Original Article

Oncogene (2011) 30, 2570–2580; doi:10.1038/onc.2010.623; published online 24 January 2011

Regulation of transcription of hypoxia-inducible factor-1α (HIF-1α) by heat shock factors HSF2 and HSF4

R Chen1, J E Liliental1, P E Kowalski1,3, Q Lu1,4 and S N Cohen1,2

  1. 1Department of Genetics, Stanford University School of Medicine, Stanford, CA, USA
  2. 2Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA

Correspondence: Dr SN Cohen, Department of Genetics, Stanford University, 300 Pasteur Drive, Stanford, CA 94305, USA. E-mail sncohen@stanford.edu

3Current address: Developmental and Stem Cell Biology Program, The Hospital for Sick Children, Toronto, Ontario, Canada M5G1L7.

4Current address: Department of Environmental Health, Harvard School of Public Health, Boston, MA 02115, USA.

Received 30 July 2010; Revised 2 December 2010; Accepted 10 December 2010; Published online 24 January 2011.

Top

Abstract

Hypoxia-inducible factor-1α (HIF-1α) is a principal regulator of angiogenesis and other cellular responses to hypoxic stress in both normal and tumor cells. To identify novel mechanisms that regulate expression of HIF-1α, we designed a genome-wide screen for expressed sequence tags (ESTs) that when transcribed in the antisense direction increase production of the HIF-1α target, vascular endothelial growth factor (VEGF), in human breast cancer cells. We discovered that heat shock factor (HSF) proteins 2 and 4—which previously have been implicated in the control of multiple genes that modulate cell growth and differentiation and protect against effects of environmental and cellular stresses—function together to maintain a steady state level of HIF-1α transcription and VEGF production in these cells. We show both HSFs bind to discontinuous heat shock element (HSE) sequences we identified in the HIF-1α promoter region and that downregulation of either HSF activates transcription of HIF-1α. We further demonstrate that HSF2 and HSF4 displace each other from HSF/HSE complexes in the HIF-1α promoter so that HIF-1α transcription is also activated by overexpression of either HSFs. These results argue that HSF2 and HSF4 regulate transcription of HIF-1α and that a critical balance between these HSF is required to maintain HIF-α expression in a repressed state. Our findings reveal a previously unsuspected role for HSFs in control of VEGF and other genes activated by canonical HIF-1α-mediated signaling.

Keywords:

heat shock factor; heat shock element; angiogenesis; HSE; VEGF; HIF-1α

Top

Introduction

Angiogenesis—the process of formation of new blood vessels—is regulated in both normal tissues and tumors largely by an oxygen-sensitive signaling pathway that involves hypoxia-inducible factor-1 (HIF-1) and vascular endothelial growth factor (VEGF) (Mazure et al., 1996, 1997; Semenza, 2003; Ng et al., 2006; Pouyssegur et al., 2006; Brahimi-Horn and Pouyssegur 2007; Fraisl et al., 2009). Under normoxic conditions, the steady state level of the α subunit of HIF-1 (HIF-1α) is tightly controlled by prolyl hydroxylation. During hypoxia, hydroxylation and consequent degradation of the HIF-1α protein is inhibited and HIF-1α accumulates—interacting with the β subunit of HIF-1 and binding to specific DNA sequences (hypoxia-response elements; HREs) in the transcription-regulating region of genes that promote cell growth—and activating these genes. Among the genes activated by HIF-1α is VEGF, which together with other HIF-1-regulated genes encoding fibroblast growth factors, glucose transporter 1 and erythropoietin, facilitates the survival and proliferation of O2-deprived cells (Semenza, 2003; Pouyssegur et al., 2006; Brahimi-Horn and Pouyssegur, 2007; Fraisl et al., 2009). Upregulation of HIF-1α expression or blockade of the degradation can also occur under normoxic conditions in cancer cells, contributing to tumor progression (Jiang et al., 1997; Isaacs et al., 2002; Chi and Karliner, 2004; Busca et al., 2005). Although the regulation of HIF-1α expression at the translational and post-translational levels has been studied extensively (Semenza, 2003; Pouyssegur et al., 2006; Brahimi-Horn and Pouyssegur, 2007; Fraisl et al., 2009), relatively little is known about mechanisms that affect HIF-1α transcription (Jiang et al., 1997; Page et al., 2002; Busca et al., 2005; Pipinikas et al., 2008; Nardinocchi et al., 2009).

Our laboratory has developed a function-based strategy that uses antisense RNAs complementary to a collection of lentiviral-based ESTs to randomly inactivate chromosomal genes (Lu et al., 2004). Using this approach, together with a fluorescence-activated cell sorting (FACS) screen that enables the isolation of cell clones showing elevated VEGF expression (Liliental and Cohen, in preparation), we discovered the previously unsuspected role of heat shock transcription factors (heat shock factors; HSFs) in the regulation of canonical HIF-1α/VEGF signaling. We show that heat shock factors HSF2 and HSF4, which were known previously for their ability to modulate cell growth and differentiation as well as the cellular response to stress (Pirkkala et al., 2001; Chi and Karliner, 2004; Chang et al., 2006; Akerfelt et al., 2007; Morimoto, 2008; Zhang et al., 2008), function together to maintain the steady state level of HIF-1α mRNA by binding to heat shock element (HSE) sites in the promoter region of the HIF-1α gene and regulating HIF-1α transcription.

Top

Results

Upregulation of VEGF expression by HSF4 deficiency

A library of MCF-7G cells infected with a lentiviral-based collection of 40000 sequenced ESTs under control of a tetracycline-repressed promoter (Lu et al., 2004) was subjected to FACS to isolate clones in which expression of a VEGF gene fusion with a GFP reporter gene protein was increased (Liliental and Cohen, in preparation). Among these was clone C16, in which cells showed fourfold greater GFP fluorescence than the parental cell line MCF-7G, as determined by FACS analysis (Figure 1a). DNA sequencing revealed that clone C16 contains a chromosomally inserted EST (GenBank image clone 3395089) corresponding to part of the protein-coding region of heat shock factor 4 (HSF4), a transcriptional regulator previously implicated in expression of heat shock proteins (HSPs)—chaperones that assist protein folding and protect cells against proteotoxic stress (Nakai et al., 1997; Pirkkala et al., 2001; Chi and Karliner, 2004; Min et al., 2004; Fujimoto et al., 2004, 2008; Akerfelt et al., 2007). This HSF4 EST was oriented in the antisense direction relative to the lentiviral promoter.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Effect of HSF4 knockdown on VEGF expression. (a) FACS: the left peak (black) indicates the basal fluorescence mediated by the GFP reporter gene in parental MCF-7G (that is, pVEGF-GFP/MCF-7) cells. The right peak (gray) indicates the GFP-mediated signal in clone #16. (b) Western blot: lane 1, wild-type MCF-7 cells cultured for 24h in media containing 1μM tetracycline (WT, Tet+); lane 2, HSF4 EST-integrated MCF-7 cells cultured as in lane 1 (EK-F4, Tet+); lanes 3 and 4, wild-type MCF-7 or EK-F4 cells, respectively, cultured for 24h in media lacking tetracycline. Numbers represent relative HSF4 levels after normalization for α-tubulin expressed as fold induction over control (wild-type MCF-7 cells cultured in media lacking tetracycline). (c) Semi-quantitative RT–PCR. GUSB (β-glucuronidase) was used as the amplification control for VEGF. Culture conditions in the presence (Tet+) or absence (Tet–) of tetracycline were as in (b). Lane 1, wild-type MCF-7 cells; lane 2, HSF4 EST-integrated MCF-7 cells; lane 3, wild-type MCF-7 cells; lane 4, HSF4 EST-integrated MCF-7 cells. (d) Real-time PCR (see Materials and methods). The templates were RNA isolated from wild-type MCF-7 cells cultured for 24h in media lacking tetracycline (WT) or MCF-7 cells containing a chromosomally integrated HSF4 EST cultured under the same conditions (EK-F4). Numbers represent VEGF mRNA levels after normalization for GUSB mRNA level as fold induction over control (wild-type MCF-7 cells without tetracycline). (e) Western blot of total protein isolated from wild-type MCF-7 cells treated with the transfection buffer as indicated in Materials and methods (Mock) or with siRNA consisting of a randomly scrambled sequence (SC), an antisense oligonucleotide corresponding to a segment of the HSF4-coding sequence (AS-F4), or siRNA corresponding to a segment of the HSF4-coding sequence. The numbers represent relative HSF4 levels expressed as fold induction over the Mock control after normalization for α-tubulin. (f) Semi-quantitative RT–PCR. Mock, Scramble, AS-F4 and si-F4 are as in (e). (g) Real-time PCR was carried out as in (d). Mock, si-F4 and AS-F4 are as in (e). Numbers represent VEGF mRNA levels upon normalization for GUSB mRNA level as fold induction over the Mock control. (h) Enzyme-linked immunosorbent assay measurement of secreted VEGF concentration in MCF-7 conditioned medium. Mock and si-F4 are as in (e).

