Circular RNA circFBXW4 suppresses hepatic fibrosis via targeting the miR-18b-3p/FBXW7 axis

Abstract

Rationale: Circular RNAs (circRNAs) are a new form of noncoding RNAs that play crucial roles in various pathological processes. However, the expression profile and function of circRNAs in hepatic fibrosis (HF) remain largely unknown. In this study, we show a novel circFBXW4 mediates HF via targeting the miR-18b-3p/FBXW7 axis. Methods: We investigated the expression profile of circRNAs, microRNAs and mRNAs in hepatic stellate cells (HSCs) from HF progression and regression mice by circRNAs-seq and microarray analysis. We found a significantly dysregulated circFBXW4 in HF. Loss-of-function and gain-of-function analysis of circFBXW4 were performed to assess the role of circFBXW4 in HF. Furthermore, we confirmed that circFBXW4 directly binds to miR-18b-3p by luciferase reporter assay, RNA pull down and fluorescence in situ hybridization analysis. Results: We found that circFBXW4 downregulated in liver fibrogenesis. Enforcing the expression of circFBXW4 inhibited HSCs activation, proliferation and induced apoptosis, attenuated mouse liver fibrogenesis injury and showed anti-inflammation effect. Mechanistically, circFBXW4 directly targeted to miR-18b-3p to regulate the expression of FBXW7 in HF. Conclusions: circFBXW4 may act as a suppressor of HSCs activation and HF through the circFBXW4/miR-18b-3p/FBXW7 axis. Our findings identify that circFBXW4 serves as a potential biomarker for HF therapy.

Keywords: FBXW7; circFBXW4; circular RNAs; hepatic fibrosis; miR-18b-3p.

Figure 1

circRNAs profiling in HSCs from vehicle, HF and HF recovery mice. (A) The procedure of mouse acute toxic liver injury model, HF model and HF recovery model used for animal. (B) Pathology observation of H&E and sirius red staining of liver tissues, scale bar, 100 μM; immunohistochemistry staining for α-SMA in liver tissues, scale bar, 100 μM; immunofluorescence staining of α-SMA and collagens I in liver tissues, nuclei were stained with DAPI, scale bar, 200 μM. The positive staining areas were measured by Ipwin32 software. (C) Protein expression of α-SMA and collagens I in HSCs detected by western blot. (D) mRNA levels of α-SMA, PDGF, TIMP-1 and collagens I in HSCs detected by qRT-PCR. (E) A heatmap showed differentially expressed circRNAs with same expression pattern in vehicle and HF recovery mice, which opposite to HF mice. (F) circRNA composition in terms of genes distribution. (G, H) The coverage and distribution of differentially expressed circRNAs on the mouse chromosomes. Data represent the mean ± s.e.m (n=8). *p<0.05, **p<0.01 versus Oil; ^#p<0.05, ^##p<0.01 versus CCl₄.

Figure 2

Expression of circFBXW4 decreased in HF and rescued in HF recovery mice. (A) Eight differentially expressed circRNAs confirmed by qRT-PCR in HSCs (n=10-12). (B) Pathology observation of H&E staining from acute toxic liver injury tissue, scale bar, 100 μM; immunohistochemistry staining for α-SMA, scale bar, 100 μM. The positive staining areas were measured by Ipwin32 software. (C) mRNA level of TGF-β1 and (D) relative expression of circFBXW4 in HSCs (n=6). (E) Relative expression of circFBXW4 in HSCs and (F) in serum. (G) mRNA level of FBXW4 in HSCs. (H) Expression pattern of circFBXW4 in HF progression and regression stages. (I) Pathology observation of H&E and masson staining of human liver tissue, scale bar, 100 μM; immunohistochemistry staining for α-SMA, scale bar, 100 μM. The positive staining areas were measured by Ipwin32 software. (J) Relative expression of circFBXW4, α-SMA, TIMP-1 and collagens I in human liver tissue detected by qRT-PCR (Control, n=10; Fibrosis, n=14). Data represent the mean ± s.e.m. *p<0.05, **p<0.01 versus Oil; ^#p<0.05, ^##p<0.01 versus CCl₄; ^&p<0.05, ^&&p<0.01 versus Control.

