ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2013

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 884

Syndecan - Regulationand Function of itsGlycosaminoglycan Chains

ANNA S. ERIKSSON

ISSN 1651-6206ISBN 978-91-554-8637-2urn:nbn:se:uu:diva-197691

Dissertation presented at Uppsala University to be publicly examined in A1:107a, BMC,Husargatan 3, Uppsala, Friday, May 17, 2013 at 13:15 for the degree of Doctor of Philosophy(Faculty of Medicine). The examination will be conducted in English.

AbstractEriksson, A. S. 2013. Syndecan - Regulation and Function of its Glycosaminoglycan Chains.Acta Universitatis Upsaliensis. Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 884. 54 pp. Uppsala. ISBN 978-91-554-8637-2.

The cell surface is an active area where extracellular molecules meet their receptors andaffect the cellular fate by inducing for example cell proliferation and adhesion. Syndecansand integrins are two transmembrane molecules that have been suggested to fine-tune theseactivities, possibly in cooperation. Syndecans are proteoglycans, i.e. proteins with specifictypes of carbohydrate chains attached. These chains are glycosaminoglycans and either heparansulfate (HS) or chondroitin sulfate (CS). Syndecans are known to influence cell adhesion andsignaling. Integrins in turn, are important adhesion molecules that connect the extracellularmatrix with the cytoskeleton, and hence can regulate cell motility. In an attempt to study how thetwo types of glycosaminoglycans attached to syndecan-1 can interact with integrins, a cell basedmodel system was used and functional motility assays were performed. The results showed thatHS, but not CS, on the cell surface was capable of regulating integrin-mediated cell motility.

Regulation of intracellular signaling is crucial to prevent abnormal cellular behavior. In thesecond part of this thesis, the aim was to see how the presentation of glycosaminoglycanchains to the FGF signaling complex could affect the cellular response. When attached tothe plasma membrane via syndecan-1, CS chains could support the intracellular signaling,although not promoting as strong signals as HS. When glycosaminoglycans were attached tofree ectodomains of syndecan-1, both types of chains sequestered FGF2 from the receptors tothe same extent, pointing towards functional overlap between CS and HS.

To further study the interplay between HS and CS, their roles in the formation of pharyngealcartilage in zebrafish were established. HS was important during chondrocyte intercalation andCS in the formation of the surrounding extracellular matrix. Further, the balance between thebiosynthetic enzymes determined the ratio of HS and CS, and HS biosynthesis was prioritizedover CS biosynthesis.

The results presented in this thesis provide further insight into the regulation of HSbiosynthesis, as well as the roles of both HS and CS on the cell surface. It is evident, that incertain situations there is a strict requirement for a certain HS structure, albeit in other situationsthere is a functional overlap between HS and CS.

Keywords: heparan sulfate, chondroitin sulfate, integrin, cell motility, fibroblast growth factorsignaling, pharyngeal cartilage, biosynthesis regulation

Anna S. Eriksson, Uppsala University, Department of Medical Biochemistry andMicrobiology, Box 582, SE-751 23 Uppsala, Sweden.

© Anna S. Eriksson 2013

ISSN 1651-6206ISBN 978-91-554-8637-2urn:nbn:se:uu:diva-197691 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-197691)

Fly me to the moon, let me play among the stars

Supervisors: Dorothe Spillmann, Senior Lecturer Department of Medical Biochemistry and Microbiology, Uppsala University Ulf Lindahl, Professor Department of Medical Biochemistry and Microbiology, Uppsala University Faculty opponent: Catherine Merry, Senior Lecturer Stem Cell Glycobiology Group, School of Materials, University of Manchester, Manchester, United Kingdom Examining Committee: Anna Dimberg, Associate Professor Department of Immunology, Genetics and Pathology, Uppsala University Karin Forsberg-Nilsson, Professor Department of Immunology, Genetics and Pathology, Uppsala University Anders Hjerpe, Professor Department of Laboratory Medicine, Karolinska Institute Staffan Johansson, Professor Department of Medical Biochemistry and Microbiology, Uppsala University Johan Lennartsson, Associate Professor Ludwig Institute for Cancer Research, Uppsala Chairperson: Lena Kjellén, Professor Department of Medical Biochemistry and Microbiology, Uppsala University The cover picture shows CHO-K1 cells overexpressing uncleavable syndecan-1 in green, the cell nucleus in blue and actin cytoskeleton in red.

List of Papers

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

I Eriksson, A.S, Spillmann, D. (2012) The mutual impact of syndecan-1 and its glycosaminoglycan chains – A multivariable puzzle. Journal of Histochemistry and Cytochemistry, 60(12):936–42

II Eriksson, A.S, Reyhani, V, Spillmann, D. (2013) Cell surface hepa-

ran sulfate chains are important for integrin-mediated cell motility. Manuscript.

III Eriksson A.S, Spillmann, D. (2013) Role of glycosaminoglycans on

syndecan-1 in fibroblast growth factor signaling. Manuscript.

IV Holmborn, K*, Habicher, J*, Kasza, Z, Eriksson, A.S, Filipek-Gorniok, B, Gopal, S, Couchman, J.R, Ahlberg, P.E, Wiweger, M, Spillmann, D, Kreuger, J*, Ledin, J*. (2012) On the roles and regu-lation of chondroitin sulfate and heparan sulfate in zebrafish pharyn-geal cartilage morphogenesis. Journal of Biological Chemistry, 287(40):33905-16 * These authors contributed equally to the work.

Reprints were made with permission from the respective publishers.

Contents

Introduction ..................................................................................................... 9  

Background ................................................................................................... 10  Proteoglycans ............................................................................................ 10  

Syndecans ............................................................................................ 11  Glycosaminoglycans ................................................................................. 13  

Biosynthesis of glycosaminoglycans ................................................... 14  Regulation of glycosaminoglycan biosynthesis ................................... 19  

Integrins .................................................................................................... 20  Integrins and cell migration ................................................................. 21  

Fibroblast growth factors and their receptors ........................................... 22  Fibroblast growth factors and heparan sulfate ..................................... 23  Fibroblast growth factor induced signaling ......................................... 23  

Glycosaminoglycans in pathology ............................................................ 26  

Present Investigations .................................................................................... 29  Aim ........................................................................................................... 29  Model systems .......................................................................................... 30  

Chinese hamster ovary cells ................................................................. 30  Zebrafish .............................................................................................. 31  

Results and Discussion ............................................................................. 32  Paper I .................................................................................................. 32  Paper II ................................................................................................. 33  Paper III ............................................................................................... 35  Paper IV ............................................................................................... 36  

Concluding remarks ...................................................................................... 38  

Populärvetenskaplig sammanfattning ........................................................... 40  

Acknowledgements ....................................................................................... 41  

Thanks/Tack .................................................................................................. 42  

References ..................................................................................................... 44  

Abbreviations

CHO Chinese hamster ovary CS Chondroitin sulfate DS Dermatan sulfate ECM Extracellular matrix Erk1/2 Extracellular-signal-regulated kinase 1 and 2 EXT Exostosin FGF Fibroblast growth factor FGFR Fibroblast growth factor receptor FRS2 Fibroblast growth factor receptor substrate 2 GAG Glycosaminoglycan GalN Galactosamine GalNAc N-acetyl-galactosamine GlcA Glucuronic acid GlcN Glucosamine GlcNAc N-acetyl-glucosamine GlcNS N-sulfated glucosamine HS Heparan sulfate IdoA Iduronic acid KS Keratan sulfate MAPK Mitogen activated protein kinase MMP Matrix metalloproteinase NDST N-deacetylase/N-sulfotransferase OST O-sulfotransferase PAPS 3’-phosphoadenosine 5’-phosphosulfate PG Proteoglycan RGD Arginine-Glycine-Aspartic acid Sulfs Endo-O-sulfatases

9

Introduction

A large number of interactions take place at the surface of cells that deter-mine cellular fate by initiating processes such as proliferation, differentia-tion, or migration, that are crucial during both embryogenesis and homeosta-sis in the adult. However, inability to regulate the signals that control these events can lead to pathological conditions, such as the formation of tumors. Molecules at the cell surface are therefore of great importance in controlling and fine-tuning the flow of information from the surrounding environment to the nucleus of the cells.

Glycosaminoglycans are long carbohydrate chains that are highly nega-tively charged which infers interaction with a large number of molecules in the surrounding of cells, for example extracellular matrix proteins such as fibronectin and collagen, and cytokines like fibroblast growth factors and vascular endothelial growth factors. Glycosaminoglycan chains are attached to proteins to form proteoglycans. They are found in virtually every mamma-lian cell, at different locations, for example in the cell membrane. One mem-brane-associated proteoglycan is syndecan, and it has been suggested that syndecan has the ability to cooperate with other membrane bound molecules to control and fine-tune the flow of signals. The work in this thesis focuses on two types of glycosaminoglycan chains, namely heparan sulfate and chondroitin sulfate, and how they can influence the cellular behavior when they are attached to syndecan-1.

10

Background

Proteoglycans Proteoglycans (PGs) are proteins with glycosaminoglycan (GAG) chains covalently attached that are expressed in all mammalian cells. They can be found inside cells in vesicles, attached to the plasma membrane, or secreted into the extracellular matrix (ECM), Figure 1 [1, 2]. Due to the negative charge of the attached GAG chains the PGs can interact with a number of cell surface and matrix proteins, chemokines and cytokines, and thereby affect both physiological and pathological processes [3].

PGs can be characterized depending on which kind of GAGs that are at-tached and further divided into subgroups based on their localization. Hepa-ran sulfate proteoglycans (HSPGs) can be divided into four subclasses: the plasma membrane spanning, the glycosylphosphatidylinositol (GPI)-anchored, the extracellular and the intracellular, Figure 1. An example of a membrane spanning PG is syndecan. The four members of the syndecan family can be modified by both HS and CS chains. The glypicans are GPI-anchored and in humans the family comprises six members known to carry two to three HS chains that are attached close to the plasma membrane [4]. Among the extracellular HSPGs, perlecan and agrin are the two major ones found in basement membranes [5]. Intracellular PGs can be found in hema-topoietic cells such as mast cells and neutrophils, and are of serglycin type. In mast cells, serglycin is present in granules where it acts as a storage mole-cule for histamine and a variety of proteases. Depending on the tissue source, serglycin carries heparin or CS chains [6, 7].

Chondroitin sulfate proteoglycans (CSPGs) are important in structuring the ECM, especially in cartilage and in the brain, but they are also capable of affecting for example cytokinesis, proliferation and organogenesis [8]. Ag-grecan is a CSPG found in the ECM with more than hundred CS chains at-tached. It can also be modified with keratan sulfate (KS) chains, Figure 1 [9]. Aggrecan molecules bind non-covalently to a hyaluronan backbone to form large aggregates that are important in keeping the osmotic pressure in the tissue [10]. Other important CSPGs in the ECM are biglycan and decorin that belong to the small leucine-rich repeat PGs that can bind collagen fibrils and affect scaffold formation in connective tissues [11]. The transmembrane receptor NG2, also known as melanoma CSPG, has one CS chain attached

11

and has been shown to be upregulated in proliferative cells as well as present in many tumor types [12].

Figure 1. Examples of proteoglycans. PGs are found membrane associated, inside, and outside of the cell. They consist of a protein part (white) with HS/heparin (black), CS/DS (grey) or both types of chains attached. Aggrecan PGs bind to hyalu-ronan (black) and large extracellular aggregates are assembled. Aggrecan carries CS or KS chains (dashed grey).

Syndecans Syndecans are type I transmembrane PGs that can serve as co-receptors on the cell surface and also provide a link between the ECM and the cytoskele-ton by directly interacting with the cytoskeleton or via other molecules like integrins [13]. Syndecans are known to be involved in for example wound healing and inflammation [14]. In humans, there are four syndecans (Figure 2) that are expressed in a spatial and temporal manner during development and homeostasis. Syndecan-1 is mostly found in epithelial and mesenchymal cells together with syndecan-2, the latter being also found in neuronal cells. Syndecan-3 is expressed almost exclusively in neuronal and musculoskeletal tissue. Syndecan-4 is found in almost every cell type and is involved in focal adhesions [15]. When knocking out syndecan-4 in mice, the animals are viable and fertile but have defective wound healing and angiogenesis [16]. The removal of syndecan-1 also affects wound healing, and these mice show abnormal mammary gland development [17, 18].

12

The syndecans have variable extracellular domains with several attach-ment sites for GAG chains. Both syndecan-1 [19] and syndecan-4 [20] can be substituted with HS and CS and the same has been proposed for syndecan-3 [12, 21].

Figure 2. The syndecan family. All four members of the mammalian syndecan family are transmembrane proteins (white) with variously sized ectodomains where the attachment sites for HS (black) and CS (grey) are found.

Whereas the ectodomain of the four syndecans is different, the transmem-brane domain is highly conserved and has a sequence (GXXXG) that gives high affinity for self-association [22]. Syndecans form homodimers however there is still no evidence for heterodimers within the family [12]. The cyto-plasmic part also shows homology. It can be divided into three regions; two highly conserved parts, C1 and C2 and one variable region, V. C1, close to the plasma membrane, is thought to be important for dimerization. It is also a binding site for intracellular molecules such as ezrin and tubulin [15]. C2 is situated at the C-terminus of the protein and carries the PDZ-binding domain EFYA [12]. The V region located between the two conserved sequences is very heterogeneous among the mammalian syndecans, but identical in syndecan-1 from human, mouse, rat, and hamster [23]. In the V region only syndecan-4 has a binding site for phosphatidylinositol-4,5-bisphosphate. This site is involved in dimerization of syndecan-4 and plays a critical role in the activation of protein kinase Cα [12, 15].

A common feature of the ectodomain of all syndecans is their susceptibil-ity to cleavage by different proteases that release the ectodomain from the cell surface. This phenomenon is called shedding (Figure 3) and it is known to promote cell migration [24, 25]. Matrix metalloproteinases (MMPs), and in particular MMP-7, MMP-9 and MMP-14, have been suggested as the proteases responsible for shedding [24, 26]. A recent study provides new evidence by showing that also MMP-2 and MMP-3 are capable of shedding

13

both syndecan-1 and syndecan-4. Multiple cleavage sites were also found and mapped using mutagenesis [27]. What controls this process is not fully understood, with suggested regulatory mechanisms both on the inside and outside of the cell membrane. One study shows that the small GTPase Rab5 may regulate the shedding intracellularly by binding to the cytoplasmic do-main of the syndecan [28]. A more recent study suggests that HS chains attached to the extracellular part are responsible for controlling the shedding process [29].

Figure 3. Shedding of syndecan-1. The ectodomain of syndecan-1 (core protein in white) can be released from the cell surface by proteolytic cleavage. One cleavage site (marked with an X) is located close to the plasma membrane and is recognized by a variety of MMPs.

