TYPE XVIII COLLAGEN HARRI Characterization of the primary structure and expression ELAMAA pattern of different variants in Xenopus laevis, characterization of the human gene structure and analysis of transgenic mice Faculty of Medicine, expressing endostatin Department of Medical Biochemistry and Molecular Biology, Biocenter Oulu, University of Oulu

OULU 2004

HARRI ELAMAA

TYPE XVIII COLLAGEN Characterization of the primary structure and expression pattern of different variants in Xenopus laevis, characterization of the human gene structure and analysis of transgenic mice expressing endostatin

Academic Dissertation to be presented with the assent of the Faculty of Medicine, University of Oulu, for public discussion in the Auditorium of the Department of Pharmacology and Toxicology, on December 3rd, 2004, at 1 p.m.

OULUN YLIOPISTO, OULU 2004 Copyright © 2004 University of Oulu, 2004

Supervised by Professor Taina Pihlajaniemi

Reviewed by Docent Klaus Elenius Doctor Takako Sasaki

ISBN 951-42-7568-3 (nid.) ISBN 951-42-7569-1 (PDF) http://herkules.oulu.fi/isbn9514275691/ ISSN 0355-3221 http://herkules.oulu.fi/issn03553221/

OULU UNIVERSITY PRESS OULU 2004 Elamaa, Harri, Type XVIII collagen. Characterization of the primary structure and expression pattern of different variants in Xenopus laevis, characterization of the human gene structure and analysis of transgenic mice expressing endostatin Faculty of Medicine, Department of Medical Biochemistry and Molecular Biology, and Biocenter Oulu, University of Oulu, P.O.Box 5000, FIN-90014 University of Oulu, Finland 2004 Oulu, Finland

Abstract In this work the type XVIII collagen has been studied by using several approaches, such as different animal models. The primary structure of frog, Xenopus laevis, type XVIII collagen and the expression pattern of its variants during early embryogenesis have been elucidated. The gene structure of human type XVIII collagen was characterized and the localization and processing of its longest variant was studied by generated antibodies. In addition, the function of the proteolytically released C-terminal part of type XVIII collagen, endostatin, was studied by generating transgenic mice expressing endostatin. The primary structure of X. laevis type XVIII collagen is comprised of three N-terminal variants resembling their mammalian counterparts. The sizes of the polypeptides are 1285, 1581, and 1886 residues. The most conserved regions are the C-terminal endostatin region and the cysteine-rich domain in the N-terminus. Whole-mount in situ hybridization reveals different expression patterns for variants during embryogenesis. The short variant is the most abundant, whereas the two longest variants exhibit more restricted expression. The gene structure of human type XVIII collagen reveals an exon-intron organization that is conserved with mouse. The length of the human gene is about 105 kb and contains 43 exons. The third variant of type XVIII collagen has a conserved cysteine-rich domain with homology to the extracellular part of frizzled proteins. This third variant is localized to developing muscle and lung, and is also found in serum. In cell culture, the proteolytic fragments of the N-terminus, including the cysteine-rich motif, are also detected. Endostatin function was studied by generating mouse lines expressing endostatin under the keratin-14 promoter, which drives the expression mainly in the skin. Three independent transgenic mouse lines were achieved with varied expression levels. The phenotype was seen in the eye with lens opacity and abnormal morphology of epithelial cells in the lens. In the skin, a broading of the basement membrane in the epidermis dermis junction was detected. Immunoelectron microscopy analysis revealed a polarized orientation of type XVIII collagen in the basement membrane. In transgenic mice, altered localization of endogenous type XVIII collagen was seen, suggesting displacement of the endogenous type XVIII collagen with transgenic endostatin leading to disorganized basement membrane.

Keywords: basement membrane, collagen type XVIII, endostatins, protein isoforms

Acknowledgements

This research was carried out at the Department of Medical Biochemistry and Molecular Biology. I wish to thank my supervisor, Professor Taina Pihlajaniemi, for giving me the opportunity to work in her group and introducing me to the field of matrix biology. I wish to express my appreciation and respect for Research Professor emeritus Kari Kivirikko for creating, together with Professor Taina Pihlajaniemi, excellent conditions for many scientists. I also wish to thank Professor emeritus Ilmo Hassinen, Professor Leena Ala- Kokko, Docent Johanna Myllyharju, and Professor Peppi Karppinen for their contributions to department management. I am very grateful to Docent Klaus Elenius and Dr. Takako Sasaki for their accurate review and valuable comments on my thesis manuscript. Sandra Hänninen, M.Sc. is acknowledged for her careful revision of the language of the manuscript. The collaboration between different laboratories and researchers has been crucial for my thesis. I would like to thank all my co-authors for their invaluable efforts. I wish to express my deepest gratitude to Dr. Olivier Destrée and the members of his research group in the Hubrecht laboratory in Utrecht. Dr. Destrée and his group wished me warmly welcome to their group and introduced me to the frog world during my several visits. I wish to thank Docent Helena Autio-Harmainen for her contribution to this work. I also wish to thank Docent Raija Sormunen and her excellent team for sharing their expertise in ultrastructural analysis. I am grateful for the valuable comments from Docent Raija Soininen, and expert in transgenic technology. Particularly, I am very grateful to Marko Rehn for helping me to take my first steps in the scientific field, and his support and advice during these years. Especially I wish to thank Päivi Tuomaala for her excellent technical assistance, which has been extremely valuable for my thesis. I also wish to thank Outi Laurila, Aira Harju, Sirpa Kellokumpu and Anna-Liisa Oikarinen for their valuable technical assistance. I am also grateful to the people in the laboratory animal center, Liisa Asikkala, Sirpa Tausta and Päivi Moilanen. In addition, special thanks go to Marja-Leena Kivelä and Auli Kinnunen for their kind help in secretarial matters, Seppo Lähdesmäki and Pertti Vuokila their help in a variety of practical matters, and Ari-Pekka Kvist and Risto Helminen for helping me with the computers. I have been privileged to work with pleasant researchers who are willing to offer help and consultation. I would like to thank all members of Taina´s group for their support. I am grateful to Ritva Heljasvaara for her comments on my thesis manuscript. I would especially like to thank my friends Joni Mäki and Lauri Eklund. We have shared a lot during our work and leisure time. I also want to thank all of my colleagues from the old downstairs office, L023, for creating a friendly atmosphere. I owe my deepest gratitude to my parents Maire and Eero for their love and support during these years, and my warmest thanks to my brother Kari and sister Satu and their families for their friendship and many hilarious moments. Above all, I owe my greatest gratitude to my wife Sari for her love and caring, and sharing everyday life. You have encouraged me and reminded me of the existence of intriguing life outside of work. I am deeply indebted to our new family member, the sweet daughter, who came to brighten our family life and cheered up her father during the thesis process. This work was supported by grants from the Emil Altonen Foundation, the Research and Science Foundation of Farmos, and the Finnish Cultural Foundation. Oulu, October 2004 Harri Elamaa Abbreviations

BM basement membrane bp basepairs C- carboxy- COL collagenous domain CRD cysteine-rich domain fz Frizzled domain GAG glycosaminoglycan HSPG heparin sulphate proteoglycan IEM immunoelectron microscopy kb kilobase kDa kilodalton K14 keratin 14 N- amino- NC non-collagenous PAGE polyacrylamide gel electrophoresis PCR polymerase chain reaction RT-PCR reverse transcription polymerase chain reaction SDS sodium dodecyl sulphate VEGF vascular endothelial growth factor

List of original articles

This thesis is based on the following articles, which are referred to in the text by their Roman numerals: I Elamaa H, Peterson J, Pihlajaniemi T & Destrée O (2002) Cloning of three variants of type XVIII collagen and their expression patterns during Xenopus laevis development. Mech Dev 114: 109-13. II Elamaa H, Snellman A, Rehn M, Autio-Harmainen H & Pihlajaniemi T (2003) Characterization of the human type XVIII collagen gene and proteolytic processing and tissue location of the variant containing a frizzled motif. Matrix Biol 22: 427-42. III Elamaa H, Sormunen R, Rehn M, Soininen R & Pihlajaniemi T (2004) Endostatin overexpression specifically in the lens and skin leads to cataract and ultrastructural alterations in basement membranes. Am. J. Pathol., In press.

Contents

Abstract Acknowledgements Abbreviations List of original articles Contents 1 Introduction ...... 13 2 Review of the literature ...... 14 2.1 Cryptic angiogenesis inhibitors ...... 14 2.2 Collagen superfamily...... 15 2.3 Type XVIII collagen...... 19 2.3.1 Primary structure of type XVIII collagen...... 19 2.3.2 Gene structure of the mouse α1 (XVIII) chain ...... 20 2.3.3 Type XVIII collagen is a heparan sulphate proteoglycan ...... 20 2.3.4 Frizzled domain /cysteine-rich domain...... 21 2.3.5 Expression and localization of type XVIII collagen...... 22 2.3.6 Mutations in the type XVIII collagen gene ...... 24 2.3.7 Type XVIII collagen and the eye...... 24 2.4 Endostatin...... 25 2.4.1 Structure of endostatin...... 25 2.4.2 Proteolytic generation of endostatin and inhibition of matrix metalloproteinases by endostatin...... 26 2.4.3 Binding activities of endostatin ...... 26 2.4.4 Intracellular signalling mechanisms ...... 28 2.4.5 Biological activities of monomeric and oligomeric endostatin ...... 30 2.4.6 Insoluble endostatin...... 30 2.4.7 Endostatin and cancer...... 31 2.5 Type XV collagen...... 32 2.6 Xenopus as a model organism...... 33 2.7 Basement membranes...... 34 2.7.1 Assembly of BMs ...... 35 2.8 Development of the eye and cataractogenesis ...... 40 3 Outlines of the present research...... 42 4 Materials and methods...... 43 4.1 Isolation and characterization of cDNA and genomic clones (I, II) ...... 43 4.2 Reverse transcription (RT) and polymerase chain reaction (PCR) (I, II, III)...... 44 4.3 Whole-mount in situ hybridization (I) ...... 44 4.4 Expression of polypeptide fragments in E. coli for antigen production (II) ...... 45 4.5 Preparation and affinity purification of antibodies (II)...... 45 4.6 Immunoprecipitation (II) ...... 46 4.7 SDS-PAGE and Western blot (II, III) ...... 46 4.8 Cell culture, transfections, and immunofluorescence of cells (II) ...... 47 4.9 Generation of transgenic mouse lines (III) ...... 47 4.10 Immunohistology (II, III)...... 48 4.11 Histology and light microscopy (I, II, III) ...... 48 4.12 Electron microscopy (III) ...... 49 4.13 Immunoelectron microscopy (III)...... 49 5 Results ...... 50 5.1 Xenopus laevis type XVIII collagen (I) ...... 50 5.2 The longest variant of human type XVIII collagen (II) ...... 51 5.3 Human type XVIII collagen gene structure (II)...... 52 5.4 Transgenic mice overexpressing endostatin (III)...... 53 5.5 Transgenic endostatin in the lens of the eye (III)...... 53 5.6 Transgenic endostatin in the skin (III) ...... 54 5.7 Polarized orientation of type XVIII collagen in the skin BM (III) ...... 55 6 Discussion ...... 57 6.1 Xenopus laevis type XVIII collagen ...... 57 6.2 Human type XVIII collagen ...... 58 6.3 Transgenic mice overexpressing endostatin ...... 60 7 Future perspectives...... 63 References

1 Introduction

The extracellular matrix is a complex of macromolecules that underlies the epithelia and endothelia and surrounds the connective tissue cells, and it has been the prerequisite for the development of multicellular organisms during evolution. Tissue matrices are typically comprised of collagens, elastic fibers, glycoproteins, and proteoglycans. The collagens are the most abundant proteins of the extracellular matrix, maintaining the architecture of tissues and organs and conferring strength to them. However, they are also involved in other functions, such as organogenesis. Type XV and type XVIII collagen form a subgroup within the collagen superfamily. The C-terminal domain of type XVIII collagen contains the endostatin fragment, shown to inhibit endothelial cell proliferation and migration, and thereby angiogenesis and tumor growth. The highest homology of type XV and XVIII collagen is encompassed in their C-terminal anti-angiogenic portions. This thesis work deals with type XVIII collagen. We have studied the function of endostatin by generating mouse lines overexpressing this fragment. We have also characterized the primary structure and embryonal expression of type XVIII collagen in Xenopus laevis. Moreover, the gene structure of human type XVIII collagen has been characterized, and a variant of type XVIII collagen, previously undetected in humans, has been identified and shown to be released into cell culture medium. 2 Review of the literature

2.1 Cryptic angiogenesis inhibitors

Over the past years, there has been a dramatic increase in the identification of anti- angiogenic fragments from larger proteins with unrelated functions. These cryptic anti- angiogenic molecules are largely derived from matrix components, but also from intracellular proteins like the naturally occurring alternative splice variant of tryptophanyl-tRNA synthetase (TrpRS) and its C-terminal fragment (Wakasugi et al. 2002, Otani et al. 2002). The list of anti-angiogenic factors derived from larger molecules that have functions unrelated to angiogenesis includes angiostatin, a fragment of plasminogen (O´Reilly et al. 1994); PEX, a noncatalytic metalloproteinase fragment that has integrin-binding activity (Brooks et al. 1998); vasostatin, a fragment of calreticulin (Pike et al. 1998); and cleavage products of angiotensinogen and prolactin (Celerier et al. 2002, Ferrara et al. 1991). Several cryptic anti-angiogenetic agents have been isolated from extracellular matrix components, such as the non-collagenous domains of collagen IV, including arresten, which is generated from the α1 chain of type IV collagen (Colorado et al. 2000); canstatin, a fragment of the α2 chain of type IV collagen (Kamphaus et al. 2000); tumstatin, a fragment of the α3 chain of type IV collagen and the NC1 domain of the α6 chain of type IV collagen (Petitclerc et al. 2000); vastatin, the NC1 domain of α1 chain of type VIII collagen (Xu et al. 2001); anastellin, a peptide derived from the first type III fibronectin repeat (Yi & Ruoslahti 2001); as well as endorepellin, the C-terminal part of perlecan, a component of basement membranes (Mongiat et al. 2003). These cryptic anti-angiogenic agents block a variety of endothelial functions, including proliferation and migration, and display significant inhibition of tumor growth via their vessel-blocking properties. The release of anti-angiogenic fragments could be related to local proteolysis as part of the normal matrix turnover process, generating the anti-angiogenic activity that contributes locally to vessel stabilization. An excess of proteolysis, as occurs in the case of inflammation or tumor vascularization, might lead to high circulating levels of anti-angiogenic agents. If these molecules have vessel-blocking activity, they could be useful as therapeutic agents, whether or not they naturally play this role in vivo. So, it is essential to understand the 15 physiological roles of these fragments in normal and pathological processes (D'Amore & Ng 2002). The C-terminal parts of types XV and XVIII collagen, restin and endostatin, respectively, also belong to the list of cryptic anti-angiogenetic fragments. Endostatin has been suggested as a potential therapeutic agent for inhibiting tumor growth and tumor angiogenesis (O´Reilly et al. 1997). In understanding the physiological role of endostatin it is crucial to understand the physiological role of the parental molecule, type XVIII collagen, in normal processes in the body, which includes the elucidation of the role of type XVIII collagen during development. The existence of three N-terminal variants of type XVIII collagen adds to the complexity and is likely to point to roles in several biological processes. This extracellular matrix field consists of a huge amount of data and it is impossible to deal with this field without making exclusions and compromises. This review will highlight the general aspects type XVIII collagen and its endostatin fragment. Type XVIII collagen belongs to the collagen superfamily and so a short section concerning different collagens is given in this review, as well as the basic aspects of basement membranes.

2.2 Collagen superfamily

Collagens are the major macromolecules of most connective tissues and the most abundant proteins in the body. Tissues such as bone, skin, tendon, cartilage, ligaments, and vascular walls are especially rich in collagen. The main function of collagens is to maintain the architecture of tissues and organs and to confer strength to them. Collagens are also involved in early development and organogenesis, cell attachment, chemotaxis, and filtration through basement membranes. The collagens are defined as structural proteins of the extracellular matrix that contain one or more characteristic triple-helical domains consisting of three polypeptides with Gly-X-Y amino acid triplets. The triple- helical conformation is made possible by the occurrence of glycine as every third amino acid residue. Each triple-helical region has a high proline content in the X position and a hydroxyproline in the Y position of the repeated Gly-X-Y triplet. All collagen molecules consist of three polypeptide chains, called α chains, which wind around each other in a right-handed helix to form the characteristic collagen triple-helix. In some collagens all three α chains are identical, whereas in others the molecules contain two or even three different α chains. All collagen molecules contain non-collagenous sequences at their termini, and some collagens also contain them in the collagenous regions interrupting the Gly-X-Y repeats. The critical role of collagens illustrates a very wide spectrum of diseases caused by a large number of collagen genes (for reviews, see van der Rest & Garrone 1991, Kivirikko 1993, Brown & Timpl 1995, and Brodsky & Shah 1995). Vertebrates have at least 27 collagen types with 42 distinct α-chains (Fig. 1, Table 1). In addition, several other proteins which are not included in the collagen family contain triple-helical domains, such as the pulmonary surfactant proteins, the mannan-binding protein, the macrophage scavenger receptor, and ectodysplasin. The collagen superfamily can be divided into the following subfamilies on the basis of their polymeric structures or other features (Fig. 1) (for reviews, see Van Der Rest & Garrone 1991, Kivirikko 1993, 16 Brown & Timpl 1995, Brodsky & Shah 1995, Prockop & Kivirikko 1995, Pihlajaniemi & Rehn 1995, Myllyharju & Kivirikko 2001, and Myllyharju & Kivirikko 2004). A. Fibrillar collagens (types I, II, III, V, XI, XXIV, and XXVII). These collagens aggregate into fibrils containing large triple-helical domains. The classical fibrillar collagens are types I, II, and III, which occur in high amounts in the body. They are first synthesized as large precursors; then specific proteinases cleave the N- and C- propeptides, allowing their assembly into cross-striated fibrils. Collagen types V and XI are relatively minor components of the extracellular matrix, and are implicated in regulating the diameter of types I and II collagen, respectively. Types XXIV and XXVII are recently characterized collagens that structurally resemble fibrillar collagens. B. Fibril-associated collagens with interrupted triple helices, FACIT collagens (types IX, XII, XIV, XVI, XIX, XXI, XXII, and XXVI). The collagens of this subgroup are multidomain molecules with similar structural features. These collagens do not form fibrils themselves, but are found attached to the surface of pre-existing fibrils of the fibril-forming collagens. All of these collagens are characterized by short triple-helical domains interrupted by short non-collagenous sequences. Type IX collagen is found on the surface of fibrils of type II collagen, and types XII and XIV are found mainly in tissues rich in type I collagen. C. Collagens that form hexagonal networks (types VIII and X). These hexagonal- lattice-assembling collagens are short-chain collagens that are just half the size of fibrillar collagens. They also have a condensed gene structure, a similar triple-helical domain, and a highly conserved C-terminal non-collagenous domain. D. Type IV collagens. These collagens are the major components of the basement membranes. Type IV collagens aggregate to form network-like structures essential for creating the sheet-like architecture of basement membranes. E. Type VI collagen. This collagen, which forms beaded filaments, has a short-triple helical core and extensive C-and N-terminal globular domains. F. Type VII collagen. Type VII collagens are the major component of the anchoring fibrils that link basement membranes into the underlying extracellular matrix. G. Transmembrane collagens (types XIII, XVII, XXIII, XXV). These collagens contain hydrophobic transmembrane segments and are integral plasma membrane proteins. The transmembrane collagens have a type II orientation, with the N-terminus being situated intracellularly and the collagenous domain situated in the extracelular space. H. Types XV and XVIII collagens. Collagen types XV and XVIII form the multiplexin (multiple triple-helix domains and interruptions) collagen family, having an interrupted collagenous region flanked by N- and C-terminal non-collagenous domains. 17

Fig. 1. The members of the collagen family of proteins and their assembled structures. The rods indicate triple-helical collagenous domains in each molecule and solid circles indicate non-collagenous domains. GAG, glycosaminoglycan; PM, plasma membrane. Modified with permission from Myllyharju & Kivirikko (2004). 18 Table 1. Collagen types, their constituent chains, mutations and tissue distributions.

Type Chains Human disease Occurence I α1(I), α2(I) Osteogenesis imperfecta, osteoporosis, Ehlers- Most connective tissues, especially bone, Danlos syndrome types VIIA and VIIB cornea, dermis, and tendon II α1(II) Several chondrodysplasias Cartilage, vitreous body III α1(III) Ehlers-Danlos syndrome type IV, arterial Like collagen I, but absent in bone and aneurysm tendon, IV α1(IV), α2(IV), Basement membrane α3(IV), α4(IV) Autosomal Alport syndrome α5(IV) X-linked Alport syndrome α6(IV) Alport syndrome with leiomyomatosis (also in α5(IV)) V α1(V), α2(V) Ehlers-Danlos syndrome types I and II Tissues containing collagen I α3(V) α4(V) VI α1(VI), α2(VI), Bethlem myopathy, Ullrich muscular dystrophy Most connective tissues like dermis, cornea, α3(VI) and cartilage VII α1(VII) Dystrophic forms of epidermolysis bullosa Anchoring fibrils like in skin and mucosal epithelium VIII α1(VIII), α2(VIII) Corneal epithelial dystrophy Many tissues like Descement´s membrane in the eye and endothelium of vasculature IX α1(IX), α2(IX), Epiphyseal dysplasia, osteoarthrosis, Tissues containing type II collagen α3(IX) intervertebral disc disease (α2(IX), α3(IX)) X α1(X) Schmid metaphyseal chondrodysplasia Hypertrophic cartilage XI α1(XI), α2(XI), Chondrodysplasia Tissues containing type II collagen α3(XI)* XII α1(XII) Tissues containing collagen I XIII α1(XIII) Many tissues like skin and skeletal muscle XIV α1(XIV) Tissues containing collagen I XV α1(XV) Many tissues in BM zone like skeletal muscle XVI α1(XVI) Many tissues like heart, kidney and intestine XVII α1(XVII) Junctional form of epidermolysis bullosa Skin hemidesmosomes XVIII α1(XVIII) Knobloch syndrome Many tissues in BM zone XIX α1(XIX) Many tissues in BM zone XX α1(XX) Corneal epithelium, skin, cartilage and tendon XXI α1(XXI) Many tissues like heart, kidney, and skeletal muscle XXII α1(XXII) Tissue junctions XXIII α1(XXIII) Metastatic tumor cells XXIV α1(XXIV) Developing bone and cornea XXV α1(XXV) Brain, neurons XXVI α1(XXVI) Testis, ovary XXVII α1(XXVII) Cartilage, skin and lung * α3(XI) is posttranslational variant of α1(II). 19 2.3 Type XVIII collagen

2.3.1 Primary structure of type XVIII collagen

Prior to the commencement of this thesis study, the primary structure of type XVIII collagen was determined from mouse and partially also from human (Oh et al. 1994, Rehn & Pihlajaniemi 1994, Rehn et al. 1994, Rehn & Pihlajaniemi 1995, Muragaki et al. 1995, Saarela et al. 1998b). The collagen α1(XVIII) chain cDNA sequences revealed open reading frames for three different N-terminal variants in mouse and two in human (Fig. 2). In the mouse, the sizes of the different N-terminal non-collagenous domains (NC1) are 301, 517 and 764 residues and in the human, 303 and 493 residues. All three different N-terminal variants contain a thrombospondin-1 homology region in the NC1 domain. The rest of the type XVIII collagen molecule consists of a collagenous sequence, having ten collagenous domains (COL1-10) and nine short interruptions (NC2-10) and a C-terminal non-collagenous domain (NC11). In the mouse, the size of the interrupted collagenous domain is 674 residues and the size of the C-terminal non-collagenous domain is 315 residues, while in the human the sizes of the same domains are 688 and 312 residues, respectively. The short variant has its own specific signal sequence and two residues preceding the region common for all three variants. The two longest variants share 247 residues, having also a common signal sequence. The middle variant results from alternative splicing of the RNA transcripts. The longest variant of type XVIII collagen includes 10 conserved cystein residues having homology with the extracellular part of the Wnt-binding frizzled receptor. Altogether, the three type XVIII collagen polypeptide variants are 1315, 1527, and 1774 residues in size in mouse, and 1336 and 1516 residues in human (Oh et al. 1994, Rehn & Pihlajaniemi 1994, Rehn et al. 1994, Rehn & Pihlajaniemi 1995, Muragaki et al. 1995, Saarela et al. 1998b). Type XVIII collagen is one of the very few collagen genes conserved between vertebrates and nematodes (Ackley et al. 2001). The Caenorhabditis elegans gene cle-1 is a homolog to the vertebrate collagen XV and XVIII genes, the homology being the highest with collagen XVIII. The polypeptides encoding cle-1 have three different N- terminal variants like mammalian collagen XVIII, and the protein isoforms are named CLE-1A, CLE-1B, and CLE-1C. All three variants contain a C-terminal non-collagenous domain having 55 % similarity with the mouse type XVIII collagen and a collagenous region in the middle of molecule. The collagenous sequence is distinctly shorter than in the mammalian genes. Isoforms CLE-1A and B have a thrombosbondin domain at the N- terminus with 35% similarity to vertebrate type XV/XVIII collagen. The CLE-1A isoform is the longest variant and contains two fibronectin type III repeats at the N- terminus, but does not include a frizzled homology region as do vertebrates (Ackley et al. 2001). 20

Fig. 2. Schematic picture of type XVIII collagen. Putative signal sequences are shown by vertical and horizontal hatching, non-collagenous (NC) domains in black and collagenous (COL) domains in a white spindle pattern. A sequence specific to the longest variant is shown with a brick pattern. Fz, frizzled domain; Tsp-1, thrombosbondin motif; HS, heparan sulphate binding site. Endostatin is part of the C-terminal domain which can alternatively be marked as NC1 when the domains are numbered from the C terminus to the N terminus.

