Interaction between the Interferon Signaling Pathway and the Human Glucocorticoid Receptor Gene 1A Promoter
http://www.100md.com
内分泌学杂志 2005年第3期
Department of Biochemistry and Molecular Biology and Stanley S. Scott Cancer Center, Louisiana State University Health Sciences Center, New Orleans, Louisiana 70112
Address all correspondence and requests for reprints to: Dr. Wayne V. Vedeckis, Department of Biochemistry and Molecular Biology , Louisiana State University Health Sciences Center, 533 Bolivar Street, New Orleans, Louisiana 70112. E-mail: wvedec@lsuhsc.edu.
Abstract
The newly described 1A promoter of the human glucocorticoid receptor (hGR) gene contains an interferon (IFN) regulatory factor element (IRF-E), a binding motif for the family of proteins termed IFN regulatory factors (IRFs) that are regulated by IFNs. To examine the in vivo role of IFNs in hGR gene regulation, human T cell lines (CEM-C7 and Jurkat) were treated with IFN. IFN rapidly induces the expression of IRF-1 proteins in a dose- and time-dependent manner. Luciferase expression is induced by IFN treatment in Jurkat cells transfected with an hGR 1A promoter IRF-E/luciferase reporter gene, but induction is lost with deletion of the IRF-E. Electrophoretic mobility shift and supershift analyses indicate an increase in the binding of IRF-1 to oligonucleotides containing the hGR 1A promoter IRF-E after IFN treatment, whereas IRF-2 binding to this oligonucleotide is unchanged. Human IRF-1 and IRF-2 proteins expressed in Chinese hamster ovary cells bind to the hGR 1A promoter IRF-E; however, only IRF-1 activates transcription. Although IFNs clearly activate a transfected reporter gene containing the hGR 1A promoter in T cells, they do not alter the sensitivity of CEM-C7 cells to glucocorticoid-induced apoptosis. Additional studies revealed that the glucocorticoid steroid hormone, dexamethasone (DEX), completely blocks IFN induction of IRF-1 mRNA levels. This could explain the lack of any greater apoptotic response to a combination of DEX plus IFN compared with the response to DEX alone. In addition, treatment with IFN alone does not alter endogenous GR mRNA levels (including exon 1A-containing transcripts derived from the hGR 1A promoter) in T lymphoblast cells, even though IRF-1 levels are induced. The difference in IRF-1-driven transcription between the hGR 1A reporter construct and the endogenous hGR 1A promoter could potentially be due to epigenetic effects, such as methylation.
Introduction
ALTHOUGH THE HUMAN glucocorticoid receptor (hGR) is ubiquitously expressed, regulation of the hGR gene is complex, leading to differences in the absolute concentration of GR transcript and protein in individual cells (1, 2, 3). For instance, treatment with glucocorticoids elicits a down-regulation of GR expression in most cells, a process that may help avoid the detrimental effects associated with chronic exposure to these steroids (reviewed in Ref.4). T cells are unique, in that glucocorticoids induce hGR mRNA and protein, eventually leading to programmed cell death (apoptosis) (5). Because of their effect on T cells, glucocorticoids are important therapeutic agents in pathological conditions involving neoplastic T cells, such as acute lymphoblastic leukemia (ALL) (reviewed in Ref.6). Corticosteroid treatment is also effective in patients with pre-B cell and early pre-B cell leukemia as well (7, 8). Although most patients are initially responsive to glucocorticoid therapy, many are glucocorticoid resistant upon relapse. In a cell line derived from a T-cell ALL patient, deleterious mutations in the GR-coding sequence result in glucocorticoid resistance (reviewed in Ref.9). However, very few GR mutations have been identified to date in human patients (reviewed in Ref.10). In other steroid-resistant cells isolated from this cell line that do not harbor such mutations, glucocorticoid resistance may be due to decreased concentrations of functional GR per T cell (11). This suggests a disruption of normal GR gene regulation and expression. Because it is clear that the level of functional GRs correlates positively with the responsiveness of leukemia patients to corticosteroid therapy (7, 8, 12, 13), elevating the GR level in patient blast cells would probably improve the clinical outcome.
Transcription of the mouse GR gene initiates from at least five different promoter regions (14, 15). Only a single promoter region (1C) for the hGR gene had been described until recently, when two other hGR promoter regions (1A and 1B) were identified (3, 16, 17). Regions within the hGR 1A promoter are responsible for the autologous up-regulation of the hGR gene by glucocorticoids (18). Surprisingly, this promoter also contains a site that could confer interferon (IFN) responsiveness.
Upon binding their respective membrane receptor, IFNs initiate a phosphorylation cascade that results in the activation of signal transducers and activators of transcription (STATs) (reviewed in Ref.19). These transcription factors, in turn, regulate a number of IFN-responsive genes, including a family of transcription factors termed IFN regulatory factors (IRFs) that are necessary for downstream induction of many other IFN-responsive genes. The consensus IFN response element (IRF-E) for two members of the IRF family (IRF-1 and IRF-2) is G(A)AAA(G or C)(T or C) GAAA(G or C)(T or C) (20). Although the IRF proteins share significant homology within their DNA-binding domains, they display very different biochemical characteristics and activities. IRF-1 has a very short half-life within a cell (30 min), whereas the half-life of IRF-2 is much longer (8 h) (21). This difference in protein longevity plays a role in regulating the cellular response to IFN. Normally, the more abundant and long-lived IRF-2 (which binds to, but does not activate, transcription from an IRF-E) effectively competes with IRF-1 (which acts as a transcriptional activator from an IRF-E) and blunts IRF-1 activity. However, IFNs act to rapidly increase the intracellular concentration of IRF-1 protein to levels that can effectively compete for IRF-E-binding sites and thus induce IFN-responsive genes (22).
IFNs enhance the immune response to viral and bacterial challenges. IFNs are also commonly used in the treatment of leukemia (reviewed in Ref.23). There are indications that IRF-1 may act as a tumor suppressor, and loss of function mutations in IRF-1 have been implicated in some forms of leukemia and other neoplasias (reviewed in Ref.24). However, the link between IFN, IRF-1, and such pathologies is still unclear. The discovery of an IRF-E within the hGR gene 1A promoter was intriguing, given the importance of glucocorticoids to T cell physiology.
The studies described here were performed to determine whether the IRF-E in the hGR 1A promoter is functional, and if it can be used as a target to improve the apoptotic response of T lymphoblasts to steroid therapy. Although the IRF-E of the hGR 1A promoter is functional when driving the expression of a reporter gene, IFN treatment does not enhance glucocorticoid-mediated apoptosis. Rather, glucocorticoids cause a previously unknown inhibition of IFN-induced IRF-1 expression. Endogenous GR transcripts arising from the hGR 1A promoter are not up-regulated by IFNs, suggesting that some cellular mechanism(s) renders the IRF-E in the hGR 1A promoter unresponsive to IFNs.
Materials and Methods
Cell culture
Human CEM-C7 T cell ALL cells (a gift from Dr. E. Brad Thompson, University of Texas Medical Branch, Galveston, TX) were grown in RPMI 1640 supplemented with 10% dialyzed, heat-inactivated, fetal bovine serum (Invitrogen Life Technologies, Inc., Gaithersburg, MD). Human Jurkat T cell ALL cells (American Type Culture Collection, Manassas, VA) were grown in RPMI 1640 supplemented with 10% fetal bovine serum. Chinese hamster ovarian cells infected with the polyoma virus (CHOP cells) (25) were grown in DMEM supplemented with 10% fetal bovine serum and 2 mM L-glutamine.
Western blot analysis of IFN-treated cells
CEM-C7 and Jurkat cells (2 x 105 cells/ml) were seeded in six-well tissue culture plates. Twenty-four hours later, they were treated with IFN, IFN?, or IFN (human; Sigma-Aldrich Corp., St. Louis, MO) at 10–100 U/ml. Cells were collected at various times and lysed in Laemmli sample buffer containing 5 μg/ml each of antitrypsin, aprotinin, chymostatin, leupeptin, and pepstatin A and 1 mM phenylmethylsulfonylfluoride (all from Sigma-Aldrich Corp.). The protein concentration in each sample was determined using the DC Protein Assay System (Bio-Rad Laboratories, Inc., Hercules, CA) and BSA as a protein standard. Twenty-five micrograms of total protein were electrophoresed in a denaturing acrylamide gel, and the proteins were transferred to nitrocellulose. Antihuman IRF-1 and IRF-2 antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were used for Western blot analysis. After characterizing the IFN dose response of IRF-1 in these cells, the rate of induction was examined by treating the cells in T-150 flasks with IFN at a final concentration of 100 U/ml. Aliquots of these cells were collected after 0, 1, 2, 4, and 24 h of IFN treatment and subjected to Western blot analysis.
RT-PCR
Quantification of IRF-1, IRF-2, and GR transcripts in samples of total RNA was performed by real-time quantitative RT-PCR (QRT-PCR) with TaqMan probes (Applied Biosystems, Foster City, CA) containing 6-carboxyfluorescein at the 5' end and the quencher 6-carboxy-N,N,N',N'-tetramethylrhodamine at the 3' end. Total GR transcripts and GR transcripts containing exon 1A3 were determined as described previously (26). For the IRF-1 assay, the forward primer, reverse primer, and the TaqMan probe had the sequences 5'-TGAGGAAGGGAAATTACCTGAGG-3', 5'-CACGTTTGTTGGCTGCCA-3', and 5'-TCCGACTGCTCCAAGAGCTTCATGATG-3', respectively. For the IRF-2 assay, the sequences were 5'-GCGAGCTCTACCCTCTGCAG-3', 5'-TGGGCACACTATCAGTCGTTTC-3', and 5'-TCTCCCCCGTGTCTTCCTATGCAGAAAG-3'. Dilutions of total RNA from CEM-C7 cells were used as relative standards for the IRF-1 and IRF-2 assays. The transcripts were quantified relative to 18S rRNA determined by real-time QRT-PCR as described previously (26).
DNA constructs
RT-PCR was used to amplify sequences encoding human IRF-1 and IRF-2 proteins using CEM-C7 total RNA. Oligonucleotide primers (sense, 5'-AATCGATCCTGCAGCACAGC-3'; antisense, 5'-CCAGGATCCCTGCTACGGT-3') were designed to amplify the entire coding sequence of human IRF-1. PCR products were ligated into pCRII (Invitrogen Life Technologies, Inc.) and sequenced to confirm IRF-1 identity and fidelity. An expression construct (pIRF1) was created by subcloning the coding sequence for human IRF-1 into the EcoRI and HindIII restriction sites of plasmid cytomegalovirus 5 (pCMV5). Oligonucleotide primers (sense, 5'-AGCAACACTGGTACCTTGCG-3' (KpnI site is underlined); antisense, 5'-AGTCAGAAGCTTAACAGCTC-3' (Hind III site is underlined) were designed to amplify the entire coding sequence of human IRF-2. An expression construct (pIRF2) was created by digesting the PCR product with Kpn I and HindIII and ligating it into the same sites in pCMV5. Clones were sequenced to confirm IRF-2 identity and fidelity. The hGR 1A IRF-E/pXP-1 luciferase reporter construct (sequence corresponding to nucleotides +41 to +269 of hGR exon 1A) and the hGR IRF-E deletion construct (hGR 1A IRF-E/pXP-1) were created as previously described (3).
Transfections and luciferase assays
Luciferase and ?-galactosidase (?-gal) assays were performed as described previously (16) with a Fluoroskan Ascent luminometer (Lab Systems, Helsinki, Finland). Luciferase gene expression was normalized to ?-gal activity.
To examine the effect of IFN treatment on hGR 1A promoter activity in T cells, Jurkat cells were seeded in six-well plates (2 x 106 cells/well) and transfected with 1.5 μg reporter construct (either hGR 1A IRF-E/pXP-1 or hGR 1A IRF-E/pXP-1) and 1 μg pCMV/?-gal construct using Superfect transfection reagent (Qiagen, Valencia, CA). Transfected cells were allowed to grow for 24 h, at which time 100 U/ml IFN, IFN?, or IFN were added to individual wells. Cells were collected 24 h after IFN addition for determination of luciferase and ?-gal activities.
