miRNA-487a Promotes Proliferation and Metastasis in Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is the fifth highest cause of new cancer cases and the second leading cause of cancer-related death worldwide. Every year an estimated 782,500 new liver cancer cases and 745,500 deaths occur worldwide, and half of these new cases and deaths occur in China (1). After intended curative therapy, the 5-year recurrence rates remain over 70% after tumor resection and 15% to 30% after liver transplantation, often leading to mortality (2, 3). Therefore, comprehensive elucidation of the mechanisms governing metastasis and recurrence of HCC are urgently needed.

MicroRNA (miRNA) is a type of short noncoding RNA (19–22 nucleotides), which function as an important regulator in tumorigenesis and tumor development (4, 5). Many studies have demonstrated that miR-422a, miR-26a, miR-195, miR-10a, and others are deregulated in HCC tissues and are associated with metastasis of HCC (6–9). Currently, 2,588 mature miRNAs have been found in humans (http://www.mirbase.org/). However, the role of most of these miRNAs in HCC remains unclear.

In our previous work, we defined a subtype of HCC named solitary large hepatocellular carcinoma (SLHCC), which has a relatively better prognosis than other HCC subtypes (3). In addition, we screened for up-regulated or downregulated miRNAs in HCCs and demonstrated that miRNAs play important roles in the proliferation and metastasis of HCC (10–12). In this study, we focused on a specific miRNA called Hsa-miR-487a (miR-487a) which is upregulated in SLHCC and is even more so in nodular HCC (NHCC). Because NHCC has a higher metastatic potential (3), we speculated that miR-487a plays a key role in the progression of HCC. Recently, Ma and colleagues (13) showed that miR-487a promotes the migration and invasion of breast cancer cells by directly targeting membrane associated guanylate kinase inverted 2, which indicates that miR-487a has a role in regulating cancer migration. However, there has been no prior report about the role of miR-487a in the metastasis of HCC or its regulatory mechanisms.

In this study, we explored the regulatory mechanism and clinical significance of miR-487a in HCC. Upregulated miR-487a was an independent prognostic marker for reduced survival and early recurrence for HCC patients. miR-487a promoted proliferation and metastasis both in vitro and in vivo by regulating phosphoinositide-3-kinase regulatory subunit 1 (PIK3R1) and sprouty-related EVH1 domain containing 2 (SPRED2) expression, respectively. Furthermore, high miR-487a expression in HCC tissues was induced by ectopic expression of heat shock factor 1 (HSF1). Inhibition of miR-487a with in vivo-morpholino effectively reduced tumor growth and metastasis in vivo. These findings, for the first time, reveal the function of miR-487a in HCC, which provides an explanation for the association between elevated miR-487a and a poor prognosis in HCC patients.

A total of 600 cases of liver tumor were randomly selected from patients who had undergone liver resection at the Department of Surgery, Xiangya Hospital of Central South University from January 2003 to December 2013. Among these cases, 86 were excluded because histopathological examination showed the tumor was intrahepatic cholangiocarcinoma or hemangioma. Patient samples acquired from January 2008 to December 2013 were divided into one group and the remaining samples were the second group. A total of 198 cases were randomly selected from each subset. The ratio of training cohort/number of validation cohort was 2:1. A total of 132 cases were included in the training cohort and 66 cases were included in the validation cohort (Supplementary Fig. S1). All research protocols strictly complied with REMARK guidelines for reporting prognostic biomarkers in cancer (14). Prior informed consent was obtained from all patients. The study was approved by the Ethics Committee of Xiangya Hospital of Central South University.

These assays were performed with HCC cells infected with lentivirus or transfected with vectors to reveal the proliferation and metastasis of each cell. All these assays were performed as described in our previous studies (10, 12).

All data were analyzed using the statistical software SPSS 17 for Windows (SPSS Inc.). Comparisons were made for the differences in clinicopathological features. The association between miR-487a expression and clinicopathological parameters was analyzed by Spearman rank analyses. Kaplan–Meier survival curves were evaluated for differences between groups using a log-rank test. The Cox proportional hazard regression model was used to identify the risk factors that were independently associated with overall survival (OS) and disease-free survival. Only factors with a P < 0.05 in univariate analysis could be analyzed in multivariate analysis. Continuous data in this study were presented as mean ± SD and analyzed by the Student t test. A comparison among three groups was performed using a one-way ANOVA. Categorical data were analyzed with a Fisher exact test. All tests were two-sided and P < 0.05 was considered statistically significant.

