Microarray Analysis Uncovers the Induction of the Proapoptotic BH3-only Protein Bim in Multiple Models of Glucocorticoid-induced Apoptosis*

Despite being one of the earliest recognized and most clinically relevant forms of apoptosis, little is known about the transcriptional events that mediate glucocorticoid-induced apoptosis. Therefore, we used oligonucleotide microarrays to identify the pattern of dexamethasone-induced changes in gene expression in two well characterized models of glucocorticoid-induced apoptosis, the murine lymphoma cell lines S49.A2 and WEHI7.2. Dexamethasone treatment induced a diverse set of gene changes that evolved over a 24-h period preceding the onset of cell death. These include previously reported changes in the expression of genes regulating prosurvival signals mediated by c-Myc and NFκB. Unexpectedly, we discovered that glucocorticoid treatment increases expression of the gene encoding Bim, a BH3-only member of the Bcl-2 family that is capable of directly activating the apoptotic cascade. Induction of Bim was confirmed by immunoblotting not only in S49.A2 and WEHI7.2 cells but also in the human leukemia cell line CEM-C7 and in primary murine thymocytes. All three prototypical isoforms of Bim (BimEL, BimL, and BimS) were induced by dexamethasone. Because elevated expression of Bim initiates the execution phase of cell death, this report that Bim is induced by dexamethasone provides novel insight into the mechanism through which glucocorticoid-mediated changes in gene expression induce apoptosis in lymphoid cells.

The ability of glucocorticoid hormones to induce apoptosis in leukemia and lymphoma cells has been utilized by physicians for nearly a half-century (1,2). Today glucocorticoids are common components in many chemotherapeutic protocols for lymphoid malignancies, including multiple myeloma, acute lymphoblastic leukemia, chronic lymphocytic leukemia, and non-Hodgkin's lymphoma (3)(4)(5)(6). In fact, glucocorticoid therapy is essential for the successful treatment of childhood acute lymphoblastic leukemia (5).
Apoptosis is an orderly process of cell death typified by protease and endonuclease activation and is characterized morphologically by condensed chromatin, a reduction of cell volume, and plasma membrane blebs (7). Whether or not a cell will initiate the apoptotic cascade depends upon the relative expression of the pro-and antiapoptotic members of the Bcl-2 family (8). Members of the Bcl-2 family of proteins all are related by the conservation of at least one Bcl-2 homology (BH) 1 domain (9). Bcl-2 and Bcl-xL contain four BH domains and are capable of inhibiting apoptosis. Family members, such as Bax and Bak, share multiple BH domains with the antiapoptotic Bcl-2 proteins; however, these proteins are proapoptotic and are required for nearly all forms of apoptosis (10). A host of proapoptotic proteins containing only the BH3 domain also exists within the Bcl-2 family. These include Bim, Bbc3/PUMA, Bad, and Bid (9,11). BH3-only proteins are required for the initiation of apoptosis by multiple stimuli and have been identified in species as primitive as Caenorhabditis elegans (12). They initiate apoptosis either by inhibiting antiapoptotic Bcl-2 proteins (12,13) or by activating proapoptotic proteins such as Bax and Bak (14).
Many death-inducing signals activate constitutively expressed apoptotic machinery capable of triggering apoptosis without altering gene expression. Unlike these signals, glucocorticoid-induced apoptosis requires the activation of an incompletely defined program of transcriptional regulation initiated by the glucocorticoid receptor (15). The glucocorticoid receptor belongs to the nuclear steroid hormone receptor family of zinc finger transcription factors. Upon ligand binding, the receptor dissociates from its cytosolic chaperone complex, homodimerizes, and translocates into the nucleus. Once in the nucleus, the glucocorticoid receptor can stimulate or repress gene expression either directly by binding glucocorticoid response elements on regulatory regions of target genes or indirectly through protein-protein interactions with other transcription factors, including NFB and AP-1 (16 -18).
