UNIVERSITY OF CALGARY. Molecular Basis of Articular Cartilage Boundary Lubrication: Role of PRG4 Structure & Multimerisation.

Transcription

1 UNIVERSITY OF CALGARY Molecular Basis of Articular Cartilage Boundary Lubrication: Role of PRG4 Structure & Multimerisation by Saleem Abubacker A THESIS SUBMITTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY GRADUATE PROGRAM IN BIOMEDICAL ENGINEERING CALGARY, ALBERTA JUNE, 2015 Saleem Abubacker 2015

2 Abstract Proteoglycan 4 (PRG4) is a mucin-like glycoprotein found in synovial fluid (SF) and at the articular cartilage surface, where it is required for joint lubrication and health. Hyaluronan (HA), a glycosaminoglycan polymer, is another SF constituent that contributes to SF s viscosity and cartilage lubrication properties. PRG4 and HA function effectively as friction reducing boundary lubricants at a cartilage-cartilage interface, though both have been studied at other interfaces with varying results. PRG4 can exist in SF as disulfide-bonded multimers, a structurally determinant characteristic of mucins, which may be necessary for its cartilage adsorption and boundary lubricating ability. A recently developed full-length recombinant human PRG4 (rhprg4) has demonstrated appropriate higher order structure, O-linked glycosylations, and boundary lubricating ability at the ocular surface. However, it remains unclear if this rhprg4 is able to adsorb to and function as a cartilage boundary lubricant. The objectives of this thesis were to (1) determine the effect of different sliding interface materials on the lubricating ability of PRG4 and HA by measuring the kinetic coefficient of friction, (2i) assess the cartilage adsorption and boundary lubricating ability of disulfide-bonded PRG4 multimers and PRG4 monomers, (2ii) evaluate the cartilage boundary lubricating ability of PRG4 multimers and PRG4 monomers with HA, and (3) assess the cartilage adsorption of rhprg4 and the in vitro cartilage boundary lubricating properties of rhprg4, with and without HA. PRG4 demonstrated boundary lubricating function at both cartilage-cartilage and cartilage-glass interfaces, while HA demonstrated friction reducing ability only at the cartilagecartilage interface. The inter-molecular disulfide-bonded multimeric structure of PRG4 was important for its ability to adsorb to a cartilage surface and function as a boundary lubricant. i

3 Finally, rhprg4 demonstrated cartilage adsorption and boundary lubricating function, with and without HA, equivalent to native PRG4. Collectively, these results demonstrate the effectiveness of putative cartilage boundary lubricants can be affected by the counterface, contribute to a greater understanding of the molecular basis of articular cartilage boundary lubrication of PRG4, and provide the foundation and motivation for future clinical evaluation of rhprg4 as a biotherapeutic treatment for osteoarthritis. ii

4 Preface This thesis is presented in a manuscript based format; therefore there is some repetition in Introductions and Methods between chapters. Chapters 2, 3 and 5 and the Appendix are either in preparation for submission, submitted, or published. Chapter 2, is in preparation for submission to the Journal of Orthopaedic Research: Abubacker S, McPeak AE, Dorosz SG, Egberts P, Schmidt TA. Effect of Counterface on Cartilage Boundary Lubricating Ability by Proteoglycan 4 and Hyaluronan: Cartilage-Glass vs. Cartilage-Cartilage. Chapter 3, is in preparation for submission to Connective Tissue Research: Abubacker S, Ponjevic D, Matyas JR, Schmidt TA. Effect of disulfide bonding and multimerisation on Proteoglycan 4 s cartilage boundary lubricating ability and adsorption. Chapter 5, is in revision at Annals of Biomedical Engineering: Abubacker S, Dorosz SG, Ponjevic D, Jay GD, Matyas JR, Schmidt TA. Cartilage Boundary Lubricating Ability of Full- Length Recombinant Human Proteoglycan 4 Alone and in Combination with Hyaluronan. The Appendix, is published in Osteoarthritis and Cartilage: Abubacker S, Ham HO, Messersmith PB, Schmidt TA. Cartilage Boundary Lubricating Ability of Aldehyde Modified Proteoglycan 4 (PRG4-CHO). Osteoarthritis Cartilage. 2013; 21(1): iii

5 Acknowledgements Firstly, I would like to thank my supervisor, Dr. Tannin Schmidt, I will be forever grateful for his willingness to take a gamble on me by offering me a position in his lab as his first PhD student, even though he admittedly did not understand me fully during my phone interview due to my accent. Dr. Schmidt is unbelievably patient and supportive, has a great enthusiasm for science which is a great motivation for me, and at such a young age with so much achieved and much more to come he is a great role model for me and others under his supervision. I am especially thankful for his support on an academic level, going through all of my experiments and working on my writing skills, reeling me in at times when I got ahead of myself, aiding my development as a researcher, and showing me the importance of knowledge transfer. I am proud to have been part of this lab at its early stages and look forward to following its development in the future. Dr. John Matyas as my co-supervisor was a great mentor and brought a great alternative perspective to my research. I am especially thankful to him and Dragana Ponjevic, MSc. for their help in training me on the immunohistochemistry part of my research, and getting those great images. Thank you to my committee members (Drs. David Hart and Arin Sen) for their continuous support and advice, and my defence examiners (Drs. Roman Krawetz, Ed Nowicki, Jennifer Elisseeff) for your time and suggestions. Collaborators who have aided in the iv

6 completion of my work; Dr. Gregory D. Jay, Dr. Phillip B. Messersmith, Dr. Hyun Ok Ham, Dr. Heather Sheardown, and Ms. Lina Liu, MSc. My thanks to Dr. Duncan Shepherd, Master s supervisor, who has continued to keep contact with me beyond my time in Birmingham and provided me with countless reference letters. Thank you to the McCaig Institute and Alberta OA Team for their continuous support throughout my PhD, not only financially but also the interaction with all the collaborators and by providing my fifteen minutes of fame in the media and all the skills I gained from that experience such as communication skills and knowledge transfer. Thank you to the Arthritis Society and Faculty of Graduate Studies for support through studentships. Thanks to the BME graduate program for helping me develop through the course and BMEG, and a special thanks to Dr. Mike Kallos and Lisa Mayer. To my lab mates past and present, thank you for all your help and fun times, especially Sam Dorosz and Michael Samsom for the laughs, Dr. Bridgett Steel, Sheila Morrison and Cecilia Alvarez Veronesi for aiding me in my early lab life, and the several students who made summer research periods fun and helped complete my work. I would like to specially thank future Dr. Drs. Taryn Ludwig and Kristen Barton, who kept my dream career choice ray of light alive and the many Menchies trips. Taryn for her unreserved patience in training me in my early days and not hanging me off the HPL stairs with my numerous questions. More importantly thank you for taking me under your wing, helping my transition to Canadian life, and being a great friend. Kristen for her continuous upbeat support in and out of the lab, continuous supply of baked goods and making conference trips that much more fun. Thank you and your family for making my life in Canada that much easier by adopting me during the festive holidays. Thanks to all the members of the HPL who have made my life fun in and out of the science world. To the guys from Avalanche, for bringing regular football (soccer) back into my v

7 life. I would especially like to thank Ryan Madden, Sarah Akierman, Nicky Schrier, Hendrik Enders, Jennifer Baltich, Erica Buckeridge, Alissa Nicolucci, Chris Newell, Francisco Uribe and Alex Jalba for being great friends and also helping me during my broken leg period, keeping me entertained and getting me back into shape with post-injury training, food, chocolate brioches and tea, your support made my recovery that much easier and my time here even better. Shamadul Haque, Amrit Bagha and Sonia Bangar, your friendship from across the pond is so appreciated after all these years. You guys have always inspired the boffin inside of me and push me to do great stuff. Thank you for bringing a slice of home to my Canadian life, looking out for me and the many hours of FM. Definitely the best friends a guy could ask for. My family both in UK and India, thank you for your continuous support throughout my life. My aunt, Dr. Majeeda Kamaluddeen, and cousin, Nabila Ruhi, for looking after me whilst I have been here in Calgary by providing a home away from home, having family in a new city makes for a much easier transition. Finally, my parents Sadiq and Waheeda Abubacker, and my sister Naureen Abubacker, for their continuous support into my full time career of studying. Naureen thank you for the updates from home and keeping me grounded whenever everyone else thought I was a golden child. My parents have provided me with a great foundation and have always pushed my educational interests, and have supported my life decisions and provided advice every step of the way. My achievements, work ethic and my personal character are all in part thanks to them, and I hope one day to be able to pay them back in some small part. Thank you. vi

8 Dedication I would like to dedicate this thesis, work and achievements to my late grandfather, Utthakan Abubacker. He taught me that education and hard work are key to life, and always encouraged me to aim for my life goals and was supportive of my accomplishments. The stories of his life he told me were a great inspiration for me and he looked out for me throughout my life. I hope he would be proud of this achievement and my future accomplishments. vii

9 Table of Contents Abstract... i Preface... iii Acknowledgements... iv Dedication... vii Table of Contents... viii List of Figures and Illustrations... xii List of Symbols, Abbreviations and Nomenclature...xv CHAPTER 1: INTRODUCTION Introduction to the Thesis Synovial Joints Articular Cartilage Synovial Fluid (SF) Osteoarthritis (OA) Modes of Lubrication Experimental Analysis of Lubrication Proteoglycan 4 (PRG4) Structure Function PRG4 and Osteoarthritis Hyaluronan (HA) Structure Function HA and Osteoarthritis Aims...28 CHAPTER 2: EFFECT OF COUNTERFACE ON CARTILAGE BOUNDARY LUBRICATING ABILITY BY PROTEOGLYCAN 4 AND HYALURONAN: CARTILAGE-GLASS VS. CARTILAGE-CARTILAGE Abstract Introduction Methods Solution Preparation PRG HA Boundary Lubrication Test Cartilage-cartilage test sample preparation Cartilage-glass test sample preparation Lubrication Test Data Analysis Results Test Characterisation Cartilage-on-Glass Friction...37 viii

10 2.4.3 Cartilage-on-Cartilage Friction Discussion Acknowledgements...43 CHAPTER 3: EFFECT OF DISULFIDE BONDING AND MULTIMERISATION ON PROTEOGLYCAN 4 S CARTILAGE BOUNDARY LUBRICATING ABILITY AND ADSORPTION Abstract Introduction Methods Lubricant Preparation and Characterisation PRG PRG4 separation PRG4 Characterisation PRG4 Reduction and Alkylation Boundary Lubrication Tests Sample Preparation Lubrication Test Lubricant Test Sequences Immunohistochemistry Sample preparation Specimen processing Statistical Analysis Results Boundary Lubrication Tests Cartilage Boundary Lubricating Ability of PRG4Multi+ vs. PRG4Multi Concentration Dependent Cartilage Boundary Lubricating Ability of PRG4Multi+ and PRG4Multi Cartilage Boundary Lubricating Ability of R/A PRG4Multi and NR PRG4Multi Immunohistochemistry Discussion...66 B1. Acknowledgements...70 CHAPTER 4: CARTILAGE BOUNDARY LUBRICATING ABILITY OF PROTEOGLYCAN 4 MULTIMER ENRICHED AND MULTIMER DEFICIENT PREPARATIONS WITH HYALURONAN Abstract Introduction Methods Lubricant Preparation and Characterisation PRG4 Preparation and Characterisation PRG4 separation HA Preparation...76 ix

11 4.3.2 Cartilage Boundary Lubrication Tests Sample Preparation Lubrication Test Lubricant Test Sequence Statistical Analysis Results Cartilage Boundary Lubrication Tests Discussion Acknowledgements...85 CHAPTER 5: CARTILAGE BOUNDARY LUBRICATING ABILITY OF FULL-LENGTH RECOMBINANT HUMAN PROTEOGLYCAN 4 ALONE AND IN COMBINATION WITH HYALURONAN Abstract Introduction Methods (rh)prg4 Preparations PRG rhprg Immunohistochemistry (IHC) Immunohistochemistry (IHC) Samples Specimen processing Boundary Lubrication Tests Sample Preparation Lubrication Test Lubrication Test Sequences Statistical Analysis Results SDS-PAGE Immunohistochemistry Boundary Lubrication Tests rhprg4 vs. PRG rhprg4 vs. rhprg4+ha Discussion Acknowledgements CHAPTER 6: CONCLUSIONS Summary of Findings Limitations Discussion and Future Work PRG4 structure in disease progression and personalised treatment PRG4 and HA relationship PRG4 localisation at the articular cartilage surface REFERENCES x

12 APPENDIX A: CARTILAGE BOUNDARY LUBRICATING ABILITY OF ALDEHYDE MODIFIED PROTEOGLYCAN 4 (PRG4-CHO) A1. Abstract & Introduction A2. Methods A3. Results A4. Discussion A5. Acknowledgements A6. Re-print permission for Appendix A, published in Osteoarthritis and Cartilage 2013; 21(1) p APPENDIX B: TANDEM MASS SPECTROMETRY (MS/MS) ON PURIFIED PROTEOGLYCAN 4 (PRG4) OBTAINED FROM SIZE EXCLUSION CHROMATOGRAPHY B1. Introduction B2. Methods B3. Results B4. Discussion xi

13 List of Figures and Illustrations Figure 1-1: Schematic of the human knee with important anatomical features and motions (A), with particular focus on the articular cartilage aspect of the knee (B). Adapted from [3] Figure 1-2: Schematic of healthy articular cartilage with cellular organisation (A) and collagen fibre orientation (B), with the percentage values representing approximate space occupied by each zone of the total cartilage thickness (STZ superficial tangential zone). Adapted from [7, 12] Figure 1-3: Molecular weight, duration of efficacy and treatment schedule of available intraarticular Hyaluronan formulations. Adapted from [20] Figure 1-4: Schematic of the basic engineering Stribeck curve (red) with corresponding lubrication film thicknesses (blue). Adapted from [34] Figure 1-5: Schematic of various theories developed based around the fluid-film (A-F), boundary mode of lubrication (G) and adaptive mechanically controlled lubrication mechanism (H). Adapted from [3, 39, 40, 42] Figure 1-6: Various testing protocols and interfaces used for lubricant assessment (blue arrow load applied; red arrow motion of counterface). Adapted from [43, 44, 47] Figure 1-7: Schematic of key features of the PRG4 molecule. Adapted from [58] Figure 1-8: Kinetic coefficient of friction (<μ kinetic,neq >) values for PBS, HA, PRG4, HA + PRG4, HA + PRG4 + SAPL, and SF. All lubricants of interest are at physiological concentrations; HA at 3.33 mg/ml, PRG4 at 450 μg/ml and SAPL at 200 μg/ml. Adapted from [13] Figure 1-9: Chemical structure of HA. The linear polymer is built from alternating units of Glucuronic acid and N-acetylglucosamine. Adapted from [27, 81] Figure 2-1: Friction data acquired on cartilage-glass interfaces. Torque (τ) profiles obtained during two rotations at three velocities (0.01, 0.3 and 10 mm/s) on PBS (A) and SF (B). Instantaneous axial load (N) profiles acquired simultaneously for PBS (C) and SF (D) Figure 2-2: Friction data acquired on cartilage-cartilage interfaces. Torque (τ) profiles obtained during two rotations at three velocities (0.01, 0.3 and 10 mm/s) on PBS (A) and SF (B). Instantaneous axial load (N) profiles acquired simultaneously for PBS (C) and SF (D) Figure 2-3: Cartilage-glass; average kinetic friction coefficient <μ kinetic > (A) and average normalised kinetic friction coefficient <μ kinetic,neq > (B). Cartilage-cartilage; average xii

14 kinetic friction coefficient <μ kinetic > (C) and average normalised kinetic friction coefficient <μ kinetic,neq > (D) Figure 3-1: SEC chromatogram showing three peaks of interest. Peak 2 is shown to be PRG4Multi+ and peak 3 is shown to be PRG4Multi Figure 3-2: SDS-PAGE of PRG4Multi and PRG4Multi+ preparations; Protein stain (A) and western blotting with Ab LPN (B) and mab 4D6 (C) Figure 3-3: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients PRG4Multi and PRG4Multi+ at 450 µg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p < Figure 3-4: Kinetic <μ kinetic,neq > at Tps = 1.2 s dose-response friction coefficients of PRG4Multi and PRG4Multi+ at 45, 150 and 450 µg/ml. Where * is p < 0.01 and ** is p < 0.05 for PRG4Multi+. There are no significant differences across the different concentrations of PRG4Multi. ANOVA with Fisher s post-hoc. Sample size, n = Figure 3-5: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients NR PRG4Multi and R/A PRG4Multi at 450 µg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p < Figure 3-6: Immunolocalisation of PRG4 at an articular cartilage surface. Key: Fresh samples taken directly from joint and snap-frozen (control). All other samples were shaken overnight in PBS at 4 o C, frozen over night at -80 o C, shaken again in fresh PBS at 4 o C and soaked in solutions of interest overnight at room temperature. Solutions: PBS (negative control), SF (positive control), PRG4Multi and PRG4Multi+, with PRG4 at physiological concentrations of 450 µg/ml. ( ) signifies negative samples that lacked primary mab 4D6 but contained secondary goat-anti-mouse Ab. Signal to anti-prg4 mab 4D6 is depicted by the red staining, chondrocyte cells are depicted by the blue DAPI staining Figure 4-1: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients PRG4Multi+ + HA, PRG4Multi + HA and unseparated PRG4 + HA, with PRG4 preparations at 450 µg/ml and HA at 3.33 mg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p < Figure 4-2: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients HA, Recombined-PRG4 with HA and unseparated PRG4 with HA, with PRG4 preparations at 450 µg/ml and HA preparations at 3.33 mg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p < xiii

15 Figure 5-1: Protein stain (A) and Western Blot (B) probed with mab 9G3 of non-reduced rhprg4 (NR) to reduced rhprg4 (R) Figure 5-2: Immunolocalisation of PRG4 at an articular cartilage surface. Fresh samples were taken directly from joint and snap-frozen (control). All other samples were shaken overnight in PBS at 4 o C, frozen at -80 o C, and shaken overnight again in PBS at 4 o C. Samples were then incubated in solutions of interest overnight at room temperature: PBS (negative control), SF (positive control), rhprg4 and PRG4 both at physiological concentrations of 450 µg/ml. Signal to anti-prg4 mab 9G3 is depicted by the red staining, chondrocyte cells are depicted by the blue DAPI staining Figure 5-3: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients in PBS, rhprg4 and PRG4 both at 450 µg/ml, and SF. Different letters signify statistically significant differences (p < 0.05), n = Figure 5-4: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients in PBS, rhprg4, rhprg4 + hyaluronan (HA) (rhprg4 at 450 µg/ml and HA (1.5 MDa) at 3.33 mg/ml), and SF. Different letters signify statistically significant differences (p < 0.05), n = Figure 6-1: schematic illustration of possible localisation of PRG4 monomers and multimers at an articular cartilage surface. C-terminal (blue) and N-terminal (red). Adapted from concepts by [34, 41, 58, 100, 122, 123] Figure 7-1: Protein stain of non-reduced (NR) and reduced (R) (A, B), and western blotting of NR (C) and R (D) PRG4 samples Figure 7-2: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients PRG4-SHAM, PRG4-CHO and PRG4 at 450 µg/ml. Sample size, n = xiv

16 List of Symbols, Abbreviations and Nomenclature Symbol Definition τ Torque τ max Maximum axial torque in the first 20 of rotation Axial torque averaged over the last 360 of τ rotation N Instantaneous axial load N eq R eff v eff Equilibrium axial load Effective radius Effective sliding velocity µ Coefficient of friction <μ kinetic > Kinetic coefficient of friction calculated using instantaneous axial load <μ kinetic,neq > Kinetic coefficient of friction calculated using equilibrium axial load μ static,neq ω 9G3 4D6 Ab ACL ANOVA BCA C-terminal CACP CHAPS CHO Static coefficient of friction calculated using equilibrium axial load Angular velocity Anti-PRG4 antibody, mucin domain Anti-PRG4 antibody, unknown binding site Anti-body Anterior cruciate ligament Analysis of variance Bicinchoninic acid assay Carboxy terminal Camptodactyly-arthropathy-coxa vara-pericarditis 3-[(3-cholamidopropyl) dimethylammonio]-1- propanesulfonate, zwitterionic detergent Chinese hamster ovary xv

17 CI Confidence interval DEAE Diethylaminoethyl dh 2 O Distilled water EtOH Ethanol F Friction force FPLC Fast protein liquid chromatography GAG Glycosaminoglycan HA Hyaluronan HAS Hyaluronan synthases IgG Immunoglobulin G IHC Immunohistochemistry LPN Anti-PRG4 antibody, C-terminal LUB:1 Truncated human recombinant lubricin mab Monoclonal anti-body MS/MS Tandem mass spectrometry MSF Megakaryocyte stimulating factor MW Molecular weight n One friction testing replicate N-terminal Amino terminal NR Non-reduced NSAIDs Non-steroidal anti-inflammatory drugs OA Osteoarthritis PBS Phosphate buffered saline PRG4 Proteoglycan 4 PRG4-CHO Aldehyde modified PRG4 PRG4Multi+ PRG4 Multimer Enriched PRG4Multi PRG4 Multimer Deficient R Radius R inner (or R i ) Inner radius R outer (or R O ) Outer radius xvi

18 R/A (or R) rhprg4 RMS s SAPL SD SDS-PAGE SEC SEM SF SZP TGF-β1 Tps Reduced and alkylated Recombinant human PRG4 Root mean square Seconds Surface active phospholipid Standard deviation Sodium dodecyl sulfate polyacrylamide gel electrophoresis Size exclusion chromatography Standard error of the mean Synovial fluid Superficial zone protein Transforming growth factor β1 Pre-sliding duration xvii

19 CHAPTER 1: Introduction 1.1 Introduction to the Thesis Proteoglycan 4 (PRG4), a mucin-like glycoprotein, and hyaluronan (HA), a glycosaminoglycan, are synovial fluid (SF) constituents that are necessary for joint health and lubrication. PRG4 and HA are both found at the surface of articular cartilage. PRG4 has consistently been shown to function as a boundary lubricant at various synthetic and biological surfaces. Conversely, HA has demonstrated variable boundary lubricating properties at synthetic and biological surfaces. At a cartilage-cartilage interface, PRG4 and HA alone function as dosedependent boundary lubricants, and when combined function synergistically to reduce friction to levels near that of whole SF. PRG4 is similar to other mucins in that it is highly glycosylated and is able to form entanglements and disulfide-bonded multimers, both of which are functionally determinant properties. Indeed, the ability to form multimers has been shown to be important for PRG4 s cartilage boundary lubricating ability. However, it has yet to be fully elucidated what effects inter- and intra-molecular disulfide bonds have on PRG4 s boundary lubricating ability at a cartilage interface, both alone and in combination with HA, as well as its ability to adsorb to the articular cartilage surface. Full-length recombinant human PRG4 (rhprg4) has recently become available for study. The successful expression of rhprg4 was motivated by efforts towards developing an improved biotherapeutic treatment for patients with osteoarthritis. However, the ability of this rhprg4 to function as a cartilage boundary lubricant, both alone and in combination with HA, and to adsorb to the articular cartilage surface remains to be determined. 1

20 The hypotheses of this thesis were: 1. HA functions as a boundary lubricant at a cartilage-cartilage interface but not a cartilageglass interface. 2i. PRG4 s disulfide-bonded multimeric structure is necessary for its ability to adsorb to a cartilage surface and function as a friction-reducing cartilage boundary lubricant. 2ii. PRG4 s disulfide-bonded multimeric structure is necessary for its ability to interact with HA to synergistically reduce friction at a cartilage-cartilage interface. 3. rhprg4 is able to adsorb to the surface of articular cartilage and function as an effective friction-reducing cartilage boundary lubricant, both alone and in combination with HA. These hypotheses were investigated using a combination of previously characterised biomechanical, biochemical and histological methodologies. Overall, this work contributed to the understanding of PRG4 s structure-function relationship, as related to disulfide-bonded structure, and (rh)prg4 s cartilage boundary lubricating function both alone and in combination with HA. Aspects of these contributions are discussed below. This thesis is presented in a manuscript based format; therefore there is some repetition in the Introductions and Methods between chapters. Chapters 2, 3 and 5 and the Appendix are either in preparation for submission, submitted, or published as described in the Preface. Chapter 1 provides further information regarding the synovial joint anatomy and the roles of articular cartilage and SF. As well as expanding on the composition and function of SF, with a particular focus on PRG4 and HA, the different modes of lubrication, and methods used to analyse the lubricating ability of these molecules. Lastly, the PRG4 structure and function, and its relationship with HA are discussed. 2

