UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Modulation of versican expression by smooth muscle cells and myofibroblasts in cardiovascular disease McDonald, Paul Christopher 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2003-792803.pdf [ 21.77MB ]
JSON: 831-1.0091418.json
JSON-LD: 831-1.0091418-ld.json
RDF/XML (Pretty): 831-1.0091418-rdf.xml
RDF/JSON: 831-1.0091418-rdf.json
Turtle: 831-1.0091418-turtle.txt
N-Triples: 831-1.0091418-rdf-ntriples.txt
Original Record: 831-1.0091418-source.json
Full Text

Full Text

Modulation of Versican Expression by Smooth Muscle Cells and Myofibroblasts in Cardiovascular Disease by P A U L CHRISTOPHER M C D O N A L D B.Sc , The University of Victoria, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES Department of Pathology and Laboratory Medicine; Faculty of Medicine We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A December 2002 © Paul Christopher McDonald, 2002 U B C Rare Books and Special Collections - Thesis Authorisation Form Page 1 of 1 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the Un i v e r s i t y of B r i t i s h Columbia, I agree that the Li b r a r y s h a l l make i t f r e e l y a v a i l a b l e for reference and study. I further agree that permission for extensive copying of t h i s thesis for s c h o l a r l y purposes may be granted by the head of my department or by h i s or her representatives. It i s understood that copying or p u b l i c a t i o n of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of The U n i v e r s i t y of B r i t i s h Columbia Vancouver, Canada Date http://www.library.ubc.ca/spcoll/thesauth.html 12/20/2002 Abstract Aberrant accumulation of proteoglycans, particularly versican, is a hallmark of vascular lesions and is potentially important in the development of myxomatous cardiac valve lesions. My overarching hypothesis is that the dysregulation of proteoglycan expression plays an important role in the pathogenesis of accelerated vascular disease and valvular heart disease. I tested three specific hypotheses relating to this overarching framework. First, that the overexpression of versican by smooth muscle cells (SMCs) is due to modulation of its expression by specific molecules present within the tissue microenvironment. Second, that versican is spatially associated with modified lipoproteins in human atheromatous disease. Third, that the exposure of cardiac valves to certain anorexigens (appetite suppressants) produces a proteoglycan-rich lesion morphologically distinct from lesions characterizing other valve diseases and related to the dysregulation of versican expression by myofibroblasts (MFBs). To address the first hypothesis, growth factor-mediated modulation of versican expression by rat aortic SMCs was examined. Treatment of SMCs with either epidermal growth factor (EGF) or insulin-like growth factor-I (IGF-I) significantly upregulated versican mRNA levels and increased core protein expression. Stimulation of SMCs with a combination of EGF and IGF-I further increased versican expression. The oxidative status of lipoproteins present in coronary arteries from human heart allografts and their localization relative to versican and vessel wall cells were examined to address the second hypothesis. Lipoproteins in mild and severe lesions were oxidatively modified, although certain forms of oxidative modification only appeared in severe lesions. The oxidized lipoproteins were often associated with versican and intimal SMCs. The third hypothesis was addressed by examining the geometric and compositional characteristics of normal and diseased heart valves. Valves excised from patients exposed to anorexigens exhibited a glycosaminoglycan-rich lesion distinct from lesions characterizing other myxomatous valve diseases. The myxomatous process was present in both the valve onlays and ii the valve proper. Myxomatous regions were rich in versican and MFBs cultured from human cardiac valve tissue expressed versican. In summary, these data provide key insights into the modulation of versican expression and lay the groundwork for future experiments aimed at further understanding the regulation of versican expression and development of potential novel therapies for the treatment of cardiovascular disorders. iii T A B L E OF C O N T E N T S Abstract ii List of Tables vii List of Figures viii Abbreviations ix Acknowledgements xii C H A P T E R I B A C K G R O U N D 1 1.1 The Extracellular Matrix of the Cardiovascular System 1 1.1.1 Normal Structure and Function of the Vessel Wall E C M 1 1.1.2 Wound Repair and Response to Injury 3 1.1.3 Importance of Versican in Tissue Pathology 5 1.2 Versican 5 1.2.1 Protein Structure 6 1.2.2 Gene Structure 8 1.2.3 Tissue Distribution During Embryonic Development 11 1.2.4 Distribution in Adult Tissue 11 1.2.5 Ligands for Versican 12 1.2.6 Modulation of Versican Expression by Physical Stimuli 15 1.2.7 Modulation of Versican Expression by Growth Factors and Cytokines 16 1.3 Functions of Versican in Atherogenesis 20 1.3.1 Cell Adhesion, Migration and Proliferation 20 1.3.2 Lipid Retention and Modification 22 "Response-to-Retention" Hypothesis 23 Interactions Between Proteoglycans and Low-Density Lipoprotein 23 Colocalization of Versican and Lipoproteins in the Arterial Wall 25 Alterations in Lipoprotein Structure Affect Proteoglycan-LDL Interactions 26 Alterations in Proteoglycan Structure Affect Proteoglycan-LDL Interactions 28 1.4 Biological Significance of Versican in Cardiovascular Pathology 29 1.4.1 Transplant Vascular Disease 29 Importance of Lipids in T V D 33 1.4.2 Anorexigen-Associated Valve Disease 36 1.4.3 Rheumatic Heart Valve Disease 39 1.4.4 Floppy Mitral Valve Disease 41 1.4.5 Carcinoid Heart Valve Disease 42 1.5 Dysregulation of Vascular Smooth Muscle Cells in Cardiovascular Disease 42 1.5.1 Intimal SMC Accumulation in T V D 43 1.5.2 Extracellular Matrix Production 46 1.6 Role of Myofibroblasts in Cardiovascular Disease 48 1.6.1 Definition and Phenotypic Characterization of MFBs 49 1.6.2 Mechanism of Phenotypic Modulation in M F B Activation 50 IV 1.6.3 Role of MFBs in Wound Repair and Inflammation 51 1.6.4 Cardiac Valvular Interstitial Cells are Tissue-Specific MFBs 52 CHAPTER II HYPOTHESES AND EXPERIMENTAL AIMS 54 CHAPTER III MATERIALS AND METHODS 56 3.1 Selection of Patient Material 56 3.1.1 Human Coronary Arteries 56 3.1.2 Human Cardiac Valves 56 3.2 Tissue Processing 57 3.2.1 Human Coronary Arteries 57 3.2.2 Human Cardiac Valves 59 3.3 Cell Culture 61 3.3.1 Rat Aortic Smooth Muscle Cells 61 3.3.2 Human Cardiac Valve Myofibroblasts 62 3.3.3 Treatment of Vascular SMCs with Growth Factors 63 3.4 Immunoassays 63 3.4.1 Antibodies 64 3.4.2 Immunocytochemistry 67 3.4.3 Immunohistochemistry 69 3.5 Reverse Transcription - Polymerase Chain Reaction (RT-PCR) 70 3.5.1 Primers 70 3.5.2 Total RNA Extraction 71 3.5.3 RT-PCR 71 3.6 Digital Imaging 72 3.6.1 Cell Layers 72 3.6.2 Tissue Sections 73 3.6.3 Agarose Gels 73 3.7 Quantitation 73 3.7.1 Immunocytochemical Staining of Cultured Vascular SMCs 74 3.7.2 Geometric Analysis of Cardiac Valves 74 3.7.3 Compositional Analysis of Cardiac Valves 79 3.8 Statistical Analysis 82 3.8.1 Quantitative Immunocytochemistry 82 3.8.2 Geometric and Compositional Analyses of Cardiac Valves 83 3.8.3 Discriminant Analysis 83 CHAPTER IV EFFECTS OF EGF AND IGF-I ON VERSICAN EXPRESSION 85 4.1 Introduction 85 4.2 Results 87 4.2.1 PDGF-BB-Mediated Versican Core Protein Expression by Vascular SMCs 87 4.2.2 EGF Upregulates Versican Expression by Vascular SMCs 89 4.2.3 IGF-I Upregulates Versican Expression by Vascular SMCs 95 4.2.4 EGF + IGF-I Further Upregulate Versican Expression by Vascular SMCs 97 4.3 Discussion 101 CHAPTER V OXIDIZED LDL AND VERSICAN IN TVD 107 5.1 Introduction 107 5.2 Results 110 5.2.1 Oxidized Lipoproteins in Vessel s with Mild T V D 110 v 5.2.2 Lipoprotein Oxidation in Severe T V D 112 5.2.3 Association of OxLDL with Vessel Wall Cells 114 5.2.4 Colocalization of O x L D L and Versican 115 5.3 Discussion 118 CHAPTER VI PROTEOGLYCAN EXPRESSION IN CARDIAC VALVES 124 6.1 Introduction 124 6.2 Results 126 6.2.1 Geometry 126 6.2.2 Composition 131 6.2.3 Discriminant Analysis 134 6.2.4 Versican Expression by Cardiac Valve Myofibroblasts 140 6.3 Discussion 143 CHAPTER VII CONCLUSIONS AND FUTURE PROSPECTS 151 REFERENCES 154 vi List of Tables Table 1 Versican isoforms produced through alternative splicing 9 Table 2 Comparison of gene structure for human and murine versican 10 Table 3 Patient characteristics for valve morphology study 58 Table 4 Antibodies used for immunocytochemistry and immunohistochemistry 65 Table 5 Adjusted means from mixed effect model analyses for leukocytes and vessels 133 vii List of Figures Figure 1 Versican core protein structure 7 Figure 2 Putative roles of versican in the diseased vessel wall 13 Figure 3 Photomicrographs of proteoglycans and lipids in T V D 30 Figure 4 Hypothetical framework for vascular injury and repair in allografts 32 Figure 5 Histologic section of a valve leaflet from a female exposed to anorexigens 37 Figure 6 Leukocytes and blood vessels in a valve leaflet exposed to anorexigens 38 Figure 7 Important features of diseased cardiac valves 40 Figure 8 Role of soluble mediators in SMC migration, growth and E C M production 44 Figure 9 Sectioning procedure for human cardiac valves 60 Figure 10 Characterization of vascular SMCs isolated from rat aorta 66 Figure 11 Color segmentation profile for quantiation of versican expression 75 Figure 12 Methodology for obtaining total area measurements with ImagePro Plus 77 Figure 13 Color masks for valve study 80 Figure 14 Color segmentation files for valve compositional analysis 81 Figure 15 PDGF-BB exposure enhances versican expression by SMCs 88 Figure 16 Quantitation of versican expression by SMCs after PDGF-BB exposure 90 Figure 17 Versican expression by SMCs after stimulation with EGF or IGF-1 91 Figure 18 EGF quantitatively increases versican core protein expression by SMCs 92 Figure 19 EGF alters the level of versican mRNA expressed by SMCs 94 Figure 20 IGF-I-mediated upregulation of versican core protein expression by SMCs 96 Figure 21 IGF-I alters the level of versican mRNA expressed by SMCs 98 Figure 22 Exposure to EGF + IGF-I further upregulates versican expression by SMCs 99 Figure 23 EGF + IGF-I quantitatively enhances versican expression by SMCs 100 Figure 24 Addition of EGF + IGF-I increases versican mRNA levels in SMCs 102 Figure 25 Interactions between lipoproteins and proteoglycans in the vessel wall 108 Figure 26 OxLDL in the vessel wall of a coronary artery with mild T V D I l l Figure 27 OxLDL in the vessel wall of a coronary artery with severe TVD 113 Figure 28 Colocalization of oxLDL and cells in mild versus severe T V D 116 Figure 29 Colocalization of versican and oxLDL in mild versus severe T V D 117 Figure 30 Mixed effect model analyses of areas for mitral and aortic valves 127 Figure 31 Mixed effect model analyses of onlay size and thickness for cardiac valves 128 Figure 32 Illustrative cross sections of normal and diseased mitral and tricuspid valves 129 Figure 33 Illustrative cross sections of normal and diseased cardiac valves 130 Figure 34 Tissue constituents of normal and diseased cardiac valves 132 Figure 35 Leukocytes and vessels in diseased cardiac valves 135 Figure 36 Discriminant analysis of leaflets of normal, rheumatic, and floppy valves 136 Figure 37 Discriminant analysis of mitral valves 138 Figure 38 Discriminant analysis of aortic valves 139 Figure 39 Spatial association between GAGs and versican in cardiac valve disease 141 Figure 40 Versican expression by cultured human cardiac valve MFBs 142 Figure 41 PDGF-BB-mediated stimulation versican expression by MFBs 144 viii Abbreviations Abbreviation Definition % percent > greater than or equal to "sulfate 5-HT 5-hydroxytryptamine A B C - A P streptavidin biotin complex - alkaline phosphatase Ang II angiotensin II A N O V A analysis of variance apo(a) apolipoprotein (a) apoB apolipoprotein B apoE apolipoprotein E AT-1 R angiotensin type 1 receptor A T - 2 R angiotensin type 2 receptor bFGF basic fibroblast growth factor BSA bovine serum albumin C4S chondro itin-4 -sulfate C6S chondroitin-6-sulfate Caz+ calcium CD31 cluster of differentiation 31 CD44 cluster of differentiation 44 cDNA clonal D N A CREB cAMP responsive element-binding CRP C-reactive protein CS chondroitin sulfate CSPG chondroitin sulfate proteoglycan CTGF connective tissue growth factor dexfen dexfenfluramine D N A dideoxyribonucleic acid DPBS Dulbecco's phosphate-buffered saline DS dermatan sulfate E C M extracellular matrix EDTA ethylenediaminetetraacetic acid EGF epidermal growth factor EGF-R epidermal growth factor receptor E R K extracellular regulated kinase fen fenfluramine g gram G l N-terminal globular domain G3 C-terminal globular domain G A G glycosaminoglycan GITC guanidine isothiocyanate glu glutamate GM-CSF granulocyte macrophage-colony stimulating factor H A hyaluronan ix H A B R hyaluronan-binding region H'NE 4-hydroxynonenal H N E - L D L hydroxynonenal-modified low-density lipoprotein HS heparan sulfate HSI hue/saturation/intensity i.u. international units IGF-I insulin-like growth factor I IGF-I R insulin-like growth factor receptor IgG immunoglobulin G IL-1P interleukin-1 beta IVUS intravascular ultrasound JNK c-jun NH2-terminal kinase kba kilobase kDa kiloDalton kg kilogram L liter L A D left anterior descending L D L low-density lipoprotein L D L - R low-density lipoprotein receptor L O P - L D L linoleic acid oxidation product-modified low-density lipoprotein Lp(a) lipoprotein (a) LpL lipoprotein lipase LRP LDL-receptor-related protein lys lysine lysoPC lysophosphatidylcholine M molar M A P K mitogen-activated protein kinase M C D B Molecular and Cellular Developmental Biology M D A malondialdehyde M D A - L D L malondialdehyde-modified low-density lipoprotein M F B myofibroblast mg milligram ml milliliter mm millimeter mM millimolar mRNA messenger ribonucleic acid ng nanogram °C degrees Celsius oxLDL oxidized low-density lipoprotein PDGF platelet-derived growth factor phen phentermine PI3K phosphatidylinositol-3-kinase PKB protein kinase B PKC protein kinase C RER rough endoplasmic reticulum RGB red/green/blue RNA ribonucleic acid X rRNA ribosomal ribonucleic acid RT room temperature R T K receptor tyrosine kinase RT-PCR reverse transcriptase - polymerase chain reaction S E M standard error of the mean SLC secondary lymphoid tissue chemokine S M myosin smooth muscle myosin SMase sphingomyelinase SMC smooth muscle cell TBST Tris-buffered saline + tween 20 TGF-(3 transforming growth factor beta TNF-a tumor necrosis factor alpha T V D transplant vascular disease U units U B C University of British Columbia VO versican isoform 0 V I versican isoform 1 V2 versican isoform 2 V3 versican isoform 3 V E G F vascular endothelial growth factor V L D L very low-density lipoprotein w/v weight/volume X R E xenobiotic responsive element a alpha a -SM actin alpha smooth muscle actin P beta p-actin beta actin e-amino epsilon amino y-actin gamma actin microgram jim micrometer xi Acknowledgements The study of versican in vascular disease is a research focus that is both very complex and highly intriguing. As this area of research continues to expand, it has become increasingly difficult and even counterproductive to "go it alone". Likewise, the successful completion of my thesis would not have been possible without the help and support of many individuals. I would like to take a moment to thank several people who have contributed in various ways to this body of work. First, I would like to thank my Doctoral Supervisor, Dr. Bruce McManus. During my tenure as a graduate student, Bruce provided me with all of the tools necessary for successful completion of my degree. Bruce ensured the presence of a secure, productive and team-oriented environment in which I could learn and grow as a trainee. Further, he ensured that no boundaries, financial or scientific, existed that might impede the forward momentum of my project. Bruce also made certain that all doors remained open for me, providing me with many opportunities to enrich my experience as a graduate student. Most importantly, despite an increasingly hectic schedule of his own, Bruce always made himself available to help and support me as I tackled both scientific and personal challenges. I consider myself very fortunate to have been mentored by Bruce and, over the years, he and I have developed a strong professional bond and, more importantly, a lasting friendship. I would also like to take the opportunity to thank Bruce's wife, Janet McManus. Janet played a pivotal role in my success as a graduate student. In addition to "showing me the ropes" in practical matters, including how to negotiate Bruce's demanding schedule, Janet spent countless hours helping me work on and complete numerous projects. From the preparation of formal presentations to technical and scientific input on many projects related to the work set out in my thesis, Janet taught me a tremendous amount over the course of my degree. Further, she xii always played the gracious host, providing me with nourishment at parties and during weekend work sessions at her home. Most importantly, Janet helped to support me through several professional and personal challenges and has become both a trusted colleague and a close friend. I would like to thank the members of my Graduate Supervisory Committee, Drs. Rick Hegele, Issy Laher, Haydn Pritchard, and Jiri Frohlich. Their support and guidance throughout my tenure as a graduate student has been invaluable. Together, the Committee helped to steer my project and my training in the appropriate direction. They were always available for and open to my questions and concerns. Moreover, they provided me with a rich resource of scientific knowledge and experience. Finally, the Committee provided a system of checks and balances that helped to ensure appropriate progress was made. I would like to thank the members of the Cardiovascular Research Laboratories and the McDonald Research Laboratories/iCAPTUR4E Center for all of their help and support during my training. The process of science is, without doubt, a team effort and I am indebted to the members of these laboratories for their technical expertise, scientific knowledge and helpful advice. Special thanks must be given to the Summer Students, Directed Studies Students, Cooperative Education Students, Graduate Students, Laboratory Technicians and Administrative Assistants who have dedicated countless hours and various degrees of scientific and administrative expertise to the projects related to my thesis. I would like to acknowledge the funding agencies that have been critical to the successful completion of my thesis. In particular, I would like to thank the Heart and Stroke Foundation of Canada for providing me with salary support. I would also like to thank the Heart and Stroke Foundation of British Columbia and Yukon, the Canadian Institutes of Health Research, the British Columbia Transplant Society, St. Paul's Hospital Foundation and Wyeth Ayerst Research for their financial contributions to my scientific endeavors. Without the combined support of these agencies, the work contained within my thesis would not have been possible. xiii The completion of my degree would not have been possible without the love and support of my family. I would like to give special thanks to my Mom and Dad. They have always put my desires and my well-being ahead of their own and have made many sacrifices over the years so that I could follow my dreams. I deeply appreciate their steadfast determination to see me through my degree. I would also like to thank my Mother-in-law and Father-in-law for their support. They have always shown a keen interest in my studies and have stood by me patiently, helping me where needed as I have confronted and surmounted the challenges set before me. My biggest "thank you" is reserved for my wife, Treena. Her inexhaustible love for me and her enduring faith and trust in me have provided me with the courage and determination t pursue my goals. Together with her amazing strength of character, Treena's passion for life and for science have taught me volumes and have provided me with the ability to surmount the toughest of obstacles. I am forever indebted to Treena for her understanding during the many evenings, weekends and holidays spent working. Thank you for everything. Finally, I dedicate this thesis to my wife, Treena and my son, Justin. xiv 1 CHAPTER I BACKGROUND 1.1 The Extracellular Matrix of the Cardiovascular System The tissues of the cardiovascular system, like all vertebrate tissues, are composed of cellular constituents situated within and interconnected by a composite molecular framework called the extracellular matrix (ECM). The exact make-up of the E C M depends on the specific structure being examined, but its basic biochemical composition includes collagens, elastic fibers, glycoproteins and proteoglycans (*). Far from a static structural entity, the compositional characteristics of the E C M are constantly in flux and are controlled by complex mechanisms related to the differential regulation of synthesis and degradation of each component C1* 2\ The E C M is responsible for many physical and mechanical properties ascribed to tissues, including tensile strength, elastic recoil, compressibility, and viscoelasticity 0). These mechanical properties are not only essential for normal tissue function, but also serve to protect resident cells from external mechanical forces such as shear and elastic recoil (3X Further, the E C M is an important participant in the regulation of cell adhesion, migration and proliferation, particularly during development and disease processes 0). Two vital components of the cardiovascular system, blood vessels and cardiac valves, rely on a properly developed E C M to carry out their structural and functional duties. 1.1.1 Normal Structure and Function of the Vessel Wall E C M The normal vessel wall consists of three basic layers called the tunica intima, tunica media and tunica adventitia (4). The E C M in each of these layers is specially formulated to provide specific functions. The adventitia, the outermost connective tissue layer, contains largely fibrillar collagen and some elastic fibers to provide strength and rigidity as well as a certain degree of elasticity 0). The media of muscular arteries is composed of smooth muscle 1 cells (SMCs) arranged circumferentially and surrounded by limited amounts of E C M containing collagen, elastic fibers and proteoglycans (4X The media is bounded at the intimal surface by a fenestrated sheet of elastic fibers called the internal elastic lamina and on the adventitial side by a less defined external elastic lamina (4). Thus, the E C M of the media offers a combination of tensile strength and viscoelasticity. The proteoglycan- and glycosaminoglycan (GAG)-rich intimal layer imparts compressible, viscoelastic properties on the vessel wall ( l \ This subendothelial layer is very thin in most normal vessels, although an intimal cushion consisting of intimal SMCs embedded in a loose matrix is present in some vessels of both infants and adults (5> 6). Maintaining a balanced network of components in each layer of the vessel wall is critical to its physiologic function, and deviations from this balance lead to a variety of vascular pathologies. The histologic structure of the normal cardiac valve, like that of the vessel wall, displays a layered architecture (7X In the mitral valve, for example, each leaflet is completely surrounded by a layer of endothelial cells that is continuous with the endocardial surface of the heart. Elastic lamellae are present immediately beneath the endothelium on the atrial aspect of the valve (the atrialis), together with a discontinuous layer of SMCs (8X The subendothelial connective tissue layer of the atrialis, termed the spongiosa, is composed of myofibroblasts (MFBs) embedded in a loose, hydrated E C M enriched in proteoglycans and GAGs (?). The spongiosa is thinnest near the base of attachment and expands toward the free margin of the valve. The proteoglycan-rich composition of the spongiosa provides the valve with compressive and viscoelastic properties that allow it to function normally while protecting embedded cells from the forces generated as blood flows from the left atrium to the left ventricle. In contrast, the fibrosa, the connective tissue layer on the ventricular aspect of the valve, is enriched in fibrillar collagen, providing the valve with structural rigidity (8). The chordae tendineae are attached to the ventricular surface and are composed of a collagen-rich core covered by an endothelial layer (7). 2 As with the vasculature, maintenance of the delicate balance among the E C M components in each layer of the cardiac valve is essential if this tissue is to perform properly. Disruptions or alterations in the normal composition of the valvular E C M through mechanical or other injury may lead to pathologic remodeling processes that alter valve structure and function. Chronic rheumatic valve disease, floppy mitral valve disease and carcinoid disease are examples of clinically relevant valve pathologies involving substantial E C M remodeling and valve dysfunction. Exposure of valves to specific drugs may also result in E C M remodeling and valve dysfunction (9> 1 0 \ In particular, exposure of morbidly obese individuals to anorexigens, pharmacological agents that control appetite, has been associated recently with myxomatous changes in valvular E C M leading to valvular regurgitation ( n X 1.1.2 Wound Repair and Response to Injury The generic "wound healing" or "wound repair" paradigm embodies several important concepts that underlie the pathogenesis of many diseases, including those of the cardiovascular system. The normal wound repair response, as seen with a full-thickness dermal laceration for example, involves three major stages: inflammation, granulation tissue formation, and scar formation (12> 1 3 X Immediately after injury, clot formation and inflammation result in the release of cytokines and growth factors that act as chemotactic and mitogenic agents for adjacent fibroblasts. Activated fibroblasts migrate into the wounded area through a provisional E C M composed of the fibrin clot, its associated proteins, and matrix components secreted by the activated fibroblasts themselves ( 1 3X Upon becoming established within the wounded area, activated fibroblasts differentiate into MFBs and are responsible for the formation of granulation tissue through proliferation and E C M deposition. Importantly, E C M deposition occurs early and precedes differentiation, leading to suggestions that the differentiation process may be an adaptive response to modified E C M O3). These cells participate in repairing the wound by 3 remodeling the E C M , a dynamic process requiring synthesis of matrix components as well as production of proteases involved in matrix degradation ( 1 3 \ MFBs are also responsible for wound contracture, a process by which the E C M is actively retracted to limit the wound area and aid in the repair process ( 3\ Under normal conditions, re-epitheialization occurs and MFBs further remodel and degrade the E C M , ultimately disappearing through apoptotic mechanisms as scar formation is completed ( 1 3 \ Pathological wound healing results from continued activation of MFBs and inappropriate E C M accumulation and remodeling (3). Similar mechanisms involving vascular MFBs and SMCs may be operative in cardiovascular tissues in response to a variety of injurious stimuli, a process called the "response to injury". The "response to injury" hypothesis heralded by the late Russell Ross (14) and John Glomset in the early 1970's (l5> 1 6) provides a classic paradigm describing the mechanism of vessel wall remodeling in response to injurious events. It incorporates many stereotypic components of the wound repair process in an attempt to define the initiation and progression of occlusive vascular disease. In this model, initiation of healing follows repeated or chronic endothelial cell injury and is characterized by the upregulation of numerous molecules involved in leukocyte activation, recruitment, adhesion, and transmigration into the vessel wall ( 1 7X Injurious stimuli also lead to direct secretion of inflammatory mediators by endothelial cells. The production of cytokines and growth factors by endothelial cells and infiltrating inflammatory cells activates medial SMCs and induces phenotypic modulation. Activated SMCs migrate along a chemokine gradient into the neointimal space where they are stimulated to proliferate. These events require that SMCs produce and secrete a provisional E C M consisting of numerous components, especially proteoglycans. Deposition of the provisional E C M is regulated by a delicate balance between stimulatory and inhibitory mediators present in the vessel wall microenvironment. Dysregulation of the E C M remodeling process in the neointimal space is a key event in the development, progression and complication of vascular diseases such as 4 atherosclerosis. Although not as well defined, similar mechanisms are likely present during the progression of valvular diseases. 1.1.3 Importance of Versican in Tissue Pathology The dysregulated response to injury and repair requires an orchestra of participants of which versican, among other interstitial proteoglycans such as biglycan and decorin, is an essential player. The accumulation of versican in a variety of diverse diseases, including vascular disease, valvular heart disease and cancer, underscore its critical role in pathobiology and highlight the importance of understanding its modulation and regulation. In this thesis I have taken advantage of two clinically burdensome diseases characterized by dramatic overexpression of proteoglycans to better define and understand the role of versican and its modulation. The accelerated atheromatous process evident in transplant vascular disease (TVD) results in striking and aberrant accumulation of interstitial proteoglycans, particularly versican, in association with subendothelial accumulation of lipoproteins (18> 1 9 X As such, it offers an excellent model in which to study the modulation of versican by mediators such as growth factors and oxidized lipid moieties. The myxomatous nature of cardiac valve lesions associated with anorexigen exposure as well as exposure to other injurious stimuli provides a unique setting in which to study versican expression by MFBs, the prototypic cell type involved in matrix remodeling subsequent to tissue injury. 1.2 Versican Versican is an interstitial proteoglycan belonging to a family of proteoglycans termed hyalectans or lecticans ( 2 0 \ As a family, lecticans exhibit a modular core protein structure composed of highly conserved N - and C-terminal globular domains separated by variable central regions that house the G A G chains (21). The name versican alludes to the substantial versatility imparted both by its structure and its wide tissue distribution ( 2 0X Recent studies have suggested 5 that versican plays a central role in modulating cell adhesion, migration and proliferation, highlighting its importance in normal tissue development and maintenance, as well as in numerous pathologies. 1.2.1 Protein Structure The full-length isoform of the versican core protein is composed of an N-terminal globular domain (Gl) , a central linear region, and a C-terminal globular domain (G3) (22). The N-terminal globular domain is a link-protein-like structure called the hyaluronan-binding region (HABR) and is comprised of an immunoglobulin-like loop and a tandem repeat domain (Figure 1). The C-terminal globular portion houses a lectin-like region containing two epidermal growth factor (EGF)-like elements, a C-type lectin domain and a complement regulatory protein-like region (Figure 1). The central region of versican is composed of two adjacent segments to which chondroitin sulfate (CS) G A G chains attach (Figure 1). These two chains are called G A G - a and GAG-p, respectively (22> 2 3 ) . The core protein of the largest of the versican isoforms has a calculated molecular weight of 370 kiloDaltons (kDa) and an apparent molecular weight of approximately 550 kDa (22> 2 3 X The discrepancy between the calculated and apparent molecular weights may be due to the high number of N - and O-linked oligosaccharides attached to the core protein and/or its the low isoelectric point, either of which may lead to aberrant gel migration properties (20). The molecular weight of the intact proteoglycan is difficult to determine due to the heterogeneity of the attached G A G chains, but is estimated to be in excess of 1.5 x 103 kDa (2°). Indeed, the CS side chains display marked heterogeneity, differing in size and composition depending on the tissue of origin or culture conditions. For example, the CS chains of chick limb bud versican are reported to be 60 kDa (24), while versican from monkey aortic SMCs has CS chains of 40-45 kDa ( 2 5 \ 6 e>c II Vo V l •ocxx , V2 V3 •ocxx Figure 1. A schematic diagram of the versican core protein and its alternatively spliced isoforms. See text for details. (Modified from ref. 20). 7 The structural diversity of versican is derived, to a large degree, from alternative mRNA splicing events that produce at least four splice variants in human (23> 2 6 \ murine (26> 2 7 ) and bovine (28> 2 9 ) tissues, named VO, V I , V2 and V3 (Figure 1). The differential splicing occurs in the central portion of the molecule, producing isoforms with altered GAG-binding potential. VO represents the full-length protein, containing both GAG-cc and GAG-p (23X V I and V2 have G A G - a and GAG-(3 spliced out, respectively, while V3 lacks both GAG-binding regions (23). Differential splicing leads to versican core proteins of variable molecular weight and G A G -binding capacity (Table 1) and leads to potential differences in the functions of these isoforms (30-33) Recent studies have even called into question the validity of classifying V3 as a bona fide proteoglycan, given its inability to support the attachment of G A G chains ( 3 3 \ 1.2.2 Gene Structure The versican gene, designated CSPG2, extends over a 90 kilobase (kb) region of chromosome 5ql2-14 in the human genome (34>35) and chromosome 13 in the mouse genome (36). The single gene has a structural organization that mirrors the modular units of the core protein and is composed of 15 exons, including two large exons encoding for the G A G chain binding domains (Table 2) (35X Alternative splicing events produce versican transcripts of 12 kb, 9 kb, 6.5 kb and 3 kb for VO, V I , V2 and V3, respectively (35> 37X Furthermore, the presence of several polyadenylation signals in the 3-prime untranslated region (Table 2) leads to multiple bands for each isoform upon Northern blotting (35X Versican gene expression is controlled by a gene-specific promoter housing a T A T A box and multiple potential transcription factor binding sites, including AP2, SP-1, CCAAT-binding transcription factor, C C A A T enhancer binding protein, xenobiotic responsive element (XRE)-binding protein and cAMP responsive element-binding (CREB) protein (35). Previous studies have demonstrated activity of the versican promoter in cells of epithelial and mesenchymal origin (35X 8 Table 1 Versican isoforms produced through alternative splicing VO V I V2 V3 G A G Chains G A G - a , GAG-p GAG-p G A G - a none Number of CS Sites 17-23 12-15 5-8 0 Calculated M W * (kDa) 370 262 180 72 Apparent M W * * (kDa) 550 500 400 unknown Abbreviations: G A G - glycosaminoglycan, CS - chondroitin sulfate, M W - molecular weight, kDa - kiloDalton *The MWs reported are those of the versican core protein. **The apparent M W refers to the size of the versican core protein as determined by gel electrophoresis. 9 Table 2 Comparison of gene structure for human and murine versican Human Versican Murine Versican Chromosomal Location 5ql2-14 13 Size of Gene (kb) 90 90 Number of Exons 15 15 Size of GAG-oc/GAG-3 (kb) 3/5.3 2.9/5.2 Number of Poly-A Sites 3 4 10 1.2.3 Tissue Distribution During Embryonic Development Versican is a major component of the E C M in the developing embryo (30> 38> 3 9 X During embryonic development in the chicken, versican appears first in the subectodermal region dorsal-lateral to the neural tube, in association with the basement membrane of the early epithelial somites and the neural tube, in the early perinotochordal mesenchyme and in the posterior halves of the sclerotomes (2°). It is expressed later in all prechondrogenic tissues, being replaced by aggrecan as cartilage formation occurs. Versican is also present in the developing gut (2°) and is transiently expressed in the retina (4°) and optic tectum (41) of the central nervous system. Versican is expressed to a high degree in the heart (42> 4 3 ) and blood vessels (2°), suggesting its importance in the development of the cardiovascular system. In late embryonic development, a wide distribution of versican similar to that seen in adult tissues is observed (20> 4 4 \ 1.2.4 Distribution in Adult Tissue Versican is ubiquitously expressed in adult tissues and is found in the loose connective tissue of most organs, especially those harboring SMCs ( 4 4X Versican expression is apparent in fibrous and elastic cartilage (45-47)? tendon (48> 4 9 \ skeletal muscle (50) and skin (51>52). Limited amounts are present in the central nervous system, the peripheral nervous system and on the lumenal surface of some gladular epithelia (44> 5 1 X Qualitative investigations have demonstrated a close association between versican and elastic fibers, including colocalization with elastin (44> 5 1 ) , fibulins (53-55) a n d fibrillin-1 (56). The close association between versican and elastic tissue has led to the suggestion that it may inhibit interactions between cells and the elastic network, reducing mechanical strain on the cell and modulating elastic fiber orientation. Recent studies have established that V2 is exclusive to the nervous system where it is the predominant component of the mature brain E C M (28> 57, 58) Importantly, versican is present in all blood 11 vessels in the normal vascular system (44> 5 9 \ It localizes to all three wall layers in elastic arteries and veins, but is confined to the adventitia of muscular arteries. Versican present in the vascular wall is derived largely from arterial SMCs (25> 59> 6°), although limited expression is also apparent in endothelial cells (27> 61> 6 2 \ 1.2.5 Ligands for Versican Versican associates with a variety of E C M and cell surface molecules through specific protein-protein and protein-carbohydrate interactions. Ligands for versican can be divided into three separate categories: 1) those molecules that bind the N-terminal domain, 2) those molecules that bind the G A G regions, and 3) those molecules that bind the C-terminal domain (Figure 2). The best-known interaction is that between hyaluronan (HA) and versican. Recombinant full length and N-terminal versican constructs have been used to demonstrate that HA, but not heparin or CS, specifically binds to the H A B R at the N-terminus of versican ( 6 3 \ The interaction between H A and versican is supported by several immunohistochemical and biochemical studies demonstrating either colocalization of the two macromolecules or the presence of versican-HA aggregates ( 6 4- 6 8). A host of ligands, including selectins, matrix components and chemokines are capable of specific interactions with the negatively charged G A G chains present in the central region of versican (Figure 2). The leukocyte adhesion molecules L-selectin and P-selectin have been shown to recognize overlapping binding sites on versican, an interaction that involves carbohydrate-binding domains on the selectins and CS moieties on the versican G A G chains, specifically those side chains containing dermatan sulfate (DS) and C6S ( 6 9 ' 7 0 ) . A specific glycoform of versican has also been reported as an extravascular ligand for L-selectin in the kidney, both in vitro and in situ ( 6 9X The interaction between versican and selectins may relate to 12 13 leukocyte adhesion and trafficking, as versican is shed from kidney tubuli and deposited in adjacent vascular bundles around leukocyte infiltrates after unilateral ureteral obstruction ( 6 9 \ The association of versican with other adhesion molecules present in the E C M may also involve the G A G chains. Substrate mixtures of fibronectin, but not laminin or collagen, with versican isolated from mouse sciatic nerve specifically inhibit adhesion of cultured neurons and Schwann cells, a function mediated by the CS chains on versican and the cell-binding domains of fibronectin (71X Thus, interactions between the G A G chains of versican and specific ligands may represent an important mechanism by which versican modulates cellular activities such as adhesion. In addition to binding molecules related to cell adhesion, the G A G chains attached to versican interact with specific chemokines and either inhibit (72> 7 3 ) or potentiate (72) their activity. Studies in embryonic mice have shown that versican can bind midkine, a heparin-binding chemokine, with high affinity via polysulfated domains of CS G A G chains containing a region of DS structure (72X The interaction between midkine and versican has been proposed to modulate midkine activity by either concentrating the chemokine or competing with its receptor for binding (72X On the other hand, G A G chain-mediated binding of secondary lymphoid tissue chemokine (SLC) to versican inhibits both a4p7 integrin-dependent binding to mucosal addressin cell adhesion molecule-l-IgG and calcium (Ca 2 +) mobilization by chemokine receptor-expressing lymphoid cells (73X These studies demonstrate the ability of proteoglycans such as versican to act as "sinks" for soluble mediators, either increasing their local concentration or effectively removing them from the active environment. Examination of the ligand-binding capacity of the C-terminal portion of versican has revealed a number of protein-carbohydrate and protein-protein interactions (Figure 2). A recombinant construct comprised of the 2 EGF domains, the C-type lectin domain and the CRP-like region has been used to demonstrate versican's capacity to bind D-mannose, D-galactose, L-14 fucose, N-acetyl-D-glucosamine, heparin and heparan sulfate. These binding events are Ca -dependent and require the CRP-like region. The E C M glycoproteins tenascin-R, fibulin-1 and 2_j_ fibulin-2 also bind the versican lectin domain in a specific, high affinity, Ca -dependent manner (54, 55, 74) indeed, fibulin-2 and tenascin-R may share the same binding site on versican ( 5 5 \ Binding of these molecules to versican occurs via protein-protein interactions and, in the case of the fibulins, involves the C-type lectin domain of versican and the EGF-like repeats of fibulin-1 and fibulin-2 (54>55). Recent ultrastructural and biochemical examination of fibrillin-containing microfibrils from human normal placental membranes has revealed that they bind covalently to the C-type lectin domain of versican ( 5 6X Of particular interest, the binding site for versican appears to be the region between cbEGF domains 11-21 on fibrillin-1 which coincides with mutations resulting in severe forms of Marfan's syndrome (56). The C-terminal domain of versican also interacts with cell surface integrins. Cell-binding and immunoprecipitation assays on astrocytoma cell lysates transfected with mini-versican have demonstrated that pi-integrins bind the C-terminal region of versican in a Ca2+-dependent, RGD-independent manner ( 7 5 \ These studies also showed that the integrin-versican interaction activates focal adhesion kinase, increases integrin expression, promotes glial cell adhesion and inhibits apoptosis, suggesting that versican may modulate integrin-mediated intracellular signaling events ( 7 5 \ 1.2.6 Modulation of Versican Expression by Physical Stimuli Versican synthesis by mesenchymal cells can be regulated by physical stimuli, including cell density and mechanical stress. Serum-mediated stimulation of cultured human arterial SMCs increases CS proteoglycan synthesis in a density-dependent manner, with cells at low density producing more CS proteoglycans than cells at confluency ( 7 6X Furthermore, quiescent cells situated in confluent monolayers produce fewer proteoglycans than similarly serum-deprived cells plated at low density (76). Decreasing levels of versican mRNA and protein 15 concomitant with increasing cell density have also been observed in human normal dermal fibroblasts (51). In addition to cell density, the stretching forces related to blood flow may have an impact on versican accumulation in the vessel wall. Recently, studies using homogeneous biaxial strain as a stimulant demonstrated that mechanical deformation increases versican mRNA levels and core protein synthesis by human arterial SMCs ( 7 7 \ Substantive increases in versican expression were noted even when the forces applied remained within a range deemed physiological ( 7 7X These data suggest that the dynamic physical environment present within tissues such as the vessel wall, especially the distribution of cells and the presence of mechanical disruption or distortion, may play a role in modulating versican expression under both physiological and pathological conditions. 