Synovium and the Innate Inflammatory Network in Osteoarthritis Progression

Abstract

This review focuses on the recent advancements in the understanding of innate immunity in the pathogenesis of osteoarthritis, particularly with attention to the roles of damage-associated molecular patterns (DAMPs), pattern recognition receptors (PPRs), and complement in synovitis development and cartilage degradation. Endogenous molecular products derived from cellular stress and extracellular matrix disruption can function as DAMPs to induce inflammatory responses and pro-catabolic events in vitro and promote synovitis and cartilage degradation in vivo via PRRs. Some of the DAMPs and PRRs display various capacities in driving synovitis and/or cartilage degradation in different models of animal studies. New findings reveal that the inflammatory complement cascade plays a key in the pathogenesis of OA. Crosstalk between joint tissues such as synovium and cartilage communicated at the cellular level within the innate immune inflammatory network is implicated to play an important role in OA progression. Further studies on how the innate immune inflammatory network impacts the OA disease process at different stages of progression will lead to the development of new therapeutic strategies.

Introduction

Osteoarthritis (OA) is the most common form of arthritis, and is a leading cause of pain and disability [1••]. There is an increasing recognition that OA is a disease of the whole joint [1••]. The major pathological changes of OA include cartilage breakdown, the formation of osteophytes, subchondral bone sclerosis, variable degrees of inflammation of the synovium, degeneration of ligaments and the menisci in the knee, and alteration of the joint capsule [1••]. Multiple lines of evidence suggest that low-grade articular inflammation contributes to OA progression [1••, 2].

Inflammation of the synovium results in synovitis, which can occur in early stages of OA [2]. Although it is generally of a lower grade than that observed in rheumatoid arthritis (RA), synovitis directly contributes to several clinical signs and symptoms including joint swelling and effusion, and reflects the structural progression of the disease [1••, 2, 3]. Histologically, OA synovium displays hyperplasia with an increase in number of synovial lining cells, accompanied by infiltration of inflammatory cells consisting of macrophages and lymphocytes [2]. It is still unclear whether morphological changes that occur in the OA synovium are primarily due to a systemic immune response or occur secondarily to cartilage degradation and lesions of the subchondral bone. Nevertheless, soluble inflammatory mediators are detected in OA synovial fluid, including a variety of cytokines and chemokines such as IL-1β TNFα, IL-6, IL-8, IL-15, and IL-17 that may be involved in promoting synovitis [1••, 2]. The molecular products derived from cellular stress and extracellular matrix (ECM) disruption can also cause synovitis [1••, 2, 3]. Innate immunity has been implicated as an active player in the development of synovitis and activation of downstream inflammatory and catabolic events in articular cartilage that lead to OA progression. This review focuses on recent advancements in the understanding of the roles of innate immune players, including pattern recognition receptors (PPRs), damage-associated molecular patterns (DAMPs), and complement in synovitis development and cartilage degradation in OA.

Role of PRRs and DAMPs in OA

The innate immune system plays an essential role not only in the host defense against microbial invasion but also in modulation of multiple forms of tissue injury and repair, and involves the recognition of distinct pathogen-associated molecular patterns (PAMPs) and DAMPs, respectively, by PRRs [4]. Activation of PRRs triggers cell signaling that leads to the production of pro-inflammatory cytokines and chemokines, and the induction of inflammatory responses [4]. These primordial innate immune inflammatory responses are established to mediate many acute and chronic forms of tissue injury [4]. The tissue degradation seen in the OA joint resembles a chronic injury.

Toll-like receptors (TLRs), the type I transmembrane glycoproteins, are members of the largest PRRs [4, 5]. Their extracellular domain contains leucine-rich repeats, which are primarily responsible for mediating ligand recognition [4, 5]. To date, 10 functional TLRs have been identified in humans [4, 5]. All the TLRs, except TLR3, utilize MyD88-dependent signaling pathways to activate transcription factors with NF-κB playing a major role, leading to production of pro-inflammatory mediators including cytokines and chemokines [4, 5]. Many members of the TLR family have been detected in synovial macrophages [6••] and chondrocytes [7, 8] in both OA and RA.