21 Chapter 2, which is in preparation for submission to the Journal of Orthopaedic Research, examines HA and PRG4 s lubricating ability at cartilage-glass and cartilage-cartilage interfaces by measuring the kinetic coefficient of friction over a wide range of sliding velocities. Chapter 3, which is in preparation for submission to Matrix Biology, employs biochemical methods to separate and characterise preparations of PRG4 multimers and monomers, then assess their cartilage adsorption and boundary lubricating ability. Chapter 4, examines the cartilage boundary lubricating ability of PRG4 multimers and monomers with HA. Chapter 5, which is in revision at Osteoarthritis and Cartilage, assesses the adsorption of rhprg4 to the cartilage surface and the cartilage boundary lubricating ability of rhprg4 with and without HA. Chapter 6 provides an overall summary, discussion of the major findings and suggestions for future work. The Appendix, which is published in Osteoarthritis and Cartilage, examines the effects of aldehyde modification on PRG4, previously shown to improve adsorption to depleted cartilage surfaces, on PRG4 s cartilage boundary lubricating ability. 3

22 1.2 Synovial Joints Synovial joints are a specific type of joint that are freely movable and can vary in size and shape within the human body; however, they do have similar characteristics[1-3]. They consist of a joint capsule that encapsulates the synovial fluid (SF). This consists of two parts: the stratum fibrosum (external), which is composed of ligaments that hold the surrounding bones together, and the stratum synoviale (internal), that is lined with the synovial membrane which produces SF (Figure 1-1A)[1-3]. Another common feature of diarthrosis joints is the presence of an articular cartilage lining at the end of each bone. This arrangement of SF in between the articulating cartilage helps lubricate and distribute loads across the joint, thus preventing the degeneration of the joint (Figure 1-1B)[1-3]. Figure 1-1: Schematic of the human knee with important anatomical features and motions (A), with particular focus on the articular cartilage aspect of the knee (B). Adapted from [3] Articular Cartilage Articular cartilage lines the ends of bones in synovial joints, providing a near frictionless bearing system through excellent lubrication and wear characteristics[4]. Articular cartilage also acts as a shock absorber by spreading applied loads over a larger area therefore spreading the 4

23 load across the subchondral bone[3, 5]. Articular cartilage is approximately 1 6 mm thick in human knees and differs in thickness depending on its location in the joint, the site of the joint within the body, as well as between species[3]. However, the features of articular cartilage are fairly similar. Cartilage is devoid of nerves, lymphatic vessels, and is generally avascular[6, 7]. Chondrocytes are the only active cells found in articular cartilage, making up approximately 1% of the cartilage volume in humans[6]. Chondrocytes are surrounded by an abundance of extracellular matrix, which provides the majority of articular cartilage s characteristics[8]. Although the chondrocytes and matrix are structurally independent, to maintain homeostasis they are co-dependent. The activity of the cells is required for extracellular matrix production and the matrix is required to maintain the cells environment, which provides protection for the cells via the pericellular matrix[6-8]. The matrix contains large amounts of water, a meshwork of predominantly type II collagen fibres, proteoglycans (glycosaminoglycans (GAGs) chains anchored to a protein core), a non-fibrous filler (carbohydrates and non-collagenous proteins), and a small amount of lipids and organic chemicals[3, 7-9]. The extracellular matrix is a porous-permeable solid matrix containing interstitial fluid and dissolved electrolytes. The structure of the matrix is essential to its functions; providing nutrition and removing waste from the chondrocytes via the interstitial fluid, providing water intake and elastic resistance to compression due to the presence of proteoglycans and GAGs, distributing load as explained by the biphasic theory and providing a high tensile strength thus minimising shear stresses[3, 6-11]. Throughout articular cartilage, the chondrocytes structure and the morphology of the matrix vary with depth. There are essentially four zones within articular cartilage (Figure 1-2)[6-9]. 5

24 - Zone 1 Superficial Zone: section of cartilage adjacent to the joint cavity. Chondrocytes are arranged in a compact area and form a flat-discoidal shape, parallel to the surface. The collagen fibres are aligned tangentially and parallel to the articular surface with low amounts of GAG. The uppermost part of the superficial zone is known as the surface lamina or lamina splendens. - Zone 2 Intermediate or Transitional Zone: chondrocytes are evenly spaced and form a spheroidal shape. The fibres create a meshwork with no predominant direction with increasing levels of GAG compared to the superficial zone. - Zone 3 Deep Zone: chondrocytes are spheroidal and create columnar groups of four to eight cells. The fibres create a tight linear meshwork perpendicular to the articular surface and contain highest levels of GAG. - Zone 4 Calcified Zone: adjacent to the subchondral bone. There are limited chondrocytes and the matrix begins to show signs of calcification from the crystals of calcium salts. 6

25 Figure 1-2: Schematic of healthy articular cartilage with cellular organisation (A) and collagen fibre orientation (B), with the percentage values representing approximate space occupied by each zone of the total cartilage thickness (STZ superficial tangential zone). Adapted from [7, 12] Synovial Fluid (SF) SF is a viscous ultrafiltrate of blood plasma and contains specific functional molecules, including various glycoproteins, phospholipids and proteins. One of the key functions of SF is to lubricate articular joints, which is facilitated by several putative lubricating macromolecules such as glycosaminoglycan polymers (hyaluronan (HA)), glycoproteins (proteoglycan 4 (PRG4)) and lipids (surface-active phospholipids (SAPL))[13]. The role of SF in the joint is manifold; SF reduces the friction between the articulating joint, provides nutrients to articular cartilage, and has a dampening effect when compressive forces are exerted due to its viscoelastic nature. When a sudden force is exerted, SF is able to deform elastically and resist the force[11]. The viscosity of SF in a healthy joint can vary from 7

26 60 x 10-3 Pa.s to 1400 x 10-3 Pa.s[14]. This variation is due to factors such as age, trauma of the joint and disease. An increase in age has been shown to lead to a decrease in viscosity[14]. Trauma of the joint and diseases, such as osteoarthritis (OA), also leads to a reduction of viscosity, generally less than 100 x 10-3 Pa.s[14]. The key reason for these large variations in viscosities is due to SF being a non-newtonian fluid. Thus, as shear rate increases (for example in exercise; described as a thixotropic pseudoplastic property) or the temperature increases the viscosity of the fluid decreases[3, 11] Osteoarthritis (OA) OA is a degenerative joint disease that results in articular cartilage degradation which can lead to bone-on-bone interaction; this can be extremely painful and reduce quality of life. The exact cause of OA is unknown but can be associated to multiple factors including mechanical and chemical aspects[15-17]. OA can be classed as primary OA, also known as idiopathic OA, which occurs in the absence of a known cause, and/or secondary OA, which occurs as a result of joint trauma[15-18]. In the USA, an estimated 27 million individuals over the age of 25 suffer from OA[19]. The estimated costs related to OA in 2007 was in the excess of US $60 billion per year[20]. In Canada, an estimated 10% of the population (approximately 4.4 million individuals) are affected by the disease, which costs approximately CAD $27.5 billion per year[21]. Currently, there are no cures for OA and existing treatments are only implemented when the first signs of pain are detected, which in most cases are symptoms of late stage OA[22]. There are several marketed treatments available which attempt to resolve the issues related to OA. The cheapest options are non-pharmacological treatments, such as weight loss regimes and 8

27 physical therapy[23, 24]. Non-steroidal Anti-inflammatory Drugs (NSAIDs) and other painrelief drugs are a pharmacological solution to the pain associated with OA, but can be associated with several side-effects and do not resolve the issue of cartilage degeneration[20]. Artificial joints relieve pain in most cases and can solve the problem of joint degeneration, to a certain extent, by completely replacing the bearing surface. Nevertheless, artificial joints are not a permanent solution since they succumb to both wear and loosening and only have an average life of years[25], which are unsuitable for young and active individuals. Another solution to reduce pain is by providing a new lining of articular cartilage through tissue engineering, which in principle resolves the issue of cartilage degeneration. However, the efficiency of tissue engineered cartilage, such as the mechanical properties of the cartilage construct and its durability, still needs to be improved before application in vivo[26]. Biotherapeutic based treatments, where therapeutic materials are produced via biological means, are another source of treatments for OA. Currently, HA injections are used as a form of viscosupplementation, to rejuvenate SF s lubricating capabilities and relieve pain[20, 27-31]. However, the efficacy of these OA biotherapeutics has been an issue[30, 31]; the retention time of the HA solutions is short (see Figure 1-3)[20] and has been found to be cleared from the joint within one day, even though the effects from the treatment can last several months[31]. 9

28 Figure 1-3: Molecular weight, duration of efficacy and treatment schedule of available intraarticular Hyaluronan formulations. Adapted from [20]. This highlights the debate as to whether HA treatments only reduce the amount of pain and not necessarily minimise the rate of degeneration of the cartilage[32]. Indeed, in vitro studies analysing the use of a commercially available hyaluronan injection, Hylan G-F 20, found a higher kinetic coefficient of friction compared to human SF, and similar to that of saline[33]. Hylan G-F 20 was able to reduce caspase-3 expression, an apoptosis marker, in chondrocytes compared to cartilage lubricated with saline. However, there was still a significant up-regulation of caspase-3 in the chondrocytes compared to cartilage lubricated with human SF. This suggests the removal of human SF, before intra-articular injection of HA alone, could be unfavourable[33] and that HA alone may not be sufficient in preventing cartilage breakdown. 10

29 1.3 Modes of Lubrication Friction is the resistance to motion between two surfaces in contact[3]. Friction is defined by: F = μ N Where; F = Frictional Force (N), μ = Coefficient of Friction (dimensionless), and N = Normal Applied Load (N) Although there are different forms of friction and, hence, various derivations of this formula, two of interest for articular cartilage are surface friction and plowing friction [3]. Surface friction occurs during interaction between two surfaces or due to the viscosity of a lubricant film as it is sheared between two surfaces[3]. Plowing friction occurs internally within the cartilage and when a load is applied to the joint resulting in the flow of the interstitial fluid[3, 4]. This facilitates load support and transfer of nutrients to chondrocytes[3]. The presence of SF within synovial joints provides lubrication to articular cartilage. There are three main modes of lubrication that can be considered: fluid-film lubrication, mixed lubrication and boundary lubrication (Figure 1-4)[34]. Fluid-film lubrication occurs when there is a layer of fluid separating apposing bearing surfaces leading to lower friction. Alternatively, boundary lubrication occurs when there is a surface-to-surface contact resulting in higher friction at the bearing surfaces. Mixed lubrication occurs when there is a mixture of both fluid-film and boundary lubrication[34-36]. The occurrence of fluid-film in joints, such as the knee and hip, generally occurs during minimal loads at relatively high speeds[3], which are infrequent, whereas boundary lubrication occurs during high loads at relatively slow speeds[4]. It has been found that the opposing articular cartilage in a synovial joint makes contact with over 10% of the total area[37], which increases the amount of friction occurring at the interface and leads to the 11

30 possible degradation of the cartilage. However, the analogies of describing the modes of lubrication are generally derived from basic engineering principals related to typical engineering surfaces that are non-porous with high elastic moduli and are based around the Stribeck curve[34, 36]. Figure 1-4: Schematic of the basic engineering Stribeck curve (red) with corresponding lubrication film thicknesses (blue). Adapted from [34]. With the presence of articular cartilage, a soft, hydrated and porous structure, the normal Stribeck curve seen with classic engineering interfaces cannot be mimicked[38]. Although the exact type of lubrication that occurs within synovial joints is not fully understood, there are various theories that have been developed based around the basic lubrication terminologies, specifically fluid-film and boundary modes of lubrication (Figure 1-4)[3, 36]. 12

31 Fluid-film - Hydrodynamic lubrication occurs when high-speeds are generated across the bearing surfaces with a layer of viscous fluid in between, which prevents contact between the opposing articular cartilages. Lubrication is provided through the hydrodynamic pressure of the SF forcing the articulating surfaces apart. However, this mode of lubrication requires continuous high-speed motion, which can be intermittent depending on the location of the joint and the motion being conducted (Figure 1-5A)[3]. - Elastohydrodynamic lubrication essentially follows the same theory as the hydrodynamic lubrication, with the added consideration of articular cartilage deformation (Figure 1-5B)[3]. - Squeeze Film lubrication occurs when two surfaces are approaching each other via a normal direction. Due to SF being viscoelastic, it will not immediately squeeze out from the enclosing gap and result in a time-varying pressure gap with the SF being exuded from the gap over time. However, for this to occur a larger surface area is required, such as those found in the lower-extremities of the body, in combination with a lighter load (i.e. swing phase of the gait cycle) (Figure 1-5C)[3, 39]. - Hydrostatic or Weeping lubrication is based on liquid being squeezed out of the porous articular cartilage due to pressure from the two articulating surfaces being compressed. In classic engineering theory, this theory is supported by an external pump providing continuous flow of fluid (Figure 1-5D)[3]. - Self-generating mechanism relies on fluid being exuded from the leading and trailing edges of the loading areas, and being re-absorbed at the centre. The fluid at the leading 13

32 edge would theoretically provide a supply of lubricant between the bearing surfaces (Figure 1-5E)[3]. - Boosted lubrication is hypothesised to occur when the articulating surfaces approach one another creating pressure. The solvent part of SF (i.e. water) passes through the cartilage, which leaves behind a concentrated gel-layer of lubricating molecules to provide lubrication. However, this is reliant on the thickness of the gel-layer to prevent the bearing surfaces coming into contact and therefore would only work over a short period of time (Figure 1-5F)[3]. Boundary - Molecular protection is based on a theory that there is a layer of boundary lubricants attached to the articulating surfaces providing lubrication[3, 34, 36, 39, 40]. This theory allows for the development of other ideas on how boundary lubricants, such as PRG4 and HA, lubricate at these surfaces. This includes the sacrificial mechanism theory, where after the layer of boundary lubricant is sheared off it is replenished by lubricants in the surrounding SF[39-41]. The importance of the boundary mode of lubrication is the emphasis on the layer of lubricating molecules, such as PRG4 and HA, at the cartilage surface that facilitates lubrication (Figure 1-5G)[3, 40]. Recently, an adaptive mechanically controlled lubrication mechanism has been postulated at the articular cartilage surface[42]. This mechanism is based on the HA-PRG4 complex having an adaptive role under varying compression. Under low loads and shear velocities, PRG4 is adsorbed to the surface and is entangled with HA to create a HA-PRG4 complex, thus providing a source of boundary lubrication[42]. During higher loads, HA is mechanically trapped providing 14

33 wear protection, and PRG4, due to its weaker surface attachment, is able to be sheared and redistributed to areas of higher shear forces and normal pressures (Figure 1-5H)[39-42]. 15

34 Figure 1-5: Schematic of various theories developed based around the fluid-film (A-F), boundary mode of lubrication (G) and adaptive mechanically controlled lubrication mechanism (H). Adapted from [3, 39, 40, 42]. 16

35 1.4 Experimental Analysis of Lubrication To assess the lubricating ability of molecules such as PRG4 and HA, several methods and protocols have been developed (Figure 1-6). In vivo studies are generally considered the gold standard for research. An ex vivo joint motion pendulum[43] has been previously utilised to analyse lubricants of interest; an animal knee joint would be excised, with the femoral head fixed and the tibia end allowed to act in a pendulum motion. The deceleration of the pendulum would be analysed to calculate the coefficient of friction[43]. While using this method of analysis is the most representative of physiological movements and a wide range of lubrication modes are operative, it is not possible to isolate or define a single mode of lubrication for study. The use of in vitro test setups with a stationary area of contact enables analysis of lubricants in the boundary mode of lubrication. A stationary contact area is where the opposing surfaces remain in contact, therefore minimising ploughing friction losses and as a result fluid pressure effects are minimal at relatively slow velocities[44]. However, these in vitro experimental setups can vary in the protocols and interfaces used. A latex-glass interface has been used with a rotation test geometry[45, 46]. The materials used in this setup are easy to work with and readily available, however the materials may not allow for physiologically relevant interactions of surfaces and lubricants of interest. Moving towards use of physiological tissue, a cartilage-glass interface using linear reciprocating test geometry has been utilised[42, 47]. As aforementioned, the use of glass makes these interfaces easier to work with and also enables analysis of lubrication in the presence of cartilage; however, the synthetic counterface may still not allow for all physiological interactions to occur. The linear reciprocating test geometry enables classic Stribeck analysis (by varying velocity of articulation, viscosity of test lubricant, and axial load) and therefore defining lubrication modes of interest[48]. However, Stribeck 17

36 analysis, originally developed for the study of metal bearings, may not be appropriate for or applicable to the soft, porous hydrogels such as articular cartilage[38]. Indeed, material choice has been shown to affect boundary lubrication analysis of lubricants[38]. Stribeck analysis has been utilised to assess the lubrication modes of stiff and impermeable synthetic surfaces against other high elastic moduli surfaces or against a cartilage surface[38, 42, 45-48], but this may not actually be appropriate for comparisons to articular cartilage due to its soft and porous interface[38]. The use of a material with soft and permeable properties, with or without a synthetic hard probe, was unable to recreate the normal characteristics of a Stribeck curve[38]. Hence, the use of a cartilage-cartilage interface using a stationary point of contact has been employed[44]. Preparation of cartilage samples can be challenging and require particular care, and the use of a physiologically relevant cartilage system may not enable comparisons to a classic Stribeck curve. Nevertheless, the use of a cartilagecartilage system does enable physiological interaction of lubricants at the cartilage surface that may not occur at a synthetic surface. Several studies have examined the boundary lubricating properties of both PRG4 and HA using the in vitro systems outlined above. PRG4 has consistently demonstrated low friction properties under boundary lubrication conditions at latex-glass[46], cartilage-glass[47, 48] and cartilage-cartilage[13] interfaces. Conversely, HA has been reported to function as a boundary lubricant for only the cartilage-cartilage interface[13, 49], but not for latex-glass[46] or cartilageglass[48] interfaces. Further studies of HA alone at a mica interface showed HA was unable to lubricate, and even with a lipid bilayer coated mica which enables the HA to biologically bind or with the covalent attachment of HA to a mica surface, HA was unable to lubricate[50]. However, with the presence of articular cartilage, HA has been shown to act as a boundary lubricant. It was 18

37 postulated that although HA cannot form a physical interaction with the cartilage surface, immobilisation could be achieved through mechanical trapping where entangled HA chains become trapped within collagen fibrils as the cartilage is compressed[42]. Recently, peptidemediated binding of HA was demonstrated to enhance lubrication of tissues and biomaterial surfaces, as further evidence for HA s ability to function as a cartilage boundary lubricant[51]. In summary, while each test interface (latex-glass, cartilage-glass and cartilage-cartilage) have their experimental advantages and disadvantages, for the study of cartilage boundary lubrication, a cartilage-cartilage test interface, which allows for physiological interactions between lubricants and cartilage surfaces, with a rotational test geometry and stationary area of contact, is an ideal test setup for studying cartilage boundary lubricants of interest[38]. Figure 1-6: Various testing protocols and interfaces used for lubricant assessment (blue arrow load applied; red arrow motion of counterface). Adapted from [43, 44, 47]. 19

38 1.5 Proteoglycan 4 (PRG4) Structure PRG4 is a mucin-like glycoprotein synthesised and secreted by cells at the surface of articular cartilage (specifically superficial zone chondrocytes[52, 53]), meniscus, synovial lining and tendons, and is found mainly in SF and at the surface of articular cartilage[54]. The PRG4 gene codes for the production of PRG4, and is also known as megakaryocyte-stimulating factor (MSF), superficial zone protein (SZP) or lubricin[55, 56]. Full-length PRG4 contains 12 exons, with a central mucin-like domain and is flanked by a carboxy-terminal (C-terminal) and an amino-terminal (N-terminal) (Figure 1-7)[56-58]. PRG4 possesses structurally determinant properties similar to that of mucins. Mucins are heavily glycosylated with a majority of its mass due to O-linked oligosaccharides[57, 59]. The mucin domain of PRG4 is also highly glycosylated which provides PRG4 the ability to reduce friction occurring at an interface due to repulsive hydration forces or charge repulsion[56] and provides SF the ability to dissipate strain energy[54]. Mucins, in general, can exist in a monomeric form as well as multimeric forms through disulfide bonds that occur between the cysteine-rich domains. The mucin monomers form disulfide bonds through post-translational processes initially in the endoplasmic reticulum, usually forming dimers, before higher order multimerisation occurs in the Golgi apparatus[57, 59]. The molecular weight (MW) of PRG4 can vary depending on its ability to form multimers and monomers. Monomers and multimers have been identified in bovine synovial fluid, and their MW reported to range from kda for multimers and and kda for monomers when purified from media conditioned by bovine articular cartilage explants[54, 60, 61]. Structural changes to mucins in general have shown a difference in functional characteristics, and this appears to also be true for PRG4[54, 57, 20

39 59, 62]. Pilot studies have indicated the reduction/alkylation of a PRG4 preparation, which breaks the disulfide bonds, resulted in approximately 34% decrease in lubricating function[63]. However, it remains unclear what role inter- vs. intra-molecular disulfide-bonding plays in this functional dependency. Figure 1-7: Schematic of key features of the PRG4 molecule. Adapted from [58] Function PRG4 plays a critical role in reducing the friction occurring between cartilage bearing surfaces, which prevents the degradation and adhesion of cartilage surfaces when boundary lubrication occurs. The interaction between PRG4 and the other constituents of SF, specifically HA, also plays a key role in lubricating joints. It has been found that the coefficient of friction of PRG4 and HA alone is significantly lower than PBS (negative control), and the combination of both PRG4 and HA resulted in a further decrease in friction, which was close to that of SF (positive control) (Figure 1-8)[13]. SAPL was also analysed but showed no significant reduction in friction, however only one type of SAPL, dipalmitoyl-phosphatidylcholine, was assessed[13]. 21

40 There are various types of SAPLs[64] and it is possible that one of these may have a role in the boundary mode of lubrication[65-67] and warrants further investigation. Figure 1-8: Kinetic coefficient of friction (<μ kinetic,neq >) values for PBS, HA, PRG4, HA + PRG4, HA + PRG4 + SAPL, and SF. All lubricants of interest are at physiological concentrations; HA at 3.33 mg/ml, PRG4 at 450 μg/ml and SAPL at 200 μg/ml. Adapted from [13]. Mutations in the PRG4 gene causes an autosomal recessive disorder in humans, camptodactylyarthropathy-coxa vara-pericarditis (CACP), which results in juvenile-onset, noninflammatory, precocious joint failure[68]. Indeed, Prg4-knockout mice show similar early cartilage degeneration in addition to synovial cell overgrowth, abnormal protein deposits and loss of the superficial zone chondrocytes[69]. PRG4-deficient human CACP SF lacks normal boundary lubricating ability[43]. However, an in vitro study showed replenishment of CACP SF 22

41 with PRG4 improved boundary lubrication as well as reducing chondrocyte apoptosis[70]. Generally most CACP patients are unable to produce PRG4 due to a nonsense mutation[58]. However, in some particular cases of CACP, post-translational mutation of the final coding exon in the C-terminal led to the production of non-functional PRG4 most likely due to a loss of cysteine residues[58]. This has shown to lead to loss of function, as it prevents the formation of disulfide-bridges[58] and compliments the aforementioned study where reduction/alkylation of PRG4 resulted in a detrimental effect on PRG4 s lubricating ability[54]. This further supports the notion that normal disulfide-bonding is necessary for PRG4 function and therefore joint health in vivo PRG4 and Osteoarthritis Research has demonstrated that after sustaining serious trauma or injury to joints, such as suffering an anterior cruciate ligament (ACL) injury, there is a higher risk of OA later in life[16, 71, 72]. One study collected SF from patients with ACL injury ranging from days (average 103 days), and found that there is a reduction in PRG4 concentration levels during the initial stage of trauma[71]. However, after a sufficient amount of time, approximately a year, the levels of PRG4 concentration returns to average levels of a healthy joint[71]. Other models looking at earlier time points have found that SF from ovine with ACL transection had a higher concentration of PRG4 in the first 2-4 weeks[73], before returning to normal at 20 weeks[73, 74]. In addition, the coefficient of friction values of the SF in the early weeks was significantly higher than normal SF, but did return to normal after 20 weeks[73]. It has also been shown that some patients suffering from chronic OA, who were deficient in PRG4, may benefit from PRG4 supplementation as a method of viscosupplementation[75]. 23

42 The concentration of PRG4 in some established OA patients can vary and has been shown to be normal[75] or even elevated[76] relative to normal human SF PRG4 concentrations[76]. A further confounding factor beyond concentration differences is potential variations in structural composition of PRG4 (e.g. multimers and monomers) in OA. Recently full length recombinant PRG4 (rhprg4) has been produced, in partnership with Lubris LLC. Early in vitro analysis of rhprg4 at the cornea has shown promise for its potential use as eye drops and for patients of dry eye syndrome[77, 78]. Given the known ability of PRG4 to lubricate at a cartilage interface and reduce cartilage damage in OA models, rhprg4 has the potential application as a biotherapeutic, with the possibility of improving current HA viscosupplementation treatments for OA due to their synergistic friction reducing relationship. In previous studies, a shorter form of the rhprg4 structure, LUB:1, had been produced and analysed[79]. LUB:1 provided lubricating ability and showed signs of having similar characteristics to that of native PRG4. LUB:1 was shown to have disease modifying and chondroprotective capabilities with respect to OA, thus preventing the progression of posttraumatic OA in animal models. In addition, LUB:1 was shown to localise to joint tissues for up to 28 days following a single injection[79]. LUB:1 is approximately one-third shorter than native PRG4[79], but it is unknown what the direct effects of a truncated PRG4 version have in terms of treatment for OA. The biotechnology company (Wyeth, now Pfizer) producing LUB:1 was unable to produce full-length rhprg4 at a high yield that would be feasible for large scale production. Whereas the current full-length rhprg4 produced by Lubris LLC has been produced in large quantities, approximately 1 g/l[78]. 24

