The Roles of Catabolic Factors in the Development of Osteoarthritis

Abstract

Osteoarthritis (OA) is the most prevalent disease of articular joints characterized by joint space narrowing on X-ray, joint pain, and a loss of joint function through progressive cartilage degradation and intermittent synovial inflammation. Current in vitro models of OA are often monolayer cultured primary cells exposed to high concentrations of cytokines or chemokines, usually IL-1β or TNF-α. IL-1β could play a role in the early progression or even initiation of OA as evidenced by many of the in vitro studies. However, the inconsistent or outright lack of detectable IL-1β combined with high concentrations of the natural inhibitor IL-1Ra in the OA synovial fluid makes the idea of OA being IL-1β-driven questionable. Further, other stimulants, including IL-6 and matrix fragments, have been shown in vitro to cause many of the effects seen in OA at relevant concentrations found in the OA synovial fluid. More work with these stimulants and IL-1β-independent models needs to be done. Concurrently, research should be conducted with patients with OA as early as possible in the progression of their disease to be able to potentially identify, target, and treat the initiation of the disease.

Introduction

Osteoarthritis (OA) is the most common type of arthritis and the major cause of chronic musculoskeletal pain and mobility disability in elderly populations worldwide, affecting nearly 40% of people older than 65 years.¹ Joint pain is the symptom most likely to present first, but OA is generally characterized by this pain, joint space narrowing, and loss of joint function.² In a normal joint, there is a balance between the synthetic and degradative activity. In an OA joint, there is progressive cartilage degradation induced by loss of proteoglycans and eventual breakdown of the collagen network of the extracellular matrix (ECM).³ Whereas pain therapies are available, unfortunately there is not much for patients in the way of structural benefit. For most patients, the only treatment option is inevitably joint replacement.^4,5

Although the facets of OA are not well understood, it is clear that OA is chronic, slow progressing destruction of articular joints, including the cartilage, synovium, and underlying bone. The catabolic environment of OA is marked by increases in degradative enzymes and proinflammatory cytokines. Matrix metalloproteinases (MMPs) play a large role in the destruction of the osteoarthritic joint. MMP-1 and MMP-3, an aggrecanase, are present in high levels in the OA synovial fluid.^6,7 In addition, although detected at much lower levels in the OA synovial fluid (287-fold lower than MMP-1 over 14,000-fold lower than MMP-3), the collagenase MMP-13 has been implicated due to the high gene expression in OA cartilage.⁶ Although not known as an inflammatory disease, the OA synovial fluid is full of proinflammatory cytokines and chemokines, including interleukin 8 (IL-8),^8,9 monocyte chemoattractant protein 1 (MCP-1),^10,11 macrophage inflammatory protein 1α (MIP-1α),^12,13 regulated and normal T cell expressed and secreted (RANTES),^14,15 and vascular endothelial growth factor,^15,16 among others. These proteins have been found at detectable levels, but the concentration of IL-1β and tumor necrosis factor α (TNF-α) in the OA synovial fluid is much more varied. Whereas most patients have extremely low or undetectable concentrations of IL-1β or TNF-α, some patients have considerably higher concentrations of these proteins in their synovial fluid.^17–24 Due to this wide variation in the OA synovial fluid, these proinflammatory cytokines remain controversial.

With no current treatment for structural changes in OA, research into the mechanisms for the changes involved in OA continues. The difficulty in researching OA begins with the idea that OA is not a single disease, but rather a complex combination of metabolic processes affecting the joint and surrounding tissues (Fig. 1).^25,26 This complexity makes it difficult to develop accurate models. There are animal models that are used for studying OA, including surgical instability models, such as anterior cruciate ligament transection or spontaneous models, such as the STR/ort mouse model.^27,28 The cost, time, and animals associated with performing in vivo studies mean a new OA target has to be sufficiently studied in vitro first. In vitro, OA is most often studied through monolayer cultured primary cells exposed to high concentrations of cytokines or chemokines, usually IL-1β.²⁹ However, IL-1β may not be the best stimulant when modeling clinical OA. Further, although many studies have used 2D models, cartilage is a 3D structure with chondrocytes existing in an ECM of proteoglycan and collagen. Cell–cell as well as cell–matrix interactions are important for the study of cartilage.³⁰ As shown in recent studies,^31–33 much of this can be mimicked by better tissue engineered models, including clinically relevant stimulants at physiological concentrations (Table 1).

