Background
The intervertebral disc (IVD) is composed of a central gelatinous nucleus pulposus (NP), a peripherally located annulus fibrosus (AF), and the cranial and caudal cartilaginous endplates associated with its capillary beds. In the degenerative IVD, alterations in cellular activity occur, changing the composition and concentration of extracellular matrix proteins: the synthesis of type II collagen and proteoglycans decreases and the synthesis of type I collagen increases, leading to tissue dehydration and fibrosis. Degenerative IVD cells upregulate tissue degradative enzymes, such as the matrix metalloproteinases (MMPs) and a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTSs), which play a role in matrix degradation during disc degeneration [
1‐
3]. Interleukin-1β (IL-1β) and tissue necrosis factor-α (TNF-α) are considered to strongly influence these degenerative processes [
4]. Proinflammatory cytokines, including IL-1β and TNF-α, are overexpressed in degenerated and herniated IVDs [
5,
6], resulting in the loss of tissue cellularity by upregulation of genes involved with the apoptotic pathway [
7,
8]. Proinflammatory cytokines also stimulate the expression of matrix-degrading enzymes and decrease the synthesis of matrix proteins, consequently leading to homeostatic imbalance, followed by disruption of the extracellular matrix that characterizes IVD degeneration (see review in [
8]). These biochemical and molecular changes within the degenerated IVD are considered to be associated with low back pain and degenerative disc diseases [
9,
10].
The receptor activator of nuclear factor kappa B (NF-κB) ligand (RANKL), initially identified as a member of the TNF ligand superfamily, is well known to regulate bone metabolism [
11,
12]. Previous studies have shown that the signal from RANKL and its receptor (RANK) play an essential role in the differentiation, activity and survival of osteoclasts [
10]. RANKL interacts with RANK, which is expressed on the membranes of mature osteoclasts and osteoclast precursors, to promote differentiation and activation of osteoclasts. Osteoprotegerin (OPG), which is secreted by stromal cells and osteoblasts, acts as a soluble decoy receptor for RANKL. By binding to RANKL, OPG inhibits the interaction of RANKL-RANK, thereby preventing RANK activation and subsequent osteoclastogenesis [
11,
12].
The RANK/RANKL/OPG system has been shown to be expressed in human articular cartilage; however, its functional role and relevance to the pathogenesis of knee osteoarthritis remains unknown [
13,
14]. On the other hand, RANKL and OPG have also been reported to be expressed by the human IVD and are considered to be involved with the degeneration process of IVDs [
15‐
18]. We previously found the expression of the RANK/RANKL/OPG system in the rat IVD. In the rat IVD, the expression of RANKL is regulated by stimulation with IL-1β and the expression of catabolic factors, such as IL-1β, MMP-3, and MMP-13, is enhanced by stimulation with RANKL in the presence of IL-1β [
18]. Therefore, we hypothesized that the RANK/RANKL/OPG system may play a part in the complex molecular mechanism of human IVD degeneration by interactions with the proinflammatory cytokines network, and that this system would be involved in the degenerative process of the human IVD.
The purposes of this study were (1) to examine the expression of the RANK/RANKL/OPG system by human IVD cells with or without pro-inflammatory stimulation, and (2) to evaluate the effects of the RANK/RANKL/OPG system on the expression of catabolic factors, including proinflammatory cytokines and matrix-degrading enzymes, by human IVD cells under stimulation by pro-inflammatory cytokine (IL-1β).
Discussion
In this study, the constitutive expression of RANK, RANKL and OPG by human AF and NP cells was identified. In particular, the basal expression levels of the three molecules in NP cells were higher than those in AF cells. Stimulation with exogenous IL-1β remarkably upregulated the expression of RANK, RANKL and OPG by both AF and NP cells. The administration of exogenous RANKL alone did not induce a change in the expression of catabolic factors, including proinflammatory cytokines (e.g., IL-1β) and matrix-degrading enzymes (e.g., MMP-3 and -13) by human AF and NP cells. Interestingly, these catabolic factors were however significantly upregulated by treatment with RANKL in the presence of IL-1β. On the other hand, the administration of rhOPG in the presence of IL-1β suppressed the expression of catabolic factors by human AF and NP cells upregulated by IL-1β stimulation. A similar effect was identified by treatment with anti-human RANKL monoclonal antibody (ahRANKL-mAB) in the presence of IL-1β.