The receptor for advanced glycation end-products (RAGE) is also a member of the PRR family. It is a type I single-pass transmembrane protein, which belongs to the immunoglobulin (Ig) superfamily of cell surface receptors [9]. In the majority of healthy adult tissues, RAGE is expressed at low basal levels. However, up-regulation of RAGE has been associated with diverse pathological events including OA [10], where an inflammatory process is commonly present [9]. Engagement of RAGE triggers multiple signaling pathways leading to inflammatory gene expression via several transcription factors including NF-κB [9].

Endogenous molecular products derived from cellular stress and extracellular matrix disruption can function as DAMPs to activate TLRs and RAGE [11, 12]. Concentrations of some of DAMPs are increased in joint synovial fluid or tissues during joint injury and OA. These include alarmins such as high mobility group box protein 1 (HMGB1) and S100/calgranulin family (e.g., S100A4, the S100A8/A9 heterodimer, S100A11, S100A12), altered or degraded ECM components such as tenascin C (TN-C) and low molecular weight hyaluronan (LMW-HA), modified extracellular proteins (advanced glycation end-products (AGEs), and certain plasma proteins (Table 1) [6••, 11–16, 17•, 18–22]. Numerous studies have shown that these DAMPs stimulate synovial cell proliferation, influence hypertrophic chondrocyte differentiation and induce inflammatory and pro-catabolic events in vitro, and promote synovitis and cartilage degradation in vivo in murine OA models [11–16, 17•, 18–25]. Interestingly, HMGB1 and S100A8/A9 heterodimer can function through both TLR and RAGE signaling pathways [26]. A crosstalk between TLR and RAGE has been revealed [27]. Specifically, blocking the functions of TIRAP and MyD88, two intracellular adaptors for both TLR2 and TLR4, inhibits ligand-activated RAGE intracellular signaling [27]. The signaling pathways of TLR and RAGE also converge at some common points [27]. The cooperation between TLR and RAGE may depend on the cell type and particular pathological context [27]. It would be of interest to determine if there is a crosstalk between TLR and RAGE at the cellular level in the OA joint and whether the physiological significance of such cooperation exists.

Table 1.

Innate immunity in OA

Innate immune players	Agonists or DAMPs implicated	Significance (in vivo testing)
Complement: C5, C6, MAC (C5b-9)	COMP Fibromodulin Aggrecan	C3 knockout: No chondroprotection in the medial meniscectomy model [41••] C5, C6 knockout (disruption of MAC formation): Decreased OA in the medial meniscectomy model [41••] CD59 (the MAC inhibitor) knockout: Increased OA in the medial meniscectomy, DMM and spontaneous aging models [41••]
PRRs: TLRs, RAGE, sCD14, NLRP3	S100A8/A9 (TLR4, RAGE), S100A4 and S100A11 (RAGE) S100A12 (RAGE) HMGB1 (TLR2, TLR4, RAGE) TN-C (TLR4) LMW-HA (TLR2, TLR4) Plasma proteins from OA SF (TLR4) AGEs (RAGE) Hydroxyapatite (TLR4, NLRP3) CPPD (TLR2, NLRP3) Other potential candidates (not reviewed): Biglycan (TLR2, TLR4) Necrotic cell debris, oxidized phospholipids (TLR2, TLR4) HSPs (e.g., Hsp96) (TLR2, TLR4) Fn extra domain A (TLR4)	S100A9-knockout: Decreased OA in the collagenase- induced model [30] TN-C knockout: Increased OA in the ACL and MCL transection model [28] TLR2 knockout: Increased OA in the collagenase-induced model; no significant effect in the DMM model [29] TLR2/TLR4 double knockout: Mild chondroprotection in the meniscectomy model (Liu-Bryan, unpublished data) RAGE knockout: No chondroprotection in the ACL transection model [31] Increased SF sCD14: Seen in early and advanced human OA [36] Knockout of NLRP3 in ANK-deficient mice: Less joint inflammation and cartilage degeneration [54•]

ACL anterior cruciate ligament, AGEs advanced glycation-end products, COMP cartilage oligomeric matrix protein, CPPD calcium pyrophosphate dihydrate, DMM destabilization of the medial meniscus, Fn fibronectin, HA hyaluronan, HSP heat shock protein, LMW low molecular weight, MAC membrane attack complex, MCL medial collateral ligament, NLRP3 NACHT LRR and PYD domains-containing protein 3, PPRs pattern recognition receptors, RAGE receptor for AGEs, sCD14 soluble CD14, SF synovial fluid, TLR toll-like receptor, TN-C tenascin-C