More details are available in the Supplementary Materials and Methods.

Previously, we found that miR-487a is highly expressed in NHCC and SLHCC with a miRNA array (Fig. 1A). Real-time PCR was performed on 198 patient samples in both the training and validation cohorts, as illustrated in Supplementary Fig. S1. miR-487a expression in HCC tissues was higher than that in adjacent non-tumorous liver tissues (ANLT) in 85.6% (113/132) and 81.8% (54/66) of cases in the training cohort and validation cohort, respectively (Fig. 1B). The median miR-487a expression in HCC tissues was significant higher than that in ANLT (Fig. 1C). In addition, the median miR-487a expression in NHCC was higher than that in SHCC and SLHCC (Fig. 1D). In situ hybridization also showed miR-487a overexpression in HCC tissues (Supplementary Fig. S2). In summation, our data show that miR-487a is frequently upregulated in HCC tissues.

To determine whether miR-487a expression correlates with the prognosis of HCC patients, we divided patients of the training cohort and validation cohort into two additional subgroups. Patients whose miR-487a expression in their HCC tissues were 2-fold more than that in their ANLT were included in the relative high expression group and the remaining samples were included in the relative low expression group. The clinicopathological characteristics of the patients in the training and validation cohorts were comparable (Supplementary Table S1). The association between miR-487a and clinicopathological features of HCC patients was analyzed in each group. miR-487a expression was directly associated with tumor size, tumor nodule number, capsular formation, microvascular invasion, and tumor-node-metastasis classification in the training cohort (Supplementary Table S2). In the validation cohort, miR-487a expression was associated with tumor size, tumor nodule number, and microvascular invasion (Supplementary Table S3). Given that tumor nodule number and microvascular invasion are important prognostic factors (15, 16), we further analyzed whether miR-487a is also a prognostic factor. The data showed that the OS and disease-free survival in the high miR-487a expression group was poorer than in the low miR-487a expression group in both cohorts (Fig. 1E and F). Univariate and multivariate analyses showed that in the training cohort, liver cirrhosis, tumor nodule number, microvascular invasion, and miR-487a expression were independent risk factors for OS (Supplementary Table S4) and disease-free survival (Supplementary Table S5). Similarly, in the validation cohort, tumor nodule number, microvascular invasion, tumor-node-metastasis classification and miR-487a expression were independent risk factors for OS (Supplementary Table S6). Only microvascular invasion and miR-487a expression were independent risk factors for disease-free survival (Supplementary Table S7). This indicates that upregulated miR-487a is an independent risk factor, and predicts poor survival outcomes for HCC patients.

miR-487a expression in all HCC cell lines tested was higher than that in liver cell lines (Fig. 2A). Among these HCC cell lines, the two with the highest miR-487a expression (HCCLM3 and MHCC97H) and the two with the lowest miR-487a expression (HepG2 and PLC/PRF/5) were selected for further in vitro assays. These cell lines were infected with lentivirus to overexpress or silence miR-487a expression for the following investigation. The infection efficiency of HCCLM3^{Anti-miR-487a}, HCCLM3^NC, HepG2^miR-487a, HepG2^NC, MHCC97H^{Anti-miR-487a}, MHCC97H^NC, PLC/PRF/5^miR-487a and PLC/PRF/5^NC were determined by fluorescence and real-time PCR (Supplementary Fig. S3).

A cell proliferation assay showed that miR-487a inhibition significantly suppressed proliferation of HCC cells and that miR-487a overexpression promoted proliferation (Fig. 2B). Colony formation assays indicated that significantly fewer colonies were developed by seeded HCCLM3^{Anti-miR-487a} cells than by seeded HCCLM3^NC cells. Similarly, significantly more colonies were developed by HepG2^miR-487a cells than by HepG2^NC cells (Fig. 2C). A 5-ethynyl-2′-deoxyuridine (EdU) assay revealed that the positive expression of EdU was higher in HCC cells with higher miR-487a expression (Supplementary Fig. S4A and S4B). Flow cytometry determined that reduced expression of miR-487a was associated with G₁ cell-cycle arrest (Fig. 2D). However, miR-487a did not induce apoptosis without the presence of antitumor drug (Supplementary Fig. S4C and S4D).

