Overexpression of hypoxia-inducible factors (HIF), HIF-1α and HIF-2α, leads to the up-regulation of genes involved in proliferation, angiogenesis, and glucose metabolism and is associated with tumor progression in several cancers. However, the contribution of HIF-1α versus HIF-2α to vascular endothelial growth factor (VEGF) expression and other HIF-regulated target genes under different conditions is unclear. To address this, we used small interfering RNA (siRNA) techniques to knockdown HIF-1α and/or HIF-2α expression in response to hypoxia, insulin-like growth factor (IGF)-I, or renal carcinoma cells expressing constitutively high basal levels of HIF-1α and/or HIF-2α due to loss of von Hippel-Lindau (VHL) function. We found that HIF-1α primarily regulates transcriptional activation of VEGF in response to hypoxia and IGF-I compared with HIF-2α in MCF-7 cells. We also observed a reciprocal relationship between HIF-1α and HIF-2α expression in hypoxia in these cells: HIF-2α siRNA enhanced HIF-1α–mediated VEGF expression in MCF-7 cells in response to hypoxia, which could be completely blocked by cotransfection with HIF-1α siRNA. In contrast, in renal carcinoma cells that constitutively express HIF-1α and HIF-2α due to loss of VHL function, we found that high basal VEGF, glucose transporter-1, urokinase-type plasminogen activator receptor, and plasminogen activator inhibitor-1 expression was predominantly dependent on HIF-2α. Finally, we showed that a newly identified small-molecule inhibitor of HIF-1, NSC-134754, is also able to significantly decrease HIF-2α protein expression and HIF-2α–regulated VEGF levels in renal carcinoma cells. Our data have important implications for how we target the HIF pathway therapeutically. (Cancer Res 2006; 66(12): 6264-70)
- small-molecule inhibitor
- renal carcinoma cells
Hypoxia-inducible factor (HIF)-1 and HIF-2 are heterodimeric transcription factors composed of HIF-1α and HIF-2α, respectively, which dimerize with a constitutively expressed β subunit and subsequently bind to hypoxia response elements in the promoters of target genes. HIF-α protein expression in cells is regulated by a variety of stimuli, including changes in cellular oxygen concentration, growth factors, oncogenic activation, or loss of tumor suppressor function. Under normal oxygen tension, the α subunits are hydroxylated at conserved prolyl and asparaginyl residues and are targeted for degradation by the von Hippel-Lindau (VHL) ubiquitin E3 ligase complex. In hypoxia, inhibition of hydroxylation results in stabilization of HIF-1α and HIF-2α and leads to transcriptional activation of target genes (see refs. 1, 2 for review).
Growth factors induce HIF-1α and HIF-2α protein synthesis ( 3, 4). This occurs independently of oxygen tension by regulatory mechanisms that involve mitogen-activated protein kinase and phosphatidylinositol 3-kinase/Akt signaling pathways ( 3, 5). In addition, loss or inactivation of the tumor suppressor VHL in renal carcinoma cells results in high levels of basal HIF-1α and HIF-2α and can lead to angiogenic renal cell carcinomas and hemangioblastomas ( 6). An important question in HIF biology is the role that the two HIF-α subunits exert on target gene activation, particularly the extent to which they cooperate, overlap, or have distinct roles. Although HIF-1α and HIF-2α seem to be regulated in a similar fashion, many lines of evidence suggest that there is little redundancy between the two α subunits and that they exhibit distinct roles in hypoxic expression and activation of target genes ( 7, 8). In the cancer setting, attempts to address this question have shown that, at least in the case for hypoxia induction, HIF-1α predominantly regulates transcriptional targets ( 9). In addition to hypoxia, both insulin-like growth factor (IGF) signalling and loss of VHL function are important for tumor progression and contribute to high levels of vascular endothelial growth factor (VEGF) and increased angiogenesis. In this study, we assess the contribution of HIF-1α and HIF-2α to the expression of VEGF and other HIF target genes [glucose transporter-1 (GLUT-1), urokinase-type plasminogen activator receptor (uPAR), and plasminogen activator inhibitor-1 (PAI-1)] in response to hypoxia, IGF-I, or loss of VHL function using small interfering RNA (siRNA) to specifically knockdown the expression levels of each subunit. We show that HIF-1α primarily contributes to VEGF induced by hypoxia and IGF-I, whereas HIF-2α primarily regulates VEGF, GLUT-1, uPAR, and PAI-1 gene expression in renal carcinoma cells that express both HIF-1α and HIF-2α. In addition, we show for the first time that there is a reciprocal relationship between HIF-1α and HIF-2α protein expression in hypoxia in MCF-7 breast carcinoma cells and that loss of HIF-2α in these cells results in a clear increase in HIF-1α-dependent VEGF induced by hypoxia. Finally, we show that a small-molecule inhibitor of HIF-1α expression that we identified recently ( 10) also significantly ablates HIF-2α–mediated VEGF expression in renal carcinoma cells. Taken together, our study supports an effort to identify therapeutics that can target both HIF-1α and HIF-2α. We propose that the therapeutic effects of targeting the HIF pathway will depend on the expression of either or both α subunits, stimulus of HIF activation, and cell type.
