Regulating P21 Expression to Increase Chondrogenic Potential in Human Mesenchymal Progenitor Cells

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2016 Regulating p21 Expression to Increase Chondrogenic Potential in Human Mesenchymal Progenitor Cells

Bertram, Karri

Bertram, K. (2016). Regulating p21 Expression to Increase Chondrogenic Potential in Human Mesenchymal Progenitor Cells (Unpublished master's thesis). University of Calgary, Calgary, AB. doi:10.11575/PRISM/27580 http://hdl.handle.net/11023/3397 master thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Regulating p21 Expression to Increase Chondrogenic Potential in Human Mesenchymal

Progenitor Cells

Karri Bertram

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF MASTER OF SCIENCE

GRADUATE PROGRAM IN BIOMEDICAL ENGINEERING

CALGARY, ALBERTA

SEPTEMBER, 2016

Abstract

Cartilage does not regenerate in humans, and therefore cartilage degeneration is a problem that affects a significant percentage of the population, including those with diseases such as

Osteoarthritis (OA). The p21 knockout (p21-/-) mouse contains the only known single mutation in mammals that can induce a cartilage regenerative phenotype. Work in this thesis aims to identify p21 expression inhibitors for use in humans and to characterize their effects on human synovial mesenchymal progenitor cells (MPCs) during culture and chondrogenesis. I have identified one putative p21 expression inhibitor (acting through HSP90), that induces human synovial MPCs to display phenotypic properties similar to fibroblasts from p21-/- mice.

Additionally, this inhibitor promotes cartilage formation in a mouse cartilage injury model.

These results indicate that p21 inhibition through HSP90 may be a potential pharmaceutical target for stimulating chondrogenic regeneration for the treatment of cartilage defects or in cartilage degenerating diseases such as OA.

Acknowledgements

First and foremost, I would like to thank my supervisor, Dr. Roman Krawetz for his tremendous guidance through four summer studentships and especially through my graduate work. The individual time that you spend on each one of your students is remarkable and you have made this journey a truly fun and rewarding process. I would also like to thank the members of my supervising committee, Dr. Tina Rinker and Dr. Jeff Biernaskie, thank you for your expertise and guidance. I also want to thank all of the members of the Krawetz lab, my fellow graduate students (Asmaa, Christina, Nedaa, and Guomin), our post docs (Priya and Saleem), our lab techs (Catherine and Pankaj) and summer students (Nadia and Ted). Thank you all for your technical assistance throughout this project as well as moral support. I would like to thank my

BMEG peers for giving me a sense of community as we all struggle through the same milestones. I would also like to thank Dr. Jim Powell and the Southern Alberta Organ and Tissue

Donation Program for tissue that made this project possible. I would like to acknowledge the funding that made this work possible from The Stem Cell Network and the Biomedical

Engineering Graduate Program. Finally, I would like to thank all of my friends and family that have encouraged me throughout this journey.

iii

Table of Contents

Abstract ...... ii Acknowledgements ...... iii List of Tables ...... vi List of Figures ...... vii List of Symbols, Abbreviations, and Nomenclature ...... x Chapter One: Background ...... 1

1.1 OSTEOARTHRITIS ...... 1

1.2 ROLES OF CELLS AND SIGNALLING PATHWAYS IN OA ...... 3

1.3 MESENCHYMAL STEM CELLS ...... 5

1.4 CURRENT TREATMENTS ...... 6

1.5 CURRENT AREAS OF RESEARCH IN MSC CLINICAL TRIALS ...... 9

1.6 CARTILAGE REGENERATION ...... 12

1.7 CELL CYCLE ...... 18

1.8 HYPOTHESIS AND SPECIFIC AIMS ...... 23 1.8.1 Hypothesis ...... 23 1.8.2 Specific Aims ...... 23 1.8.3 Experimental design ...... 24 Chapter Two: Methods ...... 26

2.1 ETHICS STATEMENTS...... 26 2.1.1 Human Ethics Statement ...... 26 2.1.2 Animal Ethics Statement ...... 26

2.2 CELL STRAINS ...... 26

2.3 DRUG SCREENING ...... 28

2.4 DRUG LIBRARY ...... 28

2.5 DRUG INDUCED METABOLIC TOXICITY ...... 29

2.6 CELL CYCLE ANALYSIS ...... 29 2.6.1 EDU Analysis of Proliferation ...... 29 2.6.2 PI Analysis of Cell Cycle ...... 30

2.7 DRUG KINETICS ...... 30 2.7.1 p21 protein nuclear localization ...... 30 2.7.2 p21 mRNA (qPCR) ...... 32

2.8 CHONDROGENESIS ...... 32 2.8.1 Culture ...... 32 2.8.2 qPCR (mRNA) ...... 33 2.8.3 Histology (proteoglycan)/ Pellet Size ...... 34

2.9 IN VIVO MOUSE WOUND HEALING MODEL ...... 34 2.9.1 Mice ...... 34 2.9.2 Histology ...... 35

2.10 STATISTICS ...... 36 Chapter Three: Results ...... 37

3.1 IDENTIFICATION OF P21 EXPRESSION INHIBITORS ...... 37 3.1.1 Drug screening ...... 37 3.1.2 Metabolic toxicity of drugs on XMAN cells ...... 40

3.2 CHARACTERIZATION OF CHOSEN COMPOUNDS ON MPCS DURING CULTURE ...... 42 3.2.1 Metabolic toxicity on MPCs ...... 42 3.2.2 Cell Cycle ...... 44 3.2.3 mRNA p21 ...... 47 3.2.4 p21 nuclear localization ...... 49

3.3 CHARACTERIZATION OF CHOSEN COMPOUNDS ON THE CHONDROGENESIS OF HUMAN MPCS ...... 51 3.3.1 Pellet size ...... 51 3.3.2 Alcian blue staining ...... 52 3.3.3 Chondrogenic differentiation and analysis ...... 53

3.4 CHARACTERIZATION OF DRUG 70 ON IN VIVO CHONDROGENESIS ...... 55 3.4.1 In vivo Cartilage Regeneration (0.1µM treatment group) ...... 55 3.4.2 100µM Ear hole closure ...... 58 Chapter Four: Discussion ...... 62 Chapter Five: Conclusion ...... 82 Works Cited ...... 83 Appendix A ...... 91

List of Tables

Table 1: Patient demographics for cell strains used throughout this project...... 27

Table 2: Selected compounds for p21 transcription inhibition, name, known mechanism of action, and IC50 according to literature...... 40

Table 3: Summarized alamarBlue results. Grey shaded box is the only significant result (p<0.05) compared to the DMSO carrier control...... 44

Table 4: Percentage of MPCs in each stage of the cell cycle. Shading indicates significant difference relative to the DMSO control at 24 hours, p<0.05 ...... 47

Table 5: Small molecule library list of drugs tested and their known activity. Yellow highlighted compounds are the five chosen for this study...... 91

List of Figures

Figure 1: Anatomy of the knee joint, normal versus osteoarthritic characteristics. Adapted from Uth and Trifonov, World J Stem Cells, 201410 ...... 2

Figure 2: Inflammatory factors involved in the onset of OA in the cartilage, subchondral bone and synovium. Adapted from Loeser et al., Arthritis Rheum. 2012 20...... 5

Figure 3 Pyramidal treatment scheme for the treatment of OA. Adapted from Hochburg et al., Medicographia, 201332 ...... 8

Figure 4: p21 knockout mouse ear hole closure WT (B,C) vs p21-/- (D,E) Permission for republication has been granted by the publisher 61...... 16

Figure 5: The role of p21 in cell cycle regulation. p21 suppresses the cell cycle at multiple checkpoints including G1, S and G2 by interacting with cyclins and CDKs. Adapted from Karimian et al2...... 20

Figure 6: Direct and indirect protein-protein interactions in which p21 is involved. Adapted from Dotto et al, 200078...... 21

Figure 7: Representative example of p21 fluorescent staining of a nuclear negative and nuclear positive result...... 31

Figure 8: Identification of potential p21 expression inhibitors on XMAN® cell line. Drug screening of library of 146 small molecule inhibitors targeting the complete spectrum of kinases implicated in cancer cell growth and survival. XMAN® cells were plated in each drug condition in duplicate for each drug condition, the luminescence of the plate- specific media only control was subtracted from each condition and plotted after 24 hours of drug treatment...... 38

Figure 9: Comparison of the five selected p21 inhibiting compounds. Luminescence of XMAN® cells after 24 hour treatment with each drug at four different concentrations. Each compound is normalized to the respective plate specific control. *p<0.05 ...... 39

Figure 10: Drug induced metabolic toxicity in XMAN cells. AlamarBlue absorbance and p21 luminescence readings after 96 hours treatment of each drug using four concentrations on XMAN™ cell line. AlamarBlue was tested in triplicate and normalized to the plate specific media only control. Raw luminescent data is displayed here, samples were run in triplicate on the same plate. *p<0.05...... 41

Figure 11: AlamarBlue assay for metabolic toxicity on human synovial MPC strains (n=4, pt 1, pt 2, pt 3, pt 4). Samples were plated in triplicate for each condition and normalized to the plate specific media only control and significance was analyzed compared to the DMSO only control. Statistics were performed using a standard one-way ANOVA with *p<0.05 ...... 43

vii

Figure 12: Proliferation of human MPC strains (n=3, pt 8, pt9, pt 10) following 24 hour drug treatment using Click IT® EDU. Samples were assessed using flow cytometry with one replicate each. One-way ANOVA with multiple comparisons indicates that drugs 70 and 93 increase MPC proliferation compared to the DMSO control. Statistics were done using a one-way ANOVA and p<0.05. *p<0.05, ***p<0.0001 ...... 45

Figure 13: Cell cycle analysis of MPC cell strains (n=4, pt 8, pt 9, pt 11, pt 12) using propidium iodide assay while exposed to each drug for 4 and 24 hour treatment. Cells were assessed by flow cytometry in one replicate each and statistics were assessed compared to the DMSO treated condition at 24 hours. *p< 0.05...... 46

Figure 14: p21 mRNA in human synovial MPC strains (n=3, pt 4, pt 11, pt 12) at four time points over 24 hours relative to a no treatment control. Samples analyzed in triplicate and normalized to the endogenous/housekeeping control 18S and then media only control. Statistics were assessed compared to the DMSO only control. No drugs caused a significant decrease in p21. Drug 93 increased p21 at all time points and drug 111 increased p21 at 0.5 and 4 hour time points. *P value was evaluated at < 0.05 for these tests...... 48

Figure 15: Percent of cells with p21 nuclear localization relative to the DMSO control in synovial patient MPC strains (n=4, pt 8, pt 9, pt 11, pt 12) over 24 hours of drug treatment. Cells were plated in triplicate, 9 pictures for each well were taken and the results of each picture were summed, the average of the three wells for each treatment condition were used to assess significance relative to the DMSO control. None of the time points were significantly different than the DMSO control, *p < 0.05...... 50

Figure 16: Pellet size of MPCs after 21 day pellet culture in the respective drug treated media significantly increased in area when treated with chondrogenic growth factors (chondro), drug 70, 93, 102, and 111 when compared to the undifferentiated control. Two pellets for each condition are assessed here. *p<0.05...... 51

Figure 17: Pellets under all drug treatments stain positive for alcian blue after 21-day pellet culture compared to the undifferentiated negative control...... 52

Figure 18: MPCs from patient strains (n=6, pt1, pt 5, pt 6, pt 7, pt 11, pt 12) were analyzed for mRNA content using RT-qPCR after 21 days of pellet culture in each drug condition. p21 and chondrogenic markers SOX9 and ACAN were examined. Statistics were evaluated compared to the DMSO carrier control *p<0.05 ...... 54

Figure 19: In vivo results of drug treatment on through-and through ear hole closure. Drug treated C57/BL6 mice (n=4) showed no significant difference to the control DMSO treated mice group (n=3). p21-/- mice show a significant reduction in ear hole size at week 1, 3 and 4. *p<0.05...... 56

Figure 20: Safranin-O and Fast green staining of treated (0.1 µM Drug 70 in DMSO gel) and untreated (DMSO gel) injured mouse ears after four weeks of treatment. Black arrows indicate cartilage formation. Images are provided to orient the histological sections relative the position of the injury...... 57 viii

Figure 21: High dosage ear hole closure results. p21-/- treated and untreated show no significant difference between groups with the exception of week 1 with a slight increase in hole size in the treated group compared to the untreated group. C57s show no difference in wound closure between treated and untreated groups. *p<0.05...... 59

Figure 22: Safranin-O and Fast green staining of C57 untreated controls (DMSO gel only), low dose (0.1 µM Drug 70 in DMSO gel) and high dose (100 µM Drug 70 in DMSO gel) injured mouse ears after four weeks of treatment. Black arrows indicate cartilage formation...... 60

Figure 23: Safranin-O and Fast green staining of p21-/- untreated controls (DMSO gel only), and high dose (100 µM Drug 70 in DMSO gel) injured mouse ears after four weeks of treatment. Black arrows indicate cartilage formation...... 61

Figure 24: p21 knockout mouse ear hole closure over 4 weeks compared to C57 wildtype controls and MRL mouse ear hole closure (Unpublished data from the Krawetz lab)...... 64

Figure 25: Cartilage regeneration in p21-/- mice. A) Uninjured cartilage (histological grading score = 14), B) 4 week FTCD in C57 mouse (histological grade score = 2), C) 4 week FTCD in p21 knockout mouse showing smooth surface and matrix regeneration (histological grade score = 10). (Unpublished data from the Krawetz lab) ...... 65

Figure 26: Relevant transcriptional binding sites in the p21 (CDKN1a) promoter region adapted from information provided by The Champion ChiP Transcription Factor Search Portal by Qiagen110...... 67

List of Symbols, Abbreviations, and Nomenclature

ACI Autologous chondrocyte implantation ACR American College of Rheumatology ANOVA Analysis of variables BMP Bone morphogenic protein C57 C57 BL/6 CDK Cyclin dependent kinase CDKN1a Cyclin dependent kinase 1a Cip/kip CDK interacting protein/Kinase inhibitory protein CKI Cyclin dependent kinase inhibitor DAMPS Damage associated molecular patters DMEM Dulbecco’s Modified Eagle Medium DMOADs Disease modifying osteoarthritis drugs DMSO Dimethyl sulfoxide ECM Extracellular matrix EGFR Epidermal growth factor receptor ERK Extracellular signal regulated kinases ETOH Ethanol FBS Fetal bovine serum FGF Fibroblast growth factor FLNA Filamin a HER2 Human epidermal growth factor 2 HSP90 Heat shock protein 90 IA Intraarticular injection IL Interleukin JNK C-jun n-terminal kinase KGN Kartogenin MAPK Mitogen-activated protein kinase MMPs Matrix metalloproteinases MPC Mesenchymal progenitor cell MRL Murphy Roth's large MSC Mesenchymal stem cell NBF Neutral buffered formalin

NF-kB nuclear factor kappa-light-chain-enhancer of activated B cells OA Osteoarthritis OARSI Osteoarthritis research society international p21-/- p21 knockout PCNA Proliferating cell nuclear antigen RT-qPCR Real time qualitative polymer chain reaction shRNA Short hairpin Ribonucleic Acid Homologs of both the Drosophila protein Mothers against SMAD decapentaplegic (MAD) and the Caenorhabditis elegans protein SMA (for small body size) SPRP Secondary small proline rich protein TLRs Toll-like receptors TNF-a Tumor necrosis factor alpha VEGF Vascular endothelial growth factor WOMAC Western Ontario and McMaster Universities Osteoarthritis Index

Chapter One: Background

1.1 Osteoarthritis

Osteoarthritis (OA) is the most common form of musculoskeletal disease, currently affecting 1 in

8 Canadians1. With our aging population and growing life expectancy, this number is expected to rise to 1 in 4 by 2020, placing a huge burden on our health care system1. OA is a disease characterized by the progressive degeneration of articular cartilage2. There are three types of cartilage in the body, elastic, fibrous, and articular (or hyaline). Each type of cartilage is unique in composition, although all are comprised of chondrocytes, collagen, and extracellular matrix

(ECM). Articular cartilage is mainly composed of collagen type II, in which fibrils are oriented and interact with proteoglycans in order to absorb shock. Articular cartilage covers the articular surfaces of synovial joints and provides protection for the underlying bone and a smooth sliding surface that promotes pain-free movement of joints. Unlike bone, cartilage is aneural and avascular. Therefore, cartilage does not sense pain upon impact (or injury), however being avascular results in a very limited intrinsic healing ability3.

With the onset of OA, cartilage begins to degrade, which results in the loss of this protective cartilage layer overlaying the bone4. It still remains unclear whether cartilage degradation is initially caused by injury, inflammation, mechanical wear, or other factors5. Additionally, new insights into OA suggests that one or more risk factor(s) may be cumulative in a given patient5.

When cartilage is damaged/lost, the underlying bone, which is rich in nerve endings, is exposed to the joint environment causing chronic pain6. OA is characterized by visibly worn articular cartilage, a narrowed joint space, osteophyte formation, abnormal cartilage morphology, bony sclerosis and cysts as depicted in Figure 1. While OA has been actively researched for decades in humans and animal models, little is known about why in some people the disease progresses

rapidly after joint injury or another type of joint insult, while other individuals almost appear to be resistant to the onset and progression of the disease7. Indeed, patients that suffer an inter-articular joint injury who are overweight or are female have a greater risk of developing OA with aging8,9.

As mentioned, progression of OA leads to debilitating joint pain, which has a negative impact on quality of life and affects the ability of these individuals to work, impacting the economy in turn.

