| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
and Hypoxia-Inducible Factor 1 to the Vascular Endothelial Growth Factor Promoter
Department of Physiology, University of Maryland School of Medicine, Baltimore, Maryland 21201
Address all correspondence and requests for reprints to: Robert D. Koos, Ph.D., Department of Physiology, University of Maryland School of Medicine, 655 West Baltimore Street, Baltimore, Maryland 21201-1559. E-mail: rkoos{at}umaryland.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
), rapidly stimulates VEGF expression in the uterus at the transcriptional level. The VEGF gene promoter, however, lacks a consensus estrogen response element (ERE), and attempts to identify the site through which E2 induces VEGF expression have yielded contradictory results. To address this question, we modified the chromatin immunoprecipitation method to identify transcription factors that interact with the VEGF promoter in the rat uterus in response to E2. Chromatin immunoprecipitation showed that both Sp1 and Sp3 were associated with a proximal, GC-rich region of the promoter before E2 treatment. E2 induced an increase in Sp1 binding and the recruitment of ER, and the coactivator p300 to this region. The association of ER persisted, however, after VEGF mRNA levels had declined again (at 4 h), indicating that other factors might be involved in that expression. Western analysis showed that both the - and ß-subunits of the transcription factor hypoxia-inducible factor 1 (HIF-1), which regulates VEGF expression in response to hypoxia and several hormones and growth factors, were present in the uterus. Furthermore, E2 rapidly induced their recruitment to the –944 to –611 bp region of the VEGF promoter, which contains the hypoxia response element to which HIF-1 binds. This binding was transient, matching the pattern of E2-induced VEGF expression. These results indicate that HIF-1 is an important mediator of E2-induced VEGF expression in the uterus. In addition, E2 also induced a later increase in HIF-1 mRNA and protein expression in the uterus, suggesting that it may be required for longer term effects of E2 on the uterus as well. In contrast to the uterus, HEC1A cells cultured in 95% air-5% CO2 (and therefore 20% O2) contained no HIF-1, consistent with the inability of E2 to stimulate the expression of VEGF by these and other cell types in vitro. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
interactions not with EREs but with Sp1 and Sp3 on a proximal, GC-rich segment (–66 to –47) of the promoter. Interestingly, however, this site mediated E2 suppression of VEGF expression in HEC1A endometrial cancer cells but induction in ZR-75 breast cancer cells. In the latter cells, ER and Sp1/3 were already associated with this region before E2 treatment. These results suggest that additional activation events or cellular factors are involved.
VEGF expression is also strongly induced in cells by hypoxia. As we have shown, this occurs via a hypoxia response element on the VEGF promoter, which binds the transcription factor hypoxia-inducible factor 1, or HIF-1 (14). HIF-1 is a heterodimeric protein consisting of a hypoxia-induced -subunit and a constitutively expressed ß-subunit, also known as the aryl hydrocarbon receptor nuclear translocator (ARNT) (15). Under normoxic conditions, HIF-1 is produced but rapidly degraded, resulting in low basal levels in normal tissues (16, 17). Hypoxia inhibits HIF-1 ubiquitination and degradation, causing its rapid accumulation in cells (18, 19). Interestingly, it has recently been shown that HIF-1 mediates the induction of VEGF expression by several hormones, growth factors, and cytokines. These include insulin (20), angiotensin II (21), thrombin (22), endothelin-1 (23), IGF-1 (24), prostaglandin E2 (25), epidermal growth factor (26), IL-1 (27), nitric oxide (28), and the catecholestrogen 4-OHE2 (29). There have been, however, no studies linking it to E2-induced VEGF expression.
Recently, a method called chromatin immunoprecipitation (ChIP) was developed to analyze the interaction of transcription factors with target gene promoters in native chromosomal DNA (reviewed in Ref. 30). Performed on cells collected at intervals after experimental treatment, ChIP can reveal the order and timing of transcriptional complex assembly and disassembly. It has been used, for example, to demonstrate dynamic ER interaction with coactivator proteins and target gene promoters (31) and with Sp1 (32) in breast cancer cells in vitro. ChIP results are more physiological than those obtained using reporter gene, mobility shift, or DNase footprinting assays because they reflect interactions between endogenous nuclear proteins and chromatin DNA in cells. We hypothesized that ChIP could also be performed on uterine tissue isolated from rats after E2 treatment in vivo. The feasibility of this approach was subsequently demonstrated in studies on liver (33, 34). We adapted ChIP, therefore, to examine the interaction of transcription factors with the VEGF promoter in the uteri of rats after E2 treatment in vivo.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). VEGF mRNA levels had declined again to the pretreatment level by 4 h. Uterine wet weight increased by 43% during this period (from 28.2 ± 2.3 g at 0 h to 39.6 ± 3.3 g at 4 h; P < 0.01, n = 9/group). As we have demonstrated previously (3), this edema is dependent on VEGF action.
|
association with the promoter of the pS2 gene in MCF-7 cells (31). As shown in Fig. 2, an increase in ER binding to a region of the pS2 promoter (–436 to +19) that contains a nonconsensus ERE essential for E2 induction was detected. E2 also induced a rapid association of the transcriptional coactivator p300 with the same region. The rapid occupancy of the pS2 promoter by ER and p300 after E2 exposure was similar to that reported by Shang et al. (31). When nonimmune serum was substituted for the specific antisera, no PCR products were generated.
|
and Sp1 to a region of the CKB promoter that contains a nonconsensus ERE flanked by two Sp1 sites (32). Using chromatin isolated from homogenates of whole uteri fixed immediately after removal from rats at intervals after treatment with E2, we performed ChIP assays using antibodies to rat ER or Sp1 and PCR primers encompassing this region of the rat CKB promoter (–678 to –319). ChIP showed that the association of both Sp1 and ER with this region increases within 1 h after E2 treatment (Fig. 3). This agrees with results obtained using cultured MCF-7 cells (32) and confirmed that ChIP can be used to examine the association of transcription factors with gene promoters in the rat uterus after E2 treatment in vivo.
