Investigations of Articular Cartilage Delamination Wear and a Novel Laser Treatment Strategy to

Increase Wear Resistance

Krista M. Durney

Submitted in partial fulfillment of the requirements for the degree of

Doctor of Philosophy in the Graduate School of Arts and Sciences

COLUMBIA UNIVERSITY 2018

© 2018

KRISTA M. DURNEY

ALL RIGHTS RESERVED

ABSTRACT

Investigations of Articular Cartilage Delamination Wear and a Novel Laser Treatment Strategy to Increase Wear Resistance

Krista M. Durney

There are limited treatment options available today to slow down progression of osteoarthritis in

its early stages and most interventions, such as highly invasive partial and total joint

replacement surgeries, are performed only at the late stages of the disease. Understanding the

mechanism of early articular cartilage stress-mediated wear and failure can aid in the design of

new treatment options that are introduced at earlier stages of the disease, presenting the

potential to slow down osteoarthritis progression and thus significantly improve patient

outcomes. This dissertation aims to provide a basic science understanding of wear propagation

and repair of articular cartilage in the absence of traumatic events under the normal reciprocal

sliding motion of the articular layers at physiologic load magnitudes. In this dissertation there

are three main thrusts: (1) characterize cartilage delamination wear under normal sliding (2)

define a chemical environment that promotes cartilage explant homeostasis to enable long-term

wear-and-repair studies (3) investigate a practical treatment modality capable of stopping or

slowing down structural degeneration of articular cartilage in OA.

We hypothesize that the mode of cartilage damage is delamination wear that progresses by

fatigue failure of the extracellular matrix (ECM) under physiologic sliding, even when cartilage

layers are subject to physiologic load magnitudes and contact stresses and even when the

friction coefficient µ remains low (H1a). Based on prior literature findings regarding the role of

synovial fluid (SF) boundary lubricants on the reduction of friction and wear, we also test the

hypothesis (H1b) that SF delays the onset of cartilage delamination when compared to

physiological buffered saline (PBS). We then test a third hypothesis (H1c) that loading cartilage

against cartilage delays the onset of delamination wear compared to testing glass on cartilage,

since contacting porous cartilage layers exhibit a much smaller solid-on-solid contact area

fraction than impermeable glass contacting porous cartilage.

Next, we hypothesize that the homeostatic dysregulation previously observed in cultured

immature cartilage explants results from the presence of non-physiologic levels of important

metabolic mediators in the culture medium. To this end, we hypothesize that: (H2a) immature

bovine cartilage explants cultured in native synovial fluid will maintain homeostasis as

characterized by maintenance of their mechanical properties and ECM contents at initial (post-

explantation) levels, and (H2b) explants cultured in a physiologic-based medium, consisting of

physiologic levels of key metabolic mediators, will maintain a similar homeostasis over long-

term culture.

Finally, a laser treatment strategy is explored that has the capability to reform collagen

crosslinks, replacing those lost during OA progression. This novel therapy acts without injuring

the cells and without any chemical additive or thermal ablation. The laser treatment protocol

used in this application can specifically target the subsurface region, located 200 μm of the

articular surface. By strengthening this region with enhanced crosslinking, we hypothesize (H3a)

that cartilage will demonstrate greater resistance to fatigue failure than untreated controls. We

then hypothesized (H3b) that this treatment protocol would also be effective on devitalized

fibrillated human articular cartilage from OA joints with overall Outerbridge score OS1-3.

We find that for both cartilage-on-cartilage and glass-on-cartilage sliding configurations

at physiologic applied loads, long-term sliding with a low friction coefficient causes wear in the

form of delamination. We show that the use of synovial fluid as a lubricant delays the onset of

wear; and, similarly, that sliding with a cartilage counterface also reduces the incidence of wear.

In subsequent studies we fully characterize a homeostatic culture medium to emulate cartilage

in vitro behavior in synovial fluid. We show that explants cultured in this medium can maintain

their properties for at least one month and have no loss in cell viability. Laser treatment is then

tested on both living and devitalized bovine and devitalized human cartilage and the treatment is

shown to improve the wear resistance of the tissue without harming embedded cells.

Overall this work has led to novel insights that have clinical applicability. One strength of

the in vitro investigations described in this body of work is the ability to separate out

mechanically-mediated events from biochemically-mediated events, which would be impossible

in vivo. Parsing out such specific mechanisms of cartilage wear can help guide better

understanding of disease progression and drive therapeutic intervention. Intervening during the

early stages of OA offers the promise of preventive care that currently does not exist and could

provide significant benefits to a patient’s quality of life. This dissertation asserts that focusing on

delaying or preventing wear by improving the resiliency of the extant intact cartilage in early OA

is a viable strategy to improve patient outcomes and offers an innovative approach over existing

regenerative techniques.

i

Table of Contents

List of Figures .......................................................................................................................... ivList of Tables .......................................................................................................................... viiiAcknowledgements ................................................................................................................. ixChapter 1 – Introduction to Cartilage: Background & Hypotheses ....................................... 1

1.1. The Musculoskeletal System ......................................................................................... 11.2. Osteoarthritis (OA) ......................................................................................................... 21.3. Early OA treatment: Standard of Care .......................................................................... 31.4. Articular Cartilage Structure ......................................................................................... 41.5. Articular Cartilage Mechanics ....................................................................................... 61.6. Articular Cartilage Lubrication ...................................................................................... 71.7. The role of friction and wear in OA progression .......................................................... 91.8. Cartilage Delamination Wear ....................................................................................... 111.9. Live Cartilage Explant Studies ................................................................................... 141.10. Preventive treatment over regenerative treatment .................................................. 151.11. Hypotheses and Specific Aims ................................................................................. 16

Chapter 2 – Immature Bovine Cartilage Wear by Fatigue Failure and Delamination ......... 202.1. Introduction .................................................................................................................. 202.2. Hypotheses ................................................................................................................... 222.3. Materials and Methods ................................................................................................. 23

2.3.1 Experimental Design ................................................................................................ 232.3.2. Specimen Preparation ............................................................................................. 272.3.3 Testing Devices ........................................................................................................ 282.3.4. Contact Area and Stresses ...................................................................................... 292.3.5. Friction Measurements ............................................................................................ 302.3.6. Laser Scans and Surface Roughness ..................................................................... 312.3.7. Polarized Light Microscopy and Histological Staining .............................................. 312.3.8. Detecting Onset of Delamination ............................................................................. 32

2.4. Results .......................................................................................................................... 332.4.1. Damage Morphology: .............................................................................................. 332.4.2. Friction Coefficient (H1) ........................................................................................... 352.4.3. Boundary Lubrication (H2) ....................................................................................... 372.4.4. Frictional Shear S-N Curve ...................................................................................... 392.4.5. The effect of decreasing solid-on-solid contact area fraction (cartilage as counterface, H3)................................................................................................................ 402.4.6. Increasing applied stress by decreasing contact area.............................................. 41

2.5. Discussion .................................................................................................................... 422.6. Acknowledgements ..................................................................................................... 50

ii

Chapter 3 – Development a Physiologic Chemical Environment that Maintains the Homeostasis of Immature Bovine Articular Cartilage Explants in Long-Term in vitro Culture ..................................................................................................................................... 51

3.1. Introduction .................................................................................................................. 513.2. Critical chemical identification and Hypotheses ....................................................... 533.3. Materials and Methods ................................................................................................. 56

3.3.1. Cartilage Explants and Synovial Fluid ..................................................................... 583.3.2. Culture Medium Preparation .................................................................................... 583.3.3. Tissue Characterization Methodology ...................................................................... 59

3.4. Culture Experiments .................................................................................................... 603.4.1. Cartilage Explant Glucose Consumption ................................................................. 603.4.2. Cartilage Explant Culture in Synovial Fluid .............................................................. 613.4.3. Critical Chemical Mediator Experiment #1: Corticosteroids ..................................... 623.4.4. Critical Chemical Mediator Experiment #2: Insulin and Glucose .............................. 633.4.5. Critical Chemical Mediator Experiment #3: Amino Acids ......................................... 633.4.6. Explant Dynamic Loading in Varying Culture Media ................................................ 63

3.5. Statistical Analyses ..................................................................................................... 643.6. Results .......................................................................................................................... 64

3.6.1. Cartilage Explant Glucose Consumption ................................................................. 643.6.2. Cartilage Explant Culture in Synovial Fluid .............................................................. 653.6.3. Critical Chemical Mediator Experiment #1: Corticosteroids ..................................... 653.6.4. Corticosteroid-Free Explant Swelling: Enzyme Secretion ........................................ 673.6.5. Critical Chemical Mediator Experiment #2: Insulin and Glucose .............................. 683.6.6. Critical Chemical Mediator Experiment #3: Amino Acids ......................................... 693.6.7. Explant Dynamic Loading in Varying Culture Media ................................................ 70

3.7. Discussion .................................................................................................................... 703.8. Conclusions and Future work ..................................................................................... 813.9. Acknowledgments ....................................................................................................... 81

Chapter 4: Laser-based treatment strategy to delay OA progression ................................ 824.1. Introduction .................................................................................................................. 82

4.1.1. Loss of Collagen Crosslinks leads to OA ................................................................. 824.1.2. Current laser treatment therapies for cartilage ......................................................... 834.1.3. Femtosecond laser induces collagen crosslinks ...................................................... 844.1.4. Reactive Oxygen Species ....................................................................................... 854.1.5. Hypotheses ............................................................................................................. 85

4.2. Materials and Methods ................................................................................................. 864.2.1. Experimental Design Study 1: Laser treatment immature bovine at different depths 864.2.2. Experimental Design Study 2: Laser treatment OA human ...................................... 864.2.3. Experimental Design Study 3: Cell Viability after treatment in long-term culture ...... 874.2.5. Laser Treatment ...................................................................................................... 884.2.6. Equilibrium Compressive Young’s modulus ............................................................. 904.2.7. Wear tests ............................................................................................................... 904.2.8. Cell viability & histology ........................................................................................... 90

4.3. Results .......................................................................................................................... 914.3.1. Results Study 1: Laser treatment immature bovine at different depths .................... 914.3.2. Results Study 2: Laser treatment OA human ........................................................... 924.3.3. Results Study 3: Cell viability in long-term culture ................................................... 924.3.4. Results Study 4: Vitamin C for sequestration of ROS .............................................. 93

iii

4.4. Discussion .................................................................................................................... 934.5. Future Work .................................................................................................................. 984.6. Conclusions ................................................................................................................. 99

REFERENCES ....................................................................................................................... 101Appendix I: Supplemental Figures ...................................................................................... 113Appendix II: List of Publications during graduate studies................................................. 118

Full-length Papers:............................................................................................................ 118Conference Podium Presentations: ................................................................................. 118Conference Posters: ......................................................................................................... 119

iv

List of Figures

Figure 1: Schematic representation of bovine calf articular cartilage harvest locations and both

(A) Stationary Contact Area, SCA and (B) Migrating Contact Area, MCA friction testing configurations. ..................................................................................................................... 7

Figure 2: Slides of representative bovine cartilage plugs after wear testing against various materials (Safranin-O staining for GAG), showing delamination and damage (oungoulian, 2015, reproduced with permission). * denotes cartilage canals, typical in juvenile tissue. .. 11

Figure 3: Images of cartilage samples after migration against a flat glass slide under SCA config. for 8 hours. The superficial zone has swollen and sheared off, indicating fatigue failure of cartilage MZ. (Oungoulian, 2015, reproduced with permission). .......................... 13

Figure 4: Mean friction coefficient for a stationary contact area (SCA) configuration (cartilage plug against glass slide) and two migrating contact area (MCA) configurations: bovine knee condyle against tibia, and glass lens against tibia (n=6 per group) (Caligaris, 2008; reproduced with permission). ............................................................................................. 21

Figure 5: MCA experimental testing configurations for Study 1 (top) condyle as counterface (bottom) glass lens as counterface. ................................................................................... 23

Figure 6: Four different views of the ankle condyle counterface utilized in Study 2. .................. 25Figure 7: Cross-section views of the “popped” cartilage-only immature bovine condyles (A)

femoral condyle (Oungoulian, 2014, reproduced with permission) (B) ankle condyle. Biopsy punches from (A) were used as the counterface for C-SF-8-L1 and C-PBS-8-L1 of Study 1. The full “unpopped” bone+cartilage sample from (B) was used as the counterface in Study 2. ....................................................................................................................................... 26

Figure 8: Pressure-Senstive Fuijifilm contact area for representative groups. A femoral condyle is shown at 45 N to demonstrate the increase in contact area relative to ankle, which is significantly smaller under 25 lbs. ...................................................................................... 27

Figure 9: Representative temporal evolution of the friction coefficient µ from each 8 hour group (G-PBS-8-L1, G-SF-8-L1, C-PBS-8-L1, C-SF-8-L1). .......................................................... 31

Figure 10: Average µeff for samples from Study 1. Data binned by incidence of damage. P-values for comparisons from left to right: 0.0132, 0.0147, 0.003. ....................................... 35

Figure 11: Post-test 3D scans (A,C) and photographs of representative cartilage strips (B,D). A glass lens was migrated against the cartilage strip in PBS (A,B) and SF (C,D). When PBS was the lubricant the strip exhibited significant damage, in the form of a blister while strips migrated in SF showed no evidence of damage. Scale bar: 5 mm. ................................... 36

Figure 12: Polarized light microscopy depicting damage onset and damage progression in cross-sections taken through the travel paths of six strips halted mid-test at spaced intervals (N+M): (A) no damage (Fz = 4.5 N, SF, 2800 cycles) (B-F) damage (Fz = 4.5 N in PBS). Illuminated regions indicate collagen fiber alignment (SZ, surface zone; DZ, deep zone); dark regions indicate random collagen organization (MZ, middle zone). ................. 37

Figure 13: Hysteresis curves for four strips (A,B,D,F; Figure 12) depicting two cycles: cycle 10 (baseline) and the cycle at which the test was halted (final cycle). The vectors represent the first principal eigenvectors of the final cycle which exceeded the threshold and determined N. ....................................................................................................................................... 38

Figure 14: Picrosirius Red stained histological sections of strips with (A) no damage (B) blister (C) ruptured blister. Arrow indicates travel path. ............................................................... 39

v

Figure 15: Friction coefficient ratio (last hour/ first hour). Binned by different faactors. p-value from left to right: 0.0428, 0.2547, 8.2E-2 ............................................................................ 40

Figure 16: Fatigue curve (S-N) correlating frictional shear stress with cycle to damage. (Top) Data for thirty-two damaged strips is shown, all from groups using Glass as the counterface. Frictional shear stress, S, is calculated from measured frictional traction divided by the contact area measured at the beginning of the test. .................................... 41

Figure 19: Cartilage strip from Group 8 (cartilage-on-cartilage) stopped mid-test. Blister indicative of delamination is apparent (white arrows) ......................................................... 43

Figure 20: Two damaged samples from Study 2 with friction coefficients plotted and hysteresis loops for select cycles. ....................................................................................................... 44

Figure 21: Two undamaged samples from Study 2 with friction coefficients plotted and hysteresis loops for select cycles. Images of strips were taken immediately after unloading, therefore the travel path is evident. .................................................................................... 45

Figure 22: Remaining glucose levels in culture medium after 48 hours of explant incubation for varying media to tissue volume ratios. Dashed line represents initial supplemented glucose level (1 mg/mL). ................................................................................................................. 65

Figure 23: Mechanical properties and biochemical content of cartilage explants initially and after 20 days of culture in synovial fluid: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, and (D) GAG content per final weight wet. *p<0.05 represents significant variation from day 0 value. ............................................................... 67

Figure 24: Critical chemical mediator experiment #1: Mechanical properties, biochemical contents, and histology of cartilage explants cultured for 32 days in medium with varying levels of cortisol: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, (D) GAG content per final weight wet, (E) collagen content, and (F) cell content. Insulin (1 ng/mL), glucose (1 mg/mL), amino acid (1.0X), and ascorbic acid (20 µg/mL) levels were maintained constant for all groups. Red arrow indicates physiologic level of cortisol present in synovial fluid. Dashed lines represent day 0 values. *p<0.05 represents significant variation from day 0 value. (G) Representative images of cartilage explants initially (day 0) and after 32 days of culture with varying cortisol levels. Scale bar = 1.0 mm. (H) Representative safranin-O stained sections of explants initially and after 32 days of culture with varying cortisol levels. Arrow represents crack formation in tissue center of 0 ng/mL cortisol samples, which is distinct from cartilage vascular canals (* labeled) that are present in other groups and characteristic of immature cartilage tissue. Scale bar = 1.0 mm. ........................................................................................................... 68

Figure 25: Swelling ratio of cartilage explants cultured for 32 days in cortisol-free medium while subjected to devitalization, cortisol supplementation, or specific enzyme inhibitors. *p<0.05 represents statistical variation from control value. .............................................................. 69

Figure 26: Critical chemical mediator experiment #2: Mechanical properties, biochemical contents, and histology of cartilage explants cultured for 32 days in medium with varying levels of insulin and glucose: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, (D) GAG content per final weight wet, (E) collagen content, and (F) cell content. Cortisol (25 ng/mL), amino acid (1.0X), and ascorbic acid (20 µg/mL) levels were maintained constant for all groups. Red arrow indicates physiologic levels of insulin and glucose present in synovial fluid. Dashed lines represent day 0 values. *p<0.05 represents statistical variation from day 0 value. †p<0.05 represents significant variation from corresponding 1 ng/mL insulin value. (G) Linear correlation between swelling ratio and GAG content. (H) Representative images of cartilage explants initially (day 0) and after 32 days of culture for each media group. Scale bar = 1.0 mm. (I) Representative safranin-O stained sections of explants initially and after 32 days of culture for 6000 ng/mL

vi

insulin and 4.5 mg/mL glucose group. Arrow represents crack-induced tissue void in tissue center. Scale bar = 1.0 mm. ............................................................................................... 71

Figure 27: Critical chemical mediator experiment 3: Mechanical properties, biochemical contents, and viability of cartilage explants cultured for 7, 16, or 32 days in medium with varying levels of amino acids: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, and (D) GAG content per final wet weight. Insulin (1 ng/mL), glucose (1 mg/mL), cortisol (25 ng/mL), and ascorbic acid (20 µg/mL) levels were maintained constant for all groups. Red arrow indicates physiologic level of amino acids present in synovial fluid. 1X amino acid level represents concentrations present in standard DMEM. Dashed lines represent day 0 values. *p<0.05 represents significant variation from day 0 value. †p<0.05 represents significant variation from corresponding day 7 value. (E) Representative viability images (live cells: green, dead cells: red) of cartilage explants after 32 days of culture for 0.25X and 0.03X amino acid groups. ............................................... 72

Figure 28: Secretion rate of endogenous cartilage growth factor, latent TGF-b1, from cartilage explants cultured in medium of varying levels of amino acids. Secretion rate is normalized to explant volume. Red arrow indicates physiologic level of amino acids present in synovial fluid *p<0.05 represents significant variation from 1.0X value. ........................................... 73

Figure 29: Mechanical properties and biochemical contents of cartilage explants cultured for 32 days in a corticosteroid-free, physiologic, or supraphysiologic culture medium while subjected to dynamic mechanical loading or free swelling conditions: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight. *p<0.05 †p=0.07 represents statistical comparison between corresponding free swelling value. .................. 74

Figure 30: Human OA distal femur with overall Outerbridge score 3. Dashed line represents histology section of Fig 31b. Circles on lateral condyle (left) are sites of tissue harvest for laser treatment study, rectangles show paired samples. Inset (bottom right) shows harvested sites of samples used for results reported in Figure 34. ..................................... 87

Fig 31. (a) Human femoral condyle (OS3), with evidence of delamination (arrow). (b) Human femoral condyle (OS3) showing missing surface zone, with typical fibrillation. (c) Delaminated bovine cartilage strip (see Fig 8) shows fibrillated surface (arrow). (d) Typical OA plug from Fig 2, graded OS2, test results in Fig 9b. (Picrosirius Red for collagen.) ...... 88

Fig 33. Laser setup consisting of Hi-Q femtosecond oscillator coupled with the 3- axis motorized actuators. Beam delivery optics are mounted onto the motion system ............................... 89

Figure 34: Stiffening of articular cartilage with laser treatment: (a) immature bovine treated from 0-200 µm and 100-300 µm from surface, (b) mature human OA cartilage with Outerbridge scores 1 and 2, treated from 0-200 µm from surface (PS2), *p<0.005, † p<0.05. ............... 91

Figure 35: Control bovine cartilage explant (Æ3 mm ´ 1 mm), before and after wear test, demonstrates greater visual evidence of surface damage (arrows) compared to laser-treated explant (side view). ................................................................................................ 92

Figure 36: Live/Dead staining of immature bovine cartilage explants (PS3): (a) control and (b) laser-treated, at 24 h; (c) control and (d) laser-treated at 2 wks. ........................................ 93

Figure 37: Young’s Modulus for samples after laser treatment in either PBS + Vitamin C or PBS alone. ................................................................................................................................. 94

Figure 38. LEFT: (a) Photographs and (b) laser surface scans of Æ10 mm X 1.2 mm bovine cartilage plugs after sliding for 12 h, 4 mm travel at 1 mm/s, against a spherical glass lens (R12.7 mm) under 4.45 N load. Arrows: blister due to delamination wear in control sample (left). Treated sample (right) was lased over a 3 mm x 3 mm central patch and showed no damage. RIGHT: Friction coefficient µ remains low over MCA 12 h wear test (µ=0.008

vii

damaged control, µ=0.005 undamaged treated). Typical results from SCA 4 h test (Fig 9a) show much higher, time-dependent µ for both control and treated. .................................... 95

Figure 39: Envisioned treatment process: (1) Femtosecond laser irradiates articular cartilage creating an ionization field; (2) ionization generates reactive oxygen species (ROS) which interact with collagen fibrils in the ECM; (3) biochemical reactions with ROS result in CxL formation, which enhance cartilage mechanical properties, potentially slowing down OA progression. ....................................................................................................................... 96

Fig 40. (a) Laser setups for fiber optic probe (red) & free-space optics (blue). A. Femtosecond laser; B. Tissue sample area; C. 3-axis translation mount; D. Fiber optic cable; E. Probe tip; F. Focusing optics; G. 3-axis motorized actuator. Gaussian laser beam profile with ~60 mW power is obtained from (b) free-space objective and (c) fiber-based probe. ....................... 99

Figure S1: Chapter 2 - Strip before and immediately after wear test from Study 2 using ankle-condyle as counterface. ................................................................................................... 113

Figure S2: Chapter 2 - Friction time courses for two samples from Study 1 that both experienced delamination wear. ...................................................................................... 113

Figure S3: Chapter 2 - Wear test of glass lens against bovine cartilage strips: (a) Photograph showing delaminated blister (arrow). (b,c,d) Cross-sections (PLM) show delamination progression in travel path over time (3 different strips). .................................................... 114

Figure S5: Representative viability images (live cells: green, dead cells: red) of cartilage explants after 32 days of culture in medium with physiologic glucose levels (1 mg/mL) at either a 30:1 or 100:1 media to tissue volume ratio. Yellow arrows represent explant surfaces that were exposed to culture medium. The explant bottom surface was in contact with the polystyrene culture dish throughout culture. Explants cultured at the lower media volume ratio exhibit a central region of significant viability loss. This central region exhibits the highest characteristic length from the media exposed surface and therefore is expected to possess the lowest glucose levels, as illustrated in prior work [174]. ............................ 115

Figure S6: Mechanical properties and biochemical contents of cartilage explants cultured for 32 days in medium supplemented with cortisol (Cort) or dexamethasone (Dex) corticosteroid at 50 ng/mL level: (A) swelling ratio, (B) compressive Young’s modulus, , (C) GAG content per initial wet weight, and (D) GAG content per final wet weight. Insulin (1 ng/mL), glucose (1 mg/mL), amino acid (1.0X), and ascorbic acid (20 µg/mL) levels were maintained constant for both groups. The different corticosteroids had no statistical effect on measured mechanical or biochemical properties (p>0.05). Dashed lines represent day 0 values. ... 116

Figure S7: Mechanical properties and biochemical contents of cartilage explants cultured for 32 days in medium supplemented with or without 10% fetal bovine serum (FBS): (A) swelling ratio, (B) DNA content per initial wet weight, (C) collagen content per initial wet weight, (D) GAG content per initial wet weight. Insulin (1ng/mL), glucose (1mg/mL), amino acid (1.0X), cortisol (25 ng/mL), and ascorbic acid (20 µg/mL) levels were maintained constant for both groups. FBS supplementation induced increased explant swelling and cell proliferation but had no effect on COLi or GAGi. Dashed lines represent day 0 values. *p<0.05 represents statistical difference between groups. ............................................................................ 117

viii

List of Tables

Table 1. List of test groups used in Study 1. For all groups, sliding was performed using the specified ‘Counterface’ against a cartilage strip, using the specified ‘Lubricant’ and contact ‘Load’, over the given ‘Duration’. The number of specimens used in each group is given under ‘Sample Size.’ .......................................................................................................... 24

Table 2: Average friction coefficient (over the entire test duration) and measurement of damage for all groups. ..................................................................................................................... 34

Table 3. Results of test groups used in Study 1. The number of specimens used in each group is given under ‘n’. The number of strips that were morphologically damaged and undamaged as determined by observation and measures of surface roughness changes under the given testing configuration are listed in those respective columns. The number of samples that had disagreement between observed damage and predicted damage by hysteresis examination are listed under false positives. The column N, is the cycle at which initiated failure occurred as predicted in post-hoc interrogation of hysteresis curves. ......... 42

Table 4: Concentrations of metabolic mediators in conventional media formulations and those present in native synovial fluid. Physiologic glucose, amino acid, and ascorbic acid levels are reported from plasma measurements under the general assumption that they exhibit relatively unobstructed transport across the synovium. Amino acid concentrations are listed as the average concentration among the 15 amino acids included in standard DMEM (see Table S4 for details). Cortisol levels (total and free) are obtained from direct experimental measurements performed on synovial fluid in the current investigation. ............................. 54

Table 5: Mediator concentration variation experiments performed in this study. 1.0X amino acid concentration is the level present in standard DMEM. ........................................................ 57

Table 6: Levels of critical metabolic mediators in physiologic-based culture medium required to maintain articular cartilage explant homeostasis. ............................................................... 75

Table S4: Concentrations of essential (e) and non-essential (n) amino acids in Dulbecco’s Modified Eagle’s Medium (DMEM), Ham’s F12 Medium, Minimum Essential Medium (MEM) formulations, and in blood plasma according to two prior investigations [150, 151]. Several non-essential amino acids (alanine, asparagine, aspartic acid, glutamic acid, and proline) have been omitted from the table due to their absence from standard DMEM formulations. ........................................................................................................................................ 114

ix

Acknowledgements

First, to the giants whose shoulders I stand upon, thank you curious humans who have

led the way in the pursuit of observation, understanding and expression. Second, to those

people who broke down barriers to give women a voice in directing human progress, we are all

in your debt.

“I have learned that people will forget what you said, people will forget what you did,

but people will never forget how you made them feel.” -Maya Angelou

Despite setbacks and disappointments, both in the lab and in society, the academic

bubble incubates a universe, a family, of freedom of thought and expression that I am so

grateful to have been a part of here at Columbia. I have so many people to acknowledge

because this has been such a substantial journey. At this university, I have had a number of

great teachers – from high school students exploring research for the first time to professor

emeriti, who built the very department from which I will obtain this PhD degree. I conclude my

graduate studies and enter the world feeling empowered and optimistic both personally and

professionally because of each and every one of the following individuals.

This work would not have existed without the assistance and talents of a great number

of collaborators. I’d like to first thank the undergraduate and summer students who offered their

brain, sweat and, most importantly, their time to realizing the work presented here. These

students include: AJ Singh, Jason Suh, Hyeon-Jin (Ashley) Koo, Marina Morrow, Wing-Sum

Law, Michael Anne Bolene, Katherine Guan, Deyanisse Monegro, Adriano Atallah, Mike

x

Parrish, Katherine Cavanaugh, and Anna Bute. Your effort is inestimably valued and

appreciated.

To all the other students I have worked with throughout the years including these stellar

Master’s students: James Boorman-Padgett, Shivam Sonar, Sonia Donde, Ricardo Palacios,

Polina Smirnova, and GT Salamone. I am proud to have worked with you and I am so grateful

for all the hours and brain-power you put into making the impossible, possible. Although this is

my dissertation, I share the honor with you all.

To ‘Equipe 2’ at INSERM U747 in Paris: Francois Rannou, Maite Corvol, Marie-France,

Lydia Tsagris, Caroline Chauvet, Francois Etienne, Irene, Elissar El-Hayek, Thuy-anh Auxietre,

Nabila Hassaine, and Clementine Perriere. From all of you I learned invaluable technical skills

and built completely new perspectives on science, engineering, culture and society. You are

the reason I committed to work towards a PhD. That experience stands out as a critical turning

point in my life and I was beyond fortunate to have joined such a welcoming, supportive team.

To the wonderful students from Ecole Polytechnique who had the reverse-experience of coming

from Paris to work in a lab in New York: Lucie Karbowski, Antoine Dusseaux, Lea Montegut and

Martin Xiberras – I am consistently impressed at the highest level with each and every one of

you. I’m inspired not only by your work ethic, but by your charisma and commitment to

advancing science for the greater good.

To the members of the Cellular Engineering Laboratory, my gateway into the world of

research, I am eternally grateful. Dr. Eric Lima orphaned me on the doorstep of the lab in 2008

and I never looked back. To Andrea Tan, Liz Oswald, Grace O’Connell, Adam Nover, Sonal

Sampat, Terri-Ann Kelly and Elena Algere-Ageuron – thanks for showing me the ropes with

patience, trusting my judgement, and helping me to build confidence as a researcher. To my

contemporary CEL lab generation: Eben Estell, Rob Stefani, Amy Silverstein, Brendan Roach,

xi

and Andy Li – thank you for always being there as a sounding board for ideas and for sharing

your skills, knowledge, resources, and fantasy football.

