Response of fibroblasts and chondrosarcoma cells to mechanical and chemical stimuli

Juha Piltti

Department of Integrative Medical Biology Umeå 2017

Responsible publisher under Swedish law: the Dean of the Medical Faculty This work is protected by the Swedish Copyright Legislation (Act 1960:729) ISBN: 978-91-7601-710-4 ISSN: 0346-6612 New series No. 1896 Electronic version available at http://umu.diva-portal.org/ Printed by: UmU-tryckservice, Umeå University Umeå, Sweden 2017

Dedicated to my family

Table of Contents

Table of Contents i Abstract iii Abbreviations v Tiivistelmä vii List of original papers ix 1. Background 1 1.1 1 1.2 3 1.3 6 1.4 11 1.5 Current osteoarthritis therapies 13 1.6 Strategies for novel osteoarthritis therapies 15 1.7 Mechanical stresses in articular cartilage 19 1.8 Oxygen tension in articular cartilage 24 1.9 Ras homolog gene family member A and Rho-associated kinase signaling 27 2. Aims of the thesis 30 3. Materials and methods 31 3.1 Cyclic stretching (paper I) 31 3.2 Hypoxic cell cultures (paper III, IV) 32 3.3 Rho-kinase inhibitor Y-27632 treatments (paper II, IV) 32 3.4 Phosphoprotein enrichment 33 3.5 Two-dimensional gel electrophoresis and mass spectrometric protein identification 33 3.6 Western blot analyses 34 3.7 Time-lapse cell monitoring 35 3.8 Immunofluorescence imaging 36 3.9 Mass spectrometric quantification with Orbitrap Elite 36 3.10 Mass spectrometric quantification with Synapt G2-Si HDMS 37 3.11 Bioinformatic protein analysis 37 3.12 Enzyme linked immunosorbent assay (ELISA) of type II 38 3.13 Dimethylmethylene Blue assay (DMMB) for glycosaminoglycans 38 3.14 qRT-PCR analysis 38 3.15 Statistical analysis 39 4. Results 40 4.1 Cyclic stretching induced phosphoproteomic changes in HCS-2/8 cells (paper I) 40 4.2 Effects of Rho-kinase inhibitor Y-27632 to fibroblasts (paper II) 40 4.3 Hypoxia-induced responses in the proteome of HCS-2/8 cells (paper III) 43

i

4.4 Hypoxia and Rho-kinase inhibitor Y-27632 induced cartilage matrix production in chondrocytic cells (paper IV) 45 5. Discussion 49 5.1 The phosphoproteomic analysis of cyclically stretched HCS-2/8 cells (paper I) 49 5.2 Cyclic stretching induced responses in chondrocytic cells (paper I) 50 5.3 Short-term Rho-kinase inhibition did not induce -specific cartilage production in fibroblasts (paper II) 51 5.4 Rho-kinase inhibition induced fibroblast migration and transient proliferation (paper II) 52 5.5 Increased cellular migration was potentially mediated by mDia and Rho-kinase mediated signaling routes (paper II) 54 5.6 Long-term hypoxia supported chondrocyte specific gene expression (paper III) 55 5.7 Long-term hypoxia induced known and novel protein responses in chondrocytic cells (paper III) 55 5.8 A short-term Rho-kinase inhibition at normoxic or at hypoxic conditions did not induce HCS-2/8 cell proliferation (paper IV) 58 5.9 A long-term exposure of the Rho-kinase inhibitor Y-27632 at hypoxic atmosphere enhanced cartilage extracellular matrix production (paper IV) 59 5.10 Long-term Y-27632 exposure induced several-fold protein changes at both oxygen tensions (paper IV) 59 5.11 Long-term exposure of the Rho-kinase inhibitor Y-27632 influenced changes to S100 protein synthesis (paper IV) 61 6. Conclusions 65 Acknowledgements 67 References 68

ii Abstract

Osteoarthritis is an -related disease that progressively destroys joint cartilage. This disease causes pain and stiffness of the joints, and at advanced stages, limitations to the movement or bending of injured joints. Therefore, it often restricts daily activities and the ability to work. Currently, there is no cure to prevent its progression, although certain damaged joints, such as fingers, knees and hips, can be treated with joint replacement surgeries. However, joint replacement surgeries of larger joints are very invasive operations and the joint replacements have a limited lifetime.

Cell-based therapies could offer a way to treat cartilage injuries before the ultimate damage of osteoarthritis on articular cartilage. The development of novel treatments needs both a good knowledge of articular cartilage biology and tissue engineering methods. This thesis primarily investigates the effects of mechanical cyclic stretching, a 5% low oxygen atmosphere and the Rho- kinase inhibitor, Y-27632, on protein responses in chondrocytic human chondrosarcoma (HCS-2/8) cells. Special focus is placed on Rho-kinase inhibition, relating to its potential to promote and support extracellular matrix production in cultured and its role in fibroblast cells as a part of direct chemical . The means to enhance the production of cartilage-specific extracellular matrix are needed for cell-based tissue engineering applications, since cultured chondrocytes quickly lose their cartilage-specific phenotype.

A mechanical 8% cyclic cell stretching at a 1 Hz frequency was used to model a stretching rhythm similar to walking. The cellular stretching relates to stresses, which are directed to chondrocytes during the mechanical load. The stretch induced changes in related, e.g., to certain cytoskeletal proteins, but also in enzymes associated with protein synthesis, such as eukaryotic elongation factors 1-beta and 1-delta. Hypoxic conditions were used to model the oxygen tension present in healthy cartilage tissue. Long- term hypoxia changed relative amounts in a total of 44 proteins and induced gene expressions of and type II collagen, in addition to chondrocyte differentiation markers S100A1 and S100B. A short-term inhibition of Rho- kinase failed to induce extracellular matrix production in fibroblasts or in HCS-2/8 cells, while its long-term exposure increased the expressions of chondrocyte-specific genes and differentiation markers, and also promoted the synthesis of sulfated glycosaminoglycans by chondrocytic cells. Interestingly, Rho kinase inhibition under hypoxic conditions produced a more effective increase in chondrocyte-specific gene expression and synthesis of extracellular matrix components by HCS-2/8 cells. The treatment induced changes in the synthesis of 101 proteins and ELISA analysis revealed a sixfold higher secretion of type II collagen compared to control cells. The secretion of sulfated glycosaminoglycans was simultaneously increased by 65.8%. Thus, Rho-kinase inhibition at low oxygen tension can be regarded as a potential way to enhance extracellular

iii matrix production and maintain a chondrocyte phenotype in cell-based tissue engineering applications.

iv Abbreviations

2D PAGE Two-dimensional polyacrylamide gel electrophoresis 4F2hc 4F2 cell-surface antigen heavy chain ACAN Aggrecan core protein ACI Autologous chondrocyte implantation technique ADAMTS-5 A disintegrin and metalloproteinase with thrombospondin motifs 5 (aggrecanase 5) Akt Protein kinase B (PKB) ATDC5 A teratocarcinoma derived chondrogenic cell line ATP Adenosine triphosphate BCA Bicinchoninic acid BMP morphogenetic protein CDC42 Cell division control protein 42 homolog CI Confidence interval CID Collision induced dissociation COL1A1 Collagen alpha-1(I) chain () COL2A1 Collagen alpha-1(II) chain (type II collagen) COX-2 Cyclooxygenase-2 DMEM Dulbecco's Modified Eagle Medium DMMB Dimethylmethylene Blue assay DMOAD Disease-modifying drug therapy for osteoarthritis eEF Eukaryotic elongation factor ELISA Enzyme linked immunosorbent assay ERK Extracellular signal–regulated kinases ESI Electrospray ionization FAK Focal kinase FBS Fetal bovine serum FDA Food and Drug Administration FGF Fibroblast FOXO Forkhead box protein O GalNAc N-Acetylgalactosamine GlcNAc N-Acetylglucosamine GAP GTPase activating factor GAPDH Glyceraldehyde 3-phosphate dehydrogenase GDI Guanosine nucleotide dissociation inhibitor GDP Guanosine diphosphate GTP Guanosine triphosphate HCS-2/8 Human chondrosarcoma cell line-2/8 HDAC6 Histone deacetylase 6 HIF Hypoxia-inducible factor HRP Horseradish peroxidase Hz Hertz (unit of frequency)

v IL Interleukin Inhba Inhibin beta A chain IPA® Ingenuity® Pathway Analysis LAT1 L-type amino acid transporter 1 LC Liquid chromatography MACI Matrix-assisted autologous chondrocyte implantation mDia Mammalian diaphanous formin Mig6 Mitogen-inducible gene 6 MMP Matrix metalloproteinase MPa Megapascal mRNA messenger RNA MS/MS Tandem mass spectrometry m/z Mass-to-charge ratio NADH Nicotinamide adenine dinucleotide, reduced NDRG1 N-myc downstream regulated gene-1 ODDD Oxygen-dependet degradation domain PI3K Phosphatidylinositol 3-kinase PBS Phosphate buffered saline PKA Protein kinase A pVHL von Hippel-Lindau tumor suppressor qRT-PCR Quantitative reverse transcription polymerase chain reaction Rac1 Ras-related C3 botulinum toxin substrate 1 RAGE Receptor for advanced glycation endproducts RhoA Ras homolog gene family member A ROCK Rho-associated protein kinase RPLP0 60S acidic ribosomal protein P0 RVD Regulatory volume decrease SDS Sodium dodecyl sulfate sGAG sulfated SOX5/6/9 Sex determining region Y -box 5/6/9 transcription factors TFA Trifluoroacetic acid TGF-β Transforming growth factor β TNF-α -α TOF Time-of-flight TRPV4 Transient receptor potential vanilloid 4 VCAN Versican core protein VEGFA Vascular endothelial growth factor A Y-27632 Rho-kinase inhibitor ((1R,4r)-4-((R)-1- aminoethyl)-N-(pyridin-4-yl) cyclohexanecarboxamide)

vi Tiivistelmä

Nivelrikko on nivelten tulehdussairaus, johon liittyy voimakkaita kipuja, ja joka johtaa edetessään nivelruston vaurioitumiseen. Nivelten jäykkyyden lisäksi nivelrikko aiheuttaa pahimmillaan liikerajoitteita vaurioituneissa nivelissä. Tämän takia nivelsairaudet heikentävät elämänlaatua ja nuoremmilla potilailla ne voivat aiheuttaa myös työkyvyn rajoittuneisuutta. Nivelrikon tai vaurioituneiden nivelpintojen hoitoon ei toistaiseksi ole käytettävissä sairauden parantavaa tai sen edistymisen pysäyttävää lääkkeellistä hoitoa. Tavanomaiset nivelrikon yhteydessä käytetyt lääkehoidot ovat kipua lievittäviä. Pahiten vaurioituneita niveliä esimerkiksi polvissa, lantiossa tai sormissa voidaan hoitaa leikkauksilla, jolloin potilaan oma vaurioitunut nivel korvataan keinonivelellä. Nämä operaatiot ovat luonteeltaan invasiivisiä, ja menetelmien rajoitteena on mm. keinonivelten rajallinen kesto ja haasteet niiden uusimisessa.

Uusien solupohjaisten terapioiden toivotaan mahdollistavan nivelvaurioiden hoitamisen jo ennen niiden etenemistä vaikea-asteiseen tilaan, jossa kudoksen alkuperäinen hyaliinirusto on tuhoutunut. Näiden uusien hoitomuotojen kehittäminen edellyttää hyvää tietämystä sekä nivelrustosta että erilaisista kudosteknologisista korjausmenetelmistä. Väitöskirjatyössäni on selvitetty mekaanisen venytyksen, matalan happiosapaineen ja rustokudoksen soluväliainetuotantoon liittyviä vasteita proteiinitasolla. Väitöskirjatyön pääpainotus on liittynyt Rho-kinaasi-inhibiittori Y-27632:n indusoimiin vasteisiin tukea rustosolujen ja niiden kaltaisten solujen soluväliainetuotantoa sekä ylläpitää solujen ilmiasua. Rustosolujen kasvattaminen in vitro -olosuhteissa johtaa niiden ilmiasun taantumiseen ja rustosoluille tyypillisen soluväliainetuotannon loppumiseen. Rustospesifisen ilmiasun säilyttäminen ja myös soluväliainetuotannon tehostaminen ovat keskeisiä edellytyksiä solupohjaisten hoitomuotojen kehittämisessä.

Tässä väitöskirjatyössä käytettiin jaksoittaista 8 %:n venytystä 1 Hz:n taajuudella mallintamaan mekaanisen venytyksen aiheuttamia muutoksia rustosolujen kaltaisissa HCS-2/8-soluissa. Mekaaninen venytys välittyy soluihin niiden rakenteen kautta, ja tutkitun kuormituksen havaittiin aktivoivan soluissa ERK-solusäätelyreitin. Tämän lisäksi jaksoittainen mekaaninen venytys indusoi muutoksia esimerkiksi tiettyihin solutukirangan proteiineihin ja proteiinisynteesiin liittyviin eukaryoottisiin elongaatiotekijöihin. Mekaanisen soluvenytyksen lisäksi väitöskirjatyössäni selvitettiin rustokudoksessa pitkäaikaisesti vallitsevan matalan happiosapaineen vaikutuksia rustosoluihin proteiinitasolla. Matala happiosapaine indusoi muutoksia yhteensä 44:n proteiinin synteesitasoihin

vii verrattaessa niitä vastaaviin tasoihin normaali-ilmakehässä. Hypoksiaolosuhde lisäsi tyypin II kollagenin, aggrekaanin ja rustosolujen erilaistumisastetta kuvaavien S100A1 ja S100B geenien ilmentymistä, mitkä ovat suotuisia vasteita solujen rustonkaltaisen ilmiasun säilymiselle. Ilmiasun ylläpidosta huolimatta näiden geenitason muutosten ei havaittu kuitenkaan lisäävän soluväliainetuotantoa proteiinitasolla. Rho-kinaasi-inhibition vaikutusten selvittäminen sekä fibroblastisoluissa että rustosolujen kaltaisissa jatkuvakasvuisissa HCS-2/8-soluissa on ollut keskeisessä osassa väitöskirjatyötäni. Lyhytaikaiset pari vuorokautta kestäneet Y-27632-käsittelyt indusoivat erilaisia toiminnallisia vasteita tutkituissa soluissa, mutta ne eivät indusoineet rustosolulle tyypillisen soluväliainetuotannon käynnistymistä fibroblastisoluissa, eikä niiden havaittu vaikuttavan kummankaan solutyypin ilmiasuun. Sen sijaan pitkäaikaisen, neljä viikkoa kestäneen, hypoksiaolosuhteessa toteutetun Y- 27632-altistuksen havaittiin lisäävän rustosoluspesifistä tyypin II kollageenin ja sulfatoitujen glykosaminoglykaanien soluväliainetuotantoa HCS-2/8-soluissa. Pitkäkestoinen Y-27632-käsittely indusoi muutoksia yhteensä 101:n proteiinin synteesitasoihin. Immunologisen ELISA-analyysin perusteella Rho-kinaasin inhibitio 5 % happiosapaineessa lisäsi jopa kuusinkertaisesti tyypin II kollageenituotantoa. Sulfatoitujen glykosaminoglykaanien synteesi lisääntyi samaan aikaan noin 65 %. Lopputuloksena voidaan todeta, että hypoksiaolosuhteissa toteutettava Rho- kinaasi-inhibitio vaikuttaa potentiaalisesti sekä ylläpitää rustosolujen ilmiasua että lisää niiden soluväliainetuotantoa, mitkä ovat keskeisiä edellytyksiä kudosteknologisten rustonkorjausmenetelmien kehittämiselle.

viii List of original papers

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

I. Piltti J, Häyrinen J, Karjalainen HM, Lammi MJ (2008) Proteomics of chondrocytes with special reference to phosphorylation changes of proteins in stretched human chondrosarcoma cells. Biorheology 45(3-4), 323-335

II. Piltti J, Varjosalo M, Qu C, Häyrinen J, Lammi MJ (2015) Rho- kinase inhibitor Y-27632 increases cellular proliferation and migration in human foreskin fibroblast cells. Proteomics 15(17), 2953-2965

III. Piltti J, Bygdell J, Lammi MJ (2017) Effects of long-term hypoxia in human chondrosarcoma cells. Manuscript submitted

IV. Piltti J, Bygdell J, Fernández-Echevarría C, Marcellino D, Lammi MJ (2017) Rho-kinase inhibitor Y-27632 and hypoxia synergistically enhance chondrocytic phenotype and change the S100 protein profile in human chondrosarcoma cells. Manuscript submitted

ix

1. Background

1.1 Cartilage tissue

Cartilage is a special that is predominantly present in the skeletal system and joints, but is also in found in -based organs, such as the ears and nose. Three different cartilage types are expressed in the human body, namely , fibrocartilage and hyaline cartilage. Elastic cartilage is found in soft tissues, while collagen fiber-based fibrocartilage is found, e.g., in intervertebral discs, and the most common cartilage type, hyaline cartilage is typical in diarthrodial joints. Hyaline cartilage makes movements of joints possible and absorbs and also distributes weight-bearing stress more evenly between skeletal (Tyyni and Karlsson 2000). Articular cartilage tissue differs in its structure from the other human tissues in a way that there are no blood vessels or nerves present in cartilage. Moreover, cartilage tissue does not have lymph nodes. Due to its avascularity, there is a relatively slow diffusion between and cartilage tissue and this is the only mechanism for nutrient and metabolite transport and for gas exchange. In accordance to these special features of flexible connective tissue, the structural composition of cartilage is divided into two phases, namely the solid extracellular matrix and a water phase. The correct composition and structural organization of the extracellular matrix components are important for the correct function of healthy joints. The natural turnover rate of mature articular cartilage components is slow. If cartilage tissue is damaged, the avascularity of cartilage tissues reduces the activation and progression of an individual’s own mechanisms. In addition to this, the surface of articular cartilage is not covered by , a dense and vascularized connective tissue membrane of, e.g., tracheal cartilage. This type of membraneous structure could offer a cell source of which would be needed for new differentiated chondrocytes to support and/or enhance the production of extracellular matrix. This type of mechanism could provide, at least to some extent, repair to damaged articular cartilage. Since this is missing in articular cartilage it means that it has poor ability to self-repair any damage. The lack of proper self-repair mechanisms can lead to the progression of outstanding damages, especially in articular cartilage. Continuous mechanical loading can both induce and enhance the formation of mechanical degenerative injuries especially in the case when joint structure is not intact and lubrication is not sufficient. In addition to mechanical factors, some genetic disorders and diseases can induce damages in articular cartilage.

1

The structure and the composition of the extracellular matrix

Articular cartilage is a thin tissue, with a thickness of only few millimeters in large joints like the knee and hip. The correct structure and organization of specific zones is critical for its function. Articular cartilage consists of five zones (Fig 1A), namely the superficial zone on the top of the tissue, then the intermediate and middle zones, the deep zone and finally the calcified zone that is located next to the subchondral bone (Sophia Fox et al. 2009). The calcified zone connects articular cartilage to subchondral bone. Different zones have specialized structural features. Chondrocytes are morphologically more flattened in the superficial zone than in the other zones. In this zone, collagen fibers are thin and they have a horizontal, surface-oriented direction. The water content is highest in the superficial zone while the concentration is the lowest. The orientation of collagen fibers start to curve in the intermediate zone, and they are vertical in the deep zone. The collagen fibers are also thicker in the middle and the deep zones than in the superficial zone (Hedlund et al. 1993). Proteoglycan concentration is the highest in the deep zone, and there the chondrocytes located around the collagen fibers. A calcification line isolates the deep and the calcified zone from each other.

Figure 1. The structure of articular cartilage (A). Type II collagen network builds up the skeleton of cartilage tissue (B).

Collagen network

Chondrocytes synthesize and maintain the extracellular matrix of cartilage. The extracellular matrix maintains the functional properties of cartilage. Although, the number of the chondrocytes is relatively low, even less than 2% of the total volume of cartilage tissue, chondrocytes secrete and maintain all the necessary extracellular matrix components in healthy cartilage (Hunziker et al. 2002) The main structural components of extracellular matrix are type II collagen [known also as procollagen alpha-1(II) chain] and

2

proteoglycans. The form 20% of the wet weight of articular cartilage, the proteoglycans approximately form 10%, and the remaining part is water (Sophia Fox et al. 2009). The collagens build up the skeleton of cartilage and regulate its structure and its dimensions (Fig 1B). The proteoglycans provide the flexibility of cartilage. Type II collagen represents roughly 90% of all collagens in articular cartilage and, therefore, it is referred as cartilage- specific collagen type (Eyre et al. 2002). The synthesis of type II collagen occurs in two steps. The first step involves the chondrocytes’ biosynthesis of procollagen, which are then secreted in the second step and enzymatically modified to form type II collagen by the removal of certain terminal propeptides (Prockop et al. 1979). The collagen chains are typically assembled into triple helices, which form collagen structures. These collagen are further organized to form stronger collagen fibers, which provide the excellent tensile strength of cartilage tissue. The other collagen types in the articular cartilage are mostly types III, VI, IX and XI collagens, however, there are also small amounts of types X, XII and XIV also present in articular cartilage (Eyre et al. 2002). Collagen types IX and XI are needed together with type II collagen to form a crosslinking network that functions as the basic skeleton of cartilage tissue (Mendler et al. 1989, Wu et al. 1992, Wu and Eyre 1995). The pericellular collagen type VI can connect the collagen network to other extracellular matrix components (Eyre et al. 2002). The majority of the minor collagens take part in the formation of different types of fibrillary structures present in cartilage tissue. However, the precise role of minor cartilage collagens is not known.

1.2 Glycosaminoglycans

Glycosaminoglycans are long unbranched polysaccharides that are built-up by repeated disaccharide units (Gandhi and Mancera 2008). These disaccharide units contain one or two kinds of modified hexosaminecarbohydrates, namely N-acetylgalactosamine (GalNAc) or N- acetylglucosamine (GlcNAc), in addition to uronic acid, which can either be glucuronic or iduronic acid. However, in keratan sulfate the uronic acid is replaced by galactose. Glycosaminoglycans have highly negative charges and their structure affects their viscosity. In articular cartilage five different types of glycosaminoglycans are synthesized: chondroitin sulfate, dermatan sulfate, heparin sulfate, keratan sulfate and hyaluronan. Chondroitin sulfate, dermatan sulfate, keratan sulfate and hyaluronan are the most abundant and are considered important glycosaminoglycans for cartilage structure and function. Heparin, like heparan sulfate, is also found in articular cartilage but this form might be connected to a variety of normal heparan sulfate production in cartilage (Parra et al. 2012).

3

Chondroitin sulfate

Chondroitin sulfate is the most abundant glycosaminoglycan in articular cartilage and it has a major role as a component of aggregating proteoglycan, aggrecan (Kiani et al. 2002). It is a long, acidic and sulfated polysaccharide chain that consists of repeating units of glucuronic acid and N- acetylgalactosamine (Gandhi and Mancera 2008). Chondroitin sulfate chains are covalently bound to core proteins via the hydroxyl groups of serine residues. The sulfate groups of chondroitin sulfate are enzymatically added by specific sulfotransferases primarily at the 4- and/or 6-position (C4 and/or C6) of galactosamine units, but non-sulfated disaccharide units are also known to exist. The distribution of C4 and C6 chondroitin sulfates is related to developmental status of articular cartilage. In mature human articular cartilage tissue C6 chondroitin sulfate is the major form (Bayliss et al. 1999). During maturation of articular cartilage, C4 content decreases and C6 content increases, but after this step the proportion between C4 and C6 remains relatively stable. However, certain diseases, such as osteoarthritis, change this ratio.

It is noteworthy to mention that chondroitin sulfate has been tested to treat osteoarthritis, more specifically, to provide relief from osteoarthritis-related pain. Several clinical studies involving chondroitin sulfate to treat osteoarthritis have been performed. Although the first reports showed a good bio-availability of chondroitin sulfate in humans, later studies ended up with conflicting results providing doubt to its therapeutic potential (Mazieres et al. 2001, Clegg et al. 2006, Roman-Blas et al. 2017). The most recent long- term efficacy and safety evaluations of chondroitin sulfate and glucosamine revealed that these agents are unable to heal cartilage injuries, unable to improve function of damaged cartilage or relieve pain (Sawitzke et al. 2010, Roman-Blas et al. 2017). These evaluations were based on comprehensive double-blind placebo controlled studies. However, the potential of chondroitin sulfate and glucosamine to treat damaged cartilage is still frequently discussed (Hochberg et al. 2016, Bishnoi et al. 2016). Chondroitin sulfate can be used as a prescribed medicine in some countries, however in the United States, the Food and Drug Administration (FDA) has classified chondroitin sulfate only as a dietary supplement.

Dermatan sulfate

Dermatan sulfate is a glycosaminoglycan that is a major component of small non-aggregating proteoglycans in articular cartilage (Roughley and White 1989). This glycosaminoglycan is more abundant in connective tissues, such as , the walls of blood vessels, heart valves and . Dermatan

4

sulfate proteoglycans are involved, e.g., in , and it has anticoagulant activity. In cartilage, dermatan sulfate proteoglycans function is connected with collagen fibril organization and the inhibition of fibrillogenesis. Structurally, dermatan sulfate is composed of glucuronic acid and iduronic acid, and N-acetylgalactosamine. Two dermatan sulfate epimerases can convert glucuronic acid into iduronic acid making iduronic acid a major structural component of dermatan sulfate (Malmström et al. 2012). Sulfation occurs mainly at positions 4 and/or 6 of GalNAc, but sulfation is also possible at position 2 of glucuronic or iduronic acid.

Keratan sulfate

Keratan sulfate is a highly abundant glycosaminoglycan in cornea but is also abundantly present in articular cartilage and, to a smaller amount, in other tissues, such as bone and the brain (Krusius et al. 1986, Funderburgh 2000). Keratan sulfates consist of repeating galactose and N-Acetylglucosamine units with the pattern -3Galβ1–4GlcNAcβ1-. Sulfation occurs at carbon 6 both at galactose and/or N-Acetylglucosamine. There are three different known types of linking areas, which connect keratan sulfate to core proteins. Differences in these linking areas are used to designate keratan sulfates to either Type I, II or III (Krusius et al. 1986, Funderburgh 2002). Type I keratan sulfate is the typical form in cornea, while type II is predominantly in articular cartilage. Type III keratan sulfate is located in brain tissue. Despite their relative abundancies, keratan sulfate types I to III are not tissue-specific. Type I keratan sulfate (KS-I) is attached to asparagine residue of core protein, type II keratan sulfate (KS-II) to serine or threonine residues, and type III (KS-III) to serine residues. This amino acid-specific binding mechanism is based on N- or O-links of specific carbohydrate chains. KS-I attaches to asparagine via N-acetylglucosamine, KS-II attaches to serine or threonine via N-acetylgalactosamine and KS-III to serine via mannose (Funderburgh 2002).

Heparan sulfate and heparin

Heparan sulfate is a common glycosaminoglycan component of proteoglycan structures and is found in the extracellular matrix of a number of tissues (Sarrazin et al. 2011). Heparan sulfate consists of repeating units of glucuronic acid or iduronic acid and N-acetylglucosamine (Garg et al. 2003). Epimerization can convert glucuronic acids into iduronic acids similarly to dermatan sulfate and this provides structural variability to heparan sulfate chains. This epimerization needs C5-epimerase enzyme (Kreuger and Kjellén 2012). The sulfation of heparan sulfate can occur at N-acetylglucosamine positions 3 and 6 (C3 and C6) and iduronic acid position 2 (C2). Heparin is

5

chemically similar to heparan sulfate, but its N- and O-sulfo groups of N- acetylglucosamine are different. The ratio of glucuronic and iduronic acids is different in heparan sulfate and heparin. Heparan sulfate contains more glucuronic acid than heparin. Heparin does not have important role in articular cartilage, but it acts as an efficient anticoagulant due to its binding to antithrombin. The profound biological role of heparin might relate to protection of injury sites from invading microorganisms. However, it is notable that heparin is a highly sulfated molecule, which has the highest density of negative charges among glycosaminoglycans. Heparin disaccharide unit consists typically of 2-O-sulfated iduronic acid and 6-O- sulfated, N-sulfated glucosamine (Garg et al. 2003).

Hyaluronan

Hyaluronan (also known as ) is a nonsulfated glycosaminoglycan molecule, which is not covalently attached to a protein core (Bastow et al. 2008). It is an abundant glycosaminoglycan in articular cartilage, but is also found in skin, synovial fluid and soft tissues (Fraser et al. 1997). Hyaluronan consists of repeating units of glucuronic acid and N- acetylglucosamine (Bastow et al. 2008). The nonsulfated nature of hyaluronan differs from the other cartilage-specific glycosaminoglycans. Due to the lack of sulfation, hyaluronan does not form disulfide bond based crosslinks with other molecules. On the other hand, hyaluronan has an important role together with glycoprotein lubricin and dipalmitoyl phosphatidyl choline (DPPC) to reduce friction on hyaline cartilage surfaces in synovial joints (Wang et al. 2013, Seror et al. 2015). Alhough hyaluronan provides a clear role for synovial fluid viscoelasticity and lubrication, it also has a structural role in the extracellular matrix of cartilage. Hyaluronan molecules specifically bind to aggrecan and thereby assemble larger aggregate structures. Unlike other glycosaminoglycans, hyaluronan synthesis occurs at the plasma membrane, while other glycosaminoglycans are synthesized in the Golgi apparatus (Prehm 1984). There are three hyaluronan synthase enzymes that synthesize hyaluronan chains (Itano et al. 1999). These enzymes catalyze the synthesis of hyaluronan at different rates and also affect the length of hyaluronan polymer chains.

1.3 Proteoglycans

Proteoglycans refers to macromolecules that consist of a protein core with at least one glycosaminoglycan covalently attached (Hardingham and Fosang 1992). So-called lecticans form large aggregates with hyaluronan. The proteoglycans have high density of negative charges under normal physiological conditions due to their sulfate groups and glucuronic acid.

6

(Roughley et al. 2006). These negative charges have a great influence on cartilage structure and function since they attract positively charged sodium ions that produce the osmotic pressure that adjusts prevailing water phase in articular cartilage. Water phase provides the typical swollen nature of cartilage tissue and the network of collagen acts in a cohesive way to maintain its firm structure. The water phase has an important role for the elasticity of cartilage. Mechanical load pushes water out from tissue and when the load is released, the osmotic pressure draws water back into tissue.

Aggrecan

Aggrecan is the most abundant proteoglycan in articular cartilage. It is a typical proteoglycan in tissues exposed to high compression, such as cartilage or intervertebral discs (Kiani et al. 2002, Sivan et al. 2014). Aggrecan belongs to group of lecticans, which also contains versican and two neuronal proteoglycans, brevican and neurocan. Structurally, aggrecan is a high-molecular weight proteoglycan that consists of three globular domains, an extended core protein, and contains more than 100 sulfated glycosaminoglycan molecules. The average molecular weight of one non- aggregated aggrecan monomer can vary from 1 to 3 million daltons. Two of the globular domains are located at the N-terminus and one globular domain at the C-terminus. The molecular weight of aggrecan core protein is roughly 300 kDa (Chandran and Horkay 2012). The C-terminal globular domain of aggrecan contains one EGF-like domain, one lectin-like domain and one CRP-like domain (Fülöp et al. 1993). The glycosaminoglycans attached to the aggrecan core protein are chondroitin and keratan sulfates. One aggrecan molecule contains roughly twice as many chondroitin sulfate molecules as keratan sulfates. The non-globular domains of aggrecan also bind a variety of different O- and N-linked oligosaccharides. The size of aggrecan aggregates formed together with hyaluronan traps them inside the collagen network of articular cartilage.

