Estrogen regulation of vascular endothelial growth factor ... - Nature

Oncogene (2004) 23, 1052–1063

& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00 www.nature.com/onc

Estrogen regulation of vascular endothelial growth factor gene expression in ZR-75 breast cancer cells through interaction of estrogen receptor a and SP proteins Matthew Stoner1, Mark Wormke1, Brad Saville2, Ismael Samudio1, Chunhua Qin1, Maen Abdelrahim1 and Stephen Safe*,1 1 2

Department of Veterinary Physiology and Pharmacology, Texas A&M University, College Station, TX 77843-4466, USA; Department of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843-4466, USA

Vascular endothelial growth factor (VEGF) is expressed in multiple hormone-dependent cancer cells/tumors. Treatment of ZR-75 breast cancer cells with 17b-estradiol (E2) induced a greater than fourfold increase of VEGF mRNA levels. ZR-75 breast cancer cells were transfected with pVEGF1, a construct containing a 2018 to þ 50 VEGF promoter insert, and E2 induced reporter gene (luciferase) activity. Deletion and mutation analysis of the VEGF gene promoter identified a GC-rich region (66 to 47) which was required for E2-induced transactivation of pVEGF5, a construct containing the minimal promoter (66 to þ 54) that exhibited E2-responsiveness. Interactions of nuclear proteins from ZR-75 cells with the proximal GC-rich region of the VEGF gene promoter were investigated by electrophoretic mobility shift and chromatin immunoprecipitation assays. The results demonstrate that both Sp1 and Sp3 proteins bound the GCrich motif (66 to 47), and estrogen receptor a (ERa) interactions were confirmed by chromatin immunoprecipitation. Moreover, E2-dependent activation of constructs containing proximal and distal GC/GT-rich regions of the VEGF promoter was inhibited in ZR-75 cells transfected with small inhibitory RNAs for Sp1 and Sp3. These results were consistent with a mechanism of hormone activation of VEGF through ERa/Sp1 and ERa/Sp3 interactions with GC-rich motifs. Oncogene (2004) 23, 1052–1063. doi:10.1038/sj.onc.1207201 Published online 1 December 2003 Keywords: estrogen; VEGF; transactivation; ERa/Sp1; ERa/Sp3

Introduction Vascular endothelial growth factor (VEGF) or vascular permeability factor is an angiogenic growth factor that plays a critical role in neovascularization in normal and neoplastic tissues (Hanahan and Folkman, 1996; Fer*Correspondence: Stephen Safe; E-mail: [email protected] Received 3 June 2003; revised 30 July 2003; accepted 8 September 2003

rara and Davis-Smyth, 1997; Zetter, 1998). Antiangiogenic therapies for treatment of multiple tumor types have been developed to block one or more steps in angiogenesis, and these include antibodies directed against VEGF or the VEGF receptors (VEGFR) Flt-1 (VEGFR1) and KDR/Flk-1 (VEGFR2) (Schlaeppi and Wood, 1999; Bae et al., 2000; Brekken et al., 2000). VEGF is expressed in multiple tumors, and protein and/ or mRNA levels serve as prognostic factors. Several reports have demonstrated correlations between expression of VEGF alone or in combination with other factors with the histopathological grade of cancer and prognosis for disease-free survival (Brown et al., 1993; Boocock et al., 1995; Viglietto et al., 1995). For example, in breast cancer patients, increased expressions of VEGF and mutant p53 tumor suppressor gene were correlated with an unfavorable outcome (Linderholm et al., 2001). VEGF is expressed as multiple forms including isoforms of 121, 145, 165, 189 and 206 amino acids which interact with the tyrosine kinase receptors, Flt-1 or KDR/Flk-1. VEGF expression is induced by hypoxia and also by a variety of stimuli including cytokines, NO, cobaltous ion, and mitogenic growth factors (Pepper et al., 1992; Goldman et al., 1993; Gille et al., 1996; Ryuto et al., 1996; Enholm et al., 1997; Laitinen et al., 1997; Pal et al., 1998; Kimura et al., 2001). Transcriptional upregulation of VEGF in cells/tissues grown under hypoxic conditions has been extensively characterized. Hypoxia inducible factor 1a (HIF1a) is induced in cells grown under hypoxia, and subsequent interactions of the HIF1a/HIF1b heterodimer with hypoxic response elements (HREs) in the VEGF promoter result in increased gene expression (Kimura et al., 2001). A recent study reported that HIF1a protein was expressed in early stages of breast cancer and increased levels were associated with more aggressive mammary tumors (Bos et al., 2001). Interestingly, there was also a correlation between increased HIF1a, VEGF and ER levels and this was associated with a higher proliferation rate of tumors. VEGF and the estrogen receptor (ER) are coexpressed in many breast tumors and several studies demonstrate that estrogen and other steroid hormones modulate VEGF expression in multiple tissues

Estrogen induces VEGF expression in ZR-75 cells M Stoner et al

1053

(Cullinan-Bove and Koos, 1993; Hyder et al., 1996, 1997, 1998, 2000; Nakamura et al., 1996; Yoshiji et al., 1997; Ruohola et al., 1999; Classen-Linke et al., 2000; Mueller et al., 2000; Stoner et al., 2000; Bogin and Degani, 2002; Buteau-Lozano et al., 2002). For example, 17b-estradiol (E2) and antiestrogens induce VEGF expression in the rodent uterus, and there is also evidence that E2 induces VEGF in 7,12-dimethylbenz[a]anthracene-induced mammary tumors in female Sprague–Dawley rats (Yoshiji et al., 1997). In breast and endometrial cancer cells, hormonal regulation appears to be more complex and dependent on cell context. T47D breast cancer cells express high levels of the progesterone receptor as well as other steroid hormone receptors, and progestins, but not estrogens, androgens, or glucocorticoids, induced expression of VEGF protein (Hyder et al., 1998). In contrast, progestins did not induce VEGF in other ER-positive breast (MCF-7 and ZR-75) or Ishikawa endometrial cancer cells or in ER-negative MDA-MB-231 cells (Hyder et al., 1998). Recent studies showed that E2 induced VEGF gene expression in Ishikawa cells (Mueller et al., 2000) and in MCF-7 breast cancer cells (Ruohola et al., 1999; Buteau-Lozano et al., 2002), whereas E2 decreased gene expression in HEC1A endometrial cancer cell lines that are also ER-positive (Stoner et al., 2000). This study reports that E2 induces VEGF gene expression in ZR-75 cells, and molecular analysis of this response has identified a proximal GCrich region (66 to 47) in the VEGF promoter that interacts with both ERa/Sp1 and ERa/Sp3 and is required for transactivation. The role of Sp3 in mediating the estrogenic response contrasts to studies in Ishikawa and HEC1A1 cells where in the latter cell line, ERa/Sp3 was required for E2-dependent decreased gene expression. Thus, hormone regulation of VEGF in E2-responsive breast and endometrial cancer cells is complex and strongly influenced by cell context.