43 1.6 Hyaluronan (HA) Structure HA is an extracellular matrix component and is a high MW glycosaminoglycan that is synthesised by a class of membrane proteins in chondrocytes, hyaluronan synthases (HAS). In humans, there are three types: HAS1, HAS2 and HAS3, which are found in chondrocytes[27]. The HASs produce the long linear polymers of repeating disaccharide structure composed of the monosaccharides Glucuronic acid and N-acetylglucosamine, resulting in HA having a wide range of MWs, up to 8 MDa (Figure 1-9). Synthetic HA solutions have also been developed, which are cross-linked HA with much higher MWs that is thought to provide a more effective HA injectable treatment for users[27, 80, 81]. Figure 1-9: Chemical structure of HA. The linear polymer is built from alternating units of Glucuronic acid and N-acetylglucosamine. Adapted from [27, 81] Function HA has multiple roles in the body; with respect to synovial joints, it maintains SF s viscoelastic properties, provides tissue hydration and transport of water, enables assembly with proteoglycans in the matrix, has many roles and interactions with receptors within cells in the joint and aids in cartilage lubrication[27, 80, 82]. In normal joints, HA can vary in concentration 25

44 (1 4 mg/ml) and MW (4 kda 8 MDa)[49]. HA has been shown to reduce friction in a dosedependent manner[13]. In addition, the MW of HA appears to also affect its cartilage boundary lubricating ability when alone, as increasing HA MW reduced friction at a cartilage-cartilage interface[49]. However, when combined with PRG4 there was no variation in lubricating ability across the varying concentrations and MWs[49, 83] HA and Osteoarthritis Patients suffering from chronic OA have been shown to have diminished MW of HA but not necessarily altered concentration when compared to healthy SF[75, 76, 84]. These chronic OA patients tend to have a higher concentration of lower MW HA[75, 76, 84]. Indeed, the odds of knee OA progression increases as HA MW distribution shifts lower[84]. In addition, some PRG4-deficient human OA SF has been shown to have a reduced lubricating ability compared to healthy SF at an in vitro cartilage-on-cartilage friction system[75]. However, supplementation of PRG4 restored the lubricating ability of the OA SF similar to that of healthy SF[75]. The variation in HA concentration and MW have also been shown to be similar in ACL transected ovine and equine models. In the ovine models, during early stages (2-4 weeks) the MW of HA shifted towards lower ranges[73] but appeared to return to a normal distribution at 20 weeks[74]. In the equine model, high MW HA (4 MDa) supplementation was able to restore the lubricating ability of HA-deficient SF close to that of normal SF[85]. HA is currently administered as a viscosupplement treatment for early OA. Different HA MW injections are currently used and although these injectable treatments can provide painrelief for several months, the residency time of the injected HA within in the joint can vary from 17 hours to 9 days[86]. This appears to be partially dependent on the HA MW and also the 26

45 location of the injection administered to the patient or model[51, 86-88]. The efficacy of the HA treatment in terms of modifying OA progression are conflicting[33, 86], so much so the American Academy of Orthopaedic Surgeons recently recommended that HA should not be used for patients with OA of the knee[89]. However, as described previously, PRG4 has a synergistic friction reducing relationship with HA. Hence, the addition of PRG4 to current HA treatments could potentially be an improvement on the current viscosupplements in slowing or preventing the progression of OA. 27

46 1.7 Aims Aim 1: Determine the effect of different sliding interface materials, cartilage-glass and cartilage-cartilage, on the cartilage lubricating ability of PRG4 and HA. Aim 2i: Assess the adsorption of disulfide-bonded PRG4 multimers and monomers to a native articular cartilage surface, and evaluate their in vitro cartilage boundary lubricating ability. Aim 2ii: Assess the in vitro cartilage boundary lubricating ability of disulfide-bonded PRG4 multimers and monomers in combination with HA. Aim 3: Assess the adsorption of rhprg4 to a native articular cartilage surface, and evaluate the in vitro cartilage boundary lubricating ability of rhprg4, alone and in combination with HA. 28

47 CHAPTER 2: Effect of Counterface on Cartilage Boundary Lubricating Ability by Proteoglycan 4 and Hyaluronan: Cartilage-Glass vs. Cartilage-Cartilage 2.1 Abstract The objective of this study was to determine the effect of different sliding interface materials (counterface) on the cartilage lubricating ability of proteoglycan 4 (PRG4) and hyaluronan (HA) by measuring the kinetic coefficient of friction on cartilage-glass and cartilagecartilage interfaces over a wide range of sliding velocities. The lubrication properties of PRG4 and HA were assessed at cartilage-glass and cartilage-cartilage interfaces using a previously described test setup with a stationary area of contact. Samples were articulated at varying effective sliding velocities of 10, 3, 1, 0.3, 0.1, 0.01 mm/s. The response of PRG4 and HA as effective friction-reducing cartilage boundary lubricants was varied and was dependent primarily on the test counterface. At a cartilage-glass interface HA demonstrated no friction reducing ability compared to PBS, and PRG4 appeared just as effective as SF. Conversely, at a physiological cartilage-cartilage interface, HA and PRG4 both effectively reduced friction compared to PBS. These results suggest that use of synthetic surfaces in studying cartilage boundary lubrication may not be appropriate. As such, care should be taken when interpreting such data, specifically when comparing to in vitro data obtained at a cartilage-cartilage interface, and especially when extrapolating to in vivo situations. 29

48 2.2 Introduction Synovial fluid (SF) is a viscous ultrafiltrate of blood plasma found within diarthrosis joints. The role of SF in the joint is manifold; it reduces the friction and wear between the articulating interfaces, provides nutrients to the articular cartilage, and has a dampening effect when compressive forces are exerted due to its viscoelastic feature[11, 90]. SF is composed of lubricating macromolecules such as proteoglycan 4 (PRG4) and hyaluronan (HA)[13]. PRG4, also known as lubricin[91] and superficial zone protein[52], is a mucin-like glycoprotein that exists in SF and at the surface of the articular cartilage; PRG4 is required for joint lubrication and health[56]. HA, a glycosaminoglycan polymer that is also present in SF and at the surface of cartilage, is another cartilage lubricating molecule that is key to the viscoelastic properties of SF[27]. Both PRG4 and HA have both been shown to reduce friction in a boundary mode of lubrication at a cartilage-on-cartilage interface[13, 49]. Furthermore, PRG4 and HA synergistically reduce friction at a cartilage-on-cartilage interface to levels close to that of whole SF[13]. Joint lubrication can vary depending on the load applied, speed of articulation and/or the viscosity of the fluid between the articulating surfaces[39]. The opposing articular cartilage in a synovial joint makes contact over 10% of the total area[37], and the low sliding speeds at the articular cartilage surfaces result in a boundary mode of lubrication being predominant in these situations[46]. The boundary mode of lubrication occurs when surface contact occurs between two articulating interfaces, and lubricating molecules bound at the interface are required to reduce friction[3]. This is classically defined and characterised using Stribeck curve analysis by varying speed, load, and/or viscosity, which enables observations of changes to the coefficient of 30

49 friction at different stiff and impermeable synthetic interfaces[38]. However, this may not be appropriate for soft and porous tissues such as the articular cartilage. It has recently been shown that there are differences in lubricity measurements when comparing a glass probe (stiff and impermeable) against a hydrogel (soft and permeable) versus measuring the interface between two hydrogels[38]. Indeed, the presence of a material with hydrogel properties, with or without the glass probe, was unable to recreate the normal characteristics of a Stribeck curve. This suggests Stribeck analysis may not be feasible and/or appropriate for a cartilage-cartilage interface. To assess the lubricating ability of molecules such as PRG4 and HA, several methods and protocols have been developed. Ex vivo methods are generally considered the most physiologically relevant in studying joint friction, such as a whole joint pendulum[43]. When using this method of analysis a wide range of lubrication modes are operative enabling physiological relevant analyses of lubricants, although it is not possible to define the exact mode of lubrication. The use of in vitro stationary area of contact test setups, with slow articulating velocities, enables analysis of lubricants in the boundary mode of lubrication. A latex-glass interface has been used with a rotational geometry[45, 46] and cartilage-glass interface using linear reciprocating geometry has also been utilised[47, 48]. The use of a cartilage-cartilage stationary area test setup with a rotational geometry enables analysis of boundary lubricants at a physiologically relevant interface[44]. The use of such a rotational cartilage-cartilage setup does not allow for Stribeck analysis, as has been performed with the cartilage-glass linear reciprocating system[48], due to the soft and porous nature of cartilage[38] and stationary rotational contact area[44]. However, the use of a cartilage-cartilage interface enables interaction of lubricants with articular cartilage surfaces, that may not occur with synthetic materials[42]. 31

50 The results from these various studies, and conclusions drawn, related to the boundary lubricating properties of PRG4 and HA have varied. PRG4 has consistently demonstrated friction reducing ability under boundary lubrication conditions at cartilage-cartilage[13], cartilage-glass[47, 48], as well as latex-glass[46] interfaces. Conversely, HA has been reported to function as a boundary lubricant for only the cartilage-cartilage interface[13, 49], but not for a latex-glass interface with a rotational geometry under slow sliding velocities[46] or in a preliminary study at a linear reciprocating cartilage-glass[48] interface. In the latter study, HA s lubricating abilities were attributed entirely to its viscous properties using Stribeck type analysis. These reported discrepancies between HA s cartilage boundary lubricating ability at a cartilagecartilage interface, and those using a synthetic interface, may have been due to an inability of HA to functionally interact with the test surfaces. Indeed, while each set of test interfaces has their experimental advantages and disadvantages for the study of cartilage boundary lubrication, those interfaces containing synthetic surfaces may allow for some, but perhaps not all of the physiological interactions that occur between lubricants and cartilage surfaces[38, 42]. The objective of this study was to determine the effect of different sliding interface materials (counterface) on the cartilage lubricating ability of PRG4 and HA by measuring the kinetic coefficient of friction on cartilage-glass and cartilage-cartilage interfaces over a wide range of sliding velocities. 32

51 2.3 Methods Solution Preparation PRG4 PRG4 was prepared from fresh skeletally mature bovine stifle joints obtained from a local abattoir (Calgary, AB, Canada)[13]. Cartilage discs were cultured in Dulbecco s Modified Eagle s Medium (Life Technologies, Carlsbad, CA) with 0.01% bovine serum albumin, with the addition of 25 μg/ml of ascorbic acid and 10 ng/ml of recombinant human transforming growth factor-β[92]. Purification of the media was then performed using diethylaminoethanol anion exchange chromatography (GE Healthcare Life Sciences, Baie d Urfe Quebec, Canada)[53, 92] using a gravity flow and a salt gradient. Batch elutions at M NaCl were retained, filtered with a 100 kda filter and stored at -80 o C. Total concentration was determined by bicinchoninic acid assay (Sigma-Aldrich, St. Louis, MO). The purity of the concentrated and filtered solution was confirmed using 3-8% Tris-Acetate sodium dodecyl sulfate polyacrylamide gel electrophoresis followed by Simply Blue protein stain (Life Technologies, Carlsbad, CA) and densitometry analysis (Image J, Bethesda, MD)[54, 61]. PRG4 was prepared in phosphate buffered saline (PBS) at a physiological concentration of 450 μg/ml[93] HA HA of 1.5 MDa (Lifecore Biomedical, Chaska, MN) was prepared in PBS at a physiological concentration of 3.33 mg/ml[30]. 33

52 2.3.2 Boundary Lubrication Test Cartilage-cartilage test sample preparation Cartilage cores (radius (R) = 6 mm) and annuli (R outer = 3.2 mm and R inner = 1.5 mm) were prepared from osteochondral blocks harvested from mature bovine stifle joints, obtained as above[13, 44]. Samples were rinsed vigorously overnight in PBS at 4 C to rid the articular surface of residual SF[13, 44], prior to lubrication testing in PBS. Samples were then bathed in 0.3 ml of the subsequent test lubricants (core bathed in 0.2 ml, annulus bathed in 0.1 ml), completely immersing the cartilage surface, and left at 4 C overnight prior to the next day s lubrication test. The samples were again rinsed with PBS after each test before incubation in the next test lubricant Cartilage-glass test sample preparation Cartilage annuli were prepared as described above. Borosilicate glass cores (R = 6 mm) were polished to a root mean square (RMS) surface roughness of 6.06 ± 0.76 nm[47] (measured using the ZeScope Optical Profiler, Zygo, Santa Clara, CA) Lubrication Test The Bose ELF 3200 (BOSE ElectroForce Systems Group, Eden Prairie, MN) was used to analyse the boundary lubrication ability of test solutions at each interface, essentially as described previously[13, 44]. Briefly, in each test setup all samples were compressed to 18% of the total cartilage thickness. Samples were allowed to stress-relax for 40 minutes to enable fluid depressurisation of the interstitial fluid. Samples were then rotated +/- 2 revolutions at varying 34

53 effective sliding velocities (v eff = ω R eff, where R eff is effective radius (= 2.4 mm) and ω is angular velocity[13, 44]) of 10, 3, 1, 0.3, 0.1, 0.01 mm/s; with a pre-sliding duration of 120 s between rotation. Two sequential test sequences (n=4) were employed at both cartilage-glass and cartilage-cartilage interfaces: 1) PBS, PRG4, SF; and 2) PBS, HA, SF Data Analysis Two kinetic coefficients of friction were calculated from the measured torque (τ) during the 2 nd revolution and the axial load (N). The first kinetic coefficient of friction was derived from the classic formula, and was determined as <μ kinetic > = τ/(r eff N) and normalised to the load during the 2 nd revolution. The second kinetic coefficient of friction was determined as <μ kinetic,neq > = τ/(r eff N eq ), normalised to the equilibrium axial load in cartilage (N eq )[44]. N eq is the equilibrium normal force measured after compression and stress relaxation prior to articulation of the test surfaces, which neglects the viscoelastic and hydrodynamic forces that cause variations in the instantaneous measured N during the revolutions. Data are presented as mean ± SEM. Two-way ANOVA was used to assess the effect of lubricant and velocity on kinetic friction, with Tukey post-hoc testing to compare lubricants. 35

54 2.4 Results Test Characterisation The characteristics of the measured torque were similar at both counterfaces, while axial load varied. Measured τ data reached a steady state during the 2 nd revolution for both configurations, shown for 0.01, 0.3, and 10 mm/s in PBS and SF (Figure 2-1A & B, Figure 2-2A & B). The axial load also reached a steady-state value for cartilage-glass (Figure 2-1C & D), but not for cartilage-cartilage where it demonstrated cyclical behaviour during revolutions (Figure 2-2C & D). Figure 2-1: Friction data acquired on cartilage-glass interfaces. Torque (τ) profiles obtained during two rotations at three velocities (0.01, 0.3 and 10 mm/s) on PBS (A) and SF (B). Instantaneous axial load (N) profiles acquired simultaneously for PBS (C) and SF (D). 36

55 Figure 2-2: Friction data acquired on cartilage-cartilage interfaces. Torque (τ) profiles obtained during two rotations at three velocities (0.01, 0.3 and 10 mm/s) on PBS (A) and SF (B). Instantaneous axial load (N) profiles acquired simultaneously for PBS (C) and SF (D) Cartilage-on-Glass Friction For all tests <μ kinetic > and <μ kinetic,neq > values were highest in PBS and lowest in SF, values in HA were similar to PBS and values in PRG4 were similar to SF. <μ kinetic > values appeared to vary slightly over increasing speed, but followed similar trends across all lubricants (Figure 2-3A). <μ kinetic > values varied with test lubricant and velocity (both p < 0.010), with no interaction (p = 0.701). Values of <μ kinetic > in HA were higher than SF (p < 0.001) and similar to PBS (p = 0.602), while <μ kinetic > in PRG4 was lower than PBS (p < 0.001) and similar to SF (p = 37

56 0.999). Values of <μ kinetic > in PRG4 were significantly lower than in HA (p < 0.001). <μ kinetic,neq > values appeared to increase as the effective velocity increased (Figure 2-3B). <μ kinetic,neq > values varied with test lubricant and velocity (both p < 0.001), with no interaction (p = 0.899). Differences in lubricants were similar to observed in <μ kinetic >, where values of <μ kinetic,neq > in HA were higher than in SF (p = 0.001) and similar to PBS (p = 0.399), while <μ kinetic,neq > PRG4 was lower than PBS (p < 0.001) and similar to SF (p = 0.868). Values of <μ kinetic,neq > in PRG4 were again significantly lower than in HA (p < 0.001) Cartilage-on-Cartilage Friction For all tests, <μ kinetic > and <μ kinetic,neq > values were highest in PBS and lowest in SF, and values in HA and PRG4 were intermediate. <μ kinetic > values appeared to decrease as velocity increased (Figure 2-3C). Values of <μ kinetic > varied with test lubricant and velocity (both p < 0.010), with no interaction (p = 0.947). Values of <μ kinetic > in HA, PRG4 and SF were all lower than in PBS (p < 0.010). Values in SF were lower than in PRG4 (p = 0.001), but were similar to in HA (p = 0.132), and values in PRG4 and HA were similar (p = 0.456). <μ kinetic,neq > values varied dependent on lubricant; those in PBS and PRG4 appeared to increase as velocity increased whereas HA and SF values appeared to decrease with increasing velocity (Figure 2-3D). <μ kinetic,neq > values varied with test lubricant (p < 0.001) but remained unchanged with respect to velocity (p = 0.335), with no interaction (p = 0.655). Values of <μ kinetic,neq > in HA, PRG4 and SF were again all lower than PBS (p < 0.05). Values in SF were lower than PRG4 (p < 0.001), but statistically similar to HA (p = 0.408), and values in PRG4 and HA were statistically different in this case (p < 0.05) with those in HA appearing to be lower at higher velocities compared to PRG4. 38

57 Figure 2-3: Cartilage-glass; average kinetic friction coefficient <μ kinetic > (A) and average normalised kinetic friction coefficient <μ kinetic,neq > (B). Cartilage-cartilage; average kinetic friction coefficient <μ kinetic > (C) and average normalised kinetic friction coefficient <μ kinetic,neq > (D). 39

58 2.5 Discussion The results of this study demonstrate the friction-reducing ability of putative cartilage boundary lubricants can be affected by the counterface of in vitro tests; specifically, in terms of magnitude and velocity dependent behaviour of kinetic coefficients of friction, and their relative effectiveness to reduce friction. At a cartilage-glass interface, HA demonstrated no friction reducing ability compared to PBS, and PRG4 appeared just as effective as SF. Conversely, at a physiological cartilage-cartilage interface, HA and PRG4 both effectively reduced friction compared to PBS, with HA demonstrating similar friction reducing ability to that of whole SF at higher velocities. This study analysed lubricants of interest, HA and PRG4, at cartilage-cartilage and cartilage-glass interfaces using an in vitro rotational test setup. This varies from other in vitro test setups such as a linear reciprocating geometry[47] or other rotational setups that use nonphysiological interfaces[45, 46]. The in vitro setup employed here does not wholly represent the ex vivo test setup, where many modes of lubrication are operative[43], however, it does allow for the friction measurement of lubricants of interest in a boundary mode. Furthermore, the use of the same test protocol[44] with varying interfaces over a range of velocities enabled direct comparisons of the friction reducing ability of PRG4 and HA under varying conditions. Two friction coefficient calculations were employed due to the difference in the measured axial load behaviour between interfaces during revolutions. The cyclical load observed for cartilagecartilage has been observed previously and attributed in large part to fluid pressurisation[44]. Although the resulting values between <μ kinetic > and <μ kinetic,neq > varied, the trends observed between different lubricants and conclusions drawn between both results were consistent. 40

59 The ability of PRG4 to lubricate, over varying speeds, at both cartilage-glass and cartilage-cartilage interfaces agrees with previous studies at cartilage-glass[48], latex-glass[46] and cartilage-cartilage[13] interfaces. PRG4 was able to lubricate efficiently at lower speeds at a cartilage-cartilage interface, which is consistent with previous results[13]. Interestingly, at higher speeds where a transition away from the boundary mode of lubrication could occur, even with a stationary contact area, and the viscosity of the lubricant of interest could be important, the lubricating ability of PRG4 diminished and was similar to that of PBS. The observed boundary lubricating ability of HA in this present study was dependent on the test interface. At a cartilage-glass interface, HA was unable to reduce friction and was similar to that of PBS, which in general agrees with previous studies thus demonstrating HA has no boundary lubricating properties at synthetic surfaces[46, 48]. However, at a cartilage-cartilage interface HA was able to reduce friction compared to PBS, and was similar to that of PRG4 at low speeds[13]. Furthermore, at higher speeds HA functioned as a more effective lubricant than PRG4 and had coefficient of friction values closer to that of whole SF. This could be due to HA s viscosity, which is much greater than that of PRG4 and similar to that of SF, contributing to a greater extent at higher effective velocities of articulation where some type of mixed or fluid-film mode of lubrication could be operative in the test setup[3]. Collectively, these results are consistent with previous studies demonstrating HA does not function as a boundary lubricant at a cartilage-glass interface, but does so at a cartilage-cartilage interface in a similar manner to PRG4 at slow velocities. By definition a boundary lubricant must be able to bind to an interface to function[42, 56]. An adaptive mechanically controlled lubrication mechanism was postulated to occur at the articular cartilage surface[42]. This proposed mechanism is based on a PRG4 + HA complex 41

60 having an adaptive role under varying compression. Under low loads and shear velocities PRG4 is adsorbed to the cartilage surface and is entangled with HA to create a PRG4 + HA complex, thus providing a source of boundary lubrication[42]. During higher loads the HA is mechanically trapped providing wear protection, and PRG4, due to its weaker surface attachment, is able to be sheared and redistributed to areas of higher shear forces and normal pressures[39-42]. Recently, non-covalently bound HA-binding peptides have been shown to enable improved adsorption of HA to tissues and biomaterial surfaces and provided friction reducing capabilities at both healthy and osteoarthritic human cartilage interfaces in a boundary mode of lubrication[51], further emphasising HA s ability to function as a cartilage boundary lubricant. Similarly, PRG4 has also been chemically modified to enhance its cartilage adsorption without altering its cartilage boundary lubricating ability[94, 95]. Whether through a putative adaptive multimodal mechanism or chemical modifications strategies to enhance surface adsorption, the friction reducing ability and mechanism of PRG4 and HA at tissues and biomaterials, such as cartilage and contact lenses respectively, are an area of great interest that warrants future research. In conclusion, the results obtained in this study suggest that the use of synthetic surfaces in experimentally studying cartilage boundary lubrication may not be appropriate for in vitro friction testing. A cartilage-cartilage test interface allows for physiological interactions between lubricants and cartilage surfaces. In addition, care should be taken when interpreting data from different in vitro test geometries and test interfaces, specifically when comparing to data obtained from an in vitro cartilage-cartilage interface, and especially when extrapolating it to in vivo situations. 42

61 2.6 Acknowledgements This chapter is in preparation for submission to the Journal of Orthopaedic Research: Abubacker S, McPeak AE, Dorosz SD, Egberts P, Schmidt TA. This work was supported by funding from the Natural Sciences and Engineering Research Council of Canada, the Canadian Arthritis Network, and the Alberta Innovates Health Solutions Team in Osteoarthritis. All authors contributed to the conception and design of the original study and approved the final submitted manuscript. SA, AM and SD were responsible of the data acquisition and analysis. Data analyses and interpretation was performed by SA, AM, SD, PE and TS. The article was first drafted by SA, and critically reviewed by all. TS obtained funding for the study, and all authors takes full responsibility for the integrity of the work as a whole. We would also like to thank the following individuals for their contributions to this manuscript: Dr. Bernard Guest, Mr. William Matthews, Mr. Mickey Horvath and Mr. Mark Toonen for aiding in the production of the glass pieces and use of ZeScope equipment. 43