To Dr. Kristin Myers, Dr. Karen Kasza, Dr. Kyoko Yoshida, and Andrea Westervelt,

thank you for all of your support throughout the years. All four of you lead by example in your

own ways, and I couldn’t have asked for better role models of women who are reshaping

mechanical engineering (and the world).

To Dr. Sinisa Vukelic and Chao Wang, our all-star collaborators, I have learned so much

about teamwork and integrity from working alongside you both. I am so proud of what we have

accomplished and I thank you for your creativity and flexibility throughout the years.

To Arthur Autz, Bob Stark, Keith Yeager, Mohammed Haroun and Bill Miller who were

always available to help design, customize, and build the devices that enable the science –

“Making scientists dreams come true.”

To Dr. Eric Lima, my inspiration and motivation to move into the world of Biomedical

Engineering. Thank you for teaching me, believing in me, and trusting me. You are a true

educator in all senses of the word. You have a holistic, big-picture view of science and

engineering education and it’s a privilege to have been your first Master’s student and (self-

proclaimed) protégé.

To Dr. Michael Albro, thank you for letting me be your shadow in my early years in the

lab. From you, I learned everything about forming hypotheses and designing experiments. You

are a natural leader and well-rounded human. Thank you for being a mentor and a friend.

To Dr. Katherine Reuther, thank you for trusting me to help design and teach your

Master’s level device design course. Katie, you inspire me everyday not only to be a better

teacher and leader, but to be a better person. The students were truly blessed to have you join

the department and I couldn’t have been luckier to have gained such an inspiring mentor and

friend.

xii

To Dr. Chloe Bulinski, who taught me the meaning of life in your cell biology course and

(quite literally) when you bestowed upon me clippings of your beautiful office plants years ago.

Those plants continue to thrive and as I’ve watched them through the years, and upgraded them

from pot to ever-larger pot, I’m reminded of this process of growth – of building knowledge and

experience and sharing it with others. Thanks, Chloe.

I owe my biggest thank you to my labmates, past and present, Sevan Oungoulian, Bob

Nims, Brian Jones, Alex Cigan, Chieh Hou, Brandon Zimmerman, and Jay Shim. To learn from

and alongside these gentlemen was a privilege and I couldn’t have asked for a more dedicated,

resourceful, intelligent group of engineers and scientists. Each one of them has contributed

enormously to this dissertation in immeasurable ways. To name a few: Sevan and Brian spent

years building the devices I used to perform friction tests and helped me configure them for my

own experiments, Alex was my bounce-board for all things explant culture and things of a

chemical nature, Chieh committed himself to helping solve our issues with depth-dependent

measurements, Jay (a secret experimentalist at heart) built a functional plug-in for FEBio that

enabled Brandon to model the frictional damage experiments performed here, Brandon also

helped me enormously with the human tissue laser experiments, and finally, Bob (soon to be

Professor Nims and my fellow Baby Bloomer) was always available to talk science and

brainstorm with me, whether about cartilage or breast pumps – and he was the python wizard

who enabled the post-hoc examination of hysteresis loops. In addition to practical help and

guidance, these guys were great co-workers, teammates, and friends to me over the years. I

want to officially welcome Courtney Schaeffer to the team, who helped me through my final

semester and the last leg of my thesis by expertly performing damage experiments and laser

treatments with precision and ease – I wish you all the best as you embark on your own PhD

journey.

xiii

I was blessed to be co-advised by two professors who complement each other well – Dr.

Clark Hung and Dr. Gerard Ateshian. I am so grateful to Clark for allowing me to work in the lab

when I was an undergraduate with limited lab experience and for supporting me during the work

toward my Master’s degree. Clark offered me my first teaching assistant experience during his

summer high school course, and I remember feeling valued and being trusted for the first time

with the great and serious task of teaching. That summer I grew roots at Columbia in addition to

a love of research and teaching and I thank Clark for that life-changing opportunity. I knew and

respected Gerard all-the-while, but wasn’t aware he knew me until I learned much later that he

knew me as “being outspoken in lab meetings” (accurate). I’m so happy and fortunate that

Gerard took a chance on this outspoken scientist when he decided to grow the lab in 2012. I

have learned so much from Gerard, things spanning both the technical and the non-technical:

blending disciplines, hypothesis-driven research, sound experimental design, grant writing, oral

and written presentations, story-telling, teaching, and a love and dedication to one’s craft. Since

Gerard can reliably always be found brooding in his office, he has been a terrific mentor who is

willing to drop his own work to help his students advance theirs. Going forward I take with me a

lot of hard and soft skills from both Clark and Gerard and I am so grateful for their guidance and

mentorship over the years.

To my friends Anina Stanton, Dr. Michael Waite, Dr. Charlotte Gartenberg, Ivan

Himanen, Tanya Neuman, Teresa Farrell, Alexa Mattiacci, and Dr. Susanna Makela who have

encouraged me, pushed me, kept me sane and kept me questioning. Thank you for helping me

to get out of my niche, to get out of my own head and to think big and think broad.

To my grandparents Kay and John Scipione who have been full of only support and love

throughout this journey (best roommates ever). To my grandparents Mary and Edward Durney,

who I know are proud of me. To my parents Joanne and Gerard who truly embody both what it

means “to parent” and “to be a parent.” These two have both held me close and let me fly

xiv

through the years and there’s no better feeling than making them proud after all they’ve done

and sacrificed for our family. To my second family: Stacey, Rich, Daniel, and Gregory Antonelli.

It’s an honor to be your daughter and I thank you for all of the support, encouragement and love.

I’m so blessed to be sandwiched in between two phenomenal sisters, Kerri and Kaitlin, who

have taught me more than graduate school ever could. We three are so different, and yet so

similar, and I’m eternally grateful for all the challenges and love you’ve thrown my way over the

years. To my nephews Mikey and Happy; spending time with you guys and watching you grow

is the best reminder to stay curious and stay questioning. Keep it up you little scientists.

To my best friend and lifemate, Stephen Antonelli. You are a constant inspiration of

honesty, self-awareness and self-reflection. Thank you for constantly setting the bar high for

yourself and pursuing your passions with hard work, energy, integrity, attention-to-detail and

follow-through. You are a man of your word, who knows yourself, strengths and weaknesses,

and fortunately a bit of that has started to rub off on me. Thank you for having a heart the size

of an ocean and a spirit the weight of a feather. You and I can get through anything.

New York, June 2018

xv

For my parents.

1

Chapter 1 – Introduction to Cartilage: Background & Hypotheses

1.1. The Musculoskeletal System

The human body is made up of hard and soft tissues. The hard bones that comprise our

skeletons are aligned and joined by soft connective tissues and taken together this unit is called

the musculoskeletal system. Functions of this system include: supporting the body, allowing

motion, and protecting vital organs. The musculoskeletal system is inherently both a biological

system and a mechanical one. The study of this system in this hybrid-fashion is called

“musculoskeletal biomechanics” and overlaps multiple disciplines including: biology, cell biology,

molecular biology, mechanics, elasticity, physical chemistry, electrostatics, electrodynamics,

and physiology. Furthermore, it can be studied at many scales: molecular, cellular, tissue,

organ, and organism. Investigations into the field of musculoskeletal biomechanics originated

with the goal of understanding both the healthy function and breakdown of this system, with the

ultimate aim being to develop solutions to prolong years of human health. Breakdown of the

musculoskeletal system usually results in disability, immobility, and often pain for the patient.

According to the World Health Organization, musculoskeletal conditions are the second

largest contributor to disability worldwide; between one in three and one in five people are

currently living with a painful or disabling musculoskeletal condition [1]. The most common,

disabling musculoskeletal conditions are osteoarthritis, back and neck pain, and inflammatory

conditions such as rheumatoid arthritis [1]. This dissertation will focus on strategies to study

articular cartilage in the lab, with the ultimate aim of translating these findings toward developing

clinical therapies for osteoarthritis.

2

1.2. Osteoarthritis (OA)

Osteoarthritis (OA) is a progressive, complex, multi-tissue joint disease with degenerative

changes in the articular cartilage and subchondral bone [2], with a long asymptomatic early

development and debilitating late stages. More than 27 million Americans suffer from

degenerative diseases of articular cartilage, such as osteoarthritis (OA) [3]. The primary

function of articular cartilage is to transmit loads across the joint surfaces while simultaneously

minimizing friction and wear. Progressive OA is a years-long process comprised of a loss of the

articular cartilage and exposure of the underlying bone. The initiating events in OA are believed

to be multi-faceted; risk factors may be associated with age, sex, obesity, diet, genetics,

race/ethnicity, joint laxity, misalignment, abnormal loading, or traumatic injury [4]. Regardless of

the initiating factor(s), once the synovial environment and cartilage tissue homeostasis have

been compromised, it is believed that articular cartilage degeneration and wear is mediated by

the mechanical stress to which it is subjected under normal joint function. There are limited

treatment options available today to slow down progression of OA in its early stages and most

interventions, such as highly invasive partial and total joint replacement surgeries, are

performed only at the late stages of the disease.

The focus of this dissertation will be on the articular cartilage of the knee joint, one of the

most common joints affected by OA. The articular cartilage of the femoral condyles (medial and

lateral) interacts with the articular cartilage of the tibial plateaus (medial and lateral) at the knee

joint. Besides articular cartilage, there are other soft tissue components that comprise the knee

joint and the musculoskeletal system and enable its motion. Ligaments extend from bone to

bone, acting as ropes to keep joints in alignment. Muscles transition seamlessly into tendons,

which anchor firmly to bones and power skeletal movement. The meniscus is a specialized

structure in the knee joint that provides additional stabilization and protection of underlying

3

structures. OA today is viewed as a disease that afflicts the joint as an entire system, with

inflammation, injury, and changes in bone, articular cartilage, and synovial fluid as potential

driving forces. Regardless of the driving force behind OA, degradation of the extracellular

matrix (ECM) components of articular cartilage is key to the progression of the disease and

therefore warrants investigation [5, 6].

The onset of OA is highly variable from person to person and largely unpredictable. It is

widely believed that regardless of OA initiation (injury, age, or other), there is a cascade of both

mechanical and biochemical factors that influences OA progression [7]. One strength of the in

vitro investigations described in this body of work is the ability to separate out mechanically-

mediated events from cellular-mediated and biochemically-mediated events, which would be

impossible in vivo. Parsing out such specific mechanisms of OA progression can help guide

therapeutic intervention. Intervening during the early stages of OA offers the promise of

preventive care that currently does not exist and could provide significant benefits to a patient’s

quality of life.

1.3. Early OA treatment: Standard of Care

There are limited arthroscopic treatment options for early OA to delay progression. For a long

time, arthroscopic chondroplasty (or debridement) represented the standard treatment for early

OA. Chondroplasty involves arthroscopically trimming damaged, fibrillated cartilage regions,

with the expectation that managing loose edges will help prevent further wear. Starting from the

early 2000s, rigorous outcome studies began to show that chondroplasty was no more effective

than a sham control for the purpose of relieving pain or improving function [8, 9]. Nowadays,

chondroplasty is still performed, but it is coupled with the restoration technique called

microfracture, or marrow stimulation, where small holes are drilled into regions where cartilage

has been worn away allowing mesenchymal stem cells (MSCs) to enter this region from the

4

marrow and shield the exposed bone. MSCs create fibrocartilage (scar tissue) in the defect, but

they do not restore the native biomechanics of articular cartilage that was once there [10]. In

the short term, microfracture has been shown to reduce patient pain for small lesions, but

limited success is observed in larger lesions [11-15]. In a 2017 clinical review article [16], the

authors summarized their findings on microfracture: “Meta-analyses confirm that results for

larger lesions deteriorate over time (after 2 years) and that after 5 years any size lesion

treatment can be expected to fail [15, 17, 18].” Other treatment modalities, such as

viscosupplementation with injectable hyaluronic acid, have also produced mixed results, with no

significant difference against placebo [19]. Therefore, when a surgeon detects local cartilage

fibrillation or softening during arthroscopy, no reliable modalities exist today to provide localized

repair of damaged extant tissue to prevent disease progression.

1.4. Articular Cartilage Structure

Articular cartilage lines the ends of bones at joints (where one bone meets another) in order to

cushion that interaction and facilitate smooth, near frictionless sliding. All diathrodial joints

(knee, hip, elbow etc.) are enclosed in a strong fibrous capsule [20]. The inner lining of this

capsule is called the synovium and is a highly vascularized tissue that secretes the synovial

fluid into the joint space, providing the nutrients required by articular cartilage cells [21]. The

synovium, synovial fluid and cartilage layers comprise the joint space and with supporting bone

in the mature animal system, form a “closed” biomechanical system [22]. The main function of

articular cartilage is to uniformly transmit the high stresses associated with joint loading and limit

friction within the articulating joint. In order to provide support in such a demanding loading

environment, cartilage is dependent on its unique material properties and dense extracellular

matrix (ECM) components. Cartilage can be regarded as a multi-phasic material with two major

phases: a fluid phase made up of water and electrolytes and a solid phase made up of

5

collagen, proteoglycans, other proteins, and chondrocytes [23-27]. The composition of articular

cartilage is typically: 68-85% water, 10-20% collagen (by wet weight), 5-10% proteoglycans (by

wet weight) [28]. Cartilage is known to have a low cell density (1-10% tissue volume),

containing chondrocytes that are very active during skeletal development and exhibit a

metabolism that slows down once skeletal maturity is reached. Chondrocytes respond to and

maintain their environment, meaning that they synthesize and degrade all components of the

ECM in response to environmental cues [29]. Adult cartilage is avascular (has no direct

connection to the blood supply) which means, as previously mentioned, that nutrients,

metabolites, and waste products are all exchanged through diffusion via the synovial fluid (SF),

which also acts to lubricate all cartilage-on-cartilage interactions [5].

Another important feature of articular cartilage is the inhomogeneity of its biochemical

components and the corresponding inhomogeneity of the mechanical properties throughout the

tissue depth. Articular cartilage is traditionally divided into three main zones: the superficial

zone (SFZ), the middle zone (MZ), and the deep zone (DZ). The SZ interacts with the opposing

cartilage surface and the DZ transitions to become bone. The water content in articular

cartilage decreases as a function of depth, with the SZ having the highest water content of the

three zones [25, 30, 31]. Collagen content also decreases through the depth from the SZ to the

DZ while PG content increases [27, 30, 32]. The organization of these molecules also varies

with tissue depth. Collagen in the SZ is tangentially oriented along the articular surface, the

collagen of the MZ is less-tightly packed and randomly oriented [33-39] and DZ collagen is

oriented perpendicular to the articular surface. This tightly organized structure enables cartilage

to perform its unique mechanical functions.

6

1.5. Articular Cartilage Mechanics

Healthy cartilage functions as an effective shock absorbing tissue with minimal friction and wear

for smooth joint movement. The load-bearing properties of articular cartilage are largely

achieved through the interaction of its two major ECM components: collagen and proteoglycans

(PGs). Collagen provides the tensile strength while PGs provide compressive stiffness. The

highly hydrophilic PGs attract water and swell, while the interwoven and crosslinked collagen

fibril network resists the swelling pressure of the PGs with its high tensile strength, yielding a

“mechanically taut composite structure capable of sustaining compressive loads” [40].

The function of articular cartilage is mainly to bear compressive loads. However, it is

well understood that articular cartilage has a low compressive Young's modulus (EY; 0.1–1

MPa) [41-43] relative to its high tensile Young's modulus (EY,T; 3–6 MPa) [44-46]. This

tension-compression non-linearity regulates the mechanical response of cartilage in unconfined

compression [47-52] and under sliding contact configurations [53]. The viscoelasticity of

cartilage in compression arises due to a combination of (1) the high EY,T that resists lateral

expansion (2) pressurization and flow of interstitial fluid whereby frictional drag is experienced

between the solid matrix and the fluid phase [26, 54]. Upon application of a compressive load,

the interstitial fluid cannot exude rapidly, yielding a nearly ischoric response of the tissue. This

remarkable fluid pressurization not only plays a role in the dynamic response of the tissue, but

also has been indicated as the primary mechanism that imparts low friction and low wear to the

tissue [26, 54-56].

From the cartilage mechanics literature we know that interstitial fluid pressure subsides

over time under loading configurations of creep or stress relaxation and that the friction

coefficient increases progressively with time [26, 53, 54, 56-59]. These types of loading

configurations are called stationary contact area loading configurations (SCA) (Figure 1A). It

7

has been shown by our group that an alternative loading configuration called migrating contact

area (MCA) can both maintain fluid pressurization and a low friction coefficient [60] (Figure 1B).

Figure 1: Schematic representation of bovine calf articular cartilage harvest locations and both (A) Stationary Contact Area, SCA and (B) Migrating Contact Area, MCA friction testing configurations.

1.6. Articular Cartilage Lubrication

Fluid film lubrication was long believed to be the primary mode of cartilage lubrication. Fluid film

lubrication is a standard mode of lubrication where a lubricant forms a thin layer between

bearing surfaces, creating a gap between the two, separating them from direct contact.

Because of the viscous nature of synovial fluid, this type of lubrication seemed like a good

candidate to describe cartilage lubrication, however extensive studies have failed to verify this

mode of lubrication [61, 62].

Another form of lubrication that has been proposed for cartilage is boundary lubrication

wherein molecules in the synovial fluid coat the cartilage surface and produce low friction due to

molecular-level repulsive forces. A number of studies have suggested that the friction

coefficient of cartilage is lower when experiments are conducted in SF instead of saline. This

8

reduction ranges approximately from a factor of 2 to 6 [60, 63, 64], although the occurrence of

rolling friction (viscoelastic energy dissipation in cartilage) sometimes masks the beneficial

effects of SF over saline [60]. This reduction in µ has been attributed to the presence of

boundary lubricants in SF. Several constituents of SF have been hypothesized to serve as

boundary lubricants, including hyaluronan [65, 66], lubricin/PRG4 [67-71], and phospholipids

[72, 73]. Interactions of some of these molecular constituents have also been reported in the

reduction of µ against controls, along with a dependence on their concentration [64, 74].

A third form of cartilage lubrication that has been extensively researched by our group is

lubrication by interstitial fluid pressurization. Through this mode of lubrication, it has been

proposed that the articular layers come in direct contact and that the pressurized interstitial fluid

within the tissue supports most of the contact load. Since cartilage has a high water content in

the SZ (~90% [75]), two articulating layers of this constitution experience mostly fluid-fluid

contact or fluid-solid contact (98%), and the remaining 2% can then be accounted for by

frictional solid-solid interactions [58]. The friction coefficient will therefore rise as fluid load

support diminishes, and this portion of the load will then shift to the support of frictional solid-

solid interactions. Thus, as long as the fluid pressure remains high, frictional interactions

remain low. There has been strong agreement in the literature between theoretical predictions

and experiments regarding the role of fluid load support in lubrication [53, 58, 59, 76-78].

In articular cartilage, boundary lubrication has been shown to complement the role of

interstitial fluid load support, as the friction coefficient µ in SF and saline has been shown to

exhibit the exact same temporal response associated with loss of fluid load support [63].

Cartilage lubrication is therefore currently understood to be mostly a function of the combined

effect of interstitial fluid pressure and boundary lubrication, both of which are explored in this

dissertation.

9

1.7. The role of friction and wear in OA progression

Regardless of the initiating factor(s) (PTOA or aging-related idiopathic OA), once the synovial

environment and cartilage tissue homeostasis have been compromised, it is believed that

articular cartilage wear is mediated by the mechanical stress to which it is subjected under

normal joint function. This process is thus aptly described as “wear and tear.” From the

tribology literature mechanical wear is defined as the removal of material from contacting

surfaces due to mechanical action between them [79]. From biotribology, and specifically in the

case of cartilage, mechanical wear can occur by either abrasive wear, adhesive wear (where

debris is involved in wear generation), or fatigue wear [80]. In addition to mechanical wear,

biological tissues like cartilage experience concomitant biochemical degradation (which can be

driven by a host of factors). It is unclear which of these mechanisms of wear prevail in OA, and

under what conditions.

Studies of the frictional properties of healthy articular cartilage have been on-going since

the 1930s [54]. In order to examine wear mechanisms there are many in vitro models of PTOA,

but very few in vitro models that examine non-injurious models of idiopathic OA at relevant

loads, speed, configurations, and testing durations [81-83]. At first glance this makes sense,

because we do not have 70 years (a whole human lifespan) to run such a study in the lab.

Therefore, compromises are made either experimentally (to accelerate damage) or surrogate

damage endpoints are used. Utilization of in vitro frictional sliding test bed helps to isolate and

examine modes of cartilage wear under normal, near-physiologic loading of living and

devitalized samples. These in vitro and in situ investigations require complex testing

environments and sophisticated control over, and measurements of, applied loads and cartilage

response. The friction coefficient, µ, is collected as the typical standardizing metric for the

smoothness or roughness of the sliding interaction between two surfaces and is defined as the

ratio of the tangential force (FT) to the normal contact force (FN) acting across the bearing

10

surfaces. The friction coefficient of articular cartilage is not constant [54]. The lowest reported

value of μ for cartilage against glass is typically μ≈0.002 [76], which is exceptionally low.

However, μ may rise over time, depending on loading conditions, to achieve values as high as

μ≈0.15 against glass [76], or even μ≈0.5 against stainless steel [57, 84]. These values have

been shown to be detrimental to cartilage. When cartilage is loaded against cartilage, the

friction coefficient remains constant for sustained durations; for human cartilage, it is typically

μ≈0.020 in synovial fluid (SF) and μ≈0.025 in saline [60].

The cartilage mechanics literature has mostly focused on examining the friction

coefficient μ as a surrogate for understanding wear and tear in cartilage [28]. Implicitly, a low

value of μ has been assumed to produce low wear while an elevated value could eventually

lead to significant wear. This assumed link between µ and wear has driven in vitro studies of

cartilage wear to focus on detection methods that anticipate abrasive wear: surface

topographical changes and wear debris characterization – [80, 82, 85]. The primary quantitative

approaches to studying wear proposed to date include biochemical assaying of cartilage and

test solutions [86], characterization of changing articular layer thickness, and changes in surface

roughness [57].

However, based on a recent study [83] (described in Section 1.8.), we believe that the

primary wear mechanism in cartilage is not abrasive wear, but instead fatigue wear with

delamination of the surface zone (Figure 2). Delamination of cartilage layers is a clinically

recognized symptom of OA [87-89] although some prior reports may also conflate subsurface

cartilage delamination with delamination at the cartilage–bone interface. To our knowledge, our

recent study was the first to recapitulate this damage mechanism in vitro under pure sliding,

providing significant evidence in support of this damage mechanism [83]. The distinction

between abrasive wear and fatigue wear is significant because it implies that treatments aimed

11

at targeting exclusively boundary lubricants and/or the surface zone of cartilage may be limited,

since the tissue’s true vulnerability may lie beneath the surface.

Figure 2: Slides of representative bovine cartilage plugs after wear testing against various materials (Safranin-O staining for GAG), showing delamination and damage (oungoulian, 2015, reproduced with permission). * denotes cartilage canals, typical in juvenile tissue.

1.8. Cartilage Delamination Wear

In a recent study, we investigated wear of cartilage against materials used in joint

hemiarthroplasties [19], a procedure in which only one of the articular surfaces of a joint shows

12

signs of wear, so only that half is replaced and the other apposing surface is left intact. This

study was motivated by the observation that the native cartilage layer in a hemiarthroplasty

undergoes accelerated cartilage wear [90-94] often leading to a follow-on total joint replacement

in a few years. For this study, two cobalt chromium alloys (CoCr HC and CoCr LC) and a

stainless-steel alloy (316SS) were selected as counterfaces and were slid against native

immature bovine articular cartilage plugs, with smooth glass serving as a control. A higher

surface roughness Ra was also selected for one of the cobalt chromium alloys (CoCr LC-

Ra25nm), well within the range considered acceptable by regulatory agencies (Ra ≤ 50 nm).

Immature bovine cartilage disks were subjected to a wear test in phosphate buffered saline

(PBS), in an unconfined compression configuration that accelerated loss of interstitial fluid

pressurization using a stationary contact area configuration (Figure 1A) to replicate conditions

believed to occur in hemiarthroplasties (on a much shorter time scale) [95]. The prescribed

contact stress was low (0.18 MPa) and reciprocal sliding (±5 mm at 1 mm/s) was performed for

4 h (N=1,440 cycles). Results showed that considerably more damage occurred in cartilage

samples tested against 316SS (Ra=10 nm) and CoCr LC-Ra25nm compared to glass (Ra=10

nm), and smoother CoCr HC and CoCr LC (Ra=10 nm) (Figure 2). The two materials producing

the greatest damage also exhibited higher equilibrium friction coefficients.

An inherent assumption of this study was the expectation that the primary mode of

cartilage wear was abrasive. Instead, it was observed that cartilage damage occurred primarily

in the form of delamination at the interface between the superficial zone (SZ) and the

transitional middle zone (MZ) (Figure 2). Based on histology and low particulate volume, there

was little evidence of abrasive wear at the articular surface. In the cartilage samples tested

against 316SS and CoCr LC-Ra25nm, the delamination was occult, becoming visible only on

histological sections. However, evidence of sub-surface, pre-delamination damage was also

found for other counterface materials (Figure 2). Since the samples migrated against glass

13

showed no damage after 4 hours, experiments were repeated on a separate set of samples for

8 hours (N=2,880 cycles). For that duration, glass exhibited major damage in some instances

causing the superficial layer to shear completely off (Figure 3). This finding is consistent with

the concept that damage is caused by fatigue failure: even when the friction coefficient is lower,

a longer testing duration (more cycles of loading) may accumulate damage until it becomes as

extensive as higher friction materials tested for shorter durations (fewer cycles of loading). Only

a few investigations of fatigue damage in articular cartilage have been reported to date [96-100].

Figure 3: Images of cartilage samples after migration against a flat glass slide under SCA config. for 8 hours. The superficial zone has swollen and sheared off, indicating fatigue failure of cartilage MZ. (Oungoulian, 2015, reproduced with permission).

This observation of delamination at the SZ-MZ interface is consistent with recently

reported findings that this region exhibits the lowest shear modulus across the full thickness of

the articular layer, in bovine and human cartilage [101-103]. Thus, this region is subjected to

the greatest amount of reciprocal shear strain under the action of frictional shear, causing

fatigue failure. It is well established that the collagen matrix in the middle zone exhibits a more

14

random orientation than the highly organized superficial zone fibrils (oriented tangential to the

articular surface) and deep zone fibrils (oriented perpendicular to the cartilage–bone interface)

[104, 105]. This collagen ultrastructure is a likely reason that the middle zone exhibits the least

resistance to shear and fails under reciprocal sliding.

These results point to a delamination mechanism [83], however there are questions of

the physiologic relevance of those results since the experiment was designed to simulate a

hemiarthroplastic condition, and thus the wear tests were performed using the stationary contact

area configuration (SCA), which intentionally accelerates the tissue decrease in fluid pressure

(Section 1.5) (Figure 1A). It could be argued that the elevated friction coefficient achieved

under those conditions would not occur under a more physiological migrating contact area

lconfiguration (MCA) (Figure 1B). Another concern may be edge effects since the cylindrical

plugs were cut out of the native connected cartilage surface and exposure of the cut edge may

be causing the observed delamination. Therefore, subsequent frictional wear tests (presented

in this Chapter 2 of this dissertation) were designed to use cartilage strips, instead of cylindrical

plugs, under the MCA instead of SCA (Section 1.5).

1.9. Live Cartilage Explant Studies

Articular cartilage is metabolically active, meaning that it is constantly remodeling its

environment via anabolic and catabolic activities. Anabolic events refer to synthesis,

organization, and assembly of the ECM and catabolic refers to the breakdown and loss of ECM

components. Chondrocytes control this process and strike a balance between the two, causing

the tissue to remain in a state of homeostasis, or maintenance of the ECM throughout an

organism’s life. However, things like injury or disease alter the chondrocyte environment and

they readily respond. Research has shown that chondrocytes are sensitive to soluble chemical

mediators (cytokines, signaling factors, etc.), and their mechanical environment (stresses,

15

strains, flow velocities, osmotic pressures, and electrical currents) [106-111]. Similarly,

explantation of cartilage tissue for its subsequent study in the lab stimulates a similar regulatory

response, making metabolically active cartilage tissue difficult to study in the lab under native,

homeostatic conditions. The biggest challenge is that it is difficult to recreate control conditions

that maintain the normal homeostatic behavior of articular cartilage explants in culture. Without

an adequate control, cartilage samples are subject to the dysregulation associated with being

removed from its in vivo environment. This fact, in addition to the dearth of studies on cartilage

fatigue damage, means that investigations of natural cartilage repair following episodes of

frictional loading are also very limited [112].

1.10. Preventive treatment over regenerative treatment

Over the past decade, improvements in orthopedic replacement implants have increased the

willingness of surgeons to perform partial and total joint replacement surgery on younger

patients. In 2014, the American Academy of Orthopedic Surgeons (AAOS) reported that from

2000-2009 the incidence of total knee replacement (TKR) had increased by 188% for patients

ages 45-64 [113]. This has been encouraging progress because the decision to undergo TKR

earlier can spare patients not only years of suffering while waiting for surgery, but also spares

patients those years of disability which have long-term negative impacts on health and future

mobility. However, a problem remains: artificial joint lifetimes are 20 years, at best, indicating a

future revision for a 45 year old at age 65, and yet another at age 85, in the best case scenario.

Therefore, despite increases in the rates of early TKR surgeries, surgeons will still exhaust all

alternatives first, reserving TKR as the last resort option. If arthroscopic chondroplasty and

microfracture have not been successful, cartilage restoration techniques may be considered.

