MicroRNA-26a suppresses angiogenesis in human hepatocellular carcinoma by targeting hepatocyte growth factor-cMet pathway

Authors

  • Xin Yang,

    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    Search for more papers by this author
  • Xiao-Fei Zhang,

    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    Search for more papers by this author
  • Xu Lu,

    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    Search for more papers by this author
  • Hu-Liang Jia,

    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    Search for more papers by this author
  • Lei Liang,

    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    Search for more papers by this author
  • Qiong-Zhu Dong,

    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    Search for more papers by this author
  • Qing-Hai Ye,

    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    Search for more papers by this author
  • Lun-Xiu Qin

    Corresponding author
    1. Liver Cancer Institute & Zhongshan Hospital, Institutes of Biomedical Science, Fudan University, Shanghai, China. Key Laboratory of Carcinogenesis & Cancer Invasion, Ministry of Education, China
    • Address reprint requests to: Lun-Xiu Qin, M.D., Ph.D., Liver Cancer Institute and Zhongshan Hospital, Fudan University, 180 Feng Lin Road, Shanghai 200032, China. E-mail: qin.lunxiu@zs-hospital.sh.cn; fax: +86-21-5423 7960.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • Supported by China National Key Projects for Infectious Disease (2008ZX10002-021, 2012ZX10002-012), the National Key Basic Research Program of China (2013CB910500; 2009CB521701), Program for Changjiang Scholars and Innovative Research Team (the Ministry of Education of China; IRT1118), Program of Shanghai Chief Scientist (08XD14008), China National Natural Science Foundation (81120108016; 30600589), Program for Outstanding Medical Academic Leader (LJ10010).

Abstract

MicroRNA (miR)-26a can suppress tumor growth and metastasis of hepatocellular carcinoma (HCC). Since angiogenesis is important for tumor growth and metastasis, we investigated the possible roles of miR-26a in tumor angiogenesis. Down-regulation of miR-26a was found to correlate with an increased angiogenic potential of HCC. Through gain- and loss-of-function studies, miR-26a was demonstrated to significantly inhibit vascular endothelial growth factor A (VEGFA) expression in HCC cells and then suppress the promoting effects of HCC cells on in vitro proliferation, migration, and capillary tube formation of endothelial cells, as well as in vivo tumor angiogenesis of HCC. Hepatocyte growth factor (HGF) was identified as a target of miR-26a. HGF simulation antagonized the effects induced by miR-26a up-regulation. In contrast, silencing HGF induced similar effects to miR-26a. We further found that miR-26a exerted its antiangiogenesis function, at least in part, by inhibiting HGF-hepatocyte growth factor receptor (cMet) and its downstream signaling pathway, in turn, suppressing VEGFA production in HCC cells and impairing VEGFR2-signaling in endothelial cells. HCC patients who had high miR-26a, low HGF, low VEGFA, or low microvessel density (MVD) in tumor tissues had a better prognosis with longer overall survival (OS) and time to recurrence (TTR). In multivariate analysis, miR-26a, or in combination with HGF, was demonstrated to be an independent prognostic indicator for OS and TTR of HCC patients. Conclusion: miR-26a could suppress tumor angiogenesis of HCC through HGF-cMet signaling, and it is a new hopeful therapeutic target and prognostic marker for HCC. (Hepatology 2014;59:1874–1885)

Abbreviations
AFP

alpha-fetoprotein

AKT

protein kinase B

ALT

alanine aminotransferase

BCLC

Barcelona Clinic Liver Cancer

cMet

hepatocyte growth factor receptor

DMEM

Dulbecco's modified Eagle's medium

ERK

extracellular regulated protein kinases

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

HbsAg

hepatitis B surface antigen

HCC

hepatocellular carcinoma

HIF-1α

hypoxia-inducible factor-1alpha

HGF

hepatocyte growth factor

miR-26a

microRNA-26a

miRNA

microRNA

mTOR

mammalian target of rapamycin

MVD

microvessel density

NC

negative control

OS

overall survival

PI3K

phosphoinositide 3-kinase

qRT-PCR

quantitative real-time polymerase chain reaction

S6K

ribosomal protein S6 kinase

TNM

tumor-node-metastasis

TTR

time to recurrence

UTR

untranslated region

VEGFA

vascular endothelial growth factor A

VEGFR2

vascular endothelial growth factor receptor 2

Hepatocellular carcinoma (HCC) is a highly vascularized tumor with frequent intrahepatic and extrahepatic metastasis, which is responsible for the rapid recurrence and poor survival of HCC.[1-3] Therefore, it is urgent to explore the molecular mechanisms of underlying metastasis and tumor angiogenesis of HCC, which would provide potential effective therapeutic targets to improve the survival of HCC patients. MicroRNAs (miRNAs) are involved and play critical roles in a variety of biological processes through suppressing the expression of target messenger RNAs (mRNAs).[4] Deregulation of miRNAs has been shown to result in aberrant gene expression, and contributes to invasion and metastasis of many kinds of human tumors including HCC.[2, 5-9]