In pathological conditions, such as the environment surrounding a tumor, many factors can induce the shedding of syndecans. An increased level of the HS degrading enzyme heparanase, on its own able to cleave off parts of the HS chains, has been shown to promote shedding by stimulating Erk phosphorylation that in turn leads to upregulation of MMP-9 [30]. There is also a study showing that FGF2 can induce the shedding of syndecan-1 [31]. When the ectodomain is shed, the effect is not only a released core protein but also the removal of its attached GAG chains and any bound growth fac-tors from the cell surface. The released ectodomain also has the capacity of interacting with various molecules and other cells in the surrounding [32].

Glycosaminoglycans Glycosaminoglycans (GAGs) are linear carbohydrate chains, attached to core proteins to form PGs. Due to their high negative charge, GAGs have the ability to interact with a large number of proteins and hence function in many biological processes [33]. GAGs are composed of repeating disaccha-ride units consisting of a hexosamine and a hexuronic acid or a hexose unit. Depending on the building blocks GAGs can be divided into different groups [2]; (i) chondroitin/dermatan sulfate (CS/DS) with D-galactosamine (GalN)

14

and D-glucuronic acid (GlcA) or its epimer L-iduronic acid (IdoA) [8], (ii) heparan sulfate (HS)/heparin with D-glucosamine (GlcN) and GlcA or IdoA [5], or (iii) keratan sulfate that consists of galactose and GlcN. All of these GAGs are sulfated at various positions, however not evenly along the chains resulting in different structural and functional domains [1, 34]. There is also an unsulfated GAG, hyaluronan, made of repeating units of N-acetylated GlcN (GlcNAc) and GlcA. Hyaluronan is not attached to a core protein but instead exists freely in the ECM [2, 35].

Among the GAGs, both CS/DS and HS/heparin are O-linked to selected serine residues in the core protein. Depending on the amino acids flanking the serine residue, the attached glycans may either preferentially be HS or CS chains [36, 37], and thus many core proteins carry one type of GAG chain [12].

Biosynthesis of glycosaminoglycans HS/heparin and CS/DS are produced in the Golgi apparatus by two distinct sets of biosynthetic enzymes. Due to substrate specificity, this creates the respective polysaccharide chain with heterogeneous distribution of sulfate groups (Figure 4) [2, 8, 38]. Although the biosynthetic pathways are sepa-rated, they are still linked together not only through the common linkage region, but also through the use of the same building blocks like the sulfate donor (PAPS) and donor sugars [39, 40].

Heparan sulfate HS is produced in virtually every mammalian cell and is important in both embryonic development and homeostasis in the adult, mainly due to interac-tion with various protein ligands [41]. The biosynthesis of HS is initiated when xylose is transferred from a UDP-sugar precursor to a serine residue in the core protein by xylosyltransferase. Two galactose units followed by a GlcA are consecutively added by specific glycosyltransferases to form the linkage region: xylose-galactose-galactose-glucuronic acid (Figure 4A) [34]. Up to this point the biosynthetic pathways of HS/heparin and CS/DS are common, but the addition of the fifth sugar determines which type of chain that will be elongated (Figure 4B). Addition of glucosamine by exostoses (multiple)-like 3 (EXTL3) to the linkage region will initiate polymerization of the rest of the HS chain (Figure 4C), accomplished by the action of HS polymerases known as exostosin (EXT) enzymes that add alternating saccha-ride units of GlcA and GlcNAc [42]. When knocking out EXT1 and EXT2, mice embryos fail to gastrulate and hence, loss of the EXT genes results in embryonic death at day 6 [43, 44]. Human individuals, heterozygous for either EXT1 or EXT2 suffer from hereditary multiple exostoses, a disease with abnormal overgrowth of bones [42].

15

Figure 4. The biosynthetic pathways of HS and CS. The major steps are initiation of biosynthesis, polymerization and subsequent modifications: sulfation by various O-sulfotransferases (OSTs) and epimerization by C5epimerases (C5-epi). A) Both chain types are polymerized onto a tetrasaccharide linker sequence, O-linked to a serine residue in the core protein. B) The addition of the fifth sugar determines which kind of GAG that will be extended and subsequently modified as shown in C) for HS/heparin and D) for CS/DS. For further details see the text.

As the HS chain is elongated by the EXTs, a series of modifications occur. First, acetyl groups are removed from some of the GlcNAc units and sulfate groups are added to create N-sulfated glucosamine (GlcNS) in a non-uniform manner along the chain. This is performed by a group of enzymes called glucosaminyl N-deacetylase/N-sulfotransferases (NDSTs) that utilize sulfate donated from 3’-phosphoadenosine 5’-phosphosulfate (PAPS) [45]. PAPS is synthesized in the cytosol from inorganic sulfate that has been transported across the plasma membrane. The synthesized PAPS is then transported into the Golgi by a PAPS transporter, where it is used as a substrate in for exam-ple GAG biosynthesis [46]. It has been suggested that PAPS may have a regulatory role in HS biosynthesis by affecting the action of NDST enzymes [47]. The conversion of GlcNAc to GlcNS creates a substrate for the rest of the modification enzymes, since they act in the vicinity of GlcNS [45, 48]. Removal of NDST1 in mice leads to neonatal lethality with a phenotype characterized by lung and skull defects [49]. If, on the other hand NDST2 is knocked out, mice are viable and fertile, but contain defective mast cells due to lack of sulfation of heparin chains [50]. These knockouts suggest that NDST1 can compensate for NDST2 in almost all tissues except connective tissue type mast cells, but not vice versa. Further, NDST1 seems to be re-sponsible for HS biosynthesis and NDST2 for heparin, which was shown in

16

mast cells where oversulfated heparin was produced in the absence of NDST1 [51].

Some GlcA residues are then C5-epimerized to IdoA, leading to higher degree of chain flexibility [52], and the uronic acid can be further modified by sulfation at C2 by the 2-O-sulfotransferase. 2-O-sulfate groups are most often found in NS domains of HS chains and the 2-O-sulfotransferase has a preference for IdoA over GlcA [5, 53]. GlcNAc and GlcNS can be O-sulfated at carbon 3 and 6 by 3-O-sulfotransferase and 6-O-sulfotransferase respectively [5]. The knock out of C5-epimerase is neonatally lethal with mice suffering from renal agenesis, lung defects and also skeletal malfor-mations [54]. A similar phenotype is seen when knocking out the only 2-O-sulfotransferase in mice [55].

Since sulfation and epimerization do not occur evenly along the chain, domains with different degree of modification are formed. NA-domains are characterized by a very low sulfation degree whereas NS-domains are highly sulfated. There are also regions with alternating N-acetylated and N-sulfated units called NA/NS-domains as shown in Figure 4C [1]. The biosynthetic reactions that produce HS occur very rapidly in the Golgi and the production of a chain with all modifications takes approximately one minute [56], and a synthesized PG appears on the cell surface within approximately 15 minutes [57]. How the biosynthetic machinery is organized is still not fully under-stood, but there are studies suggesting the cooperation of many of the en-zymes in a heterogenic complex referred to as the GAGosome [34]. Indeed, EXT1 and EXT2 have been shown to form a heterodimeric complex and the same has been shown for C5-epimerase and 2-O-sulfotransferase [58-61].

Heparin is an extremely highly sulfated variant of HS that is produced solely by connective tissue type mast cells. The physiological role of heparin is to bind proteases and histamine in the mast cells, i.e. to act as a scaffold molecule where the mast cell effector molecules can be stored [50]. When added exogenously to the circulation, heparin has anticoagulant activity be-cause of its binding to antithrombin via a specific pentasaccharide sequence. Heparin derivatives are therefore used clinically as anticoagulants [33, 62].

After biosynthesis, the HS chains can be further modified by endo-6-O-sulfatases 1 and 2 (Sulfs) that remove 6-O-sulfate groups from GlcNS, pref-erably from the NS domains and in particular from trisulfated IdoA2S-GlcNS6S disaccharide units [63]. The Sulfs are extracellular enzymes but seem to be associated to the cell surface and not released into the extracellu-lar environment [64]. Degradation of HS occurs by the cooperation of many enzymes in the lysosomes. One of these enzymes is the endo-β-D-glucuronidase heparanase that cleaves the chains into smaller fragments, usually 10-15 disaccharides in size. Heparanase can also be active extracel-lularly, where it modifies the HS structure on the cell surface [65, 66].

17

Chondroitin sulfate CS is one of the major components in the ECM, for example in the brain where one of its tasks is to maintain plasticity [67]. CS is also found to a large extent in cartilage where the negative charge helps to keep the osmotic resistance to bear heavy compressions [68]. The biosynthesis of CS involves a different set of enzymes than HS, but the CS chains are polymerized onto the same linkage region as HS, attached to a serine residue in the core pro-tein (Figure 4A). Addition of GalNAc by the action of CSGalNAc transfer-ase I induces the assembly of CS (Figure 4B), which is then further polymer-ized by chondroitin synthases (Figure 4D) that add GlcA and GalNAc in an alternating sequence [69, 70]. Six different polymerases have been identi-fied; chondroitin sulfate synthase-1 (CSS1)/chondroitin synthase-1 (ChSy-1), chondroitin sulfate synthase-2 (CSS2)/chondroitin polymerizing factor (ChPF), chondroitin sulfate synthase-3 (CSS3)/chondroitin synthase-2 (ChSy-2), chondroitin sulfate glucuronyl transferase/chondroitin synthase-3 (ChSy-3), and chondroitin N-acetylgalactosaminyltransferase-1, and -2 [71]. The first four are able of adding both GlcA and GalNAc, with the exception of ChPF that lacks enzymatic activity, and the last two have only GalNAc transferase activity. None of the enzymes are capable of assembling the CS chains alone. Instead it has been shown that there is cooperation between at least two, where the most efficient polymerization has been shown with the heterodimeric complex of CSS1 and CSS2. Further, CSS3 has been shown to interact with both CSS2 and CSS1. It has also been suggested that depending on which complex that is formed the length of the chains will differ [72, 73].

Due to the large number of synthases with overlapping enzymatic activi-ties their exact function has not been determined in higher organisms with the exception of CSS2 that was recently knocked out in mice. This lead to viable and fertile mice with shorter CS chains in the cartilage [73]. Knock out studies in Caenorhabditis elegans of the only chondroitin synthase (SQV-5) and the orthologue of CSS2 (PAR2.4) showed the importance of chondroitin in cytokinesis, cell division, organogenesis of the vulva, migra-tion of distal tip cells and neuronal development. These studies also con-firmed the cooperation of different polymerases in vivo [74, 75].

Similar to HS biosynthesis, a series of modifications of the CS chains oc-curs as they are polymerized, the first of which is epimerization of GlcA to IdoA by the dermatan sulfate (DS) epimerases [76, 77]. This modification does not occur evenly along the chain, resulting in blocks with more than six IdoA residues within the chains. The presence of a single disaccharide unit with IdoA changes the terminology and a chain with IdoA is referred to as a DS chain. Many chains are found to have both CS and DS residues, so called hybrid structures [78]. Deficiency of DSepi-1 in mice leads to a decrease in IdoA that affects the structure of collagen in the skin [79]. DSepi-2-/- mice are normal and although a reduction of IdoA in CS/DS chains isolated from

18

the brain was detected, there was no developmental effect, probably due to compensational activity by DSepi-1 [80].

Figure 5. The most common CS disaccharides in mammals. The basic disaccharide in CS is called the O unit and consists of GlcA and GalNAc. As the chains are as-sembled modification reactions occur which lead to epimerization of some GlcA residues to IdoA, giving rise to the iO and iA-iE units. 2-O-sulfotransferases can sulfate both GlcA and IdoA and the GalNAc can be sulfated in both 4-O, and 6-O-position. Based on Malmström [81].

Further modifications in form of sulfation occur and give rise to heterogenei-ty among the chains [8]. The most common CS disaccharides in mammals are shown in Figure 5. It has been shown that DS epimerase-1 co-localizes with a DS specific 4-O-sulfotransferase and hence blocks with a high content of IdoA are often also 4-O-sulfated resulting in presence of iA, iB and iE units as shown in Figure 5 [81]. Disaccharide structures with GlcA are also sulfated by sulfotransferases resulting in units A to E (Figure 5). In mice, there are three CS 4-O-sulfotransferases, one specific DS 4-O-sulfotranferase (as mentioned above), two CS 6-O-sulfotransferases and one CS 4-O-sulfate-6-O-sulfotransferase [82]. Removal of a CS 4-O-sulfotransferase in mice lead to severe chondrodysplasia [83] and the same phenotype was seen in humans suffering from spondyloepiphyseal dysplasia due to a mutation in one of the CS 6-O-sulfotransferase genes [84].

19

Regulation of glycosaminoglycan biosynthesis Knockouts of core proteins as well as biosynthetic enzymes demonstrate the importance of HS and CS chains during development and that the regulation of the produced structures is crucial. A recent knock out of GlcA transferase in mice provides further evidence that HS and CS are critical for cytokinesis during embryogenesis [85]. As shown in Figure 4, the biosynthetic pathways are closely related, starting with the same linkage region, and sharing UDP-sugars and the sulfate donor PAPS [86]. How the production of CS/DS and/or HS/heparin is controlled is still an enigma, and each step in the bio-synthetic pathways is a point of possible variety.

One aspect concerning formation of the linker regions is the production level of core proteins. More available serine residues require higher frequen-cy of xylosylation and a higher throughput of UDP-sugars. The amino acid sequence that surrounds the serine residue also plays a role in regulating which GAG is made. Studies have shown that serine-glycine repeats are favorable for attachment of HS chains and that also hydrophobic and acidic residues in the vicinity are enhancing factors for HS synthesis [36, 37, 87]. The next limiting step in the production of both GAGs is the xylosylation. The extent of this reaction determines how many linker regions that are made. There are competing reactions that can take place on the serine resi-due instead of xylosylation, such as sulfations or other glycosylations [88]. Further, modifications of the linker region can occur, with 2-O-phosphorylation of xylose and 4-O- or 6-O-sulfation of galactose. The first is common for both HS and CS, but a sulfated galactose is found only in CS chains [89, 90].

After completion of the linker, the point of separation between the chains arises and depending on whether EXTL3 or CSGalNAcT1/2 adds the fifth residue the nature of the chain is determined. Here, the availability of en-zyme is a limiting factor and the balance between the two may also affect which GAG is made. It has however been suggested that certain amino acids surrounding the serine residue might favor EXTL3 binding [88]. A recent study suggests that EXTL2 has a regulatory role by adding the fifth residue, a GlcNAc to phosphorylated linker regions, which terminates the polymeri-zation [91]. As already mentioned, the chains are subjected to modifications that occur unevenly, which itself causes further heterogeneity among the chains since each enzyme has substrate specificities and hence recognizes different saccharide units [5].

