TYPE XIII COLLAGENCharacterization of ectodomain shedding and its biological implications in mammalian cells, characterization of type XIII collagen expression in human cancers

MARJA-RIITTAVÄISÄNEN

Faculty of Medicine,Department of Medical Biochemistry

and Molecular Biology,Collagen Research Unit,

Biocenter Oulu,University of Oulu

OULU 2005

MARJA-RIITTA VÄISÄNEN

TYPE XIII COLLAGENCharacterization of ectodomain shedding and its biological implications in mammalian cells, characterization of type XIII collagen expression in human cancers

Academic Dissertation to be presented with the assent ofthe Faculty of Medicine, University of Oulu, for publicdiscussion in Auditorium F101 of the Department ofPhysiology (Aapistie 7), on December 1st, 2005, at 12 noon

OULUN YLIOPISTO, OULU 2005

Copyright © 2005University of Oulu, 2005

Supervised byProfessor Taina Pihlajaniemi

Reviewed byProfessor Staffan JohanssonDocent Sirkku Peltonen

ISBN 951-42-7907-7 (nid.)ISBN 951-42-7908-5 (PDF) http://herkules.oulu.fi/isbn9514279085/

ISSN 0355-3221 http://herkules.oulu.fi/issn03553221/

OULU UNIVERSITY PRESSOULU 2005

Väisänen, Marja-Riitta, Type XIII collagen. Characterization of ectodomainshedding and its biological implications in mammalian cells, characterization of typeXIII collagen expression in human cancersFaculty of Medicine, Department of Medical Biochemistry and Molecular Biology, CollagenResearch Unit, Biocenter Oulu, University of Oulu, P.O.Box 5000, FIN-90014 University of Oulu,Finland 2005Oulu, Finland

AbstractType XIII collagen is an integral membrane protein in type II orientation. In cells and tissues typeXIII collagen has been located in various adhesive structures, like focal adhesions. Due to this, itsbiological role has been implicated in cell adhesion. This collagen also exists as a soluble protein dueto the release of the ectodomain from the plasma membrane.

In this thesis, ectodomain shedding, i.e. enzymatic release of the extracellular domain, was studiedin detail, focusing on the phenomenon as it occurs in mammalian cells. It was found that theectodomain is released by members of the mammalian proprotein convertase family, e.g. furin.Shedding was shown to take place at the cell surface, but based on additional observations, thiscleavage may also take place intracellularly in the Golgi apparatus. Various intracellular mechanisms,depending on cell type, were found to be involved in the regulation of ectodomain shedding.Apparently, due to the liberation of the ectodomain, the level of type XIII collagen on the plasmamembrane is maintained at a relatively even amount.

The released ectodomain was shown to retain biological activity. It showed distinct matrix-specificity so that on vitronectin its influence on cell functions was anti-adhesive, anti-migratory,anti-proliferative and non-supportive of cell spreading. It was also demonstrated to affect thefibronectin matrix assembly in a manner that resulted in reduced amounts of the fibrillar fibronectinmatrix.

A large collection of human epithelial and mesenchymal cancer samples were screened for typeXIII collagen mRNA expression and compared to the expression levels of pre-malignant and normalsamples. It was discovered that malignant transformation upregulates the expression of type XIIIcollagen in mesenchymal cancers and particularly in the stroma of epithelial cancers, more so than incancer epithelia. TGF-β1 was demonstrated as one factor contributing to the stimulation ofexpression. Based on cell culture experiments in this study, it was also deduced that the upregulatedexpression of type XIII collagen and the concomitant shedding of the ectodomain can remodel thetumour stroma, making it inauspicious for adhesion-dependent cell functions, particularly invitronectin-rich milieu.

Keywords: cancer, collagen, ectodomain, extracellular matrix, fibronectin, shedding

To my late father

Acknowledgements

This study has been done in the Collagen Reasearch Unit of the Department of Medical Biochemistry and Molecular Biology, University of Oulu.

I wish to express my sincere gratitude to my supervisor Professor Taina Pihlajaniemi for giving me the opportunity to work in her group and in the fascinating field of matrix biology. We all get to enjoy the benefits of the great research facilities, to the creation of which she has contributed most. On a personal level I would also like to thank for her patience and faith in this project despite the periods of slow progress. I would also like to express my appreciation to Research Professor emeritus Kari Kivirikko. I consider it a personal honour to have had a chance to be a member of the Collagen Research Unit. All the other group leaders in the Department of Medical Biochemistry and Molecular Biology, namely Professor Leena Ala-Kokko, Professor Seppo Vainio and Docent Johanna Myllyharju, as well as Acting Professor Peppi Karppinen and Professor emeritus Ilmo Hassinen, are also gratefully acknowledged.

I am grateful to Professor Staffan Johansson and Docent Sirkku Peltonen for the very careful, thorough and quick review of this thesis, despite a very tight time schedule. Aaron Bergdahl is also acknowledged for an equally thorough and rapid revision of the language of the manuscript. I would like to expess my sincere thanks to my collaborators in the studies of this thesis: Ph.D. Hongmin Tu for the ever so quick delivery of the recombinant ectodomain, Docent Raija Sormunen for the help with the transmission electron microscopy and Ph.D. Päivi Pirilä for the help with the Biacore™. Somehow I feel it is an understatement just to say “thanks” to Docent Helena Autio-Harmainen, because her contribution has been so extraordinarily valuable. Not only have I got a chance to learn something totally new and different when doing the tissue microarray work, but her encouragement and expertise were downright vital throughout the cancer study - especially during the black period of raging mycoplasma infection. Already at this point, I want to thank my closest collaborator M.Sc. Timo Väisänen. I am short of words in trying to express how unquestionably crucial his contribution to the completion of the many joint studies has been. I am the first one ready to recognize that without Timo´s ideas, expertise, knowledge and persistency this thesis would never have seen the daylight. Never.

I wish to thank all the collageagues in our department, especially the members of the type XIII collagen group, but also others outside this group with whom I have worked during these years, for instance when teaching. Ph.D. Anne Snellman and Ph.D., M.D. Pasi O. Hägg are acknowledged for introducing me to the intial steps of the actual work, both practical and scientific, in this group many years ago.

I have had the priviledge to work with three scillful lab technicians, namely Maija Seppänen, Ritva Savilaakso and Sirkka Vilmi. I wish to thank them all for their excellent and flawless work, with all the flexibility in the world always ready to adjust to changing instructions and plans. Lab technicians Tuulikki Moilanen and Heli Auno from the Department of Pathology are also warmly thanked for their expert work in the cancer study. I thank Pertti Vuokila, Seppo Lähdesmäki, Marja-Leena Kivelä, Auli Kinnunen and Marja Leena Karjalainen for their friendly and efficient way of attending to the many everyday matters concerning our lab. I am truly grateful to the skillful and speedy help of our present computer support personel, namely Risto Helminen, Marko Kervinen and Antti Viklund. The Biocenter personel has always helped readily in numerous matters, for which I wish to thank them.

Outside this lab world, I wish to thank my mother Eila-Leena Keränen for her love, support and encouragement throughout my life in all matters, not only related to this thesis. I am deeply indebted to my sister Tarja-Leena Kircher and her family as well as my brother Juha-Matti Keränen for their love, friendship and neverfailing support. The lovely little Mei Kircher is such a source of joy! I also owe my thanks to my relatives in the Väisänen family for their encouragement and support. I warmly thank my good friend Karin Gebel in Germany for the many years´ of true and wonderful friendship.

And finally, the most important person last. Beyond the work, I have the unique priviledge to share the everyday life also outside the lab with Timo. Again, I find myself totally deprived of words in trying to express my thanks for his love, support and thorough understanding during these years. However, I take comfort in knowing that even if I fail to put it in so many and so right words, you know it, Timo. I know that you know. PP/M.

This work has been supported by the Center of Excellence Programme of the Academy of Finland, the Sigrid Jusélius Foundation, The Cancer Society of Finland and the European Commission Framework 6 Programme (504743).

Oulu, November 2005 Marja-Riitta Väisänen

Abbreviations

BSA bovine serum albumin C- carboxy- cDNA complementary deoxyribonucleic acid CLAC collagen-like Alzheimer amyloid plaque component CLAC-P collagen-like Alzheimer amyloid plaque component precursor CM conditioned medium COL collagenous CWFS cold water fish skin del deletion DMSO dimethylsulphoxide DTT dithiotreitol EC extracellular ECM extracellular matrix EDTA ethylenediaminetetraacetic acid disodium salt EGTA ethyleneglycol-bis(β-aminoethyl)N,N,N´,N´-tetraacetic acid EHS Englebreth-Holm-Swarm Fn fibronectin Fn70K a recombinant fibronectin fragment of MW of 70 kDa (and other fragments accordingly) GAM goat anti-mouse GAR goat anti-rabbit HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulphonic acid Hsp heat shock protein IC intracellular ISH in situ hybridization KD dissociation constant kDa kilodalton mRNA messenger RNA MOPS morpholino-propane sulphonic acid MW molecular weight N- amino-

NC non-collagenous PAGE polyacrylamide gel electrophoresis PBS phosphate-buffered saline PC proprotein convertase PFA paraformaldehyde PMA phorbol 12-myristate 13-acetate RT room temperature RT-PCR reverse transcriptase-polymerase chain reaction SAR swine anti-rabbit SDS sodium dodecyl sulphate SSC citrate-buffered saline TAPI tumour necrosis factor-α protease inhibitor TGF-β transforming growth factor-β TMA tissue microarray Tris tris(hydroxymethyl)aminomethane TRITC tetramethylrhodamine β-isothiocyanate tRNA transfer RNA X (in Gly-X-Y) any amino acid Y (in Gly-X-Y) any amino acid

List of original articles

This thesis is based on the following articles, which are referred to by their Roman numerals in the text.

I Väisänen M-R, Väisänen T, Pihlajaniemi T (2004) The shed ectodomain of type XIII collagen affects cell behaviour in a matrix-dependent manner. Biochem J 380: 685-693

II Väisänen M-R, Väisänen T, Tu H, Pirilä P, Sormunen R, Pihlajaniemi T (2005) The shed ectodomain of type XIII collagen associates with the fibrillar fibronectin matrix and may interfere with its assembly in vitro. Biochem J, in press

III Väisänen T*, Väisänen M-R*, Autio-Harmainen H, Pihlajaniemi T (2005) Type XIII collagen expression is induced during malignant transformation in various epithelial and mesenchymal tumours. J Pathol 207: 324-335 (*equal contribution)

Contents

Abstract Acknowledgements Abbreviations List of original articles Contents 1 Introduction ...................................................................................................................17 2 Review of literature .......................................................................................................18

2.1 Superfamily of collagens ........................................................................................18 2.1.1 Definition of collagens ....................................................................................18 2.1.2 Classification of collagens...............................................................................19 2.1.3 Biosynthesis of collagens ................................................................................20 2.1.4 Functions of collagens .....................................................................................21

2.2 Transmembrane collagens ......................................................................................22 2.2.1 Type XIII collagen...........................................................................................23 2.2.2 Other transmembrane collagens ......................................................................27

2.2.2.1 Type XVII collagen ..................................................................................27 2.2.2.2 Type XXIII collagen .................................................................................28 2.2.2.3 Type XXV collagen ..................................................................................28

2.3 Ectodomain shedding of proteins ...........................................................................29 2.3.1 Proteinases.......................................................................................................29

2.3.1.1 The proprotein convertases.......................................................................30 2.3.2 Ectodomain shedding of type XIII collagen ....................................................31 2.3.3 Ectodomain shedding of other transmembrane collagens................................32

2.3.3.1 Type XVII collagen ..................................................................................32 2.3.3.2 Type XXIII collagen .................................................................................33 2.3.3.3 Type XXV collagen ..................................................................................33

2.4 Fibronectin extracellular matrix .............................................................................34 2.4.1 Fibronectin.......................................................................................................34 2.4.2 Fibronectin matrix assembly ...........................................................................35 2.4.3 Collagens and the fibronectin matrix assembly ...............................................36

2.5 The tissue microenvironment concept of cancer.....................................................38 2.5.1 Tumour stroma.................................................................................................38

3 Outline of the present study...........................................................................................41 4 Materials and methods...................................................................................................42

4.1 Construction of the expression plasmids (I, II, III).................................................42 4.2 Production of type XIII collagen antibodies (II, III)...............................................43 4.3 Western blotting (I, II, III) ......................................................................................43 4.4 Image acquisition (I, II, III) ....................................................................................43 4.5 Cell culture (I, II, III)..............................................................................................44

4.5.1 Culturing of cells (I, II, III)..............................................................................44 4.5.2 Transfection (I, II, IIII) ....................................................................................44 4.5.3 Coating (I, III) .................................................................................................44 4.5.4 Cell adhesion assays (I, III) .............................................................................45 4.5.5 Cell spreading assay (III).................................................................................45 4.5.6 Cell migration assays (I)..................................................................................45 4.5.7 Cell proliferation assays (I, III) .......................................................................46

4.6 Analysis of the type XIII collagen ectodomain shedding (I) ..................................46 4.6.1 Whole-cell shedding assay (I, II, III) ...............................................................46 4.6.2 Cell-surface biotinylation (I) ...........................................................................47 4.6.3 Streptavidin pull-down (I) ...............................................................................47

4.7 Analysis of the fibronectin matrix assembly (II) ....................................................47 4.7.1 Ectodomain association with the fibronectin matrix (II) .................................47 4.7.2 Immunofluorescence staining (II)....................................................................48 4.7.3 Transmission electron microscopy (II) ............................................................48 4.7.4 Determination of the binding sites (II) ............................................................49

4.7.4.1 Determination of the type XIII collagen binding site (II) .........................49 4.7.4.2 Determination of the fibronectin binding site (II).....................................49

4.7.5 Quantification of the fibronectin matrix formation (II) ...................................50 4.7.5.1 The immunofluorescence method (II) ......................................................50 4.7.5.2 The insolubility method (II)......................................................................50

4.8 Analysis of the type XIII collagen expression in malignant transformation (III)................................................................................................51

4.8.1 Construction of tissue microarrays (III)...........................................................51 4.8.2 In situ hybridization (III) .................................................................................51

4.8.2.1 Construction of the type XIII collagen probes (III) ..................................51 4.8.2.2 The in situ hybridization protocol (III) .....................................................51 4.8.2.3 Scoring of the in situ hybridization results (III)........................................52

4.8.3 Immunohistochemical staining (III) ................................................................52 4.8.4 Conditioned medium experiment (III).............................................................52 4.8.5 Northern blotting (III)......................................................................................53

5 Results ...........................................................................................................................54 5.1 Characterization of type XIII collagen ectodomain shedding

in mammalian cells (I)...........................................................................................54 5.1.1 Shedding of ectodomain by proprotein convertases (I) ...................................54 5.1.2 Intracellular regulation of ectodomain shedding (I) ........................................55 5.1.3 Time-dependence of ectodomain shedding (I).................................................55

5.1.4 Implications of Golgi involvement in ectodomain shedding (I) ......................56 5.2 Characterization of the biological activity of the cleaved ectodomain (I, II, III)....56

5.2.1 Cell adhesion (I, III) ........................................................................................56 5.2.2 Cell migration (I) .............................................................................................57 5.2.3 Ectodomain as a substratum (I) .......................................................................57 5.2.4 Cell proliferation (I).........................................................................................58 5.2.5 Cell spreading (III) ..........................................................................................58 5.2.6 Interference of the fibronectin matrix assembly (II)........................................58

5.3 Association of ectodomain with the fibrillar fibronectin (II) ..................................59 5.4 Characterization of the binding sites (II) ................................................................59

5.4.1 The type XIII collagen binding site to fibronectin (II) ....................................59 5.4.2 The fibronectin binding site to type XIII collagen (II) ....................................60

5.5 Type XIII collagen expression in cancer (III) .........................................................60 5.5.1 Expression of type XIII collagen mRNA in various cancers (III) ...................61 5.5.2 Expression of type XIII collagen in the invasive fronts (III) ...........................61

5.6 Induction of type XIII collagen expression (III).....................................................62 5.6.1 The conditioned medium experiment (III).......................................................62 5.6.2 Role of TGF-β1 on type XIII collagen expression (III)...................................62

6 Discussion .....................................................................................................................63 7 Future perspectives........................................................................................................68 References

1 Introduction

A single cell is the functional unit of a given tissue. Numerous cells make tissues, which organize into bigger functional entities, organs. However, tissues are not only composed of cells. Instead, they contain material that is synthesized and deposited by the cells into the space surrounding them. This material is called extracellular matrix (ECM). It consists of non-cellular material, mainly proteins, glycoproteins and proteoglycans. The ECM defines the structural integrity of tissues and organs in multicellular organisms. It plays a fundamental role in the development of cells and tissues, and its correct build-up, composition and function are critical requirements for the flawless functioning of cells, tissues and organs. Despite the diverse and specific composition in each tissue, a common feature of the ECM is the abundance of fibrillar proteins like collagens and fibronectin in it. In fact, these two substances make up the major components of the ECM. This thesis concentrates on type XIII collagen, which resides in the plasma membranes of cells. Its localization allows it the unique capability to link a cell and its surrounding pericellular milieu. Other features of type XIII collagen, namely the release of its membrane-bound extracellular domain, i.e. ectodomain, into the ECM allow this collagen an equally challenging possibility to influence the synthesis, composition and dynamics of the surrounding matrix. Therefore, in this thesis the shedding of the type XIII collagen ectodomain in mammalian cells, and the biological ramifications brought upon the ECM and cell behavior by the soluble ectodomain have been explored in more detail. Related to these themes, literature overviews on the collagen superfamily, and in particular, on the family of transmembrane collagens, are given. Also, ectodomain shedding as a phenomenon, with a specific focus on that of the transmembrane collagens, is reviewed. So far, little has been known about type XIII collagen expression in various human pathological conditions, which is why this thesis also describes how the expression of this collagen is changed during tissue-level alterations leading to cancer. Related to the topic of cancer, a more recent concept of the non-epithelial stromal compartment contributing to the initiation of tumours is also presented.

2 Review of literature

2.1 Superfamily of collagens

2.1.1 Definition of collagens

One of the most abundant components of the vertebrate connective tissue is collagens. Despite their ubiquitous presence in a number of tissues, like bones, tendons, cartilage, ligaments, cornea, kidney and skin, many of them show a distinct specificity in their localization. Collagens make a large and structurally diverse protein family. Currently, 27 collagen types (I-XXVII) have been reported in humans. (van der Rest & Garrone 1991, Myllyharju & Kivirikko 2001, Myllyharju & Kivirikko 2004.) However, there are indications as to the existence of an additional one, namely type XXVIII collagen (Koch et al. 2004). The defining characteristic structure of the collagens is the triple-helical collagenous domain, and by definition, all members of the collagen family form these supramolecular structures (Brodsky & Shah 1995, Gelse et al. 2003). However, all collagens also contain non-collagenous domains, either interspersed within the collagenous domain or in the precursor forms flanking the triple-helical portion (Prockop & Kivirikko 1995, Snellman & Pihlajaniemi 2003). The characteristic collagenous domain is formed from three α polypeptide chains in an extended left-hand helix, which supercoil around each other and a central axis in a right-handed manner to form the triple-helix. Relative to each other, the α chains can be identical (thus making a homotrimer) or different (heterotrimer). Another fundamental characteristic of collagens is the Gly-X-Y repeat making the collagenous domain. In this repeat proline and 4-hydroxyproline often preoccupy the X and Y positions, respectively. (Brodsky & Shah 1995, Exposito et al. 2002.) This amino acid composition meets the critical steric and stability prerequisites of the triple-helix formation (Myllyharju & Kivirikko 2001, Gelse et al. 2003).

19

2.1.2 Classification of collagens

Based on their supramolecular organization, collagens are categorized into subgroups (see for reviews van der Rest & Garrone 1991, Myllyharju & Kivirikko 2001, Gelse et al. 2003). One way to categorize collagens is to divide them into either fibrillar or non-fibrillar collagens. However, such a division may be over-simplified in terms of not fully appreciating the huge structural and functional versatility of all collagens, so usually collagens are subcategorized into more numerous classes, which could, in short, be characterized as follows: (i) collagens type I-III, V and XI arrange themselves as fibres, hence the group name fibrillar collagens. Type XXIV and XXVII collagens were recently described to be members of this subgroup (Boot-Handford et al. 2003, Koch et al. 2003, Pace et al. 2003). (ii) Characteristic to the FACIT, i.e. fibre-associated collagens with interrupted triple-helices, including collagens type IX, XII, XIV, XVI, XIX, XX, XXI and XXII, are interruptions of collagenous domains by non-helical domains and the fact that they share certain homologous domains, including the thrombospondin N-terminus-like (TSPN) domain, the von Willebrand factor-like A (VWFA) domain and the fibronectin repeat type III-like domain. In addition, some of them are located on the surfaces of fibrillar collagens. (Olsen 1997, Fitzgerald & Bateman 2001, Koch et al. 2001, Tuckwell 2002, Kassner et al. 2003, Myers et al. 2003, Koch et al. 2004.) Type XVI collagen is unique in its capability of forming different supramolecular structures like fibrils or beaded microfibrils, depending on the tissue localization (Kassner et al. 2003). Type XXVI collagen bears structural similarity to FACIT collagens. However, the lack of a thrombospondin N-terminus-like domain renders this collagen a FACIT-related collagen (Sato et al. 2002). (iii) Type VIII collagen, the major component of the Descemet´s membrane (Kvansakul et al. 2003), and type X collagen, only present in hypertrophic cartilage (Kielty et al. 1985), are short-chain non-fibrillar collagens forming hexagonal lattices. (iv) The six different α(IV) chains have been found to associate into three different heterotrimers, thus forming a family of the type IV collagens present as a network-like organization exclusively in basement membranes (Gelse et al. 2003). (v) Type VI collagen forms beaded filaments, and (vi) type VII collagen anchoring fibrils for basement membranes. (vii) The group of the transmembrane collagens consists of collagens type XIII, XVII, XXIII and XXV (Pihlajaniemi et al. 1987, Hopkinson et al. 1992, Hashimoto et al. 2002, Banyard et al. 2003). (viii) Due to the numerous interruptions in the collagenous domains, collagens type XV and XVIII make their own distinct group of so-called multiplexins, i.e. multiple triple-helical domains and interruptions, the latter giving rise to the strongly antiangiogenic fragment endostatin (Myers et al. 1992, Kivirikko et al. 1994, Rehn & Pihlajaniemi 1994, Rehn et al. 1994). In addition to the numerous different collagen types, more variability in this protein superfamily results from the occasional capability of collagens to form hybrids consisting of various collagen types. This is exemplified by collagens type I, III, V and XII, collagens type II, IX and XI, and collagens type V and XI. (Myllyharju & Kivirikko 2004.) Furthermore, alternative splicing of the primary transcripts of several collagen polypeptide α chains, as best described for type XIII collagen (Juvonen & Pihlajaniemi 1992, Juvonen et al. 1992, Peltonen et al. 1997), and usage of alternative promoters, like with type XVIII collagen (Rehn & Pihlajaniemi 1995), can result in the creation of an

20

extensive array of variants. In addition, this heterogenous protein superfamily also includes a subgroup of collagenous transmembrane proteins that do contain collagen-like triple-helical domains, but which have not been recognized as true collagens due to the lack of involvement in the ECM structure (Exposito et al. 2002, Myllyharju & Kivirikko 2004). These proteins include, for instance, the subcomponent C1q of complement, the tail structure of acetylcholinesterase, the pulmonary surfactants SP-A and SP-D, mannan-binding protein, collectin-43, conglutinin, ficolins, type I and II macrophage scavenger receptors, a macrophage receptor MARCO, an adipose-specific collagen-like receptor apM1, a src-homologous-and-collagen (SHC) protein, aggretin and ectodysplasin A (see for reviews Myllyharju & Kivirikko 2001, Snellman & Pihlajaniemi 2003, Myllyharju & Kivirikko 2004). The variability in their polymeric structure and organization allows the collagens to conform to the many functional and structural requirements of various tissues in which they are found. Such a pliant adaptability in structural organization allows this protein superfamily to be involved in a plethora of tissue and organ functions.

