Opposite functions of HIF-α isoforms in VEGF induction by TGF-β1 under non-hypoxic conditions

Abstract

Transforming growth factor (TGF)-β1 has biphasic functions in prostate tumorigenesis, having a growth-inhibitory effect in the early stages, but in the late stages promoting tumor angiogenesis and metastasis. We demonstrate here that tumor-producing TGF-β1 induces vascular endothelial growth factor (VEGF) in prostate cancer cells, and hypoxia-inducible factor (HIF)-1α and HIF-2α has opposite functions in TGF-β1 regulation of VEGF expression under non-hypoxic conditions. The promoter response of VEGF to TGF-β1 was upregulated by the transfection of HIF-2α or siHIF-1α but downregulated by HIF-1α and siHIF-2α. Both HIF-1α and HIF-2α were induced by TGF-β1 at mRNA and protein levels, however, their nuclear translocation was differentially regulated by TGF-β1, suggesting its association with their opposite effects. VEGF induction by TGF-β1 occurred in a Smad3-dependent manner, and the Smad-binding element 2 (SBE2, −992 to −986) and hypoxia response element (−975 to −968) in the VEGF promoter were required for the promoter response to TGF-β1. Smad3 cooperated with HIF-2α in TGF-β1 activation of VEGF transcription and Smad3 binding to the SBE2 site was greatly impaired by knockdown of HIF-2α expression. Moreover, the VEGF promoter response to TGF-β1 was synergistically elevated by co-transfection of Smad3 and HIF-2α but attenuated by HIF-1α in a dose-dependent manner. Additionally, TGF-β1 was found to increase the stability of VEGF transcript by facilitating the cytoplasmic translocation of a RNA-stabilizing factor HuR. Collectively, our data show that tumor-producing TGF-β1 induces VEGF at the both transcription and post-transcriptional levels through multiple routes including Smad3, HIF-2α and HuR. This study thus suggests that autocrine TGF-β1 production may contribute to tumor angiogenesis via HIF-2α signaling under non-hypoxic conditions, providing a selective growth advantage for prostate tumor cells.

Introduction

Transforming growth factor-β1 (TGF-β1) is a 25-kDa homodimeric protein, which regulates a variety of cellular responses, such as proliferation, differentiation, migration and apoptosis (Massague, 1990; Heldin et al., 1997). TGF-β1 signal is transduced from plasma membrane to the nucleus via serine/threonine kinase receptor type I (TβR-I) and type II (TβR-II). In canonical Smad-dependent pathway, heterodimerization of TβR-II and TβR-I provokes the phosphorylation of Smad2 and Smad3, nuclear translocation of the Smads being complexed with Smad4, and eventual transcriptional regulation of target genes (Shi and Massague, 2003). TGF-β1 signal is also transduced via Smad-independent pathways through activation of mitogen-activated protein kinase (MAPK), nuclear factor-κB (NF-κB) and phosphoinositide-3-kinase (PI3K–Akt; Derynck and Zhang, 2003).

TGF-β1 functions as a growth inhibitor in normal cells of the epithelial lineage, and the TGF-β1 signaling pathway is disrupted in a number of human cancer cells by mutational alterations of TβR-I and TβR-II, or Smads. In the prostate, TGF-β1 inhibits growth of carcinoma and normal epithelial cells, and aberrant function of TβR-I and TβR-II correlates with tumor aggressiveness (Kim et al., 1996; Williams et al., 1996). However, both intracellular and serum TGF-β1 levels are elevated in cancer patients and further increased in patients with metastatic carcinoma (Truong et al., 1993). Several studies have shown that prostate cancer cells are resistant to TGF-β1-induced growth inhibition, and TGF-β1 actually promotes tumor growth, migration and metastasis, indicating that TGF-β1 action is oncogenically conversed in the process of tumorigenesis (Stearns et al., 1999; Zhu and Kyprianou, 2005; Lu et al., 2007; Pu et al., 2009). We reported that TGF-β1 promotes cellular proliferation of prostate carcinomas through the activation of Ras/MAPK signaling and induction of tumor-promoting genes, such as interleukin-6, indicating that prostate cancers have a selective growth advantage by autocrine TGF-β1 production (Park et al., 2000, 2003).

Vascular endothelial growth factor (VEGF) is a 45-kDa heparin-binding homodimeric glycoprotein with a strong proangiogenic activity (Ferrara, 2004). VEGF exists as five alternatively spliced isoforms (VEGF121, VEGF145, VEGF165, VEGF189 and VEGF209), and VEGF121 and VEGF165 are most physiologically relevant in human cells. VEGF transcription is strongly activated by hypoxic conditions, and a variety of transcriptional factors such as AP-1, AP-2 and Sp1 have been known to modulate VEGF expression (Mukhopadhyay et al., 1995; Ferrara, 2004; Loureiro and D’Amore, 2005). Expression levels of VEGF and its receptors are increased in prostate cancer cells compared with normal prostate epithelial cells, and elevated VEGF is implicated in angiogenic and non-angiogenic events, which are associated with tumor progression, including early tumor spread (Jackson et al., 2002; Pallares et al., 2006). TGF-β1 has an important role in angiogenesis, and one of the tumor-promoting functions of TGF-β1 is associated with its ability to induce VEGF (Benckert et al., 2003; Sugano et al., 2003; Wang et al., 2004; Jeon et al., 2007). Recently, it was reported that TGF-β1 activates cell migration and invasion by coordinated induction of VEGF receptor type 3 and matrix metalloproteinase-2/9 in intermediary basal epithelial cells of the prostate, suggesting that TGF-β1 is involved in phenotype changes associated with transitions to a malignant behavior (Goodyear et al., 2009).

Hypoxia is a stress for all mammalian cells and also a hallmark of aggressive solid tumor cells, wherein they acquire the capacity to survive and proliferate at conditions of low oxygen tensions (Brown and Wilson, 2004). To cope with hypoxia, tumor cells use the adaptation mechanisms including stabilization and activation of hypoxia-inducible factors (HIFs), a family of transcription factors that have been identified as an important regulator of the cellular response to hypoxia (Semenza, 1998). HIF-1 and HIF-2 are central components to set forth a range of cellular response to hypoxia and exist as a heterodimer of one oxygen-sensitive α-subunit (HIF-1α and HIF-2α) and one β-subunit. On responding to hypoxia, each of these heterodimeric transcription factors translocates to the nucleus and activates transcription of genes involved in cellular and physiologic adaptation to low oxygen supply. A direct and important role of hypoxia in survival and progression of solid tumors is to stimulate tumor angiogenesis through HIF-induced expression of proangiogenic factors, such as VEGF (Ryan et al., 1998; Blancher et al., 2000). HIF-1α and HIF-2α are not only regulated by hypoxia, and can be induced by non-hypoxic mechanisms involving growth factor-induced or oncogenic activation of signaling pathways (Semenza, 2003; Löfstedt et al., 2007). Several growth factors, including IGF-1 and EGF, have been reported to augment HIF-1α level via induction of de novo protein synthesis or reduction of degradation of the protein (Zhong et al., 2000; Laughner et al., 2001; Fukuda et al., 2003). A previous study showed that hypoxia and TGF-β1 can synergize in the induction of the promoter activity of VEGF (Sanchez-Elsner et al., 2001).

In this study, we demonstrate that TGF-β1 activates VEGF transcription through the functional cooperation of Smad3 and HIF-2α under normoxic condition. Both HIF-1α and HIF-2α are induced by TGF-β1, but they have opposite roles in VEGF expression. TGF-β1 also enhances the stability of VEGF transcript via activation of an mRNA-stabilizing factor HuR. Together, our data show that TGF-β1 activates VEGF expression via multiple routes including Smad3, HIF-2α and HuR signaling, indicating that autocrine TGF-β1 production might contribute to angiogenesis under non-hypoxic conditions and thus provide a selective growth advantage for prostate tumor cells.

