© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Carcinogenesis vol.35 no.5 pp.1144–1153, 2014doi:10.1093/carcin/bgu021Advance Access publication January 24, 2014

The oncoprotein HBXIP enhances angiogenesis and growth of breast cancer through modulating FGF8 and VEGF

Fabao Liu, Xiaona You1, Yue Wang, Qian Liu, Yunxia Liu1, Shuqin Zhang1, Lingyi Chen2, Xiaodong Zhang1 and Lihong Ye*

Department of Biochemistry, 1Department of Cancer Research and 22011 Collaborative Innovation Center of Tianjin for Medical Epigenetics, College of Life Sciences, Nankai University, Tianjin 300071, P.R. China

*To whom correspondence should be addressed. Department of Biochemistry, College of Life Sciences, Nankai University, 94 Weijin Road, Tianjin 300071, P.R. China. Tel/Fax: +86-22-23501385; Email: yelihong@nankai.edu.cnCorrespondence may also be addressed to Xiaodong Zhang. Tel/Fax: +86 -22-23501385; Email: zhangxd@nankai.edu.cnCorrespondence may also be addressed to Lingyi Chen. Tel/Fax: +86-22 -23505821; Email: lingyichen@nankai.edu.cn

Tumor angiogenesis plays an important role in the development of cancer. Previously, we reported that hepatitis B X-interacting protein (HBXIP) functioned as an oncoprotein in breast can-cer. However, the role of HBXIP in angiogenesis in breast can-cer remains poorly understood. In the present study, we show that the oncoprotein HBXIP plays crucial roles in the event. We observed that the expression levels of HBXIP were positively correlated with those of fibroblast growth factor 8 (FGF8) or vascular endothelial growth factor (VEGF) in clinical breast cancer tissues. Then, we demonstrated that HBXIP was able to upregulate FGF8 through activation of its promoter involv-ing direct binding to cAMP response element-binding protein (CREB) in breast cancer cells and thereby increased its secre-tion. Strikingly, we identified another pathway that HBXIP upregulated FGF8 and VEGF through inhibiting miRNA-503, which directly targeted 3′ untranslated region of FGF8 or VEGF mRNA in the cells. Moreover, we revealed that HBXIP-induced FGF8 could upregulate VEGF expression through acti-vating phosphoinositide 3-kinase (PI3K)/Akt/hypoxia-inducible factor 1-alpha (HIF1α) signaling and increase its secretion. In function, matrigel angiogenesis assay and hemoglobin content analysis uncovered that HBXIP-enhanced FGF8/VEGF boosted tumor angiogenesis and growth in breast cancer in vitro and in vivo in a paracrine/autocrine manner. Thus, we conclude that HBXIP enhances angiogenesis and growth of breast cancer through modulating FGF8 and VEGF. Our finding provides new insights into the mechanism of tumor angiogenesis in breast cancer. Therapeutically, HBXIP may serve as a novel target of tumor angiogenesis.

Introduction

Angiogenesis, the sprouting of new capillaries from preexisting blood vessels, is a vital process in physiologic and pathophysiologic condi-tions, including embryonic development, wound healing, atheroscle-rosis, diabetic retinopathy and tumor progression (1). In cancer, the tumor microenvironment is composed of numerous cell types that influence tumor angiogenesis. Tumor cells modulate their micro-environment by releasing growth factors and cytokines to activate

quiescent, normal cells surrounding them and trigger a cascade of events that rapidly becomes deregulated (2).

One of the best characterized modulators of angiogenesis is the heparin-binding fibroblast growth factor (FGF), which contributes to autocrine and paracrine growth stimulation to regulate cell prolifera-tion, survival, migration and differentiation during development and carcinogenesis (3,4). It has been reported that dysregulation of FGF signaling is associated with many developmental syndromes and human cancers (3,5). FGF8, originally isolated from the conditioned medium of an androgen-dependent mouse mammary carcinoma line (SC-3), plays a vital role in embryogenesis and organogenesis (6). High frequency of FGF8 expression has been detected in clinical breast cancer tissues (7,8). FGF8b is the primary isoform detected in breast cancer (8). Overexpression of FGF8b leads to increased tumor angiogenesis, growth and metastasis (9,10).

Vascular endothelial growth factor (VEGF), another best charac-terized modulator of angiogenesis, is essential for angiogenesis and metastatic growth in pathophysiology, which has been associated with poor prognosis in breast cancer (11,12). Its expression is modulated not only by many transcriptional factors and cytokines but also by several growth factors, such as transforming growth factor-β, epi-dermal growth factor, platelet-derived growth factor-BB and FGFs (12–15). Previous reports showed that VEGF and FGF system, acting through distinct receptor kinases, might function in a synergistic man-ner to enhance angiogenesis (16,17).

Mammalian hepatitis B X-interacting protein (HBXIP) is originally identified because of its interaction with the C terminus of the hepa-titis B virus X protein (18). It can collaborate with cytosolic survivin to control cell apoptosis and division (19). HBXIP is also a regula-tor of centrosome dynamics and cytokinesis to mediate cell growth (20). In addition, HBXIP is identified as a regulator component that is required for mTORC1 activation by amino acids (21). Besides, HBXIP is involved in angiogenesis of hepatoma HepG2 cells (22), but the molecular mechanisms are not well documented. Our group has reported that HBXIP is able to import into the nucleus in breast cancer cells and serves as a coactivator of transcriptional factors to promote cell proliferation and migration of breast cancer (23–27). However, the role of HBXIP and its mechanism in promoting angiogenesis in breast cancer remain unclear.

In this study, we are interested in the role of HBXIP in angiogenesis in breast cancer. Our data show that HBXIP can markedly enhance angiogenesis and growth of breast cancer through modulating FGF8 and VEGF. Our findings provide new insights into the mechanism of tumor angiogenesis in breast cancer.

Materials and methods

Cell culture, tissue specimen and oligonucleotidesHuman umbilical vein endothelial cells (HUVECs), SK-BR-3, MCF-7 and MCF-7-HBXIP (MCF-7 stably transfected with the HBXIP) cells were cul-tured in RPMI medium 1640 (Gibco, Grand Island, NY) with 10% fetal bovine serum. HBL-100, MDA-MB-231 and MDA-MB-435 were cultured in Dulbecco’s modified Eagle’s medium (Gibco) with 10% fetal bovine serum. Forty pairs of tumorous and adjacent non-tumorous breast tissues were collected from patients undergoing resection for breast cancer in Cancer Hospital of Tianjin Medical University. Informed consent was obtained from each patient and the study was approved by the Institutional Research Ethics Committee in Nankai University.

