MicroRNA-29b suppresses tumor angiogenesis, invasion, and metastasis by regulating matrix metalloproteinase 2 expression

Authors


  • Potential conflict of interest: Nothing to report.

  • Supported by grants from Ministry of Science and Technology of China (2010CB912803, 2011CB811305), National Natural Science Foundation of China (30925036, 30700993), Ministry of Health of China (2008ZX10002-019), Natural Science Foundation of Guangdong Province (8451027501001496).

Abstract

Hepatocellular carcinoma (HCC) is a highly vascularized tumor with frequent intrahepatic metastasis. Active angiogenesis and metastasis are responsible for rapid recurrence and poor survival of HCC. We previously found that microRNA-29b (miR-29b) down-regulation was significantly associated with poor recurrence-free survival of HCC patients. Therefore, the role of miR-29b in tumor angiogenesis, invasion, and metastasis was further investigated in this study using in vitro capillary tube formation and transwell assays, in vivo subcutaneous and orthotopic xenograft mouse models, and Matrigel plug assay, and human HCC samples. Both gain- and loss-of-function studies showed that miR-29b dramatically suppressed the ability of HCC cells to promote capillary tube formation of endothelial cells and to invade extracellular matrix gel in vitro. Using mouse models, we revealed that tumors derived from miR-29b-expressed HCC cells displayed significant reduction in microvessel density and in intrahepatic metastatic capacity compared with those from the control group. Subsequent investigations revealed that matrix metalloproteinase-2 (MMP-2) was a direct target of miR-29b. The blocking of MMP-2 by neutralizing antibody or RNA interference phenocopied the antiangiogenesis and antiinvasion effects of miR-29b, whereas introduction of MMP-2 antagonized the function of miR-29b. We further disclosed that miR-29b exerted its antiangiogenesis function, at least partly, by suppressing MMP-2 expression in tumor cells and, in turn, impairing vascular endothelial growth factor receptor 2-signaling in endothelial cells. Consistently, in human HCC tissues and mouse xenograft tumors miR-29b level was inversely correlated with MMP-2 expression, as well as tumor angiogenesis, venous invasion, and metastasis. Conclusion: miR-29b deregulation contributes to angiogenesis, invasion, and metastasis of HCC. Restoration of miR-29b represents a promising new strategy in anti-HCC therapy. (HEPATOLOGY 2011;)

The discovery of microRNAs (miRNAs) has expanded our knowledge regarding the complex control of gene expression and cellular activity.1 miRNAs belong to a class of phylogenetically conserved noncoding RNAs that regulate diverse cellular processes by suppressing the expression of protein-coding genes. It is well known that dysfunction of miRNAs can result in uncontrolled cell proliferation and resistance to apoptosis.2-5 Emerging evidence also suggests that deregulation of miRNAs may contribute to tumor angiogenesis and metastasis.3 Ideally, the biomedical significance of miRNAs should be studied based on not only in vitro assays, but also in vivo models as well as human specimens. Few studies using these approaches have identified miRNAs that have proangiogenic (miR-296/93/132)6-8 activity, or possess prometastatic (miR-10b/103/107/9/30d)9-12 or antimetastatic (miR-31/200/139/122)9, 13, 14 function.

Hepatocellular carcinoma (HCC) is a highly vascularized tumor with frequent intrahepatic metastasis.15 Active angiogenesis and metastasis are responsible for rapid recurrence and poor survival of HCC.15 In a previous study we found that miR-29b down-regulation was a prevalent event in HCC tissues and was significantly associated with worse recurrence-free survival of HCC patients.2 To date, the role of miR-29b in tumor angiogenesis and metastasis is still unclear, although other groups have employed the in vitro transwell system to clarify the suppressive effect of miR-29 family on invasion of non-HCC tumor cells.16-18 In this study, both gain- and loss-of-function analyses showed that miR-29b dramatically suppressed the ability of HCC cells to promote capillary tube formation of endothelial cells (ECs) and to invade extracellular matrix (ECM) gel in vitro. We further confirmed the suppressive function of miR-29b on tumor angiogenesis, invasion, and metastasis in vivo. A recent study showed that overexpression of miR-29b suppressed MMP-2 expression in prostate cancer cell line.19 Here, we revealed that matrix metalloproteinase-2 (MMP-2) was a direct target of miR-29b in HCC cells using both in vitro and in vivo systems. We also provided evidence to demonstrate that miR-29b repressed angiogenesis, invasion, and metastasis by suppressing MMP-2. Our findings highlight the importance of miR-29b dysfunction in promoting tumor progression and recurrence, and implicate miR-29b as a potential therapeutic target for HCC.