Subsequently, we assessed the role of miR-487a in metastasis with a Transwell Matrigel assay and wound-healing assay. The Transwell assay revealed that the down-regulation of miR-487a significantly reduced invasion of HCC cells, whereas the upregulation of miR-487a increased cell invasion (Fig. 2E). The wound healing assay showed that HCCLM3^NC cells and HepG2^miR-487a cells obtained quicker closure when compared with HCCLM3^{Anti-miR-487a} and HepG2^NC cells (Fig. 2F). Besides, to further confirm the role of miR-487a in HCC, we re-performed some assays with other two cell lines (MHCC97H and PLC/PRF/5). The results also indicated that miR-487a can significantly promote proliferation and metastasis of HCC cells (Supplementary Fig. S5). Taken together, we found that upregulated miR-487a significantly increases the proliferation and metastasis of HCC cells and vice versa.

To identify the direct target of miR-487a, the miRNA target predicted database (www.microRNA.org) was used. Among hundreds of predicted targets, SPRED2 and PIK3R1, which were reported as anti-oncogenes, were relevant because of their predicted binding sites and relatively good mirSVR scores of predicted binding sites. pGL3 vectors with or without wild type or mutant 3′-untranslated regions (UTR) of SPRED2 mRNA or PIK3R1 mRNA were built as illustrated in Fig. 3A and B. In SPRED2 associated assays, the luciferase activity of the SPRED2 wild-type group was significantly inhibited by miR-487a. The luciferase activity in the SPRED2 M1 group and SPRED2 M2 group was partially inhibited, but luciferase activity of SPRED2 M1+M2 group was not inhibited by miR-487a (Fig. 3C). These results indicate that miR-487a may inhibit SPRED2 expression by binding to both the 2,527–2,549 site and 2,575–2,597 site of the 3′-UTR of SPRED2 mRNA. In PIK3R1 associated assays, the luciferase activity in the PIK3R1 wild-type group was significantly inhibited by miR-487a, but luciferase activity in the PIK3R1 M1+M2 group was not inhibited. Luciferase activity in the PIK3R1 M1 group was also significantly inhibited by miR-487a, which was similar to PIK3R1 wild type. Contrarily, luciferase activity in the PIK3R1 M2 group wasn't inhibited (Fig. 3C). These data indicate that miR-487a may inhibit PIK3R1 expression by binding to the 2,802–2,824 site but not the 167–189 site of the 3′-UTR of PIK3R1 mRNA. Western blotting also showed that SPRED2 and PIK3R1 were highly expressed in HCC cells with relatively low miR-487a expression (Fig. 3D). To investigate SPRED2 and PIK3R1 expression in HCC cells and tissue, HCC cells and 40 HCC samples were randomly selected from the training and validation cohorts. The data revealed that SPRED2 mRNA and PIK3R1 mRNA was expressed at low levels in HCC cells (Supplementary Fig. S6A and S6B). The expression of SPRED2 mRNA and PIK3R1 mRNA in tumor tissue were also mostly lower than that in ANLT (Supplementary Fig. S6C and S6D). Immunohistochemistry of 40 HCC tissues indicated that miR-487a expression is associated with SPRED2 expression and PIK3R1 expression (Supplementary Fig. S6E and S6F). These data indicate that SPRED2 and PIK3R1 are direct targets of miR-487a. Thus, we have determined that miR-487a can directly bind to both 2,527–2,549 site and 2,575–2,597 site of the 3′-UTR of SPRED2, and to site 2,802–2,824 of the 3′-UTR of PIK3R1 to regulate their protein expression.