The siRNAs targeting human HBXIP (si-HBXIP)/cAMP response element-binding protein (CREB) (si-CREB)/FGF8 (si-FGF8)/VEGF (si-VEGF), con-trol siRNA (si-Control), miR-503 mimics (miR-503), negative control (NC), miR-503 inhibitor (anti-miR-503) and inhibitor negative control (anti-miR-NC) were synthesized by RiboBio (Guangzhou, China). All oligonucleotide sequences are listed in Supplementary Table  1, available at Carcinogenesis Online.

Abbreviations: CREB, cAMP response element-binding protein; ELISA, enzyme-linked immunosorbent assay; EMSA, electrophoretic mobility shift assay; FGF8, fibroblast growth factor 8; HBXIP, hepatitis B X-interacting pro-tein; HIF1α, hypoxia-inducible factor 1-alpha; HUVEC, human umbilical vein endothelial cell; IHC, immunohistochemistry; NC, negative control; PI3K, phosphoinositide 3-kinase; qRT–PCR, quantitative reverse transcription–PCR; UTR, untranslated region; VEGF, vascular endothelial growth factor.

1144

The oncoprotein HBXIP enhances angiogenesis and growth of breast cancer

Immunohistochemistry analysis, western blot analysis, reverse transcription–PCR, quantitative reverse transcription–PCR and enzyme-linked immunosorbent assayThese experiments of immunohistochemistry (IHC) analysis, western blot analysis, reverse transcription–PCR (RT–PCR), quantitative RT–PCR (qRT–PCR) and enzyme-linked immunosorbent assay (ELISA) were carried out as described previously (23,28). The breast cancer tissue microarrays (No. 08C14) were obtained from Xi’an Aomei Biotechnology Co., Ltd (Xi’an, China). These microarrays were composed of 49 breast cancer tissue samples and 1 normal breast tissue. For IHC analysis, tumor slides (thickness: 5–10 µm) were fixed in 4% paraformaldehyde for 2 days. Sections were stained using antibodies directed against Ki67 (Santa Cruz Biotechnology, Santa Cruz, CA) and CD31 (Abcam, Cambridge, UK). For western blot analysis, the primary antibod-ies were mouse anti-HBXIP (Abcam), rabbit anti-FGF8 (Proteintech Group, Chicago, IL), rabbit anti-VEGF (Proteintech Group), mouse anti-CREB (Santa Cruz), rabbit anti-Phospho-CREB (pS133) (Epitomics, Burlingame, CA), rab-bit anti-HIF1α (Proteintech Group), rabbit anti-Akt (Proteintech Group), rab-bit anti-phospho-Akt (Ser473) (Cell Signaling Technology, Beverly, MA) and mouse anti-β-actin (Sigma–Aldrich, St Louis, MO). Forskolin is an activator of adenylyl cyclase (Beyotime Institute of Biotechnology, Haimen, China). 8-Bromoadenosine 3′5′-cyclic monophosphate (8-Br-cAMP) is a cyclic AMP analog (Sigma–Aldrich). Wortmannin is a phosphoinositide 3-kinase (PI3K) inhibitor (Santa Cruz). For RT–PCR and qRT–PCR, total RNA was extracted from the cells using Trizol reagent (Invitrogen, Carlsbad, CA). The first-strand cDNA was synthesized by PrimeScript reverse transcriptase (TaKaRa Bio, Dalian, China) following the manufacturer’s instructions. The primers are listed in Supplementary Table  1, available at Carcinogenesis Online. qRT–PCR was performed using double-stranded DNA-specific SYBR Premix Ex TaqTM II Kit (TaKaRa Bio) according to the instructions of manufacturer. For ELISA, breast cancer cells were seeded in six-well plates at a density of 2 × 105 cells per well and treated with plasmids (or siRNAs or reagents), the con-ditioned media were collected after 24 h of incubation. ELISA was employed to quantitatively examine FGF8 and VEGF levels in the conditioned media, according to the manufacturer’s instructions with a detection limit of 5 pg/ml.

Luciferase reporter constructs and luciferase reporter gene assaysThese experiments were carried out as described previously (23,26,28). The constructs of FGF8 promoter fragment were named pGL3-FGF8-P1 (−1454/−317), pGL3-FGF8-P2 (−1294/−317), pGL3-FGF8-P3 (−1080/−317) and pGL3-FGF8-P4 (−904/−317), respectively. The mutant construct of pGL3-FGF8-P3 (termed pGL3-FGF8-P3-MUT) carried a substitution of four nucleotides within the binding sites of CREB. The construct of VEGF pro-moter was named pGL3-VEGF-P. Four nucleotides within the binding sites of hypoxia-inducible factor 1-alpha (HIF1α) were deleted as the mutant con-struct of pGL3-VEGF-P (termed pGL3-VEGF-P-ΔHRE). The 3′ untranslated regions (UTRs) of FGF8 (134 bp) and VEGF (405 bp) were subcloned into the XbaI/FseI site of the pGL3-Control vector to generate FGF8 3′UTR and VEGF 3′UTR. Mutant constructs of FGF8 and VEGF 3′UTRs were termed FGF8 3′UTR-MUT and VEGF 3′UTR-MUT. All primers are listed in Supplementary Table 1, available at Carcinogenesis Online.

Chromatin immunoprecipitation, coimmunoprecipitation, pull-down and electrophoretic mobility shift assaysThese experiments were described previously (23,29). For the chromatin immunoprecipitation assay, protein–DNA complexes were immunoprecipi-tated with anti-HBXIP (or anti-CREB) antibody. Normal mouse IgG was used as a negative control. For the coimmunoprecipitation assay, cells were washed three times with phosphate-buffered saline and then lysed. The lysates incu-bated at 4°C with respective primary antibodies for 4 h. And then the protein G beads were added and incubated for an additional 4 h. After six rounds of washes with the same lysis buffer, coprecipitated proteins were resuspended and then analyzed by sodium dodecyl sulfate–polyacrylamide gel electropho-resis and visualized by immunoblotting. For pull-down assays, recombinant GST-tagged CREB and His-tagged HBXIP proteins were used. For the elec-trophoretic mobility shift assay (EMSA), nuclear extracts from MCF-7 and MCF-7-HBXIP cell lines were prepared. Oligonucleotides of the FGF8 pro-moter were labeled with 32P-γATP using T4 polynucleotide kinase. The EMSA reactions were carried out as described previously (23).