Versican

Versican is a large hyaluronan binding chondroitin sulfate containing proteoglycan, which is expressed in various tissues (Zimmermann and Ruoslahti 1989). Structurally, versican shares many similarities to aggrecan although it lacks a second globular domain and its binding to hyaluronan and its protein link occurs in a different manner (Matsumoto et al. 2003). Versican binds fewer glycosaminoglycan chains than aggrecan and its C- terminal globular region does not exhibit alternative splicing like aggrecan (Grover and Roughley 1993). The C-terminal globular domain of versican contains two EGF-like domains, one lectin-like domain and one CRP-like

7

domain (Zimmermann and Ruoslahti 1989). Versican has a role in articular cartilage development, but its synthesis is decreased after this process (Shibata et al. 2003). Instead of versican aggregates, mature cartilage tissue matrix is based on aggrecan aggregates, and due to this reciprocal synthesis profile, it is suggested that versican could act as a temporary matrix for developing cartilage structure (Grover and Roughley 1993). Versican gene is also expressed in mature articular cartilage, although its expression level is substantially less than expression levels of aggrecan. Versican transcriptional expression in mature cartilage tissue is suggested to specifically relate to a role in articular cartilage surfaces (Matsumoto et al. 2006).

Small leucine-rich repeated proteoglycans

Biglycan, , epiphycan, fibromodulin and lumican are minor proteoglycans in articular cartilage that belong to a family of small leucine- rich repeat (LRR) proteoglycans (Kobe and Deisenhofer 1994, Chakravarti S and Magnuson T 1995, Grover et al. 1995). Depending on the type of glycosaminoglycan substituents, the number of LRRs, and the gene structures, small LRR proteoglycans are classified into different subfamilies (Roughley 2006). Fibromodulin and decorin are the most abundant minor proteoglycans in cartilage tissue. These two proteoglycans have ten LRRs similar to and lumican. Structurally, biglycan and decorin share close similarities due to the 8th exon structure, and they both are classified as dermatan sulfate proteoglycans. Fibromodulin and lumican have similarities in their 3rd exon and are classified as keratan sulfate proteoglycans. More accurate than previous classification, decorin usually contains one dermatan sulfate chain, which can also be a chondroitin sulfate chain depending on its tissue origin. Biglycan has two glycosaminoglycan chains, which can be either chondroitin or dermatan sulfates. Fibromodulin and lumican have up to four N-linked keratan sulfates. Epiphycan has one dermatan or chondroitin sulfate chain, and it has only seven LRRs, which is less in compared to the other LRR protein family members presented here (Johnson et al. 1997). The role of epiphycan relates to embryogenesis, when it is expressed in epiphysis. Epiphycan is also suggested to stabilize cartilage matrix and maintain joint integrity (Nuka et al. 2010). Decorin can bind to collagens and interact with certain growth factors such as transforming growth factor-β1 (TGF-β1) (Ferdous et al. 2007). The function of decorin and fibromodulin was suggested to relate to the organization of collagen fibrils and the further formation of collagen structure (Burton-Wurster et al. 2003).

8

Agrin and perlecan, type proteoglycans

Agrin is a heparan sulfate proteoglycan, which is typically associated to the development of neuromuscular junctions via low-density lipoprotein receptor-related protein 4 (LRP4) or α-dystroglycan mediated and basement membrane connections (Zhang et al. 2008, Campanelli et al. 1996). However, this large proteoglycan has also been found in cartilage tissue and the non-neuronal isoform (y0, z0) affects both chondrocyte differentiation and cartilage formation in vivo (Eldridge et al. 2016). Agrin mediates cartilage differentiation both in a LRP4- and α-dystroglycan- dependent manner prior to the activation and upregulation of Sox9.

In articular cartilage, chondrocytes are surrounded by a thin layer of a pericellular matrix. This pericellular matrix separates chondrocytes from the extracellular matrix. Perlecan (also known as basement membrane-specific heparan sulfate proteoglycan core protein) is a proteoglycan with several domains containing a core protein and three glycosaminoglycan chains (Costell et al. 1999). These glycosaminoglycans are typically heparan sulfates, but chondroitin sulfate can also occur in some cases. Although perlecan belongs to the minor set of proteoglycans of articular cartilage, it has been demonstrated to play an essential role in the biochemical and mechanical properties of the pericellular matrix in articular cartilage (Wilusz et al. 2012).

Cell surface proteoglycans

Syndecans form a proteoglycan family, which consists of four structurally close members, namely syndecan 1 (syndecan), syndecan 2 (fibroglycan), syndecan 3 (N-syndecan) and syndecan 4 (amphiglycan/ryudocan) (Kim et al. 1994, Elenius and Jalkanen 1994). Syndecans 1-3 are described to be predominantly expressed in a tissue-specific manner, however syndecan 4, also known as amphiglycan or ryudocan, is more commonly expressed in a wide variety of tissues (Morgan et al. 2007). Syndecans are classified to heparan sulfate proteoglycans, although syndecans 1 and 3 also contain minor chondroitin sulfate chains in their membrane proximal regions. (Gould et al. 1992, Kokenyesi et al. 1994). The syndecans have a single transmembrane spanning core protein and their structural differences are based upon the different lengths of extracellular regions and the variable intracellular region unique to each member (Morgan et al. 2007). The core proteins of syndecans interact with different kinds of ligands, such as extracellular matrix glycoproteins and growth factors. Although the syndecans are only minor proteoglycans in cartilage, both syndecan 1 and 4 have been detected in articular cartilage and their expression was observed

9

to change in osteoarthritis (Barre et al. 2000). Syndecan 3 is only transiently expressed during certain stages of chondrogenesis (Hall and Miyake 2000). When the expression of syndecan 1 was studied in a transgenic osteoarthritic mouse model, it was suggested that syndecan 1 could be involved in cartilage fibrillations and take part in this way to certain repair mechanism (Salminen-Mankonen et al. 2005). However, this possible repair mechanism is still unknown and further studies are needed. A recent study revealed that inhibition of syndecan 4 in osteoarthritic cartilage reduced matrix metalloproteinase 3 (MMP-3) expression and ADAMTS-5 aggrecanase activity (Echtermeyer et al. 2009). Thereby, syndecan 4 inhibition can be regarded as a potential target for therapeutic treatment of osteoarthritis.

Glypicans form another large heparan sulfate proteoglycan family (Filmus et al. 2008). Mammals have a total of six glypicans (glypicans 1-6), which all have a relatively similar size of their globular core protein. The core proteins are not transmembrane spanning proteins, rather they are attached to cell membranes via glycosyl phosphatidylinositol linkage (GPI). Glypicans act as co-receptors and function to regulate Wnt, Hedgehog, and bone morphogenetic protein signals. Dependent on the biological context, glypicans have either inducing or inhibiting roles. Glypican expression has been detected in human articular chondrocytes, but with much lower expression levels than in fibroblast cells (Grover and Roughley 1995).

Betaglycan (also known as transforming growth factor beta receptor III; TGFBR3) is a chondroitin and heparan sulfate containing cell surface proteoglycan (López-Casillas et al. 1994). The core protein of betaglycan acts as a TGF-β receptor. The betaglycan gene is marginally expressed in human chondrocytes and it has been suggested that betaglycan could form a heteromeric complex that regulates TGF-β signaling in chondrocytes (Grover and Roughley 1995, Parker et al. 2003). TGF-β signaling has an important role in chondrocytes, one that induces both the proliferation and the other that inhibits their differentiation to a hypertrophic stage (Li et al. 2005). A quite recent finding revealed that signaling of TGF-β/SMAD pathway relates to the development of osteoarthritis (Shen et al. 2014).

In addition to syndecans, glypicans and betaglycan, cell surface proteoglycans also contain two additional members, namely phosphacan and chondroitin sulfate proteoglycan 4 (also known as neuron glial antigen 2; CSPG4/NG2) (Iozzo and Schaefer 2015). Despite the general classification, phosphacan is not a true transmembrane bound proteoglycan, but its splice variant, transmembrane receptor protein tyrosine phosphatase beta, (RPTPβ/PTPζ) is synthesized with a transmembrane spanning domain

10

(Maurel et al. 1994, Peles et al. 1998, Nishiwaki et al. 1998). Altogether, phosphacan and its isoforms are chondroitin sulfate proteoglycans that are typically expressed in nervous tissues. Chondroitin sulfate proteoglycan 4 contains chondroitin sulfate chains as the name suggests. This proteoglycan is also primarily connected to interactions in nervous tissues, but it is expressed both in human fetal and adult articular chondrocytes, although cartilage specific function has not been recognized (Midwood et al. 1998).

1.4 Osteoarthritis

Osteoarthritis (OA) is an inflammation-related musculoskeletal disease, which can lead to the progressive destruction of articular cartilage and, ultimately, the entire joint. The etiology of this common articular cartilage disease is complex even though many risk factors are known (Michael et al. 2010). Individual-related risk factors, such as age, gender and genetic predisposition, can increase the likelihood of developing osteoarthritis. Obesity, different kinds of joint traumas, such as impact injuries and lifestyle habits can also increase the risk of developing osteoarthritis. Still, there is usually not just one factor that causes osteoarthritis and hereditary factors are considered to relate specifically to induction of osteoarthritis in majority of cases (Valdes et al. 2008). Osteoarthritis causes pain, swelling and stiffness of the joints and, in severe cases, it can limit the bending or the movement of the injured articular joint (Hunter et al. 2008, Michael et al. 2010). The associated pain and the reduced joint function can restrict both routine activities and the ability to work.

Osteoarthritis gradually destroys articular cartilage until the injury also affects the underlying bone. Osteoarthritis typically develops in joints of the knees, hips and fingers, but it may also occur in other synovial joints, such as the lower back and neck (Nuki 1999). The normal maintenance of healthy articular cartilage is based on the correct balance between anabolic and catabolic activities. A correct balance is critical since the number of chondrocytes is small and the normal turnover rate of extracellular matrix components is slow. If the normal maintenance of articular cartilage is disturbed, chondrocytes and synovial cells begin to produce and secrete interleukin-1, tumor necrosis factor-alpha, chemokines and other factors that upregulate the expression and activity of proteolytic enzymes (Goldring and Otero 2011). Increased catabolic activity induces inflammation and the gradual breakdown of cartilage tissue. Inflammation is a normal response to tissue injury, but prolonged inflammation is harmful and can produce even more damage. In addition, the small degraded matrix components of the extracellular matrix act as feedback regulators, which further enhance inflammation responses. Mechanistically the progress of osteoarthritis can

11

be divided into different steps (Dijkgraaf et al. 1995). If the extracellular matrix structure is damaged, the proteoglycan content starts to decrease and the water phase volume will increase. This kind of structural change makes articular cartilage more vulnerable to abrasion and mechanical impacts. The chondrocytes may respond to the change in environment by increasing the production of extracellular matrix. However, all the risk factors cannot be eliminated and at some point the chondrocytes are unable to maintain the balance between anabolic and catabolic activities, and cartilage-degrading processes start to predominate. Although the physiological responses to the progressive osteoarthritis are mostly known, the comprehensive mechanism of the origin of osteoarthritis is still not known in all parts. Thereby, it is challenging to prevent and treat articular cartilage damage.

Diagnosis of osteoarthritis

Although pain is the most characteristic symptom of osteoarthritis, it does not correlate well to the actual stage and radiographic findings of osteoarthritis (Hannan et al. 2000). Pain is a subjective feeling; since articular cartilage is aneural tissue, pain is mediated via the nerves of surrounding tissues. Indirect sources of joint pain have been studied and it has been suggested that there is a complex, even biopsychosocial framework, which plays a role in osteoarthritis and pain sensation (Dieppe and Lohmander 2005). The only reliable diagnosis of osteoarthritis is based on interviews with patients and clinical investigation (Hunter et al. 2008). Clinical examinations focus on movement ranges of joints. Commonly the diagnosis is confirmed by radiography, but additional imaging techniques, such as magnetic resonance imaging (MRI), ultrasound or optical coherence tomography (OCT) can also be used nowadays (Braun and Gold 2012). Radiography is still an important tool for osteoarthritis diagnosis, although there are certain risks to misinterpret the radiographic findings. The technique is commonly known to be insensitive to early stage changes of osteoarthritic cartilage. On the other hand, all positive radiographic findings do not explain the source of pain, which might be related, for example, to bursitis (Hannan et al. 2000). However, in later stages of osteoarthritis, radiography is a potential method that can reveal the narrowed joint spaces, bony projections, known as osteophytes and possible subchondral sclerosis and cysts in cartilage underlying bones (Hunter et al. 2008). Magnetic resonance imaging is becoming more common in osteoarthritis diagnostics. There are different kinds of contrast-agent methods for specific imaging purposes, and the technique has been validated to identify cartilage lesions (Braun and Gold 2012). Moreover, novel modifications in MRI imaging have shown potency to even interpret the changes in a physiological context. This kind of information may provide insight into the early stages of cartilage

12

breakdown. Laboratory testing can be used to support osteoarthritis diagnosis, but it is important to remember that tests rather reveal inflammation status of tissue than the actual progression of the injury.

The progression of osteoarthritis can be described using a five-stage arthroscopic grading system that is based on International Cartilage Repair Society (ICRS) guidelines (Brittberg and Winalski 2003). Normal healthy articular cartilage is classified as grade 0. Grades 1 to 4 represent cartilage injuries ranging from superficial wears to severe and completely damaged cartilage tissue. The size and the thickness of the damage are the key factors that determine the grade of the injury. Furthermore, the location of cartilage damage and how the damage affects the function of the joint influence the stage of osteoarthritis.

1.5 Current osteoarthritis therapies

Currently, there is no pharmacology-based therapy that can prevent or cure osteoarthritis. Lifestyle changes, such as moderate physical exercise and weight loss, might relieve symptoms. Anti-inflammatory and analgesic drugs are commonly prescribed to patients to relieve pain and inflammation, yet, these do not stop the progression of degenerating cartilage (da Costa et al. 2016). Certain patients get temporary relief by hyaluronan injections, which most commonly are injected to a synovial capsule of a knee (Richette et al. 2015). Hyaluronan injections are used to increase lubrication, but these might have additional beneficial effects since hyaluronan binds to CD44 and could possibly take part in chondroprotective mechanisms (Altman et al. 2015). However, non-surgical methods are usually used to postpone invasive surgical interventions. Early and middle stage cartilage damage can be treated with microfracture surgery and by the use of autologous osteochondral plugs or chondrocyte implantation methods (Falah et al. 2010). More severely or completely damaged joints are usually treated by total joint replacement surgeries (Katz et al. 2010). Although replacement surgeries are highly successful, the method is only applicable for certain joints, the operations are large and rather invasive, and joint replacements have a limited lifetime.

Microfracture surgery and autologous repair methods

Small lesions, less than 2 cm2 of cartilage, can be restored using microfracture surgery (Gudas et al. 2005). This method is based on the body’s own capability to restore injured tissue. The operation consists of making arthroscopically small holes to the underlying bone of the damaged cartilage. Blood transports both -derived stem cells and growth

13

factors to the cartilage, which subsequently induce extracellular matrix production and restoration of cartilage. The feasibility of this method is usually limited since, instead of native hyaline cartilage, a functionally weaker fibrocartilage repair tissue is normally produced during healing (Knutsen et al. 2004). This method might have potential in young patients with minor injuries, but the true evidence of long-term functionality are exiguous (Asik et al. 2008, Mithoefer et al. 2009).

Osteochondral autografts offer a potent way to repair larger cartilage damage than microfracture surgery (Patil and Tapasvi 2015). In this surgical operation, a cylinder-shaped cartilage and bone plug, or several plugs, are taken from a patient’s healthy and less weight-bearing cartilage and transferred to the damaged area. When several plugs are transplanted, the method is called mosaicplasty. Although the cartilage of the plugs is intact, the method has its own challenges. The correct transplantation of several plugs to a specific area to be treated is difficult. This can lead to an uneven positioning, which exposes the plugs to loosening and rapid abrasion. Secondly, the harvesting of the plugs produces permanent damage to a patient’s articular cartilage and, if the treated area is large, there are difficulties to obtain enough material to be harvested and transplanted (Jakob et al. 2002). However, short-term response to mosaicplasty are usually promising. A follow-up study of 19 patients demonstrated that the transplanted osteochondral autografts worked well in the majority of patients (Oztürk et al. 2006). On the other hand, another study using a larger number of patients did not reveal an increase in long-term response than what is achieved with microfracture surgeries (Krych et al. 2012). In a 10-year follow-up study, mosaicplasty operations were observed to have great variability (Solheim et al. 2013). Some operations were successful, but others were not at all. The main reasons for this variability were age, gender and the original size of the damaged cartilage. Altogether, a recent review of osteochondral autograft operations concluded that the method is practical for young patients who have lesions less than 4 cm2 in their articular cartilage (Richter et al. 2016).

Chondral and osteochondral damage can be also treated with an autologous chondrocyte implantation technique (ACI) (Brittberg et al. 1994). This method is based on the use of chondrocyte cells, which are isolated from arthroscopic biopsies taken from the patient’s healthy cartilage in the less weight-bearing areas, and later grown in culture until a number of cells is large enough to be implanted in damaged cartilage. After implantation, the chondrocytes are covered and sealed with a periosteal cover to maintain the cells in the damaged area.

14

Alternatively, matrix-assisted ACI method (MACI) can be used (Kon et al. 2013). In MACI, harvested healthy cells are seeded directly into a matrix, such as collagen. These matrices maintain the cells in the damaged area and work as a structural support for implanted cells. The MACI method reduces the need of operations in comparison to original ACI since it does not require the isolation of an autologous periosteal cover.

These autologous chondrocyte implantation methods can be used to treat damages to cartilage larger than those in which microfracture surgeries or mosaicplasty are suitable. Although both ACI techniques are potentially better to induce cartilage formation than previous methods, they have certain challenges. The synthesis of new cartilage is a slow process and operated joints will take several months to over a year to heal, in addition, these tecniques require rehabilitation (Hambly et al. 2006). Therefore, the true success of operations can only be evaluated after a long-term period of monitoring. ACI-based techniques are expensive and need both invasive operations and a special laboratory for in vitro cell culturing.

Evaluations of the outcomes of ACI operations have generally shown promising results at early- and mid-term follow-ups, which shows suitability of these methods to treat cartilage damages, e.g., in the knee (Ossendorf et al. 2011). However, poor clinical results have also been observed which indicate that more extensive and carefully designed long-term follow-up studies are needed (Filardo et al. 2013, Kon et al. 2013). The predisposition of MACI repair tissue to the onset of osteoarthritis is not known (Jacobi et al. 2011). The absolute comparison of the different surgical-based cartilage repair methods is challenging since methods are technically quite different and they are targeted for different kinds and sizes of lesions.

There has been no success to develop a cartilage-based therapy able to restore all of the features of native hyaline cartilage, such as biologically correct cartilage zones and the correct organization of extracellular matrix components. One of the challenges also relates to poor attachment of repair tissues to the treated area. Individual risk factors make it even more difficult to develop an efficient cartilage therapy suitable for all patients. Therefore, there is still a need for novel therapies that regenerate cartilage lesions and injuries in an efficient way.

1.6 Strategies for novel osteoarthritis therapies

Present strategies to develop novel osteoarthritis treatments mostly rely on improving an existing therapy or in developing a new strategy to treat osteoarthritis and related tissue damages. The strategies are often divided

15

into pharmacological attempts, such as osteoarthritis drug development and various regenerative therapies (Zhang et al. 2016). The group of the regenerative therapies is extensive and includes applications, autologous transplantation techniques and combinations of each with novel biomaterials. A better understanding of the maintenance of articular cartilage and the specific role of microenvironmental factors, such as mechanical forces, a low oxygen atmosphere, growth factors, etc., could offer cell-based cartilage therapies more effectiveness and simultaneously recognize putative biomarkers for osteoarthritis.

The aims of disease-modifying drug therapies for osteoarthritis

There have been a lot of investment to develop disease-modifying drug therepies for osteoarthritis (DMOADs). Unfortunately present approaches have not succeeded to stop the progresion of osteoarthritis. It has been speculated that previous attempts have not taken into consideration all biological factors of articular cartilage and the surrounding tissues, such as cascades, nitric oxide, proteinases, cellular and apoptosis (Qvist et al. 2008). Furthermore, the lack of proper biomarkers impedes the monitoring of the progression of osteoarthritis in test subjects. Although the need for applicable biomarkers has been recognized for a long time, it is still one of the key issues to be solved (Karsdal et al. 2016). The benefits of systemic and local dosing should also be taken into consideration. Some previous pharmacotherapic failures have been related to improperly executed clinical trials (Yu and Hunter 2015, Karsdal et al. 2016). These kinds of failures are expensive setbacks and highlight the importance of careful design of clinical studies to ensure the possibibility to reveal true effects. Moreover, the complexity of osteoarthritis should be recognized and it should be accepted that a single pharmacotherapy will not be effective for all cases and patients. Articular cartilage has a close relationship to subchondral bone. Therefore, it has been suggested that the subchondral bone could also be a potential target of osteoarthritis pharmacotherapy (Castañeda et al. 2012). For example, a potential pharmacotherapy could be developed to inhibit both osteoclast activity and induce bone vascularization (Tonge et al. 2014). The current DMOAD strategies and studies have been primarily focused on molecules such as MMP inhibitors, aggrecanase inhibitors (ADAMTS inhibitors), cytokine inhibitors, bisphosphonates, calcitonin, cathepsin K inhibitors, strontium ranelate, bone morphogenetic protein 7 (BMP-7), different growth factors and enzymes that can inhibit the production of nitric oxide (Qvist et al. 2008, Tonge et al. 2014, Wang et al. 2015). The underlying purpose of these strategies is to maintain anabolic responses and/or reduce any cartilage degrading catabolic activity with a

16

pharmacologically active molecule. These goals have proven to be difficult to reach in practice, and DMOADs are still in experimental research use.

The studies of novel regenerative therapies for osteoarthritis

There are numerous ongoing studies relating to regenerative therapies for osteoarthritis. Their main goal is to discover a cell-based solution, one that would produce long-lasting, native-like repair tissue to replace damaged articular cartilage. Regenerative cartilage therapies can be based on the use of matrices and scaffolds, but there are also scaffold-free solutions. Scaffold- free applications might enhance proteoglycan production, but they usually have certain mechanical weaknesses (Duarte Campos et al. 2012). Matrices and scaffolds can consist of both biological and synthetically manufactured materials. Common natural materials used are collagens, fibrin, and carbohydrate-based compounds, such as agarose, alginate, chitosan and hyaluronan. Polyethylene glycol, N-isopropylacrylamide, polycaprolactone, polyvinyl alcohol, polylactic acid and its’ derivatives are examples of synthetic materials studied in cartilage tissue engineering (Duarte Campos et al. 2012, Gentile et al. 2014). It is also possible to combine certain natural and synthetic components to obtain hybrid materials. Material sciences closely relate to tissue engineering, and it is suggested that novel nanomaterials, such as specific carbon nanotubes, could be applicable for efficient scaffold manufacturing (Antonioli et al. 2013). It is noteworthy that nanotechnology is not only limited to structural applications, such as scaffolds, but it might also offer a novel way to target drug delivery for treatment of osteoarthritis and bone diseases (Kang and Im 2014). Scaffolds are used as a three-dimensional support for cell growth, cell attachment and migration aimed to facilitate the production of repair tissue (Duarte Campos et al. 2012). An optimal scaffold structure should mimic the features of natural cartilage tissue. One of the challenges relate to the development of scaffold material, which would support cartilage structure until chondrocytes can replace the scaffold with its own cartilage repair tissue. Thereby, scaffolds should be made of safe and biologically degradable materials. Although most of the matrices and scaffolds are designed to be used with cells, different kinds of cell-free scaffold systems have also been tested to treat damaged cartilage (Siclari et al. 2012, Di Martino et al. 2015). The true long-term results of these studies are still awaited.

ACI technologies belong to regenerative therapies and they have been used for over 20 years. During the past two decades the method has improved. Instead of ACI, the MACI methods are currently used to avoid periosteal delamination and to reduce hypertrophic differentiation and the number of operations (Zhang et al. 2016). Although present ACI techniques appear to

17

work relatively well for a large amount of patients, there is still need for improvement. A limited number of autologous chondrocytes restrict this type of treatment. Arthroscopic biopsies cause damage to cartilage surfaces; moreover, the chondrocytes tend to lose their phenotype during in vitro expansion process. Novel cell-based strategies are considered to minimize both the number of invasive surgical operations and the damage of less weight-bearing areas during the collection of chondrocytes. Therefore, other types of cell sources are actively studied.

Stem cells might solve many of the problems relating to the optimal cell sources. Mesenchymal, embryonic and induced pluripotent stem cells are considered potential cells for chondrogenic differentiation (Zhang et al. 2016). Embryonic stem cells could be an ideal cell source for cell-based therapies since they have a wide capability to differentiate into many somatic cell types and since they are not limited in their ability to proliferate. Both scaffold-based and scaffold-free studies using embryonic stem cells have been performed, but there are still several concerns that limit their medical use. Today it is not known which is the best way to selectively differentiate embryonic stem cells into chondrocytes and there is a risk of teratoma formation. Furthermore, the source of blastocysts causes ethical questions. The use of induced pluripotent stem cells offer certain source- and risk- related benefits in comparison to embryonic stem cells, but the controlled dedifferentiation of somatic cells to the induced pluripotent stage and further redifferentiation to the chondrocytes is complex. The reprogramming processes require virus-mediated gene transfers and/or complex cocktails of different reprogramming factors, such as proteins, peptides, mRNAs, miRNA and small molecules (Singh et al. 2015, Han et al. 2016). These strategies are fascinating and probably achievable at some stage, but currently their use is focused on experimental research instead of safe and practical human studies. Furthermore, the risk of tumorigenesis also relates to the induced pluripotent stem cells since tumorigenic genes of these cells are not properly known (Zhang et al. 2012).

Thereby, more realistic expectations from stem cell-based therapies rely on the use of mesenchymal stem cells, such as bone marrow- or - derived stem cells. It has been demonstrated that these can be used with or without different types of scaffolds to treat mild osteoarthritis in animal models (Cui et al. 2009, Toghraie et al. 2012). Besides their use with scaffolds, direct and simple intra-articular injections of mesenchymal stem cells and their combinations with synthetic extracellular matrices have been also tested (Liu et al. 2006, Jo et al. 2014). injections have revealed certain improvement in joint functions in addition to pain relief and, in some cases, even hyaline-like cartilage formation (Koh et al.

18

2013, Jo et al. 2014). The influence of certain carrier media, such as platelet- rich plasma, has also been studied for stem cell injection therapies. This multifactorial component has been shown to promote type II collagen synthesis, suppress apoptosis in chondrocytes and improve cell integration to surrounding tissue and, thereby should be considered for additional trials (Mifune et al. 2013). However, in one of the most recent reviews, it was criticized that the outcome evaluation of injection studies were not long-term and pointed to the need for true evidence of long-term efficacy of stem cell injections to treat osteoarthritis (Freitag et al. 2016). Altogether, a large amount of both preclinical and clinical human studies using mesenchymal stem cells have been performed to treat osteoarthritis. Many results have been promising, but certain drawbacks, such as the hypertrophic differentiation of chondrocytes have also been observed. Therefore, comprehensive long-term clinical trials are still needed until all the advantages, disadvantages, and the safety of mesenchymal stem-based therapies will be known.

1.7 Mechanical stresses in articular cartilage

The weight-bearing function and movement of joints expose chondrocytes to increased hydrostatic pressure and mechanical stress, such as compression and stretching. These mechanical factors induce secondary responses that include fluid flow in cartilage tissue. Normal osmotic pressure produces a constant hydrostatic pressure in articular cartilage (Urban 1994). The mechanical forces cause structural tension to cartilage tissue while the porous and permeable extracellular matrix restricts the rapid flow of the fluid phase (water-based interstitial fluid) from the tissue (Sophia Fox et al. 2009). When a mechanical load is removed, the fluid phase returns to the previously unloaded stage. This mechanism enhances the structural resilience of cartilage tissue. Generally, the viscoelastic mechanisms of articular cartilage can be divided into to flow-dependent and flow- independent mechanisms. Flow-independent mechanisms are based on the viscoelastic nature of collagen and proteoglycan networks.

Hydrostatic pressure

Hydrostatic pressure is considered to be one of the most important physical factors together with mechanical compression in articular cartilage (Elder and Athanasiou 2009a). Hydrostatic pressure is generated during compression by the resistance of fluid flow in articular cartilage matrix. The effects of hydrostatic pressure on chondrocytes and extracellular matrix have been widely investigated. Models have been used to study the effects of normal physiological loads in addition to higher, non-physiological, loads.

19

Depending on the experiment, both constant and cyclic hydrostatic pressure loads have been applied. The lengths of the experiments have varied from a few minutes to hours and even up to days or weeks. Long-term experiments are usually based on the arrangements, in which a cyclic load is produced during a certain time scale per day and the procedure is repeated for days or even weeks using the same pattern. Due to high experimental variability, the absolute comparison between experimental results is challenging. However, even the low constant hydrostatic load, 0.25 kPa for three weeks induced a proliferation of chondrocytes and revealed the mechanosensitive nature of chondrocytes (Lee et al. 2005). A static and physiologically relevant hydrostatic load (5-15 MPa) can support proteoglycan synthesis in chondrocytes and cartilage explants, while those over 20 MPa have been determined to be rather catabolic (Hall et al. 1991, Lammi et al. 1994). Compression and cyclic hydrostatic loads produce a reversible re- organization of the cytoskeleton, and this mechanism regulates the mechanosensitivity of chondrocytes (Knight et al. 2006). Cyclic and intermittent hydrostatic loads at physiological magnitudes can also control synthesis of the extracellular matrix in cartilage explants and in chondrocytes (Parkkinen et al. 1993, Smith et al. 2000). Intermittent hydrostatic pressures (1, 5 and 10 MPa at 1 Hz) with time intervals of 4 h per day for one or four days enhanced cartilage-specific extracellular matrix production in normal human articular chondrocytes (Ikenoue et al. 2003). Furthermore, long-term cyclic hydrostatic loading patterns supported the development of self-assembled articular cartilage constructs (Hu and Athanasiou 2006, Elder and Athanasiou, 2009b). It is obvious that hydrostatic loading patterns with appropriate time scales can control extracellular matrix production, but the optimal parameters for a full range of extracellular matrix production and tissue assembly are unknown. It has also been noticed that the anabolic responses continue after mechanical loading has stopped (Tatsumura et al. 2013). This supports the hypothesis that the role of hydrostatic pressure in controlling cartilage-specific matrix production is a complex process.