Full figure and legend (130K)

Two alternatively spliced transcripts encoding distinct HSF4 isoforms, HSF4a and HSF4b, which possess different transcriptional activity have been described (Nakai et al., 1997; Pirkkala et al., 2001; Chi and Karliner, 2004; Min et al., 2004; Akerfelt et al., 2007; Fujimoto et al., 2004, 2008); whereas RT–PCR results showed that both HSF4a and HSF4b mRNAs are expressed in MCF-7 cells, only the protein isoform encoded by HSF4a was detected in MCF-7 cells by western blotting. The ability of the HSF4 EST to regulate VEGF expression was shown directly, by use of an MCF-7G-derived cell clone, EK-F4 that contains the HSF4 EST inserted into the chromosome under control, in antisense orientation, of a tetracycline-repressed cytomegalovirus (CMV) promoter (Figures 1b–d). In the absence of tetracycline repression, the abundance of HSF4 protein in EK-F4 cells was about one fifth the abundance in the parental cell line (Figure 1b). Similarly, RT–PCR analysis showed that VEFG mRNA was elevated more than threefold (Figures 1c and d) when the CMV promoter driving antisense transcription of the HSF4 EST was activated. Turndown of antisense transcription of the HSF4 EST by the addition of tetracycline returned VEGF mRNA to normal levels (Figure 1c).

Further evidence of the stimulatory effect of HSF4 downregulation on VEGF expression was obtained by introducing an antisense oligodeoxynucleotide (AS-F4) complementary to HSF4, or alternatively, an HSF4-specific double-stranded siRNA (si-F4) into MCF-7 cells. Both events, which decreased HSF4a protein to ~40% of normal in a pool of transfected cells, produced a reciprocal increase in VEGF mRNA abundance, as indicated by RT–PCR analysis (Figures 1e–g). Enzyme-linked immunosorbent assay measurements showed greater accumulation of VEGF protein in the culture media during a 96-h growth period of cells transfected with HSF4 siRNA than in cells similarly transfected with a randomly scrambled siRNA (Figure 1h). Collectively, these experiments indicate that HSF4 deficiency in MCF-7 cells increases the production of VEGF mRNA and protein.

The effects of HSF4 in VEGF expression are mediated through HIF-1α

The production of VEGF is controlled largely by transcription factors that include HIF-1, HIF-2 and SP1 (Jiang et al., 1997; Mazure et al., 1996, 1997; Tsuzuki et al., 2000; Bos et al., 2001; Semenza, 2003; Olenyuk et al., 2004; Pouyssegur et al., 2006; Brahimi-Horn and Pouyssegur, 2007; Fraisl et al., 2009). We observed that a cis-acting mutation known to prevent transcriptional activation by HIF-1α of a luciferase reporter gene linked to an HRE of the VEGF promoter (Mazure et al., 1996, 1997; Tsuzuki et al., 2000; Olenyuk et al., 2004; Pore et al., 2004) abrogated the elevation of luciferase activity otherwise seen during downregulation of HSF4 (Figure 2a). The implication of this finding, namely that the effects of HSF4 downregulation on VEGF production are mediated through HIF-1α, was confirmed by western blotting (Figure 2b) and RT–PCR analysis (Figure 2c); these analyses showed that HIF-1α mRNA and protein are both elevated in EK-F4 cells, and that the addition of tetracycline, which turns down antisense transcription of the chromosomally inserted HSF4 EST in these cells, reversed this increase. Consistent with these results, a population of MCF-7 cells transfected with a single-strand DNA oligonucleotide complementary to HSF4 mRNA (AS-F4) or with siRNA directed against HSF4 mRNA (si-F4) also showed elevation of HIF-1α mRNA (Figure 2c).

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Effect of HSF4 knockdown on HIF-1α expression. (a) Schematic representation (upper panels) and results of VEGF-promoter luciferase reporter assay (lower panels). Wild-type (WT) MCF-7 or HSF4 EST-integrated (EK-F4) cells were transfected with either a wild-type 1.4-kb VEGF-promoter-driven LUC reporter (1.4-kb-wt) or an HRE-mutated 1.4-kb VEGF-promoter-driven LUC reporter (HRE-mt) for 24h; aliquots containing half of the EK-F4 cells or 1/5 of WT cells were untreated. Half of EK-F4 cells and 1/5 of WT cells were treated with 1μM tetracycline (Tet+) for 24h and 3/5 of WT cells were treated with HSF4 antisense oligo (AS-F4) for 8h or HSF4 siRNA (si-F4) for 36h, and luciferase activity was then assessed in triplicate. Numbers represent luciferase activity upon normalization for CMV promoter-driven β-galactosidase activity as fold induction over the WT Tet+ control. (b) Western blot: lane 1, protein was isolated from wild-type MCF-7 cells cultured in the presence of 1μM tetracycline for 24h (WT, Tet+); lane 2, MCF-7 cells containing a chromosomally integrated HSF4 EST cultured under the same conditions (EK-F4, Tet+); lane 3, wild-type MCF-7 cells cultured without tetracycline (WT, Tet−); lane 4, as in lane 2, except that cells were cultured in media lacking tetracycline (EK-F4, Tet−). Numbers represent HSF4 levels upon normalization for α-tubulin expressed as fold induction over control (wild-type MCF-7 cells without tetracycline). Lane 5, transfection buffer-treated wild-type MCF-7 cells (Mock); lane 6, MCF-7 cells treated with HSF4 antisense oligo (AS-F4); lane 7, MCF-7 cells treated with siRNA to HSF4 (si-F4). Numbers represent HIF-1α levels upon normalization for α-tubulin expressed as fold induction over the Mock control. (c) Real-time PCR. WT, EK-F4, Mock, AS-F4, si-F4 and Tet+/− are as in Figure 1. Numbers represent HIF-1α mRNA levels upon normalization for GUSB mRNA level as fold induction over the Mock control.