Figure 3

Characterizations of circFBXW4. (A) Schematic diagram showed the genomic location and back-splicing pattern of circFBXW4. (B) Sanger sequencing of circFBXW4, the arrow represents the “head-to-tail” splicing site. (C) Divergent primers amplified circFBXW4 from cDNA by PCR and an agarose gel electrophoresis, rather than from gDNA, GAPDH was used as a linear control (n=3). (D) circFBXW4, rather than linear FBXW4 or GAPDH, resisted to Rnase R digestion (n=6). (E) circFBXW4 from cDNA was detected with divergent primers even exposed to Rnase R digestion, the opposite result showed from gDNA (n=3). (F) Relative expression of circFBXW4 and linear FBXW4 in HSCs treated with ActD+ (actinomycin D) at the indicated time points (n=6). (G) Cytoplasmic and nuclear fractionation assay revealed that circFBXW4 mainly localized in the cytoplasm (n=4). Data represent the mean ± s.e.m. ^&p<0.05, ^&&p<0.01 versus Rnase R- group; *p<0.05, **p<0.01 versus time 0.

Figure 4

Effects of circFBXW4 overexpression on the activation, proliferation and apoptosis of LX-2 cells. (A) Stable overexpression efficiency of circFBXW4, and mRNA levels of FBXW4, TGF-β1, TIMP-1, α-SMA and collagens I detected by qRT-PCR. (B) Protein expression of α-SMA and collagens I in LX-2 cells (n=4). (C) Immunofluorescence staining for α-SMA, nuclei were stained with DAPI, scale bar, 50 μM. (D) Effect of circFBXW4 overexpression on the proliferation of LX-2 cells detected by CCK8 assay at the indicated time points (n=6). (E) Assessment of DNA synthesis using Edu assay in LX-2 cells, nuclei were stained with Hoechst33342, scale bar, 200μM. (F) Flow cytometry showed the cell cycle distribution in G₀/G₁ phase increased following overexpression of circFBXW4. (G) Overexpression of circFBXW4 induced apoptosis of LX-2 cells with indicated treatment. Data represent the mean ± s.e.m. *p<0.05, **p<0.01 versus Control; ^#p<0.05, ^##p<0.01 versus LV-vector.

Figure 5

Anti-fibrotic effects of circFBXW4 in HF mice. (A) Mouse in vivo imaging analysis showed luciferase-labelled pHBAAV-circFBXW4 was specific located in mouse liver tissue. (B) Overexpression efficiency of circFBXW4 in HSCs from mice following pHBAAV-circFBXW4 administration. (C) Pathology observation of H&E staining, sirius red staining and immunohistochemistry staining for α-SMA in liver tissue, scale bar, 100 μM. The positive staining areas were measured by Ipwin32 software. (D) Protein expression of α-SMA and collagens I in HSCs from mice after circFBXW4 overexpression. (E, F) Test of HYP in mouse liver tissue and ALT in serum. (G) mRNA levels of TGF-β1, α-SMA, collagens I and TIMP-1 in HSCs from mice following pHBAAV-circFBXW4 treatment. Data represent the mean ± s.e.m (n=8). ^$p<0.05, ^$$p<0.01 versus Saline; *p<0.05, **p<0.01 versus Oil; ^#p<0.05, ^##p<0.01 versus CCl₄.