Results Hormonal regulation of VEGF gene expression in ZR-75 cells Studies in ZR-75 breast cancer cells showed that VEGF gene expression was induced by E2. Treatment of ZR-75 cells with 10 nM E2 significantly induced VEGF mRNA levels 1, 6 and 12 h after treatment, and transcript levels decreased to background levels after 24 h (Figure 1a). pVEGF1 contains the 2018 to þ 50 region of the VEGF gene promoter linked to the bacterial luciferase gene (in pCDNA3.1) and 10 nM E2 did not induce luciferase activity in ZR-75 cells transfected with pVEGF1 alone. However, cotransfection with ERa

Figure 1 Hormonal activation of VEGF in ZR-75 cells. (a) Transcriptional activation of VEGF by E2. Cells were seeded in 2.5% charcoal-stripped serum-containing medium for 1 day, cultured under serum-free conditions for 3 days, then treated with DMSO (vehicle) or 10 nM 17b-estradiol (E2) for the indicated times. VEGF mRNA levels were determined by RT–PCR. Results are expressed as mean7s.e. (*, Po0.05, n ¼ 3). (b) Effects of hERa titration and E2 treatment on VEGF gene promoter-luciferase reporter activity in ZR-75 cells. Cells were transiently transfected with 500 ng pVEGF1, 0–500 ng hERa expression plasmid and 500 ng b-galactosidase expression vector, then treated for 48 h with DMSO or 10 nM E2. Luciferase activity was normalized to bgalactosidase activity. Results are expressed as mean7s.e. and significant induction by E2 is indicated (*, Po0.05, n ¼ 3). (c) Effects of antiestrogens on E2-induced luciferase activity in ZR-75 cells. Cells were transiently transfected with 500 ng of pVEGF1, 500 ng hERa and 500 ng b-galactosidase expression vector. Cells were then treated with DMSO and 10 nM E2, or 1 mM ICI 182,780 (ICI), 1 mM tamoxifen (TAM) or 4OH-tamoxifen (4OH-TAM), with or without 10 nM E2, for 48 h. As a negative control, cotransfection of 500 ng human progesterone receptor B-form (hPR-B) and treatment with 10 nM progesterone (P) did not alter luciferase activity. Results expressed as mean7s.e. and significant induction by E2 is indicated (*, Po 0.05, n ¼ 3) Oncogene


1054

expression plasmid resulted in E2-responsiveness (Figure 1b), and similar results have been reported by several laboratories for many other E2-responsive gene promoters containing diverse response elements (Berry et al., 1989; Cavailles et al., 1989, 1993; Weisz and Rosales, 1990; Augereau et al., 1994; Gillesby et al., 1997; Paech et al., 1997; Sathya et al., 1997; Safe, 2001). The requirement for cotransfection of ERa is due to overexpression of the promoter–reporter construct in the transfected cells which results in limiting levels of ERa (Augereau et al., 1994). The results in Figure 1c show that E2 induced activity in ZR-75 cells transfected with pVEGF1, and the antiestrogens ICI 182,780 and 40 hydroxytamoxifen (but not tamoxifen) inhibited hormone-induced transactivation. In contrast, progesterone did not induce luciferase activity in ZR-75 cells transfected with pVEGF1 and PR-B expression plasmid showing that this response was hormone receptorspecific. Deletion analysis of the VEGF gene promoter The results in Figure 2a and b summarize the E2induced transactivation observed in transfection studies with constructs containing a series of 5-deletions. In ZR75 cells treated with DMSO and transfected with pVEGF1, pVEGF2 and pVEGF3, there was a progressive loss of basal activity associated with stepwise deletion of the 2018 to 131 region on the VEGF gene promoter; however, luciferase activity was highly induced (10- to 14-fold) after treatment with E2. A comparison of results obtained using pVEGF3 and pVEGF5 suggests that the 267 to 132 region of the VEGF promoter may contribute to the hormoneresponsiveness of VEGF. Further 50 -deletion analysis of the VEGF promoter using PVEGF5, pVEGF6 and pVEGF7 indicated that loss of E2-responsiveness was observed with stepwise deletion of the GC-rich 131 to 52 region of the promoter; pVEGF6 contained the minimal promoter insert (66 to þ 54), and inducibility by E2 was 3.0–5.5-fold over several experiments. Deletion of the GC-rich 66 to 52 region (to give pVEGF7) resulted in loss of hormone-induced transactivation. Moreover, a construct containing the 66 to 47 VEGF promoter insert (pVEGF8) was E2-responsive (Figure 2c) and mutation of the GC-boxes within this sequence (pVEGF8m) resulted in loss of hormoneinduced activity. The results suggest that the GC-rich 131 to 52 region of the VEGF promoter exhibited maximal E2-responsiveness. Role of Sp1 family proteins in hormone activation of VEGF E2 induces luciferase activity in ZR-75 cells transfected with pVEGF5 and wild-type (wt) ERa/Sp1 and transactivation is typically mediated by wt ERa or HE11C (a DBD deletion mutant) but not by HE15C or HE19C which contain deletions of activation function 2 (AF2) and AF1, respectively. Similar results have previously been observed in cells transfected with other E2Oncogene

Figure 2 Deletion and mutation analysis. (a) Deletion analysis of the VEGF promoter-distal region. ZR-75 cells were transiently transfected with 500 ng of pVEGF constructs, 500 ng of hERa and 500 ng b-galactosidase expression plasmid, treated with DMSO or 10 nM E2 for 48 h, and luciferase activity determined as described in the Materials and methods. Results are expressed as mean7s.e. and significant induction by E2 is indicated (*, Po 0.05, n ¼ 3). Empty vector pGL2 was not affected by E2 treatment. (b) Deletion analysis of the VEGF promoter-proximal region. ZR-75 cells were treated with DMSO or 10 nM E2, transfected with constructs, and luciferase activity determined as described in (a). Results are expressed as mean7s.e. and significant induction by E2 is indicated (*, Po0.05, n ¼ 3). (c) Mutational analysis of pVEGF6. ZR-75 cells were transiently transfected with 500 ng of pVEGF6, pVEGF7 (GC-rich sequence deleted), pVEGF8 or pVEGF8m and ERa, and luciferase activity was determined as described above. Results are expressed as mean7s.e. and significant induction by E2 is indicated (*, Po0.05, n ¼ 3)

responsive GC-rich promoters (Safe, 2001) indicating that the DBD of ERa is not required for ERa/Sp1mediated transactivation. Direct interactions of nuclear


1055

extracts (E2-induced) from ZR-75 cells with [32P]66/ 47 (66 to 47 VEGF oligonucleotide) were determined in gel-mobility shift assays (Figure 3b). Three major retarded bands were observed [Sp1–DNA (1) and Sp3–DNA (2)] (lane 2). Retarded band intensities were decreased after coincubation with 100-fold excess of unlabeled wt 66/47 (lane 3) or consensus Sp1 (lane 5) oligonucleotides, whereas intensities were unchanged after coincubation with unlabeled mutant 66/47 mutant Sp1 or ERE oligonucleotides (lanes 4, 6 and 7,

respectively). The retarded bands were supershifted after incubation with Sp1 or Sp3 antibodies (lanes 8 and 10, respectively), whereas incubation with nonspecific IgG did not supershift the retarded bands (lanes 9 and 11). Sp3 antibodies gave only a weak supershifted complex; however, immunodepletion of both Sp3–DNA bands was apparent (lane 8). Nuclear extracts from ZR-75 cells bound [32P]66/47 to give less mobile Sp1/Sp3 and more mobile Sp3 bands (lane 2) as indicated with arrows (Figure 3c). Coincubation of nuclear extracts with ERa

