TCF7/SNAI2/miR-4306 feedback loop promotes hypertrophy of ligamentum flavum

This study was approved by the Institutional Research Ethics Committee of the Zhujiang Hospital of Southern Medical University. A total of 30 LF tissues samples, including hypertrophied and non-hypertrophied LF tissues, were collected from patient under-going lumbar spine surgery in Zhujiang Hospital of Southern Medical University. The hypertrophied LF tissues (> 4 mm thickness) were obtained from patient with LSCS due to LF hypertrophy, and non-hypertrophied LF tissues (≤ 4 mm thickness) were obtained from age- and gender-matched lumbar disc herniation (LDH) patient without LF hypertrophy as control. All enrolled patients were excluded from diseases such as cancer, heart disease, kidney disease, rheumatism and autoimmune diseases. All patients underwent magnetic resonance imaging scan to confirm the thickness of the LF before surgery, and all LF samples were obtained from the anatomical region (L4/5). Fibrosis score was assessed by Masson's trichrome staining according to previously report [38]. Informed consent was obtained from each patient prior to this study. Patient information in this study is summarized in Additional file 5: Table S1.

Total RNAs were extracted from three independent samples of hypertrophied or non- hypertrophied LF using TRIzol reagent (Invitrogen) according to the manufacturer’s recommended protocol, and the RNA quantity was assessed with a NanoDrop ND-2000 spectrophotometer (NanoDrop Technologies). After purifying the mRNA using RiboZero Magnetic Gold Kit, the cDNA libraries were constructed for the KAPA Stranded RNA-Seq Library Prep kit (Illumina, Inc.) according to the manufacturer’s instructions. Subsequently, we used Agilent 2100 and qPCR to assess the quality and quantification of the cDNA library. Finally, the RNA-sequencing was performed by Next-Generation Sequencing with an Illumina HiSeq Xten platform. Clean data were obtained from the raw data by removing reads containing adapters, reads containing over 10% poly N, and low-quality reads, which were aligned to the specified reference genome (Homo sapiens. GRCh38, NBCI) to obtain the mapped data. The differentially expressed mRNAs between hypertrophied LF tissues and non- hypertrophied LF tissues were performed using EBseq R package. The fold changes (FCs) ≥ 2 or − 2 and false discovery rates (FDRs) < 0.05 served as the screening criteria to get differentially expressed mRNA.

According to our previously described method [4], LF cells were isolated and cultured from the LF tissues of patients with LSCS or LDH. Briefly, the obtained LF samples were cut into small pieces and digested for 2 h at 37 °C using Dulbecco’s modified Eagle’s medium (DMEM, Gibco) with 0.2% type I collagenase (Gibco), then seeded on to cell culture dish and incubated with DMEM containing 10% fetal bovine serum (FBS, Gibco) and 1% penicillin/streptomycin (Invitrogen). The isolated cells were observed for fibroblast morphology and identified the expression of specific markers [collagen I (1:100, #72026, Cell Signaling Technology) and Vimentin (1:500, ab16700, Abcam)] by immunofluorescence staining. Subsequent experiments were conducted using cells from the third passage to forth passage of LF cells.

The adenovirus of TCF7 overexpression/knockdown or SNAI2 overexpression/knockdown were constructed by Genechem (Nanjing, China). The adenovirusl particles of miR-4306 mimic/inhibitor and negative control (NC) mimics/inhibitor were obtained from Ribobio Inc. (Guangzhou, Guangdong, China). All cells were infected with the adenovirus according the manufacturer’s instructions, and the infection efficiency was verified by real-time quantitative PCR (RT-qPCR) or Western blotting analyses. All shRNA and miR-4306 mimics/inhibitor sequences are listed in Additional file 6:Table S2.

Total RNA from the cells or tissues sample was extracted using Trizol (Invitrogen, Carlsbad, CA). Reverse transcription and RT-qPCR for miR-4306 were carried out using miRNA 1st Strand cDNA Synthesis Kit (MR101-01, Vazyme) and miRNA Universal SYBR qPCR Master Mix (MQ101-01, Vazyme) on a LightCycler^®96 (Roche). Reverse transcription and RT-qPCR for the mRNAs were performed as described in our previous studies [4].U6 small nuclear RNA and GAPDH were used as an internal control for miR-4306 and mRNAs, respectively. The relative RNA expression of genes was calculated according to the Ct (2^−ΔΔCt) method. All experiments were performed in triplicate. All specific primers sequence in study are listed in Additional file 7:Table S3.