2.3.2 Gene structure of the mouse α1 (XVIII) chain

The gene encoding the α1(XVIII) chain (COL18A1) is located at the human chromosome site 21q22.3 and on the mouse chromosome 10 (Oh et al. 1994). The complete exon- intron organization for type XVIII collagen was first characterized from mouse (Rehn et al. 1994, Rehn et al. 1996). The length of the mouse gene is over 102 kb and it contains 43 exons. The different N-terminal variants are derived from two promoters. The short variant, 1315 amino acid residues in length, is derived from promoter 1 in conjunction with exons 1 and 2, which mainly encode the signal sequence of the short variant. The second promoter is located 50 kb downstream from promoter 1. Promoter 2 directs the expression of the two longest variants, 1527 residues and 1774 residues, and is in conjunction with exon 3, which is specific for the two longest variants. Exon 3 undergoes alternative splicing resulting in the 1527-residue variant, and the 3´ end of exon 3 is specific for the 1774-residue variant. Exons 4-9 encode the rest of the NC1 domain shared by all three variants. Exons 9-37 encode the interrupted collagenous regions and, exons 37-43 encode the C-terminal non-collagenous domain (Rehn et al. 1994, Rehn et al. 1996).

2.3.3 Type XVIII collagen is a heparan sulphate proteoglycan

Type XVIII collagen is a heparan sulphate proteoglycan (HSPG) (Halfter et al. 1998, Saarela et al. 1998a). Other HSPGs occurring in basement membranes include perlecan and agrin (Noonan et al. 1991, Tsen et al. 1995). HSPGs are composed of a core protein and heparan sulphate (HS) side chains. HS chains are composed of disaccharide units (D- glucuronic acid and D-glucosamine) with modifications like sulphation, which increase the negative charge of HSs. The binding of proteins to HS chains is critical in a variety of 21 biological phenomena, such as immobilization or the protection of proteins against proteolytic degradation or in acting as co-receptors, thus modulating biological activities (Lindahl et al. 1998, Salmivirta et al. 1996). In Western blotting analysis, type XVIII collagen appears as a smear with a molecular mass of 300 kDa. After treatment with heparitinase, the protein was reduced to the 180 kDa core protein (Halfter et al. 1998, Saarela et al. 1998a). Type XVIII collagen has several potential glycosaminoglycan (GAG) attachment sites in its primary structure, but only three attachment sites and their adjacent amino acid residues are conserved between human, mouse, and Xenopus (Dong et al. 2003). All of these sites are shown to carry GAG chains in site–directed mutagenesis, confirming the HSPG nature of collagen XVIII. These GAG sites are located in the middle and N-terminal part of type XVIII collagen (Fig. 2). Type XVIII collagen binds to L-selectin, which is involved in leukocyte migration (Kawashima et al. 2003). The interaction occurs between the lectin domain of L-selectin and the HS chains of collagen XVIII. Also, the monocyte chemoattractant protein-1 (MCP-1), which is involved in inflammation processes, binds to the HS chains of type XVIII collagen (Kawashima et al. 2003). The receptor tyrosine phosphatase sigma (RPTPσ) involved in nervous system development interacts with high affinity with type XVIII collagen, and this interaction is totally dependent upon HS chains (Aricescu et al. 2002). Double-mutant mice lacking type XVIII collagen and HS-depleted perlecan develop lens degeneration earlier than mice carrying the perlecan mutation alone, implying that HS chains of different components can substitute for each other to some extent in the lens capsule (Rossi et al. 2003).

2.3.4 Frizzled domain /cysteine-rich domain

The longest variant of type XVIII collagen contains 10 conserved cysteine residues having homology with the extracellular part of frizzled proteins. The frizzled proteins are seven-transmembrane receptors for the Wnt molecules, and the frizzled motif (fz)/cysteine-rich domain (CRD) in the frizzled protein has been shown to be responsible for Wnt binding activity (Bhanot et al. 1996). Wnt proteins form a family of highly conserved secreted signalling molecules that regulate cell-to-cell interactions during embryogenesis. In the human there are ten different frizzled proteins with the ability to bind certain Wnts. Other proteins in the Wnt-signalling pathway having fz motifs are secreted forms of frizzled proteins (SFRP/FrzB), which contain the cysteine-rich domain but not the transmembrane segments. They may bind to Wnt proteins and change the activity of Wnts. At present, six different SFRP/FrzB proteins have been reported in the human. (Wodarz & Nusse 1998, http://www.stanford.edu/~rnusse/wntwindow.html/ ). Other proteins are also characterized as having a fz/CRD domain (Rehn et al. 1998). The muscle-specific receptor kinase (MuSK), which has an essential role in the formation of the neuromuscular junction, contains a fz/CRD domain in its extracellular part (DeChiara et al. 1996, Masiakowski & Yancopoulos 1998). The receptor tyrosine kinases Ror1 and Ror 2 also contain a fz/CRD domain in their extracellular region. Rors are conserved in the animal phyla, and in the mouse a lack of the Rors causes neonatal 22 lethality (Yoda et al. 2003, Takeuchi et al. 2000, DeChiara et al. 2000, Nomi et al. 2001). In the human, mutations in the Ror2 gene cause recessive Robinow syndrome with severe skeletal dysplasia. Two of these mutations, R184C and R189W, occur in the fz/CRD domain, causing recessive Robinow syndrome (Afzal et al. 2000). The Ror2 fz/CRD domain is able to bind Wnt5a and associate with the rat frizzled2 protein, which is a putative receptor for Wnt5a. Like Wnt5a, Ror2 activates the non-canonical pathway by activation of JNK, and its co-expression has an additive effect on JNK activity (Oishi et al. 2003). In the metallocarboxypeptidase gene family, the carboxypeptidase z contains a fz/CRD domain (Song & Fricker, 1997). In early chicken embryos, carboxypeptidase z is expressed in the somites and later restricted to the sclerotome, and retroviral expression in the region evokes dysmorphogenesis of the scapula and ribs (Moeller et al. 2003). The fz/CRD domain of carboxypeptidase z is able to bind Wnt4 in an immunoprecipitation assay (Moeller et al. 2003). The other described proteins having a fz/CRD domain are corin, which is a transmembrane serine protease, and membrane-type Frizzled-related protein (MFRP), which is mutated in retinal degeneration 6 (rd6) mouse strain suffering progressive retinal dysfunction (Yan et al. 1999, Katoh 2001, Kameya et al. 2002). Also, the smoothened protein is a seven-transmembrane protein that has a cysteine-rich extracellular domain with some homology to the fz/CRD domain (van den Heuvel & Ingham 1996, Rehn et al. 1998).

2.3.5 Expression and localization of type XVIII collagen

Type XVIII collagens are found to be ubiquitous basement membrane (BM) components occuring prominently at vascular and epithelial BM throughout the body (Muragaki et al. 1995, Saarela et al. 1998a). In in situ hybridization staining during early mouse development, the expression of type XVIII collagen is seen in 8-day-old embryos in the mesoderm and in 10-day-old embryos widely in the epithelial and mesenchymal cells, in structures such as the developing head mesenchyma, the gut, the neural tube and ganglia, and heart and the blood vessels (Miosge et al. 2003). From day 12 onwards, expression is seen in the mesenchymal and epithelial cells of all major organs early development stage, like kidney, liver and lung, and at day 16 expression is also seen in the epithelial cells of the skin (Miosge et al. 2003). In the vascular system, the endothelial cells of capillaries express type XVIII collagen mRNA at different developmental stages (Miosge et al. 2003). In these developing organs, type XVIII collagen has been localized by immunostaining to the BM, starting from an 8-day-old early embryo, where the three germ layers, endoderm, ectoderm and mesoderm, are separated by BM. From 10- to 18- day-old developing mice, localization of type XVIII collagen is seen in most of the epithelial BM zones, as well as in the endothelial BM zones of capillaries and larger blood vessels (Muragaki et al. 1995, Sasaki et al. 1998, Miosge et al. 2003). In the adult, immunohistochemical staining localizes type XVIII collagen to different BM types. Type XVIII collagen is seen in the subepithelial BM of kidney, placenta, lung, skin, and liver (Muragaki et al. 1995, Saarela et al. 1998a, Miosge et al. 1999, Tomono et al. 2002). More specifically, in the kidney, type XVIII collagen is seen in the Bowmans´s capsule, the tubules, the glomerular BM, and the mesangial matrix. In placenta, signals 23 are seen in the BM of the villous epithelium, around cytotrophoblasts, and the BM of capillaries and larger vessels. In the lung, staining is seen in the bronchiolar subepithelial BM. In the skin, type XVIII collagen is localized at dermoepidermal BM, and the BMs of sweat glands and hair follicles, and in the liver staining is seen in the BM of bile ducts and in the BM of the sinusoidal space. In muscular structures, type XVIII collagen is present in low amounts in the skeletal muscle around muscle cells and at the musculotendinal junction, in cardiac muscle in very small amounts around the muscle cells, and in the smooth muscle cells in the skin pili muscle cells and around the arteries and veins (Saarela et al. 1998a, Tomono et al. 2002). Type XVIII collagen is seen in the subendothelial BMs of capillaries in a wide range of tissues, such as in the continuous capillaries of the kidney, skin, placenta, and muscular tissue, and in specialized capillaries such as the kidney glomerular capillaries, liver sinusoids, and capillaries of the alveolar wall in the lung (Saarela et al. 1998a, Tomono et al. 2002). The three different variants of type XVIII collagen show characteristic tissue-specific expression patterns. The short variant is the most abundantly expressed of all three variants and is found in most of the conventional BMs, including blood vessels, various epithelial structures, and around muscular structures. Northern blot analysis for the middle variant reveals the strongest signals in the liver, lung, muscle, and kidney in the mouse, and in the human very strong signals are seen in fetal and adult liver samples (Rehn & Pihlajaniemi 1995, Murakagi et al. 1995, Saarela et al. 1998b). In situ hybridizations reveal that hepatocytes express the middle variant in the liver, and immunostaining has localized the middle variant to the liver sinusoids (Saarela et al. 1998a, Musso et al. 1998, Musso et al. 2001). The middle variant is also localized to the BM zone of the epidermal-dermal junction and the hair follicles of the skin. The vascular smooth muscle structures around large vessels also contain the middle variant, but not other smooth muscle structures. In the kidney, immunostaining is seen in glomerular BM and mesangium and some signals are also seen at the musculotendinal junction in skeletal muscle. Immunostaining is also seen in the placenta around the fibroblast of the villi (Saarela et al. 1998a). The middle variant has been shown to exist as a plasma protein that can be immunoprecipitated as a major 150-kDa product (Musso et al. 2001). The long form of type XVIII collagen having the conserved cysteine-rich domain is expressed at the lowest level of the three variants. Northern blot analysis of mouse RNAs with a probe for the longest variant reveals signals in the liver, lung, kidney and muscle, and RT- PCR analysis of human RNAs has shown signals in fetal brain, liver, kidney and retina (Rehn & Pihlajaniemi 1995, Murakagi et al. 1995, Suzuki et al. 2002b). The Caenorhabditis elegans collagen XVIII homolog, CLE-1, encodes three developmentally regulated protein isoforms. Notably, the CLE-1A and B isoforms are expressed only in neurons. CLE-1 is broadly distributed in the basement membranes of C. elegans, but accumulates at the highest levels in association with the nervous system (Ackley et al. 2001, Ackley et al. 2003). 24 2.3.6 Mutations in the type XVIII collagen gene

In humans, mutations in the gene encoding the α1(XVIII) chain are found in patients suffering from Knobloch syndrome. Knobloch syndrome is a rare autosomal recessive disorder defined by the occurrence of high myopia, vitroretinal degeneration with retinal detachment, macular abnormalities, and occipital encephalocele (Knobloch & Layer 1971). In addition to the eye changes, patients may suffer from midface hypoplasia, generalized hyperextensiblity of joints, and hypoplasia of the right lung. An A -> T transversion in a canonic acceptor splice site of the COL18A1 intron 1 (IVS1-2A->T) results in the truncation of the short form of type XVIII collagen (Sertie et al. 2000). A deletion of 2 bp in exon 41 (c3514-3515delCT), a deletion of 1 bp in exon 23 (c2105delC), a deletion of 10 bp in exon 36 (c2969-2978delCAGGGCCCCC), and a nonsense mutation in exon 40 (c3277C -> T) leading to the creation of a premature stop codon for all variants of the type XVIII collagen gene have also been described in Knobloch syndrome patients (Suzuki et al. 2002b). A nucleotide substitution from G to C at position 4181 causing the mutation A1381T is also reported to cause Knobloch syndrome (Kliemann et al. 2003). The combination of two mutations, insertion of 1bp in exon 35 (c3363-3364insC) creating premature stop codon in one allele and transversion in exon 42 (c4309G->A) in another allele, have also been described in Knobloch syndrome patients (Menzel et al. 2004). Individuals having a single nucleotide polymorphism in exon 42 (c4349G -> A) converting aspartic acid (D) to aspargine (N) at the endostatin region were reported to have a 2,5 times increased chance of developing prostate cancer as compared to homozygous D1437 (Iughetti et al. 2001). A mouse model has been generated where the synthesis of type XVIII collagen was disturbed by a neoR casette in exon 30. These mice lacking collagen XVIII/endostatin have abnormalities in the eye, including a delay in the normal postnatal regression of hyaloid vessels along the inner limiting membrane of the retina and abnormal outgrowth of the retinal vasculature (Fukai et al. 2002). The growth of injectable B16F10 melanoma and T241 fibrosarcoma tumor cells were assessed in wild-type and XVIII-null mice, but no differences were detected (Fukai et al. 2002). These mice have also been reported to have a fragile iris and atrophy of the ciliary body (Ylikärppä et al. 2003a). In the iris, the double layer of epithelial cells separates at the apical cell contacts, leading to the rupture of the posterior pigment epithelial cell layer, and extracellular material accumulates in the basement membrane zones both in iris and ciliary body. The XVIII/endostatin-deficient mice are also characterized by abnormal migration of pigmented cells, accumulation of subretinal pigment epithelium deposits, and age-dependent loss of vision associated with abnormal vitamin A metabolism in the retinal pigment epithelium (Marneros & Olsen 2003, Marneros et al. 2004). The mice lacking collagen XVIII/endostatin is also reported to have ultrastructural changes in the BMs of several tissues (Utriainen et al. 2004).

2.3.7 Type XVIII collagen and the eye

Mutations in the type XVIII collagen gene most notably affect the eye. Human Knobloch patients, with mutations in the type XVIII collagen gene, suffer from the occurrence of 25 high myopia and vitroretinal degeneration with retinal detachment (Sertie et al. 2000). In the mouse model with lacking type XVIII collagen expression, the identified defects concentrate in the eye (Fukai et al. 2002, Ylikärppä et al. 2003a, Marneros et al. 2003, Marneros et al. 2004). In situ hybridization reveals the expression of type XVIII collagen in the embryonic chick eye in the future ciliary body, the optic disc and the lens epithelial cells (Halfter et al. 2000). In the embryonic mouse eye, the type XVIII collagen protein is localized in the BM on the outside of the retina in the inner limiting membrane. Staining is also seen in the BM of the hyaloid vessels along the inner limiting membrane of the retina (vasa hyaloidea properia, VHP), in the BM of the vessels around the lens capsule (tunica vasculosa lentis, TVL), and in the lens capsule itself (Fukai et al. 2002). In the adult mouse eye, immunofluorescence staining for type XVIII collagen stains the iris BM, the ciliary body BM, the ciliary body capillaries, the lens capsule, the inner limiting membrane and Bruch´s membrane, some of the choroidal capillary BMs, the corneal epithelium, and Descemet´s membrane (Ylikärppä et al. 2003b, Fukai et al. 2002, Kato et al. 2003).

2.4 Endostatin

Endostatin is a 20-kDa C-terminal fragment of type XVIII collagen (Fig. 2), and it was first identified from murine hemangioendothelioma (EOMA) cell culture medium as a heparin binding fragment having the ability to strongly suppress endothelial cell proliferation in cell culture (O´Reilly et al. 1997). Endostatin has been shown to have the ability to inhibit growth of primary tumors and metastasis. Endostatin inhibits angiogenesis affecting endothelial cell proliferation, migration and apoptosis, and thus it has been a subject of considerable interest as a potential therarapeutic agent (O´Reilly et al. 1997, Sasaki et al. 1998, Dhanabal et al. 1999a, b, c, Sasaki et al. 1999b).

2.4.1 Structure of endostatin

The crystal structure of endostatin reveals a compact single globular domain composed β- sheet structures and loops, but also containing two α-helices (Hohenester et al 1998, Ding et al. 1998, Hohenester et al. 2000). Endostatin contains two disulphide bridges, linking Cys164 with Cys304 and Cys266 with Cys 296. The structure of endostatin resembles the carbohydrate recognition domain (CRD) of mammalian C-type lectins, having the highest homology with the CRD domain of E-selectin (Hohenester et al. 1998). In the parent molecule, type XVIII collagen, endostatin is part of the C-terminal non-collagenous domain NC1 (NC11 when the domains are numbered from the N terminus to the C terminus). NC1 domain (38 kDa) is assembled non-covalently into a trimeric structure. The NC1 consists of three different segments which include the globular 180-residue endostatin domains, each 3 nm in diameter at the C terminus, and an association domain of 50 residues at the N terminus. These two domains are connected by a hinge region, 70 residues, which is sensitive to endogenous proteolysis (Sasaki et al. 1998). 26 2.4.2 Proteolytic generation of endostatin and inhibition of matrix metalloproteinases by endostatin

The proteolytic release of endostatin from the C-terminal part includes at least two steps in the EOMA cell culture (Wen et al. 1999). The first step is metal-dependent and the final step is elastase-dependent. The first cleavage step generates NC1 (32 kDa) and possibly involves matrix metalloproteinases (MMP). The second step results in the cleavage of the NC1 domain to endostatin (20 kDa) at the Ala-His site and involves elastases such as porcine pancreatic elastase and human neutrophil elastase (Wen et al. 1999). The cystein protease cathepsin L also generates endostatin (22 kDa) in the EOMA cell culture, and this does not require the formation of a 30-kDa MMP-generated fragment (Felbor et al. 2000). Several proteinases are thought to cleave the human recombinant NC1 (hNC1) domain into the 25-kDa endostatin. Cathepsins L, B, and K belonging to the cysteine proteinase family, MMP-3, MMP-9, MMP-12, MMP-13, and MMP-20 from the MMP family, and pancreatic elastase from the serine proteinase family all generate a 25-kDa fragment, while catepsin D from the aspartic proteinase family generates a fragment larger than 34 kDa (Ferreras et al. 2000). hNC1 is cleaved to yield endostatin most efficiently by cathepsin L and elastase. These proteinases are also able to degrade endostatin into smaller fragments. The most efficient proteinases in degrading human recombinant endostatin are cathepsins L and D, but cathepsins B and K also degrade endostatin. Pancreatic elastase, however, is not efficient and the MMPs appear totally inefficient in its degradation (Ferreras et al. 2000). MMP-7 has also been shown to cleave hNC1, generating a 28-kDa fragment (Lin et al. 2001a). Human plasma contains 23 to 26 kDa- size endostatin fragments, and extracts from different mouse tissues contain endostatin fragments of 22-38 kDa (Sasaki et al. 1998, Miosge et al. 1999). Endostatin inhibits the activation of proMMP-2 in endothelial and HT1080 cell cultures by direct interaction with proMMP-2, inhibiting the catalytic activities of the active forms of MMP-2 and membrane-type 1 MMP (MT1-MMP) and thereby blocking the invasiveness of endothelial cells and tumor cells (Kim et al. 2000). Endostatin is shown to bind to the catalytic domain of MMP-2, and an unoccupied active site of MMP- 2 is required for endostatin binding (Lee et al. 2002). Endostatin inhibits also the activation of native and recombinant proMMP-9 and recombinant proMMP-13, but it does not affect proMMP-8 (Nyberg et al. 2003). Endostatin is shown to bind to proMMP- 9 during co-immunoprecipitation. Interestingly, many MMPs which are able to generate endostatin are also inhibited by endostatin, suggesting a regulatory feedback loop.

2.4.3 Binding activities of endostatin

Due to its ability to bind heparin, endostatin was described at its discovery as an anti- angiogenic fragment from hemangioendothelioma cell culture (O´Reilly et al. 1997). Recombinant mouse endostatin produced in mammalian cells binds to heparin with moderate affinity (Sasaki et al. 1999b, Ricard-Blum et al. 2004). Endostatin contains a 27 large number of basic residues, in particular arginines, and their distribution on the protein surface is expected to play a major role in heparin binding. Alanine mutagenesis studies indicate that the major heparin binding sites are the clustered arginines in positions 155, 158, 184, and 270, and the ariginines in positions 193 and 194 are also essential for binding (Sasaki et al. 1999b). Analysis of heparin fragments demonstrates that a 12-mer is the minimum size for efficient binding with endostatin (Sasaki et al. 1999b, Blackhall et al. 2003). Mutated endostatin, which has a reduced capacity to bind heparin, also shows a reduced ability to inhibit FGF-2-induced angiogenesis in a chick chorioallantoic membrane (CAM) assay compared to intact endostatin (Sasaki et al. 1999b). The presence of divalent cations has been found to improve the binding of endostatin to heparin and heparan sulphate (Ricard-Blum et al. 2004). The binding of several extracellular matrix proteins to immobilized recombinant trimeric NC1 and endostatin has been determined by solid-phase assays. NC1 shows a strong interaction with perlecan and laminin-1, while endostatin binding is 100-fold less active (Sasaki et al. 1998). A strong and comparable binding of both NC1 and endostatin is observed with fibulin-1 and fibulin-2 (Sasaki et al. 1998). In another study laminin-1 is found to co-precipitate with oligomeric endostatin, and in a solid-phase immunoassay the NC1 domain binds to laminin-1, but not to monomeric endostatin (Javaherian et al. 2002). Strong immunofluorescence and immunogold stainings show the association of endostatin with the elastic fibers of large vessels (Miosge et al. 1999). Solid-phase assays demonstrate the binding of monomeric endostatin to the elastic fiber proteins fibulin-2, fibulin-1, and nidogen-2 and 10-100 times more weakly to nidogen-1 and perlecan, but not to tropoelastin (Miosge et al. 1999). Peptide phage-display library screenings have revealed an endostatin-binding peptide, and it has been suggested that endostatin binds to tropomyosin (MacDonald et al. 2001). In an in situ binding assay, endostatin predominantly binds to blood vessels and some epithelial BM. The binding is resistant to treatment with heparitinase, demonstrating that the binding is not heparin-mediated (Chang et al. 1999). In In vitro assays, endostatin binds to endothelial cell surfaces revealing two binding affinities (Karumanchi et al. 2001). Expression cloning identified glypicans as low affinity receptors. Glypican-1 and glypican-4 are able to bind endostatin through their GAG chains on the surface of the cell (Karumanchi et al. 2001). Glypicans are necessary for endostatin inhibition of endothelial cell migration induced by VEGF, and glypicans are also needed for endostatin inhibition of renal tubular cell branching (Karumanchi et al. 2001, Karihaloo et al. 2001). In another study a mutant oligomeric endostatin, with two mutated arginines, lacked the ability to bind heparin and the ability to abrogate motogenic activity in the endothelian tube assay (Javaherian et al. 2002). Endostatin is not capable of inhibiting angiogenesis in mice lacking either plasma fibronectin or vitronectin, supporting the hypothesis that the activity of endostatin depends on interactions with adhesion proteins (Yi et al. 2003). It has been also demonstrated that E-selectin is required for the antiangiogenic activity of endostatin (Yu et al. 2004). In corneal micropocket assays, recombinant endostatin inhibits angiogenesis in wild-type mice, but not E-selectin-deficient mice. Similarly, endostatin inhibits endothelial sprout formation from aortic rings dissected from wild-type but not from E- selectin-deficient mice. 28 2.4.4 Intracellular signalling mechanisms

