Estrogen Inducibility of c-Ha-ras Transcription in Breast Cancer Cells

Although mutation of ras gene is rare in human breast cancer, overexpression of normal c-Ha-rasgene is frequently observed. Using a mouse mammary metastasis model consisting of genetically related mammary tumor sublines with variant metastatic potential, we have previously (i) demonstrated a direct correlation between c-Ha-ras mRNA and protein levels and metastatic potential and (ii) identified a novel hormone-responsive transcriptional regulatory element in intron 1 of the mouse c-Ha-ras gene that contains the consensus half-site of a glucocorticoid response element and flanking consensus half-sites for estrogen response element. Here, we have examined the functionality of intron 1 sequence in context of upstream sequences by using transient transfection assays with plasmids expressing chloramphenicol acetyltransferase. Intron 1 sequence and sequences similar to intron 1 element located in exon 1 function as transcriptional regulatory elements that confer hormonal inducibility to chloramphenicol acetyltransferase gene expression both independently and in context of 5′-flanking sequences. Measurement of c-Ha-rastranscription rates and protein expression by nuclear run-on and metabolic labeling assays showed a 5–12-fold enhancement, respectively, following treatment with 17β-estradiol that was blunted by ICI 182,780 in the nonmetastatic variant. In contrast, constitutive overexpression of c-Ha-ras transcripts and protein in the metastatic subline was unaffected by estrogen and ICI 182,780. Gel shift assays demonstrated specific interaction of c-Ha-rasexon 1 sequence with nuclear proteins of human breast cancer MCF-7 cells with formation of two complexes, one of which contains estrogen receptor. Our data demonstrate a direct (i) interaction of c-Ha-ras sequence with estrogen receptor and (ii) stimulatory effect of estrogen on c-Ha-ras gene transcription and suggest that alteration in transcriptional regulation of c-Ha-ras gene by estrogen may play an important role in progression of breast cancer.

Although mutation of ras gene is rare in human breast cancer, overexpression of normal c-Ha-ras gene is frequently observed. Using a mouse mammary metastasis model consisting of genetically related mammary tumor sublines with variant metastatic potential, we have previously (i) demonstrated a direct correlation between c-Ha-ras mRNA and protein levels and metastatic potential and (ii) identified a novel hormone-responsive transcriptional regulatory element in intron 1 of the mouse c-Ha-ras gene that contains the consensus half-site of a glucocorticoid response element and flanking consensus half-sites for estrogen response element. Here, we have examined the functionality of intron 1 sequence in context of upstream sequences by using transient transfection assays with plasmids expressing chloramphenicol acetyltransferase. Intron 1 sequence and sequences similar to intron 1 element located in exon 1 function as transcriptional regulatory elements that confer hormonal inducibility to chloramphenicol acetyltransferase gene expression both independently and in context of 5-flanking sequences. Measurement of c-Ha-ras transcription rates and protein expression by nuclear run-on and metabolic labeling assays showed a 5-12-fold enhancement, respectively, following treatment with 17␤-estradiol that was blunted by ICI 182,780 in the nonmetastatic variant. In contrast, constitutive overexpression of c-Ha-ras transcripts and protein in the metastatic subline was unaffected by estrogen and ICI 182,780. Gel shift assays demonstrated specific interaction of c-Ha-ras exon 1 sequence with nuclear proteins of human breast cancer MCF-7 cells with formation of two complexes, one of which contains estrogen receptor.

Our data demonstrate a direct (i) interaction of c-Ha-ras sequence with estrogen receptor and (ii) stimulatory effect of estrogen on c-Ha-ras gene transcription and suggest that alteration in transcriptional regulation of c-Ha-ras gene by estrogen may play an important role in progression of breast cancer.
Ras proteins are responsible for regulating the flow of infor-mation that is triggered by diverse extracellular signals that stimulate their respective cell surface receptors. The relay of their signals via ras proteins ultimately regulate the activities of nuclear transcription factors that control the expression of key genes that regulate cell growth and differentiation (1). Thus, regulation of transcription of members of the ras gene family plays an important role in controlling cell growth. The ras genes are classified as "housekeeping genes," which are expressed at relatively constant levels in all tissues and stages of development. Alteration in structure and expression of the ras gene family has been found in human tumors (2) as well as in animal tumor model systems (3)(4)(5). Although mutation of ras genes is rare in human breast cancers, 50% of human breast carcinomas express elevated levels of normal Ha-ras protein (6 -8). Experimental induction of ras levels has been shown to lead to transformation of certain recipient cell types (9). Tumors that lack activated Ha-ras genes frequently show overexpression of ras proteins, perhaps a result of transcriptional deregulation of the ras gene (10).
The 5Ј region of the murine Ha-ras gene is highly homologous with the 5Ј upstream region of both the rat (11) and human (12) c-Ha-ras genes. The Ha-ras promoter regions for mouse, rat, and human are all very G-C rich and contain numerous repeats of the core consensus sequences for binding Sp1 transcription factor, lack TATA box, and contain a consensus sequence for the CAAT box. The cloned mouse Ha-ras upstream region has been shown to possess powerful transcriptional machinery when tested in transient gene expression assays in primary cultures of mouse epidermal cells (13). This high level of expression in transfected cells is surprising in view of the low level of expression of the endogenous mouse Ha-ras gene in normal mouse epidermis (14,15). This suggests that transcription of c-Ha-ras is tightly regulated in normal tissues, which when lost or deregulated could result in constitutive overexpression of Ha-ras that is frequently observed in tumor tissues.
Estrogen plays a crucial role in cell growth and proliferation of reproductive tissues such as uterus and mammary gland (16,17). Although the precise role of estrogen in the biology of breast carcinogenesis is not known, the effects of estrogen on proliferation of target breast cells are believed to be mediated through transactivation of specific genes that are recognized by estradiol-estrogen receptor (E 2 -ER) 1 complex. This process stimulates DNA synthesis, cell division, and production of biologically active proteins, such as pS2, tumor growth factor-␣, and epidermal growth factor (18), that influence cell growth and differentiation. Although a sequence motif, 5Ј-GGTCGN 3 TGACC-3Ј, having 92% homology to the vitellogenin estrogen response element (ERE) and differing from the consensus sequence by one base, is present in the Ha-ras gene at position Ϫ1420 (numbered according to Brown et al. (19)), no significant induction of transcription of reporter gene from the mouse Ha-ras promoter was observed in presence of estrogen in estrogen-dependent human breast cancer MCF-7 cells (13).