There are still controversies about the roles of miR-26a in human malignancies. Many studies show that miR-26a serves as a potential tumor suppressor in several distinct cancer types including HCC,[10-15] but some others indicate that it exhibits oncogenic properties in glioma, lung cancer, and cholangiocarcinoma.[16-18] It is possible that miR-26a plays organ-specific roles, in part due to the different cellular contexts of tumors. Recently, we found that miR-26a could suppress tumor growth and metastasis of HCC through IL-6-Stat3 signaling.[14] In this study, using both gain- and loss-of-function analyses, we further find the suppressive function of miR-26a on tumor angiogenesis. Moreover, we reveal that Hepatocyte growth factor (HGF) is a target of miR-26a, and miR-26a exerted its antiangiogenesis function, at least in part, by inhibiting the HGF-hepatocyte growth factor receptor (cMet) signaling pathway. Our findings provide the underlying mechanism of miR-26a in angiogenesis of HCC and implicate miR-26a as a potential therapeutic target for HCC.

Materials and Methods

Patients and Specimens

A total of 120 patients who underwent hepatectomy for HCC at Liver Cancer Institute and Zhongshan Hospital, Fudan University (Shanghai, China) between January 2005 and December 2006 were enrolled in this study. Their clinicopathological characteristics are presented in Supporting Table S1. All patients were diagnosed with primary HCC, and none had received any preoperative cancer treatment. Clinical samples were collected from patients after obtaining informed consent in accordance with a protocol approved by the Ethics Committee of Fudan University (Shanghai, China).

In Vitro Cell Proliferation and Migration Analyses

Cell proliferation analyses were performed using Cell Counting Kit-8 (Dojindo, Kumamoto, Japan). Wound-healing assays were used to examine cell migration of human umbilical vein endothelial cells (HUVECs). The details are described in the Supporting Materials and Methods.

In Vitro Capillary Tube Formation Analyses

HUVECs (5 × 103 per well) were cultured at 37°C for 3, 6, 12, and 24 hours in a 96-well plate coated with Matrigel (BD Pharmingen, San Jose, CA) in the absence or presence of culture medium of HCC cells. The formation of capillary-like structures was captured under a light microscope. The number of the formed tubes, which represent the degree of angiogenesis in vitro, were scanned and quantitated in five low-power fields (100× magnification).

In Vivo Tumor Angiogenesis Assays

All experimental procedures involving animals were approved by the Animal Care and Use Committee of Fudan University, China. HCC-LM3 or MHCC97-H cells (2 × 106) which were stably transfected with a miR-26a expression or control vector were subcutaneously injected into nude mice (BALB/c nu/nu, 4 weeks old) (SLRC, Shanghai, China) to establish flank xenograft models. Six weeks later, tumors were removed, fixed in formalin, and embedded in paraffin. Vascular endothelial growth factor A (VEGFA) and CD34 were detected in tumor tissues by immunohistochemistry (Epitomics, Burlingame, CA). The microvessel density (MVD) in tumor tissues was evaluated based on the staining for CD34.

Luciferase Reporter Assay

HEK293T cells were cotransfected with 120 ng of miR-26a or its inhibitor expression vector and 30 ng of the wild-type or mutant 3′ untranslated region (UTR) of HGF, 0.45 μL of Fugene (Promega, Madison, WI), and then applied to luciferase reporter assay as previously described.[14]

Detection of Target Proteins

HGF and VEGFA levels in serum or culture medium were detected by enzyme-linked immunosorbent assay (ELISA) (Cusabio, Wuhan, China) according to the manufacturer's instruction. The protein levels of HGF, p-Met, p-P85, protein kinase B (Akt), p-Akt, mammalian target of rapamycin (mTOR), p-mTOR, ribosomal protein S6 kinase p-(S6K), HIF-1α, and VEGFA in HCC cell lines and tumor tissues were detected by western blot and immunohistochemistry, respectively. The details are present in the Supporting Materials.