Figure 1: Anatomy of the knee joint, normal versus osteoarthritic characteristics. Adapted from

Uth and Trifonov, World J Stem Cells, 201410

1.2 Roles of Cells and Signalling Pathways in OA

It is becoming increasingly clear that OA is not solely a mechanical disease. What began as an observation of synovitis (inflammation of the synovium), in many patients with OA has led to the understanding that there is interplay between the mechanical and biochemical processes that act to perpetuate joint failure11. Synovitis has now been shown to be present before radiographic changes in the cartilage can be observed and has become one indicator of early OA12,13. With the inflammation of synovitis, the innate immune system is set into motion. Mono-nucleated inflammatory cells migrate into the synovium, which include mostly macrophages but can also include t-cells, b-cells, and natural killer cells11. When cartilage begins to degrade the fragments cause an increase in damage associated molecular patterns (DAMPs) in the synovial fluid, such as fibronectin and hyaluronan, which propagate the inflammatory response11. Toll-like receptors

(TLRs) also become elevated with the onset of synovitis14. These inflammatory responses promote the breakdown of extracellular matrix (ECM), increasing cartilage damage by releasing cytokines including IL-6, IL-1b, TNF-a, and vascular endothelial growth factor (VEGF)11. It was identified in 1993 that IL-6, IL-1 and TNF-a cytokines are involved in cartilage ECM breakdown15. These cytokines are upregulated in OA synovial fluid and influence chondrocytes to secrete matrix metalloproteinases (MMPs), a family of enzymes, including collagenases and aggrecanases, responsible for breaking down ECM5,11,15. Essentially, cartilage injury/degeneration leads to inflammation which in turn, promotes more cartilage breakdown. These inflammatory pathways involved in OA are actively being studied in order to determine appropriate early biomarkers of

OA and also for potential therapeutic targets.

Chondrocytes are the only cell type found in articular cartilage16. Under normal conditions chondrocytes are in a quiescent state and there is very little tissue turn over17. With the onset of

OA, chondrocytes begin to produce of both matrix modeling proteins and matrix degrading enzymes17. Inappropriate hypertrophy-like, cartilage maturation occurs as well as cartilage calcification, resulting in osteophytes. This chondrocyte activation is primarily through mechanical and inflammatory signals through the NF-kB and MAPK pathways18,19. Upon activation, the ERK, JNK, and p38 MAPK cascades coordinate the induction and activation of transcription factors that regulate the expression of genes involved in catabolic and inflammatory events20.

Figure 2 depicts some of the known factors involved in the inflammatory onset of OA, demonstrating the complexity of the system. OA is considered a whole joint disease as the many tissues affected by the disease process are also contributing factors (e.g. sites of pro-inflammatory cytokine release) including the cartilage, synovium, and subchondral bone.

Figure 2: Inflammatory factors involved in the onset of OA in the cartilage, subchondral bone and synovium. Adapted from Loeser et al., Arthritis Rheum. 2012 20.

1.3 Mesenchymal Stem Cells

Mesenchymal stem cells are a category of adult stem cells that have the ability to differentiate into cell types of the mesenchyme lineage21. These include osteoblasts, adipocytes, chondrocytes, myocytes, tenocytes, and stromal cells22. Mesenchymal stem cells can also be found in a partially differentiated state where they have become lineage restricted (e.g. can turn into a few or one cell type(s)). In this case they are commonly referred to as mesenchymal progenitor cells. Various types of mesenchymal stem cells and progenitor cells are currently being studied for regenerative medicine as they are readily available in adult tissues and they are already specified into a specific lineage (e.g. mesoderm), and therefore may have less risk for tumor formation upon transplantation 5

than pluripotent stem cells23. Furthermore, mesenchymal stem cells have been shown to have regenerative potential in vivo and can contribute to repairing many tissues including skin, gastric epithelium, and the heart24–26.

Mesenchymal progenitor cells are present in both normal and arthritic synovial membrane and synovial fluid27. These cells have the ability to differentiate into osteoblasts, adipocytes, and chondrocytes, in vitro and are thought to contribute to the normal healing of cartilage in healthy joints28. Kurth et al showed that in mice there are slow cycling, non-hematopoietic, non-endothelial stromal cells, present in the synovium29. After a cartilage defect, these slow-cycling cells were found in the cartilage-defect and stained for chondrocyte-lineage markers SOX9 and type II collagen, indicating that these cells can differentiate and contribute to cartilage repair in vivo in mice29. It has been shown by a few groups, including our own, that MPCs from OA joints have an increased proliferative capacity but reduced chondrogenic potential compared to those derived from non-diseased joints, suggesting that MPCs in OA joints may be ineffective in repairing any damage that occurs to the cartilage 27,28,30. This loss of differentiation ability could be due to a number of factors including changes in the biological, chemical, or joint mechanical environment.

Therefore, these tissue resident MPCs have become a potential target for joint/cartilage disease treatment due to their proliferation capacity and ability to differentiate into articular cartilage both in vitro and in vivo.

1.4 Current Treatments

Unlike Rheumatoid arthritis, there are currently no disease-modifying drugs or early diagnosis tools available for OA1. OA is currently diagnosed symptomatically; pain, swelling, and stiffness in joints, and by radiographic changes; joint space narrowing, osteophytes, and bony sclerosis2.

By the time chronic pain is felt or an X-ray can detect cartilage changes irreversible damage has already been done and the condition has progressed well beyond its initial stages (e.g. cartilage damage without bone involvement). Therefore, physicians have to rely on symptom management until invasive surgical strategies such as total joint replacement is an option (e.g. typically only performed in older individuals).

Physicians typically adhere to a pyramidal scheme for treatment of OA depending on the progression and severity of the disease described in Figure 3. The Osteoarthritis Research Society

International (OARSI) has recently published a list of non-surgical strategies for treatment of knee

OA31. These include: land based exercises, water based exercises, weight management, strength training, self-management and education, biomechanical interventions, intra-articular corticosteroids, topical nonsteroidal anti-inflammatory drugs, walking cane, capsaicin, Duloxetine

(a serotonin-norepinephrine reuptake inhibitor), and acetaminophen. While there are options available, efficacy of these treatments are still debated and all have side effects, especially for long term use; which is necessary for chronic diseases such as OA.

Figure 3 Pyramidal treatment scheme for the treatment of OA. Adapted from Hochburg et al.,

Medicographia, 201332

When the disease progresses beyond the ability to be managed with non-surgical options, clinicians turn to surgical interventions. Marrow stimulation, such as micro-fracture, in which the subchondral bone below the damaged cartilage is punctured, can be done in order to induce blood supply into the joint and, in turn, allow bone marrow mesenchymal stem cells (bmMSCs) to enter the cartilage lesion33. This technique is often used in athletes and/or active individuals who are looking to achieve a high level of activity as soon as possible. The bmMSCs form fibrocartilage

in the lesion, which provides stability in the short term34. Fibrocartilage is denser and has a higher tensile strength than articular cartilage and therefore cannot integrate properly with the surrounding tissues35,36. This leads to uneven wear, and the fibrocartilage in the affected region will preferentially break down and again result in pain during movement. A recent review of OA surgical techniques revealed that beneficial clinical outcomes after micro-fracture deteriorated after 18-24 months37. Autologous chondrocyte implantation (ACI) is another clinical method used currently38. Cartilage is taken from a non-load bearing area of the joint and implanted into the lesion, usually contained with a flap of periosteum or porcine collagen membrane, with or without the support of other natural/synthetic biomaterials. While this method replaces the lesion with the correct type of cartilage, drawbacks include the requirement of two surgeries (chondrocyte harvest and implantation), in vitro culture (possible introduction of pathogens/environmental changes to cells), and donor site morbidity34,35. A number of recent studies have also begun to explore the use of mesenchymal stem cells (MSCs) in the clinical treatment of OA. These studies have either used

MSCs that have been differentiated into a chondrogenic phenotype in vitro for seeding a biomaterial scaffold for implantation39, or in some cases, autologous undifferentiated MSCs from the bone marrow, adipose or synovial tissues for transplantation into patients. While these clinical studies have shown some repair on the macro and micro levels, there is debate in the field about the best source of MSCs for this purpose as well as the long term patient outcomes 40,41. Finally, if all previous treatments have failed and the patient is considered a good candidate for surgery (age, severity of disease, etc), then a total joint replacement is performed.

1.5 Current areas of research in MSC clinical trials

As mentioned, there have been very few clinical trials on MSC injections for OA to date. The current research is mostly still at the pre-clinical stage. A proof of concept phase I/II clinical trial

of autologous adipose derived mesenchymal stem cell (AD-MSC) intra-articular (IA) injections for the treatment of osteoarthritis was reported in 201442. In the phase I study 9 patients with idiopathic osteoarthritis, grade 2 or higher according to the Kellgren-Lawrence criteria (OA severity) were spilt into 3 cohorts and given 3 doses of IA injections of autologous AD-MSCs; low-1x107, mid-5x107, high 1x108. Patients were followed up at 1, 2, 3, and 6 months after injection. No adverse effects were noted in any of the patients treated. The high dose cohort showed a mean reduction of 39% in the Western Ontario & McMaster Universities Osteoarthritis Index

(WOMAC) score (quality of life index) at 6 months after injection compared to the baseline of this cohort, whereas patients in the low- and mid-dose groups did not improve over 6 months. The

WOMAC score is an evaluation of pain, stiffness, and physical function in patients with hip or knee OA. A lower score indicates less pain, stiffness and physical limitations. The study then moved to a phase II clinical trial with 9 more patients, treated with the high-dose of AD-MSCs and followed up at 1, 2, 3, and 6 months. These patients consented to a biopsy of the medial femoral condyle before injection and at 6 months’ post injection. The treatment showed an increase in cartilage thickness from 0.4 ± 0.3 mm before injection to 1.6 ± 0.8 mm after injection, including 4 patients with no cartilage before injection regenerating 1.6 ± 0.5 mm of cartilage after. Although this study was based on a small sample size, it suggests that undifferentiated MSCs may have the ability to regenerate cartilage in vivo in humans and may be a realistic treatment option for OA.

In another autologous MSC clinical trial, 18 patients were given autologous MSC transplants for the treatment of OA and followed up for 24 months43. In this study adipose synovium was harvested from the inner side of the patients’ infrapatellar fat pad. An average of 1.18 x106 (range,

0.3 x 106 to 2.7 x 106) MSCs were isolated and injected with 3.0 ml of platelet-rich plasma. The

injection was performed the same day, approximately 3-4 hours after harvest. This study found a negative correlation between WOMAC score and the number of cells injected, indicating that a greater number of cells injected results in improved outcomes. Overall, WOMAC scores decreased from 49.9 points preoperatively, to 38.3 points at the 1-year follow-up and to 30.3 points at the 2- year follow-up. These results suggest that MSCs from intraarticular infrapatellar fat pad also appear to have the ability to treat OA over a longer term follow-up.

At the pre-clinical level, there are groups working on how to best use various sources of MSCs for cartilage regeneration, both at the undifferentiated or differentiated state42,44–46. Various biomaterials are being assessed in parallel in order to determine the ideal method for implantation in a mobile joint47,48. And finally groups, such as our own, are looking at pharmaceutical methods for an endogenous method of OA treatment49,50, through activation of joint resident stem/progenitor cell populations.

Overall, it is essential to understand the physiologic pathways involved in OA onset and progression more thoroughly in order to treat the disease appropriately. Due to the inflammatory nature of the disease, the cause of the cartilage degradation cascade must be identified and corrected in order to have a long term effect. If these mechanisms are well understood, then they can be targeted directly. While MSC injections have shown some positive results, the major drawbacks of cell source scarcity and invasiveness of the procedure has led our group to pursue pharmaceuticals in order to target the MPCs already present in the joint.

1.6 Cartilage Regeneration

Although long debated in the field, it has recently been confirmed that collagen in articular cartilage does not regenerate endogenously to a measurable extent after skeletal maturity in both normal and diseased knees 51. Approaches to cartilage repair and regeneration can be categorized into three areas: disease-modifying osteoarthritic drugs (DMOADs), cell based therapies, and scaffold based therapies. This section will focus on outlining major advances in DMOADs.

DMOADs can be further broken down into groups with three aims: inhibiting matrix degrading enzymes, protecting chondrocytes from catabolic activity, and the development of anabolic agents to promote chondrocyte or chondrocyte progenitor cell proliferation.

Caspase inhibitors have been shown to protect against chondrocyte apoptosis in vitro as well as when injected intraarticularly in rabbits52,53. Other cartilage enzyme inhibitors have been studied for chondro-protection, however the few that have made it to clinical trials have had issues with efficacy and adverse side effects, the most common of which is musculoskeletal syndrome54. This could potentially be avoided if the drugs were injected intraarticularly for a short period of time, i.e. the critical time following joint injury or in early OA. Pro-inflammatory cytokine inhibitors have also been assessed for the treatment of OA54. IL-1Ra is therapeutically effective in animal models of OA and has moved on to clinical trials55. The anti-inflammatory cytokine IL-10 has also been shown to stimulate collagen II and proteoglycan expression, inhibit MMPs, and protect against chondrocyte apoptosis56.

Growth factors are another area for pharmaceutical targets for cartilage regeneration. Bone morphogenetic proteins (BMPs), insulin-like growth factor-1 and fibroblast growth factors (FGFs)

have all been shown to have chondroprotective abilities54. FGF-18 and BMP-7 are currently in clinical trials54.

Only a select few have looked into small molecule inhibitors for chondrogenic regeneration. Zhang et al. recently showed that systemic injection of a prolyl hydroxylase (PDD) inhibitor increased the stability of hypoxia-inducible factor 1a (HIF-1a) and this resulted in non-regenerating mice to close ear punch holes in 35 days57. Some areas of cartilaginous condensation were observed in the treated ears; however fully regenerated cartilage was not observed.

Another group has used a high-throughput image based screen to identify small molecule compounds that induce human bone marrow derived mesenchymal stem cells to differentiate into cartilage tissue58. Of 22,000 compounds screened one, kartogenin (KGN), was found to promote chondrocyte differentiation in a dose dependent manor. KGN was able to protect cartilage degradation when induced both in vitro and in vivo in two mouse OA models. The group further looked at the binding ability of KGN and found that it binds to filamin A (FLNA), an actin-binding protein that crosslinks actin filaments, thereby regulating cyto-skeletal network organization and dynamics. Knockdown of FLNA caused a five-fold increase in chondrocyte formation, confirming the relation to chondrogenesis. The pathway was further examined to find that KGN specifically blocks the interaction of FLNA with core binding factor b subunit, in turn regulating the RUNX family of transcription factors58.

Although there have been many different approaches to induce cartilage regeneration, while often shown in hydra and amphibians, it is not normally seen in mammalian species. This is why it was

very exciting for the field of cartilage regeneration when the Murphy Roth’s Large (MRL) mouse was shown to regenerate adult tissues, including cartilage by Clark et al59. MRL mice, also known as ‘superhealer’ mice have been shown to regenerate ear pinnae hole, digit tips, peripheral nerves, alkali-burned cornea, cardiac wounds, articular cartilage, intraarticular fracture, and surgical skin wounds 60. This strain of mice is able to heal through and through ear hole punches with normal tissue architecture, including angiogenesis and chondrogenesis after 28 days compared to the wildtype which did not close the hole over this time frame.

In regards to ear vs. joint cartilage, it has previously been demonstrated that chondrocytes from human auricular (ear) cartilage produce type II collagen, similarly to chondrocytes of the articular

(joint) cartilage6. Auricular cartilage can also stimulated to proliferate and secrete ECM by similar growth factors as articular cartilage is responsive to, such as TGFbeta and fibroblast growth factor

(FGF)6. There are, however, some important differences in the environments of the two types of cartilage. For instance, auricular cartilage does not have the mechanical or biological support of the underlying subchondral bone that articular cartilage does. In articular cartilage the chondrocytes become hypertrophic and calcify, creating the transition between deep zone cartilage and ossified bone7, while auricular cartilage does not calcify. In fact, the mechanical environment surrounding the auricular cartilage consists of mainly elastin, which serves to disperse force in the ear. Both types of cartilage are in close proximity to a population of mesenchymal progenitor cells, articular cartilage is near synovial progenitor cells and auricular cartilage is surrounded by progenitor cells in the perichondrium8–10. Along with the perichondrium progenitor cells, auricular cartilage is in close proximity to blood vessels, which contain pericytes and haematopoietic stem cells, as well as stem cells from the skin and hair follicles such as from the bulge and basement

membrane11,12. Overall, the cartilage found within the ear and joint demonstrate more similarities than differences. Specifically, one group has demonstrated in mice that there is a correlation between the ability to heal ear hole wounds and articular cartilage wounds13. This group used ten strains of mice with varying healing ability. Full thickness articular cartilage defects and 2 mm ear punch wounds were induced in each strain of mice and a linear correlation was observed between articular cartilage regeneration and ear wound closure according to a five parameter articular cartilage regeneration score and the ear wound healing diameter13.

One mouse strain that demonstrates superior ear and joint cartilage healing is the MRL mouse.

The MRL mouse was the result of cross breeding AKR, C3H, C57BL/6(B6), and LG strains of mice for a large murine lupus model61. In order to determine if there was a single point-mutation responsible for this regenerative capacity, over the next 10 years extensive research was done into the cell biology and genetics of the MRL mouse. It was observed that MRL fibroblast-like cells had an unusual accumulation in the G2/M cell cycle phase61. G2/M cell cycle accumulation has been associated with regeneration in hydra and amphibians, as well as mammalian livers62.

Bedelbaeva et al also found that the MRL cells had a marked increase in DNA damage, but somehow were able to overcome this in order to complete regeneration 61. This DNA damage was consistent with reports showing that mouse embryonic stem cells display endogenous DNA damage, due to a faulty G1 checkpoint, which had been attributed to the lack of p21 induction.

Under the hypothesis that the MRL mice had a similar mechanism of healing, the group found that

MRL cells display near undetectable levels of p21 expression, and was able to show that with the deletion of only CDKN1A (gene encoding p21) in a non-regenerating strain of mice, this lead to remarkably similar regeneration capabilities as the MRL mice, as seen in Figure 461.

Figure 4: p21 knockout mouse ear hole closure WT (B,C) vs p21-/- (D,E) Permission for republication has been granted by the publisher 61.

Unfortunately, also similar to the MRL mice, these p21 knockout mice develop various types of cancer and autoimmune disorders prematurely and do not live beyond approximately 16 months, compared to the 2 year lifespan of wild type mice63,64. Cells normally go through checkpoints during division to ensure DNA damage has not occurred during duplication of cellular contents.

