Signaling via the Akt/mammalian target of rapamycin pathway influences CD4+ T cell differentiation; low levels favor regulatory T cell induction and high levels favor Th induction. Although the lipid phosphatase phosphatase and tensin homolog (PTEN) suppresses Akt activity, the control of PTEN activity is poorly studied in T cells. In this study, we identify multiple mechanisms that regulate PTEN expression. During Th induction, PTEN function is suppressed via lower mRNA levels, lower protein levels, and an increase in C-terminal phosphorylation. Conversely, during regulatory T cell induction, PTEN function is maintained through the stabilization of PTEN mRNA transcription and sustained protein levels. We demonstrate that differential Akt/mammalian target of rapamycin signaling regulates PTEN transcription via the FoxO1 transcription factor. A mathematical model that includes multiple modes of PTEN regulation recapitulates our experimental findings and demonstrates how several feedback loops determine differentiation outcomes. Collectively, this work provides novel mechanistic insights into how differential regulation of PTEN controls alternate CD4+ T cell fate outcomes.
Ligand binding to the TCR induces a cascade of signaling events that culminates in gene transcription, T cell proliferation, and differentiation. We (1) and others (2, 3) have demonstrated that Foxp3-expressing regulatory T cells (Tregs) are induced when naive T cells are exposed to low Ag doses, and it has been shown that Tregs expanded following stimulation with low-dose Ag prevent autoimmune diabetes in vivo (4, 5). Treg induction is inversely correlated with the degree of signaling via the Akt/mammalian target of rapamycin (mTOR) pathway (1, 6, 7).
We developed a mathematical model to determine how TCR signal strength contributes to Treg induction (8). Regulation of the phosphatase and tensin homolog (PTEN) protein emerged as a critical control point in Treg induction (8). Expression and function of PTEN are tightly regulated via transcriptional control and posttranslational modifications (PTMs) (9, 10). Naive T cells express high levels of PTEN and following T cell activation PTEN protein disappears (11), but the mechanism for this downregulation is not known. Tregs are known to express higher levels of PTEN (12) and this changes the signaling pathways triggered following IL-2 activation (13).
In this study, we identify multiple modes of PTEN regulation that affect CD4+ T cell differentiation. Th induction is characterized by reduced PTEN transcription, degradation of PTEN protein, and higher C-terminal phosphorylation. However, in Treg induction, PTEN mRNA and protein levels are maintained. Additionally, we identify a loop connecting Akt/mTOR signaling to the regulation of PTEN via FoxO1 as a novel mechanism for regulating CD4+ T cell differentiation. Taken together, these findings define how TCR signal strength differentially tunes PTEN function to drive Th versus Treg differentiation.
Materials and Methods
C57BL/6 mice were obtained from The Jackson Laboratory. All mice were housed in a specific pathogen-free facility at the University of Pittsburgh and treated under Institutional Animal Care and Use Committee–approved guidelines in accordance with approved protocols.
CD4+ T cell isolation and activation
CD4+ T cells were isolated from C57BL/6 spleens using a CD4+ negative selection kit (Miltenyi Biotec). In some experiments CD25+ T cells were removed using CD25 microbeads. T cell activation assays were performed with low-dose (0.25 μg/ml) or high-dose (1 μg/ml) plate-bound anti-CD3 mAb in the presence of soluble anti-CD28 mAb (1 μg/ml). For inhibition studies, T cells were treated for 1 h with the following drugs prior to stimulation: caspase inhibitor (ZVAD, 80 μM), PTEN inhibitor (SF1670, 10 μM), and Akt inhibitor (Akti1/2, 10 μM).
Activated CD4+ T cells were stained with the following mAbs: anti–CD3-allophycocyanin-eFluor 780, anti–CD4-allophycocyanin, anti–CD25-PE, anti–Foxp3-Pacific Blue (eBioscience), and anti–pS6 (Ser235)-FITC (Cell Signaling Technology) using buffers from eBioscience. The stained cells were analyzed on an LSR II flow cytometer and data were analyzed with the FlowJo software package.
Western blotting was performed using the following Abs from Cell Signaling Technology: β-actin, PTEN (138G6), phospho-PTEN (Ser380/Thr382/383), FoxO1 (C29H4), phospho-FoxO1 (Thr24), and ubiquitin.
PTEN RNA levels
Total RNA was isolated from CD4+ T cells the Aurum total RNA mini kit (Bio-Rad). Quantitative RT-PCR (qPCR) was performed on a StepOnePlus real-time PCR system using SYBR Green (Applied Biosystems). The primer pair used was 5′-ACACCGCCAAATTTAACTGC-3′ and 5′-TACACCAGTCCGTCCCTTTC-3′.