In vivo studies of deficiency of TLR or RAGE and their agonists in animal models of OA have given mixed results (Table 1). For example, TN-C induces synovitis in a TLR4-dependent manner when injected into the mouse knee [25]. TN-C is not involved in the initiation of inflammation but is required for maintenance of joint inflammation [25]. In contrast, knockout of TN-C leads to worsened cartilage degeneration in the aging-related spontaneous OA and mechanical injury-induced OA mouse models [28•]. Deficiency of TN-C also delays articular cartilage repair in mice [28•]. TLR2 deficiency increased the extent of OA in the collagenase-induced model of mouse knee OA, which is associated with significant synovitis [29••]. However, in the surgically-induced destabilized medial meniscus (DMM) model, which is not associated with marked inflammation, knockout of TLR2 did not decrease experimental OA in mice [29••]. Similarly, S100A9 deficiency did not reduce focal OA cartilage destruction in the DMM model, but significantly decreased synovitis and cartilage destruction in the collagenase-induced model [30••]. RAGE deficiency also had no protective effect in the ACL tear-induced mouse knee OA [31]. The various capacities of these PRRs and DAMPs in driving synovitis and/or cartilage degradation in studies of different animal models could be informative. Each of these PRRs and DAMPs may have differential effects in specific OA phenotypes, and/or in different stages of OA.

Although deficiency of TLR2 and TLR4 attenuates catabolic responses to HMGB1 and LMW-HA in chondrocytes in vitro [20], mice deficient in both TLR2 and TLR4 had only a mild protection against OA in the medial meniscectomy model in vivo (R. Liu-Bryan, unpublished data). Immunohistochemistry analysis of the knee cartilage of these mice showed decreased expression of antioxidant enzymes such as SOD2, catalase, and thioredoxin (R. Liu-Bryan, unpublished data), suggesting a potential role of TLR2 and TLR4 in chondrocyte redox homeostasis. Interestingly, decreased SOD2 expression is seen in the superficial layer of human OA knee cartilage [32], and in the medial tibial condyle cartilage before and after the development of OA-like lesions in the spontaneous OA guinea pig model [33]. Knockdown of SOD2 expression in chondrocytes leads to oxidative damage and mitochondrial dysfunction in chondrocytes in vitro [34]. These data suggest that redox imbalance in chondrocytes could contribute to OA development and progression, and that TLR2 and TLR4 may act as the “double-edged swords” during the process of the disease.

CD14 also belongs to the PRR family. It is a co-receptor for several TLRs including TLR2 and TLR4 [35]. It is anchored into the membrane by a glycosylphosphatidylinositol tail (mCD14). CD14 acts not only as a transporter of ligands but also as a signal amplifier by moving TLRs into kinase-rich environments of lipid raft microdomains, where they associate with Src family kinases and heterodimeric G proteins [35]. A soluble form of CD14 (sCD14) also exists, arising from shedding of mCD14 or directly from intracellular vesicles through secretion [35]. Recent studies have revealed that sCD14 is increased in synovial fluid in early stage OA following meniscectomy, as well as in advanced knee OA at the time of total joint replacement, compared with asymptomatic postmortem donors [36•]. The SF levels of sCD14 were significantly higher than in paired serum samples [36•]. In addition, the SF of early OA patients can modulate the inflammatory responses of cultured synovial lining cells to exogenous microbial TLR2 and TLR4 ligands in vitro [36•]. However, it is not known whether CD14 is needed for the signaling of the endogenous TLR2 and TLR4 agonists in the OA joint.

Role of Complement in OA

The complement system is another essential component of innate immunity [37]. It comprises more than 30 plasma and membrane-bound proteins, and can be activated through three pathways: the classical, the alternative, and the lectin pathways [37]. All these result in turnover of C3 and, subsequently, in activation of the terminal pathway and formation of the membrane attack complex (MAC, C5b-9) [37]. The three activation pathways also require an efficient system of regulatory proteins to ensure that complement is not exhausted and to prevent damage to host cells [37]. The main biological function of the complement system is to recognize “foreign” particles and macromolecules, and to promote their elimination either by opsonization or lysis. Complement also plays a role in clearance of damaged or “used” host components [37]. Because complement proteins opsonize or lyse cells, complement can damage healthy host cells and tissues when activated excessively or improperly [37]. This has been linked to certain degenerative diseases including Alzheimer’s disease and age-related macular degeneration [37]. Activation of complement is also among the innate immune drivers of inflammatory arthritis such as RA [38].