The role of intermittent hydrostatic pressure is not only limited to matrix production in articular cartilage, but it is also connected to the chondrogenic differentiation of cartilage progenitor cells (Li et al. 2016). A cyclic or intermittent hydrostatic pressure, with low or physiological amplitudes, support the maintenance of chondrocyte and cartilage phenotype by increasing, e.g., type II collagen gene expression, proteoglycan synthesis, decreasing the gene expression of the matrix degrading MMP-13 enzyme and inducing other chondroprotective responses that counteract some inflammatory agents (Wong et al. 2003, Lee et al. 2003, Gavénis et al. 2007). Intermittent hydrostatic pressure regimens have proved to enhance the

20

differentiation of mesenchymal stem cells in differentiation supporting conditions, such as in the presence of regulatory factors like TGF-β (Miyanishi et al. 2006, Wagner et al. 2008). The combination of hydrostatic pressure and low concentrations of TGF-β3 enhanced chondrogenesis and additionally down-regulated the expression of Indian hedgehog and type X collagen genes related to chondrocyte hypertrophy and terminal differentiation (Vinardell et al. 2012). In addition to chondrogenic differentiation, hydrostatic pressure can regulate both cellular proliferation and stress fiber formation in mesenchymal stem cells (Zhao et al. 2015). These mechanosensitive mechanisms appear to relate to Ras homolog gene family member A (RhoA) and Ras-related C3 botulinum toxin substrate 1 (Rac) mediated pathways. Previous findings indicate that hydrostatic pressure has many beneficial responses for development and maintenance of cartilage tissue. Optimally, hydrostatic pressure and other supporting factors, such as chondrogenic growth factors, could be used together to enhance controlled extracellular matrix production in cell-based tissue engineering applications.

Mechanical stretching

Mechanical forces mediated across diarthrodial joints tend to compress and change the shape of cartilage tissue. During compression, the normal weight- bearing function of cartilage, both tissue and chondrocytes are exposed to different types of mechanical forces. In addition to compression, related fluid flow increases in hydrostatic pressure and movement of charged molecules, articular chondrocytes are exposed to tensile and shear stresses (Soltz and Ateshian 2000, Jin et al. 2001).

The role of shear stress has been studied using different kinds of technical arrangements. It has been shown to induce both type II collagen and proteoglycan synthesis and to influence chondrocyte (Lane Smith et al. 2000, Jin et al. 2001, Waldman et al. 2003). During mechanical compression of joints, the form and volume of chondrocytes are stretched towards an oval shaped morphology. These structural changes can be visually detected using confocal microscopy (Guilak 1995). Visualization showed that stretching does not only change the morphology of cells, but also the shape of their nuclei. The precise mechanism, how the stretch affects cells and their responses, is not completely known. However, it has been suggested that chondrocytes can sense compression as stretching force since compression leads to stretching of cellular membranes (Lewis et al. 2011). In addition to mechanical compression, hypo-osmotic shocks increase volume of chondrocytes and cause cellular stretching of chondrocyte plasma membranes.

21

Changes in the ionic composition can also cause a shrinking of the cells. Thereby, the chondrocytes require ion channels and aquaporins to actively maintain their cellular volumes under different conditions (Lewis et al. 2011, Barrett-Jolley et al. 2010). Moreover, ion channel function can modulate proliferation, chondrogenic differentiation and cell death associated signals (Wu and Chen 2000, Kurita et al. 2015). Cyclic mechanical stretch was shown to induce the expression of large conductance potassium, BK channels in chondrocytes, while the expression of transient receptor potential vanilloid 4 (TRPV4) channels that are known to mediate metabolic responses in chondrocytes was not affected (Hdud et al. 2014). The normal function of TRPV4 channels is required for osmotransduction-related responses and the channels have an important role in the maintenance of healthy joints (Clark et al. 2010). Calcium signaling via TRPV4 channels can enhance the chondrogenic differentiation of chondroprogenitor cells (Muramatsu et al. 2007). The precise role of stretch-sensitive BK channels is not known in chondrocytes. However, BK channels are known to allow potassium ion efflux to decrease intracellular osmotic pressure and potentially participate in regulatory volume decrease (RVD) (Mobasheri et al. 2010).

In addition to ion channels, chondrocytes also have other mechanosensitive mechanisms that can mediate and activate cellular responses. Primary cilia are non-motile cellular structures that can mediate ATP-induced Ca2+ signaling in mechanically stressed chondrocytes (Wann et al. 2012). Although they are ubiquitously expressed in eukaryotic cells, the cilia- mediated control of ATP reception was first recognized in compressed chondrocytes. Later, it was recognized that cyclic stretching activates cilia- mediated hedgehog signaling in chondrocytes in two-dimensional (Thompson et al. 2014). However, an increase in mechanical strain from a 10% to 20% magnitude (stretching with 0.33 Hz frequency) led to a reduction in cilia length and a further inhibition of hedgehog signaling and ADAMTS-5 synthesis. These mechanosensitive responses were shown to occur due to the decreased activity of histone deacetylase 6 (HDAC6), the major tubulin deacetylase. This mechanism is suggested to act as a potential chondroprotective response, which limits cartilage degrading enzymatic activity under mechanically stressful conditions.

Integrins are heterodimeric transmembrane proteins that mediate both cell- cell and cell-extracellular matrix interactions (Loeser 2014). Normal adult chondrocytes primarily express seven different integrins, namely α1β1, α3β1,

α5β1, α10β1, αVβ1, αVβ3, and αVβ5. The expression pattern of integrins changes in osteoarthritic tissue. Cyclical pressure-induced strain can induce

22

membrane hyperpolarization in chondrocytes and signal transduction is mediated by the mechanosensitive type α5β1 integrin (Wright et al. 1997).

Previous studies demonstrated that mechanical tension can induce changes in type II collagen expression, proteoglycan aggrecan synthesis, MMP synthesis and cartilage related glycosaminoglycans, such as chondroitin-6- sulfate, hyaluronan and dermatan sulfate in chondrocytes (Carvalho et al. 1995, Honda et al. 2000, Yu et al. 2015). Many different kinds of technical arrangements and varying origins of chondrocytes have been used to study the effects of mechanical loading. In some studies, the synthesis of cartilage matrix components, type II collagen and aggrecan, have increased, while the opposite has also been described even including catabolic responses (De Witt et al. 1984, Fukuda et al. 1997, Fujisawa et al. 1999, Huang et al. 2007, Ueki et al. 2008, Tanimoto et al. 2010).

Stretching-induced catabolic outcomes are typically associated with increased expression of different MMPs (e.g. MMP-1, MMP-2, MMP-3, MMP-9, and MMP-13), proinflammatory cytokine interleukin-1β (IL-1β), cathepsin B, tumor necrosis factor α (TNF-α) and hyaluronidases (Fujisawa et al. 1999, Honda et al. 2000, Doi et al. 2008, Tanimoto et al. 2010). Moreover, a 7% cyclic stretching with 10 cycles per minute was found to induce nitric oxide (NO) production and a further decrease in proteoglycan synthesis (Matsukawa et al. 2004). Generally, short-term (less than 24 hours) stretching experiments with less than a 10% tension have shown greater potential for an increased production of extracellular matrix components compared to high frequency and high magnitude stretching (De Witt et al. 1984, Fukuda et al. 1997, Zhou et al. 2007). A constant 3% tension with 0.25 Hz frequency for 4-24 hours significantly inhibited IL-1β mediated pro-inflammatory gene expression, including MMP-9 and MMP-13, inducible NO synthase, cyclooxygenase-2 (COX-2), and potentially supported anabolic responses in articular chondrocytes (Madhavan et al. 2006). Both the magnitude and frequency of stretching can modify responses in chondrocytes (Ueki et al. 2008).

The exact parameters for beneficial stretching strain are difficult to determine since stretching can has been found to induce both anabolic and catabolic responses in chondrocytes at the same time. In addition, even small changes in magnitude or frequency of the strain can induce a downregulation of extracellular matrix-specific genes (Huang et al. 2007, Mawatari et al. 2010). Certain stretch-induced catabolic effects can be suppressed in experimental models through a pre-treatment of IL-4 (10 ng/ml) (Doi et al. 2008). This supports the idea for a combination of mechanical stretching and biologically active small molecules to be used in

23

tissue engineering to enhance anabolic chondrocyte responses. In contrast to hydrostatic pressure, the role of mechanical tension in chondrogenic differentiation has not been widely studied. This might relate to the controversial, or only partially known, role of tension strain to maintain or induce anabolic responses in chondrocytes.

1.8 Oxygen tension in articular cartilage

Articular cartilage is structurally organized into different zones and cartilage tissue is avascular, therefore, synovial fluid is the main supply of oxygen for cartilage tissue. Normal oxygen tension of synovial fluid is estimated to be roughly 10% in human knee joints (Zhou et al. 2004). Within the cartilage tissue itself, it has been estimated to be in the range from 1% to 10% and is dependent upon the oxygen content of synovial fluid (Brighton and Heppenstall 1971). However, the lowest 1% oxygen level is considered less frequent in healthy cartilage tissue than originally thought (Zhou et al. 2004). The actual level of oxygen depends on the zone of articular cartilage and its depth, thereby oxygen levels are highest in the surface layers (6-10%) and smallest in the deep zone (2-3%) (Zhou et al. 2004, Gibson et al. 2008). An estimated average oxygen tension in articular cartilage is close to 5%. Certain inflammatory diseases, such as , are known to decrease oxygen tension in articular cartilage (Lund-Olesen 1970).

Hypoxia influences to chondrocytes

Low oxygen tension in articular cartilage is normal for chondrocytes and they are adapted to this (Grimshaw and Mason 2000). A hypoxic environment supports the maintenance of a normal chondrocyte phenotype and enhances extracellular matrix production in comparison to a normoxic environment (Lafont et al. 2008, Coyle et al. 2009, Ströbel et al. 2010). The hypoxia-induced phenotype maintenance and cartilage matrix production is based on a hypoxia-driven balance of both anabolic and catabolic responses (Thoms et al. 2013). In addition to maintenance and synthesis of cartilage extracellular matrix components, low oxygen tension also induces other responses in chondrocytes. It can induce changes in energy production and metabolism in addition to the expression of certain growth factors and cytokine-based pro-inflammatory mediators (Fermor et al. 2010). Morphologically, hypoxic conditions appear to support the preservation of the typical round morphology of chondrocytes (Gibson et al. 2008). Glycolysis is the main energy production mechanism in chondrocytes and these cells only consume small amounts of oxygen (Lee and Urban 1997). Although hypoxic conditions are natural for chondrocytes, and they can survive even in anoxic conditions for a few days, chondrocytes need oxygen

24

to avoid the anoxia-induced negative Pasteur effect to maintain both glycolysis and extracellular matrix production (Grimshaw and Mason 2000, Lee and Urban 1997, Gibson et al. 2008). A 5% oxygen tension has been suggested as the most optimal condition for chondrocyte energy production since it induces higher ATP levels and lower pAMPK levels compared to normoxic or strongly hypoxic (1%) environments (Fermor et al. 2007, Fermor et al. 2010). In addition, a 5% oxygen tension appears to provide some protection against IL-1 and nitric oxide-induced autophagy. Furthermore, 5% oxygen tension has also reported to be beneficial for cartilage specific extracellular matrix production (Gibson et al. 2008).

In addition to responses of the mature chondrocytes to hypoxia, the role of the hypoxic conditions for the chondrogenesis has been studied. The topic is interesting because hypoxia-driven chondrogenic differentiation may be used for cell-based tissue engineering. Hypoxia can support the chondrogenic differentiation of the mesenchymal stem cells, and PI3K/Akt/FOXO signaling route was suggested to have a role in this differentiation process (Lee et al. 2013, Shang et al. 2014). Hypoxia also supported the chondrogenic differentiation of human embryonic stem cells at certain level (Koay and Athanasiou 2008). However, the applications, which would base purely on hypoxic differentiation are not known.

Hypoxia inducible factors 1α and 2α are the central hypoxia mediators in chondrocytes

Hypoxia-induced responses can be mediated by different mechanisms. Hypoxia inducible factors HIF-1α and HIF-2α have a crucial role in oxygen tension-related homeostasis and cellular responses in chondrocytes (Schipani et al. 2001, Pfander et al. 2006, Lafont et al. 2008). Although these two isoforms share many structural and functional similarities, they have divergent, factor-specific, roles (Hu et al. 2003). There is also a third member of HIF-α family: the HIF-3α, but this isoform does not contain a C- terminal transactivation domain similar to HIF-1α and HIF-2α, therefore, it probably is limited in its ability to act as a transcription factor (Hara et al. 2001, Heikkilä et al. 2011). In some cases, HIF-3α has been shown to inhibit the other HIF α-isoforms, but some observations suggest that HIF-3α acts as a transcription factor and increases HIF target gene expressions (Hara et al. 2001, Heikkilä et al. 2011, Zhang et al. 2014, Ravenna et al. 2016). Both HIF- 1α and HIF-2α transcription factors function by forming specific dimers with constitutively expressed aryl hydrocarbon receptor nuclear translocator 1 or 2 (HIF-β1 and HIF-β2) and induce the nuclear hypoxia response elements

25

(HREs) of hypoxia-inducible genes (Wenger and Gassmann 1997). Oxygen regulates the activity of all HIF α-isoforms. Both HIF-1α and HIF-2α transcription factors are normally degraded by proteasomes when oxygen tension is high. Instead, they are not degraded under hypoxic conditions. The oxygen-controlled degradation of HIFs is based on specific oxygen- dependent degradation domains (ODDDs) and iron-dependent von Hippel- Lindau tumor suppressor (pVHL) E3 ligase complex-induced ubiquitylation (Huang et al. 1998, Maxwell et al. 1999, Jaakkola et al. 2001). The interaction of pVHL and the oxygen-dependent domains is enzymatically regulated by HIF-α prolyl-hydroxylase, which can induce targeting of specific hydroxylation of the HIF-α subunit. Due to the lack of another hydroxylatable proline on the ODDD-domain, the normal normoxic degradation of different variants of HIF-3α is suggested to be inefficient (Ravenna et al. 2016).

In chondrocytes, the role of HIF-1α has been primarily connected to mediating inhibitory actions against catabolic activities, while HIF-2α is associated with Sox9-regulated synthesis of cartilage components (Thoms et al. 2013). Transcription factor HIF-1α maintains cell survival of differentiated cells under hypoxic conditions. Moreover, it has a role in chondrocyte proliferation, differentiation and growth arrest. Hypoxic conditions increase vascular endothelial growth factor A (VEGF-A) expression in a HIF-1α-mediated manner in chondrocytes and chondrosarcoma cells (Lin et al. 2004).

Hypoxia-induced up-regulation of extracellular matrix gene expression involves HIF-2α activity. This up-regulation is based both on Sox9- dependent and -independent mechanisms (Lafont et al. 2008, Thoms et al. 2013). HIF-2α-induced Sox9-independent factors participate to define the differentiated chondrocyte phenotype, i.e., mitogen-inducible gene 6 protein (Mig6; also known as ERBB receptor feedback inhibitor 1) and inhibin beta A chain (InhbA). Although many studies have recognized a hypoxia-induced anabolic role of HIF-2α in chondrocytes, this transcription factor appears to have more complex and context-related role than previously thought. It has been observed to induce both catabolic genes, such as different MMPs, aggrecanase-1 and nitric oxide synthase-2, and to induce Fas-mediated apoptosis of chondrocytes during the osteoarthritic process (Yang et al. 2010, Ryu et al. 2012). In addition, interleukin-6 has a central role in mediating the catabolic activity of HIF-2α in osteoarthritic degeneration (Ryu et al. 2011). Moreover, the overexpression of HIF-2α in joint tissues has been observed and relate to the development of rheumatoid arthritis kinds of changes (Ryu et al. 2014).

26

Hypoxia can also up-regulate HIF-3α gene expression and it has been connected to chondrocyte redifferentiation, although the absolute roles of various HIF-3α splice variants in chondrocyte biology are not known (Schrobback et al. 2012). One of the most recent studies suggested that HIF- 3α could have a role in the regulation of chondrocyte hypertrophic terminal differentiation, but further studies are needed to fully understand this mechanism (Markway et al. 2015).

1.9 Ras homolog gene family member A and Rho- associated protein kinase signaling

Rho-associated protein kinases (ROCKs) belong to the AGC family of serine- threonine kinases (Pearce et al. 2010). The two isoforms of ROCKs, namely ROCK1 and ROCK2, are synthesized in mammals and they share 65% identity (Nakagawa et al. 1996). Rho kinases are key downstream mediators for RhoA (Ras homolog gene family, member A; Figure 2) (Riento and Ridley 2003). Although ROCKs have an important role in RhoA signaling, RhoA also has other downstream mediators, such as the mammalian diaphanous formin (mDia), protein kinase N, focal adhesion kinase, Citron kinase, phospholipase D1, myosin light chain and Rhophilin (Amano et al. 1996, Kimura et al. 1996, Madaule et al. 1998, Cai et al. 2001, Kloc et al. 2014). Small GTPases are enzymes that bind and hydrolyze guanosine triphosphate molecules (GTPs). The small Rho family GTPases RhoA, Cdc42 and Rac1 are well-known factors that mediate different kinds of actin polymerization and reorganization signaling in cells (Sit and Manser 2011). Various factors, such as G-protein connected , growth factors, hormones and receptors, regulate the activity of RhoA and induce further cellular responses. Like normal G-proteins, the active form of RhoA is RhoA-GTP and its activity is controlled by reciprocally affecting guanine nucleotide exchange factors and GTPase activating factors (GAPs) (Garcia-Mata et al. 2011). Guanine nucleotide exchange factors catalyze a release of guanosine diphosphate (GDP) and allow the replacement of GDP with GTP (Figure 2). Controversially, GAPs induce hydrolysis of GTP to form inactive GDP. A guanosine nucleotide dissociation inhibitor (GDI) is a third regulator and participates in RhoA activation by binding to the GDP form of small GTPases to suppress their activity. Moreover, GDI can affect the cellular localization of small GTPases and also prevent their activity. A previous fibroblast migration study indicated a complex spatio-temporal nature of RhoA signaling route, which is based on crosstalk between Rho GTPases and GTPase activities to regulate specific effector pathways (Martin et al. 2016). However, the rapid kinetics of signaling makes it challenging to decipher all of the components of the pathway and their specific roles.

27

Figure 2. The schematic figure of RhoA signaling has been build up from previously published data of RhoA and ROCK mediated signaling pathways (Watts et al. 2006, Garcia-Mata et al. 2011, Shum et al. 2011 and Kloc et al. 2014).

RhoA/ROCK signaling maintain chondrocyte differentiation

Common RhoA-mediated responses are linked to a reorganization of the cellular cytoskeleton, the formation of actin stress fibers, actin polymerization, cellular proliferation, cytokinesis, phagocytosis, and vesicular transport in addition to the Rac1-coordinated formation lamellipodia (Guilluy et al. 2011, Kloc et al. 2014). ROCKs have a crucial role to maintain cellular morphology, which is defined by the organization of actin cytoskeleton in chondrocytes and other cell types (Woods et al. 2005, Sit and Manser 2011). Mechanical strains, such as hydrostatic pressure, compression and osmotic pressure, are known to induce a rapid remodeling of actin cytoskeleton in chondrocytes and this reorganization is linked to changes in gene expression (Parkkinen et al. 1995, Chao et al. 2006, Knight et al. 2006, Haudenschild et al. 2008). Actin remodeling occurs via a ROCK downstream effector, LIM kinase, which phosphorylates cofilin (Maekawa et al. 1999). The active and phosphorylated cofilin subsequently depolymerizes actin filaments. RhoA/ROCK signaling is known to suppress hypertrophic chondrocyte differentiation and a selective inhibition of this pathway induces the expression of type X collagen (Wang et al. 2004). In chondrogenic ATDC5 cells, the overexpression of Rac1 and Cdc42 induced chondrocyte differentiation and expression of hypertrophic markers (Wang and Beier 2005). There is an antagonistic signaling cascade between RhoA/ROCK and Rac1/Cdc42/p38 modules. Inhibition of RhoA/ROCK

28

signaling induces the expression of hypertrophic markers, but it also affects actin organization and increases gene expression of Sox9 transcription factor, thereby inducing the expression of cartilage-specific extracellular matrix genes in primary monolayer cultures (Woods et al. 2005, Woods and Beier 2006). The selective ROCK inhibitor Y-27632 inhibited RhoA/ROCK signaling and conserved a chondrocytic phenotype of cultured articular chondrocytes (Matsumoto et al. 2012, Furumatsu et al. 2014). This molecule inhibits both ROCK1 and ROCK2. Importantly, the Ki values for Y-27632 are at least ten times less for ROCK1 and 2 than unspecific kinase targets (Ishizaki et al. 2000).

RhoA/ROCK signaling has a central role in managing chondrogenesis and conserving a chondrocyte phenotype. Osteoarthritis is known to increase the expression of TGF-α and this has been linked to both RhoA/ROCK and MEK/ERK signaling (Appleton et al. 2010). These signaling routes can be considered potential targets to suppress osteoarthritic cartilage degradation. The loss of RhoA activity and related depolymerization of actin filament structures can induce transcriptional activities. Actin depolymerization induces Sox9 gene expression via protein kinase A (PKA) signaling (Kumar and Lassar 2009). In addition to PKA, ROCKs can directly phosphorylate Sox9 in a LIM kinase-independent way at Ser181 and enhance the transcriptional activity of Sox9 (Kumar and Lassar 2009, Haudenschild et al. 2010). It has been suggested that Sox9 may take part in its own gene expression regulation in differentiated chondrocytes by a positive autoregulatory loop mechanism. Moreover, it is suggested that ROCK- and mDia-mediated RhoA signaling inhibits the activity of this autoregulatory loop of Sox9 gene expression. Thereby, the inhibition of RhoA signaling can induce Sox9-dependent chondrocyte differentiation and cartilage matrix production. Sox9 is the main transcription factor for cartilage extracellular matrix-specific genes, such as type II collagen and aggrecan (Bell et al. 1997, Lefebvre and Dvir-Ginzberg 2017). The Sox9 transcription factor has two co- activators, namely Sox5 and Sox6. This trio supports chondrogenesis and also induces the expression of S100A1 and S100B, which appear to suppress the terminal differentiation of chondrocytes (Saito et al. 2007). Previously, S100A1 and S100B expression patterns have been suggested as potential markers of human articular chondrocyte differentiation status (Diaz-Romero et al. 2014).

29

2. Aims of the thesis

Currently, no pharmacological-based thearapy is available to cure cartilage disorders. Although minor damages can be treated using microfracture surgery and autologous cellular repair techniques, their ability to provide native hyaline-like repair tissue is limited. Endoprosthesis surgeries are reserved to treat severe and irreparable damage. Novel regenerative medicine techniques are actively developed to improve currently available treatment options. These applications require both a comprehensive understanding of chondrogenic differentiation and the ability to maintain a chondrocyte phenotype in culture. Mechanical strains and low oxygen conditions both belong to the normal environment of chondrocytes. RhoA/ROCK signaling has been demonstrated to play a crucial role in chondrocyte differentiation status.

The aims of this thesis were to study the following topics

1. Mechanical stretching induced phosphoproteomic responses in chondrocytic cells (paper I)

2. The effects of cartilage matrix supporting Rho-kinase inhibitor Y-27632 in fibroblast cells (paper II)

3. Long-term hypoxia induced protein responses in chondrocytic cells (paper III)

4. The long-term maintenance of chondrocytic phenotype in hypoxic and Rho-kinase inhibited conditions (paper IV)

30

3. Materials and methods

Human chondrosarcoma 2/8 cells (HCS-2/8) and human foreskin fibroblasts were used in this thesis. HCS-2/8 cells were originally isolated and characterized in Professor Masahiru Takigawa’s research group in Okayama University (Takigawa et al. 1989). Fibroblasts (HFF-1, SCRC- 1041™) were purchased from ATCC®. HCS-2/8 cells were used in studies I, III and IV and human fibroblasts in study number II.

3.1 Cyclic cell stretching (paper I)

The cyclic cell stretching experiments were performed using a High Strain Cell Stretching Apparatus (Figure 3, Neidlinger-Wilke et al. 1994). The HCS- 2/8 cells were grown on silicone stretching plates (Elastosil 601-Me silicone; Wacker Chemie) coated with 10 µg/ml poly-D-lysine for 72 hours before the start of the experiment in order to allow sufficient adhesion of the cells. Control cells were cultured on similar plates, but not exposed to stretching.

Figure 3. High Strain Cell Stretching Apparatus.

Cyclic biaxial 8% surface stretch was generated at a frequency of 1 Hz. The stretching periods were either 2 hours for one day or 2 hours per day for a total of three days. The stretching strain was modified from the previously published study (Karjalainen et al. 2003). After completion of the stretching experiments, the cells were harvested from the silicone stretching plates. HEPES buffer (10 mM) was used for washing of the cell samples before phosphoproteomic analysis. Chondrosarcoma cells were cultured in HyQ®

31

MEM Alpha modification (including L-glutamine, ribonucleosides and deoxyribonucleosides) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/ml streptomycin. Before the start of the stretching experiments, the medium was removed and replaced by HyQ® MEM Alpha modification supplemented with 1% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin. Culturing and stretching experiments were o performed in a humidified normoxic (5% CO2) environment at 37 C.

3.2 Hypoxic cell cultures (paper III, IV)

Hypoxic cell cultures were maintained in a humidified 5% oxygen atmosphere using a Galaxy R incubator (RS Biotech Laboratory Equipment Ltd.). The 5% hypoxic environment was used to model the average oxygen strain in normal articular cartilage tissue (Zhou et al. 2004, Gibson et al. 2008). Fresh culture media was equilibrated to 5% oxygen conditions for 24 hours prior to its addition to hypoxic cultures. Culturing periods for chondrosarcoma cells in hypoxic environment were 2 days, 7 days and 28 days. HCS-2/8 cells were cultured in high glucose (4.5 g/l) DMEM with 10% FBS, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM L-glutamine and 50 µg/ml 2-phospho-L-ascorbic acid trisodium salt. Temperature was maintained at 37oC.

3.3 Rho-kinase inhibitor Y-27632 treatments (paper II, IV)

RhoA/ROCK signaling was inhibited with the selective Rho-kinase inhibitor Y-27632. Solid Y-27632 was dissolved in phosphate-buffered saline (PBS) solution and sterile-filtered before its use. The final concentration of Y- 27632 in culture media was 10 µM for all ROCK inhibition experiments. This final concentration was chosen based of previous studies (Woods and Beier 2006, Matsumoto et al. 2012, Furumatsu et al. 2014). Moreover, a 10 µM Y- 27632 concentration was reported to result in the maximal inhibition of ROCKs with a minimal loss of kinase specificity (Sahai et al. 1999, Zhou et al. 2010). Fresh medium was changed on every third day in culture. Human fibroblast cells were grown in a low glucose (1.0 g/l) DMEM supplemented with 10% FBS, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM L- glutamine, and 50 µg/ml 2-phospho-L-ascorbic acid trisodium salt. Human chondrosarcoma cells were cultured in high glucose (4.5 g/l) DMEM containing identical supplements to the low glucose medium used to cultre fibroblasts. Cells were cultured either in a humidified incubator at normoxia or at a 5% oxygen strain at 37°C temperature. Fibroblasts were cultured from 1 to 3 days in Rho-kinase inhibited conditions and HCS-2/8 cells 28 days.

32

3.4 Phosphoprotein enrichment

Cells were lysed using the lysis buffer provided in the Pro-Q Diamond Phosphoprotein Enrichment Kit (Molecular Probes) (paper I) and protein concentrations were determined using a Bradford assay (Bradford 1976). A total of one milligram per sample of total protein was applied to enrichment columns and phosphoprotein enrichment was performed according to the manufacturer’s instructions. The eluted phosphoprotein fractions were pooled and concentrated with Vivaspin® filtration concentrators.

3.5 Two-dimensional gel electrophoresis and mass spectrometric protein identification

Concentrated phosphoprotein samples were purified using the methanol- chloroform precipitation method (Wessel and Flügge 1984). Dried protein pellets were dissolved in rehydration buffer (8 M urea, 2% immobilized pH gradient (IPG) buffer 3-10 NL, 0.5% Triton-X 100, 0.5% Biolyte 3-10, 0.2% DTT and Bromophenol Blue stain). When the rehydrated proteins were absorbed by the IPG strip gel material, isoelectric focusing was performed at a total of 17,000 Volt-hours. Isoelectrically focused samples were reduced by equilibration solution I [6 M urea, 50 mg/ml dithiothreitol, 0.06 M Tris-HCl (pH 6.8), 10% glycerol, 2% SDS, Bromophenol Blue stain] and alkylated by equilibration solution II [6 M urea, 45 mg/ml iodoacetamide, 0.06 M Tris- HCl (pH 6.8), 10% glycerol, 2% SDS, Bromophenol Blue stain] prior to second dimension electrophoretic protein separation on 12% SDS- polyacrylamide gels (PAGE). Both Coomassie Brilliant Blue (CBB) and silver nitrate staining were performed on 2D SDS-PAGE gels (Shevchenko et al. 1996). The CBB-stained PAGE gels were scanned and protein abundancies of selected spots were compared using ImageJ software (National Institute of Health). Protein spots of interest were excised from 2D gels and in-gel digested using 1 µg/µl mass spectrometry grade trypsin (Promega) with slight modifications to a previously published protocol (Shevchenko et al. 2007). The digested peptides were freeze-dried and reconstituted in 0.1% trifluoroacetic acid (TFA). The digested peptides were freeze-dried and reconstituted in 0.1% trifluoroacetic acid (TFA).

Ultimate/Famos capillary LC system (LC Packings) and reverse-phase column (PepMap C18 precolumn, 300 mm i.d. 1 mm; Dionex) was used for desalting and purification of trypsin-digested peptide samples. The peptide separation was made using PepMap C18 analytical column (3 µm, 75 µm i.d., 50 mm; Dionex), and the eluation was carried out using a linear gradient of acetonitrile. This linear gradient of acetonitrile started from 0.1% formic acid/5% acetonitrile rising to 30% of 0.1% formic acid/95% acetonitrile over

33

35 minutes. Finally, the gradient was increased to 100% of 0.1% formic acid/95% acetonitrile for 40 minutes. The eluent flow rate was set to 200 nl/min, and the total time for one run was 80 minutes. The LC system was connected to a nano-ESI ion source (MDS Sciex) of the QSTAR XL hybrid quadrupole TOF instrument (Applied Biosystems). The mass spectra were recorded in a positive ion mode using information dependent acquisition (IDA) method. Tandem mass spectrometry data (MS/MS) was obtained for sequence tag-approach (Mann and Wilm 1994). For each run, one second time-of-flight (TOF) MS survey scans were recorded for mass range m/z 400–2000 followed by a four second MS/MS scans for the two most intense 2+ and 3+ charged ions. The voltage of the electrospray needle voltage was set to 2.3-2.4 kV. Nitrogen was used as the curtain and collision gas. Monoisotopic peaks of 2+ and 3+ charged ions were calibrated with trypsin- digested myoglobin and adrenocorticotropic hormone clip. The spectral data was automatically processed with Analyst® QS software (Applied Biosystems) and database searches were performed manually using MASCOTSearch script (v1.6b13) on-line at www.matrixscience.com (Perkins et al. 1999). The reliability of the mass spectral protein identifications were estimated with Mascot score parameter, which was automatically calculated from the mass values by Mascot search algorithms. The sequence coverages of the identified proteins varied from 7 to 44%. Some protein identifications were performed with MALDI-TOF mass spectrometry as a commercial service by Alphalyse A/S.