We have previously reported identification of a novel dexamethasone-responsive transcriptional enhancer element in the intron 1 of the mouse c-Ha-ras gene (20). This regulatory sequence contains the consensus half-site of a glucocorticoid response element (GRE) that is point mutated in a metastatic mammary tumor line and has three 5Ј-flanking ERE half-sites in its immediate vicinity. To investigate the relevance of sequences similar to ras intron 1 element in the hormonal regulation of Ha-ras gene expression, we carried out functional analysis of intron 1 in the presence of intron 0, a region that is conserved in mouse, rat, and human c-Ha-ras genes and previously demonstrated to possess strong transcriptional regulatory activity (21). Our data show that sequences similar to the ras intron 1 element are present both in exon 1 and intron 1 and confer strong estrogen and dexamethasone inducibility to CAT reporter gene expression independently or in context of 5Ј-flanking sequences. The presence of the naturally occurring point mutation in the consensus GRE half-site of the ras intron 1 element selectively eliminates inducibility by dexamethasone. Pretranslational control of Ha-ras synthesis by estrogen or dexamethasone could be exerted at transcriptional and/or posttranscriptional levels. To investigate the molecular sites of action of estrogen, we estimated the relative rates of Ha-ras gene transcription following treatment with 17␤-estradiol in two genetically related mouse mammary tumor sublines with variant metastatic capacities. Our data show that E 2 enhances Ha-ras gene transcription in the otherwise low ras-expressing nonmetastatic variant and has no effect on constitutive overexpression observed in the metastatic subline. The E 2 -mediated enhancement of Ha-ras synthesis is regulated at the transcriptional level and is not dependent on de novo protein synthesis. Gel shift assays demonstrated specific interaction of Ha-ras exon 1 sequence with nuclear proteins of human breast cancer MCF-7 cells with formation of two specific complexes, one of which contains estrogen receptor. This is the first report demonstrating direct (i) interaction of Ha-ras sequence with estrogen receptor and (ii) stimulatory effect of estrogen on Ha-ras gene transcription and suggests that alteration in transcriptional regulation of Ha-ras gene by estrogen may play an important role in progression of breast cancer.

EXPERIMENTAL PROCEDURES
Cell Lines and Culture-In this study we have utilized MCF-7 human breast cancer cell line and two mouse mammary tumor sister sublines that are highly tumorigenic but differ in their ability to complete specific steps of the metastatic cascade (22). Line 168FAR is nonmetastatic because of a defect in its ability to extravasate, whereas the related 4T1 subline is highly metastatic and metastasizes spontaneously from the orthotopic site to lungs and liver (22). 168FAR and 4T1 sublines are grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 1 mM nonessential amino acids, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Both sublines used for the study had been maintained in culture for Ͻ20 passages. MCF-7 cells were maintained in Ham's F-12/Dulbecco's modified Eagle's medium (1:1) supplemented with 10% fetal calf serum. For hormone response studies, cells were stripped of endogenous steroids by successive passages in phenol red-free medium supplemented with sera that were unable to support growth of the ER-positive estrogen-dependent cell line, MCF-7, indicating absence of biologically significant levels of E 2 or other estrogenic compounds.
DNA Constructs-To generate 5Ј deletions within the proximal mouse c-Ha-ras regulatory region, fragments containing various lengths of the previously characterized positive regulatory region, intron 0, but including the entire intron 1 were amplified by polymerase chain reaction (PCR) from genomic DNA isolated from normal mouse liver. Fig. 1 illustrates the position of primers used for PCR amplification of the ras fragments, YQ, ZQ, WQ, and PQ, used in the study. PQras containing the naturally occurring A 3 G nucleotide change at base 210 of the mouse Ha-ras gene was amplified by PCR from genomic DNA isolated from the 4T1 subline. The minimal glucocorticoid-responsive transcriptional regulatory sequence, ras intron 1 element, was used as a 30-bp double-stranded oligonucleotide (base 191-220) with and without the A 3 G nucleotide change at base 210 of the mouse Ha-ras gene (Fig. 1). The human c-Ha-ras exon 1 sequence, 5Ј-GT-GCGCTGACCATCCAGCTGATCCAGAACC-3Ј (base 1713-1742; Gen-Bank TM accession number J00277; Ref. 23) that shares 80% homology with the 10 base pair palindrome (underlined regions) in the mouse Ha-ras intron 1 was also used as a 30-bp double-stranded oligonucleotide. DNA fragments were blunt end-ligated upstream of the Ϫ37-base thymidine kinase (TK) gene promoter driving the expression of the bacterial chloramphenicol acetyltransferase (CAT) gene (p-37TK-CAT; Ref 24). The minimal TK promoter contains only the binding site for RNA polymerase II (24). All plasmids were sequenced to verify their insertion, sequence, and orientation. The following primers were used for PCR amplification: 5Ј-CGCGTTGGGCCCGAACCAG-3Ј (Yϩ); 5Ј-GGAGGGTCCTTCTCCAGC-3Ј (Zϩ); 5Ј-GGTTCCTCACACAGCGAT-TAAG-3Ј (Wϩ); 5Ј-GCCCTGACCATCCAGCTGATC-3Ј (Pϩ); and 3Ј-TCTCCCCATCAATGACCACC-5Ј (Q-).
Transient Transfections and CAT Assays-168FAR, 4T1, and MCF-7 cells were transfected by the calcium phosphate procedure (25). One day prior to transfection, 5 ϫ 10 5 cells were seeded in 60-mm dishes in phenol red-free medium supplemented with 10% fetal bovine serum or fetal calf serum as discussed above. Cells were fed with fresh phenol red-free medium 3 h prior to transfection with 10 g of the appropriate recombinant or empty plasmid DNAs. When ␤-galactosidase expression was used as an internal control, 5 g of plasmid pSV␤-gal (Promega Corp, Madison, WI) was added. Following 6 h of transfection, medium was removed, and cells were shocked with 10% glycerol. After washes with phosphate buffered saline, cells were grown in phenol red-free media supplemented with E 2 (0.001-10 nM; Sigma), dexamethasone (0.01-1000 nM; Sigma), or vehicle (ethanol to final concentration of 0.01%, v/v). The specificity of the hormone-induced effects were tested by including the inhibitors, ICI 182,780 (a gift from Dr. A. Wakeling, Zeneca Pharmaceuticals, Cheshire, UK) or RU486 (Center de Recherche Roussel-Uclaf, Romainville, France) at 100-fold molar excess of E 2 or dexamethasone, respectively. Cells were harvested 48 h later for measurement of CAT activity.
