been shown to attain excellent removal of Giardia cysts over a broad range of operating conditions. Cyst removals
Page 98
exceeding 99%, and often 99.9%, were reported by Lange et al. (12) for filtration rates of 2.4 to 9.6 m/h, for temperatures from 3.5 to 15°C, and for four different grades of diatomaceous earth (Celite 545T M, Celite 535T M, Celite 503T M, and Hyflo SuperCelT M). Pyper (21) reported 99.97% for one DE filter run in which Giardia cysts were added. Logsdon et al. (15) reported that when sufficient DE precoat and body feed were used, removal of 9 µm radioactive beads was nearly always 99.9% or higher. Use of a precoat of at least 1.0 kg/m2 was shown to be appropriate for obtaining most effective removal of the 9 um particles. They also reported that eleven filter runs were made with G. muris cysts at filtration rates of 2.2 to 3.5 m/h, with Celite 535T M precoat and body feed. Cyst removal exceeded 99.0% in all runs, and exceeded 99.9% in five of the runs. DeWalle et al. (9) reported on four DE filter runs conducted for Giardia cyst removal. Cyst removal exceeded 99% in each of the four runs. The overall results of all research for Giardia cyst removal indicate that DE filtration is very effective for controlling Giardia cysts. Factors important to continued effective performance are using adequate precoat and body feed, and keeping the septum very clean (good cleaning at the end of each run). Removal of total coliform bacteria by DE filtration was studied extensively at Colorado State University by Lange et al. (12). Coliform removals were strongly influenced by the grade of diatomaceous earth used. Coarser grades attained removals ranging from 30% to 50% for Celite 545T M and from 50% to 70% with Celite 503T M. The fine grades, with smaller pores, were considerably more effective. Removal with Celite 512T M was 92% to 96%, and total coliform removal with Super CelT M was 99.92% to greater than 99.98%. Malina et al. (18) reported that a high percentage of removal could be attained for poliovirus when coated DE filter aid was used or when cationic polymer was added to the raw water. In one 12hour filter run, diatomaceous earth coated with 1 mg of cationic polymer per gram of DE produced filtered water in which no viruses were recovered from 11 samples (removal > 99.95%). One of 12 samples was positive, and in this instance, virus removal was 99%. In a 12hour run in which uncoated DE was used and 0.14 mg/L of cationic polymer was added to the raw water, no viruses were recovered from any of the 12 samples analyzed. Turbidity removal when treating Horsetooth Reservoir water, as reported by Lange et al. (12), was less than 20% for the grades of diatomaceous earth commonly used for water treatment (Celite 545T M, Celite 535T M, Celite 503T M, and Hyflo SuperCelT M). Turbidity of the Horsetooth Reservoir raw water ranged from 4.5 to 5.4 NTU. The finest grade tested, FilterCelT M, could reduce the turbidity by over 95%. In contrast to these results, Logsdon et al. (15) reported that turbidity reductions of 56% to 78% were attained with Celite 535T M when raw water turbidity ranged from 0.95 to 2.5 NTU, but little change was observed when raw water turbidity ranged from 0.24 to 0.45 NTU. Pyper (21) reported an average turbidity reduction of 71%, with an effluent quality of 0.5 NTU. Pyper evaluated DE filtration for removal of THM precursor in Vermont. Results showed no difference between the raw water and the filtered water, suggesting that the THM precursor material present in Coburn Pond was dissolved. The water was colored (24 CU, average), and this may explain the lack of change during filtration, because DE filtration alone does not remove color, a known precursor. Because turbidity removal with the grades of diatomaceous earth commonly used for water treatment was so low when Horsetooth Reservoir water was filtered, the Colorado State University researchers investigated the nature of the turbidity (2). When 5.6 NTU raw water was filtered, turbidity was reduced 2% by a 5 µm pore size membrane, 36% by a 1.2 µm membrane, 73% by a 0.45 µm membrane, and 91% by a 0.22 µm membrane. Most of the light scattering matter in the water (the cause of the turbidity) was made up of particles that could pass through 1.2 µm pores, and thus fine enough to pass through typical potable water grades of DE. Additional work was done at Colorado State University to improve the capabilities of DE filtration. In order to alter the surface properties of diatomaceous earth, aluminum hydroxide was precipitated to the surface of a DE slurry. With 0.05 grams of alum per gram of Celite 545T M, total coliform removal was 99.86%, as compared to 30% to 50% removal for uncoated Celite 545T M. For the same grade of DE, turbidity removal was 98%, for coated DE vs. under 20% for uncoated DE (12). These results show that the straining mechanism of removal can be augmented by a surface attachment removal mechanism if DE is given an electropositive coating. Limits on the quality of raw water that would be appropriate for DE filtration are not easy to set. The process removes particulate matter by trapping it within the filter cake. As the concentration of particulate matter in raw water increases the load applied to the filter cake increases. To maintain high permeability of the filter cake and good head loss characteristics, body feed diatomaceous earth is added to the raw water. A rule of thumb is that higher raw water particle concentrations require more body feed, if the nature of the particles does not change. The nature of the particles being removed is quite important though, especially the compressibility. Rigid turbiditycausing particles, such as very fine sand, would not block or blind the filter cake, but compressible particles, such as algae, coagulation floc, precipitated iron, or biological matter could blind the filter cake. Pilot filtration studies are advisable if the water in question is not already being treated by DE filtration. Such studies would establish the appropriate grade of DE to use to obtain the desired effluent turbidity, the amount of body feed to add under conditions of the test runs, and the approximate length of filter run to expect. Letterman and Logsdon (13) surveyed 13 DE filtration plants and reported that filtered water turbidities above 1 NTU or filter runs of 6 or fewer hours were observed at DE plants having maximum raw water turbidities of 20 NTU or greater (Figure 1). This figure shows the percentage of plants exceeding specified values for minimum, average, and maximum raw
Page 99
water turbidity. Symbols shown in the legend identify plant problems with high filtered water turbidity or short runs or both. Operation and maintenance of diatomaceous earth filters are somewhat more complex than for slow sand filters, but less complicated than coagulationfiltration. Daily monitoring would include turbidity, disinfection residual, rate of water production with adjustments if needed, filter head loss, and rate of use of body feed. Periodic chores would include preparation of body feed slurry and precoat slurry and maintenance checks on body feed and precoat pumps. Also, filters would need to be backwashed periodically, but disposal of spent filter cake should present few problems, because it is not gelatinous and dries readily. Filter elements (septa) need to be kept very clean. The cleanliness of the septa can be readily checked if vacuum filters or quickopening pressure filters are used. Because of the number of pumps, valves, and other mechanical items in use at a DE filtration plant, operators should possess good mechanical skills. Knowledge of coagulation chemistry would not be needed unless the diatomaceous earth was conditioned by the alum coating technique. Diatomaceous earth filtration has several important advantages, especially with respect to treating waters that may contain Giardia cysts. The process has been shown
Figure 1. Influence of raw water turbidity on diatomaceous earth plant performance
in four studies to be very effective for cyst removal, and the removal efficiency is not affected by very low temperatures. Different grades of diatomaceous earth can be kept on hand, giving the operator some flexibility if the grade in use passes too many turbidity causing particles. If necessary, the surface attachment properties of the coarse grades of diatomaceous earth can be markedly enhanced by the alum coating procedure. Diatomaceous earth filter plants do not require large land area, and are in use for capacities up to 50 or 60 million L/day. Among the disadvantages of diatomaceous earth are the need for high quality raw water, the inability to remove dissolved substances, and the inability to remove very fine particles with plain diatomaceous earth. Excessive suspended solids (turbidity, algae) in raw water can cause short filter runs. Bubbles may form and collapse in the filter cake if the vacuum DE filters are used to treat cold, highly oxygenated water. If pressure DE filters are used and operated to high head loss to obtain long runs and economical use of DE precoat material, high energy costs may result. CoagulationFiltration The process train used most often in the United States for filtration involves chemical pretreatment (coagulation, and frequently flocculation and sedimentation) followed by deep bed granular media filtration. Most U.S. coagulationfiltration research for Giardia cyst removal has focused on the coagulationfiltration (inline) or coagulationflocculation filtration (direct filtration) variations of the process, because waterborne giardiasis outbreaks tended to be observed in regions of the country that had low turbidity waters which were thought to be suitable for such treatment. Research by Logsdon et al. (15), DeWalle et al. (9), and AlAni et al. (1) involved coagulation with alum, or alum plus a polymer; filtration through sand or dual media at 5 to 14 m/h; and temperature ranging from 3 to 20°C. Later research (17) was conducted on conventional treatment, with alum or alum and polymer, dual media and three monomedia types (sand, anthracite, GAC), filtration at 7 m/h, and room temperatures (about 25°C). Results of the three cited direct filtration studies indicate that Giardia cyst removal can exceed 99.0% or even 99.9% when the raw water is coagulated properly and filtered. Results of Logsdon et al. (15) and DeWalle et al. (9) indicated that with proper pretreatment, cyst removal exceeded 99.0% when filtered water turbidity was below 0.30 NTU. AlAni et al. (1) showed that cyst removal of 99% or more was likely to occur if turbidity removal was 70% or more, when raw waters in the 0.2 to 1 NTU range were treated. This would produce filtered waters in the 0.06 to 0.30 NTU range. All of the above researchers showed that dependable cyst removal results can not be attained if a clear water (about 1 NTU) is filtered without being properly coagulated. Use of no coagulant, or of an improper dose, resulted in erratic cyst removal results. In addition, DeWalle et al. (9) showed that for alum coagulation,
Page 100
using the proper pH is necessary when soft, low alkalinity water is treated. They observed effective treatment at pH 5.6 and 6.2, but at pH 6.8 with alum coagulation, cyst removal was reduced from 99% to 95%. The coagulationfiltration process can remove a variety of contaminants. Robeck et al. (22) showed that direct filtration could remove 90% to 99% of viruses, while conventional treatment removals consistently were 99%. McCormick and King (19) stated that coliform removal by direct filtration was practically 100% when filtered water turbidity was 0.10 NTU or less. Cleasby et al. (6) reported that inline filtration removed more than 86% of the total coliform bacteria in raw water, after the first hour of the filter run had passed, in 10 test runs. Edzwald (10) showed that direct filtration could remove nonpurgeable total organic carbon (NPTOC) and organic precursor materials that form trihalomethanes (TTHMFP, or total trihalomethane formation potential). With cationic polymer as the primary coagulant, both NPTOC removal and TTHMFP removal were about 40% whereas with alum as the primary coagulant removals of NPTOC and TTHMFP were nearly 60%. With the same waters, when conventional treatment was employed with alum as the primary coagulant, removals of NPTOC and THMFP were about 70%. Cleasby et al. (6) reported that waters with low to moderate algal populations, water could be treated by direct filtration. Water with few algae had a chlorophylla concentration of less than 5 µg/L (7). Water with an algal population sufficient to result in a chlorophylla concentration of 130 µg/L could not be effectively treated by direct filtration without prechlorination. Suggested limits on raw water quality for sources receiving complete conventional treatment (including predisinfection, coagulation, sedimentation, rapid granular filtration, and post disinfection) were given in the "Manual For Evaluating Public Drinking Water Supplies" as a monthly geometric mean of not more than 2,000 fecal coliform per 100 mL or a monthly geometric mean of not more than 20,000 total coliform bacteria per 100 mL, color not to exceed 75 units, odor not to exceed a threshold odor number of 5, and turbidity not to be so high as to overload the water treatment works (26). Suggested limits on raw water quality for direct filtration and inline filtration are much more stringent. Cleasby et al. (6) suggested that average raw water turbidity should depend on whether the primary coagulant is alum or a cationic polymer, and on whether algal population is low or moderate. Suggested values ranged from 7 NTU for moderate algae and alum coagulation to 16 NTU for low algae and cationic polymer coagulation. The Direct Filtration Subcommittee of the AWWA Filtration Committee (5) reported that waters with less than 40 units of color, turbidity below 5 NTU, iron less than 0.3 mg/L, manganese less than 0.05 mg/L, and algae counts up to 2000 ASU/mL appeared to be "perfect candidates for direct filtration." In a survey of 17 direct filtration plants (13), short filter runs (6 or fewer hours) were occasionally observed when maximum raw water turbidity was 8 NTU or higher, and both short runs and filtered water turbidity above 1 NTU were sometimes observed when raw water turbidity was 20 NTU or higher (Figure 2). This figure shows the percentage of plants exceeding specified values for minimum, average, and maximum raw water turbidity. Symbols shown in the legend identify plant problems with high filtered water turbidity or short runs or both. From the work of Edzwald (10), it can be inferred that if the THM formation potential of a water exceeds 0.20 mg/L, direct filtration may not be able to produce a water that will meet the 0.10 mg/L MCL for trihalomethanes. Operation and maintenance for coagulationfiltration plants can be more demanding than that for DE plants or slow sand filter plants. Both conventional plants and direct filtration plants should be monitored carefully, because failure to obtain optimum coagulation can result in poor filter performance. Conventional plants are generally considered to have a "margin of safety" with respect to coagulation control. If coagulation control is lost at the chemical feed and rapid mix point, because of the hours of detention time afforded by settling basins, and if this goes unnoticed until the poorly coagulated water reaches the filters, plant operators could find themselves in the dilemma of having settling basins full of water that could not be filtered successfully. Coagulation monitoring and control are very important, whether or not the plant employs sedimentation. One traditional approach to control is jar testing. For
Figure 2. Influence of raw water turbidity on granular media plant performance
Page 101
waters of perhaps 10 NTU or higher, jar testing combined with continuous monitoring of the turbidity of the filtered water at individual filters is an approach frequently used. If raw water quality can change rapidly, or if the raw water turbidity is low (below 10 NTU), jar tests may not be very effective, because of the time required for testing, or because of the smaller differences in raw and settled water turbidities. In such instances, coagulant dose control by zeta potential instrumentation, a streaming current detector, or a pilot filter may be appropriate. Wagner and Hudson (25) suggested that filter paper filtration using Whatman No. 40 paper could give information on the treatment levels that produce acceptable water quality. Other appropriate monitoring would include pH, head loss, chemical feed, and raw and filtered water turbidity. Maintenance operations would include care of chemical mixers and feeders, perhaps flocculation basin mixers and sludge removal equipment in settling basins. Filter backwashing is necessary, and backwash water and settling basin sludge may require treatment and ultimate disposal. If sludge removal from settling basins is not done mechanically, periodic manual basin cleaning would be needed. The level of operating skill needed at coagulationfiltration plants is substantial. In order to effectively and efficiently control the coagulationfiltration process and attain low filtered water turbidity, operators need to understand the chemical aspects of coagulation. Large and medium sized plants are able to hire and keep trained operators who can effectively operate coagulationfiltration plants. On the other hand, small plants may not have the resources to hire or train operators who have a solid understanding of coagulation. This can lead to problems of poor treated water quality, if operators are unable to adjust treatment when raw water quality changes. Of the three processes discussed in this paper, coagulationfiltration has the greatest flexibility in the kind and concentration of contaminants that can be removed in the process, especially when sedimentation is employed. Conventional treatment can handle the widest range in raw water quality, and has been in use for several decades. Coagulationfiltration plants, because they employ more treatment processes, can be designed with the most flexibility in terms of the number of processes used. For example, settling might be used for muddy water but bypassed when raw water turbidity is low. Recent developments, such as use of media in the 1 to 2 mm size range, beds about 2 m in depth, and filtration rates of 25 m/h or higher provide even more treatment capability for the coagulationfiltration process, but until experience with such plants is gained, the very high rates of filtration probably should be considered only at large water utilities with well trained, full time operators and laboratory personnel. In spite of the many advantages that can be listed for coagulationfiltration, a number of drawbacks exist. The most important potential problem is this: for rapid rate granular media filtration to be an effective process for removal of particulate matter, the chemistry of the water must be manipulated so that coagulation is effective. This can be done through adjustment of pH and addition of an inorganic coagulant or polymer or both. At utilities that serve 50,000 to 100,000 persons or more, hiring one or more scientists to work in a water quality control laboratory can be considered feasible, as it is presently being done. At water utilities too small to employ a chemist, operation of the coagulantfiltration process may be less than optimum. Testing by persons who understand the process can establish the proper chemical treatment under the raw water quality conditions existing during the test period. The operator's failure to understand the implications of changing raw water quality and make proper adjustments could result in lower process efficiency, though. A fundamental concept is that coagulation chemistry is not influenced by the magnitude of the flow in a plant. Factors such as pH, alkalinity, and temperature must be considered, regardless of the size of the plant. A particular concern in northern latitudes or mountainous areas where giardiasis outbreaks may have occurred is the difficulty of effectively coagulating and filtering cold, clear waters. When the raw water turbidity is close to 1 NTU, some plant operators may question the value of adding a coagulant. Others may be discouraged by the apparent difficulty in treating a clear water at temperatures close to 0°C, and in both instances, operators may shut off the chemical feeders. Coagulant feed should never be interrupted nor shut off. Techniques are available for treating cold waters and low turbidity waters. Performing jar tests with the jars in an ice water bath is appropriate. Use of paper filters or small (2.5 cm) minifilters with beds 30 cm deep, or shallower, could be used to evaluate filterability of clear waters. Use of streaming current detectors as an online coagulant dose control device appears to work well in winter. Experience indicates that coagulationfiltration plants can produce high quality water even when temperature and turbidity in the raw water are low. The Duluth, Minnesota filtration plant consistently produced filtered water below 0.10 NTU and attained 99% to 99.99% reductions of asbestos fibers even when temperatures were in the 3 to 5°C range and raw water turbidity was 1 NTU (16). Conclusions (i) Each of the three filtration processes reviewed is different, and no single process is ideal in every circumstance. (ii) As process complexity increases, from slow sand filtration, to DE filtration, to coagulationfiltration, the skill level needed for effective operation increases. (iii) As process complexity increases, producing high quality filtered water increasingly becomes dependent on operator skill and ability. (iv) A variety of filtration processes have been used successfully either on a pilot plant scale or at full scale to remove Giardia cysts from water.
Page 102
(v) Slow sand filtration, DE filtration, and the coagulationfiltration (inline or direct filtration) processes used without sedimentation are all affected by raw water quality, with respect to both filtered water quality and plant performance characteristics, such as filter run length. Therefore, if use of any of these processes is contemplated with a water source that is not presently being treated successfully by the process, performing a pilot plant study before design and construction of the treatment plant is highly advisable. (vi) Even though important limitations exist and must be taken into account, filtration technology capable of removing 99% or more of the Giardia cysts from drinking water exists and is in use in many locations. Disclaimer Mention or use of commercial products does not constitute endorsement by the U.S. Environmental Protection Agency. Literature Cited 1. AlAni, M.Y., Hendricks, D.W., Logsdon, G.S. and C.P. Hibler, 1986. Removing Giardia cysts from low turbidity waters by rapid rate filtration. J. American Water Works Assoc. 78:5:6673. 2. Bellamy, W.D., Lange, K.P. and D.W. Hendricks. 1984. Filtration of Giardia cysts and other substances: Volume 1. diatomaceous earth filtration. EPA600/2 84114, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1967. 3. Bellamy, W.D., Hendricks, D.W. and G.S. Logsdon. 1985a. Slow sand filtration: Influences of selected process variables. J. American Water Works Assoc. 77:12:6266. 4. Bellamy, W.D., Silverman, G.P. and D.W. Hendricks. 1985b. Filtration of Giardia cysts and other substances: Volume 2. Slow sand filtration. EPA 600/2 85/026, U.S. Environmental Protection Agency, Cincinnati, Ohio. 5. Bishop, S., Craft, T.F., Fisher, D.R., Ghosh, M., Prendiville, P.W., Roberts, K.J., Steimle, S. and J. Thompson. 1980. The status of direct filtration, Committee report. J. American Water Works Assoc. 72:7: 405411. 6. Cleasby, J.L., Hilmoe, D.J. and C.J. Dimitracopoulos. 1984a. Slow sand and direct inline filtration of a surface water. J. American Water Works Assoc. 76:12:4455. 7. Cleasby, J.L., Hilmoe, D.J., Dimitracopoulos, C. and L.M. DiazBossio. 1984b. Effective filtration of small water supplies. EPA600/284083, U.S. Environmental Protection Agency, Cincinnati, Ohio. 8. Craun, G.F., 1986. Waterborne giardiasis in the United States 19651984. Lancet II: 8505:513514. 9. DeWalle, F.B., Engeset, J. and W. Lawrence. 1984. Removal of Giardia lamblia cysts by drinking water treatment plants. EPA600/284069, U.S. Environmental Protection Agency, Cincinnati, Ohio. 10. Edzwald, J.K., 1986. Conventional water treatment and direct filtration: Treatment and removal of total organic carbon and trihalomethane precursors, p. 199 236. In: Organic Carcinogens in Drinking Water: Detection, Treatment and Risk Assessment, Ram, N.M., Calabrese, E.J. and R.F. Christman eds., John Wiley & Sons, N.Y. 11. Fox, K.R., Miltner, R.J., Logsdon, G.S., Dicks, D.L. and L.F. Drolet. 1984. Pilot plant studies of slowrate filtration. J. American Water Works Assoc. 76:12:6268. 12. Lange, K.P., Bellamy, W.D., Hendricks, D.W. and G.S. Logsdon. 1986. Diatomaceous earth filtration of Giardia cysts and other substances. J. American Water Works Assoc. 78:1:7684. 13. Letterman, R.D. and G.S. Logsdon. 1976. Survey of direct filtration practice Preliminary report. Presented at American Water Works Association Annual Conference, New Orleans, Louisiana. June, 1976. 14. Letterman, R.D. and T.R. Cullen, Jr., 1985. Slow Sand Filter Maintenance. EPA/600/285/056, U.S. Environmental Protection Agency, Cincinnati, Ohio. 15. Logsdon, G.S., Symons, J.M., Hoye, R.L., Jr. and M.M. Arozarena. 1981. Alternative filtration methods for removal of Giardia cysts and cyst models. J. American Water Works Assoc. 73:2:111118. 16. Logsdon, G.S., Evavold, G.L., Patton, J.L. and J. Watkins, Jr. 1983. Filter plant design for asbestos fiber removal. J. of Environmental Engineering. 109:4:900 914. 17. Logsdon, G.S., Thurman, V.C., Frindt, E.S. and J.G. Stoecker. 1985. Evaluating sedimentation and various filter media for removal of Giardia cysts. J. American Water Works Assoc. 77:2:6166. 18. Malina, J.F., Jr., Moore, B.D. and J.L. Marshall. 1972. Poliovirus removal by diatomaceous earth filtration. Center for research in water resources, The University of Texas, Austin, Texas. 19. McCormick, R.F. and P.H. King. 1982. Factors that affect use of direct filtration in treating surface waters. J. American Water Works Assoc. 74:5:234242. 20. Pavoni, J.L., Tenney, M.W. and W.F. Echelberger, Jr., 1972. Bacterial exocellular polymers and biological flocculation. J. Water Pollution Control Federation 44:3:414431. 21. Pyper, G.R. 1985. Slow sand filter and package treatment plant evaluation: Operating costs and removal of bacteria, Giardia, and trihalomethanes. EPA/600/2 85/052, U.S. Environmental Protection Agency, Cincinnati, Ohio. 22. Robeck, G.G., Clarke, N.A. and K.A. Dostal. 1962. Effectiveness of water treatment processes in virus removal. J. American Water Works Assoc. 54:10:12751290. 23. Slezak, L.A., and R.C. Sims. 1984. The application and effectiveness of slow sand filtration in the United States. J. American Water Works Assoc. 76:12:3843. 24. Taylor, E.W. No Date. Fortyfifth report on the results of the bacteriological, chemical and biological examination of the London waters for the years 19711973. Metropolitan Water Board, London, England. 25. Wagner, E.G. and H.E. Hudson, Jr. 1982. Lowdosage highrate direct filtration. J. American Water Works Assoc. 74:5:256261. 26. Water Supply Division, U.S. Environmental Protection Agency. 1980. Manual for evaluating public drinking water supplies. EPA 430/975011. Washington, D.C.
Page 103
Monitoring as a Tool in Waterborne Giardiasis Prevention Jan L. Sykora*, Wilder D. Bancroft, Albert H. Brunwasser, Stanley J. States, Maurice A. Shapiro, Susan N. Boutros and Louis F. Conley Graduate School of Public Health, Department of Industrial Environmental Health Sciences, University of Pittsburgh, Pittsburgh PA 15261 In December 1983 a waterborne epidemic occurred in McKeesport, PA where 347 people developed giardiasis. A series of unusual events, such as exceedingly low temperatures and water line breaks were blamed for this outbreak. After the outbreak was controlled, a monitoring program for Giardia cysts was introduced and the existing turbidity monitoring program was expanded. Using the EPA technique samples were collected from raw water (Youghiogheny River) and McKeesport finished water on a fortnightly basis. The major source of Giardia cysts was treated and untreated sewage contaminating the Youghiogheny River. On January 12, 1986 a cyst was again isolated from a finished water sample. Subsequent analysis recovered cysts from a reservoir, a standpipe and tap samples. A "boil water advisory" was issued on January 17, 1986. A vigorous process of remedial actions, which included repair of filters and increased chlorine addition to achieve a CT value of 240, were instituted. No giardiasis cases above the background level were reported and the "boil water advisory" was lifted on February 21, 1986. The monitoring results showed that turbidity cannot be solely relied upon when evaluating the effectiveness of filtration plants for Giardia cyst removal. The Giardia monitoring of raw water indicated annual as well as seasonal variations in cyst counts with the greatest number of cysts detected in colder seasons. Based on this study it is recommended that public water systems determined to be at risk of Giardia contamination consider a routine monitoring program. It is further suggested that monitoring efforts be intensified during the colder season of the year.
Introduction On February 22, 1984 the Allegheny County Health Department (ACHD) declared that a waterborne giardiasis epidemic was occurring in the McKeesport Municipal Water Authority's (MMWA) service area and issued a "boil water" advisory. An epidemiological investigation confirmed the hypothesis that drinking water was the vehicle transmitting Giardia cysts to the 48,700 people in the service area. By April 20, 1984 there were 347 medicallyconfirmed cases of giardiasis in the affected population. The MMWA is located south of Pittsburgh at the confluence of the Monongahela and Youghiogheny Rivers. It is a 9 MGD (34.1 ML/d) treatment facility utilizing chemical coagulation, flocculation, sedimentation, filtration, and chlorination. The plant's raw water supply is derived from the Youghiogheny River. The treatment plant was constructed in 1907 to serve residents in the communities of McKeesport, Versailles, Port Vue, White Oak, and their major industry, the National Tube Co.. The investigation into the cause of the giardiasis outbreak has determined that it occurred as a result of a combination of several events and circumstances. In late December 1983 unusually low ambient temperatures were experienced over an extended period of time, resulting in numerous water line breaks that were not promptly identified or repaired. One break occurred in a water line in the bed of the Youghiogheny River. Another occurred in an industrial plant that was shut down. Still other breaks occurred in the water distribution network. Monitoring capabilities of the water system had not been modernized to enable expeditious identification of the location and magnitude of such problems. The resulting high water demand exceeded the treatment capacity of the plant's coagulation, flocculation, sedimentation and filtration systems. Additionally, because an elevated backwash tank was out of service at the time, all backwash water was being supplied from the distribution system. Consequently, filters were run for several days without backwashing and a largescale breakthrough led to a significant increase in turbidity (5). The second major factor contributing to the outbreak was that over several preceding decades, the water treatment plant had not been maintained properly and there were numerous treatment deficiencies. In summary, the outbreak occurred as a result of the presence of Giardia cysts in the raw water supply and excessive water demands which compromised the treatment capabilities of the plant. At the time of the giardiasis outbreak the experience of water industry regulators in Pennsylvania dealing with this disease was in its infancy. In all of the United States, there were only a few laboratories capable of analyzing water samples for Giardia cysts. The Allegheny County Health Department, in cooperation with the Federal Environmental Protection Agency and the Pennsylvania Department of Environmental Resources issued two orders to the MMWA. The first, issued on March 16, 1984, required remedial maintenance of the * Corresponding author.
Page 104
plant and the disinfection of reservoirs and the distribution system to allow the lifting of the "boil water" advisory. The second order, based on an engineering analysis by the MMWA's consultant and issued on September 17, 1984, directed the construction of a new water treatment plant. The orders also established a threephase monitoring program as a barrier against future waterborne giardiasis outbreaks. In the first phase, the Allegheny County Health Department expanded the Federal Safe Drinking Water Act's (7) turbidity monitoring protocol to require the MMWA to sample finished water every four hours. The Federal Safe Drinking Water Act requires that turbidities be less than 1 NTU for a 30 day average and not exceed 5 NTU's for a 2 day average. The American Water Works Association (1) has recommended 0.1 NTU as a quality goal for potable water. The second phase of the program established monitoring of the treatment plant's raw and finished waters for Giardia cysts on a fortnightly basis. The finished water sampling program usually included a clearwell sample and an open reservoir sample. The third phase involved identification of sources of Giardia cysts occurring in the McKeesport raw water supply, the Youghiogheny River. The results of this last phase were published elsewhere (9,10). Materials and Methods Giardia Sampling Procedures Samples collected at the water treatment plant were obtained using standard Giardia filtration equipment (3). In the case of finished water, 500600 gallons of water were filtered at a rate of 12 gal/min. In the case of raw water less than 200 gallons were filtered due to clogging. Following sample collection filters were refrigerated and maintained at 5°C until analyzed. Giardia Analysis Procedure The samples were analyzed using the Environmental Protection Agency technique by Schaefer and Rice (8) as modified by Schaefer (personal communication, 1984). In the laboratory, the filters were divided into fourths by unwinding the yarn into skeins. Each skein was rinsed in successive 1 liter aliquots of 0.01% Tween 20T M. One percent Tween solution was added as needed to maintain suds. Each sample was concentrated by sedimentation under refrigeration for 24 hours. The sediment was then centrifuged at 600Xg for 3 minutes, and the resulting pellets combined. 12 mL of sediment, resuspended in 7090 mL of 0.01% Tween 20T M, were then layered on top of 70 mL of 1.5 M sucrose solution (sp. gr. 1.18) in a 250 mL centrifuge bottle (8). After centrifugation at 800Xg for 5 minutes, the supernatant above the interface and the pellet were discarded. The interface and all underlying sucrose solution were diluted 5 times with 0.01% Tween 20T M and further processed by centrifugation at 600Xg for 2 minutes. The resulting sediment was washed twice with Tween 20T M in 50 mL centrifuge bottles, transferred to 15 mL centrifuge tubes, concentrated by centrifugation and stained by Lugol's solution using the Jakubowski and Ericksen procedure (3). The sediment obtained by this flotation process was examined microscopically using a PalmerMaloney nannoplankton counting chamber (0.1 mL) and a 45x objective on a microscope equipped with phase contrast. Positive identification of the Giardia cysts was based on dimension, shape, and detection of at least two internal morphological features. Turbidity Measurements Turbidity measurements were performed using a Hach 18900 continuous flow turbidimeter equipped with a chart recording device. The turbidimeter was calibrated using the manufacturer's operating instructions and EPA quality assurance standards. The turbidity monitoring procedure, established after the outbreak, consisted of measurements every four hours resulting in six determinations per day. An average value was calculated for each day. The turbidity was measured at the pressure side of the pump at the end of the clearwell prior to entering the distribution system. Results From November 1984 through the beginning of December 1986, a total of 43 samples were collected from the Youghiogheny River at the MMWA plant. The cyst levels in these samples averaged 73.6 cysts/100 gal (Table 1). The standard deviation (100.9) exceeded the mean indicating high variability in cysts concentration during the sampling period. The results show that most of the samples [20] contained between 11 and 100 Giardia cysts/100 gal, eleven samples contained 110 cysts/100 gal, and twelve samples 101438 cysts/100 gal. It is of interest that very low cyst counts were detected in the summer with averages ranging from 12.5 cysts/100 gal in 1986 to 25.4 cysts/100 gal in 1985 (Table 1). On the other hand, the greatest number of cysts was detected in colder seasons. The results also indicate annual variation in cyst counts. The cyst counts were much lower in 1986 than in 1985 with averages of 50.6 cysts/100 gal and 103.6 cysts/100 gal respectively. Similar decreases were noted when averages for individual seasons during 1985 and 1986 were compared (Table 1). Method quality control studies performed on sewage and described in a previous publication indicate that the filtration and sample processing procedures involved in the standard technique recover only 314% of the cysts present (3,10). These results agree with those of other authors and suggest that the actual concentrations of Giardia cysts entering the treatment plant were probably substantially higher than detected by the standard technique. As previously reported, samples collected from finished water were negative until a cyst was isolated from a clearwell sample collected on January 8, 1986 (10). Subsequent sample analyses by laboratories other than our own recovered cysts from a reservoir, a standpipe and tap samples. Table 2 summarizes the results of TABLE 1. Giardia Cyst Counts in the Youghiogheny River
Time period
Range
Mean
Standard Deviation
# of Samples
Nov 84Dec 86 JanDec 85
1438
73.6
100.9
43
1438
103.6
132.1
21
JanDec 86
3201
50.6
54.9
21
Winter 85
2416
133.0
194.6
4
Spring 85
35438
147.6
165.8
5
Summer 85
1114
25.4
49.6
5
Fall 85
12262
112.5
103.8
6
Winter 8586
32201
84.3
79.4
4
Spring 86
6148
46.7
53.4
6
Summer 86
325
12.5
9.0
6
15141
74.0
51.6
5
Fall 86
Cysts/100 gal
Page 105 TABLE 2. McKeesport Water Authority Giardia cysts/100 gallons (December 4, 1985 April 25, 1986)* Sampling Period Raw Water
December 418
January 823
February 1127
March 1327
April 1025
36103
37**
32201
67148
721
Clearwell
0
01
0
0
0
Filters
02
0
0
0
Standpipe
04
0
Reservoir
01
0
0
Distribution System
01
0
0
Total Samples
4
15
13
6
6
% Contaminated Samples***
0
66
0
0
0
* From Sykora et al. (10) ** One sample only *** Raw water excluded Samples not collected
our Giardia cyst analyses performed during and shortly after detection of Giardia in the clearwell. This table indicates that during the critical period of January 823, 1986, 66% of drinking water samples contained cysts. The first samples obtained from the distribution system on January 13, 1986 (3 samples) and one sample collected from clearwell No. 4 on January 15, 1986, were negative. However, three samples collected from the distribution system on January 23, 1986 were contaminated. The Allegheny County Health Department issued a "boil water advisory" on January 17, 1986. A vigorous program of actions which included repair of filters, addition of filter media and an increase in chlorine concentration to achieve a chlorine concentration x contact time (CT) of 240 was also instituted. No Giardia cases above the background level were reported and the "boil water advisory" was lifted on February 21, 1986. Since the revised monitoring program was established the daily average turbidity measurements have not exceeded the Federal Safe Drinking Water Standards (7). However, the quality goal of 0.1 NTU as recommended by AWWA has been exceeded most of the time (1). Table 3 and Figure 1 describe the variation in turbidity values, expressed as daily averages during the 1986 incident. TABLE 3. Results of turbidity measurements (NTU) McKeesportClearwell (December 1985 February 1986) Date
Range
Mean
S.D.
N
Dec 110
0.140.25
0.19
0.03
10
Dec 1120
0.140.70
0.28
0.17
10
Dec 2130
0.180.45
0.26
0.09
10
Jan 110
0.250.75
0.41
0.15
10
Jan 1120
0.190.52
0.30
0.11
10
Jan 2131
0.150.75
0.40
0.19
10
Feb 111
0.120.50
0.27
0.12
11
Feb 1222
0.060.24
0.15
0.06
11
S.D. = Standard deviation N = Number of daily average samples
Discussion The results show that turbidity monitoring at the clearwell cannot be solely relied upon when evaluating treatment effectiveness of filtration plants for Giardia cyst removal. The McKeesport experience demonstrates that a filter could be defective in one area allowing the passage of Giardia cysts, while the combined effluent of all filters masks the breakthrough when using turbidity as an indicator. This is consistent with earlier laboratory studies with granular filtration performed by Logsdon et al. (4) which showed that the concentration of cysts passing through a filter could be high, while finished water turbidities remained below 1.0 NTU. Logsdon et al. (4) have also indicated that the removal of G. muris cysts by diatomaceous earth (DE) filters does not seem to be closely related to turbidity removal. In a different publication directly related to the McKeesport outbreak, Logsdon et al. (5) showed substantial variations in Giardia cyst passage rates through treatment plant filters while finished water turbidities were between 0.3 and 1.2 NTU. Substantially fewer cysts passed through the filters when turbidity was 0.20 NTU or lower however. Thus, the McKeesport field experience indicates that if turbidity is monitored at the clearwell, rather than at individual filters, operators may be unaware of the conditions of individual filters that could permit passage of Giardia cysts into finished water. Other contributing factors to the waterborne giardiasis outbreak in McKeesport were high concentrations and persistent presence of cysts in the Youghiogheny River with distinct maxima in colder seasons. Marrocco et al. (6) showed that most of the Giardia lamblia contamination of unfiltered water systems in Pennsylvania was detected during the timeframe of December through June. Craun (2), who evaluated the current status of waterborne giardiasis in the United States, showed that the outbreaks associated with community water systems occurred most frequently in the spring and fall/early winter, while outbreaks affecting visitors or campers occurred most frequently during the summer months. Both McKeesport incidents had onsets in early winter and all recent
Page 106 TABLE 4. Recent giardiasis outbreaks in Pennsylvania Location Bradford Pittston
Boil Water Advisory Date October 26, 1979 December 23, 1983
WilkesBarre/Scranton Houtzdale McKeesport
March 9, 1984 November 14, 1984 February 22, 1984
giardiasis outbreaks associated with community water systems in Pennsylvania occurred during cold weather (Table 4). Thus, the results from this study support Craun's suggestion that the outbreaks affecting permanent residents may reflect not only less effective treatment during fall/early winter, but also increased contamination of raw water supplies. The high concentrations of Giardia cysts during colder seasons may be the result of several factors such as flow, temperature, water density and less effective sewage treatment, as well as increased survival of the cysts in cold water. Results from our previous studies showed that effluents from sewage treatment plants located on the Youghiogheny River contained between 50 and 5.1 × 103 Giardia cysts/100 gal while raw wastewater concentrations ranged between 6.6 × 103 and 1.5 × 105 cysts/100 gal (9,10). Thus, the discharge of raw or poorly treated sewage during winter can make a substantial difference in Giardia cyst concentrations in raw water. In conclusion, since the monitoring of finished water for the presence of Giardia cysts is limited by a laboratory technique that results in a low recovery rate, there is a continuing need for research and development in the area of sampling and laboratory analysis to improve the analytical procedure for Giardia cyst isolation. In spite of these limitations, the McKeesport experience suggests that monitoring of raw and finished water supplies can be useful and can complement turbidity measurements. Therefore, water utilities subject to Giardia contamination of their raw water source(s), may choose to adopt routine monitoring for Giardia with emphasis on cold season sampling.
Figure 1. McKeesport daily effluent turbidities from December 1, 1985 to February 22, 1986
Acknowledgements This work was supported by the Pennsylvania Department of Environmental Resources and the Bureau of Environmental Health, Allegheny County Health Department, Pittsburgh, Pennsylvania. References 1. American Waterworks Association. Quality goals for potable water adopted by American Water Works Association Board of Directors, January 28, 1968. In: 198586 Officers and Commitee Directory, AWWA, Denver, Colorado, 1986. 2. Craun, G.F.. 1984. Waterborne outbreaks of giardiasis: current status. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer, (eds.). Plenum Press. New York and London. pp. 243259. 3. Jakubowski, W., and T.H. Ericksen. 1979. Methods for detection of Giardia cysts in water supplies. In: Waterborne Transmission of Giardiasis. W. Jakubowski and J.C. Hoff (eds.), Environmental Protection Agency 600/979 001. pp. 193210. 4. Logsdon, G.S., DeWalle, F.B., and D.W. Hendricks. 1984. Filtrations as a barrier to passage of cysts in drinking water, In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer, (eds.). Plenum Press. New York and London. pp. 287309. 5. Logsdon, G.S., Thurman, V.C., Frindt, E.S., and J.G. Stoecker. 1985. Evaluating sedimentation and various filter media for removal of Giardia cysts. Journal of AWWA 77:6166. 6. Marrocco, F.A., Lengel, L.L., and D.N. Greenfield. 1987. Giardia monitoring and regulation in Pennsylvania surface Waters. In: Proceedings of 1986 AWWA Water Quality Technology Conference, Portland OR., November 1620. pp. 10551066. 7. Safe Drinking Water Act. Public Law 93523, December 16, 1974, Washington DC., U.S.A.. 8. Schaefer, F.W., and E.W. Rice. 1981. Giardia methodology for water supply analysis, In: Water Quality Technology Conference Proceedings, Seattle, December 69, 1981. American Water Works Association, Denver CO. pp. 143146. 9. Sykora, J.L., Bancroft, W.D., States, S.J., Shapiro, M.A., Boutros, S.N., Keleti, G., Turzai, M., and L.F. Conley. 1986. Giardia cysts in raw and treated sewage. In: Controlling Waterborne Giardiasis. G.S. Logsdon (ed), Environmental Engineering Division ASCE (In Press). 10. Sykora, J.L., States, S.J., Bancroft, W.D., Boutros, S.N., Shapiro, M.A., and L.F. Conley. 1987. Monitoring of Water and Wastewater for Giardia. In: Proceedings of 1986 AWWA Water Quality Technology Conference, Portland OR., November 1620. pp. 10431054.
Page 107
The Efficiency of Point of Use Devices for the Exclusion of Giardia muris cysts from a Model Water Supply System D. Roy Cullimore* and Helen Jacobsen Regina Water Research Institute, University of Regina, Regina, Saskatchewan S4S 0A2, Canada. A recycling water system was developed which could circulate water free of particles larger than 0.22 µm through various point of use water treatment devices including ultraviolet light, reverse osmosis, ozone, granulated activated carbon and proprietary faucet attachment systems. Once operating, the recycling water was charged with viable Giardia muris cysts and the ability of the various devices to kill and/or exclude the cysts was monitored under various conditions. Viability was determined by excystation and a novel staining technique reported elsewhere. The presence of the cysts in the water was monitored using a 1.2 µm cellulose acetate membrane filter. Where cysts became entrapped on the filter, this was monitored by either pressure differential shifts, discharge flow rates and/or the detection of cysts on the filter. Under abnormal operating conditions such as ultrahigh loadings and prolonged operational times, all of the devices failed to either exclude G. muris cysts from the product water or to kill all of the passaged cysts. The relative efficiencies of the devices tested for elimination of viable cysts from the product water are grouped as follows: (1) high efficiency ultraviolet irradiation, ozone; (2) moderate efficiency reverse osmosis; and (3) poor efficiency granulated activated carbon, faucet attachments.
Introduction Waterborne giardiasis outbreaks have recently been more frequently reported on the North American continent. Giardiasis is a diarrheal disease which is commonly caused by the digestion of Giardia lamblia cysts (1). Following cyst ingestion, excystment occurs in the small intestine with the emergence of bilaterally symmetrical, flagellated trophozoites which attach themselves via a sucking disk to the walls of duodenum and upper jejunum. This causes an acute or chronic diarrheal illness which sometimes progresses to steatorrhea and/or the malabsorption syndrome (2,16). In 1976, the State of Washington recorded its first confirmed outbreak of waterborne giardiasis which occurred in the City of Camas. Approximately 600 people were affected representing 1015% of the population. The city water supply was a blended mixture of surface and well water sources involving two mountain creeks and seven deep wells. While the well waters were not implicated, G. lamblia cysts were recovered from both the raw surface water supply and the distribution system (15). Clearly, the water treatment procedures which included mixedmedia filters and chlorination had failed to exclude the cysts from the distribution system. Another major outbreak in a nearby area occurred in the winter of 198182 at the Banff National Park in Alberta, Canada. A total of 121 confirmed cases of giardiasis were reported. These and other similar outbreaks highlighted the need to reevaluate the capabilities of the various options presently used in existing water treatment facilities specifically to remove Giardia cysts from raw water supplies. It has been proposed that outbreaks of giardiasis which are waterborne in origin may be the result of process deficiencies in the water treatment plants, excessive contamination of water sources for which the treatment process becomes inadequate, or that the contamination of the water occurred after treatment (8,14). In general, it has been observed that the Giardia cysts are generally more resistant to disinfection than are the normal ''hygiene indicator" organisms such as coliforms (4,12). Unfortunately, the standard disinfection (chlorination) practices recommended for water treatment have been found to be ineffective because of insufficient dosages and/or too short a contact time (6). Treatment by ultraviolet radiation has been considered as major alternative to chemical disinfection procedures for small water systems because of its simplicity and economy in operation. However, Rice and Hoff suggested in 1981 that in areas where the water supplies contain G. lamblia, the use of ultraviolet radiation at conventional dosages proved to be inadequate for the satisfactory disinfection of potable water supplies (11). Another disinfection procedure receiving increased attention is the use of ozone. Most research has focused on the control of aquatic bacterial and viral populations with very little emphasis on the control of protozoal cysts. Research at the University of Washington (3) concluded that, in general, protozoal cysts are more resistant to ozone than are bacteria and viruses. In addition, the disinfective capability of the ozone was affected very significantly by both pH and temperature. At low temperatures (1°C) and high pH (79), Giardia cysts were found to possess a greater resistance to ozone than at room temperatures and the slightly acidic waters with a pH range of 5 to 7. * Corresponding author.
Page 108
One option available to the consumer of a potable water supply in which the user, for whatever reason, considers that there is an inherent risk in consuming water is to install a point of use device to remove the perceived risk. The recent publicity on outbreaks of giardiasis has generated claims that some of these devices will specifically exclude Giardia cysts from the product water. A point of use device can be defined as a single or multiple series of water treatment devices which when interfaced with the supply will cause a designed exclusion on such selected agents as will render the final product acceptable as a potable water supply. In general, these units are either free standing batch systems, under sinkrestricted continuousflow on demand or unrestricted continuousflow set on the whole water system. This study was designed to determine the effectiveness of a limited range of point of use devices in excluding Giardia muris cysts artificially supplemented into a particulate free water. The specific selected units functioned by either ultraviolet light, activated carbon, reverse osmosis, filtration or ozonation. Methods Selection of the Point of Use Systems A variety of units were evaluated in order to determine which single representative would be tested for each system. In all cases, emphasis was placed upon units that were likely to be installed in a domestic home involving an average throughput in total of 1,000 L/day. Selection also was based upon the claims of the various manufacturers for each style of unit. These claims are summarized (in brackets) for each system: ultraviolet treatment (kills waterborne bacteria and viruses); reverse osmosis (will remove up to 95% of total dissolved solids including salt, sulfates, sediments, bacteria, algae, phosphates, mercury, arsenic, nitrates, chlorine, detergents, rust); ozonation (kills bacteria, removes odour, taste, colour and oxides many organic and inorganic impurities); granulated activated charcoal (removes taste and odor, reduces levels of potentially toxic or carcinogenic compounds, produces a water with a lower viable organism count); faucet filters (aids in the removal of chemicals, contaminants, chlorine, toxic substances such as trichloroethylene and trihalomethane, iron, rust, silt, sediment). From these recommendations it is evident that none relate specifically to the exclusion of protozoal cysts and scientific supporting evidence for some of the other claims appears tenuous. To evaluate these devices under comparable conditions to small domestic systems, a test system was designed which could utilize a recycling water system with a controllable flow into which Giardia muris cysts could be inoculated. The configuration of the system is given in Figure 1. Water was retained in a storage tank holding approximately 80 L of water and was pumped from there to a pressure tank and on through 1/2 inch P.V.C. piping through to two separate recyclable systems. The operating range for the recycling unit included fluctuating water pressure operating over 30 to 80 p.s.i. with flow rates adjustable up to a maximum of 3 gal/min (11 L/min). In most experiments, single distilled water was used unless otherwise stated. One looped system was specifically designed to test a reverse osmosis filtration system (Unbottled Water Systems manufactured by Wetco Ltd., Las Vegas, Nevada). Water was pumped through the device and the "brine" water discharge recycled back into the main system flow. Water which had passed through the membranes of the reverse osmosis unit was collected using a separate return line into which a membrane filtration housing could be interjected to allow sampling for passaged cysts. To test the other inline continuous devices, the second loop was specifically designed to allow the ultraviolet light, activated carbon and the faucet filter to be installed at different times. Before any device was tested on either loop, the water was filtered down to remove all particles greater than one µm in size. This was done by recycling the water through the loop and filtering the water through 50, 20 and 10 µm spun Orlon filters followed by membrane filtration at 2 and 1 µm. Continuous running of the loop with no plugging (i.e., pressure differential build up) was taken to indicate that the water was particle free. Problems did arise from some steel fixtures generating rust within the system but this was controlled by the application of polyurethane varnish to all exposed surfaces. Various units were installed at different times on this second loop. The first unit tested was an ultraviolet treatment system (Ultraviolet Technology Inc., San Marcos, California, U.S.A.). The discharged water was either collected separately or passed on continuously through the recycling head. Two optional filters were installed at the confluence point for the return of the water to the reservoir. One filter was used for the evaluation of granulated activated charcoal filters while the second unit was used as a general particulate filter when the system needed to be scrubbed. Pressure gauges were installed at points around the cyclic circuit in order to assure no significant pressure losses throughout the system. The flowrate through each line was measured using the drum test method (using a 1 L graduated cylinder), the filling time measured with a stop watch. A low flow (0.2 to 3.5 mL/min.) peristaltic pump was used to directly inoculate cyst suspensions into the reservoir which was constantly stirred using a rotary propeller in order to ensure an even distribution of the cysts without any settling of cysts to the bottom of the reservoir. After testing of the U.V. system was completed, the unit was exchanged in position for a faucet filter. To determine the effectiveness of ozonation, a portable batch treatment (5 L capacity) counter top ozonator (Ozonator Systems Inc., Mississauga, Ontario, Canada) was employed in which the manufacturer's specification called for a 15 minute power up in order to allow the ozonation generated within the device to treat the water. A small granulated carbon filter was employed in the channel to the faucet as a further treatment device. All studies conducted involved sampling from the midpoint of the reaction vessel to avoid the evaluation of this secondary process. Sampling Techniques Three different sampling techniques were employed in this study. These were: (1) cyst entrapment of the total flow using a membrane filter (U.V.); (2) cyst entrapment of a sample flow by membrane filtration (reverse osmosis); and (3) direct sampling from the treated water (ozonator). Other evaluations used more than one of these techniques.
Figure 1. Recycling unit format for testing the ability of various point of use devices to exclude viable Giardia muris from product water. AL loop A BL loop B BP bipass line FM flow meter CF chemical feeder to dispense cyst suspensions FS filtering system PM pressure meters PS pressure switch PU pump RT return line to ST ST sol'n mixing tank XT Xtrol pressure tank
Page 109
In order to efficiently use the membrane filters to entrap any suspended cysts in the water, the recycling system had to be particulate free. Premature plugging of the membrane filters with other particulate material reduced the efficiency of the recovery systems in two ways: (1) the cysts could no longer be seen amongst the foreign particulate material present on the membrane; and (2) the rapid plugging of the membrane filter may significantly reduce the efficiency of entrapment allowing small numbers of cysts through the treatment system. In the routine evaluation of all of the pointofuse devices, G. muris rather than G. lamblia cysts were used as the marker cells. The human infective species (G. lamblia) was considered to present a health risk, a problem in obtaining excystation and unreliable for obtaining a consistent supply of cysts. G. muris on the other hand is more convenient to reproduce in Swiss Albino mice and the cysts are more amenable to in vitro excystation (1,7). Cysts were originally obtained from the collection at the University of Washington, Seattle, Washington. These cysts were used to infect Swiss Albino mice supplied by the Animal Resources Center, University of Saskatchewan, Saskatoon. Problems with cross infections caused erratic cyst production which was corrected by the use of C3H/HE mice which resisted any transient laboratory infections and also remained infected for a longer period of time (13). These mice were obtained from Charles River, Montreal, Quebec, Canada. Cyst isolation from the mouse feces was by a modification of the sucrose gradient method described by DeWalle and Jansson at Washington University (3). One major modification of this technique was the development of a screen customized to fit in the bottom of the mice cages. The mice were maintained on a 10 mm wire screen so that the feces would drop automatically through into a 5 mm deep distilled water trough below. This method decreased cyst isolation time to about one half. Cysts were produced in a routine manner with production gradually rising during the project from 100,000 to 40,000,000 cysts per week. The mouse colony varied in size from 10 to 48 members depending on demand. The purified cysts were resuspended in distilled water and stored at 8°C (5). Determination of Cyst Viability Giardia cyst survival after the various treatments was monitored as the percentage of cysts which could be induced to undergo in vitro excystation. The procedure used was a modification of the procedure used at the University of Washington (3). In this method, the cysts were exposed to Hanks balanced salt solution enriched with carbon dioxide generated by sodium bicarbonate. Viability was determined by the ratio of cysts which excysted when exposed to a tyrodetrypsin solution, observed using a 0.5 mL counting chamber over an inverted microscope at a magnification of ×200. Results and Discussion Reverse Osmosis Efficiency The reverse osmosis drinking water system used 18 cellulose triacetate coated perforated plastic tubes to provide a semipermeable file through which the ultra filtered water would pass. These tubes were installed vertically in a concentric ring with the treated water passing vertically to a common collection system at a rate of up to 12 L per day. The remaining water returned to the downstream main water flow. Low volumes of water under regular household pressures would therefore be forced through the perforated tubes and be fed into the "pure" water line (product water). The additional brine created was retained within the major water stock and flushed through the system with the normal utilization of the water for nondrinking purposes. During experiments with the reverse osmosis unit, the monitoring for cyst passage occurred in the total volume of the product water. The nature of the unit in acting as an ultra filtration system should ensure that no cysts would pass through to the product water. To monitor this, the reservoir water was inoculated with 10 cysts/mL of Giardia muris and subjected to constant recycling through the system operating at 30 to 80 p.s.i. and the product water was all passed through a 1.2 µm membrane filter. These monitoring filters were removed on a regular basis. To determine whether cysts were present, the membranes were thoroughly washed by agitation on a wrist action shaker (50 mL sterile particle free water for 5 minutes) followed by concentration using centrifugation at 2,500 r.p.m. for 10 minutes. A Neubauer hemocytometer was used to determine whether any cysts were present. In practice, the reverse osmosis unit effectively excluded the cysts both with new and recoated tubes for a period of 80 and 100 days respectively, but after this sporadic releases of cysts occurred in a cyclic manner (Figure 2). These periodic releases of cysts suggest that some microsheering had occurred in the membrane fabric to a sufficient extent to allow the direct passage of the cysts into the product water. At the termination of each experiment, the unit was disassembled. Each tube was removed and the incumbent water inside each was examined for the presence of cysts. Direct microscopic examination revealed high numbers of cysts in one tube and low numbers in seven others after the completion of the last experiment on recoated tubes. Excystation was performed on the cysts and 30% of them were found to be still viable. Parallel research work being conducted concurrently by Morrell in 1985 (16) revealed that G. muris cysts tended to collect at low seepage (roll over) points within a recirculating biofouling gravel pack/tube system. A similar phenomenon may have happened due to the vertical tube positioning in the test unit where the slow infiltration of the water through the
Figure 2. Release of cysts through a newly installed (continuous line) and a recoated (discontinuous line) reverse osmosis unit where time is recorded in days of operation (x axis) and the number of entrapped recorded cysts per membrane filter is given (y axis).
Page 110
membrane could lead to the cysts actually precipitating and concentrating at the bottom of each of the filter tubes. In the tube releasing very high concentrations of cysts, passage of cysts into the product water probably occurred mainly, if not entirely, through a rupture in the membrane fabric of this tube with more casual contamination occurring through secondary passages to the other tubes. This erratic presence of Giardia cysts in the discharged product water may also be, in effect, a reflection of the degree of biofouling occurring around the tears within the cellulose triacetate membrane and the degree of turbulence caused by tears breaking open. If enough cysts have collected in the tube when a new tear occurs, sufficient pulse may be created to agitate the water cysts and eject some from the tube with the product water. As subsequent microbial biofouling would at least partially close off the tear, the numbers of cysts in the product water would decrease. Where the tears became completely sealed off, no more spurious turbulence would occur and the cysts would again tend to settle out in the tubes. In all probability, the efficiency of the membranes at excluding Giardia are much shorter in duration than the evidence would suggest in the expelled product water tubes. In the repeat studies using recoated tubes, product water was frequently positive for macronidia upon microscopic examination at ×200. This would imply that a mold was now actively growing in the product water system. In summary, the reverse osmosis unit appeared to be initially effective at excluding Giardia cysts from the product. Unfortunately, there is no simple field test which could rapidly determine whether the product water was free from passaged cysts or that the membranes are, in fact, still intact and functioning satisfactorily. Because of this, restraints should be placed upon the effective life span of these membranes where promoted for the effective exclusion of Giardia cysts from drinking water. Ultraviolet Light Treatment The installed ultraviolet treatment unit generated a maximum of 40,000 µwatts/cm2 of 253.7 nm wavelength. The method for testing this system involved loading the water with different concentrations of cysts at the reservoir point in the system and recycling this water continuously past the U.V. unit. All of the water that had passed through the unit was filtered through a 1.2 µm membrane filter to entrap the cysts. The cyst concentrations used were 800, 8,000 and 80,000 cysts/L. Controls were run by recycling the same concentrations of cysts through with an inactivated U.V. light source. The numbers of excysted cysts present in the controls was taken to represent the null effect with a 100% viability in the passaged cysts. To determine the efficiency of the treatment, the numbers of viable cysts found in the treated samples were adjusted accordingly. While the U.V. light treatment did function on recycling water, the cysts passaged through the unit were subjected to only a single exposure and then filtered out. In no experiments were the cysts subjected to phased or continuous reexposure. From these experiments, it would appear that the U.V. light unit tested was completely effective in controlling 800 cysts/L in the water but that efficiency declined as the cyst loadings increased, particularly beyond 8,000 cysts/L. Unfortunately, it was not possible to determine whether the higher numbers of cysts created a shielding effect which increased viability. Alternate particulate loadings using passive materials could not be conducted since this would seriously increase the rate of plugging at the interdicting filters. The loadings and retention times in these experiments were below those normally found in water systems due to decreased flow rates generated by the gradual plugging of the monitoring membrane filter with cysts. At 800 cysts/L, the input water was almost totally free of particulates with little shielding effect, so that with an adequate retention time, it was sufficient to cause the total inactivation of the Giardia cysts. Since water would normally be expected to contain far less that 800 cysts/L, it is probable that where the particulates are low and retention time is adequate, this unit could be used to control Giardia cysts potentially present in a drinking water supply used for domestic purposes. Granular Activated Carbon Filtration System Standard cartridge type filters were obtained from Water Conditioning Canada Ltd., Regina, Saskatchewan, Canada. Immediately after installation of the first filter and during the first three days of testing, some cysts were located in the product water. Upon prolonged testing over a 21 day period, no cysts were recovered from the water with two exceptions when two and four cysts were recovered. This absence of Giardia was postulated to be due to retention and possible elimination of the cysts from the water through the presence of an actively growing biofilm on the carbon particulate surfaces. To examine this hypothesis, when the next new filter was installed, a more detailed bacterial testing was conducted in conjunction with the evaluation for the cyst passage. This new filter excluded cysts through until day 13 when large clumps of cysts were sporadically released into the discharged treated water. Very high bacterial numbers occurred in the effluent water to day 9 and subsequently declined. At the time when cyst clumping was occurring, the bacterial numbers had reduced from a high of 640,000 c.f.u./mL down to a low point of 320 c.f.u./mL. This reduction in the detached (planktonic) bacterial population in the water appeared to coincide with the occurrences of clumped cysts in the effluent water. These sporadic appearances of cysts in the product water at different population densities showed signs of cyst wall destruction and lysis. These interactive effects warrant a detailed study in order to catagorize the dynamics of this potentially important reactive interface. Granulated activated charcoal filter systems would from this limited study appear to allow the passage of Giardia cysts either directly or through interaction with the bacterial biofilm coating the carbon granulated particles.
Page 111
Ozone The efficiency of ozone to inactivate Giardia muris cysts was tested using a 5 L bench top ozonator (Ozonator System Inc., Mississauga, Ontario). This device operates on a batch by batch basis in which 5 L aliquots of water are subjected to an ozone discharge on a timed basis (recommended time is normally 15 minutes). The manufacturers claim is that this exposure provides a "complete disinfection of the drinking water" rendering it safe to drink. These experiments therefore examined the relationships between the time required for the control of Giardia cysts and the time required to control bacteria. To achieve this, exposure/inactivation tests were conducted on the bacteria Pseudamonas flourescens and Escherichia coli, both isolated and identified to species at the Regina Water Research Institute, as well as on G. muris. For these trials, sterile distilled water was used for the bacterial trials and regular tap water for the G. muris studies in order to duplicate more normal conditions that would be expected in regular use. Parallel studies done on P. fluorescens and E. coli revealed a much faster rate of kill for the latter organism. The initial population of 56,000 c.f.u. E. coli/mL was reduced to 0 in 60 seconds (see Figure 3). For P. fluorescens, prolonged survival of ozonation occurred when 60,000 c.f.u. bacteria/mL were introduced with survival still being recorded after fifteen minutes of exposure. Viability tests conducted on Giardia included the direct addition of 5,000,000 cysts to 5 L of tap water in the ozonator tank. After supplementation, the ozonator was turned on and midpoint samples were taken at 0, 1, 2.5, 5, 10, and 15 minutes. These tests were run at 8°C, 22°C, and 37°C. At 8°C and 37°C the cysts were inactivated within one minute. At 22°C, 10% of the cysts in the sample taken at 1 minute remained viable. As a result of this, the experimental runs at 22°C were repeated and midpoint samples drawn at 0, 5, 10, 20, 30, 60, 70, 80, and 90 seconds (Figure 3). Ozonation for 90 seconds rendered all of the cysts unable to excyst with a paralleled reduction in viability. Faucet Filter Simple filtration devices directly attached to the faucet are being widely promoted as a means to ensure a "safe" potable water supply. The system chosen for testing was the Pure Water "99" (Associated Mills Inc., Chicago, Illinois). The filter was attached to a faucet specifically installed on the second recycling loop. Giardia muris cysts were inoculated into the recycling water with a final cell concentration of 10 cysts/mL. Water was pumped through the device continuously and the passage of cysts determined using an interdicting membrane filter (1.2 µm pore size). For 3 days, membranes were removed and examined for cysts at 24 hour intervals. After 3 days the interdiction interval time was changed to 7 days to evaluate the effect of prolonged storage between use. Cysts were washed off the filter, concentrated by centrifugation and counted using a hemocytometer.
Figure 3. The percentage of residual viable organisms (y axis) when exposed to ozonation in a batch point of use device for periods of up to 20 minutes (x axis) for the three test organisms: Escherichia coli (thin continuous line), Giardia muris (thick continuous line), and Pseudomonas fluorescens (discontinuous line).
Cysts were recovered from the filter on day 1 but by day 3 no cyst passage was evident, indicating that the filter had retained the cysts. However, after the filters were examined again at day 10, 25 cysts were entrapped in the product water. After a further 7 days of testing, the number of cysts recovered increased to 40. The faucet filter was therefor ineffective at totally excluding cysts from the postdiluvian water. Concurrent bacteriological examinations of the product (premembrane filtration) water performed daily revealed a rapid rise in the bacterial population to a plateau of 200,000 c.f.u./mL on day 5. In this device it could be that a complex interaction was occurring between the biofouling bacteria within the filter and the transient cysts entrapped within the device. General Discussion All of the devices failed to either exclude or destroy the Giardia muris cysts under some circumstances. A variety of potential interactive factors were observed which could effect this including; high cyst loadings, biofouling of the system, screening effects from other suspended particles in the water (be it other cysts, bacteria, chemical precipitates etc.), temperature, mechanical damage or plugging within the device and the method for observing cyst numbers (i.e., recovery efficiency). Most of the studies involved relatively short intensive exposure scenarios which would differ from the actual practices involved in a normal installation on a water system. Any recommendations for the ability of a given device to successfully and continuously remove or destroy Giardia cysts from a potable water supply has to be tempered by the fact that under some conditions the mechanisms can fail and no system exists which will adequately monitor for these occurrences in a economical and efficient
Page 112
manner. Further research is needed in order to more precisely identify the relationships of the various particular interactive factors to the efficiency of exclusion of viable Giardia cysts from a given treatment system. Acknowledgements The authors wish to thank the Canadian Federal Department of Health and Welfare for the provision of a research contract within which this research was performed. In particular, we wish to thank Dr. Richard Tobin of the Department for his support and critical encouragement as the project developed, Erl Jansson for his expertise and advice as the project developed and Mehdi Karamchi who assisted in the maintenance of the cultures and developed alternative methods for the staining of the cysts which will be reported elsewhere. Literature Cited 1. Bingham, A.R., and E.A. Meyer. 1979. Giardia excystation can be induced in vitro in acidic solutions. Nature (London) 277:301302. 2. Craun, G.F.. 1978. Waterborne outbreaks of giardiasis. In: Waterborne transmission of Giardiasis. Jakubowski, W. and J.C. Hoff, (eds.). EPA600/9 79 001. 3. DeWalle, F.B., Jansson, E., and D.A. Carlson. 1983. Inactivation of Giardia by chlorine, U.V. and ozone. Department of Environmental Health, University of Washington, Seattle, Washington. 4. Jarroll, E.L., Bingham, A.K., and E.A. Meyer. 1980. Giardia cyst destruction: effectiveness of six small quantity water disinfection methods. Am. J. Trop. Med. Hyg. 29:811. 5. Jarroll, E.L., Bingham, A.K., and E.A. Meyer. 1981. Effect of chlorine on Giardia lamblia cyst viability. Appl. and Environ. Microbiol. 41:483. 6. Lippy, E.C.. 1978. Water supply problems associated with a waterborne outbreak of giardiasis. In: Waterborne Transmission of Giardiasis. Jakubowski, W. and J.C. Hoff, (eds.). EPA600/9 79001. 7. Meyer, E.A., Erlandsen, S.L., and W.S. Radulescu. 1984. Animal models for giardiasis. In: Giardia and Giardiasis. Erlandsen, S.L., and E.A. Meyer, (eds.). Plenum Press. New York. 8. Moore, G.T., Cross, W.M., McGuire, D., Mollohan, C.S., Gleason, N.N., Healy, G.R., and L.H. Newton. 1969. Epidemic giardiasis at a ski resort. New England J. Med. 281(8):402. 9. Morrell, R.. 1985. Impedence effects of biofilm formation on the passage of Giardia muris cysts. Submitted B.Sc. Honours thesis, Biology Department, University of Regina. 10. Pluntze, J.C.. 1983. The significance of giardiasis on water quality standards and water utility practice in Washington state. Presented at the British Columbia Water and Wastewater Association, November 22. 11. Rice, E.W., and J.C. Hoff. 1981. Inactivation of Giardia lamblia cysts by ultraviolet irradiation. Appl. and Environ. Microbiol. 42(3):546. 12. Rice, E.W., Hoff, J.C., and F.W. Schaefer, III. 1982. Inactivation of Giardia cysts by chlorine. Appl. and Environ. Microbiol. 43:250251. 13. RobertsThompson, I.C., and G.C. Mitchel. 1978. Giardiasis in mice. I. Prolonged infections in certain mouse strains and hypothymic (nude) mice. Gastroenterology 75:42 46. 14. Shaw, P.K., Brodsky, R.E., Lyman, D.O., Wood, B.T., Hibler, C.P., Healy, G.R., MacLeod, K.I., Stahl, W., and N.G. Schultz. 1977. A community wide outbreak of giardiasis with documented transmission by municipal water. Annals Int. Med. 87(4):426. 15. Shun, D.L.. 1985. Giardia lamblia and water supply. Journal AWWA Feb.:4047. 16. Visvesvara, G.S.. 1983. Giardiasis: an overview. Illinois Med. J. July:3439.
Page 113
Diatomite Filtration: Why It Removes Giardia from Water Harris G. Walton Manville Research and Development Center, 2500 Miguleto Canyon Rd., Lompoc, California 93436, U.S.A.. In this presentation, the author utilizes scanning electron microscopy (S.E.M.) to document the positive mechanical removal of the cysts of Giardia lamblia. Data are also presented to compare electronic and laser particle size distributions of the cysts with mercury intrusion porosimetry of a typical diatomite filter cake. Finally, a theoretical pore of a sand filter is shown in juxtaposition with the same magnification of cysts and diatomite. These magnified and graphed visual comparisons are intended to demonstrate the efficacy of diatomite water filtration and should provide meaningful references for future researchers.
Introduction Because diatomite filtration usually takes place in an enclosed vessel, and the mechanisms of particle retention are microscopic, it is not possible to observe the actual separation of Giardia cysts from a water stream while it is happening. This paper presents measured and pictorial data that enables the reader to visualize the actual physical entrapment of these cysts by diatomaceous earth as it would occur in a potable water filter Discussion Figure 1 is a scanning electron micrograph (SEM) of Giardia lamblia cysts at 2000× magnification. Unfortunately, these cysts partially collapsed during sample
Figure 1. Scanning electron micrograph of Giardia cysts, 2000x.
preparation, but this has not dramatically changed their physical dimensions. Figure 2 is a plot of cyst particle size made on a Cilas model 715 granulometer. This instrument measures particle size distribution by light scatter of a laser beam. The cysts were preserved in formalin just 20 minutes prior to measurement, so their morphology is unchanged from the viable cyst. The differential curve shows them to be monodispersed between 6 and 15 µm with a median size of 8 µm.
Figure 2. Giardia cyst particle size distribution measured by light scatter with a Cilas granulometer.
Figure 3. Giardia cyst particle size distribution as measured by (Coulter) electrolytic displacement.
Page 114
Cysts from the same sample were then measured on a Coulter counter model Ta which measures particle size electronically by the displacement of electrolyte in an electric field. Figure 3 shows again that the cysts are monodispersed between 5 and 10 µm. Another observation evident in Figures 2 and 3 is the total lack of material below the size range of the cysts. This is credited to the technique of separating "super clean" specimens at Colorado State University, who were the suppliers of the cysts in this study. The principles of diatomite filtration as shown in Figure 4 are well documented in the literature (1,2,6,7) and are beyond the scope of this paper. This is shown here to orient the reader to the relationship of the precoat and the filter cake. Figure 5 is a scanning electron micrograph (SEM) of a section of an actual diatomite precoat magnified 2000×. The grade shown here (CeliteT M 545) is at the coarse end of a range of 11 grades that are available for water filtration. This material has a permeability of 4.8 darcys while the finer grades range down as tight as 0.06 darcys (1). Note that the intersticial spaces have pores ranging from several µm through 50 µm. Note also that individual diatoms have pores within their own structure that are submicron in diameter (1). The entire field of diatoms shown in this SEM is only 45 µm across. Actual precoats of diatomite used in water filters are about 6 mm or 600 µm thick. When one envisions that water flowing through the precoat layer alone, must pass through a maze approximately 12,000 times thicker than that displayed in Figure 5, it can be readily realized that a cyst particle of 8 µm will rapidly be retained by mechanical entrapment in the diatomite.
Figure 4. Schematic of diatomite filtration.
Figure 5. Scanning electron micrograph of a diatomite precoat (CeliteT M 545), 2000×.
The pore size distribution, as measured by mercury intrusion porosimetry, is shown in Figure 6. While this shows that 85 % of the pores are greater than 8 µm, their random distribution and interspersion with the finer pores produces a filter media with retention potential of even submicron particles. In Figure 7, the relationship of the cysts to a precoat surface are shown in a side by side comparison at 2000×. It is again noted that some pores formed between diatomite particles will allow the penetration of a cyst into the surface, but a few µm deeper into the precoat, the cyst will be removed by a labyrinth of finer pores. It has been documented (3,4,5) that diatomaceous earth water filters, using this same coarse grade of diatomite, effectively remove 99.9+ % of Giardia cysts from water containing up to 3.36 × 104 cysts per liter. The reason for this is demonstrated in Figure 8, which is a mix
Figure 6. Pore size distribution of the diatomite precoat CeliteT M 545.
Page 115
Figure 7. The relationship of Giardia cysts to a TM Celite 545 precoat, 2000×.
Figure 8. A Giardia cyst precoat filter cake, 2000×.
ture of cysts and diatomite displayed at 2000×. This could be taken as a theoretical representation of a filter cake; however, in the "real world" the particles would be more tightly compacted together. The frictional drag of the water passing through and around the particles causes a differential pressure which causes the compressible cyst to deform and partially extrude into the diatomite matrix. The cyst is prevented from extruding through the filter, again by the multilayer effect of the fine pore labyrinth. To further guarantee this effect, water filters for Giardia removal are recommended to use twice the thickness of a normal industrial precoat (3,7), i.e. 20 lbs/100 ft2 instead of 10 lbs/100 ft2 as would normally be applied. The literature (4) reports that many surface water sources in mountains have raw water turbidities well below the required 1.0 NTU limit. For this reason, and because the low temperature of the water makes flocculation difficult, some sand filters are operated without the use of alum. In order to put this in perspective with a diatomite filter, which uses no flocculating agents, the following comparison is made with a theoretical pore formed by idealized sand grains having a mean diameter of 0.5 mm. In order to gain the best perspective, this pore is demonstrated in two steps. Figure 9 is the pore imaged at 200× showing how the 0.5 mm diameters form a minimum contact triangle having three equal sides of 130 µm. In Figure 10, the pore is now imaged at 2000× with Giardia cysts superimposed at the same magnification. From this illustration, it can be understood why random entrapment, eddy current and sedimentation are probably the only filtration functions that will remove cysts in this filter if it is operated without a schmutzdecke layer, and passage of some cysts into the finished water is almost a certainty.
Figure 9. Theoretical interstice of 0.5 mm sand grain.
Page 116
Figure 10. Giardia cysts superimposed on theoretical interstice of 0.5 mm sand grain, 2000×.
In this final comparison (Figure 11), the SEM of the diatomite surface is shown to the same magnification as the sand pore. Once more it is dramatically demonstrated why a diatomite filter removes fine turbidity and Giardia cysts so completely. Conclusion By visual comparisons and measured pore versus particle sizes, it has been visually demonstrated that a properly operated diatomite filter will provide an efficient barrier for the removal of Giardia cysts from potable water. Acknowledgments The author gratefully acknowledges the assistance rendered by Kathy Smith, Electron Microscopy, and Hubert Attaway, Enzyme and Microbe Immobilization Sections of Manville R&D.
Figure 11. Diatomite surfaces superimposed on theoretical interstice of 0.5 mm sand grain, 2000×.
Literature Cited 1. Cain, C.W., Jr.. 1984. Filter aid, use in filtration. Encyclopedia of Chemical Processing and Design 21:348372. 2. Cummins, A.B.. 1942. Clarifying efficiency of diatomaceous filter aids. Industrial and Engineering Chemistry 34:403414. 3. Hendricks, D.W., et al.. 1982. Removal of Giardia lamblia from water supplies. Environmental Engineering Technical Report No. 5836824. 4. Hendricks, D.W., et al. 1984. Filtration of Giardia cysts and other substances. Volume 1: Diatomaceous earth filtration. E.P.A. project summary No. EPA 660/S284 114. 5. Logsdon, G.S., et al. 1983. Control of Giardia cysts by filtration: The laboratory's role. A.W.W.A. water quality conference, Norfolk, Virginia. 6. Purchase, D.B., 1967. Industrial Filtration of Liquids. Leonard Hill books, London, England. pp. 9098. 7. Svarovsky, L., et al. 1977. Solidliquid Separation. Butterworth Publishing, LondonBoston. pp. 196198.
Page 117
Small Water System Improvements For Giardia Removal —A Case Study Michael R. Alberi, Steven J. Quail*, and Robert A. Kruse Woodward and Curran Inc., 41 Hutchins Dr., Portland, Maine 04102, U.S.A.. A case study of a treatment system installed to correct a known, persistent Giardia problem in the water supply of a south central resort town of population 2000 is presented. The results of field pilot studies forming the basis for system configuration and selection of design criteria for a modified direct filtration process are presented. Summaries of daily operating data are presented for the past three years showing plant performance during various raw water quality conditions. Plant effluent turbidity has averaged 0.054 NTU during this period with alum and anionic polymer dosages averaging 2.3 mg/L and 0.48 mg/L respectively in the winter and 11.2 mg/L and 0.56 mg/L respectively in the summer. The winter and summer operating conditions representing two separate coagulation regimes; adsorbtiondestabilization during the summer, and combination sweep floc adsorbtion in the winter are presented on pH vs log alum dose diagrams. Data on Giardia cyst counts on the raw water and finished water are presented which confirm their continued presence in the source and absence in the finished water. Operating strategies, and operating costs for labor, energy and chemicals are also presented.
Introduction In the spring and summer of 1980, an outbreak of giardiasis occurred in the city of Red Lodge, Montana, a small resort community located at the base of the Bear tooth Range of the Rockies in south central Montana. Over the following year, about 860 cases of giardiasis were confirmed by local health professionals. An on site investigation of the outbreak by the Emergency Response Team of the U.S. EPA and the Center for Disease Control led to implicating the antiquated water system and untreated surface water sources as the cause of the problem although sampling of the raw water did not identify cysts. The rate of confirmation of new cases correlated well with spikes on the turbidity of the mountain stream serving as the source of supply. Figure 1 shows that the confirmation of new cases lagged each rise in turbidity by the normally expected 6 to 22 days incubation period. The existing water system consisted of a gravity supply, coarsescreened surface water intake with a hydraulic elevation about 300 feet above the city, and a constant rate chlorinator feeding water directly into the distribution system through two miles of transmission line. Neither distribution system storage nor individual service water meters existed. These circumstances resulted in grossly excessive per capita use habits, high peaking factors, widely varying chlorine residuals in the system, excessive service pressures in the lower end of the single service zone, and taste and odor problems in the fall caused by the action of chlorine on leaves taken into the system and deposited in the pipes. This system lacked the necessary barriers against waterborne, diseasecausing agents. Remedial actions taken for an interim period until implementation of permanent comprehensive corrective measures could be undertaken included the addition of flow paced chlorination on the intake and the imposition of a boil order for all enterprises servicing the public. One of the challenges was to implement an economically sized water system to serve present water users and a reasonable amount of growth. Since the system was currently unmetered and customers could use all the water they wanted, the city had to critically review how the system was managed and financed. A master plan was quickly prepared as a basis for designing, constructing, and financing system improvements that would provide treated water free from Giardia cysts, reduce distribution system pressures to reasonable
Figure 1. Correlation of spikes in raw water turbidity with development of confirmed cases of giardiasis. * Corresponding author.
Page 118
values, provide distribution system storage, and provide service meters as an equitable means of apportioning the cost of system construction and operation. These objectives were accomplished by adding a new water treatment plant, changing the distribution system to 2 pressure zones each regulated by concrete storage reservoirs with a pressure reducing valve (PRV) station separating the two. Water meters were installed on all services, and undersized distribution mains in the lower pressure zone were upgraded. Permanent financing for the system improvements was arranged through grants and loans from the Farm Home Administration, Community Development Block Grant programs, and interim financing through the sale of bond anticipation notes. These improvements were implemented in two phases. In the first phase the distribution system modifications were made to separate the 2 pressure zones, and a lower zone 0.75 MG (million gallon) storage tank was constructed and connected to an existing shallow 1 MGD well located in the lower distribution zone. This well had been developed in the early 1960's to augment low pressure in the central business district during fire flow conditions. The upper zone created by this separation, which at that time served less than 50 residential users, continued to be served from the surface water source without filtration until the treatment facility could be put into operation. During the construction of the first phase improvements, pilot studies were conducted to develop design criteria for the treatment plant. The decision to implement a surface water treatment plant instead of developing wells as the sole source of supply was based on the following considerations: 1) The city had the first water right on this stream which would have been relinquished without compensation if its use was discontinued, 2) The source feeds the service by gravity, 3) The groundwater aquifer available for wells is shallow and difficult to protect due to the presence of existing development with on site sewage disposal systems, and 4) The surface source water has a low turbidity which could be treated economically by direct filtration. Treatment Plant Design Criteria Development A review of the literature at the time that the pilot study was conducted revealed that the treatment objectives for Giardia cyst removal by rapid sand filtration should provide a finished water turbidity of 0.10 NTU or less at all times and provide 2 hours of contact time for chlorine disinfection (1). A 0.5 gal/min pilot plant with a mixed media filter obtained from Neptune Microfoc Inc. was set up at the stream intake. The pilot plant operated over an 8 month period from November 1981 to July 1982 to evaluate variations in stream conditions. The average raw water conditions encountered are presented in Table 1. Jar tests were conducted using the traditional Phipps and Bird Stirrer protocol to establish chemical dosages for the pilot unit, which was operated for a brief initial period in a complete treatment mode and then as a direct filtration unit. The jar tests were used as a qualitative measure of floc development by a specified amount and type of TABLE 1. Source characteristics Turbidity (NTU)
0.22.5
Temperature (°F)
3243
Alkalinity (mg/L CaCO3) pH
30 77.3
conditioner, coagulant and flocculant aid added, their sequence of addition, and the apparent strength of the floc. This was determined visually be resuspending the settled floc by rapid mixing for 15 seconds. If the floc completely sheared or if no significant floc redeveloped after rapid mix, the floc was considered fragile and unsuitable for direct application to the filter. This procedure revealed two problems with the pilot plant hardware that were corrected by physical modifications shown in Figure 2. Jar testing revealed that the cold low turbidity water was best coagulated by adding the chemicals sequentially to 1) provide a nucleation site for floc development, 2) provide a buffer to maintain a constant pH, 3) provide the coagulant, alum in this case, and 4) add anionic polymer as a flocculant aid. This was accommodated by constructing a 4 cell series rapid mix unit for the addition of the four reagents. The early filter runs using the paddle flocculator with and without the sedimentation unit were unable to produce a filtered water with a turbidity less than about 0.3 NTU. Since the filter was fed with a centrifugal pump located between the flocculation unit and the filter, it was surmised from the jar test floc strength tests that the flocs were destroyed by the pump. This was overcome by eliminating the paddle flocculation and sedimentation units, and pumping directly from the rapid mix train through a 1 inch coiled polyethylene tube to the filter so as to achieve a velocity gradient of 30 sec1 and a detention time of 10 minutes. This arrangement proved to be an adequate flocculator. These two modifications provided the control and flexibility required for the treatment train to produce a filtered water of less than 0.1 NTU. During the filter trials, two operating regions, as defined on the log alumpH diagram previously published by Amirtharajah (2), were evaluated. The sweep floc region for this water covers the region bounded by pH 6.8
Figure 2. Pilot plant schematic.
Page 119
to 7.3 and alum dose 12 mg/L to 25 mg/L. Operating in this region would result in relatively short filter runs of about 8 hours and thus would require sedimentation prior to the filter. The process also operated very well in the absorptiondestabilization region characterized by the region bounded by pH 6.8 to 7.3 and alum dose 12 mg/L to 7.0 mg/L. Plots of turbidity removal and head loss for typical pilot filter runs are presented in Figure 3 which show the need for filtering to waste after backwash, or after filter shutdown and restart. The following design recommendations resulted from the pilot study: 1. Chemical addition during rapid mix in the following order: (i) bentonite to increase the raw water turbidity to 1.0 NTU (ii) soda ash to maintain a pH range of pH 6.8 to 7.2 (iii) alum at 12 mg/L for adsorbtiondestabilization, 25 mg/L for combined sweep coagulant operating regimes (iv) anionic polymer as a flocculant aid at 0.5 to 1.0 mg/L. 2. Rapid mix Gt of 33,000. 3. Flocculation G of 25 to 35 sec1. 4. Flocculation HRT of 10 minutes. 5. Mixed media filters, constant rate, with surface wash. 6. Continuous turbidity monitoring for process control and to initiate and stop filter to waste. 7. Do not reclaim backwash water due to concern about Giardia cyst contamination. The design of the full scale plant was based on operating in the adsorbtiondestabilization region because it is more economical than operation in the sweep region. It was also desired that the plant be capable of operating in the combination (sweep and adsorbtion) region if future raw water conditions warranted. This would necessitate a sedimentation device in addition to a flocculator. During equipment selection, a contact flocculator was selected as
Figure 3. Water treatment plant general arrangement. TABLE 2. Water treatment plant design criteria DESIGN CRITERIA
1)
Plant design capacity
1.4 MGD
2)
Finished water turbidity
0.1 NTU
3)
Free chlorine residual, in system extremes
PLANTSYSTEM DESCRIPTION
1)
Plant influent 3 low service pump
Vertical turbine type
LSP1
1000 GPM @ 35 FT TDH
LSP2
1000 GPM @ 35 FT TDH
LSP3
550 GPM @ 35 FT TDH
2)
Chemical feed
3 dry feed units
1 dry feed unit, stand by
1 liquid feed unit
1 liquid feed unit, stand by
Feed rates Bentonite
Soda ash
6 mg/L
Alum
12 mg/L
Anionic polymer
1.0 mg/L
3)
Rapid mix
4 chamber static mixer, 12''
4)
Filtration
Filter loading rate, design media
Contact flocculator
140 SF
Filter
280 SF
Washwater troughs, per filter
Filter backwash
Rate, design
240 HP vert. turb. pumps
Filter media surface wash
Rate
90 GPM @ 230 FT
2HP subm. turb. pumps
90 GPM @ 230 FT
Flocculator backwash
Rate
Flocculator air scour
Rate
2 positive disp. blowers
5)
Disinfection
2 chlorinators
Rate, max. ea.
20 lbs/day
6)
Clearwell Capacity
137,500 gal
7)
Reservoir Capacity
253,000 gal
8)
Emergency generator
0.5 mg/L
1.0 mg/L
3.6 gal/min/ft2
1
15 gal/min/ft2 2100 GPM @ 42 FT TDH
7.2 gal/min/ft2
2 SCFM/SF 140 SCFM @ 9 PSIG
80 KW
a compact device that could perform both functions. The full scale plant design criteria is presented in Table 2. The general arrangement of the plant is shown in Figure 3. The plant is fully automated and is controlled by a programmable controller. Plant start and stop functions are initiated by low and high level signals respectively from the upper service zone storage reservoir build integral to the filter plant. Plant Performance The plant started up in March 1984, and has performed well without mechanical problems. Plant water production in terms of a seven day moving average presented in Figure 4 shows the decrease in daily production rate over time due to changing use habits driven by metered service. It also shows the decrease in the difference between influent water pumped to the
Page 120
Figure 4. Plant water production as a 7 day moving average.
Figure 5. Daily raw water and finished water turbidity.
Figure 6. Raw water turbidity probability distribution.
Figure 7. Finished water turbidity probability distribution. TABLE 3. Frequency distribution of raw and finished water turbidity values. Raw Water Turbidity <
Finished Water Turbidity
Frequency
Probability
<
Frequency
Probability
0.00
0
0.00%
0.00
0
0.00%
0.25
273
33.96%
0.02
6
0.68%
0.50
354
77.96%
0.03
49
6.68%
0.75
44
83.99%
0.04
166
27.17%
1.00
22
86.19%
0.05
206
52.66%
1.25
21
88.81%
0.06
201
77.48%
1.50
19
91.17%
0.07
96
89.36%
1.75
4
91.67%
0.08
47
95.17%
2.00
15
93.53%
0.09
13
96.72%
2.25
9
94.65%
0.10
12
98.14%
2.50
10
95.90%
0.11
1
98.21%
2.75
2
96.14%
0.12
2
98.39%
3.00
7
97.01%
0.13
1
98.51%
3.25
0
97.01%
0.14
4
98.95%
3.50
2
97.26%
0.15
1
99.01%
3.75
0
97.26%
0.16
2
99.20%
4.00
7
98.13%
0.17
3
99.50%
4.25
2
98.38%
0.18
1
99.63%
4.50
2
98.63%
0.19
1
99.69%
4.75
0
98.63%
0.20
1
99.75%
5.00
2
98.88%
0.21
1
99.81%
5.25
1
99.00%
0.22
0
99.81%
5.50
4
99.50%
0.23
1
99.88%
5.75
0
99.50%
0.24
0
99.88%
6.00
2
99.75%
0.25
0
99.88%
6.25
1
99.88%
0.26
0
99.88%
6.50
0
99.88%
0.27
0
99.88%
6.75
0
99.88%
0.28
0
99.88%
7.00
0
99.88%
0.29
0
99.88%
7.25
0
99.88%
0.30
1
99.94%
7.50
0
99.88%
7.75
0
99.88%
8.00
0
99.88%
8.25
0
99.88%
8.50
1
100.00%
0.31
1
100.00%
treatment plant and flow delivered to the system. This difference represents both backwash water and filter to waste upon filter start up. The reduction in this difference over time is due to a general reduction in water throughput, and optimization of the operation through operator familiarity with the treatment system and changes in the source on a daily and seasonal basis. Data characterizing raw water and finished water turbidity are presented on Figures 5, 6 and 7, and Table 3. These data show that the plant has met the 0.10 NTU criteria for finished water turbidity 98.14% of the time over a period comprised of 809 days of data. With the exception of 7 days in the spring of 1985 during a period of high influent turbidity, a 7 day period in the fall of 1985 when the chief plant operator was away, and 2 days in the spring of 1986, the filter turbidity has been equal to or less than 0.10 NTU. Recent research by AlAni et al. (3) indicates that achieving a finish water turbidity of 0.10 NTU or less, or achieving at least a 70% turbidity
Page 121
removal are adequate guidelines for judging the effectiveness of Giardia cyst removal. Table 4 characterizes the turbidity removals for those days that a finished water turbidity of 0.10 NTU was not achieved. For the data presented and the two criteria listed by reference 2 for cyst removal, the days April 1, 1985, September 25, 26, 28 and 29, 1985, and March 8, 1986 represent situations in which less than optimum cyst removal could have occurred if cysts were present in the raw water. The turbidity removals presented in this table are based on raw water turbidity. On 4 of the days that the 70% removal criteria was not met, bentonite was being fed which would have made the effective raw water turbidity about 1 NTU. Thus the effective removal for these 4 days could be considered to be greater than 70% leaving only two days out of the period of record when less than optimum cyst removal may have occurred. It is somewhat conjectural that the removal of artificially created turbidity applies to this case since research has hot been conducted to date that would directly confirm this. During an 18 day period in July and August 1984, the plant was operated at 150% of the design surface loading. The finished water turbidity during this period ranged between 0.04 and 0.08 NTU. During the course of a typical year the plant operates predominantly in the adsorbtiondestabilization region of the log alumpH diagrams, with some shifting into the combination region. Figures 8, 9, 10, 11 and 12 show the range of operation for the various seasons. A transition occurs in the spring and fall in terms of soda ash and bentonite addition. Soda ash, but no bentonite is fed from April through September. The balance of the year bentonite, but no soda ash, is fed. This can be explained by a change in alkalinity in the raw water from about 30 mg/L in the winter to about 15 to 20 mg/L in the spring and summer, and the fact that the winter water has fewer TABLE 4. Turbidity removal efficiency for dates when finished water turbidity exceeded 0.10 NTU
Turbidity NTU
Date
Raw
Filtered
% Removal
Bentonite Dosage mg/L
April 10, 1985
0.53
0.13
74
0.9
April 11, 1985
0.40
0.14
65
0.9
May 18, 1985
2.20
0.12
95
0
June 15, 1985
5.80
0.15
97.5
0
June 17, 1985
6.20
0.18
97
0
July 8, 1985
1.50
0.15
90
0
July 9, 1985
1.10
0.16
85
0
Sept 23, 1985
0.65
0.23
65
0
Sept 24, 1985
0.47
0.14
70
0
Sept 25, 1985
0.51
0.21
59
0
Sept 26, 1985
0.25
0.14
44
1.1
Sept 27, 1985
0.60
0.17
72
1.1
Sept 28, 1985
0.25
0.17
32
1.1
Sept 29, 1985
0.34
0.17
50
0.8
March 8, 1986
0.26
0.12
54
1.1
June 18, 1986
5.00
0.12
98
0
Figure 8. Operating region for January.