The nuclear protein extraction was performed using the Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime, Shanghai, China) according to the protocol of manufacturer, and then subjected to western blotting analysis. Western blotting was conducted as described in our previous studies [4]. Primary antibodies antibody information is as follows: cleaved Caspase-3 (1:500, ab32042, Abcam), Bax (1:1000, ab182733, Abcam), Bcl-2 (1:500, ab182858, Abcam), TCF7 (1:1000, #2203, Cell Signaling Technology), SNAI2 (1:1000, #9585, Cell Signaling Technology), collagen I (1:1000, ab260043, Abcam), collagen III (1:1000, ab7778, Abcam), MMP2 (1:500, ab181286, Abcam), MMP13 (1:1000, ab39012, Abcam), TGF-1β (1:1000, ab215715, Abcam), and GAPDH (1:1000, ab8245, Abcam).

Cell proliferation ability was detected using a Cell Proliferation Reagent Kit I (MTT, Sigma-Aldrich) according to the manufacturer’s recommended protocol. Absorbance values were measured at the wavelength of 490 nm. All experiments were performed in three to five biological duplicates.

Cells apoptosis was measured by flow cytometry using Annexin V-FITC/propidine iodide (PI) double-staining kit (Abcam) according to the manufacturer’s instructions. All experiments were performed in three biological duplicates.

A chromatin immunoprecipitation (ChIP) assay was carried out using the ChIP assay kit (ab500, Abcam) according to the manufacturer’s instructions. Briefly, the following antibodies were used to immunoprecipitate cross-linked protein-DNA complexes: rabbit anti-TCF7 (1:50, #2203, Cell Signaling Technology, USA), rabbit anti-SNAI2 (1:50, #9585, Cell Signaling Technology, USA), and normal rabbit IgG (1:50, #2729, Cell Signaling Technology, USA). After cross-linked protein-DNA complexes to free DNA, the immunoprecipitated DNA was purified for RT-qPCR analyses with specific primers.

For validation of transcription factor interactions with target genes, the TCF7-binding motif in the promoter region of SNAI2 and the SNAI2-binding motif in the promoter region of miR-4306 were predicted through JASPAR (http://​jaspar.​genereg.​net/​). According to the predicted results, different truncated plasmid of SNAI2 or miR-4306 promoter was co-transfected with corresponding transcription factors plasmid into 293 T cells. For the verification of interaction between miR-4306 and TCF7, the wild type or mutant type of TCF7 3′-UTR (containing the binding site of miR-4306) was cloned into the luciferase vector, and then transfected into 293 T cells together with miR-4306 mimics or the negative control (NC mimics), respectively. Transfer after 24 h, luciferase activities were assessed using a Dual Luciferase Assay Kit (Promega, Madison, WI, USA) in accordance with the manufacturer’s instructions. The relative luciferase activities were calculated by normalizing the signal value of Renilla luciferase to Firefly luciferase, and then compared with the negative control.

RNA immunoprecipitation (RIP) experiments were performed with a Magna RIP™ RNA-Binding Protein Immunoprecipitation Kit (Millipore, Billerica, MA) according to the manufacturer’s instructions. AGO2 antibody ((Millipore, Billerica, MA, USA)) or negative control IgG antibody (Millipore, Billerica, MA, USA) was used for RIP. Co-precipitated RNAs were finally extracted with RNeasy Mini Kit (QIAGEN, China) for RT-qPCR to demonstrate the presence of the binding targets.

All animal experimental protocols were approved by the Ethics Committee for Animal Research of the Zhujiang Hospital of Southern Medical University (Guangzhou, China). All mice purchased from the Experimental Animal Center of Southern Medical University were 8-week-old C57BL/6 male mice. The mice model with HLF was constructed by the hydrophobic characteristics of mice according to the previously described [45, 46]. In brief, the mice were placed in a beaker with 5 mm of animal water at the bottom to induce a bipedal standing posture. The mice were maintained in a bipedal standing posture for 6 h a day with an interval of 2 h of free activity, and the above operations were continued for 8 weeks to construct the HLF mice model. The control mice were placed in a same condition as the mice with HLF model except for the bottom without animal water. To verify the therapeutic effect of silencing TCF7 on the mouse HLF model, 20 mice were randomly assigned to the control group, HLF model group, HLF + AAV-shNC group, and HLF + AAV-shTCF7 group. After six weeks of bipedal induction, we made a longitudinal skin incision in the mouse lumbar spine, and removed the dorsal paravertebral muscle from the spinous processes and laminae to expose the L5/6 LF under a surgical microscope, thereby AAV-shNC or AAV-shTCF7 (1 × 10¹² vg/ml, 3 µl) from Hanbio Biotechnology (Shanghai, China) was injected into L5/6 LF in anesthetized mice with a microinjector (NF36BV 36GA, NanoFil, United States). After 4 weeks of AVV injection, mice were euthanized and L5/6 vertebrae were collected to obtain LF tissue for later histological analyses and molecular biological analyses. Histological analyses were performed to measure the area of the LF.