1.2.7 Modulation of Versican Expression by Growth Factors and Cytokines The synthesis of versican by mesenchymal cells can be regulated at the levels of transcription, translation, and post-translational modification in response to a host of soluble mediators present within the microenvironment of diseased cardiovascular tissues. These mediators include the pro-atherogenic growth factors transforming growth factor beta (TGF-p) and platelet-derived growth factor (PDGF). Studies using cultured monkey arterial SMCs have demonstrated that TGF-P 1 and PDGF-AB each increase versican mRNA levels, core protein synthesis and G A G chain length (25). Furthermore, treatment of these cells with PDGF-AB alters the G A G chain C6S:C4S ratio, while similar exposure to TGF-p has no effect (25). As with vascular SMCs, cultured human dermal fibroblasts increase their levels of versican mRNA in response to TGF-P ( 7 8 ) , while fibroblasts derived from periodontal ligament tissue demonstrate enhanced versican mRNA levels in response to PDGF-BB ( 7 9X Similarly, human normal gingival fibroblasts respond to treatment with either TGF-P or PDGF-BB by increasing versican mRNA levels (78> 7 9 \ A comparison of human gingival fibroblasts with human granulation tissue 16 fibroblasts isolated from chronically inflamed periodontal tissue revealed enhanced versican mRNA levels by both cell types in response to TGF-pl treatment, although the TGF-pl -mediated increase in versican expression by the myofibroblastic granulation tissue cells was more robust (80). Thus, PDGF and TGF-p represent potent mediators of versican expression in both SMCs and fibroblasts in vitro, and potentially play an important role in the overexpression of versican in pathologies involving E C M remodeling. In contrast to TGF-p and PDGF, only limited work has documented alterations in versican expression by SMCs and fibroblasts in response to treatment with other growth factors. Examination of pericellular CS/DS proteoglycan expression by rat A10 SMCs, as measured by trypsin-releasable 35S-labeled material, showed increased synthesis independent of cell proliferation in response to PDGF-AB and insulin-like growth factor-I (IGF-I), but not in response to epidermal growth factor (EGF) or basic fibroblast growth factor (bFGF) (81). The identities of specific proteoglycans were not elucidated in this study. Examination of proteoglycan expression by human gingival fibroblasts and periodontal ligament fibroblasts demonstrated a similar increase in versican mRNA levels by both cell types in response to IGF-I stimulation ( 7 9 \ while stimulation of either normal dermal fibroblasts or keloid fibroblasts with bFGF failed to affect versican transcript levels ( 8 2X Finally, exposure of cultured rat vascular SMCs expressing native angiotensin type 1 (ATI) receptors to angiotensin II (Ang II) significantly upregulated proteoglycan synthesis and versican mRNA expression ( 8 3X Increased versican expression was also observed in similar cells expressing recombinant angiotensin type 2 (AT2) receptors ( 8 3X These studies, taken together, illustrate an obvious link between versican upregulation and exposure to specific individual growth factors. They also raise the possibility that the presence of multiple growth factors, a situation likely to occur in vivo, may lead to cooperative modulation of versican expression. 17 A few in vitro experiments have shown that simultaneous addition of two or more specific growth factors to cultures of vascular SMCs or fibroblasts alters versican expression to a greater extent than exposure to one growth factor alone. Simultaneous treatment of vascular SMCs with TGF-P and PDGF-AB leads to an additive increase in versican protein expression ( 2 5 \ Studies using human adult lung fibroblasts show that while TGF-P and either PDGF or EGF have no effect on proteoglycan synthesis beyond that seen with the use of TGF-p alone, the triple combination of TGF-P, PDGF-BB and EGF increases proteoglycan synthesis 1.6-fold over that seen with TGF-p only ( 8 4X Furthermore, this combination of growth factors increases versican mRNA synthesis five-fold as compared to control cells (84). In these experiments EGF, PDGF-A A and PDGF-BB, when used alone, had no effect on proteoglycan synthesis, highlighting a potential difference in the regulation of versican expression by fibroblasts originating from different tissues and by fibroblasts in comparison to vascular SMCs. These studies suggest that specific growth factors may cooperate to upregulate versican expression in fibroblasts and SMCs under physiologic and pathologic conditions. In contrast to growth factor-mediated upregulation of versican expression by vascular SMCs and fibroblasts, the pro-inflammatory cytokine interleukin-l-beta (IL-lp) appears to decrease the synthesis of this proteoglycan. Human gingival fibroblasts cultured either in a monolayer or in a three-dimensional matrix experience a time-dependent decrease in versican mRNA levels and core protein expression in response to IL- lp treatment ( 8 5X Further studies have demonstrated that both human gingival fibroblasts and monkey arterial SMCs respond to IL- ip treatment by reducing proteoglycan synthesis when the cells are plated on fibronectin, but not when plated on vitronectin or collagen ( 8 6X Such studies indicate that IL- ip plays a role in the down-regulation of versican expression during specific inflammatory processes and that mediator-induced alterations in proteoglycan synthesis by fibroblasts and vascular SMCs may be 18 dependent on the overall composition of the surrounding E C M . Furthermore, the differential modulation of versican expression by specific growth factors and cytokines suggests that the regulation of injury-induced expression is complex and relies, in part, on the amount and type of soluble mediator present in the microenvironment. Studies conducted in recent years have begun to elucidate the intracellular signal transduction pathways important for growth factor-mediated versican expression by vascular SMCs (83> 8 7 \ In vitro experiments using monkey arterial SMCs have shown that PDGF-stimulated versican core protein expression and G A G chain elongation are regulated by distinct signaling pathways (87). Specifically, versican core protein expression signaling apparently occurs through a receptor tyrosine kinase (RTK)-dependent, protein kinase C (PKC)-independent pathway, while G A G chain elongation signaling occurs through a PKC-dependent, RTK-independent pathway (87). Inhibition studies have also been used to examine the signal transduction pathways involved in the Ang II-mediated increase in versican gene expression by rat vascular SMCs ( 8 3 \ The concentration- and time-dependent increase in mRNA levels by ATI-receptor-expressing cells in response to Ang II was inhibited by the ATI-receptor antagonist losartan, the EGF receptor (EGF-R) inhibitor A G 1478 and the matrix-associated protein kinase (MAPK) inhibitor PD98059, but not by the PKC inhibitors chelerythrine and staurosporine, indicating that Ang II-mediated stimulation of SMC versican expression is regulated by EGF-R-dependent tyrosine kinase pathways ( 8 3X That RTK-dependent pathways are involved in the stimulation of versican core protein expression by both PDGF and Ang II is intriguing and suggests the existence of a common pathway for growth factor-mediated upregulation of versican expression by vascular SMCs. 19 1.3 Functions of Versican in Atherogenesis In previous times, interstitial proteoglycans such as versican were generally regarded as space-filling entities whose pathologic functions included a large water-binding capacity related to tissue edema and a significant contribution to overall lesion mass. Indeed, versican does serve these functions, but more recent studies clearly indicate that this proteoglycan is a dynamic, multifunctional biochemical species capable of influencing many pro-atherogenic events. In particular, versican is fast becoming recognized as a key participant in such functions as cell adhesion, cell migration and proliferation as well as in lipid retention and modification (Figure 2). 1.3.1 Cell Adhesion, Migration and Proliferation It is becoming increasingly clear that versican plays a critical functional role in cell adhesion, migration and proliferation. Immunocytochemical studies of cultured embryonic fibroblasts from chick, murine and human sources have indicated that versican is abundantly distributed in the pericellular space where it colocalizes with hyaluronan and CD44, but is excluded selectively from focal adhesion contacts ( 6 5 \ Furthermore, versican has been found to inhibit cell-substratum adhesion, indicating the presence of anti-adhesive properties (65X Importantly, inhibition of cell adhesion is necessary for cell growth and motility. In support of this concept, studies in keratinocytes and dermal fibroblasts have demonstrated high versican mRNA levels and core protein expression in early, proliferative cultures, but down-regulated levels in differentiated cells ( 5 1 \ The upregulation of versican expression in response to stimuli that are known to affect cell migration and proliferation provides strong evidence for a role for versican in these processes (Figure 2). Human arterial SMCs induced to proliferate by serum increase their synthesis of proteoglycans, including versican ( J 6 \ Many of the growth factors implicated in the 20 upregulation of versican by vascular S M C s and fibroblasts, including P D G F , IGF-I, E G F and A n g II, are also mitogens and/or chemoattractants for these same cell types (88-91) T t j s important to note, however, that TGF-p \ a robust mediator of increased versican expression, is not a mitogenic stimulus for S M C s and, in fact, promotes differentiation ( 2 5 \ Thus, while versican may represent an important mediator of cell growth and motility, it may also serve a function in phenotypic modulation. Particle exclusion and migration assays have been employed to explore the link between the increase in versican expression mediated by P D G F and cell adhesion, proliferation and migration. When treated with P D G F - A B , human arterial S M C s initiate the synthesis of a versican-rich, HA- r i ch pericellular matrix coat ( 9 2). Dynamic formation of this hydrous coat is prevalent around the trailing cytoplasmic processes of the cells prior to cell rounding, is rapid and is specifically coordinated with cell detachment and mitotic events (68> 9 2 \ Thus, it appears that the synthesis of pericellular matrix coats rich in H A and versican in response to proliferative stimuli such as P D G F may facilitate S M C proliferation and migration by decreasing the adhesion of cells and increasing the viscosity of the surrounding E C M (68> 92X Recombinant constructs encoding for the G l and and G3 domains of versican have been utilized to assess the role of this proteoglycan in cell migration, proliferation and differentiation. Mini-versican, a recombinant construct encoding a protein composed of the versican G l and G3 domains separated by 15% of the G A G attachment region ( 9 3), stimulates proliferation when expressed in either fibroblasts or chondrocytes (93> 94X Deletion of the EGF- l ike motifs from the G3 domain of mini-versican inhibits this effect (93> 94X Furthermore, treatment of fibroblasts with antisense oligonucleotides to the E G F - R in the presence of purified mini-versican protein product also inhibits proliferation. These results suggest that versican-mediated proliferation occurs, in part, by binding of the EGF- l ike motifs to endogenous E G F - R (?3) and raises the possibility that versican itself may be able to induce a signal transduction cascade leading to cell 21 growth. Expression of the versican G l domain in these cells also may also aid in cell proliferation by destabilizing cell adhesion, thus facilitating cell rounding during proliferative events (94> 95X An important role for the G3 domain of versican in cell growth is supported by data demonstrating proliferation of astrocytoma cells in response to mini-versican and G3 protein products (96). Interestingly, deletion of the EGF-like motifs has a dominant-negative effect on proliferative activity in this system, inhibiting the secretion of endogenous versican and reducing the binding capacity of the G3 domain for the cell surface (96X In addition to cell growth, the role of these domains in cell migration and differentiation has also been assessed. In astrocytoma cells, increased versican expression has been observed in cultures with high motility in comparison to cell cultures with low motility (97X In these studies, addition of mini-versican protein to astrocytoma cultures enhanced cell migration, an effect mediated by the anti-adhesive properties of the G l domain (97X Mini-versican has also been used in mesenchymal cells from chick limb bud to assess the role of versican in mesenchymal condensation and chondrogenesis, a model of cell differentiation (98X Cultures treated with mini-versican formed fewer, smaller cartilaginous nodules and produced less link protein and collagen, while expression of the G3 domain alone was able to inhibit chondrogenesis, an effect abolished by deletion of the EGF-like motifs (98X Furthermore, treatment of cells with versican antisense oligonucleotides promoted chondrogenesis (98X Thus, the EGF-like motifs of the versican G3 domain may stimulate migration and proliferation, while they inhibit differentiation. 1.3.2 Lipid Retention and Modification There now exists substantial evidence supporting an important functional role for versican in the retention and modification of lipoproteins entering the vessel wall during early atherogenesis. Due in large part to its G A G chains and to its ability to aggregate, versican can directly bind lipoproteins in the subendothelial space, thus providing a structural framework on 22 which lipoproteins may be retained and modified. Indeed, a prevailing hypothesis in the literature, the "response-to-retention", suggests that retention of lipoproteins within the vascular wall by versican is a key initiating step in atherogenesis. "Response-to-Retention" Hypothesis The "response-to-retention" hypothesis, championed by Williams and Tabas ("> 1 0 °), stipulates that retention of atherogenic lipoproteins within the vessel wall represents the major initiating event in atherogenesis. Once trapped beneath the endothelium, these lipoproteins can initiate and potentiate a response cascade culminating in clinical disease. This hypothesis is reliant on four primary factors: the plasma concentration of low-density lipoprotein (LDL), the aberrant retention of L D L in the arterial wall, the modification of retained L D L and the inflammatory response to modified L D L ( 1 0 1). Vessel wall proteoglycans, especially CS proteoglycans synthesized by intimal SMCs, represent key molecules responsible for initially retaining atherogenic lipoproteins and increasing their suseptibility to modification (Figure 2). Interactions Between Proteoglycans and Low-Density Lipoprotein The retention of lipoproteins within the vascular wall involves interactions between lipoproteins that infiltrate the vessel wall and the surrounding E C M (101> 1 0 2 \ Low-density lipoproteins bind several E C M components, including collagen and elastin ( 1 0 3> 1 0 4), but the most important interactions occur between L D L and proteoglycans, especially CS proteoglycans (CSPGs) such as versican ( 1 0 5 X Direct binding of L D L to versican involves ionic interactions between clusters of positively charged amino acids in the C-terminal region of apolipoprotein (apo) B100 and negatively charged G A G chains attached to the versican core protein (106> 1 0 7 \ In fact, the proteoglycan-binding activity and the LDL-receptor (LDL-R)-binding activity conferred upon apoBlOO overlap in the basic C-terminal region of the molecule. The conformation-dependent binding of the L D L - R is distinguishable, however, from the charge-dependent binding 23 of proteoglycans (101> 1 0 8 ) . Site-directed mutagenesis of the human apoBlOO protein and its expression in transgenic mice has revealed the presence of a single C-terminal lysine (Lys) to glutamine (Glu) substitution that inhibits binding to proteoglycans, yet leaves the L D L - R binding activity unaffected ( 1 0 6 \ Subsequent studies comparing these mice with mice expressing wild-type L D L have demonstrated markedly reduced atherosclerosis in those animals defective in proteoglycan-binding activity, providing direct evidence of an important role for L D L -proteoglycan interactions in early atherogenesis ( 1 0 9 \ It is important to note, however, that L D L containing apoB48, a short form of apoB containing only the N-terminal 48 residues of apoBlOO, is still capable of binding vascular proteoglycans such as versican and biglycan ( 1 0 1 X This proteoglycan-binding activity is present despite the absence of the C-terminal amino acid clusters in this short apoB molecule. Moreover, transgenic mice expressing exclusively apoB48 or apoBlOO develop similar degrees of atherosclerosis, suggesting the presence of additional binding sites for proteoglycans in the N-terminal region of apoB. Conformational alteration of the apoB molecule upon loss of the C-terminal region has been proposed as one mechanism for this observation ( 1 0 2). In addition to the direct, charge-dependent interactions between lipoproteins and proteoglycans, there is evidence for indirect binding of these two components through "adaptor" or "bridging" molecules. Recent in vitro studies have demonstrated that lipoprotein lipase (LpL), an enzyme secreted by infiltrating macrophages, facilitates the binding of L D L to versican, biglycan and decorin ( n ° ) . The binding of proteoglycans by LpL is postulated to be the result of ionic interactions, while the binding of LpL to apoBlOO requires the lipid-binding domain of the enzyme ( 1 0 2). Moreover, LpL enhances the binding of oxidized L D L (oxLDL) to versican, biglycan and decorin-coated collagen, partially overriding the reduced binding affinity that occurs as extensive oxidative modification reduces the positive charge on the apoB molecule (no, i l l ) A model for the retention of L D L by proteoglycans that takes into account the direct 24 and indirect interactions between L D L and vessel wall proteoglycans has recently been proposed ( n i ) . The model involves an initial high capacity, low affinity direct interaction between the two components followed by a shift to a low capacity, high affinity indirect interaction as the lesion progresses and infiltrating macrophages secrete LpL ( n i \ Colocalization of Versican and Lipoproteins in the Arterial Wall Based on in vitro data, versican would appear to play the most significant role in retaining lipoproteins within the vessel wall. Binding studies suggest that L D L has a higher affinity for versican than for biglycan or decorin 0 ' 2 ) , a functional property likely due to the relatively large number of G A G chains that can attach to the versican core protein. Versican-HA complexes present as part of the arterial E C M should further increase the amount of L D L retained, as these macromolecular complexes are composed of hundreds of G A G chains ( 1 1 3 \ Surprisingly, analysis of the distribution of proteoglycans in primary atherosclerotic lesions has demonstrated only minimal colocalization of versican and lipoproteins. Versican and versican-H A deposition in intermediate and severe atherosclerotic lesions appear in areas adjacent to cells expressing TGF-p and PDGF as well as around edges of the atheromatous core, indicating that versican is situated close to, but not associated directly with, retained lipoproteins ( 6 7X In contrast, biglycan is highly colocalized with apoB and apoE and is prominent in the SMC-rich E C M adjacent to TGF-p-positive macrophages ( 6 7 \ These observations suggest that, in contrast to its strong binding properties in vitro, the ability of versican to retain L D L in vivo may be somewhat hindered. Although the reasons for the discrepancy between in vitro and in situ observations are not known, it may be that versican is specifically degraded after initially trapping L D L in the vessel wall ( 1 1 3). Alternatively, as yet undescribed characteristics of the versican-HA complex may prevent LDL-binding in vivo, or the timing and location of versican expression may prevent long-term retention of L D L (" 3). On the other hand, the differing 25 distribution of versican in atheromatous diseases apart from primary atherosclerosis suggests that it may play a prominent role in certain disease states or at distinct periods of disease progression. The vascular lesions associated with accelerated atheromatous diseases, including restenosis and TVD, demonstrate excessive accumulation of proteoglycans and lipoproteins. In restenosis, studies have demonstrated strong coincidence of versican and retained lipoproteins (114, 115) Arterial lesions from failed human heart allografts show dramatic overexpression of versican that correlates with the distribution of apoE and apo(a) (18> 1 9). Most recently, a study comparing oxidative stress, lipid retention and proteoglycan accumulation in a porcine model of carotid artery graft interposition demonstrated the colocalization of apoB, versican and oxidized epitopes in the cellular neointima of the saphenous vein grafts ( 1 1 6). These studies indicate the involvement of versican in lipoprotein retention and modification in certain models of atherogenesis. Alterations in Lipoprotein Structure Affect Proteoglycan-LDL Interactions Alterations in the biochemical structure of L D L may either decrease or increase its ability to bind arterial proteoglycans. Initial retention by vascular wall proteoglycans increases the suseptibility of L D L to oxidative modification and enzymatic digestion, alterations that ultimately enhance L D L aggregation in the developing lesion. L D L aggregation has a number of pro-atherogenic consequences, including the generation of bioactive molecules (eg. lysophosphatidylcholine, ceramide), the formation of foam cells, and the promotion of local inflammation, L D L retention and lesion progression ( 1 1 7> 1 1 8). Although mild oxidation does not affect the charge of the apoB component and, in fact, increases L D L affinity for heparan sulfate (HS)- and LpL-coated surfaces ( , 1 9), extensive oxidation reduces the positive charge on apoB, leading to lower binding affinity for extracellular proteoglycans ( 1 2 0 \ As mentioned previously, LpL can partially restore the binding of oxLDL to proteoglycans ( u °) . On the other hand, 26 oxLDL also induces the synthesis of proteoglycans with enhanced LDL-binding properties and specifically increases the synthesis of biglycan ( 1 1 2 \ Similar to oxidative modification, non-enzymatic glycation causes the formation of Lys adducts, reducing the net positive charge on lipoproteins and with it the binding affinity to proteoglycans ( 1 0 2 X Proteolysis of apoB by sphingomyelinase (SMase) in the presence of phospholipase 2 increases the affinity of L D L for proteoglycans, potentially by exposing previously buried GAG-binding sites on the molecule ( ] 2 1). Enzymatic digestion of L D L by SMase also induces lipoprotein aggregation, a process that leads to enhanced binding of L D L to proteoglycans ( 1 2 2). These observations suggest that initial proteoglycan-LDL interactions may induce structural changes to the lipoprotein that alter the binding and retention of L D L by proteoglycans, a process that, based on a critical balance between pro-binding and anti-binding modifications, creates a vicious cycle of enhanced lipoprotein retention. Structural differences related to the overall atherogenicity of various apoB-containing lipoproteins also have an effect on proteoglycan-LDL interactions. It has been known for some time that small, dense L D L have a higher affinity for proteoglycans than larger, more buoyant L D L ( 1 2 3). Although the reasons for this differential affinity are not understood, it is possible that the lower content of surface phospholipids present in small, dense L D L expose more G A G -binding sites or make the molecules more prone to oxidation and aggregation ( 1 1 3). Lipoprotein(a) [Lp(a)], a highly atherogenic LDL-like lipoprotein composed of apoB disulfide-linked to apo(a), binds several E C M components, especially proteoglycans ( 1 2 4 X The high affinity of Lp(a) for proteoglycans may be related to the presence of both apolipoprotein constituents. Lp(a) has been shown to bind SMC-derived E C M with high affinity and the presence of Lp(a) bound to this E C M increases the binding of native L D L to the matrix, along with L D L aggregation ( 1 2 4). Recent studies have demonstrated that Lp(a) binds defensins and 27 that the Lp(a)-defensin complex binds the E C M , suggesting that defensins, like LpL, may act as a bridge to bind Lp(a) (125X Alterations in Proteoglycan Structure Affect Proteoglycan-LDL Interactions The G A G chains attached to proteoglycan core proteins mediate interactions between the vascular E C M and lipoproteins. Thus, modulating the number of G A G chains available for LDL-binding, either by changing the amount or variant of core protein synthesized or by altering the length or the degree of sulfation of G A G chains, leads to changes in the lipoprotein-binding properties of proteoglycans. A number of mediators present in the atherogenic environment have been shown to modulate G A G chain length and sulfation. PDGF and TGF-P increase the length and overall sulfation of GAGs attached to interstitial proteoglycans synthesized by SMCs (25X PDGF also increases the C6S:C4S ratio in the GAGs produced by these cells. A l l of these alterations give rise to proteoglycans with enhanced LDL-binding properties. Serum stimulates increased synthesis of proteoglycans with increased L D L affinity (76> 1 2 6 \ Versican synthesized by monkey arterial SMCs in the presence of TGF-p exhibits enhanced LDL-binding properties attributed to increased G A G chain length (127X O x L D L also stimulates vascular SMCs to synthesize proteoglycans with longer G A G chains and these proteoglycans demonstrate increased binding capacity for L D L (112X Finally, mechanical strain due to blood vessel hemodynamics specifically increases sulfate incorporation into proteoglycans synthesized by vascular SMCs, thereby enhancing that LDL-binding properties of the vessel wall E C M (77). In addition to increasing the length and sulfation of the G A G chains, altered synthesis of core proteins can affect the binding properties of proteoglycans by modulating the relative number of G A G chains available for binding by LDL. In vitro studies have demonstrated that mechanical shear, PDGF and TGF-p stimulate increased synthesis of biglycan or versican core proteins by vascular SMCs (25> 7 7 \ Lipid moieties are also capable of altering the synthesis of 28 specific proteoglycans. Non-esterified fatty acids stimulate the production of versican and decorin by SMCs ( 1 2 8), while oxLDL has been shown to specifically upregulate biglycan synthesis ( 1 I 2). A model could thus be conceptualized whereby local pre-lesion effects, including increased TGF-P synthesis and mechanical shear, lead to increased expression of pro-retentive CSPGs such as versican and biglycan (102> 1 1 3 ) . Such proteoglycans would be increased in number and would contain long, highly sulfated G A G chains that would further retain lipoproteins. As lesion progression continues, other factors (eg. LpL, PDGF, oxLDL) may further stimulate the production of pro-retentive proteoglycans and would also aid directly in proteoglycan-LDL binding. 1.4 Biological Significance of Versican in Cardiovascular Pathology 1.4.1 Transplant Vascular Disease Solid organ transplantation is widely accepted as an effective therapy for many patients suffering from end-stage organ failure. Advances in operative techniques and immunosuppressive therapy have steadily improved one year survival rates for all transplant recipients ( 1 2 9 X Such improved survival has been accompanied by the emergence of TVD, a manifestation of chronic rejection affecting all solid organ allografts ( 1 3°), as a leading cause of death in long-term survivors O 2 9). T V D selectively affects allograft vessels while sparing host vessels, and in cardiac allografts is characterized as a diffuse, typically concentric intimal thickening of both the major epicardial vessels and the intramyocardial vessels from the base of the heart to the apex (131> 1 3 2 ) (Figure 3, a and b). While occlusive disease is angiographically evident in more than 40% of recipients who survive for at least four years post-transplant (133> 1 3 4 ) , intravascular ultrasound (IVUS) is more sensitive and can detect intimal thickening in 60% to 70%o of cardiac transplant recipients as early as one year post-transplant ( 1 3 5 X In comparison to IVUS, pathological studies reveal even more prevalent T V D characterized by histological 29 .9 -5 <D CD .2 to i s cu Cu — CA 30 reflections of immune recognition and early responses of the vessel wall to injury, and not necessarily by the presence of "significant" intimal thickening. As such, just as the ambiguous origins of native atherosclerotic disease can be traced to early childhood, so must the equally complex, molecular origins of T V D be explored in terms of elusive initiating events. The precise pathogenetic mechanisms of T V D remain to be elucidated, but it is generally considered that vascular dysfunction and injury in transplants are related to allogeneic interactions between host immune cells and donor tissues and to ischemic injury (136> 1 3 7 X The lesion-promoting events and processes involved are intimately dependent on the interactions of both endothelial cells and SMCs in the vessel wall (Figure 4). Immune-mediated recognition and injury of the endothelial cell layer provide the stimulus for the elaboration of endothelial cell-derived cytokines and chemokines, and the upregulation of adhesion molecule expression involved in increased leukocyte recruitment and attachment, altered vascular permeability to lipids and cells, stimulation of SMC migration and proliferation, and activation of apoptosis. T cells and macrophages, subsequent to activation and transendothelial migration, further amplify the concentration of soluble mediators in the allograft microenvironment and, together with activated endothelial cells, provide a chemokine gradient leading to SMC migration from the media and subsequent proliferation and E C M synthesis in the neointimal space. The combination of altered endothelial cell permeability and SMC matrix synthesis provides a mechanism for lipoprotein insudation and retention, respectively, in the developing lesion. Such lipid accumulation can lead to oxidative modifications that are potentially pro-atherogenic. OxLDL can contribute to further lipid retention, cell proliferation, foam cell formation and cell death. Thus, the activation, injury or destruction of endothelial cells and SMCs leads to the expansion of a lipid-, SMC- and matrix-rich intimal lesion, loss of normal vascular function and the compromise of allograft viability and performance. 31 32 Importance of Lipids in T V D From as early as 1856 when Virchow O 3 8) proposed the concept that vessel injury may include lipid imbibition in the initiation and early stages of atherogenesis, lipids have been recognized as playing an important role in vascular diseases. To date, direct pathological evidence for the importance of lipids in the development of human T V D remains sparse. The associations between both serum triglycerides and cholesterol and the extent of coronary intimal thickening at autopsy ( 1 3 9) stand as a unique data set on the issue. The predictive value of post-transplant obesity and serum cholesterol for the development of intimal lesions has been supported by clinical data that suggest that high serum cholesterol at six months post-transplant is a harbinger of angiographically detected vasculopathy ( 1 4°). Biochemical analysis of cholesterol, free cholesterol, and esterified cholesterol, as well as phospholipid content and concentration, revealed consistent elevation in T V D ( 1 4 1). Not only were these lipid values greater than for donor age comparable normal coronaries, but also they approached the lipid content and concentration of plaques derived from abdominal aorta intimae of autopsied New Orleans and Guatemalan men ( 1 4 2 X The entire arterial wall sample (including intima, media, and stripped adventitia) of coronary vessels from cardiac transplant recipients contained amounts of lipids similar to the amounts found in the aortic intimae alone of 35-44 year old New Orleans men ( 1 4 3). Although there was nearly three times the amount of total cholesterol in the intimal atherosclerotic plaque of 45-54 year-old New Orleans men as compared to the entirety of transplant vessel walls, the amount of intimal lipid in the Guatemalan men was modest as compared to that in the coronary vessel walls of allografts. Thus, coronary vascular lipid accumulation in transplant patients, which develops in a period of months to a few years in chronologically young hearts, approximates lipid accumulation in the aortic intimae of older 33 patients after decades of atherosclerotic insult. As well, lipid accumulation proceeds in coronary vessel walls of transplanted hearts despite a comparatively modest blood-to-vessel wall concentration gradient of cholesterol and other lipids. The corroboration of biochemical data by histochemical, immunohistochemical, and electron micrographic techniques ( 1 4 1), establishing frequent and typically diffuse intra- and extracellular accumulation of lipids in both the intimal and medial layers of cardiac allograft arteries, has affirmed that the alloimmune environment may strongly promote lipid imbibition. As previously mentioned, the "response-to-retention" hypothesis of Williams and Tabas ( 1 0°) recognizes the central pathological role of subendothelial retention of atherogenic lipoproteins. Of considerable interest, Oil Red O positivity was noted in lumen-lining endothelial cells in vessels with early and late T V D lesions, as well as in smooth muscle cells and phagocytes ( 1 4 1 X Thus, lipoprotein aggregates appear to cause endothelial cell lipid overload, which may contribute to dysfunction, injury or death of endothelial cells in TVD. In other words, endothelial cells become foam cells in their own right early in TVD. Apolipoproteins E, B, and (a) have been demonstrated immunohistochemically in transplant lesions in patterns of distribution distinct from those seen in native atherosclerotic lesions ( ] 9 \ Specifically, transplant lesions display prominent intimal apo(a) and apoE immunoreactivity, whereas apoB is a more prominent feature in native disease. The presence of both apoE and apo(a) in early disease (when intimal thickening is minimal), as well as in severe disease, suggests that lipid accumulation is an event that occurs early in T V D and persists throughout the course of disease. These apolipoprotein deposits have a high degree of spatial concordance with versican and biglycan, both of which are overexpressed in T V D (18) (Figure 3). This apolipoprotein-proteoglycan colocalization suggests an interaction that ultimately enhances lipid retention and foam cell formation in these lesions. A precedent for these phenomena is established in native atherosclerotic disease (144-147) 34 The use of immunosuppressive reagents to treat rejection in transplant recipients adds a special dimension to an already complex intra-graft milieu. Indeed, these drugs may contribute to the development of TVD, although the extent to which they affect lesion formation in humans is unresolved. Cyclosporine's lipophilic nature, its capacity to bind plasma lipid fractions, and its ability to impair V L D L and L D L clearance ( 1 4 8), make it a prime candidate for promoting the role of lipids in TVD. Indeed, cyclosporine has been associated with the development of transplant-associated hyperlipidemia (149> 1 5°) and accelerated allograft arteriosclerosis (141> 1 5 1 ) . More recent data in the setting of kidney transplantation have demonstrated that cyclosporine treatment increases the susceptibility of L D L to in vitro oxidation and enhances in vivo oxidation O 5 2). Kidney transplant recipients weaned off cyclosporine and switched to azathioprine immunosuppression exhibited a decrease in cholesterol and triglyceride levels ( 1 5 3 ' 1 5 4 X Furthermore, Lp(a) levels in patients on cyclosporine alone were found to be increased over patients treated with a combination of azathioprine and prednisone ( 1 5 5), while patients treated with FK506 had decreased LDL-cholesterol, Lp(a), and fibrinogen levels as compared to cyclosporine-treated patients ( 1 5 6 \ Using an in vivo de-endothelialized rabbit carotid artery model, Ferns et al ( 1 5 7) demonstrated that cyclosporine therapy was associated with intimal smooth muscle vacuolation, increased total intimal thickening (due to presence of macrophage-derived foam cells) and H -thymidine uptake by neo-intimal monocytes/macrophages. On the other hand, in vitro work ( 1 5 7) demonstrated that cyclosporine decreased the rate of proliferation of rabbit aortic SMCs and endothelial cells. Despite evidence implicating cyclosporine in vessel wall injury ( 1 5 1> 1 5 7), recent work has demonstrated potentially beneficial effects of cyclosporine on the progression of TVD. For example, Miller and colleagues ( l 5 8) have noted significantly less myointimal thickening by IVUS in cardiac transplant patients receiving more than 5 mg/kg/day as compared to less than 3 mg/kg/day cyclosporine (p=0.03). These data suggest that the amount of bioavailable 35 cyclosporine, especially as assessed by dose and possibly by blood level, alters the progression of intimal thickening. The possibility that injurious effects of cyclosporine on coronary vessels are offset by enhanced immunosuppression can not be excluded. A recent clinical study with coronary angioscopy found that, in addition to donor age, lower mean blood levels of cyclosporine were a significant clinical predictor of "nonpigmented" intimal thickening O 5 9 ) . This form of intimal thickening is typical of the diffuse and concentric TVD lesion seen by IVUS or pathologically in TVD. Thus, this study suggests that augmented cyclosporine immunosuppression may be beneficial for specific lesion types. 