Materials and Methods
Cell culture. All cell lines were maintained in DMEM supplemented with 10% FCS (Harlan, Oxford, United Kingdom), 100 IU/mL penicillin, 100 μg/mL streptomycin, and 2 mmol/L glutamine (all were purchased from Life Technologies, Paisley, United Kingdom). The MCF-7 breast carcinoma cell line was obtained from American Type Culture Collection (Manassas, VA). 786-O renal carcinoma cells were gifts from Prof. William G. Kaelin, Jr. (Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA) and have been described previously ( 11). RCC4 renal carcinoma cells were gifts from Prof. Patrick Maxwell (Imperial College, London, United Kingdom; ref. 12).
siRNA duplexes and transfection. Four siRNA duplexes were designed to HIF-1α (Genbank accession no. NM_001530) and evaluated. The most efficient duplex for silencing HIF-1α targeted nucleotides 1,302 to 1,322 (5′-TACGTTGTGAGTGGTATTATT) and was designed using the HiPerformance Design Algorithm (Qiagen, Crawley, United Kingdom). Three siRNA duplexes were designed to HIF-2α/EPAS1 (Genbank accession no. NM_001430). The most efficient duplex for silencing HIF-2α targeted nucleotides 4,740 to 4,760 (5′-CCCGGATAGACTTATTGCCAA) and was obtained from Qiagen. A nonsilencing control (5′-AATTCTCCGAACGTGTCACGT) was obtained from Qiagen and used in all experiments. In addition, a negative control duplex (5′-AATTCTCCGAACGTGTCACGT) labeled with Alexa Fluor 488 (Qiagen) was used to monitor transfection efficiency.
Cells were transfected with siRNA duplexes using HiPerfect transfection reagent (Qiagen) according to the manufacturer's instructions. All experiments were done in duplicate. Briefly, 24 hours before the transfection, cells were seeded onto six-well culture dishes at a confluence of 50%. Cells were washed twice with PBS, and duplexes were added to wells at 10 nmol/L in DMEM supplemented with 10% fetal bovine serum for 24 hours at 37°C and 5% CO2. Thereafter, cells were exposed to hypoxia (1% O2, 5% CO2, and 94% N2) in a London Environmental Economics Centre dual-gas incubator (GA156) at 37°C. For IGF-I experiments, duplexes were added in low-serum-containing medium (DMEM and 0.1% FCS) for 36 hours before stimulation with 100 ng/mL IGF-I (Sigma, Poole, United Kingdom). Following treatment, conditioned medium was removed and analyzed for VEGF protein levels by ELISA (QuantiGlo, R&D Systems, Minneapolis, MN), and calibration curves were done for all experiments. In parallel, whole-cell lysates were prepared as described previously and analyzed by Western blotting ( 13). The mouse anti-HIF-1α and mouse anti-VHL antibodies were purchased from BD Biosciences (Oxford, United Kingdom). The rabbit anti-HIF-2α antibody (ab199) was purchased from Abcam (Cambridge, United Kingdom), and the mouse anti-α-tubulin antibody was purchased from Sigma.
Real-time quantitative PCR assays. Total RNA was extracted from cells using the RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. Total RNA (1 μg) was used for first-strand cDNA synthesis using the SuperScript First-Strand Synthesis System for reverse transcription-PCR (Invitrogen, Paisley, United Kingdom) and random hexamers according to the manufacturer's instructions. Real-time quantitative PCR was done using DyNAmo SYBR Green Quantitative PCR kit [Finnzymes, Genetics Research Institute (GRI) Ltd., Braintree, United Kingdom] with a DNA Engine OPTICON 2 Continuous Fluorescence Detection System (GRI) for HIF-1α, HIF-2α, VEGF, GLUT-1, uPAR, PAI-1, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using primers all purchased from Invitrogen ( Table 1 ). Expression of target gene mRNA relative to reference gene mRNA (GAPDH) was calculated using the Relative Expression Software Tool, and statistical significance was determined using the Pair-Wise Fixed Reallocation Randomization Test ( 14, 15). 1
Densitometric analysis. Autoradiographs were scanned using a Fluorchem 8900 system, and band intensities were analyzed with the AlphaEaseFC software. For quantification, values were normalized against the corresponding loading control.
