Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Synthesis and deposition of proteoglycans in fibroproliferative lung disease Burke, Adrian Kevin 1999

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

Item Metadata

Download

Media
831-ubc_1999-0405.pdf [ 9.42MB ]
Metadata
JSON: 831-1.0088954.json
JSON-LD: 831-1.0088954-ld.json
RDF/XML (Pretty): 831-1.0088954-rdf.xml
RDF/JSON: 831-1.0088954-rdf.json
Turtle: 831-1.0088954-turtle.txt
N-Triples: 831-1.0088954-rdf-ntriples.txt
Original Record: 831-1.0088954-source.json
Full Text
831-1.0088954-fulltext.txt
Citation
831-1.0088954.ris

Full Text

SYNTHESIS AND DEPOSITION OF PROTEOGLYCANS IN FIBROPROLIFERATIVE LUNG DISEASE by A D R I A N K E V I N B U R K E B.Sc, The University of British Columbia, 1991 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES THE FACULTY OF MEDICINE (Department of Medicine; Experimental Medicine Program) We accept this thesis as conforming to the required standard The University of British Columbia June, 1999 ©Adrian Kevin Burke, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date DE-6 (2/88) 11 Abstract Idiopathic pulmonary fibrosis (IPF) is characterized by chronic inflammation leading to progressive deposition of collagenous extracellular matrix. Proteoglycan metabolism may be altered in inflammation and fibrosis. The aim of this study is to characterize changes in proteoglycan synthesis and deposition in the remodeling lung. Lung biopsy tissue from 6 IPF and 6 control patients was studied using histochemistry and immunohistochemistry. Histochemistry revealed dense deposits of glycosaminoglycan in fibroblast foci unique to remodeling lung, which enzyme digestion showed to contain predominantly chondroitin sulfate/dermatan sulfate. Immunohistochemistry showed that glycosaminoglycan deposition was concordant with versican, but not decorin, biglycan or hyaluronan. The distribution of versican, decorin, biglycan and hyaluronan were described in the normal and IPF lung for the first time. Versican in IPF was associated with a-smooth muscle actin (a-SMA)-posit ive myofibroblasts migrating into airspace and myofibroblasts in fibroblast foci. In control and diseased lung, versican was associated with smooth muscle in blood vessel walls, in airway walls and with oc-SMA-positive cells in alveolar entrance rings. 2 2 . 6 % ± 4 . 1 % of tissue area in IPF biopsies was versican-positive, compared to 2.6±2.0% of tissue area in control lung (p<0.00001). Decorin was localized intracellularly in the fibroblast foci. Hyaluronan was found throughout the remodeling lung including the fibroblast foci. Morphometry showed higher mononuclear phagocyte densities in versican-rich fibroblast foci in the IPF lung. Mononuclear phagocyte densities were 3.74±1.14 cells/ 10 4 um 2 in versican-4 2 rich matrix and 1.97±0.54 cells/10 urn in versican-poor matrix (p<0.01). Total non-mononuclear phagocytes leukocyte densities were 5.97±1.05 cells/ 10 4 um 2 in versican-rich matrix and 9.34±1.21 cel ls /10 4 um 2 in versican-poor matrix (p<0.003) in IPF. In a second series of tissue I l l samples, versican was localized to thickened alveolar walls in fibroproliferative adult respiratory distress syndrome and intraluminal buds in bronchiolitis obliterans organizing pneumonia, and versican deposition preceded accumulation of collagen. Versican m R N A expression was analyzed in 10IPF biopsies and 5 control lung samples collected prospectively. Reverse transcriptase-polymerase chain reaction showed a relative increase in the synthesis of the two m R N A splice variants coding for the versican isoforms Vo and V ( , in fibrosis. Vo and V i are the largest isoforms, with the highest number of potential glycosaminoglycan-attachment sites. Versican is associated with proliferating, collagen-synthesizing myofibroblasts in early lesions in fibrosis. The function of versican in these processes is still unknown, but our work suggests it may be important in the cell biology of tissue remodeling in the lung. iv Table of Contents Abstract i i List of Tables v i i i List of Figures ix Abbreviations Defined x Acknowledgements x i i C H A P T E R 1: Introduction 1 1.1 Extracellular Matrix in the Human Lung 1 1.2 Interstitial Lung Disease 2 1.3 Diffuse Alveolar Damage (associated with Adult Respiratory Distress Syndrome) 5 1.4 Bronchiolitis Obliterans Organizing Pneumonia 5 1.5 Ce l l Biology of Fibrosis 6 1.6 Proteoglycan Structure 10 1.7 Proteoglycans in Inflammation and Fibrosis in the Lung 13 1.8 Hypothesis 15 1.9 Materials and Methods 16 1.9.1 Patient Samples, Series 1 - JPF and control tissue samples for immunohistochemical analysis 16 1.9.2 Patient Samples, Series 2 - A R D S , B O O P , IPF and control tissue samples for immunohistochemical analysis 17 1.9.3 Frozen Tissue Samples, Series 3 - Fibrosis and control tissue samples for molecular biological studies 18 1.9.4 Histochemistry and immunohistochemistry 19 C H A P T E R 2: Glycosaminoglycan Histochemistry 20 2.1 Glycosaminoglycan localization in Series 1 patient samples 20 V 2.2 Glycosaminoglycan characterization in Series 1 IPF patient samples 21 2.3 Glycosaminoglycan, collagen and elastin localization in Series 1 patient samples 22 C H A P T E R 3: Proteoglycan Immunohistochemistry 24 3.1 Introduction 24 3.2 Materials and Methods 25 3.2.1 Immunohistochemistry 25 3.2.2 Versican immunohistochemistry 26 3.2.3 Decorin immunohistochemistry 27 3.2.4 Biglycan immunohistochemistry 28 3.2.5 a-Smooth muscle actin immunohistochemistry 28 3.2.6 Quantification of immunohistochemical staining for versican 29 3.2.7 Glycosaminoglycans and collagen localization in Series 2 IPF patient samples, with the pathological patterns of D A D , B O O P and UIP 29 3.3 Results 30 3.3.1 Versican immunohistochemistry in IPF 30 3.3.2 Decorin and biglycan immunohistochemistry in IPF 31 3.3.3 a-Smooth muscle actin immunohistochemistry in IPF 33 3.3.4 Glycosaminoglycan, collagen and versican localization in Series 2 IPF patient samples, with the pathological patterns of D A D , B O O P and UIP 37 3.3.4.1 D A D Patients: 37 3.3.4.2 B O O P Patients: 38 3.3.4.3 IPF (UIP) Patients: 38 3.3.5 Quantification of immunohistochemical staining for versican in IPF 41 C H A P T E R 4: Hyaluronan 43 4.1 Introduction 43 4.2 Materials and Methods 43 4.3 Results..... 46 vi C H A P T E R 5: Leukocytes in Versican-rich and Versican-poor Matrix in JPF 49 5.1 Introduction 49 5.2 Materials and Methods 49 5.2.1 Versican immunohistochemistry 49 5.2.2 Leukocyte immunohistochemistry 50 5.2.3 Mononuclear phagocyte immunohistochemistry 51 5.2.4 Quantification of versican and leukocyte co-localization 52 5.3 Results 53 5.3.1 Versican/ leukocyte double immunohistochemistry 53 5.3.2 Versican/ mononuclear phagocyte double immunohistochemistry 54 C H A P T E R 6: Versican m R N A Synthesis in Normal Lung and Fibrosis 57 6.1 Introduction 57 6.2 Materials and Methods 58 6.2.1 P C R primer sets 58 6.2.2 Human fetal lung fibroblast culture 60 6.2.3 R N A extraction from H F L - 1 cells 63 6.2.4 Reverse transcription of H F L - 1 cell R N A 63 6.2.5 M g C l 2 optimization for P C R 64 6.2.6 R N A extraction from human tissue samples 66 6.2.7 Reverse transcription of disease and control R N A 67 6.2.8 Versican P C R on disease and control tissue samples 67 6.2.9 Sequencing of P C R products 68 6.3 Results 68 6.3.1 G A P D H R T - P C R 68 6.3.2 Total versican (Ver3/Ver3a) R T - P C R 69 6.3.3 Vo and V ] versican variant R T - P C R 69 6.3.4 V 2 versican variant R T - P C R 70 6.3.5 V 3 versican variant R T - P C R 70 C H A P T E R 7: Discussion 72 v i i 7.1 Localization and Identification of Glycosaminoglycans in the Remodeling Lung 72 7.2 Localization of Proteoglycans in the Remodeling Lung 73 7.3 Hyaluronan Localization in the Remodeling Lung 74 7.4 Semi-quantitative Analysis of Versican in the Remodeling Lung 76 7.5 Characterization of Versican in the Remodeling Lung 77 7.6 Versican Structure and Function 77 7.7 Versican Influence on Leukocyte Localization 78 7.8 Myofibroblasts and Versican Association in the Remodeling Lung. 79 7.9 Decorin and Biglycan Localization in the Remodeling Lung 80 7.10 Versican and Elastic Fiber Co-localization 81 7.11 Growth Factors Like ly to Influence the Local Synthesis of Proteoglycans and Hyaluronan in Fibrosis 82 7.12 Future Directions 83 References 85 Appendix 1: Patient Data 99 Series 1 IPF Patient Samples for Immunohistochemical Analysis: 99 Series 1 Control Patient Samples for Immunohistochemical Analysis: 99 Series 3 Patient Samples for m R N A Analysis: 100 Series 3 Control Samples for m R N A Analysis: 100 Appendix 2: List of Published Papers and Abstracts 101 Papers: 101 Abstracts: 102 v i i i List of Tables Table 1: Glycosaminoglycan Composition 11 Table 2: Structure and Function of Archetypal Proteoglycans 12 Table 3: Morphometric Analysis of Versican Deposition 41 Table 4: Leukocyte Cel l Density in Relation to Versican 54 Table 5: Mononuclear Phagocyte Ce l l Density in Relation to Versican 55 Table 6: P C R Primer Sets 59 Table 7: Optimal MgCL) Concentrations for P C R Primer Sets 65 ix List of Figures Figure 1: Formation of Foci of Intraluminal Extracellular Matrix Synthesis 8 Figure 2: Toluidine Blue 0 Metachromatic Staining for Glycosaminoglycans in Idiopathic Pulmonary Fibrosis 23 Figure 3: Versican Localization in Remodeling Areas in Idiopathic Pulmonary Fibrosis and Normal Lung 34 Figure 4: Localization of Glycosaminoglycans, Versican, Decorin, Biglycan and Myofibroblasts in Idiopathic Pulmonary Fibrosis 35 Figure 5: Localization of Glycosaminoglycans, Elastic Fibers, Versican and Myofibroblasts in Idiopathic Pulmonary Fibrosis 36 Figure 6: Localization of Glycosaminoglycans, Collagen and Versican in Diffuse Alveolar Damage 39 Figure 7: Localization of Glycosaminoglycans, Collagen and Versican in Bronchiolitis Obliterans Organizing Pneumonia 40 Figure 8: Localization of Hyaluronan in Idiopathic Pulmonary Fibrosis and Normal Lung 48 Figure 9: Localization of Versican, Leukocytes and Mononuclear Phagocytes in Idiopathic Pulmonary Fibrosis 56 Figure 10: Versican Protein Structure 61 Figure 11: Structure of Versican Splice Variants 62 Figure 12: Expression of Versican m R N A from Fibroproliferative Lung Disease and Control Patients, Analyzed by R T - P C R 71 X Abbreviations Defined A P : Alkaline Phosphatase A R D S : Adult Respiratory Distress Syndrome a - S M A : Alpha Smooth Muscle Actin B A L : Bronchoalveolar Lavage B O O P : Bronchiolitis Obliterans Organizing Pneumonia B S A : Bovine Serum Albumin D M E M : Dulbecco's Modif ied Eagle Medium D T T : Dithiothreitol E B V : Epstein Barr Virus E C L : Enhanced Chemiluminescence (Amersham) E C M : Extracellular Matrix F C S : Fetal Cal f Serum G A G : Glycosaminoglycan H A : Hyaluronan (previously known as Hyaluronic Acid) H A B R : Hyaluronan Binding Region HIFCS : Heat Inactivated Fetal Cal f Serum IL : Interleukin (e.g. IL-6) IPF : Idiopathic Pulmonary Fibrosis N R S : Normal Rabbit Serum P D G F : Platelet Derived Growth Factor P G - M : Proteoglycan-Medium (chicken proteoglycan homologous to human Versican) XI P V D F : Polyvinylidene Difluoride R T - P C R : Reverse Transcriptase-Polymerase Chain Reaction SDS : Sodium Dodecyl Sulfate S D S - P A G E : Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis T B E : Tris Borate E D T A T B O : Toluidine Blue O, p H 2.5 T B S : Tris-buffered Saline, p H 7.5 TGF-J3 : Transforming Growth Factor Beta UIP : Usual Interstitial Pneumonia Xll Acknowledgements This work was funded by grants from the British Columbia Lung Association and the British Columbia Health Research Foundation to Dr. Cl ive R. Roberts. The St. Paul's Pulmonary Research Laboratory tissue registry was partly funded by a grant from the Medical Research Council of Canada to Drs. Cl ive R. Roberts and Peter Pare. I would like to thank the many people who made this project possible. A special thanks to Dr. Cl ive Roberts for his time, effort and guidance serving as my graduate supervisor. Thank you to Dr. Tony Ba i and his laboratory, especially Tracey Weir, for sharing their knowledge and experience with R T - P C R and for providing me with random hexamers and the optimized conditions for their use. Thank you to Dr. Rick Hegele and his laboratory, for supplying me with G A P D H P C R primers and the optimized conditions for their use. Thank you to Jenny Hards for her guidance and training in the techniques of histochemistry and immunohistochemistry. Thanks to the many others who provided their thoughts, insights, guidance and editorial abilities over the course of my graduate studies. Thanks to Jackie Purvis and Lee Purvis, and all the other people who encouraged and supported me in the completion this thesis and degree. I appreciate it. CHAPTER 1: Introduction 1 1.1 Extracellular Matrix in the Human Lung The basic structure of the human lung is directly related to the lung's basic function, gas exchange and transport between the bloodstream and the external environment. At it's simplest, this basic structure of the lung begins with a layer of epithelial cells in direct contact with the atmosphere with an underlying basal lamina. Beneath this epithelial basal lamina is a complex heterogeneity of extracellular molecules, including collagen, elastin and proteoglycans, maintained by fibroblasts. Underlying this extracellular zone is an endothelial basal lamina, beneath a layer of endothelial cells within the blood vessels. Gas exchange occurs across these layers as oxygen and other gases diffuse from the atmosphere into the blood, and carbon dioxide and other gases diffuse from the blood into the atmosphere. The basic architecture of the lung is critical to its function. The overall structure and function of the lung is maintained over the lifetime of the individual because of a constant, balanced synthesis and degradation of architectural components by the cells of the lung. In healthy individuals, a state of homeostasis is maintained between degradation of cells and extracellular components and their synthesis. The structure of the lung, and thereby its function is perturbed in a number of inflammatory diseases. Each involves an imbalance in the homeostasis between synthesis and degradation. For instance, excessive degradation of lung architecture is associated with emphysema. Excessive matrix deposition is associated with the interstitial lung diseases. In this study, we investigate the changes in matrix composition found in interstitial lung disease. 2 One of the functions of the extracellular matrix is to provide a basic structure or architecture. For instance, tendons consist of pure collagen, which provides tensile strength. Elastic ligaments consists primarily of elastin, which provides these structures with elasticity. In cartilage, the proteoglycan gel, through the reversible redistribution of water bound to glycosaminoglycan side chains, provides cartilage with a compressive modulus. The extracellular matrix also serves as a tool for communication between different components of tissue. Extracellular matrix components provide information to cells that influence cell proliferation, cell migration and gene expression. The composition of the extracellular matrix may influence the progression of degradation, growth and repair of tissue and perturbations of matrix synthesis may directly affect the progression of disease. In this study, we investigate the extracellular composition of lung tissue in interstitial lung disease, with a specific focus on proteoglycans. 1.2 Interstitial Lung Disease The principal pathogenic feature of interstitial lung disease is an increased and abnormal deposition of collagenous extracellular matrix, secondary to inflammation. This process is termed fibrosis. This remodeling results in impaired gas exchange, altered lung mechanics, decreased lung compliance and a decrease in lung volume. Deposition of collagen is associated with myofibroblast proliferation (Kuhn and McDonald, 1991). The pathogenic changes in pulmonary fibrosis have been classified into three groups based on the sequence of inflammatory events that precede fibrosis (Hogg, 1991). Disorders beginning with non-granulomatous inflammation give rise to the pathologic patterns termed usual interstitial pneumonia (with a clinical diagnosis of idiopathic pulmonary fibrosis), desquamative interstitial pneumonia, bronchiolitis obliterans organizing pneumonia and organizing diffuse alveolar damage. Fibrotic 3 disorders that begin with a granulomatous inflammatory reaction include sarcoidosis, sarcoid-like lung fibrosis and extrinsic allergic alveolitis. Finally, fibrotic disorders resulting from a stromal reaction of the interstitium to neoplastic infiltration include eosinophilic granuloma, lymphangiomyomatosis and lymphoproliferative disease. Idiopathic pulmonary fibrosis (EPF) is an interstitial lung disease of insidious onset. IPF is characterized by inflammation preceding fibroblast proliferation and collagen deposition, and a pattern of diffuse, patchy fibrosis in which normal lung, fibroproliferative lesions and areas of replacement of architecture with collagenous matrix may be found in the same field; this pathological pattern is termed usual interstitial pneumonia (UIP). Altered lung architecture results in ventilation-perfusion mismatching and loss of former airspace, leading directly to the clinically observed restrictive lung disease symptoms, such as shortness of breath. IPF shows no sex predominance and has a broad age incidence, with most patients being between 40 and 70 years old (review in Fraser et al, 1991). The causes of some interstitial lung diseases are known, such as the inhalation of mineral particulates or gases, and as a side effect of specific xenobiotic drug treatments, but in a significant proportion of interstitial lung diseases the precipitating cause of lung injury is unknown and the clinical diagnosis is of idiopathic pulmonary fibrosis. Immunological abnormalities, genetic factors and viral infection are believed to predispose to disease (review in Fraser et al, 1991). Interstitial pulmonary fibrosis is a critical element of a number of systemic tissue disorders, such as rheumatoid arthritis and progressive systemic sclerosis, and many IPF patients present with clinical and serological evidence of autoimmunity (Stack et al, 1965; Chapman et al, 1984). In a minority of IPF patients, familial studies (Libby et al, 1983) and studies of monozygotic twins (Solliday et al, 1973) suggest a genetic link between lung 4 inflammation and fibrosis and disturbed immunological activity. The observed inheritance pattern appears to be autosomal dominance with low penetrance (Solliday et al, 1973), but no specific gene has been identified as being responsible. In some cases, viral infection appears to initiate interstitial fibrosis. Association with specific immunoglobulins for Epstein-Barr virus ( E B V ) has been observed, but this may reflect nonspecific depression of cell-mediated immunity or a part for E B V in the etiology of IPF (Vergnon et al, 1984). The etiology of IPF may very well involve a complex combination of genetically inherited predisposition to viral or environmental damage to lung epithelium and/or a subsequent over compensatory and possibly autoimmune response to the initial stimulus. There is no satisfactory therapy to reverse the fibrotic process in interstitial lung diseases, though disease progression may be limited by corticosteroids and related therapies in some patients (Davis, 1986; Cegla et al, 1975). It has been inferred that desquamative interstitial pneumonia represents an early stage, and usual interstitial pneumonia a later stage, in the pathology of IPF (Carrington et al, 1978). A greater proportion of desquamative interstitial pneumonia patients show a favorable response to corticosteroid treatment, and this it is considered as a distinct entity, at least clinically, whether or not these two diseases are distinct in terms of pathogenesis and etiology (Carrington et al, 1978). Scadding and Hinson (1967), in a study of 16 patients with "diffuse interstitial pulmonary disease" demonstrated that both usual interstitial pneumonia and desquamative interstitial pneumonia patients show inflammation and fibrosis localized to the lung parenchyma beyond the terminal bronchioles, thickening of alveolar walls with lymphocytes and fibrosis, and the presence of mononuclear cells within the alveolar airspace. They suggest that the degree of alveolar wall thickening and numbers of macrophages in the airspace are the only histological distinctions between usual interstitial pneumonia and desquamative interstitial pneumonia and 5 that these diseases should therefore be considered as one entity, at different stages. In this study, the pathological patterns of usual interstitial pneumonia and desquamative pneumonia are considered. 