1.4.2 Anorexigen-Associated Valve Disease The pathology of heart valve lesions associated with exposure to anorexigens, particularly fenfluramine/phentermine (fen/phen) and dexfenfluramine (dexfen), exhibits great variability, ranging from near normal in appearance to markedly abnormal ( 1 6°). The limited published data on the pathologic features of such valves suggest findings similar to those seen in ergotamine-induced or carcinoid heart disease ( n ) . Gross examination of affected valves reveals irregular thickening of leaflets and the presence of opaque, glistening white nodules or "plaques" that distort the valve surface and surround the junction of the leaflet and fused chordae ( 1 6 1 X In addition to fusion of the chordae, thickening and shortening of these structures may be apparent. Commissural fusion, calcification and vegetative growths are not evident ( 1 6 1 X The histopathology of the valves reflects these macroscopic findings (Figure 5). Microscopic features of anorexigen-associated valve lesions include the presence of patchy, superficial, fibromyxoid proliferative foci composed of moderate numbers of myofibroblast-like cells embedded in abundant loose, GAG-rich E C M and some fibrous material (11,161) (Figure 6, a and b). Such plaque-like lesions are present along the valve surface and may also encase the chordae. Indeed, myxomatous "onlay" formation on the valves of patients 36 0 X ** o T 3 X> CD 1 l 0) X ! H 2 o o u 0) & > s E x * 5 S c 2 cfi JO + J ° > co u u «8 _ «s rt S T3 « « o <U m <1> £ fj cJ » t , > M ^ U O co "7! U l O a cuo s S £ CO B B S ccj * o j i j ^ < "2 i ^ -5 o o « -c 5 CJ I U i co *a » O XI 0 g O ccS so 1 ? 1 fi ccj _ '3 W -6 co ^ T3 s i § 1 « II cu ' 3 cS . O ^ g<B OX) o — ea S3 0 o ~ a T 3 oo CO co cu 13 a co j3 8 b ccS 38 treated with fen/phen or dexfen is interpreted as a prominent histopathologic finding ( u ). The elastica remains largely intact and the GAG-rich E C M appears focally prominent at points of chordal insertions. The initial report on the pathology of anorexigen-associated valve disease found the integrity of the valve proper to be unaltered ( n ) , but more recent studies, including my own, suggest that mild to marked myxoid degenerative changes are also apparent in the valve proper, separate from superficial lesions (Chapter 6) (161> 1 6 2 X Based on quantitative image analysis of Movat's pentachrome-stained tissue sections, the composition of the valve proper is GAG-rich (162). In addition to the presence of a GAG-rich connective tissue matrix, numerous thin-walled neovessels are present in anorexigen-exposed valves, distributed within the valve proper and valve onlays ( 1 6 1> 1 6 2) (Figure 6c, Figure 7b-c). Variable numbers of leukocytes are also present in both the valve proper and onlays, and are often associated with the neovessels ( 1 6 1) (Figure 6d, Figure 7b-c). Such a range of features no doubt reflects the .duration of exposure, the drug dosages, and individual patient susceptibility factors, including local standards and patterns of care and the influence of pre-existent valve disease. 1.4.3 Rheumatic Heart Valve Disease In chronic rheumatic valve disease, inflammation and repair phenomena represent major mechanisms by which valvular tissue is pathologically remodeled. This disease is related to precedent bouts of acute rheumatic fever caused by Group A beta-hemolytic streptococcus and is likely initiated by the immune response to an antigen shared by the streptococcal M-protein and valvular connective tissue (8> 163> 1 6 4 X Chronic rheumatic valve disease can occur in any of the four cardiac valves affected by rheumatic fever, but predilection for mitral valve stenosis and regurgitation exists. The morphological features are similar regardless of the valve affected and include commissural fusion, tissue fibrosis and valve thickening. Microscopic changes in these 39 \* 1 • M l M o T l E -S -40 thickened, abnormal valves include fibrotic destruction of leaflet or cusp architecture, increased collagen deposition and the occurrence of dystrophic calcification (165> 1 6 6 ) (Figure 7d). A chronic inflammatory infiltrate composed primarily of lymphocytes and macrophages is evident and thick-walled neovessels permeate the valve core ( l 6 3> 1 6 4). Phenotypic characterization of the lymphocytic component of the inflammatory infiltrate has revealed the presence of T-helper and T-suppressor cells as well as B-cells, leading to suggestions that the valvular injury may be mediated, to some degree, by delayed-type hypersensitivity mechanisms. Chronic rheumatic valve disease represents an extremely valuable model of immune-mediated injury and repair. 1.4.4 Floppy Mitra l Valve Disease Floppy mitral valves are the most common cause of isolated mitral regurgitation in the third and fourth decades of life, principally due to valve prolapse into the atrium ( 1 6 7 X The propensity to mitral valve floppiness may be present at birth and remain clinically silent until adulthood, with young women being most commonly affected by clinically silent, echocardiographic prolapse. Synonymous with myxomatous, mucinous, hooded and ballooning mitral valves, floppy valves are characterized by interchordal hooding, valve leaflet expansion and elongation, and chordal elongation (167> 1 6 8 \ Disintegration of the valve fibrosa, including collagen bundle fragmentation and degradation, thickening of the spongiosa and increased G A G deposition (myxomatous transformation) are histologic hallmarks of floppy valves (167> l69> 1 7°) (Figure 7e). Surface fibrosis and small thrombi over the atrial aspect of the leaflets may also be evident, the result of friction and leaflet contact during prolapse. Of particular importance, certain echocardiographic and histopathologic features of floppy valves are indistinguishable from those valves affected by chronic rheumatic disease, although mononuclear cell infiltrates and vascularization are typically not features of floppy mitral valves. Moreover, the 41 pathogenetic mechanisms resulting in the myxomatous changes that manifest in floppy mitral valve disease are unknown. 1.4.5 Carcinoid Heart Valve Disease Carcinoid heart disease is a frequent finding in patients who develop carcinoid syndrome, a pathological condition resulting from a metastatic carcinoid tumor, the primary site of which is typically located in the small bowel ( 1 7 1 \ These tumors metastasize to the liver where they are thought to secrete several soluble mediators, including serotonin, into the hepatic vein, ultimately resulting in distinctive lesions on right-sided heart valves ( 1 7 1 > 1 7 2 X The lesions, present on the downstream surface of the valve leaflets, are described as glistening, often large, superficial plaque-like thickenings composed of a GAG-rich connective tissue matrix devoid of elastic fibers on the surface of valvular leaflets or cusps (172-174) (Figure 7 f ) . Inflammation and neovascularization may at times be very prominent at the valve-onlay interface. Carcinoid valve disease is unique in its predominant involvement of both the tricuspid and pulmonic valves and typically results in tricuspid regurgitation and pulmonic stenosis ( 1 7 2). 1.5 Dysregulation of Vascular Smooth Muscle Cells in Cardiovascular Disease Vascular SMCs, together with endothelial cells, are the major resident cellular contributors to the initiation and progression of atherogenesis. The aberrant accumulation of SMCs in the developing neointima and the dysregulated production of E C M components by these cells are central events in several models of vascular disease. In particular, the development and expansion of the neointima in T V D involves accumulation of SMCs and the overexpression of versican. 42 1.5.1 Intimal SMC Accumulation in TVD Progressive accumulation of vascular SMCs in the neointima is a central feature of T V D (131, 175, 176) As discussed above, chronic immunological activation / injury of vascular endothelium and the ensuing inflammatory response leads to the synthesis of a variety of growth factors and cytokines that mediate neointimal hyperplasia. Expansion of the neointima requires activation and phenotypic modulation of medial SMCs, migration of the cells in response to chemotactic stimuli, and proliferation of the cells in the intimal space (177-179) (Figure 8). Cytokines and growth factors secreted by injured endothelial cells and infiltrating macrophages and T-cells, particularly TGF-P 1, PDGF and IL-lp, are believed to be key mediators in the response to vessel wall injury (177> 178> 1 8 °). Histopathologic analyses of both human and animal models of T V D demonstrating the time-dependent accumulation of SMCs in the intima and a parallel loss of medial SMCs strongly suggest that phenomena similar to those in native atherosclerosis play a vital role in the progression of vascular disease in transplanted hearts. Phenotypic modulation and subsequent migration of medial SMCs across the internal elastic lamina is a complex process requiring cooperative changes in cell-cell and cell-matrix interactions, as well as secretion of appropriate degradative enzymes (178> 1 7 9 l Evidence suggests that fibronectin derived from SMCs is important both for the initiation of phenotypic modulation and for matrix-dependent migration in native atherogenesis ( 1 7 9 X Similarly, important roles for this matrix glycoprotein have been suggested by studies in animal models of TVD. In a piglet model of cardiac TVD, intimal thickening has been linked to fibronectin-dependent SMC migration ( 1 8 1 l Increased expression of fibronectin in the subendothelium and the inner media of transplanted vessels in this model appears to be mediated by local production of IL- lp and TNF-a ( i 8 2 , 183)^  including endothelial cell-derived IL-ip ( 1 8 1 \ Interestingly, vascular elastase, a degradative enzyme important for SMC migration, may play a role in mediating IL-iP-induced 43 p I § > o 3 » g o S XJ o CS d .2 to S o .S I * oo S C3 fl 5 S'S .B o 3 G xi a fa «3 l i s u i 1 1 T3 "3 "5 § s S C D T3 (L) o OB .s "O _3 ej .£ SO o OH X I CO s a —• u CO B X CD U s a C U c o T3 C CO I ,43 o Ul & O •a £ S x — — rV) I S 0 2 1 & s 'g 1 "S S I 8 • A 5 o <D O >> .S £ ° ° § £ s a a a s J H 2 2 o xi o op-p <2I I u CO u T3 C 2 3 © O 00 oo % » I S 3 | o o co <U C3 CU .2f 1 I o i i 5 "O 5 <5 •S £ o tu <u e l CU ,3 * J CO •a o 8 & d .M o i > co CU •3 d p o 44 fibronectin synthesis ( 1 8 4). Taken together these results suggest that soluble components in the tissue microenvironment, including inflammatory cytokines and proteases, work synergistically to promote fibronectin synthesis by medial SMCs. Such actions, together with the establishment of a chemotactic gradient composed of soluble mediators such as PDGF, are important in mediating SMC motility in TVD. The role of growth factors, particularly PDGF and its receptors, has been studied extensively in culture and native atherogenesis (177> 1 7 8 X PDGF plays a dual role in SMC accumulation during atherogenesis, serving both as a powerful chemoattractant and as a potent mitogen ( I 7 9 \ Studies in the allograft setting are suggestive of an important role for PDGF in SMC migration and proliferation upon paracrine stimulation by macrophages and endothelial cells and/or upon autocrine stimulation of SMCs (185-188) increased protein levels for PDGF ligands and receptors have been demonstrated in rat cardiac allografts ( 1 8 6) and aortic allografts (185)? w hi le elevated mRNA and protein levels for the ligands have been demonstrated in human renal allografts ( 1 8 7). Infiltrating macrophages appear to be the predominant source of PDGF-B B , and both PDGF receptors are expressed by SMCs in the intima and media (185-187) other growth factors, including IGF-I and EGF, may also help mediate SMC migration and proliferation in allografts. Both EGF and IGF-I stimulate SMC proliferation in vitro and have been localized in allograft vessels (189> 1 9 °). The importance of growth factor-mediated SMC accumulation in vascular lesion development in allografts is further suggested by lesion reduction upon inhibition of growth factor synthesis or activity. Rat aortic allografts treated with the somatostatin analogue angiopeptin have decreased levels of EGF, IGF-I and PDGF and exhibit a significant decrease in intimal and medial SMC replication along with a reduction in intimal thickening ( 1 9 1). Together these studies indicate that several growth factors, particularly PDGF, are involved in the intimal accumulation of SMCs in TVD. 45 The above discussion reinforces the essential role of neointimal SMC accumulation in the progression of chronic lesions in transplanted organs. Intimal hyperplasia in T V D appears to be driven by complex interactions involving matrix proteins, growth factors, and degradative enzymes. An intragraft milieu enriched with such mediators may also initiate apoptosis. Such an environment may contribute to the observed reduction of medial SMCs and may have important implications for cells reaching the intimal space. Regardless of the potential for cell damage and death in the growing intimal layer, those SMCs accumulating in the lesion produce generous quantities of E C M , the composition of which is paramount to the ultimate nature of the disease. 1.5.2 Extracellular Matrix Production SMCs accumulating in the neointimal space of injured vessels constitute a principal source of basement membrane (eg. collagen IV, laminin, HS proteoglycans) and interstitial (eg. collagen I, collagen III, fibronectin, CS proteoglycans) E C M production and elaboration C1). The extent and composition of E C M produced by intimal cells has potentially profound effects on lesion progression and vessel function. Excessive deposition of E C M in the growing neointima is a prominent feature in both animal and human models of T V D (18>192). Histochemical analysis of coronary arteries from human heart allografts has demonstrated dramatic accumulation of GAG-rich matrix in the early and advanced intimal lesions of T V D ( 1 9 2 X Immunohistochemical examination of these lesions for specific SMC-derived interstitial proteoglycans has demonstrated a dramatic excess of versican and biglycan in the intima, especially prominent adjacent to the media, of vessels with T V D (18). On the other hand, decorin expression is more prominent in advanced T V D and native atherosclerotic lesions (18). Corroborative data demonstrates aberrant versican and biglycan expression in loose, lipid-rich regions of native atherosclerotic lesions (67) and excessive proteoglycan deposition in restenotic arteries (193> 1 9 4 X 46 A number of in vitro studies suggest that growth factors and cytokines present in the atherogenic microenvironment are important in modulating proteoglycan expression by SMCs. A series of studies using monkey arterial SMCs has shown that both PDGF and TGF-P 1 stimulate increases in versican mRNA and core protein levels (25), while TGF-pl but not PDGF stimulates biglycan mRNA synthesis ( 1 9 5 X Decorin core protein levels are not altered by either growth factor, suggesting that the interstitial proteoglycans are differentially regulated. On the contrary, decorin production is upregulated by IL- ip , while biglycan expression is unaffected O 9 6). Both PDGF- and TGF-P 1-mediated regulation of versican are associated with an increase in the length of the G A G chains (25), an effect that has also been demonstrated for biglycan and decorin O 9 5). Moreover, PDGF is able to induce changes in the composition of the G A G chains in all three proteoglycans, specifically doubling the C6S to C4S ratio (25> 1 9 5 X These results suggest that the transcriptional and post-translational regulatory mechanisms for proteoglycans are complex and that core protein synthesis and glycosaminoglycan synthesis may be differentially regulated. While the modulation of proteoglycan synthesis by SMCs appears to be controlled, at least in part, by growth factors known to participate in SMC migration and proliferation, the relationship between these events and proteoglycan production is not clearly understood. Studies have demonstrated an increase in proteoglycan synthesis concurrent to SMC proliferation (197> 1 9 8 ) , and human SMCs stimulated to proliferate increase their mRNA levels for biglycan, versican, and perlecan (76). On the other hand, as compared to proliferating cells, quiescent rat vascular SMCs have upregulated expression of biglycan O 9 9). These behaviors appear to be dependent on E C M proteins since similarly enhanced levels of biglycan mRNA are detected when cells are plated on collagen I or plastic, while either laminin or fibronectin cause a significant decrease in mRNA expression ( 1 9 9). Recent studies using a rat SMC line suggest that ' 47 PDGF-AB stimulation increases pericellular proteoglycans independent of its proliferative effect (81). As well, IGF-I stimulation increases such proteoglycans without a parallel increase in cell number, and bFGF or EGF stimulation mediates proliferation without affecting proteoglycan expression (81). It seems reasonable that proteoglycan production may be partly, but not completely coupled to growth factor- and cytokine-mediated proliferation. It is evident from the above discussion that specific, excessive accumulation of SMC-derived matrix proteoglycans in the developing intima is essential to the pathogenesis of TVD. Growth factors and cytokines derived from both endothelial cells and infiltrating lymphocytes and macrophages appear to be of paramount importance in this process, but the role of other major components present in the allogeneic microenvironment, including lipoproteins, immunosuppressive drugs, and infiltrating cells themselves, remains to be established. Importantly, the very presence of these potential mediators in the vessel wall milieu is likely the direct result of endothelial injury and altered permeability. 1.6 Role of Myofibroblasts in Cardiovascular Disease MFBs exhibit a broad tissue distribution, participating in the growth and development of normal tissue, the normal wound-healing response and a host of pathologies broadly catagorized as the response to injury and repair (inflammation and remodeling), quasineoplastic proliferative conditions and stromal responses to neoplasia ( 2 0 0 X Modulation of progenitor cells toward the M F B phenotype is best represented as a continuum of phenotypic changes and occurs as a result of cellular activation during physiologic or pathologic stress. The current dogma suggests that in adult tissues MFBs may arise from any of several progenitor cell types, the most predominant of which are fibroblasts and SMCs (201-203) indeed, MFBs may be defined as mesenchymal cells that express phenotypic properties intermediate between those of fibroblasts and SMCs (202> 2 0 4 " 2 0 6 \ In some organs, specialized mesenchymal cells may represent alternative sources of MFBs 48 when appropriately stimulated. Examples include perisinusoidal stellate cells, glomerular mesangial cells (207> 2 0 8 ) , pericytes ( 2 0 3), pulmonary septal fibroblastic cells and valvular interstitial cells (209X 1.6.1 Definition and Phenotypic Characterization of MFBs The most precise definition of MFBs is based on the characterization of its ultrastructure as compared to that of fibroblasts and SMCs (200-202, 204, 210) Such characterization highlights the simultaneous contractile and synthetic properties of this multifunctional cell. In general, MFBs are stellate in shape, are surrounded by a partial basal lamina, and are electrically interconnected by adherens and gap junctions (200> 204> 2 1 1 \ Important intracellular features include a well-developed rough endoplasmic reticulum (RER), Golgi apparatus and pinocytotic vesicles (201> 2 0 2X Stress fibers with interdispersed dense bodies run parallel to the long axis of the cell. The nucleus is deeply invaginated and contains both nuclear bodies and prominent nucleoli (201X The cells are attached to the E C M by fibronexi, specialized transmembrane complexes that link the intracellular contractile apparatus to fibronectin through cell surface integrins (200-202) jt j s important to note, however, that the histologic features of these cells parallel the ultrastructural morphology. In fact, routine histology, together with immunohistochemical data, can be used to identify MFBs in the absence of ultrastructural data (200, 202) _ The immunohistochemical characterization of MFBs is based on the expression of cytoskeletal proteins comprising the microfilament and intermediate filament systems present in eukaryotic cells (201> 2 1 °) . Of these markers, the expression of alpha-smooth muscle actin (oc-SM actin) is most important in differentiating a M F B from its progenitors (201X As with all myoid cells, MFBs express beta (P)-actin and gamma (y)-actin, while they are positive for a -SM actin and negative for alpha (oc)-sarcomeric actin and a-skeletal actin ( 2 1 2> 2 1 3). A classification system 49 based on staining for vimentin (V), a -SM actin (A), desmin (D) and S M myosin (M) has recently been proposed (200> 2 0 2 \ Using this system, MFBs may be described as being of V- , V D -, V A - , V M - , V A D - , or VA(D)M-type, although the expression of these markers varies with species, tissue, disease state and culture conditions. 1.6.2 Mechanism of Phenotypic Modulation in MFB Activation Phenotypic modulation of progenitor cells to MFBs involves complex, co-operative interactions among the cells, soluble mediators such as TGF-p arid GM-CSF and E C M components, especially fibronectin ( 2 1 4 \ These interactions culminate in alterations to the cytoskeleton and fibronexi. GM-CSF is capable of inducing oc-SM actin expression in rat fibroblasts after subcutaneous injection in vivo, but is unable to do so in vitro, suggesting that the inflammatory cell-derived mediator is an indirect inducer of phenotypic modulation (215> 2 1 6X On the other hand, TGF-P, the "master switch" of tissue repair, is a direct mediator of a-SM actin expression ( 2 1 3 . 214> 2 1 7 \ In fact, the three known TGF-p isoforms, TGF-p l , TGF-p2 and TGF-P3, directly induce the expression of a-SM actin by MFBs in vitro (218X The current mechanistic paradigm for phenotypic modulation is based on experimental evidence from rat models of subcutaneous wound healing and bleomycin-induced pulmonary fibrosis (2,9X Initial insult induces an inflammatory response that leads to secretion of GM-CSF by infiltrating inflammatory cells. The production of GM-CSF stimulates the expression and secretion of latent TGF-P by macrophages ( 2 2°). The latent TGF-P complex is activated by molecules such as thrombospondin-1 and stimulates adjacent fibroblasts to upregulate ED-A-containing fibronectin expression (221X ED-A fibronectin-derived outside-in signaling is permissive for TGF-p-mediated upregulation of a -SM actin expression and development of the M F B phenotype (221> 2 2 2X This permissive effect is necessary, but not sufficient, for modulation toward the M F B phenotype as seeding fibroblasts on ED-A FN does not stimulate a-SM actin expression, but 50 selective blocking of ED-A with specific mAbs inhibits a-SM actin expression in response to TGF-(3 (221> 2 2 2 ) . In the bleomycin-induced lung fibrosis model, TGF-p production is followed by differentiation of alveolar fibroblasts into a-SM actin-positive MFBs, leading to increased collagen production and eventual fibrosis ( 2 2 3 X 1.6.3 Role of MFBs in Wound Repair and Inflammation Activated MFBs are dynamic participants in both wound healing and inflammatory processes (210> 214> 2 2 4 X They carry out an impressive array of reparative functions, including migration, proliferation, synthesis of and response to soluble mediators, contraction, contracture and matrix remodeling (203> 2 1 3> 2 1 4). Many of these functions are common to both processes. An early event in wound repair is the migration and proliferation of PDGF-receptor-positive MFBs within a PDGF-rich tissue microenvironment ( 2 2 5). TNF-a and bFGF, produced by inflammatory cells and macrophages, respectively, are mitogens and chemoattactants for MFBs (226-228)) a s j s connective tissue growth factor (CTGF), a TGF-p-induced immediate early gene product released by granulation tissue fibroblasts ( 2 2 9). It is important to note, however, that none of these mediators are capable of inducing a-SM actin expression, suggesting that their activity is dependent on TGF-P-mediated phenotypic modulation. Many of these growth factors are also important for matrix remodeling, a key function of MFBs in normal and diseased tissues (80, 202, 209, 213) j n addition to synthesizing and responding to an array of pro-inflammatory growth factors, cytokines and chemokines, activated MFBs also express a host of cell adhesion molecules, providing attachment sites for leukocytes during inflammation and repair ( 2 1°). MFBs also contract underlying tissue and limit the exposed surface area of the wound, a function requiring calcium and intercellular communication via gap junctions (204> 2 3 0 X Critical to both wound healing and inflammation is the process of matrix remodeling. Activated MFBs are important contributors to the synthesis and secretion of matrix components. 51 MFBs can synthesize and secrete a range of matrix components, including collagen, fibronectin and proteoglycans. Collagen and fibronectin expression are upregulated both during and after phenotypic modulation and activation of MFBs (221 , 231, 232) MFBs also represent a particularly important source of proteoglycan synthesis. A comparison of human normal gingival fibroblasts and human granulation tissue fibroblasts isolated from chronically inflamed periodontal tissues has demonstrated enhanced basal levels of versican mRNA expression by the myofibroblastic cells (80). Moreover, when treated with TGF-p, the granulation tissue cells increased versican mRNA levels to a greater extent than similarly treated normal fibroblasts (80). It is also important to note that PDGF, described above as a mediator of M F B proliferation during the response to injury, is also a potent mediator of proteoglycan expression in fibroblasts (78> 79> 233) and vascular SMCs (25, 87) it i s likely, therefore, that PDGF has a regulatory effect on proteoglycan expression by MFBs. 1.6.4 Cardiac Valvular Interstitial Cells are Tissue-Specific MFBs Cardiac valves harbour a population of tissue-specific MFBs called valvular interstitial cells (VICs) or valvular MFBs. Present throughout the valve stroma and particularly within the spongiosa, valvular MFBs are the most prevalent cell type in the valve and exhibit morphological characteristics intermediate between fibroblasts and SMCs ( 2 0 9> 234, 235) Ultrastructural, histochemical and immunohistochemical descriptions of valvular MFBs in situ have revealed that they are tentacular, spindle-shaped cells possessing long cytoplasmic extensions, prominent intermediate and gap junctions, a partial basal lamina, fibronexi, stress fibers and differentially prominent RER and Golgi complexes (234-237) They express cytoskeletal markers including a -SM actin, vimentin and myosin, but not markers of endothelial cells such as Factor VHI-related antigen and CD31 (2 2 4>234,235,237,238) Cultured valvular MFBs exist as two cell phenotypes. One phenotype is composed of elongated, spindle-shaped cells that 52 display dense cytoplasmic bundles of microfilaments stipled with dense bodies and is strongly immunoreactive for a -SM actin (234> 2 3 7 X The other cell phenotype is cuboidal or "cobblestone" in appearance, contains a well-developed Golgi apparatus and prominent RER, but few microfilaments or stress fibers, and is weakly immunoreactive for a -SM actin (234> 2 3 8 X The inverse relationship between the predominance of subcellular synthetic structures and cytoskeletal components may suggest a specific structure-function relationship related to the secretory or contractile capacity of the cell. The phenotypic characteristics exhibited by valvular MFBs are of great importance in attempting to understand their role in the response to valve injury. Cardiac valvular MFBs are multifunctional, playing pivotal roles in processes as diverse as migration, proliferation, contractility, cell-cell communication and regulated matrix secretion (209, 224, 235, 239, 240) Wounding assays using heart valve organ cultures have demonstrated the activation and migration of valvular MFBs during tissue repair (241> 2 4 2 \ Conditioned medium from valvular endothelial cells has been shown to be mitogenic for MFBs, providing evidence that endothelial cells produce soluble factors that stimulate valvular M F B growth ( 2 3 9 X Stimulants such as PDGF and bFGF have also been shown to induce valvular M F B proliferation (234, 239)^  supporting a role for growth factor-mediated M F B mitogenesis in response to cardiac valve injury. Such investigations allude to the potential contribution of MFBs to injury, inflammation and repair processes, including growth, migration and E C M synthesis. 53 2 C H A P T E R II H Y P O T H E S E S A N D E X P E R I M E N T A L A I M S My overarching hypothesis is that the dysregulation of proteoglycan expression by mesenchymal cells plays an important role in the pathogenesis of accelerated vascular disease and valvular heart diease. Two pathologic conditions in which the aberrant accumulation of mesenchymal cell-derived proteoglycans are apparent, transplant vascular disease and myxomatous valve disease, were used to examine three specific hypotheses relating to the overarching framework. Hypothesis one was that the overexpression of versican by vascular SMCs is due to modulation of its expression by specific components present within the tissue microenvironment. My specific aim was: 1. To define, in vitro, the role of specific growth factors in the modulation of versican expression by vascular SMCs. Hypothesis two was that versican is spatially associated with modified lipoproteins in human atheromatous disease. This association reflects an interaction that contributes to atherogenesis. My specific aim was: 1. To define the oxidative status of lipoproteins accumulating in the allograft vessel wall and determine their association with versican. Hypothesis three was that the exposure of cardiac valves to certain anorexigens (appetite suppressants) produces a proteoglycan-rich lesion morphologically distinct from lesions characterizing other valve diseases and related to the dysregulation of versican expression by valvular MFBs. My specific aims were: 54 1. To define the distinguishing histopathologic features of valves taken from individuals exposed to anorexigens in the context of normal valve histology and the histopathology of other diseased valves. 2. To establish a model system of cultured human cardiac valvular MFBs to further explore the regulation of versican. 55 3 CHAPTER HI MATERIALS AND METHODS 3.1 Selection of Patient Material A l l human tissues examined in this thesis were obtained from the Cardiovascular Registry, St. Paul's Hospital - University of British Columbia. Tissue samples provided by institutions other than St. Paul's Hospital were initially received by the Cardiovascular Registry and archived prior to being used for study. As such, all tissues were handled according to current ethical guidelines set out by St. Paul's Hospital and were obtained using appropriate informed consent. Protocols for the selection of specific tissues are outlined in detail below. 3.1.1 Human Coronary Arteries Ten failed human cardiac allografts derived from 8 male and 2 female patients at autopsy or explantation, 4 to 1610 days post-transplant, were used to characterize the oxidative status of accumulated lipoproteins and the association between oxidized lipoproteins and versican. The patients and the protocol for handling of tissue have been described previously (18> 19> 1 4 1 > 2 4 3 l Left anterior descending (LAD) coronary artery segments were used. Coronary arteries from 10 patients with atherosclerotic disease were obtained from the Cardiovascular Registry, St. Paul's Hospital-University of British Columbia and were used as donor-age comparable, site-matched native atherosclerotic controls. 3.1.2 Human Cardiac Valves Normal aortic and mitral valves and rheumatic, floppy, and certain carcinoid valves were identified in the Cardiovascular Registry, St. Paul's Hospital - University of British Columbia in a fashion previously described ( 2 4 4 X Carcinoid valves and those valves from patients exposed to anorexigens were obtained from the Cardiovascular Registry, St. Paul's Hospital - University of 56 British Columbia, Clarkson Hospital, Northwestern Memorial Hospital, Lehigh Valley Hospital, the Mayo Clinic, Brigham and Women's Hospital, Stanford University, and the Cleveland Clinic. A total of 270 patients ranging in age from 15 to 84 years and a total of 472 valves were included in the study (Table 3). Each valve was assessed twice grossly (when a gross valve was available) and microscopically by a cardiovascular pathologist blinded to clinical parameters and initial clinical impressions to establish the diagnosis of floppy or rheumatic disease and to exclude endocarditic valves. The number of leaflets and cusps from each study group to be compared in the analysis was not delineated a priori, since many of the fen/phen-associated and carcinoid valves received in the laboratory had already been embedded in paraffin. Thus, unlike for the normal, rheumatic and floppy valves, the number of tissue pieces, their orientation upon sectioning, and the architectural completeness of sectioned leaflets or cusps of fen/phen-associated and carcinoid valves could not be controlled. 3.2 Tissue Processing Human tissues were embedded and sectioned by technical staff in the Clinical Histology Laboratory at St. Paul's Hospital. The specific protocols for handling of the human coronary arteries and human cardiac valves are outlined below. 3.2.1 Human Coronary Arteries The tissue handling procedures for the human cardiac allografts have been described elsewhere (l41> 2 4 3 X Briefly, L A D coronary arteries were harvested from failed allografts, perfusion-fixed with 10% neutral-buffered formalin solution, sectioned transversely at 5 to 10 mm intervals and embedded in paraffin. Three micrometer (um) serial sections were cut and placed on Superfrost slides (Fisher Canada, Ottawa, ON) for immunohistochemical staining. Formalin-fixed, paraffin-embedded atherosclerotic coronary arteries were matched for percent 57 1 2 o c "3 u « <n oo • m c o CN -H oo 00 i n c o Os + 1 O 00 r-o i r i T-H t> © + 1 + 1 CN CN CO + 1 + 1 r - H CN °) SO CN i n T-H 00 CN as oo + 1 i n T-H SO CN OA "H CU o , B T 3 <U O a <n 00 CN oo © -H CO so CO T-H CN CD > < o a o. o © Os r-© CN 00 -H m CN m CN oo CN Os + 1 CN CO SO SO + 1 i n so CN + 1 CN oo oo © i-H + 1 oo CN SO i n CO + 1 CO as CN CN + 1 i n CN S oo i r-© T-H -H r-© i n c o CO © oo + 1 r-CN r-so + 1 00 m T-H +1 © oo + 1 as as i n + 1 Os SO CN i n + 1 CN SO CN s © o . *•* IO e o 'I I ICS CO 00 I i n as *—' -H CN CN <n T-H T—H^  vo CN CO + 1 CO CN as + 1 r -oo T-H SO i n CN +1 Os ,2 13 a C+H o T H '1 CO h CU OO I OH CU , 60 \< <U 60 < 13 a •a 13 a, X cu t>0 CU 13 S <u H H i n T-H CN + 1 00 i n SO U 13 B (U P H a O -*-> J H bp '53 X bO T ^ CU CO +1 so CN «n 00 +1 © i n CN - H 13 a <u H H ~So PQ 58 luminal narrowing and were sectioned at 3 um onto the slides such that allograft vessels and atherosclerotic vessels with similar degrees of luminal narrowing appeared on the same slide. 3.2.2 Human Cardiac Valves The normal, rheumatic and floppy valves identified in the Cardiovascular Registry, St. Paul's Hospital - University of British Columbia were processed using a protocol standardized in our laboratory ( 2 4 4 X Specifically, the valves were fixed in 10% neutral buffered formalin and a 2 mm wide section was cut from the central area of each of the aortic cusps and each of the mitral leaflets, perpendicular to the base of attachment of the cusps or leaflets (Figure 9). The anterior mitral valve leaflets were sectioned from the interchordal space at the free margin to the base of attachment, while the posterior mitral valve leaflets were taken in a similar orientation through the center of the leaflet component closest to the posteromedial commissure. Special attention was given to establishing the locale of each mitral leaflet's base of attachment and care was taken to ensure that the valve tissue was not stretched excessively. The base of attachment for the posterior mitral valve leaflet was readily defined as the point of intersection between the leaflet connective tissue and the cardiac muscle at the atrioventricular junction. The base of attachment of the anterior mitral valve leaflet was defined as the point of intersection of the leaflet with the left cusp of the aortic valve. The section from the anterior mitral leaflet was taken in order to obtain a section wherein the tip (free margin) was easily distinguishable and was free from convergent chordal attachments. The aortic cusps were cut from their free margins to include the noduli Arantii through to their attachment to the aortic wall. This central location was sectioned since it is the point of contact for the valve cusps and the locus of central orifice valvular regurgitation. 59 60 As noted, all other valves were received already embedded in paraffin blocks. Three \im serial sections were cut and placed on Superfrost slides (Fisher Canada, Ottawa, ON) for histochemical and immunohistochemical staining. Serial sections of aortic, mitral, pulmonic and tricuspid valves were stained with hematoxylin and eosin (H & E) or modified Movat's pentachrome stain using a Tissue Tek automated staining system (Sakura Finetek, Torrance, CA). Immunohistochemical staining of tissues was performed as described below. 3.3 Cell Culture Prior to performing experiments, cell culture systems were established using material from animal and human tissues. Specifically, vascular SMCs were isolated from the aortae of Fisher rats and valve MFBs were isolated from human cardiac valves. Specific protocols are outlined below. 