CAT assays were done as described by Gorman et al. (26). ␤-Galactosidase activity was measured according to the manufacturer's directions. For quantitation, the acetylated and nonacetylated spots were scraped from the TLC plates, and radioactivity was measured by scintillation counting. Results are expressed as fold induction produced by E 2 or dexamethasone relative to control based on acetylated chloramphenicol formed per unit protein or per unit ␤-galactosidase activity. There was no significant difference in induction when radioactivity was normalized by protein or ␤-galactosidase activity.
Hybridization of Run-on Transcripts to Filter-bound cDNAs-10 g of cDNAs for c-Ha-ras and ␤-actin prepared as described above were denatured by 0.4 M NaOH treatment for 30 min at room temperature, neutralized with 2 M ammonium acetate, pH 7.5, and applied to Magna Charge nylon membranes (Microseparations Inc., Westborough, MA) using a slot blot apparatus (Bio-Rad). Membranes were prehybridized for 2 h at 42°C in 50% formamide (v/v), 6ϫ SSC, 5ϫ Denhardt's solution, 0.5% SDS, 100 g/ml yeast tRNA, and 100 g/ml denatured salmon sperm DNA and then hybridized with 5 ϫ 10 6 cpm of 32 P-labeled RNAs at 65°C for 3 days in prehybridization buffer. Membranes were extensively washed with increasing stringency and subjected to autoradiography. Several different exposure times were chosen to obtain densitometric scans in the linear response range of the x-ray film. The autoradiograms were analyzed by measurement of band densities by scanner-densitometer (model 300A, Molecular Dynamics, Sunnyvale, CA). Ha-ras band densities were expressed as a ratio to ␤-actin.
Measurement of mRNA Stability-168FAR cells were grown in phenol red-free medium as described above. Cells were treated with vehicle (0.01% ethanol, v/v), 1 nM E 2 , or a combination of 1 nM E 2 and 100-fold molar excess of pure antiestrogen, ICI 182,780, for 6 h. Media were then removed, rinsed, and replaced with fresh medium or medium containing actinomycin D (5 g/ml) or cycloheximide (10 g/ml) and incubated for another 6 h. Total RNA was isolated and analyzed by slot blot method. 10 g of total RNA was denatured by heating at 65°C in 2.2 M formaldehyde and applied on Magna Charge nylon transfer membrane (Micron Separations, Inc., Westborough, MA) by vacuum filtration using a slot blot apparatus (Bio-Rad). The integrity and specificity of Ha-ras mRNAs in the samples were tested by Northern blotting. Filters were hybridized overnight at 42°C with random primed 32 P-labeled c-Ha-ras cDNA in buffer containing 6ϫ SSC, 50% formamide, and 0.5% SDS. Filters were washed with 1ϫ SSC and subjected to autoradiography. Loading of RNA was monitored by rehybridizing stripped membranes to 32 P-labeled GAPDH probe. Amounts of c-Ha-ras mRNA relative to GAPDH bands were quantitated with a scanner-densitometer (model 300A, Molecular Dynamics, Sunnyvale CA).
Metabolic Labeling and Immunoprecipitation-Exponentially growing 168FAR or 4T1 cells (5 ϫ 10 6 cells/100-mm dish) were incubated in methionine-and phenol red-free Dulbecco's modified Eagle's medium/2% dialyzed fetal bovine serum supplemented with vehicle, 1 nM E 2 , or a combination of 1 nM E 2 plus 100-fold molar excess of ICI 182,780 and 100 Ci of [ 35 S]methionine (specific activity 1083 Ci/mmol; NEN Life Science Products). Cells were labeled for 12 h, after which the monolayers were gently rinsed twice with phosphate-buffered saline and lysed with lysis buffer (10 mM Tris-HCl, pH 7.5, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride) at 4°C. Cell lysates were cleared by centrifugation for 15 min at 1,000 ϫ g and used immediately for immunoprecipitation of c-Ha-ras protein. Lysate volumes containing equivalent amounts of radiolabel incorporated into trichloroacetic acid insoluble material (ϳ10 7 cpm) were immunoprecipitated by incubating overnight at 4°C with 1 g of mouse monoclonal anti c-Ha-ras antibody (clone F235-1.7.1; Oncogene Science, Cambridge, MA) or normal mouse IgG. Protein A/G Sepharose (Oncogene Science) was then added to the lysates and incubated for 1 h at 4°C. Sepharose beads were pelleted by centrifugation and washed four times in lysis buffer. Bound proteins were solubilized in SDS buffer and subjected to SDS-polyacrylamide gel electrophoresis on 12.5% polyacrylamide gels. After electrophoretic separation, gels were processed for fluorography and subjected to autoradiography at Ϫ70°C. Amounts of c-Ha-ras protein bands were quantitated with a scanner-densitometer as described above.
Gel Retardation Assay-Nuclear extracts were made by the procedure of Dignam et al. (28) from estrogen-dependent MCF-7 human breast cancer cells. Protein concentration was determined by the method of Bradford (Bradford protein assay; Bio-Rad). Three pairs of complimentary oligonucleotides were used for DNA binding assays: (i) wild type intron-1, 5Ј-GTTATGGGGTATGATCCATCAGGGTATCAG-3Ј, which contains the palindromic sequence (underlined region) found between nucleotides 191 and 220 of the mouse c-Ha-ras gene; (ii) exon 1 sequence, 5Ј-GTGCGCTGACCATCCAGCTGATCCAGAACC-3Ј, found between nucleotides 1713 and 1742 of the human c-Ha-ras gene (23); and (iii) NF-1, 5Ј-TGGGGCTTGGTCATGGGCCATCAGCGCATG-3Ј, a putative NF-1 binding element occurring in hepatitis B virus enhancer region (nucleotides 1209 -1238; Ref. 29) to test specificity of binding. This oligonucleotide was chosen because the underlined sequence has 80% homology to the palindrome in the ras wild type intron-1. The oligonucleotides were annealed and end-labeled with [␥-32 P]ATP using polynucleotide kinase. Binding reactions were carried out in 20 l containing 20,000 -25,000 cpm (0.5 ng) of 32 P-labeled double-stranded human exon 1 DNA fragment, 2 g of the synthetic copolymer poly(dI-dC)⅐(dI-dC) (Roche Molecular Biochemicals) in binding buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM dithiothreitol, 12% glycerol) and 2.5 g of MCF-7 nuclear proteins for 20 min at room temperature. The resulting DNA-protein complexes were resolved from the free probe by electrophoresis on a 7% nondenaturing polyacrylamide gel and visualized by autoradiography. In competition experiments, indicated amounts of unlabeled double-stranded competitor (human exon 1 motif, wild type mouse Ha-ras intron 1 element or NF-1 sequence) were added together with the labeled human exon 1 probe. For supershift experiments, nuclear extracts were preincubated for 30 min in binding buffer with 0.5 g of ER antibodies or equivalent amounts of rat or mouse preimmune IgG before the addition of the labeled double-stranded oligonucleotide. The following ER antibodies were used: rat monoclonal ER antibody H222 and D75 (a generous gift from Dr. Geoffrey Greene) and mouse monoclonal ER antibody clone 33 (Affinity BioReagents, Neshanic Station, NJ). The gels were dried and autoradiographed with intensifying screens at Ϫ80°C.

