The breast cancer resistance protein (BCRP) is an ATP-binding cassette half transporter that confers resistance to anticancer drugs such as mitoxantrone, anthracyclines, topotecan, and SN-38. Initial characterization of the BCRP promoter revealed that it is TATA-less with 5 putative Sp1 sites downstream from a putative CpG island and several AP1 sites (K. J. Bailey-Dell et al., Biochim. Biophys. Acta, 1520: 234–241, 2001). Here, we examined the sequence of the 5′-flanking region of the BCRP gene and found a putative estrogen response element (ERE). We showed that estrogen enhanced the expression of BCRP mRNA in the estrogen receptor (ER)-positive T47D:A18 cells and PA-1 cells stably expressing ERα. In BCRP promoter-luciferase assays, sequential deletions of the BCRP promoter showed that the region between −243 and −115 is essential for the ER effect. Mutation of the ERE found within this region attenuated the estrogen response, whereas deletion of the site completely abrogated the estrogen effect. Furthermore, electrophoretic mobility shift assays revealed specific binding of ERα to the BCRP promoter through the identified ERE. Taken together, we provide evidence herein for a novel ERE in the BCRP promoter.
The breast cancer resistance protein (BCRP) belongs to the ATP-binding cassette transporter G family and is also known as ABCP, MXR, or ABCG2 (1, 2, 3) . BCRP-overexpressing cancer cells have been isolated from breast, colon, gastric, lung, or ovarian carcinomas (3, 4, 5, 6) where BCRP acts as an efflux pump, resulting in decreased intracellular concentrations of anticancer drugs. It was initially found to be associated with mitoxantrone resistance (2 , 3 , 5 , 7) and later found to confer resistance to other drugs such as topotecan, SN-38, anthracyclines, and flavopiridol (4 , 7, 8, 9) . More recently, methotrexate and its polyglutamates have also been named as BCRP substrates (10, 11, 12) . However, resistance to methotrexate occurred only in cells bearing the wild-type transporter, which contains an arginine at position 482. Cells that overexpress a mutated BCRP (R482T and R482G) remained sensitive to methotrexate (10, 11, 12) .
Among normal tissues, BCRP is expressed in the syncytiotrophoblasts of the placenta, in the epithelium of the small intestines and colon, in the liver canalicular membrane, in hematopoietic stem cells, and in the ducts and lobules of the breast (13 , 14) . Whereas the precise physiological role of BCRP is still unknown, its localization in these tissues suggests that it may play a protective role against toxic substances and metabolites by extruding them across the apical membrane. A number of endogenous substrates of BCRP have been named lately, and these include 17β-estradiol-17-(β-d-glucuronide), sulfated steroidal compounds such as dehydroepiandrosterone sulfate, and folic acid (10 , 15 , 16) . The identification of more physiological substrates may eventually provide insights into the functions of BCRP in the body.
To date, little is known about the molecular mechanisms controlling BCRP expression. The BCRP gene comprises 16 exons and 15 introns spanning 66 kb in length and is located on chromosome 4q22 (17) . Initial characterization of the BCRP promoter revealed that it is TATA-less with 5 putative Sp1 sites downstream from a putative CpG island and several AP1 sites (17) . The present study was undertaken with the goal of identifying factors responsible for regulating BCRP expression. Here we examined the sequence of the 5′-flanking region of the BCRP gene and found a novel estrogen response element (ERE) in this promoter. We showed that BCRP mRNA expression was induced by 17β-estradiol (E2) in estrogen receptor (ER)-positive cell lines and that this effect can be reversed by the antiestrogen ICI 182,780. We investigated the transcriptional activities of the 5′-flanking region of the BCRP gene and confirmed that the BCRP gene is a direct target of the E2/ER complex through the identified ERE.
Materials and Methods
Hormones and Chemicals.
E2 was purchased from Calbiochem (San Diego, CA) and ICI 182,780 was a gift from AstraZeneca Pharmaceuticals (Macclesfield, United Kingdom).
All of the cells were grown at 37°C in a humidified atmosphere containing 5% CO2. Cell culture reagents were purchased from BioWhittaker (Walkersville, MD) unless otherwise stated. The ER-positive human breast cancer cell line, T47D:A18, has been described before (18) and was maintained in RPMI 1640 containing fetal bovine serum (10%; Sigma, St. Louis, MO), penicillin (100 units/ml), streptomycin (100 μg/ml), and l-glutamine (2 mm). The ER-negative human ovarian cancer cell line, PA-1, was obtained from American Type Culture Collection (Manassas, VA) and maintained in DMEM containing fetal bovine serum (10%), penicillin (100 units/ml), streptomycin (100 μg/ml), and l-glutamine (2 mm). The PA-1+ER cell line was established by transfecting PA-1 cells with an expression vector for human ERα (pSG5-HEG0; a gift from Dr. Pierre Chambon, CNRS/INSERM/ULP, Illkirch Cedex, France), and pcDNA3.1 in the ratio of 3:1 using Lipofectamine Plus (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Positive clones were maintained in G418 (2 mg/ml; Mediatech, Inc., Herndon, VA).