Multiple lines of evidence demonstrate that glucocorticoid receptor-mediated transcriptional regulation is required for apoptosis. Glucocorticoid-induced apoptosis can be blocked by cycloheximide and actinomycin D, demonstrating a dependence on de novo protein and RNA synthesis (19 -21). Resistance to glucocorticoid therapy in vivo often involves the loss of a functional glucocorticoid receptor, and glucocorticoid receptor mutants lacking the domains required for either transactivation or transrepression are unable to induce apoptosis (22)(23)(24)(25). Fur-thermore, Reichardt et al. (26) have demonstrated that binding of the glucocorticoid receptor to DNA is required for glucocorticoid induced-apoptosis in thymocytes. This involves the use of a knock-in mouse model expressing a glucocorticoid receptor (A458T) that is unable to dimerize and cooperatively bind palindromic glucocorticoid response elements.
Because glucocorticoid-induced apoptosis is mediated by the regulation of gene expression, much attention has been centered on the ability of glucocorticoids to repress the activity of prosurvival transcription factors such as c-Myc, NFB, and AP-1 (27)(28)(29)(30). These studies have elucidated transcriptional targets of the glucocorticoid receptor that regulate signal transduction pathways capable of altering the balance between life and death. Although these data clearly demonstrate that repressing prosurvival signals shifts the balance in favor of death, studies to date have not identified transcriptional targets capable of directly initiating the apoptotic cascade.
To understand better the transcriptional regulatory changes that promote apoptosis, we have used oligonucleotide microarrays to identify a set of glucocorticoid-regulated genes common to two glucocorticoid-sensitive models of T-cell lymphoma, the S49.A2 and WEHI7.2 murine cell lines. Using this paradigm, we have discovered that bim, a proapoptotic BH3-only member of the Bcl-2 family, is induced following dexamethasone treatment. This is, to our knowledge, the first report that a proapoptotic protein capable of directly initiating the apoptotic cascade is induced following dexamethasone treatment.

EXPERIMENTAL PROCEDURES
Cell Culture-WEHI7.2 and S49.A2 murine T-cell lymphoma lines were gifts of Drs. Diane Dowd and Roger Miesfeld, respectively. The CEM-C7 human acute T-cell leukemia line was a gift of Dr. Brad Thompson. WEHI7.2 and S49.A2 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 2 mM L-glutamine, 10% bovine calf serum (HyClone), 12.5 units/ml penicillin, and 12.5 g/ml streptomycin. CEM-C7 cells were cultured in RPMI 1640 medium supplemented with 2 mM L-glutamine, 10% fetal bovine serum (Atlanta Biologicals), 12.5 units/ml penicillin, and 12.5 g/ml streptomycin. Unless otherwise noted, all cell culture reagents were purchased from Invitrogen. All experimental cultures were started at a density of 1-2 ϫ 10 5 cells/ml and grown in a humidified 7% CO 2 incubator at 37°C. For kinetic analysis, all samples were treated simultaneously and then harvested individually at the appropriate time after treatment.
Expression Analysis-Gene expression analysis was performed essentially as described in the Affymetrix GeneChip Expression Analysis Technical Manual (31). Total RNA was harvested from both dexamethasone-treated (1 M) and vehicle control populations at each time point by TRIzol (Invitrogen) extraction. TRIzol cell lysates were separated into aqueous and organic phases by the addition of chloroform to a final concentration of 20% (v/v). The aqueous phase was purified and concentrated using a Qiagen RNeasy minicolumn. DNA complementary to total RNA samples was reverse transcribed using Superscript reverse transcriptase (Invitrogen) and a T7-(dT) 24 primer (Operon). This cDNA was used as a template for the synthesis of biotinylated cRNA using the T7 Megascript kit from Ambion. Biotinylated cRNA probes were fragmented and hybridized to MG-U74A(v2) GeneChips (Affymetrix) using an Affymetrix GeneChip fluidics station 400 and standard Affymetrix protocols. Fluorescence intensities were captured with a GeneArray Scanner (Hewlett-Packard).
