The Anti-Apoptotic Bcl-2 Family Member Mcl-1 Promotes T Lymphocyte Survival at Multiple Stages¹

The life of a T lymphocyte can be divided into multiple distinct stages (1). In the thymus, T lymphocytes are generated through several well-characterized steps marked by the expression of the coreceptors CD4 and CD8. Thymocyte precursors that are CD4⁻CD8⁻ double negative (DN)³ mature into CD4⁺CD8⁺ double positive (DP) and then either CD4⁺ or CD8⁺ single positive (SP) thymocytes. SP thymocytes exit to the periphery as naive T cells. When a naive T cell encounters its cognate Ag, it undergoes activation and proliferation and differentiates into effector T cells. After the peak of the immune response, 90–95% of the effector T cells are eliminated during the contraction phase, while the remaining 5–10% of effector cells become memory T cells (1).

Proteins of the Bcl-2 family function in the intrinsic death pathway, playing important roles in regulating the life and death of T lymphocytes (2, 3, 4). These proteins contain between one and four distinct regions of homology termed Bcl-2 homology (BH) domains. The members of the Bcl-2 family can be broadly divided into anti-apoptotic and pro-apoptotic groups. Anti-apoptotic Bcl-2 family members including Bcl-2, Bcl-x_L, and Mcl-1 are expressed in T lymphocytes. For example, both Bcl-2 and Mcl-1 are expressed in DN thymocytes, whereas the expression of Bcl-x_L is largely inversely correlated with that of Bcl-2 (5, 6, 7, 8, 9). Bcl-2 is down-regulated in DP and up-regulated in SP thymocytes, whereas Bcl-x_L is strongly up-regulated in DP but down-regulated in SP cells (6, 7, 8, 9). Furthermore, Bcl-2 is expressed at a relatively high level in naive T cells and down-regulated in activated/effector CD8⁺ T cells (10, 11, 12). In contrast, Bcl-x_L is not expressed in naive T cells but is rapidly up-regulated in activated/effector T cells (11, 13). The complex pattern of expression of Bcl-2 and Bcl-x_L may be related to observations that these two proteins function during distinct stages of T lymphocyte development.

Genetic studies using mice lacking Bcl-2 or conditionally lacking Bcl-x have demonstrated that these anti-apoptotic proteins promote the survival of T lymphocytes at different stages. Deletion of Bcl-2 results in rapid disappearance of naive T cells (14, 15, 16), whereas deletion of Bcl-x impairs the survival of DP thymocytes (17, 18, 19). Although Bcl-x_L is up-regulated in activated T cells, our recent data show that the development of effector T cells is not impaired in Bcl-x_L conditional knockout mice (19), suggesting that other anti-apoptotic proteins may promote the survival of activated T cells. A likely candidate for this function is Mcl-1. Mcl-1 was initially discovered as a gene induced early during differentiation of the myeloid cell line ML-1 (20). Genetic studies have demonstrated that Mcl-1 functions as an important anti-apoptotic protein in several different cell types. Deletion of Mcl-1 in mice leads to embryonic lethality owing to a failure of implantation of the blastocyst in the uterus (21). Conditional deletion studies reveal that Mcl-1 is required for early lymphoid development and naive T cell survival (22). Furthermore, Mcl-1 also promotes the survival of neutrophils and hematopoietic stem cells (23, 24).

Although it is clear that Mcl-1 is an important molecule for T cell survival, many aspects of its role remain to be addressed. In the present study, we focused on two major questions: what is the expression pattern of Mcl-1 in the different T cell populations and at what stage of T cell development it is essential for cell survival. To answer these questions, we have developed a flow cytometry-based assay for Mcl-1 expression and generated novel genetic mouse models. Our results demonstrate that unlike Bcl-2 and Bcl-x_L, Mcl-1 is expressed throughout T cell development and is up-regulated by TCR stimulation. Conditional deletion of Mcl-1 resulted in impaired cell survival not only of DN thymocytes and naive T cells as previously reported, but also of SP thymocytes and activated T cells (22). Although Mcl-1 was not essential for the survival of DP thymocytes, the combined deletion of Mcl-1 and Bcl-x_L resulted in a dramatic reduction of the size of this population in vivo. Taken together, our data demonstrates that Mcl-1 plays a critical role for the survival of T cells throughout their development and functions.