Plasmids and Cloning.
The human BCRP promoter and deletion fragments (−1285/+362, −243/+362, and −115/+362) were subcloned into a pGL3-basic reporter plasmid (Promega, Madison, WI) and were described previously (17) . The vectors −243mut1, −243mut2, −243mut3, and −243del were generated from −243/+362 using sequential PCR steps (19) and subcloned into the pGL3-basic reporter vector. All of the constructs were verified by sequencing. The vector, Triplet ERE-luciferase, containing three copies of the Xenopus vitellogenin A2 ERE upstream of the luciferase gene, was described before (20) .
RNA Isolation and Semiquantitative Reverse Transcription-PCR Analysis.
Total RNA was extracted from T47D:A18 and PA-1+ER cells using Trizol (Invitrogen). The first-strand cDNA was synthesized with 3 μg of the total RNA using the ThermoScript Reverse Transcription-PCR system (Invitrogen). Using one twentieth of the cDNAs as a template, the PCR was carried out for each transcript under these conditions, 20–35 cycles of 94°C 30 s, 60°C 60 s, and 72°C 80 s to determine the exponential phase of amplification. As an internal control, amplification of β-actin mRNA was carried out with similar cycle optimization. The following primer sets were used, sense, 5′-TTC TCC ATT CAT CAG CCT CG-3′; and antisense, 5′-TGG TTG GTC GTC AGG AAG A-3′ for BCRP, sense, 5′-TGG AGC AGA GAG GAG GCA A-3′; and antisense, 5′-GCC GAG CTC TGG GAC TAA TCA-3′ for pS2, and sense, GAG AAG ATG ACC CAG ATC ATG T-3′; and antisense, 5′-TCG TCA TAC TCC TGC TTG CAG-3′ for β-actin.
To examine the promoter activities in response to estrogenic stimulation, PA-1 cells were cultured in phenol red-free MEM (Invitrogen) supplemented with 10% charcoal/dextran-stripped fetal bovine serum (Gemini, Woodland, CA) for at least 3 days. PA-1 cells were grown at a density of 2 × 105 cells/well in six-well plates and cotransfected with 1.5 μg of the luciferase constructs, 0.25 μg hERα expression vector (pSG5-HEG0) and 0.5 μg β-galactosidase reporter plasmids by Lipofectamine Plus, according to the manufacturer’s protocol. After 16–18 h, fresh medium containing 10 nm E2 dissolved in ethanol or ethanol alone (0.1%) was added to the transfected cells. Twenty-four h later, the cells were harvested and analyzed for both the luciferase and β-galactosidase activities using assay kits from Promega. Relative luciferase activities were normalized with β-galactosidase activities. Each experiment was performed in triplicate and repeated at least twice.
Electrophoretic Mobility Shift Analysis.
Radiolabeled probes harboring the ERE of the BCRP gene were amplified using −243/+362 as a template in a standard PCR reaction to incorporate [α-32P]dCTP (Amersham Biosciences, Piscataway, NJ) into the probe. The binding reactions were carried out in the presence of 10 nm E2 with 5 pmol recombinant ERα protein (Affinity Bioreagents, Golden, CO) and 32P-labeled 110-bp probes (5 × 105cpm/reaction) in 20 μl of buffer [5 mm HEPES (pH 7.9) containing 100 mm NaCl, 0.25 mm EDTA, 0.25 mm EGTA, and 0.25 mm DTT]. Fifty μg/ml poly(deoxyinosinic-deoxycytidylic acid; Amersham Biosciences) was used as a nonspecific carrier DNA in the reaction. The mixture was placed on ice for 30 min. Protein-DNA complexes were separated from the free probe by nondenaturing 5% PAGE at 200 V with running buffer consisting of 25 mm Tris/25 mm borate/1 mm EDTA. Gels were dried, and the radioactive bands were visualized by autoradiography. For competition EMSA, a 100-fold excess of the unlabeled oligonucleotides harboring the ERE of the BCRP gene (BCRP-ERE) was added during the preincubation period. A 10–100-fold excess of unlabeled oligonucleotides with ERE deleted (BCRP-EREm) was also added in a separate reaction. Nonspecific unlabeled oligonucleotides that do not contain any known binding sequences were used as a negative control.
Identification of a Putative ERE in the BCRP Gene.
