c-Myb and Members of the c-Ets Family of Transcription Factors Act as Molecular Switches to Mediate Opposite Steroid Regulation of the Human Glucocorticoid Receptor 1A Promoter*

Steroid auto-regulation of the human glucocorticoid receptor (hGR) 1A promoter in lymphoblast cells resides largely in two DNA elements (footprints 11 and 12). We show here that c-Myb and c-Ets family members (Ets-1/2, PU.1, and Spi-B) control hGR 1A promoter regulation in T- and B-lymphoblast cells. Two T-lymphoblast lines, CEM-C7 and Jurkat, contain high levels of c-Myb and low levels of PU.1, whereas the opposite is true in IM-9 B-lymphoblasts. In Jurkat cells, overexpression of c-Ets-1, c-Ets-2, or PU.1 effectively represses dexamethasone-mediated up-regulation of an hGR 1A promoter-luciferase reporter gene, as do dominant negative c-Myb (c-Myb DNA-binding domain) or Ets proteins (Ets-2 DNA-binding domain). Overexpression of c-Myb in IM-9 cells confers hormone-dependent up-regulation to the hGR 1A promoter reporter gene. Chromatin immunoprecipitation assays show that hormone treatment causes the recruitment of hGR and c-Myb to the hGR 1A promoter in CEM-C7 cells, whereas hGR and PU.1 are recruited to this promoter in IM-9 cells. These observations suggest that the specific transcription factor that binds to footprint 12, when hGR binds to the adjacent footprint 11, determines the direction of hGR 1A promoter auto-regulation. This leads to a “molecular switch” model for auto-regulation of the hGR 1A promoter.


Steroid auto-regulation of the human glucocorticoid receptor (hGR) 1A promoter in lymphoblast cells resides largely in two DNA elements (footprints 11 and 12). We show here that c-Myb and c-Ets family members (Ets-1/2, PU.1, and Spi-B) control hGR 1A promoter regulation in T-and B-lymphoblast cells. Two T-lymphoblast lines, CEM-C7 and Jurkat, contain high levels of c-Myb and low levels of PU.1, whereas the opposite is true in IM-9 B-lymphoblasts. In Jurkat cells, overexpression of c-Ets-1, c-Ets-2, or PU.1 effectively represses dexamethasone-mediated up-regulation of an hGR 1A promoter-luciferase reporter gene, as do dominant negative c-Myb (c-Myb DNA-binding domain) or Ets proteins (Ets-2 DNA-binding domain). Overexpression of c-Myb in IM-9 cells confers hormone-dependent up-regulation to the hGR 1A promoter
reporter gene. Chromatin immunoprecipitation assays show that hormone treatment causes the recruitment of hGR and c-Myb to the hGR 1A promoter in CEM-C7 cells, whereas hGR and PU.1 are recruited to this promoter in IM-9 cells. These observations suggest that the specific transcription factor that binds to footprint 12, when hGR binds to the adjacent footprint 11, determines the direction of hGR 1A promoter auto-regulation. This leads to a "molecular switch" model for auto-regulation of the hGR 1A promoter.
Glucocorticoids (GCs) 2 specifically induce apoptosis in several types of leukemia and lymphoma, and they are used routinely in treating T-cell acute lymphoblastic leukemia (T-ALL) (1)(2)(3)(4)(5)(6). The cellular function of GCs is mediated by the glucocorticoid receptor (GR) protein, which is widely expressed in most tissues and cell lines (7)(8)(9)(10). The inactive, cytoplasmic GR binds the GC hormone to form an activated GR-ligand complex, which then translocates to the nucleus, recognizes and binds to specific DNA sequences in the gene promoter called glucocorticoid response elements (GREs), and affects the expression of these genes (7,8,11,12). It is not yet clear how GCs cause apoptosis in certain lymphoblasts, but this fact is effectively used in chemotherapy regimens for certain types of leukemias and lymphomas (13)(14)(15)(16)(17)(18)(19).
There is a strong correlation between functional cellular GR level and the sensitivity of the cell to GCs, including the apoptotic response of lymphoblasts (3,20,21). Although the initial cellular GR level is not absolutely predictive, an auto-up-regulation of GR to a certain threshold level is required for apoptosis in hormone-sensitive cells (22)(23)(24)(25)(26)(27)(28). Conversely, an auto-down-regulation of GR levels is frequently observed in cells resistant to hormone-mediated apoptosis, such as the pre-B-lymphoblast cell line, IM-9, and the lowered GR concentration actually dampens GC signaling (9,19,26). Thus, steroid auto-up-regulation of GR gene expression in vivo could provide a sensitive indicator of hormonal sensitivity of hematologic malignancies.