hematoxylin and eosin (H&E) staining and Elastica van Gieson (EVG) staining were performed to measure the area of the LF and degree of LF fibrosis, respectively. The LF samples from mice or humans were fixed overnight with 4% paraformaldehyde (Beyotime, shanghai, China), paraffin-embedded, and cut into slices (4 μm). After dewaxing and dehydration, the slices were performed by H&E kit (Beyotime, shanghai, China) or EVG kit (JianChen, Nanjing, China) according to the manufacturer’s instructions. Quantitative analyses of the LF areas and degree of LF fibrosis (the ratio of elastic fibers to collagen fibers) were obtained using ImageJ software (NIH, United States). Each sample was detected three times, and the average value was taken.

Immunohistochemical staining was conducted using Histostain®-SP Kits according to the manufacturer’s instructions. In brief, the LF samples from humans were fixed overnight with 4% paraformaldehyde (Beyotime, shanghai, China), paraffin-embedded, and then cut into slices with 4 μm thickness. After dewaxing and antigen retrieval, the slices were organization background closed using serum blocking reagent. Then these slices were incubated with TCF7 primary antibodies (1:200, #2203, Cell Signaling Technology) and β-catenin (1:100, ab16051, Abcam) overnight at 4 °C. Subsequently, these sections were incubated with the respective secondary antibodies (Proteintech) at room temperature. Immunohistochemical results were visualized using a Zeiss microscope (Carl Zeiss Meditec AG, Jena, Germany) and then analyzed by ImageJ software (NIH, United States).

Results are presented as the means ± SD. All the statistical analyses were performed using the GraphPad Prism 6.0 (GraphPad Software, La Jolla, USA). Statistical significance for comparisons between two groups were analyzed using Student’s t-test, and statistical significance for comparisons among more than two group was analyzed using one-way ANOVA. The p < 0.05 was considered as significant difference.

To explore the underlying mechanism of HLF, the mRNA sequencing was performed by in the HLF tissues and non-HLF tissues. HLF tissue or non-HLF tissue has been identified by MRI and histopathology (Fig. 1A). The results of RNA sequencing analysis showed that a total of 848 mRNAs were differentially expressed between the HLF tissues and non-HLF tissues, of which 407 were up-regulated and 441 down-regulated in HLF tissues (Fig. 1B). Then we intersected the differential mRNA with a currently known transcription factor, and found that the six transcription factor (TCF7/CDX1/FOSL1/FOSB/FOXM1/EGR1) in HLF tissues significantly higher than those in non-HLF tissues, and five transcription factor (LMO3/NR2F2/RARB/TBX2/NOTCH3) significantly downregulated (Fig. 1C). Subsequently, KEGG analysis of the above differential transcription factors showed that they were mainly enriched in WNT signaling pathway, Notch signaling pathway, and IL-17 signaling pathway (Fig. 1D). The Wnt/β-catenin pathway signaling has been shown to propagates the initiation and progression of HLF [22], and TCF7 is essential for the typical WNT/β-catenin pathway [44]. However, the relationship between TCF7 and HLF pathogenesis is unclear and needs further study. The clinical samples were expanded to verify the mRNA expression of TCF7 in the above pathways in the non-HLF tissues and HLF tissues. The results of the RT-qPCR indicated that the mRNA expression of TCF7 in HLF tissue was significantly higher than that in non-HLF tissue (Fig. 1E), which was consistent with the mRNA sequencing results. In addition, the results of the IHC also demonstrated that the protein expression of TCF7 and β-catenin in HLF tissues was significantly increased compared with that of in non-HLF tissue (Fig. 1F). Moreover, the correlation analysis revealed that the mRNA expression of TCF7 were significantly positively correlation with LF thickness and fibrosis score (Fig. 1G). In addition, the cells isolated from the HLF tissues or non-HLF tissues were identified as LF cells by the observation of cell morphology and immunofluorescence detection of LF makers expression (Additional file 1: Fig. S1), and the results showed that the isolated cells had a typical LF cells phenotype and uniformly expressed of LF cells markers (collagen I and Vimentin), suggesting that the isolated cells is the high purity of the LF cells. Then the mRNA expression of TCF7 was measured in HLF cells and non-HLF cells by RT-qPCR, and the results indicated that TCF7 expression was significantly increased in HLF cells compared with that of in non-HLF cells (Fig. 1I). Together, these data indicated that TCF7 was markedly upregulated in HLF tissue and cells, and increased TCF7 in HLF tissues associated with LF thickness and fibrosis.