Our quantitative mRNA analysis showed that the constitutive expression of the RANK/RANKL/OPG system in NP cells was higher than that in AF cells. The NP in the adult human IVD, originally derived from the central notochord in the embryo, is rich in chondrocyte-like cells, whereas the AF, derived from mesenchymal tissue surrounding the central notochord, is rich in fibrocartilage-like cells [
1,
19]. The biological difference between NP and AF cells might account for the differences in the constitutive expression of the RANK/RANKL/OPG system. Byron et al. reported that a high concentration of OPG was found in the culture media of equine articular chondrocytes compared to that of equine articular synovial fibroblasts [
20], suggesting that the production of OPG was higher in a chondrocytic phenotype than in a fibroblastic phenotype. Similar to the results of Byron’s study, the results of this study suggest that the basal metabolism activity of the RANK/RANKL/OPG system in NP cells is higher than that in AF cells.
To imitate the micro-environment in degenerated IVDs in which proinflammatory cytokines were overexpressed [
5], in this study, human IVD cells were cultured in the presence of proinflammatory cytokine IL-1β. Stimulation with IL-1β significantly upregulated the mRNA expressions of RANK/RANKL/OPG. The expression of RANKL and OPG was more upregulated by IL-1β stimulation than that of RANK. Similar findings on the stimulative effects of IL-1β have been reported for osteoblasts [
21‐
23]; the imbalance among RANK/RANKL/OPG expressions is considered to be associated with the progression of inflammatory osteolytic diseases [
24]. Therefore, we speculated that the imbalance of these molecules by human IVD cells might be related to the homeostatic imbalance that is found in the pathogenesis of human IVD degeneration.
Our previous study evaluated the expression of the RANK/RANKL/OPG system in human IVD tissues using immunohistochemical methods [
18]. Semiquantitative immunohistochemical analysis revealed the general trend that the expression of RANK/RANKL/OPG was higher in human IVD tissues in an advanced stage of degeneration compared to that in an early stage of degeneration; these results correspond to the results of pro-inflammatory responses of human IVD cells in vitro in the present study. Therefore, the results of this study further support the hypothesis that the RANK/RANKL/OPG system plays a part in the pathogenesis of human IVD degeneration in the presence of a cytokines network.
To evaluate the effect of RANKL on the matrix metabolism (especially catabolic pathways) of human IVD cells, in the present study, the cells were cultured with RANKL in the presence or absence of proinflammatory stimuli (IL-1β). Similar to the results in our previous study using rat IVD cells [
18], the administration of RANKL alone had no significant effect on the expression of catabolic factors, such as IL-1β and MMPs, by human AF and NP cells. However those expressions were significantly more accelerated by RANKL stimulation with proinflammatory stimuli (IL-1β) than those of IL-1β alone.
Komuro et al. [
13] have previously reported that RANKL alone did not induce the activation of NF-kappa B and the expression of proinflammatory mediators, and concluded that RANKL alone has no effect to stimulate the catabolic factors that are relevant to the pathogenesis of osteoarthritis. For this reason, Kwan et al. [
14] have pointed out the small proportion of RANK-positive cells in human articular chondrocytes. From the results of our and previous studies [
13,
14,
25,
26], we speculate that proinflammatory cytokines, including IL-1β, may have the ability to enhance RANKL signaling by affecting the quantity and/or quality (isoforms) [
25,
26] of RANK found in human IVD cells. In short, RANKL may have the potential to stimulate the expression of catabolic factors in the proinflammatory cytokines-rich environment of degenerated IVDs.
To evaluate the effect of OPG on the expression of catabolic factors, in this study human IVD cells were cultured with rhOPG with or without IL-1β stimulation. Previous studies by Komuro et al. [
13], in which human articular chondrocytes were cultured with OPG, showed no significant effect of OPG on the collagenase activity of human chondrocytes; however, Kwan et al. [
14] showed that the protein level of MMP-13 and PAR-2 was significantly upregulated by OPG-Fc (composed of the RANKL-binding domains of OPG linked to the Fc portion of IgG). Kadri A. et al. [
27] reported in a murine osteoarthritis model that the systemic administration of OPG suppressed the expression of ADAMTS-4 and ADAMTS-5 by murine articular chondrocytes and prevented cartilage degradation in vivo. Thus, the effect of OPG on the induction of catabolic factors by articular chondrocytes has been controversial. The results of our study in human IVD cells demonstrated a general trend that the mRNA expression of catabolic factors stimulated by IL-1β was significantly suppressed by the administration of recombinant human OPG (rhOPG). This suggests that the addition of exogenous rhOPG would suppress RANKL signaling by interacting with the RANK-ligand that had been enhanced by IL-1β stimulation. Shimizu et al. [
28] reported that the intra-articular administration of rhOPG prevented the progression of knee osteoarthritis in a murine model of osteoarthritis, supporting the inhibitory effect of OPG on the progression of matrix degradation in an inflammatory environment. The results of our and Shimizu’s [
28] studies suggest the possibility that OPG could be applied for the treatment of human IVD degeneration by inhibiting RANKL signaling.