3.6 Western blot analyses

Harvested cells were washed with PBS and lysed in RIPA buffer (1x PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 0.3% aprotinin, 100 µg/ml PMSF, 1 mM sodium orthovanadate) and protein concentrations were determined using a Bradford assay (Bradford 1976) (paper I). Equal amounts of each protein sample (15 µg) were denatured by boiling for 5 minutes with electrophoresis sample buffer and then loaded on 10% SDS-PAGE gels. Electrophoretically separated proteins were electrotransferred to Protran® Nitrocellulose Transfer Membranes (Whatman) and the membranes were blocked with either skimmed milk or bovine serum albumin. Blocked membranes were incubated for 1 hour at room temperature, or overnight at 4ºC, with specific primary antibody (see Table 1 below). Glyceraldehyde 3- phosphate dehydrogenase (GAPDH) was used as loading control. Horseradish peroxidase (HRP)-labeled secondary antibodies for primary antibody detection were manufactured by Santa Cruz Biotechnology. The membranes were incubated with secondary antibody for 1 hour at room temperature and developed after washing steps with SuperSignal West Pico Chemiluminescent substrates (Pierce). CL-XPosure films (Pierce) were used

34

for detection and densitometric analyses of the proteins were performed using ImageJ software (National Institute of Health).

Table 1. The primary antibodies for Western blot analysis.

Antibody Dilution Manufacturer anti-MAP kinase (ERK1 + ERK2) 1:1000 Zymed eEF1D (1-90) 1:1000 Abnova EF-2 (C-14) 1:4000 Santa Cruz Biotechnology GAPDH (horseradish peroxidase conjugated) 1:5000 Abcam p-Akt1 (Ser-473) 1:200 Santa Cruz Biotechnology p-PI 3-kinase p85α (Tyr-508) 1:250 Santa Cruz Biotechnology p-ERK (E-4) 1:1000 Santa Cruz Biotechnology

3.7 Time-lapse cell monitoring

Time-lapse cell monitoring was performed with the Cell-IQ model v. 2 microscope system (CM Technologies) both for Y-27632-exposed fibroblasts (paper II) and Y-27632 and/or hypoxia-exposed HCS-2/8 cells (papers III, IV). Cell monitoring was based on automated phase contrast microscopy using a tenfold magnification. The width of one monitoring grid was 974 µm with a height of 732 µm. This monitoring technique was applied to observe changes in cellular shapes, the lengths of actin filament-based filopodia and structures, and cellular proliferation. Additionally, monitoring was also used to evaluate scratch-wound assays and to quantify cellular migration. Human primary fibroblasts were seeded at a density of 50,000 cells and the slowly- dividing chondrosarcoma cells at 200,000 cells per a well in 6-well-plates. Cells were imaged every 15 minutes with a total monitoring time of 72 hours for cellular morphology and proliferation analysis. The scratch wound assays were performed when monolayer cultures reached at least 90% confluency. Cross-directional wound scratches were made to monolayer cultures by the use of a narrow pipette tip. Monitoring intervals were every 15 minutes for scratch-wound assays with a total length of 36 hours for fibroblasts and 60 hours for HCS-2/8 cells. When Rho-kinase inhibitor Y-27632 exposure was studied in fibroblasts and HCS-2/8 cells, a 10 µM final concentration was used and the medium was not changed during any time-lapse monitoring experiment. The normoxic atmosphere was studied with both fibroblasts and HCS-2/8 cells, while a hypoxic atmosphere was only evaluated with HCS-

2/8 cells. The normoxia atmosphere (5% CO2, 19% O2 and 76% N2) and the hypoxic atmosphere (5% CO2, 5% O2 and 90% N2) were maintained with premixed gases (Aga). The temperature was kept at 37°C at all times. The

35

data analysis for time-lapse imaging was performed using the Cell-IQ analyser 4 Pro-write software (AN4.2.1 HW, CM Technologies). The automatic cell recognition-related random errors of absolute cell number determinations were corrected manually. The data from scratch-wound assays were analyzed manually using the Cell-IQ analyser 4 Pro-write software.

3.8 Immunofluorescence imaging

Cells were fixed with 4% w/v paraformaldehyde for immunofluorescence imaging (paper II). Fixed cells were washed with PBS and permeabilized with 0.1% Triton-X 100. Permeabilized cells were washed with PBS and incubated with the murine monoclonal anti-vinculin (1:1000; Sigma- Aldrich), the mouse monoclonal anti-integrins α2 (1:200, Millipore), β1 (1:200; Millipore), α5β3 (1:200; Millipore) primary antibodies for 2 hours. The fluorescein isothiocyanatelabeled goat anti-mouse secondary antibody (1:200; Chemicon) was incubated for 1 hour. Actin filaments were stained with Alexa Fluor 594-conjugated phalloidin (Molecular Probes) for 30 minutes. The nuclei were stained with 4',6-diamidino-2-phenylindole (DAPI) for 20 minutes at 37oC. Cells and immunolabeled cell structures were visualized using an Olympus IX-70 fluorescence microscope.

3.9 Mass spectrometric quantification with Orbitrap Elite

Proteins were extracted by lysing the cells in 200 µl of lysis buffer (100 mM ammonium bicarbonate, 8 M urea, 0.1% RapiGest) and sonication for 5 minutes (paper II). Protein concentrations were determined using a BCA assay (Thermo Fischer Scientific). Proteins were reduced with dithiothreitol and alkylated with iodoacetamide prior to trypsin-digestion in 1 M urea solution. The tryptic peptide digests were quenched using 10% TFA and subsequently concentrated and purified by C18-reverse phase columns (Varjosalo et al. 2013). The eluation buffer was 90% acetonitrile with 0.1% TFA. Dried peptides were reconstituted in a 2% acetonitrile/0.1% TFA solution, separated chromatographically by nLCII nanoflow HPLC system (Thermo Scientific) and analyzed by Orbitrap Elite electron-transfer dissociation mass spectrometer (Thermo Scientific). The mass spectral analyses were performed in a data-dependent acquisition mode using a top 20 CID method. Dynamic exclusion time for the selected ions was 30 s. There were not used lock masses. The maximal allowed ion accumulation time on the Orbitrap Elite electron-transfer dissociation in CID mode was 100 ms for MSn in the ion trap and 200 ms in the Fourier-transform MS. Intact peptides were detected in the Orbitrap at 60,000 resolution. Proteome

36

Discover software (Thermo Scientific) was used for peak extraction and further protein identifications. SEQUEST search engine was used to search the calibrated peak files from UniprotKB/SwissProt database (www.uniprot.org). The error tolerances on precursor and fragment ions were 15 ppm and ±0.6 Da, respectively. One missed cleavage was allowed. The label-free protein quantification was based on a spectral counting method, which assumes that protein abundances correlate with the number of peptide fragmentation events (MacCoss et al. 2003, Liu et al. 2004).

3.10 Mass spectrometric protein quantification with Synapt G2-Si HDMS

For studies III and IV, protein extractions and trypsin-digestions for Synapt G2-Si HDMS analyses were based mainly on previously published procedures (Masuda et al. 2008). After trypsin-digestions, peptides were cleaned-up using C18 STAGE-tips (Rappsilber et al. 2003). Peptide concentrations were determined using a micro BCA kit (Thermo Fischer Scientific) and 600 ng of each sample were loaded on the HSS T3 C18 analytical column (75 μm i.d. × 200 mm, 1.8 μm particles; Waters). The peptides were separated using a linear 108 minute gradient of 5–41% solvent B (3:1 acetonitrile/2-propanol) balanced with 0.1% aqueous formic acid (solvent A) at a flow rate of 295 nl/min. The eluate was analyzed with a nano-ESI equipped Synapt™ G2-Si HDMS mass spectrometer (Waters) operating in resolution mode. All data were collected using ion-mobility MSE with a scan-time of 0.5 sec. and mass-corrected using Glu-fibrinopeptide B and Leucine Enkephalin as reference peptides. Data analysis was performed with ProgenesisQI (Nonlinear Dynamics). Peptide identification was based on built-in MSE search option, which employed a false discovery rate less than 1% with a mass error cut-off of 10 ppm. Two missed cleavages were allowed and carbamidomethylated cysteines were used as fixed modifications. Methionine oxidation and asparagine and glutamine deamidation were set as variable modifications. The relative protein abundances were calculated with the Hi-N method of the ProgenesisQI software.

3.11 Bioinformatic protein analysis

Qualitative proteome annotations (paper III) were performed using DAVID Bioinformatics resources 6.8 (National Institute of Allergy and Infectious Diseases (NIAID), NIH) and Reactome Pathway Database (Huang et al. 2009a, Huang et al. 2009b, Croft et al. 2014, Fabregat et al. 2016). The protein pathways analyses for the mass spectrometric protein screening data (papers II, III, IV) were performed with the Ingenuity Pathway Analysis IPA

37

application (Ingenuity Systems). For pathway analyses, the cut-off value for protein fold changes was set to 1.3 fold for both up- and downregulations. Analysis filter was set to allow experimentally observed relationships, and all tissues and primary cells were selected for the analyses.

3.12 Enzyme linked immunosorbent assay (ELISA) of type II collagen

The concentrations of secreted type II collagen were determined from the culture media using a commercial ELISA Type II Collagen Detection Kit (Chondrex Inc.) (papers III, IV). The medium samples (50 µl) were diluted 1:1 in dilution buffer provided by the ELISA kit. The analysis was performed according to the manufacturer's instructions. Absorbances were determined at 490 nm using a Vmax kinetic microplate reader (Molecular Devices) after a one-hour incubation in o-phenylenediamine solution.

3.13 Dimethylmethylene Blue assay (DMMB) for sulfated glycosaminoglycans

Sulfated glycosaminoglycans (sGAGs) were quantified from culture medium with the dimethylmethylene blue (DMMB) assay (papers III, IV). The assay was modified from the original protocol (Farndale et al. 1986). Culture medium samples (75 µl) were mixed with 250 µl of DMMB reagent in the wells of a 96 well plate. The absorbances were immediately determined at 535 nm using a Vmax kinetic microplate reader. Shark chondroitin sulfate was used as a standard and these samples were diluted in fresh culture medium. The DMMB assay was linear in a range from 0 to 5.77 µg/ml (R2 ≥ 0.97). Background noise from the medium was taken into consideration, and the detection limit of the analyte (LOD) was determined on the basis of the limit of blank (LOB) (Armbruster and Pry 2008). Both five independent blank samples and the low concentration samples with total of 15 replicates were used for LOB and LOD determination. The arbitrary limit of quantification (LOQ) was not stated for this assay. All the compared sGAG concentrations were at least three times higher than the LOD, which was 0.35 µg/ml in this spectrophotometric assay.

3.14 qRT-PCR analysis

The relative gene expression of target genes were analyzed using quantitative polymerase chain reaction (qRT-PCR) analyses (papers II, III, IV). Total RNA was extracted with TRI Reagent, and after the purification steps, 1 µg of each RNA sample was converted into cDNA by using Thermo Scientific Verso SYBR Green 2-Step qRT-PCR Kit. The qRT-PCR reactions were

38

performed either using a LightCycler 480® (Roche Diagnostics) or a CFX Connect Real-Time System (Bio-Rad), and the master mixtures used were either LightCycler 480® SYBR Green I Master mixture (Roche) or SSoAdvanced™ Universal SYBR® Green Supermix (Bio-Rad). The primers (see Table 2) were purchased from Oligomer and Thermo Fisher. The RT- PCR reaction conditions and cycle numbers were optimized for each primer pair. Specific amplification efficiencies for each primer pair were determined from the slopes of standard curves and the efficiencies only within the range of 90–110% were accepted. The relative gene expression results have been calculated using the Pfaffl-method (Pfaffl 2001). The specificities of PCR products were determined by post-PCR melting curve analysis and by agarose gel electrophoresis.

Table 2. The RT-PCR primers. Gene Forward primer (5'→3') Reverse primer (5'→3') ACAN CACTGTTACCGCCACTTCCC AACATCATTCCACTCGCCT Col1A1 CAGCCGCTTCACCTACAGC TTTTGTATTCAATCACTGTCTTGCC Col1A2 GGCAATAGCAGGTTCACGTACA CGATAACAGTCTTGCCCCACTT Col2A1 AAGGTCATGCTGGTCTTGCT GACCCTGTTCACCTTTTCCA RPLP0 AGATGCAGCAGATCCGCAT GTGGTGATACCTAAAGCCTG S100A1 GCTCTGAGCTGGAGACGGCG GCCACTGTGAGAGCAGCCACA S100A16 CAGGGAGATGTCAGACTGCTACAC CATCAGGCCAGTGCCTGGAA S100B CCGAACTGAAGGAGCTCATC AGAACTCGTGGCAGGCAGTA SOX9 GACTTCCGCGACGTGGAC GTTGGGCGGCAGGTACTG VEGFA AGGAGGAGGGCAGAATCATCA CTCGATTGGATGGCAGTAGCT VCAN CAAGCATCCTGTCTCACGAA CAACGGAAGTCATGCTCAAA

3.15 Statistical analysis

One-way analysis of variance (ANOVA) tests were performed for time-lapse imaging data. One-way ANOVA with Fisher’s Least Significant Difference post-hoc test was used for the qRT-PCR data analysis. Student’s t-test was used for the paired data analysis of ELISA and DMMB results. The statistical comparisons between corresponding protein samples and controls were tested with a Student’s t-test (paper II). The statistical significance of the mass spectrometry-based protein quantifications (papers III, IV) were determined from the normalized arcsinh transformed data with one-way ANOVA tests with ProgenesisQI software (Nonlinear Dynamics). The p- values equal or less than 0.05 were considered significant. All experiments were performed at least three times independently.

39

4. Results

4.1 Cyclic stretching induced phosphoproteomic changes in HCS-2/8 cells (paper I)

Human chondrosarcoma 2/8 cells were exposed to a cyclic 8% tension at a 1 Hz frequency for 2 hours per day during a three-day period. Cells were immediately harvested for Western blot analysis after the end of the last cyclic stretch experiment on day three. The stretching strain induced phosphorylation of Extracellular signal-regulated protein kinases 1 and 2 (Erk1/2) (Fig. 1, paper I). The induction of phosphorylated Erk1/2 was observed at all time points evaluated. The phosphorylation states of Akt serine/threonine kinase 1 (p-Akt1) or Phosphoinositide-3-kinase (p-PI3K) were found not to change in this study. When the stretching experiment was performed only once and phosphorylated proteins were screened by proteomics, changes in the expression levels of four proteins were observed (Fig. 3., Table 1, paper I). The intensities of the eukaryotic translation elongation factor 1 delta isoform 2 (eEF1D), moesin and prolyl-3- hydroxylase 1 () protein spots were increased and the intensity of eukaryotic translation elongation factor 1 beta 2 (eEF1B) decreased. These proteins were not previously known to be sensitive to mechanical stretching, although moesin connects with ezrin and radixin proteins and cellular actin filaments to the plasma membrane. In addition to these differentially expressed proteins, we identified 15 other abundantly expressed phosphoproteins in HCS-2/8 cells (Fig. 2., Table 1, paper I). Notably, four of these were different protein disulfide isomerases, their precursors or isoforms. All of the above mentioned proteins identified were statistically significant compared to control conditions. Despite the phosphoprotein enrichment strategy, the exact phosphorylation sites on each protein could not be identified by 2D electrophoresis or further in-gel digestions and MS protein identification.

4.2 Effects of Rho-kinase inhibitor Y-27632 to fibroblasts (paper II)

Proteomic analysis revealed that two days of exposure to the Rho-kinase inhibitor, Y-27632, produced changes in 56 different proteins in human fibroblasts (Fig. 1., paper II). A total of 39 proteins were up-regulated and 17 were down-regulated when a cut-off was set to a ±1.3 fold change. A bioinformatic analysis of proteomic data using an IPA® pathway analysis determined that the changes in protein expression were connected to networks related to cellular assembly and organization, cellular morphology

40

and cellular function and maintenance (Fig. 2, paper II). Additionally, the pathway analysis suggested an increase in cell proliferation based upon the the expression profiles of annexin 2, calpain proteins, 14-3-3 protein zeta/delta, eukaryotic translation initiation factor 3 and 6, vinculin, ezrin and . Altogether, the protein pathway analysis suggested that Rho-kinase inhibition produces different types of functional responses in fibroblasts.

These functional responses, including cell morphology, proliferation, actin filament polymerization and migration were studied using time-lapse imaging. Short-term exposure to 10 µM concentration of the Rho-kinase inhibitor Y-27632 produced a transient, up to two-fold, increase in cellular proliferation during the initial 6 hours of treatment. (Fig. 3 and 4A, paper II). A depolymerization of actin filaments was inhibited and, thereby validated the inhibition of this RhoA/ROCK pathway-related response by Y- 27632 in fibroblasts. The average length of actin filament networks per cell was measured and the dynamic nature of actin filament polymerization and depolymerization was described as the relative rate of cellular protrusion formation/destruction (µm/h) (Fig. 4B and 4C, paper II). The largest increase in the lengths of the protrusions induced by Y-27632 was observed after 14 hours of Rho-kinase inhibition. At 14 hours, the average length per cell was 105.9 ± 42.5 µm compared to 20.3 ± 6.3 µm in untreated controls (mean ± 95% CI). The maximal difference between the average lengths of the protrusions was over fivefold. After this maximal effect, the average length slightly decreased and stabilized to around 80 µm/cell after 45 hours of treatment. The average lengths in the control cells remained in the range of 20-25 µm/cell starting from 8 hours through the end of the experiment. Scratch-wound assays, aimed to determine effects on cellular migration, revealed that Y-27632 significantly increased the lateral migration of fibroblasts. Rho-kinase inhibition produced a recovery of over 70% the original wound area compared to less than 30% in untreated controls (Fig. 5A and 5B, paper II). Y-27632-treated cells reached close to confluency in 25 hours, whereas untreated cells took up to 36 hours. By comparing cell layer filling rates (mm2/h) we determined that the Y-27632 effect on wound filling was the fastest immediately after the addion of inhibitor to cell cultures (Fig. 5C, paper II). Untreated control cells reached their fastest wound filling rates between 15 to 18 hours.

Filamentous actin was labelled with fluorescently-tagged phalloidin and both long filaments and filamentous network structures were observed using time-lapse imaging. Phalloidin staining of actin confirmed both the differential orientation and inhibited depolymerization of filamentous actin (Fig. 6A-B, 6E-F, paper II) by Y-27632. Immunofluorescence staining was

41

used to visualize the localization of vinculin and integrins α2, β1 and α5β3. Vinculin-associated focal adhesion plaques were abundantly expressed in control fibroblasts, but were almost entirely absent from Y-27632-treated cells (Fig. 6C-D, 6G-H, paper II). The lack of focal adhesion plaques supports the observation that Y-27632 increased the rate of cellular migration. Kinase inhibition did not produce any changes in the expression or localization of integrins (Figure 4).

Figure 4. Immunofluorescence visualization of integrins α2 (row A), β1 (row B) and α5β3 (row C) in fibroblast cells. Scale bar 25 µm.

Quantitative RT-PCR analysis was performed to monitor phenotype-specific gene expression in fibroblasts. The expression of the chondrocytic phenotype-associated Sox9 transcription factor was decreased by 30%, when fibroblasts were exposed to 10 µM Y-27632 for 48 hours (Fig. 7A, paper II). Y-27632 treatment appeared to induce an increase in aggrecan gene expression, but due to a high variability, statistical significance was not reached (Fig. 7B, paper II). Basal expression levels of aggrecan mRNA was low, as is to be expected in normal control fibroblasts. Type I collagen gene expression (procollagen α2(I) chain) remained stable during kinase inhibition and indicates a constant fibroblast phenotype (Fig. 7C, paper II). Versican gene expression did not significantly change upon Y-27632 exposure (Fig. 7D, paper II).

42

4.3 Hypoxia-induced responses in the proteome of HCS-2/8 cells (paper III)

The main aim of the third paper was to evaluate protein responses produced by a long-term (28 days) culturing period in a hypoxic environment. A hypoxic environment is normal for chondrocytes. Although the oxygen strain in articular cartilage is, on average, around 5%, many cell-based chondrocyte studies are performed in normoxic, 19% oxygen tension. A total of 1453 differentially expressed proteins were identified in HCS-2/8 cells. A DAVID Bioinformatics Resources 6.8 data analysis recognized 1438 of the identified proteins. Functional annotations were performed to reveal information and to divide the proteins in different cell organs and functional groups (Fig. 1 paper III). The majority of the identified proteins were associated to metabolic activities, gene expression and signal transduction. Proteomic analysis revealed 44 differentially synthetized proteins in a hypoxic environment compared to normoxia. The cut-off to reach a significant fold change was set to ±1.3. A total of 16 proteins were up-regulated and 28 proteins were down-regulated at hypoxia (Supplementary Table 2, paper III). Fifteen with the greatest up-regulation or down-regulation are presented in Figure 2, paper III. A long-term exposure to a 5% oxygen strain revealed several hypoxia-associated proteins that have not previously been connected to hypoxic conditions in chondrocytic cells (Table 1, paper III). Protein S100-P had the highest fold change of the novel hypoxia-sensitive findings. Hypoxia did not significantly change other S100 proteins. Many S100 proteins are known to be involved in cartilage- and chondrocyte- related functions (Yammani 2012). Therefore, we performed an additional quantitative analysis that revealed the relative abundances of nine additional S100 proteins in HCS-2/8 cells (Fig. 4., paper III). Notably, S100-A8 and S100-A9 proteins were not detected in HCS-2/8 cells although these proteins have been specifically connected to hypertrophic differentiation and osteoarthritic cartilage degradation processes in chondrocytes (Yammani 2012).

In addition to these novel findings, we recognized several hypoxia-sensitive proteins identified in previous studies. This group of previously recognized proteins, up-regulated at hypoxia, included NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 4-like 2, prolyl 4-hydroxylase subunit alpha-1, protein NDRG1, migration inhibitory factor, L- lactate dehydrogenase A chain, glycogen phosphorylase, prothymosin alpha, thioredoxin reductase 1 (cytoplasmic), BAG family molecular chaperone regulator 2, thioredoxin (mitochondrial), sequestosome-1, tumor necrosis factor receptor superfamily member 1B, epoxide hydrolase 1. The identification of these proteins validated the hypoxic exposure to HCS-2/8

43

cells. A protein pathway analysis was performed using the IPA® analysis software. This analysis suggested that hypoxia-induced protein responses were potentially related to cellular movement, hematological system development and function, and humoral immune response. The protein network depicting 19 related focus molecules is presented in Figure 3, paper III. The additional networks suggested by IPA® pathway analysis had a smaller number of focus molecules and were related to post-translational modifications, cell cycle, cellular development, cellular function and maintenance.

The protein pathway analysis identified that hypoxia activated transcription regulator HIF-1α (activation z-score 2.537, p-value of overlap 1.12E-06). Therefore, the hematological system development and function was interpreted to refer to HIF mediated regulation of the hypoxia-induced responses.

A quantitative RT-PCR analysis was performed to monitor chondrocytic phenotype and cartilage matrix related gene expression during long-term hypoxic environment. VEGFA gene expression was used as a positive control to monitor hypoxic conditions. 5% oxygen tension induced more Sox9 gene expression than normoxic control environment (Fig. 5, paper III). The largest difference was observed after 28 days in culture, when the expression of Sox9 was 1.77±0.15 higher at hypoxia than at normoxia (mean ± 95% CI, p<0.05). There was a time-dependent decrease in basal levels of Sox9 gene expression. Aggrecan and type II collagen are known targets of Sox9 transcription factor. Hypoxia produced an increase in ACAN and COL2A1 gene expression. However, these increases did not reach statistical significance until 28 days in culture. At the end of the experiment, relative COL2A1 gene expression was 4.40 ± 1.18 higher at hypoxia than at normoxia and ACAN expression was 2.49 ± 0.31 higher at hypoxia than at normoxia (mean ± 95% CI, p<0.05). The transcription factor Sox9 also regulates the expression of S100A1 and S100B genes, two genes that have been linked to chondrocyte phenotype and differentiation status (Saito et al. 2007). In this study gene expression trends of S100B, but also S100A1 to a lesser extent, shared similarities with the expression of aggrecan and type II collagen (Fig. 5, paper III). A 5% low-oxygen atmosphere increased COL1A1 gene expression after one-week in culture, but the low, 1.33-fold change (1.33 ± 0.22; mean ± 95% CI, p<0.05) together with its transient nature, was considered not to be biologically significant nor suggest any phenotypic alteration. The expression of another extracellular matrix-related gene, versican, could not be quantified due to its low expression in HCS-2/8 cells. Hypoxia-induced VEGFA fold changes were 2.19 ± 0.11 at 2 days, 2.34 ± 0.18 at 7 days and 3.59 ± 0.35 at 28 days in culture (mean ± 95% CI, p<0.05).

44

VEGFA gene expression significantly increased at all time points and validated hypoxic culturing conditions of HCS-2/8 cells (Fig. 5., paper III).

Short-term (72 hours) time-lapse imaging did not reveal any statistically significant differences in cellular proliferation rates in this study. The scratch-wound assays were also performed at both oxygen atmospheres and did not show any influence on lateral migration rates of HCS-2/8 cells (data not shown).

Although long-term hypoxia increased the gene expression of cartilage matrix components, increased secretion of sulfated GAGs or type II collagen could not be detected by DMMB or ELISA assays (data not shown).

4.4 Hypoxia and Rho-kinase inhibitor Y-27632 induced cartilage matrix production in chondrocytic cells (paper IV)

Previously, the Rho-kinase inhibitor Y-27632 was reported to promote the gene expression of cartilage extracellular matrix components and influence the differentiation status of chondrocytes. The main goals of this fourth paper were to evaluate the combined effects of long-term Rho-kinase inhibition by Y-27632 together with hypoxia in HCS-2/8 cells. The main focus was placed on proteomic response and cartilage matrix production.

Preliminary short-term (72 hours) time-lapse imaging experiments were performed to monitor cellular proliferation, morphology and lateral migration of HCS-2/8 cells. 10 µM of the Rho-kinase inhibitor Y-27632 did not produce changes in cellular proliferation at normoxic or hypoxic atmospheres (Fig. 1A and C, Supplementary Fig. 1, paper IV). The incubation of cells with Y-27632 inhibited actin depolymerization in both oxygen atmospheres and validated a Rho kinase inhibition by 10 µM Y-27632 in HCS-2/8 cells (Fig. 1B and D, Supplementary Fig. 1, paper IV). Scratch- wound migration assays demonstrated that Y-27632 slightly increased wound filling rates at early experimental time point. Following 12 hours of inhibition, the relative filling rate was at least 1.5 times higher in Y-27632- treated samples than controls at both oxygen tensions (Fig. 1E-F, Supplementary Fig. 2, paper IV). At the end of 72 hour monitoring period, Y-27632-treated cells at normoxia reached 75.4% ± 3.5% of confluency compared to 78.9% ± 3.9% in control cells. Whereas at 5% oxygen tension, Y-27632-treated cells reached only 66.7% ± 2.9% of confluency compared to 75.1% ± 3.5% of confluency in control cells.

45

A proteomic analysis revealed that long-term 10 µM Y-27632 treatment produced changes in 109 proteins (62 up- and 47 down-regulated) at normoxia (Fig. 2A, Supplementary Table 2, paper IV), while at hypoxia changes were detected in 101 proteins (64 up- and 37 down-regulated) (Fig. 3A, Supplementary Table 3, paper IV). The cut-off for protein fold changes was set to ±1.3. A protein pathway analysis (IPA® analysis) was performed on MS protein data.

The IPA® pathway analysis suggested that Y-27632 activated four protein networks under normoxic conditions. The first network related to cancer, organismal injury and abnormalities (score 59, Fig. 2B, paper IV) while the second and the third networks involved cell survival and death (scores 44 and 42). Cancer-related findings are most likely due to the chondrosarcoma cell model used in this study. The fourth activated network (score 41) referred to amino acid metabolism, hematological system development and function and to tissue development. A further detailed inspection of the networks by pathway analysis indicated those linked to a decrease of binding of cells (activation z-score 2.418, p-value of overlap 9.62E-03), an increase in apoptosis (activation z-score 1.645, p-value of overlap 1.12E-03) and an increase in proliferation of cancer cells (activation z-score 1.588, p-value of overlap 6.31E-03). The suggested increase in proliferation was slightly controversial, since the pathway analysis also simultaneously suggested a decrease in proliferation of prostate cancer lines (activation z-score -1.387, p-value of overlap 9.29E-03).

Rho-kinase inhibition under hypoxic conditions produced changes in 31 proteins that followed similar up- or down-regulated tendencies as observed at normoxia (Fig. 4B, paper IV). Any additional protein changes appeared to have a hypoxia-sensitive nature. A protein pathway analysis primarily connected differentially expressed proteins to molecular transport and to cell death and survival (26 target molecules, score 59, Fig. 3B, paper IV). The second network (score 39) related to cell cycle, cellular growth and proliferation and to tissue morphology, while the third protein network (score 34) referred to cancer and to organismal injury and abnormalities. The inspection of the networks revealed protein fold changes were associated to a potential decrease in the efflux of cholesterol (activation z-score -1.969, p-value of overlap 1.71E-03) and in the transport of cholesterol (activation z- score -1.501, p-value of overlap 3.53E-03). IPA® pathway analysis also predicted hypoxia-induced VEGFA activation (activation z-score 1.491, p- value of overlap 7.40E-06).

Long-term Y-27632 exposure and Rho-kinase inhibition produced differential expression patterns of several S100 proteins (p<0.05) at both

46

oxygen atmospheres (Fig. 4A, paper IV). Y-27632 produced the up- regulation of S100-B (8.26 ± 1.29), S100-A1 (2.11 ± 0.06), S100-A13 (2.03 ± 0.14), S100-P (1.98 ± 0.62) and S100-A4 (1.51 ± 0.32), and a down regulation of S100-A16 (-1.33 ± 0.06) at normoxia. In a hypoxic environment, Y-27632 produced the up-regulation of S100-B (3.73 ± 0.18), S100-A1 (2.58 ± 0.43), S100-A13 (1.88 ± 0.06) and S100-P (1.87 ± 0.19), and the down-regulation of S100-A6 (-1.79 ± 0.21) and S100-A16 (-1.68 ± 0.02). Of all the S100 proteins, prolonged kinase inhibition increased S100-B protein synthesis the most at both oxygen tensions, while S100-A4 up- regulation was observed only at normoxia and S100-A6 down-regulation was observed only with hypoxia.