Full figure and legend (100K)

HSF2 is also required to maintain repression of HIF-1α transcription

The findings presented above indicate that HSF-4 downregulation increases the production of HIF-1α mRNA and protein, implying that HSF4 functions as a repressor of HIF-1α expression. However, during our studies we surprisingly observed that adventitious overexpression of mRNA encoding either HSF4a or HSF4b isoform also resulted in increased HIF-1α expression (Figures 3a and b)—and to the same extent as HSF4 deficiency (cf. Figures 1 and 3)—suggesting that HSF4 actions on HIF-1α expression are more complex than simple transcriptional repression. Individual HSFs are known to function in concert with, or in opposition to, other members of the HSF family (Mathew et al., 2001; Pirkkala et al., 2001; Chi and Karliner, 2004; Xing et al., 2005; Loison et al., 2006; Akerfelt et al., 2007; Ostling et al., 2007; Sandqvist et al., 2009; Yamamoto et al., 2009), and we hypothesized that HSF4 may participate with another HSF family member in the regulation of HIF-1α transcription. To investigate this possibility, we tested the effects of downregulation of two other HSF proteins known to be produced in human cells (that is, HSF1 and HSF2) (Fujimoto and Nakai, 2010). Whereas downregulation of HSF1 had no detectable affect on the HIF-1α expression, siRNA against HSF2 or an antisense oligonucleotide complementary to the HSF2 transcript (AS-F2) resulted in a two to fourfold increase in the steady state level of HIF-1α protein and mRNA, respectively (Figures 4a and b). These findings, which parallel the HIF-1α expression changes occurring during HSF4 downregulation (Figure 2), suggested that HSF2 is required along with HSF4 to maintain steady state repression of HIF-1α transcription. Two alternatively spliced transcripts encoding distinct HSF2 isoforms, HSF2α and HSF2β, possessing different transcriptional activity have been described (Pirkkala et al., 2001), and RT–PCR results showed that both HSF2α and HSF2β mRNAs are expressed in MCF-7 cells (data not shown). Overproduction from a CMV promoter of either of the two HSF2 isoforms, which are similar in molecular mass and were not resolved by western blotting under the conditions employed (Figure 4c), increased HIF-1α protein and mRNA (Figures 4c and d)—as had been observed during overproduction of HSF-4.

Figure 3.
Figure 3 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Induction of HIF-1α transcription by overexpression of HSF4 isoforms. (a) Western blot: MCF-7 cells were transfected with the vector alone (CON) or with the plasmid construct expressing HSF4a (OE4a) or HSF4b (OE4b) in 24-well plates. Numbers under the HSF4a or HIF-1α bands represent HSF4a or HIF-1α levels after normalization for α-tubulin expressed as fold induction over the control. (b) Real-time PCR. CON, OE4a and OE4b are the same as in (a). Numbers represent HIF-1α mRNA levels upon normalization for GUSB mRNA level as fold induction over the control.

Full figure and legend (68K)

Figure 4.
Figure 4 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Effect of knockdown or overexpression of HSF2 on expression of HIF-1α. (a) Real-time PCR: MCF-7 cells treated with transfection buffer alone (Mock, the transfection buffer-treated wild-type MCF-7 cells; si-F1, HSF1 siRNA treated MCF-7) or with the indicated antisense oligonucleotides or siRNAs. Numbers represent HIF-1α mRNA levels upon normalization for GUSB mRNA level as fold induction over the Mock control. (b) Western blot. Mock, si-F2 and AS-F2 are as in (a). Numbers represent HSF2 or HIF-1α levels upon normalization for α-tubulin expressed as fold induction over the Mock control. (c) Western blot: MCF-7 cells were transfected with the vector alone (CON) or with the plasmid construct expressing HSF2α (OE2a) or HSF2β (OE2b) in 24-well plates. Numbers under the HSF2α, HSF2β or HIF-1α bands represent their levels after normalization for α-tubulin expressed as fold induction over the control. (d) Real-time PCR. CON, OE2a and OE2b are the same as in (c). Numbers represent HIF-1α mRNA levels upon normalization for GUSB mRNA level as fold induction over the control.

Full figure and legend (101K)

HSF4 and HSF2 bind to HSE-containing loci in the HIF-1α promoter region

Earlier work has shown that HSF regulation of gene expression involves the binding of HSFs to double-strand regions of DNA known as HSEs; HSE sequences, which are highly conserved among a broad range of eukaryotes, consist of multiple inverted repeats of the pentamer nGAAn, or of a variant of this sequence (Somasundaram and Bhat, 2004; Xing et al., 2005; Akerfelt et al., 2008; Fujimoto et al., 2008; Murphy et al., 2008). Consistent with our evidence that HSF4 and HSF2 are regulators of HIF-1α transcription, analysis of the 2-kb DNA region upstream from the HIF-1α transcriptional start site by the Vector NTI motif program revealed two discontinuous regions rich of ‘GAA/TTC’ repeats. The positions of these loci, −901/−864 (HSE1) and −1463/−1422 (HSE2) are shown in Figure 5a relative to the previously identified site of initiation of HIF-1α transcripts (Human genome BLAT Search/The UCSC Genome Bioinformatics).

Figure 5.
Figure 5 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

HSF4 and HSF2 bind to the HIF-1α promoter. (a) Schematic representation of the HIF-1α promoter region showing locations of HSE1 and HSE2 and their sequences. Nucleotides identical in nature and position to those in consensus HSE sequences (GAA/TTC) are shown in capital letters. Positions of two pairs of PCR primers for the Ch-IP assay (P1–P2, P3–P4) are indicated by arrows. Bent arrow indicates the major transcription start site for HIF-1α (Human genome BLAT Search/The UCSC Genome Bioinformatics). (b) Ch-IP by using MCF-7 cell lysates: agarose gels showing DNA fragments amplified by using primers P1 and P2 (259bp, lanes 1–4 and 9–18), or by using primers P3 and P4 (461bp, lanes 5–8); lane D, 100bp DNA ladder; lanes 1, 5 and 18, input control; lanes 2 and 6, amplification of anti-tubulin precipitates; lanes 3 and 7, amplification of anti-HSF2 precipitates; lanes 4 and 8, amplification of anti-HSF4 precipitates; lanes 9, 10 and 11, amplification of anti-tubulin precipitates, anti-HSF2 precipitates, anti-HSF4 precipitates from scramble-control-transfected cells; lanes 12, 13 and 14, amplification of anti-tubulin precipitates, anti-HSF2 precipitates, anti-HSF4 precipitates from HSF2-siRNA-transfected cells; lanes 15, 16 and 17, amplification of anti-tubulin precipitates, anti-HSF2precipitates, anti-HSF4 precipitates from HSF4-siRNA-transfected cells; PA indicates the band from the amplification of the primer oligomers. (c) Ch-IP by using MCF-7 cell (top panel) or HeLa cell (bottom panel) lysates. Agarose gels showing DNA fragments amplified by using primers P1 and P2 (259bp). Lane D, 100bp DNA ladder; lane 1, amplification of anti-HSF2 precipitates from empty vector transfected cells; lane 2, amplification of anti-HSF4 precipitates from empty vector transfected cells; lane 3, amplification of anti-HSF2 precipitates from pcDNA3-HSF2α-transfected cells; lane 4, amplification of anti-HSF4 precipitates from pcDNA3-HSF2α-transfected cells; lane 5, amplification of anti-HSF2 precipitates from pcDNA3-HSF4a-transfected cells; lane 6, amplification of anti-HSF4 precipitates from pcDNA3-HSF4a-transfected cells; lane 7, amplification of anti-tubulin precipitates; lane 8, input control. PA indicates the band from the amplification of the primer oligomers. (d) Gel-shift assay: the arrowhead indicates the positions of DNA fragments bound by HSFs. PB indicates the positions of free [32P] HSE86 probes. Lanes 1 and 9, [32P] HSE86 probe only; lanes 2 and 10, [32P] HSE86 with MCF-7 nuclear extracts; lane 3, [32P] HSE86 with MCF-7 nuclear extracts and 50nM unlabeled 86mu5 (control oligonucleotide); lanes 4, 15, [32P] HSE86 with MCF-7 nuclear extracts and unlabeled HSE86, the concentration is 12.5nM and 50nM, respectively; lanes 5, 6, 11, 12, [32P] HSE86 with MCF-7 nuclear extracts and unlabeled synthetic HSE1 oligonucleotide, the concentration is 5.0nM, 2.5nM, 50nM and 25nM, respectively; lane 7, 8, 13, 14, [32P] HSE86 with MCF-7 nuclear extracts and unlabeled synthetic HSE2 oligonucleotide, the concentration is 5.0nM, 2.5nM, 50nM and 25nM, respectively.