3.3.1 Rat Aortic Smooth Muscle Cells A rat aortic SMC culture was established by a modification of the enzymatic dispersion technique (245> 2 4 6 X Four adult male Fisher rats (275-350g) were euthanized by CO2 asphyxiation in accordance with ethical guidelines set out by the University of British Columbia (UBC) animal care committee. The thoracic aorta was removed from each rat and immediately washed in Molecular and Cellular Developmental Biology (MCDB) 131 medium (Sigma-Aldrich, Oakville, ON) supplemented with 1.18 grams/liter (g/L) sodium bicarbonate, 100 international units/ml (i.u./ml) penicillin, 100 micrograms/milliliter (tig/ml) streptomycin and 20% volume/volume (v/v) newborn calf serum (CS; Canadian Life Technologies, Burlington, ON), and was then cleaned of excess adventitial fat and connective tissue under a dissecting microscope. Each aorta was opened longitudinally between the origin of the intercostal arteries and pinned back on a sterile silicone-coated glass plate to expose the endothelial surface. Endothelial cells were removed by gently scraping the luminal surface with a scalpel blade and 61 the aorta was rinsed in serum-free MCDB-131 medium. Enzyme I (0.5 mg/ml collagenase II; Worthington Biochemical Corp, Freehold, NJ) was applied to the exposed media and the tissue was incubated for 20 minutes at 37°C to loosen the media from the underlying adventitia. Medial strips were removed with sterile forceps, taking care not to reach the adventitial layer, transferred to a 35 mm tissue culture dish containing 500 ul of Enzyme II (0.5 mg/ml collagenase II, 0.2 mg/ml elastase; Worthington Biochemical Corp, Freehold, NJ), and minced. Medial tissue from the aortas of all rats were pooled, additional Enzyme II solution was added to the dish, and the tissue was incubated at 37°C for 2 to 3 hours with pipetting at regular intervals to disperse cells. Liberated cells were subsequently pelleted at 1000 rpm for 5 minutes, resuspended in . l ml MCDB-131 containing 20% CS, counted using a hemacytometer, assessed for viability using a trypan blue exclusion assay, and seeded into a 35 mm tissue culture dish at a density of 1.5 x 104 cells/cm2. Cells were grown to confluence, released by trypsinization and subcultured at a density of l.Ox 104 cells/cm2 in MCDB-131 supplemented with 5% CS. 3.3.2 Human Cardiac Valve Myofibroblasts Cardiac valve MFBs were isolated from human tissue using a similar technique to that described for rat aortic SMCs. Tissue sources included those mitral valves procured by the Cardiovascular Registry - St. Paul's Hospital upon resection during valve repair/replacement surgery and from explanted hearts obtained during transplantation. Tissues were placed in cold, sterile medium for transport from the OR and cells were isolated as soon as possible after retieval of the tissue. Each valve was weighed and photographed prior to culture. Half of each valve was fixed in 10%> neutral-buffered formalin, paraffin-embedded and histochemically stained for microscopic analysis. The remainder of each valve was used for culture. Myofibroblasts were isolated by collagenase digestion, grown to confluence, subcultured to expand cell numbers, 62 aliquoted into cryovials, and frozen in liquid nitrogen for long-term storage. Cells were thawed and expanded prior to performing experiments. 3.3.3 Treatment of Vascular SMCs with Growth Factors Experiments were performed with rat aortic SMCs from passage 4 to passage 10. Cells were seeded at a standard density of 1.0 x 104 cells/cm2. For experiments requiring RT-PCR, SMCs were seeded in 60 mm tissue culture dishes. Eight-well collagen-I-coated glass culture slides (Becton Dickinson Canada Inc, Mississauga, ON) were used for immunocytochemical assays. Cells were grown to 70-90% confluence in MCDB-131 medium containing 5% CS. Cell monolayers were washed and starved for 48 hours in serum-free MCDB-131 medium supplemented with 0.2%> (w/v) tissue culture grade, immunoglobulin-free bovine serum albumin (BSA; Sigma-Aldrich, Oakville, ON). Quiescent cells were provided with fresh serum-free medium containing 0.2%> BSA with or without human recombinant (hr) IGF-I (Roche Diagnostics, Laval, PQ), EGF (Roche Diagnostics, Laval, PQ) or EGF + IGF-I. The specific concentrations and time points evaluated for each growth factor are outlined in the Results section. Cells stimulated with 20 ng/ml hrPDGF-BB (Roche Diagnostics, Laval, PQ) were used as a positive control for versican expression. Samples for RT-PCR analysis were harvested by trypsinization (0.1 %> trypsin/EDTA, 5 minutes, 37°C), pelleted by centrifugation (14 000 rpm, 4°C, 1 minute), and stored frozen at -80°C prior to RNA extraction. Cells cultured in multi-well slides were washed gently in Dulbecco's phosphate-buffered saline (DPBS), fixed for 10 minutes in a freshly prepared solution of 4% paraformaldehyde, permeabilized for 2 minutes in 0.1 %> triton-X-100 and air-dried prior to immunocytochemistry. 3.4 Immunoassays The immunoassays employed in this thesis include immunocytochemistry and immunohistochemistry. The establishment and optimization of a manual method for 63 immunocytochemical staining of cells grown in 8-well chamber slides was required prior to performing experiments. In all instances, assay conditions and antibody concentrations were optimized using appropriate positive and negative controls before performing experiments. Detailed protocols for each assay are outlined below. 3.4.1 Antibodies The antibodies used for immunocytochemical and immunohistochemical assays, together with their assay conditions, are summarized in Table 4. Subcultured rat aortic SMCs were characterized and tested for homogeneity using well defined markers of vascular SMCs, including oc-SM actin (Sigma-Aldrich, Oakville, ON) and vimentin (Sigma-Aldrich, Oakville, ON) (Figure. 10). Actin was used at a dilution of 1:100 and vimentin was used at a dilution of 1:500. Cells were also assessed for versican expression under proliferating conditions to verify their ability to produce this proteoglycan (Figure 10). A rabbit polyclonal antiserum against human versican (LF-99) was a gift from Dr. Larry Fisher (National Institute of Dental Research, Bethesda, MD). The antiserum was raised against a synthetic peptide and is monospecific ( 2 4 7 \ A dilution of 1:500 was used for immunocytochemical assays. Mouse monoclonal antibodies against malondialdehyde (MDA)-modified murine L D L (MDA2) and 4-hydroxynonenal (HNE)-modified murine L D L (NA59) were generous gifts from Dr. Wulf Palinski. These antibodies react against oxidation-specific epitopes and are specific for MDA-lysine adducts and HNE-lysine adducts, respectively (248> 2 4 9 X Both antibodies have been used to demonstrate immunohistochemically the presence of oxLDL in arterial tissues and heart valves (248> 250> 2 5 A mouse monoclonal antibody was generated by immunizing mice with human L D L modified with the products of linoleic acid oxidation (252> 2 5 3 ) . This antibody is specific for proteins modified by linoleic acid oxidation products generated through thermal autooxidation (252> 2 5 3 X Specificities and sources of antibodies for apoB, apoE and apo(a) can be 64 Table 4 Antibodies used for immunocytochemistry and immunohistochemistry Antibody Target Dilution (IHC/ICC) Antigen Retrieval Source NA59 H N E - L D L 1:400 none W. Palinski M D A 2 M D A - L D L 1:800 none W. Palinski LOP23 L O P - L D L 1:800 citrate pH 6, 95°C, 1 hr U . Steinbrecher PGM1 macrophages 1:100 none D A K O a-SM actin SMC 1:200/1:100 none DAKO/Sigma LF-99 versican 1:1000/1:500 chond A B C , 37°C, 1 hr L. Fisher vimentin SMCs 1:500 none Sigma factor VIII-related antigen endothelial cells 1:4000 protease I, 37 UC, 8 min D A K O L C A leukocytes 1:25 none D A K O 65 66 found in previous studies published by our laboratory (I8> 1 9 ). Mouse monoclonal antibodies against CD68, a -SM actin and leukocyte common antigen (LCA) were purchased from D A K O (DAKO Corporation, Carpinteria, CA) and were used as markers for macrophages, SMCs and leukocytes, respectively. Immunohistochemical staining was performed on a subset of human cardiac valves for white blood cells (Leukocyte Common Antigen [DAKO, Mississaugua, ON]: a monoclonal antibody that recognizes CD45RB and CD45 epitopes) and vessels (Factor VHI-related antigen [DAKO, Mississaugua, ON]: a polyclonal antibody that recognizes endothelial cells and megakaryocytes). Antibody staining was performed on a Ventana ES automated immunostainer (Ventana, Tucson, AZ) as previously described (18> 1 9) and the conditions were optimized with respect to antigen retrieval and antibody concentrations. In brief, after deparaffinization in xylene, the slides were rehydrated in graded ethanols (100%-70%), washed with PBS, and, i f necessary, subjected to antigen retrieval. Staining for Factor VHI-related antigen was performed on slides pretreated with Ventana's protease 1 for eight minutes at 37°C. Staining for leukocyte common antigen was accomplished without antigen retrieval. Factor VHI-related antigen was used at a dilution of 1:4000, while leukocyte common antigen was used at a dilution of 1:25. 3.4.2 Immunocytochemistry Cell monolayers were subjected to immunocytochemical analysis. The plastic wells on the slides were left intact up to and including application of the streptavidin-biotin/alkaline phosphatase (ABC-AP) complex. Immunocytochemical reagents were applied at a volume of 100 ul/well and all washes were done in excess buffer with constant stirring. Cells were rehydrated in TBST (0.10 M Tris base, 0.15 M NaCl, 0.10% v/v Tween 20, pH 7.6) and treated with 0.25 U/ml chondroitinase A B C (Sigma-Aldrich, Oakville, ON) in 0.10 M Tris-HCl/0.050 M calcium acetate, pH 8.0 for 1 hour at 37°C to remove G A G chains. The monolayers were 67 washed with TBST (3x5 minutes), serum blocking buffer (10% v/v normal goat serum or normal horse serum in TBST) was applied and slides were incubated for 30 minutes at room temperature (RT) to inhibit nonspecific interactions. Excess buffer was then removed and primary antibodies were applied after appropriate dilution in serum blocking buffer. The buffer used to dilute the versican antibody included 0.3 M NaCl to inhibit background associated with charge-charge interactions. Cell monolayers were incubated for 16-18 hours at 4°C with constant shaking and then washed extensively in TBST (3x10 minutes). Secondary antibodies (Vector Laboratories, Burlingame, CA) diluted 1:200 in serum blocking buffer were applied, the slides were incubated for 1 hour at RT with constant shaking and washed with TBST (3x5 minutes). The slides were then incubated A B C - A P complex (Vector Laboratories, Burlingame, CA) diluted in 0.10 M Tris buffer, pH 7.6 for 30 minutes at RT and washed with TBST (3x5 minutes). After washing, the plastic wells were removed and the slides were further washed 2x5 minutes in 0.10 M Tris buffer, pH 8.4. Remaining reagents were applied at a volume of 1 ml/slide. Positive immunoreactivity was detected through the generation of a red reaction product upon application of Vector Red (Vector Laboratories, Burlingame, CA) to the cell monolayers for 20 minutes at RT. Cells were washed with Tris buffer, pH 8.4 (3x5 minutes), rinsed in milli-Q water and nuclei were counterstained with Hoescht 33342 fluorescent dye (Molecular Probes, Eugene, OR) diluted 1:1000 in water and applied for 10 minutes at RT. Slides were washed extensively with water and coverslips were mounted using the Slow Fade Light antifade kit (Molecular Probes, Eugene, OR) according to the manufacturer's specifications. Omission of primary antibodies and replacement of primary antibodies with species-specific normal IgG served as negative controls. Negative controls and internal positive controls were included in each run to evaluate consistency of staining from run to run. 68 3.4.3 Immunohistochemistry Immunohistochemistry was performed as previously described (18> 1 9 X Tissue sections were deparaffmized in xylene, rehydrated in a graded series of ethanol and equilibrated in distilled, deionized water. Antigen retrieval procedures were performed on tissues stained with the LOP23 and LF-99 antibodies. Heat-mediated antigen retrieval was used on sections probed with LOP23 by incubating the tissues for 60 minutes at 95°C in 0.04 M sodium citrate, pH 6.0. Antigen retrieval for LF-99 involved treating sections with 0.25 U/ml chondroitinase A B C in 0.10 M Tris-HCl/50 mM calcium acetate, pH 8.0 for 1 hour at 37°C to digest the G A G chains from the core protein. Tissues were washed 3x5 minutes in TBST wash buffer. Serum blocking buffer (10% v/v normal horse serum or normal goat serum in TBST, pH 7.6) was applied to each tissue and sections were incubated for 30 minutes at room temperature. Excess buffer was then removed and primary antibodies appropriately diluted in serum blocking buffer were applied to the tissues. Tissue sections were incubated for 16-18 hours at 4°C and subsequently washed 3x10 minutes with TBST. The sections were incubated with biotinylated secondary antibody diluted 1:200 in serum blocking buffer for 1 hour at room temperature and then washed 3x5 minutes in TBST. The slides were then incubated with A B C - A P complex diluted in 0.10 M Tris buffer, pH 7.6 for 30 minutes at room temperature and washed 2x5 minutes in TBST wash buffer followed by 2x5 minutes in 0.10 M Tris buffer, pH 8.4. Positive immunoreactivity was detected through the generation of a red reaction product following application of Vector Red chromagen (Vector Laboratories, CA) to the tissue sections for 20 minutes at room temperature. Sections were counterstained with Gill 's hematoxylin. Omission of primary antibodies and replacement of primary antibodies with species-specific, isotype-matched normal IgG served as negative controls. Tissue sections to be labeled with a particular antibody were processed in a single immunohistochemical run to eliminate potential inter-run variability in the intensity of staining 69 and to facilitate comparison of each antibody across all tissues. Negative controls and internal positive controls were included in each run to evaluate consistency of staining from run to run. Automated staining for SMCs, leukocytes and apolipoproteins was accomplished using the Ventana immunostainer (Ventana Medical Systems, Inc., Tucson, AZ) as previously described (18> 1 9 X Briefly, primary antibodies were appropriately diluted in commercial antibody-diluting buffer (Dimension Laboratories, Inc., Mississauga, ON) and incubated on the tissues for 32 minutes at 42°C. A universal biotinylated secondary antibody was then applied for eight minutes at 42°C. A Ventana DAB Detection Kit (Ventana Medical Systems, Inc., Tucson, AZ) was used to detect positive immunoreactivity. An irrelevant mouse antibody was used as a negative control to replace the primary antibodies. 3.5 Reverse Transcription - Polymerase Chain Reaction (RT-PCR) Versican mRNA levels in rat aortic SMC cultures were assessed by RT-PCR. The assay was established and verified using appropriate positive and negative controls prior to its use in experiments. The specific primers and protocols used for each step of the assay are outlined below. 3.5.1 Primers The primer sequences used for amplification of versican cDNA and the sequences of the resulting PCR products amplified from normal rat kidney have been published previously ( 2 5 4). These primers were also used to amplify versican cDNA derived from rat aortic SMCs in the studies presented in this thesis. The sequence of the forward primer used was 5'tttcctgattggcattaatgaaga3' and the sequence of the reverse primer used was 5'tcatcaaaatcaagttcacactgg3' ( 2 5 4 X The primers, originally selected on the basis of murine sequence information, amplify a 179 bp fragment corresponding to the region from 6222 to 6401 bp of murine versican ( 2 5 4). The homology between the rat sequence and the murine sequence is 70 reported to be 98% at the amino acid level ( 2 5 4 X The PCR product generated upon amplification of versican cDNA derived from rat aortic SMCs was isolated and sequenced to confirm the identity of the versican sequence. 3.5.2 Total R N A Extraction Total RNA was extracted using a RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Briefly, cell pellets were homogenized in guanidine isothiocynanate (GITC)-containing lysis buffer and centrifuged through QIAshedders (Qiagen Canada, Mississauga, ON) to shear genomic D N A and reduce lysate viscosity. The resulting solutions were diluted in 70% ethanol and centrifuged through RNeasy spin columns to adsorb total RNA. The membranes were washed to remove contaminants and total RNA was eluted in RNase-free water. Eluates were held on ice and total R N A concentrations were determined spectrophotometrically. 3.5.3 R T - P C R Five ug of total RNA extracted from each cell pellet was used for reverse transcription. Initial removal of contaminating D N A from the total RNA preparation was accomplished by DNase I digestion (Roche Diagnostics, Laval, PQ). After heat inactivation (99°C, 5 minutes), the R N A was reverse transcribed for 45 minutes at 37°C using the M - M L V reverse transcriptase enzyme (Canadian Life Technologies, Burlington, ON) and random hexamer primers (Canadian Life Technologies, Burlington, ON) in a total reaction volume of 60 ul. A reaction containing no reverse transcriptase (-RT) was used as a negative control. Five ul of the reaction mixture was subsequently subjected to 32 cycles of PCR amplification. The 100 ul PCR reaction mixture contained 9.5 ul 10X PCR buffer (Qiagen Canada, Mississauga, ON), 1 ul 25 mM M g C l 2 (Qiagen Canada, Mississauga, ON), 0.5 ul Hot-Start Taq D N A polymerase (Qiagen Canada, 71 Mississauga, ON), 1 ul each versican primer, 5ul cDNA mixture and 82 ul DEPC water. QuantumRNA 18S ribosomal R N A (rRNA) alternate standards (Ambion, Inc, Austin, TX) were used as the housekeeping control for normalization of data. Four ul of a 1:9 primer/competimer mixture was used in each reaction (water volume was reduced to 80 ul for these reactions). Reactions containing the 18S rRNA primers were run separately from those reactions containing the versican primers as preliminary experiments involving duplex reactions resulted in competitive side reactions. Samples were denatured and Taq polymerase was activated at 95°C for 15 minutes. A l l tubes were then subjected to 32 cycles of amplification (95°C, 50°C, and 72°C for 1 minute each), and finally provided with an additional 7 minutes at 72°C to ensure complete extension. Controls for the PCR reaction included a reaction lacking template and a reaction containing the - R T sample. The resultant PCR products were separated in 2% (w/v) agarose gels containing ethidium bromide and visualized using the Eagle Eye gel analysis system (Stratagene, La Jolla, CA). 3.6 Digital Imaging Digital images were captured using a SPOT2 cooled single chip color digital camera (Diagnostic Instruments, Inc) attached to a Nikon TE300 inverted microscope (Nikon Canada, Toronto, ON) equipped with epifluorescence optics. 3.6.1 Cell Layers For the purposes of quantitation, a brightfield image of the immunocytochemical stain and a fluorescent image of the cell nuclei from the identical field of view were captured under non-saturating conditions using a 4x lens. A l l images for a given experiment were captured during a single session using identical parameters. For brightfield images, white-balancing was applied to each slide independently using a well containing no cells. Brightfield digital photomicrographs were acquired as 36 bit red/green/blue (RGB) images (12 bits/pixel/color 72 channel), while fluorescent images were acquired as monochrome images at 12 bits/pixel. Images were imported into ImagePro Plus (Media Cybernetics, Silver Spring, MD) for analysis. Demonstrative digital images were captured in a similar fashion, but made use of the fluorescent properties of the Vector Red chromagen. Thus, images to be used for illustrative purposes were captured as 12 bit monochrome images. Qualitative monochrome images depicting immunofluorescent staining and cell nuclei were pseudocolored and overlayed for presentation. A l l images to be compared were captured using identical parameters, including exposure time and display ranges. 3.6.2 Tissue Sections Fluorescent digital photomicrographs were acquired as monochrome images at 12 bits/pixel and pseudocolored for the purpose of presentation. A l l images to be compared were captured using identical parameters, including exposure times and display ranges. The display ranges used have been listed in the appropriate figure legends. 3.6.3 Agarose Gels Digital images of agarose gels were used to quantitate the relative intensity of bands derived from ethidium bromide-stained PCR products. Digital images were acquired under non-saturating conditions and densitometric values were determined for each PCR product using the Eagle Eye gel analysis software package (Stratagene, La Jolla, CA). The ratio of versican to 18S rRNA was determined for each sample and the mean value of triplicate samples was then calculated. 3.7 Quantitation Digital imaging technology was employed in order to obtain quantifiable data from cells and tissues stained using immunocytochemical, histochemical and immunohistochemical 73 techniques. The protocols for each type of quantitative data derived from the digital images are detailed below. 3.7.1 Immunocytochemical Staining of Cultured Vascular SMCs Quantitation of immunocytochemistry was performed with ImagePro Plus image analysis software (Media Cybernetics, Silver Spring, MD). A color segmentation file based on the hue/saturation/intensity (HSI) color model was developed to identify the positive signal generated by the Vector Red chromagen (Figure 11). Images of cell monolayers with and without exposure to PDGF-BB and stained for versican were used as positive control templates for development of the color segmentation file. The file was composed of 2 separate ranges. Range 1 identified intense staining and Range 2 identified less intense, diffuse staining. A final range, Range 3, identified all remaining pixels in the image area. Total staining within a given image was defined as Range 1 + Range 2. The color segmentation file was applied to each image independently and the amount of staining present was recorded in urn2. The image analysis software was also used to determine the number of nuclei present in the identical field used for quantitation of staining (Figure 11). Assuming an average of 1 nucleus/cell, this value was used to derive the number of cells present in a given field. 3.7.2 Geometric Analysis of Cardiac Valves Normal, rheumatic and floppy heart valves embedded according to our laboratory protocol ( 2 4 4) were used for the geometric analysis. Again, for many valves received already embedded, particularly the majority of carcinoid and anorexigen-exposed valves, optimal embedding was not possible and valves for which a complete, full-length section could not be obtained were necessarily excluded from the geometric analysis. Thus, 310 valves from 211 patients were included in the geometric component of the study. The total number of valves studied geometrically included 186 normal valves from 105 normal patients, 62 rheumatic valves 74 T J CO CD 5 > £ -c o o ~" ctj c3 75 from 53 patients, 36 floppy mitral valves from 36 patients, 15 valves from 7 patients exposed to anorexigens, and 11 valves from 10 patients with carcinoid heart valve disease. Each Movat's pentachrome-stained valve assessed geometrically was submitted to systematic digital image analysis using ImagePro Plus software ( 2 4 4 X A digital image of each Movat's pentachrome-stained section of valve leaflet or cusp was obtained by scanning the glass slide directly using a Nikon 2x2 slide scanner (Nikon Canada, Toronto, ON) outfitted with a custom holder. A l l images were acquired as 24 bit RGB color images (8 bits/pixel/channel) at a resolution of 2700 pixels per inch. A l l measurement and composition data were acquired using the original images without the aid of digital processing techniques such as filtering and smoothing. Standardized measurements of valve area and valve thickness were made on each leaflet or cusp in a consistent manner using the ImagePro Plus image analysis software package (Media Cybernetics, Silver Springs, MD) (Figure 12). A l l measurements were made at a magnification of lx after calibration of the image analysis software using a micrometer image scanned at a magnification identical to that used for the valve tissue. Measurements were made on the valve proper, defined as all tissue internal to the elastic membrane that circumscribes the entirety of each valve, and the onlay, defined as all tissue superficial to the elastic membrane. A l l measurements were performed on each leaflet or cusp individually. The measurements were averaged for each patient when more than one leaflet or cusp was available. Measurements included valve area (total, proper, onlay), onlay characteristics (number, percent onlay area, average onlay size), and total valve and valve proper thickness (average thickness and nine measurements along the valve leaflet or cusp). Specific parameters for each of the measurements are outlined below. 1) Area. For the purposes of this study, area is a 2-dimensional area reflecting the length and thickness of the sectioned valve leaflet or cusp. The areas of longitudinal cross-sections occupied by the total valve leaflet and the valve onlays were determined by 76 W — CD CD 65 Q I £ ( J 2 p C4-1 V l_ U -p & & .Sf § G t s» ! 77 tracing around the total valve, and by subsequently tracing around each onlay (Figure 12a). The valve proper area was determined by subtracting the total onlay area from the total valve area. Area measurements are expressed in mm 2. 2) Onlay Characteristics. For the purposes of this study, the term "onlay" refers to "neo-tissue" (or thickening) occurring superficial to the elastic membrane toward a given surface of a valve, but deep to the endothelial lining (Figure 9, c-1). The onlay may develop on the atrial or ventricular side of the mitral valve leaflets and the ventricular or aortic side of the aortic valve cusps. In this study, onlays on the flow and non-flow surfaces of each valve were counted and measured. No attempt was made to distinguish between onlays on the two surfaces. The number of onlays was counted and recorded. Using the areas measured above, the percent onlay area (onlay area/ total valve area x 100) and the average size of onlays (onlay area/number of onlays) were calculated. 3) Thickness. Thicknesses of the total valve and valve proper were determined at 9 standardized sites. These measurements, extending from the base to the tip, were designated as base, base-mid 1, base-mid 2, base-mid 3, mid, mid-tip 1, mid-tip 2, mid-tip 3 and tip (Figure 12b). Each of these standardized sites was subjected to one measurement of the full valve thickness and one measurement of the valve proper thickness. The precise location at which each measurement was taken for a given valve section was a function of total valve length. The average thickness was measured by ImagePro Plus using a line drawn on either side of the total valve and the valve proper (Figure 12c). Onlay thickness was calculated by subtracting the valve proper thickness from the total valve thickness. An onlay refers to "neo-tissue" occurring superficial to the valvular elastic membrane or plate, toward a given surface of a valve leaflet or cusp, but deep to the surface endothelial lining. 78 Thus, a valve onlay is created by superficial expansion of valve tissue, with the elastic membrane serving to separate the "neo-tissue" from the remainder of the valve. Geographical separation of onlays was used as the criterion by which multiple onlays were distinguished from a single onlay. Based on this definition, an onlay with elastic fibers underlying it was considered a single onlay. 3.7.3 Compositional Analysis of Cardiac Valves The compositional portion of the study on cardiac valves included 177 patients and 242 valves. Eighty-nine normal valves from 54 patients, 66 rheumatic valves from 56 patients, 27 floppy mitral valves from 27 patients, 19 valves from 15 patients exposed to anorexigens, and 41 carcinoid valves from 25 patients were analyzed for valve composition. Movat's pentachrome-stained heart valves from recent autopsy hearts and explant hearts, and valve tissues paraffin-embedded at the time of an operative procedure were analyzed with ImagePro Plus software for valve proper and onlay composition ( 2 4 4 X To perform quantitative compositional analysis, areas of interest (AOIs) were drawn around the valve proper and each onlay. Each AOI was obtained by tracing around the appropriate tissue structure (valve proper or valve onlay) using a combination of autotrace and manual trace tools provided as part of the ImagePro Plus software package. A color segmentation file based on a hue (color), saturation (amount of color) and intensity (gray-scale) color model was developed to allow quantitation of each component (yellow = collagen, black = elastin, sea green = glycosaminoglycans (GAGs), red = muscle) in the Movat's pentachrome stain (Figures 13 and 14). Portions of tissue that could not be segmented into one of the four components were labeled as admixed (green), representing areas of tissue that contained at least two components that could not be distinguished because of the intimate relationship of tissue components in certain locations and because of low magnification (lx). White or clear spaces present in a given tissue section, resulting from lipid deposition and 79 80 tissue edema, were excluded from the analysis by constructing the color segmentation profile to exclude white or clear areas as defined by hue, saturation and intensity. It should be noted, however, that tissue structures or "strands" encroaching upon or traversing a clear space were included in the analysis of tissue composition. The total areas quantitated for each component were added together to constitute 100% of stained tissue area and the percent area occupied by each component was determined for each of valve proper and valve onlays. The Movat's pentachrome stain does not allow accurate and quantitative delineation of two important features of heart valves, inflammatory cells and blood vessels. Thus, a subset of valves used in the compositional analysis was also assessed by immunohistochemistry to establish the number of white blood cells and blood vessels present. The patients and valve numbers in this subset included 34 valves from 23 normal patients, 22 rheumatic valves from 19 patients, 17 floppy mitral valves, 18 valves from 13 patients exposed to anorexigens, and 27 carcinoid heart valves from 19 patients, for a total of 118 valves from 91 patients. White blood cells or vessels stained positively by immunohistochemistry were counted under a light microscope in the valve proper and each onlay. Thus, with ImagePro Plus area measurements, a final index of numbers of cells or vessels / mm 2 of tissue was obtained. 3.8 Statistical Analysis The quantitative data derived from digital images of cells and tissues was subjected to statistical analysis. The specific statistical tests and models used were dependent on the type of data being analyzed. The statistical analyses utilized in this thesis are as follows. 3.8.1 Quantitative Immunocytochemistry The amount of staining in a given image was first normalized to the cell number to mitigate the effect of cell proliferation (based on the number of nuclei) on the values obtained. The mean values of triplicate normalized samples were then calculated. Data was examined by 82 single factor analysis of variance ( A N O V A ) , followed by group-to-group Student's /-test comparisons on data sets with overall significant differences. Differences were determined to be significant at p<0.05. A l l p values were adjusted to account for the use of multiple /-tests on a single data set. Values are reported as mean ± standard error of the mean (SEM). 3.8.2 Geometric and Compositional Analyses of Cardiac Valves A l l statistical analyses were carried out using SAS (SAS Institute Inc, Cary, N C ) on a U N I X platform. Descriptive statistics were prepared using data for all valve types and diagnoses, including means and standard deviations. It is noteworthy that basic statistical derivations such as sample means, standard deviations and t-test-derived p values are not valid when making comparisons among groups where each group contains a varying set of complex variables. Thus, a mixed effect model approach, which takes into account the multiple valves per person, was used for making comparisons among groups. Variables included in these models, in addition to those designating the groups, were age, sex, diagnosis and valve type (anterior mitral valve leaflet, posterior mitral valve leaflet, aortic valve). Adjusted means were calculated from these models and Tukey-Kramer corrections for multiple comparisons were made to obtain adjusted p values. These analyses were performed for the geometric features using only aortic and mitral valves from the normal, rheumatic, floppy and anorexigen-exposed valve groups. Compositional features were assessed with this analytical approach for all valves in all groups. 3.8.3 Discriminant Analysis Stepwise discriminant analyses were used in an effort to identify a set of variables that best reveals the differences among the diagnostic groups. Discriminant functions were derived as linear combinations of the variables that maximize the distance between groups. The values of the estimated discriminant functions for each valve, called the discriminant scores, were 83 plotted to show the clustering of valves according to their diagnostic groups based on the identified set of variables. Discriminant analysis was carried out for the combined geometric and compositional features of the anterior mitral valve leaflet and separately for the geometric and compositional features for mitral and for aortic valves, respectively. It was also performed using data from leukocyte and vessel counts in combination with the compositional features of mitral and aortic valves, respectively. Due to limited sample data, right-sided valves were excluded from the discriminant analysis. 84 4 CHAPTER IV EFFECTS OF EGF AND IGF-I ON VERSICAN EXPRESSION 4.1 Introduction Versican is a prototypic member of the family of large, aggregating proteoglycans variably called hyalectans or lecticans (20). It is synthesized and secreted by a variety of cell types, including vascular SMCs (u< 59> 2 5 5X Named for the inherent versatility imparted by its modular structure, versican is a key participant in several pro-atherogenic processes, including lipid retention (118> 2 5 6 ) , cell migration and cell proliferation (21> 2 3 3 ) . Aberrant accumulation of versican has been observed in vascular lesions from patients with atherosclerosis (18> 2 5 7 \ restenosis (193> 194> 258> 2 5 9 ) and T V D (18> 19> 192X Recent evidence suggests that among many stimuli, specific growth factors present in the atherogenic microenvironment are important mediators of versican expression by vascular SMCs (20> 2 1 > 67X Despite specific and increased presence of versican in atheromatous diseases and documentation of its synthesis by vascular SMCs in vitro, the soluble factors, signaling pathways, and transcriptional/translational events that regulate its production and metabolism remain incompletely understood. The microenvironment of diseased vascular tissue harbours numerous growth factors secreted by injured and activated vessel wall cells, including resident endothelial cells and SMCs as well as infiltrating cells such as macrophages (14> 2 5 5 \ Growth factors are critical mediators of pro-atherogenic processes, among which is the production and elaboration of matrix proteoglycans, particularly versican. Analysis of intermediate and severe primary atherosclerotic lesions from hypercholesterolemic monkeys has identified large deposits of versican in areas adjacent to TGF-P- and PDGF-positive macrophages (67X Stimulation of monkey arterial SMCs with these two growth factors increases versican mRNA levels, protein synthesis and G A G chain length, demonstrating the capacity of growth factors to regulate versican expression at the levels 85 of transcription, translation and post-translational modification (25X PDGF-mediated stimulation of versican core protein expression and G A G chain elongation appear to be regulated by distinct signaling pathways, with signaling for core protein expression occurring through a RTK-dependent, PKC-independent pathway and signaling for G A G chain elongation occurring via a PKC-dependent, RTK-independent pathway (87). In addition to TGF-p and PDGF, the soluble mediators EGF and IGF-I are expressed, together with their receptors, under atherogenic conditions (260-264) EGF and IGF-I are potent stimuli linked to processes involved in SMC dysregulation associated with intimal1 thickening. EGF is a well-known proliferative stimulus for cultured vascular SMCs (265> 2 6 6 \ particularly for those cells originating from the atherosclerotic intima (260) o r exhibiting a synthetic phenotype ( 2 6 1). Recent work in a murine model of vascular injury has also linked EGF to SMC migration ( 2 6 7 \ IGF-I mediates migration and proliferation of vascular SMCs in vitro (89X IGF-I also appears to stimulate these processes in vivo, as rat aortic allografts treated with IGF-I show increased levels of IGF-I and IGF-IR in SMCs as well as larger amounts of intimal thickening (190> 268> 2 6 9X In addition to their roles in mediating vascular SMC growth and migration, EGF and IGF-I regulate the synthesis of E C M components, including fibronectin ( 2 7 0> 2 7 1) and proteoglycans (81). Recent in vitro evidence indicates that versican is among those proteoglyans modulated in response to EGF and IGF-I exposure. Studies using human adult lung fibroblasts showed that, although stimulation with EGF alone had no effect on versican expression, the triple combination of TGF-P, PDGF-BB and EGF elevated proteoglycan synthesis and versican mRNA levels beyond that seen with either TGF-p alone or TGF-p and PDGF-BB (84). Examination of proteoglycan expression by human fibroblasts derived from gingiva, periodontal ligament and regenerating periodontal defects demonstrated an increase in versican mRNA levels by all cell types in response to IGF-I stimulation (79X Furthermore, pericellular CS/DS proteoglycans 86 synthesized by rat A10 SMCs, as measured by trypsin-releasable J 5S-labeled material, were reported to increase independently of cell growth in response to IGF-I stimulation, although the identities of specific proteoglycans were not elucidated ( 8 1X The effect of EGF and IGF-I on the specific regulation of versican expression by vascular SMCs, however, has not been investigated. In the present study I have examined the expression of versican by rat aortic SMCs stimulated with EGF and IGF-I. I demonstrate that when stimulated with either EGF or IGF-I, these SMCs quantitatively increase their expression of versican core protein. Semi-quantitative analysis of versican mRNA levels revealed an early, significant increase of message for this proteoglycan in response to EGF or IGF-I. I further demonstrate that there is an increase in versican core protein expression and mRNA levels in response to simultaneous addition of EGF and IGF-I, a response that is significantly beyond that seen with either growth factor alone. 4.2 Results 4.2.1 PDGF-BB-Mediated Versican Core Protein Expression by Vascular SMCs Given that PDGF has previously been reported to stimulate the expression of versican by both vascular SMCs and fibroblasts (25> 8 4 ) , I selected this mediator as a positive control to validate the detection of growth factor-stimulated versican expression by our quantitative immunocytochemical assay. Visual analysis of immunoreactivity for versican demonstrated that, although expression of versican core protein was detectable in untreated rat aortic SMCs, those cells exposed to 20 ng/ml PDGF-BB for 24 hours showed a dramatically elevated amount of staining (Figure 15). Cells exposed to PDGF-BB displayed a staining pattern that included areas of very intense immunoreactivity (Figure 15; arrows) adjacent to areas of less intense, diffuse immunoreactivity (Figure 15; arrowheads). The intensely stained regions present in treated cells were sometimes, but not always, located within trailing cytoplasmic processes, as has been previously described (68). 