Role for Ets-2Thr-72 Transcription Factor in Stage-specific Thymocyte Development and Survival

Background: The transcription factor Ets-2 is highly expressed in mouse thymocytes. Results: Expression of a phosphomutant Ets-2 (T72A) in mouse thymus severely alters T-cell development. Conclusion: Transgenic thymocytes exhibited increased apoptosis and a thymocyte developmental block. Significance: Ets-2 is a critical Ras signaling target in thymocytes. Interference of Ras signaling deregulates thymocyte development in mouse models. However, the role of Ets-2, a transcription factor that is phosphorylated on a critical threonine residue (Thr-72) by the Ras/MAPK pathway in thymocyte development, has not been defined. Transgenic mice overexpressing a phosphomutant Ets-2 (T72A) in the thymus displayed reduced thymus size associated with a 60–80% reduction in thymocyte populations. The transgenic mice exhibited a 20-fold increase in a c-Kit+ CD4+ CD8+ CD3− population and a 5-fold increase in a unique CD5low population associated with a partial developmental block at the DN2-DN3 stage of thymocytes. Transgenic thymocytes exhibited increased apoptosis, and overexpression of Bcl-2 rescued the hypocellularity and associated thymocyte developmental block in double transgenic mice. The observed defects in these mice are not dependent on Ets-1 expression. These studies implicate for the first time a stage-specific Ets-1-independent regulatory role for Ets-2 in early thymocyte development and survival.

Several studies have implicated the Ras signaling pathway in T-cell signaling and development (1). A transgenic mouse model that expressed a dominant-negative Ras (H-rasN17) in thymocytes demonstrated that disruption of TCR 2 -mediated signaling impaired thymocyte-positive selection (2). Overexpression of a dominant-negative or constitutively active Raf also interrupted TCR signaling and thymocyte-positive selection. Subsequent mouse models and experiments have also demonstrated a role for MEK (3,4) and ERK (5)(6)(7)(8)(9) in thymocyte-positive and -negative selection and maturation. These models have provided invaluable information on the importance of CD3-mediated activation of the Ras signaling pathway in thymocyte development. However, despite these numerous mouse models and in vitro experiments, there is limited infor-mation about the downstream nuclear targets of the Ras signaling pathway during thymocyte development.
Previously, Ras has been shown to activate the Ets family transcription factor Ets-2 in vitro (10). Oncogenic (constitutively active) Ras in conjunction with Ets-2 super-activates promoters containing Ets-binding sites (10,11). Ras pathwaymediated activation of Ets-2 occurs through a conserved MAPK-docking motif, PLLTP, contained within amino acids 69 -73 of the Ets-2 protein. Destruction of this binding site by mutating threonine 72 (Thr-72) to an alanine (Ets-2 T72A ) results in the creation of a construct that impairs Ras-dependent activation of Ets-2 (10 -13). In addition, loss of the MAPKdocking site also abrogates constitutively activated MEK1 signaling (14), and the threonine 72 residue has been shown to be phosphorylated directly by ERK2 (14). These data implicate Ets-2, via threonine 72 phosphorylation, as a direct downstream nuclear target of Ras-mediated cellular signaling. Ets-2 is expressed at high levels in the mouse thymus and is observed as early as the CD4 Ϫ CD8 Ϫ CD3 Ϫ triple-negative stage and is expressed through the single-positive (SP) mature T-cell stage (15). Stimulation of the Jurkat T-cell line through the T-cell receptor with anti-CD3 antibody has been shown to phosphorylate Ets-2 in vitro (16). These data suggest that signaling through the TCR, which is in part mediated by Ras signaling, may utilize Ets-2 as a nuclear target.
Although Ras-mediated signaling events are critical for normal thymocyte development, the role of its downstream target Ets-2 in this process has not been examined. To determine the role of Ras/MAPK-dependent phosphorylation of Ets-2 in in vivo thymocyte development and function, we generated transgenic mice overexpressing a phosphomutant Ets-2, containing threonine 72 replaced by alanine. Here, we show that these transgenic mice exhibit a severe block in early thymocyte development associated with Bcl-2-dependent survival defects. Furthermore, these defects are independent of Ets-1 transcription factor expression, indicating a critical role for Ets-2 in specific stages of thymocyte development and survival.

