Introduction

Hepatocellular carcinoma (HCC) comprises 75–85% of all primary liver cancers and is estimated to be one of the most common malignancies worldwide. Although the incidence and mortality have been decreasing in Asia, liver cancer is still the third-highest cause of cancer-related death [1]. With an increasing number of studies focusing on molecular mechanisms, more pathways and signaling factors related to carcinogenesis and metastasis have been discovered [2,3,4,5]. However, many problems related to genetic and epigenetic alterations in hepatocarcinogenesis remain unknown, and further research is required to better understand this issue.

SET and MYND domain-containing protein 3 (SMYD3), a member of the SET and MYND domain family of lysine methyltransferases, is overexpressed in various types of human cancers, including prostate cancer, breast cancer, colon cancer, colorectal cancer, ovarian cancer, and HCC [6,7,8,9,10]. Within SMYD3, the SET domain meditates the histone methylation of lysine residues, while the MYND domain contains a zinc finger motif that binds proline-rich regions and facilitates protein–protein interactions [11, 12]. Overexpression of SMYD3 plays an important role in the proliferation, adhesion, invasion, and migration of cancer cells, whereas decreases in SMYD3 expression inhibit cell growth, migration, invasion, and apoptosis [13,3E), and S1PR1, RIN3, and RTKN2 were located at the 1 kb promoter region (−1 kb to the TSS) (Fig. 3F, Supplementary Fig. 3 and 4). As shown in the ChIP landscape, SMYD3 binding peaks around the TSS regions of S1PR1 were higher in SMYD3-overexpressing cells. Then, we analyzed the relative mRNA expression levels of these 11 genes in HepG2/LO2, SMMC-7721/Chang live, Sk-Hep1-SMYD3/Sk-Hep1-Ctrl, and Huh7-SMYD3/Huh7-Ctrl cells. The expression levels of S1PR1, MXRA8, and RTKN2, especially S1PR1, were significantly upregulated in HCC and SMYD3-overexpressing cell lines (Fig. 3G).

Based on these results, we selected S1PR1 as the candidate gene for subsequent experiments. qRT-PCR and western blotting were performed to confirm whether SMYD3 regulates S1PR1 expression, and the results showed that SMYD3 overexpression upregulated the S1PR1 mRNA and protein, whereas knockdown of SMYD3 produced the opposite outcomes (Fig. 3H, I). These results strongly support the concept that S1PR1 is regulated by SMYD3 in HCC cells.

SMYD3 regulated S1PR1 via histone lysine methylation of the S1PR1 promoter

Based on previous studies showing that the SET domain of SMYD3, a histone lysine methyltransferase, affects downstream gene expression [20, 21], we hypothesized that SMYD3 meditated lysine methylation of the S1PR1 promoter sequence to regulate S1PR1 expression in HCC cells. We investigated the specific DNA sequence that bound the motifs of SMYD3 to verify this assumption and identified three –CCCTCC– sites in the S1PR1 promoter (Fig. 4A) [20, 22]. Luciferase assays revealed increased luciferase activity in SMYD3-transfected cells in a dose-dependent manner, while promoter activities decreased in response to SMYD3 knockdown in HCC cells (Fig. 4B). Then, we mutated the three SMYD3-binding sites via site-directed mutagenesis. Compared with the control wild-type sequence, mutations at sites 2 and 3 reduced the luciferase activity, especially when site 3 was mutated (Fig. 4C, left panel). Furthermore, Huh7 cells transfected with the SMYD3 overexpression vector showed increased luciferase activity, whereas the same cells with mutations in SMYD3-binding sites showed reduced activity (Fig. 4C, right panel). Moreover, we carried out ChIP-PCR with antibodies recognizing SMYD3 and H3K4me3 at the above 3 sites (Fig. 4D). As shown in Fig. 4E, the S1PR1 promoter was indeed occupied by the SMYD3 protein, and this binding was enhanced in the presence of ectopically expressed SMYD3. Correspondingly, knockdown of SMYD3 led to its downregulation at the S1PR1 promoter region (Fig. 4F). We concluded that SMYD3 indirectly affected S1PR1 transcription by meditating the methylation of the S1PR1 promoter at site 3 in HCC cells. These results support our hypothesis that SMYD3 regulates S1PR1 expression by meditating histone lysine methylation at the S1PR1 promoter.