Figure 9. Operating region for March.
Page 122
Figure 10. Operating region for June.
Figure 11. Operating region for August.
Figure 12. Operating region for October.
and the difference between the raw and finished water total solids values. In the winter the difference between raw and finished water total solids is in the order of 70 mg/L and in the spring and summer it is about 200 mg/L. The operating region of the two raw water temperature extremes encountered for this water source are presented in Figures 13 and 14. The data points were plotted on the log alum pH diagram as originally published by Amirtharajah. Since some operating points plotted on this diagram fall outside of the adsorbtiondestabilization region in the restabilization region, a corresponding adjustment in the boundary between these 2 regions should be made for this water source. Data characterizing the overall operation of the plant with respect to water production, chemical dosages, turbidity, and total solids are presented in Table 5. The State of Montana Department of Health and Environmental Sciences has conducted a sampling and analysis program on two separate occasions to measure the presence of Giardia cysts in the raw and finished water. This data presented in Table 6 shows that cysts and large amounts of Giardia size material are present in the raw water while the finished water contained little suspended matter and no cysts were found. The contact flocculator clarifier and filter backwash was sampled, but contained too much flocculated matter to visually identify cysts. Backwash The contact flocculator is backwashed with air and coagulated influent on the basis of head loss across the media. The sand filter calls for a backwash with
Page 123 TABLE 5. Statistical data for treatment plan operation FLOW DATA (gal/day)
Influent
Backwash
Clearwell/Res
802
800
800
Maximum
2,006,000
560,000
1,699,000
Minimum
472
0
49,000
Average
788,297
22,831
669,703
Std. Dev.
349,325
33,200
303,781
Number
CHEMICAL DOSAGES (mg/L)
Alum
Soda Ash
Bentonite
Number
792
775
733
Polymer 801
Maximum
17.4
19.0
1.8
27.0
Minimum
0.5
0.0
0.0
0.3
Average
6.1
3.5
0.7
0.6
Std. Dev.
3.4
4.8
0.5
0.9
TURBIDITY (NTU)
Raw
Floc 1
Floc 2
Eff 1
Number
804
804
801
809
807
Maximum
8.50
30.00
21.00
0.23
0.31
Minimum
0.10
0.13
0.10
0.01
0.02
Average
0.61
0.40
0.35
0.056
0.057
Std. Dev.
0.91
1.28
0.74
0.02
0.02
Eff 2
TOTAL SOLIDS (mg/L)
Raw
Floc 1
Floc 2
Eff 1
Eff 2
Number
48
49
49
49
49
Maximum
450
55
70
41
50
Minimum
19
3
7
5
7
Average
121
28
29
16
16
Std. Dev.
97
11
12
9
9
finished water automatically upon reading 8 feet of head loss. As a practical matter, the operator initiates backwash once per day or every 24 hours of operation to keep the filter bed fresh. The average filter backwash use of 22,381 gal/day equates to backwashing a single filter basin each day on the average. The difference between the amount of raw water pumped to the plant and the sum of the backwash water (for filters) and water produced is about 95,800 gal/day on the average for the period of record. This difference is water filtered to waste, and unfiltered water used to flush the contact flocculator units. This amount of water has been reduced to 77,000 gal/day in 1986. Cost Data The annual costs for chemicals, power, labor, and maintenance are presented in Table 7. The project cost of construction of the water plant, upper zone, storage, and low service pump station of $1,200,000 U.S. in May 1983 has been updated to first quarter 1987 and amortized at 10% interest over 20 years. Based on a current annual delivery of 242 MG to the distribution
Figure 13. Operating region for temperature range of 33 to 34°F.
Figure 14. Operating region for temperature range of 44 to 51°F.
Page 124 3
system, the cost of production is $0.77/1,000 gal or $0.58/100 ft . These costs are summarized in Table 7. TABLE 6. Giardia cyst sampling analysis and results
12 9 1986
Turbidity
12 11 1985
Raw
Filtered
Raw
Filtered 0.02
0.2
0.03
0.2
# cysts found
3
0
0
0
Gallons filtered
400
570
900
1390
Small particle debris
+++
++
+++
++
Large amorphous debris
++
++
+
0
Cellular plant material
+++
+
+++
0
Diatoms and algae
++
+
+
+
Protozoa
++
+
+
0
Insects
+
0
+
0
Nematodes
0
+
0
0
Chemicals:
Nalco 8184
0.75 MG/L
Alum
6.0 MG/L
Bentonite
1.1 ppm
Turbidity after
bentonite addition
1.1 NTU
0 none found + rare ++ moderate +++ heavy
TABLE 7. Water production costs
Operation & Maintenance
$/year
$/1000 gal
Power
10,623
0.044
Chemicals
7,168
0.029
Labor
16,425
0.068
Total
34,216
0.141
Construction costs 1st QTR 1987 Amortized 20 years at 10%
1,304,000
153,205
Cost per 1000 gal produced
0.630
Current cost of water produced
0.771
Literature Cited 1. Logsdon, G.S., and F.B. DeWalle. Filtration as a barrier to passage of cysts in drinking water. Joint publication of U.S.E.P.A., Breidenback, Environmental Research Center and the University of Washington. 2. Amirtharajah, A., and K.J. Mills. 1982. Rapid mix design for mechanisms of alum coagulation. JAWWA 774(4):210. 3. AlAni, M.Y., et al. 1986. Removing Giardia cysts from low turbidity water by rapid rate filtration. JAWWA 78(5):66.
Page 125
Inactivation of Giardia lamblia Cysts from a Surface Water by Oxidation with Ozone Carl Nebel*, Anthony Lally, Thomas Bosher, J. William Hmurciak, Linda Hmurciak and Dorothy A. Breen PCI Ozone & Control Systems Inc., 1 Fairfield Crescent, West Caldwell, New Jersey 07006, U.S.A.. Giardia lamblia found in surface water can be inactivated by oxidation with ozone. Ozone is manufactured on the site from electric power, air and cooling water. The ozoneair mixture is dispersed in the water via fine bubble diffusers. Ozone dissolved in the water inactivated Giardia cysts via oxidation. Concurrently ozone also oxidizes colour, taste, and odour found in the water. The rate of oxidation of Giardia lamblia by ozone is faster than chlorine and chlorine dioxide. A theoretical discussion of microbial inactivation will be presented. The engineering design, startup and operation of a full scale facility will also be documented.
Introduction The Problem The Town of North Andover Massachusetts has for decades obtained its water from nearby Lake Cochichewick. Two water pumping stations supplied water into the distribution system after chlorination. Although the water was plagued with taste, colour and odour problems, it was generally accepted as biologically safe after chlorination. The taste and odour problems were mainly due to chlorination of humic materials found in this water source. These humic acids imparted a tan colour to the water and also are the result of a rather high chemical oxygen demand (COD) which is in the range of 1220 mg/L. The humic materials served as a food source to a biofilm composed of nonpathogenic microbes in the water distribution system. Attempt to control the growth of this biofilm during the summer months with chlorine and/or chlorine dioxide were unsuccessful. During the winter months, the biofilm was completely controlled by employing additional chlorination points in the distribution system. In part, the success of this effort can be attributed to lower chlorine demand due to the use of ozone and to the fact that cooler water will hold the chlorine residual longer. The Massachusetts Department for Environmental Quality Engineering found that a higher than normal number of the town's citizens were subject to Giardiasis and therefore ordered that the Town of North Andover require that all water for human consumption be boiled for five minutes prior to use. The water source, Lake Cochichewick, was tested and 3 to 4 cysts were found in a 300 gallon sample of water. The total coliform count in the lake water was in the range of 10 to 700/100 mL. Solution to the Problem Initial consideration for the control of giardiasis was given to chlorination. Increasing the chlorine dosage level should result in adequate control. A search of the literature indicated that chlorination was not successful in inactivating Giardia lamblia cysts (1). It was also found that G. lamblia cysts become more resistant to chlorine as the water temperature decreases (2). North Andover Massachusetts is located in a geographic area where the water temperatures drop substantially, therefore consideration for the use of chlorine was terminated. Sproul and coworkers (3) found that ozone is very effective in oxidizing G. lamblia cysts at both 25°C and 5°C. As expected, the inactivation of the cyst requires more ozone at a lower temperature than at higher temperatures. (Figure 1 and 2). In the above studies, Sproul et al., oxidized G. lamblia cysts having a concentration of 10,000 cysts/mL with
Figure 1. Inactivation of G. lamblia cysts by ozone at pH 7 and 5°C * Corresponding author.
Page 126
Figure 2. Inactivation of G. lamblia cysts by ozone at pH 7 and 25°C.
ozone at two different temperatures and three different ozone concentrations each. The excystation was determined by the microscopic method. The product of ozone concentration times the contact time in minutes (C × t) for 99% inactivation is 0.17 and 0.53 mgmin/L at 25°C and 5°C respectively. Table 1 shows that G. lamblia cysts are very resistant to chlorine when compared to poliovirus and E. Coli. G. lamblia is three times more resistant to ozone at 5°C than it is at 25°C. In the same temperature range G. lamblia becomes eight times more resistant to chlorine. It is for this reason that ozone was chosen over chlorine to control G. lamblia for North Andover. The use of higher concentrations of chlorine during the cold months in this type of water would have created more taste and odour problems and would have increased the trihalomethane concentration. It was desired that the time for implementing any solution to the Giardia problem be kept at a minimum because the town wanted the state imposed water boil order to be lifted as soon as possible. Before we can address the engineering aspects of utilizing this oxidant, an understanding of ozone and its generation is in order. Ozone Ozone is a low molecular weight molecule (M.W.48) which is composed of three oxygen atoms that are chemically arranged in a chain. The bond angle between the oxygen atoms is 116°, hence the chain is in the shape of a triangle. Ozone, (O3) an allotrope of oxygen (O2), is composed of the same atoms combined in a different form. The property of ozone for oxidizing microorganisms is attributed to the fact that it is the second most powerful oxidant known (2.08V). Only fluorine exceeds ozone in its oxidation potential (2.87V) and chlorine is (1.34V). The high chemical reactivity of ozone is related to the fact that it possesses an unstable electron configuration which requires it to seek electrons from other molecules. During its reaction with other molecules, ozone is destroyed. The end result of ozone oxidation of organic molecules, such as those found in microorganisms, is carbon dioxide and water. Ozone, like oxygen, is a colorless gas. Unlike oxygen, ozone is difficult to obtain in pure form. Commercial ozone is formed in concentrations of two percent by weight from air. Before the oxidation of microbes in water takes place, the ozone dispersed in air must be transferred from air into water. This implies that a means of transferring ozone into water must be incorporated into this system. The ozone contacting device will be discussed later. Ozone Generation Ozone is an unstable molecule which slowly decomposes back to oxygen from which it was made. At ambient temperatures, the halflife of ozone in air is approximately fourteen hours. Because ozone is not completely stable, it cannot be purchased as a compressed gas. It must be generated at the site of the application and used shortly after its regeneration. This onsite generation circumvents the necessity of having to transport, store and handle a compressed gas. Commercial ozone is generated by accelerating electrons between two electrically charged electrodes. When an electron is propelled to a high velocity whose energy is in the range of 67 eV, an interaction between the electron and the oxygen molecule takes place to disassociate the oxygen molecule into two oxygen atoms (4). O2 + High Energy Electron > 2 O + Low Energy Electron The oxygen atoms formed are a very reactive species and react almost immediately with oxygen molecules to form ozone. O + O2> O3 The net reaction is: 3 O2 > 2 O3 H° = 34.61 kcal/mole TABLE 1. Comparative concentration time data for 99% inactivation.
G. lamblia Cysts
E. Coli
Concentration × Time
Ozone
25°
0.17
5°
0.53
25°
15.00
5°
125.00
20°
0.08
5°
0.22
Chlorine
5°
2.00
Ozone
1°
0.02
Chlorine
5°
0.04
Chlorine
Poliovirus I
Temperature °C
Disinfectant
Ozone
Page 127
Figure 3. Ozone Generator Electrode Assembly double fluid cooled.
Materials and Methods The Ozone Generator The ozone generator is an electronic device which accelerates electrons in the presence of very dry air. A silent corona discharge is produced between two charged electrodes. The corona discharge accelerates electrons which in turn disassociates oxygen found in air molecules into oxygen atoms. Alternating current must be employed when ozone is generated in the corona discharge. If a direct current were employed, the electron would enter the corona discharge and immediately proceed to the grounded electrode where it would become unavailable for further interaction with oxygen molecules. When alternating current is used, the electron vibrates between electrodes in accordance with the frequency of the alternating current. The higher the frequency, the greater the time the electron will exist in the discharge area. Ozone production should, therefore, be a function of the frequency applied to the high voltage electrode. Generally, ozone production is doubled whenever the electrical frequency is doubled. With the use of properly designed high frequency inverters, modern ozone generators can operate at 2,500 cycles per second. The rate of ozone production can be readily changed through a ten to one turndown by altering the power applied to the electrodes. Figure 3 shows a diagram of the ozone generating electrodes which are contained in an ozone generator. The grounded 321 stainless steel 1/4" thick electrode (A) is placed in the center of the ozone generating module. The electrode is fixed in a vertical position and is cooled by passing potable water through it. The total cooling water requirement for an ozone system capable of producing 1150 pounds of ozone per day is 32 gpm and the temperature rise of the water is from 70°F to 76°F. The cooling water is discarded into the ozone contactor. Surrounding the grounded inner stainless steel electrode (A) is a glass electrode (B). The outer surface of the glass (C) is plated with silver which serves as the high potential electrode. To prevent destruction of ozone on the inner surface of the glass electrode, the outer surface of the electrode is cooled with a nonelectrical conducting fluid which is continuously recirculated in a closed loop system. The heat which is removed by the cooling fluid is transferred to water in a shell and tube heat exchanger. Attention is paid to that removed from the ozone generator because heat induces decomposition of ozone back to oxygen. 2 O3 > 3 O2 Dry air is passed through the annular space (D) between electrodes (A) and (B). It is in this area where the corona discharge takes place and where the ozone is produced. The silver glass electrode (C) is charged with 10,000 volts at a frequency of 2,500 cycles per second. The glass electrode, when operated at 10,000 volts, has an infinite life. The air which is passed through the ozone generator must be oilless, particle free and must have moisture removed to a 40°F dew point or lower. The rate of ozone production can be varied from an external 420 mA DC control signal. The start up time of air preparation unit and the ozone generator is approximately one minute. Ozone System Sizing The single most important concern in system sizing is to supply an adequate contact time and ozone concentration to kill Giardia lamblia cysts. The work of Sproul (3) has shown that this should be at least 0.53 mg × min/L. The water at North Andover exhibits a very high instantaneous ozone demand which will decrease the amount of ozone available for microbial control. This ozone demand exists because the water has been shown to have COD values as high as 20 mg/L. Considering the ozone demand of water and the residual required for G. lamblia inactivation, an applied dosage level of 5 mg/L was chosen. At a water flow rate of 2400 gpm, the corresponding quantity of ozone is 150 pounds per day. A further consideration in contactor sizing is the gas to liquid volume ratio. Ozone is generated at a concentration of 2% in air, hence ozone and 98% percent air must be diffused through the water. The quantity of gas applied to the water from 150 pounds of ozone at a 2% concentration is 70 scfm. If the contactor is too small, then the small air bubbles which are formed will coalesce into larger bubbles. This would result in a poor transfer of ozone from air into water. Factors affecting ozone contacting design are documented in the literature (5) as well as the application of ozone to potable water (6). Considering the above three factors, a contactor was constructed which has the following dimensions: 10 feet wide by 20 feet long with a water depth of 16 feet. At a water flow rate of 2400 gpm (3.5 mgd) the contact time was 10 minutes. The contactor consisted of four ports through which the water flowed on a vertical plane (Figure 4). At the bottom of each port, porous
Figure 4. Ozone Contactor.
Page 128 TABLE 2. Public Health impact of study before and after ozone.
Total reported cases of giardiasis
Before mandate made it reportable in 1985
7
Before the installation of ozone (1/1/86 10/1/86)
23
After the installation of ozone from 10/1/86
0
Total reported cases in 1987
5
stone diffusers were placed by which the ozoneair mixture was sparged through the water. In practice, it was found that the ozone residual at the end of the contactor was in the range of 0.9 1.0 mg/L. If we assume that this residual continues to build through the contactor in a linear mode, then effective residual is onehalf of the final residual or 0.5 mg/L. The product of time (10 minutes) and residual (0.5 mg/L) now becomes 5 mg × min/L. This compares favorably with the 0.53 mg × min/L established by Sproul (3) for 5°C water. Although this is nearly an order of magnitude greater than the minimum requirement, it should be noted that this residual will decrease slightly during the summer months when the water is warmer. From inception to system start up, this project took approximately two months. This time includes the engineering, equipment manufacturing and site construction. Construction time of the contactor was decreased by the use of mild steel sheet piling, wood baffles and a concrete top. The use of the mild steel and wood in an ozone contactor is certainly innovative, but was warranted on an economic and timing basis. Results and Discussion The presence of viable Giardia lamblia cysts has been eliminated via ozonation. Total coliform count entering the ozone contactor is in the range of 10 700 mg, whereas the plate count leaving the contactor is zero. When the pumping station employed only chlorine (~4mg/L), the trihalomethane levels were in the range of 8 120 mg/L. The installation of ozone prior to chlorination lowered the trihalomethane concentration to the range of 1.1 to 2.0 mg/L. The true color levels dropped 65 to 95%. The taste and odour levels also showed a substantial decrease. The public health impact of this study before and after ozone are listed in Table 2. Literature Cited 1. Craun, G.G.. 1979. Waterborne Giardiasis in the United States: a review. Am. J. Public Health 69:817. 2. Jarroll, E.A. et al.. 1981. Effect of chlorine on Giardia lamblia cysts viability. Appl. and Environ. Microbiol. 41:483. 3. Wickramanayake, G.B., Rubin, A.J., and O.J. Sproul. 1984. Inactivation of Giardia lamblia cysts with ozone. Appl. and Environ. Microbiol. 48:671. 4. Nebel, C.. 1981. Ozone. Encyclopedia of Chemical Tech. 16:683. 5. Nebel, C.. 1981. Ozone water treatment systems. Water Eng. & Manage. Reference Handbook R77. 6. Nebel, C.. 1981. Ozone treatment of potable water. Public Works 112(1):86, 112(2):68.
Page 129
A Regulatory Agency's Experience with Giardia S. McClure* and I.B. Mackenzie Alberta Department of Environment, 9820 106th Street, Edmonton, Alberta, Canada. T5K2J6. The province of Alberta has experienced two major outbreaks of giardiasis which could be attributed to public water supply transmission. Following these incidences, the province initiated an extensive Giardia monitoring program spanning three years and involving over 700 water samples from more than 40 Alberta communities. Only in three of these samples were Giardia cysts detected. It is thought that the existing monitoring method has many inherent limitations and thus cannot be effective for predicting or controlling giardiasis outbreaks. Until such time as significant improvements in Giardia cyst detection and recovery are availaable, it is felt that more traditional indicators of plant performance will be more effective in ensuring cystfree water.
Introduction The Province of Alberta has experienced two major outbreaks of giardiasis which could be attributed to public water supplies. This resulted in the development of a Giardia monitoring program and a review of the water treatment practices. This paper reviews the efforts made by the Municipal Engineering Branch, Pollution Control Division, Alberta Environment, in dealing with potential waterborne giardiasis outbreaks. The first known outbreaks occurred late in the winter of 1982 in the resort town of Banff, located in the Canadian Rockies. The town has a permanent population of approximately 4,000, with peak season fluxes reaching 20,000. Over 150 cases of giardiasis were diagnosed. The town's water supply came from a reservoir on Forty Mile Creek which was untreated except for chlorination. The Creek watershed is well protected from human activity, however, the Public Health Inspector investigating the outbreak found that beaver had colonized the reservoir, and Giardia lamblia cysts were subsequently detected in samples taken from the reservoir. The town has since converted to groundwater sources and there appears to have been no reccurrence of the disease. The second major outbreak occurred in the City of Edmonton late in the fall of 1982 and the spring of 1983. The local health unit did a followup survey to determine the extent of the infection. Their epidemiological report (Collier and Macdonald 1983) indicated there had been 895 laboratory confirmed cases. The dates on which the number of infections were reported forms a classic epidemic curve generally associated with a single infective source. Data plotted on a city map indicated that the majority of people affected lived in the downtown or University area; an area serviced by the Rossdale Water Treatment Plant. The Rossdale Water Treatment Plant is a conventional water treatment plant utilizing chlorine dioxide as a predisinfectant, chloramines as a postdisinfectant, alum coagulation, sedimentation, filtration, and lime softening. During the incident, the plant was neither predisinfecting with chlorine dioxide nor carrying out softening procedures (Lippy 1984). Although there was no conclusive evidence to implicate the water supply, no other common source for the infection was identified. In the summer of 1983, it became mandatory to report giardiasis in Alberta. Since that time it has been one of the most common diseases in the province with approximately 1300 to 1500 cases per year. However, the data base does not suggest any epidemics attributable to water supplies. Because of the high incidence of giardiasis and the aforementioned epidemics, it was decided that the province should develop a water treatment plant monitoring program. The overall objectives of the provincewide Giardia monitoring program initiated in the spring of 1983, were to: 1. Develop the capacity to predict and control giardiasis outbreaks; 2. Provide a wide coverage of the water supplies for the majority of the Alberta population to evaluate if the organisms were present; 3. Develop and refine sampling and detection techniques; and 4. Monitor water treatment plant methods and performance. Materials and Methods The Giardia sample concentration method as outlined in the 15th edition of Standards Methods (APHA 1980) which utilizes a 7µm wound orlon filter was initially investigated. Analysis were conducted by the Provincial laboratory of Public Health associated with the University of Alberta in Edmonton. This laboratory had had considerable experience in performing protozoan analysis from stool samples for diagnostic purposes. The laboratory subsequently reported that there were difficulties in utilizing the floatation method as outlined in Standard Methods and concluded that they were more successful in centrifuging to concentrate the water from the filter, and then microscopically examining the entire pellet (this resulted in approximately 60 slides per samples). Because of the difficulty and time required to process the orlon filters, and the potential cyst loss, other methods were investigated. * Corresponding author.
Page 130
Figure 1. Schematic of Giardia sampling apparatus.
The Branch next looked at a 5µm polycarbonate membrane. A housing generally used in geological work and known as a geofilter was used, however the membrane quickly clogged and increased pressure on the filter housing. The housing would then open or separate causing a washout of the membrane surface. The laboratory was also experiencing difficulties in conducting their examinations because of the amount of debris, particularly alum sludge that was being filtered out. During the summer of 1984, another concentrating device (Figure 1) was designed. It basically condisted of a 20 µm wound orlon roughing filter followed by three 5µm polycarbonate filters. Initial trials were carried out in the fall of 1984 to determine suitable flow and pressure conditions. The device was calibrated in June of 1985 using Giardia cysts obtained from dog feces and then compared to a membrane filtration device used by the Kananaskis Centre. The Centre's system consisted of the same membrance with no prefiltration device. The filtration devices were found to be generally comparable. However, the Kananaskis laboratory obtained higher recovery rates using the zinc sulphate flotation method compared to the Provincial Laboratory's method of examining the centrifuge plug. For the 1985/86 program, sampling was conducted using the Alberta Environment divice and the Kananaskis Laboratory methodology. Discussion All sampling was conducted on treated municipal surface water supplies. Larger municipalities were sampled on a weekly basis, and smaller municipalities on a monthly basis. During the 1983/84 and 1984/85 seasons, the samples were collected by Environment staff and analysis were performed by the Provincial Laboratory of Public Health. During the 1985/86 season, some of the municipalities were provided with sampling devices and required to take their own samples. These were then sent to the Kananaskis Laboratory for analysis. A major problem encountered in the sampling for Giardia cysts was plugging of the membrane by debris. Not only did this affect the concentration procedure but it also greatly hampered cyst identification. Alum sludge was the major interfering agent but other organic and inorganic debris including algae and nematodes were frequently encountered. Occasionally, air was entrained in the influent filter water and caused it to bind off. The most significant problem was that of low cyst recovery rate. The orlon filters generally had recoveries of less than 10% while recoveries for the polycarbonate membranes ranged from less than 10% to more than 40% depending on the influent cyst concentration. The recovery rates were verified several times through calibrations using both live organism and polystyrene beads. Although the low cyst recovery is considered a serious shortcoming, the techniques used were considered comparable to those practised elsewhere. Table 1 presents the results obtained from the threeyear program. For this period over 700 samples were examined from concentrating more than 450,000 L of water. From this, three possible Giardia cysts were detected. Followup intense sampling at the municipalities where they were detected revealed no further cysts, nor was there any increase in giardiasis cases reported by the local health units. Conclusions 1. The organism could not be detected in significant numbers in the treated water tested. 2. Using current procedures, the Giardia monitoring program would not be effective in predicting or controlling the outbreak of giardiasis. The tests are time consuming and expensive for the type of results they produce. 3. The sampling and analysis techniques developed are thought to be at least comparable to those practised elsewhere. 4. The process of microscreening water is also a useful indicator of the efficiency of operation of water treatment plants. This is indicated by the amount of organisms, sludge, and debris that was accumulated on the filters. Alberta Environment will continue to maintain a monitoring capability for Giardia, but will not conduct a sampling program of the magnitude of the 1983/84 and 19844/85 season. TABLE 1. Alberta Environment Giardia Monitoring Program. Number Communities Monitored
Total* Water Filtered (L)
Number** Samples
Potential*** Number of Cysts
1983/84
29
316.0 × 103
321
0
1984/85
40
79.4 × 103
337
2
1985/86
12
66.7 × 103
66
1
Year
* Using various filter ** Varying sample sizes *** Unconfirmed beyond visual identification
Page 131
The Department will embark on a program to encourage operators to improve their water treatment practices. Initially this will involve concentrating on the proper use of coagulants using turbidity as the indicator. Literature Cited 1. APHAAWWAWPCF. 1980. Standard methods for the examination of water and wastewater. 15th Edition, Washington, D.C. 2. Collier, M.K. and P. Macdonald. 1983. Giardiasis in Edmonton. Edmonton Local Board of Health, Edmonton, Alberta. 3. Lippy, E. 1984. Review of treatment practices. Rossdale Water Treatment Plant, Alberta Environment, Edmonton, Alberta.
Page 133
Effects of Chlorine on the Ultrastructure of Giardia Cysts M. Neuwirth*, P.D. Roach, J.M. BuchananMappin and P.M. Wallis Alberta Environmental Centre, Vegreville, Alberta, Canada, TOB 4L0. Giardia muris cysts were exposed to chlorine concentrations of 0, 1.3, and 4.3 mg/L for 10, 20, and 30 minutes and at 10.5 mg/L for 90 minutes. Cysts were recovered from the experimental beakers by concentration on 5µm Nuclepore membranes and centrifugation. Samples were split; half were preserved in 5% glutaraldehyde for TEM examination and the other half were subjected to viability assays by in vitro excystation. After a buffer rinse, aliquots were fixed and embedded for TEM. Control cysts were fixed without treatment. Transmission EM showed that cysts treated with 0 mg/L chlorine for 10, 20 and 30 minutes appeared ''normal". They contained 2 or 4 nuclei, ribosomes, rough ER, axonemes, microtubules, remnants of ventral discs, and vacuoles. The cysts exposed to chlorine exhibited varying degrees of morphological changes. These include some cyst wall deterioration, ruffling of the plasma membrane and progressive granularity of the cytoplasm. Nuclei and flagella appeared unaffected in most cysts studied. These ultrastructural changes are correlated with viability as determined by excystation.
Introduction Waterborne giardiasis has been well established as an important worldwide health problem. Transmission of Giardia cysts is frequently via public water supplies. The efficacy of chlorination for the inactivation of human infective Giardia duodenalis cysts is a matter of great concern to those responsible for water treatment plants. In spite of the number of studies on effects of halogens on cyst viability (1,4,5,7,10), there is no information dealing with ultrastructural changes in cysts inactivated by chlorine. In this study, the effects of chlorine on the morphology of Giardia muris cysts were observed. G. muris cysts were used since they are similar in morphology to the human G. duodenalis cysts but are noninfective to humans. Materials and Methods Cysts were separated from feccal material collected from mice by flotation on 1M sucrose (14). Chlorine stock solutions were prepared by adding sodium hypochlorite to distilled water in 20 L quantities and were allowed to equilibrate for 24 hours. Exposure medium consisted of nonchlorinated tapwater at pH 8.2 to which sufficient stock was added to provide a free residual chlorine concentration of 0, 1.3, 4.3, and 10.5 mg/L. All tests were performed at 6°C on a stirring apparatus which stirred the contents of 6 beakers simultaneously. Approximately 106 cysts were added to the chlorine solution to make up a final volume of 800 mL. Chlorine concentrations were determined by amperometric titration using a Fisher model 397 Cl titrimeter. The contents of each beaker were filtered through a 5µm Nucleopore filter after stirring for 10, 20, 30, or 90 minutes. One mL of 10% sodium thiosulphate was added during filtration to neutralize the chlorine. Cysts were washed from the filters with Triton X100 solution into 15 mL tubes and concentrated by centrifugation. Each pellet was divided into two aliquots. One was examined for excystation by the method of Rice and Schaefer (11). The other was resuspended in 5% glutaraldehyde in 0.1M cacodylate buffer, pH 7.2, rinsed in 0.1M cacodylate buffer pH 7.2, postfixed in 1% osmium tetroxide in 0.1M cacodylate buffer pH 7.2, dehydrated in ethanol series and embedded in Spurrs epoxy resin (13). The pellet obtained following 10 mg/L chlorine and 90 min exposure was also treated for excystation. Thin secctions were cut on an LKB Ultrotome, stained in uranyl acetate and lead citrate (9) and viewed and photographed with a Hitachi H600 electron microscope. Results The results of excystation of Giardia cysts after chlorine treatment are summarized in Table 1. These results show that as concentration of chlorine and exposure time increase, the inactivation of cysts increases. A concentration of 10.5 mg/L for 90 minutes ensures total inactivation of cysts. Morphologically, the cysts exposed to 0 mg/L chlorine had a normal appearance (Figure 1). The cyst wall is 0.3 0.5µm thick and is closely applied to what appears as a narrow band of cytoplasm. A space or lacuna is present between this narrow band of cytoplasm and the organism. The cyst membrane enclosed the organism which consists of 2 or 4 nuclei, axonemes, ventral disc fragments, ribosomes and occasional rough endoplasmic TABLE 1. Viability of Giardia muris cysts following chlorine treatment measured as % excystation.
Chlorine concentration (mg/L)
0 Exposure time (min)
4.3
10.5
0
100
100
100
10
100
97
86
20
100
100
73
30
100
93
63
90
* Corresponding author
1.3
0
Page 134
Figure 1. Cyst exposed to 0 mg/L chlorine for 10 minutes and vesicles with electron dense and electron lucent material (x 9000).
Figure 2. Cyst exposed to 0 mg/L chlorine for 10 minutes, showing flagella in lacuna (x 13,500).
Figure 3. Cyst exposed to 1.3 mg/L chlorine for 30 minutes showing a "normal" appearance (x 10,000).
Figure 4. Cyst exposed to 4.3 mg/L chlorine for 20 minutes showing moderate granulation (x 15,000).
Figure 5. Cyst exposed to 4.3 mg/L chlorine for 20 minutes showing granulation of cytoplasm and shrinkage of cyst (x 17,000).
Figure 6. Cyst exposed to 10.5 mg/L chlorine for 90 minutes showing "normal" appearance of cytoplasm, ruffling of narrow band of cytoplasm (17,000).
Figure 7. Cyst exposed to 10.5 mg/L chlorine for 90 minutes showing a necrotic cell (x 8,000). A, axonemes; CW, cyst wall; Er, rough endoplasmic reticulum; F, flagellum; < granulation/flocculation; L, lacuna; n, nucleus; v, vesicle; vd, ventral disc fragment.
Page 135
reticulum. A prominent feature is a layer of vacuoles just inside the cell membrane, some of which contain electron dense material but most of which contain some flocculent but not very electron dense material. Electron dense vesicles also appear in the lacunae (Figure 2). Some of the cysts exposed to chlorine also appear "normal" in morphology (Figure 3). However, in cases where chlorine has obviously entered the cysts there is damage to the cytoplasm which is evidenced by granulation of the cytoplasm with clumping of ribosomes and spaces representing leachedout material (Figure 4). The lacunar space is enlarged due to the shrinkage of the organism (Figure 5). The nuclei and flagella/axonemes are still recognizable. There appears to be a loss of the cyst wall in many cysts exposed to chlorine since the wall does not stain with uranyl acetate and lead citrate. In samples exposed for 90 minutes to mg/L chlorine, a few cysts appear "normal", however, the band of cyloplasm appears ruffled (Figure 6), but in most cysts the cytoplasm is flocculated and the cysts appear necrotic (Figure 7). Discussion Control cysts observed in this study agree with the descriptions of cysts in the literature (2,6,8,12). The cyst wall is 0.3 0.5µm thick as reported earlier and is closely applied to the outer membrane of the narrow brand of cytoplasm surrounding the organism. The inner and outer membranes of the narrow band of cytoplasm are identical in thickness to the plasma membrane. The dense vesicles, which appear to be budding off from the organism, may be involved in secretion of elements of the cysts wall (3). Chlorine appears to attack the cyst wall, slowly degrading it, since the wall changes its staining properties with heavy metals after longer exposure to chlorine. However, when the cyst wall does stain, it is not always totally or evenly degraded and/or dissolved even when the cysts appear necrotic. Therefore, chlorine must also penetrate the cyst wall and attack cellular components. This is evidenced by the flocculation (granulation) of the cytoplasm interspersed with clear areas, shrinkage of the cyst within the cyst wall and appearance of cellular debris in forms of membranous vesicles in the enlarged lacunar space. These clear areas are probably due to chlorine degradation of cytoplasm since leaching did not occur in control cysts and all samples were prepared identically for electron microscopy. The degree of flocculation varies from small areas to the whole cytoplasm. Morphologically, nuclei and flagella appear most resistant to chlorine. In cysts which were treated with 10 mg/L chlorine with viability of 0% there are some that appear "normal", that is they do not exhibit cytoplasmic flocculation, but the cell membrane appears ruffled, and vacuoles persist. This cell membrane ruffling is probably due to the acid treatment during the in vitro excystation procedure and not to chlorine exposure. However, since none of the cysts in this latter treatment are viable, a "normal" appearance of the cytoplasm does not necessarily mean that a cell is viable. Chlorine damage must therefore occur at the molecular level even before this is manifested in morphological changes. Acknowledgements We thank R. Harris and A. Oatway for their technical assistance in electron microscopy. Literature Cited 1. DeWalle, F.P. and C.R. Erland Jansson. 1983. Inactivation of Giardia by chlorine and UV. Unpublished report to Parks Canada. p. 24. 2. Feely, D.E., Erlandsen, S.L. and D.G. Chase. 1984. Structure of the trophozoite and cyst. In: Giardia and Giardiasis, Biology, Pathogenesis and Epidemiology. S.L. Erlandsen and E.A. Meyer (eds). Plenum, New York. pp. 330. 3. Friend, D.S. 1966. The fine structure of Giardia muris. J. Cell Biol. 29:317332. 4. Jarroll, E.L., Bingham, A.K. and E.A. Meyer. 1980. Giardia cyst destruction: effectiveness of six small quantity water disinfection methods. Am. J. Trop. Med. Hyg. 29:811. 5. Jarroll, E.L., Bingham, A.K. and E.A. Meyer. 1981. Effect of chlorine on Giardia cyst viability. Appl. Environ. Microbiol. 41:483487. 6. Luchtel, D.L., Lawrence, W.P. and F.B. DeWalle. 1980. Electron microscopy of Giardia lamblia cysts. Appl. Environ. Microbiol. 40:821832. 7. Meyer, E.A. 1981. Effect of halogens on Giardia cyst viability US EPA report. 600/281174. 8. Nemanic, P.S., Owen, R.L., Stevens, D.P. and J.C. Mueller. 1979. Ultrastructural observations on giardiasis in a mouse model. II. Endosymbiosis and organelle distribution in Giardia muris and Giardia lamblia. J. Infect. Dis. 140:222228. 9. Reynolds, E.S. 1963. The use of lead citrate at high pH as an electron opaque stain in electron miroscopy. J. Cell Biol. 17:208211. 10. Rice, E.W., Hoff, J.C. and F.W. Schaefer III. 1982. Inactivation of Giardia cysts by chlorine. Appl. Environ. Mirobiol. 43:250251. 11. Rice, E.W. and F.W. Schaefer III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709710. 12. Sheffield, H.G. and B. Bjorvatn. 1977. Ultrastructure of the cyst of Giardia lamblia. Am J. Trop. med. Hyg. 26:2330. 13. Spurr, A.R. 1969. A low viscosity epoxy resin embedding medium for electron microscopy. J. Ultrastruct. Res. 26:3143. 14. Wallis, P.M. and H.M. Wallis. 1986. Excystation and culturing of human and animal Giardia spp by using gerbils and TYIS33 medium. Appl. Environ. Microbiol. 51:647651.
Page 137
Removal and Inactivation of Giardia Cysts in a Mobile Water Treatment Plant under Field Conditions: Preliminary Results P.M. Wallis*, J.S. Davies, R. Nutbrown, J.M. BuchananMappin, P.D. Roach and A. van Roodselaar Kananaskis Centre for Environmental Research, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N4. This paper reports the preliminary results from a cooperative project between the Alberta Environmental Centre (Engineering) and the University of Calgary (Microbiology). The major objective of this phase of the project was to evaluate the effectiveness of various oxidants in the inactivation of Giardia cysts and the efficiency of cyst removal using the filtration process. A mobile water treatment plant was used to conduct preliminary experiments at Vegreville using municipal water and subsequent work took place under field conditions at the Kananaskis Field Station in the East Slopes of the Rocky Mountains. This paper reports i) the development of the methodology necessary to produce, recover, and evaluate the viability of cysts in quantities large enough for experimentation in a pilot plant at a flow rate of 150 L/min, and ii) preliminary results of filtration trials with and without filter aids and of the inactivation of a strain of Giardia muris using chlorine. In the first experiment, Giardia muris cysts were continuously injected into the feed water and and recovered after passage through each of the two clarifiers and the filter. The results indicated that cysts decreased in both numbers and viability over the injection period (6 h). Reductions in cyst numbers and viability were caused by excessive agitation of the stock solution, the presence of chloramine in the feed water, and adherence to surfaces inside the clarifiers and pipes. Two sets of experiments were carried out under field conditions using Barrier Lake water (average temperature 3°C, pH 8.2) to supply the mobile water treatment plant. The sand/anthracite filtration studies showed that cyst removal efficiency was a function of maturation time until the filter had been operated for 30 h. After this time, 98% of seeded cysts were consistently removed until backflushing was required. Cysts were detected in the backflush water. A final experiment using polymeric filter aid showed that cyst removal efficiency could be raised to 99% even before the maturation time had elapsed. The disinfection studies showed that 99% of Giardia muris cysts could be inactivated using a total chlorine concentration of 0.5 mg/L and a contact time of 30 minutes. A single trial using G. duodenalis cysts of human origin gave similar results. At a concentration of 0.9 mg/L of chlorine, 94.7 ±6.1% and 99 ±5.3% of these human infective cysts were inactivated after 30 and 90 minutes of contact time (CT=26 and 77) respectively. Animal infection data indicated that CT values of 35 and 77+ were more appropriate for G. muris and G. duodenalis respectively. The results of further experiments will be reported in separate publications (14,15).
Introduction Waterborne disease is now becoming more common in the United States (2,7), reversing a downward trend that persisted for many years. The situation is similar in Canada; a recent report (13) identified six waterborne outbreaks of giardiasis that have occurred in Alberta and British Columbia since 1982. In response to these problems in Western Canada, the University of Calgary and the Alberta Environmental Centre have jointly established a research program to investigate the treatment of drinking water for the inactivation and removal of Giardia cysts. The City of Edmonton, which experienced an outbreak of giardiasis in 1983, has been actively involved with this research through the provision of filter media and operational information designed to simulate their water treatment processes. The occurrence of waterborne outbreaks of giardiasis in many communities is slowly revolutionizing the water treatment industry. Water treatment engineers are questioning the effectiveness of their existing procedures and new technologies for the inactivation and removal of Giardia cysts are being examined. There is a demand for operational data that will permit treatment plants to function at maximum efficiency. The results of this study provide data relevant to the treatment of cold, alkaline water that is typically found in Canada. Three sets of experiments are described in this paper. In the first, cysts were injected into the pilot plant when it was operating with municipal water in Vegreville containing chloramine. The object of this preliminary experiment was to test the methodology for cyst injection, recovery, and viability testing. The second series of experiments * Corresponding author.
Page 138 TABLE 1. Operating specifications for each clarifier. Potable Output
Rise Rate Under Troughs
Retention Time to Overflow
L/min
IGPM
L/min/m2
GPM/ft2
(min)
45.50
10.0
14.70
0.30
250.0
91.00
20.0
29.40
0.60
125.0
113.75
25.0
36.75
0.75
100.0
136.50
30.0
44.10
0.90
83.3
182.00
40.0
58.80
1.20
62.5
227.50
50.0
73.50
1.56
50.0
tested the effect of filter maturation and conditioning on the removal of cysts and the third series was concerned with the inactivation of G. muris cysts using chlorine. Materials and Methods Development of the Mobile Water Treatment Plant To encompass various applications, the pilot plant was designed to provide optimum flexibility. The incorporation of alternative technologies into a facility based on conventional treatment (Class 4) provided the ability to subject source waters to a wide variety of treatment options. This approach lends the capability of rapid configuration of the pilot plant to meet specific requirements. A comparison of initial capital costs showed only a small difference between 136 L/min (30 gpm) and the 227 L/min (50 gpm) capacities with the overall weight and dimensions increasing slightly. A 341 L/min (75 gpm) plant, however, was significantly more expensive and the overall weight of the components nearly doubled. A capacity smaller than 100 L/min would have required custom built components, making its costs similar to that of a 227 L/min (50 gpm) plant. Also, this small a plant would have a considerable scaleup factor to full municipal sizing. Consequently, the system was tailored around a nominal capacity of 227 L/min (50 gpm). With the selection of the 227 L/min (50 gpm) plant capacity, the prospective shelter for the components was designed to provide the optimum accommodation for the selected water treatment plant. This plant consists of 2 reactorclarifiers, 2 filters and a collection reservoir, whose heights normally run about 2.44 m (8') to 3.05 m (10'). The size of the shelter which proved most feasible was 3.66 m (12') wide by 15.85 m (52') long with an overall maximum outer height of 4.75 m (15' 7"). There is a 0.305 m (1') overhang on both sides of the lowboy, but this does not cause any instability since the significant loadings were located along the trailer's centreline. The static load of the plant when full of water is 45.4 tonnes and the dynamic load is 15.0 tonnes. To ensure compatibility with municipal plants, specifications were prepared for a range of flow conditions for both clarifiers (Table 1) and filters (Table 2). The final layout of the plant is given in Figure 1. TABLE 2. Operating specifications for each filter. Potable Output
Operation Rate 2
L/min
IGPM
L/min/m
GPM/ft2
45.50
10.0
32.81
0.67
91.00
20.0
65.13
1.33
113.75
25.0
97.94
2.00
136.50
30.0
130.75
2.67
182.00
40.0
163.07
3.33
225.50
50.0
195.88
4.00
Figure 1. Layout of mobile water treatment plant.
Versatility in water treatment was increased by the insertion of extensive interconnecting piping and flow control valves. The plant contains two reactor clarifiers and two filters, so interconnecting plumbing allows the clarifiers and filters to operate in series, in parallel, or as independent clarifierfilter systems operating side by side at different chemical dosages/flow rates (Figure 2). Clarification times and chemical dosages can therefore be modified to suit different water inlet sources. This highly adaptable unit actually simulates many different water treatment plants but at a fraction of the cost. Ease of operation was increased with the addition of a microprocessor which remotely controls and monitors the output parameters. This permitted the option of running the plant unmanned for extended periods of time providing that the chemical additives are provided for. A 36 to 48 channel input microprocessor which would monitor and record signals from pumps, thermocouples, turbidimeters, colourimeters, hardness testers, pH meters, flowmeters, chlorine meters, chlorine dioxide, conductivity meters, and chemical feed equipment was therefore built in. Field Site and Water Conditions The mobile water treatment plant was moved to the Kananaskis Field Station at Barrier Lake in Kananaskis Country (80 km west of Calgary, AB, Canada). The pilot plant was situated 2 km from the microbiological laboratory at the Field Station. The site afforded easy access to an abundant supply of raw water (Barrier Lake, Table 3) and sewage treatment lagoons for disposal of water contaminated with Giardia cysts. When no cysts were present in the plant effluent, water was discharged to the Kananaskis River. All contaminated effluent was superchlorinated at a chlorine concentration of 15 mg/L. Barrier Lake is a shallow reservoir fed by the Kananaskis River. During our experiments, the water was hard (139 165 mg/L as CaCO3) and the average turbidity during these experiments was 3.6 ±0.8 NTU. The temperature was 7.0 at the beginning of the experiments in September and dropped to 2.5 °C in late November (mean 4.9±0.7). The pH of Barrier Lake ranged from 8.1 to 8.5 in the same time period. Sources of Giardia Cysts We obtained a strain of Giardia muris (GM1) from G. Faubert at the Institute of Parasitology at McGill University and maintained it in vivo using outbred Swiss Webster mice [Crl:CFW(SW)BR]. A human strain (H8) was obtained from the stool of a symptomatic male child in Canmore, AB and maintained both in culture (12) and in vivo using the mongolian gerbil (Meriones unguiculatus) model of Belosevic et al. (1). Cyst Production Cysts were produced by infecting 30 or more mice (for GM1) or gerbils (for H8) by gavage. Six to eight days after infection, animals were placed on a wire mesh suspended over a shallow layer of water in a large cage and faeces were collected overnight. Cyst production was enhanced by adding dexamethasone (Schering) at a concentration of 40 µg/mL to their drinking water. This resulted in up to a 10fold
Page 139
Figure 2. Flow schematics for mobile water treatment plant.
increase in cyst production. Mice were found to be more tolerant of dexamethasone than gerbils but both animals were found to be very sensitive to the drug and the use of dexamthasone at higher concentrations is not recommended. Faeces were homogenized using a Braun kitchen hand blender and carefully layered over 1.0 M sucrose in centrifuge tubes (50 mL capacity). The tubes were centrifuged for 10 minutes at 2500 rpm in a Sorvall RC5 refrigerated centrifuge. After centrifugation, the interface between the sucrose and the remaining liquid was siphoned off using an Oxford MacroSet pipette whose tip had been cut off to provide a larger orifice. This crude faecal isolate was diluted to 1500 mL and stirred for 30 minutes using a magnetic stirrer after which the cysts were counted using a haemocytometer. At peak production, it was possible to isolate up to 109 cysts from 30 mice using these methods. Viability Testing Initial experimentation with trypan blue dye exclusion and the Gomorri acid phosphatase methods resulted in the conclusion that these methods were unreliable. The in vitro excystation method of Rice and Schaefer (9) was therefore used for all of our viability tests. The procedure used was essentially the same as that described by Rice and Schaefer except that lyophilized trypsin (Gibco) was substituted. We counted any cyst that showed internal movement, protoplasm extrusion, or flagellar motion as a "partially excysted trophozoite"; very few fully excysted trophozoites were observed. Stock cyst solutions were always prepared on the same day as the experiment and were always 95+% viable. Viability was also tested by animal infection. Swiss Webster mice were used to test G. muris cysts and gerbils were used to test G. duodenalis cysts. Animals were infected by gavage with approximately 2500 cysts recovered using metrizamide as described below. The presence of infection was tested by faecal examination. Animals which did not pass cysts 8 days after infection were sacrificed and the gut was examined for the presence of trophozoites by three biopsy snips taken from the duodenum, jejunum, and ileum respectively. Cyst Recovery by Membrane Filtration Cysts were recovered by tapping off water from the treatment train and pumping it through 5 µm Nuclepore membrane filters in 142 mm plexiglass housings. Peristaltic pumps set to pump at the rate of 1 L/min were used. At the end of the collection period, the filters were disassembled and washed off using a fine stream of tap water containing Triton X100 and 1% sodium thiosulphate into a sample bottle. The washings were taken immediately to the laboratory, concentrated by centrifugation, and counted using a haemocytometer. In order to test the viability of cysts that were recovered from water using Nuclepore filters (described below), it was necessary to remove background silt and algae. This was accomplished by clarifying the centrate over metrizamide, a high densitylow viscosity inert medium that can be made isotonic with human serum (8). The concentrated water suspension was layered over the metrizamide solution in a 15 mL centrifuge tube and centrifuged for 5 minutes at 500 × g. The cysts were then aspirated from the interface using a glass pipette and washed. In order to quantify the effects of metrizamide on cyst recovery efficiency and viability, the following experiment was performed. Giardia muris cysts were added to 100 L of untreated streamwater at concentrations of 50, 100, 500, and 1000 cysts per L. Using continuous agitation, two 50 L batches were filtered through 5 µm Nuclepore membranes. The recovered material was then clarified by centrifugation over metrizamide and the purified cysts were excysted according to the method of Rice and Schaefer (9). The results were pooled and recovery efficiency was found to average 31.3 ±11%. The viability of the recovered cysts was 83 ±10%. In a separate experiment where known numbers of cysts were layered over metrizamide without any filtration, the recovery efficiency of the clarification step was found to be 67 ±9%. Previous work (11) demonstrated that the recovery efficiency of the filtration procedure alone was 71 ±13%. TABLE 3. Barrier lake water chemistry. Parameter
mg/L (unless otherwise noted)
pH
7.88 8.65*
Conductivity
292 346*
Alkalinity
127 133*
Hardness
139 165*
Calcium
40.6 45.0
Magnesium
9.1 13.1
Sodium
<1 1.2
Potassium
0.33 0.40
Iron
<0.01 0.04
Chloride
<1 1.5
Sulphate
27.9 30.0
Fluoride Nitrate & Nitrite Silica Total dissolved solids (calc) Total phosphorus Total Kjeldahl Nitrogen
0.06 0.16 0.024 0.055* 3.10 3.27 156 170 0.004 0.006 0.18 0.30
* pH is in pH units; conductivity noted in microsiemens/cm; Alkalinity and hardness expressed as calcium carbonate; Nitrate and Nitrite expressed as N.
Page 140 TABLE 4. Results from the first preliminary trial. Sample
Elapsed Time (min)
Volume Filt. (L)
Recovery (% of cysts added)
Viability (%)
A1
150
54.2
58
89
A2
225
54.2
84
95
B1
330
55.0
36
93
B2
410
53.8
44
90
C1
355
53.2
1
N.D.†
C2
440
36.5
1
N.D.
Stock Cyst Solution: 278 cysts/mL Injection Rate: 225 mL/min Initial Viability: 98% † Not Determined because of insufficient numbers of cysts.
Preliminary Injection and Recovery Trial The object of this demonstration was to show that cysts could be produced in adequate numbers, added to the raw water feed, quantitatively recovered, and tested for viability. Two trials were run in 1985 using G. muris (GM1) cysts. In both experiments, a concentrated solution of cysts was slowly added to the rapid mix chamber upstream of Clarifier 1 over 6 hours on the first occasion and 8 hours on the second. The feed water was Vegreville municipal water drawn from a hydrant on the grounds of the Alberta Environmental Centre. This water contained approximately 1.5 mg/L chloramine. The flow rate of water through the plant was 150 L/min for the first trial and 101 L/min for the second. The concentrated cyst solution was continuously agitated by a magnetically driven impeller pump throughout the experiment and injected into the rapid mix chamber using a peristaltic pump. A strontium chloride solution was simultaneously injected from a separate container by peristaltic pump in order to provide a tracer. Cysts were recovered several times downstream of Clarifier 1 (sample site A), downstream of Clarifier 2 (sample site B), and downstream of the sand/anthracite filter (sample site C). The recovered cysts were counted and viability was tested by excystation. Samples were taken at each of the three points using 5 µm Nuclepore filters. Counts were made on samples that were centrifuge concentrated only. Viability tests were performed on samples after clarification with metrizamide as described above. Filter Trials The sand/anthracite filter was initially backflushed and then operated at 150 L/min intermittently over a period of 17 days until the filter was too clogged to maintain this flow rate. A total of approximately 760,000 L of Barrier Lake water was filtered. Cyst removal efficiency was tested after 7, 17.5, 29, 31, 51, and 85 hours of operation. At the end of the run, the filter was backflushed and samples were collected and analyzed for cysts. A final trial was conducted when the rejuvenated filter had been operating for 18 hours. This experiment was conducted in the same manner as the preceding trials but polymer filter aid was added to the rapid mix chamber of Clarifier 1. Strontium tracer and varying numbers of GM1 cysts (6 to 140 × 106) were simultaneously injected into the water train upstream of the filter in a spike lasting 60 sec. Sequential samples were collected downstream of the filter for 30 minutes in five separate sequential fractions. The strontium concentration was measured in the feed stock, the filtrate, and in samples taken continuously from the post filter effluent. This allowed a prediction of the number of cysts that should have been present in each 6 minute portion of the injected spike if no filtration had taken place. The cysts recovered were counted in each fraction, summed and compared with the predicted numbers using the method of Wallis and BuchananMappin (11). All of the recoveries were corrected by dividing the number of cysts by 0.7 to allow for cyst loss during the filtration procedure (see above and 11). When the filter was backflushed, samples were taken of the surface scum and of the backflush water. Disinfection Experiments Disinfection trials were carried out with Giardia muris (GM1) cysts and prechlorination with gaseous chlorine. Total and free chlorine concentrations were continually monitored while the treatment plant was in operation. The loss of free chlorine as water flowed through the clarifiers was calculated in order to quantify the chlorine demand of the pilot plant. Chlorine was added to the treatment train between experiments to ensure that the chlorine demand of the clarifiers and piping was kept to a minimum. Initial studies showed that the concentration of tracer peaked downstream of Clarifier 1 30 minutes after the addition of a spike. Attenuation of the tracer concentration occurred downstream of Clarifier 2 but a plateau concentration was reached after 90 minutes. Cysts were therefore injected over 15 minutes into the rapid mix chamber above Clarifier 1 and recovered by pumping water through four Nuclepore filters at once from recovery points post Clarifier 1 (sample site A) and post Clarifier 2 (sample site B), after 30 and 90 minutes equivalent time respectively. In this way, four replicate samples were collected for viability testing at sampling points A and B. Viability was assayed using both in vitro excystation and animal infection. Results Preliminary Injection and Recovery Trials The results of the first experiment are reported in Table 4. The number of cysts injected turned out to be rather low and the recovery rates were variable. Viability, however remained high throughout the experiment. Very few cysts were recovered at site C because most of the cysts were removed by the filter. It was decided to repeat the trial with more replicate samples at each site and to drop sampling site C. The results of the second experiment are reported in Table 5 and Figure 3. A higher concentration of cysts was injected but it was found that cyst recovery dropped steadily throughout the experiment. The stock solution was checked at the end of the experiment and it was TABLE 5. Results from the second preliminary trial. Elapsed Time (min)
Volume Filt. (L)
Recovery (% of cysts added)
Viability (%)
A1
149
52.3
58
94
A2
230
52.1
22
88
A3
305
52.1
18
70
A4
373
52.5
13
59
A5
450
53.1
12
N.D.†
B1
175
53.5
44
94
B2
234
52.1
24
88
B3
305
54.4
25
71
B4
364
53.8
22
52
B5
435
52.3
11
N.D.
Sample
Stock Cyst Solution: 1302 cysts/mL Injection Rate: 92 mL/min Initial Viability: 96 % Final Viability: 54 % † Not Determined because of insufficient numbers of cysts.
Page 141
Figure 3. Recovery and viability of GM1 cysts from the second Vegreville trial.
found that the concentration of cysts had dropped from 1302 to 810 cysts/mL. At the same time viability dropped from 94 to approximately 55%. Examination of the system disclosed that the pump used to keep the concentrated cyst solution agitated was responsible for cyst loss probably because the pumping action was too harsh. This problem was compounded by the presence of chlorine compounds in the feed water and the lack of temperature control in the concentrated stock solution. It was concluded that impellor pumps should not be used for agitation and a paddle stirrer was used for all subsequent experiments. The results of these experiments also suggested that only unchlorinated water should be used to suspend cysts despite the high chlorine demand of the faecal material and that the injection time should be kept as short as possible. Filtration Trials The results of the filtration trials are reported in Figure 4. Cyst removal by the filter was very low (33%) initially (7 hours), but reached consistently high levels after 30 hours of operation. The maximum removal efficiency observed was 98.6%. The experiment with polymer filter aid took place after only 18 hours of operation in a subsequent run and resulted in a removal efficiency of 99%. Cysts were found in the surface scum on top of the filter before backflushing and in the first two samples taken of the backflush water. Eight further samples were negative. Disinfection Trials The combined chlorine level of treated Barrier Lake water was found to average 0.05 mg/L and the loss of chlorine in water flowing through the clarifiers was low (Figure 5). The excystation results of the disinfection experiments are plotted in Figure 6. Very high reductions in GM1 cyst viability were observed at concentrations above 0.15 mg/L with contact times of both 30 and 90 minutes. When the excystation data were plotted against the product of chlorine concentration and contact time (CT), a similar pattern emerged (Figure 6). Mortality rates of 99% or greater were observed at CT values above 13 based on excystation data. Mice became infected, however, with G. muris at CT levels up to 35 mg*min/L. The single trial employing G. duodenalis cysts resulted in slightly higher viability levels. At a free chlorine concentration of 0.85 mg/L, 5.3% of H8 cysts were viable after 30 minutes and 1.0% were viable after 90 minutes of contact time. Most (3/4) of the gerbils that had been inoculated with cysts recovered post clarifier 1 (CT=26)
Figure 4. Filter removal efficiency.
Page 142
Figure 5. Chlorine loss in the mobile water treatment plant during experiments at different chlorine concentrations.
became infected but only 1/4 gerbils became infected with cysts recovered post clarifier 2 (CT=77). Discussion Filter Experiments Most of the filtration experiments were carried out without using any form of coagulant. The use of coagulating agents is known to increase cyst removal efficiency so it is of interest that removal efficiencies as high as 98% could be achieved without it. However, it was concluded that the critical time for filters in the absence of filter aid was the first 30 hours after backflushing, typically that period of normal filter operation. This time period is undoubtably flexible depending upon the turbidity and the composition of the filter media. This series of experiments clearly shows the variability in cyst removal that can be encountered depending on the maturation of the filter and the need for careful disposal of filter backwash water was demonstrated. One of the drawbacks to our experiments was that the filter was operated intermittently over a period of 17 days. This does not correspond to normal treatment plant procedure and it is possible that there was some development of biological material in the filter that enhanced the removal of Giardia cysts. Future experiments will be carried out under conditions of continuous operation. Disinfection Experiments The results reported above suggest that a significant proportion of GM1 cysts (G. muris) can be killed in a water treatment plant at chlorine concentrations of approximately 0.5 mg/L and contact times as low as 30 minutes. The results of one additional trial using human Giardia cysts suggested that this strain was more resistant. Jarroll et al. (5) reported that 99.8% of G. lamblia cysts could be killed at 3 and 20 °C in cloudy river water with free chlorine concentrations of 6 to 8 mg/L. In a later paper Jarroll et al. (6) found that 99.9% of G. lamblia cysts could be killed at a concentration of 2 mg/L after 60 minutes (pH 8, 5 °C) but that concentrations up to 8 mg/L were required to achieve the same effect after 30 minutes. DeWalle and Jansson (3) tested the viability of G. muris cysts in Banff municipal water (pH 7.8, 1 °C) and reported that mortality levels in excess of 99% were possible at concentrations above 2 mg/L and contact times of 90 min. High mortality levels could be achieved at concentrations of less than 1.0 mg/L but only after contact times of 15 h. Rice et al. (10) found that G. muris was slightly more resistant than G. lamblia but on the average only about 95% of cysts could be killed after 60 minutes at 2.5 mg/L chlorine (pH 8). Faubert et al. (4) tested the viability of G. muris (the same strain used in this study, GM1) by infecting CD1 mice with cysts that had been exposed to a free chlorine concentration of 0.59 mg/L and found that animals could still be infected after 60 minutes of contact time (CT=35). These results showed that at least some of the cysts were still viable. These authors attributed this viability level to the high pH of the tapwater (pH=8.24, temp.=6 °C, turbidity=3.1 NTU). The water chemistry of Barrier Lake water is similar.
Figure 6. Viability (% excystation) as a function of the product of chlorine concentration and time (CT) for GM1 cysts.
Page 143
Results reported in the literature for bench scale experiments indicate that for the conditions existing in Barrier Lake (pH 8, 23 °C), concentrations of approximately 4 to 5 mg/L would be required to kill 99% of G. muris or G. lamblia (duodenalis) cysts with 30 to 90 minutes of contact time (CT= 50 to 360) based on excystation. Based on excystation data, our results show that 99% mortality could be obtained at a CT of approximately 13 for G. muris and 77 for G. duodenalis under similar conditions. Animal infection data suggest that CT values of 35 and 77+ are more reasonable for G. muris and G. duodenalis respectively. It must be remembered, however, that animals were infected with a minimum of 2500 cysts which is a relatively large inoculum. We conclude that chlorine was more effective in the pilot plant than data from the literature would have predicted. It is also possible that the strains of Giardia we used were less resistant to chlorine than those used by other investigators. This possiblility will be investigated in future experiments. The results from the literature summarized above were performed on the benchtop in volumes of water ranging from 200 to 1000 mL. In most cases, the cysts were suspended in phosphate buffer solution with little or no agitation. The cysts were collected with a minimum of disturbance and tested for viability immediately. These conditions could therefore be said to be optimum for cyst survival. Experiments performed in the mobile water treatment plant, however, were quite different. The volumes of water used were 2 to 3 orders of magnitude greater, were continually mixed and circulated, and were recovered by membrane filtration before transportation to the lab for viability testing. The water used was not buffered and contained all of the background contamination typically found in raw (but good quality) feedstock water. It is possible, therefore, that these factors made cysts more susceptible to chlorine. In the control experiment (no chlorine), cyst viability dropped from 88% in the stock solution to 78% after 30 min. and 75% after 90 min. (raw data) suggesting that the treatment plant itself was contributing to the mortality of cysts even in the absence of chlorine. The equipment and procedures used in the mobile water treatment plant were very similar to those used routinely in water treatment plants throughout Alberta. Conclusions 1. Methodology was developed permitting the production, quantitative recovery, and viability testing of large numbers of different strains of Giardia cysts for water treatment experiments. 2. Filtration efficiency experiments demonstrated that filter condition is very important to cyst removal. The sand/anthracite filter was able to remove up to 98% of seeded Giardia muris (GM1) cysts but only after approximately 30 h of operation. 3. Based on a single filtration trial, the use of a polymer filter aid increased the (GM1) cyst removal efficiency to 99% even though the filter had not fully matured. 4. Filter backwash water contained Giardia (GM1) cysts indicating that caution should be employed when disposing of such wastewater. 5. Mortality rates of 99% were achieved with the Giardia muris (GM1) cysts used in these experiments at CT values as low as 13 based on excystation data and 35 based on animal infection. 6. Based on a single trial, human Giardia cysts responded to chlorine in a manner similar to Giardia muris cysts. At a concentration of 0.85 mg/L free chlorine, 94.7 ± 6.1% and 99 ± 5.3% of H8 cysts were killed at contact times of 30 and 90 minutes respectively. This corresponds to a CT of 77 for 99% mortality based on excystation data but 1/4 gerbils still became infected. 7. The results from these experiments have shown that the mobile water treatment plant is a useful tool for developing water treatment protocols for removing and inactivating Giardia cysts at the pilot plant scale. Some of the experiments have yielded results that were not predicted by previous bench scale studies but the investigators believe that the results presented in this report are more applicable to full scale water treatment plants in Alberta. Literature Cited 1. Belosevic, M., Faubert, G.M., MacLean, J.D., Law, C., and N.A. Croll. 1983. G. lamblia infections in Mongolian gerbils: an animal model. J. Inf. Dis. 147:222 226. 2. Craun, G.F. 1984. Waterborne outbreaks of giardiasis: Current Status. in: Erlandsen, S.L. and E.A. Meyer (eds.). Giardia and Giardiasis, Plenum Press, NY. 3. DeWalle, F.P. and C.R. Erland Jansson. 1983. Inactivation o Giardia by chlorine and UV. Unpublished report to Parks Canada. 24p. 4. Faubert, G.M., Leziy, S.S., Bourassa, A. and J.D. MacLean. 1986. Effects of environmental conditions and standard chlorination practices on the infectivity of Giardia cysts. Dis. Aq. Org. 2:15. 5. Jarroll, E.L., A.K. Bingham, and E.A. Meyer. 1980. Giardia cyst destruction: effectiveness of six smallquantity water disinfection methods. Am. J. Trop. Med. Hyg. 29:811. 6. Jarroll, E.L., A.K. Bingham, and E.A. Meyer. 1981. Effect of chlorine on Giardia lamblia cyst viability. Appl. Environ. Microbiol. 41:483487. 7. Lippy, E.C. 1981. Waterborne disease: occurrence is on the upswing. J. Am. Wat. Wks. Assoc. 73:5762. 8. Loos, J.A. and D. Roos. 1976. Density analysis as a tool for blood separation. in: Rickwood, D. (ed.) Biological Separations in Iodinated DensityGradient Media, Information Retrieval Ltd., London. p. 100. 9. Rice, E.W. and F.W. Schaefer III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709710. 10. Rice, E.W., J.C. Hoff, and F.W. Schaefer III. 1982. Inactivation of Giardia cysts by chlorine. Appl. Environ. Microbiol. 43:250251.
Page 144
11. Wallis, P.M. and J.M. BuchananMappin. 1985. Detection of Giardia cysts at low concentrations in water using Nuclepore membranes. Wat. Res. 19:331334. 12. Wallis, P.M. and H.M. Wallis. 1986. Excystation and culturing of human and animal Giardia spp. by using gerbils and TYIS33 medium. Appl. Environ. Microbiol. 51:647651. 13. Wallis, P.M. 1987. Criteria for Giardia in drinking water. unpublished report to Health and Welfare Canada. 54 pp. 14. Wallis, P.M., Davies, J.S., Nutbrown, J.M., BuchananMappin, J.M., Roach, P.D., and A. van Roodselaar. 1988. Removal and inactivation of Giardia cysts in a mobile water treatment plant under field conditions: preliminary results. in prep. 15. Wallis, P.M., and A. van Roodselaar. 1988. Removal of Giardia cysts by filtration in a pilot water treatment plant under field conditions. in prep.
Page 145
DIFFERENTIATION OF GIARDIA ISOLATES
Page 147
The Genome of Giardia Intestinalis Peter Upcroft*, Peter F.L. Boreham and Jaqueline A. Upcroft Queensland Institute of Medical Research, Bramston Terrace, Herston Brisbane, Qld. 4006, Australia We have used orthogonal field alternation and field inversion gel electrophoresis to effect separation and identification of chromosomes from Giardia intestinalis. Most chromosomes are very large compared to those of Plasmodium falciparum. However, there is a discrete population of minichromosomes approximately 100 kb, similar in size to those present in Trypanosoma brucei. These have been isolated, analyzed by restriction endonuclease cleavage, cloned in E. coli and probed with unique and repetitive DNA segments from cDNA and genomic libraries. This study is an attempt to establish a base for genetic analysis, gene mapping and strain identification of G. intestinalis.
Introduction Although members of the genus Giardia are distributed widely throughout the world in man and other animals, there is still considerable controversy concerning speciation within the duodenalis group, which includes G. intestinalis. Morphometric analysis does not assist with the elucidation of this problem, even though it has proven useful in distinguishing the duodenalis group from the muris and agilis groups (27). At a genetic level very little is known about the Giardia genus in general. Less is known about genetic differences which may confer phenotypic distinction between isolates e.g. pathogenicity, although differences in restriction enzyme cleavage patterns of DNA from isolates have been observed (17). For all parasite groups of medical significance, the principle drawback in genetic analysis has been the lack of a consistent method of chromosome identification. This has been circumvented in the case of the Variant Surface Glycoproteins of trypanosomes, to some extent, by cloning of the genes by recominant DNA techniques (2). However, the development of techniques to separate, and hence identify, chromosomes by gel electrophoresis, has had a wider impact on parasite genetics. This involves lysing whole parasites which are embedded in an agarose block, to prevent shearing of the large DNA molecules (19). Chromosome size DNA molecules that are smaller than 3Mbp then can be separated by three basic types of electrode and electrophoretic configurations, pulsed field gel (PFG) eletrophoresis (19), orthogonal field alternation gel electrotrophoresis (OFAGE) (7) or field inversion gel electrophoresis (FIGE) (8). Using these techniques chromosomes of yeast (8,19), Trypanosoma spp. (11,12,23,24), Plasmodium falciparum (15,25), Leishmania spp. (10,23), Crithidia fasciculata (23), Herpetomonas muscarum (23) and Leptomonas ctenocephali (23) have been separated effectively. The chromosomes vary dramatically in size from 50kb to >4Mbp and from a few, to hundreds in number. Even within a genus there are great differences. For example within the genus Trypanosoma, T. cruzi has at least 20 chromosomes larger than 300kb and probably none larger than 4Mbp. T. brucei has some very large chromosomes which do not enter the gel complemented by approximately 100 minichromosomes 50 150kb in length. Moreover, differences are observed between and within strains (12,23). This variation within strains has been observed in malaria also (15,25). These basic chromosome analyses therefore have been very informative both from a genetic and a taxonomic perspective. We have used OFAGE and FIGE to attempt separation and identification of chromosomes in a number of G. intestinalis isolates; (i) to compare them with the other described protozoa (ii) to determine if there are significant strain differences that may be used for species and strain identification (iii) to establish a base for genetic analysis and mapping. The latter has been coupled to studies with cDNA and genomic DNA libraries which we have cloned in E. coli from G. intestinalis isolates. Materials and Methods G. intestinalis isolates were maintained in axenic culture in modified TYIS33 medium supplemented with 1 mg/mL bile (3). DNA was extracted from whole parasites basically as described by Ketner and Kelly for mammalian cells (16), followed by banding in CsCl (1.60 gm/mL CsCl, Beckman 75Tirotor, 45K rpm, 36h) to remove contaminating carbohydrate. For OFAGE and FIGE whole parasites were lysed in Seaplaque agarose blocks. Usually 5 × 108 trophozoites were washed twice in phosphate buffered saline (PBS) after culturing, and were resuspended in 0.5ml PBS. 0.5mL of 2% Seaplaque agarose (Marine Colloids) in saline was added and the blocks were allowed to set in LabTek 8well chamber slides (Miles Laboratories). Slices were cut to fit the wells of a Pharmacia GNA 200 electrophoresis apparatus. In the case of P. falciparum chromosome blocks, 5 × 108 parasites were collected from saponin lysed infected red blood cells. They were washed in PBS and set in Seaplaque as above. The parasites were lysed in the blocks as previously described (19,24). To lyse parasites directly in the loading slots cast in the agarose, 5 × 107 trophozoites were washed twice in PBS containing 10% glycerol. They were loaded (30µL) over 15µL of lysis buffer (5µL each of 10% sarkosyl; 100 mMTris HCl/10mM * Corresponding author.
Page 148
EDTA, pH7.5; PBS in 30% glycerol). 5µL of proteinase K at a concentration of 20 mg/mL was added also. The parasites were allowed to lyse at room temperature for 3 hours. cDNA and genomic libraries will be described in detail elsewhere. Briefly, RNA was extracted from whole parasites (28). cDNA was prepared by the method of Gubler and Hoffman (14), the ends made blunt with mung bean nuclease (13) and ligated into PUC19 (28) cleaved with SmaI. Genomic DNA was cleaved to appropriate sizes by DNase I in the presence of Mn2+ (1,20) and cloned into pUC19. The bacterial host was E. coli JM109 (28). Restriction endonucleases were purchased from New England Biolabs and used according to the manufacturer's instructions. DNA segments were extracted from agarose (Seaplaque, Marine Colloids) as described (6) or by electroelection (International Biotechnologies, Inc., Model UEA electroelutor) according to the manufacturer's instructions. Results Figure 1 shows an example of OFAGE used to separate P. falciparum chromosomes, after gentle lysis of the parasites in Seaplaque agarose. In parallel lanes G. intestinalis trophozoites were lysed and subjected to the same OFAGE. One can see clear resolution of malaria chromosomes. Most of the Giardia DNA is still in the loading slot. Some structure can be seen as faint bands in the light smear of DNA migrating into the gel, but is not particularly convincing as chromosometype structures when compared with the malaria chromosomes. The chromosomes of the malaria strain used (FC27) have been
Figure 1. G. intestinalis and P. falciparum chromosomes set in agarose blocks and separated by OFAGE. Parasite DNA set in blocks was separated in a 1% agarose gel in TBE (0.5x) recirculating buffer. Pulse times were 45 sec. Electrophoresis was carried out at 300V, 150mA for 12h at 4°C. Lane 1 contains P. falciparum K; Lane 2 P. falciparum FC27; Lane 3 P. falciparum HB3; Lane 4 G. intestinalis BRIS/3/HEPU 106/1/3.
well characterized previously (15) and the OFAGE separation here is consistent with the published PFG analysis (15). The conditions of preparation of both malaria and Giardia chromosomes in the agarose blocks were identical, so that the slight smear of DNA in the Giardia lanes is a property of that parasite. Changing the conditions of electrophoresis, for example longer running times, or longer or shorter pulse times, did not demonstrably change the electrophoresis patterns for Giardia over many OFAGE runs. However, the expected changes in malaria chromosome profile, such as further migration of large chromosomes with longer pulse times, were consistently observed. Changes in lysis conditions, such as increased EDTA, proteinase K or detergent had no observable effect either. The OFAGE separation was transferred to nitrocellulose by the Southern procedure (21) after nicking of the large molecules by acid treatment (26) and hybridized with single copy cDNA probes and repetitive sequence genomic probes. No pattern was detected which was consistent with a single band of chromosomal material (data not shown) that had migrated a significant distance from the loading slot into the gel. Figure 2 shows separation of malaria chromosomes by FIGE. Resolution is superior to OFAGE and PFG and separation is very good also. However, the Giardia chromosome preparations analyzed on the same gel are similar to those seen by OFAGE, with perhaps a little more detail or structure seen in the smear of DNA that has migrated into the gel. Some banding pattern can also be seen underlying the faint smear. Most of the DNA again remained in the loading slot. However, there is a very distinct, but broad, band of material migrating slightly slower than the lambda phage marker at 50kb. In the previous examples the chromosomes were released from whole parasites after they were entrapped in Seaplaque agarose. To determine whether this procedure had any effect on the release of the large chromosomes, we lysed parasites directly in the gel slot without agarose entrapment, prior to FIGE, in a manner similar to the analysis of large plasmids in Rhizobia (9). In this case (Figure 3) the band migrating slightly slower than lambda at 50kb was not as intense as seen in Figure 2. However, there were distinct species seen at 0.8Mbp and 2Mbp. The smear of ethidium bromide staining material was almost absent in the region between 2Mbp and the slot. Again most of the chromosomal material did not enter the gel and remains in the slot. Since there were broad bands of DNA seen in the above examples migrating at defined molecular weights and in a reproducible fashion, it was of interest to determine if these were distinct entities, or a range of species in each band. We therefore set the lysed parasites, which were mounted in Seaplaque, into the same agarose for FIGE separation, from which the appropriate band could be excised and the DNA extracted (6). The gel was run for 16h at 250V/150mA with a pulse time of 22 seconds forward and 7 seconds in the reverse direction. The broad band migrating slightly slower than 50kb (Figure 3) was excised from the Seaplaque agarose, diluted sixfold into
Page 149
10mMTris HCl, pH 7.5, 1mMEDTA and extracted twice with an equal volume of buffered phenol. The residual phenol was removed by three extractions with an equal volume of ether. Although the smeared nature of the band obviates accurate determination of DNA concentration, the yields of such large molecules were very good, estimated at greater than 50%. The nature of the extracted DNA was analysed by further gel electrophoresis under the usual uniform field conditions. Figure 4 shows the migration of a sample of this DNA in 0.8% agarose. In contrast to the appearance of the band migrating under OFAGE or FIGE conditions,
Figure 2. G. intestinalis and P. falciparum chromosomes set in agarose blocks and separated by FIGE. Parasite DNA was separated in a 1% agarose gel in TBE buffer at 300V, 150mA for 20h. Forward and reverse pulses were 22 and 7 sec., respectively. Lane 1 contains P. falciparum 7G8; lane 2 P. falciparum FC27; lane 3 G. intestinalis BRIS/3/HEPU 106 1/3.
the band was very tight and uniform. Most if not all the DNA was of a single, discrete size class, again migrating slightly slower than lambda at 50kb. Although it is difficult to estimate accurately DNA lengths in this region, even using concatamers of lambda as markers, we estimate that the size of the extracted DNA was approximately 100kb. Malaria chromosomes migrating under PFG, OFAGE, or FIGE also showed a band in the region of 50kb. When these are extracted as above and reanalyzed by gel electrophoresis, the DNA migrated as a broad smear of different size classes (data not shown): no distinct, single band was observed. The discrete species at 100kb appears to be a property of the Giardia genome, therefore, and not of the extraction conditions used. The 100kb band of DNA seen by gel electrophoresis is consistent with either a single 'minichromosome' or a collection of 'minichromosomes' of the same size. Cleavage with restriction enzymes indicated that the 100kb DNA is unlikely to be a single species (unpublished data) because a discrete series of bands which has a summation of 100kb in size was not observed. However, the 'minichromosomes' did contain repeated sequences, which were detected as distinct bands after gel electro phoresis and by hybridization to cloned repetitive sequences after Southern transfer. These repeated sequences in the 'minichromosome' are not arranged in an order which is consistent with their tandem repetition, as might be expected for rDNA, for example. Their relative mass in the 'minichromosomes' is also not consistent with the total mass of these repeats in the Giardia genome.
Figure 3. Lysis of G. intestinalis without entrapment in agarose blocks and separation of chromosomal material by FIGE. Lanes 1 and 2 contain DNA from G. intestinalis parasites which were lysed in blocks as described for Figures 1 and 2. Lanes 3 and 4 contain DNA from whole parasites which were lysed directly in the loading slots. The conditions for electrophoresis were 250V, 150mA for 16h. Pulse times were 22 sec in the forward direction and 7 sec in the reverse direction.
Page 150
Figure 4. Electrophoretic analysis of 100kb 'minichromosomes'. The broad band of approximately 100kb seen in Figure 2 was extracted from Seaplaque agarose and analyzed on 0.8% Seakem agarose. Lane 1 contains 'minichromosome' DNA from G. intestinalis; lane 2 bacteriophage lambda (48.5kb); lane 3 1kb ladder from BRL. The top band is 12kb in length; lane 4 is lambda DNA cleaved with EcoRI and Hind III. The top band is 21kb in length.
Discussion We have described an initial approach to establishing formal genetic studies in G. intestinalis. Since it is not possible to observe the chromosomes of G. intestinalis reliably by classical means, that is metaphase spreading and staining, which seems to be a frequent problem among the protozoa, we have attempted to separate the chromosomes by gel electrophoretic techniques. These have included the recently devised techniques of OFAGE and FIGE. Electrophoresis of malaria chromosomes yielded well separated species which can be compared favourably with published PFG separations. In contrast, most of the Giardia chomosomal material remained in the loading slot of the gel, even after two days of electrophoresis. The majority of the chromosomes are therefore very large (>4Mbp) or are entrapped in a network in the parasite and do not migrate through the agarose. Since the latter conclusion is probably incompatible with the necessary separation of chromosomes during binary fission of the parasite during reproduction, we conclude that the chromosomes are very large. Similar large chromosomes have been seen in T. brucei (23,24). Even if separation were possible, say by longer electrophoretic times, or by improved electrophoretic techniques unavailable at present, minor changes would not be informative for genetic analysis (there would have to be major differences in size of chromosomes which are already large, to detect any genetic changes). This is in contrast to the situation with malaria and the trypanosomes (11,12,15,23,24,25), where large differences in small chromosomes, both between species and within strains are observed. Alternatively, the large chromosomes of mammals can be identified by metaphase spreading, and regions mapped by consistent staining of bands (18) relative to the centromeres and telomeres. Furthermore, the identification of genetic changes, such as restriction fragment length polymorphisms (RFLPs), which may or may not be related to a known gene, allow mapping at both the genetic and DNA levels (4,5), and ultimately to the spread chromosomes observed by microscopy. Since none of these choices is available presently for the whole Giardia genome, it may be possible to dissect 'artificial chromosomes', which are shorter than 4Mbp and can be separated by FIGE. One should be able to map genes and RFLPs to these molecules and observe classical genetic changes, such as deletions, translocations and inversions, both by direct observation after electrophoresis and by hybridization studies. We are exploring this possibility currently, using enzymes such as NaeI and SacI, which theoretically would cleave the Giardia genome once every 62.5kb and 250kb, respectively (see ref. 5 for calculation). These enzymes recognize GC rich sequences and contain the CG doublet also. We are also using enzymes which have 8bp recognition sequences, such as SfiI and NotI, and theoretically would cleave the Giardia genome once every 400kb. Sufficient overlap of these 'artificial chromosomes' should build up to the complete genome structure. Although the majority of the Giardia genome is contained in very large chromosomes, as assessed by electrophoretic separation, DNA bands were observed which entered the gel and migrated reproducibly. Some structure was seen also in the smear of DNA migrating between 50kb and the slot under OFAGE conditions, but was not analyzed here. When the Giardia chromosomes were subjected to FIGE, a consistent band of DNA was seen migrating at approximately 100kb. When this broad band was extracted from Seaplaque and electrophoresed again, it migrated as a tight, discrete band, whereas malaria DNA extracted from a similar region of the gel, migrated as a diffuse, polydisperse smear of species. This band is not seen when DNA is directly extracted from parasites because the band migrates in the bulk of genomic DNA during electrophoresis. Since the agarose blocking of parasites has been a very reliable method of preparing chromosomes for gel analysis, the blocks can be stored for months without DNA degradation and the malaria chromosomes which we prepared migrated as expected from published results, we conclude that the 100kb band is a
Page 151
property of the Giardia genome. The diffusion of the band seen in pulsed field conditions is probably due to differential retardation of the 'minichromosomes' as they leave their site of lysis, and to the diffusion of large molecules under changing electrical fields. As to the nature of the 'minichromosomes', we conclude that they are of a very discrete size class, they contain repetitive sequences, which we have detected after restriction enzyme cleavage and by hybridization to cloned repetitive sequences, but that the majority of the repeats are not tandemly arrayed, as might be expected for amplified rDNA, for example (unpublished data). Cleavage patterns are not consistent with the 'minichromosome' being a unique species, but rather a collection of predominantly single copy sequences in molecules of the same length. They have been observed in a number of isolates (unpublished data), so they appear to be a part of the structure of the G. intestinalis genome. Their unique size differentiates them from the minichromosomes of T. brucei, where they are approximately 100 in number, varying from 50kb to 150kb in length. These are thought to play a predominant role in the activation of VSGs in the telomeres (2,11,24). The role of antigenic variation in Giardia is yet to be assessed. The separation of Giardia 'minichromosomes' leaves the majority of the DNA in the slot. The 'minichromosomes' are unlikely therefore to be generated by a 'systematic' disruption of the large Giardia chromosomes, otherwise the remaining segments should enter the gel as readily. Furthermore, their unique size suggests that they are a newly described class of minichromosomes. The larger species at 0.8Mbp and 2Mbp, as seen by FIGE after direct lysis in the loading well, have yet to be analyzed further. In conclusion, we have described very large chromosomes of G. intestinalis and smaller minichromosome counterparts of a unique size class. We have suggested how the large chromosomes can be converted into 'artificial chromosomes' which can be separated by field inversion electrophoresis. Coupled with cloned cDNA and genomic libraries which we have constructed, containing both unique and interpersed repetitive sequences, a base has been developed from which a systematic genetic analysis for Giardia can be approached. Acknowledgements We thank the National Health and Medical Research Council of Australia for support. Literature Cited 1. Anderson, S. 1981. Shotgun DNA sequencing using cloned DNase I generated fragments. Nucl. Acids. Res. 9:30153027. 2. Bernards, A., L.H.T. Van der Ploeg, W.C. Gibson, P. Leegwater, F. Eijgenraam, T. De Lange, P. Weijers, J. Calafat, and P. Borst. 1986. Rapid changes of the repertoire of variant surface glycoprotein genes in trypanosomes by gene duplication and deletion. J. Mol. Biol. 190:110. 3. Boreham, P.F.L., R.E. Phillips, and R.W. Shepherd. 1986. The activity of drugs against Giardia intestinalis in neonatal mice. J. Antimicrob. Chemother. 18:393 398. 4. Bostein, D., R.L. White, M. Skolnick, and R.W. Davis. 1980. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 32:314331. 5. Brown, W.R.A., and A.P. Bird. 1986. Longrange restriction site mapping of mammalian genomic DNA. Nature 322:477481. 6. Burns, D.M., and I.R. Beacham. 1983. A method for the ligation of DNA following isolation from low melting temperature agarose. Anal. Biochem. 135:4851. 7. Carle, G.F., and M.V. Olson. 1984. Separation of chromosomal DNA molecules from yeast by orthogonalfield alternation gel electrophoresis. Nucl. Acids Res. 12:5647 5664. 8. Carle, G.F., M. Frank, and M.V. Olson. 1986. Electrophoretic separations of large DNA molecules by periodic inversion of the electric field. Science 232:6568. 9. Djordjeciv, M.A., W. Zurkowski, and B.G. Rolfe. 1982. Plasmids and stability of symbiotic properties of Rhizobium trifolii. J. Bacteriol. 151:560568. 10. Giannini, S.H., M. Schittini, J.S. Keithly, P.W. Warburton, C.R. Cantor, and L.H.T. Van der Ploeg. 1986. Karyotype analysis of Leishmania species and its use in classification and clinical diagnosis. Science 232:762765. 11. Gibson, W.C., and P. Borst. 1986. Size fractionation of the small chromosomes of Trypanozoon and Nannomonas trypanosomes by pulsed field gradient gel electrophoresis. Mol. Biochem. Parasitol. 18:127140. 12. Gibson, W.C., and M.A. Miles. 1986. The karyotype and ploidy of Trypanosoma cruzi. EMBO J. 5:12991305. 13. Green, M.R., and R.G. Roeder. 1980. Definition of a novel promoter for the major adenovirus associated virus mRNA. Cell 22:231242. 14. Gubler, U., and B.J. Hoffman. 1983. A simple and very efficient method for generating cDNA libraries. Gene. 25:263269. 15. Kemp, D.J., L.M. Corcoran, R.L. Coppel, H.D. Stahl, A.E. Bianco, G.V. Brown, and R.F. Anders. 1985. Size variation in chromosomes from independent cultured isolates of Plasmodium falciparum. Nature 315:347350. 16. Ketner, G., and T.J. Kelly, Jr.. 1976. Integrated simian virus 40 sequences in transformed cell DNA: analysis using restriction endonucleases. Proc. Natl. Acad. Sci. U.S.A. 73:11021106. 17. Nash, T.E., T. McCutchan, D. Keister, J.B. Dame, J.D. Conrad, and F.D. Gillin. 1985. Restriction endonuclease analysis of DNA from 15 Giardia isolates obtained from humans and animals. J. Infect. Dis. 152:6473. 18. Priest, J.H. 1977. Medical cytogenetics and cell culture. 2nd ed. Lea & Febiger, Philadelphia. 19. Schwartz, D.C., and C.R. Cantor. 1984. Separation of yeast chromosomesized DNAs by pulsed field gradient gel electrophoresis. Cell 37:6775. 20. Shenk, T.E., J. Carbon, and P. Berg. 1976. Construction and analysis of viable deletion mutants of simian virus 40. J. Virol. 18:664671. 21. Southern, E.M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503517. 22. Upcroft, J., S. Chiu, and C. Kidson. 1984. In vitro translation of Plasmodium falciparum proteins. Aust. J. Exp. Biol. Med. Sci. 62:125135. 23. Van der Ploeg, L.H.T., A.W.C.A. Cornellisen, J.D. Barry, and P. Borst. 1984. Chromosomes of kinetoplastida. EMBO J. 3:31093115. 24. Van der Ploeg, L.H.T., D.C. Schwartz, C.R. Cantor, and P. Borst. 1984. Antigenic variation in Trypanosoma brucei analyzed by electrophoretic separation of chromosomesized DNA molecules. Cell 37:7784.
