Glycobiology and Medicine: Proceedings of the 7th Jenner Glycobiology and Medicine Symposium (Advances in Experimental Medicine and Biology Vol 564)

GLYCOBIOLOGY AND MEDICINE Proceedings of the 7th Jenner Glycobiology and Medicine Symposium

ADVANCES IN EXPERIMENTAL MEDICINE AND BIOLOGY Editorial Board: NATHAN BACK, State University of New York at Buffalo IRUN R. COHEN, The Weizmann Institute of Science DAVID KRITCHEVSKY, Wistar Institute ABEL LAJTHA, N. S. Kline Institute for Psychiatric Research RODOLFO PAOLETTI, University of Milan Recent Volumes in this Series Volume 561 CHEMISTRY AND SAFETY OF ACRYLAMIDE IN FOOD Edited by Mendel Friedman and Don Mottram Volume 562 CHOLINGERGIC MECHANISMS Edited by José Gonzalez-Ros Volume 563 UPDATES IN PATHOLOGY Edited by David C. Chieng and Gene P. Siegal Volume 564 GLYCOBIOLOGY AND MEDICINE Edited by John S. Axford Volume 565 SLIDING FILAMENT MECHANISM IN MUSCLE CONTRACTION: FIFTY YEARS OF RESEARCH Edited by Haruo Sugi Volume 566 OXYGEN TRANSPORT TO TISSUE XXVI Edited by Paul Okunieff, Jacqueline Williams, and Yuhchyau Chen Volume 567 THE GROWTH HORMONE-INSULIN-LIKE GROWTH FACTOR AXIS DURING DEVELOPMENT Edited by Isabel Varela-Nieto and Julie A. Chowen Volume 568 HOT TOPICS IN INFECTION AND IMMUNITY IN CHILDREN II Edited by Andrew J. Pollard and Adam Finn Volume 569 EARLY NUTRITION AND ITS LATER CONSEQUENCES: NEW OPPORTUNITIES Edited by Berthold Koletzko, Peter Dodds, Hans Akerbloom, and Margaret Ashwell

A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact the publisher.

GLYCOBIOLOGY AND MEDICINE Proceedings of the 7th Jenner Glycobiology and Medicine Symposium

Edited by

John S. Axford Stt George’s, University of London, UK

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN-10 ISBN-10 ISBN-13 ISBN-13

0-387-25514-1 (HB) 0-387-25515-X (e-book) 978-0-387-25514-99 (HB) 978-0-387-255115-6 (e-book)

Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands. www.springeronline.com

Printed on acid-free paper

All Rights Reserved © 2005 Springer No part of this work 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, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed in the Netherlands.

INTRODUCTION 7th Jenner Glycobiology and Medicine Symposium Sunday 5 – Wednesday W 8 September 2004

John S. Axford St George’s, University of London, UK

The potential for glycobiology to improve the practice of medicine has been well recognised, which is why biannual meetings concerning the association have been taking place for the last 14 years. The science of glycobiology has matured rapidly, and with it the far reaching clinical implications are becoming understood. The next decade is going to see this final frontier of science conquered. The impact this understanding of glycobiology will have upon our practice of medicine is going to be exciting. The 7th Jenner Glycobiology and Medicine Symposium was designed to reflect these advances. All the major clinical areas were involved, with contributions from pivotal players in science and medicine. As with our previous meetings, junior scientists were involved as we recognise that at the end of the next decade they will be in the driving seat. This introduction serves as a taster to whet your appetite. From embryogenesis to pathogenesis, glycosylation plays a pivotal role. Complex and hybrid N-glyans and O-fucose glycans are critical in oocyte development and function. This area must surely be a fertile ground for glycosylation research. The pathogenesis of viral infections involves sugars at every turn. Hepatitis C virus and Bovine viral diarrhoea virus are diseases that are opening themselves to scrutiny. The BVDV has proven very useful in the evaluation of the antiviral activity of molecules that inhibit morphogenesis and/or viral entry. Infection by human immunodeficiency virus type-1 is characterised by low levels of neutralising antibodies. One broadly neutralising human monoclonal antibody is 2G12.This has 3 possible combining sites and recognises a cluster of oligomannose residues on the ‘‘immunologically silent’’ face. This recognition provides exciting challenges for immunogen design. v John S. Axford (ed.), Glycobiology and Medicine, v-viii. © 2005 Springer. Printed in the Netherlands.

vi

J. S. Axford

N- and O-linked glycosylation of enveloped glycoproteins permits Ebola virus binding to host cells. It is thought that an alternative pathway to the calnexincalreticulin folding and quality control pathway is being used by the viral glycoproteins. Immune mechanisms will be a major focus for clinical intervention over the next decade. This will involve modulation of both the innate and adaptive immune defences. The innate immune system provides the first line of defence to invading pathogens, and recognition of pathogens governs the induction, and type of pathogenspecific adaptive responses. Schistosomiasis is a major tropical parasitic disease. Recently several antigen presenting cell-associated lectins, that show interaction with egg glycoproteins of S. mansoni, such as the dendritic cell-specific DC-SIGN, L-SIGN on liver sinusoid endothelial cells, MGL (macrophage galactose-type lectin) and galectin 3, have been identified. Since their glycan ligands occur on many parasitic helminths, that they may constitute parasite patterns for lectin-mediated immune recognition. For example, Lex interacts with DC-SIGN on dendritic cells and it is thought that this interaction may play a role in triggering dendritic cells s to mount to a Th2 response. Mannan binding lectin (MBL) is an oligomeric protein designed to recognize pathogen-associated molecular patterns. The biological importance of MBL was indicated when opsonin-deficient children with recurrent infections were found to be genetically deficient in MBL. Further interest in this molecule was sparked by the observation of complement activation upon binding to carbohydrates. Glycosylation is now known to play a central role in the MBL pathway of complement activation and the glycosylation of the mannose receptor determines its functional specialisation. Glycan structures that can act as potential ligands for MBL have been identified on all the immunoglobulins. In human serum only IgG-G0 and polymeric and dimeric IgA have been shown to bind MBL and initiate the lectin pathway of complement. This is thought to occur through GlcNAc-terminating glycan structures. Disease associations with sugar changes are plentiful, when the adaptive immune system is considered. This may involve fundamental processes, for example glycosylation related molecular mechanisms are thought to involve the function of the T cell co-receptor CD8; which will have far reaching implications if abnormal. Sugar associations with cancer have been recognised for some time. There continues to be new data concerning ovarian cancer and arthritis, but research is expanding into new areas. Sugars have now been shown to be associated with the pathological mechanism associated with the GPI anchorage of the prion protein, pigeon fanciers’ lung and muscular dystrophy. At least six different ff forms of muscular dystrophy are caused by genes that encode glycosyltransferases, and when malfunctioning result in a secondary deficiency in the glycosylation of dystroglycan. Autoimmune arthritis has been associated with the generation of remnant glycoepitopes by gelatinase B.gelatinase B/matrix metalloproteinase-9, which is an inflammatory mediator and effector. ff Considerable amounts of gelatinase B are released by neutrophils in the synovial cavity of patients with rheumatoid arthritis. This is thought to be linked to the pathogenesis of arthritis as gelatinase B-deficient mice are resistant to antibody-induced arthritis. Determination of T-cell reactivity

Introduction

vii

against the gelatinase B-cleaved fragments of collagen II indicates that, there are many glyco-epitopes present in collagen II, which reinforces the role of glycopeptide antigens in autoimmunity. It is however always exciting when clinical anecdotes translate into therapeutic and diagnostic possibilities. Glycobiology is at that transition. Once disease mechanisms have been understood, the next step is to determine whether this information can be used to devise therapeutic options. Predictably, therapeutic hypotheses are plentiful. For example, there is a possibility that polysaccharide may be used to block skin inflammation. There have been new developments in treating glycosphingolipid storage diseases. This may not be a common group of diseases, but for those that have it this opens the door to improved quality of life. The glycosphingolipid lysosomal storage diseases result from defects in glycosphingolipid catabolism. They are progressive disorders, the majority of which involve storage and pathology in the central nervous system. A new approach to treatment is substrate reduction therapy (SRT), using small molecule inhibitors to reduce the rate of glycosphingolipid biosynthesis, to offset ff the catabolic defect. One of these drugs, NB-DNJ has recently been approved for clinical use in type 1 Gaucher disease, following a successful clinical trial. There is also the potential of combining SRT with drugs that target the downstream consequences of storage. RA is a common disorder where the available diagnostic tests eg rheumatoid factor, anti-citrulinated cyclic peptide, lack sensitivity. The diagnostic potential of IgG glycosylation has been previously discussed and we await the results from prospective trials. Indeed, abnormal galactosylation of polyclonal IgG in ANCA associated systemic vasculitis patients has now been reported and the diagnostic potential of this technology for other autoimmune rheumatic diseases is significant. Experiments looking at the cause behind these sugar changes indicate both quantitative and qualitative changes in the RA serum GTase isoform profile. This is likely to be due to a greater proportion of hypersialylated isoforms, which have the potential to adversely affect ff the catalytic activity of the enzyme, thus providing a possible mechanism for post-translational regulation of GTase activity in RA. It also provides further evidence that RA glycosylation changes may be more general than previously indicated and encompass proteins other than IgG. These observations can only strengthen the potential of sugars as RA disease biomarkers. The selectin family of adhesion molecules mediates the initial attachment of leukocytes to venular endothelial cells at sites of tissue injury and inflammation. For example in staphylococcal arthritis. Fucoidin, a sulfated polysaccharide from seaweed, binds to and blocks the function of L-and P-selectins thereby inhibiting leukocyte rolling and adhesion to endothelial surface. Treatment with fucoidin has been shown to reduce the severity of septic arthritis within the first three days following bacterial infection. It is suggested that the efficient treatment of septic arthritis should encompass a combination of antibiotics and immunomodulation. Gastroenterologists want to know more about normal and abnormal bacteria that inhabit our bowels. O-acetyl sialic acid expression in colorectal mucosa has been shown to be regulated by enteric microflora; as demonstrated by the loss of sialic acid oligo-O acetylation after elimination of the faecal flow. There is therefore

viii

J. S. Axford

potential to use this observation to quantitate bacterial colonisation and perhaps interfere with disease associated pathology. Biological therapies will be the new treatments of the next decade. The impact of glycosylation on the structure and function of natural and recombinant (therapeutic) IgG antibody is therefore important to get to grips with. It has been shown that the in vivo micro-environment can have a profound influence on the glycosylation profile of IgG-Fc. This may reflect the unique structural relationship between the oligosaccharide and the protein. The ‘‘core’’ heptasaccharide is essential for FccRI, FccRII, FccRIII and C1 activation whilst outer arm sugar residues can influence these and other functions, e.g. FccRIII, FcRn, MBL, MR. Thus, fidelity of glycosylation is essential to the effector ff function profile of antibodies and in the future the oligosaccharide will be used to function as a structural ‘‘rheostat’’ to generate specific glycoforms exhibiting optimal effector ff activities for a particular disease target. Cellular glycoengineering for fully human glycosylation and optimised sialylation of proteins is therefore going to be important if these molecules are going to be fully and specifically active. Most pharmaceutical proteins are expressed in bacteria, yeast or mammalian cells resulting in proteins lacking glycosylation or carrying glycans which largely differ ff from human carbohydrate chains in various aspects including sialylation. However, relationships between the N-glycan structures and biological activities of, for example, recombinant human erythropoietins produced using different ff culture conditions and purification procedures are now better understood. It is nevertheless apparent that novel glycoprotein expression technology will need to be developed to address this problem and data is now available to demonstrate how this can be done. The above introduction adds up to the fact that Glycobiology is an extremely exciting science to be involved in. Additionally, if you are a clinician, it is even more gripping as you will be at the forefront of important clinical developments. I hope these proceedings stimulate you as much as they have me and I look forward to seeing you at Jenner 8!

REFERENCES Axford J Keida C Van Dijk WV, Rudd PM. 6th Jenner Glycobiology and Medicine. CPD Bulletin. Immunology & Allergy 2004; 3(3): 84–87. Alavi A, Axford J. Glycobiology of the Rheumatic Diseases: an update. Adv Exp Med Biol. 2003; 535: 271–80. Axford J, Keida C & Dikk WV. Meeting Report 5th Jenner Glycobiology and Medicine. Glycobiology 2001; 11(2): 5G-7G. Axford JS. 4th Jenner International Glycoimmunology Meeting: A Review. Immunol Today 1997; 18(11): 511–513. Axford JS. 3rd Jenner International Glycoimmunology Meeting. Immunol Today 1995; 16(5):213–215. Axford JS. 2nd Jenner International Glycoimmunology Meeting. Immunol Today 1993; 14(3): 104–106.

CONTENTS

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

1. Glycosylation: Disease Targets and Therapy . . . . . . . . . . . . . . . . . . . . . . . Nicole Zitzmann, Timothy Block, Anund Methta, Pauline Rudd, Dennis Burton, Ian Wilson, Frances Platt, Terry Butters, and Raymond A. Dwek

1

2. Long Alkylchain Iminosugars Block the HCV P7 Ion Channel . . . . . . . D. Pavlovic, W. Fischer, M. Hussey, D. Durantel, S. Durantel, N. Branza-Nichita, S. Woodhouse, R. A. Dwek, and N. Zitzmann

3

3. The Bovine Viral Diarrhoea Virus: A Model for the Study of Antiviral Molecules Interfering with N-Glycosylation and Folding of Envelope Glycoprotein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Durantel, N. Branza-Nichita, S. Durantel, R.A. Dweek, and N. Zitzmann 4. Antibody Recognition of a Carbohydrate Epitope: A Template for HIV Vaccine Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chris Scanlan, Daniel Calarese, Hing-Ken Lee, Ola Blixt, Chi-Huey Wong, Ian Wilson, Dennis Burton, Raymond Dwek, and Pauline Rudd 5. Interaction of Schistosome Glycans with the Host Immune System . . . . Irma van Die, Ellis van Liempt, Christine M. C. Bank, and Wietske E. C. M. Schiphorst 6. The Mannan-Binding Lectin (MBL) of Complement Activation: Biochemistry, Biology and Clinical Implications . . . . . . . . . . . . . . . . . . . . Jens Christian Jensenius 7. Killer Cell Lectin-like Receptors and the Natural Killer Cell Gene Complex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ø. Nylenna, L. M. Flornes, I. H. Westgaard, P. Y. Woon, C. Naper, J. T. Vaage, D. Gauguier, J. C. Ryan, E. Dissen, and S. Fossum 8. Glycosylation Influences the Ligand Binding Activities of Mannose Receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yunpeng Roc Su, Clarence Tsang, Talitha Bakker, James Harris, Siamon Gordon, Raymond A. Dwek, Luisa Martinez-Pomares and Pauline M. Rudd

ix

5

7

9

21

23

25

x

9. Human Immunoglobulin Glycosylation and the Lectin Pathway of Complement Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . James N. Arnold, Louise Royle, Raymond A. Dwek, Pauline M. Rudd, and Robert B. Sim 10. Gelatinase B Participates in Collagen II Degradation and Releases Glycosylated Remnant Epitopes in Rheumatoid Arthritis . . . . . . . . . . . P. E. Van den Steen, B. Grillet, and G. Opdenakker

27

45

11. Hyaluronan in Immune Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Alan J. Wright and Anthony J. Day

57

12. Glycosylation and the Function of the T Cell Co-receptor CD8 . . . . . . David A. Shore, Ian A. Wilson, Raymond A. Dwek, and Pauline M. Rudd

71

13. Immunogenecity of Calreticulin-bound Murine Leukemia Virus Glycoprotein gp90 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yusuke Mimura, Denise Golgher, Yuka Mimura-Kimura, Raymond A. Dwek, Pauline M. Rudd, Tim Elliott

85

14. Glycosylation and GPI Anchorage of the Prion Protein . . . . . . . . . . . . N. M. Hooper

95

15. Glycosylation Defects and Muscular Dystrophy . . . . . . . . . . . . . . . . . . . Derek J. Blake, Christopher T. Esapa, Enca Martin-Rendon, and R. A. Jeffrey ff McIlhinney

97

16. Roles of Complex and Hybrid N-Glycans and O-Fucose Glycans in Oocyte Development and Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S. Shi, S. A. Williams, H. Kurniawan, L. Lu, L., and P. Stanley 17. Mucin Oligosaccharides and Pigeon Fanciers’ Lung . . . . . . . . . . . . . . . C. I. Baldwin, A. Allen, S. Bourke, E. Hounsell, and J. E. Calvert 18. Differential ff Glycosylation of Gelatinase B from Neutrophils and Breast Cancer Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simon A. Fry, Philippe E. Van den Steen, Louise Royle, Mark R. Wormald, Anthony J. Leathem, Ghislain Opdenakker, Pauline M. Rudd, and Raymond A. Dwek 19. Detection of Glycosylation Changes in Serum and Tissue Proteins in Cancer by Lectin Blotting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R. E. Ferguson, D. H. Jackson, R. Hutson, N. Wilkinson, P. Harnden, P. Selby, and R. E. Banks

99 101

103

113

20. Carbohydrates and Biology of Staphylococcal Infections . . . . . . . . . . . . Andrej Tarkowski, Margareta Verdrengh, Ing-Marie Jonsson, Mattias Magnusson, Simon J Foster, and Zai-Quing Liu

115

21. New Developments in Treating Glycosphingolipid Storage Diseases . . Frances M. Platt, Mylvaganam Jeyakumar, Ulrika Andersson, Raymond A. Dwek and Terry D. Butters

117

22. Fucosylated Glycans in Innate and Adaptive Immunity . . . . . . . . . . . . J. B. Lowe

127

xi

23. New Insights into Rheumatoid Arthritis Associated Glycosylation Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Azita Alavi, Andrew J. Pool, and John S. Axford 24. Production of Complex Human Glycoproteins in Yeast . . . . . . . . . . . . Tillman Gerngross 25. Relationship Between the N-Glycan Structures and Biological Activitities of Recombinant Human Erythropoietins Produced Using Different ff Culture Conditions and Purification Procedures . . . . . . . . . . . C-T. Yuen, P. L. Storring, R. J. Tiplady, M. Izquierdo, R. Wait, C. K. Gee, P. Gerson, P. Lloyd, and J. A. Cremata

129 139

141

26. Glycosylation of Natural and Recombinant Antibody Molecules . . . . . Roy Jefferis ff

143

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

149

1

GLYCOSYLATION: DISEASE TARGETS AND THERAPY

Nicole Zitzmann1, Timothy Block2, Anund Methta2, Pauline Rudd1, Dennis Burton2, Ian Wilson3, Frances Platt1, Terry T Butters1, and Raymond A. Dwek1 1Oxford University Department of Biochemistry Glycobiology Institute, Oxford 1Jefferson ff Centre Doylestown, USA 3The Scripps Research Institute La Jolla, USA

Four different ff glycosylation approaches illustrate strategies for providing therapy in disease targets. Two of these are to develop antiviral therapies using iminosugar derivatives. The first approach involves targeting virus encoded protein(s), the second targets host cell encoded protein(s) necessary for virus survival. The latter could potentially prove more resistant to the development of viral escape mutants, a problem plaguing most conventional drug therapies. In the case of hepatitis C virus (HCV), which affects ff about 3% of the world population, it is possible to employ both strategies. Using bovine viral diarrhoea virus (BVDV) as a model organism for HCV we showed that inhibition of the host cell ER alpha-glucosidases I and II using the glucose analogue deoxynojirimycin (BuDNJ) leads to an antiviral activity caused by a reduction in viral secretion due to the interference with viral envelope glycoprotein folding and subsequent impairment of viral morphogenesis. However, it is also possible to target the virally encoded HCV protein p7 which can form ion channels, using long alkylchain derivatives of either DNJ or the galactose analogue deoxygalactonojirimycin (DGJ). N7oxanonyl-6deoxy-DGJ is currently in phase I clinical studies. Worldwide, more than 350 million people are chronically infected with hepatitis B virus (HBV). Glucosidase inhibitors have been shown to be antiviral against HBV in tissue culture and in the woodchuck model of chronic HBV infection. The M surface antigen glycoprotein of HBV folds via the calnexin pathway. Glucosidase 1 John S. Axford (ed.), Glycobiology and Medicine, 1-2. © 2005 Springer. Printed in the Netherlands.

2

N. Zitzmann et al.

inhibitors which prevent this interaction prevent the formation and secretion of HBV. The misfolded M surface antigen is retained within the cell and may itself act as a long lived ‘‘drug’’ which prevents virus formation. As in the case of HCV, a second class of long alkylchain iminosugars which do not inhibit glycan processing are also potent antiviral agents and may act at a stage before viral envelopment, but the mechanism is still unknown. Toxicology studies are underway to identify a compound from these classes of orally available drugs for clinical trials for the treatment of chronic HBV infection. The humoral immune response to infection by human immunodeficiency virus type-1 (HIV-1) is characterised by low levels of neutralising antibodies, particularly those which have a broad specificity against many different ff isolates. One broadly neutralising human monoclonal antibody is 2G12. This has a novel antibody structure with 3 possible combining sites and recognises a cluster of oligomannose residues on the ‘‘immunologically silent’’ face. This recognition provides exciting challenges for immunogen design. The use of imino sugars as with other viruses may also provide additional possibilities for antiviral therapy. T he glucosphingolipid (GSL) storage diseases are a family of progressive disorders in which GSL species are stored in the lysosome, as a result of defects in the enzymes of the GSL degradation pathway. Specific diseases include Gaucher, Tay-Sachs, Fabry, Sandhoff and GM1 gangliosidosis. GSL storage diseases occur at a collective frequency of 1 in 18,000 live births and are one of the most common cause of neurodegenerative disease in infants and children. Our drug based strategy for management of these diseases is to inhibit partially GSL synthesis using imino sugars. Slowing the rate of synthesis of GSLs will lead to fewer entering the lysosome for catabolism, reducing the rate of storage. This substrate reduction therapy (SRT) has lead to a world wide approved oral drug (NB-DNJ) for Gaucher type-1 disease.

2

LONG ALKYLCHAIN IMINOSUGARS BLOCK THE HCV P7 ION CHANNEL

D. Pavlovic1, W. Fischer1, M. Hussey1, D. Durantel2, S. Durantel2, N. Branza-Nichita3, S. Woodhouse1, R. A. Dwek1, and N. Zitzmann1 1Oxford Glycobiology Institute Department of Biochemistry University of Oxford, Oxford, UK 2Virus des hepatites et pathologies associees INSERM U271, Lyon, France 3Institute of Biochemistry Splaiul Independentei Bucharest, Romania

The small p7 protein of the hepatitis C virus (HCV) and the closely related bovine viral diarrhea virus (BVDV) can form ion channels in artificial membranes (see

Figure 1. Channel recordings of synthetic HCV p7 inserted into a black lipid membrane (BLM). Channel

activity is shown for +/ _ 50 mV and +/ _ 100 mV. The closed state is shown as a solid line. Channel openings are deviations from this line. Solutions are the same in the cis and trans chamber: 0.5 M KCl, 5 mM Hepes, 1 mM CaCl2, pH 7.4. HCV p7 is added on the trans side to a final concentration of approximately 50 microM. Scale bars are 10 s and 100 pA. 3 John S. Axford (ed.), Glycobiology and Medicine, 3-4. © 2005 Springer. Printed in the Netherlands.

4

D. Pavlovic et al.

Fig. 1). Ion channel activity can be suppressed by long alkylchain iminosugar derivatives, which have been shown to be antiviral in BVDV infectivity assays. Treatment with these inhibitors does not affect ff viral RNA replication. However, the infectivity of virions secreted in the presence of the inhibitors is impaired. The physiological role of the p7 ion channel during the viral life cycle is unknown and is being investigated using inhibitory iminosugars as well as a BVDV construct from which p7 has been deleted.

REFERENCES 1. Study of the mechanism of the antiviral action of iminosugar derivatives against Bovine Viral Diarrhea Virus. D. Durantel, N. Branza-Nichita, S. Carrouee-Durantel, T. D. Butters, R. A. Dwek and N. Zitzmann (2001), Journal of Virology 75 (19), 8987–8998 2. The hepatitis C virus p7 protein forms an ion channel which is inhibited by long alkylchain iminosugar derivatives. D. Pavlovic, D. C. A. Neville, O. Argaud, B. Blumberg, R. A. Dwek, W. B. Fischer and N. Zitzmann (2003), PNAS 100 (10), 6104–610 3. Effect ff of interferon, ribavirin and iminosugar derivatives on viral infection in cells persistently infected with non-cytopathic BVDV: A comparative study. D. Durantel, S. Carrouee-Durantel, N. BranzaNichita, R. A. Dwek and N. Zitzmann (2004), Antimicrobial Agents and Chemotherapy 48 (2), 497–504

3

THE BOVINE VIRAL DIARRHOEA VIRUS: A MODEL FOR THE STUDY OF ANTIVIRAL MOLECULES INTERFERING WITH N-GLYCOSYLATION AND FOLDING OF ENVELOPPE GLYCOPROTEIN

D. Durantel1, N. Branza-Nichita2, S. Durantel1, R. A. Dweek3, and N. Zitzmann3 1Laboratoire des virus hepatiques et pathologies associees INSERM U271, Lyon, France 2Institute of Biochemistry Sector 6, Bucharest, Romania 3Glycobiology Institute Department of Biochemistry University of Oxford, Oxford, U.K.

The current treatment of chronic hepatitis C combines interferon alpha and ribavirin and is effective ff in only half of the patients treated. Considerable eff fforts are being made to develop novel anti-HVC molecules with a better efficacy particularly for refractory patients. Molecules targeting specifically viral activities are the most studied. However, an antiviral strategy based uniquely on the utilisation of this type of molecules is expected to encounter problems caused by the emergence of viral escape mutants, as already widely described for HIV and HBV. Alternative approaches and molecules are needed to complement antiviral strategies based on inhibitors of viral enzyme. Ideally, new molecules should target steps of the viral cycle that are potentially less likely to give rise to resistance. The assembly and morphogenesis of HCV belong to these yet untargeted steps of the life cycle. As no cellular system able to support the morphogenesis and secretion of HCV particles is yet available, the bovine viral diarrhoea virus (BVDV), which is phyllogenetically close and shares biological features with HCV, has been used as a surrogate model for the study of antiviral molecules interfering with the N-glycosylation and folding of viral envelope glycoprotein. We have demonstrated that some analogues of glucose (deoxynojirimycin), also 5 John S. Axford (ed.), Glycobiology and Medicine, 5-6. © 2005 Springer. Printed in the Netherlands.

6

D. Durantel et al.

generically called iminosugars, are good inhibitors of morphogenesis and prevent viral re-entry by reducing the infectivity of released virions. The mechanism of action has been studied at molecular level for these iminosugars presenting antiviral activity against BVDV. Two different ff mechanism of action have been defined to explain the whole antiviral effect. ff DNJ derivatives inhibit host ER a-glucosidases, thus preventing the trimming of 2 glucoses from triglycosylated N-glycans and the subsequent interaction with lectin chaperone. This inhibition results in the misfolding of viral glycoprotein and the subsequent defect in assembly, budding and viral secretion. Moreover, DNJ derivatives induce a diminution of viral infectivity and therefore prevent re-infection of cells by neo-formed particles. This is likely due to the incorporation of non functional envelope glycoprotein complexes. In conclusion, the BVDV has proven very useful to evaluate the antiviral activity of molecules that inhibit morphogenesis and/or viral entry. The BVDV will remain an interesting model for HCV while waiting for the development of a cell culture system able to fully propagate the latter.

4

ANTIBODY RECOGNITION OF A CARBOHYDRATE EPITOPE: A TEMPLATE FOR HIV VACCINE DESIGN

Chris Scanlan1,2, Daniel Calarese2, Hing-Ken Lee2, Ola Blixt2, Chi-Huey Wong2, Ian Wilson2, Dennis Burton2, Raymond Dwek1, and Pauline Rudd1 1Glycobiology Institute University of Oxford South Parks Rd, Oxford 0X1 3QU 2The Scripps Research Institute 10550 N.Torrey Pines La Jola, CA 92037

The humoral response to HIV-1 infection is typically characterized by low levels of neutralizing antibodies, especially antibodies which can provide sterilizing immunity against a wide range of HIV isolates. However, a small number of antibodies, isolated from infected individuals, have been shown to protect against HIV challenge in

Figure 1. Most of the antigenic surface of HIV-1 gp120 is glycosylated. 7 John S. Axford (ed.), Glycobiology and Medicine, 7-8. © 2005 Springer. Printed in the Netherlands.

8

C. Scanlan et al.

animal models. As such, these antibodies are potential templates for HIV vaccine design. One such antibody is the broadly neutralizing antibody 2G12. Alanine scanning mutagenesis, glycosidase digests and competition experiments demonstrated that 2G12 binds to a cluster of alpha1
2-linked mannose residues on the outer face of gp120. Cyrstallographic studies showed that IgG 2G12 exhibits a unique domainexchanged Fab configuration. Mutagenesis of 2G12 Fab, combined with carbohydrate inhibition experiments explained how the unusual structure of 2G12 is able to recognize its neutralization epitope on gp120. Synthetic mimics of the 2G12 epitope are currently under evaluation as potential immunogens.

REFERENCES 1. Angew Chem Int Ed Engl. 2004 Feb 13;43(8):1000. 2. Science. 2003 Jun 27;300(5628):2065–71. 3. J Virol. 2002 Jul;76(14):7306–21.

5

INTERACTION OF SCHISTOSOME GLYCANS WITH THE HOST IMMUNE SYSTEM

Irma van Die, Ellis van Liempt, Christine M. C. Bank, and Wietske E. C. M. Schiphorst Department of Molecular Cell Biology and Immunology VU University Medical Center Van der Boechorststraat 7, 1081 BT Amsterdam the Netherlands

1. INTRODUCTION Schistosomiasis is a parasitic disease caused by trematodes that affects ff more than 200 million people worldwide, mostly children in developing countries. Annually, 200 000 deaths are estimated to be associated with schistosomiasis (van der Werf et al., 2002). Until now, attempts to control infection and disease have mostly failed but the disease can be effectively ff treated by chemotherapy (Praziquantel). One of the most striking features of schistosomiasis is that the worms are experts in modulation and evasion of the host immune response, to enable their survival, migration and development in different ff host tissues. It is becoming increasingly clear that schistosome glycoconjugates play a crucial role in the evasion mechanisms that are exploited by the parasites. Here we will summarize our studies that aim to increase our molecular understanding of the role of specific schistosome glycan antigens in immune modulation. We will focus on the interactions of schistosome glycans with the host immune system that result in the mounting of T helper cell 2 (Th2) responses.

2. THE LIFE CYCLE OF SCHISTOSOMES Schistosomes have a complicated life cycle, requiring two hosts. Male and female worms live in the veins of the abdomal cavity of their vertebrate host, sometimes for more than 20 years. Here they mate and produce eggs. Schistosome eggs that become 9 John S. Axford (ed.), Glycobiology and Medicine, 9-19. © 2005 Springer. Printed in the Netherlands.

10

I. van Die et al.

lodged within host tissues are the major cause of pathology. Part of the eggs escape from the body and eventually reach the water with faeces or urine. In the water miracidia hatch from the eggs and infect a suitable snail host. After asexual reproduction in the snail, cercariae leave the snail and can penetrate the skin of their vertebrate host when they come into contact with the water. While penetrating the skin, cercariae loose their tails and become schistosomula that subsequently migrate via the lungs to the veins of the abdomal cavity and develop to sexually mature adult worms. Three major schistosome species are discriminated that infect humans. Although their life cycles basically are similar, each Schistosome has a specific snail intermediate host that is essential for their development. Schistosoma mansoni that occurs mainly in Africa, the Middle East and South America requires snails of the genus Biomphalaria as intermediate host. S. haematobium, found in Africa as well as in parts of the Middle East and Asia, is transmitted by snails of the genus Bulinus, whereas S. japonicum occurs mainly in South-east Asia and China and is transmitted by snails of the genus Oncomelania. All three species can also infect rodents, which are often used as model systems and are required, in combination with the specific snail intermediate hosts, to maintain the cycle in the laboratory for scientific research.

3. SCHISTOSOME GLYCANS AND THE SYNTHESIS OF NEOGLYCOPROTEINS Glycans antigens are abundantly present on the surface of the different ff parasite stages and within their excretory/secretory products. Several reviews have summarized the structures of glycan antigens found, as well as data that demonstrate that these glycans are the major focus of the host immune response (Cummings and Nyame, 1996; Cummings and Nyame, 1999; Hokke and Deelder, 2001). Remarkable is the absence of sialylation in the schistosome glycans, and the high degree of fucosylation in structural compositions that are not found in humans. The glycans can be very large, consisting of many different ff monosaccharides that diff ffer in sequence and anomeric linkage. Therefore, one glycan molecule can encompass different ff antigenic determinants (glycan antigens, see Fig. 1A). To elucidate their functional role in the host immune response, it is essential that individual glycan antigens can be studied separately. Such glycan antigens have been synthesized in vitro using enzymatic methods and subsequently coupled to BSA or other suitable carriers to yield neoglycoconjugates (van Remoortere et al., 2000) (Fig. 1B). Basically, a chemical, enzymatic, or a combined chemo-enzymatic approach can be applied for in vitro glycan synthesis. A drawback of chemical synthesis is that it requires a complete control of the stereoselectivity. For enzymatic synthesis glycosidases and glycosyltransferases (GTs), Golgi enzymes that are involved in the in vivo biosynthesis of carbohydrate moieties on glycoproteins, are used. GTs catalyze the transfer of sugar moieties from activated nucleotide-sugars to specific acceptor molecules, and act sequentially to built an oligosaccharide on a carrier molecule. The use of GTs offers ff significant advantages because it is fast and combines a high regioand stereospecificity with the potential availability of many different ff glycosidic linkages. However, although many mammalian glycosyltransferases are cloned and available in recombinant form for glycan synthesis, the number of glycosyltransferases

Interaction of Schistosome Glycans

11

Figure 1. Schistosome glycan antigens and neoglycoconjugates. A. Structure of a schistosome N-glycan (Srivatsan et al., 1992a), with different ff glycan antigens, LDN, LDNF and core-fucose, indicated. B. Construction of a neoglycoconjugate carrying LDNF. The glycan antigen is synthesized using different ff glycosyltransferases and coupled to BSA (van Remoortere et al., 2000). Table 1. Glycosyltransferases derived from invertebrate and plant sources that may be useful for the synthesis of typical helminth glycan antigens Glycosyltransferase

Source

Recombinant

Reference

b4-GalNAcT (
GlcNAcb) b4-GalNAcT (
GlcNAcb) Core a3-FucT

L . stagnalis C. elegans A. thaliana Drosophila A. thaliana C. elegans T . ocellata L . stagnalis L . stagnalis

no yes yes yes yes yes no yes no

(Mulder et al., 1995) (Kawar et al., 2002) (Bakker et al., 2001; Fabini et al., 2001)

Core b2-XylT a2 FucT (
Xylb) a2 FucT (
Fuca) b4 GlcNAcT (
GlcNAcb) b4 GlcT (
GlcNAcb)

(Strasser et al., 2001) (Zheng et al., 2002) (Hokke et al., 1998) (Bakker et al., 1994) (Van Die et al., 2000)

that can be applied to synthesize typical schistosome or other helminth glycan structures is limited. Since schistosomes share glycan antigens with plants and other invertebrates such as snails or the free-living nematode Caenorhabditis elegans, glycosyltransferases derived from these sources are useful for synthesis purposes. In several studies, recombinant GTs cloned from C. elegans or Arabidopsis thaliana, as well as extracts from schistosomes and snails have been exploited to synthesize schistosome glycan antigens such as LDN-DF or core-xylose/ core-fucose. In Table 1 a number of non-human glycosyltransferases that have been applied (van Remoortere et al., 2003b; Remoortere et al., 2000; Nyame et al., 2004), or may be useful for the synthesis of Schistosome glycan antigen, are summarized.

12

I. van Die et al.

4. SCHISTOSOME GLYCANS ARE THE MAJOR FOCUS OF THE HOST IMMUNE RESPONSE Schistosome glycans play an important role in the hosts humoral and cellular immune responses to infection. In Fig. 1, some of the major carbohydrate antigens within Schistosome egg antigens (SEA) that are referred to in this review, are shown. A strong humoral response has been found against the fucosylated glycan epitopes GalNAcb1-4(Fuca1-3)GlcNAc (LDNF), GalNAcb1-4(Fuca1-2Fuca1-3)GlcNAc (LDN-DF) and Fuca1-3GalNAcb1-4GlcNAc (FLDN) in both infected animals and humans (Nyame et al., 2000; van Remoortere et al., 2001; Eberl et al., 2002; van Remoortere et al., 2003a; Naus et al., 2003). Sera of infected hosts also contain low amounts of antibodies against Galb1-4(Fuca1-3)GlcNAc (Lex, CD15), a glycan epitope shared by humans and schistosomes (Nyame et al., 1998; van Remoortere et al., 2001; Eberl et al., 2002). These Lex antibodies may induce autoimmune reactions, as was shown by their ability to mediate complement-dependent cytolysis of myeloid cells and granulocytes (Nyame et al., 1996; Nyame et al., 1997; Van Dam et al., 1996). Interestingly, Lex occurs on all parasite stages, i.e. cercariae, eggs, schistosomula and adult worms (Srivatsan et al., 1992b; Cummings and Nyame, 1996; van Remoortere et al., 2000). Lex determinants have been found as repeating trisaccharides in tri- and tetra-antennary N-glycans of membrane glycoproteins (Srivatsan et al., 1992b), as core-2-based O-glycans on the secreted circulating cathodic antigen (Van Dam et al., 1994) and on cercarial glycolipids (Wuhrer et al., 2000). Interestingly, the immune system discriminates between Lex expressed in monomeric and polymeric form as shown by different ff types of antibody responses in mice towards these glycan structures (Van Roon et al., 2004). Lex containing glycoconjugates also induce proliferation of B-cells from infected animals, which V and Harn, 1994), and induce the production of secrete IL-10 and PGE (Velupillai 2 IL-10 by peripheral blood mononuclear cells from schistosome-infected individuals (Velupillai V et al., 2000). In a murine schistosome model, Lex is an effective ff adjuvant for induction of a Th2 response (Okano et al., 2001), and it has been demonstrated that sensitization with Lex results in an increased cellular response towards SEAcoupled beads implanted in the liver and to the formation of granulomas (Jacobs et al., 1999).

5. DENDRITIC CELLS RECOGNIZE PARASITE-DERIVED GLYCANS Dendritic cells (DCs) form a link between innate and adaptive immunity and are therefore crucial in the defense against pathogens, such as schistosomes (Palucka and Banchereau, 2002; Janeway, Jr. and Medzhitov, 2002). They are localized in peripheral tissues throughout the body and recognize invading pathogens using pattern recognition receptors including Toll-like receptors and lectins. Lectins are proteins that contain carbohydrate recognition domains (CRDs) that specifically

Interaction of Schistosome Glycans

13

bind to a variety of sugars present on cell-wall or secreted glycoconjugates. Several different ff types of lectins are discriminated that diff ffer in carbohydrate specificity, mechanism of binding, and function. Many C-type lectins (showing Ca2+-dependent binding of their carbohydrate ligands) function in the capture of glycoconjugates and subsequent presentation of the antigen to the immune system (Weis et al., 1998; Figdor et al., 2002; Engering et al., 2002). Current views are that the principal function of at least some C-type lectins is the recognition and clearance of glycosylated self-antigens, in order to induce tolerance (‘t Hart and Kooyk, 2004). This function may then be exploited by pathogens to escape immune attack. It is becoming evident that a pathogen-DC interaction is mediated by multiple sets of ligandreceptor interactions to generate a pathogen-specific response, and several distinct DC subsets have been identified to express different ff receptors. By targeting different ff DC subsets, an invading pathogen can trigger mixed responses, and it is thought that interaction of glycans with their receptor-lectin can either enhance or oppose TLR signalling thereby modulating the DC phenotype and outcome of the induced immune response (Gantner et al., 2003; Geijtenbeek T.B. et al., 2002). Some C-type lectins contain ITIM or ITAM sequence motifs in their cytoplasmic tails that indeed suggest a potential role in either immunosuppression or immunoactivation (Figdor et al., 2002). After internalization of bound components to allow antigen-processing and presentation, DCs migrate to secondary lymphoid organs, where they present the captured antigens to resting T cells and, dependent on the received stimuli, induce tolerance or initiate adaptive immune responses. It should be noted that in addition to DC associated C-type lectins also members of other lectin classes, such as galectins (galactose-binding lectins) or siglecs (sialic acid binding lectins) present on DC or on other antigen-presenting cells, have been implicated in host-pathogen interactions (van den Berg et al., 2004; Sato and Nieminen, 2004; Jones et al., 2003). For example, we recently showed that macrophage-derived galectin-3 is highly expressed in liver granuloma’s of schistosome infected hamsters and binds to LDN glycan antigens on schistosome egg antigens. Interestingly, in vitro studies demonstrated that galectin-3 mediates phagocytosis of LDN containing neoglycoconjugates by activated macrophages indicating a role for galectin-3 in innate immunity to schistosomes (van den Berg et al., 2004). DCs are central in directing Th1-Th2 responses and molecular patterns on the pathogen that are recognized by DC that capture pathogen determinants, are crucial for biasing the Th immune response (Jankovic et al., 2002). In mouse models, DC pulsed with SEA potently stimulate Th2 responses both in vivo and in vitro while failing to undergo a conventional maturation process (MacDonald et al., 2001). In addition, in immunization experiments using a Lex neoglycoprotein, a strong Th2 response was mounted that was dependent on the presense of Lex on the protein. Surprisingly, the antibodies that were generated appeared protein-specific, indicating that Lex merely had a function as an efficient Th2-stimulating adjuvans (Okano et al., 2001). We therefore hypothesized that Lewis-x might be recognized by a specific lectin on DC, and that this interaction would be (one of ) the signal(s) that could trigger DCs to acquire a Th2 inducing phenotype.

14

I. van Die et al.

6. IDENTIFICATION OF LECTINS ON DENDRITIC CELLS THAT BIND SCHISTOSOME EGG ANTIGENS Because DC are central in directing Th1-Th2 responses, we searched for a cellsurface receptor expressed on human immature DC that interacts with S. mansoni egg glycoproteins (SEA). SEA is a mixture of glycoproteins, containing many immunogenic glycan antigens that potentially could interact with lectins on the iDC. To detect binding of SEA to human immature DC, a fluorescent-bead adhesion assay was developed. Fluorescent beads were precoated with streptavidin and then used to capture biotinylated SEA. The conjugated beads were allowed to interact with DC, similar to the approach described by (Geijtenbeek et al., 1999). We showed that the immature DCs strongly bound SEA, and that this binding could be blocked completely by EDTA, suggesting that one or more C-type lectin(s) on DC mediate the binding to SEA. Identification of the C-type lectin(s) involved showed that a substantial part of the binding was mediated via the C-type lectin DC-SIGN (van Die et al., 2003; Van Liempt et al., 2004). In addition, the C-type lectin MGL (macrophage galactose lectin) was shown to be responsible for part of the binding of iDC to SEA (SJ van Vliet, et al, manuscript submitted). These data show that different ff C-type lectins on iDCs participate in the binding to Schistosome egg antigens, which implies that each interaction may contribute to the induction of the final DC function.

7. BINDING OF DC-SIGN TO SCHISTOSOME EGG ANTIGENS IS MEDIATED THROUGH INTERACTION WITH THE GLYCANS LEXIS-X AND LDNF As DC-SIGN has been reported to display affinity to both mannose and fucose we explored whether the binding of DC-SIGN to SEA is fucose mediated. Using a competitive ELISA it was demonstrated that this binding indeed can be blocked by monoclonal antibodies (mAbs) specific for the fucosylated glycans Lex and LDNF, respectively. A combination of anti-Lex and anti-LDNF mAbs strongly blocked the binding of DC-SIGN to SEA. These binding properties have been established by direct binding studies of DC-SIGN to Lex and LDNF containing oligosaccharides/ neoglycoconjugates. Lex and LDNF glycans are both major glycan antigens within SEA and resemble each other by containing a terminal fucose (1–3 linked to a GlcNAc residue. Our data have demonstrated that other fucosylated glycan antigens within SEA, such as core-fucose and LDN-DF, do not constitute ligands for DC-SIGN (van Die et al., 2003) (E. van Liempt ea, unpublished). This indicates that binding of DC-SIGN to Lex and LDNF glycan antigens is specific and does not only depends on the presence of a fucose. Based on our binding studies and the crystal structure of DC-SIGN (Feinberg et al., 2001) we recently proposed a molecular model of the binding of DC-SIGN to Lex (van Liempt et al., 2004) that was

Interaction of Schistosome Glycans

15

Figure 2. Binding of Lewis-x to the CRD of DC-SIGN. The three monosaccharides of Lex all participate to the binding (van Liempt et al., 2004; Guo et al., 2004). The grey sphere indicates a calcium ion.

essentially similar to a model based on the crystal structure of DC-SIGN in complex with Lex (Guo et al., 2004). Both studies showed that the fucose of Lex strongly interacts with Val351 in DC-SIGN, and that both the galactose and GlcNAc show additional contacts with the CRD of DC-SIGN and contribute to the binding. From the molecular model (Fig. 2) it can be deduced that an N-acetylgroup linkage to the C2 of galactose within the Lex trisaccharide, as is found in LDNF, will not interact with the CRD of DC-SIGN, which is in agreement with the observed binding of DC-SIGN to LDNF. By contrast, there is no place in the binding pocket for an additional fucose a2-linked to the Fuca1, 3GlcNAc moiety of LDNF, such as found in LDN-DF, which explains the lack of binding of DC-SIGN to LDN-DF. Our data indicate that DC-SIGN shows a strongly increased binding to multivalently presented Lex and LDNF (Van Liempt et al, unpublished). The optimal binding most likely is dependent on the spacing and presentation of the antigens within the glycan, and further studies are underway to identify the actual schistosome ligands that bind DC-SIGN.

8. CONCLUDING REMARKS Our knowledge of the remarkable role that Schistosome glycoconjugates play in the immunobiology of schistosomiasis is rapidly growing. We now are getting

16

I. van Die et al.

insight in the receptors on dendritic cells and macrophages that interact with Schistosome glycan antigens, although the functional consequences of these interactions mostly remain to be understood. Our data indicate that Lex interacts with DC-SIGN on DC, which in combination with the data from Okano et al., (2001), suggest that this interaction may play a role in triggering DCs to direct the mounting of a Th2 response. This hypothesis is further strengthened by recent data showing that in Helicobacter pylori infection, Lex positive variants block Th1 cell development through interaction with DC-SIGN whereas Lex negative variants induce a strong Th1 cell response (Bergman et al., 2004). However, the functional role of DC-SIGNLex interaction needs to be further investigated in schistosome infections and this interaction clearly cannot explain the general potential of helminth to induce Th2 responses. Although DC-SIGN has a broad carbohydrate recognition potential (Appelmelk et al., 2003), several helminths that we tested are not recognized by DC-SIGN (unpublished results). In addition, also other glycan antigens, such as core-fucose and core-xylose that are not recognized by DC-SIGN (Van Liempt et al., manuscript in preparation), can strongly induce Th2 responses (Faveeuw et al., 2003). It is expected that other factors, such as the presence or absence of maturation signals, the glycan-carriers and cross-talk between lectins and Toll-like receptors within DCs, will contribute to define DC-function. The availability of other Th2-inducing helminth-type glycoconjugates will be important to identify their receptors, and investigate the immunomodulatory properties of these glycoconjugates. Thus the molecular understanding how parasite glycans trigger DCs to induce Th2 responses remains an issue of high priority for the next future.

ACKNOWLEDGEMENTS We thank Anne Imberty for composing the molecular model of DC-SIGN with Lewis-x (Fig. 3)

REFERENCES Bakker, H., Agterberg, M., Van Tetering, A., Koeleman, C.A.M., Van den Eijnden, D.H., and Van Die, I., 1994, A Lymnaea stagnalis gene, with sequence similarity to that of mammalian b14-galactosyltransferases, encodes a novel UDP- GlcNAc:GlcNAc b-R b1-4-N-acetylglucosaminyltransferase. J. Biol. Chem. 269: 30326–30333. Bakker, H., Schijlen, E., De Vries, T., Schiphorst, W.E., Jordi, W., Lommen, A., Bosch, D., and Van, D., I., 2001, Plant members of the alpha1-3/4-fucosyltransferase gene family encode an alpha14-fucosyltransferase, potentially involved in Lewis(a) biosynthesis, and two core alpha13-fucosyltransferases. FEBS L ett. 507: 307–312. Bergman, M.P., Engering, A., Smits, H.H., Van Vliet, S.J., Van Bodegraven, A.A., Wirth, H-P., Kapsenberg, M.L. Vandenbroucke-Grauls, M.J.E., Van Kooyk, Y. and Appelmelk, B.J., 2004, Helicobacter pylori modulates the T helper cell 1/T helper cell 2 balance through phase-variable interaction between lipopolysaccharide and DC-SIGN. J. Exp. Med. 200: 979–990. Cummings, R.D. and Nyame, A.K., 1996, Glycobiology of schistosomiasis. FASEB Journal 10: 838–848. Cummings, R.D. and Nyame, A.K., 1999, Schistosome glycoconjugates. Biochim. Biophys. Acta 1455: 363–374.