Full figure and legend (143K)

Chromatin immunoprecipitation (Ch-IP) assays were performed to determine whether HIF-1α promoter region segments containing putative HSE sequences do, in fact, bind to HSF4 and/or HSF2. In these experiments, complexes immunoprecipitated by anti-HSF4 or HSF2 antibodies were used as substrates for 30 PCR cycles primed by oligonucleotide pairs corresponding to locations bracketing HSE1 (primers P1 and P2) or HSE2 (primers P3 and P4) (Figure 5a). As seen in Figure 5b, PCR amplification of DNA immunoprecipitated from MCF-7G cells by either of the two anti-HSF antibodies yielded DNA fragments 259bp in length (Figure 5b, left panel, lanes 3 and 4) or 461bp in length (Figure 5b, left panel, lanes 7 and 8). Sequencing of these fragments (data not shown) confirmed that they contain the expected HSEs. Transfection of MCF-7 cells by HSF2 siRNA abrogated amplification of the 259-bp DNA fragment in HSF2 antibody immunoprecipitates, but not in the HSF4 antibody immunoprecipitates (Figure 5b, right panel, comparison of lanes 10 and 13, 11 and 14); similarly, transfection of MCF-7 cells by HSF4 siRNA abrogated amplification of the 259-bp DNA fragment in the HSF4 antibody immunoprecipitates, but not in the HSF2 antibody immunoprecipitates (Figure 5b, right panel, comparison of lanes 11 and 17, and lanes 10 and 16). These findings confirm that HSF2 and HSF4 each bind to at least two separate HSEs in the HIF-1α promoter.

Together, the requirement for both HSF2 and HSF4 for repression of HIF-1α transcription and the finding that overproduction of either protein can also lead to activation of HIF-1α transcription suggested a model in which (a) HSF2 and HSF4 are components of a complex that mediates repression of the HIF-1α promoter, and (b) disruption of the normal stoichiometry of the postulated repression complex interferes with the ability of the complex to repress HIF-1α transcription. Consistent with the above model, we observed, by an additional Ch-IP analysis that overexpression of either HSF2 or HSF4 reduced the amount of the other HSF present in complexes with DNA containing the HSE sites that we identified in the HIF-1α promoter. Transfection of MCF-7 cells (Figure 5c, top panel) by plasmids expressing HSF2α increased the amount of 259bp DNA fragment immunoprecipitated by HSF2 antibody (Figure 5c, comparison between lanes 3 and 1), but decreased the amount of this fragment immunoprecipitated by antibody to HSF4 (Figure 5c, comparison of lanes 4 and 2). Conversely, transfection by plasmids expressing HSF4a produced opposite results (Figure 5c, comparison of lanes 6 vs 2 and comparison of lanes 5 vs 1). These results suggest the ability of each HSF to displace the other from complexes at the HSE-binding site. Similar findings were observed using chromatin obtained from HeLa cells (Figure 5c, bottom panel).

That HSF2 and HSF4 can interact specifically with HSE sequences was shown by electrophoretic mobility shift assays using in vitro synthesized oligonucleotides. The [32P]-labeled HSE86 probe contains three perfect nGAAn pentameric units and was previously reported to bind with both HSF2 and HSF4 (Fujimoto et al., 2008). As seen in Figure 5d, incubation of [32P] HSE86 with nuclear extracts from MCF-7 cells resulted in a shift of the [32P]-labeled band to the position indicated by the arrow (lanes 2 and 10)—confirming that HSF can bind specifically to this HSE; addition of unlabeled HSE86 partially (lane 4) or at a higher concentration, totally (lane 15) competed with the band shift. Unlabeled synthetic oligonucleotides comprising the HSE1 (lanes 5 and 6) or HSE2 (lanes 7 and 8) sequence of the HIF-1α promoter also partially competed with the shift of the labeled probe, and at higher concentrations totally competed (HSE1, lanes 11 and 12; HSE2, lanes 13 and 14). These results confirm that the specific HIF-1α promoter region sites, which we have identified as discontinuous HSEs interact with both HSFs.

HSF2 and HSF4 are expressed in multiple tissues and have biologically diverse functions and tissue-specific functions (Akerfelt et al., 2007; Fujimoto and Nakai, 2010). The results presented above (Figure 5c) indicate that HSF interactions with HIF-1α promoter region HSEs also occur in vivo in HeLa cells. We tested the effect of HSF2 and HSF4 knockdown or overexpression on VEGF production in HeLa and 293T cell lines. As shown in Figure 6, these manipulations resulted in a VEGF increase in both cell types that was similar in magnitude to the increase we had observed in MCF-7 cells, indicating that the effects of HSFs on VEGF production extend beyond the human breast cancer cell line used to discover this role of HSFs.

Figure 6.
Figure 6 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Effects of knockdown or overexpression of HSF2 or HSF4 on VEGF production in 293T or HeLa cells. Con, si-F2, si-F4, OE2a, OE2b, OE4a and OE4b are same as in Figures 2,3,4. (a) 293T cells, 36h after the transfection. (b) HeLa cells, 36h after the transfection.

Full figure and legend (44K)

Top

Discussion

The findings reported here reveal a hitherto unsuspected mechanism of regulation of expression of the canonical HIF-1α/VEGF signaling pathway by HSFs 2 and 4. Our findings indicate that the effects of HSF2 and HSF4 on VEGF expression—and probably also on the expression of other genes controlled by HIF-1α/HRE-mediated signaling—is affected by the relative abundance of HSF4 and HSF2. Previous studies have shown that HIF-1α transcription can be upregulated by non-hypoxic events, as well as by hypoxia—leading to increased expression of VEGF and other HIF-1α targeted genes (Jiang et al., 1997; Page et al., 2002; Busca et al., 2005; Pipinikas et al., 2008; Nardinocchi et al., 2009). Our results raise the prospect that HSF2/HSF4 complexes at the HSE elements of the HIF-1α promoter, which potentially may vary quantitatively in composition in different types of normal or tumor cells, may be one such mechanism of non-hypoxic regulation.