GeneChip image files were processed using the Microarray Analysis suite, version 5.0 (Affymetrix). Probe cells displaying irregular fluorescence intensity over the area of the cell were excluded from subsequent analyses. To facilitate comparison between samples and experiments, the trimmed mean signal of each array was scaled to a target intensity of 1500. A comparative analysis between treatment and control samples for each time point was performed with the Affymetrix statistical algorithm using default parameters. To compensate for gene expression changes occurring in the control cultures over time, each treated sample was compared with a control sample that was split and harvested in parallel with the treated population. Metric files from expression and comparison analyses were exported to Microsoft Access XP for further filtering and analysis. In this work, genes termed "significantly changed" were those that possessed a reliably detectable signal (abso-lute call "absent" and signal Ն 500 in treatment or control samples for inductions or repressions, respectively), determined by the statistical algorithm to be changed 2-fold or greater (change call "no change" and signal Ͼ 500 in treatment or control samples for inductions or repressions, respectively). To increase stringency, genes meeting the above criteria were filtered further to include only those that also were changed in the same direction (change call "no change") in at least one adjacent time point regardless of magnitude.
Hierarchical Clustering Analysis-Genes considered significantly changed (see "Expression Analysis" above) were grouped according to the similarity of their expression changes over time in both cell lines. An uncentered correlation similarity matrix and complete linkage analysis were selected to cluster signal log 2 ratios using GeneCluster v.2.11 (Stanford University). Data were visualized using TreeView, version 1.5 (Stanford University).
Immunoblot Analysis-Cell cultures were harvested by centrifugation, washed twice in phosphate-buffered saline, and lysed in radioimmune precipitation assay buffer (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.1% sodium dodecyl sulfate, and 0.5% sodium deoxycholate), to which complete protease inhibitor tablets were added according to the manufacturer's instructions (Roche Applied Science). Protein concentration in cell lysates was quantified by the Bradford method. The volume and the protein concentration in each of the samples were normalized by the addition of radioimmune precipitation assay buffer prior to loading. Lysates were mixed with an equal volume of 2ϫ sample loading buffer (100 mM Tris-HCl, pH 6.8, 4% sodium dodecyl sulfate, 20% glycerol, and 0.2% bromphenol blue) and then boiled for 10 min. Proteins were resolved on 12.5% Tris-HCl SDS-PAGE with a 5% stacking gel and then immobilized by electrotransfer onto polyvinylidene fluoride membranes. Nonspecific protein binding was blocked prior to incubation using primary antibodies in Tris-buffered saline containing 0.1% Tween 20 and 5% nonfat dry milk. The anti-Bim polyclonal antibody was purchased from Sigma; the anti-␤-actin and anti-Bcl-2 antibodies were obtained from Pharmingen. Horseradish peroxidase-conjugated goat anti-rabbit IgG and enhanced chemiluminescence substrate (Amersham Biosciences) were used for antibody detection. Results are representative of at least three independent experiments.
Northern Blot Analysis-Total RNA was extracted from cultured cells using the TRIzol reagent (Invitrogen) followed by purification through an RNeasy minicolumn (Qiagen). Total RNA (10 g) was separated in a 1.0% agarose-formaldehyde gel and transferred to a Gene-Screen Plus membrane (PerkinElmer Life Sciences) in 10ϫ SSC (0.15 M NaCl, 0.015 M Na citrate) by capillary blotting. The RNA was fixed to the membrane by cross-linking with ultraviolet light (245 nm, 30 s, 1200 J) using a Strata-linker UV oven (Stratagene). Membranes were hybridized with a 32 P-labeled bim probe prepared from full-length bim EL cDNA (a gift from Dr. Andreas Strasser) in QuikHyb (Stratagene) at 65°C. Membranes subsequently were washed at 65°C twice for 15 min each in 2ϫ SSC, once for 30 min in 2ϫ SSC, 0.1% SDS, and once for 10 min in 0.1ϫ SSC, 0.1% SDS.
