The miRNAs are small, non-coding RNAs that regulate various biological processes, including liver fibrosis. Hepatic stellate cells (HSCs) play a central role in the pathogenesis of liver fibrosis. By microarray profiling and real-time PCR, we noted that miR-31 expression in HSCs from rats, mice and humans was significantly increased during HSC activation in culture. Overall, miR-31 expression levels were unchanged in the whole-liver RNA extracts from fibrotic rat and human samples. Nevertheless, we found that miR-31 was particularly up-regulated in HSCs but not in hepatocytes during fibrogenesis. Thus, we hypothesized that miR-31 may mediate liver fibrosis. In the present study, we found that inhibition of miR-31 expression significantly inhibited HSC activation, whereas its over-expression obviously promoted HSC activation. Moreover, over-expression of miR-31 promoted HSC migration by enhancing matrix metalloproteinase (MMP)-2 expression whereas inhibition of miR-31 has an opposite effect. The biological function of miR-31 during HSC activation might be through targeting FIH1, a suppressor of hypoxia-inducible factor (HIF-1), because a knockdown of FIH1 by shRNA could mimic the effects of miR-31. In addition, primary rat HSCs were isolated and treated with different cytokines, such as transforming growth factor β (TGF-β), vascular endothelial growth factor and platelet-derived growth factor-BB, to evaluate upstream regulators of miR-31. We found that only TGF-β, a pivotal regulator in liver fibrosis, remarkably increased miR-31 expression in HSCs. And the effects of TGF-β on HSCs can be partially counteracted by inhibition of miR-31. In addition, chromatin immunoprecipitation experiments and the luciferase reporter assay demonstrated that Smad3, a major TGF-β-downstream transcription factor, stimulated the transcription activity of miR-31 by binding directly to miR-31's promoter. In conclusion, the miR-31/FIH1 pathway associates with liver fibrosis, perhaps by participation in the TGF-β/Smad3 signalling of HSCs.
HSC activation is a key element in the pathogenesis of liver fibrosis. miR-31 has recently emerged as an important regulator in fibrosis disease. The aim of the present study was to determine its role in liver fibrosis.
Levels of miR-31 was significantly increased in human primary HSCs during fibrogenesis. In our study, we found miR-31 may contribute to HSC activation and may thereby play an important role in liver fibrosis.
In addition, inhibition of miR-31 with antagonist may be benefit for this disorder in the future.
Liver fibrosis is a scarring process that is associated with an increased and altered deposition of extracellular matrix (ECM) in the liver . Over the past decades, hepatic stellate cells (HSCs) have been commonly accepted as the main cell type responsible for ECM formation in response to accumulated levels of inflammatory signals . HSCs remain in a quiescent state under physiological conditions, but, on liver injury, they differentiate into myofibroblast-like cells marked by expression of smooth muscle α-actin (α-SMA), loss of lipid droplets, and increased proliferation, contractility and migration. Activated HSCs secrete a variety of profibrogenic cytokines, among which transforming growth factor (TGF)-β is deemed to be the most potent fibrogenic cytokine regulating HSC collagen production, differentiation, etc. . The TGF-β signal is propagated by phosphorylating Smad2/3 (R-Smad), which binds with Smad4 (co-Smad) to form multimers. Then they are transported into the cell nucleus to affect the target gene via direct DNA binding or indirect association .
The miRNAs are short non-coding RNAs of about 22 nucleotides that have recently been shown to play important roles in mammalian gene expression [5–7]. In recent years, massive progress has been made in identifying miRNAs as important regulators of gene expression in various hepatic diseases, such as acute liver injury, viral hepatitis or hepatocellular carcinoma [8,9]. The function of miRNAs in liver fibrosis had also been studied. Forced expression of miR-221/222 promotes the activation of HSCs and the progression of liver fibrosis . Over-expression of miR-214 in HSCs increased the expression of fibrosis-related genes, matrix metalloproteinase (MMP)-2, MMP-9 and α-SMA . The miRNAs miR-150 and miR-194 may suppress the fibrotic phenotype and inhibit the production of ECM . Interferon induced miR-195 to inhibit HSC proliferation by delaying cell cycle progression in the G1 to S phase . Additional studies by Ji et al.  reported the role and mechanism of miR-27 in regulating fat metabolism and cell proliferation. Although the research area keeps on developing, few studies have identified that miRNAs have a global effect on fibrogenic TGF-β signalling in the liver.
