Retraction Note: Integrin-β5, a miR-185-targeted gene, promotes hepatocellular carcinoma tumorigenesis by regulating β-catenin stability

MHCC-97 L cell lysates were incubated with IgG (control) or ITGB5 antibody. The bounded proteins were eluted, resolved by SDS-PAGE, and visualized by CBB staining (upper panel). The band in the ITGB5 lane was identified as β-catenin by mass spectrometry. The obtained peptide sequences by mass spectrometry analysis were shown on the right.(B-D)The RNA levels of miR-185 were analysed by q-RT-PCR. (TIFF 552 kb)


Background
Hepatocellular carcinoma(HCC) accounts for most (85 to 90%) of the primary liver cancers and ranks as the second leading cause of cancer-related death in men. Globally, approximately 50% of the cases occur in China [1][2][3]. Historically, the high mortality rate of HCC is due to the advanced stage and frequent recurrence after surgical resection, and the majority of the patients are virtually limited to surgical treatment or chemotherapy [4][5][6]. Therefore, an in-depth study of the molecular pathogenesis could contribute by providing reliable predictors of HCC and a basis for effective therapeutic strategies, which is a promising approach to improve the survival rate [7].
Integrins are heterodimeric cell surface receptors consisting of 18 α-subunits and 8 β-subunits regulating diverse biological functions in cancers, including proliferation, adhesion, migration and invasion. Multifarious integrins involved in the disease progression and their correlation to oncogenes make them attractive targets for tumour therapy [8]. ITGB5,which associates only with ITGαV [9], has been linked to a variety of pathological conditions by several studies. ITGB5 enhances cell survival through cooperating with endothelial growth factor receptor 2(VEGF2) to inhibit extrinsic apoptosis [10][11][12]. The contribution of ITGB5 to angiogenesis has been well established [13][14][15]. Accumulating evidence supports the pivotal role of ITGB5 in facilitating cancer cell migration, invasion and transforming growth factor β (TGF-β)-induced epithelial-mesenchymal transition(EMT) [16][17][18][19]. In 2015, a remarkable publication demonstrated that exosomal ITGB5 governed liver tropism, which was associated with liver metastasis in several malignancies, including colorectal cancer, pancreatic cancer and gastric cancer [20]. However, the role of ITGB5 in HCC is unknown.
The WNT/β-catenin signalling pathway mediates various cell biological functions in widely diverse cancer processes [21][22][23]. β-catenin is a component of the cell surface cadherin protein complex, whose stabilization is indispensable for the activation of the WNT/β-catenin pathway [22]. A large body of evidence has indicated that alterations in β-catenin activity are correlated to the development of HCC, including the proliferation, migration and differentiation [23]. The association of integrin with β-catenin has also been published in relation to melanoma and colorectal cancer [24,25]. Meanwhile, the interplay between ITGB5 and β-catenin can still not be elucidated in HCC.
The present study provides evidence that ITGB5 potently facilitates the tumorigenesis and progression of HCC cells in vitro and in vivo. For the first time, we focused on the relationship between ITGB5 and β-catenin in HCC, revealing that the expression of ITGB5 positively correlated with the stability of β-catenin leading to an increase in β-catenin. In addition, our findings implied that miR-185 participated in the mediation of the cell growth and migration in HCC by directly regulating ITGB5. Therefore, our data indicated that miR-185-ITGB5-β-catenin pathway may play an important role in HCC tumorigenesis.

Western blot analysis and co-immunoprecipitation
For western blot analysis, cells were lysed with RIPA buffer (150 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, 1% NP40, 1 mM EDTA, and 50 mM Tris (pH = 8.0)). After boiling with loading buffer for 15 min, the supernatants were loaded onto SDS-PAGE gels. Nitrocellulose membranes (Millipore) were incubated with primary antibodies, and then, they were incubated and visualized by HRP conjugate.
For immunoprecipitation (IP) experiments, cultured cells were washed in cold PBS and lysed in lysis buffer (20 mM Tris pH = 7.4, 0.15 M NaCl, 1 mM EDTA, 1 mM EGTA, 1% TritonX-100, 0.0025 MNa 2 P 2 O 7 , 1 mM β-glycerol phosphate, 1 mM Na 3 VO 4 , and 1 mM NaF) with an EDTA-free protease inhibitor mixture (Roche Applied Science) and PMSF. Cell lysates were then incubated on ice for 10 min, centrifuged, and precleared by incubation with protein A/G beads (Roche) for 1 h at 4°C. For IP bead preparation, 30 μl of protein A/G beads was incubated with 5 μg of antibodies for 2 h at 4°C. IPs were carried out overnight at 4°C, and the beads were centrifuged, washed and subsequently boiled for 5 min in SDS protein loading buffer to elute bound protein. IP lysates were subjected to Western blotting with the indicated antibodies.