|
, and p300 to the Rat VEGF Promoter/Sp1/Sp3 binding to this region. We initially determined, therefore, whether a similar interaction was also involved in E2-induced VEGF expression in the rat uterus. As shown in Fig. 4, both Sp1 and Sp3 were already associated with the –173 to +114 region of the VEGF promoter at 0 h. Constitutive binding of these two factors to the same region of the VEGF promoter has been described previously (13, 36). Treatment with E2 significantly increased the binding of Sp1 by 1 h (+3.4-fold; P < 0.05); Sp3 association also increased approximately 2-fold, but the change was not statistically significant. To demonstrate the specificity of the ChIP results, we also performed PCR using primers to a portion of the 3'-UTR of the VEGF gene (660–922, as numbered in Ref. 42). No Sp1 interaction with this region was detected (data not shown).
|
was associated with this same region at time zero, but E2 induced a marked recruitment of both ER (+3.8-fold; P < 0.01) and p300 (>2-fold; P < 0.05) to this region within 1 h (Fig. 5). By 4 h, by which time VEGF mRNA levels had returned to pretreatment levels (Fig. 1
), ER and p300 occupancy of the promoter was still elevated, whereas Sp1 binding had returned to the pretreatment level (Fig. 4).
|
with three 5'-upstream regions of the VEGF promoter (–124 to –403, –384 to –661, and –611 to –944). Although there was an enhanced yield of PCR product for the two immediate upstream regions after E2 treatment, the signal was weaker and less consistent than that obtained for the GC-rich region (data not shown). Because sonication of chromatin yields fragments of varying size, these signals most likely derive from larger chromatin fragments containing both the targeted region and the GC-rich region. No association of ER with the most 5'-region was detected (Fig. 6). The absence of equal or stronger association with the first two upstream regions, and the absence of any signal from the most distant upstream region, indicates that the major site of ER association is with the proximal GC-rich, Sp1/Sp3-binding region of the VEGF promoter, as previously indicated in studies using human cancer cells (12, 13), rather than with variant EREs (7, 8).
|
and HIF-1ß to the Rat VEGF Promoter with the –944 to –611 region was observed, this region does contain the HIF-1 binding site essential for VEGF-induction by hypoxia (14). As discussed previously, HIF-1 has also been implicated in the induction of VEGF expression by several hormones and growth factors. Therefore, the recruitment of both HIF-1 and HIF-1ß to this region of the promoter in the uterus after E2 treatment was examined. Figure 6 shows that at 0 h, little or no HIF-1 or HIF-1ß was present on this promoter region, but, after E2 treatment, both these components of the heterodimeric transcription factor were rapidly recruited to it (3.4-fold and 4.4-fold at 1 h vs. 0 h, respectively). This pattern of binding closely matched the pattern of VEGF mRNA expression (Fig. 1). Both HIF-1 components had dissociated by 4 h, again matching the pattern of VEGF mRNA expression. Like ER, HIF-1 regulation of gene transcription requires interaction with p300 (43, 44). Consistent with this, a matching increase in the association of p300 with this region was detected (Fig. 6). This correlation between HIF-1 and p300 binding to the VEGF promoter and E2-induced VEGF mRNA expression strongly suggests that HIF-1 recruitment plays an integral role in this event.
E2 Induction of HIF-1 Expression in the Uterus
HIF-1 levels increase sharply during hypoxia, leading to HIF-1-mediated expression of a wide range of genes critical for short-term adaptation to and long-term alleviation of hypoxia (45). This occurs mainly at the posttranslational level via stabilization of HIF-1 protein. To determine whether the rapid E2-induced increase in HIF-1 association with the VEGF promoter was due in any part to such an increase in the level of HIF-1, we assessed the effects of E2 on HIF-1 and HIF-1ß protein expression by Western blot analysis. As shown in Fig. 7, both HIF-1 and HIF-1ß were present in the rat uterus before E2 treatment. In both cases, the sizes of the bands were similar to the published sizes for the two proteins (
120 and 95 kDa, respectively). Similar results were obtained for HIF-1 using two different antibodies, further confirming the identity of the band, whereas the HIF-1ß band matched exactly that obtained with a HIF-1ß standard provided by the antibody supplier (see Materials and Methods). There was no significant change in the level of HIF-1 at 0.5 or 1 h. Thus, the HIF-1 associated with the VEGF promoter 1 h after E2 treatment appears to represent recruitment of existing HIF-1 protein. The level of HIF-1 was significantly higher, however, at 2 h (3-fold) and 4 h (4-fold). This suggests that higher levels of HIF-1 are required for later as well as early effects of E2 on gene expression. Similar to other tissues (16), HIF-1ß was abundant in the uterus and was unchanged after E2 treatment (Fig. 7), as is the case with hypoxia.
|
protein was due to increased protein stability or an increase in mRNA expression, HIF-1 mRNA was also measured by real-time RT-PCR. As shown in Fig. 8, the delayed increase in HIF-1 protein was matched by an increase in HIF-1 mRNA (2.5-fold at 2 h and 3.1-fold at 4 h; P < 0.01 in both cases). This indicates that E2 induction of HIF-1 expression, unlike the effect of hypoxia and most other stimuli, is at least in part at the level of transcription. Such an effect on HIF-1 transcription is rare but has been observed to occur in smooth muscle cells in response to angiotensin II (46).