1.3 Diffuse Alveolar Damage (associated with Adult Respiratory Distress Syndrome) Diffuse alveolar damage is the pathological pattern associated with the adult respiratory distress syndrome ( A R D S ) . A wide diversity of insults can result in A R D S , including acute exposure to a variety of toxins and gases, septicemia and shock, or it may be idiopathic (review in Fraser et al, 1991). In early stages of diffuse alveolar damage, Type I pneumocytes are damaged, interstitial and alveolar edema occurs and hyaline membranes form. Type II pneumocyte proliferation and progressive interstitial fibrosis are observed in non-regressing patients after a few days (Nerlich et al, 1987). Two principal observations make diffuse alveolar damage distinct from other examples of interstitial lung fibrosis. One is the very fast onset of symptoms, with interstitial pulmonary fibrosis occuring only a few days after the acute onset of A R D S , whereas EPF and other interstitial lung fibroses follow a slow, insidious progression over years. The second is that approximately half of A R D S patients respond to treatment and in some cases show complete spontaneous regression of symptoms in contrast to the refractory nature of IPF to treatment. Half of A R D S patients go on to develop pulmonary fibrosis that is indistinguishable pathologically from later stages of EPF. 1.4 Bronchiol it is Obliterans Organizing Pneumonia Bronchiolitis obliterans with organizing pneumonia (BOOP) describes a pathological pattern of intraluminal buds of loose connective tissues, called Masson bodies, within respiratory 6 bronchioles, alveolar ducts and alveoli. Obliteration of alveolar airspace, interstitial fibrosis and chronic inflammation are described in the parenchyma adjacent to the affected bronchioles (review in Fraser et al, 1991). B O O P lesions within the lung appear to be of even age and progression and the condition is thought to be the direct result of acute epithelial injury, but localized to the bronchiolar and peribronchiolar epithelium (Myers and Katzenstein, 1988). B O O P patients show gas exchange impairment and the radiographic appearance of interstitial pneumonia: Pulmonary function tests demonstrate that B O O P is a restrictive rather than obstructive disease. B O O P patients frequently respond well to corticosteroid treatment, in contrast to IPF patients (Izumi et al, 1992). 1.5 Cell Biology of Fibrosis Historically, lung fibrosis, and the increase in interstitium volume relative to airspace, has been regarded as a thickening of the interstitium. It is now known, however, that other related processes are involved, including synthesis of new interstitium in former airspace, collapse of the alveoli and tissue contraction (reviewed by Kuhn, 1991). The process is analogous to wound healing, except that lung injury continues to recur throughout the development of the disease, until the healthy lung tissue is obliterated by a collagen-rich, pathological architecture. Type I and Type ITJ collagen are well documented in fibrosis, with increases in Type m collagen occuring in the early stages of lesion development and increased deposition of Type I collagen occuring in mature fibrosis. The inflammatory response in lung fibrosis, which plays a key role in the development of disease, results in the release of oxidants from both damaged epithelial cells and lung inflammatory cells. The oxidants act as chemoattractants for further inflammatory cells, 7 resulting in a positive feedback loop that leads to further tissue destruction and fibrosis (Cantin et al, 1987). The basic cellular processes in the pathology of IPF are clearly described. The sequence of events leading to the development of mature fibrosis in IPF appears to begin with damage to the lung epithelium (Fig. 1). Epithelium is damaged by either the initiating agent or the resulting inflammatory process. Perforation of the underlying basement membrane is associated with inflammation (Raghu et al, 1985). Initial injury to the lung epithelium and denudation of the underlying basement membrane is followed by exudation of proteinaceous interstitial fluid into the airspace and hyaline membrane formation. This exudate includes fibrinogen, fibronectin and other plasma components. Release of the exudate into the individual alveoli dilutes the lung surfactant normally present, reducing the surface tension and causing partial collapse of the individual alveoli around the alveolar ducts, while retaining the architecture of the ducts themselves. The fibrinogen-rich alveolar exudate is consolidated into a fibrin clot, containing influxing inflammatory cells, including alveolar and interstitial macrophages. Hyaline membranes form over the exudate, at the surface where the exudate contacts the remaining airspace of the alveolar ducts. Type II pneumocytes, the principal cells responsible for re-epithelialization of damaged lung surface, are then able to proliferate and migrate over the exudate, forming a new epithelial layer (Fig. 1), effectively delineating those areas of active fibrosis from relatively normal appearing lung in the patient. 8 Figure 1: Formation of Foci of Intraluminal Extracellular Matrix Synthesis Denuded Alveolar Basal Lamina Healthy Epithelial Injury Fibroblast Proliferation Parenchyma and Airspace Exudation and Connective Tissue Synthesis Figure 1: Development of interstitial fibrosis begins with an initial injury to the epithelium and denudation of the underlying basement membrane. Exudation of proteinaceous interstitial fluid into the airspace and hyaline membrane formation is followed by migration of proliferating fibroblasts and myofibroblasts into the former airspace at sites of destruction of the basal lamina. Type II pneumocyte hyperproliferation occurs at points of former hyaline membrane formation. At later stages of the process, fibroblasts and myofibroblasts synthesize and deposit a collagen-rich extracellular matrix, the hallmark of lung fibrosis, in the former airspace. The former airspace is thus transformed into a part of the expanded interstitium of the fibrotic lung. (Source: Kuhn et al, 1989) Concomitant with the exudation and consolidation of interstitial fluid in the alveolar airspace is the migration of proliferating fibroblasts and myofibroblasts into the former airspace at sites of destruction of the basal lamina (Kuhn et al, 1989; Kuhn and McDonald , 1991). Fibroblasts and myofibroblasts begin synthesizing type I procollagen and fibronectin, consolidating the fluid filled airspace into new interstitium (Kuhn and McDonald , 1991). The 9 original interstitium is also thickened as a result of proliferating fibroblasts and myofibroblasts migrating to the local area of the lung and synthesizing extracellular matrix components. Inflammation, and therefore epithelial damage, and tissue repair are recurring, chronic events, resulting in a diffuse, patchy pattern of inflammation and fibrosis throughout the lung. This cycle of recurring epithelial damage, inflammation and matrix deposition discriminates pulmonary fibrosis from repair and is responsible for the progressive, irreversible course of the disease. Thus, the interstitial space of the diseased lung is expanded, relative to airspace, by a thickening of the interstitium, the partial collapse of the alveoli, the synthesis and deposition of extracellular matrix in former airspace and active contraction of fibroblasts in the former airspace, which becomes obliterated (see reviews in Kuhn, 1991 and Kuhn and McDonald, 1991). Control of this pathological process is essential for future therapy. Kuhn and McDonald (1991) showed that the myofibroblasts in the subepithelial fibroblast foci, which retain some characteristics of fibroblasts (vimentin positive and desmin negative), express alpha smooth muscle actin. This cell type is motile and is able to migrate into previously non-interstitial zones of tissue and begin synthesis of new extracellular matrix. It has been demonstrated that the alpha smooth muscle actin-positive myofibroblasts form highly aligned, actin filament bundles and are contractile (Adler et al, 1981; Adler et al, 1989; Zhang et al, 1994). These cells and parallel filament bundles are linked in a matrix of fibronectin containing fibrils. Similar morphology is seen in the contraction phase of wound healing, suggesting that wound contraction plays a part in the remodeling of the lung in pulmonary fibrosis. The expression of platelet-derived growth factor (PDGF) is increased in IPF (Martinet et al, 1987), as is transforming growth factor-f3 (TGF-[3). P D G F stimulates cellular proliferation; 10 T G F - p strongly influences the synthesis of extracellular matrix. P D G F has been localized to the fibroblast foci in IPF (Limper et al, 1991). Deposition of TGF-(3 in the fibrotic regions, as well as those areas undergoing active repair, of the IPF lung has also been demonstrated (Khali l et al, 1991 a, b). TGF-(3 greatly increases the synthesis of collagen, fibronectin and other extracellular matrix proteins by fibroblasts in vitro (Ignotz and Massague, 1986). Cytokines may play complex roles in the synthesis of extracellular matrix in the fibrotic lung and influence the synthesis and deposition of glycosaminoglycans and proteoglycans in IPF. 1.6 Proteoglycan Structure Proteoglycans are a group of proteins to which are covalently bound a class of carbohydrates called glycosaminoglycans. Glycosaminoglycans are linear polymers of repeating disaccharide units, consisting of one hexosamine and one carboxylate or sulfate ester. In contrast to the carbohydrates attached to other glycoproteins, glycosaminoglycans are heavily negatively charged at neutral p H . The different classes of glycosaminoglycans are defined by the composition of the individual subunits in each disaccharide unit, as described in the following table: 11 Table 1: Glycosaminoglycan Composition Amino Sugar N-acetyl-D-glucosamine N-acetyl-D-galactosamine D-glucuronic acid Hyaluronan (HA) Chondroitin Sulfate (CS) D-glucuronic acid or L-iduronic acid* Heparan Sulfate (HS) and Heparin** Dermatan Sulfate (DS) D-galactose Keratan Sulfate (KS) * L-iduronic acid is produced in the mature glycosaminoglycan chain by the epimerization of the carboxyl group of D-glucuronic acid, such that the resultant polysaccharide chain contains a proportion of each uronic acid. Thus, dermatan sulfate is derived from, and is a modified form of chondroitin sulfate. ** The difference in composition between heparan sulfate and heparin is in the degree of sulfation, heparan sulfate containing 0.2-2.0 sulfates and heparin containing 2.0-3.0 sulfates, per disaccharide unit. In tissues, all glycosaminoglycans except hyaluronan are found covalently linked to a diverse group of proteins, the proteoglycans. It is apparent, based on primary sequence data, that there are a number of distinct families of proteoglycans. Glycosaminoglycans are synthesized on serine residues in the core proteins of proteoglycans at serine-glycine pairs, chondroitin sulfate, heparan sulfate and dermatan sulfate being covalently O-linked to the serine residue. Chondroitin and dermatan sulfate may also be synthesized, but much less efficiently, on threonine residues (Mann et al, 1990). Keratan sulfate side chains are linked covalently to the proteoglycans by an N-linkage to asparagine residues or by an O-linkage to threonine residues (Barry et al, 1995). Hyaluronan, which ranges in size from 2 M D a to 8 M D a , is not found covalently bound to protein and occurs throughout the interstitium in hydrated matrices. Proteoglycans have been classed according to the families of related core proteins or by the class(es) of glycosaminoglycans with which they are substituted. Most tissues contain more than one type of proteoglycan. The structure and function of the archetypal proteoglycans is described in Table 2. 12 Table 2: Structure and Function of Archetypal Proteoglycans Archetypal Proteoglycan (Reference) Proteoglycan Family (Other Family Members) Core Protein Size and Characteristics Glycosamino-glycan Type Function Aggrecan (Doege et al, 1991) large aggregating proteoglycans (family includes versican) 220 kDa, N-terminal domain binds hyaluronan, central keratan sulfate-substituted domains, C-terminal epidermal growth factor-like, lectin-like, and complement regulatory protein-like domains ~ 100 X chondroitin sulfate, variable keratan sulfate water-binding properties generate osmotic swelling pressure; influences tissue compressive mechanics Decorin (Fisher et al, 1989) small interstitial proteoglycans (family includes biglycan) 38 kDa, 11 tandem leucine-rich repeats 1 chondroitin/ dermatan sulfate collagen fibril assembly, transforming growth factor p localization in matrix, modulation of cell-matrix adhesion (anti-adhesive) Perlecan (Murdoch et al, 1992) basement membrane heparan sulfate proteoglycans 467 kDa, low density lipoprotein receptor homology, laminin A chain homology, epidermal growth factor-like homology 6 X heparan sulfate assembly and maintenance of basement membranes, cell binding, growth factor localization in matrix Syndecan (Bernfield et al, 1990) cell surface heparan sulfate proteoglycans; syndecan-like integral membrane proteoglycans "SL IPs" (4 members) 31 kDa, Thr-ser-pro-rich domain (mucin homology), transmembrane domain, cytoplasmic domain with conserved tyr residues 2 X chondroitin sulfate, 3 X heparan sulfate binds various proteins to cell surface through heparan sulfate chains, including basic fibroblast growth factor, collagens, fibronectin. Intracellular domain associates with cytoskeleton Glypican (David, 1993) glypican-like integral membrane proteoglycans "GRIPs" (2 members) 62 kDa, C-terminus linked to glycosyl phosphatidylinositol (GPI) in outer face of cell membrane 4 X heparan sulfate binds various proteins to cell surface through heparan sulfate chains, including basic fibroblast growth factor, collagens, fibronectin Table borrowed with kind permission from Dr. Cl ive Roberts (Roberts et al, 1997). 13 1.7 Proteoglycans in Inflammation and Fibrosis in the Lung There is evidence for an increase in overall glycosaminoglycan content in human fibrotic lung conditions, based on biochemical analyses. Wusteman et al (1972) documented increases in all glycosaminoglycan subtypes in both progressive massive fibrosis and pleural plaques, with a greater increase in the proportion of chondroitin-6-suphate over chondroitin-4-sulfate. In contrast, Wagner et al (1975) compared progressive massive fibrosis to normal lung samples and found increases in hyaluronan, decreases in heparan sulfate and dermatan sulfate, and a static condition in chondroitin sulfate levels. Motomiya et al (1975) detected increases in dermatan sulfate in a single case of diffuse interstitial fibrosis, or idiopathic pulmonary fibrosis. These studies covered diverse stages in the development of fibrosis, so the seemingly disparate data from the previous experiments likely describe different stages in P G synthesis and deposition throughout the course of fibrotic lung disease. These studies suggested that changes in glycosaminoglycan levels accompany the morphological changes in human fibrotic lung diseases. However,, which glycosaminoglycans and which proteoglycans are specifically altered and what the effects of these changes are is not clear. The changes in glycosaminoglycan content in animal models of fibrotic lung disease are much more well documented. Karlinsky et al (1982) demonstrated increases in lung content of all glycosaminoglycan types in this model in hamsters, but found no unique changes in the levels of the glycosaminoglycan subtypes. However, Cantor et al (1980) showed specific increases per lung weight in dermatan sulfate and chondroitin-4-sulfate and decreases in heparan sulfate in this model, 1 and 3 months after initial bleomycin instillation, as well as increases in 3 5 S 0 4 uptake, representing an increase in glycosaminoglycan synthesis. The increases in glycosaminoglycan 14 synthesis in bleomycin-induced lung fibrosis were confirmed and further characterized by later studies in hamsters demonstrating a rise in glucosamine 6-phosphate synthetase activity, a major enzyme involved in the synthesis of glycosaminoglycans (Yoshida et al, 1982). Cantor et al (1983 a, b) documented a maximal glycosaminoglycan increase at day 5 after bleomycin instillation, with levels returning to control levels after 45 days. They found particular increases in chondroitin-4-sulfate early in the model and large increases in dermatan sulfate at later stages of the fibrotic changes, and overall decreases in heparan sulfate levels. Similar work in experimental silicosis, by quartz crystal instillation, (V i l im and Hurych, 1987) found a specific increase in dermatan sulfate and a gross increase in S O 4 incorporation that plateaus 20 weeks after instillation. Juul et al (1993) showed an increase in a chondroitin sulfate proteoglycan and the glycosaminoglycan hyaluronan in experimental acute hyaline membrane disease in Macaques. The free glycosaminoglycan hyaluronan has been shown to accumulate in experimental models of pulmonary fibrosis. Hyaluronan accumulation during the development of bleomycin-induced pulmonary fibrosis in rats is evident in bronchoalveolar lavage fluid (Nettelbladt and Hallgren, 1989) and in alveolar interstitial tissue (Nettelbladt, Bergh et al, 1989). The accumulation of hyaluronan in experimental pulmonary alveolitis and fibrosis was shown to occur at early stages of the process, paralleling the development of pulmonary edema (Nettelbladt, Tengblad and Hallgren, 1989) and prior to the development of mature fibrotic lesions (Hernnas et al, 1992). Increase in hyaluronidase activity, an enzyme that degrades hyaluronan, follows the early accumulation of hyaluronan in bleomycin-induced alveolitis in hamsters (Bray et al, 1991). Thus, the synthesis and degradation of hyaluronan is likely to be a factor in early stages of tissue remodeling in pulmonary alveolitis and fibrosis. 15 Hyaluronan appears to play a functional role in morphogenesis in development and repair. Increased hyaluronan synthesis by proliferating cells in culture and hyaluronan enrichment in proliferating zones in vivo, in embryonic limb development and in regenerating tissue, has been documented (see review in Toole, 1991). Some cells migrate through hydrated, hyaluronan-rich domains by interaction between hyaluronan and CD44, a cell surface receptor for hyaluronan (Thomas et al, 1993). The receptor for hyaluronan is increased in vitro in proliferating cells (Alho and Underhill, 1989). The involvement of hyaluronan in the cell biology of development makes it possible that hyaluronan has a functional role in both the inflammatory and tissue remodeling stages of interstitial lung disease. 1.8 Hypothesis These findings, in the human condition and experimental models of fibrotic lung disease, along with demonstrations of specific functions for glycosaminoglycans and proteoglycans in changes in tissue architecture, lead us to hypothesize that glycosaminoglycans and proteoglycans may serve an important functional role in the morphological changes during human lung fibrosis. M y aim is to characterize the deposition and potential roles of glycosaminoglycans and proteoglycans in the development of human lung fibrosis, specifically in idiopathic pulmonary fibrosis, and to begin to characterize the involvement of glycosaminoglycans and specific proteoglycans in the cell biology of the fibrotic process. 16 1.9 Materials and Methods 1.9.1 Patient Samples, Series 1 - IPF and control tissue samples for immunohistochemical analysis Tissue samples, both biopsy and autopsy specimens, from individuals with a clinical history consistent with idiopathic pulmonary fibrosis (IPF) and a pathologic diagnosis of usual interstitial pneumonia (UIP) were obtained from the Lung Registry of the U B C Pulmonary Research Laboratory at St. Paul's Hospital. Series 1 patients ranged in age from 38 to 83 years (mean 66 yr., S D 15 yr.'). Patient data, diagnoses and duration of symptoms, typically reported by the patient as "shortness of breath", are given in Appendix 1. Tissue from control, age-matched patients with normal lung function and no history of interstitial lung disease, for whom removal of a lobe or whole lung was necessary to remove a small peripheral tumor, was obtained from the Lung Registry of the U B C Pulmonary Research Laboratory at St. Paul's Hospital. Series 1 control patients, selected to match the age distribution of the IPF patient group, ranged in age from 42 to 81 years (mean 66 yr., S D 14 yr.). Tissue blocks of histologically normal parenchyma, well away from the tumor site, were selected for these control studies. Control patient data is given in Appendix 1. Biopsy specimens from three individuals with a pathologic pattern of desquamative interstitial pneumonia were studied immunohistochemically but were not entered into the quantitative morphometric studies. 