87 No Growth Factor PDGF-BB Figure 15. Exposure to PDGF-BB enhances the expression of versican core protein by cultured vascular SMCs. Fluorescent digital images of rat aortic SMCs with and without exposure to PDGF-BB. Immunoreactivity for versican (red) was observed in the cytoplasm of treated and untreated cells. In PDGF-BB-stimulated cells, regions of intense cytoplasmic staining (arrows) were accompanied by areas of lighter, more diffuse cytoplasmic staining (arrowheads). Cell nuclei were detected using H33342 (blue). A negative control for the immuncytochemical procedure is provided in the inset panel. Scale bar = 50 um. 88 Quantitative image analysis of control and PDGF-BB-stimulated cells confirmed my qualitative observations. Stimulated cells demonstrated a highly significant increase (p<0.001) in total versican immunoreactivity as compared to untreated control cells (Figure 16a). Significant increases (p<0.001) in versican immunoreactivity were also observed in treated as compared to untreated cells when the very intense and more diffuse staining patterns, defined as range 1 and range 2, respectively, were analyzed independently (Figure 16b). There was a dramatic increase in the area occupied by intense staining, while the amount of lighter, diffuse staining showed a more moderate, but still highly significant increase. 4.2.2 E G F Upregulates Versican Expression by Vascular SMCs Qualitative examination of immunofluorescently stained rat aortic SMCs revealed the effect of EGF on the expression of versican core protein. Cells exposed to 10 ng/ml EGF for 24 hours clearly increased their expression of versican relative to untreated control cells (Figure 17, a and b). As with PDGF-BB-stimulated cells, areas of intense immunoreactivity and lighter, more diffuse staining were observed within the cytoplasm, while less immunoreactivity was noted in the extracellular space. Staining was particularly prominent in areas containing groups of cells apparently interconnected by long cytoplasmic extensions. Staining was also stronger in actively dividing cells (data not shown). Quantitative image analysis was employed to evaluate the effect of increasing concentrations of EGF on versican expression by cultured vascular SMCs. When total staining for versican was examined in cells treated with EGF over a limited range of concentrations, all concentrations were found to significantly increase the amount of core protein expression (p<0.02) relative to baseline levels (Figure 18a). A similar pattern emerged when the intense and lighter, diffuse regions of staining were evaluated separately, although the magnitude of the increase for regions intensely immunoreactive for versican was substantially larger than that for 89 Figure 16. Exposure to PDGF-BB increases quantitatively the expression of versican core protein by cultured vascular SMCs. (a) Quantitative analysis of total versican immunostaining in control versus PDGF-BB-stimulated SMCs. (b) Quantitation of regions of intense (black bars) and diffuse (white bars) immunoreactivity for versican. Versican immunoreactivity in untreated control cells has been arbitrarily set at one. Values are expressed as fold change in versican core protein expression over these baseline levels. Data presented are from a single experiment and are representative of results obtained from three independent experiments. *p<0.001. 90 Figure 17. Exposure of vascular SMCs to either EGF or IGF-I upregulates the expression of versican core protein. Quiescent rat aortic SMCs were maintained in serum-free medium containing no growth factor (a), 10 ng/ml EGF (b), or 50 ng/ml IGF-I (c) for 24 hours. Cells were fixed and immunoctyochemically stained for versican expression (red). Regions of intense cytoplasmic staining (arrows) and less intense, diffuse cytoplasmic staining (arrowheads) were observed. Nuclei were counterstained using H33342 (blue). A negative control for the immunocytochemical procedure is illustrated in the inset panel. Scale bar = 50 um. 91 a 1.6 5 10 20 EGF Concentration (ng/ml) EGF Concentration (ng/ml) Figure 18. EGF stimulation results in a quantitative increase in versican core protein expression by vascular SMCs over a limited concentration range, (a) Quantitative analysis of total versican immunostaining in control versus EGF-stimulated SMCs. (b) Quantitation of intense (black bars) and diffuse (white bars) regions of immunoreactivity for versican. The amount of versican expressed by untreated cells has been arbitrarily set to one. Values are expressed as fold change in versican core protein expression over levels observed in untreated cells. *p<0.02; #p<0.05. 92 regions harbouring less intense, diffuse staining (Figure 18b). The most reproducible results were obtained at a concentration of 10 ng/ml EGF. Therefore, this concentration was used in subsequent experiments requiring EGF as a stimulant. Semi-quantitative RT-PCR was used to examine whether the upregulation of versican core protein expression observed in response to EGF exposure was associated with increased levels of versican mRNA. Preliminary data indicated that P-actin mRNA levels varied markedly with culture conditions, making utilization of this gene as a "housekeeping" control or normalization control inappropriate. We did find, however, that 18S rRNA provided a stable, robust housekeeping control for normalization of PCR data. A commercially available primer/competitor system was used to ensure that the 18S rRNA PCR products would be within the linear range of amplification at the cycle number required for detection of the versican product. Specifically, this system was used to reduce the amplification efficiency of 18S rRNA. The introduction of increasing concentrations of forward and reverse primers modified at their 3' ends (competimers) to block extension by D N A polymerase resulted in a reduction in the overall PCR amplification efficiency of 18S cDNA without the primers becoming limiting and without loss of relative quantitation. Thus, I was able to keep the 18S rRNA PCR product within the linear range of amplification even when amplified for 32 cycles, the cycle number required for detection of the versican product. To determine the effect of EGF on versican mRNA levels, quiescent rat aortic SMCs were incubated for 0 to 48 hours in serum-free medium with or without 10 ng/ml EGF. Cells were harvested, total RNA was extracted and RT-PCR was performed. To our surprise, a transient increase in the level of versican mRNA was observed in untreated cells upon exposure to fresh medium, although a return to baseline values occurred within 24 hours of initial exposure (Figure 19a). The levels of versican mRNA in untreated cells at later time points fell below the levels present at 0 hours, an effect likely attributable to the ever-increasing effect of 93 v — •B b 3 O s at < Z g c 3 = c La en =c it ~~ 94 serum starvation on cellular metabolism. Extended starvation did not affect cell viability, however, as indicated by the maintenance of cell monolayer integrity in growth factor-deprived cultures. In contrast, there was a more dramatic and sustained increase in versican mRNA levels by vascular SMCs in response to EGF treatment at all time-points (Figure 19a). Quantitation of PCR products from triplicate samples at each time-point revealed a statistically significant increase in versican mRNA levels at 6 hours (p<0.01) and 12 hours (p<0.01) in treated versus control cells (Figure 19b). The level of versican mRNA peaked by 12 hours and decreased at 24 and 48 hours, although it remained significantly above control levels at these later time-points (p<0.01). 4.2.3 IGF-I Upregulates Versican Expression by Vascular SMCs Visual inspection of versican immunoreactivity in untreated cells and cells treated with IGF-I revealed results similar to those observed after EGF stimulation. Exposure of rat aortic SMCs to 50 ng/ml IGF-I for 24 hours produced increased immunoreactivity for versican core protein as compared to untreated cells at the same time-point (Figure 17, a and c). The distribution of staining was also similar to that seen after treatment with the other two growth factors tested in this study. Areas of both intense and more diffuse immunoreactivity localized to the cytoplasm, while staining in the extracellular space was limited. Quantitation of total versican immunoreactivity by SMCs in response to IGF-I stimulation using an extensive concentration range (0.5 ng/ml to 50 ng/ml) revealed a concentration-dependent upregulation of core protein expression, with concentrations of > 10 ng/ml reaching statistical significance (p<0.01) when compared to baseline levels (Figure 20a). Evaluation of intensely staining regions revealed a dramatic increase in the amount of versican expression at 50 ng/ml IGF-I (p<0.01), while less intense, diffuse staining was significantly increased at concentrations at or above 10 ng/ml (p<0.05) (Figure 20b). The most consistent 95 a 2.5! e 0 0.5 1.0 2.5 5.0 10 25 50 IGF-I Concentration (ng/ml) 0 0.5 1.0 2.5 5.0 10 25 50 IGF-I Concentration (ng/ml) Figure 20. The IGF-I-mediated upregulation of versican core protein expression is concentration-dependent, (a) Quantitative analysis of total versican immunostaining in control versus IGF-I-stimulated SMCs. (b) Quantitation of intense (black bars) and diffuse (white bars) regions of immunoreactivity for versican. The amount of versican expressed by untreated cells has been set arbitrarily to one. Values are expressed as fold change in versican core protein expression over baseline levels. *p<0.01; #p<0.05. 96 data were obtained at an IGF-I concentration of 50 ng/ml and subsequent experiments were performed using this concentration. Alterations in versican mRNA levels in response to IGF-I stimulation were examined by RT-PCR. Stimulation of quiescent rat aortic SMCs with 50 ng/ml IGF-I was associated with an increase in versican mRNA levels when compared to untreated cells (Figure 21a). Peak mRNA levels were attained within 6 hours of initial exposure (p<0.05), remained significantly elevated at 12 hours (p<0.05), and subsequently decreased to baseline (Figure 21b). 4.2.4 E G F + IGF-I Further Upregulate Versican Expression by Vascular SMCs To investigate whether EGF and IGF-I, added simultaneously, would enhance the expression of versican core protein beyond that observed with either growth factor alone, cultured vascular SMCs were exposed to a combination of these two mediators for 24 hours and versican expression was evaluated using immunocytochemistry. As with exposure to EGF or IGF-I alone, cells incubated with a combination of the two growth factors demonstrated increased immunoreactivity for versican relative to untreated cells (Figure 22, a and d). Moreover, the upregulation of versican in response to a combination of EGF and IGF-I appeared to be greater than that seen with either mediator alone (Figure 22, b-d). In particular, visual examination revealed a high level of intense staining in cells stimulated with the growth factor combination. Quantitative analysis of total staining confirmed the qualitative observation that simultaneous stimulation of vascular SMCs with EGF and IGF-I enhances versican expression beyond that seen with EGF alone (Figure 23a). A trend toward increased staining over IGF-I-stimulated cells was also noted, although this comparison failed to reach statistical significance when total staining was evaluated. Significant upregulation of versican expression as compared to IGF-1-treated cells was observed, however, when intense staining was evaluated 97 Ii 98 co o U «n e in CO rj e < c o 0) _ i o co i -CD O 2 r ' 1 ft o 2 ° * I CD . « s > o « co o CD w > o g c •2 6 CO s 2° * 2 CO u. £ ° CO CD a * CD -fl II ^ CO 4 ) l l ro i_ ro p, .2 s - o O % a o -a CU .2 CO « »-i 3 CU S 11 o CU CU CU -5 cu I-H 1 & z 8 #1 CU — kH CO C « .2 c 'CO <f l CO ^ CU J"H _ J P, ^ X £ O CO c .2 co -rt .2 1 CO "*™ kH <u > cS CD H co 5 c co a Sb o c 8 J • s 8 •—i t i 2 O .2 CD £ CD , CD a * w | s 2 2 Si 1 .2 o _ ° - 1 ? w .2 . *3 ts .2 CS cO CU s-S> <U h £ cu ' a •3 co CO CO CO -o £ 2 X I =L co m cu w M § i L co co 3 o N i-a CO O I g a r-2 5 CD a CO ft CD to _ C CD _ o 2 M w 2 -r, t-i > H hi 0CO B O T3CD CO h SO ™ <o .2 99 Figure 23. The combination of EGF + IGF-I quantitatively enhances versican core protein expression by vascular SMCs as compared to either growth factor applied alone, (a) Quantitative analysis of total versican immunostaining in control versus growth factor-stimulated SMCs. (b) Quantitation of intense (black bars) and diffuse (white bars) regions of immunoreactivity for versican. The amount of versican expressed by untreated cells has been arbitrarily set to one. Values are expressed as fold change in versican core protein expression over baseline levels. *p<0.01. 100 independently of diffuse staining. Indeed, the amount of intense cytoplasmic staining for versican seen in cells exposed to EGF and IGF-I in combination was significantly greater (p<0.01) than that seen with either mediator alone (Figure 23b). In contrast, the less intense regions of immunoreactivity were increased in cells treated with the combination as compared to EGF-stimulated cells (p<0.01), but not as compared to IGF-I-stimulated cells. Versican mRNA levels were examined to determine whether a combination of the two mediators would increase the expression of versican beyond the levels seen with either growth factor alone. Quiescent rat aortic SMCs were exposed to EGF and IGF-I alone and in combination. Cells were harvested after 12 hours of exposure and mRNA levels were analyzed. A l l treatment groups expressed versican mRNA levels significantly above that of no-treatment control cells (Figure 24). Moreover, exposure of the cells to EGF + IGF-I resulted in versican mRNA levels above the levels seen upon IGF-I stimulation alone (p<0.05), but not upon singular treatment using EGF. 4.3 Discussion In this study I tested the hypothesis that EGF and IGF-I modulate the expression of versican by vascular SMCs. I observed that stimulation of rat aortic SMCs with either EGF or IGF-I results in dramatic upregulation of versican core protein expression. I also examined the effect of these two growth factors on versican mRNA levels and found that both EGF and IGF-I significantly increase the expression of versican message by vascular SMCs. Finally, I demonstrated that exposure of rat aortic SMCs to a combination of EGF and IGF-I results in further upregulation of versican core protein expression and versican mRNA levels. Interestingly, I found that versican immunoreactivity, both in quiescent and in growth factor-stimulated vascular SMCs, was largely confined to the cell cytoplasm with limited detectable protein in the extracellular space. This finding was not an artifact resulting from the 101 # 2.0 n | I Figure 24. The simultaneous addition of EGF and IGF-I increases versican mRNA levels as compared to either growth factor applied alone. Semi-quantitative analysis of versican mRNA levels. The intensities of ethidium bromide-stained bands were obtained by digital image analysis. The ratio of versican.T8S rRNA was then determined for each sample and the average value for triplicate cultures was calculated. The amount of versican expressed by untreated cells has been set arbitrarily to one. Values are expressed as the change in versican mRNA levels relative to the level in untreated cells at the same time-point. #p<0.05. 102 immunocytochemical techniques employed, since the fixation and permeabilization procedures used in this study were chosen to ensure detection of both extracellular and intracellular versican. Further, the exquisite sensitivity provided by the immunofluorescent techniques employed in this study and the striking photostability of the Vector Red chromagen reduce the probability that extracellular versican was simply below the limit of detection. My observations are also coordinate with data previously published on human keratinocytes and dermal fibroblasts (51). The limited amount of versican in the extracellular compartment is likely due to its diffusion into the culture medium upon secretion by cells in vitro (25> 5l> 1 7 \ My observations that treatment with EGF results in upregulation of both versican core protein synthesis and versican mRNA levels by vascular SMCs provides strong evidence that EGF modulates versican expression by these cells. A previous study of pericellular CS/DS proteoglycan accumulation by rat A10 SMCs, an embryonic SMC line, reported an increase in cell number, but not in the amount of trypsin-releasable 35S-labeled proteoglycans in response to EGF stimulation ( 8 1 \ The apparent discrepancy between those findings and the findings of the present study is probably the result of methodological differences related to the measurement of proteoglycan expression. The former study evaluated proteoglycans present in the cultures, and, specifically, only those deposited in the extracellular space after secretion by the cells (81). Several studies have shown that, unlike the situation in vivo, most proteoglycans secreted by vascular SMCs in culture remain in the media and are not deposited in the extracellular space (25> 7 ? ) . As discussed above, my observation of limited immunoreactivity for versican in the E C M of cultured rat aortic SMCs supports this argument. Further, the kinetics of protein synthesis and secretion are such that, given the time-points evaluated in these in vitro studies, it is likely that a significant proportion of versican destined for the extracellular environment is still present in the cell cytoplasm. Thus, studies evaluating proteoglycan accumulation only in the extracellular space probably underestimate the impact of mediator-induced proteoglycan expression. On the 103 other hand, utilization of fluorescent immunocytochemistry in the present study permitted evaluation of versican core protein deposition in both the cytoplasmic compartment and in E C M . As such, this assay provides a sensitive technique for evaluating increased expression of versican protein associated with the cell layer in response to mediators like EGF. Further, RT-PCR was used to assess altered mRNA levels of versican, providing evidence of transcriptional upregulation of versican by these cells and confirming results obtained with immunocytochemistry. Previous studies have examined the effect of EGF stimulation on the expression of versican in cell systems other than vascular SMCs. Investigators examining proteoglycan expression by human lung fibroblasts were unable to demonstrate an effect of EGF alone on versican expression ( 8 4X In contrast, treatment of two human malignant mesothelioma cell lines, one with epithelial differentiation and one with a fibroblast phenotype, with EGF resulted in increased versican synthesis ( 2 7 2), albeit at higher EGF concentrations than required in the present study. These previous studies, together with my data demonstrating an upregulation of versican expression by vascular SMCs in response to EGF, suggest that the effects of this growth factor on proteoglycan synthesis are cell-type specific. As with exposure to EGF, I observed that stimulation of aortic SMCs with IGF-I upregulates versican core protein expression and versican mRNA levels. My results are similar to those reported in studies demonstrating an IGF-I-mediated increase in versican mRNA levels in human fibroblasts (79^ and increased pericellular CS/DS proteoglycan accumulation by rat A10 SMCs ( 8 1X The addition of IGF-I to cultures of human mesothelioma cells also upregulates versican expression ( 2 7 2 \ IGF-I-mediated versican expression by rat aortic SMCs appears to be regulated in a fashion similar to that observed with other SMCs and with fibroblasts, while EGF-mediated versican expression appears to be differentially regulated, suggesting that the 104 stimulation of versican expression by these two growth factors may proceed by distinct intracellular pathways. It is unlikely that in a complex atherogenic environment, the altered synthesis of a given gene product by a cell is controlled by exposure to a single mediator. Instead, it is probable that the altered regulation of one gene product is the result of multiple mediators interacting with cells sequentially or simultaneously via several pathways to affect transcriptional, translational and post-translational events. My data indicates that the simultaneous addition of EGF and IGF-I stimulates versican synthesis by vascular SMCs to a level above that seen with each growth factor alone. Previous studies investigating the effect of TGF-P and PDGF on versican synthesis by vascular SMCs have shown a greater effect with a combination of these two growth factors than with either mediator alone ( 2 5 \ Furthermore, the triple combination of TGF-p, PDGF and EGF has been reported to increase versican mRNA levels in fibroblasts beyond that seen with stimulation by TGF-p with or without PDGF-BB ( 8 4 \ Taken together, these results suggest that multiple mediators present in the atherogenic microenvironment may, working in concert, result in the aberrant overexpression of versican by vascular SMCs. The upregulation of versican expression by rat aortic SMCs in response to EGF and IGF-I stimulation may be related to the proliferative and migratory activities that these two mediators impart on vascular cells. EGF is well-established as a mitogen for vascular SMCs (68) and recent studies examining the effects of the EGF-R-specific inhibitor EGF-genestein in a murine model of intimal hyperplasia have suggested that activation of the EGF-R may play a role in SMC migration ( 2 6 7 \ IGF-I also stimulates vascular S M C proliferation (9°) and migration ( 8 9 \ EGF-and IGF-I-mediated growth and motility of vascular SMCs have been implicated in the progression of intimal thickening during atherogenesis ( 1 9 1), a process in which proteoglycans such as versican also play an important role. 105 Versican is rapidly becoming understood as a multifunctional molecule with anti-adhesive, pro-proliferative properties. Versican is excluded from focal contacts and has been shown to inhibit cell-substratum adhesion in fibroblasts ( 6 5X Studies of astrocytoma cells have shown that cultures expressing high versican levels also exhibit high motility ( 9 7X Human arterial SMCs exposed to PDGF rapidly secrete a hydrous coat rich in versican around the trailing cytoplasmic processes prior to cell rounding, potentially facilitating cell mitotic events and cell movement by decreasing cell adhesion and increasing extracellular matrix viscosity (68> 9 2 ) . Recent studies have suggested that versican may directly stimulate cell proliferation through its G l and G3 domains, an event that can be blocked by anti-sense oligonucleotides to the EGF-R (93, 94). Limited investigation of the signal transduction pathways involved in the regulation of mediator-induced versican expression by vascular SMCs has suggested that RTK-dependent, PKC independent mechanisms are at play (83> 8 7 ) . Similar signal transduction events have been implicated in the activation of vascular SMCs upon exposure to IGF-I and EGF. Recent work on the signal transduction pathways activated via stimulation of the EGF-R expressed by vascular SMCs have focused on EGF-R transactivation by stimuli other than EGF, including angiotensin II ( 2 7 3) and endothelin-1 ( 2 7 4 \ Although somewhat indirect, these studies indicate that EGF-mediated SMC proliferation requires activation of the PI3K and M A P K pathways. Studies have shown that IGF-1-mediated stimulation of SMC migration involves both PKC and PI3K, and leads to downstream activation of PKB/Akt (89> 275> 2 7 6 X On the other hand, activation of the M A P K pathway is required for cell proliferation induced by this growth factor ( 2 7 7> 2 7 8). Whether these signaling pathways are important in the modulation of versican expression by vascular SMCs in response to EGF and IGF-I stimulation remains to be determined. 106 5 C H A P T E R V O X I D I Z E D L D L A N D V E R S I C A N IN T V D 5.1 Introduction The influx, retention and oxidative modification of L D L are key events in atherogenesis (279-282) Perhaps owing to its heterogeneous composition, oxidized L D L (oxLDL) appears to mediate a wide array of pro-atherogenic processes (Figure 25). Minimally modified L D L is chemotactic for monocytes ( 2 8 3> 2 8 4) and T-cells ( 2 8 5), but inhibits motility of tissue macrophages, suggesting that it may be involved in recruiting monocytes and trapping macrophages in the developing lesion. Extensively oxLDL binds scavenger receptors on macrophages, leading to unregulated cholesterol accumulation and foam cell formation ( 2 86-288) High concentrations of oxLDL have cytotoxic effects on endothelial cells (289> 2 9 °) and SMCs (291> 2 9 2 ) , while lower concentrations promote cell growth (293> 2 9 4 X O x L D L also modulates the expression of adhesion molecules such as VCAM-1 (295> 2 % ) and cytokines such as interleukin 6 (IL-6) ( 2 9 7) by vessel wall cells. Moreover, vascular tone may be affected by oxLDL through interference with nitric oxide-mediated vasodilation, resulting in vessel dysfunction (298> 2 9 9 X L D L oxidation produces a varied mixture of biologically active intermediates, including lipid hydroperoxides, oxysterols, lysophosphatidylcholine (lysoPC) and reactive aldehydes ( 3 0 0 ' 3 0 1 ) . Lipid peroxidation represents a major source of reactive aldehydes in vivo. The free radical-mediated oxidative breakdown of polyunsaturated fatty acids in phospholipids, cholesterol esters and triglycerides of L D L generates the highly reactive aldehydes M D A and HNE ( 3 0 2 ' 3 0 3 ) . Such aldehydes are stable compared to their free radical initiators and are considered end-products capable of diffusing into the microenvironment where they may act in a paracrine manner. These aldehydes are strong electrophiles and may undergo a Micheal addition reaction with cysteine, histidine and lysine residues of apoB or associated proteins to form stable 107 00 I '-3 c H-» . , » t t ) 1 on -o 0 fi 8 OT a § "2 £ g x> .5 o 1 ~ o o . O H * c .2 '•3 a -§ s I—I CO d Cm O a -a (U CO CO 1) H O c co a CJ —' <D *_Ct "*-» 5 o — o co .ST1 CO i—i O - T H O N S3 X S3 co c £ l 00 CO O 2 ° O H O 1 ~ CO 1 & 2 M & u a sa P S <D cO £ J -° 3 CO 5 0 o o a O f H 3 ^ -a CB 5 « I- o a ^ 3 CO O u s m .2 H^ .2 eg co co co >, 2 *« U CO c o CL) 5 , " O "d co a u S3 o CO i c o 1 CfH "3 s CO I '53 o o a T3 CU N O £ £ o CJ <u 00 IT, .£ CU I* s S 5 CO CH_! 3 ° £ 4 CJ CO o +^ CO u on 00 X o _ a B a CJ CO CJ 108 aldehyde-lysine adducts. Such adducts are strongly implicated in macrophage-derived foam cell formation during atherogenesis. HNE is a major hydroxyaldehyde product generated during linoleic acid and arachidonic acid oxidation and reacts with cysteine, histidine and lysine residues, primarily forming stable hemiacatyl adducts ( 3 0 2 X The ketoaldehyde M D A is another abundant aldehyde produced during lipid peroxidation and often forms adducts with lysine residues ( 3 0 2). Alternatively, these aldehydes may form adducts with amine headgroup-containing phospholipids such as phosphatidylserine and phosphatidylethanolamine. Both of these aldehyde adducts are present in oxLDL and are useful biological markers of oxidatively modified lipoproteins in vivo ( 3 0 1 \ The presence of oxLDL in atherosclerotic lesions has been documented in a number of studies. Antibodies against oxidation-specific epitopes have been used to localize oxLDL in atherosclerotic lesions derived from humans, rabbits and mice ( 2 4 8> 2 5 1 > 304-306) Autoantibodies reactive against oxLDL epitopes have been demonstrated in lesions and plasma of humans and animals ( 3 0 4> 507-309) Lipoproteins extracted from atherosclerotic lesions show evidence of oxidative modification (309> 3 1 °) . A series of recent studies in L D L receptor-deficient rabbits and mice has used radiolabeled M D A 2 antibodies to identify and measure the content of oxLDL in atherosclerotic arteries (311-313) j n e findings of these studies demonstrate that in vivo uptake of 1 2 5 I -M DA2 is an accurate measure of oxLDL-rich lesions, is sensitive to regression subsequent to hypocholesterolemic intervention, and correlates with circulating autoantibodies. The extent to which lipoproteins in the arterial lesions of human heart allografts are oxidatively modified, however, is unknown. Prominent intracellular and extracellular accumulation of lipids in the intima and media of coronary arteries from failed human heart allografts has been identified and extensively documented (143> 192> 255> 256> 3 1 4 ) . Immunohistochemical analysis using specific antibodies established the presence of apoE, B, and (a) in patterns of distribution distinct from those seen in 109 native atherosclerosis ( ] 9 \ The presence of both apoE and apo(a) in early T V D and in severe disease has suggested that lipid accumulation persists throughout the course of TVD. These apolipoproteins colocalized with versican and biglycan, two proteoglycans that are overexpressed in T V D (18> 1 9 \ This apolipoprotein-proteoglycan colocalization suggests an interaction that ultimately enhances lipid retention, oxidative modification and foam cell formation, but the oxidative status of this accumulated lipid and its relationship to vessel wall cells and matrix constituents in T V D has not been investigated. In the present study I have used immunohistochemistry to evaluate the oxidative status of L D L in coronary arteries from human heart allografts, using atherosclerotic coronary arteries from human native hearts as controls. I have also investigated the relationship between oxLDL and the location of versican and vessel wall cells, especially macrophages and SMCs. 5.2 Results 5.2.1 Oxidized Lipoproteins in Vessels with M i l d T V D Allograft coronary arteries with mild T V D lesions exhibited striking immunoreactivity for oxidation-specific epitopes using antibodies against both H N E - L D L and M D A - L D L (Figure 26). Staining for H N E - L D L was distributed circumferentially in cell-rich areas of the lesions, appearing especially prominent in the deeper intima (Figure 26, a, d, g). The cell-associated staining was typically focal and was distributed in an intracellular/pericellular pattern. Limited staining of the vascular media was also visible, while trace staining was observed in certain locations along the superficial intima. The overall distribution of M D A - L D L immunoreactivity was similar to that noted for HNE-LDL, but the relative extent and intensity of staining for this oxidatively modified epitope was more dramatic (Figure 26, b, e, h). In addition to intensely immunoreactive areas associated with cells deep to the intima/media boundary, cells at and just below the endothelial lining also appeared strongly positive. Limited diffuse immunoreactivity 110 Figure 26. Oxidative status of lipoproteins in the vessel wall of a representative coronary artery with mild TVD. Shown are serial cross sections of the L A D coronary artery from a 28-year-old male recipient (45-year-old female donor) explanted 411 days post-transplant and having 27% luminal narrowing. Vessels were stained for H N E - L D L (a, d, g), M D A - L D L (b, e, h), and LOP-L D L (c, f, i). Adjacent sections from the same artery are shown labeled for the specific apolipoproteins apo (a) (j), apo B (k) and apo E (1). Inset panels illustrate the negative control for each antibody (a, IgG2A; b, IgGl; c, omission of antibody). Black boxes in a-c delineate regions of interest in d-i. Brightfield images in d-f are shown in parallel with corresponding fluorescent images (g-i). Fluorescent images were captured using identical exposure times at 12 bits per pixel (12 bpp). The display range for all 3 images has been gated from 200 to 2900 grey levels. Scale bars: a-c, 500 um; d-i, 40 um; inset panels, 200 um. I l l for H N E - L D L and M D A - L D L was present in the adventitia of some vessels, as has been previously reported for these antibodies ( 2 4 8 ' 2 5 1 \ In sharp contrast to positive staining for HNE-L D L and M D A - L D L , allograft vessels with mild disease had absent to trace immunoreactivity for LOP-LDL (Figure 26, c, f, i). To demonstrate that the oxidation-specific epitopes observed in these mildly diseased allograft arteries were associated with lipoproteins, tissue sections stained for HNE-LDL, M D A -L D L and LOP-LDL were compared to adjacent tissue sections stained with antibodies against apo(a), apoB and apoE (Figure 26, j-1). As expected, significant colocalization between these lipoprotein markers and oxidation-specific epitopes, especially H N E - L D L and M D A - L D L , was observed. The high degree of coincident staining for M D A - L D L and apo(a) was particularly interesting, given the reported atherogenic potential of this lipoprotein moiety ( 3 1 5 X 5.2.2 Lipoprotein Oxidation in Severe T V D As with mild T V D lesions, immunohistochemical staining revealed the presence of oxidation-specific epitopes in allograft vessels with severe T V D (Figure 27). However, a shift in the composition and distribution of oxidatively modified lipid was apparent in severe lesions when compared to mild disease. Diffuse, extracellular staining for H N E - L D L was observed throughout the intima of severely diseased vessels and was particularly prominent in the atheromatous core, especially in the extracellular space surrounding cholesterol clefts and adjacent to lipid-laden foam cells (Figure 27, a, d, g). Regions of increased staining were also evident just beneath the endothelium. Immunoreactivity in the superficial intima was generally weak and diffuse, although patchy areas of increased staining intensity were noted. When present, the overall distribution of M D A - L D L immunoreactivity in the intima of severe lesions was similar to that observed for HNE-LDL. The staining was less intense, however, and often only trace amounts were observed (Figure 27, b, e, h). Indeed, a general 112 Figure 27. Oxidative status of lipoproteins in the vessel wall of a representative coronary artery with severe TVD. Shown are serial cross sections of the L A D coronary artery from a 66-year-old female recipient (30-year-old male donor) explanted 499 days post-transplant and having 77% luminal narrowing. Vessels were stained with HNE-LDL (a, d, g), M D A - L D L (b, e, h), and LOP-LDL (c, f, i). Adjacent sections from the same artery are shown labeled for apo (a) (j), apo B (k), and apo E (1). Black boxes in a-c delineate regions of interest in d-i. Brightfield images in d-f are shown in parallel with corresponding fluorescent images (g-i). Fluorescent images were captured using identical exposure times at 12 bpp. The display range for all 3 images has been gated from 100 to 2500 grey levels. Scale bars: a-c, 500 um; d-i, 100 um. 113 trend toward decreased extent and intensity of staining for M D A - L D L with increasing lesion severity was evident in the allograft vessels examined. A limited amount of immunoreactivity for H N E - L D L and M D A - L D L was also present in the adventitia of these severely diseased vessels. In contrast to the diffuse, extracellular staining pattern observed for H N E - L D L and M D A - L D L , staining for LOP-LDL was focally very prominent in severely diseased vessels (Figure 27, c, f, i). Striking, punctate staining was observed in foam cell-rich areas of severe lesions, adjacent to the central, atheromatous core. This intensely positive staining pattern, which was also evident in more moderate lesions, was found to both rim the cells and stipple the cytoplasm. This intracellular staining pattern was present in areas rich in intimal cells, but was absent from acellular regions close to the intima/media border. The adventitial connective tissue was largely negative for LOP-LDL, although adventitial lipid deposits were sometimes positive. Comparison of severe lesions stained for oxidation-specific epitopes with adjacent tissue sections stained for apo(a), apoB and apoE revealed substantial colocalization of H N E - L D L immunoreactivity with all three lipoproteins in the lipid core (Figure 27, j , k, 1). Diffuse staining for H N E - L D L was also apparent in the matrix underlying more cellular areas of the lesions, as was staining for the native lipoproteins. There was no colocalization between the native lipoproteins and the focally intense, punctate staining for LOP-LDL present in foam cell-rich areas. 5.2.3 Association of O x L D L with Vessel Wall Cells Immunohistochemical staining of allograft coronary arteries for oxidation-specific epitopes and particular apolipoproteins revealed the presence of a large amount of cell-associated oxLDL, depending on the epitope stained for and the lesion severity. To determine which specific cell types within the vessel wall were associated with the presence of oxLDL, a 114 comparison of serial sections stained for oxidation-specific epitopes, SMCs and macrophages was performed. In mild lesions, the staining pattern and distribution observed for H N E - L D L and M D A - L D L were consistent with the distribution of intimal and medial SMCs, but not of CD68-positive macrophages (Figure 28, a-c). Staining for LOP-LDL in severe lesions, on the other hand, was associated with foam cells derived from both intimal SMCs and CD68-positive macrophages (Figure 28, d-f). Small clusters of leukocytes were present in moderate and severe lesions in areas associated with SMC accumulation, but did not appear to be specifically associated with regions containing oxLDL (data not shown). 5.2.4 Colocalization of O x L D L and Versican In addition to the presence of oxidation-specific epitopes in association with vascular SMCs and CD68-positive macrophages, oxLDL was detected within the E C M of allograft arteries. We have previously demonstrated that lipoproteins accumulating in the vessel wall of allograft coronary arteries colocalize with interstitial proteoglycans, especially versican (19> 1 9 2 ) . In this study we observed an overlapping distribution of versican and oxLDL in T V D lesions (Figure 29). Vessels with mild lesions were intensely immunoreactive for versican, both in the neointima and vascular media (Figure 29a). The dramatic accumulation of versican in these lesions corresponded with SMC-rich areas as well as with SMC-associated staining for M D A -L D L (Figure 29, b and c). In severe lesions, focal deposits of versican were located around and immediately adjacent to intracellular L O P - L D L deposits, many of which appeared to be SMC-derived foam cells (Figure 29, d-f). Somewhat surprisingly, the large extracellular accumulation of H N E - L D L and, when present, M D A - L D L in the atheromatous core of severe lesions did not colocalize with versican expression. 115 116 CO s p o x « X § * -a -g 3 £ i—i cs C O H c e o T3 a !8 > f— GO c & CO 21 » 2 CO 1 « VH . . 1 ^ d "3 E © C5 CD >, £1 CD >-< •5 P .& > O s -CD . O tH <4_ CD T3 u G cs 2 O CO . £ X ! O H i l l a o •d CJ, <D ^ c « d 3 .2 O OH £ o CD >> K "© CD s -Co U — u-> CO cs u C+H > O CD p, CD H • •3 & C3 d o s— O u < — CD .