A quantitative RT-PCR analysis indicated that Y-27632 increased the expression of the cartilage extracellular matrix production regulating transcription factor Sox9 both at normoxic and hypoxic atmospheres (Fig. 5, paper IV). The highest Sox9 expression level was observed after two days of Rho-kinase inhibition. Despite this, the largest relative difference between between treated and untreated controls of Sox9 gene expression was observed after 28 days of exposure to Y-27632. Sox9 gene expression was 2.03 ± 0.24 times higher than controls at normoxia and 1.48 ± 0.10 times higher at hypoxia. At 28 days, the combined effect of Y-27632 and hypoxia induced a 2.60 ± 0.18 fold change (p<0.001) in Sox9 expression.

Similar to Sox9 gene expression, both aggrecan and type II collagen gene expressions were increased the most after 28 days of exposure to Y-27632 at both normoxia or hypoxia (Fig. 5, paper IV). The changes in ACAN and COL2A1 were highest at hypoxia. Long-term Y-27632 exposure at a 5% oxygen tension increased ACAN gene expression 5.33 ± 0.93 fold (p<0.001) and COL2A1 gene expression 26.6 ± 4.14 fold (p<0.001) in comparison to the corresponding untreated normoxic controls.

Versican gene expression was not possible to determine due to its low expression level. Type I collagen slightly responded in a transient way to Y- 27632 exposure at either oxygen atmosphere. The observed changes were considered to be minimal in comparison to COL2A1 gene expression. Therefore, it is suggested that differences of COL1A1 expression could have only a minor role for extracellular matrix production.

The proteomic analyses revealed several protein expression changes of S100 proteins at both oxygen atmospheres by Y-27632. Therefore, the gene expression of S100A1, S100B and S100A16 were also determined by qRT- PCR in this study. 10 µM of Y-27632 significantly increased S100A1 gene expression at all time points at both atmospheres (Fig. 5, paper IV). The gene

47

expression profile of S100B shared similarities with S100A1 (Fig. 5, paper IV). The gene expression profiles of S100A1 and S100B supported the results obtained from the proteomic analyses. The S100B gene expression profile also shared many similarities with aggrecan expression profile. A similar trend was also found between S100B and type II collagen gene expression. In contrast, long-term Y-27632 treatment at hypoxia significantly decreased S100A16 gene expression. The fold change was 0.67 ± 0.06 (p<0.001) between the treated cells and controls (Fig. 5, paper IV). VEGFA gene expression was used as a positive control to monitor the hypoxic atmosphere on HCS-2/8 cells. Hypoxia induced VEGFA gene expression was independent of Rho-kinase inhibition.

An ELISA assay was performed to quantify secreted type II collagen from the culture media. 10 µM of Y-27632 slightly increased secreted type II collagen at normoxia but did not reach significance (Fig. 6A, paper IV). The combination of prolonged Rho-kinase inhibition together with a hypoxic environment increased the secretion of type II collagen by 6.2 ± 1.1 fold (Fig. 6A-B, paper IV). In addition, the production of sGAGs was estimated using the DMMB assay. Long-term (28 days) exposure to Y-27632 at normoxia increased the synthesis and secretion of sGAGs 31% ± 16% (Fig. 6C-D, paper IV). Under hypoxic conditions, long-term Rho-kinase inhibition increased the synthesis and secretion of sGAGs by 57% ± 9% compared to untreated control cells.

48

5. Discussion

5.1 The phosphoproteomic analysis of cyclically stretched HCS-2/8 cells (paper I)

Proteomic analyses of cartilage samples are challenging since the major components of the extracellular matrix are highly abundant in comparison to cellular proteins. The insoluble and fibrillar type II collagens form networks and larger aggregate structures with aggrecan and hyaluronan. Some other proteins and glycoproteins can bind to these matrix networks, as well. The proteome analysis of cartilage tissue samples requires the appropriate tecniques for sample preparation, e.g., the use of strongly chaotropic agents that can solubilize proteins prior to analysis (Wilson et al. 2008, Önnerfjord et al. 2012). Both the highly anionic aggrecan and hyaluronan can be precipitated by cationic cetyl pyridinium chloride, which can improve protein separation when using the traditional two-dimensional gel electrophoresis (Vincourt et al. 2006). Different kinds of sample preparation and prefractioning strategies have been developed, but there are still challenges to analyze the secretome and proteome of cartilage tissue samples. The HCS-2/8 cell model was used in our cellular stretching related phosphoproteomic study. The cell model was applicable for this kind of purpose since it made it possible to study biaxial cell stretching directly with chondrocytic cells and minimize additional sample preparation steps and analytical challenges derived from interfererence from cartilage matrix components.

A phosphoprotein enrichment strategy was chosen to purify cell lysates and use the protein loading capacity of IPG strips for phosphorylated proteins as efficiently as possible. Despite this target-orientated strategy, we could not detect the actual phosphate groups of the proteins. The reasons for missing phosphate groups relates most likely to inappropriate ionization methods for labile phosphorylation side chains, limited sequence coverages of the identified proteins and, in some cases, low amounts of extracted peptides. Moreover, phosphorylation enrichment methods may work more effectively for small peptides rather than larger and structurally more complex protein structures (Fíla and Honys 2012). During the last years, novel straightforward LC-MS based techniques have been developed that offer new and better tools for phosphoproteomic screening and the detection of phosphorylation sites. Despite these technological improvements, phosphoproteomics is still regarded as challenging, both for qualitative and quantitative analyses (Solari et al. 2015).

49

5.2 Cyclic stretching induced responses in chondrocytic cells (paper I)

The cyclic cellular stretching experiments were performed to evaluate stretch-induced phosphoprotein responses in human chondrosarcoma cells. Mechanical forces have been studied with regard to chondrocytes and cartilage tissue for a long time. However, the role of mechanical stretching of chondrocytes, an event that occurs during the compression of cartilage tissue, has not been as well studied as other types of mechanical stresses such as hydrostatic pressure. Previous stretching experiments have been based upon highly variable parameters and experiments have ended up with controversial results relating to the maintenance of cartilage matrix production. The cyclic stretching study in this thesis had the goal to evaluate stretch-induced phosphorylation changes in chondrocytic cells.

Mechanical stresses, such as compression, hydrostatic pressure and tension have been shown to induce the activation of extracellular signal regulated kinases (ERKs) in chondrocytes and chondrocytic cells (Fanning et al. 2003, Takahashi et al. 2003, Kopakkala-Tani et al. 2004). In our stretching experiments, a cyclic, 8% stretching at a frequency of 1 Hz activated ERKs, detected by Western blot technique. The differential expression of phosphorylated ERKs was not observed in Coomassie Blue-stained two- dimensional gels likely due to the relatively low sensitivity of this staining method and the low expression levels of ERK signaling mediators in the cells. The activation of other signaling routes, such as PI3K and Akt were not observed in this study. However, differentially expressed phosphoproteins related to other cellular functions were detected in this study. These include eukaryotic translation elongation factors 1 delta and 1 beta referred to as EF- 1 related GDP to GTP exchange. Moreover, eEF1D (isoform 2) which is also known as eEF1BdeltaL (eEF1BδL) regulates heat shock-associated genes (Kaitsuka and Matsushita 2015). This regulation is mediated via heat shock transcription factors and heat shock promoter elements. The reciprocally expressed isoforms of eukaryotic elongation factors had potentially specific roles in HCS-2/8 cells, but they could not be recognized in this screening study.

Moesin belongs to ERM proteins and forms complexes with ezrin and radixin. ERM proteins can connect cytoskeleton to plasma membrane and they are also present in cell adhesion sites. Moesin can also take part in other cellular functions, such as endosome maturation and microvilli formation (Oshiro et al. 1998, Chirivino et al. 2011). Hypotonic shocks induce actin reorganization and recruitment of the actin monomers and phosphorylated moesin to membrane protrusions (Tamma et al. 2007). It may be that the

50

mechanical cyclic stretching stress was transferred to the cellular cytoskeleton and induced the expression of phosphorylated moesin. Mechanosensitive up-regulation of phosphorylated moesin is a novel finding and has not been previously described. The up-regulation of prolyl 3- hydroxylase 1 (leprecan) was also identified as a novel stretch-induced protein. Prolyl 3-hydroxylase 1 is known to catalyze the post-translational modification of certain proline residues into (3S)-hydroxyproline (3-Hyp) on the α-chains of different types of collagens (Vranka et al. 2004).

In addition to the differentially expressed proteins, a group of 15 other, abundantly expressed phosphoproteins was identified in HCS-2/8 cells. These proteins are linked to many different cellular functions, and include structural roles, enzymatic roles and mitochondrial activities. Notably, four different protein disulfide isomerases, or their precursors, were abundantly expressed in HCS-2/8 cells. Protein disulfide isomerases are required to catalyze the isomerization of disulfide bonds and they can work as subunits for prolyl 4 hydroxylase enzyme, which has a crucial role for collagen biosynthesis (Koivunen et al. 2004).

The cyclical stretching experiment revealed four novel tension-related protein responses. These findings were not recognized to be associated with cartilage extracellular matrix production. Due to the small number of differentially expressed proteins and qualitative analyses of the study, it was not possible to analyze activated protein networks of stretch-induced signaling mechanisms. Although the full loading capacity of the IPG strips was used for phosphoprotein loading, it is obvious that the phosphoproteomic signaling analysis would have needed a more sensitive detection method than Coomassie Blue staining. Further studies would be needed to recognize the stretch-induced cellular signaling responses. A clarification of these responses are still of great interest since the mechanisms and roles of the different loading stresses for chondrocyte maintenance are still not properly known, and even controversial results have been reported.

5.3 Short-term Rho-kinase inhibition did not induce chondrocyte-specific cartilage extracellular matrix production in fibroblasts (paper II)

In previous studies, Rho-kinase inhibition has been demonstrated to promote chondrocyte-specific gene expression (Woods et al. 2005, Woods and Beier 2006, Matsumoto et al. 2012, Furumatsu et al. 2014). Additionally, a long-term exposure of fibroblast growth factor 2 (FGF-2) at hypoxic oxygen tension has been reported to induce chondrocyte-specific extracellular

51

matrix production by fibroblasts (Page et al. 2009). The aim of the second paper was to investigate the short-term (72 hour) Y-27632-induced proteomic responses in fibroblasts and to study whether Rho-kinase inhibition could induce chondrocyte-specific gene expression in fibroblasts. A pharmacology-based induction of cartilage extracellular matrix production would be of high interest since it could be useful for cell-based regenerative cartilage repair applications. The fibroblast and chondrocyte cells appear to have a relative straight-forward connection since chondrocytes cultured as monolayers rapidly lose their chondrocyte-specific phenotype and start to express fibroblast-specific type I and III collagens (von der Mark et al. 1977, Duval et al. 2009). Thereby, a chondrogenic induction of fibroblasts could potentially offer a novel strategy for direct pharmacological differentiation of patient-derived fibroblasts to chondrocytes. Despite the previous report of inducing effects of prolonged FGF-2 exposure under a hypoxic atmosphere (Page et al. 2009), the Rho-kinase inhibitor Y-27632 was used instead of FGF-2 growth factor in this study. FGF-2 was not used since it was shown to also induce several catabolic responses, i.e., the suppression of aggrecan gene expression and the induction of cartilage matrix degrading factors, such as ADAMTS-5 and MMP-13, in chondrocytes (Ellman et al. 2013). However, the precise role of FGF-2 may be more complex since some studies performed in human mesenchymal stem cells have demonstrated that the correct sequential dosing of FGF-2 and fibroblasts growth factors 9 and 18 could be used to support chondrogenic and osteogenic differentiation (Correa et al. 2015).

Short-term Rho-kinase inhibition was not sufficient to induce chondrocyte- specific cartilage extracellular matrix production in fibroblasts in this study. Acute exposure to Y-27632 did not increase gene expression of the transcription factor Sox9. Type I collagen and versican expressions remained at relatively stable levels and the slight increase of aggrecan expression was not statistically significant.

5.4 Rho-kinase inhibition induced fibroblast migration and transient proliferation (paper II)

A short-term (72-hour) exposure of human fibroblasts to the Rho-kinase inhibitor Y-27632 induced the differential expression of several proteins. These changes in expression were analyzed and were determined to be primarily associated to functional responses, such as cellular proliferation and migration. The inhibition of Rho-associated coiled coil kinases transiently increased during the first 12 hours of fibroblast proliferation. Y- 27632 treatment increased cell number up to two-fold during this period. After an initial accelerated proliferation, the total number of kinase

52

inhibitor-exposed cells slightly decreased. The mechanism behind this proliferation profile was not investigated in this mainly proteomic study.

The inhibition of actin depolymerization was an expected response and this was experimentally confirmed in fibroblasts. Time-lapse imaging revealed the dynamic nature of actin filament-based polymerization and depolymerization processes and demonstrated a net increase of Y-27632- induced cellular protrusions. Protein pathway analysis pointed out a number of potential canonical signaling pathways. The most potential routes were linked to protein 14-3-3 mediated signaling, actin cytoskeleton signaling and calcium binding calpain protease regulated mechanisms. Calpain protease can proteolytically degrade complexes of talin, paxillin and focal adhesion kinase and lead to the further disassembly of cellular focal adhesions (Chan et al. 2010, Cortesio et al. 2011). Rho-kinase inhibition depleted the vinculin- associated focal adhesions in our experiments in fibroblasts. It was notable that the synthesis of vinculin was simultaneously increased, therefore suggesting that Y-27632 influenced the cellular localization of vinculin in a crucial way. Reduced focal adhesions correlate well with the abserved increase in cellular migration. The motility assay, for an acute single dose of Y-27632, in primary foreskin fibroblast cells has not been previously described in a time-dependent manner. The increase in cellular motility is often considered a common response to Rho-kinase inhibition, especially using scratch-wound models (Nobes and Hall 1999, Magdalena et al. 2003, Hopkins et al. 2007, Shum et al. 2011, Breyer et al. 2012, Goetsch et al. 2014). However, increase in migration is not a global phenomenon, but rather a cell type-specific response. It has previously been shown that the Rho-kinase inhibitor Y-27632 can reduce the invasion of rat MM1 hepatoma cells, while blocking the motility of human prostate cancer cells (Somlyo et al.2000, Imamura et al. 2000). Moreover, Y-27632 exposure has been demonstrated to inhibit the migration process of human corneal fibroblasts (Zhou and Petroll 2010).

Cellular migration is based on the coordinated assembly and disassembly of actin filaments. Increased motility of fibroblasts supports wound healing processes that are apparently coordinated by small GTPases Rho, Rac, Cdc42 and Ras (Nobes and Hall 1999). Our time-lapse imaging indicated an increased absolute number of cellular protrusions. The maximum difference reached in lateral migration-based wound filling rates occurred three times faster in Y-27632-treated cells than in untreated controls. However, the rate decreased in a time-dependent manner during the experiment. Cellular proliferation also had a potential role in the wound-filling process, but as the proliferation only increased in a transient way at the beginning of the

53

experiment, it suggests that increased proliferation only had a limited role in the overall wound-filling process.

5.5 Increased cellular migration was potentially mediated by mDia and Rho-kinase mediated signaling routes (paper II)

Rho-kinases are known to mediate signals via LIM kinases (type I and II) by the phosphorylation of cofilin and by inducing the phosphorylation of myosin light chain. These of both routes are connected to the reorganization of the actin cytoskeleton and are both linked to actin stress fiber formation and to cell motility (Magdalena et al. 2003). Rho-kinase-related signaling routes have also additional regulatory mechanisms, like γ-actin, which can be a potential upstream or direct regulator of the Rho-kinase signaling system (Shum et al. 2011). It is known that a scratch-induced wound activates Cdc42 that subsequently leads to the activation of Par6/aPKC complex and a further GSK-3-mediated polarization of the cytoskeleton (Cau and Hall 2005). In addition to this, the localized activation of Cdc42 and Pak kinase has been proposed to connect to the scratch- induced activation of βPIX. This would mediate the signal to the downstream mediator Rac and ultimately lead to the polarization of actin- dependent cellular protrusions. Fibroblast migration requires cellular protrusions of the lamellipodium and actin filament disassembly is a necessary part of this mechanism. Previous research has demonstrated the key role of actin depolymerizing factors (ADF) and non-phosphorylated cofilin for the polarized protrusion of lamellipodia structures (Dawe et al. 2003). Even though actin stress fibers and their connections to the extracellular matrix form the basis for cellular migration, cellular motility also needs dynamically controlled focal adhesions at different regions of the cell (i.e., the front and the back) (Ridley et al. 2003). These focal adhesions must be controlled in a coordinated way, which means that new adhesions are formed in protrusions and older adhesions are disassembled to allow cellular movement.

Vinculin is one of the actin filament-binding proteins that connect the actin cytoskeleton to the extracellular matrix (Menkel et al. 1994). The regulation of actin dynamics in focal adhesions is a complex process and a group of actin-binding proteins, like talin, tensin and α-actinin, take part in this regulation (Le Clainche and Carlier 2008). However, both vinculin and formin (mDIA1) are known to involve the elongation step of actin filaments in focal adhesions (Le Clainche et al. 2010). Rho has been proved to mediate the signal to ROCKs and affect the localization of vinculin in different cell types, as well as actin-myosin contractility that modulates the assembly

54

process of focal adhesions (Kotani et al. 1997, Dumbauld et al. 2010). Furthermore, Rho has been shown to have a role in controlling Golgi polarity and cell motility via mDIA (Satoh and Tomonaga 2001, Magdalena et al. 2003). In summary, we suggest that the increase in cellular migration of primary human foreskin fibroblasts observed using a scratch-wound model was produced by a Y-27632-mediated inhibition of RhoA activity. This inhibition of RhoA can obviously increase cellular migration by affecting both mDIA- and ROCK-mediated routes. Moreover, one previous study demonstrated that Y-27632-mediated Rho inhibition especially affects cellular motility via ROCK2 in myoblasts (Goetsch et al. 2014). A similar mechanism is also possible in fibroblasts, although the exact signaling pathway was determined from our experimental data.

5.6 Long-term hypoxia supported chondrocyte specific gene expression (paper III)

A prolonged 28-day culture of HCS-2/8 cells at a 5% low-oxygen atmosphere slightly enhanced the gene expression of the transcription factor Sox9. Moreover, this long-term cell culture at hypoxia induced Sox9 regulated gene expression of chondrocyte-specific type II collagen and aggrecan genes. In addition, expression of the chondrocyte differentiation status S100A1 and S100B genes were also increased after 28 days of exposure to a low oxygen atmosphere. Despite the promising gene expression profiles, an actual increase of cartilage extracellular matrix production was not detected in this study. However, the gene expression results indicated that hypoxia was a favorable factor for conserving a chondrocyte phenotype during prolonged in vitro culture of HCS-2/8 cells. Data suggest that the co-administration of some chondrocyte phenotype-supporting factor, or factors, at a hypoxic atmosphere would help to prevent the loss of chondrocyte phenotype of monolayer cultures and even promote cartilage extracellular matrix production. These kinds of responses would be optimal for cell-based regenerative cartilage repair techniques.

5.7 Long-term hypoxia induced known and novel protein responses in chondrocytic cells (paper III)

A number of proteomic studies have been previously performed in normal healthy and osteoarthritic chondrocytes, but our present study focused in particular on long-term hypoxia-induced proteomic responses. Proteomic analysis revealed 44 hypoxia-induced protein responses in chondrocytic HCS-2/8 cells. A number of these differentially expressed proteins have previously been recognized as hypoxia-sensitive responses, but this study also revealed a novel group of hypoxia-related proteins. Many of the

55

previously recognized hypoxia-sensitive proteins, such as lactate dehydrogenase, NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 4- like 2, protein NDRG1 and macrophage migration inhibitory factor are known to be regulated in a HIF-dependent manner. VEGFA gene expression was used as a positive control to validate hypoxic culturing conditions, and the protein pathway analysis suggested increased activation of HIF pathway.

A bioinformatic DAVID analysis identified three differentially expressed proteins linked to regulation processes of reactive oxygen species (ROS). The three proteins identified were thioredoxin 1, mitochondrial thioredoxin (thioredoxin 2) and multifunctional cytochrome C. All of these proteins were down-regulated in a hypoxic atmosphere. A strong down-regulation and a decrease in the activity of thioredoxin 1 is a well-known hypoxia-induced response (Naranjo-Suarez et al. 2012). Thioredoxin 1 involves the regulation of ROS levels in cells, however its expression is regulated in a HIF- independent manner. A further analysis of ROS production was not performed in this study since our focus was directed towards identifying factors involved in the maintenance of chondrocyte phenotype.

Articular cartilage is particular in its hypoxic nature and thereby chondrocytes have adapted to low oxygen tension. Chondrocytes are known to use glycolysis for their normal energy production, not only under hypoxic conditions but also at normoxia, and the activity of lactate dehydrogenase is a key enzyme in chondrocyte metabolism (Lee and Urban 1997). Thereby, the increased expression of L-lactate dehydrogenase A chain at hypoxia potentially relates to hypoxia-induced glycolytic energy production in HCS- 2/8 cells.

Long-term 5% oxygen strain notably increased gene expression of N-myc downstream regulated gene-1 (NDRG1) in HCS-2/8 cells. Its expression was up-regulated more than two-fold. Hypoxia-induced NDRG1 up-regulation is a known response in many types of cancer cells and it has been considered as a diagnostic marker for certain solid tumors (Fotovati et al. 2011). The protein NDRG1 has several biological functions that involve hormonal responses, cell growth, and differentiation. Interestingly, NDRG1 up- regulation has also been connected to suppressing metastasis in certain types of cancer cells and the molecular mechanisms behind this suppression have been previously investigated (Wangpu et al. 2015). Hypoxia-induced up- regulation of NDRG1 and its possible inhibition of oncogenic signaling mechanisms most likely only have minor effects on actual metastasis suppression of chondrosarcomas, but this possible anti-cancer response would be an interesting topic for future studies.

56

The novel hypoxia-sensitive identified proteins were not observed to be closely associated with each other. NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 11 (mitochondrial) takes part in complex I of the respiratory chain. Hypoxia-induced expression of this NADH dehydrogenase enzyme might be related to other hypoxia-regulated NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 4-like 2 enzyme, although this connection has not been previously reported. The role of 40S ribosomal protein S26 has been shown to involve p53 activity of damaged DNA, but chondrocyte- specific roles have not been reported (Cui et al. 2014). Previous literature is limited to connect either hypoxia or cartilage biology with protein CDV3 homolog, acyl-coenzyme A thioesterase 9 (mitochondrial) and tubulin beta-8 chain. Further studies are warranted to determine their biological roles in chondrocytes. The hypoxia-sensitive role of CD166 antigen was also one of the novel oxygen-dependent protein discoveries. The hypoxic atmosphere decreased CD166 expression by over two-fold. The exact biological role of this decrease remains unknown in the context of chondrocytes, although an up-regulation of CD166 has been previously linked to chondrogenic differentiation of pluripotent stem cells (Wu et al. 2013). This chondrogenic connection has not been studied in detail; therefore, the putative role of CD166 in chondrocyte differentiation and phenotype maintenance is worth studying. Moreover, the decrease of TNF receptor superfamily member 1B was observed in this long-term hypoxia study and identified as a novel hypoxia-sensitive protein in chondrocytic cells. Oxygen tension affects the expression of this protein via phosphorylation changes of forkhead box protein O3 (Ding et al. 2009). Tumor necrosis factor receptor superfamily member 1B acts as a receptor, which can mediate, for example, TNF-α related metabolic effects.

Proteomic analysis of HCS-2/8 cells revealed in total 1458 different proteins. These findings are unique since the proteins identified were based on at least one unique peptide that allowed the separation of closely related proteins and/or their isoforms from one another. Proteomes are mainly descriptive data sets. However, the data of the HCS-2/8 proteome was collected as a part of the third study since it could provide new information for the use of HCS- 2/8-based cellular models.

57

5.8 A short-term Rho-kinase inhibition at normoxic or at hypoxic conditions did not induce HCS-2/8 cell proliferation (paper IV)

The main goal of the fourth study focused on evaluating the long-term effects of Rho-kinase inhibition both at normoxia and at cartilage tissue modelling hypoxic conditions. The study began with a short-term (60-hour) single-dose of Y-27632 to initially monitor cell morphology, the formation of cellular protrusions and cell proliferation. This study was designed to evaluate and determine the capability of Y-27632 to induce responses in chondrosarcoma cells and to compare the functional responses with our previous short-term experiments in fibroblasts. Visual monitoring revealed that Y-27632 inhibited the depolymerization of filamentous actin. Using a scratch-wound model, results pointed to a short and transient oxygen-independent acceleration in cellular migration by Y-27632. Y-27632-induced wound filling rates decreased more rapidly than what was previously observed in fibroblasts. In contrast to our previous study in fibroblasts, Y-27632 treatment in HCS-2/8 cells did not induce similar proliferation as observed in fibroblasts (paper II). In addition to our own observations in fibroblasts, Y-27632 has been reported to increase cellular proliferation in many other cell types, such as astrocytes, limbal epithelial cells, keratinocytes and mesenchymal stem cells (Yu et al. 2012, Chapman et al. 2014, Nakamura et al. 2014, Sun et al. 2015). However, Y-27632 has not been noticed to increase the proliferation of different types of cancer cells (Routhier et al. 2010, Jiang et al. 2015, Wang et al. 2016). Inhibition of Rho-kinase coiled coil proteins has been reported to inhibit cancer cell growth and invasion (Wang et al. 2016). There appears to be a tendency for Rho-kinase inhibition to promote cellular proliferation of normal cells, but not in immortalized cancer cell lines. Studies on the Rho-kinase related cancer invasiveness have demonstrated that Y-27632 can suppress certain cancer growth and/or invasion related responses (Imamura et al. 2000, Jiang et al. 2015, Wang et al. 2016). An accurate mechanism of Y-27632-induced cell proliferation differences between normal cells and cancer cells is still not well understood. The potential proliferation decreasing effects of Y-27632 makes ROCK inhibitors interesting subjects for cancer research.

58

5.9 A long-term exposure of the Rho-kinase inhibitor Y-27632 at hypoxic atmosphere enhanced cartilage extracellular matrix production (paper IV)

The gene expression of transcription factor Sox9 increased in all Y-27632- treated samples and the expression of COL2A1 and ACAN were simultaneously increased by long-term exposure of Y-27632, especially at low oxygen tension. The effects were not so large at the level of secreted end- product, although increases in type II collagen and sGAG supported the idea of the increased production of extracellular matrix in Y-27632-treated cells in a low oxygen atmosphere.

Type II collagen and aggrecan are the major structural components in articular cartilage. The expressions of these macromolecules tend to significantly decrease when chondrocytes are cultured in monolayers. This study revealed an efficient way to support the maintenance of a chondrocytic phenotype and to enhance the synthesis of type II collagen and aggrecan in long-term cultures of chondrocytic cells. We found that 10 µM of the Rho- kinase inhibitor Y-27632 at a 5% oxygen atmosphere was optimal. This result is very interesting and gives a basis to investigate Rho-kinase inhibition in healthy human primary chondrocytes. In addition, the use of three-dimensional culture techniques in future experiments can be seen appropriate, or even necessary, since this has been shown to significantly affect cell behavior in a number of studies. Furthermore, three-dimensional cultures are required in order to promote a proper assembly of newly synthetized extracellular matrix components of cartilage.

5.10 Long-term Y-27632 exposure induced several-fold protein changes at both oxygen tensions (paper IV)

Proteomic analysis together with a protein pathway analysis was performed to determine protein responses and their roles in chondrocytic cells. The analysis revealed that long-term Y-27632 exposure produced changes in several proteins both at normoxia and at hypoxia. The long-term Y-27632 exposure at normoxia produced the highest increase in the synthesis of S100-B, adseverin, S100-A1, S100-A13 and S100-P proteins. At hypoxia, the highest increases were related to the translocator protein, Ras-related protein Rab-11B, S100-B, S100-A1 and insulin-like growth factor-binding protein 7. Rho-kinase inhibition decreased the expression of large neutral amino acids transporter small subunit 1, retinoid-inducible serine carboxypeptidase, 4F2 cell-surface antigen heavy chain, HLA class I histocompatibility antigen, A23 alpha chain and spliceosome RNA helicase DDX39B proteins the most at normoxia. In a hypoxic environment, Y-27632

59

decreased large neutral amino acids transporter small subunit 1, cyclin- dependent kinase inhibitor 1, serine hydroxymethyltransferase (cytosolic), S100-A6, glutaredoxin-1 and tubulin alpha-4A chain the most. Notably, Y- 27632 produced differential expression of several members of the S100 protein family. The tendencies of the changes in S100 proteins were relatively similar at both atmospheres, although overall, in comparison, only around 30% of all Y-27632-induced protein responses were similar at both atmospheres.

Y-27632 treatment significantly induced cartilage extracellular matrix production under hypoxic conditions, but the protein pathway analysis did not recognize this response. Although the proteomic results did not reveal an increase in type II collagen synthesis, it did detect an increase in the synthesis of 3'-phosphoadenosine 5'-phosphosulfate synthase 1 and 2 enzymes at hypoxia. These enzymes have a crucial role in the GAG sulfation process, thereby establishing a direct link to an increase in proteoglycan production. The proteomic analysis was directed to cell samples, so it is understandable that secreted cartilage extracellular matrix components may not have been particularly well-identified. The accuracy and the explanatory values from the protein pathway analysis would have benefited from a simultanueous analysis of the secretome. Pathway analysis techniques are improving and the content of bioinformatic databases is rapidly expanding. Still, different pathway analysis techniques have their limitations, which can lead even to the misinterpretion or undiscovered results (García-Campos et al. 2015).

The main aim of this fourth study was related to cartilage phenotype and extracellular matrix production. However, we also had an interest to identify additional Y-27632-induced responses in HCS-2/8 cells at both oxygen atmospheres. The potency of different classes of Rho-kinase inhibitors for various therapeutic purposes (i.e., cancer, cardiovascular and neuronal disease, diabetes-related complications) have been previously studied, and some clinical trials are currently underway (Wei et al. 2016). Currently, Y- 27632 is not therapeutically used, but other modified Rho-kinase inhibitors, such as fasudil and ripasudil, have been approved for human medical use (Shibuya et al. 1992, Garnock-Jones 2014). The novel and therapeutically favorable Y-27632-induced protein responses could direct Rho-kinase inhibition studies in novel directions.

Interestingly, long-term (28-day) Y-27632 treatment decreased the synthesis of the SLC7A2 (LAT1, CD98LC) and SLC3A2 (4F2hc, CD98HC) at both oxygen atmospheres studied. Proteins SLC7A2 and SLC3A2 form a heterodimeric amino acid transporter complex that transports large neutral

60

amino acids into cells (Fotiadis et al. 2013). This amino acid transporter can also transport amino acid-like compounds and it has specific roles in L- DOPA and triiodothyronine T3 and T4 transport in certain tissues. Previous studies have shown that LAT1 or LAT1/4F2hc are overexpressed in many cancers, and include triple negative , different sarcomas, malignant pleural mesothelioma, pancreatic cancer, renal cell carcinoma and gliomas (Kaira et al. 2011, Furuya et al. 2012, Betsunoh et al. 2012, Kaira et al. 2012, Koshi et al. 2015). LAT1/4F2hc transporter can mediate growth- related responses and increased LAT1 or LAT1/4F2hc expression has been correlated with malignant phenotype and tumor progressive activities. Previous studies have pinpointed the crucial role of LAT1 for cancer related activities and even as prognostic marker in certain cases (Kim et al. 2006, Betsunoh et al. 2012, Furuya et al. 2012, Kaira et al. 2012). Thereby, inhibition of LAT1/4F2hc complex and LAT1 have been suggested to be potential targets for novel cancer treatments (Cormerais et al. 2016, Marshall et al. 2016).