Interaction of Schistosome Glycans

17

Eberl, M., Langermans, J.A.M., Vervenne, R.A., Nyame, K.A., Cummings, R.D., Thomas, A.W., Coulson, P.S., and Wilson, R.A., 2002, Antibodies to glycans dominate the host response to schistosome larvae and eggs: is their role protective or subversive? J. Infect. f Dis. 183: 1238–1247. Engering, A., Geijtenbeek, T. B. H., and van Kooyk, Y., 2002, Immune escape through C-type lectins on dendritic cells. T rends in Immunol. 23: 480–485. Fabini, G., Freilinger, A., Altmann, F., and Wilson, I.B.H., 2001, Identification of core alpha 1, 3-fucosylated glycans and cloning of the requisite fucosyltransferase cDNA from Drosophila melanogaster. Potential basis of the neural anti-horseradish peroxidase epitope. J. Biol. Chem. 276: 28058–28067. Faveeuw, C., Mallevaey, T., Paschinger, K., et al., 2003, Schistosome N-glycans containing core alpha 3-fucose and core beta 2-xylose epitopes are strong inducers of Th2 responses in mice. Eur J Immunol 33: 1271–1281 Feinberg, H., Mitchell, D.A., Drickamer, K., and Weis, W.I., 2001, Structural basis for selective recognition of oligosaccharides by DC-SIGN and DC-SIGNR. Science 294: 2163–2166. Figdor, C.G., van Kooyk, Y., and Adema, G.J., 2002, C-type lectin receptors on dendritic cells and langerhans cells. Nature Rev. Immunol. 2: 77–84. Gantner, B.N., Simmons, R.M., Canavera, S.J., Akira, S., and Underhill, D.M., 2003, Collaborative induction of inflammatory responses by dectin-1 and Toll-like receptor 2. J. Exp. Med. 197: 1107–1117. Geijtenbeek T.B., van Vliet, Sandra J., Koppel E.A., Sanchez-Hernandez M., Vandenbroucke-Grauls C.M., Appelmelk, B., and van Kooyk, Y., 2002, Mycobacteria target DC-SIGN to suppress dendritic cell function. J Exp Med. 197: 7–17. Geijtenbeek, T.B.H., van Kooyk, Y., van Vliet, S.J., Renes, M.H., Raymakers, R.A.P., and Figdor, C.G., 1999, High frequency of adhesion defects in B-Lineage acute lymphoblastic leukemia. Blood 94: 754–764. Guo, Y., Feinberg, H., Conroy, E., Mitchell, D.A., Alvarez, R., Blixt, O., Taylor, M.E., Weis, W.I., and Drickamer, K., 2004. Structural basis for distinct ligand-binding and targeting properties of the receptors DC-SIGN and DC-SIGNR. Nat. Struct. Mol. Biol. 11: 591–598. ’t Hart, B.A. and Van Kooyk, Y., 2004, Yin-Yang regulation of autoimmunity by DCs. T rends in Immunol. 25: 353–359. Hokke, C. H. and Deelder, A. M., 2001, Schistosome glycoconjugates in host-parasite interplay. Glycoconj. J. 18: 573–587. Hokke, C.H., Neeleman, A.P., Koeleman, C.A.M., and Van den Eijnden, D.H., 1998, Identification of an a3-fucosyltransferase and a novel a2-fucosyltransferase activity in cercariae of the schistosome T richobilharzia ocellata. Glycobiology in press. Jacobs, W., Deelder, A.M., and Van Marck, E., 1999, Schistosomal granuloma modulation. II. Specific immunogenic carbohydrates can modulate schistosome-egg-antigen-induced hepatic granuloma formation. Parasitol. Res. 85: 14–18. Janeway, C.A., Jr. and Medzhitov, R., 2002, Innate immune recognition. Ann. Rev. Immunol. 20: 197–216. Jankovic, D., Liu, Z., and Gause, W.C., 2002, Th1-and Th2-cell commitment during infectious disease: asymmetry in divergent pathways. T rends in Immunol. 22: 450–457. Jones, C., Virji, M., and Crocker, P.R., 2003, Recognition of sialylated meningococcal lipopolysaccharide by siglecs expressed on myeloid cells leads to enhanced bacterial uptake. Mol. Microbiol. 49: 1213–1225. Kawar, Z.S., van Die, I., and Cummings, R.D., 2002, Molecular cloning and enzymatic characterization of a UDP-GalNAc:GlcNAcbeta -R b1, 4-N-acetylgalactosaminyltransferase from Caenorhabditis elegans. J. Biol. Chem. 277: 34924–34932. MacDonald, A.S., Straw, A.D., Bauman, B., and Pearce, E.J., 2001, CD8– dendritic cell activation status plays an integral role in influencing Th2 response development. J. Immunol. 167: 1982–1988. Mulder, H., Spronk, B.A., Schachter, H., Neeleman, A.P., Van den Eijnden, D.H., De Jong-Brink, M., Kamerling, J.P., and Vliegenthart, J.F.G., 1995, Identification of a novel UDP-GalNAc:GlcNAcb-R b1-4 N-acetylgalactosaminyltransferase from the albumen gland and connective tissue of the snail L ymnaea stagnalis. Eur. J. Biochem. 227: 175–185. Naus, C.W.A., van Remoortere, A., Ouma, J.H., Kimani, G., Dunne, D.W., Kamerling, J.P., Deelder, A.M., and Hokke, C.H., 2003, Specific antibody responses to three Schistosome-related carbohydrate structures in recently exposed immigrants and established residents in an area of Schistosoma mansoni endemicity. Infect. f Immun. 71: 5676–5681.

18

I. van Die et al.

Nyame, A.K., Debose-Boyd, R., Long, T.D., Tsang, V.C.W., and Cummings, R.D., 1998, Expression of Lex antigen in Schistosoma japonicum and S. haematobium and immune responses to Lex in infected animals: lack of Lex expression in other trematodes and nematodes. Glycobiology 8: 615–624. Nyame, A. K., Kawar, Ziad S., and Cummings, R.D., 2004, Antigenic glycans in parasitic infections: implications for vaccines and diagnosis. Arch. Biochem. Biophys. 426: 182–200. Nyame, A.K., Leppanen, A., Bogitsh, B.J., and Cummings, R.D., 2000, Antibody responses to the fucosylated LacdiNAc glycan antigen in Schistosoma mansoni-infected mice and expression of the glycan among schistosomes. Exp. Parasitol. 96: 202–212. Nyame, A.K., Pilcher, J.B., Tsang, V.C.W., and Cummings, R.D., 1996, Schistosoma mansoni infection in humans and primates induces cytolytic antibodies to surface Lex determinants on myeloid cells. Exp. Parasitology 82: 191–200. Nyame, A.K., Pilcher, J.B., Tsang, V.C.W., and Cummings, R.D., 1997, Rodents infected with Schistosoma mansoni produce cytolytic IgG and IgM antibodies to the Lewis x antigen. Glycobiology 7: 207–215. Okano, M., Satoskar, A.R., Nishizaki, K., and Harn, D.A., Jr., 2001, Lacto-N-fucopentaose III found on Schistosoma mansoni egg antigens functions as adjuvant for proteins by inducing Th2-Type response. J. Immunol. 167: 442–450. Palucka, K. and Banchereau, J., 2002, How dendritic cells and microbes interact to elicit or subvert protective immune responses. Curr. Opin. Immunol. 14: 420–431. Sato, Sachiko and Nieminen, J., 2004, Seeing strangers or announcing ‘‘danger’’: Galectin-3 in two models of innate immunity. Glycoconjugate J. 19: 583–591. Srivatsan, J., Smith, D.F., and Cummings, R.D., 1992a, Schistosoma mansoni synthesizes novel biantennary Asn-linked oligosaccharides containing terminal b-linked N-acetylgalactosamine. Glycobiology 2: 445–452. Srivatsan, J., Smith, D.F., and Cummings, R.D., 1992b, The human blood fluke Schistosoma mansoni synthesizes glycoproteins containing the Lewis X antigen. J. Biol. Chem. 267: 20196–20203. Strasser, R., Mucha, J., Mach, L., Altmann, F., Wilson, I.B.H., Glxssl, J., and Steinkellner, H., 2001, Molecular cloning and functional expression of b1,2-xylosyltransferase cDNA from Arabidopsis thaliana. FEBS L etters 472: 105–108. Van Dam, G.J., Bergwerff, ff A.A., Thomas-Oates, J.E., Rotmans, J.P., Kamerling, J.P., Vliegenthart, J.F.G., and Deelder, A.M., 1994, The immunologically reactive O-linked polysaccharide chains derived from circulating cathodic antigen isolated from the human blood fluke Schistosoma mansoni have Lewis x as repeating unit. Eur. J. Biochem. 225: 467–482. Van den Berg, T. K., Honing, H., Franke, N., van Remoortere, A., Schiphorst, W. E. C. M., Liu, F-T, Deelder, A. M., Cummings, R. D., Hokke, C. H., and Van Die, I., 2004, LacdiNAc-glycans constitute a parasite pattern for galectin-3-mediated immune recognition. J. Immunol. 173: 1902–1907. Van Die, I, Cummings, R. D., van Tetering, A., Hokke, C. H., Koeleman, C. A. M., and van den Eijnden, D. H., 2000, Identification of a novel UDP-Glc:GlcNAcb1, 4-glucosyltransferase in Lymnaea stagnalis that may be involved in the synthesis of complex-type oligosaccharide chains. Glycobiology 10: 263–271. Van Die, I., Van Vliet, S.J., Nyame, A.K., Cummings, R.D., Bank, C.M.C., Appelmelk, B., Geijtenbeek, T.B.H., and van Kooyk, Y., 2003, The dendritic cell-specific C-type lectin DC-SIGN is a receptor for Schistosoma mansoni egg antigens and recognizes the glycan antigen Lewis x. Glycobiology 13: 471–478. Van Liempt, E., Imberty, A., Bank, C.M., Van Vliet, S.J., Van Kooyk, Y., Geijtenbeek, T.B., and Van Die, I. ff in binding properties of the highly related C-type lectins (2004). Molecular basis of the differences DC-SIGN and L-SIGN to Lewis X trisaccharide and Schistosoma mansoni egg antigens. J. Biol. Chem. 279: 33161–33167. Van Remoortere, A., Hokke, C., Van Dam, G.J., Van Die, I., Deelder, A., and Van den Eijnden, D.H., 2000, Various stages of Schistosoma express Lewisx, LacdiNAc, GalNAcb1-4 (Fuca1-3)GlcNAc and GalNAcb1-4(Fuca1-2Fuca1-3)GlcNAc carbohydrate epitopes: detection with monoclonal antibodies that are characterized by enzymatically synthesized neoglycoproteins. Glycobiology 10: 601–609. Van Remoortere, A., Vermeer, H.J., Van Roon, A.M., Langermans, J.A., Thomas, A.W., Wilson, R.A., Van Die, I., Van den Eijnden, D.H., Agoston, K., Kerekgyarto, J., Vliegenthart, J.F., Kamerling, J.P., Van Dam, G.J., Hokke, C.H., and Deelder, A.M., 2003a, Dominant antibody responses to Fuca1-3GalNAc and Fuca1-2Fuca1-3GlcNAc containing carbohydrate epitopes in Pan troglodytes vaccinated and infected with Schistosoma mansoni. Exp. Parasitol. 105, 219–225.

Interaction of Schistosome Glycans

19

Van Remoortere, A., Bank, C.M.C., Nyame, A.K., Cummings, R.D., Deelder, A.M., and Van Die, I., 2003b, Schistosoma mansoni-infected mice produce antibodies that cross-react with plant, insect, and mammalian glycoproteins and recognize the truncated biantennaryN-glycan Man GlcNAc -R. 3 2 Glycobiology 13: 217–225. Van Remoortere, A., Hokke, C.H., Van Dam, G.J., Van Die, I., Deelder, A.M., and Van den Eijnden, D.H., 2000, Various stages of Schistosoma express Lewisx, LacdiNAc, GalNAcb1-4(Fuca1-3)GlcNAc, and GalNAcb1-4(Fuca1-2Fuca1-3)GlcNAc carbohydrate epitopes: detection with monoclonal antibodies that are characterized by enzymatically synthesized neoglycoproteins. Glycobiology 10: 601–609. Van Remoortere, A., Van Dam, G.J., Hokke, C.H., Van den Eijnden, D.H., Van Die, I., and Deelder, A.M., 2001, Profiles of immunoglobulin M (IgM) and IgG antibodies against defined carbohydrate epitopes in sera of Schistosoma-infected individuals determined by Surface Plasmon Resonance. Infect. f Immun. 69: 2396–2401. Van Roon, A.M., Van de Vijver, K.K., Jacobs, W., Van Marck, E.A., van Dam, G.J., Hokke, C.H., and Deelder, A.M., 2004, Discrimination between the anti-monomeric and the anti-multimeric Lewis X response in murine schistosomiasis. Microbes and Infection 6: 1125–1132. Van Dam, G.J., Claas, F.H.J., Yazdanbakhsh, M., Kruize, Y.C.M., Van Keulen, A.C.I., Ferreira, S.T.M.F., Rotmans, J.P., and Deelder, A.M., 1996, Schistosoma mansoni excretory circulating cathodic antigen shares Lewis-x epitopes with a human granulocyte surface antigen and evokes host antibodies mediating complement-dependent lysis of granulocytes. Blood 88: 4246–4251. Velupillai, P., Dos Reis, E.A., Dos Reis, M.G., and Harn, D.A., 2000, Lewisx-containing oligosaccharide attenuates schistosome egg antigen-induced immune depression in human schistosomiasis. Human Immunology 61: 225–232. Velupillai, P. and Harn, D.A., 1994, Oligosaccharide-specific induction of interleukin 10 production by B220+ cells from schistosome-infected mice: a mechanism for regulation of CD4+ T-cell subsets. Proc. Natl. Acad. Sci. USA 91: 18–22. Weis, W. I., Taylor, Maureen E., and Drickamer, K., 1998, The C-type lectin superfamily in the immune system. Immunol. Rev. 163: 19–34. Wuhrer, M., Dennis, R.D., Doenhoff, ff M.J., Lochnit, G., and Geyer, R., 2000, Schistosoma mansoni cercarial glycolipids are dominated by Lewis X and pseudo-Lewis Y structures. Glycobiology 10: 89–101. Zheng, Q., Van Die, I., and Cummings, R.D., 2002, Molecular cloning and characterization of a novel a1,2-fucosyltransferase (CE2FT-1) from Caenorhabditis elegans. J. Biol. Chem. 277: 39823–39832.

6

THE MANNAN-BINDING LECTIN (MBL) PATHWAY OF COMPLEMENT ACTIVATION: BIOCHEMISTRY, BIOLOGY AND CLINICAL IMPLICATIONS

Jens Christian Jensenius Department of Medical Microbiology and Immunology University of Aarhus, Denmark

MBL is an oligomeric protein designed to recognize pathogen-associated molecular patterns (PAMPs). It belongs to the family of collagen-like defence molecules characterized by being comprised of several subunits each composed of three polypeptides. The polypeptide of about 30 kD presents a collagen region attached to a globular head containing the recognition structure. Each subunit thus presents three binding sites. Hence the oligomer, typically comprising four subunits presents a substantial number of binding sites, each of low affinity, with the avidity and selectivity of the molecules being determined through multiple interactions. The collectins have sugarbinding, C-type-lectin globular domains, the ficolins have fibrinogen-like domains of

A

B

A. The overall structure of the human collectins, SP-A and MBL, and L-ficolin. The schematic structures represent interpretations of the electron micrographs. Two MBL oligomers are shown. The structures are referred to as sertiform (sertula=small umbel) B. Overview of the complement system with focus on the MBL pathway. Different ff MBL-MASP complexes are involved in the formation of the C3 convertase, C4bC2b, and in direct activation of C3. Complexes of ficolin and MASP also activate complement. 21 John S. Axford (ed.), Glycobiology and Medicine, 21-22. © 2005 Springer. Printed in the Netherlands.

22

J. C. Jensenius

less defined specificity, complement C1q has domains recognizing structures on immunoglobulins. The biological importance of MBL was indicated when opsonin-deficient children with recurrent infections were found to be genetically deficient in MBL. Further interest in this molecule was sparked by the observation of complement activation upon binding to carbohydrates. This activation is mediated by MBL-associated serine proteases, MASPs, which have now been found associated also with ficolins. The MASPs present domain structures identical to those of C1r and C1s of the classical complement pathway. MBL has affinity for terminal, non-reducing sugars presenting horizontal 3- and 4-OH groups, e.g., glucose, mannose and fucose. Ficolins are also cited as being lectins, but we have recently shown L-ficolin to be selective for acetyl groups on both sugars and on other molecules. The ligands for H- and M-ficolin remain undefined. Thus the widely adopted term ‘‘lectin complement pathway’’ appears inappropriate for MASP-mediated complement activation.

REFERENCE Holmskov U, Thiel S, Jensenius JC. Collections and ficolins: humoral lectins of the innate immune defence. Annu Rev Immunol. 2003;21:547–78.

7

KILLER CELL LECTIN-LIKE RECEPTORS AND THE NATURAL KILLER CELL GENE COMPLEX

Ø. Nylenna1, L. M. Flornes1, I. H. Westgaard W 1, P. Y. Woon2, C. Naper1, J. T. Vaage, V 1, D. Gauguier2, J. C. Ryan3, E. Dissen1, and S. Fossum1 1Institute of Basic Medical Sciences University of Oslo, Norway 2Wellcome Trust Centre for Human Genetics University of Oxford, U.K. 3Department of Arthritis and Immunology University of California, USA

The natural killer cell gene complex (NKC), which maps to the distal parts of mouse chromosome 6 and rat chromosome 4, and in the human to the short arm of chromosome 12, encodes type 2 membrane receptors belonging to the group V C-type lectin superfamily (CLSF), lacking the evolutionary conserved calcium/ saccharide binding amino acid residues found in other CLSF receptors. It contains all group V CLSF genes currently known, except Klrg1 which in rodents lies 6–7 Mb proximal to the NKC (see Fig. 1), and nothing but such genes, except Gabarapl1. Due to expansion of the Nkrp1 (Klrb) and in particular the L y49 (Klra) multigene families the complex is particularly large in rodents, where it can be divided into three parts: a proximal part encoding Nkrp1 and Clr receptors, a middle part encoding a variety of group V CLSF receptors, and a large distal part encoding Ly49 receptors. In the rat the NKC spans 3.3 Mb and is predicted to contain 67 CLSF loci (including some pseudogenes), distributed as indicated in the figure. To the extent ligands are known, the NKC encoded receptors do not bind saccharides (with the exception of Dectin-11), but rather MHC class I and related ligands (rev. in2). Recently, mouse Nkrp1d and -f were shown to bind Clr molecules, providing the first example of CLSF receptor/ligand pairs3. Functionally the NKC encoded receptors have opposing regulatory roles on leukocyte activation, the activating mediating their effects ff via protein tyrosine kinases and the inhibitory via protein tyrosine phosphatases. Close to, but distinct from the NKC lies a smaller gene complex (called APLEC4) encoding opposing regulatory leukocyte receptors, but 23 John S. Axford (ed.), Glycobiology and Medicine, 23-24. © 2005 Springer. Printed in the Netherlands.

Ø. Nylenna et al.

24

Figure 1. The chromosal regions in man, mouse and rat containing the NKC (lower box) and the adjacent APLEC plus the Klrg1 gene. In the rat the NKC is predicted to contain 67 CLSF loci (based on the rat genome sequence from the BN strain. Numbers to the left indicate distances in Mb, numbers to the right (bold font) indcate number of loci.

preferentially expressed by professional antigen presenting cells and neutrophils and with the calcium/saccharide binding amino acid residues conserved (hence classified as group II CLSF receptors). The presentation will concentrate on the rat, where we now have cloned close to all of the genes, and in particular on the highly dynamic Ly49 gene region, to which the functional alloreactivity gene Nka previously was mapped5,6.

REFERENCES 1. 2. 3. 4. 5. 6.

Brown, G.D. et al. J Exp Med 196, 407–412 (2002). Yokoyama,W.M. & Plougastel,B.F.M. Nature Reviews Immunology 3, 304–316 (2003). Iizuka,K. et al. Nature Immunol 4, 801–807 (2003). Flornes, L.M. et al. Immunogenetics, in press (2004) Dissen,E. et al. J. Exp. Med. 183, 2197–2207 (1996). Nylenna, Ø. et al. Submitted.

8

GLYCOSYLATION INFLUENCES THE LIGAND BINDING ACTIVITIES OF MANNOSE RECEPTOR

Yunpeng Roc Su1,2, Clarence Tsang1, Talitha Bakker2, James Harris2, Siamon Gordon2, Raymond A. Dwek1, Luisa Martinez-Pomares2 and Pauline M Rudd1 1Glycobiology Institute 2Sir William’s Dunn School of Pathology Oxford University, OX1, 3QU, Oxford, UK

Murine mannose receptor (MR) contains seven N-linked and three O-linked oligosaccharides and differential ff binding properties have been described for MR isolated from the liver and the lung. We hypothesised that these different ff binding activities could be controlled by glycosylation. In this study the relationship between MR glycosylation and its function has been investigated using MR transductants generated in both wild type CHO cells and glycosylation-deficient LEC cells. The investigation shows that glycosylation does not affect ff the subcellular distribution, proteolytic processing and endocytic capacity of the receptor, but has a major effect ff in its binding capacity. Cells bearing MR modified with Man GlcNAc sugars 5 2 (Man-5 MR) completely lost its mannose-internalisation activity, which is associated with CRD4–5 of MR. In agreement with this observation purified soluble Man-5 MR lost the capability to bind mannan in vitro. The desialylation modification of MR also results in a 70% reduction of cellular internalisation activity and a low efficient mannan binding activity in vitro. However, cells bearing MR modified with Man GlcNAc sugars or desialylated glycans do retain their sulphated sugar 2 5 internalisation activity, which is associated to the cysteine-rich (CR) domain. Interestingly, in vitro SO -3-galactore-PAA binding study indicated desialylated MR 4 has better affinity than wild-type MR. Subsequent gel filtration and BIAcore studies showed that desialylated MR tend to form self-associated structure and multiple presentation of CR domain could enhance its affinity to sulphated sugars dramatically. These results, for the first time, suggest a role for glycosylation, especially terminal sialylation of MR, in manipulating its dual ligand binding activities in vivo.

25 John S. Axford (ed.), Glycobiology and Medicine, 25-26. © 2005 Springer. Printed in the Netherlands.

26

Yunpeng Roc Su et al.

REFERENCES Fiete, D. J. et al., 1998, A cysteine-rich domain of the ‘‘mannose’’ receptor mediates GalNAc-4-SO4 binding. Proc Natl Acad Sci U S A. 2089–2093. Feinberg, H. et al. (2000). Structure of a C-type carbohydrate recognition domain from the macrophage mannose receptor. J Biol Chem. 275, 21539–21548. Rudd, P. M. et al. (1997). Oligosaccharide sequencing technology. Nature. 205–207 Stanley, P. (1989). Chinese hamster ovary cell mutants with multiple glycosylation defects for production of glycoproteins with minimal carbohydrate heterogeneity. Mol Cell Biol. 377–383.

9

HUMAN IMMUNOGLOBULIN GLYCOSYLATION AND THE LECTIN PATHWAY OF COMPLEMENT ACTIVATION

James N. Arnold1, Louise Royle2, Raymond A. Dwek2, Pauline M. Rudd2, and Robert B. Sim1 1MRC Immunochemistry Unit 2Oxford Glycobiology Institute Department of Biochemistry University of Oxford South Parks Road, Oxford OX1 3QU, UK

1. INTRODUCTION Immunoglobulins are the major secretory products of the adaptive immune system. They are glycoproteins which are found in all higher vertebrates (mammals, birds, reptiles, amphibians, bony and cartilaginous fish, but not in jawless fish (agnatha)) (Litman et al., 1999). In humans there are five classes IgG, IgM, IgA, IgE and IgD. The immunoglobulins share similar structures (Fig. 1). Each immunoglobulin molecule is composed of two identical disulphide bridged class-specific heavy chains, each disulphide bridged to a light chain of which there are two isoforms named k and l. Both heavy and light chains are composed of regions called immunoglobulin domains. The immunoglobulin fold/domain is about 105–120 amino acids long and is composed of b-sheet secondary structure (Amzel and Poljak, 1979). The role of immunoglobulins is to bind to antigens via their N-terminal (variable amino acid sequence) domains and to mediate effector ff functions, such as activation of complement (Malhotra et al., 1995; Roos et al., 2001) or binding to receptors via their constant (invariable sequence) domains (Mimura et al., 2000; Shields et al., 2001). During immunoglobulin synthesis, rearrangement of gene segments and somatic mutation creates variation in amino acid sequence in the N-terminal domains (named VH and VL domains for Variable Heavy and Light chains respectively). The light chains have one V domain and one constant sequence domain (CL). The sequence of all l chain C domains is the same, and the sequence is homologous to 27 John S. Axford (ed.), Glycobiology and Medicine, 27-43. © 2005 Springer. Printed in the Netherlands.

28

J. N. Arnold et al.

Figure 1. Immunoglobulin Structure a) The structure of IgG showing the Variable Heavy (VH) and Constant Heavy (CH), Variable Light (VL) and Constant Light (CL) domains. The diagram identifies the Fab, Fc and flexible hinge regions of the molecule. This hinge varies in length between the different ff immunoglobulin classes and is replaced by additional CH domain in IgE and IgM. The approximate positioning of the Asn-297 N-linkage site for glycans is marked. b) Diagrammatic representation of IgG1, IgD, IgA1, IgE and IgM showing N- and Olinked glycan positions, and inter-chain disulphide bridges. The domains themselves contain intra-domain disulphide bridges, although these are not marked. IgM circulates in the serum in both pentameric and hexameric forms, in which the monomeric units are disulphide bridged together. Pentameric IgM contains a single J chain but hexameric IgM does not (Weirsma et al., 1998).

Human Immunoglobulin Glycosylation

29

the C domain shared by all k chains. Heavy chains have 3 or 4 C domains. The sequences of the C domains are class or subclass specific, i.e. all IgGs have identical constant regions, as do all IgMs. Each clone of B lymphocytes secretes only one immunoglobulin molecule, which has V regions unique to that particular B cell clone. Total IgG isolated from human serum therefore contains 4 subclasses, each with similar but distinct constant regions, and with 105–106 different ff V region sequences. IgM and IgD occur both as soluble forms (in serum) and membrane-bound forms on B lymphocytes (Van Boxel et al., 1972). The membrane-bound forms have an additional trans-membrane segment, C-terminal to the constant regions. IgA, IgG, IgE are all soluble molecules: IgG is the most abundant in serum (10–15 mg/ml), while IgA is the most abundant immunoglobulin overall. Most IgA is secreted through epithelia into the mucous lining of the gastrointestinal and respiratory tract, and into tears, saliva and milk (Norderhaug et al., 1999). The secreted form is generally dimeric and contains an extra glycosylated polypeptide chain, SC (Secretory Component) and glycosylated 16KDa J chain (Johansen et al., 2001; Royle et al., 2003), which is also found in pentameric forms of IgM (Wiersma et al., 1998). The single J chain is disulphide bridged to two C-termini of both IgM and IgA molecules (Wiersma et al., 1998; Royle et al., 2003). IgA in serum is predominantly monomeric but also forms dimers and higher polymers (Delacroix et al., 1982; Roos et al., 2001). IgE is the lowest abundance immunoglobulin, occurring as a monomer at <1 mg/ml. IgD also occurs as a low abundance monomer in serum at <30 mg/ml, while IgM is at high concentrations (~2.5 mg/ml). IgM occurs predominantly as pentamers and hexamers, although a small amount of monomer also circulates (Sørensen et al., 1999). The different ff classes of immunoglobulin are distinct in their major effector ff functions. IgM is principally associated with complement classical pathway activation via binding of C1q (Wiersma et al., 1998). IgG also activates complement via classical (Duncan and Winter, 1988) and alternative pathways (Anton et al., 1989) and mediates ADCC (Antibody Dependent Cell Cytotoxicity) (Sarmay et al., 1992). IgE is associated with mast cell and basophil stimulation in allergic conditions. IgA in secretions may act mainly to agglutinate (immobilise) or neutralise micro-organisms (Lamm, 1997). No effector ff functions have been identified for IgD. In addition to their enormous diversity of amino acid sequences and antigenbinding specificity, immunoglobulins display considerable diversity in the location and number of glycosylation sites (both N- and O-linked) and great diversity in glycan structure. The glycans attached to the immunoglobulins are important for immunoglobulin solubility (Tarentino et al., 1974), subcellular transport and secretion (Gala and Morrison, 2002), conformation (Mimura et al., 2000), binding to Fc receptors (Mimura et al., 2000), normal plasma clearance (Skockert, 1995) and complement activation (Malhotra et al., 1995). This chapter discusses both the glycan structures that are attached to the normal human serum immunoglobulins and their potential roles in complement activation through the binding of the serum ‘recognition’ lectin, Mannan Binding Lectin (MBL), and the subsequent activation of the lectin pathway of the complement system.

30

J. N. Arnold et al.

2. GLYCOSYLATION OF THE IMMUNOGLOBULINS 2.1. IgG There are four subclasses of IgG, named IgG1–4, that differ ff in their heavy chain constant region sequence and disulphide bridging. The subclasses have distinctive glycan pools (Jefferis ff et al., 1990). All IgGs have a single N-linked glycosylation site on each heavy chain in the CH2 domain at Asn-297 (Fig. 1). There are no conserved glycosylation sites in the light chain or variable regions of the heavy chain. The glycan population attached at Asn-297 contains three sets of glycoforms termed IgG-G0, -G1 and -G2 (Fig. 2b). The IgG-G2 biantennary glycans occupying Asn-297 have two arms that both terminate in galactose residues. This set of glycoforms accounts for approximately 16% of total IgG glycans. Approximately 35% are IgG-G1, which lack a terminal galactose residue on one biantennary arm, exposing a GlcNAc residue. IgG-G0 glycans make up 35% and neither biantennary arm contains a galactose residue. The final 14% of serum IgG glycans consist of IgG-G2 or -G1 glycoforms which are sialylated. Within the glycans of IgG there is a diversity of structures caused by the presence of bisecting GlcNAc residues (B in Fig. 2a) (approximately 30% of total IgG1 glycan pool), core fucose (Fc in Fig. 2a) (approximately 70% of the total IgG1 glycan pool) and sialylation of the terminal 1,3 arm galactose residues (S in Fig. 2a) (14% of total IgG1 glycan pool) (Butler et al., 2003)). IgG1 is the most abundant subclass of IgG in the serum. IgG2 and IgG3 have a preferred linkage of the galactose residues to the a1,3 arm mannose, whereas IgG1 has preferential linkage of galactose to the a1,6 arm mannose. IgG4 is reported to contain predominantly fully galactosylated structures (Jefferis ff et al., 1990). There is considerable amino acid sequence diversity in the variable regions, and N-linked glycosylation sites can occur in the variable regions. These are relatively rare. A recent survey of heavy chain variable region cDNA sequences showed that only 7 out of 75 (9.3%) had a potential N-linked glycosylation site in the variable region (Zhu et al., 2002). The glycans that occupy these sites are predominantly sialylated structures, with a high incidence of bisecting GlcNAc residues (Youings et al., 1996.; Wormald et al., 1997).

2.2. IgM IgM is found predominantly in the serum as a pentameric structure disulphide bridged at the CH3 domains and at the tail piece (a flexible region following the CH4 domain) and believed to form a ring structure. Pentameric IgM also has a J chain that contains a single N-linked glycosylation site. IgM can also adopt a hexameric structure that contains no J chain (Wiersma et al., 1998). IgM heavy chain (m chain) has five N-linked glycosylation sites at Asn-171, Asn-332, Asn-395, Asn-402, and Asn-563. Asn-402 and Asn-563 have been shown to be occupied by oligomannose structures (Chapman and Kornfeld, 1979; Wormald et al., 1991). The other N-linked glycosylation sites on each m chain in normal human serum IgM are occupied predominantly by complex biantennary glycans. The most predominant glycan is FcGlcNAc A G S (26% of total glycan pool). Sialylated structures 2 2 2 1

Human Immunoglobulin Glycosylation

31

Figure 2. Glycan Structure and IgG Glycoforms. a) Shows the general nomenclature used to describe sugar residues, bond angles and sugar linkages of the different ff glycan structures that occupy glycoproteins. b) Shows the predominant glycan structures that occupy the Asn-297 site in IgG. The glycans shown may also vary by the presence of absence of a core Fucose and/or bisecting GlcNAc.

32

J. N. Arnold et al.

Figure 3. IgA Glycosylation Types. IgA has two subclasses, IgA1 and IgA2, and both have N-linked glycosylation at Asn-263 and Asn-459. IgA1 contains nine potential O-linked sites in the hinge region, of which five or six have been shown to be occupied. *The sixth O-linked site occupies one or more of Ser-224, Thr-233, Ser-238 or Ser-240 (Tarelli et al., 2004). IgA2 has no potential O-linked sites in its hinge region. IgA2 is subdivided into IgA2m(1) which has two additional N-linked glycosylation sites in the CH1 domain and CH2 domain, and IgA2m(2) which has these additional N-linked sites but also a third additional CH1 domain N-linked glycosylation site.

(61.8%), core fucosylated (65%) and bisected structures (38%) are present in the total glycan pool (J.Arnold unpublished data).

2.3. IgA IgA has two conserved N-linked glycosylation sites, at Asn-263 in the CH2 domain and Asn-459 located in the 18 amino acid tail piece on each a chain. There are two subclasses of IgA designated IgA1 and IgA2. IgA2 has two forms that contain two (IgA2m(1)) or three (IgA2m(2)) extra conserved N-linked glycosylation sites respectively (Fig. 3). IgA occurs in several different ff oligomeric forms, and is present both in serum and in secretions. Serum IgA and Secretory IgA (SIgA), have distinct populations of glycan structures. The 23 amino acid hinge region in IgA1 contains nine potential O-linked sites of which five have been shown to be occupied (Mattu et al., 1998: Baenziger and Kornfeld, 1974b). These sites are Thr-228, Ser-230, Ser-232, with Thr-225 and Thr-236

Human Immunoglobulin Glycosylation

33

Figure 4. Core I Structures, Neutral, Mono-, and Di-Sialylated. The neutral, mono- and di-sialylated Core I O-linked glycan structures, that have been identified on serum IgA1 and also IgD hinge regions. The nomenclature is explained in Fig. 2.

partially occupied (Mattu et al., 1998). Recently a sixth occupied O-linked site at one or more of Ser-224, Thr-233, Ser-238 or Ser-240 (Tarelli et al., 2004) has been identified. No O-linked glycans have been identified on IgA2. 2.3.1. Serum IgA Serum IgA consists mainly of IgA1. IgA1 and IgA2 contain similar N-linked glycan structures (Endo et al., 1994; Royle et al., 2003). Over 80% of the glycans are di-galactosylated bi-antennary complex glycans. Less than 10% are tri- and tetra-antennary structures (Mattu et al., 1998). Sixty four percent of the glycan structures are sialylated and 95% of these are linked a2–6 to galactose. The predominant glycan is GlcNAc A G S (24%). The glycan pool has 36% of the glycans 2 2 2 2 containing a core fucose residue and 25% containing a bisecting GlcNAc residue. The oligosaccharides attached to the Fab in IgA2 differ ff from those that occupy the Fc in for example, the presence of triantennary structures and outer-arm fucose residues such as GlcNAc A G FS which accounts for 3.7% of total Fab glycan 2 3 3 3 pool (Mattu et al., 1998). The O-linked glycans on the heavy chain of IgA1 have been identified (Mattu et al., 1998, Field et al., 1994 and Rudd et al., 1994)). The hinge is predominantly occupied by mono-sialylated core I structures (37%) and neutral core I structures (31%) (Mattu et al., 1998) (Fig. 4). 2.3.2. Secretory IgA SIgA is a dimer, held together with a J chain (which has one N-linked site) and Secretory Component (SC). The SC is the extracellular portion of the epithelial polymeric Ig receptor (pIgR), and is required for transcytosis of the IgA across the epithelium to the mucosal surface. The SC has seven N-linked glycosylation sites. SIgA contains both IgA1 and IgA2 populations. SIgA is present in mucosal secretions such as colostrum and milk and can bind to microorganisms, their metabolic products and toxins, preventing their attachment

34

J. N. Arnold et al.

to the epithelium and facilitating their excretion. This process is known as immune exclusion (reviewed by Lamm, 1997). The N-linked glycans from the heavy chain of colostrum SIgA consist of approximately 15% sialylated structures (solely a2–6 linked sialic acids), with over 75% of structures containing a bisecting GlcNAc and 50% being core fucosylated. Oligomannose structures account for 12% of the N-linked glycan pool. There is a lack of glycan processing of the N-linked glycans on the a-chain of SIgA, as only 20% of structures are fully galactosylated, and 66% have an exposed terminal GlcNAc residue. The major structures occupying SIgA heavy chain are FcGlcNAc A B (30%), GlcNAc A B (21%) and FcGlcNAc A BG (8%) (Royle 2 2 2 2 2 2 1 et al., 2003). There is a large diversity of O-linked glycan structures on the heavy chain of SIgA1, which contains over 50 different ff structures of up to 15 residues in size (Royle et al., 2003), in contrast to the restricted pool of structures present on serum IgA1. The glycans occupying the J chain single N-linked site are predominantly sialylated biantennary structures (75%). Fifty percent of all structures are core fucosylated, and 50% of the neutral structures contain a bisecting GlcNAc residue (Royle et al., 2003). Interestingly no bisecting GlcNAc is present on the sialylated structures (Royle et al., 2003). The seven N-linked sites of the SC are occupied by a large diversity of structures, many of which are not found on the immunoglobulin heavy chains, for example, outer-arm fucosylated glycans. The presence of these structures may be explained partially by the fact that epithelial cell glycosylation machinery glycosylates the SC, whereas the plasma cell glycosylates the immunoglobulin. The SC N-linked glycans are predominantly bi-antennary structures. Tri-antennary (11.7%) and tetra-antennary (<1%) structures are also present (Royle et al., 2003). Over 70% of the glycans are sialylated, predominantly mono-sialylated structures, and over 65% of the glycans contain a core fucose (Royle et al., 2003).

2.4. IgD IgD has three N-linked glycosylation sites in the Fc at Asn-354, Asn-445, Asn-496 (Takahashi et al., 1982). Asn-354 in the CH2 domain is occupied solely by oligomannose structures (GlcNAc Man ) (Mellis and Baenziger, 1983a; Arnold 2 5-9 et al., 2004) which represent 34% of the total glycans. The predominant oligomannose structure is GlcNAc Man . Glucosylated mannose structures (GlcNAc Man Glc , 2 8 2 9 1 GlcNAc Man Glc and GlcNAc Man Glc ) are also present (Arnold et al., 2004; 2 8 1 2 7 1 Mellis and Baenziger, 1983a). The other 66% of glycan structures have been shown to terminate in galactose or sialic acid. These glycans occupy the two CH3 domain N-linked glycosylation sites Asn-445 and Asn-496. At these two CH3 N-linked sites 71% of the oligosaccharides are sialylated; both mono- (53%) and di-sialylated (47%) glycans have been identified. Twenty nine percent of glycans terminate in galactose residues, 50% contain core fucosylation and 50% of the glycans contain a bisecting GlcNAc (Arnold et al., 2004) at these two sites. The hinge region of IgD contains several potential O-linked glycosylation sites. In an IgD myeloma protein IgD:WAH, O-linked glycans occupy Ser-106 and Thr-126, -127, -131 and -132, although it is uncertain if Thr-131 and –132 are both occupied

Human Immunoglobulin Glycosylation

35

(Mellis and Baenziger, 1983b; Takahashi et al., 1982). Another myeloma IgD:NIG-65 contains seven O-linked glycosylation sites; the five identified in IgD:WAH and also Ser-110 and Thr-113 (Takayasu et al., 1982). The O-linked glycans present on the hinge region are solely Core I structures: di-, mono-sialylated and neutral structures (Fig. 4) (Arnold et al., 2004; Mellis and Baenziger, 1983b).

2.5. IgE IgE has seven N-linked glycosylation sites in the e chain at Asn-140, Asn-168, Asn-218, Asn-265, Asn-371, Asn-383, Asn-394 (Dorrington and Bennich, 1978). The Asn-394 N-linked glycosylation site is occupied solely by oligomannose structures (Dorrington and Bennich, 1978; Baenziger and Kornfeld, 1974b). The predominant oligomannose structure is GlcNAc Man (8.3% of the total glycan pool). The other 2 5 six exposed glycosylation sites on each e chain are occupied predominantly with sialylated glycan structures (46% mono- 42% di-sialylated structures), 12% galactose terminating structures, 68% core fucosylated and 14% bisected structures (Arnold et al., 2004).

3. MANNOSE BINDING LECTIN (MBL) AND THE LECTIN PATHWAY A OF COMPLEMENT ACTIVATION 3.1. MBL MBL (Fig. 5) is a glycoprotein, also known as Mannan/Mannose Binding Protein and is member of the collectin family of proteins (Malhotra et al., 1994). Collectins are large oligomeric proteins with multiple lectin domains and collagenous regions. MBL is synthesized in the liver and secreted into the blood stream. MBL is an important component of the innate immune system, which binds calciumdependently to sugars that have hydroxyl groups on the carbon-3 and carbon-4 orientated in the equatorial plane of the pyranose ring (Weis et al., 1992). This gives MBL affinity for mannose, fucose and N-acetyl glucosamine (GlcNAc) (Turner et al., 1996). This specificity allows MBL to bind to sugar arrays on the surfaces of microorganisms, including bacteria, viruses and fungi (Holmskov et al., 1994), but not to human glycoprotein glycans, the structures of which generally terminate in galactose or sialic acid. MBL has a structure and function similar to that of C1q, the recognition molecule that initiates the classical pathway of complement. MBL binds to sugar residues via the Carbohydrate Recognition Domain (CRD) (lectin) heads. The affinity of a single CRD for carbohydrate is very weak (10−3M) (Iobst et al., 1994). Multiple CRD binding leads to a much greater avidity. Levels of MBL in human serum vary greatly between individuals (Turner, 1996), from below 50ng/ml to above 10ug/ml. The variation of MBL levels is caused by several identified polymorphisms in the coding sequence and promoter regions of the MBL gene (Madsen et al., 1995). The coding sequence polymorphisms disrupt the Gly-X-Y repeat that is found in the collagenous region destablilising the collagen triple helix formation (Sumiya et al., 1991: Lipscombe et al., 1992), and consequently heterozygotes have low levels of MBL in the blood. Low levels of MBL have been linked to severe and recurrent infections in children (Summerfield et al., 1997).

36

J. N. Arnold et al.

Figure 5. Structure of MBL. MBL is composed of identical 25kDa polypeptides that form a trimer through the formation of a triple helix of the collagen-like regions that is the basis of the MBL subunit (or monomer). This subunit can then disulphide bridge at its N-terminus to form higher order structures. MBL circulates in the serum mainly as a hexameric molecule (i.e. six subunits, 18 polypeptide chains). The collagen-like region is attached to a Carbohydrate Recognition Domain (CRD) which binds to sugar arrays that have hydroxyl groups on the carbon-3 and carbon-4 orientated in the equatorial plane of the pyranose ring (Weis et al., 1992).

MBL participates in the host defense response through two major pathways. Firstly, it acts directly as an opsonin, promoting phagocytosis of foreign material to which it has bound. There are several candidate receptors through which this process may be mediated. The main candidate receptor is cell surface calreticulin (Sim et al., 1998; Ogden et al., 2001), but there is also evidence for the participation of complement receptor 1 (CR1: CD35) (Ghiran et al., 2000). The second pathway through which MBL functions is by triggering the lectin pathway of complement activation via MBL associated serine protease-2 (MASP-2) (Vorup-Jensen et al., 2000; Hajela et al., 2002).

3.2. MASPs MBL circulates in the serum bound to the serine protease pro-enzymes, MASPs, of which three have been identified to date; MASP-1, MASP-2 (Matsushita et al., 1992; Thiel et al., 1997) and MASP-3 (Dahl et al., 2001). The MBL-MASP complex was shown to be capable of consuming the complement components C2 and C4 (Ikeda et al., 1987). It is now generally accepted from recombinant protein work that MASP-2 is solely responsibly for the cleavage of C2 and C4 to produce C4b2a (Vorup-Jensen et al., 2000). This provides MASP-2 with a function similar to that of C1s in the C1 complex. The biological roles for MASP-1 and MASP-3 are currently unknown. MASP-1 has been shown to cleave ‘dead’ C3 (C3 in which the thiolester bond has hydrolyzed) at a slow rate. Cleavage of physiological ‘live’ C3 (C3 in which the thiolester bond is intact) occurs at a very slow rate, suggested to

Human Immunoglobulin Glycosylation

37

Figure 6. The Complement System. The lectin and classical pathways rely on cleavage of complement protein C4, forming C4b, to which C2 binds and is cleaved, that leads to the formation of C4b2a, a C3 convertase that activates C3. C3 is cleaved into C3a and C3b, which is further cleaved to form the iC3b opsonin. Activation of C3 leads to the formation of the membrane attack complex which causes cell lysis. The alternative pathway relies on preformed C3b, or C3(H O) which forms spontaneously at a slow rate. C3b binds factor B, which is 2 cleaved by Factor D to form another C3 convertase, C3bBb. The C3 convertases are inactivated by decay accelerating factor, Factor H, C4b-binding protein and complement receptor I, which speed up the dissociation of the convertase. C3b and C4b when bound by cofactors such as Factor H are cleaved by Factor I and inactivated. The C3 convertases have naturally short half lives in the circulation.

be too slow to be physiologically important (Hajela et al., 2002). MASP-1 also cleaves Factor XIII (plasma transglutaminase) and fibrinogen, two substrates of thrombin, potentially implicating MASP-1 in localized coagulation (Hajela et al., 2002).

3.3. Complement and the lectin pathway of complement activation The complement system (Fig. 6) is a major part of the innate immune response that eliminates foreign and altered-self cells by opsonisation and lysis. It is the body’s first line of defense against infectious agents. The complement system recognizes foreign matter through proteins with specific binding affinities to potential Pathogen Associated Molecular Patterns (PAMPs) including lipopolysaccharide, lipoproteins, peptidoglycan, lipoarabinomannan and oligosaccharide and charge arrays. The binding of ‘recognition’ proteins MBL and C1q leads to the activation of the complement cascade which is controlled and propagated through serine proteases and regulated directly by a serpin, C1-inhibitor (Cooper, 1985) that binds and inactivates these cascade triggering proteases. There are three routes of complement activation; the classical, alternative and lectin pathways. The classical pathway is triggered by the C1 complex. The C1 complex is composed of C1q and 2 each of the serine proteases C1r and C1s (Arlaud et al., 1987).

38

J. N. Arnold et al.

Figure 7. MBL and the Immunoglobulins A summary of the interaction of MBL with the immunoglobulins.

MBL is the recognition molecule of the lectin pathways of complement activation. Binding of MBL to a target activates MASPs. Activated MASP-2 cleaves the complement protein C4, forming C4b, to which C2 binds and is also cleaved by MASP-2, leading to the formation of C4b2a, a C3 convertase that activates C3. C3 is cleaved into C3a and C3b, which is further cleaved to form the iC3b opsonin. Activation of C3 leads on to the formation of the membrane attack complex (MAC) that causes cell lysis.