Abbreviations

Dox, doxcycline; EC, endothelial cell; ECM, extracellular matrix; HCC, hepatocellular carcinoma; HUVECs, human umbilical vein endothelial cells; miRNA, microRNA; MMP-2, matrix metalloproteinase 2; MVD, microvessel density; SFM, serum-free medium for endothelial cells; TCM, tumor-cell conditioned medium; TIMP-2, tissue inhibitor of metalloproteinase 2; 3′-UTR, 3′-untranslated region.

Materials and Methods

Human Tissue Specimens and Tumor Cell Lines.

Human HCC and adjacent nontumor liver tissues were collected in our previous study2 from 127 patients undergoing resection of HCC at the Cancer Center, Sun Yat-sen University, P.R. China. The relevant characteristics of the studied subjects were previously reported.2 Informed consent was obtained from each patient and the study was approved by the Institute Research Ethics Committee at the Cancer Center.

Tumor cell lines were: LM620 and H2M21 (HCC), 95D (lung cancer), HCT116 (colorectal cancer), HEK293T (transformed human embryonic kidney cells). All lines were maintained in Dulbecco's modified Eagle's medium (DMEM, Invitrogen, NY) supplemented with 10% fetal bovine serum (FBS, Hyclone, Logan, UT). LM6 subline stably expressing miR-29b (LM6-miR-29b) and its control line (LM6-vec) were established using the Tet-off system (ClonTech, Palo Alto, CA), as described in Supporting Materials and Methods.

Human Umbilical Vein Endothelial Cells (HUVECs).

HUVECs were isolated as described in Supporting Materials and Methods. HUVECs were cultured in gelatin-coated flasks and maintained in serum free medium for endothelial cells (SFM, Invitrogen), supplemented with 20% FBS, 0.1 mg/mL of heparin and 0.03 mg/mL of endothelial cell growth supplement (Upstate Biotechnology, Lake Placid, NY). Primary HUVECs were used at passages 2-7 in all experiments.

RNA Oligoribonucleotides and Vectors.

All miRNA mimic and small interference RNA (siRNA) duplexes (Supporting Table 1) were purchased from Genepharma (Shanghai, P.R. China). Si-MMP2 and si-TIMP2 targeted mRNAs of human MMP-2 (GenBank accession no. NM_001127891.1) and tissue inhibitor of metalloproteinase-2 (TIMP-2,NM_003255.4), respectively. The negative control RNA duplex (NC) for both miRNA mimic and siRNA was nonhomologous to any human genome sequence. The sequence-specific miR-29b inhibitor (anti-miR-29b) and its control (anti-miR-C) were from Dharmacon (Chicago, IL).

Vectors (details in Supporting Materials and Methods): miR-29b expression vectors pc3-miR-29b and pRetroX-miR-29b; firefly luciferase reporter plasmids pGL3cm-MMP2-3′-untranslated region(3′UTR)-wildtype (WT) and pGL3cm-MMP2-3′UTR-MUT that contained wildtype and mutant 3′-UTR segment of human MMP-2, respectively; MMP-2 expression vectors pc3-MMP2.

Cell Transfections.

Reverse transfection of RNA oligoribonucleotides was performed using Lipofectamine-RNAi MAX (Invitrogen). Fifty nM of RNA duplex and 100 nM of miRNA inhibitor were used for each transfection. HEK293T transfection with plasmid DNA was conducted by calcium phosphate precipitation.

Preparation of Tumor Cell-Conditioned Medium (TCM).

Tumor cells (1 × 105) were reverse transfected with RNA oligonucleotides in a 12-well plate. Thirty-six hours after transfection, medium was removed. Cells were washed with 1 × phosphate-buffered saline (PBS) three times, and then cultured in 500 μL SFM for 12 hours for miR-29b-transfectants or 24 hours for anti-miR-29b-transfectants. Following the incubation period, TCM was collected and centrifuged at 500g to remove detached cells and then at 12,000g to discard cell debris (4°C, 10 minutes each). TCM was then stored in aliquots at −80°C until used. Each corresponding well was subsequently trypsinized and the number of live cells was counted to allow appropriate correction of TCM loading for cell equivalents. For the MMP-2 blocking assay, TCM was preincubated with MMP-2-neutralizing antibody (#MS-567-P1ABX, Thermo Scientific, Braunsweig, Germany) or the isotype-matched control immunoglobulin G (IgG) (MAB002, R&D Systems, Minneapolis, MN) for 1 hour at 37°C, before being applied to coculture with HUVECs.

Capillary Tube Formation Assay.

HUVECs (1.5 × 104) were grown in the absence or presence of 75% TCM for 10 hours at 37°C in a 96-well plate coated with Matrigel (3432-005-01, R&D Systems). The formation of capillary-like structures was captured under a light microscope. The branch points of the formed tubes, which represent the degree of angiogenesis in vitro, were scanned and quantitated in five low-power fields (100×).

In Vitro Assays of Migration and Invasion.