Without p21, the G1/S phase checkpoint is skipped and division is completed regardless of errors, creating a much higher likelihood of producing a genetically damaged and eventually cancerous cell65.

p21 has been examined in the context of regeneration, most extensively in the mammalian liver66–

69. The mammalian liver has a remarkable ability to regenerate. Only as little as 25% of the mass

is needed to regrow the entire mass of the organ70. It has been shown that over expression of STAT-

3 results in p21 overexpression, leading to impaired regeneration in fatty livers71. Conversely, repression of the p53/p21 pathway has been shown to enhance liver regeneration66,68. If p21 overexpression is impeding regeneration due to cell cycle inhibition, it could be proposed that similar cell cycle inhibitors would do the same. p21 is a member of the cip/kip family of cyclin dependent kinases65. These proteins are characterized by their ability to bind to both cyclin and cyclin dependent kinase (CDK) subunits and can modulate the activity of the cyclins D-, E-, A-, and B-CDK complexes65. Other members of the cip/kip family are p27 and p5765. Interestingly, p27 and p57 knockouts show very different phenotypes when compared to each other and when both are compared to p21 knockout mice. Neither p27 or p57 knockouts demonstrate the remarkable cartilage differentiation ability observed in p21 knockout mice72,73. p53 is upstream of p21 and is a known regulator of p2165. Curiously, p53 is upregulated in MRL mice and it was investigated whether p53 is necessary for regeneration74, however it was observed that p53-/- mice did not show the same regeneration as MRL and p21-/-. Other modulators of the cell cycle checkpoints, Gadd45a and p16, which act on the G2 and G1 phase respectively, were assessed for their effect on regeneration. Knockouts of these proteins also did not show regeneration74. These studies indicate that p21 is acting independently of p53 to inhibit regeneration and that p21 inhibition is unique in terms of related cell cycle regulators in its ability to promote chondrogenic regeneration.

It has been shown by our lab that synovial derived MPCs from OA patients demonstrate a higher expression of p21 in the undifferentiated state compared to MPCs derived from healthy joints75.

Interestingly, after differentiation, MPCs derived from normal synovium demonstrated an up

regulation of p21, while OA derived MPCs did not. Our results also demonstrated that normal cells had greater chondrogenic differentiation ability than their OA counterparts. Overall this suggests that p21 may also play a role in human synovial MPC chondrogenesis. Studies by other groups have seen similar effects. Knockdown of p21 has also been shown to enhance proliferation and expression of “stemness” markers in human bone marrow derived mesenchymal stem cells by Yew et al76. This study found that late passage MSCs showed increased expression of p21, therefore, they knocked down p21 using a lentivirus-mediated shRNA in these late-passage MSCs and were able to increase the proliferation capacity and osteogenic potential of the cells. Unfortunately, this study did not look at chondrogenic potential, however it suggests that p21 knockdown may act similarly in human MSCs as it does in mice. This study also looked at the effect that p21 knockdown had on the tumorigenic potential of these cells and found that there was no change in chromosome integrity or tumorigenic potential. This may be due to the basal level of p21 remaining in the cells, protecting them from unregulated cell division, compared to the complete p21 knockout observed in MRL and p21-/- mice.

1.7 Cell cycle p21(CIP1/WAF1) is a protein encoded by the CDKN1A gene and has many known functions throughout the cell and organism. Most notably, p21 is a cyclin dependent kinase inhibitor (CKI) regulated by p5365. There are two major checkpoints during the cell cycle, G1 and G2. During the

G1 checkpoint, DNA is checked for damage before replication, while during the G2 phase proper chromosome replication is verified. The M phase checkpoint allows for properly aligned kinetochore spindle fibres to be developed1. Cyclin dependent kinases (CDKs) are required for cell cycle regulation, specifically CDK1, CDK2, CDK4, and CDK62. While CDK activity remains stable throughout the cell cycle, cyclin levels fluctuate leading to activation of CDKs. CDK2,

CDK4, and CDK6 are present in the G1 phase, CDK2 in the S phase, and CDK1 in the G2 and M phases2. Cyclin Ds activate CDK4 and CDK6 in the G1 phase and similarly cyclin E activates

CDK2 leading to G1/S transition2. Cyclin E is degraded in the S phase and replaced with cyclin A leading to transition out of S phase2. Finally, at the end of the G2 phase CDK2 interacts with cyclin

B2. CDKs are regulated by CDK inhibitors (CKIs). p21 was the first CKI to be identified. p21 binds to cyclin A/CDK2, E/CDK2, D1/CDK4, and D2/CDK4 complexes leading to the inhibition of phosphorylation of pRB as seen in Figure 52. p21’s interaction with cyclin E/CDK2, D1/CDK4, and D2/CDK4 all inhibit the G1/S transition, while cyclin A/CDK2 interaction can inhibit transition through the S phase2. Overall, this demonstrates that p21 plays a major role in the progression of the cell cycle at multiple checkpoints.

Figure 5: The role of p21 in cell cycle regulation. p21 suppresses the cell cycle at multiple checkpoints including G1, S and G2 by interacting with cyclins and CDKs. Adapted from

Karimian et al2.

Although p21 can be involved in multiple checkpoints within the cell cycle, it was discovered that fibroblasts in mice lacking p21 accumulate in the G2/M phase3. Similarly, higher levels of p21 in late passage mesenchymal stem cells (MSCs) show an increase in G1/G0 phase compared to the early passage MSCs, which have a lower amount of p214. Knockdown of p21 in both human MSCs and induced pluripotent stem cells (iPSCs) showed an increase in proliferation, indicating that the cells deficient in p21 are not arresting in the G1/G0 phase4,5. Overall, this suggests that while p21 can regulate many of the cell cycle checkpoints, the loss of p21 in mice, does not allow for cells to freely pass through the G2/M checkpoint, while the other checkpoints seem to be at least partially impaired.

Recent research has shown that p21 also acts through the CDK-independent pathway using the

Proliferating Cell Nuclear Antigen (PCNA) attachment ability of the protein 78,79. PCNA is involved in DNA damage induced degradation. p21 inhibits DNA synthesis through this pathway by binding to PCNA, which competes with several proteins involved in DNA synthesis78.

Other than interacting with CDKs and PCNA, p21 has also been associated with other protein- protein interactions, summarized in Figure 6.

Figure 6: Direct and indirect protein-protein interactions in which p21 is involved. Adapted from

Dotto et al, 200078.

In cultured epidermal cells a down-regulation of p21 has been observed in late stages of differentiation into keratinocytes however, overexpression of p21 inhibited differentiation, suggesting that p21 is part of a negative feedback loop and needs to be inactivated at later stages of differentiation in this cell type78,80. This down-regulation of p21 during differentiation has also

been shown during osteogenesis, where the overexpression of p21 similarly inhibits differentiation78. However, our lab has shown that p21 is up-regulated during chondrogenesis in synovial MPCs derived from healthy patients, whereas it is down-regulated during chondrogenesis in MPCs derived from patients with OA75. Overall, these results suggest a role of p21 in the change of state of MSC/MPCs from undifferentiated to terminally differentiated cells, however, the role of p21 appears to be different depending on the initial cell type and what lineage the cell is differentiating into. To assess the effect of p21 suppression on chondrogenic potential, while mitigating the possible carcinogenic effects of p21 suppression, the goal of this project is to temporarily knock down p21 during chondrogenic differentiation rather than to permanently remove it.

p21 is involved in a diverse set of functions with many feedback loops, therefore understanding the regulation of the protein is difficult. Because of p21’s involvement in cell cycle inhibition, p21 up-regulation has often been studied for potential anti-cancer function81–83. Suppression of p21 has been studied less-so due to the potential tumorogenic consequences84,85. It is known that the p21 protein can be degraded naturally in the body by ubiquination, an enzymatic post-translational modification in which a ubiquitin protein is attached to a substrate protein86. This initiates degradation via the proteasome. Pharmaceutically, butyrolactone, LLW10, and sorofenib act similarly to degrade p21 protein by inducing proteasome degredation87–89. At the transcriptional level it has also been shown that c-Myc, a regulator protein, inhibits p21 by binding to the promoter region during transcription90. p53 is also a known activator of p2191. p63 competes with p53 binding to the p21 promoter and acts as a transcriptional repressor of p2192. Polo-like kinase 1

(Plk1) is a p53 repressor that has been shown to reduce p53-dependent p21 transcription93. Small

proline rich protein (SPRR) 2A is a gene that encodes for a skin crosslinking protein that gives structural integrity to keratinocytes and is a critical stress and wound repair modulator94. It has been shown that SPRR2A reduces the acetylation of p300, which interferes with p300-p53 binding to the p53-RE site on the p21 gene, reducing the p53-dependent transcription of p2194. Therefore, because of the various pathways that interact with p21 and the diverse cellular outcomes that can arise from these interactions, coupled with the unknown mechanism by which p21 regulates regeneration, it was decided to attempt to identify a small molecule that would decrease the expression of p21 at the transcriptional level. Additionally, to reduce the risk of tumorigenesis while inhibiting p21, this project employed a drug library consisting of inhibitors for the full range of kinases implicated in cancer. These compounds have undergone clinical trials for cancer treatment and therefore, we can be more assured that we are not delivering a potential tumorigenic agent.

1.8 Hypothesis and Specific Aims

1.8.1 Hypothesis

Pharmacological inhibition of p21will increase the chondrogenic potential of human mesenchymal progenitor cells and enhance cartilage regeneration in vivo.

1.8.2 Specific Aims

SA1. Identify p21 inhibiting compounds using a high throughput screening method and characterize the effect of the selected compounds on human synovial mesenchymal progenitor cells during in vitro culture.

SA2. Characterize the effect of the selected compounds on human mesenchymal progenitor cell chondrogenic differentiation in vitro, and determine if the(se) compounds can induce cartilage regeneration in vivo using a mouse wound healing model.

1.8.3 Experimental design

SA1. Genetically modified cells (XMAN™) derived from human colon cancer cells expressing luciferase at the promoter region of the CDKN1A gene (p21) (Horizon) were used in a high throughput screening of potential p21 inhibiting compounds. A library of 150 small molecule compounds targeting the complete spectrum of kinases implicated in cancer cell growth and survival was chosen for the initial screening due to the tumor suppressive nature of p21. XMAN™ cells were exposed to each compound at various concentrations and the luminescence was recorded. Due to the genetic modifications of these cells, luminescence could be translated into a reading of p21 transcription. Five compounds that showed the most p21 reduction were chosen to continue with the study.

Mesenchymal progenitor cells from human donors were exposed to the selected five candidate compounds. MPCs were evaluated for metabolic toxicity using alamarBlue. The level of p21 inhibition was characterized at the mRNA level using RT-qPCR and at the protein level using immunofluorescence. The effect of the drugs on the proliferation and cell cycle was also assessed using Click-iT ® EDU and propidium iodide respectively.

SA2. MPCs were pelleted by centrifugation to induce chondrogenesis while exposed to each of the candidate compounds and cultured for 21 days. Pellet size and mRNA expression of p21 and

chondrogenic markers were assessed throughout the culture period at 7-day increments. Endpoint proteoglycan was stained using alcian blue.

One candidate compound that showed promising results in the previous experiments was chosen for the in vivo chondrogenesis study using a mouse ear hole injury model. Non-regenerating mice were given a through-and-through 2mm ear hole punch and treated with the compound of interest topically once a week for 4 weeks. The diameter of the hole was assessed weekly and safranin

O/fast green histology was done on the harvest ear tissue to assess chondrogenic regeneration.

Chapter Two: Methods

2.1 Ethics Statements

2.1.1 Human Ethics Statement

This study protocol was approved by the University of Calgary Human Research Ethics Board

(REB15-0005 and REB15-0880). Normal Group (n=5): Criteria for control cadaveric donations were an age of 40 years or older, no history of arthritis, joint injury or surgery (including visual inspection of the cartilage surfaces during recovery), no prescription anti-inflammatory medications, no co-morbidities (such as diabetes/cancer), and availability within 4 hrs of death.

Knee Osteoarthritis (n = 7): Inclusion criteria was based on a diagnosis of OA performed by an orthopedic surgeon at the University of Calgary based on clinical symptoms with radiographic evidence of changes associated with OA in accordance with American College of Rheumatology

(ACR) criteria. Radiographic evidence of OA of any compartment of the knee with collapsed or near collapsed joint space of any compartment of the knee.

2.1.2 Animal Ethics Statement

Animal studies were carried out in accordance with the recommendations in the Canadian Council on Animal Care guidelines. Animal protocols (AC16-0043) and surgical procedures in this study were approved by the University of Calgary Health Sciences Animal Care Committee.

2.2 Cell Strains

Normal MPCs were derived and purified from healthy knee synovium tissue provided to us by the

Southern Alberta Tissue Donation Program. OA MPCs were derived and purified from tissue of consenting patients undergoing a total knee joint replacement from Dr. Jim Powell. The biopsies were minced and placed into DMEM:F12 medium mesenchymal stem cell stimulatory supplement

(human) (StemCell Technologies). At 7-14 days after initial seeding (approx. passage 2), the cells were processed for magnetic purification using the Human Lineage Cell Depletion Cocktail (BD

Canada, Ontario, Canada) that removes CD3, CD14, CD16, CD19, CD41a, CD56 and glycophorin

A positive cells. This cell population was then selected for the CD90+ (BD) progenitor fraction using magnetic separation30,95.

Table 1: Patient demographics for cell strains used throughout this project.

Patient Number Disease Status Gender Age

1 Normal Male 27

2 OA Male 45

3 OA Female 84

4 Normal Male 19

5 OA Female 73

6 OA Male 57

7 Normal Male 57

8 OA Male 76

9 OA Female 58

10 OA Male 66

11 Normal Male 76

12 Normal Male 77

The XMAN™ NanoLucTM – PEST p21 promoter reporter cell line was purchased from Horizon

Discovery. In this cell line NanoLucTM-PEST is placed directly downstream of the CDKN1A endogenous start codon in a HCT116 cell line. The activity of this endogenous promoter can then

be assessed using the luciferase signal. Cells were grown in DMEM:F12 medium containing 10% fetal bovine serum, 1% non-essential amino acids, 1% pen-strep and 0.1% beta-mercaptoethanol

(all Invitrogen, Carlsbad, CA). Cells were passaged by trypsinization at a ratio of 1:10 (i.e. 1 90% confluent flask was split into 10 new flasks) due to their high proliferation rate.

2.3 Drug Screening

XMAN™ cells were plated at 5x104 cells per well in Grenier white 96 well plates, allowed to adhere overnight, treated with 10 µM, 1 µM, 0.1 µM, and 0.01 µM of each drug in duplicate for

24 hours. The XMAN™ NanoLucTM Reporter Kit was then used to activate the luminescence.

Briefly, the protocol begins by preparing the Nano-GloTM Luciferase Assay Reagent by combining

1 volume of Nano-GloTM Luciferase Assay Substrate to 50 volumes of Nano-GloTM Luciferase

Assay Buffer (Horizon Discovery). One volume of the Nano-GloTM Luciferase Assay Reagent equal to the volume of media on the cells is then added directly to each well and mixed. After 3 minutes at room temperature luminescence was quantified using a Victor X3 plate reader and analyzed against carrier control, concentration and plate specific controls.

2.4 Drug library

The drug library used is a pharmaceutical pipeline targeted kinase inhibitors panel consisting of

146 small molecule compounds targeting the complete spectrum of kinases implicated in cancer cell growth and survival. The compounds are agents provided by pharmaceutical companies after review of the study proposal and legally approved MTA and agents purchased from medicinal chemists. All compounds and known activities are listed in Table 5. All therapeutic agents used in the screening analysis were synthesized, checked for purity, and provided by Chemietek.

2.5 Drug Induced Metabolic Toxicity

AlamarBlue (Invitrogen) was used to detect metabolic activity of the cells (both XMANTM

NanoLucTM and human MPCs) after exposure to selected drugs at the same four concentrations

(0.01µM, 0.1µM, 1µM, 10 µM). Five microliters of alamarBlue (Invitrogen) was added after 24 hours of drug treatment and incubated for 4 hours in the dark at 37°C. The absorption was quantified using a Benchmark Plus microplate spectrometer (Bio-Rad) at 570nm using 600nm as a reference wavelength and analyzed against a plate specific carrier control (DMSO), as well as a media only control.

2.6 Cell Cycle Analysis

2.6.1 EDU Analysis of Proliferation

The Click-iT® EdU Flow Cytometry Assay Kit (Invitrogen) was used to measure the percent of cells that passed through the S-phase over 24 hours of drug condition treatment. Human MPC cells from 3 patients were plated at 50,000 cells per well in 6-well plates for each condition, allowed to adhere overnight, and treated with the respective condition. Cells were labeled with 10 µM EdU during the first 24 hours of treatment. Cells were then washed once with 1% BSA in PBS and pelleted by centrifugation (1500 rpm for 6 minutes). The pellet was dislodged and re-suspended in

100µl of Click-iT® fixative and incubated for 15 minutes at room temperature in the dark. Cells were then washed with 1% BSA in PBS, pelleted again by centrifugation and dislodged and resuspended in 100µl of 1x Click-iT® saponin-based permeabilization and wash reagent and mixed well. Five hundred microliters of Click-iT® reaction cocktail was added to each tube and incubated for another 30 minutes at room temperature in the dark. The cells were then washed once with 1x

Click-iT® saponin-based permeabilization and wash reagent, pelleted and resuspended in 1x

Click-iT® saponin-based permeabilization and wash reagent for flow cytometry analysis. Flow cytometry was performed on an InvitrogenTM AttuneTM Flow Cytometer. At least 10,000 gated events (excluding debris) were counted at a wavelength of 488nm.