To determine whether IRF proteins have a direct role in driving transcription from the hGR 1A promoter, CHOP cells were cotransfected with human IRF expression constructs and hGR 1A promoter constructs. CHOP cells were plated in six-well plates, grown to 60–80% confluence, and cotransfected with 1 μg reporter construct (empty pXP-1, hGR 1A IRF-E/pXP-1, or hGR 1A IRF-E/pXP-1), 0.25 μg pCMV/?-gal, and expression construct (1 μg pCMV5, 0.5 μg pIRF1, 0.5 μg pIRF2, or 0.5 μg of both pIRF1 and pIRF2). Empty pCMV5 was added to ensure that all transfections received equal amounts of DNA. Cells were collected 24 h after transfection as previously described (16).
EMSA
EMSAs were performed as previously described using 10 μg nuclear extract (16). Nuclear extracts from CEM-C7, Jurkat, and CHOP cells were obtained using the NE-PER kit (Pierce Chemical Co., Rockford, IL). IRF-E consensus oligonucleotide and mutant (eliminating IRF binding) IRF-E oligonucleotide were obtained from Geneka (Montréal, Canada). The hGR 1A IRF-E double-stranded oligonucleotide was described previously (3). Supershift analysis was performed after incubation at room temperature of the nuclear extracts with 2 μg IRF-1 or IRF-2 antibody (Santa Cruz Biotechnology, Inc.) 1 h before addition of labeled oligonucleotides.
Lactate dehydrogenase (LDH) assay
CEM-C7 cells (1 x 105 cells/ml in a 200-μl total volume) in 96-well tissue culture plates were treated with IFN, IFN?, or IFN (0, 1, 10, 100, 500, and 1000 U/ml) in 2 μl PBS or dexamethasone (DEX; 0, 1, 10, 100, 500, and 1000 nM) in 2 μl ethanol. To examine the effect of combinatorial treatment, other experiments were conducted with cells treated as described above, except IFN dose-response experiments also received 10 nM DEX in 2 μl ethanol, and DEX dose-response experiments also received 100 U/ml IFN in 2 μl PBS. After a 48-h incubation, cells were collected by centrifugation, and LDH activity in the medium was determined from an aliquot of the supernatant, as described by the manufacturer (reagent 228-500P, Sigma-Aldrich Corp.) in a microplate reader (Dynatech Laboratories, Chantilly, VA). Values were normalized to the LDH activity from wells in which cells were completely lysed in 20% Tween 20 (representing 100% death).
3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay
CEM-C7 cells (1 x 105 cells/ml in 200 μl total volume) were treated with IFN and DEX as described above. After 48 h, cell viability was assessed using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega Corp., Madison, WI). In this assay a tetrazolium compound, MTS (inner salt) is converted to a formazan product by living cells. The quantity of formazan product, as measured by the amount of absorbance at 490 nm, is directly proportional to the number of living cells in culture. Cells were incubated for an additional 4 h after the addition of AQueous One reagent (20 μl). Plates were then gently shaken to completely mix their contents, and the absorbance at 490 nm was measured with a 96-well plate reader (Bio-Rad Laboratories, Inc., Hercules, CA). Results were normalized to cells that received vehicle only, representing 100% survival.
Flow cytometric viability assay
CEM-C7 cells were seeded in six-well plates in 3-ml aliquots at a cell concentration of 2.5 x 105 cells/ml and treated with IFN, DEX, or a combination of both in 0.3 ml medium. Cell viability was assessed after 48 and 72 h by flow cytometry using the Vybrant Apoptosis Assay Kit 4 (V-13243, Molecular Probes, Eugene, OR). The assay was conducted essentially according to the manufacturer’s guidelines, except that a PBS wash was not conducted before staining, because a preliminary experiment indicated that this wash had essentially no effect on staining or viability. Measurements were performed on a FACSCalibur flow cytometer (BD Biosciences, Mountain View, CA). In a forward scatter vs. side scatter density plot, a region was set up corresponding to cells. Microscopic evaluation confirmed that the particles in the chosen region corresponded to cells. An FL1-H vs. FL3-H density plot was set up for cells gated through the selected cell region in the forward scatter vs. side scatter plot. A region encompassing nonfluorescently stained cells was then selected. Cells falling in this region were scored as living cells. Cells of higher fluorescence falling in the alternate region were scored as dead cells. The same regions were used for analyzing all samples.
Results
Treatment with IFN increases IRF-1 protein and binding of IRF-1 to the hGR 1A IRF-E
The discovery of an IRF-E in the hGR 1A promoter suggested IFN regulation of GR gene expression, most likely through the transcription factor IRF-1. Treatment of CEM-C7 and Jurkat (see below) cells with IFN induced IRF-1 protein in a time- and dose-dependent manner (Fig. 1A). The two bands observed may represent different forms of IRF-1 due to posttranslational modification or partial proteolysis. Transfection studies with an IRF-1 cDNA into CHOP cells gave the same two bands, indicating that these two bands are both authentic IRF-1 (see Fig. 3 below).
FIG. 1. IFN increases IRF-1 protein levels and activity in cultured human T lymphoblasts. A, CEM-C7 T-lymphoblast cells were treated with 100 U/ml IFN for the time indicated, and the cells were collected and lysed in Laemmli sample buffer. Total protein from collected cells was subjected to Western blot analysis using an antibody to human IRF-1. The lower panel is a dose response: Western blot analysis of CEM-C7 cells treated for 24 h with the indicated concentrations of IFN. The uniformity of protein loading in each lane was confirmed by Ponceau S staining of the filter before incubating the blot with antibody. B, IFN induces binding of IRF-1 to the 1A IRF-E. Nuclear extracts were isolated from CEM-C7 and Jurkat cells treated with PBS vehicle alone (basal) or 100 U/ml IFN for 4 h. A radiolabeled hGR 1A IRF-E oligonucleotide was mixed with nuclear extract from treated and untreated cells, and the resulting complexes were resolved using native PAGE and revealed by autoradiography. Other reactions included the addition of 2 μg anti-IRF-1 or anti-IRF-2. S, Major protein-DNA complexes that exhibit specific binding to the hGR 1A IRF-E; SS, supershifted complexes containing either IRF-1 or IRF-2.
FIG. 3. Heterologously expressed human IRF proteins bind to the hGR 1A IRF-E. A, CHOP cells were transfected with either the pIRF1 or pIRF2 construct and collected at 24 or 48 h after transfection. Proteins from untransfected cells (UT) and transfected cells were analyzed using SDS-PAGE and Western blot analysis. Equal protein loading was confirmed by Ponceau S staining. B, A labeled, double-stranded, oligonucleotide containing the hGR 1A IRF-E was mixed with nuclear extracts isolated from CHOP cells transfected with empty pCMV5, pIRF-1, or pIRF-2. The resulting complexes were resolved using native PAGE. Antibody supershift analysis involved the inclusion of 2 μg of either antihuman IRF-1 or antihuman IRF-2. S, IRF/DNA complexes; SS, antibody/IRF/DNA complexes.
EMSA was used to determine whether IRF proteins bind the IRF-E within the hGR 1A promoter. Previous studies had shown that an unlabeled hGR 1A IRF-E or an unlabeled consensus IRF-E could compete for binding of nuclear proteins to the labeled hGR 1A oligonucleotide (3), whereas competition was lost if the IRF-E was deleted. The same results were obtained when a consensus IRF-E was used as the labeled probe for EMSA instead of the hGR 1A IRF-E (data not shown). In the current study proteins within CEM-C7 and Jurkat cell nuclear extracts bind to and retard the mobility of labeled 1A IRF-E oligonucleotide (Fig. 1B, bands marked S), although the binding activity is fairly modest. To determine whether IFN induces IRF-1 binding to the hGR 1A IRF-E in human T cells, CEM-C7 and Jurkat cells were treated with IFN (100 U/ml) for 4 h, a treatment sufficient to strongly induce IRF-1 in these cells (Fig. 1A and data not shown). Nuclear protein binding to the hGR 1A IRF-E oligonucleotide was greatly increased in nuclear extracts from IFN-treated CEM-C7 and Jurkat cells compared with nuclear extract from untreated cells (Fig. 1B, bracket with arrow labeled S). The profile of these proteins is complex, with three or four shifted bands being seen in about the same region of the gel. A clearer picture of the specific IRF-1 and IRF-2 complexes was obtained when the respective cDNAs were transfected into CHOP cells (see Fig. 3). In the two lymphoblast cell lines (Fig. 1B), the addition of an IRF-1 antibody resulted in a supershift, and the amount of supershifted complex formed was greater in IFN-treated cells. There was no apparent increase in the amount of supershifted complex for either IFN-treated CEM-C7 or Jurkat cells when an IRF-2 antibody was used. Similar results were obtained with a radiolabeled consensus IRF-E double-stranded oligonucleotide (data not shown). IFN and IFN? treatment of CEM-C7 cells also resulted in some increased binding of nuclear extracts to the hGR 1A IRF-E, but the response was less robust (data not shown).
These results indicate that IRF-2, present in both untreated and IFN-treated CEM-C7 and Jurkat cells, is able to bind to the hGR 1A IRF-E. IRF-1 is also able to bind the hGR 1A IRF-E, but it is unable to form protein-DNA complexes, as analyzed by EMSA, in untreated cells. However, IFN strongly induces IRF-1 levels, and this results in easily detectable IRF-1 protein complexes with the hGR 1A IRF-E.
IFNs induce transcription from an hGR 1A IRF-E reporter construct in Jurkat cells
To examine whether IFN induces transcription from the IRF-E in the hGR 1A promoter, Jurkat cells were transfected with a hGR 1A promoter reporter construct containing the IRF-E and subsequently treated with 100 U/ml each of IFN, IFN?, and IFN. Jurkat cells were used because CEM-C7 cells are very difficult to transfect. Cloning of the +41/+269 hGR 1A sequence into the pXP-1 vector increased basal transcription about 5-fold compared with the empty vector (data not shown). All three IFNs induced luciferase activity from the hGR 1A IRF-E promoter construct, although IFN was more potent than IFN or IFN? (Fig. 2). This concentration of IFN was sufficient to cause increased IRF-1 binding to the IRF-E in an EMSA and substantially induced both IRF-1 protein and mRNA in Jurkat cells (see Figs. 6 and 7 below). Deletion of the hGR 1A IRF-E from the promoter sequence completely abolished IFN induction of luciferase transcription (Fig. 2). These results indicate that IFNs induce transcription from the hGR 1A IRF-E reporter construct transiently transfected into human T lymphoblast cells.
FIG. 2. IFN induces transcription from an hGR 1A IRF-E promoter construct. Jurkat cells were cotransfected with a pCMV/?-gal construct and with an hGR 1A promoter/luciferase reporter construct, with either the IRF-E intact or the IRF-E deleted, and treated 24 h later with 100 U/ml IFN, IFN?, IFN, or PBS vehicle. Cells were collected after 24 h of IFN treatment (48 h posttransfection). Luciferase activity was determined and normalized to ?-gal activity, and the normalized luciferase values were expressed relative to the values for the PBS-treated samples. Data represent the mean and SEM of four separate experiments. Asterisks indicate samples that showed statistically significant increases after IFN treatment compared with the PBS-treated control determined by t test for two samples, assuming equal variance (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
FIG. 6. IFN induces IRF-1 protein levels in T lymphoblasts. A, Jurkat T lymphoblast cells were treated with 100 U/ml IFN for 24 h, harvested, and extracted for protein. Western blot analyses were performed to quantify relative protein levels using antibodies specific for IRF-1 and IRF-2. Bands were scanned and quantified using the Quantity One image software (Bio-Rad Laboratories), and the results are normalized vs. the respective values obtained for tubulin in each sample. The results are given as the mean ± SEM for four independent experiments. Because Jurkat cells do not contain any functional GR, experiments with DEX treatment were not performed. B, CEM-C7 T lymphoblast cells were treated with 100 U/ml IFN and/or 1 μM DEX for 24 h. Western blotting was performed as described in A. The uniformity of protein loading in each lane was confirmed by Ponceau S staining of the filter before incubating the blot with antibody. The results are given as the mean ± SEM for four independent experiments.