|
in Cultured E2-Target Cells Correlates with the Inability of E2 to Stimulate VEGF Expression in Vitro production or activity in vivo. Exposure to high, nonphysiological O2 levels, however, effectively induces destruction of cellular HIF-1 (16, 47, 48). Culturing cells under standard conditions (i.e. in 95% air-5% CO2, and thus 20% O2), therefore, would be predicted to suppress HIF-1 to very low or undetectable levels. Thus, we hypothesized that there would be no HIF-1 in the endometrial or breast cancer cells commonly used for the study of E2 action, which could explain the weak effect of E2 on the expression of VEGF by such cells, at least compared with the uterus in vivo (12, 49, 50, 51, 52). We have found, for example, that E2 has little or no stimulatory effect on the expression of VEGF by HEC1A endometrial carcinoma cells under standard culture conditions (data not shown). To demonstrate that this correlates with a lack of HIF-1, we measured HIF-1 protein levels in HEC1A cells cultured in 95% air-5% CO2. As shown in Fig. 9, these cells contain no detectable HIF-1. To demonstrate that the cells could express HIF-1 if the O2 level was reduced, they were also cultured under low O2 (1–2%). This caused a large increase in HIF-1 protein (Fig. 9). This was matched by a 3-fold increase in VEGF mRNA levels at the same time point (n = 5 independent experiments; P < 0.01), demonstrating that the lack of stimulation of VEGF expression by E2 is not due to a general inability of the cells to increase VEGF expression. Taken together, these results are in keeping with the hypothesis that HIF-1 is required for strong E2 stimulation of VEGF expression.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The major finding of this study is that E2 rapidly causes recruitment of both ER and HIF-1 to the VEGF gene promoter in cells of the uterus. The likely site of this event is the endometrial luminal epithelial cell layer, the site of E2-induced VEGF expression (see below). This is the first time that HIF-1 has been linked to E2-induced gene expression. HIF-1 association with the promoter is transient, closely matching the pattern of VEGF expression, whereas that of ER persists after VEGF mRNA levels decline. This correlation between HIF-1 binding and transcription indicates that HIF-1 plays a central role in E2 regulation of VEGF expression. The two transcription factors bind at widely spaced sites: HIF-1 to the upstream region containing the hypoxia response element, as it does in response to hypoxia, and ER to the proximal GC-rich region that contains several Sp protein binding sites. We found that both Sp1 and Sp3 were associated with the latter region before treatment, and the binding of Sp1 increased further after the addition of E2. These results agree with those of Stoner et al. (12, 13) that ER interacts with the VEGF promoter primarily via binding to Sp proteins rather than to variant EREs (7, 8). The role of Sp proteins is not specific, however, to the effect of E2 on VEGF. The proximal Sp protein-binding region has been shown previously to be essential for basal VEGF gene expression (36, 37, 41) and induction by various growth factors (36, 53). It is also not specific to VEGF, because the expression of a number of other genes involves ER/Sp1 binding to similar GC motifs (54, 55). Our results suggest that the additional recruitment of HIF-1 is necessary for the rapid E2-induced increase in the transcription of the VEGF gene in the endometrium. The demonstration of a subsequent increase in HIF-1 expression raises the distinct possibility that HIF-1 also plays an important role in later effects of E2 on the uterus. HIF-1 has been linked to the regulation of more than 50 genes involved in energy metabolism, cell proliferation/survival, cell migration, and other processes (15). Many of these same genes are expressed cyclically in the uterus, and several are induced by E2. These include adrenomedullin (56), endothelin-1 (57), erythropoietin (58), IGF-binding proteins (59), and inducible nitric oxide synthase (iNOS) (60). We predict that HIF-1 will be found to be involved in E2 induction of some of these genes as well.
HIF-1 involvement in E2-induced VEGF expression is consistent with several recent studies showing that it also mediates the stimulation of VEGF expression by a number of other hormones, growth factors, and cytokines in various cell types under nonhypoxic conditions (20, 21, 22, 23, 24, 25, 26, 27, 28, 29). These factors appear to act through kinase pathways to increase HIF-1 protein levels and/or transcriptional activity. Activation leads to translocation to the nucleus, dimerization with HIF-1ß, and recruitment of coactivators, such as p300 (61). Our results indicate that the latter three of these events are triggered by E2. Activation of HIF-1 is most frequently associated with phosphatidylinositol 3-kinase (PI3K) and MAPK activity (45, 61), both of which can be rapidly activated by E2 (62, 63, 64, 65). Whether this depends on changes in the phosphorylation of HIF-1 or that of associated transcription factors or coactivators is unclear at this time. In preliminary studies, we have not been able to detect a change in the overall phosphorylation state of HIF-1 after E2 treatment. Activation of HIF-1 could involve phosphorylation at one or more sites and dephosphorylation elsewhere, as has been shown for other transcription factors. Unfortunately, antibodies to specific phosphorylated forms of HIF-1 are not yet available. Much more work is needed in this area. Regardless, the rapidity of HIF-1 recruitment and the known role of kinases in its activation raise the possibility that these events could be initiated by membrane ERs (66), which have now been linked to activation of both the MAPK and PI3K pathways (Refs. 67 and 68 and Fig. 10). PI3K and/or MAPKs have been shown recently to play a role in hepatocyte growth factor-induced VEGF expression through phosphorylation of constitutively bound Sp1 (53), and other recent studies have also linked phosphorylation of Sp1 by MAPK to VEGF expression (69, 70). Thus, the same kinase pathways might simultaneously activate both HIF-1 and Sp1.
|
in part because it causes hypoxia in uterine cells. E2 strongly stimulates endometrial RNA and protein synthesis, glucose oxidation, DNA replication, and other biosynthetic processes (71). The increase in oxygen consumption necessary for this increase in metabolism might, in the short term, exceed the rate of oxygen delivery. On the other hand, E2 also stimulates uterine blood flow, probably via the synthesis of the vasodilator NO, thereby supplying additional oxygen. As mentioned previously, HIF-1 has been linked to iNOS gene expression (15), and E2 induces iNOS in the uterus (60). It is possible that hypoxia and E2 could act synergistically to induce maximal VEGF expression, as is the case for hypoxia and TGF-ß in cultured cells (47).