Fig. 4: SMYD3 regulated S1PR1 via methylation of the S1PR1 promoter.
figure 4

A The specific SMYD3 DNA binding site (–CCCTCC-) in the S1PR1 promoter. B Luciferase assays with the S1PR1 promoter were performed in HCC cells transfected with different amounts of pCDNA-SMYD3 vector in SK-Hep1 cells or cells with stable overexpression (Huh7 cells) or knockdown (Hep3B cells) of SMYD3. C SK-Hep1 cells were cotransfected with pGL-S1PR1-WT, pGL-S1PR1-MT1, pGL-S1PR1-MT2, pGL-S1PR1-MT3, and SMYD3 (pcDNA3.1-SMYD3) (left panel) and Huh7-Ctrl and Huh7-SMYD3 cells were cotransfected with pGL-S1PR1-WT, MT1, MT2 and MT3 (right panel). The promoter activities were measured by performing luciferase reporter assays. D The examined primer position for ChIP-PCR assays at the S1PR1 locus was indicated. E, F The ChIP assay was performed in SK-Hep1-Ctrl and SK-Hep1-SMYD3 cells (E) and Hep3B-sh-NC and Hep3B-sh-SMYD3 (F) cells using specific antibodies. The immunoprecipitated DNA was measured with real-time PCR using the primers listed above. Data are presented as the means ± SD of three separate experiments, each performed in triplicate. *P < 0.05; **P < 0.01.

Full size image

SMYD3 regulated the downstream signaling pathway of S1PR1

We performed IHC experiments on paired tissues from 148 patients with HCC to detect the differential expression of S1PR1 between HCC and adjacent normal tissues and found abundantly higher IHC scores in HCC tissues than in the paired normal tissues (Fig. 5A, B). Furthermore, higher S1PR1 expression was correlated with shorter survival (p = 0.022) (Fig. 5C). All these results support the hypothesis that S1PR1 overexpression indicates a poor prognosis for patients with HCC. Moreover, S1PR1 expression was positively correlated with SMYD3 expression, consistent with our previous results showing that S1PR1 is the downstream effector of SMYD3 (Fig. 5D).

Fig. 5: SMYD3 regulated its downstream signaling pathways via S1PR1.
figure 5

A IHC staining was performed on paired tumor and adjacent noncancerous tissues from 148 patients with HCC. B Statistical analysis of the IHC scores between HCC and adjacent normal tissues. C Kaplan-Meier analysis of overall survival in patients with HCC stratified according to the S1PR1 expression status. The median expression level of S1PR1 was used as the cut-off. D Clinical samples from the same set of tissues were analyzed in the two groups as indicated. The correlation between SMYD3 and S1PR1 expression was assessed by performing a linear regression analysis. E Western blot analysis of SMYD3, S1PR1, p-AKT, AKT, p-STAT3, STAT3, p-Erk1/2, and Erk1/2 expression levels in control cells and cells overexpressing S1PR1 and cells overexpressing SMYD3 combined with knockdown of SMYD3 or S1PR1. Images are representative of two experiments. Bar = 200 μm. *P < 0.05; **P < 0.01.

Full size image

Recent studies have suggested that the activation of S1PR1 is involved in regulating various aggressive biological phenotypes of tumors by modulating its downstream signaling pathways [23,24,25]. We investigated changes in the expression of constituents of those signaling pathways, including the AKT, STAT3, and MAPK pathways, using western blotting to validate whether the downstream signaling pathways of SMYD3 were regulated by S1PR1. Overexpression of SMYD3 or S1PR1 increased the levels of phosphorylated AKT, STAT3, and Erk1/2, while knockdown of SMYD3 or S1PR1 reduced p-AKT, p-STAT3, and p-Erk1/2 levels. Furthermore, knockdown of S1PR1 ablated the increases in the levels of p-AKT, p-STAT3, and p-Erk1/2 induced by SMYD3, and when cells with SMYD3 knockdown were transfected with an S1PR1-overexpressing plasmid, the decreases in p-AKT, p-STAT3, and p-Erk1/2 levels were rescued (Fig. 5E). Based on these results, SMYD3 increased AKT, STAT3, and Erk1/2 activation by promoting S1PR1 expression.