Thymus Isolation-C57BL/6J mice (Jackson Laboratory) were sacrificed by CO 2 asphyxiation between 6 and 12 weeks of age. Thymi were removed, rinsed in ice-cold growth medium (Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 12.5 units/ml penicillin, and 12.5 g/ml streptomycin), and then dispersed through a steel wire mesh into 5 ml of fresh, cold growth medium/thymus. The suspension of thymocytes was filtered through a nylon mesh to remove connective tissue. For each experiment, thymocytes were pooled from at least three mice, diluted to 2-3 ϫ 10 6 cells/ml in warm growth medium, treated with dexamethasone or vehicle control, and grown in a humidified 7% CO 2 atmosphere at 37°C.
Apoptosis Assays-Apoptosis was measured by quantification of cellular DNA content by flow cytometry. Culture samples containing 1.5 ϫ 10 6 cells were collected by centrifugation at 200 ϫ g for 5 min, washed once with phosphate-buffered saline, suspended in 500 l of ice-cold methanol, and then incubated for a minimum of 5 min at Ϫ20°C. Methanol-fixed samples then were collected by centrifugation at 350 ϫ g for 2 min, washed once with phosphate-buffered saline, and then incubated at room temperature for 1 h in propidium iodide staining solution containing 50 g/ml propidium iodide, 0.1% Nonidet P-40, 20 g/ml RNase A, and 0.1% sodium azide in phosphate-buffered saline. Propidium iodide fluorescence was measured using a FACScan XL flow cytometer (Coulter). Non-aggregated whole cells having a DNA content less than that of the G 1 population were scored as apoptotic. Data analysis was performed using WinList 3D, version 4.0 (Verity Software House).

Microarray Expression Analysis of Glucocorticoid-regulated
Genes Identifies bim-Oligonucleotide microarray analysis was employed to identify glucocorticoid-regulated genes in the S49.A2 and WEHI7.2 murine T-cell lymphoma cell lines. Both cell lines are well studied models of glucocorticoid-induced apoptosis, a phenomenon that is observed 24 -36 h following treatment with 1 M dexamethasone, a synthetic glucocorticoid (Fig. 1). The WEHI7.2 cell line is more sensitive to steroidinduced apoptosis, most likely because it expresses less of the antiapoptotic Bcl-2 protein than the S49.A2 cells (Fig. 1B,  inset). To determine glucocorticoid-regulated changes in gene expression that occur prior to the onset of apoptosis, RNA samples were extracted from cells treated with dexamethasone or ethanol vehicle 6, 12, 18, and 24 h after treatment. Expression profiles of both dexamethasone-treated and control populations at each time point were determined using Affymetrix MG-U74A(v2) GeneChips. A comparison of gene expression between treatment and control populations at each time point was performed using the Affymetrix Microarray Analysis suite, version 5.0. From over 10,000 genes or expressed sequenced tags represented on the array, 284 changed significantly after dexamethasone treatment in both cell lines ( Table 1 in the supplemental material; see "Experimental Procedures" for selection criteria). Of the 284 genes within this set, 70% were induced, and 30% were repressed. Protein phosphatases, major histocompatibility complex antigens, and genes involved in free radical metabolism were the categories of dexamethasone-regulated genes for which the quantity frequently was altered (data not shown).
Because both cell lines undergo apoptosis in response to dexamethasone, genes that are mediators of apoptosis are likely to be changed in the same direction in both cell lines. To identify this subset of genes, we subjected our data set containing 284 glucocorticoid-regulated genes to hierarchical clustering analysis (Fig. 2A). Clustering the data set in this manner allowed the identification of three groups of genes for which expression changed similarly in both cell lines. A predominant group of genes repressed in both cell lines contained familiar targets of glucocorticoid-mediated transrepression; these targets included phosphofructokinase, a class I major histocompatibility complex antigen, and c-myc (Fig.  2, Group I). Genes for which expression was induced by glucocorticoids in both cell lines were segregated into two distinct clusters according to the kinetics of their induction (Fig. 2B). Two inhibitors of prosurvival signals were induced with delayed kinetics; these were the NFB inhibitor, IB-␣, and the regulatory p85 subunit of phosphatidylinositol 3-kinase (Fig. 2, Group II). Group III, by comparison, contained genes for which expression was induced rapidly, reaching a plateau between 12 and 18 h. On average, expression of genes within this group was increased by ϳ2-fold as early as 6 h following dexamethasone treatment. This group contained a probe for the expressed sequence tag AA796690, a sequence that is homologous to the long splice variant of the Bcl-2 Interacting Mediator of Cell Death (bim) (Fig. 2, Group III). Because bim is capable of initiating apoptosis at the level of the mitochondrion, the induction of bim is the glucocorticoidmediated transcriptional event most directly related to apoptosis and therefore is the focus of this work.