EXPERIMENTAL PROCEDURES
Animals-Transgenic animals described in this report were generated at the Ohio State University Comprehensive Cancer Center Transgenic and Embryonic Stem Cell Core Lab at Children's Research Institute at the Children's Hospital, Columbus, OH. FVB/N mice were obtained from the Harlan Sprague-Dawley Inc. (Indianapolis, IN). C57BL/6-TgN (BCL2) 36WEHI transgenic mice were purchased from The Jackson Laboratory (Bar Harbor, ME) (17). All experiments were performed according to the guidelines and protocol approved by the Institutional Animal Care and Use Committee.
Generation of Transgenic Mice-The T-cell-specific pTEX-Ets-2 T72A transgenic vector was constructed by linearizing the pTEX transgenic vector with XhoI (18 -20). The human ETS2 T72A cDNA (that codes for mutant Ets-2 protein containing alanine instead of threonine at amino acid position 72) was cut from the pCG-ETS2 T72A expression vector with BamHI and BglII (10). The two constructs were then ligated with T4 DNA ligase (Invitrogen) to create the transgenic vector pTEX-Ets-2 T72A . The mutation in the transgenic construct was confirmed by restriction analysis with HpaI, which is created by the T72A mutation. pTEX-Ets2 T72A transgenic mice were created by microinjecting a gel-purified 9-kb NotI fragment containing the CD2 promoter/enhancer elements along with the ETS2 T72A mutation, into fertilized eggs, derived from FVB/N mice.
Southern Blot Analysis-Transgenic mice were identified by Southern blot analysis of genomic tail DNA. Tails from transgenic or nontransgenic mice were digested overnight with lysis buffer containing 50 mM Tris, pH 8, 100 mM EDTA, pH 8, 0.5% SDS, and 350 g/ml proteinase K. Genomic tail DNA (10 g) was digested with EcoRI for 16 h. The digested DNA samples were separated on a 0.7% agarose gel. The gel was then transferred for 16 h onto Hybond-N ϩ membranes (Amersham Biosciences). The membrane was incubated with hybridization solution containing denatured radiolabeled probe for 16 h at 42°C. pTEX-Ets2 T72A mice transgenic integration was detected using a transgene-specific 500-bp BamHI-XhoI fragment from the human CD2 region of the transgenic vector.
Northern Blot Analysis-Total RNA (15 g) was isolated using TRIzol reagent (Invitrogen) and loaded onto a 1% RNAagarose gel (MOPS, (2%) formaldehyde). The gel was blotted overnight onto a nylon Hybond-N ϩ membrane (Amersham Biosciences) in 10ϫ SSC. The membrane was incubated with prehybridization solution followed by hybridization solution containing denatured radiolabeled probe for 16 h at 42°C. The membrane was then exposed to Biomax MR x-ray film (Eastman Kodak). Expression was detected by using either a transgene-specific 1.2-kb NcoI-PvuII fragment or by using a 500-bp ETS2-specific PstI fragment from the pCG-ETS2 T72A expression vector.
Analysis of Cell Proliferation-Thymocytes (2.5 ϫ 10 6 cells/ ml) were incubated in 200 l of medium into a 96-well plate (Falcon). PMA (100 ng/ml) and ionomycin (100 ng/ml) were obtained from Calbiochem. Purified anti-CD3 antibody, clone 2CII (Pharmingen), was resuspended in sterile PBS at a concentration of 15 g/ml, and 200 l was used to pre-coat individual wells in a 96-well plate for 2 h at 37°C. Coated wells were washed twice with PBS before adding cell suspensions. Cells were grown in RPMI 1640 medium supplemented with 10% of heat-inactivated FCS, 50 M ␤-mercaptoethanol, 2 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, and 0.1 mM nonessential amino acid (Invitrogen) at 37°C with 5% CO 2 for 40 h. One Ci of [ 3 H]thymidine (2.0 Ci/mmol) was then added per well, and cells were grown for 8 h. The cells were harvested, and the incorporated [ 3 H]thymidine was measured by a scintillation counter.

Ets2 in T-cell Development, Survival, and Apoptosis
were obtained from Pharmingen. Flow cytometric analysis was performed on Coulter EPICS XL (Fullerton, CA).
Cell Cycle Analysis-Thymocytes (10 ϫ 10 6 ) were washed twice in sample buffer (1 g of glucose in (1000 ml) Ca 2ϩ /Mg 2ϩfree PBS) and then resuspended in 0.5 ml of sample buffer and fixed for 16 h by dropwise addition of 1.5 ml of 100% ethanol at 4°C. Fixed cells were spun at 2500 rpm for 10 min and resuspended in propidium iodide staining solution (Sample buffer, RNase A (1 mg/ml), propidium iodide (5 g/ml)) for 30 min at room temperature. Stained cells were analyzed by flow cytometry.