Figure 3 ERa-Sp protein interactions. (a) Effects of cotransfected hERa variants on E2-induced pVEGF5. Cells were transiently transfected with 500 ng of pVEGF5, 500 ng of indicated hERa wt or variant (HE11C, HE15C and HE19C) and 500 ng of bgalactosidase expression plasmid. Treatments were for 48 h with DMSO (vehicle) or 10 nM E2, and luciferase activity determined as described in the Materials and methods. Results are expressed as mean7s.e. and significant induction by E2 is indicated (*, Po0.05, n ¼ 3). (b) Binding of nuclear extracts from ZR-75 cells to [32P]66/47 (50 -GTCCCGGCGGGGCGGAGCCATG-30 ) (lanes 1–11). ZR-75 cells were treated with 10 nM E2, nuclear extracts (NE) were obtained and EMSA performed as described in Materials and methods. Sp1 and Sp3 binding to [32P]66/47 yielded specific complexes of at least three retarded bands (lane 2). Coincubation with 100-fold molar excess of unlabeled wt 66/47 and wt Sp1 oligonucleotides abolished shifted bands (lanes 3 and 5), while excess mutant (mut) 66/47 and mut Sp1 oligonucleotides did not affect band intensities (lanes 4 and 6). Competition with unlabeled consensus ERE oligonucleotide did not decrease specific Sp1/Sp3 : DNA band intensities (lane 7). Antibodies against Sp1 and Sp3 proteins resulted in specifically supershifted Sp1 : [32P]66/47 or Sp3 : [32P]66/47 complexes (lanes 8 and 10). Bands were not supershifted by normal rabbit and normal goat IgG (lanes 9 and 11). (c) hERa enhanced binding of nuclear extracts containing Sp1 and Sp3 to [32P]66/47. Nuclear extracts containing Sp1 and Sp3 proteins bound to [32P]66/47 (lanes 2 and 6) and coincubation with increasing amounts of hERa protein enhanced retarded band intensity (lanes 3–5). hERa protein did not bind directly to [32P]66/ 47 (lane 9) but bound [32P]ERE (lane 10) and the band was supershifted with ERa antibodies (lane 11); antibodies against Sp1 and Sp3 proteins supershifted the Sp1 : DNA and Sp3 : DNA complexes (lanes 6 and 7). Intensities of the higher molecular weight Sp1/Sp3– DNA complex after treatment with ERa were quantitated (relative to no ERa addition) and presented as means7s.e. for three determinations. Significantly (Po0.05) increased binding is indicated with an asterisk Oncogene


1056

did not form a supershifted ERa/Sp1 band (lanes 3–5), but there was a greater than threefold enhanced intensity of the Sp1–DNA complex using 4 ml of in vitro translated ERa (lane 5) compared to the band intensity using unprogrammed lysate (UPL). Similar observations on ERa/Sp1–DNA interactions have previously been reported (Porter et al., 1997; Duan et al., 1998), and enhanced binding was due to an ER-dependent increase in the ‘on-rate’ of Sp1 binding to GC-rich oligonucleotides. Sp1 antibodies (lane 6) supershifted the complex, whereas Sp3 antibodies immunodepleted bands assigned to Sp3 (lane 7). In vitro translated ERa did not directly bind [32P]66/47 (lane 8) but bound [32P]ERE to give a specifically-bound complex (lane 10) that was supershifted with ERa antibodies. These data are consistent with interactions of Sp1/Sp3 with the GCrich proximal region of the VEGF promoter where ERa enhances binding but does not give a supershifted complex. Previous studies in HEC1A endometrial cancer cells showed that E2 decreased VEGF gene expression, and the results were consistent with a mechanism in which ERa/Sp3 interacted with proximal GC-rich sites (Stoner et al., 2000). Moreover, treatment of HEC1A cells with E2 caused a significant increase in relative Sp3 protein levels (B40%), whereas no change in immunoreactive Sp1 protein was observed (Stoner et al., 2000). The results in Figure 4a illustrate that levels of immunoreactive Sp1 and Sp3 were unchanged in ZR-75 cells treated with 10 nM E2 for 3, 5, 8 or 26 h. Interactions of ERa, Sp1 and Sp3 with the VEGF promoter were also investigated in the ChIP assay using antibodies for ERa, Sp1, Sp3 and brahma related gene-1 (BRG-1) to determine protein binding to the proximal region of the VEGF gene promoter (Figure 4b). PCR products were confirmed by Southern blot analysis using a radiolabeled PCR product with primers that amplify the 199 to þ 3 region of the VEGF gene promoter to ensure that only the designated VEGF promoter region was detected (data not shown). ERa, Sp1, Sp3 and Brg-1 were bound to the promoter in untreated cells (0 time), and all proteins were bound to the VEGF promoter for up to 120 min after treatment with 10 nM E2. These observations for ERa, Sp1 and Sp3 were consistent with results of deletion analysis and gel-mobility shift assays. As a control experiment, we used the ChIP assay to determine interaction of ERa to a 150 bp region of exon II in the cathepsin D gene which does not contain ERa or Sp1 (GC-rich) binding sites. PCR products were not observed using two ERa antibodies confirming the specificity of the ChIP assay and the interactions of ERa, Sp1 an Sp3 with the VEGF gene promoter. Results showing coexpression of Sp1 and Sp3 in ZR75 cells coupled with their binding to proximal GC-rich sequences in the VEGF promoter is consistent with a role for ERa/Sp1 and/or ERa/Sp3 in mediating E2induced expression of VEGF. Figure 5a and b summarizes the effects of transfecting siRNAs for LMN A/C (iLMN), Sp1 (iSp1) or Sp3 (iSp3) on Sp1 and Sp3 protein levels in whole-cell lysates. Levels of Sp1 and Sp3 were significantly decreased by 450% Oncogene

after transfection of their specific siRNAs compared to that observed in cells transfected with small inhibitory RNA for lamin A/C (iLMN) (nonspecific). These results are comparable to previous studies using iSp1 in ZR-75 cells (Abdelrahim et al., 2002). ZR-75 cells were also cotransfected with pVEGF1 or pVEGF6 and iGL2 (siRNA for the luciferase reporter gene), iSp1 or iSp3, and treated with DMSO or 10 nM E2 (Figure 5c–e). E2induced luciferase activity in ZR-75 cells transfected with iLMN and pVEGF1 was decreased by 90, 51 and 59% after transfection with iGL2, iSp1 and iSp3, respectively (Figure 5c), and cotransfection with iSp1 plus iSp3 further decreased hormone-induced activity (Figure 5d). Moreover, in cells transfected with pVEGF6 and iLMN, induced activity was decreased by 82, 46 and 67% after cotransfection with iGL2, iSp1 and iSp3, respectively (Figure 5e). In contrast, the siRNA for Sp proteins did not affect hormonedependent activation of an ERE promoter construct (data not shown) as previously reported (Abdelrahim et al., 2002). These results demonstrate that hormonal activation of distal and proximal GC-rich sites in the VEGF gene promoter are dependent on both ERa/Sp1 and ERa/Sp3.