Human GR gene expression is controlled by at least three promoters, 1A, 1B, and 1C (9,10,29). Promoters 1B and 1C are ubiquitously expressed, whereas the 1A promoter is selectively expressed in hematopoietic cells (9,10). The transcripts emanating from all three promoters are auto-up-regulated in response to DEX treatment in hormonesensitive, CEM-C7 cell, T-lymphoblasts, and down-regulated in IM-9 cells that are resistant to hormone-mediated apoptosis (9,26). Because no consensus GREs are present in any of these promoters, the molecular mechanism for auto-regulation of the 1A, 1B, and 1C promoters was unknown (9,10,30). Recently, we have identified the DNA sequence that mediates up-regulation of 1A promoter activity (9,31). This DNA sequence contains a half GRE (footprint 11 (FP11)) and a sequence (FP12) containing overlapping consensus binding sites for c-Myb or c-Ets proteins. Both FP11 and FP12 are required for full hormone responsiveness of the hGR 1A promoter.
In the present study we have further investigated the roles of c-Myb and c-Ets protein members in steroid-mediated auto-regulation of the hGR 1A promoter. Our data show that c-Myb and Ets family members (Ets-1, Ets-2, PU.1, and Spi-B) can affect hGR 1A promoter activity. Western blot and functional studies indicate that c-Myb and PU.1 are the most likely candidates in mediating the opposite hormonal response of the hGR 1A promoter in T-and B-lymphoblasts. Chromatin immunoprecipitation analyses show that, in response to DEX treatment, GR and c-Myb are recruited by the hGR 1A promoter in CEM-C7 T-cells, whereas GR and PU.1 are recruited in IM-9 B-cells. These results suggest a "molecular switch" model for hGR 1A promoter regulation, in which GR binding to the hGR 1A promoter recruits cell type-specific transcription factors to an adjacent DNA sequence. c-Myb is recruited in T-lymphoblasts, and this results in GR up-regulation and apoptosis. PU.1 is recruited in B-cells that do not undergo hormone-mediated apoptosis, and the GR is down-regulated in these cells. These findings may lead to novel clinical therapies that could increase the response rate and the magnitude of the response to steroid treatment in certain hematologic malignancies.

MATERIALS AND METHODS
Cell Culture-The human Jurkat (T-ALL) and IM-9 B-lymphoblastic cell lines (both from the American Type Culture Collection, Manassas, VA) were grown in RPMI 1640 plus 10% fetal bovine serum (FBS; Invitrogen). The human, CEM-C7, acute lymphoblastic leukemia cell line was a kind gift from Dr. E Brad Thompson (University of Texas Medical Branch, Galveston, TX), and it was maintained in RPMI 1640 supplemented with 10% dialyzed FBS (Invitrogen). All of the cells were grown in 5% CO 2 at 37°C. To treat the cells, 1 M dexamethasone (Sigma) in ethanol was added to the culture medium, and the same final concentration of ethanol vehicle (0.01%) was used in the controls.
Transient Transfection and Luciferase Reporter Gene Assays-Superfect transfection reagent (Qiagen) was used to transfect Jurkat cells, according to the manufacturer's directions. The cells were treated (with EtOH or DEX) 24 h after transfection and were collected for analysis after an additional 24 h of incubation. The collected cells were lysed and measured for firefly luciferase and ␤-galactosidase activity on an Ascent Luminoskan (Labsystems, Franklin, MA) as previously described (9,31).
Electroporation-IM-9 cells were electroporated using a method modified from the laboratory of Dr. Jeffrey M. Harmon (Uniformed Services University of the Health Sciences, Bethesda, MD). 3 The cells in log phase growth were counted and harvested by centrifugation at 800 ϫ g for 10 min (4°C). After two washes with RPMI 1640 medium (without FBS), the cells were resuspended in cold RPMI 1640 (without FBS and containing 10 mM HEPES, pH 7.2) at 2.5 ϫ 10 7 cells/ml. 200 l of cell suspension and 10 g of DNA were combined in a Bio-Rad 0.4-cm electroporation cuvette, and the mixture was incubated on ice for 10 min. Electroporation was done with a Gene Pulser II (Bio-Rad) at 340 V/960 microfarad. Electroporated cells were then incubated on ice for 10 min before being transferred to 10 ml of culture medium (with FBS) at room temperature. The electroporated cells were allowed to grow in 5% CO 2 at 37°C for 24 h before DEX or ethanol treatment.