2.6.2 PI Analysis of Cell Cycle

Human MPCs from four patient strains were plated at 50,000 cells per well in 6-well plates for each condition, allowed to adhere overnight, and treated with the respective condition. After treatment, the cells were fixed by suspending in ice cold PBS and vortexed gently while cold 70%

ETOH was added dropwise to prevent clumping. The solution was stored overnight at 4°C. Cells were then washed twice with PBS, pelleted and stained with 50µl of 100 µg/ml ribonuclease

(Sigma-Aldrich) and 200 µl of 50 µg/ml propidium iodide (Sigma) and incubated for 30 minutes at room temperature. Samples were then run on an AttuneTM Flow Cytometer and analyzed using

AttuneTM software.

2.7 Drug Kinetics

2.7.1 p21 protein nuclear localization

Human MPCs from four patient strains were plated at five thousand cells per well in black Grenier uClear® 96 well microplates. Cells were allowed to adhere overnight then treated with the respective drug treatment for 24 hours, 4 hours, 1 hour, or 0.5 hours in triplicate. Cells were then fixed with 4% PFA for 15 minutes at room temperature, washed with PBS twice and stained with

50µl of 2.5µg/ml Purified Mouse Anti-Human Cip1/p21 (BD Transduction Laboratories) in 0.1%

Tween 20 for 48 hours at 4°C. Cells were then washed 3 times with PBS and stained with 0.2%

DAPI and 1% AlexaFlour® 488nm goat anti-mouse secondary antibody (Invitrogen) in 1% Tween

20 for 45 minutes at room temperature in the dark. Cells were washed 3 times with PBS and left in 100µl PBS for imaging. An InCell 2000 (GE Lifesciences) was used at 10x magnification to take 9 pictures of each well at 405nm (DAPI) and 488nm (p21). Resulting image stacks were then analyzed using the InCell Analyzer software. The nucleus was defined from the DAPI fluorescent area and the cytoplasm was defined as a 3 µm circumference around the outline of the DAPI nucleus as seen in Figure 7. The nucleus and cytoplasm areas were then analyzed on the p21 photo for staining (fluorescent intensity) of the nucleus versus staining (fluorescent intensity) of cytoplasm. If this ratio was greater than 1, then the cell was deemed “nuclear positive”, meaning there is more p21 in the nucleus than the cytoplasm. If the ratio was less than 1, then the cell was deemed “nuclear negative”, meaning there is less p21 in the nucleus than the cytoplasm. Each picture was analyzed and the results were compiled per well. The results were then analyzed for each patient using Graphpad Prism 6.0.

Figure 7: Representative example of p21 fluorescent staining of a nuclear negative and nuclear positive result.

2.7.2 p21 mRNA (qPCR)

Human MPCs from three patient strains were plated at 5000 cells per well in 96 well dishes and allowed to adhere overnight. Cells were treated with each drug condition for 24 hours, 4 hours, 1 hours, and 0.5 hours. Cells were trypsinized and pelleted. The RNA from each sample was extracted using E.Z.N.A® Total RNA Kit I (OMEGA). RNA was converted to cDNA using High

Capacity Reverse Transcriptase cDNA kit (Invitrogen). mRNA levels were then analyzed using

TaqMan® Universal PCR Master Mix (Applied Biosciences) using TaqMan® Gene Expression

Assay primers (Applied Biosciences) for human CDNK1A (p21) and 18S (endogenous control) on 7900HT Fast-Real-Time PCR System. Samples were run in triplicate and analyzed against a carrier control (DMSO) using Graphpad Prism 6.

2.8 Chondrogenesis

2.8.1 Culture

Human MPCs from six patients were pelleted at 50,000 cells per pellet in 1.5 ml Eppendorf tubes by centrifugation. Cells were cultured in DMEM-F12 (LONZA) media with 10% FBS, 1% Anti-

Anti (GIBCO), and 1% non-essential amino acids (GIBCO), as well as the respective drug condition for 21 days. Chondrogenic control (chondro) was cultured in media containing 500 ng/mL BMP-2 (Peprotech), 10 ng/mL TGF-β3 (Peprotech), 10 M dexamethasone, 50 µg/mL ascorbic acid, 40 µg/mL proline, 100 µg/mL pyruvate and supplemented with 1× insulin, transferrin and selenium (all Sigma, St. Louis, MO), and the pH of the final solution was adjusted to 7 with NaOH. Undifferentiated control was pelleted and cultured in media without any growth factors or drugs and analyzed after 3 days rather than the full 21. At least two pellets per condition

were used (for histology and mRNA analysis) as well as a carrier control and media only control.

Media was changed every 2-3 days, with special attention not to disturb the pellet.

2.8.2 qPCR (mRNA)

Pellets were lysed in TRIzol (Ambion) with an 18 or 21-gauge needle, then total RNA was extracted using E.Z.N.A® Total RNA Kit I (OMEGA). Briefly, 200 µL of chloroform per milliliter of TRIzol was added to the TRIzol solution. Samples were vortexed, incubated at room temperature for 15 minutes, then centrifuged for 15 minutes at 4°C and 12000 rpm. The top, clear, layer was transferred into a new HiBind® RNA mini column in a 2ml collection tube. The sample was centrifuged at 10,000 g for 1 minute, and the filtrate was discarded. The sample was washed with 500 µL wash buffer I, centrifuged for 30 seconds at 10,000 g and the filtrate was discarded.

The sample was washed with 500 µL of wash buffer II, centrifuged for 1 minute at 10,000 g, and the filtrate was discarded. The sample was then centrifuged at 14,000 g for 2 minutes to remove any excess liquid. The collection tube was replaced with a new one and the sample was extracted by adding 50 µL of ultra-pure water (Quality Biological) and centrifuging at 14,000 g for 2 minutes. The sample was immediately stored at -80°C until the next step could be completed.

RNA was converted to cDNA using a High Capacity Reverse Transcriptase cDNA kit (Invitrogen). mRNA levels were then analyzed using TaqMan® Universal PCR Master Mix (Applied

Biosciences) using TaqMan® Gene Expression Assay primers (Applied Biosciences) for human

CDKN1A (p21), SOX9, ACAN, and 18S (endogenous control) on a 7900HT Fast-Real-Time PCR

System. Samples were run in triplicate and resulting threshold (Ct) values were analyzed using the

DDCt method against 18S endogenous control and DMSO as the reference untreated control. The

DDCt method can be described as follows:

:∆∆<=>?@ABC !"#$%&'" )*+, "-.!"//&01234567 = 2 where,

∆∆D%234567 = (D%F34567 − D%234567 HI2) − (D%K7L − D%K7L HI2)

The relative mRNA expression was plotted using GraphPad Prism 6.0. Statistical analysis was done on the DDCt values, as the relative mRNA expression values are log transformed.

2.8.3 Histology (proteoglycan)/ Pellet Size

Pellets were fixed overnight at 4°C in 4% paraformaldehyde (BDH) and then washed 3 times with

PBS. After fixation, photos were taken using Axio Zoom® (Zeiss) of each pellet. The area of the pellet was determined for each using ImageJ software and analyzed using Graphpad Prism 6.

Alcian blue staining was used to stain for proteoglycans. Pellets were then transferred to 90% ethanol for 48 hours at 4°C, then stained with Alcian Blue (2mg Alcian Blue; Sigma, 0.8ml water,

16ml 100% ethanol, 4ml glacial acetic acid) for 48 hours at 4°C. The pellets were then rehydrated in progressively more dilute ethanol solutions (70%, 50%, 25%, 20% ethanol) and then placed in

PBS for visualization using the Axio Zoom.

2.9 In vivo Mouse Wound Healing Model

2.9.1 Mice

Ten 4-week C57Bl/6 (C57) mice and fifteen 4-week p21-/- mice were used for the ear wound healing experiments. Mice were split into four treatment groups, C57 low dose (n=4), C57 high dose (n=4), C57 control (n=2), p21-/- high dose (n=4), and p21 control (n=11). All mice were placed under isoflurane anesthesia and a 2 mm through and through ear punch was given to the

center of the left ear at week 0. Mice in the drug group were given topical treatment of low dose

(0.1 µM), high dose (100 µM) or control (0 µM) drug 70 in DMSO gel (Life Choice) adjacent to the wound. Mice were treated with the respective condition once per week for 4 weeks. Images of the wound were taken with a size standard in frame once per week for 4 weeks. Images were analyzed for wound diameter using ImageJ software.

2.9.2 Histology

Ears were harvested and fixed in neutral buffered formalin (NBF) (Protocol) for 48 hours. Samples then underwent tissue processing using a Leica TP1020 automatic tissue processing machine.

Briefly, the process entails a 1 hour water bath, 1 hour each of progressively stronger ethanol washes (80%, 95%, 95%, 100%, 100%, 100%), three 1 hour xylene washes, and two 1 hour paraffin baths. Tissue was then embedded in paraffin wax blocks and sections were cut with Leica

Automated Rotary Microtome (RM 2255) at 10µm sections and transferred onto slides. Slides were dehydrated at 37°C overnight before staining.

Safranin-O/Fast Green The slides were stained with safranin-O and fast green and imaged using an Axio Scan (Zeiss) at 10x magnification. Briefly, the slides were deparaffinized with three 10 minute washes of Citrisolve and rehydrated with an ethanol series (100%, 95%, 70%) to distilled water. Slides were placed in Gill’s hematoxylin (Fisher) for 8 minutes followed by 15 minutes in running tap water. Slides were dipped in Acid Alcohol, distilled water, and ammonia water for 1 minute each then transferred to Fast Green (Fisher) for 7 minutes followed by 1 minute in 1% acetic acid. Slides were quickly rinsed with distilled water (1 minute) and then stained with safranin-O (Fisher) for 1 minute and 45 seconds. With quick dips, slides were dehydrated with a

progressive ethanol series (70%- 30 seconds, 95%- 1 minute, 100%- 3 minutes) to 100% ethanol after which they were cleared and mounted. Images were compiled and analyzed with Zeiss Zen software.

2.10 Statistics

Statistics on alamarBlue, EDU, propidium iodide, and mouse ear wound healing was done using a standard one-way ANOVA with a Holm-Sidak multiple comparisons test used to determine significance relative to the treatment group relative to the DMSO carrier control group. The p value cut off was set to 0.05. RT-qPCR (both for p21 expression and following chondrogenesis) was analyzed with a standard two-way ANOVA for each drug condition with a Dunnett multiple comparisons test to determine the significance relative to the DMSO carrier control. Analysis of

RT-qPCR was done using the DDCt values rather than the plotted logarithmically transformed R values, as a normally distributed data set is required for this analysis. The p value cut off was set at 0.05 for these tests as well. All statistics were done using GraphPad Prism 6 software.

Chapter Three: Results

3.1 Identification of p21 expression inhibitors

3.1.1 Drug screening

Genetically modified cells (XMAN™) derived from the HCT116 human colon cancer cell line expressing luciferase under the control of the CDKN1A (p21) promoter were utilized in a high throughput screen to identify potential p21 inhibiting compounds. A drug library of 146 small molecule compounds targeting the complete spectrum of kinases implicated in cancer cell growth and survival (Table 5 in Appendix A) was chosen for the initial screening due to the tumor suppressive nature of p2179. XMAN™ cells were exposed to each compound at four concentrations

(0.01 µM, 0.1 µM, 1 µM, and 10 µM) and the luminescence was measured after 24 hours of treatment (Figure 8). From this screening the five compounds that met the criteria of lowest luminescence and a concentration dependent decrease in luminescence were chosen for further testing.

Figure 8: Identification of potential p21 expression inhibitors on XMAN® cell line. Drug screening of library of 146 small molecule inhibitors targeting the complete spectrum of kinases implicated in cancer cell growth and survival. XMAN® cells were plated in each drug condition in duplicate for each drug condition, the luminescence of the plate-specific media only control was subtracted from each condition and plotted after 24 hours of drug treatment.

Five compounds that showed the most p21 reduction were chosen to continue with the study

(Figure 9). These will be referred to as Drug 70, 93, 102, 107 and 111 throughout this thesis. Their chemical names, their known mode of action (pathways inhibited), and their IC50 according to literature are summarized in (Table 1).

Table 2: Selected compounds for p21 transcription inhibition, name, known mechanism of action, and IC50 according to literature.

Number Name Mechanism IC50 (µM)

70 17-DMAG HSP 90 inhibitor 0.062

93 CUDC-101 HDAC/EGFR/HER2 inhibitor 0.004/0.0024/0.0157

102 INK-128 mTOR kinase inhibitor 0.001 mTORC1 and mTORC2 0.022/0.065 107 OSI-027 inhibitor 111 PF-04691502 P13K/mTOR inhibitor 0.032

3.1.2 Metabolic toxicity of drugs on XMAN cells

To assess if the decrease in luminescence observed in the initial screening could be accounted for because of fewer cells remaining (due to cell death), the metabolic toxicity of each of the 5 candidate p21 expression inhibitors on the XMAN cells was assessed using an alamarBlue assay

(which measures metabolic activity of the cells) after 96 hours of treatment. None of the 5 candidate p21 expression inhibitors demonstrated a metabolically toxic effect on the XMAN cells over the range of concentrations assessed (Figure 10, black bars). Furthermore, luminescence was also re-examined in the same cells, and although some drugs did not show significant luminescent decrease after 96 hours, most drugs showed a continued decrease validating the observations in the initial screening (Figure 10, gray bars).

*p<0.05.

3.2 Characterization of chosen compounds on MPCs during culture

3.2.1 Metabolic toxicity on MPCs

The five putative p21 expression inhibitors were then examined in the relevant cell strain of interest, human mesenchymal progenitor cells (MPCs) derived from synovium tissue. The metabolic toxicity was again assessed using alamarBlue to detect the metabolic activity of the cells after drug treatment (Figure 11, Table 3). As there is no universal acceptable limit, the cut-off used for this experiment was 85% of the carrier control96. Drugs 70, 102, 107, and 111 dropped below this viability cut-off at concentrations higher than 0.1 µM in the human MPC cell strains. Drug 93 did not drop below this cut-off and therefore the highest concentration tested, 10 µM, was used.

Therefore, for the remainder of the experiments MPCs were treated with 0.1 µM of drugs 70, 102,

107, and 111, and 10 µM of drug 93. Carrier control was treated with 10 µM DMSO.

Statistics were performed using a standard one-way ANOVA with *p<0.05

Table 3: Summarized alamarBlue results. Grey shaded box is the only significant result (p<0.05) compared to the DMSO carrier control.

0.01 µM 0.1 µM 1 µM 10 µM Standard Standard Standard Standard Drug Absorbance Deviation Absorbance Deviation Absorbance Deviation Absorbance Deviation 70 99.70 5.43 86.84 8.05 83.81 11.69 81.66 15.79 93 103.01 13.01 100.57 2.40 99.90 7.32 97.13 5.61 102 101.19 3.01 85.32 3.27 83.79 8.63 74.22 17.42 107 101.58 5.22 97.44 1.69 82.66 10.00 79.79 17.75 111 96.47 2.41 91.42 8.21 76.47 17.36 81.11 12.28

3.2.2 Cell Cycle

Since p21 is a cell cycle inhibitor, acting mainly at the G1/S checkpoint to impede cells from dividing until they have undergone proofreading, the proliferation and cell cycle was analyzed while exposed to these potentially p21 inhibiting compounds. Proliferation of the human MPC cell strains (n=3) was analyzed after 24 hours of drug treatment using an EDU assay (Figure 12). Drugs

93 and 70 demonstrated a significant increase in proliferation compared to the DMSO control, while drugs 102, 107, and 111 did not show any change in proliferation after 24 hours of treatment, however there was a large amount of variability seen between patient cells when treated with drugs

102, 107, and 111, however there is no obvious demographic difference between pt 10 and pts 8 and 9.

The distribution of human MPCs within the cell cycle with drug treatment was analyzed specifically using a propidium iodide assay. MPC cell strains (n=4, pt 8, pt 9, pt 11, pt 12) were treated for 4 and 24 hours with each drug and then analyzed using flow cytometry (Figure 13). The data is summarized in

Table 4. All treatment groups except drug 70 showed an increase in the G1/G0 accumulation after the 24 hour treatment compared to the DMSO control at 4 hours. Drug 70 showed an increase in the G2/M accumulation after 24 hour treatment compared to the DMSO control after 24 hours of treatment, whereas drugs 102 and 104 significantly increased the G1/G0 accumulation compared to the same control. Since inhibition of p21 allows cells to freely enter the cell cycle, a p21 expression inhibitor should allow the accumulation of cells in G2/M, and this has been observed as a phenotype in fibroblasts from p21-/- mice61.

Table 4: Percentage of MPCs in each stage of the cell cycle. Shading indicates significant difference relative to the DMSO control at 24 hours, p<0.05

% of cells Apoptotic G0/G1 S G2/M 4hr 0.3 59.3 18.2 19.9 DMSO 24hr 0.9 75.3 10.8 12.0 4hr 0.6 65.2 15.1 17.2 70 24hr 0.8 62.3 4.6 31.1 4hr 0.3 63.1 17.0 17.5 93 24hr 0.7 74.5 6.9 16.4 4hr 0.4 65.2 15.6 17.0 102 24hr 0.7 86.4 4.0 8.3 4hr 0.7 65.5 15.4 16.9 107 24hr 0.5 83.0 7.8 7.9 4hr 0.5 65.5 14.8 17.4 111 24hr 0.9 85.7 5.0 7.8

3.2.3 mRNA p21 p21 mRNA was assessed in human MPC strains (n=3, pt 4, pt 11, pt 12) at four different time points of drug treatment over 24 hours (0.5 hours, 1 hour, 4 hours, and 24 hours) using RT-qPCR in order to determine if the drug treatments were acting to decrease p21 transcription in this cell type. Three patient MPC strains were plated and exposed to each drug and cultured for each time point. Total RNA was extracted, converted to cDNA and analyzed with RT-qPCR using 18S as the endogenous/housekeeping control. No drugs significantly differed from the DMSO control group with the exception of 111 after 4 hours of treatment and 107 after 24 hours of treatment

(Figure 14). At these time points there is a large amount of patient variability in the data and it would be of interest to add more patients to this study in order to determine if this variability is due to experimental error, patient variability, or an anomaly. Contrary to what we would expect,

these results differ from the initial drug screening results in the XMAN® cell line. This discrepancy will be discussed further in Chapter 4.

compared to the DMSO only control. No drugs caused a significant decrease in p21. Drug 93 increased p21 at all time points and drug 111 increased p21 at 0.5 and 4 hour time points. *P value was evaluated at < 0.05 for these tests.