Tube formation and anchorage-independent growth assaysTube formation, an in vitro matrigel angiogenesis assay, was performed as described previously (30). Briefly, 24-well plates were coated with 200 µl per well growth factor-reduced matrigel (BD Biosciences, Franklin Lakes, NJ). HUVECs monolayers were nutrient starved for 4 h in serum-free RPMI 1640 medium before they were resuspended and seeded at 5 × 104 cells per well in the presence of conditioned media. After 6 h, images of forming capillary-like structures were captured. For anchorage-independent growth assay on soft

agar, transfected MCF-7 cells (3 × 104) were seeded in six-well plates, with a bottom layer of 0.6% low-melting-temperature agar in RPMI 1640 and a top layer of 0.35% agar in RPMI 1640. The plates were incubated at 37°C and 5% CO2 for 21 days, after that the cultures were inspected and photographed.

Tumor xenograft in mice and hemoglobin content analysisNude mice were housed and treated according to guidelines established by the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Four-week-old female BALB/c athymic nude mice (Experiment Animal Center of Peking, China) (each group, n  =  5) were subcutaneously injected with (4 × 105) MCF-7 treated with si-Control or MCF-7-HBXIP treated with si-control (or si-FGF8 or si-VEGF) cells. Tumor growth after 10 days from injection was monitored by measuring the length (L) and width (W) with calipers every 4 days. After 4 weeks, the animals were killed to remove tumors for analysis. The tumors were measured and calculated with the formula: (L × W2) × 0.5. Hemoglobin concentration was determined as described previously (31).

Statistical analysisValues were expressed as the mean ± SD of at least three independent experi-ments. Statistical significance was assessed by comparing mean values (± SD) using Student’s t-test. P <0.05 was considered statistically significant. FGF8 expression in breast cancer tissues and their corresponding non-tumorous tissues was analyzed using Wilcoxon signed-rank test. Pearson’s correlation coefficient was used to determine the expression correlation of genes in clini-cal breast cancer tissues.

Results

HBXIP upregulates FGF8 in breast cancer cells and increases its secretionPreviously, we reported that HBXIP was overexpressed in clinical breast cancer tissues and was able to enhance proliferation and migra-tion of breast cancer cells (23–26,28). In this study, we try to explore whether HBXIP contributes to angiogenesis in breast cancer. FGF8 plays an important role in the tumorigenesis by enhancing tumor angiogenesis in prostate and breast cancers (7,9,10). Accordingly, we are interested in the relationship between HBXIP and FGF8 in breast cancer. Our data showed that the positive rates of HBXIP and FGF8 were 81.6% (40/49) and 73.5% (36/49) in clinical breast can-cer tissues by IHC staining using tissue microarrays, respectively (Figure 1A and Supplementary Table 2, available at Carcinogenesis Online). Interestingly, the positive rate of FGF8 was 87.5% (35/40) in HBXIP-positive tissues. In addition, we validated that the mRNA levels of FGF8 were much higher in breast cancer tissues relative to their non-cancerous counterparts (P  <  0.01, Wilcoxon signed-rank test, Supplementary Figure 1A, available at Carcinogenesis Online) (7,8). Additionally, qRT–PCR assays revealed that the expression lev-els of FGF8 were significantly correlated with those of HBXIP in five breast cancer cell lines (r = 0.996, P < 0.001, Pearson’s correlation coefficient, Supplementary Figure  1B, available at Carcinogenesis Online). Thus, our data suggest that the expression levels of FGF8 are significantly associated with those of HBXIP in breast cancer tissues.

On the basis of the correlation between HBXIP and FGF8 in breast cancer tissues, we hypothesized that HBXIP might upregu-late FGF8 in breast cancer cells. To investigate the effect of HBXIP on FGF8 expression, we performed overexpression or depletion of HBXIP by transfection with HBXIP plasmid or si-HBXIP in breast cancer cell lines. Our data showed that overexpression of HBXIP could remarkably upregulate FGF8 in MCF-7 cell line (Figure  1B and Supplementary Figure 1C, available at Carcinogenesis Online). Meanwhile, the secretion levels of FGF8 were increased in the conditioned media in the system (Figure  1C and Supplementary Figure 1D, available at Carcinogenesis Online). Conversely, depletion of HBXIP resulted in the downregulation of FGF8 in MDA-MB-231 and MCF-7-HBXIP cells in a dose-dependent manner (Figure  1D and Supplementary Figure 1E, available at Carcinogenesis Online). Meanwhile, the secretion levels of FGF8 were decreased in the con-ditioned media in the system as well (Figure 1E and Supplementary Figure 1F, available at Carcinogenesis Online).

1145

F.Liu et al.

Given that HBXIP was a coactivator of transcriptional factors in breast cancer cells (23–26), we speculated that HBXIP might upreg-ulate FGF8 through activating the promoter of FGF8. To test this hypothesis, we constructed the fragments of human FGF8 promoter, including −1454/−317 (pGL3-FGF8-P1), −1294/−317 (pGL3-FGF8-P2), −1080/−317 (pGL3-FGF8-P3) and −904/−317 (pGL3-FGF8-P4) (Figure 1F and Supplementary Figure 1G, available at Carcinogenesis Online). Luciferase reporter gene assays indicated that the regions of

−1454/−317, −1294/−317 and −1080/−317 exhibited similar lucif-erase activities, which were higher than the region of −904/−317 or pGL3-Basic, indicating that the fragment −1080/−904 contains the core region of FGF8 promoter (Figure 1F). To investigate the effect of HBXIP on FGF8 promoter activity, we modulated the expres-sion of HBXIP by transfection with HBXIP plasmid or si-HBXIP to measure the activities of various promoter constructs in breast cancer cells. Our results indicated that HBXIP could significantly increase

Fig. 1. HBXIP upregulates FGF8 in breast cancer cells and increases its secretion. (A) The expression levels of HBXIP and FGF8 were examined by IHC staining assay in 49 cases of breast cancer tissues, which were from the same tissue paraffin block. (B) The relative expression levels of FGF8 were detected by RT–PCR and western blot analysis in MCF-7 transiently transfected with pCMV or HBXIP. (C) The relative secretion levels of FGF8 were detected by ELISA in MCF-7 transiently transfected with pCMV or HBXIP. (D) The relative expression levels of FGF8 were detected by RT–PCR and western blot analysis in MDA-MB-231 cells transfected with si-Control or si-HBXIP. (E) The relative secretion levels of FGF8 were detected by ELISA in MDA-MB-231 cells transfected with si-Control or si-HBXIP. (F) MCF-7 cells were cotransfected with pGL3-FGF8-P1 (or pGL3-FGF8-P2, pGL3-FGF8-P3 or pGL3-FGF8-P4) and HBXIP. The pCMV vector was used as a control. The promoter activities of FGF8 were measured by luciferase reporter gene assays. Statistically significant differences are indicated: *P < 0.05, **P < 0.01, Student’s t-test. We performed separate experiment in triplicate.

