Estradiol Regulates Different Genes in Human Breast Tumor Xenografts Compared with the Identical Cells in Culture
http://www.100md.com
《内分泌学杂志》
Department of Medicine (D.M.E.H., J.K.R., C.A.S., K.B.H.), Division of Endocrinology
Department of Pathology (K.B.H.), University of Colorado Health Sciences Center at Fitzsimons, Aurora, Colorado 80045
Breast Center (D.C.A.), Baylor College of Medicine, Houston, Texas 77030
Abstract
In breast cancers, estrogen receptor (ER) levels are highly correlated with response to endocrine therapies. We sought to define mechanisms of estrogen (E) signaling in a solid breast tumor model using gene expression profiling. ER+ T47D-Y human breast cancer cells were grown as xenografts in ovariectomized nude mice under four conditions: 1) 17-estradiol for 8 wk (E); 2) without E for 8 wk (control); 3) E for 7 wk followed by 1 wk of E withdrawal (Ewd); or 4) E for 8 wk plus tamoxifen for the last week. E-regulated genes were defined as those that differed significantly between control and E and/or between E and Ewd or control and Ewd. These protocols generated 188 in vivo E-regulated genes that showed two major patterns of regulation. Approximately 46% returned to basal states after Ewd (class I genes); 53% did not (class II genes). In addition, more than 70% of class II-regulated genes also failed to reverse in response to tamoxifen. These genes may be interesting for the study of hormone-resistance issues. A subset of in vivo E-regulated genes appears on lists of clinical ER discriminator genes. These may be useful therapeutic targets or markers of E activity. Comparison of in vivo E-regulated genes with those regulated in identical cells in vitro after 6 and 24 h of E treatment demonstrate only 11% overlap. This indicates the extent to which gene expression profiles are uniquely dependent on hormone-treatment times and the cellular microenvironment.
Introduction
MORE THAN two thirds of human breast tumors express estrogen receptors (ERs) and are hormone responsive (1). Because of the importance of estrogens (E) and ER activities, multiple studies have sought to define E-regulated signaling pathways by assessment of individual genes or, more globally, by expression profiling (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16). Most of these studies have used established human breast cancer cell lines to compare E signaling in ER+ vs. ER– settings and to assess the functions of phytoestrogens, xenobiotics, or antiestrogens (2, 3, 4, 5, 7, 8, 9, 10, 12). These studies usually involve hormone-treatment times of 1–72 h using cells cultured in vitro. Generally, ER+ human breast cancer cells such as MCF-7 or ZR-75 are used, and these in vitro approaches, especially shorter treatment times of approximately 6 h, have yielded a number of novel and interesting direct E target genes (2, 3, 4, 5, 7, 8, 9, 12).
Independently, several recent studies have identified genes differentially expressed in ER+ compared with ER– solid tumors isolated from breast cancer patients (13, 14, 15, 16, 17, 18, 19, 20, 21). Collectively, the results of these studies have yielded lists of so-called ER discriminator genes. These lists, however, cannot address the question of whether the genes or transcription patterns identified are indeed controlled by, or even related to, E and ER signaling directly or whether they are simply correlative, reflecting other aspects of cancer biology such as tumor-differentiation state or growth rate that may be associated with ER without being ER regulated (22).
The purpose of the present study was to define mechanisms of E signaling in ER+ solid tumors using an in vivo approach. To do this, we used ER+ T47D human breast cancer cells (T47D-Y) grown as in vivo E-dependent solid tumor xenografts in nude mice (23). This allowed us to define E-regulated genes as ones that are differentially expressed in static, non-E-treated control tumors, compared with the 17-estradiol (E2)-treated proliferating tumors (E).
Additionally, to further refine the definition of in vivo E-regulated genes, we devised an E withdrawal (Ewd) protocol in which E2 was withdrawn for 1 wk in mice bearing E-dependent tumors, and the genes expressed in these tumors were analyzed. The Ewd data were, in turn, compared with 1 wk of E2 plus tamoxifen (Tam) (E + Tam) treatment after 7 wk of E2. One week of Ewd or Tam was chosen because previous studies have shown that in xenografts, key markers of E action are altered in the first few days after E deprivation (24). We speculated that differences in gene-expression profiles between the E sets, and between the untreated control and Ewd sets, would rigorously define the set of direct and indirect genes targeted by E in vivo and that Tam would confirm the ER dependency of the regulation. Furthermore, we wished to compare and contrast genes expressed in our chronic E-treated in vivo models with: 1) genes regulated acutely by E in vitro using the identical T47D-Y cells; 2) previously in vitro-defined E-regulated genes; and 3) ER discriminator genes defined in the clinical breast tumor microarray studies.
We show here that using in vivo tumors, the majority of E-regulated genes identified are novel. Additionally, two categories of E-regulated genes were identified: one subset that returned to basal levels after Ewd and a second subset that remained E up- or down-regulated and appeared to be resistant to 1 wk of Ewd. Two-thirds of all E-regulated genes identified were regulated by Ewd and Tam in a similar manner. A comparative analysis of gene-expression profiles in tumors and cultured cells revealed significant differences between short-term E-regulated genes in vitro and long-term E-regulated genes in vivo. Interestingly, however, the percentage of genes involved in distinct functional pathways regulated by E in vitro and in vivo are similar except for genes associated with development and organogenesis, which are preferentially regulated in vivo. We conclude that both models identify similar E-targeted pathways but that the specific genes differ extensively in the two models.
Materials and Methods
Cell culture and E treatment
T47D-Y cells [ER+, progesterone receptor (PR–)] are clonally derived sublines of wild-type T47Dco (ER+, PR+) human breast cancer cells (25). MCF-7 and MDA-231 human breast cancer cells were obtained from the American Type Culture Collection (Manassas, VA). Cells were routinely maintained in MEM supplemented with 5% fetal bovine serum (Hyclone Laboratories, Logan, UT). For microarray studies, cells were maintained 4 d in phenol red-free MEM containing 5% twice dextran-coated, charcoal-stripped fetal bovine serum prepared as described previously (26). The medium was changed on d 4, and cells were treated with 10 nM E2 or ethanol vehicle in phenol red-free MEM for 6 or 24 h.
T47D-Y tumor xenografts
All procedures were approved by the University of Colorado Institutional Animal Care and Use Committee. Ovariectomized (ovx) female athymic nu/nu mice were obtained from Harlan Sprague Dawley (Indianapolis, IN) at 5–6 wk of age. The animals were housed under controlled temperature, humidity, and lighting conditions. At approximately 7 wk of age, animals were anesthetized and implanted with a SILASTIC brand silicon tube (Dow Corning, Midland, MI) containing 2 mg of E2 or placebo (cellulose) sc at the base of the neck, as described (27). At the same time, animals were injected sc on both the left and right flank, with 5 x 106 T47D-Y human breast cancer cells suspended in 50% Matrigel. Mice were divided into the following experimental groups: E2 for 8 wk (E); E2 for 7 wk followed by surgical removal of the E2-containing implant and continuation on no hormone for 1 wk (Ewd); E2 for 8 wk and 10 mg of Tam added for the last week (E + Tam); or placebo for 8 wk or placebo for 7 wk followed by pellet removal for 1 wk (control). Data from the latter two controls were combined. Tumor growth was monitored weekly by measuring length (l) and width (w) with a digital caliper. Tumor area was calculated by the formula lw/4. Vaginal cytology was monitored weekly or daily during Ewd- or Tam-treatment periods (28). Animals were euthanized by CO2 asphyxiation. Tumors, normal mammary glands, and uteri were immediately removed, and tumors and uteri were weighed. Selected tumors were placed in TRIzol (Invitrogen, Carlsbad, CA), frozen in liquid nitrogen, and stored at –70 C. E2 levels in sera were measured, as previously described (27), by chemiluminescence microparticle immunoassay (Abbott Laboratories, North Chicago, IL) at the University of Colorado Hospital.
Immunohistochemistry and immunoblotting
Select tumors were formalin and paraffin embedded. For immunohistochemistry (IHC), sections (4 μm) were deparaffinized, antigen was retrieved, and ERs were detected with an anti-ER monoclonal antibody (6F11, Novocastra, Norwell, MA). Primary antiserum was omitted from select incubations to control for nonspecific labeling. Sections were counterstained with Meyer’s hematoxylin (Fisher Scientific, Pittsburgh, PA). Immunostained slides were blinded and scored by the method of Allred (29). For immunoblotting, cultured cells were lysed in TRIzol. Protein extracts were prepared according to the instructions of the manufacturer, and protein concentrations were determined by the method of Bradford (Bio-Rad Laboratories, Hercules, CA). Protein extracts (150 μg) were separated on a 10% SDS-PAGE gel, transferred to nitrocellulose, and probed with an anti-ER monoclonal antibody (AB15, Lab Vision, Fremont, CA), and bands were visualized by enhanced chemiluminescence (Amersham Biosciences Pharmacia Biotech, Arlington Heights, IL).
RNA isolation and expression profiling
Total RNA was extracted from tumors and cells using TRIzol and further purified using Qiagen RNeasy columns (Qiagen, Valencia, CA). RNA samples were prepared from six independent tumors removed from six individual animals per experimental condition (control, E, and Ewd). Four tumors were removed from three different animals undergoing E + Tam treatment. Three independent cell-culture samples per four experimental groups (6 or 24 h, E2 or ethanol) were harvested. RNA integrity was confirmed with an RNA-nano Bioanalyzer (Agilent Technologies, Palo Alto, CA), and cDNA was prepared and used to generate biotin-labeled cRNA (Enzo Bioarray High Yield RNA Transcript Labeling Kit). This was fragmented, hybridized to HG-U133A Affymetrix chips (Santa Clara, CA) that contain more than 22,000 human oligonucleotide probe sets and scanned (Gene Chip Scanner 3000, Affymetrix).
Data acquisition and analysis
Initial analyses from scanned image files were performed using Microarray Suite Version 5 (Affymetrix), with global scaling to correct for technical variations among chips. Thus, each chip was normalized to an arbitrary value of 500. For quality assurance, hybridization, RNA quality, and 5'- to 3'-end ratios for control genes and internal controls were examined. For subsequent analyses, data were imported from Microarrays Suite into GeneSpring version 6.1 (Silicon Genetics, Redwood City, CA), normalized, and used to identify patterns of gene expression. Genes were defined as E regulated when: 1) the gene was identified as present (P) or marginal (M) by GeneSpring in at least one sample studied (n = 6 in control, E, and E + Ewd tumors; n = 3 in cell cultures; n = 4 for E + Tam); 2) expression changes were greater than 1.5-fold between the control and E, E and Ewd, or control and Ewd sets; and 3) differences in expression were statistically significant (P > 0.05; see below). Functional classifications were based on gene ontology (GO) annotation obtained through the NetAffx server (Affymetrix, Santa Clara, CA) (30).
Real-time quantitative RT-PCR
To confirm the expression profiles of selected genes, real-time quantitative RT-PCR (qRT-PCR) was performed using an ABI Prism 7700 sequence detector (PerkinElmer Applied Biosystems, Inc., Forster City, CA), and RNA isolated separately from parallel tumors was used for microarray. cDNA was reverse-transcribed from 1 μg of total RNA using 0.4 mM random hexamers, 1x PCR buffer (5 mM MgCl2), 20 U of RNase Inhibitor, 4 mM deoxynucleotide triphosphates, and 125 U of Moloney Murine leukemia virus reverse transcriptase. The resulting cDNA reaction (5 μl) was used in real-time PCR conditions. Amplification was carried out in a 50 μl of reaction as previously described (31). Three independent cDNA samples per experimental group were used as template RNA. Primers and probes for selected genes were designed using Prism 7700 sequence detection software (Primer Express, PerkinElmer Applied Biosystems, Inc.). Primers were: ADD3 (forward, 5'-TGG AGG AGG TGT GAA TAT GGG-3'; reverse, 5'-CTC ATA AGC CCT TCA AAC TCA A-3'); AREG (forward, 5'-TGA TAC TCG GCT CAG GCC AT-3'; reverse, 5'-ATG GTT CAC GCT TCC CAG AG-3'); CXCR4 (forward, 5'-CCT GGC CTT CAT CAG TCT GG-3'; reverse, 5'-TTG GCC TCT GAC TGT TGG TG-3'); LMO4 (forward, 5'-GGA AAT AGC GGT GCT TGC A-3'; reverse, 5'-CGC CCT CAT GAC GAG TTC A-3'); RARESS (forward, 5'-CAA CAA CAG CTT GGA CCA TGA G-3'; reverse, 5'-TCC TTC GCA GAA CTG ATG ATC A-3'); and SDF-1 (forward, 5'-CCT GAG CTA CAG ATG CCC ATG-3'; reverse, 5'-TTT GAG ATG CTT GAC GTT GGC-3'). After an initial 10-min incubation at 95 C, 40 cycles of 95 C for 15 sec and 60 C for 1 min were performed. A standard curve was generated using appropriate dilutions of cDNA or plasmid controls. PCR for the target of interest was performed in parallel with reactions using glyceraldehyde-3-phosphate dehydrogenase internal reference-specific primers after the amount of both were determined from the standard curve. The target concentration was normalized to glyceraldehyde-3-phosphate dehydrogenase.