2.1.3 Biosynthesis of collagens

The biosynthesis of fibrillar collagens has often been used as a paradigm for collagen biosynthesis. In short, the intracellular events following the translation of the procollagen molecules with N- and C-terminal propeptide extensions include cleavage of the signal peptide, hydroxylation of certain proline and lysine residues and glycosylation of some hydroxylysine and asparagine residues. All this is accomplished by a set of enzymes including prolyl hydroxylases, a lysyl hydroxylase, collagen galactosyl transferases, collagen galactosylglucosyl transferases and an oxidase. More recent data also have highlighted the necessary involvement of the peptidyl proline cis-trans isomerase, the protein disulfide isomerase and a specific chaperone Hsp47 for the successful collagen biosynthesis and stabilization of the formed procollagen. After these post-translational modifications, nucleation at the C-terminus of the α chains initiates the directional folding of the collagen triple-helix towards the N-terminus in a zipper-like manner. Nascent triple-helical procollagen is secreted from the cell into transport vesicles, followed by excision of the N- and C-propeptides by N- and C-proteinases, respectively. This triggers the spontaneous self-assembly of the collagen fibrils to thicker fibres, and the formation of the stabilizing cross-links. (Prockop & Kivirikko 1995, Gelse et al. 2003.) However, novel data concerning the procollagen processing and collagen fibrillogenesis in the embryonic tendon have revealed that the conversion of procollagens to collagens and fibril formation can, in fact, also occur intracellularly, thus providing a totally new insight to this concept. (Canty et al. 2004, Canty & Kadler 2005.) Transmembrane collagens are likely to make an exception to the general principle concerning the direction of the trimerization of the α chains, as their N-terminal ends are bound to the plasma membrane. It is hence feasible that the triple-helix formation occurs in an opposite N-to-C direction. In fact, this has been shown for collagens type XIII and XVII (Snellman et al. 2000a, Areida et al. 2001). It has been proposed that proteins with collagenous domains can utilize various structural features and mechanisms to ensure their α-chain trimerization. The coiled-coil domains can act as one mechanism, but in

21

their absence, the chain association can be accomplished by long non-triple helical domains, as with type IV collagen, or by tightly packed β-sandwhich assembly domains, as with the members of the C1q family. (Kammerer 1997, McAlinden et al. 2003.)

2.1.4 Functions of collagens

Collagens were long considered to be important merely in the structural functions of tissues. The predominant tissue localizations, supramolecular structures, other specific characteristics of collagens, like their mechanical strength, and obviously, the characteristics of a number of diseases caused by mutations in collagen genes, have fostered such thinking. For instance, the resilient type I collagen is found in bones, and as osteogenesis imperfecta, resulting from mutations in type I collagen, leads to bone brittleness, the main function of type I collagen has been considered to be a structural component of the bone. Or, since type IV collagen makes networks in the basement membranes of the kidney, and as its mutations lead to haematuria and a severe kidney malfunction, the function of this collagen has been thought to be linked to the making of efficient filtering structures in the basement membranes of the kidney. (Gelse et al. 2003, Myllyharju & Kivirikko 2004.) However, in recent years it has become obvious that collagens, in particular their proteolytic non-collagenous domains, can also serve in non-structural, delicate regulatory functions. Such activity results from the release of smaller proteolytic fragments that possess cryptic biological activity that their intact ancestors do not. (Ortega & Werb 2002.) For example, the proteolytically released non-collagenous fragments of collagens type IV, XV and XVIII have been found to influence angiogenesis, tumorigenesis and cell proliferation in an inhibitory manner, at least in vitro. These fragments include the NC1 domains of α1(IV), α2(IV) and α3(IV) called arresten, canstatin and tumstatin, respectively, as well as the NC10 domain of α1(XV) called restin, and the C-terminal NC1 domain of α1(XVIII) called endostatin. They have all exhibited such beneficial anti-cancer effects that in cancer research they have subsequently been targets for high hopes of in vivo repeatability and substantial effort has been devoted to the medical exploitation of these qualities in cancer therapies. (O´Reilly et al. 1997, Ramchandran et al. 1999, Colorado et al. 2000, Kamphaus et al. 2000, Maeshima et al. 2000.) Not only small fragments of the whole collagen molecule, but the whole extracellular domains have been shown to be released from the plasma membrane of the integral membrane-bound collagens (Schäcke et al. 1998, Peltonen et al. 1999, Hashimoto et al. 2002, Banyard et al. 2003). As an example, the ectodomain of type XVII collagen is known to be cleaved, resulting in the formation of a soluble variant with biological activity (Franzke et al. 2002). For some collagens, additional bipolar functional variability is created by the proteoglycan nature of the molecule: for instance, whereas the C-terminal proteolytic fragment of type XVIII collagen, endostatin, acts as an anti-angiogenic factor, the heparan sulphate chains provide this collagen with pro-angiogenic, i.e. opposing qualities (Iozzo 2005). All these newly discovered characteristics of the collagens make them more fundamental and versatile constituents of the ECM: not only do they serve to establish structural and functional integrity to the tissues, but they can also perform subtle regulatory tasks (Ortega & Werb 2002).

22

2.2 Transmembrane collagens

The designation of a transmembrane protein relates to a protein which contains at least one short amino acid stretch of hydrophobic nature allowing the protein to be integrated into the lipid bilayer plasma membrane of a cell. The collagen superfamily includes such membrane-bound proteins forming a subgroup of their own, namely the collagenous transmembrane proteins. (Franzke et al. 2003, Snellman & Pihlajaniemi 2003, Franzke et al. 2005.) The number of members in this group of collagens has grown in recent years, and is presently comprised of the structurally similar collagens type XIII, XXIII and XXV, which could be considered as a subfamily of their own, as well as type XVII collagen and proteins with collagenous sequences, for example ectodysplasin A, the class A macrophage scavenger receptors, the MARCO receptor, the scavenger receptor with C-type lectin domain and the C1q subcomponent of the complement (Figure 1) (Myllyharju & Kivirikko 2001, Snellman & Pihlajaniemi 2003, Franzke et al. 2005). Several features are common to transmembrane collagens. In short, the collagenous stretches are α-homotrimers of a unique collagen α chain type. For a transmembrane protein, they possess a rare type II orientation on the plasma membrane (Hooper et al. 1997). Thus, the supramolecular structure of these proteins contains an N-terminal intracellular domain (endodomain), a single membrane-spanning stretch consisting usually of 21-26 amino acids (transmembrane domain), and a long extracellular C-terminus (ectodomain), made of one or more triple-helical collagenous domain(s), which are interspersed and flanked by non-collagenous domains. With the exception of type XVII collagen, the intracellular domain is usually only 10% of the length of the ectodomain (Franzke et al. 2003). The ectodomains protrude as rigid rod-like structures into the ECM. However, the interrupting non-collagenous domains separating the collagenous domains allow structural flexibility within the ectodomains (Tu et al. 2002). Juxtamembranous regions between the plasma membrane and the first extracellular collagenous domain make “linker regions” containing an α-helical coiled-coil domain (Snellman & Pihlajaniemi 2003, Franzke et al. 2005). These domains are widely occurring protein oligomerization domains thought to initiate chain trimerization and the following folding of the adjacent triple-helices. Coiled-coil domains consist of up to four amino acid heptad repeats (a-b-c-d-e-f-g) in which hydrophobic amino acids are located in positions a and d. (Kammerer 1997, McAlinden et al. 2003.) The direction of the α-helix trimerization of transmembrane collagens differs from that of fibrillar collagens. Namely, as opposed to the C-to-N direction of the fibrillar collagens, the anchorage of the N-terminal end of the ectodomain in the plasma membrane dictates the direction of trimerization to be N-to-C, as shown for collagens type XIII and XVII. (Snellman et al. 2000a, Areida et al. 2001.) An additional feature of these collagens is their ectodomain shedding, i.e. enzymatic cleavage of the extracellular domain to become a soluble molecule in the surrounding environment. This has been described for all the transmembrane collagens as well as for ectodysplasin A. The transmembrane collagens contain a furin consensus sequence within the juxtamembranous linker region, thus rendering furin susceptible to participation in the shedding process. (Franzke et al. 2003, Snellman & Pihlajaniemi 2003, Franzke et al. 2005.)

23

Fig. 1. A schematic representation of the polypeptide structures of some collagenous transmembrane proteins. For the transmembrane collagens, the collagenous domains (COL) are indicated with a number, the grey boxes inbetween and flanking the collagenous domains represent the non-collagenous (NC) domains, and the black segments stand for the transmembrane domains. NH2, aminoterminus; COOH, carboxyterminus; MARCO, macrophage scavenger receptor; SRLC, scavenger receptor with C-type lectin; C1q, the complement subcomponent; aa, amino acid.

2.2.1 Type XIII collagen

Type XIII collagen was originally characterized partially from a human cDNA library while screening it with a mouse α2(IV) cDNA probe (Pihlajaniemi et al. 1987). Later work completed the initial findings so that the supramolecular structure of this collagen is now known to consist of three collagenous triple-helical domains (COL1-COL3), which

24

are interspersed and flanked by four non-collagenous domains (NC1-NC4). A single short stretch of hydrophobic amino acids inserts this protein into the plasma membrane in a type II orientation, so that the N-terminus faces the cytosol and the C-terminus is extended to the extracellular side of the plasma membrane as a long rod (Figure 2). (Pihlajaniemi & Tamminen, 1990, Hägg et al. 1998, Tu et al. 2002.) The non-collagenous domains, which can act as hinges, create structural flexibility within this long rod. In fact, the bending of the ectodomain at the non-collagenous domains has been shown by rotatory shadowing (Tu et al. 2002). Characterization of this protein from human HT-1080 cell extracts has determined its MW to be between 85-95 kDa upon reduction and over 180 kDa under unreduced conditions (Hägg et al. 1998). The three collagenous domains COL1-COL3 make triple-helical structures with thermal stabilities of 38, 49 and 40oC, respectively (Snellman et al. 2000b). There are altogether eight cysteine residues, with six of them localizing and likely to form interchain disulphide bonds in the NC1 domain and at the junction of the COL1 and NC2 domains. The C-terminal NC4 domain contains two cysteine residues, presumably forming intrachain bonds (Snellman et al. 2000b). As is typical for type II transmembrane proteins, type XIII collagen lacks a signal sequence (Singer 1990, Hägg et al. 1998).

Fig. 2. A schematic representation for type XIII collagen. The collagenous domains are indicated with the abbreviation COL, the non-collagenous domains with NC. C stands for cysteine residues, RRRR for the proprotein convertase consensus cleavage site consisting of four arginine (R) amino acid residues. The black segment indicates the hydrophobic transmembrane domain.

Both the human and mouse type XIII collagen genes for the α1 chains have been assigned to chromosome 10 (Pajunen et al. 1989, Shows et al. 1989, Kvist et al. 1999). One striking feature of type XIII collagen is the complex alternative splicing that its primary transcripts are subjected to (Pihlajaniemi et al. 1987, Tikka et al. 1988, Tikka et al. 1991). This phenomenon involves both the collagenous and the non-collagenous sequences, mainly the COL1, NC2 and COL3 domains, whereas the central COL2 and NC3 domains are left virtually intact (Pihlajaniemi & Tamminen 1990, Juvonen & Pihlajaniemi 1992, Juvonen et al. 1992, Peltonen et al. 1997). In humans, the alternative splicing affects exons 3B, 4A, 4B and 5 of the COL1 domain, exons 12 and 13 of the NC2 domain, and exons 29, 33 and 37 of the COL3 domain (Juvonen & Pihlajaniemi 1992, Juvonen et al. 1992), whereas in mice, exons 4A, 4B, 5, 6 and 8 of the COL1 domain, exons 12 and 13 of the NC2 domain, and exons 28B, 32 and 33 of the COL3 domain are affected (Peltonen et al. 1997). Thus, both species have several alternative exons, although only some of them are the same. It follows that, due to alternative splicing, at least at the

25

mRNA level numerous different variants may exist. However, apparently there are only a few main variants for each variable domain, and there are marked tissue differences in the occurrence of some splice variants. The lengths of the different α1(XIII) collagen chains are also predicted to vary as for the whole polypeptides as well as for the individual domains. (Juvonen & Pihlajaniemi 1992, Juvonen et al. 1992, Peltonen et al. 1997.) Several aspects concerning the alternative splicing are not yet fully understood: it is not known whether all of the type XIII collagen mRNA variants are translated to proteins and how the polypeptide chains of different lengths potentially associate into trimeric molecules. Also the biological significance of the phenomenon is yet to be elucidated. It has been speculated, though, that the alternative splicing could greatly modify the potential capability of the collagenous domains for lateral interactions (Peltonen et al. 1997).

The expression of type XIII collagen has been studied using various methods including RT-PCR, Northern and in situ hybridizations. It is ubiquitously expressed in vertebrate tissues, including bone, cartilage, small intestine, skin, striated muscle, epidermis, hair follicles, liver, spleen, kidney, lung, heart, Achilles tendon, aorta and brain (Sandberg et al. 1989, Peltonen et al. 1997). The expression is pronounced in tissues of mesodermal origin, but also tissues of epithelial and neuroectodermal origins express it (Sund et al. 2001a). Human placental and ocular expressions have been characterized with scrutiny. In the placenta, type XIII collagen is expressed in the fibroblastoid stromal cells of villi, the endothelial cells of developing capillaries, the cytotrophoblastic columnar cells, the decidual cells and the stromal cells of gestational endometrium (Juvonen et al. 1993). In the fetal and adult eye, type XIII collagen can be found in the optic nerve bundle, the ganglion layer of the retina, the developing ciliary smooth muscle, the posterior corneal stroma and the striated extraocular muscles (Sandberg-Lall et al. 2000). Type XIII collagen has a unique temporo-spatial expression profile with a strong foetal expression, as opposed to a distinctly lower expression level in mature adult tissues. Sund and co-workers have described a strong expression in the developing mouse heart and the central and peripheral nervous system. However, the expression of this collagen shows a developmental shift with expression attenuating during the maturation of the tissues. Contrary to other tissues, no major differences between foetal vs. adult expression levels have been recognized in the eye. (Sandberg-Lall et al. 2000, Sund et al. 2001a.) During development, the early diffuse manifestation of this collagen is also eventually concentrated in distinct punctuated structures of mature tissues (Sund et al. 2001a). Based on immunofluorescence staining of cells and tissues, these structures have been shown to be various cell-cell contacts and cell-matrix adhesion structures. These include focal adhesions, myotendinous junctions, costameres of skeletal muscle and intercalated discs of the heart. In adherens junctions type XIII collagen co-localized with E-cadherin. (Peltonen et al. 1999, Hägg et al. 2001.) Type XIII collagen comprises an integral component of focal adhesions, as it is present in these structures when cells are adhering, spreading or moving. It also follows the turn-over of focal adhesions, as it disappears from these structures concomitantly with other components when the focal adhesions are disassembled. (Hägg et al. 2001.) Very little colocalization was found with desmosomal components, so it is unlikely that type XIII collagen is a component of desmosomes in skin (Peltonen et al. 1999).

26

The usual C-to-N direction of trimerization of fibrillar collagens (Prockop & Kivirikko 1995, Gelse et al. 2003) is not feasible for type XIII collagen, as its N-terminus is buried and bound in the plasma membrane. In fact, it has been shown that for type XIII collagen, that the trimerization occurs in an opposite direction (Snellman et al. 2000a). By using a deletion variant lacking the juxtamembranous residues, Snellman and co-workers showed that the extracellular amino acid residues 63-83 immediately adjacent to the transmembrane domain are pivotal for the nucleation of the disulphide-bonded trimers (Snellman et al. 2000a). It was later characterized in more detail that this 21-amino acid residue is necessary, but not sufficient for the NC1 domain association. In addition to this short amino acid stretch, the NC1 domain and probably other elements in the membrane-spanning domain form a α-helical coiled-coil structure, which acts as a conserved nucleation nidus for the trimerization of α chains. (Latvanlehto et al. 2003.) These domains have been characterized in other transmembrane proteins, too, in particular, in the membrane-spanning collagens type XVII and XXV (Latvanlehto et al. 2003). Additionally, another coiled-coil structure exists in the central NC3 domain, implied to act as a trimerization center for the C-terminal collagenous domains. Absence of the NC1 coiled-coil domain was shown to lead to a lack of disulphide-bonded trimers and misfolding of the COL1 domain, whereas its absence had no influence on the COL2 and COL3 domains. Latvanlehto and co-workers thus proposed that these two domains act independently, with the NC1 coiled-coil structure responsible for the trimerization of the N-terminal end of α-chains and the NC3 equivalent for the C-terminal end. (Latvanlehto et al. 2003.)

The function of type XIII collagen is considered to be related to cell adhesion. The expression of this collagen during the development in the adhesive structures of cells and tissues (Sund et al. 2001a, Sund et al. 2001b), the enhanced Sf9 insect cell adhesion to certain matrices resulting from transient expression of this collagen (Hägg et al. 2001), but most strongly, its undisputed localization in the adhesive structures of cells and tissues (Peltonen et al. 1999, Hägg et al. 2001) have all suggested that the function of type XIII collagen is linked to cell adherence. Also, the characteristic temporo-spatial expression during development, i.e. ubiquitous and abundant in the early developmental phases and the restricted low-level expression in the adult tissues, has implied an important role in the development of correct tissue architecture and the integrity of tissues (Sund et al. 2001a). Subsequent studies have supported these hypotheses. Namely, Sund and co-workers found that the expression of a truncated variant of type XIII collagen, lacking a 90-residue COL2 sequence as well as the extreme C-terminal 36 residues, led to lethal phenotypes in mice, either due to a lack of fusion of placental membranes (the early phenotype) or developmental defects in the adhesive structures of the heart (the late phenotype) (Sund et al. 2001b). Also, as Kvist and co-workers expressed another genetically manipulated type XIII collagen variant with the intracellular and transmembrane domains deleted, they detected abnormal myofibres and myofilaments as well as disorganized plasma membrane-basement membrane connections at the myotendinous junctions, suggesting participation of this collagen in the linkage between muscle fibres and basement membrane. (Kvist et al. 2001, Sund et al. 2001b.) A common denominator in these two mouse phenotypes is the impaired structure and function of various adhesive structures, resulting from the compromised structures of type XIII collagen. Such ramifications could be interpreted to imply a role

27

for type XIII collagen in important developmental adhesive interactions (Sund et al. 2001b). The importance of the correct expression of type XIII collagen during development, here in the bone modeling, was highlighted in a recent study in which overexpression of this collagen in skeletal tissues led to an abnormally high postnatal bone mass with an increased bone formation (Ylönen et al. 2005).

The cleavage of the ectodomain has been shown to take place with type XIII collagen, since the soluble ectodomain has been recovered from the culture media of human keratinocytes (Peltonen et al. 1999) and virus-infected insect cells (Snellman et al. 2000a). Furin of the proprotein convertases has been suggested to be responsible for this, as a sequence consisting of four arginine residues fitting to the recognized consensus domain for the proprotein convertases (Thomas 2002) has been found at the stem of the ectodomain (Snellman et al. 2000a). However, the biological implications or the regulation of the ectodomain shedding have still remained uncharacterized. Tu and co-workers purified the recombinant ectodomain from insect cell culture medium, and showed its in vitro binding to some ECM or basement membrane components in a solid phase assay. These interactive proteins included fibronectin, nidogen-2, perlecan, heparin and vitronectin. (Tu et al. 2002.) Whether these interactions occur in cell culture conditions or in living organisms have not so far been resolved.

2.2.2 Other transmembrane collagens

2.2.2.1 Type XVII collagen

Type XVII collagen, also designated as 180 kDa bullous pemphigoid antigen (BP180) or bullous pemphigoid antigen 2 (BPAG2), was initially characterized as an autoantigen in the blistering skin diseases bullous pemphigoid, herpes gestationis, cicatrical pemphigoid, linear IgA disease and lichen planus pemphigoides (Diaz et al. 1990, Hopkinson et al. 1992, Li et al. 1993, Zillikens & Giudice 1999). Structurally, this collagen is a homotrimer of three 180 kDa α1(XVII) chains. It is a type II transmembrane protein residing on the plasma membrane with its globular N-terminus of 466 amino acids oriented intracellularly, and its C-terminal 1008-amino acid-long ectodomain facing the extracellular milieu. The long rod-like ectodomain contains 15 collagenous subdomains (COL1-COL15), interspersed and flanked by 16 non-collagenous domains (NC1-NC16), with some of them N-glycosylated. The hydrophobic stretch of 23 amino acids facilitates the insertion of this protein into the plasma membrane as an integral protein. The NC16A domain adjacent to the plasma membrane contains coiled-coil structures, which serve as an initial nidus for the chain association and the subsequent triple-helix folding in the N-to-C-terminal direction. (Hopkinson et al. 1992, Li et al. 1993, Areida et al. 2001, Franzke et al. 2003, Franzke et al. 2005.) Type XVII collagen is a structural component of the hemidesmosomes in the skin. Its cytosolic N-terminus resides in the hemidesmosomal plaque, whereas the extracellular portion extends to the region of the anchoring filaments in the basement membrane zone. (Zillikens & Giudice 1999, Franzke et al. 2005.) These adhesive structures mediate the adhesion of basal epidermal

28

keratinocytes to the underlying basement membranes (Diaz et al. 1990, Hopkinson et al. 1992). Presence in such organelles has suggested that type XVII collagen is involved in the epithelial cell adhesion, which has been substantiated by the clinical findings in the genetic and acquired skin diseases in which the lack or loss of function of type XVII collagen results in reduced epidermal-dermal cohesion and skin blistering as a consequence of minor shear forces experienced by the skin (Zillikens & Giudice 1999). Type XVII collagen was the first transmembrane collagen for which the constitutive shedding of the ectodomain was reported. Cleavage of the ectodomain at the cell surface results in a 120 kDa soluble ectodomain, so type XVII collagen exists as two variants; namely, a membrane-bound and a soluble variant. (Hirako et al. 1998, Schäcke et al. 1998, Schumann et al. 2000.)

2.2.2.2 Type XXIII collagen

A more recently characterized member of the transmembrane collagen family is type XXIII collagen. Structurally this collagen shows a similarity to collagens type XIII and XXV with a short intracellular domain, a short single hydrophobic transmembrane domain and a long rod-like extracellular domain with three collagenous domains (COL1-COL3) interspersed and flanked by four non-collagenous domains (NC1-NC4). Its orientation on the plasma membrane is that of a type II transmembrane protein. At the protein level, the similarity between type XXIII collagen vs. type XIII and XXV collagens is 54% and 56%, respectively, primarily in the collagenous domains. However, the C-terminal 20-amino acid-long NC4 domain seems to be particularly well conserved between these three transmembrane collagens. Type XXIII collagen is expressed in heart, retina, brain and lung, and in an up-regulated manner in metastatic prostate carcinoma cells, whereas not in non-metastatic or poorly metastatic cells. This collagen is expressed at the cell surface, but is known to be cleaved by furin proprotease. The ectodomain contains a sequence fitting to the consensus cleavage site of furin. The biological properties of the cleaved ectodomain are presently unknown, but it has been demonstrated to bind to heparin with low affinity. Also, currently the biological function of this transmembrane collagen is unknown. (Banyard et al. 2003.)

2.2.2.3 Type XXV collagen

Type XXV collagen is a type II oriented transmembrane protein, which was originally found in the senile plaques in the brains of patients with Alzheimer´s disease. It also bears considerable structural resemblance to type XIII collagen by having a short cytosolic non-collagenous intracellular domain, a single transmembrane domain and an extracellular domain with three collagenous (COL1-COL3) and four non-collagenous (NC1-NC4) domains. At the amino acid level the sequence similarity between these two transmembrane proteins is 43%. Type XXV collagen is also called CLAC-P, i.e. collagen-like Alzheimer amyloid plaque component precursor. The soluble form CLAC is generated by furin-mediated cleavage of the ectodomain from CLAC-P, and it as been

29

shown to exist in a glycosylated triple-helical form. CLAC binds heparin with an intermediate affinity. (Hashimoto et al. 2002, Söderberg et al. 2005a.) Type XXV collagen/CLAC shows restricted tissue localization by being expressed solely in cerebral neurons (Hashimoto et al. 2002). CLAC has been detected to be associated with the fibrillar Aβ of the amyloid plaques, but the biological function and the contribution of CLAC to the pathogenesis of Alzheimer´s disease and its plaque formation are still largely unknown. However, it seems that CLAC can assemble the amyloid fibrils into protease resistant aggregates (Söderberg et al. 2005b), and that CLAC can bind to amyloid β peptides via positively charged amino acid clusters within the COL1 domain. This binding is mediated by electrostatic interactions and also requires the presence of the triple-helical structure of CLAC (Osada et al. 2005). Other biological functions have not yet been described for type XXV collagen.

2.3 Ectodomain shedding of proteins

Cell surface can be considered as a dynamic platform, the repertoire of expressed proteins of which is constantly and continuously renewed (Werb & Yan 1998). Ectodomain shedding refers to a phenomenon in which the extracellular domains of membrane-bound proteins are enzymatically released from the cell surface into the surrounding pericellular milieu or to bodily fluids (Hooper et al. 1997, Kheradmand & Werb 2002). This provides a given cell with a means of disposing of a spent protein. Alternatively, controlled release may be interpreted as a regulatory mechanism for down-regulating and/or optimizing the plasma membrane content of a given protein. Once soluble, ectodomains may possess biological properties that are greatly altered relative to their membrane-bound ancestors. Cells can thus generate soluble molecules with novel bioactivity, capable of refashioning the neighbouring microenvironment and/or modifying the build-up of the matrix to meet the requirements of development and/or homeostasis. A dual existence of a protein as a membrane-bound and/or a soluble variant also allows the cell to multiply the functional roles of its membrane-bound proteins. Ectodomain cleavage is widely known to occur with integral membrane proteins, with the diversity and scope of these proteins encompassing practically all functional classes of membrane proteins. (Ehlers & Riordan 1991, Werb 1997, Peschon et al. 1998, Werb & Yan 1998, Blobel 2000, Mott & Werb 2004.)