62 CHAPTER 3: Effect of disulfide bonding and multimerisation on Proteoglycan 4 s cartilage boundary lubricating ability and adsorption 3.1 Abstract Objectives: The objectives of this study were to assess the cartilage boundary lubricating ability of (1) non reduced (NR) disulfide-bonded PRG4 multimers versus PRG4 monomers, and their dose-dependency, and (2) NR versus reduced and alkylated (R/A) PRG4 monomers, as well as (3) assess the ability of NR disulfide-bonded PRG4 multimers versus PRG4 monomers to adsorb to an articular cartilage surface. Methods: PRG4 purified from bovine cartilage explant culture was separated into two preparations of interest, PRG4 multimer enriched (PRG4Multi+) and PRG4 multimer deficient (PRG4Multi ), using size exclusion chromatography (SEC) and characterised by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The cartilage boundary lubricating ability of PRG4Multi+ and PRG4Multi was compared at a physiological concentration (450 μg/ml), and assessed at a range of pathophysiological concentrations (45, 150 and 450 μg/ml). R/A PRG4Multi and NR PRG4Multi were also evaluated at a concentration of 450 μg/ml. Immunohistochemistry with anti-prg4 monoclonal anti-body (mab) 4D6 was performed to visualise the adsorption of PRG4 preparations to the surface of PRG4 depleted articular cartilage explants. Results: Separation into enriched populations of PRG4Multi+ and PRG4Multi was achieved using SEC and was confirmed by SDS-PAGE. PRG4Multi+ and PRG4Multi both functioned as effective friction-reducing cartilage boundary lubricants at 450 μg/ml; with 44

63 PRG4Multi+ being more effective than PRG4Multi. PRG4Multi+ lubricated in a dosedependent manner, however PRG4Multi did not. R/A PRG4Multi lubricated in a similar matter to NR PRG4Multi at 450 μg/ml. PRG4 containing solutions showed 4D6 immunoreactivity at the articular surface and cut edge; the immunoreactive intensity of PRG4Multi+ appeared to be similar to that of SF, whereas PRG4Multi appeared to have less intensity. Conclusions: These results demonstrate that the inter-molecular disulfide-bonded multimeric structure of PRG4 is important for its ability to adsorb to a cartilage surface and function as a boundary lubricant. These findings contribute to a greater understanding of the molecular basis of articular cartilage boundary lubrication of PRG4. Elucidating the PRG4 structure-lubrication function relationship will further contribute to the understanding of PRG4 s role in diarthrodial joint homeostasis and disease. 45

64 3.2 Introduction Proteoglycan 4 (PRG4) is a mucin-like glycoprotein synthesised by cells in articular cartilage, meniscus, synovial lining and tendons[54]. It is encoded for by the PRG4 gene[96], and is analogous to lubricin[91], superficial zone protein (SZP)[52] and megakaryocyte stimulating factor (MSF)[97]. PRG4 is present in synovial fluid (SF) and at the surface of articular cartilage where it functions as a critical boundary lubricant necessary for joint health[56], in a dosedependent manner[13]. In addition, PRG4 provides protection by preventing protein deposition and cell adhesion[69]. The role played by PRG4 is critical in reducing the friction occurring at the bearing surfaces, which prevents the degradation of cartilage and adhesion of cartilage surfaces when boundary lubrication occurs. Indeed, mutations in the PRG4 gene results in an autosomal recessive disorder in humans, camptodactylyarthropathy-coxa vara-pericarditis (CACP), which results in juvenile-onset, non-inflammatory, precocious joint failure[68]. Furthermore, alteration in PRG4 concentration within SF due to primary[71, 73, 74] and secondary OA[75, 76], in humans and animal models, has been shown to affect joint integrity and lubrication. PRG4 shares functionally determinant structural characteristics similar to that of many other mucins[54, 58, 62, 98]. PRG4 is composed of 12 exons, with exon 6 being the highly glycosylated mucin-like domain that makes up ~50% of the molecule s mass due to the extensive O-linked oligosaccharide substitutions[34, 54, 56, 99]. This mucin-like domain is functionally important and determinant as enzymatic removal of the O-linked oligosaccharides, thought to provide repulsive hydration forces, results in diminished lubricating function[34, 54, 58]. The cysteine rich N- and C-terminal domains facilitate the formation of functionally determinant intra- and inter-molecular disulfide bonds[34, 58, 99] and therefore the formation of PRG4 46

65 dimers and multimers[57, 59, 62]. Indeed, the cysteine-rich N-terminal has been shown to enable dimerisation, entanglement and self-aggregation[34, 100]. The ability to form disulfide-bonded multimers in general is critical to various mucins functions[57, 59, 62], and this also appears true for PRG4[34, 54, 56, 99]. PRG4 exists in both monomeric and multimeric forms, and these may demonstrate differential abilities to adsorb to the surface of cartilage and function as an effective frictionreducing boundary lubricant. Monomers and multimers have been identified in bovine synovial fluid, and their MW reported to range from kda for multimers and and kda for monomers when purified from media conditioned by bovine articular cartilage explants[54, 60, 61]. PRG4 preparations purified from such media, and containing both multimers and monomers, have consistently demonstrated cartilage boundary lubricating ability[13]. Reduction and alkylation (R/A) of such PRG4 preparations, which breaks both intraand inter-molecular disulfide bonds, has been shown in a preliminary study to increase the friction at a cartilage-cartilage interface by ~34%[63]. However, the use of guanidine hydrochloride, a denaturing buffer often employed for R/A, has been shown to diminish the cartilage lubricating ability of PRG4 on its own[101]. Therefore, the role of inter-molecular disulfide bonds on PRG4 s cartilage boundary lubricating ability, i.e. non-reduced (NR) multimers versus NR monomers, as well as intra-molecular bonds, i.e. NR monomers versus R/A monomers, remains unknown. Boundary lubricants by definition are able to adsorb to the articulating surfaces on which they reduce friction, indeed PRG4 is able to adsorb on the surface of cartilage[100]. The cysteine rich C-terminal is thought to enable cartilage attachment[34, 100], as well as further disulfidebridging interaction with larger polypeptide chains[58]. Consistent with this, R/A inhibits PRG4s 47

66 ability to bind to cartilage surfaces[100, 102]. However, as with cartilage lubricating function, it remains unclear what effect inter-molecular disulfide bonds, i.e. NR multimers versus NR monomers, has on PRG4 s ability to adsorb to a cartilage surface. The objectives of this study were to assess the cartilage boundary lubricating ability of 1) non reduced (NR) disulfide-bonded PRG4 multimers versus PRG4 monomers, and their dosedependency and (2) NR versus reduced and alkylated (R/A) PRG4 monomers, as well as 3) assess the ability of NR disulfide-bonded PRG4 multimers versus PRG4 monomers to adsorb to a cartilage surface. 48

67 3.3 Methods Lubricant Preparation and Characterisation PRG4 Cartilage for PRG4 preparation was obtained from fresh skeletally mature bovine stifle joints obtained from a local abattoir (Calgary, AB, Canada), as described previously[13]. In brief, bovine stifle joints with intact articular capsules were obtained and discs with the intact articular surface of the cartilage were harvested. The cartilage discs were then cultured in Dulbecco s Modified Eagle s Medium with 0.01% bovine serum albumin for 28 days, with the addition of 25 μg/ml of ascorbic acid and 10 ng/ml of recombinant human transforming growth factor-β to the media[92]. Purification of PRG4 from the conditioned media was then performed using diethylaminoethyl anion exchange chromatography[53, 92]. The PRG4-rich M NaCl eluant was retained, concentrated with a 100 kda filter (EMD Millipore, Billerica, MA) and stored at -80 o C. The purity of the concentrated and filtered solution was confirmed using 3-8% Tris-Acetate sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), followed by protein stain and western blotting with anti-prg4 anti-body LPN[54, 98] (Life Technologies, Carlsbad, CA). The concentration of this PRG4 preparation was then determined by bicinchoninic acid protein assay (BCA) (Sigma-Aldrich, St. Louis, MO)[49] PRG4 separation Size exclusion chromatography (SEC) was employed to separate PRG4 multimers and monomers. Sephacryl S-500HR High Resolution gel filtration media packed in a XK16/100 column (GE Healthcare Life Sciences, Baie d Urfe, QC, Canada) was used in conjunction with a fast protein liquid chromatography (FPLC) apparatus (ÄKTA FPLC ; GE Healthcare Life Sciences). 49

68 Phosphate buffered saline (PBS) and 0.25% 3-[(3-cholamidopropyl) dimethylammonio]-1- propanesulfonate (CHAPS), a zwitterionic detergent (Sigma-Aldrich, St. Louis, MO) was used as the column buffer. The column was equilibrated in the PBS % CHAPS buffer at a flow rate of 0.2 ml/min, then PRG4 also suspended in PBS % CHAPS was injected and 2 ml fractions were subsequently collected. UV absorbance at 280 nm was monitored, and fractions of interest from three predominant peaks were pooled together (Figure 3-1). Pooled fractions were filtered, concentrated and buffer exchanged using Amicon Ultra-15 Centrifugal 30 kda cut-off filters (EMD Millipore) using distilled water (dh 2 O), centrifuged at 37 o C, and re-suspended in PBS. The concentration was determined by BCA assay. Figure 3-1: SEC chromatogram showing three peaks of interest. Peak 2 is shown to be PRG4Multi+ and peak 3 is shown to be PRG4Multi. 50

69 PRG4 Characterisation SDS-PAGE and western blotting was used to characterise the apparent MW and size distributions of immunoreactive PRG4 species in each pool, as described previously[54, 60, 98]. Materials and equipment for 3-8% SDS-PAGE western blotting and protein staining were obtained from Life Technologies (Carlsbad, CA). Briefly, non-reduced samples were mixed in sample buffer and heated to 70 o C for 10 minutes. These were then loaded onto a 3-8% SDS- PAGE gel followed by protein stain, and/or western blotting using anti-prg4 Ab LPN and mab 4D6 (a generous gift from Dr. Messersmith, Northwestern University, IL), as described previously[54, 60, 95, 98]. Protein stain (Figure 3-2A) indicated the peak 2 preparation (PRG4Multi+) contained a distinct high molecular weight (MW) band near the top of the gel (***), a ~1MDa band (**) (as shown previously[98]), and a weaker ~460kDa monomeric[98] band (*). The peak 3 preparation (PRG4Multi ) lacked the high putative multimeric MW bands and was enriched in the 460 kda non-reduced monomer band (*). Tandem mass spectrometry (MS/MS) confirmed all these bands to be PRG4 (data not shown). Western blotting indicated these preparations contained both LPN (Figure 3-2B) and 4D6 (Figure 3-2C) immunoreactive species, with a MW distribution similar to that of the protein stain. Peak 2 preparation was enriched in the high MW species (***, **), whereas peak 3 preparation contained primarily the ~460kDa monomer (*). Peak 1 was not resolved on the 3-8% gel as the molecule was too large to enter, though MS/MS indicated PRG4 species were present in those pooled fractions (see Appendix B). 51

70 Figure 3-2: SDS-PAGE of PRG4Multi and PRG4Multi+ preparations; Protein stain (A) and western blotting with Ab LPN (B) and mab 4D6 (C). This characterisation confirms that the two preparations of interest were generated; an enriched population of high MW species of PRG4 (peak 2), which still contained some smaller species, as well as a solution that appeared to lack the high MW PRG4 and was predominantly lower MW species. Henceforth, the terms PRG4 multimer enriched preparations (PRG4Multi+) for peak 2 and PRG4 multimer deficient preparations (PRG4Multi ) for peak 3 will be used to describe these preparations. 52

71 PRG4 Reduction and Alkylation Purified PRG4Multi (1 mg) dissolved in PBS was reduced and alkylated (R/A), essentially as described before[63, 98]. The sample was incubated with 100 mm dithiothreitol in PBS for 2 hrs at 60 O C and 400 mm iodoacetate for 2 hrs at room temperature. For protein recovery, the sample was then incubated in three volumes of ice cold ethanol (EtOH) overnight at -20 O C. It was then centrifuged at RPM for 20 min, with most of the supernatant being removed. The sample was then washed three times with EtOH whilst being centrifuged at RPM for 20 min. In between each wash the supernatant was removed after each centrifuge. Sample was then centrifuged at 37 O C and re-suspended in PBS. The concentration was then determined by BCA protein assay Boundary Lubrication Tests Sample Preparation Fresh osteochondral samples were prepared for friction testing from the patellofemoral groove of skeletally mature bovine stifle joints, as described previously[13, 44]. Briefly, cores (radius = 6 mm) and annuli (R outer = 3.2 mm and R inner = 1.5 mm) were prepared from harvested osteochondral blocks[44]. Samples were then rinsed vigorously overnight in PBS at 4 C to rid the articular surface of residual SF (confirmed by lubrication testing[13, 44]) prior to lubrication testing in PBS. Samples were then bathed in 0.3 ml of the subsequent test lubricants (core bathed in 0.2 ml, annulus bathed in 0.1 ml), completely immersing the cartilage, and left at 4 C overnight prior to the next day s lubrication test. The samples were again rinsed with PBS after each test before incubation in the next test lubricant. 53

72 Lubrication Test The Bose ELF 3200 was used to analyse the cartilage boundary lubrication ability of the PRG4 preparations, using the previously described in vitro cartilage-on-cartilage friction test[13, 44]. Briefly, all samples were compressed to 18% of the total cartilage thickness. Samples were allowed to stress-relax for 40 minutes to enable fluid depressurisation of the interstitial fluid. The samples were then rotated at an effective velocity of 0.3 mm/s (shown to maintain boundary mode lubrication at a depressurised cartilage-cartilage interface)[44] at +/-2 revolutions. Samples were then left in a pre-sliding duration (Tps) of 1200, 120, 12 and 1.2 seconds (s). Samples were rotated after each subsequent stationary period in the +/- 2 revolutions. The test sequence was then repeated in the opposite direction of rotation, -/+ 2 revolutions Lubricant Test Sequences Cartilage Boundary Lubricating Ability of PRG4Multi+ vs. PRG4Multi. The effects of inter-molecular disulfide bonds on the cartilage boundary lubricating ability of PRG4 was determined by comparing NR PRG4Multi+ to PRG4Multi at a physiological concentration of 450 μg/ml in PBS[93]. Furthermore, the potential dose-dependent cartilage boundary lubricating ability of PRG4Multi+ and PRG4Multi preparations was assessed at concentrations of 45, 150 and 450 μg/ml. PBS served as the negative control lubricant and bovine SF (Animal Technologies, Tyler, TX) served as a positive control lubricant. Test Sequence 1 (n = 7): PBS, PRG4Multi, PRG4Multi+, SF. Test Sequence 2 (n = 4): PBS, 45 μg/ml, 150 μg/ml, 450 μg/ml, then SF. Test Sequence 3 (n = 4): PBS, 45 μg/ml, 150 μg/ml, 450 μg/ml, then SF. 54

73 Cartilage Boundary Lubricating Ability of R/A PRG4Multi and NR PRG4Multi. To assess the effects of intra-molecular disulfide bonds, the cartilage boundary lubricating ability of R/A PRG4Multi and NR PRG4Multi was analysed at a concentration of 450 μg/ml in PBS. Test Sequence 4 (n = 4): PBS, R/A PRG4Multi, NR PRG4Multi, SF Immunohistochemistry Sample preparation Intact articular cartilage discs (n = 15, diameter = 6 mm) were harvested from bovine stifle joints obtained from a local abattoir. Fresh intact cartilage discs acted as a natural positive control. These were embedded in media (Tissue Tek OCT, Sakura, Torrance, CA) and snap frozen in isopropanol cooled in liquid nitrogen (labelled as fresh ) Specimen processing Cartilage discs devoid of PRG4 at the articular surface were prepared as described previously[103]. Briefly, the cartilage discs were placed in PBS and shaken vigorously overnight at 4 C to rid the articular surface of residual PRG4, and subsequently frozen at -80 C to prevent further production of PRG4 from viable chondrocytes. Discs were then thawed, and again shaken overnight at 4 C, before incubation with lubricants of interest. Discs were incubated over night at room temperature in test solutions of interest[93]; PBS (negative control), 450 μg/ml, 450 μg/ml, and SF (positive control). Discs were then fixed in OCT and stored at -80 C. Sections (5 μm thick) were cut using a cryostat microtome (Microm HM550, Thermo Scientific, Waltham, MA) and placed on positively charged glass slides (Superfrost Plus 55

74 Adhesion Slides, Thermo Scientific). Sections were fixed in 4% paraformaldehyde in PBS and washed in PBS to remove OCT. Samples were blocked with 10% hydrogen peroxide in methanol, followed by 10% goat serum with 1% BSA in PBS in a humidity chamber. Samples were then incubated overnight in anti-prg4 mab 4D6[94] in 1.5% normal goat serum at a ratio of 1:100. Slides were washed again with PBS and incubated with secondary anti-body Alexa Fluor-594 rhodamine-conjugated goat-anti mouse IgG (Life Technologies, Carlsbad, CA) in 1.5% normal goat serum at a ratio of 1:1000. Finally, samples were washed with PBS, mounted with mounting medium containing the nuclear counterstain DAPI (Vectashield, Vector Laboratories, Inc., Burlingame, CA), and sealed with microscope cover slips (VWR Scientific Products, PA). Slides were imaged using Zeiss LSM 780 microscope (Carl Zeiss, Oberkochen, Germany) at a magnification of 20 objective (dry, 0.8 NA). Fluorescence images were obtained for both red (Alexa Fluor-594 rhodamine detected PRG4; excitation/emission of 590/617 nm) and blue (DAPI detected cell staining; excitation/emission of 358/461 nm) fluorescence Statistical Analysis Two coefficients of friction; static (μ static, Neq ; resistance of start-up motion from static condition) and kinetic (<μ kinetic, Neq >; resistance of steady sliding motion) were calculated for each lubricant, and were averaged between both the + and revolutions[13]. Unless indicated otherwise, data is presented as mean ± SEM. Statistical analysis was implemented with Systat12 (Systat Software, Inc., Richmond, CA). Test Sequence 1: Values of <μ kinetic,neq > at Tps = 1.2s were within 17.8 ± 3.7% of those at Tps = 1200s, therefore for clarity <μ kinetic,neq > data is presented at Tps = 1.2s only. ANOVA was 56

75 used to assess the effect of lubricant and Tps as a repeated factor on μ static,neq and <μ kinetic,neq >, with Tukey post hoc testing on <μ kinetic,neq > at Tps = 1.2 s. Test Sequence 2 & 3: Results from test sequence 2 and 3 were pooled together for clarity and ease of comparison. Values of <μ kinetic,neq > at Tps = 1.2s were within 23.5 ± 4.0% for PRG4Multi+ dose-response and 15.4 ± 3.5% for PRG4Multi dose-response of those at Tps = 1200s. ANOVA with Fisher post-hoc testing on <μ kinetic,neq > of the three PRG4 concentrations at Tps = 1.2 s were performed on Test Sequences 2 and 3. Test Sequence 4: Values of <μ kinetic,neq > at Tps = 1.2s were within 23.5 ± 4.2% of those at Tps = 1200s. ANOVA analysis was conducted similar to that of Test Sequence 1. 57

76 3.4 Results Boundary Lubrication Tests Cartilage Boundary Lubricating Ability of PRG4Multi+ vs. PRG4Multi PRG4Multi+ and PRG4Multi both functioned as effective friction reducing cartilage boundary lubricants at a concentration of 450 μg/ml, with PRG4Multi+ being more effective than PRG4Multi. μ static,neq varied with Tps and test lubricant (both p < 0.001), with no interaction (p = 0.663). Values increased with Tps and were consistently highest in PBS and lowest in SF. PRG4Multi+ and PRG4Multi were intermediate (Figure 3-3A). <μ kinetic,neq > values exhibited similar trends, varying with lubricant and Tps (both p < 0.001) increasing only slightly with Tps, with an interaction (p = 0.001). <μ kinetic,neq > values at Tps = 1.2s were greatest in PBS and lowest in SF. Values for PRG4Multi+ were significantly lower than PRG4Multi (p = 0.02), and both were intermediate and significantly different than PBS and SF (both p < 0.01) (Figure 3-3B). 58

77 Figure 3-3: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients PRG4Multi and PRG4Multi+ at 450 µg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p <

78 Concentration Dependent Cartilage Boundary Lubricating Ability of PRG4Multi+ and PRG4Multi PRG4Multi+ s effective friction-reducing cartilage boundary lubricating was dosedependent over the concentrations tested (45, 150, 450 μg/ml), while that of PRG4Multi was not. μ static,neq for PRG4Multi+ varied with Tps and test lubricant (both p < 0.001), with no interaction (p = 0.092). μ static,neq for PRG4Multi varied with Tps and test lubricant (both p < 0.01), with an interaction (p = 0.012). Values for both increased with Tps and were consistently highest in PBS and lowest in SF. <μ kinetic,neq > values exhibited similar trends. PRG4Multi+ varied with lubricant and Tps (both p < 0.001) increasing only slightly with Tps, with no interaction (p = 0.068). PRG4Multi also varied with lubricant and Tps (both p < 0.05) increasing only slightly with Tps, with no interaction (p = 0.309). <μ kinetic,neq > values at Tps = 1.2s were greatest in PBS and lowest in SF, with all PRG4 preparations being intermediate. Fischer post-hoc testing showed PRG4Multi+ at 450 μg/ml was significantly lower than PRG4Multi+ at 45 μg/ml (p = 0.008) and 150 μg/ml (p = 0.034), there was no difference between PRG4Multi+ at 45 μg/ml and 150 μg/ml (p = 0.384). No significant difference was found between PRG4Multi samples across all three concentrations of interest (p > 0.05) (Figure 3-4). 60

79 Figure 3-4: Kinetic <μ kinetic,neq > at Tps = 1.2 s dose-response friction coefficients of PRG4Multi and PRG4Multi+ at 45, 150 and 450 µg/ml. Where * is p < 0.01 and ** is p < 0.05 for PRG4Multi+. There are no significant differences across the different concentrations of PRG4Multi. ANOVA with Fisher s post-hoc. Sample size, n = Cartilage Boundary Lubricating Ability of R/A PRG4Multi and NR PRG4Multi R/A PRG4Multi and NR PRG4Multi both functioned as effective friction-reducing cartilage boundary lubricants, with no significant differences between the two lubricants. μ static,neq varied with Tps and test lubricant (both p < 0.005), with no interaction (p = 0.220). Values increased with Tps and were consistently highest in PBS and lowest in SF, with R/A PRG4Multi and NR PRG4Multi being intermediate (Figure 3-5A). <μ kinetic,neq > values 61

80 exhibited similar trends, varying with lubricant and Tps (both p < 0.001) increasing only slightly with Tps, with no interaction (p = 0.355). <μ kinetic,neq > values at Tps = 1.2s were greatest in PBS and lowest in SF. Values for R/A PRG4Multi and NR PRG4Multi were not significant between each other (p = 0.974) (Figure 3-5B). 62

81 Figure 3-5: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients NR PRG4Multi and R/A PRG4Multi at 450 µg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p <

82 3.4.2 Immunohistochemistry IHC (Figure 3-6) of fresh cartilage discs indicated a 4D6-immunoreactive layer of PRG4 localised at the articular surface but not at the cut surface. Vigorous shaking in PBS overnight appeared to remove the majority of PRG4. The 4D6-immunoreactivity was specific as no localisation was observed in the non-immune serum control. Shaken samples that were incubated in PRG4 containing solutions showed 4D6-immunoreactivity (red) at the articular surface and cut edge. The immunoreactive intensity of PRG4Multi+ appeared to be similar to that of SF, whereas PRG4Multi appeared to have less intensity. The 4D6-immunoreactive layer at the articular surface appeared to have a greater intensity to that observed at the cut edge of the explant for PRG4 containing solutions. Again, there was no immunoreactivity observed for the non-immune control. Cartilage discs incubated in PBS alone showed some immunoreactivity, though not to the same intensity as those incubated in PRG4 containing solutions. 64

83 Figure 3-6: Immunolocalisation of PRG4 at an articular cartilage surface. Key: Fresh samples taken directly from joint and snap-frozen (control). All other samples were shaken overnight in PBS at 4 o C, frozen over night at -80 o C, shaken again in fresh PBS at 4 o C and soaked in solutions of interest overnight at room temperature. Solutions: PBS (negative control), SF (positive control), PRG4Multi and PRG4Multi+, with PRG4 at physiological concentrations of 450 µg/ml. ( ) signifies negative samples that lacked primary mab 4D6 but contained secondary goat-anti-mouse Ab. Signal to anti-prg4 mab 4D6 is depicted by the red staining, chondrocyte cells are depicted by the blue DAPI staining. 65