Cell Death Analysis-Thymocytes (1 ϫ 10 6 cells/ml) were resuspended in 5 ml of media and cultured at 37°C, 5% CO 2 , for the indicated period in a 6-well plate and analyzed. Cells were washed once in PBS ϩ 0.1% BSA and then stained with the indicated fluorochrome-labeled antibodies as described above. Cells were then incubated with 7-aminoactinomycin D (7-AAD) (Pharmingen) (0.025 g) for 10 min at room temperature in 100 l of PBS ϩ 0.1% BSA. Samples were analyzed by flow cytometry for both 7-AAD and antibody fluorochrome staining.
Histochemistry-Thymi were fixed in buffered 10% formalin solution for 16 -18 h. The samples were then placed in 75% ethanol for 16 h and processed through graded alcohol (from 70 to 100%), embedded in paraffin wax, and sliced into 4-m sections and mounted on glass slides. The sections were deparaffinized in xylene and hydrated through a descending alcohol series (100 -70%) and finally distilled water. The sections were rinsed in water for 1 min and placed in hematoxylin solution for 10 min and rinsed in water for 1 min. The section was then placed in eosin solution for 2 min and rinsed with 95% ethanol for 2 min, followed by rinsing in 100% ethanol for 1 min three times. Finally, the section was washed with 100% ethanol/ americlear mixture for 2 min and then americlear for 2 min three times.
Transient Transfections-Transient transfections were carried out using Transfectin Lipid Reagent from Bio-Rad. Jurkat cells were grown at 8 ϫ 10 5 cells/ml at 37°C (5% CO 2 ) 24 h before transfection. The day of transfection, 6 ϫ 10 5 cells were plated in a 24-well plate (Falcon) in 0.5 ml of RPMI 1640 ϩ 10% FBS ϩ (2 mM) L-glutamine, and (0.1 mM) nonessential amino acid. For each well, 1 g of DNA was added to 50 l of Opti-MEM I, and 3 l of transfectin was added to 50 l of Opti-MEM I. The DNA and transfectin were mixed and incubated for 20 min at room temperature. The DNA/transfectin mixture (100 l) was added dropwise to Jurkat cells. One ml of media was added to the transfected cells 16 h after transfection. Transfected Jurkat cells were analyzed 48 h after transfection.
Luciferase Assay-Luciferase reporter analysis of transiently transfected Jurkat cells was performed using a Dual-Luciferase assay (Promega, Madison, WI). Cells from individual wells of the 24-well transient transfections were washed twice in ice-cold PBS, resuspended in 100 l of 1ϫ passive lysis buffer, and incubated at room temperature for 15 min. Twenty l of the lysed cells were transferred to a 5-ml tube. LAR II reagent (100 l) was added to the 5-ml tube to initiate the firefly luciferase reaction. The firefly luciferase reaction was measured on a luminometer. After measurement, 100 l of Stop and Glo reagent was added to stop the firefly luciferase reaction and initiate the Renilla luciferase reaction, which was measured on a luminometer. Luciferase measurement was expressed as a ratio of firefly luciferase intensity to Renilla luciferase intensity.
TUNEL Assay-Thymi were embedded in paraffin, sliced, mounted on glass slides, deparaffinized, and rehydrated as described above. Specimens were permeabilized with proteinase K (20 g/ml) for 20 min at room temperature. Slides were washed in 3% hydrogen peroxide for 5 min, and covered in 1ϫ TdT equilibrium buffer (TdT-FragEL DNA fragmentation detection kit, Calbiochem). Sixty l of TdT labeling reaction mixture was added to the specimen for 2 h. Slides were covered in 3,3Ј-diaminobenzidine solution for 5 min. Slides were washed in water and then counterstained with methyl green for 3 min.
Annexin V Apoptosis Analysis-Thymocytes were incubated in RPMI 1640 ϩ 10% FBS at 37°C for the indicated times. Thymocytes were then washed twice with PBS and resuspended in 1ϫ binding buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl 2 ) at a concentration of 1 ϫ 10 6 cells/ml. Five l of annexin V (BD Biosciences) and 2 l of propidium iodide (50 g/ml) were added to the cells. Thymocytes were then mixed and incubated at room temperature for 15 min. Binding buffer (400 l) was added to each tube, and cells were analyzed by flow cytometry.