Because these were in vivo studies, to directly demonstrate that the recruitment of HIF-1 is an absolute requirement for E2 induction of VEGF expression, by somehow blocking its synthesis or action in the uterus, is not possible at this time. In effect, however, this experiment has been performed many times on a variety of E2-responsive cells in vitro. Culturing cells under standard conditions (i.e. in 95% air-5% CO2) exposes them to 20% oxygen, a supraphysiological concentration that suppresses HIF-1 in cells to undetectable levels (Ref. 48 and Fig. 9). If HIF-1 plays an essential role in E2-induced VEGF expression, therefore, E2 should have little or no effect on cells in vitro, which is in fact the case. Most such studies either report no significant effect (49, 50, 51), small (<2-fold) stimulation (52), or inhibition (12) of VEGF expression by E2. We have likewise observed no response from HEC1A endometrial carcinoma cells and only relatively weak or delayed effects in MCF-7 and ZR-75 breast cancer cells (Ref. 72 and our unpublished work). All of the cell types used in these studies possess E2 receptors and exhibit other responses to E2. Studies using cells transfected with VEGF promoter-reporter gene constructs yield similar results, namely modest (8) or no induction by E2 (73). The weak in vitro response to E2 clearly contrasts with the rapid, robust stimulation in the uterus in vivo (Fig. 1 and Refs. 5 and 6) and with the strong induction by hypoxia in vitro. Only a few studies have reported relatively strong inductions in vitro, in one case with MCF-7 cells (74) and in another with ZR-75 cells (13). As stated previously, we have not been able to duplicate the results of either study with the same cell types in our laboratory. The difference could be due to variations between cell lines, but an alternative explanation is that differences in culture conditions may sometimes lead to the maintenance of an adequate HIF-1 concentration to mediate E2 induction. This could be caused by inadvertent exposure of cells to lower oxygen levels (e.g. as a result of using culture vessels that retard gas exchange or a large number of culture vessels per incubator) or supplementation with growth factors that stimulate HIF-1 production (24, 75). In one case where stronger stimulation was observed (76), the only time point examined was 24 h after treatment. We have found that the small induction of VEGF in MCF-7 cells by E2 is delayed (72). Such delayed effects could occur via a completely different mechanism than that operating in vivo. Studies are currently underway in our laboratory to define the culture conditions that would maintain a physiological HIF-1 level in cultured cells (i.e. a level that would be sufficient to support E2-induced VEGF expression without being stimulatory itself).
It is unknown at this time whether HIF-1 and the ER/Sp1/Sp3 complex, which bind to widely spaced sites on the VEGF promoter, might still interact in some way, perhaps through a looping mechanism. There is evidence that HIF-1 may bind to the glucocorticoid receptor (77) and HIF-1ß to HNF-4 (78), both members of the nuclear receptor family like ER. HIF-1 can also associate with Sp1 (79, 80). Interestingly, HIF-1ß physically associates with ER in cells cotransfected with expression plasmids for both genes and acts as a coactivator of E2-induced reporter gene expression (81). Furthermore, ChIP analysis showed that treatment of T47D cells with E2 caused recruitment of both endogenous ER and HIF-1ß to the pS2 promoter. Studies of the possible direct interaction of HIF-1 and/or -1ß with Sp1 or ER on the VEGF promoter are now underway in our laboratory. HIF-1 and/or -1ß can also directly interact with other transcription factors, including cAMP response element-binding protein/activating transcription factor 1, activator protein 1 (AP-1), activator protein 2 (AP-2), and Smad proteins, all of which have been implicated in the regulation of VEGF expression by other stimuli (44, 47, 48, 82, 83). Our studies do not rule out the involvement of additional transcription factors in the effects of E2 on VEGF expression in the uterus, and different combinations of factors may well be involved in specific cells and under different physiological conditions.
Studies in three different rodent species have shown that the site of increased VEGF expression in the uterus in response to E2 is the luminal epithelium (84, 85, 86). In rats, VEGF expression increases in epithelial cells approximately 15-fold during the proestrus estrogen surge (85), and in no other compartment. Similarly, we have measured a 16-fold increase in VEGF mRNA levels in epithelial cells scraped from the endometrium of immature rats at 2 h after E2 treatment, and a much smaller increase in the residual tissue, which is likely attributable to epithelial cell contamination of that fraction because it is impossible to completely remove all epithelial cells by mechanical means (our unpublished data). All three of the aforementioned studies showed little or no VEGF expression in stromal tissue until after progesterone exposure, and no significant VEGF expression in myometrium at all. The epithelial layer is also the site of sharply increased HIF-1 expression in mice after estrogen exposure (87). Because it is well known that epithelial cells express ERs, and Sp proteins and p300 are ubiquitous, all of the components identified as associating with the VEGF promoter in the uterus after E treatment are present in the epithelial cells. Thus, although the whole uterus was the starting material for these studies, it is highly likely that the events described occurred in the epithelial cell compartment. The decision to use the whole uterus was based on the fact that the ChIP method is only valid if fixation is done rapidly and before there is any significant perturbation of the cells or tissue, e.g. during manipulation to separate cell compartments, which would likely lead to rapid changes in the association of transcription factors with target gene promoters. In retrospect, this approach was especially important with regard to HIF-1, because any attempt to isolate subpopulations of endometrial cells would have exposed them to atmospheric oxygen, which causes rapid degradation of this transcription factor (Refs. 16 and 48 and Fig. 9). In studies underway in our laboratory, we have observed sharp reductions in HIF-1 levels in cancer cells cultured under hypoxic conditions after exposure to room air for periods of less than 1 min. Additional evidence that the use of the whole uterus yields results accurately reflecting changes in distinct cell populations is the fact that the our results for ER and Sp1 on both the VEGF and CKB promoters are consistent with those obtained previously using human cancer cell lines (13, 32).