On the cell membrane, endostatin has been shown also to bind integrins, in addition to glypicans through their GAG chain (Karumanchi et al. 2001, Rehn et al. 2001). Endothelial cell adhesion to immobilized endostatin is mediated by the α5- and αV integrins. In a solid-phase ligand assay, endostatin binds α5β1 integrins, and this binding is inhibited with an α5 antibody (Rehn et al. 2001). In a cell migration assay, immobilized endostatin promotes and soluble endostatin inhibits integrin-dependent endothelial cell migration and survival (Rehn et al. 2001). Another study has reported that human endostatin prevents endothelial cell migration with no effect on proliferation (Sudhakar et al. 2003). This data indicates that the activity of endostatin is mediated by α5β1 integrins and human endostatin competes with a fibronectin/RGD cyclic peptide to bind α5β1 integrin. Attempts to elucidate the mechanism of action of endostatin have linked it to a variety of intracellular signalling mechanisms. One of the signalling targets is the assembly and disassembly of the focal adhesions regulating the adhesive and migratory phenotypes of cells. In human endothelial cell cultures, endostatin causes a rapid disassembly of focal adhesions observed by immunofluorescence analysis of the focal adhesion proteins paxillin and vinculin, and disruption of the stress fiber network in the cytoskeletal architecture (Wickström et al. 2001). Endostatin has been found to down-regulate the levels of secreted soluble urokinase-type plasminogen activator (uPA) and plasminogen activator inhibitor type 1 (PAI-1) and their complex, and in addition endostatin treatment caused a redistribution of receptor-bound uPA and uPAR, removing them from focal adhesions (Wickström et al. 2001). In human endothelial cell cultures, endostatin induces the clustering of α5β1 integrin associated with actin stress fibers and becomes colocalized with caveolin-1 (Wickstrom et al. 2002). Endostatin could be co-immunoprecipitated with α5β1 integrin and caveolin-1, and exogenous recombinant endostatin has been shown to localize to the lipid rafts through interactions mediated by heparan sulphate proteoglycan and α5β1 integrin (Wickström et al. 2002, Wickström et al. 2003). The raft localization as well as interactions between endostatin, α5β1 integrin, and heparan sulphate proteoglycans are essential for the endostatin-induced Src-dependent phosphorylation of p190RhoGAP with the concomitant down-regulation of RhoA activity. This leads to the disassembly of actin stress fibers and focal adhesions, and an impaired ability to deposit fibronectin into the extracellular matrix (Wickström et al. 2002, Wickström et al. 2003). Moreover, the binding of human endostatin to α5β1 integrin leads to the inhibition of the focal adhesion kinase / c-raf / MEK1/2 / p38 / ERK1 mitogen-activated protein kinase pathway (Sudhakar et al. 2003). In contrast, it has been shown that endostatin can induce endothelial cell tyrosine phosphorylation of focal adhesion kinase (FAK) and paxillin, and promote the formation of focal adhesions and actin stress fibers, similar to fibroblast growth factor-2 (FGF-2) (Dixelius et al. 2002). In cells co-treated with endostatin and FGF-2, focal adhesion and actin stress fibers were decreased, indicating that endostatin disturbs cell-matrix adhesion (Dixelius et al. 2002). On the other hand, human endothelial cells treated with angiogenesis inhibitors including endostatin show increased phosphorylation levels of cofilin and hsp27, proteins involved 29 in actin dynamics, causing a more extensive network of actin stress fibers and an increase in focal adhesion plaques compared to untreated cells (Keezer et al. 2003). Although endostatin inhibits the FGF- and VEGF-mediated migration of primary human microvascular endothelial cells and affects vascular formation in the embryonic body model (Eriksson et al. 2003), this apparent effect of endostatin on the intracellular pathways known to regulate cellular migration and proliferation, such as phospholipase C-γ (PLC-γ), Akt/PKB, p44/42 mitogen-activated protein kinase (MAPK), p38 MAPK and p21-activated kinase (PAK) activity, has not been verified using phosphospecific antibodies. This suggests that endostatin does not affect intracellular signalling pathways known to regulate endothelial migration and proliferation (Eriksson et al. 2003). Nitric oxide (NO) production has been shown to play a role in VEGF-induced endothelial cell migration (Zachary & Gliki 2001). Endostatin has been shown to block VEGF-induced NO synthesis and prevent VEGF-induced endothelial cell migration and tube formation, suggesting that endostatin interferes with signals leading to endothelial nitric oxide synthase (eNOS) activation (Urbich et al. 2002). Thus, endostatin has been shown to prevent VEGF-induced phosphorylation of eNOS (Urbich et al. 2002), which is required for VEGF-induced cell migration (Zachary & Gliki 2001). Endostatin release has also been shown to be modulated by the NO-pathway (Deininger et al. 2003), with NOS inhibitors reducing endostatin expression and release. In addition, adenoviral endostatin transduction results in the release of endostatin from endothelial cells and in the down- regulation of inducible nitric oxide synthase (iNOS) and eNOS (Deininger et al. 2003). Endostatin has also been documented to take part in the Wnt signalling pathway. The Wnt pathway regulates β-catenin, which is known to interact with the lymphoid enhancer factor/T-cell factor (LEF/TCF) transcription factor to activate transcription through its binding site. Endostatin causes the G1 arrest of endothelial cells induced by VEGF and bFGF, inhibiting cyclin D1 RNA expression through the LEF1 site in the cyclin D1 promoter (Hanai et al. 2002a). Thus, endostatin acts at the level of β-catenin or downstream of it (Hanai et al. 2002a). The results suggest that endostatin is a potential inhibitor of Wnt signalling. Overexpression of high amounts of endostatin in Xenopus embryos results in the formation of an ectopic cement gland, which resembles the phenotypes obtained for the overexpression of other Wnt signalling components, such as glycogen synthase kinase 3 beta (GSK3β) (Hanai et al. 2002b). Endostatin has been found to partially inhibit β-catenin and axis duplication caused by Wnt-8, but it does not inhibit axis dublication mediated by TCF-VP16, a constitutively active transcriptional activator, that is independent of β-catenin. Treatment with endostatin decrease the steady- state level of β-catenin in both Xenopus and endothelial cells, suggesting that β-catenin may be a target of endostatin (Hanai et al. 2002b). DNA microarray techniques have been used to elucidate endostatin target genes. Treatment of microvascular endothelial cells with endostatin resulted in the detection of a number of genes acting either as transcription factors or inducers of apoptosis and cell cycle arrest, supporting the hypothesis that endostatin induces the apoptosis of endothelial cells and inhibits migration (Cline et al. 2002). On the other hand, endostatin is also found to down-regulate genes involved in endothelial cell proliferation and migration, thus supporting earlier expression studies of endostatin-treated endothelial cells (Shichiri & Hirata 2001). The more extensive genome-wide expression profiling of endostatin-treated human endothelial cells similarly reveal the down-regulation of many 30 signalling pathways associated with angiogenic activity and the upregulation of many anti-angiogenic genes (Abdollahi et al. 2004). The down-regulated genes include proteins which are known to affect cell growth, differentiation, and angiogenesis, such as Ids, HIF-α, Ephrins, AP-1 and Ets-1 (Abdollahi et al. 2004). Thus, endostatin does not seem to affect a distinct pathway, but instead initiates a complex network of signals.

2.4.5 Biological activities of monomeric and oligomeric endostatin

Some data suggest that monomeric endostatin, originally described by Folkman and colleagues (O´Reilly et al. 1997), and endostatin when present as a trimer, such as in the parent molecule type XVIII collagen, have different biological activities. The NC1 domain containing oligomeric endostatin, but not monomeric endostatin, inhibits the assembly of endothelial cells into tubular structures and induces cell motility on matrigels, indicating that endostatin oligomerization is necessary for its motogenic activity (Kuo et al. 2001). In the same study, oligomerized type XV collagen endostatin do not exhibit motogenic activity. The motogenic activity of the type XVIII collagen NC1 domain is not specific for endothelial cells, since oligomerized endostatin also induces motility of PC12 pheochromocytoma and 293T embryonic kidney cells on matrigels. In cell cultures, monomeric endostatin inhibits the NC1-stimulated motogenic activity, indicating a possible negative autoregulation of NC1 activity mediated by proteolytic release of monomeric endostatin (Kuo et al. 2001). Deletion of the NC1 domain of the C. elegans type XV/XVIII collagen homolog, CLE-1, results in cell migration and axon guidance defects (Ackley et al. 2001). These defects can be rescued by ectopic expression of the trimeric NC1 domain, but not by the monomeric ES domain. On the contrary, ectopic expression of monomeric endostatin inhibits cell motility, phenocopying the NC1 domain deletion. Thus, the data from C. elegans also suggest that monomeric endostatin inhibits NC1 activity. The deletions in the cle-1 gene are also described to cause defects in synaptic structures (Ackley et al. 2003).

2.4.6 Insoluble endostatin

Studies with Alzheimer´s disease brains have revealed significantly more immunoreactivity for endostatin in the pathological tissues compared to controls (Deininger et al. 2002). Immunoreactive signals are seen in the cortical neurons, and high numbers of extracellular and perivascular endostatin deposits are also detected in the cerebral hemisphere colocalizing with amyloid β and tau proteins. Experiments with recombinant endostatin reveal that insoluble endostatin could form cross-β structures and aggregate into amyloid deposits, forming unbranched fibrils, but the soluble globular form of endostatin does not form cross-β sheet structures (Kranenburg et al. 2002, Kranenburg et al. 2003). Insoluble endostatin with cross-β structures is found to stimulate tissue-type plasminogen activator (tPA) mediated plasminogen activation like other cross-β peptides, e.g. amyloid β, and both plasminogen 31 and tPA has been shown to bind the immobilized amyloid endostatin (Kranenburg et al. 2002). Amyloid endostatin stimulates plasminogen activation also in endothelial cell cultures, and co-treatment with amyloid endostatin and plasminogen results in cell detachment, increasing vitronectin degradation (Reijerkerk et al. 2003). Experiments show that amyloid endostatin binds and is toxic to neuronal cells, whereas soluble endostatin has no effect on cell viability (Kranenburg et al. 2003).

2.4.7 Endostatin and cancer

Many studies suggest the possibility to treat cancer with drugs that affect the formation of new blood vessels instead of directly attacking the malignant cells. One of these promising anti-angiogenic agents is endostatin. Endostatin was isolated from a murine hemangioendothelioma cell line as an agent with the ability to inhibit growth of bovine capillary endothelial cells by inhibiting cell proliferation (O´Reilly et al. 1997). This first report on endostatin also showed that it inhibits the growth of Lewis lung carcinoma-, T241 fibrosarcoma-, EOMA hemangioendothelioma- and B16F10 melanoma-implanted primary tumors in mice, as well as the growth of Lewis lung carcinoma metastases. Subsequently, several animal tumor assays have been performed to study the efficiency of endostatin using different tumor cell lines and delivery mechanisms into the tumor. These include inhibition of the growth of MDA-MB-435 breast cancer tumors by injection of an endostatin-coding plasmid intratumorally (Chen et al. 1999) or JC breast carcinoma by adenovirus-mediated gene transfer (Sauter et al. 2000). Endostatin inhibits lung metastases when stably transfected into renal carcinoma cells (Yoon et al. 1999). It also inhibits Lewis lung carcinoma growth and metastasis when using adenovirus- mediated gene transfer (Sauter et al. 2000). Endostatin inhibits the formation of liver metastases in SW620 human colon carcinoma cells engineered to stably secrete endostatin (Yoon et al. 1999) and in murine colon carcinoma 51b-CC cells in association with the subcutaneous administration of recombinant human endostatin (Solaun et al. 2002). Intratumoral administration of the endostatin plasmid inhibits the growth of MCa- 4 murine mammary carcinomas in mice (Ding et al. 2001). In rats, the intragastric carcinogen-induced primary mammary tumor is inhibited by subcutaneously administered recombinant rat endostatin (Perletti et al. 2000). Microencapsulated cells producing endostatin are able to inhibit the growth of BT4C glioma cells in the rat brain (Read et al. 2001, Sorensen et al. 2002). Endostatin can also inhibit the growth of C6 glioma cells engineered to produce endostatin in the mouse skin and the rat brain (Peroulis et al. 2002). Endostatin is effective against tumorigenesis in transgenic mouse models, such as pancreatic islet cell carcinogenesis in RIP-Tag2 mice (Bergers et al. 1999) and mammary adenocarcinoma in C3(1)/T-antigen transgenic mice (Yokoyama et al 2000b, Calvo et al. 2002). In addition, several other studies have established the ability of endostatin to inhibit the growth of various cell lines in animal models, such as ovarian cancer (Yokoyama et al, 2000a), colon cancer (Feldman et al. 2000a, Shi et al. 2002), renal cell carcinoma (Dhanabal et al., 1999a), follicular thyroid carcinoma (Ye et al. 2002), pancreatic cancer (Prox et al. 2003) and, leukemia (Iversen et al. 2002). 32 The mechanisms by which endostatin is suggested to inhibit the growth of different tumors in mice models include the inhibition of tumor vascular density by inhibiting endothelial cell migration and proliferation, and by causing increased apoptosis in tumor cells (O´Reilly et al. 1997, Dhanabal et al. 1999a, Sauter et al. 2000, Yokoyama et al. 2000a, Ding et al. 2001, Peroulis et al. 2002, Shi et al. 2002, Solaun et al. 2002, Sorensen et al. 2002, Ye et al. 2002, Liby et al. 2003, Prox et al. 2003). Type XVIII collagen expression is increased in tumor blood vessels, and endostatin is localized in the vessels of the tumor marginal zone, where angiogenesis is highly active (Ergun et al. 2001, Morimoto et al. 2002). Endostatin is first seen around the endothelium of tumor blood vessels, and in the later maturation phase, endostatin is seen between endothelial cells and in the lamina densa of the BM (Ergun et al. 2001). For certain cancers, elevated serum and plasma endostatin levels have been identified in human cancer patients, such as clear cell renal carcinoma patients (Feldman et al. 2000b) and acute myelogenous leukemia patients (Glenjen et al. 2002). In some human cancers, elevated serum and plasma endostatin levels also have been reported to associate with tumor recurrence after resection and poor prognosis, such as with soft tissue sarcoma (Feldman et al. 2001b), ovarian cancer (Hata et al. 2001), colorectal cancer with liver metastases (Feldman et al. 2001a), non-small cell lung cancer (Suzuki et al. 2002a), myeloid leukemia/myelodysplastic syndrome (Lai et al. 2002), renal cell carcinoma (Feldman et al. 2002) and non-Hodgkin lymphoma (Bono et al. 2003). First clinical trials have been initiated for human patients to test the effect of endostatin on cancer, (Herbst et al. 2002). Recombinant human endostatin has entered phase I trials in doses ranging from 15 to 600 mg/m2, and this amount is generally well-tolerated, with some evidence of minor anti-tumor activity (Herbst et al. 2002).

2.5 Type XV collagen

Type XV collagen is highly homologous with type XVIII collagen, and the two collagens apparently derive from a common ancestor molecule. The highest homology encompasses the C-terminal endostatin domain of type XVIII collagen and the respective restin domain of type XV collagen, the identity of the sequences is 63% (Rehn et al. 1994, Pihlajaniemi & Rehn 1995). The primary structure of type XV collagen has been characterized from the mouse (Hagg et al. 1997) and human (Huebner et al. 1992, Myers et al. 1992, Muragaki et al. 1994, Kivirikko et al. 1994). The human α1(XV) chain contains a highly interrupted collagenous region of 577 residues and non-collagenous N- and C-terminal domains of 530 and 256 residues, respectively. In the mouse α1(XV) chain, the collagenous region is 507 residues and the flanking non-collagenous N- and C- terminal regions are 579 and 256 residues, respectively. In the human, the collagenous portion consists of nine collagenous domains, whereas in the mouse there are only seven. In both species, the NC1 domain of the α1(XV) chain has an approximately 200-residue sequence homologous with the N-terminal end of thrombospondin. Type XV collagen has been shown to be a chondroitin sulphate proteoglycan (Li et al. 2000). The human gene for the α1(XV) chain (COL15A1) is about 145 kb in size, contains 42 exons, and its chromosomal location is 9q21-q22 (Kivirikko et al. 1994, Hägg et al. 1998, Huebner et 33 al. 1992). The mouse type XV gene (Col15a1) is about 110 kb in size, contains 40 exons, and is located on chromosome 4, band B1-3 (Eklund et al. 2000, Hägg et al. 1997). Northern blot analysis has revealed the expression of type XV collagen mRNAs in many tissues like heart, skeletal muscle, kidney, lung, and testis (Murakagi et al. 1994, Kivirikko et al. 1995, Myers et al. 1996, Hägg et al. 1997). In the developing mouse embryo, immunosignals for type XV collagen are first seen in the capillaries at 10.5 dpc and later also in the developing skeletal muscle, peripheral nerves, heart, kidney, and lung (Muona et al. 2002). In adult tissues, type XV collagen is concentrated in the BM zones of epithelial cells, endothelial cells, fat cells, peripheral nerves, and smooth and striated muscle cells. However, type XV collagen localization is not restricted to BM zones, since staining can also be found in the interstitium of the dermis and in the placental villi (Hägg et al. 1997, Myers et al. 1996, Sasaki et al. 2000, Tomono et al. 2002). Mice lacking type XV collagen develop and reproduce normally, but show after 3 months of age progressive histological changes characteristic of skeletal muscle disease, such as degenerative muscle fibers and an increased number of muscle cells with central nuclei (Eklund et al. 2001). In addition, ultrastructural analyses reveal collapsed capillaries and endothelial cell degeneration in heart and skeletal muscle. Double knockout mice lacking both type XV and type XVIII collagens do not reveal any major new defects and only minor compensatory effects are seen in the regression of the capillaries surrounding the lens in the eye (Ylikärppä et al. 2003b). The C-terminal restin fragment of type XV collagen has been identified in the blood circulation, and it ranges in size from 16 kDa to 21 kDa (John et al. 1999). Recombinantly produced restin is able to inhibit the migration of endothelial cells and supress the growth of tumors in a xenograft renal carcinoma model (Ramchandran et al. 1999). The C-terminal fragments of type XV collagen have similar binding properties for extracellular matrix proteins as reported for the C-terminal fragments of type XVIII collagen (Sasaki et al. 2000). The NC1-XV fragment has been shown to bind strongly to fibulin-1 and nidogen-2, and less strongly to fibulin-1, nidogen-1, laminin-1-nidogen-1 complex and perlecan. Endostatin-XV/restin has similar binding properties as NC1-XV, except it binds stronger to fibulin-1. In the chick chorioallatoic membrane (CAM) angiogenesis assay, the C-terminus of type XV collagen is able to inhibit angiogenesis, indicating the nature of the angiogenesis inhibitor (Sasaki et al. 2000).

2.6 Xenopus as a model organism

For over a century amphibian embryos have been a favored system to study development and biological mechanisms. The advantages of amphibian embryos are their external development and relatively large size, allowing the embryologist to perform microsurgery and manipulate the embryos experimentally in ways that are not as readily possible in other vertebrate embryos. One of the disadvantages of traditional amphibian species is that they are seasonal breeders, thus restricting their usage throughout the year. Xenopus laevis, the South African clawed frog, is an exception, since it has the ability to spawn when induced with an injection of a gonadotropic hormone. 34 Several advantages of X. laevis, including identifiable blastomers, the ability to stand molecular and surgical manipulations, the ability to obtain embryos large numbers throughout the year, and easy maintenance, have resulted in it becoming a convenient model system among developmental biologists. Comprehensive studies on the developmental stages of Xenopus have been carried out to describe the external morphology of the embryos and the histology of early stages of development. The first stages of X. laevis development occur relatively fast in optimal temperatures. After fertilization, the cleavages of the embryo occur during a two-hour interval. Subsequent rapid cleavages result in the formation of the blastula 4-5 hours post-fertilization (hpf), and gastrulation commences at 9 hpf. The process of gastrulation involves the movement of the mesoderm and endoderm inside the embryo. Shortly after the end of gastrulation, the neural plate becomes more prominent and neurulation proceeds. The closure of the neural tube at 22 hpf signifies a switch from neurulation to the period of organogenesis. The X. laevis system has most commonly been used to unravel processes related to axis determination, embryonic determination, morphogenesis, and signal transduction. The methods used to identify a variety of activities of gene products include mRNA injection into embryos and the incubation of embryonic explants with protein preparations (for reviews, see Hamburger 1988, Hausen & Riebesell 1991, Wu & Gerhart 1991, Nieuwkoop & Faber 1994, and Moore & Guille 1999). Certain features make X. laevis unsuitable for genetic engineering aimed at producing transgenic embryos. One disadvantage is the 1- to 2-year generation time, which reduces its suitablity for multigenerational experiments. Also, the fact that the genome of X. laevis is allotetraploid, where many genes are represented by extra copies, makes it difficult to use for genetic analysis. A frog more suitable for genetic approaches is X. tropicalis, which has all the advantages of X. laevis as an embryological system, in addition to a shorter generation time and a smaller diploid genome (for reviews, see de Sa & Hillis 1990, Beck & Slack 2001, and Hirsch et al. 2002).

2.7 Basement membranes

Basement membranes (BM) are widely distributed, sheet-like structures that underly almost all epithelia and endothelia in the body and surround many cell types, including smooth and skeletal muscle cells, peripheral nerve fibers, and fat cells. BMs separate compartments, provide structural support for cells, and serve as a selective barrier. In addition, BMs have the capability to influence various activities of the surrounding cells such as growth, migration, and differentiation, which are seen in the processes of embryonic development, wound healing, metastasis, and remodelling of tissues. Using electron microscopy, three layers can be distinguished within the classical BM. The electron-dense lamina densa is separated from epithelial and endothelial cells by the translucent lamina lucida, and a lamina fibroreticularis links them to the connective tissues. The typical thickness of BMs is 50-100 nm, but depending on their location there is marked variation in the BM ultrastructure. During mouse development, a BM can first be visualized by immunostaining at the blastocyst stage. The first extraembryonic BM, Reichert´s membrane, is seen after 35 implantation and later an ultrastructurally visible BM is seen between the ectoderm and the endoderm. Although BMs are widely spread tissue components, their fine structure and composition vary from tissue to tissue and also between different developmental stages. About 50 proteins are now known to make up the BM, and the significance of different proteins for BM assembly and function have been extensively tested in different mouse models (Table 2). The main components of BMs are type IV collagens, laminins, heparan-sulphate proteoglycans (HSPG), and nidogens/entactins. The minor components include agrin, SPARC/BM-40/osteopontin, fibulins, and type XV and type XVIII collagen (for reviews, see Timpl & Aumailley 1989, Timpl & Brown 1996, Erickson & Couchman 2000, Ghohestani et al. 2001, Miosge 2001, Quondamatteo 2002, Engbring & Kleinman 2003, McMillan et al. 2003, Kalluri et al. 2003, Sasaki et al. 2004, and Yurchenco et al. 2004).

