The Tuberous Sclerosis Complex–Mammalian Target of Rapamycin Pathway Maintains the Quiescence and Survival of Naive T Cells

Lymphocytes are generally referred to as “resting” cells prior to antigenic stimulation. However, these cells are active at molecular levels, as evidenced by constitutive phosphorylation of the TCR ζ-chain in resting T cells (1). Requirements for engagement of Ag receptors in T cell survival are inferred by the need for BCRs or TCRs for their respective survivals (2, 3). Moreover, accumulating data also support the view that survival of naive T cells requires the positive-selecting MHC (4, 5). In addition to antigenic stimulation, T cells undergo homeostatic proliferation in a lymphopenic environment (6). In the absence of Ag and lymphopenia, self-MHC ligands do not lead to overt activation of T cells. Thus, T cells remain largely quiescent in the absence of antigenic or lymphopenic stimuli. These observations raise two interesting questions. First, is there an active mechanism that maintains the quiescence of naive T cells? Second, does the mechanism that maintains the quiescence of T cells also control their survival?

The tuberous sclerosis complex (TSC)–mammalian target of rapamycin (mTOR) pathway has emerged as a central regulator for cellular metabolism (7–11). More recent studies have revealed two functionally distinct complexes with different components, TORC1 and TORC2 (12–14). Among the different components, TOR is associated with Raptor and others to form TORC1 (15). Because rapamycin–FKBP12 selectively binds to TORC1 (16, 17), it is specifically inhibited by rapamycin (12), although prolonged exposure to rapamycin may also affect TORC2 (18). In contrast, the Rictor–TOR complex formed the core of TORC2 (12). Although the TSC negatively regulates TORC1 function, recent studies suggest that defects in the TSC result in reduced TORC2 function, either directly or indirectly (14, 19, 20).

A critical role for the mTOR pathway in T cell activation was deduced from the fact that rapamycin, which specifically targets mTOR, has been used as an immunosuppressant in transplantation (21). Additionally, a role for mTOR in lymphocyte homing has also been reported (22). More recent studies have indicated that the mTOR pathway regulates generation of effector T cells and regulatory T cells (23, 24). Surprisingly, in vivo administration of rapamycin potently induced memory T cells in the presence of Ags (25, 26) and rejuvenated the aging hematopoietic stem cells (27). Because all reports on mTOR and T cells focus on T cell responses to Ag, it is of interest to establish the role of this pathway in the quiescence and survival of naive T cells in the absence of antigenic stimulation. In this study, we report that targeted mutation of Tsc1, an important regulator of mTOR, leads to loss of quiescence and cellularity in T cells, particularly in the CD8 subset. Interestingly, loss of quiescence and cellularity are mediated by distinct mechanisms.

Tsc1^fl/fl mice (28) were provided by Dr. David J. Kwiatkowski and backcrossed for six generations onto a C57BL/6 background in the animal facility at University of Michigan. CD4-Cre (29) mice were from Taconic (Hudson, NY), and Lck-Cre (30) mice were from the The Jackson Laboratory. Mice were housed under specific pathogen-free conditions. All studies involving mice have been approved by University Laboratory Animal Use and Care Committee at the University of Michigan.

The Abs and annexin V used in the study were obtained from eBioscience and BD Biosciences unless otherwise noted. Alexa Fluor 488-conjugated p-S6 Abs were obtained from Cell Signaling Technology. Intracellular staining of p-S6 was carried out as described (11, 27). Cell apoptosis and death were analyzed by staining cells with FITC- or allophycocyanin-conjugated annexin V in conjunction with 7-aminoactinomycin D (BD Pharmingen) according to the manufacturer’s protocol.

Mice were treated with either vehicle (5% polyethylene glycol 400, 5% Tween 80 in PBS) or rapamycin (4 mg/kg) every other day for 2–7 wk or as specified in the text. The splenocytes and thymocytes were analyzed at the specified time points.

Thymocytes were sorted into four subsets (CD4⁺CD8⁺, CD4⁻CD8⁺, CD4⁺CD8⁻, and CD4⁻CD8⁻) by FACSAria. The efficacy of Tsc1 deletion in genomic DNA was quantitated by real-time PCR. Genomic DNA was isolated using a PicoPure DNA extraction kit (Arcturus, Mountain View, CA), and quantitative PCR was performed with previously described TSC1 primers F4536 and F4830 using an ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA) (31).