3.2.4 p21 nuclear localization

The protein level of p21 in human MPCs (n=4, pt 8, pt 9, pt 11, pt 12) over the same 24 hour period as the mRNA level was analyzed using fluorescent antibody staining and analyzed using InCell

Analyzer software. Since p21 is actively regulating the cell cycle when it is found in the nucleus and involved in apoptosis when residing in the cytoplasm, the amount of p21 in the DAPI-defined nucleus was compared to the amount in the surrounding 3µm, representing the cytoplasm. The percent of cells with nuclear localization, as defined as nuclear/cytoplasmic fluorescent p21 staining greater than 1 was found for each condition at each time point and normalized to the respective patient’s DMSO carrier control treatment group (Figure 15). None of the drugs tested changed the nuclear location of p21 within the human synovial MPCs at any of the time points examined. Wortmannin was used as the positive control, as it increases nuclear localization of p21, however, although there seems to be a positive increase in nuclear localization after 24 hours of treatment none of the drug treatments showed a significant difference compared to wortmannin.

*p < 0.05.

3.3 Characterization of chosen compounds on the chondrogenesis of human MPCs

3.3.1 Pellet size

The putative p21 expression inhibitors were then analyzed for their effect on MPCs throughout chondrogenic differentiation. Human MPC strains (n=6, pt1, pt 5, pt 6, pt 7, pt 11, pt 12) were pelleted and cultured while in normal culture media (e.g. no chondrogenic factors – BMP2, TGF-

β) containing the drug treatment for 21 days. Pellet size is an indicator of both cell growth and

ECM deposition, which are known to occur at the beginning stages of chondrogenic differentiation97. Pellet size was assessed at the end of the 21 day differentiation period. Compared to the undifferentiated control, the cells pelleted with chondrogenic growth factors (chondro), drug

70, 93, 102 and 111 showed a significant increase in pellet area (Figure 16). Increase in pellet size could indicate an increase in cell number or more ECM deposition therefore alcian blue was then used to stain the pellets for proteoglycan content.

Figure 16: Pellet size of MPCs after 21 day pellet culture in the respective drug treated media significantly increased in area when treated with chondrogenic growth factors (chondro), drug 70, 51

93, 102, and 111 when compared to the undifferentiated control. Two pellets for each condition are assessed here. *p<0.05.

3.3.2 Alcian blue staining

The pellets were then stained for proteoglycan content using alcian blue. All pellets showed positive staining for alcian blue, and therefore proteoglycans, compared to the undifferentiated control (Figure 17). To confirm that chondrogenesis had taken place, chondrogenic marker quantification was assessed from the pellets using RT-qPCR.

Figure 17: Pellets under all drug treatments stain positive for alcian blue after 21-day pellet culture compared to the undifferentiated negative control.

3.3.3 Chondrogenic differentiation and analysis mRNA was then extracted from the pellets after the 21 day chondrogenesis, converted to cDNA, and analyzed using RT-qPCR for p21 as well as chondrogenic markers SOX9 and Aggrecan

(Figure 18). p21 was upregulated compared to the DMSO control when treated with drug 70 (4/6 pts), drug 102 (2/6 pts) and drug 93 (2/6 pts), however drug 93 also downregulated p21 with respect to the DMSO control group in two of six patients. SOX9 was upregulated by drug 70 (5/6 pts), drug 93 (3/6 pts), drug 102 (5/6 pts), drug 107 (3/6 pts), and drug 111 (4/6 pts). Interestingly, drugs

93, 107 and 111 also significantly downregulated SOX9 in one patient each. ACAN, which codes

for Aggrecan, was only detectable in two patient MPC strains, and in these two patients Aggrecan was upregulated with all drugs in patient 1 and 70, 93, and 111 in patient 2.

3.4 Characterization of Drug 70 on in vivo chondrogenesis

Drug 70 was chosen to move forward with the in vivo experiments since exposure of MPCs to this drug led to an increase in proliferation, an accumulation of the MPCs in the G2/M phase, furthermore, and it was able to induce the expression of both SOX9 and Aggrecan in 5/6 patients during in vitro chondrogenesis.

3.4.1 In vivo Cartilage Regeneration (0.1µM treatment group)

Since mouse ear tissue contains an elastic cartilage layer which is made up of collagen II, similar to articular cartilage, an injury model examining ear cartilage regeneration was explored with drug

70 at a concentration of 0.1 µM. Through-and-through (2mm) ear punches were performed on

C57BL/6 non-regenerating mice (n=7). The wound was treated with 0.1µM drug 70 (n=4) once a week for 4 weeks, and compared to DMSO control treated mice (n=3). The wound closure was tracked over this time period using repeated measures and compared to experimental data preformed previously in our lab of the same experiment on p21-/- mice (n=10). There was no significant difference between drug treated and control C57 ear injuries after four weeks, however as early as week 1 the p21-/- mice demonstrated significantly reduced wound size (Figure 19).

Figure 19: In vivo results of drug treatment on through-and through ear hole closure. Drug treated

C57/BL6 mice (n=4) showed no significant difference to the control DMSO treated mice group

(n=3). p21-/- mice show a significant reduction in ear hole size at week 1, 3 and 4. *p<0.05.

Safranin-O/Fast Green staining

The ear tissue was harvested after sacrificing the mice at 4 weeks and prepared for histological analysis. Sections were stained with safranin-O/fast-green to visualize the layer of cartilage tissue within the ear. The control sections showed very little evidence of cartilage repair/regeneration

(small cartilage condensation can be observed in one control mouse, arrow (Figure 18)), whereas in the drug treated animals, extensive cartilage formation can be observed. Specifically, a full second layer of cartilage can be observed developing adjacent to the original damaged cartilage. It appears that the cartilage has bent over on itself growing back into the main body of the ear and away from the defect area. This was not observed in the carrier control treated animals.

3.4.2 100µM Ear hole closure

A similar ear hole punch experiment was performed on 4 C57 mice with a higher concentration of drug 70 (100 µM) in DMSO gel. Four p21-/- mice were also treated in order to assess if there was any hindrance of ear/cartilage regeneration with drug 70 since it mode of action in this model is unknown. Ear hole punches (2mm) were induced at week 0 and the drug treatment was given once a week for four weeks while measuring ear hole closure (Figure 21). The mice were then sacrificed and the ears were processed for histological analysis with safranin-O/fast green as in the previous experiment (Figure 22 and Figure 23). Again, there was also no reduction in ear hole closure compared to the DMSO carrier control in neither the C57 nor the p21-/-, with the exception of p21-

/- at week 1. However, the histological staining showed cartilage formation in treated but not untreated C57 mice with high dosage in 3 of 4 mice, similarly to the low dosage experiment. p21-

/- mice showed some cartilage formation in both the untreated and treated histology, however it seemed to be less pronounced than in the C57 mice in the drug treatment groups.

Chapter Four: Discussion

As discussed previously, cartilage degeneration is extremely common in our aging population, yet cartilage regeneration is elusive. It has long been debated whether there is any amount of naturally occurring cartilage turnover over the course of a life time. Recently Heinemeier et al. elegantly showed, using radiocarbon pulse dating, that there is little to no collagen turnover in articular cartilage after skeletal maturity, both in healthy and osteoarthritic joints51. Although mesenchymal progenitor cells have been identified in the surrounding tissues of joints (such as the synovium, fat pad, meniscus and ligaments), there has been little evidence that these cells endogenously differentiate into chondrocytes in vivo. That being said however, Kurth et al. showed synovial

MSCs are capable of migrating and differentiating into chondrocytes after the induction of a focal cartilage injury in mice29. It has been demonstrated repeatedly that progenitor cells derived from the synovium and synovial fluid of humans have the capacity to differentiate into bone, fat and cartilage tissue in vitro30,45,98,99. These synovial MPCs even have a greater preference towards chondrogenic differentiation compared to bone marrow derived stem cells28. If these resident synovial MPCs can be stimulated to differentiate into new cartilage in vivo, then cartilage degenerative diseases, such as osteoarthritis, can be more appropriately treated in the future.

Cartilage regeneration is demonstrated naturally in urodeles, such as newts and salamanders100–102.

When a limb is removed a mesenchymal growth zone, or blastema, is formed locally101. This clump of cells rapidly expands then differentiates into the correct tissue101. Regeneration of articular cartilage has been demonstrated after full-thickness defects in axolotl salamanders103.

Regeneration of tissues in adult mammals is generally poor. Although many fetal tissues have the ability to regenerate properly, adult tissues tend to heal with incorrect tissue, or scarring. As discussed in Chapter 1, it was discovered that the MRL mouse has the ability to regenerate multiple

tissues59. From genetic analysis of this model, p21 was identified as the single point mutation responsible for this healing ability61. This lab showed that p21-/- mice are able to heal ear hole punch wounds over 4 weeks and our lab was able to confirm this result (Figure 24). Our lab has also shown that p21-/- mice regenerate morphologically correct articular cartilage following a full- thickness articular defect, whereas wild-type mice do not (Figure 25).

Since cartilage is present both within the ear and joint surface, these results may suggest that p21 not only plays a role in wound healing, but that it might also play an important role in cartilage and/or the process of chondrogenesis. Cheverud et al. have demonstrated the link between ear hole wound closure and protection from OA using two mouse lines (LG/J and SM/J) from the major parental strain of the MRL mouse104. Along this line of thought, it has been demonstrated that knocking down p21 using short hairpin RNA in murine induced pluripotent stem cells was able to increase their chondrogenic capacity105, strongly suggesting that p21 plays a role in the formation of cartilage. Therefore, since the p21 gene and protein are highly conserved in mammals the aim of this project was to identify a pharmacological approach that would inhibit p21 transcription thereby increasing the chondrogenic capacity of human synovial MPCs. In this thesis I have attempted to examine the hypothesis that p21 inhibition will lead to increased chondrogenic potential of human mesenchymal progenitor cells and enhance cartilage regeneration in vivo. In order to test this hypothesis, I (1) identified p21 inhibiting compounds using a high throughput screening method and characterized the effect of the selected compounds on human synovial mesenchymal progenitor cells during in vitro culture, and (2) characterized the effect of the selected compounds on human mesenchymal progenitor cells during chondrogenic differentiation

in vitro, and determined if the(se) compounds could induce cartilage regeneration in vivo using a mouse wound healing model.

Figure 24: p21 knockout mouse ear hole closure over 4 weeks compared to C57 wildtype controls and MRL mouse ear hole closure (Unpublished data from the Krawetz lab).

From the initial drug screening, five compounds were identified that inhibit p21 transcription in a genetically modified HCT116 human colon cancer cell line (XMANTM); these drugs have been referred to as 70, 93, 102, 107, and 111. Small molecule kinases are currently the most rapidly growing type of FDA approved drug106. Over the last 15 years 28 new small molecule kinases have been approved, 15 of which were approved between 2012-2015106. These have been mostly studied in the context of anti-cancer drugs, however research is beginning to emerge for the treatment of inflammatory diseases107,108. In the context of this project we have chosen small molecule kinase inhibitors as a method to modify p21 expression at the transcriptional level in order to mimic the effects of p21 loss/knockdown in a mouse model. As mentioned previously, there have been a select few p21 modulators identified in the literature. These include butyrolactone I87, LLW1085, sorafenib89, and UC228884. Butyrolactone , LLW10, and sorafenib, are known to inhibit p21 by protein degradation, whereas UC2288 modulates p21 at the transcriptional or post-transcriptional level84. UC2288 treatment on various human cancer cell lines showed a decrease in cytoplasmic

p21, but not nuclear p2184. It is known that cytoplasmic p21 has an anti-apoptotic affect, through

PCNA binding, whereas nuclear p21 is responsible for cell cycle inhibition, through CDKs79,109.

For the purposes of this study an inhibitor of all p21, both nuclear and cytoplasmic, was sought for induction of chondrogenesis. Prior to my involvement in the study, these known p21 expression inhibitors (except UC2288) were tested in human synovial MPCs and it was observed that they did not downregulate p21 transcription nor regulated the cell cycle (unpublished data from Dr.

Ricarda Hess / Krawetz lab). Therefore, it was necessary to undertake a screening approach to identify novel p21 transcription inhibitors.

Regulating p21 transcription is a complicated matter. Although p53 is the main enhancer of p21, there are many other proteins that are able to bind to the p21 promoter region to regulate transcription (Figure 26). p21 inhibition through the p53 pathway has been shown to be insufficient to induce the regeneration effect observed in p21-/- mice, as p53-/- mice do not exhibit a regenerative phenotype74. None of the aforementioned compounds identified in the screen (including the 5 selected drugs) decreased p21 transcription in human synovial MPCs. The 5 compounds chosen from the screening, have been previously characterized as inhibitors that target pathways upstream of p21. These included a HSP90 inhibitor (drug 70), HDAC/HER2/EGFR multi-inhibitor (drug

93), two mTOR inhibitors (drugs 102 and 107), and a PIK3/mTOR inhibitor (drug 111).

Figure 26: Relevant transcriptional binding sites in the p21 (CDKN1a) promoter region adapted from information provided by The Champion ChiP Transcription Factor Search Portal by

Qiagen110.

Although each of the drugs significantly decreased p21 transcription in the XMANTM cell line, we did not observe the same decrease in p21 mRNA in the human MPC strains. There are several possible explanations for this difference. Because p21 acts on the cell cycle, a cancerous cell line for the initial screening may not be ideal. The limitations of the HCT116 cell line are discussed further in the limitations section, however the most notable differences in the HCT116 cell line vs. human MPCs are the decreased amount of endogenous p21 and the mutations in multiple genes including CDKN2A, CTNNB1, PIK3CA, and KRAS111,112. These genes code for CDK4, β- catenin, PI3K, and RAS respectively. All of these genes are involved in the cell cycle and may affect the pathways that the drugs are acting on. Drugs 102, 107, and 111 all act directly on the

PIK3/AKT/mTOR pathway, which is downstream of the RAS mutation, to inhibit p21 in XMAN cells. Therefore, while the drugs decrease p21 in the cancerous cell line with the RAS mutation,

the MPC cell strains without this mutation may have a higher amount of baseline p21 expression to begin with and a functional PIK3/AKT/mTOR pathway in which to inhibit, therefore a higher concentration may be necessary for these drugs to be effective in our cell strain of interest. Drug

93, an HDAC/HER2/EGFR multi-inhibitor, also appeared to upregulate p21 mRNA at all time points examined. The EGFR pathway is upstream of the PIK3/AKT/mTOR pathway and this may be the reason inhibition of p21 is not observed with this drug treatment in MPCs. Additionally, drug 93 showed the weakest p21 inhibition in the XMANTM cells of the 5 chosen drugs, therefore suggesting that if higher concentrations are needed in MPCs, the starting concentration of drug 93 may have been particularly low in MPCs. Drug 70, the HSP90 inhibitor, showed no change in p21 mRNA with respect to the no treatment control. Although this drug treatment did decrease the p21 transcription in XMANTM cells, the higher level of endogenous p21 may require a higher concentration of drug treatment in order to see an observable decrease, yet drug 70 did demonstrate some properties of a p21 expression inhibitor in terms of cell cycle regulation.

It would be ideal to directly observe the level of p21 transcription in human MPCs in a similar fashion as the XMANTM reporter line. Towards this goal, we first attempted to transfect human

MPCs with a CRISPR construct to knockout p21 completely to mimic the cells in the p21 knockout mouse model in a human cell strain. If this were successful it would have acted as a complete knockout and positive control to compare the drug treatments to as well as assess the effect of complete p21 inhibition in a human MPC cell strain undergoing chondrogenesis (which has not been reported to date). Additionally, a similar reporter gene could have been added at the promoter region of p21 in order to directly assess the p21 transcription as we had done using the XMANTM cell strain. Unfortunately, despite attempting numerous chemical transfections (TurboFectTM,

Lipofectamine® 3000, X-tremeGENETM) as well as nucleofection, the primary MPCs could not survive the transfection process. This will be discussed further in the limitations section.

In terms of examining cell cycle regulation by the drug treatments, it is essential to understand how cell cycle control is regulated in the MRL and p21-/- mice that both demonstrate endogenous cartilage regeneration. Fibroblasts derived from MRL mice demonstrate an increase in DNA damage, increased migratory and proliferative capacity, and an accumulation in the G2/M phase of the cell cycle62. Fibroblasts from p21-/- mice show similar characteristics, and therefore we examined the indirect effects that these drug treatments have on the cell cycle of MPCs in order to indirectly assess if they might be affecting p21 control of the cell cycle. When treated with drugs

70 and 93, human MPCs significantly increased proliferation compared to the DMSO control.

Drugs 102, 107, and 111 did not induce and change in the proliferation of MPCs. An increase in

MPC/MSC proliferation is necessary in the onset of wound regeneration. In the formation of a blastema, surrounding cells dedifferentiate and MPCs actively proliferate prior to the re- differentiation phase100.

G2/M accumulation is also an important part of the regenerative phenotype in axolotls and the mammalian liver102,113,114. Interestingly, fibroblasts derived from the p21 knockout mouse were also found to accumulate in the G2/M phase62. When human MPCs were treated with the 5 selected drugs for 24 hours, only drug 70 induced a shift from G0/G1 accumulation to cells accumulating in G2/M. Because p21 mainly acts to inhibit progression through the G1/S checkpoint after DNA damage, it is counter intuitive that a p21 inhibition would arrest the cells in the G1/G0 phase.

Therefore, when treated with a compound that inhibits p21 the G1/S checkpoint should be largely

unregulated and cells can progress to the G2/M phase. As stated, this effect was only observed with drug 70 treatment.

