Identification of a novel microRNA-141-3p/Forkhead box C1/β-catenin axis associated with rheumatoid arthritis synovial fibroblast function in vivo and in vitro

Abstract

Rationale: Rheumatoid arthritis (RA) is a prototype of inflammatory arthritis in which synovial fibroblasts (SFs) play key roles in cartilage and bone destruction through tumor-like proliferation, migration, invasion and inflammation. This study aimed to research forkhead box protein C1 (FoxC1) and microRNA (miR)-141-3p, which modulate pathological changes in the synovial membrane, to find possible strategies for treating RA. Methods: FoxC1, β-catenin and miR-141-3p gene expression in synovial tissues and SFs was quantified by real-time PCR; FoxC1 and β-catenin protein levels were evaluated by immunohistochemistry, immunofluorescence, and Western blotting. We transiently transfected human SFs with FoxC1 and β-catenin overexpression and silencing vectors and assessed proliferation, migration, invasion and inflammation by cell function and enzyme-linked immunosorbent assays. We also assessed downstream signaling activation using immunofluorescence, real-time PCR and Western blotting. Double luciferase, coimmunoprecipitation and chromatin immunoprecipitation assays were used to verify miR-141-3p, FoxC1 and β-catenin gene and protein combinations. Finally, the therapeutic effects of FoxC1 silencing and miR-141-3p overexpression were evaluated in type II collagen-induced arthritis (CIA) rats. Results: We found that FoxC1 expression was significantly upregulated in synovium and SFs in both RA patients and rats with collagen-induced arthritis (CIA). FoxC1 overexpression increased β-catenin messenger RNA (mRNA) and protein levels and upregulated cyclin D1, c-Myc, fibronectin and matrix metalloproteinase 3 (MMP3) mRNA and protein expression in RA SFs (RASFs). In contrast, FoxC1 knockdown reduced β-catenin mRNA and protein levels as well as cyclin D1, c-Myc, and fibronectin mRNA and protein levels in RASFs. Furthermore, altering FoxC1 expression did not significantly change GSK3β and pGSK3β levels. FoxC1 overexpression promoted proliferation, migration, invasion and proinflammatory cytokine (interleukin (IL)-1β, IL-6, and tumor necrosis factor-α (TNF-α)) production and reduced anti-inflammatory cytokine (IL-10) levels in RASFs. FoxC1 bound to the β-catenin promoter, and β-catenin mediated the FoxC1-induced pathological changes. We also observed downregulated microRNA (miR)-141-3p expression in SFs from both RA patients and CIA rats and further found that miR-141-3p bound to the FoxC1 3'UTR and suppressed FoxC1 expression. Intra-ankle miR-141-3p agomir or FoxC1-specific siRNA injection hindered CIA development in rats. Conclusions: FoxC1 and miR-141-3p participate in RA pathogenesis by mediating inflammation and SF proliferation, migration, and invasion and thus could be novel targets for RA therapy as a nonimmunosuppressive approach.

Keywords: FoxC1; collagen-induced arthritis; miR-141-3p; rheumatoid arthritis; synovial fibroblasts.

Figure 1

FoxC1 and β-catenin are significantly upregulated in the synovium and SFs of RA patients and CIA rats. (A) The expression of FoxC1 and β-catenin in synovial tissues of RA patients (n=20) and CIA rats (n=6) was detected by immunohistochemistry, and corresponding control groups (control patients (n=10), normal rats (n=6)) were set up. Typical images and IHC scores were shown. Original magnification ×200, original magnification ×400. (B) The protein expression of FoxC1 and β-catenin in synovial tissues of RA patients (n=20) and CIA rats (n=6) was detected by WB, and corresponding control groups (control patients (n=10), normal rats (n=6)) were set up. (C) The mRNA expression of FoxC1 and β-catenin in synovial tissues of RA patients (n=20) and CIA rats (n=6) was detected by RT-PCR, and corresponding control groups (control patients (n=10), normal rats (n=6)) were set up. (D) The expression sites of FoxC1 and β-catenin in RASFs were observed by immunofluorescence and the corresponding control SFs (n=4) were set up. Original magnification ×200. (E) The protein and mRNA expression of FoxC1 and β-catenin in SFs of RA patients (n=6) were analyzed by western blotting and qRT-PCR and the corresponding control SFs (n=6) were set up. (F) The protein and mRNA expression of FoxC1 and β-catenin in SFs of CIA rats (n=6) were analyzed by western blotting and qRT-PCR and the corresponding normal SFs (n=6) were set up. Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001, t-test. Scale bars: 50 μm.