Results

VEGF expression is activated by tumor-producing TGF-β1 under non-hypoxic conditions

A majority of malignant prostate tumors produce TGF-β1 and express high level of VEGF in the absence of hypoxic stress. To explore the role for TGF-β1 in VEGF expression under normoxic conditions, we initially tested TGF-β1 effect on VEGF expression in DU145 prostate carcinoma cells. VEGF expression is strongly upregulated by TGF-β1 at both mRNA and protein levels in a time- and dose-dependent manner (Figures 1a and b). An increase of VEGF transcript level was clearly detectable as early as 3 h after treatment. Induction of two VEGF isoforms (VEGF165 and VEGF121) was verified by semiquantitative reverse transcription-PCR (RT–PCR). TGF-β1 induction of VEGF mRNA was markedly but not completely inhibited in the presence of actinomycin D, and the stability of VEGF mRNA was elevated by TGF-β1, indicating that TGF-β1 induces VEGF at both transcriptional and post-transcriptional levels (Figures 1c and d). Consistently, TGF-β1 treatment led to a dose-associated increase in secreted VEGF level in both DU145 and PC3 cells (Figure 1e). TGF-β1 induction of VEGF secretion was strongly suppressed in the presence of cyclohexamide or actinomycin D, suggesting that enhanced VEGF secretion is associated with elevated gene expression (Figure 1f). Next, we investigated whether endogenous TGF-β1 produced by tumor cells stimulates VEGF expression. Both mRNA and secreted protein levels of VEGF were significantly decreased by treatment with TGF-β1-neutralizing antibody in both cancer cells, which produce TGF-β1 and retain intact autocrine TGF-β1 signaling (Figures 1g and h). Collectively, these results demonstrate that tumor-producing TGF-β1 induces VEGF expression under physiological oxygen tensions.

Figure 1
figure1

TGF-β1 induction of VEGF in prostate cancer cells under normoxic conditions. (a) Northern blot, semiquantitative RT–PCR, and immunoblot analyses of VEGF expression in DU145 cells. For mRNA expression assay, p21Waf1 was included as a positive control for TGF-β1 action and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an endogenous expression control. VEGF165 and VEGF121 transcripts were examined by semiquantitative RT–PCR using exon-specific primers. (b) A dose-dependent induction of VEGF expression by TGF-β1. (c) TGF-β1 effect on the stability of VEGF mRNA. DU145 cells were incubated with actinomycin D (Act D; 10 μg/ml) 1 h before the TGF-β1 treatment (24 h). (d) Elevation of half-life of VEGF transcript by TGF-β1. The values indicate the percentages of mRNA signal remaining relative to the steady-state level at the time of Act D addition. Data represent means of triplicate assays (bars, s.d.; *P<0.05). (e) Enzyme-linked immunosorbent assay analysis of TGF-β1 effect on secreted VEGF expression. (f) Effect of cycloheximide (CHX, 10 μg/ml) or Act D (10 μg/ml) on TGF-β1 induction of secreted VEGF (**P<0.01). (g) Role of tumor cell-producing TGF-β1 in VEGF expression. DU145 and PC3 cells were treated with TGF-β1 neutralizing antibody (Tβ1nAb) for 24 h and its effect on VEGF expression was determined using semi-quantitative RT–PCR and immunoblot analysis. (h) Effect of Tβ1nAb on secreted VEGF expression (*P<0.05; **P<0.01).

TGF-β1 induces HIF-1α and HIF-2α expression at both protein and mRNA levels

HIF-1α and HIF-2α have a crucial role in VEGF induction by various growth factors, as well as hypoxic stress. To define the possible implication of HIF-1α and HIF-2α in TGF-β1 induction of VEGF under non-hypoxic conditions, we examined whether HIF-1α and HIF-2α is regulated by TGF-β1. As shown in Figures 2a and b, both HIF-1α and HIF-2α protein levels were markedly increased by TGF-β1 in a dose-associated manner. HIF-2α showed detectable increase at 2 h after treatment, 6 h earlier than HIF-1α induction. Both HIF-1α and HIF-2α are known to be regulated predominantly at the post-translational level, thus, we tested whether TGF-β1 affects the stability of the proteins. In the presence of cyclohexamide, HIF-1α and HIF-2α levels were rapidly decreased in control cells, but this reduction was greatly delayed in TGF-β1-exposed DU145 cells, indicating that TGF-β1 enhances the stability of both HIFα proteins in prostate tumor cells (Figures 2c and d). In addition, mRNA level of HIF-1α and HIF-2α was substantially increased by TGF-β1 in a dose-dependent manner (Figure 2e). Similarly to protein induction, mRNA induction by TGF-β1 was transient, showing a maximum induction at 6 h after treatment (Figure 2f). These results indicate that TGF-β1 induces both HIF-1α and HIF-2α expression in prostate cancer cells under normoxic conditions.

Figure 2
figure2

TGF-β1 induction of HIF-1α and HIF-2α. (a) Western blot analysis of TGF-β1 effect on HIF-1α and HIF-2α expression in DU145 and PC3 cells. (b) A dose-dependent induction of HIF-1α and HIF-2α expression by TGF-β1. (c) TGF-β1 elevation of the protein stability of HIF-1α and HIF-2α. DU145 cells were treated with TGF-β1 (2 ng/ml, 6 h) or exposed to hypoxic conditions (1% O2, 12 h) were incubated for indicated times in the presence of CHX (10 μg/ml). (d) Inhibitory effect of TGF-β1 on protein decay of HIF-1α and HIF-2α. The values indicate the levels of proteins remaining relative to the steady-state level at the time of cycloheximide (CHX; 10 μg/ml) addition. Data represent means of triplicate assays (*P<0.05; **P<0.01). (e) A dose-associated induction of HIF-1α and HIF-2α mRNA by TGF-β1. (f) Time kinetics of TGF-β1 induction of HIF-1α and HIF-2α transcripts. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Opposite roles of HIF-1α and HIF-2α in TGF-β1 induction of VEGF

We next determined the role of HIF-1α and HIF-2α in TGF-β1-mediated VEGF expression. VEGF induction by TGF-β1 was suppressed by short-interfering RNA (siRNA)-mediated knockdown of HIF-2α but rather increased by HIF-1α knockdown (Figures 3a and b). The opposite effects of HIF-1α and HIF-2α were also seen in siRNA dose-dependent fashion, whereas both HIF-1α and HIF-2α showed a stimulatory effect in hypoxia-induced VEGF expression (Figure 3c). In both DU145 and PC3 cells, TGF-β1 led to about six- to eightfolds increase in VEGF165 mRNA level at 9 h following treatment, but this induction was significantly suppressed by siHIF-2α transfection (Figure 3d). Consistently, both intracellular and secreted VEGF protein levels were decreased by siHIF-2α but increased by siHIF-1α (Figures 3e and f), indicating that HIF-1α and HIF-2α have an opposite role in TGF-β1 regulation of VEGF in normoxia. Our immunoblot assay has shown that HIF-2α level is increased earlier than HIF-1α level in response to TGF-β1 (Figure 2a). We thus evaluated whether time kinetics of the nuclear accumulation is associated with the differential role of HIF-1α and HIF-2α in VEGF induction. As shown in Figure 3g, the nuclear HIF-2α level showed a detectable increase after 2 h treatment and the highest level at 4 h, which is coincident with VEGF induction, whereas the nuclear HIF-1α level exhibited no visible increase up to 6 h and augmented after 8 h. The nuclear HIF-2α level showed a slow decrease after 6 h treatment but elevated level was detectable up to 12 h. This indicates that both HIF-1α and HIF-2α expression are induced, but their nuclear translocation is differentially regulated by TGF-β1 under normoxic conditions.

Figure 3
figure3

Opposite roles for HIF-1α and HIF-2α in TGF-β1 induction of VEGF. (a) siRNA-mediated knockdown of HIF-1α and HIF-2α expression. DU145 cells transfected with 50 nM of control siRNA (siControl) or siHIF-1/2α RNAs were treated with TGF-β1 (2 ng/ml, 9 h). (b) Effect of HIF-1 knockdown on TGF-β1 induction of VEGF mRNA expression. C, siControl transfected. (c) Opposite effects of HIF-1α and HIF-2α on TGF-β1-mediated VEGF induction. A dose-associated effect of siHIF-1α and siHIF-2α on VEGF mRNA induction was determined. (d) VEGF mRNA levels in cells transfected with siControl, siHIF-1α or siHIF-2α were analyzed using semiquantitative RT–PCR. Data represent means of triplicate assays (bars, s.d.; *P<0.05; **P<0.01). (e) Differential roles for HIF-1α and HIF-2α in VEGF regulation by TGF-β1 and hypoxia. DU145 cells transfected with 20 nM of siControl, siHIF-1α or siHIF-1α were treated with TGF-β1 (2 ng/ml, 9 h) or exposed to hypoxic conditions (1% O2, 12 h). VEGF protein expression was examined by immunoblot assay. (f) Effect of HIF-1 knockdown on TGF-β1 induction of secreted VEGF expression (*P<0.05). (g) Subcellular distribution of HIF-1α and HIF-2α following TGF-β1 treatment. DU145 cells treated with TGF-β1 (2 ng/ml) for indicated times were fractionated and the cytoplasmic and nuclear levels of HIF-1α and HIF-2α were examined by western blot analysis. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Smad3 signaling is required for TGF-β1 activation of VEGF transcription