Chromatin Immunoprecipitation Assay (ChIP Assay)-Formaldehyde cross-linking and chromatin immunoprecipitation assays of tissue culture cells were performed as described (36,37) with some modifications. IM-9 and CEM-C7 cells (1 ϫ 10 7 ) were treated for 24 h with/or without 1 M DEX. The cells were incubated in 1% formaldehyde (Sigma) for 10 min at room temperature. Cross-linking was stopped by adding glycine to a final concentration of 125 mM. The cells were spun down and washed three times with ice-cold phosphate-buffered saline before resuspending in 200 l of cell lysis buffer (5 mM Pipes-KOH, pH 8.0, 85 mM KCl, 0.5% Nonidet P-40) containing protease inhibitors (1 g/ml leupeptin, 1 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride). After incubating for 10 min on ice, the nuclei were pelleted, resuspended in 200 l of nuclear lysis buffer (50 mM Tris-HCl, pH 8.1, 10 mM EDTA, 1% SDS with protease inhibitors), and incubated on ice for 10 min. The chromatin released from the nuclei was sonicated at 4°C with a Branson sonifier to obtain DNA lengths of 0.1-1.5 kilobase pairs. The sonicated cell lysate was clarified by centrifugation at 13,000 ϫ g for 10 min at 4°C. The supernatant containing the sheared chromatin was used for the immunoprecipitation assay.

RESULTS
Hormone Responsiveness of the hGR 1A Promoter Depends upon Footprints 11 and 12-We previously identified a hGR 1A promoter that can be significantly auto-up-regulated in T-lymphoblasts (9,26), and it is much more sensitive than the other two major hGR promoters, 1B and 1C (9,26,31). Deletion analysis and in vitro DNase I footprinting using nuclear extracts from DEX-treated CEM-C7 cells identified the DNA sequences that mediate the hormone responsiveness in T-cells: FP11, a nonconsensus GRE half-site, and FP12. Computer analysis of FP12 revealed perfectly overlapping c-Myb and c-Ets binding sequences (Fig.  1A), and our previous analysis confirmed that these proteins do bind to this sequence (31). To more fully characterize the role of these DNA elements for both basal promoter activity and in hormone responsiveness, we performed internal deletions of FP11 and FP12, singly or in combination. Deletion of either of these two footprints causes nonresponsiveness of the 1A promoter to DEX, whether a full-length promoter (Ϫ964/ϩ269; Fig. 1B) or a shorter promoter that retained about 60% of the basal promoter activity (ϩ41/ϩ269; Fig. 1C) is used. The cause of the apparent down-regulation of promoter activity upon DEX treatment of the full-length promoter in which FP12 has been deleted ( Fig. 1B) is unknown, but this may result from the influence of other upstream sequences that are lacking in the shorter promoter construct. These data clearly show that FP11 and FP12 are the critical and perhaps sole sequences required for the hormone-induced up-regulation of hGR 1A promoter activity in the Jurkat T-lymphoblast line.
Besides being critical for hormonal responsiveness, these two DNA elements also contribute much to the basal promoter activity, because their deletion causes a dramatic decrease in transcriptional activity (Fig.  1, B and C). In fact, when a minimal hGR 1A promoter containing only FP11 and FP12 (ϩ242/ϩ269) was cloned into pXP1, it retained about 70% percent of the basal activity of the full-length promoter and was completely hormone-responsive (data not shown). Thus, it appears that nearly all of the protein factors that control basal promoter activity and hormonal responsiveness can be recruited to this small DNA element (FP11/FP12) in the hGR 1A promoter.
Ets Family Members Are Selectively Expressed in Lymphoid Cell Lineages-Based upon sequence analysis of FP12, we previously showed that c-Myb and two c-Ets protein family members, c-Ets-1 and c-Ets-2, are able to bind to FP12 in vitro (31). Thus, we wished to determine which c-Ets protein family members are indeed expressed in the three lymphoid cell lines used in our studies. Although there are over 20 different Ets family protein members, various c-Ets proteins are selectively expressed in certain hematopoietic lineages (39). This allowed us to focus on the most likely candidates that might be present in the three cell lines. In particular, Spi-B and PU.1 are two Ets proteins that are expressed during, and are important for, differentiation of lymphoblasts. Spi-B is selectively expressed during T-cell development, and its level drops in mature cells, and PU.1 is restricted to B-cells and macrophages (39). Using Western blotting, we analyzed the relative levels of four Ets family members that were possibly expressed in our T-cell (CEM-C7 and Jurkat cells) and B-cell (IM-9 cells) model systems (Fig.  2). We found that: 1) c-Ets-1/2 are expressed at comparable levels in all three cell lines; 2) Spi-B is present in the three lines, but it is expressed at FIGURE 1. FP11 and FP12 are critical for hormonal responsiveness of the hGR 1A promoter in lymphoblasts. A, The hGR 1A promoter/exon sequence, containing FP11 and FP12, is shown. The c-Ets and c-Myb consensus binding sequences that match with the overlapping sequences in FP12 are indicated. B, activity of the full-length hGR 1A promoter. 1.5 g of pXP1-hGR 1A e/p Ϫ964/ϩ269 or deletion constructs were cotransfected into Jurkat cells together with 1 g of pCYGR. ⌬FP11, ⌬FP12, and ⌬FP11/12 are the respective deletion mutants of these sequences from promoter hGR 1A e/p Ϫ964/ϩ269. C, activity of the ϩ41/ϩ269 hGR 1A promoter. 1.5 g of pXP1-hGR 1A e/p ϩ 41/ϩ269 or FP11, FP12, or FP11/12 deletion constructs of this plasmid were cotransfected into Jurkat cells with 1 g of pCYGR. 3 ϫ 10 6 cells were used for each transient transfection, and 1 g of a pCMV-␤-galactosidase (pCMV-␤-gal) construct was included to normalize the transfection efficiency. The luciferase activity of each sample was measured with a luminometer and normalized to the ␤-galactosidase activity. Three or four individual experiments were combined to obtain the mean and calculate the S.E. The steroid responsiveness is plotted as the percentage of the EtOH control. The basal promoter activity is given as relative luminescence units normalized to ␤-galactosidase activity. *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.005 for steroid responsiveness versus the respective ethanol control.