The 24-well Boyden chamber with 8-μm pore size polycarbonate membrane (Corning, NY) was used to analyze the migration and invasion of tumor cells. For invasion assay, the membrane was coated with Matrigel to form a matrix barrier. Wound healing assay was applied to examine the migration of HUVECs. Details are in the Supporting Materials and Methods.

Analysis of Cell Proliferation.

The proliferation of HUVECs was assessed by bromodeoxyuridine (BrdU) incorporation assay, as described in the Supporting Materials and Methods.

Nude Mice Xenograft Studies.

All experimental procedures involving animals were performed in accordance with the Guide for the Care and Use of Laboratory Animals (NIH publications Nos. 80-23, revised 1996), and according to the institutional ethical guidelines for animal experiments.

For subcutaneous xenograft model, LM6 cells (1 × 106) that were transiently transfected with miR-29b or NC duplex were suspended in 100 μL 1 × PBS and then injected subcutaneously into either side of the posterior flank of the same female BALB/c athymic nude mice at 5 weeks of age. Five nude mice were included and tumor growth was examined over the course of 35 days.

For orthotopic liver xenograft model, 3 × 106 LM6-miR-29b or LM6-vec cells were suspended in 40 μL of PBS/Matrigel (1:1) and then inoculated under the capsule of the left hepatic lobe of BALB/c nude mice. miR-29b expression was silenced by administering drinking water supplemented with 10% sucrose plus 2 mg/mL doxcycline (Dox, ClonTech).

The animals were sacrificed and tumors or livers were dissected, fixed in formalin, and embedded in paraffin. To evaluate intrahepatic metastasis, serial sections of liver were screened.

In Vivo Matrigel Plug Angiogenesis Assay.

Growth-factor-reduced Matrigel (500 μL, cat. 3433-005-01, R&D Systems) premixed with 2 × 106 LM6-miR-29b or LM6-vec cells was subcutaneously implanted into either side of the flank of the same BALB/c nude mice for 7 days, Matrigel plugs were then dissected, embedded in OCT (Miles, Elkhart, IN), and stored at −80°C.

Luciferase Reporter Assay.

HEK293T cells grown in a 48-well plate were cotransfected with 200 ng of either pcDNA3.0 or pc3-miR-29b, 10 ng of either pGL3cm-MMP2-3′UTR-WT or pGL3cm-MMP2-3′UTR-MUT, and 2 ng of pRL-TK (Promega) and then applied to luciferase assay as reported.4

Detection of MMP-2 Activity in TCM by Gelatin Zymography.

To detect the gelatinolytic activity of MMP-2, aliquots of TCM were applied to acrylamide gel containing gelatin as described in the Supporting Materials and Methods.

Semiquantitative Reverse-Transcription Polymerase Chain Reaction (RT-PCR), Quantitative Real-Time PCR (qPCR), and Western Blot.

Semiquantitative RT-PCR was performed as described,4 using primers listed in Supporting Table 1.

qPCR analysis of miR-29b was performed on a LightCycler 480 (Roche Diagnostics, Germany) using a TaqMan MicroRNA Assay kit (Applied Biosystems, Foster City, CA). All reactions were run in triplicate. The cycle threshold (Ct) values should not differ more than 0.5 among triplicates. The miR-29b level was normalized to RNU6B, which yielded a 2-ΔΔCt value.

Antibodies for western blot: mouse mAb for AKT (cat. 2967) and phospho-ser473-AKT (cat. 4051) (Cell Signaling Technology, CST, Beverly, MA); mouse mAb for ERK1/2 (cat. 610030), and phosphor-T202/Y204-ERK1/2 (cat. 612358) (BD Biosciences, Franklin Lakes, NJ); mouse mAb for MMP-2 (cat. MAB3308, Chemicon, Temecula, CA) and β-actin (cat. BM0627, Boster, Wuhan, China), rabbit polyclonal antibody against vascular endothelial growth factor receptor 2 (VEGFR2) (cat. 2479) and phosphor-Tyr1175-VEGFR2 (cat. 2478s) from CST.

Immunohistochemical (IHC) Staining.

Matrigel plug sections and paraffin-embedded tissue sections were applied to IHC using mouse mAb against human MMP-2 (cat. 35-1300Z, Invitrogen) or CD34 (cat. sc-52312, Santa Cruz Biotechnology, Santa Cruz, CA), or rat mAb for mouse CD34 (cat. 119301, BioLegend) as described in the Supporting Materials and Methods.

MMP-2 expression was evaluated under a light microscope at a magnification of 400×. For each specimen, five images of representative areas were acquired and a total of 1,000 to 2,000 tumor cells were counted. For human samples, IHC scoring was performed using a modified Histo-score (H-score), which included a semiquantitative assessment of both fraction of positive cells and intensity of staining. The intensity score was defined as no staining (0), weak (1), moderate (2), or strong (3) staining. The fraction score was based on the proportion of positively stained cells (0%-100%). The intensity and fraction scores were then multiplied to obtain H-score, which ranged from 0 to 3 and represented the level of MMP-2.