To explore signaling pathways implicated in TGF-β1 induction of VEGF, we tested the involvement of the Smad pathway, a crucial mediator of TGF-β1 signaling. In response to TGF-β1, DU145 and PC3 cells showed increased phosphorylation of Smad 2 and Smad3, indicating that the TGF-β1-Smad signaling is intact in these cells (Figure 4a). TGF-β1 induction of VEGF mRNA was significantly inhibited by transient expression of a dominant-negative (DN)-Smad3 or DN-Smad4, whereas no detectable effect was observed by DN-Smad2 expression (Figure 4b). Consistently, secreted VEGF level was decreased by DN-Smad3 or DN-Smad4 in a dose-dependent manner (Figure 4c). Transient expression of a wild-type (WT)-Smad3 or WT-Smad4, but not WT-Smad2, led to a dose-associated induction of VEGF (Figure 4d), indicating that TGF-β1 stimulation of VEGF occurs via the Smad3 pathway. It has been known that Smad-independent pathways has an important role in VEGF induction by TGF-β1 in concerted with Smad-dependent pathway (Benckert et al., 2003). We also reported that TGF-β1 induces a synergistic collaboration of p38, NF-κB, c-Jun N-terminal kinase (JNK) and Ras signaling for interleukin-6 activation in DU145 and PC3 prostate cancer cells (Park et al., 2003). We thus examined the possible involvement of other signaling pathways using various pharmacological inhibitors against MAPK/ERK kinase (MEK) (PD98059), p38 (SB202190), JNK (SP600125), PI3K (Wormanin and LY294002), NF-κB (BAY11-7082), Cox-2 (NS-398) and Src (PP2). TGF-β1 induction of VEGF was slightly inhibited by SP600125 and SB202190, but none of the inhibitors significantly affected VEGF induction at both mRNA and protein levels (Figures 4e and f), demonstrating that Smad3 signaling has a key role in TGF-β1 induction of VEGF.

Figure 4
figure4

A key role of Smad3 in TGF-β1-mediated VEGF induction. (a) Activation of Smad2 and Smad3 in DU145 and PC3 cells by TGF-β1. The cells were exposed to TGF-β1 (2 ng/ml) for the indicated times and phosphorylated forms of Smads (p-Smads) were detected by immnublot assay using p-Smad-specific antibodies. (b) Disruption of TGF-β1 induction of VEGF by transfection of DN-Smad3 or DN-Smad4. DU145 cells transfected with DN-Smads expression or empty vector (pcDNA) were treated with TGF-β1 (2 ng/ml) for 12 h, and VEGF mRNA level was determined using semiquantitative RT–PCR. Data represent means of triplicate assays (bars, s.d.; *P<0.05; **P<0.01). (c) A dose-dependent inhibition of TGF-β1 induction of VEGF by DN-Smad3 and DN-Smad4. Total amount of transfected vector was adjusted to 3 μg using empty vector. (d) A dose-dependent induction of VEGF by WT-Smad3 and WT-Smad4. (e) Effect of signaling inhibitors on TGF-β1 induction of VEGF. DU145 cells were pretreated with the specific inhibitors against MEK1 (PD98059, 10 μM), p83 (SB202190, 10 μM), JNK (SP600125, 10 μM), PI3K (wortmannin, 2 μM; LY294002, 10 μM), NF-κB (BAY11-7082, 20 μM), COX2 (NS-398, 10 μM) and PP2 (Src, 5 μM) before the TGF-β1 treatment. (f) Effect of signaling inhibitors on TGF-β1 induction of VEGF mRNA expression. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Smad3 and HIF-2α cooperate for TGF-β1 activation of VEGF promoter

We performed a promoter assay to evaluate the role for Smad3, HIF-1α and HIF-2α in the regulation of VEGF transcription by TGF-β1 (Figure 5a). The reporter vector pVEGF/2361 containing two Smad-binding elements (SBE1 and SBE2) and one hypoxia-response element (HRE) exhibited a significant increase (25- to 30-folds) of luciferase activity following TGF-β1 treatment comparable with the positive control 3TPLuc, which carries the TGF-β1-responsive elements of the PAI-1 promoter (Figure 5b). Although pVEGF/1074, which has SBE1 deletion, showed a strong induction of luciferase activity, pVEGF/983 carrying deletion of both SBE1 and SBE2 displayed negligible response to TGF-β1 (Figure 5c). As predicted, pVEGF/961, which has deletion of SBEs and HRE, showed no promoter activity. The promoter response to TGF-β1 was markedly impaired when SBE2 or HRE was removed (1074-ΔSBE2 and 1074-ΔHRE, respectively, in Figure 5a), indicating that the presence of both SBE2 and HRE are required for TGF-β1 activation of VEGF promoter. To further define the role of Smads and HIFs, we tested effect of its expression on promoter response to TGF-β1. As expected, the luciferase activity was markedly decreased and increased by DN-Smad3 and WT-Smad3, respectively, whereas DN-Smad2 and WT-Smad2 did not affect promoter response (Figure 5d). Likewise, the promoter activity was strongly increased and decreased by siHIF-2α and WT-HIF-2α, respectively, further supporting that HIF-2α has a stimulatory role in VEGF induction by TGF-β1 (Figure 5e). Consistent with the results from expression assays, TGF-β1 activation of VEGF promoter was further enhanced by HIF-1α knockdown, but decreased by WT-HIF-1α transfection, indicating that HIF-1α and HIF-2α have opposite roles in TGF-β1 activation of VEGF promoter. Collectively, these data demonstrate that SBE2 (−992 to −986) and HRE (−975 to −968) are essential for the Smad3- and HIF-2α-mediated induction of VEGF transcription by TGF-β1.

Figure 5
figure5

TGF-β1 upregulation of the VEGF promoter activity. (a) Smad-binding elements (SBE1 and SBE2) and HRE in the VEGF promoter and construction of reporter plasmids for luciferase assay. +1 indicates the transcription initiation site. (b) The responsiveness of pVEGF/2361 to TGF-β1. DU145 cells transfected with promoter constructs were exposed to TGF-β1 (2 ng/ml) for 12 h. 3TPLux was used as a positive control. Relative luciferase (Luc) activity was normalized by the β-galactosidase activity. Data represent means of triplicate assays (bars, s.d.; *P<0.05; **P<0.01). (c) Disruption of promoter responsiveness to TGF-β1 by deletion of the SBE2 and/or HRE. Promoter response to hypoxia (1% O2) was determined for comparison. (d) Suppression and further activation of the promoter responsiveness to TGF-β1 by expression of DN-Smad3 and WT-Smad3, respectively. (e) Effect of knockdown or overexpression of HIF-1α and HIF-2α on the TGF-β1- or hypoxia-mediated activation of VEGF promoter.

TGF-β1 activation of VEGF occurs through the functional collaboration of Smad3 and HIF-2α