SMYD3 promoted HCC progression through S1PR1

We performed CCK8, colony formation, cell cycle, apoptosis, and transwell assays to explore whether SMYD3-mediated S1PR1 activity was involved in cancer cell development and progression. SMYD3 or S1PR1 was stably or transiently overexpressed or knocked down in HCC cells. The proliferation capacities of SMYD3-overexpressing cells were abrogated by S1PR1 knockdown (Fig. 6A, C and D). SMYD3 knockdown reduced the proliferation and migration capacities, but overexpression of S1PR1 in cells with SMYD3 knockdown reversed the effects compared to those in the control cells (Fig. 6A, B and Supplementary Fig. 5).

Fig. 6: SMYD3 promoted HCC progression via S1PR1.
figure 6

CCK8 (A), colony formation (B), cell cycle (C), and apoptosis (D) assays were performed to evaluate cell proliferation and migration in response to the overexpression or knockdown of SMYD3 combined with the overexpression or knockdown of S1PR1 in Huh7, SK-Hep1, and Hep3B cells. Data are presented as the means ± SD of three separate experiments, each of which was performed in triplicate. E Nude mice were subcutaneously inoculated with Huh7 cells transfected with control, SMYD3 overexpression, SMYD3 shRNA, S1PR1 overexpression, and SMYD3 shRNA combined with S1PR1 overexpression vectors. Photographs show mice bearing subcutaneous tumors from each group (upper panel) or the dissected tumors (lower panel) (n = 5). F Tumor volumes were measured and recorded every 3 days, and tumor growth curves were created for each group. G The weight of the local tumors from each group was measured. H Western blots showing SMYD3, S1PR1, p-AKT, AKT, p-STAT3, STAT3, p-Erk1/2, and Erk1/2 expression levels in Huh7 tumors. Bar = 50 μm. The image is representative of two experiments. *P < 0.05; **P < 0.01.

Full size image

We assessed the effects of altered expression, including SMYD3 knockdown, SMYD3 overexpression, S1PR1 overexpression, and the combination of SMYD3 knockdown and S1PR1 overexpression, on the growth of local tumors in nude mice to test whether SMYD3 promotes HCC progression via S1PR1 in vivo (Fig. 6E). The SMYD3-overexpressing and S1PR1-overexpressing groups showed significantly increased tumor growth compared to the control group. In contrast, SMYD3 knockdown significantly suppressed tumor growth and reduced tumor proliferation; these effects were largely reversed by S1PR1 overexpression (Fig. 6F, G). In addition, the western blot results from excised tumors showed that SMYD3 regulates AKT, STAT3, and MAPK signaling pathways through S1PR1 in vivo (Fig. 6H). Our experiments revealed that S1PR1 is a downstream gene of SMYD3 and that SMYD3 regulates the expression of S1PR1 to facilitate carcinogenesis in HCC cells.

Discussion

SMYD3 was originally described as an H3K4 methyltransferase that regulates the transcriptional activities of downstream genes involved in cancer progression [20]. Furthermore, a large amount of evidence indicates critical roles for SMYD3 in the proliferation, invasion, and migration of different tumor cells [13, 14, 26,27,28]. We investigated venous invasion, pTNM stage, and overall survival in HCC samples from a SMYD-positive group and SMYD3-negative group to confirm the correlations between SMYD3 expression and the clinical characteristics and pathological parameters in patients with HCC. The results confirmed that SMYD3 overexpression is a risk factor in patients with HCC, consistent with other studies [29,30,31]. We established SMYD3-overexpressing and SMYD3-knockdown vectors and transfected them into HCC cells to investigate changes in biological phenotypes, such as cell proliferation, migration, and colony formation. SMYD3 dramatically facilitated the development and progression of HCC, and we concluded that SMYD3 is involved in the aggressive behaviors of HCC and plays a crucial role in determining the prognosis of patients.

SMYD3 methylates not only histone H3 at lysine 4 but also histone H3 at lysine 9, histone H4 at lysine 5, and histone H4 at lysine 20 [13, 18, 32]; these histone lysine methylation events meditated by the SET domain of SMYD3 exert significant effects on cancer development and progression. As shown in previous studies, aberrant histone lysine methylation, including monomethylation, dimethylation, and trimethylation, is often associated with promoters and enhancers of oncogenes in cancer cells [33], which provides strong evidence supporting the hypothesis that SMYD3 promotes carcinogenesis via histone lysine methylation. In addition, the SET domain has been shown to catalyze the trimethylation of H3K4, which is associated with increased transcription of target genes [34, 35]. Based on these studies, we used ChIP-seq to investigate the genes that are highly overexpressed in HCC and combined those results with SMYD3 target genes to explore whether H3K4me3 of downstream genes is a mechanism by which SMYD3 promotes the carcinogenesis of HCC. From the identified genes, we selected S1PR1 for further study due to its high expression in HCC.