Page 152
25. Van der Ploeg, L.H.T., M. Smits., T. Ponnadurai, A. Vermeulen, J.H.E.Th. Meuwissen, and G. Langsley. 1985. Chromosomesized DNA molecules of Plasmodium falciparum. Science 229:658661. 26. Wahl, G.M., M. Stern, and G.R. Stark. 1979. Efficient transfer of large DNA fragments from agarose gels to diazobenzyloxymethylpaper and rapid hybridization by using dextransulfate. Proc. Natl. Acad. Sci. U.S.A. 76:36833687. 27. Woo, P.K. 1984. Evidence for animal reservoirs and transmission of Giardia infection between animal species. pp. 341364 In: Giardia and Giardiasis Biology, Erlandsen, S.L., and E.A. Meyer Eds. Plenum Press, New York. 28. YanischPerron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13 mp18 and pUC19 vectors. Gene 33:103119.
Page 153
The Partial Characterization of an Immunodominant Antigen of Giardia intestinalis Jacqueline A. Upcroft*, Anthony G. Capon, Araya DharmkrongAt, Peter Upcroft and Peter F.L. Boreham Queensland Institute of Medical Research, Bramston Terrace, Herston, Brisbane, Qld, 4006, Australia. Immunoblots of antigens from 13 isolates of Giardia intestinalis have revealed a 32kD immunodominant antigen which is recognized by antisera raised against cloned isolates. This antigen has been partially purified by electroelution from polyacrylamide gels and has been used to raise a polyclonal antiserum in rabbits. When this antiserum was tested by immunofluorescence assay, flagellae and some surface components were recognized primarily and on immunoblots the 32kD antigen was the only reacting product. Several proteins produced by in vitro translation of G. intestinalis mRNA were immunoprecipitated with an antiserum raised against whole trophozoites. However, only one of these, a major translation product of 32kD was immunoprecipitated with the antiserum raised against the partially purified 32kD antigen. cDNA prepared from Giardia mRNA and genomic DNA have been cloned into expression vectors of E. coli and seven clones which express Giardia antigens have been analyzed.
Introduction The identification of one or more antigens antigens which are unique to Giardia intestinalis is a prerequisite for the development of an improved diagnostic test of high sensitivity and specificity. There is an urgent need for such a test to identify both symptomatic and asymptomatic cases for giardiasis, as current methods, depending on either parasitological examinations or invasive procedures, are time consuming and miss many positive cases (5). It is clear from several reports that differences among Giardia isolates can be detected at the antigenic (17), enzymic (1) and genomic (14) levels. Any new diagnostic test must take into account such variation and be able to detect parasites from all geographic locations. Qualitative and quantitative antigenic differences among four isolates of G. intestinalis from Afghanistan, Oregon (USA), Ecuador and Puerto Rico have been demonstrated by crossed immunoelectrophoresis (17). This study employed antisera raised against whole parasites and partially purified antigens. A major surface antigen of 82kD, common to the above four isolates, has been detected by surface labelling techniques (8). Other minor surface labelled proteins of 180, 105, 63, 55, 37, 30 and 24kD were also identified (8). Some of these proteins were secretory/excretory products and differences between the secretory/excretory proteins of two isolates have been reported (13). It is unclear on what basis any of these differences occur and it appears they are neither related to host specificity nor to the geographic location (14). We have been studying the antigens of G. intestinalis in order to identify those which are common to different isolates and which are potentially useful as diagnostic reagents. We have isolated one such antigen of molecular weight 32kD and shown it is a surface antigen. We report on partial purification, characterization and a preliminary search for a recombinant DNA clone which expresses it. Results Detection and Characterisation of an Immunodominant Antigen 1. Detection of a Common Immunodominant Antigen Our initial observations confirmed other studies (17) that analysis of Coomassie Blue stained G. intestinalis proteins separated by polyacrylamide gel electrophoresis (PAGE) reveal no obvious differences among isolates. However, immunological detection of G. intestinalis antigens which have been electroblotted from polyacrylamide gels onto nitrocellulose (immuno or Western blot) (19) revealed many differences (Figure 1). Eleven isolates from Brisbane and environs, one from Western Australia, one from Papua New Guinea and the Portland 1 isolate were used in these studies and were reacted on immunoblots with antisera raised in rabbits (2) against six cloned lines of G. intestinalis (3) (BRIS/82/HEPU/41/1/6; BRIS/83/HEPU/99/1/4; BRIS/83/HEPU/106/1/1; BRIS/83/HEPU/106/1/5; BRIS/83/106/1/7; BRIS/83/HEPU/120/1/13). All of the isolates examined have a major, immunodominant antigen at 32kD which is recognized by all of the antisera used. This antigen or group of antigens appears as a wide band in the immunoblot (Figure 1). 2. Partial Purification of the 32kD Antigen Preparative 12.5% polyacrylamide gels were used to separate the 32kD protein. 2 × 108 trophozoites were loaded onto the gels after solubilization in SDS buffer. Following separation of the proteins, the 32kD band was excised from the gel and the protein electroeluted into a 7.5M ammonium * Corresponding author.
Page 154
acetate cushion using an International Biotechnologies Inc. model UEA electroeluter following the method given by the manufacturers. The eluate was then dialyzed overnight against phosphate buffered saline, pH 7.2, and concentrated by lyophilization. 3. Immunofluorescence Studies With Antiserum Raised Against the 32kD Antigen Rabbit antiserum raised against the 32kD band was used in immunofluorescence assays (IFA) to visualize the location of the 32kD antigen in the parasite. Rabbits were inoculated directly into lymph nodes with the partially purified protein from 2 × 108 trophozoites, mixed with Freunds complete adjuvant in three successive inoculations (2). This antiserum had a titre of 1 in 4000 by IFA against fixed parasites. Two methods were used in the immunofluorescence tests. Firstly, live washed trophozoites were incubated with antibody against the 32kD antigen, gently washed to remove
Figure 1. Immunoblot of proteins of two isolates of G. intestinalis BRIS/82/HEPU/41 and BRIS/85/HEPU/449 separated by P. A. G. E.. Lanes 1, 2 and 3 were reacted with antisera raised against cloned lines of G. intestinalis: BRIS/83/HEPU/99/1/4, BRIS/83/HEPU/106/1/1 and BRIS/83/HEPU/106/1/5, respectively.
unbound antibody, allowed to adhere to glass slides and then acetone fixed. Secondly, antiserum was reacted with parasites after acetone fixation. The binding of antibody to parasites was detected by use of a second fluorescein labelled antirabbit antibody. The antibody to the 32kD protein reacted primarily with flagellae and to a lesser extent the surface of the trophozoite (Figure 2). Both techniques revealed similar fluorescence patterns. However, there was a greater intensity of fluorescence in those organisms which have been acetone fixed prior to exposure to the antibody. 4. Immunoblot of Antigens Reacted with the Anti 32kD Serum Immunoblots of parasites reacted with the anti 32kD antiserum (Figure 3) consistently showed only the major 32kD band. This result suggests that the 32kD protein or group of proteins is not a degradation product of a larger moiety. We have shown that there is a 32kD immunodominant antigen(s) present in G. intestinalis which is common to all the isolates so far examined. This protein appears to be associated with the surface of the trophozoite and in particular with flagellae and related structures. Holberton and Ward (11) who have been studying microtubules of Giardia have identified a group of proteins of approximately 30kD which are associated with the flagellar membranes and the cytoskeleton of the parasite. This group of proteins, together with tubulin (52.5kD) are derived from the ventral disc and axonemes of Giardia. They migrate as a broad band when separated by PAGE and represent a group of possibly eight related polypeptides, called giardins. In addition, further characterization of this 30kD group of proteins has revealed that some of them are unrelated to the giardins and derived from the flagellar membrane or the paraxial rods beneath the membrane. If the common 32kD protein we have identified is indeed
Figure 2. Indirect IFA of G. intestinalis trophozoites (x1000). Isolate BRIS/85/HEPU/449 reacted with rabbit antiserum raised against the partially purified 32kD protein.
Page 155
Figure 3. Immunoblot of proteins of G. intestinalis isolate BRIS/85/HEPU/449 separated by P. A. G. E.. Lane 1: reacted with antiserum prepared against a cloned line of G. intestinalis BRIS/83/HEPU/106/1/5. Lane 2: reacted with an antiserum raised against the partially purified 32kD protein from isolate BRIS/83/HEPU/449.
associated with the flagellar membranes, and is the same as that described by Crossley et al. (7), it might be expected to be a major surface antigen and therefore detected as a prominently iodinated protein in studies on surface labelling of parasites (8,13). Interestingly both Einfields and Stibbs (8) and Nash et al. (13) have identified, in addition to a major 82kD surface antigen, a minor surface protein of approximately 32kD. However, if the protein has a low content of tyrosine and less importantly is low in histidine, tryptophan and sulphydryl groups, it may not be iodinated and identified by surface labelling techniques but may still be an abundant surface antigen. Cloning of Giardia antigens 1. Identification of the 32kD Antigen in in vitro Translations of mRNA A group of proteins recognized by antibody as a broad band in immunoblots may indicate that a particular protein is glycosylated variably such that the size of the glycosylated protein is not constant or that it represents a group or family of proteins which are modified posttranslationally. Many antigens of protozoan parasites are known to be glycosylated (6,12,15) and G. intestinalis may well be similar. In vitro translation of mRNA results in the production of proteins which are synthesized directly from the mRNA and which have not undergone any intracellular post translational modifications, for instance, glycosylation or removal of peptide sequences. G. intestinalis RNA prepared by lysis of parasites in SDS buffer containing 10mM ribonucleoside vanadyl complexes (BRL) as RNase inhibitor and phenol/chloroform extraction (20), was allowed to adsorb onto Hybond messenger affinity paper (Amersham) and the mRNA which was specifically adsorbed was subsequently eluted (23). The mRNA was translated in vitro in a rabbit reticulocyte cell free system (16) and analyzed by PAGE (Figure 4). The translation products ranged in size from low to high molecular weight proteins. When these translation products were reacted with antisera raised against whole parasites using the method of Taylor and colleagues (18), several major translation products were immunoprecipitated. These included a 32kD translation product. The antiserum raised against the partially purified 32kD band immunoprecipitated only the 32kD translation
Figure 4. Immunoprecipitation of translation products of G. intestinalis mRNA. Lane 1: whole mRNA translation products. Lanes 23: immunoprecipitation of whole translation products with antisera raised against whole parasites. Lane 4: control of immunoprecipitation, minus antibody. Lane 5: immunoprecipitation with normal rabbit serum. Lane 1 was exposed for a shorter period than 25. Lane 3 had twice as many counts as lane 2. The arrow indicates the translation product immunoprecipitated with antiserum raised against the 32kD antigen.
Page 156
product (arrowed in Figure 4). 2. Cloning of Giardia intestinalis DNA Following the positive identification of immunoprecipitable translation products we synthesized cDNA (9) from the mRNA which codes for the in vitro translation of this 32kD protein and cloned the .Ix cDNA cDNA into the E. coli vector pUC19 (21). Genomic DNA from Giardia was partially cleaved with Rsa 1, Alu 1 and DNase 1 and cloned by the same method into pUC13 and pUC19 (21). Antisera raised against whole parasites from mixed isolates and antisera raised against the 32kD band have been used as the first antibody in a colony assay to detect colonies which express Giardia antigens (10). The second antibody was biotinylated and the detection system used streptavidin and biotinylated horseradish peroxidase (22). To confirm that the colonies do express Giardia antigens, bacterial lysates were inoculated into mice (4) and the resulting antiserum tested by IFA to determine which antigen the bacterial clones were expressing. At present seven bacterial clones which react with antisera raised against whole, mixed parasites have been tested against antisera raised against cloned isolates, but none appear to react with antisera raised against the 32kD band. None of the clones reacted with all of the antisera which indicates that the common 32kD protein is not represented in these clones. The rabbit which was used to raise the anti32kD antisera had a very high level of E. coli antibodies; consequently the background in the colony assays was high despite repeated adsorptions with sonicated E. coli cells carrying the plasmid vector without an insert. More recently we have raised antisera in specific pathogen free (SPF) rabbits and the background reaction against E. coli is greatly reduced. The 32kD antigen appears to be a major component of the in vitro translation products of mRNA from Giardia since the antiserum raised against the 32kD band immunoprecipitated a major translation product of the same size. These data show that some, if not all, of the antibody in the antiserum raised against the 32kD band are directed against polypeptide and not glycosylated epitopes. Thus this antiserum is likely to recognize the 32kD antigen expressed as a foreign polypeptide when cloned into an expression vector of E. coli, such as the pUC series of plasmids. Our initial search for Giardia; antigens in our libraries has proved successful and we have identified clones which express Giardia antigens in both cDNA and genomic DNA libraries. We are presently screening these libraries with antisera raised in SPF rabbits, against the 32kD antigen. Conclusion We have identified a surface antigen of G. intestinalis which appears to be associated primarily with the flagellae and surface components of the trophozoite. This antigen has a molecular weight as determined by PAGE of 32kD and it may be the same protein that Holbertan and coworkers have identified as associated with flagellar membranes. A common surface antigen, such as the 32kD antigen, associated with flagellae may prove useful in the development of a diagnostic reagent for G. intestinalis since flagellar proteins are likely to be present in stools even if intact organisms cannot be seen. Work is currently being undertaken to determine whether these proteins can be detected in stools by ELISA. The 32kD antigen is a major component of in vitro translation products of mRNA from G. intestinalis and there is no evidence of any gross posttranslational modification. We have produced libraries of G. intestinalis DNA cloned into expression vectors of E. coli and have screened these libraries for expression of G. intestinalis antigens. Preliminary searches in our cloned libraries have revealed the expression of some Giardia antigens but as yet none that can be confirmed as the 32kD common antigen. Acknowledgements The original research described in this paper has been supported by a grant from the National Health and Medical Research Council of Australia. A.G.C. is a National Health and Medical Research Council Medical Postgraduate Research Scholar. Literature Cited 1. Bertram, M.A., E.A. Meyer, J.D. Lile, and S.A. Morse. 1983. A comparision of isozymes of five axenic Giardia isolates. J. Parasitol. 69:793801. 2. Boreham, P.F.L., and G.S. Gill. 1973. Serological identification of reptile feeds of Glossina. Acta Trop. 30:356365. 3. Boreham, P.F.L., R.E. Phillips, and R.W. Shepherd. 1987. Heterogeneity in the response of clones of Giardia intestinalis to antigiardial drugs. Trans. R. Soc. Trop. Med. Hyg. 81:406407. 4. Coppel, R.L., G.V. Brown, G.F. Mitchell, R.F. Anders, and D.J. Kemp. 1984. Identification of a cDNA clone encoding a mature blood stage antigen of Plasmodium falciparum by immunization of mice with bacterial lysates. Embo J. 3:403407. 5. Craft, J.C. 1982. Giardia and giardiasis in childhood. Pediatr. Infect. Dis. 1:196211. 6. Cross, G.A.M. 1978. Antigenic variation in Trypanosomes. Proc. R. Soc. Lond. B. 202:5572. 7. Crossley, R., J. Marshall, J.T. Clark, and D.V. Holberton. 1986. Immunocytochemical differentiation of microtubules in the cytoskeleton of Giardia lamblia using monoclonal antibodies to tubulin and polyclonal antibodies to associated lowmolecularweight proteins. J. Cell Sci. 80:233252. 8. Einfeld, D.A., and H.H. Stibbs. 1984. Identification and characterization of a major surface antigen of Giardia lamblia. Infect. Immun. 46:377383. 9. Gubler, U., and B.J. Hoffman. 1983. A simple and very efficient method for generating cDNA libraries. Gene. 25:263 269. 10. Helfman, D.M., J.R. Feramisco, J.C. Fiddes, G.P. Thomas, and S.H. Hughes. 1983. Identification of clones that encode chicken tropomysin by direct immunological screening of a cNDA expression library. Proc. Nat. Acad. Sci. 80:3135. 11. Holberton, D.V., and A.P. Ward. 1981. Isolation of the cytoskeleton from Giardia. Tubulin and a lowmolecularweight protein associated with microribbon structures. J. Cell. Sci. 47:139166.
Page 157
12. Howard, R.F., and R.T. Reese. 1984. Synthesis of merozoite proteins and glycoproteins during the schizogony of Plasmodium falciparum. Mol. Biochem. Parasitol. 10:319334. 13. Nash, T.E., F.D. Gillin, and P.D. Smith. 1983. Excretorysecretory products of Giardia lamblia. J. Immunol. 131:20042010. 14. Nash, T.E., T. McCuthcan, D. Keister, J.B. Dame, J.D. Conrad, F.D. Gillin. 1985. Restrictionendonduclease analysis of DNA from 15 Giardia isolates obtained from humans and animals. J. Infect. Dis. 152:6473. 15. Newbold, C.I., D.B. Boyle, C.C. Smith, and K.N. Brown. 1982. Identification of a schizontandspeciesspecific surface glycoprotein on erythrocytes infected with rodent malarias. Mol. Biochem. Parasitol. 5:45554. 16. Pelham, R.B., and R.J. Jackson. 1976. An efficient mRNAdependent translation system from reticulocyte lysates. Eur. J. Biochem. 67:247256. 17. Smith, P.D., F.D. Gillin, N.A. Kaushal, and T.E. Nash. 1982. Antigenic analysis of Giardia lamblia from Afghanistan, Puerto Rico, Ecuador and Oregon. Infect. Immun. 36:714719. 18. Taylor, D.W, J.S. Cordingley, and A.E. Butterworth. 1983. Immunoprecipitation of surface antigen precursors from Schistosoma mansoni messenger RNA in vitro translation products. Mol. Biochem. Parasitol. 10:305318. 19. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Natl. Acad. Sci. 76:43504354. 20. Upcroft, J.A., S. Chiu, and C. Kidson. 1984. In vitro translation of Plasmodium falciparum proteins. Aust. J. Exp. Biol. Med. Sci. 62:125135. 21. Upcroft, P., and A. Healey. 1987. A rapid and efficient method for cloning of bluntended DNA fragments. Gene 51:6975. 22. Wilchek, M, and E.A. Bayer. 1984. The avidinbiotin complex in immunology. Immunol. Today 5:3943. 23. Wreschner, D.H., and M. Herzberg. 1984. A new blotting medium for the simple isolation and identification of highly resolved messenger RNA. Nucleic Acids Res. 12:13491359.
Page 159
Immunofluorescence Differentiation between Various Animal and Human Source Giardia Cysts Using Monoclonal Antibodies Henry H. Stibbs,* Edward T. Riley, Joseph Stockard, John L. Riggs, Peter M. Wallis and Judy IsaacRenton U.S. Japan Biomedical Research Laboratory, Department of Medicine, Tulane University, Hebert Centre, Bldg 30, Bel Chase, Los Angelos, California 70037, U.S.A.. Mouse monoclonal antibodies have been developed against cysts of Giardia muris and of Giardia simoni isolated from a wild Norway rat captured on the University of Washington campus. The four antiG. muris antibodies reacted positively in indirect immunofluorescence with the rat source cysts in addition to the homologous G. muris, but not with cysts isolated from beaver (2), dog (4), human (8), muskrat (3), or Richardson's vole (1). The one antirat cyst monoclonal antibody reacted only with rat and cow (CW1; Alberta) source cysts. The antiG. lamblia cyst monoclonal antibody provided by J. Riggs was found to react with all human, beaver, and dog source cysts, and with rat Giardia, but not with G. muris or cysts of muskrat or Richardson's vole origin. The results suggest that systematic differences occur in the cyst surface membrane antigens of various Giardia strains, and that monoclonal antibodies may prove useful in developing an antibody typing system for Giardia strain and animal source identification.
Introduction Waterborne Giardia cysts have been implicated as the cause of over 80 outbreaks of giardiasis in the United States over the past 30 years (4,11,15,22). Giardia infected mammals, most notably the beaver, have been implicated as the probable source of Giardia cysts infecting humans in several of these outbreaks (1,6,15,28,30). It has never, however, been possible to prove whether the infecting Giardia strains in these outbreaks were of animal or human origin or both; or to identify with absolute certainty which species of animal(s) (if any) contributed the human infective cysts. Giardia infections are common in a wide variety of wild and domestic mammals and birds (1,3,5,8,14,20,27,28,30); therefore, it is reasonable to assume that in most North American watersheds some animal species will be Giardiainfected and excreting Giardia cysts at any point in time. A portion of the excreted cysts will find their way into the surface runoff. Thus, Giardia sp. cysts can probably be considered omnipresent and ubiquitous in natural surface waters in North America. The infectivity for man of the many animal strains or species of Giardia found in North American watersheds is an almost completely unexplored subject, about which many assumptions and guesses have been made over the years. Controlled experiments involving human subjects have never been performed. Studies involving cross infectivity between animal host species have revealed great variability in hostspeciesspecificity, so that one may not confidently make predictions about the infectivity for man of any animal isolates (1,14,30). Effective methods have been developed for filtering surface water for the purpose of concentrating Giardia cysts and for identifying the cysts in the recovered particles (10,12). Giardia cysts have been identified in surface waters in many areas of North America, and in some cases quantitative data on the concentration of cysts in surface water have been obtained (19,28). However, in using light microscopy or immunofluorescence to examine cysts recovered from filtered surface water, one cannot reliably identify the animal origin, strain or species of the cysts. This is due to the fact that Giardia cysts of all animal and human origins are anatomically very similar (14) and also the fact that Giardia cysts of most, if not all, mammalian sources are equally well visualized by immunofluorescence methods performed with polyclonal anticyst serum (21,22,26). Methods capable of distinguishing antigenically, biochemically, or genetically between the Giardia cysts produced by different animals and by man would enable one to identify by animal source the Giardia cysts that have caused a human outbreak of giardiasis (by testing cysts recovered from patients), and also to identify the animal or human origin of cysts recovered from surface water. Identifying the source(s) of the cysts in water would allow one to estimate the degree to which human fecal pollution of a watershed is contributing human source cysts to the surface water, and also to estimate the relative contributions of various animal species. At present, however, almost nothing is known of the possible antigenic, biochemical, or genetic differences that may exist between cysts of various animal and human source Giardia. Riggs, however, has already reported the use of a mouse monoclonal antibody against human source G. lamblia cysts in differentiating between human and animal Giardia cysts by immunofluorescence (21). This antibody has been found to bind to all human source cyst isolates tested as well as to cysts from beavers and dogs; * Corresponding author.
Page 160
it also has reacted positively with Giardia cysts from one cattle specimen from Yosemite National Park and with cysts from a coyote and a chipmunk. We have also found (see results below) that this antibody reacts strongly with Giardia cysts of the duodenalis morphological type recovered from a wild Norway rat in Washington state. The Riggs antibody does not, however, bind to Giardia cysts produced by a wide variety of other animals, including deer, muskrats, and different types of voles and mice (see results). Smith et al. (25) also reported production of a mouse monoclonal antibody that would react in immunofluorescence with G. lamblia cysts but not cysts from mice (G. muris) or dogs (G. canis). The mice used as the source of spleen cells for the hybridoma fusion had in this case been immunized with in vitro trophozoites of G. lamblia. The existence of antigenic diversity among cysts of human and animal Giardia isolates seems likely in light of previous discoveries of antigenic diversity among trophozoites of human and animal isolates (13,16,18,24) and of differences in agarose gel electrophoresis banding patterns of restriction endonuclease fragments of DNA from human and animal isolates (17). Diversity in isozyme patterns among Giardia trophozoites of human and animal origins has also been reported (2). We have now produced mouse monoclonal antibodies against Giardia muris cysts isolated from mice and against Giardia cysts isolated from wild, livetrapped Norway rats in an effort to see if antibodies can be produced which bind to some or all of the animal source cysts but not to human source cysts. These antibodies will also be used to identify the animal origin of Giardia cysts through the creation and use of an antibody typing system for cysts of this parasite. Methods Giardia Isolates Used in Immunizing Mice Giardia muris was acquired from Dr. Martin Heyworth of the Division of Cell Biology, VA Medical Center, San Francisco, and maintained in BALB/c, Swiss Webster, and nude (nu/nu) mice. This isolate had originally been procured from an infected golden hamster at Case Western Reserve University (23). Two wild Norway rats were livetrapped along a drainage canal located on the University of Washington campus in Seattle and were found to be infected with a Giardia strain of the duodenalis morphological type (presumably G. simoni). The rats were maintained in the laboratory for several months on a diet of commercial rat chow, apples, carrots, and water. Cyst excretion seemed to be relatively constant over the months that the rats were maintained in captivity. Gerbils (Tumblebrook Farms Co., Massachusetts) were inoculated with cysts from the rats, and cysts isolated from successfully infected gerbils used to supplement the cysts obtained directly from the rats for the purpose of immunizing mice. We were not able to infect weanling SpragueDawley or LongEvans rats with this wild rat Giardia isolate. Purification of Cysts Mouse or rat source Giardia cysts were purified from feces by centrifugation over 1 M sucrose in water (10 min, 500 × g) followed by step gradient centrifugation over two layers of Percoll (Sigma Chemical Co., St. Louis, MO.) of specific gravities 1.05 and 1.09 (15 min, 500 × g). Occasionally, cysts were purified by flotation on zinc sulfate (specific gravity, 1.18) followed by centrifugation over 1 M sucrose. Production of Monoclonal Antibodies Female BALB/c mice were immunized against freshly isolated cysts through a series of four to five intraperitoneal injections of 13 × 106 cysts/animal/injection over 6 weeks followed by 1 or 2 intravenous tail injections, 4 days apart, of 23 × 106 cysts in sterile normal saline. Four or 5 days after the last intravenous injection, mice were sacrificed and spleen cell:myeloma cell fusions carried out. Spleen cells from the immunized mice were fused with NS1 (P3NS1Ag 4.1) mouse myeloma cells grown in RPMI 1640 medium with 15% fetal bovine serum, using 40% polyethylene glycol 1500, using conventional methods (9). Cells were plated onto 96 well culture plates and hybridoma growth selected for using RPMI 1640 medium (15% fetal bovine serum) supplemented with HAT (hypoxanthine/aminopterine/thymidine). Growth was enhanced by using mouse thymus cells as feeder cells. Hybridomas secreting antibodies to cysts were detected by indirect immunofluorescence using the homologous cysts airdried onto the bottoms of flatbottom, 96well ELISA plates (Falcon 3915) at 5,000 to 10,000 cysts per well, and also (with G. muris cysts) by ELISA, again with cysts dried down onto the wells of 96well plates, using 25,000 cysts per well. Stable hybridomas were cloned by limiting dilution three times. Ascites fluid was sometimes produced in BALB/c mice pretreated with Pristane (Sigma Chemicals). Both culture supernatant and ascites fluid were used in the immunofluorescence crosstesting studies described below. Immunofluorescence Crosstesting The Giardia isolates used in the immunofluorescence crosstesting experiments with the monoclonal antibodies had diverse animal origins (see Tables 1 and 2). Muskrats were livetrapped on the University of Washington campus (same location where the Norway rats were trapped) and in eastern Washington outside of Ellensburg (the NA, or Naneum, isolates), and were maintained in rabbit cages on a diet of apples, carrots, celery, lettuce, and alfalfa hay. The MR6 muskrat isolate, the C3 vole (Clethrionomys gapperi) isolate, and the CW1 cow isolate were acquired from Peter Wallis in Calgary, Alberta, and were maintained in Swiss Webster mice, or, in the case of the CW1, in gerbils. Giardia infected Microtus ochrogaster were provided by Stanley Erlandsen of the University of Minnesota; the animals had originated in Missouri. The Manastash beaver isolate was from the Ellensburg area of eastern Washington and was provided by Glen Clark of Central Washington University. Dr. Clark also provided the feces of Microtus richardsoni which had been collected at Paradise Creek in Mr. Rainier National Park. The M beaver isolate originated in British Columbia and was provided as cultured trophozoites by Judy IsaacRenton. The D3 dog isolate was adapted to culture and provided by Peter Wallis. The H2, H3, and H4 human isolates were provided by Charles Hibler of Colorado State University in infected gerbils, and were adapted to culture in our laboratory. The TB human isolate came from a patient in Seattle and was adapted to culture; on one occasion it was used to infect a Long Evans rat, from which cysts were obtained. The dog Giardia specimens listed in Table 1 came from dogs housed in the University of Washington vivarium and used for other research purposes. In a number of cases cysts for crosstesting were obtained from gerbils that had been inoculated with either cultured trophozoites or with cysts and immunosuppressed with Dexamethazone (Intensol; Roxane Laboratories, Columbus, OH.; 1.5 mg into 100 ml drinking water); these gerbilderived cysts are indicated as such in the Tables. Cysts for immunofluorescence testing were usually purified by the procedures described above; sometimes, however, cysts were spotted onto the test wells as an aqueous fecal slurry. Cysts were spotted onto eightspot Tefloncoated slides (Bellco, Vineland, N.J.), air dried, and fixed in acetone at room temperature. The slides were usually stored at 75°C with dessicant in tightly sealed boxes until testing. Hybridoma supernatant or ascites was allowed to react with the cysts for about one hour at 37°C (or overnight at 4°C) followed by two brief rinses in 0.0175 M phosphatebuffered saline, pH 7.4 (PBS), and a further onehour incubation with FITC labeled goat antimouse
Page 161
immunoglobulin antibody (Cappel Laboratories, West Chester, PA.) at 1:80 dilution in PBS with 2% normal goat serum. After rinsing, the slides were mounted with 90% glycerol/PBS containing 0.5 mg/mL pphenylenediamine, and viewed using epifluorescence. Results Four stable, cloned hybridoma lines secreting immunofluorescencepositive and ELISApositive monoclonal antibodies were produced against G. muris cysts. All were of the IgG1 class. One stable hybridoma line secreting immunofluorescencepositive IgG1 antibody against the rat source Giardia cysts was also produced. The results of crosstesting these antibodies by indirect immunofluorescence with cysts from the many animal and human origins are shown in Tables 1 and 2. The anti G. lamblia cyst monoclonal antibody of Riggs, directly labeled with fluorescein, was also tested against all of the cyst isolates; the results are included in Table 1. The four G. muris monoclonals all showed an identical pattern of reactivity with the various cyst isolates. All four reacted strongly with cysts of the homologous isolate and also with the wild Norway ratsource Giardia. No binding to cysts of human, beaver, dog, muskrat, or vole origins was observed. Riggs monoclonal reacted strongly with cysts of human, beaver, dog, and rat origins. The rat Giardia TABLE 1. Immunofluorescence reactivity patterns obtained with Giardia cysts of various animals and human sources, after incubation with four antiG. muris monoclonal antibodies and with the antiG. lamblia monoclonal of Riggs.
Source of Giardia
Antibodies 47B
16C
1010B
211B
Riggs
Beaver (M;B.C.;gerb.)
++
Beaver (Manast.;Wa.;beav.)
++
Dog 1 (U.W. vivarium)
++
Dog 2 ''
++
Dog 3 "
++
Dog 4 "
++
Human 1 (patient)
ND
++
Human 2 (TB;rat)
++
Human 3 (H2;Co.;gerb.)
++
Human 4 (H3;Co.;gerb.)
++
Human 5 (H4;Co.;gerb.)
++
Human 6 (patient)
++
Human 7 (patient)
++
Human 8 (patient)
++
Norway rat (U.W.1,2)
++
++
++
++
++
Mouse (G. muris; orig. hamster)
++
++
++
++
Muskrat (U.W.1,2)
Muskrat (E. Wa.;NA2)
Muskrat (MR6;Alb.;mice)
Vole (M. richardsoni;Mt. Rainier)
Vole (M. ochrogaster;Mo.)
Vole (C. gapperi;Alb.;mice)
++ Strong reaction between antibody and cysts Negative reaction ND No data available
TABLE 2. Immunofluorescence reactivity of MAb 6E10 (antiNorway rat Giardia cyst) with cysts of other animal and human sources. Source of Giardia Norway rat (G. simoni;UW2;in feces)
IFA reaction (,+,++) ++
Mouse (G. muris;orig. hamster)
Vole (M. richardsoni;Mt. Rainier)
Vole (C. gapperi;C3;Alb.;mice)
Vole (M. ochrogaster;Mo.)
Muskrats (UW5,UW2)
Muskrat (MR6;Alb.;mice)
Beaver (M;B.C.;gerb.)
Beaver (Manastash;E. Wa.;gerb.)
Dog (D3;Alb.;gerb.)
Cattle (CW1;Alb.;gerb.)
++
Human (2 patients; Wa.)
Human (H2;Colorado;gerb.)
monoclonal bound only to cysts of the homologous ratsource isolate and, oddly, to cysts of the cattle (CW1) isolate from Alberta. We have recently found that neither the antirat Giardia monoclonal nor the Riggs monoclonal will bind to muris type cysts that have been recovered from other Norway rats trapped from the same location as the first two. All of the antibodies, including the one of Riggs, appeared to bind strongly to the cyst wall of those isolates that reacted positively, the pattern of immunofluorescence being evenly distributed around the entire surface of the cysts. No internal structures of the cysts could be seen fluorescing. We did not observe in any case partial reactions of cyst populations with any of these antibodies. However, because some of the test cyst preparations were cysts contained in fecal slurries, it was impossible to preclude by immunofluorescence the possibility that a fraction of the Giardia cysts in some preparations were unreactive. On several occasions, cysts of either mouse or rat origin, stored in a dry state on test slides for six weeks or more at 4°C or thawed and refrozen a number of times from 75°C, failed to react with the antibodies. These cysts would continue to react positively, although weakly, with polyclonal rabbit antisera to G. lamblia cysts. Also, cysts of the H2 isolate, harvested from gerbils, did not react with the Riggs antibody after storage for several weeks in 5% formalin in PBS at 4°C. The monoclonal antibodies also tended to lose reactivity with their homologous antigen after repeated freezing and thawing. Discussion The crosstesting results show that antigenic differences exist between cysts of various animal origins and of human origin, and suggest that it may eventually be feasible to identify cysts in environmental samples (water or feces) according to their animalsource by testing their immunofluorescence reactivity with a battery of selected monoclonal antibodies prepared against cysts. The results also show that vole and muskrat (i.e.
Page 162
microtine) Giardia isolates seem to segregate antigenically from the others in that they did not react with any of the monoclonals used. The fact that the rat and the mousesource Giardia isolates both bind the antimuris monoclonals indicates antigenic relatedness; however, the muris cysts did not bind the antirat Giardia monoclonal. Incidentally, we have also found that all of the cyst isolates shown in the Tables react strongly in indirect immunofluorescence with polyclonal rabbit antiserum against human G. lamblia cysts (data not included in Tables), an observation that corroborates previous results reported by ourselves (26) and by Sauch (22). Therefore, while it appears that all of the cyst isolates share some antigens or at least some epitopes (determinants), crosstesting with monoclonal antibodies has revealed that differences exist in the distribution of certain epitopes between isolates. We still do not know anything of the physicochemical nature of the antigens recognized by these monoclonal antibodies, nor do we know at this time whether the four antimuris monoclonals recognize the same or different epitopes. The precise ultrastructural location of the antigens on or in the cysts also remains to be determined. The antigens may be components of the thin (0.150.2 µm thick) filamentous outer coat of the cyst or of the underlying outer cyst membrane, structures described by Erlandsen et al. (7). In addition, the antigens may have some biochemical or ultrastructural relationship to the chitin present in the cyst walls (29). Finally, the possibility that some of the recognized antigens, particularly if located on the outer filamentous coat of the cyst, may be hostderived cannot yet be discounted. Two other important considerations that relate to the possible practical utility of these antibodies in differentiating between Giardia cyst isolates in nature are (1) the possible cyst antigenic variability that may exist in the Giardia populations found within one host species, especially among host populations from geographically diverse origins; and (2) the possible loss of immunofluorescence reactivity of cysts after storage in various fixatives (e.g. formalin) or in water, and at various temperatures and other conditions of storage and shipping. In conclusion, it appears that antigenic differences exist between Giardia cysts of isolates from various animal and human sources. The use of monoclonal antibodies in differentiating between cyst isolates and thus in identifying the animal or human source of an unknown test cyst isolate may prove to be feasible, although much more needs to be done to study the variables mentioned above and to develop standard practical methods for testing. We are currently continuing this work by trying to develop monoclonal antibodies against the microtine types of Giardia cysts (in particular, cysts from muskrats in Washington) as well as against human and beaver isolates, so that a more extensive battery of antibodies can be used in developing an antibody typing scheme for Giardia cyst identification. While the identification scheme we are proposing would identify the Giardia isolates as to host species origin, it is possible (provided extensive antigenic variation does not occur within isolates) that this information on antigen or epitope distribution among cyst isolates may be helpful in establishing species designations within the genus Giardia. Acknowledgements This research was funded wholly by the United States Environmental Protection Agency, Office of Exploratory Research, under cooperative agreement #R81197001 to H.H.S.; this report does not however, necessarily reflect the views of that agency and no official endorsement should be inferred. The authors gratefully acknowledge the assistance of the Department of Game of the state of Washington in providing a collecting permit for capturing some of the animals used in this work. Literature Cited 1. Bemrick, W.J. 1984. Perspectives on the transmission of giardiasis, In: Giardia and giardiasis: biology, pathogenesis, and epidemiology. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press. New York. pp. 379400. 2. Bertram, M.A., Meyer, E.A., Lile, J.D., and S.A. Morse. 1983. A comparison of isozymes of five axenic Giardia isolates. J. Parasitol. 69:793801. 3. Clark, G.W., and R.E. Pacha. 1984. Animals associated with mountain streams and meadows as reservoirs of Giardia. Abstract #30, 59th Annual meeting, American Society of Parasitologists, Snowbird, Utah. 4. Craun, G.F. 1984. Waterborne outbreaks of giardiasis: current status, In: Giardia and giardiasis: biology, pathogenesis, and epidemiology. S.L. Erlandsen and E.A. Meyer (eds.). Plenum Press. New York. pp. 243261. 5. Davies, R.B., and C.P. Hibler. 1979. Animal reservoirs and crossspecies transmission of Giardia, In: Waterborne transmission of giardiasis, proceedings of a symposium. U.S. Health Effects Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio. pp. 104126. 6. Dykes, A.C., D.D. Juranek, R.A. Lorenz, S. Sinclair, W. Jakubowski, and R. Davies. 1980. Municipal waterborne giardiasis: an epidemiologic investigation. Ann. Int. Med. 92:165170. 7. Erlandsen, S.L., D.G. Schupp, M.M. Januschka, and W.J. Bemrick. 1986. Ultrastructure of Giardia spp. cysts: development and formation of membranes in the cyst wall and their loss during excystation. Abstract #95, 61st annual meeting, American Society of Parasitologists. Denver, CO. 8. Frost, F., B. Plan, and B. Leichty. 1980. Giardia prevalence in commercially trapped mammals. J. Environ. Hlth. 42:245249. 9. Galfre, G., and C. Milstein. 1981. Preparation of monoclonal antibodies: strategies and procedures. Meth. Enzymol. 73:346. 10. Hausler, W.J., Jr., W.E. Davis, and N.P. Moyer. 1984. Development and testing of a filter system for isolation of Giardia lamblia cysts from water. Appl. Environ. Microbiol. 47:13461347. 11. Hopkins, R.S., P. Shillam, B. Gaspard, L. Eisnach, and R.J. Karlin. 1985. Waterborne disease in Colorado: three years' surveillance and 18 outbreaks. Am. J. Pub. Hlth. 75:254257. 12. Jakubowski, W.. 1984. Detection of Giardia cysts in drinking water: state of the art. In: Giardia and giardiasis: biology, pathogenesis, and epidemiology. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press. New York. pp. 286363.
Page 163
13. Korman, S.H., S.M. Le Blancq, D.T. Spira, J. El On, R.M. Reifen, and R.J. Deckelbaum. 1986. Giardia lamblia: identification of different strains from man. Z. Parasitenkd. 72:173180. 14. Kulda, J., and E. Nohynkova. 1978. Flagellates of the human intestine and of intestines of other species. In: Parasitic protozoa, Vol. 2. J.P. Kreiser (ed.). Academic Press. New York. pp. 1138. 15. Lopez, C.E., A.C. Dykes, D.D. Juranek, S.P. Sinclair, J.M. Conn, R.W. Christie, E.C. Lippy, M.G. Schultz, and M.H. Mires. 1980. Waterborne giardiasis: a communitywide outbreak of disease, and a high rate of asymptomatic infection. Am. J. Epidemiol. 112:495507. 16. Nash, T.E., and D.B. Keister. 1985. Differences in excretory secretory products and surface antigens among 19 isolates of Giardia. J. Inf. Dis. 152:11661171. 17. Nash, T.E., T. McCutchan, D. Keister, J.B. Dame, J.D. Conrad, and F.D. Gillin. 1985. Restrictionendonuclease analysis of DNA from 15 Giardia isolates obtained from humans and animals. J. Inf. Dis. 152:6473. 18. Nash, T.E., and A. Aggarwal. 1986. Cytotoxicity of monoclonal antibodies to a subset of Giardia isolates. J. Immunol. 136:26282632. 19. National Park Service. 1985. Field survey of Giardia in streams and wildlife of the Glacier Gorge and Loch Vale basins, Rocky Mountain National Park. Natural Resources Report Series #853. Water Resources Division, Applied Research Branch, National Park Service, U.S., Fort Collins, CO. 20. Pacha, R.E., G.W. Clark, and E.A. Williams. 1985. Occurrence of Campylobacter jejuni and Giardia species in muskrat (Ondatra zibethica). Appl. Environ. Microbiol. 50:177178. 21. Riggs, J.L. 1984. Giardia Methods Workshop Water Quality Technology Conference. American Water Works Assoc.. Denver, CO. pp. 496. 22. Sauch, J.F. 1985. Use of immunofluorescence and phasecontrast microscopy for detection and identification of Giardia cysts in water samples. Appl. Environ. Microbiol. 50:14341438. 23. Schaefer, F.W., III, E.W. Rice, and J.C. Hoff. 1984. Factors promoting in vitro excystation of Giardia muris cysts. Trans. Roy. Soc. Trop. Med. Hyg. 78:795800. 24. Smith, P.D., F.G. Gillin, N.A. Kaushal, and T.E. Nash. 1982. Antigenic analysis of Giardia lamblia from Afghanistan, Puerto Rico, Ecuador, and Oregon. Inf. Immun. 36:714719. 25. Smith, R.D., N.S. Stimler, J.A. Sauch, and F.W. Schaefer, III. 1983. Monoclonal antibodies to Giardia lamblia: potential for identification and speciation. Abstract #173, 58th annual meeting, American Society of Parasitologists, San Antonio, TX. 26. Stibbs, H.H. 1985. Morphological and antigenic comparisons of Giardia isolated from muskrat, beaver, mice, and man. Abstract, Tenth International Symposium on Intestinal Microecology. Society for Intestinal Microbial Ecology and Disease. University of Minnesota, Minneapolis, MN. 27. Wallis, P.M., J.M. BuchananMappin, G.M. Faubert, and M. Belosevic. 1984. Reservoirs of Giardia spp. in southwestern Alberta. J. Wildl. Dis. 20:279283. 28. Wallis, P.M., R.M. Zammuto, and J.M. BuchananMappin. 1986. Cysts of Giardia spp. in mammals and surface waters in southwestern Alberta. J. Wildl. Dis. 22:115118. 29. Ward, H.D., J. Alroy, B.I. Lev, G.T. Keusch, and M.E.A. Pereira. 1985. Identification of chitin as a structural component of Giardia cysts. Inf. Immun. 49:629 634. 30. Woo, P.K. 1984. Evidence for animal reservoirs and transmission of Giardia infection between animal species. In: Giardia and giardiasis: biology, pathogenesis, and epidemiology. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 341364.
Page 165
Comparison of Giardia Isolates by DNADNA Hybridization A. Uji*, P.M. Wallis and W.M. Wenman Departments of Pediatrics and Microbiology, University of Alberta, Edmonton, Alberta, Canada. We have analyzed the DNA restriction endonuclease profiles of several Alberta Giardia duodenalis isolates by Southern blot hybridizations. With the exception of strain WB, these Giardia strains were isolated from a single geographic area in southeastern Alberta. The following strains were analyzed (original hosts are shown in parentheses): H7 (human), H8 (human), WB (human), B5 (beaver), PB1 (beaver), D3 (dog), MR4 (muskrat) and S1 (sheep). Trophozoites were harvested from 1 litre cultures and lysed using 0.5% SDS and 500 µg/mL proteinase K. DNA was extracted with chloroformisoamyl alcohol and phenol and then treated with RNAase. The DNA from all 8 G. duodenalis strains was digested with Bam H1 and electrophoresed in 1% agarose gels. The separated DNA fragments were then Southern blotted onto nitrocellulose filters and hybridized with 32P labelled DNA probes from strains D3 or H7. The Southern blot patterns of all strains analyzed, whether probed under stringent or less stringent conditions, were markedly similar. Likewise, little difference was detected between the pattern using the dog isolate probe (D3) or the human Giardia probe (H7). These data suggest that G. duodenalis strains isolated from a specific geographic area, regardless of their original host, are closely related genetically.
Introduction Giardia duodenalis, an intestinal protozoan, is considered to be a major cause of diarrhea in many countries. In the United States, G. duodenalis was the most commonly identified pathogen in waterborne outbreaks during 1972 and 1981 (5). In Alberta, Canada, giardiasis is the most frequently reported endemic intestinal infection (1). This infection can be transmitted directly from infected individuals (11), or through contaminated foods (2) or water (5). Large community outbreaks of giardiasis have been caused by waterborne transmission (5). Beavers have been implicated in waterborne transmission as a reservoir host for human infection (3), but the mechanism of Giardia transmission between animal species is not clear. In order to investigate the possibility of cross infection among different animal species, it is important to define host specificity and to classify species within Giardia. Earlier workers have investigated biochemical differences in isolates obtained from various mammalian species in an effort to classify them. Nash et al. differentiated 15 Giardia isolates into 9 groups by Southern blot analysis of chromosomal DNA using recombinant plasmids containing Giardia DNA as probes (10). Bertram et al. differentiated 6 isolates into 3 groups by comparing the electrophoretic patterns of Giardia enzymes (4). These studies were performed on Giardia isolates obtained from geographically different areas. In this report, we compared the genetic similarity between Giardia isolates obtained from various mammalian species found in one specific geographic area. Giardia DNA was analyzed by restriction endonuclease digestion and by Southern blot technique using 32P labelled whole DNA as probes. These data may have relevance in the classifiction of Giardia strains and also for understanding the possible cross transmission of giardiasis within geographic areas. Materials and Methods Giardia Isolates The original hosts and location of the Giardia strains analyzed in this study are shown in Table 1. WB, a human isolate (ATCC 30957) originally established in culture at the National Institutes of Health, was obtained from the American Type Culture Collection. Other isolates were established in culture from cysts extracted from faeces of 5 different mammals: beaver (B5, PB1), muskrat (MR4), dog (D3), sheep (S1), and human (H7, H8), found in a single geographic area. The cysts of all isolates except PB1 were used to infect mongolian gerbils (Meriones unguiculatus). Trophozoites were extracted from their duodenums and cultures were established by methods reported previously (13). A deer mouse (Peromyscus maniculatus), livetrapped and cleared of intestinal protozoa with metronidazole, was used for the extraction of PB1 trophozoites. TABLE 1. Sources of G. duodenalis isolates. Strain
Original Host
Date
WB
Human
H7
Human
850803
Calgary
H8
Human
Calgary
B5
Beaver
850619
PB1
Beaver
MR4
Location Afghanistan
Sibbald Meadows Pond Kananaskis Country, 65 km west of Calgary
841123
Lusk Cr. Pond Kananaskis Country, 80 km west of Calgary
Muskrat
850920
Same as B5
S1
Sheep (domestic)
850702
Farm near Strathmore, 45 km east of Calgary
D3
Dog
840816
Calgary animal shelter
* Corresponding author.
Page 166
Figure 1. Agarose electrophoresis of Pst Irestricted DNA from 8 different G. duodenalis isolates stained with ethidium bromide. Lane 1, Molecular size standards (Hind III digested lambda phage DNA); lane 2, WB; lane 3, H7; lane 4 H8; lane 5, B5; lane 6, PB1; lane 7, MR4; lane 8, S1; lane 9, D3.
Isolates were cultured in modified TYIS33 medium (6) using casein digest peptone instead of trypticase peptone, supplemented with penicillin (200 µg/mL), gentamicin (200 µg/mL) and ticarcillinclavulanic acid (500 µg/mL). The organisms were grown at 37°C for 7296 h. DNA Preparation DNA was isolated from oneL culture of each Giardia isolate. The cells were harvested by chilling the culture bottles in ice for 10 min and then centrifuging the medium at 1000 × g for 10 min. After washing twice by centrifugation in cold PBS (pH7.5), the harvested cells were suspended in 2 to 4 mL of 50 mM TrisHCl (pH7.9) containing 50mM NaCl and 10 mM EDTA, and stored at 20°C. Twelve mL of lysis solution [0.1M TrisHCl (pH7.9), 0.1M NaCl, 0.05M EDTA, 0.5% SDS, and 500 µg/mL proteinase K) was added to the thawed cell suspension, and the mixture was incubated overnight at 37°C. The lysate was mixed with onehalf volume of redistilled phenol saturated with 10mM TrisHCl (pH7.9) and onehalf volume of chloroformisoamyl alcohol (24:1), and rotated gently for 30 min at room temperature. To separate phases, this mixture was centrifuged for 10 min at 1000 × g, and the aqueous phase was transfered into a new tube. An equal volume of chloroformisoamyl alcohol (24:1) was added to the aqueous phase and rotated gently for 15 min at room temperature. After centrifuging, heat inactivated RNase (100 µg/mL) was added to the aqueous phase and incubated overnight at 37°C. Then proteinase K (100 µg/mL) was added and incubated for 2 h at 37°C. The mixture was extracted with an equal volume of chloroformisoamyl alcohol (24:1) and DNA was ethanol precipitated. The precipitate was dissolved in 1 mL of 10mM TrisHCl (pH7.5) containing 10mM NaCl and 1mM EDTA, and dialysed overnight against the same buffer. The DNA was ethanol precipitated again and resuspended in 0.5mL to 1mL of the same buffer. DNA Analysis The following 4 restriction endonucleases were used to digest DNA: Bam HI (N. England Biolabs, Inc.), Eco RI (N. England Biolabs, Inc.), Hind III (Boehringer Mannheim), Pst I (Boehringer Mannheim). Approximately 1 µg of DNA per sample was digested with 2 to 5 units of enzyme (7) under conditions suggested by the manufacturer, and electrophoresed overnight at 30V in a 1% agarose gel (9). After ethidium bromide staining of the gel, DNA was transferred to nitrocellulose according to the method of Southern (12). 32
P labelled whole DNA probes were prepared from D3, H7, or WB DNA (8). Specific activity of each probe was 4.69 × 107 CPM/mL for the D3 probe, 1.76 × 107 CPM/mL for the H7 probe, and 1.14 × 107 for the WB probe. The DNA restricted with each of 4 endonucleases was hybridized with D3, H7, and WB DNA probes using stringent conditions and the DNA digested with Bam HI or Eco RI was hybridized with the D3 DNA probe using less stringent conditions. Under stringent conditions, DNA fragments were hybridized with 106 107 CPM of probe at 37°C overnight in 50% formamide. Under less stringent conditions, DNA fragments were hybridized in 25% formamide. Hybridized fragments were detected by autoradiography. Results and Discussion The ethidium bromide staining of DNA fragments restricted with Pst I (Figure 1), Bam HI (Figure 2), Eco RI or Hind III failed to differentiate the Giardia strains we studied. Nash et al (10) also noted a similar restriction endonuclease digestion profile among most of the isolates which they analyzed. The Giardia genome is sufficiently complex to render differentiation on the basis of restriction endonuclease analysis alone impractical. Therefore, our inability to distinguish isolates by agarose gel electrophoresis may reflect technical limitations as much as it
Figure 2. Agarose electrophoresis of Bam HIrestricted DNA from 8 different G. duodenalis isolates stained with ethidium bromide. Lanes are the same as in Figure 1.
Page 167
Figure 3. Southern blot hybridization of Hind IIIrestricted DNA from 8 different G. duodenalis isolates hybridized to 32P labelled WB DNA probe. Lanes are the same as in Figure 1.
may indicate strain similarity. Southern blot analysis was performed using the same strains and restriction enzymes. We used whole nicktranslated DNA from two human strains (WB, H7) and one dog isolate (D3) as probes. The DNA from different isolates displayed very similar hybridization patterns, regardless which probe was employed. Likwise, the fragments generated by the 4 restriction endonucleases utilized were seldom unique with respect to original host. This was the case whether stringent or less stringent hybridization conditions were employed. Hind III appeared to be the most useful enzyme to differentiate Giardia DNA samples (Figure 3). While this enzyme produced at least 10 common bands on Southern blot, a minor fragment (~5kb) appeared to be present in the beaver, muskrat and sheep strains, but absent in DNA from the dog and 3 human strains (H7, H8, and WB). The major hybridization band comon to all DNA samples analyzed was approximately 4.0 kb in the Hind III digest. The other 3 restrictions endonucleases utilized did not clearly differentiate the DNA samples, regardless whether the H7, WB or D3 probes were reacted with these blots. These results suggest that there is very little genetic disparity among the Giardia strains which we analyzed. Nash et al (10) reported the Southern blot analysis of 15 Giardia isolates from a variety of geographic locations. While most restriction enzyme hybridization patterns did not exhibit any differences, they were able to group their Giardia strains into 9 groups. These differences from our results may reflect the number of strains analyzed, the varied geographic origin of strains or the much larger number of restriction enzymes (sixteen) which these workers employed in their study. However, our DNA hybridization results are consistent with the high degree of antigenic relatedness shown by the Giardia strains which we analyzed (14). These results suggest that G. duodenalis isolates obtained from different mammals found in one specific geographic area are genetically very close to each other regardless of the species of their original hosts, and the difference in DNA sequence is so small that whole DNA probes cannot distinguish it. It is likely that some G. duodenalis strains have a broad host specificity and can cause outbreaks of giardiasis by transmission among different animal species. Literature Cited 1. Alberta Social Services and Community Health Epidemiol. Rpt., 1981. 9:14 2. Barnard, R.J. and G.J. Jackson. 1984. Giardia lamblia: The transfer of human infections by foods. In: Giardia and giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 365378. 3. Bemrick, W.J. 1984. Some perspectives on the transmission of giardiasis. In: Giardia and giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 379400. 4. Bertram, M.A., Meyer, E.A., Lile, J.D., and S.A. Morse. 1983. A comparison of isozymes of five axenic Giardia isolates. J. Parasitol. 69:793801. 5. Craun, G.F. 1984. Waterborne outbreaks of giardiasis: current status. In: Giardia and giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 243261. 6. Keister, D.B. 1983. Axenic culture of Giardia lamblia in TYIS33 medium supplemented with bile. Trans. Roy. Ssoc. Trop. Med. Hyg. 77:487488. 7. Maniatis, T., Fritsch, E.F., and J. Sambrook. 1982. Enzymes used in molecular cloning. In: Molecular cloning a laboratory manual. Cold Spring Harbor Laboratory, New York. pp. 104106. 8. Maniatis, T., Fritsch, E.F., and J. Sambrook. 1982. Enzymes used in molecular cloning. In: Molecular cloning a laboratory manual. Cold Spring Harbor Laboratory, New York. pp. 109112. 9. Maniatis, T., Fritsch, E.F., and J. Sambrook. 1982. Gel electrophoresis. In: Molecular cloning a laboratory manual. Cold Spring Harbor Laboratory, new York. pp. 150163. 10. Nash, T.E., McCutchan, T., Keister, D., Dame, J.B., Conrad, J.D., and F.D. Gillin. 1985. Restrictionendonuclease analysis of DNA from 15 Giardia isolates obtained from humans and animals. J. Infect. Dis. 152:6473. 11. Owen, R.L. 1984. Direct fecaloral transmission of giardiasis. In: Giardia and giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 329339. 12. Southern, E.M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503517. 13. Wallis, P.M. and H.M. Wallis. 1986. Excystation and culturing of human and animal Giardia spp. by using gerbils and TYIS33 medium. Appl. Environ. Microbiol. 51:647:651. 14. Wenman, W.M., Meuser, R.U. and P.M. Wallis. 1986. Antigenic analysis of Giardia duodenalis strains isolated in Alberta. Can. J. Microbiol. 32:926929.
Page 169
Differentiation of Giardia duodenalis from Giardia muris by Immobilization in Various Sera D.L. Lehmann and P.M. Wallis* Kananaskis Centre, BioSciences 042, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N1N4. Immobilization of trophozoites by antiserum was used as a criterion to distinguish Giardia muris collected from naturally infected voles from Giardia duodenalis trophozoites obtained from cultured stocks. G. muris trophozoites were obtained from two redbacked voles (Cleithrionomys gapperi) and three meadow voles (Microtus pennsylvanicus). Cultured G. duodenalis strains were obtained from humans (WB, H7), beaver (PB1, B5), a dog (D3), and a sheep (S1). Antisera was raised in rabbits against G. muris trophozoites from a redbacked vole and G. duodenalis trophozoites from cultured human (WB) trophozoites. Serum from an uninfected human and pooled normal rabbit sera were used as controls. G. muris trophozoites were immobilized by rabbit antiWB serum, antiWB serum adsorbed on WB and G. muris, rabbit antiG. muris serum, antiG. muris adsorbed on G. muris and normal rabbit and normal human sera. G. duodenalis trophozoites were immobilized only by antiWB serum and antiWB serum adsorbed on G. muris. Immobilization was shown to be complement mediated. The main contributor to G. muris immobilization was found to be activated serum, rabbit or human. Such a simple test could be used to distinguish human infective strains with a concurrent identification of potential reservoir hosts.
Introduction The differentiation of various species of Giardia has long been a problem and of special interest have been the differences between the Giardia of rodents and those of other mammals. Filice (7) recognized two species of Giardia that were found in mammals; G. duodenalis and G. muris which he distinguished primarily on the morphology of the median body. This method has been used by Grant and Woo (9) and other workers to identify individual Giardia isolates but the method is difficult to reproduce and has not proven to be practical (4). An additional point of differentiation is that the members of the duodenalis group can be cultured while the muris group flagellates have so far proven to be refractive to cultivation (13). Several workers have shown that both mononuclear leukocytes (11,16,17) and serum (10) can be lethal to Giardia trophozoites. Pearson and Steigbigel (15) demonstrated similar results using Leishmania donovani. Vinayak et al. (19) showed that serum antibodies will react with Giardia cysts in immunodiffusion tests. It is the contention of Stevens (18) that no accurate authentication of Giardia species is possible until appropriate immunological techniques are established. Immunological studies using sodium dodecyl sulphatepolyacrylamide gel electrophoresis analysis have been carried out by various workers (6,12,16,21). These data have shown that both differences and similarities can be found between major antigens of various strains of G. duodenalis. Detailed antigen analysis requires very clean preparations that are free of contaminating protein. So far these have only been available from cultured trophozoites and, because no one has yet succeeded in culturing G. muris, these studies have been restricted to G. duodenalis. This paper is devoted to the demonstration of a method to distinguish between G. dudodenalis and G. muris based on the degree of immobilization following exposure to various sera in vitro. TABLE 1. Sources of strains of G. duodenalis Strain
Original Host
Date
B5
Beaver
850619
Sibbald Meadows Pond Kananaskis Country 65 km west of Calgary
PB1
Beaver
841123
Lusk Cr. Pond Kananaskis Country 80 km west of Calgary
D3
Dog
840816
Calgary Animal Shelter
S1
Sheep (domestic)
850702
Farm near Strathmore 45 km east of Calgary
H7
Human
850803
Calgary
WB
Human
840615
Obtained from G. Faubert ATCC No. 30957
* Corresponding author.
Location
Page 170 TABLE 2. % Immobilization of G. duodenalis trophozoites after 30 minutes. Antibody
WB
H7
PB1
B5
D3
S1
Normal AntiWB, 1:1 dilution
30
18
17
28
11
47
AntiWB adsorbed on G. muris, 1:1 dilution
30
15
13
31
8
44
AntiWB adsorbed on WB 1:1 dilution
0
5
0
0
0
6
AntiWB inactivated
2
0
0
0
0
0
Normal Rabbit serum
0
0
5
0
0
0
Rabbit serum inactivated
3
0
0
0
0
0
Normal Human serum
0
0
3
3
3
0
Human serum inactivated
0
0
0
0
0
0
Methods Strains Six stains of duodenalis group Giardia and 25 strains of muris group flagellates were employed during this study. Giardia from the following sources were assumed to belong to the duodenalis group based on their origin and their culturability: Human WB, Human H7, Beaver PB1, Beaver B5, Dog D3, and Sheep S1 (Table 1). Attempts to characterize these trophozoites based on the shape of the median body were inconclusive and no confidence could be placed in this method (4). Giardia muris were obtained from a colony of Swiss Webster mice at the University of Calgary (GM2) and from voles collected (Microtis pennsylvanicus and Cleithrionomys gapperi) in the vicinity of the Kananaskis Field Station (80 km west of Calgary, AB, in the East Slopes of the Rocky Mountains). Trophozoite Preparation For immobilization studies G. duodenalis were concentrated from cultures in TYIS33 culture medium (23 day cultures in 1012 mL medium) by centrifugation at 500 X g for 5 minutes after which they were resuspended in 2 mL of RPMI 1640 medium (Gibco). Murine Giardia were rinsed from the intestine of the host with RPMI 1640, centrifuge concentrated and resuspended in the same fluid. Controls were run for each portion of the experimental work by mixing trophozoites with RPMI and measuring immobilization over a 30 minute period. Antibody Production Antibodies to G. duodenalis and G. muris (from C. gapperi) trophozoites were produced in adult, male albino laboratory rabbits by 4 intraperitoneal injections of intact, living trophozoites at weekly intervals. Approximately 5 × 105 trophozoites were used for each injection and the first three inoculations were diluted 1:1 with Freund's complete adjuvant. One week after the final injection, the rabbit was anaesthetized with ether and exsanguinated. The blood was allowed to clot in centrifuge tubes for 30 minutes and then centrifuged at 15,000 X g for 10 minutes in an IEC ultracentrifuge. The serum was stored at 60°C. Normal human serum was obtained from one of the authors (PMW) who was not infected with G. duodenalis as shown by 5 negative stool examinations. Antibody Inactivation and Adsorption Serum was inactivated by placing it in a water bath at 56°C for 30 minutes. The adsorption of Giardia antibody was accomplished by mixing glass bead fragmented Giardia trophozoites (concentrated from culture medium or host) with equal parts of antiserum. The mixture was placed on a shaker for 30 minutes, after which the suspension was centrifuged at 2000 rpm for 2 minutes, complement (Difco) replaced, and the supernatent retained for experimental work. Immobilization Studies Immobilization studies were performed by mixing one drop of Giardia suspension with an equal sized drop of serum and covering with a slip. Normal antibody was diluted 1:1 so that the concentration of all serum factors would be similar to that found in adsorbed antibody which was diluted equally with a suspension of fragmented G. muris. Preparations were observed initially and after 30 minutes under phase contrast at 400X using a Zeiss photomicroscope. Immobilization was determined by movement, or lack of movement, of the flagellae. Observation of the parasiteserum preparation began at one corner of the slide and progressed until 50 organisms had been seen and the proportion of living forms determined. Counting required 24 minutes. Two replicates of a 50 parasite count were carried out with a total of 100 flagellates observed. This method was similar to the immobilization test (TPI) used to detect serum antibodies against Treponema pallidum (8). Hill et al. (10) used a closely related method to study the effects of human serum on trophozoites and other authors (11,16) have also used in vitro methods to investigate the interactions between leukocytes and Giardia trophozoites. Results Controls In all cases, 95100% of trophozoites suspended in RPMI alone were active after 30 minutes (5 to 0% immobilization). Additionally, parasites mixed with equal volumes of 1:10 complementsaline resulted in 95100% trophozoite activity after 30 minutes. Immobilization of G. duodenalis Trophozoites Very little difference was observed between G. duodenalis (WB) antibody adsorbed on fragmented G. muris (from Microtis pennsylvanicus) and normal antiWB (1:1 dilution). Both caused between 8 and 47% immobilization of the 6 strains of G. duodenalis and agreement within strains was close (Table 2). AntiWB antibody adsorbed on homologous WB trophozoites caused almost no immobilization as did inactivated antiWB antiserum. Normal and inactivated rabbit serum were both inactive against G. duodenalis trophozoites as were normal and inactivated human serum. Immobilization of G. muris Trophozoites Most muris group parasites were completely immobilized (Table 3) by normal antiWB rabbit serum, antiWB adsorbed on G. muris trophozoites and antiWB adsorbed TABLE 3. % Immobilization of G. muris trophozoites after 30 minutes. C. gapperi
Antibody
M. pennsylvanicus 2
1
2
3
Normal AntiWB 1:1 dilution
100
87
100
100
100
AntiWB adsorbed on G. muris, 1:1 dilution
85
82
100
100
100
AntiWB adsorbed on WB 1:1 dilution
85
82
AntiWB inactivated
0
0
Normal rabbit serum
85
90
Rabbit serum inactivated
0
1
Normal human serum
100
71
92
Human serum inactivated
0
0
0
1
Page 171 TABLE 4. % Immobilization of Giardia spp. trophozoites after 30 minutes using antibodies raised against G. muris from C. gapperi. C. gapperi
M. pennsylvanicus
H7
B5
D3
AntiG. muris, 1:1 dilution
100
100
0
0
0
AntiG. muris, inactivated
0
0
0
0
0
AntiG. muris adsorbed with G. muris 1:1 dilution
95
91
0
0
0
Antibody
on WB trophozoites. High levels of inactivation were also observed with normal rabbit and normal human serum while inactivated human and rabbit serum produced no effect. Immobilization was sometimes accompanied by distortion of the main body of the parasite which did not appear to be reversible. Immobilization of G. muris and G. duodenalis Trophozoites by AntiG. muris Serum Antibody in rabbit serum to G. muris (from C. gapperi) when exposed to G. muris trophozoites completely immobilized the organisms within 30 minutes, and, reductions of motility by 95 and 91% were noted after adsorption of the antibody on fragmented G. muris (Table 4). After heat inactivation of antiserum no immobilization effect was noted on G. muris. Normal, adsorbed, or, inactivated G. muris antibody did not immobilize any of the duodenalis group strains (Table 4). Discussion The data indicate that, for the strains of Giardia investigated, there are distinct differences between the duodenalis group and the muris group flagellates. These differences are: 1. G. duodenalis antibody (in rabbit serum) is not adsorbed onto G. muris trophozoites. 2. G. muris antibody (in rabbit serum) will immobilize murine Giardia but has no effect upon duodenalis group flagellates. 3. G. muris antibody is responsible for a small percentage of immobilized G. muris; the main contributor to immobilization of G. muris is the rabbit serum vehicle. 4. G. duodenalis antibody produces varying degrees of immobilization of 6 strains of duodenalis group parasites but does not appreciably influence muris group Giardia; the observed immobilization of G. muris by G. duodenalis antibody is, in the main, attributed to the rabbit serum bearing the antibody as immobilization was found after the G. duodenalis antibody has been adsorbed. 5. Normal sera from the rabbit and uninfected human immobilizes from 71100% of murine parasites while duodenalis group flagellates are almost completely motile. Rabbit serum, from which G. muris or G. duodenalis antibody has been adsorbed by homologous antigen, reacts as normal rabbit serum. Heat inactivation negates the immobilizing effect of various sera upon G. muris with the implication that the immobilizing action of the sera is complement mediated. This is in agreement with the results of Belosevic and Faubert (3) who found that immune serum from mice would lyse G. muris in vitro and that this effect disappeared with inactivation. Because complement is necessary for immobilization, it must be presumed that certain immunoglobulins are involved. Inasmuch as there is no adsorption of human Giardia antibody to murine Giardia antigen (Table 3), it is speculated that cross reacting antibodies may be present in normal human and rabbit sera (2,10,15). The six strains of duodenalis group Giardia investigated are immobilized to various degrees by G. duodenalis antibody (in rabbit serum) and the action is complement mediated. The duodenalis group strains are not adversely effected by G. muris antibody, adsorbed serum, inactivated human or rabbit sera, normal rabbit serum or normal uninfected human serum. Immobilization of murine Giardia by human and rabbit sera may prevent those hosts from becoming infected with muris group parasites if excystation does occur. Cohen (5) indicates that all classes of immunoglobulins can be detected in the gut and, therefore, the intraluminal location of the parasites does not protect them from humoral factors; in this instance, the humoral factors would immobilize the parasites. Aggarwal and Nash (1) showed that Giardia possessing different surface antigens have different patterns of infection and induce different immune responses. These differences are presumbly similar to those that permit the distinction of G. duodenalis from G. muris with known antisera. If flagellates are immobilized in the gut, they would be unable to attach and multiply and would, instead, be evacuated. It is also conceivable that the immobilized Giardia may be acted upon by the cytotoxic actions of intestinal epithelium macrophages as reported by Owen et al. (14) and Smith et al. (16). It is likewise possible that certain antibacterial substances secreted by the mucous membranes may act upon Giardia (20). An additional feature which can be attributed to this investigation is that by using uninfected human serum, immobilization studies can be made upon the Giardia of other species of mammals; the implication here being that potentially human infective strains of Giardia may be distinguished from those species/strains which presumably will fail to infect man, and potential reservoir hosts may be identified. Acknowledgements This work was funded by contracts from Alberta Environment and a grant from the Alberta Environmental Research Trust. Literature Cited 1. Aggarwal A. and T.E. Nash. 1987. Comparison of two antigenically distinct Giardia lamblia isolates in gerbils. Am. J. Trop. Med. Hyg. 36:325332 2. Allison, A.C. 1978. Macrophage activity and nonspecific immunity. Int. Rev. Path. 18:3038 3. Belosevic, M. and G.M. Faubert. 1987. Lysis and
Page 172
immobilization of Giardia muris trophozoites in vitro by immune serum from susceptible and resistant mice. Parasite Immunol. 9:1119 4. Bertram, M.A., E.A. Meyer, D.L. Anderson, and C.T. Jones. 1984. A morphometric comparison of five axenic Giardia isolates. J. Parasitol. 70:530535 5. Cohen, S. 1976. Survival of parasites in the immunocompetent host. In: Immunology of Parasitic Infections. Cohen, S. and K.S. Warren (eds.), pp. 144145, Blackwell Sci. Pub., Oxford, 848 p. 6. Einfeld, D.A. and H.H. Stibbs. 1984. Identification and characterization of a major surface antigen of Giardia lamblia. Infect. Immun. 46:377383 7. Filice, F.P. 1952. Studies on the cytology and life history of a Giardia from the laboratory rat. Univ. Calif. Pub. Zool. 57:53146. 8. Finegold, S.M. and J.W. Martin. 1982. Diagnostic Microbiology. 6th ed., p. 581, Mosby, St. Louis, Mo., 705 p. 9. Grant, D.R. and P.T.K. Woo. 1978. Comparative studies of Giardia spp. in small mammals in southern Ontario I. Prevalence and identity of the organisms with a taxonomic discussion of the genus. Can. J. Zool. 56:13481359 10. Hill, D.R., J.J. Burge, and R.D. Pearson. 1984. Susceptibility of Giardia duodenalis trophozoites to the lethal effect of human serum. J. Immunol., 132:2046 2052. 11. Koester, S.K. and P.G. Engelkirk. 1984. Glass cover slip technique for studying in vitro interactions between Giardia trophozoites and host leukocytes by TEM, SEM, and light microscopy. J. Parasitol. 70:443445. 12. Nash, T.E. and J. Kiester. 1985. Differences in excretory products and surface antigens among 19 isolates of Giardia. J. Infect. Dis. 152:11661171 13. Meyer, E.A. 1976. Giardia duodenalis: isolation and axenic cultivation. Expt. Parasitol. 39:101106 14. Owen, R.L., C.L. Allen, and D.P. Stevens. 1981. Phagocytosis of Giardia muris by macrophages in Peyer's Patch epithelium in mice. Infect. Immun., 33:591 601 15. Pearson, R.D. and P.T. Steinbigel. 1980. Mechanisms of the lethal effect of human serum upon Leishmania donovani. J. Immunol. 125:21952201. 16. Smith, P.D. C.C. Elson, D.B. Keister, and T.E. Nash. 1982. Human host response to Giardia duodenalis. I. Spontaneous killing by mononuclear leukocytes in vitro. J. Immunol., 128:13731376. 17. Srivastava, R.K., Agarwal, A.K. and S.R. Das. 1986. A simple technique for assessment of immuneresistance of hosts against giardiasis. Cur. Sci. 55:1141 1143. 18. Stevens, D.P. 1976. Giardiasis: Immunity, Immunopathology and Immunodiagnosis. In: Immunology of Parasitic Infections, Cohen, S. and K.S. Warren (eds.), Blackwell Sci. Pub., Oxford, 848 p. 19. Vinayak, V.K., Jain, P., and S.R. Naik. 1978. Demonstration of antibodies in giardiasis using the immunodiffusion technique with Giardia cysts as antigen. Ann. Trop. Med. Parasitol. 72:5812 20. Weir, K.P. 1983. Immunology. Churchill Livingstone. p. 14, Edinburgh, London, Melbourne and New York. 178 p. 21. Wenman, W.M., Meuser, R.U. and P.M. Wallis. 1986. Antigenic analysis of Giardia duodenalis strains isolated in Alberta. Can. J. Microbiol. 32:926929.
Page 173
Conserved Sequences of the HSP Gene Family in Giardia lamblia Anita Aggarwal*, P. Romans, Vidal F. de la Cruz and T.E. Nash NIAD, National Institutes of Health, Bethesda, Maryland 20892, U.S.A.. Transformation of Giardia lamblia from trophic form to the cystic form may be induced by a heat shock response. A factor was found in nuclear extract of Giardia trophozoites bound to heat shock element of Drosophila hsp 70. The sequence analysis of the 5' flanking area showed a TATAA box at position 28 to 33 upstream to the initiation site and consensus promoter region at 56 to 69. The level of this factor was significantly increased after heat shock.
Introduction All living organisms examined to date respond to significant temperature increases by activating a specific set of genes called heat shock genes (1). The same genes are also activated by stress signals such as anoxia, cell transformation, glucose deprivation and chemical which interfere with oxidative phosphorylation (2). Heat shock genes also play a role in normal development. The DNA sequences involved in the heat shock response are strongly conserved not only in their proteincoding sequences, but also in their regulatory sequences. The genes for the Drosophila heat shock proteins were among the first eukaryotic genes to be identified and cloned (3,4). Three families of heat shock genes have been identified in Drosophila spp., genes encoding heat shock proteins of about 83 and 70 kDa and genes encoding for small proteins of about 2330 kDa. Conserved 14 base pair (bp) DNA sequences (heat shock elements [HSE]) were first found upstream from the TATA box of the Drosophila 70 kDa heat shock protein (hsp 70) encoding genes (5). All eukaryotic species from plants to humans have HSE's with at least five of the eight consensus nucleotides conserved. Heat shock transcription factors bind to HSE and confer temperature sensitive transcription to the gene (6). The presence of a single HSE is sufficient to regulate temperatureinducible transcription. A Giardia lamblia heat shock gene was identified and its promoter region determined. The reasons for identifying the promoter area of Giardia hsp70 like proteins are the following: a. To determine analogy to other hsp genes. b. To determine the role of the promoter region in transcription of hsps. c. To employ the promoter sequence for other experiments. Materials and Methods Identification of hsp Gene in Giardia Giardia trophozoites were grown axenically and DNA isolated as described elsewhere (7). One µg of DNA was digested with HindIII, electrophoresed in a 1% agarose gel, transferred to nitrocellulose and hybridized with 32P labeled Ppw 229 (3). This vector has a Drosophila hsp70 insert (3.4 kb). After hybridization and washing under low stringency conditions (2 XS SC, 0.1% SDS at 50°C), a broad band of positive signal was found at around 4.2 kb. The positive area was cut from a similarly prepared gel and ligated into HindIII sites of the plasmid vector puc 18. E. coli DH5°C was transformed with the DNAplasmid ligation mixture and the transformants screened with a 32P labeled 3.4 kb fragment of Drosophila hsp 70 (clone 229). One strongly positive clone (pucG.2c) was purified and further characterized. Analysis of Giardia hsp Gene (pucG.2c) The restriction map of pucG.2c 4.2 kb was constructed using combination of different restriction enzymes (Figure 1). The 5' terminal position of the Giardia hsp gene was determined by hybridizing the doubly digested Giardia hsp gene with the 32P labeled 5' fragment of Dhsp70 (the XhoI and HincII fragment of Ppw229). A positive signal was obtained with the fragment numbers 24 (BglPst, PstPst, PstSal) (Figure 1) indicating the location of the 5' end of Giardia hsp 70 gene within these fragments. DNA Sequence Analysis Three fragments, Bg1IIPst, PstPst and PstSal double digests were sequenced by standard techniques using the single stranded vectors M13mp18 and M13mp19 (8). Results and Discussion Analysis of the nucleotide sequence of 1470 base pairs in all six reading frames revealed only one continuous open reading frame from 142 to 1470 bp. The first ATG, the initiation site, was located at 142 (Figure 2) and thereafter, a total of 442 amino acids were present.
Figure 1. Restriction map of hsp Giardia (pucG.2c). The hatched boxes 2, 3, and 4 hybridized with 0.9 kb 5' fragment of Drosophila hsp 70 indicating the 5' termini of Giardia hsp gene. Double digested Bgl IIPst (Box 2), PstPst (Box 3) and PstSal (Box 4) were cloned into vector M13 mp 18 and M13 mp 19 which had been digested with appropriate restriction enzymes. * Corresponding author.
Page 174
Figure 2. The amino acid sequence of the Giardia hsp fragment in 5' 3' orientation. The first methionine* is present at position 142 and a total of continuous 442 amino acids are shown.
Sequence homologies were determined using Gene bank search Dec 10 computer programs. No significant homology was present between the hsp gene of Giardia and Drosophila hsps. Sequence analysis of 5' flanking area showed a TATAA box at position 28 to 33 upstream to the initiation site (Figure 3). Conserved Sequence Element of the hsp We looked for the putative regulatory signals in the 5' flanking region of the Giardia hsp gene by comparison of the
Figure 3. Nucleotide sequence analysis of 5' flanking region of Giardia hsp. Regions corresponding to the consensus heat shock promoter (B) and TATAA box (A) are noted.
Figure 4. Comparison of the Giardia HSE consensus sequences with sequences derived from human, chicken, mouse, T. brucei and Drosophila hsp 70 genes. The underlined nucleotides constitute the bases of potential homology. Altogether, 8 of the bases are conserved. Six of eight fall under the consensus in Giardia.
sequence of this region with HSE consensus sequences (9,10). An HSE related 14 nucleotide sequence (CTTGAGGGCTCCGG) was found at positions 56 to 69 (Figure 3). Its relatedness to the consensus promoter region (CNNGAANNTTCNNG) was 6/8 (Figure 4). Since divergence in the HSE sequences exists amongst different Drosophila hsps and among other species, it is most likely that the consensus region in the Giardia hsp has a function in transcriptional regulation. Role of the Promoter Region in Transcriptional Regulation To determine the temperature dependent regulatory role of this consensus promoter region, Giardia were heat shocked from 5 min to 5 hrs and the amount of transcribed RNA to Giardia hsp determined. RNA was isolated and spotted onto nitrocellulose and hybridized with either 32P labeled 4.2 kb Giardia hsp gene or an actin gene from Dictyostelium. Northern blot analysis is shown in Figure 5. At 42°C compared to 37°C, RNA was 10fold more abundant after 5 hrs of heat shock whereas the amount of Giardia RNA to actin remained unaffected. Therefore, enhanced transcription of Giardia hsp occurred after heat shock showing temperature dependent regulation of the gene.
Figure 5. Analysis of heat shocked RNAs. Giardia were grown axenically and heat shocked for 0 min, 5 min, 10 min, 30 min, 1 hr, 3 hrs and 5 hrs at 42°C, and the RNA isolated by hot phenol nethod (11). 100 ng was spotted onto nitrocellulose in duplicates, hybridized with 32P labeled 4.2 kb Giardia hsp gene, puc 18, the actin gene from Dictyostelium, and ribosomal DNA of Plasmodium falciparum.
Page 175
Literature Cited 1. Lindquist, S. 1986. The heat shock response. Ann. Rev. Biochem. 55:11511191. 2. Atkinson, B.G. and D.B. Walden. 1985. Changes in eukaryotic gene expression in response to environmental stress. D.B. Walden (ed.) Academic Press Inc., N.Y. 3. Livak, K.J., Freund, R., Schweber, M., Wensink, P.C., and M. Meselson. 1978. Sequence organization and transciption at two heat shock loci in Drosophila. Proc. Nat. Acad. Sci. U.S.A.. 75:5613. 4. Schedl, P., ArtavanisTsakonas, S., Steward, R., Gehring, W.J., and M.E. Mirault. 1978. Two hybrid plasmids with D. melanogaster DNA sequence complementary to mRNA coding for the major heat shock protein. Cell 14:921929. 5. Pelham, H.R.B. 1982. A regulatory upstream promoter element in Drosophila Hsp 70 heat shock gene. Cell 30:517528. 6. Topol, J., Ruden, D.M. and C.S. Parker. 1985. Sequences required for in vitro transcriptional activation of a Drosophila hsp 70 gene. Cell 42:527537. 7. Nash, T.E., McCutchan, T., Keister, D., Dame, J.B., Conrad, J. and F.D. Gillin. 1985. Endocuclease restriction analysis of DNA from 15 Giardia lamblia isolates obtained from man and animals. J. Inf. Dis. 152:6473. 8. Sanger, F., Coulsen, A.R., Barrell, B.G., Smith, A.J.J. and B. Roe. 1980. Cloning in singlestranded bacteriophage as an aid to rapid DNA sequencing. J. Mol. Biol. 143:161178. 9. Corces, V., Pellicer, A., Axel, R., and M. Meselson. 1981. Integration, transciption and control of a Drosophila heat shock gene in mouse. Proc. Natl. Acad. Sci. USA 78:70387042. 10. Holmgren, R., Corces, V., Morimoto, R., Blackman, R. and M. Meselson. 1981. Sequence homologies in the 5' regions of four Drosophila heat shock genes. Proc. Nat. Acad. Sci. USA 78:37753778. 11. Maniatis, T., Fritsch, E.F. and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, New York.
Page 177
The Response of Humans to Antigens of Giardia lamblia M.G. Ortega Pierres*, R. Lascurain, R. Arguello Garcia, R. Coral Vazquez, G. Acosta and J.I. Santos Centro de Investigation y de Estudios Avanzados Del IPN, Apartado Postal 14740, Mexico 14, D.F. Surface labelled antigens from Giardia lamblia as well as immunoblotting techniques were used to analyse the humoral response of asymptomatic and symptomatic Mexican pediatric patients to components of G. lamblia trophozoites. The results demonstrated that few radiolabelled surface components are precipitated by all sera. Western blot analysis revealed that several parasite components are constantly recognized by sera from infected patients. However the level of recognition varies with individual sera. The isolation and use of commonly recognized antigens might permit a rational approach to the development of improved immunodiagnostic methods for giardiasis.
Introduction Human giardiasis is an infection caused by the intestinal protozoan Giardia lamblia. This disease has a world wide distribution (25) and is a significant health problem in developing countries (24). Infection with this parasite can be transmitted from person to person (9,18) or by ingestion of food or water contaminated with cysts (3). The clinical spectrum of infection with G. lamblia ranges from asymptomatic passage of cysts to persistent and severe diarrhoea with malabsorption (19). Several studies have indicated that G. lamblia elicits an immune response in its host. This was suggested first by the fact that individuals exposed repeatedly to G. lamblia developed resistance to infection (11). A role for humoral responses in controlling the disease has been suggested since Igdeficient patients are more susceptible to infection with G. lamblia (1). Subsequently, various reports have indicated the presence of circulating antibodies in patients with giardiasis (15,16,20,23). Of particular importance in the study of immune responses to this parasite is the identification of parasite components which activate immune mechanisms during infection. Recently, two groups have reported their findings on the reactivity of human sera from patients with giardiasis to G. lamblia components. One study showed that a G. lamblia surface antigen of 88 kDa was precipitated by two human sera (5) while another reported that a protein with a molecular weight of 31 kDa was mainly recognized in human infections (21). Here, we analysed the reactivity of 29 sera from asymptomatic and symptomatic Mexican pediatric patients to G. lamblia antigens. We used radiolabelled parasite surface antigens and Western blots to analyse a wide range of human sera. These strategies permit fine dissection of humoral responses following infection by providing quantitative and qualitative immunoprecipitation data. Our results showed that few surface G. lamblia components were immunoprecipitated by all sera tested. The patterns of reactivity, as determined by Western blot analysis varied for each individual serum. However, there were some antigens from this parasite which were constantly recognized by all sera tested. Materials and Methods Parasite Cultures Trophozoites of G. lamblia Portland 1 (P1) strain were grown axenically at 37°C in Diamond's modified TYIS33 medium (8) supplemented with 10% heat inactivated calf serum and 250 µg/mL each of streptomcin and ampicillin. Organisms were harvested in late log phase by chilling culture flasks on ice for 30 min, inverting the flasks several times and centrifuging their contents at 250×g for 10 min. The organisms were washed three times in phosphate buffered saline (PBS) and the cell pellet from the final wash was suspended in a small volume of PBS. The concentration of trophozoites was determined with a hemacytometer. Antigen Preparation Trophozoites cultured and harvested as described above were used to obtain parasite antigens. The parasite pellet was resuspended in 10 mM trisHCl pH 8.3 containing 1 mM phenylmethylsulfonylfluoride, 25 mM Nethylmaleimide and 0.5% Triton X100. This suspension was sonicated with six 15 sec bursts in an ice bath and centrifuged at 15 600×g for 30 min to remove debris. Protein concentration of the supernatants was determined by a modified Lowry assay (4). Surface Labelling of Parasites G. lamblia trophozoites grown and harvested as described above were surface labelled with 125I by the lactoperoxidase method (12). After labelling, trophozoites were sonicated under the same conditions as for antigen preparation and the radiolabelled antigens were used for immunoprecipitation assays with human sera. Patients Sera Serum samples were obtained from children (2 to 14 years old) at the Hospital Infantil de Mexico. These sera were from 5 patients with asymptomatic and from 24 patients with various symptoms of giardiasis which included mainly diarrhoea and abdominal pain. Analysis of feces from all patients at the time of blood collection demonstrated only the presence of G. lamblia cysts. Control sera were obtained from 3 children (5 to 6 years old) with no detectable cysts in their feces and no history of giardiasis or symptoms of gastrointestinal disease. Immunoprecipitation Radiolabelled G. lamblia antigens (100,000 cpm) were precipitated with 58 µL of human sera overnight at 4°C. An excess of protein ASepharose then was added and the tubes were placed at 4°C for 2 h. The protein ASepharose beads were washed three times with NET (150 mM NaCl, 5 mM EDTA, 50 mM TrisHCl) containing 0.5% Triton X100 and then were boiled in sample solution for 5 min before analysis by SDSPAGE. * Corresponding author.
Page 178
Electrophoresis Analyses of precipitates and separation of G. lamblia antigens for immunoblotting assays were performed by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDSPAGE). This was carried out by the method of Laemmli (10) using 515% gradient slabs gels. After electrophoresis 125I containing gels were dried and exposed at 70°C to Kodak XOmat AR films. Immunoblotting Antigens separated by SDSPAGE (100 µg protein per cm of gel width) were transferred to nitrocellulose paper and analysed by the method of Towbin et al. (22) as modified by Hanff et al. (7) and Pekkala and Ruoslathi (14). After the transfer, nitrocellulose sheets were transferred to diluting buffer [PBS pH 7.2, containing 4% (wt/vol) BSA and 0.5% (vol/vol) Triton X100] containing 1:50 dilution of human antiG. lamblia sera and incubated with slow shaking for 1 h followed by three 15 min washes in PBS (pH 7.2) containing 1% (vol/vol) Triton X100 (PBST). The sheets were placed in PBS containing 10% (vol/vol) heat inactivated bovine fetal calf serum and horseradish peroxidaseconjugated goat antihuman whole Igs followed by three 15 min washes in PBST. Nitrocellulose sheets were then developed by the addition of a freshly prepared solution of 0.05% (wt/vol) 4chloro1napthol and 0.01% (v/v) hydrogen peroxide in PBS. Results The ability of sera from asymptomatic and symptomatic patients to precipitate the radiolabelled antigens of G. lamblia was determined by immunoprecipitation and qualitative analysis by SDSPAGE. The analysis of total labelled parasite components revealed the presence of approximately twelve bands (Figure 1). Of these there was a single major polypeptide of 82 kDa and less prominent bands of 63, 55, 53, 49, 43, 40, 35, 32, 27, and 24 kDa. Other surface labelled components between 190 kDa and 144 kDa were sometimes detected. Figure 1 shows the precipitation patterns obtained when 5 sera from either asymptomatic or symptomatic patients were used. In these precipitations, radiolabelled antigens with molecular weights of 85, 63 kDa and 55 kDa were observed. The same radiolabelled G. lamblia components were precipitated by all the sera tested in this assay.