Complement deposition is observed in the synovium of patients with cartilage degradation [39], and the components of complement are detected in OA synovial fluid [40]. Recent studies by Wang et al. have shed light on a critical role of the complement cascade in the pathogenesis of OA [41••]. A synovial fluid proteomic screen showed that complement proteins such as C3a and C5b-9 were overexpressed in the synovial fluid (SF) from early stage OA patients, compared with SF from healthy donors [41••], suggesting that complement activation occurs early in the joint during OA development. Additionally, the MAC was also found to be present in the synovium and cartilage from end-stage OA patients [41••]. Moreover, a synovial membrane transcriptome analysis demonstrated that levels of complement effectors transcripts were markedly higher, and levels of complement inhibitor transcripts were considerably lower in the synovium from patients with OA than from healthy donors [41••]. Furthermore, deficiency of C5 (a central component effector) or C6 (a component of the MAC), or addition of complement inhibitory pharmacological agents, protected mice from development of OA following medial meniscectomy [41••]. Conversely, deficiency of CD59a, the MAC inhibitor, promoted increased spontaneous OA in aging mice, and caused more severe OA in mice in the DMM model [41••]. Interestingly, C3 knockout was not chondroprotective in mouse OA in vivo [41••]. The possible explanation is that coagulation factors compensate for the lack of C3, allowing C5 activation to proceed.

How is the complement system activated in OA? Previous studies have demonstrated that the molecular components of extracellular matrix (ECM) such as cartilage oligomeric matrix protein (COMP) and fibromodulin, which are in higher concentrations in OA joint fluids, are capable of activating the complement cascade (Table 1) [42, 43]. A recent study showed that pulverized articular cartilage can induce the formation of MAC [41••]. Interestingly, aggrecan, but not type II collagen, has the capacity to activate complement, despite that both of them are major components of ECM of articular cartilage [41••]. These results suggest that the release or exposure of cartilage ECM due to dysregulation of cartilage remodeling and repair could contribute to OA by activation of complement. Other damaged or altered “self” components such as apoptotic and necrotic cells from joint tissues may also have a potential to activate complement in OA. Cartilage damage promoted by the MAC in the studies by Wang et al [41••] is probably not due to chondrocyte lysis, as the MAC stimulates chondrocytes to express genes of proinflammatory cytokines and cartilage-degrading enzymes at sub-cytolytic concentrations [41••]. In addition, sublytic MAC also induces the expression of complement effectors in chondrocytes [41••], indicating that chondrocyte-produced complement may synergize with synovium-derived complement to amplify pathogenic responses in OA. How the MAC stimulates production of inflammatory cytokines and catabolic enzymes remains to be determined.

Is There any Crosstalk Between TLRs and Complement in the OA Joint?

Although TLRs and complement, the two main players of innate immunity, have been studied extensively as separate entities, increasing evidence indicates that there is crosstalk between TLR and complement signaling pathways [44, 45]. Molecular interplay between the two systems can be either protective against microbial invasion by enhancing the immediate immune reaction to better prime the host in mounting a more robust adaptive immune response, or harmful to the host when activated improperly or uncontrolled by synergistic amplification of the inflammatory responses. For example, certain microbial molecules such as LPS (a TLR4 agonist), zymosan (a TLR2/6 agonist), and CpG DNA (a TLR9 agonist) activate complement in addition to initiating TLR signaling [46]. Systemic administration of these molecules to mice lacking decay-accelerating factor (DAF), a major membrane-associated complement inhibitor, induces a synergistic inflammatory response, demonstrated by significantly higher production of TNF-α, IL-1β, and IL-6 relative to wild-type mice [46]. The influence of complement on TLR signaling is exerted predominantly via the C5a receptor (C5aR), involving increased mitogen-activated protein kinase (MAPK) and NF-κB activation [46]. Blockade of the TLR co-receptor CD14 inhibits complement activities [47]. Complement activation also induces up-regulation of CD14, enhancing TLR4-induced inflammatory responses [48]. Cooperative complement–TLR induction of innate responses could be amplified by TLR-induced cytokines, such as IL-6, which promote the expression of C3aR and C5aR [49]. Given the fact that many microbial ligands are capable of activating both TLRs and complement, endogenous agonists derived from damaged tissue such as components of ECM found in OA synovial fluid may also cause coincidental activation of the two pathways, leading to damaging inflammatory responses. Future studies to determine a regulatory link between TLR and complement systems at the cellular level in the OA joint are expected and may lead to novel approaches for therapy.