4. THE INTERACTION OF MBL WITH THE IMMUNOGLOBULINS The immunoglobulins contain populations of glycans, some of which terminate in mannose or GlcNAc which are potential binding ligands for lectin-like recognition proteins of the innate immune system, such as MBL, macrophage Mannose Receptor and the surfactant proteins SP-A and SP-D. The known interactions of MBL with immunoglobulins are summarised in Fig. 7. The glycans of IgG have restricted motion because of the terminal galactose residues attached to the glycan structures. The IgG CH2 domain has a hydrophobic area on the peptide surface of each heavy chain. Galactoses attached to the a1,6 arm of the glycan interact with this region, and this together with >80 other interactions such as hydrogen bonding and van der Waals interactions holds the glycan in contact with the protein surface, which also prevents further processing to attach terminal sialic acid (Wormald et al., 1997). The glycans therefore have limited

Human Immunoglobulin Glycosylation

39

mobility. In IgG-G0, the glycans do not have terminal galactose residues, but terminal GlcNAc residues. These glycans are more mobile, as the glycan-protein interactions are not sufficient to hold the glycan anchored to the protein surface (Wormald et al., 1997). MBL has been shown to bind to the terminal GlcNAc residues of the IgG-G0 glycans (Malhotra et al., 1995). IgG-G0 glycoforms have been shown to increase dramatically in Rheumatoid Arthritis (RA) (Parekh et al., 1985). This increase has been shown to correlate with disease activity (Rook et al., 1991). Garred et al. (2000) correlated MBL levels with disease onset and progression in RA patients. This was consistent with the suggestion by Malhotra et al. (1995) that activating the lectin pathway of the complement system could be a potential route to additional inflammation in RA. IgD has the same domain structure as IgG, however the glycans found at the N-linked site homologous to that in IgG (Asn-297 in IgG and Asn-354 in IgD) are solely oligomannose structures. Although these are potential ligands for MBL, MBL does not bind IgD (Arnold et al., 2004). The oligomannose glycans at Asn-354 are inaccessible to MBL because the complex glycans occupying Asn-445 on the CH3 domain block the access to the oligomannose glycans (Arnold et al., 2004). MBL has been shown to interact with certain polymeric types of IgA but not SIgA (Roos et al., 2001; Royle et al., 2003). MBL binds to polymeric and dimeric forms of IgA with the highest avidity, but MBL does not bind to monomeric serum IgA (Roos et al., 2001). The glycans in IgA with which MBL is interacting have not been identified although it has been inferred from models that all the glycans on IgA (but not SIgA) are exposed and could potentially bind (Mattu et al., 1998). SIgA contains a large array of glycans terminating in GlcNAc residues (Royle et al., 2003). However these structures are masked from lectin binding by the SC which wraps around the IgA heavy chains. The SC itself contains predominantly sialylated complex glycans (Royle et al., 2003). The SC structure blocks access of MBL to the IgA glycans, although it has been suggested that these may be revealed when SC binds to pathogens (Royle et al., 2003). IgE has a different ff domain structure from IgG, IgD and IgA. The hinge peptides are replaced by immunoglobulin domains which form a rigid dimer. The crystal structure of the Fc and CH2 hinge domain showed an asymmetrically bent quaternary structure, where the CH2 domain bends over one side of the Fc (Wan et al., 2002). Oligomannose structures occupy Asn-394 (homologous site to Asn-297 in IgG and Asn-354 in IgD) (Dorrington and Bennich, 1978; Arnold et al., 2004). MBL, however, does not bind IgE (Arnold et al., 2004). The access to these oligomannose glycans is prevented because of the CH2 hinge domain which completely blocks access to the oligomannose glycans from one side. The CH2 hinge domain is proposed to ‘flip’ between two bent quaternary conformations with the CH2 hinge domains on either side of the Fc domain, preventing access to the oligomannose glycans from both sides of the Fc (Arnold et al., 2004). IgM is found in the serum as a pentamer and a hexamer. The IgM monomer unit has a very similar structure to that of IgE, with an Ig domain replacing the hinge region. IgM, however, contains oligomannose glycans at two N-linked glycosylation sites, located at Asn-402, homologous to the Asn-394 in IgE (and Asn-297 in IgG and Asn-354 in IgD) and at Asn-563 at the C-terminus (Wormald et al., 1991). It has been shown that immobilised human IgM does not bind MBL on microtitre

40

J. N. Arnold et al.

plates (Roos et al., 2003). The oligomannose glycans at Asn-402 are predicted to be inaccessible on the basis that it is similar in structure to IgE, where the CH2 domain ‘flips,’ between two bent quaternary conformations (J. Arnold and M. Wormald, unpublished data). The accessibility of the tail piece oligomannose glycans is currently unknown and under investigation. It may be the case that the structural change that occurs upon IgM binding to antigen (referred to as the staple form of IgM), may present the oligomannose sugars to MBL for binding. There have been reports (see Fig. 7) of human IgM binding to rat MBL (Koppel and Solomon, 2001) and human, bovine and murine IgM binding to rabbit MBL (Nevens et al., 1992) (Fig. 7). In the latter case it appears that MBL may be binding only a small subpopulation of human IgM (J.Arnold, unpublished data).

5. CONCLUSIONS The glycans attached to the immunoglobulins have a great diversity in structure, location and number. The predominant complex glycan structures are biantennary, which are variably galactosylated and sialylated. There is also a high proportion of structures that contain either or both a bisecting GlcNAc and/or core fucose residue, in different ff percentages between the immunoglobulins. Glycan structures that could act as potential ligands for MBL have been identified on all the immunoglobulins. In human serum only IgG-G0 and polymeric and dimeric IgA have been shown to bind MBL and initiate the lectin pathway of complement (Malhotra et al., 1995; Roos et al., 2001). In other immunoglobulins small quantities of GlcNAc-terminating glycan structures have been identified in the glycan pool. These structures may define small subpopulations of the immunoglobulins to which MBL could bind.

REFERENCES Amzel, L. M., and Poljak, R. J. (1979). Three-dimensional structure of immunoglobulins. Annu Rev Biochem 48, 961–997. Anton, L. C., Alcolea, J. M., Sanchez-Corral, P., Marques, G., Sanchez, A., and Vivanco, F. (1989). C3 binds covalently to the C gamma 3 domain of IgG immune aggregates during complement activation by the alternative pathway. Biochem J 257, 831–838. Arlaud, G. J., and Colomb, M. G. (1987). Modelling of C1, the first component of human complement: towards a consensus? Mol Immunol 24, 317. Arnold, J. N., Radcliffe, ff C. M., Wormald, M. R., Royle, L., Harvey, D. J., Crispin, M., Dwek, R. A., Sim, R. B., and Rudd, P. M. (2004) In Press J Immunol. Baenziger, J., and Kornfeld, S. (1974). Structure of the carbohydrate units of IgA1 immunoglobulin. I. Composition, glycopeptide isolation, and structure of the asparagines linked oligosaccharide units. J Biol Chem 249, 7260–7269. Baenziger, J., Kornfeld, S., and Kochwa, S. (1974). Structure of the carbohydrate units of IgE immunoglobulin. I. Over-all composition, glycopeptide isolation, and structure of the high mannose oligosaccharide unit. J Biol Chem 249, 1889–1896. Baenziger, J., Kornfeld, S., and Kochwa, S. (1974). Structure of the carbohydrate units of IgE immunoglobulin. II. Sequence of the sialic acid-containing glycopeptides. J Biol Chem 249, 1897–1903. Bos, N. A., Bun, J. C., Popma, S. H., Cebra, E. R., Deenen, G. J., van der Cammen, M. J., Kroese, F. G., and Cebra, J. J. (1996). Monoclonal immunoglobulin A derived from peritoneal B cells is encoded by

Human Immunoglobulin Glycosylation

41

both germ line and somatically mutated VH genes and is reactive with commensal bacteria. Infect Immun 64, 616–623. Butler, M., Quelhas, D., Critchley, A. J., Carchon, H., Hebestreit, H. F., Hibbert, R. G., Vilarinho, L., Teles, E., Matthijs, G., Schollen, E., et al. (2003). Detailed glycan analysis of serum glycoproteins of patients with congenital disorders of glycosylation indicates the specific defective glycan processing step and provides an insight into pathogenesis. Glycobiology 13, 601–622. Chapman, A., and Kornfeld, R. (1979). Structure of the high mannose oligosaccharides of a human IgM myeloma protein. II. The minor oligosaccharides of high mannose glycopeptide. J Biol Chem 254, 824–828. Cooper, N. R. (1985). The classical complement pathway: activation and regulation of the first complement component. Adv Immunol 37, 151–216. Crispin, M. D., Ritchie, G. E., Critchley, A. J., Morgan, B. P., Wilson, I. A., Dwek, R. A., Sim, R. B., and Rudd, P. M. (2004). Monoglucosylated glycans in the secreted human complement component C3: implications for protein biosynthesis and structure. FEBS Lett 566, 270–274. Dahl, M. R., Thiel, S., Matsushita, M., Fujita, T., Willis, A. C., Christensen, T., Vorup-Jensen, T., and Jensenius, J. C. (2001). MASP-3 and its association with distinct complexes of the mannan-binding lectin complement activation pathway. Immunity 15, 127–135. Delacroix, D. L., Dive, C., Rambaud, J. C., and Vaerman, J. P. (1982). IgA subclasses in various secretions and in serum. Immunology 47, 383–385. Dorrington, K. J., and Bennich, H. H. (1978). Structure-function relationships in human immunoglobulin E. Immunol Rev 41, 3–25. Duncan, A. R., and Winter, G. (1988). The binding site for C1q on IgG. Nature 332, 738–740. Endo, T., Mestecky, J., Kulhavy, R., and Kobata, A. (1994). Carbohydrate heterogeneity of human myeloma proteins of the IgA1 and IgA2 subclasses. Mol Immunol 31, 1415–1422. Field, M. C., Amatayakul-Chantler, S., Rademacher, T. W., Rudd, P. M., and Dwek, R. A. (1994). Structural analysis of the N-glycans from human immunoglobulin A1: comparison of normal human serum immunoglobulin A1 with that isolated from patients with rheumatoid arthritis. Biochem J 299 (Pt 1), 261–275. Gala, F. A., and Morrison, S. L. (2002). The role of constant region carbohydrate in the assembly and secretion of human IgD and IgA1. J Biol Chem 277, 29005–29011. Garred, P., Madsen, H. O., Marquart, H., Hansen, T. M., Sorensen, S. F., Petersen, J., Volck, B., Svejgaard, A., Graudal, N. A., Rudd, P. M., et al. (2000). Two edged role of mannose binding lectin in rheumatoid arthritis: a cross sectional study. J Rheumatol 27, 26–34. Ghiran, I., Barbashov, S. F., Klickstein, L. B., Tas, S. W., Jensenius, J. C., and Nicholson-Weller, A. (2000). Complement receptor 1/CD35 is a receptor for mannan-binding lectin. J Exp Med 192, 1797–1808. ff B. E., Ferluga, J., Hajela, S., Gal, P., and Sim, Hajela, K., Kojima, M., Ambrus, G., Wong, K. H., Moffatt, R. B. (2002). The biological functions of MBL-associated serine proteases (MASPs). Immunobiology 205, 467–475. Holmskov, U., Malhotra, R., Sim, R. B., and Jensenius, J. C. (1994). Collectins: collagenous C- type lectins of the innate immune defense system. Immunol Today 15, 67–74. Ikeda, K., Sannoh, T., Kawasaki, N., Kawasaki, T., and Yamashina, I. (1987). Serum lectin with known structure activates complement through the classical pathway. J Biol Chem 262, 7451–7454. Iobst, S. T., and Drickamer, K. (1994). Binding of sugar ligands to Ca(2+)-dependent animal lectins. II. Generation of high-affinity galactose binding by site-directed mutagenesis. J Biol Chem 269, 15512–15519. Jefferis, ff R., Lund, J., Mizutani, H., Nakagawa, H., Kawazoe, Y., Arata, Y., and Takahashi, N. (1990). A comparative study of the N-linked oligosaccharide structures of human IgG subclass proteins. Biochem J 268, 529–537. Johansen, F. E., Braathen, R., and Brandtzaeg, P. (2001). The J chain is essential for polymeric Ig receptormediated epithelial transport of IgA. J Immunol 167, 5185–5192. Koppel, R., and Solomon, B. (2001). IgM detection via selective recognition by mannose-binding protein. J Biochem Biophys Methods 49, 641–647. Lamm, M. E. (1997). Interaction of antigens and antibodies at mucosal surfaces. Annu Rev Microbiol 51, 311–340. Lipscombe, R. J., Sumiya, M., Hill, A. V., Lau, Y. L., Levinsky, R. J., Summerfield, J. A., and Turner, M. W. (1992). High frequencies in African and non-African populations of independent mutations in the mannose binding protein gene. Hum Mol Genet 1, 709–715.

42

J. N. Arnold et al.

Litman, G. W., Anderson, M. K., and Rast, J. P. (1999). Evolution of antigen binding receptors. Annu Rev Immunol 17, 109–147. ff molecular events Madsen, H. O., Satz, M. L., Hogh, B., Svejgaard, A., and Garred, P. (1998). Different result in low protein levels of mannan-binding lectin in populations from southeast Africa and South America. J Immunol 161, 3169–3175. Malhotra, R., Lu, J., Holmskov, U., and Sim, R. B. (1994). Collectins, collectin receptors and the lectin pathway of complement activation. Clin Exp Immunol 97 Suppl 2, 4–9. Malhotra, R., Wormald, M. R., Rudd, P. M., Fischer, P. B., Dwek, R. A., and Sim, R. B. (1995). Glycosylation changes of IgG associated with rheumatoid arthritis can activate complement via the mannose-binding protein. Nat Med 1, 237–243. Matsushita, M., and Fujita, T. (1992). Activation of the classical complement pathway by mannose-binding protein in association with a novel C1s-like serine protease. J Exp Med 176, 1497–1502. Mattu, T. S., Pleass, R. J., Willis, A. C., Kilian, M., Wormald, M. R., Lellouch, A. C., Rudd, P. M., Woof, J. M., and Dwek, R. A. (1998). The glycosylation and structure of human serum IgA1, Fab, and Fc regions and the role of N-glycosylation on Fc alpha receptor interactions. J Biol Chem 273, 2260–2272. Mellis, S. J., and Baenziger, J. U. (1983). Structures of the O-glycosidically linked oligosaccharides of human IgD. J Biol Chem 258, 11557–11563. Mellis, S. J., and Baenziger, J. U. (1983). Structures of the oligosaccharides present at the three asparaginelinked glycosylation sites of human IgD. J Biol Chem 258, 11546–11556. ff R. Mimura, Y., Church, S., Ghirlando, R., Ashton, P. R., Dong, S., Goodall, M., Lund, J., and Jefferis, (2000). The influence of glycosylation on the thermal stability and effector ff function expression of human IgG1-Fc: properties of a series of truncated glycoforms. Mol Immunol 37, 697–706. Nevens, J. R., Mallia, A. K., Wendt, M. W., and Smith, P. K. (1992). Affinity chromatographic purification of immunoglobulin M antibodies utilizing immobilized mannan binding protein. J Chromatogr 597, 247–256. Norderhaug, I. N., Johansen, F. E., Schjerven, H., and Brandtzaeg, P. (1999). Regulation of the formation and external transport of secretory immunoglobulins. Crit Rev Immunol 19, 481– 508. Ogden, C. A., deCathelineau, A., Hoffmann, ff P. R., Bratton, D., Ghebrehiwet, B., Fadok, V. A., and Henson, P. M. (2001). C1q and mannose binding lectin engagement of cell surface calreticulin and CD91 initiates macropinocytosis and uptake of apoptotic cells. J Exp Med 194, 781–795. Parekh, R. B., Dwek, R. A., Sutton, B. J., Fernandes, D. L., Leung, A., Stanworth, D., Rademacher, T. W., Mizuochi, T., Taniguchi, T., Matsuta, K., and et al. (1985). Association of rheumatoid arthritis and primary osteoarthritis with changes in the glycosylation pattern of total serum IgG. Nature 316, 452–457. Rook, G. A., Steele, J., Brealey, R., Whyte, A., Isenberg, D., Sumar, N., Nelson, J. L., Bodman, K. B., Young, A., Roitt, I. M., and et al. (1991). Changes in IgG glycoform levels are associated with remission of arthritis during pregnancy. J Autoimmun 4, 779–794. Roos, A., Bouwman, L. H., Munoz, J., Zuiverloon, T., Faber-Krol, M. C., Fallaux-van den Houten, F. C., Klar-Mohamad, N., Hack, C. E., Tilanus, M. G., and Daha, M. R. (2003). Functional characterization of the lectin pathway of complement in human serum. Mol Immunol 39, 655–668. Roos, A., Bouwman, L. H., van Gijlswijk-Janssen, D. J., Faber-Krol, M. C., Stahl, G. L., and Daha, M. R. (2001). Human IgA activates the complement system via the mannan-binding lectin pathway. J Immunol 167, 2861–2868. Royle, L., Roos, A., Harvey, D. J., Wormald, M. R., van Gijlswijk-Janssen, D., Redwan el, R. M., Wilson, I. A., Daha, M. R., Dwek, R. A., and Rudd, P. M. (2003). Secretory IgA N- and O- glycans provide a link between the innate and adaptive immune systems. J Biol Chem 278, 20140–20153. Rudd, P. M., Fortune, F., Patel, T., Parekh, R. B., Dwek, R. A., and Lehner, T. (1994). A human T-cell receptor recognizes ‘O’-linked sugars from the hinge region of human IgA1 and IgD. Immunology 83, 99–106. Sarmay, G., Lund, J., Rozsnyay, Z., Gergely, J., and Jefferis, ff R. (1992). Mapping and comparison of the interaction sites on the Fc region of IgG responsible for triggering antibody dependent cellular cytotoxicity (ADCC) through different ff types of human Fc gamma receptor. Mol Immunol 29, 633–639. Shields, R. L., Namenuk, A. K., Hong, K., Meng, Y. G., Rae, J., Briggs, J., Xie, D., Lai, J., Stadlen, A., Li, B.,

Human Immunoglobulin Glycosylation

43

et al. (2001). High resolution mapping of the binding site on human IgG1 for Fc gamma RI, Fc gamma RII, Fc gamma RIII, and FcRn and design of IgG1 variants with improved binding to the Fc gamma R. J Biol Chem 276, 6591–6604. Sim, R. B., Moestrup, S. K., Stuart, G. R., Lynch, N. J., Lu, J., Schwaeble, W. J., and Malhotra, R. (1998). Interaction of C1q and the collectins with the potential receptors calreticulin (cC1qR/collectin receptor) and megalin. Immunobiology 199, 208–224. Sorensen, V., Sundvold, V., Michaelsen, T. E., and Sandlie, I. (1999). Polymerization of IgA and IgM: roles of Cys309/Cys414 and the secretory tailpiece. J Immunol 162, 3448–3455. Stockert, R. J. (1995). The asialoglycoprotein receptor: relationships between structure, function, and expression. Physiol Rev 75, 591–609. Sumiya, M., Super, M., Tabona, P., Levinsky, R. J., Arai, T., Turner, M. W., and Summerfield, J. A. (1991). Molecular basis of opsonic defect in immunodeficient children. Lancet 337, 1569– 1570. Summerfield, J. A., Sumiya, M., Levin, M., and Turner, M. W. (1997). Association of mutations in mannose binding protein gene with childhood infection in consecutive hospital series. Bmj 314, 1229–1232. Takahashi, N., Tetaert, D., Debuire, B., Lin, L. C., and Putnam, F. W. (1982). Complete amino acid sequence of the delta heavy chain of human immunoglobulin D. Proc Natl Acad Sci U S A 79, 2850–2854. Takayasu, T., Suzuki, S., Kametani, F., Takahashi, N., Shinoda, T., Okuyama, T., and Munekata, E. (1982). Amino acid sequence of galactosamine-containing glycopeptides in the hinge region of a human immunoglobulin D. Biochem Biophys Res Commun 105, 1066–1071. Tarelli, E., Smith, A. C., Hendry, B. M., Challacombe, S. J., and Pouria, S. (2004). Human serum IgA1 is substituted with up to six O-glycans as shown by matrix assisted laser desorption ionisation time-offlight mass spectrometry. Carbohydr Res 339, 2329–2335. Tarentino, A. L., Plummer, T. H., Jr., and Maley, F. (1974). The release of intact oligosaccharides from specific glycoproteins by endo-beta-N-acetylglucosaminidase H. J Biol Chem 249, 818– 824. Thiel, S., Vorup-Jensen, T., Stover, C. M., Schwaeble, W., Laursen, S. B., Poulsen, K., Willis, A. C., Eggleton, P., Hansen, S., Holmskov, U., et al. (1997). A second serine protease associated with mannan-binding lectin that activates complement. Nature 386, 506–510. Turner, M. W. (1996). Mannose-binding lectin: the pluripotent molecule of the innate immune system. Immunol Today 17, 532–540. Valdimarsson, H., Stefansson, M., Vikingsdottir, T., Arason, G. J., Koch, C., Thiel, S., and Jensenius, J. C. (1998). Reconstitution of opsonizing activity by infusion of mannan-binding lectin (MBL) to MBLdeficient humans. Scand J Immunol 48, 116–123. Van Boxel, J. A., Paul, W. E., Terry, W. D., and Green, I. (1972). Communications. IgD-bearing human lymphocytes. J Immunol 109, 648–651. Vorup-Jensen, T., Petersen, S. V., Hansen, A. G., Poulsen, K., Schwaeble, W., Sim, R. B., Reid, K. B., Davis, S. J., Thiel, S., and Jensenius, J. C. (2000). Distinct pathways of mannan- binding lectin (MBL)- and C1-complex autoactivation revealed by reconstitution of MBL with recombinant MBL-associated serine protease-2. J Immunol 165, 2093–2100. Wan, T., Beavil, R. L., Fabiane, S. M., Beavil, A. J., Sohi, M. K., Keown, M., Young, R. J., Henry, A. J., Owens, R. J., Gould, H. J., and Sutton, B. J. (2002). The crystal structure of IgE Fc reveals an asymmetrically bent conformation. Nat Immunol 3, 681–686. Weis, W. I., Drickamer, K., and Hendrickson, W. A. (1992). Structure of a C-type mannose- binding protein complexed with an oligosaccharide. Nature 360, 127–134. Wiersma, E. J., Collins, C., Fazel, S., and Shulman, M. J. (1998). Structural and functional analysis of J chain-deficient IgM. J Immunol 160, 5979–5989. Wormald, M. R., Rudd, P. M., Harvey, D. J., Chang, S. C., Scragg, I. G., and Dwek, R. A. (1997). Variations in oligosaccharide-protein interactions in immunoglobulin G determine the site-specific glycosylation profiles and modulate the dynamic motion of the Fc oligosaccharides. Biochemistry 36, 1370–1380. Wormald, M. R., Wooten, E. W., Bazzo, R., Edge, C. J., Feinstein, A., Rademacher, T. W., and Dwek, R. A. (1991). The conformational effects ff of N-glycosylation on the tailpiece from serum IgM. Eur J Biochem 198, 131–139. Youings, A., Chang, S. C., Dwek, R. A., and Scragg, I. G. (1996). Site-specific glycosylation of human immunoglobulin G is altered in four rheumatoid arthritis patients. Biochem J 314 (Pt 2), 621–630. Zhu, D., McCarthy, H., Ottensmeier, C. H., Johnson, P., Hamblin, T. J., and Stevenson, F. K. (2002). Acquisition of potential N-glycosylation sites in the immunoglobulin variable region by somatic mutation is a distinctive feature of follicular lymphoma. Blood 99, 2562–2568.

10

GELATINASE B PARTICIPATES IN COLLAGEN II DEGRADATION AND RELEASES GLYCOSYLATED REMNANT EPITOPES IN RHEUMATOID ARTHRITIS

P. E. Van den Steen, B. Grillet, and G. Opdenakker Laboratory of Immunobiology Rega Institute for Medical Research University of Leuven Minderbroedersstraat 10, 3000 Leuven, Belgium

1. INTRODUCTION Rheumatoid arthritis is an autoimmune disease characterized by chronic inflammation of the joints. It is associated with the activation of autoreactive T-cells and with production of autoantibodies. The main auto-antigen is collagen type II, which is a major constituent of the cartilage in the joint. The inflammation causes cartilage degradation, hyperplasia of synovial membranes, accumulation of excessive synovial fluid, and, in the worst cases, bone erosion. The exact aetiology is not known, but it is clear that inflammatory reactions and auto-antibodies, which activate the complement cascade, are main causes of the cartilage degradation. Inhibition of inflammation by interference with some of the main pro-inflammatory cytokines, interleukin-1 (IL-1) and tumor necrosis factor-a (TNF-a) has proven to be beneficial and constitutes the basis of currently approved treatments. Matrix metalloproteinases are a family of neutral Zn2+-dependent endoproteases, which share a number of homologous domains (Nagase and Woessner, 1999). These domains are the Zn2+-containing active site, kept inactive by a propeptide in the pro-enzyme form, and a hemopexin domain (Fig. 1). The latter domain is present in most MMPs except for matrilysins (MMP-7 and -26) and cysteine-array MMP (CA-MMP or MMP-23). Several MMPs contain additional domains, such as a carboxyterminal membrane anchor in membrane-type MMPs (MT-MMPs), a fibronectin-like domain in gelatinases A and B (MMP-2 and -9), and a unique mucin-type domain in gelatinase B. The latter domain is often named collagen type 45 John S. Axford (ed.), Glycobiology and Medicine, 45-55. © 2005 Springer. Printed in the Netherlands.

46

P. E. Van den Steen et al.

Figure 1. Domain structure of MMPs MMPs are a family of neutral endoproteinases with a number of conserved domains, including the prodomain and the active enzyme and Zn2+-binding domains, which together form the active site of the enzymes. Most MMPs have also a carboxyterminal hemopexin domain, except for the matrilysins (MMP-7 and –26) and cysteine-array MMP (MMP-23). MT-MMPs additionally contain a GPI membrane anchor (for MT4-MMP and MT6-MMP) or a transmembrane domain with a short cytoplasmic domain (the other MT-MMPs). Gelatinases contain a fibronectin domain, and gelatinase B contains also a unique mucin-like domain, named the collagen V domain, which is an ideal attachment site for clustered O-linked sugars. The theoretical attachment sites for N-linked sugars are indicated with a Y symbol, one of which is conserved among most MMPs (indicated by a filled diamond). On top, the conserved histidines are shown which interact with the catalytic Zn2+, and the conserved cysteine in the propeptide, also interacting with the catalytic Zn2+ to keep the enzyme latent.

Gelatinase B Participation in Collagen II Degradation

47

V domain because it contains a large number of proline residues, conferring some homology to collagen type V. However, it is not a true collagen-like domain since it does not contain a glycine at every three amino acids, which is a main characteristic of collagen. Instead, this domain in gelatinase B contains repeats of the sequence S/T-X-X-P, and is therefore an ideal site for the clustered attachment of O-linked glycans, which are abundantly present on gelatinase B (Van den Steen et al., 1998; Mattu et al., 2000; Van den Steen et al., 2001). MMPs are well-known for their ability to cleave most extracellular matrix components. In particular, the three classical collagenases, interstitial collagenase (MMP-1), neutrophil collagenase (MMP-8) and collagenase-3 (MMP-13), are able to cleave triple-helical collagen, which is otherwise highly resistant to most proteases (Jeffrey, ff 1998). Furthermore, stromelysins and matrilysins degrade proteoglycans, and gelatinases degrade a variety of other ECM components such as gelatin, elastin (Senior et al., 1991), link protein (Nguyen et al., 1993) and collagen type V (Hibbs et al., 1987). Besides cleaving ECM proteins, MMPs also cleave a variety of regulatory molecules, such as serine protease inhibitors, cytokines and chemokines (Opdenakker et al., 2001).

2. ROLE OF GELATINASE B IN RHEUMATOID ARTHRITIS 2.1. Expression of Gelatinase B in Rheumatoid Arthritis The expression of gelatinase B in synovial fluids of patients with rheumatoid arthritis has been documented more than 10 years ago (Opdenakker et al., 1991) and was confirmed in several other studies (Sopata et al., 1995; Ahrens et al., 1996; Gruber et al., 1996). The amount of gelatinase B, as measured by gelatin zymography, correlates with inflammatory markers such as IL-6 and IL-8 (Opdenakker et al., 1991; Van den Steen et al., 2002b). Furthermore, gelatinase B is present under different ff forms in synovial fluids: monomers of around 92 kDa, homodimers of 200 kDa and a complex of gelatinase B with neutrophil gelatinase B-associated lipocalin (NGAL) (Fig. 2). The latter complex is only synthesized by neutrophils (Kjeldsen et al., 1993; Triebel et al., 1992), indicating that neutrophils are the main producers of gelatinase B in the synovial fluid. In fact, the chemokine IL-8, which is also upregulated in the synovial fluids (Peichl et al., 1991; Rampart et al., 1992; Seitz et al., 1992), attracts neutrophils from the blood vessels to the synovial fluid (Akahoshi et al., 1994) and stimulates these cells to degranulate. Since gelatinase B is present in the granules in the proform, this results in the release of progelatinase B (Masure et al., 1991). The propeptide may be removed in different ff ways, including proteolysis by several proteases and oxidation of a conserved cysteine in the propeptide, which normally is bound to the Zn2+ ion (Van den Steen et al., 2002a). Which activation pathway is active in the synovial fluid of RA patients is unknown, but it is likely that stromelysin-1 contributes to the activation, since it is also increased in arthritic synovial fluids (Ribbens et al., 2000) and it is an efficient activator of gelatinase B (Ogata et al., 1992). However, it is also possible that reactive oxygen species, produced by the activated neutrophils, are contributing to the activation (Peppin and Weiss, 1986). The active form of the protease is seen in a limited number

48

P. E. Van den Steen et al.

Figure 2. Gelatin zymography of synovial fluids Synovial fluids from arthritis patients were analysed by gelatin zymography. The positions of monomeric pro- and active gelatinase B, homodimeric gelatinase B, gelatinase B complexed with NGAL and gelatinase A are indicated. In some synovial fluids (lanes 1 and 6), the activated forms of gelatinase B are present, whereas other synovial fluids only contain gelatinase A (lane 3).

of samples. However, the presence of active gelatinase B on zymography does not prove its activity in the original synovial fluid, since it may still be complexed with the tissue inhibitor of metalloproteinases (TIMP)-1 (Murphy and Willenbrock, 1995). As this complex dissociates during the electrophoresis, TIMP-1 cannot influence the activity on gelatin-zymography. Therefore, it was analysed whether net activity was present in the synovial fluids, using an activity assay with fluorescently labelled gelatin coated onto microspheres. A high activity was measured in some patients, and in serial samples from the same patients the activity varied greatly with time (Van V den Steen et al., 2002b).

2.2. Role of Gelatinase B in Cartilage Breakdown The presence of gelatinolytic activity in the synovial fluids and synovial tissues of RA patients suggests a role for this enzyme in the disease. Gelatinase B knockout mice are resistant to anti-collagen II antibody-induced arthritis (Itoh et al., 2002). This is in sharp contrast to gelatinase A knock-out mice, which are more susceptible for arthritis development. The latter finding may be explained by antiinflammatory activities of gelatinase A, in particular the cleavage of monocyte chemotactic protein-3. This cleavage of MCP-3 aborts its potential to induce signal transduction through its receptor, but does not impair receptor binding and therefore converts the chemokine in an antagonist (McQuibban et al., 2000). In contrast, gelatinase B processes the neutrophil chemoattractant interleukin-8 into a more than 10-fold more potent variant, inducing a positive feedback loop which fuels inflammation (Van den Steen et al., 2000). These studies show the importance of the cleavage of regulatory molecules by MMPs in acute and chronic inflammation (Opdenakker et al., 2001).

Gelatinase B Participation in Collagen II Degradation

49

Figure 3. Combined action of neutrophil collagenase and gelatinase B on native collagen II Collagenases, and in particular neutrophil collagenase (MMP-8), efficiently cleave native collagen II at a single site, resulting in the formation of 3 and 1 fragments. When both collagenase and gelatinase B are 4 4 added to native collagen II, the collagen is completely degraded into small fragments, as is visualised on this SDS-PAGE analysis. The positions of uncleaved native collagen II and the 3 fragment of collagen II 4 are indicated.

However, the most efficient substrate of gelatinase B is gelatin, which is heatdenatured collagen. Triple helical collagen II cannot be cleaved by gelatinase B, because the peptide bonds are at the inner side of the triple helix and not accessible for most proteases, including gelatinase B. However, collagenases can unwind the triple helix locally and subsequently cleave the peptide chain at a single site, generating the so-called 3 and 1 fragments. Neutrophil collagenase, which is usually secreted 4 4 together with gelatinase B from the granules of neutrophils, is particularly efficient in the cleavage of collagen type II (Jeffrey, ff 1998). Although this cleavage does not result in complete denaturation of the triple helix (Marini et al., 2000), the 3 and 1 4 4 fragments are susceptible to cleavage by gelatinase B (Fig. 3). The two fragments are degraded by gelatinase B into small peptides, and these were analysed after separation of the fragments by RP-HPLC (Van den Steen et al., 2002b; Van den Steen et al., 2004). To identify each peptide in each fraction, all fractions were subjected to Edman degradation and mass spectrometry. Edman degradation yields the aminoterminal sequence, and mass spectrometry defines the exact size and thus the carboxyterminus of the peptides (figure 4). In this way, 25 and 30 different ff cleavage sites were identified in bovine and human denatured collagen II, respectively. The finding and definition of the reaction products confirm that gelatinase B participates in the complete degradation of collagen II into small peptides (Table 1). Two main immunodominant T-cell epitopes occur in collagen II (Brand et al., 1994; Myers et al., 1995; Rosloniec et al., 1996; Andersson et al., 1998). The position of the cleavage sites, relative to the known immunodominant autoreactive T-cell epitopes, suggests that the degradation of collagen II by the combined action of collagenase and gelatinase B not only leads to the degradation of the cartilage, but also releases intact immunodominant epitopes. These remnant epitopes may even lay at the basis of autoimmunity, since they may be loaded onto MHC-II of antigen

P. E. Van den Steen et al.

50

Table 1. Comparison of the cleavage sites by gelatinase B in natural human type II collagen P6

P5

P4

P3

P2

P1

P1∞

P2∞

P3∞

P4∞

P5∞

P6∞

A P P R P A S P A P E L L P L A L S L A P E P P A K P L P P L

POH P* M T A R POH L POH I R A POH P(OH) P(OH) POH V POH Q Q P* T POH Q Q D P* A V A K

G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G

P P P P A P P P P P P P A P A P P A P P P P T P P P P Q P A H

Q Q R A A E A K A POH S K(OHex) R Q P(OH) S R Q R P* A POH S T POH K Q R P* R R

G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G

F 42 A 93 L 147 A 159 A 162 A 207 A 225 Q 273 E 306 E 327 L 357 A 363 L 381 A 414 L 447 F 483 E 519 L 534 L 540 L 567 A 618 T 651 F 654 V 699 A 714 A 750 L 795 I 801 L 846 I 927 F 954

Q R P(OH) A R Q S T E R A N T R R Q R Q P* Q N S A T T R A V T Q T

G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G

N F E A N P N E K A P D R Q L L F P T M E F P P F D Q L P P L

POH POH R R D R P* POH R POH K(OHex) P(OH) POH POH POH POH POH R P POH K(OH) A POH K(OHex) POH S R POH A Q Q

G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G G

Residues showing considerable consensus are shown in bold. At P1∞, hydrophobic residues are indicated in italic and the position in the sequence is indicated in subscript. POH, hydroxyproline; KOHex, glycosylated hydroxylysine, P*, Pro with probable but uncertain hydroxylation.

presenting cells. Loading of the remnant epitopes can occur either after internalisation and further processing, or at the outside of the antigen presenting cell, since empty MHC-II is expressed on the cell surface, together with HLA-DM, which is known to catalyze loading of peptides in MHC-II (Santambrogio et al., 1999; Arndt et al., 2000). Inflammatory stimuli in the joint will therefore induce degradation of collagen II by collagenases and gelatinase B, and additionally activate the antigen presenting cells to upregulate costimulatory molecules, increasing their capacity to activate T-cells (Inaba et al., 2000). Furthermore, latent autoreactive T-cells are present in many healthy individuals (Sun et al., 1991). This may mean that, with a susceptible genetic background and with the presence of autoreactive T-cells, an inflammation in the joint may induce rheumatoid arthritis. This theory was originally formulated in the context of multiple sclerosis and was named ‘‘remnant epitopes generate autoimmunity’’ or REGA model (Opdenakker and Van Damme, 1994). The REGA

Gelatinase B Participation in Collagen II Degradation

51

Figure 4. Use of Edman degradation and mass spectrometry to analyse the gelatinase B cleavage sites and posttranslational modifications in collagen II Denatured bovine and human collagen type II were digested with natural human gelatinase B. The resulting fragments were separated by RP-HPLC, and the resulting fractions were analysed by Edman degradation and mass spectrometry. The combination of these technologies enabled to determine 30 different ff gelatinase B cleavage sites and to localize a large number of posttranslational modifications in collagen type II.

model is further supported by the study of Pu et al., who showed the existence of T-cells which react only with epitopes that are destroyed after intracellular processing of the antigen, suggesting that extracellular processing and loading on MHC-II must occur (Pu et al., 2002).

3. POSTTRANSLATIONAL MODIFICATIONS OF HUMAN COLLAGEN II The use of both mass spectrometry and Edman degradation for the determination of the gelatinase B cleavage sites in type II collagen also allowed the determination of a large number of posttranslational modifications in the sequence (Fig. 4). These include proline-hydroxylation, lysine-hydroxylation and glycosylation of

52

P. E. Van den Steen et al.

hydroxylysine. With the use of the outlined procedure, 9 different ff glycosylation sites were determined, of which one is in the major immunodominant epitope and one in its close vicinity (Van den Steen et al., 2004; Van den Steen et al., 2002b). This has major implications for the immunogenicity of this immunodominant epitope, as it is conceivable that sugars on a peptide will modify its recognition by the immune system. This was shown by several groups. Glycosylation of an antigenic peptide can influence its binding in the groove of MHC-I or MHC-II, as such a sugar may make extensive contacts within the groove or also provide steric hindrance to the binding (Haurum et al., 1995). Another possibility is that the sugar points out of the groove, consequently modifying the interaction with the T-cell receptor and even leading to activation of different ff T-cell clones (Haurum et al., 1994; Corthay et al., 1998). In this context, it is interesting to note that glycosylated collagen II is more potent to elicit arthritis than a chemically deglycosylated variant (Michaelsson et al., 1994) or a recombinant underglycosylated form of collagen II (Myers et al., 2004). Furthermore, autoreactive T-cells from patients with rheumatoid arthritis react preferentially with a synthetic glycosylated form of the immunodominant epitope than with the unglycosylated peptide (Backlund et al., 2002). Therefore, the formal proof that a glycan is present on the immunodominant epitope of natural human collagen II (Van den Steen et al., 2004) is significant for the further understanding of the autoimmune process in rheumatoid arthritis patients. Furthermore, the known immunodominant epitopes in collagen II have been detected by the technique of epitope scanning with unmodified synthetic peptides. It is therefore conceivable that other, modified immunodominant epitopes are present in collagen II. This notion is further exemplified by the finding that, after RP-HPLC separation of the degradation products of collagen II after gelatinase B cleavage, some fractions without known immunodominant epitopes induce T-cell reactivity (Van den Steen et al., 2004).

Figure 5. Generation of glycosylated remnant epitopes from collagen II by neutrophil collagenase and gelatinase B Triple helical collagen is highly resistant to proteolysis. Only collagenases, e.g. neutrophil collagenase/MMP-8, are able to locally unwind the triple helix and to perform a single cleavage. Thereafter, the collagen becomes susceptible to degradation by gelatinase B. The resulting peptides, some of which contain oligosaccharides, can subsequently be loaded onto MHC-II of antigen presenting cells and be presented to autoreactive T-cells.

Gelatinase B Participation in Collagen II Degradation

53

4. CONCLUSIONS Inflammation of the joint, as seen in rheumatoid arthritis, is associated with the expression of MMPs, in particular collagenases and gelatinase B. Collagenases such as MMP-8 cleave the triple helical collagen type II, which is one of the main components of joint cartilage, at a single position. As a consequence, collagen II becomes sensitive for cleavage by gelatinase B and is fragmented into small peptides that still contain the immunodominant epitopes. These auto-antigenic peptides, named remnant peptides, may be loaded onto MHC-II and may be presented by antigen presenting cells to autoreactive T-cells (Fig. 5). This leads to the activation of these T-cells, resulting in the formation of autoantibodies, which can activate the complement cascade, further enhancing the inflammation. Furthermore, some of the remnant epitopes are modified by hydroxylation and glycosylation, which increases their immunogenicity.

ACKNOWLEDGEMENTS P.E. Van den Steen is a postdoctoral fellow of the Belgian National Fund for Scientific Research (F.W.O.-Vlaanderen). Supported by the Geconcerteerde OnderzoeksActies (GOA 2002–2006) and by the F.W.O.-Vlaanderen.

REFERENCES Ahrens, D., Koch, A.E., Pope, R.M., Stein-Picarella, M., and Niedbala, M.J. (1996). Expression of matrix metalloproteinase 9 (96-kd gelatinase B) in human rheumatoid arthritis. Arthritis Rheum. 39, 1576–1587. Akahoshi, T., Endo, H., Kondo, H., Kashiwazaki, S., Kasahara, T., Mukaida, N., Harada, A., and Matsushima, K. (1994). Essential involvement of interleukin-8 in neutrophil recruitment in rabbits with acute experimental arthritis induced by lipopolysaccharide and interleukin-1. Lymphokine Cytokine. Res. 13, 113–116. Andersson, E.C., Hansen, B.E., Jacobsen, H., Madsen, L.S., Andersen, C.B., Engberg, J., Rothbard, J.B., McDevitt, G.S., Malmstrom, V., Holmdahl, R., Svejgaard, A., and Fugger, L. (1998). Definition of MHC and T cell receptor contacts in the HLA-DR4 restricted immunodominant epitope in type II collagen and characterization of collagen-induced arthritis in HLA-DR4 and human CD4 transgenic mice. Proc. Natl. Acad. Sci. U.S.A. 95, 7574–7579. Arndt, S.O., Vogt, A.B., Markovic-Plese, S., Martin, R., Moldenhauer, G., Wolpl, A., Sun, Y., Schadendorf, D., Hammerling, G.J., and Kropshofer, H. (2000). Functional HLA-DM on the surface of B cells and immature dendritic cells. EMBO J. 19, 1241–1251. Backlund, J., Carlsen, S., Hoger, T., Holm, B., Fugger, L., Kihlberg, J., Burkhardt, H., and Holmdahl, R. (2002). Predominant selection of T cells specific for the glycosylated collagen type II epitope (263–270) in humanized transgenic mice and in rheumatoid arthritis. Proc. Natl. Acad. Sci. U.S.A. 99, 9960–9965. Brand, D.D., Myers, L.K., Terato, K., Whittington, K.B., Stuart, J.M., Kang, A.H., and Rosloniec, E.F. (1994). Characterization of the T cell determinants in the induction of autoimmune arthritis by bovine a1(II)-CB11 in H-2q mice. J. Immunol. 152, 3088–3097. Corthay, A., Backlund, J., Broddefalk, J., Michaelsson, E., Goldschmidt, T.J., Kihlberg, J., and Holmdahl, R. (1998). Epitope glycosylation plays a critical role for T cell recognition of type II collagen in collagen-induced arthritis. Eur. J. Immunol. 28, 2580–2590. Gruber, B.L., Sorbi, D., French, D.L., Marchese, M.J., Nuovo, G.J., Kew, R.R., and Arbeit, L.A. (1996). Markedly elevated serum MMP-9 (gelatinase B) levels in rheumatoid arthritis: a potentially useful laboratory marker. Clin. Immunol. Immunopathol. 78, 161–171.

54

P. E. Van den Steen et al.

Haurum, J.S., Arsequell, G., Lellouch, A.C., Wong, S.Y., Dwek, R.A., McMichael, A.J., and Elliott, T. (1994). Recognition of carbohydrate by major histocompatibility complex class I-restricted, glycopeptide-specific cytotoxic T lymphocytes. J. Exp. Med. 180, 739–744. Haurum, J.S., Tan, L., Arsequell, G., Frodsham, P., Lellouch, A.C., Moss, P.A., Dwek, R.A., McMichael, ff on class I major histocompatA.J., and Elliott, T. (1995). Peptide anchor residue glycosylation: effect ibility complex binding and cytotoxic T lymphocyte recognition. Eur. J. Immunol. 25, 3270–3276. Hibbs, M.S., Hoidal, J.R., and Kang, A.H. (1987). Expression of a metalloproteinase that degrades native type V collagen and denatured collagens by cultured human alveolar macrophages. J. Clin. Invest. 80, 1644–1650. Inaba, K., Turley, S., Iyoda, T., Yamaide, F., Shimoyama, S., Reis, Germain, R.N., Mellman, I., and Steinman, R.M. (2000). The formation of immunogenic major histocompatibility complex class II-peptide ligands in lysosomal compartments of dendritic cells is regulated by inflammatory stimuli. J. Exp. Med. 191, 927–936. Itoh, T., Matsuda, H., Tanioka, M., Kuwabara, K., Itohara, S., and Suzuki, R. (2002). The role of matrix metalloproteinase-2 and matrix metalloproteinase-9 in antibody-induced arthritis. J. Immunol. 169, 2643–2647. Jeffrey, ff J.J. (1998). Interstitial collagenases. In Matrix Metalloproteinases, W.C.Parks and R.P.Mecham, eds. Academic Press, pp. 15–42. Kjeldsen, L., Johnsen, A.H., Sengelov, H., and Borregaard, N. (1993). Isolation and primary structure of NGAL, a novel protein associated with human neutrophil gelatinase. J. Biol. Chem. 268, 10425–10432. Marini, S., Fasciglione, G.F., De Sanctis, G., D’Alessio, S., Politi, V., and Coletta, M. (2000). Cleavage of bovine collagen I by neutrophil collagenase MMP-8. Effect ff of pH on the catalytic properties as compared to synthetic substrates. J. Biol. Chem. 275, 18657–18663. Masure, S., Proost, P., Van Damme, J., and Opdenakker, G. (1991). Purification and identification of 91-kDa neutrophil gelatinase. Release by the activating peptide interleukin-8. Eur. J. Biochem. 198, 391–398. Mattu, T.S., Royle, L., Langridge, J., Wormald, M.R., Van den Steen, P.E., Van Damme, J., Opdenakker, G., Harvey, D.J., Dwek, R.A., and Rudd, P.M. (2000). O-Glycan analysis of natural human neutrophil gelatinase B using a combination of normal phase- HPLC and online tandem mass spectrometry: implications for the domain organization of the enzyme. Biochemistry 39, 15695–15704. McQuibban, G.A., Gong, J.H., Tam, E.M., McCulloch, C.A., Clark-Lewis, I., and Overall, C.M. (2000). Inflammation dampened by gelatinase A cleavage of monocyte chemoattractant protein-3. Science 289, 1202–1206. Michaelsson, E., Malmstrom, V., Reis, S., Engstrom, A., Burkhardt, H., and Holmdahl, R. (1994). T cell recognition of carbohydrates on type II collagen. J. Exp. Med. 180, 745–749. Murphy, G. and Willenbrock, F. (1995). Tissue inhibitors of matrix metalloendopeptidases. Methods Enzymol. 248, 496–510. Myers, L.K., Miyahara, H., Terato, K., Seyer, J.M., Stuart, J.M., and Kang, A.H. (1995). Collagen-induced arthritis in B10.RIII mice (H-2r): identification of an arthritogenic T-cell determinant. Immunology 84, 509–513. Myers, L.K., Myllyharju, J., Nokelainen, M., Brand, D.D., Cremer, M.A., Stuart, J.M., Bodo, M., Kivirikko, K.I., and Kang, A.H. (2004). Relevance of posttranslational modifications for the arthritogenicity of type II collagen. J. Immunol. 172, 2970–2975. Nagase, H. and Woessner, J.F. (1999). Matrix metalloproteinases. J. Biol. Chem. 274, 21491–21494. Nguyen, Q., Murphy, G., Hughes, C.E., Mort, J.S., and Roughley, P.J. (1993). Matrix metalloproteinases cleave at two distinct sites on human cartilage link protein. Biochem. J. 295, 595–598. Ogata, Y., Enghild, J.J., and Nagase, H. (1992). Matrix metalloproteinase 3 (stromelysin) activates the precursor for the human matrix metalloproteinase 9. J. Biol. Chem. 267, 3581–3584. Opdenakker, G., Masure, S., Grillet, B., and Van Damme J. (1991). Cytokine-mediated regulation of human leukocyte gelatinases and role in arthritis. Lymphokine Cytokine. Res. 10, 317–324. Opdenakker, G. and Van Damme, J. (1994). Cytokine-regulated proteases in autoimmune diseases. Immunol. Today 15, 103–107. Opdenakker, G., Van den Steen, P.E., and Van Damme, J. (2001). Gelatinase B: a tuner and amplifier of immune functions. Trends. Immunol. 22, 571–579. Peichl, P., Ceska, M., Effenberger, ff F., Haberhauer, G., Broell, H., and Lindley, I.J. (1991). Presence of NAP-1/IL-8 in synovial fluids indicates a possible pathogenic role in rheumatoid arthritis. Scand. J. Immunol. 34, 333–339.