FIG. 7. Effect of IFNs on IRF and GR mRNA levels in cultured human T lymphoblasts. CEM-C7 and Jurkat T lymphoblast cells, grown in six-well plates, were treated with IFNs for 24 h, harvested, and extracted for RNA. Two separate experiments, with each treatment given to three individual wells, were conducted. IRF-1 mRNA, IRF-2 mRNA, and exon 1A3-containing GR mRNA quantities were determined by real-time QRT-PCR, and the quantities were calculated relative to the amount of 18S rRNA, which was also determined by real-time QRT-PCR. The results are expressed relative to the vehicle control (PBS and 0.1% BSA). The error bars represent the SEM. Asterisks indicated samples that showed statistically significant increases after IFN treatment compared with the vehicle-treated control using the t test for two samples, assuming equal variance (*, P 0.05; **, P 0.01; ***, P 0.001.
Heterologously expressed IRF-1 drives transcription from the hGR 1A IRF-E
To examine whether heterologously expressed IRF-1 and IRF-2 would bind the hGR 1A IRF-E, nuclear extracts were isolated from CHOP cells 24 h after transfection with either pIRF-1 or pIRF-2 cDNA. Western blotting of cell extracts indicated that this period was sufficient for high level expression of both proteins (Fig. 3A). As shown previously for the Western blot with endogenous IRF-1 protein (Fig. 1A), two bands of IRF-1 were detected. Both the hGR 1A IRF-E and consensus IRF-E oligonucleotides (data not shown) were shifted by nuclear extracts from CHOP cells transfected with pIRF1 or pIRF2 (Fig. 3B, bracketed bands labeled S). Labeled oligonucleotides were not shifted to a significant degree by nuclear extract from CHOP cells transfected with empty pCMV5. Addition of anti-IRF-1 antibody resulted in a supershifted complex in nuclear extract from pIRF1-transfected CHOP cells, but not in nuclear extract from pIRF2-transfected cells. Similarly, addition of anti-IRF-2 antibody resulted in supershifted complexes in nuclear extract from pIRF2-transfected CHOP cells, but not in nuclear extract from pIRF1-transfected cells. These results provide unequivocal evidence that IRF-1 and IRF-2 are able to bind the hGR 1A IRF-E.
To determine whether heterologously expressed IRF proteins would drive transcription from the hGR 1A IRF-E, CHOP cells were cotransfected with IRF expression constructs and hGR 1A IRF-E/pXP-1 luciferase reporter constructs, and cell extracts were prepared 24 h after transfection. CHOP cells cotransfected with empty pCMV5 and the hGR 1A IRF-E/luciferase reporter construct displayed low basal expression of luciferase (Fig. 4). Cotransfection with pIRF1 and the hGR 1A IRF-E/luciferase reporter construct resulted in a 10-fold increase over basal luciferase gene expression (by ANOVA, P < 0.001), whereas cotransfection with pIRF-2 did not significantly increase luciferase expression compared with basal expression. Luciferase expression in CHOP cells cotransfected with both pIRF1 and pIRF2 was significantly greater than basal expression (by ANOVA, P < 0.01), but did not significantly differ from luciferase expression in cells transfected with pIRF-1 only. Luciferase gene expression was eliminated with deletion of the IRF-E from the hGR 1A IRF-E/pXP-1 promoter construct or when empty pXP-1 was used as the reporter construct regardless of the presence or absence of IRF proteins. We previously reported that deletion of the hGR 1A IRF-E also reduced basal promoter activity when this reporter gene assay was performed in Jurkat T lymphoblasts, which contain endogenous IRF-1 and IRF-2 (3). The present experiments clearly show that the hGR 1A IRF-E is a bona fide, functional IRF-E.
FIG. 4. Heterologously expressed human IRF-1 drives transcription from the hGR 1A IRF-E. CHOP cells were cotransfected with an expression construct (empty pCMV5, pIRF1, pIRF2, or both pIRF1 and pIRF2), a luciferase reporter construct (empty pXP-1, hGR 1A IRF-E/pXP-1, or mutant hGR 1A IRF-E/pXP-1 from which the IRF-E was deleted), and a CMV/?-gal construct (to normalize for transfection efficiency). Luciferase and ?-gal activity were determined 24 h after transfection. Data represent the mean and SEM of three transfections, and asterisks indicate a significant difference from basal (pCMV5-transfected) activity (**, P < 0.01; ***, P < 0.001).
IFNs do not increase the sensitivity of CEM-C7 cells to glucocorticoids
Because IFNs could potentially increase the sensitivity of T cells to glucocorticoids by increasing intracellular GR concentrations through the human GR 1A promoter, we examined the response of CEM-C7 cells to DEX in the presence and absence of IFNs. The results of the LDH assay indicated that treatment with 100-1000 nM DEX for 48 h results in 10–15% CEM-C7 cell mortality. There was no significant difference between cells treated with DEX and those treated with DEX and 100 U/ml IFN (Fig. 5A). Similar results were obtained with IFN and IFN? (data not shown). IFN treatment alone had no effect on cell viability, whereas treatment of cells with IFN and 10 nM DEX resulted in the same level of cell death as when cells were treated with 10 nM DEX alone.
FIG. 5. IFN does not increase T cell sensitivity to DEX. CEM-C7 cells were treated with the indicated concentrations of DEX alone, DEX plus 100 U/ml IFN, IFN, or IFN plus 10 nM DEX (as described in Materials and Methods). After 48 h, cell viability was determined using either an LDH assay (A) or an MTS assay (B). The percentage of cells surviving, compared with that of matched controls that received vehicle only (ethanol and PBS), is plotted against the log concentration of either DEX or IFN. There was no significant difference in cell viability between cells treated with DEX or with DEX and 100 U/ml IFN at any concentration. Data represent the mean and SEM of three (LDH assay) or four (MTS assay) experiments. C, CEM-C7 cells were treated with varying concentrations of DEX with or without 100 U/ml IFN. Results are presented as the percentage of living cells as assessed by flow cytometry of cells using the Vybrant Apoptosis Assay Kit 4, as described in Materials and Methods. Three separate experiments were performed. In each experiment, each treatment was given to two separate wells. Data represent the mean and SEM calculated for the six wells that received the same treatment.
The MTS assay, a more sensitive method of detecting cell viability, indicated that treatment with DEX resulted in higher CEM-C7 cell mortality (20–40%). However, there again were no significant differences between cells treated with DEX alone and cells treated with DEX and IFN (only the effects of IFN were examined with this assay; Fig. 5B).
Analysis of CEM-C7 cells with flow cytometry yielded similar results (Fig. 5C). Dose-response curves using DEX in the presence or absence of 100 U/ml IFN were indistinguishable at either 48 or 72 h of treatment; no left-shifting of the dose-response cell death curve occurred with IFN treatment. In a complementary experiment, CEM-C7 cells were treated with 10 nM DEX (a dose that only slightly impairs cell viability) and varying doses of IFN. Again, no effect of IFN was observed on CEM-C7 cell viability (data not shown). These results indicate that IFNs do not alter the sensitivity of these CEM-C7 T lymphoblasts to glucocorticoids.
DEX blocks IFN induction of IRF-1 mRNA and protein
We were perplexed by the fact that the hGR 1A IRF-E appeared to be perfectly functional, and yet IFN did not affect the apoptotic response of T lymphoblasts to glucocorticoids. Thus, we quantitatively reexamined the regulation of IRF-1 and IRF-2 mRNA and protein expression. Quantitative Western blots indicated a 4.6-fold increase in IRF-1 protein in Jurkat cells treated with IFN, whereas there was no significant increase in IRF-2 protein (Fig. 6A). The IFN-induced increase in IRF-1 mRNA and protein levels was consistent with the increase in IRF-1 binding to the IRF-E observed previously by EMSA (Fig. 1B). IFN also increased IRF-1 protein in CEM-C7 cells (Fig. 6B), but the induction was not as robust (2.6-fold) as in Jurkat cells. IRF-2 protein levels were not affected by IFN treatment. DEX alone reduced the amount of IRF-1 protein by about 50% and also caused a 50% increase in IRF-2 levels. Surprisingly, the IFN induction of IRF-1 protein levels was completely blocked when CEM-C7 cells were simultaneously treated with DEX (Fig. 6B), whereas there was a small, but not significant, increase in IRF-2 protein. Even though DEX blocked the induction of IRF-1, hGR protein levels were still induced 3.2-fold by DEX in the presence of IFN, similar to that observed previously with DEX treatment alone (Fig. 6B) (26). Recent studies have identified the steroid hormone-responsive sequences in the hGR 1A promoter, and these are apparently not dependent upon the activity of the hGR 1A IRF-E (18). IFN treatment alone also did not induce hGR protein levels (Fig. 6B).
These results were confirmed and extended at the mRNA level using QRT-PCR. There was an excellent correlation between the amount of IFN induction of IRF-1 mRNA levels in CEM-C7 (3.1-fold) and Jurkat (5-fold) cells (Fig. 7A) compared with protein induction levels for these cell lines (Fig. 6). Furthermore, IFN and IFN?, from a variety of sources, were not able to induce IRF-1 mRNA. IRF-2 was relatively unaffected by treatment with IFNs, although a slight increase was observed in both cell lines after IFN treatment (Fig. 7B). Finally, IFN did not up-regulate hGR transcripts. Total GR transcript concentrations, determined as the transcripts with the exon 5-exon 6 splice site by real-time QRT-PCR, were unchanged by IFN treatment in CEM-C7 cells (data not shown). Even though IRF-1 levels were significantly induced by IFN in both cell lines, the functional IRF-E in the hGR 1A promoter was apparently nonresponsive, because transcripts derived from this promoter were not affected by any IFN tested (Fig. 7C). Thus, even though transiently transfected reporter genes containing the hGR 1A promoter are responsive to IFN induction, the endogenous hGR 1A promoter IRF-E is not. Finally, we confirmed that the inhibition of IRF-1 protein levels by DEX alone and by DEX in IFN-treated cells (Fig. 6B) was mediated at the level of IRF-1 gene transcription, because DEX completely blocked IFN-induction of IRF-1 mRNA in CEM-C7 cells (Fig. 8).
FIG. 8. Effect of IFN and DEX on IRF mRNA in CEM-C7 T lymphoblasts. CEM-C7 T-lymphoblast cells were treated with different combinations of IFN and DEX for 24 h, harvested, and extracted for total RNA. Two separate experiments, with each treatment given to three individual wells, were conducted. IRF-1 mRNA and IRF-2 mRNA quantities were determined by real-time QRT-PCR, and the values were normalized to 18S rRNA, which was also determined by real-time QRT-PCR. The results are expressed relative to the vehicle control (ethanol, PBS, and 0.1% BSA). The error bars represent the SEM. Asterisks indicate samples with statistically significant changes compared with the vehicle-treated control using the t test for two samples, assuming equal variance (*, P 0.05; **, P 0.01; ***, P 0.001).
Discussion
The purpose of these studies was to determine how the human GR 1A promoter/exon sequence is regulated by IFNs and to elucidate the transcription factors and signal transduction pathways involved. Others had shown that up-regulation of GR protein levels is required for the apoptotic response in T lymphoblasts (27). Thus, a long-term goal of our research is to improve therapy in patients with T cell ALL by up-regulating GR protein levels via treatment with steroid plus other biological response modifiers that stimulate GR 1A promoter activity.