Previous studies showed that miR-31 was emerging as an important regulator in carcinoma [15–18]. In the present study, we report that miR-31, a differentially expressed miRNA during fibrogenesis, significantly promotes HSC activation and migration. Specifically, we validated miR-31 function in HSCs by complementarily binding to the 3′-UTR of FIH1. Then, we found that miR-31 is induced by TGF-β1 and mediates TGF-β-induced cell activation and migration by down-regulating FIH1 expression. Moreover, this study indicated that TGF-β1 may act by stimulating Smad3 to regulate miR-31 expression during liver fibrosis. These findings may identify the miR-31/FIH1 nexus as a novel regulator of TGF-β signalling in HSC-mediated liver fibrosis, and suggest a potential clinical diagnostic and therapeutic approach in the future.
MATERIALS AND METHODS
Rat models of liver fibrosis
Male rats (200–220 g) received 2 ml/kg of body weight of carbon tetrachloride mixed with olive oil (40%CCl4) intraperitoneally, and the first dose was doubled. To induce liver fibrosis, carbon tetrachloride was injected twice a week for 8 weeks. Control animals were injected with an equal volume of olive oil at the same time intervals. Rats, acquired from the National Resource Centre for Rodent Laboratory Animals of China, were sacrificed 48 h after the last injection. The experimental animal procedures were reviewed and approved by the university's Animal Care and Use Committee.
Human tissue samples
We studied 18 patients with liver fibrosis. The underlying disease aetiology was chronic hepatitis B virus infection. Whole-liver RNA was extracted from the liver biopsy specimens. We obtained specimens from the General Hospital of the Fuzhou Military Region. Informed consent for the research was obtained before biopsy. Each sample was divided into two parts: one was sent for the evaluation of fibrosis severity and the other used to extract total RNA. Fibrosis staging was based on the MERAVIR scoring system (F0, no fibrosis; F1, portal fibrosis without septa; F2, portal fibrosis with rare septa; F3, numerous septa without cirrhosis; and F4, cirrhosis).
Isolation of primary cells and culture
Primary rat and mouse HSCs were isolated from the liver through two steps of digestion as described previously . Cell viability was determined by the Trypan Blue exclusion method. The purity was assessed by autofluorescence and immunostaining using anti-desmin antibody (1:100, Bioworld). Moreover, the HSC phenotype was assessed by immunofluorescence staining using antibodies for α-SMA (1:100, Sigma-Aldrich). Primary human HSCs were isolated from samples obtained from patients in the Changzheng Hospital who had undergone partial liver resection. The methods adopted were those previously described . HSCs were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS. Experimental procedures were performed according to the guidelines of the local ethics committee with patients’ informed consent. Human embryonic kidney (HEK) 293T cells were purchased from the American Type Culture Collection (ATCC). The hepatic stellate cell line (HSC-T6) was purchased from the Cell Bank at the Shanghai Institutes for Biological Sciences of the Chinese Academy of Sciences. HSC-T6 and 293T cells were cultured in DMEM. All cells were maintained at 37°C with 5% CO2.
To knock down FIH1, we designed three shRNAs (shFIH1-1, -2 and -3) targeting different regions of FIH1 and cloned into the lentiviral vector pLKO.1. The shRNA oligonucleotides used are as follows: shFIH1-1: 5′-GAGAATTGAATATCCTCTCAA-3′; shFIH1-2: 5′-CAACGGAGATTTCTCTGTGTA-3′; and shFIH1-3: 5′-CCTGCAAGAGAATATTGGCAA-3′. For luciferase reporter assays, the 3′-UTR segment of wild-type FIH1 mRNA, which possessed the target sites of miR-31, was amplified and cloned into the pGL3. The mutation of the 3′-UTR segment was performed using the QuickChange Lighting Site-Directed Mutagenesis Kit (Agilent Technologies Cat#210518, Palo Alto, CA, http:://www.agilent.com). The FIH1 expression vector was obtained by cloning the FIH1-coding sequence into the AgeI and EcoRI sites for FUGW.
To knock down Smad3, we designed three shRNAs targeting different regions of Smad3. The shRNA oligonucleotides used are as follows: shSmad3,1: 5′-GCACACAATAACTTGGACCTA-3′; shSmad3,2: 5′-CATCTCCTACTACGAGCTGAA-3′; and shSmad3,3: 5′-TCCGCATGAGCTTCGTCAAAG-3′. We co-transfected these three plasmid mixes with a helper plasmid encoding gag, pol and env into HEK293T cells to generate lentivirus particles, and virus was obtained 72 h after transfection.