To define the interplay of HIF-1α, HIF-2α and Smad3 on VEGF promoter, we carried out co-transfection assay. Compared with separate transfection, co-transfection of WT-HIF-2α and WT-Smad3 led to a synergistic elevation of promoter activity and this effect was seen in a dose-dependent fashion (Figure 6a). The synergistic effect of Smad3 and HIF-2α was blocked by WT-HIF-1α expression, whereas WT-Smad2 expression exerts only mild inhibitory effect (Figure 6b), suggesting that Smad3 and HIF-2α collaborate on the promoter, and HIF-2α and HIF-1α might compete in an antagonistic mode. To determine whether Smad3 binds to the SBE2 site, and HIF-1α and HIF-2α influence this interaction, we performed an electrophoretic mobility shift analysis using a 25-bp synthetic oligonucleotide (WT-SBE2) containing the SBE2 site. One prominently shifted band was detected when the probe was incubated with nuclear extract purified from TGF-β1-treated cells, and the band was super shifted when Smad3 antibody was added (Figure 6c). The specificity of this binding was confirmed by efficient competition with an excess of unlabeled WT- but not mutant-type competitor. Moreover, the Smad3 binding to SBE2 was disrupted when HIF-2α expression is knocked down, indicating that HIF-2α is required for the Smad3 binding to VEGF promoter (Figure 6d). To elicit whether Smad3 indeed occupies the SBE2 site in living cells, we carried out a chromatin immunoprecipitation assay. A clear association was observed between Smad3 and the SBE2 site, when the cells were treated with TGF-β1, and this interaction was substantially attenuated by HIF-2α knockdown (Figure 6e). Consistent with induction kinetics shown in Figure 1a, the Smad3 binding to the VEGF chromatin was detectable at 3 h, and remained up to 12 h after TGF-β1 treatment. We also found that both HIF-1α and HIF-2α bind to the HRE site in TGF-β1-treated cells (Figure 6f). Interestingly, HIF-2α binding to the HRE site was detected at 3 h and reached to the maximum level at 6 h, whereas HIF-1α binding to the site was detectable only after 9 h treatment, indicating that HIF-2α and HIF-1α binding to the HRE site is highly associated with their nuclear levels and effects on VEGF promoter. Moreover, HIF-2α and HIF-1α interaction with the HRE site was increased by knockdown of HIF-1α and HIF-2α, respectively, suggesting that HIF-1α and HIF-2α might compete for binding to the HRE site (Figure 6g). Taken together, our data show that TGF-β1 induction of VEGF transcription in normoxia is mediated by the functional collaboration of Smad3 and HIF-2α, whereas HIF-1α exerts an antagonizing effect on this cooperation.

Figure 6
figure6

Collaboration of Smad3 and HIF-2α in VEGF promoter activation by TGF-β1. (a) Synergistic action of Smad3 and HIF-2α in activation of VEGF promoter by TGF-β1. Data represent means of triplicate assays (bars, s.d.; *P<0.05; **P<0.01). (b) HIF-1α inhibition of synergistic induction of VEGF promoter by Smad3 and HIF-2α. (c) Electrophoretic mobility shift assay of Smad3 binding to the Smad-binding element 2 (SBE2) site. A synthetic 25-bp oligonucleotide (WT-SBE2) was incubated with nuclear extract purified from TGF-β1-treated DU145 cells. For specificity verification of the probe binding, the probe was incubated with nuclear extract in the presence of an excess of unlabeled WT-SBE2 or mutant-type (MT)-SBE2 competitors. (d) Disruption of Smad3 binding to the SBE2 site in HIF-2α-knockdown cells. (e) Chromatin immunoprecipitation assay for Smad3 binding to the SBE2 site within the VEGF promoter. DU145 cells transfected with siHIF-2α or siControl were exposed to TGF-β1 for 0–12 h. Cross-linked chromatin was immunoprecipitated with antibodies against Smad3 or rabbit immunoglobulin G (IgG) and analyzed by PCR using primers that flank the SBE2 site. (f) Chromatin immunoprecipitation assay for HIF-1α and HIF-2α binding to the HRE site within the VEGF promoter. (g) Competition of HIF-1α and HIF-2α for binding to the HRE site. DU145 cells were transfected with increasing doses of siHIF-2α or siHIF-α, and its influence on TGF-β1-mediated HIF-1α and HIF-2α binding to the HRE site was determined using chromatin immunoprecipitation assay. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

TGF-β1 enhances the mRNA stability of VEGF via HuR activation

Our expression study revealed that the mRNA stability of VEGF is also enhanced by TGF-β1 (Figures 1e and f). It has been known that HuR, a member of the Elav/Hu family, which stabilizes target mRNAs and enhances their translation, binds VEGF mRNA through its AU-rich elements located in the 3′ untranslated region (Nabors et al., 2001). We thus investigated whether HuR is involved in post-transcriptional regulation of VEGF by TGF-β1. As shown in Figure 7a, the induction of VEGF mRNA by TGF-β1 was attenuated by HuR knockdown, suggesting the involvement of HuR in TGF-β1-mediated VEGF induction. Moreover, the inhibitory effect of HuR knockdown was more apparent when higher doses of TGF-β1 were treated, suggesting that HuR activity might be upregulated by TGF-β1 in a dose-dependent manner (Figure 7b). Likewise, TGF-β1 induction of secreted VEGF level was enhanced by WT-HuR transfection but reduced by HuR knockdown (Figure 7c). HuR is predominantly localized in the nucleus, but it translocates to the cytoplasm in response to various stimuli and HuR subcellular localization is intimately linked to its function (Fan and Steitz, 1998; Tran et al., 2003). Thus we examined whether TGF-β1 affects the cytoplasmic shuttling of HuR using the nuclear and cytoplasmic fractionation of the cells. The cytosolic HuR level was clearly elevated by TGF-β1 in a dose-associated manner, whereas the nuclear HuR level was decreased following treatment (Figure 7d). Furthermore, a microscopic assay using transfection of green florescence -HuR expression plasmids showed that number of cells displaying the cytoplasmic HuR was clearly increased by TGF-β1 (Figure 7e). Together, these findings indicate that TGF-β1 also increases VEGF mRNA stability by HuR activation under normoxic conditions.

Figure 7
figure7

Involvement of HuR in TGF-β1 induction of VEGF expression. (a) Attenuation of TGF-β1 induction of VEGF mRNA expression by HuR knockdown. (b) A dose-dependent effect of siHuR on TGF-β1 activation of VEGF expression (bars, s.d.; *P<0.05). (c) Effect of WT-HuR and siHuR transfection on TGF-β1 induction of secreted VEGF expression (bars, s.d.; *P<0.05). (d) TGF-β1 effect on cytoplasmic import of HuR. DU145 cells were treated with increasing doses of TGF-β1 for 12 h. The subcellular distribution of HuR was assessed using fractionation and immunoblot assays. (e) An immunofluoresence assay for TGF-β1 effect on HuR relocalization. DU145 cells were transfected with green florescence protein (GFP)-HuR and treated with TGF-β1 (2 ng/ml) for 0–24 h, and cytoplasmic HuR was counted. Data represent means of triplicate assays (bars, s.d.; **P<0.01). GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Discussion

In this study, we show that HIF-1α and HIF-2α are activated by TGF-β1 in prostate cancer cells, but have opposite roles in VEGF expression under normoxic conditions. TGF-β1 induction of VEGF occurs in a Smad3-dependent manner and HIF-2α is required for the efficient interaction of Smad3 with the SBE2 site within the VEGF promoter. The synergistic cooperation of Smad3 and HIF-2α on the VEGF promoter is suppressed by HIF-1α. Both HIF-1α and HIF-2α are activated but show different time kinetics in nuclear translocation following TGF-β1 treatment. TGF-β1 also increases the mRNA stability of VEGF by the activation of a RNA-stabilizing factor HuR. Collectively, our data demonstrates that tumor-producing TGF-β1 activates VEGF at both transcription and post-transcriptional levels, and HIF-1α and HIF-2α have opposite roles in TGF-β1 regulation of VEGF under physiological oxygen tensions.

TGF-β1 functions as a paracrine growth inhibitor in normal cells of the epithelial lineage by inducing cell cycle arrest and apoptosis, and many human cancers escape from TGF-β1-mediated growth suppression by mutational alterations of its receptors or Smads (Markowitz et al., 1995). Unlikely normal epithelial cells, carcinoma cells produce active TGF-β1, which promotes tumor growth, migration and metastasis (Stearns et al., 1999; Zhu and Kyprianou, 2005; Lu et al., 2007; Pu et al., 2009). In prostate tumorigenesis, TGF-β1 has biphasic functions, having a growth inhibitory effect at early stages, but at later stages enhancing the malignant properties of tumors. We previously showed that TGF-β1 function is mitogenically conversed in prostate carcinoma cells through the activation of the oncogenic Ras–MAPK signaling pathway (Park et al., 2000). TGF-β1 is involved in blood vessel formation and one of its oncogenic functions is to promote tumor angiogenesis, which is associated with its ability to induce VEGF expression (Benckert et al., 2003; Sugano et al., 2003; Jeon et al., 2007). A recent study demonstrated that TGF-β1 activates migration and invasion of intermediary basal epithelial cells of the prostate by induction of VEGF receptor type 3 and matrix metalloproteinase-2/9, indicating that TGF-β1 is involved in cell phenotype changes associated with transitions to a malignant behavior (Goodyear et al., 2009). It was reported that Smad3 signaling activates VEGF transcription synergistically with augmented nuclear HIF-1α in hypoxia (Sanchez-Elsner et al., 2001). However, the molecular mechanism underlying VEGF induction by TGF-β1 under normal oxygen tensions has not been well defined. In this study, we found that Smad3 and HIF-2α have a key role in the transcriptional activation of VEGF by TGF-β1 in prostate carcinoma cells under normoxic conditions. TGF-β1 induction of VEGF was blocked by either DN-Smad3 or siHIF-2α, but not severely affected by inhibitors for MEK (PD98059), p38 (SB202190), JNK (SP600125) or PI3K (LY294002). Smad3 was identified to bind to the SBE2 site on the VEGF promoter, and this interaction was abolished by knockdown of HIF-2α expression. Likewise, VEGF induction was substantially enhanced by co-transfection of Smad3 and HIF-2α compared with transfection of either Smad3 or HIF-2α, indicating that HIF-2α cooperates with Smad3 in VEGF induction. These findings indicate that under physiological oxygen tensions, TGF-β1 stimulates VEGF transcription via HIF-2α rather than HIF-1α. Our data thus support that autocrine TGF-β1 production provides selective growth advantages for prostate tumors via transcriptional activation of tumor-promoting genes, such as VEGF, and also lead to the conjecture that VEGF induction counteracts the growth suppression function of TGF-β1 and thus contributes to the oncogenic conversion of TGF-β1 effect during tumor progression.