Statistical Analysis

Statistical analysis was performed with SPSS 15.0 (Chicago, IL). The difference between groups was analyzed using a Student t test when comparing only two groups or one-way analysis of variance when comparing more than two groups. P < 0.05 was considered statistically significant.

Results

Down-Regulation of miR-26a Is Significantly Associated With Angiogenesis in HCC

Our previous study reported that the expression level of miR-26a was significantly down-regulated in metastatic HCC tissues in comparison to nonmetastatic HCC tissues.[14] In the present study, we further confirmed the association between miR-26a down-regulation and HCC metastasis and venous invasion (P < 0.001) (Supporting Fig. S1A,B).

To evaluate the association between miR-26a level and angiogenesis in HCCs, we analyzed VEGFA expression and MVD in HCC tissues with different miR-26a levels. The “low” versus “high” miR-26a expression was defined according to the cutoff values of miR-26a level, which were defined as the median of the cohort of tested patients. The range of miR-26a is shown in Supporting Fig. S2. The VEGFA and MVD levels in high-miR-26a HCC tissues were significantly decreased compared with those low-miR-26a HCCs (P < 0.01) (Fig. 1A,B). Significant negative correlations were found between miR-26a levels and VEGFA levels (r = −0.5478, P < 0.001) or MVD (r = −0.8168, P < 0.001) in HCC tissues (Fig. 1C,D), which were further confirmed by χ[2] tests (P < 0.001) (Supporting Table S2). These data suggest that miR-26a down-regulation is associated with tumor angiogenesis in HCCs.

Figure 1.

The association of miR-26a expression with angiogenic potential in HCC tissues. The VEGFA (A) and MVD (B) levels in HCC tissues with different miR-26a levels detected by immunohistochemistry staining (200× magnification). Both VEGFA and MVD levels in high-miR-26a HCCs were significantly lower than that of low-miR-26a HCCs (A,B, right). A significant negative correlation was found between miR-26a and VEGFA (C) or MVD (D) levels in HCC tissues. **P < 0.01.

MiR-26a Significantly Inhibited In Vitro Proliferation, Migration, and Capillary Tube Formation of HUVEC Cells

To explore the biological significance of miR-26a in tumor angiogenesis, we further determined the effects of miR-26a on the abilities of in vitro proliferation, migration, and capillary tube formation of HUVECs. HCC-LM3 and MHCC97-H cells which had lower intrinsic miR-26 levels were transfected with miR-26a expression vectors to up-regulate their miR-26 expression, while the high-intrinsic miR-26a HepG2 and PLC cells were transfected with anti-miR-26a to down-regulate their miR-26a levels.[14] Their culture medium was added to HUVECs. The culture medium of miR-26a-transfected HCC-LM3 and MHCC97-H cells could significantly suppress the cell proliferation, migration, and capillary tube formation of HUVECs (Fig. 2, Supporting Fig. S3). In contrast, the medium of anti-miR-26a-transfected HepG2 and PLC cells could induce obvious increases in the abilities of cell proliferation, migration, and capillary tube formation of HUVECs (Supporting Fig. S4). These results suggest that miR-26a could suppress in vitro angiogenesis of HUVECs.

Figure 2.

The inhibitory effects of miR-26a in HCC-LM3 cells on in vitro proliferation, migration, and capillary tube formation of HUVECs. After incubating with culture medium of miR-26a-transfected HCC-LM3 cells, the abilities of in vitro proliferation (A), migration (B), and capillary tube formation (C) of HUVECs were significantly decreased compared with the controls. *P < 0.05, **P < 0.01.

We also found that miR-26a up-regulation significantly decreased VEGFA expression in HCC cells. After miR-26a transfection, the VEGFA levels of HCC-LM3 and MHCC97-H cells were significantly decreased (P < 0.01) (Supporting Fig. S5A,B). In contrast, knockdown of miR-26a could significantly increase VEGFA expression levels in HepG2 and PLC cells (P < 0.01) (Supporting Fig. S5C,D). These suggest that miR-26a is an inhibitor of VEGFA expression and miR-26a suppresses tumor angiogenesis at least in part through inhibiting VEGFA expression in HCC.

MiR-26a Could Significantly Inhibit In Vivo Tumor Angiogenesis of HCC

To further determine the effects of miR-26a on in vivo tumor angiogenesis of HCC, HCC-LM3 or MHCC97-H cells stably expressing miR-26a or control vector were subcutaneously implanted into nude mice. Palpable tumors formed within 1 week. The average levels of VEGFA in both the tumor tissues and serum of the subcutaneous models bearing miR-26a-transfectant HCC cells were significantly lower than that of the controls (P < 0.01) (Fig. 3A,B). In addition, MVD levels in tumor tissues from miR-26a-transfected HCC-LM3 or MHCC97-H cells were significantly lower than those of the controls (P < 0.01) (Fig. 3C). This indicates that miR-26a can remarkably inhibit in vivo tumor angiogenesis of HCC.