Estrogen-and Dexamethasone-responsive Transcription Regulating Sequences Are Located in Exon 1 and Intron 1 of the Mouse c-Ha-ras
Gene-Previous studies have identified the presence of functional GREs and a putative ERE in the distal promoter region of the murine Ha-ras gene at nucleotides Ϫ1506, Ϫ1655, and Ϫ1420, respectively (13). Although the GREs conferred responsiveness to dexamethasone in transient reporter gene expression assays, the putative ERE failed to confer estrogen inducibility to the Ha-ras promoter in estrogendependent MCF-7 human breast cancer cells (13). We have previously reported identification of a transcription enhancing palindromic sequence in the intron 1 of the mouse c-Ha-ras gene that confers dexamethasone responsiveness to CAT reporter gene (20). In this study, we have examined the functional relevance of this sequence in the regulation of c-Ha-ras gene expression in two genetically related mouse mammary tumor subpopulations that differ in metastatic potentials (22) and Ha-ras mRNA and protein expression levels (30). We constructed reporter plasmids containing the minimal 30-bp ras intron 1 element (ϩ191/ϩ220) with (pmutras) and without (pwtras) the naturally occurring point mutation (A 3 G at nucleotide 210) identified in 4T1 subline (Fig. 1). Reporter plasmids containing the entire intron 1 and various lengths of 5Ј sequences, pYQras (Ϫ686/ϩ355), pZQras (Ϫ545/ϩ355), pWQras (Ϫ246/ϩ355), and pPQras (ϩ51/ϩ355) were also constructed (Fig. 1). All these sequences were ligated upstream of the Ϫ37 TK promoter, which in turn was linked to the bacterial CAT gene and the SV-40 poly(A) signal.
When pwtras was transfected into 168FAR cells, a 20-fold induction of CAT activity over that of control transfected with the empty vector, p-37TKCAT was observed (Table I). Because the ras intron 1 element contains the motif, TGATCC, which is identical to the half-site of GRE found in the chicken vitellogenin gene (31) and bears homology to the consensus ERE halfsite, TGACC (32), we challenged pwtras-transfected 168FAR cells with dexamethasone or E 2 . Treatment of pwtras-transfected 168FAR cells with 0.1 or 1 M dexamethasone significantly induced CAT activity ϳ3and 4-fold, respectively, over cultures treated with vehicle (Table I). In contrast, cells transfected with pmutras were not induced by similar concentrations of dexamethasone (Table I). These data suggest that the naturally occurring point mutation at base 210 that disrupts palindromic symmetry of the 10-bp sequence in ras intron 1 selectively abolishes induction of reporter gene expression by dexamethasone (Table I and Fig. 1). When 168FAR cells transfected with pwtras or pmutras were challenged with E 2 , there was no induction of CAT activity, indicating that minimal ras intron 1 element confers only dexamethasone inducibility (Table I). The observed induction of CAT activity was specific for dexamethasone because inclusion of 100-fold molar excess of RU486 blocks dexamethasone-induced responses ( Table I).
Examination of exon 1 and intron 1 of the mouse c-Ha-ras gene revealed the presence of one consensus ERE half-site, TGACC, at base 55 in exon 1, and two consensus ERE half-sites at bases 150 (GGTCA) and 163 (TGACC) in intron 1. The latter two ERE half-sites probably constitute a complete ERE sequence and differs from the consensus ERE (GGTCANNNT-GACC) by having five extra bases in the spacer separating the two half-sites. Two GRE half-site motifs (TGATCC) are located at bases 67 and 95 in exon 1, besides the functional GRE motif (base 201-206) found in the 30-bp ras intron 1 element present in the pwtras and pmutras reporter plasmids (Fig. 1). Because half-palindromic ERE motifs have been shown to act synergistically in conferring estrogen inducibility to proximal ovalbumin gene promoter and to heterologous promoters (33), we tested the functionality of these putative motifs in regulation of CAT activity. 168FAR cells were transfected with pPQras construct (with and without the point mutation at base 210), which contains the entire intron 1 (base 111-320), and portions of exon 1 (base 51-110) and exon 2 (base 321-355). When wild type or mutant pPQras-transfected 168FAR cells were exposed to 10 nM E 2 , a dose-dependent induction of CAT activity that was ϳ4.5and 3.5-fold, respectively, higher than vehicletreated cells was observed ( Fig. 2A). Exposure of wild type pPQras-transfected 168FAR cells to dexamethasone caused a dose-dependent induction of CAT activity that was ϳ4-fold higher with 1 M dexamethasone than vehicle-treated cells (Fig. 2B). However, similar exposure of mutant pPQras-transfected 168FAR cells to dexamethasone failed to stimulate an increase in CAT gene expression (Fig. 2B). The regulatory effects of E 2 and dexamethasone on CAT gene expression were completely abolished with the estrogen antagonist ICI 182,780 and glucocorticoid antagonist, RU486, respectively (Fig. 2).