17 1.9.2 Patient Samples, Series 2 - ARDS, BOOP, IPF and control tissue samples for immunohistochemical analysis Based on the results of initial studies of the patients in Series 1, we selected patients so that we could compare the remodeling process in patients with fibroproliferative diffuse alveolar damage associated with adult respiratory distress syndrome, idiopathic bronchiolitis obliterans organizing pneumonia and idiopathic pulmonary fibrosis. Series 2 tissue samples were obtained from the Lung Registry of the U B C Pulmonary Research Laboratory at St. Paul's Hospital. These tissue samples had been entered into the registry over the period of 1984 to 1994. Cases with a histologic diagnosis of diffuse alveolar damage (DAD) (n=7), bronchiolitis obliterans organizing pneumonia (BOOP) (n=5) or idiopathic pulmonary fibrosis (IPF) (n=5) were entered into the study after all cases with a specific etiologic diagnosis were excluded. Lung tissue from five patients with normal lung function who underwent lung resection for a peripheral lung tumor served as controls. Tissue from these control cases was obtained from an area distant to the tumor. Six of the seven cases with a histologic diagnosis of diffuse alveolar damage and a clinical diagnosis of adult respiratory distress syndrome had a brief respiratory illness of possible viral origin prior to the onset of respiratory failure while one case was associated with a drug overdose. The duration of symptoms at the time of the biopsy varied from 7 to 21 days (mean 13 days); duration of mechanical ventilation was 3 to 10 days (mean 6 days). The five cases of histologic diagnosis of bronchiolitis obliterans organizing pneumonia had symptoms ranging from 35 to 56 days (mean 41 days) prior to open lung biopsy. The duration of symptoms in the usual idiopathic pneumonia group ranged for 6 months to 6 years (mean 2.3 years) prior to open lung biopsy (n=3) or autopsy (n=2). None of the bronchiolitis obliterans organizing pneumonia or 18 adult respiratory distress syndrome patients had been treated with steroids prior to biopsy. None of the idiopathic pulmonary fibrosis patients biopsied had been treated with steroids prior to biopsy. One idiopathic pulmonary fibrosis autopsy specimen was from a patient who had been unsuccessfully treated with steroids several-months before death. Results of the studies of Series 2 patients were published in 1996 (Bensadoun et al, 1996) and are appended (see List of Publications in Appendix 2). 1.9.3 Frozen Tissue Samples, Series 3 - Fibrosis and control tissue samples for molecular biological studies Biopsy tissue was collected and entered into the study prospectively from 10 individuals undergoing diagnostic biopsies that confirmed idiopathic pulmonary fibrosis (IPF), and from 5 control resected lungs from individuals with small peripheral coin lesions and normal pulmonary function tests. Idiopathic pulmonary fibrosis biopsies studied were from 4 individuals with idiopathic pulmonary fibrosis and a pathologic pattern of idiopathic pulmonary fibrosis, 4 individuals with idiopathic disease of short duration and a pattern of patchy organizing pneumonia in small airways and alveoli, corresponding to bronchiolitis obliterans organizing pneumonia, and 2 individuals with organizing diffuse alveolar damage associated with idiopathic adult respiratory distress syndrome. Samples of biopsies were obtained from the Pathology Service of St. Paul's Hospital, Vancouver, within 2 hours of excision, from patients who had not received steroid therapy. The U B C Human Ethics Committee approved sample acquisition. Tissues were fast-frozen and stored at -70°C and all were subsequently processed simultaneously. 19 1.9.4 Histochemistry and immunohistochemistry Biopsy samples were obtained 1-2 hours after biopsy, or lung resection, and fixed overnight in 10% buffered formalin. After fixation, the tissue was dehydrated in ethanol, cleared in xylol and embedded in paraffin. Serial 5mm sections from the paraffin embedded samples were mounted on 3-aminopropyltriethoxysilane coated slides, heated overnight at 37°C and stored at room temperature before use. Sections for direct comparison were processed and stained together. Prior to all histochemical and immunohistochemical staining, the sections were deparaffinized in two ten minute washes of xylene and hydrated through decreasing concentrations of ethanol (2X100%, 95%, 70%, 5 minutes each) to distilled water. 20 CHAPTER 2: Glycosaminoglycan Histochemistry 2.1 G lycosaminoglycan localization in Series 1 patient samples In order to localize glycosaminoglycans histochemically, sections were stained using 0.25% Toluidine Blue 0 (Sigma, St. Louis, M O ) in 0.5% acetic acid, p H 2.5, for one hour, and destained briefly in 0.25% acetic acid, p H 2.5. Due to their high negative charge density, glycosaminoglycans stain metachromatically with this cationic stain, giving a purple color, while nuclei stain blue (Fig. 2A). The remainder of the tissue does not stain. Toluidine Blue 0 staining at p H 2.5 showed metachromasia, indicating dense deposits of glycosaminoglycans unique to remodeling areas of the IPF lung tissue (purple stain, Fig. 2A). Metachromasia occurred in the matrix in areas with the appearance of "active lesions" in the IPF lung, particularly where the remodeling process bordered airspace. The dense deposits of glycosaminoglycans are associated with highly aligned fibroblasts in a sparse matrix, underlying apparently hyperplastic epithelium (Fig. 2A) . This is characteristic of the subepithelial fibroblast foci described by Kuhn and McDonald (1991) and this represents the "active lesions" of fibrosis. Metachromasia was not observed in the interstitial matrix of normal lung, in those regions of the interstitial matrix of the IPF lung that displayed normal morphology, or even in areas of mature fibrosis characteristic of the end-stage of the fibrotic process, in the IPF lung. In normal and IPF lungs, epithelial goblet cells, mast cells, mucus and cartilage stained metachromatically, due to their high glycosaminoglycan content. 21 2.2 Glycosaminoglycan characterization in Series 1 IPF patient samples Specific glyeosaminoglycan-degrading enzymes were used to selectively degrade glycosaminoglycans in serial tissue sections in order to determine the glycosaminoglycans in the metachromatic lesions. Serial sections were pretreated with either chondroitinase A B C (specifically degrades chondroitin sulfate/ dermatan sulfate) or heparinase HI (specifically degrades heparan sulfate) prior to Toluidine Blue O staining. To remove chondroitin sulfate and dermatan sulfate, designated sections were pretreated at 37°C with concentrated Chondroitinase A B C (Sigma) in 0.1 M Tr i s -HCl , 50 m M calcium acetate, 0.01% bovine serum albumin (BSA) ( ICN Biochemical, Cleveland, OH) , p H 7.3. To degrade heparan sulfate, sections were pretreated with concentrated Heparinase HI (Sigma) in 0.1 M sodium acetate, 0.2% B S A , p H 7.0. Additional sections were incubated at 37°C with each respective buffer alone, without enzyme, to control for possible effects of buffer treatment alone. After the selective predigestion of glycosaminoglycans, the sections were stained for glycosaminoglycans with Toluidine Blue 0 at p H 2.5 as described above. Metachromatic staining of the active fibroblast foci in the IPF lung (Fig. 2A) was removed by pretreating the sections with chondroitinase A B C (Fig. 2B), but not with heparinase HI (Fig. 2C). The dense deposits of glycosaminoglycans observed in the IPF lung were selectively degraded by chondroitinase A B C treatment. Neither pretreatment with chondroitinase A B C nor heparinase HI had an effect on the Toluidine Blue 0 metachromasia observed in goblet cells, mast cells or mucus; suggesting either a heterogeneous collection of glycosaminoglycans and polysaccharides or deposition of another class of glycosaminoglycan in these areas. As an internal positive control, the metachromatic staining of airway cartilage was eliminated by 22 chondroitinase A B C pretreatment. The chondroitin sulfate-containing proteoglycan aggrecan is responsible for the Toluidine Blue 0 metachromasia observed in airway cartilage (Roberts and Pare, 1991; Roberts et al, 1998). 2.3 Glycosaminoglycan, col lagen and elastin localization in Series 1 patient samples Sections were stained using the Gomori trichrorhe and aldehyde fuchsin method (Halami, 1952) to identify areas of collagen, elastin and glycosaminoglycan deposition. With this stain, collagen stains blue-green, elastic fibers stain navy blue, nuclei stain red, red blood cells stain yellow and glycosaminoglycans stain purple. Gomori trichrome and aldehyde fuchsin staining of serial sections confirmed the presence of high concentrations of glycosaminoglycans (purple stain, Figs. 3 A & 4 A - see Chapter 3) in the fibroblast foci that stained metachromatically with Toluidine Blue 0, pH 2.5. The Gomori stain also demonstrated heavy collagen deposition surrounding the fibroblast foci (blue-green stain, Figs. 3 A & 4A) , with much lower collagen concentrations within the foci. In some sections studied, the dense deposits of glycosaminoglycans (purple stain, Fig . 5D - see Chapter 3) co-localized with abnormal-appearing elastic fibers (navy blue stain, Fig . 5D). Thus, this study shows for the first time that glycosaminoglycan deposition is specifically associated with the fibroblast foci, and that the glycosaminoglycan is predominantly of the chondroitin sulfate/ dermatan sulfate family. 23 Figure 2: Toluidine Blue 0 Metachromatic Staining for Glycosaminoglycans in Idiopathic Pulmonary Fibrosis Figure 2: Serial sections of IPF tissue, stained with toluidine blue O at p H 2.5. Section A shows a focal area of myofibroblasts in glycosaminoglycan-rich interstitium (purple metachromasia, asterix). Section B was predigested with chondroitinase A B C , and Section C was predigested with heparinase III, before toluidine blue O staining. Loss o f metachromatic staining with chondroitinase A B C , but not with heparinase III, indicates a high density o f chondroitin sulfate and/or dermatan sulfate. Bar = 100pm 24 CHAPTER 3: Proteoglycan Immunohistochemistry 3.1 Introduction In order to determine which proteoglycans are responsible for the elevated levels of chondroitin/dermatan sulfate we observed in the fibroblast foci using histochemical staining, we used immunohistochemical techniques. We investigated the deposition of versican, decorin and biglycan, three chondroitin sulfate/ dermatan sulfate proteoglycans, in the fibrotic and normal lung. We also used immunohistochemical staining for oc-smooth muscle actin to identify myofibroblasts and smooth muscle cells. Versican is a large interstitial proteoglycan with homology to the cartilage proteoglycan aggrecan. Versican has a functional hyaluronan-binding domain at its N-terminus (LeBaron et al, 1992) and a central domain to which are covalently bound a number of chondroitin sulfate (CS) side chains and N-linked oligosaccharides (Zimmermann and Ruoslahti, 1989). A t its C-terminus are two epidermal growth factor-like modules, a lectin-like module and a complement regulatory protein-like module (Zimmermann and Ruoslahti, 1989). Versican may influence the biomechanical properties of tissue such as the walls of arteries (review in Wight, 1989). Versican may act to sterically hinder interactions between cells surface receptors and extracellular matrix ligands (Yamagata et al, 1989). Versican has been observed in some developmentally active areas in chick embryos, such as perinotochordal mesenchyme, subneuroepithelial basement membrane and around condensing mesenchymal cells in limb buds (Yamagata et al, 1993). Decorin is a small interstitial proteoglycan (45kD core protein) to which is bound a single glycosaminoglycan. A proteoglycan found throughout the interstitium of the lung (Van Kuppevelt et al, 1985 a, b) and, in particular, in association with collagen fibrils (Van Kuppevelt 25 et al, 1987; review in Wight et al, 1991) is likely decorin. Decorin binds to collagen and is believed to influence collagen fibril assembly (Brown and Vogel , 1989; Bidanset et al, 1992). Decorin also binds fibronectin (Schmidt et al, 1991) and may influence the deposition of matrix in fibrosis. Decorin, through its core protein, has been shown to bind TGF-p \ neutralizing some of the cytokine's biological activities (Yamaguchi et al, 1990) and inhibiting matrix deposition in immune complex-induced kidney inflammation and fibrosis (Border at al, 1992). Biglycan is highly homologous to decorin (Fisher et al, 1983; Fisher et al, 1987) and its function remains unclear. It is a small interstitial proteoglycan and is substituted by two glycosaminoglycan chains. Fisher et al (1983; 1987) showed that biglycan is electrophoretically distinct from decorin and that their amino acid sequences are closely related. Biglycan and decorin were shown to be the products of two distinct genes, localizing to chromosomes X q l 3 -qter and 12q, respectively (McBride et al, 1990). Biglycan localizes to the surface of cells and is abundant in the adventitia and intima of blood vessels (Bianco et al, 1990). 3.2 Materials and Methods 3.2.1 Immunohistochemistry Prior to immunohistochemical staining, sections were blocked in 1 in 10 normal goat serum in 2% Bovine Serum Albumin (ICN) in Tris buffered saline, p H 7.5 for one hour. This blocking step reduces or eliminates any non-specific binding of the tissue by the secondary antibodies, which were raised in goats against the Fc fragments of non-immune IgG molecules of the primary antibody types. The concentration of normal goat serum used was optimized at 1 in 10, based on a serial dilution experiment in which we determined that this concentration of 26 blocking antibody allowed greatest primary antibody binding and least secondary antibody non-specific binding. Primary antibody incubation steps are described in sections below, specific to the target molecule. Unless otherwise mentioned, primary antibody concentrations and duration of incubations in all immunohistochemical staining procedures were first optimized in a series of serial dilution experiments, to maximize specific binding and eliminate all non-specific binding. Sections were incubated with normal rabbit serum (NRS), in place of the described primary antibodies, as negative controls. The N R S was diluted in 2% B S A in Tris buffered saline, p H 7.5 ( T B S - B S A ) to protein concentrations, measured spectrophotometrically by absorbance at 280 nm, equivalent to that of each antiserum. One hour incubations in 1 in 100 dilutions of goat anti-(mouse IgG) or anti-(rabbit IgG), as appropriate, linked to alkaline phosphatase (Bio-Rad, Hercules, C A ) were used to localize the primary antibodies. A napthol-AS-biphosphate/ new fuchsin (Sigma) color reaction, generating a red precipitate, was used to visualize the antibody-target complexes. A l l antibody incubation steps were separated by thorough washes with T B S . Immunohistochemical sections were counterstained, when appropriate, with Mayer's hematoxylin (Mayer, 1903). A l l sections were washed, dehydrated in ethanol, mounted in Entellan (BDH) and examined by light microscopy. 3.2.2 Versican immunohistochemistry Sections to be stained for versican were pretreated at 37°C with concentrated Chondroitinase A B C (Sigma) in 0.1 M Tr i s -HCl , 50 m M calcium acetate, 0.01% bovine serum 27 albumin (BSA) ( ICN Biochemical, Cleveland, OH) , p H 7.3. In preliminary experiments, this pretreatment was found to increase specific staining, consistent with the enzyme uncovering epitopes that might be masked, sterically, from antibody binding by the chondroitin sulfate side chains of this proteoglycan. Sections were stained for versican by incubation for 2 hours with a 1 in 50 dilution of rabbit antiserum in 2% B S A (ICN) in Tris buffered saline, p H 7.5. The versican antiserum (LeBaron et al, 1992) was kindly donated by Drs. Erkki Ruoslahti & Richard Le Baron, L a Jolla Cancer Research Center. The antiserum was raised against a fusion protein extracted from NTH/3T3 fibroblasts which had been transfected with p R L T 2 T - V C - l . pRIT2T-V C - 1 is a prokaryotic Protein A fusion vector containing a fragment of versican c D N A corresponding to base pairs 1425-2342 in the deduced D N A sequence of versican. This versican D N A sequence corresponds to the glycosaminoglycan attachment domain of the originally described versican protein (Zimmermann and Ruoslahti, 1989). The versican antibody was affinity purified and preabsorbed against Chinese hamster ovary cell lysates by Dr. LeBaron. 3.2.3 Decorin immunohistochemistry Sections were stained for decorin by incubation for 2 hours with 1 in 50 dilutions of the rabbit antiserum LF-30 in 2% B S A (ICN) in Tris buffered saline, p H 7.5. This antiserum was generously donated by Larry W . Fisher, N I H . LF-30 was raised against a synthetic peptide based on amino acids 5-17 in human decorin, conjugated to keyhole limpet hemocyanin (Fisher et al, 1987). The antiserum were previously shown by Fisher et al to have no crossreactivity with other proteoglycans, such as biglycan, which is closely related to decorin. Use of the LF-30 antiserum has been described in detail in previous studies (Fisher et al, 1987, Fisher et al, 1989). 28 Some sections were stained for decorin by incubation for 2 hours with 1 in 100 dilutions of the rabbit antiserum anti-PG40 in 2% B S A (ICN) in Tris buffered saline, p H 7.5. This anti-PG40 polyclonal antiserum (Krusius and Ruoslahti, 1986) was purchased through Telios. The antiserum was raised against decorin, previously known as PG40, isolated from human fetal membranes. 3.2.4 Biglycan immunohistochemistry Sections were stained for biglycan by incubation for 2 hours with 1 in 50 dilutions of the rabbit antiserum LF-51 in 2% B S A (ICN) in Tris buffered saline, p H 7.5. This antiserum was generously donated by Larry W . Fisher, NEH. LF-51 was raised against a synthetic peptide based on amino acids 11-24 of human biglycan, conjugated to bovine serum albumin (Fisher et al, 1989). LF-51 was previously shown by Fisher et al to have no crossreactivity with the previously described LF-30 , anti-decorin antiserum. Use of LF-51 has been described in detail in previous studies (Fisher et al, 1987, Fisher et al, 1989). 3.2.5 a-Smooth muscle actin immunohistochemistry Sections.were stained for alpha smooth muscle actin by incubation for 2 hours with a 1 in 100 dilution of a mouse monoclonal antibody (ascites fluid), clone 1A4 (Sigma) in 2% B S A in Tris buffered saline, p H 7.5. This antibody was raised against a synthetic peptide based on amino acids 1-10 of human alpha smooth muscle actin, conjugated to keyhole limpet hemocyanin. Sections were incubated with normal rabbit serum (NRS), in place of the 1A4 primary antibody, as a negative control. The N R S was diluted in 2% B S A in Tris buffered saline, pH 7.5 to protein concentrations, measured spectrophotometrically by absorbance at 280 nm, equivalent to that of the antiserum. 29 3.2.6 Quantification of immunohistochemical staining for versican The percentage area of IPF and normal lung tissue staining positively for versican was determined using the Bioview Colour Image Analysis System, which was developed as a collaborative effort between the Respiratory Network of Centres of Excellence, Inspiraplex and Infrascan Inc. of Richmond, B C , Canada. Five fields, selected randomly, were analyzed for each patient at 10X magnification on a Nikon Microphot-FX light microscope (Nikon, Tokyo, Japan). High resolution, true color images were captured using an analog video camera (25.4mm Vidicon, 81 series, D A G E - M T I Inc., Michigan City, IN) and displayed on a Sony Multiscan H G monitor (1024 X 1024 pixels, 24 bit, 0.23 u m 2 pixel area, Gdill-1936m, Sony, Tokyo, Japan). Versican (quantified as red precipitate from alkaline phosphatase cleavage of naphthol-AS-biphosphate/ new fuchsin substrate in immunohistochemical localization) in each microscopic field was quantified as those pixels falling within the threshold values of hue, saturation and intensity as follows: Hue range=227-255, Saturation range=97-255, Intensity range=176-255 (these thresholds defined within a 256 binary gray level scale). Total tissue area, to which versican area was directly compared (shown in Table 3 in the Results section of this chapter), in each microscopic field was quantified as those areas with red, green and blue color threshold values as follows: Red range= 134-255, Green range=l-216, Blue range=176-255 (these values again falling within a 256 value range). 3.2.7 Glycosaminoglycans and collagen localization in Series 2 IPF patient samples, with the pathological patterns of DAD, BOOP and UIP Sections were stained with hematoxylin-eosin to visualize overall architecture. 30 Sections were stained with alcian blue to localize glycosaminoglycans and picrosirius red to stain collagen. Sections were pretreated with pre-warmed alkaline alcohol (add drops of 28% ammonium hydroxide to 95% ethanol until p H is greater than pH8) for 10 minutes in a 60°C bath and then stained with 1% alcian blue in 1% acetic acid for 30 minutes, followed by a wash in 1% acetic acid. Sections were then stained in 0.0018 M Sirius Red in saturated picric acid for 1 hour and washed in 1% acetic acid. Stained sections were washed, dehydrated in ethanol, mounted in Entellan (BDH) and examined by light microscopy. 3.3 Results The tissue samples were stained for versican, decorin and biglycan, using antibodies to the core proteins of each of these three chondroitin sulfate/ dermatan sulfate proteoglycans, to identify the molecule responsible for the Toluidine Blue 0 metachromasia seen in IPF. 3.3.1 Versican immunohistochemistry in IPF The dense glycosaminoglycan staining observed in the subepithelial fibroblast foci (purple stain, Figs. 3 A & 4A) were found to correspond to dense immunoreactivity for versican in the IPF lung (red stain, Figs. 3B & 4B). In contrast, neither decorin immunoreactivity nor biglycan immunoreactivity corresponds to glycosaminoglycan histochemical staining in IPF lung. Versican immunoreactivity was most concentrated in those areas that stain metachromatically with Toluidine Blue 0, p H 2.5, the characteristic fibroblast foci of the IPF lung. Immunoreactivity between the chondroitin sulfate bound versican target and the primary antibody required pretreatment of the sections with concentrated chondroitinase A B C , which follows logically with the observed versican and chondroitin sulfate co-localization. The versican-rich fibroblast foci (Fig. 3B) in IPF are surrounded by high concentrations of collagen 31 (green/ blue stain, Fig . 