The GR 1A promoter/exon contains an intraexonic sequence that was nearly identical to the consensus DNA binding sequence for the IFN-stimulated transcription factor, IRF-1 (3). This was particularly interesting, because glucocorticoids and IFNs have opposing effects systemically. That is, corticosteroids are generally antiinflammatory, inhibit cell division, and promote cellular differentiation, whereas IFNs are proinflammatory and often stimulate cell proliferation. Thus, it was hoped that the combined treatment of T cell ALL patients with steroid and IFN would cause a synergistic or additive effect on T lymphoblast cell death, whereas the opposing actions of the two drugs systemically would spare the patient from the untoward side-effects resulting from either drug alone. Previous studies have not led to clear-cut results regarding the combined effects of corticosteroids and IFNs of lymphoid cell growth, cell division, and survival. Combined IFN? and DEX caused the greatest suppression of cell growth for a number of hemopoietic cancer cell lines, including MOLT-4 T lymphoblasts, although it was most pronounced for monocytic cells (28). IFN?1b also seemed to increase the amount of GR in a human monocyte cell line, and this enhanced the suppression of TNF production compared with the effect of DEX alone (29). An inhibitory effect of glucocorticoids on IFN signaling in human monocytes appears to be due to the inhibition of STAT1 gene expression by the steroid (30). Thus, interactions between the corticosteroid and IFN signal transduction pathways are of potential medical and biological importance.
The putative IRF-E in the GR 1A promoter/exon sequence is a bona fide IRF-E. EMSA demonstrated that it binds both IRF-1 and IRF-2. Treatment of T lymphoblasts with IFNs causes a substantial increase in IRF-1, resulting in an increase in IRF-1/hGR 1A IRF-E complexes, as revealed by EMSA. Thus, IFN treatment could potentially induce the synthesis of a stimulatory IRF (IRF-1) that could displace an inhibitory IRF (IRF-2), resulting in the increased expression of GR mRNA. The direct interaction of IRF proteins with the hGR 1A IRF-E was confirmed using a heterologous cell system (CHOP cells) that contains no detectable IRF-1 or IRF-2. Transfection of cDNAs coding for these proteins resulted in a large increase in the cellular levels of these proteins. Correspondingly, there was a large increase in nuclear protein binding to the hGR1A IRF-E, and EMSA supershift analysis clearly showed that this was due to the expressed IRF protein.
The biological activity of the hGR 1A IRF-E was also unambiguous. IFN treatment of cells transfected with a fragment of hGR 1A sequence containing the IRF-E stimulated the expression of a downstream luciferase reporter gene, and the internal deletion of the IRF-E alone eliminated IFN-mediated stimulation of the reporter gene. Cotransfection of an IRF-1 cDNA with the hGR 1A IRF-E luciferase reporter gene into CHOP cells caused a large increase in luciferase activity. The transfection of cells with an IRF-2 cDNA did not result in a statistically significant increase in luciferase expression compared with cells transfected with the empty vector, and although not statistically significant, the cotransfection of both IRF-1 and IRF-2 cDNAs seemed to diminish the effect of IRF-1. Perhaps IRF-2 was unable to compete effectively with IRF-1 because IRF-1 is expressed at higher levels than IRF-2 24 h after transfection. These experiments clearly show that the hGR 1A IRF-E is a bona fide IRF response element.
It was, thus, very surprising that the combined treatment with IFN and DEX does not result in increased cell death in CEM-C7 T lymphoblast cells. To be certain that we could detect even subtle changes in cell viability, we used three different assays. Consistent results are obtained using an LDH release assay, an MTS mitochondrial function assay, and flow cytometric analysis of apoptotic cells. IFN does not increase DEX-mediated T lymphoblast apoptosis.
Two additional studies revealed unexpected, and potentially important, results. First, DEX completely blocks the up-regulation of IRF-1 mRNA and protein in the CEM-C7 T lymphoblast cell line. A similar effect is seen in primary human monocytes, where glucocorticoids inhibit the expression of STAT1 and the expression of a downstream target of STAT1, the IRF-1 gene (30). Thus, the fact that combined DEX/IFN treatment does not cause an increase in T cell death can be explained at least in part by the DEX blockade of IFN-induced increases in IRF-1 mRNA and protein necessary for increased transcription from the hGR 1A IRF-E.
The second surprising result was that IFN alone did not increase the level of GR exon 1A3-containing transcripts in two T lymphoblast cell lines, even though IFN clearly induces IRF-1 expression in these cells, and reporter gene assays show that the hGR 1A IRF-E is completely functional. The most likely explanation for this discrepancy is that an epigenetic event has silenced the endogenous hGR 1A IRF-E. Although the nature of this endogenous gene silencing is unknown, preliminary studies indicate that gene methylation might be a likely candidate. Indeed, there is a CpG sequence in the IRF-E that would be a substrate for the methylase used in these studies, and in vitro methylation of the hGR 1A IRF-E/pXP-1 luciferase reporter construct abolishes its ability to be activated by IFN treatment (Geng, C.-d., and W. V. Vedeckis, data not shown). It is, however, possible that this IRF-E could be functional in cell types other than T cells.
In summary, the hGR gene 1A promoter contains a functional IRF-E. However, it appears that this promoter element is silenced in at least two T lymphoblast cell lines. This may be a mechanism to ensure that hormones with opposing actions (glucocorticoids and IFNs) do not yield the same response (up-regulation of the GR gene) in these cell types. A second mechanism to ensure that these two pathways do not result in the same signaling is the blockade of IFN-induced IRF-1 gene activation by glucocorticoids. Although the IFN- and steroid-activated pathways apparently cannot be used in concert to improve the killing of ALL T cells, it is hoped that future studies will reveal a signaling pathway that can activate hGR gene expression and improve the apoptotic response of these cancer cells to hormone.
References
Kalinyak JE, Dorin RI, Hoffman AR, Perlman AJ 1987 Tissue-specific regulation of glucocorticoid receptor mRNA by dexamethasone. J Biol Chem 262:10441–10444
Barrett TJ, Vedeckis WV 1996 Coordinate regulation of glucocorticoid and c-jun gene expression is cell type-specific and exhibits differential hormonal sensitivity for down- and up-regulation. Biochemistry 35:9746–9753
Breslin MB, Geng C-D, Vedeckis WV 2001 Multiple promoters exist in the human glucocorticoid receptor, one of which is activated by glucocorticoids. Mol Endocrinol 15:1381–1395
Burnstein KL, Bellingham DL, Jewell CM, Powell-Oliver FE, Cidlowski JA 1991 Autoregulation of glucocorticoid receptor gene expression. Steroids 56:52–58
Kofler R 2000 The molecular basis of glucocorticoid-induced apoptosis of lymphoblastic leukemia cells. Histochem Cell Biol 114:1–7
Gaynon PS, Lustig RH 1995 The use of glucocorticoids in acute lymphoblastic leukemia of children. J Pediatr Hematol Oncol 17:1–12
Quddus FF, Leventhal BG, Boyett JM, Pullen DJ, Crist WM, Borowitz MJ 1985 Glucocorticoid receptors in immunological subtypes of childhood acute lymphocytic leukemia cells: a Pediatric Oncology Group Study. Cancer Res 45:6482–6486
Kato GJ, Quddus FF, Shuster JJ, Boyett J, Pullen JD, Borowitz MJ, Whitehead VM, Crist WM, Leventhal BG 1993 High glucocorticoid receptor content of leukemic blasts is a favorable prognostic factor in childhood acute lymphoblastic leukemia. Blood 82:2304–2309
Hala M, Hartmann BL, B?ck G, Geley S, Kofler R 1996 Glucocorticoid-receptor-gene defects and resistance to glucocorticoid-induced apoptosis in human leukemic cell lines. Int J Cancer 68:663–668
Kofler R, Schmidt S, Kofler A, Ausserlechner MJ 1993 Resistance to glucocorticoid-induced apoptosis in lymphoblastic leukemia. J Endocrinol 178:19–27
Pui CH, Costlow ME 1986 Sequential studies of lymphoblast glucocorticoid receptor levels at diagnosis and relapse in childhood leukemia: an update. Leukemia Res 10:227–229
Bloomfield CD, Peterson BA, Zaleskas J, Frizzera G, Smith KA, Hildebrandt L, Gail-Peczalska KJ, Munck A 1980 In vitro glucocorticoid studies for predicting response to glucocorticoid therapy in adults with malignant lymphoma. Lancet 1:1952–1956
Bloomfield CD, Smith KA, Peterson BA, Munck A 1981 Glucocorticoid receptors in adult lymphoblastic leukemia. Cancer Res 41:4857–4860
Str?hle U, Schmidt A, Kelsey G, Stewart AF, Cole TJ, Schmid W, Schütz G 1992 At least three promoters direct expression of the mouse glucocorticoid receptor gene. Proc Natl Acad Sci USA 89:6731–6735
Chen F, Watson CS, Gametchu B 1999 Multiple glucocorticoid receptor transcripts in membrane glucocorticoid receptor-enriched S-49 mouse lymphoma cells. J Cell Biochem 74:418–429
Breslin MB, Vedeckis WV 1998 The human glucocorticoid receptor promoter upstream sequences contain binding sites for the ubiquitous transcription factor Yin Yang 1. J Steroid Biochem Mol Biol 67:369–381
Nunez BS, Vedeckis WV 2002 Characterization of promoter 1B in the human glucocorticoid receptor gene. Mol Cell Endocrinol 189:191–199
Geng C-d, Vedeckis WV 2004 Steroid-responsive sequences in the human glucocorticoid receptor gene 1A promoter. Mol Endocrinol 18:912–924
Stark GR, Kerr IM, Williams BR, Silverman RH, Schreiber RD 1998 How cells respond to interferons. Annu Rev Biochem 67:227–264
Tanaka N, Kawakami T, Taniguchi T 1993 Recognition DNA sequences of interferon regulatory factor-1 (IRF-1) and IRF-2, regulators of cell growth and the interferon system. Mol Cell Biol 13:4531–4538
Watanabe N, Sakakibara J, Hovanessian AG, Taniguchi T, Fujita T 1991 Activation of IFN-? element by IRF-1 requires a posttranslational event in addition to IRF-1 synthesis. Nucleic Acids Res 19:4421–4428
Vaughan PS, van Wijnen AJ, Stein JL, Stein GS 1997 Interferon regulatory factors: growth control and histone gene regulation–it’s not just interferon anymore. J Mol Med 75:348–359
Kalidas M, Kantarjian H, Talpaz M 2001 Chronic myelogenous leukemia. J Am Med Assoc 286:895–898
Taniguchi T, Lamphier MS, Tanaka N 1997 IRF-1: the transcription factor linking the interferon response and oncogenesis. Biochim Biophys Acta 1333:M9–M17
Heffernan M, Dennis JW 1991 Polyoma and hamster papovavirus large T antigen-mediated replication of expression shuttle vectors in Chinese hamster ovary cells. Nucleic Acids Res 19:85–92
Pedersen KB, Vedeckis WV 2003 Quantification and glucocorticoid regulation of glucocorticoid receptor transcripts in two human leukemic cell lines. Biochemistry 42:10978–10990
Ramdas J, Liu W, Harmon JM 1999 Glucocorticoid-induced cell death requires autoinduction of glucocorticoid receptor expression in human leukemic T cells. Cancer Res 59:1378–1385
Pan LY, Guyre PM 1988 Individual and combined tumoricidal effects of dexamethasone and interferons on human leukocyte cell lines. Cancer Res 48:567–571
Uitdehaag BM, Hoekstra K, Koper JW, Polman CH, Dijkstra CD 2001 IFN-?1b augments glucocorticoid-induced suppression of tumor necrosis factor- production by increasing the number of glucocorticoid receptors on a human monocytic cell line. J Interferon Cytokine Res 21:133–135
Hu X, Li W-P, Meng C, Ivashkiv LB 2003 Inhibition of IFN- signaling by glucocorticoids. J Immunol 170:4833–4839(B. Scott Nunez, Chuan-don)
Address all correspondence and requests for reprints to: Dr. Wayne V. Vedeckis, Department of Biochemistry and Molecular Biology , Louisiana State University Health Sciences Center, 533 Bolivar Street, New Orleans, Louisiana 70112. E-mail: wvedec@lsuhsc.edu.