2.3.1 Proteinases

Various proteinases can be involved in the release of the ectodomains of integral membrane proteins. These include the zinc-dependent metzincin superfamily of metalloproteinases comprised of the transmembrane MT-MMPs (membrane-type matrix metalloproteinase) and the soluble MMPs (matrix metalloproteinase), ADAMs (a disintegrin and metalloproteinase) and ADAM-TS (an ADAM with a thrombospondin-like motif) families. Some serine proteinases, like the urokinase plasminogen activator, the tissue plasminogen activator, plasmin and thrombin have also been implicated in the cleavage. (Hooper et al. 1997, Werb 1997, Merlos-Suárez & Arribas 1998, Werb & Yan

30

1998, Lee & Murphy 2004, Mott & Werb 2004.) Based on sequence and experimental data, furin of the proprotein convertase family has been implicated strongly in the ectodomain shedding of transmembrane collagens, either by direct contribution or via indirect activation of MMPs. Furin is known to serve in an obligate activating role for the MMPs, which are synthesized as inactive precursor proteins with a prodomain. In order to gain their biological activity, the removal of the prodomain is necessary. Many MMPs contain a furin consensus cleavage sequence between the prodomain and the catalytic domain, and the furin-dependent activation of MMP-11, MT1-MMP (MMP-14) and MT3-MMP (MMP-16) has been verified (Pei & Weiss 1995, Sato et al. 1996, Kang et al. 2002). Also, it is currently known that furin acts in such a pre-activating role in the ectodomain shedding of type XVII collagen (Franzke et al. 2002).

2.3.1.1 The proprotein convertases

Furin belongs to the mammalian subtilisin/Kex2p-like proprotein convertase (PC) family of calcium-dependent serine endoproteases (Nakayama 1997). This protease family is comprised of seven distinct members, namely PC1/PC3, PC2, furin, PC4, PACE4, PC5/PC6 in two isoforms 6A or 6B and PC7/SPC7/LPC, with furin firstly and most extensively characterized. Some members of this family, like furin, PC6A/PC6B or PC7, show a ubiquitous tissue distribution, whereas others, like PC4, PC1 or PC2 possess a much more restricted tissue or developmental expression pattern. For instance, PC4 is exclusively expressed in the germ cells of the testis and ovaries, whereas PC1 and PC2 are expressed in the neuroendocrine cells. PC5 is widely expressed during early embryonic development, whereas its tissue expression is considerably more restricted during maturity. (Seidah & Chrétien 1997, Thomas 2002, Taylor et al. 2003.) As for the subcellular localizations, furin, PC7, PC6B and PACE4 are predominantly located in the trans-Golgi network, endocytic compartment and cell surface, whereas PC6A, PC1/3 and PC2 reside in the dense-core secretory granules. Furin is known to recycle to and from the plasma membrane via the endosomal pathway (Molloy et al. 1994). At the cell surface, furin is tethered to a subcortical actin-binding filamin. (Nakayama 1997, Seidah & Chrétien 1997, Shapiro et al. 1997, Thomas 2002.) It appears that PCs show considerable redundancy for their substrates (Taylor et al. 2003).

PCs are responsible for the proteolytic maturation of various proprotein substrates in the constitutive secretory pathway. PCs recognize and cleave (↓) the target proteins on the C-terminal side of a common consensus cleavage motif Arg-Xn-Arg/Lys-Arg↓, where X is any amino acid except cysteine. Arginine residues in positions P1 and P4 are essential, whereas Arg/Lys in P2 is not and only enhances processing efficiency. Hence the minimal cleavage sequence is Arg-X-X-Arg. Furin itself is synthesized as a proprotein with a prodomain that acts as an intramolecular chaperone for the folding. After exploiting its own cleavage-site rules, furin cuts the prodomain twice in consecutive reactions to yield an active enzyme. (Nakayama 1997, Seidah & Chrétien 1997, Shapiro et al. 1997, Thomas 2002.)

The correct and timely action of furin and other PCs is fundamental in embryogenesis and organ homeostasis. PC deficiencies have been studied on a cellular level by using two cell lines known for their lack of PC (furin) expression, namely a human colon

31

carcinoma LoVo cell line (Bennett et al. 2000, Endres et al. 2003) and RPE.40 cells derived from a Chinese hamster ovary K1 cell line through mutagenization (Sucic et al. 1999). Knockout mouse phenotypes of the PCs range from undetectable to early embryonal lethality (furin, PACE4), severe growth defects (PC1/3) and grave permutations in glucose metabolism (PC2, PC1/3). It is now clear that PCs not only play a crucial role in a variety of developmental and physiological processes, but they are involved in several pathological processes as well. Pathogenicity of certain bacterial toxins, e.g. the anthrax toxin (Klimpel et al. 1992), and enveloped pathogenic viruses, e.g. human influenza viruses (Zambon 2001) and immunodefieciency virus HIV (Moulard & Decroly 2000), are dependent on proteolytic cleavage. PCs have been found to be at least partially responsible for the activating cleavage processes. Also, in several cancers furin expression has been found to be upregulated and correlated with increased aggressiveness and malignant potential of tumour behavior. (Thomas 2002, Taylor et al. 2003.) Conversely, inhibition of PCs has been reported to result in loss of growth and tumorigenicity in human colon carcinoma cells or HT-1080 cell invasiveness (Maquoi et al. 1998, Khatib et al. 2001). Obviously, there has been considerable interest in creating specifically targeted PC inhibitors. Currently, peptidyl derivations of chloroalkylketones and protein-based α1-antitrypsin variants, called α1-antitrypsin Pittsburgh or α1-antitrypsin Portland, have been used as inhibitors. For reasons of toxicity, low stability or incomplete blockage these are currently used only in research, and the unwanted side-effects have precluded their clinical in vivo utilization (Garten et al. 1994, Nakayama 1997, Thomas 2002, Taylor et al. 2003).

2.3.2 Ectodomain shedding of type XIII collagen

Peltonen and co-workers were the first to report the recovery of type XIII collagen ectodomain from the culture medium of the human keratinocytes endogenously expressing this collagen (Peltonen et al. 1999). Later, the same was detected with virus-infected insect cells (Snellman et al. 2000a). The electrophoretic motilities of the two medium-derived ectodomains were found to be almost identical, suggesting that the ectodomains had been cleaved in a similar manner at the same site. As opposed to type XVII collagen, the N-terminus of the insect cell-derived ectodomain of type XIII collagen has been amenable to amino acid sequencing, in which the N-terminus was found to begin with a residue 109, i.e. Glu-Ala-Pro-Lys-Thr-Ser-Pro-Gly-Cys-Asn (EAPKTSPGCN), in line with having been cleaved at the proposed site. Additionally, as furin inhibitor decanoyl-Arg-Val-Lys-Arg-chloromethylketone (decRVKRcmk) (Garten et al. 1994) prohibited the cleavage, this supported the scissoring at the furin consensus site by a furin-type protease. (Snellman et al. 2000a.) Tu and co-workers showed that the shedding of type XIII collagen ectodomain could be inhibited by heparin. The mechanism of this has remained obscure, but it was proposed that since clustered arginines are essential for heparin binding, binding of heparin to the arginine-rich furin cleavage site of type XIII collagen might avert the release of ectodomain. (Tu et al. 2002.) Lack of the intracellular portion of NC1 domain has been found to enhance the release of ectodomain. This was seen with the insect cell set-up (Snellman et al. 2000a), but already earlier Kvist and co-workers had reported that deletion of both the intracellular and

32

transmembrane domains led to enhanced secretion of the mutant variant into the surrounding ECM of the skeletal muscle cells when compared to an intact type XIII collagen (Kvist et al. 2001). The mechanism in-detail responsible for this is not known. Also, the biological implications of the shed ectodomain of type XIII collagen have not been fully characterized so far.

2.3.3 Ectodomain shedding of other transmembrane collagens

2.3.3.1 Type XVII collagen

Due to cleavage at the cell surface, a 120 kDa soluble ectodomain of type XVII collagen is created (Hirako et al. 1998, Schäcke et al. 1998, Schumann et al. 2000). Since the membrane-proximal NC16A domain of type XVII collagen contains a consensus furin PC cleavage motif Arg-Ile-Arg-Arg (RIRR), the shedding was originally presumed to occur via furin activity (Schäcke et al. 1998, Schumann et al. 2000). Later it was shown, however, that the metalloproteases of the ADAM family, namely TACE (TNF-α converting enzyme), ADAM-9 and ADAM-10 are actively involved. Also, as deletion of the furin consensus sequence (504)Arg-Ile-Arg-Arg(507) from type XVII collagen had no effect on shedding nor did this deletion abolish the inhibitory effect of a chloromethylketone, the current conception holds that furin acts as an activator of the ADAMs rather than being the primary sheddase for type XVII collagen. Of the ADAMs, TACE has been considered the major and physiologically relevant sheddase. (Franzke et al. 2002, Franzke et al. 2004, Zimina et al. 2005.) Based on epitope mapping, the cleavage site has been thought to reside within the NC16A domain, but experimental verification of this is still lacking as the scissile bond has persistently remained elusive. This is probably due to secondary modifications, which are thought to render the novel N-terminus of the ectodomain unaccessible for amino acid sequencing. Later studies have also shown that certain structural features and accessibility of the cleavage site in type XVII collagen are important for the successful cleavage. The ADAMs need an unfolded stalk region surrounding the target bond, and subsequently, large deletions, in particular those of amino acids at the positions 528-547, leaving short juxtamembranous stalks, have inhibited the ectodomain shedding. It is likely that elimination of this amino acid stretch causes grave structural changes in the ADAM recognition site. The hypothesis that structural motifs, like the conformation of the NC16A domain and the steric availability of the cleavage site, are crucial for the effective ectodomain shedding have been supported by findings revealing that the N-termini of the soluble ectodomains derived from different human cell cultures have been different. This could be translated to the exact cleavage site actually varying under different biological circumstances. (Franzke et al. 2002, Franzke et al. 2004.) Most recently, the ectodomain shedding of type XVII collagen was found to be dependent on the membrane raft localization. Co-patching and co-transfection experiments revealed that the majority of type XVII collagen is, whereas TACE is not, incorporated in the cholesterol-enriched lipid rafts (Zimina et al. 2005) which are cholesterol- and sphingolipid-enriched membrane microdomains in the plasma membrane (Brown & London 1998, Brown & London 2000, Lai 2003). If the lipid phase compartmentalization of the plasma membrane is altered drastically by biochemical

33

manipulation of the membrane cholesterol level, the mutual accessibility of type XVII collagen and its cleaving enzyme TACE is dramatically enhanced so that ectodomain shedding is upregulated. Thus the membrane organization, in terms of lipid raft compartmentalization, is an important determinant in the efficiency of this shedding event. (Zimina et al. 2005). The exact implications of the ectodomain shedding of type XVII collagen are not fully known yet, but the biological context is predicted to involve keratinocyte adhesion, differentiation and migration (Franzke et al. 2002).

2.3.3.2 Type XXIII collagen

The ectodomain of type XXIII collagen is released by furin convertase activity at a site conforming to the furin consensus sequence. The rat and human XXIII collagens contain suitable basic motifs (94)Lys-Leu-Arg-Thr-Val-Arg(99) and (105)Lys-Ile-Arg-Thr-Ala-Arg(110), respectively. Also in this case the furin inhibitor prevented the ectodomain shedding. As a further confirmation, only the full-length variant, but not the cleaved ectodomain, was detected in LoVo cells, a human colon carcinoma cell line genetically devoid of furin activity. The type XXIII collagen ectodomain has also shown low affinity binding to heparin. (Banyard et al. 2003.)

2.3.3.3 Type XXV collagen

Type XXV collagen has been linked to the pathogenic process of Alzheimer´s disease, which is a progressive neurodegenerative disorder mainly affecting the elderly. The ectodomain of this transmembrane protein, CLAC, is liberated from the plasma membrane as a soluble secreted form (also named sCLAC/sColXXV) to become deposited on the pathognomonic extracellular β-amyloid plaques in the brains of patients affected with Alzheimer´s disease. Again here, the protease responsible for the release of the ectodomain appears to be furin. The NC1 domain of type XXV collagen holds a cluster of basic amino acids, namely (107)Lys-Ile-Arg-Ile-Ala-Arg(112), fulfilling the criterium of a consensus sequence of furin-type convertases. In addition, experimental data have supported the role for furin. RPE.40 cells, lacking furin activity, did not liberate sCLAC, whereas the restoration of furin activity with transfection resulted in a robust secretion of sCLAC. (Hashimoto et al. 2002.) The role for furin was later confirmed by amino acid sequencing of the N-terminus of sCLAC, which revealed the cleavage to occur between Arg and Glu within the sequence (107)Lys-Ile-Arg-Ile-Ala-Arg-Glu-Ala-Pro-Ser-Glu(117). sCLAC also binds to heparin. (Söderberg et al. 2005a.) Although detectable in the amyloid plaques, the implications of sCLAC may, in fact, be surprising. Namely, very recent studies have proposed an anti-amyloidogenic role for sCLAC in the pathogenesis of Alzheimer´s disease, as the purified sCLAC has been found to inhibit the elongation phase of the Aβ fibril formation in vitro (Osada et al. 2005).

34

2.4 Fibronectin extracellular matrix

2.4.1 Fibronectin

The ECM is rich in fibrillar proteins. Collagens make one major component of this fibre-like network, the other predominant constituent being fibronectin. Fibronectin exists in two forms. The soluble dimeric variant composed of two nearly identical subunits of MW of 250 kDa each can be found in the plasma, other bodily fluids and the conditioned media of cultured cells, whereas the multimeric, high molecular weight insoluble fibrillar variant is distributed within the ECM. The plasma variant is largely produced by hepatocytes in the liver. A wide variety of other cells also produce fibronectin. Once synthesized, the fibronectin transcript is subject to alternative splicing affecting the exons IIIA, IIIB and V (or IIICS), as well as to post-translational modifications, like glycosylation, phosphorylation, sulphation and formation of intrachain disulphide bonds. Due to alternative splicing, fibronectin can exist as different variants. The C-terminal disulphide bonds hold the two subunits together in an anti-parallel manner (making a so-called fibronectin protomer). (Johansson et al. 1997, Romberger 1997, Armstrong & Armstrong 2000, Pankov & Yamada 2002.)

The fibronectin monomer (Figure 3) is a mosaic molecule composed of three types of repeating homologous structural units, designated as type I, II and III repeats. There are 12 type I repeats, 2 type II repeats and 15-17 type III repeats. Each repeat unit is composed of multiple amino acids: the type I, II and III repeats are made of about 40, 60 and 90 residues, respectively. Several repeating units make functional domains each possessing distinct biological binding activities for cells and other ECM molecules. The domains are separated by hinge regions so that different conformations can be adopted. The soluble protomer is believed to be globular, whereas the polymerized variant is most likely is an extended stretched fibre. The functional domains that have been characterized in fibronectin show multiple binding activities. These include fibrin, thrombospondin and heparin (the N-terminal domain), collagen and gelatin (the collagen/gelatin binding domain), several integrins (many domains), heparin and chondroitin sulphate (the heparin binding domain) and fibrin (the C-terminal domain). Via these binding activities fibronectin orchestrates a series of cellular processes like cell adhesion, migration, proliferation, differentiation and gene expression in tissue remodelling, wound healing, embryonic development, angiogenesis, thrombosis, inflammation and tumorigenesis. The versatility of the cellular effects of fibronectin results from its interaction with other ECM molecules and cell surface receptors belonging to the integrin and non-integrin families. (Johansson et al. 1997, Romberger 1997, Armstrong & Armstrong 2000, Plow et al. 2000, Pankov & Yamada 2002, Mao & Schwarzbauer 2005.)

35

Fig. 3. A schematic representation of the polypeptide structure for fibronectin. The functional domains with the respective binding activities for the various ligands are indicated in the picture, as compiled from references mentioned in the text. NH2, aminoterminus; COOH, carboxyterminus; S stands for a disulphide bond.

2.4.2 Fibronectin matrix assembly

Fibronectin fibrillogenesis is a cell-mediated and integrin-dependent multistep process, in which cell-synthesized, compactly folded fibronectin dimers are polymerized at the cell surfaces into long fibres, which are eventually deposited into the ECM as a net-like meshwork. It is just this insoluble fibrillar variant that most of the biological activities of fibronectin have been ascribed to. As uncontrolled spontaneous conversion of the soluble fibronectin into polymerized fibrillar network, for instance whilst in circulation, would be highly detrimental for the host organism, the process needs to be tightly regulated. (Schwarzbauer & Sechler 1999, Pankov & Yamada 2002, Wierbicka-Patynowski & Schwarzbauer 2003, Mao & Schwarzbauer 2005.) The assembly is dependent on cell surface integrin receptors, of which α5β1 integrin is recognized as a primary receptor in mediating the fibronectin fibrillogenesis (Wu et al. 1993). Other integrins may possess a similar potential (Wu et al. 1996), although not just any given integrin promotes fibronectin fibrillization, as was demonstrated for α4β1 integrin (Wu et al. 1995). Binding of fibronectin to α5β1 integrin is mediated by a specific integrin-binding tripeptide Arg-Gly-Asp (RGD) in the type III10 repeat and the flanking pentapeptide Pro-His-Ser-Arg-Asn (PHSRN), the so-called synergy site, in the type III9 repeat of fibronectin (Ruoslahti & Pierschbacher 1987, Aota et al. 1994, Sechler et al. 1997). This interaction initiates a series of events, all of which are not fully known yet in detail, but they include (i) reorganization of the intracellular actin cytoskeleton, (ii) translocation of the initially diffusely scattered α5β1 integrins on the plasma membrane first to focal adhesions, and after a potential conformational change, to fibrillar adhesions, and (iii) conformational activation and rearrangement of fibronectin, resulting in the exposure of binding sites for other fibronectin dimers. Eventually, (iv) the fibronectin fibrils are converted into long, detergent-insoluble fibres organized and deposited in the ECM as a network. (Johansson et al. 1997, Schwarzbauer & Sechler 1999, Pankov et al. 2000, Geiger et al. 2001, Wierbicka-Patynowski & Schwarzbauer 2003, Clark et al. 2005, Mao & Schwarzbauer 2005.)

36

The RGD site in the type III10 repeat is essential for the α5β1 integrin-initiated fibronectin fibrillogenesis. Once successfully initiated, its presence is not, however, necessary for the propagation of the fibronectin matrix assembly. (Ruoslahti & Pierschbacher 1987.) The synergy site PHSRN in the adjacent type III9 repeat plays an enhancing modulatory role, although the lack of it does not block fibrillogenesis and can be compensated for with Mn2+ (Danen et al. 1995, Sechler et al. 1997). The N-terminal assembly domain consisting of type I repeats1-5 is also indispensable for the process of fibrillogenesis. It serves as a binding domain for the fibronectin molecules. The lack or the exogenous excess of this domain totally abrogates the assembly. (McKeown-Longo & Mosher 1984, McKeown-Longo & Mosher 1985, McDonald et al. 1987.) Type III repeats have also been studied in conjunction with their capability to interact with fibronectin dimers. It appears though, that their role in successful fibrillogenesis is not as crucial as that of the N-terminal domain; they rather seem to have a modulatory role (Hocking et al. 1994, Hocking et al. 1996, Sechler et al. 1996, Sechler et al. 2001). Oberhauser and co-workers have used single-protein atomic force microscopy to study the folding and conformation of the type III repeats, and have found that these repeats show a distinct mechanical unfolding hierarchy. Due to this unfolding, fibronectin as a molecule is highly extensible. This may facilitate the adoption of various conformations that can eventually regulate the rate of fibrillogenesis. (Oberhauser et al. 2002.) Fibronectin extension may play an important role in exposing binding sites to allow intermolecular interactions required for the incorporation of incoming fibronectin dimers into fibrils (Mao & Schwarzbauer 2005). In the polymerization phase, the initially short juxtacellular soluble fibrils are converted into a network of thicker and longer detergent deoxycholate-insoluble fibrils. The mechanism of this detergent-insolubility is not fully understood, but the current perception holds that protein-protein interactions rather than the interprotomeric disulphide bonds stabilize the macromolecule structure against disruption with detergents. (McKeown-Longo & Mosher 1983, McKeown-Longo & Mosher 1984, Mann et al. 1988, Chen & Mosher 1996.)

2.4.3 Collagens and the fibronectin matrix assembly

The fibronectin matrix assembly is regulated by intracellular factors (see for review Wierbicka-Patynowski & Schwarzbauer 2003). On the other hand, as collagens and fibronectin comprise the two major components of the ECM, it is feasible that the correct incorporation and/or turn-over of one may reciprocally affect those of the other. Indeed, these two appear to be interrelated. Namely, Dzamba and co-workers used Mov13 mouse fibroblasts, which do not produce type I collagen, to discover that these fibroblasts were also unable to produce a normal fibronectin matrix. Cells synthesized abnormally short collagen fibrils and the fibronectin fibres were also truncated. However, restoration of type I collagen synthesis by transfection resulted in the production of an extensive fibronectin matrix with fibrils of normal length, suggesting that the polymerization phase, rather than the very initiation of the fibronectin fibrillogenesis profited from the presence of an intact type I collagen structure. (Dzamba et al. 1993.) Also, expression of a mutant form of this collagen has been found to affect both the biosynthesis and the surface

37

expression of fibronectin (Lebbe et al. 1997). Sabatelli and co-workers discovered that the absence of secreted type VI collagen in cultures of primary fibroblasts derived from Col6a1-null mutant mice resulted in an altered fibronectin fibre organization. Fibroblasts from Bethlem myopathy patients with a drastically reduced secretion of type VI collagen, or normal fibroblasts having been treated with an anti-collagen VI antibody failed to incorporate fibronectin matrix in an adequate amount or in its proper three-dimensional organization. (Sabatelli et al. 2001.) Furthermore, dermal fibroblasts derived from types I and IV Ehlers-Danlos syndrome patients with mutations in COL5A1 and COL3A1 genes, respectively, synthesized structurally aberrant collagens type V and III with defective deposition of these proteins into the ECM. Additionally, these cells exhibited reduced levels of fibronectin in the culture medium, and a concomitant lack of fibrillar fibronectin network. (Zoppi et al. 2004.) Also, in Schwann cells type IV collagen and/or ascorbate were found to be needed for an abundant accumulation of fibronectin (Chernousov et al. 1998). Very recently, Ito and co-workers reported that the reduced thermostability of mutated type II collagen resulted in an atypical binding of type II collagen to fibronectin. This subsequently translated to a deranged organization of the fibronectin network. (Ito et al. 2005.) Taken together, these examples support the concept of collagens being able to affect the structural and functional integrity of the fibronectin matrix. As for the transmembrane type XIII collagen, Tu and co-workers have reported in vitro binding between the full-length fibronectin and the recombinant ectodomain. This interaction of relatively high affinity with a KD value of 17.5 nM was shown to be mediated by the collagenous domains of type XIII collagen. The pepsin-digested collagenous domains showed individual binding to fibronectin, although less strongly than the intact full-length ectodomain. (Tu et al. 2002.) How ectodomain binding affects fibronectin polymerization was not elaborated upon in that particular study.

Reciprocally, fibronectin seems to influence the polymerization of collagens. Velling and co-workers demonstrated that, in contrast to earlier suppositions of self-polymerization of type I and III collagens, a preformed fibronectin matrix appears to be a prerequisite for collagen network formation. No collagen matrix was formed in the absence of fibronectin or collagen-binding integrin receptors, whereas the presence of fibronectin alone was sufficient for collagen polymerization. The absence of the fibronectin matrix could be circumvented to some extent, albeit not fully, by collagen receptors α11β1 and α2β1 integrins. However, in the simultaneous presence of these collagen receptors and fibronectin, a well-organized collagen network was formed. (Velling et al. 2002.) It has been proposed that the collagen fibrillogenesis is a downstream event of the fibronectin assembly, and is mediated by interactions with cell-surface integrins (Canty & Kadler 2005). Fibronectin polymerization is not only necessary for the deposition of type I collagen and thrombospondin into the ECM, but the continual fibronectin polymerization is needed for the maintenance, deposition and stability of the ECM components, including type I collagen and fibronectin itself. In the absence of maintained fibronectin polymerization, cellular intake of fibronectin via the caveolin-1-dependent endocytosis was enhanced, leading to increased fibronectin degradation. (Sottile & Hocking 2002, Sottile & Chandler 2005.) Also, the mutual binding sites both in fibronectin and type I collagen have been described (Engvall & Ruoslahti 1977, Kleinman et al. 1978, Balian et al. 1980, Ruoslahti et al. 1981, Steffensen et al. 2002, Colombi et al. 2003).