Also of crucial importance is the localization of the biosynthetic enzymes in the Golgi, as well as their concentration. By using Brefeldin A, studies have shown that enzymes involved in HS biosynthesis are located in the cis-, medial-, and trans-Golgi, but that CS biosynthetic enzymes are located in trans-Golgi [40, 92].

20

Integrins Integrins are the molecules responsible for connecting cells to their sur-rounding matrix. They work as adhesion receptors on the cell surface where they interact with ligands in the ECM and hence attach the cells to their ma-trix [93]. During embryonic development in the formation of the proamniotic cavity, survival of cells is dependent on attachment [94]. The same is true for epithelial, endothelial and muscle cells in the adult that upon loss of their connections with the ECM will undergo apoptosis [95, 96]. The importance of integrins is further proved in the β1 integrin knockout mice, which die at implantation stage [97]. However, not all integrin knockouts are lethal, which demonstrates that these molecules have different roles in a large num-ber of biological functions, and that there may be compensatory effects [98].

The integrins are heterodimeric transmembrane proteins that are non-covalently associated to each other. Each complex consists of one α and one β subunit as shown in Figure 6A, and there are 24 known different combina-tions of assembly. All integrins span across the plasma membrane separating a large ectodomain, and a short cytoplasmic domain. The latter initiates the formation of focal adhesion complexes that will connect the cytoskeleton of cells to the ECM [98, 99]. Focal adhesions, which are located at the end of actin fibers, are large assemblies of proteins including integrins, syndecan-4 and a large number of cytoplasmic proteins, like vinculin and α-actinin. Fo-cal adhesions act as centers for signaling between the ECM and cytoskeletal proteins [100].

Figure 6. Integrins. A) Illustration of a heterodimeric integrin complex comprising one α (dark grey) and one β (light grey) unit. B) Shows the group of RGD-binding integrins. Different possible heterodimers are connected with lines, like for example αvβ5.

21

It is important that the interaction between integrins and the ECM is of mod-erate strength to enable movement of cells. Because of that, integrins are present at the cell surface in the inactive state and hence have low affinity to their ligands [101]. They can become activated due to a conformational change that can be triggered in two ways. Either, intracellular signaling will induce the activation, “inside-out”, where the binding of talin to the cyto-plasmic domain of the β integrin is the final step [102, 103]. Integrins can also bind to their extracellular ligands, which leads to “outside-in” signaling events that control cell mobility, proliferation and differentiation [104].

Integrins can be divided into different groups depending on their ligands: (i) collagen type I-binding integrins, (ii) RGD-binding integrins, (iii) lam-inin-binding integrins and (iv) leukocyte specific integrins. The collagen type I-binding integrins all include the β1 subunit in complex with α1, α2, α10 or α11. RGD-binding integrins, schematically represented in Figure 6B, are classified due to the specific peptide sequence arginine (R)-glycine (G)-aspartic acid (D) recognized in their ligands. This sequence is present in many of the matricellular proteins such as vitronectin, fibrinogen and fibron-ectin [93, 98, 105]. Fibronectin is a plasma protein, encoded by one gene but due to alternative splicing 20 different isoforms are found in human. Fibron-ectin is abundant in plasma, where it circulates in an inactive form, and it is also an important part of the ECM [106]. Further, it has binding sites for different growth factors and can also bind HS chains attached to for example syndecan molecules [107].

Integrins and cell migration Movement of cells is a key factor during embryogenesis when cells migrate to form the different germ layers in the embryo. Subsequently, cells migrate further and differentiate at specific locations to form the different tissues and organs. Migration is also of great importance in the adult where cells for example migrate to close wounds or to repair tissue. The process is also im-portant in pathological situations like cancer cell metastasis [108]. There are different modes of migration, either individual movements (amoeboid or mesenchymal migration) or collective, when multicellular units migrate. The basic mechanisms are however the same, with cells translocating along or through a substrate by constantly remodeling the actin cytoskeleton and fo-cal adhesion sites. The mesenchymal migration process can be divided into three steps: cell polarization, formation of protrusions and adhesions, and retraction of the rear of the cell [101, 108, 109]. Integrins are crucial for cell adhesion and are thus present at the leading edge of a migrating cell [110]. There is however, a constant turnover of focal adhesion sites and the integ-rins are hence trafficked or recycled as these sites are disassembled in the rear of the cell while new ones are formed in the front. It has been shown that cleavage of talin is important when adhesion sites are disrupted and

22

focal adhesion kinase (FAK) and Src family kinases are also important in the turnover of these protein assemblies [111, 112].

It has been shown that depending on which integrin that is used in binding to fibronectin, the migratory pattern of cells can be altered. Cells that used α5β1 showed a more fibroblastic morphology and more random migratory behavior, whereas cells using αvβ3 migrated in a more persistent fashion with adhesion distributed evenly over the ventral surface [113].

Fibroblast growth factors and their receptors Fibroblast growth factors (FGFs) play crucial roles in embryogenesis by regulating cell proliferation, migration and differentiation [114]. In both human and mouse, the family of FGFs comprises 22 members, where 18 are capable of stimulating tyrosine kinase receptors [115]. In contrast, only one fgf gene has been found in Drosophila melanogaster and two fgf genes have been identified in Caenorhabditis elegans [116, 117]. In the adult organism the FGFs are involved in regulating for example angiogenesis, tissue repair and wound healing [118]. Studies from knockout mice indicate that the vari-ous FGFs have different roles. The loss of FGF4 and FGF8 leads to lethality at embryonic days 4-5 and 7, respectively, whereas mice without FGF1 and FGF2 are viable with mild phenotypes [117].

The majority of FGFs are secreted glycoproteins, expressed in many or-gans both during embryogenesis and in the adult [119]. FGF1 and FGF2 do not have secretory signals. Instead, these proteins are either exported out of the cells via an exocytotic pathway independent of the ER-Golgi pathway, or released as a result of cell damage [116, 120]. FGF11-FGF14 act on intracel-lular targets in a receptor-independent manner, but the mechanism underly-ing this is still unknown. Members of the FGF19 subfamily (FGF19, 21 and 23) are poor activators of FGF-receptors and are thought to be involved in metabolic regulation [116]. Disturbed expression of FGFs (e.g. 10 and 23) in human has been reported to result in hereditary diseases and some of the FGFs (3, 4 and 8) have also been proposed as oncogenes [121].

The FGF molecules bind to and activate four FGF tyrosine kinase recep-tors (FGFRs 1-4) expressed on most cell types. The structure of the four receptors is similar with an extracellular domain consisting of two or three immunoglobulin-like folds, depending on the splice variant. The extracellu-lar domain is connected to the tyrosine kinase domains on the intracellular part of the receptor via a short transmembrane sequence [122]. The different FGFs have variable affinity and selectivity for different receptor types. FGF1 is for example capable of binding all FGFRs, whereas FGF10 activates FGFR2b only [123]. Upon binding of growth factors, the receptors form dimers and trans-phosphorylation of the intracellular kinase domains occurs in a specific order [124]. The FGFRs are known to be upregulated in several

23

cancers, for example FGFR2 and FGFR3 in bladder cancer and endometrial cancer [125] and FGFR1 in breast cancer [126]. There is also a fifth receptor with a so far unknown role, FGFR5 that lacks kinase activity [127].

Fibroblast growth factors and heparan sulfate The interaction between HS and FGFs has been well studied. In many inves-tigations, FGF2 and heparin have been used as prototype molecules in at-tempts to understand the interplay between growth factors and GAGs, and how the latter may help to control morphogenic gradients during embryo-genesis [119, 128]. The binding between HS and FGFs serves both to protect the growth factors from degradation and to store them in the ECM [119]. HS is also an important component in the ternary signaling complex consisting of FGFs, FGFRs and HS/heparin [129-131].

The ternary complex that is formed has been suggested to consist of FGFs, FGFRs and HS in a symmetric 2:2:2 [132], or in an asymmetric 2:2:1 complex where the HS chain would act as a bridge in the latter model [133]. The exact structure of the complex is still under debate. Further, it has been shown that the dimerization of FGFRs is stabilized by heparin [134] and that a higher degree of O-sulfation also increases the stability of the ternary com-plex [135]. Some studies suggest sequence dependence when HS interacts with FGFs [34]. Others show that the domain organization of sulfate groups along the HS chain regulates the FGF-induced cellular response and its dura-tion [136]. Clearly, the concept of specificity is not yet fully understood, but in a recent study it was shown that FGFs from different subfamilies recog-nize binding sites in heparin of different sizes and with variable binding ki-netics [137].

Fibroblast growth factor induced signaling The phosphorylation of the kinase domains of FGFRs creates binding sites for a number of molecules intracellularly and triggers different signaling cascades as shown in Figure 7 [126]. The different pathways will lead to different cellular responses ranging from differentiation via protein kinase C to anti-apoptotic signals via protein kinase B/Akt activation [118, 122, 138]. Furthermore, FGFs are known inducers of angiogenesis and FGF2 for exam-ple, is a very potent proangiogenic factor [126]. The cellular response to an FGF molecule is not only cell type specific, but also depends on the stage that the cell is in [139]. Further, the stability of the ternary complex [136] as well as the phosphorylation pattern of the kinase domain in the receptor can modulate the response. A study has shown that FGF2 alone induces phos-phorylation of tyrosine 463 of the receptor, whereas the combination of FGF2 and heparin stimulates phosphorylation of tyrosine 766 as well [134].

24

Different adaptor molecules also bind to different tyrosine residues and hence HS/heparin may have a possible role in fine-tuning signals.

Figure 7. Intracellular signaling pathways downstream of FGFR. FGFs bind extracellularly to their transmembrane receptors with HS chains as co-receptors. Multiple signaling cascades can become activated, and intracellular mediators trans-fer the signal to the nucleus. Several different outcomes have been reported (grey boxes under the black line). Based on Dailey, Eswarakumar, Murakami and Daniele [118, 126, 138, 139].

FGF receptor substrate 2 (FRS2) is an adaptor molecule that serves as a core onto which intracellular proteins form signaling complexes. When son of sevenless (SOS) and growth factor receptor bound-protein 2 (Grb2) bind to FRS2, the MAPK pathway is activated via Ras and Raf (Figure 7, left path-way). Downstream, this will lead to phosphorylation of Extracellular signal-regulated kinase 1/2 (Erk 1/2), a transcription factor that translocates to the

25

nucleus and regulates transcription of various genes [140]. The MAPK cas-cade is proposed as the dominating pathway upon FGF stimulation [118].

If on the other hand Grb2-associated binder 1 (Gab1) binds to the Grb2-FRS2 complex, the PI3-kinase pathway is activated (Figure 7, right path-way). Via second messengers, PKB/Akt is recruited to the plasma membrane where it binds through the pleckstrin homology domain. After phosphoryla-tion, PKB/Akt is activated and leaves the membrane to either the cytosol or the nucleus where it is part of regulating for example metabolism and sur-vival of cells [141].

Phospholipase Cγ binds to phosphorylated tyrosine 766 on the FGFR via its Src homology domain. This in turn leads to the hydrolysis of phosphati-dylinositol 4,5-bisphosphate into diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). DAG in turn activates protein kinase C and IP3 induces the release of Ca2+ leading to growth control and differentiation, shown in Figure 7 (middle pathway) [142].

The effects of HS structure alterations in signaling There are many ways for a cell to regulate the intracellular signaling. One is through the internalization of receptors that leads to degradation, or negative feedback loops that lead to attenuation of signals [126]. Another is through the action of extracellular HS-modifying enzymes such as Sulfs and hepara-nase since their activity alters the fine structure of HS, or the proteolytic cleavage of core proteins from the cell surface.

Sulfs cleave 6-O-sulfate groups from NS domains, which results in chains with lower sulfate density. These, in turn show reduced affinity for ligands like Wnts (Wingless and Integration1) and instead Wnts can bind to the Frizzled receptor and induce a cellular response. Without the actions of Sulfs, Wnts bind strongly to HSPGs instead and the receptor (Frizzled) is not activated. Concerning FGFs, removal of 6-O-sulfate groups reduces the in-tracellular signaling due to less binding of FGF and hence the response from cells is lowered [63, 64, 143, 144].

Heparanase can affect and regulate the signaling in two different ways. By cleaving HS chains in the ECM, stored growth factors can be released from their deposits and stimulate receptors in the surrounding, which will lead to enhanced signaling [145]. However, cleavage of HS chains can also lead to chains that are able to bind the ligand FGF, but not long enough to stabilize the FGF/FGFR complex, which will result in decreased FGF signal-ing.

26

Glycosaminoglycans in pathology As described previously, PGs and GAGs are crucial during embryonic de-velopment and the inability to synthesize HS and CS leads to embryonic death at the 8-cell stage in mice [85]. Mutations in genes involved in the synthesis of either core proteins or GAG chains lead to several disorders such as Ehlers-Danlos syndrome, hereditary multiple exostoses and spondy-loepiphyseal dysplasia [42, 84, 146]. Disturbance in the degradation of GAGs can also lead to disease, which is seen in mucopolysaccharidoses such as mucopolysaccharidosis type I (Hurler syndrome). In Hurler syndrome oversulfated HS is accumulated in the lysosomes and positively regulates continued biosynthesis of HS [147]. The above are examples of genetic mu-tations leading to disorders, however normally produced GAGs are also in-volved in many diseases such as osteoarthritis, amyloid diseases, tumor pro-gression and metastasis.

HSPGs in tumor progression and metastasis Due to their localization and their ability to interact with matricellular mole-cules, signaling proteins and intracellular proteins, the HSPGs can affect both tumor formation as well as the invasiveness of tumor cells [148]. It has also been shown that after expression of HSPGs, alterations can occur that further affect their role in disease progression, like cleavage of the core pro-tein, removal of sulfate groups from the HS chains and cleavage or shorten-ing of the HS chains [149]. These mechanisms all have the effect of altering HS mediated signaling, and thus contribute to disease progression.

Some PGs have been more studied than others, for example the basement membrane PG perlecan, that is upregulated in invasive melanomas [149, 150]. It has been shown that perlecan has different roles in cancer progres-sion by regulating angiogenesis. One is promoting angiogenesis via interac-tions with enhancing growth factors like FGF-2 and VEGFA. Another is via the action of endorepellin, the C-terminal fragment of perlecan, that has been shown to have anti-angiogenic properties by for example interacting with α2β1 integrins and thereby disturbing the motility of endothelial cells [149].