c-Myb and c-Ets Proteins Modulate hGR 1A Promoter Expression
much higher levels in CEM-C7 and Jurkat T-cells than in the B-cells; and 3) PU.1 is expressed at a high level in IM-9 B-cells, whereas it is undetectable in T-cells. Electrophoretic mobility supershift assays using T-and B-cell lymphoblast nuclear extracts demonstrated that Spi-B and PU.1 can bind to FP12 in vitro (data not shown). Thus, it seemed feasible that Spi-B and PU.1 are the Ets-related proteins that may be involved in cell type-specific regulation of hGR 1A promoter expression in lymphoblasts. In particular, because PU.1 is only present at high levels in IM-9 Pre-B-cells, it seemed likely that this transcription factor is responsible for DEX-mediated down-regulation of the hGR 1A promoter in these cells. In addition, because the c-Ets and c-Myb binding sites in FP12 overlap (Fig. 1A), we postulated that c-Myb and certain Ets members (depending on the cell type) antagonize each other at this binding site to oppositely regulate the response of the hGR 1A promoter to hormone in different lymphoid cell types.

Regulation of the Hormonal Response of the hGR 1A Promoter in T-lymphoblasts by c-Myb and Ets
Proteins-To determine whether c-Myb and members of the c-Ets family of transcription factors can indeed affect the response of the hGR 1A promoter to DEX, we performed transient cotransfection experiments with an hGR 1A promoter/luciferase reporter gene. The overexpression of c-Myb in Jurkat cells did not greatly affect the responsiveness of the hGR 1A promoter to DEX induction (Fig. 3A). Although this might suggest that the c-Myb site in FP12 is not functional in vivo, this is not the case. The C-terminal truncated c-Myb variant (c-Myb(Ct)) lacking the negative regulatory domain can functionally stimulate the promoter activity by 3-4-fold (Fig. 3A), which means that the c-Myb-binding site is specific and capable of recruiting c-Myb protein in vivo. Thus, the inability of transfected full-length c-Myb to further increase the hormonal response (Fig. 3A) may indicate that sufficient endogenous c-Myb is already present in Jurkat cells. This is supported by the fact that overexpressing a specific functional inhibitor of c-Myb, the c-Myb DBD dominant negative protein, in Jurkat cells effectively suppresses the hGR 1A promoter response to hormone (Fig. 3B). These data suggest that FP12 is a functional c-Myb-binding site that mediates the hormonal response of the hGR 1A promoter activity in vivo.
Overexpression of c-Ets-1, c-Ets-2, or PU.1 in Jurkat cells substantially suppresses DEX induction of the hGR 1A promoter (Fig. 4). However, this also causes large increases in basal promoter activity for c-Ets-1 (2-3-fold; data not shown) and c-Ets-2 (3-4-fold; data not shown), although overexpression of PU.1 only slightly affects the basal promoter activity (data not shown). These results suggest that although Ets family members may be involved in basal promoter activity, c-Ets-1, c-Ets-2, and PU.1 block DEX activation of the hGR 1A promoter in this T-lymphoblast cell line. Interestingly, overexpression of another c-Ets  . c-Myb modulates hGR 1A promoter expression. A, effect of c-Myb overexpression on hGR 1A promoter activity. Jurkat cells were cotransfected with pXP1-1A ϩ41/ϩ269-Luc plus pcDNA3 (empty vector), pcDNA3-c-MybHA, or C-terminal truncated c-Myb, pCt. Luciferase assays were preformed as described under "Materials and Methods. " B, overexpression of a c-Myb dominant negative inhibitor blocks hGR 1A promoter stimulation by DEX. Jurkat cells were cotransfected with 1.5 g of pXP1-1A ϩ41/ϩ269-Luc, 1 g of pCYGR, 1 g of pCMV-␤-gal (for normalizing transfection efficiency), and either 1.5 g of pcDNA3 or 1.5 g of pcDNA3-c-Myb DBD. The luciferase activity was normalized to ␤-galactosidase activity, and the steroid responsiveness is plotted as the percentage of the EtOH control. Data from four separate experiments was used to obtain the mean and calculate the S.E. **, p Ͻ 0.01, and ***, p Ͻ 0.005 for steroid responsiveness versus the respective ethanol control.  DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 family member, Spi-B, does not block the response of the hGR 1A promoter to DEX (Figs. 4B and 5A), and this even results in a slight (though not statistically significant) increase in hormonal responsiveness of the promoter. PU.1 and Spi-B belong to the same Ets family subgroup, and in some cases they have overlapping activities and can be substituted for each other (40). However, Spi-B and PU.1 are able to interact with different cofactors, and they may have subtle differences in their preferred DNA-binding sequence or different affinity for the same DNAbinding site (40 -43). We currently do not know whether FP12 is a preferred DNA-binding site for either PU.1 or Spi-B.