Abstract
The newly described 1A promoter of the human glucocorticoid receptor (hGR) gene contains an interferon (IFN) regulatory factor element (IRF-E), a binding motif for the family of proteins termed IFN regulatory factors (IRFs) that are regulated by IFNs. To examine the in vivo role of IFNs in hGR gene regulation, human T cell lines (CEM-C7 and Jurkat) were treated with IFN. IFN rapidly induces the expression of IRF-1 proteins in a dose- and time-dependent manner. Luciferase expression is induced by IFN treatment in Jurkat cells transfected with an hGR 1A promoter IRF-E/luciferase reporter gene, but induction is lost with deletion of the IRF-E. Electrophoretic mobility shift and supershift analyses indicate an increase in the binding of IRF-1 to oligonucleotides containing the hGR 1A promoter IRF-E after IFN treatment, whereas IRF-2 binding to this oligonucleotide is unchanged. Human IRF-1 and IRF-2 proteins expressed in Chinese hamster ovary cells bind to the hGR 1A promoter IRF-E; however, only IRF-1 activates transcription. Although IFNs clearly activate a transfected reporter gene containing the hGR 1A promoter in T cells, they do not alter the sensitivity of CEM-C7 cells to glucocorticoid-induced apoptosis. Additional studies revealed that the glucocorticoid steroid hormone, dexamethasone (DEX), completely blocks IFN induction of IRF-1 mRNA levels. This could explain the lack of any greater apoptotic response to a combination of DEX plus IFN compared with the response to DEX alone. In addition, treatment with IFN alone does not alter endogenous GR mRNA levels (including exon 1A-containing transcripts derived from the hGR 1A promoter) in T lymphoblast cells, even though IRF-1 levels are induced. The difference in IRF-1-driven transcription between the hGR 1A reporter construct and the endogenous hGR 1A promoter could potentially be due to epigenetic effects, such as methylation.
Introduction
ALTHOUGH THE HUMAN glucocorticoid receptor (hGR) is ubiquitously expressed, regulation of the hGR gene is complex, leading to differences in the absolute concentration of GR transcript and protein in individual cells (1, 2, 3). For instance, treatment with glucocorticoids elicits a down-regulation of GR expression in most cells, a process that may help avoid the detrimental effects associated with chronic exposure to these steroids (reviewed in Ref.4). T cells are unique, in that glucocorticoids induce hGR mRNA and protein, eventually leading to programmed cell death (apoptosis) (5). Because of their effect on T cells, glucocorticoids are important therapeutic agents in pathological conditions involving neoplastic T cells, such as acute lymphoblastic leukemia (ALL) (reviewed in Ref.6). Corticosteroid treatment is also effective in patients with pre-B cell and early pre-B cell leukemia as well (7, 8). Although most patients are initially responsive to glucocorticoid therapy, many are glucocorticoid resistant upon relapse. In a cell line derived from a T-cell ALL patient, deleterious mutations in the GR-coding sequence result in glucocorticoid resistance (reviewed in Ref.9). However, very few GR mutations have been identified to date in human patients (reviewed in Ref.10). In other steroid-resistant cells isolated from this cell line that do not harbor such mutations, glucocorticoid resistance may be due to decreased concentrations of functional GR per T cell (11). This suggests a disruption of normal GR gene regulation and expression. Because it is clear that the level of functional GRs correlates positively with the responsiveness of leukemia patients to corticosteroid therapy (7, 8, 12, 13), elevating the GR level in patient blast cells would probably improve the clinical outcome.
Transcription of the mouse GR gene initiates from at least five different promoter regions (14, 15). Only a single promoter region (1C) for the hGR gene had been described until recently, when two other hGR promoter regions (1A and 1B) were identified (3, 16, 17). Regions within the hGR 1A promoter are responsible for the autologous up-regulation of the hGR gene by glucocorticoids (18). Surprisingly, this promoter also contains a site that could confer interferon (IFN) responsiveness.
Upon binding their respective membrane receptor, IFNs initiate a phosphorylation cascade that results in the activation of signal transducers and activators of transcription (STATs) (reviewed in Ref.19). These transcription factors, in turn, regulate a number of IFN-responsive genes, including a family of transcription factors termed IFN regulatory factors (IRFs) that are necessary for downstream induction of many other IFN-responsive genes. The consensus IFN response element (IRF-E) for two members of the IRF family (IRF-1 and IRF-2) is G(A)AAA(G or C)(T or C) GAAA(G or C)(T or C) (20). Although the IRF proteins share significant homology within their DNA-binding domains, they display very different biochemical characteristics and activities. IRF-1 has a very short half-life within a cell (30 min), whereas the half-life of IRF-2 is much longer (8 h) (21). This difference in protein longevity plays a role in regulating the cellular response to IFN. Normally, the more abundant and long-lived IRF-2 (which binds to, but does not activate, transcription from an IRF-E) effectively competes with IRF-1 (which acts as a transcriptional activator from an IRF-E) and blunts IRF-1 activity. However, IFNs act to rapidly increase the intracellular concentration of IRF-1 protein to levels that can effectively compete for IRF-E-binding sites and thus induce IFN-responsive genes (22).
IFNs enhance the immune response to viral and bacterial challenges. IFNs are also commonly used in the treatment of leukemia (reviewed in Ref.23). There are indications that IRF-1 may act as a tumor suppressor, and loss of function mutations in IRF-1 have been implicated in some forms of leukemia and other neoplasias (reviewed in Ref.24). However, the link between IFN, IRF-1, and such pathologies is still unclear. The discovery of an IRF-E within the hGR gene 1A promoter was intriguing, given the importance of glucocorticoids to T cell physiology.
The studies described here were performed to determine whether the IRF-E in the hGR 1A promoter is functional, and if it can be used as a target to improve the apoptotic response of T lymphoblasts to steroid therapy. Although the IRF-E of the hGR 1A promoter is functional when driving the expression of a reporter gene, IFN treatment does not enhance glucocorticoid-mediated apoptosis. Rather, glucocorticoids cause a previously unknown inhibition of IFN-induced IRF-1 expression. Endogenous GR transcripts arising from the hGR 1A promoter are not up-regulated by IFNs, suggesting that some cellular mechanism(s) renders the IRF-E in the hGR 1A promoter unresponsive to IFNs.
Materials and Methods
Cell culture
Human CEM-C7 T cell ALL cells (a gift from Dr. E. Brad Thompson, University of Texas Medical Branch, Galveston, TX) were grown in RPMI 1640 supplemented with 10% dialyzed, heat-inactivated, fetal bovine serum (Invitrogen Life Technologies, Inc., Gaithersburg, MD). Human Jurkat T cell ALL cells (American Type Culture Collection, Manassas, VA) were grown in RPMI 1640 supplemented with 10% fetal bovine serum. Chinese hamster ovarian cells infected with the polyoma virus (CHOP cells) (25) were grown in DMEM supplemented with 10% fetal bovine serum and 2 mM L-glutamine.
Western blot analysis of IFN-treated cells
CEM-C7 and Jurkat cells (2 x 105 cells/ml) were seeded in six-well tissue culture plates. Twenty-four hours later, they were treated with IFN, IFN?, or IFN (human; Sigma-Aldrich Corp., St. Louis, MO) at 10–100 U/ml. Cells were collected at various times and lysed in Laemmli sample buffer containing 5 μg/ml each of antitrypsin, aprotinin, chymostatin, leupeptin, and pepstatin A and 1 mM phenylmethylsulfonylfluoride (all from Sigma-Aldrich Corp.). The protein concentration in each sample was determined using the DC Protein Assay System (Bio-Rad Laboratories, Inc., Hercules, CA) and BSA as a protein standard. Twenty-five micrograms of total protein were electrophoresed in a denaturing acrylamide gel, and the proteins were transferred to nitrocellulose. Antihuman IRF-1 and IRF-2 antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were used for Western blot analysis. After characterizing the IFN dose response of IRF-1 in these cells, the rate of induction was examined by treating the cells in T-150 flasks with IFN at a final concentration of 100 U/ml. Aliquots of these cells were collected after 0, 1, 2, 4, and 24 h of IFN treatment and subjected to Western blot analysis.
RT-PCR
Quantification of IRF-1, IRF-2, and GR transcripts in samples of total RNA was performed by real-time quantitative RT-PCR (QRT-PCR) with TaqMan probes (Applied Biosystems, Foster City, CA) containing 6-carboxyfluorescein at the 5' end and the quencher 6-carboxy-N,N,N',N'-tetramethylrhodamine at the 3' end. Total GR transcripts and GR transcripts containing exon 1A3 were determined as described previously (26). For the IRF-1 assay, the forward primer, reverse primer, and the TaqMan probe had the sequences 5'-TGAGGAAGGGAAATTACCTGAGG-3', 5'-CACGTTTGTTGGCTGCCA-3', and 5'-TCCGACTGCTCCAAGAGCTTCATGATG-3', respectively. For the IRF-2 assay, the sequences were 5'-GCGAGCTCTACCCTCTGCAG-3', 5'-TGGGCACACTATCAGTCGTTTC-3', and 5'-TCTCCCCCGTGTCTTCCTATGCAGAAAG-3'. Dilutions of total RNA from CEM-C7 cells were used as relative standards for the IRF-1 and IRF-2 assays. The transcripts were quantified relative to 18S rRNA determined by real-time QRT-PCR as described previously (26).
DNA constructs
RT-PCR was used to amplify sequences encoding human IRF-1 and IRF-2 proteins using CEM-C7 total RNA. Oligonucleotide primers (sense, 5'-AATCGATCCTGCAGCACAGC-3'; antisense, 5'-CCAGGATCCCTGCTACGGT-3') were designed to amplify the entire coding sequence of human IRF-1. PCR products were ligated into pCRII (Invitrogen Life Technologies, Inc.) and sequenced to confirm IRF-1 identity and fidelity. An expression construct (pIRF1) was created by subcloning the coding sequence for human IRF-1 into the EcoRI and HindIII restriction sites of plasmid cytomegalovirus 5 (pCMV5). Oligonucleotide primers (sense, 5'-AGCAACACTGGTACCTTGCG-3' (KpnI site is underlined); antisense, 5'-AGTCAGAAGCTTAACAGCTC-3' (Hind III site is underlined) were designed to amplify the entire coding sequence of human IRF-2. An expression construct (pIRF2) was created by digesting the PCR product with Kpn I and HindIII and ligating it into the same sites in pCMV5. Clones were sequenced to confirm IRF-2 identity and fidelity. The hGR 1A IRF-E/pXP-1 luciferase reporter construct (sequence corresponding to nucleotides +41 to +269 of hGR exon 1A) and the hGR IRF-E deletion construct (hGR 1A IRF-E/pXP-1) were created as previously described (3).
Transfections and luciferase assays
Luciferase and ?-galactosidase (?-gal) assays were performed as described previously (16) with a Fluoroskan Ascent luminometer (Lab Systems, Helsinki, Finland). Luciferase gene expression was normalized to ?-gal activity.
To examine the effect of IFN treatment on hGR 1A promoter activity in T cells, Jurkat cells were seeded in six-well plates (2 x 106 cells/well) and transfected with 1.5 μg reporter construct (either hGR 1A IRF-E/pXP-1 or hGR 1A IRF-E/pXP-1) and 1 μg pCMV/?-gal construct using Superfect transfection reagent (Qiagen, Valencia, CA). Transfected cells were allowed to grow for 24 h, at which time 100 U/ml IFN, IFN?, or IFN were added to individual wells. Cells were collected 24 h after IFN addition for determination of luciferase and ?-gal activities.