Discussion Hormonal regulation of VEGF has been reported in multiple tissues and increased or decreased expressions are hormone- and cell context-dependent. Progestins both decreased and increased VEGF expression in fibroblasts and epithelial cells, respectively, derived from human endometrial tissue samples (Classen-Linke et al., 2000). Progestins induced VEGF in T47D cells but not in MCF-7, ZR-75 and MDA-MB-231 breast cancer cells or Ishikawa endometrial cancer cells (Hyder et al., 1998). Estrogens and antiestrogens induced VEGF expression in the ovariectomized rat uterus (Hyder et al., 1997); however, the effects of estrogens on VEGF in breast and endometrial cancer cell lines were highly variable. Estrogen decreased VEGF mRNA levels in HEC1A cells (Stoner et al., 2000), increased expression in Ishikawa cells (Mueller et al., 2000), whereas VEGF levels were unchanged in T47D cells after treatment with E2 (Hyder et al., 1998). One study reported that E2 did not induce VEGF mRNA levels in MCF-7 cells (Bogin and Degani, 2002). In contrast, other studies reported that E2 induced VEGF mRNA and protein levels (Ruohola et al., 1999; Buteau-Lozano et al., 2002) in MCF-7 cells. Differences observed among these studies may be due to the heterogeneity of wt and variant ERa expression observed in MCF-7 cells from different laboratories (Klotz et al., 1995). Cell context-dependent differences in VEGF levels and modulation of this response by hormones is due, in part, to differential expression of other nuclear factors and their interactions with cis-regulatory elements. EREs have been identified in both 50 - and 30 -regions on the VEGF promoter (Hyder et al., 2000; Mueller


1057

Figure 4 Effects of E2 on Sp1/Sp3 protein levels and their interactions with the VEGF promoter. (a) Western immuno-blot analysis of Sp1 and Sp3 protein levels in ZR-75 cells. Cells were seeded in phenol red-free, 2.5% charcoal-stripped-serum medium for 1 day, incubated in serum-free medium for 1 day and treated with DMSO or 10 nM E2 for the indicated times. Cells were harvested and Western analysis was performed as described in Materials and methods. E2 treatment did not significantly alter Sp1 or Sp3 protein protein levels. (b) Chromatin immunoprecipitation (ChIP) analysis. ZR-75 cells were treated with E2 for different times, harvested and processed as described in Materials and methods. Antibodies against ERa, Sp1, Sp3 and Brg-1 proteins immunoprecipitated associated VEGF promoter elements from ZR-75 breast cancer cells. The PCR-amplified VEGF genomic promoter element (199 to þ 3) was resolved by agarose gel electrophoresis, transferred to a nylon membrane and detected as described in the Materials and methods. ERa, Sp1, Sp3 and Brg-1 proteins were present on the VEGF gene promoter with or without E2 treatment. As a negative control, immunoprecipitated chromatin from ZR-75 cells was tested for the presence of co-precipitated genomic DNA from a region of the CathD gene ( þ 2469 to þ 2615) that does not bind ERa protein. Immunoprecipitation with two different antibodies against ERa protein or IgG failed to detected association of ERa with this region (exon II) of cathepsin D as previously described (Castro-Rivera et al., 2001)

et al., 2000; Buteau-Lozano et al., 2002), and deletion analysis of the VEGF promoter in Ishikawa and MCF-7 cells identified a 1560 to 1175 sequence that was E2responsive in transfection assays and bound ER in gelmobility shift assays (Mueller et al., 2000; ButeauLozano et al., 2002). In contrast, deletion analysis of pVEGF constructs in ZR-75 cells demonstrated that hormone responsiveness was dependent on downstream 131 to 47 GC-rich motifs including the 66 to 47 sequence which bound Sp1 and Sp3 proteins in gel-

mobility shift assays (Figures 1–3). Moreover, GC/GTrich sites within the 267 to 131 region of the promoter may also be important for hormone-responsiveness of VEGF. E2-induced transactivation in ZR-75 cells transfected with pVEGF5 was observed after cotransfection with HEC11C, a DNA binding domain deletion mutant of ERa (Figure 3a). Similar results have been reported for other E2-responsive constructs activated by ERa/Sp1 (Safe, 2001). ERa did not bind directly to the GC-rich motifs; however, ERa physically Oncogene


1058

Figure 5 Inhibition of hormonal activation of VEGF by small inhibitory RNAs for Sp1 and Sp3. Effects of cotransfected small inhibitory RNAs for Sp1 (iSp1) (a) or Sp3 (iSp3) (b) on Sp1 and Sp3 protein levels in ZR-75 cells. ZR-75 cells were transfected with iLMN, iSp1 or iSp3, and whole-cell lysates were analysed by Western blot analysis for Sp1 or Sp3 proteins. An unchanged nonspecific band was also detected and shown as a control. The experiment was repeated (3 ), and results are given as means7s.e. where the OD of the band (Sp1 or Sp3) in the DMSO treated cells was set at 100%. Significantly decreased protein levels are indicated (*, Po0.05, n ¼ 3). iSp1 and iSp3 did not affect levels of Sp3 and Sp1 proteins, respectively (data not shown). (c–e) Effects of iLMN, iGL2, iSp1 and iSp3 on E2-increased luciferase activity in ZR-75 cells transfected with pVEGF1 (c and d) and pVEGF6 (e). Cells were transfected with pVEGF1 (c and d) or pVEGF6 (e) and the siRNAs, treated with DMSO or 10 nM E2, and induced luciferase activities were determined as described in the Materials and methods. Results are expressed as mean þ s.e. and significantly decreased activity is indicated (*, Po0.05, n ¼ 3). Levels of ERa protein were not significantly changed (data not shown) as previously reported in MCF-7 and ZR-75 cells transfected with iSp1 or iLMN (Abdelrahim et al., 2002)

interacts with Sp1 and Sp3 proteins (Porter et al., 1997; Saville et al., 2000; Stoner et al., 2000) and enhances Sp1/Sp3–DNA binding in gel-mobility shift assays using the GC-rich oligonucleotide 66 to 47 from the VEGF gene promoter (Figure 3c). These data clearly demonstrate cell context-dependent differences in upreOncogene

gulation of VEGF by E2 where a distal nonconsensus ERE has been linked to gene expression in Ishikawa and MCF-7 cells (Mueller et al., 2000; Buteau-Lozano et al., 2002). Results of our deletion analysis of the VEGF gene promoter in ZR-75 cells (Figure 2a and b) do not exclude a contribution of the distal ERE. However,