Figure 1. Autoradiograph of 125I labelled proteins of G. lamblia precipitated by antiG. lamblia human sera. Soluble surface iodinated components were immunoprecipitated as described above with control human sera (track N), sera from asymptomatic patients (tracks A1A 5) or sera from symptomatic patients (tracks S1S5). Track P1T is the total surface labelled proteins of G. lamblia P1 strain. Molecular weight: (×103).
Figure 2. Western blot analysis of antigenantibody reactions detected when G. lamblia soluble extracts were overlaid with control human sera (tracks marked N), sera from asymptomatic patients (tracks 14) or sera from symptomatic patients (tracks 528) and developed with peroxidase conjugated to antihuman Ig. Molecular weight: 103.
Although high molecular weight components were found to be iodinated, no precipitation of these components was observed with the sera tested. Sera from asymptomatic and symptomatic patients were tested for the reactivity against total soluble G. lamblia antigens by immunoblotting. Control sera from normal individuals with no history of giardiasis were also tested with the same parasite extracts. The spectrum of antigen/antibody reactions detected by Western blotting can be seen in Figure 2. All the sera tested reacted to G. lamblia soluble components with molecular weights between 35 kDa and 200 kDa. This analysis revealed variations in the recognition of G. lamblia components both quantitatively and qualitatively. Among the sera from asymptomatic patients (tracks 14) there were antigens of approximately 121, 90, 46, 44 and 35 kDa which were preferentially recognized by all sera. One of these sera reacted more strongly to a 31 kDa component (track 1). Similar proteins to those recognized by asymptomatic sera were also detected with sera from symptomatic patients (tracks 528). However differences in reactivity were also observed among this sera. Some of them reacted more strongly to antigens of 23 kDa (track 7), 62 kDa and 57 kDa (track 20), 53 kDa (track 22), 44 kDa (track 23), 152 kDa and 57 kDa (track 27). The human sera used as control in this assay (tracks marked N) gave almost no reactivity with the G. lamblia extract. Discussion Several studies of human responsiveness to antigens from parasites in general are aimed towards the development of more sensitive and specific immunodiagnostic tests. Such studies involve the detection of specific antibodies circulating in the blood as well as in other host sample material such as faeces. The use of preparations with more restricted antigenic composition permits a finer dissection of host parasite interactions. This can be achieved by the isolation of radiolabelled proteins from specific parasite compartments. We have used iodinated
Page 179
surface components from G. lamblia to analyse the human antibody response to such parasite molecules. In our surface labelling experiments we found that G. lamblia components of 82, 63, 55, 53, 49, 43, 40, 35, 32, 27 and 24 kDa were accessible to radioiodination. In addition we observed in some occasions less defined bands between 190 kDa and 144 kDa. Some of the bands observed in this study have also been detected in studies reported by Einfeld and Stibbs (6) and Clark and Holberton (2). Nash et al. (13) have found a major polydisperse band of material from 94 kDa to 225 kDa present as a smear and sometimes as a discrete ladder of bands. These variable iodination patterns could be due to minor technical variations (e.g. labelling conditions, culture conditions) or parasite variation (e.g. stable subpopulations or spontaneous occurring phenotypic variants). These possibilities might explain the fact that high molecular weight components were not always detected in our iodination experiments. Further studies regarding the use of cloned organisms as well as a more defined medium and better control of the growth and multiplication of the parasite might allow for a better understanding of surface molecules of G. lamblia. In the study reported here a number of the surface labelled proteins were invariably antigenic in both asymptomatic or symptomatic patients. Thus in all cases we found precipitation of 82, 63 and 55 kDa components. There was however, a more marked response to the 82 kDa protein regardless of clinical features and time course of infection. Recently, Edson et al. (5) have reported an 88 kDa protein from G. lamblia which was precipitated by two human sera from patients with giardiasis. This protein may be similar to the 82 kDa component recognized by all Mexican sera. Hence, these radiolabelled proteins might be potentially useful for diagnosis of giardiasis. Although we detect radiolabelled G. lamblia surface components of high molecular weight (190 to 116 kDa) in our iodination experiments these were not precipitated by any of the human sera tested. Recently, Taylor and Wenman (21) reported a major 31 kDa G. lamblia antigen recognized during human infections. In the study reported here we did not detect a major reactivity of the sera tested towards this component. The analysis of the reactivity of the human sera by immunoblotting showed that several other proteins of G. lamblia with molecular weights of 121, 90 46, 44 and 35 kDa were preferentially recognized by most of the sera tested. There were, however, differences between the individual sera used. This might be due to the fact that humans are an outbred population and this might suggest a more widely varying response to the complex array of antigenic determinants presented during the course of parasitic infection. As yet we have failed to note consistent differences in the pattern of reactivity of asymptomatic and symptomatic sera. On the other hand, the observed pattern of recognition of G. lamblia antigens might also be influenced by the balance of different Ig isotypes as well as by the titre of antibodies present at the time of sample collection. In spite of these potentials for variability, the present study has allowed the identification of G. lamblia components which might potentially be useful for diagnosis of the disease. An interesting aspect will be to carry out longitudinal studies in humans with giardiasis which will give additional information regarding the recognition of G. lamblia antigens during the course of infection. An alternative approach to the diagnosis of giardiasis and parasitic infections in general is not to detect specific circulation antibodies but specific parasite antigens in suitable host samples such as feces. A recent study by Rosoff and Stibbs (17) used this approach isolating a 65 kDa G. lamblia antigen in stools of parasite positive patients. Thus, the identification and isolation of G. lamblia specific components which are immunogenic in humans or released by the parasite during infection should hopefully provide the basis for a rational improvement of diagnostic test for giardiasis. Acknowledgements We wish to acknowledge Dr. D. Peattie and Dr. R.M.E. Parkhouse for critically reading this manuscript and Mrs. Maria de Lourdes Vazquez for secretarial assistance. R. Lascurain and R. coralVazquez are recipients of a CONACYT studentship. This work was supported in part by CONACYT (Mexico) and the MACARTHUR FOUNDATION. Literature Cited 1. Ament, M.E. and C.E. Rubin. 1972. Relation of giardiasis to abnormal intestinal structures and function in gastointestinal immunodeficiency syndromes. Gastroenterology. 62:216266. 2. Clark, J.T. and D.V. Holberton. 1986. Plasma membrane isolated from Giardia lamblia: Identification of membrane proteins. J.Cell Biol. 42:200206. 3. Craun, G.F. 1979. Waterborne outbreaks of giardiasis. In: Waterborne Transmission of Giardiasis. Jakubowski and Hoff (eds). U.S. Environmental Protection Agency. pp.127147. 4. Dulley, J.R. and P.A. Greive. 1975. A simple technique for eliminating interference by detergent in the Lowry method of protein determination. Anal. Biochem. 64:136141. 5. Edson, C.M., Farthing, M.J.G., Thorley Lawson, D.A., and G.T. Keusch. 1986. An 88,000 Mr Giardia lamblia surface protein which is immunogenic in humans. Infect. Immun. 34:621625. 6. Einfield, D.A., and E.A. Stibbs. 1984. Identification and characterization of a major surface antigen of Giardia lamblia. Infect. Immun. 46:377383. 7. Hanff, P.A., Feihniger, T.E., Miller, J.N., and M.A. Lovett. 1982. Humoral immune response in humans syphilis to polypeptides of Treponema pallidum. J. Immunol. 129:12871291. 8. Keister, D.B. 1983. Axenic culture of Giardia lamblia in TYIS33 medium supplemented with bile. Trans. R. Soc. Trop. Med. Hyg. 77:487488. 9. Keystone, J.S., Krajden, S., and M.R. Warren. 1978. Person to person transmission of Giardia lamblia in daycare nurseries. Can. Med. Assoc. J. 119:241 258. 10. Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680685.
Page 180
11. Moore, G.T., Cross, W.M., McGuire, D., Mollohan, C.S., Gleason, N.N., Healy, G.R., and L.H. Newton. 1969. Epidemic giardiasis at a ski resort. N. Engl. J. Med. 281:402407. 12. Morrison, M. 1980. Lactoperoxidasecatalyzed oxidation as a tool for investigation of proteins. Meths. Enzymol. 70:214220. 13. Nash, T.E., Guillin, F.D. and P.D. Smith. 1983. ExcretorySecretory products of Giardia lamblia. J. Immunol. 131:20042010. 14. PekkataFlagan, A. and E. Ruoslathi. 1982. Unfolded transferring polypeptide chain is immunologically crossreactive with similar derivates of serum albumin and alphafetoprotein. J. Immunol. 128:11631167. 15. Radulescu, S., Iancu, L., Simionescu, D. and E.A. Meyer. 1976. Serum antibodies in giardiasis. J. Clin. Pathol. 29:863. 16. Ridley, J.J. and D.S. Ridley. 1976. Serum antibodies and jejunal histology in giardiasis associated with malabsorption. J. Clin. Pathol. 29:3034. 17. Rosoff, J.D. and H.H. Stibbs. 1986. Isolation and identification of Giardia lamblia specific stool antigen (GS65) useful in coprodiagnosis of giardiasis. J. Clin. Microbiol. 23:905910. 18. Schmerin, M.J., Jones, T.C., and H. Klein. 1978. Giardiasis: Association with homosexuality. Ann. Intern. Med. 88:801803. 19. Smith, J.W., and M.S. Wolfe. 1980. Giardiasis. Ann. Rev. Med. 31:373383. 20. Smith, P.D., Gillin, F.D., Brown, W.R. and T.E. Nash. 1981. IgG antibody to Giardia lamblia detected by enzymelinked immunosorbent assay. Gastroenterology. 80:14761480. 21. Taylor, G.D. and W.M. Wenman. 1987. Human immune response to Giardia lamblia infection. J. Infect. Dis. 155:137140. 22. Towbin, H.T., Staehelin, T., and J. Gordon. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Nat. Acad. Sci. U.S.A. 76:43504354. 23. Visvesvara, G.S., Smith, P.D., Healy, G.R. and W.R. Brown. 1980. An immunofluorescence test to detect serum antibodies to Giardia lamblia. Ann. Intern, Med. 93:802805. 24. Walsh, D.J. and K.S. Warren. 1979. Selective primary health care: An interim strategy for disease control in developing countries. N. Engl. J. Med. 301:967 974. 25. Wolfe, M.S. 1978. Current concepts in parasitology. Giardiasis. N. Engl. J. Med. 298:319321.
Page 181
Properties of Giardia Lamblia RNAs Cecilia Montanez*, Lourdes Cervantes, Cesar Ovando and Guadalupe OrtegaPierres Department of Genetics and Molecular Biology, Centro de Investigacion y de Estudios Avanzados del IPN, Apartado Postal 14740, Mexico City 07000, Mexico. RNA from the parasitic protozoan Giardia lamblia was obtained by three different methods and analysed by agarose gel electrophoresis under denaturing conditions. In all cases two prominent populations were observed. Our results showed that the large (LSrRNA) and small (SSrRNA) subunit rRNAs are approximately 2390 and 1420 nucleotides long. Analysis of sucrose gradient profiles of G. lamblia rRNAs revealed sedimentation values of 21S and 15S for the LSrRNA and SSrRNA species respectively. Small RNAs were characterized on polyacrylamide denaturing gels in which two bands of unusually small size, 130 and 118 nucleotides long, were found. Other smaller RNAs were also observed. Total RNA was translated in vitro using a rabbit reticulocyte lysate. In this case a broad spectrum of translation products was obtained. When these products were immunoprecipitated using immune sera from humans infected with G. lamblia, we found that most of the in vitro translated components were immunogenic. No differences in the immunoprecipitation patterns were observed when sera from asymptomatic or symptomatic patients were used.
Introduction Giardia lamblia is a parasitic protozoa which infects the intestinal tract of humans. This flagellated and binucleated protozoan is distributed worldwide and causes the diarrheal disease known as giardiasis (13). Despite the high prevalence of this infection, the genetics and molecular biology of G. lamblia remain poorly understood. Specifically, the structural organization and expression of the nucleic acids from this parasite have not been well characterized. Various studies concerning the isolation and characterization of ribonucleic acid (RNA), especially ribosomal RNA (rRNA) in protozoa have been reported (1,6,8,15,16). These studies have revealed considerable diversity between homologous nucleic acids species of these organisms. They have also contributed to a better understanding of the organization and expression of their nucleic acids as well as initiating the establishment of an evolutionary relationship between them. In this context, we started studies on the characterization of G. lamblia rRNAs as well as the analysis of messenger RNAs (mRNAs) which code for antigens that activate immune responses in the host. We have determined that the rRNAs of this parasite are approximately 2390 and 1420 nucleotides long, which are particularly small for eukaryotic cells. The isolated total RNA translated in vitro encode polypeptides containing antigenic determinants that are recognized by sera from patients with giardiasis. Materials and Methods Growth of Parasites G. lamblia Portland 1 (P1) and WB trophozoites (obtained from E. Weinbach and T. Nash, National Institutes of Health, respectively) were cultured axenically in vitro under anaerobic conditions. The trophozoites were grown at 37°C in Diamond's modified TYIS33 medium (18) supplemented with 10% heatinactivated calf serum and antibiotics (penicillin at 250 µ/mL and streptomycin at 250 µg/mL). Organisms were harvested in late log phase by centrifugation at 250 × g for 10 min, and the cells were washed twice with phosphate buffered saline pH 7.2 (PBS). The cell pellet from the final wash was suspended in a small volume of ice cold PBS and the concentration of trophozoites determined by counting in a hematocytometer. Cells used for RNA extraction were stored at 70°C until use. Extraction of Total RNA Total RNA from G. lamblia trophozoites was purified using the following three methods: 1) G. lamblia trophozoites were suspended in a solution of 4 M guanidine isothiocyanate in 50 mM TrisHCl pH 7.6, 10 mM EDTA (Ethylendiamine tetraacetic acid) 2% SDS (sodium dodecyl sulphate) and 0.14 M Bmercaptoethanol and centrifuged at 8000 × g for 10 minutes (7). The supernatant was layered over 1.2 mL of 5.9 M cesium chloride (CsCl) and centrifuged 17 hrs at 35,000 rpm at 16°C. The pellet was resuspended, extracted with chloroformbutanol (4:1 v/v) and treated with LiCl and ethanol to obtain pure RNA (28). 2) G. lamblia trophozoites were washed in cold PBS and lysed in 0.1 M TrisHCl, pH 9.0, 0.1 M LiCl, 1 mM EDTA, 1% SDS. Samples were extracted with phenolchloroformisoamyl alcohol (25:24:1 v/v) and RNA was precipitated with ethanol and 5M LiCl (3). 3) G. lamblia trophozoites were dissolved in 10 volumes of 8 M urea, 0.15 M sodium phosphate pH 6.8, 0.01 M EDTA, 1% SDS and extracted with phenolchloroformisoamyl alcohol (25:24:1 v/v) (11). The aqueous layer was treated with ether, and loaded on a hydroxylapatite column (23). RNA was collected by elution with 0.19 M sodium phosphate pH 6.8, dialyzed and precipitated with ethanol in the presence of 0.3 M sodium acetate pH 5.0 Agarose Gel Electrophoresis The G. lamblia total RNA obtained by the three methods described above was analyzed on denaturing slab gels: 2 to 4 µg of RNA were heated for 5 min in 25% deionized formamide, 20% glycerol, 0.25% bromophenol blue (BPB) and 0.04% xylene cyanol (XCFF) (w/v) at 65°C and * Corresponding author.
Page 182
Figure 1. Analysis of G. lamblia (P1 strain) RNA by denaturing agarose gel electrophoresis. Total G. lamblia RNA obtained by phenol extractions were run on a denaturing 1.5% agarose gel in the presence of formaldehyde and formamide and stained with ethidium bromide. Track 1: rat brain total RNA; track 2: G. lamblia RNA; track 3: E. coli total RNA.
then chilled on ice before loading. The samples were applied to a 1.5% agarose gel containing 6% formaldehyde in 2mM NaH2PO4, 18mM Na2PO4.7H2O and the gel was run at 50 V for 4.55 hrs. The RNA was visualized using a short wave ultraviolet light after the gels were stained with 0.5 µg/mL ethidium bromide for 30 min at room temperature. Polyacrylamide Gel Electrophoresis Vertical polyacrylamide gels containing 8.8% acrylamide, 2% bisacrylamide and 0.2M urea were prepared in 1X TBE buffer. The gels were prerun at 200 V for 30 min. RNA (58 µg) was dissolved in 80% formamide, 0.25% BPB, 0.04% XCFF and heated at 85°C for 5 min prior to loading. Electrophoresis was performed at 250 V for 1.5 hr, and the gels were stained with ethidium bromide as described above. Determination of RNA Sizes Approximate sizes of G. lamblia RNA were determined on denaturing gels by comparison with RNA markers of known size. The RNAs used as standards in these determinations were rat brain 28S (4802 bases) (14), 18S (1869 bases) (30), E. coli 23S (2904 bases) (4), and 16S (1541 bases) (5) and Saccharomyces cerevisiae 5.8S (158 bases) and 5S (121 bases) (12). Sucrose Gradient Centrifugation RNA samples (80 µg) suspended in 300 µL of 100 mM sodium chloride, 10 mM sodium acetate pH 5.2, and 1 mM EDTA were layered on 1035% linear sucrose gradients containing 10mM sodium acetate pH 5.2. The samples were centrifuged at 23,000 rpm for 18 hrs at 4°C. Following centrifugation the gradients were collected and their optical densities at 260 nm determined. Svedberg (S) values were estimated using rat brain RNAs as standard markers (25). In vitro Translations All in vitro translations were performed using rabbit reticulocyte lysates from Amersham Radiochemicals, Ltd according to instructions provided by the vendor in the presence of 1.2 mM Mg(OAc)2. Total RNA (10 µg), was added to 20 mL of rabbit reticulocyte lysate containing 30 µCi (35S) methionine (1190 Ci/mmol. Amersham Radiochemicals). The samples were incubated at 30°C for 90 min. Samples were used to determine the amount of (35S) methionine incorporated into trichloroacetic acid precipitable material. The remainder of the mixture was analyzed on 10% SDS polyacrylamide gels under reducing conditions (SDSPAGE) (19). Protein detection was achieved by Coomasie blue staining and autoradiography. Sera Human sera were obtained from asymptomatic and symptomatic Mexican patients with G. lamblia infections confirmed by stool examination. Control human sera were obtained from adults with no history of giardiasis. Titration of all sera samples was performed by enzyme linked immunosorbent assay (11). Immunoprecipitation of Translation Products Human sera samples collected from patients with giardiasis were used to precipitate in vitro translation products. For this, 50 µL of translation mixture in NETT (150 mM NaCl, 5mM EDTA, 50mM TrisHCL pH 8, 0.5% Triton X100) containing 200,000 cpm were incubated overnight at 4°C with suitable dilutions of sera obtained from humans as described above. After incubation immune complexes were absorbed on protein A sepharose and unbound polypeptides were removed by washing in NETT buffer supplemented with 1% bovine serum albumin (9). Radiolabelled antigens were eluted, and the samples were analyzed by SDSPAGE (19) on 1.5 mm slab gels. Protein detection was achieved by Coomasie blue staining and autoradiography. Results Extraction and Characterization of Total G. lamblia RNA Total RNA from G. lamblia trophozoites of P1 and WB strains was isolated by three different methods in order to detect all RNA populations present in these organisms. These include guanidine isothiocyanate extraction (7), phenolchloroform extraction (3), and hydroxylapatite chromatography (see materials and methods). Size Determination of rRNAs RNA samples obtained as described above were analyzed by electrophoresis under strongly denaturing conditions. Electrophoretic analysis of the RNA obtained from the P1 strain by the three different methods revealed the presence of two prominent rRNA bands which correspond to the large (LSrRNA) and small (SSrRNA) rRNA species of G. lamblia (Figure 1). The molecular weights of these rRNA species were approximated using rat brain and Escherichia coli rRNAs as standards. The sizes of the large and small rRNAs correspond to 2390 and 1420 nucleotides, respectively. Similar sizes were determined for LSrRNA and SSrRNA from WB strain (data not shown). Interestingly these RNA species from both strains of G. lamblia are smaller than most of the prokaryotic and eukaryotic rRNAs described (24). Sedimentation Velocity Measurements RNA purified by phenol extraction or by hydroxylapatite chromatography was centrifuged through sucrose gradients under nondenaturing conditions. The profiles of the O.D. measurements from the collected fractions revealed the presence of two peaks which correspond to the two rRNA species of G. lamblia (P1strain) (Figure 2). Sedimentation values of large and small rRNAs as calculated using rat brain rRNAs as standards were 21S and 15S, respectively. The sedimentation values of these two G. lamblia rRNAs are
Page 183
Figure 2. Sucrose gradient sedimentation analysis of RNA from G. lamblia (P1 strain). The G. lamblia RNA samples isolated by phenol extraction were layered onto sucrose gradients, centrifuged, and fractionated. The optical density of each sample was determined at 260 nm. The arrows indicate the sedimentation positions of the two main rRNA species of rat brain used as markers, which were run in parallel.
significantly smaller as compared to values reported for other eukaryotic rRNAs. The same sedimentation values were obtained for rRNAs from the WB strain (data not shown). Analysis of Small RNAs In order to characterize the small RNA species from G. lamblia, RNA preparations from P1 strain were fractionated by denaturing polyacrylamide gel electrophoresis. Two major RNA populations and several other less abundant species of lower molecular weights were observed (Figure 3). The sizes of the two prominent RNA molecules were determined relative to denatured rat brain and yeast 5.8S and 5S RNA standards, and correspond to approximately 130 and 118 bases. These sizes were also calculated for the WB strain small RNAs when analyzed under similar conditions. Thus, these small ribosomal RNAs, like their large 21S and 15S counterparts, are also significantly smaller than the 5.8S and 5S RNAs found in eukaryotic cells (12). Analysis of in vitro Translation Products Precipitated By Immune Sera Total RNA obtained from P1 and WB strains was translated in vitro in mRNA dependent rabbit reticulocyte lysates. Supplementation of reticulocyte lysates with 10 µg of total cellular RNA stimulated incorporation of 35Smethionine by approximately 3 to 10 fold, relative to a control without RNA. Analysis of in vitro translation products from total RNA revealed numerous proteins with molecular weights between 20 and 150 kD by SDSPAGE (Figure 4). In order to precipitate antigens recognized by sera from infected humans, total G. lamblia RNA was translated in vitro and the products were immunoprecipitated with human sera samples. Figure 5 shows that a range of in vitro translation products are recognized by sera from both asymptomatic (tracks 1 and 2) and symptomatic patients (tracks 3 and 4). All sera reacted with most of the in vitro produced polypeptides. These results demonstrate that the G. lamblia RNA obtained encodes a wide variety of antigens which are immunogenic in the human host. Discussion There are several reports concerning the isolation and characterization of rRNA in protozoa. In some protozoa the large rRNA is apparently an intact polynucleotide chain (16). In other protozoa, however, the large rRNA is labile, and dissociates under denaturing conditions into two fragments similar in size to the rRNA found associated with the smaller subunit (1,6). In this study we report the isolation of rRNA from two strains of G. lamblia: P1 and WB. In order to detect all rRNA populations we have used three different methods to isolate the total RNA. These include guanidine isothiocyanate, phenol extraction and hydroxylapatite chromatography. The analysis of the RNA isolated with these methods revealed only two rRNA populations. The sizes of these rRNAs, as determined by gel electrophoresis under denaturing conditions, were of approximately 2390 and 1420 nucleotides. These correspond to the LSrRNA and SSrRNA species respectively. As far as the small RNA populations are concerned, there are five species. Two of them, of approximately 130 and 118 nucleotides, are more abundant and probably correspond to the 5.8S and 5S RNAs of other eukaryotes. Our analysis revealed that the rRNAs of G. lamblia are therefore the smallest
Figure 3. Electrohoretic analysis of small G. lamblia RNAs. Total G. lamblia RNA was electrophoresed in an 8.8% acrylamide, 0.2M urea gel under denaturing conditions and stained with ethidium bromide as described in Materials and Methods. Track 1: total RNA from S. cerevisiae polysomes; track 2: total RNA from rat brain; track 3: total RNA from G. lamblia.
Page 184
Figure 4. Analysis of cell free products synthesized by total G. lamblia RNA. Proteins synthesized in a rabbit reticulocyte system from total G. lamblia RNAs obtained by phenol extraction were analyzed on a 10% polyaccrylamide gel. Track 1: G. lamblia total RNA; track 2: Tobacco Mosaic Virus RNA; track 3: No added RNA. Molecular weight markers are as indicated.
rRNAs reported for eukaryotic cells, and in this respect differ from other protozoa as well (12,24). These results are further supported by other studies from our group in which the rRNA genes were found to be located on a small repetitive DNA unit of 5.4 kb (manuscript submitted for publication). Recent studies by Boothroyd et al. (2), and Edlind and Chakraborty, (10) showed similar results regarding the rRNA sizes of these protozoa as well as the size of the repeated unit encoding for rRNAs. These results, together with the data obtained for other protozoa show a considerable diversity in eukaryotic rRNAs, their genes, and the processing mechanisms in these organisms. For instance, the nuclear ribosomal repeat unit of most eukaryotes includes three mature rRNA species: the 18S, 5.8S, and 28S rRNAs. These are processed from a single large primary transcript (for review see 22). However, in some organisms, other processing steps occur which result in unusual species of these three rRNAs (1,6,17,21,31). Together, all these data may reflect differences in the translation apparatus of these organisms. Further characterization of G. lamblia rRNAs and their genes will certainly contribute to a better understanding of the meaning of these particular differences. On the other hand, the ribosomal RNA genes are among some of the most conserved, universally distributed, and functionally equivalent genes of all organisms. These characteristics make these genes well suited for defining evolutionary
Figure 5. Immunoprecipitation of 35Smethionine in vitro labelled translation products by human sera. Sera from patients infected with G. lamblia were used to precipitate in vitro translation products as described in Materials and Methods. Tracks 1 and 2: in vitro translation products precipitated by sera from two asymptomatic patients. Track 3 and 4: in vitro translation products precipitated by sera from two symptomatic patients. Track 5: in vitro translation products precipitated by control human serum. Molecular weight markers are as indicated.
relationships among eukaryotic organisms (20,26,29) and will help to place G. lamblia in the correct evolutionary position. In the present study total RNAs from the G. lamblia P1 strain were used to conduct in vitro translation of polypeptides. In this case, a full spectrum of parasite components was obtained. Most of the in vitro translated products were precipitated with human immune sera, suggesting that total RNA preparations contain most of the RNA messangers which encode for immunogenic components of the parasite. No major differences were detected in the immunoprecipitation patterns obtained when sera from asymptomatic and symptomatic patients included in this study were tested. Finally, we have purified poly A+ RNA from total RNA which has been used to prepare a cDNA library. This will allow us to identify important proteins involved in the immune response of humans infected with the parasite as well as to study the induction and control of these and other components of this important pathogen. Acknowledgements We would like to thank Dr. A. Ratray, Dr. D. Peattie and Dr. I. Meza for critically reading this manuscript and Mrs. R. Barrera for secretarial assistance. C. Ovando is a
Page 185
recipient of a CONACyT studentship. This work was supported in part by CONACyT (Mexico), COSNET (SEPMexico) and MacArthur Foundation (U.S.A.). Literature Cited 1. Albach, R., Prachayasittiko, V. and G. Heebner. 1984. Mol. Biochem. Parsitol. 12:261272. 2. Boothroyd, J.C., Wang, A., Campbell, D.A. and C.C. Wang. 1987. Nucl. Acids. Res. 15:40654084. 3. Brawerman, G. 1974. In: Methods in Enzymology XXX. F.K. Moldave and L. Grossman (eds). Academic Press, New York. pp. 605612. 4. Brosius, J., Doll, T., and H.F. Noller. 1980. Proc. Natl. Acad. Sci. 77:201204. 5. Brosius, J., Palmer, M., Kennedy, P. and H.F. Noller. 1978. Proc. Natl. Acad. Sci. 75:48014805. 6. Castro, C., Hernandez, R. and M. Castaneda. 1981. Mol. Biochem. Parasitol. 2:219233. 7. Chirgwin, J., Przybyla, A., MacDonald, R. and W. Rutter. 1979. Biochem. 18:52945299. 8. Dame, J.B. and T.F. McCutchan. 1983. Mol. Biochem. Parasitol. 8:263297. 9. David, P.H., Hadley, T.J., Aikawa, M. and L.H. Miller. 1984. Mol. Biochem. Parasitol. 11:267282. 10. Edlind, T.D. and R.R. Chakraborty. 1987. Nucl. Acids. Res. 15:78897901. 11. Engvall, E. and P. Perlmann. 1972. J. Immunol. 109:129135. 12. Erdmann, U.A. 1982. Nucl. Acids. Res. 10:93115. 13. Faust, E.C., Russell, P.F., and R.C. Jung. 1970. In: Craig and Faust Clinical Parasitology. 8th ed. Lea and Febiger (eds). Philadelphia. pp 5974. 14. Hadjiolov, A.A., Georgiev, O.J., Nosikov, V.V., and L.P. Yavachev. 1984. Nucl. Acids. Res. 12:36773693. 15. Hernandez, R., Nava, G. and M. Castaneda. 1983. Mol. Biochem. Parasitol. 8:297304. 16. Hyde, J.E., Zolg, J.W. and J.S. Scaife. 1981. Mol. Biochem. Parasitol. 4:283290. 17. Jordan, B.R., LatilDamotte, M., and R. Jourdan. 1980. Nucl. Acids. Res. 8:35653573. 18. Keister, D.B. 1983. Trans. R. Soc. Trop. Med. Hyg. 77:487488. 19. Laemmli, U.K. 1970. Nature. London. 277:680687. 20. Lane, D.J., Pace, B., Olsen, G.S., Stahl, D.A., Sogin, M.L. and N.R. Pace. 1985. Proc. Natl. Acad. Sci. 82:69556959. 21. Leipoldt, M. and J. Schmidth. 1982. In: Genome Evolution. G.A. Dover and R.B. Florell (eds). Academic Press, New York, N.Y. pp. 219236. 22. Mandal, R. 1984. Progress in Nucleic Acid Research and Molecular Biology. 31:115159. 23. Markov, G.G. and I.G. Ivanov. 1974. Anal. Biochem. 59:555563. 24. Noller, H.F. 1984. Ann. Rev. Biochem. 53:119162. 25. Osterman, L.A. 1984. In: Methods of Protein and Nucleic Acid Research. Berlin SpringerVerlag, Berlin Heidelberg. pp. 241274. 26. Pace, N.R., Olsen, G.J. and C.R. Woese. 1986. Cell 45:325326. 27. Panasci, L.D., Green, D.C., Fox, P.A. and P.S. Schein. 1977. Anal. Biochem. 83:678688. 28. Rhoades, R. 1975. J. Biol. Chem. 250:80888097. 29. Sogin, M.L., Elwood, H.J. and J.H. Gunderson. 1986. Proc. Natl. Acad. Sci. 83:13831387. 30. Torezynski, R., Bollon, A.P. and M. Fuke. 1983. Nucl. Acids. Res. 11:48794890. 31. Ware, V.C., Renakawitz, R. and S. Gerbi. 1985. Nucl. Acids. Res. 13:35813597.
Page 187
Enzyme Activities of Giardia lamblia and Giardia muris Trophozoites and Cysts Donald G. Lindmark* and James J. Miller Cleveland State University, Cleveland, Ohio 44115, U.S.A. Giardia lamblia is the most common intestinal protozoan parasite in the world. It exists in two forms, the actively growing trophozoite and the infective resistant cyst. The information available on the parasite's carbohydrate and energy metabolism and hydrolytic abilities are limited to data on the trophozoite. We report here data obtained on the carbohydrate and energy metabolism and hydrolytic abilities of the cyst stage of G. lamblia. In addition, comparative data are presented on the carbohydrate and energy metabolism and hydrolytic abilities of the trophozoite and cyst stages of Giardia muris. The following enzyme activities were detected in homogenates of the trophozoites and cysts of both G. lamblia and G. muris: hexokinase, pyruvate kinase, phosphoenolpyruvate carboxykinase, pruvate:ferredoxin oxidoreductase, alcohol dehydrogenase, NADPH oxidoreductase, malate dehydrogenase, malate dehydrogenase (decarboxylating), acid phosphatase, DNase and RNase. These enzymes showed similar levels of activity and, the enzymes of carbohydrate and energy metabolism, similar characteristics, between species and stages (trophozoite and cyst). The hydrolytic enzymes were also similar in specific activities among the species and stages. No carbohydrate splitting hydrolases could be detected in any species.
Introduction Giardia lamblia is the most common intestinal protozoan parasite in the world (2). It exists in two forms, the actively growing trophozoite and the infective resistant cyst. The information available on the parasite's carbohydrate and energy metabolism and hydrolytic abilities is limited to data on the trophozoite (2). Using the gerbil (a proposed model for human giardiasis) as a source of G. lamblia cysts (1), we investigated the enzymes of carbohydrate and energy metabolism in this stage in order to better understand the potential that the cyst has for carbohydrate and energy metabolism. Since hydrolytic processes may be involved in the exit of trophozoites from the cyst (excystation), we investigated the hydrolytic abilities of both stages to gain insight into the excystation and possibily the encystation processes. Since G. muris cysts (and trophozoites obtained by in vitro excystation) can be obtained in large quantities using the mouse model (6), G. muris has been used in many studies as a substitute for G. lamblia. Hence we believed that a study of the metabolic potential of G. muris (trophozoite and cyst) was of importance and therefore have obtained comparative results on the carbohydrate and energy metabolism and hydrolytic enzymes from both the trophozoite and cyst stages of G. muris. In general, there are few differences in the specific activities of the enzymes of carbohydrate and energy metabolism and the hydrolytic enzymes between G. lamblia and G. muris and the trophozoite and cyst stage. The characteristics of the important enzymes of carbohydrate and energy metabolisms are very similar if not identifical in both parasites. Materials and Methods Organisms Trophozoites of G. lamblia (Portland I strain) were grown and harvested as described by Lindmark (2). Cysts of G. lamblia (1 × 106 per 20 animals) were obtained from the gerbil model as described by Belosevic et al. (1). Cysts of G. muris (1 × 107 per 30 animals) were obtained from the mouse model as described by RobertsThomson et al. (6). Giardia cysts were purified by sucrose gradient centrifugation and velocity sedimentation as described by Sauch (7). Trophozoites of G. muris were obtained by excystation of cysts as described by Rice and Schaefer (5) and purified by sucrose gradient centrifugation (1 × 105 per 30 animals after excystation and purification). The cyst and trophozoite preparations were stored as pellets after washing 2 times in 0.25 M sucrose at 70°C until enough were accumulated for enzyme analysis. The low quantities of cysts and G. muris trophozoites available for enzyme analysis and characterization made some experiments, such as Km determinations of substrates and cofactors, impossible. For example, a single harvest of cysts of G. muris produced enough homogenate to conduct 12 assays with controls for malate dehydrogenase (the enzyme with the highest specific activity in Giardia preparations). Enzyme Assays Homogenates to be used for enzyme analysis were prepared from trophozoites with a PotterElvehjam homogenizer as described by Lindmark (2). Homogenates of cysts were prepared in the same manner on preparations that were frozen and thawed 5X in the presence of 0.2% Triton X100 (this was done to rupture the resistant cyst wall). For assays and characterization of oxygen sensitive enzymes, homogenates were prepared as described by Lindmark under Argon in the presence of mercaptoethanol and stored under Argon. Published assays were used for malate dehydrogenase (EC 1.1.1.37), malate dehydrogenase (decarbonylating) (EC 1.1.1.39), fumarate hydratase (EC 4.2.1.2), lactate dehydrogenase (EC 1.1.1.27), catalase (EC 1.11.1.6) (3), hydrogenase (EC 1.18.3.1), pyruvate: ferredoxin oxidoreductase (EC 1.2.7.1) (2), succinate dehydrogenase (EC 1.3.99.1), acid phosphatase (EC 3.1.3.2), citrate synthase (EC 4.1.3.7), isocitrate dehydrogenase (EC 1.1.1.42), protein (4), hexokinase (EC 2.7.1.1), pyruvate kinase (EC 2.7.1.40), * Corresponding author.
Page 188
phosphoenolpyruvate carboxykinase (EC 4.1.1.32), alcohol dehydrogenase (EC 1.1.1.2), NADPH oxidoreductase (EC 1.6.99.1), acetate kinase (EC 2.7.2.1) (2), BNacetylglucosaminidase (EC 3.2.3.30), Bgalactosidase (EC 3.2.1.23), Bglucuronidase (EC 3.2.1.3), DNase (EC 3.1.4.5), and RNase (EC 2.7.7.16) (2). The pH dependence of the activities of the enzymes was determined in 100 mM trisphosphate buffer. Enzyme units were defined as the amount of enzyme necessary to form 1 µmol of product or to degrade 1 µmol of substrate per minute under the assay conditions stated. All enzymes were assayed at 30°C unless otherwise stated. Results Enzyme Activities The specific activities of enzymes assayed in homogenates are given in Table 1. The following enzymes were below the limit of detection: citrate synthase, isocitrate dehydrogenase, succinate dehydrogenase, fumarate hydratase, lactate dehydrogenases, acetate kinase, hydrogenase, catalase, Bglucuronidase, Bgalactosidase, and BNacetylglucosaminidase. Enzyme Properties As shown below, comparisons of the characteristics of the major enzymes of energy metabolism among the homogenates prepared from the different species and stages revealed many similarities. Unless otherwise stated the following holds true for the trophozoite and cyst forms of G. lamblia and G. muris. The results obtained with trophozoites of G. lamblia agree with those of Lindmark (2). Pyruvate:Ferredoxin OxidoReductase The activities require Coenzyme A (0.1mM) and thiol compounds (dithiolthreitol or mercaptoethanol) for full activity. The enzymes are oxygen sensitive with 60% loss in activity in the presence of air in 2 h. FMN (0.5 mM), FAD (0.05 mM), and ferredoxin (0.5 mg/mL) can be used as electron acceptors. NAD (5 mM) and NADP (5mM) are ineffective as electron acceptors (2). These experiments were not done with G. muris trophozoites because of the difficulty in obtaining sufficient quantities. Pyruvate Kinase The activities have a pH optimum of approximately 7.2. The enzymes require ADP and Mg2+ IDP and GDP cannot substitute for ADP. Ca2+, Mn2+, and Co2+ cannot substitute for Mg2+. Malate Dehydrogenase The activities have a pH optimum of approximately 7.0. NADH is the main electron donor. NADPH is 25% as effective. Malate Dehydrogenase (Decarboxylating) The activities have a pH optimum of approximately 7.3. NADP is required, NAD will not substitute. The enzymes are completely inhibited by 1 mM EDTA and show a requirement for a divalent cation (Mn2+, Co2+ Fe2+). Mg2+ and Ca2+ are ineffective. NADPH Oxidoreductase The activities have a pH optimum of approximately 7.3. Like pyruvate:ferredoxin oxidoreductase the activities are oxygen sensitive losing 50% activity in the presence of air in 2 h. No activity could be detected with NADH as an electron donor. Experiments were not performed with G. muris trophozoites. Alcohol Dehydrogenase NADPH is required. NADH is ineffective. The enzymes are irreversible, only utilizing acetaldehyde as a subtrate. Ethanol, isopropanol and proanol cannot be used as substrates. In summary within the limits of this study, the enzyme activities found in trophozoites of G. lamblia by Lindmark (2) are also found in the cysts of G. lamblia and the trophozoites and cysts of G. muris. Enzyme activities below the level of detection (2) in G. lamblia trophozoites, such as carbohydrate splitting hydrolases and enzymes of the Krebs cycle, are below the level of detection in cysts of G. lamblia and trophozoites and cysts of G. muris. The specific activities of the enzymes detected are similar. The characteristics of the enzymes of carbohydrate and energy metabolism are similar between species and stages. Discussion Earlier studies by Lindmark (2) have demonstrated the occurrence of many enzyme activities in trophozoites of G. lamblia. Our results presented here confirm the above TABLE 1. Specific activities of enzymes in Giardia. Activity mU/mg protein ± S.D. (No. of determinations) G. lamblia Enzyme
G. muris Cyst
Trophozoite
Cyst
25 ± 4 (4)
20 ± 6 (5)
15 ± 3 (4)
10 ± 4 (4)
140 ± 15 (3)
101 ± 20 (3)
190 ± 39 (3)
110 ± 5 (2)
20 ± 2 (5)
18 ± 14 (6)
22 ± 4 (3)
10 ± 2 (3)
Malate dehydrogenase
850 ± 32 (6)
701 ± 15 (5)
600 ± 40 (5)
510 ± 18 (5)
Malate dehydrogenase (decarbonylating)
120 ± 15 (8)
85 ± 10 (3)
100 ± 30 (3)
60 ± 10 (2)
Pyruvate:ferredoxin oxidoreductase
350 ± 60 (10)
210 ± 54 (11)
200 ± 41 (2)
240 ± 33 (3)
Alcohol dehydrogenase
310 ± 10 (5)
200 ± 35 (4)
390 ± 26 (2)
210 ± 40 (4)
NADPH oxidoreductase
390 ± 30 (6)
300 ± 20 (3)
300 ± 38 (2)
200 ± 51 (3)
Acid phosphatase
80 ± 4 (10)
60 ± 10 (12)
85 ± 10 (3)
42 ± 10 (4)
DNase
58 ± 15 (8)
40 ± 18 (3)
84 ± 17 (2)
30 ± 18 (2)
RNase
42 ± 11 (4)
34 ± 5 (2)
60 ± 20 (2)
28 ± 3 (2)
Hexokinase Pyruvate kinase Phosphoenolpyruvate carboxykinase
Trophozoite
All assays were conducted at 30°C except alcohol dehydrogenase (19°C).
Page 189
mentioned research and demonstrate that tremendous similarities exist among the enzymes of carbohydrate and energy metabolism of trophozoites of G. lamblia and G. muris. The specific activities of the enzymes are similar and the characteristics of pyruvate kinase, malate dehydrogenase, malate dehydrogenase (decarboxylating), pyruvate:ferredoxin oxidoreductase, alcohol dehydrogenase and NADPH oxidoreductase are identical. Other enzyme activities possibly involved in carbohydrate and energy metabolism such as the enzymes of the Krebs cycle, catalase, hydrogenase, lactate dehydrogenase and acetate kinase were below the level of detection in homogenates. These similarities suggest that G. lamblia and G. muris have similar pathways of carbohydrate and energy metabolism. In addition, the cysts of both species have the same enzymetic potential of the trophozoite, suggesting that the enzymes of carbohydrate and energy metabolism are present in the cyst, to carry out metabolism of the trophozoite upon excystation. The cysts and trophozoites of each species have the same complement of hydrolytic enzymes. Both forms lack the enzymes needed to split complex carboydrates (B galactosidase, Bglucuroinadse, BNacetylglucosaminidase). This finding should be given consideration in future studies on the excystation process and the chemical components of the cyst wall. Our results demonstrate a large degree of similarity between G. lamblia and G. muris in the enzymes of carbohydrate and energy metabolism and hydrolytic enzyme activities, and also show that the trophozoites and cysts of each species exhibit the same metabolic potential. Acknowledgements The research was supported by an Academic Challenge Grant in Parasitology from the Ohio Board of Regents, Columbus, Ohio; The Thrasher Research Fund and WHO (PDP) (P2/181/20). Literature Cited 1. Belosevic, M., Faubert, G.M., MacLean, J.D., Law, C., and N.A. Croll. 1983. Giardia lamblia infections in Mongolian gerbils: an animal model. J. Inf. Diseases 147(2):222226. 2. Lindmark, D.G. 1980. Energy metabolism of Giardia lamblia trophozoites. Mol. Biochem. Parasitol. 1:112. 3. Muller, M. 1973. Biochemical cytology of trichomonad flagellates. I. Subcellular localization of hydrolases, dehydrogenases, and catalase in Tritrichomonas foetus. J. Cell Biol. 57:453474. 4. Muller, M., Hogg, J.F., and C. de Duve. 1968. Distribution of tricarboxylic acid cycle and glyozylate cycle enzymes in Tetrahymena pyriformis. J. Biol. Chem. 243:53855395. 5. Rice, E., and F.W. Schaefer, III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Micro. 14(6):709710. 6. RobertsThomson, I.C., Stevens, D.P., Mahmoud, A.A.F. and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroentoerology. 71:5761. 7. Sauch, J.F. 1984. Purification of Giardia muris cysts by velocity sedimentation. Appl. Environ. Micro. 48:454455.
Page 191
Studies on Giardia Lamblia Trophozoite Antigens Using Sephacryl S300 Column Chromatography, Polyacrylamide Gel Electrophoresis and Enzymelinked Immunosorbent Assay P.P. Chaudhuri, S. Pal, S.C. Pal, and P. Das* Department of Parasitology, National Institute of Cholera and Enteric Diseases, P33, C.I.T. Road, Scheme XM, Beliaghata, Calcutta 700 010 India Antigen prepared from Giardia lamblia trophozoites cultured in vitro in Diamond's TYIS33 medium was analyzed using sodium dodecyl sulfate polyacrylamide gel electrophoresis, Sephacryl S300 column chromatography, counter immunoelectrophoresis and enzymelinked immunosorbent assay. After elution through S300 column four distinct fractions were obtained. Molecular weights of these fractions were 150,000; 65,000; 50,000 and 10,000 daltons for FI, FII, FIII, and FIV respectively. The SDSPAGE analysis revealed a minimum of 28 distinct bands with crude antigen and 13, 22, 26 and 30 bands with FI, FII, FIII and FIV fractions respectively. The molecular weight of these bands ranged from 125,000 to 14,000 daltons. Antigenic activity was observed in all four fractions in the CIEP test. However, when assayed by the ELISA test the maximum antigenic activity was linked to the higher molecular weight fraction.
Introduction Giardia lamblia, a flagellated protozoan parasite that thrives in the upper intestine of humans and causes a spectrum of diseases, including asymptomatic carriage, acute fulminating diarrhea and chronic diarrhea with malabsorption. The disease most commonly occurs in infants and children (18), particularly those attending day care centers (8), travelers (3), homosexuals (13), hypogammaglobulinanaemics and back packers (3). In addition, the organism has been established as the etiologic agent of numerous outbreaks of diarrheal disease in various parts of the world (4,7). Despite the considerable morbidity caused by G. lamblia very little is known about the antigenic configuration of this protozoan. Earlier work on this parasite by various workers (11,12,15,16) suggests that the organism is antigenically a complex moiety. To define the antigenic nature of G. lamblia trophozoite (strain Portland1) further, Sephacryl S300 column chromatography for fractionation, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDSPAGE) for the comparison of protein constituents in CSA and its fractions, and counter immunoelectrophoresis (CIEP) and enzymelinked immunosorbent assay (ELISA) for immunological activity were used. The crude soluble antigen (CSA) and its fractions were polydisperse in their molecular weight when analyzed both in Sephacryl S300 column chromatography and SDSPAGE with a broad spectrum of immunogenicity. Materials and Methods Parasite Culture The P1 strain of G. lamblia was subcultured twice per week at 37°C in filter sterilized TYIS33 medium (6), supplemented with vitamins and 10% heat inactivated adult bovine serum, penicillin (50 µg/mL) and streptomycin (50 µg/mL) as antibiotics. Antigen Preparation Actively growing G. lamblia trophozoites showing exponential growth (7296 h) were dislodged from the walls of culture tubes by immersion in an ice bath for 10 minutes followed by centrifugation at 800 × g for 5 minutes. Pooled viable trophozoites were washed 5 times in sterile phosphate buffered saline (PBS pH 7.4, 0.05 M) and finally resuspended in normal saline. This suspension was then sonicated in an ice bath with eight 30 sec bursts (MSE Sonicator, U.K.). The sonicated material was centrifuged at 10,000 × g for 20 minutes at 4°C. The supernatant was collected as crude soluble antigen (CSA) and used for antigenic analysis after the protein contents were estimated by the method of Lowry et al. (10). Preparation of Antisera Albino rabbits weighing 23 kg were immunized with G. lamblia CSA. About 2 mg of antigenic protein in 0.5 mL were emulsified with an equal volume of Freund's complete adjuvant (Difco) and injected subcutaneously into the hind legs of each rabbit. A total of three such injections were given to each rabbit at weekly intervals. This was followed by three intravenous injections with CSA alone (approximately 34 mg of antigenic protein per rabbit) at 2 day intervals. Animals were bled a week after the last injection and the precipitating antibody in the immune rabbit serum was detected by a CIEP test against the homologous antigen. Fractionation of CSA by Sephacryl S300 (gelfiltration) Column Chromatography CSA of G. lamblia was subjected to gel filtration through Sephacryl S300 columns in order to separate its antigenic fractions. About 80 mL of preswollen Sephacryl S300 (Pharmacia Fine Chemicals, Sweden) wet bead diameter 40105 µm was poured in a glass column (1.6 × 40 cm, Pharmacia Fine Chemicals, Sweden). A flow rate of 20 mL per hour was maintained with a peristaltic pump (Pharmacia) throughout the experiment. The void volume of the column was determined by applying Dextran Blue2000. About 1.5 mL (i.e. 12% of bed volume of column) of CSA which was previously dialyzed with eluent buffer containing about 24 mg of protein, was applied and 3.5 mL fractions were collected in each tube with the help of an automatic fraction collector (Frac100, Pharmacia Fine Chemicals, Sweden). An elution profile was obtained by measuring the optical density (O.D.) * Corresponding author.
Page 192
at 280 nm and these values were plotted against the elution numbers (tubes). According to the O.D. values each peak and trailing eluates (Figure 1) were pooled separately and listed by fractions. These fractions were concentrated by lyophilization and dialyzed against physiological saline. Finally 1 mg protein/mL of solution was prepared and stored at 4°C until used for further studies. The molecular weight of each fraction was determined by comparing the Kav value of each fraction with the known mol. wt. Marker of gel filtration proteins supplied by Pharmacia Fine Chemicals, Sweden. SDSPAGE Analysis The protein constituents of CSA as well as the four fractions separated by S300 column chromatography were compared by SDSPAGE using 10% separating gels, in 0.5 trisHCl buffer pH 6.8. The electrode buffer and sample preparation was made after following standard procedures (9). SDSPAGE analysis was performed in a vertical slab gel electrophoresis chamber (LKB, Sweden) with a constant temperature of 15 °C. A constant 120 V was applied once the Giardia antigen had entered the separating gel. Known molecular weight protein standards (Pharmacia Fine Chemicals, Sweden) were run simultaneously. The gels were fixed, washed and stained for protein with 0.125% Coomassie Brilliant Blue dye. CIEP Test The antigenic activity of CSA and its four fractions were compared in counter immunoelectrophoresis tests against the different dilutions of rabbit antiGiardia antibodies. The CIEP test procedure was essentially the same as that described by Sharma et al. (14). Briefly, approximately 3 mL of 1% agarose (Sigma Chemicals, U.S.A.) in barbital buffer (pH 8.8, 0.05 M) was layered on micro slides and wells for antigen and antibody were punched according to the standard size. The antibody wells were placed to the anodal side and the antigen wells to the cathodal end. The experiment was run in a electrophoresis chamber (Shandon, U.S.A.) at a constant 160 V for 30 minutes. Readings were taken just after the experiment and also after 24 h of incubation at 4°C. ELISA Test The micro ELISA test performed was the same as described earlier in the serodiagnosis of amoebiasis (5), with slight modification. In this experiment a known amount of antibody (i.e. 1:200 dilution) was used to react with variable amounts of antigen. Four preimmune rabbit sera and four immunized against Giardia CSA rabbit sera were pooled in separate batches and used as reference positive and negative sera for the experiment. Before the experiment was conducted a chequer board titration was carried out to determine the optimal concentrations of antigen and antibody required. A 20 µg per mL G. lamblia protein from CSA and a 1:200 dilution of pooled positive and negative sera were found optimal and specific for obtaining a clear cut distinction between positive and negative results. For comparison of antigenic activity among CSA and its fractions, different concentrations of antigenic protein viz. 20 µg, 10 µg, 5 µg, 2.5 µg and 1.2 µg per mL were used in this test. A 1:1000 dilution of antirabbit IgG labelled with horse radish peroxidase (Sigma Chemicals, U.S.A.) was used as a conjugate. Results were read photometrically at 490 nm (O.D.) in an automatic ELISA reader (Dynatech Labs, U.S.A.). Results The Sephacryl S300 gel filtration pattern of Giardia CSA is presented in Figure 1. Two major light absorbing peaks were observed, one of which appeared in the void volume and the other at the end of total column volume. The last peak mostly consisted of yellowish colouring material (originally seen in the CSA). Eluted materials were appropriately pooled as indicated In Figure 1 to give four different fractions FI to FIV; molecular weights of FI to FIV were 150K, 65K, 50K, and 10K daltons, respectively.
Figure 1. Chromatographic pattern of axenic Giardia lamblia crude soluble antigen (CSA) on Sephacryl S300 gel. Column dimensions were 1.6 × 40 cm, sample size 24 mg/1.5 mL 0.05 M trisHCl buffer pH 7.6, flow rate was 20 mL/h.
In the CIEP test the FI fraction showed a positive precipitin reaction up to 1:64 titre of immunized rabbit serum; the other fractions as well as CSA itself failed to show reactivity beyond 1:8 titre of immunized rabbit serum, although the same concentration (1 mg/mL) of
Figure 2. Bar diagrammatic representations of the distribution of precipitin activities of crude soluble antigen (CSA) and its different fractions (FI FIV). Antigenic activity was determined by a counter immunoelectrophoresis testing of different dilutions of pooled positive immunized rabbit sera to Giardia CSA.
Page 193 TABLE 1. Comparision of optical density values for different concentrations of crude soluble antigen (CSA) and its fractions (FI FIV). Antigenic protein concentrations (µg/mL)
Mean O.D. valuesa FI
FII
FIII
FIV
20
0.243
CSA
0.556
0.410
0.256
0.156
10
0.203
0.500
0.318
0.196
0.079
5
0.138
0.426
0.235
0.110
0.051
2.5
0.120
0.250
0.135
0.128
0.023
1.2
0.039
0.165
0.092
0.059
a
O.D. values were taken at 490 nm.
antigenic protein was used in each case. In the ELISA test a clear cut distinction in O.D. value was observed when 20 µg of crude soluble antigen was used against a 1:200 dilution of positive and negative serum. The four preimmune negative sera showed an O.D. value ranging from 0.05 to 0.09 with a mean O.D. of 0.07 at 490 nm. The four immunized rabbit sera with the same dilution showed O.D. values ranging between 0.12 and 0.325 with a mean O.D. of 0.234. The cutoff value indicating a negative or a positive result was taken as 0.100 (mean O.D. of controls). The results of CSA and its fractions when compared using the ELISA test for antigenic activity against the pooled immunized rabbit sera are shown in Table 1. After analysis, fraction I was found to be most antigenic as compared to the parent (CSA) and the other three fractions. in other words, the 2.5 µg/mL protein of this fraction (FI) showed almost
Figure 3. SDSPAGE profile of Giardia lamblia crude soluble antigen (CSA), and its different fractions (FI FIV). Low molecular weight marker protein was also used in the experiment.
the same O.D. value as those obtained with 20 µg/mL of CSA, 5 µg/mL of FII, and 20 µg/mL of FIII, respectively. Fraction IV was found to be the least sensitive in comparison to CSA and other fractions. The pattern of SDSPAGE with CSA and its fractions for comparison of their constituent protein subunits are shown in Figure 3. The CSA showed several (about 28) discrete protein bands in the molecular weight region of ~12.5 × 104 to ~1.4 × 104 daltons. Fraction I showed a similar banding pattern with less number of subunits (about 13 bands). Fraction II showed protein bands in the molecular weight range of ~9.4 to ~4.3 × 104 daltons. The bands of FIII were mainly confined to the molecular weight region of < 9.4 × 104 to 2.1 × 104 daltons. However, the FIV contained low molecular weight protein subunits ranging from ~8.2 × 104 to ~1.4 × 104 daltons. Discussion The present investigation demonstrates the complex nature of CSA of G. lamblia (strain P1) trophozoites. They contain about 28 polypeptides according to SDS PAGE analysis and these findings correspond to the observations of Smith et al. (15). However, the results differ slightly from those of Moore et al. (12) who demonstrated about 20 distinct protein determinants with the same experiment. A similar study (17) using an immunoelectroprecipition test showed 24 precipitin arcs. These differences in the protein polymers may be due to different strains and methods used by different workers. The results of fractionation in HPLC (12) and Sephacryl S300 column chromatography of CSA showed almost similar observations. In both cases the maximum antigenic activity was recorded in the high mol. wt. fraction (FI). Fraction I, which showed maximum antigenicity, contained only 13.5% of the total protein present in the whole Giardia extract is serologically active. Fractions II, III and IV, although they had protein contents of 22%, 10% and 2.5% respectively of the total extract, showed significantly lower serological values. Furthermore, the precipitin reactions which have been demonstrated in whole CSA and its various fractions can be explained on the basis that fractions II, III and IV are not pure and are contaminated with the preceding fraction. Fractions III and IV which were found to be relatively less active than the other two fractions by ELISA test could perhaps throw some light on the difference between the ELISA and precipitin test for the detection of clinical Giardia cases as described in amoebiasis (5) where, the haemaglutination and IFA test showed negative observations. However, ELISA was found positive with the same sera. It is possible that the two tests detect different subclasses of IgG antibodies. The demonstration that the immunologic activity was associated with a particular fraction(s) makes it possible to isolate these particular fractions by Sephacryl S300 gel chromatography and to use these fractions for the production of a more specific antiG. lamblia antiserum. The production of more specific antigens and antibodies will
Page 194
allow for more sensitive assays to study and elucidate the role these antigens play during G. lamblia infection. Literature Cited 1. Ament, M.E., H.D. Ochs., and S.D. Davis. 1973. Structure and function of the gastrointestinal tract in primary immunodeficiency syndromes: a study of 39 patients. Med. 52:227248. 2. Barbour, A.G., C.R. Nichols, and T. Fakushima. 1976. An outbreak of giardiasis in a group of campers. Am. J. Trop. Med. Hy G. 25:384389. 3. Brodsky, R.E., H.C. Spencer, Jr., and M.G. Schultz. 1974. Giardiasis in American travellers to the Soviet Union. J. Infect. Dis. 130:319323. 4. Craun, G.F. 1979. Waterborne outbreaks of giardiasis, pp. 127149. In: Jakubowski, W. and J.C. Hoff (ed.), Waterborne Transmission of Giardiasis. U.S. Environ. Protect. Agency, Cincinnati, Ohio, EPA 600/979001. 5. Das, P., S. Pal, and S.C. Pal. 1984. Evaluation of the micro enzymelinked immunosorbent assay, indirect haemagglutination and indirect fluorescence antibody techniques for serodiagnosis of amoebiasis. J. Diar. Dis. Res. 2:238242. 6. Diamond, L.S., D.R. Harlow, and C.C. Cunnick. 1978. A new medium for the axenic cultivation of Entamoeba histolytica and other Entamoeba. Trans. R. Soc. Trop. Med. Hyg. 72:431432. 7. Juranek, D. 1979. Waterborne giardiasis, pp. 150153. In: Jakubowski, W. and J.C. Hoff (ed.), Waterborne Transmission of Giardiasis. U.S. Environ. Protect. Agency, Cincinnati, Ohio. 8. Keystone, J.S., S. Krajden, and M.R. Warren. 1978. Person to person transmission of Giardia lamblia in day care nurseries. Can. Med. Assoc. J. 119:241248. 9. Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680685. 10. Lowry, O.H., N.J. Rosebrough, A.L. Farr, and R.J. Randall. 1951. Protein measurement with folin phenol reagent. J. Biol. Chem. 193:265275. 11. Meyer, E.A., and S. Radulescu. 1979. Giardia and giardiasis. Adv. Parasitol. 17:147. 12. Moore, G.W., F.S. Bernal, and M.V. Dennis. 1982. Characterization of Giardia lamblia trophozoite antigens using polyacrylamide gel electrophoresis, high performance liquid chromatography and enzyme labelled immunosorbent assay. Vet. Parasitol. 10:229237. 13. Schmerin, M.J., T.C. Jones, and H. Klein. 1978. Giardiasis: association with homosexuality. Ann. Intern. Med. 88:801804. 14. Sharma, P., P. Das, and G.P. Dutta. 1981. Rapid diagnosis of amoebic liver abscess using Entamoeba histolytica antigen. Arch. Invest. Med. (Mex.) 12:215 218. 15. Smith, P.D., F.D. Gillin, N.A. Kaushal, and T.E. Nash. 1982. Antigenic analysis of Giardia lamblia from Afghanistan, Puerto Rico, Ecuador and Oregon. Infect. Immun. 36:714719. 16. Visvesvara, G.S., E.A. Meyer, and G.R. Healy. 1976. Antigenic analysis of Giardia lamblia. American Society of Parasitologists, Abstract No. 68, 51st Annual Meeting, San Antonio, Texas, p. 40. 17. Visvesvara, G.S. 1981. Giardia lamblia: America's number 1 intestinal parasite. Diag. Med. 4:2429. 18. Visvesvara, G.S. 1982. Giardiasis in children. J. Pediatr. Gastroenterol. Nutr. 1:463465.
Page 195
DETECTION OF GIARDIA CYSTS
Page 197
An Overview of the Techniques Used for Detection of Giardia Cysts in Surface Water Charles P. Hibler Department of Pathology, Colorado State University, Fort Collins, Colorado 80523, U.S.A.. Analysis of water samples for detection of Giardia cysts in surface water necessitates trapping particulate material on a filter cartridge, and selective separation of particulates from the cysts. The techniques used by analysts to isolate cysts are essentially similar and the results do not differ significantly. Some analysts use the ''reference" technique, others use an overlay/underlay modification of this technique, and some use immunofluorescence to visualize the cysts. All of the techniques are equally subject to and may be adversely affected by water quality. Efficiency is inversely proportional to the turbidity of the source water. High turbidity, especially organic turbidity, algae, freeliving protozoans, and other material (alum and/or polymers) in the water source severely limit effective recovery and/or visualization of the cysts.
Introduction At the Symposium on Waterborne Giardiasis in 1978, Jakubowski and Ericksen (2) reviewed the techniques previously employed by investigators attempting to recover cysts of Giardia from surface water and presented the EPA method developed for recovery. In 1980 the EPA brought a group of experienced investigators to the EPA headquaters in Cincinnati, Ohio to discuss the problems with diagnosis and to suggest a more efficient procedure. The result of this meeting was the "consensus" or "reference" method (3), the method recommended in Volume 16 of Standard Methods (1). The sampling device developed by the EPA (2) has become standard equipment for sampling water, but the filter cartridges, the porosity of these cartridges, and the methods for isolating cysts of Giardia from the particulates trapped by the cartridges have undergone considerable modification since the 1978 and 1980 meetings, primarily through trial and error efforts by a number of investigators. The basic principle requires trapping Giardia cysts on a filter, recovering these cysts from the filter, and selectively concentrating and identifying the cysts. The type of filter cartridge preferred (glass fiber/epoxy, orlon, cotton or polypropylene), the selective concentration media employed (potassium citrate, sucrose, zinc sulfate or percoll) and the specific procedures developed for analysis (direct light microscopy or immunofluorescence) usually vary among laboratories because experienced investigators develop personal preferences. However, investigators attempting to improve the current stateoftheart for isolating Giardia cysts from water generally are in frequent communication, exchanging ideas and techniques; consequently, the basic techniques employed by different laboratories and by individuals trained at one of these laboratories have many similarities. The filter cartridges preferred, the selective concentration media employed, and the specific procedures developed for analysis are all equally subject to and compromised by variations in water quality. The type and amount of material suspended in water varies with the source (river, lake, spring, etc.), season, or geographic location and are subject to specific circumstances (spring runoff, flood, construction, or thunderstorm). Any amount of suspended material interferes with the recovery of Giardia cysts and the higher the turbidity the greater the loss of cysts, especially if the turbidity is primarily organic. If a municipal filtration system uses alum or polymers, and small amounts of these chemicals pass the system, coagulation of the suspended material trapped on the filter cartridge may effectively prevent recovery of any cysts. Healthy water sources support a number of aquatic organisms (crustaceans, insects, algae, diatoms, protozoa, etc.) and large numbers of these organisms, especially some protozoa and algae, interfere with visualization of the cyst, either because of similarity in size and shape or, sometimes, due to sheer numbers, resulting in fatigue at the microscope. Giardia cysts are not evenly distributed in water and their dispersion in streams or rivers is dependent upon volume and rate of flow (4). Despite their relatively light weight (s.g. 1.0451.050) cysts settle rather quickly in slowly moving or standing water. The morphologic quality of Giardia cysts varies from sample to sample, within a sample, between seasons, and during special circumstances (spring runoff, thunder storms), generally due to a mixing of fresh cysts with older cysts. In some samples all of the cysts may be excellent and morphologic criteria necessary for identification (two to four nuclei, axonemes and median bodies) are readily visualized at 450X while in other samples these features can be visualized only with considerable difficulty, often necessitating a magnification of 1000X. Some Giardia cysts are quite resistant to the trauma associated with sampling, extraction and concentration and will continue to be morphologically excellent for several weeks, even without preservation; however in other samples, even from the same site but at a different time, only a small percentage of the cysts found initially in the
Page 198
sample can be recovered a few days later. Sometimes the cysts are dead and shrivelled to the extent they are almost unrecognizable, in other samples cysts are present and appear to be in good condition but cannot be selectively concentrated (irrespective of the concentration chemical used, or the specific gravity of that chemical) suggesting that the integrity of the cyst membrane has been compromised and the specific gravity of the cyst altered. Live, infectious cysts possess a cytoplasm that is essentially clear (hyaline) when viewed by phase contrast and/or brightfield microscopy whereas those that are dead and/or dying have a coagulated appearance to the cytoplasm and intracellular organelles are easily detected. Visual assessment correlates extremely well with the fluorogenic dyes, fluorescein diacetate and propidium iodide, used to determine cyst viability (7,8). Interpretation of morphologic quality of Giardia cysts for viability is subjective at best. Cysts that have been dead for months, or those inactivated by chlorine, ozone, etc. often appear morphologically excellent whereas those that are considered to be morphologically poor in quality (and therefore dead) could have been excellent (but not necessarily alive) at the time of sampling. Time lag between sampling and analysis, mishandling of the sample, activity of chlorine and the properties of the cysts in that particular sample can all affect cyst morphology at the time of analysis. All of the factors interfering with recovery and/or visualization must be considered when analyzing a sample of water for Giardia cysts. Recovery of a specific number of cysts indicates only that this is a percentage of the total number present, a percentage that is expected to be inversely proportional to the amount and type of suspended material present and the morphologic quality of the cysts. The basic techniques to be described here for recovery of Giardia cysts from surface and/or ground water, the problems associated with these techniques, and the modifications applied in an effort to overcome the problems have evolved from 12 years of experience with over 6500 water samples from geographic areas throughout North America, and 6 years of experience with recovery of Giardia cysts from pilot filtration systems seeded with cysts to evaluate efficiency of the systems. Discussion Sampling Device The descriptions and comments given below on the sampling techniques are derived primarily from experience; very little information on the precision, sensitivity and efficiency of these techniques appears in the published literature. The device developed by the EPA for recovery of Giardia cysts (2,1) has become standard equipment for sampling water. The filter cartridges generally used are 1 µm (nominal) porosity depthtype cartridges. Two or 3 µm (nominal) porosity depthtype cartridges would probably be suitable, but none have been tested. Five µm (nominal) porosity cartridges will trap about 98% of the cysts, but too few have been evaluated for any to be recommended. Seven and 10 µm (nominal) porosity filters will pass many of the cysts. Cartridges should be handled with rubber gloves to prevent contamination and care should be taken so their exposure is only to the water and packaging materials. Laboratories performing Giardia analysis often prefer different cartridges, but there is little difference between orlon, polypropylene, cotton or acetate for recovery or ease of processing. The surface type filters, such as the Balston epoxy fiberglass tube or the membrane filters have definite disadvantages because they are rapidly plugged by material suspended in the water, severely limiting the amount of turbid water that can be sampled. The filter housing should be highimpact clear plastic to permit visualization of the selected filter cartridge. The influent hose generally is a double female coupling high pressure hose (dishwasher or washing machine type), four to eight feet in length, whereas the effluent hose often is a garden hose measuring 25 feet or longer, a length sufficient to reach a floor drain in the facility or to be directed a considerable distance away from the influent hose when sampling from a stream or a lake. A water meter generally is attached directly to the effluent side of the filter housing and the effluent hose attached to the meter. Rate of flow can be satisfactorily regulated with the faucet to which the device is attached or, if necessary, a limiting orifice flow control. Sampling directly from a stream or lake can be done with an electric or gasolinepowered portable pump, or with a light weight 12 volt marine bilge pump (about 3 to 3.5 gallons/minute) powered directly from batteries or an automotive vehicle. If an electric or gasolinepowered pump is used the filter housing should be placed between the influent housing and the pump to prevent possible destruction of the cysts by the pump. The influent hose must be suspended in the water, usually with a flotation collar constructed from styrofoam to prevent the suction of sediment into the device. If use of the 12 volt marine bilge pump is necessary, the pump is suspended in the water by a styrofoam flotation collar and an extremely light weight hose attached between the pump and the filter housing. Often an anchor is necessary in rapidly flowing water. Monzingo and Hibler (4) have described the applications of this equipment for sampling in remote areas. The sampling device must be thoroughly cleansed, preferably with soap and hot water containing chlorine at an approximate concentration of 100 mLs of household bleach (5.25% sodium hypochlorite by weight) to 20 L of water and this must be done before the equipment has been allowed to dry. Sauch (personal communication) has observed that airdried cysts can be reconstituted with water and often retain most of the features necessary for identification. If the sampling equipment is assembled as indicated above, the housing and influent equipment only need to be washed because the 1 µm cartridge is essentially 100% effective in removing particulates greater than 1 µm in size or larger when properly seated. If the type of cleansing preferred is not possible, the influent hose and housing should be thoroughly rinsed with the water from the next source to be sampled. Prudent planning can prevent problems when sampling equipment is limited.
Page 199
For example, if only one device is available and a municipal filtration system is being sampled, the plant effluent must be sampled before the influent. If at all possible, samples should be taken before chlorine has been applied to the system. If chlorinated water must be sampled, it should be taken as near the chlorinator as possible and 0.5% sodium thiosulfate introduced by a fluid proportioner pump in accordance with Volume 16 of Standard Methods (1). An alternative method, although not as satisfactory, is to add about 100 mL of 1% sodium thiosulfate to the cartridge (in the bag with the cartridge) after sampling. Unfortunately, with the latter method any cysts trapped on the cartridge will have been exposed to chlorinated water for several hours and the Giardia cysts may be dead and unrecognizable before analysis occurs. Volume of Water Samples and Rate of Flow The amount of suspended material, the type of material (organic, inorganic), and the particle size of that suspended material will determine how much water could or should be sampled, and there are no clear guidelines. Jakubowski (3) suggested a minimum of 380 L (100 U.S. gallons). Highly turbid water, especially water with considerable organic material, severely limits the analyst's chances of recovering Giardia cysts and, as indicated earlier, recovery is inversely proportional to the amount of this material. Since Giardia cysts are not evenly distributed in water and do not move downstream continuously, the amount of water sampled is a trade off. However, samples of 3800 L (1000 gallons) are easily obtained if the turbidity is less than 1 NTU and the turbidity is primarily inorganic in composition. Generally a sample of 380 to 1520 L (100 to 400 gallons) is the best suggestion that can be offered at present; unless the water temperature is less than 5°C and the turbidity is less than 1 NTU, then 3800 L would increase the chances of recovering cysts. Some samplers closely monitor the rate of flow and when the rate sharply decreases they terminate sampling: this approach is acceptable. Since Giardia cysts are not evenly distributed in water the rate of flow, and even the time of day, are probably as critical as the volume sampled. Jakubowski (3) recommended a rate of 1 gallon/minute and this is an acceptable approach. Unfortunately a 1 gpm rate is not always possible due to other sampling commitments, weather conditions, etc. Samples taken at 1 gpm versus those taken at 3 to 4 gpm do not appear to differ significantly in the number of cysts recovered. The time of day is important. Animal populations, acting as reservoirs, frequently are more active at night and therefore sampling is preferably done overnight when animals are present. If the source of contamination is sewage, peak periods of sewage effluent should influence the sampling effort. Packaging, Labelling and Shipment Filter cartridges should be removed from the filter holder, placed into a plastic bag and secured to prevent leakage and, as an added insurance, doublebagged (zip loc type bags are excellent). Each inner cartridge bag should be clearly labelled with water proof ink. The information should include the name, address and telephone number of the individual responsible for the sample, date and time sampling was started, date and time sampling was completed, the number of gallons or litres sampled, and the source of the water and the type of sample (raw or treated water and type of treatment). Turbidity, water temperature and pH should be included. If the sample was from a chlorinated source, the amount of chlorine used (mg/L or ppm), the contact time (in minutes) before sampling occurred and the procedure used to dechlorinate the sample (inline dechlorination or dechlorinated postsampling) should be indicated. All of this information is important for the records of the municipality; moreover, this information facilitates analysis and interpretation by the laboratory. Filter cartridges must be refrigerated with wet ice and shipped in a strong, secure, leakproof container. Samples must not be exposed to freezing conditions. Samples should arrive at the laboratory no later than 48 hours after sampling is completed. The type of carrier employed for shipment is dependent upon the location of the laboratory and the sampling site. Bus lines are reliable and efficient for sites near the laboratory, and all of the air carriers are equally efficient and reliable when the distance is considerable. Shipment should be planned for samples to arrive at the laboratory during a week day, primarily because receipt of packages on weekends, irrespective of the carrier, often is difficult. As indicated earlier, cysts from that source may not survive and be recognizable unless the sample is processed immediately. Processing the Filter Cartridge for Analysis Extraction of suspended material from the filter cartridge is done by hand, using distilled water. Extraction of membranes or the epoxy fiberglass tube is performed by a backwash procedure. Some laboratories using the yarnwound (orlon) filter cartridges unwind the yarn, separate the yarn into sections, and hand wash. This laborious, timeconsuming procedure simply cannot be performed by busy laboratories processing 120 filter cartridges daily; moreover, the procedure requires an inordinate and unnecessary amount of handling with no greater cyst recovery. A fast and efficient procedure is to slice the cartridge to the core with a razor knife. Washing should begin with the inner fibers and all of the fibers should be washed in six to twelve sections depending on the amount of debris. Cartridge fiber sections should be repeatedly washed in fresh distilled water until the fibers appear clean. The number of washings necessary is dependent upon the amount and type of suspended material present. The wash water is then combined for sedimentation. The hand washing procedure is necessary and all laboratories (except those sampling with the epoxy fiberglass tube or the membranes) use this procedure. Some laboratories use a wetting agent (0.1% Tween 20 or Tween 80 in distilled water) to facilitate removal of Giardia cysts from the fibers (especially orlon) although there is no proof cysts are "sticky". Irrespective of the filter cartridge used, or the specific protocol followed by different laboratories, the first, and potentially the greatest, loss of cysts occurs during this washing procedure. Analysis of this step by Jakubowski and Ericksen (2) indicates that recovery varies from 58139%. In our
Page 200
laboratory we have discovered that recovery from orlon, polypropylene or cotton filter cartridges varies from 85109%. No differences could be detected among filter cartridges or different materials. However, in the analyses for both of the above evaluations the only suspended material in the filter cartridges was Giardia cysts: turbidity was not a factor. After the washing procedure has been performed, samples containing a high organic content, especially in the summer months, should be preserved with sufficient formalin to make a 2% (v/v) solution. This prevents further growth of algae and protozoa. After preservation, the samples can be refrigerated overnight at 46°C or concentration of the washings can be initiated immediately with a centrifuge. All cartridges generally are washed in clear glass, wide mouth gallon jars (3.78 litres) to facilitate visualization of the next procedure. If the sample has been refrigerated and allowed to settle, the supernatant should be siphoned carefully to the sediment. The amount of supernatant to be removed is dependent upon the amount of sediment. A good rule is to leave an amount of supernatant equivalent to the amount of sediment unless the sample is filtration plant effluent containing only a trace of sediment. Siphoning then should be terminated sooner, leaving about 200300 mL of supernatant, otherwise sediment could be inadvertently siphoned. Selective Concentration of the Giardia Cysts. The Consensus or Reference Method As a result of the meeting in Cincinnati, Ohio, and at the publication of the latest edition of Standard Methods (1), the recommended technique was the zinc sulfate centrifugal flotation technique, a procedure that was highly modified and improved upon by several laboratories during the lag time before publication of the technique. The technique works very well if the water is not highly turbid; however, a primary limitation is the amount of centrifugate from sample aliquots. If approximately 0.25 mL of sediment (centrifugate) in a 25 mL aliquot is present cyst recovery is about 85%, providing the cysts are of excellent morphologic quality. Cyst recovery decreases proportionately with an increase in sediment (centrifugate). If the volume of centrifugate is approximately 0.75 mL, cyst recovery is about 20% and only an occasional cyst can be found when the centrifugate is approximately 1 mL. Moreover, vortexing, or other mixing of the sample to mix the centrifugate with the Lugol's iodine and zinc sulfate results in an inordinate amount of inorganic and organic material adhering to the meniscus of the fluid in the tube and the majority of this material cannot be effectively removed from the meniscus; therefore the cysts floating to the top are then combined with the material in the meniscus of the zinc. This interferes with visualization of the cyst and increases the fatigue factor. If cysts are of poor morphologic quality and the cyst membrane is compromised, losses are even greater. Regardless of the chemical used for selective concentration, losses are comparable; all are equally compromised by the amount of sediment and/or the quality of the cysts. The Overlay/Underlay and the Membrane Filter Selective Concentration Techniques A modification of the consensus method that considerably improves recovery of cysts and other living material in highly turbid water is the overlay or underlay technique. For the overlay technique 2025 mL of selective concentration media (zinc sulfate, surose, percoll or potassium citrate) is poured (or injected) into a clear 50 mL conical centrifuge tube (glass or nalgene) and an equal amount of the concentrate from the filter cartridge washings introduced (layered) onto the surface of the chemical without disturbing the interface between chemical and concentrate. Another approach, the underlay technique, is to pour 2025 mL of the filter concentrate into the centrifuge tube and then introduce 2025 mL of the selective concentration chemical beneath the filter washing concentrate without disturbing the interface. No doubt a number of approaches can be used to accomplish this step, but this laboratory uses a 30 mL disposable syringe to which is attached a 16 gauge spinal needle. The Stylex syringe (Parmaseal Americal Pharmaseal, Valencia, CA 91355) is preferred for a smoother insertion of the chemical. Although we have used both overlay and underlay, and laboratory personnel trained by us have developed preferences for one or the other, we prefer the underlay technique, primarily because filter washing concentrate can sometimes plug the barrel of the syringe, resulting in a mixing with the chemical concentrate when the filter washings are discharged; moreover, washing and cleaning a syringe and needle that has been filled with chemical is much easier and much more effective than washing equipment that has been filled with filter washings. After the overlay/underlay step, centrifugation must begin immediately. Any delay will allow the material suspended in the filter washing to settle at the interface between the chemical/filter washing, resulting in a considerable number of Giardia cysts being pulled through the interface. The tubes should be centrifuged for 5 to 8 minutes in a centrifuge that will provide about 380 g. As indicated previously, if the filter washings contain a minimum of sediment, the reference method is better than the overlay/underlay technique. The purpose of the overlay/underlay technique is to trap material suspended in the filter washing having a specific gravity of less than the chemical at the interface between the filter washings and the chemical. This includes Giardia cysts, other animal parasites, algae, diatoms, protozoa, plant debris, crustaceans, freeliving nematodes, etc. After centrifugation, the filter washing supernatant to the interface, and the surface of the interface is siphoned through a 5 µm Nuclepore membrane. We use a 47 mm membrane secured in a Millipore chamber. Slight vacuum is applied with a faucet siphon attached to the Millipore cup. Several approaches can be used to recover material at the surface. The first is to use a syringe and needle, beginning at the meniscus and removing material to the interface and about 5 mm below the interface (into the chemical) or another approach is to insert the needle at the interface and withdraw the liquid. With either technique the liquid level decreases from the meniscus to the interface. The inner perimeter of the tube should be circled as the fluid is withdrawn. In this laboratory a third
Page 201
approach is used. We break the interface with an applicator stick and then quickly pour the washings and chemical solution through the membrane. After siphoning through the membrane, the membrane is removed with forceps and held against the inner wall of a beaker to be washed with a fine, strong stream of distilled water from a squirt bottle. The suspended material in the beaker is poured into a 50 mL centrifuge tube, the beaker rinsed and the rinse added to the tube. This tube (first washing) is then centrifuged at 380g for 3 to 5 minutes and the liquid siphoned to the 6 10 mL level. This first washing material is mixed with a vortexer and transferred, with rinsing, to a 15 mL clear nalgene tube (preferably with a lip around the top) for a second washing and centrifugation at 380 g for 3 to 5 minutes. Following this centrifugation, the water is siphoned to above the pellet and two drops of Lugol's iodine added. There are several formulas in use for Lugol's iodine. We use 10 g potassium iodide and 5 g iodine to 100 mL water. Two to three mL of chemical solution (zinc sulfate, percoll, potassium citrate, sucrose) is added immediately and the tube vortexed to suspend the material. Chemical solution is then added to fill the tube until the meniscus bulges very slightly. A clean glass coverslip is added and the tube centrifuged for 3 to 5 minutes at 380g. The purpose in using a tube with a lip is to insure a better seal between tube and coverslip, preventing loss of the coverslip during centrifugation at this last, critical step. After centrifugation the coverslip is removed, placed on a glass slide and the material adhering to the coverslip examined systematically with a microscope at 100%. The first loss of cysts occurs in the processing of the filter cartridge, and the second loss occurs in the overlay/underlay technique, irrespective of the chemical employed. The third loss occurs in the final step. Not all of the cysts will adhere to the coverslip, and when lifted off the centrifuge tube some remain behind on the meniscus. The amount and type of suspended material and the quality of the cysts will cause cyst losses. When cysts are suspended in water, or in a small amount of inorganic or organic material, recovery approaches 9095%; however, as with the consensus method, large amounts of suspended material result in considerable loss of cysts. A very effective procedure to mitigate these losses is to dilute the concentrate to at least a 1:2 ratio and process more replicate samples through the membrane. This does not necessitate microscopic examination of an inordinate number of replicates; they can be combined and trapped on one membrane. Each step in this procedure must be performed carefully and thoroughly. The top lip of the 15 mL centrifuge tube must be perfectly flat to form an effective seal. Glass coverslips often are greasy; if they are greasy the glass must be thoroughly degreased, rinsed with distilled water and dried. Tom Trok (personal communication) prepares an effective cleanser by adding Ajax or a similar cleanser to distilled water, centrifuging the suspension and using the supernatant as a cleanser. Squirting alcohol on one or two coverslips from each new box will quickly determine if grease is present. The consensus method (Standard Methods, 1985) offers two alternatives, adding the coverslip to the tube before centrifugation, or adding the coverslip afterwards (touching the meniscus with the coverslip). The latter procedure is not as effective. Some Giardia cysts begin to settle into the chemical immediately after centrifugation. Two major considerations for these techniques are the quality of the centrifuge and the microscope. The centrifuge should have an accurate timer and rpm meter; moreover, it should coast slowly to a stop, not suddenly. We find that a centrifuge which stops too quickly disturbs the Giardia cysts on the coverslip or at the meniscus. The microscope must possess excellent quality lenses, excellent objectives, and a strong, bright, wellbalanced light source. While the quality of the microscope does not interfere with recovery of cysts, it can certainly limit effective visualization and quickly initiate microscope fatigue. Selective Concentration Media Several chemicals have been used to selectively concentrate water samples possibly containing cysts of Giardia. Most experienced parasitologists prefer ZnSO4, but investigators in different laboratories use sucrose, percoll, ficollhypaque, or potassium citrate. All have good and bad points. The chemical preferred by diagnosticians in the different laboratories is generally the chemical with which that individual has the most experience, and when used by that individual will probably provide results comparable to a chemical preferred by an individual in another laboratory. Zinc sulfate, potassium citrate and sucrose are all extremely hygroscopic and will shrink the cysts; however, this does not interfere with diagnosis. Percoll and ficoll hypaque do not shrink the cysts, but these chemicals are costprohibitive for routine analysis. This laboratory's experience with these various chemicals evaluated at a s.g. of 1.13 indicates that percoll and/or ficollhypaque are the best chemicals for recovery, and are about 10% better than ZnSO4, which is about 10% better than sucrose; potassium citrate was the worst of the chemicals examined. However, zinc sulfate, potassium citrate, and sucrose when used at specific gravities of 1.2 to 1.3 were better than percoll of ficollhypaque. Specific Gravity of the Chemical Investigators must be prepared to use chemicals with different specific gravities to selectively remove Giardia cysts and other living material from the filter washings. A hydrometer must be used to determine the specific gravity, because preparation of an exact solution cannot be done effectively by weight/volume. We use ZnSO4 at specific gravities of 1.1, 1.2 and 1.3. Some laboratories use 40% potassium citrate (about 1.3 sp.g.). We prefer to underlay with a chemical of 1.3 s.g. because comparison of results with 1.1, 1.2 or 1.3 ZnSO4 indicates the higher density will result in more cysts recovered (the shrinkage is also more pronounced but still not an interference). Unfortunately high organic content, alum or polymers in the filter concentrate irrespective of the amount of dilution, often nessitates use of chemicals at 1.2 s.g. or even 1.1 s.g.; otherwise material from the filter concentrate packs at the
Page 202
interface of concentrate/chemical. As the specific gravity of the chemical is decreased to selectively remove the low density material present in the concentrate, cyst losses (trapped in the material and pulled through the interface) increase proportionately. If a considerable amount of alum or polymers is present in the concentrate, selective removal of cysts is almost impossible. "No confidence in the results" is reported when organic turbidity or these chemical coagulants interfere with cyst detection. Analysts may use one or two approaches in an attempt to selectively remove cysts from the material containing excessive sediment: 1) dilute the concentrate and overlay/underlay more tubes; and/or 2) use a chemical with lower specific gravity. Using a chemical with a lower specific gravity should be the last choice because of the loss of efficiency. Microscopic Examination If the glass coverslip method is employed prior to centrifugation in the last step, it should be oriented with one corner of the coverslip in the direction of the centrifugal force and, when removed, oriented on the glass slide with a mark on the slide indicating this quadrant. Examination should begin in this quadrant because experience has shown that due to the centrifugal action and the specific gravity of the cysts versus the chemical, the majority of cysts will be present in this quadrant. Before fatique becomes a problem this quadrant should be thoroughly examined, then the remainder of the coverslip examined, field by field. During the examination the miroscopist should note the relative amounts of material (inorganic and organic debris) and organisms present to assess the risk factor, especially when comparing filter plant influent (raw water) and filter plant effluent (finished water). When examining influent and effluent to assess the efficiency of a plant to remove Giardia cysts, the presence or absence of cysts in the raw water is irrelevant; the ability of the system to remove particulates the size of or larger than Giardia cysts is extremely important, and the relative percentages removed are an assessment of risk. Many plants remove most of the small particulates (clay, etc.) but allow passage of plant debris, a nice name for the undigested fecal debris from herbivorous rodents (muskrat and beaver). Plant debris is very pliable and light in weight and is the best indicator of filter plant efficiency. If beaver and/or muskrat are present on the source then generally Giardia cysts will be found in the raw water, if not in that sample certainly in a subsequent sample. If the animals are not shedding cysts or cysts are not found, the plant debris will be present and the filter plant's ability to remove this material is important. Modifications to the Analysis Techniques Most investigators now use the overlay/underlay technique with one of the above chemicals to separate the cysts and other living material from the inorganic particulates. Some, as in our laboratory, attach the coverslip before centrifugation, some "touch" the coverslip to the meniscus of the tube after centrifugation, and others use the bacteriologic loop to secure a sample of the meniscus. With these techniques a relatively "clean" sample (about 8001000 gallons of water at 0.5 to 1.0 NTU) requires about 2 hours of effort for analysis, while a "dirty'' sample may require about 4 hours. Other investigators have chosen to wash the material from the membrane and then concentrate by centrifugation, ultimately obtaining about 0.5 to 1 mL of material to be examined drop by drop, either after staining with Lugol's iodine or with the aid of a phase contrast microscope. This procedure requires 8 24 hours of microscope time, depending on the amount of material trapped at the interface between concentrate and chemical. Eye fatigue is a very real problem and effectively compromises the results. Busy laboratories cannot afford the time for this technique and municipalities cannot afford the cost. Immunofluorescence Riggs (5) developed a direct fluorescent antibody (FA) technique and Sauch (6) developed an indirect fluorescent antibody (IFA) technique for the detection of Giardia cysts in water. Recently, Riggs developed a monoclonal antibody from human source cysts for this purpose. Sauch selectively cleans the concentrate with percoll which is the best chemical for this purpose, Riggs uses 40% potassium citrate, and this laboratory uses zinc sulfate because it is costeffective for routine monitoring. After selective concentration with the chemical, the material trapped at the interface between the two chemicals is evaluated microscopically and the volume of concentrate adjusted to provide a monolayer of particulates when 1 mL is applied to a cellulose triacetate 25 mm diameter filter (Gelman Metrical). Sauch uses a 0.2 µm absolute porosity, while we use a 5 µm absolute porosity. Diluted antiserum is applied to each membrane for 15 minutes, rinsed 5 times with phosphate buffered saline (PBS) and diluted conjugage (goat antirabbit IgG conjugated to FITC, Miles Scientific, Naperville, IL) is applied for 15 minutes and rinsed five times with PBS. Sauch uses PBS supplemented with 2% bovine serum albumin and 0.05% polyoxyethylene sorbitan monolaurate (Tween 20) while we use only the PBS. Sauch stains the material with Evans blue (0.003% Evans blue in 0.15 M KCL, Sigma Chemical Company) for 10 minutes, dehydrates with an aqueous ethanol series (v/v) containing 5% (v/v) glycerol (10, 20, 40 and 80% ethanol), then cleans the membrane with glycerol containing 0.12% (w/v) propylgallate (Sigma Chemical Company), covers the membrane with a 25 mm coverslip and seals with clear fingernail polish. Generally we forego the staining and dehydration steps and only clear the membrane, cover with a coverslip and read directly. Sauch prefers to clear the membrane because she double checks her findings by switching from fluorescence to phasecontrast microscopy to verify the results. The procedures developed by Riggs and Sauch if followed closely, will work very well providing the analyst has a good antibody properly titrated for maximum cyst fluorescence and minimal background nonspecific fluorescence. Problems with Direct Microscopy and Fluorescent Microscopy As previously mentioned, the greatest losses of cysts for any of these techniques occurs in the first two steps: washing the fibers (or backflushing the epoxyfiberglass cartridges) and in the selective concentration step. Losses are going to occur and at the current state
Page 203
oftheart there is little that can be done except attempt to minimize these losses. A problem with zinc sulfate, sucrose, or potassium citrate is the dehydration factor. The cysts are smaller (shrunken), necessitating considerable experience for recognition. Moreover, if the cyst wall is compromised the organism may fill with chemical and fail to float properly in the last flotation step because it then has the same specific gravity as the chemical. Losses may occur because of greasy coverslips, etc. One particularly difficult problem is that many different organisms will take the iodine stain, all appearing brown in color. While live, fresh Giardia cysts are highly refractile and brown in color, older or dead cysts may be light brown or even green in color. This, together with the dehydration, often necessitates careful examination at 1000x to locate the internal structures necessary to confirm that the organism is a Giardia cyst. Healthy, wellbalanced sources of water may be heavily laden with small flagellates and/or algae similar in size and staining properties of Giardia cysts; this will contribute quickly to fatigue and a failure to find or see the few cysts present. Individuals planning to use the direct microscopy procedure, irrespective of whether they use dropbydrop direct examination or a flotation procedure should be welltrained in protozoology, stream biology and fresh water algae. One of the main problems with immunofluorescence is that it is only available from one commercial source. This antibody has not been extensively evaluated. Another problem is the cost: good quality microscopes with both phasecontrast and epifluorescent capabilities are expensive. Some investigators have suggested immunofluorescence may be a better option for the inexperienced microscopist than the direct microscopy technique; however, others experienced in both techniques feel that fluorescence of other organisms may result in false positive diagnoses. There is no advantage to immunofluorescence for experienced, welltrained microscopists. Immunofluorescence is more timeconsuming than direct microscopy and fatigue becomes a factor more quickly; therefore busy laboratories examining 1 20 samples/day cannot afford the time. Both techniques are much quicker than the dropbydrop technique and are more effective because of the even greater fatigue factor associated with the latter. A distinct advantage of the direct microscope technique is the ability to evaluate filtration plant performance. The majority of our samples come from municipalities with filtration systems and we are asked to assess efficiency based on removal of risk material, organisms in the water the size of or larger than Giardia. As indicated earlier the presence or absence of Giardia is of interest, but the performance of the plant is more important. A problem for microscopists using either technique is the detection and identification of Giardia cysts from sources that probably are not infectious for humans. We find that 100% of the blackcrowned night herons are infected with Giardia and shed a cyst that is very similar to the Giardia duodenalis type found in mammals. No doubt some of the other waterfowl and/or shore birds are infected with Giardia. We have found the bird type of cyst in water from reservoirs in both the eastern and western United States. Conversation with municipal authorities has revealed that waterfowl were present on the reservoir at the time of sampling. This cyst fluoresces extremely well with the polyclonal antibody we use, but may not fluoresce with a monoclonal antibody. Currently the risk of misidentification is very real, irrespective of the technique, and could result in an unnecessary "boil water" order for a municipality. Quality Assurance If microscopists do not regularly examine sources of water contaminated with cysts, they should "seed" water with cysts to check on their ability to visualize the organisms. Seeding samples with cysts often is frustrating. Some batches (from the same source) will be excellent and cysts can be effectively recovered for days or weeks; in other batches cysts can be recovered for only a few days because the cyst numbers have decreased or the cyst quality deteriorates. We examine 1 20 filters/day and 20 40% are contaminated with cysts; however, we continue to use quality assurance samples as a control on our technique. We regularly supply quality assurance samples, actual concentrates from water samples contaminated with cysts, to persons trained in our laboratory and this procedure works reasonable well; however, as indicated earlier, the number of cysts present in a sample on the day examined may or may not be the same as what can be recovered the next day. Generally, fewer cysts will be present. This necessitates sending samples containing a higher concentration of cysts to insure that the individual will find at least a few. An animal model is used as a source for cysts. This is timeconsuming and costly; moreover, the U.S. Department of Agriculture has established extremely strict guidelines for use of laboratory animals. Most small laboratories could not comply with the standard without costly remodeling and training of personnel. Conclusions Diagnosis of waterborne Giardia is a difficult and timeconsuming task of selectively concentrating and finding a few cysts in the concentrate of 100 to 1000 gallons of water. No specific technique currently available is any better or worse than the other techniques available: they are all relatively inefficient. We find cysts in approximately 20% of the samples examined; likely closer to 30 or 40% of these samples have cysts present but due to the reasons discussed they are not recovered or visualized if recovered. Repeated sampling of a negative source usually provides positive results. We have a means by which we can determine if these cysts are alive but not if they are dead (the gerbils may not be susceptible to the cyst strain in the sample). If the cyst is of the Giardia duodenalis type we must assume potential infectivity for humans and initiate the barriers necessary to prevent infection. We must assume that all surface water sources of water are either contaminated with or will be contaminated with the cysts of Giardia.