To determine whether IRF proteins have a direct role in driving transcription from the hGR 1A promoter, CHOP cells were cotransfected with human IRF expression constructs and hGR 1A promoter constructs. CHOP cells were plated in six-well plates, grown to 60–80% confluence, and cotransfected with 1 μg reporter construct (empty pXP-1, hGR 1A IRF-E/pXP-1, or hGR 1A IRF-E/pXP-1), 0.25 μg pCMV/?-gal, and expression construct (1 μg pCMV5, 0.5 μg pIRF1, 0.5 μg pIRF2, or 0.5 μg of both pIRF1 and pIRF2). Empty pCMV5 was added to ensure that all transfections received equal amounts of DNA. Cells were collected 24 h after transfection as previously described (16).
EMSA
EMSAs were performed as previously described using 10 μg nuclear extract (16). Nuclear extracts from CEM-C7, Jurkat, and CHOP cells were obtained using the NE-PER kit (Pierce Chemical Co., Rockford, IL). IRF-E consensus oligonucleotide and mutant (eliminating IRF binding) IRF-E oligonucleotide were obtained from Geneka (Montréal, Canada). The hGR 1A IRF-E double-stranded oligonucleotide was described previously (3). Supershift analysis was performed after incubation at room temperature of the nuclear extracts with 2 μg IRF-1 or IRF-2 antibody (Santa Cruz Biotechnology, Inc.) 1 h before addition of labeled oligonucleotides.
Lactate dehydrogenase (LDH) assay
CEM-C7 cells (1 x 105 cells/ml in a 200-μl total volume) in 96-well tissue culture plates were treated with IFN, IFN?, or IFN (0, 1, 10, 100, 500, and 1000 U/ml) in 2 μl PBS or dexamethasone (DEX; 0, 1, 10, 100, 500, and 1000 nM) in 2 μl ethanol. To examine the effect of combinatorial treatment, other experiments were conducted with cells treated as described above, except IFN dose-response experiments also received 10 nM DEX in 2 μl ethanol, and DEX dose-response experiments also received 100 U/ml IFN in 2 μl PBS. After a 48-h incubation, cells were collected by centrifugation, and LDH activity in the medium was determined from an aliquot of the supernatant, as described by the manufacturer (reagent 228-500P, Sigma-Aldrich Corp.) in a microplate reader (Dynatech Laboratories, Chantilly, VA). Values were normalized to the LDH activity from wells in which cells were completely lysed in 20% Tween 20 (representing 100% death).
3-(4,5-Dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay
CEM-C7 cells (1 x 105 cells/ml in 200 μl total volume) were treated with IFN and DEX as described above. After 48 h, cell viability was assessed using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega Corp., Madison, WI). In this assay a tetrazolium compound, MTS (inner salt) is converted to a formazan product by living cells. The quantity of formazan product, as measured by the amount of absorbance at 490 nm, is directly proportional to the number of living cells in culture. Cells were incubated for an additional 4 h after the addition of AQueous One reagent (20 μl). Plates were then gently shaken to completely mix their contents, and the absorbance at 490 nm was measured with a 96-well plate reader (Bio-Rad Laboratories, Inc., Hercules, CA). Results were normalized to cells that received vehicle only, representing 100% survival.
Flow cytometric viability assay
CEM-C7 cells were seeded in six-well plates in 3-ml aliquots at a cell concentration of 2.5 x 105 cells/ml and treated with IFN, DEX, or a combination of both in 0.3 ml medium. Cell viability was assessed after 48 and 72 h by flow cytometry using the Vybrant Apoptosis Assay Kit 4 (V-13243, Molecular Probes, Eugene, OR). The assay was conducted essentially according to the manufacturer’s guidelines, except that a PBS wash was not conducted before staining, because a preliminary experiment indicated that this wash had essentially no effect on staining or viability. Measurements were performed on a FACSCalibur flow cytometer (BD Biosciences, Mountain View, CA). In a forward scatter vs. side scatter density plot, a region was set up corresponding to cells. Microscopic evaluation confirmed that the particles in the chosen region corresponded to cells. An FL1-H vs. FL3-H density plot was set up for cells gated through the selected cell region in the forward scatter vs. side scatter plot. A region encompassing nonfluorescently stained cells was then selected. Cells falling in this region were scored as living cells. Cells of higher fluorescence falling in the alternate region were scored as dead cells. The same regions were used for analyzing all samples.
Results
Treatment with IFN increases IRF-1 protein and binding of IRF-1 to the hGR 1A IRF-E
The discovery of an IRF-E in the hGR 1A promoter suggested IFN regulation of GR gene expression, most likely through the transcription factor IRF-1. Treatment of CEM-C7 and Jurkat (see below) cells with IFN induced IRF-1 protein in a time- and dose-dependent manner (Fig. 1A). The two bands observed may represent different forms of IRF-1 due to posttranslational modification or partial proteolysis. Transfection studies with an IRF-1 cDNA into CHOP cells gave the same two bands, indicating that these two bands are both authentic IRF-1 (see Fig. 3 below).
FIG. 1. IFN increases IRF-1 protein levels and activity in cultured human T lymphoblasts. A, CEM-C7 T-lymphoblast cells were treated with 100 U/ml IFN for the time indicated, and the cells were collected and lysed in Laemmli sample buffer. Total protein from collected cells was subjected to Western blot analysis using an antibody to human IRF-1. The lower panel is a dose response: Western blot analysis of CEM-C7 cells treated for 24 h with the indicated concentrations of IFN. The uniformity of protein loading in each lane was confirmed by Ponceau S staining of the filter before incubating the blot with antibody. B, IFN induces binding of IRF-1 to the 1A IRF-E. Nuclear extracts were isolated from CEM-C7 and Jurkat cells treated with PBS vehicle alone (basal) or 100 U/ml IFN for 4 h. A radiolabeled hGR 1A IRF-E oligonucleotide was mixed with nuclear extract from treated and untreated cells, and the resulting complexes were resolved using native PAGE and revealed by autoradiography. Other reactions included the addition of 2 μg anti-IRF-1 or anti-IRF-2. S, Major protein-DNA complexes that exhibit specific binding to the hGR 1A IRF-E; SS, supershifted complexes containing either IRF-1 or IRF-2.
FIG. 3. Heterologously expressed human IRF proteins bind to the hGR 1A IRF-E. A, CHOP cells were transfected with either the pIRF1 or pIRF2 construct and collected at 24 or 48 h after transfection. Proteins from untransfected cells (UT) and transfected cells were analyzed using SDS-PAGE and Western blot analysis. Equal protein loading was confirmed by Ponceau S staining. B, A labeled, double-stranded, oligonucleotide containing the hGR 1A IRF-E was mixed with nuclear extracts isolated from CHOP cells transfected with empty pCMV5, pIRF-1, or pIRF-2. The resulting complexes were resolved using native PAGE. Antibody supershift analysis involved the inclusion of 2 μg of either antihuman IRF-1 or antihuman IRF-2. S, IRF/DNA complexes; SS, antibody/IRF/DNA complexes.
EMSA was used to determine whether IRF proteins bind the IRF-E within the hGR 1A promoter. Previous studies had shown that an unlabeled hGR 1A IRF-E or an unlabeled consensus IRF-E could compete for binding of nuclear proteins to the labeled hGR 1A oligonucleotide (3), whereas competition was lost if the IRF-E was deleted. The same results were obtained when a consensus IRF-E was used as the labeled probe for EMSA instead of the hGR 1A IRF-E (data not shown). In the current study proteins within CEM-C7 and Jurkat cell nuclear extracts bind to and retard the mobility of labeled 1A IRF-E oligonucleotide (Fig. 1B, bands marked S), although the binding activity is fairly modest. To determine whether IFN induces IRF-1 binding to the hGR 1A IRF-E in human T cells, CEM-C7 and Jurkat cells were treated with IFN (100 U/ml) for 4 h, a treatment sufficient to strongly induce IRF-1 in these cells (Fig. 1A and data not shown). Nuclear protein binding to the hGR 1A IRF-E oligonucleotide was greatly increased in nuclear extracts from IFN-treated CEM-C7 and Jurkat cells compared with nuclear extract from untreated cells (Fig. 1B, bracket with arrow labeled S). The profile of these proteins is complex, with three or four shifted bands being seen in about the same region of the gel. A clearer picture of the specific IRF-1 and IRF-2 complexes was obtained when the respective cDNAs were transfected into CHOP cells (see Fig. 3). In the two lymphoblast cell lines (Fig. 1B), the addition of an IRF-1 antibody resulted in a supershift, and the amount of supershifted complex formed was greater in IFN-treated cells. There was no apparent increase in the amount of supershifted complex for either IFN-treated CEM-C7 or Jurkat cells when an IRF-2 antibody was used. Similar results were obtained with a radiolabeled consensus IRF-E double-stranded oligonucleotide (data not shown). IFN and IFN? treatment of CEM-C7 cells also resulted in some increased binding of nuclear extracts to the hGR 1A IRF-E, but the response was less robust (data not shown).
These results indicate that IRF-2, present in both untreated and IFN-treated CEM-C7 and Jurkat cells, is able to bind to the hGR 1A IRF-E. IRF-1 is also able to bind the hGR 1A IRF-E, but it is unable to form protein-DNA complexes, as analyzed by EMSA, in untreated cells. However, IFN strongly induces IRF-1 levels, and this results in easily detectable IRF-1 protein complexes with the hGR 1A IRF-E.
IFNs induce transcription from an hGR 1A IRF-E reporter construct in Jurkat cells
To examine whether IFN induces transcription from the IRF-E in the hGR 1A promoter, Jurkat cells were transfected with a hGR 1A promoter reporter construct containing the IRF-E and subsequently treated with 100 U/ml each of IFN, IFN?, and IFN. Jurkat cells were used because CEM-C7 cells are very difficult to transfect. Cloning of the +41/+269 hGR 1A sequence into the pXP-1 vector increased basal transcription about 5-fold compared with the empty vector (data not shown). All three IFNs induced luciferase activity from the hGR 1A IRF-E promoter construct, although IFN was more potent than IFN or IFN? (Fig. 2). This concentration of IFN was sufficient to cause increased IRF-1 binding to the IRF-E in an EMSA and substantially induced both IRF-1 protein and mRNA in Jurkat cells (see Figs. 6 and 7 below). Deletion of the hGR 1A IRF-E from the promoter sequence completely abolished IFN induction of luciferase transcription (Fig. 2). These results indicate that IFNs induce transcription from the hGR 1A IRF-E reporter construct transiently transfected into human T lymphoblast cells.
FIG. 2. IFN induces transcription from an hGR 1A IRF-E promoter construct. Jurkat cells were cotransfected with a pCMV/?-gal construct and with an hGR 1A promoter/luciferase reporter construct, with either the IRF-E intact or the IRF-E deleted, and treated 24 h later with 100 U/ml IFN, IFN?, IFN, or PBS vehicle. Cells were collected after 24 h of IFN treatment (48 h posttransfection). Luciferase activity was determined and normalized to ?-gal activity, and the normalized luciferase values were expressed relative to the values for the PBS-treated samples. Data represent the mean and SEM of four separate experiments. Asterisks indicate samples that showed statistically significant increases after IFN treatment compared with the PBS-treated control determined by t test for two samples, assuming equal variance (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
FIG. 6. IFN induces IRF-1 protein levels in T lymphoblasts. A, Jurkat T lymphoblast cells were treated with 100 U/ml IFN for 24 h, harvested, and extracted for protein. Western blot analyses were performed to quantify relative protein levels using antibodies specific for IRF-1 and IRF-2. Bands were scanned and quantified using the Quantity One image software (Bio-Rad Laboratories), and the results are normalized vs. the respective values obtained for tubulin in each sample. The results are given as the mean ± SEM for four independent experiments. Because Jurkat cells do not contain any functional GR, experiments with DEX treatment were not performed. B, CEM-C7 T lymphoblast cells were treated with 100 U/ml IFN and/or 1 μM DEX for 24 h. Western blotting was performed as described in A. The uniformity of protein loading in each lane was confirmed by Ponceau S staining of the filter before incubating the blot with antibody. The results are given as the mean ± SEM for four independent experiments.