Figure 3.

The inhibitory effects of miR-26a on in vivo tumor angiogenesis of HCC. VEGFA levels in tumor tissue (A) and the serum (B) of nude mice models with subcutaneous implantation of HCC were noticeably reduced when miR-26a was up-regulated. (C) MVD in tumors tissues of subcutaneous implantation models bearing HCC-LM3 and MHCC97-H (200× magnification). **P < 0.01.

HGF Is a Direct Downstream Target of miR-26a

We searched for candidate target genes of miR-26a using publicly available databases including TargetScan (http://www.targetscan.org/), PicTar (http://pictar.mdc-berlin.de/), and miRanda (microrna.org and miRbase), and found the complementary sequence of miR-26a in the 3′-UTR of mRNA of HGF which plays critical roles in tumor angiogenesis[19-22] (Fig. 4A). Therefore, HGF was selected for further experimental validations. HGF levels in high-miR-26a HCC tissues were significantly decreased compared with those with low-miR-26a (P < 0.01) (Supporting Fig. S6A). A significant inverse correlation was revealed by Spearman's correlation analysis between mRNA levels of miR-26a and HGF (r = −0.6299, P < 0.001) (Supporting Fig. S6B) and between miR-26a mRNA levels and HGF protein levels in 120 HCC tissues (r = −0.6932, P < 0.001) (Fig. 4B), which was further confirmed by Fisher's exact tests (P < 0.001 and P = 0.022) (Supporting Table S2). We also detected HGF mRNA expression levels in HCC cells (Supporting Fig. S7). Moreover, up-regulation of miR-26a could induce a significant decrease of HGF mRNA levels in HCC-LM3 and MHCC97-H cells and protein levels in the culture supernatant of HCC cells (P < 0.01) (Fig. 4C; Supporting Fig. S8A). In contrast, down-regulation of miR-26a dramatically increased HGF mRNA levels in HepG2 and PLC cells and protein levels in the culture supernatant of HCC cells (P < 0.01) (Fig. 4C, Supporting Fig. S8B).

Figure 4.

HGF is a direct target of miR-26a in HCC. (A) miR-26a and its putative binding sequence in the 3′-UTR of HGF. The mutant miR-26a-binding site was generated in the complementary site for the seed region of miR-26a (wt: wild-type; mt: mutant type). (B) HGF expression in HCC tissues detected by immunohistochemistry staining (left) (200× magnification). miR-26a levels were inversely correlated with HGF levels in HCC tissues (right). (C) Effects of miR-26 up- or down-regulation on HGF protein levels in culture supernatant of HCC cells. MiR-26a transfection significantly reduced HGF protein levels of HCC-LM3 and MHCC97-H cells, while anti-miR-26a dramatically increased supernatant HGF protein levels of HepG2 and PLC cells. (D) The changes of serum protein HGF levels in subcutaneous implantation models of HCC-LM3 or MHCC97-H cells after miR-26a was up-regulated. (E) Influence of miR-26a on transcription of HGF. miR-26a significantly suppressed the luciferase activity that carried wt but not mt 3′-UTR of HGF. Anti-miR-26a led to a noticeable increase in luciferase activity of wt 3′-UTR of HGF. **P < 0.01.

In the subcutaneous implantation nude mice models bearing human HCC, HGF mRNA levels in tumors and protein levels in the serum and tumor tissues were found to be noticeably reduced when miR-26a was up-regulated (P < 0.01) (Fig. 4D; Supporting Figs. S8C, S9). In addition, overexpression of miR-26a significantly inhibited the luciferase activity of HGF containing a wild-type 3′-UTR but did not significantly influence HGF with a mutant 3′-UTR (P < 0.01) (Fig. 4E). Knockdown of miR-26a by anti-miR-26a could significantly increase the luciferase activity of HGF (P < 0.01) (Fig. 4E). These results strongly suggest that HGF is a target of miR-26a in HCC.

