Advances in Clinical Chemistry, Volume 48

ADVANCES IN CLINICAL CHEMISTRY VOLUME 48 This page intentionally left blank Advances in CLINICAL CHEMISTRY Edited b...

Author: Gregory Makowski

45 downloads 828 Views 1MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

ADVANCES IN CLINICAL CHEMISTRY VOLUME 48

This page intentionally left blank

Advances in CLINICAL CHEMISTRY Edited by GREGORY S. MAKOWSKI Department of Laboratory Medicine University of Connecticut Health Center Farmington, CT, USA

VOLUME 48

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 32 Jamestown Road, London NW1 7BY, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands This book is printed on acid-free paper. ⬁ Copyright ß 2009, Elsevier Inc. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://www.elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-374797-6 ISSN: 0065-2423 For information on all Academic Press publications visit our website at www.elsevierdirect.com Printed and bound in USA 09 10 11 12 10 9 8 7 6

5 4 3 2

1

CONTENTS CONTRIBUTORS

................................................................................

ix

PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi

Clinical Validation of Biomarkers for Predicting Risk STANLEY S. LEVINSON 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Abstract ... ................................................................................... Introduction ................................................................................. RR/OR Ratios as Diagnostic Tools ...................................................... ROC Plots/Curves as a Diagnostic Tool ................................................. Comparison of RR/OR with ROC curves ............................................... Distributions................................................................................. Bayesian Principles.......................................................................... Weaknesses of ROC Analysis ............................................................. Weaknesses of RR/OR ..................................................................... Stand-Alone versus Synergic Biomarkers ................................................ Techniques for Improving Stratification of Synergic Biomarkers ..................... Criteria for Identifying Testing of Clinical Consequence............................... Discussion. ................................................................................... Conclusions .................................................................................. Glossary of Expressions and Explanations .............................................. References. ...................................................................................

1 2 3 4 7 9 9 12 13 15 15 16 19 20 21 22

The Potential Role of Heat Shock Proteins in Cardiovascular Disease: Evidence from In Vitro and In Vivo Studies M. GHAYOUR-MOBARHAN, A.A. RAHSEPAR, S. TAVALLAIE, S. RAHSEPAR, AND G.A.A. FERNS 1. 2. 3. 4. 5. 6.

Abstract ... ................................................................................... Introduction ................................................................................. HSPs and Atherogenesis ................................................................... HSPs and Autoimmunity in Atherogenesis .............................................. Therapeutic Implications................................................................... Conclusions .................................................................................. References. ...................................................................................

v

28 28 34 45 58 59 59

vi

CONTENTS

The Emerging Role of Symmetric Dimethylarginine in Vascular Disease ARDUINO A. MANGONI 1. 2. 3. 4. 5. 6.

Abstract....................................................................................... Introduction.................................................................................. Synthesis, Transport, and Metabolism of ADMA and SDMA ........................ ADMA and the Cardiovascular System .................................................. SDMA and the Cardiovascular System................................................... Discussion .................................................................................... References ....................................................................................

73 74 75 78 79 88 89

Melanocortin-4 Receptor Mutations in Obesity FERRUCCIO SANTINI, MARGHERITA MAFFEI, CATERINA PELOSINI, GUIDO SALVETTI, GIOVANNA SCARTABELLI, AND ALDO PINCHERA 1. 2. 3. 4. 5. 6. 7. 8. 9.

Abstract....................................................................................... Introduction.................................................................................. The Melanocortin System .................................................................. The MC4R ................................................................................... Mutations in the MC4R .................................................................... Functional Alterations of MC4R.......................................................... Clinical Phenotype of MC4R-Mutated Individuals ..................................... Implications of MC4R Mutations in the Clinical Management of Obesity........... Conclusions .................................................................................. References ....................................................................................

95 96 97 99 100 102 102 103 103 104

Proinflammatory Cytokines in CRP Baseline Regulation CARITA M. EKLUND 1. 2. 3. 4. 5. 6. 7.

Abstract....................................................................................... C-Reactive Protein and Inflammation .................................................... Demographic, Metabolic, and Socioeconomic Factors ... .............................. Proinflammatory Cytokines ................................................................ Signaling Through IL Receptors .......................................................... Genetic Polymorphisms..................................................................... Conclusions .................................................................................. References ....................................................................................

111 112 114 118 123 124 124 126

Fetal Skin Wound Healing EDWARD P. BUCHANAN, MICHAEL T. LONGAKER, AND H. PETER LORENZ 1. Abstract....................................................................................... 2. Introduction.................................................................................. 3. Development .................................................................................

138 138 140

CONTENTS 4. 5. 6. 7.

Scarless Fetal Wound Repair Specificity ................................................. Stem Cells . ................................................................................... Cellular Inflammatory Mediators ......................................................... Cytokines . ................................................................................... References. ...................................................................................

vii 141 147 149 151 155

Clinical Relevance of BNP Measurement in the Follow-Up of Patients with Chronic Heart Failure ALDO CLERICO, MARIANNA FONTANA, ANDREA RIPOLI, AND MICHELE EMDIN 1. 2. 3. 4. 5.

Abstract ... ................................................................................... Background and Aim of the Study........................................................ Biochemical and Physiological Properties of B-Type Natriuretic Peptides........... Circulating Levels of B-Type Natriuretic Peptides ...................................... Variations of Plasma B-Type Natriuretic Peptides, Dependent on Pharmacological Treatment, as Surrogate End-Point for Treatment of Patients with HF ......................................................................... 6. Prognostic Relevance of Plasma BNP/NT-proBNP Variations After Treatment ... 7. Meta-Analysis for Overall Mortality Including All Randomized Clinical Trials .... 8. BNP-Guided Therapy in Chronic Heart Failure: Instructions for Use ............... References. ...................................................................................

163 164 165 167

INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

181

168 169 174 175 176

This page intentionally left blank

CONTRIBUTORS Numbers in parentheses indicate the pages on which the authors’ contributions begin.

EDWARD P. BUCHANAN (137), Division Plastic Surgery, Department of Surgery, Stanford University School of Medicine, Stanford, California 94305, USA ALDO CLERICO (163), Scuola Superiore Sant’Anna, 56126 Pisa, Italy; and Gabriele Monasterio Foundation CNR-Regione Toscana, 56126 Pisa, Italy CARITA M. EKLUND (111), Department of Microbiology and Immunology, University of Tampere, Medical School, 33520 Tampere, Finland MICHELE EMDIN (163), Gabriele Monasterio Foundation CNR-Regione Toscana, 56126 Pisa, Italy G.A.A. FERNS (27), Postgraduate Medical School, University of Surrey, Guildford, Surrey GU2 7WG, UK MARIANNA FONTANA (163), Gabriele Monasterio Foundation CNR-Regione Toscana, 56126 Pisa, Italy M. GHAYOUR-MOBARHAN (27), Cardiovascular Research Center, Avicenna Research Institute, Mashhad University of Medical Science (MUMS), Mashhad 91376-73119, Iran; and Department of Nutrition and Biochemistry, Faculty of Medicine, MUMS, Mashhad 91376-73119, Iran STANLEY S. LEVINSON (1), Laboratory Service, Department of Veterans AVairs Medical Center, Louisville, Kentucky 40206, USA; and Department of Pathology and Laboratory Medicine, School of Medicine, University of Louisville, Louisville, Kentucky 40292, USA MICHAEL T. LONGAKER (137), Division Plastic Surgery, Department of Surgery, Pediatric Surgical Research Laboratory, Stanford University School of Medicine, Stanford, California 94305-5148, USA

ix

x

CONTRIBUTORS

H. PETER LORENZ (137), Division Plastic Surgery, Department of Surgery, Pediatric Surgical Research Laboratory, Stanford University School of Medicine, Stanford, California 94305-5148, USA MARGHERITA MAFFEI (95), Dulbecco Telethon Institute at the Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy ARDUINO A. MANGONI (73), Department of Clinical Pharmacology, School of Medicine, Flinders University, Adelaide 5001, Australia CATERINA PELOSINI (95), Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy ALDO PINCHERA (95), Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy A.A. RAHSEPAR, (27), Cardiovascular Research Center, Avicenna Research Institute, Mashhad University of Medical Science (MUMS), Mashhad 91376-73119, Iran; and Department of Nutrition and Biochemistry, Faculty of Medicine, MUMS, Mashhad 91376-73119, Iran S. RAHSEPAR (27), Cardiovascular Research Center, Avicenna Research Institute, Mashhad University of Medical Science (MUMS), Mashhad 91376-73119, Iran; and Department of Nutrition and Biochemistry, Faculty of Medicine, MUMS, Mashhad 91376-73119, Iran ANDREA RIPOLI (163), Gabriele Monasterio Foundation CNR-Regione Toscana, 56126 Pisa, Italy GUIDO SALVETTI (95), Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy FERRUCCIO SANTINI (95), Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy GIOVANNA SCARTABELLI (95), Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy S. TAVALLAIE (27), Department of Nutrition and Biochemistry, Faculty of Medicine, MUMS, Mashhad 91376-73119, Iran

PREFACE I am pleased to present volume forty-eight of Advances in Clinical Chemistry series. In this second volume for 2009, the lead chapter explores the fundamental importance of receiver operator curves as the gold standard for statistical analysis of test diagnostic performance. The next chapter probes the significance of heat shock proteins, a group of highly conserved proteins expressed under stress, as risk factors for development of cardiovascular disease. The role of symmetric and asymmetric dimethylarginine in nitric oxide synthesis is explored in the following chapter which discusses potential impact on vascular homeostasis and vascular disease. An interesting chapter on obesity seeks to explore the involvement of the melanocortin receptor system as an important mediator of leptin eVect on body weight and metabolism. The next chapter investigates the impact of low-grade inflammation and proinflammatory cytokines on C reactive protein. An interesting chapter investigates the unique restoration of extracellular matrix architecture, strength, and function in fetal wound healing. The role of the inflammatory response, cellular mediators, cytokines, and growth factors are elucidated in this fascinating process. We conclude volume forty-eight with a manuscript on the usefulness of BNP to biochemically monitor patients with chronic heart failure. I extend my appreciation to each contributor of volume forty-eight and thank colleagues who participated in the peer-review process. I extend thanks to my Elsevier editorial liaison, Gayathri Venkatasamy. I sincerely hope the second volume of 2009 will be enjoyed by our readers. As always, I invite comments and suggestions for future review articles for the Advances in Clinical Chemistry series. In keeping with the tradition of the series, I would like to dedicate volume forty-eight to my brother Keith. GREGORY S. MAKOWSKI

xi

This page intentionally left blank

ADVANCES IN CLINICAL CHEMISTRY, VOL. 48

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK Stanley S. Levinson*,†,1 *Laboratory Service, Department of Veterans Affairs Medical Center, Louisville, Kentucky 40206, USA † Department of Pathology and Laboratory Medicine, School of Medicine, University of Louisville, Louisville, Kentucky 40292, USA

1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11. 12. 13. 14.

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RR/OR Ratios as Diagnostic Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ROC Plots/Curves as a Diagnostic Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of RR/OR with ROC curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bayesian Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Weaknesses of ROC Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1. Diagnostic Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2. Prognostic Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Weaknesses of RR/OR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stand‐Alone versus Synergic Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Techniques for Improving Stratification of Synergic Biomarkers . . . . . . . . . . . . . . . . Criteria for Identifying Testing of Clinical Consequence . . . . . . . . . . . . . . . . . . . . . . . . Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Glossary of Expressions and Explanations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 3 4 7 9 9 12 12 13 13 15 15 16 19 20 21 22

1. Abstract Background: A useful biomarker should improve clinical management in an economically reasonable way. This should be determined from well‐ designed outcome studies that show clinical management can be altered on 1

Corresponding author: Stanley S. Levinson, e-mail: [email protected] 1

0065-2423/09 $35.00 DOI: 10.1016/S0065-2423(09)48001-6

Copyright 2009, Elsevier Inc. All rights reserved.

2

STANLEY S. LEVINSON

the basis of the biomarker. It is important not to confuse results from testing prior to outcome with outcome studies. Content: This chapter reviews statistical tests used to evaluate studies performed prior to final outcome studies and criteria that assess whether or not a biomarker should be considered for outcome studies at each step. I review how relative risk and odds ratios are related to receiver operator characteristic (ROC) plot analysis. Other statistical techniques such as reclassification and the Hosmer Lemeshow test that have been suggested for evaluation of diagnostic usefulness are considered. Weaknesses of each technique are discussed. Summary: I consider ROC analysis to be a mainstay against which other statistical tests of diagnostic performance should be compared. The importance of expressing data in terms of predictive values is emphasized. Tests showing weak diagnostic associations with a disease are diYcult to evaluate for outcome study application, because there is usually great diVerence in between‐study variance so that the true relationship between the biomarker, its diagnostic ability, and predictive capability are unclear.

2. Introduction The final criterion for defining biomarkers for clinical prognosis should be: can clinical management be altered on the basis of test results leading to improved care in an economically reasonable way while reaching a dimension of quality that is critical for preventing wrong results and wrongful treatment of patients? [1]. This information should be obtained from well‐ designed prospective outcome studies [2, 3]. A recent editorial questioned the overzealous emphasis on p values of 0.05 [4]. But, for clinical studies, even well‐defined statistical significance is not suYcient to impart usefulness, since there is a breach between statistical significance and meaningful diagnostic discrimination. In this chapter, I postulate that the statistical tests used to evaluate studies performed prior to final outcome studies (preliminary studies) help in pointing toward which biomarkers might show suYcient diagnostic discrimination to be evaluated in large expensive prospective studies, not to determine whether these tests should be put into clinical service. In recent years, a plethora of new biomarkers have been proposed, most for predicting risk or progression of disease [5–8], some of which, it seems to me, have been suggested for clinical use on the basis of preliminary studies alone [6, 9–11]. This chapter reviews strengths and weaknesses of statistical techniques that help us decide which biomarkers might be appropriate for application to outcome studies and reviews diagnostic criteria that can be applied at each

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

3

step of the evaluation. Central to these preliminary decisions is an estimation of positive predictive value (PPV). Traditionally, receiver operator characteristic (ROC) analysis has been used to describe and compare the clinical accuracy of biomarkers. ROC analysis expresses data as diagnostic sensitivity and specificity that can easily be translated into useful predictive values (PPV and negative predictive value, NPV). Often, the performance of new markers are gauged by relative risk (RR) or odds (OR) ratios. I will review how RR/OR are related to ROC analysis. I will also review other approaches for assessing discrimination that have been proposed to be more sensitive than ROC analysis that may have merit [12]. In spite of several weaknesses that I will discuss, I am of the opinion that illustration of data in terms of ROC analysis is a good way to appreciate the strength of the diagnostic relationship (or association) between the biomarker and the disease. I will discuss how weak diagnostic associations between disease and a biomarker, that fall into the area of low accuracy by ROC analysis, usually show great interstudy variability so that the true relationships are unclear and predictive capability is poor. Such inconsistent behavior produces a great challenge in terms of cost and likely results for prospective outcome studies. I hope this chapter will provide a better insight for the clinical laboratory practitioner as to how the statistics used for preliminary evaluation of biomarkers translate into meaningful clinical discrimination. Although, in most examples, the focus is on inflammatory biomarkers, these principles apply to biomarkers in general.

3. RR/OR Ratios as Diagnostic Tools In a cohort study, the association between a factor and the occurrence of an event is often depicted as the RR (see Glossary) that requires a known incidence. OR (see Glossary) is the odds that a case is exposed divided by the odds that a control is exposed. OR are usually applied to studies where incidence rates cannot be established, especially cross‐sectional studies. Usually, OR are derived from multivariate logistic type computations, where adjustment of risk for covariant markers (possible confounders) can be made [13]. When determining the usefulness of new biomarkers, it is important to adjust risk for established markers (covariant or possible confounders) to decide if the new marker is independent of the established markers and, therefore, whether or not it adds additional diagnostic discrimination (not necessarily useful diagnostic discrimination). In cohort studies, the adjusted RR can be calculated from the OR [14]. An adjusted RR, known as the hazard ratio, can be obtained for survival analysis censored outcomes [15].

4

STANLEY S. LEVINSON

4. ROC Plots/Curves as a Diagnostic Tool ROC curves compare diagnostic accuracy of methods over all possible sensitivity/specificity pairs by plotting the true positive rate (TPR) which is the diagnostic sensitivity, against that of the false positive rate (FPR) which is 1 (diagnostic specificity). Figure 1 illustrates examples of an idealized series of ROC curves [16]. Each point on the curve/plot corresponds to a diagnostic sensitivity/specificity associated with a specific concentration of biomarker so that a best reference cutoV can easily be determined. ROC analysis can be statistically adjusted for possible confounders just as RR/OR can [17–19]. Each point on the ROC curve also represents a likelihood ratio (see Glossary) [19, 20]. The ROC curve is obtained by plotting the cumulative frequency of the true positive results (TP) which is the proportion of diseased subjects against the false positive results (FP) which is the proportion of nondiseased subjects each correctly diagnosed at various cutoV points. In essence, the number of subjects that are TP and FP are cumulatively added at each concentration of the biomarker and the cumulative sum at each concentration is divided by the total number of TP and FP to give the cumulative frequency. As such the concentration at each sensitivity/specificity pair is known. An example for constructing a ROC curve is shown in Table 1 and Fig. 2. As shown in Fig. 2, although the scale reflecting concentration is not linear, the concentrations for important cutoVs (sometimes called decision levels) can be denoted on a third axis. Generally, the emphasis is on the sensitivity/specificity pair that best fits the clinical paradigm—diagnostically how many FP can we tolerate to reach an optimal number of TP. For any biomarker in which the disease and nondisease distributions overlap, the sensitivity and specificity move in opposite directions over the span of the ROC curve, so there are always trade‐oVs between the two requiring a decision‐level selection that must depend on the clinical circumstances. Nevertheless, since the biomarker concentrations have been used to generate the ROC graph once a sensitivity/specificity pair that best fits the clinical situation is identified the corresponding biomarker concentrations or decision levels are known. Unfortunately, graphs in many articles do not show the biomarker concentrations on a third axis. Some have been concerned that the decision level may not be obvious from an inspection of the ROC plot. An interesting alternative is to plot the sensitivity and specificity on separate curves but on the same graph against concentration—called cumulative distribution analysis [21]. Complicated neural networks, discriminant analysis, and logistic regression techniques have also been used to compute appropriate decision levels [22]. Nevertheless, although theoretically interesting, it is the clinical

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

0 1.0

[X + – 10]

Concentration [X + 20] [X + – – 30]

5

[X + – 40]

171 36

True positive rate (sensitivity)

16 9

0.8 0.95 0.85

3 2

1.5 1.0

0.75

0.6

OR 0.5

0.4

AUROC or c-statistic

0.2

0 0

0.2 0.4 0.6 0.8 False positive rate (1– specificity)

1.0

FIG. 1. Series of idealized ROC curves showing the relationship between the area under the ROC curve (AUROC), also called the c‐statistic and ORs. If the AUROC curve is 1.0, the method is perfectly accurate. In this case, the plot follows the y‐axis into the left corner of the plot. If the area under the ROC curve is 0.5, the method shows no discrimination. The diagonal line with an AUROC of 0.5 illustrates no discrimination. Each point on the curve corresponds to a biomarker concentration (indicated on the top horizontal axis). If the biomarker concentration is increasing with the disease, the concentration on the top axis is decreasing [X] from left to right. If the biomarker concentration is decreasing with disease, the concentration on the top axis is increasing [Xþ] from left to right. Above, as indicated by the arrow are odds ratios (OR) corresponding to AUROC indicated by the arrow just below. The true positive rate is the same as the diagnostic sensitivity, while the false positive rate is equal to 1 (diagnostic specificity). The relationship between AUROC and OR was modified from Pepe et al. [16] with permission.

paradigm applied to the appropriate sensitivity/specificity pair that must ultimately determine the adequacy of an optimal cutoV(s) and no mathematical approach can be used alone to define a decision level. The area under the ROC curve, which for binary outcomes is called the c‐statistic, is equivalent to the nonparametric Mann–Whitney statistic and is not aVected by skewness of the underlying data distribution [23]. The c‐statistic can be used to globally compare tests. The test with the higher c‐statistic is considered the better test. A test with high accuracy generally shows a c‐statistic of 0.9, intermediate accuracy; 0.7–0.9, useful for some

6

STANLEY S. LEVINSON TABLE 1 EXAMPLE OF ROC CURVE CALCULATION FOR A BIOMARKER THAT IS INCREASING IN CONCENTRATION WITH DISEASE

Biomarker concentrationa (mg/dL)

Number of subjectsb

Number of normalsc

Number of diseased

CUM SUM of normals

CUM SUM of diseased

FPRd

TPRd

350 300 250 200 150 125 100 75 50 25

101 80 56 45 40 37 20 15 9 15

1 5 6 5 10 17 10 10 5 14

100 75 50 40 30 20 10 5 4 1

1 6 12 17 27 44 55 65 70 84

100 175 225 265 295 315 325 330 334 335

0.012 0.071 0.143 0.202 0.321 0.534 0.655 0.774 0.833 1.000

0.299 0.522 0.672 0.791 0.881 0.940 0.970 0.985 0.997 1.000

a The data is ranked into concentrations of the biomarker with decreasing or increasing concentration (the highest or lowest concentration first), depending on whether or not the disease increases or decreases with concentration, respectively. For example, LDLC increases with disease while HDLC decreases with disease. b The number of subject with disease and without disease at each concentration is counted. This begins a transformation from actual data to a binary‐type classification based on frequency. Transformation introduces uncertainty and is a weakness of the method. c The number of normal subjects at each concentration is added cumulatively to the next to obtain a series of cumulative (CUM) sums. The diseased subjects are likewise cumulatively summed. d The TPR and FPR is obtained by dividing the CUM SUM at each concentration by the maximum cumulative sum for diseased and normals, respectively, to derive the TPR and FPR for each concentration, each of which is plotted on a separate axis (as shown in Fig. 2).

purposes; and low accuracy of 0.5–0.7 [18]. Nevertheless, ROC curves are more than just the c‐statistic; they are a graphic depiction of all sensitivity/ specificity pairs [18]. Moreover, the curve does not always show an idealized shape [24], so that a cutoV point may show much more discrimination (or less) than implied by the c‐statistic. For example, it was shown that the replacement of low density lipoprotein (LDL) cholesterol (C) by apo B or nonhigh‐density lipoprotein cholesterol (non‐HDLC) caused an adjusted ROC curve to change its area from 0.7 to 0.73–0.74 (Fig. 3) with an increased TPR (diagnostic sensitivity) from about 0.4 to about 0.47 at a diagnostic specificity of 83% [25] which is the cutoV point that equals the National Cholesterol Education Program (NCEP) guideline [26] for an elevated LDLC (about 1300 mg/L) and non‐HDLC (about 1600 mg/L). If such an improvement be accurate, replacement of

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

7

Biomarker concentration (mg/dL) 350 300 250200 1.0

150

125

100

75 50

25

0.8

1.0

True positive rate (sensitivity)

0.8

0.6

0.4

0.2

0 0

0.2

0.4

0.6

False positive rate (1–specificity) FIG. 2. Illustrative plot of a ROC curve. Table 1 shows hypothetical concentrations for a biomarker and describes how to arrange the ranked concentrations, how to generate the cumulative sums and calculate the TPR and FPR from the frequency coincident with (generated by) concentrations of the biomarker.

LDLC by apo B or non‐HDLC in the routine lipid screen would identify 7% more high‐risk persons. Notice, the improvement in diagnostic discrimination of 7% at the critical cutoV value is apparent from the graphic depiction of the curve while the global estimate from the c‐statistic is only 3–4%.

5. Comparison of RR/OR with ROC curves Like ROC curves, OR can be related to TPR and FPR, in that the OR ¼ [TPR/(1TPR)] [(1FPR)/FPR] [16]. Using this information, an approximation as to the relationship between OR and c‐statistic has been estimated as shown in Fig. 1 [16]. Based on the figure, an OR of about 3.0 would be required to reach borderline intermediate accuracy (c‐statistic 0.7) and about 36 (c‐statistic 0.95) to reach high accuracy levels. It was suggested that an OR less than 3.0, corresponding to a c‐statistic of about 0.65, would not be adequate for individual classification [16]. Most of the tests whose

8

STANLEY S. LEVINSON

LDLC (mg/dL) concentration 130 100

0

50

1 After correction for: age, BMI, smoking, BP

True positive rate (sensitivity)

0.9

0.8

Apo B or NonHDLC

0.7 LDLC 0.6

0.5 0.47 0.40

0.4

0.3

0

Auroc curve LDLC

= 0.70

Apo B

= 0.74

NonHDLC

= 0.73

0.1 0.2 0.3 0.4 0.5 False positive rate (1–specificity)

0.6

FIG. 3. ROC analysis of apo B and lipoprotein lipids after adjustment for the traditional risk factors of age, smoking (S), hypertension (BP), and body mass index (BMI). The vertical line at about a FPR of 17% (specificity 87%) represents the sensitivity–specificity points that correspond to the NCEP’s LDLC and non‐HDLC recommended cutoV point of 1300 and 1600 mg/L, respectively. At this point, LDLC curve shows a TPR of about 0.4 (40%) while the apo B and non‐HDLC curves, that are nearly superimposed, show a TPR of about 0.47 (47%). This is a 7% diVerence while the AUROC, shown in the block, shows only a 3–4% diVerence (0.7 vs. 0.73 or 0.74). AUROC curve is a global measure that is equivalent to the c‐statistic. Modified from reference [25], with permission. The adjusted ROC curves were generated from logistic regression equations by the method described in Ref. [18].

discriminations are described in terms of RR/OR fall in the range between 1.0 and 2.0 [6]. These are equivalent to c‐statistics of <0.6 or poor diagnostic discrimination (Fig. 1). Discrimination that would produce ORs of 1:16 or greater are rarely seen in the literature since this would usually be expressed by ROC analysis with c‐statistics greater than 0.80.

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

9

6. Distributions It is important to remember that the distribution (see Glossary) over which the biomarker data are spread is also important in evaluation of discrimination. Too few values usually cause statistical significance to be insignificant, but when many values are presented statistical significance may be reached, but with poor diagnostic discrimination. Even an apparently strong biomarker can be a poor screening test. This is because both RR/OR and ROC analysis often evaluate data that is lumped together on each end of the distribution for the biomarker, so that the eVect of being highly exposed to a biomarker is compared with being slightly exposed while most people in the middle of the distribution are being ignored [27]. This occurs either because there are a few values in the middle or because a few extreme values at the end are causing the biomarker to appear statistically strong when, in fact, diVerentiation in the important area where most of the results fall is poor. For these reasons, it is important to be able to examine the distribution of the data, so that one can assess if it is uniformly distributed. This can be accomplished by displaying the data as scattergrams (plots) or box and whisker plots. Figure 4 illustrates a case in which when the data is expressed in five intervals, it appears that the relationship between CRP concentration and risk of disease is linear, as denoted by the dotted line. But when the data are expressed in 10 intervals (Fig 4, below), a plateau, as indicated by the dotted line, is observed between concentrations of CRP from about 0.64 to 5.17 mg/L. This discrepancy occurs because the extreme values on either end are causing the RR to reach significance, so that when the data are plotted in too few intervals it appears linear, but when more intervals are used, it becomes apparent that there is little discrimination in the important middle where most of the women’s results fall.

7. Bayesian Principles The PPV is the probability of disease in a patient with an abnormal biomarker. The NPV is the probability of not having the disease when the biomarker is normal. Predictive values are sometimes called posterior (or posttest) probability [28]. Prevalence is the pretest probability or proportion of persons in a population at any given point having the disease. Prevalence is also called the prior probability [28]. In the eighteenth century, the Reverend Thomas Bayes developed an equation to assess the probability that an event will actually occur on the basis of prior probability. For medical decision making, Bayesian statistics relate the pretest probability to the posttest

10

STANLEY S. LEVINSON

Range of hs-CRP concentrations (mg/L) < 0.49 3.0

0.49 – 1.08 > 1.08 – 2.09 > 2.09 – 4.19

1

2

3

4

> 4.19 5

Quintiles of hs-CRP

2.5 2.0

Relative risk (RR)

1.5 1.0 3.0

2.5 2.0

1.5 1.0 0.5

1

2

3

Decile of hs-CRP 4 5 6

7

8

9

10

< 0.36 0.36 – 0.64 – 1.00 – 1.46 – 2.02– 2.74 – 3.71 – 5.17– > – 7.73 < 0.64 < 0.1.0 < 1.46 < 2.02 < 2.74 < 3.71 < 5.17 7.73

Range of hs-CRP concentrations (mg/L) FIG. 4. Framingham risk scores adjusted RR for high sensitivity (hs)‐CRP (from the Women Health Study of 27,939) binned into quintiles (above) and deciles (below). The numbers 1–5 (above) and 1–10 (below) indicate each quintile and decile, respectively. The range of hs‐CRP concentrations within each quintile or decile are listed just above or below the appropriate interval, respectively. The dotted line, above, illustrates a linear trend over the entire range of values with quintiles. The dotted line below for deciles illustrates a plateau between about 0.64 and 5.17 mg/L of hs‐CRP, a range that includes about 50% of women between ages 30 and 49 [50]. Sixty percent of women show hs‐CRP concentrations >1 mg/L [51]. Figures were constructed from data expressed in tables [46, 49] modified from Ref. [6], with permission.

probability. With prognostic testing, the prevalence is often not known since a disease has not yet developed, so the incidence may be substituted. The TPR or diagnostic sensitivity and FPR or 1 (diagnostic specificity) are characteristics of the test. Whether drawn from ROC analysis or other means, these parameters are not Bayesian in their own right. When

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

11

diagnostic sensitivity and specificity values are used to calculate predictive values from population prevalence, they enter the realm of Bayesian statistics. As illustrated in Table 2, the following relationships allow classification between prevalence and diagnostic testing [29]: (1) true positives (TP), the number of diseased patients correctly classified by the tests; (2) false positives (FP), the number of patients without the disease misclassified by the test; (3) false negatives (FN), the number of diseased patients misclassified by the test; and (4) true negatives (TN), the number of patients without the disease correctly classified by the test. From these definitions, the diagnostic relationships shown in Table 2 can be derived. It is important to remember, as show in Table 2, that two types of predictive values can be calculated from preliminary testing: a conditional predictive value and a revised or actual predictive value. The diagnostic sensitivity and specificity are inherent properties of the test, while the revised PPV and NPV are dependent on the prevalence of the disease in the population. When a test sample is evaluated, the conditional predictive values will denote the PPV and NPV in the test sample only. Unless a cohort representative of the actual population to be tested is studied, it is necessary to calculate a predictive value that is revised to fit the actual prevalence of the population to be tested. For a disease with a prevalence of 2%, a biomarker with a sensitivity of 99% and a specificity of 99% will give a PPV of only 66.9%, and a biomarker with a specificity and sensitivity of only 50% will give a NPV of 98% [30]. As the prevalence of a disease approaches zero, the PPV of a biomarker approaches zero, while as prevalence approaches 100%, the NPV approaches zero. Generally, the prevalence of disease in population screening is low so that the NPV is very high while the PPV tends to be low. For this reason, even a TABLE 2 DEFINITIONS OF DIAGNOSTIC RELATIONSHIPS Diagnostic sensitivity or TPR ¼ TP/(TP þ FN) Diagnostic specificity or (1 FPR) ¼ TN/(FP þ TN) a Conditional predictive values: Predictive value of a positive test result (PPV) ¼ TP/(TP þ FP) Predictive value of a negative test result (NPV) ¼ TN/(TN þ FN) Revised PPV ¼ a

ðdiagnostic sensitivityprevalenceÞ ðdiagnostic sensitivity prevalenceÞþ ð1specificityÞ ð1prevalenceÞ

Unless a cohort representative of the actual population to be tested is studied, it is necessary to calculate a revised predictive value to fit the actual prevalence of the population tested and since the prevalence of disease in populations is usually low, it is the revised PPV that is all important.

12

STANLEY S. LEVINSON

test with a good sensitivity and specificity of 95% will yield a low PPV and most positive tests will be false positives. Recall, if a test has a diagnostic sensitivity of 95% and a specificity of 95%, and there is a prevalence of 1%, there will be 100 people with disease in 10,000 and the test can identify 95 of them correctly. But, 5% or 495 people of the 9900 without disease will also have positive test results, so the PPV will only be 95/590 100 ¼ 16.1% or 83.9% of the positive results will be false. Thus, in general screening, the revised PPV that is calculated as shown in Table 1 is of utmost importance. As a result, weak diagnostic relationships between a biomarker and disease give rise to low diagnostic sensitivities and thus very low revised PPV.

8. Weaknesses of ROC Analysis Although the focus of this chapter is on prognosis, it is worthwhile to briefly compare diagnostic models with prognostic models since a comparison helps one to better understand how various degrees of accuracy fit into the overall spectrum of clinical discrimination and predictability.

8.1. DIAGNOSTIC MODELS There are some accurate biomarker tests; nevertheless, even those in the range of high accuracy must be carefully evaluated as to its ability for definitive diagnosis. Troponin I for diagnosis of myocardial infarction showed a c‐statistic of 0.99 and a diagnostic sensitivity of 96% at a specificity of >99% [31]. This was confirmed and positive troponins are now required for definitive diagnosis of myocardial infarction [32]. B‐type naturetic peptide (BNP) was shown to have a c‐statistic of about 0.91 for diagnosis of congestive heart failure [33]. At the recommended cutoV point of 100 pg/mL BNP shows a diagnostic sensitivity of about 90% but a diagnostic specificity of only about 75% (FPR of 25% at TPR of 90% midway between c‐statistic of 0.85 and 0.95, Fig. 1). Depending on the prevalence in a selected population, this would result in a large number of false positive results, but more importantly about 10% of people with the disease would be missed. For this reason, this test was considered inaccurate for initial diagnosis of heart failure where echocardiography remains the test of choice [3], but BNP has application in the emergency department [34]. Thus, in spite of a c‐statistic in the high accuracy range, BNP did not meet the requirements for definitive diagnosis, illustrating again the importance of examining the ROC curve for sensitivity/specificity pairs rather than relying on the c‐statistic alone.

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

13

8.2. PROGNOSTIC MODELS In the above examples, ROC analysis was used for diagnostic purposes. Prognosis is more problematic than diagnosis and ROC analysis shows various weaknesses. It has been estimated that prognostic models usually cannot reach a maximum c‐statistic of 1.0 [12]. This is not surprising since in disease a trait is already present, while in prognosis, the condition may not have evolved suYciently to yet express the trait. Moreover, even genetic traits leading to disease remain unexpressed or partially expressed. Calibration compares the actually observed and predicted probabilities. For tests with high accuracy, the c‐statistic for perfectly calibrated models has been estimated to be only between 0.75 and 0.9 [12]. The c‐statistic may be insensitive for detecting small but clinically useful discrimination. The c‐statistic is a ranked nonparametric score that is minimally aVected by the shape of the distribution. A weakness of this approach is that the actual scores are usually transformed into a binary‐type classification (see Table 1 and Fig. 2). If a small number of persons in a cohort exhibit high risk while the preponderance of individuals are at low risk, binary rank‐ based measure do not take this distribution diVerence into account. Moreover, the influence of two pairs on the c‐statistic would be the same although one pair might have a much larger diVerence than the other [12, 35]. It was noted that if, as suggested [16], an RR/OR of 3.0 was required as a strict criterion for inclusion of each additional biomarker in risk prediction, then, most components of the Framingham risk score would be ineligible for inclusion [12]. None of the traditional risk markers of blood pressure, smoking, or lipids achieved a RR 3.0 [12], although modification of each reduces heart disease [26]. ROC analysis does not easily summarize survival relationships as does a hazard ratio [15]. Nevertheless, ROC plots can be constructed from Kaplan– Meier plots [36] or other time to event analysis [37, 38] which has become common practice [5, 39–44].

9. Weaknesses of RR/OR When the relationship between the consequence and the biomarker is weak, the RR/OR is small and the between study variance is large so that diVerent studies may show a wide range of RR/ORs and the true value is unclear. For example, an earlier study showed an adjusted RR for CRP in the fourth quartile of 4.1 [45]. This was a nested case–control study that only measured 366 samples. When all 27,000 samples were measured for CRP, the adjusted RR in the fifth quintile was a much lower 2.3 [46].

14

STANLEY S. LEVINSON

Fig. 1 shows that this diVerence would be reflected by a c‐statistic of about 0.6 at a RR of 2.3 and 0.75 at a RR of 4.1, the diVerence between low and intermediate accuracy, respectively. A later metastudy examining the relationship between CRP and coronary disease found a combined adjusted OR of 1.49 when limiting the data to later studies and concluded that higher RR/ ORs found in earlier study was due to publication bias [47]. Figure 1 indicates an OR of 1.49 would be equivalent to a c‐statistic of about 0.53 which shows little or no useful diagnostic discrimination. Very large samples are currently encouraged for clinical studies. Since the confidence level or confidence interval (see Glossary) [48] is dependent on the sample size, studies of new biomarkers using very large samples often show statistically significant diVerences between disease and nondisease persons with narrow confidence intervals that do not overlap after adjustment in one study but show narrow confidence intervals with no statistically significant relationship in another study. If the relationship between the new biomarker and the disease is weak, this ambiguity can be explained in part by small diVerences in the samples—such as diVerences in sample selection, methods bias, and confounding bias [28]. When the relationship is weak, these covariants may actually show more interstudy variation than the association between the biomarker and the disease. Also, persons at high risk may be on more medications than those at lower risk that might further confound weak relationships. Moreover, it is important to remember that like the ROC curve an OR/RR represents a spectrum of cutoV values. For example, if a biomarker has an OR of 3.0 at a 10% FPR (a good 90% diagnostic specificity), Fig. 1 indicates it would only identify about 25% of the positive cases. On the other hand, if the same biomarker identified 80% of the cases then it would have a FPR of about 60% or specificity of only 40%. Even at an OR of 36, Fig. 1 indicates that if a cutoV with a TPR of about 0.95 was chosen so that almost all positive cases would be identified, the FPR would be very large, about 0.5 (specificity only 50%). This type of exercise illustrates the advantage of examining a spectrum of cutoVs. Often, the RR/OR data are broken into intervals—tertiles, quartiles, or quintiles. The lowest interval is considered normal and given a RR/OR of 1. If there is a proportional association between the disease and the biomarker, the higher the interval the greater the risk. This allows assessment of risk at various concentrations (sort of a poor man’s ROC curve). Results are often expressed in only a few intervals. Expression of data in less than 10 intervals may lead to misinterpretation. This problem is illustrated in Fig. 4, where data showing the relationship between CRP and coronary disease from a cohort of 27,939 women after adjustment from two diVerent publications are shown. When the data are expressed in five intervals [46], it appears that the relationship between CRP

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

15

concentration and risk of disease is linear, as denoted by the dotted line (Fig 3, above). But when the data are expressed in 10 intervals [49](Fig 3, below), a plateau [6], as indicated by the dotted line, is observed between concentrations of CRP from about 0.64 to about 5.17 mg/L. More the 60% of women have CRP concentrations greater than 1 mg/L [50, 51]. Since the prevalence of coronary disease is very low in these women, examination of the data expressed as deciles would suggest that the biomarker would have a very low PPV which was demonstrated as being less than 1% [52], but would not be as apparent from the data expressed as quintiles. Misinterpretation due to too few intervals is not limited to RR/OR, but also applies to ROC curves. The diVerence is that programs that are now available for ROC analysis are usually expressed as a continuous plot of all of the data and intervals are rarely used.

10. Stand‐Alone versus Synergic Biomarkers Generally, a stand‐alone biomarker must have a very high level of discrimination to accurately detect a disease with a c‐statistic greater than 90% such as troponins. Since this level of discrimination is unusual for prognosis, these risk markers are generally synergistic in that several markers are combined to achieve a high discrimination for total risk. Thus, risk for coronary disease is assessed using the established risk factors of age, hypertension, smoking history, lipid status, and body mass index (BMI) [26]. For example, a recent test for polymorphism and progression to type 2 diabetes showed an overall RR of 1.54 [53]. Clearly, as a stand‐alone marker, it would show little or no clinical usefulness since Fig. 1 indicates that an OR of 1.5 is about equivalent to a c‐statistic of about 0.55. This means that at a 15% FPR the TPR would be about 20%. Assuming a prevalence of disease of 1%, this would translate into a PPV of 1.3% ([true positives/true positives þ false positives] 100) (98.7 of every 100 persons identified would be false positive). Nevertheless, such a marker could be useful if it added additional real clinical prediction value to existing markers for diabetes such as BMI or borderline elevated glucose.

11. Techniques for Improving Stratification of Synergic Biomarkers One problem with analyzing associations between genes and other global approaches to identifying relevant biomarkers (i.e., Proteomics) is that these approaches are not hypothesis driven but rather the associations are defined from the data and the hypothesis formulated later [7]. This means the

16

STANLEY S. LEVINSON

relationship is more likely to be fortuitous. On the other hand, the relationship between many new inflammatory markers and disease are hypothesis driven. There is substantial evidence that atherosclerosis is in part an inflammatory disease [54]. As a result, it is not surprising that investigators examined inflammatory markers to determine if these might add additional discrimination to traditional markers for coronary disease [6]. Although hypothesis driven, generally, new inflammatory biomarkers have shown weak incremental diagnostic relationships with coronary disease when added to existing markers (RR/OR < 1.5) [5, 41, 42, 44] that places them in the region of poor accuracy by ROC analysis (Fig. 1). Nevertheless, the RR/ORs may show statistically significant increase. Clinical reclassification has been used to assess the clinical value of synergic prognostic biomarkers that show statistical significance by RR/OR but little diVerentiation according to the c‐statistic [12, 35]. Clinically relevant risk categories are defined and the ability of a new marker to correctly reclassify patients who have been assigned on the basis of an old marker(s) alone is evaluated. The change in estimated risk can then be compared for fit using the Hosmer–Lemeshow test (see Glossary) [35]. Although this approach may have merit, when diagnostic associations with the biomarker are weak, it is limited by the same types of interstudy variability that aVects the RR/OR. For example, in one article, reclassification of persons at intermediate risk for coronary disease to high risk on the basis of CRP was limited to 2.7% [55], whereas, another study showed a reclassification of 12% for those at intermediate risk [12]. Moreover, when the prevalence of a disease is low, reclassification may cause many more persons without the disease to be reclassified into the high‐risk group than persons with the disease, giving rise to worse prediction [56]. Also, the Hosmser–Lemeshow test shows no evidence of lack of fit if the test statistic is p 0.05, with evidence of fit above the 95 percentile. This means there is a good probability of a Type II error (see Glossary) that, according to some, makes this test unsuitable as a mean to assess precise model‐fit [57].

12. Criteria for Identifying Testing of Clinical Consequence It is important to remember that the conclusions drawn from any study are dependent on the samples being studied and many assumptions made regarding the design and appropriateness of the study [28]. Questions that should be asked include: Does the sample of patients, controls, or other comparison groups truly represent the population that one wishes to study? Is the study well designed? and Are those factors that may confound the study appropriately included as covariant markers or otherwise controlled for? If a

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

17

treatment is involved there may be a need for randomization. Outcome studies are expensive and biomarkers selected to be tested must be carefully considered after examination of all of the preliminary data. The weaker the diagnostic relationship between the biomarker and the disease, the larger will be the sample needed and the more expensive the outcome study. If there is no clear treatment based on the concentration of the biomarker, continued evaluation of the test may be pointless. Assuming that these crucial factors are appropriate, criteria for determining the practical usefulness for biomarkers have been published in the form of questions [3], upon which Table 3 is based. Question 1 is usually answered by a preliminary study conducted among a group of patients with the disease and a group without disease. Generally, if the average value of those with disease or the RR/OR is statistically significantly diVerent from those without disease, it is concluded the answer is yes. Nevertheless, before going on to assess question 2, it is important to consider the overlap between the two groups. If there is a great deal of overlap between the groups, as illustrated in Fig 4, below, the test may not have much practical value, an instance of the breach between statistical significance and diagnostic discrimination. If question 1 seems true, question 2 may be tested by examining some patients who have the disease and others who do not [3]. Again, if there is a statistically significant diVerence, it may be concluded that the answer to question 2 is yes, but before going on to test question 3, cutoV values should be drawn from ROC curves or other means so that improvements in diagnostic sensitivity and specificity or reclassification can be evaluated. Revised positive predictive assessment (PPV and NPV) should be calculated based on the actual prevalence [20]. If the results from question 2 suggest the test may be clinically useful, question 3 can be tested in a cohort study, usually retrospective studies on stored samples. At this point, it is important to have the data evaluated with

TABLE 3 QUESTIONS (CRITERIA) FOR DETERMINING CLINICAL CONSEQUENCE 1. Do test results in patients with the target disorder diVer from those in normal people or do the results identify those at increased risk? 2. Does the test identify patients who are suspected of having the disorder or synergistically add discrimination to other testing (can the test be used to make or improve a diagnosis or risk assessment)? 3. Does the test result improve the diagnostic eYciency or risk assessment beyond current testing in what appears to be a clinically useful way? 4. Do patients who undergo this diagnostic test fare better (in their ultimate health outcomes) than similar patients who are not tested and is the test economically reasonable?

18

STANLEY S. LEVINSON

respect to population norms (reference range) for nonaVected people [58]. Great overlap between the normal reference range and the disease group as shown in Fig. 4, below, may reduce the predictive values, producing a test of questionable usefulness. The purpose of the testing and statistical analysis to this point is to decide if it is clinically and economically reasonable to address question 4 that can only be answered by prospective outcome studies. These are expensive studies and biomarkers selected must be carefully considered after examination of all of the preliminary data. Besides the questions proposed in Table 3, another suggestion in helping to decide which tests should be examined in outcome studies is whether or not the test is specific for the condition targeted [58]. It is not necessarily required that the biomarker be specific (plays an etiologic or causal roll in the disease), called a risk factor [59], but, if it is highly nonspecific, it is likely that covariate biomarkers, especially true risk factors, will reduce its eVect. For example, traditional biomarkers for coronary disease include blood pressure and measures of cholesterol [26]. Although the exact mechanisms are not known and synergic increases in c‐statistic may be small [12, 35], evidence indicates that these are, not only predictors, but true risk factors [7, 59, 60]. Moreover, abundant evidence indicates therapeutic moderation of these biomarkers reduces the risk [26, 60, 61]. Clearly, it would be good to have a new test with specificity that adds to or even more reliably predicts coronary disease than the traditional tests but, in spite of evidence that inflammation is an important factor in atherosclerosis, inflammatory biomarkers examined to date do not seem to meet this standard. It is unclear that the markers being measured have any causal relationship. This would classify them as risk markers rather than risk factors [59]. They show little specificity, since they correlate with cancer, liver disease, hormone therapy, various other heart conditions, and chronic and acute diseases [6]. Besides, because of weak associations, they show RR/ORs between 4.1 and 1.0 after adjustment for traditional risk markers [5, 6, 41, 42, 45, 47], so that the true relationship is unclear. A related question that is important is that of absolute versus relative risk (RR). The NCEP guidelines use absolute risk. This risk varies up to >20% per 10‐year period. The use of RR is much more tenuous. For example, in the Woman’s Health Study there were 121 events in about 28,000 women that would result in about 0.4% of women having an event [46]. This is a very low incidence that would not be apparent from the RR data. Thus, it is important in evaluating the meaning of the results to consider the prevalence so that accurate predictive assessments can be made. Such a low prevalence is apt to give rise to many false positives results. The safety‐to‐benefit ratio becomes a greater issue when false positive results that represent low‐risk persons are treated [60].

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

19

Frequentists have been contrasted with Bayesians [62]. Frequentists are considered objectivists who in essence are classical statisticians who do not wish to make statistical inferences beyond which the parameters of the experiment prescribed such as the standard error of the mean, confidence interval, etc. Bayesians have been called subjectivists since they are prepared to make inferences based on prior events or probability. It has been said that in the extreme Frequentists must repeat an experiment an infinite number of times to define exact parameters, which cannot be achieved. Others have said that Bayesians learn from experience, but what makes us think the future will be like the past. What is of interest to us is that classical statistics was intended to apply to a few hundred data point and only to compare a few parameters [62] which means that the classical confidence level of 0.05 was certainly not intended to define diagnostic discrimination when in some cases thousands of subjects are tested for multiple parameters. Yet, Bayesian principles are only needed for preliminary experiments when appropriate cohorts are not available. In appropriate cohort studies and well‐designed outcome studies, the results are in accord with classical objective statistics in that inferences are only from within the experiment. In fact, since very large numbers of subjects are often tested in clinical trials, the estimated parameters should be very tight and, if the entire trial was repeated, although the repeat result cannot be exactly the same as the first, the results should be very similar. If in cohort studies, the results disagree, it must be that the cohorts were suYciently diVerent such that weak relationships between the biomarker and disease discrimination was altered as a result of confounding parameters.

13. Discussion It is important not to confuse the means with the end. That is to say, results from testing prior to outcome studies should not be used to determine which biomarkers are diagnostically useful, but only to determine which biomarkers appear to lend themselves for outcome studies. This chapter has focused only on clinical criteria and has not considered the reliability of the measuring process where it has been shown that common biomarkers may not always be analytically reliable [1, 8]. Some have referred to guidelines similar to those listed in Table 3 as the truth, the whole truth, and nothing but the truth [1]. Many biomarkers can meet the first two criteria. On the average, they diVerentiate between aVected and unaVected persons, even in a sample of persons suspected of having the disease. This may be called the truth. Other tests may appear to meet the criteria listed in question 3 in preliminary studies. This may be considered

20

STANLEY S. LEVINSON

the whole truth. But the acid test is whether the biomarker not only improves the eYciency beyond current testing, but also whether patients who undergo this biomarker test fare better (in their ultimate health outcomes) than similar patients who are not tested. This is nothing but the truth. Few recent biomarkers have been shown to reach this standard.

14. Conclusions Especially important in preliminary analysis are predictive assessments which reach beyond statistical significance. Whether ROC analysis is used or reclassification, predictive estimates based on the actual prevalence or incidence can be made [56]. It has long been recognized that whenever the number of unaVected persons is very large and the test is poorly accurate, PPV will be very poor [30]. In spite of some weaknesses, my view is that ROC analysis is fundamental for assessing prognostic discrimination, and ROC graphs denoting pertinent cutoVs should be published with all papers. With survival studies, although ROC analysis may not summarize the data in a single term as can hazard ratio, for prognosis ROC plots can be developed for important times from the survival analysis [5, 39, 40, 42] that for coronary disease may be about 1, 5, or 10 years. By providing the maximal amount of information in a report, and especially information on predictive assessment, the experimenter is producing the ammunition needed for convincing others. If it can be shown that, although there is little improvement in the ROC curve, there is diVerentiation between groups as defined by RR/OR and reclassification shows improved predictions with results that are consistent between studies, this may be acceptable for initiating an outcome study. Moreover, even if a ROC analysis is not provided in a report, a rough assessment of the discrimination and predictive values can be made by referring to plots similar to those shown in Fig. 1. These conclusions beg the question: would it be more reasonable to test for a disease in an outcome study that shows a 7% increase in diagnostic sensitivity at a critical cutoV value as shown in Fig. 2 for non‐HDLC and apo B, if such a relationship has been confirmed in several studies, or to test for a disease with a weak relationship defined by an increase in RR/OR less than 1.5 but little or no increase in c‐statistic and whose increment in predictive benefit is unclear. This does not mean that the latter test may not have clinical usefulness, but when diagnostic specificity and sensitivity are this poor, the possible usefulness of the test must be very carefully examined. The weaker the relationship, the greater the challenge.

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

21

Glossary of Expressions and Explanations Confidence intervals (CI), distributions of data and sample size: CI contain information similar to confidence levels. They define statistical significance. CI represent a range of values for each variable of interest. If the CI for the diseased group overlaps the average value reflecting no disease, this represents a statistically not significant result. If there is no overlap, the results are significantly diVerent. Thus, the 95% CI is similar to the 0.05 p‐value confidence level obtained from hypothesis testing. CI have an advantage that they emphasize the size of the eVect. If the data distribution is Gaussian, the 95% width of the CI for a two‐sided test is calculated as 1.96 the standard error of the mean (SEM), where the SEM ¼ standard deviation for the distribution/square root of the number of observations (n) [48] and, after calculation, this value is added to and subtracted from the mean to produce the CI. For a ROC curve, a one sided calculation is used so the width of the CI is 1.64 SEM added to and subtracted from the c‐statistic. As with confidence levels, CI is an inverse function of n, becoming very small as the number of observations become very large. If there are too few observations, the CI will be large and the CI for the average eVect for the disease and control groups may overlap one another. As the sample increases in size both the confidence level and CI become smaller. If the distributions are truly diVerent, a statistically significant diVerence will become apparent with suYcient size. A sample of appropriate size can be calculated from the power formula (see Type I and Type II Errors, below). It is important to remember that as the size of the sample grows, if the average eVects stay the same and are truly diVerent, the confidence levels and CI become smaller and eventually show a statistically significant diVerence, but the diagnostic discrimination stays the same so that the c‐statistic and RR/OR do not change. For RR/OR, if the biomarker shows a significant diVerence between disease and no disease, the CI should not overlap 1.0. CI should also accompany c‐statistics when biomarkers are being compared. When data are displayed as RR/OR in binned groups, CI for each bin should be presented and CI can be obtained for sensitivity/specificity points of interest on a ROC curve. This allows accuracy at selected cutoVs to be better evaluated. Hosmer Lemeshow test: A goodness of fit test in which observations are sorted and binned into about 10 groups. Within each group, the estimated observed proportion and average expected frequencies are compared. The statistic has a w2 distribution with g 2 degrees of freedom, where g is the number of bins (24). Like all goodness of fit tests, there is a high chance of making a Type II error.

22

STANLEY S. LEVINSON

Likelihood ratio: Likelihood ratios can be defined as the ratio between the probability of a test result in persons that have the disease and the probability of that result in persons without the disease. Likelihood ratios correspond to slopes on the ROC curve, as such they are defined by the (sensitivity)/(1 sensitivity) at any point [19,20]. Likelihood ratios can be a powerful tool in confirming a diagnosis because the joint likelihood ratio is the product of each individual biomarker. Thus, the likelihood ratios for several diVerent independent biomarkers can be combined to help in confirming or refuting a disease. It is important to remember that like other parameters of ROC analysis, likelihood ratios are not Bayesian until adjusted for prevalence. Thus, there is a conditional and revised likelihood and the revised likelihood can be expressed as a percent [20]. Odds ratio: Odds is the probability of those with disease divided by those without disease, expressed as: subjects with disease/(1 subjects with disease). The odds ratio is the odds that the cases have particular test results divided by the odds that the controls have the particular test result. Relative risk: RR is measured as the ratio of disease incidence in those positive for a particular test (above some cutoV value) and the incidence in those negative for the test (below the cutoV value). Expressed as: Incidence of disease in exposed/Incidence of disease in unexposed. Type I and Type II Error: Type I or alpha (a) error is expressed by the conventional p value. This is the 95% confidence level (p 0.05). This means there is one chance in 20 of making a Type I error. On the other hand, a study might conclude there is no diVerence between the disease group and the control group when, in fact, there is a diVerence. This is a type II or beta (b) error—often referred to as the false negative rate. This usually occurs because the samples were too small. The probability for an adequate sample size is determined by calculating the power: Power ¼ (1 b). ACKNOWLEDGMENT This work was supported by the Department of Veteran AVairs, Louisville, Kentucky, USA.

REFERENCES [1] J.O. Westgard, T. Darcy, The truth about quality: medical usefulness and analytical reliability of laboratory tests, Clin. Chim. Acta 346 (2004) 3–11. [2] B. Bozkurt, D.L. Mann, Use of biomarkers in the management of heart failure: are we there yet? Circulation 107 (2003) 1231–1233. [3] D.L. Sackett, R.B. Haynes, The architecture of diagnostic research, BMJ 324 (2002) 539–541. [4] N. Rifai, D.G. Altman, P.M. Bossuyt, Reporting bias in diagnostic and prognostic studies: time for action, Clin. Chem. 54 (2008) 1101–1103.

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

23

[5] A.R. Folsom, L.E. Chambless, C.M. Ballantyne, J. Coresh, G. Heiss, K.K. Wu, et al. An assessment of incremental coronary risk prediction using C‐reactive protein and other novel risk markers: the atherosclerosis risk in communities study, Arch. Intern. Med. 166 (2006) 1368–1373. [6] S.S. Levinson, Inflammatory and long‐term risk markers, Clin. Lab. Med. 26 (2006) 553–570. [7] J. Loscalzo, Association studies in an era of too much information: clinical analysis of new biomarker and genetic data, Circulation 116 (2007) 1866–1870. [8] D.A. Morrow, J.A. de Lemos, Benchmarks for the assessment of novel cardiovascular biomarkers, Circulation 115 (2007) 949–952. [9] P.M. Ridker, P.W. Wilson, S.M. Grundy, Should C‐reactive protein be added to metabolic syndrome and to assessment of global cardiovascular risk? Circulation 109 (2004) 2818–2825. [10] N. Rifai, P.M. Ridker, Proposed cardiovascular risk assessment algorithm using high‐ sensitivity C‐reactive protein and lipid screening, Clin. Chem. 47 (2001) 28–30. [11] Y.M. Smulders, A. Thijs, J.W. Twisk, New cardiovascular risk determinants do exist and are clinically useful, Eur. Heart J. 29 (2008) 436–440. [12] N. Cook, Use and misuse of the receiver operator characteristic curve in risk prediction, Circulation 115 (2007) 928–935. [13] B. Dawson‐Saunders, R.G. Trapp, Statistical Methods for Multiple Variables in Basic and Clinical Biostatistics. Appleton and Lange, Norwalk, CN, 1990. [14] J. Zhang, K.F. Yu, What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes, JAMA 280 (1998) 1690–1691. [15] S.R. Rao, D.A. Schoenfeld, Survival methods, Circulation 115 (2007) 109–113. [16] M.S. Pepe, H. Janes, G. Longton, W. Leisenring, P. Newcomb, Limitations of the odds ratio in gauging the performance of a diagnostic, prognostic, or screening marker, Am. J. Epidemiol. 159 (2004) 882–890. [17] E.K. Shultz, Multivariate receiver‐operating characteristic curve analysis: prostate cancer screening as an example, Clin. Chem. 41 (1995) 1248–1255. [18] M.H. Zweig, S.K. Broste, R.A. Reinhart, ROC curve analysis: an example showing the relationships among serum lipid and apolipoprotein concentrations in identifying patients with coronary artery disease, Clin. Chem. 38 (1992) 1425–1428. [19] M.H. Zweig, G. Campbell, Receiver‐operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine, Clin. Chem. 39 (1993) 561–577. [20] S.G. Pauker, R.I. Kopelman, Interpreting hoofbeats: can Bayes help clear the haze? N. Engl. J. Med. 327 (1992) 1009–1013. [21] J.S. Krouwer, Cumulative distribution analysis graphs—an alternative to ROC curves, Clin. Chem. 33 (1987) 2305–2306. [22] V.A. DeBari, Computation of decision levels from diVerentiated logistic regression probability curves, Ann. Clin. Lab. Sci. 36 (2006) 194–200. [23] J.A. Hanley, B.J. McNeil, The meaning and use of the area under a receiver operating characteristic (ROC) curve, Radiology 143 (1982) 29–36. [24] J.A. Swets, Measuring the accuracy of diagnostic systems, Science 240 (1988) 1285–1293. [25] S.S. Levinson, Comparison of apolipoprotein B and non‐high‐density lipoprotein cholesterol for identifying coronary artery disease risk based on receiver operating curve analysis, Am. J. Clin. Pathol. 127 (2007) 449–455. [26] Expert panel on detection, evaluation, and treatment of high blood cholesterol in adults, Executive summary of the third report of the National Cholesterol Education Program (NCEP). Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III, JAMA 285 (2001) 2486–2497.

24

STANLEY S. LEVINSON

[27] N.J. Wald, A.K. Hackshaw, C.D. Frost, When can a risk factor be used as a worthwhile screening test? BMJ 319 (1999) 1562–1565. [28] R. Fletcher, S.W. Fletcher, Clinical Epidemiology, Lippincott Williams and Wilkins, Baltimore, 2005. [29] R. John, M.S. Lifshitz, J. Hang, D. Fink, Post‐analysis: medical decision‐making, Henry’s Clinical Diagnosis and Management by Laboratory Methods, 21 ed Saunders Elsevier, Philadelpia, 2007, 69–75. [30] T.J. Vecchio, Predictive value of a single diagnostic test in unselected populations, N. Engl. J. Med. 274 (1966) 1171–1173. [31] J.E. Adams, K.B. Schechtman, Y. Landt, J.H. Ladenson, S. JA, Comparable detection of acute myocardial infarction by creatine kinase MB isoenzyme and cardiac troponin I, Clin. Chem. 40 (1994) 1291–1295. [32] K. Thygesen, J.S. Alpert, H.D. White, A.S. Jaffe, F.S. Apple, M. Galvani, et al. Universal definition of myocardial infarction, Circulation 116 (2007) 2634–2653. [33] A.S. Maisel, P. Krishnaswamy, R.M. Nowak, J. McCord, J.E. Hollander, P. Duc, et al. Rapid measurement of B‐type natriuretic peptide in the emergency diagnosis of heart failure, N. Engl. J. Med. 347 (2002) 161–167. [34] R. Pfister, C.A. Schneider, Natriuretic peptides BNP and NT‐pro‐BNP: established laboratory markers in clinical practice or just perspectives? Clin. Chim. Acta 349 (2004) 25–38. [35] N. Cook, Statistical evaluation of prognostic versus diagnostic models: beyond the ROC curve, Clin. Chem. 54 (2008) 17–23. [36] J.M. Bland, D.G. Altman, Survival probabilities (the Kaplan–Meier method), BMJ 317 (1998) 1572. [37] D.G. Altman, J.M. Bland, Time to event (survival) data, BMJ 317 (1998) 468–469. [38] P.J. Heagerty, T. Lumley, M.S. Pepe, Time‐dependent ROC curves for censored survival data and a diagnostic marker, Biometrics 56 (2000) 337–344. [39] M.J. Bos, C.M. Schipper, P.J. Koudstaal, J.C. Witteman, A. Hofman, M.M. Breteler, High serum C‐reactive protein level is not an independent predictor for stroke: the Rotterdam Study, Circulation 114 (2006) 1591–1598. [40] J.J. Cao, A.M. Arnold, T.A. Manolio, J.F. Polak, B.M. Psaty, C.H. Hirsch, et al. Association of carotid artery intima‐media thickness, plaques, and C‐reactive protein with future cardiovascular disease and all‐cause mortality: the Cardiovascular Health Study, Circulation 116 (2007) 32–38. [41] A.C. St-Pierre, B. Cantin, J. Bergeron, M. Pirro, G.R. Dagenais, J.P. Despres, et al. Inflammatory markers and long‐term risk of ischemic heart disease in men. A 13‐year follow‐up of the Quebec Cardiovascular Study, Atherosclerosis 182 (2005) 315–321. [42] T.J. Wang, P. Gona, M.G. Larson, G.H. Tofler, D. Levy, C. Newton-Cheh, et al. Multiple biomarkers for the prediction of first major cardiovascular events and death, N. Engl. J. Med. 355 (2006) 2631–2639. [43] S. Blankenberg, M.J. McQueen, M. Smieja, J. Pogue, C. Balion, E. Lonn, et al. Comparative impact of multiple biomarkers and N‐Terminal pro‐brain natriuretic peptide in the context of conventional risk factors for the prediction of recurrent cardiovascular events in the Heart Outcomes Prevention Evaluation (HOPE) Study, Circulation 114 (2006) 201–208. [44] I. Tzoulaki, G.D. Murray, A.J. Lee, A. Rumley, G.D. Lowe, F.G. Fowkes, Relative value of inflammatory, hemostatic, and rheological factors for incident myocardial infarction and stroke: the Edinburgh Artery Study, Circulation 115 (2007) 2119–2127. [45] P.M. Ridker, J.E. Buring, J. Shih, M. Matias, C.H. Hennekens, Prospective study of C‐reactive protein and the risk of future cardiovascular events among apparently healthy women, Circulation 98 (1998) 731–733.