Next, the amount of p21 protein in the nucleus and cytoplasm was assessed using immunofluorescent staining over 24 hours of treatment. Unfortunately, the relative amount of staining (i.e. fluorescent intensity) could not be compared between images taken due to software constraints. The relative amount of p21 in the nucleus vs the cytoplasm of each cell was assessed using InCell analyzer software and the percentage of cells with more staining in the nucleus than in the cytoplasm was calculated for each condition. This number was then normalized to the percentage of nuclear positive cells at time 0 for each cell strain analyzed. Overall there were no significant differences between the treatment groups and time 0. This could indicate that p21 protein is not being selectively depleted or enhanced either the nucleus or cytoplasm. This is could be interpreted as a positive or negative result, as the goal of the study is to deplete p21 transcriptionally, and therefore the overall p21 protein should be decreased in turn. Many of the known p21 expression inhibitors act at the protein level by degrading or destabilizing one aspect of p21. In order to mimic the cellular environment that has caused enhanced regeneration in mice, p21 must be knocked down to a greater degree. To determine the actual amount of p21 protein, a quantitative analysis of protein levels before and after drug inhibition would be useful in identifying if the protein level is changing, although 24 hours of treatment may not be a sufficient length to observe a change, as transcriptional regulation will not affect protein level until the current protein has been turned over.

To characterize the effect of these drugs during chondrogenesis we began with an in vitro approach. Diekman et al. recently demonstrated that knocking down p21 in induced Pluripotent

Stem Cells (iPSCs), increased chondrogenesis105. Similarly, our results also demonstrated that all drugs tested did increase the chondrogenic factors SOX9 and ACAN in some human MPCs. SOX9 is a transcription factor that is known to activate collagen IIa transcriptionally115. SOX9 is upregulated in the early stages of chondrogenic differentiation. Aggrecan is a large chondroitin sulphate proteoglycan, and is a major component of articular cartilage116. However, Aggrecan is not specific to articular cartilage. It is also a component of fibrous cartilage found in the intervertebral discs117. One study showed that micromass culture, rather than pellet culture, is superior at forming hyaline-like cartilage in vitro and may be worth looking into for future studies118. Alternatively, it may be possible that a longer differentiation period was warranted.

While most studies use between 14-28 days under chondrogenic differentiation stimuli, 21 days was selected in this study based on previous experience using growth factor enhanced media, however, it is possible that if we extended the differentiation period to 28 days, collagen IIa expression may be detected as it is a later stage marker. Along with chondrogenic markers, p21 was also upregulated in many of the drug conditions after chondrogenesis. While this is counterintuitive for potential p21 inhibiting compounds, the inhibition affect was tested short term in a different metabolically different cell strain in the studies leading up to this result.

Unfortunately, we do not have a method of directly testing the activity time of the drugs, as it has not been assessed how the drugs interact with p21. It is unclear whether the decrease in activity is due to dosage or differing interactions due to the change in cell type. Additionally, it has been shown that pelleting the cells alone is sufficient to induce chondrogenesis, and therefore can induce changes in the cell cycle119. The induction of chondrogenesis increases BMP-2 and TGF-b, which

are known activators of p21120–122. Our lab recently showed that p21 is induced with chondrogenesis in normal MPCs, but reduced in OA MPCs75, these results suggest that this change in state from undifferentiated to chondrocyte effects p21 expression. Because the drugs examined in this project were initially selected during a single state (undifferentiated) it is difficult to assess how the induction of chondrogenesis affects the interactions of the drugs.

To continue with the characterization of p21 inhibition on chondrogenesis we continued with an in vivo approach in a mouse model of chondrogenic regeneration. Drug 70, the HSP90 inhibitor, was chosen to continue on to the in vivo study due to its increase in proliferation, accumulation in the G2/M phase, and its induction of SOX9 and Aggrecan. While the literature on HSP90 inhibition and chondrogenesis is minimal, a recent study has demonstrated that injections of the

HSP90 inhibitor BIIB021 protected against biomechanically induced OA in rats123. Heat shock proteins are expressed as a protective response when cartilage is under stress, specifically HSP70 is thought to play a protective role in the early stages of OA124. HSP90 acts antagonistically to

HSP70 and creates an inflammatory response, which may explain the chondroprotective response of HSP90 inhibition123.

Although HSP90 inhibitors are currently in clinical trials for cancer treatments, common side effects of the systemic doses required include nausea, vomiting, fatigue, liver enzyme disturbances, blurred vision, dry eye, keratitis, and conjunctivitis125. While these side effects may be tolerated at low grades for cancer treatment, this is not appropriate for the treatment of a non- life threatening disease such as OA. The recommended dosage for 17-DMAG during the phase I clinical trial was 80mg/m2 administered weekly by IV125. Using standard conversion factors this

translates to a 26 mg/kg dosage in mice, or 0.5mg per 20g mouse126. Our study began by using the same concentration of 17-DMAG as was used in the cell culture and chondrogenesis studies, 0.1

µM. At 20µL per weekly application this calculates to 1.2x10-6 mg 17-DMAG per mouse ear. The high dose, of 100 µM, calculates to 1.2x10-3 mg per mouse ear, still over 400 times less than the phase I clinical trial dosage.

The mouse ear defect model was used for its non-invasive means of generating and following a defined cartilage injury over time to assess healing. Mouse pinna contain a thin layer of elastic cartilage and it has been shown that cartilage progenitor cells reside in the surrounding perichondrium127. Furthermore, as with articular cartilage, cartilage of the ear also demonstrates very little repair after injury, potentially due to the lack of blood supply into the cartilage layer.

Ear wound healing has been linked with articular cartilage regeneration potential and OA protection in mouse models in a study done by Cheverud et al.104. Therefore, although there are limitations with this model, as will be discussed later, it is an appropriate model for an initial in vivo study of cartilage healing.

While treatment with this drug did not show an increase in wound closure rate over the 4-week study, the histological results show cartilage-like growth in the treated ears of the C57 mice in both the low dose and high dose groups. The safranin-O/Fast green staining show red staining surrounding open vacuoles, or lacunae, similar to the cartilage layer that we see in the unaffected section of the ear, which is indicative of cartilage staining128. This indicates that the drug is promoting cartilage growth and that further studies should look into the mechanism and effect on articular cartilage and the translation to human cartilage repair.

As well as cell cycle regulation, p21 has also been shown to play a role in the regulation of the inflammatory response.. Interestingly, MRL mice were originally bred as a model for the autoimmune disorder Lupus and similarly, p21-/- mice develop severe lupus-like syndrome14 suggesting a role for p21 in the development of Lupus. More recently, it has been demonstrated that p21 is a suppressor of Lupus, and it has been shown that p21 is involved in effector/memory

T-cell activation15. It has also been demonstrated that p21 has the ability to suppress IL-1b mediated inflammation in activated macrophages16. Furthermore, mice lacking p21 block stromal derived factor-1 (Sdf-1) upregulation and the subsequent recruitment of Cxcr4-expressing leukocytes 17. Sdf-1 binds to Cxcr4, a chemokine receptor, and an increase in this signalling has been linked to increased fibrosis and scar formation in tissues including the lung, liver and heart17.

In articular cartilage Sdf-1 activates calcium, ERK, and p38 MAPK signaling pathways, inducing the release of MMPs and other cartilage degrading proteins18. In addition, Cxcr4 is overexpressed in OA chondrocytes18. Therefore, the concept of blocking Sdf-1/Cxcr4 signalling in OA joints is gaining attraction as a potential therapeutic and additionally, p21 may be working partly through inhibition of this pathway to promote scarless regeneration.

In specific regards to the cartilage observed in the histological sections of the drug treated mouse ears, while the tissue appears to be structurally similar to the existing cartilage in the pinna sections, the tissue has not been examined explicitly to determine the cell source of the cartilage or to confirm that is has recently been formed (vs. a change in existing tissue orientation). If this truly is new cartilage induced by the drug treatment, it could be generated from a variety of cell sources including de-differentiating chondrocytes, epithelial cells, differentiating mesenchymal, dermal or peripheral stem cells from the surrounding tissues or from the vascular system. Staining the

slides Ki67 could indicate if the cartilage or surrounding tissue was proliferating, and therefore could suggest if the cartilage tissue observed had recently proliferated or if it has somehow been rearranged due to the injury and treatment.

Although there have been many attempts to pharmaceutically protect cartilage from progressive

OA, most drugs have fallen out of clinical trials due to their inability to significantly reduce joint space narrowing, the current standard for clinical assessment of OA severity. Pharmaceuticals aimed at inhibiting ECM degrading enzymes such as MMPs and aggrecanases have shown successful results in pre-clinical and phase I trials, however so far none have made it past phase II clinical trials due to their inability to effectively reverse the disease129. Similarly, to prevent chondrocyte apoptosis, iNOS inhibitors have made it to phase II clinical trials, yet again, this treatment did not show a significant reduction in joint space narrowing compared to the placebo group and has since been discontinued130. Instead of targeting the processes that are degrading cartilage, some groups are more focused on promoting new cartilage growth, similar to this project.

To stimulate chondrogenic differentiation in vitro growth factors are added to the culture media of

MPCs/MSCs such as BMP-2 and TGF-β. This technique has been used to induce chondrogenic formation in in vivo as a therapeutic treatment. Human recombinant BMP-7 showed promising results in pre-clinical trials and has moved onto a phase II trial131. Similarly, a clinical trial of fibroblast growth factor-18 (FGF-18) has published successful results in a Phase I clinical trial132.

Interestingly, both BMPs and TGF-b are known to induce p21 transcription through Smad interactions133.

One promising substance for stimulating chondrogenic differentiation, as mentioned in Chapter 1, is Kartogenin (KGN). This is a small molecule that was selected from a small molecule screening that induce chondrogenesis in human MSCs. Although this compound is not in human clinical trials yet, a study was recently published showing promising results of KGN treatment by photo cross-linked IA injections in a rabbit model of OA134. In this study rabbits were given a full thickness cartilage defect in the knee joint. After surgery, KGN was loaded into PLGA nanoparticles and then suspended in m-hyaluronic acid (m-HA) for IA injection into the injured rabbit knee joint. The injected knee was then UV radiated for 1 minute, causing the PLGA to crosslink with the m-HA, creating a solid hydrogel containing KGN with a predictable and sustained release profile. After 12 weeks of recovery, the treated joints had formed hyaline-like cartilage where the defect was induced, compared to the untreated control, which did not express proteoglycans or collagen II in the defect. The group further looked at the binding ability of KGN and found that it binds to filamin A (FLNA), an actin-binding protein that crosslinks actin filaments, thereby regulating cyto-skeletal network organization and dynamics. Knockdown of

FLNA caused a five-fold increase in chondrocyte formation, confirming the relation to chondrogenesis. The pathway was further examined to find that KGN specifically blocks the interaction of FLNA with core binding factor b subunit, in turn regulating the RUNX family of transcription factors58. Interestingly, it is known that p21 is repressed transcriptionally by RUNX1, indicating that KGN may be indirectly targeting p21 transcription, potentially attributing some level of p21 regulation to the success observed with KGN in regulating chondrogenic differentiation potential135.

Limitations

This study was not without limitations. Although we had a CRISPR modified cell line with a luciferase reporter of p21 transcription activity available for our initial drug screening, this cell line demonstrates significant differences from the cell strain of interest for this project (synovial

MPCs). The cell line that was used in the initial screening, referred to in the text as XMANTM, is derived from the HCT116 human colon cancer cell line and was modified by the addition of a luciferase reporter on the promoter region of the CDKN1A (p21) gene. Human colon cancer is known to have a reduced amount of endogenous p21111 . While this does not affect the outcome of our screening results, as all drugs were tested on the same cell line, it may identify compounds that do not reduce p21 as efficiently in a cell strain that has a higher endogenous p21 level, such as the human MPCs used in this study. It is also known that there are mutations in the CDKN2A,

CTNNB1, PIK3CA, and KRAS genes in HCT116 leading to the cancerous nature of this cell line112. Products of these mutated genes, b-catenin, PI3K, and RAS all act on the

PI3K/AKT/mTOR pathway, in which 3 of the drugs chosen for this study directly inhibit. This p21 interaction may be amplified in comparison to non-mutated cells due to competing interactions. In human MPCs these genes are not mutated under normal circumstances, and therefore the drugs may act on p21 transcription less effectively in human MPCs compared to

XMANTM cells. To mitigate these differences and ensure that the chosen drugs downregulate p21 transcription we initially aimed to genetically modify human MPCs to report p21 transcription similarly to the XMANTM reporter line. Unfortunately, the technology available to transfect human primary bone marrow stem cells was not conducive with synovial MPCs and even though a number of attempts were made, no transfected populations were obtained. One limitation of the initial drug screening is that there was no control to indicate that the decrease in transcription of p21 was

specific to p21. The incorporation of a universal transcription inhibitor, such as an inhibitor of

RNA polymerase (α-amanitin, actinomycin D, DRB, flavopiridol, triptolide) would have been a useful and simple addition to the experiment in order to ensure specific p21 expression inhibition.

One possible limitation of the p21 mRNA quantification using RT-qPCR is the choice of a single reference gene, 18S. Often times, for increased reproducibility and standardization, multiple genes are used for reference in RT-qPCR. Reference genes typically need to meet three criteria: they must be unaffected by the experimental factors, show minimal variability in expression between tissues and physiological states of the organism, and it should show similar threshold cycle to the gene of interest19. Housekeeping genes were the first to be evaluated as reference genes in RT-

PCR. It is recommended to always use at least two reference genes, as using only one may lead to large errors, as was done in our experiments. Although 18S is the most widely used reference gene, there are a few fundamental problems with 18S. As 18S is a ribosomal RNA, it degrades slower than mRNA, and therefore it does not show a similar threshold cycle to the genes of interest as the inclusion criteria recommends19. Along these same lines ribosomal RNA is controlled by RNA polymerase I while mitochondrial RNA is controlled by RNA polymerase II, making the synthesis reactions independent of each other19. Adding a second reference gene, such as GADPH, would help to minimize any potential errors in RT-qPCR data due to housekeeping gene fluctuation.

Another limitation of this project is using mouse ear wound healing as our model of chondrogenic regeneration. Although our lab has the ability to perform full-thickness articular cartilage defects, the scope of this project did not allow for us to create a suitable hydrogel or other material for injection of the chosen drug for a sustained release into an articular cartilage defect. Systemic

injection or oral application of the drugs was intentionally avoided as one of the aims of this project was to identify a compound for local treatment. For these reasons topical application of the drug was chosen. Unlike a skin wound, the ear pinnae do contain a small layer of cartilage. However, this model does not closely mimic the articular cartilage because the pinnae is vascularized and the cartilage is elastic rather than articular. Ideally, once the drug has been further characterized, a system to retain it in the articular joint (i.e. a hydrogel) will be developed and a study can be undertaken assessing the effect of the drug treatment on articular cartilage defects in a rodent model. Also, the injury may influence the drug mechanics as the pathways targeted by the drug(s) might also be differentially regulated with injury. Therefore, a non-injury control should be done to assess if the topical application of the drug on the ear has any effect on stimulating cartilage growth without injury.

Future Directions

With the continuation of this project, experiments should be done to determine the effect of drug

70 in a mouse/rat articular cartilage defect or OA model. This will require the development of a delivery method, as it is well known that small molecules are not retained in the joint space for very long (minutes to hours)136. The activity of time of the drug should be directly assessed, as well as the release kinetics from the delivery method and joint retention time. Our lab is currently collaborating with Dr. Molly Schoichet’s group at the University of Toronto in order to load 17-

AAG, an analog of 17-DMAG, into their proprietary Hyaluronan methyl cellulose (hamc) hydrogel system for intra-articular injection137.

While this is being developed, another simple mouse ear model should be assessed. Does the injury

(through and through ear punch) affect the response of the drug? Studies have shown that in the joint, exposure to blood inhibits proteoglycan turnover and induces chondrocyte apoptosis138,139.

Similarly, in wound regeneration the skin relies on a blood clot formation to begin the healing process140. Assessing the effect of the drug in a more non-vascularized environment, similar to articular cartilage damage, would be telling in the mechanism of action. This question can be answered simply by applying the topical drug to the ear pinnae of uninjured ears and histologically assessing if new cartilage growth is stimulated. This experiment is currently being conducted in the lab.

Along with animal model testing, the mechanism of action should be understood in more detail.

As mentioned previously, if the technology to transfect human primary MPCs is available it would be very useful to modify the cell strain in order to directly assess the effect that the drugs have on p21 transcription. Similarly, MPCs are not the only cell type in the joint. It is essential to assess the effects of these drugs on synovial fibroblasts, macrophages, chondrocytes, and T-cells to screen for any adverse or positive effects, as well as the differential environments of OA and normal joints. Due to the differential effect of chondrogenesis on p21 levels in normal and OA human

MPCs it would also be interesting to assess if the drug(s) have a different effect on human MPCs isolated from normal or OA patients. Finally, safety considerations must be further assessed.

Although the drugs tested have been previously developed for anti-cancer use, their involvement in changes in the cell cycle establish a basis for tumorigenic testing. One study suggested that knocking down p21incompletely allows for MSCs to have an increased chondrogenic potential, yet no changes in tumorigenic potential105. However, because these compounds are acting

indirectly on p21, rather than short hairpin RNA used in that study, it is important to assess the changes in tumorigenic potential.

Chapter Five: Conclusion

This study sought to examine the relationship between inhibiting p21 transcription and chondrogenic potential, with the hypothesis that inhibiting p21 will increase chondrogenic potential human synovial MPCs and enhance cartilage regeneration in vivo. An initial drug screening was done to elucidate putative upstream inhibitors of p21 transcription in a human cell line. The chosen compounds were then characterized on human synovial derived MPCs during in vitro culture. One compound, an HSP90 inhibitor, increased the proliferation rate of human synovial MPCs, accumulated cells in the G2/M phase of the cell cycle, and increased chondrogenic markers SOX9 and ACAN following 21 days of pellet culture. This compound was then tested on a mouse in vivo model of chondrogenic regeneration and was found to stimulate the growth of cartilage in non-regenerating mice. These results indicate that p21 inhibition through HSP90 regulation may be a potential pharmaceutical target for stimulating chondrogenic regeneration for the treatment of cartilage defects or cartilage degenerating diseases such as Osteoarthritis in the future.