84 3.5 Discussion The objectives of this study were to assess the cartilage boundary lubricating ability and the cartilage adsorption of NR disulfide-bonded PRG4 multimers versus NR PRG4 monomers. The results demonstrate the inter-molecular disulfide-bonded multimeric structure of PRG4 is important for its ability to adsorb to a cartilage surface and function as a boundary lubricant. PRG4 multimer enriched preparations (PRG4Multi+) reduced friction in a dose-dependent manner, and appeared to adsorb to the articular cartilage surface to a greater extent than multimer deficient preparations (PRG4Multi ). PRG4Multi still reduced friction and adsorbed to the articular cartilage surface, though not to the same extent as PRG4Multi+. These findings contribute to a greater understanding of the molecular basis of articular cartilage boundary lubrication of PRG4. The PRG4Multi+ and PRG4Multi preparations used in this study were generated using SEC. Chromatogram results indicated pooled peaks were not distinctly separated. As such, pooling of the chosen fractions could have resulted in cross-contamination in each of the PRG4Multi+ and PRG4Multi preparations. SDS-PAGE results indicated predominant preparations of PRG4Multi+ and PRG4Multi with some overlap between pooled samples. Subsequent evaluation of more select and fewer fractions, selected closer to the peak in each pool in an attempt to avoid cross-contamination, were performed. However, subsequent SDS- PAGE results showed no apparent improvements compared to the original pooling approach. Also, fractions collected from the peak shoulders found in peaks 2 and 3 were examined via SDS-PAGE and were again similar to those from the overall pool. Furthermore, losses were significant when trying to avoid cross-contamination and when analysing the shoulders within the peaks, therefore for practical reasons fractions were pooled as described. Additional 66

85 separation and characterisation method development would allow for further functional examination of different putative monomeric PRG4 species[61], as well as higher order (e.g. dimers, multimers) PRG4. Nevertheless, the methods developed and employed here resulted in the generation of enriched populations of PRG4 multimers/monomers that were appropriate for this study. The cartilage boundary lubricating ability of the PRG4Multi+ preparation used here appeared similar in its magnitude and dose-dependency to that previously reported for unseparated NR PRG4[13]. These results are consistent with previous studies that have shown lack of disulfide bonds reduces both PRG4s boundary lubricating ability[63] and attachment to the cartilage surface[100]. PRG4Multi did lubricate at cartilage-cartilage interface compared to PBS, however not to the same extent as PRG4Multi+ and not in a dose-dependent manner. Furthermore, NR PRG4Multi with their intra-molecular bonds intact functioned in a similar manner to R/A PRG4Multi that lacked those bonds. It remains unknown how these disulfide bonds interact to form dimers through either the C- and/or N-termini[54, 57, 59]. In addition, it is yet to be elucidated how these molecules interact with each other to form larger putative multimers and entanglements with other forms of PRG4, be it multimers or monomers, or even with other SF molecules such as hyaluronan[13, 49, 66]. Collectively, these data indicate intermolecular disulfide bonds and therefore PRG4 s multimeric structure plays a greater role in PRG4 s dose-dependent cartilage lubricating ability compared to its monomeric structure. One possible explanation for these observed differences is PRG4Multi+ s enhanced ability to adsorb and accumulate at the articular cartilage surface. PRG4Multi+ demonstrated an increased localisation at the articular cartilage surface. IHC provided qualitative information on PRG4Multi+ and PRG4Multi s ability to adsorb to an 67

86 articular cartilage surface; however quantity cannot be determined but these results do provide visual evidence for the extent of spatial adsorption at the cartilage surface. As discussed previously[103], shaken samples soaked in PBS as negative controls appeared to have minimal amounts of PRG4 at the cartilage surface. The immunoreactivity at the cartilage surface of PBS soaked samples was substantially less in intensity compared to fresh or repleted samples. The effective removal of the PRG4 at the surface by this method is further supported by the high friction coefficient values measured for these samples tested in PBS, compared to those of fresh samples tested in PBS[13, 44]. The approach of vigorous shaking in PBS to remove residual PRG4 was employed here to avoid potential alterations to the articular cartilage surface resulting from the use of other enzymatic, ionic, or mechanical methods[93, 94, 100]. Immunoreactivity was observed at the non-articular surface of samples soaked in PRG4 containing solutions. This is consistent with previous IHC results[103], demonstrating PRG4 s ability to attach to the articular cartilage surface, as well as cut surfaces. Future studies could consider incubation of osteochondral samples instead of cut cartilage explants to prevent adsorption to lower cut surfaces[103]. Future work is required to quantify the amount of PRG4Multi+ and PRG4Multi at the articular cartilage surface. One potential approach, previously employed, is the radiolabeling of PRG4[102] and quantifying the amount of PRG4 accumulated at the cartilage surfaces. Such studies would provide further insight into a possible relationship between an ability of PRG4Multi+ to accumulate at a cartilage surface and function as a dose-dependent boundary lubricant. The occurrence of altered PRG4 structural composition in OA SF, in terms of relative abundance of PRG4 multimers and monomers, and potential functional consequences in terms of cartilage boundary lubricating ability remains to be determined. Previous studies have 68

87 demonstrated that some OA SF patients lack normal levels of PRG4, have a HA MW distribution shifted lower and demonstrated diminished cartilage lubricating function that can be restored with PRG4[75, 76]. Perhaps some of the OA patients that do have normal overall PRG4 levels could lack normal PRG4 multimer/monomer distribution (i.e. diminished multimeric content) and therefore have deficient cartilage lubricating function. Future work could involve method development for the analysis of PRG4 structural composition in SF, as well as examination of the potential effect of PRG4 multimer/monomer distribution on PRG4 s synergistic friction reducing interaction with hyaluronan[13, 49]. Elucidating the PRG4 structure-lubrication function relationship in SF, will further contribute to the understanding of PRG4 s role in diarthrodial joint homeostasis and disease. 69

88 B1. Acknowledgements This chapter is in preparation for submission to Matrix Biology: Abubacker S, Ponjevic D, Matyas JR, Schmidt TA. This work was supported by funding from the Natural Sciences and Engineering Research Council of Canada, the The Arthritis Society & Canadian Arthritis Network (Grant: TGD ), the Alberta Innovates Health Solutions Team in Osteoarthritis. All authors contributed to the conception and design of the original study and approved the final submitted manuscript. SA and DP were responsible of the data acquisition and analysis. Data analyses and interpretation was performed by SA, JRM and TS. The article was first drafted by SA, and critically reviewed by all. JRM obtained funding for the IHC experiment. TS obtained funding for the study, and all authors takes full responsibility for the integrity of the work as a whole. We would also like to thank the following individuals for their contributions to this manuscript: Ms. M. Cecilia Alvarez-Veronesi, BSc, Dr. Umberto Banderali, Dr. Phillip B. Messersmith and Dr. Bridgett L. Steele. 70

89 CHAPTER 4: Cartilage Boundary Lubricating Ability of Proteoglycan 4 Multimer Enriched and Multimer Deficient Preparations with Hyaluronan 4.1 Abstract Objectives: The objective of this study was to evaluate the cartilage boundary lubricating ability of proteoglycan 4 (PRG4) preparations enriched (PRG4Multi+) and deficient (PRG4Multi ) in multimers when combined with hyaluronan (HA). Methods: PRG4 prepared from bovine cartilage explant culture was separated into two populations of interest using size exclusion chromatography (SEC); multimer enriched (PRG4Multi+) and multimer deficient (PRG4Multi ) preparations. Phosphate buffered saline (PBS) with 0.25% 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS) was used as the buffer solution for the SEC column and PRG4 solution. The PRG4Multi+ and PRG4Multi preparations were buffer exchanged from the SEC mobile phase into PBS. PRG4Multi+, PRG4Multi and unseparated PRG4 preparations were compared at a physiological concentration of 450 μg/ml with 1.5 MDa HA at 3.33 mg/ml in PBS. The cartilage boundary lubricating ability of these solutions were analysed using an in vitro cartilageon-cartilage boundary lubrication test setup. Results: The kinetic coefficient of friction of PRG4Multi+ with HA was similar to that of PRG4Multi with HA (p = 0.96), and both were significantly greater than unseparated PRG4 with HA (p < 0.05). Furthermore, PRG4 subjected to SEC then recombined did not significantly reduce the friction coefficient of HA alone (p = 0.996). 71

90 Conclusion: These results suggest that the SEC method employed, and potentially the CHAPS used in the mobile phase to generate the PRG4Multi+ and PRG4Multi preparations, altered the ability of PRG4 to interact with HA. Further research and/or method development is required to determine if PRG4 s functional interaction with HA in reducing friction is dependent upon PRG4 s disulfide-bonded multimeric structure. 72

91 4.2 Introduction Synovial fluid (SF) is a viscous ultrafiltrate of blood plasma with the presence of specific functional molecules including various glycosaminoglycans, glycoproteins, phospholipids and proteins[27]. A key role of SF is its lubrication of articular joints, and this is facilitated by boundary lubricating molecules including proteoglycan 4 (PRG4) and hyaluronan (HA)[13, 56, 67]. PRG4 is present in SF as well as at the surface of articular cartilage. It functions as a critical boundary lubricant necessary for joint health in a dose-dependent manner[13, 56], and also provides protection by preventing protein deposition and cell adhesion[69]. PRG4 is a mucinlike glycoprotein that is synthesised by cells at the articular cartilage, meniscus, synovial lining and tendons[54]. It is encoded for by the PRG4 gene[96], and is analogous to lubricin[91], superficial zone protein[52] and megakaryocyte stimulating factor[97]. HA is a glycosaminoglycan that aids cartilage boundary lubrication[13, 49] and is a key contributor to SF s viscoelastic characteristics[27]. It also contributes to tissue hydration, assembly and interaction with proteoglycan molecules, and interaction at a sub-cellular level with inflammatory receptors[27]. At a cartilage-on-cartilage boundary lubrication test setup, HA and PRG4 have been shown to work in a synergistic relationship to reduce friction close to that of whole SF[13]. Previously non-reduced PRG4 multimer enriched preparations (PRG4Multi+) have been shown to lubricate considerably better than a non-reduced PRG4 multimer deficient preparation (PRG4Multi ) using an in vitro cartilage-on-cartilage friction testing system[104]. PRG4Multi+ and PRG4Multi preparations were generated using size exclusion chromatography (SEC) with phosphate buffered saline (PBS) and 0.25% 3-[(3-cholamidopropyl) dimethylammonio]-1-73

92 propanesulfonate (CHAPS), a zwitterionic detergent, as the column buffer. However, it is unknown if inter-molecular disulfide bonds of PRG4 affects the cartilage boundary friction reducing synergism with HA. Therefore, the objective of this study was to evaluate the cartilage boundary lubricating ability of PRG4Multi+ and PRG4Multi with HA. 74

93 4.3 Methods Lubricant Preparation and Characterisation PRG4 Preparation and Characterisation PRG4 was prepared from fresh skeletally mature bovine stifle joints obtained from a local abattoir (Calgary, AB, Canada), as described previously[13]. In brief, bovine stifle joints were obtained and discs with the intact articular surface of the cartilage were harvested. The cartilage discs were cultured in Dulbecco s Modified Eagle s Medium with 0.01% bovine serum albumin for 28 days. Ascorbic acid (25 μg/ml) and recombinant human transforming growth factor-β (10 ng/ml) were added to the media[92]. The conditioned media was then purified using diethylaminoethyl anion exchange chromatography[53, 92]. The PRG4-rich M NaCl eluant was retained and concentrated with a 100 kda filter (EMD Millipore, Billerica, MA). The purity of the PRG4 solution was confirmed using 3-8% Tris-Acetate sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), followed by protein stain and western blotting with anti-prg4 anti-body (Ab) LPN[54, 98] (Life Technologies, Carlsbad, CA). The concentration of this PRG4 preparation was determined by bicinchoninic acid protein assay (BCA) (Sigma-Aldrich, St. Louis, MO)[49]. All PRG4 preparations were prepared in phosphate buffered saline (PBS) at a physiological concentration of 450 μg/ml[13, 93] PRG4 separation PRG4 was separated into PRG4Multi+ and PRG4Multi using SEC column (XK16/100 column; GE Healthcare Life Sciences, Baie d Urfe, QC, Canada) alongside a fast protein liquid chromatography (FPLC) apparatus (ÄKTA FPLC ; GE Healthcare Life Sciences), as described previously[104]. PBS and 0.25% CHAPS (Sigma-Aldrich, St. Louis, MO), a non-denaturing 75

94 zwitterionic buffer, was used as the mobile phase. Fractions of interest were filtered, concentrated and buffer exchanged using Amicon Ultra-15 Centrifugal 30 kda cut-off filters (EMD Millipore) using PBS, centrifuged at 37 o C, and re-suspended in HA solution HA Preparation HA at 1.5 MDa (Lifecore Biomedical, Chaska, MN) was also prepared in PBS at a physiological concentration of 3.33 mg/ml[13, 30] Cartilage Boundary Lubrication Tests Sample Preparation Fresh osteochondral samples were prepared for friction testing from the patellofemoral groove of skeletally mature bovine stifle joints obtained from a local abattoir (Calgary, AB, Canada), as described previously[13, 44]. Briefly, cores (radius = 6 mm) and annuli (R outer = 3.2 mm and R inner = 1.5 mm) were harvested[44]. Samples were then rinsed vigorously overnight in PBS at 4 C to rid the articular surface of residual SF prior to lubrication testing in PBS. Samples were then bathed in the subsequent test lubricants, completely immersing the cartilage, and left at 4 C overnight prior to the next day s lubrication test. The samples were again rinsed with PBS after each test before incubation in the next test lubricant Lubrication Test The Bose ELF 3200 was used to analyse the boundary lubrication ability of each of the PRG4 forms, using the previously described cartilage-on-cartilage friction test[13, 44]. Briefly, all samples were compressed to 18% of the total cartilage thickness. Samples were allowed to stress-relax for 40 minutes to enable fluid depressurisation of the interstitial fluid before initial 76

95 revolutions. Samples were then left in a pre-sliding duration (Tps) of 1200, 120, 12 and 1.2 seconds (s). Samples were rotated after each subsequent stationary period in the +/- 2 revolutions. The test sequence was then repeated in the opposite direction of rotation, -/+ 2 revolutions Lubricant Test Sequence Cartilage Boundary Lubricating Ability of PRG4Multi+ with HA and PRG4Multi with HA Compared to Unseparated PRG4 with HA. The effect of the inter-molecular disulfide bonds on PRG4 s functional synergism with HA to function as a cartilage-on-cartilage boundary lubricant was assessed with the following sequence. Test Sequence 1 (n = 8): PBS, PRG4Multi + HA, PRG4Multi+ + HA, PRG4 + HA, SF. Cartilage Boundary Lubricating Ability of Recombined-PRG4 and Unseparated PRG4 Preparations with HA. PRG4Multi+ did not enhance HA s cartilage boundary lubricating ability, therefore another test was performed to analyse if the process of SEC separation altered the ability of PRG4 to interact with HA synergistically at a cartilage-cartilage interface. PRG4 was subjected to SEC as above, the entire eluted solution was recombined, concentrated and buffer exchanged in PBS as described previously, henceforth referred to as Recombined-PRG4. Test Sequence 2 (n = 4): PBS, HA, Recombined-PRG4 + HA, PRG4 + HA, SF Statistical Analysis Two coefficients of friction; static (μ static, Neq ) and kinetic (<μ kinetic, Neq >) were calculated for each lubricant, as described previously[13]. Data is presented as mean ± SEM. Statistical analysis was implemented with Systat12 (Systat Software, Inc., Richmond, CA). For test sequence 1, values at Tps = 1.2s were within 13.0 ± 2.7% of those at Tps = 1200s and for test 77

96 sequence 2, values at Tps = 1.2s were within 28.5 ± 4.4% of those at Tps = 1200s. Therefore, ANOVA was used to assess the effect of lubricant and Tps as a repeated factor on μ static,neq and <μ kinetic,neq >, with Tukey post hoc testing on <μ kinetic,neq > at Tps = 1.2 s. 78

97 4.4 Results Cartilage Boundary Lubrication Tests Test Sequence 1. PRG4Multi+ and PRG4Multi with HA both functioned as effective friction-reducing cartilage boundary lubricants, however not to the same extent as unseparated PRG4 with HA. μ static,neq varied with Tps and test lubricant (both p < 0.001), with no interaction (p = 0.101). Values increased with Tps and were consistently highest in PBS and lowest in SF, with PRG4Multi+, PRG4Multi and unseparated PRG4 with HA being intermediate (Figure 4-1A). <μ kinetic,neq > values exhibited similar trends, varying with lubricant and Tps (both p < 0.001) increasing only slightly with Tps, with an interaction (p = 0.493). <μ kinetic,neq > values at Tps = 1.2s were greatest in PBS and lowest in SF. Values for PRG4Multi+ and PRG4Multi with HA were similar (p = 0.964), and both were significantly different than unseparated PRG4 with HA (p < 0.05). All PRG4 solutions were lower than PBS (p < 0.001), PRG4Multi+ and PRG4Multi with HA were higher than SF (p < 0.001) but unseparated PRG4 with HA was similar to SF (p = 0.533) (Figure 4-1B). 79

98 Figure 4-1: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients PRG4Multi+ + HA, PRG4Multi + HA and unseparated PRG4 + HA, with PRG4 preparations at 450 µg/ml and HA at 3.33 mg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p <

99 Test Sequence 2. All three test lubricants of interest functioned as effective friction reducing lubricants with no statistical differences between the solutions of interest. However, Recombined-PRG4 with HA appeared to lubricate in a similar manner to HA alone, and unseparated PRG4 with HA appeared to have an improved lubricating ability than these solutions and close to that of SF. μ static,neq varied with Tps and test lubricant (both p < 0.001), with an interaction (p = 0.067). Values increased with Tps and were consistently highest in PBS and lowest in SF, with HA, Recombined-PRG4 with HA and unseparated PRG4 with HA being intermediate (Figure 4-2A). <μ kinetic,neq > values exhibited similar trends, varying with lubricant and Tps (both p < 0.001) increasing only slightly with Tps, with an interaction (p < 0.001). <μ kinetic,neq > values at Tps = 1.2s were greatest in PBS and lowest in SF. Values for HA and Recombined-PRG4 with HA were similar (p = 0.996). HA was also similar to unseparated PRG4 with HA (p = 0.406) as was Recombined-PRG4 with HA (p = 0.613). All PRG4 solutions were significantly different than PBS (p < 0.01). All PRG4 with HA lubricant combinations were similar to SF (p > 0.154), with HA alone trending to being significantly higher (p = 0.082) (Figure 4-2B). 81

100 Figure 4-2: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients HA, Recombined-PRG4 with HA and unseparated PRG4 with HA, with PRG4 preparations at 450 µg/ml and HA preparations at 3.33 mg/ml. Letters that are the same signify no significant difference, differing letters signify significant difference, where p <

101 4.5 Discussion The objective of this study was to evaluate the cartilage boundary lubricating ability of PRG4Multi+ and PRG4Multi with HA compared to unseparated PRG4 with HA. PRG4Multi+ and PRG4Multi with HA lubricated in a similar manner and better than PBS, but neither were close to that of unseparated PRG4 with HA (Figure 4-1). Recombined-PRG4 with HA lubricated similarly to HA alone at a cartilage-cartilage interface (Figure 4-2). These results were surprising, as previous findings have shown that PRG4Multi+ lubricated significantly better than PRG4Multi at a cartilage-cartilage interface[104]. This led to the hypothesis that PRG4Multi+ with HA would lubricate significantly better than PRG4Multi with HA, and close to that of unseparated PRG4 with HA. These results suggest that the SEC process of separating PRG4 affected the PRG4-HA interaction. This could have been due to presence of residual CHAPS that affected the unfolding and refolding of the PRG4 molecule. It has been previously suggested the hydrophobic domains of PRG4 molecules enable interaction with HA[13, 97], and the presence of residual CHAPS may have affected this interaction. It is also possible that the SEC process may have removed or altered other link molecules, therefore disrupting the interaction of PRG4 and HA with these molecules and their ability to adsorb at an articular cartilage surface. Such link molecules could potentially be SAPLs[65-67] or other cartilage binding ligands[105] that have been previously shown to interact with PRG4 and the articular cartilage surface. The observation that Recombined-PRG4, which would contain any such putative link molecule(s) did not appear to enhance HA s cartilage boundary lubricating ability would suggest that if such a molecule exists, it was likely the residual CHAPS affecting its ability to enable PRG4+HA interaction. 83

102 In addition to having a synergistic cartilage friction reducing ability, PRG4 and HA have shown an interaction in solution. Previous studies have suggested a possible interaction between native PRG4 molecules (mixture of multimers and monomers) and HA, which was not observed with reduced and alkylated PRG4 (which are essentially unfolded monomers alone due to the breaking of inter- and intra-molecular disulfide bonds), through mechanisms such as entanglements[66, 106, 107]. This is supportive, at least in part, to PRG4 structure being important for its interaction with HA. However, additional development of separation and/or characterisation methodology is required for further functional examination of different PRG4Multi+ and PRG4Multi species with HA and any other potential binding molecules, at the surface of the articular cartilage and in solution. 84

103 4.6 Acknowledgements The experimental conception, design and data acquisition and anlysis was conducted by: Abubacker S, Haladuick J, Schmidt TA. This work was supported by funding from the Natural Sciences and Engineering Research Council of Canada, the The Arthritis Society & Canadian Arthritis Network (Grant: TGD ), and the Alberta Innovates Health Solutions Team in Osteoarthritis. 85

104 CHAPTER 5: Cartilage Boundary Lubricating Ability of Full-Length Recombinant Human Proteoglycan 4 Alone and in Combination with Hyaluronan 5.1 Abstract Objectives: Assess the (1) adsorption of full-length recombinant human proteoglycan 4 (rhprg4) to an articular cartilage surface, and (2) in vitro cartilage boundary lubricating properties of rhprg4 alone, and in combination with hyaluronan (HA). Methods: Immunohistochemistry was performed to visualise rhprg4 adsorption onto articular cartilage explants depleted of PRG4. The cartilage boundary lubricating ability of rhprg4 was first compared to bovine PRG4 (PRG4) using an in vitro cartilage-cartilage friction test under boundary lubricating conditions, and then assessed both alone and in combination with HA. Results: rhprg4 was able to adsorb to the articular cartilage surface, as well as the cut surface, of cartilage explants. The kinetic coefficient of friction of rhprg4 was similar to that of PRG4 (p = 0.674) and significantly lower than phosphate buffered saline (p < 0.001). The kinetic coefficient of friction of rhprg4+ha was significantly lower compared to rhprg4 alone (p = 0.003). Conclusions: These results demonstrate that rhprg4 is capable of adsorbing to an intact articular surface of cartilage as well as functioning as an effective friction lowering boundary lubricant, both alone and in combination with HA, in a manner similar to that of native PRG4. 86

105 5.2 Introduction Proteoglycan 4 (PRG4) is a mucin-like glycoprotein present in synovial fluid (SF) and at the surface of articular cartilage where it functions as a critical boundary lubricant[108]. PRG4 is synthesised and secreted by cells at the surface of articular cartilage, meniscus, synovial lining and tendons[54]. PRG4 is encoded for by the PRG4 gene[96], is homologous to lubricin[91] and superficial zone protein[52], and has several biophysical properties that contribute to joint health[66, 69]. These include functioning as a cartilage boundary lubricant, both alone in a dosedependent manner and in combination with hyaluronan (HA)[13, 49], in addition to inhibiting apoptosis of chondrocytes at articulating surfaces[70]. PRG4 contains a central mucin-like domain that is heavily glycosylated with O-linked oligosaccharides[108], which are necessary for PRG4 s cartilage boundary lubricating ability. Terminal globular protein domains, rich in cysteine, facilitate the intra- and inter-molecular disulfide-bonding of PRG4[54], which has been shown to be critical for its cartilage boundary lubricating ability[104], contributing to the lowfriction articulation of cartilage surfaces in vivo[70]. HA is a glycosaminoglycan present in native SF that has been used clinically as a viscosupplement in attempt to restore lubricating function. Although HA alone reduces friction at the cartilage-cartilage interface in vitro[13], the functional utility of intra-articular HA injections (as a lubricant) has been questioned in preclinical models of post-traumatic osteoarthritis (OA) (e.g., Teeple et al.[32]) and in patients suffering from knee OA (e.g. Juni et al.[31]). Nevertheless, cartilage friction is significantly reduced when HA is combined with PRG4 in vitro through a currently unknown mechanism[13, 49], and intra-articular injection of combined HA-PRG4 reduces cartilage damage in preclinical post-traumatic OA compared to HA alone[32]. 87

106 Although the precise cause of OA is unknown, one of the possible pathways leading to the development of OA amongst biomechanical mechanisms[72], involves the failure of joint lubrication[108]. Indeed, knee injury can result in decreased PRG4 concentration in SF, resulting in decreased cartilage boundary lubricating ability and contributing to the initiation of cartilage damage. Some patients with OA have diminished levels of PRG4, and their SF has reduced in vitro cartilage boundary lubricating ability compared to normal SF[75]. Decreased PRG4 concentrations have also been reported in SF of patients soon after an ACL injury[71], which is associated with an increased risk of subsequent progressive degeneration leading to OA[16, 72]. In patients with acute knee injuries or progressive chronic inflammatory arthritis, cartilage damage has been associated with decreased boundary lubricating ability of SF[109]. This association has also been observed in animal models, such as guinea pigs[110] and rabbits[109], following ACL transection which leads to the development of OA. This resulted in a reduction in PRG4 concentration, and in turn an increase in the coefficient of friction. Recent studies in rat models with ACL transections found an intra-articular injection of purified human PRG4 reduced cartilage degeneration[32, 111], improved weight bearing in the injured joint[111], and slowed cartilage damage due to the combination of ACL transection and exercise, whilst preserving the superficial zone chondrocytes viability[112]. This gives hope that improving articular boundary lubrication and preserving cartilage viability may slow or halt the initiation or progression of OA[70]. Additional studies have used recombinant versions of PRG4 to demonstrate its efficacy in preserving joint function and reducing cartilage degeneration using OA animal models. Flannery et al.[79] used a truncated version of recombinant human PRG4, LUB:1, where the mucin-like domain was approximately a third of the size of a full-length PRG4 molecule. LUB:1 was able to 88