In addition to rapidly causing existent HIF-1 recruitment to the VEGF promoter, E2 also induces HIF-1 mRNA and protein expression in the uterus. The stimulation of HIF-1 mRNA levels is rare but not unprecedented (46). It also agrees with the observation of Daikoku et al. (87), who reported a 3.5-fold increase in HIF-1 transcripts by 6 h and a peak at 12 h after E2 treatment in the ovariectomized mouse uterus. HIF-1 protein levels, however, were not examined in that study. The apparent lack of correlation between increases in HIF-1 and VEGF mRNAs led those authors to conclude that HIF-1 is not involved in the induction of VEGF by E2. Our ChIP results refute that. Only one other paper has ever examined HIF-1 expression in the uterus (88). In that study on human tissue, only weak, intermittent HIF-1 and -1ß immunohistochemical staining was seen, and no clear cyclic pattern of expression was apparent. This probably reflects the high variability commonly encountered in immunohistochemical studies of human endometrial samples.
In summary, this study implicates HIF-1in the E2-induced expression of VEGF, a factor essential for normal uterine growth and implantation (3, 4), for the first time. Furthermore, a wide range of uterine diseases and disorders, such as endometriosis, leiomyoma, dysmenorrhea, and endometrial cancer, have been linked to increased VEGF expression (4). HIF-1 is already considered a likely target for the treatment of solid tumors (15), and inhibiting HIF-1 action could be an alternative approach for the treatment of E2-dependent diseases of the uterus, breast, and other E2 target tissues. Combinatorial treatments that target both ERs (or E2 production) and HIF-1 might be especially effective for the treatment of E2-dependent cancers.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The ChIP methodology (below) was first optimized and validated using MCF-7 cells, which were provided by Dr. Angela Brodie (Department of Pharmacology and Experimental Therapeutics, University of Maryland School of Medicine). They were maintained and passaged (1:10) in improved MEM (Biosource International, Camarillo, CA) supplemented with 5% fetal bovine serum (HyClone Laboratories, Inc., Logan, UT) and penicillin-streptomycin (100 U-100 µg/ml; Invitrogen, Carlsbad, CA). They were then plated (4 x 106 cells per 150-mm dish) in improved MEM without phenol red (Biosource International) supplemented with 10% charcoal/dextran-treated fetal bovine serum (HyClone), 2 mM glutamine, and penicillin-streptomycin (as above). After 3 d, they were treated with E2 (100 nM) or vehicle. For ChIP, they were fixed in 1% formaldehyde as described by Shang et al. (31).
HEC1A human endometrial carcinoma cells were obtained from the American Type Culture Collection (Manassas, VA) and grown in DMEM/F-12 medium (CellGro; Mediatech, Herndon, VA) supplemented with 5% fetal bovine serum (HyClone), and penicillin-streptomycin (as above). For experiments, cells were plated in phenol red-free DMEM/F-12 medium with 5% charcoal-stripped fetal bovine serum (HyClone). The serum concentration was reduced to 0.5% 48 h before treatment of cells. Hypoxia was induced in an atmosphere of 5% air, 5% CO2, and 90% N2 in a tri-gas cell culture incubator (Precision Three Gas Incubator; GCA Corp., Chicago, IL), resulting in an oxygen concentration of approximately 1%.
RNA Extraction and Reverse Transcription (RT)
RNA was extracted and purified using the RNeasy Mini Kit (QIAGEN). Uterine tissue was homogenized in buffer RLT (QIAGEN) using a Mini Beadbeater and 1.0-mm zirconia/silica beads (BioSpec Products, Bartlesville, OK). The beads were removed, and the samples were further homogenized/sheared by centrifugation through QIAshredder spin columns (QIAGEN). Total RNA concentration and purity were determined from 260 nm and 280 nm absorbances. RNA was diluted with water to 0.08 µg/µl and reverse transcribed. The RT reaction mixture consisted of 6 µl diluted RNA (0.48 µg), 4 µl 5x Reverse Transcriptase buffer, 4 µl deoxynucleotide triphosphate mix (2.5 mM each of deoxy-ATP, deoxy-CTP, deoxy-GTP, deoxy-TTP), 2 µl 0.1 M dithiothreitol, 2 µl of 1 mg/ml BSA, 1 µl of 0.5 µg/µl random primers, and 1 µl (200 U) of Moloney murine leukemia virus reverse transcriptase (all from Invitrogen, San Diego, CA). The mixture was incubated for 1 h at 37 C.
PCR
Analysis of VEGF and HIF-1 mRNA expression was done by both conventional and real-time PCR. For conventional PCR, each 30-µl reaction consisted of 3 µl RT, 3 µl 10x buffer, 2.4 µl deoxynucleotide triphosphate mix (as for RT, above), 3 µl 5 µM primer mix, 0.15 µl Taq DNA polymerase (QIAGEN), and 18.45 µl molecular grade water. The optimal number of PCR cycles (the number of cycles yielding detectable product but still within the linear range of amplification) was first determined for each target mRNA (26 cycles for VEGF and 20 cycles for HIF-1). Total RNA was diluted 1:1000 before amplification of 18S ribosomal RNA (rRNA) for 16 cycles. An annealing temperature of 60 C was used in all cases. The following primers were used for both conventional and real-time PCR analysis:
rat VEGF +12 to +644: forward 5'-GCTCTCTTGGGTGCACTGGA-3' and reverse 5'-CACCGCCTTGGCTTGTCACA-3' (GenBank accession no. NM031836)
rat HIF-1 +681 to +890: forward 5'-TGCTTGGTGCTGATTTGTGA-3' and reverse 5'-GGTCAGATGATCAGAGTCCA-3' (GenBank accession nos. AF057308 and NM024359)
human 18S rRNA +364 to +647: forward 5'-CAACTTTCGATGGTAGTCGC-3' and reverse 5'-CGCTATTGGAGCTGGAATTAC-3' (GenBank accession nos. X01117 and h01593)
Both the VEGF and HIF-1 primer pairs spanned intron-exon borders. PCR products were visualized on 8% polyacrylamide gels, and the relative yield of product per sample was determined by densitometry using a GeneWizard capture and analysis system (Syngene, Cambridge, UK).