To further validate this finding, we treated the miR-26a-transfected HCC-LM3 cells with 60 ng/mL of HGF, and found that after HGF treatment the culture medium of miR-26a-transfected HCC cells could obviously increase the abilities of proliferation, migration, and capillary tube formation of HUVECs (Fig. 5A-C), which antagonized the effects induced by miR-26a. In contrast, silencing HGF in HCC-LM3 cells induced by small interfering RNA (siRNA) could promote the proliferation, migration, and capillary tube formation of HUVECs, which were similar to those of miR-26a (Fig. 5D-F). These further support that HGF is a downstream functional mediator of miR-26a in tumor angiogenesis.

Figure 5.

Alterations of HGF levels influence the in vitro effects of miR-26a on HUVECs. (A-C) HGF treatment antagonized the inhibitory effects of miR-26a on the abilities of in vitro proliferation (A), migration (B), and capillary tube formation (C) of HUVECs. (D-F) HGF down-regulation by siRNA could induce similar inhibitory effects of those by miR-26a on proliferation (D), migration (E), and capillary tube formation (F) of HUVECs. *P < 0.05, **P < 0.01.

MiR-26a Inhibits VEGFA Production Through HGF-cMet Signaling Pathway

HGF is thought to exert its function through interaction with a specific receptor c-Met.[22] The downstream phosphoinositide 3-kinase (PI3K)/Akt/mTOR/S6K and Hypoxia-inducible factor-1alpha (HIF-1α)-VEGF signaling pathway have also been found to be involved in HGF-inducing VEGF expression.[23-27] We found that the expression levels of p-Met, p-P85, p-Akt, p-mTOR, p-S6K, nuclear HIF-1α, and VEGFA were significantly decreased in HCC-LM3 and MHCC97-H cells that stably overexpressed miR-26a (Fig. 6A; Supporting Figs. S10A, S11A). They were significantly up-regulated after HGF stimulation for 48 hours (Fig. 6A; Supporting Figs. S10A, S11A). Also, pretreatment with PI3K inhibitor Ly294002 could obviously reduce the expression levels of these proteins except p-Met (Fig. 6A; Supporting Fig. S10A). Furthermore, siRNAs against HGF and cMet could significantly inhibit the expression of p-Met, p-P85, p-Akt, p-mTOR, p-S6K, nuclear HIF-1α, and VEGFA, which were similar to the effects induced by miR-26a (Fig. 6A; Supporting Figs. S10A, S11A). Moreover, the expression levels of these proteins in HCC tissues from subcutaneous implantation models of miR-26a-transfected HCC-LM3 cells were greatly decreased compared with the controls (P < 0.01) (Fig. 3A; Supporting Fig. S9). Similarly, their levels were significantly reduced in high-miR-26a HCC tissues compared with those in low-miR-26a HCCs (P < 0.01) (Fig. 1A; Supporting Figs. S6A, S12).

Figure 6.

MiR-26a inhibits VEGFA production in HCC-LM3 cells and impairs VEGFR2 signaling in HUVEC cells by suppressing the HGF-cMet pathway. (A) The expression levels of p-Met, p-P85, p-Akt, p-mTOR, p-S6K, HIF-1α, and VEGFA in HCC-LM3 cells detected by western blots with GAPDH and β-tubulin as the control for total and nuclear proteins respectively: up-regulation of miR-26a in HCC-LM3 cells significantly decreased their levels (left, lane 2) compared with the control (left, lane 1); HGF antagonized the inhibitory effects induced by miR-26a (left, lane 3); PI3K inhibitor Ly294002 decreased the expression levels of all these proteins except p-Met (left, lane 4). Knockdown HGF and cMet by siRNAs induced similar inhibitory effects to miR-26a on the expression of p-Met, p-P85, p-Akt, p-mTOR, p-S6K, nuclear HIF-1α, and VEGFA (right, lanes 2 and 3). (B) The protein levels of p-VEGFR2, VEGFR2, p-Akt, Akt, p-ERK, and ERK in HUVECs detected by western blots with GAPDH as control: the expression levels of p-VEGFR2, p-Akt, and p-ERK of HUVECs were reduced after incubating in culture medium of miR-26a-overexpressing HCC-LM3 cells (left, lane 2); HGF significantly promoted expression of these proteins (left, lane 3). HGF and cMet down-regulation by siRNAs significantly decreased the expression of p-VEGFR2, p-Akt, and p-ERK (right, lanes 2 and 3), which were similar to that induced by miR-26a. However, no significant change was in the total amounts of VEGFR2, Akt, and ERK.

Taken together, this indicates that miR-26a suppresses VEGFA production in HCC by inhibiting the HGF-cMet signaling pathway, and then influencing the downstream PI3K/Akt/mTOR/S6K and HIF-1α-VEGF signaling pathways.