HIF-1-Luc reporter plasmid (GenomeDitech) was used for the assay of hypoxia-inducible factor (HIF) transactivation. The miR-31 lentivirus vector was bought from Shanghai SBO Medical Biotechnology Co., Ltd.
Transient plasmid transfection
The miR-31 precursor and precursor control were purchased from Biotend, Shanghai, China; the miR-31 inhibitor and inhibitor control were purchased from RiboBio Co. The HSCs, 1.5×105 suspended per well in six-well plates, were transfected with precursor or inhibitors by using the GeneExpresso Max Transfection Reagent (Excellgen), according to the manufacturer's instructions. At 48–72 h post-transfection, the cells were harvested for real-time (RT)-PCR and Western blotting.
RT-PCR for miRNA and mRNA expression
To detect the expression of miRNAs, 2 μg of total RNA was used to synthesize cDNA by reverse transcription using TIANScript RT kit [Tiangen Biotech (Beijing) Co. Ltd]. Expression values were normalized to the control endogenous small RNA U6. The Bulge-Loop miRNA qPCR Primer Set (RiboBio Co., Cat#MQP-0101) was used to detect the expression of miRNAs by quantitative (q)RT-PCR assay with the SYBR Green qPCR Master Mix [Tiangen Biotech (Beijing) Co. Ltd]. To confirm mRNA expression, 500 ng of total RNA was used to synthesize cDNA by using PrimeScript RT Master Mix (Takara Bio). Primer sequences were listed as follows: glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (forward: 5′-CAGTGCCAGCCTCGTCTCAT-3′; reverse: 5′-AGGGGCCATCCACAGTCTTC-3′); collagen I (ColI) (forward: 5′-ATCCTGCCGATGTCGCTAT-3′, reverse: 5′-CCACAAGCGTGCTGTAGGT-3′); α-SMA (forward: 5′-CC-GAGATCTCACCGACTACC-3′, reverse: 5′-TCCAGAGCGAC-ATAGCACAG-3′); and FIH1 (forward: 5′-CCGTGGGT-AGAAAGATTGTCA-3′, reverse: 5′-TGATGGACAGGGTATG-GGTAG-3′).
Cells were lysed in SDS sample buffer. Proteins were separated by SDS/PAGE, transferred to membranes and detected with appropriate antibodies as described below. Antibodies used in the experiment as loading controls were: anti-collagen type Ι (Abcam, 1:2500), anti-α-SMA (Sigma-Aldrich, 1:1000), anti-Smad3 (Upstate, 1:1000), anti-FIH1 (Epitomics, 1:1000), anti-HIF-1α (BioWorld, 1:1000), anti-MMP-2 (Sigma-Aldrich, 1:1000), anti-MMP-9 (Epitomics, 1:1000) and anti-GAPDH (BioWorld, 1:5000). The signals were visualized using ImageQuant LAS 4000 Mini (GE Healthcare Life Sciences).
Cell migration assay
The cell migration assay was carried out using transwell inserts with 8.0-μm pores (Corning Life Sciences) in 24-well plates. To simulate the perisinusoidal space, we coated the upper side with Matrigel and the opposite side with ColI. A total of 4×104 cells were resuspended in serum-free medium and seeded in the upper side of the insert. Then the chambers were incubated for 26 h. Later, we removed the cells that did not migrate through the pores with cotton swabs. Cells on the lower side of the insert filter were quickly fixed by 4% paraformaldehyde for 10 min and stained using Hoechst 33342 (Life Technologies). The total number of invading cells were acquired in four representative fields using fluorescence microscopy.