CLINICAL VALIDATION OF BIOMARKERS FOR PREDICTING RISK

25

[46] P.M. Ridker, N. Rifai, L. Rose, J.E. Buring, N.R. Cook, Comparison of C‐reactive protein and low‐density lipoprotein cholesterol levels in the prediction of first cardiovascular events, N. Engl. J. Med. 347 (2002) 1557–1565. [47] J. Danesh, J.G. Wheeler, G.M. Hirschfield, S. Eda, G. Eiriksdottir, A. Rumley, et al. C‐reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease, N. Engl. J. Med. 350 (2004) 1387–1397. [48] A.R. Henderson, Chemistry with confidence: should clinical chemistry require confidence intervals for analytical and other data? Clin. Chem. 39 (1993) 929–935. [49] P.M. Ridker, N. Cook, Clinical usefulness of very high and very low levels of C‐reactive protein across the full range of Framingham Risk Scores, Circulation 109 (2004) 1955–1959. [50] E.S. Ford, W.H. Giles, A.H. Mokdad, G.L. Myers, Distribution and correlates of C‐reactive protein concentrations among adult US women, Clin. Chem. 50 (2004) 574–581. [51] S. Woloshin, L.M. Schwartz, Distribution of C‐reactive protein values in the United States, N. Engl. J. Med. 352 (2005) 1611–1613. [52] S.S. Levinson, R.J. Elin, What is C‐reactive protein telling us about coronary artery disease? Arch. Intern. Med. 162 (2002) 389–392. [53] J.C. Florez, K.A. Jablonski, N. Bayley, T.I. Pollin, P.I. deBakker, A.R. Shuldiner, et al. TCF7L2 polymorphisms and progression to diabetes in the Diabetes Prevention Program, N. Engl. J. Med. 355 (2006) 241–250. [54] R. Ross, Atherosclerosis—an inflammatory disease, N. Engl. J. Med. 340 (1999) 115–126. [55] T.J. Wang, M.G. Larson, R.S. Vasan, Biomarkers for prediction of cardiovascular events, N. Engl. J. Med. 356 (2007) 1472–1475. [56] M.S. Pepe, H. Janes, J.W. Gu, Use and misuse of the receiver operating characteristic curve in risk prediction, Letter by Pepe et al. regarding article. Circulation 116 (2007) e132; author reply e134. [57] S.J. Janket, Y. Shen, A.E. Baird, Why must new cardiovascular risk factors be carefully re‐ assessed prior to clinical application? Eur. Heart J. 29 (2008) 1336–1337; author reply 1337. [58] L. Mosca, C‐reactive protein—to screen or not to screen? N. Engl. J. Med. 347 (2002) 1615–1617. [59] T.J. Wang, New cardiovascular risk factors exist, but are they clinically useful? Eur. Heart J. 29 (2008) 441–444. [60] S.M. Grundy, Promise of low‐density lipoprotein‐lowering therapy for primary and secondary prevention, Circulation 117 (2008) 569–573; discussion 573. [61] S.L. Hardoon, P.H. Whincup, L.T. Lennon, S.G. Wannamethee, S. Capewell, R.W. Morris, How much of the recent decline in the incidence of myocardial infarction in British men can be explained by changes in cardiovascular risk factors? Evidence from a prospective population‐based study, Circulation 117 (2008) 598–604. [62] E. Bradley, Bayesians, frequentists, and scientists, J. Am. Stat. Assoc. 100 (2005) 1–5.

This page intentionally left blank

ADVANCES IN CLINICAL CHEMISTRY, VOL. 48

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS IN CARDIOVASCULAR DISEASE: EVIDENCE FROM IN VITRO AND IN VIVO STUDIES M. Ghayour-Mobarhan,*,† A.A. Rahsepar,*,† S. Tavallaie,† S. Rahsepar,*,† and G.A.A. Ferns‡,1 *Cardiovascular Research Center, Avicenna Research Institute, Mashhad University of Medical Science (MUMS), Mashhad 91376-73119, Iran † Department of Nutrition and Biochemistry, Faculty of Medicine, MUMS, Mashhad 91376-73119, Iran ‡ Postgraduate Medical School, University of Surrey, Guildford, Surrey GU2 7WG, UK

1. Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Discovery of the HSPs, Their Classification and Their Functions . . . . . . . . . . . 2.2. Atherosclerosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. HSPs and Atherogenesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. HSPs and Animal Models of Atherogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Modulation of HSP Expression in Cells Involved in Atherogenesis In Vitro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Soluble or Circulating HSPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. HSPs and Autoimmunity in Atherogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. General Consideration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Molecular Mimicry and Relation to Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Antibodies to HSPs and Infections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Antibodies to HSPs and Cardiovascular Risk Factors . . . . . . . . . . . . . . . . . . . . . . 4.5. Antibody Titers to HSPs and Their Relationship to CVD Burden . . . . . . . . . . 4.6. Changes in Titers of HSP Antibodies During Acute Coronary Syndromes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Therapeutic Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

28 28 28 28 34 36 38 44 45 45 47 47 48 54 56 58 59 59

Corresponding author: GAA Ferns, e-mail: [email protected] 27

0065-2423/09 $35.00 DOI: 10.1016/S0065-2423(09)48002-8

Copyright 2009, Elsevier Inc. All rights reserved.

28

GHAYOUR-MOBARHAN ET AL.

1. Abstract The heat shock proteins (HSPs) are highly conserved families of proteins expressed by a number of cell types following exposure to stressful environmental conditions. These conditions include several known risk factors for cardiovascular disease. A number of the HSPs have been shown to be molecular chaperones that are involved in the refolding of other damaged protein molecules. Over the past two decades there has been an increasing interest in the relationship between HSPs and cardiovascular disease, and particularly whether an autoimmune response may be implicated. The fact that microorganisms also produce HSPs, and that these are homologous to human HSPs has given rise to concept of molecular mimicry. While most of the past studies have focused on HSP 65 and 70, there has been recent interest and investigations of the possible role of the smaller HSPs, such as HSP27, in atherogenesis. Furthermore, the possibility that autoimmunity may be mediating the deleterious eVects of HSPs has led some investigators to explore tolerization as a potential therapeutic approach. 2. Introduction 2.1. DISCOVERY OF THE HSPS, THEIR CLASSIFICATION AND THEIR FUNCTIONS Approximately four decades ago, Ritossa and colleagues [1] observed that exposing larval salivary glands from Drosophila to heat induced specific genes in the giant chromosomes of the gland cells; it is now known that these genes encode proteins called HSPs. The HSPs are highly conserved families of proteins found in the cells of all organisms and several of them are known to function as molecular chaperones. The HSPs may be divided into seven major families according to their molecular weights: HSP10, small HSPs (15–30 kDa), HSP40, HSP60, HSP70, HSP90, and HSP100 (Table 1). HSP expression is increased in response to several environmental stresses in addition to heat stress; these include: certain forms of nutritional deficiency, oxidative stress, and ultraviolet radiation. This is mediated by the release of heat shock factor 1 and its binding to heat shock elements in the flanking regions of the HSP genes [2] (Fig. 1). Moreover, in addition to their role as chaperones, HSP have other putative roles [3–6]. Table 1 shows a summary of their functions. 2.2. ATHEROSCLEROSIS Atherosclerosis is a chronic multifactorial disease that underlies the pathophysiology of cardiovascular disease (CVD), stroke and peripheral vascular disease (PVD), and is the major cause of mortality worldwide [7, 8]. It is

TABLE 1 SUMMARY OF THE NOMENCLATURE, LOCATION, AND FUNCTION OF THE MAJOR HEAT SHOCK PROTEIN FAMILIES Family

Organism

HSP‐related proteins

Small HSPs

E. coli S. cerevisiae

Hsp40

E. coli S. cerevisiae Mammals E. coli

Lbp A and B HSP27 A and B crystallin HSP27 DnaJ Ydj 1 Hdj 1 and Hdj 2 GroEL

S. cerevisiae

HSP60

Plants Mammals E. coli S. cerevisiae

Cpn60 HSP60 DnaK Ssa 1–4 Ssb 1,2 Kar2 Ssc1 HSC70 HSP70 BIP MHSP70 HtpG HSP83 HSP90 GRP94

Hsp60

Hsp70

Mammals

Hsp90

E. coli S. cerevisiae Mammals

Hsp100

E. coli S. cerevisiae

Location Cytosol Cytosol Cytosol Cytosol Cytosol Cytosol/nucleus Cytosol mitochondria Chloroplasts mitochondria

Cytosol Cytosol Cytosol ER mitochondria Cytosol/nucleus Cytosol/nucleus ER mitochondria Cytosol Cytosol Cytosol ER Cytosol Cytosol

Functions Suppresses aggregation and heat inactivation of proteins in vitro; confers thermotolerance through stabilization of microfilaments; antiapoptotic activity Essential cochaperone activity with Hsp70 proteins to enhance rate of adenosine triphosphatease activity and substrate release Refolds and prevents aggregation of denatured proteins in vitro; may facilitate protein degradation by acting as a cofactor in proteolytic system; role in the assembly of bacteriophages and Rubisco (an abundant protein in the chloroplast) Roles in lambda phage replication; autoregulation of the heat shock response; interaction with nascent chain polypeptides; functions in interorganellar transport; roles in signal transduction; refolds and maintains denatured proteins in vitro; role in cell cycle and proliferation; antiapoptotic activity; potential antigen‐ presenting molecule in tumor cells

Role in signal transduction (e.g., interaction with steroid hormone receptors, tyrosine kinases, serine/threonine kinases); refolds and maintains proteins in vitro; autoregulation of the heat shock response; role in cell cycle and proliferation Role in stress tolerance; helps the solubilization of heat‐ inactivated proteins from insoluble aggregates

HSP, heat shock protein; E. coli, Escherichia coli; S. cerevisiae, Saccharomyces cerevisiae; ER, endoplasmic reticulum. Modified from Lamb et al. [2]. Publisher and year of copyright: Elsevier, 2002.

GHAYOUR-MOBARHAN ET AL.

30

Regulatory protein complex

Heat shock HSF trimerisation

HSF1 monomer

Nuclear translocation

Binding of HSF to heat shock concensus element (HSE)

TATAA GAAnnTTCnnGAA

FIG. 1. Schematic representation of the regulation of mammalian heat shock protein expression. HSF, heat shock transcription factors; HSE, heat shock consensus element; TATAA, DNA sequence containing TATAA repeats. Reference: [2]. Publisher and year of Copyright: Elsevier, 2002. Permission for reproduction/adaptation was granted by the copyright holder.

characterized by the accumulation of lipids and extracellular matrix in the intima of large and medium sized arteries. It is associated with mononuclear cell infiltration, and smooth muscle proliferation [9]. Risk factors for CVD include: age, male sex, family history of CVD, hypertension, hypercholesterolemia, smoking, diabetes mellitus, socioeconomic status, and obesity [9]. There are several emerging risk factors for CVD including markers of oxidative stress, inflammation, and autoimmunity [10]. 2.2.1. Atherosclerosis and the Role of Inflammation The inflammatory nature of atherosclerosis was first described in the 1850s [11], however, more recent interest has developed because immunocytochemical studies have allowed the cellular composition of atherosclerotic plaques to be determined and related to the onset of clinical events, such as plaque rupture [12]. Furthermore, inflammatory processes also appear to be involved in atherogenesis [13]. The earliest lesions in atherogenesis, are fatty streaks, and these are commonly found in infants and young children [14]. They are characterized by a relative paucity of lipid accumulation and comparative abundance of intimal inflammatory cells that include activated T lymphocytes (helper, suppressor, and regulator), mast cells, macrophages, dendritic cells [15], and less commonly granulocytes and NK cells [16–18].

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

31

Epidemiological studies have supported the role of inflammation in CVD. Serum C-reactive protein (CRP) concentrations have been reported to be a stronger independent predictor of coronary events than low density lipoprotein (LDL) cholesterol levels [19–22]. It has also been reported that elevated levels of soluble intercellular adhesion molecule (ICAM)-I, a marker of endothelial cell activation, are associated with increased coronary risk [23] and its expression is increased in human atherosclerotic lesions [24]. Complement activation [25–27] may play a role in endothelial injury during atherogenesis and may be a consequence of autoimmune responses to modified LDL [28] or denatured HSPs [29]. The expression of human lymphocytic antigen (HLA) class II antigen and secretion of several cytokines, within atherosclerotic lesions supports the involvement of inflammation in atherosclerosis [30]. Advanced atheromatous lesions also contain large numbers of T lymphocytes [30], most of which are T helper (h) type 1 cells bearing alpha/beta receptor [17]. Furthermore, activated T cells bearing gamma/delta receptors are abundant at the earliest stages of atherogenesis [31] and atherosclerosis can be inhibited by depletion of T lymphocytes [32]. Xu et al. [33] have suggested that CD4+ cells predominate within the T cell population in early lesions, while Van Der Wal et al. [18] have reported an increased CD8/CD4 ratio in both early and late lesions. There is a preponderance of pro-inflammatory Th1 cells expressing IFN-g and IL-2 compared to Th2 cells producing interleukin (IL)-4, IL-5, and IL-10 [34, 35]. In apolipoprotein E deficient mice it has been reported that Th1-inhibition is associated with a 60% reduction in atherosclerotic lesion area [36]. Regulatory T cells (Treg) are a subpopulation of T cells which exert important regulatory eVects on immune function [37–39], Type 1 Treg cells can inhibit immune responses by secreting TGF-b and IL-10 [40, 41], while Th2 cells suppress inflammation and dampen macrophage activity via a broader spectrum of anti-inflammatory cytokines, and may have protective eVects against atherogenesis [35, 42–44]. Switching the balance of activity from Th1 to Th2 may therefore be protective in atherogenesis [45]. Depletion of CD4+ and CD8+ T cells has been reported to reduce the formation of fatty streaks in C57BL/ 6J mice [46], which supports the importance of T cells in atherogenesis. However, there remains controversy about the precise role of cellular immunity in atherogenesis as some studies have shown that immune-suppression may result in enhanced atherogenesis in experimental models [47, 48]. 2.2.2. Atherosclerosis and the Role of Infection Several studies have shown a positive association between the degree of atherosclerosis burden and presence of chronic infectious microorganisms [19, 49], these include: the Herpes group of viruses, notably Cytomegalovirus (CMV) and herpes simplex virus type 1 (HSV-1) [50], Helicobacter (H) pylori

GHAYOUR-MOBARHAN ET AL.

32

[51], Chlamydia (C) pneumonia [52], Hepatitis A virus (HAV) [53], and infectious organisms that give rise to gingivitis [54]. These infective processes may exert a pro-atherogenic eVect in early life. Pesonen and coworkers [55] have shown that in young children, the presence of antibodies to several microorganisms was positively associated with carotid intimal thickening, a marker of atherosclerosis. It has been proposed that infection acquired during childhood may lead to atherosclerosis in later life [56]; and it has been reported that there is a positive association between the number of infectious organisms a person has been exposed to and the extent of CVD [57] (Fig. 2). Splenectomy is associated with an increased susceptibility to both infection by organisms such as C. pneumonia and more severe atherosclerosis [58–60]. Individuals with chronic infections have high serum levels of HSP60, which are also associated with severity of atherosclerosis [61]. The potential mechanisms by which infections may induce atherosclerosis and their interaction with other pro-inflammatory processes is shown in Figs. 3 and 4.

Pathogen burden 15.0 9.8

10 5

Mortality (%)

Mortality (%)

P < 0.001 15

Extent of disease

20

20

3.1

0

13.9

15 10 3.5

5 0

0 0–3

4–5

6–8

Control

Number of seropositivities

Limited disease

Advanced disease

25 20.0

Mortality (%)

20 14.9

15 10

7.7 5.9

7.0

5 0

0

0 Control

0

1.4 Limited disease

Number of seropositivities 6–8 4–5 0–3

Advanced disease

FIG. 2. Cardiovascular mortality rate according to pathogen burden and extent of atherosclerosis. Reference: [57]. Publisher and year of Copyright: American Heart Association, 2002. Permission for reproduction/adaptation was granted by the copyright holder.

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

33

Total burden of infection Oral Gastrointestinal P. Gingivalis

H. Pylori

Respiratory C. Pneumoniae

Inflammation IL - 6, TNF-a, CRP Direct infection Natural immunity Molecular mimicry

Atherosclerosis

FIG. 3. Possible mechanisms of infection-induced atherosclerosis. Reference: [234]. Publisher and year of copyright: Faculty of Dental Practitioners, 2007. Permission for reproduction/ adaptation was granted by the copyright holder.

Total burden of infection Gastrointestinal H. Pylori

Oral P. Gingivalis

Respiratory C. Pneumoniae

Obesity Molecular mimicry

Mental stress Inflammation IL – 6, TNF-a, CRP Smoking

Direct infection Natural immunity

Autoimmune disease eg RA Atherosclerosis

Diabetes

FIG. 4. Chronic inflammatory conditions such as smoking, stress, obesity, and rheumatoid arthritis may contribute to the total burden of inflammation and hence to atherosclerosis. Reference: [234]. Publisher and year of copyright: Faculty of Dental Practitioners, 2007. Permission for reproduction/adaptation was granted by the copyright holder.

34

GHAYOUR-MOBARHAN ET AL.

2.2.3. Autoimmunity in Atherogenesis Antigen presenting cells (APCs) such as macrophages and dendritic cells have been identified within atherosclerotic lesions, and autoantibodies are present in the serum of individuals with atherosclerosis, and hence it has been proposed that an autoimmune reaction may be initiated within atherosclerotic plaques [62]. Wick and colleagues [63] have hypothesized that an immune response to HSPs, either endogenously derived from cells involved in atherogenesis, or exogenously, from microorganisms, may lead to complement-mediated endothelial injury and subsequent atherosclerosis. Several other potential autoantigens have now been identified including modified LDL (oxidized LDL and malondialdehyde modified LDL) and beta-2-Glycoprotein-I [64].

3. HSPs and Atherogenesis The potential relationship between serum HSPs, HSP antibody concentrations, and CVD was initially explored in the early 1990s [17, 65, 66]. Since then most interest has focused on HSPs-60 and-70 and there have been a number of cross-sectional and cohort studies investigating the relationship between antigen and antibody concentrations in coronary and PVD [67–69]. The expression of HSP-60 and-70 in atherosclerotic lesions was first reported by Kleindienst et al. [17] and Berberian et al. [65]. HSP60 expression was found to be highest in the shoulder regions and around the necrotic core of atherosclerotic plaques [70] (Fig. 5). Pockley and colleagues [66] demonstrated that HSP60 and HSP 60 antibodies were present in the circulation of normal individuals and later studies have shown a positive relationship between serum HSP-60 and atherosclerosis burden [71–73], particularly in the early stages of disease [74]. Expression of HSP70 was shown to be most concentrated in the center of thickened atheromatous plaques; the intensity of HSP70 staining was reported to correlate with the thickness of the atherosclerotic plaque. HSP70 appears to have an athero-protective role, as indicated by several cross-sectional studies [75, 76]; this may be mediated by its eVect on the survival of smooth muscle cells (SMCs). It was subsequently shown that the localization of HSP70 expression changed during plaque evolution and was positively associated with severity of atherosclerosis and the altered patterns of HSP70 staining [77]. In advanced atherosclerotic lesions, HSP70 was found to be expressed by several cell types including SMCs, dendritic cells, and monocyte/macrophages, while in early atherosclerotic lesions only dendritic cells expressed it [78].

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

CD68

A

35

Chlamydia HSP 60 + CD68

C

Control IgG

Human HSP 60 + CD68

B

D

FIG. 5. Co-localization of Chlamydial and human HSP 60 and macrophages in human atherosclerotic lesions. Reference: [70]. Publisher and year of copyright: American Heart Association, 1998. Permission for reproduction/adaptation was granted by the copyright holder.

An increased serum HSP70 concentration is reported to be associated with a lower risk of CVD [79], which was found to be independent of CVD risk factors, although it is known that some of these risk factors can induce HSP70 expression by ECs and SMCs [80]. It has also been reported that the severity of coronary disease (number of diseased vessels) is inversely related to serum HSP70 concentrations [79], although elevated levels of HSP70 have been found in patients with chronic heart failure [81]. Studies have also reported a cellular and humoral response to HSP65 in humans with carotid and coronary atherosclerosis [82, 83]. Cellular immunity directed against HSP60 was found to be related to intima: media thickness in young male individuals but not in the elderly, suggesting a possible role of specific cellular immunity to HSP60 in the early stages of atherosclerosis [84]. However, these results are not in accord with the Bruneck study [85, 86], which showed no relationship between circulating HSP60specific T cells and late stages of atherosclerosis. However, Ramage et al. [87] have reported that the proliferative response of human T lymphocytes to highly purified hHSP60 is confined to the adult CD45RA RO+ naı¨ve subset, whereas both memory and naı¨ve T cell populations proliferated to bacterial HSP60. Increased T cell responses to microbial HSP65 (mHSP65) as well as

GHAYOUR-MOBARHAN ET AL.

36

raised levels of circulating anti-mHSP65 and HSP60 antibodies have been found in patients with diVerent autoimmune conditions, and in patients with established atherosclerosis [88]. The possible mechanisms by which HSPs may be involved in atherosclerosis are summarized in Fig. 6. 3.1. HSPS AND ANIMAL MODELS OF ATHEROGENESIS The eVects of HSPs have predominantly been studied in LDL-receptordeficient and apolipoprotein E knockout mice and cholesterol-fed rabbits. The LDL-receptor deficient (LDL-RD) mouse develops significant atherosclerosis when fed a high fat diet [89], while apolipoprotein E knockout mice spontaneously develop hypercholesterolemia with concomitant atherosclerosis [90, 91], although they are also often fed an atherogenic, high fat diet. Arterial injury models have been used in mouse, rat, and rabbit and are associated with the rapid development of intimal lesions that are SMC rich. 3.1.1. Mouse and Rat A potential protective role of HSP70 is indicated by the ability of HP70 administration to limit infarct size following the exposure of the heart to ischemia–reperfusion injury in the rat [92] and rabbit [93]. In the LDL-R knockout mouse fed a normal diet, immunization with HSP65 or with heatkilled Mycobacterium tuberculosis develop atherosclerosis more rapidly than control animals [94]. Moreover, lesion formation was also enhanced in wild-type C57BL/6J mice similarly immunized with HSP65 or mycobacterial HSP65 [95], and in the rat model of arterial injury [96], neointimal thickening has been reported to be increased following immunization with mHSP65 [97]. Antibodies directed against, and lymphocytes reactive to HSP65 have been shown to promote fatty-streak formation in LDL-RD mice, providing further evidence for the pro-atherogenic potential of cellular and humeral immunity to HSP65 [98]. In a murine model that combines hyperglycemia with diet-induced hyperlipidemia, the accelerated atherosclerotic process has been reported to be associated with a significant immune response to HSP65 and elevated levels of anti-HSP65 [99]. Further evidence supporting the role of an autoimmune response to HSP in atherogenesis comes from experiments in which HSP60 autoreactive T lymphocytes were transferred to LDL-RD mice and which led to enhanced atherosclerotic changes [98]. 3.1.2. Rabbit Immunization of normocholesterolemic rabbits with HSP65 promotes atherosclerotic lesion formation [100], although these lesions regress in the absence of additional risk factors indicating that the inflammatory response

37

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS Free radicals like Ox-LDL

Smoking

High LDL and TG, low HDL

Infections

Other risk factors

Hypertension

Induction of stressful conditions for different

Survival

Death

Apoptosis

Increased expression of soluble HSPs

Release of cell-surface HSPs from apoptotic cells

Lysis

Release of intra-cellular HSPs following cell lysis

Formation of soluble HSPs

Structured alternation of HSPs and induction of immune responses

Immune response to self-HSPs caused by molecular mimicry with infectious HSPs

Activation of antigen presenting cells

Interaction of other antigens with human HSPs and formation of immunologic-complex

Activation of innate immunity

Auto-antibodies and auto-reactive cells

Recruitment of inflammatory cells and cytokines to vascular tissue, SMCs growth

Atherosclerosis FIG. 6. Possible mechanisms of involvement of HSPs in atherosclerosis. Modified from Reference [133]. Publisher and year of copyright: Elsevier, 2004. Permission for reproduction/ adaptation was granted by the copyright holder.

38

GHAYOUR-MOBARHAN ET AL.

on its own is insuYcient to drive atherogenesis over prolonged periods of time [101]. The initial phase of lesion formation in this latter model lacks foam cells and appears to be partially reversible, however in the presence of hypercholesterolemia the lesions develop further [101]. T cells isolated from these lesions were found to respond specifically to HSP 65 in vitro [32, 102]. Immunization with the Bacillus Calmette Guerin (BCG) vaccine, which contains large quantities of HSPs, has also been reported to enhance atherogenesis in the cholesterol-fed rabbit [103] and the mechanisms that may account for this are shown in Fig. 7. Plasma levels of anti-HSP70 increased in both BCG-immunized and control rabbits following the initiation of a high cholesterol-fed diet [104]. We [105] have also demonstrated that a high-cholesterol diet can induce the expression of anti-HSP60, 65 and 70 in rabbits, and this was associated with increasing concentrations of von Willbrand factor (vWF), a marker of endothelial injury (Fig. 8). It has been reported that depletion of peripheral blood T lymphocytes results in less atherosclerosis in rabbits immunized with HSP60 [32]. Xu et al. [102] have found that a population of T lymphocytes isolated from the atherosclerotic lesions of rabbits responded specifically to HSP65; IL-2 expanded T cell lines derived from atherosclerotic lesions, showed a significantly higher HSP-65 reactivity than those from the peripheral blood of the same animal. This finding supports the proposal that HSPs are an important autoantigen recognized in atherosclerotic lesions. Furthermore, T cell lines derived from the lesions of rabbits that were not immunized but only fed cholesterol rich diet, showed hyper-reactivity to HSP65 as compared to T cells from the peripheral blood of the same animals [102]. T cells derived from rabbit atherosclerotic lesions were also found to undergo a strong proliferative response to HSP65 in vitro [102]. Table 2 summarizes the animal studies that have investigated the relationship between HSPs and HSP antibodies and atherosclerosis.

3.2. MODULATION OF HSP EXPRESSION IN CELLS INVOLVED IN ATHEROGENESIS IN VITRO Several of the cell types involved in atherosclerosis express HSPs, although the factors stimulating their expression vary; for example, HSP60 overexpression by endothelial cells may be modulated by hemodynamic factors, whereas the expression by SMCs and mononuclear cells appears to be driven by the inflammatory process [17]. There is also evidence that HSP60 and HSP70 are expressed on all major cell types in lesion‐prone sites during atherogenesis [106].

39

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

Skin BCG T-cell

IL-4 IL-5

sponse

Th2 re

HSP-65

Anti-HSP-65 Ab

1

Th B-cell

o sp

re

Macrophage

e ns Lymph node

IFN-g Activated macriophage

HSP-60

HSP-60 MHC-restricted T cytotoxic cell

Hypercholesterolaemia

Complement HSP-60

Coronary artery plaque

FIG. 7. HSP-65 from BCG is taken up by tissue macrophages within the dermis which present HSP-65 derived peptides with class II MHC molecules to either type 1 (Th1) or type 2 (Th2) T helper cells or T cytotoxic cells. HSP-65 that drains to the lymph nodes is also endocytosed and processed by B-cells who express derivatized peptides in association with MHC class II molecules. Sensitized Th2 cells recognize this complex and secrete interleukin-4 (IL-4) and interleukin-5 (IL-5). These cause the B-cell to diVerentiate and express HSP-65 specific immunoglobulin. These antibodies cross-react with HSP-60 expressed by endothelium as a result of hypercholesterolemia or other stresses, allowing complement to bind and mediating the lysis of the cell. Activated Th1 cells secrete interleukin-2 (IL-2) and interferon-g (IFN-g) which activate macrophages and may increase the activation state of macrophages within plaques. Activated T cytotoxic cells restricted to MHC class I antigen recognition may recognize endothelial HSP-60 where expressed with MHC class I molecules and mediate cell endothelial cell death. Reference: [2]. Publisher and year of copyright: Elsevier, 2002. Permission for reproduction/ adaptation was granted by the copyright holder.

3.2.1. Endothelial Cells Endothelial cells are directly exposed to stressors and cardiovascular risk factors present in blood that can lead to endothelial injury. While the subsequent increase in expression of HSPs has potential protective eVects it may also have adverse eVects. HSPs are expressed on the cell surface [107] and in the presence of cross-reacting anti-mHSP65/-hHSP60 IgG or IgM

GHAYOUR-MOBARHAN ET AL.

40 A Anti-HSP-60 (absorbance)

2.5 *

2

*

*

1.5 1 Cholesterol-fed Chow-fed

0.5 0 0

2

6

4

8

10

12

14

12

14

12

14

Weeks

B Anti-HSP-65 (absorbance)

2.5

*

* *

2 1.5 1

Cholesterol-fed Chow-fed

0.5 0 0

2

6

4

8

10

Weeks

C Anti-HSP-70 (absorbance)

2.5

*

* *

2 1.5 1 Cholesterol-fed Chow-fed

0.5 0 0

2

4

6

8

10

Weeks

FIG. 8. Time course for changes in plasma anti-Hsp-60,-65, and-70 titers in normal chow and cholesterol-fed rabbits. (A) Antibody titers to Hsp 60 were significant higher for cholesterol compared to normal chow-fed animals during the experimental period ( p = 0.0013, by ANOVA), also being significantly higher at weeks 5 ( p < 0.05), 7, and 9 ( p < 0.01) compared with baseline. (B) Antibody titers to Hsp 65 were significant higher for cholesterol compared to normal chowfed animals during the experimental period ( p = 0.001, by ANOVA), also being significantly higher at weeks 5 ( p < 0.05), 7, and 9 ( p < 0.01) compared with baseline. (C) Antibody titers to Hsp 70 were significantly higher for cholesterol compared with normal chow-fed animals during the experimental period ( p = 0.0016, by ANOVA), also being significantly higher at weeks 5 ( p < 0.05), 7, and 9 ( p < 0.01) compared with baseline. Reference: [105]. Publisher and year of copyright: Blackwell Publishing, 2007. Permission for reproduction/adaptation was granted by the copyright holder.

TABLE 2 ANIMAL STUDIES INVESTIGATING THE RELATIONSHIP BETWEEN HSPS AND HSP ANTIBODIES AND ATHEROSCLEROSIS Animal model Mouse

Rat

Aim of the study Investigation about myocardial protection and changes in gene expression following by whole body heat stress Investigation whether the expressions of HSP60 and HSP70 are correlated with the development of atherosclerotic lesions To examine the individual contribution of specific HSPs in primary rat cardiomyocytes to any protection observed following lethal heat stress or simulated lethal ischemia To test the hypothesis that the degree of protection from ischemic injury in heat-shocked rats correlates with the degree of prior HSP72 induction To examine whether the overexpression of HSP27 and alphaB-crystallin in rat cardiomyocytes would protect against ischemic injury To test whether phosphorylation of HSP27 is required for the protective role this protein plays in the cell Investigation about the possible autoantigens involved in atherosclerosis Investigation a possible relationship between HSP60 expression and the antigenic specificities of infiltrating T cells in the lesion Investigation about the immune mechanisms in atherosclerosis

Outcome

References

Increased myocardial HSP70 expression results in protection of the heart against ischemic injury

[92]

HSP60 and HSP70 are temporally expressed on all major cell types in lesion‐prone sites during atherogenesis Transfection of the inducible heat stress protein 70 was found to increase survival following a lethal heat stress and against lethal ischemia

[106]

The improved salvage after heat-shock pretreatment may be related to the amount of HSP72 induced before prolonged ischemia and reperfusion The increased expression of HSP27 and alphaBcrystallin protects against ischemic injury in adult cardiomyocytes Phosphorylation of HSP27 seems not to play a role in its ability to protect adult rat cardiomyocytes against ischemic damage Induction of arteriosclerosis in normocholesterolemic rabbits by immunization with HSP65 Increased expression of HSP65 coincides with a population of infiltrating T lymphocytes in atherosclerotic lesions of rabbits specifically responding to HSP 65 Regression of arteriosclerotic lesions induced by immunization with HSP65 containing material in normocholesterolemic, but not hypercholesterolemic rabbits

[114]

[118]

[124]

[125]

[100] [102]

[101]

(continues)

TABLE 2 (Continued) Animal model Rabbit

Aim of the study Investigation about the immune mechanisms in atherosclerosis Investigation about the immune mechanisms in atherosclerosis Investigation about the relationship between the immune responses to HSP and subsequent atherosclerosis Investigation the time course of appearance of Hsp-60,65, and-70 antibodies in the cholesterol-fed rabbit and to relate antibody titers to serum concentrations of von Willbrand factor Induction of stress proteins, such as heat-shock protein 71 (HSP71), is associated with cardioprotection in isolated ischemic myocardium

HSP, heat shock protein; CVD, cardiovascular disease.

Outcome

References

Inhibition of arteriosclerosis by T cell depletion in normocholesterolemic rabbits immunized with HSP65 Immunization with BCG vaccine increases aortic atherosclerosis in the cholesterol-fed rabbit Immune responses to HSP may be implicated in the relationship between specific infections and CVD

[32]

In cholesterol-fed rabbits, antibody titers to Hsp-60,-65, and-70 appear to rise in association with a marker of endothelial injury, peaking at approximately the same time after starting a high cholesterol diet Heat shock-induced cardioprotection is transient and delays the onset of irreversible myocardial injury caused by ischemia

[105]

[103] [104]

[113]

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

43

antibodies may be susceptible to complement-mediated cell lysis [108], or endothelial cell apoptosis [109]. 3.2.2. Smooth Muscle Cells SMCs are important players in atherogenesis and can be induced to express HSPs as part of a survival mechanism following exposure to a variety of stressors; for example, exposure to high blood pressure. Berberian et al. [65] have reported increased levels of HSP70 expression in human atherosclerotic lesions and showed a protective role for HSP70 in the survival of SMCs, which has been confirmed by others [110]. However, the reported expression of HSPs by SMCs in complex lesions appears to be inconsistent [77]. Furthermore, mechanical stresses evoke rapid activation of HSP70 expression in SMCs [111] and Kleindienst et al. [17] showed the presence of HSP60 on SMCs in aortic and carotid specimens. 3.2.3. Cardiac Myocytes HSPs have an important role in protecting myocardial cells from a number of environmental stressors. This has been investigated using in vitro and in vivo approaches in animal models [112, 113]. For example, it has been shown that overexpression of HSP70 in cultured primary cardiac cells protect these cells against ischemic or thermal stress, while overexpression of HSP60 did not have a protective eVect [114–116]. To show the cardioprotective role of HSP70, transgenic mice overexpressing HSP70 were generated and these animals were found to be more resistant to ischemic injury [92, 117]. In rats treated with whole body hyperthermia there was an induction of HSP72 and a reduction of myocardial infarction (MI) size following experimental ischemia [118]. Experimental coronary artery occlusion induces myocardial ischemia and elevation of HSP70 in heart tissues [119], and specifically in myocardial cells [120, 121]. Moreover, in other studies HSP70 has been shown to increase arterial and myocardial cell survival [92, 117, 122] and it has been proposed that HSP70 is associated with protective mechanisms in normal and diseased arteries [123]. Other HSPs also appear to protect cardiac myocytes from ischemic injury. Martin et al. [124, 125] reported that both HSP27 and HSP70 were able to protect cardiac myocytes from the eVect of ischemia and that decreasing the level of endogenous HSP27 resulted in an enhancement of the damaging eVects of a subsequent ischemic stimulus. These findings suggest that HSP27 may also be protective in myocardial cells. The authors propose that plasma HSP27 concentrations could be a potential marker of atherosclerosis, although further validation in larger patient cohorts is required. It has also been suggested that increased expression of HSP27, could be important for cardiac self-protection in cardiac allograft rejection, [126].

44

GHAYOUR-MOBARHAN ET AL.

3.2.4. Monocyte/Macrophages Monocyte/macrophages are APCs which are able to process and present HSP related peptides to lymphocytes and may be involved in the generation of an autoimmune response associated with atherosclerosis. Lymphocyte activation (particularly Th1 cell stimulation) may enhance the inflammatory process [34, 35]. Immunocytochemical studies of both early and late atherosclerotic lesions have revealed a high level of expression of both HSP60 [17] and HSP70 [65] by macrophages, particularly those adjacent to the necrotic core of advanced lesions. Furthermore, in vitro studies have shown that HSPs and anti-HSP antibodies induce the production of pro-inflammatory cytokines by macrophages that may stimulate adhesion molecule expression and thereby further enhance the inflammatory process [71, 127]. 3.2.5. Lymphocytes It has been established that atherosclerotic lesions contain large numbers of T lymphocytes [30], and that several types of T cells are involved in modulating the inflammatory response that include helper, suppressor, and regulator cells. It is also hypothesized [102] that HSP65 is an autoantigen which is recognized by these cells, and T cells, isolated from atheromatous plaque appear to be stimulated by HSP65 in vitro. 3.2.6. HSPs and Apoptosis HSP27 has been shown to bind to cytochrome C and prevent its interaction with Apaf-1 [128, 129] causing an inhibition of cell apoptosis, HSP70 can also inhibit apoptosis in a caspase independent manner [5] and further that overexpression of HSPs in cardiac myocytes has been shown to inhibit apoptosis [130]. HSP27 induces human monocytes to produce large amounts of the anti-inflammatory cytokines [131] and inhibits toll like receptor-4 (TLR-4) expression on monocytes and their diVerentiation into dendritic cells [132].

3.3. SOLUBLE OR CIRCULATING HSPS There are a number of possible sources of soluble HSPs in blood, as previously discussed [133]: (1) increased synthesis of HSP by host cells as an immune defense in infectious organisms within the host [134]; (2) release of intracellular HSPs following cell lysis [135]; (3) increased expression of soluble HSPs because of general inflammatory processes, for example, during atherogenesis; (4) release of soluble HSP from necrotic cells within the plaque; (5) and finally the release of cell-surface HSPs from apoptotic cells via the formation of microparticles [136–139]. It is reported that the TLR

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

45

4/CD14 complex is a receptor for soluble HSP [140], and hence cell surface bound HSP may also originate from exogenous sources. Some soluble HSPs may have pro-inflammatory eVects; for example, Chlamydial and human HSP60 can induce the expression of tumor necrosis factor (TNF)-a and matrix metalloproteinase (MMP)-9 [70], IL-6 [141], IL-12, IL-15 [127]; and exogenous HSP-70 has been reported to upregulate IL-1, IL-6, TNF-a expression in human monocytes [142]. Binding of soluble HSPs to the TLR 4/CD14, stimulates an innate immune response that includes the production of pro-inflammatory cytokines by macrophages and adhesion molecules in endothelial cells via NF-B activation [140]. These data suggest that soluble HSPs may serve as a danger signal for the innate immune system [143, 144].

4. HSPs and Autoimmunity in Atherogenesis 4.1. GENERAL CONSIDERATION The presence of professional APCs such as macrophages and dendritic cells within atherosclerotic plaques provides an opportunity for some plaquerelated antigens to become autoantigens. Wick et al. [145] have proposed that the accumulation of mononuclear cells within the arterial intima, could be termed the vascular-associated lymphoid tissue (VALT), and may be viewed as being analogous to the mucous associated lymphoid tissue (MALT). They have proposed that the VALT has a similar role as the MALT, which may include monitoring for potentially harmful autologous and exogenous antigenic material contained in the blood. In an attempt to identify which antigens may be implicated in early atherogenesis, Xu et al. [100] immunized normocholesterolemic rabbits with a variety of mixed antigens in complete Freund’s adjuvant. Surprisingly, all immunized animals were found to develop atherosclerotic lesions at the known predilection sites [100]. The authors proposed that this was due to the mycobacteria-derived HSP it contained. They subsequently immunized rabbits with recombinant purified mycobacterial HSP65. Animals initially developed vascular lesions that did not progress unless they were also fed a cholesterol rich diet [146]. Wick et al. [63] hypothesized that autoimmune responses to HSPs could be crucial in the initiation of atherosclerosis [95, 100] and this is supported by human studies [19, 49, 67, 82, 147]. While immune responses have a potentially important role in atherogenesis, immunosuppressed rabbits [32] or immunologically compromised mice [148–151] do nevertheless develop atherosclerosis. It is likely that both humeral and cellular responses to HSP65 are implicated, with a predominant role of Th1 cells [36, 98, 152], and it is

46

GHAYOUR-MOBARHAN ET AL.

hypothesized that T cells are involved in the initiation of disease and the humoral response plays a facilitating role [153]. It is possible that HSP27 and-90 are putative autoantigens involved during atherogenesis [154, 155]. Furthermore, autoimmunity to HSPs may lead to a systemic inflammatory response associated with elevated CRP which may also promote atherogenesis. However, autoantibodies to HSPs can be found in normal subjects. Perschinka et al. [156] found that antibodies to mHSP60/65 recognize epitopes on human HSP60; these cross-reactive epitopes were shown to serve as autoantigenic targets in incipient atherosclerosis and also HSP60 could be targeted by a proportion of anti-EC antibodies including anti-HSP antibodies. These are able to trigger apoptosis of ECs [109], which is dependent on HSP60 epitope specificity. It is reported that high antibody titers against mHSP65 are associated with increased cardiovascular morbidity and mortality [67]; there were similar findings for anti-HSP60 [157], -70 [158, 159], and -27 [160] antibodies. One possible mechanism accounting for this is via the induction of pro-inflammatory cytokines by macrophages [71, 127, 133], leading to plaque instability. However, one study has reported no significant relationship between inflammatory factors and anti-HSPs antibody titers [161]. The antiHSP antibodies could also lead to endothelial injury by antibody-dependent, complement mediated cellular cytotoxicity. We [105] have demonstrated that there is a relationship between vWF concentrations, a marker of endothelial injury or dysfunction [69], extent of atherosclerosis and antibody titers to HSP60, -65, -70 in the cholesterol-fed rabbit. Titers of these anti-HSP antibodies may be induced and maintained by several mechanisms [133]; (1) an immune response to HSP60 derived from microorganisms, but homologous to human HSP65, because of the phenomenon of molecular mimicry [162]; (2) HSP may be rendered immunogenic because of structural alteration or posttranslational modification resulting from oxidation or metabolic alteration [163]; (3) Other foreign or self-antigens could interact with HSP60 to form immunogenic complexes, and thereafter be recognized as foreign by B or T cells [164]; (4) the recognition of soluble HSPs [71] by a population of T and B cells as a non-self-antigen; and (5) genetic susceptibility (supported by the strong association between a genetic polymorphism within the IL-6 promoter and anti-HSP60 antibody levels) [165]. Binding of anti-HSP antibodies to epitopes on stressed endothelial cells is followed by complement activation [147, 166] and endothelial injury. Hence, endothelial cells that are exposed to high temperature or inflammatory cytokines (e.g., TNF-a) are particularly susceptible to complementdependent lysis by HSP60-specific antibody [167] in the presence of high concentrations of these antibodies [166]. There are marked diVerences in the ability to activate complement for diVerent anti-HSP65 and HSP60

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

47

antibodies present in patients with coronary heart disease (CHD), and this may be an issue of epitope specificity. Only HSP60–antiHSP60 immune complexes have been shown to be strong activators of complement [166], and a strong positive correlation has been reported between the degree of complement activation and the concentrations of anti-HSP60 but not antimHSP65 IgG antibodies [166]. Furthermore, it has been reported that high levels of complement-activating autoantibodies against HSP60 may be an independent cardiovascular risk factor of CHD [168]. 4.2. MOLECULAR MIMICRY AND RELATION TO INFECTION While anti-HSP antibodies are probably produced for the primary purpose of eliminating infectious organisms they may lead to endothelial injury; Mayr et al. [147] have reported that serum anti-HSP-antibodies to E. coli and C. pneumonia can mediate endothelial cell lysis of stressed, but not unstressed endothelial cells [108, 147]. Moreover, serum antibodies to HSP60/65 from subjects with atherosclerosis, appear to cross-react with hHSP60, GroEL, and Chlamydial HSP60 [147]. 4.3. ANTIBODIES TO HSPS AND INFECTIONS Mayr et al. [49] have shown that antibody titers to Mycobacterial HSP65 correlated strongly with human IgA to C. pneumonia and with IgG to H. pylori, suggesting a role for infection in inducing the production of mHSP65 antibodies. Eradication of H. pylori in patients with confirmed H. pylori infection led to a significant fall in anti-mHSP65 titers, suggesting that H. pylori infection may be a determinant of anti-mHSP65 titers [169]. High levels of anti-HSP60 and C. pneumonia antibodies were found to be independent risk factors for coronary atherosclerosis [170] and their concurrent presence substantially increased the risk of CVD [171]. It was also shown that serum levels of antihuman (h)HSP60 IgG antibody and anti-Chlamydial IgM, but not IgG or IgA antibody were significantly higher in patients with acute coronary syndrome than in patients with stable ischemic heart disease [172]. A persistent elevation in antibodies to both hHSP60 and C. pneumonia was a better predictor of coronary events than transient or individual elevations in these antibodies [173]. Mayr et al. [49] also found that IgA antibodies to C. pneumonia were correlated with the extent of carotid and femoral atherosclerosis and were associated with antibodies to mHSP65. However, these findings are not consistent with those of Hoymons et al. [174] who have reported that the antibody response to human and Chlamydial HSP60 are not associated with endothelial dysfunction, nor the presence or severity of CVD, arguing against the proposition that infection contributes to

GHAYOUR-MOBARHAN ET AL.

48

disease progression. Furthermore, Jantos et al. [175] were unable to demonstrate the value of HSP60 antibody titers to C. pneumonia in discriminating between patients with and without CVD. Deshpande et al. [176] have reported that a primary periodontopathic pathogen can invade ECs and there is evidence that antibodies to P. gingivalis GroEL in sera can cross-react with human HSP60 [177], which indicates that these antibodies could mediate endothelial cytotoxicity. It has been reported that in patients with severe periodontitis, elevated IgG antibody titers to both P. gingivalis HSP60 and human HSP60 are observed [178]. Increased levels of salivary anti-HSP65 IgA antibodies have also been reported in patients with gingivitis [179], although this has not been a consistent finding [158]; furthermore, anti-HSP-60/-65 IgA titers were found to be lower in smokers and this may be related to an impaired ability to mount a humeral responses to HSP60/65. Other possible confounding factors include the stage of disease. 4.4. ANTIBODIES TO HSPS AND CARDIOVASCULAR RISK FACTORS The association of antibodies directed against HSPs has been reviewed below for animal models and human studies, respectively. 4.4.1. Animal Models Cardiovascular risk factors may be divided into those that are modifiable and those that are nonmodifiable. It has been reported that the response to heat treatment was attenuated with increasing age in an animal study [180]. Although acute changes in blood pressure can lead to an upregulation of HSP70 in the rat aorta [181], similar changes were not seen in spontaneously hypertensive rats, even though comparable basal blood pressures were suYcient to induce HSPs levels in normotensive Wistar–Kyoto rats [181]. Nonobese diabetic mice have been reported to develop high titers of anti-HSP60 [182]. Similarly, LDL-RD mice with an induced hyperglycemia, also developed higher antibody titers to HSP65, and accelerated atherosclerosis [99]. A high cholesterol diet in the rat may lead to a significant attenuation of the protective eVects of ischemic preconditioning of their heart [183], which may be explained by the finding that hyperlipidemia can inhibit the heat shock response [184]. 4.4.2. Human Studies No significant relationship has been found between gender, positive family history of CVD and age and concentrations of soluble HSPs or anti-HSPantibody titers [67, 71, 83, 160, 185]; and in a prospective study Xu et al. [83] were unable to find an association between any traditional risk factors and

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

49

HSP65 antibody titers. Furthermore, Frostega˚rd et al. [186] found no significant correlation between serum anti-HSP 65 concentrations and several metabolic and anthropometric variables (i.e., lipoproteins, insulin, body mass index, and waist–hip ratio) and blood pressure. However, reports have been inconsistent; for example, Rea et al. [187] found that in healthy individuals aged from 20 to 96 years, there was a progressive decline in serum HSP60 and HSP70 antigen levels and a trend for an increase in serum HSP70 antibody levels with age. Similar results were found for IgG anti-HSP-27 concentrations which were strongly associated with age, gender, hypertension, and weakly with diabetes in patients with acute coronary syndrome [188]; however, other cardiovascular risk factors were not associated with anti-Hsp-27 IgG antibody concentrations. Furthermore, it was reported that anti-HSP27 antibody titers are inversely related to age but unrelated to several other established cardiovascular risk factors [189]. We were unable to demonstrate an association between anti-HSP27 antibody levels and several coronary risk factors in an Iranian cohort [160]. Blood vessels subjected to increased mechanical and shear stress express HSPs and are also more prone to the development of atherosclerosis [17, 108]. Frostega˚rd et al. [186] have demonstrated that serum anti-HSP antibodies correlate positively with hypertension, supporting the eVects of altered hemodynamic stress on HSP. Elevated levels of plasma HSP60 have also been reported in patients with borderline hypertension and were associated with increased intima-media thickness [74]. While other studies have reported that anti-hHSP60 titers were significantly lower in individuals with borderline hypertension [190], and that although anti-HSP65 titers were associated with diastolic blood pressure, they were not related to systolic blood pressure. Other investigators have reported a trend for circulating HSP60 antigen and anti-HSP65 levels to be higher and anti-HSP60 levels to be lower in patients with borderline hypertension [74]. Pockley et al. [75] found that in patients with established hypertension whose intima-media thickness was measured over a 4‐year period the smallest changes were found in subjects with high HSP70 levels, and anti-HSP70 and antimHsp65 antibodies were significantly and independently elevated in patients with established hypertension compared to normotensive controls [190, 191]. The latter result is in contrast to the findings in patients with borderline hypertension, in whom anti-HSP70 concentrations were not elevated relative to controls, and elevation in anti-mHSP65 levels was consistent between patients with borderline and established hypertension [74, 186] so the reported relationship between HSP antigen and antibody levels and blood pressure has been inconsistent in clinical studies. HSP70 antibodies appear to have a protective eVect in hypertensive subjects by modifying the progression of atherosclerosis [75], and increased levels of

50

GHAYOUR-MOBARHAN ET AL.

circulating HSP70 have been related to a lower risk of CVD and decreased intima-media thickness in hypertensive patients [75, 79]. Moreover, it has been found that peripheral blood lymphocytes from hypertensive subjects contain more HSP70 mRNA compared with normotensive individuals [192], and another study has reported that anti-HSP65 and 70 antibody levels were both associated with hypertension, independent of age, smoking habit, and blood lipids [190]. With respect to HSP27, Shams et al. [188] found an inverse relationship between hypertension and Hsp-27 IgG antibody concentrations in patients with chest pain. We have previously reported similar results for HSP-65 antibodies [193]. The combination of hypertension and presence of high anti-HSP60 titers was associated with a >4-fold higher risk of CVD compared to normotensive subjects with low concentrations anti-HSP60 [194]. A similar additive eVect on CVD risk was observed with the combination of diabetes and high concentrations of anti-HSP60 [194]. Smoking induces a necrotic and, hence, pro-inflammatory type of cell death in endothelial cells, that may lead to the release HSP60 [195–197]. Individuals who had never smoked or who were not current smokers were found to have higher serum HSP60 concentrations than individuals who were smokers [185]. Frostega˚rd et al. [186] found that smokers with atherosclerotic lesions and borderline hypertension had significantly decreased antibody titers to HSP65 compared with age-matched normotensive smokers, but not nonsmokers. Although smoking may be expected to cause an induction of HSP expression, it may also lead to increased cell necrosis with the clearance of HSP65 via the formation of immune complexes with HSP65 antibodies which would lead to a subsequent decrease of antibody titers [186]. It has also been proposed that high concentrations of plasma HSP60 may lead to a suppression of the anti-HSP65 immune response [198]. Another possible reason for low HSP60 titers is that smoking stimulates HSP60 clearance from the plasma by increased catabolism or cellular uptake [185]. Kervinen et al. [199] have shown that while a high anti-HSP60-antibody level in hypertensive patients increased coronary risk by approximately 50%, smoking more than doubled the risk, indicating the important role of smoking in the promotion of atherosclerosis. In diabetic patients anti-HSP60 plasma concentrations were reported to be higher than for nondiabetics [194]. IgA anti-HSP70 antibody concentrations were also shown to be significantly higher in type I and II diabetics than in nondiabetics [200]. Similarly, anti-HSP70 and 90 have also been reported to be higher in diabetic patients [201], and a significantly higher proportion of diabetic patients with CVD had measurable levels of plasma HSP60 compared with those with no evidence of CVD [185]. Anti-HSP60 antibody

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

51

concentrations were found to be independently associated with CVD risk, and the combination of diabetes and high concentrations of anti-HSP60 was associated with a substantially increased CVD risk [194]. A positive association between soluble HSP60 and total serum cholesterol [202] and LDL cholesterol [71] has been reported. Plasma HSP27 concentrations have also been reported to be correlated with total serum cholesterol concentrations in patients with acute coronary syndrome [203]. In contrast, a negative association was reported between anti-HSP60 antibody titers and serum HDL cholesterol [199]. Among dyslipidemic patients, serum high sensitivity (hs-) CRP concentrations and anti-HSP60, 65, 70 titers were significantly higher than for controls, but there was no significant association between HSP antibody titers and serum hs-CRP concentrations [204]. Huittinen et al. [205] were unable to find a significant relationship between plasma HSP60 antibody titers and coronary risk amongst dyslipidemic middle-aged males. Moreover, patients with hyperlipidemia and hypertension had lower levels of anti-HSP27 antibody than those with neither [188]. The presence of both dyslipidemia and high anti-HSP60 antibody titers was associated with a high risk of coronary atherothrombotic events [199]. We have recently reported a significant relationship between HSP-60, -65 and -70 antibody titers with specific dietary constituent [105, 204]; in subjects with dyslipidemia, plasma antibody titers to HSP60, -65, -70 were associated with dietary antioxidant vitamins and saturated fat [204]. We and others have also found a significant relationship between antibody titers to HSP60 versus HSP65, HSP60 versus HSP70, and HSP65 versus HSP70 [206, 207], although these findings do not accord with those of Kocsis et al. [208]. Statins are used to lower LDL cholesterol concentrations and have been shown to reduce the risk of CVD. They also appear to be associated with a reduction in anti-HSP-antibody titers [206, 207]. In addition, cardiac rehabilitation therapy was also found to be associated with a significant reduction in the antibody titers to HSP60 and 70 [207]. Statins are also inhibitors of MHC class II mediated T cell activation [209], and it is therefore possible that some beneficial eVects of compounds such as these may be due to their immunemodulatory eVects rather than their action on cholesterol metabolism. Psychological factors, such as stress are known to contribute to CVD risk [210]. Lewthwaite et al. [211] have reported an inverse association between serum HSP60 concentrations, social isolation, low socioeconomic status, and psychological distress, a finding that was confirmed in a larger cohort [202]. Table 3 provides a summary of the studies that have investigated the relationship between HSPs and anti-HSP antibodies with known coronary risk factors.

TABLE 3 STUDIES INVESTIGATING THE CORRELATION BETWEEN HSPS AND ANTI-HSP ANTIBODIES WITH CORONARY RISK FACTORS Risk factor

Subjects

Nonmodifiable and modifiable risk factors

750 subjects from general population

Nonmodifiable and modifiable risk factors Nonmodifiable and modifiable risk factors

826 subjects from general population

Age and Hypertension

60 patients with acute cardiac events

Age

255 initially healthy participants from a cohort study 72 men with borderline hypertension

Hypertension

94 patients with CVD

Outcome Negative association with anti-HSP65 antibody titers between these risk factors in these subjects ( p > 0.05) Negative association with soluble HSP60 between these risk factors in these subjects ( p > 0.05) Negative association with anti-HSP27 antibody titers between these risk factors in these subjects ( p > 0.05) Positive association between anti-HSP27 antibody titers and age and hypertension( p < 0.001) Inverse association between anti-HSP27 antibody titers and age ( p < 0.001) Positive association between circulating HSP60 ( p = 0.001) and anti-HSP65 antibody levels ( p < 0.001), Negative association with HSP70 and anti-HSP70 antibody levels ( p > 0.05)

References [67]

[71] [160]

[188] [189] [74]

Hypertension

66 men with borderline hypertension

Hypertension

111 men with established hypertension

Smoking Diabetes mellitus

855 patients (17.2% type I and 82.8% type II) 67 patients (27 type I and 40 type II)

Total cholesterol

27 patients with acute coronary syndrome

HDL-cholesterol

233 middle-aged men from a cohort study

High levels of hs-CRP

238 dyslipidemic patient

Psychological factors

126 men and 103 women

Psychological factors

541 men and 319 women

Positive association between anti-HSP65 antibody titers and hypertension ( p < 0.05) Positive association between anti-HSP65 and 70 antibody titers and hypertension ( p < 0.001) Positive association between HSP60 concentrations in nonsmoker group ( p = 0.01) Positive association between IgA antibody to HSP70 and diabetes mellitus type II ( p < 0.05) Positive association between HSP27 titers and total cholesterol ( p < 0.05) Negative association between anti-HSP60 and HDL ( p > 0.05) Negative association between anti-HSP60, 65, 70, and hs-CRP ( p > 0.05) Inverse association between soluble HSP60 levels and Psychological factors in women ( p < 0.05) Positive association between soluble HSP60 levels and Psychological factors in women and men ( p < 0.05)

HSP, heat shock protein; CVD, cardiovascular disease; Ig, Immunoglobulin; hs-CRP, high-sensitive C-reactive protein.

[186] [190] [185] [200] [203] [199] [204] [211] [202]

54

GHAYOUR-MOBARHAN ET AL.