Transcriptional regulation of HSPs by HSF family members in response to heat, infection and inflammation, and to exposure to pharmacological agents and other stresses is a process that is evolutionarily conserved in eukaryotic cells (Pirkkala et al., 2001; Chi and Karliner, 2004). In addition to containing a highly conserved DNA-binding domain, HSFs all include hydrophobic regions that mediate the formation of dimers or trimers (Yamamoto et al., 2009; Fujimoto and Nakai, 2010). In mammalian cells, HSF1 is the major factor controlling stress-inducible HSP expression, whereas HSF4 and HSF2 have been reported to be more selective transcriptional regulators that do not control expression of classical HSPs (Nakai et al., 1997; Mathew et al., 2001; Zhang et al., 2001; Loison et al., 2006 for recent review, see Fujimoto and Nakai, 2010). Notwithstanding their common structural features, their ability to form homotrimers (Yamamoto et al., 2009), and the functionally important interactions that occur between HSF2 and HSF4 at the HSEs of the HIF-1α promoter region, HSFs that are not bound to HSEs have not been observed in earlier work to interact with each other (Mazure et al., 1997; Tsuzuki et al., 2000; Pirkkala et al., 2001; Fujimoto et al., 2004; Min et al., 2004); similarly, interaction between unbound HSFs was not detected by us in co-immunoprecipitation experiments (data not shown).

Our evidence that HSF2 and HSF4 can regulate transcription of HIF-1α together with our discovery of discontinuous HSE-like sequences in the HIF-1α promoter region suggested that these HSFs would bind to the sites identified by sequence analysis as possible HSEs. Results from our Ch-IP and electrophoretic mobility shift assays confirmed the role of these sites as HSEs. Whereas HSF1, which is the major stress-activated regulator, binds to continuous HSEs, both HSF2 and HSF4 bind preferentially to discontinuous HSEs. Our findings are also consistent with previous evidence that different HSFs can have targets in common (Akerfelt et al., 2008; Fujimoto et al., 2008; Yamamoto et al., 2009). Whereas the DNA-binding domains of the three HSFs identified in mammalian cells are highly similar (70.5% identity between HSF1 and HSF2, 76.2% identity between HSF1 and HSF4 and 62.4% identity between HSF2 and HSF4), HSFs nevertheless show different DNA-binding preferences (Somasundaram and Bhat, 2004; Fujimoto et al., 2004, 2008; Yamamoto et al., 2009). Potentially, such preferences may have a role in determining the effects of HSF/HRE complexes in the HIF-1α promoter region.

Our data show that HSF2 and HSF4 can each interact with both HSEs in the HIF-1α promoter and both are required to maintain steady state repression of HIF-1α transcription. They also show that overexpression of either HSF2 or HSF4 reduced the presence of the other HSF at each of the HIF-1α HSEs and additionally resulted in activation of HIF-1α transcription. The binding of HSF4 to discontinuous HSEs has been shown to be dependent on trimerization of this HSF (Yamamoto et al., 2009), and—as noted earlier—individual members of the HSF's family can function in concert with, or in opposition to other HSF family members. These considerations, together with our finding that disruption of the balance between the amounts of HSF2 and HSF4 abrogates steady state repression of HIF-1α transcription, lead us to speculate that the transcription-limiting interactions between HSF2 and HSF4 at HSE elements may involve heteromultimer formation. While this notion is consistent with both the observations reported here and with previous knowledge about HSF binding to HSEs, structural analysis of HSF/HSE complexes will be required for confirmation.

Given the cell-type specificity of HSF actions (Fujimoto and Nakai, 2010) it seems likely that the effects of HSFs on HIF-1α transcription observed in human breast cancer cells may differ in other cell types and may also be affected differently by heat shock and other stress conditions in different cell types (see Mathew et al., 2001; Pirkkala et al., 2001; Chi and Karliner, 2004; Xing et al., 2005; Loison et al., 2006; Akerfelt et al., 2007, 2008; Ostling et al., 2007; Sandqvist et al., 2009; Yamamoto et al., 2009). However, notwithstanding these considerations, our data show that the effects of HSF2 and HSF4 on VEGF production in at least two other cell lines—HeLa and 293T cells—are similar to those seen in MCF-7 cells. A preliminary screen of the effects of physiological stresses has not identified stress conditions that regulate HIF-1α production in an HSF-dependent manner in the cell lines we have investigated.

Earlier work has shown that the HSP90 (Fujimoto et al., 2004; Min et al., 2004), which is a target of transcriptional control by HSF4 is required for HIF-1α-mediated VEGF expression (Minet et al., 1999; Isaacs et al., 2002). Additionally, expression of another HSF4-modulated heat shock protein, HSP27 (Fujimoto et al., 2004; Min et al., 2004), is correlated with tumor-induced angiogenesis (Wartenberg et al., 2001). Whereas it was tempting to suspect a role for HSP90 or HSP27 in the increased VEGF production we observed during downregulation of HSF4 or HSF2, real-time PCR and western blot experiments failed to detect any effect of HSF2 or HSF4 siRNA on the production of HSP90 protein in MCF-7 cells during a 24-h period after transfection of these siRNAs (data not shown). While HSF4 siRNA transfection did produce a decrease in HSP27 expression at both the mRNA and protein levels under the same experimental conditions, restoration of HSP27 expression during HSF4 downregulation in MCF-7 cells did not prevent upregulation of HIF-1α transcription—indicating that the derepression of HIF-1α transcription in these cells was not a consequence of deficient HSP27 expression (data not shown). However, our results do not exclude the possibility that HSF-induced changes in HSP expression can supplement the regulatory mechanisms reported here.

Top

Materials and methods

Cell culture, pVEGF-EGFP integrated MCF-7 cells

MCF-7 cells (estrogen receptor-positive and progesterone receptor-positive; American Type Culture Collection, Manassas, VA, USA) were used in all of the experiments. The MCF-7 line containing the tet-off transactivator tTA was established as described previously (Lu et al., 2004). The VEGF 2.8kb-promoter controlled EGFP expression vector pVEGF-EGFP (Fukumura et al., 1998) is a kind gift from Dr B Seed (Massachusetts General Hospital). We substituted the puromycin resistance marker into the hygromycin resistance marker. pVEGF-GFP was transfected into MCF-7 cells and the stably transfected cells were selected by adding hygromycin 200μg/ml 40h after the start of transfection. We cultured the MCF-7 cells and pVEGF-EGFP integrated MCF-7 cells in DMEM (Invitrogen, Carlsbad, CA, USA) containing 10% FBS.

Lentivirus production and infection

We produced lentivirus by transient transfection of 293T cells (calcium phosphate precipitation method) by using library DNA along with DNAs of packaging and VSVG envelope constructs as described previously (Lu et al., 2004).

Fluorescence-activated cell sorting

We established the EST-based genome-wide gene inactivation library of pVEGF-EGFP MCF-7 cells and utilized FACS to screen and isolate pVEGF-EGFP MCF-7 clones in which GFP-fluorescent signals were increased. FACS was done in the Stanford Shared FACS Facility.

Identification of EST clones

Genomic DNA was isolated from individual clones and the EST insertion was identified by PCR and DNA sequencing as described previously (Lu et al., 2004).