c-Myb and c-Ets Proteins Modulate hGR 1A Promoter Expression
To determine whether the slightly stimulatory effect (or at least the lack of inhibition) of Spi-B is due to its direct interaction with FP12, we transfected Jurkat cells with a mutant Spi-B (⌬Spi-B) that cannot bind to DNA because of a mutated DNA-binding domain. Surprisingly, we observed not only an increase in basal activity of the hGR 1A reporter gene (data not shown) but also a significant increase in the hormonal responsiveness of the promoter (Fig. 5A) compared with wild-type Spi-B. Thus, even though both Spi-B and PU.1 can efficiently bind to FP12 using an electrophoretic mobility shift assay, a different mechanism independent of the binding of Spi-B to FP12 may be responsible for the Spi-B effect noted here. Nonetheless, DNA binding of some transcription factor to FP12 is clearly required for the stimulatory response of the hGR 1A promoter to DEX, because (as was seen for the dominant negative c-Myb DBD; Fig. 3B), the overexpression of a c-Ets-2 dominant negative DBD mutant also blocks hormonal induction in the human Jurkat T-lymphoblast line (Fig. 5B).
These studies indicate that the various c-Ets family members can have complex effects upon both basal and hormone-regulated hGR 1A promoter activity, and these effects may involve both DNA binding activity and DNA binding-independent effects. With regard to corticosteroid treatment of T-and B-lymphoblasts, the role of c-Ets family members does suggest a molecular mechanism for the opposite hormonal regulation seen in these cells types. Ets family members appear to be involved in maintaining basal hGR 1A promoter activity. The IM-9 pre-B-cell lymphoblast line, in which hGR 1A transcripts are downregulated upon hormone treatment (9,26), lacks Spi-B (which does not suppress promoter activity) but contains apparently high levels of PU.1 (Fig. 2), which strongly inhibits DEX-mediated hGR 1A promoter activity. T-cells (such as the Jurkat and CEM-C7 cell lines), in which DEX causes an auto-up-regulation of hGR 1A promoter activity, contain Spi-B but do not contain detectable levels of PU.1. Thus, it seems likely that GR binding to FP11 and PU.1 binding to FP12 in the IM-9 B-cell line result in the formation of a complex on the hGR 1A promoter that inhibits its transcription.
c-Myb Is Critical in the Auto-up-regulation of the hGR 1A Promoter by Hormone-Because FP12 has a consensus c-Myb-binding site, we determined whether c-Myb might be responsible for the hormone-induced activation of the hGR 1A promoter in T-lymphoblasts. Western blotting revealed that high levels of c-Myb protein were present in the CEM-C7 and Jurkat T-cell ALL lines but that it was undetectable in the IM-9 B-cell line (Fig. 6A). DEX treatment had no effect on the c-Myb levels in the T-lymphoblasts. This observation suggests a role for c-Myb in hGR 1A promoter up-regulation in T-lymphoblasts. If this were true, then supplying c-Myb to c-Myb-deficient IM-9 cells should result in an up-regulation of the hGR 1A promoter after DEX treatment in these cells. IM-9 cells were cotransfected with a c-Myb expression construct and the hGR 1A promoter luciferase reporter gene. Indeed, an increase in hormone-stimulated hGR 1A reporter up-regulation was obtained  Spi-B and ⌬Spi-B affect DEX induction of the hGR 1A promoter in T-lymphoblasts. A, Jurkat cells were cotransfected with 1.5 g of pXP1-1A ϩ41/ϩ269-Luc and 1.5 g of ⌬EB (empty vector), ⌬EB-Spi-B, or ⌬EB-⌬SpiB (alternative spliced Spi-B variant without the Ets domain that is required by DNA binding). All of the cultures were also transfected with 1 g of pCYGR (to provide functional GR) and 1 g of pCMV-␤-gal (normalization control). B, overexpression of the dominant negative Ets protein inhibitor, Ets-2 DBD blocks the DEX-mediated up-regulation of the hGR 1A promoter. Jurkat cells were cotransfected with 1.5 g of pXP1-1A ϩ41/ϩ269-Luc and 1.5 g of pcDNA3 (vector control) or