CYLD suppression enhances the pro-inflammatory effects and hyperproliferation of rheumatoid arthritis fibroblast-like synoviocytes by enhancing NF-κB activation

Rheumatoid arthritis (RA) is a systemic and chronic inflammatory disease characterized by synovial hyperplasia formation, which mediates cartilage and bone destruction [1].The pathogenesis of RA is extraordinarily complicated and involves various kinds of cells, such as fibroblast-like synoviocytes (FLSs), T cells, B cells, monocytes/macrophages, and osteoclasts [2]. It has been indicated that each cell type plays distinct, complex, and interrelated roles in the development of RA [3]. In RA, the synovial lining thickness at the border of the joint cavity increases to 10–15 cell layers, compared with 1–3 cell layers in normal individuals [4, 5]. Prior studies have demonstrated that FLSs are the predominant cell type in the terminal layer of the hyperplastic synovium and at the sites of invasion in adjacent cartilage and bone joints, and that they vigorously contribute to the initiation and early perpetuation of RA [4, 6].

FLSs in RA, also known as synovial fibroblasts or type B synoviocytes, occur early after onset with visible erosion at cartilage–bone junctions [7]. FLS activation involves multiple factors, including Toll-like receptor (TLR) activation, inflammatory factors, matrix degradation products, and epigenetic modifications [8]. Activated RA-FLSs are highly able to recruit, retain, and activate both cells of the immune system and resident joint cells through hyperproduction of a broad array of pro-inflammatory cytokines, matrix metalloproteinases (MMPs), and chemokines, leading to the initiation and maintenance of chronic inflammation and progressive joint destruction [9]. The pro-inflammatory cytokine tumor necrosis factor alpha (TNF-α) is regarded as one of the most central cytokines that significantly triggers inflammation and joint destruction, which is confirmed by the clinical efficacy of TNF-α-blocking agents [10]. TNF-α, interleukin-1β (IL-1β), interleukin-6 (IL-6), and interleukin-8 (IL-8), the most important cytokines, can obviously aggravate inflammatory reactions and increase the influx of additional pro-inflammatory cells into the synovium to sustain regulatory feedback loops and induce production of MMPs, cathepsins, and aggrecanases [11, 12]. The synovium is the main source of MMPs and cathepsins in RA, and in situ hybridization studies have localized collagenase messenger ribonucleic acid (mRNA) nearly exclusively to RA-FLSs [13]. Activated MMPs can significantly drive the degradation of the extracellular matrix (ECM) and facilitate cartilage and bone erosion, while inhibition of MMPs can greatly alleviate the invasiveness of RA-FLSs into cartilage [14‐16]. RA-FLSs also influence bone destruction via modulation of osteoclastogenesis by secreting receptor activator of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) ligand (RANKL) [17]. Denosumab, a human monoclonal antibody that specifically binds RANKL, has been effective for the treatment of RA-associated bone loss [18, 19]. In addition, synovial hyperplasia in RA appears to be caused at least in part by the impairment of apoptosis in RA-FLSs and synovial macrophages, while deficient apoptosis has been shown to prolong RA-FLS survival by increasing the production of anti-apoptotic molecules like B-cell lymphoma 2 (Bcl-2), sentrin-1 (SUMO-1), and Fas-associated death domain-like interleukin-1- converting enzyme inhibitory protein (FLIP) [19, 20].

Cylindromatosis (CYLD) was initially identified as a tumor suppressor that is mutated in patients with familial cylindromatosis, a genetic condition that predisposes patients for the development of skin appendage tumors (cylindroma) [21]. CYLD is a key adaptor that regulates various signaling pathways to modulate diverse physiological processes, ranging from immune response and inflammation to cell cycle progression, spermatogenesis, and osteoclastogenesis [22]. Emerging evidence suggests that CYLD primarily downregulates NF-κB signaling by deubiquitinating NF-κB essential modulator (NEMO) and several upstream regulators like TNF receptor-associated factors 2 and 6 (TRAF2 and TRAF6) and TGF-β-activated kinase 1 (Tak1) [23‐25]. The family of NF-κB transcription factors plays a large role in mediating the expression of large numbers of genes during the inflammatory response. Many previous in vitro and in vivo studies on the contribution of components of NF-κB signaling pathways to the pathogenesis of RA have shown that NF-κB plays prominent roles in inflammation, cartilage degradation, cell proliferation, angiogenesis, and pannus formation [26]. Until now, however, the role and the underlying mechanism of CYLD in local inflammation and joint destruction in RA have been poorly elucidated. Therefore, in this study, we determined CYLD expression in synovia from patients with RA and analyzed its correlation with NF-κB activation or clinical disease activity. Then we investigated CYLD expression in RA-FLSs and explored its roles and mechanisms in pro-inflammatory effects, proliferation, apoptosis, and cell cycles of RA-FLSs.