4.5. ANTIBODY TITERS TO HSPS AND THEIR RELATIONSHIP TO CVD BURDEN There have been a number of observational and prospective studies that have shown associations between antibody titers to several HSPs and atherosclerosis. 4.5.1. Observational Studies Anti-HSP60 antibodies were found to be present at particularly high levels in subjects with unstable angina or following myocardial infarction (MI) [161]. Elevated levels of serum anti-HSP60 titers have also been reported in patients with ECG abnormalities, including sinus arrhythmia, chronic myocardial ischemia, and ectopic rhythm [212]. Moreover, high antibody titers to mHSP65 but not hHSP60 were found to be associated with coronary calcification [213]. There are reports that some HSP antibody titers fall after MI or angioplasty [82, 213, 214], and this may be explained by the formation of immune complexes between the circulating antibodies with HSPs released as a consequence of tissue necrosis; these being rapidly cleared by the liver [198]. Acute cardiovascular events may therefore be associated with acute changes in anti-HSP60 antibody titers, as has been reported [82]. Serum mHSP65 antibodies have been shown to cross-react with recombinant human HSP60, homogenates from atherosclerotic plaque and HSP60 present in the endothelial cells within atheromatous lesions [215]. Several studies have also reported that anti-HSP60/65 is cross-reactive [88, 182], and this has been confirmed by Zhu [157] and Xu [67] in subjects with carotid atherosclerosis. However, it has been suggested that anti-HSP60 and antiHSP65 antibodies from CHD patients are only partially cross-reactive [166, 190], and it appears that the recognition and production of antibodies to diVerent HSP60 epitopes expressed on ECs can result in diverse consequences. For instance, the anti-hHSP60 monoclonal antibody II-13 was cytotoxic for stressed ECs, while another monoclonal antibody, ML-30, which recognizes a diVerent epitope, was not [167]. So it appears that distinct epitopes are accessible to diVerent antibodies, indicating that surface orientation of HSP60 is important, or that discrete domains of hHSP60 are present on the outer surface of the cells. There have been inconsistent reports of the importance of HSP70 antibody titers in CVD [159, 208]. Zhu et al. [79] found no association between antiHSP70 IgG sero-positivity and the prevalence or severity of CVD whereas there have been reports in which patients with coronary atherosclerosis or stable/unstable angina were found to have lower levels of anti-HSP70 antibody [216]. Herz et al. [216] found higher titers of anti-HSP70 in patients with unstable angina compared to those with stable chronic angina, and Vogt [217] reported that higher titers of anti-HSP70 were associated with cardiac

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

55

output and pulmonary capillary wedge pressure in those patients who were positive for the anti-HSP70 antibody undergoing heart surgery. It was proposed that patients whose preoperative stress levels, reached the threshold for anti-HSP70 antibody production, were protected from the subsequent CABG procedure. Kramer et al. [218] found that there was no diVerence in serum titers of either anti-HSP60 or anti-HSP65 antibodies between patients with cerebrovascular disease and age-matched healthy subjects. The authors proposed that stimuli that enhance HSP expression in coronary arteries may not have a similar eVect in carotid arteries, which may be more resistant to pro-atherogenic factors. A significant correlation between anti-HSP70 antibody levels and vascular disease severity has been reported in patients with lower limb claudication, or lower limb critical ischemia [159]. In another study, although levels of serum HSP70 were significantly elevated in 20 patients with PVD, as were serum anti-HSP60 and anti-HSP70 titers, HSP60 were not significantly related to the extent of disease [68]. 4.5.2. Prospective Studies Several studies have shown that human anti-HSP60 antibodies are positively associated with the development of atherosclerosis [83, 157, 205] and higher titers of anti-HSP antibodies were strongly associated with CVD [82, 169], particularly anti-HSP60 titers [157]. In one of the first prospective cohort studies of HSP antibody titers and vascular disease, Xu et al. [83] found that elevated levels of anti-mHSP65 antibodies were an independent prognostic marker of the incidence, severity, progression, and mortality associated with carotid atherosclerosis in a population that was initially clinically healthy [67, 83, 169, 186]. Serum soluble HSP60 concentrations and antibody titers to HSP65 were also found to predict carotid disease [19] and mortality [67], respectively. Antibody titers to HSP60 have been reported to be associated with both the presence and severity of clinically significant CVD, independent of traditional coronary risk factors [157, 171], indicating that a high anti-hHSP60 titers may be an important risk factor for coronary atherosclerosis [208, 219]. However, it appears that subtype specificity may be important; for example, it has been reported that anti-HSP60 IgA titers, but not IgG or C. pneumonia HSP60 antibodies were a significant risk factor for coronary events. An association between the hHSP60 IgA antibody titers and serum CRP concentration has also been reported [199, 205]. However, elevated concentrations of anti-HSP60 IgA antibody were not found to be a risk factor for CVD unless CRP levels were also elevated, when the presence of an elevated IgA antibodies against hHSP60 was predictive of coronary death

56

GHAYOUR-MOBARHAN ET AL.

and MI. However, antibodies titers against human and Chlamydial HSP60 do not appear to be consistent markers for coronary atherosclerosis or arterial dysfunction [174]. Some studies have shown that anti-mHSP65 antibody levels are predictive of cardiovascular events [67, 220], however, these findings are in contrast to the report of Pockley et al. [75] who found no relationship to intima-media thickness, and no association between HSPs or anti-HSP antibodies and intima-media thickness in subjects with established hypertension [190]. The explanation for these discordant findings may be patients selection, or that high HSP antibody titers are a marker of plaque instability and the risk of an acute event, rather than stable plaque formation. Anti-HSP65 titers have also been reported to be a predictive marker of outcome following coronary angioplasty; patients in whom a fall in antibody levels immediately after PTCA was observed, did not develop restenosis, while in patients who developed subsequent restenosis this decrease was not observed [213, 214]. Anti-HSP65 and anti-HSP70 titers were found to be elevated 48 h after an ischemic stroke, and elevated levels of these antibodies were found to be independent risk factors for stroke [221]. HSP70 antigen is likely to be neuro-protective during the early phases of ischemic stroke; lymphocyte-associated HSP70 is elevated in patients with cerebral infarction, and its level decreased during the period of recovery [222]. It has also been reported that detectable IgG titers against HSP60/65 is associated with an increased risk of stroke [223], although anti-mHSP65 titers appear to have a poor predictive value for atherosclerosis [169, 224]. Hence, antibody titers to HSP60, ‐65, and ‐70 have been reported to be associated with increased risk of CVD [141], the severity of cardiovascular [164], and vascular endpoints in patients with established disease [167, 225] while there are relatively few consistent data for other HSPs, including small HSPs [193]. Table 4 summarizes the clinical studies investigating the relationship between HSPs and HSP antibodies and atherosclerosis. 4.6. CHANGES IN TITERS OF HSP ANTIBODIES DURING ACUTE CORONARY SYNDROMES There have been a few studies that report changes in antibody titers to HSPs in patients with acute coronary syndrome [67, 169, 226]. There have been several reports of reductions in antibody titers after MI or angioplasty [82, 214]; the latter may be due to the formation of immune complexes with HSPs released as a consequence of tissue necrosis [198]. Furthermore, it has been reported that there is a significant increase in serum HSP27 antigen levels in patients with acute coronary syndrome [203]. Shams et al. [188]

TABLE 4 CLINICAL STUDIES INVESTIGATING THE RELATIONSHIP BETWEEN HSPS AND HSP ANTIBODY TITERS AND ATHEROSCLEROSIS Design

Observational, Cross‐sectional

Subjects N ¼ 219 CVD patients N ¼ 396 autoworkers exposed to noise N ¼ 391 CVD patients N ¼ 99 CVD patients N ¼ 61 vascular patients N ¼ 421 CVD patients N ¼ 131 CVD patients

Prospective cohort, coronary disease

Carotid atherosclerosis

Stroke

Peripheral vascular disease

N ¼ 292 cerebrovascular patients N ¼ 203 MI and CVD patients N ¼ 357 CVD patients N ¼ 136 CVD patients N ¼ 867 normal subjects N ¼ 750 subjects N ¼ 66 patients with borderline hypertension N ¼ 239 CVD patients N ¼ 180 stroke patients N ¼ 65 stroke patients N ¼ 93 stroke patients N ¼ 20 PVD patients

Outcome Positive relationship with anti‐HSP60 antibody titers ( p < 0.05) Positive relationship with anti‐HSP 60 ( p < 0.01) and ‐70 ( p < 0.05) antibody titers Positive relationship with anti‐HSP60 antibody titers ( p < 0.01) Negative relationship with anti‐HSP 70 antibody titers ( p > 0.05) Positive relationship with anti‐HSP60 and ‐65 antibody titers ( p < 0.001) Positive relationship with anti‐HSP70 antibody titers ( p < 0.05) Positive relationship between low risk of CVD and HSP70 titers ( p < 0.001) Positive relationship between lower levels of anti‐HSp70 antibody titers and CVD ( p < 0.001) Negative relationship with anti‐HSP60 and ‐65 antibody titers ( p > 0.05)

References [161] [212] [157] [208]

[159] [79] [216] [218]

Positive relationship with anti‐HSP‐65 antibody titers ( p < 0.05)

[82]

Positive relationship with anti‐HSP60 antibody titers ( p < 0.001) Positive relationship with anti‐HSP65 antibody titers ( p < 0.05) Positive relationship with anti‐HSP65 antibody titers ( p < 0.05)

[219] [169] [83]

Positive relationship with HSP65 antibody titers ( p < 0.05) Positive relationship with anti‐HSP65 antibody titers ( p < 0.05)

[67] [186]

Positive relationship with human IgA antibodies to HSP60 ( p < 0.05) Positive relationship with HSP65 and 70 antibody titers ( p < 0.0001) Positive relationship with HSP70 ( p < 0.05) Positive relationship with anti‐HSP60 and ‐65 antibody titers ( p < 0.01) Positive relationship with HSP70 ( p < 0.01) and negative relationship with HSP60 ( p > 0.05)

[205] [221] [222] [223] [68]

HSP, heat shock protein; CVD, cardiovascular disease; PVD, peripheral vascular disease; Ig, Immunoglobulin; MI, myocardial infarction.

58

GHAYOUR-MOBARHAN ET AL.

found higher antibody concentrations to HSP27 in patients with chest pain compared to healthy controls. We [160] have also reported that in patients with acute coronary syndrome HSP27 antibody titers are high during the first 12 h following the event, then fall to near normal levels after about 12 h.

5. Therapeutic Implications It appears that autoimmune responses may be generated against antigens present within the atherosclerotic plaque, and this leads to a cycle of ongoing vascular injury. It has been proposed that inducing a state of tolerance to these atherosclerosis associated antigens may inhibit atherogenesis and hence it may be a feasible therapeutic approach. It has been shown that immune tolerance can be induced by mucosal administration of these antigens. The eYciency of oral tolerization is dependent on the dose of antigen administered; at high doses mucosal administration leads to clonal deletion/anergy; whereas low doses induce T regulatory cells, capable of altering cytokine production [155]. Harats and colleagues have shown that tolerization to HSP65 led to a reduction of plaque formation in a murine model of atherosclerosis [227]. Tolerization was also associated with reduced macrophage and T cell infiltration and increased expression of the anti-inflammatory cytokine, IL-10 expression [227, 228]. However, it should be noted that the eVects of immunization may vary with the epitope of HSP65 used for immunization; some appear to enhance while others inhibit atherosclerosis [229]. It is also interesting that some HSPs may be used as eVective carriers for delivering B cell epitopes to the immune system in the absence of adjuvant [230, 231]. ‘Whole pathogen’ vaccines such as the BCG contain potentially immunogenic HSP, and while these vaccines reduce morbidity and mortality associated with infection, they may simultaneously stimulate proatherogenic mechanisms [2]. BCG is a live, whole organism vaccine attenuated from the bovine tubercle bacillus. It elicits a strong cell mediated immune response to several Mycobacterial antigens [232], principally a secreted form of HSP-65 [233]. Immunization with BCG vaccine, which contains HSPs, increases the extent of atherosclerosis in the cholesterol-fed rabbit, and the anti-HSP60 titers in BCG immunized rabbits were found to be correlated with the atherosclerotic plaque formation suggesting that the specific immune response to BCG-associated HSP might be proatherogenic [103, 104]. However, atherosclerotic lesions induced by BCG immunization alone in the absence of traditional risk factors such as hypercholesterolemia were found to regress with time suggesting that in the absence of other CVD risk factors, the inflammatory response to HSP60 is not enough to drive atherogenesis over prolonged period of time [101].

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

59

6. Conclusions Hence, cells involved in atherogenesis express large quantities of HSPs in response to exposure to several stressors that may also promote atherosclerosis. Measures of HSP expression, including serum antigen, or antibody concentrations may be useful as markers of disease susceptibility, although the reported data are inconsistent. This may be due to complex interactions between HSP production, release, clearance, and autoimmune responses. Studies using experimental animal models indicate that the role of HSPs in atherogenesis may not be straightforward with diVerent HSPs potentially having pro- or antiatherogenic roles. In addition to a potential role in the initiation of atherosclerosis, they may also be involved in the later stages of disease by inducing a pro-inflammatory autoimmune response [82, 83], and the recruitment of a HSP-specific inflammatory lymphocyte population [17]. The immune response to Ox-LDL appears to be antiatherogenic while that directed against HSP65 or b2-GPI is possibly proatherogenic. The relationship between other HSP antibodies (e.g., HSP70) is less clear and studies have reported either protective or deleterious eVects [159, 208]. Despite these previous data it is unclear whether the HSPs have a direct role in atherogenesis. Longer term prospective studies in diVerent populations with a more careful assessment of the time course of appearance of the HSPs and antibodies relative to the development of clinical events will be required. Further work on the eVects of tolerization may also be useful to elucidate the importance of the HSP immune response in atherogenesis. REFERENCES [1] F. Ritossa, A new puYng pattern induced and temperature shock and DNP in Drosophilia, Experimentia 18 (1962) 571–573. [2] D.J. Lamb, W. El-Sankary, G.A. Ferns, Molecular mimicry in atherosclerosis: a role for heat shock proteins in immunisation, Atherosclerosis 167 (2) (2002) 177–185. [3] K.A. Buzzard, A.J. Giaccia, M. Killender, R.L. Anderson, Heat shock protein 72 modulates pathways of stress-induced apoptosis, J. Biol. Chem. 273 (27) (1998) 17147–17153. [4] C. Garrido, S. Gurbuxani, L. Ravagnan, G. Kroemer, Heat shock proteins: endogenous modulators of apoptotic cell death, Biochem. Biophys. Res. Commun. 286 (3) (2001) 433–442. [5] M. Jaattela, D. Wissing, K. Kokholm, T. Kallunki, M. Egeblad, Hsp70 exerts its anti-apoptotic function downstream of caspase-3-like proteases, EMBO J. 17 (21) (1998) 6124–6134. [6] S.R. KirchhoV, S. Gupta, A.A. Knowlton, Cytosolic heat shock protein 60, apoptosis, and myocardial injury, Circulation 105 (24) (2002) 2899–2904. [7] E. Braunwald, Shattuck lecture—Cardiovascular medicine at the turn of the millennium: triumphs, concerns, and opportunities, N. Engl. J. Med. 337 (19) (1997) 1360–1369.

60

GHAYOUR-MOBARHAN ET AL.

[8] J.L. Breslow, Cardiovascular disease burden increases, NIH funding decreases, Nat. Med. 3 (6) (1997) 600–601. [9] R. Ross, The pathogenesis of atherosclerosis: a perspective for the 1990s, Nature 362 (6423) (1993) 801–809. [10] G.A. Ferns, New and emerging risk factors for CVD, Proc. Nutr. Soc. 67 (2) (2008) 223–231. [11] R. Virchow, Der Ateromatose Prozess der Arterien, Wien. Med. Wochenschr. 6 (1856) 825–827. [12] Q.B. Xu, G. Oberhuber, M. Gruschwitz, G. Wick, Immunology of atherosclerosis: cellular composition and major histocompatibility complex class II antigen expression in aortic intima, fatty streaks, and atherosclerotic plaques in young and aged human specimens, Clin. Immunol. Immunopathol. 56 (3) (1990) 344–359. [13] P. Libby, G.K. Hansson, Involvement of the immune system in human atherogenesis: current knowledge and unanswered questions, Lab. Invest. 64 (1) (1991) 5–15. [14] C. Napoli, F.P. D’Armiento, F.P. Mancini, et al., Fatty streak formation occurs in human fetal aortas and is greatly enhanced by maternal hypercholesterolemia. Intimal accumulation of low density lipoprotein and its oxidation precede monocyte recruitment into early atherosclerotic lesions, J. Clin. Invest. 100 (11) (1997) 2680–2690. [15] G. Millonig, H. Niederegger, W. Rabl, et al., Network of vascular-associated dendritic cells in intima of healthy young individuals, Arterioscler. Thromb. Vasc. Biol. 21 (4) (2001) 503–508. [16] C. Buono, C.E. Come, G. Stavrakis, G.F. Maguire, P.W. Connelly, A.H. Lichtman, Influence of interferon-gamma on the extent and phenotype of diet-induced atherosclerosis in the LDLR-deficient mouse, Arterioscler. Thromb. Vasc. Biol. 23 (3) (2003) 454–460. [17] R. Kleindienst, Q. Xu, J. Willeit, F.R. Waldenberger, S. Weimann, G. Wick, Immunology of atherosclerosis. Demonstration of heat shock protein 60 expression and T lymphocytes bearing alpha/beta or gamma/delta receptor in human atherosclerotic lesions, Am. J. Pathol. 142 (6) (1993) 1927–1937. [18] A.C. Van der Wal, P.K. Das, B.D. van de Berg, C.M. Van der Loos, A.E. Becker, Atherosclerotic lesions in humans. In situ immunophenotypic analysis suggesting an immune mediated response, Lab. Invest. 61 (2) (1989) 166–170. [19] S. Kiechl, G. Egger, M. Mayr, et al., Chronic infections and the risk of carotid atherosclerosis: prospective results from a large population study, Circulation 103 (8) (2001) 1064–1070. [20] P.M. Ridker, M.J. Stampfer, N. Rifai, Novel risk factors for systemic atherosclerosis: a comparison of C-reactive protein, fibrinogen, homocysteine, lipoprotein(a), and standard cholesterol screening as predictors of peripheral arterial disease, J. AM. Med. Assoc. 285 (19) (2001) 2481–2485. [21] P.M. Ridker, N. Rifai, L. Rose, J.E. Buring, N.R. Cook, Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events, N. Engl. J. Med. 347 (20) (2002) 1557–1565. [22] N. Rifai, P.M. Ridker, High-sensitivity C-reactive protein: a novel and promising marker of coronary heart disease, Clin. Chem. 47 (3) (2001) 403–411. [23] H. Yu, N. Rifai, High-sensitivity C-reactive protein and atherosclerosis: from theory to therapy, Clin. Biochem. 33 (8) (2000) 601–610. [24] R.N. Poston, D.O. Haskard, J.R. Coucher, N.P. Gall, R.R. Johnson-Tidey, Expression of intercellular adhesion molecule-1 in atherosclerotic plaques, Am. J. Pathol. 140 (3) (1992) 665–673. [25] J. Torzewski, R. Oldroyd, P. Lachmann, C. Fitzsimmons, D. Proudfoot, D. Bowyer, Complement-induced release of monocyte chemotactic protein-1 from human smooth muscle cells. A possible initiating event in atherosclerotic lesion formation, Arterioscler. Thromb. Vasc. Biol. 16 (5) (1996) 673–677.

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

61

[26] J. Torzewski, D.E. Bowyer, J. Waltenberger, C. Fitzsimmons, Processes in atherogenesis: complement activation, Atherosclerosis 132 (2) (1997) 131–138. [27] M. Torzewski, J. Torzewski, D.E. Bowyer, et al., Immunohistochemical colocalization of the terminal complex of human complement and smooth muscle cell alpha-actin in early atherosclerotic lesions, Arterioscler. Thromb. Vasc. Biol. 17 (11) (1997) 2448–2452. [28] E. Wieland, B. Dorweiler, U. Bonitz, S. Lieser, I. Walev, S. Bhakdi, Complement activation by oxidatively modified low-density lipoproteins, Eur. J. Clin. Invest. 29 (10) (1999) 835–841. [29] Z. Prohaszka, J. Duba, G. Lakos, et al., Antibodies against human heat-shock protein (hsp) 60 and mycobacterial hsp65 diVer in their antigen specificity and complementactivating ability, Int. Immunol. 11 (9) (1999) 1363–1370. [30] G.K. Hansson, Immune mechanisms in atherosclerosis, Arterioscler. Thromb. Vasc. Biol. 21 (12) (2001) 1876–1890. [31] G. Millonig, G.T. Malcom, G. Wick, Early inflammatory-immunological lesions in juvenile atherosclerosis from the Pathobiological Determinants of Atherosclerosis in Youth (PDAY)-study, Atherosclerosis 160 (2) (2002) 441–448. [32] B. Metzler, M. Mayr, H. Dietrich, et al., Inhibition of arteriosclerosis by T-cell depletion in normocholesterolemic rabbits immunized with heat shock protein 65, Arterioscler. Thromb. Vasc. Biol. 19 (8) (1999) 1905–1911. [33] Q.B. Xu, G. Oberhuber, M. Gruschwitz, G. Wick, Immunology of atherosclerosis: cellular composition and major histocompatibility complex class II antigen expression in aortic intima, fatty streaks, and atherosclerotic plaques in young and aged human specimens, Clin. Immunol. Immunopathol. 56 (3) (1990) 344–359. [34] S.A. Huber, P. Sakkinen, C. David, M.K. Newell, R.P. Tracy, T helper-cell phenotype regulates atherosclerosis in mice under conditions of mild hypercholesterolemia, Circulation 103 (21) (2001) 2610–2616. [35] L.J. Pinderski, M.P. Fischbein, G. Subbanagounder, et al., Overexpression of interleukin10 by activated T lymphocytes inhibits atherosclerosis in LDL receptor-deficient Mice by altering lymphocyte and macrophage phenotypes, Circ. Res. 90 (10) (2002) 1064–1071. [36] E. Laurat, B. Poirier, E. Tupin, et al., In vivo downregulation of T helper cell 1 immune responses reduces atherogenesis in apolipoprotein E-knockout mice, Circulation 104 (2) (2001) 197–202. [37] H. Groux, An overview of regulatory T cells, Microbes. Infect. 3 (11) (2001) 883–889. [38] P. McGuirk, C. McCann, K.H. Mills, Pathogen-specific T regulatory 1 cells induced in the respiratory tract by a bacterial molecule that stimulates interleukin 10 production by dendritic cells: a novel strategy for evasion of protective T helper type 1 responses by Bordetella pertussis, J. Exp. Med. 195 (2) (2002) 221–231. [39] M.G. Roncarolo, R. Bacchetta, C. Bordignon, S. Narula, M.K. Levings, Type 1 T regulatory cells, Immunol. Rev. 182 (2001) 68–79. [40] A. Doetze, J. Satoguina, G. Burchard, et al., Antigen-specific cellular hyporesponsiveness in a chronic human helminth infection is mediated by T(h)3/T(r)1-type cytokines IL-10 and transforming growth factor-beta but not by a T(h)1 to T(h)2 shift, Int. Immunol. 12 (5) (2000) 623–630. [41] H. Groux, A. O’Garra, M. Bigler, et al., A CD4+ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis, Nature 389 (6652) (1997) 737–742. [42] G. Caligiuri, M. Rudling, V. Ollivier, et al., Interleukin-10 deficiency increases atherosclerosis, thrombosis, and low-density lipoproteins in apolipoprotein E knockout mice, Mol. Med. 9 (1–2) (2003) 10–17. [43] D.J. Grainger, C.M. Witchell, J.C. Metcalfe, Tamoxifen elevates transforming growth factor-beta and suppresses diet-induced formation of lipid lesions in mouse aorta, Nat. Med. 1 (10) (1995) 1067–1073.

62

GHAYOUR-MOBARHAN ET AL.

[44] E. Lutgens, M. Gijbels, M. Smook, et al., Transforming growth factor-beta mediates balance between inflammation and fibrosis during plaque progression, Arterioscler. Thromb. Vasc. Biol. 22 (6) (2002) 975–982. [45] C. Buono, C.J. Binder, G. Stavrakis, J.L. Witztum, L.H. Glimcher, A.H. Lichtman, T-bet deficiency reduces atherosclerosis and alters plaque antigen-specific immune responses, Proc. Natl. Acad. Sci. USA 102 (5) (2005) 1596–1601. [46] E.E. Emeson, M.L. Shen, C.G. Bell, A. Qureshi, Inhibition of atherosclerosis in CD4 T-cell-ablated and nude (nu/nu) C57BL/6 hyperlipidemic mice, Am. J. Pathol. 149 (2) (1996) 675–685. [47] E.E. Emeson, M.L. Shen, Accelerated atherosclerosis in hyperlipidemic C57BL/6 mice treated with cyclosporin A, Am. J. Pathol. 142 (6) (1993) 1906–1915. [48] S.E. Roselaar, G. Schonfeld, A. Daugherty, Enhanced development of atherosclerosis in cholesterol-fed rabbits by suppression of cell-mediated immunity, J. Clin. Invest. 96 (3) (1995) 1389–1394. [49] M. Mayr, S. Kiechl, J. Willeit, G. Wick, Q. Xu, Infections, immunity, and atherosclerosis: associations of antibodies to Chlamydia pneumoniae, Helicobacter pylori, and cytomegalovirus with immune reactions to heat-shock protein 60 and carotid or femoral atherosclerosis, Circulation 102 (8) (2000) 833–839. [50] S.E. Epstein, Y.F. Zhou, J. Zhu, Infection and atherosclerosis: emerging mechanistic paradigms, Circulation 100 (4) (1999) e20–e28. [51] A.R. Folsom, F.J. Nieto, P. Sorlie, L.E. Chambless, D.Y. Graham, Helicobacter pylori seropositivity and coronary heart disease incidence. Atherosclerosis Risk in Communities (ARIC) Study Investigators, Circulation 98 (9) (1998) 845–850. [52] P. Saikku, M. Leinonen, K. Mattila, et al., Serological evidence of an association of a novel Chlamydia, TWAR, with chronic coronary heart disease and acute myocardial infarction, Lancet 2 (8618) (1988) 983–986. [53] J. Zhu, A.A. Quyyumi, J.E. Norman, et al., EVects of total pathogen burden on coronary artery disease risk and C-reactive protein levels, Am. J. Cardiol. 85 (2) (2000) 140–146. [54] K.J. Mattila, M.S. Nieminen, V.V. Valtonen, et al., Association between dental health and acute myocardial infarction, Br. Med. J. 298 (6676) (1989) 779–781. [55] E. Pesonen, I. Paakkari, J. Rapola, Infection-associated intimal thickening in the coronary arteries of children, Atherosclerosis 142 (2) (1999) 425–429. [56] M. Chmiela, M. Kowalewicz-Kulbat, A. Miszczak, et al., A link between Helicobacter pylori and/or Chlamydia spp. infections and atherosclerosis, FEMS Immunol. Med. Microbiol. 36 (3) (2003) 187–192. [57] C. Espinola-Klein, H.J. Rupprecht, S. Blankenberg, et al., Impact of infectious burden on extent and long-term prognosis of atherosclerosis, Circulation 105 (1) (2002) 15–21. [58] G. Caligiuri, A. Nicoletti, B. Poirier, G.K. Hansson, Protective immunity against atherosclerosis carried by B cells of hypercholesterolemic mice, J. Clin. Invest. 109 (6) (2002) 745–753. [59] A. Nicoletti, G. Caligiuri, G. Paulsson, G.K. Hansson, Functionality of specific immunity in atherosclerosis, Am. Heart J. 138 (5 Pt 2) (1999) S438–S443. [60] J.L. Witztum, Splenic immunity and atherosclerosis: a glimpse into a novel paradigm? J. Clin. Invest. 109 (6) (2002) 721–724. [61] Q. Xiao, K. Mandal, G. Schett, et al., Association of serum-soluble heat shock protein 60 with carotid atherosclerosis: clinical significance determined in a follow-up study, Stroke 36 (12) (2005) 2571–2576. [62] R. Ohashi, H. Mu, Q. Yao, C. Chen, Atherosclerosis: immunopathogenesis and immunotherapy, Med. Sci. Monit. 10 (11) (2004) RA255–RA260.

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

63

[63] G. Wick, G. Schett, A. Amberger, R. Kleindienst, Q. Xu, Is atherosclerosis an immunologically mediated disease? Immunol. Today 16 (1) (1995) 27–33. [64] K. Mandal, M. Jahangiri, Q. Xu, Autoimmune mechanisms of atherosclerosis, Handb. Exp. Pharmacol. 170 (2005) 723–743. [65] P.A. Berberian, W. Myers, M. Tytell, V. Challa, M.G. Bond, Immunohistochemical localization of heat shock protein-70 in normal-appearing and atherosclerotic specimens of human arteries, Am. J. Pathol. 136 (1) (1990) 71–80. [66] A.G. Pockley, J. Bulmer, B.M. Hanks, B.H. Wright, Identification of human heat shock protein 60 (Hsp60) and anti-Hsp60 antibodies in the peripheral circulation of normal individuals, Cell Stress Chaperones 4 (1) (1999) 29–35. [67] Q. Xu, S. Kiechl, M. Mayr, et al., Association of serum antibodies to heat-shock protein 65 with carotid atherosclerosis: clinical significance determined in a follow-up study, Circulation 100 (11) (1999) 1169–1174. [68] B.H. Wright, J.M. Corton, A.M. El-Nahas, R.F. Wood, A.G. Pockley, Elevated levels of circulating heat shock protein 70 (Hsp70) in peripheral and renal vascular disease, Heart Vessels 15 (1) (2000) 18–22. [69] A.D. Blann, C.N. McCollum, von Willebrand factor, endothelial cell damage and atherosclerosis, Eur. J. Vasc. Surg. 8 (1) (1994) 10–15. [70] A. Kol, G.K. Sukhova, A.H. Lichtman, P. Libby, Chlamydial heat shock protein 60 localizes in human atheroma and regulates macrophage tumor necrosis factor-alpha and matrix metalloproteinase expression, Circulation 98 (4) (1998) 300–307. [71] Q. Xu, G. Schett, H. Perschinka, et al., Serum soluble heat shock protein 60 is elevated in subjects with atherosclerosis in a general population, Circulation 102 (1) (2000) 14–20. [72] J.P. Halcox, J. Deanfield, A. Shamaei-Tousi, et al., Circulating human heat shock protein 60 in the blood of healthy teenagers: a novel determinant of endothelial dysfunction and early vascular injury? Arterioscler. Thromb. Vasc. Biol. 25 (11) (2005) e141–e142. [73] M. Knoflach, S. Kiechl, M. Kind, et al., Cardiovascular risk factors and atherosclerosis in young males: ARMY study (Atherosclerosis Risk-Factors in Male Youngsters), Circulation 108 (9) (2003) 1064–1069. [74] A.G. Pockley, R. Wu, C. Lemne, R. Kiessling, U. de Faire, J. Frostegard, Circulating heat shock protein 60 is associated with early cardiovascular disease, Hypertension 36 (2) (2000) 303–307. [75] A.G. Pockley, A. Georgiades, T. Thulin, U. de Faire, J. Frostegard, Serum heat shock protein 70 levels predict the development of atherosclerosis in subjects with established hypertension, Hypertension 42 (3) (2003) 235–238. [76] B. Dybdahl, A. Wahba, E. Lien, et al., Inflammatory response after open heart surgery: release of heat-shock protein 70 and signaling through toll-like receptor-4, Circulation 105 (6) (2002) 685–690. [77] A.D. Johnson, P.A. Berberian, M. Tytell, M.G. Bond, Atherosclerosis alters the localization of HSP70 in human and macaque aortas, Exp. Mol. Pathol. 58 (3) (1993) 155–168. [78] Y.V. Bobryshev, R.S. Lord, Expression of heat shock protein-70 by dendritic cells in the arterial intima and its potential significance in atherogenesis, J. Vasc. Surg. 35 (2) (2002) 368–375. [79] J. Zhu, A.A. Quyyumi, H. Wu, et al., Increased serum levels of heat shock protein 70 are associated with low risk of coronary artery disease, Arterioscler. Thromb. Vasc. Biol. 23 (6) (2003) 1055–1059. [80] W. Zhu, P. Roma, F. Pellegatta, A.L. Catapano, Oxidized-LDL induce the expression of heat shock protein 70 in human endothelial cells, Biochem. Biophys. Res. Commun. 200 (1) (1994) 389–394.

64

GHAYOUR-MOBARHAN ET AL.

[81] S. Genth-Zotz, A.P. Bolger, P.R. Kalra, et al., Heat shock protein 70 in patients with chronic heart failure: relation to disease severity and survival, Int. J. Cardiol. 96 (3) (2004) 397–401. [82] F. Hoppichler, M. Lechleitner, C. Traweger, et al., Changes of serum antibodies to heat-shock protein 65 in coronary heart disease and acute myocardial infarction, Atherosclerosis 126 (2) (1996) 333–338. [83] Q. Xu, J. Willeit, M. Marosi, et al., Association of serum antibodies to heat-shock protein 65 with carotid atherosclerosis, Lancet 341 (8840) (1993) 255–259. [84] M. Knoflach, S. Kiechl, B. Mayrl, et al., T-cell reactivity against HSP60 relates to early but not advanced atherosclerosis, Atherosclerosis 195 (2) (2007) 333–338. [85] S. Kiechl, J. Willeit, The natural course of atherosclerosis. Part I: incidence and progression, Arterioscler. Thromb. Vasc. Biol. 19 (6) (1999) 1484–1490. [86] S. Kiechl, J. Willeit, The natural course of atherosclerosis. Part II: vascular remodeling. Bruneck Study Group, Arterioscler. Thromb. Vasc. Biol. 19 (6) (1999) 1491–1498. [87] J.M. Ramage, J.L. Young, J.C. Goodall, J.S. Gaston, T cell responses to heat-shock protein 60: diVerential responses by CD4+ T cell subsets according to their expression of CD45 isotypes, J. Immunol. 162 (2) (1999) 704–710. [88] M.G. Danieli, M. Candela, A.M. Ricciatti, et al., Antibodies to mycobacterial 65 kDa heat shock protein in systemic sclerosis (scleroderma), J. Autoimmun. 5 (4) (1992) 443–452. [89] S. Ishibashi, M.S. Brown, J.L. Goldstein, R.D. Gerard, R.E. Hammer, J. Herz, Hypercholesterolemia in low density lipoprotein receptor knockout mice and its reversal by adenovirus-mediated gene delivery, J. Clin. Invest. 92 (2) (1993) 883–893. [90] A.S. Plump, J.D. Smith, T. Hayek, et al., Severe hypercholesterolemia and atherosclerosis in apolipoprotein E-deficient mice created by homologous recombination in ES cells, Cell 71 (2) (1992) 343–353. [91] S.H. Zhang, R.L. Reddick, J.A. Piedrahita, N. Maeda, Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E, Science 258 (5081) (1992) 468–471. [92] M.S. Marber, R. Mestril, S.H. Chi, M.R. Sayen, D.M. Yellon, W.H. Dillmann, Overexpression of the rat inducible 70-kD heat stress protein in a transgenic mouse increases the resistance of the heart to ischemic injury, J. Clin. Invest. 95 (4) (1995) 1446–1456. [93] R.W. Currie, M. Karmazyn, M. Kloc, K. Mailer, Heat-shock response is associated with enhanced postischemic ventricular recovery, Circ. Res. 63 (3) (1988) 543–549. [94] A. Afek, J. George, B. Gilburd, et al., Immunization of low-density lipoprotein receptor deficient (LDL-RD) mice with heat shock protein 65 (HSP-65) promotes early atherosclerosis, J. Autoimmun. 14 (2) (2000) 115–121. [95] J. George, Y. Shoenfeld, A. Afek, et al., Enhanced fatty streak formation in C57BL/6J mice by immunization with heat shock protein-65, Arterioscler. Thromb. Vasc. Biol. 19 (3) (1999) 505–510. [96] L.B. Kirby, J.S. Mondy, C.M. Brophy, Balloon angioplasty induces heat shock protein 70 in human blood vessels, Ann. Vasc. Surg. 13 (5) (1999) 475–479. [97] J. George, S. Greenberg, I. Barshack, I. Goldberg, G. Keren, A. Roth, Immunity to heat shock protein 65—An additional determinant in intimal thickening, Atherosclerosis 168 (1) (2003) 33–38. [98] J. George, A. Afek, B. Gilburd, Y. Shoenfeld, D. Harats, Cellular and humoral immune responses to heat shock protein 65 are both involved in promoting fatty-streak formation in LDL-receptor deficient mice, J. Am. Coll. Cardiol. 38 (3) (2001) 900–905. [99] P. Keren, J. George, A. Shaish, et al., EVect of hyperglycemia and hyperlipidemia on atherosclerosis in LDL receptor-deficient mice: establishment of a combined model and association with heat shock protein 65 immunity, Diabetes 49 (6) (2000) 1064–1069.

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

65

[100] Q. Xu, H. Dietrich, H.J. Steiner, et al., Induction of arteriosclerosis in normocholesterolemic rabbits by immunization with heat shock protein 65, Arterioscler. Thromb. 12 (7) (1992) 789–799. [101] Q. Xu, R. Kleindienst, G. Schett, et al., Regression of arteriosclerotic lesions induced by immunization with heat shock protein 65-containing material in normocholesterolemic, but not hypercholesterolemic, rabbits, Atherosclerosis 123 (1–2) (1996) 145–155. [102] Q. Xu, R. Kleindienst, W. Waitz, H. Dietrich, G. Wick, Increased expression of heat shock protein 65 coincides with a population of infiltrating T lymphocytes in atherosclerotic lesions of rabbits specifically responding to heat shock protein 65, J. Clin. Invest. 91 (6) (1993) 2693–2702. [103] D.J. Lamb, L.J. Eales, G.A. Ferns, Immunization with bacillus Calmette-Guerin vaccine increases aortic atherosclerosis in the cholesterol-fed rabbit, Atherosclerosis 143 (1) (1999) 105–113. [104] D.J. Lamb, G.A. Ferns, The magnitude of the immune response to heat shock protein-65 following BCG immunisation is associated with the extent of experimental atherosclerosis, Atherosclerosis 165 (2) (2002) 231–240. [105] M. Ghayour-Mobarhan, D.J. Lamb, S. Tavallaie, G.A. Ferns, Relationship between plasma cholesterol, von Willebrand factor concentrations, extent of atherosclerosis and antibody titres to heat shock proteins-60, -65 and -70 in cholesterol-fed rabbits, Int. J. Exp. Pathol. 88 (4) (2007) 249–255. [106] R.K. Kanwar, J.R. Kanwar, D. Wang, D.J. Ormrod, G.W. Krissansen, Temporal expression of heat shock proteins 60 and 70 at lesion-prone sites during atherogenesis in ApoEdeficient mice, Arterioscler. Thromb. Vasc. Biol. 21 (12) (2001) 1991–1997. [107] G. Pfister, C.M. Stroh, H. Perschinka, et al., Detection of HSP60 on the membrane surface of stressed human endothelial cells by atomic force and confocal microscopy, J. Cell Sci. 118 (Pt 8) (2005) 1587–1594. [108] G. Schett, Q. Xu, A. Amberger, et al., Autoantibodies against heat shock protein 60 mediate endothelial cytotoxicity, J. Clin. Invest. 96 (6) (1995) 2569–2577. [109] C. Jamin, C. Dugue, J.E. Alard, et al., Induction of endothelial cell apoptosis by the binding of anti-endothelial cell antibodies to Hsp60 in vasculitis-associated systemic autoimmune diseases, Arthritis Rheum. 52 (12) (2005) 4028–4038. [110] A.D. Johnson, P.A. Berberian, M. Tytell, M.G. Bond, DiVerential distribution of 70-kD heat shock protein in atherosclerosis. Its potential role in arterial SMC survival, Arterioscler. Thromb. Vasc. Biol. 15 (1) (1995) 27–36. [111] Q. Xu, G. Schett, C. Li, Y. Hu, G. Wick, Mechanical stress-induced heat shock protein 70 expression in vascular smooth muscle cells is regulated by Rac and Ras small G proteins but not mitogen-activated protein kinases, Circ. Res. 86 (11) (2000) 1122–1128. [112] M. Amrani, J. Corbett, S.Y. Boateng, M.J. Dunn, M.H. Yacoub, Kinetics of induction and protective eVect of heat-shock proteins after cardioplegic arrest, Ann. Thorac. Surg. 61 (5) (1996) 1407–1411. [113] R.W. Currie, R.M. Tanguay, J.G. Kingma, Jr, Heat-shock response and limitation of tissue necrosis during occlusion/reperfusion in rabbit hearts, Circulation 87 (3) (1993) 963–971. [114] D.V. Cumming, R.J. Heads, A. Watson, D.S. Latchman, D.M. Yellon, DiVerential protection of primary rat cardiocytes by transfection of specific heat stress proteins, J. Mol. Cell Cardiol. 28 (12) (1996) 2343–2349. [115] D.V. Cumming, R.J. Heads, R.S. CoYn, D.M. Yellon, D.S. Latchman, Pharmacological preconditioning of primary rat cardiac myocytes by FK506, Basic Res. Cardiol. 91 (5) (1996) 367–373.

66

GHAYOUR-MOBARHAN ET AL.

[116] R.J. Heads, D.S. Latchman, D.M. Yellon, Stable high level expression of a transfected human HSP70 gene protects a heart-derived muscle cell line against thermal stress, J. Mol. Cell Cardiol. 26 (6) (1994) 695–699. [117] A.A. Knowlton, The role of heat shock proteins in the heart, J. Mol. Cell Cardiol. 27 (1) (1995) 121–131. [118] M.M. Hutter, R.E. Sievers, V. Barbosa, C.L. Wolfe, Heat-shock protein induction in rat hearts. A direct correlation between the amount of heat-shock protein induced and the degree of myocardial protection, Circulation 89 (1) (1994) 355–360. [119] W.H. Dillmann, H.B. Mehta, A. Barrieux, B.D. Guth, W.E. Neeley, J. Ross, Jr, Ischemia of the dog heart induces the appearance of a cardiac mRNA coding for a protein with migration characteristics similar to heat-shock/stress protein 71, Circ. Res. 59 (1) (1986) 110–114. [120] O.N. Demidov, V.V. Tyrenko, A.S. Svistov, et al., Heat shock proteins in cardiosurgery patients, Eur. J. Cardiothorac. Surg. 16 (4) (1999) 444–449. [121] G. Valen, G.K. Hansson, A. Dumitrescu, J. Vaage, Unstable angina activates myocardial heat shock protein 72, endothelial nitric oxide synthase, and transcription factors NFkappaB and AP-1, Cardiovasc. Res. 47 (1) (2000) 49–56. [122] R. Mestril, W.H. Dillmann, Heat shock proteins and protection against myocardial ischemia, J. Mol. Cell Cardiol. 27 (1) (1995) 45–52. [123] A.D. Johnson, P.A. Berberian, M.G. Bond, EVect of heat shock proteins on survival of isolated aortic cells from normal and atherosclerotic cynomolgus macaques, Atherosclerosis 84 (2–3) (1990) 111–119. [124] J.L. Martin, R. Mestril, R. Hilal-Dandan, L.L. Brunton, W.H. Dillmann, Small heat shock proteins and protection against ischemic injury in cardiac myocytes, Circulation 96 (12) (1997) 4343–4348. [125] J.L. Martin, E. Hickey, L.A. Weber, W.H. Dillmann, R. Mestril, Influence of phosphorylation and oligomerization on the protective role of the small heat shock protein 27 in rat adult cardiomyocytes, Gene. Expr. 7 (4–6) (1999) 349–355. [126] I. Schimke, G. Lutsch, U. Schernes, et al., Increased level of HSP27 but not of HSP72 in human heart allografts in relation to acute rejection, Transplantation 70 (12) (2000) 1694–1697. [127] W. Chen, U. Syldath, K. Bellmann, V. Burkart, H. Kolb, Human 60-kDa heat-shock protein: a danger signal to the innate immune system, J. Immunol. 162 (6) (1999) 3212–3219. [128] J.M. Bruey, C. Ducasse, P. Bonniaud, et al., Hsp27 negatively regulates cell death by interacting with cytochrome c, Nat. Cell Biol. 2 (9) (2000) 645–652. [129] C. Garrido, J.M. Bruey, A. Fromentin, A. Hammann, A.P. Arrigo, E. Solary, HSP27 inhibits cytochrome c-dependent activation of procaspase-9, FASEB J. 13 (14) (1999) 2061–2070. [130] B.K. Brar, A. Stephanou, M.J. WagstaV, et al., Heat shock proteins delivered with a virus vector can protect cardiac cells against apoptosis as well as against thermal or hypoxic stress, J. Mol. Cell Cardiol. 31 (1) (1999) 135–146. [131] A.K. De, K.M. Kodys, B.S. Yeh, C. Miller-Graziano, Exaggerated human monocyte IL-10 concomitant to minimal TNF-alpha induction by heat-shock protein 27 (Hsp27) suggests Hsp27 is primarily an antiinflammatory stimulus, J. Immunol. 165 (7) (2000) 3951–3958. [132] R.M. Vabulas, P. Ahmad-Nejad, S. Ghose, C.J. Kirschning, R.D. Issels, H. Wagner, HSP70 as endogenous stimulus of the Toll/interleukin-1 receptor signal pathway, J. Biol. Chem. 277 (17) (2002) 15107–15112. [133] K. Mandal, M. Jahangiri, Q. Xu, Autoimmunity to heat shock proteins in atherosclerosis, Autoimmun. Rev. 3 (2) (2004) 31–37.

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

67

[134] A. Kol, T. Bourcier, A.H. Lichtman, P. Libby, Chlamydial and human heat shock protein 60s activate human vascular endothelium, smooth muscle cells, and macrophages, J. Clin. Invest. 103 (4) (1999) 571–577. [135] W.L. Beatty, R.P. Morrison, G.I. Byrne, Persistent chlamydiae: from cell culture to a paradigm for chlamydial pathogenesis, Microbiol. Rev. 58 (4) (1994) 686–699. [136] C.M. Boulanger, A. Scoazec, T. Ebrahimian, et al., Circulating microparticles from patients with myocardial infarction cause endothelial dysfunction, Circulation 104 (22) (2001) 2649–2652. [137] Z. Mallat, B. Hugel, J. Ohan, G. Leseche, J.M. Freyssinet, A. Tedgui, Shed membrane microparticles with procoagulant potential in human atherosclerotic plaques: a role for apoptosis in plaque thrombogenicity, Circulation 99 (3) (1999) 348–353. [138] Z. Mallat, A. Tedgui, Current perspective on the role of apoptosis in atherothrombotic disease, Circ. Res. 88 (10) (2001) 998–1003. [139] M. Mesri, D.C. Altieri, Leukocyte microparticles stimulate endothelial cell cytokine release and tissue factor induction in a JNK1 signaling pathway, J. Biol. Chem. 274 (33) (1999) 23111–23118. [140] A. Kol, A.H. Lichtman, R.W. Finberg, P. Libby, E.A. Kurt-Jones, Cutting edge: heat shock protein (HSP) 60 activates the innate immune response: CD14 is an essential receptor for HSP60 activation of mononuclear cells, J. Immunol. 164 (1) (2000) 13–17. [141] Q. Xu, Role of heat shock proteins in atherosclerosis, Arterioscler. Thromb. Vasc. Biol. 22 (10) (2002) 1547–1559. [142] A. Asea, S.K. Kraeft, E.A. Kurt-Jones, et al., HSP70 stimulates cytokine production through a CD14-dependant pathway, demonstrating its dual role as a chaperone and cytokine, Nat. Med. 6 (4) (2000) 435–442. [143] P. Matzinger, Tolerance, danger, and the extended family, Annu. Rev. Immunol. 12 (1994) 991–1045. [144] P.K. Srivastava, A. Menoret, S. Basu, R.J. Binder, K.L. McQuade, Heat shock proteins come of age: primitive functions acquire new roles in an adaptive world, Immunity 8 (6) (1998) 657–665. [145] G. Wick, M. Romen, A. Amberger, et al., Atherosclerosis, autoimmunity, and vascularassociated lymphoid tissue, FASEB J. 11 (13) (1997) 1199–1207. [146] Q. Xu, G. Wick, The role of heat shock proteins in protection and pathophysiology of the arterial wall, Mol. Med. Today 2 (9) (1996) 372–379. [147] M. Mayr, B. Metzler, S. Kiechl, et al., Endothelial cytotoxicity mediated by serum antibodies to heat shock proteins of Escherichia coli and Chlamydia pneumoniae: immune reactions to heat shock proteins as a possible link between infection and atherosclerosis, Circulation 99 (12) (1999) 1560–1566. [148] H.M. Dansky, S.A. Charlton, M.M. Harper, J.D. Smith, T and B lymphocytes play a minor role in atherosclerotic plaque formation in the apolipoprotein E-deficient mouse, Proc. Natl. Acad. Sci. USA 94 (9) (1997) 4642–4646. [149] A. Daugherty, E. Pure, D. Defel-Butteiger, et al., The eVects of total lymphocyte deficiency on the extent of atherosclerosis in apolipoprotein E-/-mice, J. Clin. Invest. 100 (6) (1997) 1575–1580. [150] A.I. Fyfe, J.H. Qiao, A.J. Lusis, Immune-deficient mice develop typical atherosclerotic fatty streaks when fed an atherogenic diet, J. Clin. Invest. 94 (6) (1994) 2516–2520. [151] S. Gupta, A.M. Pablo, X. Jiang, N. Wang, A.R. Tall, C. Schindler, IFN-gamma potentiates atherosclerosis in ApoE knock-out mice, J. Clin. Invest. 99 (11) (1997) 2752–2761. [152] J. Frostegard, A.K. Ulfgren, P. Nyberg, et al., Cytokine expression in advanced human atherosclerotic plaques: dominance of pro-inflammatory (Th1) and macrophage-stimulating cytokines, Atherosclerosis 145 (1) (1999) 33–43.

68

GHAYOUR-MOBARHAN ET AL.

[153] M. Knoflach, D. Bernhard, G. Wick, Anti-HSP60 immunity is already associated with atherosclerosis early in life, Ann. N. Y. Acad. Sci. 1051 (2005) 323–331. [154] R. Rigano, E. Profumo, B. Buttari, et al., Heat shock proteins and autoimmunity in patients with carotid atherosclerosis, Ann. N. Y. Acad. Sci. 1107 (2007) 1–10. [155] J. George, N. Yacov, E. Breitbart, et al., Suppression of early atherosclerosis in LDLreceptor deficient mice by oral tolerance with beta 2-glycoprotein I, Cardiovasc. Res. 62 (3) (2004) 603–609. [156] H. Perschinka, M. Mayr, G. Millonig, et al., Cross-reactive B-cell epitopes of microbial and human heat shock protein 60/65 in atherosclerosis, Arterioscler. Thromb. Vasc. Biol. 23 (6) (2003) 1060–1065. [157] J. Zhu, A.A. Quyyumi, D. Rott, et al., Antibodies to human heat-shock protein 60 are associated with the presence and severity of coronary artery disease: evidence for an autoimmune component of atherogenesis, Circulation 103 (8) (2001) 1071–1075. [158] K. Buhlin, A. Gustafsson, A.G. Pockley, J. Frostegard, B. Klinge, Risk factors for cardiovascular disease in patients with periodontitis, Eur. Heart J. 24 (23) (2003) 2099–2107. [159] Y.C. Chan, N. Shukla, M. Abdus-Samee, et al., Anti-heat-shock protein 70 kDa antibodies in vascular patients, Eur. J. Vasc. Endovasc. Surg. 18 (5) (1999) 381–385. [160] M. Ghayour-Mobarhan, A. Sahebkar, S.M. Parizadeh, et al., Antibody titres to heat shock protein 27 are elevated in patients with acute coronary syndrome, Int. J. Exp. Pathol. 89 (3) (2008) 209–215. [161] A. Ciervo, P. Visca, A. Petrucca, L.M. Biasucci, A. Maseri, A. Cassone, Antibodies to 60-kilodalton heat shock protein and outer membrane protein 2 of Chlamydia pneumoniae in patients with coronary heart disease, Clin. Diagn. Lab. Immunol. 9 (1) (2002) 66–74. [162] J. Mollenhauer, A. Schulmeister, The humoral immune response to heat shock proteins, Experientia 48 (7) (1992) 644–649. [163] A. Schattner, B. Rager-Zisman, Virus-induced autoimmunity, Rev. Infect. Dis. 12 (2) (1990) 204–222. [164] U. Feige, W. van Eden, Infection, autoimmunity and autoimmune disease, EXS 77 (1996) 359–373. [165] A. Veres, Z. Prohaszka, S. Kilpinen, M. Singh, G. Fust, M. Hurme, The promoter polymorphism of the IL-6 gene is associated with levels of antibodies to 60-kDa heatshock proteins, Immunogenetics 53 (10–11) (2002) 851–856. [166] Z. Prohaszka, J. Duba, G. Lakos, et al., Antibodies against human heat-shock protein (hsp) 60 and mycobacterial hsp65 diVer in their antigen specificity and complementactivating ability, Int. Immunol. 11 (9) (1999) 1363–1370. [167] Q. Xu, G. Schett, C.S. Seitz, Y. Hu, R.S. Gupta, G. Wick, Surface staining and cytotoxic activity of heat-shock protein 60 antibody in stressed aortic endothelial cells, Circ. Res. 75 (6) (1994) 1078–1085. [168] A. Veres, T. Szamosi, M. Ablonczy, et al., Complement activating antibodies against the human 60 kDa heat shock protein as a new independent family risk factor of coronary heart disease, Eur. J. Clin. Invest. 32 (6) (2002) 405–410. [169] D.H. Birnie, E.R. Holme, I.C. McKay, S. Hood, K.E. McColl, W.S. Hillis, Association between antibodies to heat shock protein 65 and coronary atherosclerosis. Possible mechanism of action of Helicobacter pylori and other bacterial infections in increasing cardiovascular risk, Eur. Heart J. 19 (3) (1998) 387–394. [170] K. Heltai, Z. Kis, K. Burian, et al., Elevated antibody levels against Chlamydia pneumoniae, human HSP60 and mycobacterial HSP65 are independent risk factors in myocardial infarction and ischaemic heart disease, Atherosclerosis 173 (2) (2004) 339–346. [171] K. Burian, Z. Kis, D. Virok, et al., Independent and joint eVects of antibodies to human heat-shock protein 60 and Chlamydia pneumoniae infection in the development of coronary atherosclerosis, Circulation 103 (11) (2001) 1503–1508.

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

69

[172] S. Hoshida, M. Nishino, J. Tanouchi, T. Kishimoto, Y. Yamada, Acute Chlamydia pneumoniae infection with heat-shock-protein-60-related response in patients with acute coronary syndrome, Atherosclerosis 183 (1) (2005) 109–112. [173] T. Huittinen, M. Leinonen, L. Tenkanen, et al., Synergistic eVect of persistent Chlamydia pneumoniae infection, autoimmunity, and inflammation on coronary risk, Circulation 107 (20) (2003) 2566–2570. [174] V.Y. Hoymans, J.M. Bosmans, P.L. Van Herck, M.M. Ieven, C.J. Vrints, Implications of antibodies to heat-shock proteins in ischemic heart disease, Int. J. Cardiol. 123 (3) (2008) 277–282. [175] C.A. Jantos, C. Krombach, F.N. Wuppermann, et al., Antibody response to the 60-kDa heat-shock protein of Chlamydia pneumoniae in patients with coronary artery disease, J. Infect. Dis. 181 (5) (2000) 1700–1705. [176] R.G. Deshpande, M.B. Khan, C.A. Genco, Invasion of aortic and heart endothelial cells by Porphyromonas gingivalis, Infect. Immun. 66 (11) (1998) 5337–5343. [177] K. Tabeta, K. Yamazaki, H. Hotokezaka, H. Yoshie, K. Hara, Elevated humoral immune response to heat shock protein 60 (hsp60) family in periodontitis patients, Clin. Exp. Immunol. 120 (2) (2000) 285–293. [178] J.I. Choi, S.W. Chung, H.S. Kang, et al., Epitope mapping of Porphyromonas gingivalis heat-shock protein and human heat-shock protein in human atherosclerosis, J. Dent. Res. 83 (12) (2004) 936–940. [179] G. Schett, B. Metzler, R. Kleindienst, et al., Salivary anti-hsp65 antibodies as a diagnostic marker for gingivitis and a possible link to atherosclerosis, Int. Arch. Allergy Immunol. 114 (3) (1997) 246–250. [180] M. Bongrazio, L. Comini, G. Gaia, T. Bachetti, R. Ferrari, Hypertension, aging, and myocardial synthesis of heat-shock protein 72, Hypertension 24 (5) (1994) 620–624. [181] Q. Xu, D.G. Li, N.J. Holbrook, R. Udelsman, Acute hypertension induces heat-shock protein 70 gene expression in rat aorta, Circulation 92 (5) (1995) 1223–1229. [182] D. Elias, D. Markovits, T. Reshef, R. van der Zee, I.R. Cohen, Induction and therapy of autoimmune diabetes in the non-obese diabetic (NOD/Lt) mouse by a 65-kDa heat shock protein, Proc. Natl. Acad. Sci. USA 87 (4) (1990) 1576–1580. [183] Z. Szilvassy, P. Ferdinandy, J. Szilvassy, et al., The loss of pacing-induced preconditioning in atherosclerotic rabbits: role of hypercholesterolaemia, J. Mol. Cell Cardiol. 27 (12) (1995) 2559–2569. [184] T. Csont, G. Balogh, C. Csonka, et al., Hyperlipidemia induced by high cholesterol diet inhibits heat shock response in rat hearts, Biochem. Biophys. Res. Commun. 290 (5) (2002) 1535–1538. [185] A. Shamaei-Tousi, J.W. Stephens, R. Bin, et al., Association between plasma levels of heat shock protein 60 and cardiovascular disease in patients with diabetes mellitus, Eur. Heart J. 27 (13) (2006) 1565–1570. [186] J. Frostegard, C. Lemne, B. Andersson, R. van der Zee, R. Kiessling, U. de Faire, Association of serum antibodies to heat-shock protein 65 with borderline hypertension, Hypertension 29 (1 Pt 1) (1997) 40–44. [187] I.M. Rea, S. McNerlan, A.G. Pockley, Serum heat shock protein and anti-heat shock protein antibody levels in aging, Exp. Gerontol. 36 (2) (2001) 341–352. [188] S. Shams, S. Shafi, K. Bodman-Smith, P. Williams, S. Mehta, G.A. Ferns, Anti-heat shock protein-27 (Hsp-27) antibody levels in patients with chest pain: association with established cardiovascular risk factors, Clin. Chim. Acta 395 (1–2) (2008) 42–46. [189] I. Kardys, N. Rifai, O. Meilhac, et al., Plasma concentration of heat shock protein 27 and risk of cardiovascular disease: a prospective, nested case – control study, Clin. Chem. 54 (1) (2008) 139–146.

70

GHAYOUR-MOBARHAN ET AL.

[190] A.G. Pockley, U. de Faire, R. Kiessling, C. Lemne, T. Thulin, J. Frostegard, Circulating heat shock protein and heat shock protein antibody levels in established hypertension, J. Hypertens 20 (9) (2002) 1815–1820. [191] T. Wu, J. Ma, S. Chen, et al., Association of plasma antibodies against the inducible Hsp70 with hypertension and harsh working conditions, Cell Stress Chaperones 6 (4) (2001) 394–401. [192] J. Kunes, M. Poirier, J. Tremblay, P. Hamet, Expression of hsp70 gene in lymphocytes from normotensive and hypertensive humans, Acta Physiol. Scand. 146 (3) (1992) 307–311. [193] N. Vaidya, M. Ghayour-Mobarhan, D.J. Lamb, G.A. Ferns, Anti heat shock protein27 (Hsp27) titres and serum hs-CRP in patients with traditional coronary risk factors including metabolic syndrome (MS), Clin. Chim. Acta 355 (2005) S118. [194] X. Zhang, M.A. He, L. Cheng, et al., Joint eVects of antibody to heat shock protein 60, hypertension, and diabetes on risk of coronary heart disease in Chinese, Clin. Chem. 54 (6) (2008) 1046–1052. [195] D. Bernhard, G. Pfister, C.W. Huck, et al., Disruption of vascular endothelial homeostasis by tobacco smoke: impact on atherosclerosis, FASEB J. 17 (15) (2003) 2302–2304. [196] D. Bernhard, C.W. Huck, T. Jakschitz, et al., Development and evaluation of an in vitro model for the analysis of cigarette smoke eVects on cultured cells and tissues, J. Pharmacol. Toxicol. Methods 50 (1) (2004) 45–51. [197] D. Bernhard, A. Csordas, B. Henderson, A. Rossmann, M. Kind, G. Wick, Cigarette smoke metal-catalyzed protein oxidation leads to vascular endothelial cell contraction by depolymerization of microtubules, FASEB J. 19 (9) (2005) 1096–1107. [198] G. Schett, B. Metzler, R. Kleindienst, et al., Myocardial injury leads to a release of heat shock protein (hsp) 60 and a suppression of the anti-hsp65 immune response, Cardiovasc. Res. 42 (3) (1999) 685–695. [199] H. Kervinen, T. Huittinen, O. Vaarala, et al., Antibodies to human heat shock protein 60, hypertension and dyslipidemia. A study of joint eVects on coronary risk, Atherosclerosis 169 (2) (2003) 339–344. [200] A. Figueredo, J.L. Ibarra, A. Rodriguez, et al., Increased serum levels of IgA antibodies to hsp70 protein in patients with diabetes mellitus: their relationship with vascular complications, Clin. Immunol. Immunopathol. 79 (3) (1996) 252–255. [201] T.J. Sims, A. Lernmark, L.A. Mancl, R.E. SchiVerle, R.C. Page, G.R. Persson, Serum IgG to heat shock proteins and Porphyromonas gingivalis antigens in diabetic patients with periodontitis, J. Clin. Periodontol. 29 (6) (2002) 551–562. [202] A. Shamaei-Tousi, A. Steptoe, K. O’Donnell, et al., Plasma heat shock protein 60 and cardiovascular disease risk: the role of psychosocial, genetic, and biological factors, Cell Stress Chaperones 12 (4) (2007) 384–392. [203] H.K. Park, E.C. Park, S.W. Bae, et al., Expression of heat shock protein 27 in human atherosclerotic plaques and increased plasma level of heat shock protein 27 in patients with acute coronary syndrome, Circulation 114 (9) (2006) 886–893. [204] M. Ghayour-Mobarhan, S.A. New, D.J. Lamb, et al., Dietary antioxidants and fat are associated with plasma antibody titers to heat shock proteins 60, 65, and 70 in subjects with dyslipidemia, Am. J. Clin. Nutr. 81 (5) (2005) 998–1004. [205] T. Huittinen, M. Leinonen, L. Tenkanen, et al., Autoimmunity to human heat shock protein 60, Chlamydia pneumoniae infection, and inflammation in predicting coronary risk, Arterioscler. Thromb. Vasc. Biol. 22 (3) (2002) 431–437. [206] M. Ghayour-Mobarhan, D.J. Lamb, N. Vaidya, C. Livingstone, T. Wang, G.A. Ferns, Heat shock protein antibody titers are reduced by statin therapy in dyslipidemic subjects: a pilot study, Angiology 56 (1) (2005) 61–68.