In this study, it was observed that the long-term Y-27632 treatment decreased the protein synthesis of both components of the LAT1/4F2hc transporter. Of the two, LAT1 was the most down-regulated component at both oxygen atmospheres. This down-regulation of LAT1/4F2hc transporter appears to be a novel Rho-kinase inhibition-related response. The decreased expression of this transporter most likely limited the transport of essential amino-acids in HCS-2/8 cells and thereby limited cellular growth and proliferation. The mechanisms, or consequences, of this decreased expression of LAT1/4F2hc transporter was not currently studied. However Y-27632-induced LAT1 down-regulation provides new aspects to Rho-kinase inhibition-related cancer research.

5.11 Long-term exposure of the Rho-kinase inhibitor Y-27632 influenced changes to S100 protein synthesis (paper IV)

The S100 protein family consists of 24 members (Donato et al. 2013). The majority of S100 proteins are linked to calcium signaling and have EF-hand type helix-loop-helix calcium-binding sites. The roles of the different S100 proteins are divided to intracellular, intra- and extracellular, or extracellular functions. The main roles of S100 proteins relate to various regulations of cellular and signaling processes. As reviewed by Donato et al. (2013), S100 proteins are involved in processes, such as Ca2+-signaling, proliferation, differentiation, metabolism, migration and apoptosis in many cell types.

61

Secreted S100 proteins can produce responses by binding to cell surface receptors, such as receptors for advanced glycation end products (RAGE) and toll-like receptors, G protein-coupled receptors, scavenger receptors and even to some glycan structures (Yammani 2012, Donato et al. 2013). It has previously been reported that the proteins S100-A1, S100-A2, S100-A4, S100-A8, S100-A9, S100-A11 and S100-B are all commonly connected to articular cartilage (Yammani 2012). A proteomic screening of human chondrocytes isolated from non-arthropathic knee articular cartilage and cultured in alginate, revealed that these chondrocytes expressed S100-A1, S100-A4, S100-A6, S100-A10, S100-A11, S100-A13, and S100-B proteins (Lambrecht et al. 2010). Another proteomic study using healthy human chondrocytes revealed similar protein profiles with some slight differences (Tsolis et al. 2015). In that study, the expression of S100-A16 protein was observed as a new finding and the expression of the low expressed regulator, S100-B, was not identified. These minor changes in proteomic profiles of normal human chondrocytes may relate to different cell origins, different culturing conditions or technical differences used for proteome analysis.

Our proteomic analysis (paper III) with monolayer-cultured HCS-2/8 cells revealed that this chondrosarcoma cell line expresses S100-A1, S100-A4, S100-A6, S100-A10, S100-A11, S100-A13, S100-A16, S100-B and S100-P proteins. The S100 protein profile of HCS-2/8 cells was qualitatively similar in comparison to normal chondrocyte S100 protein profiles. An exception to this was observed with the S100-P protein, one that was not previously reported to be expressed in chondrocytes. S100-P protein expression might relate to cancer cell phenotype of HCS-2/8 cells, but more likely may relate to an improved sensitivity of detection. Our study (paper III) revealed that S100-P and S100-B proteins were the lowest expressed proteins of the S100 family in HCS-2/8 cells.

Long-term Y-27632 exposure produced the up-regulation of S100-A1, S100- A13, S100-B and S100-P in HCS-2/8 cells. These up-regulation tendencies were similar at both oxygen atmospheres, although S100-B protein was significantly up-regulated to a greater extent at normoxia than at hypoxia. However, a part of the strong S100-B up-regulation at normoxia was related to higher variability of the results; therefore, the actual biological response may be smaller than what was determined in the present study. Y-27632- induced a down-regulation of S100-A16 and had a similar trend at both oxygen tensions. Oxygen-dependent differential expression of some S100 proteins was observed, such that S100A4 was up-regulated at normoxia and S100-A6 was down-regulated at hypoxia.

62

The S100 proteins have been reported to regulate many cellular responses. In cartilage tissue, diverse S100 protein functions have been connected to tissue maintenance, differentiation status, hypertrophic differentiation and calcification processes (Yammani 2012, Diaz-Romero et al. 2014). Both S100-A1 and S100-B proteins are targets of the transcription factor Sox9 and its coactivators Sox5 and Sox6 (Saito et al. 2007). These S100 proteins have been suggested to inhibit the terminal differentiation of chondrocytes. A study investigating the roles of S100-A1 and S100-B in human articular chondrocyte phenotype uncovered a divergent, but associated, role of these proteins and their potential to act as markers for chondrocyte differentiation status (Diaz-Romero et al. 2014). In this study, the increased protein syntheses of S100-A1 and S100-B were proposed to be chondrocyte phenotype supporting responses.

Protein S100-A4-based stimulation has been demonstrated to increase the synthesis of the catabolic MMP-13 enzyme in chondrocytes (Yammani et al. 2006). The long-term Y-27632 exposure induced an increase in the synthesis of S100-A4 at normoxia, but not at hypoxia. The actual expressions of the secreted MMPs were not determined in this study, but it can be suggested that Y-27632 treatment at a 5% low-oxygen tension was more favorable for a chondrocytic phenotype since the normoxia-related risk of increased S100- A4 was not observed.

Protein S100-A6, which is also known as calcyclin, was previously identified in chondrocytes (Lambrecht et al. 2010, Tsolis et al. 2015). The localization of S100-A6 has been associated to chondrocyte membranes, but its chondrocyte-specific roles are not well known (Matta et al. 2015). It appears to have a cell type-specific nature (Donato et al. 2013). Previous studies have revealed that S100-A6 can interact with the calcyclin-binding protein/Siah- 1-interacting protein (CacyBP/SIP), which involves the regulation of the Hsp90 chaperone (Schneider and Filipek 2011, Góral et al. 2016). CacyBP/SIP has been linked to many cellular functions, such as ubiquitination, proliferation, differentiation and cytoskeletal reorganizations. Therefore S100-A6 may have a regulative role for these cellular activities (Schneider and Filipek 2011). Other proteins that interact with S100-A6 are caldesmon, calponin, kinesin light chain and tropomyosin (Donato et al. 2013). The long-term Rho-kinase inhibition at hypoxic environment significantly decreased S100-A6 expression, but this reduction could not be connected to any specific cellular response in this study. The expression of S100-A6 has been shown to belong to the normal chondrocyte proteome. Therefore, the functional roles of the calcium-binding modulator S100-A6 are worth further studies.

63

The expression of S100-A8 protein is low in healthy chondrocytes (Yammani 2012). The increase of S100-A8 expression has been connected to osteoblast differentiation. Furthermore, the co-expression of S100-A8 and S100-A9 proteins has been reported to relate to the hypertrophic status of chondrocytes and further calcification and bone formation processes (Zreiqat et al. 2007). It was notable that Y-27632 treatments at normoxic or at hypoxic atmospheres did not influence the expression of these osteoarthritic inflammation and hypertrophy-related S100 proteins.

Protein S100-A13 is known to form a complex with fibroblast growth factor 1 (FGF-1) and p40 synaptotagmin and it has been proposed to function in cellular signal transduction (Ridinger et al. 2000). Like in many cases of S100 proteins, its cellular localization is relatively well-known and there are suggested roles for S100-A13. Still, the precise role of this protein in chondrocytes is currently unknown. It is possible that Y-27632-induced S100-A13 expression is a stress response that could interact with FGF-1. In endometriosis, S100-A13 and FGF-1 participate in vascular remodeling and angiogenesis (Hayrabedyan et al. 2005). Due to the origin of HCS-2/8 cells and similar S100-A13 expression tendencies at both normoxia and hypoxia, the activation of angiogenesis was not proposed as a response in this context. The actual functions of S100-A13 in articular chondrocytes and cartilage tissue will require additional studies.

One of the activation mechanisms of ezrin (ERM protein) relates to the calcium-dependent binding of ezrin and S100-P (Austermann et al. 2008). This kind of activation is involved in the increased cellular migration and metastasis of endothelial cells. Moreover, extracellular S100-P can bind to RAGE receptors and mediate tumor growth and metastasis related signaling (Arumugam and Logsdon 2011). In the chondrocyte or chondrosarcoma, the specific role of S100-P has not been previously reported. The long-term Y- 27632 influence on cellular migration was not monitored in this study, while the short-term time-lapse monitoring did not reveal a constant increase in cellular migration.

64

6. Conclusions

Both human chondrosarcoma HCS-2/8 cells and primary fibroblasts have been used in this thesis as in vitro cell models. The following conclusions can be made based on the observed results.

 An 8% cyclic cell stretching at a frequency of 1 Hz produces protein phosphorylation changes in HCS-2/8 cells. The stretching stress activates extracellular regulated kinase (ERK) signaling cascades and induces protein phosphorylation responses that can be linked, e.g., to the cytoskeleton and translational activities.

 Short-term (72 hour) Rho-kinase inhibition by Y-27632 does not induce chondrocytic gene expression in human foreskin fibroblasts but induces functional responses. These include inhibition of depolymerization of filamentous actin, increases in proliferation, reduction of focal-adhesion related vinculin plaques and stimulation of lateral cell migration.

 The long-term culture of chondrosarcoma cells at low 5% oxygen tension promotes COL2A1 and ACAN gene expression thereby supporting the conservation of their chondrocytic phenotype. A hypoxic atmosphere, similar to that of articular cartilage, is not sufficient to produce an enhancement in the synthesis of cartilage extracellular matrix proteins in HCS-2/8 cells.

 Long-term (28-day) Rho-kinase inhibition at normoxia supports the cartilage extracellular matrix gene expression in HCS-2/8 cells but does not increase type II collagen production at the protein level.

 Long-term Y-27632 exposure at 5% oxygen tension supports type II collagen and aggrecan expression at the protein level. This increased production can benefit cell-based cartilage repair techniques that require extended in vitro cell culture.

 The HCS-2/8 cellular model combining hypoxia and Rho-kinase inhibition support the previous suggestion that S100A1 and S100B gene and protein expression correlate to chondrocyte differentiation status and can be used as markers for extracellular matrix production.

65

 Long-Term (28-day) Rho-kinase inhibition decreases LAT1 and 4F2hc expression in HCS-2/8 cells. The overexpression of heterodimeric LAT1/4F2hc transporter relates to many cancers. Inhibiting this transporter is one of the goals of current cancer studies. Therefore the observed Y-27632 down-regulation of LAT1/4F2hc might provide a new direction for Rho-kinase inhibition as a cancer suppressing therapy.

66

Acknowledgements

I thank the chief supervisor Mikko Lammi for a great possibility to perform my PhD thesis project in your research group. I appreciate your way to teach and support to perform scientific studies with positive attitude and patience. Your forward-looking visions have given new sights to design experiments and utilize versatile techniques to solve scientific problems.

I thank co-supervisor Jukka Häyrinen for great introduction to protein chemistry with a special focus on mass spectrometric analyses. Your advices and criticism have been constructive and thought me to appreciate tough working and precise working methods. I thank possibilities to take part in teaching in your protein chemistry courses.

I thank all co-workers and co-authors for their advices and efforts to conduct the surveys of this thesis. Chengjuan Qu, Markku Varjosalo, Joakim Bygdell, Cecilia Fernández-Echevarría, Daniel Marcellino and Hannu Karjalainen, you all have done excellent work and made important contributions to help me perform the studies. I am grateful to Daniel Marcellino for revising the language of this thesis. I also thank technicians Elina Reinikainen, Sini Miettinen and Hanna Eskelinen of skillful technical assistance.

I thank honestly all our research group members. Most importantly, Juha Prittinen and Janne Ylärinne, you have been great friends and I appreciate greatly our on-going collaboration with biomechanical cartilage studies.

Anita Dreyer-Perkiömäki, Anna-Lena Tallander and Selamawit Kefela, I thank you all your administrative work. Thanks for Göran Dahlgren of IT support. I thank my examiner Per Lindström of your constructive opinions and Per-Arne Oldenborg for encouraging comments and advices to finish my studies in Umeå University.

I thank senior lecturers Maria Halmekytö, Sanna Ryhänen and Annikka Linnala-Kankkunen of rewarding teaching tasks. I thank staff, colleagues and friends in the Departments of Integrative Medical Biology (Umeå University) and Institute of biomedicine (University of Eastern-Finland) who has supported my thesis project during the years.

Special thanks to my parents Pirjo and Timo for your encouraging support during my studies.

I am grateful for financial support from Finnish Cultural Foundation.

67

References

Altman RD, Manjoo A, Fierlinger A, Niazi F, Nicholls M. The mechanism of action for hyaluronic acid treatment in the osteoarthritic knee: a systematic review. BMC Musculoskelet Disord. 16: 321, 2015.

Amano M, Mukai H, Ono Y, Chihara K, Matsui T, Hamajima Y, Okawa K, Iwamatsu A, Kaibuchi K. Identification of a putative target for Rho as the serine-threonine kinase protein kinase N. Science. 271(5249): 648-650, 1996.

Antonioli E, Lobo AO, Ferretti M, Cohen M, Marciano FR, Corat EJ, Trava- Airoldi VJ. An evaluation of chondrocyte morphology and gene expression on superhydrophilic vertically-aligned multi-walled carbon nanotube films. Mater Sci Eng C Mater Biol Appl. 33(2): 641-647, 2013.

Appleton CT, Usmani SE, Mort JS, Beier F. Rho/ROCK and MEK/ERK activation by transforming growth factor-alpha induces articular cartilage degradation. Lab Invest. 90(1): 20-30, 2010.

Armbruster DA and Pry T. Limit of blank, limit of detection and limit of quantitation. Clin Biochem Rev. 29 (Suppl 1):S49-S52, 2008.

Arumugam T and Logsdon CD. S100P: a novel therapeutic target for cancer. Amino Acids. 41(4): 893-899, 2011.

Asik M, Ciftci F, Sen C, Erdil M, Atalar A: The microfracture technique for the treatment of full-thickness articular cartilage lesions of the knee: midterm results. Arthroscopy. 24(11): 1214-1220, 2008.

Austermann J, Nazmi AR, Müller-Tidow C, Gerke V. Characterization of the Ca2+ -regulated ezrin-S100P interaction and its role in tumor cell migration. J Biol Chem. 283(43): 29331-29340, 2008.

Barre PE, Redini F, Boumediene K, Vielpeau C, Pujol JP. Semiquantitative reverse transcription-polymerase chain reaction analysis of syndecan-1 and - 4 messages in cartilage and cultured chondrocytes from osteoarthritic joints. Osteoarthritis Cartilage. 8(1): 34-43, 2000.

Barrett-Jolley R, Lewis R, Fallman R, Mobasheri A. The emerging chondrocyte channelome. Front Physiol. 1: 135, 2010.

68

Bastow ER, Byers S, Golub SB, Clarkin SE, Pitsillides AA, Fosang AJ. Hyaluronan synthesis and degradation in cartilage and bone. Cell Mol Life Sci. 65(3): 395-413, 2008.

Bayliss MT, Osborne D, Woodhouse S, Davidson C. Sulfation of chondroitin sulfate in human articular cartilage. The effect of age, topographical position, and zone of cartilage on tissue composition. J Biol Chem. 274(22): 15892- 15900, 1999.

Bell DM, Leung KK, Wheatley SC, Ng LJ, Zhou S, Ling KW, Sham MH, Koopman P, Tam PP, Cheah KS. SOX9 directly regulates the type-II collagen gene. Nat Genet. 16(2): 174-178, 1997.

Betsunoh H, Fukuda T, Anzai N, Nishihara D, Mizuno T, Yuki H, Masuda A, Yamaguchi Y, Abe H, Yashi M, Fukabori Y, Yoshida K, Kamai T. Increased expression of system large amino acid transporter (LAT)-1 mRNA is associated with invasive potential and unfavorable prognosis of human clear cell renal cell carcinoma. BMC Cancer. 13:509, 2013.

Bishnoi M, Jain A, Hurkat P, Jain SK. Chondroitin sulphate: a focus on osteoarthritis. Glycoconj J. 33(5): 693-705, 2016.

Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem. 72: 248-254, 1976.

Braun HJ and Gold GE. Diagnosis of osteoarthritis: imaging. Bone. 51(2): 278-288, 2012.

Breyer J, Samarin J, Rehm M, Lautscham L, Fabry B, Goppelt-Struebe M. Inhibition of Rho kinases increases directional motility of microvascular endothelial cells. Biochem Pharmacol. 83(5): 616-626, 2012.

Brighton CT and Heppenstall RB. Oxygen tension in zones of the epiphyseal plate, the metaphysis and diaphysis. An in vitro and in vivo study in and rabbits. J Bone Joint Surg Am. 53(4): 719-728, 1971.

Brittberg M, Lindahl A, Nilsson A, Ohlsson C, Isaksson O, Peterson L: Treatment of deep cartilage defects in the knee with autologous chondrocyte transplantation. N Engl J Med. 331(14): 889-895, 1994.

Brittberg M and Winalski CS. Evaluation of cartilage injuries and repair. J Bone Joint Surg Am. 85-A (Suppl 2): 58-69, 2003.

69

Burton-Wurster N, Liu W, Matthews GL, Lust G, Roughley PJ, Glant TT, Cs- Szabó G. TGF beta 1 and biglycan, decorin, and fibromodulin metabolism in canine cartilage. Osteoarthritis Cartilage. 11(3): 167-176, 2003.

Cai S and Exton JH. Determination of interaction sites of phospholipase D1 for RhoA. Biochem J. 355(Pt 3): 779-785, 2001.

Campanelli JT, Gayer GG, Scheller RH. Alternative RNA splicing that determines agrin activity regulates binding to heparin and alpha- dystroglycan. Development. 122(5): 1663-1672, 1996.

Carvalho RS, Yen EH, Suga DM. Glycosaminoglycan synthesis in the rat articular disk in response to mechanical stress. Am J Orthod Dentofacial Orthop. 107(4): 401-410, 1995.

Castañeda S, Roman-Blas JA, Largo R, Herrero-Beaumont G. Subchondral bone as a key target for osteoarthritis treatment. Biochem Pharmacol. 83(3): 315-323, 2012.

Cau J and Hall A. Cdc42 controls the polarity of the actin and microtubule through two distinct signal transduction pathways. J Cell Sci. 118(Pt 12): 2579-2587, 2005.

Chakravarti S and Magnuson T. Localization of mouse lumican (keratan sulfate proteoglycan) to distal chromosome 10. Mamm Genome. 6(5): 367- 368, 1995.

Chan KT, Bennin DA, Huttenlocher A. Regulation of adhesion dynamics by calpain-mediated proteolysis of focal adhesion kinase (FAK). J Biol Chem. 285(15): 11418-11426, 2010.

Chandran PL and Horkay F. Aggrecan, an unusual polyelectrolyte: review of solution behavior and physiological implications. Acta Biomater. 8(1): 3-12, 2012.

Chao PH, West AC, Hung CT. Chondrocyte intracellular calcium, cytoskeletal organization, and gene expression responses to dynamic osmotic loading. Am J Physiol Cell Physiol. 291(4): C718-C725, 2006.

Chapman S, McDermott DH, Shen K, Jang MK, McBride AA. The effect of Rho kinase inhibition on long-term keratinocyte proliferation is rapid and conditional. Stem Cell Res Ther. 5(2): 60, 2014.

70

Chirivino D, Del Maestro L, Formstecher E, Hupé P, Raposo G, Louvard D, Arpin M. The ERM proteins interact with the HOPS complex to regulate the maturation of endosomes. Mol Cell Biol. 22(3): 375-385, 2011.

Clark AL, Votta BJ, Kumar S, Liedtke W, Guilak F. Chondroprotective role of the osmotically sensitive ion channel transient receptor potential vanilloid 4: age- and sex-dependent progression of osteoarthritis in Trpv4-deficient mice. Arthritis Rheum. 62(10): 2973-2983, 2010.

Clegg DO, Reda DJ, Harris CL, Klein MA, O’Dell JR, Hooper MM, Bradley JD, Bingham CO 3rd, Weisman MH, Jackson CG, Lane NE, Cush JJ, Moreland LW, Schumacher HR Jr, Oddis CV, Wolfe F, Molitor JA, Yocum DE, Schnitzer TJ, Furst DE, Sawitzke AD, Shi H, Brandt KD, Moskowitz RW, Williams HJ. Glucosamine, chondroitin sulfate, and the two in combination for painful knee osteoarthritis. N Engl J Med. 354(8): 795-808, 2006.

Cormerais Y, Giuliano S, LeFloch R, Front B, Durivault J, Tambutté E, Massard PA, de la Ballina LR, Endou H, Wempe MF, Palacin M, Parks SK, Pouyssegur J. Genetic disruption of the multifunctional CD98/LAT1 complex demonstrates the key role of essential amino acid transport in the control of mTORC1 and tumor growth. Cancer Res. 76(15): 4481-4492, 2016.

Correa D, Somoza RA, Lin P, Greenberg S, Rom E, Duesler L, Welter JF, Yayon A, Caplan AI. Sequential exposure to fibroblast growth factors (FGF) 2, 9 and 18 enhances hMSC chondrogenic differentiation. Osteoarthritis Cartilage. 23(3): 443-453, 2015.

Cortesio CL, Boateng LR, Piazza TM, Bennin DA, Huttenlocher A. Calpain- mediated proteolysis of paxillin negatively regulates focal adhesion dynamics and cell migration. J Biol Chem. 286(12): 9998-10006, 2011.

Costell M, Gustafsson E, Aszódi A, Mörgelin M, Bloch W, Hunziker E, Addicks K, Timpl R, Fässler R. Perlecan maintains the integrity of cartilage and some basement membranes. J Cell Biol. 147(5): 1109-1122, 1999.

Coyle CH, Izzo NJ, Chu CR. Sustained hypoxia enhances chondrocyte matrix synthesis. J Orthop Res. 27(6): 793-799, 2009.

Croft D, Mundo AF, Haw R, Milacic M, Weiser J, Wu G, Caudy M, Garapati P, Gillespie M, Kamdar MR, Jassal B, Jupe S, Matthews L, May B, Palatnik S, Rothfels K, Shamovsky V, Song H, Williams M, Birney E, Hermjakob H, Stein L, D’Eustachio P. The Reactome pathway knowledgebase. Nucleic Acids Res. 42(D1):D472-D477 2014.

71

Cui D, Li L, Lou H, Sun H, Ngai SM, Shao G, Tang J. The ribosomal protein S26 regulates p53 activity in response to DNA damage. Oncogene. 33(17): 2225-2235, 2014.

Cui L, Wu Y, Cen L, Zhou H, Yin S, Liu G, Liu W, Cao Y. Repair of articular cartilage defect in non-weight bearing areas using adipose derived stem cells loaded polyglycolic acid mesh. Biomaterials. 30(14): 2683-2693, 2009.

Diaz-Romero J, Quintin A, Schoenholzer E, Pauli C, Despont A, Zumstein MA, Kohl S, Nesic D. S100A1 and S100B expression patterns identify differentiation status of human articular chondrocytes. J Cell Physiol. 229(8): 1106-1117, 2014.

Dieppe PA and Lohmander LS. Pathogenesis and management of pain in osteoarthritis. Lancet. 365(9463): 965-973, 2005.

Echtermeyer F, Bertrand J, Dreier R, Meinecke I, Neugebauer K, Fuerst M, Lee YJ, Song YW, Herzog C, Theilmeier G, Pap T. Syndecan-4 regulates ADAMTS-5 activation and cartilage breakdown in osteoarthritis. Nat Med. 15(9): 1072-1076, 2009. da Costa BR, Reichenbach S, Keller N, Nartey L, Wandel S, Jüni P, Trelle S. Effectiveness of non-steroidal anti-inflammatory drugs for the treatment of pain in knee and hip osteoarthritis: a network meta-analysis. Lancet. 387(10033): 2093-2105, 2016.

Dawe HR, Minamide LS, Bamburg JR, Cramer LP. ADF/cofilin controls cell polarity during fibroblast migration. Curr Biol. 13(3): 252-257, 2003.

De Witt MT, Handley CJ, Oakes BW, Lowther DA. In vitro response of chondrocytes to mechanical loading. The effect of short term mechanical tension. Connect Tissue Res. 12(2): 97-109, 1984.

Di Martino A, Kon E, Perdisa F, Sessa A, Filardo G, Neri MP, Bragonzoni L, Marcacci M. Surgical treatment of early knee osteoarthritis with a cell-free osteochondral scaffold: results at 24 months of follow-up. Injury. 46 (Suppl 8): S33-38, 2015.

Dijkgraaf LC, de Bont LG, Boering G, Liem RS. The structure, biochemistry, and metabolism of osteoarthritic cartilage: a review of the literature. J Oral Maxillofac Surg. 53(10): 1182-1192, 1995.

72

Ding B, Kirkiles-Smith NC, Pober JS. FOXO3a regulates oxygen-responsive expression of tumor necrosis factor receptor 2 in human dermal microvascular endothelial cells. J Biol Chem. 284(29): 19331-19339, 2009.

Doi H, Nishida K, Yorimitsu M, Komiyama T, Kadota Y, Tetsunaga T, Yoshida A, Kubota S, Tagikawa M, Ozaki T. Interleukin-4 downregulates the cyclic tensile stress-induced matrix metalloproteinases-13 and cathepsin B expression by rat normal chondrocytes. Acta Med Okayama. 62(2): 119-126, 2008.

Donato R, Cannon BR, Sorci G, Riuzzi F, Hsu K, Weber DJ, Geczy CL. Functions of S100 proteins. Curr Mol Med. 13(1): 24-57, 2013.

Duarte Campos DF, Drescher W, Rath B, Tingart M, Fischer H. Supporting Biomaterials for Articular Cartilage Repair. Cartilage. 3(3): 205-221, 2012.

Dumbauld DW, Shin H, Gallant ND, Michael KE, Radhakrishna H, García AJ. Contractility modulates cell adhesion strengthening through focal adhesion kinase and assembly of vinculin-containing focal adhesions. J Cell Physiol. 223(3): 746-756, 2010.

Duval E, Leclercq S, Elissalde JM, Demoor M, Galéra P, Boumédiene K. Hypoxia-inducible factor 1alpha inhibits the fibroblast-like markers type I and type III collagen during hypoxia-induced chondrocyte redifferentiation: hypoxia not only induces type II collagen and aggrecan, but it also inhibits type I and type III collagen in the hypoxia-inducible factor 1alpha-dependent redifferentiation of chondrocytes. Arthritis Rheum. 60(10): 3038-3048, 2009.

Elder BD and Athanasiou KA. Hydrostatic pressure in articular cartilage tissue engineering: from chondrocytes to tissue regeneration. Tissue Eng Part B Rev. 15(1): 43-53, 2009a.

Elder BD and Athanasiou KA. Effects of temporal hydrostatic pressure on tissue-engineered bovine articular cartilage constructs. Tissue Eng Part A. 15(5): 1151-1158, 2009b.

Eldridge S, Nalesso G, Ismail H, Vicento-Greco K, Kabouridis P, Ramachandran M, Niemeier A, Herz J, Pitzalis C, Perretti M, Dell’Accio F. Agrin mediates chondrocyte homeostasis and requires both LRP4 and α- dystroglycan to enhance cartilage formation in vitro and in vivo. Ann Rheum Dis. 75(6): 1228-1235, 2016.

73

Elenius K and Jalkanen M. Function of the syndecans-a family of cell surface proteoglycans. J Cell Sci. 107(Pt 11): 2975-2982, 1994.

Ellman MB, Yan D, Ahmadinia K, Chen D, An HS, Im HJ. Fibroblast growth factor control of cartilage homeostasis. J Cell Biochem. 114(4): 735-742, 2013.

Eyre D. Collagen of articular cartilage. Arthritis Res. 4(1): 30-35, 2002.

Fabregat A, Sidiropoulos K, Garapati P, Gillespie M, Hausmann K, Haw R, Jassal B, Jupe S, Korninger F, McKay S, Matthews L, May B, Milacic M, Rothfels K, Shamovsky V, Webber M, Weiser J, Williams M, Wu G, Stein L, Hermjakob H, D’Eustachio P. The Reactome pathway Knowledgebase. Nucleic Acids Res. 44(D1): D481-487, 2016.

Falah M, Nierenberg G, Soudry M, Hayden M, Volpin G. Treatment of articular cartilage lesions of the knee. Int Orthop. 34(5): 621-630, 2010.

Fanning PJ, Emkey G, Smith RJ, Grodzinsky AJ, Szasz N, Trippel SB. Mechanical regulation of mitogen-activated protein kinase signaling in articular cartilage. J Biol Chem. 278(51): 50940-50948, 2003.

Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and discrimination of sulphated glycosaminoglycans by use of dimethylmethylene blue. Biochim Biophys Acta. 883(2): 173-177, 1986.

Fíla J and Honys D. Enrichment techniques employed in phosphoproteomics. Amino Acids. 43(3): 1025-1047, 2012.

Fraser JR, Laurent TC and Laurent UB. Hyaluronan: its nature, distribution, functions and turnover. J Intern Med. 242(1): 27-33, 1997.

Ferdous Z, Wei VM, Iozzo R, Höök M, Grande-Allen KJ. Decorin- transforming growth factor- interaction regulates matrix organization and mechanical characteristics of three-dimensional collagen matrices. J Biol Chem. 282(49): 35887-35898, 2007.

Fermor B, Christensen SE, Youn I, Cernanec JM, Davies CM, Weinberg JB. Oxygen, nitric oxide and articular cartilage. Eur Cell Mater. 13: 56-65, 2007.

Fermor B, Gurumurthy A, Diekman BO. Hypoxia, RONS and energy metabolism in articular cartilage. Osteoarthritis Cartilage. 18(9): 1167-1173, 2010.

74

Filardo G, Vannini F, Marcacci M, Andriolo L, Ferruzzi A, Giannini S, Kon E. Matrix-assisted autologous chondrocyte transplantation for cartilage regeneration in osteoarthritic knees: results and failures at midterm follow- up. Am J Sports Med. 41(1): 95-100, 2013.

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

Fotiadis D, Kanai Y, Palacín M. The SLC3 and SLC7 families of amino acid transporters. Mol Aspects Med. 34(2-3): 139-158, 2013.

Fotovati A, Abu-Ali S, Kage M, Shirouzu K, Yamana H, Kuwano M. N-myc downstream-regulated gene 1 (NDRG1) a differentiation marker of human breast cancer. Pathol Oncol Res. 17(3): 525-533, 2011.

Freitag J, Bates D, Boyd R, Shah K, Barnard A, Huguenin L, Tenen A. Mesenchymal stem cell therapy in the treatment of osteoarthritis: reparative pathways, safety and efficacy - a review. BMC Musculoskelet Disord. 17: 230, 2016.

