Collagen-induced arthritis (CIA) was induced in 6–8-week-old male DBA/1J, BALB/c, and C57BL/6 mice (Orient, Korea). To generate Fra1/Junb Tg mice, a pcDNA3.1+HA (Invitrogen, CA, USA) vector containing the FRA1 and JUNB proteins coupled to a linker peptide (3 × GGGGS) was constructed. The Fra1/Junb fragment was synthesized by GenScript Corporation (NJ, USA), with codon optimization for expression in mammalian cells. Fra1/Junb Tg mice were bred from the C57BL/6 line and maintained in facilities at Macrogen Laboratories (Seoul, Korea). All mice were maintained under specific-pathogen-free conditions at the Institute of Medical Science, The Catholic University of Korea. The presence of the transgene in the founders was confirmed by PCR of genomic DNA extracted from the tail samples. Genotyping was performed by PCR analysis of genomic DNA obtained from mice at 3 weeks of age. All experimental procedures were examined and approved by the Animal Research Ethics Committee at the Catholic University of Korea.

Raw RNA-seq data have been deposited in the NCBI Sequence Read Archive (SRR6320798 and SRR6320799).

A detailed description of all other experimental procedures and the statistical analysis is provided in the Section “Supplementary Materials and Methods” in Data Sheet S1 in Supplementary Material.

Potential STAT3-binding sites were identified using publicly available chromatin immunoprecipitation sequencing (ChIP-Seq) data cross-referenced with differentially expressed genes in Th17 cells (14). We sequenced mRNA obtained from Th17 cells and naïve T cells and compared the results with potential STAT3-binding targets identified by ChIP-Seq to identify STAT3-regulated genes involved in Th17 cell differentiation. The literature was systematically reviewed for downstream STAT3 targets in humans and mice in multiple biological contexts. By combining these three datasets, we searched for genes with STAT3-binding sites that are upregulated during Th17 cell differentiation (Figure 1A; Table S1 in Supplementary Material). RNA-Seq was performed to identify genes differentially expressed between Th17 cells and naïve T cells. This integrated approach suggested that Fosl1 (the gene locus of Fra1) and Junb are Th17 cell differentiation factors downstream of STAT3. As shown in Figures 1B–F and Table S2 in Supplementary Material, Fos and Jun were the predominant AP-1 subtypes expressed in naïve (CD4⁺CD62L⁺) T cells, whereas Fosl1 and Junb were the predominant subtypes in Th17 cells. Other AP-1 subtypes were downregulated, suggesting that Fosl1 and Junb AP-1 subtypes play an important role in Th17 cell differentiation.

Quantitative real-time reverse transcription PCR (qRT-PCR) and immunoblot analyzes were used to confirm Fra1 and Junb expression. Among the various AP-1 subtypes, only Fra1 and Junb mRNA levels were significantly increased in Th17 cells compared with naive T cells (Figures 2A,B). These findings supported the RNA-Seq results. Immunoblot analysis also showed high levels of FRA1 and JUNB expression in Th17 cells (Figure 2C). Co-immunoprecipitation studies were performed using antibodies against Jun family proteins (i.e., JUNB and c-JUN), followed by immunoblotting using antibodies against FOS family proteins (i.e., FRA1, c-FOS, and FOSB) to determine whether FRA1 and JUNB form a heterodimeric AP-1 complex. It was found that FRA1 forms a complex with JUNB, but not with c-JUN (Figure 2D).

Next, we investigated the functional association between the FRA1–JUNB complex and STAT3. The STAT3 signaling pathway was inhibited using genetically engineered STAT3-deficient CD4⁺ T cells and STA-21, which blocks the DNA-binding activity of STAT3 (26). As shown in Figure 3A, CD4⁺ T cells harboring Stat3^fl/flCd4-Cre had markedly decreased levels of Fra1 and Junb mRNA under both neutral and Th17-polarizing conditions. Treatment with STA-21 profoundly decreased the expression of FRA1, JUNB, and STAT3 (Figure 3B). Moreover, STA-21 decreased the levels of STAT3 phosphorylation under Th17-polarizing conditions (Figure 3C). STAT3 overexpression increased the relative levels of Fra1 and Junb mRNA (Figure 3D), while in vitro treatment with IL-6 upregulated Fra1 and Junb mRNA (Figure 3E).