Colony formation assays
Huh-7 and MHCC-97 L cells with the treatment as indicated (1 × 10 3 cells per well) were plated into 6 well plates and cultured at 37°C equipped with 5% CO 2 . Cells were fed with fresh growth medium every 3 days. Colonies were allowed to form for 2 weeks and were fixed with 4% paraformaldehyde, stained with crystal violet, washed with water to remove excess stain, and counted using Image J software. Each experiment was repeated three times.

Transwell assays
The transwell system was used for the determination of cell migration. Cells were seeded in serum-free medium in chambers (8.00 mm pores, BD, Biosciences), and then, they were allowed to migrate across the uncoated inserts using serum-containing medium for 24 h. Cells on the apical surface of the insert were scraped off, and the membranes invaded with cells were fixed with 1% paraformaldehyde and stained with crystal violet. Cell counts are expressed as the average number of cells per field of view. Three independent experiments were performed.
-3′ and 5′-gcGGATCCTCAGTCCACAGTGCCATT-3′ were used to generate plasmids encoding fulllength ITGB5. ITGB5 cDNA was purchased from Genecopoeia,Inc. (A8639). To generate lentivirus expressing ITGB5, HEK 293 T cells grown on a 6 cm dish were transfected with 2 μg pCDH-ITGB5 or control vector (pCDH), 1.5 μg psPax2, and 0.5 μg pMD2G. 24 h after the transfection, cells were cultured with DMEM containing 10% FBS for an additional 24 h. The culture medium containing lentiviral particles was centrifuged at 1000 g for 5 min. Viral particles collected in the supernatant were used for infection. In order to establish the stable cell line, the puromycin was used as a selection marker for the infected cells. The expression efficiency was evaluated by western blot analysis.

Introduction of microRNA mimics and inhibitors
Mimics and inhibitors of miRNA-185 were synthesized by the Genepharma Company (Shanghai, People's Republic of China). For each transfection in a six-well plate, 100 nM miRNA mimics, mimic control or inhibitor, or inhibitor control were used. The transfection of melanoma cells by oligofectamine (Invitrogen) was performed according to the manufacturer's instructions.

Protein stability assay and in vivo ubiquitination assay
To assess protein stability, cells were treated with cycloheximide (CHX) (20 mmol/L) for the indicated times(2,4, or 6 h). Cell extracts were prepared, and the expression levels of the indicated proteins were detected with western blotting. The ubiquitination assay was performed as previously described [28,29].
Promoter reporters and dual-luciferase assay MHCC-97 L and Huh-7 cells were transfected with the TOP-FLASH and FOP-FLASH reporter plasmids together with pRL-TK. Luciferase activity was measured in a 1.5-ml Eppendorf tube with the Promega Dual-Luciferases Reporter Assay kit (Promega E1980) according to manufacturer's protocols after transfection. Relative renilla luciferase activity was normalized to firefly luciferase activity. The assay was performed as previously described [29][30][31].