Statistics
Data represent the mean ± SEM. Differences between means were assessed using one-way ANOVA and individual comparisons using Student’s t test. For microarray analyses of tumors, P values were derived from nonparametric t test and ANOVA. For cultured cells, P values were derived from parametric t test and ANOVA. Both data sets were analyzed using GeneSpring statistical software. Genes that passed both the t test and ANOVA were scored as candidate genes. Genes identified in tumors that were at least 1.5-fold regulated are shown in the tables. Gene sets derived from cultured cells using a 1.5-fold cutoff are available as supplemental data, published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org. Statistical analyses of qRT-PCR were performed on normalized target values using one-way ANOVA and P < 0.05 was considered significant.
Results
The Ewd model
We investigated mechanisms of E-mediated signaling in breast cancers using both in vivo and in vitro approaches. T47D-Y human breast cancer cells that were ER+ but lacked PR were grown in vivo as solid tumors in ovx nude mice implanted with: 1) E2 for 8 wk; 2) E2 for 7 wk followed by 1 wk of Ewd; or 3) cellulose for 7 or 8 wk with or without sham withdrawal (control). In the absence of E, viable but nonproliferating tumor cells are retained at the injection site, whereas 8 wk of E-dependent growth yields small to intermediate sized tumors. The 1-wk time point for Ewd was chosen to target genes that change upon E cessation under conditions of minimal tumor regression. Other in vivo studies have shown that E-dependent markers change significantly in this time (24). In parallel studies, T47D-Y cells were also cultured in vitro and treated with or without E2 for 6 and 24 h to compare these typically studied experimental conditions (2, 4, 7, 32) with the in vivo results.
To document in vivo estrogenization and Ewd, vaginal cytology was monitored weekly and daily during the last 2 wk of the study. Mice treated with E2 for 8 wk had estrogenized vaginal epithelia that converted to diestrus after 1 wk of Ewd (not shown). At the end of the study, circulating E2 levels, uterine wet weights and morphology (Fig. 1), mammary gland morphology, and ER expression (not shown) were assessed as additional indicators of estrogenization and Ewd. Circulating E2 levels in untreated control mice averaged 55 pg/ml. They rose to 129 pg/ml after 8 wk of E2 and returned to control levels within 1 wk of Ewd (Fig. 1A). The uterus is a major E target. At the end of the studies, average uterine weights were 21, 114, and 29 mg, respectively, in control, E-supplemented, and Ewd mice. The gross morphology of the uterus mirrored these changes (Fig. 1B). Lastly, E delivery was monitored by assessment of normal mammary gland ER levels. ER expression levels were high in control glands, extensively down-regulated in E-supplemented mice, and restored to control levels by Ewd (not shown). Together, these data demonstrate the physiological validity of the E-treatment regimen and of the 1 wk Ewd approach to alter or reverse the estrogenization state of mice. The same protocols were then used to identify E-regulated genes in tumors.
E-dependent gene regulation in solid tumors
Tumor biology.
Continuous E treatment induced slow tumor growth in ovx nude mice inoculated with T47D-Y cells, whereas tumors failed to grow without E. However, nongrowing tumor nodules persisted at the injection site of untreated control mice, with the average residual tumor area starting and remaining at 27–30 mm2 over the course of the study (Fig. 2A). These nodules served as important E– controls. In mice supplemented with E, tumor area averaged 51 mm2 after 8 wk and all mice responded; 1 wk of Ewd reduced tumor area to 43 mm2, and this was statistically significant (P < 0.05). At necropsy, tumor mass averaged 71, 152, and 119 mg in control, E, and Ewd mice, respectively; however, differences between E and Ewd were not statistically significant (Fig. 2B), demonstrating that our aim of reversing the estrogenization state of the tumors could be achieved (see above) without marked tumor regression. Overall, these data indicate that the xenografts are E dependent for growth, but E is not required for cell survival, as shown by the residual tumor nodules in control mice.
ER in tumors and cells.
To additionally document E and Ewd states, ER levels were assessed in tumors and cultured cell lines. ER immunoblotting data for MCF-7 cells (ER+), MDA-231 (ER–), and untreated or 6- or 24-h E-treated T47D-Y cells are shown in Fig. 3A. Untreated MCF-7 cells express high ER levels; these are absent in MDA cells and lower in T47D-Y cells. ER were down-regulated in breast cancer cells by 6 or 24 h of E2 treatment in vitro as previously described (33). Tumor ER levels after chronic in vivo E treatment were assessed by IHC and showed a different pattern. ER, quantified by the method of Allred (29) were 3.0 in ovx, 4.0 in E, and 3.3 in Ewd; these did not differ significantly from one another. Clearly, the down-regulation pattern observed after acute E treatment in vitro, as assessed by immunoblotting, was not observed after chronic E treatment in vivo, as assessed by IHC. The differences may be due to technical differences in assay methods and antibodies or may represent real differences in ER response to transient vs. chronic E exposure.
E-regulated genes in tumors in vivo.
To identify E-regulated genes in vivo, expression profiles were assessed in tumors of control mice, after 8 wk of E2, and after 7 wk of E2 followed by 1 wk of Ewd. RNA was isolated from six independent tumors per experimental group and expression profiling was performed with Affymetrix U133A gene chips interrogating approximately 22,000 genes. Two types of statistical analyses, nonparametric ANOVA and Student’s t test, were performed. E-regulated genes were defined as ones that differ between: 1) control vs. E; 2) E vs. Ewd or; 3) control vs. Ewd. All genes passed both statistical tests and a more than 1.5-fold change was set as cutoff. One hundred eighty-eight genes passed both statistical tests and were scored as significantly regulated by E (Tables 1 and 2). One hundred eighty-six differed significantly between control and E. Among these, the majority 114 (61%) were up-regulated; the remaining 72 were (39%) were down-regulated. Two genes were unchanged between control and E but were down-regulated by Ewd and could not be classified. We defined them as E-regulated based on their sensitivity to E deprivation.
Patterns of gene regulation in tumors in response to Ewd.
Gene cluster analysis was performed using GeneSpring software and three main patterns—designated as classes I, II, and III—were observed (Fig. 4). Of the 188 E-regulated genes, 87 or approximately 46% were up- or down-regulated at least 1.5-fold by E relative to control, and then either completely or partially returned to baseline in a statistically significant manner after Ewd. These were designated as class I genes and represent the expected pattern. Table 1 lists these genes, their fold regulation, and whether they went up or down in response to E. Seventeen of these genes (italicized in Table 1) were previously reported to be E regulated.
Approximately half of the E-regulated genes (99 genes or 53%) were unaffected in a statistically significant manner by 1 wk of Ewd, and several E up-regulated genes paradoxically rose further after Ewd. These were designated as class II genes and are listed in Table 2. Twenty-five of these were previously reported to be E-regulated and are italicized in Table 2. A minor subset (1%) of two genes, GABPB2 and an EST, were designated as class III, because they did not change in response to E but were affected by Ewd.
Ewd compared with Tam.
The antiestrogen Tam is used clinically to suppress actions of ER in E-dependent breast cancers. Effects on E-regulated genes of 1 wk of Tam treatment (E + Tam) were compared with the effects of 1 wk of Ewd. One week of Tam partially suppressed E-stimulatory effects on uterine weights demonstrating efficiency of Tam delivery. Like with Ewd, E + Tam minimally suppressed tumor growth (data not shown). Expression profiling was assessed using four E + Tam tumors using Affymetrix U133A gene chips interrogating approximately 22,000 genes. Of the 188 E-regulated genes, 124 (66%) responded in the same class I, class II (supplemental Tables 1A and 1B), or class III manner to Tam as they did to Ewd (data not shown). The remaining 35% of genes either suppressed E effects when Ewd did not or failed to suppress E effects when Ewd did. That two-thirds of chronically E-regulated genes respond to Tam suggests that after 8 wk of E treatment, ERs remain engaged in regulating these genes. Interestingly, like the failure of class II genes to respond to Ewd, approximately 72% also failed to respond to Tam, suggesting that E removal, as accomplished by aromatase inhibitors, has extensive (but not complete) overlapping gene regulatory effects as Tam-induced ER blockade.
Confirmation by real-time RT-PCR.
To confirm the expression-profiling results among the two major gene classes in tumors, selected class I and II genes were reanalyzed by qRT-PCR. (The expression level of class III genes was too low for accurate quantitation.) As shown in Fig. 5, SDF-1 (CXCL12), LMO4, RARESS, ADD3, AREG, and CXCR4 showed similar expression patterns by both array profiling and qRT-PCR. Although gene profiling data indicate that SDF-1 is not significantly up-regulated by E in tumors, it was of interest because it is the ligand for the CXCR4 receptor and is a protein previously shown to be E-regulated in vitro (7).
Functional categories of E-regulated genes in tumors.
To define molecular or biological processes among the two major classes, genes were sorted into functional categories based on detailed bioinformatics analysis using the NetAffx Gene Ontology Mining Tool. The 87 class I and 99 class II genes were compared for their expression in functional categories as defined by GO gene annotations (data not shown). GO terms that predominated among genes in both classes were metabolism, cell death, development, and transport. Immune response genes were the main down-regulated class I and class II genes. Interestingly, cell proliferation, tumor aggressiveness, and cell-cycle genes were 2–7-fold more prevalent among the class II genes that fail to respond to Ewd. These include genes involved in cell cycle regulation (cyclin G2), differentiation (SKIL), development (SOX9), vascularization (VEGF), metastasis (CXCR4), oncogenesis (LMO4), and ER signaling (ESR-1). We speculate that long-term E treatment alters the hormone sensitivity of these genes, with implications for tumor aggressiveness and resistance. Note that ER (ESR-1) transcripts fall into this Ewd-resistant category, which supports the failure of ER protein to be down-regulated by E as measured by IHC (Fig. 3B).
In vitro gene regulation by E.
Because previous studies addressing E-regulated gene expression by global profiling used cultured cells exposed to short-term hormone treatments, we analyzed in vitro effects of 6 and 24 h E2 using the same ER+, T47D-Y cells from which the tumors were generated. These time points fall into the mid-range for in vitro studies, commonly between 1–72 h (2, 3, 4, 5, 7, 8, 9, 10, 12). A total of 266 genes were defined as E-regulated in vitro at one or both time points using a 1.5-fold cutoff; 74% up-regulated, and the rest down-regulated. A complete list is available online as Supplemental Tables 2A and 2B. Among in vitro up-regulated genes were: SDF-1 (CXCL12) (21.3-fold), CYP2B6 (16.6-fold), and MYC (11-fold). Down-regulated genes include: RNF32 (4.2-fold), P53AIP1 (4.1-fold), and SERHL (3.2-fold). Forty-one of these genes (italicized and referenced in the Supplemental Tables 2A and 2B) were previously reported to be E regulated. In general, fold changes observed with short-term E treatment in vitro tended to greatly exceed the fold changes measured in vivo.
Comparison of genes regulated by acute E in vitro vs. chronic E in vivo.
To determine the extent to which gene expression was modulated similarly by short-term E in vitro vs. long-term E in vivo, expression patterns between the two experimental conditions were compared. Of 266 in vitro E-regulated genes and 188 in vivo E-regulated genes, only 20 are common to both conditions: approximately 11% of the total number of genes defined as E-regulated in vivo are also E-regulated in vitro (Fig. 6 and Supplemental Tables IIA and IIB). Differences in expression profile patterns between in vitro and in vivo states are shown for four genes (Fig. 7). The CA12 (carbonic anhydrase XII) transcript is up-regulated by E2 at 6 and 24 h in vitro and is a class I gene in tumors. This is an example of a gene that is similarly regulated in both states. PFKFB3 (6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3) is up-regulated by E2 at 6 and 24 h in vitro but is a class II gene in vivo. Some genes were regulated in opposite directions in vitro vs. in vivo. For example, CITED 2 (Cbp/p300-interacting transactivator, with Glu/Asp-rich carboxy-terminal domain, 2) and SOX9 (sex-determining region Y-box 9) are E down-regulated in vitro but up-regulated in vivo.
Development and organogenesis genes are overrepresented in tumors.
To understand the differences between in vitro and in vivo gene profiles, E-regulated genes identified in both models were organized into functional categories. Of 14 functional GO terms searched, most were represented to a similar extent in cultured cells and tumors. This suggests that, whereas individual genes in a functional pathway may differ in vitro vs. in vivo, in general, similar functions are maintained under both conditions. However, several functional sets were more common (>2.5-fold) to tumors: ones involved in development/morphogenesis, differentiation, cell death, and immune response (Fig. 8). Together, the data indicate that, whereas E regulates unique genes in vitro vs. in vivo, functional pathways tend to remain similar in both states, with the exception of genes involved in processes that are unique to a three-dimensional, in vivo environment.
ER discriminator genes.
Recent studies (4, 13, 14, 15, 16, 17, 18, 19, 20, 21) of breast cancer patients have assessed the gene expression profiles of ER+ tumors. They have generated poorly overlapping lists of ER discriminator genes: genes that are coexpressed with ER, but that may or may not be under control of E signaling pathways. Of the 188 genes we define as E regulated in our experimental solid tumor models, 25 genes (Table 3), including ER (ESR1), appear at least once on one of four lists (13, 14, 15, 16). They may be particularly interesting targets for future clinical studies focused on E signaling.