Mice containing conditional Mcl-1^f/f or Bcl-x^f/f alleles were generated in our laboratory as previously described (19, 23). Mcl-1^f/f and/or Bcl-x^f/f mice were bred with Lck-Cre (25) (The Jackson Laboratory), CD4-Cre (26) (Taconic Farms), and ER-Cre (27) mice. The phenotype of the Mcl-1^f/f mice is indistinguishable from those of wild-type (Mcl-1^+/+) C57BL/6 × 129 mice and these mice were used as controls throughout the study. For some of the experiments, Mcl-1^f/fCD4Cre mice were backcrossed to C57BL/6 for seven generations. Bcl-2 transgenic mice (28) were provided by Dr. Motonari Kondo at Duke University (Durham, NC). All mice were used at 6–10 wk of age. Animal usage was conducted according to protocols approved by the Duke University Institutional Animal Care and Use Committee.

Splenocytes, thymocytes, and lymph node cells were obtained by disrupting the respective organs with a syringe plunger. Bone marrow cells were flushed out of the femurs and tibiae. This was followed by filtering through 90-μm nylon mesh (SEFAR Filtration) to generate single-cell suspensions. Single-cell suspensions of thymus, spleen, lymph nodes, and bone marrow were lysed of RBC by treatment with ACK buffer. FcRγ-receptor binding was blocked with 2.4G2 hybridoma supernatant. The following Abs were used for cell surface staining: CD4, CD8, CD3, CD44, CD25, TCRβ, CD62L, CD69, Qa-2, B220, CD43, CD11b, GD11c, and Gr-1 conjugated to FITC, PE, PE/cy5, PE/cy7, allophycocyanin, allophycocyanin/cy7, or allophycocyanin/Alexa Fluor 750 from Pharmingen, BioLegend, or eBioscience. The cells were resuspended in FACS buffer containing propidium iodide (Sigma-Aldrich). Data were collected on a FACScan, FACStarPLUS, or FACSVantage SE Diva flow cytometer (BD Biosciences) and analyzed using CellQuest software (BD Biosciences). Apoptosis was measured by Annexin V and 7-aminoactinomycin D (7-AAD) staining using the Annexin V-PE kit (BD Pharmingen) according to the manufacturer’s instructions.

Intracellular staining for Mcl-1 was performed by first staining for cell surface markers following the above protocol. The cells were then washed twice and fixed with 2% paraformaldehyde (Sigma-Aldrich) in PBS for 20 min on ice. Permeabilization was achieved by incubating the cells in 0.1% Saponin (Sigma-Aldrich) in FACS buffer containing 5% donkey serum for 20 min on ice. Mcl-1 was detected with polyclonal rabbit anti-Mcl-1 Ab (Rockland Immunochemicals) at 0.1 μg/10⁶ cells in permeabilization buffer for 1 h on ice. Normal rabbit Ig (Sigma-Aldrich) was used as an isotype control. A secondary donkey anti-rabbit-FITC Ab (Jackson ImmunoResearch Laboratories) was used at 1/100 in permeabilization buffer for 20 min on ice for final detection.