The syndecan family has also been found involved in many different can-cers such as breast and ovarian cancer [151], multiple myeloma [152], hepa-tocellular carcinoma and colon carcinoma [153], to mention some. It appears that both overexpression and loss of syndecan-1 can lead to more aggressive tumors and increased invasiveness in cancers of the pancreas and breast, and throat and lung, respectively [154-157].

Syndecan-1 has different roles in breast cancer progression, depending on whether it is bound to the plasma membrane or released after proteolytic cleavage. One study has shown that soluble syndecan promoted the inva-siveness of the human breast cancer cell line MCF-7, whereas the membrane bound form suppressed it [158]. The proteolytic shedding of syndecan-1,

27

mediated by MMP-14, also promoted the proliferation of another human breast carcinoma cell line (T47D) [159, 160]. Furthermore, the ectodomain of syndecan-1 could promote the growth of myeloma tumors, which was shown in a study where severe combined immunodeficient (SCID) mice developed tumors in human femur (thigh bone) grafts after injection of lym-phoid cells that produced the soluble form of syndecan-1 [161]. When syndecan-1 was overexpressed in malignant mesothelioma cells on the other hand, it resulted in hampered proliferation of these cells, caused by both full length and truncated forms of syndecan-1 [162]. Clearly, the various forms of syndecan-1 have different roles in cancer, and further knowledge about these mechanisms can help in developing treatments.

The MCF-7 cell line was used to establish the effect of HSPGs in FGF2-induced signaling, where HS chains isolated from the breast carcinoma cells could promote the ternary FGF-signaling complex and thereby determined the mitogenic response [163]. Further, when mouse embryonic fibroblasts were co-cultured with invasive human breast carcinoma cells (MDA-MB-231) syndecan-1 expression was induced by the carcinoma cells, and cell growth was stimulated via the attached HS chains [164]. The requirement of syndecan-1 in formation of Wnt-1-induced mammary gland tumor was also shown using syndecan-1-/- mice [17].

A study from 2011 using human breast tissue showed that syndecan-1 and syndecan-4 are independent markers in breast cancers. Expression of syndecan-1 was associated with poor prognosis in breast carcinomas, espe-cially when detected in both the epithelium and the stroma. Syndecan-4 ex-pression was however more related to the status of estrogen and progester-one receptors, which could indicate a better prognosis [165]. A clinical study from 2004 also showed that syndecan-1 expression in the epithelium and the stroma was associated with poor outcome for the patients [166].

The microenvironment surrounding tumors is a dynamic area where many molecules are dysregulated and thereby affect the growth of tumor cells and their metastatic behavior. One example is the syndecan-1/heparanase axis [167]. Heparanase is often upregulated in the tumor surroundings and can enhance the expression of syndecan-1. Further, it can also induce the shed-ding of syndecan-1 by upregulating the expression of MMP-9 via the ERK signaling pathway, possibly mediated by the insulin receptor [30, 168, 169]. It is possible that the role of the attached HS chains is to protect the syndecan core protein from proteases, and hence after heparanase digestion of HS chains the protein is more susceptible to cleavage. In this context, several heparin mimetics have reached clinical trials, but their effect is still not evaluated [167, 170].

28

CSPGs in tumor growth In investigations concerning the role of CSPGs in cancer, most studies have focused on their expression in the tumor stroma. Many of the CSPGs are large molecules known to be architectural components of the ECM where they can affect proliferation and migration of cancer cells by interacting with a large number of growth factors and chemokines [8, 81]. Aggrecan is a CSPG that has been found upregulated in many cartilage tumors. Versican is another example of a CSPG that is expressed in the stroma of almost all hu-man cancers, especially in less differentiated tumors [171]. This was con-firmed in a breast cancer study, where versican was found only in the pe-riphery of the tumors and not in the centers [172].

Decorin belongs to the family of small leucine-rich PGs and like syndecan it has been found both upregulated and repressed in different can-cers. Decorin is for example overexpressed in osteosarcoma but shows low-ered expression in breast cancer [171]. Overexpression of decorin seems to also affect the fine structure of attached GAGs, so that the polysaccharide sulfation is altered [173].

NG2 is a transmembrane CSPG also known as melanoma-associated PG, that is upregulated in proliferative cells and also found in many tumor cells [12]. NG2 has been most studied in melanoma, but is also found in various tumors of the nervous system such as glioma and astrocytoma [174].

Generally, the stroma surrounding tumors as well as fibrotic tumor tissues have a high content of PGs, possibly with more CS than HS [8], and it has been shown that CSPGs such as decorin and biglycan can inhibit growth of some cancer cell lines [171]. Further, treatment of melanoma or endothelial cells with chondroitinases to remove CS/DS indicated that the chains may regulate metastasis, since proliferation and invasion was inhibited after re-moval of the chains [175]. Considering growth factor binding, alterations in fine structure of CS/DS resulting in a higher content of highly sulfated di-saccharides can also lead to situations where the CS chains compete for hep-arin binding sites in ligand molecules. Such increase in sulfation has been shown in invading melanoma cells [176]. CS/DS chains are potential targets for therapeutic drugs, however the structural details must be resolved first [8, 177].

29

Present Investigations

Aim The aim of this study was to investigate the role of HS and CS at the cell surface, how they may affect different cellular activities, and whether there is a dependence on specific structures or a functional overlap (summarized in Figure 8). The specific aim for each project was to:

• Investigate the role of HS and CS in cell adhesion and migration, i.e. how these GAGs can interact with matricellular proteins and integ-rins, and influence their functions.

• Study how the mode of presentation of HS, and possibly CS, to the FGF signaling complex affects the response from cells.

• Elucidate the role of GAGs in pharyngeal cartilage formation in

zebrafish and further investigate the regulation of GAG biosynthesis.

Figure 8. A schematic representation of the aims of this study. Involvement of GAGs in adhesion and migration include interactions with the ECM (1) as well as integrins (2). In fibroblast growth factor induced signaling, HS chains are a prereq-uisite for a stable ternary complex formation (3).

30

Model systems

Chinese hamster ovary cells In 1985 an article reported on Chinese hamster ovary (CHO) cell mutants that were defective in the biosynthesis of GAGs. These mutants were ob-tained by chemical mutagenesis followed by replica screening and resulted in a library of mutants with different phenotypes [178]. It is worth noting that the wild type cells CHO-K1, express a limited number of biosynthetic modifying enzymes and core proteins, and hence the produced PGs may not show all activities of biological importance. However, there are NDSTs as well as O-sulfotransferases, albeit not as many isoforms as in human and mouse. The only exception is the complete lack of 3-O-sulfotransferases. Further, it has been shown that levels of produced GAGs in CHO-K1 corre-spond to, or are moderately less than, the levels produced in for example human fibroblasts [179].

Along with the wild type cells CHO-K1, two mutants have been used in this project (Paper II and III): the GAG deficient pgsA (CHO-745) and the HS deficient pgsD (CHO-677). CHO-745 cannot produce GAGs due to a mutation in xylosyltransferase-2. CHO-677 have instead a mutation in EXT1 rendering the cells unable to polymerize HS [179]. CHO-K1 and CHO-677 were further modified to express various forms of syndecan-1, shown in Figure 9. One of these syndecan isoforms (Figure 9, middle panel) has a mutation in the protease cleavage site and will stay bound to the plasma membrane, i.e. the uncleavable (uc) syndecan-1. The second lacks the cyto-plasmic domain, which results in a syndecan-1 that will be constitutively shed (se) into the culture medium (Figure 9, lower panel) [32].

Figure 9. A schematic presentation of syndecan-1 constructs. The top panel shows wild type syndecan-1 that consists of an extracellular domain (white) where attach-ment sites for GAGs are situated, a membrane spanning domain (TM) and a short cytoplasmic domain (CD). The uncleavable (uc) form of syndecan-1 (middle panel) has a mutated protease cleavage site and can hence not be removed by MMPs. The constitutively shed ectodomain (se) (lower panel) is produced without both the TM and CD and is constitutively released into the surrounding of the cell.

31

Zebrafish The zebrafish (Danio rerio) has been used as a research tool and model sys-tem for many years and there are many advantages to using zebrafish as compared to other model systems such as mice. Zebrafish are easily bred and are relatively cheap and easy to keep [180, 181]. Further, the embryos are transparent which allows visualization of internal structures, and they have a relatively fast developmental cycle where heart beating appears 24 hours post fertilization. They start to swim two days post fertilization and are ca-pable of eating after five days [182]. Figure 10 shows a zebrafish embryo four days post fertilization. Another advantage of this model is the large number of progeny that can be obtained from a single breeding pair, which can be used for comparison in large-scale experiments [183].

The possibility to easily modify the fish genetically is another great ad-vantage. Genes can be overexpressed by injection of RNA or DNA, where the first is preferred to avoid mosaic expression patterns. Genes can also be transiently silenced by injections of morpholinos, which are synthetic anti-sense oligonucleotides that bind to the corresponding mRNA and hence block translation [184]. One problem with morpholino injections is the pos-sible off-target effects that can be induced, which can lead to cell death in the nervous system. It has been shown that these apoptotic effects are medi-ated via p53, and co-injections of morpholinos against p53 can therefore help overcome this problem [185]. Other off-target effects can also be screened for through the injection of non-specific morpholinos.

In the study presented in this thesis (Paper IV), we have used a number of mutant lines of zebrafish. uxs1 and b3gat3, have impaired biosynthesis of GAGs due to mutations in genes that are important for formation of the tetrasaccharide linker sequence. extl3 and ext2 on the other hand, have im-paired HS biosynthesis. We also used morpholinos to silence the CS syn-thases, and by combining these models we could study the influence of HS and CS in cartilage formation, both separately and combined.

Figure 10. A zebrafish embryo four days post fertilization. At this time point the embryo is approximately 3.7 mm in length [182]. Photo from Paper IV.

32

Results and Discussion

Paper I

The Mutual Impact of Syndecan-1 and Its Glycosaminoglycan Chains – A Multivariable puzzle In this review we tried to highlight the complexity that surrounds the study of PGs in general, and syndecan-1 in particular. PGs are known to interact with a large number of molecules, mostly assigned to the negative charge of the attached GAG chains [3, 33]. However, there are also indications that the core protein on its own interacts with ligand molecules [186], and it is not always an easy task to distinguish the origin of the effects seen in a complex system. Further, depending on which core protein the GAGs are attached to, they will end up in different parts of the plasma membrane, with for example glypicans in lipid rafts [187] and syndecans in focal adhesion sites [188]. Their respective distance relative to the plasma membrane may also be al-tered depending on the core protein, where HS chains attached to syndecans could be further from the membrane, compared to glypican-attached HS chains.

An additional level of complexity is added when the same core protein is modified with two types of GAG chains, like syndecan-1 that can carry both HS and CS chains. Studies have shown that there might be amino acid se-quences that favor binding and action of HS biosynthetic enzymes and hence the regulation of which GAG is attached would actually be controlled on DNA level [37]. Furthermore, these chains may have different structures depending on when and where they are produced. Thus, it has been shown that GAGs can be different in size and fine structure when produced in dif-ferent cell types [189]. However, HS chains isolated from syndecan-1 and syndecan-4 in normal murine mammary gland (NMuMG) cells were indis-tinguishable, while the CS chains showed heterogeneity [190-193].

Why are different types of chains attached? Is it to reach specific purposes or is there a functional overlap between HS and CS? It is not unlikely that there are different roles for HS and CS on syndecan-1. It might be so that HS chains are attached at the outer N-terminal tip of the molecule to trap and interact with growth factors, whereas CS chains are localized closer to the plasma membrane to hinder or interact with for example MMPs or integrins. But, also CS chains can interact with growth factors [8], and further, some of the MMPs have been shown to be heparin-binding proteins and hence able to bind HS chains [26]. The question of a functional overlap versus selectivity is indeed intriguing, complex, and still not resolved.

Yet another aspect is the role of the core protein in presenting GAG chains. The syndecans have the ability to oligomerize [194] which leads to clusters of GAG chains that may in turn increase the chances of encounter-

33

ing binding partners and enhance affinity due to polyvalency. Using syndecan-4-/- fibroblast it has been demonstrated that only multiple HS chains could rescue a defect in focal adhesions, indicating that valency is indeed crucial for the function of syndecan-4 in focal adhesion sites [195].

The mutual effects that the two parts of the same molecule have on each other might be the basis for the involvement of syndecans in such a large number of biological functions. Future studies of the molecular interactions underlying cellular regulation are needed and the influence from glycosyla-tion should not be forgotten.

Paper II

Cell Surface Heparan Sulfate Chains are Important for Integrin-Mediated Cell Motility Cells are under constant influence of their surrounding, and molecules on the cell surface take part in interpreting the signals. Along this line, the question of cooperation on the cell surface between syndecans and integrins has been raised [196]. This is not an unlikely thought since they are both present at the cell surface, both span across the plasma membrane and both are involved in signaling events [12, 197]. In tumor microenvironments, a large number of molecules have altered expression patterns, leading to uncontrolled growth and movement. For therapeutic means, an increased understanding in the roles of both syndecan-1 and its attached GAG chains is of importance. The aim of this study was to further investigate the impact of HS and CS attached to syndecan-1 on the cell surface, in processes involved in cell motility, like adhesion and migration.

By transfecting CHO-K1 and CHO-677 with an uncleavable (uc) form of syndecan-1 (Figure 9) we could retain the PGs and hence the GAGs, on the cell surface. Upon transfection, it was obvious that the expression of uc syndecan-1 affected the morphology and adherence of the cells. This was seen both in CHO-K1 and HS deficient CHO-677 cells, where expression of uc syndecan-1 made cells more elongated. We investigated the expression levels of uc syndecan-1 and found that different clones expressed various amounts and we therefore continued the study with one high (H) and one low (L) expressing clone from each cell line: K1ucL, K1ucH, 677ucL and 677ucH along with the untransfected wild type cells (wt).

Since an overexpression of a core protein can lead to alterations in the glycosylation, the amounts of GAGs that were produced were analyzed. We could see that overexpression of uc syndecan-1 did not affect levels of re-covered HS drastically, but instead resulted in a greater increase in the amount of CS. We interpret, that an increase in core protein, and hence an increase in available linker regions does not trigger the HS biosynthesis to an overproduction, possibly due to tight regulation. Instead, the CS biosynthetic

34

machinery, believed to be in the later compartments of the Golgi, can use available linker regions and building blocks [92, 198].

In the functional assays we could see a clear influence of the HS chains attached to uc syndecan-1 in both adhesion, cell mediated gel contraction as well as migration. In the adhesion assay we could confirm that CHO cells in general rely on RGD-binding integrins such as αvβ5 [199], since they ad-hered on both fibronectin and vitronectin. Moreover, K1wt and K1ucH were fully spread (with lamellipodia) on fibronectin.