38

2.5 The tissue microenvironment concept of cancer

The contribution and role of an intact ECM in the correct development of tissues and organs during morphogenesis is well recognized (Vu 2001, Behonick & Werb 2003, Kleinman et al. 2003). Now its role in carcinogenesis has been under reconsideration, leading to new insights about the concept concerning the initiation and progession of cancer. Not too long ago the prevailing theory on the initiation of epithelial-derived cancers held that the cumulative mutations in oncogenes and tumour suppressor genes of the epithelium were the cause of cancer. However, with the emergence of a more profound understanding of the role of the stroma, it is currently acknowledged that not only is the epithelium altered in cancers, and is still likely the primary source of it, but due to contextual cues from the tumour, the surrounding stromal compartment also undergoes a profound change to create a microenvironment that can actively foster tumour growth. This occurs via exploitation of resident “primed” cells, i.e. cells with tumorigenic potential, yet a low innate pre-disposition to develop a tumour. In a non-tumour stroma without exogenous tumour-promoting stimuli, these cells may remain indolently in dormancy in respect to tumorigenesis, whereas in the new context of the activated surroundings, these cells can be confined to the neoplastic agenda. Thus, the reactive stroma makes a pivotal contribution to carcinogenesis by creating a new context favorable to tumorigenesis. Hence, the tumour stroma is thought to act more as a coconspirator rather than a silent bystander in the carcinogenic process: while the normal stroma may protect the epithelium from tumorigenesis, an aberrant stroma can be centrally involved in the initiation of it. The interaction between two altered compartments exacerbates the ignited chain of changes, leading to a “vicious cycle” of recurrent deranged responses ever more favorable of tumour development. The end result is an enhanced tumour cell survival, uninhibited progression and invasion or metastasis of the cancer. (Seljelid et al. 1999, Berking et al. 2001, Bissell & Radisky 2001, Sung & Chung 2002, Quaranta & Giannelli 2003, Bhowmick et al. 2004a, Weaver & Gilbert 2004, Amatangelo et al. 2005, Littlepage et al. 2005.)

2.5.1 Tumour stroma

The stromagenesis, or a stromal reaction, is a host reaction of the tumour-adjacent connective tissue, which renders it eventually permissive, progressive and supportive to the neoplastic process of the epithelium (Amatangelo et al. 2005). This results from the emergence of cells with novel phenotypes and the subsequent reciprocal exchange of altered information between the epithelial and stromal compartments. The information exchange may occur either by direct cell-to-cell contacts, by indirect contacts through the ECM or via the action of soluble or insoluble signaling molecules secreted by both parties. The tumour stroma consists of a variety of cells, which include fibroblasts, myofibroblasts, endothelial cells, macrophages, inflammatory cells, mast cells, adipose cells and smooth muscle cells, as well as the non-cellular ECM produced by the cells. (Sung & Chung 2002, Quaranta & Giannelli 2003.) The emergence of the reactive stroma seems to be a very early event in malignant transformation, as features of it can be

39

detected preceding malignancy. In prostate cancer the conversion of the normal stroma to an activated one was found to begin already during the preneoplastic stages, then evolving with cancer progression to displace the normal stroma (Tuxhorn et al. 2002a).

One characteristic feature of the stromal reaction as a response to the tumour-derived cues is the transition of normal fibroblasts to myofibroblasts, or “activated fibroblasts”. Myofibroblasts are mesenchymal cells with features of both smooth muscle cells and fibroblasts (Oda et al. 1988, Oda et al. 1990, Eyden 2001). α-Smooth muscle (α-SMA) and vimentin are recognized as marker proteins for myofibroblasts, whereas they usually lack expression of desmin. The up-regulated expression of α-SMA and myosin indicates cellular differentiation towards a smooth muscle cell phenotype with capability to contract. (Serini & Gabbiani 1999, Eyden 2001, Petrov et al. 2002, Tomasek et al. 2002.) The conversion of normal to tumour stroma is characterized by a concurrent expression switch of many fibrillar extracellular matrix proteins (desmoplasia) (Löhr et al. 2001). Myofibroblasts are implicated in the elevated production of MMPs and fibrillar proteins of the interstitial matrix, like fibronectin, glycosaminoglycans, tenascin and collagens type I, III, IV, VI and VIII, their cellular receptors integrins, and soluble factors, the seemingly endless list of which covers cytokines, growth factors (notably TGF-β1), chemokines and inflammatory mediators. (Faouzi et al. 1999, Powell et al. 1999, Bridle et al. 2001, Petrov et al. 2002, Tomasek et al. 2002.) The ability to produce such a large gamut of structural and regulatory proteins not only makes myofibroblasts key cells in organo- and morphogenesis, but also empowers them in the remodelling of the cancer stroma (Powell et al. 1999). The stromal cells of prostate cancer were found to be of a myofibroblast phenotype, as determined by positive immunohistochemistry for α-SMA (Tuxhorn et al. 2002a). Also, in a very recent study by Orimo and co-workers, CAFs, i.e. (breast) cancer-associated fibroblasts were found to be more competent than their non-cancerous counterparts in supporting tumour growth and vascularization of tumours when co-mingled with breast cancer cells. Indicative of myofibroblast phenotype, the CAFs exhibited α-SMA expression and contractility. (Orimo et al. 2005.) Stromal cells also contribute to the angiogenic switch, i.e. they secrete both MMPs and angiogenic factors facilitating an increased microvessel density, especially proximal to precancerous and cancerous foci (Tuxhorn et al. 2002b).

The factors leading to acquisition of a myofibroblast phenotype are not fully known yet, but growth factor TGF-β1 has been implicated as an inducer of the switch to myofibroblast phenotype as well as the upregulation of α-SMA and collagen expression, both in vitro and in vivo (Ignotz & Massague 1986, Ignotz et al. 1987, Serini & Gabbiani 1999, Eyden 2001, Petrov et al. 2002, Tomasek et al. 2002, Tuxhorn et al. 2002a). Berking and co-workers showed that in melanoma TGF-β1 stimulates the tumour-adjacent stromal cells to express and deposit fibrillar ECM proteins, which then convey a survival advantage and an increased metastatic capacity of melanoma cells (Berking et al. 2001). Recently, Bhowmick and co-workers demonstrated how TGF-β signaling in stromal fibroblasts modulates the oncogenic potential of the nearby epithelium. By conditional inactivation of the TGF-β type II receptor gene and hence TGF-β responsiveness in mouse fibroblasts, either dysplastic changes in the prostate or an invasive carcinoma in the forestomach of the mouse resulted. In addition to highlighting the role of TGF-β, this study also demonstrated the critical influence of the varying contexts and the plasticity of the modulation: depending on the tissue, i.e. here prostate or

40

forestomach, the same stimulus of the stromal cells can lead to either pre-cancerous or fully cancerous changes of the adjacent epithelia. (Bhowmick et al. 2004b.) Orimo and co-workers recently addressed the question of stroma modulation by showing that the breast cancer stromal fibroblasts secrete factors, here SDF-1, i.e. stromal-cell derived factor-1, which promote tumour angiogenesis by recruiting endothelial progenitor cells into the tumour masses, and additionally, supporting tumour growth by paracrine stimulation via CXCR4 chemokine receptors displayed by the tumour cells (Orimo et al. 2005).

3 Outline of the present study

Type XIII collagen was originally characterized as a membrane-bound transmembrane collagen integrated in the cell membrane by its hydrophobic domain (Hägg et al. 1998). On the plasma membrane its location was pinpointed to the adhesive structures called focal adhesions. This was deducted from findings where type XIII collagen was found to emerge in and disappear from the focal adhesions simultaneously with the well-established focal adhesion marker proteins vinculin and talin (Hägg et al. 2001). Soon enough it was discovered that the long ectodomain, without the hydrophobic transmembrane or the intracellular domains, could be recovered in the cell culture media of human keratinocytes and infected insect cells (Peltonen et al. 1999, Snellman et al. 2000a), suggesting that it might, in fact, be released from the cell surface. As the membrane-proximal part of the ectodomain was found to contain a sequence conforming to the consensus cleavage sequence of proprotein convertases, it was suspected to be shed from the plasma membrane by furin in an enzymatic cleavage reaction (Snellman et al. 2000a). Also, as the techniques for the purification of the released ectodomain were developed, subsequent in vitro assays demonstrated the binding of the recombinant ectodomain to a number of ECM proteins, including fibronectin (Tu et al. 2002). Previously, the expression of type XIII collagen had been studied on cellular and tissue levels with comparisons made between the expression levels of normal vs. cancer cells. However, hardly anything was known about the expression of type XIII collagen in various human pathological conditions. Based on all this, the goals of this thesis were defined as follows:

1. To characterize in more detail the ectodomain shedding of type XIII collagen in mammalian cells.

2. To characterize the biological implications of the ectodomain shedding of type XIII collagen with respect to earlier in vitro binding results.

3. To characterize the expression level of type XIII collagen in various human pathologies, here cancer and its preceding stages, and also, to elaborate on the cellular mechanisms leading to these expression levels found.

4 Materials and methods

Descriptions of the methods and materials used in the studies of this thesis can also be found in the original articles I-III.

4.1 Construction of the expression plasmids (I, II, III)

For the transient expression in mammalian cells of mouse type XIII collagen (bases 1-2727) and its deletion variants delIC (bases 580-2727, with an in-frame artificial ATG to facilitate translation), delEC (bases 1-655), delCOL3-NC4 (bases 1-1873) and delCOL2-NC4 (bases 1-1424), expression plasmids encoding these variants were constructed in mammalian pcDNA3.1(-)/Myc-HisTag expression vector (Invitrogen). Site-directed mutagenesis (GeneEditor™ in vitro Site-Directed Mutagenesis System, Promega) was first applied to the full-length type XIII collagen cDNA to eliminate the stop codon and to include an in-frame vector-derived Myc tag, after which the modified type XIII collagen cDNA was inserted into the pcDNA3.1(-)/Myc-HisTag vector version B. For cloning purposes, new EcoRI restriction sites at either the 5´ end (delIC) or the 3´ end (delEC, delCOL3-NC4, delNC4) of the respective type XIII collagen cDNA sites were created by mutagenesis. The ensuing clones were digested with EcoRI (delIC) or with NotI and EcoRI (delEC, delCOL3-NC4, delNC4) for the cloning into the pcDNA3.1(-)/Myc-HisTag vector B. The deletion variant delCOL2-NC4 was created by ligating the NotI and BamHI-digested fragment of the full-length type XIII collagen cDNA into the pcDNA3.1(-)/Myc-His vector version C. All the expression plasmids contained an in-frame vector-born Myc fusion tag. The primers (Sigma Genosys) that were used in the mutagenesis were: 5´GACAGTGGCGCTGGCTGAGCAGTGCGCTAGGGAATTCATGCTGCCGAGCCCCGGGTGCTGCGGACTGCTGGC3 (delIC) 5´TTCGCTGGCGCTCAGCCTGCTCGCCCACTTTGAATTCCGGACCGCGGAGCTGCAGGCCCGGGTGCTG3´ (delEC) 5´GCTTCAGGAGATCAGGACGCTGGCCTTGATGGAATTCGGGCCTCCTGGTCTTCCTGGACAAACAGGCCC3´ (delCOL3-NC4) 5´AAGGGGGACCAAGGAGCACCTGGACTAGACGAATTCGCCCCCTGCCCGCTGGGAGAAGATGGC3´ (delNC4).

43

The correct cloning was checked by sequencing the 5´ and 3´ insertion sites of the plasmids with an ABI Prism 377 DNA Sequencer (Applied Biosystems). The plasmids to be transfected were purified by an alkaline lysis method (Sambrook et al. 1989).

4.2 Production of type XIII collagen antibodies (II, III)

Synthetic peptides corresponding to the type XIII collagen NC2 and NC4 domains (amino acid residues 229-241 and 709-727, respectively) (Pihlajaniemi & Tamminen 1990) were obtained from the peptide synthesis core facility at Biocenter Oulu, University of Oulu. The NC4 peptide was cyclized at cysteine residues and coupled to a carrier BSA (Imject® BSA, Pierce), and used as an antigen to raise the anti-NC4 domain antibodies (T22B or 777B) in rabbits (Harlow & Lane 1998). Antibodies were affinity-purified as described earlier (Hägg et al. 1998). Their specificity was analysed with immunoblotting by using purified type I, III, IV, V and VI collagens, lysates of HT-1080 cells and K562 cells expressing full-length type XIII collagen or the delNC4 variant. The recognition specificity of the antibodies was tested by a 4-fold peptide antigen competition.

4.3 Western blotting (I, II, III)

Cell lysates and precipitated media samples were analyzed for the contents of a given protein by immunoblotting. The protein concentrations of the samples were determined using a BCA Protein Assay Kit (Pierce). Equal amounts of protein were fractionated using 8, 10 or 12% SDS-PAGE gels, depending on the expected protein size, blotted to nitrocellulose filters and hybridized with a primary antibody of interest. The antibodies used in the various Western blotting experiments included anti-c-Myc oncoprotein (NeoMarkers, Abcam), anti-fibronectin (Sigma), anti-PCNA, anti-actin (used as a loading control), anti-α4, anti-α5, anti-αV, anti-α2, anti-β1 integrins and anti-cadherin (all from Santa Cruz Biotechnologies Inc.), anti-vimentin (Abcam), anti-α-smooth muscle actin (Abcam), and anti-cytokeratin (Dako). The horseradish peroxidase-conjugated anti-rabbit, anti-mouse and anti-goat (BioRad) were used as secondary antibodies. For the detection, enhanced chemiluminescence was utilized, using either ECL (Amersham Pharmacia) or Lumilight (Hoffman-La Roche) Western Blotting Detection Kits.

4.4 Image acquisition (I, II, III)

In Western blotting sample bands were quantified with BioRad GS-710 densitometer with QuantityOne-4.2.2. software (BioRad). The different staining specimens were viewed with an Olympus BX51 microscope (Olympus), photographed with a DP50 digital camera (Olympus). All exposures were made digitally with the analySIS software

44

(Olympus). Cultured cells were viewed with Leica DMIL microscope and photographed with Coolpix 990 digital camera (Nikon).

4.5 Cell culture (I, II, III)

4.5.1 Culturing of cells (I, II, III)

The following human cell lines were used (the cultivation medium given in parenthesis): primary dermal fibroblasts, fibrosarcoma HT-1080 cells, cervix adenocarcinoma HeLa cells, osteosarcoma MG-63 cells, mammary gland adenocarcinoma MCF-7 cells (all in Dulbecco´s modified Eagle´s (DME), Biochrom KG), colon adenocarcinoma LoVo cells (Ham´s F12K, BioWhittaker), normal colon FHC fibroblasts (1:1 mixture of Ham´s F12K and DME with 25 mM HEPES, 10 ng/ml cholera toxin, 5 μg/ml insulin, 5 μg/ml transferrin and 100 ng/ml hydrocortisone), colon adenocarcinoma SW-480 cells (Leibovitz´s L-15, Gibco), colorectal adenocarcinoma HT-29 cells, transitional carcinoma T24 cells of the urinary bladder, endometrial adenocarcinoma HEC-1-A cells (all in McCoy´s5a, Gibco), chronic myelogenic leukaemia K562 cells (Iscove’s Modified Dulbecco’s Medium, Gibco), mammary gland ductal carcinoma T-47D cells and ovary adenocarcinoma OVCAR-3 cells (RPMI-1640, Gibco). Cells were grown at +37oC in a humidified atmosphere with 5% CO2, except for SW-480 cells which were grown without CO2. K562 cells were cultivated in suspension. All the media contained 10% heat-inactivated foetal bovine serum (FBS, EuroClone), except 20% FBS for OVCAR-3 cells, and were supplemented with 100 U/ml penicillin-streptomycin (EuroClone), 0.25 μg/ml amphotericin B (Gibco) and 2 mM L-glutamine (Gibco). MCF-7 and OVCAR-3 cells were with 5-10 μg/ml insulin. Chinese hamster ovary (CHO) cells and African green monkey kidney fibroblasts (COS-1) were grown in DME with 10% FBS.

4.5.2 Transfection (I, II, IIII)

Cells to be transfected were either counted with a Coulter™ cell counter (Coulter Inc.) (K562 cells) or split from a single cell pool (adherent cells) 24 hours prior to transfection to ensure equal cell numbers between parallel experiments. FuGENE6 transfection reagent (Boehringer) was used as a transfection reagent in a 1:1.5 ratio for the amount (μg) of the plasmid and the transfection reagent (μl). Transfection time varied between 24-48 hours.

4.5.3 Coating (I, III)

96-well plates (MaxiSorp™; Nunc Nalge), Boyden chambers (Transwell®, Costar) or glass cover slips were coated with fibronectin, vitronectin, Englebreth-Holm-Swarm (EHS) laminin, poly-D-lysine, type I and IV collagens (all from Becton Dickinson

45

Collaborative Biochemical Products) or the recombinant human type XIII collagen ectodomain (Tu et al. 2002), using the final concentration of 2 or 5 μg/cm2, depending on the assay. Coating was performed in a moist chamber at RT for two hours, after which the coated surfaces were rinsed with PBS, blocked for two hours at RT with heat-inactivated 2% BSA (Serva/Invitrogen Life Technologies) and again rinsed with PBS. In case of double coating, the primary surface was re-coated with the recombinant type XIII collagen ectodomain, following the same guidelines.

4.5.4 Cell adhesion assays (I, III)

Primary fibroblasts in serum-free DME medium containing soluble recombinant type XIII collagen ectodomain or BSA were plated on 96-well cell plates with various coatings for at least 10 minutes. When determining how type XIII collagen ectodomain supports cell adhesion as a substratum, cells were allowed to adhere for three hours, without supplementation of the medium with the soluble ectodomain. Alternatively, equal numbers of primary dermal fibroblasts, HT-1080 cells or HeLa cells in serum-free DME medium were plated for 30 minutes on fibronectin- or vitronectin-coated or type XIII collagen ectodomain-double coated 96-well plates. Unbound cells were removed by washing with PBS. The number of adhered cells was determined with a CyQuant Cell Proliferation Kit (Molecular Probes) and quantified using a 1420 Victor™ multi-label plate reader (Perkin-Elmer Life Sciences/Wallac).

4.5.5 Cell spreading assay (III)

Primary dermal fibroblasts, HT-1080 cells or HeLa cells were plated on fibronectin- or vitronectin-coated or type XIII collagen ectodomain-double coated cover slips, and were allowed to adhere and spread for up to 24 hours. During this time the cells were viewed with Leica DMIL microscope and photographed with Coolpix 990 digital camera (Nikon) at regular intervals.

4.5.6 Cell migration assays (I)

In the haptotactic migration assay (Yamashiro et al. 1998, Ebert et al. 2000), primary fibroblasts were plated in serum-free DME medium supplemented with the recombinant type XIII collagen ectodomain or BSA on vitronectin or fibronectin-coated Boyden chamber inserts, and permitted to transmigrate for three hours. Cells on the upper side of the membrane were removed with a cotton swab, followed by fixation of the membranes with 1% paraformaldehyde, permeabilization in ethanol, staining with hematoxylene and mounting on microscope slides. When determining how type XIII collagen ectodomain supports cell migration, the cells were plated on ectodomain-coated inserts in serum-free medium without additional supplementation of the medium with it. In the scratch wound assay, denuded areas were created by scraping the monolayer cell cultures with a pipette

46

tip. After rinsing to remove the detached cells, the remaining adherent cells were incubated with serum-free cell-derived conditioned medium or with serum-free medium with the recombinant type XIII collagen ectodomain. Cells were allowed to migrate over the denuded area for 4 or 24 hours, after which they were photographed.

4.5.7 Cell proliferation assays (I, III)

Primary fibroblasts were split on vitronectin or fibronectin-coated 96-well plates in serum-free medium with the recombinant type XIII collagen ectodomain or BSA, and were allowed to grow for 4, 24, 48 or 72 hours. Alternatively, equal numbers of primary fibroblasts, CM(HT-1080) or CM(HeLa) cells in serum-containing DME medium were plated on uncoated 96-well plates, and were allowed to grow for 4, 24, 48 and 72 hours. The culture media were refreshed with daily. Cell proliferation was determined with a CyQuant Cell Proliferation Kit (Molecular Probes. Cell phenotypes were monitored by photographing with a phase contrast microscope at regular intervals.

4.6 Analysis of the type XIII collagen ectodomain shedding (I)

4.6.1 Whole-cell shedding assay (I, II, III)

The cells were split and plated to sub-confluency, after which the serum-containing medium was replaced with a serum-free medium and supplemented with the desired reagent for a given period of time. These included phorbol 12-myristate 13-acetate (PMA), brefeldin A, a calcium ionophore A23187, monensin and 1,10-phenantroline (all from Sigma), TAPI (Calbiochem), GM6001 (Chemicon), leupeptin and aprotinin (Roche), decanoyl-RVKR-chloromethylketone (Bachem), EDTA (Fluka) and various protease inhibitors, including antipain, bestatin, chymostatin, E-64, phosphoramidon and EGTA (Boehringer Mannheim). Pervanadate was prepared prior to use by mixing equal volumes of 2 M sodium orthovanadate (Sigma-Aldrich) and 1 M H2O2 (Merck) (Codony-Servat et al. 1999, Gutwein et al. 2000). The final concentrations of the reagents used are described in the original articles. The shed type XIII collagen ectodomain from the media was precipitated at -20°C with an equal volume of ice-cold methanol, collected by centrifugation and suspended in a lysis buffer of 150 mM NaCl, 2 mM EDTA, 1 mM PMSF, 1% Triton X-100 and 20 mM Tris, pH 7.6. The cell lysates were made by detaching the cells by scraping, suspending and homogenizing the cells manually in the lysis buffer. When evaluating the shedding efficiency of the type XIII collagen deletion variants, equal numbers of CHO cells were first transfected with 3 μg of a given plasmid cDNA and 4.5 μl of FuGENE6 transfection reagent for 24 hours, after which they were left in serum-free medium for an additional 24 hours, after which the released type XIII collagen deletion variants were recovered from the media by methanol precipitation. Samples were analyzed by immunoblotting with anti-Myc antibody.

47

4.6.2 Cell-surface biotinylation (I)

The cell surfaces were biotinylated prior to the shedding assay with FluoReporter Cell-Surface Biotinylation Kit (Molecular Probes). Equal numbers of cells were washed three times with ice-cold PBS, allowed to sit on ice for 30 minutes in Biotin-XX SSE in DMSO (200 ng/ml) diluted with ice-cold PBS, and washed three times with ice-cold PBS. The biotinylated cells were subjected to inhibitor treatments of interest, performed the same way as the non-biotinylated cells.

4.6.3 Streptavidin pull-down (I)

After biotinylation, the cell membranes were fractionated (Hägg et al. 1998). Cells were detached by scraping, and after freezing and thawing, the precipitable fraction of the ruptured cells was pelletted by centrifugation at 15000 rpm for 10 minutes, and resuspended in 5 mM HEPES, 1 mM EDTA, 0.25 M sucrose (HES) buffer. The cell suspension was homogenized thoroughly, diluted in HES buffer and centrifuged at 2200 rpm for 10 minutes. The postnuclear supernatant was centrifuged at 17600 rpm for 60 minutes, and the pelletted membranes were suspended in a small volume of 1% Triton X-100 in HES. Protein concentrations were determined to ensure loading of equal amounts of protein to the streptavidin pull-down with 50% (v/v) avidin-conjugated Protein A Sepharose (Pharmacia) overnight at +4oC. The streptavidin-Sepharose with the bound biotin-containing cell surface proteins was washed five times with PBS at +4oC, boiled in a buffer containing 1% Triton X-100, β-mercaptoethanol and 0.1% SDS and spinned. Equal volumes of the supernatants were applied to Western blotting. (Arribas et al. 1996, Baldwin & Ostergaard 2002).

4.7 Analysis of the fibronectin matrix assembly (II)

4.7.1 Ectodomain association with the fibronectin matrix (II)

In order to study how the exogenous recombinant type XIII collagen ectodomain associates with the fibrillar fibronectin matrix formed from exogenous fibronectin, α5-positive CHO cells were first cultured for 24 hours with 20 μg/ml exogenous human fibronectin (Collaborative Biomedical Products) in serum-free DME medium to allow formation of a fibrillar fibronectin matrix. After washing with PBS, the cells were incubated with 20 μg/ml recombinant type XIII collagen ectodomain in serum-free DME medium for three hours, followed by fixing and staining. When studying how the endogenously cleaved type XIII collagen ectodomain associates with the endogenously synthesized fibronectin matrix, primary fibroblasts were cultured in serum-containing medium for 72 hours to allow deposition of a fibrillar fibronectin matrix and a concomitant endogenous expression and cleavage of type XIII collagen ectodomain, and

48

its binding to the fibronectin matrix. After incubation, the cell cultures were fixed and stained.

4.7.2 Immunofluorescence staining (II)

For immunofluorescence visualization, the cells and the surrounding matrix were fixed with 3% PFA at RT for 15-30 minutes, followed by rinsing with PBS and blocking with 2% BSA in PBS at RT for 60 minutes. For the staining, an appropriate dilution of the primary antibody against the target of interest was made in 2% heat-inactivated BSA in PBS and incubated in a moist chamber at RT for 60 minutes. After washing with PBS, the sample was incubated with a suitable dilution of the non-isotype secondary antibody at RT for 30-60 minutes. After washing with PBS, the samples were mounted in Immu-Mount™ medium (Shandon Inc). Primary antibodies used in the immunofluorescence stainings included anti-Fn (IST-4), anti-type XIII collagen NC3 domain antibody (Hägg et al. 1998) and anti-Myc (Abcam). The secondary antibodies used included GAM-Alexa 488 (Molecular Probes), TRITC-labelled SAR (DAKO A/S), GAM-Cy2 and GAR-Cy3 (Jackson Lab).