3A) , but these versican-rich foci themselves contain relatively low concentrations of collagen. Versican deposition occurred on the edges of frankly fibrotic regions in the characteristic fibroblast foci (Fig. 3C) and extensively underneath apparently hyperplastic epithelium (Fig. 3D). Versican co-localized with myofibroblasts migrating into airspace (Fig. 3E), a process characteristic of the early stages of fibrotic lesions. In some sections, areas of dense versican staining (Fig. 5C) co-localized with abnormal-appearing elastic fibers observed with the Gomori stain (navy blue stain, Fig . 5D). Versican is localized to airway walls and large blood vessels in normal lung. In both normal (Fig. 3F) and IPF lungs, versican was found pericellularly in the smooth muscle bundles of airway walls and associated with smooth muscle in the intima of blood vessels. Versican was also deposited to a lesser extent in the media and adventitia of blood vessels. In normal lung parenchyma, versican was detected in thicker portions of alveolar walls, at septal tips and in alveolar entrance rings, in association with cells that stained positively for a-smooth muscle actin. Very light staining of airway cartilage was observed with the versican antibody. This may indicate detection of versican in airway cartilage, consistent with the demonstration that articular cartilage chondrocytes synthesize versican m R N A (Le Baron et al, 1992). Other parts of the interstitium of the normal lung stained very weakly for versican. 3.3.2 Decorin and biglycan immunohistochemistry in IPF Decorin is localized intracellularly in the fibroblasts, in the fibroblast foci of the IPF lung. Decorin was observed intracellularly, but not extracellularly in the highly aligned fibroblasts common to the subepithelial fibroblast foci of the IPF lung (Fig. 4D). Thus the chondroitin sulfate/ dermatan sulfate in the extracellular matrix of these lesions are not covalently linked to 32 decorin. These fibroblasts have been shown by others to be synthesizing type I collagen (McDonald et al, 1986 and Zhang et al, 1994). Decorin is associated with collagen fibrils, including those in lung (van Kuppevelt et al, 1987) and skin (Fleischmajer et al, 1991) and has been shown to have specific affects on the rate of and organization of collagen fibril assembly (Vogel et al, 1984). Localization of decorin intracellularly in the fibroblasts may be consistent with a function related to collagen synthesis. Decorin was found intracellularly in alveolar macrophages in IPF tissue, but not in normals, and this was confirmed using an additional polyclonal antibody to decorin obtained from Telios (data not shown). Decorin and biglycan are deposited throughout the fibrotic interstitium, in areas of frank fibrosis. Biglycan was found to occur throughout the expanded interstitium of the IPF lung (Fig. 4E), but was most concentrated in regions other than those that stained strongly for versican (Fig. 4B). Decorin was most prominent in those areas, that stained strongly for collagen in the Gomori stain (Figs. 4 A and 4D). This is consistent with decorin's association with collagen fibrils (Scott and Orford, 1981; Vogel et al, 1984; Brown and Vogel , 1989; Fleischmajer et al, 1991). In both the normal and IPF lungs, decorin localized to the epithelium, submucosa and smooth muscle bundles of airways, to the adventitia of blood vessels and to alveolar walls. Biglycan was localized to the epithelium and submucosa of airways, to the intima of blood vessels and, to a lesser extent than decorin, to alveolar walls. Biglycan was most concentrated in the adventitia of blood vessels. Neither biglycan nor decorin was detected in high extracellular concentrations in the chondroitin sulfate/ dermatan-rich fibroblast foci of the IPF lung (Figs. 4 A , 4D and 4E). 33 3.3.3 a-Smooth muscle actin immunohistochemistry in IPF Versican is associated with myofibroblasts of the fibroblast foci, the active lesions of IPF. The matrix of the fibroblast foci stains intensely for glycosaminoglycans (Fig. 4A) and versican (Figs. 4B & 5A). The fibroblasts in these foci showed weak intracellular staining for alpha smooth muscle actin (Figs. 4 C & 5B), compared to the strong alpha smooth muscle actin staining observed in smooth muscle cells of airway walls and blood vessels (as seen in the small vessels at the edges of Fig. 5B). The myofibroblasts in the fibroblast foci are highly aligned, spindle shaped cells and Figure 5B shows alpha-smooth muscle actin-stained cells surrounded by dense versican staining (Fig. 5A). The weak alpha smooth muscle actin positivity identifies the spindle shaped cells in the versican-rich zones as myofibroblasts. Versican is associated with smooth muscle cells, as identified by alpha smooth muscle actin positivity. Smooth muscle cells in the airway muscle bundles and blood vessel walls of the normal and IPF lungs (Fig. 4C) stained very heavily for alpha smooth muscle actin. The smooth muscle cells in the blood vessels and airway smooth muscle bundles were surrounded by versican-rich deposits, implying that the smooth muscle cells may be the source of the versican in these areas. Blood vessels permeating the remodeling regions of the IPF lung stained strongly for alpha smooth muscle actin (Fig. 5C). This strong alpha smooth muscle actin staining identifies smooth muscle cells in these areas in both normal and EPF lungs. 34 Figure 3: Versican Localization in Remodeling Areas in Idiopathic Pulmonary Fibrosis and Normal Lung Figure 3: A , B , C, D , E : IPF lung and F: normal lung. A : modified aldehyde-fuschin Gomori trichrome stain and B , C , D , E , F: Immunolocalization of versican (red). A : Glycosaminoglycans stain purple (asterix), elastin navy blue (arrowhead) and collagen aqua. B : Serial section with A . Glycosaminoglycan-rich domains (asterix, purple stain in A ) correspond to areas of versican deposition (asterix, red stain in B) . C , D : Areas of sub-epithelial versican deposition. E : Myofibroblasts (asterix) apparently migrating into airspace. F: Versican is localized to thick sections of the alveolar wall , to airway wall and to the intima of an artery in normal lung. A , B , E Bars = 50um; C Bar = 300um; D, F Bars = lOOum 35 Figure 4: Localization of Glycosaminoglycans, Versican, Decorin, Biglycan and Myofibroblasts in Idiopathic Pulmonary Fibrosis Figure 4: A l l IPF sections are serial. A : modified aldehyde-fuschin Gomori trichrome stain, B : immunostaining for versican, C: immunostaining for a-smooth muscle actin, D : immunostaining for decorin, E : immunostaining for biglycan and F: negative control, incubated with normal rabbit serum in place o f primary antibody. A , B : Glycosaminoglycan deposition in the IPF lung (asterix, purple stain in A ) corresponds to dense versican deposition (asterix, red stain in B) . A , B , C, D , E : Versican, and not decorin or biglycan, is responsible for the dense concentrations of glycosaminoglycans localized to regions of the fibrotic lung. B a r = lOOum 36 Figure 5: Localization of Glycosaminoglycans, Elastic Fibers, Versican and Myofibroblasts in Idiopathic Pulmonary Fibrosis Figure 5: A , B : Serial sections of IPF lung, same field as seen in Figure 4, C , D : Serial sections of IPF lung. A , C : Immunolocalization of versican, B : Immunolocalization of a-smooth muscle actin and D : modified aldehyde-fuschin Gomori trichrome stain. A : Areas o f subepithelial versican (red) deposition beneath hyperplastic epithelium in IPF lung. B : Myofibroblasts (arrows) stain positively for a-smooth muscle actin. The myofibroblasts are localized to the area that stains strongly for versican (A) . Vascular smooth muscle cells stain strongly for a-smooth muscle actin in vessel walls (asterix, red stain). C : Areas o f versican deposition correspond to areas o f high elastic fiber content (D). D : Versican, evident by staining for glycosaminoglycans (purple), surrounds elastic fibers (arrows, navy blue). Versican does not co-localize with collagen (aqua). A , B Bar = 50pm; C , D Bar = 100pm 37 3.3.4 Glycosaminoglycan, collagen and versican localization in Series 2 IPF patient samples, with the pathological patterns of DAD, BOOP and UIP A s described above, versican localized to early lesions in idiopathic pulmonary fibrosis, including thickened alveolar walls and airway lesions. In Series 2 patients, in order to more precisely delineate versican's involvement in early stages of the process, we compared glycosaminoglycan, collagen and versican localization in the remodeling lung, in characteristic lesions of diffuse alveolar damage ( D A D ) , bronchiolitis obliterans organizing pneumonia (BOOP) and idiopathic pulmonary fibrosis (IPF). 3.3.4.1 D A D Patients: Hematoxylin-eosin staining showed the characteristic features of the fibroproliferative phase of A R D S , including thickened interstitium. Alcian blue staining localized glycosaminoglycans to the thickened interstitium of D A D and showed that the thickened interstitium was actually composed of multiple collapsed alveoli, the lumens of which were obliterated by organizing exudate (Fig. 6B). Glycosaminoglycans localized specifically to the thickened walls of these alveoli and to areas of more advanced organization where the underlying architecture could no longer be delineated (Fig. 6B). Picrosirius red staining localized collagen to the walls of arteries and airways (red, Fig. 6B). The thickened alveolar walls showed little staining for collagen with picrosirius red (Fig. 6B). Collagen was sparsely distributed in the thickened interstitium of the D A D lung. The thickened alveolar walls stained intensely for versican (red, F ig . 6C). Immunohistochemical staining for versican (red, Fig . 6B) was highly concordant with histochemical staining for glycosaminoglycans (blue, Fig. 6C). 38 3.3.4.2 BOOP Patients: Hematoxylin-eosin staining showed the characteristic buds of connective tissues within distal bronchioles, alveolar ducts and alveoli in B O O P . Alc ian blue staining localized glycosaminoglycans (blue, F ig . 7B) to the characteristic intraluminal buds of bronchiolitis obliterans organizing pneumonia. Picrosirius red staining showed that collagen (red, F ig . 7B) was sparsely distributed within the glycosaminoglycan-rich intraluminal buds. Immunohistochemical staining for versican (red, Fig . 7C) localized versican to the intraluminal buds, in a pattern that was highly concordant with histochemical staining for glycosaminoglycans (blue, Fig. 7B). 3.3.4.3 IPF (UIP) Patients: Alcian blue staining localized glycosaminoglycans to the fibroblast foci of IPF. Alcian blue staining in Series 2 IPF sections confirmed the presence of high density glycosaminoglycan deposits in the fibroblast foci, as observed in Toluidine Blue 0 staining in Series 1 IPF sections. Collagen, visualized with picrosirius red staining, is sparsely distributed within the fibroblast foci of IPF. Collagen is densely deposited in the frankly fibrotic regions of the IPF lung, away from the dense glycosaminoglycan deposits. This second series of studies confirmed our previous localization of versican to the characteristic fibroblast foci of the IPF lung, prior to replacement of these lesions with collagenous matrix. In addition, the fibroblasts in these foci stained positively for N-terminal propeptides of Type I collagen, consistent with early collagen synthesis (Bensadoun et al, 1996) as has been shown by others previously. In normal lung sections, alcian blue stained only the bronchial cartilage and the intima of some of the pulmonary arteries. 39 Figure 6: Localization of Glycosaminoglycans, Collagen and Versican in Diffuse Alveolar Damage Figure 6: Proteoglycan localization in D A D (the early fibroproliferative phase of A R D S ) . B : Alc ian blue and picrosirius red. Glycosaminoglycans (blue) localize to the thickened interstitium revealing that the thickened interstitium is actually composed of multiple alveoli where the lumens are obliterated by organizing exudate and alveolar collapse. C : Versican. The field matches that in B and shows that versican (red) localizes to the same thickened alveolar walls as the glycosaminoglycans (blue stain in B) . 40 Figure 7: Localization of Glycosaminoglycans, Collagen and Versican in Bronchiolitis Obliterans Organizing Pneumonia Figure 7: Proteoglycan localization in B O O P . B : Alc ian blue and picrosirius red. Glycosaminoglycans (blue) localize to the intraluminal buds while collagen (red) is sparsely distributed within these glycosaminoglycan-rich areas. C: Versican. The field matches that in B and localizes versican (red) to the same intraluminal buds o f B O O P as the glycosaminoglycans (blue stain in B) . 41 3.3.5 Quantification of immunohistochemical staining for versican in IPF The percentage of lung tissue staining positively for versican was quantified in randomly selected fields from the IPF biopsies and sections of normal lung, using immunohistochemistry and morphometry. Areas staining intensely (red, F ig . 4B) for versican in the immunohistochemical staining procedure and total area of all tissue, staining for versican or otherwise, in a microscopic field, at 10 X magnification, were quantified using the Bioview Colour Image Analysis System. The versican-rich matrix in IPF lung was found to occupy an average of 22 .6±4 .1% of the lung tissue area. In the normal lung, versican-rich matrix occupied 2.6±2.0% of the lung tissue area. Results were analyzed statistically by a paired t-test and shown to be statistically significant (p<0.00001). Individual results of this assay are displayed in the following table: Table 3: Morphometric Analysis of Versican Deposition Patient Number Patient Group Versican/ Tissue Mean Versican/ Ratio (%) Tissue Ratio (%) 1380 IPF 27.9 % 1446 IPF 23.9 % 1533 IPF 17.7 % 22.6 + 4.1 %** 1595 IPF 23.1 % (n=6) 1920 IPF 25.4 % 2089 IPF 18.0 % 1430 Control 4 .4% 1464 Control 0 .4% 1653 Control 3.6 % 2.6 + 2.0%** 1773 Control 0.8 % (n=6) 1876 Control 1.1 % A-8984 * Control 5.1 % * Identified by autopsy number (not in lung registry) * * p < 0.00001 The total tissue area in IPF lung sections was approximately 45% greater than the total tissue area in normal lung sections. Thus, the increase in versican content in the IPF lung is not accountable simply by the increase in tissue volume due to the disease process. This dramatic difference in content of versican-rich matrix (p<0.00001) indicates that versican is a key and specific feature in the pathologic pattern in IPF biopsies. 43 CHAPTER 4: Hyaluronan 4.1 Introduction A s outlined in the introduction, animal models show increased synthesis of hyaluronan in inflammation and fibrosis. However, the localization of hyaluronan in human lung disease has never been established. Therefore, hyaluronan was localized in tissue sections from IPF and control patients. We developed and characterized a specific probe that would allow localization of hyaluronan in tissue sections. Sections of diseased and normal lung were stained for hyaluronan, using a biotin-labeled protein that specifically binds hyaluronan; this was detected using an enzyme-linked method. 4.2 Materials and Methods Hyaluronan was localized in tissue sections using the hyaluronan-binding protein, link protein, which was purified and biotinylated for this purpose. Bovine nasal cartilage was diced with razor blades and dissolved at 100 mg/ml in 4 M guanidine-HCl, 0.1 M sodium acetate, p H 6.0 with 5 m M E D T A , 10 Ug/ml pepstatin, 10 ug/ml leupeptin, 1 m M P M S F and 10 ug/ml E-64 (proteinase inhibitors). This solution was incubated at 4°C for 72 hours, after which the solution was centrifuged at 15,000 rpm for 30 minutes in the ultracentrifuge (Beckman). The supernatant was collected and high molecular weight hyaluronan ("Healon"; Pharmacia L K B , Uppsala, Sweden) was added to 1.0% wt/wt. The extract was dialyzed to 0.5 M guanidinium chloride and proteoglycan aggregates, consisting of a complex of aggrecan, link protein and hyaluronan, were purified using cesium chloride (Accurate, Westbury, New York) density gradient ultracentrifugation as previously described (Roberts et al, 1989). 44 Link protein was purified to homogeneity from this complex, by cesium chloride density gradient centrifugation in the presence of 4 M guanidinium chloride (Roberts et al, 1989). Purified link protein was assayed to determine protein concentrations using the Bio-Rad Protein Assay (at 595 nm) against a standard curve of B S A in an identical buffer solution. Purified link protein was identified using sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( S D S - P A G E ) and Western blotting (Roberts et al, 1989). A doublet of 45 kDa and 41 kDa was detected by S D S - P A G E with coomasie blue staining. In Western blots, a doublet at 45kDa and 41 kDa was specifically bound by the antibody 9/30/8-A-4 (NTH), a mouse monoclonal raised against purified rat chondrosarcoma link protein that crossreacts with human link protein (Caterson et al, 1985; Goetinck et al, 1987). The purified link protein was then biotinylated as a complex with 1% w/w high molecular weight hyaluronan (Pharmacia L K B ) by reaction and linkage of the amine groups of lysine residues in the protein with an N -hydroxysuccinimide ester form of biotin, containing an extended hydrocarbon spacer arm to reduce steric hindrances in avidin-biotin binding (Amersham). Biotinylated link protein was dissociated from hyaluronan in 4 M guanidine-HCl (ICN) and purified on a Sepharose C L 2 B chromatography column (Pharmacia L K B ) . Identity of the biotinylated probe was reconfirmed by Western blotting, using the link protein specific antibody, 9/30/8-A-4 and by parallel localization of biotinylated protein with avidin-horseradish peroxidase (Bio-Rad), using chemiluminescence (Amersham). Biotinylated link protein was stored in T B S - B S A containing 5 m M E D T A . Sections to be stained for hyaluronan were blocked for two hours in 2% B S A (ICN) in Tris buffered saline, p H 7.5 and incubated overnight with 50 ug/ml biotinylated link in 2% B S A (ICN) in Tris buffered saline, p H 7.5. These sections were then incubated with streptavidin, conjugated to alkaline phosphatase (Bio-Rad), in 2% B S A (ICN) in Tris buffered saline, p H 7.5. 45 The alkaline phosphatase produced a red precipitate, using naphthol-AS-biphosphate/ new fuchsin (Sigma) as a substrate. In initial experiments, using avidin-alkaline phosphatase as a secondary antibody complex, we observed non-specific staining of intravascular serum and airway mucus, and an intracellular staining of cells later identified as macrophages (serial sections were stained for hyaluronan and CD45, a monocyte/macrophage specific, cell surface antigen). Negative control sections using avidin-alkaline phosphatase alone (no biotinylated link "primary") showed similar, but lighter, intracellular staining. Streptavidin-alkaline phosphatase, a more specific biotin-binding protein complex, was used in later experiments, in place of the avidin-alkaline phosphatase. This replacement eliminated the intracellular macrophage staining, as well as the mucus and serum staining. Thus, streptavidin-alkaline phosphatase was used for all further histochemical localization of hyaluronan. The hyaluronan probe was characterized as follows: Sections were probed with 50 plg/ml biotinylated link protein that had been pre-incubated with 10% wt/wt high molecular weight hyaluronan (Pharmacia L K B ) overnight. This treatment eliminated all the staining (Fig. 8C), demonstrating that the link protein probe binds hyaluronan. A second set of sections were pretreated with streptomyces hyaluronidase, Type LX (Sigma), in 100 m M sodium acetate ( B D H , Toronto, ON), pH 5.0, in I m M iodoacetic acid (BDH), l m M P M S F (BDH) , I m M E D T A (Sigma) and I m M pepstatin A (Sigma), prior to incubation with biotinylated link protein. Again, all staining was eliminated, demonstrating that the link protein probe was specific for a hyaluronidase-sensitive target in our sections. This is a modification of the principle developed by Tengblad (1980) and adapted for immunohistochemistry by Ripellino et al (1985) and Green et al (1988 a, b), who used a 46 trypsinized mixture of link protein and aggrecan to localize and quantitate hyaluronan. Their probe thus consisted of biotinylated fragments of link protein and the hyaluronan-binding region of aggrecan. In preliminary experiments, we trypsinized the aggrecan-hyaluronan-link protein complex, purified and biotinylated hyaluronan-binding protein fragments and used this mixture as a probe, as described by Green et al (1988 a, b). However, we found that biotinylated link protein gave similar results, but with stronger specific staining. Thus, we elected to use biotinylated link protein as our probe. A l l sections were counterstained with Mayer's hematoxylin (Mayer, 1903). A l l sections were washed, dehydrated in ethanol, mounted in Entellan (BDH) and examined by light microscopy. 