Gelatinase B Participation in Collagen II Degradation

55

Peppin, G.J. and Weiss, S.J. (1986). Activation of the endogenous metalloproteinase, gelatinase, by triggered human neutrophils. Proc. Natl. Acad. Sci. U.S.A. 83, 4322–4326. Pu, Z., Carrero, J.A., and Unanue, E.R. (2002). Distinct recognition by two subsets of T cells of an MHC class II-peptide complex. Proc. Natl. Acad. Sci. U.S.A. 99, 8844–8849. Rampart, M., Herman, A.G., Grillet, B., Opdenakker, G., and Van Damme, J. (1992). Development and application of a radioimmunoassay for interleukin-8: detection of interleukin-8 in synovial fluids from patients with inflammatory joint disease. Lab. Invest. 66, 512–518. Ribbens, C., Andre, B., Jaspar, J.M., Kaye, O., Kaiser, M.J., De Groote, D., and Malaise, M.G. (2000). Matrix metalloproteinase-3 serum levels are correlated with disease activity and predict clinical response in rheumatoid arthritis. J. Rheumatol. 27, 888–893. Rosloniec, E.F., Whittington, K.B., Brand, D.D., Myers, L.K., and Stuart, J.M. (1996). Identification of MHC class II and TCR binding residues in the type II collagen immunodominant determinant mediating collagen-induced arthritis. Cell Immunol. 172, 21–28. Santambrogio, L., Sato, A.K., Carven, G.J., Belyanskaya, S.L., Strominger, J.L., and Stern, L.J. (1999). Extracellular antigen processing and presentation by immature dendritic cells. Proc. Natl. Acad. Sci. U.S.A. 96, 15056–15061. Seitz, M., Dewald, B., Ceska, M., Gerber, N., and Baggiolini, M. (1992). Interleukin-8 in inflammatory rheumatic diseases: synovial fluid levels, relation to rheumatoid factors, production by mononuclear cells, and effects ff of gold sodium thiomalate and methotrexate. Rheumatol. Int. 12, 159–164. Senior, R.M., Griffin, G.L., Fliszar, C.J., Shapiro, S.D., Goldberg, G.I., and Welgus, H.G. (1991). Human 92– and 72-kilodalton type IV collagenases are elastases. J. Biol. Chem. 266, 7870–7875. Sopata, I., Wize, J., Filipowicz-Sosnowska, A., Stanislawska-Biernat, E., Brzezinska, B., and Maslinski, S. (1995). Neutrophil gelatinase levels in plasma and synovial fluid of patients with rheumatic diseases. Rheumatol. Int. 15, 9–14. Sun, J.B., Olsson, T., Wang, W.Z., Xiao, B.G., Kostulas, V., Fredrikson, S., Ekre, H.P., and Link, H. (1991). Autoreactive T and B cells responding to myelin proteolipid protein in multiple sclerosis and controls. Eur. J. Immunol. 21, 1461–1468. Triebel, S., Blaser, J., Reinke, H., and Tschesche, H. (1992). A 25 kDa alpha 2-microglobulin-related protein is a component of the 125 kDa form of human gelatinase. FEBS Lett. 314, 386–388. Van den Steen, P., Rudd, P.M., Dwek, R.A., and Opdenakker, G. (1998). Concepts and principles of O-linked glycosylation. Crit. Rev. Biochem. Mol. Biol. 33, 151–208. Van den Steen, P.E., Dubois, B., Nelissen, I., Rudd, P.M., Dwek, R.A., and Opdenakker, G. (2002a). Biochemistry and molecular biology of gelatinase B or matrix metalloproteinase-9 (MMP-9). Crit Rev. Biochem. Mol. Biol. 37, 375–536. Van den Steen, P.E., Opdenakker, G., Wormald, M.R., Dwek, R.A., and Rudd, P.M. (2001). Matrix remodelling enzymes, the protease cascade and glycosylation. Biochim. Biophys. Acta 1528, 61–73. Van den Steen, P.E., Proost, P., Brand, D.D., Kang, A.H., Van Damme, J., and Opdenakker, G. (2004). Generation of Glycosylated Remnant Epitopes from Human Collagen Type II by Gelatinase B. Biochemistry 43, 10809–10816. Van den Steen, P.E., Proost, P., Grillet, B., Brand, D.D., Kang, A.H., Van Damme, J., and Opdenakker, G. (2002b). Cleavage of denatured natural collagen type II by neutrophil gelatinase B reveals enzyme specificity, post-translational modifications in the substrate, and the formation of remnant epitopes in rheumatoid arthritis. FASEB J. 16, 379–389. Van den Steen, P.E., Proost, P., Wuyts, A., Van Damme, J., and Opdenakker, G. (2000). Neutrophil gelatinase B potentiates interleukin-8 tenfold by aminoterminal processing, whereas it degrades CTAP-III, PF-4, and GRO-alpha and leaves RANTES and MCP-2 intact. Blood 96, 2673–2681.

11

HYALURONAN IN IMMUNE PROCESSES

Alan J. Wright and Anthony J. Day MRC Immunochemistry Unit Department of Biochemistry University of Oxford South Parks Road, Oxford OX1 3QU, UK

1. INTRODUCTION The linear polysaccharide hyaluronan (HA) is a glycosaminoglycan consisting entirely of a repeating disaccharide of glucuronic acid and N-acetyl glucosamine (see Fig. 1), which unlike other glycosaminoglycans (e.g., heparin and chondroitin sulphate) is neither sulphated, nor attached to a protein core (Day and Sheehan, 2001; Fraser et al., 1997; Tammi et al., 2002). While HA molecules are usually of very high molecular weight (i.e., 105 to 107 Da), smaller fragments and oligosaccharides of HA have also been detected under certain physiological or pathological conditions. HA is highly polar existing as a polyanion and, in aqueous solutions (such as synovial fluid), its large hydrodynamic volume and viscoelastic properties give rise to its important space filling, filtering and lubricating functions (Hardingham, 2004; Laurent et al., 1996). It is also a vital structural component of extracellular matrix, where it has diverse roles in development, wound healing,

Figure 1. The structure of a trisaccharide of hyaluronan showing the alternating b(1–4) and b(1–3) linkages. 57 John S. Axford (ed.), Glycobiology and Medicine, 57-69. © 2005 Springer. Printed in the Netherlands.

58

A. J. Wright and A. J. Day

ovulation and the immune response (Blundell et al., 2004b; Tammi et al., 2002;). For example, at sites of inflammation HA is involved in the preliminary stages of leukocyte adhesion and transendothelial migration. Given the simplicity of HA’s chemical structure it might seem, at first glance, surprising that it is implicated in so many different ff processes. However, there is increasing evidence that this diversity results from the interaction of HA with specific HA-binding proteins (known as hyaladherins) to form HA-protein complexes with distinct architectures and functional activities (see Blundell et al., 2004b). In this regard, it has been proposed that individual hyaladherins are able to capture and propagate particular conformations of the polysaccharide, leading to the formation of different ff higher order structures (Day and Sheehan, 2001); in the absence of proteins, solution HA is thought to correspond to a stiffened ff random coil, existing as a highly dynamic ensemble of chaotically interchanging, semi-ordered states (reviewed in Blundell et al., 2004a). The repeating nature of HA (and its chemical fidelity), coupled with its conformational repertoire, makes it a perfect scaffold ff to form periodic protein arrays, especially where cooperative interactions lead to clustering of hyaladherins along the HA chain (Day and Sheehan, 2001); complexes of this type are well known components of cartilage matrix. Furthermore, the crosslinking of multiple HA strands, via protein-protein interactions, can result in huge cable-like structures, which have been shown to form during inflammation and have distinctive leukocyte-binding properties (de la Motte et al., 2003, 2004; Majors et al., 2003). The majority of hyaladherins are members of the Link module superfamily (Blundell et al., 2004b; Day and Prestwich, 2002) containing a common protein domain of ~100 amino acids that mediates the interaction with HA. To date threedimensional structures have only been determined for the HA-binding domains of two members of this superfamily; i.e., TSG-6 (tumor necrosis factor-stimulated gene-6) and CD44. The interaction domain of TSG-6, an inflammation-associated protein (see Milner and Day, 2003), consists of a single independently folded Link module that alone can support high affinity HA binding (Blundell et al., 2003; Kohda et al., 1996). In the case of CD44, a cell surface receptor for HA, additional N- and C-terminal sequences flanking the Link module are necessary to form a stable fold, giving rise to an HA-binding domain of ~150 amino acids (Teriete et al., 2004). This review will focus on CD44 and TSG-6 describing their involvement in immune processes, in particular their roles in the HA-mediated interactions of leukocytes with the vascular endothelium at sites of inflammation.

2. CD44 CD44 is ubiquitously expressed on leukocytes, and many other cell types (including tumour cells), where it represents the major cell surface receptor for hyaluronan (Arrufo et al., 1990; Lesley et al., 1997; Pure´ and Cuff, ff 2001; Teriete et al., 2004; Toole, 2004). Although CD44 constitutes a family of hyaluronan receptors varying in size from ~80 to 250 kDa arising from alternate splicing (Pure and Cuff, ff 2001), the major species on leukocytes is the standard or hematopoetic form (CD44H) that does not utilise variant exons. As can be seen in Fig. 2, CD44 consists of an N-terminal extracellular HA-binding domain (highly conserved in mammals with

Hyaluronan in Immune Processes

59

Figure 2. The domain structure of CD44.

~85% amino acid identity; Isacke and Yarwood, 2002), a membrane proximal extracellular domain, a transmembrane domain and a C-terminal cytoplasmic tail. As part of the immune response, circulating leukocytes become localised at sites of inflammation via their adhesion to, and rolling along, the vascular endothelium; this allows sampling of the local environment potentially leading to tight adhesion and transendothelial migration (Butcher, 1991; Kubes and Kerfoot, 2001; Siegelman et al., 2000). In the case of lymphocytes and monocytes the initial tethering and rolling can be mediated by CD44 (on their cell surfaces) interacting with HA immobilised on the blood vessel wall (DeGrendele et al., 1996; Mohamadzadeh et al., 1998). Importantly, CD44 on circulating leukocytes does not bind HA constitutively, but can be induced to do so in response to inflammatory signals (Lesley et al., 1993; 1997), which also lead to increased expression of HA by endothelial cells; tight regulation of these processes is essential to avoid host tissue damage due to inappropriate extravasation (e.g., as seen in autoimmune disease; Siegelman et al., 1999). The induction of CD44’s HA-binding activity can be mediated in a number of different ff ways. For instance, activated leukocytes change their cell surface CD44 to a HA-binding form after stimulation by cytokines, such as tumour necrosis factor and interferon-c, or through T cell receptor triggering (DeGrendele et al., 1997; Pure´ and Cuff, ff 2001). Alternatively, CD44 can be induced to bind HA by factors that interact directly with the receptor; e.g., there are certain monoclonal antibodies against the CD44_HABD that can induce it to bind HA (Lesley et al., 1993; Zheng et al., 1995). As discussed in detail below our recent studies on CD44 and TSG-6 have provided novel insights into the molecular basis of CD44 regulation.

3. ANALYSIS OF THE HA-BINDING SURFACE OF CD44 The N-terminal HA-binding domain of CD44 (CD44_HABD) contains a Link module, together with flanking N- and C- terminal sequences that are necessary to

60

A. J. Wright and A. J. Day

Figure 3. Leukocytes can adhere to and roll along the blood vessel endothelium via CD44-HA interactions.

Figure 4. Cartoons showing the backbone folds of (A) the Link module of TSG-6 (Link_TSG6) and (B) the ~150 amino acid CD44 HA-binding domain (CD44_HABD). In CD44 the N- and C-terminal flanking sequences (coloured black) extend the b sheet structure of the Link module.

form a stable fold (Peach et al., 1993; Banerji et al., 1998). The recent determination of the 3D structure of the CD44_HABD, by both X-ray crystallography and NMR (Teriete T et al., 2004), has revealed that the N- and C- terminal extensions, which are linked by a disulphide bond, form an extra lobe of structure in intimate contact with one side of the Link module (Fig. 4). This structure has given new information on how CD44 may bind HA and how this interaction may be regulated by N-linked glycosylation (Teriete et al., 2004). Site-directed mutagenesis of CD44 (Peach et al., 1993; Bajorath et al. 1998)

Hyaluronan in Immune Processes

61

revealed that four residues in the Link module were essential for HA binding (i.e., Arg41, Tyr42, Arg78 and Tyr79), where mutation of any one of these caused complete loss of functional activity. In addition, a further ten ‘important’ residues were implicated in HA recognition, with five located in the Link module (Lys38, Lys68, Asn100, Asn101 and Tyr105) and five in the extra lobe formed from the N- (Arg29) and C-terminal extensions (Arg150, Arg154, Lys158 and Arg162). Mapping these amino acids onto the CD44_HABD structure shows that the four key binding residues form a continuous patch on one face of the molecule (which also includes Lys38), and that Asn100, Asn101 and Tyr105 are clustered close by, extending the binding site to one side (Teriete et al., 2004). The basic amino acids in the C-terminal extension are also on this face of the CD44_HABD, but are widely spaced, and Arg29 and Lys68 are on the opposite side of the structure (see Fig. 5). NMR studies on CD44 have determined that almost an entire face of the HABD is highly perturbed on its interaction with HA; the chemical shift changes are centred round Arg41 and Tyr42, and overlap appreciably with the ligand-binding residues determined experimentally (Teriete et al., 2004). However, these perturbations are too widespread to all be caused solely by the direct interaction of a single HA molecule and a more likely explanation is that the protein undergoes a conformational change on binding, which is consistent with the observed perturbation of the hydrogen bond network (Teriete et al., 2004). The widespread distribution of amino acids implicated by mutagenesis is also difficult to reconcile with the recognition of a single HA molecule. Therefore, it has been suggested (Teriete et al., 2004) that CD44 could accommodate HA in two different, ff mutually exclusive, binding positions or modes (see Fig. 5), where mode 1 is likely to represent the HA-binding site conserved across the Link module superfamily (Blundell et al., 2003). While this hypothesis does not explain the involvement of Arg29 or Lys68 in the interaction with HA, it seems unlikely that, given their locations, they could play a direct role in binding. In this regard, it cannot be ruled out that the mutation of these, or other residues implicated as functional, might have a deleterious effect ff on the formation of an optimal binding conformation.

4. REGULATION OF CD44 FUNCTION BY N-GLYCOSYLATION N As mentioned above, CD44, although present on the surface of circulating leukocytes, does not bind HA constitutively. However, during inflammation, active CD44 molecules are formed that can mediate cell attachment and rolling (reviewed in Lesley et al., 2004; Teriete et al., 2004). While there are several possible ways in which this could occur, one of the major mechanisms of receptor activation has been shown to involve remodelling of N-linked glycans on CD44. Importantly, the level of cell surface N-glycosylation dictates the activation state of CD44, and the enzymatic removal of sialic acid (e.g., by an endogenous sialidase; Gee et al., 2003; Katoh et al., 1999), or inhibition of N-glycan biosynthesis, induces HA binding (English et al., 1998; Katoh et al., 1995; Lesley et al., 1995). In this regard, there are five Nlinked glycosylation sites in the HABD of human and mouse CD44 (residues 25, 57, 100, 110, 120 in the former), and the individual mutation of two of these (i.e., Asn25

62

A. J. Wright and A. J. Day

Figure 5. A) A surface representation of the CD44_HABD (in two orientations; one 180° around a vertical axis from the other) showing the relative positions of the essential binding residues (dark grey) and those implicated as functionally important (lighter grey). B) The widespread nature of the amino acids implicated ff modes of HA binding, which are represented by from mutagenesis could be explained by two different arrows; the HA is shown such that the arrowheads are at its non-reducing termini, which is the polarity determined previously for HA binding to Link_TSG6 (Blundell et al., 2003). The CD44_HABD in (B) is shown in the same orientation as the left-hand molecule in (A).

or Asn120) converts CD44 from an inducible to constitutively active state on lymphocyte cell lines (English et al., 1998). The recent elucidation of the tertiary structure of the HA-binding domain has allowed the position of these key regulatory glycosylation sites to be determined (Teriete et al., 2004). As can be seen from Fig. 5, Asn25 is present on the HA-binding face of CD44 in close proximity to the mode 1 interaction site. It is not difficult to imagine, therefore, how the presence of glycosylation at this location could sterically block ligand recognition (Fig. 6). In the case of Asn120, it is less obvious how the attachment of an N-glycan could inhibit binding since it is located on the opposite face of the protein. The most likely explanation is that glycosylation at Asn120 has the potential to interfere with CD44 self-association and so prevent receptor clustering (see Fig. 6), a process that is known to be

Hyaluronan in Immune Processes

63

Figure 6. A schematic diagram showing the effects ff of N-linked glycosylation on HA-CD44 interactions. The presence of certain N-glycans at Asn25 and Asn120 may have an inhibitory effect ff on HA binding by CD44+ cells and their removal (or remodelling) could either unlock the binding groove or facilitate receptor clustering, respectively.

important in CD44 regulation. For example, it is well established that cross-linking of CD44 by certain antibodies (such as IRAWB14) can trigger large increases in its HA-binding activity on leukocyte (Lesley et al., 1993; 2000). Interestingly, the epitope for IRAWB14 (Zheng et al., 1995) lies immediately adjacent to Asn120 in the CD44 structure and is centred on Lys68, one of the two residues on the rear face of the HADB whose mutation affects ff HA binding (Bajorath et al., 1998). It seems likely, therefore, that this face of CD44 plays an important role in the interaction with HA, but further work is required to determine its contribution to CD44 oligomerisation and exactly how this is modulated by changes in glycosylation. Conflicting reports on the effect ff of glycosylation on CD44-HA interactions (Bartolazi et al., 1996; English et al., 1998; Zheng et al. 1997) indicates there is considerable complexity to be understood, much of which is likely to result from the tissue and cell specific nature of CD44 biology.

5. TSG-6 TSG-6 is a 35 kDa HA-binding protein that is not generally present in healthy adult tissues, but is expressed in response to inflammatory cytokines and certain

64

A. J. Wright and A. J. Day

growth factors (Lee et al., 1992; Milner and Day, 2003; Wisniewski and Vilcek, 2004). Although TSG-6 has been found to be associated with diseases such as arthritis, it is likely to have an anti-inflammatory role, acting as part of a negative feedback loop to resolve inflammation (Getting et al., 2002). For example, TSG-6 is a potent down regulator of cytokine-induced neutrophil migration, via its inhibition of polymorphonuclear-endothelial cell interactions (Cao et al., 2004). In this regard, deletion of the TSG-6 gene leads to increased extravasation of neutrophils into synovial tissues in murine models of arthritis causing extensive joint destruction, whereas overexpression of TSG-6 has a chondroprotective effect ff (see Szanto´ et al., 2004). However, it should be noted that the inhibition of neutrophil migration by TSG-6 is likely to be independent of its HA-binding function (Getting et al., 2002). TSG-6 is also expressed in the context of inflammation-like processes such as ovulation, where it serves to stabilise the HA-rich extracellular matrix that forms around the oocyte during cumulus matrix expansion via its interaction with intera-inhibitor and PTX3 (Day et al., 2004; Fulop et al., 2003; Salustri et al., 2004). HA cross-linking of this type is likely to occur at inflammatory sites (e.g., in articular joint disease), where the HA/protein complexes formed could potentially have altered leukocyte-binding properties (see Day et al., 2004). The HA-binding domain of TSG-6 consists of a single Link module (denoted Link_TSG6) that has been extensively characterized by NMR spectroscopy and sitedirected mutagenesis (Blundell et al., 2003; Getting et al., 2002; Kahmann et al., 2000; Kohda et al., 1996; Mahoney et al., 2001). For instance, we have recently determined the solution structures of its free and HA-bound forms, revealing that a conformational change occurs in the Link module on its interaction with HA (Blundell et al., 2003). This exposes a HA-binding groove, lined with the key functional amino acids implicated by mutagenesis studies (Getting et al., 2002; Mahoney et al., 2001), which is in a similar position to the proposed mode 1 interaction site in CD44 (Teriete et al., 2004). Although the locations of the HA-interaction surfaces are likely to be conserved in TSG-6 and CD44, there appear to be some major differences ff in the details of the residues and sequence positions involved in mediating HA binding in the two proteins (Mahoney et al., 2001). It is likely therefore, that the interaction networks will be distinct in these HA-protein complexes, potentially involving the stabilisation of HA in different ff bound conformations.

6. TSG-6 ENHANCES HYALURONAN BINDING TO CELL SURFACE CD44 Given that TSG-6 is a HA-binding protein, which is expressed by a wide range of cell types in response to inflammatory stimuli (e.g., monocytes, macrophages, neutrophils, dendritic cells, microvascular endothelium, vascular smooth muscle cells and fibroblasts (Lesley et al., 2004; Milner and Day, 2003)), it was thought possible that it may act as a competitive inhibitor of CD44-mediated leukocyte migration (Lee et al., 1992). This hypothesis has recently been tested (Lesley et al., 2004) and it was found that, rather than being an inhibitor of CD44-HA adhesion, TSG-6 is able to enhance or induce HA binding to CD44 on constitutive or inducible lymphoid cell backgrounds, respectively. In the latter case, EL4 cells (derived from a T cell

Hyaluronan in Immune Processes

65

Figure 7. The interaction of TSG-6 with HA may lead to the formation of cross-linked HA fibres. TSG-6/HA complexes of this type could induce receptor clustering, leading to an increase in HA binding.

lymphoma), which express CD44 on their cell surface but do not interact constitutively with HA (i.e., they can be considered a model for circulating lymphocytes), could be induced to bind HA in a manner reminiscent of the superagonist monoclonal antibodies (e.g., IRAWB14) discussed above (Lesley et al., 2004). Importantly, enhancement by TSG-6 of the CD44-mediated interaction of lymphoid cell lines with HA was seen under conditions of flow at shear forces comparable to those that occur in post-capillary venules. In order for TSG-6 to have its modulatory effect ff it was found necessary to preincubate the HA with TSG-6 at concentrations that saturate the majority of the protein-binding sites on the polysaccharide. These mixtures retained their activity over time, even after dilution, suggesting the formation of stable TSG-6/HA complexes. In this regard, the recombinant Link module domain (i.e., Link_TSG6) was also able to enhance/induce CD44’s interaction with HA (albeit with less potency than the full length protein), where mutation of HA-binding residues impaired its modulatory effect. ff This, and the lack of a detectable interaction of TSG-6 with CD44+ cells in the absence of HA, support the conclusion that the enhancing/inducing activity depends solely on the interaction of TSG-6 with HA. The demonstration that amino acids outside the HA-binding site are also involved in this activity gave rise to the hypothesis that TSG-6 may self-associate leading to the cross-linking of multiple HA chains to form extended fibres (Lesley et al., 2004). As shown in Fig. 7 engagement of TSG-6/HA fibres with cell surface CD44 could promote receptor clustering leading to an increase/activation of HA binding due to avidity effects ff or, potentially, changes in CD44 conformation. A similar mechanism

66

A. J. Wright and A. J. Day

is likely to be involved in the binding of HA cable-like structures (produced for example by mucosal smooth muscle cells in inflamed human colon) to CD44 on non-activated monocytes, which are not able to interact constitutively with free HA (see de la Motte et al., 2004). The finding that CD44-expressing EL4 cells (that do not constitutively bind HA) are able to interact with TSG-6/HA complexes and exhibit enhanced rolling on TSG-6/HA substrates, provides a mechanism whereby circulating lymphocytes might become adhesive in an inflammatory milieu where such complexes could be present (Lesley et al., 2004); both HA and TSG-6 are upregulated in blood vessels during inflammation. For example, it is possible that retention of TSG-6/HA complexes on endothelial cells could facilitate the CD44-mediated recruitment of leukocytes and thus be proinflammatory. Alternatively, release of TSG-6/HA complexes into the local circulation and their binding to CD44 on leukocytes could inhibit the adhesive interactions of circulating cells with the vascular endothelium, and thereby have an anti-inflammatory effect. ff In this regard, TSG-6/HA complexes were found to be effective ff competitors of CD44+ cell attachment and rolling on immobilised HA (Lesley et al., 2004). However, further research is necessary to address the exact role of TSG-6 in the regulation of CD44-mediated leukocyte migration.

7. CONCLUSIONS The polysaccharide HA is important in mediating the initial attachment and rolling of mononuclear leukocytes on the vascular endothelium at sites of inflammation through its interaction with CD44, a process that necessitates tight regulation. Recent structural studies have provided novel insights into how changes in the Nlinked glycosylation of CD44 could either unblock the HA-binding site or allow receptor clustering, thus switching CD44 into an active state. The inflammationassociated hyaladherin TSG-6 has also been implicated in the positive modulation of CD44-HA interactions on lymphocytes via a mechanism that is likely to involve the formation of cross-linked HA fibres which can engage and cluster multiple receptor molecules on the leukocyte surface. Further work is now required to investigate the formation and structure of these and other multi-molecular HA/protein complexes, to allow a better understanding of their roles in the regulation of immune cell adhesion.

8. ACKNOWLEDGEMENTS We would like to thank Dr. Peter Teriete for providing Figs. 2 and 6, and Dr. Caroline M. Milner for review of the manuscript. AJD acknowledges the support of the ARC and MRC, and AJW was the recipient of a MRC Studentship.

REFERENCES Aruffo, ff A., Stamenkovic, I., Melnick, M., Underhill, C.B., and Seed, B., 1990, CD44 is the principal cell surface receptor for hyaluronate., Cell, 61:1303–1313.

Hyaluronan in Immune Processes

67

Banerji, S., Day, A.J., Kahmann, J.D. and Jackson, D.G., 1998, Characterization of a functional hyaluronan-binding domain from the human CD44 molecule expressed in Escherichia coli. Protein Expr. Purif., 14:371–381. Bajorath, J., Greenfield, B., Munro, S.B., Day, A.J. and Aruffo, ff A., 1998, Identification of CD44 residues important for hyaluronan binding and delineation of the binding site. J. Biol. Chem. 273:338–343. ff A., Spring, F. and Stamenkovic I., 1996, Glycosylation of CD44 is Bartolazzi, A., Nocks, A., Aruffo, implicated in CD44-mediated cell adhesion to hyaluronan. J. Cell Biol., 132:1199–1208. Blundell, C.D., DeAngelis, P.L., Day, A.J. and Almond, A., 2004a, Use of 15N-NMR to resolve molecular details in isotopically-enriched carbohydrates: sequence-specific observations in hyaluronan oligomers up to decasaccharides. Glycobiology, 14:999–1009. Blundell, C.D., Mahoney, D.J., Almond, A., DeAngelis, P.L., Kahmann, J.D., Teriete, P., Pickford, A.R., Campbell, I.D. and Day, A.J., 2003, The link module from ovulation- and inflammation-associated protein TSG-6 changes conformation on hyaluronan binding. J. Biol. Chem., 278:49261–49270. Blundell, C.D., Seyfried, N.T. and Day, A.J., 2004b, Structural and functional diversity of hyaluronanbinding proteins. In: Chemistry and Biology of Hyaluronan (H.G. Garg and C. A. Hales eds), Elservier, pp189–204. Butcher, E.C., 1991, Leukocyte-endothelial cell recognition: three (or more) steps to specificity and diversity. Cell, 67:1033–1036. Cao, T.V., La M., Getting, S.J., Day, A.J. and Perretti, M., 2004, Inhibitory effects ff of TSG-6 Link module on leukocyte-endothelial cell interactions in vitro and in vivo. Micocirculation, 11:615–624. Day, A.J. and Prestwich, G.D., 2002, Hyaluronan-binding proteins: tying up the giant. J. Biol. Chem., 277:4575–4579. Day, A.J., Rugg, M.S., Mahoney, D.J. and Milner, C.M., 2004, The role of hyaluronan-binding proteins in ovulation. In: HA2003 Proceedings at http://www.matrixbiologyinstitute.org/ha03/toc.htm. Day, A.J. and Sheehan, J.K., 2001, Hyaluronan: polysaccharide chaos to protein organisation. Curr. Opin. Struc. Biol., 11:617–622. de la Motte, C.A., Hascall, V.C., Drazba, J., Bandyopadhyay, S.K. and Strong, S.A., 2003, Mononuclear leukocytes bind to specific hyaluronan structures on colon mucosal smooth muscle cells treated with polyinosinic acid:polycytidylic acid: inter-alpha-trypsin inhibitor is crucial to structure and function. Am. J. Pathol., 163:121–133. de la Motte, C.A., Drazba, J., Bandyopadhyay, S., Majors, A., Hascall, V.C. and Strong, S.A., 2004, Viral stimuli induce novel hyaluronan cable structures on colon smooth muscle cells that bind leukocytes externally and nuclei internally. In: HA2003 Proceedings at http://www.matrixbiologyinstitute.org/ ha03/toc.htm. DeGrendele, H.C., Estess, P., Picker, L.J. and Siegelman, M.H., 1996, CD44 and its ligand hyaluronate mediate rolling under physiological flow: A novel lymphocyte-endothelial cell primary adhesion pathway. J. Exp. Med., 183:1119–1130. DeGrendele, H.C., Kosfiszer, M., Estess, P. and Siegelman, M.H., 1997, CD44 activation and associated primary adhesion is inducible via T cell receptor stimulation. J. Immunol., 159:2549–2553. English, N.M., Lesley, J.F. and Hyman, R., 1998, Site-specific de-N-glycosylation of CD44 can activate hyaluronan binding, and CD44 activation states show distinct threshold densities for hyaluronan binding. Cancer Res., 58:3736–3742. Fraser, J.R., Laurent, T.C. and Laurent, U.B., 1997, Hyaluronan: its nature, distribution, functions and turnover. J. Intern. Med., 242:27–33. Fulop, C., Sza´nto´, S., Mukhopadhyay, D., Bardos, T., Kamath, R.V., Rugg, M.S., Day, A.J., Salustri, A., Hascall, T.T. and Mikecz K., 2003, Impaired cumulus mucification and female sterility in tumor necrosis factor-induced protein-6 deficient mice. Development, 130:2253–2261. Getting, S.J., Mahoney, D.J., Cao, T., Rugg, M.S., Fries, E., Milner, C.M., Perretti, M. and Day, A.J., 2002, The link module from human TSG-6 inhibits neutrophil migration in a hyaluronan- and inter-ainhibitor-independent manner. J. Biol. Chem., 277:51068–51076. Gee, K., Kozlowski, M. and Kumar, A., 2003, TNFa induces functionally active hyaluronan-adhesive CD44 by activating sialidase through p38 mitogen-activated protein kinase in lipopolysaccharidestimulated human monocytic cells. J. Biol. Chem., 278:37275–37287. Hardingham, T.E., 2004, Solution properties of HA. In: Chemistry and Biology of Hyaluronan (H.G. Garg and C. A. Hales eds), Elservier, pp1–19. Isacke, C.M. and Yarwood, H., 2002, The hyaluronan receptor, CD44. Int. J. Biochem. Cell Biol., 34:718–721.

68

A. J. Wright and A. J. Day

Kahmann, J.D., O’Brien, R., Werner, J.M., Heinegard, D., Ladbury, J.E., Campbell, I.D. and Day A.J., 2000, Localization and characterization of the hyaluronan-binding site on the link module from human TSG-6. Structure Fold Des., 8:763–774. Katoh, S., Miyagi, T., Taniguchi, H., Matsubara, Y., Kadota, J., Tominaga, A., Kincade, P.W., Matsukura, S. and Kohno, S., 1999, An inducible sialidase regulates the hyaluronic acid binding ability of CD44-bearing monocytes. J. Immunol., 162:5058–5061. Katoh S., Zheng Z., Oritani K., Shimozato T. and Kincade P.W. (1995). Glycosylation of CD44 negatively regulates its recognition of hyaluronan. J. Exp. Med., 182:419–429. Kohda, D., Morton, C.J., Parkar, A.A., Hatanaka, H., Inagaki, F.M., Campbell, I.D. and Day A.J., 1996, Solution structure of the link module: a hyaluronan-binding domain involved in extracellular matrix stability and cell migration. Cell, 86:767–775. Kubes, P. and Kerfoot, S.M., 2001, Leukocyte recruitment in the microcirculation: the rolling paradigm revisited. News Physiol. Sci., 16:76–80. Laurent, T.C, Laurent, U.B and Fraser, J.R., 1996, The structure and function of hyaluronan: An overview. Immunol. Cell Biol., 74:A1–7. Lee, H.L., Wisniewski, H.G. and Vilcek, J., 1992, A novel secretory tumor necrosis factor-inducible protein (TSG-6) is a member of the family of hyaluronan binding proteins, closely related to the adhesion receptor CD44. J. Cell Biol., 116:545–557. Lesley, J., English, N., Perschl, A., Gregoroff, ff J. and Hyman, R., 1995, Variant cell lines selected for alterations in the function of the hyaluronan receptor CD44 show differences ff in glycosylation. J. Exp. Med., 182:431–437. Lesley, J., Gal, I., Mahoney, D.J., Cordell, M.R., Rugg, M.S., Hyman, R., Day, A.J. and Mikecz, K., 2004, TSG-6 modulates the interaction between hyaluronan and cell surface CD44. J. Biol. Chem., 279:25745–25754. Lesley, J., Hascall, V.C., Tammi, M. and Hyman, R., 2000, Hyaluronan binding by cell surface CD44. J. Biol. Chem., 275:26967–26975. Lesley, J., Hyman, R., English, N., Catterall, J.B. and Turner, G.A., 1997, CD44 in inflammation and metastasis. Glycoconj. J., 14:611–622. Lesley, J., Kincade, P.W. and Hyman, R., 1993, Antibody-induced activation of the hyaluronan receptor function of CD44 requires multivalent binding by antibody. Eur. J. Immunol., 8:1902–1909. Mahoney, D.J, Blundell, C.D and Day, A.J., 2001, Mapping the hyaluronan-binding site on the link module from human tumor necrosis factor-stimulated gene-6 by site-directed mutagenesis. J. Biol. Chem., 276:22764–22771. Majors, A.K., Austin, R.C., de la Motte, C.A., Pyeritz, R.E., Hascall, V.C., Kessler, S.P., Sen, G. and Strong, S.A., 2003, Endoplasmic reticulum stress induces hyaluronan deposition and leukocyte adhesion. J. Biol. Chem., 278:47223–47231. Milner, C.M. and Day, A.J., 2003, TSG-6: a multifunctional protein associated with inflammation. J. Cell. Sci., 116:1863–1873. Mohamadzadeh, M., DeGrendele, H., Arizpe, H., Estess, P. and Siegelman, M., 1998, Proinflammatory stimuli regulate endothelial hyaluronan expression and CD44/HA-dependent primary adhesion. J. Clin. Invest., 101:97–108. Peach, R.J., Hollenbaugh, D., Stamenkovic, I. and Aruffo, ff A., 1993, Identification of hyaluronic acid binding sites in the extracellular domain of CD44. J. Cell Biol., 122:257–264. Pure´, E. and Cuff, ff C.A., 2001, A crucial role for CD44 in inflammation. Trends. Mol. Med., 7:213–221. Salustri, A., Garlanda, C., Hirsch, E., De Acetis, M., Maccagno, A., Bottazzi, B., Doni, A., Bastone, A., Mantovani, G., Beck Peccoz, P., Salvatori, G., Mahoney, D.J., Day, A.J., Siracusa, G., Romani, L. and Mantovani A., 2004, PTX3 plays a key role in the organization of the cumulus oophorus extracellular matrix and in in vivo fertilization. Development, 131:1577–1586. Siegelman, M.H., DeGrendele, H.C. and Estess, P., 1999, Activation and interaction of CD44 and hyaluronan in immunological systems. J. Leukocyte Biol., 66:315–321. Siegelman, M.H., Stanescu, D. and Estess, P., 2000, The CD44-initiated pathway of T-cell extravasation uses VLA-4 but not LFA-1 for firm adhesion. J. Clin. Invest., 105:683–691. Sza´nto´, S., Bardos, T., Gal, I., Glant, T.T. and Mikecz, K., 2004, Enhanced neutrophil extravasation and rapid progression of proteoglycan-induced arthritis in TSG-6-knockout mice. Arthritis Rheum., 50:3012–3022. Tammi, M.I., Day, A.J. and Turley, E.A., 2002, Hyaluronan and homeostasis: a balancing act. J. Biol. Chem., 277:4575–4579.

Hyaluronan in Immune Processes

69

Teriete, P., Banerji, S., Noble, M., Blundell, C.D., Wright, A.J., Pickford, A.R., Lowe, E., Mahoney, D.J., Tammi, M.I., Kahmann, J.D., Campbell, I.D., Day, A.J. and Jackson, D.G., 2004, Structure of the regulatory hyaluronan binding domain in the inflammatory leukocyte homing receptor CD44. Mol. Cell, 13:483–496. Toole, B.P., 2004, Hyaluronan: from extracellular glue to pericellular cue. Nat. Rev. Cancer, 4:528–539 Wisniewski, H.G. and Vilcek, J., 2004, Cytokine-induced gene expression at the crossroad of innate immunity, inflammation and fertility: TSG-6 and PTX3/TSG-14. Cytokine Growth Factor Rev., 15:129–146. Zheng, Z., Cummings, R.D., Pummill, P.E. and Kincade, P.W., 1997, Growth as a solid tumor or reduced glucose concentrations in culture reversibly induce CD44-mediated hyaluronan recognition by Chinese hamster ovary cells. J. Clin. Invest., 100:1217–1229 Zheng, Z., Katoh, S., He, Q., Oritani, K., Miyake, K., Lesley, J., Hyman, R., Hamik, A., Parkhouse, R.M. and Farr, A.G., 1995, Monoclonal antibodies to CD44 and their influence on hyaluronan recognition. J. Cell. Biol., 130:485–95.

12

GLYCOSYLATION AND THE FUNCTION OF THE T CELL CO-RECEPTOR CD8

David A. Shore1,2, Ian A. Wilson2,3, Raymond A. Dwek1 and Pauline M. Rudd1 1The Glycobiology Institute Department of Biochemistry University of Oxford Oxford, OX1 3QU, UK 2Department of Molecular Biology 3Skaggs Institute for Chemical Biology The Scripps Research Institute 10550 North Torrey Pines Road La Jolla, CA 92037, USA.

1. INTRODUCTION The CD8 glycoprotein functions at the surface of cytotoxic T-lymphocytes (CTLs) as an essential co-receptor in T cell activation in response to peptide antigen complexed with the class I major histocompatibility complex (pMHC). CD8 interacts with the class I pMHC in a bidentate manner with the T cell receptor (TCR), to initiate and augment T-cell signalling through the CD3 subunits of the TCR complex (reviewed in Devine, 1999; Garcia, 1999; Gao, 2000; Wang, 2000; van der Merwe, 2003). In the absence of CD8 at the cell surface, the CTL response to peptide antigen is significantly impaired (Luescher, 1995). Likewise, the CD8/pMHC interaction greatly enhances TCR-mediated T cell sensitivity to antigen (Purbhoo, 2001). The CD8 glycoprotein comprises two distinct subunits, alpha(a) and beta(b), and is expressed at the cell surface as a mixture of covalently associated aa homodimers and ab heterodimers. Activation of the genes encoding the a and b subunits of CD8 is under the control of separate promoters, and the expression of CD8aa and CD8ab is highly cell-type specific (Gangadharan, 2004 and references therein). The ab heterodimeric form of CD8 is the abundant form of the co-receptor at the surface thymocytes and mature conventional class I restricted T cells (reviewed 71 John S. Axford (ed.), Glycobiology and Medicine, 71-84. © 2005 Springer. Printed in the Netherlands.

72

D. A. Shore et al.

in Zamoyska, 1994). CD8ab is the primary co-receptor for abTCR-mediated signalling in response to antigen, and is 100 fold more effective ff than CD8aa (Arcaro, 2000; Arcaro, 2001; Cawthon, 2001). Mice deficient in CD8b exhibit a substantially reduced number of peripheral CD8+ T cells, implicating CD8b in the development and differentiation ff of such cells (Crooks, 1994, Fung-Leung, 1994). The aa and ab isoforms of CD8 are functionally distinct (reviewed in Gangadharan, 2004). In contrast to CD8ab, expression of CD8aa is not limited to class I-restricted T cells. The aa receptor has been identified on the surface of a number of different ff cell types, including CD4+ T cells, lymphoid-related dendritic cells, cdTCR cells, NK cells and intraepithelial lymphocytes of the gut. It is thought that CD8aa acts in a class I pMHC-independent manner to elicit a regulatory role in T cell function. Specifically, a strong interaction between CD8aa and the thymic leukaemia antigen (TL) has been demonstrated (Devine, 2002; Liu, 2003). In this instance, the TCR-independent interaction between CD8aa and the non-antigen presenting TL is thought to modify TCR activation signals received by antigenstimulated cells (Cheroutre, 1995; Leishman, 2001).

2. STRUCTURE OF THE CD8 GLYCOPROTEIN Despite a sequence identity of only 20%, the a and b subunits are predicted to have a similar topology (Norment, 1988). Each subunit comprises an ectodomain of 150–170 amino acids, a type-2 single pass transmembrane domain and a short cytoplasmic region (Fig. 1a). The CD8 ectodomain consists of an amino terminal immunoglobulin V-set domain, tethered to the membrane by a 30–50 residue flexible stalk. The intracellular region of the a subunit of CD8 interacts with the protein tyrosine kinase LckP56 through a conserved CxCP motif (Fung-Leung, 1993; Arcaro, 2001; Kim, 2003). The stalk region of both CD8 subunits is rich in proline, serine and threonine residues, and is highly O-glycosylated in all species studied. These characteristics likely confer an extended but relatively rigid conformation to the stalk region, as in leukosialin and mucins (reviewed in Rudd, 1999). Amino acid sequencing of the membrane-distal region of rat CD8a has determined that four threonine residues, T122, T126, T132 and T134 are post-translationally modified with O-linked carbohydrates (Classon, 1992). The distribution of potential N-linked glycosylation sites between the a and b subunits is highly variable across species. The presence of at least one N-linked glycosylation site on the b subunit is common to all species, whereas N-glycosylation of the a subunit appears to be species specific (Merry, 2003).

3. INTERACTION OF CD8 WITH CLASS I MHC Binding studies involving soluble forms of the murine and human CD8 and class I MHC have shown that the affinities for class I pMHC of both the aa and ab dimers in solution are equivalent (Sun, 1997, Kern, 1999). Furthermore, CD8 interactions are generally an order of magnitude weaker than those involving the TCR, although TCR/pMHC complex affinity is highly dependent upon TCR haplotype

Glycosylation and the Function of the T Cell Co-receptor CD8

73

Figure 1. The overall topology and function of the CD8ab co-receptor. (a) The a and b subunits of CD8 are covalently associated via disulphide bonds within the stalk region of the receptor. The a and b Ig-like regions dimerise to form the class I pMHC binding site. The a subunit of CD8 is associated with the intracellular phospho-tyrosine kinase p56LcK. The b subunit stalk is shorter than that of the a subunit, potentially producing a bowed effect ff in the ab dimer. (b) The IgSF head group of CD8 spans the intercellular gap to bind the membrane proximal region of the class I MHC. The flexibility inherent within the CD8 stalk region likely enables the head group to adopt the optimal orientation for pMHC interaction. Interaction between CD8 and class I pMHC brings the associated p56LcK tyrosine kinase into close association with the ITAMs of the TCR/CD3 complex, such that an intracellular signaling cascade is initiated/augmented and maintained for the duration of the CD8/MHC interaction.

74

D. A. Shore et al.

and antigen (i.e. strong or weak agonist) and varies substantially (Kern, 1999; Garcia, 1996; 2003 van der Merwe). Hence, although recruitment of CD8 to the nascent pMHC/TCR may somewhat contribute to the overall affinity of the TCR for pMHC, it is unlikely that the primary co-receptor function of CD8 is the addition of structural support. It seems more probable that CD8 functions predominantly to bring the protein kinase LcKp56 into close association with the intracellular region of the CD3 subunits, giving rise to TCR/CD3 signal transduction when bound to pMHC (Fig. 1b) (Arcaro, 2001; Purbhoo, 2001; Doucey, 2003; Palacios, 2004). The crystal structure of a murine CD8aa immunoglobulin super family (IgSF) domain dimer in complex with the class I MHC H-2Kb has been determined (Kern, 1998). This structure describes an interaction between the membrane proximal region of the MHC and the six CDR-equivalent loops of the CD8 dimer in an antibodylike manner, perpendicular to the long axis of the MHC. In this model, the contribution of the individual CD8 subunits to the binding interface is asymmetric, with one of the CD8a subunits in the ‘‘upper’’ a1 position accounting for ~70% of the binding site (Fig. 2). Both CD8 subunits contact the protruding acidic loop of the nonpolymorphic MHC a3 domain which is considered the primary binding site; however, the a1 subunit of the CD8 dimer makes additional contacts with the MHC a2 domain, as well as with b2-microglobulin (Kern, 1998; Li, 1998). This mode of interaction was also observed in the crystal structure of the human form of CD8aa in complex with the human class I MHC molecule HLA-A2 (Gao, 1997). Currently no structural data are available to determine the orientation of the CD8b subunit in complex with class I MHC, although given the predicted similarities between IgSf domains of the two subunits, it is likely that binding of CD8ab to class I MHC is equivalent to that of CD8aa. Mutations in the CDR-like regions of the CD8a subunit have a dramatic effect ff upon the ability of CD8ab to interact with pMHC, whereas similar mutations in the b subunit have little effect ff upon complex formation (Devine, 1999a). Furthermore, transfection studies to express a chimeric human bb homodimer at the cell surface suggest that this form of the co-receptor is incapable of interacting with conventional class I MHC, indicating a critical role for the a subunit in co-receptor binding (Devine, 2000). These findings suggest that the a subunit dominates the binding site with MHC, whereas the b subunit occupies the ‘‘lower’’ a2 position in the interface. However, the shorter b subunit stalk would appear to favour a model in which the b subunit occupies the ‘‘upper’’ a1 position in the CD8/MHC complex (as in Fig. 1). The precise position of the CD8ab dimer relative to the MHC, therefore, remains a contentious issue which must await a CD8ab/pMHC complex crystal structure.

4. FUNCTION OF THE CD8b SUBUNIT It is apparent that the b subunit contributes significantly to the function of CD8 as a co-receptor to TCR mediated signalling (Irie, 1998; Witte, 1999; Bosselut, 2000). However, the mechanism underlying the enhanced effector ff function of CD8b over that of CD8a is unclear. It has been suggested that binding of the b subunit to class I pMHC induces a conformational change within the class I receptor, which in turn enhances the interaction between the pMHC and TCR (Garcia, 1996). Given the

Glycosylation and the Function of the T Cell Co-receptor CD8

75

Figure 2. The interaction between the murine CD8aa head group and the class I pMHC H-2Kb. The domain architecture of the CD8aa head group dimer is similar to that of an immunoglobulin Fab variable region. The 6 CDR-like loops of the CD8aa dimer clamp around an extended loop between the C and D strand of the class I pMHC a3 domain. Binding of the CD8a subunits to class I pMHC is asymmetric; the ‘‘upper’’ (CD8a1) subunit accounts for ~70% of the solvent exposed area within the CD8 binding pocket. Interaction between the CD8a1 subunit and the pMHC are strengthened by additional interactions with the MHC a2 domain, as well as with the CD8a DE loop and the b M domain. 2 The interaction between CD8 and pMHC is of relatively low affinity (~60 mM) and as such is likely to be reversible and sensitive to avidity effects ff in vivo.

current structural information for the class I pMHC, both in complex with the TCR and in complex with the CD8, it is not clear how such a change may be instigated at the CD8 binding site and transmitted through the MHC to the peptide binding region. The cytoplasmic tail of the CD8b subunit is palmitoylated (Arcaro, 2000), which is thought to enable CD8 to interact with cell membrane rafts more readily, thereby enhancing interaction between CD8 and raft-associated receptor molecules, such as TCR/CD3 (Arcaro, 2001). However, other work has shown that the enhanced activity of the b subunit over the a subunit of CD8 resides primarily within the stalk-like region of the ectodomain (Witte, 1999). Mutant CD8a subunits, comprising the a IgSF domain and the b stalk region are capable of restoring CD8b(−) CTL activity in the presence of class I pMHC to nearly that of wildtype, as assessed by interleukin production and cellular proliferation (Witte, 1999.) Furthermore, the presence of the b stalk region enhances the co-receptor activity, and produces a distinct topology in the resulting CD8/MHC complex (Wong, 2003).