To explore whether TCF7 participates in the biological function of HLF cells, we analyzed the effect of overexpressing or silencing TCF7 (Fig. 2A) on viability, apoptosis, and fibrosis in HLF cells. The cell viability was detected using the CCK-8 assay, and the results showed that overexpression of TCF7 significantly promoted viability of HLF cells, whereas knockdown of TCF7 resulted in the opposite result (Fig. 2B). Then cell apoptosis was measured by using the flow cytometry and western blot. The results of flow cytometry indicated that overexpressing TCF7 inhibited apoptosis ratios of HLF cells, whereas knockdown of TCF7 promoted apoptosis ratios of HLF cells (Fig. 2C and D). Consistent with the results of flow cytometry, the results of western blot demonstrated that overexpression of TCF7 in HLF cells involved in inhibition of Bax and cleaved caspase-3 protein expression and activation of Bcl-2 protein expression, whereas knockdown of TCF7 resulted in the opposite result (Fig. 2E). Furthermore, to confirmed the relationship between the TCF7 expression and fibrosis in HLF cells, western blot was used to assess the effect of overexpressing or silencing TCF7 on the expression of fibrosis-related proteins (Collagen I, Collagen III, MMP13, MMP2, and TGF-β1) in HLF cells. The results of western blot revealed that overexpression or knockdown of TCF7 enhanced or suppressed the expression of fibrosis-related proteins (Collagen I, Collagen III, MMP13, MMP2, and TGFβ), respectively (Fig. 2F), suggesting TCF7 promoted fibrogenesis in HLF cells. Together, the above data demonstrated that TCF7 enhanced the proliferation, anti-apoptosis, and fibrosis in HLF cells.

Current studies have shown that the AAV vector system is a promising delivery vehicle of gene therapy because of its safe and long-term efficacy [23, 33]. In order to further confirm the role of the TCF7 in LF, we evaluated the therapeutic effect of TCF7 knockdown on mice with HLF by in situ injection of AAV-shNC or AAV-shTCF7. The results of HE staining and EVG staining indicated that LF area and the ratio of collagen fibers to elastic fibers significantly increased in the bipedal standing-induced HLF mice (Fig. 3A and B), which were consistent with the previous studies in humans and animals [30, 45, 46], suggesting that the mouse HLF model was successfully constructed in the study. The results of western blot demonstrated that the administration of the AAV-shNC had no effect on the high expression of TCF7 in mouse HLF tissues, while the administration of the AAV-shTCF7 significantly inhibited TCF7 expression in mouse HLF tissues (Fig. 3C and D). Further our data showed that no significant difference was observed in the LF area between the HLF mouse and HLF mouse treated with AAV-shNC (Fig. 3A and B), indicating that the administration of the AAV-shNC had no effect on HLF. Compared with the HLF mouse treated with AAV-shNC, the administration of the AAV-shTCF7 obviously decreased LF area and the ratio of collagen fibers to elastic fibers (Fig. 3A and B), suggesting that administration of the AAV-shTCF7 inhibited hypertrophy and fibrosis of LF induced by bipedal standing posture in mice. Moreover, the expression levels of apoptosis-related proteins (Bax, Bcl-2, and cleaved caspase-3) and fibrosis-related proteins (Collagen I, Collagen III, MMP13, MMP2, and TGFβ1) were detected by western blot in LF tissues of mice. The results indicated that the protein expression of the Bax and cleaved caspase-3 was decreased and the protein expression of Bcl-2, Collagen I, Collagen III, MMP13, MMP2, and TGFβ1 was increased in mice with HLF compared with that of in mice with non-HLF, and the above phenomena were reversed by the administration of the AAV2-shTCF7 (Fig. 3C, E, F, and G). Together, our data revealed that TCF7 knockdown could restrain the hypertrophy and fibrosis of LF in bipedal standing mouse model.