In cell mediated gel contraction, we used composite gels consisting of a mixture of fibronectin and collagen type I to further investigate how cell surface HS and CS could influence the clustering and possibly trafficking of integrins. In K1ucH cells, the integrins stayed active on the cell surface and the cells contracted the gels efficiently. This was not due to the amount of uc syndecan-1 alone with any GAG chains, since 677ucH were unable to con-tract the gels. 677wt contracted the gels as well, and although this effect was not significant, the role of CS may not be ruled out. Cells that were efficient in contracting the gels were also poor in closing the wound in the scratch wound assay, which supports the idea that HS chains on syndecan-1 can interact with integrins and affect their stability on the cell surface and possi-bly their recycling. There are contradictory results showing that the core protein instead is the crucial component when syndecan-1 on mammary car-cinoma cells interacts with αvβ3 integrin. This was shown by enzymatic removal of GAGs in adhesion experiments of cells on vitronectin [200, 201]. It should be noted that biosynthesis of PGs and GAGs is a rather rapid pro-cess, and to exclude the effect of GAGs based on enzymatic removal may not be feasible. However, it is also possible that interactions between syndecan-1 and different integrins in different cells are regulated in different ways and hence depend on different parts of the molecule.

The results presented in this manuscript point towards a role for HS chains on syndecan-1 at the cell surface, where they may participate in in-ducing clustering of integrins, which in turn would reduce the cells ability to migrate and hence might be an important regulator in cancer metastasis. Considering studies regarding the role of syndecan-1 in breast cancer it is intriguing that depending on whether syndecan-1 is bound to the plasma membrane or released from the cell it has different roles [158]. It is also possible that small changes in the biosynthetic machinery can lead to chang-es in what syndecan-1 can do, depending on the fine structure of its attached GAG chains.

35

Paper III

Role of Glycosaminoglycans on Syndecan-1 in Fibroblast Growth Factor Signaling HS is a known co-receptor in the ternary complex that is formed upon bind-ing of FGFs to their receptors [129, 131]. By stimulating cells with soluble GAG chains or heparin oligosaccharides we have previously shown that the sulfation degree is important for the stability of the signaling complex and further, that the fine structure of HS can influence the duration of the intra-cellular signals [135, 136]. But how does the presentation of HS chains, ei-ther attached to a membrane bound core protein or presented in the ECM, affect the ternary complex formation and cell stimulation? To this end, we used CHO cells, both wild type (CHO-K1) and HS deficient (CHO-677) and transfected them with various forms of syndecan-1, shown in Figure 9. This resulted in a collection of clones, where syndecan-1 molecules with or with-out HS were either released into the culture medium or remained bound to the cell surface. These cells were stimulated with FGF2 and the induced intracellular signals were recorded as phosphorylation of Erk1/2.

In this model system, HS deficient CHO-677 cells were not able to re-spond to FGF2, while the wild type cells were. Upon expression of uc syndecan-1 however, a response could be detected from both cell types, with a higher signal from CHO-K1. This was seen at low expression levels of uc syndecan-1 and shows that GAGs that are presented at the cell surface, and stay there, may induce the stability of the ternary complex or the growth factors themselves, even if they are of CS type. Increased expression of uc syndecan-1 did not elevate the signals since the low and high expressing clones in both CHO-K1 and CHO-677 responded to at least the same level.

In CHO-K1 with low expression of uc syndecan-1 we could see a strong-er signal after stimulation with FGF2 as compared to the high expressing cells. This lead us to further look into the fine structure of the attached GAGs. Detailed chain composition analysis by RPIP-HPLC showed that the high expressing clone made chains that were slightly less sulfated, due to a decrease in trisulfated disaccharides. We also used size exclusion chroma-tography to determine chain length, and in combination with enzymatic and chemical cleavage we identified the different GAG populations. The HS chains were shorter than CS chains but of similar size in both the high and low expressing CHO-K1 clones. The CS chains on the other hand, differed more in length between low and high uc syndecan-1 expressers, where the first presented longer chains. This is contradictory to the classic picture of syndecan-1, where HS chains are often presented as longer in the N-terminus and CS as short chains close to the plasma membrane. It is not unlikely that there is great variability in the chains, both depending on the cell type that they are produced in, but also depending on the production pressure and

36

building block availability. Shorter HS chains can also be a result of hepara-nase action.

To test the influence of mode of presentation we purified syndecan-1 ec-todomains that were produced in a clone each from CHO-K1 and CHO-677 with similar mRNA levels of shed ectodomain (se) syndecan-1. These ecto-domains were used together with FGF2 to stimulate the GAG deficient CHO-745 cells. Interestingly, both kinds of ectodomains were equally effi-cient in scavenging FGF2 from the cell surface and hence the phosphoryla-tion of Erk1/2 decreased. Again, pointing towards a functional overlap be-tween HS and CS, at least to a certain extent.

The intracellular signaling pathways that are triggered in a cell upon FGF-stimulation are many with quite diverse outcomes (Figure 7). We hypothe-sized that HS chains, as part of the ternary complex, might have a role in fine-tuning the signals. Here we could see that both HS and CS were able to induce a response from cells after FGF2 stimulation, and that the mode of presentation had an influence. The uncleavable form of syndecan-1 en-hanced the stability of the complex and we could detect elevated signaling levels in these cells. The ectodomain bound GAGs on the other hand, were efficient in scavenging the growth factors from reaching the cell surface and hence the cellular response was decreased. But, in a longer time frame it is possible that they can stimulate surrounding cells with the GAG-bound growth factors to trigger a response. Indeed, many breast cancer studies have shown that elevated levels of the ectodomain of syndecan-1 is correlated with poor prognosis, and that the shed ectodomain can increase the invasive-ness of breast cancer cells [151, 158-160]. Seemingly, it is important how the GAGs are presented and where they are located, when cells respond to FGF2.

Paper IV

On the Roles and Regulation of Chondroitin Sulfate and Heparan Sulfate in Zebrafish Pharyngeal Cartilage Morphogenesis In cell based systems we could see a connection between overexpression of a core protein and the amount of recovered GAGs, as well as a possible func-tional overlap between HS and CS. To further study the interplay between the two biosynthetic pathways we used mutant lines of zebrafish as well as morpholinos to suppress gene expression, with the aim to elucidate the role of GAGs in the formation of pharyngeal cartilage. The pharyngeal cartilage is part of the architecture that provides support for the head and jaw in the zebrafish. It forms after migration of cells from the neural crest during the first embryonic day [202]. After differentiation, the chondrocytes secrete ECM consisting of PGs and hyaluronan, and they are then organized into

37

cartilage at approximately 50 hours post fertilization, forming stacks of flat-tened chondrocytes.

Mutants uxs1 and b3gat3 that cannot form the linkage tetrasaccharide and hence have disturbed synthesis of both HS and CS, showed a disturbed mor-phogenesis in the formation of the pharyngeal cartilage. They were shorter and thicker compared to the control larvae. This was also the case for ext2 mutants that cannot polymerize HS. The HS-initiation mutant, extl3 howev-er, showed a milder phenotype than the other three, and produced increased amounts of CS. We could determine that the levels of CS correlated well with the appearance of the cartilage. Further, the increased amounts of CS in extl3 mutants did not result in longer chains as determined by size exclusion chromatography, but were accompanied by an increased expression of bi-glycan and aggrecan, two CSPG core proteins. The linkage tetrasaccharide mutants, uxs1 and b3gat3 still showed production of HS, to a level of ap-proximately 50 % of the control larvae, and the ext2 mutants made less HS than extl3 mutants. In an axon sorting study, similar results were obtained showing less HS in the ext2 mutants compared to extl3 mutants, explained by maternal contribution [203].

A detailed analysis of the chain composition in the mutant fish showed an increase in sulfation of HS chains in the ext2 and extl3 mutants. The linkage tetrasaccharide mutants, uxs1 and b3gat3 on the other hand, showed a de-crease in 6-O-sulfation accompanied by an increase in 4-O-sulfation in their CS chains. Crossing ext2 with uxs1 generated a double mutant where very low amounts of GAGs were produced. These mutants also showed the strongest cartilage phenotype.

From the linkage tetrasaccharide mutants we could see that there was in-terplay between the EXTL3 and CSGalNAcT1/2 enzymes that affected the ratio of HS and CS production. To further study this event, we transfected HEK293 cells with a syndecan-4 construct where only one attachment site for GAGs was left. In combination with siRNA against Extl3, we could demonstrate that the availability of enzymes at the crossroad of the HS and CS biosynthetic pathways (Figure 4B) affected the produced syndecan-4. Indeed, when EXTL3 was silenced we could see an increase in attached CS to syndecan-4 and hence the ratio between HS/CS had been shifted.

In conclusion, by using a set of mutant zebrafish lines we could determine that chondrocyte intercalation is dependent on HS and that CS is crucial for the formation of the ECM surrounding the chondrocytes. This study also provided some insight into the interplay between the HS and CS biosynthetic enzymes, as well as evidence for a prioritized HS biosynthesis.

38

Concluding remarks

The involvement and importance of HS and CS in development, homeostasis and disease is evident. Still many questions remain, and for therapeutical agents to be developed it is important to understand how these molecules regulate cellular behavior. PGs are expressed in all mammalian cells at dif-ferent locations and hence have the ability to control many biological pro-cesses. This ability is usually ascribed the negative charge of the GAG chains, but the core proteins play a role as well, and it seems likely that the two parts influence each other. Some of these issues were raised in Paper I, where the aim was to shed some light on the complexity that underlies the study of PGs. Further we wanted to highlight some of our thoughts on how glycosylation affects the behavior of syndecan-1 and that there may also be defined roles for the CS chains.

Considering the tumor microenvironment, an area where a large number of molecules are dysregulated with loss of control, it is apparent that trans-membrane cell surface molecules such as syndecans and integrins are key players and may cooperate [196]. In Paper II we could show that the HS chains that are attached to syndecan-1, retained at the cell surface, are of importance in processes that are known to be integrin dependent, like adhe-sion, migration and cell-mediated matrix contraction.

Clearly, in some situations there is a requirement for a certain GAG chain with a certain sulfate distribution [34, 62]. In other situations though, the question of functional overlap between HS and CS has been raised [41, 204]. In Paper III we showed that also CS chains could support FGF-induced signaling, when attached to uc syndecan-1. Considering the similarities that these two types of GAGs have as linear, highly sulfated polysaccharides, and attached to core proteins, the possibility of one replacing the other seems logical. On the other hand, there are also differences in their hexosamine building blocks and sulfate distribution, expression patterns, locations, and which core proteins they are attached to. Thus, these structural differences may have developed to allow different functions or selective regulation. Formation of pharyngeal cartilage in zebrafish demonstrates distinct roles for HS and CS, Paper IV. However, CS could also compensate for the lack of HS in extl3 mutants, again showing that there are some overlapping func-tions. We could further demonstrate in vivo that the availability of EXTL3 and CSGalNAcT1/2 at the crossroad of the two biosynthetic pathways (Fig-ure 4B) determines the ratio of HS and CS, and that HS biosynthesis has a

39

higher priority than CS biosynthesis. Upon overexpression of syndecan-1 in the cell models used in Paper II and Paper III, we have seen similar results, which further strengthens the idea of a tighter regulation of HS biosynthesis.

Coming back to the complicated tumor microenvironment, many growth factors are upregulated (e.g. FGF and VEGF), leading to enhanced and/or deregulated signaling. In this context, we looked at how the GAG chains, attached to the membrane or free in the extracellular milieu, could influence the signaling. As seen in Paper III, the ectodomain of syndecan-1 was effi-cient in scavenging FGFs from the receptors, independent on the type of glycosylation. Both GAGs and core protein are thus important in regulating signaling processes. Uncontrolled signaling can lead to upregulation of the syndecan sheddases, and also to upregulation of the HS modifying enzymes heparanase and Sulfs. In an already complex situation these molecules will affect the signaling events even more by editing the sulfation pattern and the length of HS chains.

In this study, we have provided some further insight into the regulation of HS biosynthesis both by overexpressing one of many core proteins and look-ing at the consequences, and by investigating the biosynthesis interplay in the zebrafish project. It is evident, that in certain situations there is a strict requirement for HS chains with structural features that CS cannot account for. Because of this, specific structures of HS have to be synthesized to con-trol important biological processes. However, in other situations CS is capa-ble of replacing HS. This is yet another complicated issue in the field of GAGs, and will probably be important when more GAG-related molecules enter the therapeutic market. Clearly, more studies are needed before we can fully understand the complex mechanisms behind the regulation and function of GAGs.

40

Populärvetenskaplig sammanfattning

På cellytan förekommer ett ständigt samspel mellan molekyler på cellytan och i deras omgivning som påverkar cellernas öde. Beroende på vilka signa-ler cellerna får, svarar de genom att exempelvis dela eller förflytta sig. Några av molekylerna på cellytan spänner över cellväggen och kan därmed binda samman cellernas inre med miljön utanför. Två exempel på sådana moleky-ler är proteoglykanerna i syndecan-familjen (Figur 2) och integrinerna (Fi-gur 6A). Först studerade vi hur kolhydratkedjorna heparansulfat (HS) och chondroitinsulfat (CS), som sitter fast på syndecan kan påverka samspelet mellan syndecan och andra molekyler (på eller utanför cellytan) såsom in-tegriner och fibronektin. Vi undersökte också hur syndecaner och integriner samarbetar på cellytan för att cellerna ska fästa till sitt underlag och förflytta sig. Vi använde oss av celler som kan tillverka både HS och CS eller bara CS och lät dem uttrycka oklyvbart syndecan. Våra försök visade att om syn-decan inte kan klyvas, i celler som producerar både HS och CS påverkades samspelet mellan syndecan och integrin så att cellerna satt hårdare fast och förflyttade sig sämre. Vår slutsats är att HS-kedjorna på syndecan är viktiga för hur integrinerna bildar grupper på cellytan, samt hur de återanvänds i cellerna. I de här situationerna kunde HS inte ersättas av CS.

Vi undersökte också hur HS och CS på syndecan påverkar hur cellerna svarar på stimulering med FGF2. FGF2 är en tillväxtfaktor som använder HS för att bilda ett stabilt komplex med sin receptor på cellytan, och därmed framkalla ett svar från cellerna (Figur 7). Vi kunde se att när HS-kedjorna satt fast på syndecan på cellytan bibehölls signalen. Det mest spännande var dock att CS kunde ersätta HS till viss del, vilket tyder på funktionellt över-lapp mellan de två kolhydraterna.

Till sist har vi också tittat på betydelsen av HS och CS i utvecklingen av brosk i zebrafisk. Även här kunde vi se att det fanns ett funktionellt över-lapp, eftersom CS kunde kompensera till viss del när HS saknades, men att de två typerna av kolhydratkedjor har olika roller i broskets uppbyggnad. Genom att använda en muterad form av syndecan-4 i cellförsök kunde vi dessutom lägga fram bevis för att i biosyntesen av HS och CS finns det en rangordning som gör att tillverkningen av HS-kedjor prioriteras.