In vivo BrdU labelings were performed according BD Pharmingen’s suggested protocol. Briefly, 8-wk-old Cre⁺ and Cre⁻ Tsc1^fl/fl littermates were i.p. injected with 200 μl 10 mg/ml BrdU in PBS followed by feeding the mice with drinking water containing 1 mg/ml BrdU for 24 h before mice were euthanized. Single-cell suspensions were prepared from thymus and spleen. Cells were stained with anti-CD4 and anti-CD8 Abs from eBioscience (San Diego, CA). BrdU⁺ populations were stained using FITC BrdU flow kits from BD Pharmingen according to the manufacturer’s recommended protocol. FACS analyses were preformed on a BD Biosciences LSRII.

To isolate lymphocytes, lungs were first perfused by injecting 10 ml PBS through the right ventricle of the heart with dissected hepatic vein and excised. The lung tissues were then cut into 1- to 2-mm³ pieces in 1× HBSS and incubated in 1 mg/ml collagenase II (Invitrogen, Carlsbad, CA) in 1× HBSS for 1 h at 37°C. The digested tissues were minced with frosted microscope slides and pass through a 70-μm cell strainer. Cells were then washed twice with staining buffer (1× HBSS with 2% FBS and 0.04% NaN₃) before staining with Abs. Anti-CD45 Abs were included in the Ab cocktails. Cells were gated on CD45⁺ population when analyzed by flow cytometry.

Peyer’s patches were removed and minced with glass slides to get single cell suspension for further analysis. Intraepithelial lymphocytes (IEL) were isolated according to protocols described in Lycke and Lefrançois (32). Briefly, the small intestines were cut open horizontally and rinsed at least four times in ice-cold CMF solution (1× HBSS, Ca²⁺- and Mg²⁺-free, 10 mM HEPES, 25 mM sodium bicarbonate [pH 7.2], 2% FBS). Cleaned small intestines were cut into 0.5-cm pieces with scissors. IEL were extracted by rotating at 37°C with 20 ml CMF solution with 5 mM EDTA for 20 min. The cells in the supernatants were collected after the tissues had settled down. The extraction procedure was repeated one more time. The supernatants were combined and cells were washed twice with RPMI 1640 supplemented with 10% FBS, 100 U/ml penicillin, 100 U/ml streptomycin, 2 mM l-glutamine, and 0.0005% 2-ME and resuspended in 8 ml 44% Percoll (GE Healthcare) in PBS solution, which were overloaded onto 5 ml 67% Percoll PBS solution and then centrifuged at 600 × g for 20 min at room temperature with low acceleration and deceleration settings. IEL were collected by carefully harvesting 1–2 cm of the 44% and 67% Percoll interface.

A Student t test was used to determine the statistical significance of the differences. All tests are two-tailed. Significance was defined as follows: *0.01 < p ≤ 0.05, **0.001 < p ≤ 0.01, ***p ≤ 0.001.

To study the function of Tsc1in T cells, we backcrossed the floxed Tsc1 allele into the C57BL/6 background for six generations and induced deletion by crossing the Tsc1^fl/fl mice (31) to the Tsc1^fl/fl;Lck-Cre^+/− mice or Tsc1^fl/fl;CD4-Cre^+/− mice that express the recombinase Cre under the control of either proximal Lck (30) or CD4 promoter (29). The Cre⁺ and Cre⁻ littermates were used for the study.

To verify the efficacy of gene deletion, we sorted the thymic CD4⁻CD8⁻, CD4⁺CD8⁺, CD3⁺CD4⁺CD8⁻, and CD3⁺CD4⁻CD8⁺ subsets from Lck-Cre⁺ and Lck-Cre^- Tsc1^fl/fl littermates. Real-time PCR using genomic DNA as templates was used to determine the relative amounts of deleted and undeleted DNA forms. As shown in Fig. 1A, the efficacy of deletion reached ∼90% among CD4⁺CD8⁺, CD3⁺CD4⁺CD8⁻, and CD3⁺CD4⁻CD8⁺ subsets. Because the Lck promoter is activated at DN3 (29), little deletion is observed within the CD4⁻CD8⁻ subset (Fig. 1A). Consistent with the functional role of Tsc1 in cell size regulation, the Cre⁺ thymocytes were significantly larger than the Cre⁻ thymocytes (Fig. 1B). A mild but statistically significant decrease in the percentage of the CD3⁺CD8⁺CD4⁻ and CD3⁺CD8⁻CD4⁺ thymocytes was observed in Cre⁺ mice (Fig. 1C).