76

D. A. Shore et al.

In all species studied, the CD8b stalk region is shorter than that of the a subunit by 10–13 amino acids. This differentiation ff in stalk length would potentially produce a bowed effect ff within the stalk region of the ab heterodimer, in contrast with that of the symmetric CD8aa homodimer. Although the purpose of this structural bifurcation is unknown, it may well contribute to the enhanced effecter ff function of the b subunit by orientating the ab heterodimer at the cell surface such that class I MHC interaction is favoured.

5. CD8/MHC INTERACTION IS MODULATED BY O-LINKED GLYCOSYLATION The glycosylation state of T cell surface markers is associated with the T cell maturation level and activation state (Piller, 1988; Wu, 1996). Specifically, immature thymocytes and activated, mature T cells exhibit lower levels of cell-surface sialylation than mature resting T cells. Several lines of evidence indicate developmental regulation of CD8/class I MHC interaction that is independent of TCR specificity (reviewed in Daniels, 2002; Baum, 2002; Gascoigne, 2002). Work from the laboratories of Ellis Reinhertz and Steve Jameson has shown that changes in O-glycosylation associated with T cell maturation have a direct effect ff upon the ligand binding properties of CD8 in vivo (Moody, 2001; Daniels, 2000; Daniels, 2001). Daniels et al. demonstrated differences ff in the affinity of CD8 for soluble tetramers of non-cognate pMHC by comparing CD8’s ability to interact with soluble pMHC before and after T cell maturation. A marked reduction in class-I MHC/CD8 interaction is associated with the positive selection of CD8+ T cells from ff CD4+/CD8+ double positive (DP) thymocytes in this assay. To determine the effect of sialylation of cell-surface glycoproteins upon MHC/CD8 interaction, thymocytes and mature T cells extracted from the lymph nodes of mice were treated with neuraminidase. Strikingly, a significant increase in MHC/CD8 interaction was observed following neuraminidase treatment of mature CD8+ T cells. Hence, Daniels and co-workers propose a mechanism whereby CD8 interaction with pMHC is diminished following T cell maturation in a glycosylation-dependent manner; specifically, changes in the surface sialylation of T cells that occur in thymocyte development have a dramatic impact upon the ability of CD8 to bind class I MHC (Daniels, 2001). Moody et al. have observed a similar TCR-independent variation in the affinity of CD8 for pMHC tetramers in DP thymocytes as compared to resting, mature CD8+ T cells. Furthermore, Moody and et al. report that CD8ab heterodimers account for the overwhelming majority of CD8 at the thymocyte cell surface, and demonstrate that glycosylation of the b subunit is critical to the variable activity of CD8. Immunoprecipitation studies indicate that CD8b is subject to the most substantial modification in O-linked glycosylation during thymocyte maturation. Variation in T cell sialylation is brought about in part by developmentallyregulated expression of the enzyme ST3 Gal-1, which catalyses addition of sialic acid to core 1 O-linked glycans (Priatel, 2000). Moody et al. demonstrated that ST3 Gal-1 is largely responsible for the increase in sialylation of CD8 at the surface of T cells following maturation (Moody, 2001).

Glycosylation and the Function of the T Cell Co-receptor CD8

77

Further work implicated addition of sialic acid moieties to O-linked carbohydrates at specific locations of the CD8b stalk region as responsible for the effects ff of reduced CD8ab/MHC interaction, resulting from an increase in ST3 Gal-1 expression which accompanies the thymic development of CD8+ T cells from the CD4+/CD8+ double-positive to CD8+ single-positive stage (Moody, 2003). The sialylation of Olinked carbohydrates attached to peptide fragments derived from tryptic digestion of the murine CD8b subunit was also analysed by nanospray ES-MS. Five of fourteen potential O-glycosylation sites in the stalk region of CD8b were found to be posttranslationally modified, of which three are conserved in all species so far studied. The modified threonines (T120, T121, T124, T127 and T128) are clustered in the membrane distal portion of the CD8b stalk, in close association to the IgSF domain. Surprisingly, core-1 O-glycans of the CD8b subunits of both thymocytes and mature CD8+ T cells were found to be mono-sialylated. Additional sialylation of a single O-glycan at position T120-T124 was observed solely in CD8+ SP thymocytes and is, therefore, likely responsible for the enhanced CD8/MHC binding activity of these cells over immature CD4+/CD8+ DP thymocytes.

6. THE INFLUENCE OF O-GLYCOSYLATION UPON THE EXTENSION OF THE CD8 STALK The finding that a decrease in CD8/pMHC binding results from a developmentally mediated increase in CD8b O-glycan sialylation upon T cell maturation accounts for the predominance of CD8ab over CD8aa in T cell selection and development. However, it is not immediately clear how changes to O-glycans in the stalk region of CD8 can have such a dramatic effect ff upon the CD8/MHC binding site, which is ˚ away. It has also been proposed that the presence of sialic acid situated some 30 A in the stalk region gives rise to an electrostatic repulsion effect ff between CD8 and class I pMHC. However, thermodynamic considerations make this mechanism extremely unlikely. Moody et al. explain their results by suggesting that the addition of sialic acids to O-glycans in the stalk region stabilizes or re-orientates the CD8 head group in such a way that interaction with class I pMHC is no longer favoured. There is evidence that O-glycans can have an influence upon the structural characteristics of polypeptides to which they are attached (Gerken, 1989, Shogren, 1989, Rudd, 1999a) and it is possible that changes to the structure of O-glycans in the membrane distal region of the CD8 stalk influence the extent of the extension and conformation of the stalk polypeptide. To investigate whether the structure of the CD8 stalk is sensitive to modification by the addition of sialic-acid adducts to O-glycans in the membrane distal region, a structure-function analysis of CD8 carbohydrates was carried out in our laboratory. To this end, soluble forms of the CD8a and b subunit ectodomains were designed (as described elsewhere) and expressed in the Chinese hamster ovary (CHO) cell lines K1 and Lec3.2.8.1 (Merry, 2003). The CHO Lec3.2.8.1 cell line lacks the enzyme GlcNAc transferase T1, necessary for the processing of complex-form oligosacharides in the golgi apparatus, and so produces only oligomannose N-linked glycans (see Chen, 2003). The profile of O-linked glycans of proteins produced in the Lec3.2.8.1

78

D. A. Shore et al.

cell was thus expected to be significantly different ff to that of proteins expressed in the K1 cell line. The use of a glycosylation-deficient expression system was chosen as a preferred means of producing differentially-glycosylated ff protein over the alternatives of genetic modification and chemical inhibition of glycosylation due to the potential problems inherent in these systems, such as protein mis-folding (Butters, 1998). An unglycosylated form of the CD8aa IgSF domain dimer lacking the stalk region, produced in the E.coli expression system, was used as a control for these experiments. Given the significant functional activity of O-glycosylation specific to the b subunit of CD8 as opposed to that of the a subunit, it was reasonable to expect some differences ff between the N-linked glycosylation profiles of the two subunits. However, analysis of N-glycans removed from separated CD8a and b constructs produced in CHO K1 cells detected no significant differences ff between the two subunits, despite the very low level of protein sequence conservation. A comprehensive O-glycan analysis of CD8 proteins produced in both the CHO K1 and CHO LecR systems was carried out by HPLC analysis. O-glycans in The stalk region of CD8 produced in the CHO K1 system contained O-glycans of the type 1 core Gal-b1,3 GalNAc structure, the majority of which were mono and di-sialylated, although a small amount (15%) of unsialylated Gal-b1,3 GalNAc was also observed. In contrast, sialylation of O-glycans attached to CD8 produced in the LecR system was extremely limited, and was restricted to the infrequent addition of a single sialic acid to the core-1 disaccharide. An analysis of O-glycan occupancy in CD8 constructs was carried out by electrospray mass spectrometry and subsequent MS/MS fragmentation of tryptic peptide fragments of the rat CD8 stalk region. Three threonine residues (T126, T132 and T134) within the stalk-like region of CD8 constructs produced in both K1 and LecR systems were occupied, although the addition of hexose residues to the GalNAc core was extremely limited in proteins derived in LecR cells (see Fig. 3). To examine the extent of polypeptide chain extension induced by O-glycans in the stalk region of CD8 constructs produced in the CHO K1, CHO LecR and E.coli expression systems, sedimentation (s) and Perrin (P) values were determined for each construct by analytical ultracentrifugation (AUC). The (s) values determined for each ff sizes of the variably glycosylated sample varied significantly, due to the different forms of the protein. Critically, the experimentally-derived P value, which gives information regarding the shape of the molecule, was roughly equivalent for both the CHO K1 and CHO LecR derived constructs. However, a substantial difference ff was observed between the E.coli derived form of soluble IgSF CD8 and those expressed in the mammalian systems, which confirms that the system is indeed sensitive to both molecular mass and shape. Taken together, these data suggest that the presence of additional hexose residues and/or sialic acids attached to the Nacetylgalactosamine core of O-glycans in the stalk of CD8 have little or no overt affect ff upon the length of the polypeptide stalk region (Fig. 4).

7. CONCLUSIONS AND FUTURE PERSPECTIVES The majority of cell surface glycoproteins are modified by terminal sialylation (Rudd, 1999a; 2001; 2004). The sialylation of T cell surface glycoproteins is directly

Glycosylation and the Function of the T Cell Co-receptor CD8

79

Figure 3. O-glycan analysis of CHO K1 and CHO Lec 3.2.8.1-derived CD8 constructs. O-glycans released from the CHO K1 derived protein consisted of the type 1 core Gal-b1,3GalNAc disaccharide, and its sialylated tri- and tetrasaccharide derivatives. Positive ion MALDI mass spectra of the desialylated tryptic peptides from CHO Lec 3.2.8.1 derived protein indicated that ~70% of the Oglycans present consisted of only a single GalNAc residue, whilst the remaining 30% where extended b1,3-galactose to form the type 1 core Gal-b1,3GalNAc. The O-glycosylation profile of CHO Lec-derived protein was characterized by the absence of the di-sialylated structure as well as a substantial decrease in the relative amounts of the mono-sialylated structure; 82% of the CHO Lec 3.2.8.1-derived type 1 core Gal-b1,3GalNAc disaccharides were non-sialylated.

associated with the developmental state and activation state of the T cell (Piller, 1988; Harrington, 2000; Starr, 2003; Pappu, 2004). The increase in T cell surface sialylation that accompanies T cell maturation from the immature double-positive stage to mature CD8+ T cells has a negative effect ff upon the ability of CD8 to recognize and interact with class I MHC (Moody, 2001; Daniels, 2001). This variation in CD8/MHC interaction is independent of MHC haplotype and TCR specificity. The stalk polypeptide of CD8 extends the IgSF head group of the co-receptor across the intercellular junction to contact its binding site on the membrane proximal region of the class I MHC. O-glycans present in the membrane distal region of the stalk polypeptide of both subunits likely act to rigidify this region of the polypeptide by reducing the overall extension of the stalk from the theoretical maximum ˚ /residue to ~2.6 A ˚ /residue as observed in mucins, as compared to of ~3.7 A ˚ /residue in the a-helical conformation. Data from our laboratory suggest that ~1.5 A the addition of a single GalNAc moiety to threonine residues in the membrane distal stalk polypeptide is sufficient to induce this rigidification in the CD8 stalk (Merry, 2003). The presence of additional carbohydrate residues attached to the terminal ff upon the conformation of the CD8 stalk GalNAc moiety has no further effect polypeptide. However, the addition of sialic acids to these O-glycans provides a level of post-translational control over the function of CD8+ T cells with regards to class I pMHC interaction (reviewed in Daniels, 2002). Given the apparent inability of sialic acids in the CD8 stalk to directly influence the conformation of this polypeptide, the upregulation of sialylation during thymocyte development must have alternative outcomes for CD8 activity.

80

D. A. Shore et al.

Figure 4. Molecular bead modeling of the soluble forms of CD8aa produced in (a) E.coli, (b) CHO Lec3.2.8.1 and (c) CHO K1. In this simulation, the structure of the CD8aa head group is based on the previously solved crystal structure of murine CD8aa in complex with classI p MHC (Kern, 1998). The stalk region of mammalian CD8aa was modeled according to the amino acid sequence described in Merry, ˚ per residue, as observed for mucins. N-linked and O-linked oligosacch2003, assuming an extension of 2.6 A arides have been modeled according to the glycan analysis detailed in Merry, 2003. The S and P values given for each protein species are the calculated sedimentation (s) coefficients and Perrin (P) functions applicable to each bead model, generated using a computational simulation program, ˚ /residue. S and P values are those determined experimentally assuming a stalk region extension of 2.6 A exp exp by AUC. Critically, the calculated P values, which give information regarding the shape of eth molecule, are exp similar for both CHO Lec and CHO K1 derived forms of the protein. These data suggest that, although the presence or absence of the N-terminal stalk region per se is significant to the overall shape and extended conformation of the molecule in solution, the presence of additional sialic acids in the CHO K1 derived protein has little influence upon the extension of the stalk region as compared with that of the CHO Lec R derived protein.

It has been reported that clusters of O-linked carbohydrates attached to mucinlike polypeptides have the effect ff of restricting the conformational plasticity within such polypeptides (Gerken, 1989; Shogren, 1989). It is tempting to speculate that the addition of sialic acids to O-glycans in the highly extended CD8 stalk region may function to further stiffen ff the polypeptide and, thereby, limit the flexibility of the co-receptor head group. The effects ff of subtle changes to the rigidity of the polypeptide stalk in the b subunit may be enhanced by the difference ff in length between the a and b subunits, such that the orientation of the IgSF dimer is restricted, resulting in decreased interaction with the binding site of pMHC. It is also possible that an increase in sialic acid content within the stalk region acts to reduce CD8 co-localisation at the cell surface, which would account for the reduced pMHC tetramer binding observed by Moody et al and Daniels et al. Given the low affinity binding between CD8 and pMHC, it is likely that these interactions are under the

Glycosylation and the Function of the T Cell Co-receptor CD8

81

influence of avidity effects, ff such as density-dependent binding, which are independent of any structural changes within the CD8 or pMHC. Further investigation of the effects ff of increased CD8b sialylation upon individual CD8ab/MHC complex formaff of sialylation upon this tion is required in order to establish the structural effects interaction. In view of the functional significance of the stalk region of the CD8b subunit, it is interesting to speculate on the function of the equivalent stalk region of the CD8a subunit, which is similarly rich in threonine, serine and proline residues. To date, however, there have been no reports of an analogous glycosylation dependent functional variation in the homodimeric CD8aa isoform of the co-receptor. Likewise, there has been no evidence to suggest that the CD4 co-receptor is subject to glycosylation-dependent classII MHC binding, suggesting that activation and maturation state-dependent control is specific to the CD8ab/pMHC interaction. The CD4 and CD8 co-receptors are structurally distinct, despite their apparently similar biological functions as a co-receptor for TCR activation (see Gao, 2002). The CD4 co-receptor has evolved to address the problem of spanning the inter-cellular junction by association of four IgSF domains, whereas the CD8 receptor utilises an extended linker domain to serve this purpose. The reasons underlying the structural differences ff between CD4 and CD8 have yet to be determined. It is possible that the presence of O-glycans in the CD8 stalk amenable to post-translational modification may be constitutive to CD8+ T cell function and, therefore, fundamental to the structural differentiation ff of the CD4 and CD8 co-receptors. Although it is widely accepted that sialylation of O-linked carbohydrates at the T cell surface has a profound effect ff upon T cell activities, the significance of variation in the sialylation of individual glycoproteins in vivo is difficult to determine. Recent work examined the effects ff of the global de-sialylation of T cell surface markers on the ability of mature T cells to react to peptide antigen (Starr, 2004). T cells exhibit a specific loss in sensitivity to low affinity ligands following maturity, while sensitivity to higher affinity ligands is maintained. Starr and colleagues report that the level of the response by mature T cells to low-affinity peptide antigen is restored to the level of immature thymocytes in the presence of neuraminidase, suggesting a role for sialic acids in the fine tuning of the TCR activation threshold. Although these results may be explained in part by an enhanced CD8 activity following desialylation, it is apparent that the variable sialylation of other T cell surface glycoproteins is equally critical to T cell function in response to peptide antigen.

ACKNOWLEDGEMENTS IAW is supported by the National Institutes of Health Grants AI42266 and CA58896.

REFERENCES Arcaro, A., Gregoire, C., Bakker, T. R., Baldi, L., Jordan, M., Goffin, L., Boucheron, N., Wurm, F., van der Merwe, P. A., Malissen, B., and Luescher, I. F. (2001). CD8b endows CD8 with efficient coreceptor

82

D. A. Shore et al.

function by coupling T cell receptor/CD3 to raft-associated CD8/p56(lck) complexes. J Exp Med 194, 1485–95. Arcaro, A., Gregoire, C., Boucheron, N., Stotz, S., Palmer, E., Malissen, B., and Luescher, I. F. (2000). Essential role of CD8 palmitoylation in CD8 coreceptor function. J Immunol 165, 2068–76. Baum, L. G. (2002). Developing a taste for sweets. Immunity 16, 5–8. Bosselut, R., Kubo, S., Guinter, T., Kopacz, J. L., Altman, J. D., Feigenbaum, L., and Singer, A. (2000). Role of CD8b domains in CD8 coreceptor function: importance for MHC I binding, signaling, and positive selection of CD8+ T cells in the thymus. Immunity 12, 409–18. Butters, T. D., Sparks, L. M., Harlos, K., Ikemizu, S., Stuart, D. I., Jones, E. Y., and Davis, S. J. (1999). Effects ff of N-butyldeoxynojirimycin and the Lec3.2.8.1 mutant phenotype on N-glycan processing in Chinese hamster ovary cells: application to glycoprotein crystallization. Protein Sci 8, 1696–701. Cawthon, A. G., Lu, H., and Alexander-Miller, M. A. (2001). Peptide requirement for CTL activation reflects the sensitivity to CD3 engagement: correlation with CD8ab versus CD8aa expression. J Immunol 167, 2577–84. Chen, W., and Stanley, P. (2003). Five Lec1 CHO cell mutants have distinct Mgat1 gene mutations that encode truncated N-acetylglucosaminyltransferase I. Glycobiology 13, 43–50. Cheroutre, H., Holcombe, H. R., Tangri, S., Castano, A. R., Teitell, M., Miller, J. E., Cardell, S., Benoist, C., Mathis, D., and Huse, W. D. (1995). Antigen-presenting function of the TL antigen and mouse CD1 molecules. Immunol Rev 147, 31–52. Classon, B. J., Brown, M. H., Garnett, D., Somoza, C., Barclay, A. N., Willis, A. C., and Williams, A. F. (1992). The hinge region of the CD8a chain: structure, antigenicity, and utility in expression of immunoglobulin superfamily domains. Int Immunol 4, 215–25. Crooks, M. E., and Littman, D. R. (1994). Disruption of T lymphocyte positive and negative selection in mice lacking the CD8 beta chain. Immunity 1, 277–85. Daniels, M. A., Devine, L., Miller, J. D., Moser, J. M., Lukacher, A. E., Altman, J. D., Kavathas, P., Hogquist, K. A., and Jameson, S. C. (2001). CD8 binding to MHC class I molecules is influenced by T cell maturation and glycosylation. Immunity 15, 1051–61. Daniels, M. A., Hogquist, K. A., and Jameson, S. C. (2002). Sweet ‘n’ sour: the impact of differential ff glycosylation on T cell responses. Nat Immunol 3, 903–10. Daniels, M. A., and Jameson, S. C. (2000). Critical role for CD8 in T cell receptor binding and activation by peptide/major histocompatibility complex multimers. J Exp Med 191, 335–46. Devine, L., and Kavathas, P. B. (1999a). Molecular analysis of protein interactions mediating the function of the cell surface protein CD8. Immunol Res 19, 201–10. Devine, L., Kieffer, ff L. J., Aitken, V., and Kavathas, P. B. (2000). Human CD8b, but not mouse CD8 beta, can be expressed in the absence of CD8 alpha as a beta beta homodimer. J Immunol 164, 833–8. Devine, L., Rogozinski, L., Naidenko, O. V., Cheroutre, H., and Kavathas, P. B. (2002). The complementarity-determining region-like loops of CD8a interact differently ff with beta 2-microglobulin of the class I molecules H-2Kb and thymic leukemia antigen, while similarly with their alpha 3 domains. J Immunol 168, 3881–6. Devine, L., Sun, J., Barr, M. R., and Kavathas, P. B. (1999a). Orientation of the Ig domains of CD8a beta relative to MHC class I. J Immunol 162, 846–51. Doucey, M. A., Goffin, L., Naeher, D., Michielin, O., Baumgartner, P., Guillaume, P., Palmer, E., and Luescher, I. F. (2003). CD3 delta establishes a functional link between the T cell receptor and CD8. J Biol Chem 278, 3257–64. Fung-Leung, W. P., Kundig, T. M., Ngo, K., Panakos, J., De Sousa-Hitzler, J., Wang, E., Ohashi, P. S., Mak, T. W., and Lau, C. Y. (1994). Reduced thymic maturation but normal effector ff function of CD8+ T cells in CD8b gene-targeted mice. J Exp Med 180, 959–67. Fung-Leung, W. P., Louie, M. C., Limmer, A., Ohashi, P. S., Ngo, K., Chen, L., Kawai, K., Lacy, E., Loh, D. Y., and Mak, T. W. (1993). The lack of CD8a cytoplasmic domain resulted in a dramatic decrease in efficiency in thymic maturation but only a moderate reduction in cytotoxic function of CD8+ T lymphocytes. Eur J Immunol 23, 2834–40. Gangadharan, D., and Cheroutre, H. (2004). The CD8 isoform CD8aa is not a functional homologue of the TCR co-receptor CD8ab. Curr Opin Immunol 16, 264–70. Gao, G. F., and Jakobsen, B. K. (2000). Molecular interactions of coreceptor CD8 and MHC class I: the molecular basis for functional coordination with the T-cell receptor. Immunol T oday 21, 630–6. Gao, G. F., Rao, Z., and Bell, J. I. (2002). Molecular coordination of ab T-cell receptors and coreceptors CD8 and CD4 in their recognition of peptide-MHC ligands. T rends Immunol 23, 408–13.

Glycosylation and the Function of the T Cell Co-receptor CD8

83

Gao, G. F., Tormo, J., Gerth, U. C., Wyer, J. R., McMichael, A. J., Stuart, D. I., Bell, J. I., Jones, E. Y., and Jakobsen, B. K. (1997). Crystal structure of the complex between human CD8a(a) and HLA-A2. Nature 387, 630–4. Garcia, K. C., Scott, C. A., Brunmark, A., Carbone, F. R., Peterson, P. A., Wilson, I. A., and Teyton, L. (1996). CD8 enhances formation of stable T-cell receptor/MHC class I molecule complexes. Nature 384, 577–81. Garcia, K. C., Teyton, L., and Wilson, I. A. (1999). Structural basis of T cell recognition. Annu Rev Immunol 17, 369–97. Gascoigne, N. R. (2002). T-cell differentiation: ff MHC class I’s sweet tooth lost on maturity. Curr Biol 12, R99-R101. Gerken, T. A., Butenhof, K. J., and Shogren, R. (1989). Effects ff of glycosylation on the conformation and dynamics of O-linked glycoproteins: carbon-13 NMR studies of ovine submaxillary mucin. Biochemistry 28, 5536–43. Harrington, L. E., Galvan, M., Baum, L. G., Altman, J. D., and Ahmed, R. (2000). Differentiating ff between memory and effector ff CD8 T cells by altered expression of cell surface O-glycans. J Exp Med 191, 1241–6. Irie, H. Y., Mong, M. S., Itano, A., Crooks, M. E., Littman, D. R., Burakoff, ff S. J., and Robey, E. (1998). The cytoplasmic domain of CD8 beta regulates Lck kinase activation and CD8 T cell development. J Immunol 161, 183–91. Kern, P., Hussey, R. E., Spoerl, R., Reinherz, E. L., and Chang, H. C. (1999). Expression, purification, and functional analysis of murine ectodomain fragments of CD8aa and CD8ab dimers. J Biol Chem 274, 27237–43. Kern, P. S., Teng, M. K., Smolyar, A., Liu, J. H., Liu, J., Hussey, R. E., Spoerl, R., Chang, H. C., Reinherz, E. L., and Wang, J. H. (1998). Structural basis of CD8 coreceptor function revealed by crystallographic analysis of a murine CD8aa ectodomain fragment in complex with H-2Kb. Immunity 9, 519–30. Kim, P. W., Sun, Z. Y., Blacklow, S. C., Wagner, G., and Eck, M. J. (2003). A zinc clasp structure tethers Lck to T cell coreceptors CD4 and CD8. Science 301, 1725–8. Leishman, A. J., Naidenko, O. V., Attinger, A., Koning, F., Lena, C. J., Xiong, Y., Chang, H. C., Reinherz, E., Kronenberg, M., and Cheroutre, H. (2001). T cell responses modulated through interaction between CD8aa and the nonclassical MHC class I molecule, TL. Science 294, 1936–9. Li, S., Choksi, S., Shan, S., Hu, X., Gao, J., Korngold, R., and Huang, Z. (1998). Identification of the CD8 DE loop as a surface functional epitope. Implications for major histocompatibility complex class I binding and CD8 inhibitor design. J Biol Chem 273, 16442–5. Liu, Y., Xiong, Y., Naidenko, O. V., Liu, J. H., Zhang, R., Joachimiak, A., Kronenberg, M., Cheroutre, H., Reinherz, E. L., and Wang, J. H. (2003). The crystal structure of a TL/CD8aa complex at 2.1 A resolution: implications for modulation of T cell activation and memory. Immunity 18, 205-15. Luescher, I. F., Vivier, E., Layer, A., Mahiou, J., Godeau, F., Malissen, B., and Romero, P. (1995). CD8 modulation of T-cell antigen receptor-ligand interactions on living cytotoxic T lymphocytes. Nature 373, 353–6. Merry, A. H., Gilbert, R. J., Shore, D. A., Royle, L., Miroshnychenko, O., Vuong, M., Wormald, M. R., Harvey, D. J., Dwek, R. A., Classon, B. J., Rudd, P. M., and Davis, S. J. (2003). O-glycan sialylation and the structure of the stalk-like region of the T cell co-receptor CD8. J Biol Chem 278, 27119–28. Moody, A. M., Chui, D., Reche, P. A., Priatel, J. J., Marth, J. D., and Reinherz, E. L. (2001). Developmentally regulated glycosylation of the CD8ab coreceptor stalk modulates ligand binding. Cell 107, 501–12. Moody, A. M., North, S. J., Reinhold, B., Van Dyken, S. J., Rogers, M. E., Panico, M., Dell, A., Morris, H. R., Marth, J. D., and Reinherz, E. L. (2003). Sialic acid capping of CD8b core 1-O-glycans controls thymocyte-major histocompatibility complex class I interaction. J Biol Chem 278, 7240–6. Norment, A. M., and Littman, D. R. (1988). A second subunit of CD8 is expressed in human T cells. Embo J 7, 3433–9. Palacios, E. H., and Weiss, A. (2004). Function of the Src-family kinases, Lck and Fyn, in T-cell development and activation. Oncogene 23, 7990–8000. Pappu, B. P., and Shrikant, P. A. (2004). Alteration of cell surface sialylation regulates antigen-induced naive CD8+ T cell responses. J Immunol 173, 275–84. Piller, F., Piller, V., Fox, R. I., and Fukuda, M. (1988). Human T-lymphocyte activation is associated with changes in O-glycan biosynthesis. J Biol Chem 263, 15146–50.

84

D. A. Shore et al.

Priatel, J. J., Chui, D., Hiraoka, N., Simmons, C. J., Richardson, K. B., Page, D. M., Fukuda, M., Varki, N. M., and Marth, J. D. (2000). The ST3Gal-I sialyltransferase controls CD8+ T lymphocyte homeostasis by modulating O-glycan biosynthesis. Immunity 12, 273–83. Purbhoo, M. A., Boulter, J. M., Price, D. A., Vuidepot, A. L., Hourigan, C. S., Dunbar, P. R., Olson, K., Dawson, S. J., Phillips, R. E., Jakobsen, B. K., Bell, J. I., and Sewell, A. K. (2001). The human CD8 coreceptor effects ff cytotoxic T cell activation and antigen sensitivity primarily by mediating complete phosphorylation of the T cell receptor zeta chain. J Biol Chem 276, 32786–92. Rudd, P. M., Elliott, T., Cresswell, P., Wilson, I. A., and Dwek, R. A. (2001). Glycosylation and the immune system. Science 291, 2370–6. Rudd, P. M., Wormald, M. R., and Dwek, R. A. (2004). Sugar-mediated ligand-receptor interactions in the immune system. T rends Biotechnol 22, 524–30. Rudd, P. M., Wormald, M. R., Harvey, D. J., Devasahayam, M., McAlister, M. S., Brown, M. H., Davis, S. J., Barclay, A. N., and Dwek, R. A. (1999). Oligosaccharide analysis and molecular modeling of soluble forms of glycoproteins belonging to the Ly-6, scavenger receptor, and immunoglobulin superfamilies expressed in Chinese hamster ovary cells. Glycobiology 9, 443–58. Rudd, P. M., Wormald, M. R., Stanfield, R. L., Huang, M., Mattsson, N., Speir, J. A., DiGennaro, J. A., Fetrow, J. S., Dwek, R. A., and Wilson, I. A. (1999). Roles for glycosylation of cell surface receptors involved in cellular immune recognition. J M Mol Biol 293, 351–66. Shogren, R., Gerken, T. A., and Jentoft, N. (1989). Role of glycosylation on the conformation and chain dimensions of O-linked glycoproteins: light-scattering studies of ovine submaxillary mucin. Biochemistry 28, 5525–36. Starr, T. K., Daniels, M. A., Lucido, M. M., Jameson, S. C., and Hogquist, K. A. (2003). Thymocyte sensitivity and supramolecular activation cluster formation are developmentally regulated: a partial role for sialylation. J Immunol 171, 4512–20. Sun, J., and Kavathas, P. B. (1997). Comparison of the roles of CD8 aa and CD8 ab in interaction with MHC class I. J Immunol 159, 6077–82. van der Merwe, P. A., and Davis, S. J. (2003). Molecular interactions mediating T cell antigen recognition. Annu Rev Immunol 21, 659–84. Wang, J., and Reinherz, E. L. (2000). Structural basis of cell-cell interactions in the immune system. Curr Opin Struct Biol 10, 656–61. Witte, T., Spoerl, R., and Chang, H. C. (1999). The CD8b ectodomain contributes to the augmented coreceptor function of CD8ab heterodimers relative to CD8aa homodimers. Cell Immunol 191, 90–6. Wong, J. S., Wang, X., Witte, T., Nie, L., Carvou, N., Kern, P., and Chang, H. C. (2003). Stalk region of b-chain enhances the coreceptor function of CD8. J Immunol 171, 867–74. Wu, W., Harley, P. H., Punt, J. A., Sharrow, S. O., and Kearse, K. P. (1996). Identification of CD8 as a peanut agglutinin (PNA) receptor molecule on immature thymocytes. J Exp Med 184, 759–64. Zamoyska, R. (1994). The CD8 coreceptor revisited: one chain good, two chains better. Immunity 1, 243–6.

13

IMMUNOGENICITY OF CALRETICULIN-BOUND MURINE LEUKEMIA VIRUS GLYCOPROTEIN gp90

Yusuke Mimura1,2, Denise Golgher2, Yuka Y Mimura-Kimura1, Raymond A. Dwek1, Pauline M. Rudd1, and Tim Elliott2 1Glycobiology Institute Department of Biochemistry University of Oxford South Parks Road, Oxford, OX1 3QU, UK 2Cancer Sciences Division School of Medicine University of Southampton Tremona Road, Southampton, SO16 6YD, UK

1. INTRODUCTION Class II molecules of the major histocompatibility complex (MHC class II molecules) bind peptides derived from protein antigens delivered into endocytic compartments and present these peptides to CD4+ T cells (Cresswell, 1994; Germain, 1994). In the endocytic pathway antigens are unfolded and cleaved into fragments during transport through the increasingly acidic endosomal network, from early endosomes to lysosomes. Newly synthesized MHC class II molecules associated with the invariant chains (Ii) are transported to a late endocytic compartment where the Ii is removed by proteolytic cleavage. MHC class II molecules bind polypeptide antigens by removal of the Ii-derived peptide CLIP (class II invariant chain-derived peptide) through a peptide-exchange process that is catalyzed by MHC-encoded DM molecules (Fig. 1) (Denzin and Cresswell, 1995; Sherman et al., 1995; Sloan et al., 1995; Wolf and Ploegh, 1995). The initial form of antigen that binds to class II molecules may be short peptides 10-20 amino acids in length, generated by acid endopeptidases (Watts, 2001). This pathway could be termed ‘‘cut/trim first, bind later’’ model and became the paradigm for the binding of peptides to MHC class II molecules. However, there is an alternative pathway independent of DM molecules that involves early capture of unfolded, extended sequence by MHC class II molecules (Pinet et al., 1994; Pinet et al., 1995; Sercarz et al., 1993). Subsequent trimming of 85 John S. Axford (ed.), Glycobiology and Medicine, 85-94. © 2005 Springer. Printed in the Netherlands.

86

Y. Mimura et al.

Figure 1. Overview of the MHC class II processing and presentation pathway.

Figure 2. Two postulated pathways for antigen presentation by MHC class II molecules.

the MHC-bound peptide would give rise to a short peptide epitope that roughly corresponded to the ‘‘footprint’’ of the MHC binding site. The ability of MHC class II molecules to bind extended polypeptide sequences is facilitated by the fact that its peptide binding groove is open at each end, unlike the MHC class I binding groove which terminates with deep pockets at either end. This could be termed ‘‘bind first, cut/trim later’’ model (Fig. 2) (Sercarz and Maverakis, 2003). Although this latter pathway is an attractive proposition with respect to epitope selection in an aggressive proteolytic environment, whether this model is the canonical mode of operation remains unclear (Watts, 2004) because this model has been investigated mostly using well-defined hen egg lysozyme (HEL) system whose immunodominant determinant HEL(52-61) can be presented by the DM-dependent classical pathway (Castellino et al., 1998; Lindner and Unanue, 1996; Pinet et al., 1995). This chapter will be concerned with presentation by MHC class II molecules of unique tumor

Immunogenicity of Leukemia Virus Glycoprotein gp90

87

antigen gp90 (Huang et al., 1996) whose antigenicity is dependent on the conformation and calreticulin (CRT) binding (Golgher et al., 2001). We will review evidence for the ‘‘bind first’’ model and describe how this notion is related to the conformational dependence of gp90 antigenicity. As CD4+ T cells play an important role in anti-tumor immune response (Pardoll and Topalian, 1998), our findings may form a basis for the development of cancer vaccine design by manipulating the conformation of tumor antigens. An exogenous antigen is taken up by an antigen presenting cell (APC) through endocytosis, unfolded and cleaved during transport from early endosomes to lysosomes by acid endoproteases including cathepsin D and cathepsin S (Cat). Nascent MHC class II molecules are associated with Ii trimers in the endoplasmic reticulum (ER). This complex with CLIP in the class II binding site is competent for exit from the ER and moves through the Golgi apparatus to the trans-Golgi network (TGN). Most of MHC class II - Ii complexes traffic directly to later endocytic locations such as MHC class II loading compartments (MIIC) and lysosomes from the TGN (Pathway (1)), others traffic to lysosomes after internalization from the plasma membrane (Pathway (2)) and from the TGN to early endosomes before their transport to lysosomes (Pathway (3)), with the relative contribution of each pathway varying from APC to APC (Hiltbold and Roche, 2002). The exchange of CLIP for more stable antigenic peptides is catalyzed by the DM molecules. The bound peptides undergo proteolytic trimming to form a final size of 15 to 20 residues. The peptideloaded MHC class II molecules are transported to the cell surface to be recognized by CD4+ T cells. Pre-existing surface peptide - MHC class II complexes can also internalize and recycle through endosomes, where peptides can be exchanged in a DM-independent manner (Pathway (4)).

2. INFLUENCE OF ANTIGEN CONFORMATION ON PRESENTATION BY MHC CLASS II MOLECULES Since the original studies of T cell recognition of protein antigen where recognition of denatured antigen was shown to be as efficient as native antigen (Chesnut et al., 1980), it has long been assumed that antigen conformation is not an important factor in determining antigenicity because T cells generally recognize protein antigens that have been processed and subsequently bound to MHC class II molecules. The processing requirement could reflect the incapacity of native proteins to interact with MHC molecules and form MHC-antigenic peptide complexes. However, many studies have shown that MHC class II molecules can engage peptides of much greater length than those typically eluted from MHC class II molecules and that this can occur during normal processing. Sette et al. (1989) detected positive binding of urea-denatured and reduced versions of ovalbumin, bovine serum albumin, HEL, and transferrin to appropriate MHC class II molecules but not for their native forms, indicating that an unfolding step by exposure to low pH or proteolysis in endocytic compartments is a prerequisite for MHC class II binding to antigens. Although such tightly folded globular molecules do not bind to MHC class II molecules, a native protein molecule as large as fibrinogen (340 kDa) has been shown to bind to prefixed APCs in the absence of processing. The epitope recognized was localized to a

88

Y. Mimura et al.

region of the molecule with conformational flexibility that gained access to MHC class II peptide-binding grooves (Lee et al., 1988). Furthermore, some intact proteins such as HEL, beef insulin and pigeon cytochrome c have been shown to be presented to specific T cells by MHC class II molecules bound at low pH in the presence or absence of reducing agent without a requirement of proteolytic cleavage or irreversible denaturation (Jensen, 1993). These results indicate the ability of MHC class II molecules to bind and present intact or even native proteins to CD4+ T cells, depending on determinant accessibility. As a noticeable consequence of class II capture of unfolding antigens, the immunogenic T cell determinants could become proteolytically inaccessible. This protection hypothesis has been tested using a 34 amino acid HEL peptide containing the high-affinity I-Ak binding site HEL(52-61) in the core with extension of 12 unnatural D-amino acids at both sides so that this peptide would be resistant to proteolysis except in the core region (Donermeyer and Allen, 1989). The authors showed that the determinant became resistant to chymotrypsin digestion if HEL(40-73) was allowed to first bind to I-Ak, otherwise it was totally destroyed. The protection hypothesis suggests that MHC class II binding to a large antigen fragment can increase the possibility to present T cell determinants.

3. ANTIGEN PRESENTATION BY TWO DISTINCT POPULATIONS OF MHC CLASS II MOLECULES In addition to the classical DM-dependent pathway, recycling of cell-surface MHC class II molecules has been suggested as a pathway for presentation of some antigens including RNase A, influenza virus hemagglutinin and myelin basic protein (Adorini et al., 1989; Harding and Unanue, 1989; Nadimi et al., 1991; Pinet et al., 1994; Pinet et al., 1995). Cell surface MHC class II molecules have been shown to internalize and return to the cell surface rapidly. Pinet et al. (1995) have reported that truncation of either one of the _ or _ cytoplasmic tails of HLA-DR molecules eliminates internalization of HLA-DR and presentation of influenza hemagglutinin while Ii-dependent presentation of matrix antigen from the same virus particles is unaffected ff by the truncations. Thus, the compartment for peptide loading on recycling class II molecules seemed likely to be distinct from the one where newly synthesized class II molecules are transported. This alternative processing pathway was also described with the finding that intact partially-folded HEL forms stable complexes with mature I-Ak molecules in low pH compartments independently of proteases and DM (Lindner and Unanue, 1996). By subcellular fractionation it has been shown that early endosomes generate RNase(42-56)-I-Ak complexes by the DM-independent processing pathway (Griffin et al., 1997). Zhong et al. (1997) addressed the involvement of newly synthesized (Ii-associated) versus mature class II (Ii-free) in effective ff presentation of distinct determinants in HEL by truncations of leucine-based cytoplasmic tails in Ii and MHC class II molecules, demonstrating that a requirement for MHC class II internalization is inversely correlated with a requirement for Ii expression. These results suggest that the classical pathway requires Ii and DM expression for peptide loading onto newly synthesized MHC class II molecules in late endocytic compartments whereas the alternative pathway utilizes recycling, mature MHC class II molecules independently of protein synthesis, Ii and

Immunogenicity of Leukemia Virus Glycoprotein gp90

89

DM for peptide loading in earlier compartments than the classical pathway. In addition to recycling mature MHC class II molecules, Ii-free MHC class II molecules have been shown to be generated by removal of Ii in an early endosomal compartment without involvement of cathepsin S or DM, allowing the direct binding to high molecular-weight polypeptides (Villadangos et al., 2000). The existence of multiple endocytic compartments for peptide loading would be advantageous since each compartment would provide a unique environment to accommodate the processing of a wide variety of class II-restricted epitopes. Some epitopes may be revealed in early endosomes and destroyed before reaching later compartments, whereas others may require the more active proteolytic environment of late endosomes and lysosomes for release. The simultaneous presence of these mechanisms increases the possiblilties of displaying foreign epitopes that will be recognized by CD4+ T cells.

4. MHC-GUIDED PROCESSING AND IMMUNODOMINANCE ‘‘MHC-guided processing’’ was originally hypothesized by Sercarz et al. (1993) and a consequence of ‘‘bind first, trim later’’ model. The notion is closely related to immunodominance and stresses the importance of intramolecular competition between the multiple determinants on a single long peptide for binding to an MHC molecule (Deng et al., 1993). Protein antigens typically contain multiple epitopes capable of binding to MHC class II molecules, yet T cell responses are limited to only a small number of these determinants. The ability of the immune system to regulate and focus T cell responses to a select number of epitopes is termed immunodominance. The hierarchy of T cell responses observed in vivo, namely dominant, subdominant, and cryptic (silent), reflects in part this selective presentation of epitopes by MHC class II molecules, as well as the influences of T cell responsiveness and repertoire. Biochemical and functional studies of MHC class II molecules have revealed the preferential display of immunodominant epitopes, with conversely lower levels of MHC-restricted presentation of subdominant or cryptic peptides derived from the same Ag (Ma et al., 1999; Nelson et al., 1992; Viner et al., 1995). The specific reactions within APC which influence epitope selection remain poorly defined, with both Ag processing and MHC binding potentially playing key roles. Early capture of unfolded, extended sequence by an MHC class II molecule outlined by the notion of MHC-guided processing is an attractive proposition for the epitope selection in an aggressive proteolytic environment. Castellino et al. (1998) showed, using the HEL model, that large complexes of 120 kDa could be found in the endocytic pathway that comprised a single HEL polypeptide chain of about 70 amino acids, bound to two different ff MHC class II isotypes, I-Ak and I-Ek. This clearly shows that binding of MHC class II molecules can occur before excessive antigen processing. HIV envelope glycoprotein (gp140) epitopes recognized by HIVspecific CD4+ T cells have been found to be located in exposed, non-helical loops or strands on one face of the molecule (Surman et al., 2001). Presumably, MHCguided processing may play a role in the selection of immunodominant gp140 epitopes.

90

Y. Mimura et al.

5. INCOMPLETE FOLDING IN THE ER CAN ALTER ANTIGENICITY OF A GLYCOPROTEIN Golgher et al. (2001) have reported an interesting MHC class II-restricted tumor antigen, endogenous murine leukemia virus envelope protein gp90 (Fass et al., 1997; Lenz et al., 1982; Pinter and Fleissner, 1977), that is expressed in many mouse tumor lines while antigenicity varies among the cell lines. gp90 is synthesized in the endoplasmic reticulum (ER) as a precursor protein, where its amino-terminal signal sequence is cleaved. The polypeptide then undergoes disulfide bonding and is modified by the addition and processing of multiple asparagine-linked high mannose-type sugars, and oligomerizes, prior to transport to the Golgi apparatus. In the Golgi apparatus, the high mannose-type sugars are modified to complex-type, and the precursor protein polypeptide backbone is cleaved to form an extracellular glycoprotein gp70 and a transmembrane protein p15E. p15E anchors gp70 to the plasma membrane by way of noncovalent interactions and in some instances, a disulfide bond (Gliniak et al., 1991; Pinter et al., 1978). CD4+ T cell hybridoma clones specific for gp90 were obtained from mice immunized with colon adenocarcinoma cell line CT26 genetically engineered to secrete granulocyte/macrophage colonystimulating factor. Importantly, the recognition of gp90 by the CD4+ T cell hybridomas is strictly dependent on the conformation of gp90 (Golgher et al., 2001). Unfolding of the protein by disulfide bond reduction and alkylation markedly reduces antigenicity, and thermal denaturation abrogates T cell recognition, which is different ff from presentation of other model antigens such as HEL and ovalbumin described above. Furthermore, the T cell hybridomas do not respond to its mature form, gp70. Interestingly, the gp90 from other tumor lines of the same histological origin is not antigenic although the nucleotide sequence of their gp90 cDNAs was identical (unpublished data). In CT26 cells gp90 is found to be retained in the ER, associated mostly with the ER chaperone calreticulin (CRT) that recognizes monoglucosylated high mannose-type oligosaccharides (Glc1Man5-9GlcNAc ) on nascent glycopro2 teins (Helenius and Aebi, 2001; Parodi, 2000; Sitia and Braakman, 2003; Trombetta, 2003). It was noted that the CRT-bound gp90 is highly antigenic (Golgher et al., 2001). Incomplete folding of antigenic gp90 is evidenced by the presence of the CRTligand, Glc1Man9GlcNAc , in the antigenic gp90 of CT26 (YM, manuscript in 2 preparation). Why is gp90 presentation dependent on the conformation? Generally, denaturation and unfolding have a minimum effect ff on antigen presentation by MHC class II molecules (Allen and Unanue, 1984; Streicher et al., 1984) because an intact protein antigen eventually is always recognized as a denatured protein or fragment, as a result of intracellular processing by an APC. Therefore, it is very likely that the gp90 epitope is susceptible to proteolysis, i.e., destructive processing. Trypsin digestion abrogated gp90 presentation by pre-fixed APCs, in contrast to ovalbumin where prior tryptic digestion liberates the immunodominant epitope. However, if native gp90 is incubated with pre-fixed APCs, the epitope became resistant to trypsin and is successfully presented to the T cell hybridoma (YM, manuscript in preparation). Therefore, the gp90 epitope needs to be bound and protected by MHC class II molecules before processing. Furthermore, gp90 presentation was not affected ff by leupeptin or chloroquine treatment of APCs, contrary to ovalbumin (YM, manuscript

Immunogenicity of Leukemia Virus Glycoprotein gp90

91

in preparation), suggesting that gp90 binds to recycling mature MHC class II molecules in a DM-independent manner. Taken together, gp90 presentation is consistent with the alternative pathway involving MHC-guided processing. It should be noted that differential ff processing by APCs of the same glycoprotein can occur, depending on the folding status of the glycoprotein and cell lines that synthesize it. This phenomenon cannot be under-evaluated because MHC-guided processing can convert a cryptic epitope into an immunodominant one by protecting the determinant from destructive processing. Such a situation may permit induction of an anti-tumor immune response as well as an autoimmune reaction by the presentation of ‘‘self ’’ epitopes to which tolerance has not been established. The mechanism by which incomplete folding of a glycoprotein is promoted in CT26 cells remains unclear although oxidoreductases (e.g., ERp72, ERp57, PDI, Ero-1), ER molecular chaperones (e.g., Bip, CRT, CNX) or differential ff N-glycosylation site occupancy may be involved. If conformational changes of a glycoprotein antigen could be induced by modulating the activities of oxidoreductases in the ER, it would allow for manipulation of hierarchy of T cell response to the antigen, which may form a basis for the improvement of cancer vaccine design.