The microvessel density (MVD) in tumor tissues or Matrigel plug, which represents the degree of angiogenesis in vivo, was evaluated by staining for CD34, an endothelial cell marker. Any discrete cluster or single cell stained for CD34 was counted as one microvessel.

Enzyme-Linked Immunosorbent Assay (ELISA).

The levels of VEGFA in TCM were detected by ELISA kits (R&D Systems), as instructed by the manufacturer.

Statistical Analysis.

Data are expressed as the mean ± standard error of the mean (SEM) from at least three independent experiments. Values for capillary tube formation and luciferase activity assays were from three independent experiments performed in duplicate. Unless otherwise noted, the differences between groups were analyzed using Student's t test when only two groups were compared or by one-way analysis of variance (ANOVA) when more than two groups were compared. All statistical tests were two-sided. Differences were considered statistically significant at P < 0.05. All analyses were performed using SPSS software (Chicago, IL).

Results

miR-29b Inhibits the Capacity of Tumor Cells to Promote Capillary Tube Formation and to Invade ECM Gel In Vitro.

To explore the biological significance of miR-29b in tumor angiogenesis, in vitro capillary tube formation assay was first analyzed using LM6 and H2M. TCM from tumor cells without transfection or from cells transfected with NC or miR-29b was supplied to the culture medium for HUVECs. In the presence of TCM derived from NC- or nontransfected cells, HUVECs developed more capillary-like structures compared with those cultured in SFM (Fig. 1A,B). However, when HUVECs were incubated with TCM of miR-29b-transfectants, the branch points of capillary-like structures dramatically decreased to a level comparable to HUVECs grown in SFM (Fig. 1A,B). These results were further confirmed using 95D and HCT116 cells (Supporting Fig. 1). Moreover, TCM from miR-29b-tranfectants display much lower ability to promote the proliferation and migration of HUVECs, compared with that from NC- and nontransfected cells (Supporting Figs. 2, 3). We next examined the role of miR-29b on the motility and invasive capacity of HCC cells using transwell chamber without or with Matrigel. Restoration of miR-29b substantially reduced the number of LM6 and H2M cells that invaded through Matrigel (Fig. 1C,D). However, the total numbers of cells without transfection or transfected with NC or miR-29b were similar at the end of experiment (data not shown). Although miR-29b-transfectants displayed a tendency of reduced motility compared with control cells, the difference did not reach statistical significance (data not shown).

Figure 1.

miR-29b represses tumor angiogenesis and invasion in vitro. (A,B) Restoration of miR-29b inhibited HCC cell-promoted tube formation of HUVECs. HUVECs were cultured in the presence of SFM (panel 1) or 75% TCM from LM6 (A) or H2M (B) cells without transfection (panel 2) or from cells transfected with NC (panel 3) or miR-29b duplex (panel 4). Representative images of tube formation and the number of branch points of HUVECs are presented. (C,D) Overexpression of miR-29b inhibited invasion of HCC cells. The 2 × 105 LM6 (C) or 8 × 104 H2M (D) cells without transfection (panel 1) or transfected with NC (panel 2) or miR-29b duplex (panel 3) were applied to transwell chamber coated with Matrigel and then incubated for 24 hours. Representative images and the number of invaded cells are shown. The 8-μm pores of the transwell polycarbonate membranes are visible in the background. Scale bar = 100 μm. *P <0.05; **P <0.01.

To verify the findings from gain-of-function study, loss-of-function analysis was carried out using anti-miR-29b, which obviously decreased the endogenous miR-29b level (Supporting Fig. 4). Suppression of cellular miR-29b not only enhanced the proangiogenic effect but also promoted the invasive activity of HCC cells (Supporting Fig. 5). These data suggest the suppressive effect of miR-29b on tumor angiogenesis and invasion.

miR-29b Inhibits Tumor Angiogenesis and Metastasis In Vivo.

To validate the role of miR-29b in vivo, we conducted subcutaneous injection and orthotopic liver implantation of HCC cells in nude mice. Subcutaneous tumor xenografts grown from NC or miR-29b duplex-transfected cells were first analyzed. Consistent with our previous observation in HepG2 cell,2 miR-29b restoration significantly repressed the subcutaneous growth of LM6 cells (Supporting Fig. 6). Moreover, miR-29b-transfectant-derived tumors displayed much smaller and fewer blood vessels compared with those of control group (Fig. 2A). Matrigel plug assay was then used to confirm the in vivo antiangiogenesis effect of miR-29b. Matrigel containing LM6 subline with Tet-off-inducible miR-29b expression (LM6-miR-29b, Supporting Fig. 7) showed much lower number of vessels compared with that contained LM6-vec (Fig. 2B).