2.7.1 Assembly of BMs

The first BM components detected in developing mouse embryos are the laminin β1 and γ1 chains at the two-cell stage. The laminin α1 chain and nidogen are detected at the 8-16 cell stage (Cooper & MacQueen 1983, Dziadek & Timpl 1985), while collagen IV appears later in the inner cell mass of 3- to 4-day-old blastocysts (Leivo et al. 1980). It has been thought that BMs are formed spontaneously by a process of self-assembly of their components, but the cell membrane receptors for BM components also have an essential role in the formation of BM stability (Quondamatteo 2002). A number of interactions contribute to the molecular assembly of BMs. The current BM model proposes the assembly of two networks, one consists of type IV collagens and another of laminins, and the major link between these networks is nidogen-1, but also other interactions are likely to occur (Timpl & Brown 1996). Six α(IV)-chains make up several forms of type IV collagen. Each of the six chains has three domains: an amino-terminal 7S domain, a middle triple-helical domain, and a C-terminal globular non-collagenous NC1 domain. The triple-helical domain contains many interruptions in the collagenous sequence and is responsible for the flexibility of the molecules. The most abundant form of type IV collagen is the heterotrimer [α1(IV)2α2(IV)], and [α3(IV)α4(IV)α5(IV)] and [α5(IV)2α6(IV)] heterotrimers are also known to exist. Type IV collagens have the ability to self-aggregate into dimers by the association of their carboxy-terminal (NC1) domains, and tetramers by the association of their amino-terminal (7S) domains. The C-terminal interactions are stabilized by disulphide bonds, providing high mechanical stability. In addition, the lateral association of molecules results in a complex network of type IV collagen (for reviews, see Kuhn 1995, Hudson et al. 2003, and Kalluri et al. 2003). The laminins are cross-shaped molecules composed of three different classes of polypeptides, α, β, and γ, with each class containing several variants. In the C-terminus, the α-chains contain laminin globular (LG) modules, which are lacking from the β- and γ- chains. Domains I and II of the α, β, and γ chains contain a triple α-helical coiled-coil (CC) forming a long rod-like structure. While domains III and V contain several laminin epidermal growth factor-like (LE) modules that form rigid rod-like structures, domain 36 IV(L4) is globular in structure and its N-terminal part contains the globular domain VI (LN). To date five different α subunits (α1, α2, α3, α4, α5), three different β subunits (β1, β2, β3) and three γ subunits (γ1, γ2, γ3) have been described, and they can form at least 15 different laminins. The laminins are able to self-aggregate, forming hexagonal arrays by three monomers. The N-terminal LN-domain in the short arms is crucial for the polymerization of laminins. The interactions between the monomers are non-covalent, resulting in a more dynamic structure compared to the network formed by type IV collagen (for reviews, see Colognato & Yurchenco 2000, Ericson & Couchman 2000, and Ekblom et al. 2003). The laminin γ1-null mice are embryonically lethal (Smyth et al. 1999). In laminin γ1- null embryonic bodies (EB), neither BMs nor epiblast epithelium forms, whereas the endodermal cells differentiate (Smyth et al. 1999, Murray & Edgar 2000). Epiblast differentiation could be rescued by exogenously added laminin-1 (Murray & Edgar 2000). Laminin β1-null mice are similar to laminin γ1 knockout embryos in that they lack BMs and do not survive embryonic day 5.5 (Miner et al. 2004). In the laminin α1-null embryos, the BM forms, the embryonic ectoderm cavitates, and the parietal endoderm differentiates, apparently because laminin-10 (α5β1γ1) partially compensates for the absent laminin-1 (Miner et al. 2004). Howerever, compensation does not occur for Reichert´s membrane, which is absent, and embryos die by embryonic day 7. The β1 integrin-null cells in EB cultures produce distinctly lower levels of laminin-1 than control cells. The mutant cells also fail to develop properly in the EB culture, while exogenously applied laminin-1 rescues epiblast development of β1 integrin-null ES cells (Aumailley et al. 2000, Li et al. 2002). Interestingly, the inhibition of FGF-signalling by a dominant- negative FGFR2-receptor in ES cells also results in the failure to form the two epithelial sheets, the inner epiblast and the outer endoderm of the embryoid body, the BM, and no detectable amount of laminin-1 and collagen IV is produced (Li et al. 2001). The application of laminin-1 could partially restore the epithelial phenotype, indicating a crucial role for laminin-1 in BM assembly. The laminin and type IV collagen networks are connected by nidogen-1 linking the laminin γ1 subunit to type IV collagen (Fox et al. 1991). Two nidogens, nidogen-1 and nidogen-2, have been characterized, and despite low sequence homology both have a three-globular-domain structure separated by a link region between G1 and G2, and a rod between the G2 and G3 domains (Kohfeldt et al. 1998, Kimura et al. 1998). The C- terminal G3 domain of nidogen-1 links to the single LE domain of domain III of the laminin γ1 chain (Pöschl et al. 1996), while its G2 N-terminal globule binds with high affinity to type IV collagen (Aumailley et al. 1989, Fox et al. 1991). In addition, the G2 domain of nidogen-1 binds to perlecan and fibulin-1 (Battaglia et al. 1992, Reinhardt et al. 1993), whereas the G3 domain of nidogen-1 is able to interact with fibulin-2 (Aumailley et al. 1998). Nidogen-2 has a strikingly lower affinity for the III domain of the laminin γ1 chain compared to nidogen-1, but nidogen-2 binds to collagen type IV and perlecan with the same affinity as nidogen-1. In contrast, nidogen-2 is unable to bind fibulin-1 and -2 (Kohfeldt et al. 1998, Sasaki et al. 1999a). Perlecan is the major heparan sulphate proteoglycan (HSPG) of the basement membrane. The core protein is divided into five domains. The N-terminal domain I is unique to perlecan and includes three glycosaminoglycan (GAG) attachment sites. Domain II is homologous to the class A low-density lipoprotein (LDL) receptor (LA), and 37 domain III is homologous to part of the laminin α-chains, containing domain IV-like and LE modules. Domain IV contains a long series of Ig-like repeats, and the C-terminal domain V has similarities with the globular domain of the laminin α-chain and agrin (Jiang & Couchman 2003). The HS chains interact with a number of extracellular matrix (ECM) molecules including laminin, type IV collagen, and fibronectin (Battaglia et al. 1992, Ettner et al. 1998). The perlecan domain IV can bind nidogens-1 and -2 and fibulin-2 (Hopf et al. 2001), and domain V binds fibulin-2 and the laminin/nidogen complex (Friedrich et al. 1999). The role of integrins in BM organization is suggested by the mouse model lacking the α3-integrin subunit, which displayed aberrant and disorganized BMs (Kreidberg et al. 1996, DiPersio et al. 1997). Also, selective deletion of the β1 integrin gene in the basal cells of the epidermis leads to defects in the BM (Aumailley et al. 2000, Brakebusch et al. 2000, Raghavan et al. 2000). The integrins, the major family of ECM adhesion molecules, are transmembrane proteins composed of one α and one β subunit. To date, 18 α subunits and 8 β subunits are known giving rise to 24 different integrins (Bouvard et al. 2001). Several mechanisms by which the integrins may be involved in BM integrity have been postulated. The binding of BM components to integrin receptors may have some organizational role in the process of BM formation. Integrins α1β1 and α2β1 are the major receptors for type IV collagen and bind to different epitopes within the central region of the triple helix (Timpl & Brown 1996). These collagen-binding integrins can also bind laminin-1, and their binding sites are located in the N-terminal region of the laminin α1 chain (Ekblom et al. 2003). Other integrins binding to the laminin α-chains include α3β1, α6β1, α7β1, and α6β4 (Colognato & Yurchenco 2000, Ericson & Couchman 2000, Hynes 2002, Aumailley et al. 2003). The binding sites of integrins α3β1 and α6β4 have been localized to the C-terminal LG1-3 domain of the laminin α3 chain, while integrin α6β1 has been found to bind the first three LG modules of the laminin α1 chain, a site located opposite to that which integrins α1β1 and α2β1 bind, at the N- terminus of laminin α1 chain. Dystroglycan is a widely distributed dimeric receptor for BM components such as laminin-1 and laminin-2 (Timpl et al. 2000). The α-dystroglycan components bind to the last two LG modules of the laminin α1 and α2 chains, which also bind heparin. In addition, agrin and perlecan bind to dystroglycan (Talts et al. 1999). Agrin is a heparan sulphate proteoglycan that exists as several isoforms in various tissues and is known as a crucial organizer of postsynaptic differentiation at neuromuscular junctions (Bezakova & Ruegg 2003). The role of dystroglycan in providing stability for the BM is controversial. Dystroglycan null mice die during the early embryonic period and display ruptures in Reichert´s membrane (Williamson et al. 1997). In contrast, it has been shown that dystroglycan null embryoid bodies do not have a defect in BM assembly, suggesting that dystroglycan is not required for BM stability (Li et al. 2002). Even organisms in the coelenterates phylum have a BM as do all more complex animals. The basic constituents of the BM, type IV collagen, laminin, nidogen/entactin, and proteoglycans, are highly conserved in lower animals. D. melanogaster and C. elegans have three laminin subunits (2α, 1β, 1γ) that are related to the vertebrate laminins. Similarly, both the fly and worm have a single pair of type IV collagen genes, which are arranged in a head-to-head orientation on a single chromosome in Drosophila, resembling the vertebrates genes. Also, perlecan and nidogen/entactin have homologs in 38 C. elegans and D. melanogaster, suggesting that these four proteins, type IV collagen, laminin, nidogen/entactin, and perlecan, have formed the basis of an early BM. Collagen XV/XVIII and SPARC/osteonectin are also well-conserved among the phyla. With respect to integrins, C. elegans has one β subunit and two α subunits, whereas D. melanogaster has two β subunits and five α subunits. In C. elegans, one integrin binds the ligands at a recognition site containing the tripeptide sequence RGD, resembling the mammalian α5β1, α8β1, and αVβ1 integrins, while another integrin binds to laminins in a way similar to the mammalian α3β1, α6β1, and α7β1 integrins. The Drosophila integrin subunits represent similar classes, but include also others that are non-orthologous to the mammalian subunits (For reviews, see Hynes & Zhao, 2000, and Brown, 2000). 39 Table 2. A selected list of inactivated BM components in mouse

Gene Examples of phenotype References Integrin β1 chain Lethal, E5.5 Failure of blastocyst development, failure of Stephens et al. 1995, Fässler BM assembly & Meyer 1995 Integrin α3 chain Lethal, birth Defects in kidneys and lungs, skin blistering, Kreidberg et al. 1996, rupture of basement membrane Dipersio et al. 1997 Integrin β4 chain Lethal, birth Skin blistering, absence of hemidesmosomes Van der Neut et al. 1996, Dowling et al. 1996 Integrin α6 chain Lethal, birth Skin blistering, absence of hemidesmosomes Georges-Labouesse et al. 1996 Dystroglycan Lethal, E6.5 Failure of Reichert´s membrane Williamson et al. 1997 Laminin γ1 chain Lethal, E5.5 Failure of blastocyst development, failure of Smyth et al. 1999 BM assembly Laminin-α1 chain Lethal, E6.5 Failure of Reichert´s membrane Miner et al. 2004 Laminin-β1 chain Lethal, E5.5 Failure of blastocyst development, failure of Miner et al. 2004 BM assembly Laminin γ2 chain Lethal, postnatal Skin blistering, abnormal hemidesmosomes Meng et al. 2003 5d Laminin α3 chain Lethal, postnatal Skin blistering, abnormal hemidesmosomes Ryan et al. 1999 Laminin β3 chain Lethal, birth Skin blistering, abnormal hemidesmosomes Kuster et al. 1997 Laminin α5 chain Lethal, E17 Failure of anterior neural tube closure, Miner et al. 1998 syndactyly, abnormalities in BM structure Perlecan Lethal, E10-12, Myocardial defects, abnormal cephalic Arikawa-Hirasawa et al. birth development, chondrodysplasia 1999, Costell et al. 1999 Nidogen-1/entactin-1 Viable No obvious phenotype Murshed et al. 2000 Nidogen-1/entactin-1 Viable Neurological defects, abnormal BM in brain Dong et al. 2002 capillaries and lens capsule Nidogen-2 Viable No obvious phenotype Schymeinsky et al. 2002 Deletion of γ1III4 Lethal, birth Renal agenesis, interrupted BM in Wolffian Willem et al. 2002 domain of laminin γ1 duct, impaired lung development Sparc/osteonectin/BM Viable Lens opacity, ruptures in lens capsule Gilmour et al. 1998, Norose 40 et al. 1998, Yan et al. 2002 Agrin Lethal, birth Disruption of neuromuscular function, reduced Gautam et al. 1996 number of AChR in NMJ Fibulin-1 Lethal, birth Hemorrhages, rupture of small vessels Kostka et al. 2001 Type IV collagen α3 Lethal, adult Renal failure, attenuated and laminated Cosgrove et al. 1996, Miner chain glomerular BM & Sanes 1996 Type IV collagen Lethal, E10.5 Impaired BM stability Pöschl et al. 2004 α1/α2 chains Type XV collagen α1 Viable Skeletal myopathy and cardiovascular defects Eklund et al. 2001 chain Type XVIII collagen Viable Eye abnormalities and altered BM Fukai et al. 2002, Utriainen et α1 chain al. 2004 Type XIX collagen α1 Lethal, 3 wk Smooth muscle motor dysfunction and Sumiyoshi et al. 2004 chain hypertensive sphincter in the esophagus 40 2.8 Development of the eye and cataractogenesis

Development of the eye begins during brain vesicle formation. Two bulges from the neuroectodermal walls of the diencephalons, the optic vesicles, protrude laterally towards the head ectoderm. The optic vesicles remain attached to the brain wall by a hollow optic stalk. When the vesicles contact the head ectoderm, the ectoderm thickens into the lens placode. The lens placode reciprocates and causes changes in the optic vesicle, whereafter the vesicle invaginates and forms a double-walled optic cup and the two layers of the optic cup begin to differentiate. The outer layer produces pigment and becomes the pigmented retina. The inner cell layer proliferates and forms the neural retina. The axons from the ganglion cells of the neural retina meet at the base of the eye and travel down the optic stalk; this stalk is then called the optic nerve. On the outer surface of the optic cup, the mesenchymal shell differentiates into the vascular choroid of the eye and the fibrous components of the sclera and cornea. In the development of the lens, the lens placode rounds up and contacts the overlaying ectoderm, inducing the ectoderm to form the transparent cornea. The cells in the inner portion of the lens vesicle elongate and, under the influence of the neural retina, produce the lens fibers. Cells in the posterior half of the vesicle elongate and differentiate to form the primary fibers, whereas cells in the anterior vesicle differentiate into the epithelium. Cell divisions predominantly occur in the epithelial region just above the lens equator known as the germinative zone. The progeny of the cell divisions migrate below the equator into the transitional zone, where they elongate and differentiate into the fiber cells. Thus, the lens consists of three components: the lens capsule, the lens epithelium, and the lens substance, which consist of cortical and nuclear lens cell fibers. The lens capsule is a thick and specialized BM structure enclosing the lens. Beneath the anterior portion of the capsule is a single layer of cuboid epithelial cells that extend posteriorly up to the equatorial region. In the cortical region of the lens, elongated and concentrically arranged cells arise from the anterior epithelium at the equator region. Cortical lens fibers contain nuclei and organelles. The nucleus and organelles eventually disappear when the cortical lens fiber approachs the center of the lens, as in the nuclear lens fiber region (Fig. 3) (for reviews, see McAvoy et al. 1999, and Francis et al. 1999). 41

Fig. 3. Schematic picture of lens morphogenesis during the development of the eye. The upper part indicates the formation of the lens placode and the lower part, the formation and arrangement of lens fibers during embryogenesis.

The lens capsule originates from the BM of the surface ectoderm, which invaginates to form the lens. The lens capsule is the thickest BM in the body and contains typical BM components. It continues to grow throughout most of the organism´s life, growing anteriorly in thickness. The anterior lens capsule is produced by the lens epithelium, but the synthesis and growth of the posterior lens capsule is not fully understood. Studies on the growth of the anterior lens capsule have shown that the capsule tissue is deposited in a lamellar fashion at the inner surface, and mixing of the newly formed collagen with collagen formed earlier does not occur (Krag & Andreassen 2003). A cataract is opacity of the lens caused by a change in the solubility of lens proteins. Based on morphological appearance, the mouse cataract is divided into four major forms: nuclear, cortical, capsular-epithelial, and lens extrusion. Animal models are valuable tools for understanding biological mechanisms associated with human disease, and the fact that cataracts are easily identified in the mouse, makes them ideal candidates for the study of cataractogenesis. About 100 mouse cataract mutants have now been engineered with abnormalities in development, immunity, growth, and physiology (Smith et al. 1997, Francis et al. 1999). Some of the BM genes in the mouse models reveal defects in the lens capsule leading to the onset of the cataract. SPARC/osteonectin/BM40-null mice appear normal and fertile while young, but as they age they develop severe eye pathologies characterized by cataract formation and rupture of the lens capsule (Gilmour et al. 1998, Norose et al. 1998, Yan et al. 2002). In mice lacking exon 3 of perlecan, which results in the loss of attachment sites for three HS-side chains, the major defect is degeneration of lens at an early age leading to congenital cataract (Rossi et al. 2003). Also, mice lacking entactin-1/nidogen-1 have structural alterations in the lens capsule (Dong et al. 2002). 3 Outlines of the present research

At the time when this work was started, the primary structure and genomic structure of mouse type XVIII collagen had just been elucidated, revealing three different N-terminal variants, and the primary structure and genomic structure of human type XVIII collagen had been partially resolved. Comparisons of the homologous type XV collagen primary structure with the human and mouse type XVIII collagen primary structure reveal conserved regions between these collagens and suggested possible functional domains. To study in more detail the conserved regions and to delineate the functional domains of type XVIII collagen, we set to characterize the primary structure of type XVIII collagen from lower organisms and to study the function of this protein in animal models. Thus, new genetically modified mouse strains were generated. The following aims were set during my work to understand the structure and function of type XVIII collagen: 1. To characterize the gene structure of the human type XVIII collagen gene by isolating and sequencing genomic clones covering the whole gene region; to study the longest variant of human type XVIII collagen by cloning the cDNAs of this variant, elucidate the tissue distribution of the corresponding mRNA, and generate an antibody to study the localization and processing of this variant. 2. To characterize the primary structure of type XVIII collagen from Xenopus laevis, study the expression pattern of the different variants during X. laevis embryogenesis, and study the function of type XVIII collagen by disturbing its expression. 3. To generate a genetically altered mouse model to study the function of the endostatin domain of type XVIII collagen. 4 Materials and methods

4.1 Isolation and characterization of cDNA and genomic clones (I, II)

To isolate cDNA clones of X. laevis type XVIII collagen, a 465-bp fragment was obtained from stage 30 cDNA library (Stratagene) using degenerate primers for the endostatin region. The use of this fragment as a probe for screening an X. laevis stage 30 cDNA library revealed clones corresponding to the short variant. To characterize 5´ clones, a cDNA pool was obtained using the Time-Saver cDNA synthesis kit (Amersham Pharmacia Biotech Inc.) and an RNA mixture from adult heart, kidney, brain, and liver. Linkers were ligated to cDNA pools and PCR reactions were performed with these linkers and type XVIII collagen-specific primers. This led to the isolation of clones corresponding to the 5´end of the middle and long variants. To obtain a cDNA clone for the long variant of the human α1(XVIII) chain, EBNA- 293 cells were cultured in 1 x Dulbecco´s Eagles´s medium containing 10 % fetal bovine serum under standard conditions. Total cellular RNA was isolated using the Tri-Pure system (Boehringer), and reverse transcriptase reactions were carried out using a standard protocol with 0.5 µg/µl of RNA as the template and 40 µM hexamers as primers. The PCR reaction was performed using the primers 5´-ATAAGCTTCTCTCTCCTCC TTGCT–3´and 5´-GGAAGTGGTACCGGGCCACTTGGC-3´ at 96°C for 1 min and 65°C for 30 sec for 30 cycles, followed by an incubation at 72°C for 10 min. The ensuing PCR fragment was cloned in the pSP72 vector and sequenced. To isolate and characterize genomic clones of the human α1(XVIII) chains, a human genomic library (Stratagene 951 203) was screened with the cDNA clones Hp19.3, FL 7.1.1, and HuL8.2 (Saarela et al. 1998a, Saarela et al. 1998b), resulting in the isolation of cosmid clones Kos4.1, Kos14.1, Kos19.1, and Kos12.5. Genomic clone I23 was isolated from a cosmid library (Lawrence Livemore National Laboratory) with an EcoRI/AvaI fragment of the cDNA clone HFK1.1 (Saarela et al. 1998b). The clones were analyzed by restriction enzyme mapping and Southern blotting, and appropriate fragments were subcloned in pBluescript SK (Stratagene) or pSP72 (Promega). Nucleotide sequences of genomic and cDNA clones were determined either directly from the cosmid clones or from subcloned plasmids using vector-specific or sequence- 44 specific primers. DNA sequences were determined by the dideoxynucleotide chain termination method by manual or automated sequencing (Applied Biosystems). DNASIS, PROSIS (Pharmacia Biotech.), CLUSTAL W (EMBL outstation), BOXSHADE (ch.embnet.org), and ExPASy (SIB) were used to analyze the nucleotide and amino acid sequence data.

4.2 Reverse transcription (RT) and polymerase chain reaction (PCR) (I, II, III)

RT-PCR was used to analyze the expression of different N-terminal variants of X. laevis type XVIII collagen at different developmental stages. Total RNA was isolated from embryos of different stages, and 5 µg was used for RT reactions performed with hexamer primers following the manufacturer’s instructions (Gibco BRL). The downstream primer for all variants was 5´-ATG GTA CTG GGC CAC CTG GCC-3´, and the specific upstream primers were 5´-CAC CGC TCC ACA CAG CAG CAC-3´ for the short variant (243 bp), 5´-ACA CCT GTA GCG ATG GAT GTA ACC-3´ for the middle variant (504 bp), and 5´-TTG CTG CTG GCT CCG AGC TGT-3´ for the long variant (369 bp). The control oligonucleotides were ornithine decarboxylase (ODC) primers 5´-GGG TGC CTA CAC TGT TGC TGC A-3´ and 5´-TCC ATT CCG CTC TCC TGA GCA C-3´ (180 bp). All the reactions were carried out at 94ºC for 1 min and 65°C for 30 sec for 30 cycles, followed by incubation at 72°C for 10 min. Human Multiple Tissue cDNA Panels (Clontech #K1425-1) were screened with PCR reactions. RT reactions for human skin samples were carried out with 2 µg total RNA. PCR reactions was performed with specific primers for exon 4, 5´-GGGAGTGGTA CCGGGCCACTTGGC-3´, and for exon 3, 5´-CCACCCTGCTGCCAGTTCTGCGAGG- 3´, of the longest α1(XVIII) chain variant at 94C for 1 min and 65°C for 30 sec for 35 cycles, followed by an incubation at 72°C for 10 min. The expected size of the PCR fragment was 283 bp. Expression of transgenic endostatin at the RNA level was studied by RT-PCR. Total RNA was isolated from whole mouse eyes using Tri Reagent (Sigma) according to the manufacture´s instructions, and RT-reactions were carried out using a standard protocol with 2 µg total RNA as the template and 0.5 µg of hexamers as primers. The PCR reaction was performed using the primers 5´-GCGCAGAGGCTCTCACTGCCCTGAT- 3´and 5´-GAGATAGCACCTCGTCCTTCAGGTT-3´ at 94°C for 1 min and 65°C for 30 sec for 30 cycles, followed by incubation at 72°C for 10 min.

4.3 Whole-mount in situ hybridization (I)

At different X. laevis developmental stages, the expression pattern of the N-terminal variants of type XVIII collagen were studied by whole-mount in situ hybridization using antisense digoxigenin-labeled RNA. In situ hybridizations with antisense and sense probes were performed as described (Harland 1991) with some modifications. The probe 45 for all variants was a 4.3 kb cDNA clone amplified in Bluescript. A 1 kb insert in a pSP72 vector was used to generate a probes for the middle and long variants. Probe 3, which was specific for the long variant, was an 800 bp PCR fragment cloned in the pSP72 vector. After whole-mount in situ hybridization, the embryos were embedded in paraffin and sectioned at 5 µm.

4.4 Expression of polypeptide fragments in E. coli for antigen production (II)

A 348 bp fragment corresponding to the sequence preceding the cysteine-rich domain of the long variant of the human α1(XVIII) chain was generated by PCR using a genomic fragment as the template and the primers 5´-ATGGTACCTGCTTCTCTCTCCTCCTTGC TG-3´ and 5´-ATAAGCTTAGCGACCGGCGGGGGCCTC-3´. Following a KpnI/HindIII digestion, the PCR fragment was subcloned in pQE-41 (Qiagen, Inc.), which allows the expression of polypeptides as dihydrofolate reductase (DHFR) fusion proteins with an N- terminal His-tag. This clone, named long18PQE-41, was transformed into the E. coli strain M15, positive clones were checked by sequencing, and protein expression was performed as suggested by Qiagen. The bacterial culture was pelleted, resuspended in 6 M guanide-HCl, 0.5 M NaCl and 20 mM Tris-HCl, pH 7.9, and frozen at -70°C. The cells were lysed by rotation at room temperature for 2 h, lightly sonicated, and centrifuged for 20 min at 12 000 g. The supernatants were run through ProBond columns (Invitrogen) pre-equilibrated with 8 M urea, 0.5 M NaCl and 20 mM Tris HCl, pH 7.9, and proteins were eluted in a 0 to 500 mM stepwise imidazole gradient in 0.5 M NaCl and 20 mM Tris-HCl, pH 7.9. The eluted fractions were analyzed by SDS-PAGE and Coomassie staining. The positive fractions were dialyzed against 1 x PBS and concentrated by ultrafiltration (MWCO 10 kDa, Millipore).

4.5 Preparation and affinity purification of antibodies (II)

Antibodies to the fusion protein long18PQE-41 were raised in rabbits. The sera were tested by Western blotting using crude cell lysates from bacteria expressing the long18PQE-41 and long18PQE-31 clones. For affinity purification of the crude sera, an expression clone, long18PQE-31, was made similarly to long18PQE-41, except that the vector pQE-31 (Qiagen, Inc.), which contains the N-terminal His-tag but not the DHFR leader sequence, was used. This fragment was expressed and purified as described above, except that an additional purification step was included. After 1x PBS dialysis, the fragment was dialyzed against 50 mM Na2HPO4, pH 7.0 and applied to a HiTrap Sp cation exchange column according to the manufacturer´s instructions (Pharmacia). The bound polypeptides were eluted with a stepwise gradient of NaCl from 0-1 M in 50 mM Na2HPO4, pH 7.0. The eluted fractions were characterized by SDS-PAGE and Coomassie staining. The positive fraction was dialyzed against 0.5 M NaCl in 0.2 M NaHCO3, pH 46 8.6, after which the material was coupled to CNBr-Sepharose (Pharmacia) according to the manufacturer´s protocol. The antiserum termed anti-fz was diluted 1:1 with 1 x PBS, pH 7.2, and applied to the affinity column, which was subsequently washed with 2 M NaCl-1 x PBS, pH 7.2. The bound antibody molecules were eluted with 150 mM glycine-HCl, pH 2.5, and then with 100 mM triethylamine, pH 11.0. The fractions containing protein were detected at absorbance 280 nm, neutralized immediately with 2 M Tris-HCl, pH 7.5, pooled, and concentrated (MWCO 100 kDa, Millipore). The concentration was determined by the absorbance at 280 nm, where 1 mg/ml of IgG gives a value of 1.35.

4.6 Immunoprecipitation (II)

Human serum was immunoprecipitated with three antibodies for type XVIII collagen, anti-all (Saarela et al. 1998a), anti-long (Saarela et al. 1998a), and anti-fz. The antibodies were bound to sheep anti-rabbit magnetic beads (Dynal A.S) for 3 hours in 0.1 % BSA- PBS, pH 7.4, and washed in 0.1 % BSA-PBS, pH 7.4. After several washes, the antibody- coated beads were incubated with 1 ml aliquots of serum overnight at +4°C. Unbound proteins were extensively washed with 1 % Triton X-100, 0.1 % SDS, and 1 x PBS, pH 7.4, and diluted in 60 µl of 50 mM Tris-HCl, 5 mM CaCl2, and 0.1 % BSA, pH 7.4. The aliquots were divided into two halves, one of which was exposed to digestion by 15 U of bacterial collagenase (Worthington), and both aliquots were incubated for 3 h at 37°C. The samples were analyzed by 5 % SDS-PAGE and Western blotting. To re-use the blot, it was stripped with 100 mM β-mercaptoethanol, 2 % SDS in 62.5 mM Tris-HCl, pH 6.7, for 30 min at 50°C, and washed.