FIG. 7. Effect of IFNs on IRF and GR mRNA levels in cultured human T lymphoblasts. CEM-C7 and Jurkat T lymphoblast cells, grown in six-well plates, were treated with IFNs for 24 h, harvested, and extracted for RNA. Two separate experiments, with each treatment given to three individual wells, were conducted. IRF-1 mRNA, IRF-2 mRNA, and exon 1A3-containing GR mRNA quantities were determined by real-time QRT-PCR, and the quantities were calculated relative to the amount of 18S rRNA, which was also determined by real-time QRT-PCR. The results are expressed relative to the vehicle control (PBS and 0.1% BSA). The error bars represent the SEM. Asterisks indicated samples that showed statistically significant increases after IFN treatment compared with the vehicle-treated control using the t test for two samples, assuming equal variance (*, P 0.05; **, P 0.01; ***, P 0.001.
Heterologously expressed IRF-1 drives transcription from the hGR 1A IRF-E
To examine whether heterologously expressed IRF-1 and IRF-2 would bind the hGR 1A IRF-E, nuclear extracts were isolated from CHOP cells 24 h after transfection with either pIRF-1 or pIRF-2 cDNA. Western blotting of cell extracts indicated that this period was sufficient for high level expression of both proteins (Fig. 3A). As shown previously for the Western blot with endogenous IRF-1 protein (Fig. 1A), two bands of IRF-1 were detected. Both the hGR 1A IRF-E and consensus IRF-E oligonucleotides (data not shown) were shifted by nuclear extracts from CHOP cells transfected with pIRF1 or pIRF2 (Fig. 3B, bracketed bands labeled S). Labeled oligonucleotides were not shifted to a significant degree by nuclear extract from CHOP cells transfected with empty pCMV5. Addition of anti-IRF-1 antibody resulted in a supershifted complex in nuclear extract from pIRF1-transfected CHOP cells, but not in nuclear extract from pIRF2-transfected cells. Similarly, addition of anti-IRF-2 antibody resulted in supershifted complexes in nuclear extract from pIRF2-transfected CHOP cells, but not in nuclear extract from pIRF1-transfected cells. These results provide unequivocal evidence that IRF-1 and IRF-2 are able to bind the hGR 1A IRF-E.
To determine whether heterologously expressed IRF proteins would drive transcription from the hGR 1A IRF-E, CHOP cells were cotransfected with IRF expression constructs and hGR 1A IRF-E/pXP-1 luciferase reporter constructs, and cell extracts were prepared 24 h after transfection. CHOP cells cotransfected with empty pCMV5 and the hGR 1A IRF-E/luciferase reporter construct displayed low basal expression of luciferase (Fig. 4). Cotransfection with pIRF1 and the hGR 1A IRF-E/luciferase reporter construct resulted in a 10-fold increase over basal luciferase gene expression (by ANOVA, P < 0.001), whereas cotransfection with pIRF-2 did not significantly increase luciferase expression compared with basal expression. Luciferase expression in CHOP cells cotransfected with both pIRF1 and pIRF2 was significantly greater than basal expression (by ANOVA, P < 0.01), but did not significantly differ from luciferase expression in cells transfected with pIRF-1 only. Luciferase gene expression was eliminated with deletion of the IRF-E from the hGR 1A IRF-E/pXP-1 promoter construct or when empty pXP-1 was used as the reporter construct regardless of the presence or absence of IRF proteins. We previously reported that deletion of the hGR 1A IRF-E also reduced basal promoter activity when this reporter gene assay was performed in Jurkat T lymphoblasts, which contain endogenous IRF-1 and IRF-2 (3). The present experiments clearly show that the hGR 1A IRF-E is a bona fide, functional IRF-E.
FIG. 4. Heterologously expressed human IRF-1 drives transcription from the hGR 1A IRF-E. CHOP cells were cotransfected with an expression construct (empty pCMV5, pIRF1, pIRF2, or both pIRF1 and pIRF2), a luciferase reporter construct (empty pXP-1, hGR 1A IRF-E/pXP-1, or mutant hGR 1A IRF-E/pXP-1 from which the IRF-E was deleted), and a CMV/?-gal construct (to normalize for transfection efficiency). Luciferase and ?-gal activity were determined 24 h after transfection. Data represent the mean and SEM of three transfections, and asterisks indicate a significant difference from basal (pCMV5-transfected) activity (**, P < 0.01; ***, P < 0.001).
IFNs do not increase the sensitivity of CEM-C7 cells to glucocorticoids
Because IFNs could potentially increase the sensitivity of T cells to glucocorticoids by increasing intracellular GR concentrations through the human GR 1A promoter, we examined the response of CEM-C7 cells to DEX in the presence and absence of IFNs. The results of the LDH assay indicated that treatment with 100-1000 nM DEX for 48 h results in 10–15% CEM-C7 cell mortality. There was no significant difference between cells treated with DEX and those treated with DEX and 100 U/ml IFN (Fig. 5A). Similar results were obtained with IFN and IFN? (data not shown). IFN treatment alone had no effect on cell viability, whereas treatment of cells with IFN and 10 nM DEX resulted in the same level of cell death as when cells were treated with 10 nM DEX alone.
FIG. 5. IFN does not increase T cell sensitivity to DEX. CEM-C7 cells were treated with the indicated concentrations of DEX alone, DEX plus 100 U/ml IFN, IFN, or IFN plus 10 nM DEX (as described in Materials and Methods). After 48 h, cell viability was determined using either an LDH assay (A) or an MTS assay (B). The percentage of cells surviving, compared with that of matched controls that received vehicle only (ethanol and PBS), is plotted against the log concentration of either DEX or IFN. There was no significant difference in cell viability between cells treated with DEX or with DEX and 100 U/ml IFN at any concentration. Data represent the mean and SEM of three (LDH assay) or four (MTS assay) experiments. C, CEM-C7 cells were treated with varying concentrations of DEX with or without 100 U/ml IFN. Results are presented as the percentage of living cells as assessed by flow cytometry of cells using the Vybrant Apoptosis Assay Kit 4, as described in Materials and Methods. Three separate experiments were performed. In each experiment, each treatment was given to two separate wells. Data represent the mean and SEM calculated for the six wells that received the same treatment.
The MTS assay, a more sensitive method of detecting cell viability, indicated that treatment with DEX resulted in higher CEM-C7 cell mortality (20–40%). However, there again were no significant differences between cells treated with DEX alone and cells treated with DEX and IFN (only the effects of IFN were examined with this assay; Fig. 5B).
Analysis of CEM-C7 cells with flow cytometry yielded similar results (Fig. 5C). Dose-response curves using DEX in the presence or absence of 100 U/ml IFN were indistinguishable at either 48 or 72 h of treatment; no left-shifting of the dose-response cell death curve occurred with IFN treatment. In a complementary experiment, CEM-C7 cells were treated with 10 nM DEX (a dose that only slightly impairs cell viability) and varying doses of IFN. Again, no effect of IFN was observed on CEM-C7 cell viability (data not shown). These results indicate that IFNs do not alter the sensitivity of these CEM-C7 T lymphoblasts to glucocorticoids.
DEX blocks IFN induction of IRF-1 mRNA and protein
We were perplexed by the fact that the hGR 1A IRF-E appeared to be perfectly functional, and yet IFN did not affect the apoptotic response of T lymphoblasts to glucocorticoids. Thus, we quantitatively reexamined the regulation of IRF-1 and IRF-2 mRNA and protein expression. Quantitative Western blots indicated a 4.6-fold increase in IRF-1 protein in Jurkat cells treated with IFN, whereas there was no significant increase in IRF-2 protein (Fig. 6A). The IFN-induced increase in IRF-1 mRNA and protein levels was consistent with the increase in IRF-1 binding to the IRF-E observed previously by EMSA (Fig. 1B). IFN also increased IRF-1 protein in CEM-C7 cells (Fig. 6B), but the induction was not as robust (2.6-fold) as in Jurkat cells. IRF-2 protein levels were not affected by IFN treatment. DEX alone reduced the amount of IRF-1 protein by about 50% and also caused a 50% increase in IRF-2 levels. Surprisingly, the IFN induction of IRF-1 protein levels was completely blocked when CEM-C7 cells were simultaneously treated with DEX (Fig. 6B), whereas there was a small, but not significant, increase in IRF-2 protein. Even though DEX blocked the induction of IRF-1, hGR protein levels were still induced 3.2-fold by DEX in the presence of IFN, similar to that observed previously with DEX treatment alone (Fig. 6B) (26). Recent studies have identified the steroid hormone-responsive sequences in the hGR 1A promoter, and these are apparently not dependent upon the activity of the hGR 1A IRF-E (18). IFN treatment alone also did not induce hGR protein levels (Fig. 6B).
These results were confirmed and extended at the mRNA level using QRT-PCR. There was an excellent correlation between the amount of IFN induction of IRF-1 mRNA levels in CEM-C7 (3.1-fold) and Jurkat (5-fold) cells (Fig. 7A) compared with protein induction levels for these cell lines (Fig. 6). Furthermore, IFN and IFN?, from a variety of sources, were not able to induce IRF-1 mRNA. IRF-2 was relatively unaffected by treatment with IFNs, although a slight increase was observed in both cell lines after IFN treatment (Fig. 7B). Finally, IFN did not up-regulate hGR transcripts. Total GR transcript concentrations, determined as the transcripts with the exon 5-exon 6 splice site by real-time QRT-PCR, were unchanged by IFN treatment in CEM-C7 cells (data not shown). Even though IRF-1 levels were significantly induced by IFN in both cell lines, the functional IRF-E in the hGR 1A promoter was apparently nonresponsive, because transcripts derived from this promoter were not affected by any IFN tested (Fig. 7C). Thus, even though transiently transfected reporter genes containing the hGR 1A promoter are responsive to IFN induction, the endogenous hGR 1A promoter IRF-E is not. Finally, we confirmed that the inhibition of IRF-1 protein levels by DEX alone and by DEX in IFN-treated cells (Fig. 6B) was mediated at the level of IRF-1 gene transcription, because DEX completely blocked IFN-induction of IRF-1 mRNA in CEM-C7 cells (Fig. 8).
FIG. 8. Effect of IFN and DEX on IRF mRNA in CEM-C7 T lymphoblasts. CEM-C7 T-lymphoblast cells were treated with different combinations of IFN and DEX for 24 h, harvested, and extracted for total RNA. Two separate experiments, with each treatment given to three individual wells, were conducted. IRF-1 mRNA and IRF-2 mRNA quantities were determined by real-time QRT-PCR, and the values were normalized to 18S rRNA, which was also determined by real-time QRT-PCR. The results are expressed relative to the vehicle control (ethanol, PBS, and 0.1% BSA). The error bars represent the SEM. Asterisks indicate samples with statistically significant changes compared with the vehicle-treated control using the t test for two samples, assuming equal variance (*, P 0.05; **, P 0.01; ***, P 0.001).
Discussion
The purpose of these studies was to determine how the human GR 1A promoter/exon sequence is regulated by IFNs and to elucidate the transcription factors and signal transduction pathways involved. Others had shown that up-regulation of GR protein levels is required for the apoptotic response in T lymphoblasts (27). Thus, a long-term goal of our research is to improve therapy in patients with T cell ALL by up-regulating GR protein levels via treatment with steroid plus other biological response modifiers that stimulate GR 1A promoter activity.