4.3 Results The matrix in IPF stains intensely for hyaluronan. Both normal and IPF lung tissue showed extensive staining for hyaluronan. In the normal lung, hyaluronan was seen in the submucosa of airways, in cartilage and in the adventitia, media and intima of blood vessels (Fig. 8B). In normal alveolar interstitium, hyaluronan was most concentrated in relatively thicker portions of the alveolar walls. This included the junctions (wall trifurcations) between individual alveoli, the thickened wall at the points where individual alveoli join the terminal bronchiole, and in juxtaposition to capillaries in the alveolar walls. In the IPF lung, hyaluronan was detected in these same areas and extensively throughout the expanded matrix volume that is particular to IPF (Fig. 8A). Hyaluronan in these remodeled regions showed a diffuse staining pattern. A reciprocal staining pattern occurs between hyaluronan and extremely high density deposits of versican in some IPF tissue sections. A reciprocal staining pattern was observed 47 between hyaluronan and the Toluidine Blue metachromatic fibroblast foci of the EPF lung in one patient in initial experiments (Patient #1595 only). These regions stained very lightly or not at all for hyaluronan. Further experiments including the other five patients demonstrated co-localization of hyaluronan with the glycosaminoglycan-rich/ versican-rich zones in the EPF lung. This reciprocal staining in patient 1595 indicates that the areas of dense glycosaminoglycan deposition are either (i) devoid of hyaluronan or (ii) contain proteoglycan that inhibits the binding of biotinylated link protein probe to its target, hyaluronan. Possibility (ii) would be analogous to a phenomenon that occurs in cartilage, where the high content of hyaluronan-associated aggrecan and endogenous link protein occupy the binding sites for the biotinylated link protein probe. The fibroblast foci in the biopsy from patient 1595 stained more intensely that other patients' biopsies for both glycosaminoglycans, histochemically, and versican, immunohistochemically, in these foci. Therefore, we hypothesized, but could not prove, that the reciprocal staining observed in this one patient was due to occupation of the link protein probe binding sites on the deposited hyaluronan by exceptionally high concentrations of versican. 48 Figure 8: Localization of Hyaluronan in Idiopathic Pulmonary Fibrosis and Normal Lung Figure 8: A : Section of IPF lung, B , C: Serial sections of normal lung. A , B : Localization of hyaluronan using biotinylated link probe, C: Negative control, localization o f hyaluronan using biotinylated link probe, blocked by preincubation of section with hyaluronan (Healon). A : Hyaluronan is extensively deposited in the fibrotic regions of the IPF lung. B : Hyaluronan is localized to the airway wall and alveolar walls of the normal lung. C: Biotinylated link probe is specifically absorbed by hyaluronan as evidenced by the elimination of staining by preincubation with the target molecule. Bar = 200pm 49 CHAPTER 5: Leukocytes in Versican-rich and Versican-poor Matrix in IPF 5.1 Introduction In the course of our initial experiments, in which we co-localized versican and a-smooth muscle actin, we noted that the versican-rich zones in the IPF lung sections contained few cells other than the observed myofibroblasts. These relatively acellular zones were often bordered by regions densely populated by cells whose morphologic appearance suggested that they were inflammatory cells. The morphology of the cells was as follows: spherical cell shape, dark staining nuclei and high nucleus to cytoplasm ratio. We used immunohistochemistry techniques to investigate the relationships between versican and inflammatory cell localization. Sections were double stained for (i) versican and the mononuclear phagocyte marker CD68 and (ii) versican and leukocyte common antigen. We quantified the positive cells in each case within the versican-rich and versican-poor areas of the fibrotic lung samples. 5.2 Materials and Methods 5.2.1 Versican immunohistochemistry The versican antiserum described earlier was used (Chapter 3). The antiserum used in this series of experiments was a different batch from that used for the initial analyses of versican in the IPF lung. The antiserum in this series was used at a much lower concentration than the earlier antiserum because (i) it appeared to have a stronger affinity for the target in initial tests and (ii) the alkaline phosphatase anti-alkaline phosphatase ( A P A A P ) secondary antibody system was used, which increases the resulting signal dramatically. This allows use of less antibody for 50 experiments, but the disadvantage of this is that amplification in the process is greater and results are less quantitative. Sections were pretreated with 0.5 units/ml of chondroitinase A B C in 0.1 M Tris, 0.05 M calcium acetate, 0.1% B S A at 37°C for 1 hour. Sections were then blocked for 1 hour in 10% normal goat serum in 2% B S A in Tris-buffered saline (TBS), p H 7.5. Sections were probed with 1 in 500 rabbit anti-versican antibody in 2% B S A , T B S , p H 7.5 for 2 hours. The antigen-antibody complexes were detected using a rabbit alkaline phosphatase-anti-alkaline phosphatase ( A P A A P ) system, 1 in 20 goat anti-rabbit IgG (Dako A / S , Glostrup, Denmark) in 2% B S A , T B S , p H 7.5 for 30 minutes, followed by 1 in 50 rabbit alkaline phosphatase anti-alkaline phosphatase (Serotec, Kidlington, Oxford, England) in 2% B S A , T B S , p H 7.5 for 30 minutes. Naphthol-AS-biphosphate/ new fuchsin (Sigma) was used as the alkaline phosphatase substrate, generating a red precipitate. 5.2.2 Leukocyte immunohistochemistry Versican-immunostained sections were treated with 0.3% hydrogen peroxide in 100% methanol for 30 minutes (to remove endogenous peroxidase activity). This was followed by an enzymatic digestion with pronase E (Type X I V from Streptomyces griseus, Sigma) at 1 mg/ml at 37°C for 1 hour. Sections were blocked for 1 hour in 10% normal goat serum in 2% B S A , T B S , p H 7.5 and probed with 1 in 35 anti-leukocyte common antigen ( L C A ) antibody (Dako) in 2% B S A , T B S , p H 7.5 for 2 hours. L C A antigen/ an t i -LCA antibody complexes were detected with 1 in 2000 biotinylated goat anti-mouse IgG (Sigma) in 2% B S A , T B S , p H 7.5 for 30 minutes. The Vectastain™ peroxidase substrate kit (Vector Laboratories, Burlingame, California) was used to visualize results, incubating sections with an avidin/ biotinylated horse radish peroxidase mixture 51 (Vectastain™ A B C Reagent) for thirty minutes, followed by a 3,3',5,5'-tetramethylbenzidine and hydrogen peroxide mixture ("True Blue" peroxidase substrate, Kirkegard & Perry Laboratories, Gaithersburg, Maryland) for 10 minutes, giving an alcohol soluble, blue precipitate. A l l sections were allowed to air dry ("True Blue" substrate/ precipitate is not resistant to organic solvent treatment) and mounted in Entellan ( B D H Inc., Toronto, Ontario). A full series of control sections were probed in parallel with the above sections, substituting 1 in 2000 normal rabbit serum (NRS) for the anti-versican antibody and/or non-immune mouse IgG for the an t i -LCA antibody. 5.2.3 Mononuclear phagocyte immunohistochemistry Versican-immunostained sections were treated with 0.3% hydrogen peroxide in 100% methanol for 30 minutes (to remove endogenous peroxidase activity) and predigested for 25 minutes at 37°C with 0.1% trypsin (Worthington, Freehold, NJ) in 0.1% calcium chloride (BDH) , pH7.6. Sections were then blocked with 10% normal goat serum in 2% B S A in Tris-buffered saline, p H 7.5 for 60 minutes and probed with a 1 in 50 dilution of the antibody KP1 (Dakopatts, Glostrup, Denmark) in 2% B S A in Tris-buffered saline, pH7.5. KP1 is a mouse monoclonal antibody that specifically binds the CD68 molecule, which is unique to the subcellular fraction of human monocytes/ macrophages (Pulford et al, 1989). CD68-KP1 antibody complexes bound to the sections were detected with 1 in 2000 biotinylated goat anti-mouse IgG (Sigma) in 2% B S A , T B S , p H 7.5 for 30 minutes. The Vectastain™ peroxidase substrate kit (Vector Laboratories, Burlingame, California) was used to visualize results, incubating sections with an avidin/ biotinylated horseradish peroxidase mixture (Vectastain™ A B C Reagent) for thirty minutes, followed by a 3,3',5,5-tetramethylbenzidine and hydrogen 52 peroxide mixture ("True Blue" peroxidase substrate, Kirkegard & Perry Laboratories, Gaithersburg, Maryland) for 10 minutes, giving an alcohol soluble, blue precipitate. A l l sections were allowed to air dry ("True Blue" substrate/ precipitate is not resistant to organic solvent treatment) and mounted in Entellan ( B D H Inc., Toronto, Ontario). A full series of control sections were probed in parallel with the above sections, substituting 1 in 2000 normal rabbit serum (NRS) for the anti-versican antibody and/or non-immune mouse IgG for the anti-CD68 antibody. 5.2.4 Quantification of versican and leukocyte co-localization In the versican-leukocyte common antigen double immunostained sections, cells staining positively for the leukocyte common antigen were visually counted and designated as being within versican-rich zones or within versican-poor zones of the IPF lung. These cells did not include interstitial or alveolar macrophages. Mononuclear phagocytes express very low levels of leukocyte common antigen and, identified in the sections based on morphology and serial sections with CD68 staining, mononuclear phagocytes stained extremely lightly or not at all with the L C A immunohistochemical protocol used. In the versican-CD68 double immunostained sections, cells staining positively for CD68 were visually counted and designated as being within versican-rich zones or within versican-poor zones of the IPF lung. Only CD68-positive cells localized within the lung tissue were counted, thus eliminating alveolar macrophages from this analysis. Sections were visualized at 200X (10X eyepiece lens and 20X objective lens) magnification on a light microscope (Nikon Labophot-2, Nikon , Tokyo, Japan) with the assistance of a 20 by 20 grid system overlaid on a 10X eyepiece lens (Leitz, Wetzlar, Germany). 53 This system provided a grid measuring 0.3025 square millimeters per field analyzed. Cells were counted, and designated as lying within versican-rich zones or within versican-poor zones, using an electronic tabulator (American Dade, Miami, Florida). Five visual fields per test section were examined and the totals per slide compiled (Table 5), thus counting cells within a 1.5125 square millimeter total area per slide/ patient. Results wereanalyzed statistically by a paired t-test. 5.3 Results 5.3.1 Versican/ leukocyte double immunohistochemistry Non-mononuclear phagocyte leukocytes are less abundant in versican-rich extracellular matrix than in versican-poor regions (Fig. 9). Cells that stained positively for the leukocyte common antigen (LCA or CD45) occurred at a lower cell density in the versican-rich zones of the IPF lung than in the versican-poor zones. LCA-positive cell densities in versican-rich and versican-poor zones were calculated for each patient (Table 4), quantified in five microscopic fields per patient/ slide, and the overall difference in densities determined. Non-mononuclear phagocyte leukocytes, defined as strongly LCA-positive cells, were found at a concentration of 5.97 cells per 104 p,m2 of tissue in areas that stained strongly for versican (versican-rich) and 9.35 cells per 104 urn2 of tissue in areas that stained relatively lightly or not at all for versican (versican-poor). LCA-positive cells were only 64% as abundant within versican-rich zones than within versican-poor zones. LCA-positive cell density data, in versican-rich and versican-poor areas, were analyzed by the paired t-test and found to be significantly different (p=0.003). These 54 observations and values were confirmed by an independent analysis performed by a second observer, who had no knowledge of the original results. The exclusion of leukocytes from versican-rich zones is apparent in color photomicrographs of these foci within the context of the IPF fibrotic lung (Fig. 9A). The majority of the cells within the versican-rich zones have spindle shaped nuclei and low nucleus to cytoplasm ratios. These cells stain positively for a-smooth muscle actin (Fig. 4C), identifying them as the myofibroblasts most likely responsible for the synthesis of the new extracellular matrix. The non-mononuclear phagocyte leukocytes tend to be more highly concentrated in the collagen-rich, versican-poor zones of the IPF lung (Figs. 9A) . O f particular note is the observation that these cells tend to be clustered about the versican-rich zones (Figs. 4B & 9A), in the collagen-rich areas that border these versican-rich fibroblast foci. Table 4: Leukocyte Cell Density in Relation to Versican IPF Patient Leukocyte Density in Versican- Leukocyte Density in Versican-Number rich Matrix poor Matrix (cells per 104 urn2) (cells per 104 um2) 1380 5.64 11.5 1446 4.72 9.40 1533 5.23 8.21 1595 6.69 8.95 1920 5.90 8.29 2089 7.62 9.74 Mean ± SD 5.97 ± 1.05* 9 . 3 4 ± 1.21* significantly different, p<0.003 5.3.2 Versican/ mononuclear phagocyte double immunohistochemistry Mononuclear phagocytes are more abundant in the versican-rich extracellular matrix of the IPF lung than in the versican-poor regions (Fig. 9). Tissue mononuclear phagocytes were defined as those cells in the tissue sections that stained positively for CD68 and that did not 55 reside within the alveolar space. Tissue mononuclear phagocytes occurred at higher cell density within versican-rich regions of the IPF lung than in versican-poor regions. Mononuclear phagocyte cell densities in versican-rich and versican-poor zones were calculated for each patient (Table 5), quantified in five microscopic fields per patient/ slide, and the overall differences determined. Mononuclear phagocytes were found at a concentration of 3.74 cells per 10 4 u m 2 of tissue in areas that stained strongly for versican (versican-rich) and 1.97 cells per 10 4 urn 2 of tissue in areas that stained relatively poorly or not at all for versican (versican-poor) (Table 5). Mononuclear phagocytes were 190% as concentrated in versican-rich zones than in versican-poor zones. Mononuclear phagocyte cell density data, in versican-rich and versican-poor areas, was analyzed by the paired t-test and found to be significantly different (p=0.01). These observations and values were confirmed by an independent analysis performed by a second observer, who had no knowledge of the original results. Table 5: Mononuclear Phagocyte Cell Density in Relation to Versican IPF Patient Mononuclear Phagocyte Density Mononuclear Phagocyte Density Number in Versican-rich Matrix in Versican-poor Matrix (cells per 104 um 2 ) (cells per 104 (xm2) 1380 4.22 2.96 1446 3.11 1.97 1533 3.54 1.44 1595 3.55 2.11 1920 2.34 1.59 2089 5.70 1.73 Mean ± SD 3.74 ± 1.14* 1.97 ± 0 . 5 4 * *significantly different, p<0.01 56 Figure 9: Localization of Versican, Leukocytes and Mononuclear Phagocytes in Idiopathic Pulmonary Fibrosis A B 'Era*. , 1 Figure 9: Versican, leukocyte and mononuclear phagocyte localization in IPF. A : immunostaining for versican (red) and leukocyte common antigen (blue). Non-mononuclear phagocyte leukocytes (blue, arrowhead) localize to versican-poor extracellular matrix and cluster at the edges of versican-rich sub-epithelial fibroblast foci (red). B : Alc ian blue (aqua) and immunostaining for CD68-positive mononuclear phagocytes (purple). Mononuclear phagocytes (purple, arrowhead) localize within glycosaminoglycan-rich matrix (aqua). Glycosaminoglycan-rich matrix corresponds to the versican-rich fibroblast foci previously described. 57 CHAPTER 6: Versican mRNA Synthesis in Normal Lung and Fibrosis 6.1 Introduction We have demonstrated an increased deposition of the proteoglycan versican, detected immunohistochemically, in the fibrotic regions of IPF lungs as compared to control lungs. Increased deposition of versican protein could be due to increased synthesis or decreased degradation, or both. We next investigated the level of m R N A for versican in tissue samples from disease lungs compared to control lungs. The structure of the gene for versican was published in December, 1994, as we were completing our immunohistochemical demonstration that versican was abundant in the fibroblast foci of IPF lung. A novel element in this was the demonstration that the human versican gene contained an additional exon, coding for an additional glycosaminoglycan-attachment domain (Naso et al, 1994), not present in the original versican c D N A (Zimmermann and Ruoslahti, 1989) and the demonstration that four alternately spliced variants could exist (Dours-Zimmermann and Zimmermann, 1994). The N-terminus of the versican protein (Fig. 10) contains a hyaluronan binding region ( H A B R ) domain (LeBaron et al, 1992). The two central domains of the versican protein contain multiple glycosaminoglycan attachment sites, S G X G (serine-glycine-X-glycine amino acid sequences). There are also S G and GS (serine-glycine and glycine-serine) amino acid pairs representing additional putative glycosaminoglycan attachment sites (Zimmermann and Ruoslahti, 1989) within these two central domains. The C-terminus of the versican consists of two epidermal growth factor-like (EGF-like) modules, a lectin-like module and a complement regulatory protein-like (CRP-like) module (Zimmermann and Ruoslahti, 1989). There exist four potential versican m R N A splice variants (Fig. 11) that code for one, both, or neither of two 58 central glycosaminoglycan-attachment domains (Dours-Zimmermann and Zimmermann, 1994), identified as G A G - a and G A G - p \ The splice variants are designated V 0 (containing G A G -a + GAG-P), V i (GAG-(3), V 2 (GAG-a) and V 3 (no glycosaminoglycan domains) and contain 17-23 ( V 0 ) , 15 (V)), 6 (V 2 ) or 0 ( V 3 ) putative glycosaminoglycan-attachment sites. The specific charge density of each of the splice variants is in the order V 0 > V ! > V 2 > V 3 , assuming that all putative glycosaminoglycan-attachment sites are functional. Within this terminology, the original versican c D N A isolated by Zimmermann and Ruoslahti (1989) was V l ; containing GAG-p. To determine the nature of the versican splice variants expressed in fibroproliferative lung disease samples and control lung samples, we had to develop a method to study splice variant m R N A and total versican m R N A in R N A samples from very small tissue samples. Other studies in our laboratory had previously demonstrated that cultured human fetal lung fibroblasts (HFL-1) express high levels of versican m R N A by Northern blotting and protein by immunohistochemistry (Roberts C R , Hammil l D and Burke A K , unpublished). We optimized m R N A splice variant-specific reverse transcriptase-polymerase chain reaction (RT-PCR) techniques using these cultured human fetal lung fibroblasts and then a panel of lung samples was studied, using semi-quantitative domain-specific R T - P C R for versican variants. 6.2 Materials and Methods 6.2.1 PCR primer sets P C R primers were synthesized by the U B C Oligonucleotide Laboratory, based on previously published primer sequences. The following table indicates the targets (compare to Fig . 11) and original references for each of the primer sets used for these experiments: 59 Table 6: PCR Primer Sets Set P C R Pr imers Target P C R Product Size (bp) Reference 1* Ver 1 and V e r l a Versican; hyaluronan binding region ( H A B R ) 780 Grover and Roughley, 1993 2* Ver 3 and Ver3a Versican; lectin-like and complement regulatory protein-like regions 530 Grover and Roughley, 1993 3 B - U p and B - L o w Versican V 0 variant; G A G -a and GAG-(3 regions 351 Dours-Zimmermann and Zimmermann, 1994 4 A - U p and B -L o w Versican V ! variant; H A B R and GAG-(3 regions 386 Dours-Zimmermann and Zimmermann, 1994 5 B - U p and C-Low Versican V 2 variant; G A G -oc and EGF- l ike regions 373 Dours-Zimmermann and Zimmermann, 1994 6 A - U p and C -L o w Versican V 3 variant; H A B R and EGF- l ike regions 408 Dours-Zimmermann and Zimmermann, 1994 7 G A P D H #1 and G A P D H #2 Human glyceraldehyde 3-phosphate dehydrogenase 445 Dr. Rick Hegele, Pulmonary Research Lab, U B C (manuscript in preparation) * Note that primer sets 1 and 2 are to sequences present and constant in all versican variants. P C R primer sets specific to each of the four splice variants (Dours-Zimmermann and Zimmermann, 1994) were used to determine levels of m R N A of these variants in disease and control patient samples. Initially, the V e r l and V e r l a P C R primer pair, which bind to a region of versican c D N A coding for part of the hyaluronan-binding region, was used to determine total versican m R N A levels. The V e r l / V e r l a P C R product is more than twice the size of the versican splice variant products ( V 0 , V , , V 2 , V 3 ) . Therefore, the V e r l / V e r l a polymerization rate would be considerably lower than that of the splice variant P C R products and it was deemed unsuitable for direct comparison. The Ver3/Ver3a pair of P C R primers, which span the lectin-like and complement regulatory protein-like regions (Grover and Roughley, 1993) and give a 530 base pair P C R product, were used to determine total versican m R N A levels in both disease patients 60 and controls. Levels of glyceraldehyde 3-phosphate dehydrogenase m R N A (a housekeeping gene) were quantified, using R T - P C R , for all samples, as a denominator for versican m R N A quantification and to verify the consistency of the technical aspects of the R T - P C R techniques used. A l l of these R T - P C R reactions were first optimized using R N A purified from a human fetal lung fibroblast cell line (HFL-1 , A T C C ) and these cells were also used as positive controls for the disease and control patient R T - P C R reactions. 