1146

The oncoprotein HBXIP enhances angiogenesis and growth of breast cancer

the activities of FGF8 promoter and the regulation sites of HBXIP covered the region of −1080/−904 (Figure  1F and Supplementary Figure 1H, available at Carcinogenesis Online). Thus, we conclude that HBXIP is able to upregulate FGF8 in breast cancer cells and increases its secretion.

HBXIP activates FGF8 promoter through interacting with transcriptional factor CREBTo dissect the mechanism by which HBXIP activates FGF8 promoter, we analyzed the identified core region of −1080/−904 by bioinfor-matics using Genomatix software suite. Interestingly, we observed that the region contained several transcriptional factor-binding sites, such as CREB, c-Myc and MZF1, etc. HBXIP is essential for CREB-dependent transcriptional activity of HBx (32), therefore, we wonder whether HBXIP activates FGF8 promoter through CREB.

As expected, the promoter activity of FGF8 could be abolished by si-CREB in MCF-7-HBXIP and MDA-MB-231 cells (Figure  2A). Moreover, we showed that HBXIP failed to work when the CREB-binding site was mutated in FGF8 promoter (Figure 2B). Next, we are interested in whether FGF8 is cAMP responsive. Our data showed that the level of CREB phosphorylation was increased in the forskolin (an activator of adenylyl cyclase)-treated cells, but the promoter activ-ity of FGF8 exhibited no response to the treatment (Supplementary Figure 2A and B, available at Carcinogenesis Online). In addition, we observed the similar results in the cells treated with 8-Br-cAMP (data not shown), suggesting that FGF8 is not regulated by cAMP. To test whether HBXIP binds to FGF8 promoter via CREB, we exam-ined the interaction of HBXIP with the region of FGF8 promoter by chromatin immunoprecipitation assay. Our data showed that HBXIP and CREB could bind to FGF8 promoter, but both of them failed to

Fig. 2. HBXIP activates FGF8 promoter through interacting with transcriptional factor CREB. (A) MCF-7-HBXIP or MDA-MB-231 cells were transiently cotransfected with pGL3-FGF8-P3 and si-CREB. The si-Control was used as a control. The promoter activities of FGF8 were detected by luciferase reporter gene assays. (B) The wild-type or point-mutant FGF8 promoter was transiently cotransfected with HBXIP into MCF-7 cells. The pCMV vector was used as a control. The promoter activities of FGF8 were measured by luciferase reporter gene assays. (C) The interaction of HBXIP (or CREB) with promoter region of FGF8 in MCF-7-HBXIP cells was examined by chromatin immunoprecipitation (ChIP) assays. (D) The interaction of HBXIP with CREB in MCF-7-HBXIP cells was examined by coimmunoprecipitation (co-IP). (E) The interaction of His-tagged HBXIP with GST-tagged CREB was examined by pull-down assays. (F) The CRE site was predicted using Genomatix software suite. The interaction of HBXIP with CRE element located in the region of FGF8 promoter was examined by EMSA. (G) The interaction of CREB with CRE element located in the region of FGF8 promoter was examined by EMSA. Statistically significant differences are indicated: *P < 0.05, **P < 0.01, Student’s t-test. We performed separate experiment in triplicate.

1147

F.Liu et al.

work when the cells were treated with si-CREB (Figure 2C), suggest-ing that HBXIP may bind to FGF8 promoter through interacting with CREB. Then, coimmunoprecipitation and pull-down assays indicated that HBXIP was able to directly bind to CREB in the cells (Figure 2D and E). EMSA showed that the nuclear proteins were able to interact with 32P-labeled CRE oligonucleotide and the band was enhanced by HBXIP (Figure 2F). In contrast, incubation with an unlabeled oligo-nucleotide decreased the specific binding of nuclear protein to the 32P-labeled CRE oligonucleotide. Moreover, the supershift band was observed when the anti-HBXIP or anti-CREB antibody was added into the system (Figure 2F and G). It suggests that HBXIP is a coac-tivator of CREB in regulation of FGF8 promoter. Thus, our data sup-port the notion that HBXIP is capable of activating FGF8 promoter through CREB in breast cancer cells.

HBXIP upregulates FGF8 through inhibiting miR-503To further identify other mechanisms by which HBXIP upregulates FGF8, we predicted the miRNAs targeting FGF8 by TargetScan (http://www.targetscan.org). Interestingly, we observed that there was a target site of miR-503, an antiangiogenic agent, in the 3′UTR of FGF8. Then, we examined the effect of HBXIP on miR-503 expression in breast cancer cells. As shown in Figure 3A, overex-pression of HBXIP suppressed miR-503 expression in MCF-7 cells, conversely, knockdown of HBXIP resulted in the increase of miR-503 in MCF-7-HBXIP cells in a dose-dependent manner. Moreover, we found that the expression levels of miR-503 were negatively correlated with those of FGF8 in clinical breast cancer tissues by qRT–PCR analysis (r  =  −0.671, P  <  0.01, Pearson’s correlation coefficient, Figure 3B). To ascertain whether FGF8 was one of the targets of miR-503, we cloned the 3′UTR of FGF8 mRNA and its mutant into the pGL3-Control vector, respectively (Figure  3C). Luciferase reporter gene assays elucidated that miR-503 could remarkably decrease the luciferase activity of FGF8 3′UTR in MCF-7-HBXIP cells. However, it failed to work when the miR-503-tar-geting site of FGF8 3′UTR was mutated (Figure 3D). Meanwhile, the anti-miR-503 was capable of increasing the luciferase activity of FGF8 3′UTR in MCF-7 cells, but the enhancement could be attenuated when the target site of miR-503 was mutated (Figure 3E). Subsequently, we observed that miR-503 was able to downregulate FGF8 in MCF-7-HBXIP cells, but anti-miR-503 resulted in upregu-lation of FGF8 in MCF-7 cells (Figure 3F and G). Therefore, our data suggest that HBXIP is able to upregulate FGF8 through inhibit-ing miR-503 in breast cancer cells.