In Vivo 5-Bromo-2-deoxyuridine (BrdU) Incorporation-Nontransgenic and transgenic Ets-2 T72A mice were injected intraperitoneally with 1 mg of BrdU (Pharmingen). Ninety minutes later, the mice were sacrificed, and thymocytes were stained for the surface antigens CD4, CD8, CD5, and CD117, as described above. Thymocytes were fixed and permeabilized by incubation in Cytofix/Cytoperm (BD Biosciences) buffer for 30 min on ice. Thymocytes were then washed once in 1ϫ Perm/Wash buffer (BD Biosciences). Cells were spun at 450 ϫ g for 5 min, and the supernatant was discarded. Thymocytes were resuspended in Cytoperm Plus Buffer (BD Biosciences) and incubated for 10 min on ice. The cells were washed in 1ϫ Perm/Wash buffer (BD Biosciences), spun, and resuspended in 100 l of DNase (300 g/ml) for 1 h at 37°C. The cells were then resuspended in 50 l of Perm/Wash buffer (BD Biosciences) with anti-BrdU FITC antibody or an isotype control (Pharmingen) for 20 min at room temperature. Samples that were analyzed for cell cycle status were resuspended in 20 l of 7-AAD, before addition of staining buffer. Samples were analyzed by flow cytometry.

Ras-mediated Activation of Ets-2 in T-cells Is Dependent on
Threonine 72-In a variety of cell types Ets-2 is regulated by the Ras-dependent MAPK ERK (10,11). Activated ERK phosphorylates Ets-2 at threonine 72, which results in transcriptional transactivation. Mutation of threonine 72 to an alanine in Ets-2 protein (Ets-2 T72A ) prevents Ras-dependent MAPK-mediated activation of Ets-2 and thus inhibits Ets-2-dependent transcription. To determine whether Ets-2 is regulated in a Ras-dependent mechanism in T-cells, Jurkat T-cells were transiently cotransfected with either wild type Ets-2 or a phosphomutant Ets-2 T72A , along with a constitutively active Ras, and a luciferase reporter gene driven by the Ets-2-dependent urokinase plasminogen promoter (Fig. 1) FEBRUARY 17, 2012 • VOLUME 287 • NUMBER 8 2-mediated transcription of the luciferase reporter gene in Jurkat T-cells (Fig. 1). Mutation of threonine 72 to alanine in Ets-2 resulted in the inhibition of Ras-dependent transactivation of Ets-2. These data suggest the importance of threonine 72 in Ras-dependent activation of Ets-2 ( Fig. 1).

Ets2 in T-cell Development, Survival, and Apoptosis
Generation and Characterization of a Phosphomutant Ets-2 T72A Transgenic Mice-Ets-2 protein is expressed constitutively in thymocytes and purified splenic T-cells ( Fig. 2A). To determine the in vivo importance of the Ras-mediated Ets-2 activation in thymocyte development, we generated transgenic mice that overexpressed a phosphomutant human Ets-2 (Ets-2 T72A ) under the control of the CD2 promoter/enhancer element. The human CD2 promoter/enhancer cassette has been used previously to express transgenes in mouse thymocytes (18). A schematic representation of the pTEX-Ets-2 T72A transgenic construct is shown in Fig. 2B. The contiguity of the construct and the replacement of threonine 72 to alanine were confirmed by sequencing and restriction analysis ( Fig. 2C and data not shown).
Four independent founder lines, T72A-1, T72A-2, T72A-3, and T72A-4, with 3, 10, 30, and 100 copies of the transgene, respectively, were identified by Southern blot analysis of tail DNA digested with EcoRI and probed with the CD2 enhancer probe shown in Fig. 2D. RT-PCR analysis of total RNA from Tg and Ntg thymocytes with human Ets-2-specific primers detected the transgene-specific transcripts in founder lines T72A-3 and T72A-4 (Fig. 2E, founder line T72A-3). Ets-2 has been previously shown to have up to three major RNA bands at 4.7, 3.2, and 2.7 kb (21). Northern blot analysis of thymocyte RNA with an ETS2 probe detected a 4.5-kb transcript in transgenic but not nontransgenic littermates (Fig. 2F, founder line T72A-3). Consistent with the RNA expression, a rabbit polyclonal Ets-2 antibody that recognizes human and mouse Ets-2 protein detected transgenic Ets-2 proteins in the transgenic thymocytes, while detecting only the endogenous Ets-2 in thymocytes from nontransgenic littermate controls (Fig. 2G, founder line T72A-3). The founder line T72A-3 expressed the highest levels of transgenic protein and was primarily used for detailed analysis described in this report. A similar phenotype was confirmed independently in founder line T72A-4.