4.7.3 Transmission electron microscopy (II)

CHO cells were supplemented with 20 μg/ml fibronectin in serum-free medium for 24 hours, after which the medium was changed for three hours to serum-free one containing 15 μg/ml recombinant type XIII collagen ectodomain. Following washing with PBS, samples were fixed with 3% paraformaldehyde in PBS for 10 minutes on ice, washed with PBS and blocked first with 0.05% glycine in PBS for 15 minutes and then with 5% BSA in 0.1% CWFS gelatin-PBS (Aurion) for 30 minutes. After washing with 0.1% BSA in cationic BSA (c/BSA), the cell culture was incubated for 60 minutes with 20 μg/ml anti-fibronectin antibody (IST-4), washed with 0.1% BSA in c/PBS, incubated with diluted rabbit anti-mouse antibody (Zymed) for 30 minutes and washed with 0.1% BSA in c/PBS. Fibronectin-bound antibody complexes were detected by incubation for 60 minutes with Protein A-conjugated 5 nm gold particles (purchased from G. Posthuma, University of Utrecht, Netherlands). After washing, cells were fixed for 5 minutes with 1% glutaraldehyde in PBS. Cells were again blocked with 0.1% glycine in PBS and then with 1% BSA in PBS, followed by washing with 0.1% BSA in c/PBS. Cells were then incubated for 60 minutes with an appropriate dilution of anti-type XIII collagen NC3 domain antibody, followed by washing with 0.1% BSA in c/PBS. The ectodomain-bound antibodies were detected by incubation for 60 minutes with protein A-conjugated 10 nm gold particles (purchased from G. Posthuma, University of Utrecht, Netherlands). Cells were washed with 0.1% BSA in c/PBA. After immunolabelling, the cells were fixed in 2.5% glutaraldehyde in 0.1 M sodium phosphate buffer (pH 7.4) post-fixed in 1% osmium tetroxide (Electron Microscopy Sciences) and embedded in Epon Embed 812 (Electron Microscopy Sciences). Thin sections were cut with a Reichert Ultracut ultramicrotome and examined with a Philips CM100 transmission electron microscope. Images were captured by means of a CCD camera equipped with TCL-EM-Menu version

49

3 (Tietz Video and Image Processing Systems GmbH). As controls to correct immunostaining of fibronectin and type XIII collagen ectodomain, control stainings were done simultanously with omission of either of the primary antibodies. (van Bergen en Henegouwen 1989).

4.7.4 Determination of the binding sites (II)

4.7.4.1 Determination of the type XIII collagen binding site (II)

CHO cells were grown with exogenously added 20 μg/ml fibronectin in serum-free DME medium for 24 hours, whereas primary fibroblasts were allowed to synthesize and deposit fibronectin matrix endogenously by growing them in serum-containing medium for 48 hours. Prior to adding to the cell cultures, the recombinant type XIII collagen ectodomain was pre-incubated with a 3-molar excess of the anti-NC4 domain antibody T22B or rabbit non-immune IgG (Jackson ImmunoResearch Laboratories) used as an isotype control. The cells were then subsequently incubated for three hours, followed by fixing and staining. The experiment was repeated with the antigen peptide for T22B, for which the peptide corresponding the NC2 domain was used as a control. Primary fibroblasts were grown in serum-containing medium for 48 hours, washed with PBS, pre-incubated for 30 minutes with 20 μg/ml of the NC4 or NC2 peptide, followed by incubation with 20 μg/ml recombinant type XIII collagen ectodomain and either of the peptides for three hours, after which cells were fixed and stained.

4.7.4.2 Determination of the fibronectin binding site (II)

The surface plasmon resonance technology (Biacore™) was used to determine the fibronectin domain(s) responsible for the interaction with type XIII collagen ectodomain. 20 μg/ml recombinant type XIII collagen ectodomain in 10 mM sodium acetate, pH 5.5 was covalently immobilized as a ligand by amine coupling to a carboxymethyldextran CM5 sensor chip activated by serial application of N-hydroxysuccinimide, N-ethyl-N´-(3-dimethylaminopropyl)carbodiimide and ethanol-amine. The binding experiments were run at +25oC using different analyte concentrations of the full-length fibronectin or its recombinant fragments (Fn120K, Fn70K, Fn45K, Fn40K, Fn30K, FnIII1-C; Sigma or Chemicon). In the preliminary tests all the samples were run at concentrations of 25 nM, 50 nM and 100 nM, whereas with those fibronectin fragments that had initially shown binding to type XIII collagen ectodomain were run in the kinetic tests at concentrations of 0 nM, 20 nM, 50 nM, 100 nM, 300 nM and 900 nM diluted in the running buffer of 50 mM Tris-HCl, pH 7.4 and 150 mM NaCl in order to reduce the bulk effect. The analytes were passed over the immobilized ligand at a constant flow-rate of 30 μl/minute for 180 seconds. At the end of each cycle of the analysis, the sensor chip was regenerated with 2.5 M NaCl and 5 mM NaOH. The amount of analytes bound to the immobilized ligand was monitored by measuring the variation in the plasmon resonance angle as a function of time, and expressed as resonance units (RU). The background signal was subtracted from the

50

recorded values by a simultaneous injection of analytes over a blank surface (reference cell) composed of ethanolamine-substituted dextran and activated like the sample cells but without the ligand. The fitting of the crude data, preparation of the overlay plots and determination of KD values were performed using the Biaevaluation 3.0 software (Biacore™).

4.7.5 Quantification of the fibronectin matrix formation (II)

4.7.5.1 The immunofluorescence method (II)

To ensure proper cell adherence, equal numbers of primary fibroblasts were plated on cover slips in fibronectin-depleted medium for two hours. The fibronectin-depleted growth medium was produced by replacing FBS with serum form which fibronectin had been extracted by using affinity chromatography with Gelatin Sepharose® 4B (Pharmacia Biotech) (Engvall & Ruoslahti 1977). The culture medium was changed to serum-free DME medium with exogenously added 10 μg/ml fibronectin and 50-100 μg/ml recombinant type XIII collagen ectodomain or type I collagen, and allowed to stay on cells for three hours, followed by fixation and staining with anti-fibronectin antibody (IST-4). Capturing of the fluorescence images was done with identical settings. The quantitative evaluation of fibronectin matrix-associated fluorescence was done from 10 randomly selected viewing fields per sample (Zoppi et al. 2004) with MetaMorph 6.1 image analysis software (Universal Imaging Corporation). Results were expressed as integrated optical densities (IOD) related to the area containing the fluorescent signal.

4.7.5.2 The insolubility method (II)

Equal numbers of primary fibroblasts were plated on cover slips in serum-containing medium, and were allowed to adhere properly for two hours. The cells were rinsed with PBS and incubated with 50 μg/ml recombinant type XIII collagen ectodomain, type I or VI collagens in serum-free DME medium for 24 hours. The cells were disrupted with hypotonic 25 mM NH4OH. After removal of the soluble material by washing with PBS, the attached fibronectin matrices were scraped off and recovered with 10 mM Tris-HCl, 8.0, 8 M urea, 1% SDS and 15% β-mercaptoethanol. After boiling for 5 minutes, equal volumes of samples were separated on a 5% SDS-PAGE gel under reducing conditions, followed by Western blotting with anti-Fn antibody (IST-4). (Dang et al. 2004).

51

4.8 Analysis of the type XIII collagen expression in malignant transformation (III)

4.8.1 Construction of tissue microarrays (III)

The patient material was gained from the archives of the Department of Pathology, University of Oulu. The tissue microarrays consisted of normal, dysplastic and tumour specimens from colon (n=109), cervix (n=113), urinary bladder (n=64), endometrium (n=86), ovary (n=63), and mesenchymal cancers fibrotic histiocytoma, dermatofibroma, fibrosarcoma, dermatofibrosarcoma, leiomyosarcoma, rhabdomyosarcoma and chondrosarcoma (n=89). After selection and marking of the representative foci in the hematoxylin-eosin stained donor tissue slices, the donor blocks were oriented with respect to the slices for the perfect relocation of the target sites. The 1 mm core needle “biopsy” samples from the donor blocks were drilled with a specially constructed drilling machine (Department of Pathology), and re-assembled into 96-sample array blocks. (Kononen et al. 1998). 4 μm sections were cut for the various applications.

4.8.2 In situ hybridization (III)

4.8.2.1 Construction of the type XIII collagen probes (III)

The human type XIII collagen sense and anti-sense probes for the in situ hybridization (ISH) were constructed by ligating the BamHI-SmaI fragment (bases 1538-2505) of the full-length human type XIII collagen cDNA into a pSP72 expression vector (Promega). Sequencing of the insertion ends was used to verify the correct cloning. The hybridization specificity of the probe was tested by Northern blotting using total RNA isolated from transiently transfected COS-1 cells overexpressing type XIII collagen. Linearization of the plasmid with either HindIII or EcoRI resulted in anti-sense or sense type XIII collagen templates, respectively. Transcribing them with T7 or SP6 RNA polymerases (Promega) yielded the respective digoxigenin-11-UTP-labelled (Roche) RNA probes.

4.8.2.2 The in situ hybridization protocol (III)

The hybridization was preceded by dewaxing of the paraffin-embedded samples with xylene and ethanol. The deparaffinated samples were incubated in 0.2 M HCl for 20 minutes and digested with 10 μg/ml proteinase K (Merck) for 30 minutes at +37oC, which was stopped by 0.2% glycine (Riedel-de-Haën) in PBS. Samples were postfixed with 4% PFA for 20 minutes, acetylated with 0.1 M triethanolamine (Riedel-de-Haen) in 0.25% acetic acid anhydride, and balanced by a phosphate-based Tris buffer, pH 6.8 for

52

15 minutes. The prehybridization was carried out at +60oC for two hours in 4x SSC, 10% dextrane sulphate, 0.01 M DTT, 2x Denhardt´s solution, 50% formamide and 250 μg/ml yeast tRNA. The probes were denatured at +80oC, after which the hybridization was carried out at +60oC overnight in the prehybridization buffer. The samples were washed serially with a decreasing concentration of SSC at +60oC. After blocking with a blocking buffer (Roche) for 30 minutes at RT, the bound probes were detected with alkaline phosphatase-conjugated anti-digoxigenin antibody (Roche) and the color reaction was performed using Fast Red color substrate (Roche). The slides were counterstained with 0.3% methyl green (Sigma) in sodium acetate, pH 4.0. The sense probe was used as a control for unspecific hybridization.

4.8.2.3 Scoring of the in situ hybridization results (III)

The ISH staining results were semi-quantified manually by means of a four-step grading score for the overall staining intensity (Bachmeier et al. 2000). The grading categories included score 0 for no staining signal, score 1 for a faint positive one, score 2 for a moderate positive one and score 3 for a strong positive staining signal. The actual scoring for individual samples was a combined result derived from viewing of three observers. The varying sample numbers within the series were taken into account in the overall signal intensity scoring by expressing the results as means of staining intensity.

4.8.3 Immunohistochemical staining (III)

Paraffin-embedded tissue sections were deparaffinated with xylene and ethanol. The endogenous peroxidase activity was quenched by incubating the samples in 0.3% H2O2 in methanol for 30 minutes, after which samples were serially treated with a decreasing concentration of ethanol. After washes with TBS, the samples were subjected to antigen retrieval by boiling for 15 minutes in 1 mM EDTA, pH 8, then blocked with 1% BSA for 60 minutes and incubated overnight at +4oC with a suitable dilution of the antibody 777B. The specimens were incubated with the biotinylated secondary anti-rabbit antibody and the avidin complexes, as prepared and instructed by the manufacturer (Vectastain ABC Kit, Vector Laboratories). Diaminobenzidine was used as a chromogen and hematoxylin as the counterstain.

4.8.4 Conditioned medium experiment (III)

Conditioned medium (CM) from HT-1080 and HeLa cells were prepared by incubating primary fibroblasts in serum-free DME medium overnight. The media were collected, centrifuged at 1000 rpm for 5 minutes, filtered through a 0.22 μm-filter, supplemented with 10% FBS and diluted 1:1 with serum-containing DME medium (Löhr et al. 2001, Arnold et al. 2002). The cells were cultivated in the CM throughout the experiment, in a

53

same way as their untreated control equivalents in normal serum-containing medium for the same length of time. The cell lines obtained were named CM(HT-1080) and CM(HeLa), respectively. In the CM-reverse experiment, the CM cell lines were deprived of their CM, and instead cultivated in normal growth medium. In the anti-TGF-β1 blocking experiment, the undiluted CM from HT-1080 cells was supplemented with 200 ng/ml anti-human TGF-β1 antibody (R&D Systems) throughout the entire experiment.

4.8.5 Northern blotting (III)

Primary human fibroblasts were serum-starved for 24 hours, after which they were incubated in serum-free medium with 0.1% BSA, and either with 1 or 5 ng/ml TGF-β1 (Chemicon) or 100 ng/ml phorbol 12-myristate 13-acetate (PMA, Sigma) for 24 hours (Kissin et al. 2002, Kubota et al. 2003). The mRNAs were isolated with the MicroFastTrack® 2.0 mRNA Isolation Kit (Invitrogen). Total RNAs were isolated with SV Total RNA Isolation System (Promega) from COS-1 cells transiently transfected with pcDNA3.1(-)/Myc-HisTagB/COLXIII (type XIII collagen) cDNA. Untreated or untransfected cells, depending on the assay, were used as control samples. Equal amounts of mRNAs or total RNAs run on a 1% MOPS agarose gel with 1x MOPS as a running buffer. Transfer to nitrocellulose was done according to standard protocols, hybridizations were performed at +42oC overnight using a nick-translated [α-32P]dCTP-labelled human type XIII collagen cDNA digestion fragment (bases 1538-2505) as a probe. Membranes were washed extensively at +65oC in a decreasing concentration of SSC and 0.1% SDS. Loading of mRNAs was normalized by hybridization with a β-actin control probe (Clontech).

5 Results

5.1 Characterization of type XIII collagen ectodomain shedding in mammalian cells (I)

5.1.1 Shedding of ectodomain by proprotein convertases (I)

One of the main goals of the studies in this thesis was to characterize the shedding of type XIII collagen ectodomain in mammalian cells and to address its implications on cell behavior and dynamics of the ECM.

Human fibroblasts of the skin and colon, as well as HT-1080 fibrosarcoma cells, were used to determine the size of the cleaved ectodomain. All three cell lines released type XIII collagen ectodomain into the culture medium, although with different efficiencies, as determined by methanol precipitation of the culture media. As analysed by Western blotting with an anti-NC3 domain antibody, the recovered ectodomains migrated as a slightly smaller fragment than the intact type XIII collagen of the cell lysates. The size difference between the two was estimated to be about 10 kDa, irrespective of the cell line. In order to specify the protease(s) responsible for ectodomain shedding, a series of protease inhibitors were tested for the blocking of ectodomain shedding in HT-1080 cells. Serum-free whole-cell shedding assays were supplemented with either decRVKRcmk, antipain, bestatin, chymostatin, E-64, leupeptin, pepstatin, phosphoramidon, EDTA, aprotinin or EGTA, after which the methanol-precipitated media were analyzed in Western blotting with anti-type XIII collagen NC3 domain antibody. This showed that effective blocking of shedding in the HT-1080 cells was seen only with decRVKRcmk. The potential involvement of MMPs or ADAMs was excluded by lack of an inhibitory effect on the shedding by 1,10-phenantroline, GM6001 or TNF-α convertase (TACE) inhibitor TAPI-1. In order to corroborate this and to study the shedding event at the cell membrane, the cell surface proteins of HT-1080 cells were biotinylated. Again, the only effective inhibitor of shedding was decRVKRcmk. Cleavage by MMPs, ADAMs or active plasmin was ruled out, as ectodomain was released from the cell surface-biotinylated HT-1080 cells in spite of 1,10-phenantroline, GM6001, aprotinin or

55

leupeptin. Taken together, these inhibitor experiments strongly proposed the solitary involvement of the PCs.

Furin-deficient LoVo cells expressing only PC7 and PACE4 (Bennett et al. 2000, Ebert et al. 2000) were used to discriminate between various PCs in ectodomain shedding. The low endogenous expression of type XIII collagen in LoVo cells necessitated their transfection with a Myc-tagged type XIII collagen expression plasmid prior to exposure to the shedding phase. Despite the furin-null background, the transfected LoVo cells showed ectodomain release into the medium, as detected by anti-NC3 domain and anti-Myc-tag antibodies in Western blotting of the precipitated media. The fact that decRVKRcmk blocked the ectodomain shedding here as well is explained by the lack of strict furin-specificity of decRVKRcmk. This inhibitor is redundant in its specificity within the PC group.

5.1.2 Intracellular regulation of ectodomain shedding (I)

PMA and pervanadate are recognized stimulators of ectodomain shedding. PMA activates the intracellular protein kinase C signaling cascade and pervanadate inhibits the intracellular phosphotyrosine phosphatases (Beer et al. 1999, Codony-Servat et al. 1999, Messageo et al. 2003). To characterize if these intracellular mechanisms are involved in modulating the ectodomain shedding of type XIII collagen, HT-1080 cells were treated with PMA, which was found to stimulate the release of the ectodomain especially when the basal rate of shedding was low, as with adherent cells at early time points of two to six hours. At a later time point of 24 hours and in detached cells in suspension, the enhancing effect of PMA was not equally as strong. As opposed to HT-1080 cells, PMA did not up-regulate ectodomain shedding in SW-480 cells. On the other hand, pervanadate treatment stimulated the ectodomain shedding in SW-480 cells, but had only an indeterminate effect in HT-1080 cells. Regardless of the cell line, detachment of the cells from the substratum into suspension with EDTA resulted in a notable up-regulation of ectodomain shedding. Consistent with the earlier results, decRVKRcmk reduced the ectodomain release in both adherent and detached cells, regardless of the cell line.

5.1.3 Time-dependence of ectodomain shedding (I)

In order to determine the time-dependence of the ectodomain release, the ectodomain shedding of HT-1080 cells was followed for up to 24 hours. Over this time frame, the cleaved ectodomain was found to accumulate in the medium concomitantly with the growing cellular expression of type XIII collagen, i.e. as the cell expression of type XIII collagen increased, the cells released more ectodomain into the medium. However, isolation of the membrane fractions revealed that the amount of type XIII collagen on the plasma membrane of the cells remained relatively constant throughout the follow-up period, with only a slight increase at the later time points, suggesting that the cell membrane level of type XIII collagen is maintained relatively even, with the excess effectively secreted out of cells. By comparison, the relative increase in the amount of the

56

membrane-bound type XIII collagen was smaller than the increase in the total cellular amount or that of the shed ectodomain in the medium. In line with this, the late-phase cultures were found to release more ectodomain into the medium relative to the cell number than the new early-phase ones during the same time frame. Cell surface biotinylation was applied to study the origin of the cleaved ectodomain. The shed ectodomain in the medium was found to be derived at least partly from cleavage at the plasma membrane, as the surface-biotinylated HT-1080 cells exhibited a time-dependent release pattern of biotinylated ectodomain into the medium.

5.1.4 Implications of Golgi involvement in ectodomain shedding (I)

Based on the results of the earlier experiments in this study, type XIII collagen ectodomain is shed at the cell surface. This is consistent with the known locations of furin, but since furin, PACE4 and PC7 are also known to be located in the trans-Golgi network, in fact predominantly so (Seidah & Chrétien 1997, Thomas 2002, Endres et al. 2003), the possibility of Golgi involvement in type XIII collagen ectodomain shedding was also explored. The assumption for this experiment was that if the Golgi apparatus is a site for type XIII collagen ectodomain processing, its biochemical manipulation leading to a malfunctioning Golgi may subsequently affect the shedding of ectodomain. Brefeldin A and monensin are known to hamper the intracellular trafficking of maturing proteins by causing fusion of the Golgi saccular structure with the endoplasmic reticulum (brefeldin A) or by disrupting vesicular transport within and from the trans-Golgi network to the plasma membrane (monensin). The calcium ionophore A23187 causes depletion of the intracellular calcium stores, rendering conditions for the calcium-dependent serine proteases suboptimal without substantially affecting intracellular trafficking (Ebert et al. 2000). As HT-1080 cells were exposed to brefeldin A, monensin or A23187 for six hours, followed by recovery of the cleaved ectodomain in the medium, the shedding of the ectodomain was clearly hampered. These results suggested that the Golgi apparatus may indeed be involved in the ectodomain shedding of type XIII collagen.

5.2 Characterization of the biological activity of the cleaved ectodomain (I, II, III)

A second goal of this thesis was to define if the cleaved type XIII collagen ectodomain possess any biological activity, which might affect cell behaviour and/or composition and dynamics of the ECM.

5.2.1 Cell adhesion (I, III)

Primary fibroblasts were used to study how the soluble recombinant ectodomain affects cell adhesion. Cover slips were coated with vitronectin, fibronectin, EHS laminin, poly-

57

D-lysine or type I or IV collagens, on which substrata fibroblasts were then seeded. The ectodomain inclusion showed time- and matrix-specificity in its effects, as the soluble ectodomain reduced adhesion of the primary fibroblasts only in a short-term experiment (t=10 minutes) reflecting the very initial phases of cell adherence and on vitronectin substratum. The inclusion of the soluble ectodomain had no effect on cell adherence when prolonged adhesion times or other substrata than vitronectin were tested.

As the soluble ectodomain had shown very restricted vitronectin-specific effects, an additional experimental set-up was created to model a situation where the cell-surface shed ectodomain binds to a vitronectin-enriched matrix (fibronectin was used as comparison). Cell adherence between the single- and double-coated surfaces was then compared. The double coating with ectodomain, when compared to single surfacing, was found to reduce the adhesion of primary fibroblasts, HT-1080 and HeLa cells, again only on vitronectin but not on fibronectin.

5.2.2 Cell migration (I)

The effect of the soluble ectodomain on cell migration was examined with two experimental set-ups. In a wound closure assay cells were allowed to migrate over a denuded crevasse made by scraping of the monolayer of cells with a pipette tip. The effect of the ectodomain was examined by either supplementing the cells with the recombinant soluble ectodomain or by HT-1080 cell-derived conditioned medium (CM) enriched in the shed ectodomain. At 24 hours, the denuded space between the edges of the “wound” had been covered more efficiently by migrating HT-1080 cells in a serum-free medium devoid of ectodomain than by those in the CM or with the exogenously added recombinant ectodomain. The reduced motility of HT-1080 cells was deduced to result from the effect of the soluble ectodomain, as the inclusion of only purified ectodomain was enough to hamper cell motility. As the cells exhibited a relatively typical migratory phenotype, it is likely that the ectodomain rendered the conditions suboptimal, although not entirely intolerable, for cell migration. In the haptotactic transmigration assay, i.e. Boyden chamber assay, the effect of the recombinant ectodomain on cell motility on isolated substrata was studied. Again, the exogenously added ectodomain showed a distinct matrix-specificity by interfering with the motility of primary fibroblasts specifically on vitronectin.

5.2.3 Ectodomain as a substratum (I)

The ability of the ectodomain to support cell adhesion and migration as a substratum was studied by plating primary fibroblasts on ectodomain coating. During a prolonged adhesion of three hours, primary fibroblasts showed little adherence to the ectodomain surface relative to vitronectin, fibronectin, EHS laminin and collagens type I and IV. When compared with the other surfaces, the ectodomain also provided poor support for cell migration, as studied by the haptotactic transmigration assay in Boyden chambers. The few cells that had migrated showed insufficient spreading on ectodomain.

58

5.2.4 Cell proliferation (I)

The effect on cell growth was studied by including the recombinant ectodomain in the serum-supplemented cell culture medium. Here the ectodomain compromised the growth rate of primary fibroblasts distinctly within the first 48 hours. The number of cells in the control culture had doubled within this time frame, whereas in the ectodomain-supplemented fibroblast culture only a marginal increase in the cell number was seen. At 72 hours, however, the ectodomain-treated cell culture exhibited an almost equal rate of cell proliferation as the control culture. Since fibronectin and vitronectin are the major cell attachment factors in serum (Stenn et al., 1983, Basara et al. 1985, Hayman et al. 1985, Romberger 1997, Schwartz et al. 1999), the experiment was repeated under serum-free conditions on isolated fibronectin or vitronectin. Again, the effect of ectodomain was substratum-specific as inhibition of cell proliferation was only seen on vitronectin, not on fibronectin. It was deduced that this proliferation effect resulted from a transient distraction of the early phase of cell adherence and spreading. Namely, at four hours the ectodomain-treated cells exhibited an aberrant morphology with inadequate spreading as compared with the BSA-treated control cells. The round cells displayed a total lack of spreading, while the others showed initial spreading. None of the cells exhibited the characteristic fibroblast-like phenotype as was seen in the control cells. The morphological differences had, however, become less distinct by 24 hours, at which time cells in both cultures showed a fully extended fibroblast-like morphology.