Figure 2.

miR-29b suppresses tumor angiogenesis and metastasis in vivo. (A) Subcutaneous xenograft tumors of miR-29b duplex-transfectants displayed lower microvessel density. Tumor tissues grown from NC or miR-29b duplex-transfected LM6 cells were immunohistochemically stained for murine CD34 (mCD34). (B) Matrigel plug with LM6-miR-29b cells showed lower microvessel density. Cryostat frozen sections were stained for mCD34. (C,D) Restoration of miR-29b inhibited tumor angiogenesis and metastasis in orthotopic xenograft model. LM6-miR-29b or its control LM6-vec was transplanted in the left hepatic lobe. In (C), sections of orthotopic primary tumors were stained for mCD34. For (A-C), the three most intensely vascularized areas or the whole area of a small tumor were evaluated under light microscopy at a magnification of 100×. Representative images and the quantitative MVD measurement of control and miR-29b groups are presented. In (D) hematoxylin-eosin-stained sections of intrahepatic metastatic nodules are presented. The maximum diameter of intrahepatic metastatic nodules of each case was determined by screening serial sections of liver from the case that developed tumor in the implanted site. Arrows indicate metastatic nodules. For (C,D), miR-29b expression was induced and maintained for 27 days (early) or 9 days (late) before mice were sacrificed. Scale bar = 50 μm. *P <0.05; **P <0.01

To further confirm the above findings, an orthotopic liver xenograft model was applied. Mice injected with LM6-miR-29b cells were further divided into two groups: miR-29b-early and -late induction groups, based on the timepoint when miR-29b expression was induced. For miR-29b-early induction group (n = 14), miR-29b expression was silenced by Dox for the first 14 days after implantation, then induced and maintained for 27 days by Dox withdrawal. For miR-29b-late induction groups (n = 9), miR-29b was induced at day 33 and maintained for 9 days before mice were sacrificed. Compared with the control group (n = 14), tumor incidence was significantly lower in the miR-29b-early induction group (11/14 versus 9/14 mice), but a similar rate was found in the miR-29b-late induction group (11/14 versus 7/9 mice). Tumor size was also reduced in the miR-29b expression group in a dose-dependent manner (Supporting Fig. 8). Furthermore, compared with control, both miR-29b expression groups showed much less MVD (Fig. 2C), significantly decreased occurrence of intrahepatic metastasis (control versus miR-29b-late versus -early induction groups: 8/11 versus 4/7 versus 4/9), and reduced size of metastatic nodules (Fig. 2D).

Collectively, these findings indicate that miR-29b suppresses both tumor angiogenesis and metastasis in vivo.

miR-29b Directly Regulates MMP-2 Expression.

We then explored the molecular mechanisms responsible for the multiple function of miR-29b. Potential target genes of miR-29b were first predicted using databases, including TargetScan, PicTar, and miRanda. Among them, MMP-2 was chosen for further experimental validation, not only because it was identified as a target of miR-29b by all three databases, but also due to its frequent overexpression in tumor tissues and well-known importance in both tumor angiogenesis and metastasis.22-25 Dual-luciferase reporter analysis showed that coexpression of miR-29b significantly inhibited the activity of firefly luciferase that carried wildtype but not mutant 3′-UTR of MMP-2 (Fig. 3A,B), indicating that miR-29b may suppress gene expression through its binding sequence at 3′-UTR of MMP-2. Moreover, introduction of miR-29b diminished the expression of cellular MMP-2 protein (Fig. 3C). Furthermore, gelatin zymography showed that, compared with TCM obtained from control cells, those from miR-29b-transfectants displayed a significant reduction in MMP-2 activity (Fig. 3D), whereas TCM from anti-miR-29b-transfectants revealed up-regulated MMP-2 activity (Fig. 3E). Consistently, in the orthotopic liver implanted model primary tumors of LM6-miR-29b cells showed much lower MMP-2 expression, compared with those of control cells (Fig. 3F). These findings indicate that miR-29b may negatively regulate the expression of MMP-2 by directly targeting its 3′-UTR.

Figure 3.