Some studies have reported a critical role for S1PR1 in cancer biology [23, 36, 37]. However, the mechanism by which SMYD3 mediates S1PR1 expression has not been elucidated. We verified that the expression and function of S1PR1 in HCC were regulated by SMYD3. The results implied that S1PR1 is the downstream gene of SMYD3. Furthermore, ChIP-PCR indicated that a critical site in the S1PR1 promoter exhibited a high level of H3K4me3 when targeted by SMYD3. At the same time, luciferase assays revealed that mutating these sites reduced the activity of SYMD3. Therefore, SMYD3 methylates lysines in histones at the S1PR1 promoter to regulate S1PR1 expression.

We performed western blotting to detect the levels of p-AKT, p-STAT3, and p-Erk1/2 in different treatment groups compared with those in the control group and to determine whether the downstream signaling pathways of S1PR1 are regulated by SMYD3. As expected, signaling pathways mediated by S1PR1 were upregulated by SMYD3 overexpression, whereas SMYD3 knockdown downregulated their levels. In vivo experiments were conducted to compare the tumor size in the treatment groups and the control group. According to the results, S1PR1 is involved in SMYD3-related HCC progression.

Collectively, these studies support the hypothesis that SMYD3 plays a crucial function in HCC progression, partially by mediating histone methylation at the downstream gene S1PR1, which affects critical signaling pathways associated with the carcinogenesis and progression of HCC.

Patients and methods

Patient samples

Eighty fresh HCC tissues and corresponding adjacent noncancerous liver tissues (more than 2 cm from the margin) were obtained from patients undergoing liver resection at Sun-Yat-Sen Memorial Hospital of Sun-Yat-Sen University between January 2008 and December 2012. Another 148 pairs of paraffin-embedded HCC samples and adjacent noncancerous tissues were obtained from January 1999 to December 2007. Eleven fresh normal liver samples and 10 paraffin-embedded normal liver samples were obtained from the specimen bank of the Department of Hepatobiliary Surgery at Sun-Yat-Sen Memorial Hospital. The use of clinical specimens was approved by the Ethical Institutional Review Board of the Sun-Yat-Sen Memorial Hospital.

HCC cell lines

HEK293T cells and the hepatoma cell lines HepG2, Huh7, SK-hep1, PLC/PRF/5 (derived from HBV-infected liver), and Hep3B (derived from HBV-infected liver) were purchased from the Chinese Academy of Sciences Cell Bank of Type Culture Collection (CBTCCCAS). LO2, SMMC-7721, and HepG2.2.15 (HepG2-derivative with the integration of the HBV genome) cells [38, 39] were purchased from BioHermes. All cell lines had guaranteed authenticity through short tandem repeat profiling and comparison to DNA profiles of known cell lines. Cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and incubated at 37 °C in an atmosphere containing 5% CO2.

RNA extraction, reverse transcription, and quantitative real-time polymerase chain reaction (qRT-PCR)

Total RNA was extracted using TRIzol reagent according to the manufacturer’s instructions (Invitrogen, USA). Complementary DNA was synthesized from total RNA using reverse transcriptase and random primers. Quantitative real-time PCR was performed in the GoTaq® qPCR System (Promega, Bei**g, China). The gene-specific primers are as follows: SMYD3, F 5′-GTCTTCAAACTTATGGATGGAGC-3′, 5′-GGCATCCTGTATTTCTTCTCTCA-3′; S1PR1, F 5′-CAGCAAATCGGACAATTCCT-3′; R 5′-GCCAGCGACCAAGTAAAGAG-3′; β-actin, F 5′-CTCCCTGGAGAAGAGCTACG-3′, R 5′-ACAGGACTCCATGCCCAG-3′; and GAPDH, F 5′-GAAGGTGAAGGTCGGAGTCAACG-3′; R 5′-TGCCATGGGTGGAATCATATTGG-3′.