pFN-Ets-2 DBD. All of the cultures were also transfected with 1 g of pCYGR (to provide functional GR) and 1 g of pCMV-␤-gal (normalization control). Three or four individual experiments were combined to obtain the mean and calculate the S.E. The steroid responsiveness is plotted as the percentage of the EtOH control. **, p Ͻ 0.01, and ***, p Ͻ 0.005 for steroid responsiveness versus the respective ethanol control. FIGURE 6. c-Myb is required for steroid-mediated up-regulation of the hGR 1A promoter in lymphoblasts. A, Western blot analysis of c-Myb protein levels in CEM-C7, Jurkat, and IM-9 cells, which were treated with EtOH (ET) vehicle or 1 M DEX for 24 h. 40 g of total protein were subjected to 8% PAGE. After transfer to nitrocellulose membrane, the blots were probed with anti-c-Myb or anti-actin antibodies (loading control) and developed using ECL. B, IM-9 cells were electroporated with 3 g of pXP1-1A ϩ41/ ϩ269-Luc, 2 g of pCMV-␤-gal (for normalization), and increasing amounts of the c-Myb expression construct, pcDNA3-c-MybHA (0, 0.5, 1.0, 2.0, or 5.0 g). For each transfection, the pcDNA3 empty vector was added to maintain the total amount of transfected DNA at 10 g/reaction. Three individual experiments were combined to obtain the mean and calculate the S.E. The steroid responsiveness is plotted as the percentage of the EtOH control. *, p Ͻ 0.05 for steroid responsiveness versus the respective ethanol control.

c-Myb and c-Ets Proteins Modulate hGR 1A Promoter Expression
with increasing levels of transfected c-Myb cDNA (Fig. 6B). These data suggest that the recruitment of the GR to FP11 along with c-Myb recruitment to FP12 is critical for the auto-up-regulation of hGR 1A promoter in T-lymphoblasts, whereas hormone-mediated down-regulation of the 1A reporter is due to the recruitment of PU.1 to FP12 in IM-9 B-lymphoblasts.
c-Myb and PU.1 Are Involved in Regulating Hormonal Responsiveness of the hGR 1A Promoter in Lymphoblast Cells in Vivo-Using chromatin immunoprecipitation analysis, we investigated the involvement of specific transcription factors in hormone-treated CEM-C7 and IM-9 cells in vivo. After 24 h of DEX treatment, both the GR and c-Myb were recruited to the hGR 1A promoter region containing FP11 and FP12 in CEM-C7 cells, compared with the EtOH-treated sample (Fig. 7A). No recruitment of GR or c-Myb was observed in the negative controls (hGR 1A promoter upstream region and the PGK exon sequence), indicating that the recruitment of the GR and c-Myb was specific for the hGR 1A ϩ146/ϩ316 sequence (which contains FP11 and FP12). Together with the results obtained using overexpression and reporter gene analysis, it appears that the hGR and c-Myb, acting in concert, may assist in forming an active transcription complex on the hGR 1A promoter during DEX induction in CEM-C7 cells.
As expected, no c-Myb was recruited to the hGR 1A promoter in IM-9 cells, which lack this transcription factor, whereas the GR was recruited in response to hormone treatment (Fig. 7B). However, an increased recruitment of PU.1 to this hGR 1A promoter region was observed in DEX-treated IM-9 cells (Fig. 7B). Because the overexpression of PU.1 can repress hGR 1A ϩ41/ϩ269 promoter activity in T-lymphoblasts (Fig. 4B), the DEX-induced increase of PU.1 recruitment to the hGR 1A promoter in IM-9 B-lymphoblasts may represent the in vivo mechanism for hormone-mediated down-regulation of the hGR 1A promoter in this cell line. Similar to the corecruitment of GR and c-Myb in CEM-C7 cells, the simultaneous recruitment of the GR (to FP11) and PU.1 (to FP12) may assist in forming a complex capable of repressing transcription of the hGR 1A promoter in IM-9 B-cells. Thus, we propose that the expression pattern of specific transcription factors, such as PU.1 and c-Myb, in different lymphoblast cell lineages may be largely responsible for determining the direction of the auto-regulatory response of the hGR 1A promoter to hormone.