>: *3 * 3 I 1*5 "O cS CS T j -O o 00 to £ w d d cs o e C3 fa ^ > ' o d o 1 & N CS _d d cs o 9 fa ON J CN CD CV £ j M s £ OD § fa * a OH CO 3 Co EH CO o PH CO 00 CS d T3 5 2 cS d _ <D - J Q - J i c-0 —i C cS S i ^ . i—I < ft Q co /—v CD ••a 117 5.3 Discussion Previous investigations by members of the laboratory have documented prominent intracellular and extracellular accumulation of lipids in coronary arteries from failed human heart allografts (14l> 143> 192> 255> 2 5 6 ) . Subsequent analysis of these lesions using specific antibodies established the presence of apoE, B, and (a) in patterns of distribution distinct from those seen in native atherosclerosis and demonstrated the colocalization of these lipoproteins with interstitial proteoglycans, particularly versican (19). The present study has, for the first time, defined the oxidative status of these lipid accumulations and demonstrated specific associative relationships among oxLDL, vessel wall cells and versican. A panel of monoclonal antibodies against oxidation-specific epitopes was used to define the oxidative status of lipid accumulations in human cardiac allograft arteries. Vessels affected by mild T V D were strongly positive for H N E - L D L and M D A - L D L . In general, these epitopes shared a similar distribution, localizing to intimal and medial SMCs. The distribution of HNE-L D L and M D A - L D L shifted toward more diffuse, extracellular staining in severe lesions, although some cell-associated staining remained evident in certain arteries. M y observations regarding the differential staining in mild and severe lesions using these two antibodies is congruent with observations reported by other investigators in animal and human models of native atherosclerosis (248> 250> 2 5 J ) . Examination of oxLDL in the Watanabe heritable hyperlipidemic rabbit revealed the presence of macrophage-associated, particulate and annular staining in early and transitional lesions, while advanced lesions exhibited a shift toward extracellular staining ( 2 4 8 X In studies using human fetal aortas from normocholesterolemic and hypercholesterolemic patients as a model of early atherosclerosis, staining for oxLDL was 118 described as intracellular and associated with macrophages, although some diffuse extracellular staining was present ( 2 5 1 X Although the general distribution of H N E - L D L and M D A - L D L in T V D as seen in our study mirrors that described in other investigations, an apparent distinction does exist with regard to the specific cell types associated with positive staining. While observations in rabbit and human atherosclerotic lesions have implicated the macrophage as the major cell type associated with the presence of oxLDL ( 2 4 8 ' 2 5 ] ) , our observations suggest that vascular SMCs represent the major cell type associated with aldehyde-modified lipoproteins in human allograft arteries. Substantial in vivo and in vitro data now exists describing the role of vascular SMCs in lipoprotein modification, uptake and retention ( 2 5 5 X It is interesting that in human TVD, M D A - L D L staining was prominently immunoreactive in mild lesions but was much weaker in severe lesions, while the reverse appeared true for HNE-LDL. It is tempting to speculate that the difference is biologically important, but caution must be used when interpreting observations of relative intensity between two different antibodies, especially when technical limitations require that the antibodies be used in separate runs, as was the case in this study. In addition to antibodies directed against the aldehyde-protein adducts H N E - L D L and M D A - L D L , I utilized another monoclonal antibody specific for oxidatively modified human L D L , called LOP-LDL, to more fully characterize the oxidative status of lipoproteins in human TVD. I found that while mild T V D lesions were largely negative for LOP-LDL, pockets of intense punctate intracellular immunoreactivity were observed in the foam cell-rich areas of severe lesions. This particulate staining was associated with intimal SMCs and some CD68-positive macrophages. My results regarding the pattern and distribution of staining for LOP-L D L in human heart allografts are similar to observations reported in a rat model of focal glomerulosclerosis ( 2 5 3). In that study, oxLDL present in the glomeruli of diseased rats was 119 described as granular and punctate, and was diffusely distributed within the cell cytoplasm of ED-1-positive macrophages. The punctate, cytoplasmic staining pattern for oxidatively modified L D L observed in foam cells has led to speculation that such products enter the cell via phagocytosis or receptor-mediated entry and are resistant to enzymatic degradation, thus providing a mechanism for their accumulation ( 2 5 3 X Our results showing the presence of LOP-L D L in association with SMC-derived foam cells, in addition to macrophages, provide new evidence of an important role for SMCs in accumulation of oxLDL in human TVD. The presence of oxLDL in the allograft microenvironment and its proximity to intimal and medial SMCs, especially in mild lesions, suggests it may play an important role in regulating pro-atherogenic cellular processes. Evidence from in vitro studies has suggested that, as an active component of oxLDL, H N E plays a role in vascular SMC proliferation, gene expression and apoptosis. Exposure of vascular SMCs to relatively low concentrations of HNE, either alone or together with mediators such as serotonin (5-HT), Ang II and urotensin, results in activation of c-Jun NH2-terminal kinase (JNK), p38 and extracellular signal-regulated kinase (ERK), increased gene and protein expression levels for c-fos and c-jun, and increased levels of AP-1 DNA-binding activity ( 2 9 4> 316-320) HNE has been shown to derivatize and activate the EGF-R, suggesting that the EGF-R can react to lipid and highlighting a potential mechanism for partial blockade of HNE-mediated cell proliferation by erbstatin A , an EGF-R inhibitor (320> 3 2 1 ) . Studies have also demonstrated that HNE is capable of inducing vascular endothelial growth factor (VEGF) secretion in SMCs and localization of V E G F in balloon-injured baboon aortas correlates with the generation of HNE, suggesting that reactive oxygen species can regulate V E G F expression ( 3 2 2 \ a potentially important mediator of endothelial permeability and lipid insudation in T V D ( 2 5 5 \ While low concentrations of HNE mediate proliferation, higher concentrations induce apoptosis in vascular SMCs, indicating a bifunctional role for H N E in atherogenesis ( 3 2 3 X 120 HNE-modified L D L also exhibits pro-atherogenic actions on macrophages. Lipid accumulation and degradation by macrophages and SMCs is decreased in the presence of L D L derivatized by low concentrations of H N E ( 3 2 4 X Increased formation of HNE-apoB crosslinks and blockage of the e-amino groups on apoB lysine residues are among the modifications made under these conditions. On the other hand, modification of L D L by higher concentrations of HNE results in L D L aggregation ( 3 2 4 \ Based on biochemical and ultrastructural studies, HNE-induced L D L aggregation leads to increased uptake by macrophages via a phagocytotic pathway and, ultimately, foam-cell formation (324-326) Although the panel of antibodies used in this study has been characterized in previous studies as recognizing oxidatively modified L D L , the antibodies actually recognize a range of oxidation-specific epitopes derived during lipid peroxidation. Thus, in addition to recognizing oxLDL, these antibodies may react with other oxidatively modified moieties in the vessel wall (248, 249) To verify that the oxidation-specific epitopes observed within the allograft arteries represented oxLDL, we compared tissue sections stained for HNE-LDL, M D A - L D L and LOP-L D L with adjacent sections stained for the specific apolipoproteins apo(a), apoB and apoE. This comparison revealed a co-distribution of the apolipoproteins, particularly apo(a) and apoE, and oxidation-specific epitopes in mild and severe lesions, strongly suggesting that oxidation-specific antibodies recognized oxLDL in these lesions. It is noteworthy that, particularly in severe lesions, colocalization of apolipoproteins and oxLDL epitopes occurred to the greatest extent in the E C M and not in association with cells. This observation is supported by previous work demonstrating colocalization between M D A - L D L and apoB in advanced atherosclerotic lesions (248). The response-to-retention hypothesis ("> 1 0°) argues that retention of atherogenic lipoproteins within the vessel wall represents the major initiating event in early atherogenesis. Once trapped beneath the endothelium, these lipoproteins can initiate and potentiate a response 121 cascade leading to disease. Arterial proteoglycans, especially SMC-derived interstitial proteoglycans such as versican, represent key molecules capable of binding, retaining and modifying lipoproteins ( 1 1 8). Specifically, L D L is known to form complexes with versican through ionic interactions between specific positively charged domains of apoB and the negatively charged G A G chains of the proteoglycans ( 1 0 5). This interaction contributes to lipid retention and enhances internalization by macrophages. Moreover, the association of L D L with proteoglycans enhances the suseptibility of L D L to oxidative modification (327> 3 2 8 X In the present study I found that versican, shown previously to be dramatically overexpressed and co-distributed with specific lipoproteins in human allograft coronary arteries, was widely expressed and colocalized with SMC-associated oxLDL in mild T V D lesions. This associative evidence supports the contention that proteoglycan-mediated retention of lipoproteins plays an important pro-atherogenic role in early vascular disease. A similar relationship between cell-associated oxLDL and versican could be demonstrated in severe lesions where focal increases in versican accumulation colocalized with SMC-derived foam cells. It is possible that the interaction between versican and LDL, and the oxidative modifications that ultimately result from this interaction, lead to the formation of SMC-derived foam cells. Recent studies have demonstrated the uptake of aggregated or fused L D L by vascular SMCs through the LDL-receptor-related protein (LRP), a cell-surface protein overexpressed in atherosclerotic lesions ( 3 2 9). Further, the versican-LDL interaction leads to the production of both monomeric L D L particles and fused L D L particles capable of increasing cholesterol ester accumulation in vascular SMCs ( 3 3°). The former particles are internalized via the L D L receptor, while the latter particles are taken in through the LRP ( 3 3°). These data suggest a role for versican in the mechanism for oxLDL uptake by vascular SMCs and provides a rationale for the association of versican with foam cells containing oxLDL. 122 It is interesting that the diffuse, extracellular staining for oxLDL, particularly H N E - L D L , in the atheromatous core of severe lesions localized to areas largely devoid of versican. The reasons for this pattern of oxLDL localization remain speculative, but may be related to changes in the interaction between oxLDL and versican as lesion development progresses. Recent in vitro studies have suggested that oxidation of L D L complexed with CSPGs effectively weakens the ionic interaction between L D L and the proteoglycan, potentially resulting in the release of the oxidatively modified lipoprotein ( 1 2°). The altered lipoprotein, now at increased risk for further oxidation, may then aggregate and become trapped in the lipid-rich core of the growing lesion. The reduction in positive charge on the apoB molecule due to increasing oxidation may ultimately overwhelm the increased binding capacity of versican arising from G A G chain elongation ( 1 1 2) and the presence of bridging molecules such as LpL ( 1 1 0 X One study has suggested recently that nonproteoglycan components of the E C M may be involved in the binding of oxidized lipoproteins ( 3 3 1), providing a potential mechanism for the extracellular retention of lipoproteins in more advanced lesions. Weakened interactions between L D L and versican may also facilitate the uptake of oxLDL into macrophages and SMCs, effectively removing oxLDL from the versican-rich extracellular space and simultaneously enhancing foam-cell formation. In summary, I have demonstrated the presence of oxLDL in vascular lesions from failed human heart allografts. Oxidatively modified lipids were present in both mild and severe lesions, but the precise oxidative status and location were dependent, to a certain degree, on lesion severity. OxLDL was coincident with intimal SMCs and versican accumulation in mildly diseased arteries. In severely diseased vessels, intracellular accumulation of oxLDL was evident in SMC- and macrophage-derived foam cells in association with versican deposition. Diffuse, extracellular accumulation of oxLDL also occurred in the acellular, lipid-rich areas of severe lesions. These studies suggest that a delicate balance between bound and free oxidized lipoproteins may be important in the pathogenesis of atheromatous diseases such as TVD. 123 6 CHAPTER VI PROTEOGLYCAN EXPRESSION IN CARDIAC VALVES 6.1 Introduction The need for greater detail regarding the clinical and morphological features of normal and diseased human heart valves has been fostered by the recent discussions and controversy arising from anorexigen-associated valvular dysfunction. When the therapeutic combination of fenfluramine (fen) and phentermine (phen) was linked by investigators at the Mayo Clinic in Fargo, North Dakota to regurgitation of both mitral and aortic valves ( n ) , pharmaceutical companies rapidly and voluntarily withdrew fen and dexfenfluramine (dexfen) from the market. A large number of clinical studies have since been initiated in an attempt to define the existence (332-337)^ n a t u r e (338-340), severity (341-344) a n d potential reversibility (345-351) 0 f a n y valvular dysfunction (or lesion) that might arise in association with the use of anorexigens. These investigations have led to better definition of many clinical aspects, including disease prevalence, but the potentially unique histopathology of such valves has received far less attention ( 1 6°). At the time the present study was initiated, clinical findings suggested that left-sided heart valves were affected to a greater extent by anorexigenic agents than right-sided heart valves, including evidence of mild aortic valve regurgitation ( u > 3 3 3 ,342 ) a n ( j a lesser frequency of mitral valve regurgitation ( H \ No data from controlled studies suggested tricuspid or pulmonic valve dysfunction. Information has begun to emerge that the occurrence of valvular dysfunction is dependent upon duration of therapy and drug dosages ( 3 4 0> 352) Gardin and colleagues ( 3 4 0) have reported a higher prevalence of aortic regurgitation in patients exposed to fen/phen or dexfen versus controls (13.7% fen/phen, 8.9%, dexfen, 4.1% untreated group), with no differences in valvular dysfunction when treatment was less than three months. Similarly, Jollis and coworkers ( 3 5 2) evaluated 1163 patients who had taken fen/phen and 672 control patients. 124 Valvular abnormalities occurred primarily after patients were on the drug regimen for more than six months and dysfunction was characterized predominantly by mild aortic regurgitation (9% of patients as compared to 4% of normal controls, p<0.001). Mild regurgitation was also evident in the mitral, tricuspid and pulmonic valves in the treated patients, but this tendency was not statistically different from controls. Importantly, valvular dysfunction associated with fen/phen may be reversible upon cessation of the drugs ( 3 4 5> 3 4 6 > 3 4 8-350,353) Weissman and associates ( 3 3 3 ' 3 4 9 ) have demonstrated that aortic and mitral regurgitation were in excess in dexfen-treated patients versus controls wherein the average duration of therapy was 72 days and the median time of follow-up echocardiogram after cessation of drugs was 34 days. Recently, Gardin and associates ( 3 5°) have completed a one-year follow-up on the original cohort of dexfen-treated and control subsets, and their data indicate that reversibility of dysfunction is to be expected in the vast majority of patients. Despite the work published to date, much of which relies on echocardiographic data, few studies have described the pathological changes to heart valves that result in the reported regurgitant valve dysfunction associated with exposure to anorexigens ( 1 6°). Indeed, a very limited number of preliminary and anecdotal pathological observations have been published on valves from patients taking anorexigens who, upon clinical evaluation for valvular regurgitation, underwent either aortic and/or mitral valve replacement (n> 1 6 1> 3 5 4), or who had right ventricular endomyocardial biopsy ( 3 5 5). Most recently, a larger study was published describing the histopathology of aortic and mitral valves removed from 64 patients treated with anoretic agents (356) j n e a [ m 0 f the present study was to more completely define the distinguishing histopathologic features of valves taken from individuals exposed to anorexigens in the context of normal valve histology and the histopathology of other diseased valves. I have examined quantitatively the geometry and composition of mitral and aortic valves removed surgically from 125 patients treated with anorexigens in comparison to normal valves and in comparison to those from patients with rheumatic and floppy valve disease. In addition, I have compared compositional features from a small number of tricuspid and pulmonic valves from patients treated with anorexigens and a substantial number of right-sided heart valves affected by carcinoid valve disease to the floppy, rheumatic, and anorexigen-exposed aortic and mitral valves. An effort was made through both univariate and discriminant analyses of quantitative variables to establish the distinctive microscopic features that separate one valve group from another. Finally, MFBs were isolated from cardiac valve tissue and the expression of versican by these cells was documented. 6.2 Results 6.2.1 Geometry In Figures 30 and 31, the group means adjusted for age, sex, and valve type from the mixed effect model analyses are presented for those cases with aortic and mitral valves. The cross-sectional area of the total valve and the valve proper regions were not statistically different among comparison groups (Figure 30, a and b). Differences in the average valve onlay area among the groups were, however, highly significant (pO.OOl). Specifically, rheumatic valves had a larger average onlay area than any of the other valve categories, while floppy valve onlay areas were larger than those of normal valves (Figure 30c; Figure 32, a-c, Figure 33, a and b). When comparing the percent onlay area, I found a similar trend (Figure 30d), with rheumatic valves having the greatest percent onlay area (35%), followed by floppy valves (25%), anorexigen-exposed valves (23%), and normal valves (6%). Rheumatic valves had the greatest 2 2 average onlay size (7 mm) (Figure 31a), followed by floppy valves (4 mm), anorexigen-exposed valves (2 mm2), and normal valves (1 mm2). The number of onlays (Figure 31b) was the greatest in the anorexigen-exposed valves, with approximately three onlays per valve 126 1 i C <u ft ft o T 3 U GO o c u I a u 2 P cu o • es 4) -ca e O u 0. U TO s i co e -5 Sb M co .3 <u CO <u <u cu ? i CO 5 © 2 . 3 <» 0L| "< CO . >~ co J T Co —I cO O S X J 127 Pi CH o IB C O IM o u 43 E 3 Z CD t/3 O 93 i C CU bp "3 93 e2 E <! I • • CJ 1= -o a. o C Cu <J <+-> CU cu O op 93 cu 00 o c Co ^ 2 £ <U fa cS cn nj cu * Ik. * Co CCS Co Co CJ Co E a "c. 09 W cu -S .a GO cu OJD 93 < > > o H O £ •o e 128 s Sf ** CU -o S < CU > > u -a <u en e a x w cu ii, re -5 ^ 3 £ CD re £ 3 CU CU 5-2 > to r] 53 <D >> C B <D J H 9 § J CD .ES « O f > . ca co co > I I I s o S J3 0! s o z CU > 3 I S *S £ 129 T3 > i > « 5 .55 9> © o Z ^^^^^^^^^^^^^^^^^^ co T3 S I js* ca $ -0 S 5 T 3 O o 1 o 3 OH 5, X _7 « 9 2 ca <D o .S? o < X 0 ) I-o § co > fO, "3 > o g i .S 3 o <u 3 -S 1? <U 3 co H a | X c U C+H a ° <U C O .gp C * .2 u _ b Q 0 <D 3 co CS co ~ C O 2 o 1 o 2 u I i "3 g S « I I co •"" s C+H .2 ° ri co <D U C O 00 C O 2 C O H O -3 0 T3 *o 2 43 ca * i > >> 1 ' a-3 is? • 2 2 s 0 5 -3 :? 2 a ft ca o 130 (Figure 32d and Figure 33c), while normal valves had the fewest (~1 onlay/valve). Figure 31c and 3Id depict the average thickness of the total valve and valve proper, respectively. The average thickness of the rheumatic, floppy, and anorexigen-exposed valves was found to be greater than the normal valves, with the rheumatic valves being the thickest. 6.2.2 Composition I recently carried out a compositional analysis of human normal heart valves and their onlays as they age using a digital imaging system ( 2 4 4). I have now extended that analysis to the assessment of valves in different comparison groups. The valve proper and valve onlays were analyzed separately for their composition. Adjusted means are shown in Figure 34 for both the valve proper and onlays, including each constituent distinguishable by digital image analysis, for the mitral or tricuspid and aortic or pulmonic valves. The mean compositional contributions to the valve onlays (Figure 34, a and c) and the valve proper (Figure 34, b and d) are illustrated to show the similarities and differences among comparison groups. In the valve proper, it can be seen that the anorexigen-exposed, carcinoid and floppy valves have a greater relative amount of GAGs than the normal and rheumatic valves, while the rheumatic and normal valves have a greater percentage of collagen than the other valve categories. Features seen in the valve onlays of different valve categories are in approximate concordance with findings in the respective^ valve proper. In addition to alterations in basic tissue composition, the presence or absence of blood vessels and inflammatory cells can be associated with important events during certain valvular disease processes (172> 357-359) j found vessels to be frequently prominent in both onlay and valve proper regions of carcinoid valves (Table 5) and, with the exception of the valve proper of rheumatic valves, were found with statistically greater frequency than in any of the other valve groups (onlay: C vs N , p=0.008; C vs R, p=0.002; C vs F, p=0.002; C vs A , p=0.03. proper: C 131 cu 1 < fa 0 a> us = © • aa 3 c o i OH cd (-1 u > o cu ~ c3 si 1 § rS © H S O a S o "S 0 s § .2 3 8 1 g § •§ CJ CO cu i CO CO 'H-J C M o d o I CO O , 3 U .2 o •3 0) s co g o co — . OH " © 8 & .2 1 -9 .2 -fa '— ' 2 5 co -TH CUH . o co 11 CJ — CO CO 0 CD > - U "3 § & o •3 > < > % o U * 3 H-» <2 "S o ' a 9 8 1 u I s § 5 -d CO > H-H O §11 111 S ? (j 8 * o 2 u £ c o O 1 8 cd 8 S O CJ > . CO X ? CD H2 ?f> > fa d 5 CD . 8* 8 & o U O „ > >^  60 *3\ cd CO 3 O C cu e3 60 -S > a "o u 3 ° rH CO ^ 53 * "S > d 3 S ** co cD 2 11 P h i T 3 I n - CD I & CO o CO > CD X co cO *a > C 4 _ CD O f= .J2 "© § c3 co 5=5 £3 O CB U CO CD w 3 co CO co cO _J <H_ '•— 1 ^H fa 8 2 d .2 3^ fa-» II •S CD CO 43 ^ a (U - - H .d co H g . d I c2 co "© CD w & § 8 -8 CO g co O OH 1-H 3 a M I i 3 x > <D CO CU > o . co CO CD °r > p s •d ± d > b c © 8 -2 sS § L8 •© CD CO O QH X tD S T3 CD J 3 Ccu 60 4 3 CD •d S CD 2 % u O CD U X5 (-1 H-> O H p O. CO CO 132 Table 5 Adjusted means from mixed effect model analyses for leukocytes and vessels Valve Group Leukocytes* Vessels* Onlay Proper Onlay Proper Normal 7.9 5 0 0 Rheumatic 28.7 4 0 5 Floppy 21.1 3 0 0 Anorexigen 88.2 4 3.8 1.5 Carcinoid 21.5 16.8 15 12.5 2 Data are expressed as number of cells or vessels per area of tissue (mm ). 133 vs N , p=0.01; C vs F, p=0.02; C vs A , p=0.01). Importantly, the geographical distribution of blood vessels and inflammatory cells was also distinctive for certain valve categories (Figure 35). The vessels of rheumatic valves were located mostly in the spongiosa of the valve proper (Figure 35 c), while vessels with surrounding inflammatory cells were more widespread and superficial in both the valve proper and onlays of carcinoid (Figure 35, d-f) and anorexigen-exposed valves (Figure 35, a and b). Floppy and normal valves rarely had vessels in either the onlays or valve proper. Although leukocytes were present to varying degrees in the valve onlays of all comparison groups, they were typically more prominent in the carcinoid and anorexigen-exposed valves in both the valve proper and onlay regions (Table 5; Figure 35, a and d). Substantial numbers of inflammatory cells were also evident in the onlays of rheumatic valves (Table 5). The variability in number for each case prevented statistical significance from being reached. Overall, the anorexigen-exposed valves exhibited the greatest range of any of the comparison groups for inflammatory cells and vessels in the valve proper or the valve onlays (Table 5). 6.2.3 Discriminant Analysis Discriminant analysis uses linear combinations of independent variables, termed linear discriminant functions, to classify cases into one of the comparator groups. The values of these estimated functions for each valve are the discriminant scores and these scores can be plotted to describe the separation between groups. Discriminant analysis was carried out first for the three valve diagnoses, normal, floppy, and rheumatic. The geometric and compositional data sets were combined for the anterior mitral valve leaflets as illustrated in Figure 36. Effective discrimination among the three groups was obtained using the variables identified in the discriminant function. 134 V 0> - 4 • * * S i ' *r 3 s O T 3 - H o S 3 £ 2 2 « 'oo S Q .SP £ > > | .2 o CO 135 a: a: on LL LL a: oc or LU-DC * 1 , ^ - | 1 1 1 1 r e z i o L-h o CN i m <u 1 £ ~ o CD co cu CO o ca > v CH -3 <S • — co ro 3 £ E w ca .2 "3 la *o •e JS ^ 3 II co C ca (U u _> ca > E * 1-ca <u d u CU X O 00 'co T3 <u ca q CO ca c ca 3 C I .22 ? •g o CD OO a . 2 CD o co = 3 ^ 60 .2 ca T3 T3 xi HH CD co i O CD ° "2 i i 136 Discriminant analyses were then carried out separately for the geometric data and compositional data for the mitral and aortic valves. Floppy mitral valves could not be distinguished using the discriminant analysis that included normal, rheumatic, floppy, and anorexigen-exposed valves. Thus, the results for floppy valves based on the latter analysis are not presented. Discriminant analysis based on geometry of normal, rheumatic, and anorexigen-exposed anterior mitral valve leaflets (Figure 37a) included six variables. These variables were, in order of importance, onlay area, percent onlay area, base-mid 3 thickness, mid-tip 1 thickness, mid-tip 2 thickness, and the tip thickness. Variables for the aortic valves, in order of importance, included percent onlay area, mid-tip 3 valve proper thickness, base-mid 2 valve proper thickness, onlay area, mid valve thickness, mid-tip 1 thickness, and onlay number (Figure 38a). Thus, for both the mitral and aortic valve, the most important geometric features included onlay size and valve thickness toward the free margin. Using this statistical model, separation of the valves into three distinct groups was achieved, particularly for the aortic valve. In a similar fashion, discriminant analysis for color composition included several discriminating variables for each valve (Figure 37b and Figure 38b). In the case of the mitral valve, these variables were percent GAGs in the valve proper, percent GAGs in the valve onlays, percent elastin in the valve onlays, and percent muscle-like cells in the valve proper. The discriminant aortic valve variables were only slightly different and included the percent GAGs in the valve proper, the percent elastin in the valve proper, the percent muscle-like cells in the valve onlays, and the percent collagen in the valve proper. Again, reasonable separation of the valves into three groups was achieved, especially for the aortic valve. Of particular interest, the percent of GAGs present in the valve proper was a prominent distinguishing feature for both the aortic and mitral valves. 137 z z Z L 2 Z Z z < < z < < i 1 1 1 1 1 r-e z \- o i- z- e-< • < r t tr* E tr " *z z * z • % z * z trtr —1 1 1 1 1 1 -£ 2 1 0 1 - 3 -• < < i i r Z I 0 I- Z- £-.2 ^ « Q .2 p4 CO T3 co • "S 2 tu c s C3 Q 138 a z 2£E Z _ Z zz 2-< < Z Z - 1 — z-—I— z z • % z < • < z-<U * [fl « l a •=! cd U P 80 . S * i i d 8 53 00 o •S d ,. 3 I * 5? O H o a> x i a cu <u OH — ' X> 52 cd Cd •{-; 2 | O "d <u d t» Sb 03 « OH ^ cu • H w S co -d o d s l •d .2 .2 "° *HS 13 d O (U co ed > a < C O ^ ~ cu "d <U C O O OH X CU I co d CU 1 £ SP 5 '£ > <u U l o o € I cd "d 0 ~» _ I 1 .23 (± C O C J HU d cd cd g d <u cd j s d "5 1 c 5 I 2, CJ to 50 CU co t i 3 cd • cd X> X ! W CJ CU cj > 'B it CU £ -2 o 1 CU O 00 cd C+H cu p M co O >> « "cd *d d ju <! | • c2 . 2 « co CU O H C J _> cd > cj o cd CJ X3 cS T3 CJ co r l <U .22 O 5 -2 CO O I O CJ — O CU OH CO cd S3 CJ Ed C J i CO CO ^ C O H? OH OH 3 co H d ® 1 cu 5 > 00 .d 'co 3 -d cu X! co o CJ d CHH HB O J -•2 .22 5 S 2 -d «H rt d cu o I I • a & CU co C O cu X> cd -d cu d «44 O co CJ CJ . 2 00 d <u o 3 CJ oxi d co X ) U cd > -q I d 3 > > O H 00 O H O O -d. « 3 2 » cd > OH O w % "— o 00 cd o OH I -s o -d O CU 11 •m cu "3 O co Q, CU o cj — "o cj c+-O d •c cd > 00 d 1 _d 1 i 139 Discriminant analysis was also performed for the subset of valves that had both color composition based on the Movat's pentachrome staining and leukocyte and vessel counts (Figure 37c and Figure 38c). The two most discriminating features for the mitral valve were the percent area of GAGs and the number of vessels in the valve proper, while the three discriminant features for the aortic valve were the percent GAGs in the valve proper and the percent muscle-like cells and number of vessels in the valve onlays. Separation of the valves was even more distinct than when color composition alone was used. 6.2.4 Versican Expression by Cardiac Valve Myofibroblasts The data on valve composition demonstrating the GAG-rich nature of anorexigen-exposed, carcinoid and floppy valves suggested the presence of excessive amounts of proteoglycans such as versican. Indeed, preliminary observations of diseased cardiac valve tissue stained immunohistochemically for specific interstitial proteoglycans demonstrated spatial association between GAGs and versican (Figure 39). Further, examination of cultured human cardiac valve MFBs stained immunofluorescently for the versican core protein demonstrated that MFBs do indeed produce this proteoglycan (Figure 40a). The distribution of staining in the MFBs was similar to that observed in rat aortic SMCs. Versican staining was generally confined to the cytoplasm and was distributed in a diffuse manner with regions of increased intensity. Double immunolabelling for versican and a-SM actin revealed that, in larger stellate cells, versican was present in areas adjacent to the nucleus, but was absent from more distal cytoplasmic regions (Figure 40b). In contrast, elongated cells expressed versican throughout the cell cytoplasm. Taken together, these results suggest that cardiac valve MFBs may be responsible for the excessive production and accumulation of versican in myxomatous valve lesions. 140 Figure 39. Spatial association between GAGs and versican in cardiac valve disease. A representative histologic section of the anterior mitral valve leaflet from a 47-year-old female exposed to anorexigens is shown. Adjacent sections of valve tissue were stained histochemically with Movat's pentachrome (a) and immunohistochemically for versican (b). The elastic membrane separates the valve onlay from the valve proper. Scale bar = 100 um. 141 Figure 40. Expression of versican by cultured human cardiac valve MFBs. Cells were fixed and stained immunocytochemically for versican (a) or versican and a-SM actin (b). Versican is shown in green and a-SM actin is shown in red. Nuclei were counterstained with H33342 (blue). Images were captured on a epifluorescence microscope equipped with a digital camera and three dimensional deconvolution software. 142 To determine whether growth factors influence the expression of versican by cardiac valve MFBs, quiescent cells were cultured in the absence or presence of 10 ng/ml PDGF-BB and stained immunocytochemically for versican. In comparison tb untreated cultures, cells exposed to PDGF-BB for 24 hours showed increased versican expression (Figure 41). The staining intensity was particularly prominent in elongated cells, as noted above. Thus, at least in vitro, growth factors such as PDGF-BB upregulate versican core protein expression by valve MFBs. 6.3 Discussion In the first report on the echocardiographic and pathologic findings of heart valves with fen/phen exposure in obese patients O1), echocardiography revealed the presence of both right-sided and left-sided abnormalities in 24 female patients. The issue of cardiac valvular dysfunction in association with anorexigen exposure has since been the subject of several clinical investigations (333> 340> 346> 3 4 9 X Many of these studies have relied heavily upon echocardiography (340, 344, 351, 352) ^  a technique that detects primarily physiological events and not necessarily fine structural features that are potentially pathologically important. In contrast, few reports exist describing the gross or histologic pathology of such functionally regurgitant valves. Right- and left-sided valves removed surgically from three patients in the original study by Connolly and colleagues ( u ) revealed histopathologic features said to be similar to those seen in ergotamine-induced or carcinoid heart disease, namely thickened, glistening leaflets as well as an appearance of shortened chordae tendineae, with matrix-rich expansion of the valve tissue. "Myxomatous" onlay formation on the valves of patients treated with fen/phen or dexfen was interpreted as a prominent histopathologic finding in these patients. Only two full-length reports have provided any details about the valve lesions associated with anorexigen use. Based on assessment of a small number of surgically excised valves (two mitral and one aortic), the first of these studies described focal thickening associated with increased valvular opacity as major 143 Figure 41. Stimulation of valve MFBs upregulates versican expression. Digital images of human cardiac valve MFBs cultured in the absence (a) or presence (b) of 10 ng/ml PDGF-BB are shown. The cells were exposed to the growth factor for 24 hours. Positive staining for versican is indicated by the red fluorescent signal. Cell nuclei were counterstained with H33342 (blue). Scale bar = 100 um. 144 gross features ( 1 6 1 X Microscopically, superficial "fibromyxoid" tissue was observed, as was the presence of CD3-positive leukocytes. The second, larger study described a "typical plaque" as containing a myxoid stroma, proliferative myofibroblastic cells and often small vessels and lymphocytic accumulations ( 3 5 6 \ This study also highlighted dramatic interpatient variability in the severity of lesions associated with anorexigen use. While these descriptive investigations have acknowledged the presence of myxomatous and inflammatory lesions exhibiting histopathologic features similar to other valve pathologies, they have not made direct, quantitative comparisons between valve lesions associated with anorexigen exposure and normal or diseased valves. The current study was undertaken in an effort to establish the presence of distinctive microscopic features that separate human heart valves exposed to anorexigens from normal valves and valves exhibiting rheumatic, floppy and carcinoid lesions. It is the first attempt at discrimination of comparison valve groups on the basis of fine compositional and geometric characteristics at the histopathological level. It is important to note that several changes occur during the natural aging process of heart valves (360-362)^ making the separation of lesions brought on by normal aging from those associated with anorexigen use necessary. My recent quantitative assessment of normal valves has demonstrated that they are typically thin, with very few but definite small onlays ( 2 4 4 X With increasing age, normal valves become thicker, more onlays appear and the percentage of total valve area attributable to onlays becomes greater ( 2 4 4). Normal valves are composed of roughly equal amounts of collagens and GAGs, and rarely harbour leukocytes and vessels, even as they age. These structural changes affect the entire valve and are most likely the result of hemodynamic, rheologic, and mechanical factors operative in the valve leaflets and cusps over time. The anorexigen-exposed valves exhibited great variability, ranging from near normal in appearance to markedly abnormal. Still, I found several quantitative features that effectively 145 separated such valves from their normal counterparts. Specifically, valves taken from individuals exposed to anorexigens exhibited a greater percent onlay area, more onlays per valve, a greater total valve thickness, a higher percentage of GAGs and a lower percentage of collagen. Numerous vessels and variable numbers of leukocytes were distributed within the valve proper and valve onlays of anorexigen-exposed valves, while these entities were rarely present in normal valves. Thus, valve thickness, valve onlay number, G A G content, vascularity and inflammation serve to separate normal and aging valves from those exposed to anorexigens. Similar features were also important in distinguishing anorexigen-exposed valves from rheumatic valves. Previous observational studies have described leaflets or cusps of valves affected by rheumatic disease as fibrotic and collagenous with associated neovessels in the valve core and inflammatory cell infiltrates consisting primarily of lymphocytes and macrophages (164> 3 6 3 X In the present study, rheumatic heart valves were found to be the thickest of the valve groups analyzed, a characteristic that was particularly evident toward the valve tip. These valves had the greatest cross-sectional valve area, onlay area, percent onlay area and onlay size, but yielded the smallest number of onlays per valve. The composition of the valve proper and onlays was similar, consisting primarily of collagen, a few GAGs and some SMCs and MFBs. The collagen-rich nature of these valves was not unexpected, given that collagen is the characteristic matrix component associated with late healing events and valvular lesions seen in rheumatic disease by the time of valve excision would typically be considered well-healed. These valves also harbored numerous vessels in the valve proper. In comparison to rheumatic valves, valves associated with anorexigen exposure had smaller onlay areas, percent onlay areas and average onlay sizes, greater numbers of onlays, more GAGs and less collagen. Furthermore, vessels and leukocytes, while present in both valve categories, were differentially distributed. Vessels in rheumatic valves were numerous, but were 146 limited to the spongiosa in the valve core, while vascular structures in anorexigen-exposed valves were superficial and widespread. The greater onlay number, the smaller onlay size and the GAG-rich nature of valves associated with anorexigen exposure as compared to rheumatic valves may reflect lesion age, with the small, numerous, provisional matrix-rich lesions in anorexigen-exposed valves representing young, succulent lesions. Alternatively, these differences may signal the presence of a distinct pathogenetic process in anorexigen-exposed valves. Published qualitative gross and microscopic features of floppy mitral valve leaflets include lengthening of leaflets and chordae, interchordal hooding, increased GAGs (myxomatous transformation) and thickening of the spongiosa ( 1 6 7 X In the present study, the thickness, average onlay area and percent onlay area of floppy valves were intermediate between that seen in normal and rheumatic valves, while the number of onlays per valve was greater in floppy as compared to rheumatic valves. In contrast to the collagen-rich rheumatic valves, floppy valves were composed of roughly equal amounts of collagen and GAGs. This tissue composition was reflected in both the valve proper and the valve onlays, although a trend toward increased GAGs in the valve proper was noted. Distinctive microscopic features separating anorexigen-exposed valves from floppy valves were limited to valve composition and vascularity, as none of the geometric features outlined in this study were significantly different between the two valve groups. Valves exposed to anorexigens were more GAG-rich than floppy valves, both in the valve onlay and the valve proper. In addition, while blood vessels were generally absent from floppy valves, they were a prominent feature of anorexigen-exposed valves. Thus, while both floppy mitral valves and anorexigen-exposed valves are considered to have undergone myxomatous degeneration, the greater proportion of GAGs and the presence of vascular structures in anorexigen-exposed valves aids in separating these two valve pathologies. Furthermore, the differences suggest that the pathobiological mechanisms involved in the 147 generation of valve lesions associated with anorexigen use may be, in part, different from those operant in floppy mitral valves. Valvular lesions associated with carcinoid disease have been described as glistening, superficial "plaque-like" thickenings composed of GAG-rich tissue on the surface of valvular leaflets and cusps (171-173) s u c n lesions have been reported to be similar to the myxomatous valve lesions associated with anorexigen use. In my study, carcinoid valves were evaluated based on composition alone. The valve proper of the carcinoid valves contained approximately equal proportions of GAGs and collagen, while the onlays of these valves were significantly GAG-rich. Both the valve proper and onlays of the carcinoid valves had a large number of leukocytes and vessels per square millimeter of tissue area. In comparison to valves affected by carcinoid disease, valves associated with anorexigen exposure exhibited even greater G A G deposition in the valve proper, while significantly fewer vessels were present. Such features indicate the presence of a particularly myxomatous process within the valve itself and further suggest the existence of a distinctive pathologic process in valves associated with anorexigen exposure. Recognizing that a distinctive pathological process may be involved in the lesions of anorexigen-exposed valves, I sought to validate our data demonstrating differences in geometry and composition among valve groups using an independent statistical approach. Discriminant function analysis provided a statistical tool that enabled us to separate the valves into distinguishable groups. Discriminant function analyses were prepared for the aortic and mitral valve separately, using the quantitative measures of tissue geometry and composition discussed above. The presence of onlays on the valves and the size of such onlays represented important geometric features that could