Treatment of cells with antisense oligos and siRNAs

Phosphorothioate antisense (AS) oligos used in this study were synthesized from the Stanford PAN Facility. The sequence of HSF4 AS oligo is 5′-AGAGGCTGTAAGTAGAAGGC, the sequence of HSF2 AS oligo is 5′-TGGGTTTCCTCCACAAGCGT and the sequence of HSF1 AS oligo is 5′-GGTCGAACACGTGGAAGCTG. siRNAs were purchased from Dharmacon (Lafayette, CO, USA). The sequence of HSF4 siRNA is 5′-GAGACAAAUUUGGGCCUUAUU (sense), the sequence of HSF2 siRNA is 5′-UAUCGACUCUGGAAUUGUAUU (sense) and the sequence of HSF1 siRNA is 5′-CCACUUGGAUGCUAUGGAC. A scrambled siRNA duplex (5′-CAGCGCUGACAACAGUUUCAU) was used as a control.

MCF-7 cells plated at the density of 4000 cells per 1.9cm2 were treated 1 day later for 1 or 2 days with AS oligo or siRNA, respectively. Effectene (Qiagen, Valencia, CA, USA) and Lipofectamine2000 (Invitrogen), were used to increase the AS oligo or siRNA uptake into the cells. The cells were treated with 0.8μM AS oligo or 0.1μM siRNA after a pre-incubation for 20min with 3mg/ml Effectene or Lipofectamine2000 in serum-free OPTI-MEM (Invitrogen). After 4h, the medium was replaced with standard culture medium as described above. For the AS oligo experiments, cells were harvested after another 4h incubation. For siRNA experiments, cells were harvested after 32h incubation.

RNA isolation and reverse transcription

Total RNA was prepared from harvested cells with the RNeasy Midi Kit (Qiagen) and treated with RNase-free DNase (Qiagen) according to the manufacturer's instructions.

The SuperScript II (Invitrogen) reagent set was used for reverse transcription reaction, which was carried out using standard protocol provided by the manufacturer. The reverse transcription reaction was performed at 42°C for 50min and was stopped by heating up to 70°C for 15min.

Semi-quantitative PCR and real-time quantitative PCR

The reverse transcription and real-time PCR primers were designed to span introns by using Vector NTI 9 (Invitrogen). GUSB (β-glucuronidase) mRNA is amplified as the control. The specific primers sequences are HIF-1α 5′-GTCTCACGAGGGGTTTCCCG-3′ (forward) and 5′-GCCGAGATCTGGCTGCATCT-3′ (reverse); VEGF 5′-TGCACCCATGGCAGAAGGAG-3′ (forward) and 5′-TGTGCTGGCCTTGGTGAGGT (reverse); GUSB 5′-CGGCCTGTGACCTTTGTGAG-3′ (forward) and 5′-ATTCCCCAGCACTCTCGTCG-3′ (reverse).

Real-time PCR was performed by using Bio-Rad iQ SYBR Green Supermix (to 1 × ), 250nM each PCR primer in 20μl and a Bio-Rad fluorescent real-time PCR instrument. The reaction started with 3min of pre-incubation at 95°C followed by 40 amplification cycles. The threshold cycle (Ct) was determined by use of the maximum-second-derivative function of the software. Formation of expected PCR product was confirmed by agarose gel electrophoresis (2%) and melting curve analysis.

Transfection and assays for luciferase and β-galactosidase

A 1.4-kb wild-type VEGF-promoter-driven luciferase construct and an HRE-mutated VEGF-promoter-driven luciferase construct (Pore et al., 2004), are kind gifts from Dr A Maity (University of Pennsylvania School of Medicine, Philadelphia, PA, USA). A 5-HRE-driven luciferase construct is a kind gift from Dr Giaccia (Cancer Biology Research Laboratory, Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA, USA). For promoter activity analysis, transient transfection was carried out using Lipofectamine2000 (Invitrogen). Unless specified in the figure legends, the cells were plated in 24-well tissue culture plates at 4 × 104/well and cultured for 18h before being transfected with 0.1μg/well of VEGF/HRE promoter luciferase reporter construct. As the control for transfection efficiency 0.1μg/well of pCMV-β-Gal expression vector was simultaneously transfected. For co-expression assays, a total 0.02−0.2μg/well of expression vector for transcription factors were used, empty pcDNA3 is used as the control. The cells were harvested 36h after transfection, and the luciferase activity and β-galactosidase expression levels were assayed according to the manufacturer's protocols (Promega, Sunnyvale, CA, USA). The promoter activities were normalized in relative light units/milliunit of β-galactosidase activity.

The pcDNA3-HSF4 expressing plasmid (Zhang et al., 2001) was kindly offered by Dr NF Mivechi (Medical College of Georgia, Augusta, GA, USA). The pcDNA3-HSP27 expressing plasmid (Schafer et al., 1999) was kindly offered by Dr R Benndorf (University of Michigan, Ann Arbor, MI, USA).

Western blotting

Unless specified in the figure legends, the cells were plated in 6-well tissue culture plates at 2 × 105/well and cultured for 18h before being used. For overexpression assay, a total 0.8μg/well of expression vector for transcription factors were used, blank pcDNA3 was used as the control. We prepared the nuclear or whole cell extracts and performed the western blot analysis by using a standard protocol. Antibodies were used at 1:1000 dilution for the polyclonal rabbit anti-HSF4b antibody (a kind gift from Dr SP Bhat, University of California, Los Angeles, CA, USA), 1:1000 for the monoclonal mouse anti-HSF4a (BD Biosciences, San Jose, CA, USA), 1:250 for the monoclonal mouse anti-HIF-1α (BD Biosciences), 1:1000 for the polyclonal rabbit anti-HSP90 (Cell Signaling Technology, Beverly, MA, USA), 1:1000 for the monoclonal mouse anti-HSP27 (Cell Signaling Technology), 1:250 for the monoclonal rat anti-HSF2 (Millipore, Billerica, MA, USA) and 1:5000 for the monoclonal mouse anti-α-tubulin (Sigma-Aldrich, St Louis, MO, USA). Antigen–antibody complexes were detected by chemiluminescence (Western Lightning, GE Healthcare, Piscataway, NJ, USA). Density of the bands was analyzed using ImageJ software 1.41b (Image Processing and Analysis in Java, National Institutes of Health, Bethesda, MD, USA). The results are expressed as a protein/α-tubulin density ratio.

Chromatin immunoprecipitation

Non-stimulated MCF-7 cells, MCF-7 cells transfected with HSF4 or HSF2 siRNAs, and MCF-7 cells transfected with pcDNA3-HSF4 were treated with 10% formaldehyde for 10min at 37°C. Then the cells were washed twice with ice-cold PBS and suspended in the lysis buffer (0.5% NP-40, 150mM NaCl, 50mM Tris pH 7.5, w/protease inhibitors). After 10min incubation on ice, cells were centrifuged to pellet the nuclei. Nuclei were then suspended in the nuclei lysis buffer (1% SDS, 10mM EDTA, 50mM Tris pH 8.1, w/protease inhibitors), incubated on ice for 10min and sonicated for 1.5min (30s each time, then cooled on ice and repeated twice). The sheared chromatin was then immunoprecipitated by using a specific monoclonal anti-HSF2 antibody or a specific monoclonal anti-HSF4 antibody. After extensive washing and then elution, the crosslink was reverted by heat treatment (65°C overnight and protease K digestion). The captured genomic DNA fragments were then purified by using Qiagen PCR purification kit. Identification of the captured HIF-1α promoter fragments was performed by PCR analysis by using the promoter primers. The sequences of primers covering region −915 to −658, are 5′-ACTCTTTGCCACGGAGCACA (forward) and 5′-GCTTGCAAAGTTGCCAAAGG (reverse); the sequences of primers covering region −1455 to −995, are 5′-TTGAGCCCAACAAAGTAGCATT (forward), 5′-CTTCTCTTCAGGCATTTCCCA (reverse). Thirty cycles of PCR were performed and the amplified products were analyzed on a 2% agarose gel.