In normoxic state, HIF-1α is constitutively hydroxylated on specific proline residues, ubiquitinated by von Hippel Lindau protein, and degraded via proteosomal degradation pathway. Hypoxia impedes hydroxylation of HIF-1α, and eventually increases its protein level. In contrast to protein regulation, mRNA level of HIF-1α is generally not fluctuated by hypoxia or other growth factors except for some cases (Wang et al., 1995). We observed that TGF-β1 upregulates protein expression of both HIF-1α and HIF-2α. In the presence of cyclohexamide, the degradation of HIF-1α and HIF-2α were suppressed by TGF-β1, indicating that TGF-β1 enhances the stability of both HIFα proteins. Consistent with our finding, a recent study demonstrated that TGF-β1 selectively inhibits PHD2 expression, thereby impairs the HIF-1α degradation in normoxia (McMahon et al., 2006). Interestingly, we also observed that mRNA expression of HIF-1α and HIF-2α is strongly induced by TGF-β1 in a dose- and time-dependent manner. A detectable increase of HIF-1α and HIF-2α transcript levels at 3 h after TGF-β1 treatment suggests that TGF-β1 might directly activate the transcription of both genes in normoxia. We are exploring the molecular mechanism underlying TGF-β1 activation of HIF-α gene transcription under normoxic conditions and the possible involvement of other signaling pathways, such as PI3K, NF-κB and MAPK.

Despite of well-established HIF-1α role in tumor angiogenesis, several lines of evidences have supported differential or contrasting roles of HIF-1α and HIF-2α in VEGF expression (Löfstedt et al., 2007; Noguera et al., 2009). It was demonstrated that HIF-1α and HIF-2α are regulated differently in a time- and oxygen-dependent fashion, and that hypoxia-induced VEGF expression is primarily driven by HIF-2α at prolonged hypoxia, whereas HIF-1α functions as a major VEGF inducer during acute hypoxic conditions (Uchida et al, 2004; Holmquist-Mengelbier et al., 2006; Helczynska et al., 2008). HIF-2α expression in apparently well vascularized tumor areas where the cells are well oxygenized also suggest that HIF-2α can be active at physiologic oxygen tensions (Holmquist-Mengelbier et al., 2006; Pietras et al., 2008). In addition, HIF-1α uniquely activates a few glycolytic enzymes, such as CA9 and BNip3, whereas HIF-2α essentially contributes to Oct-4, cyclin D1 and erythropoietin (EPO). Interestingly, in renal cell carcinoma cells, BNip3 was identified to be induced by either HIF-1α overexpression or HIF-2α knockdown, demonstrating a contrasting role of both HIFα (Raval et al., 2005). We found that HIF-1α and HIF-2α are induced by TGF-β1, but they have opposite roles in VEGF expression in prostate cancer cells under normoxic conditions. Both mRNA and secreted protein levels of VEGF were elevated by siHIF-1α but decreased by siHIF-2α in a dose-dependent manner. Likewise, the VEGF promoter response to TGF-β1 was down- and upregulated by ectopic expression of HIF-1α and HIF-2α, respectively. It is therefore, conceivable that HIF-2α might compete with HIF-1α in interplay with Smad3 on the VEGF promoter. This was supported by findings that Smad3 interaction with the SBE site on the VEGF promoter and its effect on the promoter activity are regulated positively and negatively by HIF-2α and HIF-1α, respectively. Our results also suggest that the opposite role of HIF-1α and HIF-2α might be associated with the different time kinetics of their nuclear translocation in response to TGF-β1. An obvious increase in nuclear HIF-2α was observed at 2 h following TGF-β1 treatment, which is coincides with VEGF mRNA induction, whereas HIF-1α translocation was detectable at 6 h after treatment. It is thus plausible that HIF-2α and HIF-1α have a reciprocal role in the Smad3 binding to the VEGF chromatin under normoxic conditions, exerting differential effect on VEGF promoter response to TGF-β1.

It is becoming clear that HIF-2α has an important role in tumor development and progression. In xenograft models, overexpression of mutant HIF-2α, which lost the von Hippel Lindau-binding activity, thus stabilized in nomorxia, was found to significantly promote the tumor growth, whereas the same kind mutant HIF-1α inhibited tumor growth, implying that HIF-2α but not HIF-1α is the critical target of von Hippel Lindau (Kondo et al., 2002; Maranchie et al., 2002). Furthermore, high level of HIF-2α, but not HIF-1α, associates with distant metastasis, an immature stem-cell-like phenotype and poor outcome in some tumor types (Giatromanolaki et al., 2003; Yoshimura et al., 2004; Pietras et al., 2008). HIF-2α is associated with breast-cancer-specific death and distant metastasis, whereas in neuroblastoma, HIF-1α correlates negatively to vascularization and shows an association with low tumor stage and favorable patient prognosis (Holmquist-Mengelbier et al., 2006; Helczynska et al., 2008; Noguera et al., 2009). These findings raise an important issue that if these HIF-2α-positive cells indeed are tumor-initiating or stem cells, the capacity to produce VEGF and attract blood vessel formation under physiologic oxygen tensions might be crucial during an early phase of establishment of solid tumors (Noguera et al., 2009). These observations thus indicate that although expression of both HIF-1α and HIF-2α correlate with VEGF expression, they are not coherently regulated in vivo and associate with differently with patient outcome. In this context, our data suggest that TGF-β1 could enhance the invasive and metastatic potential of tumor cells through HIF-2α signaling, which has a wide range of oncogenic roles in prostate tumor progression. It was reported that HIF-1α level is higher in the intraepithelial neoplasia than the respective normal epithelium, stromal cells and benign prostate hyperplasia (Zhong et al., 2004). However, a study using 149 radical prostatectomy specimens demonstrated that correlation of HIF-2α and VEGF is higher than that of HIF-1α and VEGF (Boddy et al, 2005). These studies raise the possibility that HIF-1α upregulation is an early event in prostate tumorigenesis, but HIF-2α could be a dominant isoform in advanced prostate cancers. Considering that intracellular and serum TGF-β1 levels correlate with tumor malignancy in prostate cancer patients, our study suggests that HIF-2α might take the dominant role in the process of malignant tumor progression.

In this study, it was found that VEGF mRNA level is increased by TGF-β1 in the presence of actinomycin D, indicating that TGF-β1 also upregulated VEGF at the posttranscriptional level. Our study showed that HuR, a member of the Elav/Hu family that binds to target mRNA subsets bearing an AU-rich element, is activated by TGF-β1. HuR undergoes bidirectional nuclear–cytoplasmic shuttling, with mRNA stabilizing activity presumed to occur in the cytoplasm, possibly in association with translational machinery (Fan and Steitz, 1998). A growing body of evidence points to the key role of HuR in regulating various cellular responses, including cell proliferation, differentiation, inflammation, replicative senescence, immune cell activation and oncogenesis (Wang et al., 2001; Benjamin and Moroni, 2007). HuR interacts with AU-rich elements of many angiogenic factors and immunomodulating cytokines, including VEGF, COX-2, c-Myc, interleukin-8, TGF-β and tumor necrosis factor-α (Wang et al., 2000; Dixon et al., 2001). Accordingly, misregulated association of HuR with AU-rich element-containing transcripts could allow for enhanced expression of growth-related genes that can influence the neoplastic and angiogenic potential of cancer cells. HuR expression is aberrantly elevated in many human malignancies, including colon, brain and gastric cancers, and correlates with advancing stages of malignancy, supporting a role for HuR in tumorigenesis (Nabors et al., 2001; Kang et al., 2008). We found that TGF-β1 stimulates the cytoplasmic translocation of HuR, and siRNA-mediated HuR knockdown leads to a detectable decrease in VEGF expression induced by TGF-β1, indicating that TGF-β1 induces VEGF expression at the post-transcriptional level via HuR activation. Recently, we reported that the expression and cytoplasmic HuR import of HuR is activated by the PI3K–Akt–NF-κB signaling in human gastric cancer (Kang et al., 2008). Considering that Akt and NF-κB signaling are activated by TGF-β1 in many tumor cells, it will be valuable to examine to what extent HuR activation contributes to oncogenic effects exerted by autocrine TGF-β1 production in human tumors.