c-Myb May Also Regulate Hormonal Responsiveness of the hGR 1B and 1C Promoters in Lymphoblasts-At least three promoters (1A, 1B, and 1C) control the expression of the human GR gene (9,10,29). In contrast to the hGR 1A promoter, the 1B and 1C promoters are GC-rich and resemble promoters found in "housekeeping" genes (10,29,30,45,46). Although DEX induces significant in vivo up-regulation of all three promoters in CEM-C7 cells, no consensus GREs have been identified in any of them (9,26,27,31). The hGR 1A promoter is the most hormonesensitive in T-lymphoblasts. Because all three promoters are coordinately up-regulated (albeit to different levels), we wished to determine whether the same molecular mechanism involving c-Myb may be operative for all three promoters. Fig. 8 shows that luciferase reporter genes containing hGR promoter 1A, 1B, or 1C are all up-regulated by hormone in Jurkat T-ALL cells. The coexpression of the c-Myb DBD dominant negative mutant protein was effective in blocking the up-regulation of all three promoters. Thus, it appears that a similar molecular mechanism as that proposed for the hGR 1A promoter (see below) may control expression of the 1B and 1C promoters as well. Future experiments will determine whether composite response elements, as was found for the hGR 1A promoter (Fig. 1A), may exist for the hGR 1B and 1C promoters as well.

DISCUSSION
Corticosteroids are effective agents in treating certain types of leukemia, because they trigger apoptosis in sensitive cells. Whereas the abso- To analyze the specificity of the ChIP assay, two additional controls for each sample were included; an hGR 1A upstream (Ϫ3416/Ϫ3616) amplicon is located 3.6 kilobase pairs upstream of hGR the 1A (ϩ146/ϩ316) sequence, and it contains no known GRE-, c-Myb-, or c-Ets-binding site; and 2) a phosphoglycerate kinase 1 coding exon (ϩ396/ϩ798) sequence, which serves as a negative control. The PCR was performed as described under "Materials and Methods. " For the hGR 1A (ϩ146/ϩ316) and hGR 1A upstream sequences, 28 cycles (IM-9 cells) or 31 cycles (CEM-C7 cells) were used. For the PGK gene coding exon, 24 cycles were used. After gel electrophoresis, the EtBr-stained PAGE gels were photographed. A, ChIP assays of EtOH-or DEX-treated CEM-C7 cells using the GR and c-Myb antibodies. B, ChIP analysis of EtOH-or DEX-treated IM-9 cells using a c-Myb (top panel) or PU.1 and GR antibodies (bottom panels). Each ChIP experiment has been repeated at least four times to confirm the constant and reproducible patterns shown here.  DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 lute concentration of GR in cells prior to hormone treatment correlates somewhat to the sensitivity of the cell (13,16,21,24,25,28), the relationship between the steroid-sensitive phenotype and the GR level after auto-up-regulation by hormone treatment is much stronger (9,22,23,26,31,47). Indeed, the overexpression of functional GR can convert hormone-resistant T-ALL cells to a hormone-sensitive phenotype (18,21,23). Thus, it is important to understand the molecular mechanism for auto-up-regulation of hGR promoters in lymphoblasts (1,15,22,23,26,27,31,48).

c-Myb and c-Ets Proteins Modulate hGR 1A Promoter Expression
Steroid treatment of the hormone-sensitive, CEM-C7, T-cell ALL line causes auto-up-regulation of GR expression and apoptosis. Conversely, the IM-9, pre-B-cell, lymphoblastoid line auto-down-regulates GR expression and is resistant to hormone-mediated cell death (9,19,26,27,31). We had identified a composite hGR 1A promoter element (FP11/FP12) containing a nonconsensus GRE located adjacent to a DNA site containing overlapping core sequences for c-Myb and c-Ets protein family members. This element appears to be necessary and sufficient in conferring hormone responsiveness in T-lymphoblasts (9,31,39,49,50). We proposed that the direction of GR gene regulation by the binding of the GR to FP11 was determined by which transcription factor (c-Myb, c-Ets family member) simultaneously bound to FP12 (31). The results reported here strongly support this hypothesis.
c-Myb and c-Ets family members play pivotal roles in development and gene expression regulation in hematopoietic cells (39,40,42). Based upon the selective expression of c-Ets members in certain lymphoid cells, we focused on Ets-1, Ets-2, PU.1, and Spi-B in our studies. Of these, PU.1 appears to be the best candidate for controlling the auto-downregulation of GR gene expression, because it is present in a cell line that exhibits down-regulation (IM-9) and absent in T-cell lines (CEM-C7, Jurkat) that auto-up-regulate the hGR 1A promoter. PU.1 can interact with numerous other transcription factors and cofactors including TATA-binding protein (51), pRb (52), NF-EM5/Pip (53), and NF-IL6␤ (C/EBP␦) (54), CREB-binding protein (55), c-Myb, C/EBP ␣ (56), c-Fos (57), c-Jun (58), AML1 (59), and GR (60). PU.1 also causes transcrip-tional repression via its binding to the corepressor mSin3A and the formation of a complex including HDAC1 and via a direct interaction of PU.1 with MeCP2 in a mSin3A-HDAC repression complex (44). Thus, PU.1 is a bi-functional transcription factor that can activate or repress gene expression, depending upon the specific cofactors that it recruits to a promoter in a certain cell type. Because PU.1 is recruited to the hGR 1A promoter in DEX-treated IM-9 cells that auto-down-regulate promoter activity and suppresses hGR 1A promoter up-regulation in Jurkat T-cells, it is likely that PU.1 acts as a transcriptional repressor by recruiting corepressors that inhibit transcription of the hGR 1A promoter (Fig.  9). Future studies will reveal the other protein cofactors that are recruited to the hGR 1A promoter in IM-9 cells and that contribute to hGR 1A promoter auto-down-regulation.