bim Is Induced by Dexamethasone-Based upon the hybridization signals for the AA796690 expressed sequence tag, the expression of bim was induced as early as 6 h following dexamethasone treatment and reached a maximum induction of greater than 2-fold after 24 h in both S49.A2 and WEHI7.2 cells (Figs. 2 and 3A). Northern blotting confirmed the elevated level of bim in WEHI7.2 cells (Fig. 3B). By immunoblot analysis, we observed the induction of all three prototypical isoforms, Bim EL , Bim L , and Bim S , in both S49A2 and WEHI7.2 cells (Fig.  3, C and D, respectively). An increase in protein expression was apparent within 24 h of dexamethasone treatment in both cell lines. Bim not only was induced at the time dexamethasonetreated cells began to undergo apoptosis but also was induced only by doses of dexamethasone sufficient to induce apoptosis (Fig. 4).
The Induction of bim Requires de Novo Transcription and Translation-Because de novo RNA and protein synthesis is required for dexamethasone-induced apoptosis in thymocytes, we tested whether the induction of bim expression similarly required de novo transcription and translation. WEHI7.2 cells were treated with dexamethasone in the presence and absence of the transcriptional inhibitor actinomycin D or the protein synthesis inhibitor cycloheximide. In the presence of actinomycin D, the abundance of bim falls to barely detectable levels.
Also, cycloheximide appears to block dexamethasone-induced bim expression (Fig. 5A). These results suggest that de novo transcription and translation are required for the induction of bim expression by dexamethasone. Furthermore, in the presence of actinomycin D, dexamethasone did not induce Bim protein expression after 36 h (Fig. 5B). These data indicate that the elevation of Bim is due to either an increase in bim transcription or the transcriptional induction of a protein that stabilizes the bim transcript.

Bcl-2 Overexpression Does Not Prevent the Induction of Bim and Reveals the Induction of Bim S -Bcl-2 overexpression in-
hibits glucocorticoid-induced apoptosis in the WEHI7.2 cell line (32). To verify that the induction of Bim is an event that precedes apoptosis, we tested whether Bcl-2 could prevent or delay its induction following dexamethasone treatment. WEHI7.2 cells stably overexpressing Bcl-2 were treated with dexamethasone and harvested for immunoblot analysis from 6 to 24 h following steroid treatment. As shown in Fig. 5C, Bcl-2 overexpression neither prevents nor alters the kinetics of Bim induction by dexamethasone. This demonstrates that the induction of Bim occurs prior to the onset apoptosis. These data are consistent with the evidence that Bim promotes apoptosis at a point in the apoptotic cascade that is inhibited by Bcl-2 (33). Furthermore, the protection offered by Bcl-2 overexpression enhanced the induction of Bim S , the most potent Bim isoform, which is localized exclusively at the mitochondria and is capable of activating Bax (14).
Dexamethasone Induces Bim in Multiple Models of Glucocorticoid-induced Apoptosis-We next examined whether the induction of Bim by dexamethasone is limited to murine T-cell lymphoma lines or is a general characteristic of dexamethasone-induced apoptosis. Immunoblot analysis revealed that three isoforms, Bim EL , Bim L , and Bim S , were induced in the human acute T-cell leukemia line CEM-C7 as early as 24 h after dexamethasone treatment (Fig. 6A). Thus the induction of

FIG. 2. Microarray expression analysis of two models of glucocorticoid-induced apoptosis reveals common glucocorticoidregulated gene changes.