Reduced Thymic Size, Weight, and Cell Number in Ets-2 T72A Transgenic Mice-Comparison of intact thymi from nontransgenic and transgenic Ets-2 T72A littermates revealed a marked reduction in organ size (Fig. 3A) associated with decreased organ to body weight ratio by ϳ50% (0.32 Ϯ 0.06% in nontransgenic and 0.16 Ϯ 0.06% (p Ͻ 0.01) in transgenic mice) (Fig. 3B). In agreement with the reduced thymic size and weight, ϳ70% decrease in total thymocyte number was observed in the transgenic mice compared with nontransgenic littermates (29 Ϯ 17.4 ϫ 10 6 in transgenic mice versus 104.3 Ϯ 20.4 ϫ 10 6 in nontransgenic littermates (p Ͻ 0.005)) (Fig. 3C). With the exception of the size of the thymus, examination of the stained sections revealed no distinct morphological changes in the Ets-2 T72A transgenic thymus (Fig. 3D). Transgenic Mice-To determine whether the decreased thymocyte numbers were associated with altered thymocyte mat-uration, thymocyte development was analyzed by flow cytometry. Although total cell number was significantly reduced, the percentage of CD4 ϩ CD8 ϩ double-positive and CD4 ϩ CD8 Ϫ and CD4 Ϫ CD8 ϩ single-positive populations within the thymus were not significantly altered between Tg and Ntg littermates (Fig. 4A).
Interestingly, two unique populations existed within the transgenic thymocytes that were significantly different from that observed in their nontransgenic littermates. Approximately 30 -40% of the total thymocyte population in the transgenic mice were c-Kit ϩ (CD117), compared with ϳ2% in nontransgenic littermates (36.0 Ϯ 9% for Tg and 2.8 Ϯ 1.1% for Ntg, p Ͻ 0.0001, n ϭ 8) (Fig. 4A, middle panels). Thymocytes from the low expressing T72A-4 line also exhibited significant increase in c-Kit ϩ cells (data not shown). A unique CD5 low thymocyte population (MFI 35 Ϯ 7) was also identified in transgenic mice compared with nontransgenic littermates. The CD5 low population represented ϳ25-30% of the total thymocyte population in the transgenic mice (36.2 Ϯ 2.1% for Tg and 8.4 Ϯ 2.4% for Ntg, p Ͻ 0.0001, n ϭ 7). In Ntg animals, the largest populations of thymocytes express CD5 at an intermediate (MFI 150 Ϯ 15) or high (MFI 800 Ϯ 25) expression level (Fig. 4A, bottom panel).
Normal Thymocyte Proliferation in Ets-2 T72A Transgenic Mice-To determine whether the reduction in transgenic thymocyte numbers was due to defective proliferation, total thymocytes from transgenic or nontransgenic littermates were cultured in vitro with either immobilized anti-CD3 or the protein kinase C activator PMA and the calcium ionophore ionomycin. [ 3 H]Thymidine incorporation was measured to determine cell proliferation. In vitro, the transgenic thymocytes were found to proliferate at levels comparable with their nontrans-  FEBRUARY 17, 2012 • VOLUME 287 • NUMBER 8 genic littermates, in response to anti-CD3, or PMA ϩ ionomycin stimulation (Fig. 5A).

Ets2 in T-cell Development, Survival, and Apoptosis
To examine if the thymocytes exhibited proliferative differences in vivo, nontransgenic and Ets-2 T72A transgenic mice were injected i.p. with 1 g of BrdU. Ninety minutes after injection, the mice were sacrificed, and the thymocytes were analyzed for BrdU incorporation and total DNA content (Fig. 5B). BrdU incorporation into the thymocytes was readily observed, and each stage of the cell cycle was examined. No significant differences were observed between nontransgenic and transgenic thymocytes at any cell cycle stage. In the G 0 /G 1 stage, we found nontransgenic thymocytes 87.3 Ϯ 0.2% and 86.8 Ϯ 2.6% of transgenic thymocytes. 11.2 Ϯ 1.5% of Ntg and 12.6 Ϯ 2.4% of Tg thymocytes were in the S phase. 0.46 Ϯ 0.2% of Ntg and 0.42 Ϯ 0.15% of Tg thymocytes were in G 2 /M stage. 0.15 Ϯ 0.2% of Ntg and 0.05 Ϯ 0.01% of Tg mice were in a sub-G 0 phase. These data along with the in vitro data suggest normal proliferation in Ets-2 T72A transgenic thymocytes.