Discussion
Models
These studies sought to define E-regulated gene expression profiles in an ER+, E-dependent xenograft model of human breast cancer grown in ovx nude mice. To our knowledge, no studies have analyzed E signaling in solid breast tumors or after long-term hormone treatments under conditions that resemble physiological and clinical states. In vivo E-regulated genes were defined by assessing differences in expression profiles in residual tumor nodules of ovx, E-untreated mice compared with E-treated mice, and/or by assessing differences in expression profiles in 8-wk E-treated tumors compared with ones in which E had been withdrawn for 1 wk or in which E-stimulated tumors were treated with Tam for 1 wk. We speculated that the genes expressed in the control and Ewd states would be similar, adding a second layer of control and minimizing the possibility that differences between control and E were due to generic growth signaling pathways rather than to pathways under E control. Tam treatment also indicates ER-dependent pathways.
We additionally assessed E-dependent gene expression profiles in cultured cells treated acutely with E. This allowed us to perform a detailed comparison of data generated in the same cells, under solid tumor conditions, against: 1) in vitro conditions commonly used to profile gene regulation and 2) clinical ER discriminator genes. We found that despite the fact that specific genes regulated acutely in vitro were substantially different from ones regulated chronically in vivo, functional pathways regulated by E in vitro and in vivo were similar. These data demonstrate the extent to which specific gene expression in human breast cancer cells is modulated by the microenvironment in which they reside.
Two classes of E-regulated genes in vivo
We identified a total of 188 genes regulated by E in vivo. Eighty-seven (46%) of these returned to pre-E levels after Ewd (class I). The majority of these class I genes (53, or 61%) also responded to Tam, suggesting that they are direct or indirect ER target genes. Because expression changes in class I genes correlated with initiation of tumor regression they may be involved in E-dependent tumor growth. Interestingly, approximately half (53%) of E-regulated genes did not return to basal levels after Ewd. These genes, which we call class II (Table 2), include transcripts for proteins thought to be involved in E-mediated growth or tumor progression, including ER (ESR1) (34); CXCR4 (chemokine C-X-C motif receptor 4) (35); CXCL12, the receptor for SDF-1; VEGF (vascular endothelial growth factor) (36, 37); and LMO4 (LIM domain only 4) (38). We speculate that this phenomenon plays a role in the resistance that invariably accompanies long-term hormone treatments of tumors. Importantly, many (>70%) of these class II genes were also Tam resistant. If the clinical objective of Ewd using AI, or ER blockade using antiestrogens, is to shut down E signaling, then our data would indicate that after chronic E treatment, only half of the signaling activity is effectively blocked; the rest remains activated and insensitive to suppression.
With regard to the class II genes, establishment of a permanently estrogenized phenotype could be due to epigenetic changes (39) in specific cell lineages, analogous to developmental imprinting (40, 41). Such changes can be hormonally driven (40, 42). Alternatively, it is possible that the class II gene subset is controlled not only by E but also by other pathways, including translational factors (43), growth factors (44, 45, 46, 47, 48, 49, 50), etc. This suggests mechanisms for resistance in which control over key genes, initially mediated by E, is seized by other factors over time (51). Under these conditions, E suppression or anti-ER therapies would fail to work even in ER+ settings. Class II transcripts may be excellent targets for studies aimed at understanding how tumors acquire resistance to long-term, ER-targeted treatments. Long-term Ewd studies of several weeks are necessary to further define the class II genes that do not return to baseline. This would distinguish more sensitive genes that can be regulated after longer periods of Ewd, from those whose expression is indeed permanently altered. However, the extensive overlap in class II genes exhibiting resistance to both Ewd and Tam suggests that the 1 wk test is valid.
E-regulated genes in tumors and cell lines
The majority of E-regulated genes defined to date—whether as single genes or by global profiling—are based on in vitro studies, mostly using ER+ MCF-7 cells. Short-term hormone treatments, in the range of hours to several days, are used in an effort to ensure that targets are direct. Gene expression profiles in these cells have sought to identify novel downstream E effectors in response to E, Tam, other selective ER modulators or xenoestrogens (2, 4, 7, 9, 32, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61). Most of the studies conclude that the gene profiles described provide new therapeutic targets for breast cancer prevention or treatment.
Of the total number of genes we define as regulated by long-term E treatment in T47D human breast cancer cells in vivo, only 11% overlap with acutely E-regulated genes in the identical cells grown in vitro (Fig. 6 and Supplemental Tables IIA and IIB). This is likely the result of the disparity in the time points examined, as well as differences in the microenvironment of solid tumors compared with cells in culture. Occasionally, we find that the same gene is regulated in opposite directions in vitro compared with in vivo. Despite these differences in specific genes regulated under the two conditions, we also show that both gene sets tend to fall into the same broad functional categories, with the exception that in tumors, E is more likely to regulate genes involved in nucleic acid metabolism, morphogenesis/development, differentiation, and immune responses. It has been suggested that different sets of genes can specify a single pathway (e.g. growth), so that it is possible to combine genes in many ways to achieve the same end (62). This suggests that cancer cell lines grown in vitro accurately retain information about functional pathways (e.g. E regulation of growth) while, compared with the same cells grown in vivo, perhaps inaccurately specifying any one gene in that pathway (e.g. gene X as an obligate mediator of E-dependent growth). Here we show that pathways are correctly identified by both in vitro and in vivo profiling but that it may be difficult to determine specific genes to target therapeutically using only in vitro models.
E-regulated genes in our tumor model vs. clinical ER discriminator genes
That human breast cancer cell lines are excellent models to study the human breast tumor biology has been demonstrated by recent gene expression studies that compare normal vs. malignant breast tissues against immortalized breast cancer cell lines. They find, remarkably, that the established cell lines faithfully reflect the unique properties of tumors, rather than of the cognate normal tissue of origin (63). Because it is difficult to study E signaling in clinical samples directly, alternate approaches to assessing E-regulated pathways focus on genes that cocluster with ER in microarray analyses of breast tumor biopsies. A number of studies have attempted to define ER discriminator gene clusters in human breast tumors (14, 15, 16, 18, 19, 20, 64, 65, 66). However, the linkage between ER discriminator genes and genes that are E regulated is unclear. One possibility is that ER discriminator genes correlate with enhanced differentiation, lack of aggressiveness, slower growth, and other indices of tumor biology that are only partly related to estrogenization state. Alternatively, coclustering of these genes with ER may indeed be due to their regulation by E. We have identified 25 genes (Table 3) that are E regulated in our solid tumor models and that have also been previously reported to be ER discriminator genes. These genes may represent one set to target for E-suppression therapies or to use as markers to monitor the effectiveness of such therapies.
Acknowledgments
We thank Dr. Britta Jacobsen for help with data analysis and review of the manuscript and acknowledge use of the Gene Expression, Tissue Culture, and Quantitative PCR Core Laboratories of the University Colorado Cancer Center.
Footnotes
This work was supported by the National Institutes of Health (National Research Service Award 5 F32 CA103511-02) and a United Negro College Fund/Pfizer grant (to D.M.E.H). Support was also provided by National Institutes of Health Grant RO1 CA26869, the National Foundation for Cancer Research, the Avon Foundation, and the Breast Cancer Research Foundation (to K.B.H.).
First Published Online October 20, 2005
Abbreviations: E, Estrogen(s); ER, estrogen receptor; Ewd, estrogen withdrawal; E2, 17-estradiol; GO, gene ontology; IHC, immunohistochemistry; ovx, ovariectomized; PR, progesterone receptor; qRT-PCR, real-time quantitative RT-PCR; Tam, tamoxifen.
Accepted for publication October 12, 2005.
References
Clark GM, Osborne CK, McGuire WL 1984 Correlations between estrogen receptor, progesterone receptor, and patient characteristics in human breast cancer. J Clin Oncol 2:1102–1109
Charpentier AH, Bednarek AK, Daniel RL, Hawkins KA, Laflin KJ, Gaddis S, MacLeod MC, Aldaz CM 2000 Effects of estrogen on global gene expression: identification of novel targets of estrogen action. Cancer Res 60:5977–5983
Soulez M, Parker MG 2001 Identification of novel oestrogen receptor target genes in human ZR75–1 breast cancer cells by expression profiling. J Mol Endocrinol 27:259–274
Inoue A, Yoshida N, Omoto Y, Oguchi S, Yamori T, Kiyama R, Hayashi S 2002 Development of cDNA microarray for expression profiling of estrogen-responsive genes. J Mol Endocrinol 29:175–192
Lobenhofer EK, Bennett L, Cable PL, Li L, Bushel PR, Afshari CA 2002 Regulation of DNA replication fork genes by 17-estradiol. Mol Endocrinol 16:1215–1229
Pedram A, Razandi M, Aitkenhead M, Hughes CC, Levin ER 2002 Integration of the non-genomic and genomic actions of estrogen. Membrane-initiated signaling by steroid to transcription and cell biology. J Biol Chem 277:50768–50775
Frasor J, Danes JM, Komm B, Chang KCN, Lyttle CR, Katzenellenbogen BS 2003 Profiling of estrogen up-and down-regulated gene expression in human breast cancer cells: insights into gene networks and pathways underlying estrogenic control of proliferation and cell phenotype. Endocrinology 144:4562–4574
Cicatiello L, Scafoglio C, Altucci L, Cancemi M, Natoli G, Facchiano A, Iazzetti G, Calogero R, Biglia N, De Bortoli M, Sfiligoi C, Sismondi P, Bresciani F, Weisz A 2004 A genomic view of estrogen actions in human breast cancer cells by expression profiling of the hormone-responsive transcriptome. J Mol Endocrinol 32:719–775
Frasor J, Stossi F, Danes JM, Komm B, Lyttle CR, Katzenellengoben BS 2004 selective estrogen receptor modulators: discrimination of agonistic versus antagonistic activities by gene expression profiling in breast cancer cells. Cancer Res 64:1522–1533
Lin CY, Strom A, Vega VB, Kong SL, Yeo AL, Thomsen JS, Chan WC, Doray B, Bangarusamy DK, Ramasamy A, Vergara LA, Tang S, Chong A, Bajic VB, Miller LD, Gustafsson JA, Liu ET 2004 Discovery of estrogen receptor alpha target genes and response elements in breast tumor cells. Genome Biol 5:R66
Stossi F, Barnett DH, Frasor J, Komm B, Lyttle CR, Katzenellenbogen BS 2004 Transcriptional profiling of estrogen-regulated gene expression via estrogen receptor (ER) alpha or ERbeta in human osteosarcoma cells: distinct and common target genes for these receptors. Endocrinology 145:3473–3486
Wang DY, Fulthorpe R, Liss SN, Edwards EA 2004 Identification of estrogen-responsive genes by complementary deoxyribonucleic acid microarray and characterization of a novel early estrogen-induced gene: EEIG1. Mol Endocrinol 18:402–411
van ’t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, Bernards R, Friend SH 2002 Gene expression profiling predicts clinical outcome of breast cancer. Nature 415:530–536
West M, Blanchette C, Dressman H, Huang E, Ishida S, Spang R, Zuzan H, Olson Jr JA, Marks JR, Nevins JR 2001 Predicting the clinical status of human breast cancer by using gene expression profiles. Proc Natl Acad Sci USA 98:11462–11467
Dressman MA, Walz TM, Lavedan C, Barnes L, Buchholtz S, Kwon I, Ellis MJ, Polymeropoulos MH 2001 Genes that co-cluster with estrogen receptor alpha in microarray analysis of breast biopsies. Pharmacogenomics J 1:135–141
Kun Y, How LC, Hoon TP, Bajic VB, Lam TS, Aggarwal A, Sze HG, Bok WS, Yin WC, Tan P 2003 Classifying the estrogen receptor status of breast cancers by expression profiles reveals a poor prognosis subpopulation exhibiting high expression of the ERBB2 receptor. Hum Mol Genet 12:3245–3258
van de Vijver MJ, He YD, van’t Veer LJ, Dai H, Hart AA, Voskuil DW, Schreiber GJ, Peterse JL, Roberts C, Marton MJ, Parrish M, Atsma D, Witteveen A, Glas A, Delahaye L, van der Velde T, Bartelink H, Rodenhuis S, Rutgers ET, Friend SH, Bernards R 2002 A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999–2009
Bertucci F, Houlgatte R, Benziane A, Granjeaud S, Adelaide J, Tagett R, Loriod B, Jacquemier J, Viens P, Jordan B, Birnbaum D, Nguyen C 2000 Gene expression profiling of primary breast carcinomas using arrays of candidate genes. Hum Mol Genet 9:2981–2991
Bertucci F, Nasser V, Granjeaud S, Eisinger F, Adelaide J, Tagett R, Loriod B, Giaconia A, Benziane A, Devilard E, Jacquemier J, Viens P, Nguyen C, Birnbaum D, Houlgatte R 2002 Gene expression profiles of poor-prognosis primary breast cancer correlate with survival. Hum Mol Genet 11:863–872
Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO, Botstein D 2000 Molecular portraits of human breast tumours. Nature 406:747–752
Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS, Thorsen T, Quist H, Matese JC, Brown PO, Botstein D, Eystein Lonning P, Borresen-Dale AL 2001 Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA 98:10869–10874
Weisz A, Basile W, Scafoglio C, Altucci L, Bresciani F, Facchiano A, Sismondi P, Cicatiello L, De Bortoli M 2004 Molecular identification of ERalpha-positive breast cancer cells by the expression profile of an intrinsic set of estrogen regulated genes. J Cell Physiol 200:440–450
Sartorius CA, Shen T, Horwitz KB 2003 Progesterone receptors A and B differentially affect the growth of estrogen-dependent human breast tumor xenografts. Breast Cancer Res Treat 79:287–299
Detre S, Salter J, Barnes DM, Riddler S, Hills M, Johnston SR, Gillett C, A’Hern R, Dowsett M 1999 Time-related effects of estrogen withdrawal on proliferation- and cell death-related events in MCF-7 xenografts. Int J Cancer 81:309–313
Sartorius CA, Groshong SD, Miller LA, Powell RL, Tung L, Takimoto GS, Horwitz KB 1994 New T47D breast cancer cell lines for the independent study of progesterone B- and A-receptors: only antiprogestin-occupied B-receptors are switched to transcriptional agonists by cAMP. Cancer Res 54:3868–3877
Horwitz KB, Mockus MB, Lessey BA 1982 Variant T47D human breast cancer cells with high progesterone-receptor levels despite estrogen and antiestrogen resistance. Cell 28:633–642
Sartorius CA, Harvell DME, Shen T, Horwitz KB 2005 Progestins initiate a luminal to myoepithelial switch in estrogen-dependent human breast tumors without altering growth. Cancer Res 65:9779–9788
Gallo D, Cantelmo F, Distefano M, Ferlini C, Zannoni GF, Riva A, Morazzoni P, Bombardelli E, Mancuso S, Scambia G 1999 Reproductive effects of dietary soy in female Wistar rats. Food Chem Toxicol 37:493–502
Harvey JM, Clark GM, Osborne CK, Allred DC 1999 Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 17:1474–1481
Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis A, Lewis S, Matese JC, Richardson JE, Ringwald M, Rubin GM, Sherlock G 2000 Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25:25–29
Haugen BR, Jensen DR, Sharma V, Pulawa LK, Hays WR, Krezel W, Chambon P, Eckel RH 2004 Retinoid X receptor -deficient mice have increased skeletal muscle lipoprotein lipase activity and less weight gain when fed a high-fat diet. Endocrinology 145:3679–3685
Ghosh MG, Thompson DA, Weigel RJ 2000 PDZK1 and GREB1 are estrogen-regulated genes expressed in hormone-responsive breast cancer. Cancer Res 60:6367–6375
Horwitz KB, McGuire WL 1980 Nuclear estrogen receptors. Effect of inhibitors on processing and steady state levels. J Biol Chem 255:9699–9705
Platet N, Cathiard AM, Gleizes M, Garcia M 2004 Estrogens and their receptors in breast cancer progression: a dual role in cancer proliferation and invasion. Crit Rev Oncol Hematol 51:55–67
Walser TC, Fulton AM 2004 The role of chemokines in the biology and therapy of breast cancer. Breast Dis 20:137–143
Desruisseau S, Palmari J, Giusti C, Romain S, Martin PM, Berthois Y 2004 Clinical relevance of amphiregulin and VEGF in primary breast cancers. Int J Cancer 111:733–740
Bachelder RE, Wendt MA, Mercurio AM 2002 Vascular endothelial growth factor promotes breast carcinoma invasion in an autocrine manner by regulating the chemokine receptor CXCR4. Cancer Res 62:7203–7206
Visvader JE, Venter D, Hahm K, Santamaria M, Sum EY, O’Reilly L, White D, Williams R, Armes J, Lindeman GJ 2001 The LIM domain gene LMO4 inhibits differentiation of mammary epithelial cells in vitro and is overexpressed in breast cancer. Proc Natl Acad Sci USA 98:14452–14457
Muller C, Leutz A 2001 Chromatin remodeling in development and differentiation. Curr Opin Genet Dev 11:167–174
Nelson KG, Sakai Y, Eitzman B, Steed T, McLachlan J 1994 Exposure to diethylstilbestrol during a critical developmental period of the mouse reproductive tract leads to persistent induction of two estrogen-regulated genes. Cell Growth Differ 5:595–606
Paulsen M, Ferguson-Smith AC 2001 DNA methylation in genomic imprinting, development, and disease. J Pathol 195:97–110
Ginger MR, Gonzalez-Rimbau MF, Gay JP, Rosen JM 2001 Persistent changes in gene expression induced by estrogen and progesterone in the rat mammary gland. Mol Endocrinol 15:1993–2009
Pentecost BT, Song R, Luo M, DePasquale JA, Fasco MJ 2005 Upstream regions of the estrogen receptor alpha proximal promoter transcript regulate ER protein expression through a translational mechanism. Mol Cell Endocrinol 229:83–94
Allar MA, Wood TL 2004 Expression of the insulin-like growth factor binding proteins during postnatal development of the murine mammary gland. Endocrinology 145:2467–2477
Lei J, Wu C, Wang X, Wang H 2005 p38 MAPK-dependent and YY1-mediated chemokine receptors CCR5 and CXCR4 up-regulation in U937 cell line infected by Mycobacterium tuberculosis or Actinobacillus actinomycetemcomitans. Biochem Biophys Res Commun 329:610–615
Loureiro RM, Maharaj AS, Dankort D, Muller WJ, D’Amore PA 2005 ErbB2 overexpression in mammary cells upregulates VEGF through the core promoter. Biochem Biophys Res Commun 326:455–465
Mori T, Kim J, Yamano T, Takeuchi H, Huang S, Umetani N, Koyanagi K, Hoon DS 2005 Epigenetic up-regulation of C-C chemokine receptor 7 and C-X-C chemokine receptor 4 expression in melanoma cells. Cancer Res 65:1800–1807
Read LD, Greene GL, Katzenellenbogen BS 1989 Regulation of estrogen receptor messenger ribonucleic acid and protein levels in human breast cancer cell lines by sex steroid hormones, their antagonists, and growth factors. Mol Endocrinol 3:295–304
Richer JK, Lange CA, Wierman AM, Brooks KM, Tung L, Takimoto GS, Horwitz KB 1998 Progesterone receptor variants found in breast cells repress transcription by wild-type receptors. Breast Cancer Res Treat 48:231–241
Shupnik MA, Gordon MS, Chin WW 1989 Tissue-specific regulation of rat estrogen receptor mRNAs. Mol Endocrinol 3:660–665
Osborne CK, Shou J, Massarweh S, Schiff R 2005 Crosstalk between estrogen receptor and growth factor receptor pathways as a cause for endocrine therapy resistance in breast cancer. Clin Cancer Res 11:865s–870s
Gadal F, Starzec A, Bozic C, Pillot-Brochet C, Malinge S, Ozanne V, Vicenzi J, Buffat L, Perret G, Iris F, Crepin M 2005 Integrative analysis of gene expression patterns predicts specific modulations of defined cell functions by estrogen and tamoxifen in MCF7 breast cancer cells. J Mol Endocrinol 34:61–75
Hayashi S 2004 Prediction of hormone sensitivity by DNA microassay. Biomed Pharmacother 58:1–9
Coser KR, Chesnes J, Hur J, Ray S, Isselbacher KJ, Shioda T 2003 Global analysis of ligand sensitivity of estrogen inducible and suppressible genes in MCF7/BUS breast cancer cells by DNA microarray. Proc Natl Acad Sci USA 100:13994–13999
Omoto Y, Hayashi S 2002 A study of estrogen signaling using DNA microarray in human breast cancer. Breast Cancer 9:308–311
Singleton DW, Feng Y, Chen Y, Busch SJ, Lee AV, Puga A, Khan SA 2004 Bisphenol-A and estradiol exert novel gene regulation in human MCF-7 derived breast cancer cells. Mol Cell Endocrinol 221:47–55
Kuang WW, Thompson DA, Hoch RV, Weigel RJ 1998 Differential screening and suppression subtractive hybridization identified genes differentially expressed in an estrogen receptor-positive breast carcinoma cell line. Nucleic Acids Res 26:1116–1123
Levenson AS, Svoboda KM, Pease KM, Kaiser SA, Chen B, Simons LA, Jovanovic BD, Dyck PA, Jordan VC 2002 Gene expression profiles with activation of the estrogen receptor alpha-selective estrogen receptor modulator complex in breast cancer cells expressing wild-type estrogen receptor. Cancer Res 62:4419–4426
Montano MM, Katzenellenbogen BS 1997 The quinone reductase gene: a unique estrogen receptor-regulated gene that is activated by antiestrogens. Proc Natl Acad Sci USA 94:2581–2586
Hall JM, Korach KS 2002 Analysis of the molecular mechanisms of human estrogen receptors alpha and beta reveals differential specificity in target promoter regulation by xenoestrogens. J Biol Chem 277:44455–44461
Inadera H, Hashimoto S, Dong HY, Suzuki T, Nagai S, Yamashita T, Toyoda N, Matsushima K 2000 WISP-2 as a novel estrogen-responsive gene in human breast cancer cells. Biochem Biophys Res Commun 275:108–114
Ein-Dor L, Kela I, Getz G, Givol D, Domany E 2005 Outcome signature genes in breast cancer: is there a unique set Bioinformatics 21:171–178
Beck CA, Weigel NL, Moyer ML, Nordeen SK, Edwards DP 1993 The progesterone antagonist RU486 acquires agonist activity upon stimulation of cAMP signaling pathways. Proc Natl Acad Sci USA 90:4441–4445
Gruvberger S, Ringner M, Chen Y, Panavally S, Saal LH, Borg A, Ferno M, Peterson C, Meltzer PS 2001 Estrogen receptor status in breast cancer is associated with remarkably distinct gene expression patterns. Cancer Res 61:5979–5984
Nagahata T, Onda M, Emi M, Nagai H, Tsumagari K, Fujimoto T, Hirano A, Sato T, Nishikawa K, Akiyama F, Sakamoto G, Kasumi F, Miki Y, Tanaka T, Tsunoda T 2004 Expression profiling to predict postoperative prognosis for estrogen receptor-negative breast cancers by analysis of 25,344 genes on a cDNA microarray. Cancer Sci 95:218–225
Nagai MA, Da Ros N, Neto MM, de Faria Junior SR, Brentani MM, Hirata Jr R, Neves EJ 2004 Gene expression profiles in breast tumors regarding the presence or absence of estrogen and progesterone receptors. Int J Cancer 111:892–899
Hunt SM, Clarke CL 1999 Expression and hormonal regulation of the Sox4 gene in mouse female reproductive tissues. Biol Reprod 61:476–481
Bouras T, Southey MC, Chang AC, Reddel RR, Willhite D, Glynne R, Henderson MA, Armes JE, Venter DJ 2002 Stanniocalcin 2 is an estrogen-responsive gene coexpressed with the estrogen receptor in human breast cancer. Cancer Res 62:1289–1295
Martinez-Lacaci I, Saceda M, Plowman GD, Johnson GR, Normanno N, Salomon DS, Dickson RB 1995 Estrogen and phorbol esters regulate amphiregulin expression by two separate mechanisms in human breast cancer cell lines. Endocrinology 136:3983–3992
Watanabe H, Suzuki A, Kobayashi M, Takahashi E, Itamoto M, Lubahn DB, Handa H, Iguchi T 2003 Analysis of temporal changes in the expression of estrogen-regulated genes in the uterus. J Mol Endocrinol 30:347–358
Simard J, de Launoit Y, Haagensen DE, Labrie F 1992 Additive stimulatory action of glucocorticoids and androgens on basal and estrogen-repressed apolipoprotein-D messenger ribonucleic acid levels and secretion in human breast cancer cells. Endocrinology 130:1115–1121
Mitev YA, Wolf SS, Almeida OF, Patchev VK 2003 Developmental expression profiles and distinct regional estrogen responsiveness suggest a novel role for the steroid receptor coactivator SRC-1 as discriminative amplifier of estrogen signaling in the rat brain. FASEB J 17:518–519
Robertson JA, Farnell Y, Lindahl LS, Ing NH 2002 Estradiol up-regulates estrogen receptor messenger ribonucleic acid in endometrial carcinoma (Ishikawa) cells by stabilizing the message. J Mol Endocrinol 29:125–135
Henttu P, Liao SS, Vihko P 1992 Androgens up-regulate the human prostate-specific antigen messenger ribonucleic acid (mRNA), but down-regulate the prostatic acid phosphatase mRNA in the LNCaP cell line. Endocrinology 130:766–772
Stoner M, Wormke M, Saville B, Samudio I, Qin C, Abdelrahim M, Safe S 2004 Estrogen regulation of vascular endothelial growth factor gene expression in ZR-75 breast cancer cells through interaction of estrogen receptor and SP proteins. Oncogene 23:1052–1063
Reese J, Das SK, Paria BC, Lim H, Song H, Matsumoto H, Knudtson KL, DuBois RN, Dey SK 2001 Global gene expression analysis to identify molecular markers of uterine receptivity and embryo implantation. J Biol Chem 276:44137–44145
Islamov RR, Hendricks WA, Katwa LC, McMurray RJ, Pak ES, Spanier NS, Murashov AK 2003 Effect of 17-estradiol on gene expression in lumbar spinal cord after sciatic nerve crush injury in ovx mice. Brain Res 966:65–75(Djuana M. E. Harvell, Jennifer K. Richer)
Department of Pathology (K.B.H.), University of Colorado Health Sciences Center at Fitzsimons, Aurora, Colorado 80045
Breast Center (D.C.A.), Baylor College of Medicine, Houston, Texas 77030
Abstract
In breast cancers, estrogen receptor (ER) levels are highly correlated with response to endocrine therapies. We sought to define mechanisms of estrogen (E) signaling in a solid breast tumor model using gene expression profiling. ER+ T47D-Y human breast cancer cells were grown as xenografts in ovariectomized nude mice under four conditions: 1) 17-estradiol for 8 wk (E); 2) without E for 8 wk (control); 3) E for 7 wk followed by 1 wk of E withdrawal (Ewd); or 4) E for 8 wk plus tamoxifen for the last week. E-regulated genes were defined as those that differed significantly between control and E and/or between E and Ewd or control and Ewd. These protocols generated 188 in vivo E-regulated genes that showed two major patterns of regulation. Approximately 46% returned to basal states after Ewd (class I genes); 53% did not (class II genes). In addition, more than 70% of class II-regulated genes also failed to reverse in response to tamoxifen. These genes may be interesting for the study of hormone-resistance issues. A subset of in vivo E-regulated genes appears on lists of clinical ER discriminator genes. These may be useful therapeutic targets or markers of E activity. Comparison of in vivo E-regulated genes with those regulated in identical cells in vitro after 6 and 24 h of E treatment demonstrate only 11% overlap. This indicates the extent to which gene expression profiles are uniquely dependent on hormone-treatment times and the cellular microenvironment.