4.7 SDS-PAGE and Western blot (II, III)

Different samples were analyzed by SDS-PAGE and Western blot analysis. To test the generated antibodies, EBNA-293 cells transfected with full-length cDNAs for the long and middle forms of human collagen XVIII were collected and suspended in PBS. The samples were mixed in 4 x SDS sample buffer and β-mercaptoethanol was added to 5 %, followed by boiling for 5 min. The aliquots were analyzed by 7 % SDS-PAGE, followed by staining with Coomassie Brilliant Blue or Western blotting onto polyvinylidene fluoride membranes. The filters were blocked with 5 % fat-free milk powder-0.1 % Tween 20 in TBS, pH 7.4. The antibodies, anti-all and anti-fz, were used at dilutions of 0.5 µg/ml, and the reactions were detected with a horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Bio-Rad) and enhanced chemiluminescence detection reagents (Amersham Corp.) as recommended. To detect the N-terminal fragments, EBNA-293 cell culture media were concentrated to 1/20 volume using a centrifugal 30K NMWL membrane (Millipore). Forty microliter aliquots were analyzed by 10 % SDS- PAGE and Western blotting as above with antibodies for the N-terminal variants, anti-all (Saarela et al. 1998a), anti-long (Saarela et al. 1998a) and anti-fz. 47 Homogenate from trangenic tissues was isolated from whole eyes using Tri Reagent (Sigma) according to the manufacturer´s instructions. The precipitated proteins were dissolved in 32 µl of 4 x SDS sample buffer, and β-mercaptoethanol was added to 5 %. A 1 cm2 skin sample was homogenized in 1 ml 8 M urea, 0.5 M NaCl, 20 mM Tris, pH 7.9 with 2 % β-mercaptoethanol using an ultra turrex homogenizer and centrifuged at 13 000 rpm for 10 min. Ten microliters of the skin supernatant with 4 x SDS sample buffer and 10 ul of eye protein homogenate was analyzed by 15 % SDS-PAGE and Western blotting as above with the polyclonal antibody against endostatin, RES (Kawashima et al. 2003). Antibody dilutions of 0.5 µg/ml were used with an overnight incubation at +4°C.

4.8 Cell culture, transfections, and immunofluorescence of cells (II)

EBNA-293 cells were cultured on 35 x 20 mm plates to 80 % confluency in 1.5 ml medium containing 10 % fetal bovine serum. The cells were transfected using the Fugene6 transfection reagent (Boehringer Mannheim) with 5 µg of full-length cDNA for the long and middle forms of human type XVIII collagen, which had been ligated in the plasmid pREP7 (Invitrogen). After 24 hours the cells were collected and suspended in 1 ml 1 x PBS. To detect the N-terminal fragments, EBNA-293 cells were cultured on 100 x 20 mm plates to 80 % confluency in 10 ml medium supplemented with 10% FBS as above. The cells were washed with 1 x PBS, and medium without serum was applied. After 3 and 5 days the medium was collected and analyzed by Western blotting. For immunostaining, transfected EBNA-293 cells grown on glass coverslips were fixed in –20°C methanol for 5 min and incubated in 5 % FBS-PBS for one hour. After blocking, they were treated with the antibody (0.5 µg/ml) overnight, then washed with 1 x PBS and incubated with a 1:300 dilution of a Cy3-conjugated goat anti-rabbit secondary antibody (Jackson Immuno Research). After washing, the coverslips were mounted with immuno-mount (Shandon) on microscope slides and analyzed under a light microscope.

4.9 Generation of transgenic mouse lines (III)

A transgenic cDNA construct was generated that encoded the signal sequence and C- terminal endostatin part of the mouse α1(XVIII) chain. The signal sequence and the endostatin region were generated by PCR using a mouse cDNA clone (Rehn et al. 1994, Rehn & Pihlajaniemi 1995) as the template and the partially complementary primers 5´- GGATCCGGCCCAGCGCA-3´; 5´-ATGGTGATGGTGATGGTGATCAGCGCTGGCA GGCA-3´; and 5´-CACCATCACCATCACCATACTCATCAGGACTTT-3´; 5´-GGATC CTATTTGGAGAAAGAGGTCATGAAG-3´. In the second step, PCR was carried out using DNA fragments from the first PCR reactions as the template and the primers 5´- GGATCCGGCCCAGCGCA-3´ and 5´-GGATCCTATTTGGAGAAAGAGGTCATGA AG-3´. The ensuing PCR fragment was digested with BamH I, ligated into the keratin 14 promoter, K14-promoter, expression cassette (Vassar et al. 1989), and the cloned 48 expression vector was injected into fertilized mouse oocytes of the strain FVB/NIH. Three independent transgene mouse lines with germline transmission were generated. The genotypes of the transgene mice were confirmed by Southern blotting and PCR with the specific primers 5´-GCGCAGAGGCTCTCACTGCCCTGAT-3´ and 5´- GAGATAGCACCTCGTCCTTCAGGTT-3´ at 94°C for 1 min and 65°C for 30 sec for 30 cycles, followed by incubation at 72°C for 10 min. Col18a1-/-;K14-endostatin double- mutant mice were generated by cross-breeding the Col18a1-/- and the endostatin transgenic mice (Fukai et al. 2002).

4.10 Immunohistology (II, III)

Immunohistochemistry was performed using frozen and paraffin sections. Frozen sections for immunofluorescence staining were cut into 5 µm and 9 µm cryosections and fixed in ethanol for 10 min at -20°C. Sections were blocked by incubation with 1 % blocking reagent (Boehringer Mannheim) in PBS, pH 7.4, for 1 h at room temperature followed by overnight incubation at 4°C with a specific antibody. An anti-His-tag monoclonal mouse antibody (34660, Qiagen), an anti-endostatin rabbit polyclonal antibody, RES (Kawashima et al. 2003) and anti-fz rabbit polyclonal antibody were used in concentrations of 1-5 µg /ml. After washing with PBS, pH 7.4, a Cy3-conjugated goat anti-rabbit antibody (Jackson Immunoresearch Laboratories Inc.) for endostatin antibody and an Alexa 568-conjugated goat anti-mouse antibody (Molecular Probes, Eugene, OR) with a His-tag antibody were applied and samples were incubated for 60 min in the dark at room temperature. After extensive washing, slides were mounted in immuno-mount (Shandon Inc.) and examined under an epifluorescence microscope. For paraffin sections, tissues were fixed in 4 % paraformaldehyde, dehydrated, and embedded for paraffin and sectioned at 5 µm. Immunostaining of paraffin sections was performed using the Histomouse-kit or Histostain-SP kits (Zymed) according to the manufacturer´s instructions. The same consentrations of antibodies were used as above.

4.11 Histology and light microscopy (I, II, III)

X. laevis embryos were embedded in paraffin after whole-mount in situ hybridization and sectioned at 5 µm. Eyes from transgenic and control mice of different ages were obtained and fixed in phosphate-buffered 4 % paraformaldehyde for 24 hours. After fixation, the samples were dehydrated and transferred to the infiltration solution of the Leica Historesin Embedding Kit (Leica) for 3 days, and the infiltration solution were changed a couple of times. Samples were embedded in embedding medium (Leica) and attached to a block holder after polymerization by Leica historesin mounting medium (Leica). Plastic sections were obtained at 1.5-3 µm and stained with hematoxylin and eosin by routine methods. 49 4.12 Electron microscopy (III)

For transmission electron microscopy, skin and eye samples from control and transgenic mice were fixed in 2.5 % glutaraldehyde in 0.1 M phosphate buffer, postfixed in 1 % osmiumtetroxide, dehydrated in acetone, and embedded in Epon Embed 812 (Electron Microscopy Sciences). Thin sections were cut with a Reichert Ultracut ultramicrotome and examined in a Philips CM100 transmission electron microscope. Images were captured and the thickness of the BMs was measured by a CCD camera equipped with TCL-EM-Menu version 3 from Tietz Video and Image Processing Systems GmbH. The measured BMs included both the lamina lucida and lamina densa. For statistical analyses (t-test and F-test), the analyzed numbers were 11 transgenic mice and 7 wild-type mice, with 25 measurements for each mouse. The analyzed mice varied from 2.5 to 5.5 months of age.

4.13 Immunoelectron microscopy (III)

For immunogold labeling, skin and eye samples were prepared from transgenic mice and wild-type mice, fixed in 4 % paraformaldehyde in 0.1 M phosphate buffer with 2.5 % sucrose for two hours at room temperature, immersed in 2.3 M sucrose, and frozen in liquid nitrogen. Thin cryosections were cut with a Leica Ultracut UCT microtome. For the immunolabeling, sections were first incubated in 0.05 M glycine in PBS, followed by incubation in 5% BSA with 0.1% CWFS (cold water fish skin) gelatin (Aurion) in PBS. Antibodies and gold conjugate were diluted in 0.1 % BSA-C (Aurion) in PBS. All washings were performed in 0.1% BSA-C in PBS. Sections were then incubated with polyclonal antibodies against endostatin, RES (Kawashima et al. 2003) and against the N-terminal part of type XVIII collagen, ELQ (Saarela et al. unpublished results) for 60 min, followed by protein A-gold complex (size 10 nm) for 30 min. Controls were prepared by carrying out the labeling procedure without the primary antibody. The sections were embedded in methylcellulose and examined as described above. 5 Results

5.1 Xenopus laevis type XVIII collagen (I)

Using degenerate primers we obtained cDNA clones encoding the endostatin region of X. laevis type XVIII collagen. cDNA library screening and RT-PCR reactions led to the isolation of cDNA clones covering the entire length of type XVIII collagen from X. laevis. X. laevis type XVIII collagen contains three variant forms, differing in their N- terminal non-collagenous sequences. The shortest variant consists of a 22-residue signal peptide and a 1285-residue mature portion, with only the first three residues of the latter being specific to this variant. The middle variant (1581 residues excluding the signal peptide) and the long variant (1886 residues excluding the signal peptide) both contain a 23-residue signal sequence and a 299-residue N-terminal non-collagenous portion. Specific to the long variant is a 305-residue domain characterized by a cysteine-rich region sharing homology with the ligand-binding portion of frizzled receptors. All three variants share the rest of the molecule in common, including an N-terminal thrombospondin-1 homology region, a highly interrupted collagenous domain, and a C- terminal non-collagenous domain. The best conserved region when compared to mammalian homologs is the C-terminal endostatin domain. RT-PCR analyses of frog embryos at different developmental stages indicate that the short variant is most strongly expressed, being detectable after gastrulation at stage 10, the first time point analyzed, and markedly increasing in expression level until stage 12. The RT-PCR patterns for the middle and long variants show a low expression at early stages that increases at later stages. Since only a very short sequence is specific to the short variant, a probe recognizing all variants was used first for whole-mount in situ hybridization. Due to the high expression level of the short variant, the probe for all variants detected mainly these transcripts. Expression was detected in the eye anlage and the neural tube, and at the boundaries between somites at the end of neurulation, stage 20. At stage 29, expression was found in the eye, at the mid-hindbrain boundary, in the otic vesicle, head mesenchyme, heart, branchial arches and pronephros and between the somites. 51 Using a probe that detected both the middle and long variants, signals were first seen at the tail bud stages. Strong signals were found in the pronephros and pronephric duct, near the olfactory placodes, and in the mandibular, hyoid, and subsequent branchial arches. Sectioning revealed staining in the mesenchyme underneath the olfactory placodes and in the center of the branchial arches. Signals were also seen in the region of cranial sensory ganglia V, VII, IX and X, extending into the mandibular, hyoid, and third branchial arches, with the strongest expression seen ventrally. These signals mainly represented the middle variant, as they were very faint with the probe for the longest variant. Signals mainly representing the middle variant were also found between the somites, most likely in the endothelium of the intersomitic arteries. Signals specific to the long variant were weak, with the strongest staining at the tadpole stages being found in the head and branchial arches. In the head, expression was strongest in the mesenchyme underneath the olfactory placodes and in the mandibular and hyoid arches, as also seen in embryo sections. Faint signals at the tadpole stages were seen around the otic vesicle and in the pronephros and pronephric duct.

5.2 The longest variant of human type XVIII collagen (II)

The cDNA structures of two variants of the human α1(XVIII) chain have been elucidated in earlier studies (Saarela et al. 1998b), but the structure and localization of the longest variant of type XVIII collagen remained undetermined. PCR reactions from EBNA-293 cell-derived cDNAs revealed an 850 bp fragment which encoded a longer NC1 domain than the previously identified NC1-303 and NC1-493 domains (Saarela et al. 1998b). The longest α1(XVIII) variant consisted of 1751 amino acid residues, including the 728- residue NC1 domain and a 23-residue putative signal sequence. The NC1-728 domain contains 11 cysteine residues, 10 of which are conserved with the extracellular part of frizzled proteins. Comparison of the amino acid sequences encoded by exon 3 between the human, mouse, and frog type XVIII collagen genes revealed that the region specific to the NC1-728 variant was more conserved than the part of exon 3 that encodes sequences shared by the NC1-493 and NC1-728 variants. In particular, the cysteine-rich region of the long variant showed the highest conservation, which suggest it may be of some functional importance RT-PCR was carried out from various human fetal tissues and adult skin to investigate the distribution of transcripts encoding the NC1-728 variant. PCR reactions with specific primers for exons 3 and 4 of the longest α1(XVIII) chain variant revealed clear signals in fetal liver, lung, skeletal muscle, and spleen as well as in adult skin, and weak signals were found from the kidney and thymus. To generate an antibody for the third variant of human type XVIII collagen, a fragment specific to the longest α1(XVIII) chain variant was expressed in bacteria and the recombinant protein was purified and used to immunize rabbits. The affinity-purified antibodies were characterized using crude cell lysates from EBNA-293 cells expressing a full-length cDNA construct encoding the long (1751aa) and middle (1516aa) forms of the human α1(XVIII) chains. Cells transfected with the middle variant were used as a negative control. No immunosignal was visible in the control sample with the anti-fz antibody, as expected, whereas cells transfected with the full-length cDNA for the long 52 variant gave a strong signal of about 200 kDa along with some minor bands. The previously described antibody that detects all three variants of type XVIII collagen, anti- all (Saarela et al. 1998a), gave strong signals in both transfected cell lysates, indicating that the anti-fz antibody detects specifically the long form of type XVIII collagen. The specificity of the anti-fz antibody was also confirmed by immunofluorescence staining of EBNA-293 cells transfected with the same full-length cDNA construct as above. Strong immunostaining was detected in the transfected cells when incubated with the anti-fz antibody, while only background staining was seen in untransfected EBNA-293 cells. Human serum samples were immunoprecipitated with the antibodies anti-fz, anti-long, previously shown to be specific for the long variant of type XVIII collagen (Saarela et al. 1998a), and anti-all, and detected with anti-long and anti-fz using Western blotting. The anti-long antibody detected a 173-kDa polypeptide immunoprecipitated with all three antibodies, and a 144-kDa polypeptide immunoprecipitated with anti-long, and anti-all. Both fragments were sensitive to bacterial collagenase, confirming their collagenous nature. The anti-fz antibody detected only the 173-kDa fragment in the immunoprecipitates of anti-fz, anti-long and anti-all. Since the 144-kDa fragment contains the anti-long and anti-all epitopes but not the epitopes for anti-fz, it was identified as derived from the middle variant. Since the 173-kDa band was positive with all three antibodies, it was identified as being derived from the longest variant. For the determination of tissue location of the longest variant, paraffin sections from fetal and adult tissues were analyzed by immunostaining using the affinity-purified anti- fz antibody. The clearest staining with this antibody could be seen around the branching bronchioles in the developing fetal lung, on the luminal side of the epithelial cells, and also weakly in the cell BM zone lining the mesenchyme. The developing fetal skeletal muscle of the fingers exhibited some granular or dot-like staining on the forming myotubes. This staining could be detected throughout the muscle, but the staining appeared strongest at sites where the muscle attached to the surrounding connective tissue.

5.3 Human type XVIII collagen gene structure (II)

The gene structure of mouse type XVIII collagen has been characterized earlier (Rehn et al. 1996), and we showed that the human gene strongly resembles its mouse counterpart. The human COL18A1 gene contains two promoters, an upstream one adjacent to exons 1 and 2, encoding the signal peptide and the first two residues specific to the short N- terminal variant, NC1-303, and a downstream promoter adjacent to exon 3, encoding another signal peptide and sequences specific to the previously described (Saarela et al. 1998b) NC1-493 and a third variant, NC1-728. Exon 3 encodes 192 residues shared by the NC1-493 and NC1-728 variants, together with 235 residues that are predicted to be present only in the NC1-728 variant due to the alternative splicing of exon 3 sequences. The distance between promoters 1 and 2 was 50 kb. Characterization of the overlapping genomic clones revealed that the human type XVIII collagen gene contained 43 exons and was about 105 kb in length. The lengths of the exons and the introns varied from 27 to 1350 bp and from 79 bp to 50 kb, respectively. Intron 2, which separates exons 2 and 53 3, corresponded to nearly half of the total size of the gene. The exon-intron boundaries conformed well to the consensus sequence AG-exon-GT. The N-terminal variant-specific exons 1-3 were followed by exons 4 to 9, encoding the common regions of the NC1 domain, with exon 9 also encoding the COL1 and NC2 domains and the beginning of the COL2 domain. Exons 10-37 encoded the collagenous domain with multiple interruptions. Exon 37 was a junction exon which encoded the rest of the NC10 domain, the last collagenous domain COL10, and the first two residues of the extreme C-terminal non-collagenous NC11 domain. The rest of the NC11 domain was encoded by exons 38-43, containing the association domain, the hinge region and the endostatin domain. Comparison of the mouse and human genes revealed differences in the lengths of some exons. The lowest identity percentages were seen in the first half of exon 3, in exon 6, encoding the end of NC1, and at the beginning of the collagenous region, particularly in exon 9, which corresponded to sequences shown to be subjected to alternative splicing (Saarela et al. 1998b). The highest identity percentages were seen in the exons encoding the NC11 domain which included the endostatin domain and the exons encoding collagenous region.

5.4 Transgenic mice overexpressing endostatin (III)

Transgenic mice overexpressing endostatin were generated using a construct encoding the extreme 184 C-terminal amino acid residues encompassing the endostatin domain of mouse type XVIII collagen, a signal sequence, and six histidine residues for the specific detection of the transgenic protein. The cDNA was inserted into an expression vector with a keratin 14 promoter, K14-promoter, and the plasmid was injected into fertilized oocytes to create transgenic mice lines. Transgene mice were identified by Southern blotting and PCR, and three independent transgenic founders, K18, G20 and J4, were mated with FVB/N mice to generate separate lines.

5.5 Transgenic endostatin in the lens of the eye (III)

Transgenic mice expressing endostatin under the K14-promoter developed opacity of the eyes in line J4 at four to eight months of age, and after ten months the penetrance was nearly complete. To analyze this unexpected finding, the expression of the transgene in the eyes was tested by performing RT-PCR from whole eyes of three- and eight-month- old transgenic mice. RT-PCR signals were seen in all three transgenic mouse lines. The signal was weaker in the K18 line than in the two other lines. At the protein level, expression of the transgene in the eyes was analyzed by Western Blotting from three- and eight-month-old transgenic mice, revealing 20 kDa signals with an anti-endostatin antibody, RES, in both three- and eight-month-old mice from lines G20 and J4. The strongest of the transgene signals was seen in the J4 line and somewhat weaker ones in the G20 line, and in the K18 line the endostatin signal was undetectable. All three mouse 54 lines revealed expression of the transgene at the RNA level, but Western blot analysis confirmed clear protein expression only in lines G20 and J4. The localization of the transgene product was analyzed in lens immunofluorescence stainings of frozen eye sections using the RES antibody and an anti-His-tag antibody specific for the transgene product. RES antibody staining was seen in the wild-type sections at the outer side of the lens capsule, detecting the endogenous endostatin domain. The anti-His-tag antibody detects the transgene product but not endogenous endostatin. Anti-His-tag staining was seen throughout the lens capsule. Immunoelectron microscopy with the RES antibody confirmed endogenous collagen XVIII/endostatin localization at the outer surface of the lens capsule in wild-type mice, and in the transgenic mice clear immunoelectron microscopic signals were seen throughout the lens capsule, indicating a broader localization compared to the endogenous collagen XVIII/endostatin. Light microscopic analysis of plastic-embedded and hematoxylin eosin-stained eye sections from four-month-old transgenic mice revealed vacuoles in the lens epithelium. These changes were most prominent in line J4, with the highest transgene expression, but changes in the epithelial layer could also be seen in line G20, albeit to a somewhat lesser extent. Over one-year-old G20 mice contained an increased number of vacuoles between the epithelial cells and the lens capsule. The ruptures first appeared in the central epithelium and later extended towards the equatorial epithelium. At the ultrastructural level, the ruptures were localized to the anterior side of the lens epithelial cells at the border between the cell layer and the lens capsule. These abnormalities were first detectable in 2.5-month-old J4 mice. Additional defects were seen in the eyes of over one-year-old mice, including disorganization of the lens epithelial layer on the anterior side. The transgene samples contained multilayer cell clumps, where the epithelial cells had lost their normal shape and became spindle-like. These severe defects were most readily seen line J4.

5.6 Transgenic endostatin in the skin (III)

The K14-promoter drives the expression of the transgene mainly in the basal cells of the skin. The expression of transgenic endostatin in the skin was first confirmed by SDS- PAGE fractionation and Western blotting of skin samples from eight-month-old control mice of all three transgenic mouse lines. The anti-endostatin antibody, RES, recognized a 20-kDa fragment in samples from the G20 and J4 lines, while samples of wild-type and K18 mice were negative. The differences in the intensities of the signals corresponded to the previously noted differences in the transgene expression in the eyes of the three lines. In the K18 mice line, the amount of transgenic endostatin was undetectable by Western blotting reflecting a very low expression level, while in lines J4 and G20 the amount of the transgene was greater. The highest level of transgene-derived endostatin was seen in line J4. Immunofluorescene staining was performed to localize the transgenic endostatin in the skin. The RES antibody was used to detect both transgene and endogenous gene products. The immunostainings were performed with 2-, 4-, and 8-month-old mice. In wild-type 55 skin, endogenous endostatin/type XVIII collagen was localized to the BM regions in the epidermis-dermis junction and the hair follicles. In the skin of transgenic mice, the staining was localized to the same structures as endogenous endostatin/type XVIII collagen, but the intensity of the staining was stronger. To ascertain the expression pattern of the transgene-derived endostatin, the J4 mice were mated with Col18a1-/- mice to obtain offspring lacking endogenous type XVIII collagen/endostatin expression, but expressing the transgene-derived endostatin. The anti-endostatin signal was observed in these Col18a1-/-; K14-endostatin mice in a manner similar to that of the endogenous endostatin signal, whereas no signal was detected in the Col18a1-/- mice. Thus, the transgene-derived endostatin was located in the same structures as the endogenous type XVIII collagen/endostatin. In spite of the strong expression of the K14 promoter-derived transgene in the skin, obvious changes at the light microscope level were not observed. However, at the electron microscopic level we observed abnormal thickening of the epidermal basement membrane in the transgenic mice compared to age-matched wild-type mice. The thicknesses of the BMs and the lamina densa were measured from 2.5-, 3.5- and 5.5- month-old mutant and control mice. The BM was 48.9±13.0 nm in control, 57.1±15.7 nm in G20, and 93.3±62.5 nm in J4 mice. The thicknesses of the lamina densa were 33.7±9.9 nm in control, 44.2±11.5 nm in G20, and 78.5±53.1 nm in J4 mice. The differences between control and G20 samples were statistically significant (p<0.001) as well as between control and J4 samples (p<0.001) in both groups. Particularly in the J4 mice there were areas of highly abnormal structures of the lamina densa with undulating and very thick areas. There were also regions where the lamina lucida was not detectable.

5.7 Polarized orientation of type XVIII collagen in the skin BM (III)

The localization of the transgenic endostatin to the same structures as endogenous type XVIII collagen in skin raised the possibility of dominant negative effects of transgenic endostatin. To understand the role of type XVIII collagen and endostatin in this process, we carried out immunolocalization of type XVIII collagen using two antibodies, one against the N-terminus of type XVIII collagen, ELQ, and the other against the C-terminal endostatin domain, RES. Under the light microscope, there were no clear differences in the immunofluorescene staining with the ELQ antibody between the control and transgenic skin samples. Immunoelectron microscopy (IEM) with the RES antibody revealed signals in the lamina densa of wild-type skin, and with the ELQ antibody under the lamina densa. This staining pattern suggests a polarized orientation of type XVIII collagen in the epidermal BM, with the C-terminal part of type XVIII collagen, endostatin, being localized in the lamina densa and the N-terminal part of type XVIII collagen being oriented to the dermal zone of BM. In the transgenic skin, the RES antibody can be expected to react both with endogenous type XVIII collagen and transgenic endostatin. The IEM of transgenic skin revealed strong signals and a somewhat dispersed pattern for RES, reflecting the higher endostatin level and broadened BM in endostatin-overexpressing mice. Staining with the ELQ antibody was diminished in the transgene skin compared to control samples. Moreover, the localization of the 56 signals were typically altered, gold labels being often seen deeper in the dermis of the transgenic samples compared to control samples. This staining result suggested that an excess of endostatin disturbs at least partly the localization of endogenous type XVIII collagen, possibly causing disorganization and broading of the BM. 6 Discussion

6.1 Xenopus laevis type XVIII collagen

We characterized the primary structure of X. laevis type XVIII collagen. The deduced amino acid sequence revealed a protein structure resembling its mammalian counterpart. X. laevis type XVIII collagen contained three different N-terminal variants, and all of these included an interrupted collagenous region flanked by non-collagenous domains in the N- and C-termini of the polypeptide. The common region for all three variants included a thrombosbondin homology region in the N-terminus and an endostatin homology region in the C-terminus. The short variant had its own signal sequence, indicating the usage of a separate promoter for the short variant, thus resembling the mammalian ones. The two longest variants shared the same signal sequence and N- terminal region, which were specific to these two forms. The longest variant contained, in addition, a specific cysteine-rich motif having homology with the extracellular part of frizzled proteins. The best conserved domains were the endostatin region and the cysteine-rich domain. The similarity of the X. laevis protein structure with human and mouse indicated an essential role for these domains for the function in type XVIII collagen. The short variant was first expressed during embryogenesis and was shown to be the most abundant one of these variants. Whole embryo in situ hybridizations revealed the strong expression of the short variant all over the body. In mammals the short variant has been shown to localize at BM structures in many tissues (Murakagi et al. 1995, Saarela et al. 1998a, Saarela et al. 1998b, Miosge et al. 1999). In Xenopus, expression of the short variant was first seen in the neurula stage, with a strong signal at the eye analage. Interestingly, in mice lacking type XVIII collagen, defects have been identified in the eye indicating its essential role in eye development (Fukai et al. 2002, Ylikärppä et al. 2003a, Marneros et al. 2003, Marneros et al. 2004). The expression of the middle and long variants were seen at later stages and in lower amounts compared to the short variant. The signals were seen in the head mesenchyme, branchial arches, cranial ganglia, between the somites, and in the pronephros duct. The role of type XVIII collagen in the development of these organs is not known, but the cysteine-rich domain in the longest variants raises 58 the possibility of interactions with Wnt-proteins, which are know to play crucial role in embryogenesis (Wodarz & Nusse 1998, Lin et al. 2001b).