In conclusion, our study reveals that Smad3 and HIF-2α functionally cooperates and function synergistically in TGF-β1 induction of VEGF in prostate cancer cells under normoxic conditions. HIF-1α has an antagonistic function on the collaboration of Smad3 and HIF-2α on the VEGF promoter. TGF-β1 also increases VEGF mRNA stability by activation of a RNA-stabilizing factor HuR. Therefore, our study demonstrates that tumor-producing TGF-β1 activates VEGF at both transcription and post-transcriptional levels in prostate carcinoma cells under physiological oxygen tensions.

Materials and methods

Cell lines and reagents

Human prostate carcinoma cell lines DU145 and PC3 were maintained at 37 °C in RPMI-1640 medium supplemented with 10% fetal bovine serum (GIBCO BRL, Gaithersburg, MD, USA). Porcine TGF-β1 and anti-TGF-β1 neutralizing antibody were obtained from R&D Systems (Minneapolis, MN, USA). Cycloheximide (protein synthesis inhibitor), actinomycin D (RNA synthesis inhibitor), PD98059 (MEK1 inhibitor), SB202190 (p38 MAPK inhibitor), wortmannin and LY294002 (PI3K inhibitors), BAY 11-7082 (IKK-β inhibitor), NS-398 (COX-2 inhibitor), and PP2 (Src inhibitor) were purchased from Calbiochem (San Diego, CA, USA). Inhibitors were added to cultures 1 h before treatment of TGF-β1. Other reagents were obtained from Sigma (St Louis, MO, USA), if not indicated.

Semi-quantitative RT–PCR analysis

Total cellular RNA (1 mg) prepared by single-step method was reverse transcribed to complementary DNA using MoMuLV reverse transcriptase (GIBCO BRL) and random hexamer primers. For quantitative evaluation by RT–PCR, we initially performed the PCR reaction over a range of cycles (20–40 cycles), and 1:4 diluted complementary DNA (12.5 ng/50 ml PCR reaction) undergoing 24–36 cycles was observed to be within the logarithmic phase of amplification with all primers used for VEGF121, VEGF165, p21Waf1, HIF-1α, HIF-2α and an endogenous expression standard gene glyceraldehyde-3-phosphate dehydrogenase. Primer sequences are available upon request. The complementary DNA was subjected to 28–34 cycles of PCR at 95 °C (1 min), 58–60 °C (0.5 min) and 72°C (1 min) in 1.5 mM MgCl2-containing reaction buffer (PCR buffer II, Perkin Elmer, Wellesley, MA, USA). RT–PCR products (10 μl) were resolved on 2% agarose gels. Quantitation of gene expression levels was achieved by densitometric scanning of the ethidium bromide-stained gels. Absolute area integrations of the bands representing each sample were then compared after adjustment for glyceraldehyde-3-phosphate dehydrogenase expression. Integration and analysis were performed using Molecular Analyst software program (Bio-Rad, Hercules, CA, USA).

Northern blotting analysis

Total RNA (5 mg) was separated on formaldehyde–agarose gel. RNA was transfered to nylon membrane using TurboBlotter (Schleicher & Schulell, Keene, NH, USA) and the membrane was baked in the CL-1000 UV crosslinker (UVP, Upland, CA, USA). Dioxigenin-labeled probes for VEGF165, p21Waf1 and glyceraldehyde-3-phosphate dehydrogenase were prepared using PCR. Hybridization was conducted with Dig-labeled probes in the high SDS hybridization solution at 50 °C for 12 h. Detection was performed using DIG chemiluminescent detection kit (Roche, Basel, Switzerland) according to the manufacturer’s protocol. The same blot was stripped and hybridized with Dig-labeled glyceraldehyde-3-phosphate dehydrogenase probe.

Enzyme-linked immunosorbent assay

The levels of VEGF protein in the conditioned media were measured using human VEGF enzyme-linked immunosorbent assay kit (R&D Systems), which is based on sandwich enzyme immunoassay technique. The supernatants of cultures from various conditions and VEGF standard provided by the manufacturer were assayed according to the manufacturer’s instruction. The absorbance was determined using a microplate reader (Bio-Rad).

Immunoblotting analysis

Cells were lysed in a lysis buffer containing 20 mM Tris (pH 7.4), 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM sodium fluoride, 2 mM sodium pyrophosphate, 1 mM sodium orthovanadate, protease inhibitor cocktail and 1 mM phenylmethylsulfonyl fluoride (PMSF). The cell lysate was clarified by centrifugation and 20–50 μg of total protein was supplemented with Laemmli buffer and loaded on an 8% SDS–polyacrylamide gel for electrophoresis. Immunoblotting analyses were performed using antibodies specific for VEGF (Santa Cruz Biotechnology, Santa Cruz, CA, USA), HIF-1/2α (Novus Biologicals, Littleton, CO, USA), pSmads (Millipore Corp, MA, USA), HuR (Santa Cruz Biotechnology) and Tubulin (Santa Cruz Biotechnology). Antibody binding was detected by enhanced chemiluminescence (Santa Cruz Biotechnology) using a secondary antibody conjugated to horseradish peroxidase. For reprobing with other antibodies, the membranes were incubated in a stripping buffer (0.2 M glycine pH 2.2, 0.1% SDS, 1% Tween-20) at 50 °C for 60 min.

siRNA, plasmid constructs and transfection

siRNA duplex against HIF-1α (5′-IndexTermCTGATGACCAGCAACTTGA-3′), HIF-2α (5′-IndexTermCAGCATCTTTGATAGCAGT-3’) and control siRNA duplex which served as negative control were synthesized by Invitrogen (Carlsbad, CA, USA). For transfection, 2 × 105 cells were plated on 60-mm diameter dishes 24 h and incubated with a siRNA–oligofectamine mixture at 37 °C for 4 h. Fresh medium containing 1% fetal bovine serum was added and incubated for 20 h. Expression plasmids encoding DN-Smad2, DN-Smad3 and DN-Smad4 were described previously (Park et al., 2000). Construction of WT-Smad2, WT-Smad3 and WT-Smad4 was described in a previous report (Park et al., 2003). Transfection of expression plasmids was performed using Effectene (Qiagen, Hilden, Germany) according to the instruction of manufacturer. Briefly, 6 × 104 cells were transfected with plasmids–Effectene mixture for 4 h. The transfected cells were incubated with RPMI-1640 medium with 1% serum for 20 h and then TGF-β1 was added for indicated duration. Detectable toxicity and apoptosis by reagent or vectors was rare. Each transfection experiment was carried out in triplicate. The transfection efficiency was monitored using a fluorescence microscopy for green florescence protein or CAT assay (Roche) according to the instruction of manufacturer.

Reporter constructs and luciferase assay

The reporter plasmid 3TP-Lux, which contains TGF-β1-responsive elements of the PAI-1 promoter was kindly provided by Dr Joan Massague (Memorial Sloan-Kettering Cancer Center, USA). The VEGF promoter regions were cloned into the pGL3-basic vector (Promega, Madison, WI, USA). Briefly, the pVEGF/2361 vector, which contains VEGF promoter region (nucleotides −2361 to +298), and other truncated or deletion constructs were amplified using PCR with primers which have linker sequences of KpnI and XhoI at 5′- and 3′-end, respectively, and amplicons were ligated to a pCR 2.1-TOPO vector (Invitrogen), and finally subcloned into pGL3 reporter vector. Cells (6 × 104 cells/well) were transfected with 500 ng of the promoter constructs using Lipofectamine 2000 (Invitrogen). When the reporter vector was co-transfected with expression vectors and pCMV-β-galactosidase vector, the whole amount of DNA in each transfection was adjusted to be same by using the empty expression vector. Luciferase activity was determined by luminometer (Junior LB 9509, Berthold Technologies, Bad Wildbad, Germany) and relative luciferase activity was calculated by normalization using β-galactosidase activity as an indicator of transfection efficiency. The mean and s.e.m. of luciferase acitivity were calculated and displayed as arbitrary unit.