Growth curve, cell cycle and apoptosis assay
Two days after transfection, cells were transferred into 96-well plates at a density of 3×103 cells/well. Cell counts were taken after periods of 24, 48 and 72 h. Cultured cells were incubated with a Cell Counting Kit-8 (CCK-8) solution (Dojindo) for 1 h at 37°C and the absorbance at 450 nm was measured with a spectrophotometer. Flow cytometric analysis was used for quantification of cell cycle distribution and apoptosis
Luciferase reporter assay
The FIH1 3′-UTR segment containing the target site for miR-31 was amplified by PCR from genomic DNA and inserted into the PGL3 control vector. We used 293FT cells to detect the relationship between miR-31 and FIH1. The cells were co-transfected in 24-well plates using FuGENE (Roche) according to the manufacturer's protocol with 300 ng of the firefly luciferase report vector and 10 ng of the control vector containing Renilla luciferase. For each well, 0.3 nM pre-miR-31 (Biotend) or pre-control miRNA was used. Firefly and Renilla luciferase activities were measured consecutively using the Dual-Luciferase Reporter Assay System (Promega) 48 h after transfection
For miR-31 promoter-binding activities, HSCs were infected with Repo lentivirus particles designed for Smad3. The miR-31 promoter reporters of lentiviral constructs, which drive the expression of firefly luciferase, were generated by cloning a PCR-amplified cassette containing wild-type or mutated Smad3-binding sites from the miR-31 promoter. Virus particles were collected after the pGMLV-miR-31-Luc or pGMLV-Luc vectors had been co-transfected with the vectors PAX2 and VSVG into HEK293T cells. HSCs plated in 24-well plates at a density of 5×104 cells/ml in culture medium were infected with the miR-31 promoter lentiviral firefly reporter (wild-type or Smad3-binding site mutant) and lentiviral Renilla. After infection for 3 days, the cells were stimulated with TGF-β1 at different concentrations for another 24 h. Cells were harvested to measure dual luciferase activities using the luciferase assay system (GenomeDitech).
We performed chromatin immunoprecipitation (ChIP) assays using a modified kit from Millipore. We fixed cells (5×106) with 1% formaldehyde, sonicated the chromatin preparations and pre-cleared the samples by incubation with protein A/G agarose and salmon sperm DNA (Millipore). We immunoprecipitated the chromatin by incubation with 10 μg of anti-Smad3-specific antibody or rabbit IgG monoclonal antibody at 4°C overnight, followed by incubation with protein A/G agarose and salmon sperm DNA for 1 h. We then renatured the immunoprecipitates and purified the DNA. We quantified the amount of immunoprecipitated DNA by RT-PCR with the ABI PRISM 7500 Sequence Detection System (Applied Biosystems) using SYBR Green with the forward primer 5′-GCCCAGCCATCTTGACTGC-3′ and the reverse primer 5′-GTGACTCCGCCGGACAGAG-3′.
miR-31 is particularly up-regulated in HSCs in liver fibrosis
Primary rat HSCs at days 2, 7 and 14 were collected to evaluate miR-31 expression during HSC activation. As our previous study showed, HSCs cultured at day 2 were deemed to be quiescent, at day 7 to be partially activated and at day 14 to be fully activated . The expression of miR-31 was significantly increased at day 14 during HSC activation, validated by RT-PCR, consistent with our previous results for miRNA microarray. Similar phenomena were also observed in primary mouse HSCs and primary human HSCs (Figures 1A, 1B and 1C). For further validation of the alteration of miR-31 in fibrogenesis, we initially detected the expression of miR-31 in an animal cirrhotic model and human cirrhotic samples. Whole miRNAs were extracted from livers of human cirrhotic samples and Sprague–Dawley rats which were injected with carbon tetrachloride for 8 weeks. Two fibrosis markers, ColI and α-SMA, were also significantly increased in carbon tetrachloride model rats (Figure 1G). Unexpectedly, there was no significant change in miR-31 expression in liver tissue during liver fibrosis in either a carbon tetrachloride-induced hepatic fibrosis animal model or human cirrhotic samples compared with the controls (Figures 1D and 1E).
miR-31 is particularly up-regulated in HSCs in liver fibrosis
We then hypothesized that miR-31 may, in particular, be up-regulated in HSCs during liver fibrogenesis. Primary activated HSCs and hepatocytes were isolated from the carbon tetrachloride-induced hepatic fibrosis rat model. It is of interest that our results showed that miR-31 expression was increased in primary activated HSCs from the hepatic fibrosis rat model compared with primary quiescent HSCs from normal rats (Figure 1F, third and fourth columns). There was a bit of a decline in miR-31 expression in hepatocytes isolated from the hepatic fibrosis rat model compared with normal rats (Figure 1F, first and second columns). The expression of miR-31 in the hepatocytes was higher than that in HSCs in the carbon tetrachloride-induced hepatic fibrosis rat model (Figure 1F, fifth and sixth columns). In short, our results suggested that miR-31, in particular, was increased in activated HSCs during liver fibrosis.