Page 204
Literature Cited 1. A.P.H.A., A.W.W.A., W.P.C.F. 1985. Standard methods for the examination of water and waste water. 16th edition, A.P.H.A. Washington, D.C. pp. 1268. 2. Jakubowski, W. and T.H. Ericksen. 1979. Methods for detection of Giardia cysts in water supplies. In: Waterborne Transmission of Giardiasis. W. Jakubowski and J.C. Hoff (eds). USEPA, Office of Research and Development, Environmental Research Center. Nat. Tech. Info. Service, Springfield, VA. EPA. 3. Jakubowski, W. 1985. Detection of Giardia cysts in drinking water: state of the art. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York, NY, pp. 263285. 4. Monzingo, D.L. and C.P. Hibler. 1987. The prevalence of Giardia in a beaver colony and the resulting environmental contamination. J. Wildl. Dis. Vol. 23. pp. 576585. 5. Riggs, J.L., Nakamura, K. and J. Crook. 1984. Identifying Giardia lamblia by immunofluorescence. In: Proc. of the 1984 Specialty Conf., Environmental Engineering. M. Pirbazari and J.S. Devinny (eds). 6. Sauch, J.F. 1985. Use of immunofluorescence and phasecontrast microscopy for detection and identification of Giardia cysts in water samples. Appl. and Environ. Microbiol. 50(6):14341438. 7. Schupp, D.C. and S.L. Erlandsen. 1987. A new method to determine Giardia cyst viability: correlation of fluorescein diacetate and propidium iodide staining with animal infectivity. Appl. Environ. Microbiol. 53:704707. 8. Schupp, D.C., Januschka, M.M. and S.L. Erlandsen. 1987. Assessment of Giardia cyst viability with fluorogenic dyes: comparisons to animal infectivity and cyst morphology by light and electron microscopy. Advances in Giardia Research. University of Calgary Press.
Page 205
Methods for the Recovery of Giardia and Cryptosporidium from Environmental Waters and their Comparative Occurrence Joan B. Rose*, Dima Kayed, Mary S. Madore, Charles P. Gerba, Michael J. Arrowood, Charles R. Sterling and John L. Riggs The University of Arizona, College of Agriculture, Department of Nutrition and Food Science, 309 Shantz Building, Tucson, Arizona 85721, U.S.A.. Cryptosporidium has recently been associated with a waterborne disease outbreak in the United States. Information on the occurrence of this parasite in domestic sewage and water, however, is almost nonexistant. Recently, our group developed methods for the detection of Cryptosporidium in large volumes of sewage and water. This study applied these methods to determine the relative occurrence of both Giardia and Cryptosporidium in surface waters and sewage effluents in the western United States. Both parasites were concentrated from water using polypropylene spun fiber filters, and identified with the aid of monoclonal antibodies. Giardia cyst levels in raw sewage averaged 51/L. Giardia was only occasionally observed in treated sewage effluents (1.3 cysts/L) and contaminated surface waters (0.35 1.2 cysts/L), while Cryptosporidium was observed in almost all samples examined. Average concentrations of oocysts in raw sewage were 5,200/L and 1,400/L in treated sewage effluents. In surface water, oocyst concentrations averaged from 0.08 to 28.5/L from polluted to pristine waters respectively. In summary, water sources free of Giardia cannot be assumed to be free of Cryptosporidium.
Introduction The ability to recover pathogenic microorganisms from environmental samples can aid the microbiologist, epidemiologist and water engineer in defining waterborne outbreaks of disease (11). Because bacterial indicator systems used to judge water quality do not accurately predict the presence or absence of some specific pathogens, such as Giardia, it may become particularly important to have methods which can detect these organisms (12). Routine sampling procedures used for bacteria are not applicable to Giardia. Specialized techniques are required for the sampling, recovery, and detection of Giardia cysts from water. Early efforts employed the use of membrane filtration for collection of cysts from sewage with limited success (13). This procedure, however, was not applicable to other types of water (11). A pool sand filter was also used to collect larger volumes of water and cysts were recovered by filter backflushing, and alum coagulation (19). For routine use, however, this method proved too complicated and laborious. The use of the yarnwound cartridge filter for cyst concentration has now proven to be a usable system (10). Large volumes of water can be processed, yet, overall recovery efficiencies were still poor (6.3%). Although much attention has been given to the development and modification of methods for the isolation of Giardia cysts from environmental samples (11), improvements in cyst detection by microscopic methods have lagged. Under most circumstances, the accurate identification of the Giardia cysts requires an experienced individual. Recognizing cysts in environmental samples, however, may prove difficult to even the experienced parasitologist due to presence of artifacts, algae cells, or other debris resembling Giardia cysts. Cyst quantification may also be difficult. Spaulding et al. (20) found that membrane filtration was an accurate means for determining cyst concentrations. Sauch (18) used immunofluorescence in conjunction with membrane filtration to detect Giardia from environmental samples. This combination of techniques has proven reliable for cyst detection and quantitation. Giardia isolation methods have been adopted for use in the recovery of another enteric protozoan, Cryptosporidium (14, 15). Polypropylenewound filters were used to collect oocysts and a sucrose gradient centrifugation method was used to purify them. Direct immunofluorescence was used to identify and quantitate oocysts. Although initial oocyst recoveries were low (14.5%), improvements have increased recoveries to a 59% average (16). A similar system has been used to detect Giardia in water samples using a potassium citrate medium during the purification step (17). It is now possible to examine waters for the presence of enteric protozoa such as Cryptosporidium and Giardia. Improvements and thorough evaluations of the methods are needed, however, before these techniques can have widespread applicability. Materials and Methods Sample Collection Water samples were collected using a portable gasolinedriven water pump and filtered through teninch spun polypropylene cartridge filters (Micro Wynd II, * Corresponding author.
Page 206
AMF/CUNO Division, Meriden, CT) with a nominal porosity of 1 µm. Flow rates were adjusted to four to five gallons per minute and 100 to 400 gallons of water were filtered. Filter Elution Filters were initially processed by backflushing the filter with 2700 mL of eluent [deionized (DI) water containing 0.1% Tween 80]. The filter was cut longitudinally, separated from the core, teased apart, and washed three times; each time in onethird (900 mL) of the eluent. The washing was done on an automatic shaker for 10 minutes in a onegallon container. The sample was concentrated and combined into a single pellet by centrifugation (1200 × g for 10 minutes). The final pellet was divided in half and resuspended in 10% formalin or 2.5% potassium dichromate. Pellet Processing Pellets suspended in formalin were washed and resuspended with a detergent solution [DI water with 1% Tween 80 and 1% sodium dodecyl sulfate (SDS)] and then homogenized. One drop of antifoam was added to facilitate total sample recovery. Next, the sample was washed and resuspended with the detergent solution, or DI water, and finally sonicated immediately prior to the layering onto a clarification gradient. Centrifugation on Sheather's (500 g sucrose, 320 mL DI water, and 9.7 mL liquid phenol, 1.29 g/mL) was used to separate oocysts from sediments. Five or ten mL of sample was layered onto 10 or 30 mL of a 3/5 or 4/5 dilution of Sheather's solution (1.17 g/mL and 1.24 g/mL, respectively). The tubes were centrifuged at 1200 × g for 10 minutes and the supernatants were recovered. Potassium citrate (40% solution, 1.16 g/mL) was used to separate Giardia cysts from sediments. The samples suspended in DI water were used for this centrifugation. A 1:3 ratio of sample to media (10 mL to 30 mL) was used. The tubes were centrifuged at 800 × g for two minutes and supernatants were recovered. Oocyst and Cyst Detection Recovered supernatants were washed. Samples were filtered, either directly or after 1:10 dilutions through 13 mm cellulose nitrate membrane filters (pore size, 1.2 µm, 5.0 µm for oocysts and cysts, respectively). Monoclonal antibodies, directly conjugated to fluorescein isothiocyanate (FITC), were used to detect both Giardia (2) and Cryptosporidium (21). An indirect immunofluorescent technique was also used for Cryptosporidium identification (9). The Giardia filter was counterstained with 0.003% Evans' Blue. The filters were rinsed and mounted on glass slides in a glycerol solution. The organisms were enumerated on the filter using epifluorescent microscopy. Number of organisms per liter of water was then calculated. Efficiency Studies The efficiency of the separation procedures was evaluated for both Cryptosporidium and Giardia. One to five mL of sedimented pellet from water were seeded with Cryptosporidium oocysts previously purified on percoll gradients (3). After concentrations were determined, each was subjected to the full strength sucrose centrifugation and various dilutions, as previously described. A stool sample containing Giardia cysts was suspended in the detergent solution, and after counts were made, six centrifugation solutions, (full strength Sheather's, 4/5 and 3/5 dilutions of Sheather's, zinc sulfate, 40% potassium citrate, and percollsucrose (18) were evaluated. A number of variables were tested including sonication and detergents. The total supernatant was collected and recoveries were calculated. The efficiency of Cryptosporidium oocyst recovery from tap water and sewage was evaluated. One hundred gallons of dechlorinated tap water or secondary effluent were seeded with known levels of Cryptosporidium oocysts. This water was filtered and the sample processed as previously described. The initial number of oocysts was determined by addition of oocysts from a stock suspension (in deionized water with 1% Tween 80) to 50100 mL of the tap water. After a sample was taken for assay, the tap water suspension was then added to the 100 gallon volume. Results and Discussion To improve protozoa recovery from water samples, the entire procedure was evaluated at four distinct steps. These include: 1) sample collection; 2) filter elution; 3) sample reconcentration and clarification; and 4) parasite detection. Each step was optimized to enhance the overall method efficiency. During the collection step, two filters were evaluated for their ability to recover Cryptosporidium oocysts from secondary effluent. Both were teninch cartridge, polypropylene, yarnwound filters with a 1 µm nominal porosity. Filter 1 was made as a continuous spiral of a single strand, woven back and forth. In contrast, the polypropylene for Filter 2 was applied in a blanket form. Filter 2 was superior with an average oocyst recovery of 36.6% as compared to a 3% recovery with Filter 1. The superior performance of Filter 2 may be due to increased oocysts entrapment or due to increased recovery of oocysts from the filter during backflushing and washing. The latter is suspected since Filter 2 washed cleaner during elution and almost twice the volume of pellet was recovered. The filter elution step as compared to the other procedures is the most cumbersome and time consuming and possibly may be where the greatest loss in oocyst or cyst recovery occurs. Previously it had been shown that backflushing of the filter enhanced elution (14). During seeded tap water studies, it was found that initial backflushing recovered 16% of the oocysts. Further processing of the filter was, therefore, necessary. After cutting the filter apart, an initial washing recovered 20% more oocysts. Second and third washings achieved a 58% recovery (data not shown). Additional washings with greater volumes did not appreciably increase the efficiency of the method. Currently, three replicate washings of the filter, after backflushing, with volumes of approximately 900 mL each are recommended for optimal elution. The total eluent volume of 2700 mL can be concentrated to a single pellet in large volume capacity centrifuge, making sample processing more convenient. The third step in the procedure, clarification of the sample, is necessary to remove interfering debris without concomitant loss of the organism. Musial et al. (14) found that Cryptosporidium oocyst recoveries using a Sheather's media were dramatically improved by the addition of Tween 80 and SDS to the sample. It was speculated that the detergents may act by disrupting hydrophobic and electrostatic interactions between oocysts and sediment. Although Sheather's gradients with detergents were efficient (82% recoveries), detection of low oocyst numbers was difficult since samples were never totally cleared of debris. Dilution of fullstrength Sheather's solution by 4/5, 3/5, 2/5, and 1/5 (1.24 g/mL, 1.17 g/mL, 1.11 g/mL and 1.06 g/mL) decreased recoveries to 72, 76, 67 and 18%, respectively (Table 1). The sample was sufficiently cleared, however, using the 4/5 and 3/5 dilutions to detect as low as 0.06 oocysts/L. In addition, an
Page 207 TABLE 1. Recoveries of Cryptosporidium Oocysts with Sheather's. Sheather's
Specific Gravity (g/mL)
Number of Trials
Average % Recovery
Full strength1
1.29
3
82
4/5*
1.24
4
72
3/5*
1.17
6
76
2/5*
1.11
6
67
1/5*
1.06
3
18
1 500 gm sucrose/320 mL deionized water, 9.7 mL phenol * Dilutions of full strength
equivalent of 378 L could be filtered through a single membrane filter. The clarification step was also evaluated for Giardia recovery. Six gradient solutions were tested (Table 2) and the entire supernatant was collected and examined in each case. Recoveries averaged 76, 77, 70, 68, 66 and 40% for potassium citrate, percollsucrose, Sheather's, 4/5 Sheather's, 3/5 Sheather's, and zinc sulfate, respectively. Using ANOVA, no statistical significant difference was observed between the various media. Potassium citrate and 4/5 Sheather's resulted in cleaner preparations, however, when used for environmental samples. These two gradient solutions were also evaluated for Giardia recovery without the use of detergents or sonication. A significant decrease in recovery at the 95% confidence limit was found for both potassium citrate and 4/5 Sheather's when used without detergents, 41 and 35.6% as compared to 76 and 68% with detergents, respectively (Table 3). Sonication was not found to statistically increase recoveries. The final step in the procedure is the detection and enumeration of oocysts and cysts. The use of monoclonal antibodies has increased the sensitivity of parasite detection and the ability to accurately detect the targeted microorganisms (2, 9, 14, 17, 21). In addition, the accurate enumeration of Giardia on membrane filters had been previously reported (20). Various membrane filters, including cellulose nitrate, cellulose triacetate and polycarbonate, were examined for their ability to retain oocysts and cysts. Polycarbonate filters were not satisfactory for this application because many organisms were TABLE 2. Recovery of Giardia cysts from stools using various solutions. Centrifugation Solutions
Specific Gravity (g/mL)
Number of Trials
Average % Recovery
40% Potassium Citrate
1.16
5
76
Percollsucrose
1.09
5
77
Full Strength Sheather's
1.29
6
70
4/5 Sheather's
1.24
5
68
3/5 Sheather's
1.17
5
66
Zinc Sulfate
1.18
5
40
1 500 gm sucrose/320 mL deionized water, 9.7 mL phenol
TABLE 3. Effect of Detergents on the Recovery of Giardia cysts during clarification.
Average Recoveries(%)
Solution Potassium Citrate (40%) 4/5 Sheather's
2
With Tween 80 SDS1
Without Tween 80 SDS
76.2
41.0
68.0
35.6
1
Sodium Dodecyl Sulfate
2
500 gm sucrose/320 mL deionized water; 9.7 mL phenol
lost during the antibody washing part of the protocol. Cellulose nitrate and triacetate filters performed well. Filter pore size was also important. Membrane filters with a 5.0 µm porosity were 100% efficient in the retention of Giardia cysts, while 98.2% of the Cryptosporidium oocysts were lost. Oocyst losses of 57.5% were found using a 3.0 µm porosity, filter; thus filters with a pore size of 1.2 µm, giving 100% retention of the oocysts, were used for Cryptosporidium (Table 4). The foregoing improvements in methodology were put to use to study Giardia and Cryptosporidium occurrence in environmental waters. Wastewaters and surface waters were sampled and cyst and oocyst concentrations were determined. Raw sewage samples were positive for both protozoa (Table 4). Cryptosporidium oocysts averaged 5191/L while Giardia cysts averaged 51/L. Cryptosporidium was detected in all treated effluents (average 1374 oocysts/L) while Giardia was detected in only 40% of the samples (average 1.3 cysts/L). Based on this preliminary data, it appears that Cryptosporidium may be less efficiently removed by secondary sewage treatment processes than Giardia. Cryptosporidium concentrations were 100 times higher than Giardia in raw sewage. There may be a number of possible explanations for this: 1) there may be more individuals infected with Cryptosporidium; or 2) those infected are excreting large numbers of oocysts; or 3) perhaps the overall efficiency of recovery and detection is lower for Giardia. Three different surface water sources have been examined for the parasites (Table 6). Of 21 samples, TABLE 4. Enumeration of cysts and oocysts on membrane filters.
Organism
Number Of Trials
Percent Recovered
Type
Pore Size µ
Cryptosporidium
4
1001
CN2
1.2
Cryptosporidium
2
42.5
CN
3.0
Cryptosporidium
2
1.9
CN
5.0
3
Giardia
4
100
CT
5.0
Giardia
3
46
PC4
5.0
1 Counts on 0.2 µ filters chosen as 100% for comparison. 2 Cellulose nitrate. 3 Cellulose triacetate. 4 Polycarbonate.
Filter
Page 208 TABLE 5. Concentrations of Cryptosporidium and Giardia in Sewage in Arizona. Range of Volumes Sampled (L)
Oocysts/L Range (Average)
Cysts/L Range (Average)
Raw sewaged
34170
84513,738 (5191)
0.7198 (51)
Chlorinated Secondary Effluent
121757
1433699 (1374)
02.6 (1.3)
Type of Sample
Cryptosporidium and Giardia were detected in 19, and 12 of the samples respectively, and oocyst concentrations were 10 to 20 times higher than cyst levels. Source 1, a river (originating from source 2) which ran through an area concentrated with cattle pastures had the greater numbers of both parasites. Source 2, a lake that was receiving domestic effluents had lower numbers (7.1 oocysts and 0.35 cysts/L) of both parasites. Source 3 was a river in a protected watershed. The low numbers of oocysts and cysts may have come from indigenous animal species since both Cryptosporidium and Giardia can be found in other mammals (5,8,22). Waterborne transmission of Giardia is well established and recently (6) has been documented for Cryptosporidium (7). Sewage contamination of drinking water has been responsible for many of these outbreaks. In a few cases, outbreaks have occurred in chlorinated water systems, free of coliform bacteria. In addition, Giardia outbreaks have occurred where animals other than man have been incriminated as the contaminating source of a water supply. It has been suggested, therefore, that all surface waters used for potable purposes be filtered and disinfected to prevent the transmission of Giardia (1). Similar concerns and recommendations may be stated in the future for Cryptosporidium. Waters free of Giardia can not be assumed to be free of Cryptosporidium since this coccidian protozoan was always found frequently and in greater numbers. During one investigation of an outbreak of Giardia in campers, the suspected stream sample yielded no Giardia, however, coccidian oocysts in small numbers were observed (4). Further research with improved methods is necessary to document the environmental occurrence of Cryptosporidium and Giardia. TABLE 6. Concentrations of Cryptosporidium and Giardia in surface waters. Source Description
Number of Samples
1 River with domestic effluent discharges and cattle pasture runoff
8
2 Lake with domestic effluent discharges
10
1.2
7.1
3
Cysts/L
28.5
3 River in a protected watershed
Oocysts/L
0.35
0.08
0.009
Literature Cited 1. Akin, E.W. and W. Jakubowski. 1986. Drinking water transmission of Giardiasis in the United States. Water Sci. & Tech. 18:219226. 2. American Water Works Association. 1985. Giardia Methods Workshop In: Water Supplies Detection, Occurrence and Removal. AWWA. p. 49. 3. Arrowood, M.J. and C.R. Sterling. 1987. Isolation of Cryptosporidium oocysts and sporozoites using discontinuous sucrose and isopycnic percoll gradients. J. Parasit. 73:314319. 4. Barbour A.G., Nichols, C.R. and T. Fukushima. 1976. An outbreak of Giardiasis in a group of campers. Am. J. Trop. Med. Hyg. 25:384389. 5. Centers for Disease Control. 1982. Human cryptosporidiosis: Alabama. MMWR 31:252254. 6. Craun, G.F. 1986. Waterborne diseases in the United States. CRC Press, Boca Raton, FL. 7. D'Antonio, R.G., Winn, R.E., Taylor, J.P., Gustafson, T.L., Current, W.L., Rhodes, M.W., Gary, G.W. and R.A. Zayac. 1985. A waterborne outbreak of cryptosporidiosis in normal hosts. Ann. Inter. Med. 103:886888. 8. Davies, R.B. and C.P. Hibler. 1979. Animal reservoirs and crossspecies transmission of Giardia. In: Waterborne Transmission of Giardiasis. W. Jakubowski and J.C. Hoff. (eds). Report EPA600/979001. Cincinnati, Ohio U.S. Environmental Protection Agency. pp. 104126. 9. Garcia, L.S., Brewer, T.C. and D.A. Bruckner. 1987. Fluorescent detection of Cryptosporidium oocysts in human fecal specimens using monoclonal antibodies. J. Clin. Microbiol. 25:119121. 10. Jakubowski, W. and T.H. Ericksen. 1979. Methods for detection of Giardia cysts in water supplies. In: Waterborne Transmission of Giardiasis. W. Jakubowski and H.C. Hoff. (eds). Report EPA600/979001. Cincinnati, Ohio U.S. Environmental Protection Agency. pp. 193210. 11. Jakubowski, W. 1984. Detection of Giardia cysts in drinking water: State of the Art. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 263285. 12. Lippy, E.C. and S.C. Waltrip. 1984. Waterborne disease outbreaks 19461980: A thirtyfive year perspective. J. Amer. Water Works Assoc.. 76:6067. 13. Moore, G.T., Cross, W.M., McGuire, D., Mollohan, C.S., Gleason, N.W., Healy, G.R., and L.H. Newton. 1969. Epidemic Giardiasis at a ski resort. N. Engl. J. Med. 281:402407. 14. Musial, C.E., Arrowood, M.J., Sterling, C.R., and C.P. Gerba. 1986. Development of a method for the detection of Cryptosporidium in water. Appl. Environ. Microb. 53:687692. 15. Rose, J.B., Musial, C.E., Arrowood, M.J., Sterling, C.R., and C.P. Gerba. 1985. Development of a method for the detection of Cryptosporidium in drinking water. In: Advances in Water Analysis and Treatment. Water Quality Technology Conference, Houston, TX, Amer. Water Works Assoc., Denver, CO. Dec. 8 11. p. 117. 16. Rose, J.B., Cifrino, A., Madore, M.S., Gerba, C.P., Sterling, C.R. and Arrowood, M.J. (1986). Detection of Cryptosporidium from wastewater and fresh water environments. Water Sci. Techn. 18:233239. 17. Rose, J.B., Madore, M.S., Riggs, J.L., and C.P. Gerba. 1986. Detection of Cryptosporidium and Giardia in environmental waters. Water Quality Technology Conference, Amer. Water Works Assoc. Denver, CO Nov. 1619, Portland, OR. 18. Sauch, J.F. 1985. Use of immunofluorescence and phasecontrast microscopy for detection and identification of Giardia cysts in water samples. Appl. Environ. Microbiol. 50(6):14341438.
Page 209
19. Shaw, P.K., Brodsky, R.E., Lyman, D.O., Wood, B.T., Hibler, C.P., Healy, G.R., MacLeod, K.I.E., Stahl, W. and M.G. Schultz. 1977. A community wide outbreak of Giardiasis with evidence of transmission by a municiple water supply. Ann. Intern. Med. 87:426432. 20. Spaulding, J.J., Pacha, R.E., and G.W. Clark. 1983. Quantitation of Giardia cysts by membrane filtration. J. Clin. Microbiol. 18(3):713715. 21. Sterling, C.R. and M.J. Arrowood. 1986. Detection of Cryptosporidium sp. infections using a direct immunofluorescent assay. Ped. Infect. Dis. 5:51395142. 22. Woo, P.K. 1984. Evidence for animal reservoirs and transmission of Giardia infection between animal species. In: Giardia and Giardiasis. S.L. Erlandsen and E. A. Meyer (eds). Plenum Press, New York, pp. 341364.
Page 211
Comparison of Five Procedures for the Sedimentation of Giardia Lamblia and Other Protozoan Cysts D.R. Pennell*, J.F. Stoebig, D.E. Sampson, and R.F. Schell. State Laboratory of Hygiene, University of Wisconsin, 465 Henry Mall, Madison, Wisconsin 53706, U.S.A.. Three commercial sedimentation products—FeKal CONTrate (Trend Scientific), Fecal Parasite Concentrator (Evergreen Scientific), and ParaPak MacroCon (Meridian Diagnostics) were evaluated for their ability to concentrate Giardia lamblia and other protozoan cysts from Formalinpreserved specimens. All three products, utilizing ethyl acetate as a digestion agent and the (manufacturer supplied) detergent, were compared with the standard Formalinacetate (FA) sedimentation procedure with and without addition of a detergent, Trition X100. Initial and final cyst counts allowed for calculation of the concentration efficacy of each procedure. FA was superior to all three commercial products for the concentration of G. lamblia cysts and for cysts of Entamoeba coli, Entamoeba histolytica, Endolimax nana, and Chilamastix mesnili. Addition of detergent to the standard FA procedure lowered the cyst yields. Although the commercial products simplify the procedure, FA sedimentation results in optimal recovery of G. lamblia and other protozoan cysts.
Introduction The classical Formalinether (FE) sedimentation procedure for concentration of Giardia lamblia, other protozoan cysts, and helminth eggs in fecal specimens was described by Ritchie in 1948 (5). An important modification by Young et al. in 1979 (8) replaced diethyl ether with a less volatile solvent, ethyl acetate. The resulting Formalinacetate (FA) procedure is safer for laboratory use and produces comparable results (1,7). FA is the standard sedimentation procedure for the concentration of these parasites today. Recently three commercial sedimentation products have been introduced. Fecal Parasite Concentrator (FPC), originally described and evaluated against FE sedimentation in 1978 (9), has since been modified and is now marketed by Evergreen Scientific. FeKal CONTrate (FCT; Trend Scientific) has been evaluated against FA (3). ParaPak MacroCon (PPMC; Meridian Diagnostics) has not been previously evaluated. The manufacturers of these commercial products either recommend or stipulate the use of acetate as a lipid digestion agent. Each product is also supplied with a second reagent, described as one which reduces adhesive forces and/or helps to break down fecal aggregates. This reagent was identified as 20% Triton X100 for FPC. It is not identified for FCT or PPMC. The use of Triton X100, in a sedimentation procedure, has been previously described (9). In this study the three commercial sedimentation procedures were compared with FA sedimentation, alone and with the detergent Triton X100 (FAD), for their ability to concentrate protozoan cysts. Materials and Methods Specimens Human fecal specimens were transported to our laboratory in 10% buffered Formalin for routine parasitologic examination. Specimens established as positive for protozoan cysts by conventional FA or FE were used for this study within 1 week of receipt. Only specimens with enough cysts to permit an initial count were used. Of the 43 specimens, 19 contained cysts of G. lamblia, 11 Entamoeba coli, 4 Entamoeba histolytica, 8 Endolimax nana, and 1 Chilomastix mesnili. Fecal Homogenates Each fecal homogenate was prepared by suspending the specimen in 10% buffered Formalin and adjusting the suspension to nine parts liquid, one part sediment. Sediment volumes were determined with a Wassermann graduated centrifuge tube. Sedimentation Procedures With continuous mixing on a magnetic stirrer, a specified volume of each fecal homogenate was transferred for each sedimentation procedure (Figure 1). FA was performed in accordance with the recommendations of the Centers for Disease Control, United States Department of Health and Human Services (4). FA was also performed with the incorporation of 20% Triton X100 (FAD). The commercial procedures were performed according to the manufacturers' instructions. Details of all five procedures are shown in Table 1. Final sediment volumes were determined and adjusted to nine parts liquid, one part sediment. Cyst Counts Cyst counts were performed on all original fecal homogenates and final sedimentation suspensions by making duplicate 1:4 or 1:16 dilutions of each suspension and transferring 10µL of each dilution to a hemocytometer. Physiological saline containing Dobell and O'Conner's iodine was used as a diluent. All counts were adjusted for the dilution. Counts were performed by a trained microbiologist without knowledge of the sedimentation procedure utilized. Calculations For each sedimentation procedure performed with each specimen, a concentration coefficient (CC) was calculated as the ratio of the concentration in the final sedimentation suspension (FC) to the concentration in the intial homogenate (IC): CC=FC/IC. Thus CC>1 shows a higher concentration of cysts in the final suspension. The higher the CC, the greater the concentration efficacy. The mean CC for each procedure was calculated, and the means were compared by the two tailed Ttest for paired samples (6). P<0.05 was defined as significant. Precision was calculated as the coefficient of variation (CV) of the final cyst counts (n=4 for each specimen). * Corresponding author.
Page 212 TABLE 1. Procedural steps and conditions for the five sedimentation procedures.
Procedural steps
Formalin (FA)
Formalin Trend, with Acetate Acetate detergent Trate (FAD) (FCT)
Evergreen, FeKal CON Meridian, Fecal Concentrator Parasite Macro (FPC) Con(PPMC)
7.5 mL
7.5 mL
7.5 mL
7.5 mL
12.5 mL1
2. Detergent added (amount, type)
3 drops, 20% Triton X100
2 drops supplied2
3 drops, 20% Triton X100
10 drops supplied2
3. Acetate added (mL):
3.0
4. Shaking(s)
30 sec.
30 sec.
30 sec.
60 sec.
2 layers gauze
2 layers gauze
metal screen3
plastic mesh3
plastic mesh3
6. First wash sedimentation
650g, 2 min.
650g, 2 min.
500g 2 min.
7. Second wash sedimentation
650g, 2 min.
650g, 2 min.
3.0
3.0
3.0
5.0
1. Fecal homogenate used (mL)
5. Filtration (type)
8. Acetate added (mL): 9. Shaking(s) 10. Final sedimentation
30 sec.
30 sec.
30 sec.
60 sec.
500g, 2 min.
500g, 2 min.
500g, 2 min.
500g, 2 min.
500g, 2 min.
1
Initial volume greater for PPMC than for other procedures, but the homogenate/acetate ratios were constant.
2
Manufacturersupplied, formulation not given.
3
Manufacturer supplied.
Results The precision of the reference FA procedure with four replicates was CV 7.2 and 10.6% for final concentrations of 229,000 and 841,000 cysts/mL respectively (Table 2). For G. lamblia FA without Triton X100 had greater concentration efficiency than FAD (CC 7.5 vs. 6.7; Table 3). The difference between these means was statistically significant (P<0.03). Similar results were obtained for cysts of other protozoans (P<0.02). FA was more effective than any of the commercial procedures (P<0.001 for FA vs. each procedure with G. lamblia and P<0.003 with other protozoan cysts). For both types of specimens the relative efficacy of the commercial products was FCT>FPC>PPMC (P<0.02). Each procedure produced a wide range of CC (Table 3). For PPMC occasional specimens produced CC<1.
Figure 1. Method of Analysis of Sedimentation Procedures.
Discussion Many recent evaluations of parasite sedimentation procedures utilized a method of evaluation whereby the sedimentations and slide preparations were performed as they would be in a routine clinical parasitology laboratory (13, 7, 8). The presumed advantage of this approach is that the findings are applicable to routine practice. However, the routine practice for transfer of a portion of the final sediment and for slide preparation entails a risk of variable results. If cyst concentations vary throughout the sediment pellet, the portion transferred to the slide may not be representative. The ratio of liquid to particulate material may also vary, and the depth of material under a standard slide coverslip can vary both from slide to slide and from one area to another within a slide. In an attempt to eliminate these factors of potential imprecision, the method of analysis used in this study incorporated a standardized final sediment suspension with transfer of a constant volume of well mixed suspension to a hemocytometer for cyst count determination. In addition, for each sedimentation, two suspensions of final sediment were prepared and counted. A precision analysis when applied to the FA procedure produced CV 7.2 and 10.6% at final concentrations of 229,000 and 841,000 cysts/mL, respectively (Table 2). All five sedimentation procedures produced a wide range of CC, yet the relative efficacy of these procedures remained very consistent from one specimen to another. Thus the wide CC range for each procedure appears to reflect variations in the inherent properties of the fecal specimens. For PPMC, occasional specimens yielded a TABLE 2. Precision evaluation for the method of analysis used to compare the sedimentation procedures. Sedimentation Procedure
Mean Cyst Count in Final Sediment Suspension (cysts/mL)
CV
FA
4
229,000
7.2%
FA
4
841,000
10.6%
Replications
Page 213 TABLE 3. Efficacy of sedimentation procedures.
G. lamblia Cysts (n = 19)
FA
FAD
FCT
FPC
PPMC
Mean CC
7.5
6.7
5.6
3.5
2.7
CC range high
16.3
15.7
12.2
8.6
7.3
low
1.8
1.7
1.8
1.1
0.6
Other Protozoan Cysts (n = 24)
Mean CC
6.8
6.0
5.7
4.0
2.5
CC range high
13.0
10.9
9.1
8.0
4.7
low
2.4
1.9
2.3
1.6
0.8
CC<1; i.e., the cyst counts were higher in the initial fecal homogenate than in the final sediment suspension. Incorporation of Triton X100 into the FA procedure decreased FA's concentration efficacy for cysts. This is contrary to the reported efficacy for use of detergent in the concentration of parasite eggs (9). This discrepancy poses a problem for the clinical parasitology laboratory, where optimal recovery of both cysts and eggs is ideal, yet performance of two sedimentations for each specimen is costly. Confirmation of these findings would be useful, as would an investi gation of alternative detergents and/or detergent concentrations. FCT was the most efficacious of the commercial products tested, followed by FPC and then PPMC. However, FA was superior to all three commercial products. This finding is inconsistent with a previous report comparing FCT and FA (3). One apparent difference between the two studies is in the methods used to analyze the procedures. All three commercial products made sedimentation easier, partly through unique product design, especially at the level of specimen filtration, and partly by omission of steps that are required in the standard FA (Table 1). In addition, both FPC and PPMC function as an enclosed processing unit during much of the procedure, improving laboratory safety by reducing the chances of exposure to the specimen, Formalin, and acetate. These conveniences, however, are gained at the expense of sensitivity in cyst detection. Literature Cited 1. Erdman, D.D. 1981. Clinical comparison of ethyl acetate and diethyl ether in the Formalinether sedimentation technique. J. Clin. Microbiol. 14:483485. 2. Garcia, L.S., and R. Shimizu. 1981. Comparison of clinical results for the use of ethyl acetate and diethyl ether in the Formalinether sedimentation technique performed on polyvinyl alcoholpreserved specimens. J. Clin. Microbiol. 13:709713. 3. Long, E.G., Tsin A.T., and B.A. Robertson. 1985. Comparison of the Fekal CONTrate System with the Formalinethyl acetate technique for detection of intestinal parasites. J. Clin. Microbiol. 22:210211. 4. Melvin, D.M., and M.M. Brooke. 1974. Laboratory procedures for the diagnosis of intestinal parasites. U.S. Department of Health and Human Services publication no. (CDC) 758282. U.S. Government Printing Office, Washington, D.C. p. 104106. 5. Ritchie, L.S. 1948. An ether sedimentation technique for routine stool examinations. Bull. U.S. Army Med. Dept. 8:326. 6. Snedecor, G.W., and W.G. Cochran. 1967. Statistical methods. 6th ed. Iowa State University Press, Ames, Iowa. p. 6264, 9197. 7. Truant, A.L., Elliott, S.H., Kelly, M.Tm., and J.H. Smith. 1981. Comparison of Formalinethyl ether sedimentation, Formalinethyl acetate sedimentation, and zinc sulfate flotation techniques for detection of intestinal parasites. J. Clin. Microbiol. 13:882884. 8. Young, K.H., Bullock, S.L., Melvin, D.M., and C.L. Spruill. 1979. Ethyl acetate as a substitute for diethyl ether in the Formalinether sedimentation technique. J. Clin. Microbiol. 10:852853. 9. Zierdt, W.S. 1978. A simple device for concentration of parasite eggs, larvae, and protozoa. Am. J. Clin. Pathol. 70:8993.
Page 215
Comparison of the Modified ''Reference Method" and the Indirect Fluorescent Antibody Technique for Detection of Giardia Cysts in Water Benny E. Quinones*, Charles P. Hibler and Carrie M. Hancock. Department of Pathology, Colorado State University, Fort Collins, Colorado 80523, U.S.A.. Sixty municipal water samples from various geographic areas were analysed by both the modified reference method (Zinc Sulfate) and the Indirect Fluorescent Antibody (IFA) technique to monitor surface water sources for cysts of Giardia. Statistical analysis with the Student's t test indicates no statistically significant difference between techniques if analysts were highly skilled and experienced with the technique they were using. Differences noted were often personal preferences. Inexperienced analysts usually found more cysts with IFA. Both techniques were affected by water quality; recovery of cysts was inversely proportional to the turbidity of the source and high organic turbidity samples containing considerable algae severely limited the efficiency of both techniques. IFA required more preparation time/sample and the microscope fatique factor was greater, therefore IFA could not be used effectively by laboratories monitoring 10 or more samples/day. The cost factor for a quality epifluorescent microscope with phase contrast capabilities and the cost of antibody would be a consideration for laboratories planning to use IFA. Confidence in the results would be contingent upon the availability and reliability of the antibody.
Introduction During the past 11 years we have examined over 6000 water samples for cysts of Giardia. Most of these are from municipalities, County and State Health Departments requesting this service, some are from research on filtration systems and some are from foreign countries. Initially we used the "reference method" described by Jakubowski in Volume 16, Standard Methods (1) and then switched to the overlay/underlay modification of this technique. Sauch (1985) (3) developed an indirect fluorescent antibody technique (IFA) for the detection of Giardia cysts in surface water sources, and with her assistance we incorporated the technique and offered this type of analysis along with the modified reference method we generally use for analysis of water samples. Our laboratory routinely monitors 120 samples/day or 80100 samples/month for municipalities requesting this service. Samples are sent from geographic areas across the United States. Since immunofluorescence has not been used to any great extent for monitoring surface water supplies for cysts of Giardia, we felt that an evaluation of the advantages and disadvantages of both techniques for monitoring purposes was necessary. Recovery of Giardia cysts from contaminated sources during an epidemic of waterborne giardiasis is relatively simple because cyst numbers often reach 1015/gallon of water; however, routine monitoring for their presence frequently means attempting to find one cyst/100 gallons of water and this can be a time consuming procedure necessitating numerous hours of microscope time. Materials and Methods Samples of concentrate from 60 different municipal sources were analyzed by the modified reference method, the technique we generally use with ZnSO4 for monitoring, with a coverslip used in the final step, and with Sauch's technique (3). The only changes made in her technique was the elimination of the dehydration steps and the dye. The results were compared by the Paired Difference Student's t test. Samples for comparison were selected at random. Whenever possible, the same amount of concentrate was used and the chemical used for selective concentration during the first step in the clean up procedure was at the same specific gravity. On occasion further dilution of the concentrate was necessary with the immunofluorescence technique because 25 mm diameter membranes would plug whereas the 47 mm membranes used in the modified reference method did not plug. Moreover, a lower specific gravity sometimes was necessary to overcome the plugging encountered with the smaller membrane size used in the IFA technique. Lower specific gravities of the chemical result in lower cyst recovery (2). Results The samples analyzed varied from western water sources containing primarily inorganic turbidity to eastern water sources containing high levels of organic turbidity and numerous algae, flagellates, etc. in the sample. Irrespective of the technique employed, increasing turbidity interfered with the selective concentration of cysts and other organisms. If the turbidity was primarily organic, or the sample contained alum and/or polymers, considerable dilution was necessary. For a few of the samples, use of chemicals with a lower specific gravity (e.g. 1.1) was necessary. Apparently both techniques were * Corresponding author.
Page 216
equally compromised by the water quality because results of the analysis were statistically insignificant for the 60 samples compared. A total of 446 cysts, ranging from 0 to 50 cysts/samples, averaging 7.4 cysts/sample was found by the modified reference method; and 383 cysts ranging from 0 to 48 cysts/sample, and averaging 6.4 cysts/sample was found by immunofluorescence. The mean of the differences between techniques for each paired sample was 1.07, range +21. This value is not statistically different from 0 (p<0.01). The difference was less than +21 in 17 (28%) of the samples. Discussion The results of this comparison indicated that if analysts were highly skilled and experienced with the technique used for analysis results were not significantly different. However, comparison of the two techniques using less experienced analysts indicated that inexperienced individuals sometimes would find more cysts by immunofluorescence, but the differences observed throughout the trial were not significant. Use of direct microscopy or the modified reference method for diagnosis of Giardia cysts in water necessitates considerable training in a number of academic disciplines and experience with the microscope. Preferably individuals have training and experience in protozoology, parasitology, stream or fresh water biology and botany. Our experience from training individuals indicates that students with a broad background in biology or microbiology become proficient in a short period of time and gain confidence as they analyze more samples providing they frequently see positive samples or use quality assurance samples. If an individual is not academically trained in these disciplines immunofluorescence may be a better option, but currently we doubt that effective analysis can be performed by either technique without considerable experience and academic training. For the average sample, preparation time and analysis by IFA required approximately twice the time necessary to prepare and analyze a sample by the modified reference method. If laboratories are analyzing 10 or more samples/day the time factor would be prohibitive, especially if the analyst had other duties. Analysts experienced with both techniques felt that the microscope fatigue factor was considerably greater with IFA than with the modified reference method. Laboratories planning to use either technique must realize that investment in an excellent, high quality microscope with a strong, wellbalanced light source is a must. The cost of an epifluorescent microscope with phasecontract capabilities is much greater than a quality brightfield microscope, a financial consideration for private or municipal laboratories contemplating analysis with immunofluorescence. Currently polyclonal or monoclonal antibody is not commercially available to laboratories planning to use immunofluorescence for analysis of water samples. No doubt commercial laboratories will soon make antibody available. The polyclonal antibodies in use by several laboratories are certainly specific for Giardia cysts, however, our antibody will not distinguish between cysts from mammals and those from birds. Future research may show that several "strains" of Giardia are present in human and/or animal populations, and more than one strain may be present in any given human or animal. What criteria will be used for the selection of "strains" to produce these antibodies? Will monoclonal antibody distinguish between Giardia cysts infectious or potentially infectious for humans, as opposed to those from animals or birds that are not infectious for humans? These are questions needing answers before immunofluorescence can be recommended as better than any other technique. Another key problem besides the quality of antibody is the availability of a reliable antibody. An important concern with the IFA technique is that particulate analysis cannot be performed and this analysis is essential to assessing filter plant efficiency. Plant efficiency in removing particulates the size of and larger than Giardia is more important than detecting Giardia. A problem with the modified reference method is that ZnSO4, potassium citrate and sucrose, the chemicals generally used to selectively separate cysts, are hygroscopic and cause shrinkage. This necessitates experience to recognize the cysts, however, cysts were readily visualized and identified by experienced individuals. If percoll was used shrinkage did not occur, unfortunately this chemical is not costeffective for routine monitoring of surface water supplies. Another problem with the reference method, especially when using ZnSO4, is the necessity of staining with Lugol's iodine if a brightfield microscope is used for analysis. The iodine also effectively stains algae and flagellates. Visualizing a cyst among myriads of similar shaped and similar stained organisms quickly initiates microscope fatigue, however, the background fluorescence from algae also interfered with analysis by the immunofluorescent technique and with comparable results. Occasionally algae completely prohibits analysis by IFA. As was shown by the statistical analysis, both techniques apparently were effectively compromised if samples had these organisms present. Personal preferences and biases was a problem for one of the analysts. Since this individual (CPH) had experience with more than one technique he disliked sitting in a darkened room and switching from epifluorescence to phase contrast microscopy to confirm identification of the cysts; visualization and confirmation by Lugol's stained samples using ZnSO4 was quicker and easier. Moreover, evaluation of the quality of the cyst (possibly alive, definitely dead, etc.) was easier with the ZnSO4 technique. Often the immunofluorescence technique would detect organisms that may have been a cyst at one time but was not recognizable and had to be listed as a "cystlike structure". Literature Cited 1. A.P.H.A., A.W.A.A., W.P.C.F. 1985. Standard Methods for the Examination of Water and Waste Water. 16th Edition, A.P.H.A. Washington D.C.. pp. 1268.
Page 217
2. Hibler, C.P. 1987. An overview of the techniques used for detection of Giardia cysts in surface water. (Calgary Giardia Conference). 3. Sauch, J.F. 1985. Use of immunofluorescence and phasecontrast microscopy for detection and identification of Giardia cysts in water samples. Appl. Environ. Microbiol. 50(6):14341438.
Page 219
Giardia Detection using Monoclonal Antibodies Recognizing Determinants of In Vitro Derived Cysts C.R. Sterling*, R.M. Kutob, M.J. Gizinski, M. Verastegui, and L. Stetzenbach Department of Veterinary Science, University of Arizona, Tucson, Arizona 85721, U.S.A.. Monoclonal antibodies (MAbs) were made to in vitro derived cysts of an axenically cultured Peruvian Giardia isolate. Encystment of the trophozoites was accomplished using Keister's Modified TYIS33 medium without bile incubated under a 90:10 nitrogen:carbon dioxide atmosphere. BALB/c mice were immunized with sonicated in vitro derived cysts. Hybridoma fusions were performed using spleen cells of sensitized mice and P3x63Ag8.653 myeloma cells. Antibody to cysts was detected in 11 of 144 wells by indirect immunofluorescence. Positive wells were cloned and MAbs were recovered in ascites fluid. One MAb (2B3G6) an IgG, gave an IIF titer of 100,000. This MAb and culture supernatant from uncloned positive wells (6C1 and 2B3) were highly specific for Giardia cysts and did not cross react with Entamoeba histolytica, E. coli, E. hartmanni, Eimeria sp., Cryptosporidium, Candida sp., Rhodotorula sp., or algae. The MAb and the culture supernatants were used to detect Giardia cysts in fecal smears from symptomatic and asymptomatic patients, primary, secondary, and tertiary sewage effluents from Peru, and surface waters from Arizona using indirect immunofluorescence. The specificity of these MAbs greatly enhanced Giardia detection and allowed for the rapid screening of fecal and environmental samples.
Introduction New and improved diagnostic methodologies have recently been developed to detect Giardia in symptomatic and asymptomatic individuals. Enhanced Giardia detection has largely resulted from the development of several highly sensitive immunodiagnostic procedures including indirect immunofluorescent detection of serum antibodies (20,21), counterimmunoelectrophoresis (CIE) of feces (5), and enzymelinked immunosorbent assays (ELISA) for detection of serum antibodies (21) or the presence of Giardia antigens in fecal specimens (8,19). The principle technique for Giardia detection, however, remains conventional microscopy (4). Microscopic examination of multiple stool samples, flotation concentrates, duodenal contents, and intestinal biopsies for characteristic trophozoites and cysts of Giardia may be required for confirmation of infection (4,13,15,22). Positive identification of organisms from these samples is definitive. Unfortunately, not all laboratories are skilled in performing many of these techniques, resulting in variable rates of accurate diagnosis. Detection of Giardia in environmental waters has been even more difficult. First, it is usually necessary to process large volumes of water to detect the few cysts likely to be present. Second, detection and reporting errors may result from the presence of algal cells, artifacts or other cell types resembling Giardia cysts in size and shape. Immunofluorescent staining procedures have been used rather sparingly to detect parasites despite the fact that they have been available for over 40 years (3). Recently, several workers have developed polyclonal antibodies to Giardia and Cryptosporidium and described their use in either indirect (IIF) or direct (DIF) immunofluorescent assays to detect these organisms in fecal and environmental samples (14,16,18). In addition, monoclonal antibodies (MAbs) have been produced to determinants of the oocyst wall of Cryptosporidium and similarly used in IIF and DIF assays (7). The use of the MAbs was considered to be at least ten times more sensitive than other specialized staining techniques. The successful development and application of MAbs for the diagnosis of Cryptosporidium within our own laboratory led us to attempt a similar strategy at developing MAb reagents which could be used to identify Giardia. In the present study, we describe the development of these MAbs using in vitro derived Giardia cysts for mouse sensitization prior to hybrid fusions. In addition, we describe the use of these MAbs to detect Giardia in both fecal and environmental samples using an indirect immunofluorescent assay. Methods In vitro Encystment An axenically cultured human Giardia isolate of Peruvian origin was induced to encyst in vitro. Four day old trophozoite cultures grown in Keister's TYIS33 medium (10) were placed on ice for 10 minutes and harvested by centrifugation for 10 min at 250 × g.. The pellets of two culture tubes were combined and incubated under a 90:10 nitrogen:carbon dioxide atmosphere for 48 hrs. in 1.5 mL of Keister's Modified TYIS33 medium without bile. Cultures were then placed at 23°C for 24 hrs. after which they were centrifuged and the pellets resuspended in 1.0 mL phosphate buffered saline (PBS, pH 7.4) for examination by phase contrast microscopy. * Corresponding author.
Page 220
Figure 1. Phase contrast image of in vitro derived cysts. Bar, 10µm.
Scanning Electron Microscopy In vitro derived cysts were preserved with 4% formalin, 1% glutaraldehyde in 0.1 M. phosphate buffer, pH 7.2 (12) and filtered onto 0.1µm polycarbonate filters. Filters were postfixed with 2% osmium tetroxide (OsO4), dehydrated through a graded ethanol series, and critical point dried using ethanol:Freon TF (50:50) (2). Processed filters were mounted on stubs and coated with gold:paladium. Samples were viewed with an ISI DS 130 scanning electron microscope. Giardia Cyst Viability The in vitro derived cysts were tested for viability using in vivo and in vitro assays. Twelve 45 day old CF1 mice were selected for the in vivo viability study. Nine of the mice were fed 8 × 104 cysts/mouse and the remaining three served as controls. One control mouse and three test mice were sacrificed on day 4, 6, and 8. Sections of intestine were excised, placed in PBS, and examined for Giardia trophozoites by phase microscopy. In vitro excystation, as a measure of viability, was attempted using the technique of Bhatia and Warhurst (1). In vivo derived cysts were added to hydrochloric acid (pH 2) for 15 minutes, centrifuged (250 × g, 10 min.), and washed twice in Keister's Modified Medium Solution A. Cysts were then suspended in Keister's complete medium and incubated at 37°C. The suspension was examined for trophozoites at timed intervals (0, 10, 20, 60, 90 min., 2 hrs., and 24 hrs.) using phase microscopy. Monoclonal Antibodies Monoclonal antibodies recognizing Giardia cyst wall determinants were produced using the in vitro derived cysts as a source of antigen for mouse immunizations. BALB/c mice were immunized with sonicated suspensions of 106cysts/mouse at day 0 (Freund's complete adjuvantI.M.), 14 (Freund's incomplete adjuvantI.M.), and 27 (I.V.). The mice were bled from the retroorbital plexus on day 30 and the serum tested for IIF reactivity to Giardia cysts obtained from a human fecal sample. Spleen cells of seropositive immunized mice were fused on day 31 with P3×63Ag8.653 myeloma cells using polyethylene glycol 4000 (6). Antibody to the human source Giardia cysts was detected in 11 of 144 fusion wells by IIF (11). Culture supernatants from two uncloned expanded cell lines (6C1 and 2B3) were used for initial Giardia screening. Cells from fusion well 2B3 were cloned by limiting dilution and put into pristane primed BALB/c mice for ascites production (6). Ascites fluid from this cloned cell line and culture supernatants from uncloned cell lines were screened for cross reactivity with various protozoa, yeast, and algae in an IIF assay. Immunoglobulin isotype and subclass was determined by IIF using biotinylated subclass specific antisera and strepavidin labelled fluorescein isothiocyanate (FITC). Detection of Giardia in Fecal Samples Formalin fixed human fecal samples submitted by Maricopa County Health agencies were screened for Giardia by IIF using the MAbs. Samples were obtained from routine submissions by newly
Figure 2. Scanning electron micrograph of in vitro derived cysts. Note trophozoite on the left. Bar, 10µm.
immigrated Asian nationals and from hospital and clinic patients presenting with diarrhea. Samples were identified only by accession number. Fecal smears were air dried, heat fixed and stored at 23°C until further processed by IIF. Canine fecal samples submitted by local veterinarians as suspect Giardia cases were similarly tested. Surface Water and Sewage Effluent Sampling Surface waters used for recreation and as drinking water sources were sampled using a modified Environmental Protection Agency (EPA) filtration technique (8). Filter processing for Giardia cyst detection was carried out using modifications of techniques designed to isolate Cryptosporidium oocysts (C.E. Musial, Ph.D. Dissertation, University of Arizona, Tucson, 1985). Pellets from processed samples were smeared on glass slides, heat fixed, and stained for IIF using MAbs. Primary, secondary, and tertiary sewage effluents collected in Lima, Peru were similarly treated with the exception that filtered effluents were formalin fixed on site for transport back to the United States for processing. Results Rates of Giardia encystment ranging from 10 25% were observed during the conduct of our experiments. In vitro cyst formation was observed at the end of the 48 hr. incubation at 37°C under a 90:10 N2:CO2 atmosphere in medium without bile. Maximum encystment was observed following a further 24 hr. incubation at 23°C. Prolonged incubation beyond 24 hrs. at 23°C did not result in an increased number of cysts. Phase contrast microscopy showed cystlike structures with discernible walls (Figure 1). Characteristic internal cyst structures, i.e., axonemes and median body, were not observed. Scanning electron microscopy of in vitro cultures showed smooth walled cyst forms distinct from trophozoites (Figure 2). Viability of the in vitro derived cysts as assessed by in vivo and in vitro techniques was not demonstrated. Culture supernatants from uncloned cell lines (6C1 and 2B3) and MAb from ascites of a cloned line in mice (2B3G6) were specific for Giardia and did not cross react with Entamoeba histolytica, E. coli, E. hartmanni, Eimeria sp., Cryptosporidium, Candida, Rhodotorula, or algae. Peripheral cyst fluorescence was observed using both the uncloned culture supernatants and the cloned ascites fluid in IIF assays (Figure 3). In identical microscopic fields
Page 221
Figure 3. Epifluorescent illumination of cysts stained with MAb 2B3G6. Bar, 10µm.
viewed by phase contrast microscopy it was often difficult to discern the cysts (Figure 4). Monoclonal antibody 2B3G6 gave an IIF titer of 100,000 and was determined to be an IgG1 subclass. Culture supernatant 6C1 and MAb 2B3G6 were used to detect Giardia in 6% of over 900 human fecal samples submitted to our laboratory during the month of October, 1986. A portion of the samples positive by IIF for Giardia (16%) had been reported as negative following routine microscopic examination (personnel communication, Maricopa County Health Department). All of these positives were reported as 1+ infections by IIF. Two canine fecal samples were also positive for Giardia cysts using IIF. Giardia cysts were detected in four of 72 surface water sites sampled within Arizona (5.5%) and they were observed in high numbers in sewage effluents from Peru using IIF. Discussion Examination of feces by conventional microscopy is currently the most available and widely used clinical diagnostic test for detecting Giardia infections. The preparation required for such an examination may be quite time consuming and involve the collection of multiple fecal samples with subsequent processing by various flotation techniques to concentrate cysts. Even then, cyst detection may not readily be accomplished because of
Figure 4. Phase contrast image of cysts in Figure 3 (arrows). Bar, 10µm.
technician inexperience in recognizing Giardia, or in not being able to distinguish them from other protozoan cysts or artifacts. Permanent stains, which enhance internal cyst features, are not widely used because of the additional expense and time involved in specimen preparation and technician training. Examination of specimens obtained by procedures such as the EnteroTest and intestinal biopsies are likely to increase the chances of detecting infection, yet suffer from being invasive (13). Serologic testing, either by immunofluorescence (20,21) or the ELISA (21) has the disadvantage of not always being able to detect active infection because of variable antibody persistance. The CIE assay to detect Giardia antigens in feces (5) depends on subjective analysis and is not readily available for clinical use. Only the ELISA for Giardia antigen detection (8,19) appears to overcome many of the problems associated with the other detection methods. It, however, may still be cumbersome in unskilled hands because of the need for continual standardization of the numerous reagents required to conduct the test. Immunofluorescent visualization of cysts using MAbs, as reported herein, offers a highly specific and sensitive method for detecting Giardia in symptomatic and asymptomatic individuals and in environmental samples. The identification of low level infections using IIF in individuals reported as negative by conventional microscopic examination underscores the sensitivity of this technique.
Page 222
In addition, fecal or environmental slides prepared for IIF initially could be screened using a low power objective (10×) to identify fluorescing cysts. Thereafter, a higher magnification (40×) could be used to confirm cysts shape and size and characteristic internal structures after switching to phase contrast microscopy. The absence of cross reactivity of the MAbs with other commonly encountered fecal protozoan cysts, yeasts, or algae, however, makes phase contrast confirmation of internal structures unnecessary. Polyclonal antisera prepared against Giardia cysts (14,16) and Cryptosporidium oocysts (18) have similarly been used to identify these infectious agents from individuals and environmental samples. Such sera, however, are not easily produced in large quantity and often lack the specificity of MAbs. The use of in vitro derived cysts to generate MAbs recognizing epitopes of human and animal Giardia cysts indicates that the two forms share common determinants. Further confirmation of identity between these two cyst forms comes from the observation of identical immunofluorescent staining patterns following use of the IIF assay. Unfortunately, it appears that the in vitro derived cysts were not viable. The factors responsible for inducing encystation remain unknown despite our success at getting 1025% of trophozoites growing in vitro to encyst. Overall, we consider the immunofluorescent approach to the identification of Giardia cysts from fecal or environmental samples to be superior to other microscopic methods of detection. The superiority of this technique employing the use of MAbs as primary antibodies in the IIF assay has previously been demonstrated in a clinical study (7). Because cysts fluoresce so brightly, even when viewed with a 10× objective, only a few need be present on a slide for detection. Under most circumstances this would virtually eliminate the need for the use of time consuming flotation techniques required to concentrate cysts. Enhanced detection would also make it easier to identify asymptomatic cyst excretors or patients with variable cyst excretion patterns. Likewise, use of this technique will make it easier to detect the few cysts that are likely to be found in positive environmental samples. Finally, because the MAb we are using is an IgG1, we may assume, based on past experience (17), that the immunofluorescent assay can easily be converted to a direct procedure. This will serve to shorten procedural time considerably. Acknowledgements This investigation was supported in part by Arizona Disease Control Research Commission contract 827700000011AQ6622, USDAAnimal Health and Disease Research Award ARZT360458A0204 and Thrasher Research Fund 27985. The authors are grateful to ShanAnne Edwards, Humberto Mena, Staci Matlock Mena, Shannath Merbs, and Lisa Shubitz for their technical assistance. The authors are also grateful to Marilyn Marshall for project environmental sampling coordination and Jim Topping of the Maricopa County Health Department for patient fecal samples and data. Literature Cited 1. Bhatia, V.N., and D.C. Warhurst. 1981. Hatching and subsequent cultivation of Giardia intestinals in Diamond's medium. J. Trop. Med. Hyg. 84:45. 2. Cohen, A.L.. 1974. Critical point drying. In: Principles and Techniques of Scanning Electron Microscopy, Vol. I. M.A. Hayat (ed.), Van Nostrand Reinhold Co., New York. 3. Coons, A.H., Creech, H.J., and R.N. Jones. 1941. Immunological properties of an antibody containing a fluorescent group. Proc. Soc. Exp. Biol. 47:200202 4. Craft, J.C.. 1982. Giardia and giardiasis in childhood. Ped. Inf. Dis. 1:196211. 5. Craft, J.C., and J.D. Nelson. 1982. Diagnosis of giardiasis by counterimmunoelectrophoresis of feces. J. Inf. Dis. 145:499504. 6. Fazekas De St. Groth, S., and D. Scheidegger. 1980. Production of monoclonal antibodies: strategy and tactics. J. Immunol. Methods. 35:121. 7. Garcia, L.S., Brewer, T.C. and D.A. Brnckner. 1987. Fluorescent detection of Cryptosporidium oocysts in human fecal specimens using monoclonal antibodies. J. Clin. Microbiol. 25:119121. 8. Green, E.L., Miles, M.A., and D.C. Warhurst. 1985. Immunodiagnostic detection of Giardia antigen in faeces by a rapid visual enzymelinked immunosorbent assay. The Lancet ii: 691693. 9. Jakubowski, W. 1984. Detection of Giardia cysts in drinking water: State of the art. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyers (eds). Plenum Press, New York. pp. 263286. 10. Keister, D.B.. 1983. Axenic culture of Giardia lamblia in TYIS33 medium supplemented with bile. Trans. Roy. Soc. Trop. Med. Hyg. 77:487488. 11. Kuvin, S.F., Tobie, J.E., Evans, C.B., Coatney, G.R., and P.G. Contacos. 1962. Antibody production in human malaria as determined by the fluorescence antibody technique. Science. 135:11301131. 12. McDowell, E.M. 1978. Fixation and processing. In: Diagnostic Electron Microscopy. B.F. Trump and R.J. Jones (eds.). John Wiley and Sons. New York. 13. Pickering, L.K.. 1985. Problems in diagnosing and managing giardiasis. Ped. Inf. Dis. 4:s6s10. 14. Riggs, J.L., Dupuis, K.W., Nakamura, K., and D.P. Spath. 1983. Detection of Giardia lamblia by immunofluorescence. Appl. Environ. Microbiol. 45:698700. 15. Rosenthal, P. & W.M. Liebman. 1980. Comparative study of stool examinations, duodenal aspiration, & pediatric Enterotest for giardiasis in children. J. Pediatr. 96:27879. 16. Sorenson, S.K., Riggs, J.L., Dileanis, P.D., and T.J. Suk. 1986. Isolation & detection of Giardia cysts from water using direct immunofluorescence. Water Res. Bull. 22:843845. 17. Sterling, C.R., and M.J. Arrowood. 1986. Detection of Cryptosporidium sp. infections using a direct immunofluorescent assay. Ped. Inf. Dis. 5:s139s142. 18. Stibbs, H.H., and J.E. Ongerth. 1986. Immunofluorescent detection of Cryptosporidium oocysts in fecal smears. J. Clin. Microbiol. 24:517521. 19. Unger, B.L.B., Yolken, R.H., Nash, T.E., and T.C. Quinn. 1984. Enzymelinked immunosorbent assay for the detection of Giardia lamblia in fecal specimens. J. Inf. Dis. 149:9097. 20. Visvesvara, G.S., Smith, P.D., Healy, G.R., & W.R. Brown. 1980. An immunofluorescence test to detect serum antibodies to Giardia lamblia. Ann. Intern. Med. 93:802805. 21. Wittner, M., Maayan, S., Farrer, W., and H.B. Tanowitz. 1983. Diagnosis of giardiasis by two methods: Immunofluorescence and enzymelinked immunosorbent assay. Arch. Pathol. Lab. Med. 107:524527. 22. Wolfe, M.S. 1979. Giardiasis. Pediatr. Clin. North Am. 26:295303.
Page 223
Routine Monitoring of Watersheds for Giardia Cysts in Northeastern Pennsylvania Sally A.M. McFarlane Water Quality Laboratory (No. 35114), Pennsylvania Gas and Water Company, 135 Jefferson Avenue, Scranton, Pennsylvania, 185031799 U.S.A. Pennsylvania Gas and Water Company, a large investorowned utility company located in Northeastern Pennsylvania, was faced with an outbreak of Giardiasis in December, 1983. At that time, two reservoirs were deemed contaminated and boil advisories were issued that affected 250,000 customers. Since that time, an in house Giardia testing laboratory was incorporated. Routine monitoring of twentyseven reservoirs in the Company's distribution system was initiated in January of 1985. Laboratory personnel obtained training from Dr. Charles P. Hibler, Colorado State University, Fort Collins, Colorado. The Company's Giardia laboratory presently performs both the zinc flotation or reference method (1) and the indirect fluorescent antibody technique. To date, approximately 1300 filters have been analyzed on both raw and distribution waters for our reservoirs as well as contract work. We are currently recognized by the Pennsylvania State Department of Environment Resources as one of the few "accepted" laboratories for detection of waterborne Giardia. We believe our experiences and data will aid other laboratories and researchers in understanding complexities facing laboratories that may be required to routinely monitor for Giardia.
Introduction Monitoring the Watershed Environment: Bacteriological and Chemical Indicators of Possible Sewage Contamination After the Giardiasis outbreaks, safeguards were implemented by The Company to assure the other twentyfive reservoirs servicing the 1700 miles of pipeline distribution would not be affected. However, only two reservoirs have complete filtration and utilize postchlorination as the means of disinfection. The remainder rely solely on chlorination to remove or inactivate potentially harmful microorganisms. With wildlife abundant throughout the 265 square miles of watershed, cyst multiplication through cross transmission (2) became a constant threat. Elaborate measures taken by the Company after the outbreaks to protect and monitor the watershed included the following: 1. Wildlife control was increased through patrols, establishment of beaver and muskrat trapping programs, and removal of feedstock to discourage beaver habitation. 2. Hired additional laboratory personnel to monitor for bacteriological and chemical indicators of possible sewage contamination upstream as well as routine monitoring of all reservoirs for Giardia cysts (4). 3. Constructed a 2.4 mile pipeline from Nesbitt Reservoir to bypass Spring Brook Reservoir after contamination was found. This increased the chlorine contact time to levels that would inactivate cysts (6). 4. Selective use of reservoirs, where possible, to reduce the risk of Giardiasis. New treatment facilities were built to increase contact times at reservoirs found to be contaminated. 5. Began construction of three new water filtration plants at an estimated cost of 52 million dollars. In January of 1984, an extensive monitoring program of 60 reservoir inlets (performed monthly) and 48 tributaries (performed quarterly) for various bacteriological and chemical parameters commenced. If the tests and consensus of the data accumulated from any of the 108 monitored points are found to be suspicious, it is reported to the Pennsylvania State Department of Environmental Resources. (Table 1). After analyzing the accumulated data for the past three years, we concluded that out of 108 sites, only 8 tributaries indicated high readings of bacterial counts and some chemical parameters analyzed. However, of the 27 reservoirs monitored monthly for Giardia cysts, 22 have at various instances been found harboring cysts. Those found to be contaminated are listed in Table 2. TABLE 1. Monitoring from 1984, 1985, and 1986.
Tests Performed
Ranges
Heterotrophic Plate Count
20 to 2,000,000 colonies/1.0mL
MF Procedure Total Coliform
20 to 2,000,000 colonies/100 mL
MF Procedure Fecal Coliform
<1 to 190,000 colonies/100 mL
MF Procedure Streptococcal
<1 to 20,000 colonies/100 mL
Color
<5 to 165
Turbidity
0.25 to 60.0 NTU
pH
4.1 to 7.7
Temperature
1°C to 24°C (seasonal)
Nitrogen as Nitrate (NO3N)
<0.01 to 3.00 mg/L
Nitrogen as Ammonia (NH3N)
<0.10 to 0.50 mg/L
Biochemical Oxygen Demand
<1.0 to 13.00 mg/L
Total Phosphate
<0.01 to 3.00 mg/L
Total Suspended Solids
<1.0 to 150 mg/L
Detergents (MBS)
<0.01 to 0.02 mg/L
Iron (Soluble)
<0.01 to 1.00 mg/L
Manganese (Soluble)
<0.01 to 0.50 mg/L
Page 224 TABLE 2. Contaminated Sites. Brace Brook(+)
Lake Scranton
Olyphant
Ceasetown(*)
La Rue(+)
Pine Run
Edgerton
Laurel Run(+)
Plymouth Relief
Elmhurst(*)(*)
Laurel Run No. 2(+)
Spring Brook(*)(*)(+)
Fallbrook
Mill Creek
Stillwater(F)
Gardner Creek
Nesbitt(*)
Watres(*)
Huntsville(*)(F)
No. 5 Dam(+)
White Oak(+)
(*) Reservoir with tributary showing higher than normal readings. (F) Reservoir with filtration. (+) Reservoir taken out of service. Important Points: 1. Data indicates NO threat of possible sewage contamination at 15 of the 22 reservoirs known to harbor cysts. 2. Monitoring in two cases was indicative of possible contamination. Elmhurst and Spring Brook are the two reservoirs which were involved in the Giardiasis outbreaks.
Procedures Used in Monitoring Indirect Fluorescent Antibody vs. Zinc Flotation As you are well aware, detection of waterborne Giardia is quite involved and much an art (5). Our laboratory was initially trained by Dr. Charles P. Hibler in July, 1984. After six months of quality control samples, we acquired enough experience and were deemed "accepted" by the Pennsylvania State Department of Environmental Resources. With the Company conscious of Giardia in so many of its reservoirs, as well as the upgrading of the Safe Drinking Water Act (8), training stateoftheart procedures is essential. After additional training at Colorado State University in June 1986, the indirect fluorescent antibody technique was incorporated in the monitoring laboratory (7). Following are our sampling procedures, zinc flotation, and indirect fluorescence techniques. We will elaborate on problems associated with them both. Methods Sampling Procedure Apparatus: Hose, hose adapter, filter housing, flow meter, filter type 1 micron polypropylene Cuno DPPY (1). Location: Water pressure 3035 psi regulator 15 psi, creek/reservoir use gasoline water pump. Approximately 400 gallons over a few hours dependent on algae growth. Maximum 1000 gallons. Zinc Flotation Procedure 1. Process filters in 4 L of deionized water, refrigerate for 24 hours. 2. Aspirate to pellet note volume of filtrate. 3. Underlay with 20 mL 1.28 (sp.g.) ZnSO4 with 20 mLs. of filtrate. 4. Centrifuge 58 minutes. 1500 RPM. 5. Take applicator stick, break interface, aspirate onto 5 µm Nuclepore membrane. 6. Rinse with deionized water, pour into 50 mL tube. 7. Repeat step 4. 8. Aspirate to 6 mLs., transfer into 15 mL tube. 9. Repeat step 4. 10. Aspirate to pellet, add 3 drops Lugol's iodine, fill tube with 1.20 (sp.g.) ZnSO4, add glass coverslip. 11. Repeat step 4. 12. Place coverslip on clean glass slide, scan at 10X. Verify cysts with 40X or 100X. Note ciliates, plant debris, total number of cysts. A minimum of 2 slides per filter should be analyzed. If no cysts are present, report as NONE DETECTABLE. If cysts are present, polaroid pictures are taken. We categorize cysts by appearance per order of the Pennsylvania State Department of Environmental Resources. When considering issuing boil advisories, they arbitrarily chose a minimum contact time of 1 hour at 4.0 mg/L free chlorine as a working dosage for cyst inactivation. Understandably, categorizing cyst quality is solely dependent on the analysts' experience and judgement (Table 3). Indirect Fluorescent Antibody Technique 1. Acquire optimum dilutions of rabbit antiserum and goatantirabbit FITC conjugate purchased from Colorado State university and Miles Scientific, Inc., respectively. 2. Underlay 10 mL of filtrate with 20 mL 1.28 (sp.g.) ZnSO4. 3. Centrifuge 58 minutes at 1500 RPM. 4. Aspirate onto 25 mm 5 µm Nuclepore membrane in Gelman filter holder. 5. Aspirate 200 µl of ultraclean Giardia cysts onto a positive control membrane. 6. Apply 2 increments of 125 µm diluted antiserum onto membranes. 7. Incubate at 37°C for 30 minutes. 8. Clean membranes with approximately 600 mL of PBS (7.07.2). 9. Repeat steps 6, 7, and 8 using diluted conjugate. 10. On clean glass slide, place one drop 4% glycerol solution. Place membrane on glass slide, add drop of glycerol solution. Place glass cover slip on top, remove all air bubbles. Seal with clear fingernail polish. Examine slide immediately at 20X with microscope with fluorescent apparatus. Giardia cysts appear greenapple in color, ranging from 815 µm. In some instances, detection of some internal structures is possible. Discussion As of August of 1986, our laboratory performs both techniques on samples that are found to be positive or suspicious when routinely monitoring with zinc sulfate (Table 4). Each technique has both advantages and disadvantages (Table 5). TABLE 3. Categorization of Cysts. Excellent
cyst clearly reveals three inner structures cysts wall well intact color is bright and very refractive
2.
Good
cyst reveals two or three inner structures cyst wall pulled slightly from outer membrane color is refractive
3.
Fair
cyst reveals two or three inner structures cyst wall pulled away from outer membrane color is less refractive; indicative of an older cyst
4.
Poor
cyst reveals one or two inner structures cyst wall is wrinkled color is either dark brown or green refractiveness is minor
5.
Hardly Recognizable
cyst reveals one inner structure cyst wall is collapsed color is either brown or green refractiveness is slight
1.
Page 225 TABLE 4.
Sample Date
Number Cysts ZnSO4 20 mL Condition
Number Cysts F.A. 10 mL Fluorescence
Laurel Run Raw 81286
None Detectable
1/10 (yes)
Huntsville Raw 81386
3/20 (FairExcellent)
8/10 (yes)
Laurel Run Raw 91686
1/20 (Poor)
Background Fluorescence Monitoring Impossible
Fallbrook Raw 92086
None Detectable
1/10 (yes)
Q.C. Sample C.S.U. 92386
4/20 (Poor)
9/10 (yes)
Edgerton Raw 10986
8/20 (Excellent)
4/10 (yes)
Huntsville Raw 101486
3/20 (Excellent)
11/10 (yes)
Watres Distribution 101786
1/20 (Fair)
3/10 (yes)
Watres Distribution 102186
None Detectable
2/10 (yes)
Watres Raw 102886
2/20 (Poor)
8/10 (yes)
Fallbrook Raw 102986
None Detectable
13/10 (yes)
Edgerton Raw 111086
2/20 (Good)
4/10 (yes)
Zinc Sulfate
Indirect Fluorescence
12 Hours
45 Hours
TABLE 5. Advantages and Disadvantages.
1. Analysis Time 2. Detection of Cyst Quality 3. Expense 4. Cyst Verification 5. Detectability 6. Microscope Fatigue 7. Background Algae Growth
Yes
No
$50.00/Filter
$175.00/Filter
Yes
Not Always
>Zinc Sulfate
Minimal
Excessive
Can be Eliminated
Cannot be Eliminated
When routinely monitoring reservoirs, the zinc sulfate procedure is more time and cost efficient. The fluorescent procedure is effective in assuring against contamination and safeguarding outbreak situations. Acknowledgements Particular thanks is given to Dr. Charles P. Hibler and his staff at Colorado State University for their expertise and dedication in training personnel and extending their generosity and hospitality over the past few years. Additional thanks to the laboratory personnel who all played an important role in preparing this report: Karen Caparo, Richard Ungvarsky, Robert Notartomaso, Lynne Rogers, Suzanne Swartz. Literature Cited 1. APHAAWWAWPCF. 1985. Standard Methods for the Examination of Water and Wastewater, 16th Ed. American Public Health Association, Washington D.C.. 2. Davies, R.B., Fukutaki, K., and C.P. Hibler. 1983. Cross transmission of Giardia. EPA 600/S182013. 3. Jakubowski, W.. 1981. Detection of Giardia cysts in drinking water: stateoftheart. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 263286. 4. Jakubowski, W.. 1984. A reference method for detecting Giardia cysts in water. In: Advances in Water Analysis and Treatment. (12th Annual AWWA WQTC 1985), AWWA. pp. 7378. 5. Jarroll, E.L., Bingham, A.K., and E.A. Meyer. 1981. Effect of chlorine on Giardia lamblia cyst viability. Appl. Environ. Microbiol. 41:483487. 6. Sauch, J.A. 1984. Detection and identification of Giardia cysts using immunofluorescence and phase contrast microscopy. In: Advances in Water Analysis and Treatment. (12th Annual AWWAWQTC1985), AWWA. pp. 7986. 7. Thompson, J.C.. 1986. Updating the safe drinking water act and the drinking water regulations. Water Engineering and Management. pp. 2124.
Page 227
Waterborne Giardiasis: Sources of Giardia Cysts and Evidence Pertaining to their Implication in Human Infection Stanley L. Erlandsen* and William J. Bemrick Department of Cell Biology and Neuroanatomy, 4135 Jackson Hall, University of Minnesota School of Medicine, Minneapolis, Minnesota 55455, U.S.A.. In some waterborne outbreaks of giardiasis, the contamination of water with Giardia cysts has been attributed to the presence of beavers living in the watershed and in reservoirs. Our studies of aquatic mammals, at sites having experienced outbreaks of waterborne giardiasis in 5 New England states, revealed that about 17% of livetrapped beavers and 100% of livetrapped muskrats were infected with Giardia. Crossspecies transmission studies of human origin Giardia, into either beavers or muskrats, were successful in establishing Giardia infections. Despite the high incidence of natural infection in these aquatic mammals and their susceptibility to infection with human Giardia, it is possible that many outbreaks of giardiasis may have been caused by Giardia cysts from other sources, because 1) Giardia cysts from muskrats and other rodents, such as voles, had a morphologically distinct binary appearance (contained 2 mature trophozoites) entirely different from human cysts and were also immunologically dissimilar. 2) Trophozoites isolated from livetrapped beavers possessed a morphological characteristic, a short pair of caudal flagella, distinctly different from human Giardia. 3) Wading migratory birds sharing the same habitat of aquatic mammals were found to contain Giardia. 4) The potential for human waste contamination existed at all waterborne outbreak sites, but in varying degrees. Finally, no apparent alteration in aquatic mammal populations has occurred at these sites, yet no site has experienced a second outbreak of giardiasis. Thus, Giardia cysts in water may be derived from multiple sources.
Introduction The proliferation and interaction of the intestinal protozoan Giardia within the gastrointestinal tract mucosa of both man and animals leads to signs and symptoms of intestinal dysfunction known as giardiasis. The incidence of giardiasis has reached endemic proportions throughout the world and today this disease is one of the leading causes of waterborne epidemic disease in the United States (67). Giardia cysts passed by animals, including man, have been shown to remain viable for months at 410°C in water (38). Water contaminated by animal or human fecal waste may serve as a source for infecting campers, hikers, or even large segments of the population that derive drinking water from a contaminated watershed (9). The sources of waterborne Giardia cysts which infect man may be of human origin, but it has been suggested that some animals may act as reservoirs (13,22) and over 40 different animal species have been reported to harbor this protozoan (31). Animals suggested as being reservoirs for Giardia lamblia from man included rodents, dogs, cats, cattle, muskrats, and beavers, with the latter being considered the most frequent suspect in waterborne outbreaks of giardiasis (13,22,37,38). Discussion Animals Implicated in Waterborne Giardiasis In waterborne outbreaks of giardiasis the most logical animals to be considered as potential reservoirs for Giardia cysts that produce human infection would include not only aquatic mammals such as beavers, but also muskrats, domestic animals, birds and man. Each of these potential animal sources will be discussed in terms of the evidence for their implication. Beavers Within the last decade several suggestions in the literature have implicated the beaver in outbreaks of waterborne giardiasis. First, epidemiological studies indicated that individuals consuming mountain water in Utah and Colorado were at high risk for infection with Giardia (3,69). Secondly, an outbreak of giardiasis in Camas, WA was linked to the presence of beavers in the watershed and it was reported that cysts isolated from beaver feces were capable of inducing giardiasis in specific pathogen free beagles (15). Also, it was reported by Davies and Hibler (13), that two of three humans ingesting Giardia cysts isolated from beavers became infected and displayed Giardia cysts in their stools. Subsequently, in other waterborne outbreaks of giardiasis, the cysts of Giardia were either detected within samples of municipal water obtained from beaver inhabited watersheds or within samples of beaver feces or tissues (Table 1). Despite the attention being focused on the beaver as a ''possible" reservoir for infecting man with giardiasis, it was not at all clear that the beaver should be considered guilty (4). Examination of Table 1 has revealed that only a limited number of beavers in watersheds that supplied water to * Corresponding author.
Page 228 TABLE 1. Relationship between Waterborne Outbreaks of Giardiasis and Naturally Infected Animals Waterborne Giardiasis Outbreak Site* City, State
Ground Fecal Collection Beaver
Muskrat
Livetrapped Animals
Killtrapped Animals Carcass Fecal Sample Beaver
Muskrat
Colonic Fecal Sample Beaver
Muskrat
Small Intestine Smear/Histology Beaver
Muskrat
Giardia Cysts in Municip: Water Supply
Camas, WA
—
—
2/3+
—
1/1+
—
—
—
+
Estes Park, CO
—
—
—
—
—
—
0/10+
0/7+
+
Berlin, NH
—
—
—
—
—
—
1/4+(infected, not shedding)
Zig Zag, OR
+ (# unknown)
—
—
—
—
—
—
—
+
Government Camp, OR
+(# unknown)
—
—
—
—
—
—
—
—
Graeagle, CA
—
—
2/3+
—
—
—
—
—
0
Bradford, PA
—
—
—
—
1/1+
—
—
—
+
Reno, NV
—
—
—
—
1/1+
—
—
—
+
Dover Foxcroft, ME
—
—
—
—
1/2+
—
1/2+
—
+
Pittsfield, MA
—
—
1/9+
—
—
2/7+
—
7/7+
+
+
*based on data from Dykes et al. (1980), Lippy (1979), Keifer (1980), and unpublished data.
human populations having experienced giardiasis outbreaks were tested for the presence of Giardia. Other animal species, including muskrats, birds and man were almost totally ignored as potential sources of cysts. Although the detection of Giardia cysts in the municipal water supplies at the outbreak sites (Table 1) appeared to correlate with the presence of Giardia in the few beavers examined, in retrospect, the lack of a systematic survey of other animals has made this correlation inconclusive. Assessment of the natural prevalence of Giardia in beavers has been carried out in five distinctly different geographical locales as shown in Table 2. A close correlation was seen between the prevalence of Giardia in fecal samples from killtrapped beavers from Washington and New England where approximately 11% of the samples were positive for cysts. On the other hand, the prevalence of Giardia was seen to be somewhat higher in necropsy studies (17%) based on the detection of trophozoites within intestinal scrapings from livetrapped beavers. Analysis of fecal samples from these same beavers (not shown) resulted in a slightly lower prevalence (14%), indicating that the examination of intestinal scrapings at necropsy was a more sensitive method for determining Giardia prevalence in a beaver population. TABLE 2. Natural Prevalence of Giardia in Beavers and Muskrats. Intestinal** Trophozoites
Cysts in Fecal Samples
Geographical Location
Beavers
Muskrats
Beavers
Muskrats
1. New England* (ME,NH,NY,VT)
17% (N=171)
100% (N=97)
11.7% (N=369)
33.1% (N=432)
2. Minnesota*
10.1% (N=89)
100% (N=49)
3. Colorado (13)
18% (N=244)
0% (N=21)
4. Washington (22)
10.7% (N=529)
42% (N=133)
5. British Columbia (26)
14.7% (N=299)
40% (N=20)
* Erlandsen and Bemrick, unpublished observations. ** Detected by light microscopic examination of intestinal scrapings.
Page 229
Figures 12. Scanning electron micrographs of a Giardia trophozoite isolated from the intestine of a beaver livetrapped in Vermont (figure 1) and a Giardia trophozoite seen adherent to the intestinal mucosa obtained from a muskrat livetrapped in New York (figure 2). In figure 1 the trophozoite obtained from an indigenous infection had a short pair of caudal flagella (arrowheads), which stand out in sharp contrast to the long caudal flagella (arrowheads) seen in Giardia trophozoites obtained from either muskrat intestine (figure 2) or human origin (not illustrated). Bar equals 2 microns, figures 1 and 2.