be used for discrimination. GAGs as a percent of tissue area was the most important discriminating factor for analyses using either color composition alone or 148 color composition in concert with leukocyte and vessel counts. The presence of vessels also had a prominent influence on discriminating valve groups. It is interesting that, at least based on preliminary observations, there is spatial association between GAGs and versican in myxomatous valve lesions. Moreover, MFBs isolated from human cardiac valves synthesize the versican core protein, the expression of which can be upregulated by growth factors such as PDGF-BB. Given the important role played by MFBs in pathologic matrix remodeling and the probability that myxomatous valve lesions represent tissue microenvironments rich in growth factors and other mediators, it is tempting to speculate that growth factor-mediated upregulation of versican by valve MFBs is a major mechanism involved in the formation of such GAG-rich lesions. Experiments aimed at testing this hypothesis remains an important focus of ongoing research. The majority of mitral valves from patients exposed to anorexigens were distinguishable from the other valve groups based on geometry and composition using discriminant function analysis. Two anorexigen-exposed mitral valves were, however, visibly different in histologic makeup when compared to all other anorexigen-exposed valves. In fact, these two valves could not be distinguished from rheumatic valve disease on the Movat's pentachrome stain when discriminated using color composition alone or color composition in combination with vessels and leukocytes. The inability to separate these two valves may be related to the presence of pre-existent valve disease, although constraints on the acquisition of clinical information necessarily requires that such a conclusion remains speculative. These two valves also highlight the diagnostic dilemma involved in separating some valves exposed to anorexigens from valves with post-rheumatic scarring and demonstrate the importance of defining distinctive, quantitative features that differentiate anorexigen-associated valvulopathy from other valve pathologies. In conclusion, I have demonstrated distinctive microscopic features that separate human heart valve lesions associated with anorexigen exposure from normal valves and valve lesions 149 associated with other pathologies. The number of valve onlays, their size, the degree of G A G deposition and the presence and location of vessels and leukocytes were important features distinguishing anorexigen-exposed valves from normal and rheumatic valves, and enabled their separation by discriminant function analysis. While discriminant function analysis was unable to clearly separate anorexigen-exposed valves from floppy mitral valves in this study, greater deposition of GAGs and the presence of vessels and leukocytes in the former valve group aid in distinguishing between these myxomatous pathologies. Further, while certain of these features are reminiscent of those seen in carcinoid valves, the phenotype observed in the anorexigen-exposed valves available for analysis was less dramatic and more variable from the point of view of average onlay size and insofar as the extent of inflammation and vascularity. The presence of a myxomatous process in both the valve core and the onlay suggests the presence of pathogenetic mechanisms distinct from those involved in floppy mitral valve disease and carcinoid disease. The range of features found in the anorexigen-exposed valves no doubt reflects the duration of exposure, the drug dosages, and individual patient susceptibility factors, including local standards and patterns of care and the influence of pre-existent valve disease. Unfortunately, because of constraint on the acquisition of clinical information, ascertainment of the number of patients with pre-anorexigen valvular dysfunction could not be pursued. Finally, the provisional matrix-rich, inflamed and neovascularized nature of the onlays and of the valve proper of anorexigen-exposed valves suggest the opportunity for regressive structural and functional changes following discontinuation of the drugs (345> 346> 3 5 3 \ 150 7 CHAPTER VII CONCLUSIONS AND FUTURE PROSPECTS The data that I have presented in this thesis provide the groundwork for future experiments aimed at better understanding the regulation of versican expression by vascular SMCs and valvular MFBs under physiologic and pathologic conditions. I have shown that EGF and IGF-I, alone and in combination, upregulate versican mRNA levels and core protein expression by vascular SMCs. I have demonstrated that lipoproteins accumulating in the vessel wall of coronary arteries from human heart allografts are oxidatively modified and are associated geographically with intimal SMCs as well as versican. Further, I have provided evidence that human cardiac valves excised from patients exposed to aorexigens exhibit a GAG-rich and versican-rich lesion distinct from lesions that characterize other myxomatous valve diseases. Finally, I have established a model system of cultured human mitral valve MFBs and demonstrated that such cells represent a useful tool in which to further explore the regulation of versican. The data presented here showing the upregulation of versican expression by vascular SMCs in response to a combination of EGF and IGF-I suggest that growth factors synthesized by injured endothelial cells may act in concert to modulate versican synthesis during atherogenesis. Although such regulation is presumed to occur through receptor-mediated mechanisms, the specific signal transduction pathways involved in the regulation of versican by SMCs remain to be elucidated. Data from other laboratories implicate the M A P K cascade as one component in this regulatory process, but EGF and IGF-I are capable of activating multiple pathways upon binding their respective receptors and it is probable that cross-talk among a variety of pathways is required for appropriate signaling to occur. Experiments using specific inhibitors designed to block these signals at the level of the receptors as well as at downstream locations should help 151 clarify this issue. Strategic, directed approaches utilizing genomic and proteomic technologies may also be valuable in identifying multiple activated signaling pathways as well as the potential for cross-talk among such pathways. The targeted regulation of gene expression by the M A P K pathways requires the translocation of ERK1/2 from the cytoplasm to the nucleus and its subsequent interaction with the cell's transcriptional machinery at the site of gene promoter. The versican gene contains a promoter upstream of its transcription start site that houses a number of putative binding sites for transcription factors (35). Information regarding the regulation of the versican promoter, both at a basal level and in response to stimulation, remains extremely limited. Further, there is currently no published evidence detailing the actual binding of transcription factors to the promoter under any condition. It is very likely that the regulation of versican expression by mesenchymal cells occurs at the level of the gene-specific promoter. Experiments aimed at elucidating the specific cis-acting sequences and transcription factors involved in the modulation of versican transcriptional regulation by vascular SMCs and mitral valve MFBs, particularly in response to modulation induced by mediators such as EGF and IGF-I, represent an important investigative direction. Such experiments will require molecular tools such as transfection-based promoter expression systems and specific deletional promoter constructs, both of which are currently established in our laboratory. The observed association between oxLDL and versican in proximity to intimal SMCs in human heart allografts is intriguing and lays the foundation for further investigations related to the molecular structure of versican in these lesions as well as its regulation by vascular SMCs. Future experiments will need to address the quality of the E C M present in these lesions, particularly the specific splice variants of versican expressed by intimal SMCs. As alternative splicing occurs in the regions of versican housing the G A G chain attachment sites, versican isoforms expressing both G A G attachment domains would, in theory, have an increased capacity 152 for binding and retaining lipids in the vessel wall. Strategies for gathering such data may include the use of primers, probes and antibodies specific for each of the known versican isoforms. Laser capture microdissection of SMC populations from within lesions of human and animal T V D coupled with real-time RT-PCR technology for specific versican isoforms should provide quantitative information on the quality of the matrix produced. The ability of aldehyde-modified lipoproteins to modulate versican expression by vascular SMCs should also be explored. The knowledge that H N E itself can activate the EGF receptor on vascular SMCs, together with my data showing the upregulation of versican by EGF, raises the possibility that HNE-modified lipoproteins may modulate versican expression through receptor-mediated mechanisms. To date the specific lipid moieties oxidatively modified subsequent to interactions with versican and the capacity of these moieties to regulate the expression of versican remain unknown. 153 8 R E F E R E N C E S 1. Wight TN. The vascular extracellular matrix. In: Atherosclerosis and Coronary Artery Disease, edited by V . Fuster, R. Ross and E. J. Topol. Philadelphia: Lippincott-Raven, 1996, p. 421-440. 2. Galis ZS, Khatri JJ: Matrix metalloproteinases in vascular remodeling and atherogenesis: the good, the bad, and the ugly. Circ Res 2002, 90:251-262. 3. Tomasek JJ, Gabbiani G, Hinz B, Chaponnier C, Brown RA: Myofibroblasts and mechano-regulation of connective tissue remodelling. Nat Rev Mol Cell Biol 2002, 3:349-363. 4. Burkitt HG, Young B, Heath JW. Circulatory system. In: Wheater's Functional Histology: a Text and Colour Atlas (Third ed.). London: Churchill Livingstone, 1999, p. 143-145. 5. Zhdanov VS, Sternby N H , Drobkova IP, Galakhov IE: Hyperplasia of coronary intima in young males in relation to development of coronary heart disease in adults. Int J Cardiol 2000, 76:57-64. 6. Haerem JW: Cushion-like intimal lesions in intramyocardial arteries of man. Their relation to age, sex, coronary atherosclerosis, and certain diseases. Acta Pathol Microbiol Scand 1969, 77:598-608. 7. Davies MJ. The mitral valve. In: Pathology of Cardiac Valves. London: Butterworth & Co. Ltd., 1980, p. 62-69. 8. Gulizia JM, Cunningham MW, McManus B M : Immunoreactivity of anti-streptococcal monoclonal antibodies to human heart valves. Evidence for multiple cross-reactive epitopes. A m J Pathol 1991, 138:285-301. 154 9. Hauck A J , Edwards WD, Danielson GK, Mullany CJ, Bresnahan DR: Mitral and aortic valve disease associated with ergotamine therapy for migraine. Report of two cases and review of literature. Arch Pathol Lab Med 1990,114:62-64. 10. Hendrikx M , Van Dorpe J, Flameng W, Daenen W: Aortic and mitral valve disease induced by ergotamine therapy for migraine: a case report and review of the literature. J Heart Valve Dis 1996, 5:235-237. 11. Connolly H M , Crary JL, McGoon M D , Hensrud DD, Edwards BS, Edwards WD, Schaff HV: Valvular heart disease associated with fenfluramine-phentermine. N Engl J Med 1997,337:581-588. 12. Burkitt HG, Stevens A , Lowe JS, Young B. Acute inflammation, healing and repair. In: Wheater's Basic Histopathology: a Colour Text and Atlas (Third ed.). London: Churchill Livingstone, 1996, p. 16-20. 13. Lorena D, Uchio K, Costa A M , Desmouliere A : Normal scarring: importance of myofibroblasts. Wound Repair Regen 2002, 10:86-92. 14. Ross R: Atherosclerosis—an inflammatory disease. N Engl J Med 1999, 340:115-126. 15. Ross R, Glomset JA: The pathogenesis of atherosclerosis (second of two parts). N Engl J Med 1976, 295:420-425. 16. Ross R, Glomset JA: The pathogenesis of atherosclerosis (first of two parts). N Engl J Med 1976, 295:369-377. 17. Libby P: Changing concepts of atherogenesis. J Intern Med 2000, 247:349-358. 18. Lin H, Wilson JE, Roberts CR, Horley KJ , Winters GL, Costanzo MR, McManus B M : Biglycan, decorin, and versican protein expression patterns in coronary arteriopathy of human cardiac allografts: distinctness as compared to native atherosclerosis. J Heart Lung Transplant 1996, 15:1233-1247. 155 19. Lin H , Ignatescu M , Wilson JE, Roberts CR, Horley KJ , Winters GL , Costanzo MR, McManus B M : Prominence of apolipoproteins B, (a), and E in the intimae of coronary arteries in transplanted human hearts: geographic relationship to vessel wall proteoglycans. J Heart Lung Transplant 1996,15:1223-1232. 20. Zimmermann DR. Versican. In: Proteoglycans: Structure, Biology and Molecular Interactions, edited by R. V . Iozzo. New York: Marcel Dekker, 2000, p. 327-341. 21. Iozzo RV: Matrix proteoglycans: from molecular design to cellular function. Annu Rev Biochem 1998, 67:609-652. 22. Zimmermann DR, Ruoslahti E: Multiple domains of the large fibroblast proteoglycan, versican. Embo J 1989, 8:2975-2981. 23. Dours-Zimmermann MT, Zimmermann DR: A novel glycosaminoglycan attachment domain identified in two alternative splice variants of human versican. J Biol Chem 1994, 269:32992-32998. 24. Kimata K, Oike Y , Tani K, Shinomura T, Yamagata M , Uritani M , Suzuki S: A large chondroitin sulfate proteoglycan (PG-M) synthesized before chondrogenesis in the limb bud of chick embryo. J Biol Chem 1986, 261:13517-13525. 25. Schonherr E, Jarvelainen HT, Sandell LJ, Wight TN: Effects of platelet-derived growth factor and transforming growth factor-beta 1 on the synthesis of a large versican-like chondroitin sulfate proteoglycan by arterial smooth muscle cells. J Biol Chem 1991, 266:17640-17647. 26. Zako M , Shinomura T, Ujita M , Ito K, Kimata K: Expression of PG-M(V3), an alternatively spliced form of P G - M without a chondroitin sulfate attachment in region in mouse and human tissues. J Biol Chem 1995, 270:3914-3918. 156 27. Ito K, Shinomura T, Zako M , Ujita M , Kimata K: Multiple forms of mouse PG-M, a large chondroitin sulfate proteoglycan generated by alternative splicing. J Biol Chem 1995,270:958-965. 28. Schmalfeldt M , Dours-Zimmermann MT, Winterhalter K H , Zimmermann DR: Versican V2 is a major extracellular matrix component of the mature bovine brain. J Biol Chem 1998,273:15758-15764. 29. Niederost BP, Zimmermann DR, Schwab M E , Bandtlow CE: Bovine CNS myelin contains neurite growth-inhibitory activity associated with chondroitin sulfate proteoglycans. J Neurosci 1999, 19:8979-8989. 30. Landolt R M , Vaughan L, Winterhalter K H , Zimmermann DR: Versican is selectively expressed in embryonic tissues that act as barriers to neural crest cell migration and axon outgrowth. Development 1995, 121:2303-2312. 31. Paulus W, Baur I, Dours-Zimmermann MT, Zimmermann DR: Differential expression of versican isoforms in brain tumors. J Neuropathol Exp Neurol 1996, 55:528-533. 32. Lemire JM, Braun KR, Maurel P, Kaplan ED, Schwartz S M , Wight TN: Versican/PG-M isoforms in vascular smooth muscle cells. Arterioscler Thromb Vase Biol 1999, 19:1630-1639. 33. Lemire JM, Merrilees MJ , Braun KR, Wight TN: Overexpression of the V3 variant of versican alters arterial smooth muscle cell adhesion, migration, and proliferation in vitro. J Cell Physiol 2002,190:38-45. 34. lozzo RV, Naso MF, Cannizzaro L A , Wasmuth JJ, McPherson JD: Mapping of the versican proteoglycan gene (CSPG2) to the long arm of human chromosome 5 (5ql2-5ql4). Genomics 1992,14:845-851. 157 35. Naso MF, Zimmermann DR, Iozzo R V : Characterization of the complete genomic structure of the human versican gene and functional analysis of its promoter. J Biol Chem 1994, 269:32999-33008. 36. Naso MF, Morgan JL, Buchberg A M , Siracusa LD, Iozzo R V : Expression pattern and mapping of the murine versican gene (Cspg2) to chromosome 13. Genomics 1995, 29:297-300. 37. Shinomura T, Zako M , Ito K, Ujita M , Kimata K: The gene structure and organization of mouse P G - M , a large chondroitin sulfate proteoglycan. Genomic background for the generation of multiple P G - M transcripts. J Biol Chem 1995, 270:10328-10333. 38. Shinomura T, Jensen K L , Yamagata M , Kimata K, Solursh M : The distribution of mesenchyme proteoglycan (PG-M) during wing bud outgrowth. Anat Embryol 1990, 181:227-233. 39. Yamagata M , Shinomura T, Kimata K: Tissue variation of two large chondroitin sulfate proteoglycans (PG-M/versican and PG-H/aggrecan) in chick embryos. Anat Embryol 1993, 187:433-444. 40. Zako M , Shinomura T, Miyaishi O, Iwaki M , Kimata K: Transient expression of PG-M/versican, a large chondroitin sulfate proteoglycan in developing chicken retina. J Neurochem 1997, 69:2155-2161. 41. Yamagata M , Herman JP, Sanes JR: Lamina-specific expression of adhesion molecules in developing chick optic tectum. J Neurosci 1995,15:4556-4571. 42. Capehart A A , Mjaatvedt C H , Hoffman S, Krug EL: Dynamic expression of a native chondroitin sulfate epitope reveals microheterogeneity of extracellular matrix organization in the embryonic chick heart. Anat Rec 1999, 254:181-195. 158 43. Mjaatvedt C H , Yamamura H, Capehart A A , Turner D, Markwald RR: The Cspg2 gene, disrupted in the hdf mutant, is required for right cardiac chamber and endocardial cushion formation. Dev Biol 1998, 202:56-66. 44. Bode-Lesniewska B, Dours-Zimmermann MT, Odermatt BF, Briner J, Heitz PU, Zimmermann DR: Distribution of the large aggregating proteoglycan versican in adult human tissues, J Histochem Cytochem 1996, 44:303-312. 45. Jakob M , Demarteau O, Schafer D, Hintermann B , Dick W, Heberer M , Martin I: Specific growth factors during the expansion and redifferentiation of adult human articular chondrocytes enhance chondrogenesis and cartilaginous tissue formation in vitro. J Cell Biochem 2001, 81:368-377. 46. Martin I, Jakob M , Schafer D, Dick W, Spagnoli G, Heberer M : Quantitative analysis of gene expression in human articular cartilage from normal and osteoarthritic joints. Osteoarth Cartil 2001, 9:112-118. 47. Milz S, Schluter T, Putz R, Moriggl B, Ralphs JR, Benjamin M : Fibrocartilage in the transverse ligament of the human atlas. Spine 2001, 26:1765-1771. 48. Robbins JR, Evanko SP, Vogel K G : Mechanical loading and TGF-beta regulate proteoglycan synthesis in tendon. Arch Biochem Biophys 1997, 342:203-211. 49. Waggett A D , Ralphs JR, Kwan AP, Woodnutt D, Benjamin M : Characterization of collagens and proteoglycans at the insertion of the human Achilles tendon. Matrix Biol 1998,16:457-470. 50. Carrino D A , Sorrell J M , Caplan A l : Dynamic expression of proteoglycans during chicken skeletal muscle development and maturation. Poult Sci 1999, 78:769-777. 51. Zimmermann DR, Dours-Zimmermann MT, Schubert M , Bruckner-Tuderman L: Versican is expressed in the proliferating zone in the epidermis and in association with the elastic network of the dermis. J Cell Biol 1994,124:817-825. 52. du Cros DL, LeBaron RG, Couchman JR: Association of versican with dermal matrices and its potential role in hair follicle development and cycling. J Invest Dermatol 1995, 105:426-431. 53. Miosge N , Sasaki T, Chu M L , Herken R, Timpl R: Ultrastructural localization of microfibrillar fibulin-1 and fibulin-2 during heart development indicates a switch in molecular associations. Cell Mol Life Sci 1998, 54:606-613. 54. Aspberg A , Adam S, Kostka G, Timpl R, Heinegard D: Fibulin-1 is a ligand for the C-type lectin domains of aggrecan and versican. J Biol Chem 1999, 274:20444-20449. 55. Olin A l , Morgelin M , Sasaki T, Timpl R, Heinegard D, Aspberg A: The proteoglycans aggrecan and versican form networks with fibulin-2 through their lectin domain binding. J Biol Chem 2001, 276:1253-1261. 56. Isogai Z, Aspberg A , Keene DR, Ono RN, Reinhardt DP, Sakai L Y : Versican interacts with fibrillin-1 and links extracellular microfibrils to other connective tissue networks. J Biol Chem 2002, 277:4565-4572. 57. Schmalfeldt M , Bandtlow CE, Dours-Zimmermann MT, Winterhalter K H , Zimmermann DR: Brain derived versican V2 is a potent inhibitor of axonal growth. J Cell Sci 2000, 113:807-816. 58. Yamaguchi Y : Lecticans: organizers of the brain extracellular matrix. Cell Mol Life Sci 2000, 57:276-289. 59. Yao L Y , Moody C, Schonherr E, Wight TN, Sandell LJ : Identification of the proteoglycan versican in aorta and smooth muscle cells by D N A sequence analysis, in situ hybridization and immunohistochemistry. Matrix Biol 1994, 14:213-225. 60. Lemire JM, Potter-Perigo S, Hall K L , Wight TN, Schwartz SM: Distinct rat aortic smooth muscle cells differ in versican/PG-M expression. Arterioscler Thromb Vase Biol 1996,16:821-829. 160 61. Morita H , Takeuchi T, Suzuki S, Maeda K, Yamada K, Eguchi G, Kimata K: Aortic endothelial cells synthesize a large chondroitin sulphate proteoglycan capable of binding to hyaluronate. Biochem J 1990, 265:61-68. 62. Kinsella M G , Tsoi CK, Jarvelainen HT, Wight TN: Selective expression and processing of biglycan during migration of bovine aortic endothelial cells. The role of endogenous basic fibroblast growth factor. J Biol Chem 1997, 272:318-325. 63. LeBaron RG, Zimmermann DR, Ruoslahti E: Hyaluronate binding properties of versican. J Biol Chem 1992, 267:10003-10010. 64. Iwata M , Wight TN, Carlson SS: A brain extracellular matrix proteoglycan forms aggregates with hyaluronan. J Biol Chem 1993, 268:15061-15069. 65. Yamagata M , Saga S, Kato M , Bernfield M , Kimata K: Selective distributions of proteoglycans and their ligands in pericellular matrix of cultured fibroblasts. Implications for their roles in cell-substratum adhesion. J Cell Sci 1993, 106:55-65. 66. Bensadoun ES, Burke A K , Hogg JC, Roberts CR: Proteoglycans in granulomatous lung diseases. Eur Respir J 1997, 10:2731-2737. 67. Evanko SP, Raines EW, Ross R, Gold LI, Wight TN: Proteoglycan distribution in lesions of atherosclerosis depends on lesion severity, structural characteristics, and the proximity of platelet-derived growth factor and transforming growth factor-beta. A m J Pathol 1998, 152:533-546. 68. Evanko SP, Angello JC, Wight TN: Formation of hyaluronan- and versican-rich pericellular matrix is required for proliferation and migration of vascular smooth muscle cells. Arterioscler Thromb Vase Biol 1999, 19:1004-1013. 69. Kawashima H , L i Y F , Watanabe N , Hirose J, Hirose M , Miyasaka M : Identification and characterization of ligands for L-selectin in the kidney. I. Versican, a large chondroitin sulfate proteoglycan, is a ligand for L-selectin. Int Immunol 1999,11:393-405. 161 70. Kawashima H , Hirose M , Hirose J, Nagakubo D, Plaas A H , Miyasaka M : Binding of a large chondroitin sulfate/dermatan sulfate proteoglycan, versican, to L-selectin, P-selectin, and CD44. J Biol Chem 2000, 275:35448-35456. 71. Braunewell K H , Pesheva P, McCarthy JB, Furcht LT, Schmitz B , Schachner M : Functional involvement of sciatic nerve-derived versican- and decorin-like molecules and other chondroitin sulphate proteoglycans in ECM-mediated cell adhesion and neurite outgrowth. Eur J Neurosci 1995, 7:805-814. 72. Zou K, Muramatsu H, Ikematsu S, Sakuma S, Salama R H , Shinomura T, Kimata K , Muramatsu T: A heparin-binding growth factor, midkine, binds to a chondroitin sulfate proteoglycan, PG-M/versican. Eur J Biochem 2000, 267:4046-4053. 73. Hirose J, Kawashima H , Yoshie O, Tashiro K, Miyasaka M : Versican interacts with chemokines and modulates cellular responses. J Biol Chem 2001, 276:5228-5234. 74. Aspberg A , Binkert C, Ruoslahti E: The versican C-type lectin domain recognizes the adhesion protein tenascin-R. Proc Natl Acad Sci U S A 1995, 92:10590-10594. 75. Wu Y , Chen L, Zheng PS, Yang BB: Betal-integrin mediated glioma cell adhesion and free radical-induced apoptosis are regulated by binding to a C-terminal domain of PG-M/versican. J Biol Chem 2002, 277:12294-12301. 76. Tao Z, Smart FW, Figueroa JE, Glancy DL , Vijayagopal P: Elevated expression of proteoglycans in proliferating vascular smooth muscle cells. Atherosclerosis 1997, ' 135:171-179. 77. Lee RT, Yamamoto C, Feng Y , Potter-Perigo S, Briggs WH, Landschulz KT, Turi TG, Thompson JF, Libby P, Wight TN: Mechanical strain induces specific changes in the synthesis and organization of proteoglycans by vascular smooth muscle cells. J Biol Chem 2001,276:13847-13851. 162 78. Kahari V M , Larjava H, Uitto J: Differential regulation of extracellular matrix proteoglycan (PG) gene expression. Transforming growth factor-beta 1 up-regulates biglycan (PGI), and versican (large fibroblast PG) but down-regulates decorin (PGII) mRNA levels in human fibroblasts in culture. J Biol Chem 1991, 266:10608-10615. 79. Haase HR, Clarkson RW, Waters MJ , Bartold P M : Growth factor modulation of mitogenic responses and proteoglycan synthesis by human periodontal fibroblasts. J Cell Physiol 1998,174:353-361. 80. Hakkinen L, Westermarck J, Kahari V M , Larjava H : Human granulation-tissue fibroblasts show enhanced proteoglycan gene expression and altered response to TGF-beta 1. J Dent Res 1996, 75:1767-1778. 81. Emoto N , Onose H, Yamada H , Minami S, Tsushima T, Wakabayashi I: Growth factors increase pericellular proteoglycans independently of their mitogenic effects on A10 rat vascular smooth muscle cells. Int J Biochem Cell Biol 1998, 30:47-54. 82. Tan E M , Hoffren J, Rouda S, Greenbaum S, Fox JWt, Moore JH, Jr., Dodge GR: Decorin, versican, and biglycan gene expression by keloid and normal dermal fibroblasts: differential regulation by basic fibroblast growth factor. Exp Cell Res 1993, 209:200-207. 83. Shimizu-Hirota R, Sasamura H , Mifune M , Nakaya H, Kuroda M , Hayashi M , Saruta T: Regulation of vascular proteoglycan synthesis by angiotensin II type 1 and type 2 receptors. J A m Soc Nephrol 2001,12:2609-2615. 84. Tiedemann K , Malmstrom A , Westergren-Thorsson G: Cytokine regulation of proteoglycan production in fibroblasts: separate and synergistic effects. Matrix Biol 1997, 15:469-478. 85. Qwarnstrom EE, Jarvelainen HT, Kinsella M G , Ostberg CO, Sandell LJ , Page RC, Wight TN: Interleukin-1 beta regulation of fibroblast proteoglycan synthesis involves a decrease in versican steady-state mRNA levels. Biochem J 1993, 294:613-620. 86. Ostberg CO, Zhu P, Wight TN, Qwarnstrom EE: Fibronectin attachment is permissive for IL-1 mediated gene regulation. FEBS Lett 1995, 367:93-97. 87. Schonherr E, Kinsella M G , Wight TN: Genistein selectively inhibits platelet-derived growth factor-stimulated versican biosynthesis in monkey arterial smooth muscle cells. Arch Biochem Biophys 1997, 339:353-361. 88. Yamamoto M , Yamamoto K: Growth regulation in primary culture of rabbit arterial smooth muscle cells by platelet-derived growth factor, insulin-like growth factor-1, and epidermal growth factor. Exp Cell Res 1994, 212:62-68. 89. Pukac L, Huangpu J, Karnovsky MJ: Platelet-derived growth factor-BB, insulin-like growth factor-I, and phorbol ester activate different signaling pathways for stimulation of vascular smooth muscle cell migration. Exp Cell Res 1998, 242:548-560. 90. Duan C, Bauchat JR, Hsieh T: Phosphatidylinositol 3-kinase is required for insulin-like growth factor-I-induced vascular smooth muscle cell proliferation and migration. Circ Res 2000, 86:15-23. 91. Facchiano A, De Marchis F, Turchetti E, Facchiano F, Guglielmi M , Denaro A , Palumbo R, Scoccianti M , Capogrossi M C : The chemotactic and mitogenic effects of platelet-derived growth factor-BB on rat aorta smooth muscle cells are inhibited by basic fibroblast growth factor. J Cell Sci 2000,113:2855-2863. 92. Evanko SP, Johnson PY, Braun KR, Underhill CB, Dudhia J, Wight TN: Platelet-derived growth factor stimulates the formation of versican-hyaluronan aggregates and pericellular matrix expansion in arterial smooth muscle cells. Arch Biochem Biophys 2001, 394:29-38. 93. Zhang Y , Cao L, Yang B L , Yang B B : The G3 domain of versican enhances cell proliferation via epidermial growth factor-like motifs. J Biol Chem 1998, 273:21342-21351. 164 94. Yang B L , Zhang Y , Cao L, Yang BB: Cell adhesion and proliferation mediated through the G l domain of versican. J Cell Biochem 1999, 72:210-220. 95. Zhang Y , Cao L, Kiani C, Yang B L , Hu W, Yang BB: Promotion of chondrocyte proliferation by versican mediated by G l domain and EGF-like motifs. J Cell Biochem 1999, 73:445-457. 96. Wu Y , Zhang Y , Cao L, Chen L, Lee V , Zheng PS, Kiani C, Adams M E , Ang LC, Paiwand F, Yang B B : Identification of the motif in versican G3 domain that plays a dominant-negative effect on astrocytoma cell proliferation through inhibiting versican secretion and binding. J Biol Chem 2001, 276:14178-14186. 97. Ang LC, Zhang Y , Cao L, Yang B L , Young B, Kiani C, Lee V , Allan K, Yang B B : Versican enhances locomotion of astrocytoma cells and reduces cell adhesion through its G l domain. J Neuropathol Exp Neurol 1999, 58:597-605. 98. Zhang Y , Cao L, Kiani CG, Yang B L , Yang B B : The G3 domain of versican inhibits mesenchymal chondrogenesis via the epidermal growth factor-like motifs. J Biol Chem 1998,273:33054-33063. 99. Williams KJ , Tabas I: The response-to-retention hypothesis of early atherogenesis. Arterioscler Thromb Vase Biol 1995,15:551-561. 100. Williams K J , Tabas I: The response-to-retention hypothesis of atherogenesis reinforced. Curr Opin Lipidol 1998, 9:471-474. 101. Boren J, Gustafsson M , Skalen K, Flood C, Innerarity TL: Role of extracellular retention of low density lipoproteins in atherosclerosis. Curr Opin Lipidol 2000,11:451-456. 102. Chait A , Wight TN: Interaction of native and modified low-density lipoproteins with extracellular matrix. Curr Opin Lipidol 2000,11:457-463. 103. Pentikainen M O , Oorni K, Lassila R, Kovanen PT: The proteoglycan decorin links low density lipoproteins with collagen type I. J Biol Chem 1997, 272:7633-7638. 104. Kovanen PT, Pentikainen MO: Decorin links low-density lipoproteins (LDL) to collagen: a novel mechanism for retention of L D L in the atherosclerotic plaque. Trends Cardiovasc Med 1999,9:86-91. 105. Camejo G, Hurt-Camejo E, Wiklund O, Bondjers G: Association of apo B lipoproteins with arterial proteoglycans: pathological significance and molecular basis. Atherosclerosis 1998, 139:205-222. 106. Boren J, Olin K, Lee I, Chait A , Wight TN, Innerarity TL: Identification of the principal proteoglycan-binding site in L D L . A single-point mutation in apo-BlOO severely affects proteoglycan interaction without affecting L D L receptor binding. J Clin Invest 1998, 101:2658-2664. 107. Camejo G, Olofsson SO, Lopez F, Carlsson P, Bondjers G: Identification of Apo B-100 segments mediating the interaction of low density lipoproteins with arterial proteoglycans. Arteriosclerosis 1988, 8:368-377. 108. Boren J, Lee I, Zhu W, Arnold K, Taylor S, Innerarity TL: Identification of the low density lipoprotein receptor-binding site in apolipoprotein B100 and the modulation of its binding activity by the carboxyl terminus in familial defective apo-BlOO. J Clin Invest 1998,101:1084-1093. 109. Veniant M M , K i m E, McCormick S, Boren J, Nielsen L B , Raabe M , Young SG: Insights into apolipoprotein B biology from transgenic and gene-targeted mice. J Nutr 1999, 129:451S-455S. 110. Olin K L , Potter-Perigo S, Barrett PH, Wight TN, Chait A : Lipoprotein lipase enhances the binding of native and oxidized low density lipoproteins to versican and biglycan synthesized by cultured arterial smooth muscle cells. J Biol Chem 1999, 274:34629-34636. 166 111. Pentikainen M O , Oorni K, Kovanen PT: Lipoprotein lipase (LPL) strongly links native and oxidized low density lipoprotein particles to decorin-coated collagen. Roles for both dimeric and monomeric forms of LPL. J Biol Chem 2000, 275:5694-5701. 112. Chang M Y , Potter-Perigo S, Tsoi C, Chait A , Wight TN: Oxidized low density lipoproteins regulate synthesis of monkey aortic smooth muscle cell proteoglycans that have enhanced native low density lipoprotein binding properties. J Biol Chem 2000, 275:4766-4773. 113. Williams KJ : Arterial wall chondroitin sulfate proteoglycans: diverse molecules with distinct roles in lipoprotein retention and atherogenesis. Curr Opin Lipidol 2001, 12:477-487. 114. Galis ZS, Alavi M Z , Moore S: Co-localization of aortic apolipoprotein B and chondroitin sulfate in an injury model of atherosclerosis. Am J Pathol 1993, 142:1432-1438. 115. Shi Y , Niculescu R, Wang D, Ormont M , Magno M , San Antonio JD, Williams KJ , Zalewski A : Myofibroblast involvement in glycosaminoglycan synthesis and lipid retention during coronary repair. J Vase Res 2000, 37:399-407. 116. Shi Y , Patel S, Davenpeck K L , Niculescu R, Rodriguez E, Magno M G , Ormont M L , Mannion JD, Zalewski A : Oxidative stress and lipid retention in vascular grafts: comparison between venous and arterial conduits. Circulation 2001,103:2408-2413. 117. Williams KJ , Tabas I: Atherosclerosis—an inflammatory disease. N Engl J Med 1999, 340:1928. 118. Williams K J : Interactions of lipoproteins with proteoglycans. Methods Mol Biol 2001, 171:457-477. 119. Makoveichuk E, Lookene A, Olivecrona G: Mild oxidation of lipoproteins increases their affinity for surfaces covered by heparan sulfate and lipoprotein lipase. Biochem Biophys Res Commun 1998, 252:703-710. 167 120. Oorni K, Pentikainen M O , Annila A , Kovanen PT: Oxidation of low density lipoprotein particles decreases their ability to bind to human aortic proteoglycans. Dependence on oxidative modification of the lysine residues. J Biol Chem 1997, 272:21303-21311. 121. Oorni K, Hakala JK, Annila A , Ala-Korpela M , Kovanen PT: Sphingomyelinase induces aggregation and fusion, but phospholipase A2 only aggregation, of low density lipoprotein (LDL) particles. Two distinct mechanisms leading to increased binding strength of L D L to human aortic proteoglycans. J Biol Chem 1998, 273:29127-29134. 122. . Schissel SL, Tweedie-Hardman J, Rapp JH, Graham G, Williams KJ , Tabas I: Rabbit aorta and human atherosclerotic lesions hydrolyze the sphingomyelin of retained low-density lipoprotein. Proposed role for arterial-wall sphingomyelinase in subendothelial retention and aggregation of atherogenic lipoproteins. J Clin Invest 1996, 98:1455-1464. 123. Griffin B A : Lipoprotein atherogenicity: an overview of current mechanisms. Proc Nutr Soc 1999,58:163-169. 124. Lundstam U , Hurt-Camejo E, Olsson G, Sartipy P, Camejo G, Wiklund O: Proteoglycans contribution to association of Lp(a) and L D L with smooth muscle cell extracellular matrix. Arterioscler Thromb Vase Biol 1999,19:1162-1167. 125. Bdeir K, Cane W, Canziani G, Chaiken I, Weisel J, Koschinsky M L , Lawn R M , Bannerman PG, Sachais BS, Kuo A , Hancock M A , Tomaszewski J, Raghunath PN, Ganz T, Higazi A A, Cines DB: Defensin promotes the binding of lipoprotein(a) to vascular matrix. Blood 1999, 94:2007-2019. 126. Camejo G, Fager G, Rosengren B, Hurt-Camejo E, Bondjers G: Binding of low density lipoproteins by proteoglycans synthesized by proliferating and quiescent human arterial smooth muscle cells. J Biol Chem 1993, 268:14131-14137. 168 127. Little PJ, Tannock L, Olin K L , Chait A , Wight TN: Proteoglycans synthesized by arterial smooth muscle cells in the presence of transforming growth factor-betal exhibit increased binding to LDLs. Arterioscler Thromb Vase Biol 2002, 22:55-60. 128. Olsson U , Bondjers G, Camejo G: Fatty acids modulate the composition of extracellular matrix in cultured human arterial smooth muscle cells by altering the expression of genes for proteoglycan core proteins. Diabetes 1999, 48:616-622. 129. Hosenpud JD, Bennett LE , Keck B M , Fiol B, Novick RJ: The Registry of the International Society for Heart and Lung Transplantation: fourteenth official report— 1997. J Heart Lung Transplant 1997,16:691-712. 130. Radio S, Wood S, Wilson J, Lin H, Winters G, McManus B: Allograft vascular disease: comparison of heart and other grafted organs. Transplant Proc 1996, 28:496-499. 131. Billingham M E : Histopathology of graft coronary disease. J Heart Lung Transplant 1992, ll:S38-44. 132. Lin H, Wilson JE, Kendall TJ, Radio SJ, Cornhill FJ, Herderick E, Winters GL, Costanzo MR, Porter T, Thieszen SL, McManus B M : Comparable proximal and distal severity of intimal thickening and size of epicardial coronary arteries in transplant arteriopathy of human cardiac allografts. J Heart Lung Transplant 1994, 13:824-833. 133. Gao SZ, Schroeder JS, Alderman EL, Hunt SA, Silverman JF, Wiederhold V , Stinson EB: Clinical and laboratory correlates of accelerated coronary artery disease in the cardiac transplant patient. Circulation 1987, 76:V56-61. 134. Gao SZ, Alderman EL, Schroeder JS, Silverman JF, Hunt SA: Accelerated coronary vascular disease in the heart transplant patient: coronary arteriographic findings. J A m Coll Cardiol 1988,12:334-340. 135. Yeung A C , Davis SF, Hauptman PJ, Kobashigawa JA, Miller LW, Valantine HA, Ventura HO, Wiedermann J, Wilensky R: Incidence and progression of transplant 169 coronary artery disease over 1 year: results of a multicenter trial with use of intravascular ultrasound. J Heart Lung Transplant 1995, 14:S215-220. 136. Andersen HO, Holm P, Stender S, Hansen BF, Kjeldsen K, Nordestgaard BG: Relative importance of ischemic injury and immunological injury on the development of transplant arteriosclerosis in rabbit aortic allografts. Transplantation 1995, 60:631-638. 137. Weis M , von Scheidt W: Cardiac allograft vasculopathy: a review. Circulation 1997, 96:2069-2077. 138. Virchow R: Phlogose und thrombose im gefabsystem. Gesammelte Abhandlungen zur Wissenschaftlichen Medicin Frankfurt, Meidinger Sohn 1856, 458. 139. Winters GL, Kendall TJ, Radio SJ, Wilson JE, Costanzo-Nordin MR, Switzer B L , Remmenga JA, McManus B M : Posttransplant obesity and hyperlipidemia: major predictors of severity of coronary arteriopathy in failed human heart allografts. J Heart Transplant 1990, 9:364-371. 140. Eich D, Thompson JA, Ko D, Hastillo A, Lower R, Katz S, Katz M , Hess M L : Hypercholesterolemia in long-term survivors of heart transplantation: an early marker of accelerated coronary artery disease. J Heart Lung Transplant 1991,10:45-49. 141. McManus B M , Horley KJ , Wilson JE, Malcom GT, Kendall TJ, Miles RR, Winters GL, Costanzo MR, Miller LL , Radio SJ: Prominence of coronary arterial wall lipids in human heart allografts. Implications for pathogenesis of allograft arteriopathy. Am J Pathol 1995, 147:293-308. 142. Malcom GT, Strong JP, Restrepo C: Atherosclerosis and lipid composition of the abdominal aorta. Comparison of autopsied New Orleans and Guatemalan men. Lab Invest 1984, 50:79-86. 170 143. Wilson JE, Wood S, McDonald P, Kenyon J, Dong C, McManus B M : Contribution of lipids to the pathogenesis of transplant vascular sclerosis. Transpl Immunol 1997, 5:247-250. 144. Amenta JS, Waters L L : The precipitation of serum lipoproteins by mucopolysaccharides extracted from aortic tissue. Yale J Biol Med 1960, 33:112-121. 145. Wight TN, Ross R: Proteoglycans in primate arteries. I. Ultrastructural localization and distribution in the intima. J Cell Biol 1975, 67:660-674. 146. Camejo G: The interaction of lipids and lipoproteins with the intercellular matrix of arterial tissue: its possible role in atherogenesis. Adv Lipid Res 1982,19:1-53. 147. Ismail N A E , Alavi M Z , Moore S: Lipoprotein-proteoglycan complexes from injured rabbit aortas accelerate lipoprotein uptake by arterial smooth muscle cells. Atherosclerosis 1994,105:79-87. 148. Kobashigawa JA, Kasiske B L : Hyperlipidemia in solid organ transplantation. Transplantation 1997, 63:331-338. 149. Ballantyne C M , Podet EJ, Patsch WP, Harati Y , Appel V , Gotto A M , Jr., Young JB: Effects of cyclosporine therapy on plasma lipoprotein levels. J A M A 1989, 262:53-56. 150. Schorn TF, Kliem V , Bojanovski M , Bojanovski D, Repp H, Bunzendahl H , Frei U : Impact of long-term immunosuppression with cyclosporin A on serum lipids in stable renal transplant recipients. Transplant Int 1991, 4:92-95. 151. Uretsky BF, Murali S, Reddy PS, Rabin B, Lee A , Griffith BP, Hardesty RL , Trento A , Bahnson HT: Development of coronary artery disease in cardiac transplant patients receiving immunosuppressive therapy with cyclosporine and prednisone. Circulation 1987, 76:827-834. 171 152. van den Dorpel M A , Ghanem H , Rischen-Vos J, Man in't Veld A J , Jansen H , Weimar W: Conversion from cyclosporine A to azathioprine treatment improves L D L oxidation in kidney transplant recipients. Kidney Int 1997, 51:1608-1612. 153. Harris KPG, Russell G l , Parvin SD, Veitch PS, Walls J: Alterations in lipid and carbohydrate metabolism attributable to cyclosporin A in renal transplant recipients. Br Med J 1986, 292:16. 154. Versluis DJ, Wenting GJ, Derkx F H M , Schalekamp M A D H , Jeekel J, Weimar W: Who should be converted from cyclosporine to conventional immunosuppression in kidney transplantation, and why. Transplantation 1987, 44:387-389. 