Promotion of HSC activation by
As we discovered, miR-31 was significantly increased during the process of activation; we used the lentivirus vector of knockdown (KD)-miR-31 to knock down its expression at the early stage of primary HSCs (day 2). On day 14, the KD-miR-31 cells showed a more irregular shape with more peripheral protrusions (Figure 2A). The total cell RNA was extracted and some fibrosis-related genes were detected. In our study, we found that, compared with the cells infected with KD-mock virus, those infected with KD-miR-31 expressed a lower level of ColI and α-SMA (Figure 2B).
miR-31 promotes HSC activation and migration
To investigate the role of miR-31 in regulating HSC activation further, we tested the effect of miR-31 over-expression and inhibition on HSC-T6 cells, a useful tool for studying cell mechanisms and biology . The miR-31 precursor and inhibitor were separately transfected in HSC-T6 cells to increase and decrease the expression of miR-31 individually. The results revealed that inhibition of miR-31 significantly decreased α-SMA and ColI expression whereas its over-expression obviously increased α-SMA and ColI expression, assessed by qRT-PCR and Western blotting (Figures 2C and 2D). These results suggested that inhibition of miR-31 significantly inhibited HSC activation whereas its over-expression obviously promoted HSC activation.
FIH1 is validated as the direct target of
miR-31 in HSCs and miR-31 promotes HSC activation and migration possibly through targeting FIH1
miR-31 accelerates HSC migration by increasing MMP-2 expression
Previous studies have demonstrated that HSCs cultured in Matrigel could remain in a quiescent state. Primary HSCs were seeded on dishes coated with a thin layer of Matrigel, and then we infected cells with the indicated lentivirus on day 2. Most of the cells seemed to be phenotypically quiescent, with a rounded shape and prominent lipid droplets. However, some of the cells infected with miR-31-lentivirus (31OP) began to spread out with more prominent protrusions. We suspected that the partially activated cells digest the matrix and migrate from the surface to the plastic below (Figure 2E, top).
To study the effect of miR-31 on HSC migration further, we used modified Costar's chambers which could partially mimic a microenvironment in health and disease. The upper layer of the permeable membrane was coated with Matrigel to mimic the normal space of Disse, whereas the lower membrane was coated with ColI to mimic the fibrillar matrix (Figure 2E, bottom). Then we detected whether cell migration was changed when miR-31 expression was altered. HSC-T6 cells were transfected separately with an miR-31 precursor or inhibitor for 48 h before being seeded in the upper compartment. The results suggested that over-expression of miR-31 significantly promoted cell migration. On the contrary, inhibition of miR-31 significantly decreased cell migration (Figures 2F and 2G). Previous findings have revealed that MMPs may be important for HSC migration , so the effects of miR-31 on MMP-2 and MMP-9, two major gelatinases contributing to cell migration, were studied. Our results showed that MMP-2 expression was significantly decreased by inhibition of miR-31 and increased by over-expression of miR-31, whereas MMP-9 expression was not affected (Figure 2H). In addition, cell proliferation and the cell cycle were evaluated using a CCK-8 and flow cytometry. We found that miR-31 had no effect on HSC proliferation and the cell cycle (results not shown). Collectively, over-expression of miR-31 promoted HSC migration by increasing MMP-2 expression.
Promotion of liver fibrosis by
miR-31 via direct targeting of FIH1
According to the TargetScan, miRDB, PicTar, miRnada and miRwalk algorithms, FIH1 was predicted to be a putative target of miR-31. The predicted sequences of interaction are shown in Figure 3(A). The expression of FIH1 was significantly reduced during HSC activation, which presented the completely opposite tendency to miR-31 (Figure 3B). Then we engineered luciferase reporters with either a wild-type 3′-UTR of FIH1 or a mutant 3′-UTR of FIH1 with a 7-bp anti-sense mutation in the target site of miR-31. The precursor of miR-31 apparently suppressed wild-type FIH1 3′-UTR luciferase activities compared with an empty vector control, whereas a mutant 3′-UTR of FIH1 disrupted base pairing between miR-31 and FIH1 (Figure 3C). In addition, FIH1 protein levels were reduced after transfection of the miR-31 precursor and increased after transfection of the miR-31 inhibitor (Figure 3D). Collectively, these results strongly suggest that miR-31 binds to the 3′-UTR of FIH1 and down-regulates its expression.