However, osteochondral autografts, autologous chondrocyte implantation, and donor allografts

all are largely considered to be sub-optimal options due to: donor-site morbidity, requirement of

16

a second procedure, need for donor availability and limited shelf life, respectively [114, 115].

Since OA in younger patients (likely post-traumatic OA) may be localized to one of the three

compartments of the knee (medial, lateral, or patellofemoral), surgeons will often recommend a

unicompartmental replacement. This type of replacement is preferable to the TKR because it

replaces only the damaged compartment with metal and plastic leaving the healthy cartilage of

the remaining compartments intact – reserving the option for TKR down the line, if necessary. A

TKR following a unicompartmental replacement becomes necessary most typically due to

eventual failure of one of the other previously mildly arthritic compartments. Thus, at the time of

the partial replacement, a treatment modality that could strengthen the mildly arthritic tissue in

order to delay OA progression, would be incredibly valuable. One could envision such a

preventive treatment modality to be utilized at various stages throughout the patient cycle of

care: from the first arthroscopic procedure up to the partial joint replacement procedure. This

dissertation asserts that focusing on delaying or preventing wear by improving the resiliency of

the extant intact cartilage in early OA is a viable strategy to improve patient outcomes and offers

an innovative approach over existing regenerative techniques.

1.11. Hypotheses and Specific Aims

This dissertation aims to provide a basic science understanding of wear propagation and repair

of articular cartilage in the absence of traumatic events under the normal reciprocal sliding

motion of the articular layers at physiologic load magnitudes. Characterizing mechanical wear

in vitro enables an engineering approach to parsing out distinct modes of cartilage tissue failure.

In Chapter 2, articular cartilage is considered as a devitalized biomaterial, allowing for the

assessment of failure and mechanical wear propagation in the absence of cellular mediators.

Chapter 3 defines a homeostatic chemical environment in which future wear and repair studies

can be performed on living cartilage explants. In Chapter 4, we use the developed

17

methodologies of wear assessment and homeostatic culture conditions to test the effectiveness

of a novel treatment strategy to prevent or slow down cartilage wear progression.

The first hypothesis of this dissertation (H1a) is that cartilage damage may be initiated

by fatigue failure of the extracellular matrix (ECM) under the normal reciprocal sliding motion of

the articular layers, even when these layers are subject to physiologic load magnitudes and

contact stresses and even when the friction coefficient µ remains low. Based on prior literature

findings regarding the role of synovial fluid (SF) boundary lubricants on the reduction of friction

and wear [64, 116, 117], we also test the hypothesis (H1b) that SF delays the onset of cartilage

delamination when compared to physiological buffered saline (PBS). Based on our previously

reported model for the dependence of the frictional force on interstitial fluid load support and the

solid-on-solid contact area fraction [77, 118], we also test the hypothesis (H1c) that loading

cartilage against cartilage delays the onset of delamination wear compared to testing glass on

cartilage, since contacting porous cartilage layers exhibit a much smaller solid-on-solid contact

area fraction than impermeable glass contacting porous cartilage.

Next, we suspect that the homeostatic dysregulation previously observed in cultured

immature cartilage explants results from the presence of non-physiologic levels of important

metabolic mediators in the culture medium. To this end, we hypothesize that: (H2a) immature

bovine cartilage explants cultured in native synovial fluid will maintain homeostasis as

characterized by maintenance of their mechanical properties and ECM contents at initial (post-

explantation) levels, and (H2b) explants cultured in a physiologic-based medium, consisting of

physiologic levels of key metabolic mediators, will maintain a similar homeostasis over long-

term culture.

Finally, a laser treatment strategy is explored that has the capability to reform collagen

crosslinks, replacing those lost during OA progression. This novel therapy acts without injuring

the cells and without any chemical additive or thermal ablation. The laser treatment protocol

18

used in this application can specifically target the subsurface region, located 200 μm of the

articular surface. By strengthening this region with enhanced crosslinking, we hypothesized

(H3a) that cartilage will demonstrate greater resistance to fatigue failure than untreated controls.

We then hypothesized (H3b) that this treatment protocol would also be effective on devitalized

fibrillated human articular cartilage from OA joints with overall Outerbridge score OS1-3.

We tested these hypotheses in vitro with the following specific aims:

Specific Aim 1 (SA1): Develop a protocol for in vitro frictional sliding such that cartilage wear

can be observed and measured while maintaining physiologic contact stresses, loads, and

sliding speeds. Perform frictional wear experiments on devitalized immature bovine cartilage

strips to evaluate wear initiation of healthy tissue in the absence of cellular mediators.

Specific Aim 2 (SA2): Optimize tissue culture medium to maintain homeostasis of living juvenile

bovine cartilage explants in vitro for up to one month. Identify key biochemical components that

influence cellular metabolism and catabolism of components and the effect on tissue

biomechanical properties: swelling and stiffness. SA2 will establish a baseline from which

damage and repair studies on living tissue can be performed without the influence of the

confounding factors of ex vivo conditions.

Specific Aim 3 (SA3): Implement a protocol for selectively crosslinking articular cartilage

collagen via treatment with a femtosecond laser. Examine morphological, structural, and

functional modification of the cartilage ECM in both devitalized bovine and human tissue

explants. (a) Evaluate overall and local mechanical properties in addition to wear resistance of

laser-treated tissues compared against controls. (b) Repeat key experiments from SA3 (a) with

living immature bovine cartilage explants and culture for several weeks to determine cell viability

and permanency of laser treatment. (c) Scale-up treatment strategy to cover a clinically-

relevant (10 mm) patch on a devitalized cartilage strip and evaluate effectiveness of wear

19

prevention. SA3 introduces a novel early-stage OA treatment strategy, leveraging the damage

model to demonstrate effectiveness of the treatment at preventing or delaying cartilage wear.

20

Chapter 2 – Immature Bovine Cartilage Wear by Fatigue Failure and

Delamination

2.1. Introduction

It is well recognized that regardless of the causative factors of OA, cartilage stresses

produced by sliding contact of the articular layers mediates the progression of tissue

degeneration. The mechanisms by which stresses produce tissue degeneration via mechanical

failure remain poorly understood. OA is often described as a natural process of wear and tear

associated with aging, or an initiating traumatic event. The cartilage mechanics literature has

mostly focused on examining the friction coefficient µ as a surrogate for understanding wear and

tear in cartilage [119] (Section 1.5).

When cartilage is loaded against cartilage, it produces a migrating contact area (MCA)

configuration that sustains elevated interstitial fluid load support [120]. As a result, the friction

coefficient remains relatively low for sustained durations (Figure 4); for human cartilage, it is

typically µ»0.020 in synovial fluid (SF) and µ»0.025 in saline [60]. Implicitly, a low value of µ has

been assumed to produce low wear while an elevated value could lead to significant wear.

Despite the prominence of this hypothesized mechanism, it is only recently that cartilage wear

studies have started to be performed under controlled conditions, most notably by investigating

PRG4 knockout mice [117], since PRG4/lubricin has been shown to reduce the friction

coefficient of cartilage in vitro [64] and prevent degeneration in vivo [116]. Few studies have

investigated the ability of living cartilage to repair, following episodes of tissue wear [121].

Wear is a generally complex phenomenon that may manifest itself in different ways. In

the engineering tribology literature, a broad range of wear mechanisms are reported, many of

which are mostly applicable to metals and other artificial surfaces [79]. However, some of these

mechanisms may also be candidates for wear of biological tissues, such as abrasive wear,

21

which removes particulates of matter from the bearing surfaces; third-body wear, where

particulate matter causes further abrasion of the bearing surfaces; fatigue wear with

delamination, where the load-bearing material fails below the surface due to fatigue and the

failure propagates until a lamina shears off; and chemical wear, where breakdown of the

bearing material is initiated by chemical reactions, such as proteolysis in biological tissues. It is

unclear which of these mechanisms prevail in articular cartilage, and under what conditions.

Figure 4: Mean friction coefficient for a stationary contact area (SCA) configuration (cartilage plug against glass slide) and two migrating contact area (MCA) configurations: bovine knee condyle against tibia, and glass lens against tibia (n=6 per group) (Caligaris, 2008; reproduced with permission).

In our recent study on immature bovine cartilage [83], wear tests were performed on cartilage

plugs sliding against glass or various metals used in orthopaedic implants producing

delamination of the superficial zone with negligible abrasive wear. These results were

consistent with the fact that delamination is a clinically recognized symptom of OA [88, 89].

22

However, a potential limitation of that study was our adoption of a stationary contact area (SCA)

testing configuration, which promoted loss of interstitial fluid pressurization over time. It could

be argued that the elevated friction coefficient achieved under those conditions would not occur

under more physiological loading conditions. Furthermore, with prolonged wear testing,

complete delamination and removal of the top layer of these plugs was observed (Figure 3),

raising the possibility that the initiating failure resulted from edge effects between the flat

counterface material and the circular edge of the plug surface.

2.2. Hypotheses

In this study, we perform frictional sliding experiments on immature bovine cartilage to test our

primary hypothesis (H1) that delamination wear occurs even when the friction coefficient µ

remains low under a migrating contact area configuration (MCA). We use large, rectangular

cartilage strips harvested from the medial or lateral tibial plateau, loaded with a glass lens under

low physiological contact stresses, such that the contact area remains well within the strip

boundaries to avoid edge effects. Based on prior literature findings regarding the role of

synovial fluid (SF) boundary lubricants on the reduction of friction and wear, we also test the

hypothesis (H2) that SF delays the onset of cartilage delamination when compared to

physiological buffered saline (PBS). Based on our previously reported model for the

dependence of the frictional force on interstitial fluid load support and the solid-on-solid contact

area fraction, we also test the hypothesis (H3) that loading cartilage against cartilage delays the

onset of delamination wear compared to testing glass on cartilage, since contacting porous

cartilage layers exhibit a much smaller solid-on-solid contact area fraction (Section 1.6) than

impermeable glass contacting porous cartilage.

23

2.3. Materials and Methods

2.3.1 Experimental Design

Hypotheses H1-H3 were investigated in a series of friction experiments the we called Study 1.

The experimental set-up is shown schematically in Figure 1 and experimentally in Figure 5.

Seven test groups were included in Study 1, as summarized in Table 1 using either a glass lens

(G) or a condylar cartilage (C) counter-face (similar radius of curvature) sliding against a tibial

cartilage strip. The choices of load magnitudes magnitudes (L1 = 4.45 N or L4 = 17.8 N) and

testing durations (8 h or 24 h) were motivated by observations from preliminary studies.

Figure 5: MCA experimental testing configurations for Study 1 (top) condyle as counterface (bottom) glass lens as counterface.

24

Friction experiments were performed using three different friction-testing devices. Two

of these testers were identical (Tester A) and can apply loads up to 45 N; the other tester

(Tester B) can apply up to 220 N of contact load. Study 1 was performed using Tester A and

either a glass lens (G) or a femoral condyle plug (C) (similar radius of curvature) were slid

against a tibial cartilage strip (Figure 1B). Hypothesis H1 was tested by examining whether the

friction coefficient remained low for samples that exhibited delamination in any of the eight

groups. Hypothesis H2 was tested by comparing the incidence of delamination failure, as well

as the number of reciprocal sliding cycles to failure, in Group 1 versus Groups 2 (G-PBS-8-L1

versus G-SF-8-L1), Group 3 versus Group 4 (G-PBS-24-L1 versus G-SF-24-L1), and Group 2

versus Group 5 (G-PBS-8-L1 versus G-SF-8-L4). Hypothesis H3 was tested by examining the

incidence of delamination in Group 1 versus Group 6 (G-PBS-8-L1 versus C-PBS-8-L1), and in

Groups 2,4,5 versus Group 7 (G-SF-8-L1, G-SF-24-L1, G-SF-8-L4 versus C-SF-8-L1).

Group Counter-face Lubricant Duration Load Sample Size

G-PBS-8-L1 Glass lens PBS 8 h 4.45 N 12

G-SF-8-L1 Glass lens PBS 8 h 4.45 N 7

G-PBS-24-L1 Glass lens PBS 24 h 4.45 N 10

G-SF-24-L1 Glass lens SF 24 h 4.45 N 6

G-SF-8-L4 Glass lens SF 8 h 17.8 N 14

C-PBS-8-L1 Cartilage PBS 8 h 4.45 N 7

C-SF-8-L1 Cartilage SF 8 h 4.45 N 6

Table 1. List of test groups used in Study 1. For all groups, sliding was performed using the specified ‘Counterface’ against a cartilage strip, using the specified ‘Lubricant’ and contact ‘Load’, over the given ‘Duration’. The number of specimens used in each group is given under ‘Sample Size.’

25

Since no damage was observed for the first two cartilage-on-cartilage groups (C-PBS-8-

L1 and C-SF-8-L1), Study 2 was added (C-PBS-24-L22) to examine both the resilience of

cartilage-on-cartilage to wear under high loads and the nature of that damage. Preliminary

studies for cartilage-on-cartilage showed that increasing the testing duration to 24 h for the 4.45

N loading conditions did not lead to delamination and, due to the thick articular layer of the

femoral condyle counterface, increasing the load led to a dramatic increase in contact area over

the duration of the test causing unavoidable edge effects. Therefore, Study 2 was performed

using Tester B, which can apply up to 220 N of contact load, using an osteochondral

counterface from a different joint, the ankle. This counterface was selected for its small radius

of curvature (high curvature) and the thickness of the articular layer (~1 mm, compared to ~5

mm of the femoral condyle) (Figures 6 & 7).

Figure 6: Four different views of the ankle condyle counterface utilized in Study 2.

26

Whereas under 98 N the femoral condyle experiences a dramatic increase in contact

area over the duration of the test, the ankle condyle contact area stays relatively constant,

allowing higher stresses (2 MPa, well within physiologic range) to be imparted to the strip

throughout the duration of the test (Figures 7 & 8). PBS was selected as the lubricant and a 98

N load was applied over a 24 h testing duration. Most tests ran to completion, but some tests

were stopped early when gross damage was evident.

Figure 7: Cross-section views of the “popped” cartilage-only immature bovine condyles (A) femoral condyle (Oungoulian, 2014, reproduced with permission) (B) ankle condyle. Biopsy punches from (A) were used as the counterface for C-SF-8-L1 and C-PBS-8-L1 of Study 1. The full “unpopped” bone+cartilage sample from (B) was used as the counterface in Study 2.

27

An additional subset of cartilage strips was used to examine the progression of damage

using the experimental conditions of GSPBS-8-L1. These tests were used to validate a method

of real-time damage detection described in Section 2.3.8. below.

Figure 8: Pressure-Senstive Fuijifilm contact area for representative groups. A femoral condyle is shown at 45 N to demonstrate the increase in contact area relative to ankle, which is significantly smaller under 25 lbs.

2.3.2. Specimen Preparation

Full-thickness cartilage strips (10 mm ´ 30 mm) were harvested from the tibial plateau of knee

joints of immature bovine calves (2-3 months old). For cartilage-on-cartilage groups C-PBS-8-

L1 and C-SF-8-L1, full-thickness cylindrical plugs of large diameter (Æ15.875 mm) were

harvested from the opposing femoral condyle. For cartilage-on-cartilage group C-PBS-24-L22,

ankle condyle osteochondral samples (Figure 6) were collected from the same leg via band-

saw, frozen for no more than 7 days prior to testing and mounted for testing with no further

preparation. Experiments were performed on tissue collected from 41 bovine stifle joints (a

maximum of three strips were collected per knee). Cartilage strips and cartilage-only condylar

plugs were placed articular surface down on a freezing stage sledge microtome (Leica

Instruments #SM2400, Nusslock, Germany), embedded and frozen in a water-soluble

28

sectioning gel (Thermo-Fisher Scientific #1310APD, Rockford, IL). Specimens were trimmed to

remove bone, in order to obtain cartilage-only specimens with an intact superficial zone. Strips

were cut to a final thickness of 1.4 mm ± 0.2mm, and plugs were cut to a final thickness (from

the condyle to the microtomed surface) of 10 mm. Samples were rinsed and stored in a 0.22

µm filtered phosphate-buffered saline (PBS) solution supplemented with an isothiazolone-based

biocide (Proclin 950, Sigma-Aldrich #46878-U, St. Louis, MO) and protease inhibitor (0.05 M

Ethylenediaminetetraacetic acid, EDTA) and frozen at −20 °C for no more than 3 weeks prior to

testing. On the day of testing, samples were thawed, aligned, and affixed with a cyanoacrylate

adhesive to their respective stages.

2.3.3 Testing Devices

Study 1 was performed on two custom testing devices of identical design, adapted from [83]

(Tester A, operating load range 1 N – 45 N), while Study 2 employed a different custom testing

device adapted from [122] (Tester B, operating load range 4.5 N – 220 N). Driven under load

control, both testers introduced sliding contact between the two apposing surfaces. Tester A

imparted pure motorized translation to produce sliding between the counter-face and the strip

whereas Tester B imparted motorized rotation on the counter-face, about an axis located a large

radius away, producing negligible rolling and predominant sliding at the contact interface. All

devices were controlled with custom LabVIEW software (National Instruments Corporation

#LabVIEW 2010, Austin, TX) and associated motion controllers (National Instruments

Corporation #7354, Austin, TX) to provide continuous measures of relevant displacements,

applied loads and reaction loads.

Tester A: Two identical two-axis frictional sliding devices were used; each of these devices were

equipped with a six-axis load cell (JR3 Inc. #20E12A4, Woodland, CA) mounted on a stepper

motor driven (Orientalmotor, Vexta line #PK266-03B) horizontal translation stage (JMAR XY

29

motorized linear stage) equipped with linear encoder (RSF Electronics MSA 65x series, 5-10 µm

resolution) to record the horizontal displacement of the stage (, and a voice-coil actuated

(H2Wtechnologies #NCC20-15-027-1RC), vertically-oriented loading platen equipped linear

encoder (Renishaw #T1031-30A) to record the vertical displacement () of the loading platen.

Reciprocal sliding motion was prescribed at 1 mm/s over a range ±5 mm, producing a total

travel path of 10 mm per reciprocal cycle. Normal loads (4.45 – 17.8 N) were prescribed with

the voice-coil actuator operated under a feedback control loop. The test duration was either 8 h

or 24 h. Thus, the maximum number of reciprocal cycles of sliding was 2880 for the 8 h test,

and 8640 for the 24 h test. For experiments where a plano-convex glass lens was used as the

counter-face, the lens radius of curvature (R12.7 mm) was selected to best match that of the

femoral condylar plugs (R13.0 ± 1.6 mm as measured from laser scans).

Tester B: In this rotary testing apparatus, cartilage strips were mounted over a 6-axis load cell

(JR3 Inc. #45E15A4), while the counter-face was mounted on the end of a rotating pendulum

arm (length = 145.7 mm + sample thickness), actuated by a rotary stepper motor (Orientalmotor

#ARM98AC-N25). The load cell was secured to a stepper motor-actuated vertical translation

stage (Maxon Motor #241410), which moved up and down throughout the stroke of the

pendulum to maintain the prescribed load via a feedback control loop. The angle of rotation of

the pendulum arm was prescribed as to yield a ±5 mm travel distance on the strip, with a

relative sliding speed of 1 mm/s.

2.3.4. Contact Area and Stresses

To estimate the average contact stresses under the applied loads, contact areas were

measured using pressure-sensitive film (Fujifilm, prescale 4LW; 0.05 – 0.2 MPa) (Figure 8).

The average contact area between the glass lens and cartilage strip was 8.7 mm2 after 60 s of

loading under 4.45 N and 19.9 mm2 under 17.8 N, equivalent to average contact stresses of

30

0.52 MPa and 0.89 MPa, respectively. The average contact area between the femoral condylar

plugs and cartilage strip varied in the range from 10.6 – 21.3 mm2 for different samples after 60

s of loading under 4.45 N (depending on the exact location from which the plug was harvested

along the curvature of the condyle (Figure 7), yielding average contact stresses in the range

0.44 – 0.21 MPa. The average contact area between ankle condyle plugs and cartilage strips

varied from 43.1 – 57.5 mm2 after 60 s of loading under 98 N for different samples (yielding

average contact stresses within the range from 1.93 – 2.58 MPa) (Figure 8).

2.3.5. Friction Measurements

For both types of testers, the cartilage strip was placed in a bath mounted on top of the

multiaxial load cell and the axes of the load cell were aligned with the direction of relative motion

and the normal direction to the articular surface. An encoder recorded the relative tangential

displacement (!" ) of the contacting surfaces and this information was used to detect the

beginning of each new cycle of reciprocal sliding. The load cell components tangential (#") and

normal (#$ ) to the articular surface were recorded. A point-wise friction coefficient % was

calculated at each position !" using % = (##" − #%")/(##$ + #%$), where + and – refer to values

of these forces on the forward and reverse strokes of each cycle. This type of calculation

minimizes the effects of slight alignment errors and fluctuations in strip surface slope over its

length. For the purpose of plotting the friction coefficient over the 8 h or 24 h duration of a test,

% was averaged over each cycle and plotted against the mean value of time over that entire

cycle. This time-dependent cycle-averaged value was called%. Individual cycle-averaged % for

representative samples over time can be seen in Figure 9. Results report: % (the average % for

the first hour of a test), % (the average % for the last hour of a test of any length), % (the average

for the entire test), and %ratio (the average % for the last hour / the average % for the first hour).

31

Figure 9: Representative temporal evolution of the friction coefficient µ from each 8 hour group (G-PBS-8-L1, G-SF-8-L1, C-PBS-8-L1, C-SF-8-L1).

2.3.6. Laser Scans and Surface Roughness

At the completion of sliding tests, cartilage strips and condylar plugs were unloaded and allowed

to recover in PBS (at 4 °C) for no less than the testing duration. The articular surface of each

sample was then laser scanned (NextEngine Inc., Santa Monica, CA) with a rated accuracy of

125 µm, producing a dense point cloud (160 points/cm). The 3D coordinates of the articular

surface of the strip were fitted to a plane while those of the condylar plugs were fitted to a

sphere. The root-mean-square deviation between the measured 3D coordinates and the fitted

surface was used as a measure of surface roughness (Figure 11A,C).

2.3.7. Polarized Light Microscopy and Histological Staining

After the laser scans, cartilage strips were lifted from their plate and bisected through the travel

path: half of the strip was frozen for later imaging via polarized light microscopy (PLM) and the

other half of the strip was fixed in 70% acid formalin for 24 h and stored in 70% ethanol for

32

histological processing. Samples for histology (n=4 per group), were embedded in paraffin wax

and 5 µm thick sections were obtained. Sections were stained with both 0.1% Safranin-O + 1%

Fast Green for GAG and 1% Picrosirius Red for collagen. Samples reserved for PLM were

sectioned to 120 µm thick on a freezing stage sledge microtome. These unfixed sections were

mounted on glass slides and imaged under polarized light, allowing visualization of collagen

alignment (Olympus BX60 microscope, drop-in U-POT polarizer, Olympus DP72 camera) [123].

2.3.8. Detecting Onset of Delamination

In a post-hoc analysis of the data from Study 1, several parameters were explored for

use in detecting real-time fatigue damage (% vs. time, % vs. !", #" vs. time, and #" vs. !").

Principal component analysis (PCA) of hysteresis loops plotting #" vs. !" proved to be the most

sensitive and reliable indicator of the onset of delamination wear. Families of hysteresis loops

were analyzed using PCA on a per cycle basis, by computing the covariance matrix of #" versus

!" data and calculating its eigenvalues and eigenvectors. The first eigenvalue was determined

to be the most sensitive to hysteresis loop changes over time, therefore magnitudes of the first

eigenvalues were recorded for all tests. During the post-hoc analysis of Study 1, a threshold

was determined to identify the onset of damage and extract the number & of cycles to damage.

This threshold was determined by calculating a baseline initial eigenvalue at the beginning of

the test; all subsequent eigenvalues were compared against this initial value. The cycle at

which the current value exceeded 20% of the initial value was determined to be the failure cycle.

The initial eigenvalue was determined by taking the average eigenvalue from cycle 10 through

5% of the total cycles in that test (for an 8 hour test, cycles 10 – 144 were averaged; for a 24

hour test, cycles 10 through 432 were averaged). This “training” phase for the program used

data from 32 damaged strips. For all strips in Study 1, the PCA-based diagnostic of

delamination was verified against direct visualization and laser scan of the cartilage strip at the

33

completion of the test; discrepancies were noted as ‘PCA false positive’ (there were no false

negatives) and the direct visualization assessment was used as the main indicator of damage.

To validate this method against direct visualization of damage, a subset of sliding experiments

was halted mid-test, either at & or at subsequent cycles, & +', using real-time visualization of

the hysteresis data. Strips were sectioned (120 µm thick) and imaged under polarized light

microscopy (PLM) to assess damage. For all other strips, the PCA-based diagnostic of

delamination was verified against direct visualization and laser scan of the cartilage strip at the

completion of the test; discrepancies were noted as ‘PCA false positive’ (there were no false

negatives) and the direct visualization assessment was selected as the standard.

2.3.9. Statistics

Unpaired t-tests were performed to examine the role of lubricant, damage, and counterface on

%-ratio. Unpaired t-tests were also performed to examine the role of damage on %, % and % data.

2.4. Results

2.4.1. Damage Morphology:

Based on 3D scans (Figure 11A,C), RMS roughness data (Table 2), and visual inspection

(Figure 11B,D), significant surface damage was observed in G-PBS-8-L1, G-PBS-24-L1, G-SF-

24-L1 and G-SF-8-L4 whereas G-SF-8-L1, C-PBS-8-L1, C-PBS-8-L1 exhibited no evident

damage. Study 2 was analyzed by visual inspection and it was found that 6 out of 10 samples

had been damaged over duration of the test. All samples that were damaged had signs of

failure by delamination wear that manifested in the form of a blister (Figure 11). For some

samples this blister ruptured by the end of the test, and for other samples it did not (Figure 14).

For group 8 (C-PBS-24-L22), 6 out of 10 samples showed similar ruptured-blister damage, while

the remaining 4 out of 10 showed either no damage, or incomplete relaxation after 24 hours

(Figure S1) prior to storing/freezing. One extra ankle-strip pair (same parameters as Study 2)

34

was friction tested and interrupted after 5 hours for visual inspection. This strip had an intact

blister and is further evidence of the initiation of delamination wear for the cartilage-on-cartilage

configuration (Figure 19). For all damaged groups, PLM consistently reinforced that damage

occurred as a result of delamination at the interface between the superficial zone, SZ, and

middle zone, MZ (Fig.13, 19, S3). The observed delamination was only apparent in the contact

region of the strip, whereas the unloaded region of the strip remained intact (Fig. 12 and 19).

Group Counterface Lubricant RMS, damage [mm] %-()) - overall [ ]

G-PBS-8-L1

Glass lens

PBS 0.25 ± 0.22 0.022 ± 0.006

G-SF-8-L1 SF 0.06 ± 0.03 0.011 ± 0.001

G-PBS-24-L1 PBS 0.32 ± 0.10 0.020 ± 0.005

G-SF-24-L1 SF 0.27 ± 0.21 0.015 ± 0.005

G-SF-8-L4 SF 0.35 ± 0.11 0.014 ± 0.002

C-PBS-8-L1

Cartilage

PBS 0.04 ± 0.01 0.018 ± 0.011

C-SF-8-L1 SF 0.06 ± 0.01 0.006 ± 0.001

Ankle PBS N/A 0.058 ± 0.017

Table 2: Average friction coefficient (over the entire test duration) and measurement of damage for all groups.

When visual damage to the strip was observed, hysteresis loops progressively rotated

and elongated, with corresponding increases in magnitude of associated eigenvalues (Figure

12). Experimental conditions that did not result in visual damage to the cartilage strip generated

hysteresis curves that maintained the same unchanged hysteresis loop over the course of the

test (Figure 20, 21). The hysteresis program was built based on inspection of the hysteresis

loops of both the damaged and undamaged samples that comprise Study 1. During testing of

the program, thresholding was deemed acceptable when samples were appropriately sorted

35

into their respective damage/no damage category (based on visual inspection of the strip after

the test). These results for the determination of N can be found in Table 5. Good agreement

was found between observed damage and predicted damage. There were six samples that did

not have good agreement - the program predicted damage, but none was observed. These are

listed in the table under false positives.

Figure 10: Average µeff for samples from Study 1. Data binned by incidence of damage. P-values for comparisons from left to right: 0.0132, 0.0147, 0.003.

Since this was a destructive test, six samples were used (the PLM images shown in

Figure 13 are of several samples, each with a distinct value of N). Figure 13 depicts

corresponding snapshots of damage progression, showing clear delamination of the MZ and

SZ. Strip B in Figure 13 was stopped at cycle N, and though this strip showed no macroscopic

signs of damage, PLM revealed a jagged an atypical MZ-DZ transition within the travel path.

2.4.2. Friction Coefficient (H1)

The cycle-averaged friction coefficient averaged over the entire duration of the test % is shown

in Table 2. It can be seen that % remains low for all groups. Regardless of the incidence of

36

damage, no individual sample had a friction coefficient that exceeded % = 0.03. On close

inspection of the data, Study 1, that experienced damage (GPBS-8-L1, G-PBS-24-L1, G-SF-24-

L1, G-SF-8-L4), the average% (first, last, and overall) was observed to be statistically greater

when compared against the binned average% (first, last, and overall) of the undamaged groups

(G-SF-8-L1, C-PBS-8-L1, C-SF-8-L1) (Figure 10). The % ratio (% –last /% –first) was also

examined. It was shown that the choice of counterface had no statistical effect on% ratio

whereas both the incidence of damage and the use of PBS increased the% –ratio significantly.

The effect of lubricant had a more significant effect on% –ratio and, interestingly, SF had% ratio

< 1, indicating that on average,% –last <% –first, regardless of incidence of damage (Figure 15).