Real-time PCR analysis was done using a DNA Opticon system (MJ Research, Boston, MA). Each 30-µl reaction mixture included 3 µl cDNA, 15 µl DyNAmo SYBR green qPCR mix (MJ Research), 1.2 µl primer mix (5 µM each primer), and 10.8 µl molecular grade water. Each sample was assayed in duplicate. A standards curve was generated by serially diluting uterine cDNA. The yield of product for each unknown sample was calculated by applying its threshold cycle, or C(T), value (the cycle at which the sample’s fluorescence trace exceeds background noise and begins to increase linearly) to the standard curve using the Opticon Monitor analysis software (version 1.01, MJ Research). Values were normalized to corresponding 18S rRNA values and expressed as the fold increase relative to 0 h.
In Vivo ChIP
Uterine horns were placed in 1 ml cold Dulbecco’s PBS containing protease inhibitors (Complete Mini EDTA-free Protease Inhibitor Cocktail, 1 tablet/10 ml; no. 1 836 170; Roche Applied Science, Indianapolis, IN) and phosphatase inhibitors (cocktails I and II, 1:100 each; Sigma Chemical Co., St. Louis, MO) briefly before homogenizing in radioimmunoprecipitation buffer [50 mM Tris-HCl (pH 7.4); 1% Nonidet P-40; 0.25% deoxycholic acid; 1 mM EDTA; and protease and phosphatase inhibitors, as described above]. Homogenates were centrifuged at 12,000 rpm for 5 min at 4 C, and the supernatants were removed and discarded. At this point, the ChIP protocol described by Shang et al. (31) was followed with minor modifications. Briefly, pellets were resuspended in 600 µl lysis buffer [50 mM Tris-HCl (pH 8.1); 5 mM EDTA; 1% sodium dodecyl sulfate; and protease inhibitors (as above)] and incubated on ice for 15 min. Samples were sonicated on ice for 10 x 10 sec cycles, with 20-sec pauses between each cycle, using a Microson Ultrasonic Cell Disruptor (Misonix, Farmingdale, NY) at power level 2.5. Sonicated samples were then divided into 100-µl aliquots and stored at –80 C.
Sonicated sample aliquots were thawed on ice and diluted 1:1000 with dilution buffer [20 mM Tris-HCl (pH. 8.1), 150 mM NaCl, 2 mM EDTA, 1% triton X-100, and protease inhibitors (as above)] before being immunocleared in a solution containing 45 µl of a 50% slurry of either Protein A or Protein G Sepharose 4 Fast Flow (Amersham Biosciences, Piscataway, NJ) in Tris-EDTA (TE) buffer, pH 8, 2 µg salmon sperm DNA (Invitrogen), and 20 µl normal serum (Sigma) for 2 h at 4 C; use of protein A or protein G was determined based on which had stronger affinity for each particular antibody (based on supplier’s recommendation). Supernatants were collected and incubated overnight at 4 C with one of the following antibodies: 1 µg mouse monoclonal human ER antibody (Ab-10; NeoMarkers/Lab Vision, Fremont, CA), 2 µg mouse monoclonal human HIF-1 antibody (BD Biosciences, Palo Alto, CA), 5 µg mouse monoclonal human HIF-1ß antibody (BD Biosciences), 5 µg anti-p300 (N-15) rabbit polyclonal antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), 4 µg rabbit polyclonal Sp1 (H-225 or PEP 2) antibody (no. sc-14027 or sc-59; Santa Cruz Biotechnology), and 5 µg rabbit polyclonal Sp3 (H-225) antibody (no. sc-13018; Santa Cruz Biotechnology). For the "no antibody" control, an equal volume of normal mouse or rabbit serum was substituted for the specific mouse or rabbit antibody. Protein A or protein G sepharose beads (45 µl of a 50% slurry in TE buffer) and salmon sperm DNA (2 µg) were then added and incubated for 1 h at 4 C. The beads were then washed sequentially with 1 ml each of TSE I, TSE II, and buffer III (31), and then 2 x 1 ml of TE buffer (pH 8.0); each wash was for 10 min each at 4 C. The protein-DNA complexes were then eluted by twice incubating beads in 100 µl of elution buffer (1% sodium dodecyl sulfate, 0.1 M NaHCO3) for 10 min at room temperature with vigorous mixing. To separate immunoprecipitated protein and DNA, the pooled eluates were incubated at 65 C overnight. The DNA was purified using the Qiaquick PCR Purification kit (QIAGEN). The final volume was 50 µl (10 mM Tris-HCl, pH 8.5).