We examined the sequence of the 5′-flanking region of the BCRP gene and found a putative ERE in the promoter (Fig. 1A) ⇓ using our own computer software, PAGen@UIC, 4 to search for known DNA-binding protein recognition sites. Previous analyses of the 5′-flanking region of the human BCRP gene had not revealed any indication of an ERE (17) . This sequence is highly similar to the ERE found in genes for human coagulation Factor XII (21 , 22) and human c-fos (23) , as shown in Fig. 1B ⇓ . Although not perfectly palindromic as with most elements identified to date, at least 10 bp of the putative ERE of the BCRP gene are identical to the consensus ERE, in particular, to that found in human coagulation Factor XII (21 , 22) and human c-fos (23) . Furthermore, the imperfect palindrome is separated by exactly 3 bp, thereby retaining the length of the ERE critical for high affinity ER-ERE binding (24) .
Estrogen Treatment Increases BCRP mRNA Levels.
To determine whether BCRP mRNA levels are regulated by E2, we placed ER-positive T47D:A18 breast cancer cells in a medium containing charcoal/dextran-stripped serum for 3 days before treating them with 10 nm E2 for 24 h. Semiquantitative reverse transcription-PCR was performed and β-actin was used as a loading control. As shown in Fig. 2 ⇓ A, BCRP mRNA levels were increased significantly with E2 treatment. The estrogen-inducible gene, pS2, was used as a positive control and showed a similar increase. This effect was effectively abolished with concomitant treatment with a pure steroidal antiestrogen, ICI 182,780.
Next we asked if a similar trend would be observed in an ER-negative ovarian cancer cell line PA-1, stably transfected with ERα, PA-1+ER. As shown in Fig. 2 ⇓ B, BCRP mRNA levels were also increased in PA-1+ER cells with E2 treatment. As expected, no change was observed in the parental PA-1 cells after E2 treatment (data not shown).
BCRP Gene Contains a Functional ERE.
These results encouraged us to analyze whether BCRP promoter activity is E2-regulated. We transfected PA-1 cells with the full-length promoter-luciferase construct −1285/+362, together with the human ERα expression plasmid, and treated them with 10 nm E2 only, or with a combination of 10 nm E2 and ICI 182,780 at concentrations ranging from 0.01 to 1 μm. As shown in Fig. 3 ⇓ A, the promoter activity was increased up to 6.2-fold in the presence of E2 alone and decreased in a dose-dependent manner in the presence of ICI 182,780, reflecting a blockade of ER activity. As a positive control, we transfected the cells with a Triplet ERE-luciferase plasmid and found a dramatic increase in promoter activity with E2 treatment of 22-fold (Fig. 3B) ⇓ .
Deletion and Site-Directed Mutagenic Analysis of the Responsiveness of BCRP Promoter to Estrogen.
To identify the region necessary for the response to estrogen in the BCRP promoter, three fragments of the 5′-flanking region of the BCRP gene, which were cloned into the luciferase reporter vector, were transiently transfected into PA-1 cells. The luciferase activities of −1285/+362 and −243/+362 increased 6.1–6.6- fold with estrogen treatment, but the activity of −115/+362 did not increase significantly (Fig. 3C) ⇓ . This result indicates that the region that responds to estrogen treatment exists between positions −243 and −115 in the human BCRP promoter. To additionally confirm the involvement of the region between −243 and −115 in the induction of the BCRP promoter, we introduced mutations in the ERE sequence, transfected the resulting constructs in PA-1 cells, and determined the luciferase activities. As shown in Fig. 3 ⇓ D, the mutation in the left arm of the palindromic ERE decreased the luciferase activity to 3.8-fold, whereas that in the right arm resulted in a more dramatic decrease to 2.8-fold. When the mutations were introduced in both arms, the luciferase activity was decreased to 2.2-fold. More importantly, the deletion of the putative ERE completely abrogated the induction of activity with E2 treatment.
The ERE in the BCRP Promoter Binds ERα.
To determine whether the ERE in the BCRP gene can be recognized by ERα like a known ERE, electrophoretic mobility shift analyses were performed using purified recombinant ERα protein incubated with 10 nm E2 and radiolabeled oligonucleotides harboring the putative ERE of the BCRP gene. Fig. 4 ⇓ (Lane 2) shows efficient binding to the radiolabeled oligonucleotides. The binding was competed by adding a 100-fold excess of unlabeled ERE probe (Fig. 4 ⇓ , Lane 3). When a 10–100-fold excess of unlabeled oligonucleotides with the ERE deleted was added, no competition of binding occurred (Fig. 4 ⇓ , Lanes 4 and 5), indicating that binding only occurred with specific recognition of the ERE by the recombinant ERα protein. A nonspecific unlabeled competitor used also did not result in any changes in binding (Fig. 4 ⇓ , Lane 6). Collectively, these data show that the ERE of the BCRP gene behaves similarly as the consensus ERE and that it can bind ERα.