Figure 11: Post-test 3D scans (A,C) and photographs of representative cartilage strips (B,D). A glass lens was migrated against the cartilage strip in PBS (A,B) and SF (C,D). When PBS was the lubricant the strip exhibited significant damage, in the form of a blister while strips migrated in SF showed no evidence of damage. Scale bar: 5 mm.

37

Figure 12: Polarized light microscopy depicting damage onset and damage progression in cross-sections taken through the travel paths of six strips halted mid-test at spaced intervals (N+M): (A) no damage (Fz = 4.5 N, SF, 2800 cycles) (B-F) damage (Fz = 4.5 N in PBS). Illuminated regions indicate collagen fiber alignment (SZ, surface zone; DZ, deep zone); dark regions indicate random collagen organization (MZ, middle zone).

2.4.3. Boundary Lubrication (H2)

Hypothesis 2 tests the role of the boundary lubricant SF in delaying the onset of wear. This was

done by comparing the incidence of delamination failure as well as the number of reciprocal

38

Figure 13: Hysteresis curves for four strips (A,B,D,F; Figure 12) depicting two cycles: cycle 10 (baseline) and the cycle at which the test was halted (final cycle). The vectors represent the first principal eigenvectors of the final cycle which exceeded the threshold and determined N.

sliding cycles to failure in G-PBS-8-L1 versus G-SF-8-L1, G-PBS-8-L1 versus G-SF-8-L4, and

G-PBS-24-L1 versus G-SF-24-L1. For the first comparison, the only difference between the two

groups was the lubricant. At the end of testing, G-SF-8-L1 had 0 out of 7 strips with damage

and 1 false positive, whereas G-PBS-8-L1 had 11 out of 12 damaged strips, 1 undamaged, and

1 false positive. For the next comparison, the same conditions are used as in the previous,

except that the number of cycles was increased (8 hr = 2,880 cycles to 24 hr = 8,640 cycles).

Again, the only difference between G-PBS-24-L1 and G-SF-24-L1 is the presence of SF. This

time, while Group 3 had 10 out of 10 samples get damaged with no false positives, G-SF-24-L1

39

had 4 out of 6 get damaged, with 2 false positives. For the next comparison, G-SF-8-L4

experienced wear and was compared against the undamaged G-SF-8-L1 from the first

comparison. The only difference between these groups was the applied load – which was 4X

higher in G-SF-8-L4 than that applied in G-SF-8-L1. G-SF-8-L4 had 13 strips damaged, 1

undamaged, and 1 false positive as compared to the 7 undamaged strips from G-SF8-L1.

Figure 14: Picrosirius Red stained histological sections of strips with (A) no damage (B) blister (C) ruptured blister. Arrow indicates travel path.

2.4.4. Frictional Shear S-N Curve

For samples that exhibited delamination, in an effort to identify the measured parameter most predictive of the number of cycles to failure, we examined the correlation coefficient between various measures (#", #$, #$/*, #"/*, %()), ℎ, max|!$|) and & (where max|!$| is the maximum

creep deformation of the contacting materials). The only measure that produced a statistically significant correlation was the average tangential shear stress + , estimated from cycles preceding the onset of delamination (cycle 10 to cycle &− 1), as the ratio of the tangential force

40

%())#$ and contact area * (R2 = 0.33) (Figure 16). This measure was used to produce an

experimental +− & curve for fatigue failure from frictional shear.

Figure 15: Friction coefficient ratio (last hour/ first hour). Binned by different faactors. p-value from left to right: 0.0428, 0.2547, 8.2E-2

2.4.5. The effect of decreasing solid-on-solid contact area fraction (cartilage as counterface, H3)

Hypothesis 3 tests the role of fluid load support in delaying the onset of delamination wear. This

was achieved by comparing G-PBS-8-L1 versus C-PBS-8-L1, and G-SF-8-L1, G-SF-24-L1, G-

SF-8-L4 versus C-SF-8-L1. For the first comparison the only difference between the two groups

is the type of counterface. At the end of testing, G-PBS-8-L1 had 11 out of 12 damaged strips,

1 undamaged, and 1 false positive, whereas C-PBS-8-L1 had 0 out 7 damaged strips with 0

false positives, highlighting the clear advantage of using cartilage as a counterface to prevent

wear. Interestingly, no statistical difference was found between µ for these two groups (µ =

0.022 ± 0.006 vs. µ = 0.018 ± 0.011; p=0.176). For the next comparison, the effect of fluid load

support was then examined in presence of synovial fluid. Two of the groups have identical

conditions and only the type of counterface is varied (G-SF-8-L1 versus C-SF-8-L1). Since

neither group experiences damage, the more extreme conditions are explored (G-SF-24-L1, G-

SF-8-L4 versus C-SF-8-L1; G-SF-24-L1, G-SF-8-L4 versus C-SF-8-L1). As is known from

exploring H2, G-SF-24-L1 had 4 out of 6 strips experience damage within 24 hours at L1, with 2

false positives and G-SF-8-L4 had 13 strips experience damage, 1 undamaged, and 1 false

positive within 8 hours at L4. C-SF-8-L1, on the other hand, had 0 out of 6 strips fail for 8 hours

41

at L1. Anecdotally, several cartilage-strip pairs were tested for the incidence of damage on this

device with parameters ranging from 4.5-40 N and 24-120 hours. Cartilage damage was not

observed in any of these cases, even for the most extreme case of C-SF-120-L10 (n=1),

(SF+antibiotics refreshed daily).

Figure 16: Fatigue curve (S-N) correlating frictional shear stress with cycle to damage. (Top) Data for thirty-two damaged strips is shown, all from groups using Glass as the counterface. Frictional shear stress, S, is calculated from measured frictional traction divided by the contact area measured at the beginning of the test.

2.4.6. Increasing applied stress by decreasing contact area

Hysteresis loops and µ time course are shown for 4 samples. In Figures 20, 21 (two damaged

and two undamaged). As per Table 3, 6 out of 10 strips failed by apparent delamination wear

42

within 24 hours whereas the other 4 did not show signs of damage beyond post-test regions of

high depressurization (Figure S1). One additional sample was tested and halted so

delamination onset could be visualized. The test was stopped based on the appearance of the

strip and an intact blister was observed (Figure 19). After relaxation, PLM was subsequently

performed for sub-surface characterization of the blister (Figure 19).

Group n Damaged Undamaged False

Positives

N (failure cycle)

G-PBS-8-L1 12 11 1 1 552 ± 370

G-SF-8-L1 7 0 7 2 N/A

G-PBS-24-L1 10 10 0 0 1515 ± 1280

G-SF-24-L1 6 4 2 2 2925 ± 1928

G-SF-8-L4 14 13 1 1 489 ± 414

C-PBS-8-L1 7 0 7 0 N/A

C-SF-8-L1 6 0 6 0 N/A

Table 3. Results of test groups used in Study 1. The number of specimens used in each group is given under ‘n’. The number of strips that were morphologically damaged and undamaged as determined by observation and measures of surface roughness changes under the given testing configuration are listed in those respective columns. The number of samples that had disagreement between observed damage and predicted damage by hysteresis examination are listed under false positives. The column N, is the cycle at which initiated failure occurred as predicted in post-hoc interrogation of hysteresis curves.

2.5. Discussion

This study provides strong evidence that cartilage damage (or “wear and tear”) under a

migrating contact area may be recreated in the laboratory at physiological levels of contact

stresses (typically 0-6 MPa). The observed damage occurred as a result of delamination of the

SZ at its interface with the MZ, a location known to exhibit the lowest shear modulus across the

43

articular layer thickness [102]. For all samples in Study 1, damaged or undamaged, the average

% was no greater than 0.02 (Table 3). This finding is consistent with previous reports of MCA

friction coefficients for both healthy and OA human tissue [60], confirming that an MCA was

maintained over the course of these tests even up to 24 h. Taken together, these two findings

allow us to accept H1 – that delamination wear occurs even when the friction coefficient %

remains low under a migrating contact area configuration (MCA).

Figure 19: Cartilage strip from Group 8 (cartilage-on-cartilage) stopped mid-test. Blister indicative of delamination is apparent (white arrows)

On close inspection, there are statistically significant differences between %, %, % when

Study 1 data is binned by incidence of damage (Figure 10). This is in agreement with our

findings that although % is insufficient to predict the onset of damage, damage is not observed

completely independently of changes to %. The literature has mostly focused on examining the

friction coefficient % as a surrogate for understanding wear and tear in cartilage [119]. Implicitly,

a low value of % has been assumed to produce low wear while an elevated value could lead to

significant wear; however, a specific wear-producing threshold value for % has never been

proposed. These results paint a more complex picture of the relationship between % and

damage. This complexity is better understood when the friction profile of an individual test is

plotted with time (Figure S2). In several instances, % was on the low end of the spectrum (% =

44

0.015) constantly throughout the test, despite damage. In other instances a damage event

temporarily increased % for several hours, but initial low friction coefficient values were

recovered by the end of the test. Therefore, shifting the focus away from abnormally elevated

values of % may help parse out the many of the mechanical factors at play and help decompose

the process that leads to wear and tear in primary and secondary OA.

Figure 20: Two damaged samples from Study 2 with friction coefficients plotted and hysteresis loops for select cycles.

As evidenced by this work, fatigue damage occurs in cartilage subjected to sliding

friction under low-to-moderate physiological contact stresses, a finding most relevant to

understanding mechanical factors in OA. Fatigue failure is made readily apparent because of

the clear distinction that PLM provides to distinguish between the articular cartilage zones. Both

the SZ and DZ (deep zone) appear bright white under PLM due to the highly organized collagen

45

fibrils present in those zones. The MZ, instead, appears as a dark band, sandwiched between

the bright SZ and DZ. The MZ appears black due to the random organization of the fibrils in

that zone.

Figure 21: Two undamaged samples from Study 2 with friction coefficients plotted and hysteresis loops for select cycles. Images of strips were taken immediately after unloading, therefore the travel path is evident.

Delamination is indicative of sub-surface fatigue failure, resulting from the reciprocal

shearing of this SZ-MZ interface due to the tangential frictional shear stress, S, acting at the

articular surface. Based on our previously validated friction model for articular cartilage [120,

124], S is dependent on the magnitude of interstitial fluid load support and the presence of

boundary lubricants. Those two parameters were therefore experimentally varied in Study 1:

Glass compared with cartilage to explore the role of fluid load support and SF compared with

PBS to explore the role of boundary lubricants at preventing delamination wear. Since Study 2

46

was tested on a different loading device at a much higher load, direct statistical comparisons

were not made with any of the other presented groups. Study 2 instead serves as validation

that delamination by fatigue failure is a relevant failure mechanism for the cartilage-on-cartilage

configuration and that this type of wear can be recreated in the lab under the appropriate

physiologic experimental conditions and not an artifact of using a glass lens as a counterface.

The comparison made to test H2 looks at experimental G-PBS-8-L1 and G-SF-8-L1,

which differed only in that Group 2 friction tests were performed in SF. Lubrication by synovial

fluid had a clear effect on the incidence of damage, resulting in 0 out of 10 strips with damage

for the SF group and 10 out of 10 strips with damage for the PBS group. These results allow us

to accept H2, and further experiments were performed to test the limits of SF wear-resistant

benefits. In the next two comparisons (G-SF-8-L1 vs. G-SF-8-L4 and G-PBS-24-L1 vs. G-SF-

24-L1), the results indicate that while SF prevents wear for up to 8 hours at the lowest load

tested, this benefit can be overcome by both increasing the number of cycles or increasing the

applied load. Examination of the average failure cycles for the 24 hour lubricant comparison (G-

PBS-24-L1 vs. G-SF-24-L1) supports this conclusion (Table 3). Strips from G-PBS-24-L1,

tested in PBS, were computed to have N =1515 ± 1280, where strips from G-SF-24-L1, tested

identically in SF, required a significantly higher N =2925 ± 1928 (p = 0.006). These results

indicate that under these conditions, boundary lubricants delay the onset of delamination wear

by increasing the failure cycle by a factor of 2. The prior literature has shown lower % for SF

versus saline, and PRG-4 knockout mice show cartilage degeneration at skeletal maturity

compared to wild-type. This result demonstrates that SF reduces wear and that there is a limit

to such benefits.

In addition, the finding that the effect of lubricant had a more significant effect on% ratio

than any other factor (counterface or incidence of damage) is interesting. For groups that had

SF as the lubricant, it was shown that on average that% –last <% –first, regardless of the

47

incidence of damage. This is in agreement with our previous findings where the friction

coefficient of moderate-OA human cartilage was shown to be significantly lower than the friction

coefficient of mild-OA cartilage in PBS [60].

Hypothesis 3 tested that the cartilage-on-cartilage loading configuration delays the onset

of delamination wear compared to testing glass-on-cartilage. We accept Hypothesis H3 on the

ground that cartilage-on-cartilage damage was not obtained in Study 1, whereas experiments

using the same loading conditions with a glass lens as the counterface did experience damage.

We posit that delayed wear for cartilage-on-cartilage occurs because contacting porous

cartilage layers exhibit a much smaller solid-on-solid contact area fraction than impermeable

glass contacting porous cartilage. A higher solid-on-solid contact area fraction implies that a

higher proportion of ECM components of the SZ experience direct frictional interaction with the

sliding glass platen, causing greater transmission of the applied sliding contact load to deeper

tissue regions, namely the vulnerable middle zone. While this is our working theory, of course

there are other differences between a glass lens and a femoral condyle of similar curvature.

The ability of the condyle to deform, and spread the contact area, means that for the same

applied load, the glass lens is necessarily applying more contact stress (and thus, more

frictional stress, since % is very similar in all the test groups) over the test duration. Beneficial

biomolecular interactions across the cartilage-cartilage interface may also play a role, especially

in the presence of the boundary lubricants of synovial fluid.

Study 2 was added to verify that (1) migration under a physiologic MCA can cause

damage for cartilage-on-cartilage under the appropriate physiological loading conditions and (2)

the mode of observed damage would be similar to that of glass-on-cartilage (fatigue failure and

delamination). We carefully selected the condyle from the ankle joint, because of its high

curvature and anatomically thin articular layer (~1 mm, Figure 6), which minimized spreading of

the contact area, causing stresses to remain elevated (~2 MPa) throughout the test duration, yet

48

remaining in the low-to-moderate range of contact stresses. Performing ten 24 hour tests in

PBS under a 98 N load for this ankle-strip configuration confirmed that loading in vitro

parameters could be tuned to impart delamination wear under a physiologic loading

configuration. A final ankle-strip pair was halted mid-test to examine if the onset of delamination

wear could be visualized. This strip indeed captured the onset of a blister, providing further

evidence of the initiation of delamination wear for the cartilage-on-cartilage configuration (Figure

19). While the mode of failure is the same, there is one observed difference between the

damage observed for the ankle-strip and the damage observed for glass-on-strip. For glass-on-

strip the blister manifests directly under the center of travel of the glass counterface. For ankle-

on-strip, instead the blister was observed to originate near the outskirts of the travel path (Figure

19). This may be due to the dissimilar contact areas and contact pressures between the two

types of tests - with glass having a small (9 mm2), rigid, hemispherical migrating contact area,

and the ankle having a wide (50 mm2), more congruent, non-rigid, migrating contact area.

Our data point to delamination wear (instead of abrasive wear) as a primary damage

mechanism in cartilage. The presence of delamination is a strong indication that wear is

initiated by fatigue damage, caused by a combination of the tangential force #", and the number

of loading cycles N. The ability to accurately identify N for various cartilage loading

configurations allows for improved fatigue failure characterization; the formation +-N curves that

could serve as tools for predicting fatigue-life estimates (Figure 16). We examined the

correlation coefficient between various measures (#", #$ , #$/*, #"/*, %()), ℎ, max!$)and&

(where, max|!$| is the maximum creep deformation of the contacting materials). The only

measure that produced a statistically significant correlation was the average tangential shear

stress +. The + -N curve is only valid for those samples that experienced damage and have a

computed value for N. The figure showing the S-N curve (Figure 16) also displays the value of

49

S for undamaged samples, at the highest tested cycle number. These results show that the

great majority of these samples have a low S, falling beneath the S-N curve.

Another consideration of this investigation was the appropriateness of using the SCA

model to accelerate wear in vitro. We had questioned whether the delamination observed in a

previous study from our laboratory was an artifact of the experimental set-up [83]. One major

conclusion from the current study is the confirmation that delamination wear occurs also under

MCA, for both glass-on-cartilage and for cartilage-on-cartilage. Not only is this interesting as it

applies to how we understand the manifestation of mechanical wear in OA, but it also lends

credibility to the future use of the SCA model for investigating accelerated in vitro wear on

smaller tissue samples. The experiments performed in this study were long (between 8 hours

and 24 hours) and required large cartilage strips; the SCA model provides a more efficient

means to study wear and repair in the laboratory.

Another important outcome of this investigation is that from post-hoc loading data from

Study 1, we identified a method by which to detect damage onset in real-time. Using this

method we were able to identify and evaluate wear milestones (i.e. pre-wear, wear initiation,

delamination onset, initial separation, failure by delamination) (Figure 12,13). The damage

progression was predicted using our method of hysteresis curve examination (Figure 12). This

result demonstrated that variations in principal components of hysteresis curves could identify

physical disruption of the tissue. At the earliest indication of damage (at N), strip B had no

visible damage (Figure 13B). This suggests that PCA of the frictional hysteresis curves detects

not only visual tissue damage, but early mechanical breakdown of the sub-surface collagen

network prior to the onset of macroscopic damage, highlighting the sensitivity of this method of

determining N. On close examination of the strips (Figure 13), there is an intensified bright

signal in the fatigued MZ coincident with dark spindles extending into the deep zone, perhaps

50

indicating the recruitment of fibers from the highly organized DZ to resist shearing and swelling

of the MZ.

One limitation of this study and an important consideration for future work is the inability

to track contact area changes over the testing duration. Contact area measurements were

performed using pressure-sensitive film, which provides a measurement of the contact region

created between the two sliding surfaces (glass-on-strip, cartilage-on-strip). From this

measurement the applied stress and the frictional stress were calculated. Reported applied

stresses (0.21-2.58 MPa) are therefore average contact stresses, indicating that the centrally-

loaded region of migrating contact necessarily experiences higher stresses, at least during the

initial testing cycles. Over time, as the contact area changes (especially for cartilage-on-

cartilage), it is important to account for these changes in our analysis. Contact area changes

may account for reductions in S, which may have a measurable influence on the shape of the S-

N curve.

2.6. Acknowledgements

Thanks to Dr. X.E. Guo and the Bone biomechanics laboratory for PLM support.

51

Chapter 3 – Development a Physiologic Chemical Environment that

Maintains the Homeostasis of Immature Bovine Articular Cartilage

Explants in Long-Term in vitro Culture

3.1. Introduction

Due to the dearth of studies on fatigue of cartilage damage, investigations of natural cartilage

repair following episodes of frictional loading are also very limited [112]. The performed studies

set the stage for future work in this area that will shed significant new light into the ability of

cartilage to repair itself in response to physiological loading conditions. Before we can perform

experiments that evaluate in vitro cartilage repair in response to damage, a homeostatic

baseline chemical environment for living cartilage explants must be established.

The ability to maintain living articular cartilage tissue in long-term culture without

significant alterations to tissue composition can serve as a highly valuable analytical tool for

cartilage research. The culture of explanted cartilage tissues in a tightly controlled environment

allows for the direct examination of mechanical or chemical perturbations on cell-mediated

tissue changes. A fundamental challenge for this technique is being able to recreate the salient

environmental conditions of the synovial joint in culture that are required to maintain native

cartilage homeostasis. In principle, the tissue should ideally be maintained in an in situ

homeostatic state where cartilage tissue synthesis is in balance with degradation, as is

experienced by cartilage in its native setting. Interestingly, conventional media formulations

used in explanted cartilage tissue culture investigations often consist of levels of metabolic

mediators that deviate greatly from their concentrations in synovial fluid. Despite a growing

interest in the effect of the mechanical loading environment on cartilage biosynthesis [125], far

52

fewer studies have investigated the effect of the chemical environment as applied by culture

media formulations on tissue homeostasis. Because of our interest in examining the role of the

frictional sliding and delamination on intrinsic tissue repair, we first characterize a chemically-

defined environment for homeostatic cartilage explant culture.

3.2. Background

Live cartilage explant culture models have been used extensively in attempts to provide

important insights into cartilage behavior in response to a multitude of conditions, such as

mechanical injury [126], drug exposure [127], or tissue repair strategies [128, 129].

Conventionally, articular cartilage explants have been procured from a host of sources,

including skeletally mature [130, 131] and immature [128, 132-137] animal joints, and from

human joint sources (from post-mortem donors or tissue discarded from arthroplasty

procedures) [138, 139]. Articular cartilage procured from skeletally immature large animals,

(e.g. bovine or porcine), has served as a critically important tissue source for the cartilage

biomechanics and mechanobiology community, due to its ease of procurement, abundant

availability, and large explant size. In particular, its sizeable characteristic thickness (on the

same scale as human articular cartilage (2-4 mm) [140]) makes these tissues particularly well-

suited for use in cartilage mechanobiology studies, as they can undergo well-controlled

mechanical stimulation protocols and precise measurements of their mechanical properties. For

all these reasons, the use of immature bovine tissue for experimental investigations of living

cartilage tissue is a convenient choice. Another important characteristic of immature tissue is its

high metabolic activity which makes it appropriate for detecting cell and tissue changes on an

relatively short time scale (on the order of days) in the laboratory. The advantage of this high

metabolic activity also comes at a cost: immature cartilage tissue responds not only to targeted

53

experimental conditions, but also undergoes rapid tissue remodeling in response to its ex vivo

environment.

To date, there exists no standard media formulation for in vitro cartilage explant tissue

culture. The challenge of identifying suitable culture media formulations for articular cartilage

has been strikingly evident from several prior investigations, particularly those involving

metabolically active juvenile cartilage explants. Culture in a traditional media formulation,

consisting of Dulbecco’s Modified Eagle Medium (DMEM) supplemented with fetal bovine serum

(FBS), has been shown to induce a rapid and pronounced tissue degeneration, marked by

tissue swelling, softening, and extracellular matrix (ECM) loss [134]. It has previously been

surmised that this degeneration may result from the action of catabolic mediators from serum

that are not similarly present in synovial fluid. Motivated by this outcome, more recent

investigations have performed explant culture in a chemically-defined medium formulation that

has previously been shown to promote chondrogenesis and cartilage growth: DMEM

supplemented with a combination of the cell stimulatory supplement, ITS+ premix, and the

highly potent synthetic corticosteroid, dexamethasone [134, 141]. Interestingly, while this

chondrogenic medium greatly attenuates tissue degeneration, it instead leads to a rapid

synthesis of cartilage ECM and a pronounced stiffening of the tissue [134]. These non-

physiologic, pronounced physical changes in explants may be highly undesirable, as they may

significantly confound the interpretation of results in experimental studies involving cartilage

explants.

3.2. Critical chemical identification and Hypotheses

Immature bovine cartilage explants undergo rapid physical changes when placed in in vitro

culture. These rapid physical changes suggest that in vitro culture conditions may not be

suitably mimicking the tissue’s native environment, leading to an imbalance in anabolic and

54

catabolic processes or a breakdown of tissue homeostasis. While this homeostatic

dysregulation can potentially result from a multitude of factors, including tissue damage from

explant excision or alterations to the tissue’s ex vivo mechanical environment, it is interesting

that media formulations used in prior culture studies often consist of levels of several highly

influential metabolic mediators that deviate greatly from their concentrations in SF (Table 4).

Metabolic Mediator

Levels Utilized in Conventional Media

Formulations

Physiologic Levels in Synovial Fluid (SF)

Glucose - Low glucose DMEM: 1.0 mg/mL [126, 132, 142, 143]

- High glucose DMEM: 4.5 mg/mL [129, 133, 134]

- MEM: 1.0 mg/mL [144] - Ham’s F-12: 1.8 mg/mL [144]

1.1 mg/mL [145]

Insulin - Unsupplemented: 0 ng/mL [129, 130, 132, 133, 135, 137, 139, 142-144, 146, 147]

- 6000 ng/mL via ITS+ premix [128, 134, 148]

1.2+0.3 ng/mL [149]

Corticosteroids

- Unsupplemented: 0 ng/mL [126, 129-131, 133, 135-137, 139, 142-144, 146, 147]

- Dexamethasone: 50 ng/mL [128, 134, 148]

Cortisol (mature SF) Total: 32.1±4.7 ng/mL; Free: 8.7±3.5ng/mL Cortisol (immature SF) Total: 55.2±11.4 ng/mL; Free: 25.6±11.9 ng/mL

Amino acids - DMEM: 107 mg/L - MEM: 48 mg/L - Ham’s F-12: 35 mg/L

20.9 mg/L [150, 151]

Ascorbic acid - Unsupplemented: 0 µg/mL [133, 136] - 20-25 µg/mL [126, 130, 137,

139, 142, 147] - 50 µg/mL [128, 134, 148]

8.0 µg/mL [152]

Table 4: Concentrations of metabolic mediators in conventional media formulations and those present in native synovial fluid. Physiologic glucose, amino acid, and ascorbic acid levels are reported from plasma measurements under the general assumption that they exhibit relatively unobstructed transport across the synovium. Amino acid concentrations are listed as the average concentration among the 15 amino acids included in standard DMEM (see Table S4 for

55

details). Cortisol levels (total and free) are obtained from direct experimental measurements performed on synovial fluid in the current investigation.

In this study, we hypothesize that the homeostatic dysregulation previously observed in

cultured immature cartilage explants results from the presence of non-physiologic levels of

important metabolic mediators in the culture medium. To this end, we hypothesize that: (H2a)

immature bovine cartilage explants cultured in native synovial fluid will maintain homeostasis as

characterized by a maintenance of their mechanical properties and ECM contents at initial

(post-explantation) levels, and (H2b) explants cultured in a physiologic-based medium,

consisting of physiologic levels of key metabolic mediators, will maintain a similar homeostasis

over long-term culture.

The metabolic mediators targeted in this investigation are glucose, amino acids, insulin,

and corticosteroids, all selected based on their presence in synovial fluid and established strong

influence on cartilage ECM biosynthesis. Glucose (180 Da) and amino acids (57-186 Da) are

basic nutrients that are required for cell survival, as they provide a supply of energy to cells for

protein synthesis. Further, these nutrients serve as the fundamental building blocks for the

assembly of the major cartilage ECM constituents; amino acids form the peptide backbone of

type II collagen and the proteoglycan core protein and glucose is a direct precursor for the

monomers present in glycosaminoglycans (GAGs) [153]. Insulin is a peptide hormone (5.8 kDa)

that can increase the cellular uptake of glucose [154] and, at supraphysiologic doses, can

activate the insulin-like growth factor receptor, leading to pronounced biosynthetic

enhancements [155]. Corticosteroids are a class of powerful steroid hormones that serve to

regulate a wide range of physiologic processes. The major systemic natural corticosteroid,

cortisol (~350 Da), has a strong multifaceted effect on cartilage metabolism. At low

concentrations, corticosteroids can inhibit the secretion of matrix degrading proteases [156]; at

56

high levels, corticosteroids can potentially suppress chondrocyte ECM biosynthesis [157]. As

such, levels of cartilage ECM biosynthesis are likely to be dependent on the concentrations of

these critical metabolic mediators in culture medium.

3.3. Materials and Methods

Baseline homeostasis of articular cartilage explants, in their in situ chemical environment, was

determined by subjecting immature bovine explants to long-term culture in bovine synovial fluid.

In order to create a chemically-defined, homeostatic culture medium that would recapitulate this

chemical environment, the effect of individual critical chemical mediators on cartilage was

investigated in a series of parametric studies. Explants were cultured in media formulations of

subphysiologic, physiologic, and supraphysiologic levels (Table 4) of synovial fluid components

known to be critical to cartilage homeostasis: cortisol, insulin, glucose, and amino acids (Critical

Chemical Mediator Experiments #1-3; Table 5). While ascorbic acid similarly plays a critical role

in chondrocyte metabolism [158], its concentration variation assessment was not performed

since it is typically supplemented in culture media at near-physiologic levels (Table 4).

Following culture, explants were analyzed for their mechanical properties and biochemical

contents. Finally, in order to demonstrate the importance of media selection on cartilage

explant experimentation, an additional experiment was performed that examined the effect of

dynamic mechanical loading on cartilage properties while cultured in a physiologic-based media

formulations versus media with subphysiologic or supraphysiologic levels of the aforementioned

metabolic mediators.

At the end of tissue culture, cylindrical cartilage explants were analyzed for changes in

viability, mechanical integrity and ECM content. Mechanical integrity was assessed via the

compressive equilibrium Young’s modulus (EY) and the degree of tissue swelling, measured via

the swelling ratio (SR), evaluated from the tissue wet weight ratio relative to day 0. Changes in

57

explant ECM content were assessed by measuring levels of sulfated glycosaminoglycans

(GAGi), collagen (COLi), and cells (DNAi), all normalized to the initial tissue wet weight. GAG

measurements were further normalized to the tissue final wet weight (GAGf), allowing for an

assessment of the final concentration of GAG in the tissue, accounting for concomitant

volumetric changes to the tissue as a result of swelling.

Furthermore, in addition to influencing the synthesis of ECM constituents, chemical

mediators can also effect the secretion of growth factors and other signaling molecules into the

tissue. While these molecules are secreted at very low levels, they can act as strong autocrine

mediators and have an important impact on the metabolic activity of cartilage [159]. In order to

better understand the impact of metabolic mediators in culture medium on signaling molecule

secretion, we further analyzed the secretion of transforming growth factor beta (TGF-beta).