CYLD was initially reported as a tumor suppressor with deubiquitinating enzyme activity that is mutated in familial cylindromatosis. It suppresses the activation of NF-κB, which plays central roles in inflammation, immune responses, carcinogenesis, protection against apoptosis, and promotion of cell proliferation [21]. Our current study showed that CYLD expression in synovia from RA patients was significantly downregulated compared with that in synovia from OA patients. Conversely, NF-κB expression was distinctly upregulated in the RA synovium compared with the OA synovium. This result is consistent with that of a past study which found that NF-κB activation is higher in RA than in OA [36]. In addition, we found a significant inverse correlation between CYLD and NF-κB in the RA synovium, suggesting that NF-κB activation is suppressed when CYLD function is normal, but that it tends to be enhanced in the absence of functional CYLD. Moreover, synovial CYLD expression correlates significantly with ACPA, CRP, and RF. Above all, our results support the possibility that NF-κB activation accompanied by loss of CYLD may be a crucial step in the development and progression of RA. Thus, we postulated that functional relevant loss of CYLD expression may contribute to RA development and progression and may provide a new target for therapeutic strategies.

RA is a chronic inflammatory disease characterized by persistent inflammation of the diarthrodial joints with synovial hyperplasia and progressive joint destruction. In the terminal layer of the hyperplastic rheumatoid synovium, FLSs are the dominant cell types. Similar to tumor cells, they can be expanded in cell culture over several passages and escape contact inhibition to invade and degrade the adjacent cartilage and bone [37]. Activated RA-FLSs can be detected early after RA onset by visible erosions at cartilage–bone junctions [7].There, RA-FLSs can be activated by (among other things) inflammatory factors, TLR activation, matrix degradation products, and epigenetic modifications [8]. Recent studies have revealed that at an initial stage of synovial activation, TLRs are the key recognition structures of innate immune stimuli, such as microbial components, endogenous ligands, liberated cellular RNA, and DNA fragments from necrotic cells within the synovial fluid [38, 39]. TLRs are stimulated by their respective ligands; they dimerize and recruit downstream adaptor molecules, which propagate signals that converge at NF-κB through the TRAF6, TRAF3, and IL-1- and IL-1R-associated kinase 1/4 (IRAK1/4) complex and induce the production cytokines such as TNF-α, IL-1β, and IL-6 [40]. In addition, RA-FLS activation is not only the response to active inflammation within the synovium and the presence of inflammatory cells but constitutes an intrinsic feature of RA-FLSs. This inflammation-independent activation has been elegantly confirmed by studies in which human RA-FLSs implanted into a severe combined immunodeficient (SCID) mouse model of RA attached to and invaded co-implanted healthy human cartilage without cells of the immune system [41]. In addition, Murphy Roths Large mice with or without lymphoproliferation (MRL − lpr/lpr mice) can spontaneously develop RA-like arthritis, indicating that synovial cell proliferation and invasion into joint structures occur even before inflammatory cells migrate into the synovium [42]. Hence, RA-FLSs potentially contribute to the initiation and early progression of RA and are thus key factors in the pathogenesis of RA. Therefore, targeting the aggressive, joint-destructive, and pro-inflammatory properties of RA-FLSs is a recent approach to RA treatment.

Activated NF-κB plays a central role in regulating the expression of numerous genes during the inflammatory response, which is the fundamental process of RA. Generally, NF-κB is present in the cytoplasm as a heterodimeric complex composed of the p50 and p65 subunits, which are associated with inhibitory κB proteins (IκBs) [43]. NF-κB activation typically occurs following specific serine/threonine kinase IκB kinase (IKK) signalosome activation, leading to the phosphorylation and subsequent dispatch of IκBs to the proteasome of protein degradation [44]. IκB degradation enables NF-κB to translocate to the nucleus, stimulating the transcription of specific genes. The IKK complex consists of at least three subunits, including the kinases IKK-α and IKK-β and the associated regulatory subunit IKK-γ [44]. It has been demonstrated that NF-κB and IKK are abundant in RA-FLSs [26, 36]. NF-κB activation can robustly induce pro-inflammatory cytokines, MMPs, and RANKL production in RA-FLSs, which enhances the inflammatory cycle and contributes to joint erosion in RA [45, 46]. NF-κB has been shown to mediate fibronectin fragment-induced chondrocyte activation and to increase pro-inflammatory cytokines, chemokines, MMPs such as IL-6 and IL-8, and methyl-accepting chemotaxis protein-1 (MCP-1), growth-related oncogenes a, b, and g (GRO-α, GRO-b, and GRO-g, respectively), and MMP-13 production in human articular chondrocytes [47, 48]. TNF-α and IL-1β can strongly promote NF-κB activation and enhance the production of IL-6, IL-8, intercellular adhesion molecule 1 (ICAM-1), and collagenase in RA-FLSs [49]. TNF-α, IL-1β, and IL-6 are prominent pro-inflammatory cytokines in the inflammatory cascade of RA and have been widely studied as therapeutic interventions [5, 50].