Figure 2

In RASFs, FoxC1 promotes proliferation, migration, and invasion of the cells, stimulates proinflammatory cytokine production and reduces anti-inflammatory cytokine levels. (A) The CCK-8 assay was used to evaluate cell proliferation after overexpression (LVFoxC1, n=4) or inhibition (SiFoxC1, n=4) of FoxC1 in RASFs, and corresponding control groups (n=8) were set up. (B) Wound healing through RASFs (LVFoxC1 (n=4), SiFoxC1 (n=4) and control (n=8)) migration was observed sequentially (0 h, 24 h and 48 h) by microscopy. Original magnification ×100. The black lines were used to mark the general area of the wound. Wound area was quantified by Image J. (C) Invasion of RASFs (LVFoxC1 (n=3), SiFoxC1 (n=3) and control (n=6)) was determined by transwell assays (48 h). Original magnification ×200. (D) Inflammatory cytokines concentration in culture supernatant of RASFs (LVFoxC1 (n=4), SiFoxC1 (n=4) and control (n=8)) was measured by ELISA. Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, t-test. Scale bars: 100 μm.

Figure 3

FoxC1 induces pathological changes in RASFs through β-Catenin. (A) WB and qRT-PCR methods were used to detect protein and gene changes in FoxC1, β-catenin, GSK3β, cyclin D1, c-Myc, fibronectin, and MMP3 in FoxC1 inhibited RASFs (SiFoxC1, n=6), and corresponding control group (n=6) was set. (B) WB and qRT-PCR methods were used to detect protein and gene changes in FoxC1, β-catenin, GSK3β, cyclin D1, c-Myc, fibronectin, and MMP3 in FoxC1 overexpressed RASFs (LVFoxC1, n=6), and corresponding control group (n=6) was set. (C) WB and qRT-PCR methods were used to detect protein and gene changes in β-catenin, cyclin D1, c-Myc, fibronectin, and MMP3 in overexpression β-catenin SiFoxC1-RASFs (n=5), and corresponding control group (n=5) was set. (D) WB and qRT-PCR methods were used to detect protein and gene changes in β-catenin, cyclin D1, c-Myc, fibronectin, and MMP3 in β-catenin-knockdown LVFoxC1-RASFs (n=5), and corresponding control group (n=5) was set. Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001, t-test.

Figure 4

FoxC1 induces pathological changes in RASFs through β-Catenin. (A) The CCK-8 assay was used to evaluate cell proliferation after overexpression β-catenin SiFoxC1-RASFs (n=4), β-catenin- knockdown LVFoxC1-RASFs (n=4) in RASFs, and corresponding control groups (n=8) were set up. (B) Wound healing through RASFs (overexpression β-catenin SiFoxC1-RASFs (n=4), β-catenin-knockdown LVFoxC1-RASFs (n=4) and corresponding control groups (n=8) migration were observed sequentially (0 h, 24 h and 48 h) by microscopy. Original magnification ×100. The black lines were used to mark the general area of the wound. (C) Invasion of RASFs (overexpression β-catenin SiFoxC1-RASFs (n=3), β-catenin-knockdown LVFoxC1-RASFs (n=3) and corresponding control groups (n=6)) was determined by transwell assays (48 h). Original magnification ×200. (D) Inflammatory cytokines concentration in culture supernatant of RASFs (overexpression β-catenin SiFoxC1-RASFs (n=4), β-catenin-knockdown LVFoxC1-RASFs (n=4) and corresponding control groups (n=8)) was measured by ELISA. Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, t-test. Scale bars: 100 μm.

Figure 5

FoxC1 binds to the β-catenin promoter and induces pathological changes in RASFs. (A)The co-immunoprecipitation assay was used to detect the interaction of FoxC1 and β-catenin in RASFs (n=3). (B) Luciferase reporter assays were used to prove that the relationship between FoxC1 and β-catenin promoter in RASFs (n=3). (C) ChIP-qRT-PCR analysis confirmed the immunoprecipitation of FoxC1 and BS1 (n=3). Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, t-test, one-way ANOVA.

Figure 6

MiR-141-3p binds to the FoxC1 3′UTR and influences the regulation of β-catenin by FoxC1. (A) The expression of miRNAs in synovial tissues of RA patients (n=20) and RASFs (n=6) was detected by qRT-PCR, and corresponding control groups (control patients synovial tissue (n=10), control SFs (n=6)) were set up. (B) Dual-luciferase assay confirmed the binding of MiR-141-3p to FoxC1 3′UTR (n=3). (C) WB and qRT-PCR methods were used to detect protein and gene changes in FoxC1, β-catenin in overexpression miR-141-3p RASFs (n=5), miR-141-3p knockdown RASFs (n=5), and corresponding control groups (n=10) were set. (D) WB and qRT-PCR methods were used to detect protein and gene changes in FoxC1, β-catenin in overexpression miR-141-3p LVFoxC1-RASFs (n=5), miR-141-3p knockdown SiFoxC1-RASFs (n=5), and corresponding control groups (n=10) were set. Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001, t-test, one-way ANOVA.