To further elucidate the underlying mechanism by which TCF7 promotes LF fibrosis, the TRRUST ver.7.3 (https://​www.​grnpedia.​org/​trrust/​result.​php) database and JASPAR database (https://​jaspar.​genereg.​net/​search?​advanced=​true) were used to predict the promoters of the SNAI2 genes. The results showed that promoter region of the SNAI2 contained three TCF7 binding sites (Fig. 4A). While the results of RT-qPCR and western blot showed that overexpressing TCF7 markedly increased the mRNA and protein expression of SNAI2 in HLF cells, while silencing TCF7 displayed the opposite effects (Fig. 4B). These results implied that TCF7 might bind to the SNAI2 promoter to positively regulate SNAI2 expression in HLF cells. ChIP assay confirmed that TCF7 can directly associate with the SNAI2 promoters, and observed successful recruitment of TCF7 by binding sites #1 rather than binding sites #2 and #3 (Fig. 4C). Consistently, the results of luciferase reporter assay showed that overexpressing TCF7 significantly increased SNAI2 promoter-driven reporter activity HLF cells (Fig. 4D), further confirming TCF7 directly targets SNAI2 and activates its transcription expression.

Next, we explored whether the overexpression of TCF7 enhanced the proliferation and fibrosis by activation of SNAI2 in HLF cells. The increased viability of HLF cells induced by overexpressing TCF7 was also dramatically abrogated by the silencing SNAI2 (Additional file 2: Fig. S2A and Fig. 4E). Furthermore, the results of flow cytometry showed that silencing SNAI2 significantly decreased the inhibitory effect of TCF7 overexpression on the apoptosis rate (Fig. 4F). Consistently, the results of western blot indicated that silencing SNAI2 significantly reversed the effect of TCF7 overexpression on the protein expression of apoptosis-related markers (Bax, cleaved caspase-3, and Bcl-2) in HLF cells (Additional file 2: Fig. S2B). These results demonstrated that TCF7 inhibited apoptosis of HLF cells by upregulating SNAI2. Furthermore, the increased protein expression of fibrosis-related proteins (Collagen I, Collagen III, MMP13, MMP2, and TGFβ1) induced by overexpressing TCF7 was also markedly reversed by the silencing SNAI2 (Fig. 4G). Moreover, our data also showed that overexpressing SNAI2 significantly reversed the inhibition effect of TCF7 knockdown on the cell viability, anti-apoptosis, and the protein expression of fibrosis-related markers (Additional file 3: Fig. S3), indicating that silencing TCF7 inhibit the proliferation and fibrosis of HLF cells by downregulating SNAI2. Interestingly, we also found that the overexpressing SNAI2 in HLF cells significantly enhanced the cell viability, anti-apoptosis, and the protein expression of fibrosis-related markers (Additional file 3: Fig. S3 and Fig. 4E–G). Whereas silencing SNAI2 in HLF cells remarkably decreased the cell viability, anti-apoptosis, and the protein expression of fibrosis-related markers compared with vector group (Additional file 3: Fig. S3). Together, our data suggested that activation of SNAI2 contributes to TCF7-mediated LF hypertrophic and fibrotic phenotype in vitro.

As shown in Fig. 4 and Additional file 3: Fig. S3, SNAI2 is involved in the malignant biological phenotype of HLF cells, such as abnormal proliferation and fibrosis. However, the molecular mechanism by which SNAI2 regulated the malignant biological phenotype of HLF cells remained unclear. As a transcription factor, SNAI2 could inhibit the transcription by directly binding to miRNAs promoters [8, 9, 34]. The TransmiR v2.0 database (http://​www.​cuilab.​cn/​transmir) database was used to predict miRNAs that motif of SNAI2 binds to miRNAs promoters, and found that motif of SNAI2 had binding sites with the promoters of three miRNAs (miR-4306, miR-641, and miR-2278). The results of RT-qPCR showed that overexpressing or silencing SNAI2 markedly decreased or increased the miR-4306 expression in HLF cells, but the expression levels of miR-641 and miR-2278 were not regulated by SNAI2 (Fig. 5A). These data suggested that SNAI2 might bind to miR-4306 promoter to negatively regulate miR-4306 expression in HLF cells. To confirmed the binding regions of miR-4306 promoter regulated by SNAI2, two binding sites for SNAI2 motif on miR-4306 promoter region were predicted by the JASPAR database (Fig. 5B). ChIP assay demonstrated that SNAI2 could bind the promoter of miR-4306 directly, and observed successful recruitment of SNAI2 by binding sites #1 or site #2, and the effect of binding site #1 was better than binding site #2 (Fig. 5C). Consistently, the results of luciferase reporter assay showed that overexpressing SNIAI2 significantly decreased miR-4306 promoter-driven reporter activity (Fig. 5D), further confirming SNIAI2 directly targets miR-4306 and activates its transcription expression. Importantly, we found that overexpressing or silencing TCF7 markedly decreased or increased miR-4306 expression, the phenomenon was dramatically abrogated by the silencing or overexpressing SNAI2 (Fig. 5E), suggesting that TCF7 negatively regulated miR-4306 expression by mediating SNAI2 expression. Together, these data demonstrated that SNAI2 positively regulated by TCF7 inhibited the transcription of miR-4306 by binding to the promoter region of miR-4306.