(i) Geographic Distribution of Beavers versus Sites of Giardiasis Outbreaks The beaver has been shown to inhabit most of the North American continent from Alaska and Canada to the Mexican border, except in Western California and the Southeastern United States, including the Gulf coast (44). Due to its extensive geographical distribution, the physical presence of beavers would appear to coincide closely with known sites of giardiasis outbreaks, however, waterborne outbreaks of human giardiasis have occurred in sites where beavers were nonexistent (14,28,36) and Giardia cysts have been detected in water uninhabited by beavers used for drinking purposes (C.P. Hibler, personal communication). Thus, the mere physical presence of beavers in a reservoir or watershed experiencing a giardiasis outbreak may not necessarily have any epidemiological significance regarding zoonotic transmission. (ii) Are the Giardia in Beavers the Same As Those in Man? The assumption has been made in previous studies implicating beavers in waterborne outbreaks of giardiasis that the Giardia organisms found within the beaver were identical to those producing disease in man. Statements have been made that beaver Giardia, presumably cysts, were morphologically identical to Giardia from infected humans (15,35) but no evidence was presented to substantiate these claims, and these findings have been questioned (4). However, evidence has been obtained suggesting morphological dissimilarity between Giardia trophozoites of beaver and human origin. Our light and scanning electron microscopic studies of Giardia trophozoites recovered from naturally infected beavers in New England and Minnesota have revealed a type of Giardia trophozoite that was morphologically distinct from Giardia of human origin. Giardia trophozoites from naturally infected beavers possessed a short pair of caudal flagella (3.3 ± 2.5µm; Figure 1) and was the predominant form, being found in nine of eleven (81%) infected animals studied (Table 3). The length of caudal flagella found in Giardia of human origin was 9.1 ± 2.9µm. A comparison of the length of caudal flagella in beaver Giardia with those from other sources clearly demonstrated that Giardia from indigenous infections in beavers, and one of the two axenically cultured Giardia strains from beavers (IP0482:1), were morphologically distinct in that their caudal flagella were significantly shorter (p=<0.05) than those of other Giardia derived from 1) man, 2) waterborne cysts of unknown origin from a giardiasis outbreak in Pittsfield, MA, 3) indigenous infections in muskrats, and 4) from an axenic beaver culture obtained from another laboratory (IP0583:1). Although it may seem inconsistent that the latter beaver culture (IP0583:1) resembled the flagellar length of human origin, this cultured beaver isolate has been shown in endonuclease restriction experiments to have a DNA banding pattern identical to that of human
Page 230 TABLE 3. Distribution of Trophozoites from Beaver, Muskrats, and Man, Based on Length of Tail Flagella (um) TROPHOZOITE SOURCE
LENGTH OF TAIL FLAGELLA (um)
(n)* <1 2 3 4 5 6 7 8 9 10> Human** (56)
x ± sd
12%
11%
21%
9%
20%
11%
16%
7.8 ± 2.2
13%
7%
21%
9%
35%
9.1 ± 2.3
Human** (60)
11%
Muskrat Livetrapped (98)
1%
3%
12%
10%
12%
14%
11%
36%
8.6 ± 2.4
Beaver Livetrapped (56)
9%
9%
21%
14%
10%
13%
5%
6%
3.3 ± 2.5
Beaver axenic (50) Culture #IP0482:1
2%
14%
16%
12%
32%
14%
4%
4%
2%
4.2 ± 1.6
Beaver axenic (50) Culture #IP0583:1
6%
12%
6%
4%
14%
4%
6%
48%
10.5 ± 4.5
* Number of flagella measured. ** Trophozoites obtained from gerbils infected with Giardia cysts of human origin (courtesy of Dr. F.W. Schaeffer III and Mr. Walt Jakubowski, U.S. Environmental Protection Agency, Cincinnati, OH). *** Trophozoites obtained from beavers infected with Giardia cysts of human origin at University of Minnesota.
origin Giardia, whereas the beaver culture (IP0482:1) possessing the short flagellar length, similar to that found in beavers with indigenous infections, had a DNA banding pattern that was distinctly different (42). The interesting correlation between DNA banding patterns and flagellar length suggested that two different types of Giardia trophozoites may have been isolated from beavers, one resembling the Giardia seen in man while the other type of trophozoite in beavers possessed short caudal flagella and was the predominant type found in indigenous infections. In a variety of protozoa, including Chlamydomonas, Euglena, Ochromonas, Polytomella, and Tetrahymena, the control of the length of flagella and cilia may be regulated by a variety of factors, particularly genetic ones, rather than being determined solely by the availability of assembly competent proteins (33). The presence of two distinct populations of Giardia trophozoites in beavers, based on flagellar length, raised interesting questions as to whether or not this reflected genetic control and therefore, represented different types, perhaps species, of Giardia. The presence of multiple species of Giardia within a single host is not without precedence since rats appear to harbor not only G. muris, but also a second species, G. simoni, that closely resembles the species in man (20,32,58). Also, the recent development of the mongolian gerbil model for Giardia infection demonstrates that Giardia cysts derived from a variety of hosts including mice, muskrats, beavers and man, can successfully infect these animals, and that the Giardia trophozoites or cysts recovered have the same morphological features as the original inoculum. However, it remains to be determined whether or not flagellar length can serve as a marker for different types (species) of Giardia and future developments in this field await the discovery of immunologic probes that can differentiate the different subtypes of Giardia belonging within the G. duodenalis classification of Filice (20). (iii) Infection of Beavers with Giardia Cysts of Human Origin Beavers can be successfully infected with Giardia cysts derived from human fecal samples. In their review of animal reservoirs for Giardia, Davies and Hibler (13) reported that they infected two beavers using an inoculum of 10,000 human cysts derived from an unknown donor(s), but no experimental details were provided. Recently, our laboratory has described crossspecies transmission experiments in which 38 beavers were used to investigate the infectivity of G. lamblia cysts (Bemrick,
Page 231
W.J. et al. 1986). Crossspecies transmission of giardiasis: infection of beavers with human Giardia lamblia, 61st Annual Meeting of the American Society of Parasitologists, abstract 96). Viable human cysts from multiple donors were obtained by fluorescence cell sorting using fluorogenic dyes (55). Successful Giardia infections were obtained in twelve of twenty beavers challenged with oral doses ranging from 5 × 102, 5 × 103, and 5 × 105 cysts per animal. The successful infection of beavers with human origin Giardia indicate that it is theoretically possible that beavers may act as a reservoir of human cysts and potentially could serve as a vehicle for the spread of waterborne cysts. To date, in our studies only a minority of beavers (17% of the livetrapped animals) are known to contain Giardia, and a majority of these animals (81%) appear to have Giardia trophozoites morphologically different from those seen in humans. If the human type of Giardia can be detected within beavers, the presence of such an infection should serve as an impetus for the investigation of other species of water birds or mammals as contributing sources of Giardia cysts, since the presence of cysts in the water is probably required for the establishment of an infection within beavers in the first place. However, since the infection of beavers requires their exposure to human type cysts (endogenous beaver Giardia may be different), the presence of the human type of Giardia in these animals may be an important indicator of contaminated water, derived from other sources, such as raw waste water from sewage spills, septic systems, human recreational usage of the watershed, and possibly, from other infected animals or birds. Muskrats. (i) Animal Distribution and Prevalence of Giardia The muskrat shares the same aquatic environment as the beaver, but has received little attention as a potential source of Giardia cysts, even though our data indicates the prevalence of Giardia infection in the muskrat to be 100% in livetrapped animals (Table 2). Compared to the beaver, population estimates of muskrat density in a watershed are difficult to determine (19). Signs of habitation or feeding activity of muskrats are not as obvious as that of beavers. Many watersheds appear to be capable of supporting a much larger population of muskrats than beavers, because of the formers' lesser impact on the environment in terms of food depletion and also due to their greater reproductive capacity since one pair of muskrats can yield as many as 100 or more progeny in one year versus 46 for beavers. Analysis of Giardia cysts on a per gram basis revealed estimates ranging from 7.4 × 101 to 1.2 × 105 in muskrat feces, while up to 1.4 × 106 [Erlandsen and Bemrick, unpublished observations (60)] cysts per gram were detected in beaver feces. It would appear that beaver feces may have contained more Giardia cysts than that of muskrats, however, none of these studies took into account the daily fecal output of each animal, nor was homogeneity of fecal cyst distribution determined. (ii) Morphology and Immunoreactivity Muskrats are infected with the G. duodenalis species (20) and the trophozoites are similar, if not identical, to Giardia trophozoites of human origin. The median body does have the typical clawhammer shape that characterizes this species. However, the cysts of Giardia isolated from naturally infected muskrats are completely different from human cysts in that each mature cyst contains two fully formed trophozoites, each possessing a formed adhesive disc and a total complement of flagella. Boeck (7) named these binary cysts and they have been detected only in microtine rodents (voles and muskrats). The binary appearance of the Giardia cyst from muskrats is not host dependent since crossspecies transmission of these cysts to either mongolian gerbils or mice, performed in two separate laboratories, results in excretion of Giardia cysts with the same binary morphology (Erlandsen, S.L., Bemrick, W.J., Schaefer, F.W. III, and W. Jakubowski, unpublished observations). The consistent finding of these types of cysts in microtine rodents and their absence in other species, including man, strongly suggests that they may be a separate type or even a different species of Giardia. Immunologic differences have been observed in Giardia cysts from muskrats (and other microtine rodents) when compared to G. duodenalis cysts obtained from dogs, beavers, and man. Immunofluorescent testing of antisera, raised in our laboratory and obtained from others (52,53) directed against Giardia cysts revealed immunostaining of all Giardia cysts tested including those from mouse, beaver, muskrat, dog and man. One antiserum (52) recognized G. duodenalis cysts from dog, beaver, and man, but did not react with G. duodenalis cysts from microtine rodents, including the muskrat, or with G. muris cysts from mice. This immunologic difference clearly indicated that not all cysts within the G. duodenalis species, based on trophozoite morphology, should be considered identical. It also provided additional support for the idea that binary cysts may have been derived from a separate species of Giardia. (iii) CrossSpecies Transmission of Human or Beaver Giardia Cysts to Muskrats Studies in our laboratory on the crossspecies transmission of Giardia cysts isolated from feces obtained from either beavers or humans demonstrate that both types of cysts can successfully establish Giardia infections in muskrats. These results suggest that it may be feasible for muskrats to serve as a vector for the spread of human (or beaver) cysts. Before any conclusions can be drawn regarding the muskrats involvement in human giardiasis, more detailed investigations of the morphological and immunological properties of Giardia cysts collected from naturally infected muskrats need to be performed to determine whether or not muskrats (or other animals) can simultaneously support multiple infections including Giardia indigenous to muskrats and those derived from other species. Domestic Animals and Birds In the past, a number of studies have indicated that Giardia derived from a variety of domestic animal hosts were implicated as possible zoonotic sources of human giardiasis. The experiments of Padchenko and Stolyarchuck (47), Shaw et al. (56), and Hewlett et al. (24) utilized canine Giardia cysts in their crossspecies transmission experiments. Their
Page 232
experimental methodology has been criticized (4) and the suggestion that dogs were involved in the transmission of Giardia to humans, based upon existing data, was considered invalid. Cats have been postulated to play a role in zoonotic transmission of Giardia based on the ability of both cat and human Giardia to infect gerbils (6). This inference, derived from the extrapolation of infectivity information from one animal species to others, may have led to erroneous conclusions. Similarly, other inferences that involved the cat in the transmission of human giardiasis have been made (5) but were based on inconclusive evidence. Woo and Paterson (68) recently reported a series of detailed experiments on crossspecies transmission that involved attempts to infect puppies, kittens, and mice with human Giardia from an axenic culture. They were unable to infect any of these animal species with this human Giardia. Other animals including cattle, sheep, and elk (13) have been suggested as playing some role in the transmission of human giardiasis, but without any data to support this contention other than the detection of Giardia cysts observed in fecal samples. The lack of any information regarding the presence/absence of morphologic or immunologic similarities of Giardia derived from these hoofed animals with other known species of Giardia, has made it impossible to assess their role in human giardiasis. It should be noted that none of these animals have been shown to have a geographical distribution that can be correlated with a majority of known giardiasis outbreak sites, therefore, it seemed unlikely that their role was of general significance, although the possibility of a contributory role at individual sites could not be ruled out. Several species of birds have been shown to harbor species of G. duodenalis (2,31). Giardiasis outbreaks have been reported in birds (48,54) and Box (8) has suggested that the budgerigar may be involved in human giardiasis. The latter seems unlikely since a scanning electron microscopy study (18) of Giardia psittaci trophozoites from budgerigars has shown that this was a separate species, morphologically distinct from any other type of Giardia, and that this species has not been reported in ultrastructural studies of Giardia in animals (16,17,21,25,45,46) or man (41,51,57,63) A recent report (23) of Giardia in a great blue heron, suggested that this species of bird should be considered as a potential source of Giardia cysts in waterborne giardiasis. We have corroborated this finding, and also have found Giardia in green herons and egrets. All of these wading birds have trophozoites with a clawhammer shape median body; a characteristic shared with the type of Giardia found in man and classified as the species, G. duodenalis (20). Wading birds, such as herons and egrets, share the same habitat as beavers and muskrats, therefore, they should be given the same consideration as beavers, muskrats, and man, as potential vectors for the spread of waterborne giardiasis. This is especially true, since they have the capability of moving from one watershed to another over great distances and thus can easily establish new foci of infections without the physical limitations of travel imposed on mammals by the terrain. However, in northern states wading birds may only be temporary or seasonal inhabitants of the watershed, and no information has been reported on the prevalence within any avian species. Man. (i) Prevalence of Giardia in Man and Sewage Giardia lamblia has been considered the most common human intestinal parasite in the United States, being reported in one survey in 3.8% of fecal samples examined (40). A prevalence rate of 26% has been reported throughout the world, although in some areas it has approached 30% (49). A higher prevalence rate, ranging from 2050% was often encountered among infants and children, many of them being asymptomatic (50). Transmission between children has been considered to be of a fecaloral route due to inadequate hygiene. Persons ill with giardiasis have been reported to shed as many as 7.1 × 108 cysts per day (64). Based on a prevalence rate of as little as 1% or as much as 25% of the population, Jakubowski (27) has estimated that for an average city, shedding of cysts by infected persons would lead to levels of Giardia cysts ranging from 3.6 × 104 to 9.0 × 105 per gallon of raw sewage. Values close to the lower estimate have been confirmed recently by Sykora et al. (61) who analyzed raw and treated sewage in McKeesport, PA, and demonstrated that raw waste water contained 5.0 × 103 to 1.5 × 105 Giardia cysts per gallon, while treated sewage released from the plant contained as many as 5.0 × 102 cysts per gallon. The viability of the cysts in raw or treated sewage was not determined. (ii) Waterborne Outbreaks of Giardiasis Human versus Animal Origin for Cysts In the United States, 90 outbreaks of waterborne giardiasis were detected from 19651984, and 73% were related to community water systems using surface water as the source for drinking water (11,12). Waterborne outbreaks of giardiasis have occurred mainly at sites that have traditionally depended upon surface water, where water treatment consisted of disinfection with chlorine without filtration. The sources of cysts in these waterborne outbreaks appeared to have been derived from either contamination with human waste, or possibly, as just discussed, from animals living in the watershed. Human sewage has been shown to contain from 104 to 105 Giardia cysts per gallon (27,61) and was shown to be responsible in the United States for 15% of the cases of waterborne giardiasis during 19651984 (11,12). In 1946, a waterborne outbreak of giardiasis was attributed to sewage contamination of the water supply in a Tokyo apartment building (18). Waterborne outbreaks in the United States suggested as being due to human sewage contamination of surface water supplies included outbreaks during 196465 in Aspen, CO (39), 197475 in Rome, NY (56), 1979 in Bradford, PA (29), and during 1984 in McKeesport, PA (62). Waterborne outbreaks involving filtered surface water supplies occurred during 1977 in Berlin, NH (35), and 1978 in Vail, CO (10). The latter outbreak was the largest involving a filtered water system, having effected an estimated 5000 cases.
Page 233
The direct involvement of animals in waterborne giardiasis occurred in Camas, WA, in 1976, which was reported as the "first substantiated case of wild animals (beavers) contaminating a human population with Giardia. Thirty Giardia cysts were recovered from the raw water in two reservoirs, water influent to the water treatment plant, and in Boulder Creek, one of two feeder streams to the water treatment plant. Beavers were trapped on Jones Creek, the other feeder stream, however, no Giardia cysts were detected in the water at this site. Although three beavers infected with Giardia were trapped on Jones Creek, they were trapped further downstream from the location of the city water intake, than were three beavers that were negative for the parasite (15). The watershed was subject to some logging activity but otherwise had extremely limited human usage, although it had been postulated that the beavers may have been infected with Giardia through exposure to water contaminated with human feces, obtained further downstream. Giardia cysts isolated from beaver feces were reported to be infective in four SPF beagle puppies (15). However, crossspecies transmission of Giardia cysts between animals should not be considered as evidence that it will necessarily occur from animals to man (4). Also, a claim (30,35) was made that the Giardia cysts recovered from beavers and the city water supply appeared morphologically identical to G. lamblia cysts from infected humans, but this should not be surprising since most cysts described (at that time) in warmblooded animals had a striking, if not identical, morphological appearance. Thus, upon reevaluation, the evidence for implication of beavers in the waterborne giardiasis outbreak was not totally convincing, especially in light of the acknowledgement of possible human contamination of water downstream. Subsequent to the waterborne giardiasis outbreaks in Camas, WA and Berlin, NH [an outbreak later attributed to sewage contamination (34)], waterborne outbreaks were in some instances, attributed to the presence of beavers in the watersheds. This occurred even though the potential for human contamination was present, and in certain cases, a more likely explanation. In Red Lodge, MT, a giardiasis outbreak in 1983 was related to increased melt runoff due to the Mt. St. Helens volcanic eruption (66). The investigation revealed beayer, dogs, and cattle within the watershed and they were postulated as potential sources of Giardia cysts. However, the presence of fecal coliforms in samples of tapwater together with human usage of the watershed, including potential overflow from septic systems upstream from the city water intake, suggested that the water was contaminated with human waste, theoretically a more logical source of Giardia cysts infective to man. In Reno, NV, a giardiasis outbreak occurred in 1982 and both Giardia cysts and an infected beaver were detected in one of two water supply reservoirs (43). Giardia cysts were detected in 22/27 samples of raw water from the Truckee River diversions that supplied both reservoirs. As stated by the authors (43) the presence of the beaver in the reservoir may have been coincidental since the rise in giardiasis cases occured several months prior to the arrival of the beaver in the reservoir. Also, Giardia cysts were detected in water from Hunter Creek reservoir, where no beavers were found, but were not detected in Highlands reservoir which was inhabited by the infected beaver. No systematic search was undertaken for other animals infected with Giardia, including the large populations of birds known to frequent these reservoirs. More recently, a large waterborne giardiasis outbreak occured in Pittsfield, MA, in December, 1985 and beavers were immediately implicated as the source of waterborne cysts, resulting in headlines in one of the local papers reading, "Beavers are no laughing matter" (65). The analysis of fecal samples from killtrapped beavers revealed that one of nine animals (11%) was infected with Giardia. This infected animal was one of three trapped in the vicinity (downstream) of the Ashley reservoir, which was indicated as the source of cyst contaminated water. Three muskrats livetrapped in the Ashley reservoir were infected with Giardia. However, humans could also have been a source of Giardia cysts at this site because of human activity in the watershed, including canoeing and fishing. There were also obvious signs of young adult recreational use along the shoreline, even though the area was posted as a restricted access area. The Sources of Giardia Cysts in Water The host source of Giardia cysts in water cannot be attributed to a single specific species of animal based on the use of the current criteria of light microscopic morphology and existing serological methods. Most watersheds and reservoirs have animals that are capable of being naturally infected with Giardia. These animals include beavers, muskrats, various species of voles and rodents, domestic animals and pets. Also, it includes humans and a variety of avian species including wading birds. The cysts from most of these animal sources, as has been previously stated, are essentially morphologically indistinguishable from one another, except for the characteristic binary cyst morphology seen in microtine rodent cysts. Serological tests, based on immunofluorescence (52,53,59) have been used for the identification of Giardia cysts in concentrated water samples, but none of these or any other immunological methods reported to date, have been shown to be capable of distinguishing various species of Giardia based on host origins. They are also incapable of adequately differentiating subpopulations within the G. duodenalis classification (20), the predominant form in mammals and birds. Therefore, since Giardia cysts cannot be easily distinguished from one another on the basis of their morphology, the mere detection of Giardia cysts within water samples and their correlation with the presence of infected animals (and sometimes uninfected ones) in a watershed can lead to erroneous conclusions pertaining to the origin of the cysts. This is particularly important because many animals may have endogenous infections of Giardia, in some cases with a prevalence rate of greater than 90%, that may be entirely unrelated to the Giardia species producing disease in man. In support of this idea, a reevaluation of the use of animal stool surveys in watersheds has also indicated that such
Page 234
surveys may be of little value in assessing the risk for waterborne transmission of giardiasis (12). Evidence has accumulated, over the past ten years, that supported the concept that, within limitations, crossspecies transmission of giardiasis can occur between animals (4). The successful infection of gerbils with Giardia cysts derived from a variety of hosts including mouse, beaver, muskrat, and man, and also the successful infection of beavers and muskrats with human cysts indicated that crossspecies transmission could occur. It was tempting to extrapolate from successful crossspecies transmission of giardiasis within animal models and assume that transmission could occur from animals, such as the beaver or muskrat, to man. However, without any supportive evidence that the cysts in these animals were the same species as that found within the water and within man, it would seem presumptuous to state that any particular animal is the source of the cysts infecting man. If it can be shown that infected animals within the watershed were harboring organisms capable of producing disease in man, then it might be feasible that they may have acted as hosts of a zoonotic infection. However, since it was likely that these animals would obtain their infection from human cysts present in the water, it might be possible that their infection was only a secondary sign reflecting the contamination of the water from an unidentified primary source. Giardia cysts in water would seem to have been derived either from humans or from animals frequenting or inhabiting the watershed. The potential for the presence of large numbers of Giardia cysts in raw human waste [3.6 9.0 × 105 gallon (27)] and the occurrence of human recreation or habitation near or within watersheds has made man a likely suspect as the primary source of Giardia cysts in water. If however, we assumed that animals other than man, such as the beaver, were the major source of cysts in water, then we would be faced with the following paradox. Waterborne outbreaks of giardiasis have never occured more than once at any specific site, despite the fact that the populations of animals, beaver or otherwise, have essentially remained unchanged over the past decade (4). Theoretically, the animals should have been acting as a constant source of cysts. If, subsequent to the initial Giardia outbreak, the same situation regarding the quality of water treatment (chlorine level, contact time, turbidity) were to recur, then why had a second outbreak of giardiasis not occured at any of these sites? It has been postulated (1) that the lack of repetition of giardiasis outbreaks at any site may have been due to either a) the use of corrective measures, such as installing filtration or effective disinfection, or b) some immunity acquired by the population affected. The role of immunity in preventing reoccurrences of giardiasis outbreaks may be questionable, since if animals have always been a continuing source of cysts, then immunity to Giardia should have been ongoing, and thus, prevented the initial outbreak. Another plausible explanation could be that the repeated detection of Giardia cysts in water, together with the lack of a second outbreak, would suggest that not all cysts present within water were possibly infectious to man. Determining the Source of Giardia Cysts in Waterborne Outbreaks of Giardiasis To answer the question as to which animals may be the source of Giardia cysts infectious to man, we would recommend the following approach: 1) there should be a systematic sampling of all suspected animals, mammals and birds, found within the watershed, 2) trophozoites and cysts collected from necropsied animals, cysts concentrated from water samples, and cysts from giardiasis patients, should be used to establish in vitro cultures, 3) the continued development of the technique for in vitro encystation should permit the collection of cysts free of bacterial contamination, 4) using antigens isolated from in vitro induced cysts and each axenic culture of Giardia, attempts should be made to produce immunologic and molecular probes selective or specific for species of Giardia, and 5) Giardia isolated from all sources should be examined and the species determined by means of morphological, biochemical, immunological, and metabolic techniques. The information derived should be compared to results obtained from crossspecies transmission experiments done whenever feasible. The complexity and need for further development of several of the methods described above would preclude their accomplishment within any one laboratory, but a cooperative effort utilizing the expertise available in various specialized laboratories may be successful. Acknowledgements The authors wish to express their appreciation to Ms. Lee Ann Sherlock, Ms. Mary Januschka, and Ms. Lisa Kamp for their excellent technical assistance. We would like to thank Dr. Louis Diamond, Laboratory of Parasitic Diseases, NIH, for providing the axenic Giardia cultures, IP0581:1 and IP0482:1. We also wish to acknowledge the cooperation and assistance provided by Mr. Ed Bogges, Minnesota Department of Natural Resources, Mr. Henry Hilton, Maine Inland Fisheries and Wildlife, Mr. Eric Orff, New Hampshire Department of Fish and Game, Mr. Paul Bishop, New York Department of Environmental Conservation, Mr. Jim Distefano, Vermont Department of Fish and Game, Mr. Jack Korlath, Minnesota Department of Public Health, and Mr. Richard Buech, U.S.D.A. Forest Service at the University of Minnesota. We would also like to thank Mr. Warren Mathis (NY), Mr. Warren Anderson (NH), Mr. Henry Laramie (NH), and Mr. Clarence Rademacher (MN) for their skilled assistance in trapping animals, as well as others too numerous to mention. The research described in this article has been funded by the U.S. Environmental Protection Agency through cooperative agreement No. CR811834, but it has not been subjected to the Agencys' review, and therefore, does not necessarily reflect the views of the Agency and no official endorsement should be inferred.
Page 235
Literature Cited 1. Amirtharajah, A. 1986. Variance analyses and criteria for treatment regulations. J. Am. Waterworks Assoc. 78:3449. 2. Ansari, M.A.R. 1952. Contribution a l etude du genre Giardia Kunstler, 1882 (Mastigophophora: Octimitidae). Am. Parasitol. Hum. Comp. 27:421479. 3. Barbour, A.G., Nichols, C.R., and T. Fukushima. 1976. On outbreaks of giardiasis. Am. J. Trop. Med. Hyg. 25:384389. 4. Bemrick, W.J. 1984. Some perspectives of the transmission of giardiasis. In Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. Erlandsen, S.L. and E.A. Meyer. (eds). Plenum Press, New York, NY. pp. 379400. 5. Belosevic, M., Faubert, G.M., Guy, R., and J.D. MacLean. 1984. Observations on natural and experimental infections with Giardia isolated from cats. Can. J. Comp. Med. 48:241244. 6. Belosevic, M., Faubert, G.M., Maclean, J.D., Law, C., and N.A. Croll. 1983. Giardia lamblia infections in mongolian gerbils: an animal model. J. Infect. Dis. 147:222224. 7. Boeck, W.L. 1919 Studies on Giardia microti. Univ. Calif. Pub. Zool. 19:85134. 8. Box, E.D. 1981. Observations on Giardia of budgerigars. J. Protozool. 28:491494. 9. Craun G.F. 1979. Waterborne outbreaks of giardiasis. In: Waterborne Transmission of Giardiasis. Jakubowski, W. and J.C. Hoff. (eds). U.S. Environmental Protection Agency, Cincinnati, OH. pp. 127149. 10. Craun, G.F. 1984. Waterborne outbreaks of giardiasis: current status. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. Erlandsen, S.L., and E.A. Meyer. (eds). Plenum Press, New York, NY. pp. 243261. 11. Craun, G.F. 1986. Waterborne giardiasis in the United States, 19651985. Lancet ii(8505):513514. 12. Craun, G.F., and W. Jakubowski. 1987. Status of waterborne giardiasis outbreaks and monitoring methods. Water Resources Bull., In press. 13. Davies, R.R., and C.P. Hibler. 1979. Animal reservoirs and crossspecies transmission of Giardia. In: Waterborne Transmission of Giardiasis. Jakubowski, W. and J.C. Hoff (eds). U.S. Environmental Protection Agency, Cincinnati, OH. pp. 104126. 14. Davis, C. and L.S. Ritchie. 1948. Clinical manifestations and treatment of epidemic amebiasis ocurring in occupants of the Mantestsa apartment building, Tokyo, Japan. Am. J. Trop. Med. 28:817822. 15. Dykes, A.C., Juranek, D.D., Lorenz, R.A., Sinclair, S., Jakubowski, W., and R. Davis. 1980. Municipal waterborne giardiasis: an epidemiological investigation. Ann Int. Med. 92:165170. 16. Erlandsen, S.L. 1974. Scanning electron microscopy of intestinal giardiasis: Lesions of the microvillous border of the villus epithelial cells produced by trophozoites of Giardia. In: Scanning Electron Microscopy. 1973. O. Johari (ed). IIT Research Institute, Chicago. pp. 775782. 17. Erlandsen, S.L., and D.G. Chase. 1974. Morphological alterations in the microvillous border of villous epithelial cells produced by intestinal microorganisms. Am. J. Clin. Nutr. 27:12771286. 18. Erlandsen, S.L., and W.J. Bemrick. 1987. SEM evidence for a new species, Giardia psittaci. J. Parasitology. 73:623629. 19. Errington, P.L. 1963. Muskrat Populations. Iowa State University Press, Ames, IA. pp. 664. 20. Filice, F.P. 1952. Studies on the cytology and life history of a Giardia from the laboratory rat. Univ. Calif. Publ. Zool. 57:53143. 21. Friend, D.S.. 1966. The fine structure of Giardia muris. J. Cell Biol. 29:317332. 22. Frost, F., Plan, B. and B. Leichty. 1980. Giardia prevalence in commercially trapped animals. J. Environ. Health 42:245249. 23. Georgi, M.E., Carlisle, M.S., and L.E. Smiley. 1986. Giardiasis in a great blue heron (Ardea herodiasis) in New York State: another potential source of waterborne giardiasis. Am. J. Epidemiology 123:916917. 24. Hewlett, E.L., Andrews, J.S., Ruffin, J., and F.W. Schaefer III. 1982. Experimental infection of mongrel dogs with with Giardia lamblia cysts and cultured trophozoites. J. Infect. Dis. 145:8993. 25. Holberton, D.V.. 1974. Attachment of Giardia A hydrodynamic model based on flagellar activity. J. Exp. Biol. 60:207221. 26. IsaacRenton, J.L., Moriez, M.M., and E.M. Proctor. 1987. A Giardia survey of furbearing mammals in British Columbia. J. Environ. Health. 50:8083. 27. Jakubowski, W.. 1984. Detection of Giardia cysts in drinking water: state of the art. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. Erlandsen, S.L. and E.A. Meyer. (eds). Plenum Press, New York, NY. pp. 263286. 28. Jephcott, A.E., Begg, N.T., and I.A. Baker. 1986. Outbreak of giardiasis associated with mainswater in the United Kingdom. The Lancet 29:730732. 29. Keifer, A., Roberto, R.R., Powers, L., Gaston, J., Sherwood, R.W., Johnston, L., Blair, J., Early, B., Hopkins, R.S., Scott, E., Smede, R., Osterud, H., Slifman, N., Shilke, J., Hills, L., Goodins, J.A., Burkhart, T.J., Jenkins, M.T., Witte, E.J., McCarthy, M., de Melfi, T., Erb, J., and E.C. Lippy. 1980. Waterborne giardiasis in California, Colorado, Oregon, Pennsylvania. Morbid. Mortal. Weekly Rep. 29:121123. 30. Kirner, J.C., Littler, J.D., and L.A. Angelo. 1978. A waterborne outbreak of giardiasis in Camas, Wash. J. Am. Waterworks Assoc. 70:3540. 31. Kulda, J., and E. Nohynkova. 1978. Flagellates of the human intestine and of intestines of other species. In: Parasitic Protozoa. J.P. Kreier (ed). Academic Press, New York, NY. pp. 69104. 32. Lavier, G. 1924. Deux especes de Giardia du rat d'egout Parisien. (Epimys norvegicus). Ann. Parasitol. Hum. Comp. 11:161168. 33. Lefebrve, P.A., and J.L. Rosenbaum. 1986. Regulation of the synthesis and assembly of ciliary and flagellar proteins during regeneration. Ann. Rev. Cell Biol. 2:515544. 34. Lin, S.D. 1985. Giardia lamblia and water supply. J. Am. Waterworks Assoc. 77:4047. 35. Lippy, E.C.. 1978. Tracing a giardiasis outbreak at Berlin, New Hampshire. J. Am. Waterworks Assoc. 70:512520. 36. Lopez, C.E., Juranek, D.D., Sinclair, S.P., and M.G. Schultz. 1978. Giardiasis in american travelers to Madeira Island, Portugal. Am. J. Trop. Med. Hyg. 27:11281132. 37. Meyer, E.A., and M. Radulescu. 1979. Giardia and giardiasis. Adv. Parasitol. 17:147. 38. Meyer, E.A., and E.L. Jarroll. 1980. Giardiasis: review and commentary. J. Epidemiology 111:112. 39. Moore, G.T., Cross, W.M., McGuire, D., Mollohan, C.S., Gleason, N., Healy, G.R., and L.H. Newton. 1969. Epidemic giardiasis at a ski resort. New Eng. J. Med. 281:402407. 40. Morbidity and Mortality Weekly Report, Center for Disease Control. 1978. Intestinal parasite surveillance United States 1976. 27:167168. 41. Mueller, J.C., Jones, A.L., and L.L. Brandborg. 1974. Scanning electron microscope observations in human giardiasis. In: Scanning Electron Microscopy. O. Johari (ed). IIT Research Institute, Chicago. pp. 557564. 42. Nash, T.E., McCutchan, T., Keister, D., Dame, J.B., Conrad, J.D., and F.D. Gillin. 1985. Restrictionendonuclease analysis of DNA from 15 Giardia isolates obtained from humans and animals. J. Infect. Dis. 152:6473. 43. Navin, T.R., Juranek, D.D., Ford, M., Minedew, D.J., Lippy, E.C. and R.A. Pollard. 1985. Case control study of waterborne Giardiasis in Reno, Nevada. Am. J. Epidemiol. 122:269275.
Page 236
44. Nentyl, J.A.. 1983. The Beaver. Crestwood House, Mankato, MN. pp. 147. 45. Owen, R.L., Nemanic, P.C., and D.P. Stevens. 1979. Ultrastructural observations on giardiasis in a murine model. I. Intestinal distribution, attachment, and relationship to the immune system of Giardia muris. Gastroenterology. 76:757769. 46. Owen, R.L. 1980. The ultrastructural basis of Giardia function. Trans. R. Soc. Trop. Med. Hyg. 74:429433. 47. Padchenko, J.K., and N.G. Stolyarchuk. 1970. Dogs as a spontaneous carrier of Lamblia and probable source and vector of lambliasis in nature. Vestnik Zoologiya 4:5661. 48. Panigraphy, B., Mathewson, J.J., Hall, C.F., and L.C. Grumbles. 1981. Unusual disease conditions in pet and aviary birds. J. Am. Vet. Med. Assoc. 178:394 395. 49. Peterson, H. 1972. Giardiasis (lambliasis). Scand. J. Gastroenterology 7 (supplement 14):144. 50. Pickering, L.K, Woodward, W.E., DuPont, H.L., and P. Sullivan. 1984. Occurrence of Giardia lamblia in children in day care centers. J. Pediatrics. 104:522526. 51. Poley, J.R., and S. Rosenfield. 1982. Malabsorption in giardiasis: Presence of a luminal barrier (mucoid pseudomembrane). A scanning and transmission electron microscope study. J. Pediatr. Gastroenterol. 1:6380. 52. Riggs, J.L., Dupuis, K.W., Nakamura, K., and D.P. Spath. 1983. Detection of Giardia lamblia by immunofluorescence. Appl. Environ. Micro. 45:698700. 53. Sauch, J.F. 1985. Use of immunofluorescence and phasecontrast microscopy for detection and identification of Giardia cysts in water samples. Appl. Environ. Micro. 50:14341438. 54. Scholtens, R.G., New, J.C., and S. Johnson. 1982. The nature and treatment of giardiasis in parakeets. J. Am. Vet. Med. Assoc. 180:170173. 55. Schupp, D.G., and S.L. Erlandsen. 1987. A new method to determine cyst viability: correlation between fluorescein diacetate/propidium iodide staining and animal infectivity. J. Parasitol. 53:704707. 56. Shaw, P.K., Brodsky, R.Z., Lyman, D.O., Wood, B.T., Hibler, C.P., Healy, G.R., McLeod, K.I.E., Stahl, W., and M.G. Schultz. 1977. A communitywide outbreak of giardiasis with evidence of transmission by a municipal water supply. Ann. Int. Med. 87:426432. 57. Sheffield, H. 1979. The ultrastructural aspects of Giardia. In: Waterborne Transmission of giardiasis. Jakubowski, W. and J.C. Hoff. (eds). U.S. Environmental Protection Agency, Cincinnati, OH. pp. 921. 58. Simon, C.E. 1922. A critique of the supposed rodent origin of human giardiasis. Am. J. Hyg. 1:406434. 59. Sorenson, S.K., Riggs, J.L., Dileanis, P.D., and T.J. Suk. 1986. Isolation and detection of Giardia cysts from water using direct immunofluorescence. Am. Water Res. Bull. 22:843845. 60. Spaulding, J.J., Pacha, R.E., and G.W. Clark. 1983. Quantitation of Giardia cysts by membrane filtration. J. Clin. Microbiol. 18:713715. 61. Sykora, J.L., Bancroft, W.D., States, S.J., Shapiro, M.A., Boutros, S. N., Keleti, G., Turzai, M., and L.F. Conley. 1987. Giardia cysts in raw and treated sewage. In: Controling Waterborne giardiasis. Chapter 3. G.S. Logsdon (ed). Environmental Engineering Division, American Society of Civil Engineers. pp. 2233. 62. Sykora, J.L., States, S.J., Bancroft, W.D., Boutros, S.N., Shapiro, M.A., and L.F. Conley. 1986. Monitoring of water and waste water for Giardia. Proc. of 1986 American Waterworks Association Water Quality Technology Conference, Portland, OR. pp. 10431054. 63. Takano, J., and J.H. Yardley. 1965. Jejunal lesions in patients with giardiasis and malabsorption: An electron microscopic study. Bull. Johns Hopkins Hosp. 116:413429. 64. Tsuchiya, H. 1931. A study on the variabilities in dimensional and numbers of discharged cysts in Giardia lamblia (Stiles, 1915) from day to day under normal conditions. Am. J. Hyg. 13:544567. 65. Tynan, T. 1985. Beavers are no laughing matter in Mass. New Hampshire Sunday News, Sunday, Dec. 29. p. 4a. 66. Weniger, B.G., Blaser, M.J., Gedrose, J., Lippy, E.C., and D.D. Juranek. 1983. An outbreak of waterborne giardiasis associated with heavy water runoff due to warm weather and volcanic ashfall. Am. J. Publ. Health 73:868872. 67. Wolfe, M.S. 1978. Current concepts in giardiasis. New Eng. J. Med. 298:319321. 68. Woo, P.T.K., and W.B. Paterson. 1986. Giardia lamblia in children in daycare centres in southern Ontario, Canada, and susceptibility of animals to G. lamblia. Roy. Soc. Trop. Med. Hyg. 80:5659. 69. Wright, R.A., Spencer, H.C., Brodsky, R.E., and T.M. Vernon. 1977. Giardiasis in Colorado. An epidemiology study. Am. J. Epidemiology. 105:330336.
Page 237
Analysis of Municipal Water Samples for Cysts of Giardia Charles P. Hibler CH Diagnostic & Consulting Service Inc., 2012 Derby Court, Fort Collins, Colorado 80526, U.S.A.. Analysis of 4423 water samples from 301 municipal sites in 28 states between 1979 and 1986 has shown that 102/301 (34%) sites sometimes were positive. The 102 sites accounted for 3633/4423 (82%) of these samples; the 199 negative sites sent 651 samples, or 3.27 samples/site. The remaining 139 samples were from unknown sites, unknown sources, or of unknown type. Raw samples include 512/1968 (26%) positive, finished samples include 267/2372 (11%) positive, and unknown type 15/83 (18%) positive. Positive samples (both raw and finished) were distributed: creeks 346/1218 (28%), rivers 212/828 (26%), lakes 193/1983 (10%), springs 16/84 (19%), and wells 2/63 (3%) (mixed sources excluded from total). Positive finished samples were distributed: unfiltered 80/1214 (6.6%), direct filtration 148/615 (24%), conventional treatment 12/357 (3.4%), commercial manufactured 4/33 (12%), slow sand 0/11, diatomaceous earth 0/7, mechanical 11/51 (21.6%) and infiltration gallery 7/37 (18.9%) (unknowns excluded from total).
Introduction More than 6500 water samples have been examined for cysts of Giardia the past 12 years by the author's laboratory. Samples have been analyzed from 28 of the contiguous 48 United States. A total of 4423 water samples from 301 sites (municipalities) are included in this report. Data from September, 1979 through September, 1986 have been included; data obtained prior to 1979 are not complete because thorough records were not kept during these years. Over 1000 water samples from foreign countries and research samples on pilot filtration systems have not been included. Materials and Methods Sampling and Diagnostic Techniques The modified reference method was used primarily for these analyses and the indirect immunofluorescent technique was occasionally used as a comparison. A description of the techniques, the modifications or improvements in those techniques and the associated problems, past and present, have been presented in another article in this book. Recovery of cysts was highly dependent upon the turbidity, the type of turbidity (organic or inorganic) and the quality of the cysts. Recovery of cysts was inversely proportional to the turbidity of the source. Cysts were effectively recovered if the turbidity was primarily inorganic and the concentrate from the filter cartridges was diluted to the extent that the centrifugate from a 25 mL aliquot was equal to or less than 0.25 mL in volume. If the turbidity was primarily organic, dilution of the concentrate together with use of a lower than optimum specific gravity of the chemical was usually necessary. If cysts were alive and in good condition, recovery from low inorganic turbidity sources was excellent. If cysts were alive and in good condition and the turbidity was primarily organic, recovery was poor. If cysts were dying and/or dead, recovery was very poor, irrespective of the type and amount of turbidity. If the source contained large numbers of algae, diatoms and free living flagellates, their presence interfered with visualization and quickly led to eye fatigue. If the sample had been chlorinated, even though inline dechlorination may have been applied to the sample, the cysts were usually inactivated. The time lag between sampling, shipping, processing and analysis of samples from chlorinated sources invariably resulted in decreased recovery. If an unfiltered source used presedimentation with alum, or if a filtration system used alum or polymers, and these chemicals were present in the sample, the samples were essentially impossible to analyze effectively. Alum and polymers would very effectively coagulate debris, cysts and chemical into an inseparable mass in the concentrate obtained from washing the filter cartridges. Assessment of Risk If a water sample was obtained from a municipality using some type of filtration, both a raw sample and a finished (filter plant effluent) sample were necessary to assess filter plant efficiency. The presence or absence of Giardia cysts was secondary to the ability of the filtration plant to remove material the size of or larger than Giardia cysts. A risk assessment of 0 to 100% removal was used to evaluate plant efficiency. Plant debris (the undigested fecal material from herbivorous animals), some species of algae and diatoms, coccidia (both fish and mammal), parasitic nematode eggs (usually beaver) and crustaceans or crustacean eggs were all excellent indicators. Plant debris was the best indicator because it is extremely light in weight and filtration plants having problems removing risk material invariably passed plant debris. The most common source of the plant debris was beaver or muskrat (muskrat and/or smaller rodents). Types of Giardia Cysts The majority of the Giardia cysts found in surface water sources were the G. duodenalis type. A few sources, especially large rivers receiving sewage effluent from large municipalities, had both G. duodenalis and G. muris. Necropsy of 28 blackcrowned herons, a migratory waterfowl, revealed 100% infection with a species of Giardia very similar morphologically to G. duodenalis. The cyst, however, is near spherical in shape. Likely other waterfowl also are infected. The ''bird type" Giardia cyst has been found in lakes, ponds and large, open springs throughout the United States. Results and Discussion All Samples The 4423 water samples originated from springs, creeks, rivers, lakes and wells. This includes 1968 raw water samples and 2372 finished water samples. A total of 83 were unknown type samples (Table 1, Figures 1, 1A). Eighteen percent (794/4423) of the samples from all sources were contaminated with cysts of Giardia, and 15/83 (18%) of the unknowns were contaminated. Of the 301 sites from which samples were received, 102/301 (34%) sites sent samples often contaminated with Giardia cysts; 3633/4423 (82%) of the total samples
Page 238 TABLE 1. Detection of Giardia cysts in water samples: classification by type of water*. Water type
# of samples examined
% of total samples
# of samples positive
% of samples positive
Raw
1968
44.5
512
26
Finished
2372
53.6
267
11
Unknown
83
1.9
15
18
Totals
4423
100
794
18
* Samples originated from 301 municipal sites in 28 states; 102 (34%) sites were positive for Giardia cysts. Negative samples were obtained from 199 sites, but ony 651 samples (average 3.27 samples/site) were examined.
originated from these 102 sites and 794/3633 (22%) were positive. Raw water samples consisted of 1968/4423 (44.5%) of the samples examined and 512/1968 (26%) were contaminated with cysts; 512/794 (64%) of the positive samples were raw water (Table 1). The 102 sites that had samples often contaminated with cysts sent 1655/1968 (84%) of these raw water samples for analysis and 512/1655 (31%) were contamined. Finished water samples consisted of 2372/4423 (54%) of the samples examined and 267/2372 (11%) were contaminated with cysts, 267/794 (34%) of the positive samples were finished water (Table 1). The 102 sites that had samples often contaminated with cysts sent 1978/2372 (83%) of these finished samples for analysis and 267/1978 (13.5%) were contaminated. The 199 negative sites from which samples were obtained sent 651 samples for analysis, an average of 3.27 samples/site. While positive samples frequently were obtained at a single sampling, often repeated sampling was necessary. With few exceptions, repeated sampling provided positive results. Some sites consistently had cysts present in the water while in others the contamination was sporadic. The
Figure 1. Composite of samples from all sources: 794/4423 (18%) positive, 19791986. Negative
number of cysts recovered varied considerably between samples from the same site and between different sites at the same source. On occasion, samples would be compared from the same site but taken at different times over a 24 hour period. If only a few cysts were present, recovery would vary from none to a number similar to the number found previously. If considerable numbers consistently were present, little difference was noted in recovery. Numbers varied from as low as 0.07 cysts/100 gallons of water sampled to as high as 1472 cysts/100 gallons of water sampled during epidemics of waterborne giardiasis. In those few instances where samples were obtained during the course of an epidemic, 10 to 15 cysts/100 gallons of water sampled was not unusual. Sources contaminated with raw or treated sewage often had cysts too numerous to count effectively. The data from all sites indicate that Giardia cysts are more numerous in the late winter through early spring months than the other months of the year (Figures 1, 1A). If, indeed, cysts are more numerous during these months, systems would be at greater risk because winter temperatures would necessitate use of more chlorine and/or time of contact to inactivate the cysts. However, a number of factors can influence these data: 1) turbidity is generally lower in all sources during the winter, increasing the efficiency of the techniques; 2) numbers of algae, diatoms, flagellates, pollen, etc. are considerably reduced, facilitating visualization of the cysts; 3) volume of water is often considerably reduced in creeks and rivers, essentially concentrating the cysts; and 4) cysts live longer in cold water. As in analysis of raw water, many factors influence cyst recovery in finished water. Many of the finished water samples (filtered or unfiltered) and all of the distribution samples had been exposed to chlorine and even though municipalities usually dechlorinated their samples with 1% sodium thiosulfate, either metered inline or subsequent to disconnecting the sampling cartridge, the results of these analyses obviously were compromised by the chlorine and the time lag between sampling, shipping, processing and analysis. Some of the finished water samples originated from sources using only presedimentation
Figure 1A. Composite samples from all sources: 794/4423 (18%) positive, 19791986.
Page 239 TABLE 2. Detections of Giardia cysts in water samples. Classification by source and type of water. Number of sites
Number of sites positive
Percent of sites positive
Number of samples
Number of samples positive
Percent of samples positive
Creeks (total)
75
38
51
1218
346
28
Raw
444
181
41
Finished
774
165
21
Rivers (total)
74
38
51
828
212
26
Raw
449
163
36
Finished
379
49
13
Lakes (total)
49
19
39
1983
193
10
Raw
829
138
17
Finished
1154
55
5
Springs (total) **
36
5
14
84
16
19
Wells (total)**
40
2
5
63
2
3
Source*
* Municipalities using a combination of sources (i.e. creek and spring) have been excluded from totals. ** Most water from springs and wells is unfiltered and may/may not be disinfected before consumption.
and the alum present made effective analysis essentially impossible. Filtered sources using chemical pretreatment with alum and/or polymers often passed these chemicals through the filtration system and these samples also were often impossible to analyze. Some municipalities initiated animal control (beaver and muskrat) on the source if a considerable number of cysts were consistently present. Frequently animal control was initiated even when the municipality had a superb filtration system. Generally the author recommended animal control around a lake or within the channel leading from a creek or river to the plant if cysts were numerous. More often than not the animal control program was then initiated without informing the laboratory. This resulted in a loss of potentially important data regarding the efficiency of the control program. Creeks The definition "creek" varies between geographic regions in the United States. A "river" in the west may be a "creek" in the east. A total of 1218 creek samples were analyzed and Giardia cysts were found in 346/1218 (28%) of the raw and finished samples (Figures
Figure 2. Composite of creeek samples: 346/1218 (28%) positive, 19791986. Negative
2, 2A) (Table 2). Creeks are subject to runoff and/or thunderstorm conditions much the same as rivers, but on a smaller scale. In general, creeks provide an excellent habitat for aquatic type mammals because of the smaller volume of water and the proximity of vegetation for food and cover. The presence of Giardia cysts in creek samples generally was the result of animal contamination because many creeks originated in pristine environments rarely, if ever, visited by humans, especially when they were under 810 feet of snow in the winter. Considerable numbers of beaver (51% infected) and muskrat (100% infected) living on creeks in Colorado have been examined. Small rodents (voles, mice) have not been found infected in Colorado, but they have been found infected in the state of Washington and in western Canada. About 20% of the cattle on creeks are infected, and about 5% of the domestic sheep (not accurate for sheep because sheep fecal matter dries quickly), but infection has not been found in freeranging horses, mule deer, bighorn sheep, etc. Captive moose, captive mule deer and pastured horses have been found infected. Infection has been found in elk in the state of Washington. Cattle, because
Figure 2A. Composite of samples from creeks: 346/1218 (28%) positive, 19791986.
Page 240
of their habit of frequenting streams can be a limited threat, generally functioning to spread infection between watersheds. Moose, because of their aquatic habits could also pose a potential threat. Creek samples, from their alpine origins to the plains, have been examined for contamination. Cysts have been found in samples from alpine habitats, but re examination usually showed a transient beaver trying to survive. Samples taken below the alpine, in timbered areas, frequently are contaminated. In creeks, as in rivers, lakes, etc. young animal populations begin moving after weaning activities in the summer to establish a home after the runoff. They also begin winter preparations in midsummer. As late as November many young are still trying to find a home and often will move directly into municipal intakes. Animal populations stabilize in the winter and then begin moving again in the spring, when they stabilize until weaning in the summer. Samples taken in beaver ponds or below beaver dams invariably results in recovery of cysts, but samples taken in the next pond below a beaver dam (still water) are usually negative. Samples taken at or below campsites on creeks frequently are contaminated with Giardia, however, these cysts could be either animal or human origin. If creeks (or rivers) traverse urban areas to reach municipal intakes, parasites of cats and dogs (especially ascarids and coccidia) are commonly found. Residents along the course of this water often toss fecal refuse into the water, effectively fouling the source. Some creek sources have been a dilemma for municipalities because careful examination has not revealed the animal(s) responsible for contamination. These watersheds are usually under 810 feet of snow most of the winter, yet cysts are always present in the raw water. Smaller municipalities generally used creeks as a source for municipal supply, as did summer camps, guest ranches or lodges, etc. Many types of filtration systems were employed by these camps and smaller communities, from mechanical pointofuse to conventional treatment. Finances available usually dictate the type of system, and the functional efficiency of the system.
Figure 3. Composite of river samples: 212/828 (26%) positive, 19791986. Negative
Figure 3A. Composite of river samples: 212/828 (26%) positive, 19791986.
Filtration systems found on creeks were usually more at risk for contamination with cysts, primarily because the municipalities could not afford to build and maintain systems that would effectively remove the cysts. Creeks generally had an abundant aquatic animal population and, more often than not, a high quality source of water that met or exceeded existing state or federal regulations. Rivers A total of 828 river samples were examined and 212 (26%) of the raw and finished water samples had cysts of Giardia (Table 2, Figures 3, 3A). Rivers in the west, northwest and southwest may be called creeks in the east; moreover, frequently an open channel or ditch from the main river source, often several miles long, supplies the plant. For all practical purposes this ditch or channel is a creek and not a river. The ditch or channel provides a habitat for aquatic animals much like a creek and, if it traverses urban areas, is subjected to the same fecal refuse from domestic animals as a creek; therefore distinguishing between creeks and rivers is difficult at best. Rivers in the west, northwest and southwest were subject to more severe runoff conditions than creeks and the effects of thunderstorms was compounded. Rivers in the above geographic areas generally (but not always) were contaminated with cysts from animal sources whereas in the larger rivers in the east often this contamination obviously was from sewage plants. Municipalities using river sources for water were usually larger, metropolitan type areas with the financial resources to construct plants capable of removing the cysts, especially in the east and along the plains areas in the west. The comments regarding animal populations, campsites, etc. relating to creeks also applies to rivers, especially in the western part of the United States. Unlike creeks, the turbidity of river water, during runoff or thunderstorms was much higher. While the percentage of contamination with cysts was about the same, likely the turbidity in rivers compromised these results. Lakes A total of 1983 lake samples were analyzed and 193 (10%) raw and finished water samples had Giardia cysts present (Table 2, Figures 4, 4A). Lakes varied from manmade to natural, some no larger than ponds
Page 241
Figure 4. Composite of lake samples: 193/1983 (10%) positive, 19791986. Negative
(also subject to interpretation in different geographic areas) to those covering several square miles. Lakes often were municipally owned and protected. In the western United States, lakes were situated near or above the limit of aquatic animal activity. While these timberline lakes sometimes have cysts present, contamination is usually sporadic. High mountain, cold water lakes generally were very clear, with little turbidity even after thunderstorm activity. These lakes did not support very much algae growth (or anything else). Lakes in the east and northeast varied from those that were very clear, with little organic turbidity to those with high levels of organic material and algae, flagellates, crustaceans, ciliates, amoeba, etc. The latter is a very healthy water source but diagnosis of Giardia cysts in samples from these lakes, especially in the summer and early fall was essentially impossible, irrespective of the techique used. Consequently Figures 4 and 4A, indicating that cysts are more common in the winter, are probably a reflection of the turbidity and other organisms interfering with effective diagnosis during the summer. Sometimes summer samples from lakes are almost impossible to process when an algal bloom is underway; the algae and myriads of flagellates quickly caused eye fatigue. Some lakes were very deep, others were shallow. Shallow lakes, when municipal demand is considerable, frequently developed a channel that functioned much like a creek or river; moreover, shallow lakes seemed to be more attractive to aquatic animals. Large lakes, with the influent at one end and the effluent at the other, and without any "shortcircuiting", seldom had cysts unless aquatic animals were living close to or frequenting the effluent area. Municipalities using a lake as a primary source often initiated animal control on the lake but not on the source supplying the lake. Generally this was a more common practice when the finished water was unfiltered. If the lake was deep this was effective; if the lake was shallow it was not effective. Giardia cysts settle very quickly in undisturbed water. In lakes with a rich organic bottom, cysts quite likely are consumed by fauna living on the bottom or become locked into the thick detritus and
Figure 4A. Composite of lake samples: 193/1983 (10%) positive, 19791986.
eventually decompose. Monzingo (personal communication) found that cysts settling to the bottom of beaver ponds, a rich organic situation, cannot be found a short period of time after settling. Springs A total of 84 spring sources were examined and 16 (19%) raw and finished samples were contaminated with Giardia; others were at risk even though cysts were not found (Table 2, Figures 5, 5A). Some were not true springs but natural infiltration galleries containing organisms that normally live on the surface and need sunlight to survive. Some spring sources were efficiently protected, others were haphazardly protected and some were unprotected. If the spring was open, Giardia cysts often were present. While users often chlorinated spring water, very seldom was any filtration used. Wells A total of 63 well samples were examined and 2 (3%) had cysts of Giardia present (Table 2). One well was essentially an infiltration gallery of the creek about 25 feet from the well. The creek near this well always has considerable Giardia, both from the animal populations and the sewage plant upstream. The other was a true well, 250 feet deep, that had been contaminated by priming with river water from a source that was heavily contamined with Giardia. Since cysts were found in the sample sent from the true well, this had to be recorded. Some of the well samples had surface type organisms present, such as crustaceans; therefore these systems were not true wells. The well systems at risk are not included because cysts were not found. Finished Water A total of 2372/4423 (54%) of the samples were finished water and 267/2372 (11%) of these were contaminated with cysts; 267/794 (34%) of the positive samples were finished water (Table 1). Finished water samples from the 102 sites that often had positive samples accounted for 1978/2372 (83%) of the finished water samples and 267/1978 (13.5%) of these were contaminated. Unfiltered Unfiltered, finished water samples came from 94 sites and 16/94 (17%) sites had samples often contaminated with cysts. A total of 1214/2372 (51%) of the finished water samples were unfiltered and 1046/1214 (86%) of the samples came from the 16 sites; 80/1046
Page 242
Figure 5. Composite of spring samples: 16/84 (19%) positive, 19791986. Negative
(7.7%) of these samples had cysts present (Table 3). Many of the unfiltered samples had been treated with chlorine; moreover, a considerable number of these finished, unfiltered samples were taken within the distribution system. The data presented were compromised because of the chlorine and the time lag. Most of the unfiltered samples originated from sites using lakes as a source for municipal supply, but a few samples came from sites using creeks, rivers and springs as sources. The creek and river sites often used presedimentation with chemicals as the only treatment prior to chlorination. Fewer positive raw water samples were obtained from lakes, especially deep lakes, than from creek or river sources and fewer cysts were found in these positive raw samples from lakes, no doubt a result of the "settling effect" in still water. Irrespective of the chlorine and the time lag that probably compromised the results, 80/1046 (7.7%) or even 80/1214 (6.6%) positive indicates that a lake or a reservoir is a good barrier for unfiltered systems providing adequate chlorine and/or time is available or can be built into the system.
Figure 5A. Composite of spring samples: 16/84 (19%) positive, 19791986.
Direct Filtration Of the 2372 finished water samples analyzed, 615/2372 (26%) were direct filtration (Table 3). A total of 92 sites used direct filtration, either throughout the year or in the winter months and 17/92 (18.5%) of these sites often had cysts in the filtered water. A total of 336/615 (55%) of the samples came from these 17 sites and 148/336 (44%) had cysts present. Usually cysts were present in those systems that did not use chemical pretreatment (Figure 7) rather than in systems using chemical pretreatment (Figure 6). Many sites switch from conventional treatment to direct filtration when turbidities decrease and stabilize near 1 NTU in the winter months; others use direct filtration all year, especially smaller sites. Some sites used chemical pretreatment with alum or polymers before filtration; in others the system functioned as a screen, generally because the water met (and frequently exceeded) existing regulations. Several of those sites not using chemical pretreatment began to use chemicals when analysis revealed passage of cysts and/or risk material. TABLE 3. Detections of Giardia cysts in finished drinking water supplies. Classification
Number of sites
Number of sites positive
Percent of sites positive
Number of samples
Unfiltered, chlorinated
94
Direct filtration*
92
Conventional treatment
86
Commercial and/or Pressure filters
12
Slow sand and Diatomaceous earth
Number of samples positive
Percent of samples positive
16
17
1214
80
6.6
17
18.5
615
148
24.0
5
5.8
357
12
3.4
2
16.7
33
4
12.1
3
0
0
18
0
0
Mechanical or Cartridge Filters
13
7
53.8
51
11
21.6
Infiltration gallery
16
5
31.3
37
7
18.9
Filter type unknown**
24
6
25.0
132
15
11.4
* Most of the positive samples originated from systems that were not using coagulation prior to filtration. ** Many of these samples are also included in the 83 samples (Table 1) where the type of sample (raw of finished) is also unknown.
Page 243
Figure 6. Composite of direct filtration systems, 19791986 (Chemical pretreatment).
Data used in Figures 6 and 7 was taken from direct filtration systems, generally in the fall and winter months. Most of these originated from sites with surface water temperatures of less than 1 to near 5°C and turbidities of less than 0.5 to near 1.0 NTU. Each data point of Figures 6 and 7 represents a minimum of 4 and a maximum of 8 raw and finished samples. Data are expressed as the percentage of turbidity reduction and the percentage of cysts removed in the filtration process. Figure 7 reveals the problems associated with removal of Giardia cysts from cold water, low turbidity sources in actual filtration plant situations. Figure 6 is an excellent demonstration of the increased efficiency when alum and/or polymers is added. Surprisingly, removal of 8085% of the incoming turbidity effectively removed the cysts if chemicals were added while removal of 8095% of the turbidity did not remove cysts when no chemicals were added. Data in Figures 6 and 7 was developed from filtration plants with cysts in the raw water; a number of plants using direct filtration without chemical pretreatment are not included because no cysts were found in the raw water; however, analysis for the risk material indicates most of these systems were at risk; Giardia cysts simply were not present at the time of sampling. Addition of chemicals in effective combinations and/or amounts was difficult when water conditions were fluctuating. For example, municipalities switching to direct filtration when temperatures decreased in the winter, and before switching to conventional treatment in the spring when temperatures increased, were often at risk for several days or even weeks because temperature and pH were widely fluctuating. Conventional Treatment Of the 2372 finished water samples examined, 357/2372 (15%) came from sites using conventional treatment. A total of 86 sites used conventional treatment at least part of the year and 5/86 (5.8%) of these sites sometimes had cysts present in the finished water (Table 3). A total of 75/357 (21%) of the samples
Figure 7. Composite of direct filtration systems, 19791986 (No Chemical pretreatment).
came from these 5 sites and 12/75 (16%) had cysts in the sample. Samples obtained from 3 of these 5 sites appeared to be from direct filtration without chemical pretreatment because nothing was being removed from the raw water; therefore these three samples were not included in Figure 8. Conventional treatment data indicate that this procedure is superb for removing Giardia cysts and all particulates the size of or larger than Giardia (Figure 8). Generally we attempted to discourage municipalities from sending samples because they generally were wasting their money and our time. Commercial and/or Pressure Filters Of the 2372 finished water samples examined, 33/2372 (1.5%) came from commercially manufactured filtration systems. Often it was difficult to determine if the system was a pressure filter or a small commercial filter designed similar to a direct filtration system because it was usually listed by company name or simply as "commercial", "readybuilt", "metalbox", etc. A total of 4/33 (12%) of these filters were passing cysts (Table 3). Obviously 33 samples is an inadequate number to use for any conclusions regarding their efficiency; however most of the samples analyzed from these systems indicated that the systems were at risk because they were not removing particulates the size of or larger than the cysts. Slow Sand and Diatomaceous Earth Only 11 samples from slow sand filters (one site) and 7 samples from diatomaceous earth filters (two sites) were examined for Giardia cysts and while this is too small a number to provide a high degree of confidence, both types of systems performed superbly (Table 3). One of the diatomaceous earth systems was in place on a guest ranch using river water from a source that was always contaminated with cysts. The slow sand samples also came from a source that usually has cysts in the raw water. Cartridge or Mechanical Filters Several types of mechanical filters were employed and 11 of the 51
Page 244
Figure 8. Composite of conventional treatment samples, 19791986.
(21.6%) samples examined had Giardia cysts (Table 3). Some of these filters removed all particulates the size of or larger than Giardia, others did not function as intended by the manufacturer. More often than not, the problem was not with the filter, or the cartridges, but the operators and the sales personnel. Sometimes the cartridges were not 1 micron, material in the effluent indicated about a 10 micron porosity. Other times the operator "stacked" 9.75 inch wound filters in drum or kettle type housings without using spacers to seal individual cartridges (spacers often are not needed for pleated type cartridges, only the wound). Sometimes the lid or cover was not sufficiently tight to make a good seal, or gaskets needed replacement and/or alignment. Infiltration Galleries Some of the infiltration galleries were superb, others were not and 7 of the 37 (18.9%) samples had Giardia cysts (Table 3). Most of these were used as the only means of filtration, others were in place ahead of either mechanical filters, pressure, or commercial filters and one was the initial screening mechanism before conventional treatment. Most of them appeared to have had similar construction and used a similar distribution of sand, gravel and rocks for filtration. Filter Type Unknown Only 15 of the 132 (11.4%) samples where the type of filtration was not given were infected with cysts of Giardia (Table 3). Samples of this nature are not valuable for analysis when important information has not been included. However, they represent only 3% of the total samples examined. Conclusions A number of limiting factors must be considered in the interpretation of these results: 1) efficiency of the diagnostic techniques used over the last 12 years probably is the most important of these factors because results are only as good as the technique employed and the techniques, despite improvements are only about 50% effective; 2) turbidity and type of turbidity; 3) time of year; 4) time of day samples were taken (some animals are active only at night); 5) handling, shipping the samples, the time factor; 6) eye fatigue, a direct result of the above factors and 7) quality of cysts in the sample. Giardia cysts probably are present in most if not all of the surface water sources in the United States. Repeated sampling of negative sources usually provided positive results. A high percentage of the negative sources sent only 12 samples. Giardia cysts are not uniformly distributed in surface waters and 1 or 2 samples from a site is inadequate. The source of surface water contamination sometimes is difficult to determine, at other times the source is clearly human sewage or clearly wild or domestic animal; unfortunately, in several outbreaks wild animals (beaver) were incriminated and found guilty without a trial. Investigators seem reluctant to place any blame on sewage. Man and many wild and domestic animals frequently are infected with the Giardia duodenalis type parasite. The animals responsible for most of the cysts found in surface water are humans, especially indirectly through sewage, beaver, muskrat, other large and small rodents, cattle, and moose. This laboratory and other investigators (Erlandson, personal communication) have found infection in a high percentage of herons. The risk to humans is unknown but to the analyst, it is a cyst and it is in water. While it has been found in deer, elk, horses, dogs, cats, coyotes, wolves, etc., they likely do not constitute a significant threat because of their habits and habitats. Watershed management generally is practiced by municipalities when they own or control the reservoirs; however, many sites obtain water from publicowned sources (e.g., U.S. Forest Service) that practice the multipleuse concept (agriculture, forestry and recreation). Management of these watersheds for municipal water supplies is not always consistent with the policy of the agency in control. Grazing permits, timber management, fish and wildlife management and recreational use (campsites, water skiing, etc.) may indirectly affect and prohibit effective management for potential giardiasis problems. If the creek and/or river traverses privately owned (agricultural) land, and the owner has rights to a certain percentage of that water, management is at the owner's discretion. The number of cases of giardiasis documented in backpackers, hikers, campers, etc. who drink raw surface water does indeed point out that wild and/or domestic animals can be a source of infection, although the limited facilities available for personal hygiene should make the camp cook suspect in some of these instances! Wild and domestic animals are susceptible to infection with human source cysts. Likely the reverse is true in many but not all instances. Too many crosstransmission trials have failed when attempts have been made to further crosstransmit the organism. Crosstransmission results are not always reliable and should be interpreted with caution. In some outbreaks wild animals definitely are to blame, in others there is sufficient reason for doubt. We have been directly or indirectly involved in most of the
Page 245
outbreaks in the United States, and a few of those in Canada. Some of these have been publicized and documented, others have not. The author is not at liberty to provide details of the unreported outbreaks only to point out the results of his investigations. In one community outbreak, the surface water source originated in a well protected, pristine environment. This outbreak was written and published but the epidemiologists did not make an effort to investigate the watershed. This laboratory did investigate after the outbreak; the only animals present on the watershed were beavers and muskrats and most of those were infected and shedding cysts. No septic system or other possible human contamination was observed. Cysts in the finished (filtered) water were almost too numerous to count over a period of several weeks. A rather large guest ranch in the western United States had approximately 100 cases over a period of two weeks. A creek originating in the alpine was the source of water. The water went through an infiltration gallery and a commercial filter (neither of which removed any of the cysts). This creek was thoroughly investigated: the only inhabitants were beavers and all were infected. Beavers were removed, the cyst count dropped to zero, and the outbreak subsided. The infiltration gallery and the filter are no more effective now than during the outbreak. While these two outbreaks clearly were of animal origin, in most of the outbreaks in which we have been directly or indirectly involved the source, human or animal, was never determined. Sewage was always a good possibility. The presence of Giardia cysts in a finished water sample does not necessarily indicate that an outbreak is imminent. Many of the sites from which this laboratory has received samples the past 12 years had as many if not more cysts in the finished water than sites in the midst of an outbreak; moreover, the chlorine and/or contact time was minimal (often none!) yet no cases of giardiasis other than the usual background cases were present. These cysts were alive: frequently they were introduced into mongolian gerbils (Meriones unguiculatus) and caused infection. In some of the smaller communities all of the natives probably were immune, but for other communities the indication is that these cysts were not infectious for humans. Several of these sites have been investigated and the animals present invariably were beavers and muskrats, usually living peacefully at the municipal intakes. Obviously there are strains(?) of G. duodenalis that do not readily crossinfect between animals and humans. Unfortunately, there is no means of determining which sources will and which sources will not infect humans. Acknowledgements I should like to acknowledge Mr. Walter Jakubowski, USEPA, Cincinnati, Ohio and Dr. Martin J. Allen, American Water Works Association Research Foundation, Denver, Colorado for the unrelenting but always diplomatic pressure they applied to convince me to compile these data without compromising the municipal sites, our customers who sent these samples for analysis. This study was funded by Mr. Stig Regli, USEPA, Office of Drinking Water, Washington, D.C. To my many graduate students, who often spent evenings and weekends at the microscope analyzing these samples, you know how much I appreciate your efforts. Over the years these samples have been processed and analyzed by: Dr. Robert Davies, Mr. Steve Henry, Dr. Don Monzingo, Dr. John Wegrzyn and Ms. Donna Howell. Ms. Carrie Hancock, Research Associate, and Ms. Diane Swabby, Laboratory Coordinator, have been and continue to be an integral part of the staff.
Page 247
VIABILITY TESTING
Page 249
A Review of Methods that Are Used to Determine Giardia Cyst Viability Frank W. Schaefer, III. Health Effects Research Laboratory, United States Environmental Protection Agency, 26 West M.L. King Drive, U.S.A.. Over the past 55 years a number of methods have been described for excysting Giardia cysts as a means of determining viability. The excystation methods for G. muris cysts are reliable and reproducible. However, methods published to date for the excystation of G. lamblia cysts, the human pathogen, have not yielded reliable, reproducible results. For both Giardia cyst types the stimulatory factors promoting excystation are low pH, carbon dioxide, a temperature around 37°C, and a final neutralizing step at pH 7.0. These excystation promoters induce numerous ectoplasmic vacuoles to dump their contents between the trophozoite and the cyst wall. Unfortunately, large numbers of cysts are required for all the excystation methods. Recently, other methods have been developed for determining Giardia cyst viability. These include several fluorescent vital dyes, differential interference contrast microscopy, and Mongolian gerbil infectivity tests. The fluorescent vital dye and differential interference contrast microscopy methods are new and may require validation. The animal infectivity method has several disadvantages. Not all G. lamblia cyst isolates from humans will infect gerbils. Cyst isolates which do infect do not necessarily produce cysts. This requires animal necropsy to verify the infection. Also, the lower the G. lamblia cyst dosage, the lower the probability is of infection being produced. Clearly, there is a great need for further study in this area of Giardia research.
Introduction Giardia species are flagellated protozoan parasites found in the intestinal tract of many vertebrates. Two forms of the parasite are usually found within the host's intestinal tract. The trophozoite, which is an active, vegetative form, lives in the upper third of the small intestine. As trophozoites get caught in the luminal flow of nutrients, they are swept down the intestinal tract and are induced to transform into cysts. The cyst is a dormant, transmission form of the parasite. Transmission is direct by the fecal oral route. The mature cyst is a rounded trophozoite, which is surrounded by a cyst wall, and has completed karyokinesis, but usually not cytokinesis. The transformation from trophozoite to cyst is usually complete by the time the organism reaches the distal portion of the small intestine. Besides transmission by contaminated food (2,28) and person to person contact (6,22,27), transmission can occur by ingestion of contaminated water and has been responsible for numerous waterborne outbreaks in the United States (9,24). Determining trophozoite viability generally is not a problem, because flagellar and caudal movement can be detected by conventional microscopy. Cysts, however, do not exhibit rapid movement unless they are induced to excyst. Since cysts are the usual infective form, it is especially important to know whether cysts are viable or not. Viability information on a cyst population is necessary for cyst disinfection studies and in determining risk assessments for contaminated drinking water supplies. Excystation procedures have been instrumental in the successful completion of Giardia cyst disinfection studies (16,20,37). To date the procedures for determining Giardia cyst viability have included in vivo excystation, in vivo infectivity, in vitro excystation, vital stains, and differential interference contrast microscopy. The following discussion will address various aspects of these viability testing procedures. A more detailed discussion of excystation may be found in the review by Meyer and Schaefer (26). Discussion Excystation Hegner (13,14,15) described in vivo excystation of human source Giardia cysts that had been injected into the stomach of, and later removed from, rats. Hegner noted that excystation occurs in the jejunum and ileum and speculated that the factors responsible for excystation include a temperature of 37°C and moisture. He also suggested (15) that host digestive juices are unnecessary for excystation. Armaghan (1) attempted to determine the site of excystation by placing G. muris cyst suspensions in various parts of the rat gastrointestinal tract (stomach, duodenum, ileum, and cecum) by laparotomy followed by injection to determine the site of excystation. Excystation was considered positive if cysts could be isolated from the fecal material and/or trophozoites could be isolated from the gastrointestinal tract post exposure. Infections were detected only in those rats injected either in the stomach or duodenum. The development of in vitro excystation made the application of quantitative procedures for assessing survival of Giardia cysts exposed to chemical factors easier. Excystation has been quantified two ways depending primarily on whether G. lamblia or G. muris cysts were being excysted. Bingham et al. (4) and Rice and Schaefer (30) determined the percentage excystation of G. lamblia
Page 250
cysts by counting the number of intact cysts (IC), partially excysted trophozoites (PET), and totally excysted trophozoites (TET) and applying the following formula: % excystation = (TET/2 + PET) + (TET/2 + PET + IC) × 100. The number of totally excysted trophozoites was divided by 2 in this formula, because each excysted organism yielded a pair of trophozoites. Empty cyst walls were not counted, because they were difficult to detect. Each excystation percentage derived by Bingham et al. (4) involved 3 to 5 separate counts in which an average of 708 cysts were counted to obtain each value. Feely (11) has used this procedure for G. muris excystation. On the other hand, Schaefer et al. (34) counted full and empty G. muris cyst walls rather than totally excysted trophozoites, because empty cyst walls were easier to count than swimming trophozoites. They calculated percentage excystation using the following equation: % excystation = (ECW + PET) ÷ (ECW + PET + IC) × 100, where ECW is the number of empty cyst walls, PET is the number of partially excysted trophozoites, and IC is the number of intact cysts. Each excystation percentage determination involved a total of 100 cyst types. These workers determined the thermal death point of G. muris cysts using this procedure. At exposure temperatures of 50°C and 52°C no empty cyst walls were found on counting 100 cysts. If, however, the entire slide which held approximately 100,000 cysts was scanned for mobile trophozoites, some were found at both of these temperatures. This indicated that counting only 100 cysts did not sample a large enough portion of the population to distinguish percentages of less than 1. In determining percentage excystation by either quantitative procedure, there is no universally accepted minimum number of cysts that must be counted to insure precision and accuracy of the result. Further research on this point is needed. Bingham and Meyer (5) reported the first successful in vitro excystation of Giardia cysts from humans, monkeys, and dogs, and G. muris cysts from rats and mice. Their procedure, which was carried out at 37°C, consists of three steps: a low pH induction step for 1 hour, a wash step to remove the acid, and an incubation step for 1 hour in a nutrient medium to complete the excystation process. Synthetic gastric juice (pepsin and hydrochloric acid), aqueous acids, and water were tried as induction solutions. For incubation media, they used water and HSP3 (Hanks' balanced salt solution, phytone, serum, and NCTC135; 25). Excystation occurred when the induction solution was synthetic gastric juice or aqueous acid, and the incubation medium was HSP3. Salts and pepsin in the synthetic gastric juice did not significantly alter the excystation level indicating that the acid alone was critical to the process. Although a number of inorganic acids were tested in induction solutions and were found effective, hydrochloric acid was selected for routine use. Peak excystation (2226%) occurred when the pH was between 1.3 and 2.7 in the induction solution and the incubation medium was HSP3. Bingham et al. (4) studied other factors in the excystation process of Giardia cysts isolated from a single human donor. Of these factors, time, temperature, pH, and incubation medium were shown to affect excystation levels. The induction solution exposure time decreased as the pH decreased. A hydrochloric acid induction solution at pH 2.0 only required an exposure time between 10 and 30 minutes. The highest excystation levels always occurred when both the induction solution and the incubation medium were at 37°C. The pH of the HSP3 incubation medium was critical, too. No excystation occured when the HSP3 was in the 0.5 to 4.0 range; excystation started at pH 6.2 and was best at pH 6.8. The excystation of acidinduced cysts in various components of HSP3 medium and complete HSP3 medium was compared (A.K. Bingham, M.S. thesis, Oregon Health Sciences University, Portland, 1979). Significantly lower excystation levels occurred in Hanks' phytone broth than in Hanks' phytone broth plus serum or complete HSP3 medium. The factors Bingham and Meyer (5) found favoring in vitro excystation closely approximated the host's in vivo environment in the stomach (induction step) and the duodenum (incubation step). They also reported that a period of cyst maturation varying from 2 to 7 days post isolation was required for G. lamblia cysts before maximal excystation could be achieved with their procedure. Rice and Schaefer (30) reported their experience with the BinghamMeyer procedure for excysting G. lamblia cysts isolated from human donors. They attempted to use the procedure for G. lamblia and G. muris cysts but usually obtained only 23% excystation. The RiceSchaefer procedure for G. lamblia cysts includes induction, wash, and incubation steps like the BinghamMeyer procedure. The induction step involves the addition of hydrochloric acid (pH 2), Hanks' balanced salt solution supplemented with Lcysteine hydrochloride and glutathione, and sodium bicarbonate to the cyst preparation in that order. Immediately thereafter, the tube containing the cysts is capped tightly, mixed, and then incubated for 30 minutes at 37°C. The cysts are washed by centrifugation in trypsinTyrode's incubation medium and resuspended in trypsinTyrode's incubation medium which has the pH adjusted to 8.0 with sodium bicarbonate. The resultant suspension is incubated 1 hour at 37°C. In 1981, Rice and Schaefer (30) reported the results of this procedure used in 28 G. lamblia excystation trials. Ten trials used cysts from two symptomatic donors, while 18 trials employed cysts from three asymptomatic donors. The lowest observed percent excystation (40%) occurred with cysts from an asymptomatic donor; the highest percentage (95%) occurred with cysts from both symptomatic and asymptomatic donors. The mean percentage excystation from the symptomatic donors (87%) was higher than that from the asymptomatic donors (70%). The lowest percentage excystation Rice and Schaefer observed was 10% higher than the maximum reported by Jarroll et al. (19) using the BinghamMeyer procedure.
Page 251
These data indicate that the RiceSchaefer method may be successfully used to excyst G. lamblia cysts from both asymptomatic and symptomatic donors, and that higher percentages of excystation may be expected from symptomatic donors. They also confirmed Bingham et al.'s (4) observation that G. lamblia cysts exhibited higher excystation rates after 7 days maturation. Schaefer et al. (34) reported the results of excystation experiments employing G. muris cysts isolated from mouse feces. Their procedure resulted in excystation levels consistently greater than 90% as compared to levels of less than 5% with the BinghamMeyer procedure. All of the G. muris excystation procedures consist of the same steps: induction, wash, and incubation. The main difference in the Schaefer et al. procedure was that induction was promoted by exposure of G. muris cysts to an induction solution consisting of 1 part reducing solution (Hanks' balanced salt solution supplemented with glutathione and 1cysteine hydrochloride; 25) and 1 part 0.1 M sodium bicarbonate for 30 minutes at 37°C in a sealed test tube. In this case, no inorganic acid is used. When the induction components are mixed, carbon dioxide is evolved, the pH is 2, and the oxidationreduction potential is 120 mV. A gradual decline in excystation to 2% or less occurred as the pH and the oxidation reduction potential were changed to 7 and 57 mV, respectively. The induction step was followed by a wash and incubation step both carried out in trypsinTyrode's solution (30). The trypsinTyrode's solution is reported by these investigators to be crude and tedious to make. The trypsin they used was an impure lyophilized extract of hog pancreas which does not completely dissolve after vigorous mixing in Tyrode's solution for 30 minutes at ambient temperature. The undissolved trypsin was always removed by high speed centrifugation followed by positive pressure filtration. When the trypsinTyrode's solution was boiled, G. muris excystations declined to the 70th percentile, indicating the enzymatic activity of the solution is not crucial and confirming Hegner's (15) and Bingham and Meyer's (5) observations that digestive enzymes are not obligatory. Attempts by Schaefer et al. to make trypsinTyrode's solution with purified pancreatin resulted in little if any G. muris excystation. Similar results were obtained by Marchin (personal communication, 1982) when he used purified trypsin in his trypsinTyrode's solution. Only when he substituted crude trypsin for purified trypsin was he able to obtain G. muris excystations in the 90th percentile. These data imply nutrients were supplied by the complex trypsinTyrode's incubation medium that were not supplied by the highly purified solutions. Feely (11) recently published a procedure which he used to excyst G. muris cysts. Induction of the cysts was done in Hanks' balanced salt solution at pH 2 for 30 minutes at 37°C. This procedure specifies that the Hanks' balanced salt solution must be prepared fresh daily and the pH adjusted to neutrality with sodium bicarbonate before being acidified to pH 2 with 2 N hydrochloric acid. This allows slow liberation of carbon dioxide from the induction medium during the incubation step like in the Rice and Schaefer (30) and Schaefer et al. (34) methods. The wash is in Hanks' balanced salt solution at pH 7.2 followed by incubation in TYI medium (salts, trypticase, yeast extract, glucose, cysteine hydrochloride, ascorbic acid, ferric ammonium citrate, bovine bile, and serum; 21) at 37°C for 30 minutes. No exogenous enzymes are required in this procedure. The excystation rates reported for this procedure were always greater than 90%. The advantages reported for this procedure are that it eliminates the tediously prepared trypsinTyrode's incubation medium preparation and requires no reducing agents. In these proceedings, Sauch (33) reports procedures for excysting G. lamblia and G. muris cysts in proteinfree media. Her induction solution for G. lamblia cysts was the same as Rice and Schaefer's (30) and for G. muris cysts was the same as Schaefer et al.'s (34); however, exposure to the induction solution at 37°C was increased from 30 to 45 minutes in the case of G. lamblia. In her procedures, trypsinTyrode's solution was replaced with Hanks' balanced salt solution supplemented with cysteine, sodium bicarbonate, and either proteose peptone for G. muris cysts or phytone peptone for G. lamblia cysts. Comparison of her methods with that of Rice and Schaefer for G. lamblia cysts and that of Schaefer et al. for G. muris cysts showed no significant differences in percent excystation. The advantage reported for this procedure is that excystation can be completed in the absence of enzymes which may destroy antigens on cysts. Excystation of G. lamblia and G. muris cysts is an active process on the part of the parasite (5,7,8,34). The caudal flagella and distal ends of the other flagella protrude from one end of the cyst. Flagellar movement starts slowly but rapidly increases within the first five to ten minutes of their emergence. The rapid flagellar movement seems to help pull and/or break the trophozoite out of the cyst wall. Scanning electron microscopy of this process has shown a tear in the cyst wall of G. muris cysts from the polar opening toward the opposite cyst pole (7). This opening appeared to be subsequently enlarged, presumably by flagellar action. Transmission electron microscopy of induced G. muris cysts by Coggins and Schaefer (8) has shown a cytoplasm devoid of endoplasmic reticulum, golgi bodies, and mitochondria. In uninduced cysts, large membrane bound ectoplasmic vacuoles are seen. After induction these ectoplasmic vacuoles appear to dump their contents into the peritrophic space. Speculation (8) that the contents of these vacuoles are enzymatic is supported indirectly by current information concerning the presence of lysosomal enzymes (10,23) along with evidence for mucus material secreted during excystation (5,7). More conclusive evidence was recently reported by Feely and Dyer (12) who demonstrated acid phosphatase activity in the ectoplasmic (peripheral) vacuoles of G. lamblia and G. muris trophozoites by cytochemical light and electron microscopy. Newly emerged trophozoites appear disorganized, oval in shape, and associated with mucoidlike material. Quadrinucleate trophozoites quickly begin to flatten, elongate, and undergo cytokinesis within 30 minutes of emergence to form binucleate trophozoites with completely organized adhesive disks. The characteristic ectoplasmic vacuoles
Page 252
documented in mature Giardia cysts and trophozoites still are rare in excysted trophozoites undergoing cytokinesis and in daughter trophozoites 30 minutes into the incubation period. These observations are consistent with the suggestion that these ectoplasmic vacuoles play a role in excystation. A number of conclusions may be drawn from the excystation studies done to date on G. lamblia and G. muris cysts: 1) the temperature of the induction and incubation reagents should be 37°C; 2) the induction medium should have a pH around 2 and evolve carbondioxide; 3) the wash step should help neutralize the acidity from the induction step; 4) the incubation medium should contain nutrients but need not contain serum, bile, or trypsin; 5) presently, no obligatory need for exogenous enzymes in the excystation process has been demonstrated; 6) a cyst maturation period is not needed for G. muris cysts; 7) a cyst maturation period for G. lamblia excystation is needed; 8) G. muris cysts can be routinely excysted with efficiencies in the 90th percentile; 9) G. lamblia excystation efficiencies are erratic and usually much less than 90%; and 10) excystation is a process requiring active participation by the parasite. Other Methods Giardia cyst densities in surface waters have not been reported to any great extent. One study which gathered data from the states of Oregon, Idaho, Wyoming, and Pennsylvania reported densities ranging from 0.05 to 680 Giardia cysts per 380 liters (Craun, G.F. and W. Jakubowski, Water Resour. Bull., in press). This indicates that Giardia cyst densities are usually lower than a thousand cysts per 380 liters of sample. Even if all the Giardia cysts could be recovered efficiently from the sample, which they cannot, there would be insufficient cyst numbers to determine viability by current excystation procedures which require between 1 and 5 × 105 cysts. In addition, the procedures for excysting Giardia cysts require several hours to complete and are not equally successfully in every laboratory. These problems have forced the development of other viability methods which can be used with small numbers of cysts. Among these alternate methods are vital dyes, fluorogenic dyes, differential interference contrast microscopy, and animal infectivity procedures. Bingham et al (4) compared the BinghamMeyer (5) procedure for excysting G. lamblia cysts with eosin dye exclusion and found that there was no correlation between the two techniques. The eosin dye exclusion technique always indicated greater viability in the cyst population than excystation did. This could have been for various reasons. The excystation technique may not have been optimal. It is also possible that there were cysts in the population that were alive but could not excyst. Furthermore, the cyst suspension could have contained dead cysts which excluded the dye. In these proceedings, Hudson et al. (18) report fluorescent dye exclusion as a method for determining Giardia cyst viability. In this procedure, as in the eosin procedure, live cysts exclude the dye. These workers utilized FluoraBora I (3(dansylamido)phenylboronic acid) in G. muris and G. lamblia cyst suspensions and compared the results with excystation. They found a high degree of correlation between the dye exclusion and excystation methods in the case of G. muris cysts. In the case of G. lamblia cysts, however, the correlation was low. This again suggests that viability as determined by excystation of G. lamblia cysts is being either overestimated or underestimated. Schupp and Erlandsen (Abstr. Annu. Meet. Am. Soc. Parasitol., 1986, 16, p. 35) have developed a method of determining G. muris cyst viability which employs fluorescein diacetate and propidium iodide stains. Viable cysts fluoresce green due to uptake and metabolic conversion of the stain to fluorescein. Dead cysts fluoresce orange to bright red depending on the excitation wavelength, due to staining by propidium iodide. Their results with freshly isolated cysts revealed greater than 85 to 90% of the cyst population stained green and less than 10% of the cysts did not stain at all. Freshly isolated fluorescein diacetate positive cysts and cysts that did not stain with either dye were inoculated into separate groups of mice. Similarly, heatkilled propidium iodide stained cysts were inoculated into another group of mice. Infections occurred in those mice receiving either fluorescein diacetate stained or unstained cysts but did not occur in those mice receiving the propidium iodide stained cysts. Necropsies of the mice exposed to either fluorescein diacetate or propidium iodide stained cysts always revealed trophozoites in the fluorescein diacetate exposed mice but never in the mice receiving propidium iodide stained cysts. These investigators have said that fluorescein diacetate positive stained cysts are capable of excystation, but they did not provide details. These data demonstrate that fluorescein stained cysts were viable as determined both by staining and infectivity; and possibly by excystation as well. Rotman and Papermaster (32) reported that enzymatic hydrolysis of fluorogenic esters in tissue culture cells varied as much as 80 fluorescein fluorimeter units. Some variability in the fluorescence of fluorescein diacetate stained G. muris cysts has been noted by Schupp and Erlandsen. In these proceedings, Schupp et al. (35) compared the light microscopic morphology of fluorescein diacetate stained G. muris cysts to propidium iodide stained G. muris cysts. Viable, fluorescein positive cysts had clearly defined cyst walls, a peritrophic space, and flagella at one pole. Occasionally axonemes were also observed, but no other cytoplasmic organelles were seen in these viable cysts. On the other hand, in nonviable, propidium iodide stained cysts the cytoplasmic organelles (axonemes, median bodies, portions of the adhesive disk, and nuclei) were seen. The peritrophic space in nonviable cysts was not observed, because the cytoplasm appeared to be closely adherent to the cyst wall. These differences were observed with both differential interference contrast and bright field microscopy. These observations imply that structural differences exist between these two types of G. muris cysts and may be a useful way to assess viability. Both the fluorescein diacetatepropidium iodide and the FluoroBora I viability staining methods are not
Page 253
selective staining methods. These methods are easily used with pure or highly concentrated Giardia cyst suspensions. However, in cyst suspensions containing other macro and microorganisms, the reaction of the Giardia cysts could easily be masked by the other living organisms. There is a need to link selective identification methods with these viability methods. Animal models of giardiasis have been developed in the mouse (31) and the Mongolian gerbil, Meriones unguiculatus (3). Both models have been successfully used to determine cyst viability and infectivity; however, both have several unsatisfactory aspects. Neither model manifests overt pathology like that seen in some humans. The mouse model produces G. muris cysts continuously for several weeks post exposure, unlike the human condition, which produces cysts intermittently. Although the Mongolian gerbil intermittently excretes cysts for several weeks post exposure, like humans do, not all human derived G. lamblia cysts will infect gerbils (Visvesvara, G.S., J. Dickerson, and G.R. Healy, Abstr. Ann. Meet. Am. Soc. Trop. Med. Hyg., 1983, K139, p. 117). Some human derived G. lamblia cysts infect the gerbil as evidenced by the production of trophozoites but not cysts. Between 1 and 15 cysts are necessary to infect a mouse with G. muris cysts (17) and as few as 10 cysts are required to infect man with G. lamblia cysts (29). In the case of Mongolian gerbils, the probability of infection is directly related to the dosage (Visvesvara, G.S., J. Dickerson, and G.R. Healy, Abstr. Annu. Meet. Am. Soc. Trop. Med. Hyg., 1983, K139, p. 117); cyst dosages should be greater than 100 cysts per gerbil to insure 70% of the gerbils become infected, when G. lamblia cysts are used. Similar results have recently been obtained when G. lamblia cysts from humans were given to beavers (Bemrick, W.J., S.L. Erlandsen, L.A. Kemp, L.F. Sherlock, and D.G. Schupp, Abstr. Ann. Meet. Am. Soc. Parasitol., 1986, 96, p. 55). Wallis and Wallis (36) recently successfully infected gerbils with Giardia cyst isolates from meadow voles, humans, dogs, and beavers. Both cysts and trophozoites were capable of infecting gerbils. Overall rates of infection were 89% for meadow voles, 46% for human isolates, 50% for dog isolates, and 91% for beaver isolates. Cyst dosages ranged between 4000 and 1.5 million cysts per animal. To date the Mongolian gerbil is the best experimental host for G. lamblia, but the system is subject to errors. Just because an infection did not take in the gerbil does not mean that the cysts used were not alive or infective to humans. The methods used to determine Giardia cyst viability included dye exclusion, fluorogenic dyes, excystation, and animal infectivity. There are disadvantages to each of these methods. Excystation requires large numbers of cysts and 2 to 3 hours to complete. G. lamblia cysts do not excyst routinely at high levels. Animal infectivity studies require weeks to complete. Furthermore, animals are expensive to buy, feed, and house. Also, only positive results can be conclusively used in animal infectivity studies. Eosin dye exclusion does not correlate well with excystation. Fluorogenc dyes seem to correlate well with G. muris excystation but not with G. lamblia excystation. Morphological disparity between living and dead cysts correlate well with infectivity in G. muris cysts. To date, however, this technique has not been reported for G. lamblia cysts. If differential interference contrast microscopy is used to differentiate living from dead cysts, a large investment is required for the optics. While fluorogenic dyes and differential interference contrast microscopy are promising techniques for determining cyst viability, they require further evaluation. Clearly, there is a great need for further study in this interesting area of Giardia research. Notice This document has been reviewed in accordance with U.S. Environmental Protection Agency policy and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. Literature Cited 1. Armaghan, V. 1937. Biological studies on the Giardia of rats. Am. J. Hyg. 26:236258. 2. Barnard, R.J. and G.J. Jackson. 1984. Giardia lamblia: the transfer of human infections by foods. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 365378. 3. Belosevic, M., Faubert, G.M., MacLean, J.D., Lau, C. and N.A. Croll. 1983. Giardia lamblia infections in Mongolian gerbils: an animal model. J. Inf. Dis. 147:222226. 4. Bingham, A.K., Jarroll, Jr., E.L., Meyer, E.A. and S. Radulescu. 1979. Giardia, sp: physical factors of excystation in vitro, and excystation vs. eosin exclusion as determinants of viability. Exp. Parasitol. 47:284291. 5. Bingham, A.K. and E.A. Meyer. 1979. Giardia excystation can be induced in vitro in acidic solutions. Nature (Lond.) 277:301302. 6. Black, R.E., Dykes, A.C., Sinclair, S.P. and J.G. Wells. 1977. Giardiasis in daycare centers: evidence of person to person transmission. Pediatrics 60:486491. 7. Coggins, J.R. and F.W. Schaefer, III. 1984. Giardia muris: scanning electron microscopy of in vitro excystation. Exp. Parasitol. 57:6267. 8. Coggins, J.R. and F.W. Schaefer, III. 1986. Giardia muris: ultrastructural analysis of in vitro excystation. Exp. Parasitol. 61:219228. 9. Craun, G.F. 1986. Waterborne giardiasis in the United States, 19651984. Lancet 2:513514. 10. Feely, D.E. 1982. Histochemical localization of acid phosphatase in Giardia. Anat. Rec. 202:54A. 11. Feely, D.E. 1986. A simplified method for in vitro excystation of Giardia muris. J. Parasitol. 72:474475. 12. Feely, D.E. and J.D. Dyer. 1987. Localization of acid phosphatase activity in Giardia lamblia and Giardia muris tropozoites. J. Protozool. 34:8083. 13. Hegner, R. 1927. Excystation and infection in the rat with Giardia lamblia from man. Am. J. Hyg. 7:433447. 14. Hegner, R. 1927. The viability of cysts of Giardia lamblia from man in the stomach of the rat. Am. J. Hyg. 7:782785. 15. Hegner, R. 1927. Excystation in vitro of human intestinal protozoa. Science (Wash. D.C.) 65:577578. 16. Hoff, J.C., Rice, E.W. and F.W. Schaefer, III. 1984. Disinfection and the control of waterborne giardiasis. In: Environmental Engineering. M. Pirbazari and J.S. Devinny (eds). American Society of Civil Engineers, New York. pp. 239244. 17. Hoff, J.C., Rice, E.W. and F.W. Schaefer, III. 1985. Comparison of animal infectivity and excystation as measures of Giardia muris cyst inactivation by chlorine. Appl. Environ.