Expt #

Varied Mediator Concentrations

Constant Mediator Concentrations

Analysis Time Points

1 Cortisol: 0, 5, 10, 25, 50, 200, 1000 ng/mL

Glucose: 1 mg/mL Insulin: 1 ng/mL Amino acids: 1.0X Ascorbic acid: 20 µg/mL

Days 0, 32

2 Glucose: 1, 4.5 mg/mL Insulin: 1, 60, 600, 6000 ng/mL

Cortisol: 25 ng/mL Amino acids: 1.0X Ascorbic acid: 20 µg/mL

Days 0, 32

3 Amino acids: 0.03X, 0.12X, 0.25X, 1.0X

Cortisol: 25 ng/mL Glucose: 1 mg/mL Insulin: 1 ng/mL Ascorbic acid: 20 µg/mL

Days 0, 7, 16, 32

Table 5: Mediator concentration variation experiments performed in this study. 1.0X amino acid concentration is the level present in standard DMEM.

58

3.3.1. Cartilage Explants and Synovial Fluid

Live cylindrical cartilage explants were aseptically harvested from the femoral condyles of

healthy 2-month old calf knee joints (Green Village Packing Co, NJ) at 24 hours after sacrifice.

Explants were extracted via a Æ 3mm biopsy punch (20 explants per condyle from high load

bearing regions) and immediately transferred to 4.5mg/mL glucose Dulbecco’s Modified Eagle’s

medium (HG-DMEM; Thermo Fisher #11965) + 1% antibiotic-antimycotic. Within 24 hours of

collection, explants were trimmed to remove both the underlying bone and a majority of the

cartilage deep zone. Explant cutting was performed using a custom cutting device that

maintains sterility, viability, and an intact articular surface, yielding cartilage disks (Æ 3 mm ´ 1.7

±0.05 mm), consisting of superficial, middle, and upper deep zone cartilage. Day 0 mechanical

properties were assessed and explants were randomized into their respective test groups. In

total, 24 joints were used for this study, with samples evenly distributed from six joints for each

experiment. Healthy synovial fluid (SF) was procured from calf bovine wrist joints (n=8; Green

Village Packing Co), solely for cortisol measurements, and from adult bovine joints (100 mL

pooled lots (n=5); 2-3 year old; Animal Technologies), for cortisol measurements and tissue

culture experiments. Upon receipt, mature SF was thawed, decellularized via centrifugation

(3000 ´ g for 20 min), supplemented with antibiotics (1% antibiotic-antimycotic), and refrozen in

aliquots for tissue culture.

3.3.2. Culture Medium Preparation

The basal medium used throughout these studies was low glucose Dulbecco’s Modified Eagle’s

medium (LG-DMEM; Thermo Fisher #11885). The following media supplements were present in

all cultures: 1% antibiotic-antimycotic (100 U/mL penicillin, 100 µg/mL streptomycin, and 0.25

µg/mL amphotericin-B), 20 µg/mL ascorbic acid, 10 µg/mL sodium pyruvate, and 1 mg/mL

bovine serum albumin (BSA). The metabolic mediators targeted in this investigation were

59

administered at varying levels that encompass those present in physiologic synovial fluid (Table

1). Physiologic levels of insulin, glucose, and amino acids in SF were determined from literature

assessments, as summarized in Table 1. Physiologic levels of cortisol in SF (free and total)

were experimentally measured in both mature and immature SF samples. Cortisol level

variation was achieved through the supplementation of medium with hydrocortisone (Sigma

H0888). Glucose level variation was achieved through the use of low glucose (1 mg/mL) or high

glucose (4.5 mg/mL) DMEM. Insulin level variation was achieved through the supplementation

of medium with ITS+ premix (BD Biosciences). The transferrin and selenium present in ITS+

premix have been shown to have a negligible effect on chondrocyte biosynthesis [158] and,

therefore, were not considered in this analysis. Standard DMEM contains a mixture of amino

acids (all nine essential and six nonessential) at concentrations that are generally higher than

those found in plasma or synovial fluid (107 mg/L versus 21 mg/L on average for each amino

acid; Table 1 & S1). For this study, amino acid level variation was achieved through the

selective mixing of basal DMEM with an amino acid-free DMEM (custom-made, Invitrogen). This

mixing yields formulations with varying fractions of standard DMEM amino acid levels (1.0X,

0.25X, 0.12X, and 0.03X), representing respective concentrations of 107 mg/L, 27 mg/L, 13

mg/L, 3 mg/L when expressed as the average level per amino acid. For all culture experiments,

explants were maintained in culture (1 mL medium per explant) at 37oC and 5% CO2, and, to

avoid nutrient unstirred boundary layer formation, explants were maintained on a slowly orbiting

shaker (30 rpm). Media was changed three times weekly.

3.3.3. Tissue Characterization Methodology

The equilibrium compressive Young’s modulus of cylindrical cartilage explants was measured in

unconfined compression via a custom-testing device, as described [134]. Explants were initially

equilibrated under a 10 g tare load for 5 minutes, followed by a stress relaxation test, consisting

60

of 10% compressive strain (relative to pre-tare thickness) and 20 minutes of relaxation. The

equilibrium compressive stress was calculated based on measured explant dimensions,

allowing for a determination of EY. Biochemical content of explants was determined via the

dimethylmethaline blue (GAG), orthohydroxyproline (collagen), and Picogreen (DNA) assays

after proteinase-K digestion (0.5 mg/mL at 56°C for 20 hours), as described previously [134].

Cortisol levels in SF were determined via ELISA (Parameter, R&D Systems). For cortisol ELISA

measurements, SF samples were initially treated with hyaluronidase (4 µg/mL) for 30 minutes

before assay. For measurements of total (bound + free) cortisol levels, samples were further

treated with proteinase-K (0.5 mg/mL at 56 °C for 20 hours). This protein degradation step

liberates cortisol from binding proteins and has no detrimental effect on the ability of the assay

to detect cortisol (Fig S2). Glucose levels in synovial fluid and culture media were measured

using the Amplex Red Assay (Invitrogen). The TGF-beta secretion rate was assessed by

measuring the latent TGF-b1 concentration in media via ELISA (Duoset, R&D Systems), as

described previously [158, 159]. Chondrocyte viability in explants was assessed via the Live-

Dead assay (Invitrogen) and confocal imaging (Leica, 10X objective).

3.4. Culture Experiments

3.4.1. Cartilage Explant Glucose Consumption

In order to determine the required volume of culture medium to maintain sufficient glucose

availability and tissue viability, a preliminary test was performed. Here, explants were

maintained in 1 mg/mL glucose medium in volumes of 0.3, 0.6, 1.0, 2.0, and 4.0 mL of medium

per explant. After two days of culture, media was assayed for remaining glucose levels.

61

3.4.2. Cartilage Explant Culture in Synovial Fluid

Cartilage explants were cultured in mature bovine synovial fluid (SF) for 20 days (1 mL SF per

explant). SF was changed three times per week. For this experiment, procured SF was

strategically modified in order to better mimic the nutrition environment that is experienced by

immature cartilage explants in situ. For one, it is important to consider that, for these

experiments, immature bovine cartilage explants are cultured in mature bovine SF. According to

our analysis, mature SF possesses significantly lower levels of cortisol than immature SF (Table

1). This analysis is consistent with prior reports about age-related changes on systemic cortisol

levels [160]. This measured concentration disparity can potentially alter explant behavior [161].

To account for this disparity and better replicate the in situ environment of immature cartilage,

mature SF cortisol levels were adjusted to immature SF levels, via the supplementation of an

additional 17 ng/mL of cortisol. Second, SF was further modified to account for nutrient losses

during fluid procurement. During the 24-hour pre-procurement period between animal sacrifice

and fluid aspiration, nutrients in SF likely continue to be consumed by synovial joint cells but are

no longer replenished to the fluid as a result of the cessation of blood flow. To illustrate this

issue, we measure glucose levels at 0.52 mg/mL in our procured adult SF, substantially below

levels routinely observed in vivo (~1.0 mg/mL [145]). Explants cultured in SF consisting of this

sub-physiologic glucose level can potentially undergo extensive viability loss. Therefore, to

account for these pre-procurement losses, our procured SF was replenished with an additional

0.5 mg/mL glucose and other nutrients that are likely to be similarly consumed during the pre-

procurement period (0.15X essential and nonessential amino acids, and 20 µg/mL ascorbic

acid). Lastly, synovial fluid was diluted with an additional 15% v/v phosphate buffered saline

(PBS) in order to reduce its viscosity in an attempt to recreate the convection-induced mixing

62

experienced by the fluid in vivo, while it is situated on our slowly orbiting shaker (30 rpm) within

the incubator.

3.4.3. Critical Chemical Mediator Experiment #1: Corticosteroids

Cartilage explants were cultured for 32 days in basal medium supplemented with a range of

cortisol levels (0, 5, 10, 25, 50, 200, or 1000 ng/mL), bracketing the physiologic level observed

in immature synovial fluid (25 ng/mL; Table 4). An additional set of explants was cultured in

medium supplemented with the synthetic corticosteroid dexamethasone (potency ~ 25-fold

above natural cortisol [162]) at levels typically utilized for tissue culture (50 ng/mL [134]). For all

groups in this experiment, medium contained physiologic levels of glucose (1 mg/mL) and

insulin (1 ng/mL), and supraphysiologic levels of amino acids (1X). Explant properties were

analyzed initially (day 0) and after 32 days of culture.

An additional experiment was performed to determine the enzymes involved in the

pronounced swelling of cartilage explants in the absence of corticosteroids (as shown in

results). Explants were cultured for 14 days in the presence of cortisol-free media consisting of

1 mg/mL glucose and supraphysiologic levels of insulin (6000 ng/mL) and amino acids (1X) in

order to induce a pronounced degree of cortisol-free swelling (consistent with swelling observed

in Fig S1). Media was supplemented with either cortisol (25 ng/mL), or 25 µg/mL of one of the

following specific MMP inhibitors (Santa Cruz Biotechnology): N-[[(4,5-Dihydro-5-thioxo-1,3,4-

thiadiazol-2-yl)amino]carbonyl]-L-phenylalanine for MMP-3 (#SC-311432), (3R)-N-hydroxy-2-(4-

methoxyphenyl)sulfonyl-3,4-dihydro-1H-isoquinoline-3-carboxamide for MMP-8 (#SC-311436),

4-N,6-N-bis[(4-fluoro-3-methylphenyl)methyl]pyrimidine-4,6-dicarboxamide for MMP-13 (#SC-

205756). An additional group of explants was devitalized via a freeze-thaw cycle and

maintained in culture in parallel with the other groups. Following culture, SR of explants was

measured and their viability was assessed.

63

3.4.4. Critical Chemical Mediator Experiment #2: Insulin and Glucose

Cartilage explants were cultured for 32 days in basal medium supplemented with a range of

insulin levels (1, 60, 600, or 6000 ng/mL), including and exceeding levels observed in SF (1

ng/mL; Table 1) and with physiologic (1 mg/mL; Table 1) or supraphysiologic (4.5 mg/mL)

glucose levels. For all groups, medium contained physiologic levels of cortisol (25 ng/mL) and

supraphysiologic levels of amino acids (1X). Explant properties were analyzed initially (day 0)

and after 32 days of culture.

3.4.5. Critical Chemical Mediator Experiment #3: Amino Acids

Cartilage explants were cultured for 32 days in medium supplemented with a range of amino

acid levels (0.03X, 0.12X, 0.25X, or 1X), bracketing the physiologic level observed in synovial

fluid (~0.25X; Table 1). For all groups, medium contained physiologic levels of cortisol (25

ng/mL), glucose (1 mg/mL), and insulin (1 ng/mL). Explant properties and viability were

analyzed initially (day 0) and after 7, 16, and 32 days of culture. The secretion rate of latent

TGF-beta was analyzed over the initial 14 days of culture.

3.4.6. Explant Dynamic Loading in Varying Culture Media

Cartilage explants were cultured in a media formulation, consisting of either physiologic-based,

cortisol-free, or supraphysiologic levels of several of the critical metabolic mediators

investigated in this study: corticosteroid-free (1 mg/mL glucose, 0.25X amino acids, 1 ng/mL

insulin, 0 ng/mL cortisol), physiologic-based (1 mg/mL glucose, 0.25X amino acids, 1 ng/mL

insulin, 25 ng/mL cortisol), and supraphysiologic (1 mg/mL glucose, 1X amino acids, 1 ng/mL

insulin, 50 ng/mL dexamethasone). For the supraphysiologic formation, the synthetic

corticosteroid dexamethasone was used, given its increased potency above natural cortisol. In

each formulation, explants were either maintained under free swelling conditions or subjected to

64

a physiologic dynamic loading regimen (10% platen-to-platen strain, 1 Hz frequency, 1 hour

duration, 3 days per week). Explant properties were analyzed and compared to free swelling

values after 21 days of culture.

3.5. Statistical Analyses

For experiments on the mediator level variation of cortisol (experiment #1) and amino acids

(experiment #3), two-way ANOVAs (a=0.05) were performed to determine the effects of culture

time and mediator level on explant properties: EY, SR, GAGi, GAGf, COLi, DNAi. A three-way

ANOVA (a=0.05) was performed to determine the effect of culture duration, insulin level, and

glucose level on explant properties (experiment #2). One-way ANOVAs were performed to

determine the effect of: 1) enzyme inhibition on explant swelling and 2) amino acid level on

latent TGF-beta secretion. Tukey’s post hoc test was used to detect differences in the means.

For explant dynamic loading experiments, an unpaired Student’s t-test (two-tailed) was used to

compare explant properties between dynamic loading and free swelling within each media

group. Statistical significance was set at p<0.05.

3.6. Results

3.6.1. Cartilage Explant Glucose Consumption

In the presence of cartilage explant tissues, medium glucose levels decreased after 2 days of

culture due to cellular consumption (Figure 22). Remaining glucose concentrations varied

nonlinearly with the medium to tissue volume ratio; at a ratio of 25:1, only 2.5% of the original

glucose was remaining in medium, while at a 330:1 ratio, 98% of the glucose remained. After 32

days of culture, explants maintained at a 30:1 volume ratio exhibited a substantial loss of cell

viability in the tissue central region, while a ratio of 100:1 resulted in no viability loss (Fig S5).

65

For all experiments of this study, a 100:1 ratio was adopted in light of its ability to maintain a

majority (>60%) of medium glucose level between media change periods.

3.6.2. Cartilage Explant Culture in Synovial Fluid

After 20 days of culture in synovial fluid, explants exhibited no viability loss (results not shown),

no statistically significant swelling (SR=1.02±0.04; p>0.05; Fig 23A-SF culture), and no

statistically significant change in EY (p>0.05; Fig 23B-SF culture). Explants did exhibit a small,

but statistically significant increase in GAG content (p<0.05; Fig 23C-D-SF culture) with GAGi,

and GAGf increasing by 23% and 22%, respectively.

Figure 22: Remaining glucose levels in culture medium after 48 hours of explant incubation for varying media to tissue volume ratios. Dashed line represents initial supplemented glucose level (1 mg/mL).

3.6.3. Critical Chemical Mediator Experiment #1: Corticosteroids

For subphysiologic cortisol levels (0 ng/mL and 5 ng/mL), explants exhibited statistically

significant swelling after 32 days of culture (p<0.05; Figs 3A, 3G, 3H); SR increased by 30%

and 26% above day 0 values, respectively. At high swelling levels, explants exhibited crack

formation in the tissue center (Fig 24H-cortisol). Swelling decreased with increasing cortisol

66

levels. For physiologic and all supraphysiologic cortisol levels, SR remained at day 0 values

(p>0.99). Conversely, EY increased with increasing cortisol levels (Fig 24B-cortisol): In the

absence of cortisol (0 ng/mL) EY decreased by 60% below day 0 values (p=0.16), at other

subphysiologic cortisol levels (5 ng/mL, 10 ng/mL) EY was maintained at day 0 values (p>0.94),

and at physiologic and supraphysiologic cortisol levels (25-1000 ng/mL) EY increased above day

0 values (p<0.05). GAGi increased above day 0 values for all groups with the exception of 0

ng/mL and 50 ng/mL cortisol levels (Fig 24C-cortisol). GAGf was maintained at day 0 values for

0 ng/mL, 5 ng/mL and 1000 ng/mL cortisol levels and increased above day 0 values for all other

groups (p<0.03; Fig 24D-cortisol). COLi (p>0.99) and DNAi (p>0.59) did not differ from day 0

values for all groups (Fig 24E-F-cortisol). The response of explants to 50 ng/mL dexamethasone

was not significantly different from that of 50 ng/mL cortisol in regards to EY, SR, GAGi, and

GAGf (Fig S6).

67

Figure 23: Mechanical properties and biochemical content of cartilage explants initially and after 20 days of culture in synovial fluid: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, and (D) GAG content per final weight wet. *p<0.05 represents significant variation from day 0 value.

3.6.4. Corticosteroid-Free Explant Swelling: Enzyme Secretion

Live control explants exhibited a large degree of swelling in response to cortisol-free medium

(SR= 2.68±0.64; Fig 25-enzyme). Devitalization prevented explant swelling (SR= 1.03±0.08).

Swelling of live explants was statistically reduced in the presence of cortisol (SR= 1.37±0.33) or

the specific MMP-13 inhibitor (SR= 1.14±0.24), but not in the presence of the specific MMP-3

(SR=2.09±0.43) or MMP-8 (SR= 2.49±0.56) inhibitors. MMP inhibitors had no effect on cell

viability (results not shown).

68

Figure 24: Critical chemical mediator experiment #1: Mechanical properties, biochemical contents, and histology of cartilage explants cultured for 32 days in medium with varying levels of cortisol: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, (D) GAG content per final weight wet, (E) collagen content, and (F) cell content. Insulin (1 ng/mL), glucose (1 mg/mL), amino acid (1.0X), and ascorbic acid (20 µg/mL) levels were maintained constant for all groups. Red arrow indicates physiologic level of cortisol present in synovial fluid. Dashed lines represent day 0 values. *p<0.05 represents significant variation from day 0 value. (G) Representative images of cartilage explants initially (day 0) and after 32 days of culture with varying cortisol levels. Scale bar = 1.0 mm. (H) Representative safranin-O stained sections of explants initially and after 32 days of culture with varying cortisol levels. Arrow represents crack formation in tissue center of 0 ng/mL cortisol samples, which is distinct from cartilage vascular canals (* labeled) that are present in other groups and characteristic of immature cartilage tissue. Scale bar = 1.0 mm.

3.6.5. Critical Chemical Mediator Experiment #2: Insulin and Glucose

SR increased with increasing levels of insulin and glucose, reaching values significantly above

day 0 for 600 ng/mL and 6000 ng/mL insulin at low glucose and 60 ng/mL, 600 ng/mL, and 6000

69

ng/mL insulin at high glucose (Fig 26A & 26H-insulin). At the high glucose and insulin levels,

explants exhibited a dramatic swelling-induced void in the tissue center (Fig 26I). EY values

increased above day 0 for all groups (p<0.05; Fig 26B). GAGi increased with increasing levels of

insulin and glucose, reaching values statistically above day 0 for 60 ng/mL, 600 ng/mL, and

6000 ng/mL insulin at both low and high glucose (Fig 26C). GAGf increased above day 0 values

for 600 ng/mL, and 6000 ng/mL insulin at low glucose and 60 ng/mL, 600 ng/mL, and 6000

ng/mL insulin at high glucose (Fig 26D). Levels of insulin and glucose had no statistically

significant effects on COLi (p=0.95; Fig 26E) but DNAi increased for 600 ng/mL and 6000 ng/mL

insulin at high glucose (Fig 26F). A strong linear correlation (p=0.93) was observed between SR

and GAGi (Fig 26G).

Figure 25: Swelling ratio of cartilage explants cultured for 32 days in cortisol-free medium while subjected to devitalization, cortisol supplementation, or specific enzyme inhibitors. *p<0.05 represents statistical variation from control value.

3.6.6. Critical Chemical Mediator Experiment #3: Amino Acids

For physiologic amino acid levels (0.25X), SR and EY were statistically maintained at their day 0

values for up to 32 days (p>0.58; Fig 27A-B-AAs). While GAGi and GAGf increased above day 0

70

values, they did not further increase above day 7 values at subsequent time points (p>0.77; Fig

27C-D-AAs). In contrast, for supraphysiologic amino acid levels (1.0X), all properties increased

above day 0 values and continued to rise above day 7 values. For subphysiologic amino acid

levels (0.12X and 0.03X), all properties exhibited no statistically significant change from day 0

values with the exception of GAGi at 0.12X level on day 16. Explants from all groups exhibited

no cell viability loss (Fig 27E).

Explants secreted similar levels of latent TGF-b1 when cultured in physiologic (0.25X) or

supraphysiologic (1.0X) amino acid levels (Fig 28-LTGFsec). At subphysiologic AA levels, latent

TGF-b1 was secreted at a statistically significant decreased rate (p<0.05).

3.6.7. Explant Dynamic Loading in Varying Culture Media

The response of explant properties to physiologic dynamic loading was dependent on the

culture medium type (Fig 29-DL+homeo). Dynamic loading of explants in physiologic medium

increased EY by 80% (p=0.03) and GAGi by 15% (p=0.07) relative to free swelling control

values. In contrast, explants cultured in subphysiologic (corticosteroid-free) or supraphysiologic

medium exhibited no statistical change in EY, or GAGi relative to free swelling control values.

3.7. Discussion

The capacity to maintain living cartilage explants in long-term culture under homeostatic

conditions can serve as a highly valuable tool for cartilage research. The results of this study

have led to the identification of four key metabolic mediators—glucose, amino acids, insulin, and

cortisol—that have strong influences on explant mechanical properties and biochemical

contents, and, when administered at near-physiologic levels, can maintain homeostasis of

immature cartilage explants for at least up to one month in culture.

71

Figure 26: Critical chemical mediator experiment #2: Mechanical properties, biochemical contents, and histology of cartilage explants cultured for 32 days in medium with varying levels of insulin and glucose: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, (D) GAG content per final weight wet, (E) collagen content, and (F) cell content. Cortisol (25 ng/mL), amino acid (1.0X), and ascorbic acid (20 µg/mL) levels were maintained constant for all groups. Red arrow indicates physiologic levels of insulin and glucose present in synovial fluid. Dashed lines represent day 0 values. *p<0.05 represents statistical variation from day 0 value. †p<0.05 represents significant variation from corresponding 1 ng/mL insulin value. (G) Linear correlation between swelling ratio and GAG content. (H) Representative images of cartilage explants initially (day 0) and after 32 days of culture for each media group. Scale bar = 1.0 mm. (I) Representative safranin-O stained sections of explants initially and after 32 days of culture for 6000 ng/mL insulin and 4.5 mg/mL glucose group. Arrow represents crack-induced tissue void in tissue center. Scale bar = 1.0 mm.

72

Figure 27: Critical chemical mediator experiment 3: Mechanical properties, biochemical contents, and viability of cartilage explants cultured for 7, 16, or 32 days in medium with varying levels of amino acids: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight, and (D) GAG content per final wet weight. Insulin (1 ng/mL), glucose (1 mg/mL), cortisol (25 ng/mL), and ascorbic acid (20 µg/mL) levels were maintained constant for all groups. Red arrow indicates physiologic level of amino acids present in synovial fluid. 1X amino acid level represents concentrations present in standard DMEM. Dashed lines represent day 0 values. *p<0.05 represents significant variation from day 0 value. †p<0.05 represents significant variation from corresponding day 7 value. (E) Representative viability images (live cells: green, dead cells: red) of cartilage explants after 32 days of culture for 0.25X and 0.03X amino acid groups.

An important foundation of this investigation is the novel use of synovial fluid to determine

the baseline behavior of cartilage explants in vitro. These results demonstrate that explants do

indeed maintain near homeostasis for several weeks when cultured in synovial fluid, as

evidenced by their maintenance of day 0 mechanical properties and only minor increases

(~20%) in their biochemical contents (Fig 23). As such, these findings provide the important

conclusion that the dramatic homeostatic dysregulation observed in several prior explant culture

studies can be almost completely attributed to the presence of a non-physiologic chemical

73

environment rather than artifacts induced during the physical explantation of the tissue. In

contrast to some prior reports [163], we observed no loss of chondrocyte viability during

synovial fluid culture, suggesting the importance of performing several modifications to synovial

fluid to improve glucose availability during in vitro culture: 1) replenishing fluid procurement

glucose losses, 2) reducing the viscosity of the fluid, and 3) maintaining cultures mixed on an

orbital shaker.

Figure 28: Secretion rate of endogenous cartilage growth factor, latent TGF-b1, from cartilage explants cultured in medium of varying levels of amino acids. Secretion rate is normalized to explant volume. Red arrow indicates physiologic level of amino acids present in synovial fluid *p<0.05 represents significant variation from 1.0X value.

The metabolic mediators investigated in this study have an intriguing and complex role on

the behavior of cartilage explants, influencing both the mechanical stability of the tissue ECM

and the biosynthetic rates of new ECM products. Changes to ECM stability and rates of

biosynthesis manifest in alterations to the biochemical contents, degree of tissue swelling, and

the tensile and compressive mechanical properties of explants. The controlled examination of

the concentration dependence of each metabolic mediator, as performed in this study, has

74

allowed for the development of insights into the mechanistic contribution of each mediator

towards explant properties and homeostasis.

Figure 29: Mechanical properties and biochemical contents of cartilage explants cultured for 32 days in a corticosteroid-free, physiologic, or supraphysiologic culture medium while subjected to dynamic mechanical loading or free swelling conditions: (A) swelling ratio, (B) compressive Young’s modulus, (C) GAG content per initial wet weight. *p<0.05 †p=0.07 represents statistical comparison between corresponding free swelling value.

Perhaps the most striking demonstration of the current investigation is the dramatic

response of cartilage explants when cultured in the absence of corticosteroids, whereby

explants exhibit pronounced tissue swelling (Fig 24A & 24G-cortisol) that can lead to cracking in

the tissue central region (Fig 24H-cortisol). This swelling is not observed during culture in

synovial fluid (Fig 23A-SF). In order to understand the mechanism of corticosteroid-free explant

swelling, it is important to note that the degree of swelling of cartilage tissue is dependent on the

well-documented balance between the GAG-induced Donnan osmotic swelling pressure and the

tensile integrity of the collagen matrix, which acts to physically resist this swelling pressure

[164]. Consequently, changes in the degree of tissue swelling result from alterations to either

the tissue GAG content or the tensile integrity of the collagen matrix. Since explants cultured in

the absence of corticosteroids exhibit no change from their day 0 GAG content (GAGi; Fig 24C-

cortisol), explant swelling can be predominantly attributed to a loss of collagen tensile integrity,

75

consistent with prior direct mechanical characterizations [45]. This attribution is further

consistent with the observed softening of explants in the absence of corticosteroids (Fig 3B),

given the dependence of the tissue modulus in unconfined compression on a combination of

both the GAG content and collagen matrix tensile properties [165].

Metabolic Mediator

Medium supplemented

level Glucose 1 mg/mL Amino acids 27 mg/L Insulin 1 ng/mL Cortisol 25 ng/mL Ascorbic acid 20 µg/mL

Table 6: Levels of critical metabolic mediators in physiologic-based culture medium required to maintain articular cartilage explant homeostasis.

The attribution of corticosteroid-free swelling to collagen integrity loss is consistent with

prior reports that corticosteroids can suppress chondrocyte secretion of matrix-degrading

enzymes, such as matrix metalloproteinases (MMPs) [156, 166], and, thus, inhibit cartilage

degradation [167]. In fact, it has long been known that explanted cartilaginous embryonic limbs

can uptake substantial amounts of water and swell in culture, a phenomenon that is inhibited by

corticosteroids [168, 169]. The swelling of immature cartilage in the current investigation likely

results from similar mechanisms. In the current investigation, tissue swelling is greatly

attenuated in the presence of a specific MMP-13 inhibitor and does not occur in devitalized

explants (Fig 25-enzyme), suggesting that cell secretion of this specific collagenase enzyme

plays a major role in collagen integrity loss. It is important to note that in the absence of

corticosteroids, we do not observe a decrease of collagen levels in explants (Fig 24E-cortisol).

This suggests that explant swelling results not from a loss of collagen from the tissue, but rather

from a breakdown of the integrity of the collagen network that was originally present. As such,

76

we suspect that the observed loss of collagen tensile integrity during swelling likely results from

MMP induced “damage” to type-II collagen [170]. However, we do not entirely discount the

possibility that a loss of tensile integrity can partially result from enzyme-induced degradation of

collagen crosslinking molecules (e.g. pyridinoline, glycoproteins, and minor collagen molecules),

which can also influence tensile properties [45, 171, 172].

The collagen network appears to be greatly stabilized through the supplementation of

physiologic levels of the natural corticosteroid, cortisol. However, while cortisol is necessary to

maintain explant homeostasis in tissue culture medium, in itself, it appears to be insufficient.

Even in the presence of physiologic cortisol levels, explants in this study still underwent

homeostatic dysregulation, as exhibited by pronounced increases in GAG content (GAGi

increase of 78%; Fig 24C-cortsiol) and mechanical stiffening (EY increase of 70%; Fig 24B).

GAG accumulation was far greater than for explants cultured in synovial fluid (Fig 23C-SF) and

can likely be attributed to the presence of supraphysiologic levels of other anabolic mediators in

the culture medium that promote ECM biosynthesis.

The anabolic mediators insulin, glucose, and amino acids were targeted in this investigation

based on their documented influence on chondrocyte GAG biosynthesis [138, 158, 173]. The

results of this study confirm that supraphysiologic levels of these mediators can greatly

contribute to homeostatic dysregulation in cartilage explants. In general, increased levels of

these mediators led to increased GAG accumulation, tissue swelling, and stiffening. At the

highest levels of insulin, glucose, and amino acids utilized in the study, the changes to explant

properties were strikingly evident: GAGi, SR, and EY increased by 3.5-fold, 2.2-fold and 2-fold,

respectively (Fig 26-insulin). It is important to note that the mechanism of supraphysiologic

anabolic mediator-induced swelling differs from swelling observed in the absence of

corticosteroids. Whereas corticosteroid-free swelling appears to be initiated by collagen network

breakdown and marked by tissue softening, swelling induced by supraphysiologic anabolic

77

mediators appears to alternatively result from an increased synthesis and accumulation of GAG

in the tissue, leading to an increased Donnan osmotic swelling pressure. This swelling pressure

produces an initial stiffening of the tissue; at high levels of accumulation, the increased osmotic

pressure overcomes the native level of tensile resistance by the collagen network, resulting in a

pronounced swelling effect. This attribution is strongly supported by the high correlation

between tissue swelling (SR) and GAG content (GAGi; Fig 26G-insulin).