THE POTENTIAL ROLE OF HEAT SHOCK PROTEINS

71

[207] Y.O. Shin, J.S. Bae, J.B. Lee, et al., EVect of cardiac rehabilitation and statin treatment on anti-HSP antibody titers in patients with coronary artery disease after percutaneous coronary intervention, Int. Heart J. 47 (5) (2006) 671–682. [208] J. Kocsis, A. Veres, A. Vatay, et al., Antibodies against the human heat shock protein hsp70 in patients with severe coronary artery disease, Immunol. Invest. 31 (3–4) (2002) 219–231. [209] B. Kwak, F. Mulhaupt, S. Myit, F. Mach, Statins as a newly recognized type of immunomodulator, Nat. Med. 6 (12) (2000) 1399–1402. [210] A. Rosengren, S. Hawken, S. Ounpuu, et al., Association of psychosocial risk factors with risk of acute myocardial infarction in 11119 cases and 13648 controls from 52 countries (the INTERHEART study): case-control study, Lancet 364 (9438) (2004) 953–962. [211] J. Lewthwaite, N. Owen, A. Coates, B. Henderson, A. Steptoe, Circulating human heat shock protein 60 in the plasma of British civil servants: relationship to physiological and psychosocial stress, Circulation 106 (2) (2002) 196–201. [212] J. Yuan, M. Yang, H. Yao, et al., Plasma antibodies to heat shock protein 60 and heat shock protein 70 are associated with increased risk of electrocardiograph abnormalities in automobile workers exposed to noise, Cell Stress Chaperones 10 (2) (2005) 126–135. [213] J. Zhu, R.J. Katz, A.A. Quyyumi, et al., Association of serum antibodies to heat-shock protein 65 with coronary calcification levels: suggestion of pathogen-triggered autoimmunity in early atherosclerosis, Circulation 109 (1) (2004) 36–41. [214] M. Mukherjee, B.C. De, D. Jewitt, V.V. Kakkar, Association of antibodies to heat-shock protein-65 with percutaneous transluminal coronary angioplasty and subsequent restenosis, Thromb. Haemost. 75 (2) (1996) 258–260. [215] Q. Xu, G. Luef, S. Weimann, R.S. Gupta, H. Wolf, G. Wick, Staining of endothelial cells and macrophages in atherosclerotic lesions with human heat-shock protein-reactive antisera, Arterioscler. Thromb. 13 (12) (1993) 1763–1769. [216] I. Herz, R. Rosso, A. Roth, G. Keren, J. George, Serum levels of anti heat shock protein 70 antibodies in patients with stable and unstable angina pectoris, Acute Card. Care 8 (1) (2006) 46–50. [217] S. Vogt, I. Portig, B. Kusch, et al., Detection of anti-hsp70 immunoglobulin G antibodies indicates better outcome in coronary artery bypass grafting patients suVering from severe preoperative angina, Ann. Thorac. Surg. 78 (3) (2004) 883–889. [218] J. Kramer, P. Harcos, Z. Prohaszka, et al., Frequencies of certain complement protein alleles and serum levels of anti-heat-shock protein antibodies in cerebrovascular diseases, Stroke 31 (11) (2000) 2648–2652. [219] Z. Prohaszka, J. Duba, L. Horvath, et al., Comparative study on antibodies to human and bacterial 60 kDa heat shock proteins in a large cohort of patients with coronary heart disease and healthy subjects, Eur. J. Clin. Invest. 31 (4) (2001) 285–292. [220] A. Veres, G. Fust, M. Smieja, et al., Relationship of anti-60 kDa heat shock protein and anticholesterol antibodies to cardiovascular events, Circulation 106 (22) (2002) 2775–2780. [221] G. Gromadzka, J. Zielinska, D. Ryglewicz, U. Fiszer, A. Czlonkowska, Elevated levels of anti-heat shock protein antibodies in patients with cerebral ischemia, Cerebrovasc. Dis. 12 (3) (2001) 235–239. [222] X. Jin, C. Xiao, R.M. Tanguay, et al., Correlation of lymphocyte heat shock protein 70 levels with neurologic deficits in elderly patients with cerebral infarction, Am. J. Med. 117 (6) (2004) 406–411. [223] H.L. Staub, G.L. Norman, T. Crowther, et al., Antibodies to the atherosclerotic plaque components beta2-glycoprotein I and heat-shock proteins as risk factors for acute cerebral ischemia, Arq. Neuropsiquiatr. 61 (3B) (2003) 757–763. [224] G.K. Hansson, Immunological markers of atherosclerosis, Lancet 341 (8840) (1993) 278.

72

GHAYOUR-MOBARHAN ET AL.

[225] I.U. Khan, R. Wallin, R.S. Gupta, G.M. Kammer, Protein kinase A-catalyzed phosphorylation of heat shock protein 60 chaperone regulates its attachment to histone 2B in the T lymphocyte plasma membrane, Proc. Natl. Acad. Sci. USA 95 (18) (1998) 10425–10430. [226] R. Gruber, S. Lederer, U. Bechtel, S. Lob, G. Riethmuller, H.E. Feucht, Increased antibody titers against mycobacterial heat-shock protein 65 in patients with vasculitis and arteriosclerosis, Int. Arch. Allergy Immunol. 110 (1) (1996) 95–98. [227] D. Harats, N. Yacov, B. Gilburd, Y. Shoenfeld, J. George, Oral tolerance with heat shock protein 65 attenuates Mycobacterium tuberculosis-induced and high-fat-diet-driven atherosclerotic lesions, J. Am. Coll. Cardiol. 40 (7) (2002) 1333–1338. [228] R. Maron, G. Sukhova, A.M. Faria, et al., Mucosal administration of heat shock protein65 decreases atherosclerosis and inflammation in aortic arch of low-density lipoprotein receptor-deficient mice, Circulation 106 (13) (2002) 1708–1715. [229] G. Foteinos, A.R. Afzal, K. Mandal, M. Jahangiri, Q. Xu, Anti-heat shock protein 60 autoantibodies induce atherosclerosis in apolipoprotein E-deficient mice via endothelial damage, Circulation 112 (8) (2005) 1206–1213. [230] C. Barrios, A.R. Lussow, E.J. Van, et al., Mycobacterial heat-shock proteins as carrier molecules. II: the use of the 70-kDa mycobacterial heat-shock protein as carrier for conjugated vaccines can circumvent the need for adjuvants and Bacillus Calmette Guerin priming, Eur. J. Immunol. 22 (6) (1992) 1365–1372. [231] R. Perraut, A.R. Lussow, S. Gavoille, et al., Successful primate immunization with peptides conjugated to purified protein derivative or mycobacterial heat shock proteins in the absence of adjuvants, Clin. Exp. Immunol. 93 (3) (1993) 382–386. [232] B. Sander, U. Skansen-Saphir, O. Damm, L. Hakansson, J. Andersson, U. Andersson, Sequential production of Th1 and Th2 cytokines in response to live bacillus Calmette– Guerin, Immunology 86 (4) (1995) 512–518. [233] A.S. Mustafa, K.E. Lundin, F. Oftung, Human T cells recognize mycobacterial heat shock proteins in the context of multiple HLA-DR molecules: studies with healthy subjects vaccinated with Mycobacterium bovis BCG and Mycobacterium leprae, Infect. Immun. 61 (12) (1993) 5294–5301. [234] P.J. Ford, K. Yamazaki, G.J. Seymour, Cardiovascular and oral disease interactions: What is the evidence? Prim. Dent. Care 14 (2) (2007) 59–66.

ADVANCES IN CLINICAL CHEMISTRY, VOL. 48

THE EMERGING ROLE OF SYMMETRIC DIMETHYLARGININE IN VASCULAR DISEASE Arduino A. Mangoni1 Department of Clinical Pharmacology, School of Medicine, Flinders University, Adelaide 5001, Australia

1. Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Synthesis, Transport, and Metabolism of ADMA and SDMA . . . . . . . . . . . . . . . . . . . 3.1. Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Transport and Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. ADMA and the Cardiovascular System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Endothelial Function and Hemodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Arterial StiVness, Cardiac Function, Atherosclerosis, and Inflammation . . . . 4.3. Cardiovascular Outcomes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. SDMA and the Cardiovascular System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Endothelial Function and Cardiac Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Cardiovascular Homeostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Cardiovascular Outcomes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

73 74 75 75 77 78 78 78 79 79 79 84 87 88 89

1. Abstract Asymmetric dimethylarginine (ADMA), an endogenous methylated form of the amino acid L‐arginine, inhibits the activity of the enzyme endothelial nitric oxide synthase (eNOS), with consequent reduced synthesis of nitric oxide (NO). An increased synthesis and/or a reduced catabolism of ADMA might contribute to the onset and progression of atherosclerosis and thrombosis. The detrimental eVects of ADMA on endothelial function, cardiovascular homeostasis, and cardiovascular outcomes have been extensively 1

Corresponding author: Arduino A. Mangoni, e‐mail: [email protected] 73

0065-2423/09 $35.00 DOI: 10.1016/S0065-2423(09)48003-X

Copyright 2009, Elsevier Inc. All rights reserved.

74

ARDUINO A. MANGONI

investigated. However, little attention has been paid to another methylated form of L‐arginine, symmetric dimethylarginine (SDMA), as a potential modulator of vascular homeostasis and vascular disease. The first part of this chapter discusses the synthesis, transport, and metabolism of ADMA and SDMA and summarizes the evidence linking ADMA with vascular disease and adverse cardiovascular outcomes. The second part describes the results of recent studies highlighting the important role of SDMA in modulating vascular homeostasis and vascular damage. Suggestions for future research directions on SDMA are also discussed.

2. Introduction The endothelium plays a crucial role in regulating vascular homeostasis through the synthesis of the endogenous vasodilator NO by the enzyme eNOS [1]. NO exerts important eVects on vascular tone, such as vasorelaxation and reduction of arterial stiVness, as well as significant anti‐ inflammatory, antithrombotic, and antiatherosclerotic eVects [2, 3]. Not surprisingly, vascular disease states such as hypertension and diabetes are characterized by a reduced synthesis of NO by eNOS that is endothelial dysfunction [2]. This leads to a shift in the balance between vasodilatation and vasoconstriction with predominance of the latter [2]. Moreover, it favors the onset and progression of atherosclerosis and thrombosis [2]. There is very good evidence that an impairment of endothelial function, in the form of reduced endothelium‐dependent vasodilatation, predicts adverse cardiovascular outcomes. A multivariate analysis of published studies conducted on 2500 patients with various cardiovascular conditions, with a follow‐up ranging from 1 to 92 months, has demonstrated that endothelial dysfunction is strongly and independently associated with increased cardiovascular morbidity and mortality [4]. Endothelial dysfunction could be considered as the integration of the diVerent traditional risk factors for a number of reasons. First, the presence and the severity of individual cardiovascular risk factors are both associated with impaired endothelial function. Second, the coexistence of diVerent risk factors has additive detrimental eVects on endothelial function. Third, both the management and the elimination of risk factors generally lead to an improvement in endothelial function [5–9]. The identification of biochemical, pharmacological, and hemodynamic factors regulating NO synthesis is currently one of the main areas of interest in cardiovascular research [10–19]. Studies conducted over the last 20 years have demonstrated that ADMA, an endogenous methylated form of the amino acid L‐arginine, modulates eNOS activity by exerting powerful inhibitory eVects [20]. An increase in the plasma concentrations of ADMA has

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

75

been widely shown to exert detrimental eVects on vascular homeostasis by impairing endothelial function, increasing arterial stiVness, and promoting vascular inflammation [21–23]. Not surprisingly, a number of prospective studies have demonstrated that elevated plasma ADMA concentrations independently predict adverse outcomes in several cardiovascular patient groups [24–33]. Unlike ADMA, the biological activity of SDMA, another methylated form of L‐arginine, has been poorly investigated. The opinion that SDMA is relatively inert has been the main reason for the lack of specific studies on this methylarginine. However, recent in vitro and in vivo investigations have suggested that SDMA possesses important biological activities. Scope of this chapter is to (1) discuss the synthesis, metabolism, and main biological eVects of ADMA and SDMA; (2) describe the results of recent studies demonstrating the potential role of SDMA in cardiovascular homeostasis and vascular disease; and (3) propose future research directions in this area.

3. Synthesis, Transport, and Metabolism of ADMA and SDMA 3.1. SYNTHESIS ADMA and SDMA are synthesized within cells following the methylation of arginine residues in proteins by protein arginine methyltransferases (PRMTs type I and II) [34]. As a result two methyl groups are added to the guanidine nitrogens of arginine. The PRMT type I methylates only one guanidine nitrogen group, resulting in the formation of ADMA, whereas the PRMT type II methylates both groups and leads to the synthesis of SDMA (Figs. 1 and 2) [34]. Both types of PRMT can also add only one methyl group with the formation of NG‐monomethyl L‐arginine (L‐NMMA) [34]. PRMTs use S‐adenosylmethionine, which is synthesized from methionine and adenosine triphosphate, as methyl group donor [34] (Fig. 2). The methylation of arginine is virtually irreversible. However, this concept has been recently challenged [35, 36]. After the transfer of the methyl donor S‐adenosylmethionine is converted firstly into S‐adenosylhomocysteine and then into homocysteine. The latter is either metabolized in the trans‐sulfuration route or remethylated to methionine [37]. Proteolysis leads to the liberation of free ADMA, SDMA, and L‐NMMA into the cytoplasm [38]. There is no evidence for a direct route of synthesis of these methylated forms from free arginine. Therefore, the amount produced depends on both the extent and the rate of arginine methylation in proteins and protein turnover. However, it is unknown whether these processes are constant. There is evidence that changes in PMRT expression correlate with

76

ARDUINO A. MANGONI

CH2 NH

NH2

NH2

NH

NH

CH2 NH

CH2 N

CH2

CH2

NH

NH

C

C

C

C

NH

NH

NH

NH

CH2

CH2

CH2

CH2

CH2

CH2

CH2

CH2

CH2

CH2

CH2

CH2

CH

CH

CH

CH

NH2

COOH

Arginine

COOH

L-NMMA

NH2

COOH

NH2

ADMA

COOH SDMA

FIG. 1. Arginine and its methylated forms. L‐NMMA, NG‐monomethyl L‐arginine; ADMA, asymmetric dimethylarginine; SDMA, symmetric dimethylarginine.

Methionine

Protein

SAM

PRMT (transmethylation)

Remethylation

Homocysteine

Methylated protein

SAH

Proteolysis DDAH Citrulline + dimethylamine

ADMA + SDMA

CAT

Plasma ADMA + SDMA FIG. 2. Intracellular synthesis, metabolism, and transport of asymmetric dimethylarginine (ADMA). SAM, S‐adenosylmethionine; SAH, S‐adenosylhomocysteine; PRMT, protein arginine methyltransferase; DDAH, dimethylarginine dimetylaminohydrolase; CAT, cationic amino acid transporter.

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

77

changes in the liberation of free ADMA, but not SDMA, in the cytosol [39]. These findings suggest a diVerential regulation of ADMA and SDMA formation. Recent animal studies have demonstrated that proteins containing methylated arginine residues are degraded at a relatively high rate with liberation of 66% of ADMA after 60 min [40]. By contrast, proteins containing arginine are degraded at a lower rate [40]. The mechanisms responsible for the rapid degradation of proteins containing methylated arginine are unknown, although a trigger phenomenon leading to the activation of the 20S proteasome has been speculated [41]. 3.2. TRANSPORT AND METABOLISM Both intracellular ADMA and L‐NMMA are metabolized to citrulline and dimethylamine by the enzyme dimethylarginine dimethylaminohydrolase (DDAH) [42]. There are two DDAH isoforms, DDAH‐1 and DDAH‐2, with diVerent tissue distribution [43]. DDAH‐1 is particularly expressed in the liver and the kidney. Not surprisingly, these organs are the major sites of ADMA metabolism [44, 45]. DDAH‐1 is also expressed in the pancreas, aorta, and the forebrain [46, 47]. By contrast, DDAH‐2 is highly expressed in the vascular endothelium, heart, placenta, and the kidney [47]. The modulation of DDAH activity plays a pivotal role in regulating intracellular ADMA concentrations, with important eVects on vascular homeostasis. For example, impairment in DDAH activity, resulting in elevated ADMA concentrations and reduced NO synthesis, can promote the onset and progression of atherosclerosis [48]. By contrast, DDAH overexpression can significantly reduce ADMA concentrations and its detrimental eVects on endothelial function [49]. The Km of DDAH is 180 mmol/L, much higher than the physiological intracellular ADMA concentrations [42]. Therefore, under physiological conditions, the rate of ADMA metabolism is proportional to its concentration. The metabolism of ADMA by DDAH seems to involve a nucleophilic attack on the guanidine portion of the ADMA molecule by a cysteine held in an activated state in the tertiary structure of the enzyme [50]. Of note, this cysteine can be oxidated by NO. There is evidence that increased NO production nitrosates DDAH and inhibits its activity [51]. This provides an important negative feedback system whereby an increase in NO concentrations can switch oV further NO synthesis through increased concentrations of ADMA. Although most of the ADMA (>90%) is metabolized in the cytosol, a small fraction escapes local degradation and is transported, together with SDMA, across the cell membrane into the circulation via the yþ cationic amino acid transport system [52]. Changes in the expression of this transport

78

ARDUINO A. MANGONI

system can significantly influence this process [52]. Physiological reference ranges for plasma ADMA and SDMA concentrations have been proposed [53–55]. However, they are significantly influenced by the diVerent analytical approaches available [56]. The fraction of ADMA that escapes intracellular metabolism is metabolized in the liver and in the kidney through DDAH, although a small fraction of ADMA can also be excreted in the urine [45, 57]. A significant increase in ADMA plasma concentrations is commonly observed in patients with either renal or liver disease [20, 45, 58, 59]. SDMA is almost entirely eliminated by renal excretion [57]. However, this concept has been recently challenged as the liver abundantly expresses the yþ cationic amino acid transporter [60]. Siroen et al. have demonstrated a significant uptake of SDMA in the human liver in patients undergoing hepatic surgery [59].

4. ADMA and the Cardiovascular System 4.1. ENDOTHELIAL FUNCTION AND HEMODYNAMICS There is overwhelming evidence that ADMA inhibits eNOS activity and increases the Km of this enzyme [20, 58, 61]. The IC50 of ADMA is similar to L‐NMMA, that is, 2–5 mM [58]. However, it also depends on arginine concentrations as the inhibitory eVects can be reversed by adding excess arginine [62]. Intravenous ADMA administration in animals increases renal, mesenteric, and peripheral vascular resistance with a concomitant increase in blood pressure [63, 64]. Intravenous ADMA infusion in humans induces, in a dose‐dependent fashion, a sustained reduction in NO synthesis, cardiac output, and renal perfusion and a concomitant increase in peripheral vascular resistance, blood pressure, and sodium reabsorption in the kidney [65, 66]. Moreover, increased plasma ADMA concentrations have been observed in animal models of hypertension as well as in patients with arterial or pulmonary hypertension [33, 67–69]. 4.2. ARTERIAL STIFFNESS, CARDIAC FUNCTION, ATHEROSCLEROSIS, AND INFLAMMATION Intravenous infusion of ADMA in healthy volunteers significantly reduces cerebral blood perfusion and increases arterial stiVness [22]. An increase in arterial stiVness leads to a higher systolic blood pressure and a lower diastolic blood pressure [70]. The consequent increase in cardiac afterload can favor both the development of left ventricular hypertrophy and a critical reduction of coronary perfusion pressure with the onset of myocardial ischemia [70].

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

79

Increased plasma ADMA concentrations are significantly correlated with markers of abnormal left ventricular relaxation and diastolic dysfunction in patients with chronic heart failure [32]. In vitro studies have also demonstrated that ADMA plays a pivotal role in favoring the onset and progression of atherosclerosis by (1) enhancing monocyte adhesion induced by angiotensin‐II; (2) activating chemokine receptors by the reactive oxygen species/NF‐kappaB pathway; (3) reducing erythrocyte deformability; and (4) promoting vascular inflammation [23, 71, 72]. 4.3. CARDIOVASCULAR OUTCOMES Several prospective studies of various sample size and follow‐up have demonstrated that elevated plasma ADMA concentrations strongly and independently predict increased cardiovascular morbidity and/or mortality in healthy women and the general population [73, 74], patients with renal failure [25], ischemic heart disease [24, 27, 30, 31], peripheral vascular disease [26], chronic heart failure [28, 32], idiopathic pulmonary hypertension [33], and diabetes [29, 75] (Table 1). Notably, the predictive eVect of ADMA in these studies has been shown to be independent of age, gender, and other established cardiovascular risk factors.

5. SDMA and the Cardiovascular System Until recently, there was no evidence for a significant biological role of SDMA [20]. The results of recent studies on the role of SDMA in modulating cardiovascular homeostasis and renal function are discussed in the following paragraphs (Table 2). 5.1. ENDOTHELIAL FUNCTION AND CARDIAC FUNCTION Bode‐Bo¨ger et al. studied the eVects of increasing concentrations of SDMA (0, 2, 5, 10, and 100 mmol/L) on NO synthesis, eNOS expression, and the content of reactive oxygen species in endothelial cells cultured in medium containing arginine 70 mmol/L and incubated for 24 h [76]. SDMA significantly reduced NO synthesis and increased intracellular oxidative stress in a dose‐dependent fashion (P < 0.05 vs control). However, eNOS expression was not significantly modified by SDMA. Notably, the eVects of SDMA both on NO synthesis and on oxidative stress were already present at concentrations (2 mmol/L) close to those measured in healthy subjects and in patients with vascular disease [53, 76, 77]. The eVects of SDMA on NO synthesis and reactive oxygen species were abolished by adding the eNOS

TABLE 1 STUDIES ASSESSING THE PREDICTIVE ROLE OF ELEVATED ADMA AND SDMA PLASMA CONCENTRATIONS ON CARDIOVASCULAR OUTCOMES

Reference

Study population

Follow‐up

End‐points

Method

[25]

End‐stage renal failure (n ¼ 225)

33.4 months

[73]

Healthy women (n ¼ 880)

24 years

[31]

Cardiogenic shock (n ¼ 79)

30 days

All‐cause mortality and cardiovascular events All‐cause mortality, cardiovascular mortality, noncardiovascular mortality, and cardiovascular events All‐cause mortality

[30]

Myocardial infarction (n ¼ 249)

1 year

All‐cause mortality

HPLC

[27]

Ischemic heart disease (n ¼ 2543)

5.4 years

All‐cause mortality and cardiovascular mortality

HPLC

[24]

Ischemic heart disease (n ¼ 1874)

2.6 years

Cardiovascular death and cardiovascular events

ELISA

[26]

Peripheral artery disease (n ¼ 496)

19 months

Cardiovascular events

HPLC

HPLC HPLC

MS

Results—predictive power ADMA: yes SDMA: no ADMA: yes SDMA: not assessed ADMA: yes SDMA: trend (P ¼ 0.08) ADMA: yes SDMA: no ADMA: yes SDMA: not assessed ADMA: yes SDMA: not assessed ADMA: yes SDMA: not assessed

[28]

Chronic heart failure (n ¼ 253)

14.2 months

[75]

11.3 years

[29]

Type 1 diabetes and diabetic nephropathy (n ¼ 397) Type 2 diabetes (n ¼ 125)

21 months

[32]

Chronic heart failure (n ¼ 132)

33 months

[33]

Idiopathic pulmonary hypertension (n ¼ 57)

26 months

[74]

General population (n ¼ 572)

5 years

Cardiovascular decompensation, cardiovascular events, and all‐cause mortality Cardiovascular disease and cardiovascular mortality Cardiovascular events

HPLC

ADMA: yes SDMA: not assessed

HPLC

ADMA: yes SDMA: no ADMA: yes SDMA: not assessed ADMA: yes SDMA: trend (P ¼ 0.17)

HPLC

All‐cause mortality, cardiac transplantation, or hospitalization for heart failure All‐cause mortality

MS

Cardiovascular death and cardiovascular events

MS

MS

ADMA: yes SDMA: not assessed ADMA: yes SDMA: yes

ADMA, asymmetric dimethylarginine; SDMA, symmetric dimethylarginine; HPLC, high‐performance liquid chromatography; ELISA, enzyme‐linked immunosorbent assay; MS, mass spectrometry.

82

ARDUINO A. MANGONI TABLE 2 STUDIES ON THE ROLE OF SDMA IN VASCULAR HOMEOSTASIS AND VASCULAR DISEASE

Reference [76]

Methods/population Patients undergoing coronary angiography (n ¼ 147) Human umbilical vein endothelial cells

[78]

Monocyte oxidative burst activity from whole blood (healthy volunteers)

[32]

Patients with chronic heart failure (n ¼ 132)

[68]

Rat model of hypertension (postunilateral or bilateral ureteral occlusion)

[67]

Patients with idiopathic pulmonary hypertension (n ¼ 11)

[81]

Patients with mild‐moderate renal failure (n ¼ 227)

[82]

Patients with chest pain with or without overt coronary artery disease (n ¼ 145) Rodents and patients with renal failure

[83]

[84]

Patients with renal failure (n ¼ 135)

[85]

Patients with mild‐moderate renal failure (n ¼ 93)

Results SDMA is a marker of glomerular filtration rate and extent of coronary artery disease SDMA inhibits nitric oxide synthesis and increases the production of reactive oxygen species. The eVects are dose dependent SDMA stimulates the production of reactive oxygen species by monocytes. This eVect is mediated by an increased calcium entry from the extracellular milieu SDMA is significantly correlated with echocardiographic parameters of impaired left ventricular relaxation and with serum concentrations of plasma aminoterminal pro‐B‐type natriuretic peptide Higher SDMA concentrations observed in hypertensive rats with unilateral and bilateral ureteral occlusion versus controls Higher SDMA concentrations observed in patients with idiopathic pulmonary hypertension versus controls SDMA significantly correlated with glomerular filtration rate and serum creatinine concentrations SDMA significantly correlated with glomerular filtration rate SDMA significantly correlated with creatinine clearance both in rodents and in patients with renal failure SDMA significantly correlated with creatinine clearance SDMA significantly correlated with creatinine clearance (continues)

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

83

TABLE 2 (Continued) Reference

Methods/population

[86]

Patients with renal failure (n ¼ 44)

[87]

Children and adolescents with hypertension (n ¼ 38) and controls (n ¼ 9)

[79]

Meta‐analysis of 18 studies (n ¼ 2136)

[80]

Children and adolescents with renal disease (n ¼ 28)

[77]

Patients with end‐stage renal disease undergoing hemodialysis (n ¼ 52)

Results SDMA significantly correlated with glomerular filtration rate SDMA significantly correlated with glomerular filtration rate SDMA significantly higher in hypertensive patients SDMA concentrations significantly correlated with markers of renal function and glomerular filtration rate SDMA concentrations and SDMA/ ADMA ratio significantly correlated with glomerular filtration rate and blood pressure load SDMA concentrations independently predict the onset of intradialytic hypotension

substrate L‐arginine [76]. The results of this elegant study suggest that (1) SDMA inhibits NO synthesis by reducing L‐arginine availability and (2) the eVects of SDMA are mediated by reactive oxygen species. The eVects of SDMA on reactive oxygen species have been further investigated. Schepers et al. studied the eVects of SDMA and ADMA on intracellular Ca2þ concentrations in monocytes isolated from whole blood in healthy volunteers [78]. The stimulating eVect of SDMA on reactive oxygen species production was accompanied by an increase in Ca2þ entry from extracellular compartments. This resulted in higher amplitude of the peak changes in cytoplasmatic Ca2þ [(Ca2þ)i]. Interestingly, ADMA did not exert any significant eVects on either reactive oxygen species or changes in cytoplasmic Ca2þ [78]. Bode‐Bo¨ger et al. also investigated the relationship between SDMA plasma concentrations and the severity of coronary atherosclerosis, documented with angiography, using a stepwise regression analysis approach [76]. The parameters entered into the analysis were ADMA, SDMA, and renal function parameters such as the glomerular filtration rate (GFR) and serum concentrations of parathyroid hormone. Only SDMA was retained in the final model (R2 ¼ 0.075; P ¼ 0.003) [76]. These results indicate that SDMA

84

ARDUINO A. MANGONI

plasma concentrations are independently associated with the severity of coronary atherosclerosis [76]. Wilson Tang et al. observed a significant positive relationship between higher plasma SDMA concentrations, estimates of reduced left ventricular relaxation (mitral E/A ratio, mitral deceleration time, and pulmonary vein S/D ratio), and serum concentrations of plasma aminoterminal pro‐B‐type natriuretic peptide in patients with chronic heart failure [32]. Although no cause–eVect relationship was determined, these results suggest a potential role for SDMA in modulating cardiac remodeling in heart failure. A recent study has demonstrated the possible involvement of SDMA in the development of hypertension. Carlstro¨m et al. measured plasma concentrations of SDMA and blood pressure in rats rendered hydronephrotic after unilateral or bilateral ureteral obstruction [68]. Hypertensive rats with bilateral ureteral obstruction demonstrated higher plasma SDMA concentrations (0.67 0.15 mmol/L) compared to hypertensive rats with unilateral ureteral obstruction (0.32 0.02 mmol/L) and normotensive controls (0.29 0.01 mmol/L) [68]. Similarly, increased plasma SDMA concentrations have been observed in patients with idiopathic pulmonary hypertension vs. healthy controls (1.46 0.24 vs 0.53 0.07 mmol/L, P < 0.05) [67].

5.2. CARDIOVASCULAR HOMEOSTASIS 5.2.1. Renal Function Several studies have demonstrated that SDMA is a reliable marker of renal function in humans [76, 79–87]. Fliser et al. measured plasma ADMA and SDMA concentrations in 227 patients with mild‐moderate renal failure [81]. Plasma SDMA concentrations were significantly correlated both with GFR (R ¼ 0.837, P < 0.01) and with serum creatinine concentrations (R ¼ 0.894, P < 0.01). Similar correlations were observed with plasma ADMA concentrations (GFR: R ¼ 0.591, P < 0.01; serum creatinine: R ¼ 0.595; P < 0.01) [81]. Bode‐Bo¨ger et al. measured L‐arginine, ADMA, and SDMA plasma concentrations in 97 patients with ischemic heart disease by liquid chromatography–mass spectrometry [76]. Renal function was assessed by measuring GFR (estimated GFR, eGFR) using the established modification of diet in renal disease formula [88]. Stepwise regression analysis adjusted for ADMA, extent of coronary artery disease, and parathyroid hormone concentrations, showed that SDMA was the only parameter independently associated with eGFR and accounted for more than 30% of the eGFR variance (R2 ¼ 0.35, P < 0.001) [76].

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

85

Wang et al. measured plasma homocysteine, ADMA, and SDMA concentrations in 145 patients with and without overt coronary artery disease [82]. A significant negative correlation was observed between these compounds and eGFR (homocysteine: R ¼ 0.22, P < 0.01; ADMA: R ¼ 0.21, P < 0.05; SDMA: R ¼ 0.45, P < 0.01) [82]. Al Banchaabouchi et al. investigated the relationship between the concentration of methylated arginines and creatinine clearance in mice, rats, and in patients with renal failure [83]. There was a significant correlation between plasma ADMA and SDMA and creatinine clearance in humans (ADMA: R ¼ 0.703, P < 0.0001; SDMA: R ¼ 0.714, P < 0.0001) and rats (ADMA: R ¼ 0.711, P < 0.0001; SDMA: R ¼ 0.836, P < 0.0001). However, in mice there was a significant correlation only between SDMA and creatinine clearance (R ¼ 0.398, P < 0.05) [83]. Marescau et al. measured the serum concentrations of ADMA, SDMA, and other 13 guanidino compounds in 135 patients with renal failure [84]. The serum guanidine compounds showing the highest correlation with creatinine clearance were guanidinosuccinic acid (R ¼ 0.926, P < 0.0001), SDMA (R ¼ 0.916, P < 0.0001), methylguanidine (R ¼ 0.871, P < 0.0001), and guanidine (R ¼ 0.827, P < 0.0001). A lower correlation was observed with ADMA (R ¼ 0.700, P < 0.0001) [84]. Nanayakkara et al. measured plasma ADMA and SDMA concentrations in 93 patients with mild‐moderate renal failure [85]. SDMA concentrations were strongly correlated with creatinine clearance both in univariate and in multivariate analysis after adjusting for age, body mass index, total cholesterol, and history of smoking (b ¼ 0.850, P < 0.0001) [85]. Similarly, Kielstein et al. observed a strong correlation between plasma SDMA concentrations and GFR in 44 patients with diVerent degrees of renal failure (R ¼ 0.78, P < 0.0001) [86]. By contrast, the correlation between ADMA and GFR failed to reach statistical significance [86]. Goonasekera et al. measured plasma ADMA, SDMA, and L‐NMMA in 38 hypertensive children and adolescents and in 9 healthy controls [87]. Both plasma ADMA and SDMA concentrations were higher in hypertensive subjects (0.23 0.03 vs 0.10 0.01 mmol/L, P < 0.001; 1.37 0.06 vs 1.18 0.06 mmol/L, P ¼ 0.03, respectively) whereas L‐NMMA concentrations were not significantly diVerent. There was a significant correlation between lower GFR and higher concentrations of ADMA (R ¼ 0.77, P < 0.001), SDMA (R ¼ 0.38, P ¼ 0.02), and L‐NMMA (R ¼ 0.35, P ¼ 0.03) [87]. Kielstein et al. performed a meta‐analysis of 18 studies containing data on plasma SDMA concentrations and renal function conducted in 2136 subjects [79]. SDMA concentrations were significantly and positively correlated with inulin clearance (R ¼ 0.85, 95% CI 0.76–0.91, P < 0.0001), various estimates of GFR (R ¼ 0.77, 95% CI 0.65–0.85, P < 0.0001), and with serum creatinine

86

ARDUINO A. MANGONI

concentrations (R ¼ 0.75, 95% CI 0.46–0.89, P < 0.0001) [79]. In contrast to SDMA, ADMA plasma concentrations were significantly correlated with renal function in only some studies [79]. Brooks et al. analyzed the relationship between ADMA, SDMA, eGFR, and 24‐h blood pressure load in children and adolescents with renal disease and in controls [80]. eGFR explained 42–60% of the variability of SDMA/ADMA ratio and SDMA plasma concentrations. Moreover, SDMA concentrations and SDMA/ADMA ratio explained 27–40% of the diastolic blood pressure variability (P ¼ 0.013 and 0.032, respectively) [80]. Notably, plasma ADMA concentrations were not predictive of blood pressure load in this study. The results of these studies suggest that SDMA is a reliable marker of renal function and a potential modulator of blood pressure load. These findings might have pathophysiological and clinical implications for at least two reasons. First, estimates of GFR using serum creatinine concentrations are insensitive to even‐moderate reductions in GFR and heavily rely on a parameter which is strongly influenced by muscle mass, protein intake, age, and gender [89]. Second, large prospective studies have provided solid evidence that renal dysfunction is a major risk factor for cardiovascular morbidity and mortality in several patient groups [90–92]. It is possible that SDMA might serve as a combined marker of renal function, vascular disease, and cardiovascular risk in view of its relationship with GFR and the detrimental eVects on endothelial function previously discussed. Clearly more research is needed to test this hypothesis. 5.2.2. Intradialytic Hypotension As previously discussed, ADMA and SDMA plasma concentrations are elevated in patients with renal failure [58]. However, an acute reduction in plasma concentrations is observed in patients undergoing hemodialysis. This is due to the loss of these molecules through the dialysis membrane [93, 94]. Mangoni et al. investigated whether the acute fluctuation in the plasma concentration of ADMA and SDMA might mediate the onset of intradialytic hypotension (IDH), as a result of the acute removal of factors favoring vasoconstriction [77]. After adjusting for systolic blood pressure, clinical variables, ADMA, L‐NMMA, and L‐arginine plasma concentrations, the odds of IDH occurring were higher with increased prehemodialysis SDMA plasma concentrations (OR ¼ 1.31 per 0.1 mmol/L SDMA increase; 95% CI 1.04–1.65, P ¼ 0.02) and with decreases in SDMA during hemodialysis (OR ¼ 1.39 per 0.1 mmol/L SDMA decrease; 95% CI 1.02–1.91, P ¼ 0.04) [77]. Therefore, both elevated prehemodialysis SDMA plasma concentrations and their reductions during hemodialysis independently predict the onset of IDH. Although the exact mechanism responsible is unknown, these findings might reflect the previously discussed inhibitory eVects of

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

87

SDMA on NO synthesis, leading to endothelial dysfunction. The latter is associated with an impairment of arterial and cardiopulmonary baroreflexes [95]. Impairment in these reflexes may diminish the ability of vasoconstrictive mechanisms to maintain blood pressure within a physiological range under conditions of stress, that is, a change in volume status during hemodialysis, thereby increasing the risk of IDH. The relationship between reductions in SDMA plasma concentrations and IDH might be related to the rapid fall in the blood concentration of an inhibitor of NO synthesis with increased NO production and greater vasodilatation. This hypothesis, however, warrants further research. 5.3. CARDIOVASCULAR OUTCOMES Until recently, SDMA plasma concentrations have not been shown to predict cardiovascular outcomes in prospective studies apart from a trend toward statistical significance in the studies by Nicholls et al. in patients with cardiogenic shock after myocardial infarction and by Wilson Tang et al. in patients with chronic heart failure [31, 32] (Table 1). In a very recent study, Kiechl et al. measured ADMA and SDMA plasma concentrations in 572 subjects followed up for 5 years [74]. Regression analysis showed that both ADMA and SDMA predicted the primary end‐point which was a composite of all cardiovascular events, including ischemic stroke, TIA, myocardial infarction, vascular death, and revascularization procedures [74]. The hazard ratios over the 5‐year period were of similar strength, 3.86 (1.36–10.9) for ADMA and 7.91 (1.94–32.3) for SDMA [74]. It should be noted that SDMA plasma concentrations were not measured, or their impact on outcomes not assessed, in most studies [24, 26–29, 33, 73]. In the study by Zoccali et al., SDMA failed to predict outcomes in patients with end‐stage renal failure undergoing hemodialysis [25]. However, it must be emphasized that both creatinine and SDMA plasma concentrations rise exponentially with decreasing renal function. This could have reduced the inter‐subject variability resulting in very high baseline SDMA concentrations in most study subjects, thus minimizing the impact of this variable in predicting adverse outcomes. Another important point to be considered when interpreting the results of these studies is the diVerent methodology for the measurement of ADMA and SDMA plasma concentrations (Table 1). While high‐performance liquid chromatography has been widely used in the last few years both the enzyme‐ linked immunosorbent assay and mass spectrometry techniques oVer significant advantages in terms of rapidity. However, mass spectrometry is unsurpassed in its unique ability to identify ADMA and SDMA on the basis of their distinct mass spectrums [56].

88

ARDUINO A. MANGONI

6. Discussion ADMA has been extensively shown to inhibit eNOS activity, to promote the onset and progression of atherosclerosis and thrombosis, and to predict adverse cardiovascular outcomes. The results of recent studies suggest that SDMA can also exert significant eVects on cardiovascular homeostasis. In particular, the recent evidence that SDMA inhibits NO production and increases the production of reactive oxygen species sheds new lights on the potential role of this methylated arginine in contributing to vascular disease and, perhaps, allowing a better stratification of cardiovascular risk. A reliable biomarker of cardiovascular risk should possess the following characteristics: (1) be easily measurable in the population; (2) be related to cardiovascular risk in a linear, or other predictable fashion; and (3) be modifiable by means of pharmacological and/or nonpharmacological interventions. Although SDMA can be easily measured in the blood, its relationship with cardiovascular risk is far from being established. A recent study has demonstrated that SDMA levels can be modulated by means of pharmacological intervention. Aslam et al. investigated the eVects of two antihypertensive drugs, amlodipine and valsartan, on plasma ADMA and SDMA concentrations in patients with end‐stage renal disease [96]. Both treatments significantly reduced both ADMA (38.1–39.2%) and SDMA (37.5–41.6%) concentrations [96]. The authors speculated a direct inhibitory eVect of valsartan and amlodipine on PRMTs as a result of their antioxidant properties [96]. This would lead to decreased synthesis and liberation of ADMA and SDMA into the cytoplasm and, ultimately, in the circulation. More research is urgently warranted in the following areas: (1) the eVects of SDMA on endothelial cells’ homeostasis, redox state, and modulation of arterial tone and arterial stiVness; (2) the interplay between SDMA, ADMA, as well as other biochemical markers of vascular damage such as C‐reactive protein, homocysteine, and endothelin, in regulating eNOS activity and NO synthesis; (3) the relationship between SDMA and blood pressure control; (4) the potential advantages of SDMA, as compared with other established markers, in determining GFR and renal function; and (5) the eVects of pharmacological and nonpharmacological interventions in modulating plasma SDMA concentrations. Clinical studies on patients with diVerent cardiovascular risk profiles should be performed to identify diVerences, if any, in serum concentrations of SDMA as well as other methylated forms of L‐arginine and other established risk factors of vascular disease. Then, prospective studies should be appropriately powered on the statistical diVerences obtained from these cross‐sectional studies. The method for measuring the plasma concentrations of L‐arginine and its methylated derivatives is of pivotal importance to

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

89

correctly interpret the results of such studies. There is good evidence that mass spectrometry is currently the technique of choice because of its inherent high specificity and the ability to measure low‐serum concentrations of the methylated arginines and distinguish between the diVerent subfractions. In conclusion, SDMA has the potential to represent a strong and reliable marker of vascular disease, renal function, and cardiovascular risk. However, a correct scientific approach coupled with rigorous study design are both mandatory to clarify the role of SDMA in vascular homeostasis and vascular disease. This is an important, yet exciting, area of research in the years to come. REFERENCES [1] L.J. Ignarro, Nitric oxide as a unique signaling molecule in the vascular system: a historical overview, J. Physiol. Pharmacol. 53 (2002) 503–514. [2] L.J. Ignarro, C. Napoli, Novel features of nitric oxide, endothelial nitric oxide synthase, and atherosclerosis, Curr. Atheroscler. Rep. 6 (2004) 281–287. [3] I.B. Wilkinson, S.S. Franklin, J.R. Cockcroft, Nitric oxide and the regulation of large artery stiVness: from physiology to pharmacology, Hypertension 44 (2004) 112–116. [4] A. Lerman, A.M. Zeiher, Endothelial function: cardiac events, Circulation 111 (2005) 363–368. [5] P.O. Bonetti, L.O. Lerman, A. Lerman, Endothelial dysfunction: a marker of atherosclerotic risk, Arterioscler. Thromb. Vasc. Biol. 23 (2003) 168–175. [6] H. Drexler, B. Hornig, Endothelial dysfunction in human disease, J. Mol. Cell Cardiol. 31 (1999) 51–60. [7] M. Hashimoto, K. Kozaki, M. Eto, M. Akishita, J. Ako, K. Iijima, et al., Association of coronary risk factors and endothelium‐dependent flow‐mediated dilatation of the brachial artery, Hypertens. Res. 23 (2000) 233–238. [8] H. Moreno, Jr., S. Chalon, A. Urae, O. Tangphao, A.K. Abiose, B.B. HoVman, et al., Endothelial dysfunction in human hand veins is rapidly reversible after smoking cessation, Am. J. Physiol. 275 (1998) H1040–H1045. [9] W.H. Leung, C.P. Lau, C.K. Wong, Beneficial eVect of cholesterol‐lowering therapy on coronary endothelium‐dependent relaxation in hypercholesterolaemic patients, Lancet 341 (1993) 1496–1500. [10] B. Braam, M.C. Verhaar, Understanding eNOS for pharmacological modulation of endothelial function: a translational view, Curr. Pharm. Des. 13 (2007) 1727–1740. [11] C.D. Searles, Transcriptional and posttranscriptional regulation of endothelial nitric oxide synthase expression, Am. J. Physiol. Cell Physiol. 291 (2006) C803–C816. [12] D.M. Dudzinski, J. Igarashi, D. Greif, T. Michel, The regulation and pharmacology of endothelial nitric oxide synthase, Annu. Rev. Pharmacol. Toxicol. 46 (2006) 235–276. [13] N.J. Alp, K.M. Channon, Regulation of endothelial nitric oxide synthase by tetrahydrobiopterin in vascular disease, Arterioscler. Thromb. Vasc. Biol. 24 (2004) 413–420. [14] R.D. Minshall, W.C. Sessa, R.V. Stan, R.G. Anderson, A.B. Malik, Caveolin regulation of endothelial function, Am. J. Physiol. Lung Cell Mol. Physiol. 285 (2003) L1179–L1183. [15] Y.C. Boo, H. Jo, Flow‐dependent regulation of endothelial nitric oxide synthase: role of protein kinases, Am. J. Physiol. Cell Physiol. 285 (2003) C499–C508.

90

ARDUINO A. MANGONI

[16] H. Li, T. Wallerath, T. Munzel, U. Forstermann, Regulation of endothelial‐type NO synthase expression in pathophysiology and in response to drugs, Nitric Oxide 7 (2002) 149–164. [17] H. Li, T. Wallerath, U. Forstermann, Physiological mechanisms regulating the expression of endothelial‐type NO synthase, Nitric Oxide 7 (2002) 132–147. [18] T. Wallerath, G. Deckert, T. Ternes, H. Anderson, H. Li, K. Witte, et al., Resveratrol, a polyphenolic phytoalexin present in red wine, enhances expression and activity of endothelial nitric oxide synthase, Circulation 106 (2002) 1652–1658. [19] R. Govers, T.J. Rabelink, Cellular regulation of endothelial nitric oxide synthase, Am. J. Physiol. Renal. Physiol. 280 (2001) F193–F206. [20] P. Vallance, J. Leiper, Cardiovascular biology of the asymmetric dimethylarginine:dimethylarginine dimethylaminohydrolase pathway, Arterioscler. Thromb. Vasc. Biol. 24 (2004) 1023–1030. [21] J.P. Cooke, ADMA: its role in vascular disease, Vasc. Med. 10 (Suppl. 1) (2005) S11–S17. [22] J.T. Kielstein, F. Donnerstag, S. Gasper, J. Menne, A. Kielstein, J. Martens‐LobenhoVer, et al., ADMA increases arterial stiVness and decreases cerebral blood flow in humans, Stroke 37 (2006) 2024–2029. [23] G.G. Zhang, Y.P. Bai, M.F. Chen, R.Z. Shi, D.J. Jiang, Q.M. Fu, et al., Asymmetric dimethylarginine induces TNF‐alpha production via ROS/NF‐kappaB dependent pathway in human monocytic cells and the inhibitory eVect of reinioside C, Vascul. Pharmacol. 48 (2008) 115–121. [24] R. Schnabel, S. Blankenberg, E. Lubos, K.J. Lackner, H.J. Rupprecht, C. Espinola‐Klein, et al., Asymmetric dimethylarginine and the risk of cardiovascular events and death in patients with coronary artery disease: results from the AtheroGene Study, Circ. Res. 97 (2005) e53–e59. [25] C. Zoccali, S. Bode‐Boger, F. Mallamaci, F. Benedetto, G. Tripepi, L. Malatino, et al., Plasma concentration of asymmetrical dimethylarginine and mortality in patients with end‐ stage renal disease: a prospective study, Lancet 358 (2001) 2113–2117. [26] F. Mittermayer, K. Krzyzanowska, M. Exner, W. Mlekusch, J. Amighi, S. Sabeti, et al., Asymmetric dimethylarginine predicts major adverse cardiovascular events in patients with advanced peripheral artery disease, Arterioscler. Thromb. Vasc. Biol. 26 (2006) 2536–2540. [27] A. Meinitzer, U. Seelhorst, B. Wellnitz, G. Halwachs‐Baumann, B.O. Boehm, B.R. Winkelmann, et al., Asymmetrical dimethylarginine independently predicts total and cardiovascular mortality in individuals with angiographic coronary artery disease (the Ludwigshafen Risk and Cardiovascular Health study), Clin. Chem. 53 (2007) 273–283. [28] C. Duckelmann, F. Mittermayer, D.G. Haider, J. Altenberger, J. Eichinger, M. Wolzt, Asymmetric dimethylarginine enhances cardiovascular risk prediction in patients with chronic heart failure, Arterioscler. Thromb. Vasc. Biol. 27 (2007) 2037–2042. [29] K. Krzyzanowska, F. Mittermayer, M. Wolzt, G. Schernthaner, Asymmetric dimethylarginine predicts cardiovascular events in patients with type 2 diabetes, Diabetes Care 30 (2007) 1834–1839. [30] M. Zeller, C. Korandji, J.C. Guilland, P. Sicard, C. Vergely, L. Lorgis, et al., Impact of asymmetric dimethylarginine on mortality after acute myocardial infarction, Arterioscler. Thromb. Vasc. Biol. 28 (2008) 954–960. [31] S.J. Nicholls, Z. Wang, R. Koeth, B. Levison, B. DelFraino, V. Dzavik, et al., Metabolic profiling of arginine and nitric oxide pathways predicts hemodynamic abnormalities and mortality in patients with cardiogenic shock after acute myocardial infarction, Circulation 116 (2007) 2315–2324. [32] W.H. Wilson Tang, W. Tong, K. Shrestha, Z. Wang, B.S. Levison, B. DelFraino, et al., DiVerential eVects of arginine methylation on diastolic dysfunction and disease progression in patients with chronic systolic heart failure, Eur. Heart J. 29 (2008) 2506–2513.

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

91

[33] J.T. Kielstein, S.M. Bode‐Boger, G. Hesse, J. Martens‐LobenhoVer, A. Takacs, D. Fliser, et al., Asymmetrical dimethylarginine in idiopathic pulmonary arterial hypertension, Arterioscler. Thromb. Vasc. Biol. 25 (2005) 1414–1418. [34] S. Clarke, Protein methylation, Curr. Opin. Cell Biol. 5 (1993) 977–983. [35] Y. Wang, J. Wysocka, J. Sayegh, Y.H. Lee, J.R. Perlin, L. Leonelli, et al., Human PAD4 regulates histone arginine methylation levels via demethylimination, Science 306 (2004) 279–283. [36] G.L. Cuthbert, S. Daujat, A.W. Snowden, H. Erdjument‐Bromage, T. Hagiwara, M. Yamada, et al., Histone deimination antagonizes arginine methylation, Cell 118 (2004) 545–553. [37] J.A. Duerre, C.H. Miller, G.G. Reams, Metabolism of S‐adenosyl‐L‐homocysteine in vivo by the rat, J. Biol. Chem. 244 (1969) 107–111. [38] Y. Kakimoto, S. Akazawa, Isolation and identification of N‐G, N‐G‐ and N‐G, N0 ‐G‐ dimethyl‐arginine, N‐epsilon‐mono‐, di‐, and trimethyllysine, and glucosylgalactosyl‐ and galactosyl‐delta‐hydroxylysine from human urine, J. Biol. Chem. 245 (1970) 5751–5758. [39] R.H. Boger, K. Sydow, J. Borlak, T. Thum, H. Lenzen, B. Schubert, et al., LDL cholesterol upregulates synthesis of asymmetrical dimethylarginine in human endothelial cells: involvement of S‐adenosylmethionine‐dependent methyltransferases, Circ. Res. 87 (2000) 99–105. [40] T. Teerlink, ADMA metabolism and clearance, Vasc. Med. 10 (Suppl. 1) (2005) S73–S81. [41] J. Mehlhase, T. Grune, Proteolytic response to oxidative stress in mammalian cells, Biol. Chem. 383 (2002) 559–567. [42] T. Ogawa, M. Kimoto, K. Sasaoka, Purification and properties of a new enzyme, NG, NG‐ dimethylarginine dimethylaminohydrolase, from rat kidney, J. Biol. Chem. 264 (1989) 10205–10209. [43] J.M. Leiper, M.J. Santa, A. Chubb, R.J. MacAllister, I.G. Charles, G.S. Whitley, et al., Identification of two human dimethylarginine dimethylaminohydrolases with distinct tissue distributions and homology with microbial arginine deiminases, Biochem. J. 343 (Pt 1) (1999) 209–214. [44] R.J. Nijveldt, T. Teerlink, C. van Guldener, H.A. Prins, A.A. van Lambalgen, C.D. Stehouwer, et al., Handling of asymmetrical dimethylarginine and symmetrical dimethylarginine by the rat kidney under basal conditions and during endotoxaemia, Nephrol. Dial. Transplant. 18 (2003) 2542–2550. [45] R.J. Nijveldt, T. Teerlink, M.P. Siroen, A.A. van Lambalgen, J.A. Rauwerda, P.A. van Leeuwen, The liver is an important organ in the metabolism of asymmetrical dimethylarginine (ADMA), Clin. Nutr. 22 (2003) 17–22. [46] M. Kimoto, H. Tsuji, T. Ogawa, K. Sasaoka, Detection of NG, NG‐dimethylarginine dimethylaminohydrolase in the nitric oxide‐generating systems of rats using monoclonal antibody, Arch. Biochem. Biophys. 300 (1993) 657–662. [47] C.T. Tran, M.F. Fox, P. Vallance, J.M. Leiper, Chromosomal localization, gene structure, and expression pattern of DDAH1: comparison with DDAH2 and implications for evolutionary origins, Genomics 68 (2000) 101–105. [48] H. Dayoub, V. Achan, S. Adimoolam, J. Jacobi, M.C. Stuehlinger, B.Y. Wang, et al., Dimethylarginine dimethylaminohydrolase regulates nitric oxide synthesis: genetic and physiological evidence, Circulation 108 (2003) 3042–3047. [49] H. Dayoub, R. Rodionov, C. Lynch, J.P. Cooke, E. Arning, T. Bottiglieri, et al., Overexpression of dimethylarginine dimethylaminohydrolase inhibits asymmetric dimethylarginine‐induced endothelial dysfunction in the cerebral circulation, Stroke 39 (2008) 180–184. [50] J. Murray‐Rust, J. Leiper, M. McAlister, J. Phelan, S. Tilley, M.J. Santa, et al., Structural insights into the hydrolysis of cellular nitric oxide synthase inhibitors by dimethylarginine dimethylaminohydrolase, Nat. Struct. Biol. 8 (2001) 679–683.

92

ARDUINO A. MANGONI

[51] J. Leiper, J. Murray‐Rust, N. McDonald, P. Vallance, S‐nitrosylation of dimethylarginine dimethylaminohydrolase regulates enzyme activity: further interactions between nitric oxide synthase and dimethylarginine dimethylaminohydrolase, Proc. Natl. Acad. Sci. USA 99 (2002) 13527–13532. [52] E.I. Closs, F.Z. Basha, A. Habermeier, U. Forstermann, Interference of L‐arginine analogues with L‐arginine transport mediated by the yþ carrier hCAT‐2B, Nitric Oxide 1 (1997) 65–73. [53] A. Meinitzer, M. Puchinger, B.M. Winklhofer‐Roob, E. Rock, J. Ribalta, J.M. Roob, et al., Reference values for plasma concentrations of asymmetrical dimethylarginine (ADMA) and other arginine metabolites in men after validation of a chromatographic method, Clin. Chim. Acta 384 (2007) 141–148. [54] G.G. Hov, E. Sagen, A. Bigonah, A. Asberg, Health‐associated reference values for arginine, asymmetric dimethylarginine (ADMA) and symmetric dimethylarginine (SDMA) measured with high‐performance liquid chromatography, Scand. J. Clin. Lab. Invest. 67 (2007) 868–876. [55] F. Schulze, R. Maas, R. Freese, E. Schwedhelm, E. Silberhorn, R.H. Boger, Determination of a reference value for N(G), N(G)‐dimethyl‐L‐arginine in 500 subjects, Eur. J. Clin. Invest. 35 (2005) 622–626. [56] E. Schwedhelm, Quantification of ADMA: analytical approaches, Vasc. Med. 10 (Suppl 1) (2005) S89–S95. [57] R.J. Nijveldt, P.A. van Leeuwen, C. van Guldener, C.D. Stehouwer, J.A. Rauwerda, T. Teerlink, Net renal extraction of asymmetrical (ADMA) and symmetrical (SDMA) dimethylarginine in fasting humans, Nephrol. Dial. Transplant. 17 (2002) 1999–2002. [58] P. Vallance, A. Leone, A. Calver, J. Collier, S. Moncada, Accumulation of an endogenous inhibitor of nitric oxide synthesis in chronic renal failure, Lancet 339 (1992) 572–575. [59] M.P. Siroen, J.R. van der Sijp, T. Teerlink, C. van Schaik, P.A. van Leeuwen, The human liver clears both asymmetric and symmetric dimethylarginine, Hepatology 41 (2005) 559–565. [60] Y. Hattori, K. Kasai, S.S. Gross, Cationic amino acid transporter gene expression in cultured vascular smooth muscle cells and in rats, Am. J. Physiol. 276 (1999) H2020–H2028. [61] R. MacAllister, P. Vallance, Nitric oxide in essential and renal hypertension, J. Am. Soc. Nephrol. 5 (1994) 1057–1065. [62] R.J. MacAllister, G.S. Whitley, P. Vallance, EVects of guanidino and uremic compounds on nitric oxide pathways, Kidney Int. 45 (1994) 737–742. [63] S.M. Gardiner, P.A. Kemp, T. Bennett, R.M. Palmer, S. Moncada, Regional and cardiac haemodynamic eVects of NG, NG, dimethyl‐L‐arginine and their reversibility by vasodilators in conscious rats, Br. J. Pharmacol. 110 (1993) 1457–1464. [64] J.S. Jin, L.G. D’Alecy, Central and peripheral eVects of asymmetric dimethylarginine, an endogenous nitric oxide synthetase inhibitor, J. Cardiovasc. Pharmacol. 28 (1996) 439–446. [65] J.T. Kielstein, B. Impraim, S. Simmel, S.M. Bode‐Boger, D. Tsikas, J.C. Frolich, et al., Cardiovascular eVects of systemic nitric oxide synthase inhibition with asymmetrical dimethylarginine in humans, Circulation 109 (2004) 172–177. [66] J.T. Kielstein, S. Simmel, S.M. Bode‐Boger, H.J. Roth, H. Schmidt‐Gayk, H. Haller, et al., Subpressor dose asymmetric dimethylarginine modulates renal function in humans through nitric oxide synthase inhibition, Kidney Blood Press. Res. 27 (2004) 143–147. [67] S. Pullamsetti, L. Kiss, H.A. Ghofrani, R. Voswinckel, P. Haredza, W. Klepetko, et al., Increased levels and reduced catabolism of asymmetric and symmetric dimethylarginines in pulmonary hypertension, FASEB J. 19 (2005) 1175–1177. [68] M. Carlstrom, R.D. Brown, J. Edlund, J. Sallstrom, E. Larsson, T. Teerlink, et al., Role of nitric oxide deficiency in the development of hypertension in hydronephrotic animals, Am. J. Physiol. Renal. Physiol. 294 (2008) F362–F370.

SYMMETRIC DIMETHYLARGININE AND VASCULAR DISEASE

93

[69] J.T. Kielstein, S.M. Bode‐Boger, J.C. Frolich, E. Ritz, H. Haller, D. Fliser, Asymmetric dimethylarginine, blood pressure, and renal perfusion in elderly subjects, Circulation 107 (2003) 1891–1895. [70] M.E. Safar, B.I. Levy, H. Struijker‐Boudier, Current perspectives on arterial stiVness and pulse pressure in hypertension and cardiovascular diseases, Circulation 107 (2003) 2864–2869. [71] M. Chen, Y. Li, T. Yang, Y. Wang, Y. Bai, X. Xie, ADMA induces monocyte adhesion via activation of chemokine receptors in cultured THP‐1 cells, Cytokine 43 (2008) 149–159. [72] H.B. Xiao, Z.C. Yang, S.J. Jia, N.S. Li, D.J. Jiang, X.H. Zhang, et al., EVect of asymmetric dimethylarginine on atherogenesis and erythrocyte deformability in apolipoprotein E deficient mice, Life Sci. 81 (2007) 1–7. [73] T. Leong, D. Zylberstein, I. Graham, L. Lissner, D. Ward, J. Fogarty, et al., Asymmetric dimethylarginine independently predicts fatal and nonfatal myocardial infarction and stroke in women: 24‐year follow‐up of the population study of women in Gothenburg, Arterioscler. Thromb. Vasc. Biol. 28 (2008) 961–967. [74] S. Kiechl, T. Lee, P. Santer, G. Thompson, S. Tsimikas, G. Egger, et al., Asymmetric and symmetric dimethylarginines are of similar predictive value for cardiovascular risk in the general population, Atherosclerosis (2009) (in press). [75] M. Lajer, L. Tarnow, A. Jorsal, T. Teerlink, H.H. Parving, P. Rossing, Plasma concentration of asymmetric dimethylarginine (ADMA) predicts cardiovascular morbidity and mortality in type 1 diabetic patients with diabetic nephropathy, Diabetes Care 31 (2008) 747–752. [76] S.M. Bode‐Boger, F. Scalera, J.T. Kielstein, J. Martens‐LobenhoVer, G. Breithardt, M. Fobker, et al., Symmetrical dimethylarginine: a new combined parameter for renal function and extent of coronary artery disease, J. Am. Soc. Nephrol. 17 (2006) 1128–1134. [77] A.A. Mangoni, C.L. Hewitson, R.J. Woodman, M.J. Whiting, B. McAteer‐Carr, J.A. Barbara, Symmetric dimethylarginine is an independent predictor of intradialytic hypotension, Am. J. Hypertens. 21 (2008) 955–959. [78] E. Schepers, G. Glorieux, A. Dhondt, L. Leybaert, R. Vanholder, Role of symmetric dimethylarginine in vascular damage by increasing ROS via store‐operated calcium influx in monocytes, Nephrol. Dial. Transplant. 24 (2009) 1429–1435. [79] J.T. Kielstein, S.R. Salpeter, S.M. Bode‐Boeger, J.P. Cooke, D. Fliser, Symmetric dimethylarginine (SDMA) as endogenous marker of renal function—a meta‐analysis, Nephrol. Dial. Transplant. 21 (2006) 2446–2451. [80] E.R. Brooks, C.B. Langman, S. Wang, H.E. Price, A.L. Hodges, L. Darling, et al., Methylated arginine derivatives in children and adolescents with chronic kidney disease, Pediatr. Nephrol. 24 (2009) 129–134. [81] D. Fliser, F. Kronenberg, J.T. Kielstein, C. Morath, S.M. Bode‐Boger, H. Haller, et al., Asymmetric dimethylarginine and progression of chronic kidney disease: the mild to moderate kidney disease study, J. Am. Soc. Nephrol. 16 (2005) 2456–2461. [82] J. Wang, A.S. Sim, X.L. Wang, C. Salonikas, D. Naidoo, D.E. Wilcken, Relations between plasma asymmetric dimethylarginine (ADMA) and risk factors for coronary disease, Atherosclerosis 184 (2006) 383–388. [83] M. Al Banchaabouchi, B. Marescau, I. Possemiers, R. D’Hooge, O. Levillain, P.P. De Deyn, NG,NG‐dimethylarginine and NG,NG‐dimethylarginine in renal insuYciency, Pflugers Arch. 439 (2000) 524–531. [84] B. Marescau, G. Nagels, I. Possemiers, M.E. De Broe, I. Becaus, J.M. Billiouw, et al., Guanidino compounds in serum and urine of nondialyzed patients with chronic renal insuYciency, Metabolism 46 (1997) 1024–1031.

