miRNA-338-3p/CDK4 signaling pathway suppressed hepatic stellate cell activation and proliferation

The isolation method of primary rat HSCs was according to the previous literature [29]. Primary Rat HSCs, HSC-T6 and HEK293T (human embryonic kidney cell line) were kindly gifted from Dr. Gao (Tongji University, Shanghai, China). The primary cells and cell lines were cultured in DMEM (Dulbecco’s modified Eagle’s medium, Thermo, Waltham, MA, USA) containing 10% FBS (Fetal bovine serum, FBS, Gibco, Grand Island, NY, USA) at 37 °C in a humidified atmosphere of 5% CO2.

Wild-type 3’UTR containing predicted miR-338-3p binding sites were amplified from HSC-T6 genomic DNA and inserted into the PGL3 luciferase reporter vector. Mutant 3’UTR was generated using the Quick Change Lighting Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA, USA). The CDK4 expression vector was obtained by cloning the CDK4-coding sequence into the pcDNA.

Luciferase assay was performed with the Dual Luciferase Reporter Assay System (Promega, Madison, WI). Transfection was carried out in 48 well plates using Fugen (Roche). There were two groups. One was co-transfected with 200 ng wild-type-CDK4-3’UTR, 20 nm miR-338 precursor, and 20 ng Renilla. Another was co-transfected with 200mutant CDK4 3’UTR without binding site of miR-338-3p, 20 nm miR-338 precursor, and 20 ng Renilla. 48 h later, Firefly and Renilla luciferase activities were tested.

The statistical analysis in our study was performed by using SPSS 22.0. Data were given as mean ± SEM. Two tailed t-test was used to determine between two groups. Statistical significance level was set at p < 0.05.

Based on the microarray data, multitude ectopic miRNAs were detected in the HSCs. In our previous study, we isolated rat primary HSCs and extracted total RNA to perform miRNA microarray assay. Our attention was focused on miR-338-3p, a new underlying member of liver fibrosis. The microarray data showed that the expression of miR-338-3p was sharply reduced by 90% at day 7 (partially activated HSCs) [14]. To validate this finding, RT-PCR was carried out to measure the endogenous miR-338-3p expression in a quiescent state and an activated state. As primary HSCs were gradually activated during culture, we assessed miR-338-3p expression at day 2, day 7 and day 14 after isolation. Our data indicated that endogenous miR-338-3p expression was obviously reduced at day 7 and day 14 (Fig. 1a). Meanwhile, collagen type I (Col1) and α-sma (α-smooth muscle actin, α-SMA), two key biomarkers of HSCs activation, was gradually increased (Fig. 1b, c). In addition, we found that treatment with transforming growth factor (TGF-β, 2 ng/ml) in quiescent HSCs (day 2), the expression of miR-338-3p was reduced rapidly (Fig. 1d). When cells were fully activated (day 14), we treated them with SB431542, a potent and specific inhibitor of TGF-β and detected the level of miR-338-3p. The results suggested that the expression of miR-338-3p was increased compared to that in control group (Fig. 1e).

As we discovered, miR-338-3p was significantly decreased during the process of activation, we transfected miR-338 precursor to repair its loss at the early stage of primary HSCs. On day 7, cells transfected with miR-338 precursor showed a more original shape with less peripheral protrusions. However, the control group cells showed a more irregular shape with more peripheral protrusions (Fig. 2a). As results showed in Figs. 1b, c and 2a, it seemed that there was a negative relationship between miR-338 expression and HSCs activation. We assumed that the expression of miR-338-3p was involved in this activation process. To confirm our hypothesis, we alter the miR-338-3p expression in primary HSC by transfecting miR-338 precursor. Interestingly, the expression of Col 1 and α-sma was slightly decreased due to the upregulation of miR-338-3p (Fig. 2b). Due to the low efficiency of transfection in primary cells, HSC-T6 cell line was further used to conduct the following studies. To replicate the results, HSC-T6 cell line was transfected with miR-338 precursor or negative control. 48 h later, cells were collected and transfection efficiency was confirmed by RT-PCR (Fig. 2c). Then the expression of Col1 and α-sma in two groups was measured using RT-PCR. As expected, the data showed miR-338-3p inhibited HSCs activation. The expression of Col1 and α-sma was respectively reduced as a result of miR-338-3p overexpression (Fig. 2d, e). Furthermore, inhibition of miR-338-3p could upregulate Col1 and α-sma which confirm their association on the other direction (Fig. 2f, g). Their expression was also confirmed at the protein level (Fig. 2h).

Next, we examined the impacts of miR-338 on HSCs proliferation. HSC-T6 were transfected with miR-338 precursor, inhibitor or corresponding negative control. In the CCK-8 assay, the cell growth curves suggested that overexpression of miR-338 significantly restrained HSC-T6 proliferation in a time-dependent manner (Fig. 3a.). Moreover, cells transfected with miR-338 inhibitor showed a higher proliferative ability (Fig. 3b). Edu incorporation assay also demonstrated that miR-338 precursor reduced the proliferation of HSC-T6 (Fig. 3c-d). Besides proliferation, we also assessed the role of miR-338-3p in HSCs migration. As Fig. 3e, f showed, miR-338-3p has no effects on cell migration.

Based on the prediction of Bioinformatics (TargetScan, PicTar), some genes are predicted to be the targets of miR-338-3p. Among them, we hypothesized CDK4, an oncogene in liver cancer, might be a putative target gene of miR-338-3p in liver fibrosis. The predicted sequence of interaction was showed in Fig. 3g. To test this prediction, 3’UTR with miR-338-3p binding sites were cloned into the PGL3 luciferase reporter vector. A mutant 3’UTR of CDK4 with anti-sense mutation in the predicted sites was also constructed. The reporter construct was co-transfected into HEK293T with Renilla plasmid and miR-338 precursor or negative control. The wild-type CDK43’UTR luciferase activity was suppressed due to miR-338-3p overexpression. By contrast, the activity of mutant-3’UTR-CDK4 remained relatively unaffected (Fig. 3h, i). Additionally, CDK4 expression was decreased in the miR-338 precursor group while increased in the miR-338 inhibitor group (Fig. 3j, k). miR-338-3p could also regulate CDK4 at protein level (Fig. 4l). Taken together, these data strongly suggested that miR-338-3p repressed CDK4 expression by directly binding to its 3’UTR region.

Since miR-338 regulated cell growth, cell migration and cell invasion in liver cancer and colorectal carcinoma by targeting CDK4 [21, 30], the role of CDK4 in liver fibrosis remained unclear.

We conduct CCK-8 assay to determine the proliferation of HSC-T6 transfected with CDK4 plasmid or control empty vector. The results showed that cell transfected with CDK4 plasmid displayed a higher proliferative capacity compared with the control group (Fig. 4a). Besides, we found cells transfected with CDK4 could promote cell activation with upregulation of Col1 and α-Sma (Fig. 4b, c). We co-transfected HSC-T6 cells with miR-338 precursor and CDK4 plasmid to investigate whether CDK4 would rescue the inhibition effect of miR-338 on the cell activation and proliferation or not. The results suggested that overexpression of CDK4 partially block the repression effect of miR-338 on the activation and proliferation of HSC-T6. The growth curves of three groups (Control, Pre-miR-338/pcDNA, Pre-miR-338/CDK4) were shown in Fig. 4d. The expression of activation associated markers, Col1 andα-Sma, was showed in Fig. 4e-f.