RETRACTED ARTICLE: Integrin-β5, a miR-185-targeted gene, promotes hepatocellular carcinoma tumorigenesis by regulating β-catenin stability

The human HCC cell line Huh-7 was obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China), and the MHCC-97 Lcell line was obtained from Zhongshan Hospital, Fudan University (Shanghai, China). These cell lines were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal bovine serum (FBS), 1% penicillin, streptomycin and 0.1% Savelt ™. They were maintained in a humidified atmosphere with 5% CO₂ at 37 °C. The following antibodies and reagents were used: the antibody to ITGB5 (1:500, Abcam, #ab15459), beta-catenin (1:1000, Proteintech Group, #51067-2-AP), beta-actin (1:1000, Proteintech Group, #60008-1-Ig), ubiquitin (1:1000,Cell Signalling Technology,#3393), Cyclin D1 (1:500, Cell Signaling Technology,#2978), c-Myc (1:500, Cell Signaling Technology,#13,987), MG132 (Calbiochem), puromycin (Mediatech), cycloheximide (CHX) (Inalco Spa Milano Italy), Savelt™ (Hanbio Co.LTD 1:1000) and polybrene (Sigma).

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 MNa₂P₂O₇, 1 mM β-glycerol phosphate, 1 mM Na₃VO₄, 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.

Huh-7 and MHCC-97 L cells with the treatment as indicated (1 × 10³ cells per well) were plated into 6 well plates and cultured at 37 °C equipped with 5% CO₂. 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.

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.

RNA interference was performed as previously described [26, 27]. The shRNA was purchased from Sigma. The various targeting sequences were as follows:ITGB5-1,5-CCCGCTATGAAATGGCTTCAA-3; ITGB5-2, 5-GCATCCAACCAGATGGACTAT-3;and β-catenin, 5-AGGTGCTATCTGTCTGCTCTA-3. The primers 5′- gcGAATTCATGCCGCGGGCCCCGGCG.

-3′ and 5′- gcGGATCCTCAGTCCACAGTGCCATT-3′ were used to generate plasmids encoding full-length 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.

The total RNA was isolated using Trizol (Invitrogen). One microgram of total RNA was used to synthesize cDNA using the PrimeScript™ RT reagent kit (Takara, RR047A) according to the manufacturer’s instructions. The primers were as follows:ITGB5up: 5-GTCTGCTAATCCACCCAAAATG-3,dn:5- TCTCTATCTCACCTCCACAGC-3;and actin up: 5-GACCTGACTGACTACCTCATGAAGAT-3, dn: 5-GTCACACTTCATGATGGAGTTGAAGG-3.The primers for mature miR-185 were purchased from Takara.

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.

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].

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 [29‐31].

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 β-catenin (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 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 × 10⁷) 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.

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 data-dependent 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.

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 one-way 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.

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).

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).

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 β-catenin 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 ITGB5-targeting 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 ITGB5 knockdown cells compared to that from control-treated 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).

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 non-responsive (FOP-FLASH) reporter. We found that overexpression of ITGB5 induced an approximately three-fold higher transcription from the b-catenin/TCF/LEF-dependent promoter (TOP-FLASH) and had little effect on FOP-FLASH (Fig. 4a). Whereas, inhibition of ITGB5 notably decreased the b-catenin/TCF/LEF-dependent 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).

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 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 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).

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 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.

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 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).