Increased in Vitro and in Vivo Apoptosis in Ets-2 T72A Transgenic Thymocytes-The finding of comparable levels of total thymocyte proliferation yet decreased cell numbers in the Ets-2 T72A transgenic thymocytes suggested a potential increase in apoptosis. To test if cell death rate played a role in the decrease in transgenic thymocyte cell number, thymocytes were cultured in vitro and stained with propidium iodide to determine the cell cycle profile based on cellular DNA content by flow cytometry (Fig. 6A). In agreement with the BrdU/7-AAD data, both freshly isolated nontransgenic and transgenic Ets-2 T72A thymocytes exhibited comparable cell cycle profiles. However, Ets-2 T72A thymocytes cultured in vitro for 12, 24, and 36 h revealed a difference in the percentage of cell death as early as 12 h as measured by the sub-G 0 hypodiploid population (14.3% Ntg to 27.2% Tg). The propensity to die increased in a time-dependent manner in the transgenic Ets-2 T72A population. Total transgenic thymocytes cultured for 36 h in vitro displayed an ϳ2-fold increase in the sub-G 0 population as compared with nontransgenic thymocytes (38.7% Ntg to 71.3% Tg) (Fig. 6A).
To examine the apoptotic properties of the thymocyte subpopulations, nontransgenic and transgenic Ets-2 T72A thymocytes were cultured in vitro in media at 37°C. After 6 h, the NTg and Tg thymocytes were analyzed for annexin V staining, an indicator of early apoptotic cellular events. Dual staining with the vital dye propidium iodide was used to distinguish live cells,
To determine whether any of the thymocyte subpopulations exhibited a selective increase in in vitro apoptosis, nontransgenic and transgenic Ets-2 T72A thymocytes were analyzed by 7-aminoactinomycin D (7-AAD) staining. 7-AAD allows for the determination of cell death, along with the ability to stain by fluorescently labeled antibodies for different cell populations. NTg and Tg thymocytes were analyzed for cell death after 6 h of in vitro culture in media. In nonstimulated cultures transgenic thymocytes showed a 2-fold increase in cell death in the CD4, CD8, and c-KitϪ populations, when compared with nontrans-genic littermates (Fig. 6C). These data suggest that all transgenic thymocyte populations are more susceptible to increased apoptosis compared with thymocytes from nontransgenic littermate controls.

Analysis of Apoptosis in Vivo by TUNEL Assay in Thymus-
To determine whether the above-mentioned in vitro apoptosis is observed in vivo, thymi from 4-to 6-week-old nontransgenic and transgenic mice were analyzed for 3Ј apoptotic DNA breaks by TUNEL (Fig. 6D). Hematoxylin and eosin staining of similar sections was used to identify the thymic structure of examined sections (Fig. 6D). To quantify the rates of apoptosis in nontransgenic and transgenic thymi, TUNEL sections were analyzed under ϫ1000 bright field microscopy. Fifteen random fields from three nontransgenic and three transgenic TUNEL sections were analyzed for apoptotic HRP-positive cells (Fig.  6D). In a field of ϳ1000 cells, nontransgenic thymus contained 6.8 Ϯ 0.3 apoptotic cells and the Tg thymus contained 11.6 Ϯ 1 apoptotic cells (p Ͻ 0.015). Thus, the in vivo analysis concurred with the in vitro data.
Overexpression of Bcl-2 Partially Rescues the Developmental Defects in the Ets-2 T72A Transgenic Thymus-The anti-apoptotic protein Bcl-2 has been implicated in thymocyte development. Intracellular staining of Ets-2 T72A thymocytes revealed decreased Bcl-2 expression, suggesting a potential role for Bcl-2 in the observed developmental block and hypocellularity (Fig.  7). To test if overexpression of Bcl-2 will rescue the observed defects in the Ets-2 T72A transgenic mice, hemizygous transgenic Ets-2 T72A mice were bred to E-Bcl-2-36 transgenic hemizygous mice. The E-Bcl-2-36 transgenic mice overexpress the anti-apoptotic protein Bcl-2 in thymocytes and B-cells (22). The Bcl-2 ϫ Ets-2 T72A double transgenic mice were compared with both single transgenic Ets-2 T72A and Bcl-2 transgenic littermates (Fig. 8). Thymocyte cell population in the double transgenic mice, when compared with single transgenic Ets-2 T72A mice, was rescued to the levels observed in nontransgenic mice. The average total thymocyte cell number increased from 29 ϫ 10 6 to 110 ϫ 10 6 , along with an increase in overall thymic organ size in double (Ets-2 T72A ϫ Bcl-2) transgenic mice. Furthermore, the c-Kit ϩ population decreased 2-fold, from 30% of the population seen in single Ets-2 T72A transgenic mice to 15% in the Ets-2 T72A x Bcl-2 double transgenic mice (Fig. 8, A and B). However, not all unique transgenic Ets-2 T72A populations were altered by overexpression of Bcl-2. The CD5 low population was still present in Bcl-2/Ets-2 T72A double transgenic mice at levels comparable with single transgenic Ets-2 T72A mice (data not shown), suggesting that the overexpression of Bcl-2 could rescue the c-Kit ϩ population in vivo, but not reverse all phenotypes associated with the expression of the Ets-2 T72A phosphomutant.