HBXIP-induced FGF8 upregulates VEGF through activating PI3K/Akt/HIF1α signaling in breast cancer cells and increases its secretionIt has been reported that VEGF is one of the best known angiogenic factors and miR-503 is able to target VEGF in hepatoma cells (33). Therefore, we are interested in whether HBXIP upregulates VEGF through suppressing miR-503 in breast cancer cells. We found that the expression levels of miR-503 were inversely correlated with those of VEGF in clinical breast cancer tissues. Then, we demonstrated that miR-503 downregulated VEGF through directly targeting its 3′UTR in breast cancer cells (Supplementary Figure 3A–F, available at Carcinogenesis Online). Accordingly, we examined the expression relationship between HBXIP and VEGF in clinical breast cancer tis-sues. We found that the positive rate of VEGF was 80.0% (32/40) in HBXIP-positive tissues by IHC staining analysis (Figure 4A and Supplementary Table  2). Moreover, we demonstrated that HBXIP was able to upregulate VEGF expression in MCF-7 cells in a dose-dependent manner (Figure  4B). Meanwhile, the secretion levels of VEGF were remarkably increased in the media of HBXIP-transfected MCF-7 cells (Figure 4C). Thus, our data suggest that HBXIP upregu-lates VEGF through downregulating miR-503 in breast cancer cells and increases its secretion.

Given that FGF2 could influence VEGF expression through PI3K/Akt/HIF1α signaling (14), we concerned about whether

HBXIP-induced FGF8 was involved in the upregulation of VEGF in breast cancer. Interestingly, we showed that the expression levels of VEGF mRNA were positively correlated with those of FGF8 in clini-cal breast cancer tissues (r = 0.84, P < 0.001, Pearson’s correlation coefficient, Figure 4D). Moreover, we found that depletion of FGF8 reduced the upregulation of VEGF mediated by HBXIP in MCF-7-HBXIP cells and MCF-7 cells transiently transfected with HBXIP (Figure 4E), suggesting that HBXIP-induced FGF8 is able to upregu-late VEGF in breast cancer cells.

It has been reported that HIF1 is overexpressed in primary breast cancer and FGF8 could activate PI3K/Akt signaling (34,35). Accordingly, we supposed that the PI3K/Akt/HIF1α signaling might be responsible for the upregulation of VEGF mediated by HBXIP-elevated FGF8 in the cells. As expected, our data showed that the upregulation of p-Akt, HIF1α and VEGF mediated by HBXIP could be significantly reduced when the cells were treated with PI3K inhibitor or anti-FGF8 antibody (Figure 4F). In addition, we constructed the VEGF promoter containing a reported HRE site (36). Interestingly, we found that overexpression of HBXIP was able to increase the promoter activity of VEGF. Conversely, the effect could be abrogated when the HRE site was mutated. Moreover, the enhancement of VEGF promoter activity mediated by HBXIP could be reduced by PI3K inhibitor or si-FGF8 (Supplementary Figure 4A and B, available at Carcinogenesis Online). Meanwhile, the HBXIP-enhanced VEGF secretion was abrogated in the condi-tioned media in the system (Figure 4G). Therefore, we conclude that HBXIP-induced FGF8 is able to upregulate VEGF through activat-ing PI3K/Akt/HIF1α signaling in breast cancer cells and increases its secretion.

HBXIP promotes angiogenesis and cell proliferation in breast cancer through FGF8 and VEGF in vitroTo determine the role of HBXIP in angiogenesis in breast cancer through FGF8 and VEGF, we performed matrigel angiogenesis assays using the conditioned media from HBXIP-overexpressing MCF-7 cells. Our data showed that the formation of capillary-like structures was increased when the HUVECs were cultured in 70% (vol/vol) con-ditioned media from HBXIP-transfected MCF-7 cells relative to con-trol (Figure 5A). Meanwhile, the neutralization of FGF8 (or VEGF) activity in the conditioned media from HBXIP-transfected MCF-7 cells by anti-FGF8 (or anti-VEGF) antibody resulted in a significant reduction of formation of capillary-like structures (Figure 5A), sug-gesting that HBXIP promotes tumor angiogenesis through FGF8 and VEGF. Furthermore, we tested whether HBXIP enhanced prolifera-tion of MCF-7 cells through FGF8 or VEGF by colony formation and MTT assays. Interestingly, we found that the enhancement of cell proliferation mediated by HBXIP could be abolished by si-FGF8 or si-VEGF (Figure 5B and C and Supplementary Figure 5A and B, available at Carcinogenesis Online). Therefore, our data suggest that HBXIP promotes angiogenesis and cell proliferation in breast cancer through FGF8 and VEGF in vitro.

HBXIP increases angiogenesis and growth of breast cancer in vivoTo better understand the contribution of HBXIP to tumor angiogene-sis and growth, we performed tumor xenograft analysis. Figure 6A–C and Supplementary Figure  6, available at Carcinogenesis Online, showed that overexpression of HBXIP enhanced the tumor growth in mice transplanted with MCF-7 cells, while the enhanced growth could be attenuated by knockdown of FGF8 or VEGF. The forma-tion of neovessels in tumor tissues was determined by the presence of red blood cells and mouse endothelial cells. As shown in Figure 6D, the hemoglobin content in tumor tissues derived from MCF-7-HBXIP cells was 1-fold higher than that in the tissues derived from control cells. Consistent with the results of the hemoglobin content measure-ment, we found that the tumor tissues derived from MCF-7-HBXIP cells were much more intensively vascularized relative to controls as determined by more intense staining of the marker CD31 (Figure 6E), suggesting that overexpression of HBXIP is intimately associated with

1148

The oncoprotein HBXIP enhances angiogenesis and growth of breast cancer

Fig. 3. HBXIP upregulates FGF8 through inhibiting miR-503. (A) The relative expression levels of miR-503 were detected in MCF-7 cells transfected with pCMV or HBXIP (or in MCF-7-HBXIP cells transfected with si-Control or si-HBXIP). (B) The correlation between miR-503 and FGF8 was detected by qRT–PCR in 20 cases of clinical breast cancer tissues (r = −0.671, P < 0.01, Pearson’s correlation coefficient). (C) MiR-503-binding site was predicted at 84–90 nt of the 3′UTR of FGF8. The miR-503-binding site mutant was shown. (D and E) The relative luciferase activities of wild-type or mutant FGF8 3′UTR (indicated as FGF8 3′UTR or FGF8 3′UTR-MUT) were detected in MCF-7-HBXIP cells transfected with NC or miR-503 (or in MCF-7 cells transfected with anti-miR-NC or anti-miR-503). (F and G) The relative expression levels of FGF8 were detected by qRT–PCR and western blot analysis in MCF-7-HBXIP cells transfected with NC or miR-503 (in MCF-7 cells transfected with anti-miR-NC or anti-miR-503). Statistically significant differences are indicated: *P < 0.05, **P < 0.01, Student’s t-test. We performed separate experiment in triplicate.