Tissue microarray, immunohistochemistry (IHC) and histological studies
The HCC tissue microarrays containing 61 HCC tissues and 35 normal tissues were purchased from Shanghai Outdo Biotech (Shanghai, China). The characteristics of the patients and their tumors were collected through review of medical records and pathologic reports. Informed consent with approval of the ethics committee of Taizhou Hospital of Zhejiang Province was obtained. All of the methods in this study were in accordance with the approved guidelines. Sections were stained with Masson's trichrome and H&E (haematoxylin and eosin) for histopathological examination. For immunohistochemistry, sections were subjected to antigen retrieval using microwave heating at 95°C in citrate buffer (pH = 6.0, for β-catenin) or boiling in Tris-EDTA buffer(pH = 9.0, for ITGB5). The indicated antibodies specific forITGB5 (1:100, abcam, #ab15459) and βcatenin (1:50, Proteintech Group, #51067-2-AP) were diluted according to the manufacturer's instructions.
The degrees of immunostaining were reviewed and scored by two independent observers. The proportion of the stained cells and the extent of the staining were used as criteria of evaluation. For each case, at least 1000 tumor cells were analyzed. For each sample, the proportion of ITGB5 and β-catenin -expressing cells varied from 0 to 100%, and the intensity of staining varied from weak to strong. One score was given according to the percentage of positive cells as:< 5% of the cells:1 point; 6-35% of the cells:2 point; 36-70% of the cells:3 point; > 70% of the cells: 4 point. Another score was given according to the intensity of staining as: negative staining: 1 point; weak staining (light yellow): 2 point; moderate staining (yellowish brown): 3 point; and strong staining (brown): 4 point. A final score was then calculated by multiple the above two scores. If the final score was equal or bigger than four, the protein expression in the tumor was considered high; otherwise, the protein expression in the tumor was considered low. The high expression of miR-185 in HCC tissues was defined to more than 2 fold higher than in normal tissues. On the contrary, we defined it low expression.

Animal experiments
Animal studies were carried out in accordance with the National Institute of Health's Guide for the Care and Use of Laboratory Animals, with the approval of the Animal Research Committee of Dalian Medical University. Male nude mice (4-6 weeks old, 18-20 g) were obtained from the SPF Laboratory Animal Centre of Dalian Medical University (Dalian, China). The mice were used for experiments after they had been acclimatized for 1 week. Stable MHCC-97 L cells (1 × 10 7 ) that were suspended in 200 μl of PBS were subcutaneously inoculated in mice. Five mice (n = 5) was used in each of the experiments. After 5 weeks, all animals were killed by cervical decapitation, the tumour weights were measured and the tumour tissues were excised aseptically. The protocol was approved by the Animal Care and Ethics Committee of Dalian Medical University.

Mass spectrometry
Immunoprecipitates were resolved on 12% SDS page. Gels were minimally stained with Coomassie Brilliant Blue to differentiate IgG bands. Each lane was then cut into three molecular weight regions. These bands were digested with 100 ng of trypsin overnight. LTQ Orbitrap Velos (Thermo Fisher Scientific) was run in a datadependent mode, where each sample was eluted in 5-30% acetonitrile gradient. Spectral data were then searched against the human protein RefSeq database in Proteome Discoverer1.3 Suites with Mascot software.

Statistical analysis
All results are shown as the mean ± S.D. of multiple independent experiments, not technical replicates. Detailed P values for each panel in the figures are stated in the corresponding legends. A Student's t-test, a Mann-Whitney test (for two group comparisons) or a Kruskal-Wallis oneway ANOVA followed by Dunn's multiple comparison tests (for more than two group comparisons) was used for statistical analyses. All statistical analyses were performed with GraphPad Prism 5 and SPSS 19.0 software. All statistical tests were two-sided, and P values < 0.05 were considered to be statistically significant.

ITGB5 promotes HCC tumorigenesis
To investigate the role of ITGB5 in HCC tumorigenesis, we used Huh-7 and MHCC-97 L cell lines and knocked down ITGB5 expression with 2 independent shRNAs. As shown in Fig. 1a, stable cell lines expressing these shRNAs showed significantly reduced ITGB5 levels and the percent of ITGB5 knockdown was nearly 70% in Huh-7 and 90% in MHCC-97 L. In the subsequent colony formation assays, the numbers of colonies formed by Huh-7 and MHCC-97 L cells with ITGB5 knockdown were remarkably reduced compared to control cells ( Fig. 1b and c). Furthermore, in transwell assays, the migratory capabilities of Huh-7 and MHCC-97 L cells depleted of ITGB5 were also apparently inhibited (Fig. 1d, e, and f ). We then turn to the opposite approach to investigate the effects of ITGB5 overexpression on HCC cells and generated Huh-7 cells stably overexpressing ITGB5 (Fig. 1g). Consistent with data from knockdown studies, Huh-7 cells overexpressing ITGB5 showed increased colony numbers in colony formation assays and enhanced migratory ability in transwell assays (Fig. 1h-k). To examine the involvement of ITGB5 expression in HCC tumorigenesis in vivo, we implanted MHCC-97 L cells stably expressing control shRNA or shRNA targeting ITGB5 into nude mice. As illustrated in Fig. 1l and m, the size and weight of xenograft tumours were significantly reduced by ITGB5 knockdown. Tumour tissues dissected were also subjected to immunohistochemistry analysis that confirmed the efficiency of ITGB5 depletion (Fig. 1n).

