Discussion
RNA sequencing technology has been used to detect differences in gene expression profiles in RA patients for the last few years. Several gene expression profiling studies of synovial tissues and PBMCs from RA patients have revealed marked variation in gene expression profiles that have facilitated the identification of distinct molecular disease mechanisms involved in RA pathology(18-21). The heterogeneity of RA was demonstrated by the presence of distinct autoantibody specificities such as ACPA in serum, differential responsiveness to treatment, and variability in clinical presentation. RA patients can be stratified into two subgroups defined by the presence or absence of ACPA, and ACPA+ patients exhibit more severe inflammation and radiographic damage than ACPA- patients(5-7). In the present study, various genes were compared in ACPA+ and ACPA- patients via RNA sequencing, and two significantly increased chemokines, CXCL2 and CXCL7 were identified in ACPA+ patients. The next more focused analysis using PCR technology on many more patients found that only CXCL2 was differentially expressed in ACPA+ RA patients and ACPA-RA patients. There was no significant difference in CXCL7 expression between ACPA+ RA patients and ACPA- RA patients.
CXCL2 was first identified as a major chemokine produced by endotoxin-treated macrophages(22), which bind to the G-protein coupled receptor CXCR2 expressed on macrophages, neutrophils, and epithelial cells(23). In previous studies, CXCL2 level was found to be higher expressed in RA patients than in normal controls(24). Xiaokun et al. downloaded microarray datasets of GSE1919, GSE12021, and GSE21959 (35 RA samples and 32 normal controls) from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) and identified DEGs in RA samples. They found that CXCL2 was strongly associated with DEGs involved in RA progression(24). Jacobs et al. (25) performed a microarray analysis to characterize the molecular events underlying pathology in autoantibody-mediated arthritis and reported that CXCL2 was up-regulated in parallel with the disease. Jeongim et al. (26) reported that CXCL2 was significantly higher in synovial fluid and sera from RA patients compared with corresponding samples from osteoarthritis patients. In the present study, serum CXCL2 was significantly increased in RA patients compared with healthy controls, and serum CXCL2 was higher in ACPA+ RA patients than in ACPA- RA patients. Compared with ACPA− patients, ACPA+ patients had longer disease duration and higher positive rate of RF. To exclude the impact of these two factors on CXCL2 level, correlation analysis was applied and showed that CXCL2 was irrelevant to disease duration and RF titer.
CXCL2 was involved in various biological progresses, such as angiogenesis, inflammation and cancer biological behaviors(23, 27-30). CXCL2 was also considered to be a proinflammatory factor in RA(24). In line with this, our study showed that CXCL2 was positively correlated with DAS28, ESR and CRP. Moreover, recent studies revealed CXCL2 was also involved in the process of osteoclastogenesis(26, 31). In the present study, we also found serum CXCL2 was significantly higher in RA patients with bone erosion than in RA patients without bone erosion. Thus, we hypothesized that CXCL2 could be one of the key molecules upregulated in RA progression, especially in ACPA+ RA which exhibit more radiographic damage, and focused on the CXCL2 for subsequent analyses.
As the most important osteoclast precursors, CD14+ monocytes were derived from ACPA+ patients and exhibited elevated CXCL2 secretion compared with those derived from ACPA- patients. CXCR2 is known to be the major receptor for CXCL2(32). It has been suggested that CXCL2 acts as a chemoattractant for various types of cells by binding to CXCR2, and monocytes constitutively express CXCR2(32). However, whether there is a difference in the expression of cell surface CXCR2 in CD14+ monocytes from ACPA+ and ACPA- RA patients remains unclear. In the present study, we found that CXCR2 expression was higher on the surfaces of CD14+ monocytes derived from ACPA+ patients than those from ACPA- patients, but the difference was not statistically significant.