MMP-2 is a direct target of miR-29b. (A) miR-29b and its putative binding sequence in the 3′-UTR of MMP-2. The mutant miR-29b-binding site was generated in the complementary site for the seed region of miR-29b. (B) Analysis of luciferase activity. HEK293T cells were cotransfected with pcDNA3.0 (pc3) or pc3-miR-29b, firefly luciferase reporter containing either a wildtype or a mutant 3′-UTR (indicated as Wt or Mut on the X axis), and a Renilla luciferase expressing construct (as internal control to calibrate the differences in both transfection and harvest efficiencies). The firefly luciferase activity of each sample was normalized to the Renilla luciferase activity. The normalized luciferase activity of pc3-transfectants in each experiment was set as relative luciferase activity 1. Therefore, no error bar is shown for pc3-transfectants. *P <0.05, compared with pc3-transfectants. (C) Effects of miR-29b overexpression on the level of cellular MMP-2. LM6 and H2M cells without transfection (lane 1) or cells transfected with NC (lane 2) or miR-29b (lane 3) were analyzed by western blot 48 hours after transfection. β-Actin, internal control. (D,E) Analysis of MMP-2 activity in TCM by gelatin zymography. TCM was collected from tumor cells without transfection (lane 1) or from cells transfected with NC or anti-miR-C (lane 2) or with miR-29b or anti-miR-29b (lane 3). For (C-E), the results were reproducible in three independent experiments and the representative images are shown. (F) Analysis of MMP-2 expression in orthotopic primary tumors by IHC. Brown signal in IHC was considered as positive staining for MMP-2. Scale bar = 50 μm. ***P <0.001, comparison between control and miR-29b-early induction groups.

miR-29 Exerts Its Function by Suppressing MMP-2 Signaling.

The role of MMP-2 in miR-29b-mediated phenotypes was then evaluated. First, we examined whether blockage of MMP-2 could mimic the effect of miR-29b expression. As expected, TCM that was preincubated with MMP-2-neutralizing antibody displayed a decreased capacity to promote tube formation of HUVECs (Fig. 4A). Also, LM6 cells treated with this antibody display less invasive activity (Fig. 4B). These results phenocopied those of enhanced miR-29b expression. On the other hand, overexpression of MMP-2 in miR-29b-transfectants recovered MMP-2 activity in TCM (Supporting Fig. 9), and attenuated the inhibitory effect of miR-29b on angiogenesis (Fig. 4C) and invasion (Fig. 4D).

Figure 4.

MMP-2 is involved in the dual inhibitory function of miR-29b on tumor angiogenesis and invasion. (A,B) Blockage of MMP-2 mimicked the antiangiogenesis and antiinvasion effect of miR-29b overexpression. In (A), LM6-derived TCM without treatment (panel 1) or preincubated with control IgG (panel 2) or MMP-2-neutralizing antibody (panel 3) were added to HUVECs at a final concentration of 75%. In (B) LM6 cells without treatment (panel 1) or incubated with control IgG (panel 2) or MMP-2-neutralizing antibody (panel 3) were applied to transwell chamber coated with Matrigel and then incubated for 24 hours. (C,D) Reintroduction of MMP-2 antagonized the antiangiogenesis and antiinvasion effect of miR-29b. In (C) TCM derived from LM6 cells cotransfected with NC/vector (lane 1), miR-29b/vector (lane 2), NC/MMP2 (lane 3), or miR-29b/MMP2 (lane 4) were added to HUVECs at a final concentration of 75%. In (D) LM6 cells transfected as in (C) were applied to transwell chamber coated with Matrigel. Scale bar = 100 μm. *P <0.05; **P <0.01; ***P <0.001.

We further analyzed the associations among miR-29b level, MMP-2 expression, angiogenesis, and venous invasion in human HCC tissues. Samples from 127 HCC cases, whose miR-29b levels had been analyzed previously,2 were stained immunohistochemically for MMP-2 and CD34 (Fig. 5A). Obviously, the miR-29b level was inversely correlated with MMP-2 expression (Fig. 5B; Supporting Fig. 10A); miR-29b down-regulation was significantly associated with higher MVD (Fig. 5C; Supporting Fig. 10B); HCC with venous invasion displayed much lower miR-29b expression compared with those without venous invasion (Fig. 5D). Together with our previous observation that a decreased miR-29b level was associated with recurrence of HCC,2 we suggest that down-regulation of miR-29b may be responsible for the increased level of MMP-2 in human HCC tissues, which in turn promotes angiogenesis, invasion, and metastasis of HCC.

Figure 5.

The level of miR-29b is inversely correlated with MMP-2 expression, MVD, and venous invasion in human HCC tissues. (A) The representative images of IHC staining for MMP-2 and CD34. Images were captured at ×400 for MMP-2 and ×200 for CD34 staining. Scale bar = 50 μm. (B) Tumors with lower miR-29b level displayed higher MMP-2 expression. (C) HCC with higher MVD showed lower miR-29b expression. (D) HCC with venous invasion revealed lower miR-29b expression. The level of mature miR-29b was previously analyzed by qPCR,2 and the median of all 127 cases was chosen as the cutoff point for separating low-miR-29b (n = 64) from high-miR-29b expressing tumors (n = 63) in (A-C). Expression of MMP-2 was quantified based on IHC staining, using a modified Histo-score as described in the Materials and Methods. MVD was evaluated based on CD34 staining and was determined from the five most intensely vascularized areas of each section at a magnification of 200×. The median of all 127 cases was chosen as the cutoff point for separating low-MVD (n = 63) from high-MVD tumors (n = 64) in (C). In (D), absence (n = 64) or presence (n = 63) of macroscopic or microscopic venous invasion was indicated as “−” and “+”, respectively. For (B,D), the Mann Whitney test was performed. The central horizontal line, mean value; error bar, SEM. **P <0.01; ***P <0.001.