1059

GC-rich sites are important for E2-responsiveness in ZR-75 cells since small inhibitory RNAs for Sp1 and Sp3 significantly decrease hormone-responsiveness of pVEGF1 and pVEGF6 (Figure 5). Analysis of the VEGF gene promoter in the ChIP assay also shows that Sp1, Sp3 and ERa are bound to the proximal region of the promoter (Figure 4b), and these results are consistent with the deletion analysis and gel-mobility shift assays. Previous studies have identified ERa and Sp proteins associated with GC-rich promoters using DNA footprinting or ChIP assays (Stoner et al., 2000; Castro-Rivera et al., 2001; Samudio et al., 2001); however, it has also been reported that interactions of ERa and Brg-1 with cathepsin D promoter are hormone-dependent (DiRenzo et al., 2000). The reasons for these difference are unclear; however, for ERachromatin interactions, we have observed significant variability that is antibody-dependent. In this study, ERa (D12) antibodies detected ERa-VEGF promoter interactions at time 0, whereas other ERa antibodies gave more or less intense bands with the VEGF and other GC-rich E2-responsive gene promoters (CastroRivera et al., 2001; Samudio et al., 2001). Current studies are further investigating the influence of multiple antibodies on results of ChIP assays. Several studies have also demonstrated the importance of Sp1 interactions with the proximal GC-rich sites for constitutive and induced expression of VEGF in different cancer cell lines (Ryuto et al., 1996; Finkenzeller et al., 1997; Gille et al., 1997; Salimath et al., 2000; Shi et al., 2001). However, in ZR-75 cells, both Sp1 and Sp3 bind the 66 to 47 VEGF oligonucleotide, and similar results have been observed for extracts from HEC1A (Stoner et al., 2000), MCF-7 and T47D cell lines (data not shown). Previous studies in multiple cell lines including Drosophila Schneider SL-2 cells suggest that ERa/Sp1 is associated with hormone-induced upregulation of gene expression (reviewed in Safe, 2001), whereas ERa/Sp3 was linked to decreased VEGF expression in HEC1A cells (Stoner et al., 2000). Although Sp3 interactions with GC-rich promoters are associated with downregulation of some genes, there are also reports that Sp3 alone or in combination with Sp1 is important for high constitutive gene expression (Xiao et al., 2000). Results in Figure 5 show that ectopic expression of small inhibitory mRNAs for Sp1 or Sp3 decreased levels of these proteins in whole cell lysates and significantly decreased E2-induced activation of pVEGF1 and pVEGF6. These results indicate that proximal and distal GC/GT sites are involved in E2dependent induction of VEGF gene expression by ERa/ Sp1 and ERa/Sp3. Thus, E2-dependent activation of ERa/Sp3 can result in induction (ZR-75) or repression (HEC1A) of VEGF in different cell lines, and this may be related to differential expression of key nuclear coactivators, corepressors and other factors. Cell context also plays an important role in ligand activation of ERa/Sp1-mediated responses. For example, ERa/Sp1 interaction with GC-rich sites in breast cancer cells is important for induction of several E2-responsive genes (reviewed in Safe, 2001). In contrast, E2-dependent

downregulation of insulin-like growth factor 1 receptor expression in rat aortic smooth muscle cells is mediated through ERa/Sp1 (Scheidegger et al., 2000). GC-rich motifs in several gene promoters are cisacting elements for other ligand-activated nuclear and orphan receptors which interact with Sp1 and related proteins (Rohr et al., 1997; Monte et al., 1998; Owen et al., 1998; Pipaoń et al., 1999; Simmen et al., 1999; Suzuki et al., 1999). Thus, Sp-family proteins are important DNA-bound transcription factors that mediate transactivation by nuclear receptors and contribute to their cell-context-dependent up- or downregulation of selected target genes such as VEGF. Current studies are focused on characterizing important interacting nuclear cofactors that are required for ER/Sp-family complexdependent gene expression and have not yet been identified. Identification of nuclear cofactors and Sp proteins that regulate basal and hormone-inducible VEGF expression are also potential therapeutic targets for inactivation since this could lead to inhibition of tumor cell growth and/or metastasis.

Materials and methods Cell lines, chemicals and biochemicals ZR-75 human breast cancer cell line was obtained from American Type Culture Collection (ATCC, Rockville, MD, USA) and was cultured in RPMI 1640 medium (Sigma Chemical Co., St Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS) (Summit Biotechnology, Fort Collins, CO, USA, JRH Biosciences, Lenexa, KS, USA or Atlanta Biologicals, Inc., Norcross, GA, USA), 2.2 g/l sodium bicarbonate, 4.8 g/l HEPES, 4.5 g/l dextrose, 0.22 g/l sodium pyruvate. Cells were maintained in 371C incubators under humidified 5% CO2 : 95% air. Sp1, Sp3, ERa (D12) and Brg-1 antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Dimethyl sulfoxide (DMSO), E2, progesterone and other chemicals were obtained from commercial sources at the highest quality available. Oligonucleotides, plasmids and small inhibitory RNAs VEGF promoter derived oligonucleotides, PCR primers and primers employed in plasmid construction were synthesized by Genosys/Sigma (The Woodlands, TX, USA), Gene Technologies Laboratory [Texas A&M University (TAMU), College Station, TX, USA], or by the laboratory of Dr James Derr (Department of Veterinary Pathobiology, TAMU) or Integrated DNA Technologies (IDT) (Coralville, IA, USA). Consensus and mutant GC-rich oligonucleotides were designed according to published sequences in Promega Co. (Madison, WI, USA) and Santa Cruz Biotech (Santa Cruz, CA, USA) catalogs. A consensus estrogen responsive element (ERE) probe used in EMSA assays was 50 -GTC CAA AGT CAG GTC ACA GTG ACC TGA TCA AAG TT-30 . VEGF promoter luciferase constructs pVEGF1, pVEGF2 and pVEGF3 were previously named p2068Luc, p840Luc and p318Luc, and were generously provided by Drs Gerhard Siemeister and Gunter Finkenzeller (Institute of Molecular Medicine, Tumor Biology Center, Freiburg, Germany). pVEGF5 (131/ þ 54), pVEGF6 and pVEGF7 were produced by PCR amplification (Stoner et al., 2000). pVEGF8 and pVEGF8m were constructed by insertion of oligonucleotides Oncogene


1060 66/47 (50 -CT CCC GGC GGG GCG GAG CCA TG-30 ) and 66/47 m (50 -CT CCC GGC TTT TCG GAG CCA TG30 ) into TATA-pGL2 digested with KpnI and NheI. The 66/ 47 and 66/47m oligonucleotides used for EMSA analysis were 50 -GGT CCC GGC GGG GCG GAG CCA TG-30 and 50 -GGT CCC GGC TTT TCG GAG CCA TG-30 , respectively. Human estrogen receptor (hERa) expression plasmid was kindly provided by Dr Ming-Jer Tsai (Baylor College of Medicine, Houston, TX, USA). hERa deletion constructs hE11C, hE15C and hE19C were originally obtained from Dr Pierre Chambon (GCBMC, Illkirch, France). hERa cDNAs were also inserted into vectors pCDNA3 and pCDNA3.1/His C for transient expression in mammalian cells. Small inhibitory RNA (siRNA) duplexes were prepared by Dharmacon Research (Lafayette, CO, USA) and targeted coding regions of Sp1, Sp3, lamin A/C (608–626) and GL2 (luciferase) (153–171). Sequences of individual siRNA duplexes are summarized below, and approaches used for selecting specific sequences were comparable to those described (Harborth et al., 2001).