5.2.5 Cell spreading (III)

The cell spreading assay was performed on either solitary vitronectin or on ectodomain double-coated vitronectin. On vitronectin the majority of HT-1080 cells were well spread at three hours after plating, whereas on double-coated vitronectin only few cells showed even initial spreading. HeLa cells exhibited a generally slower spreading pace, but at 6.5 hours a similar spreading deficiency on double-coated vitronectin was seen. This effect was again a transient one, since after 24 hours no marked morphological differences were seen.

5.2.6 Interference of the fibronectin matrix assembly (II)

This experiment was based on an earlier finding of strong binding between the type XIII collagen ectodomain and fibronectin in a solid phase assay (Tu et al. 2002). Therefore the follow-up question for the experiment was whether the ectodomain might bind to a fibrillar fibronectin matrix in cell culture conditions, and if so, whether it might affect its assembly. Recombinant type XIII collagen ectodomain or type I collagen, the latter used as a collagen control, and exogenous fibronectin were added to fibroblast cultures for 3 hours. The formed matrices were stained with an anti-fibronectin antibody, followed by the measurement of the matrix-associated fluorescence intensity from randomly selected

59

viewing fields of the samples. Compared to the untreated sample, inclusion of the recombinant ectodomain, but not that of type I collagen, had interfered with the assembly of the fibronectin matrix. The estimated reduction in the matrix formation was about 35%. As for the fibre morphology, the fibronectin fibre structure in the ectodomain-treated culture seemed unaffected. However, with higher ectodomain concentrations the polymerizing fibronectin fibres aggregated to non-fibrillar precipitations.

Another method was also used for the same purpose. The effect of the ectodomain on the endogenous fibronectin matrix polymerization was analyzed semi-quantitatively with Western blotting. Here recombinant ectodomain was included in serum-containing medium for a given period of time, with collagens type I and VI used as collagen controls. Again, the presence of the soluble ectodomain had diminished the amount of the endogenously formed fibronectin matrix, here about 25% when compared to the untreated control. Consistent with the earlier results in this study, the influence of the ectodomain was distinctly different from that of the other collagens.

5.3 Association of ectodomain with the fibrillar fibronectin (II)

Immunofluorescence staining was used to study how the ectodomain, either exogenously added or endogenously shed, associates in vitro with the fibrillar fibronectin matrix, also either endogenously deposited or polymerized from exogenous sources. The first set-up was used to study how the exogenous ectodomain associates with the fibrillar fibronectin matrix formed out of exogenously added fibronectin. CHO cells were cultivated with fibronectin, followed by supplementation of the recombinant ectodomain. Double immunofluorescence staining with anti-type XIII collagen and anti-fibronectin antibodies showed that the recombinant ectodomain had associated congruently with the fibronectin fibres. In the second set-up, primary fibroblasts were used to study how endogenously synthesized and cleaved ectodomain associates with the fibrillar fibronectin matrix made by the cells. In line with the results from the CHO cell experiment, the immunofluorescence staining showed that the endogenously synthesized and cleaved ectodomain had organized in a fully overlapping manner with the endogenously deposited fibrillar fibronectin matrix. The ultrastructural analysis by the immunogold technique and transmission electron microscopy verified the impeccable linear alignment of the recombinant ectodomain along the fibronectin fibres.

5.4 Characterization of the binding sites (II)

5.4.1 The type XIII collagen binding site to fibronectin (II)

A series of expression plasmids encoding Myc-tagged full-length type XIII collagen and different deletion variants were constructed for the characterization of the domain(s) of type XIII collagen involved in mediating the association with fibronectin. First, following transient transfection of the CHO cells, the comparable shedding of the ectodomain

60

variants was verified in a whole cell shedding assay. Subsequently, immunofluorescence double staining with anti-Myc (for the detection of the Myc-tagged ectodomain variants) and anti-fibronectin antibodies revealed that the cleaved ectodomains of the full-length type XIII collagen and the variant lacking the intracellular domain (delIC) had fully associated with the fibronectin matrix. On the other hand, the variant lacking the conserved NC4 domain (delNC4) showed a markedly reduced association with the fibronectin matrix, and the ones with longer truncations (delCOL2NC4 and delCOL3NC4) did not associate at all.

The immunofluorescence data implied that the conserved C-terminus of type XIII collagen is involved in the association. This was studied by producing a specific polyclonal antibody, named T22B, against the NC4 domain. It was found to be specific for its antigen domain with no detectable cross-reactivity against other collagens. In distinct contrast to non-immune rabbit IgG used as a control, T22B antibody almost completely blocked the binding of ectodomain to the fibrillar fibronectin matrix, which was either endogenously deposited by primary fibroblasts or polymerized from exogenous fibronectin by CHO cells. The epitope peptide corresponding to the NC4 domain did not show a similar blocking effect, nor did a control peptide against the NC2 domain.

5.4.2 The fibronectin binding site to type XIII collagen (II)

The surface plasmon resonance method (Biacore™) was used to pinpoint which domain(s) of fibronectin mediate the binding of fibronectin to type XIII collagen ectodomain. A series of commercial recombinant fibronectin fragments, covering almost the entire length of the fibronectin molecule and corresponding to the various functional domains, were run over the immobilized ectodomain. Of all the fibronectin fragments studied, only the amino-terminal Fn70K fragment and the Fn45K fragment containing the collagen/gelatin binding domain showed association with the immobilized ectodomain. In more detailed kinetic tests, the mean KD values were estimated to be about 0.5 μM for the Fn70K and 15.6 μM for the Fn45K fragment, with χ2 values for the KD fittings in the range of 2.1-3.8 and 0.38-0.47, respectively, as analyzed in a Langmuir 1:1 model.

5.5 Type XIII collagen expression in cancer (III)

Since essentially no data has existed thus far on the expression of type XIII collagen in different human pathological conditions, the third main goal of the studies in thesis was to begin a survey on this topic. In view of the findings related to cell adhesion, migration and proliferation and type XIII collagen ectodomain, cancer was chosen as the disease with which to initiate this task.

61

5.5.1 Expression of type XIII collagen mRNA in various cancers (III)

The evaluation of expression levels was performed using a tissue microarray (TMA) approach, which as a method possesses considerable benefits, amongst the best of which is an efficient through-put of a high number of samples without variations in detection (i.e. staining, hybridization) conditions. TMAs were constructed of epithelial and mesenchymal tumours, also including dysplastic and normal samples. The carcinomas included colon adenocarcinoma, cervical squamous cell carcinoma, urinary bladder carcinoma, endometrial adenocarcinoma and ovarial cystadenocarcinoma, whereas the mesenchymal tumours included fibrotic histiocytoma, dermatofibroma, fibrosarcoma, dermatofibrosarcoma, leiomyosarcoma, rhabdomyosarcoma and chondrosarcoma. The constructed TMAs were analyzed using ISH with a human type XIII collagen anti-sense probe, the hybridization specificity of which was first evaluated by Northern blotting to be correct. Also, test hybridizations with the sense probe excluded unspecific hybridization. With the epithelial cancers, TMA ISH revealed an enhanced stromal expression of type XIII collagen mRNA, increasing progressively from normal to dysplastic to cancer stroma. All the mesenchymal cancers, apart from dermatofibrosarcoma, showed an upregulated expression of type XIII collagen mRNA relative to the control tissue. The overall staining signal intensities for each TMA ISH were scored semi-quantitatively and analyzed statistically (Bachmeier et al. 2000). With the carcinomas, the difference between normal vs. cancer stromal mRNA expression of type XIII collagen proved to be statistically significant in all cases, with no marked differences in the stromal mRNA levels between the cancer grades. The transition from the dysplastic to the malignant stroma induced a notable rise in type XIII collagen mRNA expression. The expression induction was distinct in the stromal compartment, whereas the dysplastic and malignant epithelia showed only slightly enhanced type XIII collagen mRNA expression relative to the normal epithelium. The magnitude of the epithelial expression enhancement was statistically non-significant and always diminutive by comparison with that found in the stromal compartments.

5.5.2 Expression of type XIII collagen in the invasive fronts (III)

The newly produced anti-type XIII collagen antibody, named 777B, was used for the immunohistochemistry of the cancer tissues. 777B correctly detected a protein of similar size as the previously reported anti-type XIII collagen NC3 domain antibody (Hägg et al. 1998). A membrane-associated type XIII collagen protein expression was detected in the invasive foci of the carcinomas, coinciding with the type XIII collagen ISH positivity. The epithelial origin of the cells was verified by cytokeratin positivity. Interestingly, upregulated type XIII collagen mRNA expression was also found in the physiologically invasive cytokeratin-positive cytotrophoblasts of the placenta. These findings showed that the upregulation of type XIII collagen expression is related to invasiveness, whether for malignant or physiological reasons. In addition to this, type XIII collagen expression was also evident deeper in the tumour mass. Typically, the cancer foci were surrounded by a locally upregulated type XIII expression in the adjacent stromal cells, suggestive of a paracrine expression induction by the cancer cells.

62

5.6 Induction of type XIII collagen expression (III)

5.6.1 The conditioned medium experiment (III)

To study the effect of cancer-derived growth factors and cytokines on the expression of type XIII collagen, an in vitro model was set up where primary fibroblasts were cultured in the CM from either epithelial (HeLa) or mesenchymal (HT-1080) cancer cells. The CM cultivation lead to an altered morphology of the target cells within the whole cultures, with the CM(HT-1080) cells acquiring a less branched shape and the CM(HeLa) cells a more rounded and flattened phenotype with distinctive membrane ruffling at the cell edges. However, extended deprivation of the cancer-derived factors in CM did not revert the phenotype of the CM cells, indicating irreversibility of this alteration. The level of type XIII collagen protein expression was markedly higher in the CM cells than in the original ancestor cells, and it surpassed expression of the cancer cells. Based on Northern blotting, this upregulation originated at a transcriptional level. This change also affected the expression of other proteins, including various cell-matrix and cell-cell binding receptors, cytoskeletal and matrix proteins. Interestingly, the expression of α-SMA, the marker protein for the myofibroblasts, was also upregulated and followed the same temporal expression profile as type XIII collagen. This upregulation suggested a myofibroblast-like transition. Both the CM cells showed a higher proliferation rate and more effective shedding of ectodomain than their ancestors.

5.6.2 Role of TGF-β1 on type XIII collagen expression (III)

As TGF-β1 has been suggested as having a role in the fibroblast-myofibroblast transition and in the upregulation of collagen expression, the effect of TGF-β1 on the regulation of type XIII collagen expression was assessed in primary fibroblasts. This was compared to the effect of PMA. After being rendered quiescent by serum starvation, the cells were treated with TGF-β1 or PMA for 24 hours. mRNA analysis with Northern blotting, normalized to β-actin, showed that TGF-β1 induced type XIII collagen mRNA expression in a concentration-dependent manner up to about 1.8-fold, whereas the induction by PMA was more moderate (about 1.3-fold). The role of TGF-β1 in the phenotypic transition during the CM treatment was assayed by inhibiting the TGF-β1 function with a blocking antibody. In the absence of it, CM from HT-1080 cells stimulated the change of cellular morphology and type XIII collagen expression, whereas when present, the blocking antibody prevented the induction of type XIII collagen expression and the concomitant morphological change of cells.

6 Discussion

An increasing number of membrane-bound proteins are known to liberate their extracellular domains from cell membranes, either to the pericellular space of the ECM or to bodily fluids. The diversity of such proteins is large, comprising most functional classes of the membrane proteins. As for the biological function of ectodomain shedding, instead of merely ridding cell membranes of proteins expired of their purpose, shedding may also represent a means of multiplying the functional roles of the integral membrane proteins for a given cell. The release may result in exposure of new functional epitopes with novel biological activity not possessed by the ancestors. On the other hand, liberation of ectodomains may modulate the surrounding milieu greatly. Both of these factors can subsequently drastically alter cell behaviour. (Ehlers & Riordan 1991, Werb 1997, Werb & Yan 1998.) As an example, Giannelli and co-workers showed that the specific cleavage of a basement membrane protein laminin-5 by MMP-2 in breast cancer cells exposes a cryptic promigratory site, which triggers cell motility (Giannelli et al. 1997). Endo and co-workers reported that the cleavage of the syndecan-1 ectodomain by MMP-1 results in a stimulation of cell migration (Endo et al. 2003). Lately, it has become evident that ectodomain shedding also occurs with the transmembrane collagens (Schäcke et al. 1998, Peltonen et al. 1999, Snellman et al. 2000a, Hashimoto et al. 2002, Banyard et al. 2003). The released ectodomains of the transmembrane collagens have shown interaction capability with various ECM proteins, like fibronectin, nidogen-2, perlecan, heparin and the fibrillar β-amyloid peptide, while such properties have not so far been characterized for their uncleaved membrane-bound variants (Hashimoto et al. 2002, Tu et al. 2002, Banyard et al. 2003). Also, in some cases the shed ectodomain has affected cell migration (Franzke et al. 2002), but as a whole, the function and/or the resulting biological implications of the ectodomain shedding of the transmembrane collagens are currently not fully known.

Based on inhibitor experiments and with the transfected furin-null LoVo cells (Bennett et al. 2000, Ebert et al. 2000), the family of the mammalian Kex/subtilisin-like proprotein convertases was shown to be responsible for ectodomain shedding of type XIII collagen. In addition to type XIII collagen, the involvement of furin in ectodomain shedding has been shown for collagens type XXIII and XXV, whereas for type XVII collagen TACE, ADAM-9 and ADAM-10 are responsible for liberating the ectodomain (Franzke et al. 2002, Hashimoto et al. 2002, Banyard et al. 2003). No other proteases were shown to be

64

involved in the shedding of ectodomain of type XIII collagen. The cell surface biotinylation experiments with the successful abolition of ectodomain shedding by decRVKRcmk showed that the ectodomain is liberated at the cell surface. However, the trans-Golgi network, being a predominant site for the PCs (Thomas 2002), was also implicated in the intracellular processing of this collagen. This was based upon a finding whereby perturbation of the function of the trans-Golgi network by brefeldin A and monensin, or suboptimal intracellular Ca2+ levels brought about by A23187 (Codony-Servat et al. 1999, Messageo et al. 2003) compromised ectodomain shedding. Such a finding would pose no contradiction to the known cellular locations of the PCs because, although mainly in the Golgi, for instance furin and PC7 are continuously reshuffled between the trans-Golgi network and the cell surface (Nakayama 1997, Thomas 2002). Based on the studies concerning cleavage at the cell surface, the cells show a tendency to maintain the plasma membrane level of type XIII collagen content balanced despite growing cellular amounts.

This leads to the question of the function of ectodomain shedding of type XIII collagen. The fact that cells prefer to maintain an even plasma membrane level of type XIII collagen with the excess being expelled suggests that the preservation of the cell surface level of this collagen may serve an important functional purpose. On the other hand, if the ECM becomes saturated with a wealth of the released soluble ectodomain with binding capacity to various matrix proteins, this may result in an altered environment and may influence some other phenomenon. Ectodomain liberation into the ECM may take place as a result of a physiological, controlled and timely response to a given stimulus experienced by the cell, or maybe due to mere unloading of type XIII collagen that has outlived its purpose or that has been synthesized excessively due to derailed control. One could further theorize that the full-length membrane-bound type XIII collagen may show a dualistic phenotype with the membrane-bound and the soluble isoforms, both functioning actively under specific conditions, or that this collagen, when membrane-bound, could represent a precursor of the cleaved functional type XIII collagen ectodomain. Detaching cells from the underlying substrata was found to enhance shedding of the ectodomain. This is noteworthy as cell detachment could represent a regulated, carefully timed stimulus for the cells to liberate the extracellular portion of type XIII collagen, for instance during repeated cell attachment-detachment cycles in cell migration. Interestingly, cell motility has often been found to be altered in conjunction with ectodomain shedding. This was seen with laminin-5 and syndecan-1 (Giannelli et al. 1997, Endo et al. 2003). The cleaved ectodomain of type XVII collagen has been shown to decrease cell motility (Franzke et al. 2002), and also here, the ectodomain hampered cell migration. Considering the known cellular location of the intact type XIII collagen (Peltonen et al. 1999, Hägg et al. 2001, Sund et al. 2001b), the biological context of the shed ectodomain can be surmised to be related to adhesive-dependent cell functions. Supporting this hypothesis, the effects of the shed ectodomain were found not only anti-migratory, but also anti-adhesive, anti-proliferative and non-supportive of cell spreading. Remarkably, these effects of the ectodomain were highly matrix-specific, being discernible only on the vitronectin surface. Vitronectin is one of the major serum-derived proteins recognized to support cell adhesion, spreading, migration and proliferation (Stenn et al., 1983, Basara et al. 1985, Hayman et al. 1985, Schwartz et al. 1999). Both vitronectin and fibronectin contain collagen-binding domains and an RGD/cell adhesion

65

domain, which flank one another in vitronectin but lie further apart in fibronectin (Romberger 1997, Schwartz et al. 1999). As cells use the αv integrins to attach to vitronectin via the RGD sequence (Suzuki et al. 1984, Suzuki et al. 1985), consequently the steric masking of the RGD/cell adhesion domain by ectodomain binding to the juxtaposed collagen-binding domain might be able to reduce availability of the necessary integrin binding sites, and might eventually distract integrin-mediated cell adhesion and other adhesion-dependent cell functions. (Kjøller et al. 1997, Palecek et al. 1997.) The window of opportunity for the ectodomain influence apparently involves only the very initial stages of cell attachment, during which the cells also exhibited transiently defective spreading.

Tu and co-workers have earlier demonstrated that the ectodomain of type XIII collagen can interact with fibronectin in a solid phase assay (Tu et al. 2002). A new follow-up scenario emanated from this finding, as in that the ectodomain could bind to the fibrillar fibronectin matrix in cell culture conditions and possibly affect the assembly of the fibronectin matrix. Also, others have described interaction between various other collagens and fibronectin, mutual binding sites have been characterized, and there is data to suggest reciprocity and interdependence between the assemblies of the fibronectin and collagen fibrillar networks. (Engvall & Ruoslahti 1977, Balian et al. 1980, Kleinman et al. 1978, Ruoslahti et al. 1981, McDonald et al. 1987, Dzamba et al. 1993, Romberger 1997, Pankov & Yamada 2002, Sottile & Hocking 2002, Velling et al. 2002.) On one hand, the correct incorporation of collagens has been shown to be reflected in the structural and functional integrity of the fibronectin matrix. It has been demonstrated that a lack of synthesis of type I collagen or expression of a mutated variant of it, reduced synthesis and secretion of type VI collagen, or expression of mutated collagens type III and V have all resulted in abnormal assembly of the fibronectin matrix in fibroblasts (Dzamba et al. 1993, Lebbe et al. 1997, Sabatelli et al. 2001, Zoppi et al. 2004). On the other hand, the correct polymerization of collagens type I and III into the ECM was recently shown to be dependent on intact preformed fibrillar fibronectin (Sottile & Hocking 2002, Velling et al. 2002). The recombinant exogenous ectodomain and the endogenously released cell-derived ectodomain of type XIII collagen were shown to be able to bind to fibrillar fibronectin. Consistent with earlier data (Engvall & Ruoslahti 1977, Kleinman et al. 1978, Balian et al. 1980, Ruoslahti et al. 1981, McDonald et al. 1987, Dzamba et al. 1993, Romberger 1997, Pankov & Yamada 2002), the N-terminal domain(s) of fibronectin were corroborated as the collagen binding unit. As for type XIII collagen, the conserved C-terminal NC4 domain and its adjoining collagenous domain(s) were implicated in this interaction. The precise detailed molecular-level mechanism of the fibronectin-collagen interaction is still incompletely understood (Steffensen et al. 2002), as is still the case in this particular interaction as well. However, it was demonstrated that the ectodomain of type XIII collagen may actually interfere with the assembly of the fibronectin matrix in a way that results in alterations in the amount, yet not the structure, of the matrix. The results presented here are thus in line with the earlier observations by others in that various soluble collagens in the ECM, now also including type XIII collagen, can affect the incorporation of the fibronectin matrix assembly.

Very little information has existed so far about the expression of type XIII collagen in various human diseases. Such a survey was thus initiated, and cancer was chosen as a starting point. Several epithelial and mesenchymal tumours accompanied by their normal

66

and pre-neoplastic, i.e. dysplastic, equivalents were screened using the TMA methodology and in situ hybridization. The main findings of the cancer TMA survey on the expression of type XIII collagen mRNA can be summarized as follows: i) the expression of type XIII collagen is clearly upregulated in cancer, and it is consistently stimulated particularly in the stromal compartment of carcinomas, ii) it is step-wise upregulated during malignant transformation, i.e. it follows an ascending order of normal < dysplastic < cancer stroma, iii) the epithelial expression is diminutive when compared to that in the stroma, iv) the induction of the expression occurs at an early stage during tumour progression in response to malignant transformation in the epithelia, v) the expression in cancer reverts to an embryonal pattern, and vi) the expression in tissues organizes in a gradient-like pattern, with the expression strongest next to the transformed tumour foci, suggesting paracrine stimulus by the transformed cells. When compared to other collagens, a similar stromal desmoplasia has been described for several other collagens, including other transmembrane collagens (Brown et al. 1999, Powell et al. 1999, Bode et al. 2000, Burns-Cox et al. 2001, Fischer et al. 2001, Parikka et al. 2001, Theret et al. 2001, Banyard et al. 2003, Parikka et al. 2003). Using immunohistochemistry, type XIII collagen expression was found to be somewhat upregulated in the invasive fronts of tumours. As this was also the case in the invasive cytotrophoblasts of the placenta, the expression upregulation in the invasive fronts is likely to be linked to the invasiveness of the cells or tissues rather than to malignancy only.

Cancer cells are known to induce an irreversible phenotypic change in stromal fibroblasts to myofibroblasts via the secretion of soluble factors (Ronnov-Jensen & Petersen 1993, Powell et al. 1999, Valenti et al. 2001, Sung & Chung 2002). Myofibroblasts appear to be responsible for the creation of the stromal desmoplasia (Faouzi et al. 1999, Petrov et al. 2002, Tuxhorn et al. 2002a). In the cell culture model established to study this, primary fibroblasts eventually differentiated into morphologically stable cells, which exhibited not only an altered expression of a number of matrix proteins, an enhanced proliferation rate and a clearly upregulated expression of type XIII collagen, but also changes attributable to a myofibroblast-like transdifferentiation, based on the expression of the myofibroblast marker proteins α-SMA and vimentin (Powell et al. 1999, Serini & Gabbiani 1999, Eyden et al. 2001, Tomasek et al. 2002). The results showed that cancer cells can profoundly influence normal stromal cells in a paracrine manner so as to induce morphological changes and to upregulate type XIII collagen expression. Many earlier studies have suggested a central role for TGF-β1 in the transdifferentation of fibroblasts to α-SMA-positive myofibroblasts and in the induction of desmoplasia (Ignotz et al. 1987, Kähäri et al. 1991, Ronnov-Jensen & Petersen 1993, Frazer et al. 1994, Powell et al. 1999, Serini & Gabbiani 1999, Tomasek et al. 2002, Tuxhorn et al. 2002a). Here the role of TGF-β1 as an inducer of type XIII collagen expression and the transition of the cellular phenotype was verified by the tumour cell-derived CM experiment with function blocking antibodies to TGF-β1. The results shown here imply that at least TGF-β1 can increase type XIII collagen expression during desmoplastic stroma formation and that the expression of this collagen can rapidly respond to such growth factor stimulation.

The tumour stroma is recognized to be strikingly different from a normal stroma (Seljelid et al. 1999, Sung & Chung 2002, Tuxhorn et al. 2002a, Quaranta & Giannelli

67

2003, Bhowmick et al. 2004a, Weaver & Gilbert 2004). The upregulation of type XIII collagen expression in the myofibroblast-like cells was shown to be accompanied by a concomitantly enhanced shedding of the ectodomain. This was shown to have distinct repercussions to the composition and dynamics of the tumour stroma, as the ectodomain was found to i) alter the binding characteristics of vitronectin-rich surroundings, and ii) influence the fibronectin matrix assembly in a manner that affects its amount. These influences may be of biological significance when contributing to the many other changes of the altered tumour stroma. The invasive fronts releasing the type XIII collagen ectodomain at their surfaces can contribute to the development of an ectodomain-rich microenvironment, which is unfavourable and non-supportive for cell adhesion, spreading and migration, and which can thus expedite the dissemination or metastasis of cancer cells. In line with the tissue microenvironment concept of malignancy, aberrant expression of the matrix components, here type XIII collagen, may critically affect the progression of a neoplastic agenda in tissues.

