Nitric Oxide Inhibits Glucocorticoid-induced Apoptosis of Thymocytes by Repressing the SRG3 Expression*

Nitric oxide (NO) plays many roles in the immune system. It has been known that NO rescues thymocytes from glucocorticoid (GC)-induced apoptosis. However, the downstream target of NO in the protection from GC-induced thymocyte apoptosis has yet to be identi-fied. We previously reported that GC sensitivity of developing thymocytes is dependent on the expression level of SRG3. In the present report, we found that NO repressed the SRG3 expression in both primary thymocytes and 16610D9 thymoma cells. Specifically, NO down-regulated the transcription of SRG3 via the inac-tivation of the transcription factor Sp1 DNA-binding activity to the SRG3 promoter. In addition, overexpression of SRG3 by a heterologous promoter reduced NO-mediated rescue of thymocytes from GC-induced apoptosis. These observations strongly suggest that NO may be involved in protecting immature thymocytes from GC-induced apoptosis by repressing the SRG3 expression in thymus.

NO 1 is a diffusible, highly reactive molecule with many regulatory roles under physiological and pathological conditions. During the past decade, it has been recognized that NO plays many roles in the immune system as well as in other organ systems. It is involved in the pathogenesis and control of infectious diseases, tumors, autoimmune processes, and chronic degenerative diseases (1). NO and its reaction products can exert opposite effects on apoptosis on the cell systems and conditions involved (2). They induce apoptosis in a variety of cell types, including macrophages, dendritic cells, thymocytes, and neuronal cells (3)(4)(5). In contrast, NO inhibits apoptosis in hepatocytes (6,7), endothelial cells (8), and thymocytes (9). Because of its capacity to induce or inhibit apoptosis in thymocytes, it is thought that NO may play a crucial role as an effector molecule in the selection and development of thymocytes in the thymus (1). Involvement of NO in T cell development has been indicated by several observations. Many immune cells, including epithelial and dendritic cells in the corticomedullary junction and medullary region of the thymus, constitutively express inducible NO synthase (iNOS), which is further up-regulated after contact with self-or allo-antigens or with thymocytes activated by T cell-receptor (TCR) stimulation (1,10). After stimulation by TCR, double-positive (DP) thymocytes are highly sensitive to the killing by NO, whereas singlepositive (SP) thymocytes remain viable upon exposure to NO (11). This result suggests that NO released by iNOS-positive thymic stromal cells is one of the factors mediating deletion of DP thymocytes (1,3). In contrast, NO is implied in T cell survival. Pretreatment with NO protects thymocytes from glucocorticoid (GC)-induced apoptosis (3), in the same way that activation of TCR signaling antagonizes GC-induced apoptosis (12)(13)(14)(15). However, its exact role in transduction of anti-apoptotic signals to thymocytes in thymus is not well documented.
GCs, which are produced in the thymus as well as in the adrenal gland (12), are known to play complex roles during thymocyte development. The levels of endogenous GCs in normal (non-stressed) conditions directly regulate T cell pool size and the CD4 ϩ :CD8 ϩ T cell ratio (16), and they contribute to kill DP thymocytes in times of stress (17,18). It has been postulated that GCs play crucial roles in removing thymocytes expressing TCR without recognizing the self-major histocompatibility complex antigen (14, 15, 19 -22). TCR and glucocorticoid receptor (GR) signals are linked via the Ras/MEK signaling components, and they antagonize each other's apoptosis-inducing process (12)(13)(14)(15). Immature DP thymocytes are extremely sensitive to GC-induced apoptosis, whereas mature SP thymocytes are relatively resistant (23). Therefore, it appears that the immature thymocytes to be positively selected might be protected from apoptotic actions of GCs (14,15). GC-induced apoptosis in thymocytes are inhibited by TCR-or Notch-mediated signals (24,25). Many studies have addressed the GCinduced apoptotic pathways. Meanwhile, it was reported that NO, one of the most versatile players in the immune system, inhibits GC-induced apoptosis of developing thymocytes (3). However, the specific downstream target of NO signaling conferring a GC resistance on thymocytes remains unclear.