Previous studies have shown that TCF7 is negatively regulated by miRNAs [12, 28]. To explore whether miR-4306 negatively regulated TCF7 expression in HLF cells, the RAID and miRDB database were used to predict the interaction between TCF7 and miR-4306. The results showed that TCF7 3’UTR had potential binding sites with miR-4306. And the overexpressing miR-4306 markedly inhibited the mRNA and protein expression of TCF7 in HLF cells, while silencing miR-4306 displayed the opposite effect (Fig. 5F and G). these results suggested that TCF7 may be a target gene of miR-4306 in HLF cells. The relationship between miR-4306 and TCF7 was evaluated through RIP experiments, the results revealed the interaction between miR-4306 and TCF7 (Fig. 5H). Further dual-luciferase reporter assay was used to confirm the specific binding site of miR-4306 and TCF7. The sequence of TCF7 3′-UTR including the miR-4306 binding site (TCF7 3′-UTR-WT) or its corresponding mutation sequence (TCF7 3′-UTR-MTU) was inserted into the psiTM-Check2 luciferase vector, and then co-transfected 293 T cells with NC mimics or miR-4306 mimic and TCF7 3′-UTR-WT plasmids or TCF7 3′-UTR-MTU plasmids, respectively. The results showed that miR-4306 overexpression greatly inhibited the luciferase activity of TCF7 3′-UTR-WT reporter, but luciferase activity of TCF7 3′-UTR-MUT was not significantly impacted by miR-4306 overexpression (Fig. 5I), indicating that miR4306 directly binds to the TCF7 3′-UTR region to regulate TCF7 transcription. In summary, our data demonstrated that SNAI2 enhanced TCF7 expression by transcriptionally inhibiting miR-4306 expression.

It has been reported that miR-4306 is significantly down-regulated in HLF tissues (fold change = − 1.26, P = 0.005), and its expression level was significantly negatively correlated with LF/spinal canal area ratio (LSAR), suggesting that miR-4306 may be involved in the occurrence and development of HLF [22]. However, the role of miR-4306 in HLF cells has not been explored. HLF cells were transfected with miR-4306 mimics or miR-4306 inhibitor to evaluate the effects of miR-4306 on the viability, apoptosis, and fibrosis of HLF cells. We found that overexpressing miR-4306 inhibited viability of HLF cells, this inhibition effect was reversed by overexpressing SNAI2 (Fig. 6A and B). The results of flow cytometry revealed that overexpressing miR-4306 significantly increased the apoptosis ratios of HLF cells, this effect was weakened by overexpressing SNAI2 (Fig. 7C). Consistently, the results of western blot also indicated that overexpressing miR-4306 significantly increased the protein expression of cleaved caspase-3 and Bax and markedly decreased the Bcl-2 protein expression, this phenomenon was reversed by overexpressing SNAI2 (Fig. 6D). Furthermore, the overexpressing miR-4306 significantly decreased the protein expression of fibrosis-related proteins (Collagen I, Collagen III, MMP13, MMP2, and TGFβ1), the inhibition function was also markedly reversed by overexpressing SNAI2 (Fig. 6E). In addition, our data showed that silencing miR-4306 significantly promoted cell viability, anti-apoptosis, and the protein expression of fibrosis-related markers in HLF cells, this phenomenon was weakened by silencing SNAI2 (Additional file 4: Fig. S4). In summary, miR-4036 negatively regulated by SNAI2 inhibited the proliferation and fibrosis of HLF cells.