FIG. 1.

Disease perpetuation, catabolic cycle. Disease initiation of clinical osteoarthritis (OA) is due to several possible causes, including trauma, poor alignment, or inflammation. After initiation, joints undergo a catabolic cycle as shown. This catabolic cycle includes synovitis, pain, increases in proinflammatory cytokines and degradative enzymes, and extracellular matrix degeneration. Interrupting this cycle is likely essential to slowing the progression of OA.

Table 1.

Effects of Catabolic Stimulants in Models and Clinical Osteoarthritis

Effects	IL-1β	IL-6	Matrix fragments
Loss of proteoglycans	30–31	66–67	78–79, 84
Procatabolic	34–37	62	78, 85
Antianabolic	38–41	69–71	78, 84
Clinical OA	47–53	64–66	73–77, 80–83, 86–88

OA, osteoarthritis.

Role of IL-1β in Catabolism In Vitro

Although OA is not considered to be an inflammatory disease, studies have found elevated levels of cytokines and growth factors in the synovial fluid of individuals with OA.^34–36 Given that cartilage breakdown is one of the most recognizable changes in OA, much research has been done looking at this destruction in vitro. The only cell type in hyaline cartilage is the chondrocyte. It follows then that, to study the effects of OA on cartilage, chondrocytes have been the focus of research efforts. Although it is clear the cartilage biology is changed in OA, the reason for the change is not fully understood. To study these effects in vitro, chondrocytes have been stimulated with IL-1β or TNF-α. Both IL-1β and TNF-α have been shown to downregulate the synthesis of major ECM components as well as inhibiting the anabolic activity of chondrocytes.^37,38

Two hallmarks of OA are the loss of glycosaminoglycan (GAG) and collagen II content in the cartilage.³⁹ Human OA chondrocytes cultured in alginate suspension were treated with 100 pg/mL of IL-1β. In these cultures, aggrecan gene expression was downregulated two- to threefold compared with media alone controls.⁴⁰ Similar results were seen with human and porcine cartilage explants.⁴¹ Monolayer cultures of rabbit articular chondrocytes treated with 10 ng/mL of IL-1β had greater than a 40% reduction in COL2A1 mRNA.⁴² In human monolayer cultures of chondrocytes, IL-1β reduced COL2A1 expression.⁴³ It should be noted that all culture systems used very high concentrations of IL-1β as the stimulant, several orders of magnitude higher than what is found in the OA synovial fluid. This is described in more detail below.

In addition to reducing the anabolic activity, IL-1β has been shown to increase catabolism. IL-1β stimulated chondrocytes release several MMPs, including MMP-1, MMP-3, and MMP-13, and aggrecanases, including aggrecanase 1 and 2 (ADAMTS-4 and ADAMTS-5). These degradative enzymes with aggrecanase and collagenase activity have been shown important in OA.^44–47 Further, osteoarthritic cartilage has been shown to express these MMPS⁴⁸ and aggrecanases.⁴⁹ Stimulation of chondrocytes with IL-1β has been shown to induce the production of many of the proinflammatory cytokines commonly found in the OA synovial fluid, including IL-6,⁵⁰ IL-8,⁵¹ MCP-1,⁵² and RANTES.⁵³ It can be shown that chondrocytes stimulated with IL-1β appear to produce an environment similar to the OA synovial fluid. Further, due to IL-1β, this increase in cartilage breakdown factors in combination with the reduction in repair capabilities mimics many aspects of clinical OA. This dual role makes IL-1β appear to be a logical target for OA therapies.