High expression of ITGB5 in HCC
To reveal the clinical relevance of ITGB5 expression in HCC, we collected 61 HCC specimens and 35 normal tissue samples adjacent to tumours. Through immunohistochemistry assays, we examined the expression levels of ITGB5 in these samples. Results showed that ITGB5 was significantly overexpressed in tumour tissues compared to control normal samples (Fig. 2b), and representative immunohistochemistry data are shown in Fig. 2a. The associations of ITGB5 expression with the age, gender, tumour size, and clinical stage of the patients were also analysed. Positive correlations of ITGB5 expression to patient age, tumour size, and the clinical stage were observed, which showed that ITGB5 was significantly overexpressed in patients with ages of over 50, with tumour sizes of greater than 5 cm, and those classified as in advanced stages (Fig. 2c-f ).

ITGB5 interacts with β-catenin and enhances its stability
To gain more mechanistic insights into the ITGB5 regulation of HCC tumorigenesis, we carried out mass spectrometry analysis to study the partners interacting with ITGB5 and identified β-catenin as a top candidate (Additional file 1: Figure S1A). To validate the interaction between β-catenin and ITGB5, we immunoprecipitated endogenous β-catenin and ITGB5 separately from Huh-7 and MHCC-97 L cells. In both HCC cell lines, β-catenin co-immunoprecipitated with ITGB5 and vice versa ( Fig. 3a and b). Interestingly, when we knocked down ITGB5 expression in Huh-7 and MHCC-97 L cells, the protein levels of β-catenin were significantly reduced, although its mRNA levels remained stable (Fig. 3c-f). On the contrary, the overexpression of ITGB5 in MHCC-97 L cells led to an upregulation of β-catenin without changing its mRNA levels ( Fig. 3g and h). These findings collectively suggest that ITGB5 regulates βcatenin levels at the post-transcriptional level. Therefore, we speculated that ITGB5 might affect the degradation of β-catenin protein. We treated Huh-7 and MHCC-97 L cells with the proteasome inhibitor MG132 to stop proteasomal protein degradation and observed the protein levels of β-catenin in these cells expressing control or ITGB5targeting shRNAs. As shown in Fig. 3i and j, MG132 treatment efficiently rescued the protein levels of β-catenin in Huh-7 and MHCC-97 L cells depleted of ITGB5. Furthermore, we treated MHCC-97 L cells stably expressing control or ITGB5-targeting shRNAs with cycloheximide to stop protein synthesis and then examined the turnover of β-catenin. It appeared that β-catenin was degraded at a much faster rate in cells with ITGB5 knockdown than in control shRNA-expressing cells (Fig. 3k). Meanwhile, the overexpression of ITGB5 in MHCC-97 L cells stabilized the protein levels of β-catenin under cycloheximide treatment conditions (Fig. 3l). Since the proteasomal degradation of cellular proteins is frequently mediated by ubiquitination, we measured the ubiquitination status of β-catenin in cells with ITGB5 knockdown or overexpression. As shown in Fig. 3m, a greatly enhanced ubiquitin signal was detected β-catenin immunoprecipitated from were detected by western blotting. Cell proliferation was examined by a colony formation assay. Data represent the mean ± SD of three independent experiments. ***p < 0.001 vs. control. d-f Cell migration was detected by a transwell assay. Data represent the mean ± SD of three independent experiments. **p < 0.01 and ***p < 0.001 vs. control. g ITGB5 was overexpressed in Huh7 cells using the lentivirus expressing vector pCDH, and the expression levels of ITGB5 were detected by western blotting. h-i Cell growth was examined by colony formation assays. Data represent the mean ± SD of three independent experiments. ***p < 0.001 vs. control. j-k Cell migration was detected by a transwell assay. Data represent the mean ± SD of three independent experiments. ***p < 0.001 vs. control. l-n The tumour-forming abilities of MHCC-97 L cells with or without ITGB5 knockdown were measured in vivo (l). The tumour weight (m) was assessed. The ITGB5 expression level was examined by immunohistochemistry (n) ITGB5 knockdown cells compared to that from controltreated cells. We also immunoprecipitated β-catenin from Huh-7 and MHCC-97 L cells with or without ITGB5 overexpression. Consistent with data from knockdown cells, in both HCC cell lines, the overexpression of ITGB5 greatly diminished the ubiquitination of β-catenin (Fig. 3n).