We previously isolated SRG3, the murine homolog of yeast SWI3 and human BAF155, as a gene highly expressed in thymus but at a low level in peripheral lymphoid organs (17). It may play essential roles as a core component of SWI/SNF complex by remodeling chromatin structure (26,27). Recently, we also found that the expression level of SRG3 is downregulated right after positive selection and is critical in determining GC sensitivity in T cells (14,15,17,28). Furthermore, TCR/CD3 and Notch1 signaling inhibit GC-induced apoptosis of developing thymocytes by down-regulating SRG3 expression (14,15,28). In response to positively selecting signals, DP thymocytes may down-regulate SRG3 expression and thus become GC-resistant. Otherwise, DP thymocytes without survival signals may still express a high level of SRG3 and thus may be eliminated by GC-induced apoptosis (14,28). These results suggest that SRG3 may play an important role in protecting positively selected immature DP thymocytes from GCmediated apoptosis and the differentiation of DP thymocytes into mature SP thymocytes (14,15,18,28).
In this report, we investigate the possibility that the inhibitory effect of NO on GC-induced apoptosis of thymocytes may be through the down-regulation of SRG3 expression. Here we show that NO repressed SRG3 expression of primary thymocytes and 16610D9 thymoma cells. As known in the previous results, pretreatment of NO inhibits GC-induced apoptosis of thymocytes in wild-type mice. However, thymocytes, in transgenic mice overexpressing SRG3, were not as efficiently rescued by NO. We also found that NO suppressed the SRG3 transcription by reducing the DNA-binding activity of transcription factor Sp1 on the SRG3 promoter. Taken together, these observations strongly suggest that NO may inhibit GCinduced apoptosis of immature thymocytes by down-regulating the SRG3 expression.
Cell Culture, Transfections, and Reporter Assay-The murine thymoma cell line, S49.1, was purchased from the American Type Culture Collection (ATCC, Manassas, VA) and cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (HyClone). The murine DP thymoma cell line, 16610D9, was kindly provided by Dr. C. Murre (University of California at San Diego, La Jolla, CA) and cultured in Opti-Mem (Invitrogen) supplemented with 10% fetal bovine serum. Transfections were performed using Gene Pulse-II RF (Bio-Rad). Luciferase reporter assays were performed with the luciferase assay system (Promega), and ␤-galactosidase activity was used to normalize results for luciferase activity.
Western and Northern Blot Analyses-SDS-PAGE and Western blot analysis were performed as described previously (17). For Northern analysis, total RNAs from 16610D9 were isolated with TRIzol (Invitrogen). Each RNA sample (10 g) was electrophoresed, transferred to membrane, and then hybridized with specific probes. Probes were obtained as described previously (17).
Flow Cytometry-After mechanical disruption of thymic fragments from 3-to 4-week-old male mice, isolated cells were treated with 500 M SNAP and/or 10 Ϫ7 M DEX for 12 h. Single-cell suspension of thymocytes were stained with FITC-conjugated 53-6.72 and PE-conjugated GK1.5. Stained cells were analyzed with the CellQuest TM software using a FACStar plus (BD Biosciences). One million cells of 16610D9 were stained with AnnexinV-FITC (BD Pharmingen) and propidium iodide and analyzed by fluorescence-activated cell sorting.
Analysis of GC Sensitivity of Thymocytes-Thymocytes were harvested and stained with anti-CD4-PE and anti-CD8-FITC monoclonal antibody, and 20,000 cells were analyzed by flow cytometry. A viable cell gate was used to count the number of viable DP cells. The percentage of relative survival was calculated according to the equation previously described (28). Relative survival of DP thymocytes (%) ϭ 100 ϫ A/B, where A is the survival rate of DP thymocytes (%) in medium supplemented with reagents and B is the survival rate of DP thymocytes (%) in medium alone.