To verify whether miR-487a promotes proliferation and metastasis by targeting PIK3R1 and SPRED2, gain-and-loss assays were performed. The data showed that simultaneous inhibition of SPRED2 and PIK3R1 expression in HCCLM3^{Anti-miR-487a} cells can restore their proliferation and metastatic capacity, while simultaneous overexpression of both SPRED2 and PIK3R1 expression in HepG2^miR-487a cells eliminated the role of miR-487a in proliferation and metastasis (Fig. 4, Supplementary Figs. S7 and S8A and S8B). However, reversion of SPRED2 or PIK3R1 alone in HCC cells resulted in a different phenotype. In proliferation and colony formation assays, we found that inhibition or overexpression of PIK3R1 effectively mimicked or blocked miR-487a–mediated proliferation, respectively. However, inhibition or overexpression of SPRED2 did not play a notable role in the proliferation of HCC cells (Fig. 4A and B; Supplementary Fig. S7). A similar result was also observed in the EdU assay (Supplementary Fig. S8). Conversely, in the wound healing and Transwell assays with Matrigel or collagen IV, inhibition or overexpression of SPRED2 in HCCLM3 or HepG2 cells significantly eliminated miR-487a mediated HCC cell metastasis. However, alteration of PIK3R1 didn't significantly increase or suppress metastasis of HCC cells (Fig. 4C–F, Supplementary Fig. S8C and S8D). These results indicate that miR-487a may promote proliferation via PIK3R1, but promote metastasis via SPRED2. To our knowledge, PIK3R1 is a crucial regulatory subunit of (17), whereas SPRED2 is a member of the Sprouty/SPRED family of proteins that regulate growth factor–induced activation of the mitogen-activated protein kinase (MAPK) cascade (18). We then detected phospho-protein kinase B (p-AKT), AKT and p-ERK expression in HCC cells to analyze whether AKT or MAPK signaling was involved in the mechanism of miR-487a. Western blotting showed that p-AKT and PIK3R1 were contrarily expressed and were associated with the proliferation capability of HCC cells. p-ERK expression and SPRED2 were also contrarily expressed and were correlated with the metastatic capability of HCC cells (Supplementary Fig. S9). Our data indicated that miR-487a promotes proliferation by regulating PIK3R1-mediated AKT signaling, and facilitates metastasis by regulating SPRED2-mediated MAPK signaling in HCC cells.

To further investigate the role of miR-487a in vivo, HCC orthotopic implantation mouse models were developed as description in Supplementary Materials and Methods. Tumor volume was lower in the HCCLM3^{Anti-miR-487a} group compared with the HCCLM3^NC group, but the tumor volume in HepG2^miR-487a group was larger than in the HepG2^NC group (Supplementary Fig. S10A). Immunohistochemistry indicated that the expression of Ki67 in the HCCLM3^{Anti-miR-487a} and HepG2^NC administered mice was lower than that in the HCCLM3^NC and HepG2^miR-487a administered mice, respectively (Supplementary Fig. S10B and S10C). This was consistent with the finding that Ki67 expression was higher in HCC tissues with high miR-487 expression, and that Ki67 expression was lower in HCC tissues with low miR-487a expression (Supplementary Fig. S10D). Serial sections of liver and lung specimens stained with hematoxylin and eosin demonstrated that cells with lower miR-487a expression had less HCC intrahepatic and pulmonary metastasis (Supplementary Fig. S10E). A greater number and larger lung metastasis nodules were found in the HCCLM3^NC group and HepG2^miR-487a group than in the HCCLM3^{Anti-miR-487a} group and HepG2^NC group (Supplementary Fig. S10F). Immunohistochemistry also showed that, in concordance with our in vitro data, SPRED2 and PIK3R1 were highly expressed and p-ERK and p-AKT were lowly expressed in the HCCLM3^{Anti-miR-487a} group (Supplementary Fig. S11). These data support the conclusion that miR-487a promotes growth and metastasis of tumors in vivo.

To find out the underlying mechanisms that lead to ectopic miR-487a expression, we analyzed the region of 2,000 nucleotides upstream of miR-487a using Genomatix software. Among 141 predicted binding transcription factors, four transcription factors [hypoxia-inducible factor 1 (HIF1), HSF1, STAT3, paired box 5 (PAX5)] were considered because of their multiple predicted binding sites and their known roles in proliferation and metastasis of HCC. As illustrated in Fig. 5A, nine pairs of primers were designed to check the DNA fragments containing these predicted binding sites. Subsequently, chromatin immunoprecipitation real-time PCR was performed to detect whether these transcription factors can bind to predicted binding segments by using these specific primers. Chromatin immunoprecipitation real-time PCR was performed in HCCLM3 cells which have high HIF1, HSF1, STAT3, and PAX5 expression. The results showed that the regions amplified by primer 1 or primer 3 were significantly enriched by the HIF1 antibody or HSF1 antibody, respectively (Fig. 5B), indicating that miR-487a may be the direct target of HIF1 or HSF1. To determine whether HIF1 and HSF1 induces miR-487a expression and to identify the functional binding sites, we built vectors with a wild type or mutant promoter region of miR-487a as illustrated in Fig. 5C. A dual luciferase assay showed that mutated HSF1 predicted binding position 1 (−1,287∼−1,263) and position 2(−1,116∼−1,092) or HSF1 predicted binding position 1 (−1,287∼−1,263) alone significantly suppressed luciferase activity (Fig. 5C), which indicates that miR-487a may be mediated by HSF1 through binding at its promoter region −1,287∼−1,263 site. Subsequently, we detected miR-487a expression in HCCLM3 cells or HepG2 cells infected with HSF1 inhibiting lentivirus or overexpressing lentivirus, respectively. miR-487a expression was inhibited in HCCLM3^shHSF1 cells and its expression was up-regulated in HepG2^HSF1 cells (Fig. 5D). In addition, the dual luciferase assay showed that luciferase is only significantly activated by HSF1 in a pGL3 vector with the wild type HSF1 binding site (−1,287∼−1,263). The luciferase with mutant HSF1 binding sites (−1,287∼−1,263) could not be activated by HSF1 (Fig. 5E). This indicates that HSF1 may mediate miR-487a expression by binding to the −1,287∼−1,263 region before the transcriptional start site of miR-487a.