107 reduce cartilage degradation and structural damage in a rat OA meniscus model[79], and subsequently was demonstrated to have a residency time within the joint of approximately 28 days following intra-articular administration in a rat model[79, 113]. Full-length recombinant human lubricin (rhlub) was also produce in small quantities, and demonstrated cartilage boundary lubricating ability in an in vitro cartilage-on-glass friction system[114]. Full-length rhlub also demonstrated an ability to bind to a depleted cartilage surface[100]. In another study, full-length recombinant human PRG4 (rhprg4) demonstrated the ability to effectively reduced cartilage damage after an ACL transection in a rat model[115]. However, it was unclear if this recombinant version of PRG4 was able to multimerise into the functionally important multimeric form of PRG4[54, 104], and more importantly was not able to be expressed at high levels necessary for future (clinical) evaluation[115]. Recently, full-length PRG4 has been successfully expressed at high (> 1g/L) levels[78] using Chinese Hamster Ovary (CHO) cells[116]. This rhprg4 demonstrates the appropriate higher order structure, O-linked glycosylations, and possesses a boundary lubricating ability similar to bovine PRG4 at the ocular surface[78]. While effective at the ocular surface, it remains unclear if rhprg4 is able to form an effective friction-reducing boundary lubricant at the articular cartilage surface. Hence, the specific objectives of this study were to assess the adsorption of rhprg4 to a native articular cartilage surface, and to evaluate the in vitro cartilage boundary lubricating properties of rhprg4, alone and in combination with HA. 89

108 5.3 Methods (rh)prg4 Preparations PRG4 PRG4 was prepared from fresh skeletally mature bovine stifle joint obtained from a local abattoir (Calgary, AB, Canada), as described previously[13]. Cartilage discs were cultured in Dulbecco s Modified Eagle s Medium (Life Technologies, Carlsbad, CA) with 0.01% bovine serum albumin, with the addition of 25 μg/ml of ascorbic acid and 10 ng/ml of recombinant human transforming growth factor-β to the medium[92]. Purification of the media was then performed using diethylaminoethanol anion exchange chromatography (GE Healthcare Life Sciences, Buckinghamshire, UK)[53, 92] using a gravity flow and a salt gradient, with batch elutions at M NaCl retained, filtered with a 100 kda filter and stored at -80 o C. Total concentration was determined by bicinchoninic acid assay (BCA) (Sigma-Aldrich, St. Louis, MO). The purity of the concentrated and filtered solution was confirmed using 3-8% Tris- Acetate SDS-PAGE (Life Technologies, Carlsbad, CA) followed by Simply Blue protein stain and densitometry analysis (Image J, Bethesda, MD)[54, 61, 111] rhprg4 rhprg4 was prepared as described previously[78]. Culture media conditioned by a CHO line transfected with the PRG4 gene was provided by Lµbris, LLC (Framingham, MA) in collaboration with Selexis SA (Geneva, Switzerland). Briefly, the gene encoding the full length 1404 amino acid human PRG4 was inserted into plasmid vectors commercially available at Selexis SA for enhanced gene expression in mammalian CHO cells. rhprg4 rich media was obtained from a shake flask culture with fed-batch cultivation (SFM4CHO medium (Hyclone, 90

109 Logan, UT)) supplemented with 8 mm L-Glutamine, hypoxanthine and thymidine (1x HT, Life Technologies, Carlsbad, CA) of a high expressing rhprg4 clonal cell line. Enriched rhprg4 preparations were prepared for immunohistochemistry (IHC) and friction testing, described below, by concentrating the rhprg4 rich media using a 100 kda MW cut off centrifugal filter. The purity of high MW rhprg4 species was assessed to be 50% by SDS-PAGE, protein stain and densitometry, as described above. The concentration of the enriched rhprg4 preparation was then determined by BCA and adjusted to take into account the level of purity. SDS-PAGE western blotting was also used to characterise the size distributions of immunoreactive species of non-reduced (NR) and reduced (R) (using NuPAGE Sample Reducing Agent, Life Technologies) rhprg4 samples, essentially as described previously, with anti-prg4 monoclonal Ab (mab) 9G3 (which recognises a combined sugar-peptide epitope in the mucin domain, unpublished data, G.D. Jay)[54, 60, 61, 111] Immunohistochemistry (IHC) Immunohistochemistry (IHC) Samples Cartilage discs with the intact articular cartilage surface (n = 15, diameter = 6 mm) were harvested from the bovine stifle joints of bovine knees obtained from a local abattoir. Three intact cartilage discs were embedded in media (Tissue Tek OCT, Sakura, Torrance, CA) and snap frozen in isopropanol cooled with liquid nitrogen, and served as positive control samples (labelled as fresh ) Specimen processing All other samples were placed in phosphate buffered saline (PBS) and shaken vigorously overnight at 4 C to rid the articular surface of residual PRG4, and subsequently frozen at -80 C 91

110 to prevent further production of PRG4 from viable chondrocytes. Samples were then thawed, and again shaken overnight at 4 C, before incubation with lubricants of interest. Samples were incubated over night at room temperature in a 48-well plate (VWR Scientific Products, PA); PBS (negative control), rhprg4 at a physiological concentration of 450 μg/ml, PRG4 at 450 μg/ml and bovine synovial fluid (SF) (Animal Technologies, Tyler, TX) (positive control), and left to incubate at room temperature for 24 hours[93]. Samples were then fixed in OCT and stored at - 80 C. Five-micron-thick sections were cut using a cryostat microtome (Microm HM550, Thermo Scientific, Waltham, MA) and placed on positively charged glass slides (Superfrost Plus Adhesion Slides, Thermo Scientific). Sections were then fixed in 4% paraformaldehyde in PBS and washed in PBS to remove OCT[93]. Samples were blocked with 10% hydrogen peroxide in methanol, followed by 10% goat serum with 1% BSA in PBS in a humidity chamber. Samples were then incubated overnight in anti-prg4 mab 9G3 in 1.5% normal goat serum at a ratio of 1:200[111]. Samples were then washed with PBS and incubated with secondary anti-body Alexa Fluor-594 rhodamine-conjugated goat-anti mouse IgG (Life Technologies, Carlsbad, CA) in 1.5% normal goat serum at a ratio of 1:100[94]. These were then washed with PBS, mounted with mounting medium containing the nuclear counterstain DAPI (Vectashield, Vector Laboratories, Inc., Burlingame, CA) and sealed with microscope cover slips (VWR Scientific Products, PA). Results were imaged using a Zeiss LSM 780 microscope (Carl Zeiss, Oberkochen, Germany) at a magnification of 20 objective (dry, 0.8 NA). Fluorescence images were obtained for both red (Alexa Fluor-594 rhodamine detected PRG4; excitation/emission of 590/617 nm) and blue (DAPI detected cell staining; excitation/emission of 358/461 nm) fluorescence. 92

111 5.3.3 Boundary Lubrication Tests Sample Preparation Fresh osteochondral samples (n = 11) were prepared for friction testing from the patellofemoral groove of skeletally mature bovine stifle joints, as described previously[13, 44]. Briefly, cores (radius = 6 mm) and annuli (R O = 3.2 mm and R i = 1.5 mm) were harvested from osteochondral blocks, both with central holes (radius = 0.5 mm) to enable fluid depressurisation[44]. Osteochondral samples were then rinsed vigorously overnight in ~40 ml of PBS at 4 C to rid the articular surface of residual SF (a procedure confirmed previously by lubrication testing[13, 44]) prior to lubrication testing in PBS. Samples were then frozen in PBS with proteinase inhibitors at -80 C as described previously[85]. Samples were re-shaken overnight in PBS to further deplete the surface of any residual PRG4 at the surface upon thawing (as described above) before friction testing. Samples were then bathed in ~0.3 ml of the subsequent test lubricants, completely immersing the cartilage, at 4 C overnight prior to the next day s lubrication test, and were again rinsed with PBS after each test before incubation in the next test lubricant Lubrication Test The Bose ELF 3200 was used to analyse the boundary lubrication ability of each of the PRG4 forms, using the previously described cartilage-on-cartilage friction test[13, 44]. Briefly, all samples were compressed to 18% of the total thickness at a constant rate of mm/s, and were allowed to stress-relax for 40 minutes to enable fluid depressurisation of the interstitial fluid. The samples were then rotated at an effective velocity of 0.3 mm/s (shown to maintain boundary mode lubrication at a depressurised cartilage-cartilage interface)[44] at ±2 revolutions. 93

112 Samples were then left in a pre-sliding duration (Tps) of 1200, 120, 12 and 1.2 seconds (s). Samples were then rotated after each subsequent stationary period, +/- 2 revolutions. The test sequence was then repeated in the opposite direction of rotation, -/+ 2 revolutions Lubrication Test Sequences Two test sequences were employed to assess the cartilage boundary lubricating ability of rhprg4, both alone and in combination with HA. In both test sequences, PBS served as the negative control lubricant and bovine SF served as a positive control lubricant. Both rhprg4 and PRG4 were prepared in PBS at a physiological concentration of 450 μg/ml[30]. HA at 1.5 MDa (Lifecore Biomedical, Chaska, MN) was prepared in PBS at a physiological concentration of 3.33 mg/ml[30]. Lubricants were tested in presumed increasing order of lubricating function. Test Sequence 1 (n = 7): rhprg4 vs PRG4. The lubrication test sequence was PBS, rhprg4, PRG4, SF. Test Sequence 2 (n = 4): rhprg4 vs rhprg4+ha. The lubrication test sequence was PBS, rhprg4, rhprg4+ha, SF Statistical Analysis The two coefficients of friction; static (μ static, Neq ) (resistance of start-up motion from static condition) and kinetic (<μ kinetic, Neq >) (resistance of steady sliding motion) were calculated for each lubricant as described previously[13]. <μ kinetic,neq > increased only slightly with Tps, with values at Tps = 1.2 s being on average within 23 ± 3% (mean ± SD) for test sequence 1 and 15 ± 6% (mean ± SD) for test sequence 2 compared to those at Tps = 1200s. Therefore, for brevity and clarity <μ kinetic,neq > data is presented at Tps = 1.2s only, as done previously[13]. Unless otherwise indicated, data is presented as mean with a 95% confidence interval (CI) (lower limit, 94

113 upper limit). ANOVA was used to assess the effect of lubricant and Tps as a repeated factor on μ static,neq and <μ kinetic,neq >, with Tukey post hoc testing on <μ kinetic,neq > at Tps = 1.2 s. Statistical analysis was implemented with Systat12 (Systat Software, Inc., Richmond, CA). 95

114 5.4 Results SDS-PAGE rhprg4 demonstrated high MW bands and immunoreactivity as assessed by protein stain and western blot. Protein staining (Figure 5-1A) of the NR rhprg4 preparation showed two prominent high MW bands with apparent MW of ~1 MDa and ~460 kda, and a single high MW ~460kDa band for the R rhprg4. Western blotting (Figure 5-1B) indicated high MW 9G3- immunoreactive bands in the NR and R rhprg4 preparations with similar apparent MW to those observed in the protein staining. 96

115 Figure 5-1: Protein stain (A) and Western Blot (B) probed with mab 9G3 of non-reduced rhprg4 (NR) to reduced rhprg4 (R) Immunohistochemistry IHC (Figure 5-2) of whole cartilage discs indicated a discrete 9G3-immunoreactive layer of PRG4 at the articular surface of fresh explants, the majority of which could be removed with vigorous shaking in PBS. This immunoreactivity was localised to the articular surface of fresh 97

116 explants, with none observed at the cut edges, and was specific as no immunoreactivity was observed in the non-immune serum control. Shaken articular discs incubated in PRG4 containing solutions showed 9G3 immunoreactivity (red) at the articular surface and cut edge, with the immunoreactivity of rhprg4 being similar to that observed for PRG4 as well as SF. The 9G3 immunoreactive layer at the articular surface appeared to have a greater intensity to that observed at the cut edge of the explant incubated in (rh)prg4 containing solutions (i.e. PRG4, rhprg4, and SF). Again, there was no immunoreactivity observed for the non-immune control. Immunoreactivity was also observed at the cartilage surface with cartilage discs incubated in PBS, though not to the same intensity as those incubated in PRG4 containing solutions. 98

117 Figure 5-2: Immunolocalisation of PRG4 at an articular cartilage surface. Fresh samples were taken directly from joint and snap-frozen (control). All other samples were shaken overnight in PBS at 4 o C, frozen at -80 o C, and shaken overnight again in PBS at 4 o C. Samples were then incubated in solutions of interest overnight at room temperature: PBS (negative control), SF (positive control), rhprg4 and PRG4 both at physiological concentrations of 450 µg/ml. Signal to anti-prg4 mab 9G3 is depicted by the red staining, chondrocyte cells are depicted by the blue DAPI staining Boundary Lubrication Tests rhprg4 vs. PRG4 rhprg4 functioned similarly to PRG4 as an effective friction-reducing cartilage boundary lubricant. μ static,neq varied with Tps and test lubricant (both p < ), with no interaction (p = 0.834). Values increased with Tps and were consistently highest in PBS and 99

118 lowest in SF. rhprg4 and PRG4 were intermediate (Figure 5-3A). <μ kinetic,neq > values exhibited similar trends, varying with lubricant and Tps (both p < ), with an interaction (p < ). <μ kinetic,neq > values at Tps=1.2s were greatest in PBS and lowest in SF. Values for rhprg4 (0.106 (0.093, 0.119)) and PRG4 (0.087 (0.068, 0.106)) were again intermediate and not significantly different (p = 0.674). PBS was significantly different than PRG4 (p < ) and rhprg4 (p = 0.001), and SF was also significantly different than PRG4 (p = 0.009) and rhprg4 (p = 0.001) (Figure 5-3B). 100

119 Figure 5-3: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients in PBS, rhprg4 and PRG4 both at 450 µg/ml, and SF. Different letters signify statistically significant differences (p < 0.05), n =

120 rhprg4 vs. rhprg4+ha rhprg4 functioned as a more effective friction-reducing cartilage boundary lubricant in the presence of HA. μ static,neq varied with Tps and test lubricant (both p < ), with an interaction (p < ). Values increased with Tps and were consistently highest in PBS and lowest in SF. rhprg4 and rhprg4+ha were intermediate (Figure 5-4A). <μ kinetic,neq > values exhibited similar trends, varying with lubricant and Tps (both p < ), with no interaction (p = 0.133). <μ kinetic,neq > values at Tps=1.2s were greatest in PBS and lowest in SF. Values for rhprg4 (0.104 (0.093, 0.116)) and rhprg4+ha (0.063 (0.050, 0.075)) were again intermediate but significantly different (p = 0.003). PBS was significantly different than rhprg4 (p < ) and rhprg4+ha (p < ), and SF was also significantly different than rhprg4 (p < ) and rhprg4+ha (p = 0.013) (Figure 5-4B). 102

121 Figure 5-4: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients in PBS, rhprg4, rhprg4 + hyaluronan (HA) (rhprg4 at 450 µg/ml and HA (1.5 MDa) at 3.33 mg/ml), and SF. Different letters signify statistically significant differences (p < 0.05), n =

122 5.5 Discussion The objective of this study was to analyse the ability of full-length rhprg4 to adsorb to the surface of articular cartilage and function as an effective friction-reducing boundary lubricant at the articular cartilage surface. The rhprg4 examined here demonstrated immunoreactivity, adsorption to the cartilage surface and cartilage boundary lubricating function equivalent to that of native PRG4. SDS-PAGE showed high MW immunoreactivity of rhprg4 with the anti-prg4 mab 9G3, similar to that of PRG4 suggesting the rhprg4 has functionally determinant O-linked glycosylations consistent with those of native PRG4. The NR rhprg4 also demonstrated a higher apparent MW species of ~1MDa in addition to ~460KDa, consistent with previous descriptions as a putative dimer and monomer respectively[61]. Lastly, the rhprg4 preparation used functioned synergistically with HA to further reduce friction at the cartilage-cartilage biointerface to levels approaching that of whole SF. The results of this study agree with and extend a previous study that used a similar enriched rhprg4 preparation, whose protein identity was confirmed by tandem mass spectrometry, generated through 100 kda cut off centrifugal filtration[78]. As discussed in that study, which demonstrated the appropriate higher order structure, glycosylation, and ocular surface boundary lubricating ability of the rhprg4, this preparation likely contained smaller non-prg4 proteins. However, these non-prg4 proteins did not inhibit the ability of the high MW rhprg4 to function as an effective boundary lubricant. The present study extended the characterisation to the high MW immunoreactivity to mab 9G3, presenting as a broad distribution typical of glycosylation-dependent epitopes[54], and demonstrated that rhprg4 was indeed able to bind to the articular surface and function as an effective cartilage boundary lubricant. This suggests any impurities present did not significantly inhibit function in these in 104

123 vitro tests. As such, this study, along with the previous study, provides the basis for future in vitro, in vivo, and clinical evaluation of highly purified rhprg4 as a cartilage lubricant and potential biotherapeutic. rhprg4 was able to bind to the surface of articular cartilage depleted of PRG4, a property required for function as a boundary lubricant. The method employed to remove the PRG4 from the articular surface, vigorous shaking in PBS, was chosen over other enzymatic, ionic, or mechanical methods employed as to avoid any potential alterations to the articular cartilage surface[93, 94, 100]. Some residual PRG4 was visualised on the articular surface by IHC with fluorescence detection, though with substantially less intensity compared to fresh or repleted samples. The effective removal of the PRG4 at the surface by this method is further supported by the high friction coefficient values measured for these samples tested in PBS, compared to those of fresh samples tested in PBS[13, 44]. Indeed, simple incubation in PBS without agitation has been shown previously by immunohistochemical staining to not significantly remove PRG4 from the articular surface[93], suggesting the protein is firmly bound to the articular surface. Enzymatic methods have been used to remove PRG4 from the surface, though it is unclear what potential effects this has on the functional articular surface[117]. A high salt extraction has been used to remove PRG4 from the articular surface, which was able to be repleted by incubation with the aforementioned rhlub[100]. In that study, in contrast to the present study, adsorption was specific to the articular surface, with no observed PRG4, via immunohistochemical staining, on the cut edges. However, this difference could be due to cartilage samples being formalin-fixed, embedded in paraffin, sectioned, and mounted on a glass slide prior to incubation[100], compared to the incubation of the whole explants done in the present study. While the adsorption observed in this study to the non-articular surfaces was 105

124 perhaps somewhat unexpected, it was observed in all PRG4 containing solutions, including SF, and was not visualised in the non-immune control. Future studies could consider incubation of osteochondral samples instead of cut cartilage explants to prevent adsorption to lower cut surfaces. The IHC data presented here demonstrates the ability of rhprg4 to not only adsorb to the surface of articular cartilage, as assessed by 9G3 immunoreactivity, but also interestingly to cut surfaces in a manner similar to that of native PRG4 and whole SF. This adsorptive/adhesive property is valuable for a biotherapeutic aimed at improving surface lubrication. This study also agrees with and extends previous studies examining the lubricating ability of recombinant PRG4. The studies by Jones et al.[100] and Gleghorn et al.[114] demonstrated rhlub bound to articular surfaces[100], and reduced friction in a boundary mode at a cartilageglass interface[114]. Consistent with these previous studies, the present study demonstrated fulllength recombinant human PRG4 was also able to bind to the surface of articular cartilage and reduce friction in a boundary mode at a cartilage-cartilage biointerface. Furthermore, rhprg4 s ability to synergistically function with HA and further reduce friction, like that of native PRG4 through a currently unknown mechanism[49], was also demonstrated. This functional interaction was also demonstrated using a preliminary purification process involving previously described anion exchange chromatography[118] (data not shown). While this purification scheme resulted in unacceptable losses of protein, it provides the framework for further development and demonstrates enriched preparations of rhprg4 is able to function synergistically with HA as a cartilage boundary lubricant. Collectively, these results extend upon the initial characterisation of the rhprg4 and reported ocular surface boundary lubricating ability, as well as potential in vitro studies, and demonstrate rhprg4 s ability to function as a boundary lubricant at various biological tissues[78]. 106

125 Given the abundance of data on PRG4 s efficacy in preventing cartilage degradation in animal models of post traumatic OA[32, 79, , 115], the potential use of this recently available rhprg4 to preserve joint function is of great promise. Future in vivo large animal studies will be paramount in evaluating the efficacy of this newly available rhprg4, appropriately purified for intra-articular injection. Such studies combined with continued efforts towards industrial production, scale up, and characterisation will provide the foundation and motivation for clinical evaluation of rhprg4 as a biotherapeutic treatment to prevent or slow the progression of OA, therefore improve the quality of life of those that suffer from the disease. 107

126 5.6 Acknowledgements We would also like to thank the Umberto Banderalli for his help with the confocal microscopy. All authors contributed to the conception and design of the original study and approved the final submitted manuscript. SA, SD and DP were responsible for the data acquisition and analysis was conducted by SA, SD, JM and TS. The article was first drafted by SA, and critically reviewed by SD, DP, JM, GDJ and TS. TS obtained funding for the study, and all authors takes full responsibility for the integrity of the work as a whole. This work was supported by funding from the Natural Sciences and Engineering Research Council of Canada (355591), The Arthritis Society, and the Alberta Innovates Health Solutions Team in Osteoarthritis. TAS and GDJ have a financial interest in, and are named inventors on issued patents held by a commercial entity (Lubris) developing rhprg4 for therapeutic uses. TAS is also a paid consultant for the same entity. 108

127 CHAPTER 6: Conclusions 6.1 Summary of Findings The overall goals of this thesis work were to first determine the effects of different counterfaces (cartilage-cartilage versus cartilage-glass) on the lubricating ability of PRG4 and HA. Secondly, to assess the ability of disulfide-bonded PRG4 multimers and PRG4 monomers to localise at an articular cartilage surface and function as a boundary lubricant, with and without HA. Lastly, to evaluate the newly available full-length rhprg4 compared to native PRG4 in terms of its cartilage adsorption and boundary lubricating ability. The major findings were: 1. The lubricating ability of putative cartilage boundary lubricants can be affected by the counterface in in vitro tests. At a cartilage-glass interface, HA demonstrated no friction reducing ability compared to PBS, and PRG4 appeared just as effective as SF. Conversely, at a physiological cartilage-cartilage interface, HA and PRG4 both effectively reduced friction compared to PBS, with HA demonstrating similar friction reducing ability to that of whole SF at higher velocities. 2. Inter-molecular disulfide-bonded multimeric structure of PRG4 is important for PRG4 s ability to adsorb to the articular cartilage surface and function as a boundary lubricant. PRG4 multimer enriched (PRG4Multi+) preparations reduced friction in a dose-dependent manner, and appeared to adsorb to the articular surface to a greater extent than multimer deficient preparations (PRG4Multi ). The PRG4Multi preparations still reduced friction and adsorbed to the articular surface, though not to the same extent as the PRG4Multi+ preparations. 109

128 3. PRG4Multi+ and PRG4Multi preparations used in this study did not enhance HA s cartilage boundary lubricating ability, as has been observed with native unseparated PRG4. This suggests the separation method used here altered PRG4 s ability to interact with HA to synergistically reduce friction. 4. The rhprg4 examined in this study demonstrated immunoreactivity, adsorption to the articular cartilage surface and cartilage boundary lubricating function, with and without HA, equivalent to that of native PRG4. This thesis work has highlighted the importance of using physiologically relevant tissues when analysing the lubricating ability of PRG4 and HA in vitro, and suggests being cautious when comparing data between other methodological approaches that use synthetic surfaces. With respect to the PRG4 structure-function relationship, these findings contribute to a greater understanding of the molecular basis of articular cartilage boundary lubrication of PRG4. Elucidating the PRG4 structure-lubrication function relationship in SF will further contribute to the understanding of PRG4 s role in diarthrodial joint homeostasis and disease. Finally, given the abundance of data on PRG4 s efficacy in preventing cartilage degradation in animal models of post traumatic OA, the potential use of rhprg4 to preserve joint function is of great promise. Such studies combined with continued efforts towards rhprg4 industrial production, scale up, and characterisation will provide the foundation and motivation for clinical evaluation of rhprg4 as a biotherapeutic treatment to prevent or slow the progression of OA. 110