6.2.2 Human fetal lung fibroblast culture Human fetal lung fibroblasts (HFL-1 ; A T C C ) were used as positive controls for versican expression and for the optimization of each of the versican splice variant specific P C R primer sets. H F L - 1 cells, at passage 13, were acquired from the American Type Culture Collection. The fibroblasts were grown in D M E M containing 10% heat-inactivated fetal calf serum (HIFCS) in T75 culture flasks (Gibco B R L ) in an atmosphere of 5% C02/95% air. The medium was changed every 72 hours. Confluent cultures were harvested by trypsinization, with 0.25% trypsin, and subcultured weekly at a dilution of 1:20. The total number of passages for this culture, prior to R N A extraction, was 15. Figure 10: Versican Protein Structure 61 COOH Putative cell-binding domain GAG-attachment domain Hyaluronan-binding domain Complement-regulatory protein homology Lectin homology Epidermal growth factor homology Chondroitin sulphate Hyaluronan Figure 10: Illustration of versican based on c D N A from Zimmermann and Ruoslahti (1989) showing N-terminal hyaluronan binding domain, central glycosaminoglycan-attachment domains with chondroitin sulfate side chains and C-terminal epidermal growth factor-like, lectin-like and complement regulatory protein-like modules. 62 Figure 11: Structure of Versican Splice Variants D O M A I N S T R U C T U R E O F S P L I C E V A R I A N T S O F H U M A N V E R S I C A N c D N A A N D P R O T E I N S , S H O W I N G L O C A T I O N S O F P R I M E R S F O R D O M A I N - S P E C I F I C P C R . c D N A A-UP H A B R B-UP B-LOW G A G a G A G p Ve r 3 C-LOW V e r 3 A E-L-C P R O T E I N N I VO 17-23 G A G s G A G a G A G ( i P r imers B-Up/B-Low Spec i f i c product 351 bp V1 12-15 G A G s C G A G p Pr imers A-Up/B-Low Spec i f i c product 386bp V2 II I I I I G A G a Pr imers B-Up/C-Low Spec i f i c product 373bp V 3 No G A G s Pr imers A-Up/C-Low Spec i f i c product 408bp All ve rs i can var iants give a 530bp product with pr imers Ve r3 and V e r 3 A Primers are as reported by Grover and Roughley (1993) [Ver3 and Ver 3A] and Dours-Zimmermann and Zimmermann (1994) [A-UP, B-UP, B-LOW and C-LOW]. All primer sets are intron-spanning, to exclude the possibility of interference in analysis by genomic DNA. 63 6.2.3 RNA extraction from HFL-1 cells Total R N A was extracted from cultured human fetal lung fibroblasts (HFL-1 ; A T C C ) . Medium from the cultured cells was removed and R N A was extracted using 8ml of TRIzol reagent (Gibco B R L ) per 75cm flask and the accompanying single step methodology (Chomczynski, 1993). R N A was quantified spectrophotometrically and the A 2 6 o / A 2 8 o ratio was more than 1.8 for all samples. l.Oug of purified R N A from each sample was reverse transcribed and the resultant c D N A used for P C R reactions with primer sets specific for each of the four splice variants and for two constant regions of the versican c D N A . 6.2.4 Reverse transcription of HFL-1 cell RNA "First Strand" synthesis reactions for each of the H F L - 1 cell R N A samples were run, each at a volume of 40 ul (2X original recipe). A l l R N A had been dissolved in DEPC-treated water. A n additional sample tube was run containing only DEPC-treated water (no R N A ) as a negative control (labeled " R T - H 2 0 " ) . We used 100 ng of random hexamer per (ig of R N A . Each reverse transcriptase reaction contained: I X first strand synthesis buffer (50mM Tr i s -HCl , pH 8.3, 75 m M KC1, 3 m M M g C l 2 ) 10 m M D D T , 20 units R N A s i n 0.5 m M of each of the four dNTPs 100 ng random hexamers 1.0 ug sample R N A 400 units of M M L V reverse transcriptase (all reagents, except random hexamers, supplied by Gibco B R L ) . 64 A l l reactions were incubated at 37°C for 1.5 hours, heated at 95°C for five minutes to inactivate the reverse transcriptase, aliquoted at 4 p i per tube for use in P C R reactions and stored at -70°C. 6.2.5 MgCI 2 optimization for PCR A l l the optimization experiments, for all primer sets, were run on H F L - 1 R N A as a positive control for versican m R N A . We found, in pilot experiments using Northern blotting, Western blotting and immunohistochemistry, that this cell line produces and deposits large amounts of versican in culture. A negative control (no c D N A ) was run in each optimization experiment, replacing the volume of c D N A (4pl per tube) with DEPC-treated water. A n additional sample tube containing only DEPC-treated water that had gone through the reverse transcription step (no R N A or c D N A ) was also run as a negative control. A third negative control was run using H F L - 1 R N A that had not been reverse transcribed. Negative controls were run using a magnesium chloride concentration of 5.0 m M . H F L - 1 c D N A (4ul per reaction tube) was subjected to P C R with each of the P C R primer sets at a range of magnesium chloride concentrations (1.5 m M , 2.5 m M , 5.0 m M and 10.0 mM) . Each reaction contained, in addition to each specific magnesium chloride concentration: I X P C R buffer 0.2 m M of each of the four dNTPs 4pl of template c D N A 1.0 p M of each P C R primer pair 2.5 units of Taq polymerase (all P C R reagents supplied by Gibco-BRL) 65 Each optimization reaction was run for 35 P C R cycles with an annealing temperature specific to each primer set, as indicated by the original publications describing their use (Grover and Roughley, 1993 and Dours-Zimmermann and Zimmermann, 1994). P C R products were run (25 u l per lane in I X P C R loading buffer [2.0% Ficol l 400, 10 m M N a 2 E D T A , 0.1% SDS and 0.025% (w/v) bromophenol blue] and containing 0.2 ug ethidium bromide) on a 2.0% D N A grade agarose (Bio-Rad) gel in 0.5X T B E . Results were photographed and the optimal magnesium chloride concentration for each primer set determined, based on maximum specific band intensity and least non-specific banding appearance on the gel. The following table indicates the optimal magnesium chloride concentration for each P C R primer set as determined using H F L - 1 c D N A : Table 7: Optimal MgCl 2 Concentrations for PCR Primer Sets P C R Pr imer Set: Op t ima l M g C l 2 Concentration Ver 1 and Ver l a 2.5 m M Ver 3 and Ver 3 a 1.5 m M V 0 (B-up and B-low) 2.5 m M V , (A-up and B-low) 1.5 m M V 2 (B-up and C-low) — V 3 (A-up and C-low) 5.0 m M Note: The V 2 primer set (B-up and C-low) gave no product using H F L - 1 c D N A , so 2.5 m M magnesium chloride was used for patient samples for possible optimization at a later date, if a positive c D N A sample could be found. A l l negative controls were completely negative for all R T - P C R reactions tested. 66 A l l disease and control patient P C R reactions were run using the indicated optimal magnesium chloride concentration. The G A P D H P C R primer set had been previously optimized by the suppliers using 2.5 m M magnesium chloride. 6.2.6 RNA extraction from human tissue samples Frozen samples from the previous immunohistochemically studied lung samples, embedded in paraffin, were unavailable for the use in the R T - P C R studies. We obtained frozen tissue samples prospectively from a new set of disease and control patients. These patients were included based on the histological classification of their biopsy by a service pathologist at St. Paul's Hospital and were all in their seventh decade. We investigated and compared the expression of versican splice variants between purified R N A samples of fibroproliferative lung disease patients (2 cases of Diffuse Alveolar Damage [DAD] in association with Adult Respiratory Distress Syndrome [ A R D S ] ; 4 cases of Bronchiolitis Obliterans with Organizing Pneumonia [BOOP] and 4 cases of Idiopathic pulmonary fibrosis [IPF]). 5 control patients with normal lung function tests were chosen and tissues examined histologically and confirmed as normal-appearing. Control patient samples were from patients for whom the resection of a lung or lobe was necessary for the removal of a localized, peripheral coin lesion, and for whom pulmonary function tests were normal, before resection (see Appendix 1 for information on patients). Total R N A was extracted from fibroproliferative lung disease patient tissue samples and control patient tissue samples using the TRIzol reagent (Gibco B R L ) and the accompanying single step methodology (Chomczynski, 1993).). R N A was quantified spectrophotometrically and the A260/A280 ratio was more than 1.7 for all samples. l.Oug of purified R N A from each 67 sample was reverse transcribed and the resultant c D N A used for P C R reactions with appropriate, specific primer sets. 6.2.7 Reverse transcription of disease and control RNA "First Strand" synthesis reactions for each of the isolated patient and control R N A samples were run, each at a volume of 40 ul. Two of the H F L - 1 R N A samples were also run as positive controls for this series of experiments. A l l R N A had been dissolved in DEPC-treated water. A n additional sample tube was run containing only DEPC-treated water (no R N A ) as a negative control (labeled " R T - H 2 0 " ) . The reverse transcription of the patient and control R N A were run using the methods and conditions previously described for the HFL-1 R N A reverse transcription. The resultant c D N A was aliquoted at 4 ul per tube for use in P C R reactions and stored at -70°C. 6.2.8 Versican PCR on disease and control tissue samples A l l P C R reactions were run using the parameters described in the magnesium chloride optimization section of the materials and methods, using a concentration of magnesium chloride as indicated by the results of the optimization experiments. Each reaction contained 4ul of template c D N A from each reverse transcriptase reaction. Each reaction set included two negative controls, labeled " P C R - H 2 0 " and " R T - H 2 0 " , as previously described and two positive control samples, each containing 4ul of HFL-1 c D N A . A l l P C R reactions were run for 30 cycles. P C R products were run on 2.0% agarose gels, as previously described, and photographed. The P C R methods used here are not truly quantitative, but are semi-quantitative. In order to ensure consistent starting levels of R N A and c D N A , R N A extractions were performed 68 simultaneously for normal and diseased tissue samples. c D N A was synthesized simultaneously for all samples using the same starting quantity of R N A (1.0 ug). In order to ensure comparable data, the number of P C R cycles was reduced from 35 cycles to 30. Pilot experiments, using H F L -1 cells and tissue samples demonstrated that 35 cycles of P C R maximized the reactions and the resultant P C R products. At 30 cycles the P C R reactions had not reached a maximum production limit and therefore could be compared semi-quantitatively. 6.2.9 Sequencing of PCR products The resulting P C R products for each of the versican P C R reactions giving positive results (Vo, V ] , V~2 and Ver3/Ver3a products) were sequenced to verify their identity. P C R bands were excised from the 2% agarose gel. The strips of gel were sealed in dialysis tubing, P C R products were electroeluted and collected. For each P C R product, bi-directional sequencing was performed by the Nucleic A c i d and Protein Sequencing Service (NAPS) at U B C , using our P C R primers. 6.3 Results 6.3.1 GAPDH RT-PCR R T - P C R for G A P D H m R N A resulted in the consistent production of the 445bp P C R product expected between all patient and control samples (Fig. 12). H F L - 1 cells (positive control) produced slightly higher levels of G A P D H P C R product per R N A concentration than all patient samples. A l l negative controls for the G A P D H R T - P C R were completely negative. 69 6.3.2 Total versican (Ver3/Ver3a) RT-PCR The Ver 3/3a P C R product is derived from c D N A in the lectin-like and complement regulatory protein-like modules of versican, and is found in all versican variants. These modules • are found in separate exons of the versican gene. This reaction gave the expected product of 530 bp for all fibroproliferative disease patients and the H F L - 1 positive controls (Fig. 12). A l l "normal" control tissue samples were very lightly positive for this R T - P C R product. In addition to the expected 530bp band, the H F L - 1 positive controls and one disease sample, a diffuse alveolar damage patient, gave an additional P C R product, a doublet of 690-7 lObp in size. The 530bp P C R band was electroeluted and sequenced, confirming its identity as derived from c D N A for the versican lectin-like and complement regulatory protein-like modules. Attempts to electroelute and sequence the 690bp and 710bp bands were unsuccessful as the quantity of D N A eluted for each band was insufficient for sequencing. Variations in R T - P C R product levels, between the Ver3/Ver3a reactions in the disease samples, were consistent with the slight variations observed in G A P D H m R N A levels in these samples (Fig. 12). A l l negative controls were completely negative. 6.3.3 V 0 and versican variant RT-PCR R T - P C R for the V 0 and V , splice variants was positive for the H F L - 1 positive controls and all disease samples. A l l the "normal" control samples were very faintly positive for the expected P C R products (Fig. 12) and at higher cycle numbers or higher concentrations of c D N A than shown in Figure 12, gave clear, positive results. V 0 R T - P C R gave a positive product of the expected size, 351 bp, whereas the product for the V , R T - P C R appeared slightly larger than the 70 expected 386 bp size on the gel. The P C R products from the 10 IPF patients ( D A D , UIP and B O O P ) were electroeluted, purified and sequenced to confirm their identities. The sequencing results verified the identity of both the V 0 and V | R T - P C R products, despite the V i product running at a larger size on the gel, presumably due to aberrations in the gel running procedure. A l l negative controls were completely negative for both P C R primer sets. 6.3.4 V 2 versican variant RT-PCR R T - P C R for the V 2 versican variant was intermittently positive in some disease and some normal samples, but consistently negative in the H F L - 1 cells (Fig. 12). The R T - P C R product was of the expected size, 373bp, but upon electroelution (both disease and control samples), purification and sequencing, it was discovered that the product was not versican, but the product of an unrelated unknown gene, with low homology to known members of the pregnancy specific glycoprotein family. The P C R product had approximately 70% homology with each of the two P C R primers, but no further homology with the versican gene. A l l samples were therefore determined to be negative for the V 2 splice variant. A l l negative controls for the P C R reaction were completely negative. 6.3.5 V 3 versican variant RT-PCR R T - P C R for the V 3 splice variant of versican was negative for all patient and "normal" control samples. H F L - 1 positive controls were either negative or very faintly positive for this P C R product (Fig. 12), consistent with the very light positive product in the magnesium chloride optimization experiments, using these cells. We were unable to sequence this faint band. A l l negative controls were completely negative. 71 Figure 12: Expression of Versican mRNA from Fibroproliferative Lung Disease and Control Patients, Analyzed by RT-PCR F DAD UIP S izG Msrkors BOOP F CONTROLS H , 0 (base pairs) Versican V„ 400 bp 300 bp Versican V, PMMM *•* Mft •4- 500 bp + - 400 bp Versican V , 400 bp 300 bp Versican V , « - 500 bp + - 400 bp LEC/CRP Domain (all versican variants) 600 bp 500 bp GAPDH 500 bp 400 bp Figure 12: Versican variant-specific R T - P C R . F: Human fetal lung fibroblast samples, D A D : Diffuse alveolar damage patient samples, UIP: Usual interstitial fibrosis (pathological description of idiopathic pulmonary fibrosis) patient samples, B O O P : Bronchiolitis obliterans organizing pneumonia patient samples, Controls: Control patient samples, H 2 0 : Negative control (no c D N A ) . 72 CHAPTER 7: Discussion In these studies we investigated the synthesis and deposition of glycosaminoglycans and proteoglycans in fibroproliferative lung disease. We used histochemistry, immunohistochemistry and the reverse transcription-polymerase chain reaction (RT-PCR) to study these molecules in idiopathic pulmonary fibrosis (IPF), diffuse alveolar damage ( D A D ) associated with adult respiratory distress syndrome ( A R D S ) and bronchiolitis obliterans organizing pneumonia (BOOP). We hypothesized that the synthesis and deposition of specific proteoglycans in these diseases is a part of the remodeling process. We sought to clarify the role of proteoglycans in the progression of lung fibrosis in general and to differentiate this role within the different studied examples of lung fibrosis. 7.1 Localization and Identification of Glycosaminoglycans in the Remodeling Lung Histochemical staining was used to localize glycosaminoglycans to the active lesions of human lung fibrosis. Glycosaminoglycans, detected using alcian blue, toluidine blue 0 and the modified aldehyde-fuschin Gomori trichrome method, were abundant in the subepithelial fibroblast foci. The glycosaminoglycan-rich fibroblast foci were observed at the edge of the densely fibrotic regions of the EPF lung, particularly where mature fibrosis bordered on airspace and normal appearing lung. These fibroblast foci were unique to the remodeling lung, not being observed in normal interstitium. The glycosaminoglycan-rich fibroblast foci have a very characteristic appearance and correspond to fibroblast foci identifiable on hematoxylin and eosin staining. These foci have been documented in the pulmonary pathology literature as new or early areas of tissue remodeling. Consistent with this, modified aldehyde-fuschin Gomori trichrome 73 staining and picrosirius red staining demonstrated that the glycosaminoglycan-rich fibroblast foci are collagen-poor relative to the dense, collagen-rich matrix in areas of mature fibrosis, collagen being the hallmark of fibrosis. In this patchy, diffuse disease, the glycosaminoglycan-rich fibroblast foci appear on the leading edge of an advancing wave of fibrotic tissue. Histochemical staining, coupled to predigestion of glycosaminoglycans with specific glycosidases, was used to identify the glycosaminoglycans deposited in the active lesions of idiopathic pulmonary fibrosis as chondroitin sulfate/ dermatan sulfate. The described histochemical staining for glycosaminoglycans was eliminated by pretreatment of tissue sections with a chondroitin sulfate/ dermatan sulfate glycosidase, chondroitinase A B C , but not by pretreatment with another glycosidase, heparinase HI. Chondroitin sulfate/ dermatan sulfate is the most abundant glycosaminoglycan deposited in the characteristic fibroblast foci of the remodeling lung. 7.2 Localization of Proteoglycans in the Remodeling Lung Immunohistochemical staining, using an antibody raised against the versican core protein, showed that immunohistochemical staining for versican and histochemical staining for glycosaminoglycans were highly concordant, demonstrating that the proteoglycan versican, a chondroitin sulfate bound proteoglycan, is likely responsible for much of the dense glycosaminoglycan deposits. The concordance of versican and glycosaminoglycan staining demonstrated that the increased staining for versican in idiopathic pulmonary fibrosis represents true proteoglycan deposition, rather than deposition of core protein alone. The very high degree of staining of the proteoglycan-rich fibroblast foci in idiopathic pulmonary fibrosis with the versican antibody and very light staining of cartilage with the versican antibody are completely 74 inconsistent with the possibility that the proteoglycan in idiopathic pulmonary fibrosis is aggrecan rather than versican. In contrast, immunohistochemical staining for the proteoglycans decorin and biglycan showed a different distribution. Decorin and biglycan were localized to the collagen dense, glycosaminoglycan-poor regions of the remodeling lung, in areas of dense fibrosis. Decorin was also localized intracellularly within the glycosaminoglycan-rich fibrotic foci, which may be consistent with its documented role in collagen assembly (Vogel et al, 1984; Brown and Vogel , 1989) and the fact that these cells synthesize Type I procollagen (Kuhn and McDonald , 1991; Bensadoun et al, 1996). The localization of intracellular decorin and extracellular versican to the collagen-poor fibroblast foci supports the conclusions that versican synthesis and deposition in the fibroblast foci is an early phase of the remodeling process. This study is the first localization and characterization of versican deposition in the normal lung and in the fibrotic foci of the remodeling lung. Versican is likely the major proteoglycan responsible for the presence of much of the glycosaminoglycans in the fibrotic foci of the remodeling lung. 7.3 Hyaluronan Localization in the Remodeling Lung Versican's ligand, hyaluronan, was found to be distributed ubiquitously through the extracellular matrix of the lung in idiopathic pulmonary fibrosis. In particular, hyaluronan co-localized with versican in the characteristic fibroblast foci of the remodeling lung. A number of factors might lead to underestimation of hyaluronan localization by our localization technique, particularly in the versican-rich fibroblast foci. Both versican and our hyaluronan probe, biotinylated link protein, bind hyaluronan. In areas of the lung that are rich in versican, the number of sites on hyaluronan available for binding by the hyaluronan probe may be reduced. 