1149

F.Liu et al.

the presence of a higher density of vasculature and endothelial cells in the tumor tissues as xenografts. Meanwhile, knockdown of FGF8 or VEGF remarkably reduced the HBXIP-enhanced hemoglobin con-tent and CD31 in tumor tissues (Figure 6D and E). Besides, the Ki67 staining supported that overexpression of HBXIP was able to promote proliferation of MCF-7 cells. However, downregulation of FGF8 (or

VEGF) by si-FGF8 (or si-VEGF) could block the HBXIP-enhanced proliferation of breast cancer cells in tumor tissues, suggesting that HBXIP can promote proliferation of breast cancer cells through FGF8 and VEGF in tumor tissues (Figure 6F). Taken together, our data sug-gest that HBXIP promotes tumor angiogenesis and growth of breast cancer through FGF8 and VEGF in vivo.

Fig. 4. HBXIP-induced FGF8 upregulates VEGF through activating PI3K/Akt/HIF1α signaling in breast cancer cells and increases its secretion. (A) The expression levels of VEGF were examined by IHC staining assay in 49 cases of clinical breast cancer tissues, which were from the same tissue paraffin block. (B) The expression levels of VEGF were detected by RT–PCR and western blot analysis in MCF-7 cells transfected with pCMV or HBXIP. (C) The relative secretion levels of VEGF were detected by ELISA in MCF-7 cells transfected with pCMV or HBXIP. (D) Correlation between VEGF and FGF8 mRNA levels was detected by qRT–PCR in 40 cases of clinical breast cancer tissues (r = 0.84, P < 0.001, Pearson’s correlation coefficient). (E) Depletion of FGF8 with siRNA, the protein levels of VEGF were detected in stably or transiently HBXIP-transfected MCF-7 cells. (F) After transfection with HBXIP, the levels of p-Akt, Akt, HIF1α and VEGF were detected in MCF-7 cells treated with PI3K inhibitor or anti-FGF8 antibody. (G) After transfection with HBXIP, the secretion levels of VEGF were detected in MCF-7 cells treated with PI3K inhibitor or anti-FGF8 antibody. Wortmannin is the inhibitor of PI3K. Statistically significant differences are indicated: *P < 0.05, **P < 0.01, Student’s t-test. We performed separate experiment in triplicate.

1150

The oncoprotein HBXIP enhances angiogenesis and growth of breast cancer

Discussion

Tumor angiogenesis plays a vital role in the development of cancer. FGF and VEGF families are the most potent promoters of angiogen-esis. Numerous reports have demonstrated that dysregulation of FGF–FGF receptor and VEGF–VEGF receptor signaling, whatever it is at the ligand and/or receptor level, is able to result in tumor development

and progression (2,3,13,15,17). Recently, we have reported that the oncoprotein HBXIP significantly promotes proliferation and migra-tion of breast cancer cells through upregulating S100A4, LMO4, Skp2 and Lin28B (23–26). In this study, we are interested in the role of HBXIP in angiogenesis in breast cancer.

It has been reported that the mammalian FGF family comprises 18 ligands, which exert their actions through 4 highly conserved transmembrane tyrosine kinase receptors (3). The high frequency of FGF8 expression is detected in the clinical breast cancer tissues, which increases tumor growth, invasion and angiogenesis (7–9); however, its regulatory mechanism is poorly understood. In this study, we showed that the positive rate of FGF8 expression was 73.5% in clinical breast cancer tissues. Interestingly, we found that the positive rate of FGF8 was 87.5% (35/40) in HBXIP-positive breast cancer tissues. Moreover, our data showed that HBXIP was able to upregulate FGF8 in breast cancer cells. Our previous studies showed that HBXIP acted as a coactivator of transcriptional fac-tor (23–26). Accordingly, we identified that HBXIP was capable of activating FGF8 promoter through directly interacting with tran-scriptional factor CREB, suggesting that HBXIP upregulates FGF8 through acting as a coactivator of CREB. Previous studies showed that CREB was able to activate the promoters of BRCA1 and bcl-2, which were not stimulated by forskolin (37,38). Oct-1 potentiated CREB-dependent cyclin D1 transcriptional activity via a phospho-CREB and CBP-independent mechanism (39). In this study, we found that FGF8 was not regulated by cAMP. That may be due to lack of canonical TATA box near the CREB-binding sites, which is essential for cAMP responsiveness (40).

As we know, microRNAs are small non-coding RNAs that can act as oncogenes or tumor suppressor genes in human cancers. We further investigated the mechanism by which HBXIP upregulated FGF8 at the posttranscriptional level. Interestingly, we observed that the miR-503-binding site existed in 3′UTR of FGF8. A grow-ing number of studies have proposed that miR-503 might be a mas-ter regulator of the cell cycle, metastasis and angiogenesis (41,42). We found that HBXIP was able to suppress miR-503 expression in breast cancer cells. Indeed, we showed that FGF8 was one of the target genes of miR-503. Recent study has reported that hyper-methylation of the miR-503 CpG island regions is associated with downregulation of miR-503 (33). We previously found that overex-pression of HBXIP significantly elevated DNA (cytosine-5-)-meth-yltransferase 1 (DNMT1) levels in MCF-7 cells (23). It suggests that HBXIP may suppress miR-503 expression via epigenetic modulation of its promoter. Moreover, we validated that VEGF was another target gene of miR-503 in breast cancer cells as well. This finding is consistent with the report that miR-503 is able to target VEGF in hepatoma cells (33). Next, we demonstrated that HBXIP upregulated VEGF through suppressing miR-503 in breast cancer cells. Taken together, our finding provides evidence that HBXIP upregulates both FGF8 and VEGF through suppressing miR-503 in breast cancer cells.

Emerging evidence shows that FGF2 induces VEGF expression to promote tumor angiogenesis (14,43). Therefore, we are inter-ested in whether FGF8 regulates VEGF in breast cancer cells. Strikingly, we revealed that HBXIP-induced FGF8 was able to upregulate VEGF through activating PI3K/Akt/HIF1α signaling. It is well established that tumor growth is angiogenesis-dependent and FGFs may play an integral role in resistance to anti-VEGF therapy (2,17). Dual blockade of FGF and VEGF signaling are cur-rently being evaluated in clinical trials. In this study, we found that HBXIP strongly increased both expression and secretion of FGF8 and VEGF, resulting in tumor angiogenesis and growth of breast cancer in vitro and in vivo.