In the spleen, the percentage of CD4 and CD8 T cells was further reduced in Cre⁺ mice (Fig. 2A). The percentage of CD4 T cells was reduced by ∼2-fold, whereas those of the CD8 T cells was reduced by ∼2- to 5-fold (Fig. 2B). As a result, an increase of the CD4/CD8 ratio was observed after deletion of Tsc1 (Fig. 2B). The cell loss can be observed in both young (2-wk-old) and adult mice (Fig. 2B). Corresponding to the dramatic reduction of CD8 percentage of T cells, the numbers of CD8 T cells were also significantly reduced at all time points tested (Fig. 2C). Owing to larger variations in cell counting, the reduction of CD4 T cells was significant only in the 2-wk-old mice (Fig. 2C). In addition to the change of T cell numbers, the remaining T cells were more active as judged by their cell surface markers. As shown in Fig. 2D and 2E, an overall increase of CD44^hi and CD62L^lo T cells was observed in the Tsc1-deficient mice. Similar changes in the cell number and phenotypes were observed in lymph nodes (data not shown).

To substantiate our basic observation, we also used CD4-Cre to delete the Tsc1 gene and carried out a parallel analysis. The data, presented in Supplemental Fig. 1, support the conclusion that deletion of the Tsc1 gene in T cell lineages caused loss of both quiescence and cellularity in mature T cells. One exception is that the central memory CD8 T cells, which are CD44^hi CD62L^hi, were found to be increased in Lck-Cre⁺ mice (Fig. 2E, middle lower panel) but were decreased in CD4-Cre⁺ mice with Tsc1 deletion (Supplemental Fig. 1E, middle lower panel). This difference remains to be explained. Because CD4-Cre is known to be more efficient in gene deletion than Lck-Cre, one potential explanation is that heterozygous deletion of Tsc1 by Lck-Cre results in partial activation that leads to production of central memory, whereas more complete deletion by CD4-Cre results in a higher level of T cell activation, leading to a transition to effector phenotype.

We measured the percentage of proliferating T cells by BrdU incorporation. As shown in Fig. 3A, whereas deletion of Tsc1 showed a slight trend of increased T cell proliferation in the thymus, this is statistically insignificant. In contrast, the peripheral CD4 and CD8 T cells from Cre⁺ mice had a significantly higher percentage of BrdU⁺ cells. Therefore, deletion of Tsc1 increased proliferation of T cells in the periphery (Fig. 3B, 3C).

To understand the mechanism of the loss of Tsc1^−/− T cells, we stained the thymocytes (Fig. 4A) and the spleen cells (Fig. 4B) with annexin V and 7-aminoactinomycin D, in conjunction with CD4 and CD8. Representative FACS profiles are shown in the left panels of Fig. 4A and 4B while the summary data involving five to eight mice per group are shown in the right panels. As shown in Fig. 4A and 4B, significant increases of the apoptotic T cells were observed among the CD8 T cells. Additionally, Tsc1 deletion caused loss of mitochondria potential in both CD4 and CD8 T cells, an early indicator of apoptosis (Fig. 4C). The lack of increase in apoptosis of CD4 T cells may be explained by efficient removal of apoptotic cells in vivo (33). To confirm this prediction, we cultured spleen cells in vitro for 6 h. As shown in Fig. 4D, Tsc1 deletion doubled the percentage of apoptotic CD4 and CD8 T cells in the culture. Lack of increase in apoptotic Tsc1^−/− CD4 T cells by annexin V staining may therefore be attributable to clearance in vivo.

It has been suggested that activated effector memory T cells preferentially home to mucosal tissues such as the lung and the gut (34). To rule out the possibility that the reduced cellularity in spleen was due to increased homing to mucosal tissues, we examined the CD4 and CD8 T cells from the lung and Peyer’s patches and IEL from small intestine. As shown in Fig. 5A and 5B, the percentages of CD4 and CD8 T cells were comparable in lung and Peyer’s patches between Cre⁺ and Cre⁻ mice, whereas the T cells in spleen were decreased in Cre⁺ mice. Histological examination of lung (data not shown) and intestine (Supplemental Fig. 2A) showed no increased lymphocyte infiltration. Moreover, the frequency of CD4 and CD8 T cells among IEL was similar in number between Cre⁺ and Cre⁻ mice (Supplemental Fig. 2B). Furthermore, in the OT-1 transgenic mice, no increase of transgenic T cells was attributable to deficiency of TSC1 (Supplemental Fig. 2C). Therefore, loss of mature T cells in the spleen and lymph node was not due to their migration in the mucosal tissue. Perhaps because activated T cells have a tendency to migrate to target tissues, no reduction in T cell frequencies was observed in lung and Peyer’s patches.

Thus, deletion of Tsc1 specifically in T cell lineages caused loss of quiescence and cellularity in mature T cells by increasing the proliferation and apoptosis.