The GR 1A promoter/exon contains an intraexonic sequence that was nearly identical to the consensus DNA binding sequence for the IFN-stimulated transcription factor, IRF-1 (3). This was particularly interesting, because glucocorticoids and IFNs have opposing effects systemically. That is, corticosteroids are generally antiinflammatory, inhibit cell division, and promote cellular differentiation, whereas IFNs are proinflammatory and often stimulate cell proliferation. Thus, it was hoped that the combined treatment of T cell ALL patients with steroid and IFN would cause a synergistic or additive effect on T lymphoblast cell death, whereas the opposing actions of the two drugs systemically would spare the patient from the untoward side-effects resulting from either drug alone. Previous studies have not led to clear-cut results regarding the combined effects of corticosteroids and IFNs of lymphoid cell growth, cell division, and survival. Combined IFN? and DEX caused the greatest suppression of cell growth for a number of hemopoietic cancer cell lines, including MOLT-4 T lymphoblasts, although it was most pronounced for monocytic cells (28). IFN?1b also seemed to increase the amount of GR in a human monocyte cell line, and this enhanced the suppression of TNF production compared with the effect of DEX alone (29). An inhibitory effect of glucocorticoids on IFN signaling in human monocytes appears to be due to the inhibition of STAT1 gene expression by the steroid (30). Thus, interactions between the corticosteroid and IFN signal transduction pathways are of potential medical and biological importance.
The putative IRF-E in the GR 1A promoter/exon sequence is a bona fide IRF-E. EMSA demonstrated that it binds both IRF-1 and IRF-2. Treatment of T lymphoblasts with IFNs causes a substantial increase in IRF-1, resulting in an increase in IRF-1/hGR 1A IRF-E complexes, as revealed by EMSA. Thus, IFN treatment could potentially induce the synthesis of a stimulatory IRF (IRF-1) that could displace an inhibitory IRF (IRF-2), resulting in the increased expression of GR mRNA. The direct interaction of IRF proteins with the hGR 1A IRF-E was confirmed using a heterologous cell system (CHOP cells) that contains no detectable IRF-1 or IRF-2. Transfection of cDNAs coding for these proteins resulted in a large increase in the cellular levels of these proteins. Correspondingly, there was a large increase in nuclear protein binding to the hGR1A IRF-E, and EMSA supershift analysis clearly showed that this was due to the expressed IRF protein.
The biological activity of the hGR 1A IRF-E was also unambiguous. IFN treatment of cells transfected with a fragment of hGR 1A sequence containing the IRF-E stimulated the expression of a downstream luciferase reporter gene, and the internal deletion of the IRF-E alone eliminated IFN-mediated stimulation of the reporter gene. Cotransfection of an IRF-1 cDNA with the hGR 1A IRF-E luciferase reporter gene into CHOP cells caused a large increase in luciferase activity. The transfection of cells with an IRF-2 cDNA did not result in a statistically significant increase in luciferase expression compared with cells transfected with the empty vector, and although not statistically significant, the cotransfection of both IRF-1 and IRF-2 cDNAs seemed to diminish the effect of IRF-1. Perhaps IRF-2 was unable to compete effectively with IRF-1 because IRF-1 is expressed at higher levels than IRF-2 24 h after transfection. These experiments clearly show that the hGR 1A IRF-E is a bona fide IRF response element.
It was, thus, very surprising that the combined treatment with IFN and DEX does not result in increased cell death in CEM-C7 T lymphoblast cells. To be certain that we could detect even subtle changes in cell viability, we used three different assays. Consistent results are obtained using an LDH release assay, an MTS mitochondrial function assay, and flow cytometric analysis of apoptotic cells. IFN does not increase DEX-mediated T lymphoblast apoptosis.
Two additional studies revealed unexpected, and potentially important, results. First, DEX completely blocks the up-regulation of IRF-1 mRNA and protein in the CEM-C7 T lymphoblast cell line. A similar effect is seen in primary human monocytes, where glucocorticoids inhibit the expression of STAT1 and the expression of a downstream target of STAT1, the IRF-1 gene (30). Thus, the fact that combined DEX/IFN treatment does not cause an increase in T cell death can be explained at least in part by the DEX blockade of IFN-induced increases in IRF-1 mRNA and protein necessary for increased transcription from the hGR 1A IRF-E.
The second surprising result was that IFN alone did not increase the level of GR exon 1A3-containing transcripts in two T lymphoblast cell lines, even though IFN clearly induces IRF-1 expression in these cells, and reporter gene assays show that the hGR 1A IRF-E is completely functional. The most likely explanation for this discrepancy is that an epigenetic event has silenced the endogenous hGR 1A IRF-E. Although the nature of this endogenous gene silencing is unknown, preliminary studies indicate that gene methylation might be a likely candidate. Indeed, there is a CpG sequence in the IRF-E that would be a substrate for the methylase used in these studies, and in vitro methylation of the hGR 1A IRF-E/pXP-1 luciferase reporter construct abolishes its ability to be activated by IFN treatment (Geng, C.-d., and W. V. Vedeckis, data not shown). It is, however, possible that this IRF-E could be functional in cell types other than T cells.
In summary, the hGR gene 1A promoter contains a functional IRF-E. However, it appears that this promoter element is silenced in at least two T lymphoblast cell lines. This may be a mechanism to ensure that hormones with opposing actions (glucocorticoids and IFNs) do not yield the same response (up-regulation of the GR gene) in these cell types. A second mechanism to ensure that these two pathways do not result in the same signaling is the blockade of IFN-induced IRF-1 gene activation by glucocorticoids. Although the IFN- and steroid-activated pathways apparently cannot be used in concert to improve the killing of ALL T cells, it is hoped that future studies will reveal a signaling pathway that can activate hGR gene expression and improve the apoptotic response of these cancer cells to hormone.
References
Kalinyak JE, Dorin RI, Hoffman AR, Perlman AJ 1987 Tissue-specific regulation of glucocorticoid receptor mRNA by dexamethasone. J Biol Chem 262:10441–10444
Barrett TJ, Vedeckis WV 1996 Coordinate regulation of glucocorticoid and c-jun gene expression is cell type-specific and exhibits differential hormonal sensitivity for down- and up-regulation. Biochemistry 35:9746–9753
Breslin MB, Geng C-D, Vedeckis WV 2001 Multiple promoters exist in the human glucocorticoid receptor, one of which is activated by glucocorticoids. Mol Endocrinol 15:1381–1395
Burnstein KL, Bellingham DL, Jewell CM, Powell-Oliver FE, Cidlowski JA 1991 Autoregulation of glucocorticoid receptor gene expression. Steroids 56:52–58
Kofler R 2000 The molecular basis of glucocorticoid-induced apoptosis of lymphoblastic leukemia cells. Histochem Cell Biol 114:1–7
Gaynon PS, Lustig RH 1995 The use of glucocorticoids in acute lymphoblastic leukemia of children. J Pediatr Hematol Oncol 17:1–12
Quddus FF, Leventhal BG, Boyett JM, Pullen DJ, Crist WM, Borowitz MJ 1985 Glucocorticoid receptors in immunological subtypes of childhood acute lymphocytic leukemia cells: a Pediatric Oncology Group Study. Cancer Res 45:6482–6486
Kato GJ, Quddus FF, Shuster JJ, Boyett J, Pullen JD, Borowitz MJ, Whitehead VM, Crist WM, Leventhal BG 1993 High glucocorticoid receptor content of leukemic blasts is a favorable prognostic factor in childhood acute lymphoblastic leukemia. Blood 82:2304–2309
Hala M, Hartmann BL, B?ck G, Geley S, Kofler R 1996 Glucocorticoid-receptor-gene defects and resistance to glucocorticoid-induced apoptosis in human leukemic cell lines. Int J Cancer 68:663–668
Kofler R, Schmidt S, Kofler A, Ausserlechner MJ 1993 Resistance to glucocorticoid-induced apoptosis in lymphoblastic leukemia. J Endocrinol 178:19–27
Pui CH, Costlow ME 1986 Sequential studies of lymphoblast glucocorticoid receptor levels at diagnosis and relapse in childhood leukemia: an update. Leukemia Res 10:227–229
Bloomfield CD, Peterson BA, Zaleskas J, Frizzera G, Smith KA, Hildebrandt L, Gail-Peczalska KJ, Munck A 1980 In vitro glucocorticoid studies for predicting response to glucocorticoid therapy in adults with malignant lymphoma. Lancet 1:1952–1956
Bloomfield CD, Smith KA, Peterson BA, Munck A 1981 Glucocorticoid receptors in adult lymphoblastic leukemia. Cancer Res 41:4857–4860
Str?hle U, Schmidt A, Kelsey G, Stewart AF, Cole TJ, Schmid W, Schütz G 1992 At least three promoters direct expression of the mouse glucocorticoid receptor gene. Proc Natl Acad Sci USA 89:6731–6735
Chen F, Watson CS, Gametchu B 1999 Multiple glucocorticoid receptor transcripts in membrane glucocorticoid receptor-enriched S-49 mouse lymphoma cells. J Cell Biochem 74:418–429
Breslin MB, Vedeckis WV 1998 The human glucocorticoid receptor promoter upstream sequences contain binding sites for the ubiquitous transcription factor Yin Yang 1. J Steroid Biochem Mol Biol 67:369–381
Nunez BS, Vedeckis WV 2002 Characterization of promoter 1B in the human glucocorticoid receptor gene. Mol Cell Endocrinol 189:191–199
Geng C-d, Vedeckis WV 2004 Steroid-responsive sequences in the human glucocorticoid receptor gene 1A promoter. Mol Endocrinol 18:912–924
Stark GR, Kerr IM, Williams BR, Silverman RH, Schreiber RD 1998 How cells respond to interferons. Annu Rev Biochem 67:227–264
Tanaka N, Kawakami T, Taniguchi T 1993 Recognition DNA sequences of interferon regulatory factor-1 (IRF-1) and IRF-2, regulators of cell growth and the interferon system. Mol Cell Biol 13:4531–4538
Watanabe N, Sakakibara J, Hovanessian AG, Taniguchi T, Fujita T 1991 Activation of IFN-? element by IRF-1 requires a posttranslational event in addition to IRF-1 synthesis. Nucleic Acids Res 19:4421–4428
Vaughan PS, van Wijnen AJ, Stein JL, Stein GS 1997 Interferon regulatory factors: growth control and histone gene regulation–it’s not just interferon anymore. J Mol Med 75:348–359
Kalidas M, Kantarjian H, Talpaz M 2001 Chronic myelogenous leukemia. J Am Med Assoc 286:895–898
Taniguchi T, Lamphier MS, Tanaka N 1997 IRF-1: the transcription factor linking the interferon response and oncogenesis. Biochim Biophys Acta 1333:M9–M17
Heffernan M, Dennis JW 1991 Polyoma and hamster papovavirus large T antigen-mediated replication of expression shuttle vectors in Chinese hamster ovary cells. Nucleic Acids Res 19:85–92
Pedersen KB, Vedeckis WV 2003 Quantification and glucocorticoid regulation of glucocorticoid receptor transcripts in two human leukemic cell lines. Biochemistry 42:10978–10990
Ramdas J, Liu W, Harmon JM 1999 Glucocorticoid-induced cell death requires autoinduction of glucocorticoid receptor expression in human leukemic T cells. Cancer Res 59:1378–1385
Pan LY, Guyre PM 1988 Individual and combined tumoricidal effects of dexamethasone and interferons on human leukocyte cell lines. Cancer Res 48:567–571
Uitdehaag BM, Hoekstra K, Koper JW, Polman CH, Dijkstra CD 2001 IFN-?1b augments glucocorticoid-induced suppression of tumor necrosis factor- production by increasing the number of glucocorticoid receptors on a human monocytic cell line. J Interferon Cytokine Res 21:133–135
Hu X, Li W-P, Meng C, Ivashkiv LB 2003 Inhibition of IFN- signaling by glucocorticoids. J Immunol 170:4833–4839(B. Scott Nunez, Chuan-don)