To further identify the inhibitory effect of RANKL signaling on the expression of catabolic factors by human IVD cells, the activity of RANKL was next specifically neutralized by anti-human RANKL monoclonal antibody (ahRANKL-mAb). The results of this study have shown, for the first time, that the ahRANKL-mAb has the potential to decrease the mRNA expression of catabolic factors upregulated by proinflammatory stimuli, including IL-1β. In addition, the quantity of MMP-3 released into the medium was also significantly suppressed by ahRANKL-mAb. These results also support our finding that RANKL has a potential role in regulating the expression of catabolic factors that are relevant to the pathogenesis of disc degeneration in the presence of proinflammatory stimuli.
It is now well known that a fully human monoclonal IgG2 antibody that specifically targets RANKL, ‘denosumab’, remarkably improves systemic bone mineral density in patients with osteoporosis [
29]. Recent clinical studies have shown that ‘denosumab’ inhibited the progression of bone erosion in inflammatory diseases, such as rheumatoid arthritis [
30,
31]. More, recently, anti-RANKL monoclonal antibody was systemically administered in a delayed-type hypersensitivity arthritis model; the treatment reduced the destruction of the subchondral bone and, in addition, serum levels of serum amyloid P component and MMP-3 [
32]. Although these studies showed evidence that the systemic administration of anti-RANKL antibody inhibited bone destruction or erosion in inflammatory diseases or conditions, its inhibitory effect on cartilage degradation has not been proved [
30‐
32].
Interestingly, Sato et al. [
33] reported that the administration of anti-RANKL antibody to punctured rat intervertebral discs significantly downregulated the expression of proinflammatory cytokines, such as IL-6 and TNF-α, in dorsal root ganglion (DRG) neurons innervating the injured disc. Their results suggest that anti-RANKL antibody may have the potential to suppress the inflammatory response that is associated with pain transmission within the degenerated discs. The results of this study and Sato’s studies suggest that anti-RANKL antibody might have inhibitory effects against the inflammatory responses that are relevant to the progression of human IVD degeneration.
In 2008, Neogi et al. [
34] reported in the secondary analysis of data from the fracture intervention trail (FIT) that the alendronate (ALN) treatment group showed lower spinal osteophyte and disc height narrowing scores than the placebo group, suggesting that the administration of bisphosphonate (ALN) would be responsible for the inhibitory effect on the progression of disc degeneration.
Luo and colleagues [
35] evaluated the effect of ALN on lumbar IVD degeneration related to osteoporosis using an overiectomized (OVX) rat model. They reported that administration of alendronate in the OVX rat model significantly prevented cartilage endplate thickening, and improved histological scores of disc degeneration, suggesting that ALN treatment was effective in delaying the process of rat disc degeneration [
35]. Thereafter, Song et al. [
36] reported in the same animal model that the systemic administration of ALN inhibited disc height narrowing and the alterations of extracellular matrix and cellular components characterized in the degenerative IVD.
Considering these inhibitory effects of ALN on IVD degeneration in osteoporotic animal models, it can be speculated that anti-RANKL treatment may have a similar effect to suppress the progression of IVD degeneration in the patients with osteoporosis.
The inhibition of RANKL signaling would be more specific than the inhibition of IL-1β or TNF-α signaling on inflammatory reactions. Therefore, the authors speculate the possibility that RANKL inhibitors may have the advantage of reducing multiple adverse effects that have been identified by the clinical use of nonspecific immunosuppressive agents, such as TNF-α inhibitors [
37,
38]. In addition, the anti-RANKL antibody, ‘denosumab’, has been clinically used for osteoporosis patients, and its safety [
29‐
31] and cost-effectiveness compared to other osteoporosis treatments [
39,
40] has been reported. Therefore, the clinical applicability of anti-RANKL antibody treatment for low back pain patients would be high.
There are several limitations in this study. First, all data were obtained from in vitro studies that mimic the in vivo conditions of human IVD degeneration. Second, human IVD samples with different MRI grades of disc degeneration were utilized in this study; therefore, a potential bias regarding the gene expression of RANK/RANKL/OPG and the response of IL-1β, RANKL, OPG or ahRANKL-mAb may exist. Third, this study focused on catabolic factors related to IVD degeneration, thus, further study is needed to evaluate anabolic factors to provide a detailed analysis of the mechanisms of disc degeneration. Fourth, although the effects of IL-1β, RANKL, OPG or ahRANKL-mAb were mainly evaluated by quantitative-PCR, additional evaluations of those effects by examining protein expression, such as by western blot, would further support the results of those mRNA expressions.