HIF-1α primarily contributes to hypoxia- and IGF-I–induced VEGF expression in MCF-7 cells. Recent evidence has shown the predominance of HIF-1α compared with HIF-2α in regulating target genes in response to hypoxia in human breast cancer cell lines ( 9). Using siRNA techniques to HIF-1α and/or HIF-2α, we also found that VEGF expression was primarily dependent on HIF-1α following stimulation by hypoxia in MCF-7 cells ( Fig. 1 ). A recent study has shown that there is a reciprocal relationship between the expression of HIF-1α and HIF-2α protein in renal carcinoma cells in normoxia upon loss of VHL function ( 16). Therefore, we were interested to assess whether this also occurred in response to hypoxia. Indeed, we observed a reciprocal relationship between HIF-1α and HIF-2α protein levels in MCF-7 cells in hypoxia, such that HIF-1α siRNA increased hypoxia-induced HIF-2α protein levels compared with the nonsilencing control and vice versa ( Fig. 1A and B). Consequently, HIF-2α siRNA significantly enhanced VEGF expression in response to hypoxia compared with nonsilencing control ( Fig. 1C). This could be completely blocked by cotransfection with HIF-1α siRNA, suggesting that this effect was mediated by the enhanced HIF-1α protein ( Fig. 1).
The IGF pathway plays an important role in tumor progression and also significantly up-regulates VEGF expression ( 3, 17), but the dependence of VEGF on HIF activation is unclear. In MCF-7 cells, we observed a 35-fold increase in VEGF protein levels compared with untreated cells after 24 hours of stimulation with IGF-I ( Fig. 2A ). This corresponded to an increase in HIF-1α and HIF-2α protein levels, although HIF-2α protein was induced to a much lesser extent ( Fig. 2A). HIF-1α siRNA blocked IGF-I–induced VEGF expression by ∼50% in MCF-7 cells ( Fig. 2B), whereas HIF-2α siRNA had no significant effect ( Fig. 2B). The inhibitory effects of both HIF-1α and HIF-2α siRNA on IGF-I–mediated VEGF expression were similar to those obtained for HIF-1α siRNA alone, further indicating that, on IGF-I stimulation, HIF-2α did not contribute to VEGF expression in these cells. Interestingly, the enhanced VEGF expression that we observed following HIF-2α siRNA expression in hypoxia ( Fig. 1) was not observed following IGF-I stimulation ( Fig. 2B). In summary, our data suggest that HIF-1α primarily regulates VEGF expression in response to hypoxia and IGF-I in MCF-7 cells and that other HIF-independent pathways contribute, in part, to IGF-I–mediated VEGF expression.
HIF-2α primarily contributes to constitutive VEGF expression in renal carcinoma cells. Loss of VHL function leads to stabilization of HIF-1α and HIF-2α and correspondingly high levels of VEGF in renal carcinoma cells ( 12). To assess the contribution of HIF-1α and HIF-2α to constitutive VEGF expression, we used the renal carcinoma cell lines, RCC4 and 786-O, which have been described previously ( 7, 11, 12, 16). We found that loss of VHL function in both RCC4 and 786-O cells led to high basal VEGF expression ( Fig. 3A ). Interestingly, we observed that the 786-O cells had consistently higher basal levels of VEGF ( Fig. 3A) compared with RCC4 cells, although the 786-O cells express only the HIF-2α subunit compared with RCC4 cells that express high levels of both HIF-1α and HIF-2α ( Fig. 3A). This led us to hypothesize that HIF-2α may be the predominant HIF-α regulating VEGF expression upon loss of VHL function. To test this directly, we analyzed the effects of transient expression of HIF-1α or HIF-2α siRNA in RCC4 cells. We found that HIF-1α siRNA had no significant effect on VEGF protein expression in RCC4 cells but was significantly inhibited by HIF-2α siRNA ( Fig. 3B).