155. Brown JH, Murphy B G , Douglas A F , Short CD, Bhatnagar D, Mackness MI, Hunt LP, Doherty CC, Durrington PN: Influence of immunosuppressive therapy on lipoprotein(a) and other lipoproteins following renal transplantation. Nephron 1997, 75:277-282. 156. Hohage H , Arlt M , Bruckner D, Dietl K H , Zidek W, Spieker S: Effects of cyclosporin A and FK 506 on lipid metabolism and fibrinogen in kidney transplant recipients. Clin Transplant 1997, 11:225-230. 157. Ferns G, Reidy M , Ross R: Vascular effects of cyclosporine A in vivo and in vitro. A m J Pathol 1990,137:403-413. 158. Miller L, Kobashigawa J, Valantine H, Ventura H, Hauptman P, O'Donnell J, Wiedermann J, Yeung A: The impact of cyclosporine dose and level on the development and progression of allograft coronary disease. J Heart Lung Transplant 1995, 14:S227-234. 159. Mehra MR, Ventura HO, Jain SP, Ramireddy K, Al i A , Stapleton DD, Smart FW, Ramee SR, Collins TJ, White CJ: Heterogeneity of cardiac allograft vasculopathy: clinical insights from coronary angioscopy. J A m Coll Cardiol 1997, 29:1339-1344. 172 160. McDonald PC, Wilson JE, McManus B M : Fen-phen valve disease: a real entity? Pathol Case Rev 2001,6:281-286. 161. Steffee C H , Singh H K , Chitwood WR: Histologic changes in three explanted native cardiac valves following use of fenfluramines. Cardiovasc Pathol 1999, 8:245-253. 162. McDonald PC, Wilson JE, Gao M , McNeill S, Spinelli JJ, Williams OD, Harji S, Kenyon JA, McManus B M : Quantitative analysis of human heart valves: Does anorexigen exposure produce a distinctive morphological lesion? Cardiovasc Pathol 2002, 11:251-262. 163. Becker A E : Acquired heart valve pathology. An update for the millennium. Herz 1998, 23:415-419. 164. Rose A G : Etiology of valvular heart disease. Curr Opin Cardiol 1996,11:98-l 13. 165. Edwards WD, Peterson K, Edwards JE: Active valvulitis associated with chronic rheumatic valvular disease and active myocarditis. Circulation 1978, 57:181-185. 166. Lis Y , Burleigh M C , Parker DJ, Child A H , Hogg J, Davies MJ: Biochemical characterization of individual normal, floppy and rheumatic human mitral valves. Biochem J 1987,244:597-603. 167. Virmani R, Atkinson JB, Forman M B : The pathology of mitral valve prolapse. Herz 1988,13:215-226. 168. Wigle ED, Rakowski H , Ranganathan N , Silver M C : Mitral valve prolapse. Ann Rev Med 1976,27:165-180. 169. Olsen EG, Al-Rufaie HK: The floppy mitral valve. Study on pathogenesis. Br Heart J 1980, 44:674-683. 170. Cole WG, Chan D, Hickey A J , Wilcken DE: Collagen composition of normal and myxomatous human mitral heart valves. Biochem J 1984, 219:451-460. 173 171. Kulke M H , Mayer RJ: Medical Progress: Carcinoid Tumors. N Engl J of Med 1999, 340:858-868. 172. Roberts WC: A unique heart disease associated with a unique cancer: carcinoid heart disease. A m J Cardiol 1997, 80:251-256. 173. Lundin L, Funa K , Hansson HE, Wilander E, Oberg K: Histochemical and immunohistochemical morphology of carcinoid heart disease. Pathol Res Pract 1991, 187:73-77. 174. Strickman NE, Rossi PA, Massumkhani GA, Hall RJ: Carcinoid heart disease: a clinical pathologic, and therapeutic update. Curr Prob Cardiol 1982, 6:1-42. 175. Adams D H , Karnovsky MJ: Hypercholesterolemia does not exacerbate arterial intimal thickening in chronically rejecting rat cardiac allografts. Transplant Proc 1989, 21:437-439. 176. Armstrong AT, Strauch AR, Kardan A , Starling RC: Morphometric and immunocytochemical analysis of coronary arterioles in human transplanted hearts. J Heart Lung Transplant 1996, 15:818-826. 177. Ross R: The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature 1993, 362:801-809. 178. Schwartz SM, deBlois D, ER OB: The intima. Soil for atherosclerosis and restenosis. CircRes 1995,77:445-465. 179. Sanders M : Molecular and cellular concepts in atherosclerosis. Pharmaco Therapeu 1994, 61:109-153. 180. BashaBJ, Sowers JR: Atherosclerosis: an update. A m Heart J 1996, 131:1192-1202. 181. Molossi S, Clausell N , Rabinovitch M : Coronary artery endothelial interleukin-1 beta mediates enhanced fibronectin production related to post-cardiac transplant arteriopathy in piglets. Circulation 1993, 88:11248-256. 174 182. Clausell N , Rabinovitch M : Upregulation of fibronectin synthesis by interleukin-1 beta in coronary artery smooth muscle cells is associated with the development of the post-cardiac transplant arteriopathy in piglets. J Clin Invest 1993, 92:1850-1858. 183. Molossi S, Clausell N , Rabinovitch M : Reciprocal induction of tumor necrosis factor-alpha and interleukin-1 beta activity mediates fibronectin synthesis in coronary artery smooth muscle cells. J Cell Physiol 1995,163:19-29. 184. Cowan B, Baron O, Crack J, Coulber C, Wilson GJ, Rabinovitch M : Elafin, a serine elastase inhibitor, attenuates post-cardiac transplant coronary arteriopathy and reduces myocardial necrosis in rabbits afer heterotopic cardiac transplantation. J Clin Invest 1996, 97:2452-2468. 185. Akyurek L M , Paul L C , Funa K, Larsson E, Fellstrom BC: Smooth muscle cell migration into intima and adventitia during development of transplant vasculopathy. Transplantation 1996, 62:1526-1529. 186. Lemstrom K B , Koskinen PK: Expression and localization of platelet-derived growth factor ligand and receptor protein during acute and chronic rejection of rat cardiac allografts. Circulation 1997, 96:1240-1249. 187. Alpers CE, Davis C L , Barr D, Marsh CL, Hudkins K L : Identification of platelet-derived growth factor A and B chains in human renal vascular rejection. A m J Pathol 1996, 148:439-451. 188. Waltenberger J, Akyurek M L , Aurivillius M , Wanders A , Larsson E, Fellstrom B, Funa K: Ischemia-induced transplant arteriosclerosis in the rat. Induction of peptide growth factor expression. Arterioscler Thromb Vase Biol 1996,16:1516-1523. 189. Lou H , Kodama T, Wang Y N , Katz N , Ramwell P, Foegh M L : L-arginine prevents heart transplant arteriosclerosis by modulating the vascular cell proliferative response to 175 insulin-like growth factor-I and interleukin-6. J Heart Lung Transplant 1996, 15:1248-1257. 190. Motomura N , Lou H, Orskov H, Ramwell PW, Foegh M L : Exposure of vascular allografts to insulin-like growth factor-I (IGF-I) increases vascular expression of IGF-I ligand and receptor protein and accelerates arteriosclerosis in rats. Transplantation 1998, 65:1024-1030. 191. Hayry P, Aavik E, Myllarniemi M : Blockade of growth factor synthesis and growth factor action: two possible sites of interference in allograft vessel disease and coronary bypass or balloon injury. Metabolism 1996, 45:101-103. 192. McManus B M , Malcolm G, Kendall TJ: Lipid overload and proteoglycan expression in chronic rejection of the human transplanted heart. Clin Transplant 1994, 8:336-340. 193. Wight TN, Lara S, Riessen R, Le Baron R, Isner J: Selective deposits of versican in the extracellular matrix of restenotic lesions from human peripheral arteries. A m J Pathol 1997, 151:963-973. 194. Lin H , Kanda T, Hoshino Y , Takase S, Kobayashi I, Nagai R, McManus B M : Versican, biglycan and decorin protein expression patternis in coronary arteries: analysis of primary and restenotic lesions. Cardiovasc Pathol 1998, 7:31-37. 195. Schonherr E, Jarvelainen HT, Kinsella M G , Sandell LJ, Wight TN: Platelet-derived growth factor and transforming growth factor-beta 1 differentially affect the synthesis of biglycan and decorin by monkey arterial smooth muscle cells. Arterioscler Thromb 1993, 13:1026-1036. 196. Edwards IJ, Xu H, Wright MJ , Wagner WD: Interleukin-1 upregulates decorin production by arterial smooth muscle cells. Arterioscler Thromb 1994, 14:1032-1039. i 176 197. Deudon E, Berrou E, Breton M , Picard J: Growth-related production of proteoglycans and hyaluronic acid in synchronous arterial smooth muscle cells. Int J Biochem 1992, 24:465-470. 198. Breton M , Berrou E, Brahimi-Horn M C , Deudon E, Picard J: Synthesis of sulfated proteoglycans throughout the cell cycle in smooth muscle cells from pig aorta. Exp Cell Res 1986, 166:416-426. 199. Dreher K L , Asundi V , Matzura D, Cowan K: Vascular smooth muscle biglycan represents a highly conserved proteoglycan within the arterial wall. Eur J Cell Biol 1990, 53:296-304. 200. Schurch W, Seemayer TA, Gabbiani G. Myofibroblast. In: Histology for Pathologists (Second ed.), edited by S. S. Sternberg. Philadelphia: Lippincott-Raven, 1997, p. 129-165. 201. Gabbiani G: Some historical and philosophical reflections on the myofibroblast concept. Curr Top Pathol 1999, 93:1-5. 202. Schurch W, Seemayer TA, Gabbiani G: The myofibroblast: a quarter century after its discovery. A m J Surg Pathol 1998, 22:141-147. 203. Gabbiani G: The cellular derivation and the life span of the myofibroblast. Pathol Res Practl996, 192:708-711. 204. Eyden B: The myofibroblast: an assessment of controversial issues and a definition useful in diagnosis and research. Ultrastruct Pathol 2001, 25:39-50. 205. Gabbiani G: Evolution and clinical implications of the myofibroblast concept. Cardiovasc Res 1998, 38:545-548. 206. Gabbiani G: The biology of the myofibroblast. Kidney Int 1992, 41:530-532. 207. Johnson RJ, Floege J, Yoshimura A , Iida H , Couser WG, Alpers CE: The activated mesangial cell: a glomerular "myofibroblast"? J A m Soc Nephrol 1992, 2:SI90-197. 177 208. Boukhalfa G, Desmouliere A , Rondeau E, Gabbiani G, Sraer JD: Relationship between alpha-smooth muscle actin expression and fibrotic changes in human kidney. Exp Nephrol 1996,4:241-247. 209. Mulholland DL, Gotlieb A l : Cell biology of valvular interstitial cells. Can J Cardiol 1996,12:231-236. 210. Powell DW, Mifflin RC, Valentich JD, Crowe SE, Saada JI, West A B : Myofibroblasts. I. Paracrine cells important in health and disease. A m J Physiol 1999, 277:Cl-9. 211. Gabbiani G, Majno G: Dupuytren's contracture: fibroblast contraction? An ultrastructural study. A m J Pathol 1972, 66:131-146. 212. Desmouliere A : Factors influencing myofibroblast differentiation during wound healing and fibrosis. Cell Biol Int 1995,19:471-476. 213. Desmouliere A , Geinoz A , Gabbiani F, Gabbiani G: Transforming growth factor-beta 1 induces alpha-smooth muscle actin expression in granulation tissue myofibroblasts and in quiescent and growing cultured fibroblasts. J Cell Biol 1993, 122:103-111. 214. Serini G, Gabbiani G: Mechanisms of myofibroblast activity and phenotypic modulation. Exp Cell Res 1999, 250:273-283. 215. Andreutti D, Gabbiani G, Neuville P: Early granulocyte-macrophage colony-stimulating factor expression by alveolar inflammatory cells during bleomycin-induced rat lung fibrosis. Lab Invest 1998, 78:1493-1502. 216. Vyalov S, Desmouliere A , Gabbiani G: GM-CSF-induced granulation tissue formation: relationships between macrophage and myofibroblast accumulation. Virchows Arch B Cell Pathol Incl Mol Pathol 1993, 63:231-239. 217. Serini G, Gabbiani G: Modulation of a-smooth muscle actin expression by transforming growth factor-p isoforms: An in vivo and in vitro study. Wound Rep Reg 1996, 4:277-287. 178 218. Petroll W M , Jester JV, Bean JJ, Cavanagh HD: Myofibroblast transformation of cat corneal endothelium by transforming growth factor-betal, -beta2, and -beta3. Invest Ophthalmol Vis Sci 1998, 39:2018-2032. 219. Phan SH: Role of the myofibroblast in pulmonary fibrosis. Kidney Int Suppl 1996, 54:S46-48. 220. Xing Z, Tremblay G M , Sime PJ, Gauldie J: Overexpression of granulocyte-macrophage colony-stimulating factor induces pulmonary granulation tissue formation and fibrosis by induction of transforming growth factor-beta 1 and myofibroblast accumulation. Am J Pathol 1997,150:59-66. 221. Serini G, Bochaton-Piallat M L , Ropraz P, Geinoz A , Borsi L, Zardi L, Gabbiani G: The fibronectin domain ED-A is crucial for myofibroblastic phenotype induction by transforming growth factor-betal. J Cell Biol 1998, 142:873-881. 222. Bochaton-Piallat M L , Kapetanios A D , Donati G, Redard M , Gabbiani G, Pournaras CJ: TGF-betal, TGF-beta receptor II and ED-A fibronectin expression in myofibroblast of vitreoretinopathy. Invest Ophthalmol Vis Sci 2000, 41:2336-2342. 223. Vyalov SL, Gabbiani G, Kapanci Y : Rat alveolar myofibroblasts acquire alpha-smooth muscle actin expression during bleomycin-induced pulmonary fibrosis. A m J Pathol 1993, 143:1754-1765. 224. Taylor P M , Allen SP, Yacoub M H : Phenotypic and functional characterization of interstitial cells from human heart valves, pericardium and skin. J Heart Valve Dis 2000, 9:150-158. 225. Martin P: Wound healing-aiming for perfect skin regeneration. Science 1997, 276:75-81. 226. Bachem M G , Sell K M , Melchior R, Kropf J, Eller T, Gressner A M : Tumor necrosis factor alpha (TNF alpha) and transforming growth factor beta 1 (TGF beta 1) stimulate 179 fibronectin synthesis and the transdifferentiation of fat-storing cells in the rat liver into myofibroblasts. Virchows Arch B Cell Pathol Incl Mol Pathol 1993, 63:123-130. 227. Schneider E, Schmid-Kotsas A , Zhao J, Weidenbach H , Schmid R M , Menke A , Adler G, Waltenberger J, Grunert A , Bachem M G : Identification of mediators stimulating proliferation and matrix synthesis of rat pancreatic stellate cells. A m J Physiol Cell Physiol 2001, 281:C532-543. 228. Berndt A , Kosmehl H , Mandel U , Gabler U , Luo X , Celeda D, Zardi L, Katenkamp D: TGF beta and bFGF synthesis and localization in Dupuytren's disease (nodular palmar fibromatosis) relative to cellular activity, myofibroblast phenotype and oncofetal variants of fibronectin. Histochem J 1995, 27:1014-1020. 229. Igarashi A , Okochi H , Bradham D M , Grotendorst GR: Regulation of connective tissue growth factor gene expression in human skin fibroblasts and during wound repair. Mol Biol Cell 1993,4:637-645. 230. Spanakis SG, Petridou S, Masur SK: Functional gap junctions in corneal fibroblasts and myofibroblasts. Invest Ophthalmol Vis Sci 1998, 39:1320-1328. 231. Nakatsuji S, Yamate J, Sakuma S: Macrophages, myofibroblasts, and extracellular matrix accumulation in interstitial fibrosis of chronic progressive nephropathy in aged rats. Vet Pathol 1998,35:352-360. 232. Zhang K , Rekhter M D , Gordon D, Phan SH: Myofibroblasts and their role in lung collagen gene expression during pulmonary fibrosis. A combined immunohistochemical and in situ hybridization study. A m J Pathol 1994,145:114-125. 233. Kresse H, Schonherr E: Proteoglycans of the extracellular matrix and growth control. J Cell Physiol 2001,189:266-274. 180 234. Messier RH, Jr., Bass B L , Aly H M , Jones JL, Domkowski PW, Wallace RB, Hopkins RA: Dual structural and functional phenotypes of the porcine aortic valve interstitial population: characteristics of the leaflet myofibroblast. J Surg Res 1994, 57:1-21. 235. Filip DA, Radu A, Simionescu M : Interstitial cells of the heart valves possess characteristics similar to smooth muscle cells. Circ Res 1986, 59:310-320. 236. Lester W M , Damji A A , Gedeon I, Tanaka M : Interstitial cells from the atrial and ventricular sides of the bovine mitral valve respond differently to denuding endocardial injury. In Vitro Cell Dev Biol 1993, 29A:41-50. 237. Lester W, Rosenthal A , Granton B, Gotlieb A l : Porcine mitral valve interstitial cells in culture. Lab Invest 1988, 59:710-719. 238. Zacks S, Rosenthal A , Granton B, Havenith M , Opas M , Gotlieb A l : Characterization of Cobblestone mitral valve interstitial cells. Arch Pathol Lab Med 1991, 115:774-779. 239. Johnson C M , Hanson M N , Helgeson SC: Porcine cardiac valvular subendothelial cells in culture: cell isolation and growth characteristics. J Mol Cell Cardiol 1987, 19:1185-1193. 240. Roy A, Brand NJ, Yacoub M H : Expression of 5-hydroxytryptamine receptor subtype messenger R N A in interstitial cells from human heart valves. J Heart Valve Dis 2000, 9:256-260. 241. Lester W M , Gotlieb A l : In vitro repair of the wounded porcine mitral valve. Circ Res 1988, 62:833-845. 242. Lester W M , Damji A A , Tanaka M , Gedeon I: Bovine mitral valve organ culture: role of interstitial cells in repair of valvular injury. J Mol Cell Cardiol 1992, 24:43-53. 243. Gulizia JM, Kandolf R, Kendall TJ, Thieszen SL, Wilson JE, Radio SJ, Costanzo MR, Winters GL , Miller L L , McManus B M : Infrequency of cytomegalovirus genome in coronary arteriopathy of human heart allografts. Am J Pathol 1995, 147:461-475. 181 244. McDonald PC, Wilson JE, Gao M , Spinelli JJ, Wiebe H, McNeill S, McManus B M : The challenge of defining normality of human mitral and aortic valves: geometrical and compositional analysis. Cardiovasc Pathol 2002, 11:193-209. 245. Thyberg J: Differentiated properties and proliferation of arterial smooth muscle cells in culture. Int Rev Cytol 1996,169:183-265. 246. Campbell JH, Campbell GR: Culture techniques and their applications to studies of vascular smooth muscle. Clin Sci 1993, 85:501-513. 247. Fisher LW, Stubbs JT, 3rd, Young MF: Antisera and cDNA probes to human and certain animal model bone matrix noncollagenous proteins. Acta Orthop Scand Suppl 1995, 266:61-65. 248. Rosenfeld M E , Palinski W, Yla-Herttuala S, Butler S, Witztum JL: Distribution of oxidation specific lipid-protein adducts and apolipoprotein B in atherosclerotic lesions of varying severity from W H H L rabbits. Arteriosclerosis 1990, 10:336-349. 249. Palinski W, Yla-Herttuala S, Rosenfeld M E , Butler SW, Socher SA, Parthasarathy S, Curtiss L K , Witztum JL: Antisera and monoclonal antibodies specific for epitopes generated during oxidative modification of low density lipoprotein. Arteriosclerosis 1990,10:325-335. 250. Olsson M , Thyberg J, Nilsson J: Presence of oxidized low density lipoprotein in nonrheumatic stenotic aortic valves. Arterioscler Thromb Vase Biol 1999, 19:1218-1222. 251. Napoli C, DArmiento FP, Mancini FP, Postiglione A , Witztum JL, Palumbo G, Palinski W: 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 1997, 100:2680-2690. 182 252. Zhang H, Yang Y , Steinbrecher UP: Structural requirements for the binding of modified proteins to the scavenger receptor of macrophages. J Biol Chem 1993, 268:5535-5542. 253. Magil A B , Frohlich JJ, Innis S M , Steinbrecher UP: Oxidized low-density lipoprotein in experimental focal glomerulosclerosis. Kidney Int 1993, 43:1243-1250. 254. Pyke C, Kristensen P, Ostergaard PB, Oturai PS, Romer J: Proteoglycan expression in the normal rat kidney. Nephron 1997, 77:461-470. 255. McDonald PC, Wong D, Granville DJ, McManus B M : Emerging roles of endothelial cells and smooth muscle cells in transplant vascular disease. Transplant Rev 1999, 13:109-127. 256. McDonald PC, Kenyon JA, McManus B M : The role of lipids in transplant vascular disease. Lab Invest 1998, 78:1187-1201. 257. Gutierrez P, O'Brien K D , Ferguson M , Nikkari ST, Alpers CE, Wight TN: Differences in the distribution of versican, decorin and biglycan in atherosclerotic human coronary arteries. Cardiovasc Pathol 1997, 6:271-278. 258. Matsuura R, Isaka N , Imanaka-Yoshida K, Yoshida T, Sakakura T, Nakano T: Deposition of PG-M/versican is a major cause of human coronary restenosis after percutaneous transluminal coronary angioplasty. J Pathol 1996,180:311-316. 259. Merrilees MJ, Beaumont B, Scott LJ: Comparison of deposits of versican, biglycan and decorin in saphenous vein and internal thoracic, radial and coronary arteries: correlation to patency. Coron Artery Dis 2001,12:7-16. 260. Mitsumata M , Gamou S, Shimizu N , Yoshida Y : Response of atherosclerotic intimal smooth muscle cells to epidermal growth factor in vitro. Arterioscler Thromb 1994, 14:1364-1371. 183 261. Saltis J, Thomas A C , Agrotis A , Campbell JH, Campbell GR, Bobik A : Expression of growth factor receptors in arterial smooth muscle cells. Dependency on cell phenotype and serum factors. Atherosclerosis 1995,118:77-87. 262. Wilson V J , Ward JP, Burnand K G , Thomas CR: Upregulation of IGF-I and collagen I mRNA in human atherosclerotic tissue is not accompanied by changes in type 1 IGF receptor or collagen III mRNA: an in situ hybridization study. Coron Artery Dis 1996, 7:569-572. 263. Grant M B , Wargovich TJ, Bush D M , Player DW, Caballero S, Foegh M , Spoerri PE: Expression of IGF-1, IGF-1 receptor and TGF-beta following balloon angioplasty in atherosclerotic and normal rabbit iliac arteries: an immunocytochemical study. Regul Peptl999, 79:47-53. 264. Okura Y , Brink M , Zahid A A , Anwar A , Delafontaine P: Decreased expression of insulin-like growth factor-1 and apoptosis of vascular smooth muscle cells in human atherosclerotic plaque. J Mol Cell Cardiol 2001, 33:1777-1789. 265. Blay J, Hollenberg M D : Heterologous regulation of EGF receptor function in cultured aortic smooth muscle cells. Eur J Pharmacol 1989,172:1-7. 266. D'Amore PA, Smith SR: Growth factor effects on cells of the vascular wall: a survey. Growth Factors 1993, 8:61-75. 267. Trieu V N , Narla RK, Myers DE, Uckun F M : EGF-genistein inhibits neointimal hyperplasia after vascular injury in an experimental restenosis model. J Cardiovasc Pharmacol 2000, 35:595-605. 268. Motomura N , Lou H , Maurice P, Foegh M L : Acceleration of arteriosclerosis of the rat aorta allograft by insulin growth factor-I. Transplantation 1997, 63:932-936. 184 269. Lou H, Ramwell PW, Foegh M L : Estradiol 17-beta represses insulin-like growth factor I receptor expression in smooth muscle cells from rabbit cardiac recipients. Transplantation 1998, 66:419-426. 270. Kaiura TL, Itoh H, Kubaska SM, 3rd, McCaffrey TA, Liu B, Kent K C : The effect of growth factors, cytokines, and extracellular matrix proteins on fibronectin production in human vascular smooth muscle cells. J Vase Surg 2000, 31:577-584. 271. Lo CS: PDGF-BB and IGF-I use different signaling pathways to induce fibronectin synthesis in cultured rat thoracic aortic smooth muscle cells. Cell Biol Int 2001, 25:371-373. 272. Syrokou A, Tzanakakis G N , Hjerpe A , Karamanos N K : Proteoglycans in human malignant mesothelioma. Stimulation of their synthesis induced by epidermal, insulin and platelet-derived growth factors involves receptors with tyrosine kinase activity. Biochimie 1999, 81:733-744. 273. Bokemeyer D, Schmitz U , Kramer HJ: Angiotensin Il-induced growth of vascular smooth muscle cells requires an Src-dependent activation of the epidermal growth factor receptor. Kidney Int 2000, 58:549-558. 274. Iwasaki H, Eguchi S, Marumo F, Hirata Y : Endothelin-1 stimulates D N A synthesis of vascular smooth-muscle cells through transactivation of epidermal growth factor receptor. J Cardiovasc Pharmacol 1998, 31 Suppl LS182-184. 275. Imai Y , Clemmons DR: Roles of phosphatidylinositol 3-kinase and mitogen-activated protein kinase pathways in stimulation of vascular smooth muscle cell migration and deoxyriboncleic acid synthesis by insulin-like growth factor-I. Endocrinology 1999, 140:4228-4235. 185 276. Yano K, Bauchat JR, Liimatta M B , Clemmons DR, Duan C: Down-regulation of protein kinase C inhibits insulin-like growth factor I-induced vascular smooth muscle cell proliferation, migration, and gene expression. Endocrinology 1999, 140:4622-4632. 277. Arnqvist HJ, Borafeldt K E , Chen Y , Lindstrom T: The insulin-like growth factor system in vascular smooth muscle: interaction with insulin and growth factors. Metabolism 1995, 44:58-66. 278. Bayes-Genis A , Conover CA, Schwartz RS: The insulin-like growth factor axis: A review of atherosclerosis and restenosis. Circ Res 2000, 86:125-130. 279. Yla-Herttuala S: Macrophages and oxidized low density lipoproteins in the pathogenesis of atherosclerosis. Ann Med 1991, 23:561-567. 280. Yla-Herttuala S: Is oxidized low-density lipoprotein present in vivo? Curr Opin Lipidol 1998,9:337-344. 281. Berliner JA, Heinecke JW: The role of oxidized lipoproteins in atherogenesis. Free Radic Biol Med 1996, 20:707-727. 282. Berliner JA, Navab M , Fogelman A M , Frank JS, Demer L L , Edwards PA, Watson A D , Lusis AJ : Atherosclerosis: basic mechanisms. Oxidation, inflammation, and genetics. Circulation 1995, 91:2488-2496. 283. Ishikawa K , Navab M , Leitinger N , Fogelman A M , Lusis AJ : Induction of heme oxygenase-1 inhibits the monocyte transmigration induced by mildly oxidized LDL. J Clin Invest 1997,100:1209-1216. 284. Cushing SD, Berliner JA, Valente A J , Territo M C , Navab M , Parhami F, Gerrity R, Schwartz CJ, Fogelman A M : Minimally modified low density lipoprotein induces monocyte chemotactic protein 1 in human endothelial cells and smooth muscle cells. Proc Natl Acad Sci U S A 1990, 87:5134-5138. 186 285. Murphy A A , Santanam N , Morales A J , Parthasarathy S: Lysophosphatidyl choline, a chemotactic factor for monocytes/T-lymphocytes is elevated in endometriosis. J Clin Endocrinol Metab 1998, 83:2110-2113. 286. Dhaliwal BS, Steinbrecher UP: Scavenger receptors and oxidized low density lipoproteins. Clin Chim Acta 1999, 286:191-205. 287. Dhaliwal BS, Steinbrecher UP: Cholesterol delivered to macrophages by oxidized low density lipoprotein is sequestered in lysosomes and fails to efflux normally. J Lipid Res 2000,41:1658-1665. 288. Lougheed M , Moore ED, Scriven DR, Steinbrecher UP: Uptake of oxidized L D L by macrophages differs from that of acetyl L D L and leads to expansion of an acidic endolysosomal compartment. Arterioscler Thromb Vase Biol 1999, 19:1881-1890. 289. Sata M , Walsh K: Endothelial cell apoptosis induced by oxidized L D L is associated with the down-regulation of the cellular caspase inhibitor FLIP. J Biol Chem 1998, 273:33103-33106. 290. Kuzuya M , Ramos M A , Kanda S, Koike T, Asai T, Maeda K , Shitara K, Shibuya M , Iguchi A: V E G F protects against oxidized L D L toxicity to endothelial cells by an intracellular glutathione-dependent mechanism through the K D R receptor. Arterioscler Thromb Vase Biol 2001, 21:765-770. 291. Hsieh CC, Yen M H , Liu HW, Lau YT: Lysophosphatidylcholine induces apoptotic and non-apoptotic death in vascular smooth muscle cells: in comparison with oxidized LDL. Atherosclerosis 2000, 151:481-491. 292. Hughes H , Mathews B, Lenz M L , Guyton JR: Cytotoxicity of oxidized L D L to porcine aortic smooth muscle cells is associated with the oxysterols 7-ketocholesterol and 7-hydroxycholesterol. Arterioscler Thromb 1994,14:1177-1185. 187 293. Chatterjee S, Ghosh N : Oxidized low density lipoprotein stimulates aortic smooth muscle cell proliferation. Glycobiology 1996, 6:303-311. 294. Watanabe T, Pakala R, Katagiri T, Benedict CR: Mildly oxidized low-density lipoprotein acts synergistically with angiotensin II in inducing vascular smooth muscle cell proliferation. J Hypertens 2001,19:1065-1073. 295. Takei A , Huang Y , Lopes-Virella MF: Expression of adhesion molecules by human endothelial cells exposed to oxidized low density lipoprotein. Influences of degree of oxidation and location of oxidized L D L . Atherosclerosis 2001, 154:79-86. 296. Khan B V , Parthasarathy SS, Alexander RW, Medford R M : Modified low density lipoprotein and its constituents augment cytokine-activated vascular cell adhesion molecule-1 gene expression in human vascular endothelial cells. J Clin Invest 1995, 95:1262-1270. 297. Massy ZA, Kim Y , Guijarro C, Kasiske B L , Keane WF, O'Donnell MP: Low-density lipoprotein-induced expression of interleukin-6, a marker of human mesangial cell inflammation: effects of oxidation and modulation by lovastatin. Biochem Biophys Res Commun 2000, 267:536-540. 298. Hein TW, Liao JC, Kuo L: oxLDL specifically impairs endothelium-deperident, NO-mediated dilation of coronary arterioles. A m J Physiol Heart Circ Physiol 2000, 278.H175-183. 299. Tanner FC, Noll G, Boulanger C M , Luscher TF: Oxidized low density lipoproteins inhibit relaxations of porcine coronary arteries. Role of scavenger receptor and endothelium-derived nitric oxide. Circulation 1991, 83:2012-2020. 300. Uchida K: Cellular response to bioactive lipid peroxidation products. Free Radic Res 2000,33:731-737. 188 301. Palinski W, Witztum JL: Immune responses to oxidative neoepitopes on L D L and phospholipids modulate the development of atherosclerosis. J Intern Med 2000, 247:371-380. 302. Esterbauer H, Schaur RJ, Zollner H: Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free Radic Biol Med 1991, 11:81-128. 303. Esterbauer H , Dieber-Rotheneder M , Waeg G, Striegl G, Jurgens G: Biochemical, structural, and functional properties of oxidized low-density lipoprotein. Chem Res Toxicol 1990,3:77-92. 304. Palinski W, Ord V A , Plump AS, Breslow JL, Steinberg D, Witztum JL: ApoE-deficient mice are a model of lipoprotein oxidation in atherogenesis. Demonstration of oxidation-specific epitopes in lesions and high titers of autoantibodies to malondialdehyde-lysine in serum. Arterioscler Thromb 1994,14:605-616. 305. Haberland M E , Fong D, Cheng L: Malondialdehyde-altered protein occurs in atheroma of Watanabe heritable hyperlipidemic rabbits. Science 1988, 241:215-218. 306. Boyd HC, Gown A M , Wolfbauer G, Chait A : Direct evidence for a protein recognized by a monoclonal antibody against oxidatively modified L D L in atherosclerotic lesions from a Watanabe heritable hyperlipidemic rabbit. A m J Pathol 1989,135:815-825. 307. Palinski W, Horkko S, Miller E, Steinbrecher UP, Powell HC, Curtiss L K , Witztum JL: Cloning of monoclonal autoantibodies to epitopes of oxidized lipoproteins from apolipoprotein E-deficient mice. Demonstration of epitopes of oxidized low density lipoprotein in human plasma. J Clin Invest 1996, 98:800-814. 308. Palinski W, Miller E, Witztum JL: Immunization of low density lipoprotein (LDL) receptor-deficient rabbits with homologous malondialdehyde-modified L D L reduces atherogenesis. Proc Natl Acad Sci U S A 1995, 92:821-825. 189 309. Yla-Herttuala S, Palinski W, Rosenfeld M E , Parthasarathy S, Carew TE, Butler S, Witztum JL, Steinberg D: Evidence for the presence of oxidatively modified low density lipoprotein in atherosclerotic lesions of rabbit and man. J Clin Invest 1989, 84:1086-1095. 310. Palinski W, Rosenfeld ME, Yla-Herttuala S, Gurtner GC, Socher SS, Butler SW, Parthasarathy S, Carew TE, Steinberg D, Witztum JL: Low density lipoprotein undergoes oxidative modification in vivo. Proc Natl Acad Sci U S A 1989, 86:1372-1376. 311. Tsimikas S, Palinski W, Witztum JL: Circulating autoantibodies to oxidized L D L correlate with arterial accumulation and depletion of oxidized L D L in L D L receptor-deficient mice. Arterioscler Thromb Vase Biol 2001, 21:95-100. 312. Tsimikas S, Shortal BP, Witztum JL, Palinski W: In vivo uptake of radiolabeled M D A 2 , an oxidation-specific monoclonal antibody, provides an accurate measure of atherosclerotic lesions rich in oxidized L D L and is highly sensitive to their regression. Arterioscler Thromb Vase Biol 2000, 20:689-697. 313. Tsimikas S, Palinski W, Halpern SE, Yeung DW, Curtiss L K , Witztum JL: Radiolabeled M D A 2 , an oxidation-specific, monoclonal antibody, identifies native atherosclerotic lesions in vivo. J Nucl Cardiol 1999, 6:41-53. 314. McManus B M , Horley KJ , Wilson JE, Malcom GT, Kendall TJ, Miles RR, Winters GL, Costanzo MR, Miller L L , Radio SJ: Prominence of coronary arterial wall lipids in human heart allografts. Implications for pathogenesis of allograft arteriopathy. A m J Pathol 1995,147:293-308. 315. Skalen K, Gustafsson M , Rydberg EK, Hulten L M , Wiklund O, Innerarity TL, Boren J: Subendothelial retention of atherogenic lipoproteins in early atherosclerosis. Nature 2002,417:750-754. 190 316. Ruef J, Rao G N , L i F, Bode C, Patterson C, Bhatnagar A , Runge MS: Induction of rat aortic smooth muscle cell growth by the lipid peroxidation product 4-hydroxy-2-nonenal. Circulation 1998, 97:1071-1078. 317. Kakishita H , Hattori Y : Vascular smooth muscle cell activation and growth by 4-hydroxynonenal. Life Sci 2001, 69:689-697. 318. Koba S, Pakala R, Watanabe T, Katagiri T, Benedict CR: Synergistic interaction between thromboxane A2 and mildly oxidized low density lipoproteins on vascular smooth muscle cell proliferation. Prostaglandins Leukot Essent Fatty Acids 2000, 63:329-335. 319. Watanabe T, Pakala R, Katagiri T, Benedict CR: Lipid peroxidation product 4-hydroxy-2-nonenal acts synergistically with serotonin in inducing vascular smooth muscle cell proliferation. Atherosclerosis 2001, 155:37-44. 320. Watanabe T, Pakala R, Katagiri T, Benedict CR: Synergistic effect of urotensin II with mildly oxidized L D L on D N A synthesis in vascular smooth muscle cells. Circulation 2001,104:16-18. 321. Sue I, Meilhac O, Lajoie-Mazenc I, Vandaele J, Jurgens G, Salvayre R, Negre-Salvayre A: Activation of EGF receptor by oxidized LDL. Faseb J 1998, 12:665-671. 322. Ruef J, Hu Z Y , Yin L Y , Wu Y , Hanson SR, Kelly A B , Harker L A , Rao G N , Runge MS, Patterson C: Induction of vascular endothelial growth factor in balloon-injured baboon arteries. A novel role for reactive oxygen species in atherosclerosis. Circ Res 1997, 81:24-33. 323. Ruef J, Moser M , Bode C, Kubler W, Runge MS: 4-hydroxynonenal induces apoptosis, NF-kappaB-activation and formation of 8-isoprostane in vascular smooth muscle cells. Basic Res Cardiol 2001, 96:143-150. 191 324. Hoff HF, J ON, Chisolm G M , 3rd, Cole TB, Quehenberger O, Esterbauer H, Jurgens G: Modification of low density lipoprotein with 4-hydroxynonenal induces uptake by macrophages. Arteriosclerosis 1989, 9:538-549. 325. Hoff HF, Cole TB: Macrophage uptake of low-density lipoprotein modified by 4-hydroxynonenal. An ultrastructural study. Lab Invest 1991, 64:254-264. 326. Hoff HF, Whitaker TE, J ON: Oxidation of low density lipoprotein leads to particle aggregation and altered macrophage recognition. J Biol Chem 1992, 267:602-609. 327. Camejo G, Hurt E, Thubrikar M , Bondjers G: Modification of low density lipoprotein association with the arterial intima. A possible environment for the antiatherogenic action of beta-blockers. Circulation 1991, 84:VI17-22. 328. Hurt-Camejo E, Camejo G, Rosengren B, Lopez F, Ahlstrom C, Fager G, Bondjers G: Effect of arterial proteoglycans and glycosaminoglycans on low density lipoprotein oxidation and its uptake by human macrophages and arterial smooth muscle cells. Arterioscler Thromb 1992, 12:569-583. 329. Llorente-Cortes V , Martinez-Gonzalez J, Badimon L: L D L receptor-related protein mediates uptake of aggregated L D L in human vascular smooth muscle cells. Arterioscler Thromb Vase Biol 2000, 20:1572-1579. 330. Llorente-Cortes V , Otero-Vinas M , Hurt-Camejo E, Martinez-Gonzalez J, Badimon L: Human coronary smooth muscle cells internalize versican-modified L D L through L D L receptor-related protein and L D L receptors. Arterioscler Thromb Vase Biol 2002, 22:387-393. 331. Chang M Y , Potter-Perigo S, Wight TN, Chait A: Oxidized L D L bind to nonproteoglycan components of smooth muscle extracellular matrices. J Lipid Res 2001, 42:824-833. 332. Khan M A , Herzog CA, St. Peter JV, Hartley GG, Madlon-Kay R, Dick CD, Asinger RW, Vessey JT: The prevalence of cardiac valvular insufficiency assessed by transthoracic 192 echocardiography in obese patients treated with appetite-suppressant drugs. N Engl J Med 1998, 339:719-724. 333. Weissman NJ, Tighe JF, Jr., Gottdiener JS, Gwynne JT: An assessment of heart-valve abnormalities in obese patients taking dexfenfluramine, sustained-release dexfenfluramine, or placebo. Sustained-Release Dexfenfluramine Study Group. N Engl J Med 1998, 339:725-732. 334. Burger A J , Sherman HB, Charlamb MJ, Kim J, Asinas L A , Flickner SR, Blackburn GL: Low prevalence of valvular heart disease in 226 phentermine-fenfluramine protocol subjects prospectively followed for up to 30 months. J A m Coll Cardiol 1999, 34:1153-1158. 335. Kancherla M K , Salti HI, Mulderink TA, Parker M , Bonow RO, Mehlman DJ: Echocardiographic prevalence of mitral and/or aortic regurgitation in patients exposed to either fenfluramine-phentermine combination or to dexfenfluramine. Am J Cardiol 1999, 84:1335-1338. 336. Lepor N E , Gross SB, Daley WL, Samuels B A , Rizzo MJ , Luko SP, Hickey A , Buchbinder N A , Naqvi TZ: Dose and duration of fenfluramine-phentermine therapy impacts the risk of significant valvular heart disease. A m J Cardiol 2000, 86:107-110. 337. Teramae C Y , Connolly H M , Grogan M , Miller FA: Diet drug-related cardiac valve disease: the Mayo Clinic echocardiographic laboratory experience. Mayo Clin Proc 2000, 75:456-461. 338. Fisher JE, Orlandi Q, Ruden R, Price D, Istfan N , Joziatis BT: Valvular regurgitation in subjects receiving anti-obesity agents: A consecutive series with follow-up. J A m Soc Echocardiogr 1999,12:370. 193 339. Kimmel SE, Keane M G , Crary JL, Jones J, Kinman JL, Beare J, Sanimel M , Sutton MS, Strom B L : Detailed examination of fenfluramine-phentermine users with valve abnormalities identified in Fargo, North Dakota. A m J Cardiol 1999, 84:304-308. 340. Gardin J M , Schumacher D, Constantine G, Davis K D , Leung C, Reid CL: Valvular abnormalities and cardiovascular status following exposure to dexfenfluramine or phentermine/fenfluramine. J A M A 2000, 283:1703-1709. 341. Wee CC, Phillips RS, Aurigemma G, Erban S, Kriegel G, Riley M , Douglas PS: Risk for , valvular heart disease among users of fenfluramine and dexfenfluramine who underwent echocardiography before use of medication. Ann Intern Med 1998,129:903-905. 342. Jick H, Vasilakis C, Weinrauch L A , Meier CR, Jick SS, Derby LE: A population-based study of appetite-suppressant drugs and the risk of cardiac-valve regurgitation. N Engl J Med 1998, 339:719-724. 343. Eichelberger JP, Pearson TA, Gajary E: Fifteen year outcome data on patients treated with fenfluramine-phenermine combination. J A m Soc Echocardiogr 1999, 12:355. 344. Mast S, Jollis J, Ryan T, Landolfo C, Crary J: The natural histology of fenfluramine-associated valvulopathy assessed by echocardiography. J A m Coll Cardiol 2000, 35:523A. 345. Cannistra L B , Cannistra AJ : Regression of multivalvular regurgitation after the cessation of fenfluramine and phentermine treatment. Neuroreport 1998, 9:2989-2993. 346. Hensrud DD, Connolly H M , Grogan M , Miller FA, Bailey KR, Jensen M D : Echocardiographic improvement over time after cessation of use of fenfluramine and phentermine. Mayo Clin Proc 1999, 74:1191-1197. 347. Mancini M C , Leite CC, Grinberg M , Halpern A: Regression of heart-valve abnormalities after discontinuation of dexfenfluramine treatment. Int J Obes Relat Metab Disord 1999, 23:S178. 194 348. Ryan D H , Bray G A , Helmcke F, Sander G, Volaufova J, Greenway F, Subramaniam P, Glancy DL: Serial echocardiographic and clinical evaluation of valvular regurgitation before, during, and after treatment with fenfluramine or dexfenfluramine and mazindol or phentermine. Obes Res 1999, 7:313-322. 349. Weissman NJ, Tighe JF, Jr., Gottdiener JS, Gwynne JT: Prevalence of valvular-regurgitation associated with dexfenfluramine three to five months after discontinuation of treatment. J A m Coll Cardiol 1999, 34:2088-2095. 350. Gardin JM, Weissman NJ, Leung C, Panza JA, Fernicola D, Davis K D , Constantine G, Reid C: One year echocardiographic follow-up of patients treated with anorexigens. Int J Obes Met Dis 2000, 24:S188. 351. Mast ST, Jollis JG, Ryan T, Anstrom KJ , Crary JL: The progression of fenfluramine-associated valvular heart disease assessed by echocardiography. Ann Intern Med 2001, 134:261-266. 352. Jollis JG, Landolfo CK, Kisslo J, Constantine GD, Davis K D , Ryan T: Fenfluramine and phentermine and cardiovascular findings: effect of treatment duration on prevalence of valve abnormalities. Circulation 2000, 101:2071-2077. 353. Asinger RW: The fen-phen controversy: is regression another piece of the puzzle? Mayo Clin Proc 1999, 74:1302-1304. 354. Garon CF, Oury JH, Duran C M : Virus-like particles in the mitral and tricuspid valves explanted from a patient treated with fenfluramine-phentermine. J Heart Valve Dis 1999, 8:232. 355. Fowles RE, Cloward TV, Yowell RL: Endocardial fibrosis associated with fenfluramine-phentermine. N Engl J Med 1998, 338:1316. 356. Volmar K E , Hutchins G M : Aortic and mitral fenfluramine-phentermine valvulopathy in 64 patients treated with anorectic agents. Arch Pathol Lab Med 2001,125:1555-1561. 195 357. Hanson TP, Edwards BS, Edwards JE: Pathology of surgically excised mitral valves. One hundred consecutive cases. Arch Pathol Lab Med 1985, 109:823-828. 358. Olson LJ , Subramanian R, Ackermann D M , Orszulak TA, Edwards WD: Surgical pathology of the mitral valve: a study of 712 cases spanning 21 years. Mayo Clin Proc 1987, 62:22-34. 359. Anderson AS, Krauss D, Lang R: Cardiovascular complications of malignant carcinoid disease. Am Heart J 1997, 134:693-702. 360. McMillan JB, Lev M : The aging heart. II. the valves. J Gerontol 1964,19:1-14. 361. Schenk K E , Heinze G: Age-dependent changes of heart valves and heart size. Recent Adv Stud Cardiac Struct Metab 1975,10:617-624. 362. Sahasakul Y , Edwards WD, TSIaessens JM, Tajik AJ: Age-related changes in aortic and mitral valve thickness: implications for two-dimensional echocardiography based on an autopsy study of 200 normal human hearts. A m J Cardiol 1988, 62:424-430. 363. Roberts WC: Valvular, subvalvular and supravalvular aortic stenosis: morphologic features. Cardiovasc Clin 1973, 5:97-126. 196 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items