Figure 6

Anti-inflammation of circFBXW4 in HF mice. (A) Expression pattern of circFBXW4 in liver macrophages from vehicle and HF mice. (B) pHBAAV-circFBXW4 successful elevated circFBXW4 level in liver macrophages. (C) Levels of pro-inflammatory cytokines TNF-α and IL-1β in serum detected by ELISA. (D) mRNA levels of CCL2 and CCL9 in liver macrophages (n=8). (E) The up panel is a schematic representation of mouse inflammation antibody array containing 40 inflammation antibodies, positive control, negative control and blank spots with duplicates. Array showed the expression of inflammatory factors in mouse liver lysates, fold changes of differentially expressed 5 inflammatory factors (CCL2, CD153, IL-12 p70, TNF-RII, CCL9) were showed in the down panel. (F) Immunohistochemistry staining for F4/80 in liver tissue, scale bar, 50 μM. The positive staining areas were measured by Ipwin32 software. (G) Flow cytometric analysis showed the reduction of F4/80⁺CD11b⁺ liver macrophages in HF mice with pHBAAV-circFBXW4 treatment (n=4). Data represent the mean ± s.e.m. ^$p<0.05, ^$$p<0.01 versus Saline; *p<0.05, **p<0.01 versus Oil; ^#p<0.05, ^##p<0.01 versus CCl₄.

Figure 7

Microarray analysis of HSCs and circFBXW4 binds miR-18b-3p. (A) Heatmap of differentially expressed miRNAs in HSCs from vehicle, HF and HF recovery mice. (B) circRNA-miRNA network showed differentially expressed miRNAs with their circRNA targets. (C) Differentially expressed miRNAs combine with circFBXW4 predicted miRNA targets, and a schematic showed the binding sites of miRNAs bind to circFBXW4. (D) The renilla luciferase activity analysis of wild type or mutant circFBXW4 and miRNAs mimics or negative control co-transfected into HEK-293T cells, respectively. Data are presented as the ratio of renilla luciferase activity to firefly activity (n=3). (E) Biotinylated circFBXW4 probe could capture 3 miRNA candidates by RNA pull-down analysis (n=4). (F) Schematic of miR-18b-3p sites in circFBXW4 based on complementary sequences. (G) Co-localization of circFBXW4 with miR-18b-3p was determined in mouse liver tissue by FISH, nuclei were stained with DAPI, scale bar, 200 μM and 100 μM. (H) Level of miR-18b-3p was suppressed in HSCs following circFBXW4 overexpression (n=7). Data represent the mean ± s.e.m. *p<0.05, **p<0.01 versus Negative mimics+circFBXW4-wt; ^#p<0.05, ^##p<0.01 versus miRNA-mimics+circFBXW4-wt; ^&p<0.05, ^&&p<0.01 versus circFBXW4 probe; ^$p<0.05, ^$$p<0.01 versus Oil; ^@p<0.05, ^@@p<0.01 versus CCl₄.

Figure 8

Identification of circFBXW4/miR-18b-3p/FBXW7 axis. (A) Scatter plots showed differentially expressed mRNAs in HSCs. (B) circRNAs-micorRNAs-mRNAs interaction network showed negative regulatory relationships combine the differentially expressed miRNAs with their differentially expressed circRNAs and mRNAs targets. (C, D) GO and KEGG enrichment analysis of differentially expressed genes in HSCs. (E) Binding sites of miR-18b-3p and 3'UTR of FBXW7. (F) Overexpression efficiency of miR-18b-3p in HSCs with miR-18b-3p-Up-agomir treatment. (G) Immunohistochemistry staining for FBXW7 and Yap1 in liver tissue, scale bar, 50 μM; Double-immunofluorescence staining showed co-localization of FBXW7 with α-SMA in liver tissue, nuclei were stained with DAPI, scale bar, 200 μM and 100 μM (up). (H, I) Protein expression and mRNA levels of FBXW7 and Yap1 in HSCs following circFBXW4 overexpression and miR-18b-3p-Up-agomir treatment. Data represent the mean ± s.e.m (n=6). ^$p<0.05, ^$$p<0.01 versus Control; *p<0.05, **p<0.01 versus Oil; ^#p<0.05, ^##p<0.01 versus CCl₄; ^&p<0.05, ^&&p<0.01 versus CCl₄+pHBAAV-circFBXW4.