The Level of SRG3 Expression Is Down-regulated by NO-It
has been independently reported that NO inhibits GC-induced apoptosis in developing thymocytes (3) and that there is a clear correlation between GC sensitivity and the expression level of SRG3 in T cells (18,28). Based on these results, we hypothesized that NO might inhibit GC-induced apoptosis by repressing the expression of SRG3 in thymocytes. To test this possibility, we analyzed the effects of NO on the expression of SRG3 in murine DP thymoma cell line 16610D9 and normal thymocytes. Cells were treated with SNAP, and SRG3 protein levels were measured by Western blot analysis. The SRG3 protein level was decreased in a time-dependent manner after addition of SNAP to the culture media (Fig. 1A). At 18 h of SNAP treatment, SRG3 protein expression was decreased to ϳ40% of control in 16619D9 cells and 20% in thymocytes. Similar results were obtained by Northern blot analysis showing that the SRG3 mRNA level was decreased in a time-dependent manner after SNAP treatment in 16610D9 cells (Fig. 1B). These results suggest that the SRG3 expression is down-regulated by NO pathway at the transcription level. To examine the effects of NO on the SRG3 promoter activity, we used pSRG3-Luc reporter construct containing a 1.1-kb SRG3 promoter fragment described previously (28). 16610D9 cells were transiently transfected with pSRG3-Luc by electroporation, treated with various concentrations of SNAP, and subjected to reporter activity assay. Treatment of the transfected cells with SNAP resulted in the suppression of the SRG3 promoter activity in a dose-dependent manner (Fig. 1C). To eliminate the possibility that potential secondary metabolites other than NO derived from SNAP might effect the SRG3 expression, we used a SNAP solution that was allowed to fully decompose for 1 week and found that there was no significant effect on the SRG3 protein levels and promoter activity (data not shown). These results suggest that NO reduces the expression of SRG3 in 16610D9 cells and thymocytes by repressing its promoter activity.
NO Inhibits the SRG3 Expression by Redox Regulation-A classicmechanismforNO-mediatedgeneregulationisthecGMPdependent pathway. NO activates the soluble guanylyl cyclase to produce cGMP, which acts as a second messenger (30 -32). However, it has also been found that many other targets (e.g. caspases, zinc finger proteins such as Sp1, EGR-1, mRNA-binding proteins) are affected by redox regulation of NO (31). Therefore, we investigated to determine in which path-ways NO represses the SRG3 expression. For this, we analyzed the effects of SNAP and the cGMP analog, 8-bromo-cGMP, on the activity of the SRG3 promoter. The SRG3 promoter activity was significantly inhibited by SNAP treatment (49% of control), but this inhibitory effect was not notably observed in cells treated with the cGMP (Fig. 2A). Therefore, it appears that NO represses SRG3 promoter activity by a cGMP-independent mechanism. Meanwhile, It is well known that the reducing agent DTT effectively restores the thiol modification induced via redox regulation by NO (6,29). To determine whether the repression of SRG3 by NO is mediated by a redox control, 16610D9 cells were treated with SNAP (750 M) for 5 h after transfection with pSRG3-Luc and then treated with 1 mM DTT. In these cells, the reduced SRG3 promoter activity caused by NO was significantly restored (SNAP, 49% versus SNAP/DTT, 78%) ( Fig. 2A). However, the addition of Me 2 SO as control had no effect on the inhibition of SRG3 promoter activity by NO. Moreover, we analyzed SRG3 protein expression under the same conditions. In agreement with the reporter gene assays, the level of SRG3 protein was not reduced by cGMP treatment, and the repression of SRG3 expression by NO was restored when DTT was added after SNAP treatment (Fig. 2B). These observations suggest that the inhibition of SRG3 promoter activity by NO occurs through the redox regulation of NO.

Repression of SRG3 Expression by NO Occurs through the Inhibition of Sp1
Binding in the SRG3 Promoter-A characteristic feature of many transcription factors is their remarkable redox sensitivity (31). Numerous studies have analyzed the regulatory effect of NO on mammalian transcription factors such as AP-1, NF-B, Sp1, Egr-1, YY1, the vitamin D 3 receptor, and the retinoid X receptor (30,31). Because NO inhibited SRG3 expression by redox regulation (Fig. 2), it was assumed that NO might inhibit SRG3 promoter activity by modulating the activity of the transcription factors that bind to the SRG3 promoter. By sequence analysis of the 1.1-kb SRG3 promoter fragment, we found a potential binding site for YY1 and four putative sites for Sp1 zinc finger transcription factors (Fig. 3A). Zinc finger proteins have been suggested to be primary targets of NO-induced disruption of their functional structures (28). To investigate any specific roles played by these factors during the repression of the SRG3 promoter activity by NO, we performed a systematic deletion analysis of the SRG3 promoter, removing binding sites of these transcription factors as shown in Fig. 3A. D1 contains four Sp1 binding sites but lacks the YY1 binding site, and D2 contains only the two 3Ј Sp1 binding sites. All DNA fragments were linked to the luciferase reporter gene, and their activity was measured after transfection into 16610D9 cells in FIG. 1. SRG3 expression is down-regulated by NO. A, total cell extracts from 16610D9 cells and murine thymocytes treated with 500 M SNAP were analyzed by Western blotting with anti-SRG3 antibody. The same extracts were analyzed with an antiserum against Lck or actin as control. Relative expression of SRG3 protein was assessed by densitometric analysis of the bands and plotted. B, 16610D9 cells were treated with 500 M SNAP, and the expression of the SRG3 transcript was analyzed by Northern blot analysis. The same blot was hybridized with a glyceraldehyde-3-phosphate dehydrogenase probe as control. C, 16610D9 cells were transfected with the pSRG3-Luc reporter construct. Luciferase activity was measured in cells treated with Me 2 SO or SNAP for 13 h. ␤-Galactosidase activity was measured to normalize the transfection efficiency.