ITGB5 promotes HCC tumorigenesis via regulating β-catenin
To understand the functional role of this ITGB5-β-catenin interaction, we first evaluated the effect of ITGB5 on the transcriptional activity of b-catenin/TCF/LEF by using a TCF/LEF-responsive luciferase (TOP-FLASH) and nonresponsive (FOP-FLASH) reporter. We found that overexpression of ITGB5 induced an approximately three-fold higher transcription from the b-catenin/TCF/LEFdependent promoter (TOP-FLASH) and had little effect on FOP-FLASH (Fig. 4a). Whereas, inhibition of ITGB5 notably decreased the b-catenin/TCF/LEFdependent promoter activity (Fig. 4b). Next we analyzed the expression of the downstream target genes of β-catenin and found overexpression of ITGB5 elevated the protein and mRNA levels of Cyclin D1 and c-Myc ( Fig. 4c and d). Conversely, depletion of ITGB5 reduced the expression of Cyclin D1 and c-Myc ( Fig. 4e and f).
β-catenin has been indicted to play an important role in tumorigenesis. To investigate whether the effects of ITGB5 on HCC tumorigenesis were relied on the regulation of β-catenin. We stably overexpressed β-catenin in MHCC-97 L cells expressing control or ITGB5 targeting shRNAs. As shown in Fig. 4g, exogenous overexpression elevated β-catenin levels in both control MHCC-97 L cells and those depleted of ITGB5. Colony formation assays were performed using these β-catenin overexpressing cells, with results showing increased colony numbers with β-catenin overexpression despite of ITGB5 knockdown (Fig. 4h and i). Additionally, in transwell assays, β-catenin overexpression also rescued the defect of cell migration caused by ITGB5 depletion (Fig. 4j and k). We then used these β-catenin overexpressing cells together with corresponding parental ones to grow xenograft tumors in nude mice. Relative to MHCC-97 L control and ITGB5 knockdown cells, β-catenin overexpression led to significantly increased size and weight of tumors formed by MHCC-97 L cells with or without ITGB5 knockdown (Fig. 4l and m).