Gene

siRNA Duplex

GL2

50 -CGU ACG CGG AAU ACU UCG ATT TT GCA UGC GCC UUA UGA AGC U-50

LMN

50 -CUG GAC UUC CAG AAG AAC ATT TT GAC CUG AAG GUC UUC UUG U-50

Sp1

50 -AUC ACU CCA UGG AUG AAA UGA TT TT UAG UGA GGU ACC UAC UUU ACU-50

Sp3

50 -GCG GCA GGU GGA GCC UUC ACU TT TT CGC CGU CCA CCU CGG AAG UGA-50

Transfection with siRNAs was essentially carried out as previously described in ZR-75 and MCF-7 cells (Abdelrahim et al., 2002), and the final concentration of siRNAs was 140 nM. For transactivation studies, cells were transfected using Oligofectamine reagent (InVitrogen, Carlsbad, CA, USA) with pVEGF6 and siRNAs for 36 h. Cells were then treated with DMSO or 10 nM E2 for 20 h, and luciferase activity was determined as described above. Western blot analysis of whole-cell lysates were obtained from ZR-75 cells transfected with siRNAs for 48 h as described (Abdelrahim et al., 2002; Wormke et al., 2003). Transient transfection assays VEGF-luciferase-reporter plasmids, hERa expression plasmid and pCDNA3.1-His-LacZ. expression plasmid (Invitrogen, Carlsbad, CA, USA) (for normalization of transfection efficiency) were transiently transfected into ZR-75 cells using the calcium phosphate–DNA co-precipitation method. pCDNA3 or pCDNA3.1 (InVitrogen, Carlsbad, CA, USA) empty vectors were transfected to maintain DNA mass balance among different transfection groups. A pC3-Luc highly E2inducible control plasmid containing the promoter region of the human complement-3 gene (C3), generously supplied by Dr Donald P McDonnell (Duke University Medical School, Durham, NC, USA), was often used as a positive control in experiments to confirm estrogen-responsiveness of the cells. Following transfection, cells were shocked with a 25% glycerol : 75% PBS mixture to increase transfection efficiency, Oncogene

then cells were washed once with PBS and were treated for 48 h with fresh serum-free medium containing 10 nM E2 dissolved in DMSO. Control treatments were with DMSO alone. Antiestrogen treatments were: 1 mM ICI 182,780 (Alan Wakeling, Zeneca Pharmaceuticals) and 1 mM tamoxifen or 4OHtamoxifen (Sigma). Following treatment, cells were harvested by scraping in 1 lysis buffer (Promega) and soluble protein was extracted by one cycle of freezing/thawing the cells. A volume of 20 ml aliquots of transfected cell lysates was assayed for luciferase activity using Luciferase Assay Reagent (Promega) and b-galactosidase activity using Tropix Galacto-Light Plus Assay System (Tropix, Bedford, MA, USA) in a Packard Lumicount micro-well plate reader (Packard Instrument Co., Downers Grove, IL, USA). Relative luciferase activity was normalized to relative b-galactosidase units to obtain transfection values. Semiquantitative reverse transcriptase–polymerase chain reaction (RT–PCR) analysis The following VEGF and GAPDH primers were used: 249 bp VEGF(F) 50 -CCA TGA ACT TTC TGC TGT CTT-30 , 249 bp VEGF(R) 50 -ATC GCA TCA GGG GCA CAC AG-30 , 558 bp GAPDH(F) 50 -AAT CCC ATC ACC ATC TTC CA-30 , 558 bp GAPDH(R) 50 -GTC ATC ATA TTT GGC AGG TT-30 , ZR75 cells were cultured in serum-free medium for 3 days prior to treatment with DMSO or 10 nM E2 for different times. Total RNA was obtained with RNAzol B (Tel-Test, Friendswood, TX, USA), according to the manufacturer’s protocol. RNA concentration was measured by UV 260/280 nm absorption ratio, and 100–400 ng/ml RNA was used in each reaction for RT–PCR. RNA was reverse transcribed at 421C for 25 min using oligo d(T) primer (Promega), and subsequently PCR amplified of RT product using 2 mM MgCl2,, 1 mM of each gene-specific primer, 1 mM dNTPs, and 2.5 U AmpliTaq DNA polymerase (Promega). VEGF and GAPDH primer sets were both added to each mixture and the gene products coamplified using 22–25 cycles (951C, 30 s; 581C, 30 s; 721C, 30 s). Following amplification in a PCR Express thermal cycler (Hybaid US, Franklin, MA, USA), 18 ml of each sample was loaded on a 2% agarose gel containing ethidium bromide. Electrophoresis was performed at 80 V in 1 TAE buffer for 1 h and the gel was photographed on by UV-transillumination using Polaroid film. GAPDH and VEGF band intensity values were obtained by scanning the Polaroid on a Sharp JX-330 scanner (Sharp Electronics, Mahwah, NJ, USA), followed by image negation in Adobe Photoshop 3.0 (Adobe Systems Inc., San Jose, CA, USA), background signal subtraction and densitometric analysis using Zero-D software (Scanalytics, Sunnyvale, CA, USA) software. Results were expressed as VEGF band intensity values normalized to GAPDH values, then by averaging duplicate determinations for each treatment group. Preparation of ZR-75 cell nuclear extracts Cells were cultured in medium without phenol red, supplemented with 2.5% fetal bovine serum, pretreated with dextrancoated charcoal, to remove growth factors and endogenous estrogens. Treatments were added to cells for 1–24 h prior to harvesting by scraping in HED (25 mM HEPES, 1.5 mM EDTA, 1 mM dithiothreitol, pH 7.6). Cells were allowed to swell on ice for 15 min, followed by cellular homogenization in HEGD (HED with 10% glycerol) using a Type B Dounce homogenizer. Cell viability was verified by staining with trypan-blue, and homogenization was stopped with approximately 80–90% of cells broken. Centrifugation of the cellular


1061 homogenate yielded a pellet containing crude nuclei. Nuclei were incubated in high salt buffer HEGDK þ (HEGD containing 0.5 M KCl) on ice, for 30 min to 1 h. All buffers were supplemented with a mixture of protease inhibitors. Nuclei in high salt buffer were centrifuged at high speed for 15–30 min at 41C and extracted nuclear proteins in the supernatant were aliquoted in small volumes and stored at 801C for use in Electrophoretic mobility shift assays (EMSA).

followed by incubation with horseradish peroxidase conjugated anti-rabbit secondary antibody (1 : 5000 dilution). Blots were exposed to chemiluminescent substrate (ECL, NENDuPont, Boston, MA, USA) and placed on Kodak X-O-Mat autoradiography film (Eastman Kodak Co., Rochester, NY, USA). Sp1 and Sp3 protein band intensities were measured by scanning, densitometry and background subtraction using Zero-D Scanalytics software (Scanalytics Corp.).