A total of 5 × 10⁶ cells were lysed with 100 μl 1×SDS sample buffer. Equal amounts of protein (20 μg) were separated on 10% polyacrylamide gel and transferred to polyvinylidene difluoride membrane (PerkinElmer) for chemiluminescent detection. The membranes were blocked for 1 h at room temperature with 5% nonfat milk (Carnation) in 0.02% Tween 20 (EMD Biosciences) in PBS (PBS-T) followed by overnight incubation at 4°C with the primary Ab diluted in 3% BSA (Sigma-Aldrich) in PBS-T. The primary Abs were rabbit anti-Mcl-1 used at 1/10,000 dilution (Rockland Immunochemicals). Rabbit anti-ERK2 and goat anti-actin were used at 1/1000 dilution (Santa Cruz Biotechnology). The membranes were incubated with the secondary Abs for 1–3 h at room temperature. The secondary Abs were anti-rabbit-HRP or anti-goat-HRP (Santa Cruz Biotechnology) used at 1/10,000 dilution. Proteins were detected with West Pico Chemiluminescence substrate (Pierce) for 5 min at room temperature. The membranes were exposed on Sterling high-speed x-ray film (BioWorld). For fluorescent Western blots, polyvinylidene difluoride-FL membranes (Millipore) were used. After primary Ab incubation, membranes were incubated in anti-rabbit Alexa-Fluor 680 (Invitrogen) or anti-goat IRDye 800 (Rockland Immunochemicals) at 1/5000 dilution for 1 h at room temperature. Proteins were detected using the Odyssey system (LI-COR).

Total thymocytes were lysed of RBC and 5 × 10⁵ cells were cultured in 200 μl complete RPMI 1640 (Invitrogen) medium for different periods of time in the presence of 2 ng/ml PMA, 20 ng/ml ionomycin, or 1 nM dexamethsone (all from Sigma-Aldrich). FACS sorted DP thymocytes were cultured at 1.5 × 10⁵ cells in 200 μl complete medium for 24 or 48 h in the presence of 2 ng/ml PMA, 1 nM dexamethasone, or 0.6 μM Tunicamycin (Sigma-Aldrich), or plate-bound anti-CD3 (2C11) and anti-CD28 (BioLegend).

Mcl-1^f/f-ER-Cre T cells were purified from total splenocytes using the EasySep T cell Enrichment kit (Stem Cell Technologies) and cultured at 10⁶ cells/ml for different periods of time in RPMI 1640 complete medium. The cells were stimulated with plate-bound anti-CD3 (2C11) and anti-CD28 Abs or 1 ng/ml IL-7. To induce deletion of Mcl-1 in vitro, Mcl- 1^f/fER-Cre T cells were treated with 0.2 μM 4-hydroxytamoxifen (4-OHT) (Sigma-Aldrich) dissolved in ethanol. Cell death was evaluated by propidium iodide staining.

Statistical analysis was performed using unpaired two-tailed Student’s t test using the GraphPad Prizm software (GraphPad).

The expression pattern of Mcl-1 was previously examined by RT-PCR and Western blot analysis at the total population level (22). This study showed that both T and B cells express Mcl-1. However, it was not known whether the expression level of Mcl-1 protein in T lymphocytes is differentially regulated throughout T cell development. To test this, we developed a flow cytometric assay to measure the intracellular levels of Mcl-1. We used Mcl-1-deficient T cell populations wherever possible as negative controls (DP, CD4⁺ SP, and CD8⁺ SP thymocytes from Mcl-1^f/fCD4-Cre mice as shown in Fig. 1,A). The fluorescence intensity of Mcl-1-deficient T cells stained with anti-Mcl-1 Ab was similar to the background staining seen in cells stained with an isotype-matched control Ab (Fig. 1,A). As the degree of background staining varied in the different T cell populations, we used the ratio of the fluorescence intensity of Mcl-1 staining to that of the background staining to represent the expression levels of Mcl-1 (Fig. 1).