the presence or absence of SNAP treatment. D1 and D2 showed reduced promoter activity (40 and 30%, respectively) compared with the 1.1-kb Luc-containing 1.1-kb SRG3 promoter fragment. Treatment of SNAP inhibited the promoter activity of 1.1 kb (47%), D1 (50%), and D2 (52%) at similar levels compared with each control (Fig. 3B). Because the D2 construct showing a basal promoter activity (data not shown), was still downregulated by SNAP treatment, we concluded that Sp1 binding sequences located at Ϫ144 and Ϫ107 regions (III and IV) might be potential candidates as NO response elements.
To assess directly the Sp1 binding sites III and IV as NO response elements, we have constructed plasmids containing site-directed mutations in the Sp1 sites (Fig. 3A). A mutant construct (m IIIϩIV), with mutations in the Sp1 III and IV, showed reduced SRG3 promoter activity (53%) compared with the control construct (1.1 kb), and, importantly, this mutation showed a complete loss of NO responsiveness (m IIIϩIV, Fig.  3C). However, mutations in YY1 (mYY1) or the other Sp1 elements (m IϩII) displayed no significant effect on the SRG3 promoter activity and did not reduce the inhibitory effect of NO on the promoter activity (1.1 kb, 49%; m YY1, 50%; and m IϩII, 53% reduction) (Fig. 3C). Taken together, these results indicate that the Sp1 binding elements, located at Ϫ144 and Ϫ107 on the SRG3 promoter, are responsible for the responsiveness to NO.
NO Disrupts the DNA-binding Activity of Sp1 Transcription Factor on SRG3 Promoter-NO destroys [ZnS] clusters and mediates the release of Zn 2ϩ from zinc-storing protein (32)(33)(34). Subsequently, the DNA-binding activities of the eukaryotic zinc finger transcription factors were found to be inhibited by NO (35), and these NO-mediated inhibitions turned out to be reversible by post-treatment with DTT (32,33). We investigated the regulation of Sp1 binding to SRG3 promoter by NO using gel mobility shift assay. Double-stranded synthetic oligonucleotides encompassing the regions from III (Ϫ144) and IV (Ϫ107) of the SRG3 promoter were radiolabeled and incubated with nuclear extract from 16619D9 cells in the presence or absence of specific anti-Sp antibodies. Two major DNA-protein complexes were detected with both probes (Fig. 4A, left). One hundred-fold excess of unlabeled Sp1 consensus oligonucleo- tides blocked Sp protein binding to the 32 P-labeled probe (Fig.  4A, lane 2). The presence of Sp1 and Sp3 in their corresponding complexes was confirmed by supershift assays with specific antibodies (Fig. 4A, lanes 3 and 4). Similar results were obtained with the putative Sp1 binding sequence located at Ϫ107 of SRG3 promoter (Fig. 4A, right). Nuclear extracts from 16619D9 cells treated for 6 h in the presence of 500 M SNAP exhibited a reduced DNA-binding activity of Sp1 (and Sp3) compared with the untreated control (Fig. 4B, lanes 1 versus 2  and lanes 5 versus 6). The impaired Sp1 (and Sp3) binding activity caused by NO was completely restored upon addition of 1 mM DTT to the SNAP-treated nuclear extracts for 1 h before the DNA-binding reaction (Fig. 4B, lanes 3 and 6). Nitrosylation of Sp1 by the thiol modification may result in reduced binding of the transcription factor to its consensus sequence in the SRG3 promoter. With the TFIID binding consensus oligonucleotides, SNAP had no inhibitory effect on formation of the EMSA complex (Fig. 4B). These results strongly indicate that NO represses SRG3 expression by inhibiting Sp1 (and Sp3) binding to SRG3 promoter.