6.2 Human type XVIII collagen

The gene encoding human type XVIII collagen was found to be about 105 kb in length. It contained 43 exons and two promoters separated by a distance of 50 kb. Promoter 1 was associated with exons 1 and 2 and drived the expression of the shortest, 1336-residue α1(XVIII) chain with the N-terminal domain NC1-303. Promoter 2, on the other hand, was associated with exon 3 and drived the expression of the 1516- and 1751-residue α1(XVIII) chains with N-terminal domains NC1-493 and -728, respectively, the two differing due to the alternative splicing of exon 3. The human gene had a structure similar to the gene described for mouse type XVIII collagen (Rehn et al. 1996). The lowest homology between the human and mouse exons was detected in exon 3, which involves sequences upstream of the frizzled motif. The investigation also led to the isolation of cDNAs for the longest N-terminal variant of the human α1(XVIII) chains. The cDNA sequences covering this variant revealed a similar amino acid composition to those characterized earlier for the mouse and frog (Rehn & Pihlajaniemi 1995, Elamaa et al. 2002). This third N-terminal variant, NC1-728, had a cysteine-rich domain, which was homologous to the extracellular ligand-binding part of the frizzled proteins known to bind Wnt signalling molecules (Bhanot et al. 1996). The Wnt signalling transduction pathway functions in various developmental processes by affecting cell proliferation, polarity, and differentiation (Wodarz & Nusse 1998). The Wnt proteins are secreted from the cell, and their receptors, the frizzled proteins, are seven-span transmembrane proteins (Wodarz & Nusse 1998). A similar cysteine-rich motif occurs in the secreted FRP/FrzB proteins, which are involved in the Wnt signalling pathway through their action in antagonizing Wnt molecules (Wodarz and Nusse 1998). Other proteins found to contain the same cysteine-rich motif include several receptor tyrosine kinases, carboxypeptidase Z, Corin, and MFRP (Saldanha et al. 1998, Masiakowski &Yancopoulos 1998, Song & Fricker 1997, Yan et al. 1999, Katoh 2001). Earlier studies suggest that the third variant of type XVIII collagen is expressed at very low levels (Saarela et al. 1998b). This was confirmed here by RNA dot blot hybridization of a wide panel of human tissues showing only faint signals, but RT-PCR reactions with specific primers for exon 3 and exon 4 of the longest variant revealed signals in fetal liver, lung, skeletal muscle, and spleen and more faintly in some other tissues. This variant was also recently shown to be expressed in human brain and retina (Suzuki et al. 2002b). This longest variant has also been shown to be the rarest one in the mouse and frog (Rehn & Pihlajaniemi 1995, Elamaa et al. 2002). An affinity-purified polyclonal antibody specific to the human NC1-728 variant of type XVIII collagen was prepared in order to study the expression of this variant further. Cells transfected with a cDNA construct directing the synthesis of full-length α1(XVIII) chains containing NC1- 728 revealed a 200-kDa band in Western blots of cell extracts. Western blotting of human serum revealed 173-kDa and 144-kDa bacterial collagenase-sensitive bands corresponding to the NC1-728 and NC1-493 variants of the α1(XVIII) chains. The 59 present data demonstrates that also both long forms of type XVIII collagen occur as soluble serum proteins. It should be noted that the polypeptides detected in the serum were positive with three N-terminal antibodies but not with the C-terminal antibody recognizing the endostatin domain, indicating that human serum contains nearly full- length NC1-728 and NC1-493 variants of the α1(XVIII) chains with an intact N-terminus and the collagenous region but lacking the endostatin region. This may either be due to unwanted proteolysis during sample preparation, or the endostatin domain may be proteolytically cleaved by the cells and tissues, producing circulating type XVIII collagen. The NC1-493 variant of type XVIII collagen, which is derived from the same promotor as NC1-728, is known to be strongly expressed in the liver (Saarela et al. 1998a, Saarela et al. 1998b, Musso et al. 2001). It is likely that the liver could be the source of the circulating NC1-493 variant as well as the NC1-728 variant, as we detected mRNAs for the latter in this tissue. Interestingly, we were also able to demonstrate here that the NC1-728 variant gives rise to a 45-kDa N-terminal fragment in cultured EBNA-293 cells that is detectable in concentrated serum-free medium from the untransfected cell cultures. This fragment was positive for all three N-terminal-specific antibodies, anti-long, anti-fz, and anti-all, demonstrating that it includes all of the NC1-728 domain-specific sequences including the sequences shared by the NC1-728 and NC1-493 variants and those shared with the shortest variant, NC1-303. The manner in which the 45-kDa fragment is produced was not studied here, but the present data suggest that the N-terminal portion may be proteolytically processed, resulting in a soluble fragment containing the frizzled motif. This raises the intriguing possibility that this fragment derived from type XVIII collagen may antagonize Wnt molecules in a manner similar to that shown for secreted FRP/FrzBs (Ladher et al. 2000). The clearest immunostaining for the NC1-728 variant in fetal tissues was seen in the developing lung and muscle. These signals coincide with type XVIII collagen staining observed previously in lung and muscle using antibodies detecting all type XVIII collagen variants (Saarela et al. 1998a, Sajnani-Perez et al. 2003, Miosge et al. 2003). In an earlier stage of mouse lung development, type XVIII collagen has been shown to localize to the epithelial bud and the stalk region, but it is lost at a later stage from the stalk region while persisting in the epithelial bud (Lin et al. 2001b). The localization of the longest variant of type XVIII collagen under the epithelial cells and on their luminal side in the developing lung of a 16-week human fetus supports the finding that type XVIII collagen is involved in the process of lung development. In the forming myotubes the signal was seen as a granular or dot-like pattern, accumulating in the junction region between the myotubes and the surrounding connective tissue. These cell-associated staining patterns could reflect structural functions and possibly also signalling roles for the longest variant of type XVIII collagen. The Wnt pathway is known to be active in the developing lung and to participate in epithelial-mesenchymal interactions. Interestingly, some of the components of the Wnt signalling pathway show similar expression patterns in the developing lung (Tebar et al. 2001) as seen here for type XVIII collagen. Previous studies with mice suggest that type XVIII collagen is involved in the regulation of differential epithelial morphogenesis of early lung and kidney development (Lin et al. 2001b). The soluble type XVIII collagen- derived fragment containing the frizzled motif and the cell associated staining pattern in 60 human tissues, both identified here, and co-localization of staining with Wnt molecules suggest that the frizzled-containing N-terminal fragment of type XVIII collagen may be involved in Wnt signalling.

6.3 Transgenic mice overexpressing endostatin

Transgenic mice overexpressing endostatin were generated in order to assess the physiological role of this special C-terminal domain of type XVIII collagen. To this end, a cDNA encoding endostatin was inserted under the keratin K14 promoter known to drive expression mainly in the basal cells of the epidermis (Vassar et al. 1989). The transgenic endostatin started with residues HTH, which were identified as the N-terminal end of endostatin isolated from EOMA-conditioned medium (O´Reilly et al. 1997). The construct did not encode the association and hinge domains of the type XVIII collagen NC1 domain (Sasaki et al. 1998), thus leading to the synthesis of monomeric endostatin. Three independent transgenic mouse lines were generated, varying in the expression levels of transgenic endostatin. The phenotypic consequences of the overexpression were found to reflect the expression levels. The J4 mice were characterized by lens opacity due to progressing nuclear cataract. Light microscopy revealed the progressive occurrence of vacuoles in the epithelial cell layer, and in ultrastructural analysis there were ruptures at the interphase of the lens epithelial cells and the lens capsule. In older mice, the epithelial cells became spindle- shaped and formed clusters at the anterior side of the lens. This process resembled TGFβ- induced epithelial-mesenchymal transition (McAvoy et al. 1999). Prompted by the strong lens phenotype in the transgenic mice, we examined the expression of the transgene in the eye. The K14 promoter is known to be expressed in the lens epithelium in neonates (Nguyen et al. 2002). RT-PCR, Western blotting, and immunofluorescence staining all confirmed the presence of transgenic endostatin in the mouse lens. The expected 20-kDa endostatin fragment was detected in lens homogenates of lines G20 and J4, but not in line K18, which gave only weak RT-PCR signals. Comparison of the location of the transgenic endostatin with that of the endogenous gene product revealed both in the lens capsule, but in higher amounts and present more broadly throughout the capsule in the transgenic samples. The lens capsule is a specialized BM providing support for the attachment of epithelial cells (Bassnett et al. 1999, Zampighi et al. 2000). The overexpression of endostatin in the lens appeared to first lead to the loosening of the epithelial cell contact from the lens capsule, and at later stages to the abnormal proliferation and clustering of the cells. Also the capsule appeared thickened. The excess of endostatin and its abnormal localization could cause disorganization of the capsule. However, the lens capsule was not ruptured in the transgenic mice. Integrin α5β1 has been suggested as a functional receptor for endostatin (Rehn et al. 2001, Sudhakar et al. 2003), and endostatin is known to bind to glycosaminoglycans in a specific manner (Kreuger et al. 2002) including binding with the cell surface heparan sulphate proteoglycans, glypicans 1 and 4 (Karumanchi et al. 2001). Endostatin is also known to cause the disassembly of the actin cytoskeleton via downregulation of RhoA activity (Wickström et al. 2003). When expressed under the 61 K14 promoter, oncogenes E6 and E7, which affect the cell cycle protein p53 and the retinoblastoma protein (pRB), cause the deregulation of the cell cycle of lens epithelial cells and lead to cataract development in mice (Nguyen et al. 2002). Along these lines it is of interest to note that endostatin can cause the cell cycle arrest of endothelial cells through the inhibition of cyclin D1 (Hanai et al. 2002a). The changes affecting the structure of the lens capsule in the endostatin overexpressing mice could affect the ability of the epithelial cells to anchor to the capsule, possibly via integrins or cell surface heparan sulphate proteoglycans, leading to the occurance of vacuoles and the clustering of cells. It is also conceivable that excess endostatin leads to alterations in intracellular signalling and thereby to changes in epithelial cell viability. Western blotting analysis of transgene expression in the skin revealed an approximately 20-kDa fragment. The expression level was highest in line J4, whereas in the K18 line the transgene-derived signal was comparable to the endogenous level. In contrast to the lens, transgene expression in the skin did not lead to obvious defects detectable at the light microscopic level. The ultrastructural analysis revealed a broadened epidermal BM in lines J4 and G20. In J4 mice, the thickness of the lamina densa of the epidermal BM was twice that of controls, and in some areas it was distinctly abnormal in appearance. Nevertheless, the phenotypic consequences were less severe in the case of the skin compared to the lens phenotype. It has been shown that a lack of type XVIII collagen leads to broadening of the epithelial BM of the iris and ciliary body and an accumulation of extracellular material has been detected in the irises of Col18a1-/- mice by immunofluorescence and EM studies (Ylikärppä et al. 2003a). The lack of type XVIII collagen in mice also results in ultrastructural changes in the BMs of the choroids plexuses in the brain, BMs in the epidermis, heart atrioventricular valves and kidney proximal tubules (Utriainen et al. 2004). We show here that overexpression of the endostatin domain similarly led to broadening of the skin BM. We used antibodies against major BM components to test whether the broadening of the BM observed in skin is due to dispersion or compensatory overproduction of some other BM components, but immunostaining of some BM components (perlecan, type IV collagen, nidogen-1, nidogen-2, fibulin-2, and collagen VII) did not reveal any changes in their expression patterns between the transgenic and wild-type samples in either light or electron microscopy. However, the polarized location of type XVIII collagen molecules, with the C-terminal endostatin domain being embedded in the BM and the N-terminal non-collagenous domain occurring at the BM-fibrillar matrix interphase, and the apparent displacement of type XVIII collagen molecules to the fibrillar matrix in the event of overexpression of monomeric endostatin are intruiging findings. It is known that the type XVIII collagen NC1 domain, which contains the endostatin domain in trimeric form, binds the BM proteins laminin-1 and perlecan more strongly than does monomeric endostatin, while both bind fibulin-1, fibulin-2, and heparin equally well (Sasaki et al. 1998). It is thus conceivable that the large amount of transgene-derived endostatin in the mice that overexpress this component may compete for binding sites with the full-length type XVIII collagen molecules, leading to their displacement. The broadening of the BMs in the absence of this molecule and the apparent phenocopying of this BM defect when monomeric endostatin is overexpressed in mice synthesizing the full-length type XVIII collagen molecules are suggestive of a molecular explanation for the role of type XVIII 62 collagen. We propose that this collagen, and in particular its trimeric endostatin domain, has a role in binding other BM components together, leading to a tightening of the network of BM molecules. When monomeric endostatin is introduced, it competes with these binding sites, but is not as efficient as trimeric endostatin in supporting the molecular network. Alternatively, displacement of the full-length type XVIII collagen molecules affects the anchorage of the lamina densa to the dermis. Thus, both a lack of type XVIII collagen and overexpression of its endostatin domain could affect the structural integrity of the BMs. Further work is needed to validate this model and to identify the key binding partners. It can also be envisaged that the proposed model may help us to understand the multiple roles that type XVIII collagen appears to have, not only in supporting BM integrity, but also in potentiating the destabilization of BMs, for example in situations of BM synthesis and degradation when proteolytic enzymes such as the matrix metalloproteinases undergo increased expression. These enzymes are known to be able to release monomeric endostatin (Wen et al. 1999, Felbor et al. 2000, Ferreras et al. 2000). 7 Future perspectives

Type XVIII collagen is a multi-domain protein including functional domains which can be released from the parent molecule, namely the endostatin domain and, as suggested here, the cysteine-rich frizzled-like domain. The fact that type XVIII collagen is conserved between species ranging from C. elegans to X. laevis and mammals suggests that it has a central role in biological processes. The expression patterns of different variants of type XVIII collagen in early X. laevis embryogenesis suggest independent roles for the variants. New techniques make it possible to study the functions of the different variants in the frog model system. Using the morpholino antisense technique, it is possible to block the translation of specific target genes. This technique is particularly well-suited to study the function of genes expressed in early embryogenesis. In the case of type XVIII collagen, antisense morpholino treatment could shed light on the role of type XVIII collagen in the early phases of embryogenesis. In addition, the frog model system gives the opportunity to test the interactions between different proteins by expressing them together in the developing frog embryo. In particular, the frog system has been used successfully to test the players of the Wnt-signalling cascade. In the case of type XVIII collagen, it would be interesting to identify the Wnt-proteins possibly interacting with the cysteine-rich motif in type XVIII collagen based on their expression patterns, e.g. in pronephric kidney, and test the interactions by expressing different combinations of Wnt-proteins and type XVIII collagen. The elucidation of the complete structure of the human type XVIII collagen gene gives the tools to screen for mutations in patient samples for candidate diseases. The role of the long variant containing the cysteine-rich motif and an apparent proteolytic cleavage site raise questions about the enzymes which are involved in this process and the role of the cleavage. The conservation of this motif and expression of the long variant during embryogenesis, although at low levels, suggest a role for this variant in organogenesis. The animal models, lacking or overexpressing this variant, could help to understand the function of the cysteine-rich domain. Endostatin, the C-terminal proteolytic fragment of type XVIII collagen, has been subject to much interest because of its ability to inhibit tumor growth by inhibiting angiogenesis. The high conservation of the endostatin domain of type XVIII collagen in avascular organisms, like C. elegans, suggests that it is more than just a player in the 64 regulation of angiogenesis. The existence and conservation of endostatin throughout the phyla are suggestive of an essential role for it which already exists in the lower organisms. The polarized orientation of type XVIII collagen in BMs and the localization of the endostatin part to the lamina densa raise the question about binding parters for endostatin. The affinity in vitro of endostatin for heparan sulphate chains could signify potential partners as well as its binding with certain proteins in the BM. Altogether, although numerous actions have been identified for endostatin in various studies, further work is needed to understand its significance in normal tissues and in pathological conditions. Our data suggest that monomeric endostatin, in addition to directly affecting cellular processes, may interfere with the structural role of type XVIII collagen in BMs. References