94

ARDUINO A. MANGONI

[85] P.W. Nanayakkara, T. Teerlink, C.D. Stehouwer, D. Allajar, A. Spijkerman, C. Schalkwijk, et al., Plasma asymmetric dimethylarginine (ADMA) concentration is independently associated with carotid intima‐media thickness and plasma soluble vascular cell adhesion molecule‐1 (sVCAM‐1) concentration in patients with mild‐to‐moderate renal failure, Kidney Int. 68 (2005) 2230–2236. [86] J.T. Kielstein, R.H. Boger, S.M. Bode‐Boger, J.C. Frolich, H. Haller, E. Ritz, et al., Marked increase of asymmetric dimethylarginine in patients with incipient primary chronic renal disease, J. Am. Soc. Nephrol. 13 (2002) 170–176. [87] C.D. Goonasekera, D.D. Rees, P. Woolard, A. Frend, V. Shah, M.J. Dillon, Nitric oxide synthase inhibitors and hypertension in children and adolescents, J. Hypertens. 15 (1997) 901–909. [88] A.S. Levey, J.P. Bosch, J.B. Lewis, T. Greene, N. Rogers, D. Roth, A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group, Ann. Intern. Med. 130 (1999) 461–470. [89] S. Herget‐Rosenthal, A. Bokenkamp, W. Hofmann, How to estimate GFR‐serum creatinine, serum cystatin C or equations? Clin. Biochem. 40 (2007) 153–161. [90] T. Almquist, L. Forslund, N. Rehnqvist, P. Hjemdahl, Prognostic implications of renal dysfunction in patients with stable angina pectoris, J. Intern. Med. 260 (2006) 537–544. [91] N.S. Anavekar, J.J. McMurray, E.J. Velazquez, S.D. Solomon, L. Kober, J.L. Rouleau, et al., Relation between renal dysfunction and cardiovascular outcomes after myocardial infarction, N. Engl. J. Med. 351 (2004) 1285–1295. [92] J.F. Mann, H.C. Gerstein, J. Pogue, J. Bosch, S. Yusuf, Renal insuYciency as a predictor of cardiovascular outcomes and the impact of ramipril: the HOPE randomized trial, Ann. Intern. Med. 134 (2001) 629–636. [93] D.S. Raj, B. Vincent, K. Simpson, E. Sato, K.L. Jones, T.C. Welbourne, et al., Hemodynamic changes during hemodialysis: role of nitric oxide and endothelin, Kidney Int. 61 (2002) 697–704. [94] C.L. Hewitson, M.J. Whiting, J.A. Barbara, A.A. Mangoni, Acute eVects of haemodialysis on biochemical modulators of endothelial function, J. Intern. Med. 262 (2007) 571–580. [95] M.W. Chapleau, J.T. Cunningham, M.J. Sullivan, R.E. Wachtel, F.M. Abboud, Structural versus functional modulation of the arterial baroreflex, Hypertension 26 (1995) 341–347. [96] S. Aslam, T. Santha, A. Leone, C. Wilcox, EVects of amlodipine and valsartan on oxidative stress and plasma methylarginines in end‐stage renal disease patients on hemodialysis, Kidney Int. 70 (2006) 2109–2115.

ADVANCES IN CLINICAL CHEMISTRY, VOL. 48

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY Ferruccio Santini,*,1 Margherita Maffei,† Caterina Pelosini,* Guido Salvetti,* Giovanna Scartabelli,* and Aldo Pinchera* *Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy † Dulbecco Telethon Institute at the Department of Endocrinology and Kidney, University Hospital of Pisa, 56124 Pisa, Italy

1. 2. 3. 4. 5. 6. 7. 8. 9.

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Melanocortin System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The MC4R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mutations in the MC4R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Functional Alterations of MC4R . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clinical Phenotype of MC4R‐Mutated Individuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Implications of MC4R Mutations in the Clinical Management of Obesity . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

95 96 97 99 100 102 102 103 103 104

1. Abstract The current alarming spread of obesity in many parts of the world is caused by a sudden environmental shift characterized by replacement of a frugal diet with low cost, energy dense food, and little requests for physical activity during work and leisure time. Yet, not all people exposed to an obesogenic environment become obese, and individual diVerences in the propensity to gain weight as well as the occurrence of diVerent obese phenotypes within the same environment indicate that the genetic heritage in this 1

Corresponding author: Ferruccio Santini, e-mail: [email protected] 95

0065-2423/09 $35.00 DOI: 10.1016/S0065-2423(09)48004-1

Copyright 2009, Elsevier Inc. All rights reserved.

96

SANTINI ET AL.

regard is significant and heterogeneous. The central melanocortin circuit has received much attention during the past decade, since mutations of genes expressing some key molecules in neurons of this system were discovered, which may cause monogenic forms of obesity in animals and humans. Within the arcuate nucleus of the hypothalamus the prohormone proopiomelanocortin is posttranslationally cleaved to produce the a‐melanocyte stimulating hormone, a peptide with anorexigenic eVects upon activation of the melanocortin‐4 receptor (MC4R) expressed on the surface of target neurons. Studies regarding the frequency of MC4R mutations associated with human obesity have provided variable results (up to 6% of obese subjects). Various findings suggest an oligogenic and codominant mode of inheritance for MC4R deficiency, with modulation of expressivity and penetrance of the phenotype. The yield of MC4R testing in clinical diagnosis and treatment of obesity is at present undefined since the relatively low prevalence of MC4R pathogenic variants in the general population, along with the high number of sequence variants, has so far compromised the devising of systematic controlled intervention studies. Hopefully, in the future, MC4R testing will have practical implications for the development of new mechanism‐based therapy of obesity as well as for the design of specific and more eVective protocols, based on lifestyle intervention and current pharmacological or surgical approaches, for management of obesity in MC4R‐mutated individuals.

2. Introduction Obesity is a heterogeneous condition that develops when excessive lipid accumulation occurs at various sites of the body, including but not limited to the adipose tissue [1]. The amount and the distribution of lipid depots determine the obesity phenotype and the consequences on the health status [2]. Lipid accumulation derives from an excess of energy intake over expenditure, and originally represented a positive survival trait that enabled the building of fuel storages whenever the food supply was suYcient to allow it. The current alarming spread of obesity in many parts of the world, including poorer emerging countries, is caused by a sudden environmental shift characterized by replacement of a frugal diet with low cost, energy dense food, and little requests for physical activity during work and leisure time. This has made the natural human thrifty metabolism inadequate to cope with a new set of external conditions, and obesity can then be viewed as the predictable response of our biological system, evolved in specific ecological niches, to the abundant energy oVer typical of modern societies [3, 4]. Yet, not all people exposed to an obesogenic environment become obese, and individual diVerences in the propensity to gain weight as well as the

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY

97

occurrence of diVerent obese phenotypes within the same environment indicate that the genetic heritage in this regard is significant and heterogeneous. Much evidence has been provided by family studies clearly demonstrating that a substantial background of susceptibility to obesity is determined by genetic factors [5]. Although these concepts have been known for a long time, the genetics of obesity was born in recent years. The discovery of leptin in 1994 [6] provided the first evidence for a hormonal signal that originates from adipocytes and acts on specific neurons within well‐defined centers, by relaying information regarding the level of energy stored in the adipose tissue. Serum leptin concentrations are proportional to the total fat mass [7] and leptin/leptin receptor deficiency results in severe obesity [5]. These discoveries have produced a large body of research that allowed the identification of several critical pathways in the neuroendocrine regulation of energy homeostasis. Among them, the central melanocortin circuit has received much attention during the past decade, since mutations of genes expressing some key molecules in neurons of this system were discovered, which may cause monogenic forms of obesity in animals and humans. The literature concerning mutations of the melanocortin‐4 receptor (MC4R) in humans is herein reviewed, trying to call attention on those aspects that might have a clinical relevance in the management of obesity.

3. The Melanocortin System The central melanocortin system is a pivotal mediator of leptin eVects on the control of body weight and energy expenditure [8]. The central melanocortin system includes two distinct populations of neurons in the arcuate nucleus of the hypothalamus—the proopiomelanocortin (POMC) and cocaine‐ and amphetamine‐related transcript prepropeptide (CARTPT) expressing neurons and the agouti‐related peptide (AgRP) and neuropeptide Y (NPY) expressing neurons—and their downstream target neurons located in various hypothalamic nuclei and expressing the melanocortin receptors (Fig. 1). POMC is a prohormone that is synthesized in various tissues and yields several biologically active peptides with diVering functions, depending upon tissue‐specific endoproteases (proconvertases) [9]. Within the arcuate nucleus of the hypothalamus POMC is posttranslationally cleaved to produce the a‐melanocyte stimulating hormone (a‐MSH), a peptide with anorexigenic eVects upon activation of the melanocortin‐4 receptor (MC4R) expressed on the surface of target neurons. Leptin stimulates POMC expression and eventually results in MC4R stimulation. Like leptin deficiency, deficiency in the POMC–MC4R system results in weight gain [5].

SANTINI ET AL.

98

PC Other

a MSH

POMC

+ +

MC4R

AGRP

ARC

PVN

BDNF/TrkB Leptin

?

?

Regulation of energy balance (food intake-energy expenditure) Adipose tissue FIG. 1. Schematic representation of the melanocortin system. Leptin, an adipokine produced by the adipose tissue, stimulates proopiomelanocortin (POMC) expression in the arcuate nucleus (ARC) of the hypothalamus. POMC is posttranslationally cleaved by a specific proconvertase (PC) to produce the a‐melanocyte stimulating hormone (a‐MSH), a peptide with anorexigenic eVects upon activation of the melanocortin receptor 4 (MC4R), expressed on the surface of target neurons in the paraventricular nucleus (PVN). At the same time leptin reduces the expression of AgRP, an endogenous molecule that antagonizes the action of a‐MSH at the MC4R, thus amplifying the anorexigenic eVect due to stimulation of POMC neurons. Brain‐ derived neurotrophic factor (BDNF), its tyrosine kinase receptor B (TrkB) and possibly other eVectors mediate the melanocortinergic signaling in the control of energy balance.

AgRP is an endogenous molecule that antagonizes the action of a‐MSH at the MC4R, thereby inducing an orexigenic eVect [8]. Furthermore, evidence has been provided that MC4R exhibits a constitutive activity on which AgRP acts as an inverse agonist [10, 11]. Leptin reduces the expression of AgRP, thus amplifying the anorexigenic eVect due to stimulation of POMC neurons. Aberrant AgRP overexpression causes obesity while AgRP deficiency is associated with hypophagia and leanness in animals [12]. Thus, regulation of body weight and energy homeostasis is influenced by the interplay of MC4R agonists or antagonists within the hypothalamic network, and single gene defects altering these neuroanatomical relationships may produce obesity. Although the melanocortin system is regarded as a main mediator of leptin signaling, there is evidence that the two pathways may have some independent and additive eVects [13, 14], and additional hormones (e.g., ghrelin) may intervene to modulate the activity of the system.

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY

99

Among the five known melanocortin receptors, the melanocortin‐3 receptor (MC3R) is the only other expressed at various sites of the central nervous system, but its relevance in the pathogenesis of human obesity is still unclear [15–17]. It has been suggested that MC3R is part of a feedback loop that negatively regulates the anorexic tone of POMC expressing neurons [18, 19]. A recent study indicates that MC3R is required for expression of anticipatory patterns of activity and wakefulness during periods of limited nutrient availability [20].

4. The MC4R The MC4R is a G protein‐coupled seven‐transmembrane segment receptor that shows high structural and functional conservation throughout vertebrate evolution [21]. The MC4R is a member of the family A (rhodopsin/b2 adrenergic receptor‐like receptors) [22, 23], and it is encoded by a single exon gene localized on human chromosome 18q22. The seven‐transmembrane‐spanning a‐helical segments are connected by alternating intra‐ and extracellular loops, with the amino terminus located on the extracellular side and the carboxy terminus on the intracellular side. Upon activation by specific agonists the receptor recruits the activity of intracellular heterotrimeric G proteins and stimulates cAMP production (Fig. 2). The third intracellular loop plays an essential function in G protein coupling specificity and in the maintenance of an optimal constitutive activity of the receptor [24, 25]. This constitutive activity, in addition to the agonist/antagonist mediated regulation of MC4R, might be important for preservation of the most favorable energetic balance. The MC4R is largely expressed in the central nervous system, and it has been found in many sites involved in autonomic and endocrine functions, such as the paraventricular nucleus of the hypothalamus, the dorsal motor nucleus of the vagus and the raphe [26]. Several lines of evidence suggest an important role for melanocortin signaling within the brainstem, beside that played within the hypothalamic network, and a remarkable degree of coordination between the brainstem and hypothalamic melanocortin systems for control of hunger/satiety signals [27]. Although a‐MSH has been classically considered the endogenous stimulating ligand of MC4R, recent studies indicate that b‐MSH could also be involved in the central regulation of energy balance and body weight by its interaction with MC4R [28–30]. The downstream eVectors of MC4R activation are largely unknown. Accumulating evidence suggests that brain‐derived neurotrophic factor and its tyrosine kinase receptor B could be important mediators of the melanocortinergic signaling in the control of energy balance [31–33].

SANTINI ET AL.

100

N-terminus

AC a

a g

b

ATP GTP

GDP

C-terminus cAMP

FIG. 2. Schematic representation of the melanocortin 4 receptor (MC4R). Like other G protein‐coupled receptors, the MC4R is composed of seven transmembrane domains connected by alternating intra‐ and extracellular loops, with the amino terminus located on the extracellular side and the carboxy terminus on the intracellular side. Upon activation by specific agonists the receptor recruits the activity of intracellular heterotrimeric G proteins and stimulates cAMP production. AC: adenylate cyclase.

5. Mutations in the MC4R Mutations in MC4R associated with dominantly inherited obesity in man were first reported in 1998 [34, 35]. Mutations cosegregated with the obesity phenotype in the probands’ families, suggesting a new cause of monogenic obesity in humans. Since then, the MC4R gene has been extensively investigated, and more than 100 gene variants have been cumulatively reported by many research groups. Some of these variants were equally found in obese and lean subjects and considered polymorphisms, but recent studies indicate that at least one of them [V103I] may confer a protective role against obesity and the metabolic syndrome [36–40], thus suggesting that the MC4R should be considered as a multifaceted modulator of body weight. Studies regarding the frequency of MC4R mutations associated with obesity have provided variable and sometimes conflicting results. The highest prevalence (5–6%) has been found in subjects from United Kingdom with severe (BMI standard deviation scores of more than three), early‐onset (before 10 years of age) obesity [41, 42]. A lower prevalence has been reported in Czech (2.4%) [43], German (1.9%) [44], Finnish (1.8%) [45], French (1.7%) [46], Chinese (1.5%) [47], Asian (1.3%) [48], Spanish (1.2%) [49], Austrian (1%) [50], and Italian or Belgian (<0.5%) [51, 52] obese children and/or adolescents.

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY

101

Studies in obese adults also revealed a variable prevalence of MC4R mutations, ranging from 5.1% in Switzerland [53] to 2.6% in France [54], 2.5% in Denmark [55], 2.2% in North America [17], 1.7–2.5% in Italy [56, 57], 1.3% in China [58], 1% in Pima Indians [59], and <1% in UK [60], Japan [61], Sweden [62], Spain [63], Belgium [52], or Germany [64]. Very recently, a study in a large population of obese individuals of European origin allowed calculating a 1.8% prevalence of loss‐of‐function MC4R mutations in obese children and 1.6% in obese adults [65]. A rare mutation in the promoter region of the MC4R gene has also been described, which was associated with early‐onset obesity [66] but its relevance in the pathogenesis of common obesity remains controversial [67, 68]. A study in an unselected European British population estimated obesity‐ causing MC4R mutations at 1 in 1100 [69], making it one of the most common autosomal disorders in humans. Loss‐of‐function MC4R mutations display an autosomal dominance inherited pattern since most heterozygous individuals are obese, and haploinsuYciency of MC4R has been proposed as the pathogenic mechanism. Yet, there are several controversial findings that need to be considered when evaluating the functional relevance of naturally occurring variants of MC4R: (1) Only part (68%) of nonsynonymous mutations in MC4R results in a loss‐of‐function in vitro [65]. (2) Although the cumulative frequency of MC4R mutation is relatively high, loss‐of‐function mutations are individually infrequent. (3) Mutations entailing impaired function of the receptor can be found in lean subjects [62, 64, 65, 70]. (4) Although a specific MC4R mutation usually associates with obesity within a pedigree, the degree and age of onset of obesity will vary across and within families, and relatives harboring the same mutation as the index patient may be simply overweight or even lean [56, 65, 66, 71, 72]. (5) A decrease in obesity penetrance between children and adult carriers of MC4R mutations has been observed [65]. (6) Subjects carrying homozygous mutations exhibit a more severe phenotype as compared to the heterozygous carrier of the same mutation [41], but no diVerence in BMI could be demonstrated between heterozygous carrier of loss‐of‐function MC4R mutations from the general population as compared with homozygotes for the wild‐type allele [64]. Altogether, these findings suggest an oligogenic and codominant mode of inheritance for MC4R deficiency, with modulation of expressivity and penetrance of the phenotype [5, 65, 73]. The specific eVect of a mutation on the receptor function, the individual genetic background and concurrent environmental obesigenic factors all contribute to development of the phenotype in individuals’ carriers of MC4R mutations. In the light of these considerations, the term ‘‘major gene eVect’’ for MC4R mutations, instead of monogenic, has been proposed [44, 74].

102

SANTINI ET AL.

6. Functional Alterations of MC4R Mutations in the MC4R can aVect the receptor function by multiple mechanisms. Mutations can alter the expression and traYcking of the receptor to the cell surface, thus causing intracellular retention of the mutant protein [46]. Mutations can also alter the binding of the agonist and/or the antagonist to the receptor, its signaling properties (i.e., the ability to modulate adenylate cyclase activity in response to ligand binding) or its constitutive activity (i.e., the basal activity in the absence of the ligand). In vitro assays have been developed to characterize the functional properties of several diVerent MC4R variants and some classifications have been proposed [42, 75–77]. These studies are important to establish a causative relationship between various mutations and obesity or other clinical phenotypes, such as eating disorders, and eventually might allow the development of specific therapies aimed at recovering the receptor function (e.g., by increasing the cell‐surface expression of transport defective mutants) [78, 79].

7. Clinical Phenotype of MC4R‐Mutated Individuals Individuals carrying homozygous mutations display a more severe phenotype than the heterozygous carriers [41, 80], and a correlation between receptor dysfunction and the severity of the clinical phenotype has been observed in childhood obesity [41]. In the latter study, MC4R deficiency was characterized by early‐onset hyperphagia, accelerated linear growth, increased fat and lean mass, increased bone mineral content and hyperinsulinemia, with no evidence of defective energy expenditure, abnormal development of puberty, reduced fertility, or other major hormonal abnormalities. Some of these findings were not confirmed in later studies [43, 54, 81]. Discrepancies could be related to age, ethnic, or selection criteria of the study groups. In various studies, the obese phenotype seemed to become less prominent with age [41, 64, 65]. This could depend on a reduced impact of MC4R mutations in adulthood or to an enhanced contribution of the environment in the pathogenesis of obesity in the young generations. Prospective studies will be necessary to elucidate the age‐dependent penetrance of MC4R mutations in the onset of obesity. A gender‐related eVect of MC4R mutations on the obese phenotype has also been reported, with a greater pathogenic impact in women than in men [65, 74]. Early studies suggested that binge eating and bulimia nervosa could be a major phenotypic characteristic of subjects with a mutation in MC4R [53, 71, 82]. This association could not be confirmed in subsequent studies

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY

103

[54, 70, 83], and the relationship between MC4R mutations and eating disorders in humans remains a matter of investigation. A very recent study indicates that the melanocortin system is implicated in the control of blood pressure in humans through an insulin‐independent mechanism [84]. Both prevalence of hypertension and mean blood pressure were lower in adults with MC4R haploinsuYciency as compared with matched controls, while administration of a MC4R agonist produced a dose‐dependent increase in blood pressure. These data suggest that MC4R signaling could influence blood pressure by increasing the activity of the sympathetic nervous system.

8. Implications of MC4R Mutations in the Clinical Management of Obesity At present, there is no specific drug therapy for the treatment of MC4R deficiency. Several selective small molecules and peptide agonists have been developed [85–88]. These ligands could rescue the potency and stimulatory response of abnormally functioning MC4Rs, but therapeutic experience so far is mainly restricted to animal models [88, 89]. Furthermore, since most patients are heterozygous for MC4R mutations, it is possible that MC4R agonists might be eVective treatment for this disorder just acting on the one functional intact allele. Conversely, MC4R antagonists could be employed to treat anorectic and cachectic conditions [90]. MC4R variants could influence an individual’s response to environmental stimuli. Indeed, it has been recently shown that children with MC4R mutations leading to reduced receptor function were able to lose weight in a lifestyle intervention but had much greater diYculties to maintain this weight loss [91]. The eVect of MC4R variants on complications and treatment outcome after bariatric surgery is still controversial [92, 93].

9. Conclusions Mutations in MC4R, in concert with allelic variations in other genes involved in the regulation of body weight and energy balance, contribute to the development of nonsyndromic obesity in humans. The question has been raised whether MC4R deficiency could be part of a thrifty genotype to facilitate storage of fat under selective environmental pressure, thus promoting survival and reproductive eYciency across evolutionary time [94]. In this regard, we now know that while MC4R mutations as a whole are relatively common in obese subjects, each single variant is rare and scattered in various

104

SANTINI ET AL.

populations. Comparative studies in vertebrates [21, 95] indicate that MC4R is highly conserved, having experienced a strong negative selection throughout evolution and an apparent increase in selective constraint along the primate phylogeny. Most functionally relevant mutations in humans occur at positions that are highly conserved during evolution while the majority of those with no functional eVect are found in highly variable residues. These observations suggest that MC4R mutations are mostly deleterious, producing a detriment to an individual’s fitness, and on their way to being removed from the population. This would imply that MC4R mutations that reduce the receptor signaling are ‘‘pathogenic,’’ although the penetrance and extent of obesity depend upon a complex interaction between genotype, genetic background, environment and maybe epigenetic factors. Yet, taking into account that subjects carrier of MC4R mutations are fertile and that detrimental consequences of obesity usually occur later in life, the evolutionary constraints causing removal of MC4R mutations are not clear at this time. The yield of MC4R testing in clinical diagnosis and treatment of obesity is at present undefined [96]. Theoretically, obese or lean carriers of functionally relevant MC4R mutations might respond diVerently to various therapies. However, the relatively low prevalence of MC4R pathogenic variants in the general population, along with the high number of sequence variants, has so far compromised the devising of systematic controlled intervention studies. Hopefully, in the future, MC4R testing will have practical implications for the development of new mechanism‐based therapy of obesity as well as for the design of specific and more eVective protocols, based on lifestyle intervention and current pharmacological or surgical approaches, for management of obesity in MC4R‐mutated individuals. REFERENCES [1] J.P. Despre´s, I. Lemieux, Abdominal obesity and metabolic syndrome, Nature 444 (2006) 881–887. [2] M.D. Jensen, Role of body fat distribution and the metabolic complications of obesity, J. Clin. Endocrinol. Metab. 93 (2008) S57–S63. [3] A. Prentice, F. Webb, Future perspectives on obesity, Obes. Metab. 1 (2005) 1–13. [4] M. Siervo, J.C.K. Wells, G.A. Cizza, Evolutionary theories, psychosocial stress and the modern obesity epidemic, Obes. Metab. 4 (2008) 131–142. [5] S. Farooqi, S. O’Rahilly, Genetics of obesity in humans, Endocr. Rev. 27 (2006) 710–718. [6] Y. Zhang, R. Proenca, M. MaVei, M. Barone, L. Leopold, J.M. Friedman, Positional cloning of the mouse obese gene and its human homologue, Nature 372 (1994) 425–432. [7] M. MaVei, J. Halaas, E. Ravussin, R.E. Pratley, G.H. Lee, Y. Zhang, et al., Leptin levels in human and rodent: measurement of plasma leptin and ob RNA in obese and weight-reduced subjects, Nat. Med. 1 (1995) 1155–1161.

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY

105

[8] R.D. Cone, Anatomy and regulation of the central melanocortin system, Nat. Neurosci. 8 (2005) 571–578. [9] L.E. Pritchard, A. White, Neuropeptide processing and its impact on melanocortin pathways, Endocrinology 148 (2007) 4201–4207. [10] C. Haskell‐Luevano, E.K. Monck, Agouti‐related protein functions as an inverse agonist at a constitutively active brain melanocortin‐4 receptor, Regul. Pept. 99 (2001) 1–7. [11] W.A. Nijenhuis, J. Oosterom, R.A. Adan, AgRP(83–132) acts as an inverse agonist on the human‐melanocortin‐4 receptor, Mol. Endocrinol. 15 (2001) 164–171. [12] J.S. Flier, AgRP in energy balance: will the real AgRP please stand up? Cell Metab. 3 (2006) 83–85. [13] B.A. Boston, K.M. Blaydon, J. Varnerin, R.D. Cone, Independent and additive eVects of central POMC and leptin pathways on murine obesity, Science 278 (1997) 1641–1644. [14] J.L. Trevaskis, A.A. Butler, Double leptin and melanocortin‐4 receptor gene mutations have an additive eVect on fat mass and are associated with reduced eVects of leptin on weight loss and food intake, Endocrinology 146 (2005) 4257–4265. [15] Y.S. Lee, L.K. Poh, B.L. Kek, K.Y. Loke, The role of melanocortin 3 receptor gene in childhood obesity, Diabetes 56 (2007) 2622–2630. [16] M. Mencarelli, G.E. Walker, S. Maestrini, et al., Sporadic mutations in melanocortin receptor 3 in morbid obese individuals, Eur. J. Hum. Genet. 16 (2008) 581–586. [17] M.A. Calton, B.A. Ersoy, S. Zhang, et al., Association of functionally significant melanocortin‐4 but not melanocortin‐3 receptor mutations with severe adult obesity in a large North-American case control study, Hum. Mol. Genet. 18 (2008) 1140–1147. [18] S. Je´gou, I. Boutelet, H. Vaudry, Melanocortin‐3 receptor mRNA expression in proopiomelanocortin neurones of the rat arcuate nucleus, J. Neuroendocrinol. 12 (2000) 501–505. [19] M.A. Cowley, J.L. Smart, M. Rubinstein, et al., Leptin activates anorexigenic POMC neurons through a neural network in the arcuate nucleus, Nature 411 (2001) 480–484. [20] G.M. Sutton, D. Perez-Tilve, R. Nogueiras, et al., The melanocortin‐3 receptor is required for entrainment to meal intake, J. Neurosci. 28 (2008) 12946–12955. [21] D.A. Hughes, A. Hinney, H. Brumm, et al., Increased constraints on MC4R during primate and human evolution, Hum. Genet. 124 (2009) 633–647. [22] K.G. Mountjoy, L.S. Robbins, M.T. Mortrud, R.D. Cone, The cloning of a family of genes that encode the melanocortin receptors, Science 257 (1992) 1248–1251. [23] U. Gether, Uncovering molecular mechanisms involved in activation of G protein-coupled receptors, Endocr. Rev. 21 (2000) 90–113. [24] C.S. Kim, S.H. Lee, R.Y. Kim, et al., Identification of domains directing specificity of coupling to G-proteins for the melanocortin MC3 and MC4 receptors, J. Biol. Chem. 277 (2002) 31310–31317. [25] D.H. Kim, S.W. Shin, J.H. Baik, Role of third intracellular loop of the melanocortin 4 receptor in the regulation of constitutive activity, Biochem. Biophys. Res. Commun. 365 (2008) 439–445. [26] K.G. Mountjoy, M.T. Mortrud, M.J. Low, R.B. Simerly, R.D. Cone, Localization of the melanocortin‐4 receptor (MC4‐R) in neuroendocrine and autonomic control circuits in the brain, Mol. Endocrinol. 8 (1994) 1298–1308. [27] R.D. Cone, Studies on the physiological functions of the melanocortin system, Endocr. Rev. 27 (2006) 736–749. [28] Y.S. Lee, B.G. Challis, D.A. Thompson, et al., A POMC variant implicates beta-melanocyte-stimulating hormone in the control of human energy balance, Cell Metab. 3 (2006) 135–140.

106

SANTINI ET AL.

[29] H. Biebermann, T.R. Castan˜eda, F. van Landeghem, et al., A role for beta-melanocytestimulating hormone in human body-weight regulation, Cell Metab. 3 (2006) 141–146. [30] L.Z. Yan, H.M. Hsiung, M.L. Heiman, et al., Structure–activity relationships of beta-MSH derived melanocortin‐4 receptor peptide agonists, Curr. Top. Med. Chem. 7 (2007) 1052–1067. [31] B. Xu, E.H. Goulding, K. Zang, et al., Brain-derived neurotrophic factor regulates energy balance downstream of melanocortin‐4 receptor, Nat. Neurosci. 6 (2003) 736–742. [32] G.S. Yeo, C.C. Connie Hung, J. Rochford, et al., A de novo mutation aVecting human TrkB associated with severe obesity and developmental delay, Nat. Neurosci. 7 (2004) 1187–1189. [33] B. Bariohay, J. Roux, C. Tardivel, J. Trouslard, A. Jean, B. Lebrun, BDNF/TrkB signalling is a downstream eVector of the brainstem melanocortin system in food intake control, Endocrinology 150 (2009) 2646–2653. [34] G.S. Yeo, I.S. Farooqi, S. Aminian, D.J. Halsall, R.G. Stanhope, S. O’Rahilly, A frameshift mutation in MC4R associated with dominantly inherited human obesity, Nat. Genet. 20 (1998) 111–112. [35] C. Vaisse, K. Clement, B. Guy-Grand, P. Froguel, A frameshift mutation in human MC4R is associated with a dominant form of obesity, Nat. Genet. 20 (1998) 113–114. [36] R. Rosmond, M. Chagnon, C. Bouchard, P. Bjo¨rntorp, A missense mutation in the human melanocortin‐4 receptor gene in relation to abdominal obesity and salivary cortisol, Diabetologia 44 (2001) 1335–1338. [37] F. Geller, K. Reichwald, A. Dempfle, et al., Melanocortin‐4 receptor gene variant I103 is negatively associated with obesity, Am. J. Hum. Genet. 74 (2004) 572–581. [38] E.H. Young, N.J. Wareham, S. Farooqi, et al., The V103I polymorphism of the MC4R gene and obesity: population based studies and meta-analysis of 29 563 individuals, Int. J. Obes. (Lond) 31 (2007) 1437–1441. [39] F. Stutzmann, V. Vatin, S. Cauchi, et al., Non-synonymous polymorphisms in melanocortin‐4 receptor protect against obesity: the two facets of a Janus obesity gene, Hum. Mol. Genet. 16 (2007) 1837–1844. [40] I.M. Heid, C. Vollmert, F. Kronenberg, et al., Association of the MC4R V103I polymorphism with the metabolic syndrome: the KORA Study, Obesity (Silver Spring) 16 (2008) 369–376. [41] I.S. Farooqi, J.M. Keogh, G.S. Yeo, E.J. Lank, T. Cheetham, S. O’Rahilly, Clinical spectrum of obesity and mutations in the melanocortin 4 receptor gene, N. Engl. J. Med. 348 (2003) 1085–1095. [42] K. Tan, I.D. Pogozheva, G.S. Yeo, et al., Functional characterization and structural modeling of obesity associated mutations in the melanocortin 4 receptor, Endocrinology 150 (2009) 114–125. [43] I. Hainerova´, L.H. Larsen, B. Holst, et al., Melanocortin 4 receptor mutations in obese Czech children: studies of prevalence, phenotype development, weight reduction response, and functional analysis, J. Clin. Endocrinol. Metab. 92 (2007) 3689–3696. [44] A. Hinney, S. Hohmann, F. Geller, et al., Melanocortin‐4 receptor gene: case – control study and transmission disequilibrium test confirm that functionally relevant mutations are compatible with a major gene eVect for extreme obesity, J. Clin. Endocrinol. Metab. 88 (2003) 4258–4267. [45] K. Valli-Jaakola, M. Lipsanen-Nyman, L. Oksanen, et al., Identification and characterization of melanocortin‐4 receptor gene mutations in morbidly obese Finnish children and adults, J. Clin. Endocrinol. Metab. 89 (2004) 940–945. [46] C. Lubrano-Berthelier, E. Durand, B. Dubern, et al., Intracellular retention is a common characteristic of childhood obesity-associated MC4R mutations, Hum. Mol. Genet. 12 (2003) 145–153.

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY

107

[47] C.L. Wang, L. Liang, H.J. Wang, J.F. Fu, J. Hebebrand, A. Hinney, Several mutations in the melanocortin 4 receptor gene are associated with obesity in Chinese children and adolescents, J. Endocrinol. Invest. 29 (2006) 894–898. [48] Y.S. Lee, L.K. Poh, B.L. Kek, K.Y. Loke, Novel melanocortin 4 receptor gene mutations in severely obese children, Clin. Endocrinol. (Oxf) 68 (2008) 529–535. [49] M.C. Ochoa, C. Azcona, H. Biebermann, et al., A novel mutation Thr162Arg of the melanocortin 4 receptor gene in a Spanish children and adolescent population, Clin. Endocrinol. (Oxf) 66 (2007) 652–658. [50] E. Rettenbacher, P. Tarnow, H. Brumm, et al., A novel non-synonymous mutation in the melanocortin‐4 receptor gene (MC4R) in a 2‐year‐old Austrian girl with extreme obesity, Exp. Clin. Endocrinol. Diab. 115 (2007) 7–12. [51] E. Miraglia Del Giudice, G. Cirillo, V. Nigro, et al., Low frequency of melanocortin‐4 receptor (MC4R) mutations in a Mediterranean population with early-onset obesity, Int. J. Obes. Relat. Metab. Disord. 26 (2002) 647–651. [52] S. Beckers, I. Mertens, A. Peeters, L. Van Gaal, W. Van Hul, Screening for melanocortin‐4 receptor mutations in a cohort of Belgian morbidly obese adults and children, Int. J. Obes. (Lond) 30 (2006) 221–225. [53] R. Branson, N. Potoczna, J.G. Kral, K.U. Lentes, M.R. Hoehe, F.F. Horber, Binge eating as a major phenotype of melanocortin 4 receptor gene mutations, N. Engl. J. Med. 348 (2003) 1096–1103. [54] C. Lubrano-Berthelier, B. Dubern, J.M. Lacorte, et al., Melanocortin 4 receptor mutations in a large cohort of severely obese adults: prevalence, functional classification, genotypephenotype relationship, and lack of association with binge eating, J. Clin. Endocrinol. Metab. 91 (2006) 1811–1818. [55] L.H. Larsen, S.M. Echwald, T.I. Sørensen, T. Andersen, B.S. WulV, O. Pedersen, Prevalence of mutations and functional analyses of melanocortin 4 receptor variants identified among 750 men with juvenile-onset obesity, J. Clin. Endocrinol. Metab. 90 (2005) 219–224. [56] F. Santini, M. MaVei, G. Ceccarini, et al., Genetic screening for melanocortin‐4 receptor mutations in a cohort of Italian obese patients: description and functional characterization of a novel mutation, J. Clin. Endocrinol. Metab. 89 (2004) 904–908. [57] P. Buono, F. Pasanisi, C. Nardelli, et al., Six novel mutations in the proopiomelanocortin and melanocortin receptor 4 genes in severely obese adults living in southern Italy, Clin. Chem. 51 (2005) 1358–1364. [58] R. Rong, Y.X. Tao, B.M. Cheung, A. Xu, G.C. Cheung, K.S. Lam, Identification and functional characterization of three novel human melanocortin‐4 receptor gene variants in an obese Chinese population, Clin. Endocrinol. (Oxf ) 65 (2006) 198–205. [59] L. Ma, P.A. Tataranni, C. Bogardus, L.J. Baier, Melanocortin 4 receptor gene variation is associated with severe obesity in Pima Indians, Diabetes 53 (2004) 2696–2699. [60] T. Gotoda, J. Scott, T.J. Aitman, Molecular screening of the human melanocortin‐4 receptor gene: identification of a missense variant showing no association with obesity, plasma glucose, or insulin, Diabetologia 40 (1997) 976–979. [61] Y. Ohshiro, T. Sanke, K. Ueda, et al., Molecular scanning for mutations in the melanocortin‐4 receptor gene in obese/diabetic Japanese, Ann. Hum. Genet. 63 (1999) 483–487. [62] P. Jacobson, O. Ukkola, T. Rankinen, et al., Melanocortin 4 receptor sequence variations are seldom a cause of human obesity: the Swedish Obese Subjects, the HERITAGE Family Study, and a Memphis cohort, J. Clin. Endocrinol. Metab. 87 (2002) 4442–4446. [63] A. Marti, M.S. Corbala´n, L. Forga, J.A. Martinez, A. Hinney, J. Hebebrand, A novel nonsense mutation in the melanocortin‐4 receptor associated with obesity in a Spanish population, Int. J. Obes. Relat. Metab. Disord. 27 (2003) 385–388.

108

SANTINI ET AL.

[64] A. Hinney, T. Bettecken, P. Tarnow, et al., Prevalence, spectrum, and functional characterization of melanocortin‐4 receptor gene mutations in a representative population-based sample and obese adults from Germany, J. Clin. Endocrinol. Metab. 91 (2006) 1761–1769. [65] F. Stutzmann, K. Tan, V. Vatin, et al., Prevalence of melanocortin‐4 receptor deficiency in Europeans and their age-dependent penetrance in multigenerational pedigrees, Diabetes 57 (2008) 2511–2518. [66] K. Valli-Jaakola, J.J. Palvimo, M. Lipsanen-Nyman, et al., A two-base deletion‐439delGC in the melanocortin‐4 receptor promoter associated with early-onset obesity, Horm. Res. 66 (2006) 61–69. [67] C. Lubrano-Berthelier, M. Cavazos, C. Le StunV, et al., The human MC4R promoter: characterization and role in obesity, Diabetes 52 (2003) 2996–3000. [68] K. Cle´ment, Mutation in the MC4R gene promoter: an elusive or eVective role? Horm. Res. 66 (2006) 70–72. [69] K.K. Alharbi, E. Spanakis, K. Tan, et al., Prevalence and functionality of paucimorphic and private MC4R mutations in a large, unselected European British population, scanned by meltMADGE, Hum. Mutat. 28 (2007) 294–302. [70] Y.X. Tao, D.L. SegaloV, Functional analyses of melanocortin‐4 receptor mutations identified from patients with binge eating disorder and nonobese or obese subjects, J. Clin. Endocrinol. Metab. 90 (2005) 5632–5638. [71] M. Sina, A. Hinney, A. Ziegler, et al., Phenotypes in three pedigrees with autosomal dominant obesity caused by haploinsuYciency mutations in the melanocortin‐4 receptor gene, Am. J. Hum. Genet. 65 (1999) 1501–1507. [72] C. Vaisse, K. Clement, E. Durand, S. Hercberg, B. Guy-Grand, P. Froguel, Melanocortin‐4 receptor mutations are a frequent and heterogeneous cause of morbid obesity, J. Clin. Invest. 106 (2000) 253–262. [73] A. Oswal, G.S. Yeo, The leptin melanocortin pathway and the control of body weight: lessons from human and murine genetics, Obes. Rev. 8 (2007) 293–306. [74] A. Dempfle, A. Hinney, M. Heinzel-Gutenbrunner, et al., Large quantitative eVect of melanocortin‐4 receptor gene mutations on body mass index, J. Med. Genet. 4 (2004) 795–800. [75] Y.X. Tao, D.L. SegaloV, Functional characterization of melanocortin‐4 receptor mutations associated with childhood obesity, Endocrinology 144 (2003) 4544–4551. [76] C. Govaerts, S. Srinivasan, A. Shapiro, et al., Obesity-associated mutations in the melanocortin 4 receptor provide novel insights into its function, Peptides 26 (2005) 1909–1919. [77] R.G. MacKenzie, Obesity-associated mutations in the human melanocortin‐4 receptor gene, Peptides 27 (2006) 395–403. [78] Y.X. Tao, Molecular mechanisms of the neural melanocortin receptor dysfunction in severe early onset obesity, Mol. Cell Endocrinol. 239 (2005) 1–14. [79] H.F. Mendes, J. van der Spuy, J.P. Chapple, M.E. Cheetham, Mechanisms of cell death in rhodopsin retinitis pigmentosa: implications for therapy, Trends Mol. Med. 11 (2005) 177–185. [80] B. Dubern, S. Bisbis, H. Talbaoui, et al., Homozygous null mutation of the melanocortin‐4 receptor and severe early-onset obesity, J. Pediatr. 150 (2007) 613–617. [81] J. KrakoV, L. Ma, S. Kobes, et al., Lower metabolic rate in individuals heterozygous for either a frameshift or a functional missense MC4R variant, Diabetes 57 (2008) 3267–3272. [82] J. Hebebrand, M. Fichter, G. Gerber, et al., Genetic predisposition to obesity in bulimia nervosa: a mutation screen of the melanocortin‐4 receptor gene, Mol. Psychiatry 7 (2002) 647–651. [83] J. Hebebrand, F. Geller, A. Dempfle, et al., Binge-eating episodes are not characteristic of carriers of melanocortin‐4 receptor gene mutations, Mol. Psychiatry 9 (2004) 796–800.

MELANOCORTIN‐4 RECEPTOR MUTATIONS IN OBESITY

109

[84] J.R. Greenfield, J.W. Miller, J.M. Keogh, et al., Modulation of blood pressure by central melanocortinergic pathways, N. Engl. J. Med. 360 (2009) 44–52. [85] P.J. Emmerson, M.J. Fisher, L.Z. Yan, J.P. Mayer, Melanocortin‐4 receptor agonists for the treatment of obesity, Curr. Top. Med. Chem. 7 (2007) 1121–1130. [86] Z. Xiang, I.D. Pogozheva, N.B. Sorenson, et al., Peptide and small molecules rescue the functional activity and agonist potency of dysfunctional human melanocortin‐4 receptor polymorphisms, Biochemistry 46 (2007) 8273–8287. [87] A.P. Coll, EVects of pro-opiomelanocortin (POMC) on food intake and body weight: mechanisms and therapeutic potential? Clin. Sci. (Lond) 113 (2007) 171–182. [88] X. Tian, A.G. Switzer, S.A. Derose, et al., Discovery of orally bioavailable 1,3,4‐trisubstituted 2‐oxopiperazine‐based melanocortin‐4 receptor agonists as potential antiobesity agents, J. Med. Chem. 51 (2008) 6055–6066. [89] H.M. Hsiung, J. Hertel, X.Y. Zhang, et al., A novel and selective beta-melanocyte-stimulating hormone-derived peptide agonist for melanocortin 4 receptor potently decreased food intake and body weight gain in diet-induced obese rats, Endocrinology 146 (2005) 5257–5266. [90] J.E. Wikberg, F. Mutulis, Targeting melanocortin receptors: an approach to treat weight disorders and sexual dysfunction, Nat. Rev. Drug. Discov. 7 (2008) 307–323. [91] T. Reinehr, J. Hebebrand, S. Friedel, et al., Lifestyle intervention in obese children with variations in the melanocortin 4 receptor gene, Obesity (Silver Spring) 17 (2009) 382–389. [92] N. Potoczna, R. Branson, J.G. Kral, et al., Gene variants and binge eating as predictors of comorbidity and outcome of treatment in severe obesity, J. Gastrointest. Surg. 8 (2004) 971–981. [93] R. Peterli, T. Peters, M. von Flu¨e, M. Hoch, A.N. Eberle, Melanocortin‐4 receptor gene and complications after gastric banding, Obes. Surg. 16 (2006) 189–195. [94] R.D. Cone, HaploinsuYciency of the melanocortin‐4 receptor: part of a thrifty genotype? J. Clin. Invest. 106 (2000) 185–187. [95] C. Sta¨ubert, P. Tarnow, H. Brumm, et al., Evolutionary aspects in evaluating mutations in the melanocortin 4 receptor, Endocrinology 148 (2007) 4642–4648. [96] B.M. Kublaoui, A.R. Zinn, Editorial: MC4R mutations—weight before screening! J. Clin. Endocrinol. Metab. 91 (2006) 1671–1672.

This page intentionally left blank

ADVANCES IN CLINICAL CHEMISTRY, VOL. 48

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION Carita M. Eklund1 Department of Microbiology and Immunology, University of Tampere, Medical School, 33520 Tampere, Finland

1. 2. 3. 4.

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C‐Reactive Protein and Inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Demographic, Metabolic, and Socioeconomic Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . Proinflammatory Cytokines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. IL‐6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. IL‐1 Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. TNF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. IL‐17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Signaling Through IL Receptors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Genetic Polymorphisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

111 112 114 118 118 119 121 122 123 124 124 126

1. Abstract Low‐grade inflammation, a minor elevation in the baseline concentration of inflammatory markers such as C‐reactive protein (CRP), is nowadays recognized as an important underlying condition in many common diseases. Concentrations of CRP under 10 mg/l are called low‐grade inflammation and values above that are considered as clinically significant inflammatory states. Epidemiological studies have revealed demographic and socioeconomic factors that associate with CRP concentration; these include age, sex, birth weight, ethnicity, socioeconomic status, body mass index (BMI), fiber consumption, alcohol intake, and dietary fatty acids. At the molecular level, 1

Corresponding author: Carita M. Eklund, e-mail: [email protected] 111

0065-2423/09 $35.00 DOI: 10.1016/S0065-2423(09)48005-3

Copyright 2009, Elsevier Inc. All rights reserved.

112

CARITA M. EKLUND

production of CRP is induced by proinflammatory cytokines IL‐1, IL‐6, and IL‐17 in the liver, although extra hepatic production most likely contributes to systemic concentrations. The cytokines are produced in response to, for example, steroid hormones, thrombin, C5a, bradykinin, other cytokines, UV‐light, neuropeptides and bacterial components, such as lipopolysaccharide. Cytokines exert their biological eVects on CRP by signaling through their receptors on hepatic cells and activating diVerent kinases and phosphatases leading to translocation of various transcription factors on CRP gene promoter and production of CRP protein. Genetic polymorphisms in the interleukin genes as well as in CRP gene have been associated with minor elevation in CRP. As minor elevation in CRP is associated with both inflammatory and noninflammatory conditions, it should be noticed that the elevation might just reflect distressed or injured cells homeostasis maintenance in everyday life, rather than inflammation with classical symptoms of redness, swelling, heat, and pain.

2. C‐Reactive Protein and Inflammation CRP is classified as one of the classical acute‐phase proteins by its biological properties. It is synthesized and secreted to the blood by the liver after initiating signals from the body, for example, infection, trauma, or tissue damage, mediated by inflammatory cytokines. By structure, CRP belongs to pentraxin protein family and is part of the soluble innate immune system where it plays an important role as a pattern‐recognition molecule [1]. After binding a ligand, for example, phosphocholine [2], chromatin [3], histones [4], fibronectin [5], laminin [6], and small nuclear ribonucleoprotein particles [7], CRP binds to Fcg–receptors on phagocytic cells and mediates ligand elimination from the body. CRP is also able to activate the complement cascade by binding C1q [8]. In higher animals PhC is ubiquitous in the phospholipids of cellular membranes, and also in the circulating plasma lipoproteins, in particular low‐density lipoprotein and very low‐density lipoprotein [9]. However, in eukaryotic membranes PhC is a constituent of sphingomyelin and phosphatidylcholine, but in a form that cannot bind to CRP; the head groups of these phospholipids are normally inaccessible to CRP. Despite of this, CRP can bind these molecules on eukaryotic membranes in damaged and apoptotic cells [10–14]. Similarly complement activation resulting in membrane damage can expose sites for CRP binding [15], as well as damage due to phospholipases [16]. Although microorganisms express PhC and are likely to be important targets of CRP, the interaction of CRP with PhC in damaged membranes may even be biologically more important as clearance of host apoptotic and necrotic cells occurs via this route [17]. One of the suggested

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

113

main functions of CRP, in fact, is waste management and the prevention of autoimmunity [18]. Thus, CRP has a physiological role in homeostasis maintenance in addition to its role as an acute‐phase reactant. By definition, inflammation is a coordinated process induced by tissue damage or microbial infections [19]. Tissue damage can be sterile, that is, without infection causing microbe, like trauma or ischemia, or can be accompanied by infection. It was recently suggested that the distinct inflammatory outcomes might depend on the type of ligand recognized and the context in which the recognition takes place [20]. For example, the response to infection might require additional destruction of self‐tissue to contain the microbe, whereas sterile injury might focus more on repair [20]. However, sterile tissue damage and infection still most likely share many events in the inflammatory reaction, for example, dendritic cell activation and neutrophil and monocyte infiltration [21]. Tissue injury, both sterile and concomitant with infection, unleashes several signals; neurons release bioactive peptides in response to pain and broken cells release constitutively expressed intracellular proteins that trigger cytokine production when found in the extracellular space. If microbes are involved, their shed or secreted products are sensed through binding of their conserved molecular constituents to receptors such as complement, mannose‐binding protein and lipopolysaccharide‐binding protein, and to cell‐surface receptors such as Toll family members. All of these activate inflammatory cytokine production from immune cells. A minor elevation in the baseline level of inflammatory markers in blood is called low‐grade inflammation, where the body is constantly under very mild chronic inflammation but not to the extent of acute inflammation. However, drawing a precise line between these two is often diYcult. One of the simplest and widely used ways to measure low‐grade inflammation is to measure C‐reactive protein concentration from the blood. For practical reasons, CRP concentrations under 10 mg/l are considered as low‐grade inflammation and values above that as clinically significant inflammatory states. Low‐grade inflammation has recently caught much attention from many researchers, as it is slowly being accepted as the underlying condition in many common diseases/conditions. For example, low‐grade inflammation is implicated in obesity and in many obesity related diseases, for example, metabolic syndrome [22, 23], type 2 diabetes [24], insulin resistance syndrome [25], atherosclerosis [26, 27], and some cancers [28]. However, minor elevation in CRP levels is also associated with many apparently noninflammatory medical conditions [29]. The American Hearth Association (AHA) and the CDC have evaluated CRP as a risk assessment tool and suggested that cut points of below 1 mg/l, between 1 and 3 mg/l, and greater than 3 mg/l be used to identify these at lower, average and high relative risk, respectively, for CVD events [30]. The reference values were largely derived from white cohorts of European descent

114

CARITA M. EKLUND

[31–33]. Besides underlying many common conditions/diseases, low‐grade inflammation is also reported to predict the development of a variety of diseases in healthy people, including colon cancer, cardiovascular events, diabetes mellitus, rheumatoid arthritis, and type 2 diabetes [29]. Minor elevation in CRP can also predict undesired outcomes or complications in various medical conditions, or likelihood of dying in diVerent diseases. In acute‐phase reaction, blood CRP concentration can increase up to 1000‐fold. An acute‐phase reactant has been defined as one whose plasma concentration increases or decreases by at least 25% during inflammatory responses [34]. The magnitude of the increases varies from about 50% in the case of complement components to 1000‐fold in the case of CRP. When the inflammatory stimulus is eliminated, CRP concentration may fall rapidly or remain elevated, depending on the type and chronicity of the stimulus. Of the many soluble factors that initiate and maintain the inflammatory response at the molecular level, three cytokines specifically signal the transcription of human CRP from the liver, namely IL‐1b, IL‐6, and newly found IL‐17. Other cells besides hepatic cells are also capable of CRP expression and secretion, possibly contributing to baseline level of blood CRP as well, but the exact level of the contribution is yet to be determined [35–44]. The same cytokines induce CRP production from the liver under both acute and low‐grade inflammation. In low‐grade inflammation, the traditional symptoms of inflammation, namely redness, swelling, heat, and pain are all not necessarily present, even though phagocytes and various other inflammatory cells might be. Low‐grade inflammation may not mean that the full human armory of inflammation related molecules and events are activated. Epidemiological studies have revealed demographic, metabolic, and socioeconomic factors that aVect the baseline CRP concentrations. These include age, sex, cigarette smoking [45], BMI, birth weight [46, 47], ethnic background [48], educational achievements [49], socioeconomic position [48], dietary fatty acids [50], coVee consumption [51], alcohol intake [52–54], and dietary fiber intake [55–57]. Many of these factors are also related to increased cytokine levels. Some of these factors are briefly described below.

3. Demographic, Metabolic, and Socioeconomic Factors Population studies have shown that serum CRP concentrations are broadly distributed and highly skewed to the right in apparently healthy people [32, 58], that is, most individuals are grouped together in the lowest values measured, the rest being dispersed up to 10 mg/l or more. CRP concentrations have been found to increase slightly with age and to diVer between males and females [59, 60], females tending to have slightly increased

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

115

concentrations compared to males [61, 62]. However, the sex diVerence is explainable in apparently healthy subjects by female oral contraceptive usage or hormone replacement therapy and adjustment for these factors generally abolishes the diVerence [58, 63]. Cigarette smoking has been associated with increased CRP values [45]. A strong and independent dose–response relationship between cigarette smoking and CRP was found in a large population‐based study of over 4000 current smokers, almost 5000 former smokers, and >8000 never‐ smokers with smoking status based on cotinine levels [64]. CRP levels seem to be primarily related to lifetime exposure of smoking (pack‐years) and not to years since quitting smoking, and in several studies CRP was still significantly raised 10–19 years after smoking cessation [65–67]. This suggests an ongoing low‐grade inflammatory response persisting in former smokers, although most of other (noninflammatory) smoking induced changes are reversible after quitting [45]. Birth weight and CRP have been found to associate inversely [46, 47]. A 1 kg increase in birth weight was associated with 11–16% decrease in adult CRP, the eVect being similar in both men and women [46, 47]. It seems that small size at birth combined with excessive weight gain during adolescence and young adulthood may predispose to low‐grade inflammation [46]. Several mechanisms might explain the relationship between low‐grade inflammation in adulthood and small size at birth; an activated inflammatory response as a result of stress response programming during pregnancy or early life. Maternal stressors—diet, physiological stress among others—might aVect the fetus by trans‐placental passage of maternal hormones including cortisol and lead to consistent glucocorticoid signaling. An alternative explanation might be that restricted overall growth might lead to impaired liver or kidney development in critical early phase and thus might aVect adult IL‐6 concentrations [46]. As multiethnic studies have pointed to significant diVerences between racial/ethnic groups in prevalence and incidence of cardiovascular disease [68–71], interests to study diVerences in CRP concentration between ethnic/ racial groups also aroused. In 2003, the AHA/CDC panel identified a major and urgent need for data to address the question about ethnic diVerences in CRP concentration [30]. Recently, a systematic review of population‐based studies focusing on ethnic diVerentials of CRP concentrations was reported by Nazmi and Victora [48]. Variables related to race, skin color, or ethnicity were analyzed. The authors reported that of 15 studies, 14 found diVerences between racial/ethnic groups. Mostly whites had the lowest CRP concentrations while blacks, Hispanics and South Asians tended to have the highest concentrations. Most of the studies included adjustment for potential mediating variables in the causal chain between race/ethnicity and CRP, after which the association in most cases attenuated but remained still

116

CARITA M. EKLUND

significant [48]. The conclusion was that nonwhite race was associated with elevated CRP concentration among adults and that most analyses in the literature are underestimating the true eVects of racial/ethnic factors due to adjustment for mediating factors (sex hormone concentrations, markers of socioeconomic status, anthropometric parameters, etc.). Much, but not all, of the ethnic diVerences seem to be accounted for the modifiable risk factors, particularly BMI [72]. In a recent systematic review of population‐based studies on socioeconomic diVerentials of CRP, the authors found that of 20 studies, 19 found inverse association between CRP and socioeconomic position [48]. Another report linked socioeconomic status to classic markers of inflammation, where lower household income and educational background were associated with increased blood levels of both CRP and IL‐6 [73]. The classical Western diet of developed countries is characterized by high intake of fat, red meat, cereal grains, and refined sugars. It has been estimated that the present Western diet is ‘‘deficient’’ in omega‐3 fatty acids with a ratio of omega‐6 to omega‐3 of 16/1, instead of 1/1 as is the case with wild animals and presumably human beings [74]. A high omega‐6/omega‐3 ratio promotes cardiovascular and inflammatory diseases, whereas increased levels of omega‐3 fatty acids exert suppressive eVects. Several observational studies have reported a positive correlation between diets with a high content of saturated and trans fatty acids and biomarkers of inflammation including CRP [75, 76], while intervention trials with trans fatty acids have been somewhat conflicting [50, 77]. Recently, palmitic acid, a common saturated fatty acid was shown to promote inflammation by increasing NF‐kB binding activity and TNF production in murine adipocytes [78]. The polyunsaturated fatty acids, in particularly, omega‐3 fatty acids, modulate the inflammatory response by multiple mechanisms. Ferrucci and colleagues studied the relationship of plasma polyunsaturated fatty acids to circulating inflammatory markers in 1123 persons aged 20–98 years in a community‐based sample [79]. The total omega‐3 fatty acids were independently associated with lower levels of proinflammatory markers [IL‐6, TNF, CRP], and higher anti‐inflammatory markers (soluble IL‐6R, IL‐10, transforming growth factor‐a (TGFa)) independent of confounders. The omega‐6/ omega‐3 ratio was a strong negative correlate of IL‐10. The authors concluded, ‘‘Omega‐3 fatty acids are beneficial in patients aVected by diseases characterized by active inflammation.’’ High omega‐3 consumption is also reported to decrease CRP values in healthy older persons [80]. A Mediterranean‐style diet, high in oleic acid or monounsaturated fatty acid content, fiber, and antioxidants have been reported to reduce inflammation in metabolic diseases in both epidemiological and interventional studies [81–84]. In particular, lower levels of CRP were found in asymptomatic subjects at high

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

117

cardiovascular risk following the Mediterranean diet after only 3 weeks [85]. A study on overall dietary patterns, endothelial dysfunction and inflammatory markers revealed that the alternate Mediterranean Diet Index among the alternate Healthy Eating Index had the strongest inverse association with these markers, compared with other dietary patterns [76]. One of the strongest correlates of blood CRP concentration is the BMI. High BMI is associated with high CRP and overweight and obese individuals tend to have higher CRP concentrations than lean or normal weight individuals [86]. Overweight and obese individuals have higher CRP due to higher adipose mass [86]. Weight loss induces favorable changes in CRP, the decline in CRP reported to be on average 0.13 mg/l for each kilogram of weight lost [87]. It was long thought that adipose tissue was solely an inert mass storing fat, but now it has been recognized that it is an active endocrine organ capable of secreting a large number of adipokines, cytokines, and chemokines. Obesity is considered as a low‐grade inflammation state resulting from the secretion of cytokines, chemokines, and hormone‐like factors. Increased adipose mass has been linked with increases in many inflammatory molecules; CRP, TNF, serum amyloid A, migration inhibitory factor, resistin, IL‐6, and inducible nitric oxide synthase, to mention a few [88–90]. It has been estimated that adipose tissue may produce about 25% of the total IL‐6 in circulation [91]. Adipose tissue of obese individuals seems to be infiltrated by an increased number of macrophages [88] and there are reports showing that macrophages are the main source of the increase in the circulating inflammatory molecules detected in obesity [92, 93], even though adipose cells themselves are also able to secrete CRP [41, 94]. The majority of the macrophages localize and aggregate to dead adipocytes [95]. Although macrophages infiltrate both visceral and subcutaneous adipose tissue, it is postulated that visceral macrophages would be the main source of systemic cytokine levels [96]. The source of macrophages in adipose tissue is not entirely clear, however; besides infiltration from blood there is also evidence suggesting that preadipocytes may convert to macrophages under favorable conditions [97]. In morbidly obese people macrophages form crown‐like structures, a hallmark of chronic inflammation, express activation markers and scavenge residual adipocyte lipid released from the dead adipocytes [95]. The increased concentration of IL‐6 in circulating blood induces CRP secretion from the liver and maintains a low‐grade inflammatory state. Consumption of dietary fiber has been inversely associated with CRP in cross‐sectional studies of mixed‐gender populations [55, 56] and in two smaller studies [98, 99]. However, association was not found in older, postmenopausal women [57], but instead association between dietary fiber intake and IL‐6 and TNF‐a‐R2 was [57]. Randomized controlled trials have not, however, been very promising. A trial of dietary fiber showed no significant relationship between the dietary fiber intervention and CRP concentrations

118

CARITA M. EKLUND

except in a subgroup of lean normotensive individuals [100], and another recent study did not find psyllium fiber supplementation to reduce CRP levels in overweight or obese individuals from baseline to three months [101]. Consequently, the impact of dietary fiber or fiber supplementation on inflammation is still somewhat unclear, although epidemiological evidence suggests CRP reduction by dietary fiber [55, 56]. The exact mechanism how fiber exerts its eVect on inflammation is not clear, but there are a few hypotheses. Dietary fiber might decrease the oxidation of lipids, which in turn may reduce inflammation [55] or it might reduce inflammation by altering adipocytokines in adipose tissue and increasing enterohepatic circulation of lipids and lipophilic compounds [102].

4. Proinflammatory Cytokines The main proinflammatory cytokine inducers of CRP in hepatic cells are the interleukins‐1 (IL‐1) and ‐6 (IL‐6) and recently found IL‐17. Generally, cytokines transduce signals from outside the cell into the cells via specific receptors. Compared to hormone or growth or factor receptors, the number of cytokine receptors on cell surfaces is usually a hundred time less than that of hormone or growth factor receptors [103]. Inside the cells, a given set of kinases, phosphatases and transcription factors are activated that are constitutively associated with the receptor, the set being determined by the cytoplasmic tail of the cytokine receptor. However, the target cell receives multiple signals through independent receptors, and inside the cell these transduction pathway components may engage in crosstalk. Furthermore, the combination and concentration of cytokines at specific sites diVer, and the microenvironment of the target cell also aVects the action of cytokines on their target cells. Put together, a single cytokine can exert pleiotrophic functions on cells, the result of the action being dependent on various factors both outside and inside the cell.