Sammantaget visar resultaten i avhandlingen på vikten av HS-kedjor för att celler ska förstå sin omgivning på ett korrekt sätt, och därmed förhindra att sjukdomstillstånd uppkommer. De visar också att i vissa lägen kan CS ersätta HS, men att biosyntesen av HS har högst prioritet.

41

Acknowledgements

First, I would like to express my sincere gratitude to my principle supervisor, Dorothe Spillmann. Not only have you encouraged me to develop and grow as a researcher and a person, but you have also shared your knowledge and enthusiasm, and I am extremely grateful for your help over the years. All the late evenings at BMC when I have been writing and you, very patiently, have waited, just to give constructive comments the next morning (or even within hours!) are much appreciated. I would also like to thank my co-supervisor Ulf Lindahl. Over the years, you have pushed me in the right direction, probably without knowing it. I am truly impressed by your knowledge and the way you still keep up to date. Special thanks also to my examiner Johan Kreuger. I appreciate the way you have encouraged me on this journey, both scientifically and personally, and I have truly enjoyed our late evening conversations! I would also like to thank the collaborators and co-authors that have contrib-uted to the projects, especially Johan Kreuger and Zsolt Kasza for valuable input, positive feedback, and trust during the zebrafish and miRNA projects. Thanks to Johan Ledin, Katarina Holmborn Garpenstrand and Judith Habicher for nice collaboration on the zebrafish project, and Vahid Rey-hani for valuable input on the syndecan project. Special thanks also to the members of the Spillmann group, Nadja Jastrebova, Rashmi Ramachan-dra and Ramesh Babu Namburi, for keeping a great spirit during this time, at meetings and book club. And of course to the students that have been part of my projects; Peter Johansson, Natalia Strid and Kai Böker. A big thanks to Andrew Hamilton, Elina Hjertström and Vahid Reyhani for helpful and highly appreciated contribution to the writing of this thesis. My co-workers in D9:4, under supervision of Lena Kjellén, Jin-ping Li, Maria Ringvall, Pär Gerwins, Johan Kreuger, and Cecilia Annerén, for fruitful discussions, Friday seminars and many good cakes at 2:30! The IMBIM administration, Barbro, Erika, Eva, Kerstin, Malin, Olav, Marianne, Rehné, Susanne L, and Susanne T for providing great service and help with everything from big to small, as well as encouraging smiles!

42

Thanks/Tack

During my years in Uppsala many people have colored my life and shared not only knowledge, thoughts and reagents, but also fun times and laughter. All my friends are superheroes!!! From the bottom of my heart, I would like to express my sincere gratitude to all of you for helping me on the way:

Without you, Vahid my dear friend, this thesis would probably not have looked the way it does. Not only have you taught me everything there is to know about Fn/CGC, but you have also kept me alive by providing energy in various forms, both nutritional and spiritual, from near and far J! Thanks for invaluable support and encouragement!!! Elina, thanks for giving me a second chance! In you I have found one of my best friends and it feels like you have always been there. No words can de-scribe how much I appreciate our friendship and I expect it to last forever. Thanks also for HS knowledge and harsh truths about both science and life! Thank you Peder, for leading the way towards this day (on the verge of elec-trocution in our cave!). In some strange way it made me calm knowing that you always had two weeks less J. A special thanks to Sophia Schedin-Weiss and Wei Sun who showed me what science is all about, it could not have started in a better way! And Eva Gottfridsson of course, who actually made it possible!

43

Thank you to all past and present members of D9:4 and IMBIM, especial-ly: Audrey, Australia. Forever in my heart and memory. Anh-tri, for laugh-ter and good times in the gym. And hot pot. And guys night. And Manly beach J. Inger, the best office mate anyone can have! Thanks for conversa-tions, laughter, and dreams about Irish setters. I hope you know how much I appreciate you! Zsolt for tsin (!!), jokes and triple jump in clogs. See you in Rio de Janeiro 2016! Jimmy for raising the stress level, and for letting me be a small part of your beautiful family. In many different situations, both pleasant and awkward J.

I also need to thank five weeks of ISR, for introducing me to Josefin. You are incredibly warmhearted and make me feel so welcome in your home that I forget to leave J. You were the only good thing (almost!) that came out of that course! And Nathalie, for patiently keeping in touch and making sure that we meet! You and our evenings are invaluable to me! Barndomsvänner i mitt hjärta och era fina familjer, alla underbara vänner från tiden i Karlstad: Tack för allt stöd och uppmuntran. Tack för att ni finns! Världens bästa syster yster och mina små solstrålar Ebba och Nora. Famil-jen Melberg är den bästa fan club man kan önska sig! Och så du, älskade mamma! Utan dig hade det inte gått! Jag är oändligt tacksam för allt stöd och hjälp (på alla nivåer) genom både lätt och svårt. Men framförallt för att du alltid tror på mig och tycker att jag är bäst i världen.

44

References

1. Esko, J.D. and U. Lindahl, Molecular diversity of heparan sulfate. J Clin Invest, 2001. 108(2): p. 169-73.

2. Kjellen, L. and U. Lindahl, Proteoglycans: structures and interactions. Annu Rev Biochem, 1991. 60: p. 443-75.

3. Gotte, M., Syndecans in inflammation. Faseb J, 2003. 17(6): p. 575-91. 4. Bernfield, M., et al., Functions of cell surface heparan sulfate proteoglycans.

Annu Rev Biochem, 1999. 68: p. 729-77. 5. Lindahl, U. and J.P. Li, Interactions between heparan sulfate and proteins-

design and functional implications. Int Rev Cell Mol Biol, 2009. 276: p. 105-59.

6. Kolset, S.O. and H. Tveit, Serglycin--structure and biology. Cell Mol Life Sci, 2008. 65(7-8): p. 1073-85.

7. Abrink, M., M. Grujic, and G. Pejler, Serglycin is essential for maturation of mast cell secretory granule. J Biol Chem, 2004. 279(39): p. 40897-905.

8. Yamada, S. and K. Sugahara, Potential therapeutic application of chondroitin sulfate/dermatan sulfate. Curr Drug Discov Technol, 2008. 5(4): p. 289-301.

9. Doege, K., et al., Complete primary structure of the rat cartilage proteoglycan core protein deduced from cDNA clones. J Biol Chem, 1987. 262(36): p. 17757-67.

10. Heinegard, D., Proteoglycans and more--from molecules to biology. Int J Exp Pathol, 2009. 90(6): p. 575-86.

11. Iozzo, R.V., The biology of the small leucine-rich proteoglycans. Functional network of interactive proteins. J Biol Chem, 1999. 274(27): p. 18843-6.

12. Couchman, J.R., Transmembrane signaling proteoglycans. Annu Rev Cell Dev Biol, 2010. 26: p. 89-114.

13. Kirn-Safran, C., M.C. Farach-Carson, and D.D. Carson, Multifunctionality of extracellular and cell surface heparan sulfate proteoglycans. Cell Mol Life Sci, 2009. 66(21): p. 3421-34.

14. Alexopoulou, A.N., H.A. Multhaupt, and J.R. Couchman, Syndecans in wound healing, inflammation and vascular biology. Int J Biochem Cell Biol, 2007. 39(3): p. 505-28.

15. Tkachenko, E., J.M. Rhodes, and M. Simons, Syndecans: new kids on the signaling block. Circ Res, 2005. 96(5): p. 488-500.

16. Echtermeyer, F., et al., Delayed wound repair and impaired angiogenesis in mice lacking syndecan-4. J Clin Invest, 2001. 107(2): p. R9-R14.

17. Alexander, C.M., et al., Syndecan-1 is required for Wnt-1-induced mammary tumorigenesis in mice. Nat Genet, 2000. 25(3): p. 329-32.

18. Forsberg, E. and L. Kjellen, Heparan sulfate: lessons from knockout mice. J Clin Invest, 2001. 108(2): p. 175-80.

19. Rapraeger, A., et al., The cell surface proteoglycan from mouse mammary epithelial cells bears chondroitin sulfate and heparan sulfate glycosaminoglycans. J Biol Chem, 1985. 260(20): p. 11046-52.

45

20. Shworak, N.W., et al., Characterization of ryudocan glycosaminoglycan acceptor sites. J Biol Chem, 1994. 269(33): p. 21204-14.

21. Gould, S.E., W.B. Upholt, and R.A. Kosher, Syndecan 3: a member of the syndecan family of membrane-intercalated proteoglycans that is expressed in high amounts at the onset of chicken limb cartilage differentiation. Proc Natl Acad Sci U S A, 1992. 89(8): p. 3271-5.

22. Dews, I.C. and K.R. Mackenzie, Transmembrane domains of the syndecan family of growth factor coreceptors display a hierarchy of homotypic and heterotypic interactions. Proc Natl Acad Sci U S A, 2007. 104(52): p. 20782-7.

23. Carey, D.J., Syndecans: multifunctional cell-surface co-receptors. Biochem J, 1997. 327 ( Pt 1): p. 1-16.

24. Bass, M.D., M.R. Morgan, and M.J. Humphries, Syndecans shed their reputation as inert molecules. Sci Signal, 2009. 2(64): p. pe18.

25. Endo, K., et al., Cleavage of syndecan-1 by membrane type matrix metalloproteinase-1 stimulates cell migration. J Biol Chem, 2003. 278(42): p. 40764-70.

26. Selleck, S.B., Shedding light on the distinct functions of proteoglycans. Sci STKE, 2006. 2006(329): p. pe17.

27. Manon-Jensen, T., H.A. Multhaupt, and J.R. Couchman, Mapping of matrix metalloproteinase cleavage sites on syndecan-1 and syndecan-4 ectodomains. FEBS J, 2013.

28. Hayashida, K., P.D. Stahl, and P.W. Park, Syndecan-1 ectodomain shedding is regulated by the small GTPase Rab5. J Biol Chem, 2008. 283(51): p. 35435-44.

29. Ramani, V.C., et al., Heparan sulfate chains of syndecan-1 regulate ectodomain shedding. J Biol Chem, 2012. 287(13): p. 9952-61.

30. Purushothaman, A., et al., Heparanase stimulation of protease expression implicates it as a master regulator of the aggressive tumor phenotype in myeloma. J Biol Chem, 2008. 283(47): p. 32628-36.

31. Ding, K., et al., Growth factor-induced shedding of syndecan-1 confers glypican-1 dependence on mitogenic responses of cancer cells. J Cell Biol, 2005. 171(4): p. 729-38.

32. Wang, Z., et al., Constitutive and accelerated shedding of murine syndecan-1 is mediated by cleavage of its core protein at a specific juxtamembrane site. Biochemistry, 2005. 44(37): p. 12355-61.

33. Lindahl, U., Heparan sulfate-protein interactions - A concept for drug design? Thromb Haemost, 2007. 98(1): p. 109-15.

34. Esko, J.D. and S.B. Selleck, Order out of chaos: assembly of ligand binding sites in heparan sulfate. Annu Rev Biochem, 2002. 71: p. 435-71.

35. Laurent, T.C. and J.R. Fraser, Hyaluronan. FASEB J, 1992. 6(7): p. 2397-404. 36. Zhang, L. and J.D. Esko, Amino acid determinants that drive heparan sulfate

assembly in a proteoglycan. J Biol Chem, 1994. 269(30): p. 19295-9. 37. Zhang, L., G. David, and J.D. Esko, Repetitive Ser-Gly sequences enhance

heparan sulfate assembly in proteoglycans. J Biol Chem, 1995. 270(45): p. 27127-35.

38. Ledin, J., et al., Heparan sulfate structure in mice with genetically modified heparan sulfate production. J Biol Chem, 2004. 279(41): p. 42732-41.

39. Presto, J., et al., Heparan sulfate biosynthesis enzymes EXT1 and EXT2 affect NDST1 expression and heparan sulfate sulfation. Proc Natl Acad Sci U S A, 2008. 105(12): p. 4751-6.

40. Uhlin-Hansen, L. and M. Yanagishita, Differential effect of brefeldin A on the biosynthesis of heparan sulfate and chondroitin/dermatan sulfate

46

proteoglycans in rat ovarian granulosa cells in culture. J Biol Chem, 1993. 268(23): p. 17370-6.

41. Kreuger, J., et al., Interactions between heparan sulfate and proteins: the concept of specificity. J Cell Biol, 2006. 174(3): p. 323-7.

42. Busse, M., et al., Contribution of EXT1, EXT2, and EXTL3 to heparan sulfate chain elongation. J Biol Chem, 2007. 282(45): p. 32802-10.

43. Lin, X., et al., Disruption of gastrulation and heparan sulfate biosynthesis in EXT1-deficient mice. Dev Biol, 2000. 224(2): p. 299-311.

44. Stickens, D., et al., Mice deficient in Ext2 lack heparan sulfate and develop exostoses. Development, 2005. 132(22): p. 5055-68.

45. Kjellen, L., Glucosaminyl N-deacetylase/N-sulphotransferases in heparan sulphate biosynthesis and biology. Biochem Soc Trans, 2003. 31(2): p. 340-2.

46. Klaassen, C.D. and J.W. Boles, Sulfation and sulfotransferases 5: the importance of 3'-phosphoadenosine 5'-phosphosulfate (PAPS) in the regulation of sulfation. FASEB J, 1997. 11(6): p. 404-18.

47. Carlsson, P., et al., Heparin/heparan sulfate biosynthesis: processive formation of N-sulfated domains. J Biol Chem, 2008. 283(29): p. 20008-14.

48. Grobe, K., et al., Heparan sulfate and development: differential roles of the N-acetylglucosamine N-deacetylase/N-sulfotransferase isozymes. Biochim Biophys Acta, 2002. 1573(3): p. 209-15.

49. Ringvall, M., et al., Defective heparan sulfate biosynthesis and neonatal lethality in mice lacking N-deacetylase/N-sulfotransferase-1. J Biol Chem, 2000. 275(34): p. 25926-30.

50. Forsberg, E., et al., Abnormal mast cells in mice deficient in a heparin-synthesizing enzyme. Nature, 1999. 400(6746): p. 773-6.

51. Dagalv, A., et al., Lowered expression of heparan sulfate/heparin biosynthesis enzyme N-deacetylase/n-sulfotransferase 1 results in increased sulfation of mast cell heparin. J Biol Chem, 2011. 286(52): p. 44433-40.

52. Casu, B., et al., Conformational flexibility: a new concept for explaining binding and biological properties of iduronic acid-containing glycosaminoglycans. Trends Biochem Sci, 1988. 13(6): p. 221-5.