129 6.2 Limitations This study utilised a cartilage-cartilage in vitro rotational test setup, which varies to other in vitro test setups such as a linear reciprocating geometry[47] or other rotational setups that use non-physiological interfaces[45, 46]. The in vitro setup employed here does not completely represent the whole joint ex vivo test setup, where many modes of lubrication are operative[43], however, it does allow for the friction measurement of lubricants of interest in a boundary mode at a physiologically relevant interface. The PRG4Multi+ and PRG4Multi preparations used in this study were generated using SEC. Chromatogram results indicated pooled peaks were not distinctly separated. As such, pooling of the chosen fractions could have resulted in cross-contamination in each of the PRG4Multi+ and PRG4Multi preparations. SDS-PAGE results indicated enriched preparations of multimers (PRG4Multi+) and monomers (PRG4Multi ) with some overlap between pooled samples. Subsequent evaluation of more select and fewer fractions, selected closer to the peak in each pool in an attempt to avoid cross-contamination, showed no apparent improvements but it did result in significant sample losses. Additional separation and characterisation method development would allow for further functional examination of different putative monomeric PRG4 species[61], as well as higher order (e.g. dimers, multimers) PRG4. Nevertheless, the methods developed and employed here resulted in the generation of enriched populations of PRG4 multimers/monomers that were appropriate for this study. The use of IHC provided a qualitative insight to the localisation of various PRG4- containing preparations at the cartilage surface. However, future work is required to quantify the amount of PRG4Multi+, PRG4Multi and rhprg4 at the articular cartilage surface. One 111

130 potential approach, previously employed, is the radiolabeling of PRG4[102] and quantifying the amount of PRG4 accumulated at the cartilage surfaces. The SEC process of separating PRG4 appeared to affect the PRG4-HA interaction. This could have been due to presence of residual CHAPS post SEC and buffer exchange that affected the unfolding and refolding of the PRG4 molecule. The observation that Recombined-PRG4 did not appear to enhance HA s cartilage boundary lubricating ability would suggest that residual CHAPS affecting its ability to enable PRG4+HA interaction. Future work would require extensive buffer exchange of the PRG4 preparations, such as exhaustive dialysis. However, given the sample loses observed with dialysis, for this preliminary study assessing the importance of PRG4 structure, the current methodologies were appropriate. The rhprg4 preparation used in this study likely contained smaller non-prg4 proteins. Further purification would be required to remove these contaminants, similar to the aforementioned dialysis, however the losses due to these purification methodologies would need to be monitored. These non-prg4 proteins did not appear to inhibit the ability of the high MW rhprg4 to function as an effective boundary lubricant or adsorb to the cartilage surface. This suggests any impurities present did not significantly inhibit function, as assessed in these in vitro tests. As such, this study, along with the previous study, provides the basis for future in vitro, in vivo, and potentially clinical evaluation of highly purified rhprg4 as a cartilage lubricant and potential biotherapeutic. 6.3 Discussion and Future Work PRG4 structure in disease progression and personalised treatment The exact cause of OA is unknown but can be associated to multiple factors including mechanical and chemical elements[15-17]. Joint lubrication, specifically that provided by PRG4, 112

131 is one key aspect of maintaining joint health and integrity[56, 79]. However, PRG4 levels alone may not always be representative of PRG4 s friction reducing ability at the articular cartilage surface. SF from primary and secondary OA human patients and animal models have varied in concentration and lubricating ability, but not always in a manner directly proportional to PRG4 content[56, 71, 73-75]. Indeed, PRG4 concentration has been shown to increase in SF of ACL deficient ovine models soon after surgery (2-4 weeks) before returning to normal (20 weeks)[73, 74]. The lubricating ability of this ovine SF was diminished during these early stages of injury compared to normal SF, and by 20 weeks the SF returned to normal lubricating ability[73, 74]. These results were similar to those reported in an equine model, where acute injury SF had the highest levels of PRG4 followed by chronic injury SF and normal SF. The kinetic coefficient of friction of SF from equine models was also similar in trends to that of the ovine model, where friction values of the acute injury SF being 39% higher than normal SF and values of chronic injury SF being 20% higher than normal SF[85]. The variability in SF lubricating ability and PRG4 levels is also observed in humans. Studies have shown PRG4 concentrations drop after ACL injury and recover after one year[71]. A similar reduction of PRG4 is seen with synovitis patients resulting in diminished lubricating ability[56]. Patients suffering from late-stage OA have shown elevated[76], normal and diminished concentrations of PRG4[75]. Those OA SF samples with elevated levels of PRG4 showed equivalent function compared to normal SF[76]. Conversely, OA SF with reduced levels of PRG4 resulted in diminished lubricating ability[75]. These results collectively suggest that the PRG4 content is not necessary and sufficient on its own for SF s cartilage boundary lubricating ability. 113

132 PRG4 structure, specifically inter-molecular disulfide bonds, may contribute to diarthrodial joint homeostasis and therefore alterations could contribute to disease initiation and/or progression. A lack of normal PRG4 multimer/monomer distribution in SF could be an explanation for those patients who have normal or elevated concentrations of PRG4 but lack normal lubricating ability. This potential alteration in normal distribution of PRG4 multimer/monomer, resulting from a joint injury or in chronic OA, could be due to the joint attempting to compensate with an excessive production of PRG4 in an effort to protect the articular cartilage[76]. This elevated synthesis rate of PRG4 over a shorter period of time in a response to a change in the joint environment, could result in a bypass of the high energy process in which monomers are organised into a higher level of structural complexity[119]. Hence the post-translational higher order multimerisation process at the Golgi apparatus[57, 59] may not be able to be maintained during this period of PRG4 up-regulation. Therefore, although the concentration of PRG4 may be normal or even high in post injury or chronic disease states, there could be higher levels of PRG4 monomers. PRG4 monomers are able to lubricate to a certain extent but not to the same levels as PRG4 multimers. This could effectively result in an overall diminished lubricating ability of the SF. Further method development is required to facilitate the analysis of human SF (normal, joint injury and OA) in terms of PRG4 monomer/multimer composition, to see if there are differences in PRG4 structural composition between patients. Understanding the PRG4 structurelubrication function relationship in SF will provide further insight to PRG4 s role in fundamental joint lubrication and alteration in disease. In addition, understanding deficiencies in PRG4 structural composition on a patient specific basis could provide an opportunity for personalised 114

133 treatment with the rhprg4 multimers and/or monomers for both primary and secondary OA patients to restore normal PRG4 multimer/monomer distribution PRG4 and HA relationship PRG4 and HA have been shown to interact in solution, possibly through entanglement or crowding[66, 106, 107], and at an articular cartilage interface to reduce friction close to that of whole SF[13, 49]. In addition, PRG4 and HA s rheological characteristics, has been shown to be dependent in part by PRG4 s multimeric structure[106]. PRG4 solutions demonstrate shearthinning properties at a physiological concentration (450 μg/ml) alone; however, when altered through removal of disulfide bonds via reduction and alkylation, PRG4 solutions demonstrate Newtonian characteristics[106]. The addition of HA to even a low concentration of PRG4 (45 μg/ml) showed an enhancement to its viscosity, however when the disulfide bonds were removed the viscosity of HA remained the same[106]. Collectively, this suggests that the higher level structure of PRG4 may be important for the PRG4 and HA interaction. The results from the PRG4Multi+/PRG4Multi with HA study did not concur with the hypothesis that the disulfide-bonded multimeric structure of PRG4 is necessary for normal cartilage boundary lubricating ability with HA. However, the size exclusion chromatography (SEC) process employed to separate the PRG4 appeared to be a factor in affecting PRG4 s ability to interact with HA at a cartilage surface, although the exact cause of this alteration is unknown. It is possible the separation process may have resulted in incorrect refolding of PRG4 after separation with CHAPS, thus preventing functional interaction with HA through entanglement or crowding. Indeed, the presence of pool 1 (Figure 3-1), confirmed to contain PRG4 through MS/MS (data not shown) would suggest higher order structuring beyond 115

134 dimerisation, possibly a micelle or aggregate of PRG4 molecules. Since mucins contain both hydrophilic and hydrophobic domains, they are able to adhere strongly to a wide range of surfaces, as well as self-aggregate[34]. The ability to self-aggregate was further supported when reduced and alkylated PRG4 was subjected to the SEC. This resulted in no pool 2 (putative dimer) and an abundance of pool 3 (putative monomer), as expected, but unexpectedly pool 1 was still present and enhanced compared to the non-reduced PRG4. It is possible that this increase in concentration in pool 1 was due to exposure of hydrophobic domains from reduction/alkylation, resulting in excessive aggregation. It is unclear if such a species would naturally be present in SF, and if it would be a function of overall concentration of PRG4, but it is unlikely it has the ability to adsorb or lubricate at the articular cartilage[63, 100]. Despite these observations, the fact that PRG4Multi+ and PRG4Multi preparations exhibited an ability to function as a friction reducing lubricants suggests the PRG4 was refolded appropriately post SEC. Another possible explanation is the SEC separation process irreversibly altered a putative link molecule, such as SAPLs or other cartilage binding ligands[65-67, 105], that may be required for PRG4 and HA to interact at a cartilage surface[120]. As aforementioned, it was observed that Recombined-PRG4, which would contain any such putative link molecule(s), did not appear to enhance HA s cartilage boundary lubricating ability. This would suggest that if such molecule(s) existed, it is likely the residual CHAPS affected the link molecule(s) s ability to enable PRG4+HA interaction. The use of CHAPS in the SEC was chosen as it is a mild zwitterionic detergent that preserves the biological structure and activity of the molecule[121]. However, further improvement on this separation methodology is required to enable functional analysis in the 116

135 presence of HA. Indeed, one possible, perhaps likely, explanation for the altered ability of PRG4 multimers to interact with HA is the presence of residual CHAPS post buffer exchange. Even at small levels, CHAPS bound to the PRG4 could interfere with its ability to interact with the hydrophobic domains of HA; which has been postulated as a mechanism of interaction of PRG4 and HA[13, 97]. One possible solution to this is to employ exhaustive dialysis to completely buffer exchange after SEC and remove CHAPS, rather than just the use of the 100 kda cut-off filters alone. This approach was not feasible in these studies due to the limited amount of PRG4 multimers and monomers, and the significant cost/time required to generate such preparations due to typical losses associated with SEC that would be exacerbated with dialysis for buffer exchange. Although, dialysis could be employed in the future given the abundance of rhprg4 that has become available PRG4 localisation at the articular cartilage surface IHC suggested PRG4Multi+ had an increased localisation at the articular cartilage surface compared to PRG4Multi. However, these IHC results are qualitative and further work is required to provide quantitative outcomes. Previously 125 Iodine-labelled PRG4 has been used to assess localisation at the articular cartilage surface[102]. This method could be utilised to provide quantification of PRG4Multi+ and PRG4Multi at the articular cartilage surface. Preliminary results have shown that PRG4Multi+ showed ~20% increased localisation at the cartilage surface compared to PRG4Multi (data not shown). In addition to this, when compared to a fibrinogen monolayer the increased accumulation of the PRG4 preparations at the articular cartilage suggests there is multilayer adsorption occurring at the surface of the articular cartilage. It is unknown if a thicker layer of PRG4 necessarily would result in improved lubrication. 117

136 Though, if PRG4 is functioning through a sacrificial layer mechanism[41, 122] of boundary lubrication, thickness and the amount at the surface could be a factor. Indeed, the notion of a viscous boundary layer at the surface of cartilage has been suggested by others[123]. The appropriate concentration of PRG4 in solution is an important factor for cartilage-oncartilage lubrication, as PRG4 functions in a dose-dependent manner until a saturation point[13, 83]. The results from this study demonstrate the structure of PRG4 plays a role in cartilage-oncartilage lubrication, as PRG4Multi+ acted in a dose-dependent manner while PRG4Multi did not (though both still functioned as effective lubricants compared to PBS). These results suggest that not only is the appropriate concentration of PRG4 required for effective boundary lubrication, but also the appropriate structure of PRG4 in terms of disulfide-bonded multimeric species. 118

137 Monomers Multimers Figure 6-1: schematic illustration of possible localisation of PRG4 monomers and multimers at an articular cartilage surface. C-terminal (blue) and N-terminal (red). Adapted from concepts by [34, 41, 58, 100, 122, 123]. A schematic illustration of possible PRG4 monomers and multimers localisation at the articular cartilage surface is presented in Figure 6-1 (adapted from concepts by[34, 41, 58, 100, 122, 123]). The cysteine rich C-terminal (blue) is thought to enable attachment to the articular cartilage surface and further disulfide-bridging interaction with larger polypeptide chains. The N- terminal (red) has been shown to enable dimerisation, entanglement and self-aggregation. As 119

138 such, it is possible that the PRG4 multimers and monomers adsorb to the cartilage surface until saturation (~450 μg/ml). However, when there is a higher concentration of PRG4 monomers, only a thin layer is provided compared to PRG4 multimers. Therefore, some lubricating ability is provided by PRG4 monomers, but not to the same extent as PRG4 multimers. This potential mechanism would be consistent with PRG4 disulfide-bonded multimers at the right concentration being required to provide normal cartilage boundary lubrication. Lastly, concurrent analysis of cartilage localisation of PRG4 and HA could also yield an understanding in a possible interaction of PRG4Multi+/PRG4Multi and HA at an articular cartilage surface. Preliminary IHC analysis have been conducted to this extent[124], but was limited by the observation that the HA binding protein used to visualise HA also interacted with PRG4[125], thus preventing an appropriate visualisation of the two lubricants at the cartilage surface. Future method development and/or alternative approaches would need to be explored to further simultaneously examine the localisation of PRG4 and HA at the articular cartilage surface. 120

139 References 1. Gray, H., Anatomy of the Human Body. Vol. Twentieth Edition. 1918, Philadelphia: Lea & Febiger. 2. Blackburn, T.A. and E. Craig, Knee Anatomy: A Brief Review. Physical Therapy, : p Ateshian, G.A. and V.C. Mow, Friction, lubrication, and wear of articular cartilage and diarthrodial joints, in Basic Orthopaedic Biomechanics and Mechano-Biology, V.C. Mow and R. Huiskes, Editors. 2005, Lippincott Williams & Wilkins: Philadelphia. p Mow, V.C., G.A. Ateshian, and R.L. Spilker, Bioemechanics of diarthodial joints: a review of twenty years of progress. J Biomech Eng, : p Mow, V.C., Wang, C. C., Hung, C. T., The extracellular matrix, interstitial fluid and ions as a mechanical signal transducer in articular cartilage. Osteoarthritis Cartilage, : p Buckwalter, J.A., H.J. Mankin, and A.J. Grodzinsky, Articular cartilage and osteoarthritis. Instructional Course Lectures, : p Fox, A.J.S., A. Bedi, and S.A. Rodeo, The Basic Science of Articular Cartilage: Structure, Composition, and Function. Sports Health, (6): p Freeman, M.A.R., ed. Adult Articular Cartilage. 2nd. ed. 1979, Pitman Medical: Tunbridge Wells, England. 9. McDevitt, C.A., Biochemistry of articular cartilage: Nature of proteoglycans and collagen of articular cartilage and their role in ageing and in osteoarthrosis. Annals of the Rheumatic Diseases, (4): p

140 10. Mow, V.C., et al., Biphasic creep and stress relaxation of articular cartilage in compression: theory and experiment. J Biomech Eng, : p Hussainova, I., and Ghaemi, H., Biobearings: where mechanics meets biology. Proceedings of the Estonian Academy of Seciences, (4): p Buckwalter, J.A., V.C. Mow, and A. Ratcliffe, Restoration of injured or degenerated articular cartilage. J Am Acad Orthop Surg, (4): p Schmidt, T.A., et al., Boundary lubrication of articular cartilage: role of synovial fluid constituents. Arthritis Rheum, (3): p Jebens, E.H. and M.E. Monk-Jones, On the viscosity and ph of synovial fluid and the ph of blood. The Journal of Bone and Joint Surgery, B(2): p Buckwalter, J.A., et al., Bone biology. Part II: formation, form, modeling, remodeling, and regulation of cell function. J Bone Joint Surg Am, A: p Gelber, A.C., et al., Joint injury in young adults and risk for subsequent knee and hip osteoarthritis. Ann Intern Med, (5): p Fleming, B.C., et al., Ligament Injury, Reconstruction and Osteoarthritis. Curr Opin Orthop, (5): p Buckwalter, J.A. and H.J. Mankin, Articular cartilage. Part II: degeneration and osteoarthrosis, repair, regeneration, and transplantation. J Bone Joint Surg Am, A: p Lawrence, R.C., et al., Estimates of the prevalence of arthritis and other rheumatic conditions in the United States. Part II. Arthritis Rheum, (1): p Waddell, D.D., Viscosupplementation with hyaluronans for osteoarthritis of the knee: clinical efficacy and economic implications. Drugs Aging, (8): p

141 21. Canada, S.C.-G.o. Musculoskeletal diseases - Osteoarthritis. 2010; Available from: Abusara, Z., Krawetz, R., et al., Muscular loading of joints triggers cellular secretion of PRG4 into the joint fluid. Journal of Biomechanics, (7): p Bosomworth, N.J., Exercise and knee osteoarthritis: benefit or hazard? Canadian Family Physician, (9): p Pelletier, J., Martel-Peletier, J., and Raynauld, J., Most recent developments in strategies to reduce the progression of structural changes in osteoarthritis: today and tomorrow. Arthritis Research & Therapy, (206): p Jin, Z.M., D. Dowson, and J. Fisher, Analysis of fluid film lubrication in artificial hip joint replacements with surfaces of high elastic modulus. Proc Inst Mech Eng [H], : p Pelttari, K., Wixmerten, A., and Martin, I., Do we really need cartilage tissue engineering? Swiss Medical Weekly, (41-42): p Necas, J., et al., Hyaluronic acid (hyaluronan): a review. Veterinarni Medicina, (8): p Balazs, E.A. and J.L. Denlinger, Viscosupplementation: a new concept in the treatment of OA. J Rheumatol, S39: p Balazs, E.A., Therapeutic use of hyaluronan. Struct Chem, : p Watterson, J.R. and J.M. Esdaile, Viscosupplementation: therapeutic mechanisms and clinical potential in osteoarthritis of the knee. J Am Acad Orthop Surg, (5): p

142 31. Juni, P., et al., Efficacy and Safety of Intraarticular Hylan or Hyaluronic Acids for Osteoarthritis of the Knee. Arthritis Rheum, (11): p Teeple, E., et al., Effects of supplemental intra-articular lubricin and hyaluronic acid on the progression of posttraumatic arthritis in the anterior cruciate ligament-deficient rat knee. Am J Sports Med, (1): p Waller, K.A., Zhang, L. X., Fleming, B. C., and Jay, G. D., Preventing Friction-induced Chondrocyte Apoptosis: Comparison of Human Synovial Fluid and Hylan G-F 20. The Journal of Rheumatology, (7): p Coles, J.M., D.P. Chang, and S. Zauscher, Molecular mechanisms of aqueous boundary lubrication by mucinous glycoproteins. Current Opinion in Colloid & Interface Science, (6): p Pylios, T., and Shepherd, D. E. T., Prediction of lubrication regimes in wrist implants with spherical bearing surfaces. Journal of Biomechanics, (3): p Dowson, D., Bio-tribology. Faraday Discuss., : p Morrell, K.C., et al., Corroboration of in vivo cartilage pressures with implications for synovial joint tribology and osteoarthritis causation. Proc Natl Acad Sci U S A, (41): p Dunn, A.C. and W.G. Sawyer, Gemini Interfaces in Aqueous Lubrication with Hydrogels. Tribology Letters, (1): p Neu, C.P., K. Komvopoulos, and A.H. Reddi, The interface of functional biotribology and regenerative medicine in synovial joints. Tissue Eng Part B Rev, (3): p Wright, V. and D. Dowson, Lubrication and cartilage. J Anat, : p

143 41. Chan, S.M.T., et al., Tribological altruism: A sacrificial layer mechanism of synovial joint lubrication in articular cartilage. Journal of Biomechanics, (14): p Greene, G.W., et al., Adaptive mechanically controlled lubrication mechanism found in articular joints. Proc Natl Acad Sci U S A, (13): p Jay, G.D., et al., Association between friction and wear in diarthrodial joints lacking lubricin. Arthritis Rheum, (11): p Schmidt, T.A. and R.L. Sah, Effect of synovial fluid on boundary lubrication of articular cartilage. Osteoarthritis Cartilage, (1): p Davis, W.H.J., S.L. Lee, and L. Sokoloff, A proposed model of boundary lubrication by synovial fluid: Structuring of boundary water. J Biomech Eng, : p Jay, G.D., Characterization of a bovine synovial fluid lubricating factor. I. Chemical, surface activity and lubricating properties. Connect Tissue Res, : p Gleghorn, J.P., et al., Boundary mode frictional propertiesof engineered cartilaginous tissues. European Cells and Materials, : p Bonnevie, E.D., et al., Stribeck Analysis Of Synovial Lubricants: Lubricating Mechanisms And Interaction Of Hyaluronic Acid And Lubricin. Trans Orthop Res Soc, Poster No: Kwiecinski, J.J., et al., The effect of molecular weight on hyaluronan's cartilage boundary lubricating ability--alone and in combination with proteoglycan 4. Osteoarthritis Cartilage, (11): p

144 50. Benz, M., N. Chen, and J. Israelachvili, Lubrication and wear properties of grafted polyelectrolytes, hyaluronan and hylan, measured in the surface forces apparatus. J Biomed Mater Res A, A(1): p Singh, A., et al., Enhanced lubrication on tissue and biomaterial surfaces through peptide-mediated binding of hyaluronic acid. Nature Materials, : p Schumacher, B.L., et al., A novel proteoglycan synthesized and secreted by chondrocytes of the superficial zone of articular cartilage. Arch Biochem Biophys, (1): p Schmidt, T.A., et al., Synthesis of proteoglycan 4 by chondrocyte subpopulations in cartilage explants, monolayer cultures, and resurfaced cartilage cultures. Arthritis Rheum, (9): p Schmidt, T.A., A.H. Plaas, and J.D. Sandy, Disulfide-bonded multimers of proteoglycan 4 (PRG4) are present in normal synovial fluids. Biochim Biophys Acta, (5): p Schmidt, T.A., et al., Chemo-mechanical coupling in articular cartilage: IL-1a and TGFß1 regulate chondrocyte synthesis and secretion of proteoglycan 4, in Physical Regulation of Skeletal Repair, R.K. Aaron and M.E. Bolander, Editors. 2005, American Academy of Orthopaedic Surgeons: Chicago. p Jay, G.D., Lubricin and surfacing of articular joints. Curr Opin Orthop, (5): p Gum, J.R., Jr., Human mucin glycoproteins: varied structures predict diverse properties and specific functions. Biochem Soc Trans, (4): p

145 58. Rhee, D.K., et al., Consequences of disease-causing mutations on lubricin protein synthesis, secretion, and post-translational processing. J Biol Chem, (35): p Perez-Vilar, J. and R. Mabolo, Gel-forming mucins. Notions from in vitro studies. Histol Histopathol, (4): p Alvarez, M.C., J. Kooyman, and T.A. Schmidt, Synthesis of proteoglycan 4 (PRG4) disulfide-bonded multimers by chondrocytes in cartilage explants. Trans Orthop Res Soc, : p Steele, B.L., Alvarez-Veronesi, M. C., Schmidt, T. A., Molecular weight characterization of PRG4 proteins using multi-angle laser light scattering (MALLS). Osteoarthritis Cartilage, (3): p Thornton, D.J. and J.K. Sheehan, From mucins to mucus: toward a more coherent understanding of this essential barrier. Proc Am Thorac Soc, (1): p Schmidt, T.A., et al., PRG4 boundary lubrication of articular cartilage: role of disulfide bonding and chondroitin sulfate. Trans Orthop Res Soc, : p Rabinowitz, J.L., J.R. Gregg, and J.E. Nixon, Lipid composition of the tissues of human knee joints. II. Synovial fluid in trauma. Clin Orthop Rel Res, 1984(190): p Sarma, A.V., G.L. Powell, and M. LaBerge, Phospholipid composition of articular cartilage boundary lubricant. J Orthop Res, (4): p Jay, G.D., et al., The role of lubricin in the mechanical behavior of synovial fluid. Proc Natl Acad Sci U S A, (15): p Hills, B.A., Boundary lubrication in vivo. Proc Inst Mech Eng [H], (1): p