Introduction
MORE THAN two thirds of human breast tumors express estrogen receptors (ERs) and are hormone responsive (1). Because of the importance of estrogens (E) and ER activities, multiple studies have sought to define E-regulated signaling pathways by assessment of individual genes or, more globally, by expression profiling (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16). Most of these studies have used established human breast cancer cell lines to compare E signaling in ER+ vs. ER– settings and to assess the functions of phytoestrogens, xenobiotics, or antiestrogens (2, 3, 4, 5, 7, 8, 9, 10, 12). These studies usually involve hormone-treatment times of 1–72 h using cells cultured in vitro. Generally, ER+ human breast cancer cells such as MCF-7 or ZR-75 are used, and these in vitro approaches, especially shorter treatment times of approximately 6 h, have yielded a number of novel and interesting direct E target genes (2, 3, 4, 5, 7, 8, 9, 12).
Independently, several recent studies have identified genes differentially expressed in ER+ compared with ER– solid tumors isolated from breast cancer patients (13, 14, 15, 16, 17, 18, 19, 20, 21). Collectively, the results of these studies have yielded lists of so-called ER discriminator genes. These lists, however, cannot address the question of whether the genes or transcription patterns identified are indeed controlled by, or even related to, E and ER signaling directly or whether they are simply correlative, reflecting other aspects of cancer biology such as tumor-differentiation state or growth rate that may be associated with ER without being ER regulated (22).
The purpose of the present study was to define mechanisms of E signaling in ER+ solid tumors using an in vivo approach. To do this, we used ER+ T47D human breast cancer cells (T47D-Y) grown as in vivo E-dependent solid tumor xenografts in nude mice (23). This allowed us to define E-regulated genes as ones that are differentially expressed in static, non-E-treated control tumors, compared with the 17-estradiol (E2)-treated proliferating tumors (E).
Additionally, to further refine the definition of in vivo E-regulated genes, we devised an E withdrawal (Ewd) protocol in which E2 was withdrawn for 1 wk in mice bearing E-dependent tumors, and the genes expressed in these tumors were analyzed. The Ewd data were, in turn, compared with 1 wk of E2 plus tamoxifen (Tam) (E + Tam) treatment after 7 wk of E2. One week of Ewd or Tam was chosen because previous studies have shown that in xenografts, key markers of E action are altered in the first few days after E deprivation (24). We speculated that differences in gene-expression profiles between the E sets, and between the untreated control and Ewd sets, would rigorously define the set of direct and indirect genes targeted by E in vivo and that Tam would confirm the ER dependency of the regulation. Furthermore, we wished to compare and contrast genes expressed in our chronic E-treated in vivo models with: 1) genes regulated acutely by E in vitro using the identical T47D-Y cells; 2) previously in vitro-defined E-regulated genes; and 3) ER discriminator genes defined in the clinical breast tumor microarray studies.
We show here that using in vivo tumors, the majority of E-regulated genes identified are novel. Additionally, two categories of E-regulated genes were identified: one subset that returned to basal levels after Ewd and a second subset that remained E up- or down-regulated and appeared to be resistant to 1 wk of Ewd. Two-thirds of all E-regulated genes identified were regulated by Ewd and Tam in a similar manner. A comparative analysis of gene-expression profiles in tumors and cultured cells revealed significant differences between short-term E-regulated genes in vitro and long-term E-regulated genes in vivo. Interestingly, however, the percentage of genes involved in distinct functional pathways regulated by E in vitro and in vivo are similar except for genes associated with development and organogenesis, which are preferentially regulated in vivo. We conclude that both models identify similar E-targeted pathways but that the specific genes differ extensively in the two models.
Materials and Methods
Cell culture and E treatment
T47D-Y cells [ER+, progesterone receptor (PR–)] are clonally derived sublines of wild-type T47Dco (ER+, PR+) human breast cancer cells (25). MCF-7 and MDA-231 human breast cancer cells were obtained from the American Type Culture Collection (Manassas, VA). Cells were routinely maintained in MEM supplemented with 5% fetal bovine serum (Hyclone Laboratories, Logan, UT). For microarray studies, cells were maintained 4 d in phenol red-free MEM containing 5% twice dextran-coated, charcoal-stripped fetal bovine serum prepared as described previously (26). The medium was changed on d 4, and cells were treated with 10 nM E2 or ethanol vehicle in phenol red-free MEM for 6 or 24 h.
T47D-Y tumor xenografts
All procedures were approved by the University of Colorado Institutional Animal Care and Use Committee. Ovariectomized (ovx) female athymic nu/nu mice were obtained from Harlan Sprague Dawley (Indianapolis, IN) at 5–6 wk of age. The animals were housed under controlled temperature, humidity, and lighting conditions. At approximately 7 wk of age, animals were anesthetized and implanted with a SILASTIC brand silicon tube (Dow Corning, Midland, MI) containing 2 mg of E2 or placebo (cellulose) sc at the base of the neck, as described (27). At the same time, animals were injected sc on both the left and right flank, with 5 x 106 T47D-Y human breast cancer cells suspended in 50% Matrigel. Mice were divided into the following experimental groups: E2 for 8 wk (E); E2 for 7 wk followed by surgical removal of the E2-containing implant and continuation on no hormone for 1 wk (Ewd); E2 for 8 wk and 10 mg of Tam added for the last week (E + Tam); or placebo for 8 wk or placebo for 7 wk followed by pellet removal for 1 wk (control). Data from the latter two controls were combined. Tumor growth was monitored weekly by measuring length (l) and width (w) with a digital caliper. Tumor area was calculated by the formula lw/4. Vaginal cytology was monitored weekly or daily during Ewd- or Tam-treatment periods (28). Animals were euthanized by CO2 asphyxiation. Tumors, normal mammary glands, and uteri were immediately removed, and tumors and uteri were weighed. Selected tumors were placed in TRIzol (Invitrogen, Carlsbad, CA), frozen in liquid nitrogen, and stored at –70 C. E2 levels in sera were measured, as previously described (27), by chemiluminescence microparticle immunoassay (Abbott Laboratories, North Chicago, IL) at the University of Colorado Hospital.
Immunohistochemistry and immunoblotting
Select tumors were formalin and paraffin embedded. For immunohistochemistry (IHC), sections (4 μm) were deparaffinized, antigen was retrieved, and ERs were detected with an anti-ER monoclonal antibody (6F11, Novocastra, Norwell, MA). Primary antiserum was omitted from select incubations to control for nonspecific labeling. Sections were counterstained with Meyer’s hematoxylin (Fisher Scientific, Pittsburgh, PA). Immunostained slides were blinded and scored by the method of Allred (29). For immunoblotting, cultured cells were lysed in TRIzol. Protein extracts were prepared according to the instructions of the manufacturer, and protein concentrations were determined by the method of Bradford (Bio-Rad Laboratories, Hercules, CA). Protein extracts (150 μg) were separated on a 10% SDS-PAGE gel, transferred to nitrocellulose, and probed with an anti-ER monoclonal antibody (AB15, Lab Vision, Fremont, CA), and bands were visualized by enhanced chemiluminescence (Amersham Biosciences Pharmacia Biotech, Arlington Heights, IL).
RNA isolation and expression profiling
Total RNA was extracted from tumors and cells using TRIzol and further purified using Qiagen RNeasy columns (Qiagen, Valencia, CA). RNA samples were prepared from six independent tumors removed from six individual animals per experimental condition (control, E, and Ewd). Four tumors were removed from three different animals undergoing E + Tam treatment. Three independent cell-culture samples per four experimental groups (6 or 24 h, E2 or ethanol) were harvested. RNA integrity was confirmed with an RNA-nano Bioanalyzer (Agilent Technologies, Palo Alto, CA), and cDNA was prepared and used to generate biotin-labeled cRNA (Enzo Bioarray High Yield RNA Transcript Labeling Kit). This was fragmented, hybridized to HG-U133A Affymetrix chips (Santa Clara, CA) that contain more than 22,000 human oligonucleotide probe sets and scanned (Gene Chip Scanner 3000, Affymetrix).
Data acquisition and analysis
Initial analyses from scanned image files were performed using Microarray Suite Version 5 (Affymetrix), with global scaling to correct for technical variations among chips. Thus, each chip was normalized to an arbitrary value of 500. For quality assurance, hybridization, RNA quality, and 5'- to 3'-end ratios for control genes and internal controls were examined. For subsequent analyses, data were imported from Microarrays Suite into GeneSpring version 6.1 (Silicon Genetics, Redwood City, CA), normalized, and used to identify patterns of gene expression. Genes were defined as E regulated when: 1) the gene was identified as present (P) or marginal (M) by GeneSpring in at least one sample studied (n = 6 in control, E, and E + Ewd tumors; n = 3 in cell cultures; n = 4 for E + Tam); 2) expression changes were greater than 1.5-fold between the control and E, E and Ewd, or control and Ewd sets; and 3) differences in expression were statistically significant (P > 0.05; see below). Functional classifications were based on gene ontology (GO) annotation obtained through the NetAffx server (Affymetrix, Santa Clara, CA) (30).
Real-time quantitative RT-PCR
To confirm the expression profiles of selected genes, real-time quantitative RT-PCR (qRT-PCR) was performed using an ABI Prism 7700 sequence detector (PerkinElmer Applied Biosystems, Inc., Forster City, CA), and RNA isolated separately from parallel tumors was used for microarray. cDNA was reverse-transcribed from 1 μg of total RNA using 0.4 mM random hexamers, 1x PCR buffer (5 mM MgCl2), 20 U of RNase Inhibitor, 4 mM deoxynucleotide triphosphates, and 125 U of Moloney Murine leukemia virus reverse transcriptase. The resulting cDNA reaction (5 μl) was used in real-time PCR conditions. Amplification was carried out in a 50 μl of reaction as previously described (31). Three independent cDNA samples per experimental group were used as template RNA. Primers and probes for selected genes were designed using Prism 7700 sequence detection software (Primer Express, PerkinElmer Applied Biosystems, Inc.). Primers were: ADD3 (forward, 5'-TGG AGG AGG TGT GAA TAT GGG-3'; reverse, 5'-CTC ATA AGC CCT TCA AAC TCA A-3'); AREG (forward, 5'-TGA TAC TCG GCT CAG GCC AT-3'; reverse, 5'-ATG GTT CAC GCT TCC CAG AG-3'); CXCR4 (forward, 5'-CCT GGC CTT CAT CAG TCT GG-3'; reverse, 5'-TTG GCC TCT GAC TGT TGG TG-3'); LMO4 (forward, 5'-GGA AAT AGC GGT GCT TGC A-3'; reverse, 5'-CGC CCT CAT GAC GAG TTC A-3'); RARESS (forward, 5'-CAA CAA CAG CTT GGA CCA TGA G-3'; reverse, 5'-TCC TTC GCA GAA CTG ATG ATC A-3'); and SDF-1 (forward, 5'-CCT GAG CTA CAG ATG CCC ATG-3'; reverse, 5'-TTT GAG ATG CTT GAC GTT GGC-3'). After an initial 10-min incubation at 95 C, 40 cycles of 95 C for 15 sec and 60 C for 1 min were performed. A standard curve was generated using appropriate dilutions of cDNA or plasmid controls. PCR for the target of interest was performed in parallel with reactions using glyceraldehyde-3-phosphate dehydrogenase internal reference-specific primers after the amount of both were determined from the standard curve. The target concentration was normalized to glyceraldehyde-3-phosphate dehydrogenase.