Inhibitory Effect of NO on GC-induced Apoptosis Is Blocked by SRG3 Overexpression-As described previously, pretreatment of thymocytes with SNAP rendered them resistant to GC-induced apoptosis (3). To test a possibility that the antiapoptotic effect of NO on GC-induced apoptosis was due to the down-regulation of SRG3 expression, we analyzed the effects of NO on GC-induced apoptosis of 16610D9 thymoma cells and thymocytes from transgenic mice (CD2-SRG3) overexpressing SRG3 in T lineage cells by a heterologous promoter (18). The level of SRG3 protein in thymocytes from these mice was about two times higher than that of cells from littermate control mice (28). 16610D9 cells were highly sensitive to DEX treatment (Fig. 5A, closed bar). However, the pretreatment of SNAP followed by DEX treatment resulted in reduced cell death in 16610D9 cells (Fig. 5A, open bar). Similar results were obtained with another thymoma cell line, S49.1 (data not shown). Treatment of SNAP alone had no significant effect on survival of thymocytes from either littermate control or the transgenic mice (Fig. 5B, left), whereas DEX treatment alone induced cell death in both cells (Fig. 5B, middle). Pretreatment of SNAP followed by DEX treatment also resulted in reduced cell death (42% of DEX alone) in control mice, indicating a protective effect of NO from GC-induced cell death in thymocytes (Fig. 5B,  right, open bar). However, such a protective effect of NO on GC-induced apoptosis in thymocytes was not observed in transgenic thymocytes overexpressing SRG3 (Fig. 5B, right lane,  closed bar). These results strongly suggest that NO inhibits GC-induced apoptosis in thymocytes through the down-regulation of SRG3 expression in control cells, but not in transgenic cells overexpressing SRG3 by a heterologous promoter. DISCUSSION We report here that NO protects thymocytes from GC-induced apoptosis by down-regulating the expression of SRG3. We have previously shown that the expression level of SRG3 is critical for determining the sensitivity to GC-induced apoptosis (17,18,28). Overexpression of SRG3 in peripheral T cells, which express a low level of SRG3 and are relatively resistant to GCs, rendered them more sensitive to GCs (18). On the other hand, the expression of antisense SRG cDNA by the lck proximal in transgenic (lck-␣SRG3 ϩ ) mice rendered thymocytes resistant to the GC treatment (28). In time-course and doseresponse studies, the NO-generating compound SNAP treatment inhibited the SRG3 promoter activity as well as reduction of SRG3 expression in thymocytes and 16610D9 thymoma cells. Pretreatment of SNAP protected thymocytes from GC-induced apoptosis, but not those overexpressing SRG3. These results strongly suggest that the protective effect of NO was due to the repression of SRG3 expression (Fig. 5B). Although 8-bromo-cGMP had no effect on the expression and promoter activity of SRG3, addition of a reducing agent, DTT, restored NO-induced SRG3 repression (Fig. 2). These data indicate that repression of SRG3 expression upon NO treatment occurs not through a cGMP-dependent process but through a redox-regulated mechanism. Deletion analysis of the SRG3 promoter revealed that the responsiveness to NO is through the Sp1 binding regions located at Ϫ144 (III) and Ϫ107 (IV) in the SRG3 promoter (Fig.  3). Introduction of site-specific mutations into the Sp1 elements III and IV resulted in complete loss of NO responsiveness. Finally, we confirmed the specific binding of the transcription factor Sp1 (and Sp3) to elements III (Ϫ144) and IV (Ϫ107) in the SRG3 promoter (Fig. 4). Furthermore, we found that pretreatment of SNAP inhibited DNA-binding activity of the transcription factor Sp1 (and Sp3) in the SRG3 promoter, but DTT FIG. 4. NO disrupts the DNA-binding activity of Sp1 transcription factor. A, binding of Sp1 to their putative binding sites, located at Ϫ144 and Ϫ107 on the SRG3 promoter, was performed with the nuclear extracts from 16610D9 cells and the 32 P-labeled III and IV oligomers. Specific binding was assessed by supershift using the anti-Sp1 or Sp3 polyclonal antibody (lanes 3 and 7). B, nuclear extracts from 16610D9 cells treated with 500 M SNAP for 6 h (lanes 2 and 5) were analyzed using EMSA to assess the specific Sp1 DNA-binding activity. These extracts were incubated in the absence or presence of the reducing agent 100 M DTT (lanes 3 and 6). As negative control, the binding reaction was performed with 32 P-labeled TFIID consensus oligomers (lanes 7 and 8).
treatment reversed this inhibition. Thus, we conclude that NO inhibits apoptosis of thymocytes in response to GCs by repressing the SRG3 expression.