Electrophoretic mobility shift and chromatin immunoprecipitation assay

Synthetic oligonucleotides containing the SBE2 site (5′-IndexTermGCGCCGACCGGAAGTCCCGCCTCTC-3′) were biotin-labeled at its 3′-end. The annealed probes were incubated for 40 min with nuclear extract in a gel shift-binding buffer (Amersham Pharmacia Biotech. Piscataway, NJ, USA). Unlabeled probes were used 100-folds excess as a competitor. The reaction products were resolved using a 6% non-denaturing polyacryamide gel at 100 V for 2.5 h at room temperature and transfered onto nylon membrane. The probe–protein complex was detected using anti-streptavidin antibody-HRP conjugate. For supershift assay, 1 μg of anti-Smad3 antibody was added to the incubation solution. For chromatin immunoprecipitation, cells were incubated in 1% of formaldehyde solution for 20 min. The cells were lysed and the pellet was resuspended in nuclei lysis buffer and sonicated. Immunoprecipitation was performed with antibodies specific for Smad3 (Millipore Corp), HIF-1/2α (Novus Biologicals). PCR was performed using following primer pairs; HL2-5 (5′-IndexTermCGAGGGCCGGAACCCAGTTCGC-3′) and HL-6 (5′-IndexTermGCGGCTCCTCCTCAGCGCGCA-3′).

Statistical analysis

Statistical analysis was conducted using a one-way analysis of variance and statistical values were presented as a mean±s.e.m. Throughout, P<0.05 was regarded as significant. All of the experiments were conducted in duplicate and repeated at least three times.

Abbreviations

TGF-β1:

transforming growth factor-β1

VEGF:

vascular endothelial growth factor

HIF:

hypoxia-inducible factor

NF-κB:

nuclear factor-κB

PI3K:

phosphoinositide-3-kinase

RT–PCR:

reverse transcription–PCR

siRNA:

short-interfering RNA

References

  1. Benjamin D, Moroni C . (2007). mRNA stability and cancer: an emerging link? Expert Opin Biol Ther 7: 1515–1529.

  2. Benckert C, Jonas S, Cramer T, Von Marschall Z, Schäfer G, Peters M et al. (2003). Transforming growth factor β1 stimulates vascular endothelial growth factor gene transcription in human cholangiocellular carcinoma cells. Cancer Res 63: 1083–1092.

  3. Blancher C, Moore JW, Talks KL, Houlbrook S, Harris AL . (2000). Relationship of hypoxia-inducible factor (HIF)-1α and HIF-2α expression to vascular endothelial growth factor induction and hypoxia survival in human breast cancer cell lines. Cancer Res 60: 7106–7113.

  4. Boddy JL, Fox SB, Han C, Campo L, Turley H, Kanga S et al. (2005). The androgen receptor is significantly associated with vascular endothelial growth factor and hypoxia sensing via hypoxia-inducible factors HIF-1α, HIF-2α, and the prolyl hydroxylases in human prostate cancer. Clin Cancer Res 11: 7658–7663.

  5. Brown JM, Wilson WR . (2004). Exploiting tumour hypoxia in cancer treatment. Nat Rev Cancer 4: 437–447.

  6. Derynck R, Zhang YE . (2003). Smad-dependent and Smad-independent pathways in TGF-β family signaling. Nature 425: 577–584.

  7. Dixon DA, Tolley ND, King PH, Nabors LB, McIntyre TM, Zimmerman GA et al. (2001). Altered expression of the mRNA stability factor HuR promotes cyclooxygenase-2 expression in colon cancer cells. J Clin Invest 108: 1657–1665.

  8. Fan XC, Steitz JA . (1998). HNS, a nuclear-cytoplasmic shuttling sequence in HuR. Proc Natl Acad Sci USA 95: 15293–15298.

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

  10. Fukuda R, Kelly B, Semenza GL . (2003). Vascular endothelial growth factor gene expression in colon cancer cells exposed to prostaglandin E2 is mediated by hypoxia-inducible factor 1. Cancer Res 63: 2330–2334.

  11. Jackson MW, Roberts JS, Heckford SE, Ricciardelli C, Stahl J, Choong C et al. (2002). A potential autocrine role for vascular endothelial growth factor in prostate cancer. Cancer Res 62: 854–859.

  12. Jeon SH, Chae BC, Kim HA, Seo GY, Seo DW, Chun GT et al. (2007). Mechanisms underlying TGF-β1-induced expression of VEGF and Flk-1 in mouse macrophages and their implications for angiogenesis. J Leukoc Biol 81: 557–566.

  13. Helczynska K, Larsson AM, Holmquist Mengelbier L, Bridges E, Fredlund E, Borgquist S et al. (2008). Hypoxia-inducible factor-2α correlates to distant recurrence and poor outcome in invasive breast cancer. Cancer Res 68: 9212–9220.

  14. Heldin CH, Miyazono K, ten Dijke P . (1997). TGF-β signalling from cell membrane to nucleus through SMAD proteins. Nature 390: 465–471.

  15. Holmquist-Mengelbier L, Fredlund E, Löfstedt T, Noguera R, Navarro S, Nilsson H et al. (2006). Recruitment of HIF-1α and HIF-2α to common target genes is differentially regulated in neuroblastoma: HIF-2α promotes an aggressive phenotype. Cancer Cell 10: 413–423.

  16. Giatromanolaki A, Sivridis E, Kouskoukis C, Gatter KC, Harris AL, Koukourakis MI . (2003). Hypoxia-inducible factors 1α and 2α are related to vascular endothelial growth factor expression and a poorer prognosis in nodular malignant melanomas of the skin. Melanoma Res 13: 493–501.

  17. Goodyear SM, Kheyfets SB, Garcia FU, Stearns ME . (2009). Role of the VEGFR3/VEGFD receptor axis in TGFβ1 activation of primary prostate cell lines. Prostate 69: 982–990.

  18. Kang MJ, Ryu BK, Lee MG, Han J, Lee JH, Ha TK et al. (2008). NF-κB activates transcription of the RNA-binding factor HuR, via PI3K-AKT signaling, to promote gastric tumorigenesis. Gastroenterology 135: 2030–2042.

  19. Kim IY, Ahn HJ, Zelner DJ, Shaw JW, Lang S, Kato M et al. (1996). Loss of expression of transforming growth factor β type I and type II receptors correlates with tumor grade in human prostate cancer tissues. Clin Cancer Res 2: 1255–1261.

  20. Kondo K, Klco J, Nakamura E, Lechpammer M, Kaelin Jr WG . (2002). Inhibition of HIF is necessary for tumor suppression by the von Hippel-Lindau protein. Cancer Cell 1: 237–246.

  21. Laughner E, Taghavi P, Chiles K, Mahon PC, Semenza GL . (2001). HER2 (neu) signaling increases the rate of hypoxia-inducible factor 1α (HIF-1α) synthesis: novel mechanism for HIF-1-mediated vascular endothelial growth factor expression. Mol Cell Biol 21: 3995–4004.

  22. Löfstedt T, Fredlund E, Holmquist-Mengelbier L, Pietras A, Ovenberger M, Poellinger L et al. (2007). Hypoxia inducible factor-2α in cancer. Cell Cycle 6: 919–926.

  23. Loureiro RM, D'Amore PA . (2005). Transcriptional regulation of vascular endothelial growth factor in cancer. Cytokine Growth Factor Rev 16: 77–89.

  24. Lu S, Lee J, Revelo M, Wang X, Lu S, Dong Z . (2007). Smad3 is overexpressed in advanced human prostate cancer and necessary for progressive growth of prostate cancer cells in nude mice. Clin Cancer Res 13: 5692–5702.

  25. Maranchie JK, Vasselli JR, Riss J, Bonifacino JS, Linehan WM, Klausner RD . (2002). The contribution of VHL substrate binding and HIF1-α to the phenotype of VHL loss in renal cell carcinoma. Cancer Cell 1: 247–255.