Co-immunoprecipitation studies were performed using anti-FRA1 and anti-JUNB antibodies followed by immunoblotting with antibodies against STAT3 to determine whether FRA1 and JUNB physically interact with STAT3. Both FRA1 and JUNB were found to form a complex with STAT3 (Figure 3F). We performed ChIP and qPCR to examine whether STAT3 binds directly to the promoter regions of Fosl1 and Junb. The ChIP-qPCR results were compared with publicly available ChIP-Seq data from Th17 cells. We found that STAT3 binds to the predicted recognition sites upstream of Fosl1 and Junb, suggesting that it regulates their expression by binding directly to their promoters (Figure 3G). Using confocal microscopy, we found that FRA1 and JUNB were colocalized with phosphorylated STAT3 in CD4⁺ splenic T cells and Th17 cells from arthritic mice (Figures 3H,I). These findings suggest that STAT3 signaling is critical for the expression and activation of FRA1 and JUNB during Th17 cell differentiation.

The AP-1 family plays a versatile role in T-cell development (27). Levels of IL-17 were measured in Fra1- and Junb-overexpressing and silenced cells. The relative levels of IL17 mRNA were significantly increased in cells overexpressing Fra1 and Junb. The relative mRNA levels of other genes associated with Th17 cells, such as IL21, IL22, and Rorc (the RORγt gene locus), were also increased (Figures 4A,B). CD4⁺IL-17⁺ cells and IL17 mRNA levels were significantly decreased when Fra1 and Junb were silenced under Th17-polarizing conditions (Figures 4C,D). By contrast, when c-Fos and c-Jun, the dominant AP-1 subtypes in naïve T cells, were overexpressed, IL17 mRNA levels were decreased significantly (Figures 4E,F). These findings support the specificity of the role of the FRA1–JUNB complex in Th17 cell differentiation.

We next investigated whether the FRA1–JUNB complex directly regulates IL-17 expression by binding to the IL17a promoter. The interaction between the FRA1–JUNB complex and the AP-1-binding motif within the IL17a promoter region was examined by electrophoretic mobility shift assay. The anti-FRA1 and anti-JUNB antibodies caused a super shift of the labeled DNA probe, indicating that the AP-1-binding motif interacts specifically with FRA1 and JUNB (Figure 4G). This finding was confirmed by an analysis of the IL17a promoter using a luciferase reporter system. Overexpression of both FRA1 and JUNB increased luciferase activity in EL4 cells. Simultaneous overexpression of FRA1 and JUNB caused luciferase activity to increase twofold (Figure 4H). Silencing of Fra1 and Junb caused a greater than twofold reduction in luciferase activity. Simultaneous overexpression of c-FOS and c-JUN caused a significant decrease in the luciferase activity induced by the IL17a promoter (Figure 4I). ChIP-qPCR analyses showed that FRA1 and JUNB bind directly to the IL17a gene loci and to gene loci encoding other Th17 cell differentiation activators, such as IL21, IL22, and Rorc (Figure 4J). These findings indicate that the FRA1–JUNB complex binds to the IL17a promoter region, and that this binding activates IL17 gene transcription.