DISCUSSION
Here we have shown a critical role for Ets-2 T72 activation in T-cell development and survival. Transgenic mice overexpressing a phosphomutant Ets-2 T72A under the control of the CD2 promoter/enhancer showed altered thymocyte development associated with a small, hypocellular thymus. A unique c-Kit ϩ population, consisting of ϳ30 -40% of the total thymocytes, exists in Ets-2 T72A transgenic mice. A second population defined by low levels of CD5 represents an ϳ6-fold increase over nontransgenic littermates. Furthermore, there is a block at the DN2-DN3 stage of thymocyte development in Ets-2 T72A transgenic animals. In vitro and in vivo survival of transgenic thymocytes is also decreased. The Ets-2 T72A transgenic thymocyte c-Kit ϩ population and the CD5 low population are either Bcl-2-negative or Bcl-2-low. Ets-2 T72A transgenic mice crossed to E-Bcl-2-36 transgenic mice resulted in a rescue of thymocyte cell number and a partial rescue of the c-Kit ϩ population.
Ets-2 T72A mice appear to have a developmental block that is associated with the DN2 and DN3 stage of thymocyte development. In the Ets-2 T72A transgenic mice, we observed an ϳ8-fold decrease of DN2 stage thymocytes and an ϳ4-fold decrease in DN3 stage thymocytes (Fig. 4B). During these stages of development, TCR␤ selection occurs, and signaling events are initiated by the pre-TCR. A similar developmental block is observed

Ets2 in T-cell Development, Survival, and Apoptosis
in thymocytes that have targeted disruption of ZAP-70, Syk (29), and other downstream signaling molecules such as LAT, SLP-76, and PLC-␥1 (30). These data suggest that Ets-2 is involved in either ␤-selection or is a downstream target of pre-TCR signaling. This block may, in part, explain the size and cellularity of the Ets-2 T72A transgenic thymus.
Interestingly, analysis of double-negative thymocytes revealed that both transgenic and nontransgenic mice had an equal percentage of CD44 Ϫ and CD25 Ϫ (DN4) thymocytes. During thymocyte development, DN4 stage thymocytes undergo a proliferative burst (31). In agreement with our in vitro and in vivo proliferation data (Fig. 5, A and B), Ets-2 may not be involved in thymocyte proliferation, and thus transgenic thymocytes that successfully mature to DN4 stage are able to proliferate and populate the thymus with functionally competent thymocytes.
Ras signaling events are involved in multiple stages of thymocyte development. The importance of Ras signaling in the double-positive stage of thymocyte development has been demonstrated in transgenic mice that overexpress Ras/MAPK dominant-negative constructs in thymocyte development. Ras, Raf, and MEK all block thymocyte development at the double-  positive stage (2,4,8,32,33). Inhibition of Ras signaling appears to affect positive selection but not negative selection (34). Because Ets-2 has been demonstrated to be both a downstream target of Ras (10) and Ets-2 is activated during CD3 signaling (16), it will be interesting to determine whether there are defects in positive selection in transgenic thymocytes similar to those observed in dominant-negative Ras transgenic mice.
Two major population changes are observed in the Ets-2 T72A thymocytes, overexpression of c-Kit and a decrease in the MFI of CD5. c-Kit expression has primarily been characterized as being expressed in very early immature thymocytes. The c-Kit overexpression observed in transgenic thymocytes may result from the DN2-DN3 developmental block. However, c-Kit has also been suggested to play an important role in thymocyte selection. c-Kit ϩ and c-Kit Ϫ thymocyte expression patterns have been suggested to represent two critical pathways in thymocyte-positive and -negative selection (35,36). In this model, c-Kit ϩ cells are able to proceed through positive selection toward single-positive mature T-cells. In Ets-2 T72A transgenic mice the c-Kit population, which consists of up to 40% of the population, may be the result of an alternative developmental pathway, allowing progression through the double-positive stage and/or during positive selection.