The yield of target region DNA in each sample after ChIP was analyzed by both conventional and real-time PCR, as described previously. In both cases, 3 µl of each 50-µl sample was amplified. For real-time PCR, standard curves were generated by serially diluting an input chromatin sample. The following primers were used for ChIP PCR analysis (30 cycles at 60 C annealing temperature):
rat VEGF –944 to –611: forward 5'-TCTGCCAGACTCCACAGTG-3' and reverse 5'-TGCGTGTTTCTAACACCCAC-3' (GenBank accession no. U22373)
rat VEGF –661 to –384: forward 5'-GTTTCCGAGGTCAAACAAGC-3' and reverse 5'-CACACTATACCCAGACACAC-3' (GenBank accession no. U22373)
rat VEGF –403 to –124: forward 5'-GTGTGTCTGGGTATAGTGTG-3' and reverse 5'-GCCACTACTGCGAAATAGAAA-3' (GenBank accession no. U22373)
rat VEGF –173 to +114: forward 5'-CAGGCTATGGACCCTGGTAA-3' and reverse 5'-ATAGTCTGCCTTGTCGCTGC-3' (GenBank accession no. U22373)
rat VEGF +666 to +928: forward 5'-TGAGTCAAGAGGACAGAGAG-3'and reverse 5'-ATTACCAGGCCTCTTCTTCC-3' (GenBank accession no. U22372)
rat CKB –678 to –319: forward 5'-GGAAAGAACCTGGGGATTTG-3'and reverse 5'-GTTAGCACTTGAGGTTCCTG-3' (GenBank accession no. M18668)
human pS2 –436 to +19: forward 5'-GGCCATCTCTCACTATGAATC-3'and reverse 5'-GGCAGGCTCTGTTTGCTTAAA-3' (GenBank accession no. X05030)
Western Blot Analysis
Rat uterine tissue was homogenized on ice in 200 µl radioimmunoprecipitation buffer (as described above) using a Tissue-Tearor rotor-stator homogenizer (Biospec Products) at power setting 1.5. The homogenate was then centrifuged at 3000 rpm for 30 min at 4 C, and the supernatant was collected, aliquoted, and stored at –80 C. The protein concentration was determined using the BCA Protein Assay (Pierce Chemical Co., Rockford, IL). Protein samples (20–40 µg depending on antibody used) were loaded onto either a 7.5% (for HIF-1) or 10% (for HIF-1ß) SDS-PAGE gel and transferred overnight onto polyvinylidene fluoride membrane. Membranes were then blocked with 5% nonfat dry milk/1x TTBS (1x Tris-buffered saline with 0.1% Tween 20) for 1 h before being incubated for 2.5 h at room temperature with mouse monoclonal antibodies to either HIF-1 [no. 610958 (BD Biosciences) at 1:250; or NB100–105 (Novus Biologicals, Littleton, CO) at 1:10,000] or HIF-1ß (no. 611078, BD Biosciences at 1:1,000). After three 5-min washes in 1x TTBS, the membranes were incubated at room temperature for 30 min with 1:2000 goat antimouse IgG horseradish peroxidase-conjugated antibody (Upstate, Charlottesville, VA). The membranes were then washed three times for 5 min in 1x TTBS and two times for 5 min in 1x Tris-buffered saline at room temperature. Western Lightning Chemiluminescence Reagent Plus (PerkinElmer Life and Analytical Sciences, Boston, MA) was used for visualization of protein bands, according to the manufacturer’s instructions.
Statistical Analysis
Statistical analyses were done using factorial ANOVA and appropriate post hoc tests (StatView Version 4.5, Abacus Concepts, Berkeley, CA).
|
FOOTNOTES |
|---|
First Published Online March 17, 2005
Abbreviations: ChIP, Chromatin immunoprecipitation; CKB, creatine kinase B; E2, 17ß-estradiol; ER, estrogen receptor ; ERE, estrogen response element; HIF-1, hypoxia-inducible factor 1; HRE, hypoxia response element; iNOS, inducible nitric oxide synthase; PI3K, phosphatidylinositol 3-kinase; RT, reverse transcription; TE, Tris-EDTA; UTR, untranslated region; VEGF, vascular endothelial growth factor.
Received for publication September 29, 2004. Accepted for publication March 8, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and ß. Proc Natl Acad Sci USA 97:10972–10977 and ERß at AP1 sites. Science 277:1508–1510 and Sp3 proteins. J Biol Chem 275:22769–22779 and SP proteins. Oncogene 23:1052–1063[CrossRef][Medline]
targeted for VHL-mediated destruction by proline hydroxylation: implications for O2 sensing. Science 292:464–468 to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation. Science 292:468–472/ARNT. EMBO J 17:5085–5094[CrossRef][Medline]
in vascular smooth muscle cells. J Biol Chem 275:26765–26771 in ovarian carcinoma cells. J Biol Chem 277:27850–27855 by interleukin-1ß involves the extracellular signal-regulated kinase 1/2 pathway in normal human cytotrophoblast cells. Biol Reprod 70:1822–1827 and vascular endothelial growth factor A through phosphatidylinositol 3-kinase/Akt/FRAP pathway in OVCAR-3 and A2780-CP70 human ovarian carcinoma cells. Toxicol Appl Pharmacol 196:124–135[CrossRef][Medline]
-induced transcriptional activation of the vascular permeability factor (VPF/VEGF) gene requires AP-2-dependent DNA binding and transactivation. EMBO J 16:750–759[CrossRef][Medline]
in human glioma cells. Possible roles of SP-1. J Biol Chem 271:28220–28228 by transcriptional and translational mechanisms. J Biol Chem 277:48403–48409 and ß. Cancer Res 62:4977–4984 (ER) via interaction between ER and PI3K. Cancer Res 61:5985–5991 with phosphatidylinositol 3-OH kinase. Steroids 67:935–939[CrossRef][Medline]
mediates rapid signaling by estrogen. J Neurosci Res 74:1–11[CrossRef][Medline]