We added more information involving our previous papers and other reports (44–50) in Supplementary Figure  7, available at Carcinogenesis Online. In this model, we show that HBXIP plays an important role in regulation of tumor angiogenesis and growth in a network manner, in which FGF8 may influence Lin28B (or LMO4, S100A4) through PI3K/Akt signaling (23,46,48,49);

Fig. 5. HBXIP promotes angiogenesis and cell proliferation in breast cancer through FGF8 and VEGF in vitro. (A) HUVECs were cultured in the conditioned media from MCF-7 cells transfected with pCMV or HBXIP or in the conditioned media from MCF-7 cells transfected with HBXIP containing the presence of non-related antibody (IgG), anti-FGF8 antibody (FGF8 Ab) or anti-VEGF antibody (VEGF Ab), and then representative examples of the formation of capillary-like structures were shown as a low-power image (magnification, ×10). (B) Representative photomicrographs were shown as a low-power image (magnification, ×10) depicting colony formation in soft agar for MCF-7 transiently transfected with HBXIP or cotransfected with HBXIP and si-FGF8 (or si-VEGF). The pCMV vector or si-Control was used as a control. (C) Quantitative measures of colony formation consisted of the diameter and number of colony foci. Statistically significant differences are indicated: **P < 0.01, Student’s t-test. We performed separate experiment in triplicate.

1151

F.Liu et al.

Lin28B may enhance FGF8 signaling through inhibiting let-7, which targets FGF receptor (48,50); LMO4 may increase VEGF expression via downregulating p53, which reduces HIF1α expres-sion (45,47); S100A4 can upregulate VEGF expression (44). Collectively, HBXIP plays a vital role in the development of breast cancer. Therapeutically, HBXIP may serve as a novel target of tumor angiogenesis.

In summary, the central finding of this work is that HBXIP can remarkably induce tumor angiogenesis and growth through upregu-lating the expression and secretion of FGF8 and VEGF in breast cancer. HBXIP is able to activate FGF8 promoter through directly interacting with CREB. In addition, HBXIP upregulates FGF8 and VEGF through suppressing miR-503. Interestingly, HBXIP-enhanced FGF8 stimulates VEGF expression via activating PI3K/Akt/HIF1α

signaling. Thus, our finding provides new insights into the mechanism of tumor angiogenesis and growth of breast cancer.

Supplementary material

Supplementary Tables 1 and 2 and Figures 1–7 can be found at http://carcin.oxfordjournals.org/

Funding

National Basic Research Program of China (973 Program, 2011CB512113, 2010CB833603); National Natural Science Foundation of China (81071623, 81272217, 31271547, 81130044).

Conflict of Interest Statement: None declared.

Fig. 6. HBXIP increases angiogenesis and growth of breast cancer in vivo. (A) The growth curve of the tumors in nude mice transplanted with MCF-7 cells treated with si-Control or MCF-7-HBXIP cells treated with si-Control (or si-FGF8 or si-VEGF) was shown. (B) The weight of tumors was detected. (C) The protein levels of FGF8 and VEGF in tumor tissues were detected by western blot analysis. (D) The levels of hemoglobin content in tumor tissues were detected by ELISA. (E and F) The expression of CD31 and Ki67 in tumor tissues was detected by IHC assay. Statistically significant differences are indicated: **P < 0.01, Student’s t-test.

1152

The oncoprotein HBXIP enhances angiogenesis and growth of breast cancer

References

1. Carmeliet,P. (2005) Angiogenesis in life, disease and medicine. Nature, 438, 932–936.

2. Weis,S.M. et al. (2011) Tumor angiogenesis: molecular pathways and ther-apeutic targets. Nat. Med., 17, 1359–1370.

3. Turner,N. et al. (2010) Fibroblast growth factor signalling: from develop-ment to cancer. Nat. Rev. Cancer, 10, 116–129.

4. Sonvilla,G. et al. (2008) FGF18 in colorectal tumour cells: autocrine and paracrine effects. Carcinogenesis, 29, 15–24.

5. Wesche,J. et al. (2011) Fibroblast growth factors and their receptors in can-cer. Biochem. J., 437, 199–213.

6. Tanaka,A. et al. (1992) Cloning and characterization of an androgen-induced growth factor essential for the androgen-dependent growth of mouse mam-mary carcinoma cells. Proc. Natl Acad. Sci. USA, 89, 8928–8932.

7. Tanaka,A. et al. (1998) High frequency of fibroblast growth factor (FGF) 8 expression in clinical prostate cancers and breast tissues, immunohisto-chemically demonstrated by a newly established neutralizing monoclonal antibody against FGF 8. Cancer Res., 58, 2053–2056.

8. Marsh,S.K. et al. (1999) Increased expression of fibroblast growth factor 8 in human breast cancer. Oncogene, 18, 1053–1060.

9. Ruohola,J.K. et al. (2001) Enhanced invasion and tumor growth of fibro-blast growth factor 8b-overexpressing MCF-7 human breast cancer cells. Cancer Res., 61, 4229–4237.

10. Tuomela,J. et al. (2010) Fast growth associated with aberrant vasculature and hypoxia in fibroblast growth factor 8b (FGF8b) over-expressing PC-3 prostate tumour xenografts. BMC Cancer, 10, 596.

11. Linderholm,B.K. et  al. (2009) Significantly higher levels of vascular endothelial growth factor (VEGF) and shorter survival times for patients with primary operable triple-negative breast cancer. Ann. Oncol., 20, 1639–1646.

12. Shenoy,S.K. et al. (2012) β-arrestin1 mediates metastatic growth of breast cancer cells by facilitating HIF-1-dependent VEGF expression. Oncogene, 31, 282–292.

13. Presta,M. et  al. (2005) Fibroblast growth factor/fibroblast growth fac-tor receptor system in angiogenesis. Cytokine Growth Factor Rev., 16, 159–178.

14. Shi,Y.H. et al. (2005) In vitro study of HIF-1 activation and VEGF release by bFGF in the T47D breast cancer cell line under normoxic conditions: involve-ment of PI-3K/Akt and MEK1/ERK pathways. J. Pathol., 205, 530–536.

15. Ferrara,N. (2004) Vascular endothelial growth factor: basic science and clinical progress. Endocr. Rev., 25, 581–611.

16. Kano,M.R. et al. (2005) VEGF-A and FGF-2 synergistically promote neo-angiogenesis through enhancement of endogenous PDGF-B-PDGFRbeta signaling. J. Cell Sci., 118(Pt 16), 3759–3768.

17. Lieu,C. et  al. (2011) Beyond VEGF: inhibition of the fibroblast growth factor pathway and antiangiogenesis. Clin. Cancer Res., 17, 6130–6139.

18. Melegari,M. et al. (1998) Cloning and characterization of a novel hepa-titis B virus x binding protein that inhibits viral replication. J. Virol., 72, 1737–1743.