The use of physiologic levels of these key anabolic mediators greatly mitigates changes to

mechanical properties, swelling, and GAG content, giving rise to a homeostatic maintenance

similar to that achieved in synovial fluid. The slight increase in GAGi at day 7 (for 0.25X amino

acids; Fig 27C-AA) mirrors the increase observed in SF culture (Fig 23C-SF). The GAG content

does not continue to rise at subsequent time points (Fig 27C-AA), suggesting that this minor

GAG accumulation may result from the explant excision process or exposure of non-physiologic

levels of other mediators in DMEM that have not been examined in this investigation. It is

important to note that, while the reduction of anabolic mediators to physiologic levels greatly

decreases the GAG synthesis rate, it induces no loss of cell viability, suggesting that

chondrocytes are still capable of maintaining their essential metabolic activity. In fact, no viability

loss is observed in explants cultured for 32 days, even at exceptionally low, subphysiologic

amino acid levels (0.03X amino acid group, representing a 97% decrease below DMEM levels

and 86% decrease below physiologic levels; Fig 27E). Furthermore, the reduction of amino

acids from supraphysiologic (1X) to physiologic levels (0.25X) appears to have no influence on

the secretion of the critical signaling protein, latent TGF-beta (Fig 28-LTGF), suggesting that the

synthesis of essential autocrine growth factors and perhaps other signaling molecules remain

unaffected by these lowered amino acid levels.

In addition to the importance of physiologic concentrations of key mediators in culture

media, the suitable replenishment of these mediators is equally critical. While the native

78

synovial joint consists of low volumes of synovial fluid, small nutrients are replenished at high

rates via transport from the vasculature surrounding the synovium. For conventional tri-weekly

media change protocols for in vitro culture, it becomes necessary to utilize a sufficient media

volume to ensure that nutrients are not depleted during the 48-72 hour period between media

changes [174]. This consideration becomes critical when using physiologic levels of glucose

(1 mg/mL). Remaining glucose levels vary non-linearly with media volume, and the use of

insufficient media to tissue volume ratios can potentially lead to “unsafe” low glucose levels by

the end of the media change period (Fig 22-glucose). Over 32 days of culture, the use of a

media to tissue volume ratio (30:1) that exhibits pronounced glucose depletion manifests into a

significant loss of viability in the central regions of the tissue (Fig S5), presumably due to

insufficient glucose levels for interior chondrocytes. Alternatively, our results demonstrate that

the adoption of a media to tissue volume ratio (100:1) that substantially reduces glucose

depletion can prevent explant viability loss during long-term culture.

Historically, basal media formulations have focused on the ability to impart rapid

proliferation to isolated cells in culture. More recently, highly supraphysiologic chondrogenic

media formulations have been developed to impart rapid tissue growth for regenerative

medicine platforms [175] or to achieve long-term survival of live osteochondral allografts during

tissue bank storage [176]. The results of this study offer a novel perspective and underscore the

importance of the selection of the chemical environment for different applications that involve

either long-term maintenance or long-term growth of cartilage tissue. While media formulations

used in cartilage explant culture often possess physiologic levels of several of the mediators

targeted in this study, we have yet to identify a reported culture protocol where all four

mediators are administered at physiologic levels. In particular, corticosteroids are rarely

incorporated into media formulations, and DMEM, which possesses a 4-fold excess of each

amino acid on average (Table S4), appears to have become one of the most commonly utilized

79

basal formulations for cartilage culture. Alternatively, the key mediators identified in this study

can readily be administered in media formulations at physiologic levels (listed in Table 6);

cortisol, insulin, and ascorbic acid can be readily supplemented via stock solutions. For glucose,

basal media formulations (e.g. DMEM and MEM) are commercially available at physiologic

glucose levels (1 mg/mL), commercially termed as “low glucose” medium for DMEM (LG-

DMEM). For amino acids, physiologic levels can be achieved through titration of standard

DMEM with a custom amino acid-free DMEM formulation, as utilized in this study. Interestingly,

other commercial basal media formulations (e.g. MEM and Ham’s F-12) possess amino acid

concentrations that are closer to physiologic levels (Table S6) and, although they were not

investigated in this study, they may serve as attractive options for cartilage explant culture. It is

important to consider that while this physiologic-based formulation appears to mimic the native

chemical environment, synovial fluid possesses a multitude of additional metabolic mediators

that can greatly influence chondrocyte biosynthesis and may be a factor in maintaining

homeostasis when the tissue is subjected to mechanical loading or other external conditions.

Cartilage explant culture can serve as a highly valuable and versatile tool for cartilage

research, allowing for: 1) insights into tissue function and physiology and 2) assessments of the

efficacy of disease modifying and tissue repair strategies. The utilization of a physiologic-based

medium may be crucial for successfully developing biological insights from these systems. The

induction of dramatic physical changes in explants, which do not occur in the native setting,

such as tissue swelling, stiffening, and GAG accumulation, can greatly confound the

interpretation of experimental results. Most notably, large non-physiologic tissue changes can

potentially mask the effect of a target stimulus or elicit a non-physiologic tissue response. The

importance of a physiologic medium is strongly demonstrated in our experimental monitoring of

cartilage explants in response to a long-term regimen of dynamic mechanical loading. As

depicted (Fig 8-DL+homeo), dynamic loading appears to have no effect on the properties of

80

cartilage explants when utilizing a culture medium that is corticosteroid-free or possesses

supraphysiologic levels of amino acids. We suspect that the increased rate of ECM biosynthesis

induced by supraphysiologic amino acid levels effectively masks the effect of mechanical

stimulation. Further, ECM swelling/degradation in the absence of corticosteroids may alter the

transmission of mechanical loads to chondrocytes, compromising their mechanotransduction

response. In contrast, the use of a physiologic-based medium elicits cartilage’s classical

mechanobiological behavior, whereby dynamic loading enhances the GAG content and stiffness

of the tissue [177, 178]. In a further demonstration, through the use of our physiologic-based

medium, we observe that the supplementation of fetal bovine serum induces a high degree of

tissue swelling and cell proliferation in explants, but, interestingly, has no effect on ECM

synthesis (Fig S7).

The results of this study may have broad applicability for musculoskeletal tissue research.

As outlined in this study, physiologic-based media are particularly critical for the in vitro

maintenance of immature cartilage tissues, which possess high cell densities and associated

high rates of metabolic activity and ECM remodeling. However, the utilization of a physiologic

culture medium may also be critical for culture of mature cartilage, whereby a lack of

corticosteroids or anabolic mediator excesses may also induce undesirable physical tissue

changes, albeit those that are likely less pronounced than physical changes in immature

tissues. Further, given the commonality of the synovial fluid-based chemical environment for

many musculoskeletal tissues, the culture medium developed in this investigation can serve as

an important formulation for investigations into the behavior of many other musculoskeletal

tissues of the synovial joint, such as tendons, ligaments, synovium, and the meniscus.

81

3.8. Conclusions and Future work

We hypothesizes that the utilization of a culture medium consisting of near-physiologic levels of

several highly influential metabolic mediators (glucose, amino acids, cortisol, insulin, and

ascorbic acid) will maintain the homeostasis of cartilage explants as assessed by their

mechanical properties and extracellular matrix contents. Results demonstrate that the

aforementioned mediators have a strong effect on the mechanical and biochemical stability of

skeletally immature bovine cartilage explants. Most notably, 1) in the absence of cortisol,

explants exhibit extensive swelling and tissue softening and 2) in the presence of

supraphysiologic levels of anabolic mediators (glucose, amino acids, insulin), explants exhibit

increased matrix accumulation and tissue stiffening. In contrast, the administration of

physiologic levels of these mediators (as present in native synovial fluid) greatly improves the

stability of live cartilage explants over one month of culture. These results may have broad

applicability for articular cartilage and other musculoskeletal tissue research, setting the

foundation for important culture formulations required for examinations into tissue response to in

vitro wear.

3.9. Acknowledgments

Research was supported by the National Institute of Arthritis and Musculoskeletal and Skin

Diseases under Award Numbers AR043628, AR060361, AR046568.

82

Chapter 4: Laser-based treatment strategy to delay OA progression

4.1. Introduction

Combining our understanding of the manifestation of fatigue damage and our in vitro

model of cartilage wear (Chapter 2) with a homeostatic chemical environment (Chapter 3), we

have laid the groundwork to investigate novel in vitro repair strategies on living cartilage

explants. In this Chapter we demonstrate a practical treatment modality capable of stopping or

slowing down structural degeneration of articular cartilage in OA. A radically different use of

ultrafast lasers is presented to achieve this goal, which relies on inducing a photochemical effect

rather than optical breakdown [179] or photoablation [180] which departs from current laser-

based OA cartilage treatment strategies (Section 4.1.2). Although our collaborators have

recently demonstrated viability of the treatment modality in corneal stroma [181, 182], articular

cartilage is compositionally and structurally more complex. Therefore, the following experiments

serve as the initial in vitro validation of the technique as applied to cartilage and lay the

groundwork for extensive in situ and in vivo animal studies, which are envisioned as the natural

continuation to this dissertation work.

4.1.1. Loss of Collagen Crosslinks leads to OA

Whether the initiating factors of OA are acutely mechanical (trauma) or gradually developing

with age, an imbalance in ECM homeostasis is a key pathogenic pathway of OA. The disruption

of collagen network leads to its inability to withstand the swelling pressure of PGs, resulting in

increased water content. Along with swelling, increased proteolytic activity shifts the ECM

homeostasis towards degradation of matrix components, progressively leading to cartilage

degeneration and loss of function [183]. The loss of collagen crosslinks is key to collagen

network instability in OA. The onset of OA is characterized by changes to the structure, though

83

not necessarily the content, of the collagen matrix in articular cartilage. Crosslinks (CxLs)

stabilize the collagen fiber network, and their disruption leads to the loss of tensile strength and

structural integrity [184]. In cartilage, hydroxylysyl pyridinoline (PYR) is the major mature CxL

and pentosidine (PSD) is the senescent CxL. Lysyl oxidase mediates covalent crosslinking of

collagen fibrils by oxidizing hydroxylysine residues to hydroxylysyl aldehydes which then,

through several reactions, lead first to immature PSD, then to stable PYR CxLs [185, 186]. The

contribution of crosslinking to tensile strength and stability of cartilage collagen matrix has been

extensively demonstrated in tissue engineering studies [184]. In animal models of OA,

crosslinking loss determined the irreversibility of cartilage ECM degeneration, irrespective of the

level of proteoglycan loss [187]. In patients, increased levels of PYR and PSD in urine

correlated with the severity of OA and other degenerative joint diseases [188].

4.1.2. Current laser treatment therapies for cartilage Current use of lasers in OA treatment is divided into two distinct fields, disease management

[189-193] and techniques related to laser-assisted tissue ablation [194-198] such as

debridement and abrasion, used in the arthroscopic treatment of OA and cartilage lesions [199-

201]. In the latter, the laser is used as a thermal source to perform chondroplasty, aiming to

smooth the fibrillated articular cartilage surface. For some practitioners lasers are seen as a

more effective replacement to mechanical chondroplasty [202-204]. However, commonly used

lasers such as excimer and solid state lasers cause significant thermal injury to the joint [205]

that can lead to avascular necrosis of adjacent subchondral bone, affecting the viability of the

remaining articular cartilage [206, 207]. Smoothing of osteoarthritic cartilage has been

proposed to be beneficial because it may reduce friction and further wear [203, 208]. However,

shaving off fibrillated layers of cartilage may also remove biomechanically protective regions

[209]. Furthermore, shaving treatment does not improve the biomechanical properties of OA-

afflicted cartilage and the resulting pain relief may only be temporary [208]. In particular, it has

84

been shown [208] that excimer laser-treated articular cartilage in rabbits has inferior structural

integrity when compared against normal tissue, despite its cartilage-like morphological

appearance.

4.1.3. Femtosecond laser induces collagen crosslinks

A novel laser treatment strategy has been shown to induce crosslinks into the type I collagen

network of corneal tissue [181, 182]. Here we extend this platform to the more complex type II

collagen network of articular cartilage. The treatment modality utilizes an ultrafast oscillator,

which has extremely low pulse energy (~1.2 nJ), and as such does not inflict thermal damage,

as demonstrated in previous work on porcine cornea (in vitro) [182, 203] and lapine corena (in

vivo) [181]. When ultrashort pulses carrying nano-Joule (nJ) energy are relatively loosely

focused onto collagen-rich biological media, the interaction results in the formation of low

density plasma (LDP) within the focal volume and its immediate vicinity. Though this LDP

regime has been previously observed and reported, with second harmonic imaging showing that

it alters the signal from collagen rich tissues [210], today this phenomenon is mostly viewed as

an undesired side-effect occasionally present in multiphoton imaging [180]. Here, we use LDP

as a tool for the generation of an ionization field within biological media, without producing

tissue-damaging thermoacoustic [180] and shock waves [211]. The ionization field locally

ionizes and dissociates interstitial water, creating reactive oxygen species (ROS) which interact

with surrounding proteins to form collagen CxL, giving rise to spatially resolved alterations in

mechanical properties. ROS initiate photo-oxidation of proteins, which results in the formation

of chemical CxLs [212]. All amino acids are susceptible to modification by free radicals;

however tryptophan, tyrosine, histidine, and cysteine are particularly sensitive [213]. Amino

acids involved in CxL formation in corneal tissue include histidine, hydroxylysine and tyrosine

[214]. Oxidative modification of tyrosine is characterized by abstraction of phenolic hydrogen

atom from tyrosine residues. The tyrosyl radical is relatively long-lived and can react with

85

another tyrosyl radical or tyrosine to form a stable, covalent carbon-carbon bond, resulting in the

creation of 1,3-dityrosine. This formation is a product of protein oxidation which leads to intra-

or intermolecular crosslinking [215]. The reaction serves as a primer of pathways that lead to

crosslinking of ECM upon irradiation with a femtosecond oscillator.

4.1.4. Reactive Oxygen Species

Due to its avascular nature, the metabolic requirements of adult cartilage are mostly serviced by

diffusion from synovial fluid. Oxygen levels in the tissue are rather low, inhibiting glycolysis and

ECM production. In this environment, reactive oxygen species (ROS) play significant

physiological roles, including cell activation, proliferation and matrix modeling. Sustained levels

of ROS are required for the maintenance of ion homeostasis in chondrocytes. However, the

overproduction of ROS in pathological cartilage has been associated with the accumulation of

lipid peroxidation products and nitrotyrosine in situ. This well-documented role of ROS in OA-

afflicted articular cartilage requires careful consideration of potentially adverse effects of laser-

induced ROS [216]. There is also evidence that antioxidant supplementation is associated with

the slowing of OA progression [217, 218].

4.1.5. Hypotheses

The laser treatment strategy examined in this dissertation was selected because of its distinct

capability to reform collagen crosslinks, replacing those being lost during OA progression,

without injuring the cells and without any chemical additive or thermal ablation. The laser

treatment protocol used in this application can specifically target the subsurface region, located

200 μm of the articular surface. By strengthening this region with enhanced crosslinking, we

hypothesize that cartilage will demonstrate greater resistance to fatigue failure than untreated

controls. We then hypothesized that this treatment protocol would also be effective on

86

devitalized fibrillated human articular cartilage from OA joints with overall Outerbridge score

OS1-3.

4.2. Materials and Methods

4.2.1. Experimental Design Study 1: Laser treatment immature bovine at different depths

In Study 1, we tested devitalized immature bovine cartilage explants (Æ3 mm ´ 1 mm) in a two-

factorial experimental design: Factor 1 tested the effect of laser treatment dosage at a given

energy level (0, 1 and 2 rasterized laser treatment passes of the explant surface, with 0

representing the untreated control group); factor 2 tested the effectiveness of the treatment with

increasing depth from the articular surface (focusing the laser at depths of 0-200 µm, and 100-

300 µm). This 3´2 design had n=6 explants tested in each cell (total of 36 explants). The

principal outcome measure was the compressive equilibrium unconfined compression modulus

EY. A subset of paired control and treated samples from this study were wear-tested under SCA

to determine tissue resilience.

Preliminary MCA wear tests on two pairs of (Æ10 mm ´ 1.2 mm) immature bovine

cartilage discs were also performed. One disc from each pair was laser-treated at its center (3 x

3 square raster pattern). A spherical glass lens (R12.7 mm) was used as the counterface under

4.45 N load. The contact area migrated past the laser treated region, 4 mm travel at 1 mm/s.

4.2.2. Experimental Design Study 2: Laser treatment OA human

In Study 2, we tested the hypothesis that this treatment protocol would also be effective on

devitalized fibrillated human articular cartilage from OA joints with overall Outerbridge score

OS1-3 [219] (Figure 31, 32 ). Three OA joints were used (2 F and 1 M, ages 73-95). Paired

cartilage plugs (control and laser-treated, Æ3 mm ´ 1.3 mm) were harvested from regions with

local OS1 (outerbridge score (5 sample pairs) and OS2 (7 sample pairs); regions with OS3 were

87

too soft to isolate testable tissue plugs. The compressive modulus EY was measured before and

after sham (control) or laser treatments.

Figure 30: Human OA distal femur with overall Outerbridge score 3. Dashed line represents histology section of Fig 31b. Circles on lateral condyle (left) are sites of tissue harvest for laser treatment study, rectangles show paired samples. Inset (bottom right) shows harvested sites of samples used for results reported in Figure 34.

4.2.3. Experimental Design Study 3: Cell Viability after treatment in long-term culture

In Study 3 we laser-treated live immature bovine cartilage explants (Æ3 mm ´ 1 mm) (0 and 1

pass, from 100 to 300 µm, n=12 per group) and examined EY and cell viability (Live/Dead Assay

Kit, Molecular Probes) 24 h after laser treatment. Additional samples were maintained in culture

and viability was assessed after two weeks.

4.2.4. Experimental Design Study 4: ROS and Vitamin C

In Study 4, devitalized immature bovine cartilage explants (Æ3 mm ´ 1 mm) were bathed either

in PBS or PBS + 1M AAP for 24 hours at 4oC (n=4 per group). Laser treatment was

88

administered in the respective bathing solutions. The compressive modulus EY was measured

after laser treatment.

Fig 31. (a) Human femoral condyle (OS3), with evidence of delamination (arrow). (b) Human femoral condyle (OS3) showing missing surface zone, with typical fibrillation. (c) Delaminated bovine cartilage strip (see Fig 8) shows fibrillated surface (arrow). (d) Typical OA plug from Fig 2, graded OS2, test results in Fig 9b. (Picrosirius Red for collagen.)

4.2.5. Laser Treatment

A femtosecond laser High-Q (High-Q Laser, Austria) oscillator with temporal pulse width of 99 fs

and 52 MHz repetition rate and output wavelength centered around 1060 nm was used for laser

treatment. The laser was coupled with a 3-axis translational stage (Thorlabs, Inc.) configured to

89

deliver laser pulses through a high numerical aperture objective (Figure 33). Cartilage explants

(Æ3 mm ´ 1 mm) were seated in a press-fit containing ring within a petri dish. A coverslip with a

focusing target was positioned on top of the sample and a sufficient amount of PBS was added

to the dish such that the sample remained hydrated throughout treatment. The sample dish was

positioned beneath the objective and the laser was then carefully focused within (10 um

resolution) on the target region of the sample. The sample remained stationary throughout

treatment as the objective rastered the treatment pattern, moving in all three dimensions. For

single-layer treatment a single x-y plane was rastered parallel to the articular surface. For multi-

layer treatment, the sample was treated layer-wise through the depth. For cartilage discs (Æ10

mm ´ 1 mm) laser treatment was administered in a similar fashion, except discs were glued to

petri dish, not press-fit.

Fig 33. Laser setup consisting of Hi-Q femtosecond oscillator coupled with the 3- axis motorized actuators. Beam delivery optics are mounted onto the motion system

90

4.2.6. Equilibrium Compressive Young’s modulus

To measure EY across the full thickness of an explant, we will use the devices and protocols for

unconfined compression stress-relaxation described in our previous studies [47, 83] and Section

3.3.3.

4.2.7. Wear tests Wear tests were performed under both SCA and MCA configurations as described (Section

1.5). The method of MCA was described in (Section 2.3.3), tester A was used. For SCA, tester

A was also used and affixed with an impermeable flat glass platen that is much larger than the

sample (instead of the glass lens under MCA) (Figure 1). The flat glass platen is used to

perform reciprocal sliding and it remains in constant contact with the entire surface of the

cartilage plug during the duration of the test. This set-up is designed to intentionally defeat the

sustained interstitial fluid pressurization that normally exists between two native migrating

articular layers (Section 1.5). Under the SCA configuration, the cartilage plug is press-fit into a

shallow recess of the same diameter within a PBS-filled dish that is affixed to a translating stage

that performs reciprocal sliding (±5 mm at 1 mm/s). The counterface is slowly lowered until

brought in contact with the plug and a designated load is applied. Contact stress can be applied

as low as 0.18 MPa with this device and as high as 18 MPa. For control samples (Æ3 mm ´ 1

mm) wear was observed under the prescribed contact stress (0.18 MPa) and reciprocal sliding

(±5 mm at 1 mm/s) after 4 h (N=1,440).

4.2.8. Cell viability & histology Cell viability was assessed with live/dead staining (DNA [220]) under confocal imaging. Samples

for histology were fixed, sectioned, and stained with Safranin-O [221] for charged PGs and with

Picrosirius red [222].

91

4.3. Results

4.3.1. Results Study 1: Laser treatment immature bovine at different depths

Results showed that 1 and 2 passes enhanced EY by nearly 100%, producing a significantly

higher EY than the 0-pass control group (p<0.03). This enhancement was equally effective at

the two ranges of depths from the articular surface (p=0.47, Figure 34A). The wear resistance

was also examined in a subset of samples from Study 1 (n=4 per group). For control samples,

EY decreased significantly with wear testing, from 0.38±0.01 MPa down to 0.32±0.02 MPa

(p<0.03), exhibiting clear visual evidence of damage (Figure 35). In contrast, laser-treated

samples exhibited no significant difference in EY (p=0.13), going from 0.50±0.03 to 0.49±0.03

MPa before and after the wear test. These samples showed little visual evidence of damage

(Figure 35). Control discs subjected to the wear test showed a blister indicative of delamination

wear, whereas the laser-treated discs were damaged (Figure 38).

Figure 34: Stiffening of articular cartilage with laser treatment: (a) immature bovine treated from 0-200 µm and 100-300 µm from surface, (b) mature human OA cartilage with Outerbridge scores 1 and 2, treated from 0-200 µm from surface (PS2), *p<0.005, † p<0.05.

92

4.3.2. Results Study 2: Laser treatment OA human

For control samples, no differences were found before and after sham treatment (p=0.34,

EY=0.63±0.43 MPa for OS1, EY=0.14±0.05 MPa for OS2), therefore we only summarize EY

values before and after laser treatment (Figure 34B) for OS1 and OS2 (two-factor design, 1:

before/after, repeated measures, 2: OS1/OS2). We found that laser treatment significantly

enhanced EY for both OS1 (p=0.04) and OS2 (p=0.004).

Figure 35: Control bovine cartilage explant (Æ3 mm ´ 1 mm), before and after wear test, demonstrates greater visual evidence of surface damage (arrows) compared to laser-treated explant (side view).

4.3.3. Results Study 3: Cell viability in long-term culture

The compressive stiffness was confirmed to exhibit a similar increase with laser treatment

(p<0.03), from EY=0.36±0.21 MPa (0 pass) to EY=0.58±0.21 MPa (1 pass). Confocal images of

explants demonstrated no evidence of viability loss from laser treatment (Figure 36A,B).

Additional samples maintained for two weeks in culture after laser treatment, cell viability was

similarly maintained when compared to untreated controls (Figure 36C, D).

93

Figure 36: Live/Dead staining of immature bovine cartilage explants (PS3): (a) control and (b) laser-treated, at 24 h; (c) control and (d) laser-treated at 2 wks.

4.3.4. Results Study 4: Vitamin C for sequestration of ROS

It was found that lasing bovine cartilage bathed in PBS+1M AAP produced EY=0.37±0.19 MPa,

comparable to untreated cartilage (Figure 37) and significantly lower (p=0.04) than samples

treated in PBS only, EY=0.64±0.20 MPa (n=4 per group); thus, AAP can be used to actively

modulate ROS activity during laser treatment.

4.4. Discussion

Based on these results, we have shown that the novel laser treatment strategy that induces

crosslinks into the type I collagen network of corneal tissue [181, 182] can be extended to

induce collagen crosslinks to the more complex type II collagen network of articular cartilage.

94

We demonstrated crosslink induction via both mechanical property changes to the ECM of small

cartilage cylindrical plugs and large explant discs.

Figure 37: Young’s Modulus for samples after laser treatment in either PBS + Vitamin C or PBS alone.

In Study 1, we showed a doubling in stiffness of devitalized immature bovine explants

after laser treatment, independent of the number of passes (Figure 34). From these Study 1

results, we concluded that one pass was sufficient to saturate the collagen crosslinking sites

available for this modality and a doubling of the treatment dosage exhibited no mechanical

evidence of harmful effects. All subsequent tests were performed at the same energy level,

using one pass. In addition to enhancing the modulus EY, we examined the wear resistance of

control and laser-treated bovine explants, in a subset these samples from the study reported in

(n=4 per group). Explants were subjected to sliding against glass for 4 h, using the SCA

configuration (Figure 1A) [83]. From the results of Chapter 2, SCA can be used to accelerate

wear and yet still produce physiologic damage. Wear tests performed under the SCA

configuration at the designated applied stress reliably show swelling and softening consistent

with fatigue failure for immature bovine cartilage explants (Æ3 mm ´ 1.3 mm). Post-wear

mechanical tests indicated that laser-treated explants neither softened nor swelled like their

95

paired controls (Figure – swelling). These results are striking and indicate that not only has

laser-treatment enhanced the tissue compressive stiffness as measured under static stress-

relaxation measurements, but that laser-treatment has improved the tissue’s resistance to

dynamic wear.

Figure 38. LEFT: (a) Photographs and (b) laser surface scans of Æ10 mm X 1.2 mm bovine cartilage plugs after sliding for 12 h, 4 mm travel at 1 mm/s, against a spherical glass lens (R12.7 mm) under 4.45 N load. Arrows: blister due to delamination wear in control sample (left). Treated sample (right) was lased over a 3 mm x 3 mm central patch and showed no damage. RIGHT: Friction coefficient µ remains low over MCA 12 h wear test (µ=0.008 damaged control, µ=0.005 undamaged treated). Typical results from SCA 4 h test (Fig 9a) show much higher, time-dependent µ for both control and treated.

We further tested bovine cartilage discs under MCA, where laser treatment was only performed

at the center region, and compared against untreated controls, using the protocol described in

Chapter 2. Conditions known to induce delamination wear in immature bovine strips were

selected: G-PBS-12-L1 (Glass lens counterface, PBS as lubricant, 4.45 N applied load, for 12

hours). For the four samples tested (2 controls and 2 treated), the two controls showed

expected delamination wear (one in the form of a blister and the other in the form of a ruptured

bluster), whereas laser-treated samples showed no evidence of visual damage on surface laser

scans (Figure – laser scan). Representative friction curves are shown for both the MCA and

96

SCA wear tests of bovine cartilage. Plots of the friction coefficient the friction coefficient (Figure

38) remains comparable in magnitude between control and laser-treated cartilage, both for MCA

and for stationary contact area (SCA) as performed in PS1. These results suggest that the most

important functional properties of cartilage (sustaining compressive load and producing low

wear) are significantly enhanced by our novel laser treatment methodology, without detriment to

the friction coefficient. While SCA exhibits much higher friction than MCA due to loss of fluid

pressurization, eventually they both lead to similar delamination, justifying our use of the less

time-consuming SCA protocol here.

Figure 39: Envisioned treatment process: (1) Femtosecond laser irradiates articular cartilage creating an ionization field; (2) ionization generates reactive oxygen species (ROS) which interact with collagen fibrils in the ECM; (3) biochemical reactions with ROS result in CxL formation, which enhance cartilage mechanical properties, potentially slowing down OA progression.

In Study 2, we found that laser treatment significantly enhanced EY for both OS1

(p=0.04) and OS2 (p=0.004); most interestingly, OS2 samples achieved the same final EY value

as OS1 samples (EY=0.98±0.36 MPa, p=0.95), suggesting that the laser treatment may be

97

saturating available cross-linking sites, whose abundance increases with fibrillation. These

results establish that this laser treatment protocol is equally effective for cartilage from different

species and from different grades of osteoarthritic human cartilage. These results also suggest

that we could reduce the amount or duration of laser treatment for human tissue, since it

appears we have achieved saturated conditions.

The viability studies performed in Study 3 are highly encouraging. These results

demonstrate that laser treatment does not compromise cell viability for up to two weeks, while

achieving equally effective enhancement of the tissue compressive modulus as with devitalized

explants. These highly encouraging results will be examined more thoroughly in future studies

by culturing control and treated explants for up to 4 weeks post-treatment to confirm long-term

cell viability and enhancement of other properties, such as wear resistance, without

compromising tissue composition.