Fujisawa T, Hattori T, Takahashi K, Kuboki T, Yamashita A, Takigawa M. Cyclic mechanical stress induces extracellular matrix degradation in cultured chondrocytes via gene expression of matrix metalloproteinases and interleukin-1. J Biochem. 125(5): 966-975, 1999.

Fukuda K, Asada S, Kumano F, Saitoh M, Otani K, Tanaka S. Cyclic tensile stretch on bovine articular chondrocytes inhibits protein kinase C activity. J Lab Clin Med. 130(2): 209-215, 1997.

Funderburgh JL. Keratan sulfate biosynthesis. IUBMB Life. 54(4): 187-194, 2002.

Funderburgh JL. Keratan sulfate: structure, biosynthesis, and function. Glycobiology. 10(10): 951-958, 2000.

Furumatsu T, Matsumoto-Ogawa E, Tanaka T, Lu Z, Ozaki T. ROCK inhibition enhances aggrecan deposition and suppresses matrix metalloproteinase-3 production in human articular chondrocytes. Connect Tissue Res. 55(2): 89-95, 2014.

Furuya M, Horiguchi J, Nakajima H, Kanai Y, Oyama T. Correlation of L- type amino acid transporter 1 and CD98 expression with triple negative breast cancer prognosis. Cancer Sci. 103(2): 382-389, 2012.

75

Fülöp C, Walcz E, Valyon M, Glant TT. Expression of alternatively spliced epidermal growth factor-like domains in of different species. Evidence for a novel module. J Biol Chem. 268(23): 17377-17383, 1993.

Gandhi NS and Mancera RL. The structure of glycosaminoglycans and their interactions with proteins. Chem Biol Drug Des. 72(6): 455-482, 2008.

García-Campos MA, Espinal-Enríquez J, Hernández-Lemus E. Pathway Analysis: State of the Art. Front Physiol. 6: 383, 2015.

Garcia-Mata R, Boulter E, Burridge K. The 'invisible hand': regulation of RHO GTPases by RHOGDIs. Nat Rev Mol Cell Biol. 12(8): 493-504, 2011.

Garg HG, Yu L, Hales CA, Toida T, Islam T, Linhardt RJ. Sulfation patterns in heparin and heparan sulfate: effects on the proliferation of bovine pulmonary artery smooth muscle cells. Biochim Biophys Acta. 1639(3): 225- 231, 2003.

Garnock-Jones KP. Ripasudil: first global approval. Ripasudil. Drugs. 74(18): 2211-2215, 2014.

Gavénis K, Kremer A, Von Walter M, Hollander DA, Schneider U, Schmidt- Rohlfing B. Effects of cyclic hydrostatic pressure on the metabolism of human osteoarthritic chondrocytes cultivated in a collagen gel. Artif Organs. 31(2): 91-98, 2007.

Gentile P, Chiono V, Carmagnola I, Hatton PV. An overview of poly(lactic-co- glycolic) acid (PLGA)-based biomaterials for bone tissue engineering. Int J Mol Sci. 15(3): 3640-3659, 2014.

Gibson JS, Milner PI, White R, Fairfax TP, Wilkins RJ. Oxygen and reactive oxygen species in articular cartilage: modulators of ionic homeostasis. Pflugers Arch. 455(4): 563-573, 2008.

Goetsch KP, Snyman C, Myburgh KH, Niesler CU. ROCK-2 is associated with focal adhesion maturation during myoblast migration. J Cell Biochem. 115(7): 1299-1307, 2014.

Goldring MB and Otero M. Inflammation in osteoarthritis. Curr Opin Rheumatol. 23(5): 471-478, 2011.

76

Góral A, Bieganowski P, Prus W, Krzemień-Ojak Ł, Kądziołka B, Fabczak H, Filipek A. Calcyclin binding protein/Siah-1 interacting protein is a Hsp90 binding chaperone. PLoS One. 11(6): e0156507, 2016.

Gould SE, Upholt WB and Kosher RA. 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 USA. 89(8): 3271-3275, 1992.

Grimshaw MJ and Mason RM. Bovine articular chondrocyte function in vitro depends upon oxygen tension. Osteoarthritis Cartilage. 8(5): 386-392, 2000.

Grover J, Chen XN, Korenberg JR, Roughley PJ. The human lumican gene. Organization, chromosomal location, and expression in articular cartilage. J Biol Chem. 270(37): 21942-21949, 1995.

Grover J and Roughley PJ. Expression of cell-surface proteoglycan mRNA by human articular chondrocytes. Biochem J. 309(Pt 3): 963-968, 1995.

Grover J and Roughley PJ. Versican gene expression in human articular cartilage and comparison of mRNA splicing variation with aggrecan. Biochem J. 291(Pt 2): 361-367, 1993.

Gudas R, Kalesinskas RJ, Kimtys V, Stankevicius E, Toliusis V, Bernotavicius G, Smailys A. A prospective randomized clinical study of mosaic osteochondral autologous transplantation versus microfracture for the treatment of osteochondral defects in the knee joint in young athletes. Arthroscopy. 21(9): 1066-1075, 2005.

Guilak F. Compression-induced changes in the shape and volume of the chondrocyte nucleus. J Biomech. 28(12): 1529-1541, 1995.

Guilluy C, Garcia-Mata R, Burridge K. Rho protein crosstalk: another social network? Trends Cell Biol 21(12): 718-726, 2011.

Hall AC, Urban JP, Gehl KA. The effects of hydrostatic pressure on matrix synthesis in articular cartilage. J Orthop Res. 9(1): 1-10, 1991.

Hall BK and Miyake T. All for one and one for all: condensations and the initiation of skeletal development. Bioessays. 22(2): 138-147, 2000.

77

Hambly K, Bobic V, Wondrasch B, Van Assche D, Marlovits S. Autologous chondrocyte implantation postoperative care and rehabilitation: science and practice. Am J Sports Med. 34(6): 1020-1038, 2006.

Han YC, Lim Y, Duffieldl MD, Li H, Liu J, Abdul Manaph NP, Yang M, Keating DJ, Zhou XF. Direct reprogramming of mouse fibroblasts to neural stem cells by small molecules. Stem Cells Int. 2016, doi: 10.1155/2016/4304916.

Hannan MT, Felson DT, Pincus T. Analysis of the discordance between radiographic changes and knee pain in osteoarthritis of the knee. J Rheumatol. 27(6): 1513-1517, 2000.

Hara S, Hamada J, Kobayashi C, Kondo Y, Imura N. Expression and characterization of hypoxia-inducible factor (HIF)-3alpha in human kidney: suppression of HIF-mediated gene expression by HIF-3alpha. Biochem Biophys Res Commun. 287(4): 808-813, 2001.

Hardingham TE and Fosang AJ. Proteoglycans: many forms and many functions. FASEB J. 6(3): 861-870, 1992.

Haudenschild DR, Chen J, Pang N, Lotz MK, D’Lima DD. Rho kinase- dependent activation of SOX9 in chondrocytes. Arthritis Rheum. 62(1): 191- 200, 2010.

Haudenschild DR, Nguyen B, Chen J, D’Lima DD, Lotz MK. Rho kinase- dependent CCL20 induced by dynamic compression of human chondrocytes. Arthritis Rheum. 58(9): 2735-2742, 2008.

Hayrabedyan S, Kyurkchiev S, Kehayov I. FGF-1 and S100A13 possibly contribute to angiogenesis in endometriosis. J Reprod Immunol. 67(1-2): 87- 101, 2005.

Hdud IM, Mobasheri A, Loughna PT. Effects of cyclic equibiaxial mechanical stretch on α-BK and TRPV4 expression in equine chondrocytes. Springerplus. 3: 59, 2014.

Hedlund H, Mengarelli-Widholm S, Reinholt FP, Svensson O. Stereologic studies on collagen in bovine articular cartilage. APMIS. 101(2): 133-140, 1993.

78

Heikkilä M, Pasanen A, Kivirikko KI, Myllyharju J. Roles of the human hypoxia-inducible factor (HIF)-3α variants in the hypoxia response. Cell Mol Life Sci. 68(23): 3885-3901, 2011.

Hochberg MC, Martel-Pelletier J, Monfort J, Möller I, Castillo JR, Arden N, Berenbaum F, Blanco FJ, Conaghan PG, Doménech G, Henrotin Y, Pap T, Richette P, Sawitzke A, du Souich P, Pelletier JP, MOVES Investigation group. Combined chondroitin sulfate and glucosamine for painful knee osteoarthritis: a multicentre, randomised, double-blind, non-inferiority trial versus celecoxib. Ann Rheum Dis. 75(1): 37-44, 2016.

Honda K, Ohno S, Tanimoto K, Iljuin C, Tanaka N, Doi T, Kato Y, Tanne K. The effects of high magnitude cyclic tensile load on cartilage matrix metabolism in cultured chondrocytes. Eur J Cell Biol. 79(9): 601-609, 2000.

Hopkins AM, Pineda AA, Winfree LM, Brown GT, Laukoetter MG, Nusrat A. Organized migration of epithelial cells requires control of adhesion and protrusion through Rho kinase effectors. Am J Physiol Gastrointest Liver Physiol. 292(3): G806-G817, 2003.

Hu CJ, Wang LY, Chodosh LA, Keith B, Simon MC. Differential roles of hypoxia-inducible factor 1alpha (HIF-1alpha) and HIF-2alpha in hypoxic gene regulation. Mol Cell Biol. 23(24): 9361-9374, 2003.

Hu JC and Athanasiou KA. The effects of intermittent hydrostatic pressure on self-assembled articular cartilage constructs. Tissue Eng. 12(5): 1337- 1344, 2006.

Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID Bioinformatics Resources. Nat Protoc. 4(1): 44-57, 2009a.

Huang da W, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 37(1): 1-13, 2009b.

Huang J, Ballou LR, Hasty KA. Cyclic equibiaxial tensile strain induces both anabolic and catabolic responses in articular chondrocytes. Gene. 404(1-2): 101-109, 2007.

79

Huang LE, Gu J, Schau M, Bunn HF. Regulation of hypoxia-inducible factor 1alpha is mediated by an O2-dependent degradation domain via the ubiquitin-proteasome pathway. Proc Natl Acad Sci U S A. 95(14): 7987-7992, 1998.

Hunter DJ, McDougall JJ, Keefe FJ. The symptoms of osteoarthritis and the genesis of pain. Rheum Dis Clin North Am. 34(3): 623-643, 2008.

Hunziker EB, Quinn TM, Häuselmann HJ. Quantitative structural organization of normal adult human articular cartilage. Osteoarthritis Cartilage. 10(7): 564-572, 2002.

Ikenoue T, Trindade MC, Lee MS, Lin EY, Schurman DJ, Goodman SB, Smith RL. Mechanoregulation of human articular chondrocyte aggrecan and type II collagen expression by intermittent hydrostatic pressure in vitro. J Orthop Res. 21(1): 110-116, 2003.

Imamura F, Mukai M, Ayaki M, Akedo H. Y-27632, an inhibitor of rho- associated protein kinase, suppresses tumor cell invasion via regulation of focal adhesion and focal adhesion kinase. Jpn J Cancer Res. 91(8): 811-816, 2000.

Iozzo RV and Schaefer L. Proteoglycan form and function: A comprehensive nomenclature of proteoglycans. Matrix Biol. 42: 11-55, 2015.

Ishizaki T, Uehata M, Tamechika I, Keel J, Nonomura K, Maekawa M, Narumiya S. Pharmacological properties of Y-27632, a specific inhibitor of rho-associated kinases. Mol Pharmacol. 57(5): 976-983, 2000.

Itano N, Sawai T, Yoshida M, Lenas P, Yamada Y, Imagawa M, Shinomura T, Hamaguchi M, Yoshida Y, Ohnuki Y, Miyauchi S, Spicer AP, McDonald JA, Kimata K. Three isoforms of mammalian hyaluronan synthases have distinct enzymatic properties. J Biol Chem. 274(35): 25085-25092, 1999.

Jaakkola P, Mole DR, Tian YM, Wilson MI, Gielbert J, Gaskell SJ, von Kriegsheim A, Hebestreit HF, Mukherji M, Schofield CJ, Maxwell PH, Pugh CW, Ratcliffe CJ. Targeting of HIF-alpha to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation. Science. 292(5516): 468-472, 2001.

Jacobi M, Villa V, Magnussen RA, Neyret P. MACI - a new era? Sports Med Arthrosc Rehabil Ther Technol. 3(1): 10, 2011.

80

Jakob RP, Franz T, Gautier E, Mainil-Varlet P: Autologous osteochondral grafting in the knee: indication, results, and reflections. Clin Orthop Relat Res. 401: 170-184, 2002.

Jiang L, Wen J, Luo W. Rho-associated kinase inhibitor, Y-27632, inhibits the invasion and proliferation of T24 and 5367 bladder cancer cells. Mol Med Rep. 12(5): 7526-7530, 2015.

Jin M, Frank EH, Quinn TM, Hunziker EB, Grodzinsky AJ. Tissue shear deformation stimulates proteoglycan and protein biosynthesis in bovine cartilage explants. Arch Biochem Biophys. 395(1): 41-48, 2001.

Jo CH, Lee YG, Shin WH, Kim H, Chai JW, Jeong EC, Kim JE, Shim H, Shin JS, Shin IS, Ra JC, Oh S, Yoon KS. Intra-articular injection of mesenchymal stem cells for the treatment of osteoarthritis of the knee: a proof-of-concept clinical trial. Stem Cells. 32(5): 1254-1266, 2014.

Johnson HJ, Rosenberg L, Choi HU, Garza S, Höök M, Neame PJ. Characterization of epiphycan, a small proteoglycan with a leucine-rich repeat core protein. J Biol Chem. 272(30): 18709-18717, 1997.

Kaira K, Oriuchi N, Takahashi T, Nakagawa K, Ohde Y, Okumura T, Murakami H, Shukuya T, Kenmotsu H, Naito T, Kanai Y, Endo M, Kondo H, Nakajima T, Yamamoto N. L-type amino acid transporter 1 (LAT1) expression in malignant pleural mesothelioma. Anticancer Res. 31(12): 4075- 4082, 2011.

Kaira K, Sunose Y, Arakawa K, Ogawa T, Sunaga N, Shimizu K, Tominaga H, Oriuchi N, Itoh H, Nagamori S, Kanai Y, Segawa A, Furuya M, Mori M, Oyama T, Takeyoshi I. Prognostic significance of L-type amino-acid transporter 1 expression in surgically resected pancreatic cancer. Br J Cancer. 107(4): 632-638, 2012.

Kaitsuka T and Matsushita M. Regulation of translation factor EEF1D gene function by alternative splicing. Int J Mol Sci. 16(2): 3970-3979, 2015.

Kang ML and Im GI. Drug delivery systems for intra-articular treatment of osteoarthritis. Expert Opin Drug Deliv. 11(2): 269-282, 2014.

Karjalainen HM, Sironen RK, Elo MA, Kaarniranta K, Takigawa M, Helminen HJ, Lammi MJ. Gene expression profiles in chondrosarcoma cells subjected to cyclic stretching and hydrostatic pressure. A cDNA array study. Biorheology. 40(1-3): 93-100, 2003.

81

Karsdal MA, Michaelis M, Ladel C, Siebuhr AS, Bihlet AR, Andersen JR, Guehring H, Christiansen C, Bay-Jensen AC, Kraus VB. Disease-modifying treatments for osteoarthritis (DMOADs) of the knee and hip: lessons learned from failures and opportunities for the future. Osteoarthritis Cartilage. 24(12): 2013-2021, 2016.

Katz JN, Earp BE and Gomoll AH. Surgical management of osteoarthritis. Arthritis Care Res. 62(9): 1220-1228, 2010.

Kiani C, Chen L, Wu YJ, Yee AJ, Yang BB. Structure and function of aggrecan. Cell Res. 12(1): 19-32, 2002.

Kim CW, Goldberger OA, Gallo RL, Bernfield M. Members of the syndecan family of heparan sulfate proteoglycans are expressed in distinct cell-, tissue- and development-specific patterns. Mol Biol Cell. 5(7): 797-805, 1994.

Kim SG, Kim HH, Kim HK, Kim CH, Chun HS, Kanai Y, Endou H, Kim DK. Differential expression and functional characterization of system L amino acid transporters in human normal osteoblast cells and osteogenic sarcoma cells. Anticancer Res. 26(3A): 1989-1996, 2006.

Kimura K, Ito M, Amano M, Chihara K, Fukata Y, Nakafuku M, Yamamori B, Feng J, Nakano T, Okawa K, Iwamatsu A, Kaibuchi K. Regulation of myosin phosphatase by Rho and Rho-associated kinase (Rho-kinase). Science. 273(5272): 245-248, 1996.

Kloc M, Li XC, Ghobrial RM. RhoA cytoskeletal pathway to transplantation. J Immunol Clin Res. 2(1): 1012, 2014.

Knight MM, Toyoda T, Lee DA, Bader DL. Mechanical compression and hydrostatic pressure induce reversible changes in actin cytoskeletal organisation in chondrocytes in agarose. J Biomech. 39(8): 1547-1551, 2006.

Knutsen G, Engebretsen L, Ludviqsen TC, Drogset JO, Grøntvedt T, Solheim E, Strand T, Roberts S, Isaksen V, Johansen O. Autologous chondrocyte implantation compared with microfracture in the knee. A randomized trial. J Bone Joint Surg Am. 86-A(3): 455-464, 2004.

Koay EJ and Athanasiou KA. Hypoxic chondrogenic differentiation of human embryonic stem cells enhances cartilage protein synthesis and biomechanical functionality. Osteoarthritis Cartilage. 16(12): 1450-1456, 2008.

82

Kobe B and Deisenhofer J. The leucine-rich repeat: a versatile binding motif. Trends Biochem Sci. 19(10): 415-421, 1994.

Koh YG, Jo SB, Kwon OR, Suh DS, Lee SW, Park SH, Choi YJ. Mesenchymal stem cell injections improve symptoms of knee osteoarthritis. Arthroscopy. 29(4): 748-755, 2013.

Koivunen P, Salo KE, Myllyharju J, Ruddock LW. Three binding sites in protein-disulfide isomerase cooperate in collagen prolyl 4-hydroxylase tetramer assembly. J Biol Chem. 280(7): 5227-5235, 2004.

Kokenyesi R and Bernfield M. Core protein structure and sequence determine the site and presence of heparan sulfate and chondroitin sulfate on syndecan-1. J Biol Chem. 269(16): 12304-12309, 1994.

Kon E, Filardo G, Di Matteo B, Perdisa F, Marcacci M. Matrix assisted autologous chondrocyte transplantation for cartilage treatment: A systematic review. Bone Joint Res. 2(2): 18-25, 2013.

Kopakkala-Tani M, Elo MA, Sironen RK, Helminen HJ, Lammi MJ. High hydrostatic pressure induces ERK and PI3 kinase phosphorylation in human HCS-2/8 chondrosarcoma cells. Cell Mol Biol. 50(4): 485-490, 2004.

Koshi H, Sano T, Handa T, Yanagawa T, Saitou K, Nagamori S, Kanai Y, Takagishi K, Oyama T. L-type amino acid transporter-1 and CD98 expression in bone and soft tissue tumors. Pathol Int. 65(9): 460-467, 2015.

Kotani H, Takaishi K, Sasaki T, Takai Y. Rho regulates association of both the ERM family and vinculin with the plasma membrane in MDCK cells. Oncogene. 14(14): 1705-1713, 1997.

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

Krusius T, Finne J, Margolis RK, Margolis RU. Identification of an O- glycosidic mannose-linked sialylated tetrasaccharide and keratan sulfate oligosaccharides in the chondroitin sulfate proteoglycan of brain. J Biol Chem. 261(18): 8237-8242, 1986.

Krych AJ, Harnly HW, Rodeo SA, Williams RJ 3rd: Activity levels are higher after osteochondral autograft transfer mosaicplasty than after microfracture for articular cartilage defects of the knee: aretrospective comparative study. J Bone Joint Surg Am. 94(11): 971-978, 2012.

83

Kumar D and Lassar AB. The transcriptional activity of Sox9 in chondrocytes is regulated by RhoA signaling and actin polymerization. Mol Cell Biol. 29(15): 4262-4273, 2009.

Kurita T, Yamamura H, Suzuki Y, Giles WR, Imaizumi Y. The ClC-7 chloride channel is downregulated by hypoosmotic stress in human chondrocytes. Mol Pharmacol. 88(1): 113-120, 2015.

Lafont JE, Talma S, Hopfgarten C, Murphy CL. Hypoxia promotes the differentiated human articular chondrocyte phenotype through SOX9- dependent and -independent pathways. J Biol Chem. 283(8): 4778-4786, 2008.

Lammi MJ, Inkinen R, Parkkinen JJ, Häkkinen T, Jortikka M, Nelimarkka LO, Järveläinen HT, Tammi MI. Expression of reduced amounts of structurally altered aggrecan in articular cartilage chondrocytes exposed to high hydrostatic pressure. Biochem J. 304(Pt 3): 723-730, 1994.

Lambrecht S, Dhaenens M, Almqvist F, Verdonk P, Verbruggen G, Deforce D, Elewaut D. Proteome characterization of human articular chondrocytes leads to novel insights in the function of small heat-shock proteins in chondrocyte homeostasis. Osteoarthritis Cartilage. 18(3): 440-446, 2010.

Lane Smith R, Trindade MC, Ikenoue T, Mohtai M, Das P, Carter DR, Goodman SB, Schurman DJ. Effects of shear stress on articular chondrocyte metabolism. Biorheology. 37(1-2): 95-107, 2000.

Laurent TC, Laurent UB, Fraser JR. The structure and function of hyaluronan: An overview. Immunol Cell Biol. 74(2): A1-A7, 1996.

Le Clainche C and Carlier MF. Regulation of actin assembly associated with protrusion and adhesion in cell migration. Physiol Rev. 88(2): 489-513, 2008.

Le Clainche C, Dwivedi SP, Didry D, Carlier MF. Vinculin is a dually regulated actin filament barbed end-capping and side-binding protein. J Biol Chem. 285(30): 23420-23432, 2010.

Lee CY, Liu X, Hsu HC, Wang DY, Luo ZP. A modified cell culture method for autologous chondrocyte transplantation. Connect Tissue Res. 46(2):93- 99, 2005.

84

Lee HH, Chang CC, Shieh MJ, Wang JP, Chen YT, Young TH, Hung SC. Hypoxia enhances chondrogenesis and prevents terminal differentiation through PI3K/Akt/FoxO dependent anti-apoptotic effect. Sci Rep. 3: 2683, 2013.

Lee MS, Ikenoue T, Trindade MC, Wong N, Goodman SB, Schurman DJ, Smith RL. Protective effects of intermittent hydrostatic pressure on osteoarthritic chondrocytes activated by bacterial endotoxin in vitro. J Orthop Res. 21(1): 117-122, 2003.

Lee RB and Urban JB. Evidence for a negative Pasteur effect in articular cartilage. Biochem J. 321(Pt 1): 95-102, 1997.

Lefebvre V and Dvir-Ginzberg M. SOX9 and the many facets of its regulation in the chondrocyte lineage. Connect Tissue Res. 58(1): 2-14, 2017.

Lewis R, Feetham CH, Barrett-Jolley R. Cell volume regulation in chondrocytes. Cell Physiol Biochem. 28(6): 1111-1122, 2011.

Li TF, O’Keefe RJ, Chen D. TGF-beta signaling in chondrocytes. Front Biosci. 10: 681-688, 2005.

Li Y, Zhou J, Yang X, Jiang Y, Gui J. Intermittent hydrostatic pressure maintains and enhances the chondrogenic differentiation of cartilage progenitor cells cultivated in alginate beads. Dev Growth Differ. 58(2): 180- 193, 2016.

Lin C, McGough R, Aswad B, Block JA, Terek R. Hypoxia induces HIF- 1alpha and VEGF expression in chondrosarcoma cells and chondrocytes. J Orthop Res. 22(6): 1175-1181, 2004.

Liu H, Sadygov RG, Yates JR 3rd. A model for random sampling and estimation of relative protein abundance in shotgun proteomics. Anal Chem. 76(14): 4193-4201, 2004.

Liu Y, Shu XZ and Prestwich GD. Osteochondral defect repair with autologous bone marrow-derived mesenchymal stem cells in an injectable, in situ, cross-linked synthetic extracellular matrix. Tissue Eng. 12(12): 3405- 3416, 2006.

Loeser RF. Integrins and chondrocyte-matrix interactions in articular cartilage. Matrix Biol. 39: 11-16, 2014.

85

López-Casillas F, Payne HM, Andres JL, Massagué J. Betaglycan can act as a dual modulator of TGF-beta access to signaling receptors: mapping of ligand binding and GAG attachment sites. J Cell Biol. 124(4): 557-568, 1994.

Lund-Olesen K. Oxygen tension in synovial fluids. Arthritis Rheum. 13(6): 769-776, 1970.

MacCoss MJ, Wu CC, Liu H, Sadygov R, Yates JR 3rd. A correlation algorithm for the automated quantitative analysis of shotgun proteomics data. Anal Chem. 75(24): 6912-6921, 2003.

Madaule P, Eda M, Watanabe N, Fujisawa K, Matsuoka T, Bito H, Ishizaki T, Narumiya S. Role of citron kinase as a target of the small GTPase Rho in cytokinesis. Nature. 394(6692): 491-494, 1998.

Madhavan S, Anghelina M, Rath-Deschner B, Wypasek E, John A, Deschner J, Piesco N, Agarwal S. Biomechanical signals exert sustained attenuation of proinflammatory gene induction in articular chondrocytes. Osteoarthritis Cartilage. 14(10): 1023-1032, 2006.

Maekawa M, Ishizaki T, Boku S, Watanabe N, Fujita A, Iwamatsu A, Obinata T, Ohashi K, Mizuno K, Narumiya S. Signaling from Rho to the actin cytoskeleton through protein kinases ROCK and LIM-kinase. Science. 285(5429): 895-898, 1999.

Magdalena J, Millard TH, Machesky LM. Microtubule involvement in NIH 3T3 Golgi and MTOC polarity establishment. J Cell Sci. 116(Pt 4): 743-756, 2003.

Malmström A, Bartolini B, Thelin MA, Pacheco B, Maccarana M. Iduronic acid in chondroitin/dermatan sulfate: biosynthesis and biological function. J Histochem Cytochem. 60(12): 916-925, 2012.

Mann M and Wilm M. Error-tolerant identification of peptides in sequence databases by peptide sequence tags. Anal Chem. 66(24): 4390-4399, 1994.

Markway BD, Cho H, Zilberman-Rudenko J, Holden P, McAlinden A, Johnstone B. Hypoxia-inducible factor 3-alpha expression is associated with the stable chondrocyte phenotype. J Orthop Res. 33(11): 1561-1570, 2015.

86

Marshall AD, van Geldermalsen M, Otte NJ, Anderson LA, Lum T, Vellozzi MA, Zhang BK, Thoeng A, Wang Q, Rasko JE, Holst J. LAT1 is a putative therapeutic target in endometrioid endometrial carcinoma. Int J Cancer. 139(11): 2529-2539, 2016.

Martin K, Reimann A, Fritz RD, Ryu H, Jeon NL, Pertz O. Spatio-temporal co-ordination of RhoA, Rac1 and Cdc42 activation during prototypical edge protrusion and retraction dynamics. Sci Rep. 6: 21901, 2016.

Masuda T, Tomita M, Ishihama Y. Phase transfer surfactant-aided trypsin digestion for membrane proteome analysis. J Proteome Res. 7(2): 731-740, 2008.

Mawatari T, Lindsey DP, Harris AH, Goodman SB, Maloney WJ, Smith RL. Effects of tensile strain and fluid flow on osteoarthritic human chondrocyte metabolism in vitro. J Orthop Res. 28(7): 907-913, 2010.

Matsukawa M, Fukuda K, Yamasaki K, Yoshida K, Munakata H, Hamanishi C. Enhancement of nitric oxide and proteoglycan synthesis due to cyclic tensile strain loaded on chondrocytes attached to fibronectin. Inflamm Res. 53(6): 239-244, 2004.

Matsumoto E, Furumatsu T, Kanazawa T, Tamura M, Ozaki T. ROCK inhibitor prevents the dedifferentiation of human articular chondrocytes. Biochem Biophys Res Commun. 420(1): 124-129, 2012.

Matsumoto K, Kamiya N, Suwan K, Atsumi F, Shimizu K, Shinomura T, Yamada Y, Kimata K, Watanabe H. Identification and characterization of versican/PG-M aggregates in cartilage. J Biol Chem. 281(26): 18257-18263, 2006.

Matsumoto K, Shionyu M, Go M, Shimizu K, Shinomura T, Kimata K, Watanabe H. Distinct interaction of versican/PG-M with hyaluronan and link protein. J Biol Chem. 278(42): 41205-41212, 2003.

Matta C, Zhang X, Liddell S, Smith JR, Mobasheri A. Label-free proteomic analysis of the hydrophobic membrane protein complement in articular chondrocytes: a technique for identification of membrane biomarkers. Biomarkers. 20(8): 572-589, 2015.

87

Maurel P, Rauch U, Flad M, Margolis RK, Margolis RU. Phosphacan, a chondroitin sulfate proteoglycan of brain that interacts with neurons and neural cell-adhesion molecules, is an extracellular variant of a receptor-type protein tyrosine phosphatase. Proc Natl Acad Sci USA. 91(7): 2512-2516, 1994.

Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux EC, Cockman ME, Wykoff CC, Pugh CW, Maher ER, Ratcliffe PJ. The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature. 399(6733): 271-275, 1999.

Mazieres B, Combe B, Phan Van A, Tondut J, Grynfeltt M. Chondroitin sulfate in osteoarthritis of the knee: a prospective, double blind, placebo controlled multicenter clinical study. J Rheumatol. 28(1): 173-181, 2001.

Mendler M, Eich-Bender SG, Vaughan L, Winterhalter KH, Bruckner P. Cartilage contains mixed fibrils of collagen types II, IX, and XI. J Cell Biol. 108(1): 191-197, 1989.

Menkel AR, Kroemker M, Bubeck P, Ronsiek M, Nikolai G, Jockusch BM. Characterization of an F-actin-binding domain in the cytoskeletal protein vinculin. J Cell Biol. 126(5): 1231-1240, 1994.

Michael JW, Schlüter-Brust KU and Eysel P. The epidemiology, etiology, diagnosis, and treatment of osteoarthritis of the knee. Deutsch Arztebl Int. 107(9): 152-162, 2010.

Midwood KS and Salter DM. Expression of NG2/human melanoma proteoglycan in human adult articular chondrocytes. Osteoarthritis Cartilage. 6(5): 297-305, 1998.

Mifune Y, Matsumoto T, Takayama K, Ota S, Li H, Meszaros LB, Usas A, Nagamune K, Gharaibeh B, Fu FH, Huard J. The effect of platelet-rich plasma on the regenerative therapy of muscle derived stem cells for articular cartilage repair. Osteoarthritis Cartilage. 21(1): 175-185, 2013.

Mithoefer K, McAdams T, Williams RJ, Kreuz PC, Mandelbaum BR: Clinical efficacy of the microfracture technique for articular cartilage repair in the knee: an evidence-based systematic analysis. Am J Sports Med. 37(10): 2053-2063, 2009.