It has been shown that the local balance between MMPs and their physiological inhibitors affects angiogenesis process in vivo.26, 27 The VEGFR2-signaling pathway regulates proliferation, migration and survival of ECs by way of ERK and AKT. Proangiogenic signals, such as VEGF, induce the phosphorylation and activation of VEGFR2, which then phosphorylates ERK and AKT, and subsequently promotes tube formation of ECs.28, 29 The natural inhibitor of MMP-2, TIMP-2,22 can promote VEGFR2 dephosphorylation by way of protein tyrosine phosphatase Shp-1, thereby blocking VEGFR2-signaling.30-32 However, this effect is abolished when TIMP-2 is bound by pro-MMP-2.31, 32 Therefore, we first explored whether down-regulation of TIMP-2 could affect the function of miR-29b. Dramatically, TIMP-2 knockdown (Supporting Fig. 11A) abrogated the antiangiogenic effect of miR-29b (Supporting Fig. 11B).

We further evaluated whether miR-29b repressed tumor angiogenesis by inhibiting MMP-2 in tumor cells and, in turn, abrogating VEGFR2-signaling in ECs. In agreement with the above observation on tube formation, compared with the control (Fig. 6A,B, lane 1), HUVECs that were incubated with TCM from nontransfected or NC-transfected HCC cells (Fig. 6A,B, lanes 2 and 3) had significantly increased phosphorylation of VEGFR2, ERK, and AKT. However, the observed TCM-promoted VEGFR2-signaling in HUVECs was dramatically attenuated when miR-29b was restored in tumor cells (Fig. 6A,B, lane 4). This effect was replicated when MMP-2 in tumor cells was silenced by siRNA (Supporting Fig. 12A,B; Fig. 6A,B, lane 5) or when tumor cell-derived TCM was preincubated with MMP-2-neutralizing antibody (Supporting Fig. 13). In contrast, TCM from anti-miR-29b-transfectants caused an enhanced VEGFR2-signaling in HUVECs (Fig. 6C). Furthermore, TIMP-2 knockdown rescued the suppressive effect of miR-29b on VEGFR2-signaling (Fig. 6D).

Figure 6.

miR-29b exerts its antiangiogenic function by inhibiting VEGFR2-signaling in endothelial cells. (A,B) Restoration of miR-29b repressed VEGFR2-signaling. HUVECs were cultured in the presence of SFM (lane 1), or TCM from nontransfected cells (lane 2), or from NC (lane 3), miR-29b (lane 4) or si-MMP2 (lane 5) transfected LM6 (A) or H2M (B) cells. (C) Antagonism of miR-29b enhanced VEGFR2 signaling. HUVECs were cultured in the presence of SFM (lane 1), TCM from nontransfected cells (lane 2) or from anti-miR-C (lane 3), or anti-miR-29b (lane 4) transfected LM6 cells. (D) Knockdown of TIMP-2 attenuated the suppressive effect of miR-29b on VEGFR2-signaling. HUVECs were cultured in the presence of TCM from LM6 cells cotransfected with NC/NC (lane 1), NC/miR-29b (lane 2), si-TIMP2/NC (lane 3), or si-TIMP2/miR-29b (lane 4). For (A-D), HUVECs were grown in SFM supplemented with 1% FBS for 12 hours and then cultured in the presence of SFM or 75% TCM for 20 minutes, before immunoblotting analysis for phosphor-Tyr1175-VEGFR2, VEGFR2, phospho-ser473-AKT, AKT, phosphor-T202/Y204-ERK1/2, and ERK1/2 expression. All results were reproducible in three independent experiments and the representative immunoblots are shown. β-Actin, internal control.

Because VEGFA is a pivotal activator of VEGFR2 pathway, we further evaluated whether the VEGFA level in TCM of miR-29b-transfectants was different from that of control cells. ELISA assay revealed significant VEGFA accumulation in TCM, but no difference in VEGFA level was found among cells without transfection or transfected with NC, miR-29b, or si-MMP2 (Supporting Fig. 14).

Taken together, our data imply that miR-29b may suppress tumor angiogenesis, invasion, and metastasis by repressing MMP-2 signaling.

Discussion

Here we demonstrate that miR-29b is capable of repressing tumor angiogenesis, invasion, and metastasis, and miR-29b exerts its multiple inhibitory functions, at least partly, by directly suppressing MMP-2 expression. This is the first attempt to illuminate the role of miR-29b deregulation in tumor angiogenesis and metastasis, using both in vitro and in vivo models.