Works Cited

1. Arthritis Alliance of Canada. Arthritis Alliance of Canada, The Impact of Arthritis in Canada: Today and Over the Next 30 Years. Available at: http://www.arthritisalliance.ca/images/PDF/eng/Initiatives/20111022_2200_impact_of_art hritis.pdf. (Accessed: 28th October 2014) 2. Center for Disease Control. CDC - Arthritis - Basics - Definition - Osteoarthritis. (2014). Available at: http://www.cdc.gov/arthritis/basics/osteoarthritis.htm. (Accessed: 15th December 2014) 3. Ahmed, N., Stanford, W. L. & Kandel, R. A. Mesenchymal stem and progenitor cells for cartilage repair. Skeletal Radiol. 36, 909–912 (2007). 4. Moisio, K. et al. Denuded subchondral bone and knee pain in persons with knee osteoarthritis. Arthritis Rheum. 60, 3703–3710 (2009). 5. Berenbaum, F. Osteoarthritis as an inflammatory disease (osteoarthritis is not osteoarthrosis!). Osteoarthr. Cartil. 21, 16–21 (2013). 6. Johnston, S. A. Osteoarthritis. Joint anatomy, physiology, and pathobiology. Vet. Clin. North Am. Small Anim. Pract. 27, 699–723 (1997). 7. Cooper, C. et al. RISK FACTORS FOR THE INCIDENCE AND PROGRESSION OF RADIOGRAPHIC KNEE OSTEOARTHRITIS. ARTHRITIS Rheum. 43, 995–1000 (2000). 8. Wilder, F. History of acute knee injury and osteoarthritis of the knee: a prospective epidemiological assessment The Clearwater Osteoarthritis Study. Osteoarthr. Cartil. 10, 611–616 (2002). 9. Lau, E. C. et al. Factors Associated with Osteoarthritis of the Hip and Knee in Hong Kong Chinese : Obesity , Joint Injury , and Occupational Activities. 152, 855–862 (2000). 10. Uth, K. & Trifonov, D. Stem cell application for osteoarthritis in the knee joint: A minireview. World J. Stem Cells 6, 629–36 (2014). 11. Sokolove, J. & Lepus, C. M. Role of inflammation in the pathogenesis of osteoarthritis: latest findings and interpretations. Ther. Adv. Musculoskelet. Dis. 5, (2013). 12. Krasnokutsky, S. et al. Quantitative magnetic resonance imaging evidence of synovial proliferation is associated with radiographic severity of knee osteoarthritis. Arthritis Rheum. 63, 2983–2991 (2011). 13. Roemer, F. W. et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30- month follow-up: the MOST study: A Longitudinal Multicenter Study of Knee Osteoarthritis. Ann Rheum Dis 70, 1804–1809 (2011). 14. Scanzello, C. R. & Goldring, S. R. The role of synovitis in osteoarthritis pathogenesis. Bone 51, 249–257 (2012). 15. Pelletier, J. P., DiBattista, J. A., Roughley, P., McCollum, R. & Martel-Pelletier, J. Cytokines and inflammation in cartilage degradation. Rheum. Dis. Clin. North Am. 19, 545–68 (1993). 16. Hoffman, B. E. et al. Development and characterization of a human articular cartilage- derived chondrocyte cell line that retains chondrocyte phenotype. J. Cell. Physiol. 222, 695–702 (2010). 17. Goldring, M. B. Chondrogenesis, chondrocyte differentiation, and articular cartilage metabolism in health and osteoarthritis. Ther. Adv. Musculoskelet. Dis. 4, 269–85 (2012). 18. Olivotto, E., Otero, M., Marcu, K. B. & Goldring, M. B. Pathophysiology of

osteoarthritis: canonical NF-κB/IKKβ-dependent and kinase-independent effects of IKKα in cartilage degradation and chondrocyte differentiation. RMD open 1, e000061 (2015). 19. Saklatvala, J. Inflammatory signaling in cartilage: MAPK and NF-kappaB pathways in chondrocytes and the use of inhibitors for research into pathogenesis and therapy of osteoarthritis. Curr. Drug Targets 8, 305–13 (2007). 20. Loeser, R. F., Goldring, S. R., Scanzello, C. R. & Goldring, M. B. Osteoarthritis: A disease of the joint as an organ. Arthritis Rheum. 64, 1697–1707 (2012). 21. Caplan, A. I. Mesenchymal stem cells. J. Orthop. Res. 9, 641–650 (1991). 22. Pittenger, M. F., Mosca, J. D. & McIntosh, K. R. in 3–11 (Springer Berlin Heidelberg, 2000). doi:10.1007/978-3-642-57276-0_1 23. Zeidán-Chuliá, F. & Noda, M. ‘Opening’ the mesenchymal stem cell tool box. Eur. J. Dent. 3, 240–9 (2009). 24. Sasaki, M. et al. Mesenchymal Stem Cells Are Recruited into Wounded Skin and Contribute to Wound Repair by Transdifferentiation into Multiple Skin Cell Type. J. Immunol. 180, 2581–2587 (2008). 25. Okumura, T. et al. Identification of a bone marrow-derived mesenchymal progenitor cell subset that can contribute to the gastric epithelium. Lab. Investig. 89, 1410–1422 (2009). 26. Carlson, S. et al. Cardiac mesenchymal stem cells contribute to scar formation after myocardial infarction. Cardiovasc. Res. 91, 99–107 (2011). 27. Sekiya, I. et al. Human Mesenchymal Stem Cells in Synovial Fluid Increase in the Knee with Degenerated Cartilage and Osteoarthritis. 943–949 (2012). doi:10.1002/jor.22029 28. Jones, E. A. et al. Synovial Fluid Mesenchymal Stem Cells in Health and Early Osteoarthritis Detection and Functional Evaluation at the Single-Cell Level. 58, 1731– 1740 (2008). 29. Kurth, T. B. et al. Functional mesenchymal stem cell niches in adult mouse knee joint synovium in vivo. Arthritis Rheum. 63, 1289–1300 (2011). 30. Krawetz, R. J. et al. Synovial fluid progenitors expressing CD90+ from normal but not osteoarthritic joints undergo chondrogenic differentiation without micro-mass culture. PLoS One 7, e43616 (2012). 31. Mcalindon, T. E. et al. OARSI guidelines for the non-surgical management of knee osteoarthritis. Osteoarthr. Cartil. 22, 363–388 (2014). 32. Hochberg, M. C. et al. Osteoarthritis: a story of close relationship between bone and cartilage. Medicographia 35, 139–254 (2013). 33. Yen, Y. et al. Treatment of Osteoarthritis of the Knee with Microfracture and Rehabilitation. Med. Sci. Sports Exerc. 40, 200–205 (2008). 34. Oussedik, S., Tsitskaris, K. & Parker, D. Treatment of articular cartilage lesions of the knee by microfracture or autologous chondrocyte implantation: a systematic review. Arthroscopy 31, 732–744 (2015). 35. Hunziker, E. B. Articular cartilage repair: Basic science and clinical progress. A review of the current status and prospects. Osteoarthritis and Cartilage 10, 432–463 (2002). 36. Jackson, D. W., Lalor, P. A., Aberman, H. M. & Simon, T. M. Spontaneous repair of full- thickness defects of articular cartilage in a goat model. A preliminary study. J. Bone Joint Surg. Am. 83-A, 53–64 (2001). 37. Harris, J. D., Siston, R. A., Pan, X. & Flanigan, D. C. Autologous Chondrocyte Implantation: A systematic reviw. J. Bone Jt. Surg. 92, 2220–2233 (2010). 38. Minas, T., Gomoll, A. H., Bs, C. P. & Rn, T. B. Autologous Chondrocyte Implantation for

Joint Preservation in Patients with Early Osteoarthritis. 147–157 (2010). doi:10.1007/s11999-009-0998-0 39. Tuan, R. S., Chen, A. F. & Klatt, B. A. Cartilage regeneration. J. Am. Acad. Orthop. Surg. 21, 303–311 (2013). 40. Filardo, G. & Madry, H. Mesenchymal stem cells for the treatment of cartilage lesions : from preclinical findings to clinical application in orthopaedics. Knee Surg. Sport. Traumatol. Athroscopy 21, 1717–1729 (2013). 41. Kim, Y. S. et al. Clinical Outcomes of Mesenchymal Stem Cell Injection With Arthroscopic Treatment in Older Patients With Osteochondral Lesions of the Talus. Am. J. Med. 41, 1090–1099 (2013). 42. Jo, C., Lee, Y. & SHIN, W. Intra-Articular Injection of Mesenchymal Stem Cells for the Treatment of Osteoarthritis of the Knee: A Proof-of-Concept Clinical Trial. Stem Cells 32, 1254–1266 (2014). 43. Koh, Y. G. et al. Mesenchymal stem cell injections improve symptoms of knee osteoarthritis. Arthrosc. - J. Arthrosc. Relat. Surg. 29, 748–755 (2013). 44. Martínez-Lorenzo, M. J. et al. Phenotype and chondrogenic differentiation of mesenchymal cells from adipose tissue of different species. J. Orthop. Res. 27, 1499–1507 (2009). 45. De Bari, C., Dell’Accio, F., Tylzanowski, P. & Luyten, F. P. Multipotent mesenchymal stem cells from adult human synovial membrane. Arthritis Rheum. 44, 1928–42 (2001). 46. O’Sullivan, J., D’Arcy, S., Barry, F. P., Murphy, J. M. & Coleman, C. M. Mesenchymal chondroprogenitor cell origin and therapeutic potential. Stem Cell Res. Ther. 2, 8 (2011). 47. Premnath, P., Tan, B. & Venkatakrishnan, K. Bioactive interlinked extracellular matrix- like silicon nano-network fabricated by femtosecond laser synthesis. Biores. Open Access 1, 231–8 (2012). 48. Nitta, S. K. & Numata, K. Biopolymer-based nanoparticles for drug/gene delivery and tissue engineering. Int. J. Mol. Sci. 14, 1629–54 (2013). 49. Mobasheri, A. The future of osteoarthritis therapeutics: Emerging biological therapy. Curr. Rheumatol. Rep. 15, (2013). 50. Tat, S. K., Pelletier, J. P., Velasco, C. R., Padrines, M. & Martel-Pelletier, J. New perspective in osteoarthritis: The OPG and RANKL system as a potential therapeutic target? Keio Journal of Medicine 58, 29–40 (2009). 51. Heinemeier, K. M. et al. Radiocarbon dating reveals minimal collagen turnover in both healthy and osteoarthritic human cartilage. 8, (2016). 52. D’Lima, D., Hermida, J., Hashimoto, S., Colwell, C. & Lotz, M. Caspase inhibitors reduce severity of cartilage lesions in experimental osteoarthritis. Arthritis Rheum. 54, 1814– 1821 (2006). 53. Nuttall, M. E. et al. Inhibition of caspase-3-like activity prevents apoptosis while retaining functionality of human chondrocytes in vitro. J. Orthop. Res. 18, 356–363 (2000). 54. Lotz, M. K. & Kraus, V. B. New developments in osteoarthritis. Posttraumatic osteoarthritis: pathogenesis and pharmacological treatment options. Arthritis Res. Ther. 12, 211 (2010). 55. Chevalier, X. et al. Safety study of intraarticular injection of interleukin 1 receptor antagonist in patients with painful knee osteoarthritis: a multicenter study. J. Rheumatol. 32, 1317–23 (2005). 56. Schulze-Tanzil, G. et al. Interleukin-10 and Articular Cartilage: Experimental

Therapeutical Approaches in Cartilage Disorders. Curr. Gene Ther. 9, 306–315 (2009). 57. Zhang, Y. et al. Drug-induced regeneration in adult mice. 7, (2015). 58. Johnson, K. et al. A Stem Cell-Based Approach to Cartilage Repair. Science (80-. ). 336, 717–722 (2012). 59. Clark, L. D., Clark, R. K. & Heber-Katz, E. A new murine model for mammalian wound repair and regeneration. Clin. Immunol. Immunopathol. 88, 35–45 (1998). 60. Rai, M. F., Schmidt, E. J., McAlinden, A., Cheverud, J. M. & Sandell, L. J. Molecular insight into the association between cartilage regeneration and ear wound healing in genetic mouse models: targeting new genes in regeneration. G3 (Bethesda). 3, 1881–91 (2013). 61. Bedelbaeva, K. et al. Lack of p21 expression links cell cycle control and appendage regeneration in mice. Proc. Natl. Acad. Sci. U. S. A. 107, 5845–50 (2010). 62. Arthur, L. & Heber-Katz, E. The role of p21 in regulating mammalian regeneration. Stem Cell Res. Ther. 2, 30 (2011). 63. Martı, J., Flores, J. M., Garcı, P. & Serrano, M. Tumor Susceptibility of p21 Waf1 / Cip1 - deficient Mice 1. 6234–6238 (2001). 64. Balomenos, D. et al. The cell cycle inhibitor p21 controls T-cell proliferation and sex- linked lupus development. Nat. Med. 6, 171–6 (2000). 65. Besson, A., Dowdy, S. F. & Roberts, J. M. CDK inhibitors: cell cycle regulators and beyond. Dev. Cell 14, 159–69 (2008). 66. Stepniak, E. et al. c-Jun/AP-1 controls liver regeneration by repressing p53/p21 and p38 MAPK activity. Genes Dev. 20, 2306–2314 (2006). 67. Willenbring H et al. A Dual Role of p21 in Liver Regeneration and Hepatocarcinogenesis. 68. Torbenson, M. et al. STAT-3 overexpression and p21 up-regulation accompany impaired regeneration of fatty livers. Am. J. Pathol. 161, 155–61 (2002). 69. Buitrago-Molina, L. E. et al. The degree of liver injury determines the role of p21 in liver regeneration and hepatocarcinogenesis in mice. Hepatology 58, 1143–52 (2013). 70. Michalopoulos, G. K. et al. Liver regeneration. Science 276, 60–6 (1997). 71. Wu, H. et al. Targeted in vivo expression of the cyclin-dependent kinase inhibitor p21 halts hepatocyte cell-cycle progression, postnatal liver development, and regeneration. Genes Dev. 10, 245–260 (1996). 72. Susaki, E. & Nakayama, K. I. Functional similarities and uniqueness of p27 and p57: Insight from a knock-in mouse model. Cell Cycle 8, 2497–2501 (2009). 73. Nakayama, K. et al. Mice Lacking p27Kip1 Display Increased Body Size, Multiple Organ Hyperplasia, Retinal Dysplasia, and Pituitary Tumors. Cell 85, 707–720 (1996). 74. Arthur, L. M. et al. Epimorphic regeneration in mice is p53-independent. Cell Cycle 9, 3667–3673 (2010). 75. Masson, A. O. et al. Increased levels of p21(CIP1/WAF1) correlate with decreased chondrogenic differentiation potential in synovial membrane progenitor cells. Mech. Ageing Dev. 149, 31–40 (2015). 76. Yew, T. L. et al. Knockdown of p21Cip1/Waf1 enhances proliferation, the expression of stemness markers, and osteogenic potential in human mesenchymal stem cells. Aging Cell 10, 349–361 (2011). 77. Aktories, K. et al. New Perspectives in Regeneration Responsible series editor : Hilary Kropowski. 367, 78. Dotto, G. P. p21 WAF1 a Cip1 : more than a break to the cell cycle ? 1471, (2000).