T helper 17 cells and IL-17 contribute significantly to the development of RA (28, 29). To examine the in vivo role of the FRA1–JUNB complex in rheumatoid inflammation, Fra1 and Junb were silenced in CIA mice by injecting Fra1 and Junb short hairpin RNA vectors. This markedly reduced the incidence of arthritis and decreased the arthritis score (Figure 5A). The numbers of FRA1⁺, JUNB⁺, and IL-17⁺ cells were significantly lower within the joint tissues (Figure 5B). Histological examination showed that the ankle joints of Fra1/Junb-silenced mice had less inflammation and cartilage damage (Figure 5C). Inhibition of Fra1/Junb also reduced the expression of Th17 cytokines such as IL-6, IL-17, IL-1β, and tumor necrosis factor-α in joint tissue. Other mediators of joint destruction such as vascular endothelial growth factor and receptor activator of nuclear factor κB ligand (RANKL) were also downregulated (Figure S1A in Supplementary Material). In CIA mice, disease severity was correlated with the level of type II collagen (CII)-specific IgG antibodies (30). Silencing of Fra1 and Junb reduced T-cell proliferation and CII-specific antibody production (Figures S1B,C in Supplementary Material). It also substantially reduced the population of IL-17-producing CD4⁺ T cells, considered to be Th17 cells, in the spleen and draining lymph nodes (Figure S1D in Supplementary Material).

CD4⁺ T cells in Fra1/Junb Tg mice had different proportions of Th1 and Th17 cells (Figure S2 in Supplementary Material). CIA was induced in Fra1/Junb transgenic mice to investigate the in vivo effects of Fra1 and Junb overexpression. Splenocytes were examined for HA-tagged FRA1/JUNB transgene by immunoblotting (Figure 5D). The incidence of arthritis and the arthritis score were significantly increased (Figures 5E,F). There was a marked increase in the levels of IL-17, FRA1, and JUNB in the spleen (Figure 5G). The relative mRNA level of the osteoclastogenic marker tartrate-resistant acid phosphatase was significantly higher (Figure 5H), while CII-specific T-cell proliferation and IgG production were also increased (Figures 5I–K). Overexpression of FRA1/JUNB profoundly increased CIA joint pathology (Figure S3A in Supplementary Material). Immunostaining showed that the expression of various proinflammatory cytokines and RANKL, a mediator of osteoclastogenesis, was significantly greater in the joint tissues of transgenic mice with CIA (Figure S3B in Supplementary Material). IL-17 mRNA levels and Th17 cell numbers were increased markedly (Figures S3C,D in Supplementary Material). These findings suggest that the FRA1–JUNB complex plays a major role in the development of CIA, likely through the induction of Th17 cell differentiation.

We demonstrated in vivo roles of FRA1 and JUNB in mice with CIA, an animal disease model for RA. To confirm the roles of FRA1 and JUNB in patients with RA, FRA1 was overexpressed in CD4⁺ T cells isolated from healthy controls. FRA1 overexpression increased the proportion of IL-17-producing CD4⁺ T cells during Th17 polarization (Figure 6A). These cells produced significantly higher levels of IL17a mRNA and cytokines (Figure 6B). Fra1 silencing decreased IL17a mRNA levels and IL-17 secretion (Figure 6C). There were significantly increased the transcription levels of Fra1 and Junb among peripheral blood mononuclear cells (PBMCs) and synovial fluid mononuclear cells (SFMCs) from RA patients compared with PBMCs from healthy controls (Figure 6D). Junb was also upregulated in RA patients. When Fra1 was silenced in CD4⁺ T cells isolated from SFMCs of RA patients, IL-17 secretion was decreased during Th17 polarization (Figure 6E). The expression of both FRA1 and JUNB was particularly high in the inflamed synovium of RA patients, a region enriched with inflammatory immune cells (Figure 6F). Of note, FRA1, JUNB, and phosphorylated STAT3 were colocalized in the inflamed synovium, which further supported our in vitro experimental results. A schematic illustration of the inflammatory signaling pathway induced by FRA1/JUNB is shown in Figure 6G. Th17 conditioning induces an inflammatory signal by promoting FRA1/JUNB. The inflammatory response is mediated by activation of STAT3, which regulates genes involved in Th17 cell differentiation. This cascade induces excessive inflammation and autoimmunity resulting in RA.