Chromatin immunoprecipitation of FoxO1
Chromatin immunoprecipitation (ChIP) for FoxO1 was performed using the ChIP-IT high sensitivity kit (Active Motif) and an anti-FoxO1 Ab (Cell Signaling Technology). qPCR was used to quantify the amount of the PTEN 5 kb upstream element (primer pair used: 5′-CCTTGCATCAGCCATTAAAGAAA-3′ and 5′-CTGCTGTACCAGGCTATATGTG-3′) associated with FoxO1.
FoxO1 small interfering RNA knockdown
A murine FoxO1 small interfering RNA (siRNA) kit (Origene) was used to knock down FoxO1 expression. The siRNAs were introduced into isolated CD4+ T cells using a standard protocol (Lonza Nucleofector kit for mouse T cells). Western blot analysis for FoxO1, PTEN, and actin was performed after 48 h of incubation.
Nuclear/cytoplasmic isolation and localization
Nuclear and cytoplasmic fractions of activated T cells were generated using the NE-PER nuclear and cytoplasmic kit (Thermo Scientific) following the manufacturer’s instructions.
Computational modeling of T cell development
Simulations of PTEN levels in cells destined to become Th cells or Tregs were performed as described (8). A minimized version was specified as a Boolean model using the logical rules shown in Supplemental Fig. 2. For the simulations performed in this study, the logical rules were translated into the BioNetGen rule-based modeling language using an automated tool called Boolean2BNGL included in the BioNetGen package (14). The simulations, performed according to the general asynchronous update scheme (15), used a modified version of Gillespie’s (16) direct stochastic simulation algorithm.
All statistical calculations were performed in the GraphPad Prism 6 software package. Except where indicated, all analyses used the two-way ANOVA analysis with a Bonferroni postanalysis correction. Some analyses used the one-way ANOVA or Student t test. A p value <0.05 was considered significant.
Results and Discussion
PTEN protein levels are suppressed during Th differentiation but maintained during Treg differentiation
Our computational model of Th/Treg differentiation (8) predicted that, in developing Th cells, PTEN rapidly decreased and remained absent, whereas PTEN was only transiently downmodulated in cells destined to become Tregs (Fig. 1A). To test this prediction, we performed T cell activation assays with low and high doses of anti-CD3 mAb, which drive Treg and Th differentiation, respectively (Supplemental Fig. 1A, 1B) (1, 4, 17). Western blot analysis of PTEN in activated CD4+ T cells confirmed the model predictions, and PTEN levels were markedly and persistently reduced with high-dose stimulation, but only transiently reduced with low-dose stimulation (Fig. 1B, 1C). Similar results were obtained with CD25-depleted CD4+ T cells (Supplemental Fig. 1C, 1D)
To determine whether PTEN activity regulated CD4+ T cell differentiation, activation assays were performed in the presence of a PTEN inhibitor. PTEN inhibition of T cells stimulated with high-dose anti-CD3 resulted in the development of Th cells and enhancement of Akt/mTOR signaling as expected (Supplemental Fig. 1E). However, under low-dose stimulation, PTEN inhibition enhanced Th (Fig. 1D) and suppressed Treg (Fig. 1E) induction. These results are consistent with recent studies using PTEN short hairpin RNA, showing decreased Treg induction and increased pS6 in cells with PTEN knockdown (13, 18).
PTEN is ubiquitinated under conditions of Th but not Treg induction
TCR signaling perturbs PTEN protein levels (Fig. 1) (13, 18), but the molecular mechanisms that regulate PTEN protein levels in T cells are unexplored. The rapidity of PTEN degradation in Th and Treg conditions suggests that PTMs contribute to the regulation of PTEN turnover. In other cell types, polyubiquitination of PTEN by the E3 ubiquitin ligase Nedd4 leads to proteosomal degradation (19). In Treg conditions, PTEN was not ubiquitinated (Fig. 2A), whereas in Th conditions, PTEN was extensively ubiquitinated (Fig. 2B).
During Treg induction, PTEN is degraded soon after T cell activation, yet PTEN is not ubiquitinated (Figs. 1, 2A), suggesting a different mechanism for the early degradation of PTEN. Multiple caspases, including caspase-3 and caspase-9, are known to cleave PTEN (20), and caspases function in TCR signaling (21, 22). Pan-caspase inhibition with the inhibitor ZVAD blocked degradation of PTEN under Treg conditions (Fig. 2C), demonstrating that caspase-mediated degradation of PTEN occurs during CD4+ T cell activation.