Ascorbic-acid-phosphate (AAP, Vitamin C) is a known scavenger of ROS commonly

used during laser-based procedures. Adverse effects of induced radicals can be minimized via

application of AAP during and post laser treatment. In Study 4, the ability of AAP to modulate

the potential harmful ROS effects from laser treatment was investigated. To test the efficacy of

AAP at interrupting ROS generation, cartilage explants were saturated with AAP prior to laser

treatment. Our working hypothesis is that our laser treatment induces short-lived bursts of non-

nitric ROS, primarily from ionization of the cartilage interstitial water, and these ROS then mostly

interact with collagen in the ECM. Therefore, if AAP was present at the time of ionization, it

would interrupt CxL formation and downstream tissue stiffening. Indeed, samples saturated in

AAP did not respond to laser treatment, suggesting that AAP may be used as an effective ROS

scavenger for this application.

Another observation that must be appreciated is that the prior work identifying

delamination wear under SCA and MCA have been extended here as useful tools. The ability to

98

assess tissue wear resilience is only possible because of the ability to impart and assess

reliable, repeatable, physiologic wear. Therefore, the role of the test bed must be appreciated

as it enabled the important findings reported here.

4.5. Future Work

Currently, our technology allows us to treat a Æ3 mm region in 5 min. To convert our current ex

situ setup to a practical tool for in situ laser treatment, our collaborators have begun

development of a fiber optic-based probe for future in vivo arthroscopic studies (Figure 40).

Clinical and engineering considerations will be taken into account, to constrain the laser

procedure to last no longer than common cartilage debridement (e.g., 30 minutes or less). To

meet this constraint, the focusing tip of the probe must be designed such that the required

energy density is deployed at numerical aperture values that accommodate a relatively large

focal volume. Since the focal volume has the shape of a Gaussian ellipsoid [223] with both radii

being a strong function of numerical aperture, relatively loose focusing will result in a large

depth of field. This favorable characteristic increases the penetration depth. However, the laser

intensity within the focal volume also follows a Gaussian distribution, thus care must be taken

not to exceed the upper threshold identified in SA2 at the center of the ellipsoid. The successful

completion of SA3 will let us proceed with future arthroscopic studies in animals and, potentially,

in humans.

99

Fig 40. (a) Laser setups for fiber optic probe (red) & free-space optics (blue). A. Femtosecond laser; B. Tissue sample area; C. 3-axis translation mount; D. Fiber optic cable; E. Probe tip; F. Focusing optics; G. 3-axis motorized actuator. Gaussian laser beam profile with ~60 mW power is obtained from (b) free-space objective and (c) fiber-based probe.

4.6. Conclusions Current OA treatments focus on relieving pain and improving joint function. This is

achieved through a combination of pain medications, steroids, injections, physical therapy,

arthroscopic procedures and joint replacement surgery as a last resort. None of the current

options act directly to reverse or delay OA, instead they act to reduce pain and improve mobility.

Even emerging technologies that target cartilage defects are offering short-term pain relieving

options that do not interrupt disease progression. The proposed technology would be the first

OA disease-modifying treatment of its kind that could be used in tandem with existing

treatments and could delay the need for a joint replacement. In this dissertation through in vitro

100

experiments we showed the potential for this therapy to prevent wear as it applies to both

immature bovine explants and OA human cartilage.

101

REFERENCES

1. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 390(10100): p. 1211-59. 2. Ashkavand Z, Malekinejad H, and Vishwanath BS, The pathophysiology of osteoarthritis. Journal of Pharmacy Research, 2013. 7(1): p. 132-138. 3. Nlo., H., Osteoarthritis. National Institute of Arthritis and Musculoskeletal and Skin Diseases. 4. Johnson VL and Hunter DJ, The epidemiology of osteoarthritis. Best Pract Res Clin Rheumatol, 2014. 28(1): p. 5-15. 5. Dijkgraaf LC, et al., Normal cartilage structure, biochemistry, and metabolism: a review of the literature. Journal of oral and maxillofacial surgery 1995. 53: p. 924-929. 6. Loeser RF, Osteoarthritis year in review 2013: biology. Osteoarthritis Cartilage, 2013. 21(10): p. 1436-42. 7. Andriacchi TP, et al., A Framework for the in Vivo Pathomechanics of Osteoarthritis at the Knee. Annals of Biomedical Engineering, 2004. 32(3): p. 447-457. 8. Sihvonen, R., et al., Arthroscopic partial meniscectomy versus placebo surgery for a degenerative meniscus tear: a 2-year follow-up of the randomised controlled trial. Ann Rheum Dis, 2017. 9. Sihvonen, R., et al., Arthroscopic partial meniscectomy versus sham surgery for a degenerative meniscal tear. N Engl J Med, 2013. 369(26): p. 2515-24. 10. Xing, L., et al., Microfracture combined with osteochondral paste implantation was more effective than microfracture alone for full-thickness cartilage repair. Knee Surg Sports Traumatol Arthrosc, 2013. 21(8): p. 1770-6. 11. Bert JM and Maschka K, The Arthroscopic Treatment of Unicompartmental Gonarthrosis: A Five-Year Follow-up Study of Abrasion Arthroplasty Plus Arthroscopic Debridement and Arthroscopic Debridement Alone. The Journal of Arthroscopic and Related Surgery. 5(1): p. 25-32. 12. Ciccotti, M.C., et al., The prevalence of articular cartilage changes in the knee joint in patients undergoing arthroscopy for meniscal pathology. Arthroscopy, 2012. 28(10): p. 1437-44. 13. Curl WW, et al., Cartilage Injuries: a review of 31,516 knee arthroscopies. Arthroscopy: The Journal of Arthroscopic and Related Surgery, 1997. 13(4): p. 456-460. 14. Hjelle, K., et al., Articular cartilage defects in 1,000 knee arthroscopies. Arthroscopy: The Journal of Arthroscopic & Related Surgery, 2002. 18(7): p. 730-734. 15. Mithoefer, K., et al., Clinical efficacy of the microfracture technique for articular cartilage repair in the knee: an evidence-based systematic analysis. Am J Sports Med, 2009. 37(10): p. 2053-63. 16. Albright, J.C. and A.K. Daoud, Microfracture and Microfracture Plus. Clin Sports Med, 2017. 36(3): p. 501-507. 17. Goyal, D., et al., Evidence-based status of microfracture technique: a systematic review of level I and II studies. Arthroscopy, 2013. 29(9): p. 1579-88.

102

18. Kon, E., et al., Articular cartilage treatment in high-level male soccer players: a prospective comparative study of arthroscopic second-generation autologous chondrocyte implantation versus microfracture. Am J Sports Med, 2011. 39(12): p. 2549-57. 19. Jevsevar D, et al., Viscosupplementation for Osteoarthritis of the Knee: A Systematic Review of the Evidence. J Bone Joint Surg Am, 2015. 97(24): p. 2047-60. 20. Mow VC, R.A., Poole AR, Cartilage and diarthrodial joints as paradigms for hierarchical materials and structures. Biomaterials, 1992. 13(2): p. 67-97. 21. RA, S., Biology of Cartilage Cells. 1979, Cambridge, UK: Cambridge University Press. 22. Ropes MW, B.W., Synovial fluid changes in joint diseases. 1953, Cambridge, MA: Harvard University Press. 23. Lai WM, H.J., Mow VC, A triphasic theory for the swelling and deformation behaviors of articular cartilage. J Biomech Eng. , 1991. 113(3): p. 245-58. 24. Linn FC, S.L., Movement and Composition of interstitial fluid of Cartilage. Arthritis & Rheumatism, 1965. 8: p. 481-494. 25. A, M., Physicochemical properties of articular cartilage in Adult articular cartilage, F. MAR, Editor. 1979, Pitman Medical Publishing: Kent, UK. p. 215-290. 26. Mow VC, K.S., Lai WM, Armstrong CG., Biphasic creep and stress relaxation of articular cartilage in compression: Theory and experiments. J Biomech Eng. , 1980. 102(1): p. 73-84. 27. Maroudas A, B.M., Venn MF, Further studies on the composition of human femoral head cartilage. Ann Rheum Dis. 1980 Oct;39(5):514-23., 1980. 39(5): p. 514-23. 28. Mow VC and Huiskes R, Basic orthopaedic biomechanics & mechano-biology. Lubrication and wear of diarthrodial joints: Ateshian GA & Mow VC. 2005, Philadelphia, Pa: Lippincott Williams & Wilkins. 29. Mitrovic D, Q.M., Stankovic A, Ryckewaert A., Cell density of adult human femoral condylar articular cartilage. Joints with normal and fibrillated surfaces. Lab Invest. 49(3):309-16., 1983. 49(3): p. 309-16. 30. Lipshitz H, E.R.r., Glimcher MJ, Changes in the hexosamine content and swelling ratio of articular cartilage as functions of depth from the surface. J Bone Joint Surg Am. , 1976. 58(8): p. 1149-53. 31. Muir H, B.P., Maroudas A, The distribution of collagen in human articular cartilage with some of its physiological implications. J Bone Joint Surg Br., 1970. 52(3): p. 554-63. 32. Franzén A, I.S., Hejderup SO, Heinegård D., Variations in the composition of bovine hip articular cartilage with distance from the articular surface. Biochem J. 1981 Jun 1;195(3):535-43., 1981. 95(3): p. 535-43. 33. Aspden RM, Y.Y., Hukins DW., Collagen orientations in the meniscus of the knee joint. J Anat., 1985. 140(3): p. 371-380. 34. ND, B., Structural consequences of traumatizing articular cartilage. Ann Rheum Dis., 1986. 45(3): p. 225-34. 35. Broom ND, M.D., Ultrastructural evidence for fibril-to-fibril associations in articular cartilage and their functional implication. J Anat. , 1986. 146: p. 185-200. 36. Broom ND, S.-R.H., Collagen-collagen versus collagen-proteoglycan interactions in the determination of cartilage strength. Arthritis Rheum., 1990. 33(10): p. 1512-7. 37. JM., C., The organization of collagen in cryofractured rabbit articular cartilage: a scanning electron microscopic study. J Orthop Res. , 1985. 3(1): p. 17-29. 38. IC., C., Articular cartilage: a review and scanning electron microscope study. 1. The interterritorial fibrillar architecture. J Bone Joint Surg Br., 1971. 53(4): p. 732-50. 39. Lane JM, W.C., Review of articular cartilage collagen research. Arthritis Rheum., 1975. 18(6): p. 553-62.

103

40. Broom, N., Marra, D., New Structural concepts of articular cartilage demonstrated with a physical model. Connect Tissue Res., 1985. 14: p. 1-8. 41. Wang, C.C.B., et al., An Automated Approach for Direct Measurement of Two-Dimensional Strain Distributions Within Articular Cartilage Under Unconfined Compression. Journal of Biomechanical Engineering, 2002. 124(5): p. 557. 42. Wang, C.C., Guo, X.E., Sun, D., Mow, V.C., Ateshian, G.A., Hung, C.T., The functional environment of chondrocytes within cartilage subjected to com- pressive loading: a theoretical and experimental approach. Biorheology, 2002. 39: p. 11-25. 43. Park, S., Hung, C.T., Ateshian, G.A., Mechanical response of bovine articular cartilage under dynamic unconfined compression loading at physiological stress levels. . Osteoarthritis Cartilage, 2004. 12: p. 65-73. 44. Kempson, G.E., Freeman, M.A., Swanson, S.A., 1968, Tensile properties of articular cartilage. Nature, 1968. 220: p. 1127-1128. 45. Williamson, A.K., Masuda, K., Thonar, E.J., Sah, R.L., , Growth of immature articular cartilage in vitro: correlated variation in tensile biomechanical and collagen network properties. 9, 625–634. Tissue Engineering 2003. 9: p. 625-634. 46. Williamson, A.K., Chen, A.C., Masuda, K., Thonar, E.J., Sah, R.L.,, Tensile mechanical properties of bovine articular cartilage: variations with growth and relationships to collagen network components. Journal of Orthopaedic Research, 2003. 21: p. 872-880. 47. Soltz MA and Ateshian GA, A Conewise Linear Elasticity mixture model for the analysis of tension- compression nonlinearity in articular cartilage. J Biomech Eng, 2000. 122: p. 576-586. 48. Cohen NP, F.R., Mow VC., 1998 Oct;28(4):203-15. Composition and dynamics of articular cartilage: structure, function, and maintaining healthy state. J Orthop Sports Phys Ther. , 1998. 28(4): p. 203-15. 49. Soulhat J, B.M., Shirazi-Adl A., A fibril-network-reinforced biphasic model of cartilage in unconfined compression. J Biomech Eng. 1999 Jun;121(3):340-7., 1999. 121(3): p. 340-7. 50. Huang CY, S.M., Kopacz M, Mow VC, Ateshian GA., Experimental verification of the roles of intrinsic matrix viscoelasticity and tension-compression nonlinearity in the biphasic response of cartilage. J Biomech Eng. 2003 Feb;125(1):84-93., 2003. 125(1): p. 84-93. 51. Huang CY, M.V., Ateshian GA., The role of flow-independent viscoelasticity in the biphasic tensile and compressive responses of articular cartilage. J Biomech Eng., 2001. 123(5): p. 410-7. 52. Chahine NO, W.C., Hung CT, Ateshian GA., Anisotropic strain-dependent material properties of bovine articular cartilage in the transitional range from tension to compression. J Biomech. 2004, 2004. 37(8): p. 1251-61. 53. Krishnan, R., M. Kopacz, and G.A. Ateshian, Experimental verification of the role of interstitial fluid pressurization in cartilage lubrication. . Journal of Orthopaedic Research, 2004. 22(3): p. 565-70. 54. McCutchen CW, The frictional properties of animal joints. Wear, 1962. 5: p. 1-17. 55. GA, A., A theoretical formulation for boundary friction in articular cartilage. J Biomech Eng., 1997. 119(1): p. 81-6. 56. Forster H, F.J., The influence of loading time and lubricant on the friction of articular cartilage. Proc Inst Mech Eng H. 1996;210(2):109-19., 1996. 210(2): p. 109-19. 57. Forster H and F. J;, The influence of continuous sliding and subsequent surface wear on the friction of articular cartilage. Proc. Inst. Mech. Eng. [H], 1999. 213: p. 329-345. 58. Park S, K.R., Nicoll SB, Ateshian GA., Cartilage interstitial fluid load support in unconfined compression. J Biomech, 2003. 36(12): p. 1785-96.

104

59. Soltz, M.A.a.G.A.A., Experimental verification and theoretical prediction of cartilage interstitial fluid pressurization at an impermeable contact interface in confined compression. . Journal of Biomechanics, 1998. 31(10): p. 927-934. 60. Caligaris, M., et al., Investigation of the frictional response of osteoarthritic human tibiofemoral joints and the potential beneficial tribological effect of healthy synovial fluid. Osteoarthritis Cartilage, 2009. 17(10): p. 1327-32. 61. Charnley, J., The Lubrication of Animal Joints in Relation to Surgical Reconstruction by Arthroplasty. Annals of the Rheumatic Diseases, 1960. 19(1): p. 10-19., 1960. 19(1): p. 10-19. 62. Gleghorn, J.P., et al.,, Boundary mode lubrication of articular cartilage by recombinant human lubricin. J Orthop Res., 2009. 27(6): p. 771-7. 63. Caligaris, M. and G.A. Ateshian, Effects of sustained interstitial fluid pressurization under migrating contact area, and boundary lubrication by synovial fluid, on cartilage friction. Osteoarthritis Cartilage, 2008. 16(10): p. 1220-7. 64. Schmidt, T.A., et al., Boundary lubrication of articular cartilage: role of synovial fluid constituents. Arthritis Rheum, 2007. 56(3): p. 882-91. 65. Walker, P.S., et al., "Boosted lubrication" in synovial joints by fluid entrapment and enrichment. Ann Rheum Dis, 1968. 27(6): p. 512-20. 66. Bell, C.J., E. Ingham, and J. Fisher, Influence of hyaluronic acid on the time-dependent friction response of articular cartilage under different conditions. Proc Inst Mech Eng [H], 2006. 220(1): p. 23-31. 67. Swann, D.A., et al., The lubricating activity of human synovial fluids. Arthritis Rheum, 1984. 27(5): p. 552-6. 68. Swann, D.A. and E.L. Radin, The molecular basis of articular lubrication. I. Purification and properties of a lubricating fraction from bovine synovial fluid. J Biol Chem, 1972. 247(24): p. 8069-73. 69. Jay, G.D., K. Haberstroh, and C.J. Cha, Comparison of the boundary-lubricating ability of bovine synovial fluid, lubricin, and Healon. J Biomed Mater Res, 1998. 40(3): p. 414-8. 70. 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, 2001. 19(4): p. 677-87. 71. Schumacher, B.L., et al., Proteoglycan 4 (PRG4) synthesis and immunolocalization in bovine meniscus. J Orthop Res, 2005. 23(3): p. 562-8. 72. Hills, B.A., Oligolamellar lubrication of joints by surface active phospholipid. J Rheumatol, 1989. 16(1): p. 82-91. 73. Schwarz, I.M. and B.A. Hills, Surface-active phospholipid as the lubricating component of lubricin. Br J Rheumatol, 1998. 37(1): p. 21-6. 74. Forsey, R.W., et al., The effect of hyaluronic acid and phospholipid based lubricants on friction within a human cartilage damage model. Biomaterials, 2006. 27(26): p. 4581-90. 75. Oswald ES, P.-H.G., Bulinski JC, Ateshian GA, Hung CT, Dependence of zonal chondrocyte water transport properties on osmotic environment. Cell Mol Bioeng, 2008. 1(4): p. 339-348. 76. Krishnan, R., et al., Removal of the superficial zone of bovine articular cartilage does not increase its frictional coefficient. Osteoarthritis Cartilage, 2004. 12(12): p. 947-55. 77. Ateshian, G.A., H.Q. Wang, and W.M. Lai, The role of interstitial fluid pressurization and surface porosities on the boundary friction of articular cartilage. Journal of Tribology-Transactions of the Asme, 1998. 120(2): p. 241-248. 78. Soltz, M.A.a.G.A.A., Interstitial fluid pressurization during confined compression cyclical loading of articular cartilage. . Ann Biomed Eng, 2000. 28(2): p. 150-9.

105

79. Moore DF, Principles and applications of tribology 1st ed. 1975, Oxford, New York: Pergamon Press. 80. Katta J, et al., Biotribology of articular cartilage--a review of the recent advances. Med Eng Phys, 2008. 30(10): p. 1349-63. 81. Lee DW, Banquy X, and Israelachvili JN, Stick-slip friction and wear of articular joints. Proc Natl Acad Sci U S A, 2013. 110(7): p. E567-74. 82. Katta J, et al., Effect of nominal stress on the long term friction, deformation and wear of native and glycosaminoglycan deficient articular cartilage. Osteoarthritis Cartilage, 2009. 17(5): p. 662-8. 83. Oungoulian SR, et al., Wear and damage of articular cartilage with friction against orthopedic implant materials. Journal of Biomechanics, 2015. 48: p. 1957-1964. 84. H;, F. and F. J;, The influence of loading time and lubricant on the friction of articular cartilage. Proc. Inst. Mech. Eng. [H], 1996. 210: p. 109-119. 85. Oungoulian, S.R., et al., Articular cartilage wear characterization with a particle sizing and counting analyzer. J Biomech Eng, 2013. 135(2): p. 024501. 86. Lipshitz H, E.R.r., Glimcher MJ., In vitro wear of articular cartilage. J Bone Joint Surg Am. , 1975. 57(4): p. 527-34. 87. Collins DH, The Pathology of Articular and Spinal Disease. 1949, London, UK: Edward Arnold. 75-90. 88. Meachim G, Age-related degeneration of patellar articular cartilage. J. Anat., 1982. 134: p. 365-371. 89. Pritzker, K.P., et al., Osteoarthritis cartilage histopathology: grading and staging. Osteoarthritis Cartilage, 2006. 14(1): p. 13-29. 90. Champion L, Miranda V, and McNally SA, Dislocation after revision of hemiarthroplasty to total hip replacement. Injury, 2004. 35(2): p. 161-164. 91. Haidukewych GJ, Israel TA, and Berry DJ, Long-term survivorship of cemented bipolar hemiarthroplasty for fracture of the femoral neck. Clinical orthopaedics and related research, 2002. 403: p. 118-126. 92. Hasan SS, et al., Characteristics of unsatisfactory shoulder arthroplasties. Journal of Shoulder and Elbow Surgery, 2002. 11(5): p. 431-441. 93. Parsons, I.M., P.J. Millett, and J.J.P. Warner, Glenoid Wear after Shoulder Hemiarthroplasty. Clinical Orthopaedics and Related Research, 2004. 421: p. 120-125. 94. Sperling JW, C.R., Steinmann SP, Shoulder arthroplasty for osteoarthritis secondary to glenoid dysplasia. The Journal of bone and joint surgery. American volume, 2002. 84-A: p. 541-546. 95. V;, R., et al., Hemiarthroplasties Defeat Interstitial Fluid Pressurization in Cartilage and Promote Greater Friction than Natural Joints. In: 2010 Annual Meeting of the Orthopaedic Research Society, vol. 35. New Orleans, LA(ORS 2010). 96. Hosseini SM, V.M., Ito K, van Donkelaar CC, Is collagen fiber damage the cause of early softening in articular cartilage? . Osteoarthritis and cartilage / OARS, Osteoarthritis Research Society, 2013. 21(1): p. 136-143. 97. Hosseini SM, W.W., Ito K, van Donkelaar CC, A numerical model to study mechanically induced initiation and progression of damage in articular cartilage. Osteoarthritis and cartilage / OARS, Osteoarthritis Research Society, 2014. 22(1): p. 95-103. 98. Landinez-Parra NS, G.-A.D., Vanegas-Acosta JC, A phenomenological mathematical model of the articular cartilage damage. Computer methods and programs in biomedicine, 2011. 104(3): p. 58-74. 99. Simon WH, M.A., Spirt A., The effect of shear fatigue on bovine articular cartilage. J Orthop Res., 1990. 8(1): p. 86-93.

106

100. Weightman B, C.D., Jenkins EA., A second study of tensile fatigue properties of human articular cartilage. Ann Rheum Dis., 1978. 37(1): p. 58-63. 101. Buckley MR, et al., High-resolution spatial mapping of shear properties in cartilage. J Biomech, 2010. 43(4): p. 796-800. 102. Buckley MR, Bonassar LJ, and Cohen I, Localization of viscous behavior and shear energy dissipation in articular cartilage under dynamic shear loading. J Biomech Eng, 2013. 135(3): p. 31002. 103. Thambyah, A. and N. Broom, On how degeneration influences load-bearing in the cartilage-bone system: a microstructural and micromechanical study. Osteoarthritis Cartilage, 2007. 15(12): p. 1410-23. 104. Clarke IC, Articular cartilage: a review and scanning electron microscope study. 1. The interterritorial fibrillar architecture. J. Bone Jt. Surg., 1971. Br. Vol. 53: p. 732–750. 105. Xia, Y., Averaged and depth-dependent anisotropy of articular cartilage by microscopic imaging. Semin Arthritis Rheum, 2008. 37(5): p. 317-27. 106. Sah RL, D.J., Grodzinsky AJ, Plaas AH, Sandy JD., Effects of compression on the loss of newly synthesized proteoglycans and proteins from cartilage explants. Arch Biochem Biophys. 1991 Apr;286(1):20-9., 1991. 286(1): p. 20-9. 107. Ratcliffe A, B.P., Saed-Nejad F., Differential levels of synovial fluid aggrecan aggregate components in experimental osteoarthritis and joint disuse. J Orthop Res., 1994. 12(3): p. 464-73. 108. Parkkinen JJ1, I.J., Lammi MJ, Laakkonen J, Tammi M, Helminen HJ., Effects of cyclic hydrostatic pressure on proteoglycan synthesis in cultured chondrocytes and articular cartilage explants. Arch Biochem Biophys, 1993. 300(1): p. 458-65. 109. Palmoski MJ, B.K., Running inhibits the reversal of atrophic changes in canine knee cartilage after removal of a leg cast. Arthritis Rheum., 1981. 24(11): p. 1329-37. 110. Guilak F, M.B., Ratcliffe A, Mow VC., The effects of matrix compression on proteoglycan metabolism in articular cartilage explants. Osteoarthritis Cartilage, 1994. 2(2): p. 91-101. 111. Gray ML, P.A., Grodzinsky AJ, Lee RC., Mechanical and physiochemical determinants of the chondrocyte biosynthetic response. J Orthop Res, 1988. 6(6): p. 777-92. 112. Neu CP, K.K., Reddi AH., The interface of functional biotribology and regenerative medicine in synovial joints. Tissue Eng Part B Rev., 2008. 14(3): p. 235-47. 113. AAOS, Press Release: 2.5 Million Americans Living with an Artificial Hip, 4.7 Million with an Artificial Knee, 2014. 114. Abrams, G.D., et al., BioCartilage: Background and Operative Technique. Operative Techniques in Sports Medicine, 2013. 21(2): p. 116-124. 115. Vanlauwe, J., et al., Five-year outcome of characterized chondrocyte implantation versus microfracture for symptomatic cartilage defects of the knee: early treatment matters. Am J Sports Med, 2011. 39(12): p. 2566-74. 116. Flannery CR, et al., Prevention of cartilage degeneration in a rat model of osteoarthritis by intraarticular treatment with recombinant lubricin. Arthritis Rheum, 2009. 60(3): p. 840-7. 117. Jay GD, et al., Association between friction and wear in diarthrodial joints lacking lubricin. Arthritis Rheum, 2007. 56(11): p. 3662-9. 118. Soltz MA, B.I., Ateshian GA., Hydrostatic pressurization and depletion of trapped lubricant pool during creep contact of a rippled indenter against a biphasic articular cartilage layer. J Biomech Eng. , 2003. 125(5): p. 585-93. 119. Ateshian, G.A., and Mow, V. C., , Lubrication and wear of diarthrodial joints, in Basic orthopaedic biomechanics & mechano-biology, a.R.H. V. C. Mow, Editor. 2005, Lippincott Williams & Wilkins: Philadelphia, Pa.; London.

107

120. Caligaris M, A.G., Effects of sustained interstitial fluid pressurization under migrating contact area, and boundary lubrication by synovial fluid, on cartilage friction. Osteoarthritis Cartilage , 2008. 16(10): p. 1220-7. 121. Teeple E, E.K., Jay GD, Zhang L, Badger GJ, Akelman M, Bliss TF, Fleming BC., 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., 2011. 39(1): p. 164-172. 122. Jones BK, et al., The friction coefficient of shoulder joints remains remarkably low over 24 h of loading. J Biomech, 2015. 48(14): p. 3945-9. 123. Canal CE, Hung CT, and Ateshian GA, Two-dimensional strain fields on the cross-section of the bovine humeral head under contact loading. J Biomech, 2008. 41(15): p. 3145-51. 124. Krishnan R, M.E., Ateshian GA., Effect of dynamic loading on the frictional response of bovine articular cartilage. J Biomech. 2005 Aug;38(8):1665-73., 2005. 38(8): p. 1665-73. 125. Grodzinsky, A.J., et al., Cartilage tissue remodeling in response to mechanical forces. Annu Rev Biomed Eng, 2000. 2: p. 691-713. 126. Kurz, B., et al., Biosynthetic response and mechanical properties of articular cartilage after injurious compression. J Orthop Res, 2001. 19(6): p. 1140-6. 127. Lay, E., et al., Short- and long-term exposure of articular cartilage to curcumin or quercetin inhibits aggrecan loss. J Nutr Biochem, 2012. 23(2): p. 106-12. 128. Vinardell, T., et al., Chondrogenesis and integration of mesenchymal stem cells within an in vitro cartilage defect repair model. Ann Biomed Eng, 2009. 37(12): p. 2556-65. 129. Hunter, C.J. and M.E. Levenston, Maturation and integration of tissue-engineered cartilages within an in vitro defect repair model. Tissue Eng, 2004. 10(5-6): p. 736-46. 130. Dumont, J., et al., Mature full-thickness articular cartilage explants attached to bone are physiologically stable over long-term culture in serum-free media. Connect Tissue Res, 1999. 40(4): p. 259-72. 131. Osborn, K.D., S.B. Trippel, and H.J. Mankin, Growth factor stimulation of adult articular cartilage. J Orthop Res, 1989. 7(1): p. 35-42. 132. Asanbaeva, A., et al., Regulation of immature cartilage growth by IGF-I, TGF-beta1, BMP-7, and PDGF-AB: role of metabolic balance between fixed charge and collagen network. Biomech Model Mechanobiol, 2008. 7(4): p. 263-76. 133. Luyten, F.P., et al., Insulin-like growth factors maintain steady-state metabolism of proteoglycans in bovine articular cartilage explants. Arch Biochem Biophys, 1988. 267(2): p. 416-25. 134. Bian, L., et al., Mechanical and biochemical characterization of cartilage explants in serum-free culture. J Biomech, 2008. 41(6): p. 1153-9. 135. Balcom, N.T., et al., In vitro articular cartilage growth with sequential application of IGF-1 and TGF-beta1 enhances volumetric growth and maintains compressive properties. J Biomech Eng, 2012. 134(3): p. 031001. 136. Williams, G.M., et al., Differential regulation of immature articular cartilage compressive moduli and Poisson's ratios by in vitro stimulation with IGF-1 and TGF-beta1. J Biomech, 2010. 43(13): p. 2501-7. 137. Sah, R.L., et al., Differential effects of bFGF and IGF-I on matrix metabolism in calf and adult bovine cartilage explants. Arch Biochem Biophys, 1994. 308(1): p. 137-47. 138. Posever, J., F.M. Phillips, and L.A. Pottenger, Effects of basic fibroblast growth factor, transforming growth factor-beta 1, insulin-like growth factor-1, and insulin on human osteoarthritic articular cartilage explants. J Orthop Res, 1995. 13(6): p. 832-7.