miR-185 regulates ITGB5 expression
Having identified β-catenin as the downstream target of ITGB5 in HCC tumorigenesis, we sought to investigate upstream regulators of ITGB5 expression. Through combined analyses from web-based miRNA resources (miRanda and TargetScan), we identified 6 top candidate miRNAs that might regulate ITGB5 expression, including miR-122, miR-486-5p, miR-139, miR-340, miR-271, and miR-185 (Fig. 5a). To examine whether these candidates regulate the expression of ITGB5, we constructed the 3′UTR region of ITGB5 onto the luciferase reporter system and carried out luciferase assays. The results revealed that both miR-271 and miR-185 suppressed the expression of luciferase, with miR-185 showing the strongest inhibitory effect (Fig. 5b). To examine the influence of miR-271 and miR-185 on endogenous ITGB5 control. c-f Correlation analyses of ITGB5 expression with the patient age, gender, tumour size and clinical stage in the 61 HCC tissues. Data represent the mean ± SD of three independent experiments. *p < 0.05, **p < 0.01 and ***p < 0.001 vs. control levels, we treated Huh-7 cells with the corresponding miRNA mimics. Both mimics repressed the expression of ITGB5 protein (Fig. 5c). As miR-185 showed the strongest inhibition on ITGB5 levels, we picked this miRNA for further studies. Consistently, in MHCC-97 L cells, the miR-185 mimic also efficiently suppressed ITGB5 levels (Fig. 5d). Furthermore, treatment with the miR-185 inhibitor resulted in elevated ITGB5 expression in both Huh-7 and MHCC-97 L cells (Fig. 5e and f ).
We then investigated the binding site of miR-185 to the 3′UTR region of ITGB5. In doing this, we split the 3′UTR region of ITGB5 into two fragments (0-500 and 500-1166) and constructed these fragments together with the full length region (0-1166) onto pSICHEK2 vectors (Fig. 5g). Luciferase activity assays were performed using these constructs with or without miR-185 mimic treatment. The results showed that miR-185 mimic significantly inhibited the luciferase activities in the full length and 0-500 fragment groups, but the activity was not altered in the 500-1166 fragment group (Fig. 5h). These observations suggest that the binding site of miR-185 to the 3′UTR region of ITGB5 is located Fig. 3 ITGB5 regulates β-catenin levels. a-b MHCC-97 L and Huh-7cell lysates were subjected to immunoprecipitation with control IgG, anti-ITGB5 (a) or anti-β-catenin (b) antibodies. The immunoprecipitates were then detected using the indicated antibodies. c-h Protein or mRNA levels of β-catenin in ITGB5-delepted or ITGB5-overexpressed stable cell lines were analysed by western blotting and q-RT-PCR. i-j Huh-7 and MHCC-97 L cells with or without ITGB5 knockdown were treated with MG132, and then, the protein levels of β-catenin were examined by western blotting. k-l ITGB5-delepted or ITGB5-overexpressed stable cells were treated with CHX for the indicated times and then, the protein levels of β-catenin were examined by western blotting. m-n ITGB5-delepted or ITGB5-overexpressed stable cell lines were transfected with HA-Ub and then treated with MG132 as indicated. Cell lysates were subjected to immunoprecipitation with anti-β-catenin antibody. The immunoprecipitates were then detected using the indicated antibodies Fig. 4 ITGB5 promotes HCC tumorigenesis via regulating β-catenin. a The increase of b-catenin/TCF/LEF-mediated transcription by ITGB5 in MHCC-97 L and Huh-7 cells with or without ITGB5 overexpression was monitored. The luciferase activities of the reporter TOP-FLASH or non-responsive control reporter FOP-FLASH were measured. The transcriptional activity was quantified via measuring the relative firefly luciferase activity normalized to the activity of Renilla luciferase. Data represent the mean ± SD of three independent experiments. ***p < 0.001 vs. control. b The decrease of b-catenin/TCF/LEF-mediated transcription by ITGB5 in MHCC-97 L and Huh-7 cells was monitored. The luciferase activities of the reporter TOP-FLASH or non-responsive control reporter FOP-FLASH were measured. The transcriptional activity was quantified via measuring the relative firefly luciferase activity normalized to the activity of Renilla luciferase. Data represent the mean ± SD of three independent experiments. **p < 0.01 and ***p < 0.001vs. control. c-d The protein and mRNA levels of Cyclin D1 and c-Myc were measured in MHCC-97 L and Huh-7 cells with or without ITGB5 overexpression. Data represent the mean ± SD of three independent experiments. **p < 0.01 and ***p < 0.001vs. control. e-f The protein and mRNA levels of Cyclin D1 and c-Myc were measured in MHCC-97 L and Huh-7 cells with or without ITGB5 knockdown. Data represent the mean ± SD of three independent experiments. **p < 0.01 and ***p < 0.001vs. control. g ITGB5 was knocked down in MHCC-97 L cells with or without β-catenin overexpression. The protein levels of ITGB5 and β-catenin were analysed by Western blotting. h-k cell proliferation was examined by colony formation assay(h and i). Cell migration was detected by transwell assay (j-k). Data represent the mean ± SD of three independent experiments. **p < 0.01 vs. control. l-m Representative images of tumor from the experimental groups, as indicated and graphic representation of tumor weight within the sequence 0-500 of the ITGB5 3′UTR region. By the comparison of sequence similarity between miR-185 and the sequence 0-500 of the ITGB5 3′UTR region, we speculated the binding site of miR-185 and generated the corresponding binding mutant for miR-185 on the pSICHEK2-luciferase construct (Fig. 5i). In subsequent luciferase activity assays, the binding mutations conferred this construct resistance to miR-185 mimic-mediated inhibition of the luciferase activity ( Fig. 5j and k). On the contary, inhibition of endogenus miR-185 elevated the luciferase activity (Fig. 5l).