6. THE ROLE OF GLYCOSYLATION IN INFLUENCING MHC CLASS II-RESTRICTED ANTIGEN PROCESSING The extent to which MHC class II epitopes are generated via a bind/trim versus a trim/bind pathway will depend on a number of factors including competition between MHC class II and destructive proteases for a relatively short stretch of polypeptide within a larger precursor protein. Accessibility of this site to the two mutually exclusive activities may in turn depend upon the route of antigen uptake and the timing of its subsequent encounters with processing factors such as oxidoreductases and hydrolases. The presence of post-translational modifications carried by protein antigens could clearly affect ff any of these events. Although formal proof has yet to be provided for the direct influence of glycosylation on the outcome of antigen processing, compelling indirect evidence exists linking N-linked glycosylation to the antigenicity of glycoproteins for T cell responses directed towards non-glycopeptide epitopes. Thus Sjolander et al. (1996) demonstrated that HIV-1 gp160 lacking three N-linked glycans in its C-terminal CD4-binding region were unable to induce in mice T cell responses to non-glycosylated epitopes in this region of the molecule; whereas the fully glycosylated gp160 could. One explanation for this observation is that the physical location of the peptide epitope within the native protein leads to differential ff processing and consequent epitope selection and furthermore that proximal glycosylation may influence this process – perhaps by protecting regions of the molecule from destructive processing. Consistent with this idea is the observation that extensive T cell epitope mapping of the gp160 glycoprotein has shown that epitope clustering occurs in four hotspots that comprise relatively short sequences that were bordered by regions of heavy glycosylation on exposed strands (Surman et al., 2001). In other studies, the generation of MHC class II-restricted epitopes from tumor antigens (tyrosinase and gp90 described above) has been shown to be dependent on these antigens being glycosylated (Golgher et al., 2001; Housseau et al.,

92

Y. Mimura et al.

2001). Both these studies provide indirect evidence that the T cell epitope is not glycosylated. In the former case, site-directed mutagenesis of each of seven potential N-glycosylation sequons showed that four sites were required to generate ‘‘immunogenic’’ tyrosinase. In the latter example, it was shown that other structure perturbations to gp90 such as reduction and thermal denaturation (which did not affect ff glycosylation) also led to loss of antigenicity – consistent with the notion that denaturation renders the protein susceptible to ‘‘overprocessing’’.

REFERENCES Adorini, L., Appella, E., Doria, G., Cardinaux, F. and Nagy, Z.A. (1989) Competition for antigen presentation in living cells involves exchange of peptides bound by class II MHC molecules. Nature, 342, 800–803. ff requirements for antigen processing by macrophages for Allen, P.M. and Unanue, E.R. (1984) Differential lysozyme-specific T cell hybridomas. J Immunol, 132, 1077–1079. Castellino, F., Zappacosta, F., Coligan, J.E. and Germain, R.N. (1998) Large protein fragments as substrates for endocytic antigen capture by MHC class II molecules. J Immunol, 161, 4048–4057. Chesnut, R.W., Endres, R.O. and Grey, H.M. (1980) Antigen recognition by T cells and B cells: recognition of cross-reactivity between native and denatured forms of globular antigens. Clin Immunol Immunopathol, 15, 397–408. Cresswell, P. (1994) Assembly, transport, and function of MHC class II molecules. Annu Rev Immunol, 12, 259–293. Deng, H., Apple, R., Clare-Salzler, M., Trembleau, S., Mathis, D., Adorini, L. and Sercarz, E. (1993) Determinant capture as a possible mechanism of protection afforded ff by major histocompatibility complex class II molecules in autoimmune disease. J Exp Med, 178, 1675–1680. Denzin, L.K. and Cresswell, P. (1995) HLA-DM induces CLIP dissociation from MHC class II alpha beta dimers and facilitates peptide loading. Cell, 82, 155–165. Donermeyer, D.L. and Allen, P.M. (1989) Binding to Ia protects an immunogenic peptide from proteolytic degradation. J Immunol, 142, 1063–1068. Fass, D., Davey, R.A., Hamson, C.A., Kim, P.S., Cunningham, J.M. and Berger, J.M. (1997) Structure of a murine leukemia virus receptor-binding glycoprotein at 2.0 angstrom resolution. Science, 277, 1662–1666. Germain, R.N. (1994) MHC-dependent antigen processing and peptide presentation: providing ligands for T lymphocyte activation. Cell, 76, 287–299. Gliniak, B.C., Kozak, S.L., Jones, R.T. and Kabat, D. (1991) Disulfide bonding controls the processing of retroviral envelope glycoproteins. J Biol Chem, 266, 22991–22997. ff E., Edidin, M., Pardoll, D.M. and Elliott, T. (2001) Golgher, D., Korangy, F., Gao, B., Gorski, K., Jaffee, An immunodominant MHC class II-restricted tumor antigen is conformation dependent and binds to the endoplasmic reticulum chaperone, calreticulin. J Immunol, 167, 147–155. Griffin, J.P., Chu, R. and Harding, C.V. (1997) Early endosomes and a late endocytic compartment generate different ff peptide-class II MHC complexes via distinct processing mechanisms. J Immunol, 158, 1523–1532. Harding, C.V. and Unanue, E.R. (1989) Antigen processing and intracellular Ia. Possible roles of endocytosis and protein synthesis in Ia function. J Immunol, 142, 12–19. Helenius, A. and Aebi, M. (2001) Intracellular functions of N-linked glycans. Science, 291, 2364–2369. Hiltbold, E.M. and Roche, P.A. (2002) Trafficking of MHC class II molecules in the late secretory pathway. Curr Opin Immunol, 14, 30–35. Housseau, F., Moorthy, A., Langer, D.A., Robbins, P.F., Gonzales, M.I. and Topalian, S.L. (2001) N-linked carbohydrates in tyrosinase are required for its recognition by human MHC class II-restricted CD4(+) T cells. Eur J Immunol, 31, 2690–2701. Huang, A.Y., Gulden, P.H., Woods, A.S., Thomas, M.C., Tong, C.D., Wang, W., Engelhard, V.H., Pasternack, G., Cotter, R., Hunt, D., Pardoll, D.M. and Jaffee, ff E.M. (1996) The immunodominant

Immunogenicity of Leukemia Virus Glycoprotein gp90

93

major histocompatibility complex class I-restricted antigen of a murine colon tumor derives from an endogenous retroviral gene product. Proc Natl Acad Sci USA, 93, 9730–9735. Jensen, P.E. (1993) Acidification and disulfide reduction can be sufficient to allow intact proteins to bind class II MHC. J Immunol, 150, 3347–3356. Lee, P., Matsueda, G.R. and Allen, P.M. (1988) T cell recognition of fibrinogen. A determinant on the A alpha-chain does not require processing. J Immunol, 140, 1063–1068. Lenz, J., Crowther, R., Straceski, A. and Haseltine, W. (1982) Nucleotide sequence of the Akv env gene. J V irol, 42, 519–529. Lindner, R. and Unanue, E.R. (1996) Distinct antigen MHC class II complexes generated by separate processing pathways. Embo J, 15, 6910–6920. Ma, C., Whiteley, P.E., Cameron, P.M., Freed, D.C., Pressey, A., Chen, S.L., Garni-Wagner, B., Fang, C., Zaller, D.M., Wicker, L.S. and Blum, J.S. (1999) Role of APC in the selection of immunodominant T cell epitopes. J Immunol, 163, 6413–6423. Nadimi, F., Moreno, J., Momburg, F., Heuser, A., Fuchs, S., Adorini, L. and Hammerling, G.J. (1991) Antigen presentation of hen egg-white lysozyme but not of ribonuclease A is augmented by the major histocompatibility complex class II-associated invariant chain. Eur J Immunol, 21, 1255–1263. Nelson, C.A., Roof, R.W., McCourt, D.W. and Unanue, E.R. (1992) Identification of the naturally processed form of hen egg white lysozyme bound to the murine major histocompatibility complex class II molecule I-Ak. Proc Natl Acad Sci USA, 89, 7380–7383. Pardoll, D.M. and Topalian, S.L. (1998) The role of CD4+ T cell responses in antitumor immunity. Curr Opin Immunol, 10, 588–594. Parodi, A.J. (2000) Protein glucosylation and its role in protein folding. Annu Rev Biochem, 69, 69–93. Pinet, V., Malnati, M.S. and Long, E.O. (1994) Two processing pathways for the MHC class II-restricted presentation of exogenous influenza virus antigen. J Immunol, 152, 4852–4860. Pinet, V., Vergelli, M., Martin, R., Bakke, O. and Long, E.O. (1995) Antigen presentation mediated by recycling of surface HLA-DR molecules. Nature, 375, 603–606. Pinter, A. and Fleissner, E. (1977) The presence of disulfide-linked gp70-p15(E) complexes in AKR murine leukemia virus. V irology, 83, 417–422. Pinter, A., Lieman-Hurwitz, J. and Fleissner, E. (1978) The nature of the association between the murine leukemia virus envelope proteins. V irology, 91, 345–351. Sercarz, E.E., Lehmann, P.V., Ametani, A., Benichou, G., Miller, A. and Moudgil, K. (1993) Dominance and crypticity of T cell antigenic determinants. Annu Rev Immunol, 11, 729–766. Sercarz, E.E. and Maverakis, E. (2003) Mhc-guided processing: binding of large antigen fragments. Nat Rev Immunol, 3, 621–629. Sette, A., Adorini, L., Colon, S.M., Buus, S. and Grey, H.M. (1989) Capacity of intact proteins to bind to MHC class II molecules. J Immunol, 143, 1265–1267. Sherman, M.A., Weber, D.A. and Jensen, P.E. (1995) DM enhances peptide binding to class II MHC by release of invariant chain-derived peptide. Immunity, 3, 197–205. Sitia, R. and Braakman, I. (2003) Quality control in the endoplasmic reticulum protein factory. Nature, 426, 891–894. Sjolander, S., Bolmstedt, A., Akerblom, L., Horal, P., Olofsson, S., Morein, B. and Sjolander, A. (1996) N-linked glycans in the CD4-binding domain of human immunodeficiency virus type 1 envelope glycoprotein gp160 are essential for the in vivo priming of T cells recognizing an epitope located in their vicinity. V irology, 215, 124–133. Sloan, V.S., Cameron, P., Porter, G., Gammon, M., Amaya, M., Mellins, E. and Zaller, D.M. (1995) Mediation by HLA-DM of dissociation of peptides from HLA-DR. Nature, 375, 802–806. Streicher, H.Z., Berkower, I.J., Busch, M., Gurd, F.R. and Berzofsky, J.A. (1984) Antigen conformation determines processing requirements for T-cell activation. Proc Natl Acad Sci USA, 81, 6831–6835. Surman, S., Lockey, T.D., Slobod, K.S., Jones, B., Riberdy, J.M., White, S.W., Doherty, P.C. and Hurwitz, J.L. (2001) Localization of CD4+ T cell epitope hotspots to exposed strands of HIV envelope glycoprotein suggests structural influences on antigen processing. Proc Natl Acad Sci USA, 98, 4587–4592. Trombetta, E.S. (2003) The contribution of N-glycans and their processing in the endoplasmic reticulum to glycoprotein biosynthesis. Glycobiology, 13, 77R-91R. Villadangos, J.A., Driessen, C., Shi, G.P., Chapman, H.A. and Ploegh, H.L. (2000) Early endosomal maturation of MHC class II molecules independently of cysteine proteases and H-2DM. Embo J, 19, 882–891.

94

Y. Mimura et al.

Viner, N.J., Nelson, C.A. and Unanue, E.R. (1995) Identification of a major I-Ek-restricted determinant of hen egg lysozyme: limitations of lymph node proliferation studies in defining immunodominance and crypticity. Proc Natl Acad Sci USA, 92, 2214–2218. Watts, C. (2001) Antigen processing in the endocytic compartment. Curr Opin Immunol, 13, 26–31. Watts, C. (2004) The exogenous pathway for antigen presentation on major histocompatibility complex class II and CD1 molecules. Nat Immunol, 5, 685–692. Wolf, P.R. and Ploegh, H.L. (1995) How MHC class II molecules acquire peptide cargo: biosynthesis and trafficking through the endocytic pathway. Annu Rev Cell Dev Biol, 11, 267–306. Zhong, G., Romagnoli, P. and Germain, R. N. (1997) Related leucine-based cytoplasmic targeting signals in invariant chain and major histocompatibility complex class II molecules control endocytic presentation of distinct determinants in a single protein. J Exp Med, 185, 429–438.

14

GLYCOSYLATION AND GPI ANCHORAGE OF THE PRION PROTEIN

N. M. Hooper School of Biochemistry and Microbiology Leeds Institute for Genetics Health and Therapeutics University of Leeds, Leeds, UK

Prion diseases or transmissible spongiform encephalopathies are a group of neurodegenerative disorders including scrapie in sheep, bovine spongiform encephalopathy in cattle, Creutzfeldt–Jakob disease and Gerstmann–Straussler–Scheinker disease in humans. In prion diseases the normal cellular form of the prion protein (PrPC) undergoes a conformational conversion to the b-sheet-rich scrapie isoform (PrPSc). Although PrPC is critical for the development of prion disease through its conversion into PrPSc, the physiological role of PrPC is less clear. PrPC undergoes a variety of post-translational processing events, including glycosylation, GPI anchorage and

Figure 1. 95 John S. Axford (ed.), Glycobiology and Medicine, 95-96. © 2005 Springer. Printed in the Netherlands.

96

N. M. Hooper

proteolysis. PrP contains two N-glycosylation sequons at Asn180 and Asn196, both of which can be glycosylated. However, either one or both of the acceptor Asn may remain unglycosylated, giving rise to unglycosylated, mono-glycosylated and di-glycosylated proteins. The ratio of these three glycoforms appears to be characteristic for particular ‘strains’ of PrP. N-glycosylation of PrP is dramatically influenced by its membrane topology and by the distance of the Asn sequons to the C-terminus of the protein. The role of these post-translational modifications in the life cycle of PrPC will be discussed. The C-terminal signal peptide directs the addition of a glycosyl-phosphatidylinositol (GPI) anchor to the protein within the lumen of the ER. This GPI anchor, along with a determinant in the N-terminal region of the protein, promotes the association of PrPC with cholesterol-rich lipid rafts that are involved in the trafficking of the protein and its conversion to PrPSc.

REFERENCES 1. Walmsley, A. R., Zeng, F. and Hooper, N. M. (2001). Membrane topology influences N-glycosylation of the prion protein. EMBO J. 20, 703–712. 2. Walmsley, A. R. and Hooper, N. M. (2003). Distance of sequons to the C-terminus influences the cellular N-glycosylation of the prion protein. Biochem. J. 370, 351–355. 3. Walmsley, A. R., Zeng, F. and Hooper, N. M. (2003). The N-terminal region of the prion protein ectodomain contains a lipid raft targeting determinant. J. Biol. Chem. 278, 37241–37248.

15

GLYCOSYLATION DEFECTS AND MUSCULAR DYSTROPHY

Derek J. Blake1, Christopher T. Esapa1, Enca Martin-Rendon2, and R. A. Jeffrey ff McIlhinney3 1Department of Pharmacology, University of Oxford. Mansfield Road, Oxford, UK 2Stem Cell Research Laboratory, National Blood Service, Oxford Centre, Oxford, UK 3MRC Anatomical Neuropharmacology Unit, University of Oxford, Oxford, UK

Glycosylation is an important post-translational modification of many proteins in the secretory pathway. Mutations in several genes involved in glycan metabolism are known to cause different ff types of congenital disorders of glycosylation; a genetically heterogeneous group of diseases that affect ff multiple organs and are frequently associated with developmental delay, haematological and immunological anomalies. In addition to these disorders, it is now apparent that at least six different ff forms of muscular dystrophy are caused by genes that encode actual or putative glycosyltransferases. Although these diseases are clinically distinct, they are all associated with a secondary deficiency in the glycosylation of a-dystroglycan. Hypo-glycosylation of a-dystroglycan disrupts a link between the membrane and proteins in the extracellular matrix such as laminin, resulting in muscle disease and in several cases a neuronal migration disorder. At least three allelic disorders are caused by mutations in the gene encoding fukutin-related protein (FKRP). These are; congenital muscular dystrophy type 1C (MDC1C), limb girdle muscular dystrophy 2I (LGMD2I) and congenital muscular dystrophy (CMD) with brain malformations and mental retardation. These diseases result from any one of 36 different ff missense mutations in the gene encoding FKRP. FKRP is a DxD motif-containing type II membrane protein that is targeted to the medial Golgi-apparatus by an N-terminal signal anchor sequence. We have found that FKRP mutations associated with the most severe disease phenotypes result in retention of the mutant protein in the endoplasmic reticulum (Figure 1). The ER-retained mutants have a prolonged association with calenxin and are preferentially degraded by the proteasome. These data suggest that 97 John S. Axford (ed.), Glycobiology and Medicine, 97-98. © 2005 Springer. Printed in the Netherlands.

98

D. J. Blake et al.

Figure 1.

impaired intracellular trafficking and proteasomal degradation contribute to the cellular pathology of CMD caused by mutations in the FKRP gene.

REFERENCES 1. Brockington M et al., (2001) Am J. Hum. Genet. 69:1198–209. 2. Esapa CT et al., (2002) Hum. Mol. Genet. 11: 3319–3331. 3. Martin-Rendon E and Blake DJ (2003) T rends Pharm. Sci. 24: 178–183.

16

ROLES OF COMPLEX AND HYBRID N-GLYCANS AND O-FUCOSE GLYCANS IN OOCYTE DEVELOPMENT AND FUNCTION

S. Shi, S. A. Williams, H. Kurniawan, L. Lu, and P. Stanley Department of Cell Biology Albert Einstein College Medicine New York, NY 10461, USA

Roles for complex or hybrid N-glycans in oocyte maturation and function were investigated using female mice with a floxed Mgat1 gene (Mgat1F) carrying a Crerecombinase transgene under the control of the zona pellucida protein 3 (ZP3) promoter (1). Inactivation of the Mgat1 gene responsible for the synthesis of complex and hybrid N-glycans is embryonic lethal, but homozygous mutant blastocysts are rescued by maternal Mgat1 gene transcripts. Following deletion of the Mgat1 gene in oocytes, females with mutant oocytes had reduced fertility and fewer oocytes after superovulation. All mutant oocytes had a zona pellucida (ZP) that contained ZP1, ZP2 and ZP3 glycoproteins but they did not contain complex N-glycans and mutant ZP were thin, deformed and loosely attached (Fig. 1). Nevertheless, mutant oocytes were efficiently fertilized, all embryos implanted and ~70% developed to E9.5 (Mgat1−/−) or birth (Mgat1+/− or Mgat1+/+). However, ~40%–50% embryos were

Figure 1. Mgat1+/+ oocyte (left); Mgat1−/− oocyte with ZP (right). 99 John S. Axford (ed.), Glycobiology and Medicine, 99-100. © 2005 Springer. Printed in the Netherlands.

100

S. Shi et al.

retarded in development to various degrees at E3.5. Surprisingly, heterozygous blastocysts that synthesized complex N-glycans were both retarded and normal in morphology at E3.5 indicating that a proportion of Mgat1−/− oocytes are unable to support development of a zygote, even when they synthesize complex and hybrid N-glycans. Additionally unexpected, was the finding that Mgat1−/− zygotes could develop normally during pre-implantation, implant and progress to E9.5 in the absence of complex and hybrid N-glycans. Therefore sugars on complex or hybrid N-glycans are required at some stage of oogenesis for the generation of a developmentally competent oocyte, but fertilization, pre-implantation development and implantation may proceed in their absence. When the Pofut1 f gene was deleted in oocytes using the ZP3Cre transgene, a similar result was obtained. The Pofut1 f gene is responsible for initiating the synthesis of O-fucose glycans found on certain EGF repeats of Notch receptors and their ligands (2). Deletion of Pofut1 f gives embryonic lethality and a phenotype characteristic of embryos with severe Notch signaling defects (3). After deletion in oocytes, while ZP and egg morphology were normal, about 50% females had a reduced litter size. These females were found to have a high proportion of malformed and under-developed embryos at E3.5, although the total number of pre-implantation embryos was normal. Therefore a loss of Pofut1 in oocytes did not appear to inhibit oocyte maturation, ovulation or fertilization but generated a majority of oocytes that were unable to facilitate development to normal blastocysts following fertilization. The implications of these findings for the roles of Notch signaling in oocyte development will be discussed.

REFERENCES (1) Ye, Z. and Marth J. D. (2004) Glycobiology, 14, 547–558. (2) W Wang, Y., Shao, L., Shi, S., Harris, R. J., Spellman, M. W., Stanley, P., Haltiwanger, R. S. (2001) J. Biol Chem. 276, 40338–40345. (3) Shi, S. and Stanley, P. (2003) Proc. Natl. Acad. Sci. USA., 100, 5234–5239.

17

MUCIN OLIGOSACCHARIDES AND PIGEON FANCIERS’ LUNG

C. I. Baldwin1, A. Allen2, S. Bourke3, E. Hounsell4, and J. E. Calvert2 1School of Applied Sciences Northumbria University Newcastle upon Tyne, UK 2School of Cell and Molecular Biosciences Newcastle University Newcastle upon Tyne, UK 3Department of Respiratory Medicine Royal Victoria Infirmary Newcastle upon Tyne, UK 4School of Biological and Chemical Sciences Birkbeck, University of London, UK

Pigeon fanciers’ lung (PFL), a form of extrinsic allergic alveolitis, is an immunologically mediated lung disease that occurs in susceptible individuals after repeated inhalation of pigeon antigens. The disease is characterised by hypersensitivity reactions that occur in the distil bronchioles and alveoli which may lead to irreversible pulmonary fibrosis. Previous studies have shown that IgG responses to carbohydrate determinants on pigeon intestinal mucin (PIM) are key to the development of disease. To specifically identify carbohydrate epitopes on PIM O-linked oligosaccharides were released by hydrazinolysis, separated by reverse phase-HPLC and the released free reducing oligosaccharides were then coupled to poly-L-lysine (PLL). The antigenicity of the resultant polyvalent conjugates was tested with sera from pigeon fanciers by both dot blot and ELISA. Further to this two dimensional 1H NMR spectroscopy and GC-MS was used to determine the structure of one of the major antigenic oligosaccharides of PIM. The sera reacted with a large number of PLL-oligosaccharide conjugates confirming that anti-PIM IgG1 and IgG2 recognise O-linked oligosaccahrides. ELISA showed that IgG1 responses were mainly to glycoforms in the earlier fractions (2–20) whilst IgG2 responses were directed against some of the very early fractions (2–10) and the later fractions (21–35) suggesting that there are at least two major epitopes on PIM. Furthermore IgG1 responses 101 John S. Axford (ed.), Glycobiology and Medicine, 101-102. © 2005 Springer. Printed in the Netherlands.

102

C. I. Baldwin et al.

were significantly higher in symptomatic individuals against oligosaccharides found in fractions 15–17. As anti-mucin IgG1 responses correlate with disease it may be that these fractions contain disease associated epitopes. Structural analysis revealed the presence of two novel forms of LeX substituted at C-3 of the b-Gal with either b-GalNAC or b-GlcNAc. These glycoforms have not been described in mammalian glycosystems and one may expect both to be highly antigenic. These observations support the concept that PIM is a key antigen in the development of PFL.

18

DIFFERENTIAL GLYCOSYLATION OF GELATINASE B FROM NEUTROPHILS AND BREAST CANCER CELLS

Simon A. Fry1, Philippe E. Van den Steen2, Louise Royle1, Mark R. Wormald1, Anthony J. Leathem3, Ghislain Opdenakker2, Pauline M. Rudd1, and Raymond A. Dwek1 1Glycobiology Institute Department of Biochemistry University of Oxford South Parks Road, Oxford OX1 3QU, UK 2Rega Institute for Medical Research Laboratory of Molecular Immunology University of Leuven Minderbroedersstraat 10, B-3000 Leuven, Belgium 3Department of Surgery Royal Free and University College London Medical School 67–73 Riding House Street, London W1W 1EJ, UK

1. INTRODUCTION The matrix metalloproteases (MMPs) are a family of zinc-dependant endopeptidases (Nagase and Woessner 1999) that are involved in extracellular matrix (ECM) remodelling in a variety of physiological and pathological processes. MMP degradation of ECM proteins is associated with many aspects of cancer progression, including cancer cell growth, differentiation, ff apoptosis, migration and invasion, as well as regulation of angiogenesis and immune surveillance (Egeblad and Werb 2002). Gelatinase B (MMP-9) is structurally one of the most complex MMPs (Opdenakker, Van den Steen et al. 2001; Van den Steen, Dubois et al. 2002). All MMPs have a prodomain (proteolytically removed to yield active enzyme), an active domain and a zinc-binding domain. Gelatinase B also has a gelatin-binding fibronectin domain, a collagen type V-like domain and a carboxyterminal hemopexin domain. Human natural gelatinase B from neutrophils is heavily glycosylated (Rudd, Mattu 103 John S. Axford (ed.), Glycobiology and Medicine, 103-112. © 2005 Springer. Printed in the Netherlands.

104

S. A. Fry et al.

et al. 1999; Mattu, Royle et al. 2000). 3 consensus sequences for the attachment of N-linked glycans are present in gelatinase B, 1 in the prodomain and 2 in the active domain. N-glycan site analysis revealed that only 2 of the 3 sites are occupied (Van den Steen, Opdenakker et al. 2001; Kotra, Zhang et al. 2002), one each in the prodomain (Asn -Leu-Thr) and active domain (Asn -Tyr-Ser). The O-linked gly38 120 cans are probably located in the Thr, Ser and Pro rich collagen V- like domain, since this sequence contains ideal clustered attachment sites for O-linked oligosaccharides. Gelatinase B is typically stored in large amounts in neutrophil granules to allow rapid release in innate immune responses (Masure, Proost et al. 1991). Expression is increased in many cancer cell types, including breast carcinomas (Zucker, Lysik et al. 1993), and experimental metastasis assays demonstrate that it is required for effective ff cell invasion and metastasis (Hua and Muschel 1996; Itoh, Tanioka et al. 1999). Gelatinase B has these biological properties because some of its substrates are ECM proteins. As well as cleaving gelatins (denatured collagens), gelatinase B can cleave the 3 fragment of collagen type II (produced following collagenase cleavage 4 at a single site) (Van den Steen, Proost et al. 2004), collagen type V (Hibbs, Hoidal et al. 1987) and perhaps (Mackay, Hartzler et al. 1990) collagen type IV (a major component of basement membranes) (Wilhelm, Collier et al. 1989; Okada, Gonoji et al. 1992). Other gelatinase B ECM substrates include elastin (Senior, Griffin et al. 1991), aggregan (Fosang, Neame et al. 1992), link protein (Nguyen, Murphy et al. 1993) and galectin-3. (Ochieng, Fridman et al. 1994) Aberrant protein glycosylation is a common feature of cancer (Feizi 1985; Saitoh, Wang et al. 1992; Lloyd, Burchell et al. 1996; Kim and Varki 1997; Granovsky, Fata et al. 2000). This chapter discusses the known glycosylation of natural neutrophil gelatinase B compared to that of gelatinase B secreted from MCF-7 breast cancer cells.

2. MATRIX METALLOPROTEASE GLYCOSYLATION MMPs are synthesised as latent proenzymes that are activated by proteolysis. As such, MMPs act sequentially in a protease cascade that produces active proteases whose combined effects ff mediate ECM remodelling (Fig. 1). Alterations in the glycosylation of certain MMPs can alter their enzymatic activity. This has consequences in terms of individual MMP properties and more widely in terms of the protease cascade. Plasminogen has 3 N-linked glycosylation sites and occurs naturally in 2 populations that differ ff only by the absence of an N-glycan at N288. The presence of an Nglycan at this site down regulates plasminogen activation by tissue-type plasminogen activator (Rudd, Woods et al. 1995). Although the observed effect ff is only a 2-fold reduction, these enzymes initiate the protease cascade which terminates in gelatinase B activation. A similar glycosylation-induced alteration in activation in other MMPs would be amplified through the cascade providing a potentially powerful means of regulation. Increased b1–6 GlcNAc branching of N-linked glycans has long been associated with tumour metastasis (Dennis, Laferte et al. 1987; Fernandes, Sagman et al. 1991).

Differential ff Glycosylation of Gelatinase B

105

Figure 1. The protease cascade. Each enzyme has another enzyme as a substrate, arrows indicate the activation cascade of latent proenzymes. u-PA, urokinase-type plasminogen activator; t-PA, tissue-type plasminogen activator.

Increased b1–6 GlcNAc branching (mediated by UDP-GlcNAc a-mannoside b1–6–N-acetylglucosaminyltransferase) leads to increased resistance of matriptase to degradation (Ihara, Miyoshi et al. 2002). This in turn promotes matrix metalloproteinase activation through enhanced activation of urokinase-type plasminogen activator (Fig. 1). MMPs, including gelatinase B, that have been secreted and activated can still be regulated by Tissue Inhibitor of Metalloprotease (TIMP) inhibition (Murphy and Docherty 1992). TIMP-1 exerts its inhibitory effect ff by binding with high affinity to both progelatinase B and activated gelatinase B (Murphy, Houbrechts et al. 1991). However, enzymatic desialylation of gelatinase B has been demonstrated to reduce TIMP-1 inhibition by ~50% (Van den Steen, Opdenakker et al. 2001). Together, these findings highlight how intimately glycosylation is associated with gelatinase B activity and activation.

3. PRODUCTION OF MCF-7 GELATINASE B AND GLYCOSYLATION ANALYSIS In order to perform glycan analysis, 100 mg of gelatinase B was required. Phorbolmyristate-acetate (PMA) is known to upregulate expression levels of gelatinase B in

106

S. A. Fry et al.

many cell types by directly activating protein kinase C (PKC) (Masure, Proost et al. 1991; Opdenakker, Masure et al. 1991; Houde, de Bruyne et al. 1993). Activated PKC phosphorylates a range of substrates leading to enhanced transcription of the gelatinase B gene (Houde, de Bruyne et al. 1993). Optimal gelatinase B expression levels were achieved by incubating MCF-7 cells in serum free media for 48 hours in the presence of 10 ng/ml PMA (data not shown). Cell supernatants were harvested, filtered and purified by affinity chromatography on gelatin-sepharose as previously described (Masure, Proost et al. 1991). In order to liberate N-glycans from purified gelatinase B, samples were reduced and alkylated prior to SDS-PAGE. Gel bands were cut out and washed alternatively with 20mM sodium bicarbonate and acetonitrile. N-glycans were removed from protein in the gel bands by overnight incubation with Peptide N-glycosidase F at 37°C. N-linked glycans were recovered from the gel pieces with water and sonication before being dried and 2-AB labeled. As O-glycans can not be removed enzymatically, manual hydrazinolysis was performed as previously described (Royle, Mattu et al. 2002). Briefly, gelatinase B samples were dialysed against 0.1% trifluoroacetic acid, lyophilised and then cryogenically dried. O-linked glycans were released by incubation with anhydrous hydrazine for 6hr at 60°C, then re-N-acetylated and desalted before peptides were removed. Released N- and O-glycans were fluorescently labelled with 2-AB according to the method of Bigge et al. (Bigge, Patel et al. 1995) 2-AB labelled N- and O-linked glycans of gelatinase B were resolved by NP-HPLC. By comparison with a standard dextran hydolysate ladder, oligosaccharide elution positions could be expressed as glucose units (GU). Oligosaccharide structures were assigned by reference to the GU values of a database of standard sugars (Royle, Mattu et al. 2002), and confirmed by using a series of parallel exoglycosidase digestions.

4. GELATINASE B GLYCOSYLATION MCF-7 gelatinase B is post-translationally modified differently ff to natural neutrophil gelatinase B. The N- and O-linked glycans of neutrophil gelatinase B have been sequenced (Rudd, Mattu et al. 1999; Mattu, Royle et al. 2000). More than 95% of the N-linked glycans are partially sialylated, core fucosylated bianntennary structures with and without outer arm a1–3 linked fucose (Fig. 2a). The O-linked glycans (ranging from 2–10 monosaccharides) mainly consisted of type II cores with lactosamine extensions, with or without sialic acid or outer arm fucose (Fig. 2b). The N-linked glycans of MCF-7 gelatinase B are all core-fucosylated biantennary structures (Fig. 3a). There is one mono-sialylated and one disialylated N-glycan, and all outer arm fucosylation occurs through a mixture of a1–2 and a1–3 linkages. The O-linked glycans of MCF-7 gelatinase B differ ff from those of neutrophil gelatinase B more dramatically than the N-linked glycans. MCF-7 gelatinase B O-glycans are relatively small (in the range of 2–6 monosaccharides) and heavily sialylated with no fucosylation (Fig 3b). The most abundant O-glycans have type I cores (Galb1–3GalNAc) with fewer galactosylated core II structures (Galb1–4GlcNAcb1–6 [Galb1–3]GalNAc) present.

Differential ff Glycosylation of Gelatinase B

107

Figure 2. N- and O-Linked glycans of neutrophil gelatinase B. a) Diagrammatic representation of a neutrophil gelatinase B bianntennary N-linked glycan. Glycosylation analysis showed this glycan is predominantly present in a core-fucosylated form, with varying amounts of outer arm a1–3 fucosylation or sialylation. b) A typical neutrophil gelatinase B O-linked glycan. Core II O-glycans containing 1–3 lactosamine repeats predominate. Fucosylation (a1–3linked) and/or sialylation (a2–3 and a2–6 linked) are also present. The scheme for glycan representation is as follows: #, mannose; 2, N-acetylgalactosamine; 1, galactosamine; &, N-acetylglucosamine; k, fucose; 0, sialic acid; dashed line, a-linkage; full line, , linkage position. b-linkage;

5. DISCUSSION The N-linked glycans of human neutrophil and MCF-7 gelatinase B are very similar. The main difference ff is the presence of a1–2 linked outer arm fucose on MCF-7 gelatinase B N-glycans (absent from neutrophil gelatinase B N-glycans). A function is yet to be ascribed to gelatinase B N-glycans, although 1 of the 2 Nglycans is known to reside on the propeptide, so could conceivably influence proteolytic activation. The minor differences ff in neutrophil and MCF-7 N-glycosylation are unlikely to impact on the biological activity of gelatinase B. In contrast to the N-linked glycans, there are dramatic differences ff in the Oglycans. In comparison to neutrophil gelatinase B O-glycans, the O-glycans of MCF-7 gelatinase B are truncated by the addition of sialic acid to core I structures (Fig. 4). This finding is consistent with that of truncation of MUC1 O-glycans in breast cancer (Lloyd, Burchell et al. 1996; Dalziel, Whitehouse et al. 2001). The majority of neutrophil gelatinase B core I O-glycans are N-acetylglucosylated to form core II structures. These are further elongated with N-acetyllactosamine repeats to produce extended O-linked glycans that are completely absent from MCF-7 gelatinase B. Interestingly, these differences ff in O-glycosylation will alter the nature of interaction with the gelatinase B substrate, galectin-3.

S. A. Fry et al.

108

Figure 3. N- and O-linked glycans of MCF-7 gelatinase B. a) Diagrammatic representation of an MCF-7 gelatinase B bianntennary N-linked glycan. Glycosylation analysis shows this glycan is always corefucosylated, with varying amounts of outer arm a1–3 and/or a1–2 linked fucose or a2–3 linked sialic acid. b) A typical MCF-7 gelatinase B O-linked glycan. Mono- and disialylated core I O-glycans predominate. There is an absence of fucosylation. The scheme for glycan representation is as follows: #, mannose; 2, N-acetylgalactosamine; 1, galactosamine; &, N-acetylglucosamine; k, fucose; 0, sialic acid; dashed line, a-linkage; full line, b-linkage;

, linkage position.

Galectin-3 is a 30kDa b-galactoside binding lectin that has a carboxy-terminal carbohydrate recognition domain (CRD) (Seetharaman, Kanigsberg et al. 1998) and a flexible Pro, Tyr and Gly-rich amino-terminal domain that contains a gelatinase B cleavage site. This amino-terminal domain is required for the non-covalent selfassociation of between 2 and 5 galectin-3 molecules (Ochieng, Platt et al. 1993; Barboni, Bawumia et al. 1999; Ahmad, Gabius et al. 2004), allowing cross-linking of suitable glycans. Breast cancer galectin-3 expression has been reported to correlate positively with disease progression (Honjo, Nangia-Makker et al. 2001). Also, galectin-3 induced expression of cell surface adhesion molecules and the direct interaction of galectin-3 with suitably glycosylated proteins, such as laminin and integrins, may favour tumorigenesis (Ochieng, Leite-Browning et al. 1998; Matarrese, Fusco et al. 2000). A property of galectin-3 is a dramatic increase in affinity in response to increases in repeating lactosamine units in the substrate; ie, LN (K =26 mM), LN2 (1.3 mM), d LN3 (0.35 mM) (Hirabayashi, Hashidate et al. 2002). The glycan analysis performed here shows that neutrophil gelatinase B has a higher proportion of O-glycans that are galectin-3 ligands, some of which contain lactosamine repeats, than MCF-7 gelatinase B (Table 1). Indeed, changes in galectin-3 binding to neutrophil and MCF-7 gelatinase B have been demonstrated (Fry et al., in press). Truncation of gelatinase B O-glycans reduces galectin-3 binding and may therefore facilitate cancer cell invasion. Gelatinase B is the only MMP to contain a collagen type V-like domain. This

Differential ff Glycosylation of Gelatinase B

109

Figure 4. Neutrophil and MCF-7 gelatinase B O-glycans differ ff in their core structures. Core structures of O-linked glycans are represented as a % of the total O-glycan pool. a) 70.1% of neutrophil gelatinase B O-glycans have a type II core. b) 90.4% of MCF-7 gelatinase B O-glycans have a type I core. Table 1. Gelatinase B O-linked glycans that are galectin-3 ligands. O-glycans that are galectin-3 ligands are represented as a % of the total O-glycan pool. LNx N , number (x) of lactosamine repeats. % O-glycans that are galectin-3 ligands Gelatinase B source

LN1

LN2

LN3

Total

Neutrophil MCF-7

36.3 8.6

6.1 –

3.8 –

46.2 8.6

56 amino acid sequence contains repeating Ser, Thr and Pro residues and thus is thought to be the site of clustered O-linked glycans (18 potential O-glycosylation sites as predicted by the NetOGlyc 3.1 server, www.cbs.dtu.dk/services/NetOGlyc/). The function of this domain is as yet unknown, but similar peptide sequences that are heavily O-glycosylated are extended and possess an increased rigidity of the peptide chain (Lukacik, Roversi et al. 2004). Hence, a role for the repeating Ser, Thr and Pro residues may be to extend the polypeptide chain to maximise separation

110

S. A. Fry et al.

between the hemopexin and the other domains. The O-glycans of MCF-7 gelatinase B are all short and carry negatively charged sialic acid. Charge repulsion may facilitate peptide backbone extension and alter gelatinase B properties accordingly. Gelatinase B secreted from MCF-7 breast cancer cells is aberrantly O-glycosylated when compared to natural neutrophil gelatinase B. Although N-glycosylation profiles are similar, MCF-7 O-glycans are truncated and heavily sialylated compared to those of neutrophil gelatinase B. This aberrant glycosylation results in a reduction in the proportion of gelatinase B O-glycans that are galectin-3 ligands, and a complete absence of multiple N-acetyllactosamine repeats that are bound with high affinity by galectin-3. These cancer associated alterations in glycosylation may facilitate cancer cell metastasis by altering the peptide conformation of gelatinase B and/or its interaction with galectin-3.

REFERENCES Ahmad, N., H. J. Gabius, et al. (2004). ‘‘Galectin-3 precipitates as a pentamer with synthetic multivalent carbohydrates and forms heterogeneous cross-linked complexes.’’ J Biol Chem 279(12): 10841–7. Barboni, E. A., S. Bawumia, et al. (1999). ‘‘Kinetic measurements of binding of galectin 3 to a laminin substratum.’’ Glycoconj J 16(7): 365–73. Bigge, J. C., T. P. Patel, et al. (1995). ‘‘Nonselective and efficient fluorescent labeling of glycans using 2-amino benzamide and anthranilic acid.’’ Anal Biochem 230(2): 229–38. Dalziel, M., C. Whitehouse, et al. (2001). ‘‘The relative activities of the C2GnT1 and ST3Gal-I glycosyltransferases determine O-glycan structure and expression of a tumor-associated epitope on MUC1.’’ J Biol Chem 276(14): 11007–15. Dennis, J. W., S. Laferte, et al. (1987). ‘‘Beta 1–6 branching of Asn-linked oligosaccharides is directly associated with metastasis.’’ Science 236(4801): 582–5. Egeblad, M. and Z. Werb (2002). ‘‘New functions for the matrix metalloproteinases in cancer progression.’’ Nat Rev Cancer 2(3): 161–74. Feizi, T. (1985). ‘‘Demonstration by monoclonal antibodies that carbohydrate structures of glycoproteins and glycolipids are onco-developmental antigens.’’ Nature 314(6006): 53–7. Fernandes, B., U. Sagman, et al. (1991). ‘‘Beta 1–6 branched oligosaccharides as a marker of tumor progression in human breast and colon neoplasia.’’ Cancer Res 51(2): 718–23. Fosang, A. J., P. J. Neame, et al. (1992). ‘‘The interglobular domain of cartilage aggrecan is cleaved by PUMP, gelatinases, and cathepsin B.’’ J Biol Chem 267(27): 19470–4. Granovsky, M., J. Fata, et al. (2000). ‘‘Suppression of tumor growth and metastasis in Mgat5-deficient mice.’’ Nat Med 6(3): 306–12. Hibbs, M. S., J. R. Hoidal, et al. (1987). ‘‘Expression of a metalloproteinase that degrades native type V collagen and denatured collagens by cultured human alveolar macrophages.’’ J Clin Invest 80(6): 1644–50. Hirabayashi, J., T. Hashidate, et al. (2002). ‘‘Oligosaccharide specificity of galectins: a search by frontal affinity chromatography.’’ Biochim Biophys Acta 1572(2–3): 232–54. Honjo, Y., P. Nangia-Makker, et al. (2001). ‘‘Down-regulation of galectin-3 suppresses tumorigenicity of human breast carcinoma cells.’’ Clin Cancer Res 7(3): 661–8. Houde, M., G. de Bruyne, et al. (1993). ‘‘Differential ff regulation of gelatinase B and tissue-type plasminogen activator expression in human Bowes melanoma cells.’’ Int J Cancer 53(3): 395–400. Hua, J. and R. J. Muschel (1996). ‘‘Inhibition of matrix metalloproteinase 9 expression by a ribozyme blocks metastasis in a rat sarcoma model system.’’ Cancer Res 56(22): 5279–84. Ihara, S., E. Miyoshi, et al. (2002). ‘‘Prometastatic effect ff of N-acetylglucosaminyltransferase V is due to modification and stabilization of active matriptase by adding beta 1–6 GlcNAc branching.’’ J Biol Chem 277(19): 16960–7. Itoh, T., M. Tanioka, et al. (1999). ‘‘Experimental metastasis is suppressed in MMP-9-deficient mice.’’ Clin Exp Metastasis 17(2): 177–81.

Differential ff Glycosylation of Gelatinase B

111

Kim, Y. J. and A. Varki (1997). ‘‘Perspectives on the significance of altered glycosylation of glycoproteins in cancer.’’ Glycoconj J 14(5): 569–76. Kotra, L. P., L. Zhang, et al. (2002). ‘‘N-Glycosylation pattern of the zymogenic form of human matrix metalloproteinase-9.’’ Bioorg Chem 30(5): 356–70. Lloyd, K. O., J. Burchell, et al. (1996). ‘‘Comparison of O-linked carbohydrate chains in MUC-1 mucin from normal breast epithelial cell lines and breast carcinoma cell lines. Demonstration of simpler and fewer glycan chains in tumor cells.’’ J Biol Chem 271(52): 33325–34. Lukacik, P., P. Roversi, et al. (2004). ‘‘Complement regulation at the molecular level: the structure of decayaccelerating factor.’’ Proc Natl Acad Sci U S A 101(5): 1279–84. Mackay, A. R., J. L. Hartzler, et al. (1990). ‘‘Studies on the ability of 65-kDa and 92-kDa tumor cell gelatinases to degrade type IV collagen.’’ J Biol Chem 265(35): 21929–34. Masure, S., P. Proost, et al. (1991). ‘‘Purification and identification of 91-kDa neutrophil gelatinase. Release by the activating peptide interleukin-8.’’ Eur J Biochem 198(2): 391–8. Matarrese, P., O. Fusco, et al. (2000). ‘‘Galectin-3 overexpression protects from apoptosis by improving cell adhesion properties.’’ Int J Cancer 85(4): 545–54. Mattu, T. S., L. Royle, et al. (2000). ‘‘O-glycan analysis of natural human neutrophil gelatinase B using a combination of normal phase-HPLC and online tandem mass spectrometry: implications for the domain organization of the enzyme.’’ Biochemistry 39(51): 15695–704. Murphy, G. and A. J. Docherty (1992). ‘‘The matrix metalloproteinases and their inhibitors.’’ Am J Respir Cell Mol Biol 7(2): 120–5. Murphy, G., A. Houbrechts, et al. (1991). ‘‘The N-terminal domain of tissue inhibitor of metalloproteinases retains metalloproteinase inhibitory activity.’’ Biochemistry 30(33): 8097–102. Nagase, H. and J. F. Woessner, Jr. (1999). ‘‘Matrix metalloproteinases.’’ J Biol Chem 274(31): 21491–4. Nguyen, Q., G. Murphy, et al. (1993). ‘‘Matrix metalloproteinases cleave at two distinct sites on human cartilage link protein.’’ Biochem J 295 (Pt 2): 595–8. Ochieng, J., R. Fridman, et al. (1994). ‘‘Galectin-3 is a novel substrate for human matrix metalloproteinases-2 and –9.’’ Biochemistry 33(47): 14109–14. Ochieng, J., M. L. Leite-Browning, et al. (1998). ‘‘Regulation of cellular adhesion to extracellular matrix proteins by galectin-3.’’ Biochem Biophys Res Commun 246(3): 788–91. Ochieng, J., D. Platt, et al. (1993). ‘‘Structure-function relationship of a recombinant human galactosidebinding protein.’’ Biochemistry 32(16): 4455–60. Okada, Y., Y. Gonoji, et al. (1992). ‘‘Matrix metalloproteinase 9 (92-kDa gelatinase/type IV collagenase) from HT 1080 human fibrosarcoma cells. Purification and activation of the precursor and enzymic properties.’’ J Biol Chem 267(30): 21712–9. Opdenakker, G., S. Masure, et al. (1991). ‘‘Cytokine-mediated regulation of human leukocyte gelatinases and role in arthritis.’’ L ymphokine Cytokine Res 10(4): 317–24. Opdenakker, G., P. E. Van den Steen, et al. (2001). ‘‘Gelatinase B: a tuner and amplifier of immune functions.’’ T rends Immunol 22(10): 571–9. Royle, L., T. S. Mattu, et al. (2002). ‘‘An analytical and structural database provides a strategy for sequencing O-glycans from microgram quantities of glycoproteins.’’ Anal Biochem 304(1): 70–90. Rudd, P. M., T. S. Mattu, et al. (1999). ‘‘Glycosylation of natural human neutrophil gelatinase B and neutrophil gelatinase B-associated lipocalin.’’ Biochemistry 38(42): 13937–50. Rudd, P. M., R. J. Woods, et al. (1995). ‘‘The effects ff of variable glycosylation on the functional activities of ribonuclease, plasminogen and tissue plasminogen activator.’’ Biochim Biophys Acta 1248(1): 1–10. Saitoh, O., W. C. Wang, et al. (1992). ‘‘Differential ff glycosylation and cell surface expression of lysosomal membrane glycoproteins in sublines of a human colon cancer exhibiting distinct metastatic potentials.’’ J Biol Chem 267(8): 5700–11. Seetharaman, J., A. Kanigsberg, et al. (1998). ‘‘X-ray crystal structure of the human galectin-3 carbohydrate recognition domain at 2.1-A resolution.’’ J Biol Chem 273(21): 13047–52. Senior, R. M., G. L. Griffin, et al. (1991). ‘‘Human 92– and 72-kilodalton type IV collagenases are elastases.’’ J Biol Chem 266(12): 7870–5. Van den Steen, P. E., B. Dubois, et al. (2002). ‘‘Biochemistry and molecular biology of gelatinase B or matrix metalloproteinase-9 (MMP-9).’’ Crit Rev Biochem Mol Biol 37(6): 375–536. Van den Steen, P. E., G. Opdenakker, et al. (2001). ‘‘Matrix remodelling enzymes, the protease cascade and glycosylation.’’ Biochim Biophys Acta 1528(2–3): 61–73. Van den Steen, P. E., P. Proost, et al. (2004). ‘‘Generation of glycosylated remnant epitopes from human collagen type II by gelatinase B.’’ Biochemistry 43(33): 10809–16.

112

S. A. Fry et al.

Wilhelm, S. M., I. E. Collier, et al. (1989). ‘‘SV40-transformed human lung fibroblasts secrete a 92-kDa type IV collagenase which is identical to that secreted by normal human macrophages.’’ J Biol Chem 264(29): 17213–21. Zucker, S., R. M. Lysik, et al. (1993). ‘‘M(r) 92,000 type IV collagenase is increased in plasma of patients with colon cancer and breast cancer.’’ Cancer Res 53(1): 140–6.