Mcl-1 expression levels were similar in different populations of developing thymocytes including DN, DP, CD4⁺ or CD8⁺ SP (Fig. 1,A). Thymocytes at the transitional immature single positive (ISP) or post-positive selection stage (TCRβ^highCD69⁺ DP) also express similar levels of Mcl-1 (Fig. 1,A). Within the DN compartment, Mcl-1 expression was the highest in DN1 cells while thymocytes at the DN2-DN4 stages expressed slightly lower levels of Mcl-1 (Fig. 1,B). Mcl-1 protein was also detected in different populations of peripheral T cells (Fig. 1,C). Naive CD4⁺ and CD8⁺ (CD62L^highCD44⁻) T cells express levels of Mcl-1 largely comparable to those detected in memory CD4⁺ (CD62L^lowCD44⁺), effector memory CD8⁺ (CD62L^lowCD44⁺), and central memory CD8⁺ (CD62L^highCD44⁺) T cells (Fig. 1 C). Taken together, these data demonstrate that Mcl-1 protein is constitutively expressed by T lymphocytes at all developmental stages.

A previous report using conditional knockout mice has demonstrated that Mcl-1 plays a critical role in the survival of DN thymocytes and naive T lymphocytes (22). Given that Mcl-1 is expressed at all stages of T cell development, it is important to know whether Mcl-1 also promotes the survival of other T cell populations. For example, it is unknown whether Mcl-1 is required for the survival of SP thymocytes and activated effector cells, which also express other anti-apoptotic proteins, including Bcl-2 and Bcl-x_L. To address this, we generated mice lacking Mcl-1 in T lymphocytes by crossing Mcl-1^f/f mice that were previously generated in our laboratory (23) to a transgenic line expressing the Cre recombinase under control of the proximal Lck promoter (25). The Lck promoter drives Cre expression in early stages of T cell development, starting around the DN1-DN2 stage. Thus, Lck-Cre-induced deletion of Mcl-1 will allow the assessment of its role in the early stages of T cell development. Consistent with previously published results (22), thymic cellularity was greatly reduced in Mcl-1^f/fLck-Cre mice, as thymuses from these mice contained only ∼5% of the number of cells observed in a control thymus (Fig. 2,A). FACS analysis revealed that the frequencies of DN and CD4⁺ SP thymocytes were relatively increased while the frequency of DP cells was decreased (Fig. 2,B). The total cell numbers of DP, CD4⁺ and CD8⁺ SP thymocytes from Mcl-1^f/f Lck-Cre mice were only 5–10% of the controls, whereas the total number of DN thymocytes from Mcl-1^f/f Lck-Cre mice was ∼40% of the controls (Fig. 2, A and B). When the DN compartment of Mcl-1^f/fLck-Cre mice was analyzed, the frequencies of DN2/3 thymocytes were increased while the frequency of DN4 thymocytes was decreased (Fig. 2,B), suggesting a developmental blockade between the DN2/3 and DN4 stages. A likely explanation for this observation is that Mcl-1 promotes the survival of thymocytes undergoing the DN2/3 to DN4 transition. To test this, we examined the apoptotic rates of DN thymocytes from Mcl-1^f/fLck-Cre and control mice. As shown in Fig. 2,C, the apoptotic rate of DN thymocytes from Mcl-1^f/fLck-Cre mice was increased by ∼2-fold (Fig. 2 C).

We then examined the mature T cell compartment in Mcl- 1^f/fLck-Cre mice. The number of T cells in the spleen of Mcl- 1^f/fLck-Cre mice was 40–50% of that in control mice (Fig. 2, A and D). Given the dramatically reduced thymocyte number in Mcl-1^f/fLck-Cre mice, it was surprising to observe this substantial number of mature T cells. These cells could have been derived through an Mcl-1-independent pathway or have escaped deletion of Mcl-1. To distinguish between these two possibilities, we purified mature T cells from the spleen of Mcl-1^f/fLck-Cre mice and assessed them for Mcl-1 expression. Mature T cells in Mcl- 1^f/fLck-Cre mice expressed nearly the same level of Mcl-1 as control cells (Fig. 2 E), suggesting that the surviving T cells in the periphery had escaped deletion. Together, these results suggest that the requirement for Mcl-1 in the early stages of T cells development is essential for their further maturation.