An interesting issue is whether the impact of Tsc1 deletion on naive T cells depends on antigenic stimulation. This issue cannot be resolved in mice with polyclonal T cells, as many of them will be stimulated by environmental Ags. We therefore bred the floxed Tsc1 allele into transgenic mice expressing OT-1 TCR, which is specific for OVA (35). By allelic exclusion, the T cells expressing high levels of transgenic α-chain are likely to be specific for the OVA presented by MHC class I and are thus less likely to recognize environmental Ags. As shown in Fig. 6A, although the mice were not challenged with the OVA, a 3-fold reduction of Vα2^hiCD8⁺ cells was observed in the spleen of the Cre⁺ mice. This corresponded to an increased apoptosis among the transgenic T cells (Fig. 6B). Additionally, the CD8 T cells are more activated as demonstrated by a 3-fold loss in naive T cells and a 5-fold increase in CD44^hiCD62L^lo cells (Fig. 6C). Because the loss of quiescence was more robust than what was observed in mice with polyclonal T cell repertoire, the data demonstrated that the spontaneous activation was not induced by Ags, even though a small fraction of the Vα2^hi cells may still have rearranged endogenous Tcra locus.

Tsc1 forms a complex with Tsc2 to regulate mTOR function (7–11). To determine whether increased mTOR activation in T cells is responsible for losses in both quiescence and cellularity, we treated the Tsc1^fl/fl;Lck-Cre⁺ mice with rapamycin or vehicle. As shown in Fig. 7A, 2 wk rapamycin treatment (4 mg/kg, once every other day) effectively inhibited mTOR activation of the Tsc1^fl/fl;Lck-Cre⁺ T cells, based on intracellular staining of p-S6. Additionally, the treatment restored the cell size of both CD4 and CD8 T cells (Fig. 7B). Furthermore, rapamycin treatment increased the percentage of naive T cells (Fig. 7C). Surprisingly, rapamycin treatment failed to restore cellularity of mature T cells (Fig. 7D). Neither extending the treatment to 7 wk nor decreasing the dose of rapamycin to 75 μg/kg had any appreciable effect on T cell cellularity (Supplemental Fig. 3). Correspondingly, the percentage of apoptotic cells was also unaffected (Supplemental Fig. 3). To confirm that the loss of cellularity was not rescued by rapamycin, we cultured wild-type and mutant T cells in the presence of a wide range of concentrations of rapamycin. As shown in Fig. 7E, rapamycin treatment did not reduce apoptosis of spleen T cells in vitro. The differential effect of rapamycin on apoptosis and cellular activation indicated that loss of quiescence and cellularity may be mediated by different mechanisms.

Taken together, we showed both spontaneous activation and loss of cellularity of naive T cells after activation of mTOR by Tsc1 deletion. Although the Tsc1 gene controls both survival and quiescence, the downstream effector mechanisms for the two functions may be distinct. On one hand, because rapamycin effectively inhibited activation but not loss of cellularity, it is likely that loss of cellularity was not due to loss of quiescence. On the other hand, although loss of T cells may create lymphopenia that leads to homeostatic proliferation and acquisition of memory markers, the fact that Tsc1 deletion increased T cells with effector memory markers suggests that homeostatic proliferation alone, which primarily resulted in central memory T cells (6), cannot fully account for the spontaneous T cell activation .

A recent report demonstrated that FoxO1 maintains survival and trafficking of naive T cells by regulating expression of Sell, Klf2, and Il7r (36). Because none of the genes are repressed by Tsc1 deletion (Supplemental Fig. 4), it is likely that the FoxO and Tsc mediate nonoverlapping functions, both necessary for survival and quiescence of naive T cells. Additional studies are needed to elucidate how Tsc is stimulated to maintain naive T cells. One of our groups has shown that the Tsc is activated by AMP-activated protein kinase (8). Because AMP-activated protein kinase is known to be activated by TCR signaling (37), it would be of interest to test whether such activation may play a role in T cell quiescence and survival.

Interestingly, deletion of Tsc1, much like that of FoxO1 (36), reduced the number of CD8 but not CD4 T cells. The functional dichotomy of the two subsets remains to be explained. Because the TCR–MHC interaction is necessary for survival of CD8 (2, 4, 38) but apparently not CD4 T cells (39), it is intriguing that both FoxO1 and Tsc1 play major roles in mediating the protection conferred by TCR–MHC interaction.

Regardless of the mechanism, our data provide a direct link between T cell survival and mTOR activation. The molecular link helps to explain how T cells survive under physiological conditions. Moreover, because inflammatory cytokines are potent inducers of mTOR activation (27, 40), our results presented in this study may provide new insights on how inflammation associates with loss of T cells during infections (41, 42).

We thank Darla Kroft for editorial assistance.

The authors have no financial conflicts of interest.