CD5 is expressed in thymocytes from a very early CD4 Ϫ / CD8 Ϫ stage. Interestingly, CD5 is a negative regulator of TCR/ CD3 signaling (37)(38)(39). The down-regulation of CD5 observed in Ets-2 T72A transgenic mice could be a developmental response to an interruption in CD3-mediated signaling. Decreasing expression of CD5, a negative regulator, could modulate TCR sensitivity and overcome defects induced by expression of the Ets-2 T72A transgene. This would suggest that at least a component of CD3 signaling is mediated through Ets-2. Although CD3-stimulated proliferation is unaffected in vitro, other maturation or developmental pathways mediated through CD3 could involve activation of Ets-2 on threonine 72.
Our group has previously demonstrated that Ets2 can be both an activator and a repressor; phosphorylation of the Thr-72 residue acts as a molecular switch. In the unphosphorylated form, mimicked by the T72A mutation, Ets2 interacts with distinct corepressor complexes, including the BS69-nuclear receptor corepressor-histone deacetylase complex and a repressive BRG1 complex. The Thr(P)-72 protein interacts with a distinct set of coactivators, including the CREB-binding protein complex (40 -43). In addition, we and others have established that BCL family members (BCL2 and BCLx-X) are targets of Ets2 (12). Based on these data, our model for the mechanism is that the Ets2 T72A repressor complex becomes dominant in the transgenic mice, leading to a lack of responsiveness to upstream ERK signaling cues and a lack of target gene expression. A key experiment that supports this model is the BCL2 rescue experiments, i.e. restoring the pro-apoptotic response caused by Ets2 deletion rescues the T-cell phenotype. A second experiment that supports this model is expression of the Ets2 T72A transgene in mice lacking the closely related factor Ets1; the Ets2 phenotype is dominant as shown in Fig. 9, indicating that the effect of the transgene is specific for Ets2 and not a "dominant-negative" that affects closely related Ets factors.
Could Ets-2 T72A transgenic expression permit other transcription factors to bind in place of Ets-2 or result in novel gene activation? Previous promoter studies with Ets-2 T72A suggest the primary effect of the T72A mutation is the abrogation of Ras-mediated Ets-2 activation (10 -13). Although Ras/MAPKmediated activation of Ets-2 is probably the primary means of activating Ets-2, it is not the only means to activate Ets-2. Prior studies have demonstrated that Ets-2 can bind and activate the c-fms promoter in a Ras/MAPK independent mechanism (12). In those studies, wild type Ets-2 and Ets-2 T72A are both able to transactivate the c-fms promoter to a similar degree. Additionally, it is also possible that the presence of Ets-2 T72A permits other ETS family member to bind Ets-2 sites. Because of the similarity between Ets-1 and Ets-2 in both structure and function, we believe that Ets-1 would be the most likely candidate to compete for Ets-2 sites. Thymocytes from Ets-2 T72A ϫ Ets-1 Ϫ/Ϫ mice displayed a similar phenotype to Ets-2 T72A transgenic mice (Fig. 9). These data suggest that Ets-2 T72A overexpression mitigates Ras-mediated signaling without altering its binding preferences or permitting other ETS family members to inappropriately bind Ets-2 sites.
Ets-2 has previously been implicated in both cell survival and apoptosis. Overexpression of wild type Ets-2 in neuronal cells (44), the prostate tumor cell line PPC-1 (45), and in transgenic thymocytes that express Ets-2 from the sheep metallothionein promoter results in an increase in cell death (46). Zaldumbide et al. (27) also overexpressed wild type Ets-2 in developing thymocytes, but they used the proximal Lck promoter and did not observe an increase in apoptosis. However, Zaldumbide et al. (27) did observe an increase in apoptosis in thymocytes that overexpressed an Ets-2 DNA binding domain from the proximal Lck promoter. Neither of these two groups reported any of the developmental defects that were observed in the Ets-2 T72A transgenic mice reported here. It is possible that the observed increase in apoptosis in Ets-2 T72A transgenic thymocytes is a consequence of early DN developmental defects. However, the role that Ets-2 T72A plays in the expression of anti-apoptotic factors, such as Bcl-2, might directly make thymocytes more susceptible to pro-apoptotic signals. This is consistent with the ability of overexpressed Bcl-2 to rescue the thymocyte numbers comparable with the wild type nontransgenic controls. Mouse models deficient in Ets-2 have not clarified the functional role of Ets-2 in lymphoid tissue due to an embryonic lethal phenotype (47). In this context the Ets-2 T72A transgenic mice and the transgenic model overexpressing Ets-2 DNA binding domain in T-cells will be extremely useful in identifying the Ets-2 targets and their role in thymocyte development and function.