(HIF-1) synthesis: novel mechanism for HIF-1-mediated vascular endothelial growth factor expression. Mol Cell Biol 21:3995–40044 gene by an interaction between hypoxia-inducible factor-1 and Sp1. J Biol Chem 279:15025–15031NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  K. H. Molitoris, A. A. Kazi, and R. D. Koos Inhibition of Oxygen-Induced Hypoxia-Inducible Factor-1{alpha} Degradation Unmasks Estradiol Induction of Vascular Endothelial Growth Factor Expression in ECC-1 Cancer Cells in Vitro Endocrinology, December 1, 2009; 150(12): 5405 - 5414. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Dadiani, D. Seger, T. Kreizman, D. Badikhi, R. Margalit, R. Eilam, and H. Degani Estrogen regulation of vascular endothelial growth factor in breast cancer in vitro and in vivo: the role of estrogen receptor {alpha} and c-Myc Endocr. Relat. Cancer, September 1, 2009; 16(3): 819 - 834. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. A. Kazi, K. H. Molitoris, and R. D. Koos Estrogen Rapidly Activates the PI3K/AKT Pathway and Hypoxia-Inducible Factor 1 and Induces Vascular Endothelial Growth Factor A Expression in Luminal Epithelial Cells of the Rat Uterus Biol Reprod, August 1, 2009; 81(2): 378 - 387. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Wang, J. Tan, A. Coffey, J. Fehrenbacher, B. R. Weil, and D. R. Meldrum Signal transducer and activator of transcription 3-stimulated hypoxia inducible factor-1{alpha} mediates estrogen receptor-{alpha}-induced mesenchymal stem cell vascular endothelial growth factor production J. Thorac. Cardiovasc. Surg., July 1, 2009; 138(1): 163 - 171. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Sabnis, A. Schayowitz, O. Goloubeva, L. Macedo, and A. Brodie Trastuzumab Reverses Letrozole Resistance and Amplifies the Sensitivity of Breast Cancer Cells to Estrogen Cancer Res., February 15, 2009; 69(4): 1416 - 1428. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. N. Taylor, Jie Yu, P. B. Torres, A. C. Schickedanz, J. K. Park, M. D. Mueller, and N. Sidell Mechanistic and Therapeutic Implications of Angiogenesis in Endometriosis Reproductive Sciences, February 1, 2009; 16(2): 140 - 146. [Abstract] [PDF]
|
|
|
|
 |
|
 S. P. Yun, M. Y. Lee, J. M. Ryu, C. H. Song, and H. J. Han Role of HIF-1{alpha} and VEGF in human mesenchymal stem cell proliferation by 17{beta}-estradiol: involvement of PKC, PI3K/Akt, and MAPKs Am J Physiol Cell Physiol, February 1, 2009; 296(2): C317 - C326. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. C. Zachos, M. G. Burch, R. B. Billiar, C. Li, E. D. Albrecht, and G. J. Pepe Regulation of Expression of Microvillus Membrane Proteins by Estrogen in Baboon Fetal Ovarian Oocytes Biol Reprod, December 1, 2008; 79(6): 1160 - 1168. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. W. Bonagura, G. J. Pepe, A. C. Enders, and E. D. Albrecht Suppression of Extravillous Trophoblast Vascular Endothelial Growth Factor Expression and Uterine Spiral Artery Invasion by Estrogen during Early Baboon Pregnancy Endocrinology, October 1, 2008; 149(10): 5078 - 5087. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Buteau-Lozano, G. Velasco, M. Cristofari, P. Balaguer, and M. Perrot-Applanat Xenoestrogens modulate vascular endothelial growth factor secretion in breast cancer cells through an estrogen receptor-dependent mechanism J. Endocrinol., February 1, 2008; 196(2): 399 - 412. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. L. Young Complexity in Endometrial Estradiol Signaling Reproductive Sciences, October 1, 2007; 14(7): 627 - 628. [PDF]
|
|
|
|
 |
|
 V L Bosquiazzo, J G Ramos, J Varayoud, M Munoz-de-Toro, and E H Luque Mast cell degranulation in rat uterine cervix during pregnancy correlates with expression of vascular endothelial growth factor mRNA and angiogenesis Reproduction, May 1, 2007; 133(5): 1045 - 1055. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. A. Kazi and R. D. Koos Estrogen-Induced Activation of Hypoxia-Inducible Factor-1{alpha}, Vascular Endothelial Growth Factor Expression, and Edema in the Uterus Are Mediated by the Phosphatidylinositol 3-Kinase/Akt Pathway Endocrinology, May 1, 2007; 148(5): 2363 - 2374. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Hartman, K. Lindberg, A. Morani, J. Inzunza, A. Strom, and J.-A. Gustafsson Estrogen Receptor {beta} Inhibits Angiogenesis and Growth of T47D Breast Cancer Xenografts Cancer Res., December 1, 2006; 66(23): 11207 - 11213. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. O'Brien, T. J. Peterson, M. H. Tong, E.-J. Lee, L. E. Pfaff, S. C. Hewitt, K. S. Korach, J. Weiss, and J. L. Jameson Estrogen-induced Proliferation of Uterine Epithelial Cells Is Independent of Estrogen Receptor {alpha} Binding to Classical Estrogen Response Elements J. Biol. Chem., September 8, 2006; 281(36): 26683 - 26692. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S M Hyder Sex-steroid regulation of vascular endothelial growth factor in breast cancer. Endocr. Relat. Cancer, September 1, 2006; 13(3): 667 - 687. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. D.-Y. Lin, A. Mannikarottu, B. A Kogan, C. Whitbeck, P. Chichester, R. E Leggett, and R. M Levin Estrogen induces angiogenesis of the female rabbit bladder. J. Endocrinol., August 1, 2006; 190(2): 241 - 246. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Punyadeera, V.L. Thijssen, S. Tchaikovski, R. Kamps, B. Delvoux, G.A.J. Dunselman, A.F.P.M. de Goeij, A.W. Griffioen, and P.G. Groothuis Expression and regulation of vascular endothelial growth factor ligands and receptors during menstruation and post-menstrual repair of human endometrium Mol. Hum. Reprod., June 1, 2006; 12(6): 367 - 375. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