Inhibitory Effects of miR-26a on VEGFR2 Signaling Pathway in HUVEC Cells

Since proangiogenic signals are able to induce the phosphorylation and activation of VEGFR2, which regulates proliferation, migration, and tube formation of ECs by activating ERK and AKT,[28, 29] we further evaluated whether miR-26a repressed VEGFA expression by inhibiting HGF in tumor cells and, in turn, abrogating VEGFR2-signaling in ECs. We found that the expression levels of p-VEGFR2, p-Akt, and p-ERK in HUVECs were significantly decreased after incubating in culture medium of HCC-LM3 and MHCC97-H cells that stably overexpressed miR-26a (Fig. 6B; Supporting Fig. 11B). HGF treatment could significantly stimulate the expression of p-VEGFR2, p-Akt, and p-ERK in HUVECs (Fig. 6B; Supporting Fig. 11B). These effects were dramatically attenuated when HGF and c-Met in HCC cells were silenced by siRNA (Fig. 6B; Supporting Fig. 11B), which were similar to those induced by miR-26a. Furthermore, the expression levels of p-VEGFR2, p-Akt, and p-ERK in subcutaneously implanted tumors of miR-26a-transfected HCC-LM3 were greatly decreased compared with the controls (P < 0.01) (Supporting Fig. S13). In addition, the levels of these proteins were dramatically reduced in high-miR-26a HCC tissues compared with those low-miR-26a HCCs (P < 0.01) (Supporting Fig. S14). These data indicate that miR-26a inhibits VEGFR2-signaling in HUVECs to exert its antiangiogenic function.

Prognostic Values of miR-26a and HGF for HCC Patients

Consistent with our previous study,[14] we found that miR-26a levels in tumor tissues were significantly associated with tumor size (P = 0.037), tumor number (P = 0.032), and vascular invasion (P < 0.001) (Supporting Table S3), as well as prognosis of HCC patients (Supporting Fig. S15A,B). The patients with high HGF levels had shorter overall survival (OS) and higher possibilities of tumor recurrence compared with those patients with low HGF levels (Fig. 7A,B). Similarly, the patients with high VEGFA and MVD levels had poorer OS and higher possibilities of tumor relapse (Supporting Fig. S15C-F).

Figure 7.

The prognostic significances of miR-26a and HGF for HCC patients assessed by Kaplan-Meier analyses. High HGF expression in HCC tissues was significantly associated poorer OS (A) and higher probability of tumor recurrence (B). Patients in subgroup I had the best OS (C) and lowest possibility of tumor recurrence (D). The four subgroups were divided according to combinations of miR-26a and HGF levels, i.e., I, high miR-26a/low HGF; II, low miR-26a/low HGF; III, high miR-26a/high HGF; IV, low miR-26a/high HGF. For each cohort, the median of the cohort were defined as the cutoff values of miR-26a and HGF.

In the univariate and multivariate Cox hazard analyses, miR-26a, HGF, VEGFA, MVD, tumor size, and capsule were identified as independent prognostic indicators for HCC patients (Table 1).

Table 1. Univariate and Multivariate Analyses of Factors Associated With OS and TTR of HCC Patients (n = 120)
 Overall SurvivalTime to Recurrence
Clinical VariablesHR (95%CI)P ValueHR (95%CI)P Value
  1. HBsAg, hepatitis B surface antigen; AFP, alpha-fetoprotein; ALT, alanine aminotransferase; TNM, tumor-node-metastasis staging system; BCLC, Barcelona Clinic Liver Cancer; tumor differentiation, Edmondson grade; CI, confidence interval.

  2. Combination of miR-26a and HGF: I, High miR-26a and Low HGF; II, High miR-26a and High HGF; III, Low miR-26a and Low HGF; IV, Low miR-26a and High HGF.

  3. The “low” or “high” expression of miR-26a, HGF, VEGFA and MVD level was defined according to their cutoff values, which were defined as the median values of the cohort of patients tested. Analysis was conducted using univariate analysis or multivariate Cox proportional hazards regression

  4. a

    Multivariate analysis of miR-26a, HGF, VEGFA, MVD, tumor size, tumor capsule, and vascular invasion

  5. b

    Multivariate analysis of the combination of miR-26a and HGF, VEGFA, MVD, tumor size, tumor capsule, and vascular invasion.