Estrogen-and Glucocorticoid-inducible Motifs Located in Exon 1 and Intron 1 of the Murine c-Ha-ras Gene Are Functional in Context with the 5Ј-Flanking Sequences-Because the results of pPQras transfection showed the presence of functional ERE and GRE motifs that are capable of conferring estrogen-and dexamethasone-mediated transcriptional activation of the CAT reporter gene, it was of interest to determine whether the ERE and GRE half-sites identified in exon 1 and intron 1 of the Ha-ras gene can support hormone-induced transcription when placed in the context of 5Ј upstream sequences. Reporter constructs (pYQras, Ϫ686/ϩ355; pZQras, Ϫ545/ϩ355; pWQras, Ϫ246/ϩ355; Fig. 1) containing the entire intron 1 but lacking various lengths of proximal upstream sequences were transfected into 168FAR or 4T1 cells and treated with E 2 ,  1. Schematic representation of the 5-flanking region, first coding exon, and intron 1 of the mouse c-Ha-ras gene. The numbering represents the distance in base pairs from the A of the initiation codon (19). The position of the various sequence motifs, SV40 core enhancer (E), AP1/ATF (Ⅺ), ERE (ࡗ), and GRE (q) half-sites are indicated. Also shown is the sequence for the 30-bp ras intron 1 element containing the consensus GRE half-site (underlined) within the 10-bp palindrome. Presence of the naturally occurring point mutation resulting in A 3 G substitution at base 210 is indicated by ૺ. Position of primers Y, Z, W, P, and Q that were used for amplification and construction of pYQras, pZQras, pWQras, and pPQras-TK-CAT reporter plasmids are also shown. dexamethasone or a combination of E 2 and ICI 182,780, or dexamethasone and RU-486. A dose-dependent induction of CAT expression by 1 and 10 nM E 2 was observed from YQras, ZQras, and WQras DNAs ( Fig. 2A). This E 2 -stimulated increase, ϳ3-fold over control, was similar to that observed with pPQras plasmid (Fig. 2). Treatment of transfected cells with combination of 10 nM E 2 and 100-fold molar excess of ICI 182,780 significantly inhibited the E 2 -induced increase in CAT activity, indicating that E 2 -mediated stimulation occurred via the E 2 -ER pathway. Similarly, measurement of dexamethasone-induced CAT activity from pYQras, pZQras, and pWQras showed dose-dependent stimulation of CAT activity that was comparable to that observed with the pPQras plasmid (Fig.  2B), and addition of 100-fold molar excess of the glucocorticoid antagonist RU486 blocked the dexamethasone-stimulated increases in CAT activity (Fig. 2B). Because the magnitude of CAT activity induced from pYQras, pZQras, or pWQras under the influence of E 2 or dexamethasone is similar to that in cells transfected with the pPQras construct, these results indicate that activities of functional estrogen and glucocorticoid-responsive motifs in exon 1 and intron 1 of the murine c-Ha-ras gene are uninfluenced by 5Ј-flanking sequences. Results of CAT gene expression and regulation by estrogen and dexamethasone in 4T1 cells are similar to those observed in 168FAR cells and hence are not shown.

Sequences in Exon 1 of the Human c-Ha-ras Gene Bearing Homology to the Mouse Ha-ras Intron 1 Element Mediate Estrogen and Glucocorticoid-responsive Transcriptional En-
hancer Activity-To determine whether the sequence in exon 1 of the Ha-ras gene possesses transcriptional regulatory activity similar to the intron 1 element with which it bears 80% homology, we constructed reporter plasmid containing the 30-bp human Ha-ras exon 1 sequence (base 1713-1742) that was ligated upstream of the minimal Ϫ37 TK promoter, which in turn was linked to the bacterial CAT gene and the SV-40 poly(A) signal. When ph-rasexon1-TK-CAT was transfected into estrogen-dependent ER-positive human breast cancer MCF-7 cells, a 15fold induction of CAT activity over that of control transfected with the empty vector, p-37TKCAT was observed (Fig. 3). Because the ras exon 1 sequence contains the motifs, TGATCC (which is identical to the half-site of GRE found in the chicken vitellogenin gene; Ref. 31) and TGACC (which is identical to the consensus ERE half-site, TGACC; Ref. 32), we exposed ph-rasexon1-TK-CAT-transfected MCF-7 cells with dexamethasone or E 2 . Treatment of ph-rasexon1-TK-CAT-transfected MCF-7 cells with 0.1 or 1 M dexamethasone significantly induced CAT activity ϳ2and 4-fold, respectively, over cultures treated with vehicle (Fig. 3). When MCF-7 cells transfected with ph-rasexon1-TK-CAT were challenged with E 2 , a 3-5-fold induction of CAT activity was observed with 1 and 10 nM E 2 , respectively (Fig. 3). The observed induction of CAT FIG. 2. Regulation of CAT gene expression by 17␤-estradiol or dexamethasone in pYQras, pZQras, pWQras, or pPQras transfected nonmetastatic 168FAR mammary tumor cells. Cells grown in phenol red-free medium were transfected with the appropriate reporter construct and treated for ϳ40 h with E 2 (A) or dexamethasone (DEX; B) over a concentration range of 0.1-10 nM or 0.01-1 M, respectively. Specificity of E 2 or DEX-mediated effects on CAT gene activation was tested by including 100-fold molar excess of ICI 182,780 or RU-486, respectively. Control cultures were exposed to vehicle (0.01% ethanol, v/v). Results obtained from three transfections are expressed as the means Ϯ S.D. Co-transfection with pSV␤-gal, which expresses ␤-galactosidase activity at a constant level regardless of E 2 or DEX concentration, confirmed that the results were not due to differences in transfection efficiency or cell viability (data not shown).

FIG. 3. Regulation of CAT gene expression by 17␤-estradiol or dexamethasone in phrasexon1-TK-CAT-transfected human breast cancer MCF-7 cells.
Cells grown in phenol red-free medium were transfected with hrasexon1-TK-CAT or empty Ϫ37TK-CAT reporter constructs and treated for ϳ40 h with 17␤-E 2 or dexamethasone (DEX) over a concentration range of 0.1-10 nM or 0.1 and 1 M, respectively. Specificity of E 2 or dexamethasone-mediated effects on CAT gene activation was tested by including 100-fold molar excess of ICI 182,780 or RU-486 (RU), respectively. Control cultures were exposed to vehicle (0.01% ethanol, v/v). Because no changes in CAT gene expression were observed in Ϫ37TK-CAT-transfected cells exposed to steroid hormones as compared with untreated, the results are grouped together. Results obtained from three transfections are expressed as the means Ϯ S.D. Co-transfection with pSV␤-gal that expresses ␤-galactosidase activity at a constant level regardless of E 2 or DEX concentration confirmed that the results were not due to differences in transfection efficiency or cell viability (data not shown). activity was specific for dexamethasone or E 2 because inclusion of 100-fold molar excess of RU486 or ICI 182,780 blocked dexamethasone-or E 2 -induced responses, respectively (Fig. 3).