4.1. IL‐6 IL‐6 is a pleiotropic cytokine with proinflammatory, anti‐inflammatory, and endocrine functions. It is commonly produced at local tissue sites and released into the bloodstream at situations of homeostasis disturbance; typically trauma and acute infections. IL‐6 is induced together with other ‘‘alarm’’ cytokines TNF and IL‐1, which are all involved in the acute‐phase reaction elicitation [104]. IL‐6 has many biological functions; besides its role as the main acute‐phase protein synthesis regulator, it is a hemopoietic growth factor, hepatocyte stimulating factor able to induce the acute‐phase reaction, and a

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

119

diVerentiation factor for B and T cells [105]. It is also able to stimulate the hypothalamic‐pituitary‐adrenal axis by promoting CRH release, adrenocorticotrophic hormone synthesis and corticosteroid production [106, 107]. IL‐6 is a prototype of the helical cytokines of the IL‐6 family members that share a common signal‐transduction component (gp130) in their receptors. The helical cytokines, leukemia inhibitory factor (LIF), Oncostatin M (OSM), interleukin‐11 (IL‐11), ciliary neurotrophic factor (CNTF), and cardiotrophin‐1 (CT‐1) share a number of actions—an example of redundancy in cytokine biology—and participate in a variety of cellular processes [108]. IL‐6 eVects are mediated through IL‐6 receptor (IL‐6R) which consists of two molecules; the signal‐transducing IL‐6Rb (gp130), which is not ligand‐specific and is also used by the other helical cytokines, and the IL‐6 binding IL‐6Ra (gp80) chain which is specific for IL‐6 and belongs to the immunoglobulin gene superfamily. IL‐6Rb is present on almost every cell type [109], but IL‐6Ra is found mainly on monocytes, hepatocytes, T cells, and activated B cells [105, 110, 111]. Binding of IL‐6 to IL‐6Ra triggers association of IL‐6Ra with IL‐6Rb. This induces the signal transduction cascades and activation of various tyrosine kinases and transcription factors, namely JAK/STAT, Ras/Raf, and Src‐family of kinases [112, 113]. As an example of the signal trasductions, first the JAK family tyrosine kinases are activated, which phosphorylate the cytoplasmic tail of the receptor. The phosphorylated tyrosines then act as docking sites for STATs, which are also tyrosine‐phosphorylated and dimerize. After dimerization they enter the nucleus, bind DNA and act as transcription factors. A soluble form of IL‐6R also exists (sIL‐6R), which is generated either by proteolytic cleavage of the membrane bound IL‐6R or by alternative spliced mRNA [108]. It has an agonistic function to IL‐6 but may also function as an IL‐6R (after binding to IL‐6) on cells which lack IL‐6Ra but have IL‐6Rb [108]. Expression of IL6 is stimulated by a wide range of mediators, including a number of cytokines and growth factors, bacterial endotoxins, and neuropeptides. Most nucleated cells are capable of expressing and synthesizing IL‐6, but the most substantial source of IL‐6 are stimulated monocytes/macrophages [114]. For example, nonimmune cells such as fibroblasts, adipocytes, epithelial, and endothelial cells produce IL‐6. The hepatic cells respond to IL‐6 (and IL‐1) through transcription factor C/EBPb, also known as NF‐IL6. C/EBPb binds to CCAAT box motif in the promoter or enhancer regions of acute‐ phase genes, including CRP, and induces their expression [115]. 4.2. IL‐1 FAMILY The three best known members of the IL‐1 family are IL‐1a, IL‐1b, and interleukin 1 receptor antagonist (IL‐1Ra) [116]. These molecules are structurally related to one another and act in a hormone‐like fashion to regulate

120

CARITA M. EKLUND

cells involved in inflammatory responses. They all bind to IL‐1 receptors (IL‐1R) on cells. The former names of IL‐1 describe its major biological roles: endogenous pyrogen, leukocyte endogenous mediator, lymphocyte‐activating factor, B cell activating factor, osteoclast activating factor, epidermal cell‐derived thymocyte activating factor, hemopoieting‐1 and mononuclear cell factor [116, 117]. Both IL‐1b and IL‐1a are proinflammatory molecules produced by multiple cells including monocytes, macrophages, neutrophils, hepatocytes, and tissue macrophages in response to proinflammatory signals, such as LPS. IL‐1b is produced in response to systemic and local inflammation and contributes to hepatic acute‐phase protein synthesis [118]. It is also central to the pathogenesis of local and systemic, acute, and chronic inflammatory diseases of the peripheral organs as well as the central nervous system [119]. IL‐1a and IL‐1b share about 30% structural homology. They are both produced as 31 kDa precursors and can be cleaved to a 17‐kDa ‘‘mature’’ form by intracellular or extracellular proteases [116]. However, almost 90% of IL‐1a remains in the cytosol of cells in its precursor form or is transported to the cell surface [120, 121], whereas 80% of processed IL‐1b is released into the extracellular space [116]. The membrane bound form of IL‐1a may become activated and released following cleavage by an extracellular protease, perhaps to act as a paracrine messenger to adjacent cells [120–122]. However, considering that some pro‐IL‐1a may translocate to the nucleus, it is possible that intracellular pro‐IL‐1a directly functions as a gene regulator [123]. IL‐1 proteins tend to lack the hydrophobic signal sequence (i.e., leader sequence) that targets most secreted proteins to the endoplasmic reticulum, and thus secretion via the classical ER‐Golgi route is precluded. This could occur via exocytosis [124], active transport and/or following cell death [125–127]. There are two forms of IL‐1 receptors, type I (IL‐1RI) and type II (IL‐1RII). Binding of IL‐1a or IL‐1b to IL‐1RI [128] is facilitated by interaction with IL‐1 receptor accessory protein (IL‐1RAcP) [129]. IL‐1RI can be found prominently on endothelial cells, smooth muscle cells, epithelial cells, hepatocytes, fibroblasts, keratinocytes, epidermal dendritic cells, and T cells [116]. IL‐1RII is also able to bind IL‐1 but cannot signal as it lacks the intracellular domain. The role of IL‐1RII is to act as a sink for IL‐1b and has been termed a decoy receptor [130], and thus may negatively regulate cell activation. The downstream signaling pathways are induced by IL‐1R together with IL‐1RAcP, and involve many diVerent adaptor molecules and a cascade of kinases (for detailed review, see [131]), but eventually leads to liberation of NF‐kB to activate transcription in the nucleus. Induction of CRP gene expression by IL‐1b has been studied mostly in hepatoma cell lines (Hep3B, PLC/PRF/5) while studies in primary human hepatocytes are less abundant. The outcome of the in vitro studies has been

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

121

dependent on the specific cell type studied; the results vary according to cell type [132]. However, both of these cell lines have their limitations. Where hepatoma cell lines might have lost their physiological properties of primary hepatic cells, for example, the endogenous CRP gene is either dysregulated or weakly active, the primary hepatic cells are hard to handle in culture and the cell preparations may be contaminated by other cell types. Particularly, contamination by KupVer cells may decrease the reliability of the in vitro studies of primary hepatic cells [132]. Nonetheless, current knowledge on CRP expression mostly comes from studies on hepatoma cell lines. In vivo studies of acute‐phase protein synthesis in IL‐1b deficient mice have also been conducted, but as CRP is not an acute‐phase protein in mice the results do not cover CRP [133]. In Hep3B hepatoma cells the activation of CRP is most enhanced by combined stimulation by IL‐6 and IL‐1b, while moderate activation is achieved by IL‐6 stimulation alone. Stimulation by IL‐1b alone is not capable of CRP induction [132, 134, 135]. The results from PLC/PRF/5 cells are similar yet not identical to Hep3B cells; IL‐6 alone induces CRP but IL‐1b does not [136, 137], whereas IL‐1b either potentiates IL‐6 induced expression [136] or not [134]. In contrast, studies in primary hepatic cells show CRP activation for both IL‐6 and IL‐1b stimulation alone, and combined stimulation have an additive eVect on CRP activation [138]. There is evidence that IL‐1a also is able to induce CRP expression. In a study of human peripheral blood mononuclear cells, IL‐1a incubation induced CRP expression [36]. IL1‐Ra functions as a natural antagonist to both IL‐1a and IL‐1b and blocks the biological responses of the two agonists by competing for key binding sites on cell‐surface IL‐1R [139]. It is structurally related to the other ligands, but has undergone mutations leaving it incapable of interacting with IL‐1RAcP. The major isoform of IL‐1Ra is secretory IL‐1Ra (sIL‐1Ra), which is secreted largely from the same cells that release IL‐1b. However, high concentrations of IL‐1Ra, about 100‐fold excess, relative to IL‐1, are needed to block IL‐1‐mediated signal transduction [140]. IL‐1Ra has also three intracellular forms (icIL‐1Ra1,2,3). Several studies have described that serum levels of IL‐1Ra are elevated in patients with inflammatory conditions. The balance between IL‐1 and IL‐1Ra in local tissues plays an important role in the susceptibility to and severity of many diseases. The major form of IL‐1Ra, sIL‐1Ra is actively produced in acute‐phase responses by hepatic cells [141]. 4.3. TNF TNF (previously called TNF‐a) is an important immune mediator of inflammation and the host’s defense against injury or pathogens. Macrophages and monocytes are the predominant sources of TNF after activation

122

CARITA M. EKLUND

by IL‐1, hypoxia, Fc receptor cross‐linking, C5a, or bacterial components such as LPS [142]. However, other cell types are also able to secrete TNF, for example, astrocytes, cardiac myocytes, and keratinocytes. TNF is synthesized as a transmembrane protein after which it is cleaved by a protease to release the soluble form of TNF. TNF release after stimulus (e.g., LPS) is very rapid; in monocytes secretion starts in some minutes and is extinguished in 3–4 h [143]. TNF binds to two types of receptors, TNFRI and TNFRII. The response to TNF is mainly mediated through TNFRI, the main signaling receptor on responding cells. TNFRII also delivers some intracellular signals, but lacks the intracellular death domain region and is believed to act mainly by passing on the bound TNF to TNFRI. Binding of TNF stimulates the monocytes and macrophages to secrete IL‐6 and IL‐1, among other factors, which in turn induce the liver to produce CRP. IL‐1 also induces the cells to secrete more IL‐6 in paracrine and autocrine fashion. Thus, TNF can stimulate CRP expression through IL‐1 and IL‐6, but not directly in hepatoma cell lines [135], although it is able to directly induce other acute‐ phase proteins. TNF is neither able to induce CRP [132, 138], IL‐1b, or IL‐6 from the primary human hepatocytes [132], thus induction probably goes through activated immune cells and their IL‐1 and IL‐6 production. 4.4. IL‐17 IL‐17 is a quite recently found proinflammatory cytokine [144] produced by activated Th17 cells, a new T‐helper cell population diVerent from the classical Th1 and Th2 subsets [145–147]. The human IL17A gene product is a protein of 150 amino acids [144] with a molecular weight of 15 kDa, and is secreted as a disulfide linked homodimer of 30–35 kDa glycoprotein [148]. While IL‐17 is expressed only by T‐cells, its receptor (IL‐17R) is expressed in various organs including the lung, kidney, liver, and spleen [149] and in a variety of cells, for example, epithelial cells, fibroblasts, B, and T lymphocytes. Thus, IL‐17 has far reaching eVects in the body. The receptor is a single‐pass transmembrane protein of approximately 130 kDa. Binding of IL‐17 to the receptor activates extracellular signal‐regulated protein kinase (ERK), c‐jun N‐terminal kinase (JNK), and p38 MAP kinase pathways [150–152] and results in upregulation of IL‐6, IL‐1, and NF‐kb [153]. The cells start to produce chemokines and cytokines, including IL‐8, CXCL1, CXCL2, CXCL5, granulocyte colony stimulating factor (G‐CSF), and IL‐6, which result in generation and accumulation of neutrophils. IL‐17 is therefore considered as an important cytokine in neutrophil‐mediated inflammatory responses. In addition to mediating host defensive mechanisms to various infections, especially to extracellular bacteria infections, IL‐17 is involved in the pathogenesis of many autoimmune diseases.

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

123

It was recently shown that IL‐17 is capable of CRP stimulation [152]. IL‐17 could stimulate CRP expression from both hepatocytes (Hep3B cell line and primary hepatocytes) and coronary artery smooth muscle cells via p38 MAPK and ERK1/2‐dependent signal pathways. Interestingly, all these cell types were able to respond to IL‐17 independently of IL‐6 and IL‐1b.

5. Signaling Through IL Receptors Signaling from the IL receptors to CRP induction is done through several signaling molecules and transcription factors. The regulation of CRP expression is done at transcriptional level and focuses on the 300‐bp long CRP promoter. Most of the CRP expression studies have been done in Hep3B cells, and these are described below. The promoter region of CRP of Hep3B cells harbors binding sites for transcription factors HNF‐1a (hepatic nuclear factor 1a), OCT‐1 (octamerbinding transcription factor 1), STAT3 (signal transducer and activator of transcription 3), C/EBPb/d (CCAAT binding protein‐b/d; also called NFIL6), c‐Rel, and p50. In contrast to other CRP binding transcription factors, HNF‐1a and OCT‐1 are not activated by cytokines but rather present constitutively [154–156]. Transcription factors c‐Rel and p50 belong to the Rel/NF‐kB protein family, which also includes p65, RelB, and p52 members. These proteins share a 300 amino acid N‐terminal domain, which contains a DNA recognition motif, a nuclear localization and dimerization sequence and a sequence for IkB binding. Some of the proteins have also a more C‐terminal sequence for transcription activation (e.g., c‐Rel and p65). The heterodimer of p50 and p65, the classical NF‐kB, binds to a kB site centered at position‐69 on the CRP promoter [156]. The homodimer of p50 binds DNA through a nonconsensus kB site at position‐ 47, overlapping the C/EBP binding site [157, 158]. Instead of binding directly to DNA, c‐Rel homodimer binds another transcription factor, C/EBPb and enhances its binding to DNA [159]. The C/EBP binds to a nonconsensus C/EBP binding sequence in CRP promoter and without c‐Rel it binds this sequence relatively poorly. The role of c‐Rel is thus to enhance the transcription by binding to C/EBPb and potentiating its capability to bind CRP promoter. The C/EBP family members act as master regulators of many cellular responses, like cellular proliferation, diVerentiation, and inflammation [115]. The family consists of six members that are designated as C/EBP followed by a Greek letter indicating the chronological order of their discovery, C/EBPa–C/EBPz [115, 160]. Besides the liver, they are expressed in various tissues, including the lungs and adipose tissue [161]. In acute‐phase response, through the actions of cytokines such as IL‐6, IL‐1, and TNF [162],

124

CARITA M. EKLUND

the level of C/EBPa expression decreases while that of C/EBPb and ‐d increases. C/EBP is one of the main transcription factors essential for CRP expression [163]. There are two binding sites for C/EBPb/d on the CRP promoter, at positions ‐53 and ‐219 [154, 164]. STAT3 belongs to the family of signal transducers and activators of transcription and is an important mediator of the eVects of many cytokines, including IL‐6. Binding of IL‐6 to its receptor complex leads to phosphorylation of Janus kinase kinases, with subsequent phosphorylation (15–60 min), dimerization and nuclear translocation of STAT3. STAT3 binds to CRP promoter containing the motif TT(N)4AA [165], which is located at nucleotide site ‐108 [163], and enhances its transcription. A recent paper suggested that C/EBPz, which is ubiquitously expressed at low levels in proliferating cells, including Hep3B cells, is bound to CRP promoter under basal conditions, but is replaced by C/EBPb and p50 under induction [166]. Another recent study showed evidence that C/EBPb, STAT3, p50, and c‐Rel are all bound to CRP promoter under basal conditions at low levels (absence of cytokines), and that cytokine treatment (IL‐6 plus IL‐1) markedly enhances the binding of C/EBPb, after which CRP mRNA is produced [163].

6. Genetic Polymorphisms Serum CRP concentrations are influenced by genetic polymorphisms of various genes. Polymorphisms of several proinflammatory cytokine genes have been reported to associate with CRP concentration. A summary of these studies is presented in Table 1. Majority of the studies are quite small, and adjustment for covariates is lacking in many papers. Besides cytokine genes, polymorphisms in other genes also show association with CRP concentrations, including CRP, LEPR, HNF1A, APOE, GCKR, IRAK1, FTO [167–171].

7. Conclusions Low grade inflammation is being accepted to lie behind many common diseases. However, it should be kept in mind that minor elevation in one of the major inflammatory markers measured in clinical practice today, CRP, can be the result of physiological homeostasis maintenance caused by minor tissue damage in our everyday life, reflecting CRP’s role as a waste management molecule helping phagocytes remove cellular debris. Infection or a full blown inflammation with classical symptoms of redness, swelling, heat, and pain is not necessarily present. A minor elevation in CRP can be caused by

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

125

TABLE 1 ASSOCIATION STUDIES SHOWING ASSOCIATION OF THE CYTOKINE GENE POLYMORPHISMS WITH INCREASED CRP

Gene

Study population

IL1B

Caucasian (>97%) individuals undergoing coronary angiography, age 18–75 years, n ¼ 454 A subgroup from the ARIC study; Caucasian individuals, age 45–65 years, n ¼ 900 Healthy Caucasian volunteers, age19–64, n ¼ 338 Caucasian coronary heart disease patients, n ¼ 160 Patients with severe periodontal infections, n ¼ 94 A pooled analysis of three studies (CARDIA, CLEAR and CHS), n ¼ 7145 Patients with severe periodontal infections, n ¼ 94 Hypertensive Caucasian patients with their families, n ¼ 588 Patients with type 2 diabetes and peripheral artery disease, n ¼ 290 Cardiovascular Health Study, Caucasian older adults, n ¼ 4714 Japanese male transit company employees, 35–60 years, n ¼ 347 Chinese normo‐ and hypertensive subjects, mean age 54 and 57 years, respectively, n ¼ 502 Korean men with coronary artery disease, mean age 55 years, n ¼ 536 Healthy postmenopausal women, mean age 72 years, n ¼ 495 Diabetic and healthy Caucasian women, n ¼ 633 and n ¼ 692, respectively Healthy adults from the HERITAGE family study, n ¼ 688

IL1A IL1Ra

IL6

IL6R

TNF

Allele or haplotype associated with increased CRP

References

rs1143634 T‐allele

[172]

haplotypes B3/B3, B2/B3 and B3/B4

[173]

rs1143634 C‐allele

[174]

rs1143634 T‐allele

[175]

rs1800587 T‐allele

[176]

rs4251961 C‐allele rs2232354 G‐allele rs380092 A‐allele rs1800795 C‐allele

[177]

[176]

rs1800795 C‐allele

[178]

rs1800795 GG genotype

[179]

rs1800795 C‐allele rs1554606 T‐allele rs1800786 G‐allele in nonsmokers rs1800796 G‐allele

[180]

rs1800796 GG genotype in men not treated with lipid‐lowering drugs rs1800795 C‐allele

[183]

rs8192284 A‐allele þ haplotypes 21122 and 11211 rs1800629 A‐allele

[181] [182]

[184] [185]

[186]

126

CARITA M. EKLUND

several factors, for example, metabolic, demographic, socioeconomic, and genetic factors. Therefore, repeated measurements of CRP would be a more reliable tool for low grade inflammation determination than single measurement. Future research hopefully defines a more precise definition to low‐ grade inflammation, both at molecular and symptoms level, so that our overall understanding of how the immune system senses stress in the form of infection and tissue damage is widened. ACKNOWLEDGMENT I would like to thank Prof. M Hurme for valuable discussions during the manuscript preparation.

REFERENCES [1] C.A. Janeway Jr., R. Medzhitov, Innate immune recognition, Annu. Rev. Immunol. 20 (2002) 197–216. [2] J.E. Volanakis, M.H. Kaplan, Specificity of C‐reactive protein for choline phosphate residues of pneumococcal C‐polysaccharide, Proc. Soc. Exp. Biol. Med. 136 (1971) 612–614. [3] F.A. Robey, K.D. Jones, T. Tanaka, T.Y. Liu, Binding of C‐reactive protein to chromatin and nucleosome core particles. A possible physiological role of C‐reactive protein, J. Biol. Chem. 259 (1984) 7311–7316. [4] T.W. Du Clos, L.T. Zlock, R.L. Rubin, Analysis of the binding of C‐reactive protein to histones and chromatin, J. Immunol. 141 (1988) 4266–4270. [5] E.M. Salonen, T. Vartio, K. Hedman, A. Vaheri, Binding of fibronectin by the acute phase reactant C‐reactive protein, J. Biol. Chem. 259 (1984) 1496–1501. [6] S.J. Swanson, M.M. McPeek, R.F. Mortensen, Characteristics of the binding of human C‐reactive protein (CRP) to laminin, J. Cell Biochem. 40 (1989) 121–132. [7] T.W. Du Clos, C‐reactive protein reacts with the U1 small nuclear ribonucleoprotein, J. Immunol. 143 (1989) 2553–2559. [8] M.P. Stein, J.C. Edberg, R.P. Kimberly, E.K. Mangan, D. Bharadwaj, C. Mold, et al., C‐reactive protein binding to FcgammaRIIa on human monocytes and neutrophils is allele‐specific, J. Clin. Invest. 105 (2000) 369–376. [9] F.C. de Beer, A.K. Soutar, M.L. Baltz, I.M. Trayner, A. Feinstein, M.B. Pepys, Low density lipoprotein and very low density lipoprotein are selectively bound by aggregated C‐reactive protein, J. Exp. Med. 156 (1982) 230–242. [10] J.E. Volanakis, K.W. Wirtz, Interaction of C‐reactive protein with artificial phosphatidylcholine bilayers, Nature 281 (1979) 155–157. [11] A.J. Narkates, J.E. Volanakis, C‐reactive protein binding specificities: artificial and natural phospholipid bilayers, Ann. N. Y. Acad. Sci. 389 (1982) 172–182. [12] D. Gershov, S. Kim, N. Brot, K.B. Elkon, C‐Reactive protein binds to apoptotic cells, protects the cells from assembly of the terminal complement components, and sustains an antiinflammatory innate immune response: implications for systemic autoimmunity, J. Exp. Med. 192 (2000) 1353–1364. [13] M.K. Chang, C.J. Binder, M. Torzewski, J.L. Witztum, C‐reactive protein binds to both oxidized LDL and apoptotic cells through recognition of a common ligand: phosphorylcholine of oxidized phospholipids, Proc. Natl. Acad. Sci. USA 99 (2002) 13043–13048.

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

127

[14] C.E. Hack, G.J. Wolbink, C. Schalkwijk, H. Speijer, W.T. Hermens, H. van den Bosch, A role for secretory phospholipase A2 and C‐reactive protein in the removal of injured cells, Immunol. Today 18 (1997) 111–115. [15] Y.P. Li, C. Mold, T.W. Du Clos, Sublytic complement attack exposes C‐reactive protein binding sites on cell membranes, J. Immunol. 152 (1994) 2995–3005. [16] J.E. Volanakis, A.J. Narkates, Interaction of C‐reactive protein with artificial phosphatidylcholine bilayers and complement, J. Immunol. 126 (1981) 1820–1825. [17] C. Mold, R. Baca, T.W. Du Clos, Serum amyloid P component and C‐reactive protein opsonize apoptotic cells for phagocytosis through Fcgamma receptors, J. Autoimmun. 19 (2002) 147–154. [18] A.J. Szalai, C-reactive protein (CRP) and autoimmune disease: facts and conjectures, Clin. Dev. Immunol. 11 (2004) 221–226. [19] C. Nathan, Points of control in inflammation, Nature 420 (2002) 846–852. [20] G.M. Barton, A calculated response: control of inflammation by the innate immune system, J. Clin. Invest. 118 (2008) 413–420. [21] C.J. Chen, H. Kono, D. Golenbock, G. Reed, S. Akira, K.L. Rock, Identification of a key pathway required for the sterile inflammatory response triggered by dying cells, Nat. Med. 13 (2007) 851–856. [22] E.S. Ford, The metabolic syndrome and C‐reactive protein, fibrinogen, and leukocyte count: findings from the Third National Health and Nutrition Examination Survey, Atherosclerosis 168 (2003) 351–358. [23] D.E. Laaksonen, L. Niskanen, K. Nyyssonen, K. Punnonen, T.P. Tuomainen, V.P. Valkonen, et al., C-reactive protein and the development of the metabolic syndrome and diabetes in middle-aged men, Diabetologia 47 (2004) 1403–1410. [24] A.D. Pradhan, J.E. Manson, N. Rifai, J.E. Buring, P.M. Ridker, C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus, J. Am. Med. Assoc. 286 (2001) 327–334. [25] A. Festa, R. D’Agostino, Jr., G. Howard, L. Mykkanen, R.P. Tracy, S.M. HaVner, Chronic subclinical inflammation as part of the insulin resistance syndrome: the Insulin Resistance Atherosclerosis Study (IRAS), Circulation 102 (2000) 42–47. [26] G.K. Hansson, Inflammation, atherosclerosis, and coronary artery disease, N. Engl. J. Med. 352 (2005) 1685–1695. [27] R. Ross, Atherosclerosis—an inflammatory disease, N. Engl. J. Med. 340 (1999) 115–126. [28] K. Heikkila, R. Harris, G. Lowe, A. Rumley, J. Yarnell, J. Gallacher, et al., Associations of circulating C‐reactive protein and interleukin‐6 with cancer risk: findings from two prospective cohorts and a meta-analysis, Cancer Causes Control 20 (2009) 15–26. [29] I. Kushner, D. Rzewnicki, D. Samols, What does minor elevation of C‐reactive protein signify? Am. J. Med. 119 (166) (2006) e17–e28. [30] T.A. Pearson, G.A. Mensah, R.W. Alexander, J.L. Anderson, R.O. Cannon, 3rd, M. Criqui, et al., Markers of inflammation and cardiovascular disease: application to clinical and public health practice: a statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association, Circulation 107 (2003) 499–511. [31] J. Danesh, J.G. Wheeler, G.M. Hirschfield, S. Eda, G. Eiriksdottir, A. Rumley, et al., C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease, N. Engl. J. Med. 350 (2004) 1387–1397. [32] P.M. Ridker, M. Cushman, M.J. Stampfer, R.P. Tracy, C.H. Hennekens, Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men, N. Engl. J. Med. 336 (1997) 973–979. [33] P. Sakkinen, R.D. Abbott, J.D. Curb, B.L. Rodriguez, K. Yano, R.P. Tracy, C-reactive protein and myocardial infarction, J. Clin. Epidemiol. 55 (2002) 445–451.

128

CARITA M. EKLUND

[34] J.J. Morley, I. Kushner, Serum C‐reactive protein levels in disease, Ann. N. Y. Acad. Sci. 389 (1982) 406–418. [35] P. Calabro, J.T. Willerson, E.T. Yeh, Inflammatory cytokines stimulated C‐reactive protein production by human coronary artery smooth muscle cells, Circulation 108 (2003) 1930–1932. [36] D.G. Haider, N. Leuchten, G. Schaller, G. Gouya, J. Kolodjaschna, L. Schmetterer, et al., C-reactive protein is expressed and secreted by peripheral blood mononuclear cells, Clin. Exp. Immunol. 146 (2006) 533–539. [37] A.E. Kuta, L.L. Baum, C-reactive protein is produced by a small number of normal human peripheral blood lymphocytes, J. Exp. Med. 164 (1986) 321–326. [38] Q. Dong, J.R. Wright, Expression of C‐reactive protein by alveolar macrophages, J. Immunol. 156 (1996) 4815–4820. [39] S.K. Venugopal, S. Devaraj, I. Jialal, Macrophage conditioned medium induces the expression of C‐reactive protein in human aortic endothelial cells: potential for paracrine/autocrine eVects, Am. J. Pathol. 166 (2005) 1265–1271. [40] J.M. Gould, J.N. Weiser, Expression of C‐reactive protein in the human respiratory tract, Infect. Immun. 69 (2001) 1747–1754. [41] P. Calabro, D.W. Chang, J.T. Willerson, E.T. Yeh, Release of C‐reactive protein in response to inflammatory cytokines by human adipocytes: linking obesity to vascular inflammation, J. Am. Coll. Cardiol. 46 (2005) 1112–1113. [42] W.J. Jabs, B.A. Logering, P. Gerke, B. Kreft, E.M. Wolber, M.H. Klinger, et al., The kidney as a second site of human C‐reactive protein formation in vivo, Eur. J. Immunol. 33 (2003) 152–161. [43] K. Yasojima, C. Schwab, E.G. McGeer, P.L. McGeer, Human neurons generate C‐reactive protein and amyloid P: upregulation in Alzheimer’s disease, Brain Res. 887 (2000) 80–89. [44] A.M. Wilson, J.D. Swan, H. Ding, Y. Zhang, R.J. Whitbourn, J. Gurry, et al., Widespread vascular production of C‐reactive protein (CRP) and a relationship between serum CRP, plaque CRP and intimal hypertrophy, Atherosclerosis 191 (2007) 175–181. [45] D.G. Yanbaeva, M.A. Dentener, E.C. Creutzberg, G. Wesseling, E.F. Wouters, Systemic eVects of smoking, Chest 131 (2007) 1557–1566. [46] I. Tzoulaki, M.R. Jarvelin, A.L. Hartikainen, M. Leinonen, A. Pouta, M. Paldanius, et al., Size at birth, weight gain over the life course, and low-grade inflammation in young adulthood: northern Finland 1966 Birth Cohort study, Eur. Heart J. 29 (2008) 1049–1056. [47] N. Sattar, A. McConnachie, D. O’Reilly, M.N. Upton, I.A. Greer, G. Davey Smith, et al., Inverse association between birth weight and C‐reactive protein concentrations in the MIDSPAN Family Study, Arterioscler. Thromb. Vasc. Biol. 24 (2004) 583–587. [48] A. Nazmi, C.G. Victora, Socioeconomic and racial/ethnic diVerentials of C‐reactive protein levels: a systematic review of population-based studies, BMC Public Health 7 (2007) 212. [49] D. Gimeno, J.E. Ferrie, M. Elovainio, L. Pulkki-Raback, L. Keltikangas-Jarvinen, C. Eklund, et al., When do social inequalities in C‐reactive protein start? A life course perspective from conception to adulthood in the Cardiovascular Risk in Young Finns Study, Int. J. Epidemiol. 37 (2008) 290–298. [50] D.J. Baer, J.T. Judd, B.A. Clevidence, R.P. Tracy, Dietary fatty acids aVect plasma markers of inflammation in healthy men fed controlled diets: a randomized crossover study, Am. J. Clin. Nutr. 79 (2004) 969–973. [51] A. Zampelas, D.B. Panagiotakos, C. Pitsavos, C. Chrysohoou, C. Stefanadis, Associations between coVee consumption and inflammatory markers in healthy persons: the ATTICA study, Am. J. Clin. Nutr. 80 (2004) 862–867.

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

129

[52] A. Imhof, M. Froehlich, H. Brenner, H. Boeing, M.B. Pepys, W. Koenig, EVect of alcohol consumption on systemic markers of inflammation, Lancet 357 (2001) 763–767. [53] A. Imhof, M. Woodward, A. Doering, N. Helbecque, H. Loewel, P. Amouyel, et al., Overall alcohol intake, beer, wine, and systemic markers of inflammation in western Europe: results from three MONICA samples (Augsburg, Glasgow, Lille), Eur. Heart J. 25 (2004) 2092–2100. [54] M.A. Albert, R.J. Glynn, P.M. Ridker, Alcohol consumption and plasma concentration of C‐reactive protein, Circulation 107 (2003) 443–447. [55] U.A. Ajani, E.S. Ford, A.H. Mokdad, Dietary fiber and C‐reactive protein: findings from national health and nutrition examination survey data, J. Nutr. 134 (2004) 1181–1185. [56] D.E. King, B.M. Egan, M.E. Geesey, Relation of dietary fat and fiber to elevation of C‐reactive protein, Am. J. Cardiol. 92 (2003) 1335–1339. [57] Y. Ma, J.R. Hebert, W. Li, E.R. Bertone-Johnson, B. Olendzki, S.L. Pagoto, et al., Association between dietary fiber and markers of systemic inflammation in the Women’s Health Initiative Observational Study, Nutrition 24 (2008) 941–949. [58] A. Imhof, M. Frohlich, H. Loewel, N. Helbecque, M. Woodward, P. Amouyel, et al., Distributions of C‐reactive protein measured by high-sensitivity assays in apparently healthy men and women from diVerent populations in Europe, Clin. Chem. 49 (2003) 669–672. [59] E.S. Ford, W.H. Giles, G.L. Myers, D.M. Mannino, Population distribution of highsensitivity C‐reactive protein among US men: findings from National Health and Nutrition Examination Survey 1999–2000, Clin. Chem. 49 (2003) 686–690. [60] N. Rifai, P.M. Ridker, Population distributions of C‐reactive protein in apparently healthy men and women in the United States: implication for clinical interpretation, Clin. Chem. 49 (2003) 666–669. [61] M.H. Wener, P.R. Daum, G.M. McQuillan, The influence of age, sex, and race on the upper reference limit of serum C‐reactive protein concentration, J. Rheumatol. 27 (2000) 2351–2359. [62] X. Garcia-Moll, E. Zouridakis, D. Cole, J.C. Kaski, C-reactive protein in patients with chronic stable angina: diVerences in baseline serum concentration between women and men, Eur. Heart J. 21 (2000) 1598–1606. [63] M. Raitakari, K. Mansikkaniemi, J. Marniemi, J.S. Viikari, O.T. Raitakari, Distribution and determinants of serum high-sensitive C‐reactive protein in a population of young adults: the Cardiovascular Risk in Young Finns Study, J. Intern. Med. 258 (2005) 428–434. [64] L.A. Bazzano, J. He, P. Muntner, S. Vupputuri, P.K. Whelton, Relationship between cigarette smoking and novel risk factors for cardiovascular disease in the United States, Ann. Intern. Med. 138 (2003) 891–897. [65] M. Frohlich, M. Sund, H. Lowel, A. Imhof, A. HoVmeister, W. Koenig, Independent association of various smoking characteristics with markers of systemic inflammation in men. Results from a representative sample of the general population (MONICA Augsburg Survey 1994/95), Eur. Heart J. 24 (2003) 1365–1372. [66] G.D. Lowe, J.W. Yarnell, A. Rumley, D. Bainton, P.M. Sweetnam, C-reactive protein, fibrin D-dimer, and incident ischemic heart disease in the Speedwell study: are inflammation and fibrin turnover linked in pathogenesis? Arterioscler. Thromb. Vasc. Biol. 21 (2001) 603–610. [67] S.G. Wannamethee, G.D. Lowe, A.G. Shaper, A. Rumley, L. Lennon, P.H. Whincup, Associations between cigarette smoking, pipe/cigar smoking, and smoking cessation, and haemostatic and inflammatory markers for cardiovascular disease, Eur. Heart J. 26 (2005) 1765–1773.

130

CARITA M. EKLUND

[68] M.A. Winkleby, H.C. Kraemer, D.K. Ahn, A.N. Varady, Ethnic and socioeconomic diVerences in cardiovascular disease risk factors: findings for women from the Third National Health and Nutrition Examination Survey, 1988–1994, J. Am. Med. Assoc. 280 (1998) 356–362. [69] S.S. Anand, S. Yusuf, V. Vuksan, S. Devanesen, K.K. Teo, P.A. Montague, et al., DiVerences in risk factors, atherosclerosis, and cardiovascular disease between ethnic groups in Canada: the Study of Health Assessment and Risk in Ethnic groups (SHARE), Lancet 356 (2000) 279–284. [70] R. Cooper, J. Cutler, P. Desvigne-Nickens, S.P. Fortmann, L. Friedman, R. Havlik, et al., Trends and disparities in coronary heart disease, stroke, and other cardiovascular diseases in the United States: findings of the national conference on cardiovascular disease prevention, Circulation 102 (2000) 3137–3147. [71] R.S. Cooper, Social inequality, ethnicity and cardiovascular disease, Int. J. Epidemiol. 30 (Suppl. 1) (2001) S48–S52. [72] A. Kelley-Hedgepeth, D.M. Lloyd-Jones, A. Colvin, K.A. Matthews, J. Johnston, M.R. Sowers, et al., Ethnic diVerences in C‐reactive protein concentrations, Clin. Chem. 54 (2008) 1027–1037. [73] N. Ranjit, A.V. Diez-Roux, S. Shea, M. Cushman, H. Ni, T. Seeman, Socioeconomic position, race/ethnicity, and inflammation in the multi-ethnic study of atherosclerosis, Circulation 116 (2007) 2383–2390. [74] A.P. Simopoulos, Evolutionary aspects of diet, the omega‐6/omega‐3 ratio and genetic variation: nutritional implications for chronic diseases, Biomed. Pharmacother. 60 (2006) 502–507. [75] E. Lopez-Garcia, M.B. Schulze, T.T. Fung, J.B. Meigs, N. Rifai, J.E. Manson, et al., Major dietary patterns are related to plasma concentrations of markers of inflammation and endothelial dysfunction, Am. J. Clin. Nutr. 80 (2004) 1029–1035. [76] T.T. Fung, M.L. McCullough, P.K. Newby, J.E. Manson, J.B. Meigs, N. Rifai, et al., Diet-quality scores and plasma concentrations of markers of inflammation and endothelial dysfunction, Am. J. Clin. Nutr. 82 (2005) 163–173. [77] A.H. Lichtenstein, A.T. Erkkila, B. Lamarche, U.S. Schwab, S.M. Jalbert, L.M. Ausman, Influence of hydrogenated fat and butter on CVD risk factors: remnant-like particles, glucose and insulin, blood pressure and C‐reactive protein, Atherosclerosis 171 (2003) 97–107. [78] R.L. Bradley, F.F. Fisher, E. Maratos-Flier, Dietary fatty acids diVerentially regulate production of TNF-alpha and IL‐10 by murine 3T3‐L1 adipocytes, Obesity (Silver Spring) 16 (2008) 938–944. [79] L. Ferrucci, A. Cherubini, S. Bandinelli, B. Bartali, A. Corsi, F. Lauretani, et al., Relationship of plasma polyunsaturated fatty acids to circulating inflammatory markers, J. Clin. Endocrinol. Metab. 91 (2006) 439–446. [80] P.D. Tsitouras, F. Gucciardo, A.D. Salbe, C. Heward, S.M. Harman, High omega‐3 fat intake improves insulin sensitivity and reduces CRP and IL6, but does not aVect other endocrine axes in healthy older adults, Horm. Metab. Res. 40 (2008) 199–205. [81] K. Esposito, R. Marfella, M. Ciotola, C. Di Palo, F. Giugliano, G. Giugliano, et al., EVect of a mediterranean-style diet on endothelial dysfunction and markers of vascular inflammation in the metabolic syndrome: a randomized trial, J. Am. Med. Assoc. 292 (2004) 1440–1446. [82] K. Esposito, M. Ciotola, D. Giugliano, Mediterranean diet, endothelial function and vascular inflammatory markers, Public Health Nutr. 9 (2006) 1073–1076. [83] C. Chrysohoou, D.B. Panagiotakos, C. Pitsavos, U.N. Das, C. Stefanadis, Adherence to the Mediterranean diet attenuates inflammation and coagulation process in healthy adults: the ATTICA Study, J. Am. Coll. Cardiol. 44 (2004) 152–158.

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

131

[84] D. Giugliano, K. Esposito, Mediterranean diet and metabolic diseases, Curr. Opin. Lipidol. 19 (2008) 63–68. [85] R. Estruch, M.A. Martinez-Gonzalez, D. Corella, J. Salas-Salvado, V. Ruiz-Gutierrez, M.I. Covas, et al., EVects of a Mediterranean-style diet on cardiovascular risk factors: a randomized trial, Ann. Intern. Med. 145 (2006) 1–11. [86] M. Visser, L.M. Bouter, G.M. McQuillan, M.H. Wener, T.B. Harris, Elevated C‐reactive protein levels in overweight and obese adults, J. Am. Med. Assoc. 282 (1999) 2131–2135. [87] E. Selvin, N.P. Paynter, T.P. Erlinger, The eVect of weight loss on C‐reactive protein: a systematic review, Arch. Intern. Med. 167 (2007) 31–39. [88] R. Cancello, C. Henegar, N. Viguerie, S. Taleb, C. Poitou, C. Rouault, et al., Reduction of macrophage infiltration and chemoattractant gene expression changes in white adipose tissue of morbidly obese subjects after surgery-induced weight loss, Diabetes 54 (2005) 2277–2286. [89] L.K. Heilbronn, M. Noakes, P.M. Clifton, Energy restriction and weight loss on very-lowfat diets reduce C‐reactive protein concentrations in obese, healthy women, Arterioscler. Thromb. Vasc. Biol. 21 (2001) 968–970. [90] G.S. Hotamisligil, Inflammation and metabolic disorders, Nature 444 (2006) 860–867. [91] V. Mohamed-Ali, S. Goodrick, A. Rawesh, D.R. Katz, J.M. Miles, J.S. Yudkin, et al., Subcutaneous adipose tissue releases interleukin‐6, but not tumor necrosis factor-alpha, in vivo, J. Clin. Endocrinol. Metab. 82 (1997) 4196–4200. [92] J.N. Fain, A.K. Madan, M.L. Hiler, P. Cheema, S.W. Bahouth, Comparison of the release of adipokines by adipose tissue, adipose tissue matrix, and adipocytes from visceral and subcutaneous abdominal adipose tissues of obese humans, Endocrinology 145 (2004) 2273–2282. [93] S.P. Weisberg, D. McCann, M. Desai, M. Rosenbaum, R.L. Leibel, A.W. Ferrante Jr., Obesity is associated with macrophage accumulation in adipose tissue, J. Clin. Invest. 112 (2003) 1796–1808. [94] N. Ouchi, S. Kihara, T. Funahashi, T. Nakamura, M. Nishida, M. Kumada, et al., Reciprocal association of C‐reactive protein with adiponectin in blood stream and adipose tissue, Circulation 107 (2003) 671–674. [95] S. Cinti, G. Mitchell, G. Barbatelli, I. Murano, E. Ceresi, E. Faloia, et al., Adipocyte death defines macrophage localization and function in adipose tissue of obese mice and humans, J. Lipid. Res. 46 (2005) 2347–2355. [96] C.A. Curat, V. Wegner, C. Sengenes, A. Miranville, C. Tonus, R. Busse, et al., Macrophages in human visceral adipose tissue: increased accumulation in obesity and a source of resistin and visfatin, Diabetologia 49 (2006) 744–747. [97] G. Charriere, B. Cousin, E. Arnaud, M. Andre, F. Bacou, L. Penicaud, et al., Preadipocyte conversion to macrophage. Evidence of plasticity, J. Biol. Chem. 278 (2003) 9850–9855. [98] Y. Ma, J.A. GriYth, L. Chasan-Taber, B.C. Olendzki, E. Jackson, E.J. Stanek, 3rd, et al., Association between dietary fiber and serum C‐reactive protein, Am. J. Clin. Nutr. 83 (2006) 760–766. [99] D.E. King, B.M. Egan, R.F. Woolson, A.G. Mainous, 3rd, Y. Al-Solaiman, A. Jesri, EVect of a high-fiber diet vs a fiber-supplemented diet on C‐reactive protein level, Arch. Intern. Med. 167 (2007) 502–506. [100] S. Liu, Whole-grain foods, dietary fiber, and type 2 diabetes: searching for a kernel of truth, Am. J. Clin. Nutr. 77 (2003) 527–529. [101] D.E. King, A.G. Mainous, 3rd, B.M. Egan, R.F. Woolson, M.E. Geesey, EVect of psyllium fiber supplementation on C‐reactive protein: the trial to reduce inflammatory markers (TRIM), Ann. Fam. Med. 6 (2008) 100–106.

132

CARITA M. EKLUND

[102] D.E. King, Dietary fiber, inflammation, and cardiovascular disease, Mol. Nutr. Food Res. 49 (2005) 594–600. [103] T. Kishimoto, Interleukin‐6: from basic science to medicine—40 years in immunology, Annu. Rev. Immunol. 23 (2005) 1–21. [104] A. Koj, Initiation of acute phase response and synthesis of cytokines, Biochim. Biophys. Acta 1317 (1996) 84–94. [105] G. Cox, J. Gauldie, Interleukin‐6, in: D. Remick, J. Friedland (Eds.), Cytokines in Health and Disease. Marcel Dekker, Inc., New York. [106] U. Renner, E.C. De Santana, J. Gerez, B. Frohlich, M. Haedo, M.P. Pereda, et al., Intrapituitary expression and regulation of the gp130 cytokine interleukin‐6 and its implication in pituitary physiology and pathophysiology, Ann. N. Y. Acad. Sci. 1153 (2009) 89–97. [107] A. Kariagina, D. Romanenko, S.G. Ren, V. Chesnokova, Hypothalamic-pituitary cytokine network, Endocrinology 145 (2004) 104–112. [108] M. Peters, A.M. Muller, S. Rose-John, Interleukin‐6 and soluble interleukin‐6 receptor: direct stimulation of gp130 and hematopoiesis, Blood 92 (1998) 3495–3504. [109] M. Hibi, M. Murakami, M. Saito, T. Hirano, T. Taga, T. Kishimoto, Molecular cloning and expression of an IL‐6 signal transducer, gp130, Cell 63 (1990) 1149–1157. [110] T. Hirano, T. Taga, K. Yamasaki, T. Matsuda, B. Tang, A. Muraguchi, et al., A multifunctional cytokine (IL‐6/BSF‐2) and its receptor, Int. Arch. Allergy Appl. Immunol. 88 (1989) 29–33. [111] K. Yamasaki, T. Taga, Y. Hirata, H. Yawata, Y. Kawanishi, B. Seed, et al., Cloning and expression of the human interleukin‐6 (BSF‐2/IFN beta 2) receptor, Science 241 (1988) 825–828. [112] T. Kishimoto, S. Akira, M. Narazaki, T. Taga, Interleukin‐6 family of cytokines and gp130, Blood 86 (1995) 1243–1254. [113] T. Nishikawa, K. Hagihara, S. Serada, T. Isobe, A. Matsumura, J. Song, et al., Transcriptional complex formation of C‐Fos, STAT3, and hepatocyte NF‐1 alpha is essential for cytokine-driven C‐reactive protein gene expression, J. Immunol. 180 (2008) 3492–3501. [114] K. Kato, T. Yokoi, N. Takano, H. Kanegane, A. Yachie, T. Miyawaki, et al., Detection by in situ hybridization and phenotypic characterization of cells expressing IL‐6 mRNA in human stimulated blood, J. Immunol. 144 (1990) 1317–1322. [115] D.P. Ramji, P. Foka, CCAAT/enhancer-binding proteins: structure, function and regulation, Biochem. J. 365 (2002) 561–575. [116] C.A. Dinarello, Biologic basis for interleukin‐1 in disease, Blood 87 (1996) 2095–2147. [117] M. Tocci, J. Schmidt, Interleukin‐1: structure and function, in: D. Remick, J. Friedland (Eds.), Cytokines in health and disease, (1997) Marcel Dekker, Inc., New York. [118] C. Gabay, I. Kushner, Acute-phase proteins and other systemic responses to inflammation, N. Engl. J. Med. 340 (1999) 448–454. [119] C.A. Dinarello, Blocking IL‐1 in systemic inflammation, J. Exp. Med. 201 (2005) 1355–1359. [120] S. Endres, J.G. Cannon, R. Ghorbani, R.A. Dempsey, S.D. Sisson, G. Lonnemann, et al., In vitro production of IL 1 beta, IL 1 alpha, TNF and IL2 in healthy subjects: distribution, eVect of cyclooxygenase inhibition and evidence of independent gene regulation, Eur. J. Immunol. 19 (1989) 2327–2333. [121] G. Lonnemann, S. Endres, J.W. Van der Meer, J.G. Cannon, K.M. Koch, C.A. Dinarello, DiVerences in the synthesis and kinetics of release of interleukin 1 alpha, interleukin 1 beta and tumor necrosis factor from human mononuclear cells, Eur. J. Immunol. 19 (1989) 1531–1536.

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

133

[122] C.A. Dinarello, S.M. WolV, The role of interleukin‐1 in disease, N. Engl. J. Med. 328 (1993) 106–113. [123] Y. Kawaguchi, E. Nishimagi, A. Tochimoto, M. Kawamoto, Y. Katsumata, M. Soejima, et al., Intracellular IL‐1alpha‐binding proteins contribute to biological functions of endogenous IL‐1alpha in systemic sclerosis fibroblasts, Proc. Natl. Acad. Sci. USA 103 (2006) 14501–14506. [124] A. MacKenzie, H.L. Wilson, E. Kiss-Toth, S.K. Dower, R.A. North, A. Surprenant, Rapid secretion of interleukin‐1beta by microvesicle shedding, Immunity 15 (2001) 825–835. [125] K.A. Hogquist, E.R. Unanue, D.D. Chaplin, Release of IL‐1 from mononuclear phagocytes, J. Immunol. 147 (1991) 2181–2186. [126] D. Ferrari, C. Pizzirani, E. Adinolfi, R.M. Lemoli, A. Curti, M. Idzko, et al., The P2X7 receptor: a key player in IL‐1 processing and release, J. Immunol. 176 (2006) 3877–3883. [127] D. Brough, N.J. Rothwell, Caspase‐1‐dependent processing of pro-interleukin‐1beta is cytosolic and precedes cell death, J. Cell Sci. 120 (2007) 772–781. [128] J.E. Sims, M.A. Gayle, J.L. Slack, M.R. Alderson, T.A. Bird, J.G. Giri, et al., Interleukin 1 signaling occurs exclusively via the type I receptor, Proc. Natl. Acad. Sci. USA 90 (1993) 6155–6159. [129] H. Wesche, C. Korherr, M. Kracht, W. Falk, K. Resch, M.U. Martin, The interleukin‐1 receptor accessory protein (IL‐1RAcP) is essential for IL‐1‐induced activation of interleukin‐1 receptor-associated kinase (IRAK) and stress-activated protein kinases (SAP kinases), J. Biol. Chem. 272 (1997) 7727–7731. [130] F. Colotta, F. Re, M. Muzio, R. Bertini, N. Polentarutti, M. Sironi, et al., Interleukin‐1 type II receptor: a decoy target for IL‐1 that is regulated by IL‐4, Science 261 (1993) 472–475. [131] X. Li, J. Qin, Modulation of Toll-interleukin 1 receptor mediated signaling, J. Mol. Med. 83 (2005) 258–266. [132] F. Kramer, J. Torzewski, J. Kamenz, K. Veit, V. Hombach, J. Dedio, et al., Interleukin‐1 beta stimulates acute phase response and C‐reactive protein synthesis by inducing an NFkappaB- and C/EBPbeta-dependent autocrine interleukin‐6 loop, Mol. Immunol. 45 (2008) 2678–2689. [133] H. Zheng, D. Fletcher, W. Kozak, M. Jiang, K.J. Hofmann, C.A. Conn, et al., Resistance to fever induction and impaired acute-phase response in interleukin‐1 beta‐deficient mice, Immunity 3 (1995) 9–19. [134] M.K. Ganapathi, L.T. May, D. Schultz, A. Brabenec, J. Weinstein, P.B. Sehgal, et al., Role of interleukin‐6 in regulating synthesis of C‐reactive protein and serum amyloid A in human hepatoma cell lines, Biochem. Biophys. Res. Commun. 157 (1988) 271–277. [135] U. Ganter, R. Arcone, C. Toniatti, G. Morrone, G. Ciliberto, Dual control of C‐reactive protein gene expression by interleukin‐1 and interleukin‐6, EMBO J. 8 (1989) 3773–3779. [136] D.M. Steel, A.S. Whitehead, Heterogeneous modulation of acute-phase-reactant mRNA levels by interleukin‐1 beta and interleukin‐6 in the human hepatoma cell line PLC/PRF/5, Biochem. J. 277 (Pt 2) (1991) 477–482. [137] N.D. Goldman, T. Liu, K.J. Lei, Structural analysis of the locus containing the human C‐reactive protein gene and its related pseudogene, J. Biol. Chem. 262 (1987) 7001–7005. [138] S.H. Yap, H.J. Moshage, B.P.C. Hazenberg, H.M.J. Roelofs, J. Bijzet, P.C. Limburg, et al., Tumor necrosis factor (TNF) inhibits interleukin (IL)‐1 and/or IL‐6 stimulated synthesis of C‐reactive protein (CRP) and serum amyloid A (SAA) in primary cultures of human hepatocytes, Biochim. Biophy. Acta Mol. Cell Res. 1091 (1991) 405–408. [139] W.P. Arend, G. Palmer, C. Gabay, IL‐1, IL‐18, and IL‐33 families of cytokines, Immunol. Rev. 223 (2008) 20–38.

134

CARITA M. EKLUND

[140] G. Wakabayashi, J.A. Gelfand, J.F. Burke, R.C. Thompson, C.A. Dinarello, A specific receptor antagonist for interleukin 1 prevents Escherichia coli-induced shock in rabbits, FASEB J. 5 (1991) 338–343. [141] C. Gabay, J. Gigley, J. Sipe, W.P. Arend, G. Fantuzzi, Production of IL‐1 receptor antagonist by hepatocytes is regulated as an acute-phase protein in vivo, Eur. J. Immunol. 31 (2001) 490–499. [142] K.J. Tracey, Tumor necrosis factor, in: D. Remick, J. Friedland (Eds.), Cytokines in Health and Disease, (1997) Marcel Dekker, Inc., New York. [143] M. Bianchi, O. Bloom, T. Raabe, P.S. Cohen, J. Chesney, B. Sherry, et al., Suppression of proinflammatory cytokines in monocytes by a tetravalent guanylhydrazone, J. Exp. Med. 183 (1996) 927–936. [144] Z. Yao, S.L. Painter, W.C. Fanslow, D. Ulrich, B.M. MacduV, M.K. Spriggs, et al., Human IL‐17: a novel cytokine derived from T cells, J. Immunol. 155 (1995) 5483–5486. [145] C.L. Langrish, Y. Chen, W.M. Blumenschein, J. Mattson, B. Basham, J.D. Sedgwick, et al., IL‐23 drives a pathogenic T cell population that induces autoimmune inflammation, J. Exp. Med. 201 (2005) 233–240. [146] L.E. Harrington, R.D. Hatton, P.R. Mangan, H. Turner, T.L. Murphy, K.M. Murphy, et al., Interleukin 17‐producing CD4þ eVector T cells develop via a lineage distinct from the T helper type 1 and 2 lineages, Nat. Immunol. 6 (2005) 1123–1132. [147] H. Park, Z. Li, X.O. Yang, S.H. Chang, R. Nurieva, Y.H. Wang, et al., A distinct lineage of CD4 T cells regulates tissue inflammation by producing interleukin 17, Nat. Immunol. 6 (2005) 1133–1141. [148] F. Fossiez, O. Djossou, P. Chomarat, L. Flores-Romo, S. Ait-Yahia, C. Maat, et al., T cell interleukin‐17 induces stromal cells to produce proinflammatory and hematopoietic cytokines, J. Exp. Med. 183 (1996) 2593–2603. [149] T.A. Moseley, D.R. Haudenschild, L. Rose, A.H. Reddi, Interleukin‐17 family and IL‐17 receptors, Cytokine Growth Factor Rev. 14 (2003) 155–174. [150] T. Shalom-Barak, J. Quach, M. Lotz, Interleukin‐17‐induced gene expression in articular chondrocytes is associated with activation of mitogen-activated protein kinases and NF-kappaB, J. Biol. Chem. 273 (1998) 27467–27473. [151] J. Martel-Pelletier, F. Mineau, D. Jovanovic, J.A. Di Battista, J.P. Pelletier, Mitogenactivated protein kinase and nuclear factor kappaB together regulate interleukin‐17‐ induced nitric oxide production in human osteoarthritic chondrocytes: possible role of transactivating factor mitogen-activated protein kinase-activated proten kinase (MAPKAPK), Arthritis Rheum. 42 (1999) 2399–2409. [152] D.N. Patel, C.A. King, S.R. Bailey, J.W. Holt, K. Venkatachalam, A. Agrawal, et al., Interleukin‐17 stimulates C‐reactive protein expression in hepatocytes and smooth muscle cells via p38 MAPK and ERK1/2‐dependent NF‐kappaB and C/EBPbeta activation, J. Biol. Chem. 282 (2007) 27229–27238. [153] H.E. Broxmeyer, Is interleukin 17, an inducible cytokine that stimulates production of other cytokines, merely a redundant player in a sea of other biomolecules? J. Exp. Med. 183 (1996) 2411–2415. [154] S.P. Li, N.D. Goldman, Regulation of human C‐reactive protein gene expression by two synergistic IL‐6 responsive elements, Biochemistry 35 (1996) 9060–9068. [155] C. Toniatti, A. Demartis, P. Monaci, A. Nicosia, G. Ciliberto, Synergistic trans-activation of the human C‐reactive protein promoter by transcription factor HNF‐1 binding at two distinct sites, EMBO J. 9 (1990) 4467–4475. [156] B. Voleti, A. Agrawal, Regulation of basal and induced expression of C‐reactive protein through an overlapping element for OCT‐1 and NF‐kappaB on the proximal promoter, J. Immunol. 175 (2005) 3386–3390.

PROINFLAMMATORY CYTOKINES IN CRP BASELINE REGULATION

135

[157] A. Agrawal, H. Cha-Molstad, D. Samols, I. Kushner, Transactivation of C‐reactive protein by IL‐6 requires synergistic interaction of CCAAT/enhancer binding protein beta (C/EBP beta) and Rel p50, J. Immunol. 166 (2001) 2378–2384. [158] H. Cha-Molstad, A. Agrawal, D. Zhang, D. Samols, I. Kushner, The Rel family member P50 mediates cytokine-induced C‐reactive protein expression by a novel mechanism, J. Immunol. 165 (2000) 4592–4597. [159] A. Agrawal, D. Samols, I. Kushner, Transcription factor C‐Rel enhances C‐reactive protein expression by facilitating the binding of C/EBPbeta to the promoter, Mol. Immunol. 40 (2003) 373–380. [160] Z. Cao, R.M. Umek, S.L. McKnight, Regulated expression of three C/EBP isoforms during adipose conversion of 3T3‐L1 cells, Genes Dev. 5 (1991) 1538–1552. [161] S. Akira, H. Isshiki, T. Sugita, O. Tanabe, S. Kinoshita, Y. Nishio, et al., A nuclear factor for IL‐6 expression (NF-IL6) is a member of a C/EBP family, EMBO J. 9 (1990) 1897–1906. [162] V. Poli, The role of C/EBP isoforms in the control of inflammatory and native immunity functions, J. Biol. Chem. 273 (1998) 29279–29282. [163] D.P. Young, I. Kushner, D. Samols, Binding of C/EBPbeta to the C‐reactive protein (CRP) promoter in Hep3B cells is associated with transcription of CRP mRNA, J. Immunol. 181 (2008) 2420–2427. [164] B. Majello, R. Arcone, C. Toniatti, G. Ciliberto, Constitutive and IL‐6‐induced nuclear factors that interact with the human C‐reactive protein promoter, EMBO J. 9 (1990) 457–465. [165] D. Zhang, M. Sun, D. Samols, I. Kushner, STAT3 participates in transcriptional activation of the C‐reactive protein gene by interleukin‐6, J. Biol. Chem. 271 (1996) 9503–9509. [166] P.P. Singh, B. Voleti, A. Agrawal, A novel RBP-J kappa-dependent switch from C/EBP beta to C/EBP zeta at the C/EBP binding site on the C‐reactive protein promoter, J. Immunol. 178 (2007) 7302–7309. [167] P.M. Ridker, G. Pare, A. Parker, R.Y. Zee, J.S. Danik, J.E. Buring, et al., Loci related to metabolic-syndrome pathways including LEPR, HNF1A, IL6R, and GCKR associate with plasma C‐reactive protein: the Women’s Genome Health Study, Am. J. Hum. Genet. 82 (2008) 1185–1192. [168] H. Berrahmoune, B. Herbeth, G. Siest, S. Visvikis-Siest, Heritability of serum hs-CRP concentration and 5‐year changes in the Stanislas family study: association with apolipoprotein E alleles, Genes Immun. 8 (2007) 352–359. [169] P.T. Johnson, K.E. Betts, M.J. Radeke, G.S. Hageman, D.H. Anderson, L.V. Johnson, Individuals homozygous for the age-related macular degeneration risk-conferring variant of complement factor H have elevated levels of CRP in the choroid, Proc. Natl. Acad. Sci. USA 103 (2006) 17456–17461. [170] E. Fisher, M.B. Schulze, N. Stefan, H.U. Haring, F. Doring, H.G. Joost, et al., Association of the FTO rs9939609 Single nucleotide polymorphism with C‐reactive protein levels, Obesity (Silver Spring) 17 (2009) 330–334. [171] C. Eklund, M. Kivimaki, M.S. Islam, M. Juonala, M. Kahonen, J. Marniemi, et al., C-reactive protein genetics is associated with carotid artery compliance in men in The Cardiovascular Risk in Young Finns Study, Atherosclerosis 196 (2008) 841–848. [172] P. Berger, J.P. McConnell, M. Nunn, K.S. Kornman, J. Sorrell, K. Stephenson, et al., C-reactive protein levels are influenced by common IL-1 gene variations, Cytokine 17 (2002) 171–174. [173] J. Rogus, J.D. Beck, S. Offenbacher, K. Huttner, L. Iacoviello, M.C. Latella, et al., IL1B gene promoter haplotype pairs predict clinical levels of interleukin-1beta and C-reactive protein, Hum. Genet. 123 (2008) 387–398.