79. Abbas, T. & Dutta, A. P21 in Cancer: Intricate Networks and Multiple Activities. Nat. Rev. Cancer 9, 400–14 (2009). 80. Di Cunto, F. et al. Inhibitory function of p21Cip1/WAF1 in differentiation of primary mouse keratinocytes independent of cell cycle control. Science 280, 1069–72 (1998). 81. Eastham, J. A. et al. In Vivo Gene Therapy with p53 or p21 Adenovirus for Prostate Cancer’. 5151–5156 (1995). 82. Waldman, T., Kinzler, K. W. & Vogelstein, B. Advances in Brief p2 ! Is Necessary for the p53-mediated G , Arrest in Human Cancer Cells ’. 5187–5190 (1995). 83. Murray, B. W. et al. Small-molecule p21-activated kinase inhibitor PF-3758309 is a potent inhibitor of oncogenic signaling and tumor growth. Proc. Natl. Acad. Sci. U. S. A. 107, 9446–51 (2010). 84. Wettersten, H. I. et al. A novel p21 attenuator which is structurally related to sorafenib. Cancer Biol. Ther. 14, 278–85 (2013). 85. Park, S.-H., Wang, X., Liu, R., Lam, K. S. & Weiss, R. H. High throughput screening of a small molecule one-bead-one-compound combinatorial library to identify attenuators of p21 as chemotherapy sensitizers. Cancer Biol. Ther. 7, 2015–2022 (2008). 86. Micol Tillhon, O. C. I. D. L. A. S. and E. P. New Research Directions in DNA Repair. (InTech, 2013). doi:10.5772/46014 87. Sax, J. K., Dash, B. C., Hong, R., Dicker, D. T. & El-Deiry, W. S. The cyclin-dependent kinase inhibitor butyrolactone is a potent inhibitor of p21 (WAF1/CIP1 expression). Cell Cycle 1, 90–96 (2002). 88. Park, J. H., Saravanakumar, G., Kim, K. & Kwon, I. C. Targeted delivery of low molecular drugs using chitosan and its derivatives. Adv. Drug Deliv. Rev. 62, 28–41 (2010). 89. Inoue, H., Hwang, S. H., Wecksler, A. T., Hammock, B. D. & Weiss, R. H. Sorafenib attenuates p21 in kidney cancer cells and augments cell death in combination with DNA- damaging chemotherapy. Cancer Biol. Ther. 12, 827–36 (2011). 90. Gui, C., Ngo, L., Xu, W. S., Richon, V. M. & Marks, P. A. Histone deacetylase ( HDAC ) inhibitor activation of p21 WAF1 involves changes in promoter-associated proteins , including HDAC1. 1–6 (2003). 91. El-deiry, W. S. The role of p53 in chemosensitivity and radiosensitivity. 53, 7486–7495 (2003). 92. Westfall, M. D., Mays, D. J., Sniezek, J. C. & Pietenpol, J. A. The ⌬ Np63 Phosphoprotein Binds the p21 and 14-3-3 Promoters In Vivo and Has Transcriptional Repressor Activity That Is Reduced by Hay-Wells Syndrome-Derived Mutations. 23, 2264–2276 (2003). 93. Ando, K. et al. Polo-like Kinase 1 ( Plk1 ) Inhibits p53 Function by Physical Interaction and Phosphorylation *. 279, 25549–25561 (2004). 94. Mizuguchi, Y. et al. SPRR2A enhances p53 deacetylation through HDAC1 and down regulates p21 promoter activity. BMC Mol. Biol. 13, 20 (2012). 95. Bertram, K. L. & Krawetz, R. J. Osmolarity regulates chondrogenic differentiation potential of synovial fluid derived mesenchymal progenitor cells. Biochem. Biophys. Res. Commun. 422, 455–61 (2012). 96. O ’brien, P. J. High-Content Analysis in Toxicology: Screening Substances for Human Toxicity Potential, Elucidating Subcellular Mechanisms and In Vivo Use as Translational Safety Biomarkers. doi:10.1111/bcpt.12227

97. Dexheimer, V., Frank, S. & Richter, W. Proliferation as a requirement for in vitro chondrogenesis of human mesenchymal stem cells. Stem Cells Dev. 21, 2160–9 (2012). 98. Kurth, T. et al. Chondrogenic potential of human synovial mesenchymal stem cells in alginate. Osteoarthr. Cartil. 15, 1178–1189 (2007). 99. Fan, J., Varshney, R. R., Ren, L., Cai, D. & Wang, D.-A. Synovium-Derived Mesenchymal Stem Cells: A New Cell Source for Musculoskeletal Regeneration. Tissue Eng. Part B Rev. 15, 75–86 (2009). 100. Tanaka, E. M. Cell differentiation and cell fate during urodele tail and limb regeneration. Curr. Opin. Genet. Dev. 13, 497–501 (2003). 101. Stocum, D. L. The urodele limb regeneration blastema. Differentiation 27, 13–28 (1984). 102. Mescher, A. L. The cellular basis of limb regeneration in urodeles. Int. J. Dev. Biol. 40, 785–795 (2004). 103. Cosden, R. S. et al. Intrinsic repair of full-thickness articular cartilage defects in the axolotl salamander. Osteoarthr. Cartil. 19, 200–205 (2011). 104. Cheverud, J. M. et al. Fine-mapping quantitative trait loci affecting murine external ear tissue regeneration in the LG/J by SM/J advanced intercross line. Heredity (Edinb). 112, 508–18 (2014). 105. Diekman, B. O. et al. Knockdown of the cell cycle inhibitor p21 enhances cartilage formation by induced pluripotent stem cells. Tissue Eng. Part A 21, 1261–74 (2015). 106. Wu, P., Nielsen, T. E. & Clausen, M. H. FDA-approved small-molecule kinase inhibitors. Trends Pharmacol. Sci. 36, 422–439 (2015). 107. Barnes, P. J. New anti-inflammatory targets for chronic obstructive pulmonary disease. Nat. Rev. Drug Discov. 12, 543–559 (2013). 108. Clark, J. D., Flanagan, M. E. & Telliez, J.-B. Discovery and Development of Janus Kinase (JAK) Inhibitors for Inflammatory Diseases. J. Med. Chem. 57, 5023–5038 (2014). 109. Gartel, A. L. & Tyner, A. L. The role of the cyclin-dependent kinase inhibitor p21 in apoptosis. Mol. Cancer Ther. 1, 639–49 (2002). 110. Search ChIP-qPCR Assay. Available at: http://www.sabiosciences.com/chipqpcrsearch.php?species_id=0&factor=Over+200+TF& gene=CDKN1a&nfactor=n&ninfo=n&ngene=n&B2=Search. 111. Ogino, S. et al. p21 expression in colon cancer and modifying effects of patient age and body mass index on prognosis. Cancer Epidemiol. Biomarkers Prev. 18, 2513–21 (2009). 112. American Type Culture Collection. ATCC ® CELL LINES BY GENE MUTATION. (2014). 113. Sal, E. & Baguna, J. Regeneration and pattern formation in planarians I. The pattern of mitosis in anterior and posterior regeneration in Dugesia (G) tigrina, and a new proposal for blastema formation. J. Embryol. exp. Morph 83, 63–80 (1984). 114. Celton-Morizur, S. & Desdouets, C. Polyploidization of liver cells. Adv Exp Med Biol 676, 123–135 (2010). 115. Lefebvre, V., Huang, W., Harley, V. R., Goodfellow, P. N. & de Crombrugghe, B. SOX9 is a potent activator of the chondrocyte-specific enhancer of the pro alpha1(II) collagen gene. Mol. Cell. Biol. 17, 2336–46 (1997). 116. Dudhia, J. Aggrecan, aging and assembly in articular cartilage. Cellular and Molecular Life Sciences 62, 2241–2256 (2005). 117. Sivan, S. S., Wachtel, E. & Roughley, P. Structure, function, aging and turnover of aggrecan in the intervertebral disc. Biochim. Biophys. Acta - Gen. Subj. 1840, 3181–3189

(2014). 118. Zhang, L. et al. Chondrogenic differentiation of human mesenchymal stem cells: a comparison between micromass and pellet culture systems. Biotechnol. Lett. 32, 1339– 1346 (2010). 119. Bosnakovski, D. et al. Chondrogenic differentiation of bovine bone marrow mesenchymal stem cells in pellet cultural system. Exp. Hematol. 32, 502–9 (2004). 120. Datto, M. B. et al. Transforming growth factor beta induces the cyclin-dependent kinase inhibitor p21 through a p53-independent mechanism. Proc. Natl. Acad. Sci. U. S. A. 92, 5545–9 (1995). 121. Ghosh-Choudhury, N. et al. Bone morphogenetic protein-2 induces cyclin kinase inhibitor p21 and hypophosphorylation of retinoblastoma protein in estradiol-treated MCF-7 human breast cancer cells. Biochim. Biophys. Acta - Mol. Cell Res. 1497, 186–196 (2000). 122. M., J. et al. in Regenerative Medicine and Tissue Engineering (InTech, 2013). doi:10.5772/56298 123. Siebelt, M. et al. Hsp90 Inhibition Protects Against Biomechanically Induced Osteoarthritis in Rats. Arthritis Rheum. 65, 2102–2112 (2013). 124. Takahashi, K. et al. Analysis of heat shock proteins and cytokines expressed during early stages of osteoarthritis in a mouse model. Osteoarthritis Cartilage 5, 321–9 (1997). 125. Pacey, S. et al. A phase I study of the heat shock protein 90 inhibitor alvespimycin (17- DMAG) given intravenously to patients with advanced solid tumors. Clin. Cancer Res. 17, 1561–70 (2011). 126. Biomed, L. Equivalent Surface Area Dosage Conversion Factors. 127. Kobayashi, S. et al. Presence of Cartilage Stem/Progenitor Cells in Adult Mice Auricular Perichondrium. PLoS One 6, e26393 (2011). 128. Schmitz, N., Laverty, S., Kraus, V. B. & Aigner, T. Basic methods in histopathology of joint tissues. Osteoarthr. Cartil. 18, S113–S116 (2010). 129. Krzeski, P. et al. Development of musculoskeletal toxicity without clear benefit after administration of PG-116800, a matrix metalloproteinase inhibitor, to patients with knee osteoarthritis: a randomized, 12-month, double-blind, placebo-controlled study. Arthritis Res. Ther. 9, R109 (2007). 130. Hellio le Graverand, M.-P. et al. A 2-year randomised, double-blind, placebo-controlled, multicentre study of oral selective iNOS inhibitor, cindunistat (SD-6010), in patients with symptomatic osteoarthritis of the knee. Ann. Rheum. Dis. 72, 187–95 (2013). 131. Hunter, D. J. et al. Phase 1 safety and tolerability study of BMP-7 in symptomatic knee osteoarthritis. BMC Musculoskelet. Disord. 11, 232 (2010). 132. Lohmander, L. S. et al. Intraarticular sprifermin (recombinant human fibroblast growth factor 18) in knee osteoarthritis: a randomized, double-blind, placebo-controlled trial. Arthritis Rheumatol. (Hoboken, N.J.) 66, 1820–31 (2014). 133. Pardali, K., Kowanetz, M., Heldin, C.-H. & Moustakas, A. Smad pathway-specific transcriptional regulation of the cell cycle inhibitor p21(WAF1/Cip1). J. Cell. Physiol. 204, 260–72 (2005). 134. Shi, D. et al. Photo-Cross-Linked Scaffold with Kartogenin-Encapsulated Nanoparticles for Cartilage Regeneration. ACS Nano 10, 1292–1299 (2016). 135. Lutterbach, B. et al. A Mechanism of Repression by Acute Myeloid Leukemia-1, the Target of Multiple Chromosomal Translocations in Acute Leukemia. J. Biol. Chem. 275, 651–656 (2000).

136. Evans, C. H., Kraus, V. B. & Setton, L. A. Progress in intra-articular therapy. Nat. Rev. Rheumatol. 10, 11–22 (2014). 137. Führmann, T. et al. Injectable hydrogel promotes early survival of induced pluripotent stem cell-derived oligodendrocytes and attenuates longterm teratoma formation in a spinal cord injury model. (2016). doi:10.1016/j.biomaterials.2015.12.032 138. Hooiveld, M. et al. Blood-induced joint damage: longterm effects in vitro and in vivo. J. Rheumatol. 30, 339–44 (2003). 139. Hooiveld, M. et al. Short-Term Exposure of Cartilage to Blood Results in Chondrocyte Apoptosis. Am. J. Pathol. 162, 943–951 (2003). 140. Reinke, J. M. & Sorg, H. Wound Repair and Regeneration. Eur. Surg. Res. 49, 35–43 (2012).

Appendix A

Table 5: Small molecule library list of drugs tested and their known activity. Yellow highlighted compounds are the five chosen for this study.

Number Inhibitor Activity 1 CP-690550 JAK3 inhibitor 2 Docetaxel Tubulin stabilizer 3 TGX-221 PI3K inhibitor 4 Lapatinib HER1/2, EGFR inhibitor 5 PD 0325901 MEK inhibitor 6 PIK-75, Hydrochloride PI3K inhibitor 7 ABT-263 Bcl-2 inhibitor 8 Axitinib (AG-013736) VEGF-R inhibitor 9 AZD05030 (Saracatinib) Src inhibitor 10 Canertinib (CI-1033) ErbB-R inhibitor 11 GDC-0941 PI3K inhibitor 12 Bosutinib (SKI-606) Abl, Src inhibitor 13 Nilotinib Bcr-Abl inhibitor 14 FTY720, Hydrochloride Immunosuppressant 15 ABT-888 (Veliparib) PARP inhibitor 16 BIBW 2992 (Tovok) RTK inhibitor 17 GDC-0449 Hedgehog pathway inhibitor 18 Vandetanib (Zactima) RTK inhibitor 19 Vatalanib Dihydrochloride RTK inhibitor 20 AZD 2281 (Olaparib) PARP inhibitor 21 Bicalutamide (Casodex) Androgen receptor inhibitor 22 BI 2536 PLK inhibitor 23 ZM 447439 Aurora kinase inhibitor 24 Pp242 mTOR inhibitor 25 OSI-906 IGF-1R inhibitor 26 LBH-589 (Panobinostat) HDAC inhibitor 27 Laropiprant PGD2-R antagonist 28 MK-2206 Akt inhibitor 29 Ramatroban (Bay u3405) CRTH2-R antagonist 30 BMS-599626 RTK inhibitor 31 BMS-754807 IGF-1R inhibitor 32 Raltegravir HIV integrase inhibitor 33 RDEA119 MEK inhibitor

Number Inhibitor Activity 34 PF-2341066 c-MET inhibitor 35 AZD 6244 (ARRY-142886) MEK inhibitor 36 Odanacatib (MK-0822) Cathepsin inhibitor 37 MS-275 HDAC inhibitor 38 NVP-TAE684 NPM-ALK inhibitor 39 SN-38 Topoisomerase I inhibitor 40 VX702 p38 MAPK inhibitor 41 MGCD0103 HDAC inhibitor 42 Maraviroc (UK-427857) CCR5 antagonist 43 VX-680 Aurora kinase inhibitor 44 AN2728 PDE4 inhibitor 45 Bexarotene (Targretin) RXR activator 46 Capecitabine (Xeloda) NA synthesis inhibitor 47 CVT-6883 Adenosine receptor antagonist 48 Motesanib (AMG-706) RTK inhibitor 49 Imatinib TK inhibitor 50 Hypothemycin T cell activation inhibitor 51 Dimebolin Hydrochloride Antihistamine 52 FK-506 Immunosuppressant 53 Vorinostat (SAHA) HDAC inhibitor 54 Dasatinib Src inhibitor 55 Montelukast Sodium (Singulair) LTR antagonist 56 Rofecoxib (Vioxx) COX-2 inhibitor 57 Pemetrexed Disodium (Alimta) NA synthesis inhibitor 58 Gemcitabine, HCl (Gemzar) NA synthesis inhibitor 59 Doxorubicin (Adriamycin) NA synthesis inhibitor 60 Topotecan (Hycamtin) Topoisomerase I inhibitor 61 TM30089 CRTH2-R antagonist 62 Gefitinib (Iressa) TK inhibitor 63 Etoposide Topoisomerase II inhibitor 64 Bortezomib (Velcade) Proteasome inhibitor 65 ABT-737 Bcl-2 inhibitor 66 Sorafenib TK inhibitor 67 Rapamycin (Sirolimus) mTOR inhibitor 68 Erlotinib, Hydrochloride EGF-R inhibitor 69 Paclitaxel (Taxol) Tubulin stabilizer 70 17-DMAG HSP 90 inhibitor 71 Sunitinib TK inhibitor 72 Tandutinib RTK inhibitor 73 17-AAG HSP 90 inhibitor

Number Inhibitor Activity 74 BI-D1870 p90 RSK inhibitor 75 A-769662 AMPK activator 76 AC220 (Quizartinib) FLT3 inhibitor 77 AG014699 (PF-01367338) PARP inhibitor 78 ARRY-162 MEK inhibitor 79 ARQ 197 (Tivantinib) c-MET inhibitor 80 Atorvastatin Calcium (Lipitor) HMG-CoA reductase inhibitor 81 AV-951 (Tivozanib) VEGF-R inhibitor 82 AZD1152 Aurora kinase inhibitor 83 AZD1480 JAK inhibitor 84 AZD4547 FGFR inhibitor 85 AZD8055 mTOR inhibitor 86 Belinostat (PXD101) HDAC inhibitor 87 BI 6727 (Volasertib) PLK inhibitor 88 BMS-777607 c-MET inhibitor 89 BSI-201 (Iniparib) PARP inhibitor 90 CAL-101 PI3K inhibitor 91 Carfilzomib (PR-171) Proteasome inhibitor 92 CGS 21680 Adenosine receptor agonist 93 CUDC-101 HDAC/EGFR/HER2 inhibitor 94 CYT-387 JAK inhibitor 95 Dutasteride (Avodart) 5α-reductase inhibitor 96 EMD1214063 c-MET inhibitor 97 Eprosartan Mesylate (Teveten) Angiotensin receptor antagonist 98 GSK1120212 MEK inhibitor 99 GSK461364 PLK inhibitor 100 GSK690693 Akt inhibitor 101 INCB018424 (Ruxolitinib) JAK inhibitor 102 INK128 mTOR inhibitor 103 Lenalidomide (CC-5013) Immunomodulator 104 LY294002 PI3K inhibitor 105 MK-4827 PARP inhibitor 106 NVP-LDE225 SMO inhibitor 107 OSI-027 mTOR inhibitor 108 Pazopanib (Votrient) TK inhibitor 109 PD-0332991 CDK inhibitor 110 PF-04217903 c-MET inhibitor 111 PF-04691502 PI3K/mTOR inhibitor 112 PLX4032 (RG7204) B-Raf inhibitor 113 PLX4720 B-Raf inhibitor

Number Inhibitor Activity 114 Ponatinib (AP24534) Bcr-Abl inhibitor 115 Regorafenib (BAY 73-4506) Multi-kinase inhibitor 116 SR1 AHR antagonist 117 TG100-115 PI3K inhibitor 118 TG101348 JAK inhibitor 119 Tubacin HDAC inhibitor 120 Tubastatin A HDAC inhibitor 121 Varespladib (LY315920) Phospholipase A2 inhibitor 122 VX-765 ICE/Caspase-1 inhibitor 123 VX-950 (Telaprevir) Protease inhibitor 124 WZ4002 EGF-R inhibitor 125 XL-147 PI3K inhibitor 126 XL-184 (Cabozantinib) TK inhibitor 127 YM155 Survivin suppressant 128 AS703026 (MSC1936369B) MEK inhibitor 129 NVP-AUY922 (VER-52296) HSP 90 inhibitor 130 AZD7762 Checkpoint kinase inhibitor 131 Foretinib c-MET/VEGFR2 inhibitor 132 Dabrafenib B-Raf inhibitor 133 GDC-0980 PI3K/mTOR inhibitor 134 NVP-BGJ398 FGFR inhibitor 135 NVP-BKM120 PI3K inhibitor 136 PCI-32765 (Ibrutinib) Bruton's TK inhibitor 137 Dacomitinib ErbB-R inhibitor 138 SCH900776 Checkpoint kinase inhibitor 139 VX-11e ERK2 inhibitor 140 Deforolimus mTOR inhibitor 141 SGI-1776 Pim-1 inhibitor 142 AZD1208 Pan-Pim inhibitor APO Apocynin Anti-arhtritis/anti-inflammatory SP SP600125 MAPL/JNK Inhibitor OLOM Olomoucine CDK2/CDC2/MAPK inhibitor IV Butyrolactone I CDK2/CDC2 inhibitor