Statistics
Data represent the mean ± SEM. Differences between means were assessed using one-way ANOVA and individual comparisons using Student’s t test. For microarray analyses of tumors, P values were derived from nonparametric t test and ANOVA. For cultured cells, P values were derived from parametric t test and ANOVA. Both data sets were analyzed using GeneSpring statistical software. Genes that passed both the t test and ANOVA were scored as candidate genes. Genes identified in tumors that were at least 1.5-fold regulated are shown in the tables. Gene sets derived from cultured cells using a 1.5-fold cutoff are available as supplemental data, published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org. Statistical analyses of qRT-PCR were performed on normalized target values using one-way ANOVA and P < 0.05 was considered significant.
Results
The Ewd model
We investigated mechanisms of E-mediated signaling in breast cancers using both in vivo and in vitro approaches. T47D-Y human breast cancer cells that were ER+ but lacked PR were grown in vivo as solid tumors in ovx nude mice implanted with: 1) E2 for 8 wk; 2) E2 for 7 wk followed by 1 wk of Ewd; or 3) cellulose for 7 or 8 wk with or without sham withdrawal (control). In the absence of E, viable but nonproliferating tumor cells are retained at the injection site, whereas 8 wk of E-dependent growth yields small to intermediate sized tumors. The 1-wk time point for Ewd was chosen to target genes that change upon E cessation under conditions of minimal tumor regression. Other in vivo studies have shown that E-dependent markers change significantly in this time (24). In parallel studies, T47D-Y cells were also cultured in vitro and treated with or without E2 for 6 and 24 h to compare these typically studied experimental conditions (2, 4, 7, 32) with the in vivo results.
To document in vivo estrogenization and Ewd, vaginal cytology was monitored weekly and daily during the last 2 wk of the study. Mice treated with E2 for 8 wk had estrogenized vaginal epithelia that converted to diestrus after 1 wk of Ewd (not shown). At the end of the study, circulating E2 levels, uterine wet weights and morphology (Fig. 1), mammary gland morphology, and ER expression (not shown) were assessed as additional indicators of estrogenization and Ewd. Circulating E2 levels in untreated control mice averaged 55 pg/ml. They rose to 129 pg/ml after 8 wk of E2 and returned to control levels within 1 wk of Ewd (Fig. 1A). The uterus is a major E target. At the end of the studies, average uterine weights were 21, 114, and 29 mg, respectively, in control, E-supplemented, and Ewd mice. The gross morphology of the uterus mirrored these changes (Fig. 1B). Lastly, E delivery was monitored by assessment of normal mammary gland ER levels. ER expression levels were high in control glands, extensively down-regulated in E-supplemented mice, and restored to control levels by Ewd (not shown). Together, these data demonstrate the physiological validity of the E-treatment regimen and of the 1 wk Ewd approach to alter or reverse the estrogenization state of mice. The same protocols were then used to identify E-regulated genes in tumors.
E-dependent gene regulation in solid tumors
Tumor biology.
Continuous E treatment induced slow tumor growth in ovx nude mice inoculated with T47D-Y cells, whereas tumors failed to grow without E. However, nongrowing tumor nodules persisted at the injection site of untreated control mice, with the average residual tumor area starting and remaining at 27–30 mm2 over the course of the study (Fig. 2A). These nodules served as important E– controls. In mice supplemented with E, tumor area averaged 51 mm2 after 8 wk and all mice responded; 1 wk of Ewd reduced tumor area to 43 mm2, and this was statistically significant (P < 0.05). At necropsy, tumor mass averaged 71, 152, and 119 mg in control, E, and Ewd mice, respectively; however, differences between E and Ewd were not statistically significant (Fig. 2B), demonstrating that our aim of reversing the estrogenization state of the tumors could be achieved (see above) without marked tumor regression. Overall, these data indicate that the xenografts are E dependent for growth, but E is not required for cell survival, as shown by the residual tumor nodules in control mice.
ER in tumors and cells.
To additionally document E and Ewd states, ER levels were assessed in tumors and cultured cell lines. ER immunoblotting data for MCF-7 cells (ER+), MDA-231 (ER–), and untreated or 6- or 24-h E-treated T47D-Y cells are shown in Fig. 3A. Untreated MCF-7 cells express high ER levels; these are absent in MDA cells and lower in T47D-Y cells. ER were down-regulated in breast cancer cells by 6 or 24 h of E2 treatment in vitro as previously described (33). Tumor ER levels after chronic in vivo E treatment were assessed by IHC and showed a different pattern. ER, quantified by the method of Allred (29) were 3.0 in ovx, 4.0 in E, and 3.3 in Ewd; these did not differ significantly from one another. Clearly, the down-regulation pattern observed after acute E treatment in vitro, as assessed by immunoblotting, was not observed after chronic E treatment in vivo, as assessed by IHC. The differences may be due to technical differences in assay methods and antibodies or may represent real differences in ER response to transient vs. chronic E exposure.
E-regulated genes in tumors in vivo.
To identify E-regulated genes in vivo, expression profiles were assessed in tumors of control mice, after 8 wk of E2, and after 7 wk of E2 followed by 1 wk of Ewd. RNA was isolated from six independent tumors per experimental group and expression profiling was performed with Affymetrix U133A gene chips interrogating approximately 22,000 genes. Two types of statistical analyses, nonparametric ANOVA and Student’s t test, were performed. E-regulated genes were defined as ones that differ between: 1) control vs. E; 2) E vs. Ewd or; 3) control vs. Ewd. All genes passed both statistical tests and a more than 1.5-fold change was set as cutoff. One hundred eighty-eight genes passed both statistical tests and were scored as significantly regulated by E (Tables 1 and 2). One hundred eighty-six differed significantly between control and E. Among these, the majority 114 (61%) were up-regulated; the remaining 72 were (39%) were down-regulated. Two genes were unchanged between control and E but were down-regulated by Ewd and could not be classified. We defined them as E-regulated based on their sensitivity to E deprivation.
Patterns of gene regulation in tumors in response to Ewd.
Gene cluster analysis was performed using GeneSpring software and three main patterns—designated as classes I, II, and III—were observed (Fig. 4). Of the 188 E-regulated genes, 87 or approximately 46% were up- or down-regulated at least 1.5-fold by E relative to control, and then either completely or partially returned to baseline in a statistically significant manner after Ewd. These were designated as class I genes and represent the expected pattern. Table 1 lists these genes, their fold regulation, and whether they went up or down in response to E. Seventeen of these genes (italicized in Table 1) were previously reported to be E regulated.
Approximately half of the E-regulated genes (99 genes or 53%) were unaffected in a statistically significant manner by 1 wk of Ewd, and several E up-regulated genes paradoxically rose further after Ewd. These were designated as class II genes and are listed in Table 2. Twenty-five of these were previously reported to be E-regulated and are italicized in Table 2. A minor subset (1%) of two genes, GABPB2 and an EST, were designated as class III, because they did not change in response to E but were affected by Ewd.
Ewd compared with Tam.
The antiestrogen Tam is used clinically to suppress actions of ER in E-dependent breast cancers. Effects on E-regulated genes of 1 wk of Tam treatment (E + Tam) were compared with the effects of 1 wk of Ewd. One week of Tam partially suppressed E-stimulatory effects on uterine weights demonstrating efficiency of Tam delivery. Like with Ewd, E + Tam minimally suppressed tumor growth (data not shown). Expression profiling was assessed using four E + Tam tumors using Affymetrix U133A gene chips interrogating approximately 22,000 genes. Of the 188 E-regulated genes, 124 (66%) responded in the same class I, class II (supplemental Tables 1A and 1B), or class III manner to Tam as they did to Ewd (data not shown). The remaining 35% of genes either suppressed E effects when Ewd did not or failed to suppress E effects when Ewd did. That two-thirds of chronically E-regulated genes respond to Tam suggests that after 8 wk of E treatment, ERs remain engaged in regulating these genes. Interestingly, like the failure of class II genes to respond to Ewd, approximately 72% also failed to respond to Tam, suggesting that E removal, as accomplished by aromatase inhibitors, has extensive (but not complete) overlapping gene regulatory effects as Tam-induced ER blockade.
Confirmation by real-time RT-PCR.
To confirm the expression-profiling results among the two major gene classes in tumors, selected class I and II genes were reanalyzed by qRT-PCR. (The expression level of class III genes was too low for accurate quantitation.) As shown in Fig. 5, SDF-1 (CXCL12), LMO4, RARESS, ADD3, AREG, and CXCR4 showed similar expression patterns by both array profiling and qRT-PCR. Although gene profiling data indicate that SDF-1 is not significantly up-regulated by E in tumors, it was of interest because it is the ligand for the CXCR4 receptor and is a protein previously shown to be E-regulated in vitro (7).
Functional categories of E-regulated genes in tumors.
To define molecular or biological processes among the two major classes, genes were sorted into functional categories based on detailed bioinformatics analysis using the NetAffx Gene Ontology Mining Tool. The 87 class I and 99 class II genes were compared for their expression in functional categories as defined by GO gene annotations (data not shown). GO terms that predominated among genes in both classes were metabolism, cell death, development, and transport. Immune response genes were the main down-regulated class I and class II genes. Interestingly, cell proliferation, tumor aggressiveness, and cell-cycle genes were 2–7-fold more prevalent among the class II genes that fail to respond to Ewd. These include genes involved in cell cycle regulation (cyclin G2), differentiation (SKIL), development (SOX9), vascularization (VEGF), metastasis (CXCR4), oncogenesis (LMO4), and ER signaling (ESR-1). We speculate that long-term E treatment alters the hormone sensitivity of these genes, with implications for tumor aggressiveness and resistance. Note that ER (ESR-1) transcripts fall into this Ewd-resistant category, which supports the failure of ER protein to be down-regulated by E as measured by IHC (Fig. 3B).
In vitro gene regulation by E.
Because previous studies addressing E-regulated gene expression by global profiling used cultured cells exposed to short-term hormone treatments, we analyzed in vitro effects of 6 and 24 h E2 using the same ER+, T47D-Y cells from which the tumors were generated. These time points fall into the mid-range for in vitro studies, commonly between 1–72 h (2, 3, 4, 5, 7, 8, 9, 10, 12). A total of 266 genes were defined as E-regulated in vitro at one or both time points using a 1.5-fold cutoff; 74% up-regulated, and the rest down-regulated. A complete list is available online as Supplemental Tables 2A and 2B. Among in vitro up-regulated genes were: SDF-1 (CXCL12) (21.3-fold), CYP2B6 (16.6-fold), and MYC (11-fold). Down-regulated genes include: RNF32 (4.2-fold), P53AIP1 (4.1-fold), and SERHL (3.2-fold). Forty-one of these genes (italicized and referenced in the Supplemental Tables 2A and 2B) were previously reported to be E regulated. In general, fold changes observed with short-term E treatment in vitro tended to greatly exceed the fold changes measured in vivo.
Comparison of genes regulated by acute E in vitro vs. chronic E in vivo.
To determine the extent to which gene expression was modulated similarly by short-term E in vitro vs. long-term E in vivo, expression patterns between the two experimental conditions were compared. Of 266 in vitro E-regulated genes and 188 in vivo E-regulated genes, only 20 are common to both conditions: approximately 11% of the total number of genes defined as E-regulated in vivo are also E-regulated in vitro (Fig. 6 and Supplemental Tables IIA and IIB). Differences in expression profile patterns between in vitro and in vivo states are shown for four genes (Fig. 7). The CA12 (carbonic anhydrase XII) transcript is up-regulated by E2 at 6 and 24 h in vitro and is a class I gene in tumors. This is an example of a gene that is similarly regulated in both states. PFKFB3 (6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3) is up-regulated by E2 at 6 and 24 h in vitro but is a class II gene in vivo. Some genes were regulated in opposite directions in vitro vs. in vivo. For example, CITED 2 (Cbp/p300-interacting transactivator, with Glu/Asp-rich carboxy-terminal domain, 2) and SOX9 (sex-determining region Y-box 9) are E down-regulated in vitro but up-regulated in vivo.
Development and organogenesis genes are overrepresented in tumors.
To understand the differences between in vitro and in vivo gene profiles, E-regulated genes identified in both models were organized into functional categories. Of 14 functional GO terms searched, most were represented to a similar extent in cultured cells and tumors. This suggests that, whereas individual genes in a functional pathway may differ in vitro vs. in vivo, in general, similar functions are maintained under both conditions. However, several functional sets were more common (>2.5-fold) to tumors: ones involved in development/morphogenesis, differentiation, cell death, and immune response (Fig. 8). Together, the data indicate that, whereas E regulates unique genes in vitro vs. in vivo, functional pathways tend to remain similar in both states, with the exception of genes involved in processes that are unique to a three-dimensional, in vivo environment.
ER discriminator genes.
Recent studies (4, 13, 14, 15, 16, 17, 18, 19, 20, 21) of breast cancer patients have assessed the gene expression profiles of ER+ tumors. They have generated poorly overlapping lists of ER discriminator genes: genes that are coexpressed with ER, but that may or may not be under control of E signaling pathways. Of the 188 genes we define as E regulated in our experimental solid tumor models, 25 genes (Table 3), including ER (ESR1), appear at least once on one of four lists (13, 14, 15, 16). They may be particularly interesting targets for future clinical studies focused on E signaling.