  26. Massague J . (1990). The transforming growth factor-β family. Ann Rev Cell Biol 6: 597–641.

  27. McMahon S, Charbonneau M, Grandmont S, Richard DE, Dubois CM . (2006). Transforming growth factor β1 induces hypoxia-inducible factor-1 stabilization through selective inhibition of PHD2 expression. J Biol Chem 281: 24171–24181.

  28. Mukhopadhyay D, Tsiokas L, Zhou XM, Foster D, Brugge JS, Sukhatme VP . (1995). Hypoxic induction of human vascular endothelial growth factor expression through c-Src activation. Nature 375: 577–581.

  29. Nabors LB, Gillespie GY, Harkins L, King PH . (2001). HuR, a RNA stability factor, is expressed in malignant brain tumors and binds to adenine- and uridine-rich elements within the 3′ untranslated regions of cytokine and angiogenic factor mRNAs. Cancer Res 61: 2154–2161.

  30. Noguera R, Fredlund E, Piqueras M, Pietras A, Beckman S, Navarro S et al. (2009). HIF-1α and HIF-2α are differentially regulated in vivo in neuroblastoma: high HIF-1α correlates negatively to advanced clinical stage and tumor vascularization. Clin Cancer Res 15: 7130–7136.

  31. Pallares J, Rojo F, Iriarte J, Morote J, Armadans LI, de Torres I . (2006). Study of microvessel density and the expression of the angiogenic factors VEGF, bFGF and the receptors Flt-1 and FLK-1 in benign, premalignant and malignant prostate tissues. Histol Histopathol 21: 857–865.

  32. Park BJ, Park JI, Byun DS, Park JH, Chi SG . (2000). Mitogenic conversion of transforming growth factor-β1 effect by oncogenic Ha-Ras-induced activation of the mitogen-activated protein kinase signaling pathway in human prostate cancer. Cancer Res 60: 3031–3038.

  33. Park JI, Lee MG, Cho K, Park BJ, Chae KS, Byun DS et al. (2003). Transforming growth factor-β1 activates interleukin-6 expression in prostate cancer cells through the synergistic collaboration of the Smad2, p38-NF-κB, JNK, and Ras signaling pathways. Oncogene 22: 4314–4332.

  34. Pietras A, Gisselsson D, Ora I, Noguera R, Beckman S, Navarro S et al. (2008). High levels of HIF-2α highlight an immature neural crest-like neuroblastoma cell cohort located in a perivascular niche. J Pathol 214: 482–488.

  35. Pu H, Collazo J, Jones E, Gayheart D, Sakamoto S, Vogt A et al. (2009). Dysfunctional transforming growth factor-β receptor II accelerates prostate tumorigenesis in the TRAMP mouse model. Cancer Res 69: 7366–7374.

  36. Raval RR, Lau KW, Tran MG, Sowter HM, Mandriota SJ, Li JL et al. (2005). Contrasting properties of hypoxia-inducible factor 1 (HIF-1) and HIF-2 in von Hippel-Lindau-associated renal cell carcinoma. Mol Cell Biol 25: 5675–5686.

  37. Ryan HE, Lo J, Johnson RS . (1998). HIF-1α is required for solid tumor formation and embryonic vascularization. EMBO J 17: 3005–3015.

  38. Sanchez-Elsner T, Botella LM, Velasco B, Corbi A, Attisano L, Bernabeu C . (2001). Synergistic cooperation between hypoxia and transforming growth factor-β pathways on human vascular endothelial growth factor gene expression. J Biol Chem 276: 38527–38535.

  39. Semenza GL . (1998). Hypoxia-inducible factor 1: master regulator of O2 homeostasis. Curr Opin Genet Dev 8: 588–594.

  40. Semenza GL . (2003). Targeting HIF-1 for cancer therapy. Nat Rev Cancer 3: 721–732.

  41. Shi Y, Massague J . (2003). Mechanisms of TGF-β signaling from cell membrane to the nucleus. Cell 113: 685–700.

  42. Stearns ME, Garcia FU, Fudge K, Rhim J, Wang M . (1999). Role of interleukin 10 and transforming growth factor β1 in the angiogenesis and metastasis of human prostate primary tumor lines from orthotopic implants in severe combined immunodeficiency mice. Clin Cancer Res 5: 711–720.

  43. Sugano Y, Matsuzaki K, Tahashi Y, Furukawa F, Mori S, Yamagata H et al. (2003). Distortion of autocrine transforming growth factor β signal accelerates malignant potential by enhancing cell growth as well as PAI-1 and VEGF production in human hepatocellular carcinoma cells. Oncogene 22: 2309–2321.

  44. Tran H, Maurer F, Nagamine Y . (2003). Stabilization of urokinase and urokinase receptor mRNAs by HuR is linked to its cytoplasmic accumulation induced by activated mitogen-activated protein kinase-activated protein kinase 2. Mol Cell Biol 23: 7177–7188.

  45. Truong LD, Kadmon D, McCune BK, Flanders KC, Scardino PT, Thompson TC . (1993). Association of transforming growth factor-β1 with prostate cancer: an immunohistochemical study. Human Pathology 24: 4–9.

  46. Uchida T, Rossignol F, Matthay MA, Mounier R, Couette S, Clottes E et al. (2004). Prolonged hypoxia differentially regulates hypoxia-inducible factor (HIF)-1α and HIF-2α expression in lung epithelial cells: implication of natural antisense HIF-1α. J Biol Chem 279: 14871–14878.

  47. Wang GL, Jiang BH, Rue EA, Semenza GL . (1995). Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci USA 92: 5510–5551.

  48. Wang L, Kwak JH, Kim SI, He Y, Choi ME . (2004). Transforming growth factor-β1 stimulates vascular endothelial growth factor 164 via mitogen-activated protein kinase 3-p38α and p38δ mitogen-activated protein kinase-dependent pathway in murine mesangial cells. J Biol Chem. 279: 33213–33219.

  49. Wang W, Caldwell MC, Lin S, Furneaux H, Gorospe M . (2000). HuR regulates cyclin A and cyclin B1 mRNA stability during cell proliferation. EMBO J 19: 2340–2350.

  50. Wang W, Yang X, Cristofalo VJ, Holbrook NJ, Gorospe M . (2001). Loss of HuR is linked to reduced expression of proliferative genes during replicative senescence. Mol Cell Biol 21: 5889–5898.

  51. Williams RH, Stapleton AM, Yang G, Truong LD, Rogers E, Timme TL et al. (1996). Reduced levels of transforming growth factor β receptor type II in human prostate cancer: an immunohistochemical study. Clin Cancer Res 2: 635–640.

  52. Yoshimura H, Dhar DK, Kohno H, Kubota H, Fujii T, Ueda S et al. (2004). Prognostic impact of hypoxia-inducible factors 1α and 2α in colorectal cancer patients: correlation with tumor angiogenesis and cyclooxygenase-2 expression. Clin Cancer Res 10: 8554–8560.

  53. Zhong H, Chiles K, Feldser D, Laughner E, Hanrahan C, Georgescu MM et al. (2000). Modulation of hypoxia-inducible factor 1α expression by the epidermal growth factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway in human prostate cancer cells: implications for tumor angiogenesis and therapeutics. Cancer Res 60: 1541–1545.

  54. Zhong H, Semenza GL, Simons JW, De Marzo AM . (2004). Up-regulation of hypoxia-inducible factor 1α is an early event in prostate carcinogenesis. Cancer Detect Prev 28: 88–93.

  55. Zhu B, Kyprianou N . (2005). Transforming growth factor β and prostate cancer. Cancer Treat Res 126: 157–173.

Download references

Acknowledgements

This work was supported in part by grants from National Research Foundation of Korea (0000794 and 0001197), the Korea Research Foundation (2008-314-C00247) and the National Cancer Center (0820070), Republic of Korea.

Author information

Correspondence to S G Chi.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Disclaimer

All of the authors are aware of and agree to the content of the paper and their being listed as authors on the paper. This manuscript does not contain any information conveyed either by personal communication or release of unpublished experimental data.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Chae, K., Kang, M., Lee, J. et al. Opposite functions of HIF-α isoforms in VEGF induction by TGF-β1 under non-hypoxic conditions. Oncogene 30, 1213–1228 (2011). https://doi.org/10.1038/onc.2010.498

Download citation

Keywords

  • TGF-β1
  • VEGF
  • Smad3
  • HIF-2α
  • HuR
  • prostate cancer

Further reading