75 Hyaluronan synthesis by proliferating cells in culture and deposition in proliferating zones in vivo has been documented (review in Toole, 1991). Increased hyaluronan in bronchoalveolar lavage ( B A L ) fluid has been reported in IPF (Bjermer et al, 1989), A R D S (Hallgren et al, 1989), and sarcoidosis (Hallgren et al, 1985). Hyaluronan may influence fluid balance in the lung interstitium (Bhattacharya et al, 1989). In addition to water-binding properties, hyaluronan influences cell migration in morphogenic events in development (Toole, 1991) and in wound healing (Laurent and Fraser, 1992). Pulmonary macrophages (Green et al, 1988) and proliferating epithelial cells (Alho and Underhill, 1989) express the hyaluronan receptor, CD44. Migrating smooth muscle cells in development (Boudreau et al, 1991) use distinct receptors for hyaluronan ( " R H A M M " ) in locomotion through the interstitium. Hyaluronan is likely a part of the early cellular migration phase, as well as being an integral component of the new interstitium that makes up the idiopathic pulmonary fibrosis lung. Versican may bind molecules of hyaluronan to form large aggregates in the extracellular matrix. It is clear that versican and hyaluronan are abundant in the "provisional" matrix in the remodeling lung, delineating areas that wi l l later be replaced by mature, collagen-rich fibrosis. Deposition and organization of versican and hyaluronan may provide a temporary or "provisional" architectural structure in which, at later stages of lung remodeling, collagen synthesis, deposition and organization into mature collagen fibrils may occur. This "provisional" matrix is observed in the remodeling lung as the characteristic fibroblast foci that represent active lesions of lung remodeling. The initial synthesis and deposition of hyaluronan and versican by fibroblasts and myofibroblasts to form this "provisional" matrix may be followed by further migration of myofibroblasts and fibroblasts into this matrix and a change in the focus of extracellular matrix synthesis by the fibroblasts. At later stages, fibroblasts in the subepithelial 76 fibroblast foci begin to synthesis collagen and decorin, as indicated by the intracellular staining for decorin in these studies and the intracellular staining for the N-terminal propeptides of Type I collagen in additional studies (Bensadoun et al, 1996). Elucidation of the key processes in the conversion of fibroblasts and myofibroblasts from "provisional" matrix synthesis to collagen synthesis and deposition is a candidate for future study. Further experiments showed that versican is deposited in association with myofibroblasts in the alveolar walls in diffuse alveolar damage, lesions in bronchiolitis obliterans organizing pneumonia (Bensadoun et al, 1996), lesions of sarcoidosis, extrinsic allergic alveolitis and tuberculosis (Bensadoun et al, 1997), and in the thickened airway walls in fatal asthma (Roberts, 1995; Roberts and Burke, 1998). The synthesis and deposition of versican appears to be a common element in the processes of fibroproliferative lung disease. 7.4 Semi-quantitative Analys is of Versican in the Remodeling Lung The percentage of tissue occupied by versican-rich domains in disease and control lungs was quantitated using a computerized image analysis system. Versican-rich domains occupied 22.6 + 4.1% of tissue area in biopsies of idiopathic pulmonary fibrosis tissue (n=6), compared with 2.6 + 2.0% of total tissue area in age-matched control lung tissue that was processed and stained concurrently (n=6) (significantly different; p<0.00001). As total tissue area is itself increased dramatically in lung fibrosis, the increased versican deposition likely represents a considerable absolute increase in versican in idiopathic pulmonary fibrosis. It should be noted that this study represents a retrospective analysis of biopsy tissue, taken from areas of lung selected for pathologic analysis by a thoracic surgeon, and is not an analysis of random samples of lung tissue from idiopathic pulmonary fibrosis patients. There was no clear relationship 77 between the percentage of tissue area occupied by versican-rich domains in these biopsies and the duration of symptoms in idiopathic pulmonary fibrosis patients. 7.5 Characterization of Versican in the Remodeling Lung Reverse transcriptase-polymerase chain reaction, using primers common to all versican variants and primers specific to each versican variant were used to characterize versican, semi-quantitatively, in the fibrotic lung. The number of cycles of P C R used was reduced to non-maximized product levels, to allow for semi-quantitative analysis of the results. The versican gene may generate four potential m R N A splice variants. The four variants include two, one or none of two central glycosaminoglycan-binding domains (Fig. 11), labeled G A G - a and G A G - p \ We determined that the V 0 splice variant of versican, containing both the G A G - a and GAG-(3 regions of the gene and potentially 17-23 chondroitin sulfate side chains, and the V , splice variant of versican, containing the GAG-(3 region and potentially 11-15 chondroitin sulfate side chains, are the prevalent versican products found in fibroproliferative disease samples. These two variants contain the highest number of potential glycosaminoglycan attachment sites, in the mature protein product, of all the potential versican splice variants. The presence of these two versican variants correlates well with the high density deposits of glycosaminoglycans observed using histochemistry. 7.6 Versican Structure and Function Versican was named following c D N A cloning from a human fibroblast library (Zimmermann and Rouslahti, 1989). The exact function of versican is still under investigation. Versican has considerable homology in domain structure to aggrecan, the large proteoglycan of cartilage. It contains a hyaluronan binding region in its N-terminus and has been demonstrated to bind hyaluronan with a dissociation constant of 4x l0" 9 M , which is similar to the binding constant of aggrecan (LeBaron et al, 1992). This domain likely serves to keep versican in place in the tissue, in large aggregated complexes with hyaluronan. The covalently linked, negatively charged glycosaminoglycan chains in aggrecan produce an osmotic swelling pressure in tissue, and the reversible redistribution of proteoglycan-bound water under loading gives cartilage its characteristic mechanical properties (review in Hascall, 1988). Similarly, it is likely that the high density of fixed negative charges in versican's glycosaminoglycan side chains result in osmotic activity and contribute to water-binding in the interstitium. The precise mechanical correlates of this would depend on the charge density and other matrix elements associated with versican; in cartilage, for instance, aggrecan complexes are embedded in type II collagen and can swell only until osmotic swelling pressure balances tension in the collagenous network. Versican may modulate extracellular matrix hydration and mechanics, and thus influence the progression of disease and delineate architectural structures in the remodeling lung. 7.7 Versican Influence on Leukocyte Localization In initial studies, it appeared that inflammatory cells were less abundant within versican-rich domains. Double immunohistochemical staining for versican and the leukocyte common antigen showed that leukocytes were significantly less abundant in versican-rich domains. In contrast, mononuclear phagocytes, differentiated from other leukocytes by their very light leukocyte common antigen staining and positive CD68 staining, were significantly more abundant in versican-rich domains than in versican-poor domains. These data suggest that versican, or an associated molecule, may selectively influence the adhesion and/or migration of 79 specific subsets of leukocytes, and thus influence the disease process. Versican binding to hyaluronan might function to block binding of cell membrane bound hyaluronan receptors, such as CD44, by occupying binding epitopes on hyaluronan. This might allow increased cellular passage through the hyaluronan-rich zones and down regulate cellular adhesion to hyaluronan. The C-terminus of versican contains two epidermal growth factor-like repeats, a lectin-like module and a complement regulatory protein-like module (Zimmermann and Rouslahti, 1989). The lectin module binds heparan sulfate and tenascin R (through its carbohydrate) and is likely to be involved in cell-extracellular matrix recognition. Versican has been previously shown to occur in developmentally active areas (Yamagata et al, 1993) and to be involved in tissue hydration and the modulation of various cell-substrate adhesion systems in vitro (Yamagata et al, 1989). It is likely that versican is playing such a role in the development of lung fibrosis. This role in controlling cell-extracellular matrix adhesion indicates that versican may be important in the control of how and where cells are organized in newly forming fibrotic zones. 7.8 Myofibroblasts and Versican Assoc iat ion in the Remodeling Lung Myofibroblasts are differentially related to smooth muscle cells and fibroblasts, serving both contractile and synthetic functions. They express low, yet variable, levels of alpha smooth muscle actin (Foo et al, 1992). Kuhn and McDonald (1991) have previously identified myofibroblasts, based on their expression of alpha smooth muscle actin and vimentin, but not desmin, in the fibroblast foci of idiopathic pulmonary fibrosis lungs. The staining patterns and morphological appearance of the cells we observed in the versican-rich fibroblast foci of the TPF lung are consistent with the previous characterization of myofibroblasts in IPF (Kuhn et al, 1989; Kuhn and McDonald, 1991). Versican deposition in association with proliferating 80 myofibroblasts appears to be a novel component early in the progression of lung fibrosis. These cells were not observed in the mature, collagen rich areas of the fibrotic lung matrix. The intensity of versican staining in the matrix around myofibroblasts makes it very difficult to unequivocally identify intracellular staining for this proteoglycan in the myofibroblasts. However, versican is a very large, hyaluronan-binding molecule. The versican-rich zones of the IPF lung contain hyaluronan. The sheer size of versican and the fact that hyaluronan, an immobile ligand for versican, is found in these regions suggest that versican does not diffuse very far through the matrix after it is synthesized. It is therefore reasonable to suggest that the myofibroblasts identified in close association with these versican-rich fibroblast foci are the cells responsible for versican synthesis. This hypothesis remains to be confirmed with in situ hybridization localization of versican m R N A within these myofibroblasts. 7.9 Decorin and Biglycan Localization in the Remodeling Lung Immunohistochemical staining, using an antibody specific for the proteoglycan decorin was used to localize this chondroitin sulfate/ dermatan sulfate proteoglycan in the fibrotic lung. Decorin localized, extracellularly, to collagen rich zones towards the center and away from the collagen-poor, versican-rich fibrotic foci in the IPF lung. Decorin was not observed extracellularly within the versican-rich regions of the fibrotic lung, but was observed intracellularly in these regions. Decorin synthesis and deposition is likely to occur in concert with the synthesis and deposition of collagen in fibrosis. Decorin has been shown to specifically bind to collagen fibrils and may be involved in fibrillogenesis (Vogel et al, 1984). Decorin may be synthesized in conjunction with the collagen that is laid down in the later stages of fibrosis. It may be that decorin synthesis is down regulated in the early stages of fibrotic lung development 81 and up regulated in the collagen I prevalent, mature fibrotic stage, where decorin may be involved with collagen fibril formation. Immunohistochemistry, using an antibody specific for the dermatan sulfate proteoglycan biglycan was used to localize biglycan in the fibrotic lung. Biglycan localized to some of the same collagen-rich regions as decorin, in the disease lung, as well as to other distinct foci. Biglycan stained diffusely and extracellularly within the versican-rich fibroblast foci in the fibrotic lung and more intensely in the collagen dense regions of the diseased lung. Biglycan has no known function as yet, so the significance of its occurrence in fibrosis is not clear. Biglycan has a distinct localization pattern in idiopathic fibrosis, particularly to the walls of small vessels that permeate the fibrotic regions. Its staining was more intense than decorin staining, yet its pattern is not as diffuse as that of decorin. Biglycan may be involved in the tissue restructuring in either an active manner or as a purely structural component. 7.10 Versican and Elastic Fiber Co-localization Elastic fibers, visualized histochemically in fibrotic tissue using the modified aldehyde-fuchsin Gomori trichrome stain, were often surrounded by deposits of versican, suggesting a specific interaction between versican and components of lung elastic fibers and implying specific synthesis of versican around lung elastic fibers in idiopathic pulmonary fibrosis. In the normal lung, low levels of versican were detectable in proximity to airway and vascular smooth muscle, but a more widespread association with elastin is not apparent. Previous studies of versican in human dermis have demonstrated a co-distribution of versican with elastic fibers, indicating a possible association between versican and elastic microfibrils (Zimmermann et al, 1994). This co-localization of elastic fibers and versican was observed in the idiopathic pulmonary fibrosis 82 lung and poses an interesting paradox. Elastin is not believed to be synthesized to any large extent in the normal lung after the end of lung growth (Shapiro et al, 1991), but these areas are newly synthesized interstitium. The fact that these elastic fibers were larger than any observed in the normal lung suggests that they are abnormal. Whether or not these elastic fibers are newly synthesized is not known, and there is no literature documenting elastin synthesis in diseased adult human lung. 7.11 Growth Factors Likely to Influence the Local Synthesis of Proteoglycans and Hyaluronan in Fibrosis Transforming growth factor-(3 (TGF-P) and platelet-derived growth factor (PDGF) induce increased synthesis of the proteoglycan versican by smooth muscle cells in vitro and TGF-p 1 alters versican's subsequent glycosylation, increasing the attached glycosaminoglycan chain length (Schonherr et al, 1991). This has direct affects on the resultant versican molecules, effectively increasing each proteoglycan's hydrodynamic size and increasing their local charge density. TGF-(3 (Westergren-Thorsson et al, 1990), P D G F (Papakonstantinou et al, 1995) and Interleukin-1 (Sampson et al, 1992) increase the synthesis of the glycosaminoglycan hyaluronan by fibroblasts in vitro and may be responsible for increased hyaluronan in lung inflammation. Increased TGF-J3 levels are associated with increased synthesis of the proteoglycans biglycan and versican and decreased levels of the proteoglycan decorin in cultured human fibroblasts (Kahari et al, 1991). In an experimental model of kidney fibrosis, increased levels of TGF-(3 are associated with an initial increase in decorin levels, followed by an ultimate decrease. Intravenous application of decorin in this model ameliorates the extracellular matrix synthesizing 83 effects of TGF-(3 and implies a binding and inhibition of the growth factor by decorin, thus presenting a model for a negative feedback loop. 7.12 Future Directions A number of future studies are indicated as a result of our investigations. Further elucidation of the roles of versican and hyaluronan in the fibroproliferative processes of IPF might be accomplished through the use of experimental models of disease, such as bleomycin-induced pulmonary fibrosis, using transgenic mice that are deficient for either or both of these extracellular matrix molecules. Versican knock out mice die early in development (Mjaatvedt et al, 1998). Comparison of levels of induced fibrosis in transgenic, versican-deficient and/or hyaluronan-deficient mice with the levels of induced fibrosis in control, non-transgenic mice might lead to clarification of the roles of versican and hyaluronan in the disease process. These experiments, in combination with investigations of the effects of versican and hyaluronan in cell culture, might also further clarify the roles of these extracellular matrix molecules in cell-matrix adhesion and cell mobilization. The hypothesis that myofibroblasts are the cells responsible for versican synthesis could be confirmed with in situ hybridization localization of versican m R N A within these myofibroblasts found in the characteristic fibroblast foci of the IPF lung. Further investigation of the exact content of the elastic fibers found to co-localize with versican in the fibrotic foci of the IPF lung would determine i f the elastic fibers are newly synthesized and how this relates to versican synthesis and deposition. These experiments, in combination with the transgenic mice experiments might further clarify the key processes in the conversion of fibroblasts and 84 myofibroblasts from "provisional" hyaluronan and versican-rich matrix synthesis to collagen synthesis and deposition in later stages of the disease process. References 85 1. Adler K B , Craighead JE, Vallyathan N V , Evans J N . Actin-containing cells in human pulmonary fibrosis. A m J Pathol 1981; 102:427-437. 2. Adler K B , Low R B , Leslie K O , Mitchell J, Evans J N . Contractile cells in normal and fibrotic lung. Lab Invest 1989; 60:473-485. 3. Alho A M , Underhill C B . The hyaluronate receptor is preferentially expressed on proliferating epithelial cells J Ce l l B i o l 1989; 108: 1557-1565. 4. Barry FP, Rosenberg L C , Gaw J U , Koob TJ , Neame PJ. N - and O-linked keratan sulfate on the hyaluronan binding region of aggrecan from mature and immature bovine cartilage. J B i o l Chem 1995; 270(35):20516-20524. 5. Bensadoun E S , Burke A K , Hogg JC, Roberts C R . Proteoglycan deposition in pulmonary fibrosis. A m J Respir Crit Care M e d 1996; 154:1819-1828. 6. Bensadoun E S , Burke A K , Hogg JC, Roberts C R . Proteoglycans in granulomatous lung diseases. Eur Respir J 1997; 10:2731-2737. 7. Bernfield M , Sanderson R D . Syndecan, a developmentally-regulated cell surface proteoglycan, that binds extracellular matrix and growth factors. Phi l Trans R Soc B Lond 1990; 327:171-186. 8. Bhattacharya J, Cruz T, Bhattacharya S, Bray B A . Hyaluronan affects extravascular water in lungs of unanaesthetized rabbits. J Appl Physiol 1989; 66:2595-2599. 86 9. Bianco P, Fisher L W , Young M F , Termine JD, Gehron-Robey P. Expression and localization of the two small proteoglycans biglycan and decorin in developing human skeletal and non-skeletal tissue. J Histochem Cytochem 1990; 38:1549-1563. 10. Bidanset D J , Guidry C , Rosenberg L C , Choi H U , Timpl R, Hook M . Binding of the proteoglycan decorin to collagen type V I . J B i o l Chem 1992; 267:5250-5256. 11. Bjermer L , Lundgren R, Hallgren R. Hyaluronan and type HI procollagen propeptide concentrations in bronchoalveolar lavage fluid in idiopathic pulmonary fibrosis. Thorax 1989; 44:126-131. 12. Border W A , Noble N A , Yamamoto T, Harper JR, Yamaguchi Y , Pierschbacher M D , Ruoslahti E . Natural inhibitor of transforming growth factor-beta protects against scarring in experimental kidney disease. Nature 1992; 360: 361-364. 13. Boudreau N , Turley E , Rabinovitch M . Fibronectin, hyaluronan and a hyaluronan binding protein contribute to increased ductus arteriosus smooth muscle migration. Dev B i o l 1991; 143:235-247. 14. Bray B A , Sampson P M , Osman M , Giandomenico, Turino G M . Early changes in lung disease hyaluronan (hyaluronic acid) and hyaluronidase in bleomycin-induced alveolitis in hamsters. A m Rev Respir P i s 1991; 18:1466-1472. 15. Brown D C , Vogel K G . Characteristics of the in vitro interaction of a small proteoglycan (PGII) of bovine tendon with type I collagen. Matrix 1989; 9:468-478. 16. Burke A K , Roberts C R . Proteoglycans in lung fibrosis: Deposition of versican, hyaluronan, biglycan and decorin in usual form of interstitial pneumonia (UIP). (submitted). 17. Cantin A M , North S L , Fells G A . Oxidant-mediated epithelial cell injury in idiopathic pulmonary fibrosis. J C l i n Invest 1987; 79(6): 1665-1673. 87 18. Cantor JO, Bray B A , Ryan SF, Mandl I, Turino G M . Glycosaminoglycan and collagen synthesis in N-nitroso-N-methylurethane-induced pulmonary fibrosis. Proc Soc Exp B i o l M e d 1980; 164:1.-8. 19. Cantor JO, Cerreta J M , Osman M , Mott S H , Mandl I, Turino G M . Glycosaminoglycan synthesis in bleomycin-induced pulmonary fibrosis: biochemistry and autoradiography. Proc Soc Exp B i o l M e d 1983; 174:172-181. 20. Cantor JO, Osman M , Cerreta J M , Mandl I, Turino G M . Glycosaminoglycan synthesis in explants derived from bleomycin-treated fibrotic hamster lungs. Proc Soc Exp B i o l M e d 1983; 173:362-366. 21. Carrington C B , Gaensler E A , Coutu R E , Fitzgerald M X , Gupta R G . Natural history and treated course of usual and desquamative interstitial pneumonia. N Engl J M e d 1978; 298:801-809. 