EMSA

Chromatin immunoprecipitation (CHIP) assay

Oligonucleotides from the VEGF promoter were annealed and 5 pmol were 50 -end labeled using T4 kinase (Promega) and [32P]ATP (NEN-Dupont, Boston, MA, USA). A 20–30 ml EMSA reaction mixture contained 75–150 mM salt in a mixture of HEGD and HEGDK þ to achieve the proper salt concentration, 1–10 mg of nuclear protein, 500–1000 ng poly(dI-dC) or poly(dA-dT) (Roche Molecular Biochemicals, Basel, Switzerland), with or without unlabeled competitor oligonucleotide, and 10 fmol labeled probe incubated for 15 min on ice. Commercially available antibodies to Sp1, Sp3 and hERa proteins were added to reaction mixes immediately after the probe and incubated a further 15–60 min on ice as described (Stoner et al., 2000). Protein : DNA complexes were resolved by 5% PAGE at 110–200 V at room temperature or 41C for 1.5–4 h. Specific antibody/protein interactions were observed as more slowly migrating complexes in the gel. Specific nuclear extract/VEGF promoter complexes formed were separated from free probe by electrophoresis at room temperature on a 5% polyacrylamide gel in 1 TBE (0.09 M Tris-base, 0.09 M boric acid, 2mM EDTA, pH 8.3), 120 V, for 2–4 h. For EMSA of pure proteins or in vitro transcribed/translated proteins, 2 ml rabbit reticulocyte lysate containing Sp1 or Sp3 protein were incubated in 1 binding buffer (20 mM HEPES, 100 mM KCl, 1 mM EDTA, 5% glycerol (v : v) and 40 mM ZnCl2; pH 8.0). For hERa enhancement of Sp1 and Sp3 binding, pure hERa protein (PanVera, Madison, WI, USA) was diluted to 100 fmol/ml in BSA and added to aliquots of nuclear extracts from cells treated with E2. All reactions were balanced for with equal amounts of BSA and rabbit reticulocyte lysate that ‘mock’-transcribed/translated empty pCDNA3 or pCDNA3.1 vector. Polyclonal antibodies against transcription factors Sp1 and Sp3 (Santa Cruz Biotech) were added to some reaction mixtures, immediately after labeled probe, and supershifted specific protein : DNA complexes were observed.

ZR-75 breast cancer cells were grown to 95% confluence, placed under serum-free conditions for 72 h, and treated with 10 nM E2 for varying times. Cells were then crosslinked by the addition of formaldehyde to a final concentration of 1%, and the crosslinking reaction was stopped by addition of 0.125 M glycine. After washing several times with PBS containing 1 mM PMSF, cells were scraped into 50 ml conical tubes on ice, and pelleted by centrifugation. Collected cells were hypotonically lysed in the presence of a nonionic detergent and a ‘cocktail’ of protease inhibitors, and nuclei were then collected by highspeed centrifugation. Nuclei were resuspended in sonication buffer and sonicated for 35–50 s. The resulting chromatin was precleared by addition of normal rabbit IgG and protein Aconjugated beads, followed by incubation for 1 h at 41C with gentle agitation. The beads were pelleted and the precleared chromatin supernatant was immunoprecipitated with antibodies to ERa, Sp1, Sp3 or Brahma-related gene 1 (Brg-1) (Santa Cruz Biotechnology). After sample incubation for 8 h at 41C, protein–antibody complexes were collected by addition of protein A-conjugated beads at room temperature for 30 min; the beads were extensively washed, protein–DNA crosslinks were reversed, and the DNA purified. In order to increase the specificity of the PCR analysis of immunoprecipitated DNA, and to reduce primer artifacts, the reactions were carried out in the presence of 1 M betaine and/or varying amounts of DMSO (3–6%) in a stringent two-step PCR program (95–601C or 95– 651C). VEGF primers were: (forward) 50 -GGT CGA GCT TCC CCT TCA-30 and (reverse) 50 -GAT CCT CCC CGC TAC CAG-30 and amplified a 202 bp product of the VEGF gene promoter from 199 to þ 3; cathepsin D primers were: (forward) 50 -TGC ACA AGT TCA CGT CCA TC-30 and (reverse) 50 -TGT AGT TCT TGA GCA CCT CG-30 and amplified a 150 bp region of the human cathepsin D exon II. PCR was products were resolved on a 1.5% agarose gel in the presence of 0.5 mg/ml ethidium bromide and transferred to a charged nylon membrane by common capillary transfer. The membrane was probed with either a radiolabeled plasmid that contains the promoter region of interest, or with radio-labeled PCR product from the same plasmid using primers used for amplification from immunoprecipitated DNA. After extensive washing, radioactive signals were quantified by exposure to a phosphor-storage screen, followed by scanning on a STORM 860 (Molecular Dynamics, Sunnyvale, CA, USA).

Western immunoblot analysis ZR-75 cells were seeded into six-well plates in culture medium supplemented with 2.5%. charcoal stripped-FBS. Medium was changed to serum-free medium for 1 day and cells were treated with DMSO or 10 nM E2 over a time course. Cells were harvested in lysis buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 10% (v/v) glycerol, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA, 10 mg/ml aprotinin, 50 mM phenylmethylsulphonylflouride, 50 mM sodium orthovanadate) and whole-cell extract of soluble protein was obtained by centrifugation at 14 000 g for 5 min at 41C. A measure of 30 mg of protein was diluted with Laemmli’s loading buffer, boiled and loaded on a 7.5% SDS–polyacrylamide gel. Samples were resolved by electrophoresis at 150–180 V for 3–4 h, and separated proteins were transferred (transfer buffer: 48 mM Tris–HCl, 29 mM glycine and 0.025% SDS) to PVDF membrane (BioRad, Hercules, CA, USA). Specific proteins on the membrane were detected with polyclonal primary antibodies Sp1-PEP2 and Sp3-D20 (each diluted 1 : 1000) against Sp1 and Sp3 proteins,

Statistical analysis Experiments were repeated as indicated, and results expressed as mean7s.e. Data were subjected to ANOVA and Scheffe’s test, or Fisher’s Protected LSD. Acknowledgements Supported by financial assistance from the National Institutes of Health (ES04176, CA076636 and ES09106) and the Texas Agricultural Experiment Station. Oncogene