19

DETECTION OF GLYCOSYLATION CHANGES IN SERUM AND TISSUE PROTEINS IN CANCER BY LECTIN BLOTTING

R. E. Ferguson, D. H. Jackson, R. Hutson, N. Wilkinson, P. Harnden, P. Selby, and R. E. Banks Cancer Research UK Clinical Centre St. James’s University Hospital Leeds, L59 7TF, UK

Renal cancer is the tenth most common cancer but its incidence is increasing, with the 22% increase in rate in the last 10 years being the largest change for any cancer in females [1]. Surgery is the main therapy for organ-confined disease, however, over 50% of cases present with locally advanced or metastatic disease which is resistant to chemotherapy. Ovarian cancer is the fourth most common malignancy in females and due to the relatively asymptomatic progression, over 70% patients

Figure 1. 1D PAGE blot of 4 patient matched normal and malignant renal tissues probed with one of a panel of lectins. Boxed region highlights specific differences ff between the matched normal and renal cancer tissue samples. 113 John S. Axford (ed.), Glycobiology and Medicine, 113-114. © 2005 Springer. Printed in the Netherlands.

114

R. E. Ferguson et al.

present with advanced disease [2]. In both cases there is a clear need for tumour markers to enable earlier diagnosis of disease as well as the identification of novel therapeutic targets. It is estimated that over 60% of all proteins, and virtually all secreted and transmembrane proteins are glycosylated [3] and there is abundant evidence for alterations of glycosylation in cancer with glycan structures playing a role in cancer cell homing and metastasis. Lectins are proteins which bind specific glycans and an effective ff method to study cancer-associated glycosylation changes is by lectin-based profiling [4]. We have studied changes in glycosylation of serum proteins and normal and tumour tissue samples using lectin-based profiling, with the aim that such cancerassociated glycoforms will form the basis of either tumour marker assays or therapeutic targets. 1D- and 2D-PAGE blots probed with a panel of lectins has demonstrated several differences ff in glycoprotein profile for both malignancies (see Fig. 1) and has identified a potential novel ovarian cancer serum marker which is currently undergoing downstream validation to determine its diagnostic utility.

REFERENCES 1. Cancer statistics 2002 – Cancer Research UK. 2. Ries, L. A. G. et al. (2001) SEER Cancer Statistics Review 1973–1998, Section 20: Ovarian Cancer. Bethesda: National Cancer Institute. 3. Apweiler, R. et al. (1999) Biochim Biophys Acta, 1473: 2 1–34. 4. Dwek M. V., et al. (2001) Proteomics, 1:756–62.

20

CARBOHYDRATES AND BIOLOGY OF STAPHYLOCOCCAL INFECTIONS

Andrej Tarkowski1, Margareta Verdrengh1, Ing-Marie Jonsson1, Mattias Magnusson1, Simon J Foster2, and Zai-Quing Liu1 1Department of Rheumatology and Inflammation Research Goteborg University, Sweden 2Department of Molecular Biology and Biotechnology University of Sheffield, UK

The genus Staphylococcus includes more than 30 species but only three are of major clinical importance: S. aureus, S. epidermidis, and S. saprophyticus. S. aureus, that this review deals with, is a commensal found on the skin of 1/3 of the entire human population. However, in many healthy and diseased subjects S. aureus colonization reaches much higher numbers. The increasing number of immunocompromised subjects on one hand and the appearence of methicillin-resistant staphylococci on the other should increase the attempts to better understand the biology of hostbacterium interaction in staphylococcal diseases. In this abstract we focus on the impact of certain carbohydrate structures within staphylococci on their virulence and ability to cause inflammatory responses. Furthermore, we describe action of carbohydrates stemming from seaweed, being able to downregulate inflammatory responses caused by staphylococcal infection by interacting with rolling properties of endogenous leukocytes. Carbohydrate constituents of the staphylococcal cell wall include polysaccharide microcapsule and peptidoglycans. We have shown that staphylococci, defective with respect to the expression of the type 5 capsular polysaccharides triggered in murine recipients lower frequency and severity of arthritis as well as infection related mortality than the congeneic wild-type strain. Further in vitro studies suggested that this outcome was due to the enhanced phagocytosis and intracellular killing of bacteria that lacked capsular polysaccharides. This finding was employed in a recent study where vaccination with staphylococcal polysaccharides (coupled to protein carrier to increase their immunogenicity) was used in a clinical setting. It was shown that hemodialysis patients receiving such a conjugate vaccine were partially protected against staphylococcal bacteremia. Peptidoglycan component of the staphylococcal 115 John S. Axford (ed.), Glycobiology and Medicine, 115-116. © 2005 Springer. Printed in the Netherlands.

116

A. Tarkowski et al.

cell wall proved to be an extremely potent pro-inflammatory substance, even if adminstered alone in a single dose intra-articularly. Inflammation triggered by peptidoglycans was mediated by combined action of lymphocytes and monocytes as proved in a recent study. Peptidoglycans may very well prove to be important cause of staphylococcal sepsis/septic shock along with actions of superantigens. Our recent and still unpublished data (Magnusson et al.) indicate that the peptidoglycan component that probably is responsable for its inflammatory action is Pam-Cys. Indeed, Pam-Cys alone is capable to trigger severe synovitis, when injected intra-articularly. This finding opens the perspective to interact with pro-inflammatory properties of staphylococcal peptidoglycans by silencing Toll Like receptor 2, mediating the action of Pam-Cys. The selectin family of adhesion molecules mediates the initial attachment of leukocytes to venular endothelial cells at sites of tissue injury and inflammation. For this reason expression of selectins mediates inflammation also in case of staphylococcal arthritis. Fucoidin, a sulfated polysaccharide from seaweed, binds to and blocks the function of L-and P-selectins thereby inhibiting leukocyte rolling and adhesion to endothelial surface. Treatment with fucoidin was used to assess whether signs of early joint inflammation triggered by invading staphylococci could be blocked. Indeed, severity of septic arthritis was significantly decreased within the first three days following bacterial infection. Similar result was obtained using P-selectin deficient mice indicating that this selectin interaction is of crucial importance for extravasation of leukocytes during staphylococcal arthritis. Since P-selectin deficiency decreases phagocytic activity of neutrophils it should be kept in mind that the effcient ff treatment of staphylococcal arthritis should encompass combination of antibiotics and immunomodulation. Indeed, recent experimental and clinical studies indicate that this principle is clearly superior to antibiotic treatment alone.

21

NEW DEVELOPMENTS IN TREATING GLYCOSPHINGOLIPID STORAGE DISEASES

Frances M. Platt, Mylvaganam Jeyakumar, Ulrika Andersson, Raymond A. Dwek and Terry D. Butters Department of Biochemistry University of Oxford South Parks Road, Oxford OX1 3QU, UK

1. GLYCOSPHINGOLIPIDS Eukaryotic cells have complex membranes that consist of a number of lipid and protein species. The majority of the lipids present are phospholipids, with glycosphingolipids being present at much lower levels, representing only a few percent of total cellular lipid content. Glycosphingolipids are not required for membrane integrity (Ichikawa, Nakajo et al. 1994) but function in conjunction with cholesterol in the formation of membrane microdomains or rafts that facilitate certain signalling events in cells (Kobayashi and Hirabayashi 2000; Galbiati, Razani et al. 2001; Munro 2003). They have been implicated to play a role in a number of biological processes (Bektas and Spiegel 2004; Zhang and Kiechle 2004) and are required for embryogenesis (Yamashita, Wada et al. 1999). In this review we will focus on the sub-family of glycosphingolipids that have glucosylceramide (Ichikawa and Hirabayashi 1998) as their core structure, the glucosphingolipids (GSLs).

2. GSL BIOSYNTHESIS AND CATABOLISM GSLs are synthesised within the Golgi apparatus by the sequential addition of monosaccharides to ceramide (Sandhoff and Kolter 2003). The first step in this pathway is the transfer of glucose to ceramide to generate glucosylceramide (Ichikawa, Sakiyama et al. 1996; Chujor, Feingold et al. 1998; Ichikawa, Ozawa et al. 1998). This reaction is catalyzed by ceramide glucosyltransferase, an enzyme that has its catalytic domain on the cytosolic side of an early Golgi compartment 117 John S. Axford (ed.), Glycobiology and Medicine, 117-126. © 2005 Springer. Printed in the Netherlands.

118

F. M. Platt et al.

(Futerman and Pagano 1991; Trinchera, Fabbri et al. 1991). The GlcCer is then either transported directly to other sites within the cell (Warnock, Lutz et al. 1994) or flipped into the Golgi lumen and further modified by the sequential addition of monosaccharides via the action of glycosyltransferases (Sandhoff and Kolter 2003). This further processing results in two families of GSLs, the neutral GSLs and the gangliosides (Maccioni, Daniotti et al. 1999). The gangliosides contain one or more sialic acid residue and are present at high levels in the CNS (Walkley, Zervas et al. 2000). There are no human disease states associated with the early steps in GSL biosynthesis, implying this pathway may be essential for embryonic development. Consistant with this, knocking out the ceramide glucosyltransferase in the mouse is embryonically lethal (Yamashita, Wada et al. 1999). However, there has been a recent report of a severe human infantile onset epilepsy syndrome in an Amish pedigree resulting from a proven defect in the GM3 synthase gene, preventing the formation of complex gangliosides (Simpson, Cross et al. 2004). This raises the possibility that other human diseases resulting from defects in ganglioside biosynthesis exist, but have yet to be identified. GSLs typically re-cycle from and to the cell surface via the Golgi apparatus (Pagano, Puri et al. 2000) but can also be routed to the lysosome for degradation where they are subjected to the sequential action of specific glycohydrolases (Sandhoff and Kolter 2003). Disease states are known to result from defects in many of the steps in the GSL catabolic pathway. The substrate for the defective enzyme accumulates in the lysosome and leads to pathology. The resulting diseases are termed the GSL lysosomal storage disorders (Jeyakumar, Butters et al. 2002; Platt and Walkley 2004).

3. GSL LYSOSOMAL STORAGE DISEASES The GSL storage diseases are individually rare but collectively affect ff approximately 1:18,000 live births (Meikle, Hopwood et al. 1999). They result from the inherited defects in the genes that encode the lysosomal hydrolases or their co-factors required for GSL catabolism in the lysosome. The substrate(s) for the defective enzyme accumulates in the endosomal/lysosomal system leading to disease. The age of onset of clinical signs is highly variable depending on how the specific mutation affects ff enzyme function and/or stability (Wraith 2004). Relatively modest levels of residual enzyme activity can slow the rate of disease progression where as individuals almost null for the enzyme activity in question will develop disease in utero or during the early post-natal period and have the most severely attenuated life span (Winchester 2004). The diseases fall into two main categories, those in which neutral GSLs are stored (Gaucher and Fabry disease) and those in which gangliosides are stored (the GM1 and GM2 gangliosidoses).

4. SECONDARY STORAGE OF GSLs Storage of GSLs also occurs in a number of storage diseases in which the primary defect is independent of GSL catabolism (Walkley 2004). The reason for

New Developments in Glycosphingolipid Storage Diseases

119

the accumulation of these secondary species is unknown, but there is emerging evidence implicating the secondary storage of GSLs contributing to the disease process (Zervas, Somers et al. 2001). Reducing GSL storage in these diseases may therefore constitute a rational approach to their potential therapy (Platt and Butters 2004).

5. THERAPEUTIC APPROACHES FOR GSL STORAGE DISEASES As GSL storage diseases are the result of defects in the genes encoding the catabolic enzymes of the lysosome, introducing the functional wild type enzyme or gene should correct the problem. As the majority of GSL storage disorders involve GSL storage in the CNS, correction of disease in the brain is essential if these therapies are going to deal with more than the visceral manifestation of these diseases (Schiffmann ff and Brady 2002). Currently, enzyme replacement therapy is in clinical use for the non-CNS diseases, type 1 Gaucher and Fabry (Brady 2003; Neufeld 2004). However, we currently have no therapies for treating storage diseases of the brain. Gene therapy remains elusive (Cabrera-Salazar, Novelli et al. 2002), although eventual efficacy of this approach in these diseases is anticipated as the hurdle to cross to achieve therapeutic benefit in these diseases is a very low one. Even a small increase in residual enzyme activity can greatly reduce disease severity (Winchester 2004). There are also cell-based therapies such as bone marrow transplantation (BMT) (Erikson, Groth et al. 1990; Ringden, Groth et al. 1995; Dobrenis 2004). The amount of brain reconstitution with BM-derived microglial cells is low and this limits the level of functional enzyme achievable in the CNS (Krivit, Sung et al. 1995). The risks associated with the transplantation procedure and the need for HLA-matched donors limits the clinical application of BMT. When it has been used it generally tends to arrest the disease process but does not reverse pre-existing clinical symptoms. New emerging experimental cell-based therapies involve the introduction of neuronal stem cells directly into the brain to serve as a source of wild type enzyme and also to potentially replace dead or dying cell. This approach has the potential to treat all neurodegenerative diseases, including the GSL storage diseases (Snyder, Daley et al. 2004). The other therapeutic option is to decrease the synthesis of the stored substrate using small molecule enzyme inhibitors. This approach was suggested first by Radin (Inokuchi and Radin 1987; Radin 1996) and has been termed substrate reduction therapy (SRT) (Platt and Butters 2004). As GSL species cannot be completely degraded in the lysosome, as a result of the inherited enzyme deficiency, the biosynthesis of fewer GSL molecules reduces the influx of GSLs into the lysosome allowing more of the molecules synthesised by the cell to be catabolised. The aim is to balance synthesis with the impaired rate of degradation. There are several advantages to this approach. These include use of an orally acting drug, use of a drug which penetrates the CNS and by targeting an early step in the GSL biosynthetic pathway one drug could potentially be used to treat multiple GSL storage diseases (Fig. 1).

120

F. M. Platt et al.

Figure 1. Schematic representation of glycosphingolipid biosynthesis. The step in the pathway inhibited by the imino sugars NB-DNJ and NB-DGJ is the ceramide glycosyltransferase catalysed biosynthesis of glycosylceramide from ceramide.

6. SUBSTRATE REDUCTION THERAPY (SRT ) Small molecule inhibitors of the first step in GSL biosynthesis have been identified that have the potential to be used for SRT. To date there are two distinct chemical classes of these inhibitors that have been characterised, the PDMP series of compounds (Abe, Inokuchi et al. 1992) and the imino sugars (Butters, Dwek et al. 2003; Butters, Mellor et al. 2003). In this article we will be focusing on the imino sugars as these are now in clinical use (Lachmann 2003). The imino sugars are stereochemical monosaccharide mimetics that have a ring nitrogen atom instead of the oxygen (Winchester and Fleet 1992; Butters, van den Broek et al. 2000). The inhibition of ceramide glucosyltransferase by these imino sugars (Fig. 1) is critically dependant upon N-alkyation of the ring nitrogen with at least a 4 carbon alkyl chain (Butters, Mellor et al. 2003). The prototypic compound is N-butyldeoxynojirimycin (NB-DNJ) (Platt, Neises et al. 1994a). The galactose analogue N-butyldeoxygalactonojirimycin (NB-DGJ) also inhibits the ceramide glucosyltransferase (Platt, Neises et al. 1994b). The mechanism for their inhibitory property is not fully understood but may in part be due to their structural similarity to ceramide (Butters, Mellor et al. 2003).

7. SRT IN THE SANDHOFF DISEASE MOUSE MODEL Proof of principle studies using NB-DNJ were conducted in a mouse model of Sandhoff disease. The mouse lacked the b-subunit of b-hexosaminidase resulting in loss of the HexA (ab) and Hex B (bb) isoenzymes, with only very low level of residual enzyme activity conferred by Hex S (aa) (Sango, Yamanaka et al. 1995). GM2 ganglioside storage results, along with storage of GA2 due to the action of a

New Developments in Glycosphingolipid Storage Diseases

121

lysosomal sialidase that converts GM2 to its asialo derivative GA2. This mouse model has a clinical phenotype similar to human Tay-Sachs and Sandhoff disease. Sandhoff mice were fed on a diet containing NB-DNJ and were monitored for the rate of disease progression, using behavioural tests that measure motor coordination and muscle strength (Jeyakumar, Butters et al. 1999). It was found that the presymptomatic period was extended in response to SRT, the rate of clinical decline was slowed and life expectancy increased by approximately forty percent. This demonstrated that NB-DNJ mediated SRT could potentially be of therapeutic value in managing storage diseases, including those involving pathology in the CNS. A clinical trial was initiated in type 1 Gaucher disease, a disorder with welldefined clinical endpoints and no CNS involvement.

8. CLINICAL TRIALS OF NB-DNJ N Type 1 Gaucher disease is a macrophage disorder characterised by hepatosplenomegaly, anaemia and bone disease (Beutler and Grabowski 2001). Patients were recruited (Cambridge, Amsterdam, Prague and Jerusalem) into a 1-year open-label clinical trial of NB-DNJ (also termed OGT-918) (Cox, Lachmann et al. 2000). All patients were unable or unwilling to receive ERT, the standard of care in this disease (Neufeld 2004). Liver and spleen volumes and haematological parameters were measured. Biochemical markers were also assessed including chitotriosidase (Aerts and Hollak 1997), cell surface leukocyte GM1 and plasma levels of GlcCer, the storage lipid. Oral dosing was typically 100 mg OGT-918 three times daily. Pharmacokinetic profiling in a subgroup of the 28 patients showed that the drug reached maximum plasma concentrations at 2.5 hours with a plasma half-life of 6.3 hours. Steady state concentrations of OGT-918 were achieved after 15 days of dosing. The mean peak level of OGT-918 over the 12 month study was 6.8 mM with trough values of 3.9 mM (Cox, Lachmann et al. 2000; Moyses 2003).

9. CLINICAL TRIALS Spleen and liver volumes showed a significant reduction ((15%, 11.8–18.4, p<0.001) and (7%, 3.4–10.5, p<0.001) respectively) after six months of therapy. At 12 months the decrease from base line was 19% (14.3–23.7, p<0.001) and 12% (7.8–16.4, p<0.001) respectively (Cox, Lachmann et al. 2000). This was comparable to the response observed in patients of the same disease severity at baseline receiving ERT (Lachmann and Platt 2001). Chitotriosidase activity showed a time dependent reduction consistent with a reduction in the number of Gaucher cells. (Cox, Lachmann et al. 2000). Haemoglobin and platelet counts showed trends towards improvement, with a greater improvement in haemoglobin noted in patients who were anaemic at baseline. A statistically significant improvement in platelet counts was achieved following 12 months of treatment. Longer-term efficacy and safety were evaluated in patients that had completed 12 months of therapy (Elstein, Hollak et al. 2004). Eighteen of the 22 patients that were eligible entered the extension phase and were followed for a further two years.

122

F. M. Platt et al.

Continued and increasing efficacy was observed in all clinical parameters assessed. There was a marked improvement in platelet counts and haemoglobin levels relative to the 12-month time point. No serious adverse events were reported and no further cases of peripheral neuropathy emerged in the extension phase (two cases were reported and the patients withdrew in the initial 12 month study). GI tract side effects ff (due to intestinal sucrase/isomaltase inhibition by NB-DNJ) persisted in these patients, but to a lesser extent than in the first 12 months. Bone marrow fat fraction measurements were made in two patients and improvements were noted at 12 months, with further improvement found at 36 months. This parameter reflects a reduction in the number of Gaucher cells in the bone marrow in response to therapy. Consistent with this finding, chitotriosidase levels continued to decline in the extension phase.

10. REGULATORY APPROVAL In 2002 the European regulatory authority (EMEA) approved NB-DNJ (miglustat, Zavesca≤) for the treatment of type 1 Gaucher disease (mild to moderate disease, unwilling or unable to receive ERT) (Cox, Aerts et al. 2003; Lachmann 2003). In the same year approval was granted in Israel. The FDA approved miglustat in the USA in 2003, under the same label.

11. CURRENT CLINICAL TRIALS Clinical trials with miglustat are currently in progress in type 3 Gaucher disease, late onset Tay-Sachs disease and in Niemann-Pick type C. Results from these trials are anticipated in 2005.

12. SECOND-GENERATION COMPOUNDS The galactose analogue NB-DGJ is equivalent to NB-DNJ in terms of potency against the ceramide glucosyltransferase (Platt, Neises et al. 1994) but lacks many of the additional enzyme inhibitory properties associated with NB-DNJ (Andersson, Butters et al. 2000). Significantly, it does not inhibit the gut disaccharidases sucrase/isomaltase, the property of NB-DNJ that causes osmotic diarrhoea (Andersson, Butters et al. 2000). Also, NB-DGJ does not cause weight loss. NB-DGJ was evaluated in the mouse model of Sandhoff disease and dose escalation was possible with increasing benefit in terms of extended survival and improved function (Andersson, Smith et al. 2004). This compound is now in phase 1 in healthy volunteers.

13. INFLAMMATION AS AN ADDITIONAL TARGET FOR TREATING GSL STORAGE DISEASES To date two main downstream consequences of storage have been reported that are potentially amenable to pharmacological intervention, altered calcium homeostasis (Ginzburg, Kacher et al. 2004) and macrophage/microglial cell mediated

New Developments in Glycosphingolipid Storage Diseases

123

inflammation in the brain (Wada, Tifft ff et al. 2000; Jeyakumar, Thomas et al. 2003; Wu and Proia 2004). The CNS inflammation has been extensively characterised and mouse models of the GM2 gangliosidoses (Tay-Sachs, Late Onset Tay-Sachs (LOTS), Sandhoff ) and GM1 gangliosidosis have been studied to determine whether there is a common neuro-inflammatory component to these disorders (Jeyakumar, Thomas et al. 2003). During the disease course, the expression of a number of inflammatory markers have been studied in the central nervous system (CNS), including MHC class II, CD68, CD11b (CR3), 7/4, F4/80, nitrotyrosine, CD4 and CD8. Cytokine production was also profiled (TNFa, TGFb1 and IL1b) and blood-brain barrier (BBB) integrity determined. The kinetics of apoptosis and the expression of Fas and TNF-R1 were also assessed. In all symptomatic mouse models, a progressive increase in local microglial activation/expansion and infiltration of inflammatory cells was noted. Altered blood-brain barrier permeability was detected in Sandhoff and GM1 ff LOTS mice. Progressive CNS inflammation mice, but not in the more mildly affected was coincident with the onset of clinical signs in these mouse models. These data suggested that inflammation might play an important role in the pathogenesis of the gangliosidoses (Jeyakumar, Thomas et al. 2003). Recently, anti-inflammatory drugs and anti-oxidants (vitamin E and vitamin C) have been evaluated in the mouse model of Sandhoff disease and found to show efficacy as monotherapies and to synergise with SRT (Jeyakumar, Smith et al. 2004). As this approach utilises drugs already in common usage, the translation of these findings into clinical studies could potentially be quite rapid. It is likely that combination therapy will provide the most benefit for the infantile onset GSL storage diseases, which remain the most challenging group of patients to treat, due to the lack of significant levels of residual enzyme activity.

14. PERSPECTIVE Despite the relative rarity of the GSL storage diseases there has been an impressive amount of progress made in both understanding the underlying the pathobiology of these diseases and in the development of multiple therapeutic approaches of which BMT, ERT and SRT are currently in clinical use. In common with the more common neurodegenerative diseases such as Alzheimer’s, the GSL storage diseases are neuro-inflammatory diseases with inflammation in the brain contributing to disease progression. How inflammation is triggered by the storage of GSLs is not yet known. However, the efficacy of non-steroidal anti-inflammatory drugs and anti-oxidant therapies in the Sandhoff mouse offers ff the prospect that targeting inflammation as an adjunctive therapy in these and related storage disorders may be of clinical benefit. It is to be anticipated that over the next few years further new targets for storage disease therapy will emerge and be evaluated in clinical studies either alone or in combination with other proven or experimental therapies.

REFERENCES Abe, A., J. Inokuchi, et al. (1992). ‘‘Improved inhibitors of glucosylceramide synthase.’’ J Biochem Tokyo 111(2): 191–6.

124

F. M. Platt et al.

Aerts, J. M. and C. E. Hollak (1997). ‘‘Plasma and metabolic abnormalities in Gaucher’s disease.’’ Baillieres Clin Haematol 10(4): 691–709. Andersson, U., T. D. Butters, et al. (2000). ‘‘N-butyldeoxygalactonojirimycin: a more selective inhibitor of glycosphingolipid biosynthesis than N-butyldeoxynojirimycin, in vitro and in vivo.’’ Biochem Pharmacol 59(7): 821–9. Andersson, U., D. Smith, et al. (2004). ‘‘Improved outcome of N-butyldeoxygalactonojirimycin-mediated substrate reduction therapy in a mouse model of Sandhoff disease.’’ Neurobiol Dis 16(3): 506–15. Bektas, M. and S. Spiegel (2004). ‘‘Glycosphingolipids and cell death.’’ Glycoconj J 20(1): 39–47. Beutler, E. and G. Grabowski (2001). Gaucher disease. The metabolic and molecular bases of inherited diseases. C. R. Scriver, A. L. Beadet, D. Valle and W. S. Sly. New York, McGraw Hill. 3: 3636–3668. Brady, R. O. (2003). ‘‘Enzyme replacement therapy: conception, chaos and culmination.’’ Philos Trans R Soc Lond B Biol Sci 358(1433): 915–9. Butters, T. D., R. A. Dwek, et al. (2003). ‘‘Therapeutic applications of imino sugars in lysosomal storage disorders.’’ Curr Top Med Chem 3(5): 561–74. Butters, T. D., H. R. Mellor, et al. (2003). ‘‘Small-molecule therapeutics for the treatment of glycolipid lysosomal storage disorders.’’ Philos Trans R Soc Lond B Biol Sci 358(1433): 927–45. Butters, T. D., L. A. G. M. van den Broek, et al. (2000). ‘‘Molecular requirements of imino sugars for the selective control of N-linked glycosylation and glycospingolipid biosynthesis.’’ Tetrahedron Assymetry 11: 113–124. Cabrera-Salazar, M. A., E. Novelli, et al. (2002). ‘‘Gene therapy for the lysosomal storage disorders.’’ Curr Opin Mol Ther 4(4): 349–58. Chujor, C. S. N., K. R. Feingold, et al. (1998). ‘‘Glucosylceramide synthase activity in murine epidermis: quantitation, localization, regulation, and requirement for barrier homeostasis.’’ J Lipid Res 39(2): 277–285. Cox, T., R. Lachmann, et al. (2000). ‘‘Novel oral treatment of Gaucher’s disease with N-butyldeoxynojirimycin (OGT 918) to decrease substrate biosynthesis.’’ Lancet 355(9214): 1481–5. Cox, T. M., J. M. Aerts, et al. (2003). ‘‘The role of the iminosugar N-butyldeoxynojirimycin (miglustat) in the management of type I (non-neuronopathic) Gaucher disease: a position statement.’’ J Inherit Metab Dis 26(6): 513–26. Dobrenis, K. (2004). Cell-mediated delivery systems. Lysosomal Disorders of the Brain. F. M. Platt and S. U. Walkley. Oxford, Oxford University Press: 339–380. ff of oral miglustat (Zavesca, Elstein, D., C. Hollak, et al. (2004). ‘‘Sustained therapeutic effects N-butyldeoxynojirimycin, OGT-918) in type 1 Gaucher disease.’’ J Inherit Metab Dis In Press. Erikson, A., C. G. Groth, et al. (1990). ‘‘Clinical and biochemical outcome of marrow transplantation for Gaucher disease of the Norrbottnian type.’’ Acta-Paediatr-Scand 79(6–7): 680–5. Futerman, A. H. and R. E. Pagano (1991). ‘‘Determination of the intracellular sites and topology of glucosylceramide synthesis in rat liver.’’ Biochem J 280(Pt 2): 295–302. Galbiati, F., B. Razani, et al. (2001). ‘‘Emerging themes in lipid rafts and caveolae.’’ Cell 106(4): 403–11. Ginzburg, L., Y. Kacher, et al. (2004). ‘‘The pathogenesis of glycosphingolipid storage disorders.’’ Semin Cell Dev Biol 15(4): 417–31. Ichikawa, S. and Y. Hirabayashi (1998). ‘‘Glucosylceramide synthase and glycosphingolipid synthesis.’’ Tr Cell Biol 8(5): 198–202. Ichikawa, S., N. Nakajo, et al. (1994). ‘‘A mouse B16 melanoma mutant deficient in glycolipids.’’ Proc Natl Acad Sci U S A 91(7): 2703–7. Ichikawa, S., K. Ozawa, et al. (1998). ‘‘Molecular cloning and characterization of the mouse ceramide glucosyltransferase gene.’’ Biochem Biophys Res Commun 253(3): 707–11. Ichikawa, S., H. Sakiyama, et al. (1996). ‘‘Expression cloning of a cDNA for human ceramide glucosyltransferase that catalyzes the first glycosylation step of glycosphingolipid synthesis.’’ Proc Natl Acad Sci USA 93(10): 4638–43. Inokuchi, J. and N. S. Radin (1987). ‘‘Preparation of the active isomer of 1-phenyl-2-decanoylamino3-morpholino-1-propanol, inhibitor of murine glucocerebroside synthetase.’’ J Lipid Res 28(5): 565–71. Jeyakumar, M., T. D. Butters, et al. (1999). ‘‘Delayed symptom onset and increased life expectancy in Sandhoff disease mice treated with N-butyldeoxynojirimycin.’’ Proc. Natl.Acad. Sci. U.S.A. 96(11): 6388–6393. Jeyakumar, M., T. D. Butters, et al. (2002). ‘‘Glycosphingolipid lysosomal storage diseases: therapy and pathogenesis.’’ Neuropathol Appl Neurobiol 28(5): 343–57.

New Developments in Glycosphingolipid Storage Diseases

125

Jeyakumar, M., D. Smith, et al. (2004). ‘‘NSAIDS incraese survival in the Sandhoff disease mouse: synergy with NB-DNJ.’’ Annals of Neurology In Press. Jeyakumar, M., R. Thomas, et al. (2003). ‘‘Central nervous system inflammation is a hallmark of pathogenesis in mouse models of GM1 and GM2 gangliosidosis.’’ Brain 126(Pt 4): 974–87. Kobayashi, T. and Y. Hirabayashi (2000). ‘‘Lipid membrane domains in cell surface and vacuolar systems.’’ Glycoconj J 17(3 –4): 163–71. Krivit, W., J. H. Sung, et al. (1995). ‘‘Microglia: the effector ff cell for reconstitution of the central nervous system following bone marrow transplantation for lysosomal and peroxisomal storage diseases.’’ Cell-Transplant 4(4): 385–92. Lachmann, R. H. (2003). ‘‘Miglustat. Oxford GlycoSciences/Actelion.’’ Curr Opin Investig Drugs 4(4): 472–9. Lachmann, R. H. and F. M. Platt (2001). ‘‘Substrate reduction therapy for glycosphingolipid storage disorders.’’ Exp. Opin. Invest. Drugs 10: 455–466. Maccioni, H. J., J. L. Daniotti, et al. (1999). ‘‘Organization of ganglioside synthesis in the Golgi apparatus.’’ Biochim Biophys Acta 1437(2): 101–18. Meikle, P. J., J. J. Hopwood, et al. (1999). ‘‘Prevalence of lysosomal storage disorders.’’ JAMA 281(3): 249–54. Moyses, C. (2003). ‘‘Substrate reduction therapy: clinical evaluation in type 1 Gaucher disease.’’ Philos Trans R Soc Lond B Biol Sci 358(1433): 955–60. Munro, S. (2003). ‘‘Lipid rafts: elusive or illusive?’’ Cell 115(4): 377–88. Neufeld, E. F. (2004). Enzyme replacement therapy. Lysosomal disorders of the brain. F. M. Platt and S. U. Walkley. Oxford, Oxford University Press: 327–338. Pagano, R. E., V. Puri, et al. (2000). ‘‘Membrane traffic in sphingolipid storage diseases.’’ Traffic 1(11): 807–15. Platt, F. M. and T. D. Butters (2004). Inhibition of substrate synthesis: a pharmacological approach for glycosphingolipid storage disease therapy. Lysosomal Disorders of the Brain. F. M. Platt and S. U. Walkley. Oxford, Oxford University Press: 381–408. Platt, F. M., G. R. Neises, et al. (1994a). ‘‘N-butyldeoxynojirimycin is a novel inhibitor of glycolipid biosynthesis.’’ J Biol Chem 269(11): 8362–5. Platt, F. M., G. R. Neises, et al. (1994b). ‘‘N-butyldeoxygalactonojirimycin inhibits glycolipid biosynthesis but does not affect ff N-linked oligosaccharide processing.’’ J Biol Chem 269(43): 27108–14. Platt, F. M. and S. U. Walkley (2004). Lysosomal defects and storage. Lysosomal Disorders of the Brain. F. M. Platt and S. U. Walkley. Oxford, Oxford University Press: 32–49. Radin, N. S. (1996). ‘‘Treatment of Gaucher disease with an enzyme inhibitor.’’ Glycoconj J 13(2): 153–7. Ringden, O., C. G. Groth, et al. (1995). ‘‘Ten years’ experience of bone marrow transplantation for Gaucher disease.’’ Transplantation 59(6): 864–70. Sandhoff, ff K. and T. Kolter (2003). ‘‘Biosynthesis and degradation of mammalian glycosphingolipids.’’ Philos Trans R Soc Lond B Biol Sci 358(1433): 847–61. Sango, K., S. Yamanaka, et al. (1995). ‘‘Mouse models of Tay-Sachs and Sandhoff diseases differ ff in neurologic phenotype and ganglioside metabolism.’’ Nat Genet 11(2): 170–6. Schiffmann, ff R. and R. O. Brady (2002). ‘‘New prospects for the treatment of lysosomal storage diseases.’’ Drugs 62(5): 733–42. Simpson, M. A., H. Cross, et al. (2004). ‘‘Infantile-onset symptomatic epilepsy syndrome caused by a homozygous loss-of-function mutation of GM3 synthase.’’ Nature Genetics In Press. Snyder, E. Y., G. Q. Daley, et al. (2004). ‘‘Taking stock and planning for the next decade: realistic prospects for stem cell therapies for the nervous system.’’ J Neurosci Res 76(2): 157–68. Trinchera, M., M. Fabbri, et al. (1991). ‘‘Topography of glycosyltransferases involved in the initial glycosylations of gangliosides.’’ J-Biol-Chem 266(31): 20907–12 issn: 0021–9258. Wada, R., C. J. Tifft, ff et al. (2000). ‘‘Microglial activation precedes acute neurodegeneration in Sandhoff disease and is supressed by bone marrow transplantation.’’ Proc Natl Acad Sci USA 97(20): 10954–10959. Walkley, S. U. (2004). ‘‘Secondary accumulation of gangliosides in lysosomal storage disorders.’’ Semin Cell Dev Biol 15(4): 433–44. Walkley, S. U., M. Zervas, et al. (2000). ‘‘Gangliosides as modulators of dendritogenesis in normal and storage disease-affected ff pyramidal neurons.’’ Cereb Cortex 10(10): 1028–37. Warnock, D. E., M. S. Lutz, et al. (1994). ‘‘Transport of newly synthesized glucosylceramide to the plasma membrane by a non-Golgi pathway.’’ Proc Natl Acad Sci U S A 91(7): 2708–12.

126

F. M. Platt et al.

Winchester, B. (2004). Primary defects in lysosomal enzymes. Lysosomal Disorders of the Brain. F. M. Platt and S. U. Walkley, Oxford University Press: 81–130. Winchester, B. and G. W. Fleet (1992). ‘‘Amino-sugar glycosidase inhibitors: versatile tools for glycobiologists.’’ Glycobiology 2(3): 199–210. Wraith, J. E. (2004). Clinical aspects and diagnosis. Lysosomal disorders of the brain. F. M. Platt and S. U. Walkley. Oxford, Oxford University Press: 50–77. Wu, Y. P. and R. L. Proia (2004). ‘‘Deletion of macrophage-inflammatory protein 1 alpha retards neurodegeneration in Sandhoff disease mice.’’ Proc Natl Acad Sci U S A 101(22): 8425–30. Yamashita, T., R. Wada, et al. (1999). ‘‘A vital role for glycosphingolipid synthesis during development and differentiation.’’ ff Proc. Natl. Acad. Sci. USA 96: 9142–9147. Zervas, M., K. L. Somers, et al. (2001). ‘‘Critical role for glycosphingolipids in Niemann-Pick disease type C.’’ Curr Biol 11(16): 1283–7. Zhang, X. and F. L. Kiechle (2004). ‘‘Review: Glycosphingolipids in health and disease.’’ Ann Clin Lab Sci 34(1): 3–13.

22

FUCOSYLATED GLYCANS IN INNATE AND ADAPTIVE IMMMUNITY

J. B. Lowe Department of Pathology, the Life Sciences Institute, and the Howard Hughes Medical Institute University of Michigan Ann Arbor, MI, USA

Mammalian glycans are characterized by fucose modifications. Some of these are constitutively expressed, whereas others are under temporal, developmental, and lineage-specific control. These modifications include fucose in alpha1,3-linkage to aspragine-linked, lipid-linked and serine/threonine-linked glycans. Others include fucose in direct linkage to serines and threonines of EGF-like domains, and thrombospondin repeat domains, in several of proteins. To address the functions of these fucose modifications in vivo, we have created and characterized mice with targeted null mutations in genes that control glycan and protein fucosylation. Mice with deletions on a pair of alpha1,3fucosytransferases exhibit deficits in selectin-dependent neutrophil and T lymphocyte recruitment in acute inflammation, with an accompanying faulty innate immunity. Adaptive immunity in these mice is also faulty, due to defective L-selectin counter-receptor activity in HEVs of peripheral nodes, and a corresponding deficit in homing of naive T lymphocytes. In mice that are homozygous for a null allele at the FX locus, GDP-fucose synthesis is conditionally active only when fucose is supplied to the mice in their chow or water. In the absence of exogenous fucose, these mice exhibit selectin counter-receptor defects akin to those observed in the alpha1,3fucosytransferase deficient mice. Fucose-deficient FX null mice also display numerous other abnormalities, including a reversible thymic atrophy phenotype that is apparently consequent to loss of Notch signaling activity in the pathway the controls thymocyte development from multipotent thymic progenitors. These observations provide mechanistic insight into important functional roles for fucose-modifications in mammals.

127 John S. Axford (ed.), Glycobiology and Medicine, 127-128. © 2005 Springer. Printed in the Netherlands.

128

J. B. Lowe

REFERENCES 1. Haltiwanger RS and Lowe JB. Role of glycosylation in development. Annu Rev Biochem. 73:491–537, 2004. 2. Haines N and Irvine KD. Glycosylation regulates Notch signalling. Nat Rev Mol Cell Biol. 4:786–797, 2003. 3. Smith PL, Myers JT, Rogers CE, Zhou L, Petryniak B, Becker DJ, Homeister JW, and Lowe JB. Conditional control of selectin ligand expression and global fucosylation events in mice with a targeted mutation at the FX locus. J Cell Biol. 158:801–815, 2002.

23

NEW INSIGHTS INTO RHEUMATOID ARTHRITIS ASSOCIATED GLYCOSYLATION CHANGES

Azita Alavi, Andrew J. Pool and John S. Axford Biochemistry and Immunology Academic Unit for Musculoskeletal Disease St Georges Hospital Med School Cranmer Terrace, London, SW17 0RE, UK

1. INTRODUCTION The link between RA, reduced b1,4-Galactosyltransferase (GTase) enzyme activity (1–7), and immunoglobulin G (IgG) hypo-galactosylation is a well documented phenomenon that may be linked to the pathology of rheumatoid arthritis (RA; 3, 8–11). The IgG molecule represents one of the most powerful effector ff components of the immune system and as such its structural diversification via changes in its, conserved, oligosaccharide constituents is of fundamental importance in our understanding of pathological mechanisms associated with diseases such as RA. Glycosylation changes have been shown to have profound effects ff on the stability, conformation, antigenicity and, consequently, the overall function of IgG. It is therefore, not surprising that physiological IgG exists as a population of glycoforms, each conveying different ff physical and or biochemical properties that may result in functional diversity. The concise informational package of the complex glycans and the significant degree of heterogeneity that theses highly variable groups of branched ring structures confer to the protein backbone is governed by a finely tuned mechanism that relies on the action and interplay of a group of enzymes known as the glycosidases and glycosyltransferases.

1.1. b1,4-Galactosyltransferase Subfamily b1,4-Galactosyltransferase is a subfamily of the glycosyltransferase super-family, which comprise of at least seven members (12–15). The most widely distributed, 129 John S. Axford (ed.), Glycobiology and Medicine, 129-138. © 2005 Springer. Printed in the Netherlands.

130

A. Alavi et al.

principle, member of this family, and the most extensively studied in terms of both RA and the control of galactosylation, is the classical GTase which has recently been designated as GTase-I (13, 15). The other, more recent, additions (GTase-II – GTaseVII) identified by primary sequence similarities (15), are differentially ff expressed, often present at comparatively low levels, and are still under investigation with regard to their biological significance, if any, in terms of the control of galactosylation (14, 16–18). GTase has a wide range of biological functions, and has been extensively studied in relation to RA and its role in the galactosylation of the terminal Nacetylglucosamine on the complex N-linked biantennary oligosaccharides located in the CH2 domain of IgG (1–6, 12, 19). These sugars are an integral feature of IgG and are known to affect ff various, Fc-mediated, eff ffector functions (20).

1.2. Reduced GTase Activity in RA Reduced B cell GTase activity and the corresponding decreased IgG galactosylation (IgG-G0) in RA, and various animal models (3, 5, 7–8, 10), appear to be directly linked to the pathogenic features associated with RA (9, 11). For example, agalactosylation of IgG can trigger the inappropriate activation of complement (21), is an important component of rheumatoid factor-IgG complexes in RA (22, 23) and has been shown to be pathogenic in animal models of this disease (8, 11). In addition, IgG-G0 has been found to be a significant diagnostic and prognostic feature of RA, which together with rheumatoid factor status predicts a more severe disease (24–26). GTase, in particular GTase-I, has been shown to be the principal regulator of IgG galactosylation as demonstrated by the specific alteration of b1-4GTase-I expression in a human IgG-secreting cell line, via transfection with sense / antisense human b1-4GTase-I cDNA (12). GTase galactosylation of IgG is a pre-secretory event that occurs, principally, within the trans-Golgi apparatus (27, 28). In addition to its biosynthetic function within the cell, GTase is also involved in a wide variety of other complex biological processes, including cell-cell and cell-matrix interactions, and is therefore widely distributed and important in cell migration, matrix formation and signal transduction (28, 29). GTase is, therefore, present in the Golgi as well as on the cell surface and in a soluble form in various body fluids e.g. serum (30). Structurally, GTase consists of a short positively charged region attached to a tightly folded globular (catalytic) domain which is attached to the membrane by a heavily glycosylated stem region (27, 28). The enzyme is synthesised in the rough endoplasmic reticulum and has a half-life of #20 hr, as a membrane bound glycoprotein in the Golgi apparatus, after which it is released (via the proteolytic cleavage of the stem region) in a catalytically active soluble form (31). Serum GTase composition is, therefore, reflective of GTase at the cellular level (27, 30, 31).

1.3. Regulation of GTase Activity in RA Regulation of GTase activity is complex (32) and controlled in part by transcription, translation and post-synthetic modifications e.g. the degree of phosphorylation (33). However, to date, no evidence has been found to explain the reduction in GTase

Serum Galactosyltransferase Isoenzyme Changes In Rheumatoid Arthritis

131

activity in RA. Studies aimed at transcriptional and translational control have found no evidence of reduced GTase mRNA expression (19) or reduced amount of GTase protein in RA B Cells (6). Furthermore, there is also no evidence of unique B cell polymorphisms of the GTase gene or the gene controlling phosphorylation in RA patients (34), or any evidence for an intracellular inhibitor of GTase in RA (5). Together, these findings would suggest that the reduction in GTase activity in RA, is unlikely to be due to either a genetic abnormality or reduced expression of the enzyme, and points to the possibility of post-synthetic regulatory modifications. Potential mechanisms for post-translational regulation of GTase are evident (35), and could give rise to various isoforms of GTase that may exhibit altered kinetics (35, 36). In this respect it is interesting to note that a number of serum GTase isoforms exist. At least 12 isoforms with broad and similar acceptor specificities have been demonstrated in both health and disease (36–39). This heterogeneity may be due to the fact that GTase is a sialoglycoprotein. It contains #10% carbohydrates, with one N- (40) and a variety of O- linked glycan chains (mucin-type sugar chains: ABO & Lewis blood group determinants) that are expressed in accord with the blood group (41) and thus highly heterogeneous (42). Sialic acid content, however, appears to be the principal determinant of the observed charge heterogeneity of GTase isoforms, e.g. in cancer (38) and although its function is unknown, it may influence the enzymatic activity of GTase, which has been shown to be influenced by its pI value (35–36). The possibility that RA may be associated with the differential ff expression of GTase isoforms was suggested by pilot studies in which we have previously demonstrated (a) the differential ff incorporation of galactose onto a variety of different ff acceptor molecules, and (b) significant changes in the gross isoelectric focusing (IEF) profiles (pH 3–10) of i) soluble serum and ii) peripheral B cell (CD19), GTase derived from a group of patients with RA and healthy individuals (43).

1.4. IEF Profiling of Serum GTase In a recent study (44) we extended our preliminary pilot IEF investigations and used a liquid phase IEF system to determine whether there were GTase isoforms specific to, or elevated in RA, and investigated whether these changes could be attributed to possible sialylation differences. ff For this purpose we used serum samples from patients with 1) RA (n=9), 2) a disease control group (n=9) and a group of healthy individuals (n=10). We used Rotofor IEF to separate serum GTase by charge and used a 2% ampholyte solution; pH specificity of 4–6. The fractions were harvested (20 in total ), neutralised and assayed using a well characterised GTase assay that uses Uridine diphospho-d-[6-3H] galactose as the donor sugar and ovalbumin as the acceptor molecule. To standardise the data from each assay, the enzyme activity (cpm×103) of each sample was expressed as a percentage of the sum of the activity of all the fractions (% Total GTase activity). In order to ascertain whether sialylation had a role to play, both serum and purified GTase (bovine) were desialylated with Clostridium Perfringens sialidase and then assayed using an ampholyte with a pH range of 5–7.

132

A. Alavi et al.

Figure 1. Percentage difference ff in the IEF profile of the RA group compared to the 2 control groups; DC (shaded) & HI (black).

1.5. GTase Isoforms GTase isoforms have previously been identified in both HI and malignancy, where a large degree of the observed charge heterogeneity has been attributed to changes in the degree of sialylation (35, 36, 38). Our aim was to ascertain whether there were RA associated serum changes in the GTase isoform composition and whether these changes could be attributed to possible sialylation differences. ff Whole serum which represents secreted GTase from a number of different ff cell types, including lymphocytes, was used in preference to purified GTase, as the purification procedure may have selectively removed certain GTase isoforms in preference to others (37). Solution phase IEF demonstrated highly significant ( p<0.0001) differences ff between the groups studied with the RA IEF profile, being significantly different ff from that of the disease control group as well as the healthy controls (Fig. 1). Analysis of variance applied to each pair of groups demonstrated that the difference ff in the shape of the IEF profile was highly significant for RA vs. DC or HI ( p<0.0001). There was no significant difference ff between the IEF profiles of the DC and HI groups.

1.6. Differential ff Expression of GTase Isoforms in RA The IEF differences ff were found to be predominantly the result of changes in the pI of two distinct, albeit broad, peaks of GTase activity (Fig. 2) and point towards the possibility that the RA patients are synthesising larger quantities of the relatively more acidic GTase isoforms.

Serum Galactosyltransferase Isoenzyme Changes In Rheumatoid Arthritis

133

Figure 2. Representative IEF profiles: (a) a RA patient; (b) a disease control patient; and (c) a healthy individual.

1.7. RA is Associated with a More Acidic GTase IEF Profile Further analysis of the two main peaks of GTase activity revealed that the RA IEF differences ff were due to acidic shifts in the pI of the two main peak of activity (Fig. 3).

134

A. Alavi et al.

Figure 3. Percentage total GTase activity and pH value of (a) Peak 1 and (b) Peak 2 for RA patients (black) and HI (white). Bars represent mean and standard error. The RA peak 2 was significantly more acidic (p<0.01) and constituted a greater proportion of the total activity (p<0.01), when compared to the HI peak.