After validating the role and function of miR-487a (Supplementary Fig. S12), we investigated whether a miR-487a inhibitor could be a potential treatment for HCC. A phosphorodiamidate morpholino oligomer, which has been widely used in gene silencing in vivo, was adopted in our study (19, 20). A special 23-mer in vivo-morpholino designed for silencing miR-487a (Morpholino-Anti-miR-487a) and its mismatch in vivo-morpholino (Morpholino-Control) were established by Gene Tool, LLC (Corvallis; refs. 21, 22). To observe the growth and metastasis of HCC tumors, the HCCLM3 cell line labeled with luciferase (Supplementary Fig. S13) was used to construct an HCC orthotopic implantation mouse model and an HCC lung metastasis mouse model. Morpholino-anti–miR-487a or morpholino-control (12 mg/kg) was intravenously injected into mice through the tail vein twice weekly. For HCC orthotopic implantation mouse models, in vivo imaging system (IVIS) showed that tumor growth and metastases were significantly suppressed after intravenous administration of Morpholino-anti–miR-487a (Fig. 6A). The tumors in the morpholino-anti–miR-487a group displayed significant reduction in growth rates after 28 days (Fig. 6B). The sizes of the tumors in the liver samples from the morpholino-anti–miR-487a group were smaller than those of the morpholino-control group (Fig. 6C). IVIS revealed visible small metastatic lesions in two lung tissues from mice in the morpholino-anti–miR-487a group, which were fewer in number and smaller than those of morpholino-control group (Fig. 6D). Hematoxylin and eosin staining of lung tissues also showed that fewer and smaller metastatic nodules were found in the morpholino-anti–miR-487a group (Fig. 6E). For the HCC lung metastasis mouse model, IVIS showed that the metastatic lesions in morpholino-anti–miR-487a were smaller and fewer in number than those in the morpholino-control group (Supplementary Fig. S14A and S14B). There were more metastatic nodules in lung tissues collected from the morpholino-control group than in those from the morpholino-control group (Supplementary Fig. S14C and S14D). These data indicate that intravenous administration of morpholino-anti–miR-487a effectively suppresses the growth and metastasis of HCC tumors in vivo.

miR-487a, initially identified in human fetal liver in 2005 (23), when compared with other miRNAs has received little attention. Until now, there has only been reports about the role of miR-487a in tumors, and they indicated that miR-487a may affect the chemotherapeutic drug resistance or metastasis of breast cancer (13, 24). To our knowledge, this is the first study that indicates that ectopically expressed miR-487a can promote proliferation and metastasis of HCC cells and lead to a poor postoperative prognosis.

In this study, we observed several interesting targets of miR-487a, including SPRED2 and PIK3R1. We found that miR-487a promotes metastasis of HCC mainly through regulating SPRED2 induced MAPK signaling, and it also facilitates proliferation mainly via affecting PIK3R1 induced AKT signaling. Ample evidence has shown that SPRED2 is a tumor suppressor that is widely downregulated in melanoma, prostate cancer, HCC and many other cancers, and it functions mainly by inhibiting the MAPK cascade (18, 25). Our results have further confirmed the tumor suppressing role of SPRED2 in HCC. The PIK3R1 gene encodes the 85 kDa regulatory subunit alpha, which is an important subunit of PI3K. It is worthwhile to note that the role of PIK3R1 is still controversial in the HCC cell. PIK3R1, which is reported as a target of miR-486-5p or miR-376a, may promote growth, migration, and invasion of the HCC cell (26, 27). However, Taniguchi and colleagues (28) reported that PIK3R1 can exert tumor suppressor properties in HCC. Liver-specific deletion of the PIK3R1 gene leads to the development of aggressive HCC with pulmonary metastases, AKT activation and decreased phosphatase and tensin homolog (PTEN) expression. In our study, according to the proliferation assay, we also proved that PIK3R1 may inhibit the activity of AKT and function as a tumor suppressor, which supports Taniguchi's data. Chagpar and colleagues (29) showed that PIK3R1 can directly bind and enhance the lipid phosphatase activity of PTEN, making it a dual regulatory protein for both the p110-PI3-kinase and the PTEN-PI3-phosphatase. The reduction of PIK3R1 impairs PTEN-mediated PI(3,4,5)P3 dephosphorylation and leads to increased and sustained PI3K signaling. PIK3R1 can positively or negatively regulate PI3-kinase in HCC cells in different conditions, which may be the reason leading to the disagreement about the role of PIK3R1 in HCC cells.