136

CARITA M. EKLUND

[174] C. Eklund, F. Jahan, T. Pessi, T. Lehtimaki, M. Hurme, Interleukin 1B gene polymorphism is associated with baseline C-reactive protein levels in healthy individuals, Eur. Cytokine Netw. 14 (2003) 168–171. [175] G. Latkovskis, N. Licis, U. Kalnins, C-reactive protein levels and common polymorphisms of the interleukin-1 gene cluster and interleukin-6 gene in patients with coronary heart disease, Eur. J. Immunogenet. 31 (2004) 207–213. [176] F. D’Aiuto, M. Parkar, P.M. Brett, D. Ready, M.S. Tonetti, Gene polymorphisms in proinflammatory cytokines are associated with systemic inflammation in patients with severe periodontal infections, Cytokine 28 (2004) 29–34. [177] A.P. Reiner, M.M. Wurfel, L.A. Lange, C.S. Carlson, A.S. Nord, C.L. Carty, et al., Polymorphisms of the IL1-receptor antagonist gene (IL1RN) are associated with multiple markers of systemic inflammation, Arterioscler Thromb. Vasc. Biol. 28 (2008) 1407–1412. [178] M.A. Vickers, F.R. Green, C. Terry, B.M. Mayosi, C. Julier, M. Lathrop, et al., Genotype at a promoter polymorphism of the interleukin-6 gene is associated with baseline levels of plasma C-reactive protein, Cardiovasc. Res. 53 (2002) 1029–1034. [179] M. Libra, S.S. Signorelli, Y. Bevelacqua, P.M. Navolanic, V. Bevelacqua, J. Polesel, et al., Analysis of G(-174)C IL-6 polymorphism and plasma concentrations of inflammatory markers in patients with type 2 diabetes and peripheral arterial disease, J. Clin. Pathol. 59 (2006) 211–215. [180] J.D. Walston, M.D. Fallin, M. Cushman, L. Lange, B. Psaty, N. Jenny, et al., IL-6 gene variation is associated with IL-6 and C-reactive protein levels but not cardiovascular outcomes in the Cardiovascular Health Study, Hum. Genet. 122 (2007) 485–494. [181] Y. Saijo, E. Yoshioka, T. Fukui, M. Kawaharada, F. Sata, H. Sato, et al., Effects of the Interaction between Interleukin-6–634C/G Polymorphism and Smoking on Serum C-Reactive Protein Concentrations, Hypertens. Res. 30 (2007) 593–599. [182] L.Y. Wong, R.Y. Leung, K.L. Ong, B.M. Cheung, Plasma levels of fibrinogen and C-reactive protein are related to interleukin-6 gene -572C > G polymorphism in subjects with and without hypertension, J. Hum. Hypertens. 21 (2007) 875–882. [183] Y. Jang, O.Y. Kim, Y.J. Hyun, J.S. Chae, S.J. Koh, Y.M. Heo, et al., Interleukin-6–572C > G polymorphism-association with inflammatory variables in Korean men with coronary artery disease, Transl. Res. 151 (2008) 154–161. [184] S.L. Ferrari, L. Ahn-Luong, P. Garnero, S.E. Humphries, S.L. Greenspan, Two promoter polymorphisms regulating interleukin-6 gene expression are associated with circulating levels of C-reactive protein and markers of bone resorption in postmenopausal women, J. Clin. Endocrinol. Metab. 88 (2003) 255–259. [185] L. Qi, N. Rifai, F.B. Hu, Interleukin-6 receptor gene, plasma C-reactive protein, and diabetes risk in women, Diabetes 58 (2009) 275–278. [186] H.M. Lakka, T.A. Lakka, T. Rankinen, T. Rice, D.C. Rao, A.S. Leon, et al., The TNFalpha G-308A polymorphism is associated with C-reactive protein levels: the HERITAGE Family Study, Vascul. Pharmacol. 44 (2006) 377–383.

ADVANCES IN CLINICAL CHEMISTRY, VOL. 48

FETAL SKIN WOUND HEALING Edward P. Buchanan,*,1 Michael T. Longaker,† and H. Peter Lorenz† *Division Plastic Surgery, Department of Surgery, Stanford University School of Medicine, Stanford, California 94305, USA † Division Plastic Surgery, Department of Surgery, Pediatric Surgical Research Laboratory, Stanford University School of Medicine, Stanford, California 94305‐5148, USA

1. Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Fetal Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Fetal ECM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Collagen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Hyaluronic Acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Proteoglycan ECM Modulators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Scarless Fetal Wound Repair Specificity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Scarless Fetal Wound Phenotype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Collagen Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Hyaluronic Acid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. ECM Adhesion Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. ECM Proteoglycan Modulators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6. Scarless Repair is Intrinsic to Fetal Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.7. Scarless Repair Depends on Gestational Age and Wound Size . . . . . . . . . . . . . 4.8. Mechanisms of Scarless Repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Stem Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1. Hematopoietic Stem Cells (HSCs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2. Fetal and Postnatal Epidermal Stem Cells. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3. Dot Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Cellular Inflammatory Mediators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1. Platelets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

138 138 140 140 140 140 141 141 141 141 142 143 143 145 146 146 147 147 147 148 149 149 149

Corresponding author: Edward P. Buchanan, e‐mail: [email protected] 137

0065-2423/09 $35.00 DOI: 10.1016/S0065-2423(09)48006-5

Copyright 2009, Elsevier Inc. All rights reserved.

138

BUCHANAN ET AL.

6.2. Neutrophils. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3. Fibroblasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Cytokines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1. Transforming Growth Factor‐Beta (TGF‐b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2. Connective Tissue Growth Factor (CTGF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3. Vascular Endothelial Growth Factor (VEGF). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4. Fibroblast Growth Factors (FGFs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5. Platelet Derived Growth Factor (PDGF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6. Wnts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.7. Interleukins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.8. Molecular Control of Scarless Repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.9. Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

150 150 151 151 152 152 153 153 154 154 154 155 155

1. Abstract The developing fetus has the ability to heal wounds by regenerating normal epidermis and dermis with restoration of the extracellular matrix (ECM) architecture, strength, and function. In contrast, adult wounds heal with fibrosis and scar. Scar tissue remains weaker than normal skin with an altered ECM composition. Despite extensive investigation, the mechanism of fetal wound healing remains largely unknown. We do know that early in gestation, fetal skin is developing at a rapid pace and the ECM is a loose network facilitating cellular migration. Wounding in this unique environment triggers a complex cascade of tightly controlled events culminating in a scarless wound phenotype of fine reticular collagen and abundant hyaluronic acid. Comparison between postnatal and fetal wound healing has revealed diVerences in inflammatory response, cellular mediators, cytokines, growth factors, and ECM modulators. Investigation into cell signaling pathways and transcription factors has demonstrated diVerences in secondary messenger phosphorylation patterns and homeobox gene expression. Further research may reveal novel genes essential to scarless repair that can be manipulated in the adult wound and thus ameliorate scar.

2. Introduction Scar tissue is the end result of tissue injury in all human soft tissue organ systems. Cutaneous scar tissue contains diVerent quantities of elastin and collagen than normal skin, making scar functionally diVerent than normal skin [1]. For example, cleft lip and palate repair scars are inelastic and inhibit underlying muscle motion and maxillary bone growth, exposing children to

FETAL SKIN WOUND HEALING

139

numerous lip, palate, alveolar, nasal, pharyngeal, and orthognathic surgical procedures. Scarring is caused by overdeposition of interstitial collagens and diVerentiation of fibroblasts to myofibroblasts, whose cytoskeletal structure has a high contractile function through smooth muscle actin (SMA) function. In addition, pathological fibrosis in internal organs is also characterized by overdeposition of collagens and overexpression of myofibroblasts. Therefore, understanding the mechanisms of scar formation and developing treatments to reduce scarring will have dramatic benefits for human well‐being. Fetal skin is the template for skin regeneration. Fetal full‐thickness skin wounds heal with restoration of normal epidermal and dermal architecture and not with scar formation. The biology responsible for scarless wound healing, a paradigm for ideal tissue repair, is the subject of active investigation. Scarless healing has been confirmed in both animal models (mouse, rat, rabbit, pig, sheep, and monkey) and in human fetuses [2]. The wound dermis has a normal architecture in which the collagen matrix pattern is reticular and unchanged from unwounded dermis. In addition, the hair follicle and sebaceous glands regenerate in the fetal wound. In mice during the scarless time period before E17.5, hair follicles which are not fully developed have no bulge region, the niche where postnatal epidermal stem cells are located. The fetal dermis is more cellular and does not have large interstitial collagen bundles. Fetal skin wounds transition to heal with scar formation late in gestation. The transition period has been defined in fetal mouse, rat, sheep, and primate models of repair. For primate full‐ thickness lip wounds, the transition occurs in the early third trimester of gestation [3]. During the transition, hair follicle and other appendage regeneration is lost, but the wound collagen matrix pattern remains reticular and scarless. Scarring of open wounds begins on E18.5 in mice (full term is E20.5) [4]. In mouse E18.5 skin, hair follicles have begun to develop and bulge area is forming. More fibroblastic cells are in the dermis, and interstitial fibrillar collagen bundles are forming. By postnatal age, the epidermis and dermis have clearly delineated structures. In the adult, cutaneous wound healing restores tissue integrity, but the outcome is fibrosis and scarring instead of normal tissue architecture. In the mature scar, the collagen bundles are packed tightly together in large parallel fibers, unlike the reticular pattern in unwounded skin. Although scar remodeling occurs for months to years after the initial wound event, the scar remains weaker than normal skin [5] and never achieves restoration of normal ECM architecture or function [6]. An overdeposition of collagen occurs in tissue fibrosis and scarring, unlike the restoration of a normal collagen pattern in the scarless fetal skin wound.

140

BUCHANAN ET AL.

3. Development 3.1. FETAL SKIN Fetal skin structure and histology change dramatically with development. The transition from scarless to scarring wound repair occurs in the context of the developing fetal skin. Therefore, investigation of normal skin development may reveal primary regulators of cellular diVerentiation and proliferation that play a role in fetal wound healing. In early gestation, mutually inductive mechanisms between ectoderm and mesoderm stimulate development of the epidermis and dermis. Epidermal primordial cells, derived from ectoderm, proliferate at 7 weeks gestation in the human fetus, forming a squamous layer of periderm and a basal germinative layer. The periderm cells are keratinized, shed, and eventually replaced by the stratum corneum at 21 weeks. The basal germinative layer becomes the stratum germinativum, a source of new cells for dermal appendages, the intermediate layers found in mature skin, and hair germs. Hair follicles begin development in the 9–12th week. Peripheral follicle cells become the epithelial root sheath and surrounding mesenchymal cells form the dermal root sheath. Mesoderm‐derived mesenchymal cells produce collagen and elastic connective tissue fibers of the dermis by 11 weeks. Endothelial‐lined blood vessels form in the dermis and diVerentiate into arteries and veins. Skin maturation with dermal thickening continues into the postnatal period [7]. 3.2. FETAL ECM The important role of the ECM in cell adhesion, diVerentiation, and proliferation has only recently been discovered [8]. In the past, the ECM was regarded as inert scaVolding. We now know it is a dynamic layer of collagen, proteoglycans, and glycosaminoglycans which facilitates cellular migration in the fetus and serves as a reservoir for growth factors. The fetal ECM undergoes a series of changes before reaching the adult phenotype. Fetal ECM diVers from adult ECM in collagen composition, hyaluronic acid (HA) content, and proteoglycan ECM modulators. This may have implications in scarless repair. 3.3. COLLAGEN Collagen is the dominant structural protein in all human connective tissue. Although several types of collagen exist, Type I predominates and is the principal component of both adult and fetal ECM [8]. Its strength is derived from a triple helix configuration of polypeptide chains, which are

FETAL SKIN WOUND HEALING

141

cross‐linked and stabilized by lysyl oxidase. Fetal skin has a higher ratio of type III to type I collagen than adult skin [9]. With maturation, the relative amount of type III collagen in fetal skin diminishes, although the adult phenotype is not seen until the postnatal period [10].

3.4. HYALURONIC ACID HA is a negatively charged, nonsulfated glycosaminoglycan of the ECM found in soluble form or complexed with proteoglycans. Increased HA content in the ECM is noted during rapid cellular migration and angiogenesis [11]. The net negative charges of HA trap and impede water molecules, which allows resistance to deformation and facilitates cellular movement [8]. Fetal skin contains more HA than adult skin [12]. HA stimulates collagen synthesis by fibroblasts in vitro [13].

3.5. PROTEOGLYCAN ECM MODULATORS Proteoglycan ECM modulators decorin, fibromodulin, lysyl oxidase, and matrix metalloproteinases (MMPs) serve a role in collagen synthesis, maturation, and degradation. Decorin production increases by 72% during the transition period in fetal rat skin and continues to increase into the postnatal period achieving a level 300 that of early gestational fibroblasts [14]. Likewise, increased expression of enzymes lysyl oxidase and MMP occurs with development [15, 16]. Fibromodulin, another modulator of collagen fibrillogenesis in the decorin family of proteoglycans, has decreased production with maturation [17]. Fibromodulin binds and inactivates transforming growth factor beta (TGF‐b), a key cytokine implicated in adult wound healing and scar [18]. Fibromodulin has been shown to have an antiscarring eVect during wound repair [17].

4. Scarless Fetal Wound Repair Specificity 4.1. SCARLESS FETAL WOUND PHENOTYPE Phenotypic diVerences in collagen deposition and cross‐linking patterns, HA content, and diVerential expression of proteoglycan ECM modulators and adhesion proteins distinguish fetal wounds from adult wounds [14–20].

142

BUCHANAN ET AL.

4.2. COLLAGEN CONTENT In scarless fetal wounds, collagen is rapidly deposited in a fine reticular pattern indistinguishable from uninjured skin (Figs. 1 and 2). In contrast, adult scarring wounds have disorganized type I collagen bundles with more collagen cross‐linking [19–21] (Figs. 3 and 4). Type I collagen and the molecular chaperone heat shock protein 47 (HSP 47) both are increased in adult rat wounds, as shown by reverse transcriptase‐polymerase chain reaction (RT‐ PCR). In contrast, fetal wounds show no diVerence in collagen I production or HSP 47 expression [22]. Lovvorn et al. implanted PVA sponges in fetal sheep wounds and noted increased collagen cross‐linking with advancing gestational age that paralleled the transition from scarless to scar‐forming repair [23]. Although Type I collagen cross‐linking is essential for adult wound healing and strength, its rigidity may impede the movement of cellular mediators required for rapid cellular regeneration in the fetus. A

B

50mm

100mm C

D

100mm

50mm

FIG. 1. Scarless healing of E16 fetal wounds (H&E Stain). Black arrows indicate India ink tattoo made at the time of wounding in order to demonstrate scarless wound location. Healed wounds (A and C) at 72 h (100). The epidermal appendage (developing hair follicles) pattern shows numerous appendages directly in the healed wound. Magnified views of the same wounds (B and D) showing epidermal appendages (open arrows) within the wound site (200). No inflammatory infiltrate is present. Reproduced with Permission from Lippincot Williams @ Wilkins. Original Publication: Ref. [21].

FETAL SKIN WOUND HEALING

A

50 mm

143

B

10 mm

C

10 mm FIG. 2. Scarless healing of E16 fetal wounds (Confocal Microscopy). Collagen fibers are stained with sirius red and appear white. (A) Healed wound harvested at 72 h (200). The epidermis is thickened at the wound site (arrow). The collagen fiber is reticular and unchanged from the surrounding dermis. (B) Healed wound harvested at 72 h under a higher magnification (1000). The collagen fibers are thin and closely approximating each other with little interfiber space. The fibers are arranged in a wispy reticular pattern. (C) Non‐wounded E19 skin at the same magnification as B (1000). The dermal collagen fiber pattern is identical to B. Reproduced with Permission from Lippincot Williams @ Wilkins. Original Publication: Ref. [21].

4.3. HYALURONIC ACID The HA content of scarless fetal wounds increases more rapidly, is more sustained, and is overall greater than that of adult wounds [12]. Fetal wounds have greater HA stimulating activity and fewer proinflammatory cytokines, such as IL‐1 and TNF‐alpha, that downregulate HA expression [24]. 4.4. ECM ADHESION PROTEINS Scarless fetal wounds are characterized by a more rapid upregulation of ECM adhesion proteins and diVerential expression of cell surface receptors (integrins). The adhesion protein, fibronectin, mediates cellular attachment to the ECM and attracts fibroblasts, keritinocytes, and endothelial cells to

144

BUCHANAN ET AL.

A

B

50 mm

100 mm C

D

100 mm

50 mm

FIG. 3. Scar formation after transition point in E18 fetal wounds (H&E Stain). Black arrows indicate green vital dye tattoo made at the time of wounding. (A) Wound at 24 h. The wound remains open and an inflammatory cell infiltrate is present (open arrow) (100). (B) Magnified view (200) of the incompletely healed wound at 24 h shown in A. (C) Wound at 72 h. No epidermal appendages are present, consistent with adult‐type repair and scar formation (100). (D) Magnified view (200) of wound shown in C, demonstrating the lack of dermal appendages and an inflammatory cell infiltrate. Reproduced with Permission from Lippincot Williams @ Wilkins. Original Publication: Ref. [21].

the site of injury [11]. Early gestation fetal rabbit wounds express fibronectin 4 h after wounding while its expression is not seen until 12 h in the adult [25]. Whitby et al. found no diVerence in the onset of fibronectin expression, but noted more sustained expression in the adult [20]. Tenascin blocks fibronectin‐mediated cellular attachment. In upper lip wounds of mice, tenascin appears more rapidly in the fetus (1 h) compared to the adult (24 h) and precedes cellular migration [20]. This suggests a role for tenascin in the rapid closure of fetal wounds. Collagen integrin receptors in fetal fibroblasts are diVerentially expressed with increasing gestational age. Fetal fibroblasts have increased alpha 2 integrin subunit expression and decreased alpha 1 and 3 integrin subunit expression compared to the adult. This correlates with the low capacity of fetal fibroblasts to contract a collagen

FETAL SKIN WOUND HEALING

A

145

B

50 mm

10 mm

C

10 mm

FIG. 4. Scar formation after transition point in E18 wounds (Confocal Microscopy). Collagen fibers are stained with sirius red and appear white. (A) Healed wound at 72 h (200). The wound dermal collagen pattern (open arrows) is diVerent from the surrounding nonwounded dermis (striped arrow). The fibers are less densely compacted. No epidermal appendages are present. Neo‐vascularization is shown with the white arrows. (B) Healed wound at 72 h at a higher magnification (1000). The collagen fibers are thicker, but with greater interfiber spaces compared to nonwounded dermis. (C) Nonwounded skin at E21 days gestational age (1000). When compared to wound collagen fibers (B), nonwounded dermal collagen fibers are thinner with less interfiber space. Reproduced with Permission from Lippincot Williams @ Wilkins. Original Publication: Ref. [21].

gel and may have implications in the diVerences seen between fetal and adult wound contraction [26]. 4.5. ECM PROTEOGLYCAN MODULATORS Regulators of collagen organization and degradation influence the ECM architecture. Decorin, a modulator of collagen fibrillogenesis, shows no change in fetal wounds but is upregulated in adult wounds [14]. Fibromodulin facilitated cellular migration is downregulated in adult wounds and unchanged in the fetal wound. [17, 27]. This may prove useful as a marker of wound phenotype—if exogenous factors decrease scarring, they may

146

BUCHANAN ET AL.

decrease decorin and increase fibromodulin. Lysyl oxidase expression is greater during adult wound repair and has been implicated in fibrotic diseases [15]. MMPs and tissue‐derived inhibitors (TIMPs) function in ECM turnover. Levels of MMP1 and MMP9 are increased in scarless wounds while scarring wounds have downregulation of MMP2 and contain greater TIMP expression. Overall, scarless wounds have a higher ratio of MMP to TIMP expression, likely favoring remodeling and less accumulation of collagen [16]. 4.6. SCARLESS REPAIR IS INTRINSIC TO FETAL SKIN The capacity for scarless repair was initially attributed to the sterile intrauterine environment. Amniotic fluid is rich in HA and growth factors but devoid of bacteria and inflammatory stimulators, thus it was thought to be permissive for scarless repair. However, early studies demonstrated that the intrauterine environment is neither essential nor suYcient for scarless repair. Fetal marsupials develop outside the uterus in a maternal pouch and heal cutaneous wounds without scar [28]. Adult sheep skin transplanted onto the backs of fetal sheep bathed in the amniotic fluid of the intrauterine environment healed incisional wounds with scar while adjacent incisional wounds in fetal skin healed without scars [29]. Fetal scarless repair is also organ‐specific. At time points early in gestation where fetal skin heals without scar, fetal stomach, intestine, and diaphragm heal with scar formation [30, 31]. This suggests certain subpopulations of cells in skin modulate the local wound healing response. Further evidence implicates fetal dermal cells as the eVector cell responsible for scarless repair. Human fetal skin from 15 to 22 weeks gestation was transplanted subcutaneously and cutaneously onto the backs of athymic adult mice. In this adult system, wounds created in the subcutaneous fetal grafts healed without scars with human collagen from fetal fibroblasts. Conversely, wounds made in the gestationally equivalent cutaneous fetal grafts healed with scar composed of mouse collagen from adult fibroblasts [32]. 4.7. SCARLESS REPAIR DEPENDS ON GESTATIONAL AGE AND WOUND SIZE There is a developmentally regulated threshold for scarless healing based on gestational age and the extent of injury. The ontogenetic transition of rat skin has been defined in an organ culture system and confirmed in vivo with confocal microscopic analysis [21, 33]. This transition point lies between days 16.5 and 18.5 of gestation (Term ¼ 21.5 d). In a human fetal skin model, the transition point occurs after 24 weeks of gestation [32]. Wound size modulates the transition point. In fetal lambs, increasing wound size increased the frequency of scarring at a gestational age when smaller wounds healed

FETAL SKIN WOUND HEALING

147

without scars [34]. In nonhuman primates, the transition from scarless to scarring repair has been shown to proceed through an intermediate wound phenotype. Fetal monkey lip incisional wounds heal with restoration of normal epidermal appendage and dermal collagen architecture in midgestation. At the start of the third trimester, these wounds do not restore epidermal appendage (hair follicle and sebaceous gland) architecture, but still heal with a normal collagen dermal pattern. Thus, a ‘‘transition wound’’ phenotype occurs. By the mid‐third trimester, the wounds heal with a typical scar pattern—no appendages and collagen scar [35]. 4.8. MECHANISMS OF SCARLESS REPAIR Fetal wounds heal rapidly with a paucity of inflammatory cells. This key observation has stimulated interest in the role of cellular inflammatory mediators, cytokines, and growth factors in fetal wound healing. We know that in the postnatal animal, disruption of tissue integrity stimulates platelet activation, cytokine production, and chemotaxis of macrophages and neutrophils [36]. However, scarless wounds are characterized by a relative lack of inflammation [11]. Furthermore, introduction of inflammation into normally scarless wounds produces dose‐dependent increases in wound macrophages, neutrophils, collagen deposition, and scarring [37]. This suggests an important role of inflammation in scar formation. The inflammatory cell infiltrate likely disrupts the inherent ability of resident dermal cells to restore the dermis with minimal scarring.

5. Stem Cells 5.1. HEMATOPOIETIC STEM CELLS (HSCS) Bone marrow (BM)‐ or blood‐derived HSCs participate in tissue regeneration and development [38, 104]. Unfractionated BM cells can regenerate myocytes, neurons, hepatocytes, smooth muscle cells, and other tissues [39–42], indicating the presence of stem cells in BM. However, the true BM stem cell with pluri‐potential is still unknown. Most researchers believe that BM contains two groups of stem cells: (1) HSCs which express c‐kitþ lin sca‐1þ and diVerentiate into all blood cell types [8, 43–45] stromal stem cells that can develop into bone, fat, muscle, and cartilage. Stromal stem cells include multipotent adult progenitor cells (MAPCs) that express CD34 CD44 CD45 c‐kit [46], marrow‐isolated adult multilineage inducible (MIAMI) cells that express CD29, CD63, Cd81, CD122, and CD164, but not Cd34, Cd36, Cd45, and c‐kit [47], unrestricted somatic stem cells

148

BUCHANAN ET AL.

(USSC) that express CD34low CD45 c‐kitlow [48], and amniotic fluid‐ derived stem (AFS) cells that express similar surface markers as USSCs [49]. Isolation of stromal stem cells that have multilineage potential has been diYcult, despite years of attempts, and the exact cell types remain unclear and controversial [49, 50]. HSCs are thought to traYc from BM to skin wounds during repair, but they are diYcult to identify in the overlying inflammatory cell infiltrate. 5.2. FETAL AND POSTNATAL EPIDERMAL STEM CELLS The stem cell niche is the environment where stem cells reside. Postnatal epidermal stem cells from aged mice show similar plasticity compared to neonatal epidermal stem cells [51]. During postnatal scar formation, epidermal continuity is restored by epidermal stem cell proliferation [52]. The niche for postnatal epidermal stem cells remains under investigation. Postnatal epidermal stem cells are thought to be located in two niches: adjacent to the epidermal basement membrane and within the bulge region of hair follicles. The regeneration of the postnatal epidermis through basal epidermal stem cell proliferation is well documented [52]. Hair follicle formation begins during development when dermal cells send signals to epithelial stem cells [53]. Recently, one epidermal stem cell niche in adult skin has been defined to the hair bulge area [54]. But Levy et al. have shown that epidermal stem cells from the bulge do not participate in the renewal of the epidermis during normal homeostasis [55]. Cell surface markers that identify epidermal stem cells include integrin b1 [56], CD34 [57], integrin a6 [54], P63 [58], and keratin19. Integrin b1 was the first epithelial stem cell marker described [56]. Integrin a6 [54] and CD34 [57] mark epidermal stem cells in the bulge area. E‐cadherin [59] is another postnatal epidermal stem cell marker. The functional importance of E‐cadherin is underscored by the observation that E‐cadherin null murine embryos fail to form epithelium [60]. CD34 and P63 have also been identified as HSC markers [58, 61]. P63 has an inhibitory eVect on keratinocyte diVerentiation [58]. Epidermal stem cells isolated from adult mouse skin do not coexpress the HSC marker Sca‐1 [62]. Also, epidermal stem cells that have multilineage diVerentiation capability do not express CD34 nor Sca1 [63]. Little information about the stem cell type and niche, neither in fetal skin nor its function during scarless fetal wound healing, is known. Since the structure of fetal skin is diVerent from the adult skin, that is, no mature hair follicles and no bulges are present in fetal skin, the origin and location of stem cells in fetal skin are likely diVerent from those in adult skin. We hypothesize that fetal skin stem cell function is a key mechanism underlying scarless healing. Although extensive research has been done on the postnatal epithelial stem cells, little information is known about dermal‐derived

FETAL SKIN WOUND HEALING

149

stem cells. Fernandes et al. have found a group of multipotent adult skin‐ derived stem cells (SKPs), whose niche is in the papilla of hair follicles [64]. These stem cells can diVerentiate to neurons, glial, smooth muscle cells, and adipocytes. However, a recent report indicates that fetal mouse E16.5 dermal‐derived cells become epithelial stem cells and express keratins and E‐cadherin but not c‐Kit [65]. This information indicates that the lineage potential of dermal‐derived stem cells may depend on the age when the cell is isolated from skin. In addition, no specific markers for dermal stem cells have been identified. 5.3. DOT CELLS A new stem cell , which we detected in fetal and adult mouse blood, and in the fetal dermis, may be one cell responsible for scarless wound repair. Dot cells, named for their extremely small size, are thought to be primitive cells or stem cells due to their cell surface expression of stem cell markers such as E‐cadherin, integrin b1, and CD 34[66]. They are derived from BM and can be found circulating in the blood. During development in the fetus, they are distributed in many tissues and likely contribute to tissue diVerentiation. Dot cells have a strong aYnity to sites of tissue injury. They migrate to wounds and diVerentiate into dermal cells which release less scar and interstitial collagen. One mechanism of migration of Dot cells maybe due to their cell surface marker, CD184, which the receptor for stromal‐derived factor‐1 (SDF‐1). SDF‐1 is known to regulate the migration of mesenchymal stem cells to sites of injury. The ratio of Dot cells circulating in scarless healing fetal rats is more than 20 times higher than that of scarring postnatal rats. The higher number of circulating Dot cells in scarless healing animals supports their possible function in fetal skin regeneration. Furthermore, after transplantation of Dot cells into wounded adult mice, scarless wound healing occurs. Intravenous injection of labeled Dot cells into wounded adult animals results in skin regeneration [66]. Dot cells fuse to diVerentiate into wound bed cells which reduced scarring and increased hair follicle generation. However, the complete mechanism of scarless wound repair by Dot cells is still unknown. Studies are currently ongoing to elucidate this process. 6. Cellular Inflammatory Mediators 6.1. PLATELETS The absence of an acute inflammatory infiltrate in scarless wounds may be partly explained by decreased fetal platelet degranulation and aggregation in the fetus compared to the adult. Although there is no diVerence in size,

150

BUCHANAN ET AL.

organization, or granule content by transition electron microscopy in fetal compared to adult platelets, fetal platelets produce less platelet‐derived growth factor (PDGF), TGF‐b1, and TGF‐b2 than their adult counterparts [67]. Fetal platelet exposure to collagen in vitro does stimulate growth factor release; however, the platelets still do not aggregate [68]. Olutoye et al. further investigated the aggregation capabilities of adult and fetal porcine platelets after exposure to collagen and ADP. The fetal platelets responded suboptimally to collagen and showed an age‐dependent aggregation response to ADP exposure corresponding with the transition period for cutaneous scarless to scar‐forming wounds [69]. Additionally, HA suppresses aggregation and release of PDGF from fetal platelets in a dose‐dependent fashion, having the greatest eVect in the HA‐rich fetal environment [70]. This reduced function of fetal platelets may be one mechanism of fetal scarless repair. 6.2. NEUTROPHILS Neutrophils neutralize and engulf bacteria. Cytokines TGF‐b1 and PDGF recruit neutrophils to the site of injury. In turn, neutrophils release self‐ stimulating cytokines and chemoattractants for fibroblasts and macrophages [36]. Fewer neutrophils are present in the fetal wound, and an age‐dependent defect in the ability of fetal neutrophils to phagocytose pathogenic bacteria has been demonstrated in fetal sheep [71]. 6.3. FIBROBLASTS Synthesis and remodeling of the ECM by fibroblasts is essential for wound healing. Adult and fetal fibroblasts are recruited to the site of injury by soluble chemoattractants released by macrophages and neutrophils [8]. Fetal wounds characteristically have less inflammatory cells and cytokine expression yet heal more rapidly than adult wounds. This may be partly explained by intrinsic diVerences between adult and fetal fibroblasts. Fetal and adult fibroblasts display diVerences in synthetic function of collagen, HA, and other ECM components. In vitro, fetal fibroblasts synthesize more type III and IV collagens than their adult counterparts, correlating with an increase in prolyl hydrolase activity, the rate‐limiting step in collagen synthesis [72, 73]. Collagen synthesis is delayed in the adult wound while fibroblasts proliferate. In contrast, fetal fibroblasts simultaneously proliferate and synthesize collagen [8]. Increases in cell density diminish HA production in the adult but has no eVect on fetal HA synthesis [74]. Fetal fibroblasts have a greater ability to migrate into collagen gels than adult fibroblasts. A migration stimulation factor secreted by fetal fibroblasts is purported to be responsible for this enhanced migratory ability [75].

FETAL SKIN WOUND HEALING

151

Fetal fibroblasts have more surface receptors for HA, which also serves to enhance fibroblast migration [74]. Additionally, TGF‐b, which inhibits migration of confluent fibroblasts in vitro, is decreased in the fetal wound [76]. DiVerences in contractile fibroblasts, termed ‘‘myofibroblasts,’’ have also been reported. Myofibroblasts, detected by the presence of alpha SMA, appear in the adult wound 1 week after wounding. The content of myofibroblasts is greatest during the 2nd and 3rd week and then decreases with time [8]. Wounds made early in gestation have virtually no myofibroblasts. In contrast, scarring fetal and postnatal wounds have progressively more active myofibroblasts, which correlates with contraction and degree of scarring [77]. Overall, the fetal fibroblast has an intrinsic ability to synthesize the dermal ECM that is superior to the adult fibroblast in terms of its ability to generate dermis at sites of injury.

7. Cytokines 7.1. TRANSFORMING GROWTH FACTOR‐BETA (TGF‐b) The TGF‐b were linked to wound healing shortly after their discovery more than 20 years ago. TGF‐b is chemotactic for fibroblasts, keratinocytes, and inflammatory cells, and stimulates collagen I production by fibroblasts [78]. Isoforms TGF‐b1 and TGF‐b2 are thought to be profibrotic and to promote scar formation because their expression is increased in adult wounds and their exogenous administration to adult wounds increases collagen, protein, and inflammatory cell accumulation [78]. Expression is modified by decorin, fibromodulin, hypoxia, hypoxia‐inducible factor‐1 (HIF), and other ECM proteoglycan modulators [18, 79]. In turn, TGF‐b modulates MMP expression [78]. Evidence implicating TGF‐b1 as a proscarring cytokine is well established. Immunohistochemical analysis reveals no change in TGF‐b1 and ‐b2 expression in fetal rabbit wounds but increased expression in adult wounds [80]. Scarless wounds in fetal mice have less TGF‐b1 staining than neonatal or adult wounds [81]. Insertion of PVA sponges containing TGF‐b1 into rabbit wounds causes normally scarless wounds to heal with scar [81]. Treatment of adult rat wounds with neutralizing antibodies to TGF‐b1 and TGF‐b2 reduces scar formation [80, 82]. Whether this is due to decreased inflammation or a primary decrease in collagen synthesis by fibroblasts is unknown. Furthermore, the relative proportion of TGF‐b isoforms, and not the absolute amount of any one isoform, may determine the wound outcome phenotype. In scarless fetal wounds, TGF‐b3 expression is increased while TGF‐b1 expression is unchanged. Conversely, TGF‐b1 expression is

152

BUCHANAN ET AL.

increased and TGF‐b3 decreased in scarring fetal wounds [83, 84]. Treatment of adult rat wounds with exogenous TGF‐b3 reduces scar formation [85]. This suggests the ratio of TGF‐b3 to TGF‐b1 may determine whether tissue regenerates or forms scar. 7.2. CONNECTIVE TISSUE GROWTH FACTOR (CTGF) Like TGF‐b, CTGF is considered to be profibrotic. When TGF‐b binds to its receptors, it activates Smad proteins, which influence a number of target genes. One of these TGF‐b responsive genes is CTGF. CTGF stimulates the increased deposition of collagen fibers and other ECM components. Unlike TGF‐b, CTGF has no eVect on epidermal and inflammatory cells. At baseline, fetal fibroblasts were found to have lower CTGF expression compared with adult fibroblasts. However, after stimulation of adult and fetal fibroblasts with TGF‐b isoforms, fetal fibroblasts demonstrated increased expression of CTGF [86]. Thus, scarless fetal repair may also be due to a low expression of CTGF. Selective regulation of profibrotic cytokines may provide for a unique opportunity to decrease scarring and promote regeneration in postnatal wounds. Because TGF‐b has a broader role in the mileau of wound healing, its selective regulation may lead to unwarranted side eVects. Because CTGF is not known to have such a broad role in wound healing, its selective regulation could provide for a more controlled decrease in collagen scarring. 7.3. VASCULAR ENDOTHELIAL GROWTH FACTOR (VEGF) VEGF is a glycoprotein produced by keratinocytes, fibroblasts, and macrophages. Four VEGF isoforms have been identified via molecular studies, VEGF A–D. VEGF exerts its influence through receptors VEGFR‐1 and 2 in endothelial cells. VEGF is considered one of the main regulators of angiogenesis and vasculogenesis. VEGF isoforms display diVerent roles in diVerent tissues of the body. For example, VEGF‐B is found mostly in muscle and is a strong endothelial cell mitogen. VEGF‐C and ‐D regulate lymphogenesis through VEGFR‐3. Expression of VEGF is increased during adult wound and scar formation [103]. This increased expression has been associated with angiogenesis. VEGF role in fetal wound healing was studied by Colwell et al. Excisional wounds were created in fetal rats during the scarless and scarring gestational ages. In the scarless wound healing model, VEGF expression increased nearly threefold [87]. No increased expression was witnessed in the scarring wound healing model. VEGF and VEGFR 1 and 2 increased expression during skin development and dermal diVerentiation. The increased

FETAL SKIN WOUND HEALING

153

expression of VEGF during scarless wound repair may be a key factor contributing to the accelerated rate of healing in the scarless wound when compared to the scarring wound. 7.4. FIBROBLAST GROWTH FACTORS (FGFS) FGFs are a cytokines that are involved in regulation of cell proliferation, diVerentiation, and migration. Their role in scarless wound healing was evaluated by Dang et al. in 2002. It has been well established that FGFs 1,2,5,7, and 10 are upregulated during adult cutaneous wound healing. Excisional wounds in fetal rats at gestational ages for scarless (Day 16.5) and scarring (Day 19.5) wound healing were evaluated at 24, 48, and 72 h. Expression of FGF isoforms 2,5,7,9 and 10 and FGFR 1,2, and 4 were studied in relation to unwounded fetal skin [88]. In unwounded fetal skin, FGF isoform 5 doubled at birth, FGF 10 doubled at the transition period, and FGF 7 expression increased more than seven times at birth. FGF isoforms 2 and 9 had no change during fetal development while isoforms 1, 2, and 4 expression increased at birth. In the wounded models, the majority of the FGF isoforms demonstrated downregulation. FGF 2 expression decreased in both scarless and scarring wounds. FGF 7 and 10 were downregulated in scarless wound repair. These results demonstrate an overall downregulation of FGF isoforms in scarless wound healing. This finding suggests addition of FGFs to postnatal skin wounds will not reduce scarring. 7.5. PLATELET DERIVED GROWTH FACTOR (PDGF) The PDGF molecule consists of polypeptide chains A and B. When combined, they become a cytokine which is involved in wound healing in several ways. The molecule is produced and released by platelets at the site of wounding. It has shown to be a potent mesenchymal cell mitogen and chemoattractant [102]. In addition, its role in increasing ECM synthesis as well as glycosoaminoglycans synthesis in rabbit models has been demonstrated. PDGF‐BB is a recombinant human growth factor available commercially to assist with the treatment of chronic wounds by helping to promote granulation tissue. Peled et al. studied the expression of PDGF in Sprague‐Dawley rats at diVerential gestational ages representing scarless and scarring wound healing [89]. PDGF expression did not change markedly as a function of gestational age in fetal fibroblasts. PDGF did demonstrate an increase in gene expression with increasing gestational age in whole skin. There was a marked decrease in PDGF expression between the 16th and 18th gestational day, the transition period between scarless and scarring repair in rats. PDGF may play a role in the mechanism of scarless wound repair.

154

BUCHANAN ET AL.

No large change in PDGF expression has been found between scarless and scarring wounds. 7.6. WNTS The Wnt family glycoproteins are mitogens involved in cell proliferation, diVerentiation, cell–cell signaling as well as carcinogenesis. During fetal skin development, most Wnts, including Wnt‐4 are expressed. This expression, however, is lost in postnatal skin. The expression of Wnt‐4 is higher in fetal skin compared to postnatal skin [90]. After wounding, fetal skin wound Wnt expression does not increase. In contrast, Wnt expression increases during adult repair. Thus, the Wnts are likely permissive for scarless repair but cannot induce regenerative repair in adults. 7.7. INTERLEUKINS Interleukins are cytokines important in chemotaxis and activation of inflammatory cell mediators. IL‐6 stimulates monocyte chemotaxis and macrophage activation, while IL‐8 attracts neutrophils and stimulates neovascularization [91]. Wounding stimulates a rapid increase in IL‐6 and IL‐8, which persists at 72 h in the adult but disappears by 12 h in the fetus [91, 92]. PDGF induces adult fibroblast production of IL‐6 [91]. In turn, the addition of IL‐6 to fetal wounds produces scar in normally scarless wounds. Both IL‐6 and IL‐8 expression are significantly lower in early fetal fibroblasts at baseline and with PDGF stimulation compared to in adult fibroblasts [91–93]. Thus, proinflammatory ILs likely promote scarring. In contrast, IL‐10 has an anti‐inflammatory function through decreased production of IL‐6 and IL‐8. Wounds in fetal skin grafts harvested from early gestation IL‐10 knockout mice and grafted onto syngeneic adult mice heal with significant inflammation and scar [94]. In an initial study, adult mouse wounds were treated with an IL‐10 overexpression adenoviral vector. Inflammation was reduced and scarless healing occurred [95]. This will likely have potential therapeutic implications for human adult wounds. The ILs that decrease inflammation have potential antiscarring eVects. 7.8. MOLECULAR CONTROL OF SCARLESS REPAIR EVorts toward defining the scarless fibroblast phenotype have examined cellular signaling via secondary messenger receptor patterns and adapter protein Shc expression. Shc couples receptor tyrosine kinase (RTK) to mitogen‐activated protein kinase (MAPK) [96]. It serves as a key intermediate for discoid domain receptor (DDR) signaling and may contribute to

FETAL SKIN WOUND HEALING

155

hypoxia‐induced HIF protein stabilization and endothelial migration [97, 98]. Although TGF‐b signaling is mediated through intrinsic serine/ threonine kinase receptors, tyrosine kinase receptor signaling controls fundamental reaction sequences leading to gene activation [99]. DiVerent RTK phosphorylation patterns are observed between fetal and adult rat fibroblasts with increased amounts of epidermal growth factor receptor, DDR, and Shc proteins in fetal fibroblasts suggesting that RTK signaling may play a role in scarless repair [99]. The diVerent RTK phosphorylation patterns between fetal and adult fibroblasts further demonstrates their intrinsic diVerences that likely relate to their repair outcomes. Ultimately, the mechanistic diVerences between scarless and scarring repair may be regulated at the gene expression level. Homeobox genes are transcription factors that are implicated in the patterning and cell type specification events during development. These genes determine the direction taken by major developmental pathways involving activity of hundreds of genes. Their role in skin embryogenesis and wound healing is being investigated. Human homeobox genes MSX‐1, MSX‐2, and MOX‐1 are diVerentially expressed during skin development [100]. Additionally, human fetal scarless repair is associated with decreased expression of HOXB13 and increased PRX‐2 expression [101]. Given that scarless repair is inherent to developing skin, it seems likely that coordinated control of groups of genes by transcription factors, such as homeobox genes, has a crucial function during the repair process. 7.9. PERSPECTIVE Experimental data obtained in the past decade has greatly increased our knowledge of fetal wound healing, but the precise mechanism of this complex event remains unknown. Fetal wound repair is a tightly regulated process involving various cellular mediators and cytokines. In vivo up or downregulation of these repair elements interrupts the orderly sequence of regeneration resulting in scar formation. In turn, manipulation of the postnatal scarring wound cascade to decrease inflammation has decreased scarring and eventually may allow skin to regenerate. In addition, stem cell treatment has reduced scar and regenerated skin. This novel treatment will likely translate to clinical therapy. REFERENCES [1] G.S. Ashcroft, C.M. Kielty, M.A. Horan, M.W. Ferguson, Age‐related changes in the temporal and spatial distributions of fibrillin and elastin mRNAs and proteins in acute cutaneous wounds of healthy humans, J. Pathol. 183 (1997) 80–89. [2] A.S. Colwell, M.T. Longaker, H.P. Lorenz, Mammalian fetal organ regeneration, Adv. Biochem. Eng. Biotechnol. 93 (2005) 83–100.

156

BUCHANAN ET AL.

[3] H.P. Lorenz, D.J. Whitby, M.T. Longaker, N.S. Adzick, Fetal wound healing. The ontogeny of scar formation in the non‐human primate, Ann. Surg. 217 (1993) 391–396. [4] A.S. Colwell, T.M. Krummel, M.T. Longaker, H.P. Lorenz, An in vivo mouse excisional wound model of scarless healing, Plast. Reconstr. Surg. 117 (2006) 2292–2296. [5] J.W. Madden, E.E. Peacock, Jr., Studies on the biology of collagen during wound healing. 3. Dynamic metabolism of scar collagen and remodeling of dermal wounds, Ann. Surg. 174 (1971) 511–520. [6] A.S. Colwell, T.T. Phan, W. Kong, M.T. Longaker, P.H. Lorenz, Hypertrophic scar fibroblasts have increased connective tissue growth factor expression after transforming growth factor‐beta stimulation, Plast. Reconstr. Surg. 116 (2005) 1387–1390; Discussion 1391–1382. [7] K.L. Moore, The integumentary system, in: K.L. Moore, T.V.N. Persaud, (Eds.), Before We Were Born: Essentials of Embryology and Birth Defects. W.B. Saunders Company, Philadelphia, 1998, pp. 482–487. [8] R.A.F. Clark, Wound repair overview and general considerations, in: R.A.F. Clark. (Ed.), The Molecular and Cellular Biology of Wound Repair, Vol 12, Plenum Press, New York, 1996, pp. 3–50. [9] J.R. Merkel, B.R. DiPaolo, G.G. Hallock, D.C. Rice, Type I and type III collagen content of healing wounds in fetal and adult rats, Proc. Soc. Exp. Biol. Med. 187 (1988) 493–497. [10] G.G. Hallock, J.R. Merkel, D.C. Rice, B.R. DiPaolo, The ontogenetic transition of collagen deposition in rat skin, Ann. Plast. Surg. 30 (1993) 239–243. [11] R.L. McCallion, M.W.J. Ferguson, R.A.F. Clark, (Eds.), Fetal Wound Healing and the Development of Antiscarring Therapies for Adult Wound Healing, Vol xxiii, Plenum Press, New York, 1996, pp. 561–600. [12] B.A. Mast, L.C. Flood, J.H. Haynes, R.L. DePalma, I.K. Cohen, R.F. Diegelmann, T.M. Krummel, Hyaluronic acid is a major component of the matrix of fetal rabbit skin and wounds: implications for healing by regeneration, Matrix 11 (1991) 63–68. [13] B.A. Mast, R.F. Diegelmann, T.M. Krummel, I.K. Cohen, Hyaluronic acid modulates proliferation, collagen and protein synthesis of cultured fetal fibroblasts, Matrix 13 (6) (1993) 441–446. [14] S.R. Beanes, C. Dang, C. Soo, Y. Wang, M. Urata, K. Ting, et al., Down‐regulation of decorin, a transforming growth factor‐beta modulator, is associated with scarless fetal wound healing, J. Pediatr. Surg. 11 (2001) 1666–1671. [15] S.R. Beanes, C. Dang, C. Soo, H.P. Lorenz, Ontogenetic transition in the fetal wound extracellular matrix correlates with scar formation, Wound Repair Regen. 9 (2001) 151. [16] H.P. Lorenz, C. Soo, S.R. Beanes, C. Dang, X. Zhang, J.B. Atkinson, et al., DiVerential expression of matrix metalloproteinases and their tissue‐derived inhibitors in scarless fetal wound healing, Surg. Forum (2001) 397–401. [17] C. Soo, F.Y. Hu, X. Zhang, Y. Wang, S.R. Beanes, H.P. Lorenz, et al., DiVerential expression of fibromodulin, a TGF‐b modulator, in fetal skin development and scarless repair, Am. J. Pathol. 157 (2000) 423–433. [18] A. Hildebrand, M. Romaris, L.M. Rasmussen, D. Heinegard, D.R. Twardzik, W.A. Border, et al., Interaction of the small interstitial proteoglycans biglycan, decorin and fibromodulin with transforming growth factor b, Biochem. J. 302 (1994) 527–534. [19] M.T. Longaker, D.J. Whitby, N.S. Adzick, T.M. Cromblehome, J.C. Langer, B.W. Duncan, et al., Studies in fetal wound healing. VI. Second and third trimester fetal wounds demonstrate rapid collagen deposition without scar formation, J. Pediatr. Surg. 25 (1990) 63–69. [20] D.J. Whitby, M.W.J. Ferguson, The extracellular matrix of lip wounds in fetal, neonatal, and adult mice, Development 112 (1991) 651–668.

FETAL SKIN WOUND HEALING

157

[21] S.R. Beanes, F.Y. Hu, C. Soo, C.M. Dang, M. Urata, K. Ting, J.B. Atkinson, et al., Confocal microscopic analysis of scarless repair in the fetal rat: defining the transition, Plast. Reconstr. Surg. 109 (2002) 160–170. [22] Z.L. Wang, T. Inokuchi, H. Ikeda, T.T. Baba, M. Uehara, N. Kamasaki, et al., Collagen‐ binding heat shock protein HSP47 expression during healing of fetal skin wounds, Int. J. Oral Maxillofac. Surg. 31 (2) (2002) 179–184. [23] H.N. Lovvorn, 3rd, D.T. Cheung, M.E. Nimni, N. Perelman, J.M. Estes, N.S. Adzick, Relative distribution and crosslinking of collagen distinguish fetal from adult sheep wound repair, J. Pediatr. Surg. 34 (1999) 218–223. [24] C.I. Kennedy, R.F. Diegelmann, J.H. Haynes, D.R. Yager, Proinflammatory cytokines diVerentially regulate hyaluronan synthase isoforms in fetal and adult fibroblasts, J. Pediatr. Surg. 35 (2000) 874–879. [25] M.T. Longaker, D.J. Whitby, R.W. Jennings, B.W. Duncan, M.W. Ferguson, M.R. Harrison, et al., Fetal diaphragmatic wounds heal with scar formation, J. Surg. Res. 50 (1991) 375–385. [26] V. Moulin, M. Plamondon, DiVerential expression of collagen integrin receptor on fetal vs. adult skin fibroblasts: implication in wound contraction during healing, Br. J. Dermatol. 147 (5) (2002) 886–892. [27] C. Soo, S. Beanes, C. Dang, X. Zhang, K. Ting, Fibromodulin, a TGF‐b modulator, promotes scarless fetal repair, Surg. Forum 52 (2001) 578–581. [28] J.R. Armstrong, M.W.J. Ferguson, Ontogeny of the skin and transition from scar free to scarring phenotype during wound healing in the pouch young of Monodelphis domestica, Dev. Biol. 169 (1995) 242–260. [29] M.T. Longaker, D.J. Whitby, M.W.J. Ferguson, H.P. Lorenz, M.R. Harrison, N.S. Adzick, Adult skin wounds in the fetal environment heal with scar formation, Ann. Surg. 219 (1994) 65–72. [30] M.T. Longaker, D.J. Whitby, R.W. Jennings, B.W. Duncan, M.W. Ferguson, M.R. Harrison, et al., Fetal diaphragmatic wounds heal with scar formation, J. Surg. Res. 50 (1991) 375–385. [31] M. Meuli, H.P. Lorenz, M.H. Hedrick, K.M. Sullivan, M.R. Harrison, N.S. Adzick, Scar formation in the fetal alimentary tract, J. Pediatr. Surg. 30 (1995) 392–395. [32] H.P. Lorenz, R.Y. Lin, M.T. Longaker, D.J. Whitby, N.S. Adzick, The fetal fibroblast: the eVector cell of scarless wounds repair, Plast. Reconstr. Surg. 96 (1995) 1251–1259. [33] S. Ihara, Y. Motobayashi, E. Nagao, A. Kistler, Ontogenetic transition of wound healing pattern in rat skin occurring at the fetal stage, Development 110 (1990) 671–680. [34] D.L. Cass, K.M. Bullard, K.G. Sylvester, E.Y. Yang, M.T. Longaker, N.S. Adzick, Wound size and gestational age modulate scar formation in fetal wound repair, J. Pediatr. Surg. 32 (1997) 411–415. [35] H.P. Lorenz, D.J. Whitby, M.T. Longaker, N.S. Adzick, Fetal wound healing: the ontogeny of scar formation in the non‐human primate, Ann Surg 217 (1993) 391–396. [36] A.F. Singer, R.A.F. Clark, Cutaneous wound healing, N. Engl. J. Med. 341 (10) (1999) 738–746. [37] F.W. Frantz, D.A. Bettinger, J.H. Haynes, D.E. Johnson, K.M. Harvey, H.P. Dalton, et al., Biology of fetal repair: the presence of bacteria in fetal wounds induces an adult‐like healing response, J. Pediatr. Surg. 28 (1993) 428–434. [38] A.J. Wagers, J.L. Christensen, I.L. Weissman, Cell fate determination from stem cells, Gene Ther. 9 (2002) 606–612. [39] M. Sata, A. Saiura, A. Kunisato, A. Tojo, S. Okada, T. Tokuhisa, et al., Hematopoietic stem cells diVerentiate into vascular cells that participate in the pathogenesis of atherosclerosis, Nat. Med. 8 (2002) 403–409.

158

BUCHANAN ET AL.

[40] G. Ferrari, G. Cusella‐De Angelis, M. Coletta, E. Paolucci, A. Stornaiuolo, G. Cossu, F. Mavilio, Muscle regeneration by bone marrow‐derived myogenic progenitors, Science 279 (1998) 1528–1530. [41] E. Mezey, K.J. Chandross, G. Harta, R.A. Maki, S.R. McKercher, Turning blood into brain: cells bearing neuronal antigens generated in vivo from bone marrow, Science 290 (2000) 1779–1782. [42] E. Lagasse, H. Connors, M. Al‐Dhalimy, M. Reitsma, M. Dohse, L. Osborne, et al., Purified hematopoietic stem cells can diVerentiate into hepatocytes in vivo, Nat. Med. 6 (2000) 1229–1234. [43] A.J. Wagers, R.I. Sherwood, J.L. Christensen, I.L. Weissman, Little evidence for developmental plasticity of adult hematopoietic stem cells, Science 297 (2002) 2256–2259. [44] L.B. Balsam, A.J. Wagers, J.L. Christensen, T. Kofidis, I.L. Weissman, R.C. Robbins, Haematopoietic stem cells adopt mature haematopoietic fates in ischaemic myocardium, Nature 428 (2004) 668–673. [45] L. Roybon, Z. Ma, F. Asztely, A. Fosum, S.E. Jacobsen, P. Brundin, et al., Failure of transdiVerentiation of adult hematopoietic stem cells into neurons, Stem Cells 24 (2006) 1594–1604. [46] Y. Jiang, B.N. Jahagirdar, R.L. Reinhardt, R.E. Schwartz, C.D. Keene, X.R. Ortiz‐ Gonzalez, et al., Pluripotency of mesenchymal stem cells derived from adult marrow, Nature 418 (2002) 41–49. [47] G. D’Ippolito, S. Diabira, G.A. Howard, P. Menei, B.A. Roos, P.C. Schiller, Marrow‐ isolated adult multilineage inducible (MIAMI) cells, a unique population of postnatal young and old human cells with extensive expansion and diVerentiation potential, J. Cell Sci. 117 (2004) 2971–2981. [48] G. Kogler, S. Sensken, J.A. Airey, T. Trapp, M. Muschen, N. Feldhahn, A new human somatic stem cell from placental cord blood with intrinsic pluripotent diVerentiation potential, J. Exp. Med. 200 (2004) 123–135. [49] A. Chiavegato, S. Bollini, M. Pozzobon, A. Callegari, L. Gasparotto, J. Taiani, et al., Human amniotic fluid‐derived stem cells are rejected after transplantation in the myocardium of normal, ischemic, immuno‐suppressed or immuno‐deficient rat, J. Mol. Cell Cardiol. 42 (2007) 746–759. [50] C. Holden, Stem cells. Controversial marrow cells coming into their own? Science 315 (2007) 760–761. [51] L. Liang, S. Chinnathambi, M. Stern, A. Tomanek‐Chalkley, T.D. Manuel, J.R. Bickenbach, As epidermal stem cells age they do not substantially change their characteristics, J. Investig. Dermatol. Symp. Proc. 9 (2004) 229–237. [52] H. Oshima, A. Rochat, C. Kedzia, K. Kobayashi, Y. Barrandon, Morphogenesis and renewal of hair follicles from adult multipotent stem cells, Cell 104 (2001) 233–245. [53] M. Rendl, L. Lewis, E. Fuchs, Molecular dissection of mesenchymal–epithelial interactions in the hair follicle, PLoS Biol. 3 (2005) e331. [54] T. Tumbar, G. Guasch, V. Greco, C. Blanpain, W.E. Lowry, M. Rendl, et al., Defining the epithelial stem cell niche in skin, Science 303 (2004) 359–363. [55] V. Levy, C. Lindon, B.D. Harfe, B.A. Morgan, Distinct stem cell populations regenerate the follicle and interfollicular epidermis, Dev. Cell 9 (2005) 855–861. [56] P.H. Jones, F.M. Watt, Separation of human epidermal stem cells from transit amplifying cells on the basis of diVerences in integrin function and expression, Cell 73 (1993) 713–724. [57] C.S. Trempus, R.J. Morris, C.D. Bortner, G. Cotsarelis, R.S. Faircloth, J.M. Reece, et al., Enrichment for living murine keratinocytes from the hair follicle bulge with the cell surface marker CD34, J. Invest. Dermatol. 120 (2003) 501–511.

FETAL SKIN WOUND HEALING

159

[58] A. Yang, M. Kaghad, D. Caput, F. McKeon, On the shoulders of giants: p63, p73 and the rise of p53, Trends Genet. 18 (2002) 90–95. [59] R.J. Medina, K. Kataoka, M. Takaishi, M. Miyazaki, N.H. Huh, Isolation of epithelial stem cells from dermis by a three‐dimensional culture system, J. Cell Biochem. 98 (1) (2006) 174–184. [60] L. Larue, M. Ohsugi, J. Hirchenhain, R. Kemler, E‐cadherin null mutant embryos fail to form a trophectoderm epithelium, Proc. Natl. Acad. Sci. USA 91 (1994) 8263–8267. [61] D.S. Krause, T. Ito, M.J. Fackler, O.M. Smith, M.I. Collector, S.J. Sharkis, Characterization of murine CD34, a marker for hematopoietic progenitor and stem cells, Blood 84 (1994) 691–701. [62] M.R. Albert, R.A. Foster, J.C. Vogel, Murine epidermal label‐retaining cells isolated by flow cytometry do not express the stem cell markers CD34, Sca‐1, or Flk‐1, J. Invest. Dermatol. 117 (2001) 943–948. [63] L. Liang, J.R. Bickenbach, Somatic epidermal stem cells can produce multiple cell lineages during development, Stem Cells 20 (2002) 21–31. [64] K.J. Fernandes, I.A. McKenzie, P. Mill, K.M. Smith, M. Akhavan, F. Barnabe‐Heider, A dermal niche for multipotent adult skin‐derived precursor cells, Nat. Cell Biol. 6 (2004) 1082–1093. [65] R.J. Medina, K. Kataoka, M. Takaishi, M. Miyazaki, N.H. Huh, Isolation of epithelial stem cells from dermis by a three‐dimensional culture system, J. Cell Biochem. 98 (2006) 174–184. [66] W. Kong, S. Li, M. Longaker, H.P. Lorenz, Blood derived small dot cells reduce scar in wound healing, Exp. Cell Res. 314 (2008) 1529–1539. [67] O.O. Olutoye, D.R. Yager, I.K. Cohen, R.F. Deigelmann, Lower cytokine release by fetal porcine platelets: a possible explanation for reduced inflammation after fetal wounding, J. Pediatr. Surg. 31 (1996) 91–95. [68] O.O. Olutoye, E.J. Barone, D.R. Yager, I.K. Cohen, R.F. Diegelmann, Collagen induces cytokine release by fetal platelets: implications in scarless healing, J. Pediatr. Surg. 32 (1997) 827–830. [69] O.O. Olutoye, S.M. Alaish, M.E. Carr, M. Paik, D.R. Yager, I.K. Cohen, R.F. Diegelmann, Aggregatory characteristics and expression of fetal porcine platelets, J. Pediatr. Surg. 30 (1995) 1649–1653. [70] O.O. Olutoye, E.J. Barone, D.R. Yager, T. Uchida, I.K. Cohen, R.F. Diegelmann, Hyaluronic acid inhibits fetal platelet function: implications in scarless healing, J. Pediatr. Surg. 32 (1997) 1037–1040. [71] R.W. Jennings, N.S. Adzick, M.T. Longaker, B.W. Duncan, H. Scheuenstuhl, T.K. Hunt, Ontogeny of fetal sheep polymorphonuclear leukocyte phagocytosis, J. Pediatr. Surg. 26 (1991) 853–855. [72] B.L. Thomas, T.M. Krummel, M. Melang, J.W. Cawthorn, J.M. Nelson, I.K. Cohen, et al., Collagen synthesis and type expression by fetal fibroblasts in vitro, Surg. Forum 39 (1988) 642. [73] H.P. Lorenz, N.S. Adzick, Scarless skin wound repair in the fetus, West. J. Med. 159 (1993) 350–355. [74] W.Y. Chen, M.E. Grant, A.M. Schor, S.L. Schor, DiVerences between adult and foetal fibroblasts in the regulation of hyaluronate synthesis. Correlation with migratory activity, J. Cell Sci. 94 (1989) 577–589. [75] S.L. Schor, A.M. Schor, A.M. Grey, G. Rushton, Fetal and cancer patient fibroblasts produce an autocrine migration stimulating factor not made by normal adult fibroblasts, J. Cell Sci. 90 (1988) 391–399.

160

BUCHANAN ET AL.