19. Marusawa,H. et al. (2003) HBXIP functions as a cofactor of survivin in apoptosis suppression. EMBO J., 22, 2729–2740.

20. Fujii,R. et al. (2006) HBXIP, cellular target of hepatitis B virus oncopro-tein, is a regulator of centrosome dynamics and cytokinesis. Cancer Res., 66, 9099–9107.

21. Bar-Peled,L. et  al. (2012) Ragulator is a GEF for the rag GTPases that signal amino acid levels to mTORC1. Cell, 150, 1196–1208.

22. Wang,F. et al. (2012) Overexpression of hepatitis B x-interacting protein in HepG2 cells enhances tumor-induced angiogenesis. Mol. Cell. Biochem., 364, 165–171.

23. Liu,S. et al. (2012) The oncoprotein HBXIP uses two pathways to up-reg-ulate S100A4 in promotion of growth and migration of breast cancer cells. J. Biol. Chem., 287, 30228–30239.

24. Yue,L. et  al. (2013) The oncoprotein HBXIP activates transcriptional coregulatory protein LMO4 via Sp1 to promote proliferation of breast can-cer cells. Carcinogenesis, 34, 927–935.

25. Xu,F. et  al. (2013) The oncoprotein HBXIP up-regulates Skp2 via acti-vating transcription factor E2F1 to promote proliferation of breast cancer cells. Cancer Lett., 333, 124–132.

26. Liu,Q. et al. (2013) The oncoprotein HBXIP upregulates Lin28B via acti-vating TF II D to promote proliferation of breast cancer cells. Int. J. Cancer, 133, 1310–1322.

27. Zhang,Y. et  al. (2013) The nuclear import of oncoprotein hepatitis B X-interacting protein depends on interacting with c-Fos and phos-phorylation of both proteins in breast cancer cells. J. Biol. Chem., 288, 18961–18974.

28. Hu,N. et  al. (2011) miR-520b regulates migration of breast cancer cells by targeting hepatitis B X-interacting protein and interleukin-8. J. Biol. Chem., 286, 13714–13722.

29. Kanno,E. et al. (2010) Comprehensive screening for novel rab-binding pro-teins by GST pull-down assay using 60 different mammalian Rabs. Traffic, 11, 491–507.

30. Zhang,Y. et al. (2011) PGE2 promotes angiogenesis through EP4 and PKA Cγ pathway. Blood, 118, 5355–5364.

31. Shao,R. et  al. (2004) Acquired expression of periostin by human breast cancers promotes tumor angiogenesis through up-regulation of vascu-lar endothelial growth factor receptor 2 expression. Mol. Cell. Biol., 24, 3992–4003.

32. Wen,Y. et al. (2008) Interaction of hepatitis B viral oncoprotein with cel-lular target HBXIP dysregulates centrosome dynamics and mitotic spindle formation. J. Biol. Chem., 283, 2793–2803.

33. Zhou,B. et al. (2013) MicroRNA-503 targets FGF2 and VEGFA and inhib-its tumor angiogenesis and growth. Cancer Lett., 333, 159–169.

34. Tarkkonen,K. et al. (2010) Fibroblast growth factor 8 induced downregula-tion of thrombospondin 1 is mediated by the MEK/ERK and PI3K path-ways in breast cancer cells. Growth Factors, 28, 256–267.

35. Kimbro,K.S. et al. (2006) Hypoxia-inducible factor-1 in human breast and prostate cancer. Endocr. Relat. Cancer, 13, 739–749.

36. Liu,Y. et al. (1995) Hypoxia regulates vascular endothelial growth factor gene expression in endothelial cells. Identification of a 5’ enhancer. Circ. Res., 77, 638–643.

37. Atlas,E. et al. (2001) A CREB site in the BRCA1 proximal promoter acts as a constitutive transcriptional element. Oncogene, 20, 7110–7114.

38. Wilson,B.E. et al. (1996) Induction of bcl-2 expression by phosphorylated CREB proteins during B-cell activation and rescue from apoptosis. Mol. Cell. Biol., 16, 5546–5556.

39. Boulon,S. et al. (2002) Oct-1 potentiates CREB-driven cyclin D1 promoter activation via a phospho-CREB- and CREB binding protein-independent mechanism. Mol. Cell. Biol., 22, 7769–7779.

40. Conkright,M.D. et al. (2003) Genome-wide analysis of CREB target genes reveals a core promoter requirement for cAMP responsiveness. Mol. Cell, 11, 1101–1108.

41. Rissland,O.S. et  al. (2011) MicroRNA destabilization enables dynamic regulation of the miR-16 family in response to cell-cycle changes. Mol. Cell, 43, 993–1004.

42. Caporali,A. et  al. (2011) Deregulation of microRNA-503 contributes to diabetes mellitus-induced impairment of endothelial function and repara-tive angiogenesis after limb ischemia. Circulation, 123, 282–291.

43. Tsunoda,S. et  al. (2007) Fibroblast growth factor-2-induced host stroma reaction during initial tumor growth promotes progression of mouse mela-noma via vascular endothelial growth factor A-dependent neovasculariza-tion. Cancer Sci., 98, 541–548.

44. Jia,W. et al. (2013) S100A4 silencing suppresses proliferation, angiogene-sis and invasion of thyroid cancer cells through downregulation of MMP-9 and VEGF. Eur. Rev. Med. Pharmacol. Sci., 17, 1495–1508.

45. Ravi,R. et  al. (2000) Regulation of tumor angiogenesis by p53-induced degradation of hypoxia-inducible factor 1alpha. Genes Dev., 14, 34–44.

46. Montañez-Wiscovich,M.E. et al. (2009) LMO4 is an essential mediator of ErbB2/HER2/Neu-induced breast cancer cell cycle progression. Oncogene, 28, 3608–3618.

47. Zhou,X. et al. (2012) LMO4 inhibits p53-mediated proliferative inhibition of breast cancer cells through interacting p53. Life Sci., 91, 358–363.

48. Chang,T.C. et  al. (2009) Lin-28B transactivation is necessary for Myc-mediated let-7 repression and proliferation. Proc. Natl Acad. Sci. USA, 106, 3384–3389.

49. Tsai,W.B. et  al. (2012) Activation of Ras/PI3K/ERK pathway induces c-Myc stabilization to upregulate argininosuccinate synthetase, lead-ing to arginine deiminase resistance in melanoma cells. Cancer Res., 72, 2622–2633.

50. Barh,D. (2008) Let-7 replacement therapy: applicability in cancer. Cancer Ther., 6, 969–984.

Received July 18, 2013; revised January 6, 2014; accepted January 10, 2014

1153