  6. Bold indicates significant values.

Univariate analysis    
miR-26a (high versus low)0.31(0.15-0.62)0.0010.61(0.39-0.94)<0.001
HGF (high versus low)2.75(1.41-5.34)0.0013.43(2.13-5.52)0.025
VEGFA (high versus low)3.04(1.17-7.90)0.0143.43(2.13-5.52)0.035
MVD (high versus low)3.04(1.08-8.56)0.0160.61(0.39-0.94)0.019
Gender (female versus male)0.84(0.32-2.21)0.7210.78(0.42-1.48)0.322
Age (≥50 versus <50 years)1.21(0.61-2.36)0.5780.87(0.56-1.36)0.409
HBsAg (positive versus negative)0.93(0.36-2.40)0.8760.86(0.36-2.00)0.718
Cirrhosis (yes versus no)0.77(0.39-1.53)0.4550.83(0.46-1.50)0.542
ALT (≥75 versus<75U/L)1.23(0.62-2.44)0.5480.89(0.66-1.70)0.716
AFP (≥20 versus<20 ng/mL)1.40(0.70-2.78)0.3481.12(0.63-2.00)0.691
Tumor size (≥5cm versus<5 cm)3.58(1.69-7.57)0.0012.39(1.34-4.30)0.003
Tumor number (multiple versus single)1.14(0.55-2.37)0.7221.19(0.64-2.21)0.591
Tumor capsule (complete versus none)2.44(1.22-4.88)0.0122.49(1.35-4.34)0.003
Vascular invasion (yes versus no)2.65(1.32-5.30)<0.0011.59(0.89-2.85)0.118
Distance metastasis (yes versus no)1.90(0.73-4.98)0.1892.34(0.95-5.81)0.066
Tumor differentiation (III/IV versus I-II)0.78(0.40-1.51)0.4680.85(0.45-1.59)0.609
TNM stage (III/IV versus I/II)1.43(0.77-2.68)0.2621.09(0.64-1.83)0.759
BCLC stage(0 and A versus B and C)1.23(0.65-2.02)0.3411.37(0.86-2.56)0.425
Combination of miR-26a and HGF    
 I versus IV0.08(0.03-0.26)<0.0010.16(0.07-0.36)<0.001
 II versus IV0.40(0.17-0.94)0.0120.34(0.16-1.73)0.001
 III versus IV0.45(0.21-0.96)0.0120.73(0.38-1.40)0.125
Multivariate analysisa    
miR-26a (high versus low)0.35(0.19-0.61)0.0010.29(0.17-0.50)0.005
HGF (high versus low)2.78(1.53-5.05)0.0031.75(1.07-2.85)0.043
VEGFA (high versus low)2.96(1.24-7.10)0.0232.02(1.05-3.89)0.035
MVD (high versus low)2.89(1.21-6.69)0.0352.19(1.14-4.20)0.024
Tumor size (≥5cm versus<5 cm)3.06(1.52-6.16)0.0022.30(1.32-4.01)0.003
Tumor capsule (complete versus none)2.02(1.13-3.60)0.0172.02(1.23-3.31)0.005
Vascular invasion (yes versus no)3.14(1.69-5.84)0.002  
Multivariate analysisb    
Combination of miR-26a and HGF    
 I versus IV0.10(0.03-0.28)<0.0010.15(0.07-0.33)<0.001
 II versus IV0.40(0.20-0.81)0.0350.32(0.16-0.62)0.005
 III versus IV0.41(0.21-0.82)0.0380.63(0.35-1.14)0.341
VEGFA (high versus low)2.88(1.21-7.02)0.0251.99(1.02-3.76)0.038
MVD (high versus low)2.85(1.22-6.58)0.0352.16(1.05-4.12)0.026
Tumor size (≥5cm versus<5 cm)3.02(1.50-6.08)0.0022.28(1.31-3.98)0.004
Tumor capsule (complete versus none)1.96(1.09-3.52)0.0241.98(1.20-3.27)0.008
Vascular invasion (yes versus no)3.14(1.59-5.84)0.002  

HCC patients with high miR-26a and low HGF had the best prognosis, and those low-miR-26a and high-HGF patients had the poorest prognosis, with the lowest OS and highest probability of tumor recurrence (P < 0.001) (Fig. 7C,D). The combination of miR-26a and HGF was confirmed to be an independent prognostic indicator for OS (P <0.001) and time to recurrence (TTR) (P <0.001), which was even better than miR-26a or HGF alone (Table 1).