[76] I.R. Ellis, S.L. Schor, DiVerential eVects of TGF‐ 1b on hyaluronan synthesis by fetal and adult skin fibroblasts: implications for cell migration and wound healing, Exp. Cell Res. 228 (1996) 326–333. [77] J.M. Estes, J. Vandeberg, N.S. Adzick, T.E. MacGillirray, A. Desmonliere, G. Gabbiani, Phenotypic and functional features of myofibroblasts in sheep fetal wounds, DiVerentiation 56 (1994) 173–181. [78] A.B. Roberts, M.B. Sporn, Transforming growth factor‐b, in: R.A.F. Clark. (Ed.), The Molecular and Cellular Biology of Wound Repair, Vol xxiii, Plenum Press, New York, 1996, pp. 275–308. [79] A. Scheid, R.H. Wenger, L. SchaVer, I. Camenisch, O. Distler, A. Ferenc, et al., Physiologically low oxygen concentrations in fetal skin regulate hypoxia‐inducible factor 1 and transforming growth factor-beta3, FASEB J. 16 (3) (2002) 411–413. [80] R.K. Nath, M. LaRegina, H. Markham, G.A. Ksander, P.M. Weeks, The expression of transforming growth factor type beta in fetal and adult rabbit skin wounds, J. Pediatr. Surg. 29 (1994) 416–421. [81] T.M. Krummel, B.A. Michna, B.L. Thomas, M.B. Sporn, J.M. Nelson, A.M. Salzberg, et al., Transforming growth factor beta (TGFb) induces fibrosis in a fetal wound model, J. Pediatr. Surg. 23 (1988) 647–652. [82] M. Shah, D.M. Foreman, M.W. Gerguson, Neutralising antibody to TGF‐beta 1,2 reduces cutaneous scarring in adult rodents, J. Cell Sci. 107 (1994) 1137–1157. [83] M. Hsu, Z.M. Peled, G.S. Chin, W. Liu, M.T. Longaker, Ontogeny of expression of transforming growth factor‐beta‐1 (TGF‐beta‐1), TGF‐beta‐3, and TGF‐beta receptors I and II in fetal rat fibroblasts and skin, Plast. Reconstr. Surg. 107 (2001) 1787–1794. [84] C. Dang, S.R. Beanes, B.C. Soo, M.H. Hedrick, H.P. Lorenz, A high ratio of TGFb3 to TGFb1 expression in wounds is associated with scarless repair, Wound Repair Regen. 9 (2001) 153. [85] M. Shah, D.M. Foreman, M.W. Ferguson, Neutralisation of TGF‐beta 1 and TGF‐beta 2 or exogenous addition of TGF‐beta 3 to cutaneous rat wounds reduces scarring, J. Cell Sci. 108 (1995) 985–1002. [86] A. Colwell, T. Krummel, M. Longaker, H.P. Lorenz, Fetal and adult fibroblasts have similar TGF‐b-mediated, Smad dependent signaling pathways, Plast. Reconstr. Surg. 117 (2004) 2277–2283. [87] A. Colwell, S. Beanes, C. Soo, C. Dang, K. Ting, M. Longaker, Increased angiogenesis and expression of vascular endothelial growth factor during scarless repair, Plast. Reconstr. Surg. 115 (2005) 204–212. [88] C.M. Dang, S.R. Beanes, C. Soo, H.P. Lorenz, Decreased expression of fibroblast and keratinocyte growth factor isoforms and receptors during scarless repair, Plast. Reconstr. Surg. 111 (2001) 1969–1979. [89] Z. Peled, S. Rhee, M. Hsu, J. Chang, T. Krummell, M. Longaker, The ontogeny of scarless healing II: EGF and PDGF‐B gene expression in fetal rat skin and fibroblasts as a function of gestational age, Ann. Plast. Surg. 47 (2001) 417–424. [90] A. Colwell, T. Krummel, M. Longaker, H.P. Lorenz, Wnt‐4 Expression is increased in fibroblasts after TGF‐b1 Stimulation and during fetal and postnatal wound repair, Plast. Reconstr. Surg. 117 (2006) 2297–2301. [91] K.W. Liechty, N.S. Adzick, T.M. Crombleholme, Diminished interleukin 6 (IL‐6) production during scarless human fetal wound repair, Cytokine 12 (2000) 671–676. [92] K.W. Liechty, T.M. Crombleholme, D.L. Cass, B. Martin, N.S. Adzick, Diminished interleukin‐8 (IL‐8) production in the fetal wound healing response, J. Surg. Res. 77 (1998) 80–84.

FETAL SKIN WOUND HEALING

161

[93] I. Ono, H. Gunji, J.Z. Zhang, K. Maruyama, F. Kaneko, A study of cytokines in burn blister fluid related to wound healing, Burns 21 (1995) 352–355. [94] K.W. Liechty, H.B. Kim, N.S. Adzick, T.M. Crombleholme, Fetal wound repair results in scar formation in interleukin‐10 deficient mice in a syngeneic murine model of scarless fetal wound repair, J. Pediatr. Surg. 35 (2000) 866–872. [95] A.d Gordon, J. Karmacharya, M. Herlyn, T.M. Crombleholme, Scarless wound healing induced by adenoviral‐mediated overexpression of interleukin‐10, Surg. Forum 52 (2001) 568–569. [96] K.S. Ravichandran, Signaling via Shc family adapter proteins, Oncogene 20 (44) (2001) 6322–6330. [97] K. Ikeda, L.H. Wang, R. Torres, H. Zhao, E. Olaso, F.J. Eng, et al., Discoidin domain receptor 2 interacts with Src and Shc following its activation by type I collagen, J. Biol. Chem. 277 (21) (2002) 19206–19212. [98] F. Jung, J. Haendeler, J. HoVmann, A. Reissner, E. Dernbach, A.M. Zeiher, et al., Hypoxia induction of the hypoxia‐inducible factor is mediated via the adapter protein Shc in endothelial cells, Circ. Res. 91 (1) (2002) 38–45. [99] G.S. Chin, W.J. Kim, T.Y. Lee, W. Liu, P.B. Saadeh, S. Lee, et al., DiVerential expression of receptor tyrosine kinases and Shc in fetal and adult rat fibroblasts: toward defining scarless versus scarring fibroblast phenotypes, Plast. Recontr. Surg. 105 (2000) 972–979. [100] E.J. Stelnicki, L.A. Komuves, D. Holmes, W. Clovin, M.R. Harrison, N.S. Adzick, et al., The human homeobox genes MSX‐1, MSX‐2, and MOX‐1 are diVerentially expressed in the dermis and epidermis of fetal and adult skin, DiVerentiation 62 (1997) 33–41. [101] E.J. Stelnicki, J. Arbeit, D.L. Cass, C. Saner, M. Harrison, C. Largman, Modulation of the human homeobox genes PRX‐2 and HOXB13 in scarless fetal wounds, J. Invest. Dermatol. 111 (1998) 57–63. [102] D.J. Whitby, M.W.J. Ferguson, Immunohistochemical localization of growth factors in fetal wound healing, Dev. Biol. 147 (1991) 207–215. [103] S.R. Beanes, C. Dang, C. Soo, H.P. Lorenz, DiVerential expression of vascular endothelial growth factor in fetal wounds, Wound Repair Regen. 9 (2001) 154–155. [104] J.A. Shizuru, R.S. Negrin, I.L. Weissman, Hematopoietic stem and progenitor cells: clinical and preclinical regeneration of the hematolymphoid system, Annu. Rev. Med. 56 (2005) 509–538.

This page intentionally left blank

ADVANCES IN CLINICAL CHEMISTRY, VOL. 48

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP OF PATIENTS WITH CHRONIC HEART FAILURE Aldo Clerico,*,†,1 Marianna Fontana,* Andrea Ripoli,* and Michele Emdin* *Gabriele Monasterio Foundation CNR-Regione Toscana, 56126 Pisa, Italy † Scuola Superiore Sant’Anna, 56126 Pisa, Italy

1. 2. 3. 4.

5.

6. 7. 8.

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Background and Aim of the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Biochemical and Physiological Properties of B‐Type Natriuretic Peptides . . . . . . . . Circulating Levels of B‐Type Natriuretic Peptides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Analytical Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Pathophysiological Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Variations of Plasma B‐Type Natriuretic Peptides, Dependent on Pharmacological Treatment, as Surrogate End‐Point for Treatment of Patients with HF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prognostic Relevance of Plasma BNP/NT‐proBNP Variations After Treatment. . Meta‐Analysis for Overall Mortality Including All Randomized Clinical Trials . . BNP‐Guided Therapy in Chronic Heart Failure: Instructions for Use . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

163 164 165 167 167 168

168 169 174 175 176

1. Abstract The measurement of circulating brain natriuretic peptide (BNP) and its related peptide, the N‐terminal fragment of proBNP (NT‐proBNP), have a high degree of diagnostic accuracy and clinical relevance both in acute and chronic heart failure (HF). However, the role of measurement of BNP/NT‐ proBNP in the follow‐up of treated HF patients is still debated. In this

1

Corresponding author: Aldo Clerico, e‐mail: [email protected] 163

0065-2423/09 $35.00 DOI: 10.1016/S0065-2423(09)48007-7

Copyright 2009, Elsevier Inc. All rights reserved.

164

CLERICO ET AL.

chapter, authors have studied the clinical impact of B‐type natriuretic peptide assay in the follow‐up of patients with heart failure, and, in particular, the possible role of the measurement of its circulating levels in guiding the treatment. A relatively small number of randomized studies were designed to specifically evaluate the clinical use of BNP/NT‐proBNP assay in monitoring and tailoring the medical therapy in HF patients. A meta‐analysis of results reported in these studies indicate that the ineYcacy to improve the mortality rate of the peptide‐guided compared to the control group found in some studies, may depend on to the inability of current therapeutic strategies to modify prognosis, especially in the elderly subset of patients, who are characterized by more advanced disease and comorbidities. Further prospective and randomized clinical studies are necessary to definitively demonstrate whether BNP/ NT‐proBNP‐guided therapy is able to significantly improve the outcome of patients with HF.

2. Background and Aim of the Study Chronic heart failure (HF) represents a major public health problem, aVecting almost 7 million Europeans and 5 million North Americans each year; estimates regarding the prevalence of symptomatic HF in the general European and North American population range from 0.4% to 2% [1–5]. Both the incidence and prevalence of HF grow steadily with age in the European and North American population and the incidence of HF approaches 10 per 1000 population after the age of 65 [1–5]. In the USA, HF is the most common hospital discharge diagnosis, and more Medicare dollars are spent for diagnosis and treatment of HF than for any other diagnosis [4]; similar data have been reported for diVerent European countries [1–3]. Despite the enormous advances in the understanding and treatment of HF that have taken place over the last 50 years [6], HF continues to have a poor prognosis. In the European and North American populations, just less than 40% of patients diagnosed with severe HF (NYHA class IV or ACC/ AHA stage D) die within a year with survival rates similar to those of colon cancer, and worse than those of breast or prostate cancer [1–5]. HF may be considered as the fatal finishing line of all cardiovascular disorders. Some years ago, Braunwald and Bristow [6] suggested that it is possible to reverse the HF process, that had long been considered as irreversible and amenable only to palliative therapy. The idea of chronic HF as an irreversible, end‐stage process has been challenged by the experimental and clinical evidences of a possible improvement in the intrinsic defects of function and structure, aZicting the failing heart [6]. Such improvement is more rapid if intervention takes place in the very early phase of cardiac alteration.

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP

165

Diagnosis, prognosis, and follow‐up patients with HF depend on the availability of specific, accurate, and eVective disease markers. For this reason, there is an increasing interest in the development of new biomarkers and a great number of laboratory tests have recently been proposed [7]. Cardiac natriuretic peptide plasma concentration has been progressively growing in HF, with the occurrence of signs and symptoms of expanded fluid volume, worsening of myocardial structural changes, and systolic and diastolic dysfunction [8–10]. The measurement of circulating brain natriuretic peptide (BNP) and its related peptides, such as the N‐terminal fragment of proBNP (NT‐proBNP), is now considered a useful marker of clinical severity and prognosis in HF [8–16], and has been included in the first step of the diagnostic algorithm of suspicious symptoms, along with history, physical examination, electrocardiogram, and chest X‐ray film [1–3]. Recent systematic reviews and meta‐analyses [10–14] have confirmed that both BNP and NT‐proBNP assays have a high degree of diagnostic accuracy and clinical relevance both in acute and chronic HF. However, the role of BNP/ NT‐proBNP assay in the follow‐up of treated HF patients is still debated. In the present chapter, authors review the clinical impact of B‐type natriuretic peptide assay in the follow‐up of patients with HF, and, in particular, the possible role of the measurement of its circulating levels in guiding the treatment. Furthermore, this article discusses some data reported in literature and also the original results obtained in the authors’ laboratory.

3. Biochemical and Physiological Properties of B‐Type Natriuretic Peptides Human BNP is encoded by a single copy gene on the chromosome 1, consisting of three exons and two introns [17]. Unlike the atrial natriuretic peptide (ANP), where regulation seems to occur at the level of release from storage granules, BNP regulation takes place during the gene expression [8, 17]. It is believed that ANP is preferentially produced in atria, while BNP is produced in bursts in ventricular and to a smaller extent in atrial cardiomyocytes, as well as in fibroblasts [8, 17–19]. Mounting evidence from in vivo and ex vivo studies is providing supports to the hypothesis that the production/secretion of cardiac natriuretic peptides is regulated by complex interactions with the neurohormonal and immune systems, especially in the ventricular myocardium, as recently reviewed in detail [8, 17, 18]. Endothelin‐1 and angiotensin II are considered the most powerful stimulators of production/secretion of both ANP and BNP; similarly, glucocorticoids, female sex steroid hormones, thyroid hormones, some

CLERICO ET AL.

166

growth factors, and cytokines (especially TNF‐a, interleukin‐1, and interleukin‐6) share stimulating eVects on the cardiac endocrine function [8, 17, 18]. Human BNP is synthesized as a 134‐amino acid (aa) precursor protein (preproBNP) and is subsequently processed during secretion to form the 108‐aa peptide, proBNP. The pre-propeptide hormones of the cardiac natriuretic peptides can be enzymatically cleaved by at least two proprotein convertases produced in the cardiomyocytes, such as corin and furin [20, 21]. In particular, proBNP is processed to form the 76‐aa N‐terminal peptide (i.e., NT‐proBNP), and then the biologically active 32‐aa C‐terminal peptide (i.e., BNP). BNP and the NT‐proBNP are secreted in the blood in equimolar amounts; however, BNP has a shorter plasma half‐life (about 15–20 min vs. 1 or 2 h) and consequently lower plasma concentration, compared to NT‐proBNP (Fig. 1) [8, 10]. Moreover, some forms of O‐glycosylated proBNP are also present in plasma, especially of patients with heart failure [22]. Studies on structure–activity relationships have shown the importance for the binding to the specific receptors of the central ring structure of cardiac natriuretic peptides, formed by a disulfide bridge between the two cysteine residues. For this reason, only BNP, which present the disulfide bridge in the peptide chain, share the typical hormonal activity of cardiac natriuretic hormones, while the NT‐proBNP does not [8, 10].

Ventricular cardiomyocyte preproBNP (134aa)

Signal peptide (26aa)

proBNP (108aa)

NT-proBNP1–76

BNP (32aa)

Plasma

NT-proBNP

Half-life > 60 min Median 41 ng/L Range 7–220 ng/L

BNP

Half-life 15–20 min Median 7 ng/L Range 1–65 ng/L

FIG. 1. Schematic representation of BNP production and secretion.

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP

167

4. Circulating Levels of B‐Type Natriuretic Peptides 4.1. ANALYTICAL ASPECTS Despite the important physiological role played by BNP and its importance as a diagnostic analyte, little is known about the structure of the circulating forms of BNP and its related peptides [8, 21]. The precursor proBNP, the active peptide BNP, as well as the inactive N‐terminal fragment of proBNP (NT‐proBNP) are present in the circulating blood, and therefore these peptides can be assayed in the plasma or serum samples. In particular, BNP and NT‐proBNP are usually measured by fully automated platforms using noncompetitive immunometric assays [8, 23–28]. These assays are noncompetitive sandwich‐type immunoassays that use nonradioactive materials as labels for antigen/antibody reaction and two monoclonal antibodies or a combination of monoclonal and polyclonal antibodies for peptide binding [28]. Immunoassay methods for BNP use one antibody specific for the ring structure and the other antibody for the C‐ or N‐terminal end of the peptide hormone, respectively [28]; while the second generation electrochemiluminescence (ECLIA) method for NT‐proBNP uses two diVerent monoclonal antibodies against the central part of the peptide, which is the most stable part of the molecule [29]. Theoretically, setting up an immunoassay for NT‐proBNP should be easier than that for BNP, because this inactive peptide has higher plasma concentration than the active hormone (Fig. 1) [24, 25, 28]. Since BNP and NT‐proBNP have completely diVerent biochemical structure, molecular weight, biological activity, and degradation pathways, it is not surprising that BNP and NT‐proBNP assay methods may also have diVerent analytical characteristics and quality specifications [8, 23–28]. A multicenter collaborative study, including more than 100 Italian clinical laboratories, which carried out a total of 2354 determinations on 28 study samples, has recently confirmed that there are marked diVerences in analytical characteristics (such as assay imprecision) and measured values among the most popular commercial methods for BNP and NT‐proBNP [30]. Furthermore, another recent study has suggested that both native and glycosylated forms of circulating proBNP can diVerently aVect the immunoassay methods for BNP and NT‐ proBNP; thus suggesting that proBNP glycosylation is likely to interfere with peptide antibody binding, especially in the NT‐proBNP immunoassays [23]. The above‐reported data confirm that reference intervals and decision limits derived from clinical studies are only valid for the particular assay used and should not be extrapolated to other assays [8, 23–28]. To avoid misinterpretation of the results, the international guidelines [26, 27] recommend that one should consider the assay used, the available clinical evidence based on that individual assay, together with the clinical aim of an individual

168

CLERICO ET AL.

biomarker‐based study. Clinicians should give great care to compare results obtained by laboratories using diVerent methods. 4.2. PATHOPHYSIOLOGICAL CONSIDERATIONS As already stated, the cardiac endocrine function is influenced not only by hemodynamic changes leading to ventricular enlargement and/or an increase in ventricular wall stress, but also by neuroendocrine system (such as sympathetic, renin–angiotensin–aldosterone, and endothelin systems) and several cytokines, which share vasoconstrictor, sodium‐retentive, and hypertrophic eVects. In addition, the response of the cardiac natriuretic hormone system to neuroendocrine and cytokine activation is not linear (probably log shaped); therefore, in the presence of small stresses, the hormone system responds with a much greater augmentation of BNP levels [8, 10]. Small hemodynamic changes, which are hardly detectable at standard instrumental evaluation, may produce significant changes in BNP levels. Age and sex may also play a relevant role in the regulation of BNP (or NT‐ proBNP) circulating levels. It is well known that women in their fertile period show higher values (almost twofold) than their male counterparts, but after the age of 50, these values increase in both sexes, so that a 60‐year‐old man may have doubled values of BNP levels than a 30‐year‐old man [8, 10]. Nevertheless, an elevated concentration of cardiac natriuretic peptides may be found in some physiological (especially pregnancy and physical exercise) and pathological conditions and also in some therapeutic settings (female sex steroid hormones, corticosteroids, thyroid hormones, sympathomimetic agents with beta‐agonist activity, beta‐blockers, and digitalis). In some of these cases, cardiac output is often within the normal range or rather slightly increased (e.g., in case of physical exercise or hyperthyroidism). On the other hand, many diseases with increased BNP/NT‐proBNP levels may show the same symptoms of heart failure such as peripheral edema, dyspnea, and fatigue (including renal disease, hydroelectrolytic imbalance, hepatic cirrhosis, and pulmonary disease) [8, 10]. In these cases, a real discordance may be highlighted between the results of cardiac instrumental examination (first echocardiography), clinical symptoms, and BNP/NT‐proBNP assay.

5. Variations of Plasma B‐Type Natriuretic Peptides, Dependent on Pharmacological Treatment, as Surrogate End‐Point for Treatment of Patients with HF According to the international guidelines, medications such as beta‐ blockers, angiotensin‐converting enzyme (ACE) inhibitors, antialdosterone drugs, and diuretics are titrated based on targets defined by (a) large‐scale

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP

169

clinical trials, (b) patient tolerance, and (c) symptoms and signs of fluid retention [1–5]. However, despite these well established target doses, there is a lack of reliable assessment tools to monitor the response to therapies; change in dyspnea, edema, and body weight are the current, but rather subjective, markers which are used to tailor the treatment in the individual patient. As a matter of fact, there is no specific, quantitative tool, accepted as a surrogate end‐point for treatment of HF patients, able to improve their management and tailor the drug dose on the basis of the individual response [1–5, 7, 10, 31, 32]. Several authors suggested that BNP/NT‐proBNP assay is useful in monitoring and tailoring the medical therapy in HF patients [31–40]. To provide a practical and objective indicator of eVective treatment, the biomarker concentration should be significantly aVected by drugs [25]. Indeed, ACE inhibitors, valsartan, diuretics and nitrates have been shown to reduce plasma natriuretic peptide levels in parallel with hemodynamic and clinical improvement [33–49]. More variable eVects on plasma natriuretic peptide levels have been reported after beta‐blockade eVects, and are, at least in part, attributable to their diVering specificities or to ancillary properties [10, 50]. Acute administration of beta‐blockers may provide an early rise in plasma natriuretic peptides, while sustained treatment with associated improvement in cardiac function, reduction in filling pressure, and cardiac volumes should be associated with a fall in hormone levels [10, 31, 50]. As an example, in Fig. 2, we report the variation in NT‐proBNP levels during a 4‐year follow‐up of a patient with idiopathic dilated cardiomyopathy. NT‐proBNP levels decreased under the upper limit of the reference value (i.e., 150 ng/L), down to 47 ng/L, after both optimized pharmacological treatment (ACE‐inhibitor, beta‐blocking agent, and antialdosterone drugs) and physical training. It is important to note that the left ventricular ejection fraction (LVEF) was always below the normal range, although it significantly partially improved compared to the level before treatment (25% vs. 38%).

6. Prognostic Relevance of Plasma BNP/NT‐proBNP Variations After Treatment Several clinical trials in HF patients [51–56] demonstrated that either the baseline level of BNP and NT‐proBNP or its decrease after treatment hold a powerful prognostic value. In particular, the decrease in peptide concentration under baseline median concentration is associated with treatment eYcacy and clinical improvement, whereas unchanged or increased levels are associated with disease progression and worse prognosis [51, 54].

CLERICO ET AL.

170 Diagnosis

1100

LVEF 25%

NT-proBNP (ng/L)

900 Ramipril Carvedilol

700

Spironolactone Training

500

300 LVEF 35% 100

LVEF 38%

Cut-off 2004

2005

2006

2007

2008

FIG. 2. Time‐course of plasma concentration of NT‐proBNP, measured by ECLIA method [29], and left ventricular ejection fraction (LVEF) in a patient with initial diagnosis of dilative cardiomyopathy, on optimal medical treatment, on a 4‐year follow‐up, after a first event of de‐novo acute heart failure. Notice that clinical improvement is better mirrored by stabilization of peptide level within normal reference values, than by the partial improvement in left ventricular systolic function. The dashed line corresponds to the upper normal reference limit.

As illustrated by the data reported in Fig. 3, the natural history of a chronic HF patient is characterized by withdrawal of symptoms, interrupted by acute decompensation, which requires hospitalization. Disease remission is characterized by lower BNP levels, while during worsening of the clinical condition increased peptide levels are usually observed (Fig. 2). According to some recent studies [51, 54], BNP‐guided treatment should be able to distinguish ‘‘responders,’’ with a better prognosis, from ‘‘nonresponders,’’ on the basis of plasma‐concentration variation. These data suggest that if optimal medical therapy is not able to decrease the BNP levels under the median level of the studied population at baseline, it is likely to be ineVective on prognosis, but this hypothesis should be tested in randomized prospective studies. At present, a relatively small number of randomized studies [35, 36, 57–60] were designed to specifically evaluate the clinical use of BNP/NT‐proBNP assay in monitoring and tailoring the medical therapy in HF patients. Murdoch et al. [35] sought to determine whether titration of vasodilator therapy according to plasma BNP may be of value in the individual optimization of vasodilator therapy in chronic HF. Twenty patients with mild to moderate chronic HF and receiving stable conventional therapy were randomly assigned to titration of ACE‐inhibitor dosage, according to serial

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP LVEF 20%

171

LVEF 15%

Frusemide, Candesartan, Carvedilol, Spironolactone

VAD

CRT

H

3000

BNP (ng/L)

2500 2000

H H

H

H

H

H

1500

H

H

H

H

1000 500 Cut-off

0 1998

2000

2002

2004

2006

FIG. 3. Time‐course of plasma concentration of BNP, measured by an IRMA method [10, 15], and left ventricular ejection fraction (LVEF) in a patient with initial diagnosis of ischemic cardiomyopathy, and severe left ventricular dysfunction, on optimal medical and device‐ treatment, on a 8‐year follow‐up, after the first event of de‐novo acute heart failure, corresponding to hospitalization (H), and to adequate therapeutical choices, such as cardiac resynchronization therapy (CRT), and ultimately destination ventricular assist device (VAD) implantation. Notice that periodical clinical worsening leading to repeated hospitalization is always preceded by a significant increase in peptide level. The dashed line corresponds to the upper normal reference limit.

measurement of plasma BNP or to optimal empirical ACE‐inhibitor therapy for 8 weeks. Only the BNP‐driven approach was associated with significant reductions in plasma BNP concentration throughout the duration of the study and with a significantly greater suppression when compared with empiric therapy after 4 weeks. This study suggests that plasma BNP may be chronically reduced by tailored vasodilator therapy in chronic HF. Troughton et al. [36] studied 69 patients with impaired systolic function (EF <40%) and symptomatic HF (NYHA class II‐IV, on average 2.3). These patients were randomized to receive treatment guided by either plasma NT‐proBNP concentration or standardized clinical assessment. During follow up (minimum 6 months, median 9.5 months), there were fewer total cardiovascular events (death, hospital admission, or HF decompensation) in the NT‐proBNP‐guided group. This study indicates that NT‐proBNP‐guided treatment reduced total cardiovascular events, and delayed time to first event compared with intensive clinically guided treatment [36]. Jourdain et al. [57] studied 220 optimally treated NYHA class II‐III patients, who were randomized to medical treatment according to either current guidelines (clinical group) or to the goal of decreasing BNP plasma

172

CLERICO ET AL.

levels <100 ng/L. The primary combined end‐point was death or hospital stay for HF (median of the follow‐up 15 months). This study pointed out that in optimally treated patients with chronic HF, a BNP‐guided strategy reduced the risk of HF‐related death or hospital stay. The result was mainly explained by a significant increase in ACE‐inhibitor and beta‐blocker dosages in the BNP arm. Pfisterer et al. [58] compared 18‐month outcome of 499 patients aged 60 years or older, undergoing NT‐proBNP versus symptom‐guided therapy in a randomized controlled multicenter trial. These patients were characterized by systolic heart failure (ejection fraction 45%), NYHA class of II or greater. The treatment goal was to reduce symptoms to NYHA class of II or less in the patients’ group with symptom‐guided therapy, while the BNP level of two times or less than the upper limit of normal (i.e., 400 ng/L in patients younger than 75 years, and 800 ng/L in patients aged 75 years or older) and symptoms to NYHA class of II or less in the patients’ group with BNP‐ guided therapy. The data of this study indicated that NT‐proBNP‐guided strategy, as compared to symptom‐guided strategy, did not improve outcome or quality of life in all patients, taken as a whole, not in the subgroup of patients aged 75 years or older, but it improved outcomes only in the subgroup of patients aged 60–75 years [58]. Very recently, a preliminary report of the PRIMA study [59] has been made available online after presentation during an international meeting. In this study, 345 patients [mean (SD) age 72 (12) years] with chronic heart failure (NYHA class from I to III) were randomized in the NT‐proBNP‐ guided treatment arm (174 patients) or in the clinical‐guided treatment one (171 patients) for a median follow up of 702 days (range from 488 to 730). The target level of NT‐proBNP‐guided treatment was the lowest level at discharge or after 2 weeks follow‐up. No significant diVerences were found between NT‐proBNP and clinically guided groups of patients for both total mortality ( p ¼ 0.196) and number of days alive outside the hospital ( p ¼ 0.49) [59]. However, a significant diVerence was found for the mortality rate ( p < 0.001) and the number of days alive outside the hospital ( p < 0.001), when the 101 (out of 174) patients in NT‐proBNP‐guided group (58%), who maintained their target in more than 75% visits, were compared to the clinically guided group [59]. Finally, another preliminary study, called Battlescarred Trial, was recently published only as abstract [60]. In the Battlescarred Trial, 364 patients with symptomatic HF were randomized 1:1:1 to usual care, intensive clinical management or hormone‐guided care incorporating serial measurement of NT‐proBNP levels into the treatment algorithm. Minimum follow‐up was 12 months with median follow‐up of 2.8 years. The primary end‐points were

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP

173

all‐cause mortality and/or admission with decompensated HF. At 2 and 3 years follow‐up, mortality rates did not diVer overall, but significant treatment eVects were observed in those aged 75 years (p < 0.025). The composite end‐point of death and/or admission with HF was reduced only in younger patients receiving hormone‐guided care (p < 0.05 at 1, 2, and 3 years follow‐up compared with standard care). Total days ‘‘alive and not in hospital with HF’’ over 3 years of follow‐up averaged 206 days more in this subgroup than in their peers within the usual care group (p < 0.05). No benefits were observed in those aged >75 years [60]. There are of course some conflicting results when comparing these studies: three studies [35, 36, 57] reported an improved outcome (a lower number of end‐point events) in the group of patients with BNP/NT‐proBNP‐guided therapy, while the other three studies found an improvement only in the subgroup of patients with age lower than 75 years [58–60]. However, these studies were not homogeneous, diVering for the period of follow‐up, for clinical endpoints, and especially for the overall clinical condition of the patients enrolled. Patients enrolled by Pfisterer et al. [58] were older and had more severe disease (74% of patients with NYHA class III) than the other studies [36, 57, 59, 60]. Another important diVerence among these studies was the treatment goal chosen for the BNP/NT‐proBNP‐guided therapy. Jourdain et al. [57] chose a value close to the upper limit of normal reference values as the goal for BNP‐ guided therapy (i.e., 100 ng/L), while Pfisterer et al. [58] chose a value that was two times the upper limit of normal, adjusted for age (i.e., 400 ng/L for patients younger than 75 years and 800 ng/L for older patients), for the NT‐proBNP‐guided therapy group. Moreover, Pfisterer et al. [58] did not specify the number of patients who had reached the treatment goal throughout the follow‐up. The lack of prognostic improvement by the NT‐proBNP‐ guided strategy reported in patients older than 75 years in this study [58] may be due to treatment ineYcacy in this elderly subset, characterized by more advanced disease and comorbidities, such as renal failure. This hypothesis seems to be confirmed by NT‐proBNP levels before and after the treatment reported in the study by Pfisterer et al.: the median value (i.e., more than 2000 ng/L) after 6‐month follow‐up in patients older than 75 years was much higher than the NT‐proBNP value chosen as treatment goal (i.e., two times 800 ng/L), suggesting that a relatively large number of patients older than 75 did not reach the treatment goal [58]. Interestingly, in the PRIMA study [55], a highly significant diVerence in mortality rate was found when only the subgroup of patients in the NT‐proBNP‐guided group, who maintained their target in more than 75% of visits, were compared to the clinical‐guided group.

174

CLERICO ET AL.

7. Meta‐Analysis for Overall Mortality Including All Randomized Clinical Trials Despite great heterogeneity of the randomized studies [35, 36, 57–60] designed to specifically evaluate the clinical use of BNP/NT‐proBNP assay in monitoring and tailoring the medical therapy in HF patients, a meta‐ analysis of the overall mortality data, published before April 2009, may give a further insight. Due to the diVerent study protocol (three randomized groups instead of two) and the partial information on the number of patients, who were enrolled and survived in each randomized arm, we did not include the data reported by the Battlescarred Trial in the first meta‐analysis. Therefore, considering all data reported by other four studies [36, 57–59], a total of 1133 patients (568 in the BNP/NT‐proBNP‐guided group, and 565 in the control group) were enrolled. The death rate (overall mortality for cardiac and extracardiac causes) showed a trend, although not significant, to be lower in the BNP/NT‐proBNP‐guided compared to control group [62/568 (10.9%) vs. 84/565 (14.9%), p ¼ 0.0579 by Yates‐corrected chi‐square test]. Considering the number of patients enrolled (i.e., 1133) and the significance level (i.e., p ¼ 0.0579), the calculated power (1 b error probability) of the chi‐square test is 0.99. On the other hand, using a priori computation and considering an a error of 0.05 with a power of 0.95 and the same small eVect size (w about 0.1), the calculated actual sample size is about 1300 [61]. The results of this first meta‐analysis suggest that only a small increase in the number of patients studied is actually necessary to reach a significant statistical level for the overall mortality for cardiac and extracardiac causes between BNP/NT‐proBNP‐guided and clinically guided treatments, assuming that the observed diVerence in mortality rate (i.e., 10.9% BNP/NT‐ proBNP‐guided vs. 14.9 % clinically guided) between the two treatment strategies is true. As a consequence, it may be very interesting to include also the data reported by the Battlescarred Trial [60] in a new meta‐analysis. We estimated the number of patient survived at 1‐year follow‐up by the all‐ cause mortality rate concerning the standard therapy (i.e., 18.9%) and the hormone‐guided therapy (i.e., 9.1%) [60]. Adding these new data to the meta‐ analysis, we included 1376 patients in the new statistical analysis (689 in the BNP/NT‐proBNP‐guided arm and 687 in the standard treatment‐guided arm, respectively). A significant diVerence in overall mortality ( p ¼ 0.0078 by Yates‐corrected chi‐square test) was found between the standard treatment arm (687 patients with 107 deaths, mortality rate: 15.6%) and hormone‐ guided arm (689 patients with 73 deaths, mortality rate: 10.6%). These data suggest a better outcome of patients followed according to the BNP/NT‐ proBNP‐guided strategy compared to those treated with the standard care,

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP

175

especially for patients with age <75 years and considering a follow‐up lasting 12 months.

8. BNP‐Guided Therapy in Chronic Heart Failure: Instructions for Use Is it possible to draw some general indications from studies regarding the clinical impact of BNP/NT‐proBNP assay in the follow‐up of HF patients? Prognostic relevance of the serial assay of plasma B‐type natriuretic peptide concentration is now well demonstrated in HF patients [51–56]. The principal aim of monitoring natriuretic peptide levels during follow‐up is that to distinguish responder from nonresponder patients. Nonresponders, who usually show a worse prognosis even in the short and medium follow‐up period (6–12 months) [53, 54], should receive a tailored, more intensified treatment compared to the standard one. Randomized clinical trials [57–59] have demonstrated that an enhanced therapy was more frequently recommended in patients of the BNP‐guided therapy arm than in those of the standard treatment arm. This ‘‘intensified’’ therapy produced significant beneficial eVects (i.e., improved outcome and quality of life) in the BNP‐guided therapy patients, compared to controls, in one study [57], and at least in the subset of patients 75 years in two other studies [58, 60]. The results by Pfisterer et al. [58] suggest that even an optimized treatment may not be able to modify the progression of disease in many elderly patients. A recent study [62] has confirmed that NT‐proBNP and age are powerful predictive variables of vital outcome in elderly patients hospitalized for acute dyspnea. The data from the study by Pfisterer et al. [58] should be reanalyzed, subdividing patients into four diVerent groups, according to the mode of decision making (‘‘guided therapy’’ vs. ‘‘nonguided therapy’’), and to the actual response to treatment (‘‘responders’’ vs. ‘‘nonresponders’’). Irrespective of protocol modality (guided or nonguided), patients with a significant decrease in peptide level, compared to baseline level, are likely to show a better prognosis, as demonstrated by the better outcome of patients, who maintained their target, in the PRIMA study [59]. The lack of prognostic improvement in patients with changes in BNP/ NT‐proBNP level at follow‐up evidences the inability of current therapeutic strategies to modify prognosis, rather than the ineYcacy of peptide‐guided strategy. Characterization of this high‐risk subset would allow a deeper insight in the refractory form of HF, making it possible to target the eVort of the scientific community in searching novel therapeutical strategies. Further prospective and randomized clinical studies are necessary to definitively demonstrate whether BNP/NT‐proBNP‐guided therapy is able to significantly improve the outcome of HF patients, especially those with age over 75 years.

176

CLERICO ET AL.

REFERENCES [1] NICE (National Institute of Clinical Excellence), Clinical Guideline 5, in: chronic heart failure. Management of chronic heart failure in adults in primary and secondary care, (July 2003) pp. 1–44 London. [2] K. Swedberg, J. Cleland, H. Dargie, et al., Task Force for the Diagnosis and Treatment of Chronic Heart Failure of the European Society of Cardiology. Guidelines for the diagnosis and treatment of chronic heart failure: executive summary (update 2005): the Task Force for the Diagnosis and Treatment of Chronic Heart Failure of the European Society of Cardiology, Eur. Heart J. 26 (2005) 1115–1140. [3] M.S. Nieminen, M. Bo¨hm, M.R. Cowie, et al., ESC Committe for Practice Guideline (CPG). Executive summary of the guidelines on the diagnosis and treatment of acute heart failure: the Task Force on Acute Heart Failure of the European Society of Cardiology, Eur. Heart J. 26 (2005) 384–416. [4] S.A. Hunt, American College of Cardiology; American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure). ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure), J. Am. Coll. Cardiol. 46 (2005) e1–e82. [5] K.F. Adams, J. Lindenfeld, J.M.O. Arnold, D.W. Baker, D.H. Barnard, K.L. Baughman, Executive summary: HFSA 2006 comprehensive heart failure practice guideline, J. Card. Fail. 12 (2006) 10–38. [6] E. Braunwald, M.R. Bristow, Congestive heart failure: fifty years of progress, Circulation 102 (20 Suppl. 4) (2000) IV14–IV23. [7] M. Emdin, S. Vittorini, C. Passino, A. Clerico, Old and new biomarkers of heart failure, Eur. J. Heart Fail. 11 (2009) 331–335. [8] A. Clerico, F.A. Recchia, C. Passino, M. Emdin, Cardiac endocrine function is an essential component of the homeostatic regulation network: physiological and clinical implications, Am. J. Physiol. Heart Circ. Physiol. 290 (2006) H17–H29. [9] C.M. Balion, P. Santaguida, R. McKelvie, et al., Physiological, pathological, pharmacological, biochemical and hematological factors aVecting BNP and NT‐proBNP, Clin. Biochem. 41 (2008) 231–239. [10] A. Clerico, M. Emdin, Diagnostic accuracy and prognostic relevance of the measurement of the cardiac natriuretic peptides: a review, Clin. Chem. 50 (2004) 33–50. [11] J.A. Doust, P.P. Glasziou, E. Pietrzak, A.J. Dobson, A systematic review of the diagnostic accuracy of natriuretic peptides for heart failure, Arch. Intern. Med. 164 (2004) 1978–1984. [12] A. Clerico, M. Fontana, L. Zyw, C. Passino, M. Emdin, Comparison of the diagnostic accuracy of brain natriuretic peptide (BNP) and the N‐terminal part of the propeptide of BNP immunoassays in chronic and acute heart failure: a systematic review, Clin. Chem. 53 (2007) 813–822. [13] C.M. Balion, R.S. McKelvie, S. Reichert, et al., Monitoring the response to pharmacologic therapy in patients with stable chronic heart failure: is BNP or NT‐proBNP a useful assessment tool? Clin. Biochem. 41 (2008) 266–276. [14] B. Ewald, D. Ewald, A. Thakkinstian, J. Attia, Meta‐analysis of B type natriuretic peptide and N‐terminal pro B natriuretic peptide in the diagnosis of clinical heart failure and population screening for left ventricular systolic dysfunction, Intern. Med. J. 38 (2008) 101–113. [15] M. Emdin, C. Passino, C. Prontera, et al., Comparison of brain natriuretic peptide (BNP) and amino‐terminal proBNP for early diagnosis of heart failure, Clin. Chem. 53 (2007) 1264–1272.

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP

177

[16] W.H. Tang, G.S. Francis, D.A. Morrow, et al., National Academy of Clinical Biochemistry Laboratory Medicine practice guidelines: clinical utilization of cardiac biomarker testing in heart failure, Circulation 116 (2007) e99–e109. [17] H. Ramos, A.J. de Bold, Gene expression, processing, and secretion of natriuretic peptides: physiologic and diagnostic implications, Heart Fail. Clin. 2 (2006) 255–268. [18] A.J. De Bold, Cardiac natriuretic peptides gene expression and secretion in inflammation, J. Investig. Med. 57 (2009) 29–32. [19] H. Yasue, M. Yoshimura, H. Sumida, et al., Localization and mechanism of secretion of B‐type natriuretic peptide in comparison with those of A‐type natriuretic peptide in normal subjects and patients with heart failure, Circulation 90 (1994) 195–203. [20] Y. Sawada, M. Suda, H. Yokoyama, et al., Stretch‐induced hypertrophic growth of cardiocytes and processing of brain‐type natriuretic peptide are controlled by proprotein‐ processing endoprotease furin, J. Biol. Chem. 272 (1997) 20545–20554. [21] J.P. Goetze, Biochemistry of pro‐B‐type natriuretic peptide‐derived peptides: the endocrine heart revisited, Clin. Chem. 50 (2004) 1503–1510. [22] U. Schellenberger, J. O’Rear, A. Guzzetta, R.A. Jue, A.A. Protter, N.S. Pollitt, The precursor to B‐type natriuretic peptide is an O‐linked glycoprotein, Arch. Biochem. Biophys. 451 (2006) 160–166. [23] K.N. Luckenbill, R.H. Christenson, A.S. JaVe, et al., Cross‐reactivity of BNP, NT‐ proBNP, and proBNP in commercial BNP and NT‐proBNP assays: preliminary observations from the IFCC committee for standardization of markers of cardiac damage, Clin. Chem. 54 (2008) 619–620. [24] A. Clerico, S. Del Ry, D. Giannessi, Measurement of natriuretic cardiac hormones (ANP, BNP and related peptides) in clinical practice: the need for a new generation of immunoassay methods, Clin. Chem. 46 (2000) 1529–1534. [25] M. Panteghini, A. Clerico, Understanding the clinical biochemistry of N‐terminal pro‐B‐ type natriuretic peptide: the prerequisite for its optimal clinical use, Clin. Lab. 50 (2004) 325–331. [26] F.S. Apple, M. Panteghini, J. Ravkilde, et al., On Behalf of the Committee on Standardization of Markers of Cardiac Damage of the IFCC. Quality specifications for B‐type natriuretic peptide assays, Clin. Chem. 51 (2005) 486–493. [27] F.S. Apple, A.H. Wu, A.S. JaVe, et al., National Academy of Clinical Biochemistry and IFCC Committee for Standardization of Markers of Cardiac Damage Laboratory Medicine practice guidelines: analytical issues for biomarkers of heart failure, Circulation 116 (2007) e95–e98. [28] S. Vittorini, T. Prontera, G.C. Zucchelli, A. Clerico, Cardiac natriuretic hormones: methodological aspects, Immuno‐Analyse Biol. Spec. 22 (2007) 209–272. [29] C. Prontera, C.G. Zucchelli, S. Vittorini, S. Storti, M. Emdin, A. Clerico, Comparison between analytical performances of polyclonal and monoclonal electrochemiluminescence immunoassays for NT‐proBNP, Clin. Chim. Acta 400 (2009) 70–73. [30] C. Prontera, M. Zaninotto, S. Giovannini, et al., Proficiency testing project for BNP and NT‐proBNP immunoassays: the CardioOrmocheck study, Clin. Chem. Lab. Med. 47 (2009) 762–768. [31] A.M. Richards, J.G. Lainchbury, M.G. Nicholls, R.W. Troughton, T.G. Yandle, BNP in hormone-guided treatment of heart failure, Trends Endocrinol. Metab. 13 (2002) 151–155. [32] A. Clerico, S. Vittorini, C. Passino, M. Emdin, New biomarkers of heart failure, Crit. Rev. Cl. Lab. Sci. 46 (2009) 107–128. [33] I.S. Anand, L.D. Fisher, Y.T. Chiang, et al., Val‐HeFT Investigators. Changes in brain natriuretic peptide and norepinephrine over time and mortality and morbidity in the Valsartan Heart Failure Trial (Val‐HeFT), Circulation 107 (2003) 1278–1283.

178

CLERICO ET AL.

[34] A.M. Richards, R. Doughty, M.G. Nicholls, et al., Plasma N‐terminal pro‐brain natriuretic peptide and adrenomedullin: prognostic utility and prediction of benefit from carvedilol in chronic ischemic left ventricular dysfunction. Australia–New Zealand Heart Failure Group, J. Am. Coll. Cardiol. 37 (2001) 1781–1787. [35] D.R. Murdoch, T.A. McDonagh, J. Byrne, et al., Titration of vasodilator therapy in chronic heart failure according to plasma brain natriuretic peptide concentration: randomized comparison of the hemodynamic and neuroendocrine eVects of tailored versus empirical therapy, Am. Heart J. 138 (1999) 1126–1132. [36] R.W. Troughton, C.M. Frampton, T.G. Yandle, E.A. Espiner, M.G. Nicholls, A.M. Richards, Treatment of heart failure guided by plasma aminoterminal brain natriuretic peptide (N‐BNP) concentrations, Lancet 355 (2000) 1126–1130. [37] K. Kawai, K. Hata, H. Takaoka, H. Kawai, M. Yokoyama, Plasma brain natriuretic peptide as a novel therapeutic indicator in idiopathic dilated cardiomyopathy during beta‐blocker therapy: a potential of hormone‐guided treatment, Am. Heart J. 141 (2001) 925–932. [38] M.G. Nicholls, J.G. Lainchbury, A.M. Richards, R.W. Troughton, T.G. Yandle, Brain natriuretic peptide‐guided therapy for heart failure, Ann. Med. 33 (2001) 422–427. [39] R.W. Troughton, A.M. Richards, M.G. Nicholls, Individualized treatment of heart failure, Intern. Med. J. 31 (2001) 138–141. [40] G. McGeoch, J. Lainchbury, G.I. Town, L. Toop, E. Espiner, A.M. Richards, Plasma brain natriuretic peptide after long‐term treatment for heart failure in general practice, Eur. J. Heart Fail. 4 (2002) 479–483. [41] I.G. Crozier, M.G. Nicholls, H. Ikram, E.A. Espiner, T.G. Yandle, Atrial natriuretic peptide levels in congestive heart failure in man before and during converting enzyme inhibition, Clin. Exp. Pharmacol. Physiol. 16 (1989) 417–424. [42] M. Yoshimura, H. Yasue, H. Tanaka, et al., Responses of plasma concentrations of A type natriuretic peptide and B type natriuretic peptide to alacepril, an angiotensin‐converting enzyme inhibitor, in patients with congestive heart failure, Br. Heart J. 72 (1994) 528–533. [43] N.C. Davidson, W.J. Coutie, D.J. Webb, A.D. Struthers, Hormonal and renal diVerences between low dose and high dose angiotensin converting enzyme inhibitor treatment in patients with chronic heart failure, Heart 75 (1996) 576–581. [44] T. Nishikimi, H. Matsuoka, K. Ishikawa, et al., Antihypertensive therapy reduces increased plasma levels of adrenomedullin and brain natriuretic peptide concomitant with regression of left ventricular hypertrophy in a patient with malignant hypertension, Hypertens. Res. 19 (1996) 97–101. [45] C.G. Missouris, E. Grouzmann, M.G. Buckley, J. Barron, G.A. MacGregor, D.R. Singer, How does treatment influence endocrine mechanisms in acute severe heart failure? EVects on cardiac natriuretic peptides, the renin system, neuropeptide Y and catecholamines, Clin. Sci. 94 (1998) 591–599. [46] Y. Hara, M. Hamada, Y. Shigematsu, et al., EVect of beta‐blocker on left ventricular function and natriuretic peptides in patients with chronic heart failure treated with angiotensin‐converting enzyme inhibitor, Jpn. Circ. J. 64 (2000) 365–369. [47] T. Tsutamoto, A. Wada, K. Maeda, et al., EVect of spironolactone on plasma brain natriuretic peptide and left ventricular remodeling in patients with congestive heart failure, J. Am. Coll. Cardiol. 37 (2001) 1228–1233. [48] W. Johnson, T. Omland, C. Hall, et al., Neurohormonal activation rapidly decreases after intravenous therapy with diuretics and vasodilators for class IV heart failure, J. Am. Coll. Cardiol. 39 (2002) 1623–1629. [49] R. Latini, S. Masson, I. Anand, et al., Valsartan Heart Failure Trial Investigators. EVects of Valsartan on circulating brain natriuretic peptide and norepinephrine in symptomatic chronic heart failure: the Valsartan Heart Failure Trial (Val‐HeFT), Circulation 106 (2002) 2454–2458.

CLINICAL RELEVANCE OF BNP MEASUREMENT IN THE FOLLOW‐UP

179

[50] R. Latini, S. Masson, N. de Angelis, I. Anand, Role of brain natriuretic peptide in the diagnosis and management of heart failure: current concepts, (Review) J. Card. Fail. 8 (2002) 288–299. [51] P. Bettencourt, A. Azevedo, J. Pimenta, F. Frio˜es, S. Ferreira, A. Ferreira, N‐terminal‐pro‐ brain natriuretic peptide predicts outcome after hospital discharge in heart failure patients, Circulation 110 (2004) 2168–2174. [52] R.T. Yan, M. White, A.T. Yan, et al., Usefulness of temporal changes in neurohormones as markers of ventricular remodeling and prognosis in patients with left ventricular systolic dysfunction and heart failure receiving either candesartan or enalapril or both, Am. J. Cardiol. 96 (2005) 698–704. [53] R. Latini, S. Masson, M. Wong, et al., Val‐HeFT Investigators. Incremental prognostic value of changes in B‐type natriuretic peptide in heart failure, Am. J. Med. 119 (70) (2006) e23–e30. [54] S. Masson, R. Latini, I.S. Anand, et al., Val‐HeFT Investigators. Prognostic value of changes in N‐terminal pro‐brain natriuretic peptide in Val‐HeFT (Valsartan Heart Failure Trial), J. Am. Coll. Cardiol. 52 (2008) 997–1003. [55] J. Rosenberg, F. Gustafsson, W.J. Remme, G.A. Riegger, P.R. Hildebrandt, EVect of beta‐ blockade and ACE inhibition on B‐type natriuretic peptides in stable patients with systolic heart failure, Cardiovasc. Drugs Ther. 22 (2008) 305–311. [56] S.W. Waldo, J. Beede, S. Isakson, et al., Pro‐B‐type natriuretic peptide levels in acute decompensated heart failure, J. Am. Coll. Cardiol. 51 (2008) 1874–1882. [57] P. Jourdain, G. Jondeau, F. Funck, et al., Plasma brain natriuretic peptide‐guided therapy to improve outcome in heart failure: the STARS‐BNP Multicenter Study, J. Am. Coll. Cardiol. 49 (2007) 1733–1739. [58] M. Pfisterer, P. Buser, H. Rickli, et al., TIME‐CHF Investigators. BNP guided vs symptom‐ guided heart failure therapy: the Trial of Intensified vs Standard Medical Therapy in Elderly Patients With Congestive Heart Failure (TIME‐CHF) randomized trial, JAMA 301 (2009) 383–392. [59] L. Eurlings, S. van Wijk, A. Balk, et al., Can pro‐brain natriuretic peptide guided therapy of heart failure improve heart failure morbidity and mortality? in: main Outcome of the PRIMA‐Study. Presented to ACC Congress, Orlando March 29th 2009; (Presentation number 402‐14, available online at http://assets.cardiosource.com/PRIMA_Eurlings.ppt. [60] A.M. Richards, J.G. Lainchbury, R.W. Troughton, et al., Abstract 5946: NTproBNP guided treatment for chronic heart failure: results from the Battlescarred Trial, Circulation 118 (2008) S1035–S1036. [61] F. Faul, E. Erdfelder, A.G. Lang, A. Buchner, G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences, Behav. Res. Methods 3 (2007) 175–191. [62] J.L. Reny, O. Millot, T. Vanderecamer, et al., Admission NT‐proBNP levels, renal insuYciency and age as predictors of mortality in elderly patients hospitalized for acute dyspnea, Eur. J. Intern. Med. 20 (2009) 14–19.

This page intentionally left blank

INDEX A Acute-phase reactant, 114 Adipose tissue, 97–98, 117 Agouti-related peptide (AgRP), 97–98 Amlodipine, 88 Amniotic fluid-derived stem (AFS) cells, 148 Antigen presenting cells (APCs), 34 Apolipoprotein E knockout mice, 36 Asymmetric dimethylarginine (ADMA), 74–75, 88 and cardiovascular outcomes, studies on, 80–81 and cardiovascular system, 79 arterial stiVness, 78 atherosclerosis and inflammation, 79 chronic heart failure, 79 endothelial function and hemodynamics, 78 synthesis of, 75–77 transport and metabolism of, 77–78 Atherosclerosis, 28, 30 autoimmunity in, 34 HSPs involvement in, 37 and inflammation, 30–31 and role of infection, 31–33 smoking and, 50 Atrial natriureticpeptide (ANP), 165 B Bacillus Calmette Guerin (BCG) and atherogenesis, 38, 39, 58 Biomarker, of cardiovascular risk, 88. See also Symmetric dimethylarginine (SDMA) Biomarkers, clinical validation for predicting risk, 2–3, 19–20 and Bayesian statistics, 9–12 criteria for determining clinical consequence, 16–19 diagnostic models, 12 distribution of data, 9

prognostic models and, 13 ROC curves as diagnostic tool, 4–7 RR/OR comparison, with ROC curves, 7–8 RR/OR ratios as diagnostic tools, 3 weaknesses of, 13–15 synergic biomarker, 15–16 Bone marrow (BM) stem cells, 147–148 B-type naturetic peptide (BNP), 12 circulating levels of, 167–168 age and sex, eVect of, 168 assays for, 167 production and secretion of, 165–166 C Cardiovascular disease (CVD), 28 anti-HSP60 titers and, 50–51, 54–57 higher risk of, 50–51 risk factors for, 30 role of inflammation in, 31 serum HSP70 concentration and, 35 statins, use of, 51 Cardiovascular risk factors and antibodies to HSPs animal models, 48 human studies, 48–53 CD34, as HSC marker, 148 C/EBP family members, 123–124 Chronic infectious microorganisms, 31–32 Clinical reclassification and synergic biomarker, 16 Collagen, 140–141 Confidence interval, 14, 21 Connective tissue growth factor (CTGF), 152 C-reactive protein (CRP), 112 concentrations in inflammation, 113–114 age and sex factor, 114–115 birth weight, 115 BMI and, 117 cigarette smoking and, 115 dietary fiber and, 117–118 ethnic diVerences and, 115–116 181

182

INDEX

C-reactive protein (CRP) (cont. ) Mediterranean diet and, 116–117 and omega-3 fatty acids, 116 and socioeconomic position, 116 Western diet and, 116 functions of, 112–113 and genetic polymorphisms, 124, 125 interaction PhC in damaged membranes, 112 proinflammatory cytokine inducers of, 118 IL-6, 118–119 IL-17, 122–123 IL-1 family, 119–121 TNF, 121–122 signaling through IL receptors, 123–124 synthesis and structure, 112 c-statistic, 5–6, 13 Cumulative distribution analysis, 4 D Decorin, 141, 145 Dimethylarginine dimethylaminohydrolase (DDAH), 77 Discoid domain receptor (DDR) signaling, 154–155 Dot cells, 149 E E-cadherin, 148 Electrochemiluminescence (ECLIA) method, 167 Endothelial dysfunction, 74 Endothelial nitric oxide synthase (eNOS), 74 Endothelium, role in vascular homeostasis, 74 F False positive rate (FPR), 4, 7, 10 Fetal ECM, 140 Fetal skin collagen in, 140–141 development of, 140 ECM of, 140 HA in, 141 proteoglycan ECM modulators, role of, 141 Fetal skin wound healing, 139. See also Fetal skin cellular inflammatory mediators, role of

fibroblasts, 150–151 neutrophils, 150 platelets, 149–150 collagen content of, 142–144 cytokines connective tissue growth factor, 152 transforming growth factor-beta, 151–152 cytokines, role of CTGF, 152 FGFs, 153 interleukins, 154 PDGF, 153–154 VEGF, 152–153 Wnts, 154 ECM adhesion proteins, 143–145 HA content of, 143 role of ECM proteoglycan modulators, 145–146 scarless fetal wound phenotype, 141 scarless repair capacity and nature of, 146 gestational age and wound size, eVect of, 146–147 mechanisms of, 147 molecular control of, 154–155 stem cells and Dot cells, 149 fetal and postnatal epidermal stem cells, 148–149 hematopoietic stem cells, 147–148 Fibroblast growth factors (FGFS), 153 Fibromodulin, 141, 145 Fibronectin, 143–144 Framingham risk score, 10, 13 G Glomerular filtration rate (GFR), 83–86 H Hazard ratio (adjusted RR), 3 Heart failure (HF), chronic, 164–165 BNP/NT-proBNP assay, role of, 165, 169 clinical impact in follow-up of HF patients, 175 NT-proBNP levels, in idiopathic dilated cardiomyopathy, 169 randomized studies, on clinical use of BNP/NT-proBNP assay, 173

INDEX Battlescarred Trial, 172–173 BNP-guided strategy, 171–172 meta-analysis of overall mortality data, 174–175 NT-proBNP-guided treatment, 171 NT-proBNP versus symptom-guided therapy, 172 PRIMA study, 172 titration of vasodilator therapy, 170–171 variation in plasma natriuretic peptide level, 169 prognostic value, after treatment, 169–173 Heat shock protein 47 (HSP 47), 142 Heat shock proteins (HSPs), 28 acute coronary syndrome and changes in antibody titers, 56, 58 antibody titers to observational studies on, 54–55 prospective studies on, 55–57 and apoptosis, 44 and atherogenesis, 34–36 animal models, 36, 38 autoimmune responses, 45–47 eVects of tolerization, 58 classification of, 28 discovery of, 28 endothelial cells and HSP expression, 39, 43 function of, 29 HSP60 expression, 34 HSP70 expression, 34–35 HSP expression in cardiac myocytes, 43 lymphocytes and HSP expression, 44 monocyte/macrophages, HSP expression in, 44 relationship with HSP antibodies and atherosclerosis, animal studies on mouse, 41 rabbit, 42 rat, 41 serum anti-HSP-antibodies and infection, 47 smooth muscle cells (SMCs) and HSP expression, 43 soluble HSPs, 44–45 Hematopoietic stem cells (HSCS), 147–148 Hep3B hepatoma cells, 121 Homeobox genes, 155 Hosmser–Lemeshow test, 16, 21 Hyaluronic acid (HA), 141

183 I

IL-6 receptor (IL-6R), 119 IL-1 receptors (IL-1R), 120 Inflammation acute-phase reaction, 114 and atherosclerosis, 30–31 definition of, 113 low-grade inflammation, 113–114 role in scar formation, 147 Inflammatory biomarkers, 16 Integrin 1, 148 Intercellular adhesion molecule (ICAM)-I, 31 Interleukins-6 (IL-6), 118–119 Interleukins-17 (IL-17), 122–123 Interleukins-1 (IL-1) family, 119–121 Interleukins, role in scarless repair, 154 Intradialytic hypotension (IDH), 86–87 L LDL-receptor deficient (LDL-RD) mouse, 36 Left ventricular ejection fraction (LVEF), 169, 170 Leptin, role in obesity, 97–98 Likelihood ratio, 4, 22 Low-grade inflammation, 113 in common diseases/conditions, 113–114 CRP concentrations in, 113 Lysyl oxidase, 141, 146 M Matrix metalloproteinases (MMPs), 141, 146 Melanocortin-3 receptor (MC3R), 97–98 Melanocortin-4 receptor (MC4R), 97, 99–100 functional alterations of, 102 mutations in, 100–101 and clinical phenotype of individuals, 102–103 implications in clinical management of obesity, 103 loss-of-function mutation, 101 Melanocortin system, 97–99 -Melanocyte stimulating hormone (-MSH), 97 Mitogen-activated protein kinase (MAPK), 154 Mucous associated lymphoid tissue (MALT), 45 Myofibroblasts, 151

184

INDEX N

National Cholesterol Education Program (NCEP), 6 Negative predictive value (NPV), 3, 9 NG-monomethyl L-arginine (L-NMMA), 75 Nitric oxide (NO), 74 Noncompetitive immunometric assays, 167 N-terminal fragment of proBNP (NT-proBNP), 165–168 O Obesity, 96–97, 117. See also Melanocortin-4 receptor (MC4R) genetics of, 97 Odds (OR) ratio, 3, 22 P Palmitic acid, 116 P63, as HSC marker, 148 Pentraxin protein family, 112 Platelet-derived growth factor (PDGF), 150, 153–154 POMC–MC4R system, deficiency in, 97 Positive predictive value (PPV), 3, 9, 11–12 Posterior probability, 9 Postnatal epidermal stem cells, 148 Prevalence. See Prior probability Prior probability, 9 Proopiomelanocortin (POMC), 97 Protein arginine methyltransferases (PRMTs), 75

examples of series of idealized ROC curves, 4, 5 and RR/OR, comparison, 7–8 S S-adenosylmethionine, 75 Sca-1, as HSC marker, 148 Scarless healing. See Fetal skin wound healing Scarring, 139 Scar tissue, 138–139 Signal transducer and activator of transcription 3 (STAT3), 124 Skin-derived stem cells (SKPs), 149 Smoking and low HSP60 titers, 50 Splenectomy, 32 Stand-alone marker, 15 Statins, 51 Stem cell niche, 148 Stromal-derived factor-1 (SDF-1), 149 Stromal stem cells, 147–148 Symmetric dimethylarginine (SDMA), 75, 88–89 cardiac function, 83–84 cardiovascular homeostasis intradialytic hypotension, 86–87 renal function, 84–86 and cardiovascular outcomes, 87 and development of hypertension, 84 eVect on reactive oxygen species, 83 endothelial function, 79, 83 future research area, 88–89 role in vascular homeostasis and vascular disease, studies on, 82–83 synthesis of, 75–77 transportation of, 77–78

R T Receiver operator characteristic (ROC) analysis, 3 Receptor tyrosine kinase (RTK), 154–155 Regulatory T cells (Treg), 31 Relative risk (RR) ratio, 3, 22 Rel/NF-B protein family, 123 ROC curves construction of, 4–7 c-statistic, 5–6

Tenascin, 144 Th17 cells, 122 Tissue damage, 113 Tissue-derived inhibitors (TIMPs), 146 TLR 4/CD14 complex, 44–45 T lymphocytes, 31 Transforming growth factor-beta (TGF- ), 151–152

INDEX Troponin I, 12 True positive rate (TPR), 4, 7, 10

185

Vascular endothelial growth factor (VEGF), 152–153 von Willbrand factor (vWF), 38

V W Valsartan, 88 Vascular-associated lymphoid tissue (VALT), 45

Wnts, role in scarless repair, 154

This page intentionally left blank