Peak one Although, the peak in the RA group was more acidic, there were no significant differences ff in the mean pI of this peak between the three groups examined 4.49 [range 4.30–4.65], 4.63 [range 4.38–4.86] and 4.60 [range 4.40–5.00] for the RA, DC and HI respectively (Fig. 3a). There were also, no significant differences ff in the activity associated with this peak, which constituted 19.7% [range 17.2–26.0], 16.9% [range 12.1–22.8] and 17.7% [range 9.3–27.3], of the total GTase activity in the RA, DC and HI respectively. Peak two The RA peak was significantly more acidic when compared to the peak in the DC ( p<0.05) and the HI group ( p<0.01; Fig. 3b) and constituted a significantly greater proportion of the total GTase activity (RA: 16.1% [range 13.3–18.8]), when compared to the second peak in the DC (13.5% [range 10.2–18.1; p<0.05]) and HI (12.6% [range 8.9–17.7]; p<0.01) group.

1.8. Are Theses GTase Isoforms Unique to RA? The acidic isoforms seen in RA are unlikely to be unique to this disease, as a small percentage of both DC and HI show a second peak in a similar position. This is not surprising since a previous study examining GTase isoforms in HI and malignancy, has shown that HI express most isoforms of the enzyme (30). The results of our studies and those of others would, in fact, indicate that the difference ff between disease states lies in the relative quantities of each individual isoform. On comparing the IEF profile of RA in relation to the disease control group, it is apparent that the association between the acidic GTase isoform and RA is especially strong. Signifying that the RA associated changes in the GTase isoform profile are specific and not due to inflammation per se as all the DC patients investigated had active disease at the time of sampling. Interestingly, when comparing serum RA GTase profiles with those found in malignant disease, there appears to be a number of similarities, in particular, the presence of a large and prominent peak at pH 4.75 and the absence of a peak

Serum Galactosyltransferase Isoenzyme Changes In Rheumatoid Arthritis

135

Figure 4. The effect ff of sialidase treatment on the IEF profile of serum GTase in 3 patients with RA (inactive; shaded circle, active RA1; black squares, and active RA2; black triangle) and 2 HI (white symbols): (a) Pre-desialylation and (b) Post-desialylation (bracket bars represent the range of activity for each peak). Sialidase treatment resulted in similar IEF profiles of GTase in both RA and HI.

normally found at pH 5.10 (45), suggesting that isoforms with greater overall negative charge predominate in both cancer and RA patients.

1.9. What is the Principal Cause of the Observed Charge Heterogeneity? Sialidase treatment of sera from both RA patients and HI with markedly differing ff GTase isoform profiles, lead to three peaks of activity with comparable pH values (Fig. 4), which would suggest that the acid shift in the RA profile, is due in part, to the presence of isoforms with higher sialic acid content. The fact, that the differences ff in GTase isoform profiles, in both RA and malignant disease (30, 36), appear to be primarily due to the presence of increased quantities of hypersialylated GTase, suggests that there may be a common mechanism by which this occurs. The most likely explanation is the high levels of sialyltransferase found in both RA and malignancy (46, 47).

1.10. Do These Hypersialylated Isoform(s) Exhibit Altered Kinetics? The enzymatic activity of purified GTase increased by an average of 75% following desialylation (Table 1). The finding that GTase activity can be influenced by its degree of sialylation supports earlier observations, demonstrating that the enzymatic activity of GTase may, in part, be influenced by its pI value; isoforms with low pI values having decreased, and those with higher pI values having increased enzyme activity (36).

A. Alavi et al.

136

Table 1. The effect ff of desialylation on the activity of purified GTase. GTase activity is expressed as Mean counts per minute (cpm) of duplicate assays

GTase (m Units) 3.10 6.25 12.50 25.00 50.00

GTase Activity Sialylated (cpm×103)

GTase Activity Desialylated (cpm×103)

% Increase in Activity

7.94 16.12 33.97 79.15 190.61

14.38 25.79 63.24 138.46 326.30

81 60 86 75 71

CONCLUSION In conclusion, the data from our recent studies demonstrate quantitative and qualitative changes in the RA serum GTase isoform profile. These changes appear to be, predominantly, due to increased synthesis of a greater proportion of hypersialylated isoforms, which have the potential to adversely affect ff the catalytic activity of the enzyme, thus providing a possible mechanism for post-translational regulation of GTase activity in RA. It also provides further evidence that RA glycosylation changes may be more general than previously indicated and encompass proteins other than IgG.

REFERENCES 1. Axford JS, Mackenzie L, Lydyard PM, Hay FC, Isenberg DA, Roitt IM. Reduced B-cell galactosyltransferase activity in rheumatoid arthritis. Lancet 1987;2(8574):1486–8. 2. Furukawa K, Matsuta K, Takeuchi F, Kosuge E, Miyamoto T, Kobata A. Kinetic study of a galactosyltransferase in the B cells of patients with rheumatoid arthritis. Int Immunol 1990;2(1):105–12. 3. Axford JS, Sumar N, Alavi A, Isenberg DA, Young A, Bodman KB, et al. Changes in normal glycosylation mechanisms in autoimmune rheumatic disease. J Clin Invest 1992;89(3):1021–31. 4. Wilson IB, Platt FM, Isenberg DA, Rademacher TW. Aberrant control of galactosyltransferase in peripheral B lymphocytes and Epstein-Barr virus transformed B lymphoblasts from patients with rheumatoid arthritis. J Rheumatol 1993;20(8):1282–7. 5. Alavi A, Axford J. Evaluation of beta 1,4-galactosyltransferase in rheumatoid arthritis and its role in the glycosylation network associated with this disease. Glycoconj J 1995;12(3):206–10. 6. Keusch J, Lydyard PM, Berger EG, Delves PJ. B lymphocyte galactosyltransferase protein levels in normal individuals and in patients with rheumatoid arthritis. Glycoconj J 1998;15(11):1093–7. 7. Alavi A, Axford JS, Hay FC, Jones MG. Tissue-specific galactosyltransferase abnormalities in an experimental model of rheumatoid arthritis. Ann Med Interne (Paris) 1998;149(5):251–60. 8. Rademacher TW, Williams P, Dwek RA. Agalactosyl glycoforms of IgG autoantibodies are pathogenic. Proc Natl Acad Sci U S A 1994;91(13):6123–7. 9. Young A, Sumar N, Bodman K, Goyal S, Sinclair H, Roitt I, et al. Agalactosyl IgG: an aid to differential ff diagnosis in early synovitis. Arthritis Rheum 1991;34(11):1425–9. 10. Alavi A, Arden N, Spector TD, Axford JS. Immunoglobulin G glycosylation and clinical outcome in rheumatoid arthritis during pregnancy. J Rheumatol 2000;27(6):1379–85. 11. Kuroda Y, Nakata M, Hirose S, Shirai T, Iwamoto M, Izui S, et al. Abnormal IgG galactosylation in MRL-lpr/lpr mice: pathogenic role in the development of arthritis. Pathol Int 2001;51(12):909–15. 12. Keusch J, Lydyard PM, Delves PJ. The effect ff on IgG glycosylation of altering beta1, 4– galactosyltransferase-1 activity in B cells. Glycobiology 1998;8(12):1215–20.

Serum Galactosyltransferase Isoenzyme Changes In Rheumatoid Arthritis

137

13. Lo NW, Shaper JH, Pevsner J, Shaper NL. The expanding beta 4-galactosyltransferase gene family: messages from the databanks. Glycobiology 1998;8(5):517–26. 14. Guo S, Sato T, Shirane K, Furukawa K. Galactosylation of N-linked oligosaccharides by human beta-1,4-galactosyltransferases I, II, III, IV, V, and VI expressed in Sf-9 cells. Glycobiology 2001;11(10):813–20. 15. Amado M, Almeida R, Schwientek T, Clausen H. Identification and characterization of large galactosyltransferase gene families: galactosyltransferases for all functions. Biochim Biophys Acta 1999;1473(1):35–53. 16. Kido M, Asano M, Iwakura Y, Ichinose M, Miki K, Furukawa K. Normal levels of serum glycoproteins maintained in beta-1, 4-galactosyltransferase I-knockout mice. FEBS Lett 1999;464(1– 2):75–9. 17. Kotani N, Asano M, Iwakura Y, Takasaki S. Knockout of mouse beta 1,4-galactosyltransferase- 1 gene results in a dramatic shift of outer chain moieties of N-glycans from type 2 to type 1 chains in hepatic membrane and plasma glycoproteins. Biochem J 2001;357(Pt 3):827–34. 18. Kotani N, Asano M, Iwakura Y, Takasaki S. Impaired galactosylation of core 2 O-glycans in erythrocytes of beta1,4-galactosyltransferase knockout mice. Biochem Biophys Res Commun 1999;260(1):94–8. 19. Jeddi PA, Bodman-Smith KB, Lund T, Lydyard PM, Mengle-Gaw L, Isenberg DA, et al. Agalactosyl IgG and beta-1,4-galactosyltransferase gene expression in rheumatoid arthritis patients and in the arthritis-prone MRL lpr/lpr mouse. Immunology 1996;87(4):654–9. 20. Jefferis ff R, Lund J. Interaction sites on human IgG-Fc for FcgammaR: current models. Immunol Lett 2002;82(1–2):57–65. 21. Malhotra R, Wormald MR, Rudd PM, Fischer PB, Dwek RA, Sim RB. Glycosylation changes of IgG associated with rheumatoid arthritis can activate complement via the mannose-binding protein. Nat Med 1995;1(3):237–43. 22. Soltys AJ, Hay FC, Bond A, Axford JS, Jones MG, Randen I, et al. The binding of synovial tissuederived human monoclonal immunoglobulin M rheumatoid factor to immunoglobulin G preparations of differing ff galactose content. Scand J Immunol 1994;40(2):135–43. 23. Bond A, Kerr MA, Hay FC. Distinct oligosaccharide content of rheumatoid arthritis-derived immune complexes. Arthritis Rheum 1995;38(6):744–9. 24. Isenberg DA, Lydyard PM, Axford JS, Hay FC, Rook GW, Roitt IM, et al. Galactosylation of IgG associated oligosaccharides. Lancet 1988;2(8605):288. 25. van Zeben D, Rook GA, Hazes JM, Zwinderman AH, Zhang Y, Ghelani S, et al. Early agalactosylation of IgG is associated with a more progressive disease course in patients with rheumatoid arthritis: results of a follow-up study. Br J Rheumatol 1994;33(1):36–43. 26. Axford J, Gunnane G, FitzGerald O, Martin Bland J, Bresnihan B, Frears ER. Rheumatic disease differentiation ff using immunoglobulin G sugar printing by high density electrophoresis. J Rheumatol 2003; 30(12):2540–6. 27. Alavi A. The Glycosyltransferases. In Abnormalities of IgG glycosylation and immunological disorders. 1996:149–169. 28. Kleene R, Berger EG. The molecular and cell biology of glycosyltransferases. Biochim Biophys Acta 1993;1154(3–4):283–325. 29. Shur BD, Evans S, Lu Q. Cell surface galactosyltransferase: current issues. Glycoconj J 1998;15(6):537–48. 30. Davey RA, Harvie RM, Cahill EJ, Levi JA. Serum galactosyltransferase isoenzymes as markers for solid tumours in humans. Eur J Cancer Clin Oncol 1984;20(1):75–9. 31. Strous GJ. Golgi and secreted galactosyltransferase. CRC Crit Rev Biochem 1986;21(2):119– 51. 32. Dinter A, Berger EG. The regulation of cell- and tissue-specific expression of glycans by glycosyltransferases. Adv Exp Med Biol 1995;376:53–82. 33. Zhang SW, Xu SL, Cai MM, Yan J, Zhu XY, Hu Y, et al. Effect ff of p58GTA on beta-1,4– galactosyltransferase 1 activity and cell-cycle in human hepatocarcinoma cells. Mol Cell Biochem 2001;221(1–2):161–8. 34. Delves PJ, Lund T, Axford JS, Alavi-Sadrieh A, Lydyard PM, MacKenzie L, et al. Polymorphism and expression of the galactosyltransferase-associated protein kinase gene in normal individuals and galactosylation-defective rheumatoid arthritis patients. Arthritis Rheum 1990;33(11):1655–64. 35. Furukawa K, Roth S. Embryonic and adult forms of two galactosyltransferases differ ff in their degrees of sialylation. Eur J Biochem 1985;150(1):175–80.

138

A. Alavi et al.

36. Gerber AC, Kozdrowski I, Wyss SR, Berger EG. The charge heterogeneity of soluble human galactosyltransferases isolated from milk, amniotic fluid and malignant ascites. Eur J Biochem 1979;93(3):453–60. 37. Davey R, Bowen R, Cahill J. The analysis of soluble galactosyltransferase isoenzyme patterns using high resolution agarose isoelectricfocusing. Biochem Int 1983;6(5):643–51. 38. Davey R, Harvie R, Cahill J, Levi J. Serum galactosyltransferase isoenzyme patterns of cancer patients with liver involvement. Br J Cancer 1986;53(2):211–5. 39. Uemura M, Sakaguchi T, Uejima T, Nozawa S, Narimatsu H. Mouse monoclonal antibodies which recognize a human (beta 1–4)galactosyl-transferase associated with tumor in body fluids. Cancer Res 1992;52(22):6153–7. 40. Endo T, Amano J, Berger EG, Kobata A. Structure identification of the complex-type, asparaginelinked sugar chains of beta-D-galactosyl-transferase purified from human milk. Carbohydr Res 1986;150:241–63. 41. Amano J, Straehl P, Berger EG, Kochibe N, Kobata A. Structures of mucin-type sugar chains of the galactosyltransferase purified from human milk. Occurrence of the ABO and Lewis blood group determinants. J Biol Chem 1991;266(18):11461–77. 42. Malissard M, Berger EG. Improving solubility of catalytic domain of human beta-1,4– galactosyltransferase 1 through rationally designed amino acid replacements. Eur J Biochem 2001;268(15):4352–8. 43. Soltys A, Alavi A, Dalziel M, Axford J. Galactosyltransferase isoenzymes in rheumatoid arthritis and healthy individuals [abstract]. Glycosylation & Disease 1994;1:204. 44. Alavi A, Axford JS, Pool A. Serum galactosyltransferase isoform changes in rheumatoid arthritis. J Rheumatol. 2004; 31(8):1513–20. 45. Qian GX, Liu CK, Waxman S. Abnormal isoelectric focusing patterns of serum galactosyltransferase activity in patients with liver neoplasia. Proc Soc Exp Biol Med 1984;175(1):21–4. 46. Kessel D, Allen J. Elevated plasma sialyltransferase in the cancer patient. Cancer Res 1975;35(3):670–2. 47. Basset C, Durand V, Mimassi N, Pennec YL, Youinou P, Dueymes M. Enhanced sialyltransferase activity in B lymphocytes from patients with primary Sjogren’s syndrome. Scand J Immunol 2000;51(3):307–11.

24

PRODUCTION OF COMPLEX HUMAN GLYCOPROTEINS IN YEAST

Tillman Gerngross Engineering and the Department of Biological Sciences Dartmouth College, Hanover, USA

Recent advances in the Glycobiology field have helped to establish a relationship between therapeutic protein function and glycosylation structures. Most of these studies rely on the comparison of mixed glycoforms, which complicate the clear interpretation of distinct structure activity relationships. We describe the use of combinatorial genetic libraries to engineer yeast cells that perform entirely humanlike glycosylation with exceptional fidelity and uniformity. The use of these libraries to elucidate structure function relationships of glycoproteins and the ability to manufacture complex glycoproteins with unprecedented control over glycosylation will be discussed.

139 John S. Axford (ed.), Glycobiology and Medicine, 139. © 2005 Springer. Printed in the Netherlands.

25

RELATIONSHIPS BETWEEN THE N-GLYCAN STRUCTURES AND BIOLOGICAL ACTIVITIES OF RECOMBINANT HUMAN ERYTHROPOIETINS PRODUCED USING DIFFERENT CULTURE CONDITIONS AND PURIFICATION PROCEDURES

C-T. Yuen1, P. L. Storring1, R. J. Tiplady1, M. Izquierdo2, R. Wait3, C. K. Gee1, P. Gerson1, P. Lloyd1, and J. A. Cremata2 1National Institute for Biological Standards and Control Potters Bar, Herts., UK 2Centre for Genetic Engineering and Biotechnology Havana, Cuba 3Kennedy Institute of Rheumatology Hammersmith, London, UK

Eight preparations of recombinant human erythropoietin (rhEPO) with differing ff isoform compositions were produced by using different ff culture conditions and purification procedures. The N-glycan structures of these rhEPOs were analyzed using an HPLC based, with fluorescent detection profiling procedure (Yuen et al, 2002) and identified using matrix-assisted laser desorption ionization mass spectrometry. The specific activities of these rhEPOs were estimated by in vivo and in vitro mouse bioassays. The eight rhEPOs were found to differ ff in their isoform compositions, as judged by isoelectric focusing, their N-glycan profiles, and in their in vivo and in vitro bioactivities. N-glycan analyses identified at least 23 different ff structures among these rhEPOs, including bi-, tri- and tetra-antennary N-glycans, with or without fucosylation or N-acetyllactosamine extensions, and sialylated to varying degrees. Mass spectrometry also indicated the presence of N-glycans with incomplete outer chains terminating in N-acetylglucosamine residues, and of molecular masses consistent with phosphorylated or sulphated oligomannoside structures. The tetrasialylated tetraantennary N-glycan contents of the eight rhEPOs were found to be significantly and positively correlated with their specific activities as estimated by mouse in vivo bioassay, and significantly and negatively correlated with their specific activities as 141 John S. Axford (ed.), Glycobiology and Medicine, 141-142. © 2005 Springer. Printed in the Netherlands.

142

C-T. Yuen et al.

estimated by mouse in vitro bioassay. It is concluded that the tetrasialylated tetraantennary N-glycan content of rhEPO is a major determinant for its in vivo biological activity in the mouse.

REFERENCE Yuen, C-T., Gee, C.K. and Jones, C (2002). Biomedical Chromatography, 16, 247–254.

26

GLYCOSYLATION OF NATURAL AND RECOMBINANT ANTIBODY MOLECULES

Roy Jefferis ff Immunity & Infection University of Birmingham, B15 2TT UK

Antibodies are often referred to as adaptor molecules that link humoral and cellular immunity. Antibody/antigen interactions generate immune complexes that trigger a cascade of inflammatory mechanisms resulting in the elimination of pathogens and resolution of infection. These inflammatory reactions are generally protective, however, unless effectively ff regulated they can also cause ‘‘bystander’’ damage. Inflammatory cascades are triggered by interactions of complexed IgG-Fc with one or more effector ff ligands, e.g. cellular receptors (FccRI, FccRIII, FccRIII), the C1q component of complement, mannan binding lectin (MBL), the neonatal receptor (FcRn), the mannose receptor (MR) etc [1–4]. The inflammatory cascade can include antibody-dependent-cellular-cytotoxicity (ADCC), complement-dependent-cellularcytotoxicity (CDCC), phagocytosis, the oxidative burst, release of inflammatory mediators etc. The profile of inflammatory (protective) reactions mediated by the human antibody isotypes has been determined through research programmes spanning decades. However, these studies have been conducted, mostly, in vitro and employed heterologous systems, e.g. the use of guinea pig complement. It is not possible to readily extrapolate from these studies to antibody-mediated mechanisms activated in the intact animal since the response will be comprised of multiple antibody isotypes, differing ff epitope specificities, affinity etc. The advent of recombinant antibody therapeutics challenges us to anticipate the effector ff functions optimal for a disease indication and to select the antibody isotype accordingly. Further, we may generate new antibody constructs optimized to activate a given inflammatory cascade(s). A majority of antibody drugs approved for clinical use have been based on the human IgG molecule. The effector ff functions activated by chimeric or humanized antibodies can be selected, in part, through the choice of the human IgG subclass employed; it is unequivocally established that appropriate glycosylation of the IgG-Fc is essential. 143 John S. Axford (ed.), Glycobiology and Medicine, 143-148. © 2005 Springer. Printed in the Netherlands.

144

R. Jefferis ff

Figure 1. The alpha backbone structure of human IgG showing functional regions – see text; light chain: grey; heavy chains: black.

The basic structure of the IgG molecule is of two light and two heavy chains in covalent and non-covalent association to form three independent protein moieties connected thorough a flexible linker or hinge region, Fig. 1. Two of these moieties are of identical structure and each expresses an antigen specific binding site, the Fab regions; the third, the Fc, expresses interaction sites for ligands that activate clearance mechanisms. N-linked glycosylation of the IgG-Fc through Asn is a defining 297 feature and is essential to the expression of multiple effector ff activities [1–4]. For polyclonal human IgG ca. 10 – 20% of Fab moieties bear N-linked oligosaccharides. The glycosylation motifs are present in the variable regions of the light or heavy chains and may be germline encoded or introduced through somatic mutation [4, 5]. Effector ff mechanisms mediated through FccRI, FccRII, FccRIII, C1q, MBL and MR are ablated or severely compromised for aglycosylated IgG-Fc. The influence of glycosylation on FcRn binding and activation appears to be minimal.

1. IgG-Fc GLYCOSYLATION The oligosaccharide moiety is of the complex diantennary type and exhibits heterogeneity with respect to terminal sugars, Fig. 2. The minimal structure observed for normal human IgG is a heptasaccharide having terminal N-acetylglucosamine residues (full ‘‘bond’’ lines). The possible addition of fucose, galactose, bisecting N-acetylglucosamine and sialic acid (dotted ‘‘bond’’ lines), generates the multiple glycoforms present in polyclonal and monoclonal IgG preparations. There are unique features associated with IgG-Fc glycosylation. The site of attachment, Asn 297, is proximal to the N-terminal region of the C 2 domain, and H the oligosaccharide ‘‘runs forward’’ with terminal sugar residues being exposed at the C 2/C 3 domain interface. The oligosaccharide appears to be ‘‘enclosed’’ H H within the protein structure and has defined secondary/tertiary structure [6]. This

Glycosylation of Natural and Recombinant Antibody Molecules

145

Figure 2. The structures of the possible diantennary oligosaccharide structure attached to IgG-Fc at asparagine 297 (Asn ). The core ‘‘core’’ heptasaccharide present in normal human IgG is shown in 297 upright type; this generates the G0 structure. Additional sugar residues that may be attached to the ‘‘core’’ are shown in italic type; thus G0F represents a fucosylated GO (G0F) oligosaccharide.

is due to multiple non-covalent interactions between the oligosaccharide and the protein surface of the C 2 domains. In other glycoproteins oligosaccharides are H attached to asparagine residues exposed on the surface of the protein, are highly mobile and interact with the aqueous phase. Analysis of neutral oligosaccharides released from normal polyclonal IgG reveals twelve principle structures [7]. The possible addition of sialic acid provides for a total of 21 structures. Since each heavy chain may bear a unique oligosaccharide a total of>400 glycoforms can be anticipated. It remains to determine the significance of this heterogeneity for polyclonal antibody populations and its impact on the activation of inflammatory cascade(s) for antibodies to a given antigen (pathogen). Insights into the significance of IgG antibody glycoforms for functional activity is being obtained from experiences with recombinant antibody molecules Since posttranslational modifications of proteins are species, tissue and site specific the glycosylation of recombinant antibody molecules varies between production vehicles, e.g. CHO, NSO cells etc. Analysis of currently licensed therapeutic antibodies, produced in CHO, NSO or Sp2/0 reveals simple glycoform profiles in which the G0F glycoform predominates. The terminal sugar residue of G0F glycoforms is N-acetylglucosamine; there is evidence that immune complexes incorporating this glycoform may bind and activate mannan-binding protein, with the activation of a pseudo-classical complement pathway, and cellular mannose receptor, facilitating uptake by dendritic cells [8 – 10]. By contrast there is evidence that the fully galactosylated glycoform (G2F) facilitates placental transfer and complement activation [11, 12]. An additional heterogeneity may be present in the form of the presence or absence of the C-terminal lysine residue has been removed. Cell engineering has been employed to generate glycoforms of IgG present in normal polyclonal IgG that are absent or present in small yield from CHO cells.

146

R. Jefferis ff

Transfection of CHO cells with the N-acetylglucosamine III transferase gene (GTIII) generated an anti-neuroblastoma IgG antibody with high level of bisecting N-acetylglucosamine that had 10–20 fold increased FccRIII mediated ADCC efficacy [13]; a similar improvement was reported for the therapeutic anti-CD20 antibody (Rituximab) [14]. Dramatic increases in FccRIII mediated ADCC efficacy have been reported for glycoforms of antibodies lacking core fucose [15–17]. Interestingly, the increased efficacy was restricted to FccRIII mediated activities. It has also been shown that addition of sialic acid in an a(2
6) configuration, as opposed to a(2
3), results in improved FccRI and C1 activities [18]. The finding of increased FccRIII mediated ADCC for IgG antibodies: (i) with bisecting N-acetylglucosamine and (ii) the absence of fucose suggests that the most efficient glycoform would be an antibody combining both these structural attributes. It is interesting to note that such glycoforms are essentially absent from normal polyclonal human IgG. The contribution of oligosaccharides to structure and function of IgG has been probed for a series of homogeneous antibody glycoforms and revealed that a simple trisaccharide may confer both stability and function to the IgG-Fc [19 – 21]. These studies show that whilst the influence on function is considerable it results from very subtle conformational differences. ff It was concluded that the lower hinge region of the IgG molecule exhibits ‘‘plasticity’’ and may exist as an equilibrium of conformers that allows for multiple ligand binding specificities. Functional activities and glycoform profiles may also be manipulated by alanine replacement of amino acid residues that make non-covalent interactions with the oligosaccharide [22]. Interaction sites for the FccRI, FccRII, FccRIII and C1q ligands have been ‘‘mapped’’ to the amino acid residues in the lower hinge region and the hinge proximal region of the C 2 domain. Whilst the isolated IgG-Fc fragment expresses H FccRI, FccRII, FccRIII and C1q binding properties x-ray crystallographic analysis shows this hinge proximal region of protein structure to be disordered [6]. Previously predicted interaction sites for FccRIII have been validated through x-ray crystallographic analysis of a complex of IgG1-Fc and a soluble form of FccRIII. It shows asymmetric binding of the receptor to each of the lower hinge heavy chain regions such that they become ordered. Interestingly, there is no significant contact with the oligosaccharides although their presence is essential to the formation of complex [23, 24].

2. IgG-Fab GLYCOSYLATION It is established that 15 – 20% of polyclonal human IgG molecules bear N-linked oligosaccharides within the IgG-Fab region, in addition to the conserved glycosylation site at Asn 297 in the IgG-Fc (4, 25). There are no consensus sequences for N-linked oligosaccharide within the constant domains of either the kappa or lambda light chains or the C 1 domain of heavy chains, therefore, when present they are H attached in the variable regions of the kappa (Vk), lambda (Vl) or heavy (V ) H chains; sometimes both. In the immunoglobulin sequence database ~20% of IgG V regions have N-linked glycosylation consensus sequences (Asn-X-Thr/Ser; where X can be any amino acid except proline). Interestingly, these consensus sequences

Glycosylation of Natural and Recombinant Antibody Molecules

147

are mostly not germline encoded but result from somatic mutation – suggesting positive selection for improved antigen binding (4, 5). A monoclonal human IgG, isolated the serum of patient with multiple myeloma, was shown to bear oligosaccharide in the V region, in addition to the normal L IgG-Fc. The IgG-Fc oligosaccharide could be removed with PNGase F but the V L oligosaccharide was resistant; in contrast the V oligosaccharide but not the IgG-Fc L oligosaccharide was removed on exposure to endoglycosidase F. The therapeutic Erbitux (Cetuximab) also exhibits IgG-Fab glycosylation, at Asn 88 in the V region. H Only the IgG-Fc oligosaccharide could be removed with PNGase F. In each case oligosaccharide analysis of the IgG-Fc and IgG-Fab shows that whilst the IgG-Fc bears mainly G0F oligosaccharides the IgG-Fab bears complex diantennary sialylated structures. Thus, there is distinct site specificity to the glycoforms generated. Hypogalactosylation of polyclonal IgG-Fc has been reported for a number of inflammatory autoimmune diseases. An extreme example is in Wegener’s granulomatosis and microscopic polyangiitis [26]. Oligosaccharide analysis of the IgG-Fc and IgG-Fab shows that whilst the IgG-Fc bears mainly G0F oligosaccharides the IgG-Fab bears complex diantennary sialylated structures. This suggests that whilst the galactosylation machinery is intact environmental factors are exerting an influence on IgG-Fc galactosylation. The functional significance for IgG-Fab glycosylation of polyclonal IgG has not been fully evaluated but data emerging for monoclonal antibodies suggests a positive, neutral or negative influence on antigen binding. These studies demonstrate that the micro-environment, in vivo, can have a profound influence on the glycosylation profile of the IgG-Fc. This may reflect the unique structural relationship between the oligosaccharide and the protein. The ‘‘core’’ heptasaccharide is essential for FccRI, FccRII, FccRIII and C1 activation whilst outer arm sugar residues can influence these and other functions, e.g. FccRIII, FcRn, MBL, MR. Thus, fidelity of glycosylation is essential to the effector ff function profile of antibodies. However, the oligosaccharide can function as a structural ‘‘rheostat’’ to generate specific glycoforms exhibiting optimal effector ff activities for a particular disease target.

REFERENCES 1. Jefferis, ff R., Lund, J. & Pound, J. (1998) IgG-Fc mediated eff ffector functions: molecular definition of interaction sites for effector ff ligands and the role of glycosylation. Immunol.Rev. 163:50–76. 2. Jefferis, ff R. (2001) Glycosylation of human IgG antibodies: relevance to therapeutic applications. Biopharm. 14:19–26. 3. Jefferis, ff R. and Lund, J. (2002) Interaction sites on human IgG-Fc for FccR: current models. Immunol.Lett. 82:57–65. 4. Jefferis, ff R. (2005) Glycosylation of Recombinant Antibody Therapeutics. Biotechnology Progress. In press. 5. Dunn-Walters D.; Boursier L.; Spencer J.(2000) Effect ff of somatic hypermutation on potential N-glycosylation sites in human immunoglobulin heavy chain variable regions. Molecular Immunology. 37:107–13. 6. Deisenhofer, J., (1981) Crystallographic refinement and atomic models of a human Fc fragment and its ˚ resolution. complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8-A Biochemistry 20:2361–2370.

148

R. Jefferis ff

7. Jefferis, ff R, Lund, J, Mizitani, H, Nakagawa, H, Kawazoe, Y, Arata, Y & Takahashi, N. (1990) A comparative study of the N-linked Oligosaccharide structures of human IgG subclass proteins. Biochem. J. 268, 529–537. 8. Malhotra. R.; Wormald, M. R.; Rudd, P. M.; Fischer, P. B.; Dwek, R. A.; Sim R. B. (1995) Glycosylation changes of IgG associated with rheumatoid arthritis can activate complement via the mannose-binding protein. Nature Medicine. 1, 237–243. 9. Abadeh, S., Church, S., Dong, S., Lund, J., Goodall, M. and Jefferis ff R. (1997) Remodelling the oligosaccharide of human IgG antibodies: effects ff on biological activities. Biochem.Soc.Trans. 25:S661 10. Dong, X.; Storkus, W. J.; Salter, R. D. (1999) Binding and uptake of agalactosyl IgG by mannose receptor on macrophages and dendritic cells. J.Immunol. 163, 5427–5434. 11. http://www.fda.gov/cder/biologics/review/ritugen112697-r2.pdf 12. Kibe, T., Fujimoto, S., Ishida, C., Togari, H., Okada, S., Nalagawa, H., Tsukamoto, Y Takahashi, N. (1996) Glycosylation and placental transport of IgG. J.Clin.Biochem.Nutrition. 21:57–63. 13. Umana, P, Jean-Mairet J, Moudry R, Amstutz H, Bailey JE. (1998) Engineered glycoforms of an antineuroblastoma IgG1 with optimized antibody-dependent cellular cytotoxic activity. Nat Biotechnol. 17:176–80. 14. Davies, J., Jiang, L., LaBarre, MJ., Anderson, D., Reff, ff M. ( 2001) Expression of GTIII in a recombinant anti-CD20 CHO production cell line: Expression of antibodies of altered glycoforms leads to an increase in ADCC thro’ higher affinity for FcRIII. Biotech.Bioeng. 74:288–294. 15. Shields RL, Lai J, Keck R, O’Connell LY, Hong K, Meng YG, Weikert SH, Presta LG. (2002) Lack of Fucose on Human IgG1 N-Linked Oligosaccharide Improves Binding to Human Fcgamma RIII and Antibody-dependent Cellular Toxicity. J. Biol Chem. 277:26733–40. 16. Shinkawa T, Nakamura K, Yamane N, Shoji-Hosaka E, Kanda Y, Sakurada M, Uchida K, Anazawa H, Satoh M, Yamasaki M, Hanai N, Shitara K. (2003) The absence of fucose but not the presence of galactose or bisecting N-acetylglucosamine of human IgG1 complex-type oligosaccharides shows the critical role of enhancing antibody-dependent cellular cytotoxicity. J Biol Chem. 278:3466–73. 17. Okazaki A, Shoji-Hosaka E, Nakamura K, Wakitani M, Uchida K, Kakita S, Tsumoto K, Kumagai I, Shitara K. (2004) Fucose depletion from human IgG1 oligosaccharide enhances binding enthalpy and association rate between IgG1 and FcgammaRIIIa. J Mol Biol. 336:1239–49. 18. Jassal, R., Jenkins, N., Charlwood, J., Camilleri, P., Jefferis, ff R. and Lund, J. (2001) Sialylation of human IgG-Fc carbohydrate by transfected rat a(2 – 6) sialyltransferase. Biochem. Biophys.Res.Comm. 286:243–249. 19. Mimura Y., Church S., Ghirlando R,. Dong S., Goodall M., Lund J. and Jefferis ff R. (2000) The influence of glycosylation on the thermal stability and effector ff function expression of human IgG1-Fc: properties of a series of truncated glycoforms. Molecular Immunology. 37:697 – 706. 20. Mimura, Y., Sondermann, P., Ghirlando, R., Lund, J., Young, S.P., Goodall, M. and Jefferis, ff R. (2001). The role of oligosaccharide residues of IgG1-Fc in FccIIb binding. J.Biol.Chem. 276:45539–45547. 21. Krapp, S., Mimura, Y., Jefferis, ff R., Huber, R. and Sondermann, P. (2003) Structural analysis of human IgG glycoforms reveals a correlation between oligosaccharide content, structural integrity and Fccreceptor affinity. J.Mol.Biol. 325:979–989. 22. Lund, J., Winter, G., Jones, P.T., Pound, J., Tanaka, T., Walker, M.R., Artymiuk, P.J., Arata, Y., Burton, D.R., Jefferis, ff R. and Woof, J.M. (1991) Human FccRI and FccRII interact with distinct but overlapping sites on human IgG. J.Immunol. 147 2657 – 2662. 23. Sondermann P. Huber R. Oosthuizen V. Jacob U. (2000). The 3.2-A crystal structure of the human IgG1 Fc fragment-Fc gammaRIII complex. Nature. 406(6793):267–273. 24. Radaev S, Motyka S, Fridman WH, Sautes-Fridman C, Sun PD. (2001) The structure of a human type III Fcgamma receptor in complex with Fc. J Biol Chem. 276:16469–77. 25. Youings, A.; Chang, S. C.; Dwek, R. A.; Scragg, I. G. (1996) Site-specific glycosylation of human immunoglobulin G is altered in four rheumatoid arthritis patients. Biochem J. 314, 621–30. 26. Holland, M., Takada, K., Okomoto, T., Takahashi, N., Kato, K., Adu, D., Ben-Smith, A., Harper, L., Savage, C.O.S. and Jefferis, ff R. (2002) Hypogalactosylation of serum IgG in patients with ANCAassociated systemic vasculitis. Clin.exp.Immunol. 129:183–190.

INDEX

a-dystroglycan 97 adaptive immune system vi fucosylated glycans 127–8 alanine scanning mutagenesis 8 allergic alveolitis see pigeon fanciers’ lung anti-citrulinated cyclic peptide vii anti-inflammatory drugs 123 anti-oxidants 123 antibiotic treatment, staphylococcal infection 116 antibody molecules (natural and recombinant), glycosylation of 143–7 antibody-dependent-cellular-cytotoxicity (ADCC) 29, 143 antiviral therapies 1 APLEC 23–4 Arabidopsis thaliana 11 Asn-100 61 Asn-101 61 Asn-120 62, 63(fig.) Asn-25 62, 63(fig.) autoimmune arthritis vi autoimmune rheumatoid arthritis see rheumatoid arthritis

bovine viral diarrhoea virus (BVDV) v antiviral molecules interfering with N-glycosylation, model for study 5–6 model organism for HCV 1 bowels, bacteria, normal/abnormal vii C1-complex 37–8 C1-inhibitor 37 C1g component of complement 143 C-type lectin superfamily (CLSF) 23–4 Caenorhabditis elegans 11 calnexin, pathway 1 calreticulin-bound murine leukaemia virus glycoprotein, gp90, immunogenecity 85–92 cancer, associated glycosylation 113–14 carbohydrate epitope, antibody recognition, HIV vaccine design 7–8 carbohydrate recognition domain (CRD) 35 cartilage breakdown, gelatinase B 48–51 CD44 58–63 HA-binding surface, analysis 59–61 N-glycosylation, regulation by 61–3 structure 59(fig.) CD8 glycoprotein, glycosylation 71–81 CD8 b subunit, function 74–6 CD8 stalk, influence of O-glycosylation upon the extention 77–8 CD8/MHC interaction, O-linked glycosylation, modulated by 76–7 class I MHC, interactions with 72–4 structure 72, 73(fig.) cell-based therapies 119 cellular glycoengineering viii cellular receptors (Fc c RI,Fc c RII,Fc c RIII) 143 ceramide 120(fig.)

bacterial colonisation vii b1,4-Galactosyltransferase (GTas) rheumatoid arthritis 129–36 activity, reduced 130 activity, regulation 130–1 IEF profiling 131–2 isoforms 132–5 subfamily 129–30 biological therapies viii Biomphalaria 10 Bip 92 bone marrow transplantation 119 bovine spongiform encephalopathy in cattle 95

149

150

ceramide glycosyltransferase 117–18 cercariae 10 Cetuximab 147 chondroitin 57 CLIP (class II invariant chain-derived peptide) 85, 87 CNX 91 collagen type-II vi–vii rheumatoid arthritis 45–53 posttranslocational modifications 51–2 collagen type-V 47 collagenase-3 (MMP-13) 47 collectins 21–2, 35 complement classical pathway activation 27, 29 complement system 37(fig.) innate immune response 37–8 complement-dependent-cellular cytotoxicity (CDCC) 143 congenital muscular dystrophy (CMD) 97 type 1C (MDC1C) 97 ‘‘core’’ heptasaccharide viii Creutzfeldt–Jakob disease 95 CRT 91 DC-SIGN vi, 14–16 Dectin-1 23 dendritic cells parasite-derived glycans, recondition 12–13 Th1/Th2 responses 13–14 deoxygalactonojirimycin (DGJ) 1, 5–6 deoxynojirimycin (BuDNJ) 1 Ebola virus v Edman degradation 49, 51(fig.) enteric microflora vii enzyme replacement therapy 119 Erbitux 147 Ero-1 91 ERp57 91 ERp72 91 eukaryotic cells 117 Fabry disease 2, 118 factor VIII (plasma transglutaminase) 37 fibrinogen 37 ficolins 21–2 fucoidin vii, 116 fucosylated glycans, innate and adaptive immunity 127–8 fukutin-related protein (FKRP) 97

Index

2-G12 monoclonal antibody v, 2, 8 galactose-binding lectins 13 galactosylation, abnormal vii galectin-3 108 galectins 13 ganglioside biosynthesis, defects 118 Gaucher disease 2, 118 type-1 vii gelatin 49 gelatin zymography 47, 48(fig.) gelatinase A 45 gelatinase B vi differential ff glycosylation, from neutrophils and breast cancer cells 103–10 rheumatoid arthritis, role in 45–53 cartilage breakdown 48–51 gelatinase B/matrix metalloproteinase-9 vi Gerstmann–Straussler–Scheinker disease 95 glucosidase inhibitors 1–2 glucosphingolipid (GSL) storage disease 2 glycan(s) antigens 10, 11–12(figs) structures vi glycosaminoglycans 57 glycosphingolipid storage diseases vii glycosphingolipid(s) 117–18 biosynthesis and catabolism 117–18 lysosomal storage disease 118 secondary storage 118–19 glycosphingolipid(s) storage disease, new developments 117–23 inflammation as an additional target 122–3 NB-DNJ clinical trials 121–2 substrate reduction therapy 120–1 therapeutic approaches 119, 120(fig.) glycosylation, disease targets and therapy 1–2 glycosylceramide 117, 120(fig.) glycosyltransferases (GT) 10–11, 97 GM1 gangliosidosis 2 GM3 synthase gene, defect 118 gp90, immunogenicity 85–92 heparin 57 hepatitis C virus (HCV) v disease targets and therapies 1–2 P7 ion channel, long alkylchain iminosugars, blocking role 3–4 HIV vaccine design (a template), antibody

Index

recognition of a carbohydrate epitope 7–8 human immunodeficiency virus type-1 (HIV-1) v gp120 antigenic surface, glycosylated 7(fig.) humoral immune response to infection 2 human immunoglobulin G (IgG) 143 alpha backbone structure 144(fig.) human immunoglobulin glycosylation, lectin pathway of complement activation 27–40 hyaladherins 58 hyaloronan 57–66 structure 57(fig.) hypo-glycosylation 97 iminosugar(s) 120 derivatives 1 long alkylchain, HCVp7 ion channel blocking 3–4 morphogenesis inhibitors, viral re-entry prevention 5–6 immunoglobulin A (IgA) 27 glycosylation 32–4 immunoglobulin D (IgD) 27 glycosylation 34–5 immunoglobulin E (IgE) 27 glycosylation 35 immunoglobulin G (IgG) 27 glycosylation 30immunoglobulin G (IgG)-Fab glycosylation 146–7 immunoglobulin G (IgG)-Fc glycosylation 144–6 immunoglobulin M (IgM) 27 glycosylation 30–2 immunoglobulins 27–9 mannose binding lectin (MBL), interactions with 38–40 role 27 solubility 29 structure 28(fig.) synthesis 27–9 immunomodulation 116 innate immune system vi complement system 37–8 fucosylated glycans 127–8 interferon a 5 interleukin-1 45 interstitial collagenase (MMP-1) 47 isoelectric focusing (IEF) 131 killer cell lectin-like receptors 23–4

151

L-selectin vii L-SIGN vi, 14–16 lactoceramide 120(fig.) laminin, hypo-glycosylation 97 Late Onset Tay–Sachs disease 123 LDN 11–12(figs) LDNF 11–12(figs) lectin blotting, glycosylation changes in serum and tissue proteins in cancer 113–14 lectin complement pathway 22 mannose binding lectin (MBL) 35–6 MBL-associated serine proteases (MASPs) 36–8 lectins 12–13 C-type 13 cell-associated vi Lewis-x 13 binding to the CRD of DC-SIGN 14–16 Lex 3 antibodies, autoimmune reactions 12–13 limb girdle muscular dystrophy 2I (LGMD2I) 97 Link module superfamily 58, 60–1 lipoarabinomannan 37 lipopolysaccharide 37 lipoproteins 37 L y49 (Klra) 23–4 macrophage galactose-type lectin (MGL) vi major histocompatibility complex (MHC) class II molecules 85–9 antigen presentation by two distinct populations 88–9 overview and presentation pathway 86(figs) presentation, influence of antigen conformation 87–8 class II-restricted tumour antigen 90–2 processing, role of glycosylation in influencing 91–2 guided processing and immunodominance 89 Man5-GlcNAc2 sugars 25 mannan binding lectin (MBL) 3, 29, 143 pathway of complement activation, biochemistry, biology and clinical implications 21–2 mannan/mannose Binding protein 35 structure 36(fig.)

152

mannose binding lectin (MBL) immunoglobulins 38–40 lectin pathway of complement activation 35–8 structure 36(fig.) mannose receptor (MR) 38, 143 mass spectrometry 49, 51(fig.) mast cell, IgE 29 matrilysins 47 matrix metalloproteinases (MMP) 45, 103–5 structure 46(fig.) MBL-associated serine proteases (MASPs) 21–2, 36–7 lectin pathway of complement activation 36–8 MCF-7 gelatinase B, production and glycosylation analysis 105–10 methicillin-resistant staphylococci 115 minacidia 10 mucin oligosaccharides 101–2 murine mannose receptor (MR), ligand binding, glycosylation influence 25–6 muscular dystrophy vi glycosylation defects 97–8 N7-oxanonyl-6deoxy-DGJ 1 N-butyldeoxynojirimycin (NB-DNJ) vii, 2, 120–1 clinical trials 121–2 N-glycans v oocyte development and function, role in 99–100 N-linked glycosylation v CD44 function, regulation 61–3 natural killer cell gene complex (NKC) 23–4 neoglycoproteins, synthesis of 10–12 neonatal receptor (FcRn) 143 neuronal migration disorder 97 neuronal stem cells 119 neutrophil collagenase (MMP-8) 47, 49, 53(fig.) neutrophil gelatinase B-associated lipocalin (NGAL) 47, 48(fig.) Nkrp1 (Klrb) 23–4 novel anti-HVC molecules 5 O-acetyl sialic acid expression vii O-fucose glycans v oocyte development and function, role in 99–100

Index

O-glycans 76–81 analysis 79(fig.) O-linked glycosylation v O-linked glycosylation CD8 stalk extension, influence upon 77–8 CD8/MHC interaction 76–7 oligosaccharide viii, 37 Oncomelania 10 ovarian cancer vi asymptomatic progression, need for tumour markers 114 oxidative burst 143 oxidoreductases 91 p7 protein 3–4 ion channels formation/activity 3–4 P-selectin vii 1D/2D-PAGE, glycoprotein profile, malignancies 113–14 Pam-Cys 116 pathogen-associated molecular patterns (PAMPs) P 21–2, 37–8 PDI 91 peptidoglycan(s) 37 staphylococcal cell wall 115–16 phagocytosis 143 phorbol-myristate-acetate (PMA) 105–6 pigeon fanciers’ lung vi mucin oligosaccharides 101–2 pigeon intestinal mucin (PIM) 101 polysaccharide skin inflammation vii posttranslocational modifications, human collagen-II 51–2 Praziquantel 9 prion disease 95 prion protein vi glycosylation and GPI anchorage 95–6 protease, cascade 104, 105(fig.) proteoglycans 47 recombinant human erythropoietins viii recombinant human erythropoietins, N-glycan structures and biological activities 141–2 recombinant IgG antibody viii recombinant therapy viii REGA model 50–1 renal cancer incidence 113 metastatic disease, chemotherapy 113

Index

rheumatoid arthritis vi associated glycosylation changes, new insights 129–36 gelatinase B, expression in 45–53 rheumatoid factor vii ribavirin 5 Sandhoff disease 2, 120–3 Schistosoma haematobium 10 Schistosoma japonicum 10 Schistosoma mansoni 10 schistosome egg antigens (SEA) 14 major glycan antigens within 12 schistosome glycans host immune response, major focus 12–13 interaction with the host immune system 9–16 Lewis-x, binding to the CRD of the CD-sign 13–16 neoglycoproteins, synthesis 10–12 parasite-derived glycans, dendritic cells recognition 12–14 schistosome egg antigens, dendritic cells binding, lectins identification 14 schistosomes, life cycle 9–10 schistosomiasis vi, 9 schistosomula 10 scrapie in sheep 95 selectin family of adhesion molecules vii septic arthritis vii serpin 37 sialic acid 131 sialic acid-binding lectins 13 sialidase treatment 135 siglecs 13 staphylococcal arthritis vii staphylococcal infection, carbohydrates and biology 115–16 Staphylococcus aureus 115 Staphylococcus epidermidis 115 Staphylococcus saprophyticus 115

153

stromelysins 47 substrate reduction therapy (SRT) vii, 2, 119–20 Sahdhoff disease, mouse model 120–1 sugars cancer, association with vi viral infections, pathogenesis v synovial fluids, gelatinase B expression 47, 48(fig.) cartilage breakdown 48–51 systemic vasculitis vii T helper cells-2 (Th-2) responses, schistosomiasis 9–16 Tay–Sachs disease 2, 123 thrombin 37 tissue inhibitor of metalloproteinases-1 (TIMP-1) 48, 105 toll-like receptors 12 transmissible spongiform encephalopathies 95–6 trematodes 9 tropical parasitic diseases vi tumour markers, ovarian cancer serum marker 114 tumour necrosis factor a 45 tumour necrosis factor-stimulated gene-6 58, 60(fig.), 63–6 hyaluronan binding to cell surface CD44 64–6 Tyr-105 61 viral infections, pathogenesis v vitamin C, 123 vitamin E 123 van der Waals interactions 38 Wegener’s granulomatosis 147 yeast, production of complex human glycoproteins 139