Natural Inhibition of IL-1β: IL-1Ra

IL-1β appears to be a mediator of joint damage in OA, at least based on in vitro data. There is plenty of evidence implicating cytokines in the progression of rheumatoid arthritis (RA), a disease with a major inflammatory component.^54–57 The lower number of synovial cells secreting IL-1β,⁵⁸ combined with the inconsistent or outright absence of detectable IL-1β in the OA synovial fluid, serum or synovial tissue^17–24 makes the concept of IL-1β driving the degradation in OA questionable.

IL-1 receptor antagonist (IL-1Ra) is a naturally occurring, specific antagonist of either IL-1α or IL-1β. IL-1Ra binds to the IL-1 receptors with similar affinities as IL-1α or IL-1β, but does not transduce a signal.⁵⁹ It has been shown that in vitro, 10–100 times the amount of IL-1Ra is needed to inhibit IL-1β. In vivo, there needs to be 100–2000 times more IL-1Ra than IL-1β.^60,61 In a study of candidates awaiting total knee arthroplasty, the synovial fluid was analyzed for levels of IL-1β, TNFα, and IL-1Ra among others. In this study, IL-1β and TNFα were undetectable even though the detection limit was <0.1 pg/mL. In male patients, IL-1Ra was detected at 485.90 pg/mL. In female patients, IL-1Ra was found at 332.68 pg/mL.⁶² With at least 4000 times the amount of IL-1Ra to IL-1β, it is unlikely that additional anti-IL-1β treatment would be effective at this stage.

These patients clearly represent end-stage OA. However, given that the disease may be relatively painless and asymptomatic in the initial stages, by the time the patient seeks treatment, significant damage may have occurred in the cartilage. Thus, at the point a patient appears at the doctor, it may be too late to use an anti-IL-1 treatment. Anti-IL-1 treatment has been attempted in the clinic. Anakinra is an IL-1 receptor antagonist used to treat RA. A total of 160 patients with OA were randomized and treated with anakinra. At week 4, there was no improvement in knee pain, function, stiffness, or cartilage turnover in patients treated with anakinra compared with the placebo.⁶³ The major difference between RA and OA is the levels of inflammation. The level of IL-1 β is so high that it can be measured in the peripheral blood of RA patients. This is not the case with OA patients.⁶⁴ In RA patients, the synovial membrane is highly thickened and inflamed. This is not seen as frequently in OA patients.⁶⁵ The low level of inflammation, the low concentrations of IL-1β, the high concentration of the natural inhibitor to IL-1β, and the lack of benefit with anti-IL-1β therapy in the clinic suggests that the role of IL-1β in OA is limited by the time patients present in the clinic. This suggests that the use of IL-1β blocking therapy is probably limited.

IL-6: Induction of Catabolism by Other Cytokines

Whereas IL-1β is the classical proinflammatory cytokine, several others have been shown to play a role in catabolism. TNF-α, has long been hypothesized to be important in the progression of OA, although the role of TNF-α in cartilage degradation is generally regarded as less clear. TNF-α activates the production and release of MMPs, which leads to matrix breakdown and could explain cartilage degeneration.⁶⁶ IL-17 and IL-18 are two additional cytokines that are potent inducers of catabolic responses in chondrocytes. IL-17 and IL-18 stimulate the production of IL-6, iNOS, COX-2, and MMPs. Adenoviral overexpression of IL-17 induces cartilage proteoglycan loss in the joint. In animal models, IL-18 deficiency or blockade reduced cartilage destruction and inflammation.⁶⁷ Whereas much work has been done to explore the role of these catabolic cytokines, the role of IL-6 is less obvious.

IL-6 is a pleiotropic cytokine with a wide range of biological activities in immune regulation, hematopoiesis, inflammation, and oncogenesis. IL-6 is produced by many different cell types, including T cells, B cells, monocytes, fibroblasts, keratinocytes, and endothelial cells.⁶⁸ In OA patients, the level of IL-6 in the synovial fluid is quite high. Due to this high level of IL-6, its regulatory action might be crucial to the progression of OA. The exact role of IL-6 in OA is controversial. In studies, IL-6 has been shown to be both a proinflammatory cytokine⁶⁹ and an anti-inflammatory mediator.⁷⁰ Due to these confounding results, IL-6 remains an active research area in the progression of OA.