The anti-apoptotic mechanisms of NO have been understood via the expression of protective genes such as heat shock protein and Bcl-2 or the inhibition of p53 up-regulation and the apoptotic caspase family proteases by S-nitrosylation of the cysteine thiol (6 -9). Previous work found that both NO and heat treatment were able to protect thymocytes from GC-induced apoptosis and suggested that the prevention of apoptosis by both treatments was through the induction of heat shock protein expression (3). Another possibility is that NO decreased GC binding to its receptor (36). However, we observed that the anti-apoptotic effect of NO on GC-induced apoptosis was greatly reversed in thymocytes from CD2-SRG3 transgenic mice. This observation suggests that the down-regulation of SRG3 is necessary for inhibition of GC-induced apoptosis by NO. On the other hand, heat treatment did not affect SRG3 expression in 16610D9 cells (data not shown), suggesting that at least two distinct mechanisms operate in the protection from apoptosis by NO.
There are evidences that NO has a cytoprotective effect on many apoptotic stimuli (6 -9). It is conceivable that the downregulation of SRG3 by NO may be crucial for its anti-apoptotic effect to apoptotic reagents other than GCs. However, up-regulation of SRG3 in CD2-SRG3 transgenic mice rendered mature T cells more sensitive to apoptosis induced specifically by GC, but not with other pro-apoptotic reagents such as anti-Fas Ab or staurosporine (18). These results strongly suggest that the down-regulation of SRG3 by NO is crucial for the cytoprotective effect in GC-induced apoptosis of T cells but unlikely in apoptosis induced by other stimuli.
In thymus, antigen-presenting cells (APCs), especially dendritic cells express iNOS in basal conditions (3,4). The capacity of dendritic cells to generate NO is enhanced by exposure to antigens, thymocytes activated by TCR stimulation, or cytokines like interferon-␥ and lipopolysaccharide (1,10). It is well known that the interaction with APCs is important for immature thymocytes to develop in the thymus. Thus, it appears that NO produced by thymic APCs may be involved in development of immature thymocytes in the thymus. During maturation in the thymus, for immature DP thymocytes to be positively selected, they should be protected from GC-induced apoptosis (14,15). Recently, we have shown that TCR signaling, mediated through activation of Ras pathway, inhibit GC-induced apoptosis of T cells by down-regulation of SRG3, and thus SRG3 is a key modulator for controlling GC sensitivity in developing thymocytes (14,15,17,28). Here we show that NO reduces the sensitivity of immature thymocytes to GC-induced apoptosis by modulating the expression of a specific gene, SRG3. Our finding that repression of SRG3 determines the inhibition of GCinduced apoptosis by NO fortifies the possibility that SRG3 might serve as a modulator in the apoptosis of developing thymocytes in response to GCs and participate in thymocyte development in the thymus. Therefore, we provide a possible mechanism for understanding how such a pleiotropic and reactive molecule could carry out these protective roles in thymic development.
FIG. 5. Down-regulation of SRG3 is necessary for inhibition of GC-induced apoptosis by NO. A, 16610D9 cells were incubated with 500 M SNAP for 30 min prior to 10 Ϫ7 M DEX treatment (SNAPϩDEX), SNAP alone (SNAP), or DEX alone (DEX). The cells were analyzed for apoptosis using flow cytometry at the indicated time points. To determine the amount of apoptosis, cells were analyzed for AnnexinV binding. B, thymocytes prepared from CD2-SRG3 ϩ transgenic mice (C57BL/6, 3-4 weeks old) and littermate mice were incubated in medium with 500 M SNAP alone, 10 Ϫ7 M DEX alone, or SNAP for 30 min prior to DEX (SNAPϩDEX) treatment for 12 h, and analyzed by flow cytometry for viability and CD4 and CD8 expression. Results are the relative survival rate of DP thymocytes shown as an average of three independent experiments.