In summary, mammalian cells effectively shed type XIII collagen ectodomain. This was shown to happen at the cell surface, but the involvement of the Golgi apparatus is also suggested. Various intracellular mechanisms can be involved in the regulation of the phenomenon. The release of the ectodomain creates a biologically active soluble molecule with distinct matrix-specific effects, particularly on vitronectin, on adhesion-dependent cell functions. Also, the soluble ectodomain is capable of influencing the fibronectin matrix assembly. By these effects, the soluble ectodomain of type XIII collagen is bound to remodel the surrounding pericellular ECM. This may be of consequence in the case of cancer, where the expression of type XIII collagen and the shedding of its ectodomain are upregulated, especially in the stroma of epithelial cancers, and also, in the invasive fronts of the tumour mass. Consequently, the involvement of type XIII collagen in the remodelling of the stroma can reciprocally affect cell behaviour. These results concerning type XIII collagen broaden the present knowledge of the function and role of the transmembrane collagens. Not only do the results conform well to the established concept of ongoing exchange of information between cells and the ECM, and support the idea of the release of proteolytic fragments possessing cryptic biological activity, they also imply that the role of the transmembrane collagens may be of a more regulatory and remodelling nature than what has hitherto been recognized.

7 Future perspectives

The studies of this thesis have presented new information about the basic cell biologic characteristic of type XIII collagen as well as about the role of this collagen in one human disease group, cancer. Additionally, the findings in this thesis open up new topics, some of which could be summoned as:

1. Related to ectodomain shedding, the in-depth characterization of the participation of the Golgi apparatus in ectodomain shedding is of fundamental interest. Also, the circumstances and the extent to which the trans-Golgi network participates in ectodomain shedding are currently not known. There may be variation in this related to tissues, developmental stage, disease status etc. In general, determination of the extent of ectodomain cleavage taking place at the cell surface vs. intracellularly in the Golgi, and the intracellular transport routes of the cleaved ectodomain are questions to be explored.

2. Related to the influence of the cleaved ectodomain on the ECM, the detailed mechanism of the ectodomain interference on the fibronectin matrix assembly is not fully known. Also, the mechanism of its influence on vitronectin and maybe on other ECM proteins could be characterized.

3. Related to the cancer findings, the roles of the soluble ectodomain vs. membrane-bound type XIII collagen in cancer, particularly in the tumour stroma, and in the invasive fronts of cancer should be studied further. It would be of interest to determine if type XIII collagen or its soluble ectodomain could be used as early markers for malignant transformation, and if the expression of type XIII collagen in cancer is reflected upon patient outcome. The role of this collagen in cancer could be studied further by using various knock-out or over-expression mouse models for type XIII collagen, or showing expression of mutated variants of this collagen. Many proteolytically released fragments of collagens have shown anti-angiogenic features; it would thus be of interest to study if the ectodomain of type XIII collagen possesses such qualities.

References

Aota S, Nomizu M & Yamada KM (1994) The short amino acid sequence Pro-His-Ser-Arg-Asn in human fibronectin enhances cell-adhesive function. J Biol Chem 269: 24756-24761.

Amatangelo MD, Bassi DE, Klein-Szanto AJ & Cukierman E (2005) Stroma-derived three-dimensional matrices are necessary and sufficient to promote desmoplastic differentation of normal fibroblasts. Am J Path 167: 475-488.

Areida SK, Reinhardt DP, Müller PK, Fietzek PP, Köwitz J, Marinkovich MP & Notbohm H (2001) Properties of the collagen type XVII ectodomain. J Biol Chem 276: 1594-1601.

Armstrong PB & Armstrong MT (2000) Intracellular invasion and the organizational stability of tissues: a role for fibronectin. Biochim Biophys Acta 1470: O9-O20.

Arnold JT, Lessey BA, Seppälä M & Kaufman DG (2002) Effect of normal endometrial stroma on growth and differentiation in Ishikawa endometrial adenocarcinoma cells. Cancer Res 62: 79-88.

Arribas J, Coodly L, Vollmer P, Kishimoto TK, Rose-John S & Massagué J (1996) Diverse cell surface protein ectodomains are shed by a system sensitive to metalloprotease inhibitors. J Biol Chem 271: 11376-11382.

Bachmeier BE, Nerlich AG, Boukamp P, Lichtinghagen R, Tschesche H, Fritz H & Fink E (2000) Human keratinocyte cell lines differ in the expression of the collagenolytic matrix metalloproteinases-1,-8, and -13 and of TIMP-1. Biol Chem 381: 509-516.

Baldwin TA & Ostergaard HL (2002) The protein-tyrosine phosphatase CD-45 reaches the cell surface via Golgi-dependent and –independent pathways. J Biol Chem 277: 50333-50340.

Balian G, Click EM & Bornstein P (1980) Location of a collagen-binding domain in fibronectin. J Biol Chem 255: 3234-3236.

Banyard J, Bao L & Zetter BR (2003) Type XXIII collagen, a new transmembrane collagen identified in metastatic tumor cells. J Biol Chem 278: 20989-20994.

Basara ML, McCarthy JB, Barnes DW & Furcht LT (1985) Stimulation of haptotaxis and migration of tumor cells by serum spreading factor. Cancer Res. 45: 2487-2494.

Beer S, Olezewski M, Gutwein P, Geiger C & Altevogt P (1999) Metalloproteinase-mediated release of the ectodomain of L1 adhesion molecule. J Cell Sci 112: 2667-2675.

Behonick DJ & Werb Z (2003) A bit of give and take: the relationship between the extracellular matrix and the developing chondrocytes. Mech Dev 120: 1327-1336.

Bennett BD, Denis P, Haniu M, Teplow DB, Kahn S, Louis JC, Citron M & Vassar R (2000) A furin-like convertase mediates propeptide cleavage of BACE, the Alzheimer´s β-secretase. J Biol Chem 275: 37712-37717.

70

van Bergen en Henegouwen PMP (1989) Immunogold labelling of ultrathin cryosections. In: Hayat MA (ed) Colloidal Gold: Principles, methods, and applications 1: 191-216. Academic Press, San Diego.

Berking C, Takemoto R, Schaider H, Showe L, Satyamoorthy K, Robbins P & Herlyn M (2001) Transforming growth factor β1 increases survival of human melanoma through stroma remodelling. Cancer Res 61: 8306-8316.

Bhowmick NA, Neilson EG & Moses HL (2004a) Stromal fibroblasts in cancer initiation and progression. Nature 432: 332-337.

Bhowmick NA, Chytil A, Plieth D, Gorska AE, Dumont N, Shappell S, Washington MK, Neilson EG & Moses HL (2004b) TGF-β signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303: 848-851.

Bissell MJ and Radisky D (2001) Putting tumors in context. Nat Rev 1: 46-54. Blobel CP (2000) Remarkable roles of proteolysis on and beyond the cell surface. Curr Op Cell

Biol 12: 606-612. Bode MK, Karttunen TJ, Mäkelä J, Risteli L & Risteli J (2000) Type I and III collagens in human

colon cancer and diverticulosis. Scand J Gastroenterol 35: 747-752. Boot-Handford RP, Tuckwell DS, Plumb DA, Farrington Rock C & Poulsom R (2003) A novel and

highly conserved collagen (proα1(XXVII)) with a unique expression pattern and unusual molecular characteristics establishes a new clade within the vertebrate fibrillar collagen family. J Biol Chem 278: 31067-31077.

Bridle KR, Crawford DH, Powell LW & Ramm GA (2001) Role of myofibroblasts in tumor encapsulation of hepatocellular carcinoma in haemochromatosis. Liver 21: 96-104.

Brodsky B & Shah NK (1995) The triple-helix motif in proteins. Protein Motifs 8: 1537-1546. Brown DA & London E (1998) Functions of lipid rafts in biological membranes. Annu Rev Cell

Dev Biol 14: 111-136. Brown DA & London E (2000) Structure and function of sphingolipid- and cholesterol-rich

membrane rafts. J Biol Chem 275: 17221-17224. Brown LF, Guidi AJ, Schnitt SJ, van de Water L, Iruela-Arispe ML, Yeo TK, Tognazzi K &

Dvorak HF (1999) Vascular stroma formation in carcinoma in situ, invasive carcinoma, and metastatic carcinoma of the breast. Clin Cancer Res 5: 1041-1056.

Burns-Cox N, Avery NC, Gingell JC & Bailey AJ (2001) Changes in collagen metabolism in prostate cancer: a host response that may alter progression. J Urol 166: 1698-1701.

Canty EG, Lu Y, Meadows RS, Shaw MK, Holmes DF & Kadler KE (2004) Coalignment of plasma membrane channels and protrusions (fibripositors) specifies the parallelism of tendon. J Cell Biol 165: 553-563.

Canty EG & Kadler KE (2005) Procollagen trafficking, processing and fibrillogenesis. J Cell Sci 118: 1341-1353.

Chen H & Mosher DF (1996) Formation of sodium dodecyl-stable fibronectin multimers. J Biol Chem 271: 9084-9089.

Chernousov MA, Stahl RC & Carey DJ (1998) Schwann cells use a novel collagen-dependent mechanism for fibronectin fibril assembly. J Cell Sci 111: 2763-2777.

Clark K, Pankov R, Travis MA, Askari JA, Mould AP, Craig SE, Newham P, Yamada KM & Humphries MJ (2005) A specific α5β1-integrin conformation promotes directional integrin translocation and fibronectin matrix formation. J Cell Sci 118: 291-300.

Codony-Servat J, Albanelli J, Lopez-Talavera JC, Arribas J & Baselga J (1999) Cleavage of the HER2 ectodomain is a pervanadate-activable process that is inhibited by the tissue inhibitor of metalloproteases-1 in breast cancer cells. Cancer Res 59: 1196-1202.

Colombi M, Zoppi N, De Petro G, Marchina E, Gardella R, Tavian D, Ferraboli S & Barlati S (2003) Matrix assembly induction and cell migration and invasion inhibition by a 13-amino acid fibronectin peptide. J Biol Chem 278: 14346-14355.

71

Colorado PC, Torre A, Kamphaus G, Maeshima Y, Hopfer H, Takahashi K, Volk R, Zamborsky ED, Herman S, Sarkar PK, Ericksen MB, Dhanabal M, Simons M, Post M, Kufe DW, Weichselbaum RR, Sukhatme VP & Kalluri R (2000) Anti-angiogenic cues from vascular basement membrane collagen. Cancer Res 60: 2520-2526.

Danen EHJ, Aota S, van Kraats AA, Yamada KM, Ruiter DJ & van Muijen GNP (1995) Requirement for the synergy site for cell adhesion to fibronectin depends on the activation state of integrin α5β1. J Biol Chem 270: 21612-21618.

Dang D, Yang Y, Li X, Atakilit A, Regezi J, Eisele D, Ellis D & Ramos DM (2004) Matrix metalloproteinases and TGFβ1 modulate oral tumor cell matrix. Biochem Biophys Res Comm 316: 937-942.

Diaz LA, Ratrie III H, Saunders WS, Futamura S, Squiquera HL, Anhalt GJ & Giudice GJ (1990) Isolation of a human epidermal cDNA corresponding to the 180-kDa autoantigen recognized by bullous pemhpigoid and herpes gestationis sera. J Clin Invest 86: 1088-1094.

Dzamba BJ, Wu H, Jaenisch R & Peters DM (1993) Fibronectin binding site in type I collagen regulates fibril formation. J Cell Biol 121: 1165-1172.

Ebert AD, Wechselberger C, Nees M, Clair T, Schaller G, Martinez-Lacaci I, Wallace-Jones B, Bianco C, Weitzel HK & Salmon DS (2000) Cripto-1-induced increase in vimentin expression is associated with enhanced migration of human Caski cervical carcinoma cells. Exp Cell Res 257: 223-229.

Ehlers MRW & Riordan JF (1991) Membrane proteins with soluble counterparts: role of proteolysis in the relases of transmembrane proteins. Biochem 30: 10065-10077.

Endo K, Takino T, Miyamori H, Kinsen H, Yoshizaki T, Furukawa M & Sato H (2003) Cleavage of syndecan-1 by membrane type metalloproteinase-1 stimulates cell migration. J Biol Chem 278: 40764-40770.

Endres K, Anders A, Kojro E, Gilbert S, Fahrenholz F & Postina R (2003) Tumor necrosis factor-α converting enzyme is processed by proprotein-convertases to its mature form which is degraded upon phorbol ester stimulation. Eur J Biochem 270: 2386-2393.

Engvall E & Ruoslahti E (1977) Binding of soluble form of fibroblast surface protein, fibronectin, to collagen. Int J Cancer 20: 1-5.

Exposito JY, Cluzel C, Garrone R & Lethias C (2002) Evolution of collagens. Anat Rec 268: 302-316.

Eyden B (2001) The myofibroblast: an assessment of controversial issues and a definition useful in diagnosis and research. Ultrastruct Pathol 25: 39-50.

Faouzi S, Le Bail B, Neaud V, Boussarie L, Saric J, Bioulac-Sage P, Balabaud C & Rosenbaum J (1999) Myofibroblasts are responsible for collagen synthesis in the stroma of human hepatocellular carcinoma: an in vivo and in vitro study. J Hepatol 30: 275-284.

Fischer H, Stenling R, Rubio C & Lindblom A (2001) Colorectal carcinogenesis is associated with stromal expression of COL11A1 and COL5A5. Carcinogen 22: 875-878.

Fitzgerald J & Bateman JF (2001) A new FACIT of the collagen family: COL21A1. FEBS Lett 505: 275-280.

Franzke CW, Tasanen K, Schäcke H, Zhou Z, Tryggvason K, Mauch C, Zigrino P, Sunnarborg S, Lee DC, Fahrenholz F & Bruckner-Tuderman L (2002) Transmembrane collagen XVII, an epithelial adhesion protein, is shed from the cell surface by ADAMs. EMBO J 21: 5026-5035.

Franzke CW, Tasanen K, Schumann H & Bruckner-Tuderman L (2003) Collagenous transmembrane proteins: collagen XVII as a prototype. Matrix Biol 22: 299-309.

Franzke CW, Tasanen K, Borradori L, Huotari V & Bruckner-Tuderman L (2004) Shedding of collagen XVII/BP180: structural motifs influence cleavage from cell surface. J Biol Chem 279: 24521-24529.

Franzke CW, Bruckner P & Bruckner-Tuderman L (2005) Collagenous transmembrane proteins: recent insights into biology and pathology. J Biol Chem 280: 4005-4008.

72

Frazer A, Bunning RA, Thavarajah M, Seid JM & Russell RG (1994) Studies on type II collagen and aggrecan production in human articular chondrocytes in vitro and effects of transforming growth factor-β and interleukin-1β. Osteoarthritis Cartilage 2: 235-245.

Garten W, Hallenberger S, Ortmann D, Schäfer W, Vey M, Angliker H, Shaw E & Klenk HD (1994) Processing of viral glycoproteins by the subtilisin-like endoprotease furin and its inhibition by specific peptidylchloroalkylketones. Biochim 76: 217-225.

Geiger B, Berhadsky A, Pankov R & Yamada KM (2001) Transmembrane extracellular matrix-cytoskeleton crosstalk. Nat Rev Mol Cell Biol 2: 793-805.

Gelse K, Pöschl E & Aigner T (2003) Collagens – structure, function and biosynthesis. Adv Drug Deliv Rev 55: 1531-1546.

Giannelli G, Falk-Marzillier J, Schiraldi O, Stetler-Stevenson WG & Quaranta V (1997) Induction of cell migration by matrix metalloprotease-2 cleavage of laminin-5. Science 277: 225-228.

Gutwein P, Olezewski M, Mechtersheimer S, Agmon-Levin N, Krauss K & Altevogt P (2000) Role of Src kinases in the ADAM-mediated release of L1 adhesion molecule from human tumor cells. J Biol Chem 275: 15490-15497.

Harlow E & Lane D (1998) Antibodies: a laboratory manual. Cold Spring Harbor Press, Cold Sping Harbor.

Hashimoto T, Wakayabashi T, Watanabe A, Kowa H, Hosoda R, Nakamura A, Kanazawa I, Arai T, Takio K, Mann DMA & Iwatsubo T (2002) CLAC: a novel Alzheimer amyloid plaque component derived from a transmembrane precursor, CLAC-P/collagen type XXV. EMBO J 21: 1524-1534.

Hayman EG, Pierschbacher MD, Suzuki S & Ruoslahti E (1985) Vitronectin – a major cell attachment promoting protein in fetal bovine serum. Exp Cell Res 160: 245-258.

Hirako Y, Usukura J, Uematsu J, Hashimoto T, Kitajima Y & Owaribe K (1998) Cleavage of BP180, a 180-kDa pemhigoid antigen, yields a 120-kDa collagenous extracellular polypeptide. J Biol Chem 273: 9711-9717.

Hocking DC, Sottile J & McKeown-Longo PJ (1994) Fibronectin´s III-1 module contains a conformation-dependent site for the amino-terminal region of fibronectin. J Biol Chem 269: 19183-19187.

Hocking DC, Smith RK & McKeown-Longo PJ (1996) A novel role for the integrin-binding III-10 module in fibronectin matrix assembly. J Cell Biol 133: 431-444.

Hooper NM, Karran EH & Turner AJ (1997) Membrane protein secretases. Biochem J 321: 265-279.

Hopkinson SB, Riddelle KS & Jones JC (1992) Cytoplasmic domain of the 180-kDa bullous pemphigoid antigen, a hemidesmosomal component: molecular and cell biologic characterization. J Invest Dermatol 99: 264-270.

Hägg P, Rehn M, Huhtala P, Väisänen T, Tamminen M & Pihlajaniemi T (1998) Type XIII collagen is identified as a plasma membrane protein. J Biol Chem 273: 15590-15597.

Hägg P, Väisänen T, Tuomisto A, Rehn M, Tu H, Huhtala P, Eskelinen S & Pihlajaniemi T (2001) Type XIII collagen: a novel cell adhesion component present in a range of cell-matrix adhesions and in the intercalated discs between cardiac muscle cells. Matrix Biol 19: 727-742.

Ignotz RA & Massague J (1986) Transforming growth factor-β stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J Biol Chem 261: 4337-4345.

Ignotz RA, Endo T & Massague J (1987) Regulation of fibronectin and type I collagen mRNA levels by transforming growth factor-β. J Biol Chem 262: 6443-6446.

Iozzo RV (2005) Basement membrane proteoglycans: from cellar to ceiling. Nat Rev Mol Cell Biol 6: 646-656.

73

Ito H, Rucker E, Steplewski A, McAdams E, Brittingham RJ, Alabyeva T & Fertala A (2005) Guilty by association: some collagen II mutants alter the formation of ECM as a result of atypical interaction with fibronectin. J Mol Biol 352: 382-395.

Johansson S, Svineng G, Wennerberg K, Armulik A & Lohikangas L (1997) Fibronectin-integrin interactions. Frontiers in Bioscience 2: d126-146.

Juvonen M & Pihlajaniemi T (1992) Characterization of the spectrum of alternative splicing of α1(XIII) collagen transcripts in HT-1080 cells and calvarial tissue resulted in identification of two previously unidentified alternatively spliced sequences, one previously unidentified exon, and nine new mRNA variants. J Biol Chem 267: 24693-24699.

Juvonen M, Sandberg M & Pihlajaniemi T (1992) Patterns of expression of the six alternatively spliced exons affecting the structures of the COL1 and NC2 domains of the α1(XIII) collagen chain in human tissues and cell lines. J Biol Chem 267: 24700-24707.

Juvonen M, Pihlajaniemi T & Autio-Harmainen H (1993) Location and alternative splicing of type XIII collagen RNA in the early human placenta. Lab Invest 69: 541-551.

Kähäri VM, Peltonen J, Qiu Chen Y & Uitto J (1991) Differential modulation of basement membrane gene expression in human fibrosarcoma HT-1080 cells by transforming growth factor-β1. Enhanced type IV collagen and fibronectin gene expression correlates with altered culture phenotype of the cells. Lab Invest 64: 807-818.

Kammerer RA (1997) Alpha-helical coiled-coil oligomerization domains in extracellular proteins. Matrix Biol 15: 555-565.

Kamphaus GD, Colorado PC, Panka DJ, Hopfer H, Ramchandran R, Torre A, Maeshima Y, Mier JW, Sukhatme VP & Kalluri R (2000) Canstatin, a novel matrix-derived inhibitor of angiogenesis and tumor growth. J Biol Chem 275: 1209-1215.

Kang T, Nagase H & Pei D (2002) Activation of membrane-type matrix metalloproteinase 3 zymogen by the proprotein convertase furin in the trans-Golgi network. Cancer Res. 62: 675-681.

Kassner A, Hansen U, Miosge N, Reinhardt DP, Aigner T, Bruckner-Tuderman L, Bruckner P & Grässel S (2003) Discrete integration of collagen XVI into tissue-specific collagen fibrils or beaded microfibrils. Matrix Biol 22: 131-143.

Khatib AM, Siegfried G, Prat A, Luis J, Chrétien M, Metrakos P & Seidah NG (2001) Inhibition of proprotein convertases is associated with loss of growth and tumorigenicisty of HT-29 human colon carcinoma cells. J Biol Chem 276: 30686-30693.

Kheradmand F & Werb Z (2002) Shedding light on sheddases: role in growth and development. Bioessays 24: 8-12.

Kielty CM, Kwan AP, Holmes DF, Schor SL & Grant ME (1985) Type X collagen, a product of hypertrophic cartilage. Biochem J. 227: 545-554.

Kissin EY, Lemaire R, Korn JH & Lafyatis R (2002) Transforming growth factor β induces fibroblast fibrillin-1 matrix formation. Arthr Rheum 46: 3000-3009.

Kivirikko S, Heinämäki P, Rehn M, Honkanen N, Myers JC & Pihlajaniemi T (1994) Primary structure of the α1 chain of human type XV collagen and exon-intron organization in the 3´ region of the corresponding gene. J Biol Chem 269: 4773-4779.

Kjøller L, Sanlip MK, Kirkegaard T, Rodenburg KW, Ronne E, Goodman SL, Preissner KT, Ossowski L & Andreasen PA (1997) Plasminogen activator inhibitor-1 represses integrin- and vitronectin-mediated cell migration independently of its function as an inhibitor of plasminogen activator. Exp Cell Res 232: 420-429.

Kleinman HK, McGoodwin EB, Martin GR, Klebe RJ, Fietzek PP & Woolley DE (1978) Localization of the binding site for cell attachment in the α1(I) chain of collagen. J Biol Chem 253: 5642-5647.

Kleinman HK, Philp D & Hoffman MP (2003) Role of the extracellular matrix in morphogenesis. Curr Op Biotech 14: 526-532.

74

Klimpel KR, Molloy SS, Thomas G & Leppla SH (1992) Anthrax toxin protective antigen is activated by a cell surface protease with the sequence specificity and catalytic properties of furin. Proc Natl Acad Sci USA 89: 10277-10281.

Koch M, Foley JE, Hahn R, Zhou P, Burgeson RE, Gerecke DR & Gordon MK (2001) α1(XX) collagen, a new member of the collagen subfamily, fibril-associated collagens with interrupted triple helices. J Biol Chem 276: 23120-23126.

Koch M, Laub, F, Zhou P, Hahn RA, Tanaka S, Burgeson RE, Gerecke DR, Ramirez F & Gordon MK (2003) Collagen XXIV, a vertebrate fibrillar collagen with structural features of invertebrate collagens. J Biol Chem 278: 43236-43244.

Koch M, Schulze J, Hansen U, Ashwodt T, Keene DR, Brunken WJ, Burgeson, RE, Bruckner P, Bruckner-Tuderman L (2004) A novel marker of tissue junctions, collagen XXII. J Biol Chem 279: 22514-22521.

Kononen J, Bubendorf L, Kallioniemi A, Bärlund M, Schraml P, Leighton S, Torhorst J, Mihatsch MJ, Sauter G & Kallioniemi OP (1998) Tissue microarray for high-throughput molecular profiling of tumor specimens. Nat Med 4: 844-847.

Kubota K, Okazaki J, Louie O, Kent KC, Liu B (2003) TGF-β stimulates collagen (I) in vasecular smooth muscle cells via short element in the proximal collagen promoter. J Surg Res 109: 43-50.

Kvansakul M, Bogin O, Hohenester E & Yayon A (2003) Crystal structure of the collagen α1(VIII) NC1 trimer. Matrix Biol 22: 145-152.

Kvist AP, Latvanlehto A, Sund M, Horelli-Kuitunen N, Rehn M, Palotie A, Beier DR & Pihlajaniemi T (1999) Complete exon-intron organization and chromosomal localization of the gene for mouse type XIII collagen (Col13a1) and comparison with its human homologue. Matrix Biol. 18: 261-274.

Kvist AP, Latvanlehto A, Sund M, Eklund L, Väisänen T, Hägg P, Sormunen R, Komulainen J, Fässler R & Pihlajaniemi T (2001) Lack of cytosolic and transmembrane domains of type XIII collagen results in progressive myopathy. Am J Path 159: 1581-1592.

Lai EC (2003) Lipid rafts make for slippery platforms. J Cell Biol 162: 365-370. Latvanlehto A, Snellman A, Tu H & Pihlajaniemi T (2003) Type XIII collagen and some other

transmembrane collagens contain two separate coiled-coil motifs, which may function as independent oligomerization domains. J Biol Chem 278: 37590-37599.

Lebbe C, Font J, Bonaventure J, Pichon J, Wantyghem J, Rossi M, Haentjens G, Cohen-Solal L & Aubery M (1997) Altered collagen of human pathological fibroblasts impairs the synthesis of fibronectin. Matrix Biol 15: 503-507.