miR-185 regulates β-catenin through ITGB5
We next examined whether miR-185 regulates the levels of β-catenin. When Huh-7 and MHCC-97 L cells were treated with the miR-185 mimic, a concurrent decrease of ITGB5, β-catenin protein levels and the transcript activity of β-catenin was observed (Fig. 6a-c). In these assays, the miR-185 mimic dramatically upregulated the levels of miR-185 (Additional file 1: Figure S1B). Furthermore, treatment with the miR-185 inhibitor in Huh-7 and MHCC-97 L cells increased expression of both ITGB5 and β-catenin, leading to the transcription activity of β-catenin The 3′UTR of ITGB5 was constructed into a pSICHECK2 vector and was cotransfected with miR-122, miR-486-5P, miR-139,miR-271 and miR-185 individually. The luciferase activities were measured. Data represent the mean ± SD of three independent experiments. *p < 0.05 and **p < 0.01 vs. control. c-d miR-185 and miR-271 were introduced into Huh-7 cells. The protein levels of ITGB5 were examined by western blotting. e-f miR-185 inhibitor was transfected into Huh-7 and MHCC-97 L cells. The protein levels of ITGB5 were examined by western blotting. g Schematic illustration of pSICHECK2-based reporter constructs used in luciferase assays to examine the 3′UTR activity of ITGB5. The different truncations were named F1, F2 and F3 as indicated. h The different truncations were respectively transfected into MHCC-97 L cells with or without miR-185 overexpression. The luciferase activity was measured. **p < 0.01 and ***p < 0.001 vs. control. i Sequences of ITGB5 3′UTR and the potential miR-185 binding site at the 3′ UTR of ITGB5 as well as nucleotides mutated in the 3′ UTR of ITGB5, where the red indicates the mutated region. j-k The wild type 3′UTR (WT) and mutated 3′UTR (MUT) of ITGB5 were transfected into Huh-7 and MHCC-97 L cells with or without miR-185 overexpression. The luciferase activities were measured.**p < 0.01 and ***p < 0.001 vs. control. l The wild type 3′UTR (WT) of ITGB5 was transfected into Huh-7 and MHCC-97 L cells with or without miR-185 inhibition. The luciferase activities were measured.**p < 0.01 vs. control upregulation. However, when ITGB5 expression was silenced, the miR-185 inhibitor failed to increase the levels of β-catenin in Huh-7 and MHCC-97 L cells, suggesting an ITGB5-dependent mechanism of the miR-185 regulation of β-catenin expression (Fig. 6d, e, and f). In these assays, the miR-185 inhibitor also efficiently suppressed the expression of miR-185 (Additional file 1: Figures S1C and  S1D). In the following colony formation assays, the miR-185 inhibitor increased the number of colonies. However, when the expression of ITGB5 or β-catenin was knocked down, the stimulatory effect of the miR-185 inhibitor was compromised ( Fig. 6g and h). Similarly, in transwell assays, the miR-185 inhibitor promoted cell migration, in which process the expression of ITGB5 and β-catenin was also required ( Fig. 6i and j). These results indicate that miR-185 regulates the growth and migration of HCC cells through effects on ITGB5 and β-catenin.