There has been some investigation into the correlation between IL-6 and knee pain severity, development, and progression leading to inconsistent results. Stannus et al. found that an increase in IL-6 concentrations in the serum led to a significant increase in joint pain when standing. Pain was measured by a questionnaire using the Western Ontario McMaster osteoarthritis index (WOMAC).⁷¹ Conversely, Pelletier et al. found no such correlation between IL-6 and pain.⁷² It is important to note the difference in timeline and statistical analysis. In the Stannus' study, patients were followed out to 5 years as compared with 2 years in the Pelletier's study. Further, Stannus et al. adjusted results for potential confounders, including age, sex, and body mass index.

Data about the role of IL-6 in structural degradation are equally confounding. IL-6 levels were investigated in the synovial fluid from patients with increasing severity of joint destruction as measured by the Kellgren-Lawrence (KL) grading scale. In this study, the authors found a significant inverse relationship between the KL grade and concentration of IL-6.⁷³ The conclusion from this correlation could indicate that IL-6 represents a proanabolic factor able to mitigate cartilage damage early in the disease, but is decreased as OA progresses. Based on in vitro data, IL-6 may be more integral in the pathogenesis of OA. Based on that assumption, the data could suggest that active cartilage destruction occurs at the highest rate in earlier grades of OA and tapers off as the disease reaches end stage. Although Pelletier et al. did not find a correlation of IL-6 to pain, this study did support IL-6 as a possible predictor of knee OA progression. In addition, similarly to previous work, cartilage volume lost over time was associated with a decrease in IL-6 concentrations. Additionally, increased levels of IL-6 have been considered as an indicator for sarcopenia. Sarcopenia is the degenerative loss of muscle mass and physical function associated with age. Sarcopenia is an aggravating factor that increases the progression of knee OA. Similarly, IL-6 was found to correlate with the loss of strength in knee flexor muscles.⁷⁴

Many of the functions of IL-6 are mediated through its soluble receptor (sIL-6R). This soluble receptor forms a complex with IL-6, which can then transduce signals through the ubiquitously expressed glycoprotein 130 (gp130). Thus, sIL-6R-mediated signaling allow for activation of cell types that would not normally respond to IL-6 itself. In a study comparing RA and OA, levels of IL-6 and sIL-6R were examined. Although the authors found significantly higher levels of IL-6 and sIL-6R in RA, the concentration of these in the OA synovial fluid was still quite high, 6.39 ng/mL and 14.78 ng/mL for IL-6 and sIL-6R, respectively. The authors also found a significant correlation between levels of IL-6 and sIL-6R.⁷⁵

IL-6 and sIL-6R have been shown to modulate the balance of anabolism and catabolism in vitro. In a study of rabbit articular chondrocytes, IL-6, sIL-6R, or both inhibited type II collagen production.⁷⁶ In bovine articular chondrocytes, IL-6 and sIL-6R induced activation of Janus Kinase 1 (JAK1), JAK2, and signal transducer and activator of transcription 1 (STAT1)/STAT3. These conditions significantly downregulated the expression of type II collagen, aggrecan core, and link proteins.⁷⁷ This diminished anabolic activity is coupled with an increase in MMP and a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS) expression. Primary bovine chondrocytes were stimulated with IL-6 and sIL-6R for 24 h. Expression of MMP1, MMP3, and MMP13 as well as ADAMTS4 and ADAMTS5/11 were markedly increased.⁷⁸ These decreases in anabolic activity combined with the increased degradative activity are hallmarks of OA, suggesting that other cytokines could provide the stimulus needed for the progression of OA.

Matrix Fragments Induce Catabolism

Cartilage comprised chondrocytes, the resident cells, and an abundant ECM composed primarily of type II collagen, aggrecan, and hyaluronan.⁷⁹ Degradation of this matrix is a hallmark of OA. Whereas there is extensive literature discussing the role of ECM fragments as markers of cartilage turnover, there is a growing body of work exploring the possibility that these fragments may actually be part of the production/degradation pathway.