The essential role of Mcl-1 in early stages of thymocyte development precluded our analysis of its function in thymocytes at later stages as mature T cells in Mcl-1^f/fLck-Cre mice largely escaped Mcl-1 deletion. To circumvent this problem, we crossed Mcl-1^f/f mice to another transgenic line expressing the Cre recombinase under control of the CD4 promoter (26). The CD4 promoter lacking the silencer element that restricts expression in CD8⁺ T cells drives Cre expression in all T cells from the onset of CD4 expression. Unlike Lck-driven Cre expression, CD4 promoter-driven Cre expression does not begin until the DN4 stage (26), bypassing the DN2/3 to DN4 transition.

In contrast to the severely reduced thymic cellularity observed in Mcl-1^f/fLck-Cre mice, the number of total thymocytes in Mcl-1^f/fCD4-Cre was comparable to that of control mice (Fig. 3,A). FACS analysis revealed that while the frequency of DN cells in Mcl-1^f/fCD4-Cre mice was similar to that of control mice, there was a slight increase in the frequency of DP cells and a substantial decrease in the frequencies of CD4⁺ and CD8⁺ SP cells (Fig. 3,B). Mcl-1 appears to be especially critical for the development of CD8⁺ SP thymocytes, as the frequency of CD8⁺ SP cells was reduced by >80% (Fig. 3,B). Furthermore, the majority of the CD8⁺CD4⁻ thymocytes did not express high levels of TCRβ (Fig. 3,B), indicating that these cells are ISPs, an intermediate stage between DN and DP cells. Thus, the frequency of mature CD8⁺ SP thymocytes in Mcl-1^f/fCD4-Cre mice was <0.2%, which represents a >10-fold reduction when compared with controls. We further assessed the maturation of SP thymocytes in Mcl-1^f/fCD4-Cre mice using several other markers including CD69, HSA, CD62L, and Qa-2 (29). SP thymocytes that have just undergone positive selection are CD69^highHSA^highCD62L^lowQa-2^low. As they mature in the thymic medulla, SP cells up-regulate CD62L and Qa-2 and down-regulate CD69 and HSA (29). As shown in Fig. 3 C, both CD4⁺ and CD8⁺ SP thymocytes with a more mature phenotype (CD62L^highCD69^lowQa-2^high) were decreased in Mcl-1^f/fCD4-Cre mice compared with controls, indicating an impaired mature SP thymocyte compartment.

We next determined the efficiency of Mcl-1 deletion in total thymocytes of Mcl-1^f/fCD4-Cre mice. Western blot analysis demonstrated that >95% of Mcl-1 was deleted in the thymocytes of Mcl-1^f/fCD4-Cre mice (Fig. 3,D). As CD4 promoter-driven Cre expression starts at the DN4 stage, this deletion efficiency suggests that most of the thymocytes in Mcl-1^f/fCD4-Cre mice have successfully deleted the Mcl-1 gene. Indeed, FACS analysis of intracellular Mcl-1 expression in DP, CD4⁺ SP, and CD8⁺ SP thymocytes from Mcl-1^f/fCD4-Cre mice demonstrated that Mcl-1 was completely absent in these cells (Fig. 1 A).

We then examined the apoptotic rates among different thymocyte populations in Mcl-1^f/fCD4-Cre mice. The frequencies of apoptotic DP or CD4⁺ SP thymocytes in Mcl-1^f/fCD4-Cre mice were largely comparable to those of control thymocytes (Fig. 3 E). In contrast, the apoptotic rates of CD8⁺ SP thymocytes in the mutant mice were increased. That no increase in the apoptotic rate of CD4⁺ thymocytes was observed in Mcl-1^f/fCD4-Cre mice may be due to the fact that it is difficult to detect apoptotic cells in the thymus as the dead cells are cleared very rapidly by resident macrophages, and only dramatically enhanced apoptosis can be detected using this approach.