DNA-Protein Complexes Formed with Ha-ras Exon 1 Transcriptional Regulatory Sequence Contain Estrogen Receptor-
Because the Ha-ras exon 1 sequence bears homology to the mouse ras intron 1 element and confers transcriptional enhancer activity to CAT gene expression in MCF-7 cells, we examined the ability of nuclear proteins from MCF-7 cells to bind the human Ha-ras exon 1 motif in electrophoretic mobility shift assays. Fig. 4A shows that when labeled human Ha-ras exon 1 motif was used as a probe, two DNA-protein complexes (indicated as complexes A and B) are observed that are specifically competed with 10 -100 fold molar excess of unlabeled human Ha-ras exon 1 motif (third, fourth, and fifth lanes). Inclusion of unlabeled mouse ras intron 1 element caused efficient blocking of complex B formed with labeled human Ha-ras exon 1 motif, whereas a substantial amount of complex A remained even after adding 100-fold molar excess of the unla-beled ras intron 1 element (Fig. 4A, sixth, seventh, and eighth  lanes). These results suggest that protein(s) involved in complex B formation have similar affinities for the exon 1 motif and intron 1 element. Both complexes are specific as neither of them are competed with 50 -250-fold excess (Fig. 4A, ninth, tenth, and eleventh lanes) of NF-1 (putative NF-1 binding element occurring in the hepatitis B virus enhancer region; Ref. 29), although this sequence has 80% homology to the 10-bp palindrome in the ras intron 1 element.
Because the human Ha-ras exon 1 motif conferred estrogen responsiveness to CAT gene expression in estrogen-dependent MCF-7 cells, we examined whether the complexes formed with Ha-ras exon 1 motif contained estrogen receptor. As shown in Fig. 4B, incubation with monoclonal ER antibodies, clone 33 (Affinity BioReagents) or D75 with epitopes directed to the DNA-binding domain of ER, selectively inhibited complex A formation with resultant appearance of a faster migrating band (complex C) that displayed enhanced DNA binding activity (lanes 6 and 7). Incubation of binding reaction with the ER antibody H222 with epitope directed to the hormone binding domain of ER caused selective supershifting of complex A (Fig.  4B, lane 5), whereas inclusion of rat or mouse preimmune IgG had no influence on both complexes (Fig. 4B, lanes 2 and 3,  respectively). These results indicate that the complex A formed by the human Ha-ras exon 1-motif contains ER. It is interesting to note that competition assays demonstrated similar affinities of exon 1 motif and intron 1 element for protein(s) in complex B and markedly reduced affinity of ER and/or other protein(s) constituting complex A for the ras intron 1 element. These data are consistent with the observation that transcriptional enhancer activity mediated by pwtras construct that contains the minimal 30-bp ras intron 1 element is unaffected by estradiol (Table I).
Estrogen Regulates Transcription of c-Ha-ras Gene in Nonmetastatic Mammary Tumor Subline-Because our data from Fig. 2 demonstrate the presence of sequences in exon 1 and/or intron 1 of the murine Ha-ras gene that are capable of conferring strong estrogen inducibility to the CAT reporter gene, we tested whether estrogen can directly influence transcription of the murine c-Ha-ras gene in metastatic and nonmetastatic mammary tumor variants. In vitro transcript elongation or nuclear run-on assays were performed using purified nuclei prepared from 168FAR and 4T1 cells treated with vehicle (0.01% ethanol, v/v) or E 2 . When the amount of 32 P-labeled nascent RNA complementary to Ha-ras cDNA was compared with that hybridizing with ␤-actin cDNA, the transcription rate of c-Ha-ras gene in 168FAR cells more than doubled (relative to control) by 2 h after treatment with 1 nM E 2 (Fig. 5). Transcriptional activity increased to about five times control (vehicletreated cells) by 6 h (Fig. 5). Levels of Ha-ras transcription decreased gradually but remained elevated at 12 h and returned to control levels by 24 h of treatment with E 2 (Fig. 5). Inclusion of ICI 182,780 significantly blocked the E 2 -induced increase in c-Ha-ras gene transcription confirming that E 2mediated induction of Ha-ras transcription in nonmetastatic 168FAR cells resulted via the E 2 -ER pathway. The Ha-ras increase was not associated with changes in transcription rates of ␤-actin. A similar analysis of Ha-ras transcription rates in metastatic 4T1 cells showed the presence of ϳ20-fold higher basal levels of Ha-ras transcripts in control 4T1 cells as compared with control nonmetastatic 168FAR cells (Fig. 5). However, unlike 168FAR cells where E 2 exerts a stimulatory effect on the kinetics of Ha-ras transcription, estrogen failed to significantly influence the high basal levels of Ha-ras transcription observed in 4T1 cells (Fig. 5). Addition of the pure antiestrogen ICI 182,780 had no effect on Ha-ras transcription confirming the absence of E 2 /ER-mediated effects on Ha-ras transcription in 4T1 cells (Fig. 5). These data imply major differences in transcriptional regulation of the Ha-ras gene in 168FAR and 4T1 sublines, i.e. specific induction of Ha-ras transcription by estrogen in nonmetastatic 168FAR cells that otherwise transcribe very low basal levels until stimulated by estrogen versus constitutive, estrogen nonresponsive overexpression of Ha-ras transcripts in metastatic 4T1 cells.
Estrogen Up-regulates c-Ha-ras Protein Levels in Nonmetastatic 168FAR Subline-To determine whether effects of estrogen on Ha-ras protein expression correlate with estrogen induction profiles determined by nuclear run-on assays, 168FAR or 4T1 cells treated with 1 nM E 2 , 1 nM E 2 plus 100-fold molar excess of ICI 182,780 or vehicle were metabolically labeled with [ 35 S]methionine and analyzed for c-Ha-ras protein expression by immunoprecipitation with anti c-Ha-ras antibody. Control immunoprecipitations were performed with equivalent amounts of normal IgG. Nonimmune IgG did not precipitate 21-kDa proteins from the lysates of 168FAR or 4T1 cells (Fig. 6,  lanes 1 and 1Ј). Consistent with Ha-ras transcription rates determined by nuclear run-on assays, low levels of c-Ha-ras protein were detected in control 168FAR cultures that were enhanced ϳ12-fold following treatment with estradiol (Fig. 6,   compares lanes 2 and 3), and addition of estrogen antagonist effectively blocked the estrogen-mediated increase in Ha-ras protein levels (Fig. 6, lane 4). In contrast to estrogen-inducible expression of Ha-ras protein in 168FAR cells, similar analysis of Ha-ras protein in the metastatic 4T1 subline revealed ϳ7fold higher levels of Ha-ras protein relative to control 168FAR cells and demonstrated a lack of modulation by E 2 or ICI 182,780 (Fig. 6, lanes 2Ј-4Ј). Fig. 5 show that E 2 directly influences the rate of Ha-ras transcription in nonmetastatic 168FAR cells. To further characterize the effects of estrogen on Ha-ras mRNA stability, 168FAR cells were treated with vehicle or 1 nM E 2 for 6 h, and Ha-ras mRNA levels were measured 6 h later following exposure to actinomycin D or cycloheximide. Although, E 2 elicited only a 1.5-fold increase in steady-state level of Ha-ras mRNA over vehicle-treated control cultures (Fig. 7, second and third lanes), this difference was greatly enhanced to ϳ7-fold upon addition of a 100-fold molar excess of estrogen antagonist ICI 182,780 (Fig. 7, compare second and seventh lanes). Because addition of ICI 182,780 reduced Ha-ras mRNA levels to ϳ75% of that observed in vehicle-treated cells (Fig. 7, first and seventh lanes), these data imply that a portion of Ha-ras mRNA expressed in control cultures may be a result of stimulation by contaminating estrogen in the culture medium.