Figure 7

Role of the miR-141-3p/FoxC1/β-catenin axis in LPS-induced inflammatory SFs. (A) qRT-PCR was used to detect gene changes in miR-141-3p in LPS-induced SFs (n=6), and the corresponding control group (n=6) was set. (B) WB and qRT-PCR methods were used to detect protein and gene changes in FoxC1, β-catenin, cyclin D1, c-Myc, fibronectin, and MMP3 in LPS-induced SFs (n=6), and corresponding control group (n=6) was set. (C-D) WB and qRT-PCR methods were used to detect protein and gene changes in FoxC1, β-catenin, cyclin D1, c-Myc, fibronectin, and MMP3 in LPS-induced SFs (n=5), overexpression miR-141-3p LPS-induced SFs (n=5), FoxC1 knockdown LPS-induced SFs (n=5). (E) ELISA assay was used to evaluate the expression of inflammatory factors in SFs medium (n=4). Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001, t-test.

Figure 8

FoxC1 binds to the β-catenin promoter and also had a significantly stronger effect in LPS-induced SFs. ChIP and real-time PCR assays were used to measure FoxC1 bound to the β-catenin promoter in LPS-induced SFs (n=3), overexpression miR-141-3p LPS-induced SFs (n=3) and FoxC1 knockdown LPS-induced SFs (n=3). Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001, t-test

Figure 9

Intra-ankle injection of a miR-141-3p agomir /FoxC1-specific siRNA hinders CIA development in rats. (A) In vivo imaging results showed that miR-141-3p agomir and FoxC1-specific siRNA could be expressed in the ankle joint of rats for at least 7 days. (B) Clinical arthritis score and ankle joint swelling in the arthritic rats were evaluated (n=6). (C) Micro-CT was used to measure bone destruction and bone density (n=6). The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001,****p<0.0001 versus the normal group, ### p<0.001 versus the model group. one-way ANOVA.

Figure 10

Morphology and iconography study of rat treatment by miR-141-3p agomir/FoxC1-specific siRNA. The bone trabecular of the left calcaneus was chosen as region of interest (ROI). Micro-CT was used to measure ankle bone parameters. The data were expressed as mean ± SD. *p<0.05, **p<0.01, ***p<0.001 versus the normal group, #p<0.05, ##p<0.01, ###p<0.001 versus the model group. One-way ANOVA.

Figure 11

Expression levels of FoxC1 and β-catenin after injection of miR-141-3p agomir/FoxC1-specific siRNA in CIA rats. Immunohistochemistry and IHC scores were used to evaluate the expression of FoxC1 and β-catenin after injection of miR-141-3p agomir/FoxC1-specific siRNA. Original magnification ×200, original magnification ×400. Experiments were independently repeated three times. The data were expressed as mean ± SD. *p<0.05, **p<0.01, **** p<0.0001 versus the normal group, ##p<0.01 versus the model group. One-way ANOVA. Scale bars: 50μm.

Figure 12

Intra-ankle injection of a miR-141-3p agomir /FoxC1-specific siRNA hinders CIA development in rats. (A) HE staining and synovial inflammation score were used to assess synovial proliferation and inflammation (n=6). The black arrow was used to mark the synovial position of the ankle. (B) The degree of cartilage damage in the ankle was evaluated by Safranin O/Fast Green staining and cartilage erosion score (n=6). The arrows were used to mark the cartilage position. Black arrows were used to mark the cartilage surface of the ankle. All values were expressed as mean ± SD. **p<0.01, ****p<0.0001 versus the normal group, #p<0.05, ##p<0.01 versus the model group. one-way ANOVA. Original magnification ×20. Scale bars: 500 μm.

Figure 13

IL-1β and IL-6 expression in CIA rats after treatment by miR-141-3p agomir/FoxC1-specific siRNA. Immunohistochemistry and IHC scores were used to evaluate the expression of IL-1β and IL-6 after injection of miR-141-3p agomir/FoxC1-specific siRNA. Original magnification ×200, original magnification ×400. Experiments were independently repeated three times. The data were expressed as mean ± SD. ***p<0.001, **** p<0.0001 versus the normal group, ###p<0.001, ####p<0.0001 versus the model group. One-way ANOVA. Scale bars: 50μm.

Figure 14

TNF-α and IL-10 expression in CIA rats after treatment by FoxC1-specific siRNA/miR-141-3p agomir. Immunohistochemistry and IHC scores were used to evaluate the expression of TNF-α and IL-10 after the injection of miR-141-3p agomir/FoxC1-specific siRNA. Original magnification ×200, original magnification ×400. Experiments were independently repeated three times. The data were expressed as mean ± SD. ***p<0.001, **** p<0.0001 versus the normal group, ###p<0.001, ####p<0.0001 versus the model group. One-way ANOVA. Scale bars: 50μm.