In this study, we found that BCRP is induced by nanomolar (physiological) concentrations of E2 at the transcriptional level, and we revealed the underlying molecular mechanism by identifying a functional ERE in the BCRP promoter. The consensus ERE contains a palindrome of PuGGTCA motifs separated by 3 bp (24 , 25) . These elements are bound by ER dimers, with one receptor interacting with each PuGGTCA motif. This ER dimer-motif interaction is highly specific, as consensus EREs are expected to be found only in every 4 million bp in random DNA sequences (25) . Stimulation of BCRP gene expression in response to E2 is likely to be mediated by the “classical” mechanism, whereby E2-liganded ER binds directly to the identified ERE, and interacts with coactivator proteins and components of the RNA polymerase II transcription machinery, resulting in enhanced transcription (26) .
BCRP is highly expressed in the placenta, or more specifically, in the placental syncytiotrophoblast, at the apical surface of the chorionic villus (14) , where steroid hormones such as estrogens are produced and could reach a maximum concentration of 150 nm (27) . The localization of BCRP in the placenta suggests that BCRP plays a protective role for the fetus by effluxing drugs and toxic metabolites that enter the placenta back into the maternal circulation. In fact, Jonker et al. (28) showed that the fetal penetration of topotecan is increased at least 2-fold by the BCRP inhibitor GF120918 in pregnant mdr1a/1b knockout mice. Our findings suggest for the first time that estrogens may augment this protective function, by inducing the endogenous expression of BCRP in the presence of ER.
Our findings also raise the possibility that estrogens can directly stimulate the production of this multidrug resistance transporter in hormone-responsive tumors clinically. We have shown the E2 induction of the BCRP gene expression in ER-positive breast and ovarian tumor cells and the reversal of this induction with a specific antiestrogen ICI 182,780. Intratumoral concentrations of estrogens are much higher than that in the serum and adjacent tissues (27) , and this reservoir may cause induction of the BCRP gene and potentiate multidrug resistance in these tumors. Conversely, clinical treatment of hormone-responsive tumors with antiestrogens may help to reduce the incidence of drug resistance. It will be of interest, therefore, to determine whether ERα analysis in tumor biopsies correlates with BCRP expression and drug resistance.
Most studies of estrogen action have focused on ERα because ERβ was discovered more recently (29) . Studies have shown that ERα and ERβ may respond differently to ER ligands in a cell-dependent and target gene context-dependent manner (30 , 31) . The above consideration raises the concern that in tumor cells expressing both ERα and ERβ, the latter may counteract ERα induction of BCRP. Studies examining the effect of ERβ on BCRP gene expression are ongoing.
The regulation of BCRP by estrogen is a complex issue. We have shown the influence of estrogen on BCRP gene expression. To date, several studies on the effect of estrogen on BCRP transport functions have been reported (15 , 32) . Imai et al. (15) found that free estrogens are not transported by BCRP, although they are able to block intravesicular estrone-3-sulfate uptake. Estrogen agonists and antagonists were found to increase cellular accumulation of topotecan in BCRP-overexpressing K562 cells (32) . It is important to note, however, that in the latter study the BCRP gene was driven by a constitutive (cytomegalovirus) promoter (33) and not its endogenous promoter. Hence, no conclusions can be made regarding the effect of estrogen on BCRP gene expression.
In conclusion, a novel imperfect palindromic ERE was found in the BCRP gene promoter, and E2 was shown to activate the promoter through the classical pathway that involves binding of E2/ER complexes on the ERE. Given the abundance of BCRP in the placenta, our findings could shed light on its physiological regulation. At the same time, our results may have a clinical impact, because repression of BCRP expression might be an important result of antiestrogen therapy.
We thank Dr. Kim Bailey for her contribution in making the BCRP-promoter luciferase reporter gene constructs and Dr. Yin-Yuan Mo for helpful discussion. We also thank Martina Vaskova for outstanding administrative support.
Grant support: National Cancer Institute Grants CA30103 and CA40570 (W. T.Beck), by a VA Merit Review Grant (D. D. Ross), and in part by The University of Illinois at Chicago.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: William Beck, Department of Biopharmaceutical Sciences (MC865), 833 South Wood Street, University of Illinois at Chicago, Chicago, IL 60612.
↵4 S. Kamalakaran, S. K. Radhakrishnan, and W. T. Beck, submitted for publication.
- Received November 14, 2003.
- Revision received December 26, 2003.
- Accepted January 6, 2004.
- ©2004 American Association for Cancer Research.