There is an increase of fibronectin (FN) within the cartilage matrix and synovial fluid of patients with OA or RA.⁸⁰ This increase is three- to fourfold, with the average concentration of FN in normal synovial fluid increasing from 171 mg/ml to 721 and 568, in RA and OA, respectively.⁸¹ It follows then that this increase in FN allows for an increase in fragmentation of said FN. Indeed, FN fragments (FN-f) have been found in cartilage and synovial fluid from RA and OA patients ranging from 24 to 200 kDa.^82,83 In the OA synovial fluid, the concentration of FN-f has been estimated to be in μM concentrations.⁸⁴

Physiological concentrations of FN-f, 0.1 to 1 μM, rapidly decreased the cartilage proteoglycan content and suppressed proteoglycan synthesis in human cartilage explant cultures. These concentrations are consistent with or below measured concentrations in the OA synovial fluid. Over a few days, these explants released half of the total proteoglycan as compared with no FN-f controls. New proteoglycan synthesis, as measured by sulfate incorporation was also drastically suppressed to nearly 50% the level of the control explants. IL-6 and MMP-3 production were also significantly increased in explants treated with FN-f, 16.9- and 12.5-fold, respectively. These factors likely contribute to the loss of proteoglycans.⁸⁵ FN-f is capable of upregulating MMP-3, but MMP-3 is capable of a catabolic cycle. In addition to causing proteoglycan degradation, MMP-3 has been shown to cleave FN.⁸⁶ Due to this, it could by hypothesized that FN-f could create a self-sustaining catabolic cycle.

Similar effects were seen in vivo. New Zealand white rabbits were injected with 0.3 mL of 3 μM of a 1:1 mixture of 29 and 50 kDa mixture of FN-f. Upon gross examination, articular cartilage from animals injected with FN-f was dull in pale. In contrast, control animals that received no such injection had glistening white cartilage. As a percent of proteoglycan content in the left knee (injected knee) compared with the right knee (naïve knee), rabbits injected with FN-f had a markedly reduced proteoglycan content 59%±8%. For the purposes of comparison, control animals had proteoglycan ratios at 109%±16%, not unexpectedly. This difference was statistically significant (p<0.0001). In terms of histology, the most striking difference with FN-f treatment was the loss of Safranin-O and Toluidine Blue staining indicative of a loss of proteoglycans.

FN fragments are not the only matrix fragments capable of inducing changes in the cartilage. Due to the large proportion of type II collagen found in articular cartilage, there is an abundance of potential fragmented collagen. Indeed up to 20% of collagen in human OA cartilage can be at least partially degraded.⁸⁷ Cleavage of type II collagen may result from cleavage by collagenases commonly found in OA, including MMP-1 and MMP-13. Further, MMP-3, found at high concentrations in the OA synovial fluid, has been shown to cleave type II collagen within the N-telopeptide and C-telopeptide domain.⁷⁹ Whereas many different type II collagen fragments are possible in cartilage degeneration,^88–90 both collagen N-telopeptides and C-telopeptides have been shown to upregulate MMP expression in chondrocytes, enhance the degradation of matrix and change matrix synthesis.

Human cartilage explants treated with fragmented type II collagen at up to 1 mg/mL released GAG and collagen into the media resulting in less GAG and collagen content remaining in the explants. In these studies, fragmented collagen also reduced cell attachment and de novo synthesis of type II collagen in bovine chondrocytes.⁹¹ The suggestion is that the intact telopeptide signals the presence of intact collagen. Thus, the cell does not recognize the need for de novo collagen synthesis. In bovine cartilage explants, type II collagen fragments and N-telopeptides have been shown to upregulate MMP2, MMP3, MMP9, and MMP13 mRNA.⁹² These results create a twofold response where collagen telopeptides inhibit de novo collagen synthesis, while simultaneously increasing collagenases. This likely further exacerbates the progression of OA. C-telopeptides are elevated in the urine of OA patients⁹³ and have been suggested as a biomarker for OA. To that end, several markers have been studied as biomarkers of type II collagen metabolism, including degradation markers (C2C, C1, 2C) and synthesis marker (CP II).^94,95 Further, urinary C-telopeptides correlate with high cartilage turnover in OA⁹⁶ and the progression of OA as seen in magnetic resonance imaging (MRI).⁹⁷