Lastly, we examined the peripheral T cell compartment in Mcl-1^f/fCD4-Cre mice. The number of mature T cells in the spleen of Mcl-1^f/fCD4-Cre mice was <10% of that observed in control mice (Fig. 3,F). FACS analysis showed that very few CD4⁺ and CD8⁺ T cells were detected in Mcl-1^f/fCD4-Cre mice (Fig. 3 G). This dramatic reduction in the peripheral T cell compartment is consistent with the above observation that development of both CD4⁺ SP and CD8⁺ SP thymocytes depends on Mcl-1. Thus, Mcl-1 plays a critical role in the development of SP thymocytes.

The normal DP thymocyte compartment observed in Mcl-1^f/fCD4-Cre mice suggests that Mcl-1 is dispensable for in vivo DP thymocyte survival. Given that Mcl-1 is expressed in DP thymocytes, we wondered whether Mcl-1 regulates DP thymocyte survival in vitro, as the in vitro environment may provide fewer prosurvival signals than the in vivo environment. We, thus, examined the survival of Mcl-1-deficient thymocytes in vitro in the presence or absence of apoptotic stimuli. We used several well-characterized chemical agents that induce thymocyte apoptosis. Some of them, such as dexamethasone, ionomycin, and tunicamycin, are dependent on the BH3-only protein Bim, while PMA requires the BH3-only protein Puma for death induction (30, 31). In vitro cultured total thymocytes from Mcl-1^f/fCD4-Cre mice underwent apoptosis at a rate dramatically higher than control cells (Fig. 4,A). After 2 days in culture in the absence of any stimulus, <5% of Mcl-1-deficient thymocytes survived while ∼40% of control cells did (Fig. 4,A). As a positive control, thymocytes from Bcl-2 transgenic mice survived better than wild-type thymocytes (Fig. 4,A). Addition of PMA or ionomycin did not obviously change the death kinetics of these cells, while addition of dexamethasone dramatically facilitated the death of Mcl-1-deficient total thymocytes (Fig. 4 A).

To further confirm that in vitro cultured DP thymocytes depend on Mcl-1 for their survival, we sorted DP thymocytes from Mcl-1^f/fCD4-Cre and control mice and cultured them in vitro. The survival of Mcl-1-deficient DP thymocytes cultured with or without stimulation was significantly decreased when compared with control thymocytes (Fig. 4 B). These data demonstrate that Mcl-1 is required for the survival of DP thymocytes in vitro and suggests that Mcl-1-deficient thymocytes are protected by other anti-apoptotic factor(s) in vivo.

A likely candidate anti-apoptotic protein promoting DP thymocyte survival in vivo in Mcl-1^f/fCD4-Cre mice is Bcl-x_L, as Bcl-x_L but not Bcl-2 is highly expressed in DP thymocytes (6, 7, 8, 9). Our previous data using Bcl-x conditional knockout mice demonstrate that the DP thymocyte compartment is slightly reduced in the absence of Bcl-x, and that the in vitro survival of DP thymocytes depends on Bcl-x (19). These results suggest that Bcl-x_L and Mcl-1 may function redundantly to promote DP thymocyte survival in vivo. To test this, we generated conditional knockout mice lacking both Mcl-1 and Bcl-x in T cells by crossing Mcl-1^f/f and Bcl-x^f/f mice to CD4-Cre mice. Analysis of CD4/CD8 profiles of Mcl-1^f/fBcl-x^f/fCD4-Cre mice demonstrate that the frequency of DP thymocytes in these mice was reduced, as the DP compartment comprised ∼85% of thymocytes in control or single knockout mice, but only ∼45% of thymocytes in double knockout mice. (Fig. 5,A). Importantly, the absolute number of DP thymocytes in the double conditional knockout mice was only ∼10% of that in the controls (Fig. 5,B). Deletion of either Mcl-1 or Bcl-x in thymocytes significantly impaired the development of CD8⁺ SP but not CD4⁺ SP thymocytes (Fig. 5, B and C). In contrast, deletion of both Mcl-1 and Bcl-x almost completely abolished the CD4⁺ and CD8⁺ SP compartments as judged by their number and the frequency of TCRβ⁺ cells (Fig. 5, B and C). These results unequivocally demonstrate that DP thymocyte survival is promoted by both Mcl-1 and Bcl-x_L in vivo.