Up-regulation of c-Ha-ras Gene Expression by Estrogen and Modulation by Actinomycin D-Results of
To determine whether the estrogen effect on Ha-ras expression was at the transcriptional level, E 2 -treated cells were cultured with the transcriptional inhibitor actinomycin D at a concentration of 5 g/ml. Levels of Ha-ras mRNA dropped ϳ8-fold in cultures exposed to actinomycin D when compared with corresponding E 2 -treated cultures (Fig. 7, second and  third lanes). Similarly, actinomycin D repressed the basal expression of Ha-ras mRNA (Fig. 7, fifth lane). These data suggest that Ha-ras mRNA expression by E 2 is transcriptionally regulated. We next tested the effect of cycloheximide on the maintenance of E 2 -induced increase in Ha-ras mRNA levels in 168FAR cells. When 168FAR cells were cultured in the presence of cycloheximide (10 g/ml) after exposure to E 2 , no difference in levels of Ha-ras mRNA was observed when ratios of Ha-ras to GAPDH mRNAs were compared between cultures exposed to E 2 alone or in the presence of cycloheximide (Fig. 7,  second and fourth lane). Similarly, cycloheximide alone had no FIG. 5. Effect of 17␤-estradiol on c-Ha-ras gene transcription in nonmetastatic 168FAR and metastatic 4T1 mammary tumor cells; nuclear transcript elongation (run-on) assay. Cells grown in phenol red-free medium were incubated with medium in the presence of vehicle (0.01% ethanol, v/v), 1 nM E 2 , or a combination of 1 nM E 2 plus 100-fold molar excess of estrogen antagonist ICI 182,780. Treatment with E 2 was carried out for 2, 6, 12, or 24 h. Cultures were exposed to vehicle (control) or inhibitor for 6 h. Nuclei were isolated at the indicated time points, and nuclear run-on reactions were performed as described under "Experimental Procedures." 32 P-Labeled nascent RNAs isolated from run-on reactions were hybridized to immobilized probes on the membrane. c-Ha-ras band densities were quantified by densitometry and expressed as a ratio to ␤-actin; the relative transcriptional rates of c-Ha-ras gene in 168FAR (filled bar) and 4T1 (shaded bar) cells are shown graphically in panel B. effect on base-line Ha-ras mRNA levels (Fig. 7, compare first and sixth lanes), suggesting that de novo protein synthesis is not required for E 2 -dependent Ha-ras mRNA induction.

DISCUSSION
Ha-ras genes are rendered oncogenic either by mutation (23, 34 -36) or by overexpression (37). Using a mouse mammary tumor model consisting of genetically related sister sublines with variant metastatic capacities, we have previously shown a direct correlation between metastatic behavior and expression levels of normal Ha-ras mRNA and protein (30). Although ras mutations are infrequent, occurring only in about 5% of human breast cancers, there is considerable evidence to suggest that the pathways regulated by ras are deregulated in breast cancer cells (38). Elevated levels of normal Ha-ras have been shown to play a crucial role in tumorigenesis. 50% of human breast carcinomas express elevated levels of Ha-ras (6 -8). Thus, it is possible that the aberrant function of ras or ras related proteins may contribute to breast cancer development and/or progression. Overexpression of Ha-ras gene has been postulated to result from transcriptional deregulation (10). This is the first report that demonstrates existence of estrogen-mediated regulation of Ha-ras transcription in mammary tumor cells.
We have previously identified the presence of a glucocorticoid-responsive transcription enhancing element in intron 1 of the mouse Ha-ras gene. This transcriptional regulatory element is a palindromic sequence that contains the consensus GRE half-site (20). Identification of this intron 1 regulatory element has been facilitated by presence of a point mutation at the 3Ј end of the palindrome in the Ha-ras gene of the metastatic 4T1 subline (20). Although we do not yet know the exact role of this point mutation in regulation of Ha-ras gene expression, our data from transient reporter assays utilizing constructs containing only the ras intron 1 element or entire intron 1 show that the presence of this point mutation selectively abolishes transcriptional regulation by dexamethasone.
Previous functional analyses of murine and human Ha-ras genes have been limited to the promoter and intron 0 regions. The 5Ј region of the Ha-ras gene is highly conserved in the mouse, rat, and human (11,12,19). In the mouse c-Ha-ras gene, the intron 0 has been shown to contain a positive regulatory element (21). Although no obvious enhancer sequences were detected within the intron 0 of the mouse Ha-ras gene, a 12-bp motif closely matching the SV40 enhancer core element was found at position Ϫ620 (19). Within this region several copies of a sequence motif that is similar to the elements that bind AP1 and ATF are found (19). These features imply that intron 0 of the mouse Ha-ras gene may play an important part in transcriptional regulation of Ha-ras. However, our results from reporter expression assays indicate that sequences in intron 0 are not responsible for hormone-mediated transcriptional enhancer activity, because the reporter plasmid pPQras does not contain this region, yet demonstrates similar levels of E 2 -and dexamethasone-induced CAT activities as do pYQras, pZQras, or pWQras reporter plasmids that do contain intron 0. Our data show that intron 1 may play a significant role in hormone-mediated transcriptional regulation of Ha-ras gene because it can function independently or in the context of 5Ј upstream sequences. The importance of ras intron 1 element in regulation of Ha-ras gene expression is further strengthened by the fact that similar sequences are found in the mouse, rat, and human Ha-ras genes. Sequences bearing 100% (TGATCCT-GATCCATCA, ϩ373/ϩ388) and 80% (TGATCCATGC, ϩ1535/ ϩ1544 and ϩ1599/ϩ1608) identity to the ras intron 1 element are located in introns 1 and 3, respectively, of the rat Ha-ras gene (39). Two motifs bearing 80% identity to the 10-bp palindrome of ras intron 1 element are located in exon 1 of the mouse, rat, and human c-Ha-ras genes.  (19). Although the functional significance of this conserved sequence in regulation of human Ha-ras gene expression is not yet established, our data indicate that this sequence not only has transcriptional enhancer activity but is also capable of conferring estrogen responsiveness to CAT reporter gene expression. This is supported by demonstration of its ability to physically interact with ER in nuclear extracts of MCF-7 human breast cancer cells by gel retardation assays. Because the pPQras reporter plasmid includes the portion of exon 1 containing this sequence, estrogen-inducible transcriptional enhancer activity resulting from pPQras plasmid may be contributed at least in part by these hormonal response half-sites. Also, because the presence of a point mutation in intron 1 element selectively abolishes responsiveness to dexamethasone while maintaining estrogen regulation, sequences in ras intron 1 element, and not in exon 1, are required for dexamethasone-mediated stimulation of CAT activity.