Gel mobility shift assays

Gel mobility shift assays were performed as described previously (Fujimoto et al., 2008). Briefly, nuclear extracts (~20–30μg of protein) and ~20–30fmol of 32P-labeled double-stranded HSE86 oligonucleotide ([32P]HSE86) were used in a typical assay. The sequence of the forward strand is 5′-ggGAgggGAAaaTTCgaGAAgaagtga-3′. Competitive binding is done by adding unlabeled double-stranded oligonucleotides HSE1 or HSE2 to the reaction. The sequences of the forward strands of HSE1 and HSE2 are 5′-aaggccattttTTCtactctTTCcctGAAattggtt-3′ and 5′-ttcTTCtcTTCagGcAttTcCcatggTTCttTTCaag-3′, respectively. The sequences in capital letters denote the pentameric, GAA/TTC motifs. The reactions were carried out at room temperature for 20min. The entire volume of the reaction was electrophoresed on 4% acrylamide gel in a buffer containing 50mM Tris, pH 7.9, 40mM glycine and 1mM EDTA. The gels were run at 150V for ~180min. Under these conditions, no free (unbound) probe is retained in the gel. Complexes were detected by autoradiography. The oligonucleotide probes were end-labeled with polynucleotide kinase (New England BioLabs Inc Beverly MA, USA) using [γ-32P]ATP (Perkin-Elmer, Fremont, CA, USA) according to the manufacturer's protocol. The labeled oligonucleotides were annealed to the unlabeled complementary sequences in equimolar quantities, and double-stranded probes were used for gel shift.

Assay of VEGF protein levels by enzyme-linked immunosorbent assay

VEGF protein in conditioned media was assayed using a sandwich ELISA kit (Invitrogen, Camarillo, CA, USA), following the manufacturer's protocol. This assay detects human VEGF in the range of 5–1500pg/ml. Samples with VEGF concentrations greater than this range were diluted with RPMI Media 640 (Invitrogen, Carlsbad, CA, USA) and reassayed. All samples were assayed in duplicate. Results are presented as mean values±s.d.

Other reagents and treatments

Desferrioxamine, MG132, novabiocin and tetracycline were purchased from Sigma-Aldrich. The dosage and time course are indicated in the figure legend.

Data analysis

Each experiment was performed independently three times. Thus, experiments presented in the figures are representative of three or more different repetitions. Statistical analysis was performed by using Microsoft Excel 2003.

Top

Conflict of interest

The authors declare no conflict of interest.