HSF1 is an important transcription factor that can facilitate malignant transformation, cancer cell survival, and proliferation (30). In our previous study, we found that HSF1 promoted invasion and metastasis of HCC cells and was associated with a poor prognosis for HCC patients (31). Currently, more and more studies have confirmed the important role of HSF1 in tumor progression of HCC (32–34). Here, we have proven that HSF1 may bind to −1,287∼−1,263 site before transcriptional start site of miR-487a and mediate its expression. This may explain why both HSF1 and miR-487a were able to predict the prognosis of HCC patients and why they have similar roles in promoting HCC progression. A new relationship between HSF1 and miR-487a has been revealed, which may be interpreted as a new mechanism of HSF1.

Finally, we attempted to apply a miR-487a inhibitor as a gene therapy in vivo. Phosphorodiamidate morpholino oligomers are synthetic antisense oligonucleotide analogs that are designed to interfere with translational processes by forming base-pair duplexes with specific RNAs (35). Phosphorodiamidate morpholino oligomers may effectively interfere with target gene expression via exon-skipping in vivo and was preliminarily developed to be utilized in the therapy of human diseases (19). The safety and biochemical efficacy of intravenously administered AVI-4658 or AVI-7288 phosphorodiamidate morpholino oligomers has been demonstrated and they were intended to be disease modifying drugs for Duchenne muscular dystrophy or post-exposure prophylaxis for Marburg virus infection (19, 20). Therefore, we adopted an in vivo morpholino designed and synthesized by Gene Tool, LLC (21), to interfere with miR-487a expression in mice. No mice died during this assay and no sign of toxicity such as weight loss, local effects, or other visible impairments of the mice was found. IVIS and tissue samples showed that morpholino-anti-miR-487a effectively suppressed the growth and metastasis of HCC in vivo, indicating that morpholino-anti–miR-487a is potentially genetic therapeutic drug for HCC. However, the properties and safety of morpholino-anti–miR-487a in humans needs to be further investigated.

In conclusion, we have elucidated the role and molecular characteristics of miR-487a in HCC. This revealed that up-regulated miR-487a predicts poor survival outcomes for HCC patients after liver resection. In vitro studies showed that miR-487a can promote proliferation and metastasis of HCC cells, implying its mechanism in leading to the poor prognosis of HCC patients. miR-487a may promote proliferation mainly through PIK3R1, and also promote metastasis mainly through SPRED2. In addition, HSF1 is recognized as an upstream regulator leading to high expression of miR-487a in HCC. Finally, the intravenous administration of in vivo morpholino resulted in the silencing of miR-487a, which suggests that this may be a potential gene therapy for HCC.

No potential conflicts of interest were disclosed.

Conception and design: R.-M. Chang, S. Xiao, H. Yang, F. Fang, L.-Y. Yang

Development of methodology: R.-M. Chang, H. Yang, F. Fang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R.-M. Chang, S. Xiao, X. Lei, H. Yang, F. Fang, L.-Y. Yang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R.-M. Chang, S. Xiao, F. Fang

Writing, review, and/or revision of the manuscript: R.-M. Chang, S. Xiao, F. Fang, L.-Y. Yang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R.-M. Chang, S. Xiao, H. Yang, F. Fang, L.-Y. Yang

Study supervision: R.-M. Chang, L.-Y. Yang

This work was supported in part by the following grants: Clinical Subjects' Key Project of Ministry of Health (no. 2010439), and National Science & Technology Major Projects (2009ZX09103-681, 2012ZX100020122011) and National Nature Science Foundation of China (no. 81272395), Key Project of National Nature Science Foundation of China (no. 81330057), The Specialized Research Fund for the Doctoral Program of Higher Education of China (no. 20130162130007).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.