Lee MH & Murphy G (2004) Matrix metalloproteinases at a glance. J Cell Sci 117: 4015-4016. Li K, Tamai K, Tan EML & Uitto J (1993) Cloning of type XVII collagen. J Biol Chem 268: 8825-

8834. Littlepage LE, Egeblad M & Werb Z (2005) Coevolution of cancer and stromal cellular responses.

Cancer Cell 7: 499-500. Löhr M, Schmidt C, Ringel J, Kluth M, Muller P, Nizze H & Jesnowski R (2001) Transforming

growth factor-β1 induces desmoplasia in an experimental model of human pancreatic carcinoma. Cancer Res 61: 550-555.

Maeshima Y, Colorado PC, Torre A, Holthaus KA, Grunkemeyer JA, Ericksen MB, Hopfer H, Xiao Y, Stillman IE & Kalluri R (2000) Distinct antitumor properties of a type IV collagen domain derived from basement membrane. J Biol Chem 275: 21340-21348.

Mann DM, McKeown-Longo PJ & Millis AJT (1988) Binding of soluble fibronectin and its subsequent incorporation into the extracellular matrix by early and late passage human skin fibroblasts. J Biol Chem 263: 2756-2760.

Mao Y & Schwarzbauer JE (2005) Fibronectin fibrillogenesis, a cell-mediated matrix assembly process. Matrix Biol 24: 389-399.

75

Maquoi E, Noel A, Frankenne F, Angliker H, Murphy G & Foidart JM (1998) Inhibition of matrix metalloproteinase 2 maturation and HT1080 invasiveness by a synthetic furin inhibitor. FEBS Lett 424: 262-266.

McAlinden A, Smith TA, Sandell LJ, Ficheux D, Parry DAD & Hulmes DJS (2003) α-helical coiled-coil oligomerization domains are almost ubiquitous in the collagen superfamily. J Biol Chem 278: 42200-42207.

McDonald JA, Quade BJ, Broeckelmann TJ, LaChance R, Forsman K, Hasegawa E & Akiyama S (1987) Fibronectin´s cell-adhesive domain and an amino-terminal matrix assembly domain participate in its assembly into fibroblast pericellular matrix. J Biol Chem 262: 2957-2967.

McKeown-Longo PJ & Mosher DF (1983) Binding of plasma fibronectin to cell layers of human skin fibroblasts. J Cell Biol 97: 466-472.

McKeown-Longo PJ & Mosher DF (1984) Mechanism of formation of disulphide-bonded multimers of plasma fibronectin in cell layers of cultured human fibroblasts. J Biol Chem 259: 12210-12215.

McKeown-Longo PJ & Mosher DF (1985) Interaction of the 70000-mol-wt amino-terminal fragment of fibronectin with the matrix-assembly receptor of fibroblasts. J Cell Biol 100: 364-374.

Merloz-Suárez A & Arribas J (1998) Mechanisms controlling the shedding of transmembrane molecules. Biochem Soc Trans 27: 243-246.

Messageo F, Salhi S, Eon P & Rossignol JM (2003) Proteolytic processing of the hepatitis B virus e antigen precursor: cleavage at two furin consensus sequences. J Biol Chem 278: 891-895.

Molloy SS, Thomas L, Van Slyke JK, Stenberg PE & Thomas G (1994) Intracellular trafficking and activation of the furin proprotein convertase: localization to the TGN and recycling from the cell surface. EMBO J 13: 18-33.

Mott JD & Werb Z (2004) Regulation of matrix biology by matrix metalloproteinases. Curr Op Cell Biol 16: 558-564.

Moulard M & Decroly E (2000) Maturation of HIV envelope glycoprotein precursor by cellular endoproteases. Biochem Biophys Acta 469: 121-132.

Myers JC, Kivirikko KI, Gordon MK, Dion AS & Pihlajaniemi T (1992) Identification of a previously unknown human collagen chain, α1(XV), characterized by extensive interruptions in the triple-helical region. Proc Natl Acad Sci USA 89: 10144-10148.

Myers JC, Li D, Amenta PS, Clark CC, Nagaswami C & Weisel JW (2003) Type XIX collagen purified from human umbilical cord is characterized by multiple sharp kinks delineating collagenous subdomains and by intermolecular aggregates via globular, disulphide-linked, and heparin-binding amino termini. J Biol Chem 278: 32047-32057.

Myllyharju J & Kivirikko KI (2001) Collagens and collagen-related diseases. Ann Med 33: 7-21. Myllyharju J & Kivirikko KI (2004) Collgens, modifying enzymes and their mutations in humans,

flies and worms. Trends Genet 20: 33-43. Nakayama K (1997) Furin: a mammalian subtilisin/Kex2p-like endoprotease involved in

processing of a wide variety of precursor proteins. Biochem J 327: 625-635. Oberhauser AF, Badilla-Fernendez C, Carrion-Vasquez M & Fernandez JM (2002) The mechanical

hierarchies of fibronectin observed with single-molecule AFM. J Mol Biol 319: 433-447. Oda D, Gown AM, Vande Berg JS & Stern R (1988) The fibroblast-like nature of myofibroblasts.

Exp Mol Pathol 49: 316-329. Oda D, Gown AM, Vande Berg JS & Stern R (1990) Instability of the myofibroblast phenotype in

culture. Exp Mol Pathol 52: 221-234. Olsen BR (1997) Collagen IX. Int J Biochem Cell Biol 29: 555-558. O´Reilly MS, Boehm T, Shing Y, Fukai N, Vasios G, Lane WS, Flynn E, Birkhead JR, Olsen BR

& Folkman J (1997) Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 88: 277-285.

76

Orimo A, Gupta PB, Sgroi DC, Arenzana-Seisdedos F, Delauney T, Naeem R, Carey VJ, Richardson AL & Weinberg RA (2005) Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF-1/CXCL12 secretion. Cell 121: 335-348.

Ortega N & Werb Z (2002) New functional roles for non-collagenous domains of basement membrane collagens. J Cell Sci 115: 4201-4214.

Osada Y, Hashimoto T, Nishimura A, Matsuo Y, Wakayabashi T & Iwatsubo T (2005) CLAC binds to amyloid β peptides through the positively charged amino acid cluster within the collagenous domain 1 and inhibits formation of amyloid fibrils. J Biol Chem 280: 8596-8605.

Pace JM, Corrado M, Missero C & Byers PH (2003) Identification, characterization and expression analysis of a new fibrillar collagen gene, COL27A1. Matrix Biol 22: 3-14.

Pajunen L, Tamminen M, Solomon E & Pihlajaniemi T (1989) Assignment of the gene coding for the α1 chain of collagen type XIII (COL13A1) to human chromosome region 10q11→qter. Cytogen Cell Gen 52: 190-193.

Palecek SP, Loftus JC, Ginsberg MH, Lauffenburger DA & Horwitz A (1997) Integrin-ligand binding properties govern cell migration speed through cell-substratum adhesiveness. Nature 385: 537-540.

Pankov R, Cukierman E, Katz BZ, Matsumoto K, Lin DC, Lin S, Hahn C & Yamada KM (2000) Integrin dynamics and matrix assembly: tensin-dependent translocation of α5β1 integrins promotes early fibronectin fibrillogenesis. J Cell Biol. 148: 1075-1090.

Pankov R & Yamada KM (2002) Fibronectin at a glance. J Cell Sci 115: 3961-3863. Parikka M, Kainulainen T, Tasanen K, Bruckner-Tuderman L, Salo T (2001) Altered expression of

collagen XVII in ameloblastomas and basal cell carcinomas. J Oral Pathol 30: 589-595. Parikka M, Kainulainen T, Tasanen K, Väänänen A, Bruckner-Tuderman L & Salo T (2003)

Alterations of collagen XVII expression during transformation of oral epithelium to dysplasia and carcinoma. J Histochem Cytochem 51: 921-929.

Pei D & Weiss SJ (1995) Furin-dependent intracellular activation of the human stromelysin-3 zymogen. Nature 375: 244-247.

Peltonen S, Rehn M & Pihlajaniemi T (1997) Alternative splicing of mouse α1(XIII) collagen RNAs results in at least 17 different transcripts, predicting α1(XIII) collagen chains with length varying between 651 and 710 amino acid residues. DNA Cell Biol 16: 227-234.

Peltonen S, Hentula M, Hägg P, Ylä-Outinen H, Tuukkanen J, Lakkakorpi J, Rehn M, Pihlajaniemi T & Peltonen J (1999) A novel component of epidermal cell-matrix and cell-cell contacts: transmembrane protein type XIII collagen. J Invest Dermatol 113: 635-642.

Peschon JJ, Slack JL, Reddy P, Stocking KL, Sunnarborg SW, Lee DC, Russell WE, Castner BJ, Johnson RS, Fritzner JN, Boyce RW, Nelson N, Kozlowsky CJ, Wolfson MF, Rauch CT, Cerretti DP, Paxton RJ, March CJ & Black RA (1998) An essential role for ectodomain shedding in mammalian development. Sciense 282: 1281-1284.

Petrov VV, Fagrad RH & Lijnen PJ (2002) Stimulation of collagen production by transforming growth factor-β1 during differentiation of cardiac fibroblasts to myofibroblasts. Hypertension 39: 258-263.

Pihlajaniemi T, Myllylä R, Seyer J, Kurkinen M & Prockop DJ (1987) Partial characterization of a low molecular weight human collagen that undergoes alternative splicing. Proc. Natl. Acad. Sci. USA 84: 940-944.

Pihlajaniemi T & Tamminen M (1990) The α1 chain of type XIII collagen consists of three collagenous and four noncollagenous domains, and its primary transcript undergoes complex alternative splicing. J Biol Chem 265: 16922-16928.

Plow EF, Haas TA, Zhang L, Loftus J & Smith JW (2000) Ligand binding to integrins. J Biol Chem 275: 21785-21788.

77

Powell DW, Mifflin RC, Valentich JD, Crowe SE, Saada JI & West AB (1999) Myofibroblasts. Paracrine cells important in health and sidease. Am J Physiol 277: C1-C19.

Prockop DJ & Kivirikko KI (1995) Collagens: molecular biology, disease and potentials for therapy. Annu Rev Biochem 64: 403-434.

Quaranta V & Giannelli G (2003) Cancer invasion: watch your neighbourhood! Tumori 89: 343-348.

Ramchandran R, Dhanabal M, Volk R, Waterman JF, Segal M, Lu H, Knebelmann B & Sukhatme VP (1999) Antiangiogenic activity of restin, NC10 domain of human collagen XV: comparison to endostatin. Biochem Biophys Res Comm 255: 735-739.

Rehn M & Pihlajaniemi T (1994) α1(XVIII), a collagen chain with frequent interruptions in the collgenous sequence, a distinct tissue distribution, and homology with type XV collagen. Proc Natl Acad Sci USA 91: 4234-4238.

Rehn M, Hintikka E & Pihlajaniemi T (1994) Primary structure of the α1 chain of mouse type XVIII collagen, partial structure of the corresponding gene, and comparison of the α1(XVIII) chain with its homologue, the α1(VX) collagen chain. J Biol Chem 269: 13929-13935.

Rehn M & Pihlajaniemi T (1995) Identification of three N-terminal ends of type XVIII collagen chains and tissue-specific differences in the expression of the corresponding transcripts. J Biol Chem 270: 4705-4711.

van der Rest M & Garrone R (1991) Collagen family of proteins. FASEB J 5: 2814-2823. Romberger DJ (1997) Molecules in focus: Fibronectin. Int J Biochem Cell Biol 29: 939-943. Ronnov-Jensen J & Petersen OW (1993) Induction of α-smooth muscle actin by transforming

growth factor-β1 in quiescent human breast gland fibroblasts. Implications for myofibroblast generation in breast neoplasia. Lab Invest 68: 696-707.

Ruoslahti E, Hayman EG, Engvall E, Cothran WG & Butler WT (1981) Alignment of biologically active domains in the fibronectin molecule. J Biol Chem 256: 7277-7281.

Ruoslahti E & Pierschbacher E (1987) New perspectives in cell adhesion: RGD and integrins. Science 238: 491-497.

Sabatelli P, Bonaldo P, Lattanzi G, Braghetta P, Bergamin N, Capanni C, Mattioli E, Columbaro M, Ognibene A, Pepe G, Bertini E, Merlini L, Maraldi NM & Squarzoni S (2001) Collagen VI deficiency affects the organization of fibronectin in the extracellular matrix of cultured fibroblasts. Matrix Biol 20: 475-486.

Sambrook J, Fritsch EF & Maniatis T (1989) Molecular cloning: a laboratory manual. Cold Spring Harbor Press, Cold Spring Harbor.

Sandberg M, Tamminen M, Hirvonen H, Vuorio E & Pihlajaniemi T (1989) Expression of mRNAs coding for the α1 chain of type XIII collagen in human fetal tissues: comparison with expression of mRNAs for collagen types I, II and II. J Cell Biol 109: 1371-1379.

Sandberg-Lall M, Hägg PO, Wahlström I & Pihlajaniemi T (2000) Type XIII collagen is widely expressed in the adult and developing human eye and accentuated in the ciliary muscle, the optic nerve and the neural retina. Exp Eye Res 70: 401-410.

Sato H, Kinoshita T, Takino T, Nakayama K & Seiki M (1996) Activation of a recombinant membrane type 1-metalloproteinase (MT1-MMP) by furin and its interaction with tissue inhibitor of metalloproteinases (TIMP)-2. FEBS Lett 393: 101-104.

Sato K, Yomogida K, Wada T, Yorihuzi T, Nishimune Y, Hosokawa N & Nagata K (2002) Type XXVI collagen, a new member of the collagen family, is specifically expressed in the testis and ovary. J Biol Chem 277: 37678-37684.

Schumann H, Baetge J, Tasanen K, Wojnarowska F, Schäcke H, Zillikens D & Bruckner-Tuderman L (2000) The shed ectodomain of collagen XVII/BP180 is targeted by autoantibodies in different blistering skin diseases. Am J Pathol 156: 685-695.

Schwartz I, Seger D & Shaltiel S (1999) Molecules in focus: vitronectin. Int J Biochem Cell Biol 31: 539-544.

78

Schwarzbauer JE & Sechler JL (1999) Fibronectin fibrillogenesis: a paradigm for extracellular matrix assembly. Curr Op Cell Biol 11: 622-627.

Schäcke H, Schumann H, Hammami-Hauasli N, Raghunath M & Bruckner-Tuderman L (1998) Two forms of collagen XVII in keratinocytes. J Biol Chem 273: 25937-25943.

Sechler JL, Takada Y & Schwarzbauer JE (1996) Altered rate of fibronectin matrix assembly by deletion of the first type III repeats. J Cell Biol 134: 573-583.

Sechler JL, Corbett SA & Schwarzbauer JE (1997) Modulatory roles for integrin activation and the synergy site of fibronectin during matrix assembly. Mol Biol Cell 8: 2563-2573.

Sechler JL, Rao H, Cuminskey AM, Vega-Colon I, Smith MS, Murata T & Schwarzbauer JE (2001) A novel fibronectin binding site required for fibronectin fibril growth during matrix assembly. J Cell Biol 154: 1081-1088.

Seidah NG & Chrétien M (1997) Eukaryotic protein processing: endoproteolysis of precursor proteins. Curr Op Biotech 8: 602-607.

Seljelid R, Jozefowski S & Sveinbjornsson B (1999) Tumor stroma. Anticancer Res 19: 4809-4822. Serini G & Gabbiani G (1999) Mechanisms of myofibroblast activity and phenotypic modulation.

Exp Cell Res 250: 273-283. Shapiro J, Sciaky N, Lee J, Bosshart H, Angeletti RH & Bonifacino JS (1997) Localization of

endogenous furin in cultured cell lines. J Histochem Cytochem 45: 3-12. Shows TB, Tikka L, Byers MG, Eddy RL, Haley LL, Henry WM, Prockop DJ & Tryggvason K

(1989) Assignment of the human collagen α1(XIII) chain gene (COL13A1) to the q22 region of chromosome 10. Genomics 5: 128-133.

Singer SJ (1990) The structure and insertion of integral proteins in membranes. Annu Rev Cell Biol 6: 247-296.

Snellman A, Tu H, Väisänen T, Kvist AP, Huhtala P & Pihlajaniemi T (2000a) A short sequence in the N-terminal region is required for the trimerization of type XIII collagen and is conserved in other collagenous transmembrane proteins. EMBO J 19: 5051-5059.

Snellman A, Keränen MR, Hägg PO, Lamberg A, Hiltunen KJ, Kivirikko KI & Pihlajaniemi T (2000b) Type XIII collagen forms homotrimers with three triple-helical collagenous domains and its asscociation into disulphide-bonded trimers is enhanced by prolyl 4-hydroxylase. J Biol Chem 275: 8936-8944.

Snellman A & Pihlajaniemi T (2003) Collagenous transmembrane proteins. Rec Devel Biol Chem 2: 1-22.

Sottile J & Hocking DC (2002) Fibronectin polymerization regulates the composition and stability of extracellular matrix fibrils and cell-matrix adhesions. Mol Biol Cell 13: 3546-3559.

Sottile J & Chandler J (2005) Fibronectin matrix turnover occurs through a caveolin-1-dependent process. Mol Biol Cell 16: 757-768.

Steffensen B, Xu X, Martin PA & Zardeneta G (2002) Human fibronectin and MMP-2 collagen binding domains compete for collagen binding sites and modify cellular activation of MMP-2. Matrix Biol 21: 399-414.

Stenn KS, Madri JA, Tinghitella T & Terranova VP (1983) Multiple mechanisms of dissociated epidermal cell spreading. J Cell Biol 96: 63-67.

Sucic JF, Moehring JM, Inocencio NM, Luchini JW & Koehring TJ (1999) Endoprotease PACE4 is Ca2+-dependent and temperature-sensitive and can partly rescue the phenotype of a furin-deficient cell strain. Biochem J 339: 639-647.

Sund M, Väisänen T, Kaukinen T, Ilves M, Tu H, Autio-Harmainen H, Rauvala H & Pihlajaniemi T (2001a) Distinct expression of type XIII collagen in neuronal structures and other tissues during mouse development. Matrix Biol 20: 215-231.

79

Sund M, Ylönen R, Tuomisto A, Sormunen R, Tahkola J, Kvist AP, Kontusaari S, Autio-Harmainen H & Pihlajaniemi T (2001b) Abnormal adherence junctions in the heart and reduced angiogenesis in transgenic mice overexpressing mutant type XIII collagen. EMBO J 20: 5153-5164.

Sung SY & Chung LWK (2002) Prostate tumor-stroma interaction: molecular mechanisms and opportunities for therapeutic targeting. Diff 70: 506-521.

Suzuki S, Pierschbacher MD, Hayman EG, Nguyen K, Ohgren Y & Ruoslahti E (1984) Domain structure of vitronectin: alignment of active sites. J Biol Chem 259: 15307-15314.

Suzuki S, Oldberg A, Hayman EG, Pierschbacher MD & Ruoslahti E (1985) Complete amino acid sequence of human vitronectin deduced from cDNA: similarity of cell attachment sites in vitronectin and fibronectin. EMBO J 10: 2519-2524.

Söderberg L, Kakuyama H, Möller A, Ito A, Winblad B, Tjernberg LO & Näslund J (2005a) Characterization of the Alzheimer´s disease-associated CLAC protein and identification of an amyloid β-peptide-binding site. J Biol Chem 280: 1007-1015.

Söderberg L, Dahlqvist C, Kakuyama H, Thyberg J, Ito A, Winblad B, Näslund J & Tjernberg LO (2005b) Collagenous Alzheimer amyloid plaque component assembles amyloid fibrils into protease resistant aggregates. FEBS J 272: 2231-2236.

Taylor NA, van de Ven WJ & Creemers JWM (2003) Curbing activation: proprotein convertases in homeostasis and pathology. FASEB J 17: 1215-1227.

Theret N, Musso O, Turlin B, Lotrian D, Bioulac-Sage P, Campion JP, Boudjema K & Clement B (2001) Increased extracellular matrix remodelling is associated with tumor progression in human hepatocellular carcinomas. Hepatol 34: 82-88.

Thomas G (2002) Furin at the cutting edge: from protein traffic to embryogenesis and disease. Nat Rev Mol Cell Biol 3: 753-766.

Tikka L, Pihlajaniemi T, Henttu P, Prockop DJ & Tryggvason K (1988) Gene structure for the α1 chain of a human short-chain collagen (type XIII) with alternatively spliced transcripts and translation termination codon at the 5´ end of the last exon. Proc Natl Acad Sci USA 85: 7491-7495.

Tikka L, Elomaa O, Pihlajaniemi T & Tryggvason K (1991) Human α1(XIII) collagen gene. J Biol Chem 266: 17713-17719.

Tomasek JJ, Gabbiani G, Hinz B, Chaponnier C & Brown RA (2002) Myofibroblasts and mechanoregulation of connective tissue remodelling. Nat Rev Mol Cell Biol 3: 349-363.

Tu H, Sasaki T, Snellman A, Göhring W, Pirilä P, Timpl R & Pihlajaniemi T (2002) The type XIII collagen ectodomain is a 150-nm rod and capable of binding to fibronectin, nidogen-2, perlecan, and heparin. J Biol Chem 277: 23092-23099.

Tuckwell D (2002) Identification and analysis of collagen α1(XXI), a novel member of the FACIT collagen family. Matrix Biol 21: 63-66.

Tuxhorn JA, Ayala GE, Smith MJ, Smith VC, Dang TD & Rowley DR (2002a) Reactive stroma in human prostate cancer: induction of myofibroblast phenotype and extracellular matrix remodelling. Clin Cancer Res 8: 2912-2923.

Tuxhorn JA, McAlhany SJ, Dang TD, Ayala GE & Rowley DR (2002b) Stromal cells promote angiogenesis and growth of human prostate tumors in a differential reactive stroma (DRS) xenograft model. Cancer Res. 62: 3298-3307.

Valenti MT, Azzarello G, Balducci E, Sartori S, Sandri M, manconi R, Sicari U, Bari M & Vinante O (2001) Conditioned medium from MCF-7 cell line induces myofibroblast differentiation, decreased cell proliferation, and increased apoptosis in cultured normal fibroblasts but not in fibroblasts from malignant breast tissue. Histochem J 33: 499-509.

Velling T, Risteli J, Wennerberg K, Mosher DF & Johansson S (2002) Polymerization of type I and III collagens is dependent on fibronectin and enhanced by integrins α11β1 and α2β1. J Biol Chem 277: 37377-37381.

80

Vu TH (2001) Don´t mess with the matrix. Nat Gen 28: 202-203. Weaver VM & Gilbert P (2004) Watch thy neighbor: cancer is a communal affair. J Cell Sci 117:

1287-1290. Werb Z (1997) ECM and cell surface proteolysis: regulating cellular ecology. Cell 91: 439-442. Werb Z & Yan Y (1998) A cellular striptease act. Science 282: 1279-1280. Wierbicka-Patynowski I & Schwarzbauer JE (2003) The ins and outs of fibronectin matrix

assembly. J Cell Sci 116: 3269-3276. Wu C, Bauer JS, Juliano RL & McDonald JA (1993) The α5β1 integrin fibronectin receptor, but

not the α5 cytoplasmic domain, functions in an early and essential step in fibronectin matrix assembly. J Biol Chem 268: 21882-21888.

Wu C, Fields AJ, Kapteijn BA & McDonald JA (1995) The role of α4β1 integrin in cell motility and fibronectin matrix assembly. J Cell Sci 108: 821-829.

Wu C, Hughes PE, Ginsberg MH & McDonald JA (1996) Identification of a new biological function for the integrin αvβ3: initiation of fibronectin matrix assembly. Cell Adhes Commun 4: 149-158.

Yamashiro S, Yamakita Y, Ono S & Matsumura F (1998) Fascin, an actin-bundling protein, induces membrane protrusions and increases cell motility of epithelial cells. Mol Biol Cell 9: 993-1006.

Ylönen R, Kyrönlahti T, Sund M, Ilves M, Lehenkari P, Tuukkanen J & Pihlajaniemi T (2005) Type XIII collagen strongly affects bone formation in transgenic mice. Bone Min Res 20: 1381-1393.

Zambon MC (2001) The pathogenesis of influenza in humans. Rev Med Virol 11: 227-241. Zillikens D & Giudice GJ (1999) BP180/type XVII collagen: its role in acquired and inherited

disorders or the dermal-epidermal junction. Arch Dermatol Res 291: 187-194. Zimina E, Bruckner-Tuderman L & Franzke CW (2005) Shedding of collagen XVII ectodomain

depends on plasma membrane microenvironment. J Biol Chem 280: 34019-34024. Zoppi N, Gardella R, De Paepe A, Barlati S & Colombi M (2004) Human fibroblasts with

mutations in COL5A1 and COL3A1 genes do not organize collagens and fibronectin in the extracellular matrix, down-regulate α2β1 integrin, and recruit αvβ3 instead of α5β1 integrin. J Biol Chem 279: 18157-18168.