ACPA+ RA patients usually develop more severe radiological damage, and enhanced osteoclastogenesis may be involved in the bone erosion(7). Osteoclasts are polykaryocytes formed via the fusion of mononuclear monocytic precursors such as monocytes in peripheral blood. We surmise that elevated CXCL2 may recruit more monocytes to joint sites and augment the formation of osteoclasts. The present study showed that CXCL2 promoted the migration, differentiation, and function of osteoclasts in experiments using CD14+ monocytes isolated from RA patients.
The effects of CXCL2 on activating the signaling pathway during osteoclast differentiation were examined to explore the molecular mechanisms of the observed enhancement effect on osteoclastogenesis. The addition of CXCL2 resulted in dramatically increased phosphorylation of p65 and ERK1/2, while it did not affect the phosphorylation of IκBα or JNK. NFATc1 and c-Fos are critical transcription factors in the regulation of osteoclast differentiation(33). In the current study, NFATc1 and c-fos were increased significantly with stimulation of CXCL2 in the presence of M-CSF and RANKL. NFATc1 is an important regulatory factor in the process of osteoclast differentiation mediated by RANKL-activated MAPK and NFκB signalling pathways(34). Therefore, it suggests that CXCL2 may promote osteoclast differentiation and bone resorption by regulating NFATc1 expression via interfering with the phosphorylation of NFκB p65 and MAPK ERK1/2.
Jeongim et al. (26) reported that RANKL significantly increased the expression of CXCL2 in bone marrow-derived macrophages (BMMs) and that CXCL2 mediated RANKL-dependent differentiation of osteoclasts from BMMs in mice. The osteoclastogenesis-enhancing effects of CXCL2 were further corroborated by their investigation with an in vivo bone resorption model. Notably, CXCL2 alone induced significant bone loss in mice calvarial similar to that induced by RANKL. They also reported that CXCL2 mediated lipopolysaccharide (LPS)-induced osteoclastogenesis in RANKL-primed precursors (i.e. BMMs) (31). LPS stimulated the production of CXCL2 in BMMs, and the conditioned medium from LPS-treated BMMs could enhance the migration of osteoclast precursors, which was blocked by treatment with CXCL2-neutralising antibody or CXCR2 receptor antagonist. Blocking CXCL2 also reduced LPS-induced osteoclastogenesis, and in addition, CXCL2 neutralization prevented bone destruction in mice treated with LPS, suggesting a critical role of CXCL2 in the process of osteoclastogenesis. In the present study, CXCL2 was also evidently involved in the process of osteoclast differentiation in RA patients.
Therapies targeting CXCL2/CXCR2 have been tested in animal models of arthritis with concomitant reduction in neutrophil recruitment and tumor necrosis factor-α production(35, 36), and immunization against CXCL2 was reportedly efficient in delaying the onset of arthritis and reducing disease severity in a murine collagen-induced arthritis model(37). The novel orally-active non-competitive allosteric inhibitor of CXCL2 known as DF 2162 significantly ameliorates AIA in rats, an effect that is quantitatively and qualitatively similar to that of anti-tumor necrosis factor antibody treatment(35). In addition, the CXCR2 antagonist SCH563705 led to a dose-dependent decrease in clinical disease scores and paw thickness measurements, and clearly reduced inflammation and bone and cartilage degradation in a mouse model of arthritis(36). The results of our study imply that targeting CXCL2/CXCR2 as a strategy to treat RA may contribute to protection from bone destruction by directly inhibiting bone resorption, in addition to the already suggested anti-inflammatory effects.
In conclusion, we identified novel pathways associated with ACPA+ RA patients using RNA sequencing, and detected higher CXCL2 expression in ACPA+ RA patients than in ACPA- RA patients. Increased levels of CXCL2 led to NFκB activation in CD14+ monocytes from RA patients during the osteoclastogenic process. These results suggest a previously unreported role of CXCL2 during osteoclastogenesis in RA patients, and indicate that CXCL2 blockade might be a novel therapeutic strategy in RA.