HIF-2α primarily contributes to HIF target gene expression in renal carcinoma cells. Previous studies have shown that HIF-2α regulates the expression of a broad spectrum of target genes ( 7, 16). To determine whether other HIF target genes were regulated by HIF-2α in renal carcinoma cells, real-time quantitative PCR was done. We found that, in addition to VEGF mRNA, the expression of GLUT-1, uPAR, and PAI-1 mRNAs was significantly inhibited by HIF-2α siRNA in the RCC4 cells ( Fig. 4 ), whereas HIF-1α siRNA was less effective for inhibiting GLUT-1 mRNA expression and had no significant effect on VEGF, uPAR, or PAI-1 mRNA expression ( Fig. 4). Our data show that HIF-2α primarily contributes to constitutive HIF target gene expression in renal carcinoma cells that have lost VHL function.
Because we have shown that HIF-1α primarily contributes to VEGF expression in response to both hypoxia ( Fig. 1C) and IGF-I ( Fig. 2B) in MCF-7 cells, we next wanted to ascertain whether HIF-1α could contribute to a VEGF response in the RCC4 cells when exposed to hypoxia or IGF-I. Interestingly, we found that constitutively expressed HIF-1α and HIF-2α proteins were further induced in VHL-defective RCC4 cells in response to IGF-I ( Fig. 5 ) or hypoxia (data not shown), suggesting that VHL-independent mechanisms regulating the expression of HIF-α proteins are functional in these cells. However, in contrast to our observations in MCF-7 cells ( Fig. 2B), we found that HIF-2α, but not HIF-1α, contributed to VEGF expression induced by IGF-I in the RCC4 cells ( Fig. 5). Our data indicate that both stimulus of HIF activation and cell type can govern the relative contribution of HIF-1α versus HIF-2α to the HIF response.
NSC-134754 blocks HIF-2α and VEGF expression in renal carcinoma cells. Targeting the HIF pathway as a strategy for the development of new cancer, therapies has become of intense interest in recent years ( 1, 2, 18– 20). Small-molecule inhibitors ( 21, 22) or siRNA approaches ( 23) that block HIF-1α expression have been shown to inhibit tumor growth in vivo (see ref. 1 for review). Our study implies that the therapeutic effects of agents that target the HIF pathway will depend on their ability to block the expression of HIF-1α and/or HIF-2α in response to multiple stimuli or when constitutively expressed due to genetic abnormalities, such as loss of VHL function.
We have recently identified a small-molecule inhibitor, NSC-134754, which blocks HIF-1α induction in response to both hypoxia and growth factors ( 10). Because our observations suggest that targeting HIF-2α would be important for reducing VEGF expression in renal carcinoma cells, we next addressed whether NCS-134754 could also block constitutive HIF-2α and VEGF expression in these cells. Indeed, we found that NCS-134754 could significantly block HIF-1α and HIF-2α in VHL-defective RCC4 cells and HIF-2α in VHL-defective 786-O cells ( Fig. 6 ). In addition, NCS-134754 was also able to dose-dependently inhibit VEGF levels regulated by HIF-2α in VHL-defective 786-O cells ( Fig. 7 ). Our data suggest that small-molecule inhibitors that block both HIF-1α and HIF-2α expression in response to multiple stimuli or due to loss of VHL function will prove to be useful therapeutic tools.
In the cancer setting, both HIF-1α and HIF-2α protein levels are sensitive to changes in cellular oxygen, growth factors, or loss of VHL function. Generally, overexpression of both HIF-1α and HIF-2α has been found in many human cancers and is thought to correlate with a more aggressive phenotype (see refs. 1, 2 for review). Although the HIF-α subunits are structurally alike and regulated in a similar fashion, it is becoming increasingly clear that they exhibit distinct roles in both embryonic development ( 24, 25) and hypoxic regulation and activation of target genes ( 7, 9, 26). However, it is unclear which α subunit predominantly drives the transcriptional outcome in response to a specific stimulus and whether this is also determined by the genetic context of the cell. This study attempts to address this question. Here, we show contrasting effects of HIF-1α and HIF-2α in the regulation of HIF target gene expression in response to hypoxia, IGF-I, or in renal carcinoma cells that have lost VHL protein function. We found that HIF-1α, but not HIF-2α, primarily regulates VEGF expression in response to hypoxia and IGF-I in MCF-7 cells. In addition, we have also shown for the first time that targeted knockdown of HIF-2α by siRNA enhances HIF-1α protein levels and VEGF expression in response to hypoxia in MCF-7 cells. This response was completely reversed by cotransfection with HIF-1α siRNA.