Abdollahi A, Hahnfeldt P, Maercker C, Grone HJ, Debus J, Ansorge W, Folkman J, Hlatky L & Huber PE (2004) Endostatin's antiangiogenic signaling network Mol Cell13:649-63. Ackley BD, Crew JR, Elamaa H, Pihlajaniemi T, Kuo CJ & Kramer JM (2001) The NC1/endostatin domain of Caenorhabditis elegans type XVIII collagen affects cell migration and axon guidance. J Cell Biol 152: 1219-32. Ackley BD, Kang SH, Crew JR, Suh C, Jin Y & Kramer JM (2003) The basement membrane components nidogen and type XVIII collagen regulate organization of neuromuscular junctions in Caenorhabditis elegans. J Neurosci 23: 3577-87. Afzal AR, Rajab A, Fenske CD, Oldridge M, Elanko N, Ternes-Pereira E, Tuysuz B, Murday VA, Patton MA, Wilkie AO & Jeffery S (2000) Recessive Robinow syndrome, allelic to dominant brachydactyly type B, is caused by mutation of ROR2. Nat Genet 25: 419-22. Arikawa-Hirasawa E, Watanabe H, Takami H, Hassell JR & Yamada Y (1999) Perlecan is essential for cartilage and cephalic development. Nat Genet 23: 354-8. Aricescu AR, McKinnell IW, Halfter W & Stoker AW (2002) Heparan sulfate proteoglycans are ligands for receptor protein tyrosine phosphatase sigma. Mol Cell Biol 22: 1881-92 Aumailley M, El Khal A, Knoss N, Tunggal L (2003) Laminin 5 processing and its integration into the ECM. Matrix Biol 22: 49-54. Aumailley M, Pesch M, Tunggal L, Gaill F & Fässler R (2000) Altered synthesis of laminin 1 and absence of basement membrane component deposition in (beta)1 integrin-deficient embryoid bodies. J Cell Sci 113: 259-68. Aumailley M & Smyth N (1998) The role of laminins in basement membrane function. J Anat 193:1-21. Aumailley M, Wiedemann H, Mann K & Timpl R (1989) Binding of nidogen and the laminin- nidogen complex to basement membrane collagen type IV. Eur J Biochem 184: 241-8. Bassnett S, Missey H & Vucemilo I (1999) Molecular architecture of the lens fiber cell basal membrane complex. J Cell Sci 112: 2155-65. Battaglia C, Mayer U, Aumailley M & Timpl R (1992) Basement-membrane heparan sulfate proteoglycan binds to laminin by its heparan sulfate chains and to nidogen by sites in the protein core. Eur J Biochem 208: 359-66. Beck CW & Slack JM (2001) An amphibian with ambition: a new role for Xenopus in the 21st century. Genome Biol 2: 1029.1-1029.5 Bergers G, Javaherian K, Lo KM, Folkman J & Hanahan D (1999) Effects of angiogenesis inhibitors on multistage carcinogenesis in mice. Science 284: 808-12. 66 Bezakova G & Ruegg MA (2003) New insights into the roles of agrin. Nat Rev Mol Cell Biol 4: 295-308. Bhanot P, Brink M, Samos CH, Hsieh JC, Wang Y, Macke JP, Andrew D, Nathans J & Nusse R (1996) A new member of the frizzled family from Drosophila functions as a Wingless receptor. Nature 382: 225-30. Blackhall FH, Merry CL, Lyon M, Jayson GC, Folkman J, Javaherian K & Gallagher JT (2003) Binding of endostatin to endothelial heparan sulphate shows a differential requirement for specific sulphates. Biochem J 375: 131-9. Bono P, Teerenhovi L & Joensuu H (2003) Elevated serum endostatin is associated with poor outcome in patients with non-Hodgkin lymphoma. Cancer 97: 2767-75. Bouvard D, Brakebusch C, Gustafsson E, Aszodi A, Bengtsson T, Berna A & Fässler R (2001) Functional consequences of integrin gene mutations in mice. Circ Res 89: 211-23. Brakebusch C, Grose R, Quondamatteo F, Ramirez A, Jorcano JL, Pirro A, Svensson M, Herken R, Sasaki T, Timpl R, Werner S & Fässler R (2000) Skin and hair follicle integrity is crucially dependent on beta 1 integrin expression on keratinocytes. EMBO J 19: 3990-4003. Brodsky B & Shah NK (1995) Protein motifs. 8. The triple-helix motif in proteins.FASEB J : 1537- 46. Brooks PC, Silletti S, von Schalscha TL, Friedlander M & Cheresh DA (1998) Disruption of angiogenesis by PEX, a noncatalytic metalloproteinase fragment with integrin binding activity. Cell 92:391-400. Brown JC & Timpl R (1995) The collagen superfamily. Int Arch Allergy Immunol 107: 484-90. Brown NH (2000) Cell-cell adhesion via the ECM: integrin genetics in fly and worm. Matrix Biol 19: 191-201. Calvo A, Feldman AL, Libutti SK & Green JE (2002) Adenovirus-mediated endostatin delivery results in inhibition of mammary gland tumor growth in C3(1)/SV40 T-antigen transgenic mice. Cancer Res 62: 3934-8. Celerier J, Cruz A, Lamande N, Gasc JM & Corvol P (2002) Angiotensinogen and its cleaved derivatives inhibit angiogenesis. Hypertension 39: 224-8. Chang Z, Choon A & Friedl A (1999a) Endostatin binds to blood vessels in situ independent of heparan sulfate and does not compete for fibroblast growth factor-2 binding. Am J Pathol 155: 71-6. Chen QR, Kumar D, Stass SA & Mixson AJ (1999) Liposomes complexed to plasmids encoding angiostatin and endostatin inhibit breast cancer in nude mice. Cancer Res 59: 3308-12. Cline EI, Bicciato S, DiBello C & Lingen MW (2002) Prediction of in vivo synergistic activity of antiangiogenic compounds by gene expression profiling. Cancer Res 62: 7143-8. Colognato H & Yurchenco PD (2000) Form and function: the laminin family of heterotrimers. Dev Dyn. 2000 Jun;218(2):213-34. 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: 2520-6. Cooper AR & MacQueen HA (1983)Subunits of laminin are differentially synthesized in mouse eggs and early embryos. Dev Biol. 1983 Apr;96(2):467-71. Cosgrove D, Meehan DT, Grunkemeyer JA, Kornak JM, Sayers R, Hunter WJ & Samuelson GC (1996) Collagen COL4A3 knockout: a mouse model for autosomal Alport syndrome. Genes Dev 10:2981-92. Costell M, Gustafsson E, Aszodi A, Morgelin M, Bloch W, Hunziker E, Addicks K, Timpl R & Fässler R (1999) Perlecan maintains the integrity of cartilage and some basement membranes. J Cell Biol 147: 1109-22. 67 D'Amore PA & Ng YS (2002) Tales of the cryptic: unveiling more angiogenesis inhibitors. Trends Mol Med 8: 313-5. de Sa RO & Hillis DM (1990) Phylogenetic relationships of the pipid frogs Xenopus and Silurana: an integration of ribosomal DNA and morphology. Mol Biol Evol 7: 365-76. DeChiara TM, Bowen DC, Valenzuela DM, Simmons MV, Poueymirou WT, Thomas S, Kinetz E, Compton DL, Rojas E, Park JS, Smith C, DiStefano PS, Glass DJ, Burden SJ & Yancopoulos GD (1996) The receptor tyrosine kinase MuSK is required for neuromuscular junction formation in vivo. Cell 85: 501-12. DeChiara TM, Kimble RB, Poueymirou WT, Rojas J, Masiakowski P, Valenzuela DM & Yancopoulos GD (2000) Ror2, encoding a receptor-like tyrosine kinase, is required for cartilage and growth plate development. Nat Genet 24: 271-4. Deininger MH, Fimmen BA, Thal DR, Schluesener HJ & Meyermann R (2002) Aberrant neuronal and paracellular deposition of endostatin in brains of patients with Alzheimer's disease. J Neurosci 22: 10621-6. Deininger MH, Wybranietz WA, Graepler FT, Lauer UM, Meyermann R & Schluesener HJ (2003) Endothelial endostatin release is induced by general cell stress and modulated by the nitric oxide/cGMP pathway. FASEB J 17: 1267-76 Dhanabal M, Ramchandran R, Volk R, Stillman IE, Lombardo M, Iruela-Arispe ML, Simons M & Sukhatme VP (1999a) Endostatin: yeast production, mutants, and antitumor effect in renal cell carcinoma. Cancer Res 59: 189-97. Dhanabal M, Ramchandran R, Waterman MJ, Lu H, Knebelmann B, Segal M & Sukhatme VP (1999b) Endostatin induces endothelial cell apoptosis. J Biol Chem 274: 11721-6. Dhanabal M, Volk R, Ramchandran R, Simons M & Sukhatme VP (1999c) Cloning, expression, and in vitro activity of human endostatin. Biochem Biophys Res Commun 258: 345-52. Ding I, Sun JZ, Fenton B, Liu WM, Kimsely P, Okunieff P & Min W (2001) Intratumoral administration of endostatin plasmid inhibits vascular growth and perfusion in MCa-4 murine mammary carcinomas. Cancer Res 61: 526-31. Ding YH, Javaherian K, Lo KM, Chopra R, Boehm T, Lanciotti J, Harris BA, Li Y, Shapiro R, Hohenester E, Timpl R, Folkman J & Wiley DC (1998) Zinc-dependent dimers observed in crystals of human endostatin. Proc Natl Acad Sci U S A 95: 10443-8. DiPersio CM, Hodivala-Dilke KM, Jaenisch R, Kreidberg JA & Hynes RO (1997) alpha3beta1 Integrin is required for normal development of the epidermal basement membrane. J Cell Biol 137: 729-42. Dixelius J, Cross M, Matsumoto T, Sasaki T, Timpl R & Claesson-Welsh L (2002) Endostatin regulates endothelial cell adhesion and cytoskeletal organization. Cancer Res 62: 1944-7. Dong L, Chen Y, Lewis M, Hsieh JC, Reing J, Chaillet JR, Howell CY, Melhem M, Inoue S, Kuszak JR, DeGeest K & Chung AE (2002) Neurologic defects and selective disruption of basement membranes in mice lacking entactin-1/nidogen-1. Lab Invest 82: 1617-30. Dong S, Cole GJ & Halfter W (2003) Expression of collagen XVIII and localization of its glycosaminoglycan attachment sites. J Biol Chem 278: 1700-7. Dowling J, Yu QC & Fuchs E (1996) Beta4 integrin is required for hemidesmosome formation, cell adhesion and cell survival. J Cell Biol 134: 559-72. Dziadek M & Timpl R (1985) Expression of nidogen and laminin in basement membranes during mouse embryogenesis and in teratocarcinoma cells. Dev Biol 111: 372-82. Ekblom P, Lonai P & Talts JF (2003) Expression and biological role of laminin-1. Matrix Biol 22: 35-47. Eklund L, Muona A, Lietard J & Pihlajaniemi T (2000) Structure of the mouse type XV collagen gene, Col15a1, comparison with the human COL15A1 gene and functional analysis of the promoters of both genes. Matrix Biol 19: 489-500. 68 Eklund L, Piuhola J, Komulainen J, Sormunen R, Ongvarrasopone C, Fässler R, Muona A, Ilves M, Ruskoaho H, Takala TE & Pihlajaniemi T (2001) Lack of type XV collagen causes a skeletal myopathy and cardiovascular defects in mice. Proc Natl Acad Sci U S A 98: 1194-9. Engbring JA & Kleinman HK (2003) The basement membrane matrix in malignancy. J Pathol 200: 465-70. Ergun S, Kilic N, Wurmbach JH, Ebrahimnejad A, Fernando M, Sevinc S, Kilic E, Chalajour F, Fiedler W, Lauke H, Lamszus K, Hammerer P, Weil J, Herbst H & Folkman J (2001) Endostatin inhibits angiogenesis by stabilization of newly formed endothelial tubes. Angiogenesis 4: 193-206. Erickson AC & Couchman JR (2000) Still more complexity in mammalian basement membranes. J Histochem Cytochem 48: 1291-306. Eriksson K, Magnusson P, Dixelius J, Claesson-Welsh L & Cross MJ (2003) Angiostatin and endostatin inhibit endothelial cell migration in response to FGF and VEGF without interfering with specific intracellular signal transduction pathways. FEBS Lett 536: 19-24. Ettner N, Gohring W, Sasaki T, Mann K & Timpl R (1998) The N-terminal globular domain of the laminin alpha1 chain binds to alpha1beta1 and alpha2beta1 integrins and to the heparan sulfate- containing domains of perlecan. FEBS Lett 430: 217-21. Felbor U, Dreier L, Bryant RA, Ploegh HL, Olsen BR & Mothes W (2000) Secreted cathepsin L generates endostatin from collagen XVIII. EMBO J 19: 1187-94. Feldman AL, Alexander HR Jr, Bartlett DL, Kranda KC, Miller MS, Costouros NG, Choyke PL & Libutti SK (2001a) A prospective analysis of plasma endostatin levels in colorectal cancer patients with liver metastases. Ann Surg Oncol 8: 741-5. Feldman AL, Alexander HR Jr, Yang JC, Linehan WM, Eyler RA, Miller MS, Steinberg SM & Libutti SK (2002) Prospective analysis of circulating endostatin levels in patients with renal cell carcinoma. Cancer 95: 1637-43. Feldman AL, Pak H, Yang JC, Alexander HR Jr & Libutti SK (2001b) Serum endostatin levels are elevated in patients with soft tissue sarcoma. Cancer 91: 1525-9. Feldman AL, Restifo NP, Alexander HR, Bartlett DL, Hwu P, Seth P & Libutti SK (2000a) Antiangiogenic gene therapy of cancer utilizing a recombinant adenovirus to elevate systemic endostatin levels in mice. Cancer Res 60: 1503-6. Feldman AL, Tamarkin L, Paciotti GF, Simpson BW, Linehan WM, Yang JC, Fogler WE, Turner EM, Alexander HR Jr & Libutti SK (2000b) Serum endostatin levels are elevated and correlate with serum vascular endothelial growth factor levels in patients with stage IV clear cell renal cancer. Clin Cancer Res 6: 4628-34. Ferrara N, Clapp C & Weiner R (1991) The 16K fragment of prolactin specifically inhibits basal or fibroblast growth factor stimulated growth of capillary endothelial cells. Endocrinology 129: 896-900. Ferreras M, Felbor U, Lenhard T, Olsen BR & Delaisse J (2000) Generation and degradation of human endostatin proteins by various proteinases. FEBS Lett 486: 247-51. Fox JW, Mayer U, Nischt R, Aumailley M, Reinhardt D, Wiedemann H, Mann K, Timpl R, Krieg T & Engel J (1991) Recombinant nidogen consists of three globular domains and mediates binding of laminin to collagen type IV. EMBO J 10: 3137-46. Fukai N, Eklund L, Marneros AG, Oh SP, Keene DR, Tamarkin L, Niemela M, Ilves M, Li E, Pihlajaniemi T & Olsen BR (2002) Lack of collagen XVIII/endostatin results in eye abnormalities. EMBO J 21: 1535-44. Francis PJ, Berry V, Moore AT & Bhattacharya S (1999) Lens biology: development and human cataractogenesis. Trends Genet 15: 191-6. Friedrich MV, Gohring W, Morgelin M, Brancaccio A, David G & Timpl R (1999) Structural basis of glycosaminoglycan modification and of heterotypic interactions of perlecan domain V. J Mol Biol 294: 259-70. 69 Fässler R & Meyer M (1995) Consequences of lack of beta 1 integrin gene expression in mice. Genes Dev 9: 1896-908. Gautam M, Noakes PG, Moscoso L, Rupp F, Scheller RH, Merlie JP & Sanes JR (1996) Defective neuromuscular synaptogenesis in agrin-deficient mutant mice. Cell 85: 525-35. Georges-Labouesse E, Messaddeq N, Yehia G, Cadalbert L, Dierich A & Le Meur M (1996) Absence of integrin alpha 6 leads to epidermolysis bullosa and neonatal death in mice. Nat Genet 13: 370-3. Ghohestani RF, Li K, Rousselle P & Uitto J (2001) Molecular organization of the cutaneous basement membrane zone. Clin Dermatol 19: 551-62. Gilmour DT, Lyon GJ, Carlton MB, Sanes JR, Cunningham JM, Anderson JR, Hogan BL, Evans MJ & Colledge WH (1998) Mice deficient for the secreted glycoprotein SPARC/osteonectin/BM40 develop normally but show severe age-onset cataract formation and disruption of the lens. EMBO J 17: 1860-70. Glenjen N, Mosevoll KA & Bruserud O (2002) Serum levels of angiogenin, basic fibroblast growth factor and endostatin in patients receiving intensive chemotherapy for acute myelogenous leukemia. Int J Cancer 101: 86-94. Halfter W, Dong S, Schurer B & Cole GJ (1998) Collagen XVIII is a basement membrane heparan sulfate proteoglycan. J Biol Chem 273: 25404-12. Halfter W, Dong S, Schurer B, Osanger A, Schneider W, Ruegg M & Cole GJ (2000) Composition, synthesis, and assembly of the embryonic chick retinal basal lamina. Dev Biol 220: 111-28. Hamburger V (1988) The Heritage of Experimental Embryology. Oxford University Press, Oxford. Hanai J, Dhanabal M, Karumanchi SA, Albanese C, Waterman M, Chan B, Ramchandran R, Pestell R & Sukhatme VP (2002a) Endostatin causes G1 arrest of endothelial cells through inhibition of cyclin D1. J Biol Chem 277: 16464-9. Hanai J, Gloy J, Karumanchi SA, Kale S, Tang J, Hu G, Chan B, Ramchandran R, Jha V, Sukhatme VP & Sokol S (2002b) Endostatin is a potential inhibitor of Wnt signaling. J Cell Biol 158: 529-39. Harland, RM (1991) In situ hybridization: an improved whole mount method for Xenopus laevis. Methods Enzymol 36: 685-697. Hata K, Fujiwaki R, Nakayama K & Miyazaki K (2001) Expression of the Endostatin gene in epithelial ovarian cancer. Clin Cancer Res 7: 2405-9. Hausen P & Riebesell M ( 1991) The early development of Xenopus laevis. Springer-Verlag, Berlin Herbst RS, Hess KR, Tran HT, Tseng JE, Mullani NA, Charnsangavej C, Madden T, Davis DW, McConkey DJ, O'Reilly MS, Ellis LM, Pluda J, Hong WK & Abbruzzese JL (2002) Phase I study of recombinant human endostatin in patients with advanced solid tumors. J Clin Oncol 20: 3792-803. Hirsch N, Zimmerman LB & Grainger RM (2002) Xenopus, the next generation: X. tropicalis genetics and genomics. Dev Dyn 225:422-33. Hohenester E, Sasaki T, Mann K & Timpl R (2000) Variable zinc coordination in endostatin. J Mol Biol 297: 1-6. Hohenester E, Sasaki T, Olsen BR & Timpl R (1998) Crystal structure of the angiogenesis inhibitor endostatin at 1.5 A resolution. EMBO J 17: 1656-64. Hopf M, Gohring W, Mann K & Timpl R (2001) Mapping of binding sites for nidogens, fibulin-2, fibronectin and heparin to different IG modules of perlecan. J Mol Biol 311:529-41 Hudson BG, Tryggvason K, Sundaramoorthy M & Neilson EG (2003) Alport's syndrome, Goodpasture's syndrome, and type IV collagen. N Engl J Med 348: 2543-56. Huebner K, Cannizzaro LA, Jabs EW, Kivirikko S, Manzone H, Pihlajaniemi T & Myers JC (1992) Chromosomal assignment of a gene encoding a new collagen type (COL15A1) to 9q21 --> q22. Genomics 14: 220-4. 70 Hynes RO (2002) Integrins: bidirectional, allosteric signaling machines. Cell 110: 673-87. Hynes RO & Zhao Q (2000) The evolution of cell adhesion. J Cell Biol 150: 89-96. Hägg PM, Horelli-Kuitunen N, Eklund L, Palotie A & Pihlajaniemi T (1997) Cloning of mouse type XV collagen sequences and mapping of the corresponding gene to 4B1-3. Comparison of mouse and human alpha 1 (XV) collagen sequences indicates divergence in the number of small collagenous domains. Genomics 45: 31-41. Hägg PM, Hägg PO, Peltonen S, Autio-Harmainen H & Pihlajaniemi T (1997) Location of type XV collagen in human tissues and its accumulation in the interstitial matrix of the fibrotic kidney. Am J Pathol 150: 2075-86. Hägg PM, Muona A, Lietard J, Kivirikko S & Pihlajaniemi T (1998) Complete exon-intron organization of the human gene for the alpha1 chain of type XV collagen (COL15A1) and comparison with the homologous COL18A1 gene. J Biol Chem 273: 17824-31. Iughetti P, Suzuki O, Godoi PH, Alves VA, Sertie AL, Zorick T, Soares F, Camargo A, Moreira ES, di Loreto C, Moreira-Filho CA, Simpson A, Oliva G & Passos-Bueno MR (2001) A polymorphism in endostatin, an angiogenesis inhibitor, predisposes for the development of prostatic adenocarcinoma. Cancer Res 61: 7375-8. Iversen PO, Sorensen DR & Benestad HB (2002) Inhibitors of angiogenesis selectively reduce the malignant cell load in rodent models of human myeloid leukemias. Leukemia 16: 376-81. Javaherian K, Park SY, Pickl WF, LaMontagne KR, Sjin RT, Gillies S & Lo KM (2002) Laminin modulates morphogenic properties of the collagen XVIII endostatin domain. J Biol Chem 277: 45211-8. Jiang X & Couchman JR (2003) Perlecan and tumor angiogenesis. J Histochem Cytochem 51: 1393-410. John H, Preissner KT, Forssmann WG, Standker L (1999) Novel glycosylated forms of human plasma endostatin and circulating endostatin-related fragments of collagen XV. Biochemistry 38: 10217-24. Kalluri R (2003) Basement membranes: structure, assembly and role in tumour angiogenesis. Nat Rev Cancer 3: 422-33. Kameya S, Hawes NL, Chang B, Heckenlively JR, Naggert JK & Nishina PM (2002) Mfrp, a gene encoding a frizzled related protein, is mutated in the mouse retinal degeneration 6. Hum Mol Genet 11: 1879-86. 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: 1209-15 Karihaloo A, Karumanchi SA, Barasch J, Jha V, Nickel CH, Yang J, Grisaru S, Bush KT, Nigam S, Rosenblum ND, Sukhatme VP & Cantley LG (2001) Endostatin regulates branching morphogenesis of renal epithelial cells and ureteric bud. Proc Natl Acad Sci U S A 98: 12509- 14. Karumanchi SA, Jha V, Ramchandran R, Karihaloo A, Tsiokas L, Chan B, Dhanabal M, Hanai JI, Venkataraman G, Shriver Z, Keiser N, Kalluri R, Zeng H, Mukhopadhyay D, Chen RL, Lander AD, Hagihara K, Yamaguchi Y, Sasisekharan R, Cantley L & Sukhatme VP (2001) Cell surface glypicans are low-affinity endostatin receptors. Mol Cell 7: 811-22. Kato T, Chang JH & Azar DT (2003) Expression of type XVIII collagen during healing of corneal incisions and keratectomy wounds. Invest Ophthalmol Vis Sci 44: 78-85. Katoh M (2001) Molecular cloning and characterization of MFRP, a novel gene encoding a membrane-type Frizzled-related protein. Biochem Biophys Res Commun 282: 116-23. Kawashima H, Watanabe N, Hirose M, Sun X, Atarashi K, Kimura T, Shikata K, Matsuda M, Ogawa D, Heljasvaara R, Rehn M, Pihlajaniemi T & Miyasaka M (2003) Collagen XVIII, a basement membrane heparan sulfate proteoglycan, interacts with L-selectin and monocyte chemoattractant protein-1. J Biol Chem 278: 13069-76. 71 Keezer SM, Ivie SE, Krutzsch HC, Tandle A, Libutti SK & Roberts DD (2003) Angiogenesis inhibitors target the endothelial cell cytoskeleton through altered regulation of heat shock protein 27 and cofilin. Cancer Res 63: 6405-12. Kim YM, Jang JW, Lee OH, Yeon J, Choi EY, Kim KW, Lee ST & Kwon YG (2000) Endostatin inhibits endothelial and tumor cellular invasion by blocking the activation and catalytic activity of matrix metalloproteinase. Cancer Res 60: 5410-3. Kimura N, Toyoshima T, Kojima T & Shimane M (1998) Entactin-2: a new member of basement membrane protein with high homology to entactin/nidogen. Exp Cell Res 241:36-45. Kivirikko KI (1993) Collagens and their abnormalities in a wide spectrum of diseases. Ann Med 25: 113-26. Kivirikko S, Heinämaki P, Rehn M, Honkanen N, Myers JC & Pihlajaniemi T (1994) Primary structure of the alpha 1 chain of human type XV collagen and exon-intron organization in the 3' region of the corresponding gene. J Biol Chem 269: 4773-9. Kivirikko S, Saarela J, Myers JC, Autio-Harmainen H & Pihlajaniemi T (1995) Distribution of type XV collagen transcripts in human tissue and their production by muscle cells and fibroblasts. Am J Pathol 147: 1500-9. Kliemann SE, Waetge RT, Suzuki OT, Passos-Bueno MR & Rosemberg S (2003) Evidence of neuronal migration disorders in Knobloch syndrome: clinical and molecular analysis of two novel families. Am J Med Genet 119: 15-9. Knobloch WH & Layer JM (1971) J Pediatr Ophthalmol 8: 181-184. Kohfeldt E, Sasaki T, Gohring W & Timpl R (1998) Nidogen-2: a new basement membrane protein with diverse binding properties. J Mol Biol 282:99-109. Kostka G, Giltay R, Bloch W, Addicks K, Timpl R, Fässler R & Chu ML (2001) Perinatal lethality and endothelial cell abnormalities in several vessel compartments of fibulin-1-deficient mice. Mol Cell Biol 21: 7025-34. Krag S & Andreassen TT (2003) Mechanical properties of the human lens capsule. Prog Retin Eye Res 22: 749-67. Kranenburg O, Bouma B, Kroon-Batenburg LM, Reijerkerk A, Wu YP, Voest EE & Gebbink MF (2002) Tissue-type plasminogen activator is a multiligand cross-beta structure receptor. Curr Biol 12: 1833-9. Kranenburg O, Kroon-Batenburg LM, Reijerkerk A, Wu YP, Voest EE & Gebbink MF (2003) Recombinant endostatin forms amyloid fibrils that bind and are cytotoxic to murine neuroblastoma cells in vitro. FEBS Lett 539: 149-55. Kreidberg JA, Donovan MJ, Goldstein SL, Rennke H, Shepherd K, Jones RC & Jaenisch R (1996) Alpha 3 beta 1 integrin has a crucial role in kidney and lung organogenesis. Development 122: 3537-47. Kreuger J, Matsumoto T, Vanwildemeersch M, Sasaki T, Timpl R, Claesson-Welsh L, Spillmann D & Lindahl U (2002) Role of heparan sulfate domain organization in endostatin inhibition of endothelial cell function. EMBO J 21: 6303-11. Kuhn K (1995) Basement membrane (type IV) collagen. Matrix Biol 14:439-45. Kuo CJ, LaMontagne KR Jr, Garcia-Cardena G, Ackley BD, Kalman D, Park S, Christofferson R, Kamihara J, Ding YH, Lo KM, Gillies S, Folkman J, Mulligan RC & Javaherian K (2001) Oligomerization-dependent regulation of motility and morphogenesis by the collagen XVIII NC1/endostatin domain. J Cell Biol 152: 1233-46. Kuster JE, Guarnieri MH, Ault JG, Flaherty L & Swiatek PJ (1997) IAP insertion in the murine LamB3 gene results in junctional epidermolysis bullosa. Mamm Genome 8: 673-81. Ladher RK, Church VL, Allen S, Robson L, Abdelfattah A, Brown NA, Hattersley G, Rosen V, Luyten FP, Dale L, Francis-West PH (2000) Cloning and expression of the Wnt antagonists Sfrp-2 and Frzb during chick development. Dev Biol 218:183-98. 