Construction of tissue microarrays and IHC

Tissue microarray blocks consisting of 148 paired HCC samples and 10 normal liver samples were constructed using a tissue microarray (Quick-Ray, UNITMA). IHC was performed to detect the expression of the SMYD3 and S1PR1 proteins in the tissue microarray blocks using IHC kits (Biohao Biotechnology, Guangzhou). The immunostaining was assessed by counting the total and positively stained cells at a magnification of ×40 in at least randomly selected 10 fields. The staining intensity and extent of the stained area were graded using the semiquantitative scoring system described below. For the staining intensity of the cytoplasm, no staining = 0; weak staining = 1; moderate staining = 2; and strong staining = 3; for the extent of stained cells, 0% = 0, 1 to 20% = 1, 21 to 50% = 2, 51 to 80% = 3, and 81 to 100% = 4. The final immunoreactive score was determined by multiplying the intensity score by the extent of stained cells (range 0 to 12) [40].

Stable overexpression or knockdown of SMYD3 and S1PR1 in HCC Cells

The open reading frames of SMYD3 and S1PR1 were cloned into the retroviral vector pMSCV-eGFP-Puro according to the manufacturer’s instructions. The shRNAs targeting SMYD3 and S1PR1 were inserted into the lentiviral vector pGC-eGFP-Hygromycin (GV544); both constructs were generated by Genechem Company. The shRNA sequences of SMYD3 and S1PR1 are as follows: sh-SMYD3, 5′-GGATGGAGCACCTTCAGAATC-3′ and sh-S1PR1, 5′-CTGCTCAAGACCGTAATTAT-3′. Production and purification of the lentivirus and construction of the stably infected cell lines were performed as previously described [41].

Proliferation assay

The MTT assay was conducted using a cell proliferation kit according to the manufacturer’s instructions. Briefly, 400 HCC cells in 200 μl of complete culture medium were added to each well of a 96-well plate and cultured at 37 °C in an atmosphere containing 5% CO2. Culture medium alone served as the negative control. Then, the MTT solution was added 24, 48, 72, 96, 120, and 144 h after seeding. The culture medium was removed, and 150 μl of DMSO was added and incubated with shaking for 4 h. The optical density value was detected at 490 nm, and a cell growth curve was drawn. HCC cells were divided into six pairs and tested three times.

Flow cytometry assay

A total of 1 × 106 cells were seeded into 6-well plates in a DMEM medium. The cells were harvested and stained with annexin V-FITC and propidium iodide (PI) 48 h later according to the manufacturer’s instructions. The cellular apoptotic rates were evaluated using a FACS Verse™ flow cytometer (Becton Dickinson, CA, USA). Cells for cell cycle analysis were resuspended in 200 ml PBS, fixed with 70% ice-cold ethanol overnight, and stained with PI. The cell growth phase was detected by the FACSVerse™ flow cytometer.

Cell migration assay

The wound-healing test was conducted as follows: 5 × 105 cells in high-glucose DMEM supplemented with 10% FBS were seeded into 6-well plates and incubated overnight at 37 °C in an atmosphere containing 5% CO2. Pipette tips were used to scratch the monolayers, which were then washed with phosphate-buffered saline. A serum-free culture medium was added and incubated for another 48 h. Images were recorded at 0, 12, and 24 h. Transwell assays were conducted as follows: 1 × 105 cells were suspended in a serum-free culture medium and added to the upper chamber of a transwell cell migration apparatus (Corning). The lower chamber contained a culture medium supplemented with 10% FBS. Cells were incubated for 24 h at 37 °C in an atmosphere containing 5% CO2. Cells that migrated through the membrane pores to the lower surface of the membrane were fixed with methanol and stained with crystal violet. Then, the cells were visualized and counted under a microscope.

Colony formation assay

The colony formation assay was conducted as follows: 1 × 103 HCC cells stably transfected with SMYD3 or control vectors were seeded into 35-mm dishes and cultured at 37 °C in a humidified atmosphere containing 5% CO2 for 2-3 weeks until colonies formed. Then, colonies were fixed with polymerized formaldehyde and visualized by performing crystal violet staining. Colonies containing more than 10 cells were counted under a microscope. All conditions were analyzed in triplicated and independently tested three times.