146 68. Marcelino, J., et al., CACP, encoding a secreted proteoglycan, is mutated in camptodactyly-arthropathy-coxa vara-pericarditis syndrome. Nat Genet, (3): p Rhee, D.K., et al., The secreted glycoprotein lubricin protects cartilage surfaces and inhibits synovial cell overgrowth. J Clin Invest, (3): p Waller, K.A., et al., Role of lubricin and boundary lubrication in the prevention of chondrocyte apoptosis. Proceedings of the National Academy of Sciences of the United States of America, (15): p Elsaid, K.A., et al., Decreased lubricin concentrations and markers of joint inflammation in the synovial fluid of patients with anterior cruciate ligament injury. Arthritis Rheum, (6): p Roos, E.M., Joint injury causes knee osteoarthritis in young adults. Current Opinion Rheum, (2): p Atarod, M.A., et al., Diminished Cartilage Lubrication Early After ACL Injury and Increased Meniscal Loads Later After an ACL Injury May Both Contribute To OA Development In an Ovine Model. Trans Orthop Res Soc, Barton, K.I., et al., Characterization of Proteoglycan 4 and Hyaluronan Composition and Lubrication Function of Ovine Synovial Fluid Following Knee Surgery. Journal of Orthopaedic Research, (10): p Ludwig, T.E., et al., Diminished cartilage-lubricating ability of human osteoarthritic synovial fluid deficient in proteoglycan 4: Restoration through proteoglycan 4 supplementation. Arthritis Rheum, (12): p

147 76. Temple-Wong, M.M., et al., Effect of Knee Osteoarthritis on the Boundary Lubricating Molecules and Function of Human Synovial Fluid. Trans Orthop Res Soc, Abstract No Morrison, S., et al., Dose-dependent & synergistic effects of proteoglycan 4 (PRG4) on boundary lubrication at a human cornea - polydimethylsiloxane (PDMS) biointerface. Eye Contact Lens, (1): p Samsom, M.L., et al., Characterization of Full-Length Recombinant Human Proteoglycan 4 as an Ocular Surface Boundary Lubricant. Experimental Eye Research, : p Flannery, C.R., et al., Prevention of cartilage degeneration in a rat model of osteoarthritis by intraarticular treatment with recombinant lubricin. Arthritis Rheum, (3): p Laurent, T.C., U.B. Laurent, and J.R.E. Fraser, Structure and function of hyaluronan: an overview. Immunol Cell Biol, : p. A Laurent, T.C. and J.R.E. Fraser, Hyaluronan. FASEB J, (7): p Fraser, J.R., T.C. Laurent, and U.B. Laurent, Hyaluronan: its nature, distribution, functions and turnover. J Intern Med, (1): p Ludwig, T.E., M.M. Hunter, and T.A. Schmidt, Effects of Concentration on Synergistic Hyaluronan-PRG4 Cartilage Boundary Lubrication. Trans Orthop Res Soc, Band, P.A., et al., Hyaluronan Molecular Weight Distribution is Associated with the Risk of Knee Osteoarthritis Progression. Osteoarthritis and Cartilage,

148 85. Antonacci, J.M., et al., Effects of Equine Joint Injury on Boundary Lubrication of Articular Cartilage by Synovial Fluid: Role of Hyaluronan. Arthritis Rheum, (9): p Brandt, K.D., G.N.J. Smith, and L.S. Simon, Intraarticular injection of hyaluronan as treatment for knee osteoarthritis: what is the evidence. Arthritis Rheum, (6): p Brown, T.J., U.B.G. Laurent, and J.R.E. Fraser, Turnover of Hyaluronan in Synovial Joints: Elimination of Labelled Hyaluronan from the Knee Joint of the Rabbit. Experimental Physiology, : p Smith, G.N., et al., Effect of intraarticular hyaluronan injection on synovial fluid hyaluronan in the early stage of canine post-traumatic osteoarthritis. The Journal of Rheumatology, (6): p AAOS, Treatment of Osteoarthritis of the Knee: Evidence-Based Guideline 2nd Edition. Journal of American Academy of Orthopaedic Surgeons, Blewis, M.E., et al., A Model of Synovial Fluid Lubricant Composition in Normal and Injured Joints. European Cells and Materials, : p Swann, D.A., H.S. Slayter, and F.H. Silver, The molecular structure of lubricating glycoprotein-i, the boundary lubricant for articular cartilage. J Biol Chem, : p Schmidt, T.A., et al., Differential regulation of proteoglycan 4 metabolism in cartilage by IL-1alpha, IGF-I, and TGF-beta1. Osteoarthritis Cartilage, (1): p Nugent-Derfus, G.E., et al., PRG4 exchange between the articular cartilage surface and synovial fluid. J Orthop Res, (10): p

149 94. Chawla, K., et al., Molecular resurfacing of cartilage with proteoglycan 4. Acta Biomater, (9): p Abubacker, S., Ham, H. O., et al., Cartilage boundary lubricating ability of aldehyde modified proteoglycan 4 (PRG4-CHO). Osteoarthritis Cartilage, (1): p Ikegawa, S., et al., Isolation, characterization and mapping of the mouse and human PRG4 (proteoglycan 4) genes. Cytogenet Cell Genet, (3-4): p Jay, G.D., et al., Homology of lubricin and superficial zone protein (SZP): products of megakaryocyte stimulating factor (MSF) gene expression by human synovial fibroblasts and articular chondrocytes localized to chromosome 1q25. J Orthop Res, : p Steele, B.L.S., T A, Molecular Weight Characterization of PRG4 Species using Multi- Angle Laser Light Scattering (MALLS). Trans Orthop Res Soc, Zappone, B., et al., Molecular aspects of boundary lubrication by human lubricin: effect of disulfide bonds and enzymatic digestion. Langmuir, (4): p Jones, A.R., et al., Binding and localization of recombinant lubricin to articular cartilage surfaces. J Orthop Res, (3): p Kooyman, J., M.C. Alvarez, and T.A. Schmidt, Cartilage boundary lubricating properties of native proteoglycan 4 purified from normal bovine synovial fluid. Trans Orthop Res Soc, : p Swann, D.A., et al., The lubricating activity of synovial fluid glycoproteins. Arthritis Rheum, : p

150 103. Abubacker, S., et al., Cartilage Boundary Lubricating Ability of Full-Length Recombinant Human Proteoglycan 4 Alone and in Combination with Hyaluronan. Osteoarthritis and Cartilage, In revision Abubacker, S., et al., Cartilage Boundary Lubricating Ability of PRG4 Monomers versus Multimers. Trans Orthop Res Soc, Flannery, C.R., et al., Proteomic identificaiton of novel lubricin-binding ligands at cartialge surfaces. Trans Orthop Res Soc, : p Ludwig, T.E.C., M. K.; Schmidt, T. A., Viscosity of PRG4: Effects of Concentration, Structure and Interaction with Hyaluronan. Trans Orthop Res Soc, Bloom, A.K., et al., Investigating the Effect of Proteoglycan 4 on Hyaluronan Solution Properties using Confocal Fluorescence Recovery after PhotoBleaching. Trans Orthop Res Soc, Jay, G.D. and K.A. Waller, The biology of Lubricin: Near frictionless joint motion. Matrix Biology, Elsaid, K.A., et al., Association of articular cartilage degradation and loss of boundarylubricating ability of synovial fluid following injury and inflammatory arthritis. Arthritis Rheum, (6): p Teeple, E., et al., Coefficients of friction, lubricin, and cartilage damage in the anterior cruciate ligament-deficient guinea pig knee. J Orthop Res, (2): p Jay, G.D., et al., Prevention of Cartilage Degeneration and Gait Asymmetry by Lubricin Tribosupplementation in the Rat Following Anterior Cruciate Ligament Transection. Arthritis Rheum, (4): p

151 112. Elsaid, K.A., et al., The impact of forced joint exercise on lubricin biosynthesis from articular cartilage following ACL transection and intra-articular lubricin's effect in exercised joints following ACL transection. Osteoarthritis Cartilage, (8): p Vugmeyster, Y., et al., Disposition of Human Recombinant Lubricin in Naive Rats and in a Rat Model of Post-traumatic Arthritis After Intra-articular or Intravenous Administration. American Association of Pharmaceutical Scientists, (1): p Gleghorn, J.P., et al., Boundary mode lubrication of articular cartilage by recombinant human lubricin. J Orthop Res, (6): p Jay, G.D., et al., Prevention of cartilage degeneration and restoration of chondroprotection by lubricin tribosupplementation in the rat following anterior cruciate ligament transection. Arthritis Rheum, (8): p Girod, P.A., et al., Genome-wide prediction of matrix attachment regions that increase gene expression in mammalian cells. Nature Methods, (9): p Nugent, G.E., et al., Dynamic shear stimulation of bovine cartilage biosynthesis of proteoglycan 4. Arthritis Rheum, (6): p Dorosz, S.G., et al., Cartilage Boundary Lubricating Ability of Full-Length Recombinant Human PRG4 Alone and In Combination with Hyaluronan. Trans Orthop Res Soc, D'Alessio, G., The evolutionary transition from monomeric to oligomeric proteins: tools, the environment, hypotheses. Progress in Biophysics & Molecular Biology, : p

152 120. Pawlak, Z., et al., The ultra-low friction of the articular surface is ph-dependent and is built on a hydrophobic underlay including a hypothesis on joint lubrication mechanism. Tribology International, : p AppliChem, Detergents Schmidt, T.A., et al., PRG4 contributes to a "sacrificial layer" mechanism of boundary lubrication of articular cartilage. Trans Orthop Res Soc, : p Yakubov, G.E., et al., Viscous Boundary Lubrication of Hydrophobic Surfaces by Mucin. Langmuir, : p Malone, R., S. Abubacker, and T.A. Schmidt, Immunolocalization of Proteoglycan 4 and Hyaluronan on Articular Cartilage. Journal of Undergraduate Research in Alberta, (1) Ludwig, T.E., Cartilage boundary lubrication and rheology of proteoglycan 4 + hyaluronan solutions and synovial fluid.,

153 APPENDIX A: Cartilage Boundary Lubricating Ability of Aldehyde Modified Proteoglycan 4 (PRG4-CHO) A1. Abstract & Introduction Proteoglycan 4 (PRG4), also known as lubricin[91] and superficial zone protein[52], is a mucinous glycoprotein that is present in synovial fluid (SF) and at the surface of articular cartilage, where it functions as a critical boundary lubricant necessary for joint health[56]. High friction and high wear are amongst many factors that may contribute to cartilage degeneration[109]. PRG4 is critical in reducing friction and minimising surface tissue shear[13] at the surface of cartilage, thus preventing the degradation of cartilage through boundary lubrication. Indeed, an increased coefficient of friction in PRG4-knockout mice is associated with increased wear of the articular surface[43]. Furthermore, intra-articular injection of PRG4 has been shown to prevent cartilage degradation in post-injury rat models of osteoarthritis (OA)[79]. A recent study, motivated by diminished PRG4 levels in SF associated with early OA, demonstrated aldehyde modification of PRG4 (PRG4-CHO) significantly enhanced its binding to a depleted articular surface[94]. Such modification may contribute to an improved biotherapeutic treatment of early OA with PRG4 through enhanced residence time within the joint and/or binding to the articular cartilage surface. However, it remains to be determined if PRG4-CHO maintains the friction-reducing ability of unmodified PRG4. Therefore, the objective of this study was to assess the cartilage boundary lubricating ability of PRG4-CHO versus unmodified PRG4. The findings of this study indicate that aldehyde modification does not significantly affect the protein structure or lubricating function of PRG4, and PRG4-CHO is an effective friction reducing cartilage boundary lubricant. These results, together with previously published 135

154 data[94], collectively suggest PRG4-CHO may be useful in molecular resurfacing strategies for tissue surfaces requiring lubrication, and potentially other biointerfaces or biomaterials as well. A2. Methods Lubricant Preparation and Characterisation PRG4. PRG4 was prepared from bovine cartilage, as described previously[13]. Briefly, cartilage discs with an intact articular surface were harvested from fresh skeletally mature bovine stifle joints (Calgary, AB, Canada) and were cultured for 28 days in the presence of TGF-β1[13]. PRG4 was purified from the conditioned media using DEAE anion exchange chromatography and centrifugal filteration[13]. Purity was confirmed using 3-8% Tris-Acetate Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) followed by protein stain or western blotting using Invitrogen s NuPAGE system (Carlsbad, CA, USA), described below. The concentration was then determined by bicinchoninic acid assay (BCA) (Thermo Fisher Scientific; Rockford, IL, USA). Aldehyde Modified PRG4 (PRG4-CHO). Aldehyde modified PRG4 (PRG4-CHO) was prepared as described previously[13, 94]. Briefly, PRG4-CHO was prepared using succinimidyl- 4-formylbenzamide (CHO) at a PRG4:CHO molar ratio of 1:100 in 100mM PO 4 buffer[94]. A treatment control (PRG4-SHAM) was exposed to modification buffers and incubations in the absence of CHO. Confirmation of PRG4 purity and the CHO reaction was assessed by SDS- PAGE, describe below, and followed by quantification by BCA. SDS-PAGE. SDS-PAGE analysis was used to characterise the size distributions and immunoreactivity of species in PRG4, PRG4-SHAM, and PRG4-CHO. Briefly, non-reduced (NR) and reduced (R) samples were subject to 3-8% SDS-PAGE followed by protein stain 136

155 and/or western blotting using anti-prg4 antibodies LPN and J108N for NR and R, respectively, as described previously[49, 54, 61]. Boundary Lubrication Tests Osteochondral Sample Preparation. Fresh osteochondral samples (n = 8) were prepared for friction testing from the patellofemoral groove of skeletally mature bovine stifle joints, as described previously[13, 44]. Samples were rinsed vigorously overnight in phosphate buffered saline (PBS) to rid the articular surface of residual SF[13, 44, 49] prior to lubrication testing in PBS. Samples were then bathed in the subsequent test lubricants, completely immersing the cartilage surface, at 4 C overnight prior to the next day s lubrication test. Samples were rinsed with PBS after each test before incubation in the next test lubricant. Lubrication Test. A Bose ELF 3200 was used to analyse the boundary lubrication ability of the PRG4 preparations, using a previously described in vitro cartilage-on-cartilage friction test[13, 44]. Briefly, samples were compressed at a constant rate of mm/s to 18% of the total cartilage thickness and were allowed 40 minutes to stress-relax to enable fluid depressurisation of the interstitial fluid. The samples were then rotated +2 then -2 revolutions at an effective velocity of 0.3 mm/s (shown to maintain boundary mode lubrication at a depressurised cartilage-cartilage interface[44]) with pre-sliding durations (Tps) of 1200, 120, 12 and 1.2 seconds (s). The test sequence was then repeated in the opposite direction of rotation. Experimental Design. To determine the effect of aldehyde modification on the cartilage boundary lubricating of PRG4, two test sequences were performed. In both experiments, n = 4 osteochondral samples were tested sequentially in five test lubricants: PBS (negative control), 137

156 three PRG4-containing test lubricants, and then bovine SF (positive control). Each PRG4- containing lubricant was prepared in PBS at a concentration of 450 μg/ml. Test Sequence 1 (n = 4): PBS, PRG4-CHO, PRG4-SHAM, PRG4, SF. Test Sequence 2 (n = 4): PBS, PRG4-SHAM, PRG4-CHO, PRG4, SF. Statistical Analysis Static (μ static, Neq ) and kinetic <μ kinetic, Neq > friction coefficients of frictions, representing the resistance to start-up motion and steady sliding motion respectively, were calculated as described previously[13, 44]. Unless otherwise indicated, data is presented as mean with a 95% confidence interval (CI) (lower limit, upper limit). ANOVA was used to assess the effect of lubricant and Tps, as a repeated factor, on μ static,neq and <μ kinetic,neq >, with Tukey post hoc testing on <μ kinetic,neq > at Tps = 1.2 s. Statistical analysis was implemented with Systat12 (Systat Software, Inc., Richmond, CA). A3. Results Lubricant Characterisation PRG4-SHAM and PRG4-CHO had similar MW and reactivity to anti-prg4 Abs LPN and J108N compared to unmodified PRG4. Specifically, protein staining of NR PRG4, PRG4- SHAM, and PRG4-CHO showed two high MW bands slightly above (**) and below (*) the 460kDa marker (Figure 7-1A). Upon reduction, a single high MW band (**) was observed as well as a predominant lower MW band ~90kDa (x) between the 71kDa and 117kDa markers (Figure 7-1B). Western blotting of NR samples showed two high MW (**, *) LPN-reactive bands (Figure 7-1C) that co-migrated with those observed by protein stain, while R samples 138

157 showed the high (**) and lower (x) J108N immunoreactive bands (Figure 7-1D) that comigrated with those observed by protein stain. Figure 7-1: Protein stain of non-reduced (NR) and reduced (R) (A, B), and western blotting of NR (C) and R (D) PRG4 samples. Boundary Lubrication Tests All PRG4 preparations functioned as effective friction-reducing cartilage boundary lubricants. The order of testing (i.e. test sequence 1 and 2) did not significantly affect the friction coefficient values obtained for PRG4-SHAM and PRG4-CHO; PRG4-SHAM <μ kinetic,neq > values at Tps = 1.2s from test sequence 1 vs. test sequence 2, p = 0.554, and PRG4-CHO <μ kinetic,neq > values at Tps = 1.2s from test sequence 1 vs. test sequence 2, p = 0.594, when compared via ANOVA. Therefore, data was pooled for further analysis. Lubricants and Tps modulated friction. μ static,neq varied with Tps and test lubricant (both p < ), with no interaction (p = 0.216). Values increased with Tps and were consistently highest in PBS (ranging from (0.223, 139

158 0.342) to (0.453, 0.663)) and lowest in SF (ranging from (0.026, 0.037) to (0.207, 0.320)). Values in PRG4, PRG4-SHAM, and PRG4-CHO were intermediate and similar to each other (Figure 7-2A). <μ kinetic,neq > values exhibited similar trends, varying with lubricant and Tps (both p < ), with no interaction (p = 0.469), increasing only slightly with Tps (values at Tps = 1.2s were within 17 ± 6% (mean ± SD) of those at Tps = 1200s). Therefore, for brevity and clarity, <μ kinetic,neq > data is presented at Tps = 1.2s only. <μ kinetic,neq > values at Tps = 1.2s were greatest in PBS (0.232 (0.180, 0.284)) and lowest in SF (0.025 (0.018, 0.032)). Values in PRG4 (0.101 (0.072, 0.130)), PRG4-SHAM (0.150 (0.127, 0.172)), and PRG4-CHO (0.143 (0.124, 0.162)) were again similar and significantly different than PBS (PBS vs. PRG4, p < , PBS vs. PRG4-SHAM, p = 0.005; PBS vs. PRG4-CHO, p = 0.002) and SF (SF vs. PRG4, p = 0.01; SF vs. PRG4-SHAM, p < ; SF vs. PRG4-CHO, p < ), but not statistically different from each other (PRG4 vs. PRG4-SHAM, p = 0.189; PRG4 vs. PRG4- CHO, p = 0.324; PRG4-SHAM vs. PRG4-CHO, p = 0.998, Figure 7-2B). 140

159 Figure 7-2: Static (μ static,neq ) (A) and kinetic <μ kinetic,neq > at Tps = 1.2 s (B) friction coefficients PRG4-SHAM, PRG4-CHO and PRG4 at 450 µg/ml. Sample size, n =

160 A4. Discussion The SDS-PAGE western blotting and protein stain (Figure 7-1) data along with the lubrication test data (Figure 7-2) indicate that CHO modification does not significantly affect the protein structure or lubricating function of PRG4. These results, together with previously published data demonstrating aldehyde modification significantly enhances PRG4 adsorption to the surface of cartilage[94], collectively suggest PRG4-CHO may be useful in molecular resurfacing strategies for tissue surfaces requiring lubrication. The size distribution of species in PRG4, PRG4-SHAM and PRG4-CHO appeared similar (Figure 7-1). All preparations contained high MW immunoreactive species, in addition to the previously observed ~90kDa species released upon reduction[54, 61]. Any potential effect of the aldehyde modification process on relative distribution of different PRG4 structural forms, such as monomers and multimers[54], remain to be fully elucidated. However, the maintenance of lubrication function in PRG4-CHO preparations suggest higher order structure was maintained, since reduction and alkylation of PRG4 preparations has been shown to increase friction by ~34%[63]. Values of <μ kinetic,neq > values at Tps = 1.2s for PRG4-CHO/SHAM appeared slightly increased compared to PRG4; however, these apparent differences were not statistically significant and the PRG4-CHO/SHAM preparations effectively reduced friction compared to PBS. The unmodified PRG4 used in this study was prepared from skeletally mature bovine cartilage (in Calgary, AB) whereas PRG4-CHO/SHAM was prepared from immature bovine cartilage (in Evanston, IL); a potentially confounding factor. However, unmodified immature PRG4 (obtained later) demonstrated equivalent cartilage boundary lubricating ability to the unmodified PRG4 used in the present study (<μ kinetic,neq > at 1.2s: PBS (0.108, 0.234), 142

161 mature bovine PRG (0.057, 0.076), calf PRG (0.059, 0.088), and SF (0.035, 0.039) for n = 4). As such, the immature and mature bovine PRG4 tested here possess similar lubrication function, and protein structure (Figure 7-1), and provide the rationale for their combined use of immature and mature PRG4. This result also suggests development may not affect PRG4 structure and function, although additional studies would be required to confirm this proposition, especially in humans. While the mechanism of molecular interactions of PRG4 with the cartilage surface remains to be elucidated, local administration of recombinant human PRG4 in OA induced rat models has been shown to have chondroprotective effects as well as reduce the rate of cartilage degeneration and structural damage[79]. In addition, recombinant PRG4 remained localised for up to 28 days following a single injection intra-articular administration[79]. Given the fact that aldehyde modification enhances adsorption of PRG4 to the cartilage surface, without altering lubrication function, future in vivo studies could assess whether PRG4-CHO has an enhanced ability to adhere and reside on joint tissues, and potentially provide prolonged cartilage boundary lubricating function. Additionally, PRG4 is known to act synergistically with hyaluronan (HA) as cartilage boundary lubricants to reduce friction[13, 49]. HA is a currently used as biotherapeutic treatment for OA, which can provide pain-relief for several months[31], even though the retention time of the HA solutions have been shown to be short[31]. While the mechanism of molecular interactions of PRG4 and HA, whether in solution or at a surface, also remains to be elucidated, future studies could also assess whether PRG4-CHO maintains this synergistic relationship with HA and potentially enhance the efficacy of such treatments. Lubricity is a desirable characteristic for many tissues, biointerfaces, and artificial surfaces. PRG4 is found on other tissue surfaces other than articular cartilage, such as the 143

162 pericardium, pleural membrane, and recently the cornea epithelium and conjunctiva[56, 77]. Biomedical devices such as contact lenses and artificial joint surfaces, as well as catheters and endoscopes, all require low friction and wear at their interacting biointerfaces where PRG4 may be benficial[94]. Thus, aldehyde modification could potentially be used to coat a variety of biomaterials or tissue surfaces with functional PRG4 and provide low-friction properties. Furthermore, the chemical modification strategies used here could be applied to other lubricating molecules, either alone or in combination with PRG4, for the biolubrication of other tissues, biointerfaces, or non-biological materials. 144

163 A5. Acknowledgements This chapter has been published in Osteoarthritis and Cartilage 2013 (21(1): ): Abubacker S, Ham HO, Messersmith PB, Schmidt TA. This work was supported by funding from the Natural Sciences and Engineering Research Council of Canada, the Canadian Arthritis Network, the Alberta Innovates Health Solutions Team Grant in Osteoarthritis, as well as the Faculty of Kinesiology and the Schulich School of Engineering's Centre for Bioengineering Research and Education at the University of Calgary. In addition to NIH grants R01 DE and R01 DE at Northwestern University. We would also like to thank the following individuals for their contributions to this manuscript: Taryn E. Ludwig, Sheila Morrison and Kristen Barton. Author Contributions All authors contributed to the conception and design of the original study and approved the final submitted manuscript. SA and HH were responsible for PRG4 sample preparations. SA was responsible of the data acquisition and analysis. The article was first drafted by SA, and critically reviewed by HH, PM and TS. TS & PM obtained funding for the study, and TS takes full responsibility for the integrity of the work as a whole. Conflict of Interest The authors have no potential conflicts of interest. 145

164 A6. Re-print permission for Appendix A, published in Osteoarthritis and Cartilage 2013; 21(1) p : or: Publ ication: isher: : Title Cartilage boundary lubricating ability of aldehyde modified proteoglycan 4 (PRG4- CHO) Auth S. Abubacker,H.O. Ham,P.B. Messersmith,T.A. Schmidt Publ Date Osteoarthritis Cartilage Elsevier Jan 1, 2013 Copyright 2013, Elsevier and Logged in as: Saleem Abubacker Account #: Order Completed Thank you very much for your order. This is a License Agreement between Saleem Abubacker ("You") and Elsevier ("Elsevier") The license consists of your order details, the terms and conditions provided by Elsevier, and the payment terms and conditions. number License Licensed content publisher Licensed content publication Licensed content title Licensed content author Licensed content date content number License date Sep 04, 2014 Licensed volume Licensed content issue number pages Number Type of Use Portion Format of Are you the Reference confirmation for license number Elsevier Osteoarthritis and Cartilage Cartilage boundary lubricating ability of aldehyde modified proteoglycan 4 (PRG4-CHO) S. Abubacker,H.O. Ham,P.B. Messersmith,T.A. Schmidt January reuse in a thesis/dissertation full article both print and electronic Yes 146