88

Miyanishi K, Trindade MC, Lindsey DP, Beaupré GS, Carter DR, Goodman SB, Schurman DJ, Smith RL. Effects of hydrostatic pressure and transforming growth factor-beta 3 on adult human mesenchymal stem cell chondrogenesis in vitro. Tissue Eng. 12(6): 1419-1428, 2006.

Mobasheri A, Lewis R, Maxwell JE, Hill C, Womack M, Barrett-Jolley R. Characterization of a stretch-activated potassium channel in chondrocytes. J Cell Physiol. 223(2): 511-518, 2010.

Morgan MR, Humphries MJ, Bass MD. Synergistic control of cell adhesion by integrins and syndecans. Nat Rev Mol Cell Biol. 8(12): 957-969, 2007.

Muramatsu S, Wakabayashi M, Ohno T, Amano K, Ooishi R, Sugahara T, Shiojiri S, Tashiro K, Suzuki Y, Nishimura R, Kuhara S, Sugano S, Yoneda T, Matsuda A. Functional gene screening system identified TRPV4 as a regulator of chondrogenic differentiation. J Biol Chem. 282(44): 32158- 32167, 2007.

Nakagawa O, Fujisawa K, Ishizaki T, Saito Y, Nakao K, Narumiya S. ROCK-I and ROCK-II, two isoforms of Rho-associated coiled-coil forming protein serine/threonine kinase in mice. FEBS Lett. 392(2): 189-193, 1996.

Nakamura K, Yoshimura A, Kaneko T, Sato K, Hara Y. ROCK inhibitor Y- 27632 maintains the proliferation of confluent human mesenchymal stem cells. J Periodontal Res. 49(3): 363-370, 2014.

Naranjo-Suarez S, Carlson BA, Tsuji PA, Yoo MH, Gladyshev VN, Hatfield DL. HIF-independent regulation of thioredoxin reductase 1 contributes to the high levels of reactive oxygen species induced by hypoxia. PLoS One. 7(2): e30470, 2012.

Neidlinger-Wilke C, Wilke HJ, Claes L. Cyclic stretching of human osteoblasts affects proliferation and metabolism: a new experimental method and its application. J Orthop Res. 12(1): 70-78, 1994.

Nishiwaki T, Maeda N, Noda M. Characterization and developmental regulation of proteoglycan-type protein tyrosine phosphatase zeta/RPTPbeta isoforms. J Biochem. 123(3): 458-467, 1998.

Nobes CD and Hall A. Rho GTPases control polarity, protrusion, and adhesion during cell movement. J Cell Biol. 144(6): 1235-1244, 1999.

89

Nuka S, Zhou W, Henry SP, Gendron CM, Schultz JB, Shinomura T, Johnson J, Wang Y, Keene DR, Ramírez-Solis R, Behringer RR, Young MF, Höök M. Phenotypic characterization of epiphycan-deficient and epiphycan/biglycan double-deficient mice. Osteoarthritis Cartilage. 18(1): 88-96, 2010.

Nuki G. Osteoarthritis: a problem of joint failure. Z Rheumatol. 58(3): 142- 147, 1999.

Oshiro N, Fukata Y, Kaibuchi K. Phosphorylation of moesin by rho- associated kinase (Rho-kinase) plays a crucial role in the formation of microvilli-like structures. J Biol Chem. 273(52): 34663-34666, 1998.

Ossendorf C, Steinwachs MR, Kreuz PC, Osterhoff G, Lahm A, Ducommun PP, Erggelet C: Autologous chondrocyte implantation (ACI) for the treatment of large and complex cartilage lesions of the knee. Sports Med Arthrosc Rehabil Ther Technol. 3:11, 2011.

Oztürk A, Ozdemir MR, Ozkan Y. Osteochondral autografting (mosaicplasty) in grade IV cartilage defects in the knee joint: 2- to 7-year results. Int Orthop. 30(3): 200-204, 2006.

Page RL, Ambady S, Holmes WF, Vilner L, Kole D, Kashpur O, Huntress V, Vojtic I, Whitton H, Dominko T. Induction of stem cell gene expression in adult human fibroblasts without transgenes. Cloning Stem Cells. 11(3): 417- 426, 2009.

Parker WL, Goldring MB and Philip A. Endoglin is expressed on human chondrocytes and forms a heteromeric complex with betaglycan in a ligand and type II TGFbeta receptor independent manner. J Bone Miner Res. 18(2): 289-302, 2003.

Parkkinen JJ, Ikonen J, Lammi MJ, Laakkonen J, Tammi M, Helminen HJ. Effects of cyclic hydrostatic pressure on proteoglycan synthesis in cultured chondrocytes and articular cartilage explants. Arch Biochem Biophys. 300(1): 458-465, 1993.

Parkkinen JJ, Lammi MJ, Inkinen R, Jortikka M, Tammi M, Virtanen I, Helminen HJ. Influence of short-term hydrostatic pressure on organization of stress fibers in cultured chondrocytes. J Orthop Res. 13(4): 495-502, 1995.

Parra A, Veraldi N, Locatelli M, Fini M, Martini L, Torri G, Sangiorgi L, Bisio A. Heparin-like heparan sulfate from rabbit cartilage. Glycobiology. 22(2): 248-257, 2012.

90

Patil S and Tapasvi SR. Osteochondral autografts. Curr Rev Musculoskelet Med. 8(4): 423-428, 2015.

Pearce LR, Komander D, Alessi DR. The nuts and bolts of AGC protein kinases. Nat Rev Mol Cell Biol. 11(1): 9-22, 2010.

Peles E, Schlessinger J, Grumet M. Multi-ligand interactions with receptor- like protein tyrosine phosphatase beta: implications for intercellular signaling. Trends Biochem Sci. 23(4): 121-124, 1998.

Perkins DN, Pappin DJ, Creasy DM, Cottrell JS. Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis. 20(18): 3551-3567, 1999.

Pfaffl MW. A new mathematical model for relative quantification in real- time RT-PCR. Nucleic Acids Res. 29(9): e45, 2001.

Pfander D, Swoboda B, Cramer T. The role of HIF-1alpha in maintaining cartilage homeostasis and during the pathogenesis of osteoarthritis. Arthritis Res Ther. 8(1): 104, 2006.

Prehm P. Hyaluronate is synthesized at plasma membranes. Biochem J. 220(2): 597-600, 1984.

Prockop DJ, Kivirikko KI, Tuderman L, Guzman NA. The biosynthesis of collagen and its disorders (first of two parts). N Engl J Med. 301(1): 13-23, 1979.

Qvist P, Bay-Jensen AC, Christiansen C, Dam EB, Pastoureau P, Karsdal MA. The disease modifying osteoarthritis drug (DMOAD): Is it in the horizon? Pharmacol Res. 58(1): 1-7, 2008.

Rappsilber J, Ishihama Y, Mann M. Stop and go extraction tips for matrix- assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics. Anal Chem. 75(3): 663-670, 2003.

Ravenna L, Salvatori L, Russo MA. HIF3α: the little we know. FEBS J. 283(6): 993-1003, 2016.

Richter DL, Tanksley JA, Miller MD. Osteochondral Autograft Transplantation: A Review of the Surgical Technique and Outcomes. Sports Med Arthrosc. 24(2): 74-78, 2016.

91

Richette P, Chevalier X, Ea HK, Eymard F, Henrotin Y, Ornetti P, Sellam J, Cucherat M, Marty M. Hyaluronan for knee osteoarthritis: an updated meta- analysis of trials with low risk of bias. RMD Open. 1(1): e000071, 2015.

Ridley AJ, Schwartz MA, Burridge K, Firtel RA, Ginsberg MH, Borisy G, Parsons JT, Horwitz AR. Cell migration: integrating signals from front to back. Science. 302(5651): 1704-1709, 2003.

Ridinger K, Schäfer BW, Durussel I, Cox JA, Heizmann CW. S100A13. Biochemical characterization and subcellular localization in different cell lines. J Biol Chem. 275(12): 8686-8694, 2000.

Riento K and Ridley Y. Rocks: multifunctional kinases in cell behavior. Nat Rev Mol Cell Biol. 4(6): 446-456, 2003.

Roman-Blas JA, Castañeda S, Sánchez-Pernaute O, Largo R, Herrero- Beaumont G, CS/GS Combined Therapy Study group. Combined treatment with chondroitin sulfate and glucosamine sulfate shows no superiority over placebo for reduction of joint pain and functional impairment in patients with knee osteoarthritis: A six-month multicenter, randomized, double- blind, placebo-controlled clinical trial. Arthritis Rheumatol. 2016, 69(1): 77- 85, 2017.

Roughley PJ. The structure and function of cartilage proteoglycans. Eur Cell Mater. 12: 92-101, 2006.

Roughley PJ and White RJ. Dermatan sulphate proteoglycans of human articular cartilage. The properties of dermatan sulphate proteoglycans I and II. Biochem J. 262(3): 823-827, 1989.

Routhier A, Astuccio M, Lahey D, Monfredo N, Johnson A, Callahan W, Partington A, Fellows K, Ouellette L, Zhidro S, Goodrow C, Smith A, Sullivan K, Simone P, Le L, Vezuli B, Zohni M, West E, Gleason D, Bryan B. Pharmacological inhibition of Rho-kinase signaling with Y-27632 blocks melanoma tumor growth. Oncol Rep. 23(3): 861-867, 2010.

Ryu JH, Chae CS, Kwak JS, Oh H, Shin Y, Huh YH, Lee CG, Park YW, Chun CH, Kim YM, Im SH, Chun JS. Hypoxia-inducible factor-2α is an essential catabolic regulator of inflammatory rheumatoid arthritis. PLoS Biol. 12(6): e1001881, 2014.

92

Ryu JH, Shin Y, Huh YH, Yang S, Chun CH, Chun JS. Hypoxia-inducible factor-2α regulates Fas-mediated chondrocyte apoptosis during osteoarthritic cartilage destruction. Cell Death Differ. 19(3): 440-450, 2012.

Ryu JH, Yang S, Shin Y, Rhee J, Chun CH, Chun JS. Interleukin-6 plays an essential role in hypoxia-inducible factor 2α-induced experimental osteoarthritic cartilage destruction in mice. Arthritis Rheum. 63(9): 2732- 2743, 2011.

Saas J, Lindauer K, Bau B, Takigawa M, Aigner T. Molecular phenotyping of HCS-2/8 cells as an in vitro model of human chondrocytes. Osteoarthritis Cartilage. 12(11): 924-934, 2004.

Sahai E, Ishizaki T, Narumiya S, Treisman R. Transformation mediated by RhoA requires activity of ROCK kinases. Curr Biol. 9(3): 136-145, 1999.

Saito T, Ikeda T, Nakamura K, Chung UI, Kawaguchi H. S100A1 and S100B, transcriptional targets of SOX trio, inhibit terminal differentiation of chondrocytes. EMBO Rep. 8(5): 504-509, 2007.

Salminen-Mankonen H, Säämänen AM, Jalkanen M, Vuorio E, Pirilä L. Syndecan-1 expression is upregulated in degenerating articular cartilage in a transgenic mouse model for osteoarthritis. Scand J Rheumatol. 34(6): 469- 474, 2005.

Sarrazin S, Lamanna WC and Esko JD. Heparan sulfate proteoglycans. Cold Spring Harb Perspect Biol. 3(7), a004952, 2011.

Satoh S and Tominaga T. mDia-interacting protein acts downstream of Rho- mDia and modifies Src activation and stress fiber formation. J Biol Chem. 276(42): 39290-39294, 2001.

Sawitzke AD, Shi H, Finco MF, Dunlop DD, Harris CL, Singer NG, Bradley JD, Silver D, Jackson CG, Lane NE, Oddis CV, Wolfe F, Lisse J, Furst DE, Bingham CO, Reda DJ, Moskowitz RW, Williams HJ, Clegg DO. Clinical efficacy and safety of glucosamine, chondroitin sulphate, their combination, celecoxib or placebo taken to treat osteoarthritis of the knee: 2-year results from GAIT. Ann Rheum Dis. 69(8): 1459-1464, 2010.

Schipani E, Ryan HE, Didrickson S, Kobayashi T, Knight M, Johnson RS. Hypoxia in cartilage: HIF-1alpha is essential for chondrocyte growth arrest and survival. Genes Dev. 15(21): 2865-2876, 2001.

93

Schneider G and Filipek A. S100A6 binding protein and Siah-1 interacting protein (CacyBP/SIP): spotlight on properties and cellular function. Amino Acids. 41(4): 773-780, 2011.

Schrobback K, Klein TJ, Crawford R, Upton Z, Malda J, Leavesley DI. Effects of oxygen and culture system on in vitro propagation and redifferentiation of osteoarthritic human articular chondrocytes. Cell Tissue Res. 347(3): 649- 663, 2012.

Seror J, Zhu L, Goldberg R, Day AJ, Klein J. Supramolecular synergy in the boundary lubrication of synovial joints. Nat Commun. 6:6497, 2015.

Shang J, Liu H, Li J, Zhou Y. Roles of hypoxia during the chondrogenic differentiation of mesenchymal stem cells. Curr Stem Cell Res Ther. 9(2): 141-147, 2014.

Shen J, Li J, Chen D. TGF-β signaling and the development of osteoarthritis. Bone Res. 2, 14002, 2014.

Shevchenko A, Tomas H, Havlis J, Olsen JV, Mann M. In-gel digestion for mass spectrometric characterization of proteins and proteomes. Nat Protoc. 1(6): 2856-2860, 2007.

Shevchenko A, Wilm M, Vorm O, Mann M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem. 68(5): 850-858, 1996.

Shibata S, Fukada K, Imai H, Abe T, Yamashita Y. In situ hybridization and immunohistochemistry of versican, aggrecan and link protein, and histochemistry of hyaluronan in the developing mouse limb bud cartilage. J Anat. 203(4): 425-432, 2003.

Shum MS, Pasquier E, Po’uha ST, O’Neill GM, Chaponnier C, Gunning PW, Kavallaris M. γ-Actin regulates cell migration and modulates the ROCK signaling pathway. FASEB J. 25(12): 4423-4433, 2011.

Siclari A, Mascaro G, Gentili C, Cancedda R, Boux E. A cell-free scaffold- based cartilage repair provides improved function hyaline-like repair at one year. Clin Orthop Relat Res. 470(3): 910-919, 2012.

Singh VK, Kalsan M, Kumar N, Saini A, Chandra R. Induced pluripotent stem cells: applications in regenerative medicine, disease modeling, and drug discovery. Front Cell Dev Biol. 3:2, 2015.

94

Sit ST and Manser E. Rho GTPases and their role in organizing the actin cytoskeleton. J Cell Sci. 124(Pt5): 679-683, 2011.

Sivan SS, Wachtel E, Roughley P. Structure, function, aging and turnover of aggrecan in the . Biochim Biophys Acta. 1840(10): 3181- 3189, 2014.

Shibuya M, Suzuki Y, Sugita K, Saito I, Sasaki T, Takakura K, Nagata I, Kikuchi H, Takamae T, Hikada H and Mitsuyoshi N. Effect of AT877 on cerebral vasospasm after aneurysmal subarachnoid hemorrhage. Results of a prospective placebo-controlled double-blind trial. J Neurosurg. 76(4): 571- 577, 1992.

Smith RL, Lin J, Trindade MC, Shida J, Kajiyama G, Vu T, Hoffman AR, van der Meulen MC, Goodman SB, Schurman DJ, Carter DR. Time-dependent effects of intermittent hydrostatic pressure on articular chondrocyte type II collagen and aggrecan mRNA expression. J Rehabil Res Dev. 37(2): 153-161, 2000.

Solari FA, Dell’Aica M, Sickmann A, Zahedi RP. Why phosphoproteomics is still a challenge. Mol Biosyst. 11(6): 1487-1493, 2015.

Solheim E, Hegna J, Øyen J, Harlem T, Strand T. Results at 10 to 14 years after osteochondral autografting (mosaicplasty) in articular cartilage defects in the knee. Knee. 20(4): 287-290, 2013.

Soltz MA and Ateshian GA. A Conewise Linear Elasticity mixture model for the analysis of tension-compression nonlinearity in articular cartilage. J Biomech Eng. 122(6): 576-586, 2000.

Somlyo AV, Bradshaw D, Ramos S, Murphy C, Myers CE, Somlyo AP. Rho- kinase inhibitor retards migration and in vivo dissemination of human prostate cancer cells. Biochem Biophys Res Commun. 269(3): 652-659, 2000.

Sophia Fox AJ, Bedi A ja Rodeo SA: The basic science of articular cartilage: structure, composition, and function. Sports Health. 1(6): 461-468, 2009.

Ströbel S, Loparic M, Wendt D, Schenk AD, Candrian C, Lindberg RL, Moldovan F, Barbero A, Martin I. Anabolic and catabolic responses of human articular chondrocytes to varying oxygen percentages. Arthritis Res Ther. 12(2): R34, 2010.

95

Sun CC, Chiu HT, Lin YF, Lee KY, Pang JH. Y-27632, a ROCK inhibitor, promoted limbal epithelial cell proliferation and corneal wound healing. PLoS One. 10(12): e0144571, 2015.

Takahashi I, Onodera K, Sasano Y, Mizoguchi I, Bae JW, Mitani H, Kagayama M, Mitani H. Effect of stretching on gene expression of beta1 integrin and focal adhesion kinase and on chondrogenesis through cell- extracellular matrix interactions. Eur J Cell Biol. 82(4): 182-192, 2003.

Takigawa M, Tajima K, Pan HO, Enomoto M, Kinoshita A, Suzuki F, Takano Y, Mori Y. Establishment of a clonal human chondrosarcoma cell line with cartilage phenotypes. Cancer Res. 49(14): 3996-4002, 1989.

Tamma G, Procino G, Svelto M, Valenti G. Hypotonicity causes actin reorganization and recruitment of the actin-binding ERM protein moesin in membrane protrusions in collecting duct principal cells. Am J Physiol Cell Physiol. 292(4): C1476-C1484, 2007.

Tanimoto K, Kitamura R, Tanne Y, Kamiya T, Kunimatsu R, Yoshioka M, Tanaka M, Tanaka E, Tanne K. Modulation of hyaluronan catabolism in chondrocytes by mechanical stimuli. J Biomed Mater Res A. 93(1): 373-380, 2010.

Tatsumura M, Sakane M, Ochiai N, Mizuno S. Off-loading of cyclic hydrostatic pressure promotes production of extracellular matrix by chondrocytes. Cell Tissues Organs. 198(6): 405-413, 2013.

Thompson CL, Chapple JP, Knight MM. Primary cilia disassembly down- regulates mechanosensitive hedgehog signalling: a feedback mechanism controlling ADAMTS-5 expression in chondrocytes. Osteoarthritis Cartilage. 22(3): 490-498, 2014.

Thoms BL, Dudek KA, Lafont JE, Murphy CL. Hypoxia promotes the production and inhibits the destruction of human articular cartilage. Arthritis Rheum. 65(5): 1302-1312, 2013.

Toghraie F, Razmkhah M, Gholipour MA, Faghih Z, Chenari N, Torabi Nezhad S, Nazhvani Dehghani S, Ghaderi A. Scaffold-free adipose-derived stem cells (ASCs) improve experimentally induced osteoarthritis in rabbits. Arch Iran Med. 15(8): 495-499, 2012.

96

Tonge DP, Pearson MJ, Jones SW. The hallmarks of osteoarthritis and the potential to develop personalised disease-modifying pharmacological therapeutics. Osteoarthritis Cartilage. 22(5): 609-621, 2014.

Tsolis KC, Bei ES, Papathanasiou I, Kostopoulou F, Gkretsi V, Kalantzaki K, Malizos K, Zervakis M, Tsezou A, Economou A. Comparative proteomic analysis of hypertrophic chondrocytes in osteoarthritis. Clin Proteomics. 12(1): 12, 2015.

Tyyni A and Karlsson J. Biological treatment of joint cartilage damage. Scand J Med Sci Sports. 10(5): 249-265, 2000.

Ueki M, Tanaka M, Tanimoto K, Nishio C, Honda K, Lin YY, Tanne Y, Ohkuma S, Kamiya T, Tanaka E, Tanne K. The effect of mechanical loading on the metabolism of growth plate chondrocytes. Ann Biomed Eng. 36(5): 793-800, 2008.

Urban JP. The chondrocyte: a cell under pressure. Br J Rheumatol. 33(10): 901-908, 1994.

Valdes AM, Doherty M, Spector TD. The additive effect of individual genes in predicting risk of knee osteoarthritis. Ann Rheum Dis. 67(1): 124-127, 2008.

Varjosalo M, Sacco R, Stukalov A, van Drogen A, Planyavsky M, Hauri S, Aebersold R, Bennett KL, Colinge J, Gstaiger M, Superti-Furga G. Interlaboratory reproducibility of large-scale human protein-complex analysis by standardized AP-MS. Nat Methods. 10(4): 307-314, 2013.

Vinardell T, Rolfe RA, Buckley CT, Meyer EG, Ahearne M, Murphy P, Kelly DJ. Hydrostatic pressure acts to stabilise a chondrogenic phenotype in porcine joint tissue derived stem cells. Eur Cell Mater. 23: 121-132, 2012.

Vincourt JB, Lionneton F, Kratassiouk G, Guillemin F, Netter P, Mainard D, Magdalou J. Establishment of a reliable method for direct proteome characterization of human articular cartilage. Mol Cell Proteomics. 5(10): 1984-1995, 2006. von der Mark K, Gauss V, von der Mark H, Müller P. Relationship between cell shape and type of collagen synthesised as chondrocytes lose their cartilage phenotype in culture. Nature. 267(5611): 531-532, 1977.

97

Vranka JA, Sakai LY, Bächinger HP. Prolyl 3-hydroxylase 1, enzyme characterization and identification of a novel family of enzymes. J Biol Chem. 279(22): 23615-23621, 2004.

Wagner DR, Lindsey DP, Li KW, Tummala P, Chandran SE, Smith RL, Longaker MT, Carter DR, Beaupre GS. Hydrostatic pressure enhances chondrogenic differentiation of human bone marrow stromal cells in osteochondrogenic medium. Ann Biomed Eng. 36(5): 813-820, 2008.

Waldman SD, Spiteri CG, Grynpas MD, Pilliar RM, Kandel RA. Long-term intermittent shear deformation improves the quality of cartilaginous tissue formed in vitro. J Orthop Res. 21(4): 590-596, 2003.

Wang G and Beier F. Rac1/Cdc42 and RhoA GTPases antagonistically regulate chondrocyte proliferation, hypertrophy, and apoptosis. J Bone Miner Res. 20(6): 1022-1031, 2005.

Wang G, Woods A, Sabari S, Pagnotta L, Stanton LA, Beier F. RhoA/ROCK signaling suppresses hypertrophic chondrocyte differentiation. J Biol Chem. 279(13): 13205-13214, 2004.

Wang K, Xu J, Hunter DJ, Ding C. Investigational drugs for the treatment of osteoarthritis. Expert Opin Investig Drugs. 24(12): 1539-1556, 2015.

Wang M, Liu C, Thormann E, Dėdinaitė A. Hyaluronan and phospholipid association in biolubrication. Biomacromolecules. 14(12): 4198-4206, 2013.

Wang ZM, Yang DS, Liu J, Liu HB, Ye M, Zhang YF. ROCK inhibitor Y- 27632 inhibits the growth, migration, and invasion of Tca8113 and CAL-27 cells in tongue squamous cell carcinoma. Tumour Biol. 37(3): 3757-3764, 2016.

Wangpu X, Yang X, Zhao J, Lu J, Guan S, Lu J, Kovacevic Z, Liu W, Mi L, Jin R, Sun J, Yue F, Ma J, Lu A, Richardson DR, Wang L, Zheng M. The metastasis suppressor, NDRG1, inhibits "stemness" of colorectal cancer via down-regulation of nuclear β-catenin and CD44. Oncotarget. 6(32): 33893- 33911, 2015.

Wann AK, Zuo N, Haycraft CJ, Jensen CG, Poole CA, McGlashan SR, Knight MM. Primary cilia mediate through control of ATP- induced Ca2+ signaling in compressed chondrocytes. FASEB J. 26(4): 1663- 1671, 2012.

98

Watts KL, Cottrell E, Hoban PR, Spiteri MA. RhoA signaling modulates cyclin D1 expression in human lung fibroblasts; implications for idiopathic pulmonary . Respir Res. 7: 88, 2006.

Wei L, Surma M, Shi S, Lambert-Cheatham N, Shi J. Novel Insights into the Roles of Rho Kinase in Cancer. Arch Immunol Ther Exp (Warsz). 64(4): 259- 278, 2016.

Wenger RH and Gassmann M. Oxygen(es) and the hypoxia-inducible factor- 1. Biol Chem. 378(7): 609-616, 1997.

Wessel D and Flügge UI. A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids. Anal Biochem. 138(1): 141-143, 1984.

Wilson R, Belluoccio D, Bateman JF. Proteomic analysis of cartilage proteins. Methods. 45(1): 22-31, 2008.

Wilusz RE, Defrate LE, Guilak F. A biomechanical role for perlecan in the pericellular matrix of articular cartilage. Matrix Biol. 31(6): 320-327, 2012.

Wong M, Siegrist M, Goodwin K. Cyclic tensile strain and cyclic hydrostatic pressure differentially regulate expression of hypertrophic markers in primary chondrocytes. Bone. 33(4): 685-693, 2003.

Woods A and Beier F. RhoA/ROCK signaling regulates chondrogenesis in a context-dependent manner. J Biol Chem. 281(19): 13134-13140, 2006.

Woods A, Wang G, Beier F. RhoA/ROCK signaling regulates Sox9 expression and actin organization during chondrogenesis. J Biol Chem. 280(12): 11626- 11634, 2005.

Wright MO, Nishida K, Bavington C, Godolphin JL, Dunne E, Walmsley S, Jobanputra P, Nuki G, Salter DM. Hyperpolarisation of cultured human chondrocytes following cyclical pressure-induced strain: evidence of a role for alpha 5 beta 1 integrin as a chondrocyte mechanoreceptor. J Orthop Res. 15(5): 742-747, 1997.

Wu JJ and Eyre DR. Structural analysis of cross-linking domains in cartilage type XI collagen. Insights on polymeric assembly. J Biol Chem. 270(32): 18865-18870, 1995.

99

Wu JJ, Woods PE, Eyre DR. Identification of cross-linking sites in bovine cartilage type IX collagen reveals an antiparallel type II-type IX molecular relationship and type IX to IX bonding. J Biol Chem. 267(32): 23007-23014, 1992.

Wu L, Bluguermann C, Kyupelyan L, Latour B, Gonzalez S, Shah S, Galic Z, Ge S, Zhu Y, Petrigliano FA, Nsair A, Miriuka SG, Li X, Lyons KM, Crooks GM, McAllister DR, Van Handel B, Adams JS, Evseenko D. Human developmental chondrogenesis as a basis for engineering chondrocytes from pluripotent stem cells. Stem Cell Reports. 1(6): 575-589, 2013.

Wu QQ and Chen Q. Mechanoregulation of chondrocyte proliferation, maturation, and hypertrophy: ion-channel dependent transduction of matrix deformation signals. Exp Cell Res. 256(2): 383-391, 2000.

Zhang B, Luo S, Wang Q, Suzuki T, Xiong WC, Mei L. LRP4 serves as a coreceptor of agrin. Neuron. 60(2): 285-297, 2008.

Zhang G, Shang B, Yang P, Cao Z, Pan Y, Zhou Q. Induced pluripotent stem cell consensus genes: implication for the risk of tumorigenesis and cancers in induced pluripotent stem cell therapy. Stem Cells Dev. 21(6): 955-964, 2012.

Zhang P, Yao Q, Lu L, Li Y, Chen PJ, Duan C. Hypoxia-inducible factor 3 is an oxygen-dependent transcription activator and regulates a distinct transcriptional response to hypoxia. Cell Rep. 6(6): 1110-1121, 2014.

Zhang W, Ouyang H, Dass CR, Xu J. Current research on pharmacologic and regenerative therapies for osteoarthritis. Bone Res. 4: 15040, 2016.

Zhao YH, Lv X, Liu YL, Zhao Y, Li Q, Chen YJ, Zhang M. Hydrostatic pressure promotes the proliferation and osteogenic/chondrogenic differentiation of mesenchymal stem cells: The roles of RhoA and Rac1. Stem Cell Res. 14(3): 283-296, 2015.

Zhou C and Petroll WM. Rho kinase regulation of fibroblast migratory mechanics in fibrillar collagen matrices. Cell Mol Bioeng. 3(1): 76-83, 2010.

Zhou S, Cui Z, Urban JP. Factors influencing the oxygen concentration gradient from the synovial surface of articular cartilage to the cartilage-bone interface: a modeling study. Arthritis Rheum. 50(12): 3915-3924, 2004.

100

Zhou Y, Millward-Sadler SJ, Lin H, Robinson H, Goldring M, Salter DM, Nuki G. Evidence for JNK-dependent up-regulation of proteoglycan synthesis and for activation of JNK1 following cyclical mechanical stimulation in a human chondrocyte culture model. Osteoarthritis Cartilage. 15(8): 884-893, 2007.

Zimmermann DR and Ruoslahti E. Multiple domains of the large fibroblast proteoglycan, versican. EMBO J. 8(10): 2975-2981, 1989.

Zreiqat H, Howlett CR, Gronthos S, Hume D, Geczy CL. S100A8/S100A9 and their association with cartilage and bone. J Mol Histol. 38(5): 381-391, 2007.

Yammani RR. S100 proteins in cartilage: role in arthritis. Biochim Biophys Acta 1822(4): 600-606, 2012.

Yammani RR, Carlson CS, Bresnick AR, Loeser RF. Increase in production of matrix metalloproteinase 13 by human articular chondrocytes due to stimulation with S100A4: Role of the receptor for advanced glycation end products. Arthritis Rheum. 54(9): 2901-2911, 2006.

Yang S, Kim J, Ryu JH, Oh H, Chun CH, Kim BJ, Min BH, Chun JS. Hypoxia-inducible factor-2alpha is a catabolic regulator of osteoarthritic cartilage destruction. Nat Med. 16(6): 687-693, 2010.

Yu HS, Kim JJ, Kim HW, Lewis MP, Wall I. Impact of mechanical stretch on the cell behaviors of bone and surrounding tissues. J Tissue Eng. 6:2041731415618342, 2015.

Yu SP and Hunter DJ. Emerging drugs for the treatment of knee osteoarthritis. Expert Opin Emerg Drugs. 20(3): 361-378, 2015.

Yu Z, Liu M, Fu P, Xie M, Wang W, Luo X. ROCK inhibition with Y27632 promotes the proliferation and cell cycle progression of cultured astrocyte from spinal cord. Neurochem Int. 61(7): 1114-1120, 2012.

Önnerfjord P, Khabut A, Reinholt FP, Svensson O, Heinegård D. Quantitative proteomic analysis of eight cartilaginous tissues reveals characteristic differences as well as similarities between subgroups. J Biol Chem. 287(23): 18913-18924, 2012.

101