Numerous studies in vitro and in vivo have validated that CYLD mediates NF-κB activation by deubiquitinating TRAF2, TRAF6, and NEMO, making it an important regulator in the adaptive immune response. However, the expression and exact roles of CYLD in the pro-inflammatory effects of RA-FLSs are unknown. We observed CYLD expression in the lining and sublining layers of RA synovium, with intense staining found in the endochylema as well as the nuclei of lining cells (both macrophage-like synoviocytes and FLSs) and sublining of inflammatory cells, implying that CYLD was expressed in RA-FLSs. Its expression in primarily cultured RA-FLSs was significantly lower than that in OA-FLSs. Our experiments also further demonstrated that suppression of CYLD by sh-CYLD enhanced mRNA expression and the secretion of pro-inflammatory cytokines in RA-FLSs, including TNF-α, IL-1β, IL-6, and IL-8; aggravated the activity of NF-κB in RA-FLS nuclei and IκBα in RA-FLS cytoplasm; and decreased the level of total IκBα in RA-FLS cytoplasm. This suggests that downregulated CYLD plays a conspicuous role in the pro-inflammatory effects of RA-FLSs via NF-κB signaling.

In joints with early-stage RA, activated RA-FLSs attach to and overgrow the articular cartilage surface, then invade and destroy cartilage and induce bone resorption via secretion of MMPs, cathepsins, and inflammatory cytokines and regulation of monocyte-to-osteoclast differentiation [51, 52]. MMPs include collagenases (MMP-1, MMP-8, and MMP-13), gelatinases (MMP-2, MMP-9), stromelysins (MMP-3, MMP-7, MMP-10, MMP-11, and MMP-12), and membrane-type (MT) MMPs (MT1-MMP, MT2-MMP, MT3-MMP, and MT4-MMP). In situ hybridization studies have validated that lining cells are the major source of MMP-1 and MMP-3 in the synovium, especially RA-FLSs [53, 54]. MMP-1 cleaves collagens I, II, VII, and X, while inhibition of MMP-1 synthesis significantly reduces the cartilage invasiveness of FLSs in the SCID mouse model for RA [55]. MMP-3 can degrade cartilage proteoglycans and type IX collagen and activate other pro-MMPs, including pro-MMP-1 and pro-MMP-9 [56]. In addition, osteoclasts are the major course of local and systemic abnormalities of bone remodeling, including bone erosion and both focal and systemic osteoporosis [57]. RANKL is one of the key molecules for osteoclastic differentiation and activity that have been found to be strongly expressed at sites of bone erosion in RA synovium, especially in RA-FLSs [58]. Studies performed in human RA-FLSs have indicated that IL-1β mediates the production of MMP-1 and MMP-3 by regulating NF-κB activation [59]. Inhibition of NF-κB and Jun N-terminal kinase (JNK) activation significantly decreases production of pro-inflammatory cytokines (TNF-α, IL-1β, IL-6, and IL-8) and MMPs (MMP-1 and MMP-3) in RA-FLSs [60]. It has also been reported that NF-κB suppression obviously reduces RANKL expression in human RA-FLSs and in FLSs from mice with adjuvant-induced arthritis (AA) [61, 62]. As CYLD plays a pivotal role in regulating NF-κB activation, it is not clear whether it is involved in the synthesis of MMP-1, MMP-3, and RANKL in RA-FLSs. We showed that CYLD inhibition in RA-FLSs apparently enhanced MMP-1, MMP-3, and RANKL production and NF-κB activity, implying that CYLD participated in mediating RA-FLS-induced cartilage and bone destruction via NF-κB signaling.

In the pathology of RA, it is widely acknowledged that the progressive destruction of articular cartilage relies on the evolution of hyperplastic synovial tissue, while hyperplasia in FLSs depends on dysregulated proliferation and apoptosis [63]. Thus, promoting RA-FLS apoptosis and reducing RA-FLS proliferation are new treatment strategies for RA. It has been reported that CYLD attenuates NF-κB signaling by removing lysine63 (Lys63)-linked polyubiquitin chains from TRAF2 or TRAF6, leading to programmed cell death [64]. Previous studies have also revealed that CYLD suppression leads to decreased apoptosis and increased proliferation by activating NF-κB [23, 65]. In the present study, we found that CYLD inhibition distinctly enhanced proliferation, reduced apoptosis, and increased the cell division of RA-FLSs and aggravated the activity of NF-κB in these cells, suggesting that CYLD is involved in RA-FLS hyperproliferation in RA. However, a recent report indicated that the loss of CYLD in mouse keratinocytes enhances cell proliferation by increasing the nuclear activity of Bcl-30-associated NF-κB p50 and p52 subunits, rather than by increasing classical p65/p50 NF-κB action. This implies that the mechanism of CYLD-mediated NF-κB inhibition may vary by cell type [65].

In summary, downregulated CYLD may be involved in the pathogenesis of inflammation in the RA synovium by regulating classical NF-κB activation, as well as in the pro-inflammatory effects and hyperproliferation of RA-FLSs. Thus, CYLD may provide a potential target for the treatment of RA. However, the precise mechanism of interaction between CYLD and NF-κB in RA-FLSs remains unclear and requires further exploration. In addition, our study showed that CYLD is expressed not only in RA-FLSs but also in inflammatory cells, such as lymphocytes, and in plasma cells. Further studies are needed to investigate the expression and role of CYLD in regard to the pro-inflammatory effects, proliferation, apoptosis, and cell cycles of these cells.