Previous data have shown that Mcl-1 elicited in response to cytokine signaling is required for the survival of naive T cells (22). Among T cell survival cytokines, IL-7 induces the highest level of Mcl-1 expression in naive T cells (22). One important question remaining to be addressed is whether Mcl-1 expression is regulated by TCR engagement and whether Mcl-1 plays a role in the survival of activated T cells. To address these issues, we stimulated wild-type T cells with anti-CD3 and examined Mcl-1 expression by Western blot. Cell lysate from IL-7-stimulated naive T cells was used as a positive control. Consistent with the previous result (22), Mcl-1 expression was increased in naive T cells stimulated with IL-7 (Fig. 6,A). Interestingly, anti-CD3 stimulation strongly up-regulated Mcl-1 expression in T cells, and the extent of anti-CD3-induced up-regulation was greater than that induced by IL-7 after 48-h stimulation (Fig. 6,A). To further examine the kinetics of anti-CD3 induced Mcl-1 expression in T cells, we stimulated splenic T cells with anti-CD3 or IL-7 for 6–12 h. Short duration stimulation with anti-CD3 or IL-7 also up-regulated Mcl-1 expression in T cells (Fig. 6,B). The extent of Mcl-1 up-regulation in T cells induced by IL-7 was higher than that induced by anti-CD3 (Fig. 6 B). These data suggest that other factors might be involved in anti-CD3-induced Mcl-1 up-regulation at a later time point. Alternatively, anti-CD3-induced Mcl-1 up-regulation may require longer time to reach its peak level.

To test whether Mcl-1 plays a role in the survival of activated T cells, we generated an inducible deletion system by crossing Mcl-1^f/f mice to the ER-Cre line (27). ER-Cre mice express the Cre recombinase fused to a mutant form of the estrogen ligand-binding domain inserted into the ROSA26 locus. The mutation in the estrogen ligand-binding domain makes it insensitive to endogenous estrogen, but capable of activation by the drug tamoxifen. We treated purified T cells from Mcl-1^f/fER-Cre and control mice (Mcl-1^f/f) in vitro with 0.2 μM 4-OHT, which did not demonstrate significant toxicity in vitro. A 2-day treatment resulted in significantly reduced recovery of live cells (p < 0.001) from Mcl- 1^f/fER-Cre T cell culture (Fig. 6,C). Importantly, T cell activation with anti-CD3/CD28 resulted in accumulation of live cells in the control cultures, but the cell numbers recovered from Mcl-1^f/fER-Cre cultures treated with 4-OHT remained significantly lower (p < 0.001) (Fig. 6,C). Similarly, IL-7 treatment did not rescue 4-OHT-treated Mcl-1^f/fER-Cre T cells from death (p < 0.05) (Fig. 6 C). As activated T cells strongly up-regulate Bcl-x_L expression, these results suggest that the up-regulation of Bcl-x_L is not sufficient to protect activated T cell from death, and that Mcl-1 plays a critical role in the survival of activated T cells.

Our studies demonstrate that Mcl-1 promotes T lymphocyte survival at multiple stages. In contrast to the anti-apoptotic roles of Bcl-2 and Bcl-x_L at specific T cell developmental stages, Mcl-1 appears to be a universal regulator of T cell survival. Genetic studies have established an anti-apoptotic role for Bcl-2 in naive T cells and Bcl-x_L in DP thymocytes (14, 15, 16, 17, 18, 19). Our data shown here, together with a previous report (22), have established that Mcl-1 promotes the survival of DN, DP, SP, naive, and activated T cells.