A, hierarchical clustering of 284 glucocorticoid-regulated genes identified three groups of genes expressed similarly in both cell lines. Horizontal distance between branches of genes in the dendrogram indicates the relatedness of gene expression with shorter distances indicating higher similarity. Differences in expression between the dexamethasone-treated and time-matched control populations are designated by a matrix of colored rectangles. A red hue indicates an induced gene, whereas green identifies a gene for which expression was reduced following dexamethasone treatment. The magnitude of the expression difference is indicated by the color saturation with bold colors representing large differences, dark colors representing small differences, and black indicating no change. Genes annotated with bold type are discussed in the text. B, the average change in gene expression over time for Groups I, II, and III. Ordinate values are given as the signal log 2 ratio of fluorescence intensity between treatment and control populations. Thus a ratio of X represents a 2 X -fold change in expression following dexamethasone treatment.
Bim expression by dexamethasone may be relevant to therapy of human leukemia with glucocorticoids. Furthermore, we investigated whether the induction of Bim by dexamethasone is limited to transformed cells by examining its expression in primary murine thymocytes. Mouse thymocytes are acutely sensitive to dexamethasone and are a classic model system for studying the mechanism of corticosteroid-induced apoptosis (34). Freshly isolated murine thymocytes were treated with dexamethasone in vitro. Dexamethasone induced apoptosis in this model system as early as 2 h following glucocorticoid treatment (Fig. 6B). Immunoblot analysis revealed that Bim, mainly the Bim L isoform, was induced after 2 h of dexamethasone treatment (Fig. 6C). Furthermore, as little as 1 nM dexamethasone was sufficient to induce Bim within 8 h (Fig. 6D). The expression of Bim increased with steroid concentrations up to 100 nM. Higher doses of dexamethasone resulted in a reduction of Bim, an effect that is also apparent in the kinetic analysis (Fig. 6C) and likely the result of the swift induction of apoptosis and subsequent proteolysis in glucocorticoid-treated thymocytes.

DISCUSSION
Using Affymetrix oligonucleotide microarrays, we have identified a set of glucocorticoid-regulated genes in the S49.A2 and WEHI7.2 T-cell lymphoma lines. As expected in light of work published earlier (27)(28)(29)(30), we identified several genes that are glucocorticoid-regulated and capable of tipping the balance between life and death toward death. For example, glucocorticoids suppress survival signals through either the direct repression of a survival gene, c-myc, or the induction of inhibitors of survival factors, including IB-␣ and the regulatory subunit of the phosphatidylinositol 3-kinase. In addition, dexametha-sone induces the expression of both the type I and type II inositol 1,4,5-trisphosphate receptors, consistent with evidence for a role of intracellular calcium release in apoptosis (20,35,36). However, because of its documented ability to initiate apoptosis, the discovery that Bim is induced by dexamethasone is the principal achievement of the microarray experiments described in this report.
Dexamethasone induced the expression of the proapoptotic BH3-only protein Bim not only in the S49.A2 and WEHI7.2 murine T-cell lymphoma cell lines but also in the human T-cell leukemia cell line CEM-C7 and primary thymocytes. Multiple laboratories have demonstrated that elevated expression of Bim, particularly the short isoform, is sufficient to induce apoptosis in a variety of cell lines (13,33). Enforced expression of Bim leads to apoptosis in hematopoietic cells, even in the presence of cytokines, demonstrating that increased expression of Bim is sufficient to induce apoptosis in the absence of other apoptotic signals (37,38). Although the activity of Bim EL and Bim L can be regulated by interaction with the microtubuleassociated dynein motor complex, no such interaction occurs with Bim S (39). In addition, the BH3 domain of Bim as well as full-length Bim S bind Bax on the mitochondria, induce its oligomerization, and promote the release of cytochrome c (14,40). Thus the elevated expression of Bim S following dexamethasone treatment is a death-promoting stimulus capable of directly initiating the apoptotic cascade.