Angiogenesis is essential for tumor growth and metastasis, whereas metastasis is the major cause of cancer death.15, 29 Identification of novel antiangiogenesis or antimetastasis targets will, therefore, have enormous clinical applications.29 Studies based on clinical samples as well as in vitro and in vivo models have identified a limited number of miRNAs that display proangiogenic (miR-296/93/132)6-8 activity. However, the conclusion that miR-296 and miR-132 regulate angiogenesis is drawn from the observations that ectopic expression of these miRNAs in ECs themselves can affect the response of ECs to angiogenic factors. Tumor cell is the critical initiator and promoter of angiogenesis. Therefore, it is crucial to elucidate whether and how the dysfunction of miRNAs in tumor cells affects tumor angiogenesis. Our data suggest that miR-29b deregulation in HCC cells may result in enhanced MMP-2 level in the tumor microenvironment, which in turn activates the VEGFR-2 signaling in ECs and thereby promotes angiogenesis. Moreover, we also show that miR-29b exerts multiple inhibitory effects on angiogenesis, invasion, and metastasis by suppressing the expression of only one molecule. Our data not only supply novel insights regarding miR-29b function and the mechanisms of hepatocarcinogenesis, but may also have considerable implications in cancer therapy.

Based on orthotopic xenograft mouse models, tumors derived from miR-29b-transfectants are obviously smaller than that of the control group, and both tumor incidence and tumor size are inversely correlated with the duration of miR-29b expression. This inhibitory function of miR-29b on tumor growth may result from both increased apoptosis and decreased angiogenesis. Although fewer cells and less metabolic demand in smaller tumors may lead to decreased angiogenesis and metastasis, our data clearly suggest that miR-29b can directly repress tumor angiogenesis and metastasis by targeting MMP-2. This conclusion is based on the following evidence. First, miR-29b restoration not only significantly decreases both cellular expression of MMP-2 and the MMP-2 activity of TCM, but also attenuates the invasive capacity and the proangiogenic activity of HCC cells in vitro. Furthermore, MMP-2 knockdown phenocopies the effect of miR-29b expression, whereas reintroduction of MMP-2 antagonizes the function of miR-29b. Second, the Matrigel plug assay, in which tumor cells are mixed with growth-factor-reduced Matrigel and will not proliferate, also revealed significantly antiangiogenic function of miR-29b. Third, observations from in vitro cell models, in vivo mouse models and human samples, all disclose significant inverse correlation of miR-29b expression with tumor angiogenesis and metastasis.

The miR-29 family consists of three members: miR-29a, miR-29b, and miR-29c (miR29a/b/c). Like other miRNA family members, miR-29a/b/c display high sequence similarity and share a common seed sequence for target recognition. We have previously shown that all three members are frequently down-regulated in HCC, and restoration of either of them significantly increases the sensitivity of HCC cells to apoptosis.2 The in vitro studies from other groups have pinpointed the suppressive effect of miR-29 on proliferation, apoptosis, invasion, and migration of non-HCC tumor cells.16-18, 33 Here, both in vitro and in vivo analysis suggest multiple inhibitory effects of miR-29b on HCC angiogenesis, invasion, and metastasis. It is intriguing to find that a single miRNA can regulate different phenotypes of cancer cells and that such an miRNA may be a promising molecular target for anticancer therapy.

It is well known that MMP-2 activation results in degradation of ECM, which facilitates the invasion and metastasis of tumor cells.22 MMP-2 also facilitates the remodeling of ECM and the release of ECM-bound growth factors (VEGFA, FGF, etc.), which assist the migration and proliferation of ECs. MMP-2 overexpression has been observed in different types of malignancy, including HCC.34, 35 It has been shown that miR-29b can suppress MMP-2 expression in prostate cancer cells.19 Here, we demonstrate that MMP-2 is a direct functional target of miR-29b in HCC cells, based on in vitro and in vivo studies: miR-29b directly regulates MMP-2 expression by binding to its 3′-UTR; miR-29b down-regulation is associated with enhanced level of MMP-2, MVD, and venous invasion in human HCC tissues; restoration of miR-29b represses MMP-2 expression and inhibits angiogenesis and metastasis of HCC cells in a mouse model. These results suggest that miR-29b dysfunction accounts for one of the mechanisms responsible for MMP-2 overexpression, and in turn, the increased angiogenesis, invasion, and metastasis of HCC. Furthermore, our results also disclose that silencing of TIMP-2 can attenuate the antiangiogenic effect of miR-29b, indicating that the balance between MMP-2 and TIMP-2 is critical for the antiangiogenic effect of miR-29b.

In summary, we investigated the effect of miR-29b in tumor angiogenesis, invasion, and metastasis and its underlying mechanisms. Our data suggest that miR-29b deregulation may play an important role in rapid growth and recurrence of HCC. Restoration of miR-29b may represent a promising strategy for anti-HCC therapy.

Ancillary