Discussion
Models
These studies sought to define E-regulated gene expression profiles in an ER+, E-dependent xenograft model of human breast cancer grown in ovx nude mice. To our knowledge, no studies have analyzed E signaling in solid breast tumors or after long-term hormone treatments under conditions that resemble physiological and clinical states. In vivo E-regulated genes were defined by assessing differences in expression profiles in residual tumor nodules of ovx, E-untreated mice compared with E-treated mice, and/or by assessing differences in expression profiles in 8-wk E-treated tumors compared with ones in which E had been withdrawn for 1 wk or in which E-stimulated tumors were treated with Tam for 1 wk. We speculated that the genes expressed in the control and Ewd states would be similar, adding a second layer of control and minimizing the possibility that differences between control and E were due to generic growth signaling pathways rather than to pathways under E control. Tam treatment also indicates ER-dependent pathways.
We additionally assessed E-dependent gene expression profiles in cultured cells treated acutely with E. This allowed us to perform a detailed comparison of data generated in the same cells, under solid tumor conditions, against: 1) in vitro conditions commonly used to profile gene regulation and 2) clinical ER discriminator genes. We found that despite the fact that specific genes regulated acutely in vitro were substantially different from ones regulated chronically in vivo, functional pathways regulated by E in vitro and in vivo were similar. These data demonstrate the extent to which specific gene expression in human breast cancer cells is modulated by the microenvironment in which they reside.
Two classes of E-regulated genes in vivo
We identified a total of 188 genes regulated by E in vivo. Eighty-seven (46%) of these returned to pre-E levels after Ewd (class I). The majority of these class I genes (53, or 61%) also responded to Tam, suggesting that they are direct or indirect ER target genes. Because expression changes in class I genes correlated with initiation of tumor regression they may be involved in E-dependent tumor growth. Interestingly, approximately half (53%) of E-regulated genes did not return to basal levels after Ewd. These genes, which we call class II (Table 2), include transcripts for proteins thought to be involved in E-mediated growth or tumor progression, including ER (ESR1) (34); CXCR4 (chemokine C-X-C motif receptor 4) (35); CXCL12, the receptor for SDF-1; VEGF (vascular endothelial growth factor) (36, 37); and LMO4 (LIM domain only 4) (38). We speculate that this phenomenon plays a role in the resistance that invariably accompanies long-term hormone treatments of tumors. Importantly, many (>70%) of these class II genes were also Tam resistant. If the clinical objective of Ewd using AI, or ER blockade using antiestrogens, is to shut down E signaling, then our data would indicate that after chronic E treatment, only half of the signaling activity is effectively blocked; the rest remains activated and insensitive to suppression.
With regard to the class II genes, establishment of a permanently estrogenized phenotype could be due to epigenetic changes (39) in specific cell lineages, analogous to developmental imprinting (40, 41). Such changes can be hormonally driven (40, 42). Alternatively, it is possible that the class II gene subset is controlled not only by E but also by other pathways, including translational factors (43), growth factors (44, 45, 46, 47, 48, 49, 50), etc. This suggests mechanisms for resistance in which control over key genes, initially mediated by E, is seized by other factors over time (51). Under these conditions, E suppression or anti-ER therapies would fail to work even in ER+ settings. Class II transcripts may be excellent targets for studies aimed at understanding how tumors acquire resistance to long-term, ER-targeted treatments. Long-term Ewd studies of several weeks are necessary to further define the class II genes that do not return to baseline. This would distinguish more sensitive genes that can be regulated after longer periods of Ewd, from those whose expression is indeed permanently altered. However, the extensive overlap in class II genes exhibiting resistance to both Ewd and Tam suggests that the 1 wk test is valid.
E-regulated genes in tumors and cell lines
The majority of E-regulated genes defined to date—whether as single genes or by global profiling—are based on in vitro studies, mostly using ER+ MCF-7 cells. Short-term hormone treatments, in the range of hours to several days, are used in an effort to ensure that targets are direct. Gene expression profiles in these cells have sought to identify novel downstream E effectors in response to E, Tam, other selective ER modulators or xenoestrogens (2, 4, 7, 9, 32, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61). Most of the studies conclude that the gene profiles described provide new therapeutic targets for breast cancer prevention or treatment.
Of the total number of genes we define as regulated by long-term E treatment in T47D human breast cancer cells in vivo, only 11% overlap with acutely E-regulated genes in the identical cells grown in vitro (Fig. 6 and Supplemental Tables IIA and IIB). This is likely the result of the disparity in the time points examined, as well as differences in the microenvironment of solid tumors compared with cells in culture. Occasionally, we find that the same gene is regulated in opposite directions in vitro compared with in vivo. Despite these differences in specific genes regulated under the two conditions, we also show that both gene sets tend to fall into the same broad functional categories, with the exception that in tumors, E is more likely to regulate genes involved in nucleic acid metabolism, morphogenesis/development, differentiation, and immune responses. It has been suggested that different sets of genes can specify a single pathway (e.g. growth), so that it is possible to combine genes in many ways to achieve the same end (62). This suggests that cancer cell lines grown in vitro accurately retain information about functional pathways (e.g. E regulation of growth) while, compared with the same cells grown in vivo, perhaps inaccurately specifying any one gene in that pathway (e.g. gene X as an obligate mediator of E-dependent growth). Here we show that pathways are correctly identified by both in vitro and in vivo profiling but that it may be difficult to determine specific genes to target therapeutically using only in vitro models.
E-regulated genes in our tumor model vs. clinical ER discriminator genes
That human breast cancer cell lines are excellent models to study the human breast tumor biology has been demonstrated by recent gene expression studies that compare normal vs. malignant breast tissues against immortalized breast cancer cell lines. They find, remarkably, that the established cell lines faithfully reflect the unique properties of tumors, rather than of the cognate normal tissue of origin (63). Because it is difficult to study E signaling in clinical samples directly, alternate approaches to assessing E-regulated pathways focus on genes that cocluster with ER in microarray analyses of breast tumor biopsies. A number of studies have attempted to define ER discriminator gene clusters in human breast tumors (14, 15, 16, 18, 19, 20, 64, 65, 66). However, the linkage between ER discriminator genes and genes that are E regulated is unclear. One possibility is that ER discriminator genes correlate with enhanced differentiation, lack of aggressiveness, slower growth, and other indices of tumor biology that are only partly related to estrogenization state. Alternatively, coclustering of these genes with ER may indeed be due to their regulation by E. We have identified 25 genes (Table 3) that are E regulated in our solid tumor models and that have also been previously reported to be ER discriminator genes. These genes may represent one set to target for E-suppression therapies or to use as markers to monitor the effectiveness of such therapies.
Acknowledgments
We thank Dr. Britta Jacobsen for help with data analysis and review of the manuscript and acknowledge use of the Gene Expression, Tissue Culture, and Quantitative PCR Core Laboratories of the University Colorado Cancer Center.
Footnotes
This work was supported by the National Institutes of Health (National Research Service Award 5 F32 CA103511-02) and a United Negro College Fund/Pfizer grant (to D.M.E.H). Support was also provided by National Institutes of Health Grant RO1 CA26869, the National Foundation for Cancer Research, the Avon Foundation, and the Breast Cancer Research Foundation (to K.B.H.).
First Published Online October 20, 2005
Abbreviations: E, Estrogen(s); ER, estrogen receptor; Ewd, estrogen withdrawal; E2, 17-estradiol; GO, gene ontology; IHC, immunohistochemistry; ovx, ovariectomized; PR, progesterone receptor; qRT-PCR, real-time quantitative RT-PCR; Tam, tamoxifen.
Accepted for publication October 12, 2005.
References
Clark GM, Osborne CK, McGuire WL 1984 Correlations between estrogen receptor, progesterone receptor, and patient characteristics in human breast cancer. J Clin Oncol 2:1102–1109
Charpentier AH, Bednarek AK, Daniel RL, Hawkins KA, Laflin KJ, Gaddis S, MacLeod MC, Aldaz CM 2000 Effects of estrogen on global gene expression: identification of novel targets of estrogen action. Cancer Res 60:5977–5983
Soulez M, Parker MG 2001 Identification of novel oestrogen receptor target genes in human ZR75–1 breast cancer cells by expression profiling. J Mol Endocrinol 27:259–274
Inoue A, Yoshida N, Omoto Y, Oguchi S, Yamori T, Kiyama R, Hayashi S 2002 Development of cDNA microarray for expression profiling of estrogen-responsive genes. J Mol Endocrinol 29:175–192
Lobenhofer EK, Bennett L, Cable PL, Li L, Bushel PR, Afshari CA 2002 Regulation of DNA replication fork genes by 17-estradiol. Mol Endocrinol 16:1215–1229
Pedram A, Razandi M, Aitkenhead M, Hughes CC, Levin ER 2002 Integration of the non-genomic and genomic actions of estrogen. Membrane-initiated signaling by steroid to transcription and cell biology. J Biol Chem 277:50768–50775
Frasor J, Danes JM, Komm B, Chang KCN, Lyttle CR, Katzenellenbogen BS 2003 Profiling of estrogen up-and down-regulated gene expression in human breast cancer cells: insights into gene networks and pathways underlying estrogenic control of proliferation and cell phenotype. Endocrinology 144:4562–4574
Cicatiello L, Scafoglio C, Altucci L, Cancemi M, Natoli G, Facchiano A, Iazzetti G, Calogero R, Biglia N, De Bortoli M, Sfiligoi C, Sismondi P, Bresciani F, Weisz A 2004 A genomic view of estrogen actions in human breast cancer cells by expression profiling of the hormone-responsive transcriptome. J Mol Endocrinol 32:719–775
Frasor J, Stossi F, Danes JM, Komm B, Lyttle CR, Katzenellengoben BS 2004 selective estrogen receptor modulators: discrimination of agonistic versus antagonistic activities by gene expression profiling in breast cancer cells. Cancer Res 64:1522–1533
Lin CY, Strom A, Vega VB, Kong SL, Yeo AL, Thomsen JS, Chan WC, Doray B, Bangarusamy DK, Ramasamy A, Vergara LA, Tang S, Chong A, Bajic VB, Miller LD, Gustafsson JA, Liu ET 2004 Discovery of estrogen receptor alpha target genes and response elements in breast tumor cells. Genome Biol 5:R66
Stossi F, Barnett DH, Frasor J, Komm B, Lyttle CR, Katzenellenbogen BS 2004 Transcriptional profiling of estrogen-regulated gene expression via estrogen receptor (ER) alpha or ERbeta in human osteosarcoma cells: distinct and common target genes for these receptors. Endocrinology 145:3473–3486
Wang DY, Fulthorpe R, Liss SN, Edwards EA 2004 Identification of estrogen-responsive genes by complementary deoxyribonucleic acid microarray and characterization of a novel early estrogen-induced gene: EEIG1. Mol Endocrinol 18:402–411
van ’t Veer LJ, Dai H, van de Vijver MJ, He YD, Hart AA, Mao M, Peterse HL, van der Kooy K, Marton MJ, Witteveen AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, Bernards R, Friend SH 2002 Gene expression profiling predicts clinical outcome of breast cancer. Nature 415:530–536
West M, Blanchette C, Dressman H, Huang E, Ishida S, Spang R, Zuzan H, Olson Jr JA, Marks JR, Nevins JR 2001 Predicting the clinical status of human breast cancer by using gene expression profiles. Proc Natl Acad Sci USA 98:11462–11467
Dressman MA, Walz TM, Lavedan C, Barnes L, Buchholtz S, Kwon I, Ellis MJ, Polymeropoulos MH 2001 Genes that co-cluster with estrogen receptor alpha in microarray analysis of breast biopsies. Pharmacogenomics J 1:135–141
Kun Y, How LC, Hoon TP, Bajic VB, Lam TS, Aggarwal A, Sze HG, Bok WS, Yin WC, Tan P 2003 Classifying the estrogen receptor status of breast cancers by expression profiles reveals a poor prognosis subpopulation exhibiting high expression of the ERBB2 receptor. Hum Mol Genet 12:3245–3258