FN and collagen fragments likely arise as part of the disease of OA. These fragments are likely generated through cartilage matrix degeneration and turnover. As described above, many types of each FN or collagen fragments are possible and found in OA. The different fragments have been shown to have different potency in upregulating catabolism in vitro. The need for further study of these fragments is clear, given these fragments are found in OA at concentrations shown to be catabolic in vitro. In this way, the disease could be thought of as self-sustaining. Once the original damage has been done, by mechanical force, cytokine stimulation, or other stimuli, the turnover and matrix degradation can continue the degradation. Thus, when a patient arrives at the doctor, the original stimulus is gone and the disease is left continuing the cycle of degradation.

Additional Stimulants, Lipopolysaccharides, and Retinoic Acid

Expression of Toll-like receptors (TLRs) is upregulated in OA lesion areas. TLR-4, the receptor for lipopolysaccharides (LPS) is one of the TLRs upregulated in OA. LPS is the major component of the outer membrane of Gram-negative bacteria. To this end, several in vitro models of OA have included stimulation with LPS. Chondrocytes stimulated with LPS increase production of MMPs, including MMP-1, MMP-3, and MMP-13. LPS also increased expression of TLR2 and TLR4, as an example of ligand stimulating increases in receptor.⁹⁸ In OA synovial membranes treated with LPS, an increase in proinflammatory cytokines IL-1β and TNF-α as well as MMP-3 was shown.⁹⁹ LPS has also been shown to significantly reduce cartilage synthesis as measured by 35S uptake in cartilage explants treated with LPS.¹⁰⁰ Although most cases of OA are not likely to come from a bacterial infection, LPS models would seem to be an accurate representation of septic arthritis. Septic arthritis has a substantially lower incidence rate than OA, 2.6 per 100,000 or several orders of magnitude lower incidence than OA.¹⁰¹ It remains to be seen if LPS would be the best stimulant to model the broader disease.

The addition of retinoic acid to cartilage explants has been shown to enhance destruction of matrix components.^102,103 MMP-13 may at least be partially responsible for this degradation. Porcine chondrocytes cultured in the presence of retinoic acid had increased MMP-13 expression compared with controls. This increase in MMP-13 expression was also shown for human OA chondrocytes.¹⁰⁴ Further, porcine or bovine explants cultured in the presence of retinoic acid lost over 80% of the GAG over just 4 days.¹⁰⁵ Due to the hypothesis that antioxidants provide protection in OA patients from further joint damage; retinoic acid has been suggested as a treatment as well. Retinoic acid was able to inhibit MMP1 and MMP13 mRNA expression, protein production, and enzyme activity induced by either IL-1β or TNF-α in cultures of OA chondrocytes.¹⁰⁶ In a separate study, retinoic acid suppressed IL-1β-induced release of chemokines, including RANTES, MIP-1α among others.¹⁰⁷ Retinoic acid is controversial and more work needs to be done to explore the pro- and anticatabolic effects in model systems.

Conclusions

Given the in vitro evidence, it is clear that IL-1β could play a role in the early progression or even initiation of OA. However, patients do not present in the clinic until long after the initiation phase of OA. Further, the lack or at least inconsistent detection of IL-1β in OA patients, high concentrations of the natural inhibitor IL-1Ra, and existence of other stimulants at relevant concentrations makes the idea of OA being IL-1β driven questionable. Thus, although IL-1β is capable of creating cytokine profiles similar to OA, many other targets are similarly capable. Therefore, therapies need to be developed to the actual root cause not therapies to an in vitro catabolic stimulant. To screen new targets or understand the pathology of later stage OA, more work needs to be done with tissue engineering models that include stimulants at physiological concentrations as well as IL-1β-independent models. Concurrently, research should be conducted with patients with OA much earlier in the progression of the disease, thus being able to potentially identify, target, and treat disease initiation.

Acknowledgment

We thank the NIH (P41 EB002520) for support of this work.

Disclosure Statement

No competing financial interests exist.