Page 254
Microbiol. 50:11151117. 18. Hudson, S.J., Sauch, J.S. and D.G. Lindmark. 1987. Fluorescent dye exclusion as a method for determining Giardia cyst viability. In: Proceedings of the Calgary Giardia Conference. P. Wallis and B.R. Hammond (eds). University of Calgary, Calgary. pp. 265269. 19. Jarroll, E.L., Bingham, A.K. and E.A. Meyer. 1981. Effect of chlorine on cyst viability. Appl. Environ. Microbiol. 41:483487. 20. Jarroll, E.L., Hoff, J.C. and E.A. Meyer. 1984. Resistance of cysts to disinfection agents. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 311329. 21. Keister, D.B. 1983. Axenic culture of Giardia lamblia in TYIS33 medium supplemented with bile. Trans. Roy. Soc. Trop. Med. Hyg. 77:487488. 22. Keystone, J.S., Krajden, S. and M.R. Warren. 1978. Person to person transmission of Giardia lamblia in daycare nurseries. Can. Med. Assoc. J. 119:241 248. 23. Lindmark, D.G. 1980. Energy metabolism of the anaerobic protozoan Giardia lamblia. Mol. Biochem. Parasitol. 1:112. 24. Lippy, E.C. 1981. Waterborne giardiasis. In: The Proceedings of the International Conferrence on Lake Management and Restoration. Environmental Protection Agency. EPA 440/581010. 25. Meyer, E.A. 1976. Giardia lamblia: isolation and axenic cultivation. Exp. Parasitol. 39:101105. 26. Meyer, E.A. and F.W. Schaefer, III. 1984. Models for excystation. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York. pp. 131144. 27. Meyers, J.D, Kuharic, H.A. and K.K. Holmes. 1977. Giardia lamblia infection in homosexual men. Br. J. Vener. Dis. 53:5455. 28. Osterholm, M.T., Forfang, J.C., Ristinen, T.L., Dean, A.G., Washburn, J.W., Godes, J.R., Rude, R.A. and J.G. McCullough. 1981. An outbreak of foodborne giardiasis. N. Engl. J. Med. 304:2428. 29. Rendtorff, R.C. 1954. The experimental transmission of human intestinal protozoan parasites. II. Giardia lamblia cysts given in capsules. Am. J. Hyg. 59:209 220. 30. Rice, E.W. and F.W. Schaefer, III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709710. 31. RobertsThomson, I.C., Stevens, D.P., Mahmoud, A.A.F. and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterol. 71:5761. 32. Rotman, B. and B.W. Papermaster. 1966. Membrane properties of living mammalian cells as studied by enzymatic hydrolysis of fluorogenic esters. Proc. Natl. Acad. Sci. U.S.A. 55:134141. 33. Sauch, J.F. 1987. A new method for excystation of Giardia. In: Proceedings of the Calgary Giardia Conference. P. Wallis and B.R. Hammond (eds). University of Calgary, Calgary, Alberta. pp. 271274. 34. Schaefer, III, F.W., Rice, E.W. and J.C. Hoff. 1984. Factors promoting in vitro excystation of Giardia muris cysts. Trans. Roy. Soc. Trop. Med. Hyg. 78:795800. 35. Schupp, D.G., Januschka, M.M. and S.L. Erlandsen. 1987. Assessing Giardia cyst viability with fluorogenic dyes: comparisons to animal infectivity and cyst morphology by light and electron microscopy. In: Proceedings of the Calgary Giardia Conference. P. Wallis and B.R. Hammond (eds). University of Calgary, Calgary, Alberta. pp. 275279. 36. Wallis, P.M. and H.M. Wallis. 1986. Excystation and culturing of human and animal Giardia spp. by using gerbils and TYIS33 medium. Appl. Environ. Microbiol. 41:483487. 37. Wickramanayake, G.B., Ruben, A.J. and O.J. Sproul. 1985. Effect of ozone and storage temperature on Giardia cysts. J. Am. Water Works Assoc. 77:7477.
Page 255
Fluorescent Dye Exclusion as a Method for Determining Giardia Cyst Viability Susan J. Hudson, Judith F. Sauch, and Donald G. Lindmark*. Cleveland State University, Cleveland, Ohio 44115, U.S.A.. Exclusion of the fluorescent dye, 3[dansylamidophenyl boronic acid] (FluoroBora IFBI) from Giardia muris cysts was compared with excystation as a measure of cyst viability. Nonviable Giardia muris cysts accumulated the dye and fluoresced; whereas viable cysts excluded the dye and showed no fluorescence. The exclusion of FBI from viable Giardia cysts is contrary to the reaction with viable fibroblasts, Chinese hamster ovary cells and Giardia lamblia trophozoites. The effect of 22 different chemicals on % excystation was determined and compared with FBI exclusion under identical conditions. Correlation and regression analyses indicate a high degree of association between the two methods. Our data indicate that the dye exclusion method has potential for use as an alternative method to excystation as a measure of cyst viability. In the dye exclusion method, FBI was added directly to a cyst preparation. Observations were made immediately using a Nikon Episcopic Fluorescent Unit with a wide blue fluorescent cube (exciting filter 410485 nm, dichroic mirror 505 nm, and a barrier filter 515 nm). Routinely, 200300 cysts were scored within 4 min as either fluorescing or not fluorescing. This method was performed in only 5 min compared to the 23 h required for excystation. Since only fluorescence or nonfluorescence were observed, the method was less subjective than excystation. We present data which show that the FBI method is more reproducible and more precise than excystation at levels of viability above 60%. The FBI method has potential for use as a rapid screen of cyst viability and for the rapid assessment of the effectiveness of chemical and physical agents on viability. Preliminary evidence suggests that the FBI method has the potential for use in assessing the viability of Giardia lamblia cysts.
Introduction Giardia lamblia is the most common human intestinal protozoan parasite reported in the United States and England (1,2,13). The organism exists in two morphologically distinct forms, the trophozoite and the cyst. The infective cyst form is transmitted via the fecaloral route. The high incidence, symptomology and waterborne dissemination of Giardia has resulted in improved methods for cyst detection in water systems (11). However a rapid, simple, reliable method for determining the viability of cysts, detected in water systems, and exposed to water treatment procedures is needed. The viability of Giardia cysts has been assessed by eosin exclusion, in vitro excystation, and animal infectivity (3,7). While each of these methods has its own value, none are rapid, simple, and reliable. Although eosin exclusion is a rapid method, it indicates higher cyst viability than can be demonstrated by excystation and shows little correlation with excystation (3). Excystation is the most frequently used method for determining cyst viability. It is tedious (23 h), and subjective. It cannot be used for determining the viability of individual or small numbers of cysts (3). The cost, time and large numbers of cysts required to perform animal infectivity studies make the method impractical for routine use (7). The objective of this study was to develop a rapid, simple and reliable microscopic method for determining the viability of Giardia cysts which, when compared to excystation, would provide a more practical method for estimating cyst viability. To do this, a process called boronic aciddependent phase transfer, or ''boradeption," as described by Gallop et al. (5,6) was used. This fluorescent dye exclusion method was compared with excystation. In this method, specific fluorescent water insoluble boronates are excluded by nonviable cells and taken up by viable cells in vitro (6). The viable cells then exhibited fluorescence. Our results using FluoroBora I with Giardia lamblia trophozoites, were similar to those obtained by Gallop et al. (5,6) with mammalian cells, i.e. viable cells accumulated the dye and exhibited fluorescence. In contrast to the results obtained with viable trophozoites, viable Giardia cysts excluded the dye (nonviable cysts took up the dye and fluoresced). The FBI exclusion method with Giardia cysts showed a high degree of association with excystation and a higher degree of reproducibility and precision when compared to excystation. Materials and Methods Organisms Giardia lamblia cysts were collected from freshly excreted feces of the Mongolian gerbil (Meriones unguiculatus) (4). Giardia muris cysts were obtained from freshly excreted feces of female CF1 mice (10). Giardia lamblia trophozoites were grown as described by Lindmark (8). * Corresponding author.
Page 256
Cyst Purification Giardia cysts were purified according to a modification of the method of RobertsThomson et al. (10). Fecal samples were diluted with distilled water, filtered through a double layer of cheesecloth, and centrifuged at 450 × g for 5 min. The pellet was washed with distilled water by centrifuging for 2 min. until the supernatant solution was free of particulate matter. One 50 mL conical centrifuge tube was prepared for each mL of final pellet volume. The pellet was then resuspended in distilled water and divided equally among the tubes and diluted to 25 mL in each tube Each suspension was then underlaid with 25 mL of 1.0 M sucrose and centrifuged for 10 min at 800 × g. Cysts at the sucrosewater interfaces were aspirated with a Pasteur pipet into clean 50 mL centrifuge tubes and diluted 1 to 5 with distilled water and mixed by vortexing. Finally, these cysts were washed by centrifugation at 1500 × g for 5 min and suspended in 10 to 20 mL distilled water. Cyst counts were made using an improved Neubauer hemocytometer. The cysts were stored at 4°C in 20,000 IU of penicillin and 20 mg streptomycin per mL. and routinely used within one week of isolation. Cyst viability and sensitivity to chlorine as measured by excystation, were not affected during this time (Jarroll, E.L. pers. comm.). Preparation of Fluorescent Dye The boronic acid derivatives (FluoroBorasT M,FB) were purchased from Polysciences, Inc. of Warrington, Pennsylvania. Dyes were prepared as described by Gallop, et al.(6). FBI, (3dansylamidophenylboronic acid) was used in our studies because it produced the most intense fluorescence (unpublished observations, Lindmark). FBI was prepared by dissolving 2.0 mg in 200 µL of dimethylsulfoxide (DMSO). This was then mixed with 600 µL of 25 mM MOPSO [(3morpholino)2 hydroxypropane sulfonic acid] buffer, pH 7.2. Five µL of the FBI reagent were mixed with 25 µL of Giardia cyst preparation. Observations of the presence (indicating non viable cysts) or absence (indicating viable cysts) of fluorescence were made immediately (up to 4 min) with a Nikon Episcopic Fluorescence Unit equipped with a wide blue fluorescence cube (exciting filter 410485, dichroic mirror 505 nm, and a barrier filter 515 nm). Between 200 and 300 cysts were counted per assay. Excystation Experimental and control Giardia cysts were excysted as described by Rice and Schaefer (9). Cysts (106) were centrifuged in a Beckman microfuge 12 at 1500 × g for 5 min. The sedimented cysts were next resuspended in 500 µL HClsaline, 300 µL freshly prepared 100 mM NaHCO3, and 300 µL of a reducing solution consisting of 67 mM glutathione and 29 mM LcysteineHCl in 1X Hanks' balanced salt solution. Cysts were then incubated for 30 min. in a 37°C waterbath. After this incubation step, cysts were washed 3X in 5% trypsin in Tyrode's salts (9). The final pellet was resuspended in 3050 µL trypsinTyrode's salts, mixed, and placed on a depression slide sealed with vaseline. Slides were then incubated 30 60 min at 37°C in a warm air incubator. The number of intact cysts (IC), partially excysted trophozoites (PET) and totally excysted trophozoites (TET) at 400X magnification were determined. Percent excystation was calculated as follows (3): (#TET/2 + #PET) + (#TET/2 + #PET + #IC) × 100. Method Comparison Cysts were exposed to chemical disinfectant so that viability could be assessed by excystation and FBI exclusion. Disinfectants were used which should exhibit different modes of action in order to qualitatively observe chemically dependent discrepancies that might occur between the two methods. Examples of the disinfectant groups used are: phenol derivatives (ortho benzyl para chlorophenol, ortho phenyl phenol); quaternary ammonium compounds (alkyl dimethyl benzyl ammonium chloride, octyl decyl dimethyl ammonium chloride, dioctyl dimethyl ammonium chloride, didecyl dimethyl ammonium chloride, tetradecyl trimethyl ammonium bromide, benzyl dimethyl hexadecyl ammonium chloride, hexadexyl pryidinium bromide, benzyl dimethyl dodecyl ammonium bromide); chlorine containing compounds (chlor hexidine diacetate, sodium hypochlorite, Alcide); glutaraldehydes and mixtures of the above. The chemicals were used at different concentrations and contact times (proprietary recommendations of manufacturers, manuscript in preparation Lindmark, Miller and Zimmer). Approximately 1×106 cysts were placed in a 1.8 mL microfuge tube and centrifuged for 5 min at 1500 × g in a Beckman microfuge (12). The supernatant solution was aspirated and discarded. Control cysts were suspended in 1 mL of water; cysts to be exposed to disinfectants were suspended in 1 mL of the chemical disinfectant at the specified concentrations (proprietary recommendations of manufacturers). Cysts (controls and chemically treated) were incubated at 7, 21 and 24° C for 10 and 20 min (determined according to manufacturer's recommendations). After treatment, control and chemically treated cysts were washed 3X in distilled water for 5 min at 1500 × g, resuspended in 100 µL distilled water and mixed well. Fifty µL were removed and used for excystation; 25 µL of the remaining 50 µL were exposed to FBI. Enumeration of both methods was done simultaneously. The comparisons of the two methods between investigators was done utilizing the procedure given above with the following alterations. The initial amount of cysts (0 16 days old) used was 1×107. These were suspended in 1 mL water after treatment. Ten aliquots were then removed and used for replicate counting using the FBI method. Excystation for three replicates was determined simultaneously. Statistical Analysis Statistically, the two methods were compared using a least squares linear regression analysis and a correlation analysis. The linear relationship between the two methods was estimated by regressing the percentage of viable cysts from the excystation method on the percentage of viable cysts from the FBI method. The estimated slope was tested for a onetoone correspondence between the two methods (H0:B=1) and for a general linear trend between the two methods (H0:B=0). The existence and degree of association between the two methods was examined by calculating Pearson's correlation coefficient and Spearman's Rank correlation coefficient (12). The rank correlation analysis has the advantage that it does not require the distributional assumptions necessary for the validity of linear regression and the significance test of Pearson's correlation coefficient. The statistics indicate a general relationship but not the ability to predict the results of one method accurately from the other. Results Preliminary Studies with FluoroBora IF Ninety five percent (± 4%, 5 determinations) of Giardia lamblia trophozoites (72 hr old) fluoresced (indicating viability) when mixed with FBI as described in materials and methods. When the same cell preparations were exposed to heat at 50°C for 2 min [time (0 to 5 min tested) and temperature (50 to 60°C tested) needed to destroy motility as determined microscopically] the viability as determined by fluorescence decreased to 6% (± 5%, 5 determinations). These results indicate that the procedure for staining viable cells is applicable to Giardia lamblia trophozoites as could be predicted according to the results presented by Gallop et. al. (6). However, viable, freshly harvested cysts of Giardia lamblia and Giardia muris, when heattreated (50°C for 5 min, determined empirically as above) and shown to be killed as measured by excystation (Jarroll, pers. communications), reacted with FBI in an opposite manner. Viable control cysts (unheated) showed 5% (± 3%, 3 determinations) and 8% (± 4%, 3 determinations) fluorescence for Giardia lamblia and Giardia muris respectively. Heattreated cysts exhibited 91% (± 6%, 3 determinations) and 87% (± 10%, 3 determinations)
Page 257
Figure 1. Regression analysis of G. muris PEX % excystation). on PNF (% nonfluorescing). [The solid line is the estimated line and the dashed line represents a line with a slope = 1 (i.e., 11 correlation)]. Data points were obtained using the chemicals given in Materials and Methods. The concentrations, time of contact, temperature and chemical for each data point are not provided because of proprietary restrictions mandated by the manufacturers.
fluorescence for Giardia lamblia and Giardia muris respectively. Despite the difference in staining between Giardia cysts and trophozoites cultured in vitro (6), our results indicate that FBI exclusion from cysts could potentially be used as a method for determining cyst viability. Comparisons Between Excystation and the FBI Method Figure 1 displays the data from the control and chemicallytreated Giardia muris samples. Some chemical disinfectants used resulted in clumping and/or destruction of cysts. Data points produced when cyst clumping was noted (causing problems with excystation counts), and when cyst destruction caused low cyst numbers were omitted (N=8). Figure 2 excludes the points in Figure 1 with 0% excystation and 0% nonfluorescence (N=19). A separate analysis was performed omitting these points because of the possible bias they might have on the results. The plots show the raw data, the fitted regression line, and a line with a slope of one. Both figures show a good relationship between percentage excystation (PEX) and percentage nonfluorescing (PNF). The results of the statistical analyses are presented in Table 1. While the data did not fully meet the assumptions required for the parametric analysis, the results from all analyses were consistent in that a highly significant association was found between the two methods in estimating the percent of viable cysts. Precision and Reproducibility of the FBI and Excystation Methods The viability of the same preparations of Giardia muris cysts (0, 1, 2, 3, and 16 days old, with replicate samples) was assessed with FBI and excystation by two investigators. Figure 3 graphically represents the precision of the methods. In the range tested, the data indicate a high degree of precision (Standard Deviation SD = 0.017) and reproducibility (SD = 0.006) for the FBI method contrary to the precision (SD = 0.061) and
Figure 2. Regression analysis of G. muris PEX (% excystation) on PNF (% nonfluorescing) excluding 19 (0,0) points. [The solid line is the estimated line and the dashed line represents a line with a slope = 1 (i.e., 11 correlation)]. The same proprietary restrictions given for Figure 1 apply to Figure 2.
reproducibility (SD = 0.030) of the excystation method. Preliminary Studies With Giardia Lamblia Cysts Freshly harvested Giardia lamblia cysts exhibited lower excystation rates (2050%) than cysts of Giardia muris (9099%) and the excystation rates showed no correlation with FBI exclusion (Table 2). The mean excystation of these cysts (35%) was 42% of the viability as determined by the FBI method (82%). However, it was noticed that after induction, when the cysts were placed in excystation medium, the trophozoites inside the cyst exhibited motility within 2 min. If motility inside the cyst is used as a measure of viability, instead of excystation, a much better correlation with the FBI method is observed (Table 2). These preliminary data (Table 2) with Giardia lamblia, correlating motility inside the cyst during the excystation procedure, as a measure of viability, with FBI exclusion, indicate that the relationship between these two methods is more similar to that found using excystation and FBI exclusion with Giardia muris cysts. TABLE 1. Results of data analysis for G. muris samples.
Excluding Observations PNF = 0 and PEX = 0
All Data Regression Analysis:
N 2
R
19
0.97
0.86
(slope)
1.02
1.00
95% C.I.* for B
(0.96, 1.08)
(0.80, 1.21)
Correlation Analysis:
Pearson's R (p)
0.98 (<0.001)
0.93 (<0.001)
Spearman's R (p)
0.97 (<0.001)
0.89 (<0.001)
*Confidence Interval PNF = % nonfluorescing, PEX = % excystation
44
Page 258
Figure 3. Precision of FBI and excystation methods for determining G. muris cyst viability : Investigator A [PNF(% nonfluorescing) = 0, PEX(% excystation) = X) and B (PNF = , PEX = ). Replicate analysis was done on 8 replicate samples (from 116 days old).
Discussion Fluorescent dye exclusion (FBI method) shows good correlation (at cyst viabilities >60%) with excystation as a method for determining Giardia muris cyst viability. FBI is excluded from viable cysts, in contrast to the results obtained with viable fibroblasts, Chinese hamster ovary cells (6) and Giardia lamblia trophozoites in which viable cells take up FBI and fluoresce. This may be explained in part by the chemical nature (unknown at the present time) of the cyst wall and the permeability barrier caused by the cyst wall. The cyst wall may lack the lipoidal components, found in plasma membranes, which are needed for the "boradeption" reaction (permeation of FBI into viable cells). However if the cyst wall permeability is altered by chemical and heat treatment, the FBI can then enter and form complexes with hydroxyl and amino groups within the cell resulting in fluorescence. The FBI method was simple to perform, rapid (5 min), and can be microscopically read. The method shows a higher degree of precision than excystation at viability levels above 60% for control cysts. The FBI method is also less subjective than excystation. The major reasons for this lie in the facts that 1) the FBI method relies on an observation of fluorescence or nonfluorescence as a measure of viability; excystation requires the performance of a series of counts of intact cysts, totally excysted trophozoites and partially excysted trophozoites and it is difficult for all but the well trained investigator to differentiate among the different stages required in the counting procedure and 2) the excystation procedure is long and tedious (23 h) requiring control of a number of chemical and physical variables (temperature, pH, reducing conditions, etc.) not encountered in the FBI method. When the FBI method was used to determine the effect of various chemicals on cyst viability; clumping and/or cyst destruction occurred with several chemical disinfectants (phenol and chlorine containing compounds). Samples that show clumping cannot be evaluated for viability by excystation because of the difficulty in identifying the various stages needed for counting. Clumping, on the other hand, has little effect on counts by the FBI method. Cyst destruction is a problem in both methods in that it is difficult to obtain statistically accurate counts. Chemical disinfectant concentrations should be chosen that cause cyst death without destroying the cysts so that enough cysts remain to make the counts significant. The data presented in Figure 3 indicates that FBI measurements of Giardia muris viability are consistantly higher than by excystation, suggesting the possibility that the FBI method is more sensitive for determining cyst viability than excystation. Similar results were shown by Bingham et al. (3) with eosin dye exclusion using Giardia lamblia cysts. Preliminary evidence suggests that the FBI can be used to determine the viability of Giardia lamblia cysts. The data presented in Table 2 shows a good correlation between motility inside the cyst and FBI exclusion but not between excystation and FBI exclusion. The suggestion can also be made (based on our preliminary information) that the excystation procedure we routinely used for Giardia (9) may not be appropriate for Giardia lamblia in that it gives a low estimate of Giardia lamblia cyst viability compared to the FBI method. This suggestion is in agreement with the results found by Bingham et al. (3) with eosin exclusion in which excystation gave a lower estimate of viability than eosin exclusion. The suggestion should also be made that eosin exclusion might be further investigated as a method of assessing cyst viability. Giardia lamblia trophozoites inside the cysts exhibit motility (suggesting viability) during the excystation procedure but a high percent fail to exit the cyst wall. It can be hypothesized that the environment in the excystation medium is not condusive to the final steps of excystation TABLE 2. Viability of Giardia lamblia cysts measured by excystation, motility inside the cyst and FBI exclusion.
Cyst Preparation
Age*(days)
Excystation
Motility in Cyst
FBI Exclusion
1
0
35.2
72.4
77.9
2
0
40.1
71.6
80.0
3
1
20.0
75.5
75.6
4
3
39.6
85.4
89.6
5
5
41.5
91.0
90.2
6
0
25.6
72.3
71.7
7
5
27.4
71.2
78.1
8
5
51.6
90.3
95.6
9
0
29.8
75.2
79.2
10
1
42.8
84.1
85.3
*Relative to time of collection.
Per Cent of Cysts Showing
Page 259
because of lacking nutrients, cofactors, etc. or the presence of inhibitors. Further studies should investigate this phenomenon. In conclusion, we suggest that the FBI method has the potential for use in the rapid assessment of Giardia muris cyst viability (at levels of viability above 60%). It is less subjective and more precise than excystation. In order to state that "the FBI method can be used as a substitute for excystation" further work needs to be done. This research would involve an in depth comparison of the two methods with Giardia lamblia cysts and a comparison of the two methods with cyst preparations of less than 50% viability. Presently only minimal observations (unpublished) have been made between 1% and 50% viability and the data have been highly variable. Acknowledgements The authors would like to thank Dr. Judy Stober and Tammy Mills (USEPA, Cincinnati) for their statistical analysis of the data. This document has been reviewed in accordance with U.S. Environmental Protection Agency policy through cooperative agreement#CR811949010 to Cleveland State University and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. Literature Cited 1. Intestinal parasites ranging far afield in the United States. 1978. Medical News. J. Amer. Med. Assoc. 239:2756. 2. Communicable Disease Weekly Reports. 1977. Public Health Laboratory Service, London. 3. Bingham, A.K., Jarroll, E.L., Meyer, E.A., and S. Radulescu. 1979. Giardia sp.: physical factors of excystation in vitro, and excystation vs. eosin exclusion as determinants of viability. Exp. Parasitol. 47:284291. 4. Belosevic, M., Faubert, G.M., MacLean, J.D., Law, C., and N.A. Croll. 1983. Giardia lamblia infections in Mongolian gerbils: an animal model. J. Infect. Diseases. 147:222226. 5. Gallop, P.M., Paz, M.A., and E. Henson. 1982. Boradeption: a new procedure for transferring water insoluble agent across cell membranes. Science 217: 166 169. 6. Gallop, P.M., Paz, M.A., Henson, E., and S. Latt. 1984. Dynamic approaches to the delivery of reporter agents into living cells. Biotechniques 2:3236. 7. Kasprzak, W. and A.C. Majewska. 1983. Infectivity of Giardia sp. in relation to eosin exclusion and excystation in vitro. Tropenmed. Parasit. 34:7072. 8. Lindmark, D.G. 1980. Energy metabolism of the anaerobic protozoan Giardia lamblia trophozoites. Mol. Biochem. Parasitol. 1:112. 9. Rice, E.W. and F.W. Schaefer, III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709710. 10. RobertsThomson, I.C., Stevens, D.P., Mahmoud A.A.F., and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterology 71:5761. 11. Sauch, J.F. 1985. Use of immunofluorescence and phasecontrast microscopy for detection and identification of Giardia cysts in water samples. Appl. Environ. Microbiol. 50:14341438. 12. Wardlaw, A.C. 1985. Practical Statistics for Experimental Biologists. John Wiley and Sons, New York. 13. Wolfe, M.S. 1975. Giardiasis. J. Amer. Med. Assoc. 233:13621365.
Page 261
A New Method for Excystation of Giardia Judith F. Sauch U.S. Environmental Protection Agency, Health Effects Research Laboratory, 26 West Martin Luther King, Cincinnati, Ohio 45268, U.S.A.. In vitro excystation of Giardia is used to evaluate cyst viability and may also be used to obtain trophozoites and cyst walls for analysis. Recently published excystation procedures include the use of trypsin, serum or bile salts in the "excystation" step. Because Giardia muris and Giardia lamblia cyst walls and trophozoites were to be used for antigenic analysis, an alternative method was devised to avoid exposure to trypsin, serum and bile salts. An initial acid or low pH "induction" step was retained, but cyst exposure to the medium was increased to 45 min. Trypsin, serum or bile salts in the "excystation" medium was replaced by proteose peptone for G. muris and phytone peptone for G. lamblia. These reagents were added to Hanks' Balanced Salt Solution (HBSS) supplemented with cysteine and sodium bicarbonate. Increased levels of cysteine and glucose in HBSS decreased percent excystation for G. lamblia cysts (from human donors). Comparison between this method utilizing peptones and a method utilizing trypsin showed no significant differences in percent excystation for both G. muris and G. lamblia. Variation in percent excystation among seven G. lamblia cyst samples from human donors occurred but the two methods followed similar trends as they varied, with no significant differences between the two. Thus, neither trypsin, serum or bile salts is required for excystation of G. muris or G. lamblia.
Introduction Currently available excystation procedures expose Giardia cyst walls and trophozoites to complex media containing enzymes, serum, or bile, which might interfere with subsequent electrophoretic analyses. Therefore, alternative reagents were devised in order to avoid exposure to these agents and still obtain high levels of excystation. Although Hegner concluded in 1927 (6), after observing excystation in vivo, that moisture, temperature, and digestive juices were major factors in the process, it was not until the 1970's that excystation requirements for Giardia were defined more precisely. Bingham and Meyer (2) were the first to demonstrate that although hydrogen ions initiate the excystation process, transfer to a second favorable medium is critical to complete excystation. The "excystation" medium they used (HSP 3;9) contains human serum, phytone, vitamins, amino acids, and salts. Subsequent to these publications, a number of excystation procedures were developed as modifications of the original. In all of them, the acid incubation step is retained for induction, but supplementation of the excystation medium varies. Serum (1,2,4,7), bile (4,7), and enzymes such as trypsin (10,14), pepsin (7,14), and pancreatin (8) have been used with a variety of excystation levels reported. Schaefer et al. (14), however, noted that trypsin, while appearing not to be essential for high levels of excystation of G. muris cysts, seems to accentuate the process. An alternative excystation medium consisting of balanced salt solution supplemented with a minimum of complex reagents was developed. Since phytone is included in Bingham's medium and appeared to affect excystation favorably (E.W. Rice, personal communication), it was used first to replace serum enzymes and bile. Experiments in which phytone peptone, proteose peptone, glutathione, cysteine, and glucose were varied or eliminated led to the formulation of a simplified excystation medium consisting of a balanced salts solution supplemented with proteose or phytone peptone, cysteine, and glucose. High excystation levels for G. muris, comparable to published results, and variable, but sometimes high, excystation levels for G. lamblia were obtained. This suggests that complex agents such as enzymes, whole serum, and bile are not essential for Giardia excystation. Materials and Methods Organisms G. muris cysts were obtained from female Swiss albino mice (CF1) infected with cysts per os. Cysts were harvested by modifications of the procedure developed by RobertsThomson et al. (11). Mouse feces from approximately 60 animals were pooled after 1621 hours collection, mixed into a slurry, and diluted to about 2500 mL with 0.01% Tween 20 in distilled water (TDW). After mixing thoroughly, the slurry sat at room temperature for onehalf hour to sediment large debris. After this sedimentation was repeated twice, the resulting supernatant fluids were centrifuged at 800 × g for 2 minutes. The pellets were repeatedly mixed by vortexing into 0.01% TDW and centrifuged until the supernatant fluids were freed of fine particles. Finally, the pellets were pooled, suspended in about 600 mL TDW, mixed well, dispersed into six round bottom plastic 250 mL centrifuge bottles, underlayed with 80 mL 1.0 M sucrose, and centrifuged in a swinging bucket rotor for 10 minutes at 800 × g. The bands of cysts floating between the sucrose and water layers were collected by aspiration into a flask, diluted 1:5 with TDW, and washed three times in TDW by centrifugation for 5 minutes at 800 × g. Cysts were further purified by sedimentation at unit gravity (12) and stored in distilled water at 25°C. G. lamblia cysts were harvested in a similar manner, with a few modifications, from human donor stool specimens. Large floating particles were filtered out with 23 layers of cheesecloth. After washing, pellets were suspended in TDW; 25 mL of the pellet suspension (representing 1 mL of packed pellet volume) was overlayed onto 25 mL Percollsucrose, sp. gravity 1.09 (13). Cysts were found in a band between the water and Percollsucrose layers. These bands were collected, pooled, diluted 1:5 with TDW and washed three times
Page 262 7
by centrifugation for 5 minutes at 800 × g. Cysts were purified by sedimentation at unit gravity if more than 10 were collected. They were then stored in distilled water at 25°C until use. Excystation of G. lamblia The acid incubation step (10) was retained. Briefly, 0.5 mL of up to 107 cysts in distilled water was transferred to a 15 mL conical centrifuge tube to which was added 5 mL HClsaline (0.7 mL conc. HCl, 100 mL 0.85% NaCl, pH 1.5), 3 mL reducing solution (HBSS supplemented with 32 mM glutathione and 57 mM Lcysteine HCl) and 3 mL 0.1 M NaHCO3. The suspension was mixed by vortexing, incubated in a 37°C water bath for 45 minutes, and centrifuged for 2 minutes at 650 × g. The cysts in the pellets were washed once in prewarmed 0.85% NaCl by centrifugation for 2 minutes at 650 × g. The pellet was finally suspended in 15 mL prewarmed excystation medium and incubated for 30 minutes in a 37°C water bath. The excystation medium, prepared fresh, was prepared by dissolving 0.015 g NaHCO3 and 0.01 g LcysteineHCl monohydrate in 4 mL of 5% phytone peptone and 1 mL 10X HBSS; the final volume was brought to 10 mL with distilled water (final pH 7.1 and final phytone peptone 2%). Stock phytone peptone (5% W/V) was prepared by adding phytonepeptone (BBL #11906) to distilled water and boiling gently with stirring for 10 minutes. After filter sterilization, the stock solution was stored at 25°C in small volumes. Samples were removed from the excystation tube and counted in a hemocytometer. At least 200 totally excysted trophozoites (TET), partially excysted trophozoites (PET), and intact cysts (IC) were counted. Excystation was quantitated according to the formula reported by Bingham et al. (3) as follows: [TET + 2 + PET] + [TET + 2 + PET + IC] × 100. In contrast to G. muris, empty cyst walls of G. lamblia are difficult to observe; however, G. lamblia trophozoites settle quickly and can be counted (10). The number of TET's is halved because every excysted cyst yields a pair of trophozoites (3). This method was compared to that of Rice and Schaefer (10) using G. lamblia cysts isolated from several asymptomatic and symptomatic human donors. Excystation of G. muris The induction step was carried out as reported by Schaefer et al. (14) with minor modifications. Between 106 and 3 × 107 cysts were suspended in 1 mL distilled water in a plastic conical bottom centrifuge tube to which was added 20 mL of reducing solution [Hanks' Balanced Salt Solution (HBSS) supplemented with 32 mM glutathione and 57 mM Lcysteine HCl] and 20 mL of 0.1 M NaHCO3 (final pH 4.7). The suspension was mixed by vortexing, incubated in a 37°C water bath for 30 minutes, centrifuged at 650 × g for 2 minutes, and then washed once in the excystation medium by centrifugation at 650 × g for 2 minutes. Cysts were finally suspened in 15 mL (depending upon the number of cysts) prewarmed 0.5% proteose peptone in PBS, pH 7.2, and incubated for onehalf hour in a 37°C water bath. A stock proteose peptone (5% W/V) solution was prepared in distilled water, gently boiled for 10 minutes to destroy any remaining enzymes, filter sterilized, and stored at 2 5°C in small volumes. Excystation medium, prepared fresh, was made by adding 10 mL stock proteose peptone and 10 mL of 10X Phosphate Buffered Saline (PBS, 80 g NaCl, 2 g KH2PO4, 29 g Na2HPO4 ∙ 12 H2O, 2 g KCl, 1000 mL final volume) to 80 mL distilled water. This method was compared to that of Schaefer et al. (14). Both excystation methods were performed in centrifuge tubes; samples were removed and counted in a hemocytometer, using phase contrast optics at 200400 magnification. Between 100 and 200 empty cyst walls (ECW), partially excysted trophozoites (PET), and intact cysts (IC) were counted. The percent excystation was calculated according to the formula [ECW + PET] + [IC + ECW + PET] × 100 (14). Results and Discussion Development of Excystation Method for G. lamblia A single sample from an asymptomatic human donor was used to develop serum and enzymefree reagents. The acid induction step (10) was retained, but improved excystation occurred when this step was increased to 45 minutes (unpublished observations). After cysts were exposed to various excystation media consisting of 0.5% phytone or proteose peptone in either HBSS or Tyrode's medium, excystation was higher (11 and 8%, respectively) than that observed (<6%) when cysts were exposed to trypsinTyrode's solution, phytoneTRIS, or proteoseTRIS buffer, pH 7.0. Tyrode's medium or HBSS was next supplemented with various concentrations of phytone (0.252.0%). HBSS plus 2.0% phytone peptone yielded 57% excystation, whereas the other concentrations of phytone in HBSS yielded 5055%. At phytone concentrations lower than 2%, however, the TET were nonrefractile and judged to be dead. Therefore, they were not amenable to successful separation from cyst walls by isopycnic centrifugation. In the same experiment, Tyrode's medium containing trypsin or phytone resulted in less than 21% excystation. Further supplementation of HBSSphytone (2.0%) with glucose (0.05% above that routinely used for HBSS) and 1.0% cysteine increased excystation to 73%. However, elimination of the extra glucose, but not cysteine or phytone, yielded 90% excystation. Controls consisting of cysts exposed to trypsinTyrode's medium yielded only 27% excystation. In the same experiment, lower levels of excystation resulted when cysts were exposed to HBSS (14%), HBSSglucose (15%), HBSS plus 2.0% phytone (17%), and HBSS plus glucose and phytone (22%). When HBSS was supplemented with only 1.0% cysteine, but not phytone, excystation was as high as 83%. This suggested that cysteine plays an important role in the excystation process. A decrease in the concentration of cysteine to 0.1% resulted in 96% excystation, whereas, in the same experiment, elimination of cysteine by incubation of cysts in trypsinTyrode's medium alone reduced excystation to 57%. In all of these experiments, HBSS was supplemented with 0.15% NaHCO3. This is consistent with the results of Gillin and Reiner (5) which demonstrate that G. lamblia trophozoites depend upon thiol reducing agents, such as cysteine, for survival and attachment to glass surfaces. These results demonstrated that a complex medium containing enzymes is unnecessary and can be replaced with a simple medium consisting of a balanced salts solution, cysteine and phytone peptone. The concentration of phytone found to be most effective was slightly higher than that reported previously for the first successful in vitro excystations (2). These results suggest a minor role of some phytone constituent and a major role of Lcysteine. Development of Excystation Method for G. muris G. muris cysts were able to excyst when the above method developed for G. lamblia was used. Unfortunately, the TET died quickly and could not be adequately separated from the empty cyst walls by isopycnic centrifugation. This situation improved when proteose peptone was used in the excystation step. Optimal appearance of the TET occurred with 0.5% proteose peptone, even though excystation was >98% with all concentrations of proteose peptone used (0.22.0%). Cysteine and
Page 263 TABLE 1. Comparison of methods for excysting Giardia lamblia.
Mean % Excystation of Method (range)
Samplea
Cyst age (days)b
TrypsinTyrodec
CysteinePhytoned
1
8
15 (1217)
15 (1417)
2
8
82 (8084)
93 (9096)
3
12
84 (8385)
93 (8996)
4
21
38 (3343)
49 (4354)
5
13
<1 (01)
2 (2)
6
14
21 (1824)
21 (1725)
7
15
30 (2732)
19 (1820)
a
Samples 14 were from the same asymptomatic donor, collected at different times. Samples 57 were from different asymptomatic donors. b
Time from cyst purification to day of excystation.
c
Method of Rice and Schaefer (10). Results were the average of two trials.
d
Method described in text for G. lamblia. Results were the average of two trials. All samples were excysted on the same day.
NaHCO3 were eliminated from the excystation medium after it was determined that excystation remained at >99% in the absence of these reagents. In contrast, G. lamblia cysts were incapable of excystation when proteose peptone was substituted for phytone peptone, and showed reduced levels of excystation when cysteine was not present. G. muris excysted (>99%) when a simple induction medium consisting of HBSS, cysteine, glutathione, and NaHCO3 was used. Even though G. muris cysts were capable of >99% in the presence or absence of HCl in the induction step, that reagent was omitted to avoid unnecessary exposure of cysts to it. Recently, Feely (4) reported that reducing agents were not necessary for induction of G. muris. Although reducing agents had been retained in our induction step, an experiment was performed in which the standard induction medium (HBSS, cysteine, glutathione, and NaHCO3) was compared to acidified HBSS minus cysteine and glutathione (4) and HBSS to which only NaHCO3 was added. While the use of the standard induction solution yielded >99% excystation, use of the latter two media resulted in lower excystation (20% and 5%, respectively). The TYI excystation medium used by Feely (4) is supplemented with serum and bile. The difference in supplementation of the excystation media could account for the decrease in excystation when reducing agents were omitted from my induction step. Method Comparison and Use Once the excystation methods for G. lamblia and G. muris were standardized, they were compared to the methods of Rice and Schaefer (10,14). For either species of cyst, both excystation procedures were performed on the same day with all samples used. As can be seen from the results (Tables 1 and 2), no apparent differences existed between the two excystation methods for both species of Giardia. Both sampletosample variation and extremes of percent excystation occurred, but the two methods followed similar trends as they varied. It has been previously reported (10) that asymptomatic donors of Giardia cysts yield lower TABLE 2. Comparison of methods for excysting Giardia muris.
Mean % excystation of method (range) Cyst age (days)a
TrypsinTyrodeb
Proteose peptonecc
1
84 (8089)
91 (8994)
2
91 (9097)
89 (8890)
3
91 (9092)
89 (8890)
a
Time from cyst purification to excystation. The same batch of cysts was used for each day.
b
Method of Schaefer et al. (14). Results were the average of three trials.
c
Method described in text for G. muris. Results were the average of three trials.
excystation percentages than symptomatic carriers. However, data reported here (Table 3) show that the mean excystation of cysts from asymptomatic carriers (56% ± 11% standard error) was similar to that of cysts from symptomatic donors (50% ± 13% SE). This discrepancy between the previously reported data and that presented here is not surprising when data from the same donor are examined. Cysts from donors A and B (Table 3) showed extremes of percent excystation when different samples from the same donor and cysts of various ages from the same specimen were compared. Also, samples 14 TABLE 3. Giardia lamblia cysts from symptomatic and asymptomatic human donors excysted utilizing a cysteine & phytone containing mediuma. Donorb
Sample
Cyst age (days)c
Mean % excystationd
No. of trials (n); range
A
1
1
93
n=1
1
2
90
n=1
2
11
17
n = 3; 1519
2
13
1
n = 3; 12
A
3
20
96
n = 3; 9598
B
1
2
53
n = 3; 4163
1
4
75
n = 2; 74&76
1
8
4
n = 3; 35
A
C
1
1
21
n = 1
D
1
2
89
n = 1
E
1
4
78
n = 1
F
1
6
95
n = 1
G
1
6
93
n = 1
H
1
8
61
n = 1
I
1
10
9
n = 2; 7&10 n = 1
J
1
12
79
K
1
15
75
n = 1
L
1
15
31
n = 8; 2540
M
1
15
8
n = 2; 6&10
N
1
16
27
n = 1
O
1
17
93
n = 3; 9293
a
Method described in text for G. lamblia.
b
A and B were asymptomatic carriers and C through O were symptomatic donors.
c
Time from cyst purification to excystation.
d
At least 200 objects (IC, TET, PET) were counted per trial.
Page 264
(Table 1) from the same donor exhibited variable excystation levels. Cysts from several symptomatic carriers (Table 3) showed wide excystation variability. The methods reported here which included the incorporation of phytone or proteose peptone and cysteine into HBSS were found to be simpler than those previously published. The latter procedures use complicated media such as HSP3 (2) or TYI (4) and incorporate crude trypsin (10,14), bile (4,7) or whole serum (1,2,4,7) into the excystation medium. The preparation of these media is timeconsuming and exposed the trophozoites and cysts to such undesirable reagents as trypsin, whole serum, or bile. For G. lamblia and G. muris cysts, it was shown that the excystation process did not require the use of serum, specific enzymes, or bile. A heat inactivated protein or protein digest such as phytone peptone, but not proteose peptone, plus a reducing agent, such as Lcysteine, were necessary in the excystation medium for G. lamblia. G. muris does not require a reducing agent such as Lcysteine in its excystation medium. Acknowledgements I wish to thank Frank W. Schaefer, III for the G. muris cysts and Floyd Frost for the G. lamblia cysts from human donors. I also wish to acknowledge the very able technical assistance of Melanie Mormile and Debbie Flanigan. Notice This document has been reviewed in accordance with U.S. Environmental Protection Agency policy and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. Literature Cited 1. Bhatia, V.N., and D.C. Warhurst. 1981. Hatching and subsequent cultivation of cysts of Giardia intestinalis in Diamond's medium. J. Trop. Med. Hyg. 84:45. 2. Bingham, A.K., and E.A. Meyer. 1979. Giardia excystation can be induced in vitro in acidic solutions. Nature (Lond.) 277:301302. 3. Bingham, A.K., Jarroll Jr., E.L., and E.A. Meyer. 1979. Giardia sp.: physical factors of excystation in vitro, and excystation vs. eosin exclusion as determinants of viability. Exp. Parasitol. 47:284291. 4. Feely, D.E. 1986. A simplified method for in vitro excystation of Giardia muris. J. Parasitol. 72:474475. 5. Gillin, F.D., and D.S. Reiner. 1982. Attachment of the flagellate Giardia lamblia: role of reducing agents, serum, temperature, and ionic composition. Mol. Cell. Biol. 2:369377. 6. Hegner, R. 1927. Excystation and infection in the rat with Giardia lamblia from man. Am. J. Hyg. 7:433447. 7. Kasprezak, W., and A.C. Majewska. 1983. Infectivity of Giardia sp. cysts in relation to eosin exclusion and excystation in vitro. Tropenmed. Parasitol. 34:70 72. 8. Marchin, G.L., Fina, L.R., Lambert, J.L., and G.T. Fina. 1983. Effect of resin disinfectantsI3 and I5 on Giardia muris and Giardia lamblia. Appl. Environ. Microbiol. 46:965969. 9. Meyer, E.A. 1976. Giardia lamblia: isolation and axenic cultivation. Exp. Parasitol. 39:101105. 10. Rice, E.W., and F.W. Schaefer, III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709710. 11. RobertsThomson, I.C., Stevens, D.P., Mahmound, A.A.F., and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterol. 71:5761. 12. Sauch, J.F. 1984. Purification of Giardia muris cysts by velocity sedimentation. Appl. Environ. Microbiol. 48:454455. 13. Sauch, J.F. 1985. Use of immunofluorescence and phasecontrast microscopy for detection and identification of Giardia cysts in water samples. Appl. Environ. Microbiol. 50:14341438. 14. Schaefer, F.W., III, Rice, E.W., and J.C. Hoff. 1984. Factors promoting in vitro excystation of Giardia muris cysts. Trans. Roy. Soc. Trop. Med. Hyg. 78:795800.
Page 265
Assessing Giardia Cyst Viability with Fluorogenic Dyes: Comparisons to Animal Infectivity and Cyst Morphology by Light and Electron Microscopy Daniel G. Schupp*, Mary M. Januschka, and Stanley L. Erlandsen Department of Cell Biology and Neuroanatomy, 4135 Jacckson Hall, University of Minnesota, Minneapolis, Minnesota 55455, U.S.A.. The fluorogenic dyes fluorescein diacetate (FDA) and propidium iodide (PI) were used to assess Giardia cyst viability. Viable Giardia cysts, possessing an intact cell membrane, retained the dye FDA and were seen to fluoresce green at 450490 nm. Nonviable cysts incorporated the dye PI and fluoresced red at 545546 nm. Neonatal mice given FDA positive cyst inocula became infected as evidenced by fecal cyst shedding and positive postmortem examination for trophozoites. None of the mice, given cyst inocula that were PI positive, became infected. Using differential interference contrast microscopy, comparisons between Giardia muris cysts labeled with FDA and those labeled with PI showed unique morphological differences including a prominent cyst wall and peritrophic space in the FDA positive cysts, but not in the PI positive cysts. FDA cysts possessed a hyaline cytoplasmic appearance whereas PI cysts looked granular and the intracellular organelles were easily detected. Preliminary ultrastructural observations have shown that nonviable cysts possessed a cyst wall lacking structural continuity and that the trophozoite membrane adjacent to the inner aspect of the cyst wall was difficult to delineate. Based on our studies, fluorogenic dyes provide a simple and direct method of measuring Giardia cyst viability that correlates well with both morphological differences and animal infectivity.
Introduction Fluorogenic dyes have been accepted as a sensitive means for determination of cell viability since their description in 1966 by Rotman and Papermaster (16). Recently, these dyes were employed as cytochemical tools for making measurements on many cellular parameters. These ranged from cytosolic viscosity (5) and pH (21) to nuclear cytoskeletal interactions (13). Cell migration studies (4) and viability determinations in both static systems [counting absolute numbers of differentially stained cells (10)] and dynamic systems [measuring antibody dependent cell cytotoxicity or intracellular killing of bacteria (20,9)] have been made. Methods that have been used to determine Giardia cyst viability have included: dye exclusion (2,11), excystation in vitro (2,3,11,14) and the use of animal models for infectivity (1,15). Drawbacks to these techniques have recently been discussed (17) and include subjectivity, time, expense and loss of the test cyst population for further experimental manipulation. Recently, our laboratory has reported an excellent correlation between fluorogenic dye staining of Giardia cysts with both animal infectivity and cyst morphology (17,18). The incorporation of the fluorogenic dyes, fluorescein diacetate (FDA) and propidium iodide (PI), has been shown to represent either viable or nonviable G. muris cysts, respectively. This has been verified using the indirect animal infectivity test. Furthermore, it has been demonstrated that distinct morphological features enabled viability determinations of G. muris cysts to be made by differential interference contrast (DIC) microscopy. The morphological basis for structural differences observed by DIC in viable and nonviable G. muris cysts was investigated by electron microscopy. Based on these observations, we have proposed an explanation for the uptake of fluorogenic dye by viable and nonviable cysts and have also presented data showing that cyst morphology by DIC was closely coordinated with changes in fluorogenic dye uptake when viable cysts become nonviable as defined by PI incorporation. Materials and Methods Fluorogenic staining solutions were made according to the technique described by Schupp and Erlandsen (17). All cysts were viewed on either a Zeiss photomicroscope III or an Olympus BH2 light microscope equipped with epiillumination, DIC and phase optics. Cysts were isolated from the feces of infected CF1 mice by the technique of RobertsThomson et. al. (15). Stained cysts were separated using a Becton Dickinson fluorescence activated cell sorter (FACS IV). Sorted groups of either FDA or PI positive cysts were verified by fluorescence microscopy. All aspects of the animal infectivity were performed as described by Schupp and Erlandsen (17). To test whether changes had occured in cyst morphology as a function of fluorogenic dye uptake, cysts were stained with FDA/PI. They were then pelleted and placed on a microscope slide, and coverslipped. A field containing cysts stained with FDA was selected and photographed under both epiillumination and DIC optics. The same field was photographed at varying time intervals for up to 12 hours after staining. Cysts were stained with FDA and PI, then sorted on a FACS IV, and prepared for electron microscopy as described by Shands (19). * Corresponding author.
Page 266 TABLE 1. Inoculum Cysts/Animal
Fluorogenic Dye Uptake
Controls:
Experimental:
Giardia Cyst Appearance Days Post Inoculation
By DIC
0 Cysts (N = 8)
Type of Cyst Morphology
3
None
5
0/8
8
0/8
11
0/8
0/8
1,000 Cysts (N = 10)
FDA Postive
Viable
0/10
10/10
10/10
10/10
5,000 Cysts (N = 10)
PI Positive
NonViable
0/10
0/10
0/10
0/10
The sections were cut on a LKB Huxley ultramicrotome, stained with uranyl acetate and lead citrate (12) and examined with a JEOL 100 CX electron microscope. Results Fluorogenic Dye Uptake and Animal Infectivity The viability of Giardia muris cysts was investigated by administration of fluorogenic dye exposed cysts to neonatal mice. The results have been shown in Table 1. Control neonatal mice that did not receive an inoculum of Giardia muris cysts also did not pass cysts in their feces during the course of the experiment (day 3, 5, 8, 11, postinoculation) nor were any trophozoites present within their small intestine at necropsy. None of the mice (N=10) experimentally inoculated with doses of 1,000 FDA positive Giardia muris cysts, isolated by FACS, shed cysts on day three postinoculation. However, 100% of these mice were positive for fecal cysts on days 5, 8, and 11. Necropsy of the animals infected with FDA positive cysts revealed Giardia trophozoites within the small intestine. Ten mice were each inoculated with 5,000 PI positive cysts. Inoculation of PI positive Giardia cysts failed to establish an infection as evidenced by (1) the lack of Giardia cysts in the feces and (2) the absence of any trophozoites within the small intestine at necropsy. Cyst Viability by Light Microscopy The morphological appearance of Giardia muris cysts was studied by DIC microscopy and correlated with incorporation of fluorogenic dyes as seen in Figures 1a 1c. This technique showed striking differences between the nonviable and the viable cysts. The viable FDA cysts revealed by DIC a clearly delineated cyst wall, a space between cyst wall and cytoplasm [referred to as the peritrophic space (6)] and a hyaline appearance to the cyst which made the viewing of intracellular detail difficult (Figure 1a). The nonviable PI stained cysts displayed easily recognizable cytoplasmic organelles including two to four polar nuclei, intracytoplasmic flagellar axonemes and curved portions of the disassembled adhesive disc. By DIC, the nonviable PI cyst cytoplasm had a fine granular appearance, as opposed to the hyaline appearance of the viable FDA cysts. To investigate how closely in time Giardia cyst morphology by DIC was coupled to changes in the fluorogenic staining pattern, a sample of fluorogenic dye exposed G. muris cysts was sealed under a coverslip. A microscopic field was selected that contained viable FDA cysts, viable nonstaining cysts, (17) and nonviable PI cysts. A representative field containing both viable FDA staining and nonstaining viable cysts was photographed 5 hours after staining with FDA/PI by DIC (Figure 2a), at
Figure 13. Comparison between viable and nonviable G. muris cysts using DIC (1a, 2a, and 3a) and fluorescence microscopy (1b,c; 2b,c; and 3b,c). In Figure 1a, note the structural differences (nuclei, flagellar axonemes, and cytoplasmic appearance) between viable (V) and nonviable (N) cysts. n, nuclei; ps, peritrophic space. In Figure 1b, viable G. muris cysts incorporating FDA fluoresce green at an excitation wavelength of 450490 nm, while the nonviable cyst staining with PI fluoresces orange. The staining for FDA appears around the periphery of the cyst and in the peritrophic space (ps). In Figure 1c, only the nonviable cyst staining with PI fluoresces red at an excitation wavelength of 545546 nm. Cysts incorporating FDA are not excited to fluoresce at this wavelength. Figure 2a illustrates by DIC a group of G. muris cysts after 5 hours exposure to FDA/PI. All cysts are viable by morphologic criteria except for one nonviable cyst (N). Examination of this same group of cysts for FDA uptake (450490 nm excitation) is shown in Figure 2b and for PI staining (545546 nm excitation) in Figure 2c. The two viable cysts indicated with double arrows in Figure 2a are seen one hour later in Figure 3a by DIC. These cysts are now nonviable (N) by morphologic criteria and also fluoresce positively for PI as shown at 450490 nm excitation in Figure 3b and 545546 nm excitation in Figure 3c. In addition, comparison of Figures 2a and 2b to Figures 3a and 3b, shows that within the same one hour time frame, two nonstaining cysts are now accumulating FDA. Bar equals five microns for Figure 1ac; bar equals five microns for Figures 2ac and 3ac.
Page 267
Figures 45. Illustrated in Figure 4 is an example of a viable G. muris cyst isolated by FACS using the fluorogenic dye FDA. The cyst wall (arrowheads) and peritrophic space (ps) are easily distinguished. Within the cytoplasm, various organelles are seen including nuclei (n), flagellar axonemes (f), and portions of the adhesive disc (ad). Small vacuoles are present at the periphery of the organism. In Figure 5, an example of a nonviable cyst of G. muris isolated by FACS using the fluorogenic dye PI. In comparison to the viable cyst seen in Figure 4, the cyst wall (between arrowheads) appears to be increased in width and also is partially disorganized. The peritrophic space (ps) is present, but is less obvious than that seen in the viable cyst in Figure 4. In the cytoplasm, profiles of the adhesive disc (ad) and flagellar axonemes (f) are seen, with the latter also being present in the peritrophic space. Bar equals one micron for Figures 4 and 5.
450490 nm for FDA uptake (Figure 2b), and at 545546 nm for PI staining (Figure 2c). The identical field was again photographed by DIC and for fluorescence one hour later, six hours after exposure to FDA/PI (Figure 3). The comparison of FDA staining (compare Figures 2b to 3b) and PI staining (compare Figures 2c to 3c) indicated that, within one hour, the same cysts that underwent morphological changes and appeared nonviable by DIC criteria, also had incorporated the fluorogenic dye PI. Also, two of the nonstaining viable cysts became FDA positive during this same time period. Electron Microscopy of Viable and Nonviable Cysts The basis for structural differences in Giardia cyst morphology seen by DIC was examined at the ultrastructural level using FDA and PI stained Giardia cysts sorted by FACS. The morphology of FDA stained Giardia cysts, as illustrated in Figure 4, was characterized by the presence of typical cyst organelles, including nuclei, axonemes of flagella, dissembled portions of the adhesive disc, peripheral vacuoles, the peritrophic space, and a cyst wall composed of a fine fibrillar layer. The cyst wall was approximately 0.200.25 µm thick, and included two cyst membranes separated by a thin layer of cytoplasm. Cytoplasmic structures that corresponded to the DIC observations made of cyst morphology, were the welldefined cyst wall and the peritrophic space between cyst wall and cytoplasm. The PI sorted Giardia cysts, as represented in Figure 5, were quite distinct from FDA stained cysts in that the cyst wall appeared to be disorganized and to have increased in width. The ultrastructural changes in cyst wall morphology and the extremely diminished peritrophic space corresponded to the DIC image of PI stained Giardia cysts. The cytoplasm in PI stained cysts appeared less intensely stained (contrast of organelles to cytoplasm) than the FDA stained cysts, but otherwise, no obvious structural counterpart was observed that could explain the differences in cytoplasmic appearance seen by DIC, including the hyaline nature of FDA cysts versus the granular appearance of PI cysts. Discussion The fluorogenic dyes offer a rapid, direct measure of Giardia cyst viability based on the selective permeability of an intact cell membrane to the dyes FDA and PI. Figure 6 displays a diagram representing the proposed mechanism of fluorogenic dye staining in Giardia cysts. The fluorogenic dye, FDA, is a nonpolar molecule that is capable of traversing, by diffusion, an intact bilipid membrane. Upon entry into a cell, intracellular enzymes cleave the acetate groups off of FDA, leaving the polar fluorescein molecule within the intracellular compartment. The original FDA parent molecule is nonfluorescent, whereas, the hydrolyzed product, fluorescein, is highly fluorescent. Since the product in this reaction is polar the accumulation of fluorescence within the cell is favored. However, cells that do not possess an intact plasma membrane (or cyst wall membrane) are not able to retain fluorescein, and thus will not stain positively with FDA. In addition, it may be proposed that cysts without an intact membrane may have lost most of their original enzymatic activity and consequently will not be able to convert the fluorogenic substrate to the fluorescent product. The mechanism of staining for PI is quite different. The large planar phenanthridium ring structure of PI is unable to traverse the intact bilipid membrane by diffusion. Therefore, only cysts that have lost their membrane integrity will allow PI to enter the cytoplasm. Once inside the cyst, this fluorochrome may either intercalate into nucleic acids, or due to its hydrophilic nature, interact with other cytoplasmic sites (7). PI exhibits a slight fluorescence prior to interaction with nucleic acids, but upon intercalation the fluorescence intensity increases by approximately 100 fold (10).
Page 268
Figure 6. Schematic diagram of the mechanism for uptake of fluorogenic dyes by G. muris cysts. See text for explanation.
Lysosomal enzymes that have been proposed to cause hydrolysis of the ester linkage in FDA include nonspecific esterases as well as proteinases (16). The staining pattern of FDA in G. muris cysts typically showed initial staining in the region of peripheral vacuoles and it was not uncommon to observe intense staining for fluorescein in the peritrophic space. The lysosomal enzymes, acid phosphatase (8) and aryl sulfatase (Dennis Feely, personal communication) have been demonstrated in the peripheral vacuoles of Giardia and therefore they are considered to be lysosomal organelles. This, together with recent evidence that the peritrophic space was derived from lysosomallike peripheral vacuoles during cyst formation (Erlandsen et al., 61st Annual Meeting of the American Society of Parasitologists, abstract 95, 1986) has lent support to our proposal that these cellular sites may be the location of the hydrolysis of FDA. Ultrastructural examination of FDA and PI stained G. muris cysts sorted by FACS revealed that PI stained cysts (nonviable) did not possess an extensive peritrophic space and the cyst wall appeared to be partially degraded. The appearance of this cyst wall morphologically resembled a cyst wall after it has undergone degradative changes during excystation. Therefore, its appearance in PI stained cysts could reflect proteolysis resulting from damage to lysosomal compartments within the cyst. These changes would be compatible with the uptake of PI since membrane integrity was required for the exclusion of this fluorogenic dye from viable cysts. No ultrastructural correlation was found to explain the intracytoplasmic detail seen by DIC in PIstained cysts. Presumably, the unique morphology may have been related to differences in cytoplasmic refractive index caused by the loss of membrane permeability that could not be seen by electron microscopy. Previously, our laboratory (17,18) has shown that FDA and PI stained G. muris cysts had a distinct morphological appearance when viewed by bright field, phase, or DIC microscopy. Here, we have demonstrated that, when viable cysts have lost their membrane integrity and incorporated PI, the morphology of the cyst was converted from viable to nonviable within the shortest time interval studied (less than one hour). This seemed to indicate that the morphological appearance by DIC was closely coupled in time with dye uptake, therefore, either method would seem to be an accurate reflectance of cyst viability since the lag time between PI uptake and morphological change was small. Acknowledgements The authors wish to thank Ms. LeeAnn Sherlock and Ms. Andrea Toedter for their technical assistance, and Dr. W.J. Bemrick for his advice and critical review of the manuscript. Although the research described in this article has been funded in part by the U.S. Environmental Protection Agency through cooperative agreement No. CR811834 to the University of Minnesota, it does not necessarily reflect the views of the Agency and no official endorsement should be inferred. Literature Cited 1. Belosevic, M., Faubert, G.M., MacLean, J.D., Law, C., and N.A. Croll. 1983. Giardia lamblia infections in mongolian gerbils: an animal model. J. Infect. Dis. 147(2):222226. 2. Bingham, A.K., Jarroll, E.L., Meyer, E.A., and S. Radulescu. 1979. Giardia sp.: physical factors of excystation in vitro, and excystation vs. eosin exclusion as determinants of viability. Exp. Parasitol. 47:281291. 3. Bingham, A.K., and Meyer, E.A. 1979. Giardia excystation can be induced in vitro in acidic solutions. Nature (London) 277:301302. 4. Brenan, M., and C.R. Parish. 1984. Intracellular fluorescent labelling of cells for analysis of lymphocyte migration. J. Immunol. Methods 74(1):3138. 5. Cercek, L., and B. Cercek. 1972. Studies on the structuredness of cytoplasm and rates of enzymatic hydrolysis in growing yeast cells. I. Changes induced by ionizing radiation. Int. J. Radiat. Biol. 21:445453. 6. Coggins, J.R., and F.W. Schaefer, III. 1986. Giardia muris: ultrastructural analysis of the in vitro excystation. Experimental Parasitology 61:219228. 7. Cox, B.A., Yielding, L.W., and K.L. Yielding. 1984. Subcellular localization of photoreactive ethidium analogs in Trypanosoma brucei by fluoresence microscopy. J. Parasit. 70(5): 694702. 8. Feely, D.E., and J.K. Dyer. 1987. Localization of acid phosphatase activity in Giardia lamblia and Giardia muris trophozoites. J. Protozool. 34:8083. 9. Jackson, P.R., Pappas, M.G., and B.D. Hansen. 1985. Fluorogenic substrate detection of viable intracellular and extracellular pathogenic protozoa. Science 227:435438. 10. Jones K.W., and J.A. Sneft. 1985. An improved method to determine viability by simultaneous staining with fluorescein diacetatepropidium iodide. J. Histochem. Cytochem. 33:7779.
Page 269
11. Kasprzak, W., and A.C. Majewska. 1983. Infectivity of Giardia sp. cysts in relation to eosin exclusion and excystation in vitro. Tropenmed. Parasit. 34:7072. 12. Knight, D.P. 1977. Cytological staining methods in electron microscopy. In: Staining Methods for Sectioned Material. P.R. Lewis and D.P. Knight (eds). North Holland, Amsterdam. pp. 2976. 13. Ockleford, C.D., Hsi, B.L., Wakely, J., Badley, R.A., Whyte, A., and W.P. Faulk. 1981. Propidium iodide as a nuclear marker in immunofluorescence. I. Use with tissue and cytoskeleton studies. J. Immunol. Methods 43(3):261267. 14. Rice, E.W., and F.W. Schaefer. 1982. An improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709710. 15. RobertsThomsom, I.C., Stevens, D.P., Mahmoud, A.A.F., and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gasteroenterology 71:5761. 16. Rotman, B., and B.W. Papermaster. 1966. Membrane properties of living mamalian cells as studied by enzymatic hydrolysis of fluorogenic esters. Proc. Natl. Acad. Sci. 55:134141. 17. Schupp, D.G., and S.L. Erlandsen. 1987. A new method to determine Giardia cyst viability: correlation between fluorescein diacetate/propidium iodide staining and animal infectivity. Appl. Envir. Micro. 53:704707. 18. Schupp, D.G., and S.L. Erlandsen. 1987. Determination of Giardia muris cyst viability by differential interference contrast, phase, or bright field microscopy. J. Parasitol. 73:723729. 19. Shands, J.W. 1968. Embedding freefloating cells and microscopic particles: serum albumin coagulumepoxy resin. Stain technol. 43.15. 20. Szollosi, J., Kertai, P., Somogyi, B., and S. Damjanovich. 1981. Characterization of living normal and leukemic mouse lymphocytes with fluorescein diacetate. J. Histochem. Cytochem. 29:503510. 21. Thomas, J.A., Buschbaum, R.N., Zimmick, A., and E. Racker. 1979. Intracellular pH measurements in Ehrlich ascites tumor cells utilizing spectroscopic probes generated in situ. Biochemistry 18:22102218.
Page 271
PANEL DISCUSSIONS
Page 273
Panel Discussion on Excystation and Encystation Chairperson: Frances D. Gillin* Panel Members: E.A. Meyer, S. Erlandsen and C. Sterling Department of Pathology H811F, University of California, San Diego Medical Center, 225 Dickinson Street, San Diego, California 92103, U.S.A.. The Cyst Form and the Importance of Encystation and Excystation If Giardia trophozoites move ''downstream", they must complete their life cycle by encysting or they will die, since trophozoites do not naturally survive outside the host. In contrast, Giardia lamblia cysts are well adapted to survival and remain viable for months in cold water. Despite the importance of cysts, relatively little is known about this form, largely because it had not been induced in vitro until recently. Instead, investigators have relied upon cysts isolated from feces of infected humans or animals. Purification of fecal cysts is a lengthy and unpleasant process and some contaminants always remain. Moreover, both the numbers and quality of cysts seem to vary greatly from patient to patient and even from multiple isolations from a single donor. Therefore, the goal of this workshop was to summarize research to date on completing the Giardia life cycle in vitro. Excystation Infection with Giardia is initiated by ingestion of cysts from fecally contaminated water or food. Cysts pass through the stomach and into the small intestine where the newly emerged trophozoite immediately divides, giving rise to two daughters which can multiply, colonize the small intestine and cause symptoms. It is crucial that the trophozoite not emerge until the cyst passes into the duodenum, since it would be killed by gastric acid. Understanding excystation is important because interfering with this process would prevent infection. The first success in inducing G. lamblia cysts purified from feces to excyst in vitro was achieved in the laboratory of Dr. E.A. Meyer. The key observation was that excystation was triggered by exposure of cysts to low pH (~ 102M HCl). Trophozoite emergence occurs after transfer to a nutrient medium at neutral pH. This mimics exposure first to gastric acid, then passage into the duodenum. Dr. Meyer summarized published methods for achieving excystation in vitro. Studies from his laboratory focussed on the optimal time and temperature of cyst storage prior to excystation, as well as the pH and temperatures of the triggering and emergence steps. Modifications of the initial procedure (mainly at the EPA) identified other physiologic factors such as bicarbonate and reducing agents (glutathione, cysteine) which tend to improve the efficiency of excystation in vitro. Much remains to be learned about excystation, since the efficiencies observed with G. lamblia cysts from human feces are frequently low and highly variable, in contrast to cysts of G. muris. Moreover, G. muris cysts can excyst immediately after shedding, while excystation of G. lamblia cyst increases after a "maturation" period of several days at 4°C. Encystation Understanding of the process of encystation is less complete as this is a much more complex process of cellular differentiation in which the active, mobile trophozoite rounds up and secretes a new cyst wall around it. Moreover, neither the anatomic site nor the factors which induce encystation in the host had been clearly identified. Recent published and unpublished studies on inducing encystation in vitro were summarized by Dr.'s Erlandsen, Sterling, and Gillin. The approaches of each laboratory differed greatly, although a common motif was modification of the bile content of the medium. Dr. Erlandsen's group induced encystation of Giardia trophozoites from the muskrat. The cysts obtained were similar to those isolated from muskrat feces by differential interference as well as electron microscopy and uptake or exclusion of fluorogenic dyes (this volume). Dr. Sterling reported that the major goal of his group was to prepare G. lamblia cysts free of fecal contamination which could be utilized for isolation of cystspecific monoclonal antibodies (this volume). Their method was based on the hypothesis that large intestinal conditions, such as anaerobiasis, cellular crowding, and removal of water would be important in encystation. Dr. Gillin summarized quantitative studies from her laboratory which showed that large numbers of waterresistant cysts were induced by exposing cultured G. lamblia trophozoites to factors which are present in the human small intestine, such as primary bile salts and free fatty acids. Cyst specific antigens were demonstrated by Western blotting and increased activity of chitin synthetase was observed in encysting cultures. Dr. Gillin reported low levels of excystation and successful infection of suckling mice with in vitro derived human source G. lamblia cysts. Dr. Erlandsen reported excystation of in vitro derived cysts from his muskrat model, also at a low frequency. Therefore, much work remains in understanding the regulation of encystation, as well as the production of biologically active cysts. * Corresponding Author
Page 275
Panel Discussion on the Implications of Regulatory Changes for Water Treatment in the United States Stig Regli*, A. Amirtharajah, B. Borup, C Hibler, J. Hoff, and R. Tobin. Office of Drinking Water, U.S. Environmental Protection Agency, 401 M Street, S.W., Washington, D.C. 20460, U.S.A.. Introduction EPA intends to propose and promulgate Surface Water Treatment Requirements (SWTR) in 1987 which will: (a) provide criteria by which state regulatory agencies will determine which systems using surface water sources will be required to filter, (b) set disinfection treatment requirements, and (c) regulate for Giardia lamblia, viruses, heterotrophic plate count bacteria, Legionella, and turbidity. The proposed criteria will specify minimum treatment requirements for achieving at least 99.9% removal and/or inactivation of Giardia lamblia and 99.99% removal and/or inactivation of viruses by all surface water systems. Criteria for determining whether systems will be required to filter include raw water quality limits (turbidity, and total coliforms or fecal coliforms), maintenance of an effective watershed control program, disinfection which achieves a theoretical 99.9% inactivation of Giardia lamblia and 99.99% inactivation of viruses (as determined by the product of concentration [C, mg/L] multiplied by disinfectant contact time [T, minutes], i.e., CT values), compliance with the total trihalomethane Maximum Contaminant Level (MCL) for systems serving greater than 10,000 people, and compliance with a longterm total coliform standard for distribution system monitoring, also to be proposed at the same time as the SWTR. No requirements for Giardia monitoring are specified in the draft rule. EPA considered the use of a risk model (18,6) in the development of the proposed SWTR. Application of the risk model in the rule would allow utilities to provide less stringent disinfection (i.e., lower CT values) than that necessary to achieve the minimum 99.9 and 99.99 percent inactivation of Giardia lamblia, and viruses, respectively, if it could be demonstrated that the population served would not be exposed above an acceptable level of health risk. In the risk model it is assumed that an annual infection rate could be estimated as a function of the number of raw water samples taken, total sample volume, percent recovery by the analytical method, number of Giardia cysts and viruses detected, percent theoretical inactivation by disinfection, amount of water consumed per person per day, and the viability and infectivity of the organism. EPA decided not to include the risk model in the proposed regulation because: precision, efficiency, and sensitivity of any analytical method for measuring Giardia lamblia cysts have not yet been adequately defined; no standard methods, validation procedures, or laboratory certification procedures are available for assuming confidence in the analytical methodology; very large numbers of samples would be needed to ascertain that a system was providing water below a reasonable level of health risk; and disinfection data on which to base theoretical inactivation is very limited. The panel members discussed some of the issues that would need to be addressed before a risk model could be applied to regulate Giardia lamblia. Discussion A Possible Sampling Approach for Evaluating the Level of Risk (M.B. Borup) To develop a sampling plan that will be used to evaluate the quality of a product it is frequently standard procedure to specify two values of the mean number of defects per unit of product, say m0 and m1 where m0 < m1, and two small probabilities, say a0 and b1. If the mean number of defects per unit of product is equal to m0, the product is regarded as satisfactory and it is desirable to conclude it is satisfactory with some high probability, say at least 1a0. If the mean number of defects per unit of product is equal to m1, the product is unsatisfactory and it is desirable that we conclude the product is satisfactory with some very low probability, say no more than b1. The sampling plan should then be designed so that these criteria may be met. The number of samples required will depend on the values of m0, m1, a0 and b1. In the case of waterborne pathogens, m0 and m1 are the number of pathogenic organisms found per sample volume of water. The number of pathogens found in the raw water may be related to the number of infections per 10,000 people per year in the following manner:
where: mi represents either m0 or m1 and is equal to the average number of organisms per raw water sample volume. Ei = the number of infections per 10,000 per year corresponding to mi. Vs = sample volume. R = the fraction of organisms that are recovered by the analytical techniques used. V = volume of water consumed, 10,000 people/year G = the fractions of pathogens removed or inactivated in the water treatment process. * Corresponding author
Page 276
This equation is based on several simplifying assumptions. First, it is assumed that one organism can cause infection. Second, it is assumed that each organism will be consumed by a different individual. These assumptions will present a worst case and produce a conservative estimate, protecting public health. If y represents the total number of pathogenic organisms found in a series of samples taken from the raw water the objective of a sampling plan may be expressed in the following requirements: P (y c) 1 a0 when mi = m0 P (y c) b1 when mi = m1 where: P(y c) represents the probability that the number of pathogens found in a series of samples is less than or equal to c. c = the maximum allowable number of pathogens. To determine the number of samples required it is necessary to determine a statistical distribution that will describe the occurrence of pathogenic organisms. Very little information is available that can be used to determine a suitable statistical model for Giardia occurrence. A model which is often used to compute probabilities associated with the number of defects per sample is the Poisson Model. The largest body of data on Giardia occurrence (9) does not prompt the rejection of this model. Using a relationship between the Poisson distribution and the Chisquare distribution it can be shown that the following equation must be satisfied to meet the criteria described above:
where: 2
= chisquare probabilities.
n = number of samples required. Consider this example of the application of this equation. A utility would like to determine if their treatment process is adequately protecting the public health from giardiasis. It has been determined that water quality which causes less than or equal to one infection per 10,000 people per year is acceptable and that it is desirable to accept this water as satisfactory at least 90% of the time. It has also been determined that water quality which causes five infections per 10,000 people per year is unacceptable and this water should not be determined to be acceptable more than 5% of the time. Tests have shown that the treatment process used by this hypothetical utility removes or inactivates 99% of the Giardia cysts in the raw water, but the analytical method used to identify the cysts will only identify 50% of the cysts actually present in the sample. From this information: a0 = 0.10 b1 = 0.05 G = 0.99 R = 0.50 assume: 1) Vs = 500 L 2) Each person drinks 2 L water per day 3) Ingestion of one cyst will cause infection 4) Each cyst present in the treated water will cause infection in a different person then: V = 10,000 people x (2 L/(person.day)) x 365 days/year = 7,300,000 L m0 = (1 x 500 x 0.50)/(7,300,000 (1 0.99)) = 0.003425 m1 = (5 x 500 x 0.50)/(7,300,000 (1 0.99)) = 0.017125 Now the following equation must be satisfied by the selection of n and c:
This equation is satisfied when c = 3 and n 453. This means if the utility takes 453, 500 L samples of the raw water and they find less than 4 cysts, they can be 95% sure the water will produce less than 5 giardiasis infections per 10,000 people per year. Using this same technique the number of samples of the raw water required under given conditions are listed in Table 1. For example, if a system achieved 99.5% inactivation by disinfection, and a 50% recovery in the analysis for Giardia cysts, it would need to collect 1,132 raw water samples, each of 500 liters, per year and determine that there were no more than 3 Giardia cysts detected for the treated water to be of acceptable quality. The number of samples that would be required to demonstrate that a reasonable level of risk was being avoided, according to the preceding analysis, makes application of the risk model cost prohibitive. Questions/Comments Regli: If different assumptions were made than those discussed, the number of samples that would be needed to demonstrate that an acceptable risk level was being met could be significantly reduced. For example, if the average sample volume were 5,000 L versus 500 L, as appears to be possible in very low turbid waters, the number of samples that would be needed could be reduced by an order of magnitude. Also, the assumption that all cysts are viable and would cause infection if ingested is very conservative. Under a less conservative assumption, fewer samples would be required. Borup: Yes, this is true. The assumptions that were made were requested by EPA to allow for a conservative analysis.
Page 277 TABLE 1. Number of 500 L samples (n) required to determine the acceptability of a treated water.*
m0 = m1/5
m0 = m1/10
Recovery %
Removal %
m1
a0=0.1 c
b 1=0.05 n
a0=0.1 c
b 1=0.1 n
a0=0.10 c
b 1=0.05 n
a0=0.10 c
b 1=0.1 n
75
90
0.0005137
3
15,094
4
22,590
50
90
0.0003425
3
22,638
4
33,882
1
9,235
2
16,364
1
13,851
2
30
90
0.0002055
3
37,730
4
24,543
56,470
1
23,086
2
75
99
0.005137
3
1,510
40,906
4
2,260
1
924
2
50
99
0.0034247
3
1,637
2,264
4
3,389
1
1,386
2
30
99
0.0020548
2,455
3
3,773
4
5,647
1
2,309
2
75
99.5
4,091
0.010274
3
755
4
1,130
1
462
2
50
819
99.5
0.0068493
3
1,132
4
1,695
1
693
2
1,228
30
99.5
0.0041096
3
1,887
4
2,824
1
1,155
2
1,046
m1 = average number of cysts per 500 L sample of raw water. E1 = number of infections per 10,000 people per year corresponding to m1 based on the following: 1) Vs = 500 L 2) one cyst causes infection 3) each cyst will be ingested by a different individual 4) each person drinks 2 L water/day * Calculations of m1 are based on an E1 value of one infection per 10,000 people per year.
Regli: One more point for clarification. In considering application of the risk model in the rule, it was not EPA's intention to require all systems, wishing to avoid filtration, to conduct Giardia monitoring. Rather, the purpose of the risk model was to offer utilities an alternative to meeting the minimum treatment performance requirement of 99.9 percent inactivation of Giardia cysts in order to avoid filtration. If a large utility had inhouse monitoring capability it might be able to demonstrate, through monitoring the raw water, that a 99% level of inactivation would achieve a finished water quality below the acceptable risk level, say less than 1 infection per 10,000 people per year. A lower level of disinfection could save costs and minimize disinfection byproduct formation, thereby reducing health risk from carcinogens. Tobin: In your analysis, Dr. Borup, you made the assumption that Giardia cyst occurrence could be characterized as a Poisson distribution. Could you elaborate on the basis for this? Borup: There were very limited data on which to base this assumption. However, a Poisson distribution appears reasonable based on histograms of raw water data gathered by Dr. Hibler which included over 600 samples from 20 sites, including lakes, rivers, and creeks. Considering these data as a whole, one cannot reject that the data fit a Poisson distribution based on the ChiSquare hypothesis test. Determination of an Allowable Morbidity Rate (R.S. Tobin) Risk assessment has become an integral part of drinking water regulation and guideline establishment. According to the scheme proposed by the World Health Organization (28), there are four main components of risk assessment: hazard identification, risk estimation, risk evaluation and risk management. The establishment of a policy for an "acceptable" level of risk of infection by Giardia falls within the area of risk management. Previous studies have indicated that certain risks are more acceptable to the public than others (21). Some of the factors involved in the risk of giardiasis from potable water mitigate against a high degree of risk acceptance by the public. For example, the risks due to Giardia may be considered as involuntary risks, immediate risks, new risks, and risks that individuals have to pay for themselves (to some extent) in order to remedy. On the other hand, the risks which result from nonsecret (open) activities, are rather diffuse risks and result from known natural causes. Experience and common sense informs us that there are some upper and lower limits to risks that can be termed "acceptable" (Table 2). For the risk of death it has been suggested that, under certain conditions, individuals would accept an upper limit of risk of 103, whereas the negligible risk of death is considered between 106 and 107 (21). TABLE 2. Suggested upper and lower limits or risk to an individual (21) Upper Limit of Risk Imposing a continual annual risk of death of 1 in 100 should be described as unacceptable (1,000 per 100,000). An imposed risk of 1 in 1,000 (100 per 100,000) is not totally unacceptable, provided the individual knew of the situation, felt he had some benefit as a result, and knew that everything possible had been done to reduce the risk. Negligible Level of Risk Few people would commit their own resources to reduce an annual level of risk that was already as low as 105, and even fewer at 106 The figure of 106 is probably an appropriate negligible risk except for clear causal relationships with consumer products.
Page 278 TABLE 3. Waterborne outbreaks of giardiasis in the United States. Year
Outbreaks
Cases
Cases/100,000**
1972
4
124
0.06
1973
4
73
0.04
1974
4
4,930
2.47
1975
1
9
0.005
1976
3
639
0.32
1977
4
1,012
0.51
1978
4
5,171
2.59
1979
7
5,864
2.93
1980
8
1,730
0.87
1981
11
311
0.16
Mean:
4
1,986
0.99
* Modified from Craun (4) ** Approximation, based on U.S. population of 200 × 106.
A review of some of the Giardia statistics in North America reveals that the known waterborne outbreaks in the U.S. over a 10year period results in a risk of about 105 (Table 3). The total (waterborne plus nonwaterborne) disease in the U.S. is not known. On the other hand, data are available for total giardiasis in Canada (Table 4). Summary data from the years 19831986 revealed an average Canadawide risk of about 30 × 105. If we assume that waterborne giardiasis causes approximately 10% of this figure, the risk becomes 3 × 105, relatively close to the U.S. value. It is also instructive to compare the current level of risk posed by agents other than Giardia. Gerba (7) has completed a risk assessment of viruses in potable water, based on the stringency of proposed viral standards (Table 5). If a limit for viruses in potable water was set at 1 plaqueforming unit per 1,000 L, the annual risk of infection would be 1.5 × 103, or 150 cases per 100,000 persons per year. Based on the same virus limit, the predicted annual risk of death from hepatitis A virus was estimated as 1.1 × 105 Obviously, these risks would be shifted upward or downward, depending on the virus standard established. Another area where microbial risk has been quantified is for bathing beaches. Epidemiological studies have been used by U.S. EPA to develop recommendations for bathing water quality (Table 6). The criteria for fresh water predict an illness rate of 80 per 105 bathers and for marine waters, they predict 190 per 105 bathers. (8) predict between 120 and 1,500 illnesses per 105 bathers, based on the same epidemiological evidence. TABLE 4. Total giardiasis in Canada. PEI
NS
NB
QUE
ONT
MAN
SASK
ALB
BC
YT
NWT
CDA
rate/105*
Year
NFLD
1983
82
21
67
0
22
2,942
112
456
812
5
43
4,562
22.8
1984
51
11
107
7
275
3,692
258
1,362
895
16
36
6,710
30.5
1985
47
14
111
5
349
3,713
220
1,526
1,244
22
17
7,268
33.0
1986**
54
23
97
6
388
3,382
222
1,244
1,980
6
29
7,431
33.8
6
* Based on a population of 22 × 10 . ** Estimate based on data to August 30, 1986.