22. Caterson B , Baker JR, Christner JE, Lee Y , Lentz M . Monoclonal antibodies as probes for determining the microheterogeneity of the link proteins of cartilage proteoglycan. J B i o l Chem 1985;260:11348-11356. 23. Cegla U H , Kro id l R F , Meier-Sydow J, Thiel C , Czarnecki G , Schreiber F . Therapy of the idiopathic fibrosis of the lung. Experiences with three therapeutic principles corticosteroids in combination with azathioprine, D-penicillamine, and para-amino-benzoate. Pneumonologie 1975; 152:75-92. 24. Chapman JR, Charles PJ, Venables PJW. Definition and clinical relevance of antibodies to nuclear ribonucleoprotein and other nuclear antigens in patients with cryptogenic fibrosing alveolitis. A m Rev Respir P i s 1984; 130(3):439-443. 88 25. Chomczynski P. A method for the single-step simultaneous isolation of R N A , D N A and proteins from cell and tissue samples. BioTechniques 1993; 15(3):532-537. 26. David G . Integral membrane heparan sulfate proteoglycans. F A S E B J 1993; 7:1023-1030. 27. Davis G S . Idiopathic pulmonary fibrosis, in Current therapy of respiratory disease, 2 n d edition. Cherniack R M and Decker B C , eds. C V Mosby Co. St. Louis, Missouri,1986:161-170. 28. Doege K , Sasaki M , KimuraT, Yamada Y . Complete coding sequence and deduced primary structure of the human cartilage large aggregating proteoglycan, aggrecan. J B i o l Chem 1991; 266: 894-902. 29. Dours-Zimmermann M T , Zimmermann D R A novel glycosaminoglycan attachment domain identified in two alternative splice variants of human versican. J B i o l Chem 1994; 269:32992-32998. 30. Fisher L W , Hawkins G R , Tuross N , Termine JD. Purification and partial characterization of small proteoglycans I and n, bone sialoproteins I and n, and osteonectin from the mineral compartment of developing human bone. J B i o l Chem 1987; 262:9702-9708. 31. Fisher L W , Termine JD, Dejter S W , Whitson S W , Yamagishita M , Kimura J H , Hascall V C , Kleinman H K , Hassell JR, Nilsson B . Proteoglycans of the developing bone. J B i o l Chem 1983; 258:6588-6594. 32. Fisher L W , Termine JD, Young M F . Deduced protein sequence of bone small proteoglycan I (biglycan) shows homology with proteoglycan II (decorin) and several non-connective tissue proteins in a variety of species. J B i o l Chem 1989; 264:4571-4576. 89 33. Fleischmajer R, Fisher L W , MacDonald E D , Jacobs L , Perlish JS, Termine JD. Decorin interacts with fibrillar collagen of embryonic and adult human skin. J Struct B i o l 1991; 106:82-90. 34. Foo ITH, Naylor IL, Timmons M J , Trejdosiewicz L K . Intracellular actin as a marker for myofibroblasts in vitro. Lab Invest 1992; 67(6):727-733. 35. Fraser R G , Pare J A , Pare P D , Fraser R S , Genereux G P . Diagnosis of diseases of the lung, 3 r d edition. W . B . Saunders Co., 1991. 36. Goetinck PF , Stirpe N S , Tsonis P A , Carlone D . The tandemly repeated sequences of cartilage link protein contain the sites for interaction with hyaluronic acid. J B i o l Chem 1987; 105:2403-2408. 37. Green SJ, Tarone G , Underhill C B . Aggregation of macrophages and fibroblasts is inhibited by a monoclonal antibody to the hyaluronate receptor. Exp Cel l Res 1988; 178:224-232. 38. Green SJ, Tarone G , Underhill C B . Distribution of hyaluronate and hyaluronate receptors in the adult lung. J Ce l l Sci 1988; 89:145-156. 39. Grover J, Roughley PJ. Versican gene expression in human articular cartilage and comparison of m R N A splicing variation with aggrecan. Biochem J 1993; 291(2):361-367. 40. Halami N S . Differentiation of the two types of basophils in an adenohypophysis in the rat and the mouse. Stain Technol 1952; 27:61-66. 41. Hallgren R, Eklund A , Engstrom-Laurent A , Schmekel B . Hyaluronate in bronchoalveolar lavage fluid: a new marker in sarcoidosis reflecting pulmonary disease. Brit M e d J 1985; 290:1778-1781. 42. Hallgren R, Samuelsson T, Laurent T C , Mod ig J, Accumulation of hyaluronan in the lung in adult respiratory distress syndrome. A m Rev Respir P i s 1989; 139:682-687. 90 43. Hascall V C . Proteoglycans: the chondroitin sulfate/ keratan sulfate proteoglycan of cartilage. ISI Atlas of Sci : Biochem 1988; 1:189-198. 44. Hernnas J, Nettelbladt O, Bjermer L , Sarnstrand B , Malmstrom A , Hallgren R. Alveolar accumulation of fibronectin and hyaluronan precedes bleomycin-induced pulmonary fibrosis in the rat. Eur Respir J 1992; 5:404-410. 45. Hogg JC . Chronic interstitial lung diseases of unknown origin: a new classification based on pathogenesis. A m J Roent 1991; 156:225-233. 46. Ignotz R A , Massague J. Transforming growth factor-beta stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J B i o l Chem 1986;261:4337-4345. 47. Izumi T, Kitaichi M , Nishimura K , Nagai S. Bronchiolitis obliterans organizing pneumonia. Clinical features and differential diagnosis. Chest 1992; 102(3):715-719. 48. Juul SE , Kinsella M G , Wight T N , Hodson W A . Alterations in nonhuman primate (M. nemestrina) lung proteoglycans during normal development and acute hyaline membrane disease. A m J Respir Cel l M o l B i o l 1993; 8:299-310. 49. Kahari V - M , Larjava H , Uitto J. Differential regulation of extracellular matrix proteoglycan (PG) gene expression. J B i o l Chem 1991; 266(16): 10608-10615. 50. Karlinsky JB , Glycosaminoglycans in emphysematous and fibrotic hamster lungs. A m Rev Respir P i s 1982; 125:85-93. 51. Khal i l N , O'Connor R N , Unruh H W , Warren P W , Flanders K C , Kemp A , Bereznay O H , GreenbergAH. Increased production and immunohistochemical localization of transforming growth factor-beta in idiopathic pulmonary fibrosis. A m J Respir Cel l M o l B i o l 1991; 5:155-162. 91 52. Kha l i l N , O'Connor R, Unruh H , Warren P, Kemp A , Greenberg A . Enhanced expression and immunohistochemical distribution of transforming growth factor-(3 in idiopathic pulmonary fibrosis. Chest 1991; 99(3):65S-66S. 53. Krusiuis T, Ruoslahti E . Primary structure of an extracellular matrix proteoglycan core protein deduced from cloned c D N A . Proc Natl Acad Sci U S A 1986; 83:7683-7687. 54. Kuhn C . Patterns of Lung Repair. A Morphologist's View. Chest 1991; 99(3): 11S-14S. 55. Kuhn C , Boldt J, King T E Jr, Crouch E , Vartio T, McDonald J A . A n immunohistochemical study of architectural remodeling and connective tissue synthesis in pulmonary fibrosis. A m Rev Respir P i s 1989; 140:1693-1703. 56. Kuhn C , M c P o n a l d J A . The roles of the myofibroblast in idiopathic pulmonary fibrosis. Ultrastructural and immunohistochemical features of sites of active extracellular matrix synthesis. A m J Pathol 1991; 138:1257-1265. 57. Laurent T C , Fraser J R E . Hyaluronan. F A S E B J 1992; 6:2397-2404. 58. LeBaron R G , Zimmermann P R , Ruoslahti E . Hyaluronate binding properties of versican. J B i o l Chem 1992; 267(14): 10003-10010. 59. Libby P M , Gibofsky A , Fotino M . hrimunogenetic and clinical findings in idiopathic pulmonary fibrosis: Association with the B-cel l alloantigen H L A - P R 2 . A m Rev Respir P i s 1983; 127:618-622. 60. Limper A H , Broekelmann TJ , Colby T V , Mal i z i a G , McPona ld J A . Analysis of local m R N A expression for extracellular matrix proteins and growth factors using in situ hybridization in fibroproliferative lung disorders. Chest 1991; 99(3):55S-56S. 61. Mann P M , Yamaguchi Y , Bourdon M A , Ruoslahti E . Analysis of glycosaminoglycan substitution in decorin by site-directed mutagenesis. J B i o l Chem 1990; 265(9):5317-5323. 92 62. Martinet Y , Rom W N , Grotendorst G R , Martin G R , Crystal R G . Exaggerated spontaneous release of platelet-derived growth factor by alveolar macrophages from patients with idiopathic pulmonary fibrosis. N Engl J M e d 1987; 317: 202-209. 63. Mayer P. Notiz iiber hamatein und hamalaun. Zeitschrift fur wissenschaftliche mikroskopie und fur mikroskopische technick. 1903; 20:409-411. 64. McBride O W , Fisher L W , Young M F . Localization of PGI (biglycan, B G N ) and P G H (decorin, D C N , PG-40) genes on chromosome Xql3-qter and 12q, respectively. Genomics 1990; 6:219-225. 65. McDonald J, Broekelmann T, Matheke M , Crouch E , Koo M , Kuhn C . A monoclonal antibody to the carboxyterminal domain of procollagen type I visualized collagen synthesizing fibroblasts. J C l i n Invest 1986; 78:1237-1244. 66. Mjaatvedt C H , Yamamura H , Capehart A A , Turner D , Markwald R R . The Cspg2 gene, disrupted in the hdf mutant, is required for right cardiac chamber and endocardial cushion formation. Dev B i o l 1998; 202(l):56-66. 67. Motomiya M , Arai H , Sato H , Yokosawa A , Nagai H , Konno K . Increase of dermatan sulfate in a case of pulmonary fibrosis. Tohuko J Exp M e d 1975; 115:361-365. 68. Murdoch A D , Dodge G R , Cohen I, Tuan, R S , Iozzo R V . Primary structure of the human heparan sulfate proteoglycan from basement membrane (HSPG2/perlecan). J B i o l Chem 1992; 267: 8544-8557. 69. Myers JL , Katzenstein A L A . Ultrastructural evidence of alveolar epithelial injury in idiopathic bronchiolitis obliterans-organizing pneumonia. A m J Pathol 1988; 132(1): 102-109. 93 70. Naso M F , Zimmermann D R , Iozzo R V Characterization of the complete genomic structure of the human versican gene and functional analysis of its promoter. J B i o l Chem 1994; 269:32999-33008. 71. Nerlich A G , Nerlich M L , Muller P K . Pattern of collagen in acute post-traumatic pulmonary fibrosis. Thorax 1987; 42(ll):863-869. 72. Nettelbladt O, Bergh J, Schenholm M , Tengblad A , Hallgren R. Accumulation of hyaluronic acid in the alveolar interstitial tissue in bleomycin-induced alveolitis. A m Rev Respir P i s 1989; 139:759-762. 73. Nettelbladt O, Hallgren R. Hyaluronan (hyaluronic acid) in bronchoalveolar lavage fluid during the development of bleomycin-induced alveolitis in the rat. A m Rev Respir P i s 1989; 140:1028-1032. 74. Nettelbladt O, Tengblad A , Hallgren R. Lung accumulation of hyaluronan parallels pulmonary edema in experimental alveolitis. A m J Physiol 1989; 257: L379-L384. 75. Papakonstantinou E , Karakiulakis G , Roth M , Block L H . Platelet-derived growth factor stimulates the secretion of hyaluronic acid by proliferating human vascular smooth muscle cells. Proc Natl Acad Sci U S A (USA) 1995; 92(21):9881-9885. 76. Pulford K A F , Rigney E M , Mick lem K J , Jones M , Stross W P , Gatter K C , Mason P Y . K P 1 : a new monoclonal antibody that detects a monocyte/macrophage associated antigen in routinely processed tissue sections. J C l in Pathol 1989; 42:414-421. 77. Raghu G , Striker L J , Hudson L P , Striker G E . Extracellular matrix in normal and fibrotic human lungs. A m Rev Respir P i s 1985; 131 (2):281-289. 94 78. Ripellino J A , Klinger M M , Margolis R U , Margolis R K . The hyaluronic acid binding region as a specific probe for the localization of hyaluronic acid in tissue sections. J Histochem Cytochem 1985; 33:1060-1066. 79. Roberts C R . Is asthma a fibrotic disease? Chest 1995; 107:111S-117S. 80. Roberts C R , Burke A B . Remodeling of the extracellular matrix in asthma: proteoglycan synthesis and degradation. Can Resp J 1998; 5(l):48-50. 81. Roberts C R , Pare P D . Composition changes in human tracheal cartilage in growth and aging, including changes in proteoglycan structure. A m J Physiol 1991; 261:L92-L101. 82. Roberts C R , Rains J K , Pare P D , Walker D C , Wiggs B , Bert JL . Ultrastructure and tensile properties of human tracheal cartilage. J Biomechanics 1998; 31(l):81-86. 83. Roberts C R , Roughley PJ, Mort JS. Degradation of human proteoglycan aggregate induced by hydrogen peroxide. Protein fragmentation, amino acid modification and hyaluronic acid cleavage. Biochem J 1989; 259:805-811. 84. Roberts C R , Wight T N , Hascall V C . Proteoglycans, in The Lung: Scientific Foundations, 2 n d edition. Crystal R G , West JB et al, eds. Lippincott-Raven Publishers, Philadelphia P A , 1997:757-767. 85. Sampson P M , Rochester C L , Freundlich B , Elias J A . Cytokine regulation of human lung fibroblast hyaluronan (hyaluronic acid) production. Evidence for cytokine-regulated hyaluronan (hyaluronic acid) degradation and human lung fibroblast-derived hyaluronidase. J C l i n Invest 1992; 90:1492-1503. 86. Scadding J G , Hinson K F . Diffuse fibrosing alveolitis (diffuse interstitial fibrosis of the lungs). Correlation of histology at biopsy with prognosis. Thorax 1967; 22(4):291-304. 95 87. Schmidt G , 'Hausser H , Kresse H . Interaction of the small proteoglycan decorin with fibronectin. Involvement of the sequence N K I S K of the core protein. Biochem J 1991; 280:411-414. 88. Schonherr E , Jarvelainen H T , Sandell L J , Wight T N . 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 B i o l Chem 1991; 266:17640-17647. 89. Scott JE, Orford C R . Dermatan sulfate-rich proteoglycan associates with rat tail-tendon collagen at the d band in the gap region. Biochem J 1981; 197:213-216. 90. Shapiro S D , Endicott S K , Province M A , Pierce J A , Campbell E J . Marked longevity of human lung parenchymal elastic fibers deduced from prevalence of D-aspartate and nuclear weapons-related radiocarbon. J C l i n Invest 1991; 87:1828-1834. 91. Solliday N H , Will iams J A , Gaensler E A , Coutu R E , Carrington C B . Familial chronic interstitial pneumonia. A m Rev Respir P i s 1973; 108(2): 193-204. 92. Stack B H R , Grant I W B , Irvine W J . Idiopathic diffuse interstitial lung disease: A review of 42 cases. A m Rev Respir P i s 1965; 92(6):939-948. 93. Tengblad A . Quantitative analysis of hyaluronate in nanogram amounts. Biochem J 1980; 185:101-105. 94. Thomas L , Etoh T, Stamenkovic I, M i h m M C , Byers H R . Migration of human melanoma cells on hyaluronate is related to C P 4 4 expression. J Invest Permatol 1993; 100:115-120. 95. Toole, B P . Proteoglycans and hyaluronan in moiphogenesis and differentiation, in Cel l biology of the extracellular matrix 2 n d edition. Hay E , ed. New York: Plenum Press, 1991:305-342. 96 96. van Kuppevelt T H , Cremers FP , Domen JG, van Beuningen H M , van den Brule A J , Kuyper C M . Ultrastructural localization and characterization of proteoglycans in human lung alveoli. Eur J Cel l B i o l 1985; 36:74-81. 97. van Kuppevelt T M S M , Janssen H M J , van Beuningen H M , Cheung K S , Schijen M M A , Kuyper C M A , Veerkamp J H . Isolation and characterization of a collagen fibril-associated dermatan sulfate proteoglycan from bovine lung. Biochim Biophvs Acta 1987; 926:296-309. 98. van Kuppevelt T H M S M , van Beuningen H M , Rutten T L , van den Brule A J , Kuyper C M A . Further characterization of a large proteoglycan in human lung alveoli. Eur J Ce l l B i o l 1985; 39:386-390. 99. Vergnon J M , Vincent M , de The G , Mornex JF, Weynants P, Brune J. Cryptogenic fibrosing alveolitis and Epstein-Barr virus: an association? Lancet 1984; 2(8406):768-771. 100. V i l l i m V , Hurych J. Sulfated glycosaminoglycans in rat lung tissue during experimental silicotic fibrosis. Connect Tissue Res 1987; 16:27-40. 101. Vogel K G , Paulson M , Heinegard D . Specific inhibition of type I and type II collagen fibrillogenesis by the low molecular mass proteoglycan of tendon. Biochem J 1984; 223:587-597. 102. Wagner JC , Wusteman FS, Edwards J H , H i l l RJ . The composition of massive lesions in coal miners. Thorax 1975; 30:382-388. 103. Wegrowski J, Lefaix J L , Lafuma C . Accumulation of glycosaminoglycans in radiation-induced muscular fibrosis. Int J Radiat B i o l 1992; 61(5):685-693. 104. Westergren-Thorsson G Sarnstrand B , Fransson L A , Malmstrom A . TGF-beta enhances the production of hyaluronan in human lung fibroblasts but not in skin fibroblasts. Exp Cel l Res 1990;186:192-195. 97 105. Wight T N . Cel l biology of arterial proteoglycans. Arteriosclerosis 1989; 9:1-20. 106. Wight T N , Heinegard D K , Hascall V C . Proteoglycans: Structure and function in Ce l l biology of extracellular matrix (2 n d ed.). Hay E D , ed. Plenum Press, 1991:45-78. 107. Wusterman FS, Gold C, Wagner JC. Glycosaminoglycans and calcification in the lesions of progressive massive fibrosis and in pleural plaques. A m Rev Respir P i s 1972; 106:116-118. 108. 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. 109. Yamagata M , Suzuki S, Akiyama S, Yamada K , Kimata K . Regulation of cell-substrate adhesion by proteoglycans immobilized on extracellular substrates. J B i o l Chem 1989; 264:8012-8018. 110. Yamaguchi Y , Mann D M , Ruoslahti E R . Negative regulation of transforming growth factor beta by the proteoglycan decorin. Nature 1990; 346:281-284. 111. Yoshida A , Hiramatsu M , Hatakeyama K , Minami N . Elevation of glucosamine 6-phosphate synthetase activity in bleomycin-induced pulmonary fibrosis in hamsters. J Antibiotics 1982; 35(7):882-885. 112. Zhang K , Rehkter M D , Gordon D , Phan S H . 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. 113. Zimmermann D R , Dours-Zimmermann M T , 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 Cel l B i o l . 1994; 124:817-825. 114. Zimmermann D R , Ruoslahti E R . Multiple domains of the large fibroblast proteoglyi versican. E M B O J . 1989; 8:2975-2981. 99 Appendix 1: Patient Data Series 1 IPF Patient Samples for Immunohistochemical Ana lys is : Patient Number Age (years) Sex (M/F) Pathological Diagnosis Pathological Stage Duration of Symptoms (SOB) 1380 38 M Patchy interstitial fibrosis and inflammation M i d to late stage: Honeycomb lung 6-8 months 1446 65 M Idiopathic pulmonary fibrosis M i d to late stage 2+ years 1533 83 F Idiopathic pulmonary fibrosis and pleural microcysts M i d to late stage 1 year 1595 70 M Honeycomb lung (IPF) Late stage 5 years (worsening over last year) 1920 63 M Honeycomb lung (IPF) with secondary infection M i d to late stage 2 months 2089 74 F Idiopathic pulmonary fibrosis and acute DAD Late stage 1+ year (3 months of rapid deterioration) Series 1 Control Patient Samples for Immunohistochemical Ana lys is : Patient Age Sex Pathological Diagnosis Number (years) (M/F) 1430 42 F Moderately differentiated adenocarcinoma 1653 81 F Moderately differentiated adenocarcinoma 1773 77 F Poorly differentiated adenocarcinoma 1876 66 M Moderately to poorly differentiated squamous cell carcinoma 21220 61 F Moderately differentiated adenocarcinoma 8984 65 M Intracerebral hemorrhage Series 3 Patient Samples for mRNA Analys is : Patient Number Pathological Diagnosis 2432 Diffuse Alveolar Damage 2077 Diffuse Alveolar Damage 1975 Idiopathic Pulmonary Fibrosis 1966 Idiopathic Pulmonary Fibrosis 2422 Idiopathic Pulmonary Fibrosis 2752 Idiopathic Pulmonary Fibrosis 2106 Bronchiolitis Obliterans Organizing Pneumonia 2003 Bronchiolitis Obliterans Organizing Pneumonia 2443 Bronchiolitis Obliterans Organizing Pneumonia 3007 Bronchiolitis Obliterans Organizing Pneumonia Series 3 Control Samples for mRNA Analys is : Patient Number Pathological Diagnosis F E V 1 (percent predicted) * F V C (percent predicted) ** 1804 Carcinoid 96.4% 108.6% 1834 Squamous Cel l Carcinoma 98.9% 110.6% 1862 Carcinoma 97.5% 114.9% 1873 Adenocarcinoma 97.9% 98.0% 1977 Adenocarcinoma 112.5% 106.0% F E V 1 = Forced expiratory volume in 1 second * F V C = Forced vital capacity 101 Appendix 2: List of Published Papers and Abstracts Papers: 1. Bensadoun E S , Burke A K , Hogg JC, Roberts C R . Proteoglycan deposition in pulmonary fibrosis. A m J Respir Crit Care M e d 1996; 154:1819-1828. 2. Bensadoun E S , Burke A K , Hogg JC, Roberts C R . Proteoglycans in granulomatous lung diseases. Eur Respir J 1997; 10:2731-2737. 3. Roberts C R , Burke A K . Increased synthesis and altered splicing of m R N A for the interstitial proteoglycan versican is associated with remodeling in inflammatory lung disease. Proceedings of the 1997 Bioengineering Conference B E D 1997; 35:221-222. 4. Roberts C R , Burke A K . Remodeling of the extracellular matrix in asthma: proteoglycan synthesis and degradation. Cdn Respir J 1998; 5(l):48-50. 5. Roberts C R , Burke A K . Synthesis of the proteoglycan versican in pulmonary fibrosis. Int J Pathol in Press. 6. Roberts C R , Burke A K . Synthesis of the proteoglycan versican in pulmonary fibrosis. Revised manuscript submitted to A m J Respir Cel l M o l B i o l . 7. Roberts C D , Burke A K , Hammil l D . Synthesis of the proteoglycan versican associated with myofibroblasts in airway remodeling in asthma. Manuscript in preparation for submission to A m J Respir Ce l l M o l B i o l . 8. Roberts C R , Hammil l D , Burke A K . Remodeling of airway cartilage associated with fatal asthma. Revised manuscript in preparation for submission to A m J Respir Cel l M o l B i o l . 102 Abstracts : 1. Bensadoun E , Burke A K , Hogg JC, Roberts C R . Proteoglycan deposition in bronchiolitis obliterans organizing pneumonia. A m J Respir Crit Care M e d 1995; 151:A54. 2. Bensadoun E , Burke A K , Hogg JC, Roberts C R . The deposition of proteoglycans in diffuse alveolar damage. A m J Respir Crit Care M e d 1995; 151:A73. 3. Roberts C R , Bensadoun E , Burke A K , Hogg JC . The proteoglycan versican is associated with remodeling in lung fibrosis. A m J Respir Crit Care M e d 1996; 153:A310. 4. Roberts C R , Burke A K . Increased synthesis and altered splicing of versican proteoglycan m R N A in human lung fibrosis. A m J Respir Crit Care M e d 1997; 155:A185. 5. Roberts C R , Burke A K . Proteoglycans and hyaluronan in human lung fibrosis. F A S E B J 1994; 8:4003. 6. Roberts C R , Burke A K , Bensadoun E , Hogg JC . Deposition of versican and other proteoglycans in the extracellular matrix in lung inflammation and fibrosis. A m J Respir Crit Care M e d 1995; 151:A561. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088954/manifest

Comment

Related Items