To study the biological function of FIH1 in HSCs further, the effects of shRNA-mediated knockdown of FIH1 on HSCs were studied. Three shRNAs (shFIH1-1, -2 and -3) targeting different regions of FIH1 were designed and cloned into the lentiviral vector pLKO.1. Some 72 h after infection, RNA and protein were collected to test knockdown efficiency. Two shRNAs (shFIH1-1 and shFIH1-3) showed high knockdown efficiency (Figures 3E and 3G). Furthermore, we found that knockdown of FIH1 by both these two shRNAs significantly increased the expression of α-SMA and ColI, two main markers of HSC activation (Figures 3E and 3G). Besides, knockdown of FIH1 by these two shRNAs also increased MMP-2 expression and promoted HSC migration assessed by Western blot and transwell assay, respectively (Figures 3F and 3G). These results indicated that knockdown of FIH1 could mimic the effects of miR-31.
Next, we wondered whether FIH1 could block the biological function of miR-31 in liver fibrogenesis. HSC-T6 cells were infected with a lentiviral vector encoding the entire FIH1-coding sequence without the 3′-UTR complemented to miR-31. The precursors of miR-31 were transfected at the same time. Our results showed that over-expression of FIH1 without a 3′-UTR obviously counteracted miR-31-promoted HSC activation and migration (Figures 3H and 3I). In short, our studies showed that FIH1 is a valid target gene for miR-31 and involved in miR-31-regulated fibrogenesis
The action of
miR-31/FIH1 on HSCs is independent of HIF-1α activity
To ascertain whether the observed miR-31/FIH1 action on HSCs involved HIF-1α, we first transfected miR-31 precursors into HSCs containing an HIF-1 reporter. After 2 days, no change was detected in HIF-1 transcriptional activity under normoxic conditions (Figure 3K). Then we also found that exogenous miR-31 expression had no effect on HIF-1α expression in normoxic culture assessed by Western blot (Figure 3J). These results indicated that miR-31's action on HSCs was independent of HIF-1 activity. We further studied whether FIH1's action on HSCs was associated with HIF-1α activity. Our results showed that knockdown of FIH1 by shFIH1-1/shFIH1-3 lentivirus infection also had no effect on HIF-1 activity (Figure 3K). Then, we observed a 1.5-fold increase in miR-31 expression when exposed to hypoxia (1% O2) for 8 h (Figure 3L). Under hypoxic culture, the transactivation of HIF activity in HSCs and miR-31-inhibited HSCs was induced compared with a normoxic control (Figure 3M); however, inhibition of miR-31 could not affect HIF transactivation (Figure 3M, second and third columns). These results suggested that miR-31/FIH1's action on HSCs is independent of HIF-1α activity.
miR-31 in TGF-β signalling in promoting liver fibrogenesis
To assess the mechanisms of miR-31 in regulation of liver fibrogenesis further, primary HSCs were treated with various cytokines such as TGF-β, platelet-derived growth factor (PDGF)-BB, vascular endothelial growth factor (VEGF), endothelial growth factor (EGF) and basic fibroblast growth factor (bFGF). Interestingly, only TGF-β1, a key profibrogenic mediator, significantly promoted the expression of miR-31 (Figure 4A). At 24 and 48 h after TGF-β1 treatment miR-31 was markedly increased, especially at 24 h (Figure 4B). Then HSC-T6 cells were treated with TGF-β1 and followed by transfection with miR-31 inhibitor or inhibitor control. The results revealed that TGF-β significantly decreased FIH1 expression and increased MMP-2, α-SMA and ColI expression, assessed by Western blot (Figure 4C). These effects were, however, partially attenuated by inhibition of miR-31 compared with the group treated with TGF-β1 (Figure 4C). These results suggested that miR-31 might be involved in the TGF-β signalling pathway and contribute to HSC activation and liver fibrogenesis.