72 Lai R, Estey E, Shen Y, Despa S, Kantarjian H, Beran M, Maushouri T, Quackenbuch RC, Keating M & Albitar M (2002) Clinical significance of plasma endostatin in acute myeloid leukemia/myelodysplastic syndrome. Cancer 94: 14-7. Lee SJ, Jang JW, Kim YM, Lee HI, Jeon JY, Kwon YG & Lee ST (2002) Endostatin binds to the catalytic domain of matrix metalloproteinase-2. FEBS Lett 519: 147-52. Leivo I, Vaheri A, Timpl R & Wartiovaara J (1980) Appearance and distribution of collagens and laminin in the early mouse embryo. Dev Biol 76: 100-14. Li D, Clark CC & Myers JC (2000) Basement membrane zone type XV collagen is a disulfide- bonded chondroitin sulfate proteoglycan in human tissues and cultured cells. J Biol Chem 275: 22339-47. Li S, Harrison D, Carbonetto S, Fässler R, Smyth N, Edgar D & Yurchenco PD (2002) Matrix assembly, regulation, and survival functions of laminin and its receptors in embryonic stem cell differentiation. J Cell Biol 157: 1279-90. Li X, Chen Y, Scheele S, Arman E & Haffner-Krausz R, Ekblom P & Lonai P. (2001) Fibroblast growth factor signaling and basement membrane assembly are connected during epithelial morphogenesis of the embryoid body. J Cell Biol 153: 811-22. Liby K, Neltner B, Mohamet L, Burd C & Ben-Jonathan N. (2003) Endostatin expression by MDA-MB-435 breast cancer cells effectively inhibits tumor growth. Cancer Biol Ther 2: 48-52. Lin HC, Chang JH, Jain S, Gabison EE, Kure T, Kato T, Fukai N & Azar DT (2001a) Matrilysin cleavage of corneal collagen type XVIII NC1 domain and generation of a 28-kDa fragment. Invest Ophthalmol Vis Sci 42: 2517-24. Lin Y, Zhang S, Rehn M, Itaranta P, Tuukkanen J, Heljasvaara R, Peltoketo H, Pihlajaniemi T & Vainio S (2001b) Induced repatterning of type XVIII collagen expression in ureter bud from kidney to lung type: association with sonic hedgehog and ectopic surfactant protein C. Development 128: 1573-85. Lindahl U, Kusche-Gullberg M & Kjellen L (1998) Regulated diversity of heparan sulfate. J Biol Chem 273: 24979-82. MacDonald NJ, Shivers WY, Narum DL, Plum SM, Wingard JN, Fuhrmann SR, Liang H, Holland- Linn J, Chen DH & Sim BK (2001) Endostatin binds tropomyosin. A potential modulator of the antitumor activity of endostatin. J Biol Chem 276: 25190-6. Marneros AG, Keene DR, Hansen U, Fukai N, Moulton K, Goletz PL, Moiseyev G, Pawlyk BS, Halfter W, Dong S, Shibata M, Li T, Crouch RK, Bruckner P & Olsen BR (2004) Collagen XVIII/endostatin is essential for vision and retinal pigment epithelial function. EMBO J 23: 89- 99. Marneros AG & Olsen BR (2003) Age-dependent iris abnormalities in collagen XVIII/endostatin deficient mice with similarities to human pigment dispersion syndrome.Invest Ophthalmol Vis Sci 44: 2367-72. Masiakowski P & Yancopoulos GD (1998) The Wnt receptor CRD domain is also found in MuSK and related orphan receptor tyrosine kinases. Curr Biol 8: R407 McAvoy JW, Chamberlain CG, de Iongh RU, Hales AM & Lovicu FJ (1999) Lens development. Eye 13: 425-37. McMillan JR, Akiyama M & Shimizu H (2003) Epidermal basement membrane zone components: ultrastructural distribution and molecular interactions. J Dermatol Sci 31: 169-77. Meng X, Klement JF, Leperi DA, Birk DE, Sasaki T, Timpl R, Uitto J & Pulkkinen L (2003) Targeted inactivation of murine laminin gamma2-chain gene recapitulates human junctional epidermolysis bullosa. J Invest Dermatol 121: 720-31. Menzel O, Bekkeheien RC, Reymond A, Fukai N, Boye E, Kosztolanyi G, Aftimos S, Deutsch S, Scott HS, Olsen BR, Antonarakis SE & Guipponi M (2004) Knobloch syndrome: novel mutations in COL18A1, evidence for genetic heterogeneity, and a functionally impaired polymorphism in endostatin. Hum Mutat 23: 77-84. 73 Miner JH, Cunningham J & Sanes JR (1998) Roles for laminin in embryogenesis: exencephaly, syndactyly, and placentopathy in mice lacking the laminin alpha5 chain. J Cell Biol 143: 1713- 23. Miner JH, Li C, Mudd JL, Go G & Sutherland AE (2004) Compositional and structural requirements for laminin and basement membranes during mouse embryo implantation and gastrulation. Development 131:2247-56. Miner JH & Sanes JR (1996) Molecular and functional defects in kidneys of mice lacking collagen alpha 3(IV): implications for Alport syndrome. J Cell Biol 135:1403-13. Miosge N (2001) The ultrastructural composition of basement membranes in vivo. Histol Histopathol 16: 1239-48. Miosge N, Sasaki T & Timpl R (1999) Angiogenesis inhibitor endostatin is a distinct component of elastic fibers in vessel walls. FASEB J 13: 1743-50. Miosge N, Simniok T, Sprysch P & Herken R (2003) The collagen type XVIII endostatin domain is co-localized with perlecan in basement membranes in vivo. J Histochem Cytochem 51: 285-96. Moeller C, Swindell EC, Kispert A & Eichele G (2003) Carboxypeptidase Z (CPZ) modulates Wnt signaling and regulates the development of skeletal elements in the chicken. Development 130: 5103-11. Mongiat M, Sweeney SM, San Antonio JD, Fu J & Iozzo RV (2003) Endorepellin, a novel inhibitor of angiogenesis derived from the C terminus of perlecan. J Biol Chem 278: 4238-49. Moore W & Guille M (1999) Preparation and testing of synthetic mRNA for microinjection. Methods Mol Biol 127: 99-109. Morimoto T, Aoyagi M, Tamaki M, Yoshino Y, Hori H, Duan L, Yano T, Shibata M, Ohno K, Hirakawa K & Yamaguchi N (2002) Increased levels of tissue endostatin in human malignant gliomas. Clin Cancer Res 8: 2933-8. Muona A, Eklund L, Vaisanen T & Pihlajaniemi T (2002) Developmentally regulated expression of type XV collagen correlates with abnormalities in Col15a1(-/-) mice. Matrix Biol 21: 89-102. Muragaki Y, Abe N, Ninomiya Y, Olsen BR & Ooshima A (1994) The human alpha 1(XV) collagen chain contains a large amino-terminal non-triple helical domain with a tandem repeat structure and homology to alpha 1(XVIII) collagen. J Biol Chem 269: 4042-6. Muragaki Y, Timmons S, Griffith CM, Oh SP, Fadel B, Quertermous T & Olsen BR (1995) Mouse Col18a1 is expressed in a tissue-specific manner as three alternative variants and is localized in basement membrane zones. Proc Natl Acad Sci U S A 92: 8763-7. Murray P & Edgar D (2000) Regulation of programmed cell death by basement membranes in embryonic development. J Cell Biol 150: 1215-21. Murshed M, Smyth N, Miosge N, Karolat J, Krieg T, Paulsson M & Nischt R (2000) The absence of nidogen 1 does not affect murine basement membrane formation. Mol Cell Biol 20: 7007-12. Musso O, Rehn M, Saarela J, Theret N, Lietard J, Hintikka, Lotrian D, Campion JP, Pihlajaniemi T & Clement B (1998) Collagen XVIII is localized in sinusoids and basement membrane zones and expressed by hepatocytes and activated stellate cells in fibrotic human liver. Hepatology 28: 98-107. Musso O, Theret N, Heljasvaara R, Rehn M, Turlin B, Campion JP, Pihlajaniemi T & Clement B (2001) Tumor hepatocytes and basement membrane-Producing cells specifically express two different forms of the endostatin precursor, collagen XVIII, in human liver cancers. Hepatology 33: 868-76. Myers JC, Dion AS, Abraham V, Amenta PS (1996) Type XV collagen exhibits a widespread distribution in human tissues but a distinct localization in basement membrane zones. Cell Tissue Res 286: 493-505. Myers JC, Kivirikko S, Gordon MK, Dion AS & Pihlajaniemi T (1992) Identification of a previously unknown human collagen chain, alpha 1(XV), characterized by extensive interruptions in the triple-helical region. Proc Natl Acad Sci U S A 89: 10144-8. 74 Myllyharju J & Kivirikko KI (2001) Collagens and collagen-related diseases. Ann Med 33: 7-21. Myllyharju J, Kivirikko KI (2004) Collagens, modifying enzymes and their mutations in humans, flies and worms.Trends Genet 20: 33-43. Nguyen MM, Potter SJ & Griep AE (2002) Deregulated cell cycle control in lens epithelial cells by expression of inhibitors of tumor suppressor function. Mech Dev 112:101-13. Nieuwkoop PD & Faber J ( 1994) Normal table of Xenopus laevis. Garland Publishing, New York. Nomi M, Oishi I, Kani S, Suzuki H, Matsuda T, Yoda A, Kitamura M, Itoh K, Takeuchi S, Takeda K, Akira S, Ikeya M, Takada S & Minami Y (2001) Loss of mRor1 enhances the heart and skeletal abnormalities in mRor2-deficient mice: redundant and pleiotropic functions of mRor1 and mRor2 receptor tyrosine kinases. Mol Cell Biol 21: 8329-35. Noonan DM, Fulle A, Valente P, Cai S, Horigan E, Sasaki M, Yamada Y & Hassell JR (1991) The complete sequence of perlecan, a basement membrane heparan sulfate proteoglycan, reveals extensive similarity with laminin A chain, low density lipoprotein-receptor, and the neural cell adhesion molecule. J Biol Chem 266: 22939-47. Norose K, Clark JI, Syed NA, Basu A, Heber-Katz E, Sage EH & Howe CC (1998) SPARC deficiency leads to early-onset cataractogenesis. Invest Ophthalmol Vis Sci 39: 2674-80. Nyberg P, Heikkila P, Sorsa T, Luostarinen J, Heljasvaara R, Stenman UH, Pihlajaniemi T & Salo T (2003) Endostatin inhibits human tongue carcinoma cell invasion and intravasation and blocks the activation of matrix metalloprotease-2, -9, and -13. J Biol Chem 278: 22404-11. Oh SP, Warman ML, Seldin MF, Cheng SD, Knoll JH, Timmons S & Olsen BR (1994) Cloning of cDNA and genomic DNA encoding human type XVIII collagen and localization of the alpha 1(XVIII) collagen gene to mouse chromosome 10 and human chromosome 21. Genomics 19: 494-9. Oishi I, Suzuki H, Onishi N, Takada R, Kani S, Ohkawara B, Koshida I, Suzuki K, Yamada G, Schwabe GC, Mundlos S, Shibuya H, Takada S & Minami Y (2003) The receptor tyrosine kinase Ror2 is involved in non-canonical Wnt5a/JNK signaling pathway. Genes Cells 8: 645-54. 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: 277-85. O'Reilly MS, Holmgren L, Shing Y, Chen C, Rosenthal RA, Moses M, Lane WS, Cao Y, Sage EH & Folkman J (1994) Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79:315-28. Otani A, Slike BM, Dorrell MI, Hood J, Kinder K, Ewalt KL, Cheresh D, Schimmel P & Friedlander M (2002) A fragment of human TrpRS as a potent antagonist of ocular angiogenesis. Proc Natl Acad Sci U S A 99: 178-83. Perletti G, Concari P, Giardini R, Marras E, Piccinini F, Folkman J & Chen L (2000) Antitumor activity of endostatin against carcinogen-induced rat primary mammary tumors. Cancer Res 60: 1793-6. Peroulis I, Jonas N & Saleh M (2002) Antiangiogenic activity of endostatin inhibits C6 glioma growth. Int J Cancer 97: 839-45. Petitclerc E, Boutaud A, Prestayko A, Xu J, Sado Y, Ninomiya Y, Sarras MP Jr, Hudson BG & Brooks PC (2000) New functions for non-collagenous domains of human collagen type IV. Novel integrin ligands inhibiting angiogenesis and tumor growth in vivo. J Biol Chem 275: 8051-61. Pihlajaniemi T & Rehn M (1995) Two new collagen subgroups: membrane-associated collagens and types XV and XVII. Prog Nucleic Acid Res Mol Biol 50:225-62. Pike SE, Yao L, Jones KD, Cherney B, Appella E, Sakaguchi K, Nakhasi H, Teruya-Feldstein J, Wirth P, Gupta G & Tosato G (1998) Vasostatin, a calreticulin fragment, inhibits angiogenesis and suppresses tumor growth. J Exp Med 188: 2349-56. 75 Prockop DJ & Kivirikko KI (1995) Collagens: molecular biology, diseases, and potentials for therapy. Annu Rev Biochem 64: 403-34. Prox D, Becker C, Pirie-Shepherd SR, Celik I, Folkman J & Kisker O (2003) Treatment of human pancreatic cancer in mice with angiogenic inhibitors. World J Surg 27: 405-11. Pöschl E, Mayer U, Stetefeld J, Baumgartner R, Holak TA, Huber R & Timpl R (1996) Site- directed mutagenesis and structural interpretation of the nidogen binding site of the laminin gamma1 chain. EMBO J 15: 5154-9. Pöschl E, Schlotzer-Schrehardt U, Brachvogel B, Saito K, Ninomiya Y, Mayer U (2004) Collagen IV is essential for basement membrane stability but dispensable for initiation of its assembly during early development. Development 131: 1619-28 Quondamatteo F (2002) Assembly, stability and integrity of basement membranes in vivo. Histochem J 34: 369-81. Raghavan S, Bauer C, Mundschau G, Li Q & Fuchs E (2000) Conditional ablation of beta1 integrin in skin. Severe defects in epidermal proliferation, basement membrane formation, and hair follicle invagination. J Cell Biol 150: 1149-60. Ramchandran R, Dhanabal M, Volk R, Waterman MJ, 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 Commun 255: 735-9. Read TA, Sorensen DR, Mahesparan R, Enger PO, Timpl R, Olsen BR, Hjelstuen MH, Haraldseth O & Bjerkvig R (2001) Local endostatin treatment of gliomas administered by microencapsulated producer cells. Nat Biotechnol 19: 29-34. Rehn M, Hintikka E & Pihlajaniemi T (1994) Primary structure of the alpha 1 chain of mouse type XVIII collagen, partial structure of the corresponding gene, and comparison of the alpha 1(XVIII) chain with its homologue, the alpha 1(XV) collagen chain. J Biol Chem 269: 13929- 35. Rehn M, Hintikka E & Pihlajaniemi T (1996) Characterization of the mouse gene for the alpha 1 chain of type XVIII collagen (Col18a1) reveals that the three variant N-terminal polypeptide forms are transcribed from two widely separated promoters. Genomics 32: 436-46. Rehn M & Pihlajaniemi T (1994) Alpha 1(XVIII), a collagen chain with frequent interruptions in the collagenous sequence, a distinct tissue distribution, and homology with type XV collagen. Proc Natl Acad Sci U S A 91: 4234-8. 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. The longest form contains a novel motif homologous to rat and Drosophila frizzled proteins.J Biol Chem 270: 4705-11. Rehn M, Pihlajaniemi T, Hofmann K & Bucher P (1998) The frizzled motif: in how many different protein families does it occur? Trends Biochem Sci 23: 415-7. Rehn M, Veikkola T, Kukk-Valdre E, Nakamura H, Ilmonen M, Lombardo C, Pihlajaniemi T, Alitalo K & Vuori K (2001) Interaction of endostatin with integrins implicated in angiogenesis. Proc Natl Acad Sci U S A 98: 1024-9. Reijerkerk A, Mosnier LO, Kranenburg O, Bouma BN, Carmeliet P, Drixler T, Meijers JC, Voest EE & Gebbink MF (2003) Amyloid endostatin induces endothelial cell detachment by stimulation of the plasminogen activation system. Mol Cancer Res 1: 561-8. Reinhardt D, Mann K, Nischt R, Fox JW, Chu ML, Krieg T & Timpl R (1993) Mapping of nidogen binding sites for collagen type IV, heparan sulfate proteoglycan, and zinc. J Biol Chem 268: 10881-7. Ricard-Blum S, Feraud O, Lortat-Jacob H, Rencurosi A, Fukai N, Dkhissi F, Vittet D, Imberty A, Olsen BR & van der Rest M (2004) Characterization of endostatin binding to heparin and heparan sulfate by surface plasmon resonance and molecular modeling: role of divalent cations. J Biol Chem 279: 2927-36. 76 Rossi M, Morita H, Sormunen R, Airenne S, Kreivi M, Wang L, Fukai N, Olsen BR, Tryggvason K & Soininen R (2003) Heparan sulfate chains of perlecan are indispensable in the lens capsule but not in the kidney. EMBO J 22: 236-45. Ryan MC, Lee K, Miyashita Y & Carter WG (1999) Targeted disruption of the LAMA3 gene in mice reveals abnormalities in survival and late stage differentiation of epithelial cells. J Cell Biol 145: 1309-23. Saarela J, Rehn M, Oikarinen A, Autio-Harmainen H & Pihlajaniemi T (1998a) The short and long forms of type XVIII collagen show clear tissue specificities in their expression and location in basement membrane zones in humans. Am J Pathol 153: 611-26. Saarela J, Ylikarppa R, Rehn M, Purmonen S & Pihlajaniemi T (1998b) Complete primary structure of two variant forms of human type XVIII collagen and tissue-specific differences in the expression of the corresponding transcripts. Matrix Biol 16: 319-28. Sajnani-Perez G, Chilton JK, Aricescu AR, Haj F & Stoker AW (2003) Isoform-specific binding of the tyrosine phosphatase PTPsigma to a ligand in developing muscle. Mol Cell Neurosci 22: 37- 48. Saldanha J, Singh J & Mahadevan D (1998) Identification of a Frizzled-like cysteine rich domain in the extracellular region of developmental receptor tyrosine kinases. Protein Sci 7: 1632-5. Salmivirta M, Lidholt K & Lindahl U (1996) Heparan sulfate: a piece of information.FASEB J 10: 1270-9. Sasaki T, Fukai N, Mann K, Gohring W, Olsen BR & Timpl R (1998) Structure, function and tissue forms of the C-terminal globular domain of collagen XVIII containing the angiogenesis inhibitor endostatin. EMBO J 17: 4249-56. Sasaki T, Fässler R & Hohenester E (2004) Laminin: the crux of basement membrane assembly J. Cell Biol 164: 959 - 963. Sasaki T, Gohring W, Miosge N, Abrams WR, Rosenbloom J & Timpl R (1999a) Tropoelastin binding to fibulins, nidogen-2 and other extracellular matrix proteins. FEBS Lett 460: 280-4. Sasaki T, Larsson H, Kreuger J, Salmivirta M, Claesson-Welsh L, Lindahl U, Hohenester E & Timpl R (1999b) Structural basis and potential role of heparin/heparan sulfate binding to the angiogenesis inhibitor endostatin. EMBO J 18: 6240-8. Sasaki T, Larsson H, Tisi D, Claesson-Welsh L, Hohenester E & Timpl R (2000) Endostatins derived from collagens XV and XVIII differ in structural and binding properties, tissue distribution and anti-angiogenic activity. J Mol Biol 301: 1179-90. Sauter BV, Martinet O, Zhang WJ, Mandeli J & Woo SL (2000) Adenovirus-mediated gene transfer of endostatin in vivo results in high level of transgene expression and inhibition of tumor growth and metastases. Proc Natl Acad Sci U S A 97: 4802-7. Schymeinsky J, Nedbal S, Miosge N, Poschl E, Rao C, Beier DR, Skarnes WC, Timpl R & Bader BL (2002) Gene structure and functional analysis of the mouse nidogen-2 gene: nidogen-2 is not essential for basement membrane formation in mice. Mol Cell Biol 22: 6820-30. Sertie AL, Sossi V, Camargo AA, Zatz M, Brahe C & Passos-Bueno MR (2000) Collagen XVIII, containing an endogenous inhibitor of angiogenesis and tumor growth, plays a critical role in the maintenance of retinal structure and in neural tube closure (Knobloch syndrome). Hum Mol Genet 9: 2051-8. Shi W, Teschendorf C, Muzyczka N & Siemann DW (2002) Adeno-associated virus-mediated gene transfer of endostatin inhibits angiogenesis and tumor growth in vivo. Cancer Gene Ther 9: 513- 21. Shichiri M & Hirata Y (2001) Antiangiogenesis signals by endostatin. FASEB J 15: 1044-53. Smith RS, Sundberg JP & Linder CC (1997) Mouse mutations as models for studying cataracts. Pathobiology 65: 146-54. 77 Smyth N, Vatansever HS, Murray P, Meyer M, Frie C, Paulsson M & Edgar D (1999) Absence of basement membranes after targeting the LAMC1 gene results in embryonic lethality due to failure of endoderm differentiation. J Cell Biol 144: 151-60. Solaun MS, Mendoza L, De Luca M, Gutierrez V, Lopez MP, Olaso E, Lee Sim BK & Vidal- Vanaclocha F (2002) Endostatin inhibits murine colon carcinoma sinusoidal-type metastases by preferential targeting of hepatic sinusoidal endothelium. Hepatology 35: 1104-16. Song L & Fricker LD (1997) Cloning and expression of human carboxypeptidase Z, a novel metallocarboxypeptidase. J Biol Chem 272: 10543-50. Sorensen DR, Read TA, Porwol T, Olsen BR, Timpl R, Sasaki T, Iversen PO, Benestad HB, Sim BK & Bjerkvig R (2002) Endostatin reduces vascularization, blood flow, and growth in a rat gliosarcoma. Neuro-oncol 4: 1-8. Stephens LE, Sutherland AE, Klimanskaya IV, Andrieux A, Meneses J, Pedersen RA & Damsky CH (1995) Deletion of beta 1 integrins in mice results in inner cell mass failure and peri- implantation lethality. Genes Dev 9: 1883-95. Sudhakar A, Sugimoto H, Yang C, Lively J, Zeisberg M & Kalluri R (2003) Human tumstatin and human endostatin exhibit distinct antiangiogenic activities mediated by alpha v beta 3 and alpha 5 beta 1 integrins. Proc Natl Acad Sci U S A 100: 4766-71. Sumiyoshi H, Mor N, Lee SY, Doty S, Henderson S, Tanaka S, Yoshioka H, Rattan S & Ramirez F (2004) Esophageal muscle physiology and morphogenesis require assembly of a collagen XIX- rich basement membrane zone. J Cell Biol 166: 591-600. Suzuki M, Iizasa T, Ko E, Baba M, Saitoh Y, Shibuya K, Sekine Y, Yoshida S, Hiroshima K & Fujisawa T (2002a) Serum endostatin correlates with progression and prognosis of non-small cell lung cancer. Lung Cancer 35: 29-34. Suzuki OT, Sertie AL, Der Kaloustian VM, Kok F, Carpenter M, Murray J, Czeizel AE, Kliemann SE, Rosemberg S, Monteiro M, Olsen BR & Passos-Bueno MR (2002b) Molecular analysis of collagen XVIII reveals novel mutations, presence of a third isoform, and possible genetic heterogeneity in Knobloch syndrome. Am J Hum Genet 71: 1320-9. Takeuchi S, Takeda K, Oishi I, Nomi M, Ikeya M, Itoh K, Tamura S, Ueda T, Hatta T, Otani H, Terashima T, Takada S, Yamamura H, Akira S & Minami Y (2000) Mouse Ror2 receptor tyrosine kinase is required for the heart development and limb formation. Genes Cells 5: 71-8. Talts JF, Andac Z, Gohring W, Brancaccio A, Timpl R (1999) Binding of the G domains of laminin alpha1 and alpha2 chains and perlecan to heparin, sulfatides, alpha-dystroglycan and several extracellular matrix proteins. EMBO J 18:863-70. Tebar M, Destree O, de Vree WJ & Ten Have-Opbroek AA (2001) Expression of Tcf/Lef and sFrp and localization of beta-catenin in the developing mouse lung. Mech Dev 109: 437-40. Timpl R & Aumailley M (1989) Biochemistry of basement membranes. Adv Nephrol Necker Hosp 18: 59-76 Timpl R & Brown JC (1996) Supramolecular assembly of basement membranes. Bioessays 18: 123-32. Timpl R, Tisi D, Talts JF, Andac Z, Sasaki T & Hohenester E (2000) Structure and function of laminin LG modules. Matrix Biol 19: 309-17. Tomono Y, Naito I, Ando K, Yonezawa T, Sado Y, Hirakawa S, Arata J, Okigaki T & Ninomiya Y (2002) Epitope-defined monoclonal antibodies against multiplexin collagens demonstrate that type XV and XVIII collagens are expressed in specialized basement membranes. Cell Struct Funct 27: 9-20. Tsen G, Halfter W, Kroger S & Cole GJ (1995) Agrin is a heparan sulfate proteoglycan. J Biol Chem 270: 3392-9. Urbich C, Reissner A, Chavakis E, Dernbach E, Haendeler J, Fleming I, Zeiher AM, Kaszkin M & Dimmeler S (2002) Dephosphorylation of endothelial nitric oxide synthase contributes to the anti-angiogenic effects of endostatin. FASEB J 16: 706-8 78 Utriainen A, Sormunen R, Kettunen M, Carvalhaes LS, Sajanti E, Eklund L, Kauppinen R, Kitten GT & Pihlajaniemi T (2004) Structurally altered basement membranes and hydrocephalus in a type XVIII collagen deficient mouse line. Hum Mol Genet 13: 2089-99. van den Heuvel M & Ingham PW (1996) smoothened encodes a receptor-like serpentine protein required for hedgehog signaling. Nature 382: 547-51. van der Neut R, Krimpenfort P, Calafat J, Niessen CM & Sonnenberg A (1996) Epithelial detachment due to absence of hemidesmosomes in integrin beta 4 null mice. Nat Genet 13: 366- 9. van der Rest M & Garrone R (1991) Collagen family of proteins. FASEB J 5: 2814-23. Vassar R, Rosenberg M, Ross S, Tyner A & Fuchs E (1989) Tissue-specific and differentiation- specific expression of a human K14 keratin gene in transgenic mice. Proc Natl Acad Sci U S A 86:1563-7. Wakasugi K, Slike BM, Hood J, Otani A, Ewalt KL, Friedlander M, Cheresh DA & Schimmel P (2002) A human aminoacyl-tRNA synthetase as a regulator of angiogenesis. Proc Natl Acad Sci U S A 99: 173-7. Wen W, Moses MA, Wiederschain D, Arbiser JL & Folkman J (1999) The generation of endostatin is mediated by elastase. Cancer Res 59: 6052-6. Wickström SA, Alitalo K, Keski-Oja J (2002) Endostatin associates with integrin alpha5beta1 and caveolin-1, and activates Src via a tyrosyl phosphatase-dependent pathway in human endothelial cells. Cancer Res 62: 5580-9. Wickström SA, Alitalo K & Keski-Oja J (2003) Endostatin associates with lipid rafts and induces reorganization of the actin cytoskeleton via down-regulation of RhoA activity. J Biol Chem 278: 37895-901. Wickström SA, Veikkola T, Rehn M, Pihlajaniemi T, Alitalo K & Keski-Oja J (2001) Endostatin- induced modulation of plasminogen activation with concomitant loss of focal adhesions and actin stress fibers in cultured human endothelial cells. Cancer Res 61: 6511-6. Willem M, Miosge N, Halfter W, Smyth N, Jannetti I, Burghart E, Timpl R & Mayer U (2002) Specific ablation of the nidogen-binding site in the laminin gamma1 chain interferes with kidney and lung development. Development 129: 2711-22. Williamson RA, Henry MD, Daniels KJ, Hrstka RF, Lee JC, Sunada Y, Ibraghimov-Beskrovnaya O & Campbell KP (1997) Dystroglycan is essential for early embryonic development: disruption of Reichert's membrane in Dag1-null mice. Hum Mol Genet 6: 831-41. Wodarz A & Nusse R (1998) Mechanisms of Wnt signaling in development. Annu Rev Cell Dev Biol 14: 59-88. Wu M & Gerhart J (1991) Raising Xenopus in the laboratory. Methods Cell Biol 36: 3-18. Xu R, Yao ZY, Xin L, Zhang Q, Li TP & Gan RB (2001) NC1 domain of human type VIII collagen (alpha 1) inhibits bovine aortic endothelial cell proliferation and causes cell apoptosis. Biochem Biophys Res Commun 289: 264-8. Yan Q, Clark JI, Wight TN & Sage EH (2002) Alterations in the lens capsule contribute to cataractogenesis in SPARC-null mice. J Cell Sci 115: 2747-56. Yan W, Sheng N, Seto M, Morser J & Wu Q (1999) Corin, a mosaic transmembrane serine protease encoded by a novel cDNA from human heart. J Biol Chem 274: 14926-35. Ye C, Feng C, Wang S, Liu X, Lin Y & Li M (2002) Antiangiogenic and antitumor effects of endostatin on follicular thyroid carcinoma. Endocrinology 143: 3522-8. Yi M & Ruoslahti E (2001) A fibronectin fragment inhibits tumor growth, angiogenesis, and metastasis. Proc Natl Acad Sci U S A 98: 620-4. Yi M, Sakai T, Fässler R & Ruoslahti E (2003) Antiangiogenic proteins require plasma fibronectin or vitronectin for in vivo activity. Proc Natl Acad Sci U S A 100: 11435-8. 79 Ylikärppa R, Eklund L, Sormunen R, Kontiola AI, Utriainen A, Määttä M, Fukai N, Olsen BR & Pihlajaniemi T (2003a) Lack of type XVIII collagen results in anterior ocular defects. FASEB J 17: 2257-9. Ylikärppa R, Eklund L, Sormunen R, Muona A, Fukai N, Olsen BR & Pihlajaniemi T (2003b) Double knockout mice reveal a lack of major functional compensation between collagens XV and XVIII. Matrix Biol 22: 443-8. Yoda A, Oishi I & Minami Y (2003) Expression and function of the Ror-family receptor tyrosine kinases during development: lessons from genetic analyses of nematodes, mice, and humans. J Recept Signal Transduct Res 23: 1-15. Yokoyama Y, Dhanabal M, Griffioen AW, Sukhatme VP & Ramakrishnan S (2000a) Synergy between angiostatin and endostatin: inhibition of ovarian cancer growth. Cancer Res 60: 2190-6. Yokoyama Y, Green JE, Sukhatme VP & Ramakrishnan S (2000b) Effect of endostatin on spontaneous tumorigenesis of mammary adenocarcinoma in a transgenic mouse model. Cancer Res 60: 4362-5. Yoon SS, Eto H, Lin CM, Nakamura H, Pawlik TM, Song SU & Tanabe KK (1999) Mouse endostatin inhibits the formation of lung and liver metastases. Cancer Res 59: 6251-6. Yu Y, Moulton KS, Khan MK, Vineberg S, Boye E, Davis VM, O'Donnell PE, Bischoff J & Milstone DS (2004) E-selectin is required for the antiangiogenic activity of endostatin. Proc Natl Acad Sci U S A 101: 8005-10. Yurchenco PD, Amenta PS & Patton BL (2004) Basement membrane assembly, stability and activities observed through a developmental lens. Matrix Biol 22:521-38. Zachary I & Gliki G (2001) Signaling transduction mechanisms mediating biological actions of the vascular endothelial growth factor family. Cardiovasc Res 49:568-81. Zampighi GA, Eskandari S & Kreman M (2000) Epithelial organization of the mammalian lens. Exp Eye Res 71: 415-35.