TABLE 5. Estimated risks of infection due to low levels of virus in drinking water. Standard of 1 PFU per
Annual Risk
Predicted Cases per 100,000/yr.
10 L
1.5 × 101
15,000
100 L
1.5 × 102
1,500
1,000 L
1.5 × 103
150
10,000 L
1.5 × 104
15
Based on Gerba (7)
When all of these microbiological risk data are summarized (Table 6), the example of 10 allowable waterborne illnesses per 100,000 people per year due to Giardia (18) appears to be within the "expected" risk range. This value would be about onethird the current total Giardia illness rate in Canada (or about three times the current best estimate of illness due to waterborne Giardia in Canada) and about ten times the annual average waterborne disease cases in the U.S. It is, perhaps, close to the current risk levels in areas in North America where Giardia are prevalent in the water supplies. This risk level is, however, more protective than currently accepted recreational water quality guidelines. It is acknowledged that recreational bathing involves a voluntary choice where the individual may accept a certain risk which is balanced by the perceived benefits. In conclusion, a risk level of 10 per 105 persons per year is not out of proportion with other common microbiological risks. Whether or not it would be the most appropriate or "acceptable" level of risk still requires careful consideration. Questions/Comments Borup: Could you elaborate on the acceptable virus risk level of 150 infections per 100,000 people per year which you mentioned had been proposed. Shouldn't an acceptable Giardia risk level be in line with an acceptable virus risk level? Tobin: I was not suggesting that the risk level of 150/100,000 for viruses had been proposed, only what the risk would be relative to a finished water concentration of 1 virus/1,000 liters. If a standard were set at 1 virus/10,000 liters the risk would be 15/100,000 people per year. Audience: How would you compare the same risk level demonstrated by a public water system which filtered and disinfected versus a system with a protected watershed which only disinfected?
Page 279 TABLE 6. Current and proposed microbiological risks from water. Predicted illness per 100,000 persons
Criteria Recreational water:
126 E. coli or 35 enterococcus/100mL fresh water
80a
35 enterococcus/100mL, marine water
190a
200 fecal coliforms
120 1,500b
Drinking water:
10 coliforms/100mL
unknown
1 virus/1,000L
150 (infections)c
1.1 (deaths, HAV)c
Giardia in drinking water
10 ?
a
(26)
b
(8)
c
(7)
Tobin: Good question. I haven't addressed that issue. Do you set the risk for a given system based on the average risk expected across the U.S. or do you consider different systems being at much greater risk than others and regulate accordingly? Allowable Concentrations of Cysts in Raw and Finished Water (A. Amirtharajah) It is evident from recent studies that there exists at least two avenues towards establishing relationships between cyst levels in the raw and finished water from treatment plants. One possible approach is to establish the cyst concentration in the finished water on the basis of a given risk level of infections per 10,000 people per year and back calculate the concentrations in the raw water by using the expected removals and inactivations of cysts through a water treatment plant. This would then lead to estimates of raw water sample size and number of samples to meet the risk level specified. The validity of this approach needs to be confirmed by empirical results. An alternate approach is to establish a significant monitoring program for the raw and finished water cyst levels in a variety of water treatment plants, which would cover some episodes of giardiasis and hence determine the allowable cyst levels in the finished water. The underlying premise of this approach is that a system which has no history of disease outbreak, or record of increased Giardia incidence has an acceptable finished water Giardia cyst concentration. The latter approach is in fact an embryonic stage for the future development of a maximum contaminant level (MCL) for Giardia. In both approaches the estimated removals and inactivations of cysts through the sequence of processes in a water treatment plant is the essential link between raw water concentrations and finished water levels. Amirtharajah (1) has presented a review of the literature which summarizes the expected transformations of Giardia cysts in water treatment processes. Figure 1 illustrates the expected transformations (removals and inactivations) of contaminants in a conventional plant and Table 7 summarizes extensive data from pilot plant studies for the removal efficiencies of Giardia by water treatment processes. Table 7 indicates that filtration with good coagulation will remove 95 to 99.9% of the cysts, while poor coagulation or no coagulation prior to filtration will drop the removal efficiency to low levels of 10 to 70%. Including the effect of disinfection, the anticipated total removal and inactivation is 90 to 99.9% or greater, depending upon the level of disinfection provided. These transformation efficiencies may be used to couple the finished water cyst concentrations to the raw water cyst levels. Hibler (9) has collected extensive data on the occurrence of Giardia lamblia cysts in raw and finished waters of treatment plants. His data set includes approximately 6,000 samples, analyzed for 302 cities over a period of 10 years. The frequencies of sampling ranged from 12 samples per month to 510 samples per week. Hibler (9) has recently analyzed these data. A sample of his data was extracted and rearranged as shown in Table 8. Some general conclusions on cyst concentrations in the raw and finished waters are clearly evident in the data shown in Table 8. 1. The fraction of samples with no detectable cysts in the raw water sources range from 0% to 90.5% with a typical value of approximately 70%; i.e., typically 30% of the samples analyzed were positive. Therefore, cysts seem to be ubiquitous in surface water sources in the country. 2. The levels of cysts in raw waters are commonly 010 per 100 gallons. 3. In well operated conventional plants the cyst concentrations in the finished waters were zero in all of the samples tested.
Figure 1. Estimates of contaminant transformations in a "standard system".
Page 280 TABLE 7. Removal efficiencies of Giardia lamblia by water treatment processes Raw water concentration
Unit proces
Percent removal
Operating parameters
Reference**
1. Rapid Filtration with coagulationsedimentation
231,200/L
96.999.9
min. alum = 10 mg/L opt. pH = 6.5 filt. rate = 4.99.8 m/hr
DeWalle et al. (1984)
2. Direct Filtration
20 × 106
95.999.9
min. alum = 10 mg/L pH range = 5.66.8
DeWalle et al. (1984)
with coagulation
(as slug)
48
filt. rate = 4.99.8 m/hr
no coagulation
eff. NTU/inf. NTU = (0.020.5)/(0.71.9)
eff. poor during ripening
with flocculation
9599
alum = 25 mg/L polymer (Magnifloc 572 CR) = 12 mg/L temp = 5°C, 18°C eff. NTU/inf. NTU = 0.05/1.0 Filt. rate = 4.818.8 m/hr
no coagulation
1070
5
5
AlAni et al. (1984)
3. Diatomaceous Earth
1.5×10 9.0×10 /L 9999.99
filter aid = 20 mg/L body feed
DeWalle et al. (1984)
Filtration
102104
>99.9
filt. rate = 2.4 9.8 m/hr temp = 5° 13°C eff. NTU/inf. NTU = (0.130.16)/(1.0 2.0)
Bellamy et al. (1984a)
4. Slow Sand
50 5×103/L
100
filt. rate = 0.04 temp 0.4 m/hr = 0°C, 5°C, 17°C eff. NTU/inf. NTU = (37)/(410)
Bellamy et al. (1984b)
* Amirtharajah (1) ** These references are given in Amirtharajah.
4. Even conventional plants are not fail safe and would pass cysts, under poor operating conditions and/or when cyst levels increased to numbers in the range of 20 to hundreds per 100 gallons in the raw water. 5. Rapid gravity filtration without coagulation invariably passed cysts into the finished water. 6. Unfiltered supplies with chlorination ond no filtration often passed cysts at levels of 010 cysts per 100 gallons to the final water. An important conclusion from the results presented in Table 8, i.e., from plant scale data, is that it generally corroborates the data in Table 7 derived principally from pilot plant studies, and some of the assumed removals in Figure 1. Some very tentative generalizations on allowable cyst levels in the raw and finished waters may be made on the basis of the available data. A well operated conventional system, with a sampling program of one sample per week on raw water, should commonly detect 2030 percent of samples positive. If the cyst levels in these samples are less than 10 per 100 gallons then the risk from such a system is very small. If the cyst levels are greater than 10 for 100 gallons in the raw water, an immediate analysis for possible weaknesses in the treatment train should be initiated. Most unfiltered systems have finished water cyst levels of 010 per 100 gallons in the samples that were positive and higher levels in the raw water. Often only 510% of the finished water samples were positive. These removal rates may be attributed to disinfection and sedimentation since many samples were drawn from lakes. However, if the percentage of positive samples was only 2.5% (i.e., 97.5% negative as in Table 8) of the samples tested, these cities seem to have been historically safe from an outbreak of giardiasis. If the cyst levels reach into the tens or hundreds per 100 gallons, then the risk is significantly increased. If we assume that contaminated raw water supplies, with cyst concentrations as high as 100 cysts/100 gallons (or 100 cysts/378 L) would receive adequate treatment if removal removal/disinfection efficiencies were 99.999% (since systems with clean water sources must achieve 99.9% removal it would be reasonable to expect systems with contaminated water to achieve significantly greater removals), then acceptable finished water concentrations would be 100 × 105 cysts/378 L or 2.6 × 106 cysts/L. Then, if a system were able to achieve 99% inactivation by disinfection alone, an acceptable raw water concentration for such a system would be 2.6 × 104 cysts/L. This concentration level is higher than that determined by the risk model but falls within the same order of magnitude. According to the analysis presented in Table 1, if a system achieved 99% removal of Giardia cysts by treatment (e.g., by disinfection alone) it would produce an "acceptable" treated water if the average concentration of Giardia cysts in the source water was less than 1.36 × 105 cysts/L. i.e:
Page 281
The empirical data present some tentative possibilities for estimating allowable cyst concentration levels, but these need to be confirmed by an analysis of all of the data collected by Hibler (9) and additional data as they become available. In addition, the assumptions of the model need to be reassessed on the basis of the available data. Infectivity and Viability (J.C. Hoff) In determining the effects of various environmental conditions and exposure to disinfectants on Giardia cysts, it is important to consider the methods used to determine whether or not the cysts have been killed. It is important to note that the terms infectivity and viability have different meanings. Infectivity refers to the ability to produce an infection in a host, with or without the production of disease. Viability is a broader term that includes infectivity but also includes measurement of other parameters related to being alive such as physiological activity, respiration, reproduction, excystation and dye exclusion. The latter two methods, in vitro excystation and dye exclusion, have been the predominant methods used to determine Giardia cyst viability. Early studies, using dye exclusion, such as those of Mills et al. (17), indicated that cysts were extremely resistant to inactivation by disinfectants. Following the development of in vitro excystation methods for G. lamblia (2,20); and G. muris (22) and research indicating that dye exclusion did not correlate with excystation (2), in vitro excystation became the predominant method used for measuring cyst viability. Newer methods based on the use of fluorogenic dyes such as fluorescein diacetate and propidium iodide (23) and boronic acid derivatives (13) are being developed but little quantitative testing of their correlation with excystation or animal infectivity has been done. Relatively little research on the minimum number of Giardia cysts required to infect animals has been done. In an early study, Rendtorff (19) studied the minimum number of G. lamblia cysts required to infect humans. Based on these studies, he concluded that one cyst was capable of causing infection in a susceptible host. Hibler (10) determined that <8 G. lamblia cysts produced infection in 80 to 100% of Mongolian gerbils fed cysts. Using G. muris cysts, Hoff et al. (11) reported that ID50 for mice ranged from <1 to 16 cysts. While the data are limited, it is evident that very few cysts are required to produce infection in man or other animals. In the same study Hoff et al. (11) showed that excystation and infectivity assays gave similar results in indicating the degree of inactivation caused by exposure of G. muris cysts to chlorine. Infectivity and Viability (C.P. Hibler) Interpretation of infectivity and viability of Giardia cysts must be done with caution. Our laboratory has made a number of indirect and direct observations on Giardia cysts from surface water sources. Examination of 4,423 samples of water from 301 municipal sites, has revealed that 102 sites often had cysts present in finished water samples. Only 5/102 (5%) of these sites had an epidemic of waterborne giardiasis at the time of sampling. Several of the 97 sites often had as many cysts/gallon as those sites in the midst of an epidemic. Most of the 97 sites had adequate TABLE 8. Cyst levels in raw and finished watera
No. of Samples Positive % Samples Negative
Treatment
City No.
Area of Country
Source
Total No. of Samples Raw
Finished
010
Raw
1020
(Cyst levels #/100 gal.) >20 010 1020
Finished
>20
Conventional
20
W
Lake
112
83
100
9
1
0
0
0
0
32
W
River/Lake
43
86.4
100
3
0
0
0
0
0
93
NE
River
139
76.6
93.5*
3
0
16
1
3
0
109
NE
Lake/Creek
139
70.8
100
20
1
0
0
0
0
M
NE
River
87
0
66*
11
10
22
10
0
0
Rapid Gravity
8
W
Creek
95
52.4
70.3*
7
1
2
17
2
3
Filtration
74
W
Creek
149
27.3
44.7*
20
7
13
27
8
17
Without
152
W
River
100
52
90
13
10
1
5
0
0
Coagulation
Unfiltered
40
SW
River
134
50.8
97.3
29
1
1
0
2
0
41
SW
Creek
124
70.7
92.8
9
0
3
6
0
0
89
NE
Lake
206
78.9
87*
20
3
1
8
2
2
90
NE
Lake
449
90.5
96.4
16
2
1
9
0
0
91
NE
Lake
428
89.9
97.5
8
1
0
6
0
0
120
NE
Lake/Creek
102
96.3
89.6*
2
0
0
5
0
0
152
W
River
142
52.0
90
13
10
1
5
0
0
a All data presented in the table are from Hibler (7), except for city M from Sykora's (22) studies. Only data with both raw and finished water cyst levels are included inthe table. * These cities have had outbreaks of giardiasis.
Page 282
disinfection; a few did not, yet there was no reported increase of cases above the normal "background level" in these municipalities. Microscopic evaluation of cysts from some of these municipalities indicated cysts possibly were alive; consequently on 6 occasions, they were introduced into 10 Mongolian gerbils. The number of cysts introduced ranged from 1 to 10 cysts/mL/gerbil. In 4/6 (67%) attempts, 25 to 50% of the gerbils became infected with Giardia. On two occasions the amount of concentrate available necessitated an attempt with 1 to 2 cysts/mL/gerbil and previous experience had shown that this number of cysts would only establish infection in about 20% of the gerbils. None of the animals in trials with 1 to 2 cysts became infected. The only successful trials involved use of 5 to 10 cysts/mL/gerbil. The results showed that cysts were alive and infectious for Mongolian gerbils from 4/6 municipalities that did not have adequate disinfection, indicating that the cysts were infectious for humans. However, there are two factors to be considered: (1) Mongolian gerbils are susceptible to only 50 to 60% of the human sources to which they are exposed, suggesting the cysts may have been infectious for humans but not for gerbils; and (2) the cysts may have been responsible for the "background level", especially in small municipalities (<400) where most of the population was already immune. Infection of Mongolian gerbils, or any animal model with cysts from a surface water source, means only that those cyst were alive and infectious for the animal model; no other interpretations are possible. If infection is not established, the results suggest: (1) cysts were dead; (2) cysts were not infectious for the animal model; or (3) the number of cysts was inadequate to establish an infection. The results from use of animal models does not provide any useful information regarding the susceptible human population. A few municipalities from which we have received samples had enough cysts in the finished water to have caused a major epidemic. These cysts were found in the first samples ever received from these sites; undoubtedly they had been present for some time. While no efforts were made to infect animals and determine if cysts were alive or dead, microscopic evaluation indicated they were alive. Despite our lack of proof, and the cautions and/or precautions we must exercise before concluding that these cysts were not infectious for humans, the evidence certainly suggests they were not infectious because the chlorine levels and the contact time before the first customers was adequate. Tests for viability using an animal model must incorporate the constraints given for infectivity; human sources are not uniformly infectious for animal models. Microscopic evaluation requires considerable skill and experience and most microscopists do not see enough cysts from different animals and/or sources to develop this skill. Even experienced microscopists generally cannot recognize dying or recently dead cysts, especially if they have been dead only a short period of time. The eosinexclusion test and excystation are not adequate: eosin is unreliable and there are not enough cysts for excystation. Currently efforts are underway to develop an immunofluorescence test for viability. Likely such a test will have the same problems faced by the experienced microscopists; dying or recently dead cysts will appear to be alive. Infectivity and viability of Giardia cysts from surface water sources will continue to be a problem for analysts until the entire host range of Giardia from human and animal sources, has been elucidated and clarified. It is unlikely that monoclonal antibodies will distinguish between human and animal sources. There certainly are "strains" present in humans that will not infect animals, and there are "strains" in animals that will not infect humans. There may be strains in humans that will not infect other humans. Until more information is available, we must assume that Giardia cysts present in surface water possessing morphologic features compatible with G. duodenalis are potentially infectious for humans. Use of CT Values in Estimating Giardia Cyst Inactivation (J.C. Hoff) CT values represent the product of disinfectant concentration (C) and contact time (T) required to cause inactivation of a certain percentage (e.g., 99%) of a specific target pathogen by a specific disinfectant under specified pH and temperature conditions. The use of CT values is based on analogies between chemical reaction rates and microbial inactivation rates. In determining CT values to be used in establishing disinfection requirements for surface waters, the target pathogen is G. lamblia cysts, because of the high degree of resistance shown by the cysts. Based on the numbers likely to be found in water and their low infectious dose, a level of 99.9% inactivation has been determined to be adequate for control of waterborne giardiasis. The basis for the use of CT values to establish disinfection requirements, CT values for different types of pathogens, and some of the problems associated with application of this approach to establishing disinfection requirements, were reviewed by Hoff (12). A major difficulty is the limited data base of disinfection data available and uncertainties associated with extrapolations from the available data base to other pHs, temperatures, and disinfectant concentrations. Another consideration, especially with regard to cyst data, is that because of the assay method used to determine viable vs. nonviable cysts (microscopy) and limitations on the numbers of cysts that can be observed, these data are limited to determining values at the 99% level. Extrapolation is required to estimate times needed for 99.9% inactivation and because the curves do not follow first order kinetics, the calculated CT values at this level are less reliable. Hibler (10) has developed data showing CT values for 99.99% inactivation of G. lamblia cysts using Mongolian gerbils. These data should aid in establishing the needed CT values. Measurement of CT Values on Site (A. Amirtharajah) The extrapolation of laboratory CT values to field conditions requires a critical assessment of concentration of disinfectant, C in mg/L and contact time, T in minutes. The effective concentration is defined on the basis of its value, shortly before or after it reaches the first customer
Page 283
or shortly before a second disinfectant dose is applied. Contact time in pipelines can be calculated assuming plug flow conditions and dividing the length of the pipeline by the mean velocity of flow. This computation must be determined for peak flow conditions. The contact time in chlorine contact tanks is best estimated on the basis of tracer studies. Two tracers that are convenient to use, present no adverse health effects and are convenient for chemical analyses are lithium chloride or sodium fluoride. Fluoride is a convenient tracer since it is already used in water treatment plants and most plant laboratories have equipment for its analysis. In instances where fluoride is used, background corrections for fluoride in the water prior to disinfection need to be made. Several studies (15,25) have shown that design and evaluation of chlorine contact chambers are usefully evaluated on the basis of dispersion numbers. The tracer study will yield a dimensionless dispersion number, d for the chlorine contact tank. The ideal reactor classified as a plug flow reactor has a dispersion number of zero. The flow in pipelines approximates a reactor with zero dispersion number. The opposite extreme of ideal reactors, the complex mix reactor, has a dispersion number of infinity. The effluent to influent ratio Ci/Cio is plotted as a function of dimensionless time, in Figure 2a and illustrates the characteristic tracer curves for different dispersion numbers. The value of is calculated by dividing the different times of sampling, t in the tracer study by the volumetric displacement time t of the contact tank [t = (volume)/(rate of flow) = V/Q]. The tracer (lithium chloride or sodium fluoride) is normally added as a slug at the inlet end of the contact tank at the dimensionless time =0 as shown in Figure 2a, so that it produces a concentration Cio, if dispersed instantaneously in the volume of the tank. The effluent concentration Ci at various times t is measured and a dispersion curve as shown in Figure 2b is drawn. The dispersion curve that is generated is then compared with the typical curves shown in Figure 2b and an estimate is made of the dispersion number. The dispersion number can also be calculated from equations given in several references (27,25). Trussell and Chao (25) suggest that dispersion numbers can be alternatively estimated from length to width (L/W) ratios of the chlorine contact chamber by using the following equation:
in which K' = coefficient of a nonideality with values from 1 to 15, L = length of contact tank, W = width of contact tank. A conservative approach will assume K = 15, while chambers with good hydraulic design and approximating plug flow configurations, have K values of 3 to 5. By using the calculated or measured dispersion number and the curves in Figure 2a it is possible to make reasonable estimations of contact time T that can then be
Figure 2. Tracer responses in contact tank with a pulse input (after Weber, (27).
used to evaluate CT values. Figure 2a shows that contact tanks with dispersion numbers less than 0.025 are closer to plug flow conditions than to complete mix conditions. The midpoint of the increasing limb of the tracer curve could be used as a reasonably safe first approximation for the detention time T to be used in CT calculations. If contact tanks have dispersion numbers greater than 0.1, then it implies significant short circuiting in the tank and the effectiveness of disinfection may be improved by rearrangement of the flow conditions within the tank with additional baffles, etc. It is possible to combine the plug flow dispersion characteristics of a contact tank with the inactivations at various concentrations and contact times and hence make a very good estimate of the overall disinfection efficacy for viruses and Giardia cysts. This approach is very valid if disinfection kinetics can be modeled by well defined reaction kinetics. For a first order reaction with rate constant k, Figure 3 developed by Levenspiel and Bischoff (14,15) may be used for analysis. As an example, for a kT value of 10 and a dispersion number of 0.25, Figure 3 shows that (Ci)/Cio = 0.0065. Therefore, the inactivation rate of organisms is (1 0.0065) 100 = 99.65%. Additional details of the methodology for completing these computations are discussed in the references noted (14,15,27,25) These computations are not routinely made and for purposes of a conservative approach to assessment of CT values and disinfection efficacies a single lower bound value for contact time as recommended may be used.
Page 284
TABLE 3. Comparison of real and plugflow reactors for first order reactions (after Levenspiel and Biscchoff, 14,15).
Minimum Sampling (C.P. Hibler) Sampling efforts to date indicate that 25 to 30% of the samples from open springs, creeks and rivers and 10% of the lakes and reservoir samples will be contaminated with cysts of Giardia. Likely the actual percentage is much higher because the techniques for analysis are often severely compromised by water quality. Some surface water sources are always contaminated with cysts from animal or human sources. Other sources are contaminated only when animal populations are moving, as a result of sewage treatment plant breakdown or other unusual circumstances (runoff, flood). Every major source of surface water for municipalities probably is contaminated with cysts of Giardia: repeated sampling has usually provided positive results. The catch phrase here is ''repeated sampling." Three to four samples from a site generally are inadequate, especially if the source of contamination is intermittent, or the samples have been taken during high turbidity situations. Six to 12 samples from a site generally are adequate, providing the samples are taken during the late fall and winter months, the time of year when turbidity and interfering organisms (algae, flagellates) are at a minimum. If Giardia cysts are present during these circumstances, we can assume they are present during higher turbidity situations. A survey of the watershed for possible sources of contamination should provide municipalities with an assessment of the risk. If the possibility or probability for sewage contamination is present, or the watershed provides suitable habitat for aquatic animals (beaver, muskrat and smaller rodents), then contamination of the source is predictable, if not today, certainly in the forseeable future. If cysts are found in a surface water supply from a site, repeated sampling will not change this result. Municipalities must assume that contamination is a fact and take appropriate action to prevent an epidemic. Unfortunately, many municipalities continue to sample, often at a frenetic rate, hoping that subsequent samples will not be contaminated. While cysts may not be found at every sampling, it does not change the fact that they were present and will be there again. In my experience, municipalities that do not have potential problems with Giardia are those that assume the source is contaminated and take appropriate action to prevent any possibility of waterborne giardiasis. Questions/Comments (Regarding entire panel discussion) Audience: The reasons for which you dropped application of the risk model from the proposed rule do not seem insurmountable with additional research. The costs for building one additional treatment plant as a result of the proposed rule could probably pay for the additional research required to satisfactorily address the issues that have been discussed. Ultimately, you are not going to really know if you have effective treatment unless you can monitor the water. Regarding viruses, you can do a lot of monitoring. On paper, some treatment plants look great, but when you actually go out and monitor for viruses you find problems. When all these treatment plants are built, without some model of assessment, you are never going to really know how well they are performing. Could you comment on this? Regli: It would be great to have a model from which to evaluate treatment plants on a case by case basis, and then to modify and upgrade treatment to satisfy some acceptable risk level endpoint. We are just not there yet; and, therefore, are proposing treatment requirements. We believe that if a system were to comply with the proposed criteria, there would be a high probability of a safe water being provided. Regarding the question on filling in the gaps to the risk model with additional research, the problem is timenot having enough. We are mandated by Congress to have treatment requirements promulgated by December of this year. Audience: Could you comment on the turbidity criteria in the proposed rule? Regli: Our current thinking is that there will be different turbidity criteria for different situations. For filtered water systems, the turbidity limit would depend on the filtration technology in place. Systems using direct filtration or conventional treatment, where good correlation has been shown between turbidity removal and Giardia cyst removal, would be required to have filtered water less than 0.5 NTU in at least 95 percent of the measurements taken each month. For slow sand and diatomaceous earth filtration, where there is no correlation between turbidity removal and Giardia cyst removal, the limit would be less than 1 NTU in at least 95 percent of the samples; the limit of 1 NTU is to ensure a high probability that there is no interference with disinfection of viruses. Dr. Logsdon, would you like to add anything to this? Logsdon: Only to stress the point that it is essential for every filter within the plant to meet the criteria. Regli: Yes, this is a strong recommendation that we are making in guidance. We did not feel, however, that such a requirement would be appropriate in a rule because of the difficulty in enforcing such a requirement.
Page 285
Audience: Are there any special criteria for systems with low turbidity source waters? Regli: In a previous draft rule, systems with source water turbidities of less than 1 NTU were required to achieve either 70 percent turbidity removal or a filtered water turbidity of less than 0.1 NTU. This is now a recommendation in guidance rather than in the rule. Audience: How would the rule apply to systems which purchase their water from another system? Regli: The treatment requirements would apply to the system with the treatment technology in place. For example, a system which purchased its water from a filtered water supply would not be required to monitor for turbidity. If the first system met the treatment requirements such as the overall minimum performance of at least 99.9 and 99.99 percent removal and/or inactivation of Giardia cysts and viruses, the system purchasing its water from that system would not be subject to also meeting this requirement, i.e., providing filtration and disinfection as necessary. The purchased water system would be required to maintain a disinfectant residual in the distribution system. Audience: Dr. Borup, is the risk model sensitive to the number of service connections or people served? Borup: No, the model is based on finished water quality meeting some acceptable risk level. The number of samples required to demonstrate that a risk level was not being exceeded would not depend on system size. Audience: The risk model does not seem appropriate in that it only addresses risk relative to the first customer. Customers further out in the distribution system would have longer disinfectant contact times and be subject to less risk than at the first customer. Also, it seems that if the model were accurate, based on the assumptions made, in defining when a system would be subject to significant risk, we would see more people getting sick. The model seems to be very conservative in estimating risk. Borup: Much more research is needed to determine the appropriateness of the different assumptions that are needed for a model to be applicable in a rule. The purpose of the analysis presented was to determine the feasibility of demonstrating through monitoring whether adequate protection was being provided to avoid a specific level of risk. At this point there is not enough scientific data to support most of the assumptions necessary for applying the model in a regulatory context. Audience: Could you elaborate on how CT would be calculated in an unfiltered system for determining the percent inactivation? Regli: C and T would be determined during peak hourly flow, prior to the first customer. The product of C and T, CT, would depend on pH and water temperature. CT tables would be provided in the rule. The system would compare the calculated CT value with those in the rule to see if the percent inactivation was being met. Literature Cited 1. Amirtharajah, A. 1986. Variance analyses and criteria for treatment regulations. Journal of the Amer. Water Works Assoc., 78:3449. 2. Bingham, A. K., Jarroll, E. L., Jr., and E.A. Meyer. 1979. Giardia sp: Physical factors of excystation in vitro, and excystation vs. eosin exclusion as determinants of viability. Exp. Parasitol., 47:284. 3. Bingham, A. K., and E.A. Meyer. 1979. Giardia excystation can be induced in vitro in acidic solutions. Nature (Lond), 277:301. 4. Craun, G.F. 1984. Waterborne outbreaks of giardiasis: Current status. In: Giardia and Giardiasis. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York, New York, pp. 243261. 5. Craun, G.F. 1987. Health aspects of Surface water supplies. Presented at Preconference Seminar, Pacific Northwest Section, American Water Works Association, Bellevue, WA, May 6, 1987. 6. Borup, M.B. 1986. The determination of waterborne pathogen sampling requirements using statistical quality control techniques. Report prepared for Office of Drinking Water, U.S. EPA. 7. Gerba, C. P. 1984. Strategies for the control of viruses in drinking water. Unpublished document prepared under tenure of AAAS/EPA Environmental Science and Engineering Fellowship. 8. Health and Welfare Canada. 1983. Guidelines for Canadian Recreational Water Quality. Supply and Services Canada, Ottawa. 9. Hibler, C. P. 1987a. Analysis of municipal water samples for cysts of Giardia. Report submitted to Office of Drinking Water, U.S. Environmental Protection Agency. January. 10. Hibler, C.P., Hancock, C.M., Perger, L.M., Wegrzyn, J.G., and K.D. Swabby. 1987b. Inactivation of Giardia cysts with chlorine at 0.5°C to 5.0°C. American Water Works Association Research Foundation Report. 11. Hoff, J.C., Rice, E.W., and F.W. Schaefer III. 1985. Comparison of animal infectivity and excystation as measures of Giardia muris cyst inactivation by chlorine. Appl. Env. Microbiol. 50:1115. 12. Hoff, J.C. 1986. Inactivation of microbial agents by chemical disinfectants. EPA/600/286/067. 13. Hudson, S.J., Sauch, J.F., and Lindmark, D.G. 1988. Fluorescent dye exclusion as a method for determining Giardia cyst viability. Advances in Giardia Research. P.M. Wallis and B.R. Hammond (eds), pp. 265269, University of Calgary Press, Calgary, AB. 14. Levenspiel, O. and K.B. Bischoff. 1959. Backmixing in the design of chemical reactors. Ind. Engr. Chem., 51:1431. 15. Levenspiel, O. and K.B. Bischoff. 1961. Reaction rate constant may modify the effects of backmixing. Ind. Engr. Chem., 53:313. 16. Marske, D.M. and J.D. Boyle. 1973. Chlorine contact chamber design; A field evaluation. Water and Sewage Works, 120:7077. 17. Mills, R.G., Barlett, C.L., and J.F. Kessel. 1925. The penetration of fruits and vegetables by bacteria and other particulate matter and the resistance of bacteria, protozoan cysts and helminth ova to common disinfection methods. Am. J. Hyg., 5:559. 18. Regli, S., Amirtharajah, A., Hoff, J., and P. Berger. 1986. Treatment for control of waterborne pathogens: How safe is safe enough? Proc. of the 3rd Conference in Chemical Disinfection, Binghamton, NY, April, 1986.
Page 286
19. Rendtorff, R.C. 1954. The experimental transmission of human intestinal protozoan parasites. II. Giardia lamblia cysts given in capsules. Am. J. Hyg. 59:209. 20. Rice, E.A., and F.W. Schaefer, III. 1981. Improved in vitro excystation procedure for Giardia lamblia cysts. J. Clin. Microbiol. 14:709. 21. Royal Society Study Group. 1983. Risk assessment. Royal Society, London. 22. Schaefer, F.W. III, Rice, E.W., and J.C. Hoff. 1984. Factors promoting in vitro excystation of Giardia muris cysts. Trans. Royal Soc. Trop. Med. and Hyg. 78:795. 23. Schupp, D.G., and S.L. Erlandsen. 1987. A new method to determine Giardia cyst viability: correlation of fluorescein diacetate and propidium iodide staining with animal infectivity. Appl. Env. Microbl. 53:704. 24. Sykora, J.L., et al. 1988. Monitoring as a tool in waterborne giardiasis prevention. Advances in Giardia Research. P.M. Wallis and B.R. Hammond (eds), pp. 103106, University of Calgary Press, Calgary, AB. 25. Trussell, R.R., and Chao, J.L. 1977. Rational design of chlorine contact facilities. Jour. Water Poll. Control Fed. 49:659667. 26. U.S. Environmental Protection Agency. 1986. Ambient water quality criteria for bacteria; 1986. EPA 440/584002. 27. Weber, W. J., Jr. 1972. Physicochemical processes for water quality control. WileyInterscience, New York, NY. pp. 4955. 28. World Health Organization. 1985. Risk management in chemical safety. European Regional Program on Chemical Safety. Geneva. ICPCEH 506MOI 56881.
Page 287
Panel Discussion on Taxonomy of the Genus Giardia. Stanley L. Erlandsen*, E.A. Meyer, T.E. Nash. Department of Cell Biology and Neuroanatomy, University of Minnesota, Minneapolis, Minnesota 55455, U.S.A.. Nomenclature The intestinal protozoan Giardia was first seen by Leeuwenhoek in 1681 when he examined his own diarrheic stools. However, as reviewed by Erlandsen et al (10), the first morphological description and illustration of the organism was performed by Lambl in 185960. Later in 1881, a detailed description of Giardia trophozoites and cysts was presented by Grassi. The taxonomic position of the flagellate Giardia has been recently reviewed (14,15) and was classified as follows: Phylum Sarcomastigophora Honigberg & Balamuth 1963. Class Zoomastigophorea Calkins, 1909 Order Diplomonadida Wenyon, 1926 emend. Brugerolle, 1975 Genus Giardia Kunstler, 1882 The basis for using various specific names (ie. lamblia, intestinalis, duodenalis, enterica) has been discussed in detail by Filice (12), Kulda & Nohynkova (14), Levine (15), and Meyer (17). The most commonly used specific names for Giardia from man include Giardia lamblia Stiles, 1915; Giardia intestinalis Lambl, 1859; and Giardia duodenalis Filice, 1952. In spite of all of the existing arguments regarding which specific name is most appropriate, no agreeable solution appears to be in sight. However, the development of more sophisticated technology and advances in biochemical methodology, and their application to this problem may eventually provide an answer to the question of species designation. Criteria for Speciation Three major criteria (host specificity, morphology, and chemotaxonomy) have been used for the designation of species in the genus Giardia. None of these criteria has provided an irrefutable solution to the problem of speciation, but each has contributed interesting information, which collectively may eventually lead to a better understanding of this organism and a solution to its species designation. Host Specificity Early investigators held a rigid view that each host was infected by a single species of Giardia, and as discussed by Kulda & Nohynkova (14) and Meyer (17), this led to the description of more than 40 species of Giardia with very few detectable morphological differences. The development of gerbil and mouse models for giardiasis (2,20) has provided direct evidence for the crossspecies transmission of Giardia between animal hosts. This has led to a more liberal view that only one species of Giardia may exist and that giardiasis may be viewed as a zoonotic disease (11) while others believed that even though crossspecies transmission could occur, multiple species of Giardia existed in mammals and may represent the endogenous strains of Giardia detected in certain mammals and birds (3,9). Morphology The median body classification of Filice (12) has been, perhaps, the most widely accepted morphological characteristic that has been accepted for use in the determination of Giardia species. By light microscopy, the shape of the median body, a microtubularcontaining cytoplasmic organelle, may be used to distinquish three different morphological groups of trophozoites (accepted by some as species): G. agilis; with median bodies that resembled long clubshaped rods; G. muris; which have small rounded pair of median bodies seen in the midline; and G. duodenalis; shown to possess a pair (that sometimes have appeared to be single) of median bodies resembling a claw hammer and which lie transversely across the body. Although the median body classification has been of value in distinguishing major groups of Giardia, particularily rodent strains from those infecting man, this system has proven to be inadequate since multiple species of Giardia have been identified within this morphological group. For example, G. psittaci and G. ondatrae have both been shown to possess a claw hammer median body morphology, yet they have distinct morphological characteristics that have separated them from other known types of Giardia. Even Giardia trophozoites isolated from beaver intestines have been suggested as not being identical to those found in man (9). Morphometric analysis of the major dimensions of trophozoites has been proposed as a means of differentiating Giardia spp. [see review by Woo (24); Bertram et al (5)], however, its application was questioned by Filice (12) and the biological variation in trophozoite sizes encountered in cultures raised questions about the usefulness of this method. Chemotaxonomy A number of biochemical and molecular biological methods, including the detection of isoenzymes, immunological analysis of surface antigens, and characterization of chromosomes and DNA, have just begun to be utilized for the investigation and determination of Giardia spp. It has been anticipated that these new approaches, together with information obtained from crossspecies transmission and morphological studies, may provide some answers to help clarify the existing controversies of determining Giardia spp. However, as noted by Boreham et al. (6) even this evidence may need to be viewed with some caution since these methods have usually required large numbers of cells that currently may * Corresponding author
Page 288
only be obtained from cultures. Not all strains (species) of Giardia can be grown in culture and evidence has been presented that culture conditions might be selective for specific genotypes (1). Investigation of isoenzymes in Giardia have been carried out by Bertram & Meyer (5), Korman et al. (13), and Meloni et al (16). These results have shown that cultured Giardia isolates can be grouped into multiple zymodemes, but more importantly, have demonstrated that isolates from humans can have considerable heterogeneity. These results have been suggested to support the view that genetically different strains of G. lamblia can occur in the human host (13). However, others have interpreted the findings of similarity between isolates from different animals as evidence that humans could contract Giardia infections by a zoonotic mechanism (22). Immunological investigations of the antigenic composition of Giardia spp. have detected differences in surface labeling and antigenic profiles between Giardia strains obtained from different countries (7,18,19,21) while others have noted similarities between isolates from the same geographical locale (23,16). Restrictionendonuclease analysis of DNA from 15 different Giardia isolates obtained from human and animal sources also showed considerable heterogeneity in banding patterns, with six isolates (two animal and four human) forming one group, while the pattern for another group consisted of two human and one beaver (19). Interestingly, morphological investigation of these two beaver strains of Giardia suggested that one of them, in addition to sharing the human DNA pattern, was morphologically similar to human organisms, while the other strain which had a DNA banding pattern unlike that of human Giardia, resembled the phenotype of Giardia isolated from livetrapped beavers (9). This correlation between two different methodologies, provides an interesting approach for investigating the problem of Giardia spp., which needs further confirmation, but if correct, it should provide the impetus for using combined approaches for analysing what has proven to be a complex problem. In summary, the criteria for Giardia species proposed by Filice (12) would seem to be acceptable for identification of G. agilis and G. muris. The term G. duodenalis would not, however, represent a single species, but rather a group, since the original definition is based solely on a morphological entity (median body shape), and several distinct species (G. psittaci, G. ondatrae) share this same characteristic. The substitution of the term G. intestinalis for G. lamblia does not seem appropriate at this time, especially since there is so much confusion in regard to whether one or more species (strain) of Giardia can infect man. To minimize the confusion, it might be more helpful to confirm whether or not an isolated Giardia organism fits within the G. duodenalis classification, and then provide as much information as possible about its animal host, geographical location, and how it compares to other known strains. Until definitive results can demonstrate the existence of distinct species, it might also might be prudent to view biochemical or morphological similarities for individual characteristics with some skepticism, and not assume that it means a lack of host specificity. The development of better biochemical or molecular biological approaches that can be applied to smaller numbers of organisms in a direct isolate from infected animals, so as to avoid the potential problem of changing genotypes in culture, may eventually provide valuable information that will assist in the determination of species in the G. duodenalis group. Literature Cited 1. Aggarwal, A. and T.E. Nash. 1988. Antigenic variation of Giardia lamblia in vivo. Infect. Immun. 56:14201423. 2. Belosevic, M., Faubert, G.M., MacLean, J.D., Law, C., and N.A. Croll. 1983. Giardia lamblia infections in Mongolian gerbils: an animal model. J. Infect. Dis. 147:222226. 3. Bemrick, W.J. and S.L. Erlandsen. 1988. Giardiasis is it really a zoonosis? Parasit. Today 4:6971. 4. Bertram, M.A., Meyer, E.A., Lile, J. and S.A. Morse. 1983. A comparison of isozymes of five axenic Giardia isolates. J. Parasit. 69:793801. 5. Bertram, M.A., Meyer, E.A., Anderson, D.L. and C.T. Jones. 1984. A morphometric comparison of five axenic Giardia isolates. J. Parasit. 70:530535. 6. Boreham, P.F.L., Upcroft, J., Upcroft, P., and R. Andrews. 1988. Importance of giardiasis. Parasit. Today, in press. 7. Einfield, D.A. and H.H. Stibbs. 1984. Identification and characterisation of a major surface antigen of Giardia lamblia. Infect. Immun. 46:377383. 8. Erlandsen, S.L. and W.J. Bemrick. 1987. SEM evidence for a new species, Giardia psittaci. J. Parasitol. 73:623629. 9. Erlandsen, S.L. and W.J. Bemrick. 1988. Waterborne giardiasis: sources of Giardia cysts and evidence pertaining to their implication in human infection. In: Advances in Giardia Research. P.M. Wallis and B.R. Hammond (eds). pp. 237246, University of Calgary Press, Calgary, Canada. 10. Erlandsen, S.L. and D.G. Feely. 1984. Trophozoite motility and the mechanism of attachment. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. S.L. Erlandsen and E.A. Meyer. (eds). Plenum Press, New York, NY, pp. 3363. 11. Faubert, G.M. 1988. Evidence that giardiasis is a zoonosis. Parasit. Today 4:6668. 12. Filice, F.P. 1952. Studies on the cytology and life history of a Giardia from the laboratory rat. Univ. Calif. Publ. Zool. 57:53143. 13. Korman, S.H., Le Blancq, S.M., Spira, D.T., El On, J., Reifen, R.M. and R.J. Deckelbaum. 1986. Giardia lamblia: identification of different strains from man. Z. Parasitenkd. 72:173180. 14. Kulda, J. and E. Nohynkova. 1978. Flagellates of the human intestine and of intestines of other species. In: Parasitic Protozoa. J.P. Kreier (ed). Academic Press, New York, NY, pp. 69104. 15. Levine, N.D. 1978. Giardia lamblia: classification, structure, identification. In: Waterborne Transmission of Giardiasis. W. Jakubowski and J.C. Hoff (eds). U.S. Environmental Protection Agency, Cincinnati, OH, pp. 28. 16. Meloni, B.P., Lymbery, A.J. and R.C. Thompson. 1988. Isoenzyme electrophoresis of 30 isolates of Giardia from humans and felines. Am. J. Trop. Med. Hyg. 38:6573. 17. Meyer, E.A. 1985. The epidemiology of giardiasis. Parasit. Today 1:101105. 18. Nash, T.E. and D.B. Keister. 1985. Differences in excretorysecretory products and surface antigens among 19 isolates Giardia. J. Infect. Dis. 152:11661171. 19. Nash, T.E., McCutchan, T., Keister, D., Dame, J.B., Conrad, J.D., and F.D. Gillin. 1985. Restrictionendonuclease analysis of DNA from 15 Giardia isolates obtained from humans and animals. J. Infect. Dis. 152:6473. 20. RobertsThomson, I.C., Stevens, D.P., Mahmoud, A.A.F.
Page 289
and K.S. Warren. 1976. Giardiasis in the mouse: an animal model. Gastroenterology 71:5761. 21. Smith, P.D., Gillin, F.D., Kaushavl, N.A. and T.E. Nash. 1982. Antigenic analysis of Giardia lamblia from Afghanistan, Puerto Rico, Equador, and Oregon. Infect. Immun. 36:714719. 22. Thompson, R.C.A., Lymbery, A.J. and B.P. Meloni. 1988. Giardiasis a zoonosis in Australia? Parasit. Today 4:201. 23. Wenman, W.M., Meuser, R.U. and P.M. Wallis. 1986. Antigenic analysis of Giardia duodenalis strains isolated in Alberta. Can. J. Microbiol. 32:926929. 24. Woo, P.T.K. 1984. Evidence for animal reservoirs and transmission of Giardia infection between animal species. In: Giardia and Giardiasis: Biology, Pathogenesis, and Epidemiology. S.L. Erlandsen and E.A. Meyer (eds). Plenum Press, New York, NY. pp. 341364.
Page 291
Panel Discussion on Methods of Handling Giardia in the Laboratory Walter Jakubowski*, Ernest A. Meyer, Theodore E. Nash, Charles P. Hibler U.S. Environmental Protection Agency, Cincinnati, Ohio 45268, U.S.A. Objectives and Format The objective of this panel was to discuss particularly troublesome areas in working with trophozoites and cysts in the laboratory. The intent was to identify possible solutions through a sharing of experiences while underscoring significant problems requiring further research. The format was informal and the session was roughly divided into two sections concerning problems with culturing and maintaining trophozoites and problems with sources and quality of cysts. Each panel member briefly introduced a topic and questions and comments were taken from the floor. The summaries and discussion below were developed by the chairman from written material submitted by the panel members and from notes and tapes of the session. The chairman accepts full responsibility for the accuracy of their content. Summaries Theodore E. Nash Methods for the axenization of Giardia lamblia were reviewed. Trophozoites used for axenization are obtained in three ways. The most direct way is by sampling the small intestinal contents either by duodenal intubation or by use of the string test. Immediately after verifying the presence of the parasite, the samples are inoculated into TYIS33 medium containing antibiotics and viable Giardia are allowed to adhere to the glass tubes at 37°C. The medium is then decanted and replaced with fresh medium. In this way, the intestinal contents and the old medium, which are toxic to the organism, are separated from the adhered Giardia. The other two methods derive trophozoites from cysts purified from feces. Stools from patients excreting large numbers of cysts are used in order to increase the numbers of cysts obtained so that several attempts and procedures can be employed. There are a number of methods used for purifying cysts from stools but we use sucrose gradients. Perhaps the most successful method for obtaining trophozoites from cysts in our laboratory is to inject 50,000 cysts in 0.1 mL into the stomach of 1 to 2 day old suckling mice (Keister and Mattern, unpublished information). Mice are sacrificed at spaced intervals beginning 7 days after inoculation. Portions of the intestines are examined and placed into tubes containing medium and antibiotics as above. Intestinal contents and contaminated medium are quickly removed from adhered trophozoites. In the third method, trophozoites are obtained after in vitro excystment. At least 100,000 cysts in 100 µL of water are added to 1 mL of 1% pepsin in 0.85% NaCl, pH 2.0. Tubes are slowly mixed at 37°C for at least 30 minutes. Differences in numbers of trophozoites obtained after incubation from 30 to 90 minutes are not found, so we will often use longer incubation times to kill some of the accompanying bacteria. The pelleted cysts are then added to tubes containing medium and antibiotics. Trophozoites are not immediately found but appear after 6 to 24 hours. Most bacterial antibiotics have no effect on cultivation of Giardia so that a large assortment can be used. In contaminated specimens, some or all of the following antibiotics are used: penicillin (200 µg/mL), gentamycin (40 µg/mL), ticarcillin (1000 µg/mL), clindamycin (50 µg/mL), moxalactam (1000 µg/mL) and amphotericin (1 µg/mL). Problem organisms tend to be resistant gram negative rods, mostly Pseudomonas species, yeast and an occasional anaerobe. None of these methods work all of the time. However, the mouse inoculation method yields enormous numbers of trophozoites which increases the likelihood of axenization. Successful axenization of Giardia is obtained in our laboratory from about 70% of stools from different patients. Ernest A. Meyer After presenting historical information on the establishment of axenic cultures of Giardia, several problems relating to handling the organism in the laboratory were discussed. Problem areas meriting additional attention included the need for: 1. More reliable means for establishing organisms in culture. Successful adaptation to culture in 10 to 50% of attempts would seem to reflect the audience experience. The necessity to include in the culture medium such undefined and variable components as trypticase, yeast extract, phytone and serum most likely is responsible for some of our failures. The development of a culture medium that would more reliably yield Giardia cultures from a small inoculum could be welcome. Efforts to develop such an improved medium might well start by systematically varying existing Giardia culture media. 2. More strains established in culture, and promptly preserved by freezing, in order to minimize the possibility of changes occurring with continued cultivation. Our ability to develop a satisfactory system of classifying organisms in this genus depends on our ability to characterize Giardia. Clearly, the more isolates that are available for study, the better we are able to understand their host range, pathogenicity and chemotaxonomy, and to classify them. * Corresponding author.
Page 292
3. Development of methodology for the culture of G. muris type organisms. It was pointed out that the RobertsThomson mouse model is now used worldwide, and has yielded a great deal of information relative to Giardia behavior in the mouse. At the same time, we are not yet able to culture and study this or any other G. muris strain in vitro. 4. A central repository where all available cultures can be kept, and from which they can be supplied to interested investigators as needed. The findings of any worker can only be generally accepted when they can be repeated, and they can only be repeated when the particular Giardia organisms used in an experiment are available to others. 5. An improved nomenclature. Three aspects of this problem were discussed. First is what has become the "traditional" species problem, and whether species should be based on those few known morphological differences that can be observed including median body morphology, presence or absence in the trophozoite of a ventral flange, the existence of "binary" cysts, and length of caudal flagella, or on the basis of the animal host. A second aspect of the perceived problem of Giardia nomenclature has been the tendency of various authors to use different names for what may well be the same organism. Thus, it is possible to find Giardia from humans asigned by various authors the species names of lamblia, duodenalis and intestinalis. A third emerging aspect of the problem has been the absence of a uniform method of naming Giardia culture isolates. It would be most helpful if an interested, knowledgeable group could study the problem and make recommendations regarding a) the nomenclature of this group, and b) a recommended standard method of identifying Giardia isolates. The names by which these organisms are identified is of far less importance than the fact that a uniform, agreedupon system be employed that everyone can use. Charles P. Hibler Procedures and problems in obtaining and purifying cysts were discussed. Cysts may be obtained from human or animal sources. Human source Giardia cysts generally are obtained from medical hospitals, medical care centers or student health centers on university campuses. On large university campuses, students frequently are a good source. A number of problems face investigators planning to use human source Giardia. One potential problem is that the source is infected with pathogens other than Giardia, causing a risk to laboratory personnel processing the sample. A major problem for investigators is availability of a sample when needed. The age and sex of the patient, the possible source of the infection, the duration of the symptoms, and the treatment history generally are unknown. Frequently, the samples are inadequate (quantity of cysts and/or stool), and range from a few hours to several days old. Samples from some human sources (even when fresh) often contain more obviously dead than live cysts, indicating either treatment and/or an immune response. Most investigators prefer to work with human rather than animal source cysts. However, if the Giardia duodenalis type cyst is all that is required, a good source is humane shelters. Generally, 80 to 90% of the dogs housed in these situations are infected and high percentage (20 to 30%) are shedding cysts at any given sampling period. Other animal sources of G. duodenalis are more difficult to obtain. Veterinary hospitals, clinics or state diagnostic laboratories that routinely examine a number of large animals (cattle, sheep, horses, etc.) are a good source. Other animals, such as beaver, muskrat, moose, etc., necessitate a certain amount of field experience with wildlife to be successful. As with the human source Giardia, these animals (except dogs in a humane shelter) are rarely if ever available when needed and many of the same problems apply (duration of infection, history of treatment, etc.). Those investigators using G. muris as a model for G. duodenalis must exercise considerable caution because it is a different species and information obtained may not be applicable to G. duodenalis. If, for example, inactivation by chlorine, ozone or other disinfectants is different for G. muris and G. duodenalis, then data generated using G. muris may not be acceptable when inactivation data for G. duodenalis (=G. lamblia) are required in the development of drinking water regulations. Several animals, from neonatal rats to neonatal mice and other laboratory rodents, have been more or less evaluated for their susceptibility to human source cysts of G. duodenalis and for their suitablity as animal models. The Mongolian gerbil, Meriones unguiculatus, appears to be the best model for human as well as other animal sources. Certainly, these gerbils possess considerable desirable attributes: they are gentle, odorless, and with proper husbandry, are prolific. However, most of us have not made the same effort with other laboratory animals that we have put into Mongolian gerbils. Developing the Mongolian gerbil as a model is not an easy task. The initial colony must be treated to eliminate possible Giardia infection, and the commensals Trichomonas and Endamoeba. This breeding colony must then be maintained under strict governmental regulations regarding care and maintenance of laboratory animals. This involves considerable time, expense and facilities generally available only at larger universities or governmental laboratories. Separation of cysts from the stools of humans, dogs and other animal sources that are omnivorous or carnivorous, and produce fine particle waste, is very difficult, often necessitating considerable dilution of the sample and passage through gauze and/or screens several times. Separation of cysts from stools of most herbivorous animals (except cattle) is not as difficult. Recovery of cysts is much more effective because the waste consists of larger particles. If laboratory rodents, such as the Mongolian gerbil, are the source, fecal pellets must be collected in water because the removal of excess water in the rectum of rodents will result in dehydration of the cysts shortly after the fecal pellet is expelled. These pellets can be macerated and, with washing, the cysts passed through 80 to 100 mesh screens. This procedure can be used for
Page 293
stools from humans and animals and will provide a "semiclean" suspension that is ready for further cleaning or refrigerator storage. Cysts from human or dog sources generally will not remain viable more than a few days, even with antibiotics (Penicillin and streptomycin) added; however, laboratory or other (wild) rodent source cysts will remain viable for a considerable period providing investigators either add antibiotics or replace the water to the sediment daily and agitate. Daily water replacement may be as effective as antibiotics. If the suspension is carefully cleaned, the water replaced daily, and/or antibiotics added, some of the cysts will remain viable for 6 weeks to two months. The final purification step generally involves overlay/underlay of highly dilute suspensions with chemicals such as sucrose, zinc sulfate, percoll, ficollhypaque, etc. The lower the specific gravity (about 1.07 to 1.1) of the chemical flotation solution, the better. Cyst losses are greater, but the effect of the chemicals is less detrimental. Recovery of viable cysts is better with percoll than with any other chemical, but percoll is cost prohibitive for many laboratories. After the underlay/overlay step, the suspension is centrifuged for 3 to 8 minutes and the material to the interface, or all of the material to the pellet, is siphoned with slight vacuum through a 3 or 5 µm nucleporetype membrane. The membrane is then rinsed into a beaker with a strong jet of distilled water, and the cysts are refrigerated. As with the semiclean preparation, daily replacement of water will help longevity of the cysts. However, most if not all chemicals are detrimental to cyst survival; therefore, if critical experiments are planned, cysts should be used immediately. Our experiences with Giardia cysts from various human and animal sources indicates that a "maturation period" of several days to a week is actually detrimental. Fewer will excyst and more are required to infect animals than those collected and used within two or three days. Walter Jakubowski The need for standardization, both in definitions and in procedures for handling Giardia, was discussed. Health and regulatory agencies require quantitative information on the occurrence, distribution, survival and inactivation of this organism, and on the efficacy of water treatment processes for its removal. The information may be used in assessing risks and in setting standards, and for determining the appropriateness of intervention to interrupt transmission. Protocols and definitions developed in one laboratory for a given type of experiment may differ significantly from those used in other laboratories for the same type of experiment. For example, in evaluating the efficiency of water sampling techniques, one laboratory may use "fresh" cysts and another may use "preserved" cysts. Both terms are imprecise and have different meanings to different investigators. The cysts used by different laboratories may be from different sources and of different ages and species. Assay procedures may also vary and may be of unknown or undetermined precision. Differences in any of these factors could affect the results and interpretation of experiments. Standardization of cyst source, age, purity, preservation, storage conditions, quantitation and identification could improve reproducibility of results within a laboratory and simplify comparison and interpretation of results across laboratories. These same factors have relevance to all types of studies where cysts are used including crossspecies infectivity, survival and inactivation studies, speciation and evaluation of detection methodologies. Working with Giardia is very much an "art"; those factors that are important in reproducibility of results need to be identified and acceptable criteria for them must be established in order to advance the "science". Discussion Comments and questions from the floor reinforced some points made by the panel members, challenged others and raised additional questions for consideration. With respect to axenization, other investigators as well have found inoculation into suckling mice to be the procedure of choice for establishing trophozoite cultures. Gram negative organisms appear to be the most difficult contaminant to deal with and one investigator reported no success in finding a satisfactory antibiotic against Aerobacter. It was also mentioned that contaminated cultures might be rescued if the culture is inoculated into an animal. The animal may allow growth of the Giardia and not of the contaminant. It was suggested that clonal isolation and cryopreservation of isolates be performed early in culture since the organisms can be subject to selection and change upon subculture. It was agreed that this is a desirable procedure. However, existing cloning methods are very difficult to use and better methods are needed. The comment was made that cysts may lose infectivity if left in contact with feces for 24 hours. Another investigator has obtained infectivity of cysts left in stools for more than a week. The conflicting results may reflect the numbers of cysts originally present in the sample, the makeup of the stool or the animal model used. However, there appeared to be general agreement that cysts should be separated as soon as possible from fecal material in order to increase the chances of cyst survival. The problem of confusion in Giardia nomenclature was reiterated. In an Australian laboratory, they have adopted the WHO system used for designating trypanosome isolates. A description of this system may be found in the Journal of Antimicrobial Chemotherapy 14:449461, 1984. The question of a maturation period for Giardia cysts still appears to be a question. While there was agreement that G. muris does not require a maturation period, some investigators believe it is necessary for G. duodenalis while others do not. Again, differences, in the experimental protocols may account for the contrary results or there may be true differences among species. Further investigation of this question is needed.
Page 294
Cryopreservation of Giardia was discussed. The preservation of actively growing cultures of trophozoites is now readily achieved by freezing these organisms in the presence of glycerol or dimethylsulfoxide. However, cryopreservation of cysts requires tremendous numbers of cysts and the results are erratic. The comment was offered that Cryptosporidium oocysts do not survive freezing but that the sporozoites can be frozen and recovered. With respect to cysts used to evaluate drinking water treatment processes, suggestions were made that only cysts not more than 7 days old be used and that they not be preserved. Formalin preserved cysts may be more fragile than unpreserved cysts and their use can lead to falsely high efficiencies for water treatment removal processes. It was suggested that the cysts be of good quality, i.e., having cytoplasm that is almost translucent and not granular. One laboratory indicated they find more cysts in the supernatant of extracts from stool specimens and water filters than they find in the sediment after using established protocols for cyst recovery. This was contrary to results obtained in at least two other laboratories but it is a question deserving further study and again reflects our lack of knowledge on what to consider as significant factors and how to standardize procedures.
Page 295
INDEX A acid phosphatase 139, 187188, 252, 254, 268269 actin 175 activated serum rabbit or human 169 acute 1, 33, 49, 5253, 107, 191 adhesive disc 2930, 231, 266 adsorbed 46, 155, 169171 agilis 79, 287288 Alberta 1519, 6164, 7982, 107, 129131, 133, 137, 140, 143144, 159, 161, 164165, 167169, 172, 255, 289 amodiaquine 3 animal infectivity 32, 205, 237, 249, 252255, 265, 269, 281, 286 animal model 28, 3132, 49, 55, 64, 75, 77, 79, 82, 144, 189, 204, 234235, 254255, 260, 265, 268269, 282, 288289, 293294 animal reservoir 55, 69, 71, 74, 82, 152, 164, 210, 231, 289 animal shelter 61, 6365, 165 antiepidemic measurement 40 antiGiardia antibody 2, 4748, 50, 52, 192 antiGiardia immunity 47 antibodies 2, 4, 4550, 52, 5458, 66, 156, 159164, 169172, 177, 179180, 192, 194, 205207, 209, 216, 219220, 222223, 273, 283 antigen 14, 28, 4648, 50, 55, 153157, 162, 169, 171172, 177180, 191194, 220, 222, 288 antisera 1, 153156, 162, 169, 171, 191, 220, 222, 231 aquatic mammals 227 Arkansas 7172, 74 aryl sulfatase 268 assessment 34, 71, 103, 172, 198, 202, 205, 228, 237, 255, 259, 277278, 283286 asymptomatic 1, 1618, 33, 36, 65, 82, 153, 163, 177179, 181184, 191, 219, 222, 232, 251, 262263 attachment 24, 3334, 3637, 46, 49, 54, 6869, 88, 99, 107, 236, 262, 265, 288 axenic 2, 4, 7, 12, 21, 24, 2931, 33, 37, 54, 58, 81, 147, 156, 162, 167168, 172, 180, 192, 194, 223, 230, 232, 235, 254, 265, 288, 291 axonemes 26, 2930, 133134, 154, 197, 221, 253, 266 B B cell 47 bacteria 34, 7, 9, 12, 14, 2528, 48, 67, 8789, 9192, 95, 9798, 100, 103, 107108, 111112, 205, 208, 265, 275, 286, 291 basal bodies 29 beaver 29, 7477, 7980, 9091, 129, 159162, 164165, 167, 169170, 202, 204, 223, 227234, 236237, 239241, 245, 253, 285, 287288, 293 beige mice 47 berberine sulfate 34 bile 4, 9, 12, 24, 2931, 33, 37, 46, 4850, 52, 5455, 147, 168, 180, 219221, 223, 251252, 254, 261, 263264, 273 bile duct cannulation 50 bile salt 9, 33, 261, 273 biliary antibody responses 52 binary cysts 227, 231232 BinghamMeyer 250252 biochemical 9, 29, 31, 55, 160, 162, 165, 189, 223, 235, 287288 biopsy 3, 45, 68, 7980, 139 bright field microscopy 32, 253, 269 C Calgary 3, 6164, 7980, 137138, 160, 165, 169170, 205, 217, 254255, 286, 288 canine 61, 6465, 69, 220221, 232 carbohydrate 46, 68, 147, 187189 carbon dioxide 88, 109, 126, 219220, 249, 251252 cartridge filters 89, 206 categorization 224 cats 51, 65, 75, 227, 232, 235, 240, 245 cattle 16, 7577, 79, 160161, 208, 227, 232233, 240, 245, 293 cDNA 147148, 151153, 155157, 184 cell 6, 914, 22, 2427, 2930, 3337, 4549, 111, 118, 133135, 152, 155157, 160, 166, 173, 175, 177, 180181, 183, 185, 189, 219221, 227, 231, 236, 256, 258, 260, 265, 267, 287 cell free 33, 35, 155, 183 cell viability 9, 12, 265 characterization 4, 1314, 48, 94, 153154, 157, 172, 180184, 187, 194, 269, 287 chemical 29, 36, 50, 89, 95, 97, 100103, 107108, 112, 116, 118119, 121, 125126, 128, 138, 160, 162, 189, 198, 200203, 215216, 223, 237, 239, 243 244, 250, 255258, 283,
Page 296
286287, 293 chemotaxonomy 287, 292 children's homes 39 Chilomastix mesnili 40, 211 chlorine 7577, 8788, 97, 103, 105, 108, 112113, 117119, 125126, 128129, 133135, 137138, 140144, 198199, 223225, 233234, 238, 242, 246, 254, 256, 258, 282283, 286, 293 chromosomes 147152, 287 chronic 1, 7, 25, 27, 45, 49, 5355, 58, 107, 191 clinical features 179 coagulation 95, 97, 99103, 117, 124, 129, 197, 205, 242, 280281 coagulationfiltration 95, 100102 complement mediated 169, 171 contact 1518, 22, 3336, 39, 8788, 97, 105, 107, 115, 118119, 121, 123, 125, 127, 137, 141144, 199, 223224, 234, 238, 246, 249, 256257, 275, 282 286, 294 contamination 1617, 65, 74, 77, 87, 103, 106107, 110, 119, 143, 198199, 205, 208, 223, 225, 227, 233235, 238241, 245, 273, 284285 counter immunoelectrophoresis 12, 191192, 219, 222 coyotes 245 criteria 3, 9, 2223, 3031, 49, 117121, 144, 197, 216, 234235, 266, 275276, 278, 285288, 294 CT value 103, 137, 141, 143144, 275, 283284, 286 culture 1, 4, 7, 912, 2124, 2831, 3336, 7981, 97, 139, 147, 152, 160161, 165166, 168, 170, 177, 179180, 191, 219221, 223, 230, 232, 235, 253254, 288, 291292, 294 cysteine 21, 250251, 256, 261264, 273 cytokinesis 249, 252 cytometry 910, 1214 cytopathogenic 33 cytotoxic 33, 4548, 5558, 163, 171, 265 Czechoslovakia 3940 D day care 1, 4, 1516, 1819, 49, 69, 7374, 191, 194, 236 daycare centres 39, 237 detachment 3435 detection 4, 14, 27, 48, 5152, 54, 67, 103107, 129, 144, 152153, 155, 163164, 168, 179, 182, 188189, 194, 197, 202210, 213, 215, 217, 219225, 228, 232, 234, 236238, 255, 260, 265, 269, 287, 294 detergent 148, 180, 206, 211213 diagnosis 34, 40, 48, 65, 6869, 151, 179, 194, 197, 202, 204, 213, 216, 219, 222223, 241 diagnostic test 153, 179, 221 diatomaceous earth 87, 95, 9899, 102103, 113114, 237, 244, 285 diet 6869, 160 diethyl ether 80, 211, 213 differential interference contrast (DIC) 2932, 249, 252253, 265269 disinfection 95, 99100, 104, 107, 111112, 118119, 135, 137, 140141, 144, 223, 233234, 249, 254, 275, 277, 280286 distribution 2, 17, 28, 39, 4749, 6568, 71, 7374, 87, 8990, 103105, 107108, 113114, 117118, 120, 124125, 135, 162, 177, 189, 192, 223224, 228, 231232, 236, 238, 242, 244, 275277, 285, 293 DNA 0, 6, 12, 21, 2426, 28, 49, 55, 58, 69, 147153, 155157, 160, 163, 165168, 173, 175176, 183, 230, 236, 287289 dog 6169, 7576, 7980, 82, 130, 159161, 165, 167, 169170, 227, 231233, 236, 240, 245, 250, 253, 293 drug 37, 910, 12, 14, 2124, 2728, 58, 6566, 139, 156 duodenalis 25, 4954, 75, 7981, 133, 137, 139, 141, 143, 165172, 204, 231232, 234, 237, 245246, 283, 287289, 293294 Dusan Lambl 39 dye exclusion 4, 139, 252255, 258259, 265, 281, 286 E ectoplasmic vacuoles 249, 252 Edmonton 1516, 19, 6164, 129, 131, 137, 165 effector 4748 electron microscopy 25, 29, 31, 3336, 47, 113, 135, 205, 220221, 232, 236, 252, 254255, 265266, 268269, 273 electrophoresis 12, 4, 50, 55, 147153, 160, 165166, 168169, 178, 181183, 191192, 194, 289 ELISA 14, 50, 52, 156, 160161, 191194, 219, 222 elution 12, 50, 181, 191192, 206207 encystation 2930, 69, 75, 187, 222, 235, 273274 encystment 219221
Page 297
endonuclease 28, 49, 55, 69, 147148, 152, 165168, 230, 236, 289 endosymbiont 2628 energy metabolism 187189 Entamoeba histolytica 4, 12, 24, 37, 194, 211, 219 enteric 1, 15, 1819, 28, 47, 71, 191, 205 enzyme 1, 4, 6, 4852, 58, 116, 147, 151, 166167, 180, 182, 187189, 191, 194, 219, 222223, 262 eosin 28, 31, 252255, 259260, 265, 268269, 283, 286 EPA 4, 32, 65, 68, 81, 87, 9495, 102104, 112113, 116117, 135, 194, 197198, 204, 209, 220, 225, 254, 273, 275, 277278, 286 epidemiological 3941, 61, 65, 81, 103, 129, 227, 229, 236, 278 epidemiology 19, 31, 37, 39, 41, 48, 69, 135, 162164, 235237, 288289 erythromycin 34, 22 ethyl acetate 72, 211, 213 excretion 49, 5152, 54, 160, 222, 231 excystation 2829, 3132, 50, 55, 75, 80, 82, 107, 109, 112, 125, 133, 135, 139141, 143144, 163164, 168, 171, 187, 189, 220, 249265, 268269, 273 274, 281283, 286 F familial occurrence 3940 fatty acid 273 FDA 910, 2931, 265268 Fecal Parasite Concentrator 211 fecaloral 15, 40, 168, 232, 255 FeKal CONTrate 211 fibroblast cells 3536 field inversion gel electrophoresis 147 filter 1, 910, 21, 33, 75, 8793, 95110, 112, 114116, 118121, 123, 129131, 133, 137138, 140143, 163, 191, 197203, 205208, 216, 220, 224, 237, 242, 244245, 255256, 262, 275, 280, 285, 295 filtration 12, 26, 75, 8788, 92, 94105, 108119, 124, 129130, 133, 137, 139143, 145, 191192, 197200, 203, 205, 209, 212213, 215, 220, 223, 233 234, 237, 239245, 251, 277, 280281, 285 flagella 10, 2123, 3031, 35, 133135, 227231, 251, 253, 266, 292 Flagyl (see metronidazole) 75 flange 3536, 292 FluoraBora I (3(dansylamido)phenylboronic acid 252 fluorescein diacetate 9, 29, 32, 198, 205, 237, 252253, 265, 269, 281, 286 fluorescence 4, 910, 12, 30, 47, 154, 194, 203, 217, 221, 223224, 231, 253, 255258, 265268 fluorescence activated cell sorter (FACS IV) 265 fluorescent 10, 12, 25, 203, 206, 209, 215216, 222225, 231, 249, 252, 254256, 258, 267, 269, 286 fluorogenic 2932, 198, 205, 231, 252253, 255, 265269, 273, 281 food 1518, 40, 50, 66, 80, 88, 125, 177, 205, 231, 239, 249, 273 Formalinether (FE) sedimentation 211, 213 furazolidone 37, 911, 21, 2324 G gastrointestinal 34, 41, 4547, 177, 194, 227, 249 gel electrophoresis 12, 4, 147, 149, 151153, 160, 166, 168169, 178, 181183, 191192, 194 genes 51, 147, 150151, 173, 175176, 183184 genetic 28, 49, 147, 150151, 160, 165, 167, 230 genome 49, 147, 149151, 166 genotype 288 gerbil 55, 75, 7980, 139, 165, 246, 253, 256, 293 giardicidal 29 giardins 154 glutathione 3, 6, 250251, 256, 261263, 273 glutathione peroxidase 6 glutathione reductase 6 golgi 252 granular media 100101 growth 1, 9, 12, 14, 21, 24, 2829, 37, 6869, 88, 91, 96, 117, 125, 160, 179, 191, 200, 224, 241, 294 gypsies 39 H Hanks' balanced salt solution 250251, 256 HeLa 2728, 3336 helper/inducer 4548 homosexual 254 host 4, 29, 31, 33, 3637, 4849, 53, 55, 61, 65, 75, 7981, 148, 152153, 159, 162, 165, 167, 169170, 172, 177, 179, 181, 183,
Page 298
230231, 234, 249250, 253, 273, 281283, 287288, 292 hsp gene 173175 HSP3 250, 261, 264 human 4, 7, 9, 12, 17, 27, 31, 33, 35, 45, 4749, 51, 53, 55, 5758, 61, 6366, 6869, 7477, 7982, 97, 109, 125, 129, 133, 135, 137, 139, 142, 144, 159 163, 165, 167172, 175, 177180, 182184, 187, 203, 209, 211, 213, 216, 219223, 227234, 236237, 240, 245, 249250, 253255, 261264, 273274, 282284, 286, 288289, 292293 humoral 47, 49, 55, 171, 177, 180 Hutterite 1519 hydrolytic abilities 187 hypersensitivity 54 hypothymic 49, 5152, 5455, 113 I IgA 4550, 52, 5455, 57 IgE 50 IgG 2, 4, 4546, 4850, 57, 66, 180, 192, 194, 203, 219 IgM 45, 4850, 52, 5455, 57 immobilization 54, 116, 169172 immune 12, 4, 4749, 5255, 5758, 171172, 177, 180184, 191192, 236, 246, 282, 293 immunity 47, 49, 5355, 172, 234 immunodeficiency 45, 180 immunodiffusion 169, 172 immunoelectrophoresis 12, 153, 191192 immunofluorescence 12, 47, 49, 59, 66, 153154, 159162, 164, 180, 197, 203, 205206, 209210, 215217, 219, 222223, 225, 234, 236237, 260, 265, 269, 283 immunoglobulin 45, 4749, 55, 161, 220 immunological 1, 45, 47, 57, 81, 153, 157, 169, 191, 232, 234235, 287288 immunology 25, 37, 45, 47, 49, 157 immunoreactivity 29, 231 in vitro 1, 37, 21, 23, 2833, 3537, 46, 51, 5557, 6869, 75, 8082, 109, 112, 133, 135, 139140, 144, 153, 155157, 160, 164, 169172, 175, 181184, 187, 189, 191, 219222, 235, 249250, 254255, 257, 260, 263, 265, 268269, 273274, 281, 286, 291292 in vivo 4, 6, 12, 2931, 36, 46, 49, 8081, 139, 220221, 249250, 261, 288 inactivation 4, 7, 75, 77, 94, 110113, 125128, 133, 135, 137138, 144145, 170171, 224, 254, 275, 277, 280286, 293294 incidence 19, 41, 64, 66, 6869, 7174, 129, 227, 255, 279 incorporation 6, 2931, 87, 138, 183, 211, 213, 264266 incubation 2, 11, 2122, 3435, 50, 117, 161, 182, 192, 221, 250252, 256, 261262, 291 indicator 87, 105, 107, 130131, 202, 205, 231, 237 indigenous 92, 208, 228, 230, 232 indirect 4, 66, 155, 159162, 194, 203, 206, 215, 219, 223224, 237, 265, 282 induction 30, 184, 250252, 258, 261263 infection 4, 1519, 2528, 31, 39, 4557, 6166, 6869, 7577, 7982, 129, 137, 139140, 143144, 152, 163165, 171, 177, 179181, 194, 204, 210, 219, 222, 227228, 231, 234, 236237, 240, 245246, 249, 253254, 265266, 273278, 281282, 288289, 292293 infectivity 15, 29, 32, 144, 159, 204205, 231232, 237, 249, 252255, 260, 265266, 269, 275, 281283, 286, 294 infusion 50, 52, 54 inhibitor 6, 155 inoculation 33, 48, 5657, 7677, 80, 266, 291, 294 inoculum 10, 34, 56, 81, 143, 231, 266, 291 institution 71 intestinal 34, 2528, 31, 33, 3637, 4041, 4549, 52, 5455, 57, 6466, 6869, 74, 88, 164165, 171, 177, 180181, 187, 194, 213, 219, 222, 227228, 232, 236, 249, 254255, 260, 273, 286287, 291 ionic composition 265 isoenzyme 37, 49, 5354, 289 isolation 19, 24, 31, 37, 65, 8081, 106, 109, 151, 157, 163, 172, 177, 179181, 183, 205, 209, 223, 237, 250, 254, 256, 265, 273, 294 K karyokinesis 249 kD 183 killer cells 45, 4748 Kochs' postulates 29, 31 L Lcysteine hydrochloride 250 lamblia 1, 813, 17, 19, 2126, 28, 31, 3334, 39, 45, 54, 61, 6869, 7174, 77, 82, 94,
Page 299
107, 109, 125127, 129, 135, 142144, 156157, 159164, 167168, 172173, 176184, 187189, 191, 205, 211213, 223, 225, 227, 231233, 235237, 249265, 268269, 273275, 280283, 286, 288289, 293 lateral shield 3536 lectin 46, 48 length 26, 28, 31, 34, 68, 79, 81, 97, 99, 102, 147, 149151, 198, 229231, 283284, 292 lesions 236237 Lethbridge 6164 libraries 147148, 151152, 156157, 184 life cycle 29, 31, 273 light scatter 914, 98, 113 lipid 211 lymphocyte 47, 269 lysosomal enzymes 27, 252, 268 M macrophages 4647, 49, 171172 malabsorption 1, 33, 45, 107, 177, 180, 191, 236237 maturation 88, 92, 137138, 142, 250252, 273, 293294 McKeesport 103106, 233 mechanism 6, 24, 33, 48, 53, 65, 88, 91, 99, 165, 244, 267268, 288 media 22, 8792, 97, 100101, 103, 105107, 114, 118119, 121, 137, 142, 144, 197, 200, 202, 206207, 250251, 261264, 291 median body 26, 62, 169170, 221, 231232, 287288, 292 membrane 12, 2526, 36, 98, 107112, 130, 133135, 139, 143, 154, 159, 162, 180, 198, 200203, 205209, 215, 224, 237, 252, 255, 265, 267269, 293 membrane filter 107111, 139, 198, 200, 206208 membrane integrity 268 membrane permeability 268 Meriones unguiculatus 55, 75, 7980, 139, 165, 246, 253, 256, 293 metabolic 67, 10, 27, 50, 88, 187, 189, 235, 252 metabolism 6, 9, 12, 88, 187189, 254, 260 methods 3, 9, 12, 19, 21, 31, 49, 6567, 72, 89, 94, 103, 106, 112, 129131, 135, 139, 144, 153154, 159160, 162163, 165, 170, 177, 181184, 193, 197, 199201, 204205, 208209, 213, 215, 217, 222223, 225, 234235, 249, 251253, 255259, 261265, 269, 273, 275, 281, 286288, 291, 294 metronidazole (see Flagyl) 312, 16, 18, 2124, 4952, 54, 56, 75, 7980, 165 micro foci 3940 microscope 14, 26, 2931, 34, 89, 104, 109, 133, 197, 201203, 215217, 220, 224, 236, 246, 265266 microscopy 25, 2937, 47, 50, 76, 80, 113, 116, 135, 150, 159, 164, 172, 197198, 203, 205206, 209, 216217, 219223, 225, 232, 236, 249, 252255, 260, 265266, 268269, 273, 283, 287 microtine 162, 231232, 234 microtubules 26, 48, 133, 154, 156 microvillous border 36, 236 MIF centrifugation 6163 milk 17, 4548 minichromosomes 147, 149, 151 mitochondria 252 models 3, 6, 29, 49, 103, 113, 234, 253254, 265, 282, 287, 293 Mongolian gerbil (see Meriones unguiculatus) 32, 55, 58, 61, 64, 7577, 144, 189, 249, 253254, 256, 260, 282283, 288, 293 monitoring 15, 77, 89, 99, 101, 103107, 109110, 119, 129131, 203, 215216, 223225, 235, 237, 275, 277, 279, 285286 monoclonal 4549, 55, 58, 156, 159164, 203207, 209, 216, 219222, 273, 283 monolayer 3336, 203 mononuclear leukocytes 169, 172 morbidity rate 277 morphology 2831, 3336, 49, 52, 62, 79, 113, 133134, 169, 198, 205, 231232, 234, 253, 255, 265268, 287, 292 motility 910, 12, 2123, 37, 171, 257259, 288 mouse 4, 25, 28, 32, 4555, 76, 7980, 82, 91, 109, 113, 135, 159162, 165, 175176, 187, 189, 219220, 231, 234, 251, 253, 255, 260261, 265, 269, 287, 289, 291292 mRNA 152153, 155157, 175, 183 mucosa 4647, 54, 227228 murine 31, 4849, 5455, 170171, 236 muris 4, 2425, 2728, 3032, 45, 49, 5155, 79, 98, 105, 107, 109, 111, 133, 135, 137144, 159161, 169172, 187189, 230, 232, 236237, 249259, 261 266, 268269, 273, 281282, 286288, 292294 muskrat 29, 7677, 7980, 91, 159165, 167, 202, 223, 228, 231232, 234, 236237, 239240, 245, 273274, 285, 293
Page 300
N NADPH oxidase 6 neonatal 4, 7, 151, 265266, 293 nonviable 3031, 265266, 268, 283 nucleic acid 181, 268 nude mice 4649, 55 O ondatrae 287288 ornidazole 34, 8 orthogonal field alternation 147 outbreak 4, 1519, 49, 61, 6365, 74, 82, 87, 103107, 113, 117, 129130, 137, 159, 163, 194, 205, 208209, 223, 225, 227230, 232234, 236237, 245, 255, 279, 281 outer cyst wall 29 oxygen 28, 91, 125127, 187188, 223 P PAGE (see polyacrylamide gel electrophoresis) 13, 153156, 178, 182183, 191193 pancreatin 251, 261 ParaPak MacroCon 211 particulate 97, 99, 101, 108, 110111, 197, 212, 216, 256, 286 passive 6, 4950, 52, 54, 87, 110 passive transfer of bile 54 pathogenesis 31, 3637, 69, 135, 162164, 235236, 288289 pathway 6 Pentatrichomonas hominis 40 peptone 166, 251, 261264 peripheral vacuoles 2930, 266, 268 peritrophic space 2931, 252253, 265268 person to person 1516, 1819, 47, 177, 249, 254 perspectives 69, 162, 167, 235 pets 64, 234 Peyer's patch 4649 pH 12, 26, 29, 66, 89, 100102, 108, 117119, 121, 125, 133, 135, 137139, 142143, 149, 154, 161, 177178, 181182, 188, 191192, 199, 220, 223, 243, 249252, 256, 258, 261262, 265, 269, 273, 280, 283, 286, 291 phagocytosis 49, 172 phenotype 288 physiological 2, 31, 54, 192, 211, 281 phytone 250251, 261264, 291 PI (see propidium iodide) 2930, 265268 pig 7577 plasma cell 4647 polyacrylamide gel electrophoresis (see PAGE) 1, 4, 153, 169, 178, 183, 191, 194 pregnancy 3, 7 prepatent 57, 63, 75 prevalence 18, 3941, 48, 6165, 71, 79, 81, 163, 172, 181, 204, 228, 231232, 234, 236 primary 4, 6, 9, 3940, 4952, 54, 56, 68, 95, 100, 180, 184, 194, 200, 219220, 222, 234, 241, 273 production 4, 29, 32, 4547, 58, 74, 7677, 80, 87, 89, 99, 109, 120121, 124, 127, 139, 143, 155, 160, 170, 194, 220, 222223, 253, 274, 281 propidium iodide (see PI) 12, 32, 198, 205, 237, 252253, 265, 269, 281, 286 proteolysis 268 proteose peptone 251, 261264 protozoa 7, 12, 25, 28, 31, 41, 64, 80, 88, 147, 150, 163, 165, 181, 183184, 197, 200201, 205206, 208, 213, 220, 230, 236, 254, 269, 289 protozoan 1, 15, 25, 2729, 45, 54, 64, 68, 71, 130, 155, 165, 177, 181, 187, 191, 205, 208, 211213, 222, 227, 249, 254255, 260, 286287 psittaci 236, 287288 pyruvate:ferredoxin oxidoreductase 6, 188189 Q quinacrine 36, 8, 12, 2124 R rabbits 1, 153154, 156, 169170, 191 rat 1213, 46, 4855, 81, 159162, 172, 182183, 236, 249, 254, 265, 288 raw waste water 231, 233 reducing agents 251, 262263, 265, 273 removal 66, 8788, 90103, 105106, 108, 113114, 116119, 121, 137138, 140143, 145, 155, 200, 202203, 209, 223, 237, 243, 273, 275, 277, 280281, 285, 293294 Rendtorff 255, 282, 286 reservoirs 55, 65, 69, 74, 77, 8182, 104, 117, 152, 162164, 199, 204, 209210, 223, 225, 227, 231, 233235, 245, 289 resistance 37, 16, 27, 4849, 51, 5457, 108, 172, 177, 254, 283, 286
Page 301
respiration 281 risk 15, 18, 3940, 103, 108109, 202204, 212, 223, 227, 234, 237238, 241245, 249, 275, 277279, 281, 285287, 292 risk model 275, 277, 281, 285 RNA 0, 24, 148, 155, 157, 175, 181184 S Saccharomyces cerevisiae 12, 182 SAF 6163 sample 2, 16, 23, 50, 61, 63, 72, 7576, 8992, 103105, 108, 111, 113, 125, 129, 131, 139141, 149, 166, 178179, 183, 192, 197200, 202, 204, 206208, 215216, 220, 224, 237, 242, 244245, 250, 252, 262263, 266, 275277, 279280, 285, 292294 satranidazole 3 scanning electron microscope 2930, 220 secondary infection 40, 49, 5152 secretory 4849, 5456, 58, 153, 157, 163, 180, 289 sedimentation 50, 55, 88, 95, 100104, 106, 115, 118119, 129, 181, 183, 187, 189, 200, 211213, 261262, 265, 280281 SEM 2931, 3435, 37, 113115, 172, 236, 288 sensitivity 47, 21, 2324, 28, 58, 61, 153, 198, 207, 213, 222, 256, 275 serum 12, 21, 29, 33, 4546, 4850, 52, 5456, 66, 69, 139, 154155, 159161, 169172, 177178, 180182, 184, 191192, 203, 219220, 223, 250252, 261265, 269, 291 serum factor 170 slow sand 8789, 9192, 94100, 102103, 237, 244, 285 small bowel 55 sodium bicarbonate 109, 250251, 261 sodium dodecyl sulfate 1, 178, 191, 206 SPCA 6164 speciation 52, 147, 164, 287, 294 specificity 49, 153, 159, 165, 167, 219, 222, 287288 stage 91, 156, 187, 279 staining 12, 3132, 34, 47, 55, 65, 72, 107, 112, 134, 149150, 166, 182, 202203, 205, 217, 219, 222, 237, 252253, 257, 265269, 286 stimulation 69 stool analysis 63 stools 29, 39, 57, 6163, 65, 7175, 79, 156, 179, 207, 227, 287, 291, 293294 strain identification 147 stray dogs 6164 suckling mice 274, 291, 294 sucrose gradient 80, 109, 181183, 187, 205, 291 sulfasalazine 34 suppressor 4547 surface changes 35 surface water 1516, 74, 82, 8788, 95, 97, 102103, 107, 115, 117118, 125, 130, 159, 164, 197, 204205, 208, 215217, 219, 221, 233, 237, 243, 245, 252, 275, 280, 282285 survival 27, 88, 106, 109, 111, 143, 172, 250, 262, 273, 293294 symbiont 25, 2728 symptomatic 1618, 35, 37, 48, 55, 57, 65, 139, 153, 177179, 181184, 219, 222, 251, 262264 symptomatic donor 251, 263 symptomatology 69 symptoms 4, 1618, 33, 55, 57, 68, 177, 227, 273, 292 T T lymphocyte 4548, 54 Tcell 46 taxonomic 81, 147, 172, 287 TEM 2526, 30, 37, 133, 172 temperature 2, 17, 28, 66, 8891, 93, 96, 100, 102, 106, 108, 112, 115, 118, 121, 125127, 137, 139, 141, 148, 151, 161, 166, 173, 175, 182, 192, 199, 223, 243, 249252, 255, 257258, 261, 265, 273, 283, 286 tetracycline 22, 24 therapy 35, 78, 24, 48, 65, 68 tinidazole 36, 8, 12, 24 titer 219, 221 toxin 58 TPS1 3335, 37 translation 24, 152153, 155157, 181184 transmission 4, 1519, 25, 2830, 32, 3940, 47, 49, 55, 63, 65, 69, 7577, 79, 8182, 95, 106, 112113, 117, 129, 133, 152, 162165, 167168, 180, 194, 204, 208210, 223, 225, 227, 229, 231237, 245, 249, 252, 254255, 286289, 293 transmission electron microscopy 25, 29 trichomonads 4 Trichomonas vaginalis 3 Triton X100 80, 133, 139, 177178, 182, 187, 211213 trophozoites 1, 35, 912, 2125, 2731, 3337, 4548, 5051, 5457, 6163, 6569, 7576, 7982, 107, 139, 147148, 153155, 160
Page 302
161, 165, 169173, 177, 181182, 187189, 191, 193, 219222, 227232, 235236, 249250, 252253, 255262, 264266, 269, 273, 287, 291, 294 trypanosomatids 28 trypanosome 294 trypsin 33, 109, 139, 251252, 256, 261264 trypsinTyrode's 251, 256, 262 turbidity 87, 89, 93, 95106, 116121, 124, 131, 139, 142143, 197, 199200, 202, 215216, 223, 234, 237238, 241, 243, 245, 275, 284285 TYIS 1, 9, 21, 24, 31, 56, 82, 135, 144, 147, 166, 168, 170, 177, 180181, 191, 219220, 223, 254255, 291 U ultrastructural 25, 2829, 3132, 49, 133, 135, 162163, 232, 236237, 254, 265269 ultraviolet radiation 107 V ventral disc 23, 26, 29, 36, 133, 154 ventrolateral flange 3536 Vero 3336 viability 4, 910, 12, 2124, 2832, 8081, 107, 109111, 113, 128, 133, 135, 137144, 198, 205, 220221, 225, 233, 237, 249, 252261, 265269, 275, 281 283, 286 virulence determinants 53 virus 88, 96, 98, 103, 107, 152, 183, 278279 vital stain 249 W wall 2122, 26, 2931, 48, 50, 63, 111, 133134, 161, 163, 187, 189, 201, 203, 219220, 224, 249, 252253, 258259, 265268, 273 water treatment plant 8790, 102104, 107, 117118, 129130, 133, 137139, 142145, 233, 279, 283 waterborne outbreak 15, 49, 61, 95, 113, 165, 205, 209, 227229, 233234, 236, 249, 278 watershed 87, 129, 159, 208, 223, 227, 229, 231235, 245, 275, 279, 285 Western blot 153, 177178, 273 Y Youghiogheny River 103106 Z Zibethecus ondatrae 287288 zinc sulfate flotation 213, 223224 ZnSO4 6163, 76, 202, 215217, 224 zoonotic 79, 229, 232, 234, 287288 zymodemes 288