1062 References Abdelrahim M, Samudio I, Smith R, Burghardt R and Safe S. (2002). J. Biol. Chem., 277, 28815–28822. Augereau P, Miralles F, Cavailles V, Gaudelet C, Parker M and Rochefort H. (1994). Mol. Endocrinol., 8, 693–703. Bae DG, Gho YS, Yoon WH and Chae CB. (2000). J. Biol. Chem., 275, 13588–13596. Berry M, Nunez A-M and Chambon P. (1989). Proc. Natl. Acad. Sci. USA, 86, 1218–1222. Bogin L and Degani H. (2002). Cancer Res., 62, 1948–1951. Boocock CA, Charnock-Jones DS, Sharkey AM, McLaren J, Barker PJ, Wright KA, Twentyman PR and Smith SK. (1995). J. Natl. Cancer Inst., 87, 506–516. Bos R, Zhong H, Hanrahan CF, Mommers EC, Semenza GL, Pinedo HM, Abeloff MD, Simons JW, van Diest PJ and van der Wall E. (2001). J. Natl. Cancer Inst., 93, 309–314. Brekken RA, Overholser JP, Stastny VA, Waltenberger J, Minna JD and Thorpe PE. (2000). Cancer Res., 60, 5117–5124. Brown LF, Berse B, Jackman RW, Tognazzi K, Manseau EJ, Dvorak HF and Senger DR. (1993). Am. J. Pathol., 143, 1255–1262. Buteau-Lozano H, Ancelin M, Lardeux B, Milanini J and Perrot-Applanat M. (2002). Cancer Res., 62, 4977–4984. Castro-Rivera E, Samudio I and Safe S. (2001). J. Biol. Chem., 276, 30853–30861. Cavailles V, Augereau P and Rochefort H. (1993). Proc. Natl. Acad. Sci. USA, 90, 203–207. Cavailles V, Garcia M and Rochefort H. (1989). Mol. Endocrinol., 3, 552–558. Classen-Linke I, Alfer J, Krusche CA, Chwalisz K, Rath W and Beier HM. (2000). Steroids, 65, 763–771. Cullinan-Bove K and Koos RD. (1993). Endocrinology, 133, 829–837. DiRenzo J, Shang Y, Phelan M, Sif S, Myers M, Kingston R and Brown M. (2000). Mol. Cell. Biol., 20, 7541–7549. Duan R, Porter W and Safe S. (1998). Endocrinology, 139, 1981–1990. Enholm B, Paavonen K, Ristimaki A, Kumar V, Gunji Y, Klefstrom J, Kivinen L, Laiho M, Olofsson B, Joukov V, Eriksson U and Alitalo K. (1997). Oncogene, 14, 2475–2483. Ferrara N and Davis-Smyth T. (1997). Endocr. Rev., 18, 4–25. Finkenzeller G, Sparacio A, Technau A, Marme D and Siemeister G. (1997). Oncogene, 15, 669–676. Gille J, Swerlick RA and Caughman SW. (1997). EMBO J., 16, 750–759. Gille J, Swerlick RA, Lawley TJ and Caughman SW. (1996). Blood, 87, 211–217. Gillesby B, Santostefano M, Porter W, Wu ZF, Safe S and Zacharewski T. (1997). Biochemistry, 36, 6080–6089. Goldman CK, Kim J, Wong WL, King V, Brock T and Gillespie GY. (1993). Mol. Biol. Cell, 4, 121–133. Hanahan D and Folkman J. (1996). Cell, 86, 353–364. Harborth J, Elbashir SM, Bechert K, Tuschl T and Weber K. (2001). J. Cell Sci., 114, 4557–4565. Hyder SM, Chiappetta C and Stancel GM. (1997). Cancer Lett., 120, 165–171. Hyder SM, Murthy L and Stancel GM. (1998). Cancer Res., 58, 392–395. Hyder SM, Nawaz Z, Chiappetta C and Stancel GM. (2000). Cancer Res., 60, 3183–3190. Oncogene

Hyder SM, Stancel GM, Chiappetta C, Murthy L, Boettger-Tong HL and Makela S. (1996). Cancer Res., 56, 3954–3960. Kimura H, Weisz A, Ogura T, Hitomi Y, Kurashima Y, Hashimoto K, D’Acquisto F, Makuuchi M and Esumi H. (2001). J. Biol. Chem., 276, 2292–2298. Klotz DM, Castles CG, Fuqua SAW, Spriggs LL and Hill SM. (1995). Biochem. Biophys. Res. Commun., 210, 609–615. Laitinen M, Ristimaki A, Honkasalo M, Narko K, Paavonen K and Ritvos O. (1997). Endocrinology, 138, 4748–4756. Linderholm BK, Lindahl T, Holmberg L, Klaar S, Lennerstrand J, Henriksson R and Bergh J. (2001). Cancer Res., 61, 2256–2260. Monte D, DeWitte F and Hum DW. (1998). J. Biol. Chem., 273, 4585–4591. Mueller MD, Vigne JL, Minchenko A, Lebovic DI, Leitman DC and Taylor RN. (2000). Proc. Natl. Acad. Sci. USA, 97, 10972–10977. Nakamura J, Savinov A, Lu Q and Brodie A. (1996). Endocrinology, 137, 5589–5596. Owen GI, Richer JK, Tung L, Takimoto G and Horwitz KB. (1998). J. Biol. Chem., 273, 10696–10701. Paech K, Webb P, Kuiper GG, Nilsson S, Gustafsson J, Kushner PJ and Scanlan TS. (1997). Science, 277, 1508–1510. Pal S, Claffey KP, Cohen HT and Mukhopadhyay D. (1998). J. Biol. Chem., 273, 26277–26280. Pepper MS, Ferrara N, Orci L and Montesano R. (1992). Biochem. Biophys. Res. Commun., 189, 824–831. Pipaoń C, Tsai SY and Tsai MJ. (1999). Mol. Cell. Biol., 19, 2734–2745. Porter W, Saville B, Hoivik D and Safe S. (1997). Mol. Endocrinol., 11, 1569–1580. Rohr O, Aunis D and Schaeffer E. (1997). J. Biol. Chem., 272, 31149–31155. Ruohola JK, Valve EM, Karkkainen MJ, Joukov V, Alitalo K and Ha¨rko¨nen PL. (1999). Mol. Cell. Endocrinol., 149, 29–40. Ryuto M, Ono M, Izumi H, Yoshida S, Weich HA, Kohno K and Kuwano M. (1996). J. Biol. Chem., 271, 28220–28228. Safe S. (2001). Vitam. Horm., 62, 231–252. Salimath B, Marme D and Finkenzeller G. (2000). Oncogene, 19, 3470–3476. Samudio I, Vyhlidal C, Wang F, Stoner M, Chen I, Kladde M, Barhoumi R, Burghardt R and Safe S. (2001). Endocrinology, 142, 1000–1008. Sathya G, Li W, Klinge CM, Anolik JH, Hilf R and Bambara RA. (1997). Mol. Endocrinol., 11, 1994–2003. Saville B, Wormke M, Wang F, Nguyen T, Enmark E, Kuiper G, Gustafsson J-A and Safe S. (2000). J. Biol. Chem., 275, 5379–5387. Scheidegger KJ, Cenni B, Picard D and Delafontaine P. (2000). J. Biol. Chem., 275, 38921–38928. Schlaeppi JM and Wood JM. (1999). Cancer Metastasis Rev., 18, 473–481. Shi Q, Le X, Abbruzzese JL, Peng Z, Qian CN, Tang H, Xiong Q, Wang B, Li XC and Xie K. (2001). Cancer Res., 61, 4143–4154. Simmen RCM, Chung TE, Imataka H, Michel FJ, Badinga L and Simmen FA. (1999). Endocrinology, 140, 2517–2525.


1063 Stoner M, Wang F, Wormke M, Nguyen T, Samudio I, Vyhlidal C, Marme D, Finkenzeller G and Safe S. (2000). J. Biol. Chem., 275, 22769–22779. Suzuki Y, Shimada J, Shudo K, Matsumura M, Crippa MP and Kojima S. (1999). Blood, 93, 4264–4276. Viglietto G, Maglione D, Rambaldi M, Cerutti J, Romano A, Trapasso F, Fedele M, Ippolito P, Chiappetta G and Botti G. (1995). Oncogene, 11, 1569–1579.

Weisz A and Rosales R. (1990). Nucleic Acids Res., 18, 5097–5106. Wormke M, Stoner M, Saville B, Walker K, Abdelrahim M, Burghardt R and Safe S. (2003). Mol. Cell. Biol., 23, 1843–1855. Xiao H, Hasegawa T and Isobe K-I. (2000). J. Biol. Chem., 275, 1371–1376. Yoshiji H, Harris SR and Thorgeirsson UP. (1997). Cancer Res., 57, 3924–3928. Zetter BR. (1998). Annu. Rev. Med., 49, 407–424.

Oncogene