van de Vijver MJ, He YD, van’t Veer LJ, Dai H, Hart AA, Voskuil DW, Schreiber GJ, Peterse JL, Roberts C, Marton MJ, Parrish M, Atsma D, Witteveen A, Glas A, Delahaye L, van der Velde T, Bartelink H, Rodenhuis S, Rutgers ET, Friend SH, Bernards R 2002 A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999–2009
Bertucci F, Houlgatte R, Benziane A, Granjeaud S, Adelaide J, Tagett R, Loriod B, Jacquemier J, Viens P, Jordan B, Birnbaum D, Nguyen C 2000 Gene expression profiling of primary breast carcinomas using arrays of candidate genes. Hum Mol Genet 9:2981–2991
Bertucci F, Nasser V, Granjeaud S, Eisinger F, Adelaide J, Tagett R, Loriod B, Giaconia A, Benziane A, Devilard E, Jacquemier J, Viens P, Nguyen C, Birnbaum D, Houlgatte R 2002 Gene expression profiles of poor-prognosis primary breast cancer correlate with survival. Hum Mol Genet 11:863–872
Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO, Botstein D 2000 Molecular portraits of human breast tumours. Nature 406:747–752
Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS, Thorsen T, Quist H, Matese JC, Brown PO, Botstein D, Eystein Lonning P, Borresen-Dale AL 2001 Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA 98:10869–10874
Weisz A, Basile W, Scafoglio C, Altucci L, Bresciani F, Facchiano A, Sismondi P, Cicatiello L, De Bortoli M 2004 Molecular identification of ERalpha-positive breast cancer cells by the expression profile of an intrinsic set of estrogen regulated genes. J Cell Physiol 200:440–450
Sartorius CA, Shen T, Horwitz KB 2003 Progesterone receptors A and B differentially affect the growth of estrogen-dependent human breast tumor xenografts. Breast Cancer Res Treat 79:287–299
Detre S, Salter J, Barnes DM, Riddler S, Hills M, Johnston SR, Gillett C, A’Hern R, Dowsett M 1999 Time-related effects of estrogen withdrawal on proliferation- and cell death-related events in MCF-7 xenografts. Int J Cancer 81:309–313
Sartorius CA, Groshong SD, Miller LA, Powell RL, Tung L, Takimoto GS, Horwitz KB 1994 New T47D breast cancer cell lines for the independent study of progesterone B- and A-receptors: only antiprogestin-occupied B-receptors are switched to transcriptional agonists by cAMP. Cancer Res 54:3868–3877
Horwitz KB, Mockus MB, Lessey BA 1982 Variant T47D human breast cancer cells with high progesterone-receptor levels despite estrogen and antiestrogen resistance. Cell 28:633–642
Sartorius CA, Harvell DME, Shen T, Horwitz KB 2005 Progestins initiate a luminal to myoepithelial switch in estrogen-dependent human breast tumors without altering growth. Cancer Res 65:9779–9788
Gallo D, Cantelmo F, Distefano M, Ferlini C, Zannoni GF, Riva A, Morazzoni P, Bombardelli E, Mancuso S, Scambia G 1999 Reproductive effects of dietary soy in female Wistar rats. Food Chem Toxicol 37:493–502
Harvey JM, Clark GM, Osborne CK, Allred DC 1999 Estrogen receptor status by immunohistochemistry is superior to the ligand-binding assay for predicting response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 17:1474–1481
Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis A, Lewis S, Matese JC, Richardson JE, Ringwald M, Rubin GM, Sherlock G 2000 Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25:25–29
Haugen BR, Jensen DR, Sharma V, Pulawa LK, Hays WR, Krezel W, Chambon P, Eckel RH 2004 Retinoid X receptor -deficient mice have increased skeletal muscle lipoprotein lipase activity and less weight gain when fed a high-fat diet. Endocrinology 145:3679–3685
Ghosh MG, Thompson DA, Weigel RJ 2000 PDZK1 and GREB1 are estrogen-regulated genes expressed in hormone-responsive breast cancer. Cancer Res 60:6367–6375
Horwitz KB, McGuire WL 1980 Nuclear estrogen receptors. Effect of inhibitors on processing and steady state levels. J Biol Chem 255:9699–9705
Platet N, Cathiard AM, Gleizes M, Garcia M 2004 Estrogens and their receptors in breast cancer progression: a dual role in cancer proliferation and invasion. Crit Rev Oncol Hematol 51:55–67
Walser TC, Fulton AM 2004 The role of chemokines in the biology and therapy of breast cancer. Breast Dis 20:137–143
Desruisseau S, Palmari J, Giusti C, Romain S, Martin PM, Berthois Y 2004 Clinical relevance of amphiregulin and VEGF in primary breast cancers. Int J Cancer 111:733–740
Bachelder RE, Wendt MA, Mercurio AM 2002 Vascular endothelial growth factor promotes breast carcinoma invasion in an autocrine manner by regulating the chemokine receptor CXCR4. Cancer Res 62:7203–7206
Visvader JE, Venter D, Hahm K, Santamaria M, Sum EY, O’Reilly L, White D, Williams R, Armes J, Lindeman GJ 2001 The LIM domain gene LMO4 inhibits differentiation of mammary epithelial cells in vitro and is overexpressed in breast cancer. Proc Natl Acad Sci USA 98:14452–14457
Muller C, Leutz A 2001 Chromatin remodeling in development and differentiation. Curr Opin Genet Dev 11:167–174
Nelson KG, Sakai Y, Eitzman B, Steed T, McLachlan J 1994 Exposure to diethylstilbestrol during a critical developmental period of the mouse reproductive tract leads to persistent induction of two estrogen-regulated genes. Cell Growth Differ 5:595–606
Paulsen M, Ferguson-Smith AC 2001 DNA methylation in genomic imprinting, development, and disease. J Pathol 195:97–110
Ginger MR, Gonzalez-Rimbau MF, Gay JP, Rosen JM 2001 Persistent changes in gene expression induced by estrogen and progesterone in the rat mammary gland. Mol Endocrinol 15:1993–2009
Pentecost BT, Song R, Luo M, DePasquale JA, Fasco MJ 2005 Upstream regions of the estrogen receptor alpha proximal promoter transcript regulate ER protein expression through a translational mechanism. Mol Cell Endocrinol 229:83–94
Allar MA, Wood TL 2004 Expression of the insulin-like growth factor binding proteins during postnatal development of the murine mammary gland. Endocrinology 145:2467–2477
Lei J, Wu C, Wang X, Wang H 2005 p38 MAPK-dependent and YY1-mediated chemokine receptors CCR5 and CXCR4 up-regulation in U937 cell line infected by Mycobacterium tuberculosis or Actinobacillus actinomycetemcomitans. Biochem Biophys Res Commun 329:610–615
Loureiro RM, Maharaj AS, Dankort D, Muller WJ, D’Amore PA 2005 ErbB2 overexpression in mammary cells upregulates VEGF through the core promoter. Biochem Biophys Res Commun 326:455–465
Mori T, Kim J, Yamano T, Takeuchi H, Huang S, Umetani N, Koyanagi K, Hoon DS 2005 Epigenetic up-regulation of C-C chemokine receptor 7 and C-X-C chemokine receptor 4 expression in melanoma cells. Cancer Res 65:1800–1807
Read LD, Greene GL, Katzenellenbogen BS 1989 Regulation of estrogen receptor messenger ribonucleic acid and protein levels in human breast cancer cell lines by sex steroid hormones, their antagonists, and growth factors. Mol Endocrinol 3:295–304
Richer JK, Lange CA, Wierman AM, Brooks KM, Tung L, Takimoto GS, Horwitz KB 1998 Progesterone receptor variants found in breast cells repress transcription by wild-type receptors. Breast Cancer Res Treat 48:231–241
Shupnik MA, Gordon MS, Chin WW 1989 Tissue-specific regulation of rat estrogen receptor mRNAs. Mol Endocrinol 3:660–665
Osborne CK, Shou J, Massarweh S, Schiff R 2005 Crosstalk between estrogen receptor and growth factor receptor pathways as a cause for endocrine therapy resistance in breast cancer. Clin Cancer Res 11:865s–870s
Gadal F, Starzec A, Bozic C, Pillot-Brochet C, Malinge S, Ozanne V, Vicenzi J, Buffat L, Perret G, Iris F, Crepin M 2005 Integrative analysis of gene expression patterns predicts specific modulations of defined cell functions by estrogen and tamoxifen in MCF7 breast cancer cells. J Mol Endocrinol 34:61–75
Hayashi S 2004 Prediction of hormone sensitivity by DNA microassay. Biomed Pharmacother 58:1–9
Coser KR, Chesnes J, Hur J, Ray S, Isselbacher KJ, Shioda T 2003 Global analysis of ligand sensitivity of estrogen inducible and suppressible genes in MCF7/BUS breast cancer cells by DNA microarray. Proc Natl Acad Sci USA 100:13994–13999
Omoto Y, Hayashi S 2002 A study of estrogen signaling using DNA microarray in human breast cancer. Breast Cancer 9:308–311
Singleton DW, Feng Y, Chen Y, Busch SJ, Lee AV, Puga A, Khan SA 2004 Bisphenol-A and estradiol exert novel gene regulation in human MCF-7 derived breast cancer cells. Mol Cell Endocrinol 221:47–55
Kuang WW, Thompson DA, Hoch RV, Weigel RJ 1998 Differential screening and suppression subtractive hybridization identified genes differentially expressed in an estrogen receptor-positive breast carcinoma cell line. Nucleic Acids Res 26:1116–1123
Levenson AS, Svoboda KM, Pease KM, Kaiser SA, Chen B, Simons LA, Jovanovic BD, Dyck PA, Jordan VC 2002 Gene expression profiles with activation of the estrogen receptor alpha-selective estrogen receptor modulator complex in breast cancer cells expressing wild-type estrogen receptor. Cancer Res 62:4419–4426
Montano MM, Katzenellenbogen BS 1997 The quinone reductase gene: a unique estrogen receptor-regulated gene that is activated by antiestrogens. Proc Natl Acad Sci USA 94:2581–2586
Hall JM, Korach KS 2002 Analysis of the molecular mechanisms of human estrogen receptors alpha and beta reveals differential specificity in target promoter regulation by xenoestrogens. J Biol Chem 277:44455–44461
Inadera H, Hashimoto S, Dong HY, Suzuki T, Nagai S, Yamashita T, Toyoda N, Matsushima K 2000 WISP-2 as a novel estrogen-responsive gene in human breast cancer cells. Biochem Biophys Res Commun 275:108–114
Ein-Dor L, Kela I, Getz G, Givol D, Domany E 2005 Outcome signature genes in breast cancer: is there a unique set Bioinformatics 21:171–178
Beck CA, Weigel NL, Moyer ML, Nordeen SK, Edwards DP 1993 The progesterone antagonist RU486 acquires agonist activity upon stimulation of cAMP signaling pathways. Proc Natl Acad Sci USA 90:4441–4445
Gruvberger S, Ringner M, Chen Y, Panavally S, Saal LH, Borg A, Ferno M, Peterson C, Meltzer PS 2001 Estrogen receptor status in breast cancer is associated with remarkably distinct gene expression patterns. Cancer Res 61:5979–5984
Nagahata T, Onda M, Emi M, Nagai H, Tsumagari K, Fujimoto T, Hirano A, Sato T, Nishikawa K, Akiyama F, Sakamoto G, Kasumi F, Miki Y, Tanaka T, Tsunoda T 2004 Expression profiling to predict postoperative prognosis for estrogen receptor-negative breast cancers by analysis of 25,344 genes on a cDNA microarray. Cancer Sci 95:218–225
Nagai MA, Da Ros N, Neto MM, de Faria Junior SR, Brentani MM, Hirata Jr R, Neves EJ 2004 Gene expression profiles in breast tumors regarding the presence or absence of estrogen and progesterone receptors. Int J Cancer 111:892–899
Hunt SM, Clarke CL 1999 Expression and hormonal regulation of the Sox4 gene in mouse female reproductive tissues. Biol Reprod 61:476–481
Bouras T, Southey MC, Chang AC, Reddel RR, Willhite D, Glynne R, Henderson MA, Armes JE, Venter DJ 2002 Stanniocalcin 2 is an estrogen-responsive gene coexpressed with the estrogen receptor in human breast cancer. Cancer Res 62:1289–1295
Martinez-Lacaci I, Saceda M, Plowman GD, Johnson GR, Normanno N, Salomon DS, Dickson RB 1995 Estrogen and phorbol esters regulate amphiregulin expression by two separate mechanisms in human breast cancer cell lines. Endocrinology 136:3983–3992
Watanabe H, Suzuki A, Kobayashi M, Takahashi E, Itamoto M, Lubahn DB, Handa H, Iguchi T 2003 Analysis of temporal changes in the expression of estrogen-regulated genes in the uterus. J Mol Endocrinol 30:347–358
Simard J, de Launoit Y, Haagensen DE, Labrie F 1992 Additive stimulatory action of glucocorticoids and androgens on basal and estrogen-repressed apolipoprotein-D messenger ribonucleic acid levels and secretion in human breast cancer cells. Endocrinology 130:1115–1121
Mitev YA, Wolf SS, Almeida OF, Patchev VK 2003 Developmental expression profiles and distinct regional estrogen responsiveness suggest a novel role for the steroid receptor coactivator SRC-1 as discriminative amplifier of estrogen signaling in the rat brain. FASEB J 17:518–519
Robertson JA, Farnell Y, Lindahl LS, Ing NH 2002 Estradiol up-regulates estrogen receptor messenger ribonucleic acid in endometrial carcinoma (Ishikawa) cells by stabilizing the message. J Mol Endocrinol 29:125–135
Henttu P, Liao SS, Vihko P 1992 Androgens up-regulate the human prostate-specific antigen messenger ribonucleic acid (mRNA), but down-regulate the prostatic acid phosphatase mRNA in the LNCaP cell line. Endocrinology 130:766–772
Stoner M, Wormke M, Saville B, Samudio I, Qin C, Abdelrahim M, Safe S 2004 Estrogen regulation of vascular endothelial growth factor gene expression in ZR-75 breast cancer cells through interaction of estrogen receptor and SP proteins. Oncogene 23:1052–1063
Reese J, Das SK, Paria BC, Lim H, Song H, Matsumoto H, Knudtson KL, DuBois RN, Dey SK 2001 Global gene expression analysis to identify molecular markers of uterine receptivity and embryo implantation. J Biol Chem 276:44137–44145
Islamov RR, Hendricks WA, Katwa LC, McMurray RJ, Pak ES, Spanier NS, Murashov AK 2003 Effect of 17-estradiol on gene expression in lumbar spinal cord after sciatic nerve crush injury in ovx mice. Brain Res 966:65–75(Djuana M. E. Harvell, Jennifer K. Richer)