Decreased miR-185 levels in HCC
We compared the expression levels of miR-185 in HCC specimens and adjacent normal tissues. As shown in Fig. 7a, the expression of miR-185 was significantly decreased in HCC tissues compared to in normal adjacent samples. Correspondingly, the immunohistochemistry analyses of these Fig. 6 miR-185 regulates β-catenin through ITGB5. a-b miR-185 was introduced into MHCC-97 L and Huh-7 cells. Cell lysates were analysed using the indicated antibodies (a). The β-catenin protein levels were quantitatively measured (b). c The inhibition of b-catenin/TCF/LEF-mediated transcription by miR-185 in MHCC-97 L and Huh-7 cells was monitored. Data represent the mean ± SD of three independent experiments. **p < 0.01vs. control. d-f miR-185 inhibitor was transfected into MHCC-97 L and Huh-7 cells with or without ITGB5 knockdown. Cell lysates were examined by western blotting using the indicated antibodies (d). The β-catenin protein levels were quantitatively measured (f). The effect of miR-185 inhibitor on b-catenin/TCF/LEF-mediated transcription was measured. Data represent the mean ± SD of three independent experiments. **p < 0.01 and ***p < 0.001 vs. control. g-h miR-185 inhibitor was transfected into MHCC-97 L and Huh-7 with or without ITGB5 or β-catenin knockdown. Cell proliferation was examined by colony formation assay. Data represent the mean ± SD of three independent experiments. ***p < 0.001 vs. control. i-j Cell migration was analysed by transwell assay. Data represent the mean ± SD of three independent experiments. **p < 0.01 vs. control HCC samples showed elevated expressions of ITGB5 and β-catenin (Fig. 7b, c, and d). With these clinical data, we examined the association of miR-185 expression with the level of ITGB5 and observed significant reverse correlation (Fig. 7e). Meanwhile, the expression of β-catenin is positively related to the level of ITGB5 (Fig. 7f).

Discussion
HCC is the third most common cause of cancer mortality worldwide, and tumour recurrence due to de novo tumours and dissemination are the predominant causes leading to the high mortality [1,4,32,33].The present study demonstrated that ITGB5 is a critical regulator of the tumorigenic potential of HCC, not only facilitating cancer cell growth but also promoting cell migration. In clinical data, we found that ITGB5 expression was frequently increased in tumour tissues compared to adjacent normal specimens correlating with older patients, a large tumour size and an advanced clinical stage. In functional experiments, the depletion of ITGB5 by shRNA markedly inhibits HCC cell clone formation, which reflects the ITGB5 function in the replicative potential of cancer cells. Importantly, the knock down of ITGB5 effectively attenuates the migratory ability of HCC cells and vice versa.
Canonically, integrins affect intracellular signalling by activating integrin-associated kinases due to lack of enzymatic activity [8,9,34], and mechanistic studies of ITGB5 mainly focus on activating Src-FAK and MEK-ERK signalling pathways, which profoundly contribute to the tumorigenic capacity of the malignancies [19]. In our present experiments, results show the novel role of ITGB5 in HCC, which interacted with β-catenin and upregulated β-catenin expression. When ITGB5 had dysregulated expression, the protein level of β-catenin was subsequently altered in HCC cells, which was consistent with the clinical specimen for immunohistochemistry correlation analyses. Interestingly, we noticed that the increase of β-catenin by ITGB5 is in a transcriptionindependent manner, and the mRNA level of β-catenin did not change when ITGB5 was overexpressed or knocked down, suggesting that ITGB5 is likely to control β-catenin in the phase of protein translation and/or degradation. A series of experiments revealed that the regulation of ITGB5 on β-catenin mainly relied on inhibiting its degradation. Using a protein stability assay and immunoprecipitation assays, we confirmed that ITGB5depletion triggers the strong ubiquitylation of β-catenin. However, the in-depth molecular basis by which ITGB5 regulates the ubiquitylation of β-catenin is still unclear, Fig. 7 Decreased miR-185 levels in HCC are associated with ITGB5 protein expression. a Expression levels of miR-185 in HCC tissues and adjacent normal tissues were analysed by Q-PCR. Data represent the mean ± SD of three independent experiments. **p < 0.01 vs. control. b-d The expression levels of β-catenin and ITGB5 were examined by immunohistochemistry. Data represent the mean ± SD of three independent experiments. ***p < 0.001 vs. control. e Correlation analyses of miR-185 levels and ITGB5 expression in 30 HCC tissues. f Correlation analyses of β-catenin and ITGB5 expression in 30 HCC tissues