108

139. Strehl, R., et al., Long-term maintenance of human articular cartilage in culture for biomaterial testing. Biomaterials, 2005. 26(22): p. 4540-9. 140. Eckstein, F., M. Muller-Gerbl, and R. Putz, Distribution of subchondral bone density and cartilage thickness in the human patella. J Anat, 1992. 180 ( Pt 3): p. 425-33. 141. Bian L, et al., Effects of dexamethasone on the functional properties of cartilage explants during long-term culture. Am J Sports Med, 2010. 38(1): p. 78-85. 142. DiMicco, M.A., et al., Mechanisms and kinetics of glycosaminoglycan release following in vitro cartilage injury. Arthritis Rheum, 2004. 50(3): p. 840-8. 143. Hascall, V.C., et al., The effect of serum on biosynthesis of proteoglycans by bovine articular cartilage in culture. Arch Biochem Biophys, 1983. 224(1): p. 206-23. 144. Nakamura, T., T. Kuroda, and K. Noguchi, Growth behaviour of condylar cartilage and epiphyseal cartilage on the different medium of the organ culture. Bull Tokyo Med Dent Univ, 1979. 26(2): p. 57-61. 145. Dechant, J.E., W.A. Symm, and J.E. Nieto, Comparison of pH, lactate, and glucose analysis of equine synovial fluid using a portable clinical analyzer with a bench-top blood gas analyzer. Vet Surg, 2011. 40(7): p. 811-6. 146. Ramakrishnan, P.S., et al., Repeated measurement of mechanical properties in viable osteochondral explants following a single blunt impact injury. Proc Inst Mech Eng H, 2011. 225(10): p. 993-1002. 147. Williamson, A.K., et al., Growth of immature articular cartilage in vitro: correlated variation in tensile biomechanical and collagen network properties. Tissue Eng, 2003. 9(4): p. 625-34. 148. Bian, L., et al., Effects of dexamethasone on the functional properties of cartilage explants during long-term culture. Am J Sports Med, 2010. 38(1): p. 78-85. 149. Rovensky, J., et al., Hormone concentrations in synovial fluid of patients with rheumatoid arthritis. Clin Exp Rheumatol, 2005. 23(3): p. 292-6. 150. Stegink, L.D., et al., Plasma amino acid concentrations and amino acid ratios in normal adults and adults heterozygous for phenylketonuria ingesting a hamburger and milk shake meal. Am J Clin Nutr, 1991. 53(3): p. 670-5. 151. Frame, E.G., The levels of individual free amino acids in the plasma of normal man at various intervals after a high-protein meal. J Clin Invest, 1958. 37(12): p. 1710-23. 152. Hagel, A.F., et al., Plasma concentrations of ascorbic acid in a cross section of the German population. J Int Med Res, 2017: p. 300060517714387. 153. Hardingham, T.E. and C.F. Phelps, Studies on the incorporation of (U-14C)glucose and (35S)sulphate into the acid glycosaminoglycans of neonatal rat skin. Biochem J, 1970. 119(5): p. 885-93. 154. Li, Z., et al., Glucose Transporter-4 Facilitates Insulin-Stimulated Glucose Uptake in Osteoblasts. Endocrinology, 2016. 157(11): p. 4094-4103. 155. Trippel, S.B., et al., Characterization of a specific somatomedin-c receptor on isolated bovine growth plate chondrocytes. Endocrinology, 1983. 112(6): p. 2128-36. 156. Steinberg, J.J. and C.B. Sledge, Synovial factors and chondrocyte-mediated breakdown of cartilage: inhibition by hydrocortisone. J Orthop Res, 1983. 1(1): p. 13-21. 157. Mankin, H.J., A. Zarins, and W.L. Jaffe, The effect of systemic corticosteroids on rabbit articular cartilage. Arthritis Rheum, 1972. 15(6): p. 593-9. 158. Albro, M.B., et al., Accumulation of exogenous activated TGF-beta in the superficial zone of articular cartilage. Biophys J, 2013. 104(8): p. 1794-804. 159. Albro, M.B., et al., Heterogeneous engineered cartilage growth results from gradients of media-supplemented active TGF-beta and is ameliorated by the alternative supplementation of latent TGF-beta. Biomaterials, 2016. 77: p. 173-85.

109

160. Katayama, M., et al., Age-dependent decline in cortisol levels and clinical manifestations in patients with ACTH-independent Cushing's syndrome. Clin Endocrinol (Oxf), 1998. 49(3): p. 311-6. 161. Takagi, Y. and B.T. Bjornsson, Cortisol inhibits glycosaminoglycan synthesis in cultured rainbow trout cartilage. Gen Comp Endocrinol, 1997. 108(1): p. 80-6. 162. Cantrill, H.L., et al., In vitro determination of relative corticosteroid potency. J Clin Endocrinol Metab, 1975. 40(6): p. 1073-7. 163. Martin, J.A., A.L. Wilkey, and R.A. Brand, Cartilage extracellular matrix metabolism differs in serum and synovial fluid. Methods Cell Sci, 2002. 24(4): p. 139-43. 164. Maroudas A, Balance between swelling pressure and collagen tension in normal and degenerate cartilage Nature, 1976. 260: p. 808-809. 165. Romgens, A.M., C.C. van Donkelaar, and K. Ito, Contribution of collagen fibers to the compressive stiffness of cartilaginous tissues. Biomech Model Mechanobiol, 2013. 12(6): p. 1221-31. 166. Garvican, E.R., et al., MMP-mediated collagen breakdown induced by activated protein C in equine cartilage is reduced by corticosteroids. J Orthop Res, 2010. 28(3): p. 370-8. 167. Tyler, J.A., H.B. Fell, and C.E. Lawrence, The effect of cortisol on porcine articular tissues in organ culture. J Pathol, 1982. 137(4): p. 335-51. 168. Schryver, H.F., The Influence of Hydrocortisone on the Water Uptake of Embryonic Chick Tibiotarsi in Organ Culture. Exp Cell Res, 1965. 37: p. 327-31. 169. Reynolds, J.J., The effect of hydrocortisone on the growth of chick bone rudiments in chemically defined medium. Exp Cell Res, 1966. 41(1): p. 174-89. 170. Billinghurst, R.C., et al., Enhanced cleavage of type II collagen by collagenases in osteoarthritic articular cartilage. J Clin Invest, 1997. 99(7): p. 1534-45. 171. Eleswarapu, S.V., D.J. Responte, and K.A. Athanasiou, Tensile properties, collagen content, and crosslinks in connective tissues of the immature knee joint. PLoS One, 2011. 6(10): p. e26178. 172. Yan, D., et al., The impact of low levels of collagen IX and pyridinoline on the mechanical properties of in vitro engineered cartilage. Biomaterials, 2009. 30(5): p. 814-21. 173. Lopes, J.M. and M. Larbier, Influence of amino-acid concentration in the culture medium on the rate of amino-acid incorporation into protein and amino-acid oxidation of cultured chicken hepatocytes. Reprod Nutr Dev, 1994. 34(2): p. 157-64. 174. Nims, R.J., et al., Matrix Production in Large Engineered Cartilage Constructs Is Enhanced by Nutrient Channels and Excess Media Supply. Tissue Eng Part C Methods, 2015. 21(7): p. 747-57. 175. Byers, B.A., et al., Transient exposure to transforming growth factor beta 3 under serum-free conditions enhances the biomechanical and biochemical maturation of tissue-engineered cartilage. Tissue Eng Part A, 2008. 14(11): p. 1821-34. 176. Cook, J.L., et al., A novel system improves preservation of osteochondral allografts. Clin Orthop Relat Res, 2014. 472(11): p. 3404-14. 177. Parkkinen, J.J., et al., Local stimulation of proteoglycan synthesis in articular cartilage explants by dynamic compression in vitro. J Orthop Res, 1992. 10(5): p. 610-20. 178. Sah, R.L., et al., Biosynthetic response of cartilage explants to dynamic compression. J Orthop Res, 1989. 7(5): p. 619-36. 179. Vogel A, et al., Low-density plasmas below the optical breakdown threshold: potential hazard for multiphoton microscopy, and a tool for the manipulation of intracellular events. Proc. of SPIE, Multiphoton Microscopy in the Biomedical Sciences, 2002. 4620: p. 202-216. 180. Shen N, et al., Ablation of cytoskeletal filaments and mitochondria in live cells using a femtosecond laser nanoscissor. Mech. Chem. Biosyst, 2005. 2: p. 17-25.

110

181. Chao Wang, M.F., Guanxiong Miao, Mariya Zyablitskaya and Sinisa Vukelic, Femtosecond laser crosslinking of the cornea for non-invasive vision correction. Nature Photonics, 2018. 182. Wang C, et al., A New Paradigm for Use of Ultrafast Lasers in Ophthalmology for Enhancement of Corneal Mechanical Properties and Permanent Correction of Refractive Errors. Proc. of SPIE, San Francisco, CA, 2017. 10066( Energy-based Treatment of Tissue and Assessment IX). 183. Dijkgraaf LC, et al., The Structure, Biochemistry, and Metabolism of 0s teoarthritic Cartilage: A Review of the Literature. J. Oral Maxillofac. Surg., 1995. 53: p. 1182-1192. 184. Asanbaeva A, et al., Cartilage growth and remodeling: modulation of balance between proteoglycan and collagen network in vitro with beta-aminopropionitrile. Osteoarthritis Cartilage, 2008. 16(1): p. 1-11. 185. DR;, E., Collagens and Cartilage Matrix Homeostasis. Clinical Orthopaedics and Related Research, 2004. 427: p. S118-S122. 186. Eyre, D., et al., Covalent cross-linking of the NC1 domain of collagen type IX to collagen type II in cartilage. J Biol Chem, 2004. 279(4): p. 2568-74. 187. Uebelhart D, et al., Elevation in urinary levels of pyridinium cross- links of collagen following chymopapain-induced degradation of articular cartilage in the rabbit knee provides evidence of metabolic changes in bone. . Osteoarthritis and Cartilage, 1993. 1: p. 185-192. 188. Takahashi M, et al., Quantitative analysis of crosslinks pyridinoline and pentosidine in articular cartilage of patients with bone and joint disorders. Arthritis & Rheumatism, 1994. 37: p. 724-728. 189. Huang, Z., et al., Effectiveness of low-level laser therapy in patients with knee osteoarthritis: a systematic review and meta-analysis. Osteoarthritis and Cartilage, 2015. 23(9): p. 1437-1444. 190. Alayat, M.S.M., et al., Efficacy of pulsed Nd: YAG laser in the treatment of patients with knee osteoarthritis: a randomized controlled trial. Lasers in Medical Science, 2017: p. 1-9. 191. Marcos, R.L., et al., Biomechanical and biochemical protective effect of low-level laser therapy for Achilles tendinitis. Journal of the mechanical behavior of biomedical materials, 2014. 29: p. 272-285. 192. Al Rashoud, A., et al., Efficacy of low-level laser therapy applied at acupuncture points in knee osteoarthritis: a randomised double-blind comparative trial. Physiotherapy, 2014. 100(3): p. 242-248. 193. Chung, H., et al., The nuts and bolts of low-level laser (light) therapy. Annals of biomedical engineering, 2012. 40(2): p. 516-533. 194. Meister, J., et al., Ablation of articular cartilage with an erbium: YAG laser: an ex vivo study using porcine models under real conditions—ablation measurement and histological examination. Lasers in surgery and medicine, 2009. 41(9): p. 674-685. 195. Li, C., et al., Analysis of Nd: YAG laser�mediated thermal damage in rabbit nasal septal cartilage. Lasers in surgery and medicine, 2007. 39(5): p. 451-457. 196. Sobol, E., et al., Laser reshaping of cartilage. Biotechnology and Genetic Engineering Reviews, 2000. 17(1): p. 553-578. 197. Wright, R., et al., Shape retention in porcine and rabbit nasal septal cartilage using saline bath immersion and Nd: YAG laser irradiation. Lasers in surgery and medicine, 2005. 37(3): p. 201-209. 198. Su, E., et al., Preclinical investigations of articular cartilage ablation with femtosecond and pulsed infrared lasers as an alternative to microfracture surgery. Journal of biomedical optics, 2014. 19(9): p. 098001-098001.

111

199. McLaren, A., et al., Arthroscopic debridement of the knee for osteoarthrosis. Canadian journal of surgery. Journal canadien de chirurgie, 1991. 34(6): p. 595-598. 200. Linschoten, N. and C.A. Johnson, Arthroscopic debridement of knee joint arthritis: effect of advancing articular degeneration. Journal of the Southern Orthopaedic Association, 1996. 6(1): p. 25-36. 201. Bert, J., Role of abrasion arthroplasty and debridement in the management of osteoarthritis of the knee. Rheumatic diseases clinics of North America, 1993. 19(3): p. 725. 202. Hubbard, M., Arthroscopic surgery for chondral flaps in the knee. Bone & Joint Journal, 1987. 69(5): p. 794-796. 203. Kaplan, L.D., et al., Temperature Requirements for Altering the Morphology of Osteoarthritic and Nonarthritic Articular Cartilage In Vitro Thermal Alteration of Articular Cartilage. The American journal of sports medicine, 2004. 32(3): p. 688-692. 204. Kim, H., M. Moran, and R.B. Salter, The potential for regeneration of articular cartilage in defects created by chondral shaving and subchondral abrasion. An experimental investigation in rabbits. J Bone Joint Surg Am, 1991. 73(9): p. 1301-1315. 205. Janzen, D.L., et al., Osteonecrosis after contact neodymium: yttrium aluminum garnet arthroscopic laser meniscectomy. AJR. American journal of roentgenology, 1997. 169(3): p. 855-858. 206. Garino, J.P., et al., Osteonecrosis of the knee following laser-assisted arthroscopic surgery: a report of six cases. Arthroscopy: The Journal of Arthroscopic & Related Surgery, 1995. 11(4): p. 467-474. 207. Rozbruch, S.R., et al., Osteonecrosis of the knee following arthroscopic laser meniscectomy. Arthroscopy: The Journal of Arthroscopic & Related Surgery, 1996. 12(2): p. 245-250. 208. Athanasiou, K.A., et al., Effects of excimer laser on healing of articular cartilage in rabbits. Journal of orthopaedic research, 1995. 13(4): p. 483-494. 209. Fischer, R., et al., Thermal side effects of fiber-guided XeCl excimer laser drilling of cartilage. Lasers in surgery and medicine, 1994. 14(3): p. 278-286. 210. Hovhannisyan V, et al., Dynamics of femtosecond laser photo-modification of collagen fibers. Optics express, 2008. 16: p. 7958-7968. 211. Schaffer, C.B., A.O. Jamison, and E. Mazur, Morphology of femtosecond laser-induced structural changes in bulk transparent materials. Applied Physics Letters, 2004. 84(9): p. 1441-1443. 212. Meek KM and Hayes S, Corneal cross-linking--a review. Ophthalmic Physiol Opt, 2013. 33(2): p. 78-93. 213. Shen H-R, et al., Photodynamic cross-linking of proteins: V. Nature of the tyrosine–tyrosine bonds formed in the FMN-sensitized intermolecular cross-linking of N-acetyl-L- tyrosine. Journal of Photochemistry and Photobiology A: Chemistry, 2000. 133: p. 115-122. 214. Park, C.Y., J.K. Lee, and R.S. Chuck, Second Harmonic Generation Imaging Analysis of Collagen Arrangement in Human Cornea. Invest Ophthalmol Vis Sci, 2015. 56(9): p. 5622-9. 215. Teng, S.W., et al., Multiphoton autofluorescence and second-harmonic generation imaging of the ex vivo porcine eye. Invest Ophthalmol Vis Sci, 2006. 47(3): p. 1216-24. 216. Li D1, X.G., Wang W., Reactive oxygen species: the 2-edged sword of osteoarthritis. Am J Med Sci., 2012. 344(6): p. 486-90. 217. Kurz, B., B. Jost, and M. Schünke, Dietary vitamins and selenium diminish the development of mechanically induced osteoarthritis and increase the expression of antioxidative enzymes in the knee joint of STR/1N mice. Osteoarthritis and cartilage, 2002. 10(2): p. 119-126. 218. Clark, A.G., et al., The effects of ascorbic acid on cartilage metabolism in guinea pig articular cartilage explants. Matrix Biology, 2002. 21(2): p. 175-184.

112

219. Outerbridge, R.E., The etiology of chondromalacia patellae. J Bone Joint Surg Br, 1961. 43-B: p. 752-7. 220. Nover, A.B., et al., Long-term storage and preservation of tissue engineered articular cartilage. J Orthop Res, 2016. 34(1): p. 141-8. 221. Mauck, R.L., et al., Synergistic action of growth factors and dynamic loading for articular cartilage tissue engineering. Tissue Eng, 2003. 9(4): p. 597-611. 222. Kelly T-A, et al., Analysis of radial variations in material properties and matrix composition of chondrocyte-seeded agarose hydrogel constructs. Osteoarthritis Cartilage, 2009. 17(1): p. 73-82. 223. Zipfel, W.R., R.M. Williams, and W.W. Webb, Nonlinear magic: multiphoton microscopy in the biosciences. Nature biotechnology, 2003. 21(11): p. 1369-1377. 224. Jayme, D., T. Watanabe, and T. Shimada, Basal medium development for serum-free culture: a historical perspective. Cytotechnology, 1997. 23(1-3): p. 95-101.

113

Appendix I: Supplemental Figures

Figure S1: Chapter 2 - Strip before and immediately after wear test from Study 2 using ankle-condyle as counterface.

Figure S2: Chapter 2 - Friction time courses for two samples from Study 1 that both experienced delamination wear.

114

Figure S3: Chapter 2 - Wear test of glass lens against bovine cartilage strips: (a) Photograph showing delaminated blister (arrow). (b,c,d) Cross-sections (PLM) show delamination progression in travel path over time (3 different strips).

Table S4: Concentrations of essential (e) and non-essential (n) amino acids in Dulbecco’s Modified Eagle’s Medium (DMEM), Ham’s F12 Medium, Minimum Essential Medium (MEM) formulations, and in blood plasma according to two prior investigations [150, 151]. Several non-

Media levels (mg/L) Plasma levels (mg/L)

Amino acid Molecula

r weight [Da]

Dulbecco’s Modified Eagle’s Medium (DMEM)[

224]

Ham’s F12

Medium [224]

Minimum Essential Medium (MEM) [224]

(Stegink et al, 1999) [150]

(Frame, 1958) [151]

Mean

Amino acid level ratio

(Plasma:DMEM)

Glycine (n) 57 30 8 --- 20 18 19 0.63 L-arginine (n) 156 84 211 126 15 25 20 0.24 L-cystine (n) 103 63 35 31 --- 5 5 0.08 L-glutamine (n) 128 584 146 --- --- 77 77 0.13 L-histidine (e) 137 42 21 42 10 17 14 0.33 L-isoleucine(e) 113 105 4 52 7 19 13 0.12 L-leucine (e) 113 105 13 52 12 32 22 0.21 L-lysine (e) 128 146 37 73 23 49 36 0.25 L-methionine (e) 131 30 5 15 3 9 6 0.20 L-phenylalanine (e)

147 66 5 32 10 12 11 0.17

L-serine (n) 87 42 11 --- 10 15 13 0.31 L-threonine (e) 101 95 12 48 14 25 20 0.21 L-tryptophan (e) 186 16 2 10 --- 15 15 0.94 L-tyrosine (n) 163 104 8 52 10 18 14 0.13 L-valine (e) 99 94 12 46 23 35 29 0.31 Mean level: 107 35 48 20.9 0.28

115

essential amino acids (alanine, asparagine, aspartic acid, glutamic acid, and proline) have been omitted from the table due to their absence from standard DMEM formulations.

Figure S5: Representative viability images (live cells: green, dead cells: red) of cartilage explants after 32 days of culture in medium with physiologic glucose levels (1 mg/mL) at either a 30:1 or 100:1 media to tissue volume ratio. Yellow arrows represent explant surfaces that were exposed to culture medium. The explant bottom surface was in contact with the polystyrene culture dish throughout culture. Explants cultured at the lower media volume ratio exhibit a central region of significant viability loss. This central region exhibits the highest characteristic length from the media exposed surface and therefore is expected to possess the lowest glucose levels, as illustrated in prior work [174].

116

Figure S6: Mechanical properties and biochemical contents of cartilage explants cultured for 32 days in medium supplemented with cortisol (Cort) or dexamethasone (Dex) corticosteroid at 50 ng/mL level: (A) swelling ratio, (B) compressive Young’s modulus, , (C) GAG content per initial wet weight, and (D) GAG content per final wet weight. Insulin (1 ng/mL), glucose (1 mg/mL), amino acid (1.0X), and ascorbic acid (20 µg/mL) levels were maintained constant for both groups. The different corticosteroids had no statistical effect on measured mechanical or biochemical properties (p>0.05). Dashed lines represent day 0 values.

117

Figure S7: Mechanical properties and biochemical contents of cartilage explants cultured for 32 days in medium supplemented with or without 10% fetal bovine serum (FBS): (A) swelling ratio, (B) DNA content per initial wet weight, (C) collagen content per initial wet weight, (D) GAG content per initial wet weight. Insulin (1ng/mL), glucose (1mg/mL), amino acid (1.0X), cortisol (25 ng/mL), and ascorbic acid (20 µg/mL) levels were maintained constant for both groups. FBS supplementation induced increased explant swelling and cell proliferation but had no effect on COLi or GAGi. Dashed lines represent day 0 values. *p<0.05 represents statistical difference between groups.

118

Appendix II: List of Publications during graduate studies

Full-length Papers: 1. Durney KM, Sharifi Kia D, Wang T, Singh A, Karbowski L, Koo HJ, Ateshian GA, Albro

MB. (2018) Physiologic Medium Maintains the Homeostasis of Immature Bovine Articular Cartilage Explants in Long-Term Culture. J. Biomech Eng. In review.

2. Durney KM, Zimmerman BK, Nims RJ, Oungoulian SR, Jones BK, Boorman-Padgett JF,

Suh JT, Shah R, Hung CT, Ateshian GA. J. Biomechanics. (2018) Immature Bovine Cartilage Wear by Fatigue Failure and Delamination. In preparation.

3. Cigan AD, Durney KM, Nims RJ, Vunjak-Novakovic G, Hung CT, Ateshian, GA. (2016)

Nutrient Channels Aid the Growth of Articular Surface-Sized Engineering Cartilage Constructs. Tissue Eng Part A. 22(17-18): 1063-74.

4. Nims RJ, Durney KM, Cigan AD, Dusséaux A, Hung CT, Ateshian GA. (2016)

Continuum Theory of fibrous tissue damage mechanics using bond kinetics: application to cartilage tissue engineering. Interface Focus. 6(1): 20150063

5. Wang C, Durney KM, Fomovsky G, Ateshian GA, Vukelic S. (2016) Quantitative Raman

Characterizations of Cross-linked Collagen Thin Films as a Model System for Diagnosing Early Osteoarthritis. SPIE Proceedings Vol. 9704

6. Albro MB, Nims RJ, Durney KM, Cigan AD, Shim JJ, Vunjak-Novakovic G, Hung CT,

Ateshian, GA. (2016) Heterogeneous engineered cartilage growth results from gradients of media-supplemented active TGF-ß and is ameliorated by the alternative supplementation of latent TGF-ß. Biomaterials. 77:173-85.

7. Oungoulian SR, Durney KM, Jones BK, Ahmad CS, Hung CT, Ateshian GA. (2015)

Wear and damage of articular cartilage with friction against orthopedic implants. Journal of Biomechanics. 48(10): 1957-64.

8. Jones BK, Durney KM, Hung CT, Ateshian GA. (2015) The friction coefficient of

shoulder joints remains remarkably low over 24 hours of loading. Journal of Biomechanics. 48(14): 3958

9. Lima EG, Durney KM, Sirsi SR, Nover AB, Ateshian GA, Borden MA, Hung CT. (2012)

Microbubbles as biocompatible porogens for hydrogel scaffolds. Acta Biomaterialia. 8(12): 4334-41.

Conference Podium Presentations: 1. Durney KM, Nims RJ, Boorman-Padgett J, Suh JT, Koo HK, Jones BK, Oungoulian SR,

Hung CT, Ateshian GA. “Principal Component Analysis of Friction Force Hysteresis Cuvres for Detecting Fatigue Failure and Generating Frictional S-N Curves for Articular Cartilage.” Summer Biomechanics, Bioengineering and Biotransport Conference, July 2016, National Harbor, Maryland.

119

2. Durney KM, Oungoulian SR, Jones BK, Suh JT, Hung CT, Ateshian GA. “Cartilage

Wear Initiated by Fatigue Damage Under Physiologic Loading when Fluid Load Support and Boundary Lubrication are Compromised.” Summer Biomechanics, Bioengineering and Biotransport Conference, July 2015, Snowbird, Utah.

3. Wang C, Durney KM, Fomovsky G, Ateshian GA, Vukelic S. “Quantitative Analysis of

Raman Spectra for Assessment of Crosslink Concentration for Diagnosing Osteoarthritis at Early Stages.” Summer Biomechanics, Bioengineering and Biotransport Conference, July 2015, Snowbird, Utah.

4. Albro MB, Durney KM, Shim JJ, Singh A, Ateshian GA, Stevens MM. “Endogenous

stores of latent TGF-ß serve to maintain the integrity and viability of articular cartilage over long term culture in response to physiologic and excessive dynamic mechanical loading.” Annual Meeting of the Orthopaedic Research Society, March 2015, Las Vegas NV.

5. Durney KM, Sirsi SR, Nover AB, Ateshian GA, Borden MA, Lima EG, Hung CT.

“Microbubbles Improve Depth-Dependent Mechanical Properties of Cartilage Tissue Engineered Constructs,” Annual Meeting of the Orthopaedic Research Society, January 2011, Long Beach CA.

6. Durney KM, Sirsi SR, Ateshian GA, Borden MA, Lima EG, Hung CT “Using

Microbubbles to Modulate Hydrogel Scaffold Properties for Cartilage Tissue Engineering,” Annual Meeting of the Orthopaedic Research Society, March 2010, New Orleans LA.

Conference Posters: 1. Tan AR, Yu WT, Durney KM, Cai CC, Hu Y, Lee JH, Barbosa S, Ateshian GA, Guo XE,

Marra KG, Bozynski CC, Stoker AM, Cook JL, Hung CT. “Sustained Low-Dose Intra-articular dexamethasone delivery enhances clinical repair of focal cartilage lesions.” Annual Meeting of the Orthopaedic Research Society, New Orleans LA, 2018.

2. Durney KM, Nims RJ, Law WA, Cook JL, Hung CT, Ateshian GA. “Mature Canine Cells in full-scale engineered cartilage recapitulate native mechanical and biochemical properties with nutrient channels.” Annual Meeting of the Orthopaedic Research Society, San Diego CA, 2017.

3. Nims RJ, Durney KM, Law WA, Tan AR, Roach BL, Cook JL, Hung CT, Ateshian GA.

“Heterogeneity of engineered cartilage constructs using adult canine chondrocytes is reduced with higher TGF-ß concentrations.” Annual Meeting of the Orthopaedic Research Society, San Diego CA, 2017.

4. Nims RJ, Cigan AD, Durney KM, Jones BK, Law WA, Hung CT, Ateshian GA.

“Constrained tissue culture reduces collagen solubility and promotes collagen maturation to improve engineered cartilage functional properties.” Annual Meeting of the Orthopaedic Research Society, San Diego CA, 2017.

120

5. Wang C, Fomovsky M, Durney KM, Miao G, Yu J, Ateshian GA, Vukelic S. “Laser Irradiation as Novel Paradigm for Treatment of Early Osteoarthritis.” Annual Meeting of the Orthopaedic Research Society, San Diego CA, 2017.

6. Law WA, Durney KM, Nims RJ, Vunjak-Novakovic G, Hung CT, Ateshian GA.

“Chondrocytes produce native-like ECM organization using the MatriTek Tissue Spec™ Cartilage Hydrogel Kit as a scaffold for cartilage tissue engineering when nutrient channels and CAGE strategies are employed.” Annual Meeting of the Orthopaedic Research Society, San Diego CA, 2017.

7. Durney KM, Nims RJ, Albro MB, Gu T, Karbowski L, Singh A, Vukelic S, Hung CT,

Ateshian GA. “Raman spectrographic characterization of cartilage matrix swelling via lysyl oxidase inhibition in immature explants and tissue constructs.” Annual Meeting of the Orthopaedic Research Society, Las Vegas NV, 2015.

8. Albro MB, Durney KM, Shim JJ, Singh A, Ateshian GA, Stevens MM. “Endogenous

stores of latent TGF-ß serve to maintain the integrity and viability of articular cartilage over long term culture in response to physiologic and excessive dynamic mechanical loading.” Annual Meeting of the Orthopaedic Research Society, Las Vegas NV, 2015.

9. Albro MB, Nims RJ, Durney KM, Cigan AD, Shim JJ, Vunjak-Novakovic G, Hung CT,

Ateshian GA. “Heterogeneous growth of engineered cartilage results from gradients of media-supplemented active TGF-ß and is ameliorated through the alternative supplementation of latent TGF-ß.” Annual Meeting of the Orthopaedic Res. Society, Las Vegas NV, 2015.

10. Albro MB, Durney KM, Shim JJ, Cigan AD, Nims RJ, Hung CT, Ateshian GA.

“Endogenous stores of latent TGF-ß function to maintain the mechanical integrity of articular cartilage independent of physiologic mechanical loading: Assessment through the novel validation of the specificity of a small molecule TGF-ß inhibitor.” Annual Meeting of the Orthopaedic Research Society, New Orleans LA, 2014.

11. Albro MB, Durney KM, Shim JJ, Singh A, Nims RJ, Cigan AD, Hung CT, Ateshian GA.

“Synovial fluid and physiologic levels of cortisol, insulin, and glucose in media maintain the homeostasis of immature bovine cartilage explants over long-term culture.” Annual Meeting of the Orthopaedic Research Society, New Orleans LA, 2014.