In contrast to PU.1, c-Myb is expressed in Jurkat and CEM-C7 T-cell lymphoblasts (which exhibit DEX-mediated auto-up-regulation hGR1A promoter activity), whereas it is absent in IM-9 B-lymphoblasts. ChIP analysis in Jurkat cells shows the recruitment of c-Myb to the hGR 1A promoter after hormone treatment, and the expression of c-Myb in c-Myb-naïve IM-9 B-lymphoblasts via transient transfection results in the hGR 1A promoter becoming auto-up-regulated by steroid treatment. Thus, we propose that up-regulation of the hGR 1A promoter in T-lymphoblasts involves the simultaneous binding of the GR to FP11 and c-Myb to FP12, followed by the recruitment of coactivators to the hGR 1A promoter (Fig. 9).
Because of the overlapping c-Myb and c-Ets consensus binding sites, the configuration of FP12 lends itself to act as a molecular switch (Fig. 9). Thus, either c-Myb or a c-Ets family member, but not both, could bind to FP12 in a mutually exclusive manner. Because these transcription factors would compete for binding to the same DNA sequence, the relative levels and cell type specificity of c-Myb and c-Ets family members could largely control the direction of auto-regulation of GR gene expression when the GR binds to FP11. We have clearly shown that c-Myb is a transactivator of hGR 1A promoter expression, whereas PU.1 acts as a repressor. Although PU.1 (because of its restricted expression in the cell types tested) remains a prime candidate for the c-Ets family member involved in auto-down-regulation of the hGR 1A promoter, other c-Ets family members found in hematopoietic cells, such as Erg and Fli-1 (40), must also be tested. Further, the role (if any) of the unusual, DNA binding-independent activity of the c-Ets family member, Spi-B, in T-lymphoblasts must be resolved. It will also be important to identify the coactivators and corepressors that are recruited to the hGR 1A promoter by the GR/c-Myb and GR/PU.1 complexes, respectively. These studies also need to be extended to additional leukemia cell lines that respond to steroid treatment by undergoing apoptosis (as well as those that are steroid-resistant) and to fresh human patient samples to determine whether the molecular switch mechanism that we have proposed is a fundamental, common molecular mechanism for hGR auto-regulation. Our preliminary results indicate that a similar mechanism may be involved in hGR promoter 1B and 1C auto-regulation as well.
Lastly, by identifying other signaling pathways that can increase hGR promoter activity and the cellular level of GR protein, it may be possible to improve the response of certain leukemias to therapy. If the level of GR can be elevated in a hormone-independent way by stimulating the activity or levels of transcription factors that activate the various hGR promoters, then it would be possible to achieve a more complete steroid-induced apoptotic response in the blast cells at initial presentation and perhaps even convert hormone-resistant disease to the hormonesensitive phenotype by elevating the GR above the threshold level needed to trigger steroid-mediated apoptosis. In the basal state, the hGR 1A promoter expresses a basal activity that drives transcription of the GR gene. This may or may not involve the binding of a c-Ets family member to FP12 in the hGR 1A promoter. Upon binding the glucocorticoid ligand, the GR is activated and binds to FP11, probably as a monomer. In T-lymphoblasts, such as CEM-C7 cells, this causes the recruitment and stabilization of c-Myb binding to FP12 of the hGR 1A promoter. The GR/c-Myb complex then recruits coactivators to the hGR 1A promoter, resulting in the observed auto-upregulation of hGR 1A promoter activity by hormone. In B-lymphoblasts, such as IM-9 cells, the binding of the liganded GR protein to FP11 recruits and stabilizes the binding of the c-Ets family member, PU.1. The GR/PU.1 complex then recruits corepressors to the hGR 1A promoter, resulting in the observed auto-down-regulation of hGR 1A promoter activity by hormone seen in this cell type. Thus, the relative expression and level of c-Myb or c-Ets family members are largely responsible for the direction and magnitude of GR promoter auto-regulation seen in different cell types.