Top

References

  1. Akerfelt M, Henriksson E, Laiho A, Vihervaara A, Rautoma K, Kotaja N et al. (2008). Promoter ChIP-chip analysis in mouse testis reveals Y chromosome occupancy by HSF2. Proc Natl Acad Sci USA 105: 11224–11229. | Article | PubMed |
  2. Akerfelt M, Trouillet D, Mezger V, Sistonen L. (2007). Heat shock factors at a crossroad between stress and development. Ann NY Acad Sci 1113: 15–27. | Article | PubMed |
  3. Bos R, Zhong H, Hanrahan CF, Mommers EC, Semenza GL, Pinedo HM et al. (2001). Levels of hypoxia-inducible factor-1 alpha during breast carcinogenesis. J Natl Cancer Inst 93: 309–314. | Article | PubMed | ChemPort |
  4. Brahimi-Horn MC, Pouyssegur J. (2007). Harnessing the hypoxia-inducible factor in cancer and ischemic disease. Biochem Pharmacol 73: 450–457. | Article | PubMed | ISI | ChemPort |
  5. Busca R, Berra E, Gaggioli C, Khaled M, Bille K, Marchetti B et al. (2005). Hypoxia-inducible factor 1{alpha} is a new target of microphthalmia-. J Cell Biol 170: 49–59. | Article | PubMed | ISI | ChemPort |
  6. Chang Y, Ostling P, Akerfelt M, Trouillet D, Rallu M, Gitton Y et al. (2006). Role of heat-shock factor 2 in cerebral cortex formation and as a regulator of p35 expression. Genes Dev 20: 836–847. | Article | PubMed | ISI | ChemPort |
  7. Chi NC, Karliner JS. (2004). Molecular determinants of responses to myocardial ischemia/reperfusion injury: focus on hypoxia-inducible and heat shock factors. Cardiovasc Res 61: 437–447. | Article | PubMed | ISI | ChemPort |
  8. Fraisl P, Mazzone M, Schmidt T, Carmeliet P. (2009). Regulation of angiogenesis by oxygen and metabolism. Dev Cell 16: 167–179. | Article | PubMed | ISI | ChemPort |
  9. Fujimoto M, Izu H, Seki K, Fukuda K, Nishida T, Yamada S et al. (2004). HSF4 is required for normal cell growth and differentiation during mouse lens development. EMBO J 23: 4297–4306. | Article | PubMed | ISI | ChemPort |
  10. Fujimoto M, Oshima K, Shinkawa T, Wang BB, Inouye S, Hayashida N et al. (2008). Analysis of HSF4 binding regions reveals its necessity for gene regulation during development and heat shock response in mouse lenses. J Biol Chem 283: 29961–29970. | Article | PubMed | ISI |
  11. Fujimoto M, Nakai A. (2010). The heat shock factor family and adaptation to proteotoxic stress. FEBS J 277: 4112–4125. | Article | PubMed | ISI |
  12. Fukumura D, Xavier R, Sugiura T, Chen Y, Park EC, Lu N et al. (1998). Tumor induction of VEGF promoter activity in stromal cells. Cell 94: 715–725. | Article | PubMed | ISI | ChemPort |
  13. Isaacs JS, Jung YJ, Mimnaugh EG, Martinez A, Cuttitta F, Neckers LM. (2002). Hsp90 regulates a von Hippel Lindau-independent hypoxia-inducible factor-1 alpha-degradative pathway. J Biol Chem 277: 29936–29944. | Article | PubMed | ISI | ChemPort |
  14. Jiang BH, Agani F, Passaniti A, Semenza GL. (1997). V-SRC induces expression of hypoxia-inducible factor 1 (HIF-1) and transcription of genes encoding vascular endothelial growth factor and enolase 1: involvement of HIF-1 in tumor progression. Cancer Res 57: 5328–5335. | PubMed | ISI | ChemPort |
  15. Loison F, Debure L, Nizard P, le Goff P, Michel D, le Drean Y. (2006). Up-regulation of the clusterin gene after proteotoxic stress: implication of HSF1-HSF2 heterocomplexes. Biochem J 395: 223–231. | Article | PubMed | ISI | ChemPort |
  16. Lu Q, Wei W, Kowalski PE, Chang AC, Cohen SN. (2004). EST-based genome-wide gene inactivation identifies ARAP3 as a host protein affecting cellular susceptibility to anthrax toxin. Proc Natl Acad Sci USA 101: 17246–17251. | Article | PubMed |
  17. Mathew A, Mathur SK, Jolly C, Fox SG, Kim S, Morimoto RI. (2001). Stress-specific activation and repression of heat shock factors 1 and 2. Mol Cell Biol 21: 7163–7171. | Article | PubMed | ISI |
  18. Mazure NM, Chen EY, Laderoute KR, Giaccia AJ. (1997). Induction of vascular endothelial growth factor by hypoxia is modulated by a phosphatidylinositol 3-kinase/Akt signaling pathway in Ha-ras-transformed cells through a hypoxia inducible factor-1 transcriptional element. Blood 90: 3322–3331. | PubMed | ISI | ChemPort |
  19. Mazure NM, Chen EY, Laderoute KR, Giaccia AJ. (1996). Oncogenic transformation and hypoxia synergistically act to modulate vascular endothelial growth factor expression. Cancer Res 56: 3436–3440. | PubMed | ISI | ChemPort |
  20. Min JN, Zhang Y, Moskophidis D, Mivechi NF. (2004). Unique contribution of heat shock transcription factor 4 in ocular lens development and fiber cell differentiation. Genesis 40: 205–217. | Article | PubMed | ISI | ChemPort |
  21. Minet E, Mottet D, Michel G, Roland I, Raes M, Remacle J et al. (1999). Hypoxia-induced activation of HIF-1: role of HIF-1alpha-Hsp90 interaction. FEBS Lett 460: 251–256. | Article | PubMed | ISI | ChemPort |
  22. Morimoto RI. (2008). Proteotoxic stress and inducible chaperone networks in neurodegenerative disease and aging. Genes Dev 22: 1427–1438. | Article | PubMed | ISI | ChemPort |
  23. Murphy LA, Wilkerson DC, Hong Y, Sarge KD. (2008). PRC1 associates with the hsp70i promoter and interacts with HSF2 during mitosis. Exp Cell Res 314: 2224–2230. | Article | PubMed | ISI |
  24. Nakai A, Tanabe M, Kawazoe Y, Inazawa J, Morimoto RI, Nagata K. (1997). HSF4, a new member of the human heat shock factor family which lacks properties of a transcriptional activator. Mol Cell Biol 17: 469–481. | PubMed | ISI | ChemPort |
  25. Nardinocchi L, Puca R, Guidolin D, Belloni AS, Bossi G, Michiels C et al. (2009). Transcriptional regulation of hypoxia-inducible factor 1alpha by HIPK2 suggests a novel mechanism to restrain tumor growth. Biochim Biophys Acta 1793: 368–377. | Article | PubMed | ISI |
  26. Ng YS, Krilleke D, Shima DT. (2006). VEGF function in vascular pathogenesis. Exp Cell Res 312: 527–537. | Article | PubMed | ISI | ChemPort |
  27. Olenyuk BZ, Zhang GJ, Klco JM, Nickols NG, Kaelin WG, Dervan PB. (2004). Inhibition of vascular endothelial growth factor with a sequence-specific hypoxia response element antagonist. Proc Natl Acad Sci USA 101: 16768–16773. | Article | PubMed |
  28. Ostling P, Bjork JK, Roos-Mattjus P, Mezger V, Sistonen L. (2007). Heat shock factor 2 (HSF2) contributes to inducible expression of hsp genes through interplay with HSF1. J Biol Chem 282: 7077–7086. | Article | PubMed | ISI | ChemPort |
  29. Page EL, Robitaille GA, Pouyssegur J, Richard DE. (2002). Induction of hypoxia-inducible factor-1alpha by transcriptional and translational mechanisms. J Biol Chem 277: 48403–48409. | Article | PubMed | ISI | ChemPort |
  30. Pipinikas CP, Carter ND, Corbishley CM, Fenske CD. (2008). HIF-1alpha mRNA gene expression levels in improved diagnosis of early stages of prostate cancer. Biomarkers 13: 680–691. | Article | PubMed | ISI |
  31. Pirkkala L, Nykanen P, Sistonen L. (2001). Roles of the heat shock transcription factors in regulation of the heat shock response and beyond. FASEB J 15: 1118–1131. | Article | PubMed | ISI | ChemPort |
  32. Pore N, Liu S, Shu HK, Li B, Haas-Kogan D, Stokoe D et al. (2004). Sp1 is involved in Akt-mediated induction of VEGF expression through an HIF-1-independent mechanism. Mol Biol Cell 15: 4841–4853. | Article | PubMed | ISI | ChemPort |
  33. Pouyssegur J, Dayan F, Mazure NM. (2006). Hypoxia signalling in cancer and approaches to enforce tumour regression. Nature 441: 437–443. | Article | PubMed | ISI | ChemPort |
  34. Sandqvist A, Bjork JK, Akerfelt M, Chitikova Z, Grichine A, Vourc'h C et al. (2009). Heterotrimerization of heat-shock factors 1 and 2 provides a transcriptional switch in response to distinct stimuli. Mol Biol Cell 20: 1340–1347. | Article | PubMed | ISI | ChemPort |
  35. Schafer C, Clapp P, Welsh MJ, Benndorf R, Williams JA. (1999). HSP27 expression regulates CCK-induced changes of the actin cytoskeleton in CHO-CCK-A cells. Am J Physiol 277: C1032–C1043. | PubMed | ISI | ChemPort |
  36. Semenza GL. (2003). Targeting HIF-1 for cancer therapy. Nat Rev Cancer 3: 721–732. | Article | PubMed | ISI | ChemPort |
  37. Somasundaram T, Bhat SP. (2004). Developmentally dictated expression of heat shock factors: exclusive expression of HSF4 in the postnatal lens and its specific interaction with alphaB-crystallin heat shock promoter. J Biol Chem 279: 44497–44503. | Article | PubMed | ISI |
  38. Tsuzuki Y, Fukumura D, Oosthuyse B, Koike C, Carmeliet P, Jain RK. (2000). Vascular endothelial growth factor (VEGF) modulation by targeting hypoxia-inducible factor-1alpha--> hypoxia response element--> VEGF cascade differentially regulates vascular response and growth rate in tumors. Cancer Res 60: 6248–6252. | PubMed | ISI | ChemPort |
  39. Wartenberg M, Donmez F, Ling FC, Acker H, Hescheler J, Sauer H. (2001). Tumor-induced angiogenesis studied in confrontation cultures of multicellular tumor spheroids and embryoid bodies grown from pluripotent embryonic stem cells. FASEB J 15: 995–1005. | Article | PubMed | ISI | ChemPort |
  40. Xing H, Wilkerson DC, Mayhew CN, Lubert EJ, Skaggs HS, Goodson ML et al. (2005). Mechanism of hsp70i gene bookmarking. Science 307: 421–423. | Article | PubMed | ISI | ChemPort |
  41. Yamamoto N, Takemori Y, Sakurai M, Sugiyama K, Sakurai H. (2009). Differential recognition of heat shock elements by members of the heat shock transcription factor family. FEBS J 276: 1962–1974. | Article | PubMed | ISI | ChemPort |
  42. Zhang J, Goodson ML, Hong Y, Sarge KD. (2008). MEL-18 interacts with HSF2 and the SUMO E2 UBC9 to inhibit HSF2 sumoylation. J Biol Chem 283: 7464–7469. | Article | PubMed | ISI |
  43. Zhang Y, Frejtag W, Dai R, Mivechi NF. (2001). Heat shock factor-4 (HSF-4a) is a repressor of HSF-1 mediated transcription. J Cell Biochem 82: 692–703. | Article | PubMed | ISI | ChemPort |
Top

Acknowledgements

The study was supported by funds from the Kwoh-Ting Li Professorship to SNC, and by a grant from the National Foundation for Cancer Research (NFCR) to SNC.