Western blot analysis

Cells were lysed, and proteins were collected and separated by SDS-PAGE on 10% gels. The separated proteins were then transferred onto polyvinylidene fluoride membranes (Millipore), blocked with TBST containing 5% nonfat dry milk for 2 h, and incubated with antibodies against SMYD3, S1PR1, Erk, p-Erk, GAPDH (all from Santa Cruz Biotechnology) AKT, p-AKT, Stat3, and p-Stat3 (Cell Signaling Technology) overnight at 4 °C. The GAPDH antibody was used as an internal control. Then, membranes were washed 3 times with TBST and incubated with a horseradish peroxidase-conjugated secondary antibody (Biohao Biotechnology) before the protein bands were visualized with an ECL plus western blot detection kit (Beyotime Biotechnology). Quantitative increases in protein phosphorylation were evaluated by performing a densitometry analysis of the ratio of phosphorylated protein/total protein of the treated cells. All western blots were repeated 2–3 times, and the mean changes in the treated groups compared with the non-treated cells are shown under the gels in the figures.

Microarray analysis

A cell mRNA microarray was constructed according to the instructions from Affymetrix Inc. and was analyzed at the Gene ChIP Analysis Center at Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University. Changes greater than two-fold compared to the control were considered significant.

Luciferase assays

The 5′-flanking region (nucleotides −604 to +5) of S1PR1 was cloned into the pGL3-Basic vector (Promega) by Generay Biotech. The mutant constructs MT1, MT2, and MT3 were generated using a Site-Directed Mutagenesis Kit (Clontech) and sequences were verified. The wild-type and mutant fragments were inserted into a pmirGLO Dual-Luciferase vector (Promega) downstream of the luciferase gene, and the resulting constructs were designated pGL-S1PR1-WT, pGL-S1PR1-MT1, pGL-S1PR1-MT2, and pGL-S1PR1-MT3. Cells were infected with the indicated vector and then transfected with the indicated luciferase constructs, as described in the corresponding figure legend. Cell transfection was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. According to the manufacturer’s protocol, the cells were assayed 24 h after transfection to determine both firefly and Renilla luciferase activities using a luciferase assay kit (Promega). All transfections were performed in triplicate.

Chromatin immunoprecipitation

The cells were grown in three 100-mm culture dishes to 80% confluence. Then, they were fixed with formaldehyde, and the ChIP protocol was performed using a SimpleChiPTM Enzymatic Chromatin IP Kit (Agarose Beads) from Cell Signaling Technology according to their instructions. For sonication, 14 cycles of 30 s ON and 30 s OFF were conducted to obtain the required fragment sizes. The ChIP products were analyzed by Ribobio for ChIP-seq or PCR analysis for ChIP-PCR. RNA-seq data are deposited in Sequence Read Archive (SRA) database and are available through the accessions PRJNA699769. The primers for the S1PR1 promoter are listed below: a, F 5′-CAGTTGCGAGTAGCACGAGG-3′, R 5′-AGGTACGGAGGAGACAAGCAG-3′; b, F 5′-GGCGAGGGCAGTGATTTAT-3′, R 5′-CTGAACTGCTGAGACGCACT-3′; c, F 5′-CTTTCCTGGACAGTGCGTCT-3′, R 5′-CCAGTTCCCTGCCTGCTAC-3′; and d, F 5′-CCTCCCAGCCTTCCTGAA-3′, R 5′-CACTCCAATGGCCAGTCC-3′. All experiments were performed at least three times.

Animal studies

For the tumor formation assay, 4-week-old to 5-week-old BALB/c athymic nude mice were used and randomly assigned to individual experimental groups. The animal care and experimental protocols were approved by the Animal Research Committee of Sun Yat-sen University. The animal experiments were performed in accordance with established guidelines. In total, 1 × 106 cells were suspended in 200 μl of serum-free DMEM and injected into the right flanks of the mice. The tumor volume was measured every 3 days using calipers. The tumor volume was calculated using the following formula: V = (width2 × length) × 0.5. The tumor samples were collected from the mice, and protein expression was analyzed using western blotting.

Statistical analysis

Statistical analyses were performed using SPSS version 19.0 software. Continuous data are reported as the means ± SD, and the significance of differences was compared using Student’s t-tests, Fisher’s exact tests, chi-square tests and one-way ANOVA with Bonferroni post hoc corrections, as appropriate. The cumulative survival probability was evaluated using the Kaplan-Meier method, and differences were assessed using log-rank tests. P < 0.05 was considered statistically significant.