Discussion

Active tumor angiogenesis is confirmed to associate with invasion and metastasis of HCC.[1-3, 30-33] It has been reported that the dysfunction of miRNAs in tumor cells affects tumor angiogenesis, including in HCC.[2, 9] In the present study, using in vitro and in vivo assays, for the first time we identify the antiangiogenic function of miR-26a. The level of miR-26a expression was inversely associated with MVD in HCC tissues. Through up- and down-regulating miR-26a in HCC cells, we confirm that miR-26a could inhibit the abilities of in vitro proliferation, migration, and capillary tube formation of HUVECs, as well as suppress in vivo tumor angiogenesis in nude mice models bearing human HCC.

We found that the inhibitory effects of miR-26 on tumor angiogenesis of HCC might be due to the decreased expression of VEGFA, an important proangiogenic factor secreted by tumor cells.[29, 33] Cancer cells can secrete VEGF into the microenvironment to promote tumor angiogenesis to meet their need for a blood supply.[34, 35] Blocking VEGF leads to the regression of vascular network and inhibition of tumor growth and metastasis.[36] MiR-195 was reported to suppress angiogenesis of HCC by directly inhibiting the expression of VEGF, and miR-27b targets VEGFC to inhibit tumor angiogenesis of colorectal cancer.[9, 37] In this study, we found that miR-26a could inhibit the expression of VEGFA. However, VEGF is not an effective target gene of miR-26a. Therefore, it is crucial to elucidate the mechanisms of how miR-26a in tumor cells affects VEGFA expression.

Through the correlation analysis in both HCC tissues and cell lines, together with both in vitro and in vivo functional studies, we confirm that miR-26a inhibits tumor angiogenesis at least in part by suppressing HGF. An inverse correlation was found between the mRNA levels of miR-26a and HGF mRNA levels or protein levels in HCC tissues. This correlation is further confirmed by evaluating the effects of up- and down-regulating miR-26a on HGF expression in HCC cells. Moreover, HGF down-regulation in HCC cells could induce similar effects as those of miR-26a on HUVECs, and HGF stimulation could antagonize the effects of miR-26a. This suggests that HGF serves as a downstream mediator of miR-26a function in angiogenesis of HCC.

It is well known that HGF plays important roles in cell proliferation, motility, metastasis, and angiogenesis of HCC, mainly through activating cell surface receptors such as cMet.[19-22, 38] HGF can stimulate cancer cells to produce VEGF,[20, 39] which can be antagonized by blocking c-Met activation.[40] Meanwhile, its downstream PI3K/Akt/mTOR/S6K and HIF-1α-VEGF signal pathways are considered to involve in HGF-inducing VEGF expression.[23-25, 27] In this study, the solid data of both in vitro and in vivo assays indicate that miR-26a significantly decreases the expression levels of the proteins related to PI3K/Akt/mTOR/S6K and HIF-1α-VEGF signal pathways, and moreover, these effects are induced through the HGF-cMet pathway. These provide further evidence to support that miR-26a decreases VEGFA expression by suppressing HGF-cMet pathway.

Another interesting finding of this study is that miR-26a down-regulation in HCC cells increases VEGFA levels in the tumor microenvironment through the HGF-cMet pathway, which induces the activation of VEGFR2 signaling in endothelial cells and promotes tumor angiogenesis in HCC. We found that the decreased VEGFA induced by miR-26a through HGF-cMet pathway results in significant decreases in the expression levels of p-VEGFR2, p-Akt, and p-ERK in HUVECs, and subsequently inhibits activities of ECs. VEGF stimulates cellular responses by activating tyrosine kinase receptors (VEGFRs) on the cell surface.[41] As a pivotal activator of the VEGFR2 pathway, tumor-secreted VEGF binds to and induces the phosphorylation and activation of VEGFR2 on endothelial cells, and then results in phosphorylation of the extracellular signal-regulated kinase (ERK) and AKT, and promotes the proliferation, migration, and survival of endothelial cells.[29, 33]

Consistent with our previous study that miR-26a is a novel prognostic marker and therapeutic target for HCC,[14] in our present study we further confirm that miR-26a, HGF, their combination, MVD, and VEGFA are independent prognostic indicators for OS and TTR in HCC patients. The prognostic performance of the combination of miR-26a and HGF is much better than that of miR-26a or HGF alone.

miR-26a has been reported to regulate hepatocyte proliferation during liver regeneration,[42] block G1/S transition,[12, 15, 43] and suppress in vivo tumor growth and metastasis of HCC.[14, 44] This study provides further evidence to support that miR-26a is a promising prognostic marker and therapeutic target for HCC. As anti-VEGF therapies are important strategies for cancer treatment including for HCC, miR-26a could be a novel antiangiogenic upstream target of VEGF, and therefore has potential clinical significance.

Ancillary

Advertisement