In hemangioblastomas arising from loss of VHL, HIF-2α levels have been shown to strongly correlate with VEGF expression ( 27), and Sowter et al. have shown that, in 786-O renal cells that express only HIF-2α, VEGF is regulated by this subunit ( 9). In addition, previous studies have used gene expression microarray analysis to assess gene induction by HIF-2α in response to hypoxia in 786-O WT-8 (VHL expressing) renal carcinoma cells ( 7). These studies reported several novel HIF-2α-dependent hypoxically regulated genes, including adipose differentiation-related protein and chemokine growth-regulated oncogenes 2. Furthermore, our work identifies uPAR ( 28) and PAI-1 ( 29) as novel HIF-2α–regulated genes ( Fig. 4B).
Although many previous studies have focused their efforts on unraveling the role of HIF-1α in tumor progression and angiogenesis, it is becoming clear that the role of HIF-2α and the relationship between these subunits is also very important to understand. Consistent with a recent report ( 16), we show here that HIF-2α, but not HIF-1α, regulates constitutive VEGF expression in RCC4 cells that have lost VHL function. We also show that the expression of other HIF targets, including GLUT-1, uPAR, and PAI-1, is primarily driven by HIF-2α in renal cells carcinoma cells and that, even in response to hypoxia or IGF-I, HIF-2α predominantly regulates the transcriptional response in VHL-defective renal carcinoma cells. Other studies have shown that, in some cell lines, hypoxia and IGF-I induction of VEGF is mediated by HIF-2α ( 4, 30). In addition, increased tumor growth, microvessel density, and VEGF expression have been observed in mouse xenograft studies using embryonic stem cells that have been generated to express HIF-2α in replacement of HIF-1α ( 31).
Taken together, these data emphasize that the role of HIF-2α in addition to HIF-1α needs to be considered when developing agents that target the HIF pathway. Our observations highlight the possibility that cancer therapies, which specifically block HIF-2α expression, may indirectly increase HIF-1α–mediated transcriptional activity in hypoxia. Interestingly, we found that inhibiting the expression of both HIF-1α and HIF-2α could block this effect. Therefore, we would predict that targeting both HIF-1α and HIF-2α in cells that show this reciprocal response in hypoxia would be desirable for blocking VEGF expression.
In fact, recently, Raval et al. showed a reciprocal relationship between constitutively expressed HIF-1α and HIF-2α in renal carcinoma cells ( 16). Furthermore, overexpression of HIF-2α in 786-O cells could accelerate tumor growth, whereas overexpression of HIF-1α could reduce tumor growth compared with control 786-O cells presumably either by inhibiting HIF-2α–regulated genes and/or potentially increasing HIF-1α–mediated apoptotic gene expression ( 16), although the latter was not specifically addressed. Our observations suggest that targeted inhibition of constitutive HIF-2α expression in renal carcinoma cells significantly reduces VEGF expression. We would hypothesize therefore that a small-molecule inhibitor that could target HIF-2α would also do the same. Indeed, we show here that NSC-134754, which we recently identified to be a HIF-1α inhibitor ( 10), is also able to significantly inhibit HIF-2α expression and VEGF expression in renal carcinoma cells.
Targeting the HIF pathway has become an attractive approach in recent years for the generation and development of new anticancer agents. It is well established that overexpression of HIF-α occurs in many human cancers ( 2) and that targeting HIF-1α expression in tumor cells inhibits tumor growth in mouse xenograft studies ( 21, 23, 32, 33). However, recent studies have revealed that the HIF-α subunits also have a potential tumor suppressor role that may be context dependent ( 16, 34) and that inhibiting tumor growth by targeting HIF-α expression may not necessarily be mediated by a reduction in VEGF expression or tumor angiogenesis ( 35). Therefore, it will be important to assess whether inhibition of HIF-1α and HIF-2α either alone or in combination produces the best therapeutic outcome in terms of inhibiting tumor growth and angiogenesis in vivo.
Grant support: Institute of Cancer Research, Cancer Research UK program grant C309/A2187 and Association for International Cancer Research grant 03-150 (V.A. Carroll).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Prof. Paul Workman (The Institute of Cancer Research, Sutton, United Kingdom) for critical review of the article, Kati Rasanen for technical assistance, Profs. Ian Stratford (University of Manchester, Manchester, United Kingdom) and Caroline Dive (Paterson Institute for Cancer Research, Manchester, United Kingdom) and Dr. Kaye Williams (University of Manchester) for helpful discussions, Prof. Adrian Harris (Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, United Kingdom) for assistance with anti-HIF-2α antibodies, and Profs. Patrick Maxwell and William G. Kaelin, Jr., for the RCC4 and 786-O renal carcinoma cell lines, respectively.
- Received July 19, 2005.
- Revision received March 10, 2006.
- Accepted March 22, 2006.
- ©2006 American Association for Cancer Research.