Strasser and colleagues (41) have demonstrated that immature T-cells from bim Ϫ/Ϫ mice, although not completely resistant to glucocorticoid-induced apoptosis, die with delayed kinetics following dexamethasone treatment. Because thymocytes from Bax Ϫ/Ϫ /Bak Ϫ/Ϫ mice are completely resistant to dexamethasone-induced apoptosis (10), the incomplete resistance of bim-deficient thymocytes to dexamethasone indicates that another BH3-only protein is compensating for the absence of bim. Among the known BH3-only proteins, Bim is the only one we observed as significantly induced during glucocorticoid-induced apoptosis. It has been suggested that Bbc3/PUMA may promote glucocorticoid-induced apoptosis in thymocytes because its transcript is induced following treatment with dexamethasone (42). However, we have not observed a glucocorticoidmediated increase in Bbc3/PUMA protein expression in either WEHI7.2 or primary mouse thymocytes (data not shown).
Although the induction of bim by glucocorticoids has not been recognized previously, retrospective analysis of an earlier microarray report reveals evidence of bim induction by dexamethasone. Tonko et al. (43) identified eight genes or expressed sequence tags coordinately regulated in both proliferating and G 1 /G 0 arrested human CEM-C7 leukemia cells after glucocorticoid treatment. We find that one of the expressed sequence tags, AA682502, has high homology with the murine bim mRNA. Indeed, in the present report, immunoblot analysis confirmed that Bim is induced in CEM-C7 cells treated with dexamethasone.
bim is likely a participant in glucocorticoid-mediated apoptosis; however, the mechanism of its regulation by glucocorticoids is not clear. Because the human bim promoter does not contain a glucocorticoid response element and cycloheximide blocks the induction of bim, this response may not be mediated directly by the glucocorticoid receptor. Recently, bim has been identified as a target of the forkhead family of transcription factors (44). Forkhead proteins like FKHRL1 are transcription factors in which activity is repressed when phosphorylated by prosurvival kinases such as phosphatidylinositol 3-kinase and protein kinase B/Akt. Because we observed an induction of the regulatory phosphatidylinositol 3-kinase p85 subunit, the induction of bim may result from decreased survival kinase ac-FIG. 6. Dexamethasone induced the expression of Bim both in human CEM-C7 cells and in primary mouse thymocytes. A, human CEM-C7 cells were cultured in the presence of 1 M dexamethasone (ϩ) for the times indicated. Protein lysates were prepared and assayed for Bim expression by immunoblot. Dex, dexamethasone. B, thymocytes are acutely sensitive to dexamethasone. Freshly isolated thymocytes were cultured in 1 M dexamethasone for the times indicated. Apoptosis was quantified by cells containing sub-G 1 DNA. Values represent the means Ϯ S.E. of three experiments, each using pools of thymocytes from at least three mice. C, expression of Bim protein is induced in primary thymocytes treated with 1 M dexamethasone (ϩ) for the times indicated. D, thymocytes were treated with 1, 10, 100, or 1000 nM dexamethasone or ethanol control for 8 h. Protein lysates were analyzed for Bim expression by immunoblot. tivity that leads to the activation of FKHRL1 (45,46).
Because glucocorticoid-induced apoptosis requires receptormediated regulation of gene transcription, the concept that glucocorticoid-induced cell death is mediated through the induction of a "death gene" emerged over 3 decades ago (47). Although the efforts of many laboratories have identified glucocorticoid-induced genes capable of reducing prosurvival signals, a glucocorticoid-induced death gene has remained elusive. Now we report the induction of the proapoptotic protein Bim by dexamethasone. Expression of Bim, particularly Bim S , is capable of inhibiting the prosurvival activity of Bcl-2 and directly inducing the oligomerization of Bax. As bim-deficient thymocytes are not completely resistant to dexamethasone, further experiments are necessary to uncover which BH3-only protein works in concert with Bim to promote glucocorticoid-induced apoptosis. Only when all compensatory proteins are removed to generate a resistant phenotype can the role of individual BH3only proteins be investigated. Nevertheless, the identification of bim as a glucocorticoid-induced death gene provides the foundation for developing a complete model of the glucocorticoid-induced apoptotic pathway. Understanding the mechanism by which glucocorticoids induce the expression of bim will reveal novel points along the death-promoting cascade at which apoptosis may be therapeutically accelerated.