53. Rong, J., et al., Substrate specificity of the heparan sulfate hexuronic acid 2-O-sulfotransferase. Biochemistry, 2001. 40(18): p. 5548-55.

54. Li, J.P., et al., Targeted disruption of a murine glucuronyl C5-epimerase gene results in heparan sulfate lacking L-iduronic acid and in neonatal lethality. J Biol Chem, 2003. 278(31): p. 28363-6.

55. Bullock, S.L., et al., Renal agenesis in mice homozygous for a gene trap mutation in the gene encoding heparan sulfate 2-sulfotransferase. Genes Dev, 1998. 12(12): p. 1894-906.

56. Hook, M., et al., Biosynthesis of heparin. Studies on the microsomal sulfation process. J Biol Chem, 1975. 250(15): p. 6065-71.

57. Yanagishita, M. and V.C. Hascall, Cell surface heparan sulfate proteoglycans. J Biol Chem, 1992. 267(14): p. 9451-4.

58. Senay, C., et al., The EXT1/EXT2 tumor suppressors: catalytic activities and role in heparan sulfate biosynthesis. EMBO Rep, 2000. 1(3): p. 282-6.

59. Pinhal, M.A., et al., Enzyme interactions in heparan sulfate biosynthesis: uronosyl 5-epimerase and 2-O-sulfotransferase interact in vivo. Proc Natl Acad Sci U S A, 2001. 98(23): p. 12984-9.

60. McCormick, C., et al., The putative tumor suppressors EXT1 and EXT2 form a stable complex that accumulates in the Golgi apparatus and catalyzes the synthesis of heparan sulfate. Proc Natl Acad Sci U S A, 2000. 97(2): p. 668-73.

47

61. Kobayashi, S., et al., Association of EXT1 and EXT2, hereditary multiple exostoses gene products, in Golgi apparatus. Biochem Biophys Res Commun, 2000. 268(3): p. 860-7.

62. Lindahl, U., et al., Evidence for a 3-O-sulfated D-glucosamine residue in the antithrombin-binding sequence of heparin. Proc Natl Acad Sci U S A, 1980. 77(11): p. 6551-5.

63. Wang, S., et al., QSulf1, a heparan sulfate 6-O-endosulfatase, inhibits fibroblast growth factor signaling in mesoderm induction and angiogenesis. Proc Natl Acad Sci U S A, 2004. 101(14): p. 4833-8.

64. Dhoot, G.K., et al., Regulation of Wnt signaling and embryo patterning by an extracellular sulfatase. Science, 2001. 293(5535): p. 1663-6.

65. Bishop, J.R., M. Schuksz, and J.D. Esko, Heparan sulphate proteoglycans fine-tune mammalian physiology. Nature, 2007. 446(7139): p. 1030-7.

66. Fux, L., et al., Heparanase: busy at the cell surface. Trends Biochem Sci, 2009. 34(10): p. 511-9.

67. Kwok, J.C., P. Warren, and J.W. Fawcett, Chondroitin sulfate: a key molecule in the brain matrix. Int J Biochem Cell Biol, 2012. 44(4): p. 582-6.

68. Knudson, C.B. and W. Knudson, Cartilage proteoglycans. Semin Cell Dev Biol, 2001. 12(2): p. 69-78.

69. Lamari, F.N. and N.K. Karamanos, Structure of chondroitin sulfate. Adv Pharmacol, 2006. 53: p. 33-48.

70. Sato, T., et al., Differential roles of two N-acetylgalactosaminyltransferases, CSGalNAcT-1, and a novel enzyme, CSGalNAcT-2. Initiation and elongation in synthesis of chondroitin sulfate. J Biol Chem, 2003. 278(5): p. 3063-71.

71. Ogawa, H., et al., Chondroitin sulfate synthase-2/chondroitin polymerizing factor has two variants with distinct function. J Biol Chem, 2010. 285(44): p. 34155-67.

72. Izumikawa, T., et al., Involvement of chondroitin sulfate synthase-3 (chondroitin synthase-2) in chondroitin polymerization through its interaction with chondroitin synthase-1 or chondroitin-polymerizing factor. Biochem J, 2007. 403(3): p. 545-52.

73. Ogawa, H., et al., Chondroitin sulfate synthase-2 is necessary for chain extension of chondroitin sulfate but not critical for skeletal development. PLoS One, 2012. 7(8): p. e43806.

74. Mizuguchi, S., et al., Chondroitin proteoglycans are involved in cell division of Caenorhabditis elegans. Nature, 2003. 423(6938): p. 443-8.

75. Izumikawa, T., et al., Nematode chondroitin polymerizing factor showing cell-/organ-specific expression is indispensable for chondroitin synthesis and embryonic cell division. J Biol Chem, 2004. 279(51): p. 53755-61.

76. Malmstrom, A. and L.A. Fransson, Biosynthesis of dermatan sulfate. I. Formation of L-iduronic acid residues. J Biol Chem, 1975. 250(9): p. 3419-25.

77. Malmstrom, A., Biosynthesis of dermatan sulfate. II. Substrate specificity of the C-5 uronosyl epimerase. J Biol Chem, 1984. 259(1): p. 161-5.

78. Sugahara, K., et al., Recent advances in the structural biology of chondroitin sulfate and dermatan sulfate. Curr Opin Struct Biol, 2003. 13(5): p. 612-20.

79. Maccarana, M., et al., Dermatan sulfate epimerase 1-deficient mice have reduced content and changed distribution of iduronic acids in dermatan sulfate and an altered collagen structure in skin. Mol Cell Biol, 2009. 29(20): p. 5517-28.

80. Bartolini, B., et al., Mouse development is not obviously affected by the absence of dermatan sulfate epimerase 2 in spite of a modified brain dermatan sulfate composition. Glycobiology, 2012. 22(7): p. 1007-16.

48

81. Malmstrom, A., et al., Iduronic acid in chondroitin/dermatan sulfate: biosynthesis and biological function. J Histochem Cytochem, 2012. 60(12): p. 916-25.

82. Bulow, H.E. and O. Hobert, The molecular diversity of glycosaminoglycans shapes animal development. Annu Rev Cell Dev Biol, 2006. 22: p. 375-407.

83. Kluppel, M., et al., Maintenance of chondroitin sulfation balance by chondroitin-4-sulfotransferase 1 is required for chondrocyte development and growth factor signaling during cartilage morphogenesis. Development, 2005. 132(17): p. 3989-4003.

84. Thiele, H., et al., Loss of chondroitin 6-O-sulfotransferase-1 function results in severe human chondrodysplasia with progressive spinal involvement. Proc Natl Acad Sci U S A, 2004. 101(27): p. 10155-60.

85. Izumikawa, T., et al., Impairment of embryonic cell division and glycosaminoglycan biosynthesis in glucuronyltransferase-I-deficient mice. J Biol Chem, 2010. 285(16): p. 12190-6.

86. Kreuger, J. and L. Kjellen, Heparan sulfate biosynthesis: regulation and variability. J Histochem Cytochem, 2012. 60(12): p. 898-907.

87. Kokenyesi, R. and M. Bernfield, Core protein structure and sequence determine the site and presence of heparan sulfate and chondroitin sulfate on syndecan-1. J Biol Chem, 1994. 269(16): p. 12304-9.

88. Esko, J.D., K. Kimata, and U. Lindahl, Proteoglycans and Sulfated Glycosaminoglycans, in Essentials of Glycobiology, A. Varki, et al., Editors. 2009: Cold Spring Harbor (NY).

89. Ueno, M., et al., Structural characterization of heparan sulfate and chondroitin sulfate of syndecan-1 purified from normal murine mammary gland epithelial cells. Common phosphorylation of xylose and differential sulfation of galactose in the protein linkage region tetrasaccharide sequence. J Biol Chem, 2001. 276(31): p. 29134-40.

90. Tone, Y., et al., 2-o-phosphorylation of xylose and 6-o-sulfation of galactose in the protein linkage region of glycosaminoglycans influence the glucuronyltransferase-I activity involved in the linkage region synthesis. J Biol Chem, 2008. 283(24): p. 16801-7.

91. Nadanaka, S., et al., EXTL2, a member of EXT family of tumor suppressors, controls glycosaminoglycan biosynthesis in a xylose kinase-dependent manner. J Biol Chem, 2013.

92. Prydz, K. and K.T. Dalen, Synthesis and sorting of proteoglycans. J Cell Sci, 2000. 113 Pt 2: p. 193-205.

93. Hynes, R.O., Integrins: versatility, modulation, and signaling in cell adhesion. Cell, 1992. 69(1): p. 11-25.

94. Coucouvanis, E. and G.R. Martin, Signals for death and survival: a two-step mechanism for cavitation in the vertebrate embryo. Cell, 1995. 83(2): p. 279-87.

95. Frisch, S.M. and E. Ruoslahti, Integrins and anoikis. Curr Opin Cell Biol, 1997. 9(5): p. 701-6.

96. Stupack, D.G. and D.A. Cheresh, Get a ligand, get a life: integrins, signaling and cell survival. J Cell Sci, 2002. 115(Pt 19): p. 3729-38.

97. Stephens, L.E., et al., Deletion of beta 1 integrins in mice results in inner cell mass failure and peri-implantation lethality. Genes Dev, 1995. 9(15): p. 1883-95.

98. Barczyk, M., S. Carracedo, and D. Gullberg, Integrins. Cell Tissue Res, 2010. 339(1): p. 269-80.

49

99. Campbell, I.D. and M.J. Humphries, Integrin structure, activation, and interactions. Cold Spring Harb Perspect Biol, 2011. 3(3).

100. Okina, E., et al., Syndecan proteoglycan contributions to cytoskeletal organization and contractility. Scand J Med Sci Sports, 2009. 19(4): p. 479-89.

101. Huttenlocher, A. and A.R. Horwitz, Integrins in cell migration. Cold Spring Harb Perspect Biol, 2011. 3(9): p. a005074.

102. Hu, P. and B.H. Luo, Integrin bi-directional signaling across the plasma membrane. J Cell Physiol, 2013. 228(2): p. 306-12.

103. Travis, M.A., J.D. Humphries, and M.J. Humphries, An unraveling tale of how integrins are activated from within. Trends Pharmacol Sci, 2003. 24(4): p. 192-7.

104. Ginsberg, M.H., A. Partridge, and S.J. Shattil, Integrin regulation. Curr Opin Cell Biol, 2005. 17(5): p. 509-16.

105. Humphries, J.D., A. Byron, and M.J. Humphries, Integrin ligands at a glance. J Cell Sci, 2006. 119(Pt 19): p. 3901-3.

106. Pankov, R. and K.M. Yamada, Fibronectin at a glance. J Cell Sci, 2002. 115(Pt 20): p. 3861-3.

107. Hynes, R.O., The extracellular matrix: not just pretty fibrils. Science, 2009. 326(5957): p. 1216-9.

108. Ridley, A.J., et al., Cell migration: integrating signals from front to back. Science, 2003. 302(5651): p. 1704-9.

109. Friedl, P. and K. Wolf, Plasticity of cell migration: a multiscale tuning model. J Cell Biol, 2010. 188(1): p. 11-9.

110. Kiosses, W.B., et al., Rac recruits high-affinity integrin alphavbeta3 to lamellipodia in endothelial cell migration. Nat Cell Biol, 2001. 3(3): p. 316-20.

111. Franco, S.J., et al., Calpain-mediated proteolysis of talin regulates adhesion dynamics. Nat Cell Biol, 2004. 6(10): p. 977-83.

112. Webb, D.J., et al., FAK-Src signalling through paxillin, ERK and MLCK regulates adhesion disassembly. Nat Cell Biol, 2004. 6(2): p. 154-61.

113. Truong, H. and E.H. Danen, Integrin switching modulates adhesion dynamics and cell migration. Cell Adh Migr, 2009. 3(2): p. 179-81.

114. Turner, N. and R. Grose, Fibroblast growth factor signalling: from development to cancer. Nat Rev Cancer, 2010. 10(2): p. 116-29.

115. Beenken, A. and M. Mohammadi, The FGF family: biology, pathophysiology and therapy. Nat Rev Drug Discov, 2009. 8(3): p. 235-53.

116. Itoh, N., The Fgf families in humans, mice, and zebrafish: their evolutional processes and roles in development, metabolism, and disease. Biol Pharm Bull, 2007. 30(10): p. 1819-25.

117. Ornitz, D.M. and N. Itoh, Fibroblast growth factors. Genome Biol, 2001. 2(3): p. REVIEWS3005.

118. Eswarakumar, V.P., I. Lax, and J. Schlessinger, Cellular signaling by fibroblast growth factor receptors. Cytokine Growth Factor Rev, 2005. 16(2): p. 139-49.

119. Powers, C.J., S.W. McLeskey, and A. Wellstein, Fibroblast growth factors, their receptors and signaling. Endocr Relat Cancer, 2000. 7(3): p. 165-97.

120. Mignatti, P., T. Morimoto, and D.B. Rifkin, Basic fibroblast growth factor, a protein devoid of secretory signal sequence, is released by cells via a pathway independent of the endoplasmic reticulum-Golgi complex. J Cell Physiol, 1992. 151(1): p. 81-93.

121. Grose, R. and C. Dickson, Fibroblast growth factor signaling in tumorigenesis. Cytokine Growth Factor Rev, 2005. 16(2): p. 179-86.

50

122. Klint, P. and L. Claesson-Welsh, Signal transduction by fibroblast growth factor receptors. Front Biosci, 1999. 4: p. D165-77.

123. Zhang, X., et al., Receptor specificity of the fibroblast growth factor family. The complete mammalian FGF family. J Biol Chem, 2006. 281(23): p. 15694-700.

124. Lew, E.D., et al., The precise sequence of FGF receptor autophosphorylation is kinetically driven and is disrupted by oncogenic mutations. Sci Signal, 2009. 2(58): p. ra6.

125. Brooks, A.N., E. Kilgour, and P.D. Smith, Molecular pathways: fibroblast growth factor signaling: a new therapeutic opportunity in cancer. Clin Cancer Res, 2012. 18(7): p. 1855-62.

126. Daniele, G., et al., FGF receptor inhibitors: role in cancer therapy. Curr Oncol Rep, 2012. 14(2): p. 111-9.

127. Sleeman, M., et al., Identification of a new fibroblast growth factor receptor, FGFR5. Gene, 2001. 271(2): p. 171-82.

128. Hacker, U., K. Nybakken, and N. Perrimon, Heparan sulphate proteoglycans: the sweet side of development. Nat Rev Mol Cell Biol, 2005. 6(7): p. 530-41.

129. Harmer, N.J., Insights into the role of heparan sulphate in fibroblast growth factor signalling. Biochem Soc Trans, 2006. 34(Pt 3): p. 442-5.