Our results confirm the previous finding that DN thymocyte development depends on Mcl-1 as deletion of Mcl-1 early in T cell development results in a blockade at the DN2/3 to DN4 transition. The major survival factors for T cells at this stage are IL-7R and pre-TCR signaling. IL-7R signaling is known to induce both Bcl-2 and Mcl-1 (22, 28, 32). Deletion of IL-7 or IL-7R components blocks thymocyte development at the DN2 stage, while deletion or mutation of pre-TCR components impairs developmental progression at the DN3 stage (33, 34, 35, 36). The phenotype of Mcl-1-deficient mice suggests that both IL-7R and pre-TCR might rely on Mcl-1 to promote DN cell survival.

We also show that the survival of SP thymocytes depends on Mcl-1. Little is known about the molecular regulation of SP thymocyte survival in vivo. The severe impairment of the SP thymocyte compartment in Mcl-1^f/fCD4-Cre mice suggests that the function of Mcl-1 cannot be compensated by Bcl-2 despite its expression in SP thymocytes. This suggests that perhaps these two proteins operate in distinct molecular pathways. Most likely the difference between the two anti-apoptotic molecules lies in their interaction partners (37). For example, only Mcl-1 but not Bcl-2 can interact with the proapoptotic molecule Bak. Conversely, only Bcl-2 but not Mcl-1 interacts with Bcl-2 antagonist of cell death (BAD).

Similar to DN and SP thymocytes, naive T cell express high levels of Bcl-2, but are also critically dependent on Mcl-1 for their survival. This was first demonstrated by Opferman et al. (22) using an inducible Cre driven by IFN-response elements. However, this system has a caveat in that it requires induction of high levels of IFNα/β that impact T cell survival (38). We used the ER-Cre inducible recombinase to show that naive T cells require Mcl-1 for their survival. Naive T cell survival requires IL-7 and TCR/MHC-peptide contact (39). As both Mcl-1 and Bcl-2 can be up-regulated by IL-7 signaling while Bcl-2 expression is down-regulated upon TCR engagement, our data suggest that TCR/MHC-peptide engagement may mediate its survival signal through Mcl-1. It is likely, however, that Mcl-1 and Bcl-2 operate in separate molecular pathways to ensure the survival of the cell, because loss of either protein cannot be compensated by the remaining one.

An interesting finding from our study is the anti-apoptotic role of Mcl-1 in DP thymocytes and activated T cells, in which the expression of Mcl-1 parallels that of Bcl-x_L but not Bcl-2. Although DP thymocytes lacking either Mcl-1 or Bcl-x display some degree of survival defects both in vitro and in vivo, the deletion of both Mcl-1 and Bcl-x results in a drastic reduction of the DP thymocyte compartment and a nearly complete block of the development of SP thymocytes. Thus, Mcl-1 and Bcl-xL can largely compensate for each other in the DP compartment. In contrast, activated T cells from Mcl-1^f/f ER-Cre mice undergo apoptosis, suggesting that Bcl-x_L up-regulation is not sufficient to prevent cell death induced by Mcl-1 deletion. In contrast, Bcl-x_L is dispensable for the survival of activated T cells (19). As Bcl-x_L functions similarly to Mcl-1 in terms of their interaction with the proapoptotic protein Bak (37), the death of activated T cells lacking Mcl-1 may reflect a dose-dependent effect. Alternatively, Mcl-1 may have a unique role that cannot be compensated by Bcl-x_L. Further studies are required to establish the in vivo mechanisms by which Mcl-1 exerts its anti-apoptotic function throughout T cell development.

We thank Claire Gordy for critically reading the manuscript.

The authors have no financial conflict of interest.