Role of the Soluble Inflammatory Mediators in the Innate Immune Inflammatory Network

Soluble inflammatory mediators including cytokines and chemokines are increased in OA synovial fluids, and in joint tissues post-injury. Activation of innate immune signaling pathways of PRRs and complement can lead to generation of these soluble inflammatory mediators, produced by a variety type of cells in joint tissues including synovial fibroblasts, macrophages, and chondrocytes [50]. Among these inflammatory mediators, IL-1β and TNFα, the two most extensively studied, have been implicated in the pathogenesis of OA [2, 3]. Both IL-1β and TNFα are thought to diffuse into the synovial fluid and act on chondrocytes to suppress matrix synthesis [2, 3]. In addition, they induce expression of other cytokines such as IL-6 and IL-8, and promote cartilage catabolism by producing matrix-degrading enzymes in both synoviocytes and chondrocytes [2, 3]. Moreover, IL-1β up-regulates expression of certain PRRs such as TLR2 and RAGE [10, 51], induces release of DAMPs such as HMGB1 and low molecular weight hyaluronan [20], and promotes synthesis of classical pathway complement components [52] in cultured chondrocytes. These suggest that a positive feedback loop can occur between innate immune players and inflammatory cytokines at the cellular level in the OA joint, resulting in augmentation of synovial inflammation and cartilage breakdown.

It is known that production of active IL-1β requires activation of the NLRP3 (NACHT, LRR, and PYD domains-containing protein 3) inflammasome, an innate immune complex [53]. Certain DAMPs present in the OA joints, including hydroxyapatite (HA) and calcium pyrophosphate dihydrate (CPPD) crystals, can activate the NLRP3 inflammasome in macrophages [54•, 55]. In the ANK-deficient mice, HA crystals are extensively deposited in articular cartilage and synovial fluid, resulting in joint space narrowing, cartilage erosion, formation of bony outgrowths, and eventually joint immobility. However, both joint inflammation and cartilage degeneration are decreased by knockout of NLRP3 in the ANK-deficient mice [54•]. These results indicate that the NLRP3 inflammasome might be critical for mediating synovitis and joint destruction in the pathogenesis of OA in certain OA phenotypes within specific clinical settings.

Conclusions

Synovial inflammation plays a critical role in the symptoms and structural progression of OA. However, the mechanism by which synovial inflammation is triggered is still unknown. Studies have revealed that crosstalk between the joint tissues such as synovium and cartilage, communicated at the cellular levels within an innate immune inflammatory network, can promote synovitis and cartilage degradation. Innate immunity can also prime chronic and deleterious adaptive immune responses, and could contribute to the reported presence of Th1 cells expressing activation markers in OA synovium [56]. The physiologic impact of this observation warrants further investigation. It should be noted that innate immune effector mechanisms can mediate tissue repair, with protective effects on bone erosion described by Toll/interleukin-1 receptor domain-containing adaptor inducing interferon β (TRIF) modulation of IL-17 in the T cell activation-driven SCW model inflammatory arthritis [57].

Innate immune players (DAMPs, PRRs, complement) and inflammatory cytokines can mutually regulate each other in articular cells. Crosstalk between different PRRs (e.g., TLR and RAGE) and/or between PRRs and complement may exist in articular cells. Some of these innate immune players may have redundant, paradoxical, antagonistic, or various effects in distinct stages of OA, and/or in certain OA phenotypes within specific clinical settings. Targeting single molecules may be insufficient or problematic, or may not be pragmatic in the case of complement. OA is a disorder of the whole joint. Further studies on understanding crosstalk between joint tissues at the cellular level within the innate immune inflammatory network that impacts the disease process at different stages of progression will lead to the development of new therapeutic strategies.

Key Points.

• Innate immunity is an active player in the development of synovitis and activation of downstream inflammatory and catabolic events in articular cartilage.

• DAMPs and PRRs have various capacities in promoting synovitis and/or cartilage degradation, which may be dependent on specific OA phenotype or stage of OA.

• Dysregulation of complement in synovial joints plays a key role in OA pathogenesis.