Estrogen is known to be a key requirement for the normal development of the mammary gland. Although some reports have shown estrogen to affect Ha-ras expression (40 -46), there are reports contradicting such effects (47). Our data clearly indicate major differences in estrogen-mediated regulation of Ha-ras gene transcription in two genetically related mammary tumor sublines; thus the demonstration of a direct effect or lack thereof appears to be influenced by cell type. Regression of MCF-7 tumors in nude mice following estrogen ablation has been shown to be accompanied by a decrease in expression of c-Ha-ras, c-fos, and pS2 (48). Similarly, acquisition of an activated Ha-ras gene has been shown to confer hormone autonomy on the previously estrogendependent tumorigenicity of MCF-7 cells and causes up-regulation in secretion of growth factors in amounts that are comparable with estradiol stimulation (49). Kumar et al. (50) have shown the presence of Ha-ras oncogenes (H and K) in normal mammary glands of prepubescent animals 2 weeks after nitrosome thylurea treatment, where they remain latent until exposure to estrogen occurs during sexual maturation. These studies demonstrate that the presence of ras oncogenes in the mammary gland of young animals is not sufficient to trigger neoplastic development but rather that the ras oncogenes need to cooperate with physiological processes required for sexual maturation to exert their neoplastic properties (50).
The effects of estrogen on Ha-ras gene expression have not been examined in detail. Sequence motifs that have 92 and 50% FIG. 7. Effect of actinomycin D or cycloheximide on 17␤-estradiol-mediated increase in c-Ha-ras mRNA levels in nonmetastatic 168FAR mammary tumor cells. Cells were incubated with the medium in the presence of 1 nM E 2 , 1 nM E 2 plus 100-fold molar excess ICI 182,780 or vehicle (0.01% ethanol, v/v) for 6 h. The media were removed and replaced with fresh medium or medium containing 5 g/ml actinomycin D or 10 g/ml cycloheximide and further incubated for 6 h. Total RNA was isolated and c-Ha-ras mRNA analyzed by slot blot method. Loading of RNA was monitored by rehybridization of stripped membranes to 32 P-labeled GAPDH cDNA probe. homology to the vitellogenin ERE have been identified in the promoters of Ha-ras gene of the mouse (13) and human (12), respectively; however, this putative ERE failed to confer estrogen inducibility to the Ha-ras promoter (13). Our results from transient CAT gene expression assays clearly show that ERE half-sites (ϩ58/ϩ62, ϩ150/ϩ154, and ϩ163/ϩ167) present in pPQras and ph-rasexon 1 constructs are functional and respond to physiological concentrations of E 2 via ER as these responses are blocked by the pure antiestrogen, ICI 182,780. Although these sequences do not contain a complete ERE, ERE half-sites separated from each other by more than 100 bp have been shown to confer estrogen inducibility by acting synergistically on the proximal ovalbumin gene promoter or heterologous promoters (33). It is interesting to note that both introns 1 and 3 of the rat Ha-ras gene also contain ERE half-sites in the immediate vicinity of the sequence resembling the mouse ras intron 1 element.
To determine the primary site at which E 2 acts to control expression of Ha-ras gene, we measured transcription rates by nuclear run-on assays. A comparative analysis of Ha-ras transcription rates in nonmetastatic and metastatic mammary tumor sublines revealed fundamental differences not only in basal transcription rates but also in their ability to be modulated by estrogen. Estrogen enhances Ha-ras gene expression in nonmetastatic 168FAR cells at transcriptional levels, which is not dependent on new protein synthesis. Although both sublines express ER, expression of Ha-ras mRNA and protein is unaffected by estrogen or its antagonist in the metastatic subline. In the treatment of ER-positive breast cancers, intrinsic or secondary hormone resistance is a major clinical problem (51). This is reflected both in treatment of advanced disease, where the development of resistance to tamoxifen and other endocrine agents is inevitable, and in the adjuvant setting, where tamoxifen prevents some but not all relapses. A molecular understanding of this phenomenon is important to help identify those tumors that despite being ERϩ are intrinsically endocrineresistant. The fundamental role of Ha-ras in signal transduction pathways argues that processes that cause deregulation of Ha-ras gene expression may play an important role in breast cancer development and progression. Alteration(s) in transcriptional regulation of Ha-ras gene resulting in constitutive overexpression of Ha-ras may represent one such mechanism of antiestrogen resistance. Overexpression of ras proteins has been associated with resistance to chemotherapeutic agents and radiation (52-54). NIH3T3 cells transfected with normal or mutant c-Ha-ras oncogene are significantly more resistant to chemotherapeutic agents than normal cells (52). Although the detailed mechanisms responsible for these resistance phenomenona are not entirely clear, accumulating evidence indicate that ras-induced transactivation of AP-1 proteins, c-jun, c-fos, and Fra-1 (55, 56) may play a role. Experiments are in progress to correlate Ha-ras gene transcription with ER status, AP-1 levels, activity, composition, and regulation by estrogen and antiestrogens in human breast tumors.
Our data from transient in vitro assays have revealed the presence of estrogen-responsive elements in exon 1/intron 1 of the Ha-ras gene. However, definition of the exact role of exon 1/intron 1 motifs in in vivo regulation of the endogenous gene by estrogen requires a complete mapping of the Ha-ras gene by DNase I hypersensitivity or genomic footprinting assays. Because sequences bearing similarity to the exon 1/intron 1 motifs are present elsewhere in the gene, it is possible that these regulatory elements may function independently or in concert with those identified in exon 1/intron 1. However, it is important to note that in vivo regulation of Ha-ras gene expression by estrogen may be more complex and may involve other reg-ulatory elements and/or transcription activating factors. In conclusion, our data suggest that alterations in transcriptional regulation of the Ha-ras gene by estrogen may play an important role in progression of breast cancer.