130. Mohammadi, M., S.K. Olsen, and O.A. Ibrahimi, Structural basis for fibroblast growth factor receptor activation. Cytokine Growth Factor Rev, 2005. 16(2): p. 107-37.

131. Yayon, A., et al., Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell, 1991. 64(4): p. 841-8.

132. Schlessinger, J., et al., Crystal structure of a ternary FGF-FGFR-heparin complex reveals a dual role for heparin in FGFR binding and dimerization. Mol Cell, 2000. 6(3): p. 743-50.

133. Pellegrini, L., et al., Crystal structure of fibroblast growth factor receptor ectodomain bound to ligand and heparin. Nature, 2000. 407(6807): p. 1029-34.

134. Lundin, L., et al., Differential tyrosine phosphorylation of fibroblast growth factor (FGF) receptor-1 and receptor proximal signal transduction in response to FGF-2 and heparin. Exp Cell Res, 2003. 287(1): p. 190-8.

135. Jastrebova, N., et al., Heparan sulfate-related oligosaccharides in ternary complex formation with fibroblast growth factors 1 and 2 and their receptors. J Biol Chem, 2006. 281(37): p. 26884-92.

136. Jastrebova, N., et al., Heparan sulfate domain organization and sulfation modulate FGF-induced cell signaling. J Biol Chem, 2010. 285(35): p. 26842-51.

137. Xu, R., et al., Diversification of the structural determinants of fibroblast growth factor-heparin interactions: implications for binding specificity. J Biol Chem, 2012. 287(47): p. 40061-73.

138. Dailey, L., et al., Mechanisms underlying differential responses to FGF signaling. Cytokine Growth Factor Rev, 2005. 16(2): p. 233-47.

139. Murakami, M., A. Elfenbein, and M. Simons, Non-canonical fibroblast growth factor signalling in angiogenesis. Cardiovasc Res, 2008. 78(2): p. 223-31.

140. Al-Ayoubi, A., et al., ERK activation and nuclear signaling induced by the loss of cell/matrix adhesion stimulates anchorage-independent growth of ovarian cancer cells. J Cell Biochem, 2008. 105(3): p. 875-84.

141. Fayard, E., et al., Protein kinase B/Akt at a glance. J Cell Sci, 2005. 118(Pt 24): p. 5675-8.

51

142. Falasca, M., et al., Activation of phospholipase C gamma by PI 3-kinase-induced PH domain-mediated membrane targeting. EMBO J, 1998. 17(2): p. 414-22.

143. Ai, X., et al., QSulf1 remodels the 6-O sulfation states of cell surface heparan sulfate proteoglycans to promote Wnt signaling. J Cell Biol, 2003. 162(2): p. 341-51.

144. Lamanna, W.C., et al., Heparan sulfate 6-O-endosulfatases: discrete in vivo activities and functional co-operativity. Biochem J, 2006. 400(1): p. 63-73.

145. Li, J.P., Heparin, heparan sulfate and heparanase in cancer: remedy for metastasis? Anticancer Agents Med Chem, 2008. 8(1): p. 64-76.

146. Gotte, M., et al., Changes in heparan sulfate are associated with delayed wound repair, altered cell migration, adhesion and contractility in the galactosyltransferase I (beta4GalT-7) deficient form of Ehlers-Danlos syndrome. Hum Mol Genet, 2008. 17(7): p. 996-1009.

147. Holley, R.J., et al., Mucopolysaccharidosis type I, unique structure of accumulated heparan sulfate and increased N-sulfotransferase activity in mice lacking alpha-l-iduronidase. J Biol Chem, 2011. 286(43): p. 37515-24.

148. Blackhall, F.H., et al., Heparan sulfate proteoglycans and cancer. Br J Cancer, 2001. 85(8): p. 1094-8.

149. Iozzo, R.V. and R.D. Sanderson, Proteoglycans in cancer biology, tumour microenvironment and angiogenesis. J Cell Mol Med, 2011. 15(5): p. 1013-31.

150. Iozzo, R.V., Basement membrane proteoglycans: from cellar to ceiling. Nat Rev Mol Cell Biol, 2005. 6(8): p. 646-56.

151. Yoneda, A., et al., Breast and ovarian cancers: a survey and possible roles for the cell surface heparan sulfate proteoglycans. J Histochem Cytochem, 2012. 60(1): p. 9-21.

152. Sanderson, R.D. and Y. Yang, Syndecan-1: a dynamic regulator of the myeloma microenvironment. Clin Exp Metastasis, 2008. 25(2): p. 149-59.

153. Choi, S., D.H. Kang, and E.S. Oh, Targeting syndecans: a promising strategy for the treatment of cancer. Expert Opin Ther Targets, 2013.

154. Conejo, J.R., et al., Syndecan-1 expression is up-regulated in pancreatic but not in other gastrointestinal cancers. Int J Cancer, 2000. 88(1): p. 12-20.

155. Barbareschi, M., et al., High syndecan-1 expression in breast carcinoma is related to an aggressive phenotype and to poorer prognosis. Cancer, 2003. 98(3): p. 474-83.

156. Pulkkinen, J.O., et al., Syndecan-1: a new prognostic marker in laryngeal cancer. Acta Otolaryngol, 1997. 117(2): p. 312-5.

157. Nackaerts, K., et al., Heparan sulfate proteoglycan expression in human lung-cancer cells. Int J Cancer, 1997. 74(3): p. 335-45.

158. Nikolova, V., et al., Differential roles for membrane-bound and soluble syndecan-1 (CD138) in breast cancer progression. Carcinogenesis, 2009. 30(3): p. 397-407.

159. Su, G., et al., Shedding of syndecan-1 by stromal fibroblasts stimulates human breast cancer cell proliferation via FGF2 activation. J Biol Chem, 2007. 282(20): p. 14906-15.

160. Su, G., et al., Membrane type 1 matrix metalloproteinase-mediated stromal syndecan-1 shedding stimulates breast carcinoma cell proliferation. Cancer Res, 2008. 68(22): p. 9558-65.

161. Yang, Y., et al., Soluble syndecan-1 promotes growth of myeloma tumors in vivo. Blood, 2002. 100(2): p. 610-7.

52

162. Zong, F., et al., Effect of syndecan-1 overexpression on mesenchymal tumour cell proliferation with focus on different functional domains. Cell Prolif, 2009. 43(1): p. 29-40.

163. Mundhenke, C., et al., Heparan sulfate proteoglycans as regulators of fibroblast growth factor-2 receptor binding in breast carcinomas. Am J Pathol, 2002. 160(1): p. 185-94.

164. Maeda, T., C.M. Alexander, and A. Friedl, Induction of syndecan-1 expression in stromal fibroblasts promotes proliferation of human breast cancer cells. Cancer Res, 2004. 64(2): p. 612-21.

165. Lendorf, M.E., et al., Syndecan-1 and syndecan-4 are independent indicators in breast carcinoma. J Histochem Cytochem, 2011. 59(6): p. 615-29.

166. Leivonen, M., et al., Prognostic value of syndecan-1 expression in breast cancer. Oncology, 2004. 67(1): p. 11-8.

167. Ramani, V.C., et al., The heparanase/syndecan-1 axis in cancer: mechanisms and therapies. FEBS J, 2013.

168. Vlodavsky, I., et al., Mammalian heparanase: gene cloning, expression and function in tumor progression and metastasis. Nat Med, 1999. 5(7): p. 793-802.

169. Purushothaman, A., S.K. Babitz, and R.D. Sanderson, Heparanase enhances the insulin receptor signaling pathway to activate extracellular signal-regulated kinase in multiple myeloma. J Biol Chem, 2012. 287(49): p. 41288-96.

170. Ritchie, J.P., et al., SST0001, a chemically modified heparin, inhibits myeloma growth and angiogenesis via disruption of the heparanase/syndecan-1 axis. Clin Cancer Res, 2011. 17(6): p. 1382-93.

171. Wegrowski, Y. and F.X. Maquart, Chondroitin sulfate proteoglycans in tumor progression. Adv Pharmacol, 2006. 53: p. 297-321.

172. Suwiwat, S., et al., Expression of extracellular matrix components versican, chondroitin sulfate, tenascin, and hyaluronan, and their association with disease outcome in node-negative breast cancer. Clin Cancer Res, 2004. 10(7): p. 2491-8.

173. Daidouji, K., et al., Neoplastic changes in saccharide sequence of dermatan sulfate chains derived from human colon cancer. Dig Dis Sci, 2002. 47(2): p. 331-7.

174. Chekenya, M., et al., The NG2 chondroitin sulfate proteoglycan: role in malignant progression of human brain tumours. Int J Dev Neurosci, 1999. 17(5-6): p. 421-35.

175. Denholm, E.M., Y.Q. Lin, and P.J. Silver, Anti-tumor activities of chondroitinase AC and chondroitinase B: inhibition of angiogenesis, proliferation and invasion. Eur J Pharmacol, 2001. 416(3): p. 213-21.

176. Smetsers, T.F., et al., Human single-chain antibodies reactive with native chondroitin sulfate detect chondroitin sulfate alterations in melanoma and psoriasis. J Invest Dermatol, 2004. 122(3): p. 707-16.

177. Asimakopoulou, A.P., et al., The biological role of chondroitin sulfate in cancer and chondroitin-based anticancer agents. In Vivo, 2008. 22(3): p. 385-9.

178. Esko, J.D., T.E. Stewart, and W.H. Taylor, Animal cell mutants defective in glycosaminoglycan biosynthesis. Proc Natl Acad Sci U S A, 1985. 82(10): p. 3197-201.

179. Zhang, L., et al., CHO glycosylation mutants: proteoglycans. Methods Enzymol, 2006. 416: p. 205-21.

53

180. Kimmel, C.B., Genetics and early development of zebrafish. Trends Genet, 1989. 5(8): p. 283-8.

181. Westerfield, M., The zebrafish book. A guide for the laboratory use of zebrafish (Danio rerio). 4th ed ed. 2000 Eugene: Univ. of Oregon Press.

182. Kimmel, C.B., et al., Stages of embryonic development of the zebrafish. Dev Dyn, 1995. 203(3): p. 253-310.

183. Haffter, P. and C. Nusslein-Volhard, Large scale genetics in a small vertebrate, the zebrafish. Int J Dev Biol, 1996. 40(1): p. 221-7.

184. Hogan, B.M., et al., Manipulation of gene expression during zebrafish embryonic development using transient approaches. Methods Mol Biol, 2008. 469: p. 273-300.

185. Robu, M.E., et al., p53 activation by knockdown technologies. PLoS Genet, 2007. 3(5): p. e78.

186. Beauvais, D.M., et al., Syndecan-1 regulates alphavbeta3 and alphavbeta5 integrin activation during angiogenesis and is blocked by synstatin, a novel peptide inhibitor. J Exp Med, 2009. 206(3): p. 691-705.

187. Filmus, J., M. Capurro, and J. Rast, Glypicans. Genome Biol, 2008. 9(5): p. 224.

188. Woods, A. and J.R. Couchman, Syndecan 4 heparan sulfate proteoglycan is a selectively enriched and widespread focal adhesion component. Mol Biol Cell, 1994. 5(2): p. 183-92.

189. Sanderson, R.D., et al., Fine structure of heparan sulfate regulates syndecan-1 function and cell behavior. J Biol Chem, 1994. 269(18): p. 13100-6.

190. Sanderson, R.D. and M. Bernfield, Molecular polymorphism of a cell surface proteoglycan: distinct structures on simple and stratified epithelia. Proc Natl Acad Sci U S A, 1988. 85(24): p. 9562-6.

191. Zako, M., et al., Syndecan-1 and -4 synthesized simultaneously by mouse mammary gland epithelial cells bear heparan sulfate chains that are apparently structurally indistinguishable. J Biol Chem, 2003. 278(15): p. 13561-9.

192. Kato, M., et al., Cell surface syndecan-1 on distinct cell types differs in fine structure and ligand binding of its heparan sulfate chains. J Biol Chem, 1994. 269(29): p. 18881-90.

193. Deepa, S.S., et al., Chondroitin sulfate chains on syndecan-1 and syndecan-4 from normal murine mammary gland epithelial cells are structurally and functionally distinct and cooperate with heparan sulfate chains to bind growth factors. A novel function to control binding of midkine, pleiotrophin, and basic fibroblast growth factor. J Biol Chem, 2004. 279(36): p. 37368-76.

194. Lambaerts, K., S.A. Wilcox-Adelman, and P. Zimmermann, The signaling mechanisms of syndecan heparan sulfate proteoglycans. Curr Opin Cell Biol, 2009.

195. Gopal, S., et al., Heparan sulfate chain valency controls syndecan-4 function in cell adhesion. J Biol Chem, 2010. 285(19): p. 14247-58.

196. Morgan, M.R., M.J. Humphries, and M.D. Bass, Synergistic control of cell adhesion by integrins and syndecans. Nat Rev Mol Cell Biol, 2007. 8(12): p. 957-69.

197. Hynes, R.O., Integrins: bidirectional, allosteric signaling machines. Cell, 2002. 110(6): p. 673-87.

198. Bengtsson, J., I. Eriksson, and L. Kjellen, Distinct effects on heparan sulfate structure by different active site mutations in NDST-1. Biochemistry, 2003. 42(7): p. 2110-5.

54

199. Mechtersheimer, S., et al., Ectodomain shedding of L1 adhesion molecule promotes cell migration by autocrine binding to integrins. J Cell Biol, 2001. 155(4): p. 661-73.

200. Beauvais, D.M., B.J. Burbach, and A.C. Rapraeger, The syndecan-1 ectodomain regulates alphavbeta3 integrin activity in human mammary carcinoma cells. J Cell Biol, 2004. 167(1): p. 171-81.

201. Beauvais, D.M. and A.C. Rapraeger, Syndecan-1-mediated cell spreading requires signaling by alphavbeta3 integrins in human breast carcinoma cells. Exp Cell Res, 2003. 286(2): p. 219-32.

202. Kimmel, C.B., et al., The shaping of pharyngeal cartilages during early development of the zebrafish. Dev Biol, 1998. 203(2): p. 245-63.

203. Lee, J.S., et al., Axon sorting in the optic tract requires HSPG synthesis by ext2 (dackel) and extl3 (boxer). Neuron, 2004. 44(6): p. 947-60.

204. Le Jan, S., et al., Functional overlap between chondroitin and heparan sulfate proteoglycans during VEGF-induced sprouting angiogenesis. Arterioscler Thromb Vasc Biol, 2012. 32(5): p. 1255-63.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 884

Editor: The Dean of the Faculty of Medicine

A doctoral dissertation from the Faculty of Medicine, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofMedicine.

Distribution: publications.uu.seurn:nbn:se:uu:diva-197691

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2013