miR-31 in TGF-β signalling in promoting liver fibrogenesis
miR-31 expression by binding to its promoter
The original view of the TGF-β superfamily signalling pathways was that there were essentially two branches signalling via activation of Smad2/3, and a bone morphogenetic protein (BMP)/growth and differentiation factor branch signalling via Smad1/5/8. To confirm the role of Smad signalling in regulation of miR-31, SB431542 (a potent and selective inhibitor of TGF-β1 receptor kinase) and LDN193189 (a potent small molecular inhibitor of BMP type I receptors) were used to block TGF-β/Smad signalling. HSCs were incubated for 1 h with SB431542 and LDN193189, and subsequently stimulated with 2 ng/ml of TGF-β1 for 24 h. Results showed that HSCs preincubated with SB431542 completely counteracted the induction of TGF-β whereas the LDN193189 group expressed the same level of miR-31 compared with the TGF-β group (Figure 5A). Then, we investigated further how miR-31 was involved in the TGF-β/Smad signalling pathway in promoting HSC activation and liver fibrogenesis. A potential Smad3-, but not Smad2-, binding element (SBE) was predicted in the miR-31 upstream region by using ALGGEN-PROMO (http://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promo.cgi?dirDB=TF_8.3&calledBy=alggen; Figure 5B). Promoter fragments containing an SBE or mutant SBE were cloned into the vector PGL-3 to infect HSCs along with the treatment with TGF-β1. Promoter activity was measured 72 h later. In our results, the miR-31 promoter activity was apparently enhanced by TGF-β1. However, mutation of the SBE completely abrogated the TGF-β response (Figures 5C and 5D). Next, Smad3 expression was silenced by using shRNA/Smad3. TGF-mediated induction of miR-31 promoter activity was obviously suppressed by inhibition of Smad3 compared with the control (Figure 5E). Later, ChIP showed further that the promoter region of miR-31 was immunoprecipitated by the Smad3 antibody, suggesting that endogenous Smad3 directly binds to the miR-31 promoter (Figure 5F). To demonstrate that the alteration of Smad3 is critical in the TGF-β-mediated promotion of HSC activation and migration, we silenced Smad3 expression in HSCs followed by the addition of TGF-β. The results of Western blotting indicated that the TGF-β-mediated promotion of HSC activation and migration was reversed by the suppression of Smad3 (Figure 5G). These results suggested that TGF-β stimulate the promoter activity of miR-31 by its downstream signal molecule Smad3.
miR-31 expression by binding its promoter
HSCs, also known as perisinusoidal cells, are pericytes that are found in the perisinusoidal space of the liver. They are regarded as the major cell type involved in liver fibrosis. In response to liver injury, quiescent HSCs undergo activation, proliferation and migration . We presumed that HSC behaviour was probably regulated by miRNAs, at least to a certain degree. A multitude of miRNAs that were altered in liver fibrosis have been found using an miRNA microarray. However, some important miRNAs were missing because these experiments were based on the studies of whole-liver miRNA. It is probable that some miRNAs that are particularly deregulated in HSCs may be left out. An analogous situation happened recently when researchers detected miR-133 in liver fibrosis . Indeed, we failed to detect the alteration of miR-31 in cirrhotic livers, the reason being that miR-31's relatively stabilized expression in hepatocytes, the most abundant cells in the liver, covers up its variation in HSCs. Taken together, miR-31 is likely to participate in the progression of liver fibrosis. We noted that the literature had reported decreased miR-31 levels in experimental pulmonary fibrosis and in fibroblasts in idiopathic pulmonary fibrosis . In addition, Masamune et al.  reported that miR-31 was significantly up-regulated during the activation of pancreatic stellate cells. Thus, we were more curious to know its biological function in liver fibrosis. In the present study, the results showed that over-expression of miR-31 promoted HSC activation and migration whereas its inhibition suppressed them. This suggests that the same miRNAs may have divergent effects in different microenvironments. To elucidate the mechanism of miR-31 regulation, we used a luciferase report assay to confirm its target gene, FIH1. Western blotting analysis further validated miR-31 being negatively regulated by the level of FIH1 protein. FIH1 is a recently found protein that binds to HIF-1α and suppresses its transactivation function . The transcriptional regulator HIF-1 which consists of an HIF-1α and an HIF-1β protein unit is an essential mediator of oxygen homoeostasis. The biological activity of HIF-1 is determined mainly by the expression and activity of the HIF-1α subunit [29–31]. Under non-hypoxic conditions, the HIF-1α subunit is soon degraded. In hypoxic conditions, the HIF-1α subunit is stabilized and interacts with co-activators such as p300/CBP to regulate its transcriptional activity. We found for the first time that the extraneous addition of miR-31 significantly promoted HSC activation and migration as a result of FIH1 loss, and FIH1 supplementation can rescue miR-31 effects.
When treated with some fibrosis-related cytokines, miR-31 expression was significantly up-regulated by TGF-β1. Moreover, TGF-β-induced cell activation and migration were partially blocked by an miR-31 inhibitor. Over-expression of FIH1 can also restore the changes induced by miR-31. In the liver, the TGF-β signal acts as a key mediator in fibrogenesis, contributing to HSC transdifferentiation. Although both Smad2 and Smad3 bind together to regulate TGF-β signals, they have distinct effects on HSCs. TGF-β mediates activation of Smad2 primarily in early cultured cells and that of Smad3 primarily in transdifferentiated cells . Uemura et al.  found that Smad3, not Smad2, is a direct regulator of HSC matrix interactions and production. Schnabl et al.  implicated the collagen expression in Smad3 null mice reducing by about half compared with control mice, in response to an acute fibrogenic stimulus. The literature supports the notion that Smad3 plays a more important role in HSC transdifferentiation.
In the present study, luciferase reporter and ChIP assays confirmed a potential binding site in the miR-31 promoter region for Smad3. Further experiments validated that knockdown of Smad3 in HSCs inhibited ColI and α-SMA expression induced by TGF-β. At present, a regulation network between TGF-β and miRNAs in fibrogenesis is getting more and more attention. TGF-β is considered to be an important stimulator of miR-21 enrichment during fibrosis. Zhong et al.  have demonstrated that Smad3-mediated up-regulation of miR-21 promotes renal fibrosis. Chung et al.  have revealed that Smad3, but not Smad2, regulated miR-192 expression through interaction with the promoter of miR-192.
In spite of recent advances, in the treatment of viral hepatitis, medical methods to promote the reversal of liver fibrosis remain limited. The miRNA-based therapeutic approaches may represent a new alternative to current methods in treatment of liver disorders. Miravirsen, an oligonucleotide targeting miR-122, has provided the first piece of clinical evidence for miRNAs as therapeutic targets . Based on the role of an miRNA candidate and its expression in diseased tissue, two main pharmacological strategies can be applied. Antagonists can inhibit miRNAs showing a gain of function in the fibrotic tissue, whereas miRNA mimics can restore miRNAs exhibiting a loss of function in fibrotic tissue [38,39]. The miRNA-based therapeutics is still in its early stages. More research needs to be done to understand its complexity.
In conclusion, our findings may identify the miR-31/FIH1 nexus as a novel regulator of TGF-β signalling in HSC-mediated liver fibrosis and suggest a potential clinical diagnostic and therapeutic approach in the future.
Jiangfeng Hu, Chao Chen, Qidong Liu, Jiuhong Kang and Liang Zhu conceived and designed the experiments; Jiangfeng Hu, Qidong Liu and Chenlin Song performed the experiments; Jiangfeng Hu, Chenlin Song and Qidong Liu analysed the data; Jiangfeng Hu, Baohai Liu, Songchen Zhu, Hongyu Yu, Yao and Su Liu contributed reagents/materials/analysis tools; and Jiangfeng Hu, Chao Chen, Qidong Liu and Chaoqun Wu wrote the paper.
The authors want to thank L. Chen, Y. Li, Y. Liu and G.P. Li for their help in guidance with the experiments, and T.Y. Wei for her help with the statistical analysis.
This work was supported by grants from the Ministry of Science and Technology [grants 2010CB944900, 2010CB945000, 2011CB965100, 2011CBA01100, and 2011DFA30480], National Natural Science Foundation of China [grants 10672176, 10972235, 11272342, 90919028, 31071306, 31000378, 31101061, 31171432 and 30971451], and the Science and Technology Commission of Shanghai Municipality [grants 10142201000, 124119a4100, 09DZ2260100, 11ZR1438500, 11XD1405300), IRT1168 and 20110072110039] from the Ministry of Education, China. The funders were not involved in the study design, data collection and analysis, or the decision to publish.
smooth muscle α-actin
basic fibroblast growth factor
bone morphogenetic protein
Cell Counting Kit-8
collagen type I
Dulbecco’s modified Eagle’s medium
endothelial growth factor
human embryonic kidney
hepatic stellate cell
platelet-derived growth factor
transforming growth factor β
vascular endothelial growth factor
These authors contributed equally to this work.