TCR activation promotes PTEN phosphorylation
Constitutive phosphorylation of the C terminus of PTEN suppresses its enzymatic activity (23, 24), and casein kinase 2 has been shown to phosphorylate these sites in T cells (25, 26). We explored the phosphorylation state of PTEN during T cell differentiation. PTEN was phosphorylated at a cluster of sites, including Ser380, Thr382, and Thr383 during both Treg and Th induction (Fig. 2D), but relative levels of PTEN phosphorylation were significantly higher during Th induction (Fig. 2E). To our knowledge, these results are the first demonstration in T cells that levels of PTEN and enzymatic activity are controlled by multiple PTMs, which are differentially regulated by TCR signal strength.
PTEN transcription is regulated by FoxO1
High levels of TCR signaling have been correlated with a reduction in PTEN mRNA (13, 18). Therefore, we followed the expression of PTEN mRNA during Th and Treg induction. PTEN mRNA decreased in both Treg and Th activation conditions (Fig. 3A); however, whereas ∼20% of PTEN mRNA remained under Treg conditions, little PTEN mRNA was detected under Th conditions.
The mechanisms used to maintain PTEN mRNA levels under Treg induction conditions are not known. In our T cell differentiation model (8), the recovery of PTEN in Treg conditions arises from a link coupling Foxp3 to PTEN. However, recent data implicate the transcription factor FoxO1 in the induction of Treg, and phosphorylation of FoxO1 by Akt was established as a mechanism by which Akt/mTOR signaling inhibits Treg differentiation (27).
Previous ChIP sequencing data identified a FoxO1 binding site 5 kb upstream of the PTEN promoter (28), and we hypothesized that FoxO1 might be responsible for maintenance of PTEN transcription in Treg conditions. We performed ChIP assays to determine whether FoxO1 bound to this region. In both resting CD4+ T cells and under Treg conditions, FoxO1 bound the 5 kb upstream element, but it did not exhibit significant binding under Th conditions (Fig. 3B). siRNA knockdown of FoxO1 resulted in a concomitant decrease in PTEN protein levels, consistent with a positive role for FoxO1 in the regulation of PTEN levels (Fig. 3C).
Phosphorylation and nuclear localization of FoxO1 depend on TCR signal strength
Previous work demonstrated that unphosphorylated FoxO1 is localized in the nucleus and that Akt-mediated phosphorylation of Thr24 drives FoxO1 into the cytoplasm (29). We found that FoxO1 remained unphosphorylated under Treg conditions, but it was rapidly phosphorylated under Th conditions (Fig. 3D). As expected, FoxO1 remained in the nucleus in resting CD4+ T cells and under Treg conditions, but it was largely in the cytoplasm following high-dose stimulation (Fig. 3E), consistent with our observations on regulation of PTEN transcription.
Logical model that includes FoxO1 shows how positive and negative feedback loops determine T cell fate
Using components and interactions from our previous model of T cell differentiation (8), we developed a more limited model focusing on the regulation of PTEN, Akt, and FoxO1 (Fig. 3F, left). Simulations exhibited rapid and sustained loss of PTEN (Fig. 3G) at high Ag dose, as well as transient and incomplete loss of PTEN at low dose, in agreement with experimental findings. Under both conditions there was a rapid initial increase in Akt activity (Fig. 3G), sustained by positive feedback involving the loss of nuclear FoxO1 (Fig. 3F, bottom right). At high dose, the activation of casein kinase 2 blocked PTEN activity and sustained Akt activation. At low dose, PTEN activity rebounded as negative feedback through mTOR complex 2 (mTORC2)–decreased Akt activity (Fig. 3F, top right), increasing nuclear FoxO1 and, consequently, PTEN transcription. PTEN activity then further suppressed Akt, permitting full restoration of PTEN expression. Insertion of our reduced model back into our previous model of T cell differentiation resulted in sustained induction of Foxp3 in ∼70% of cells at low dose and only transient induction of Foxp3 at high dose (Fig. 3G), in agreement with results from the previous model and experiments (8).
Transient suppression of PTEN at low dose arises from the inclusion of two regulatory mechanisms in the model: the requirement for MEK1 to activate PTEN (30), and inhibition of mTORC2 by Akt (31). Prior to MEK1 activation by TCR (32), PTEN is inactive, allowing the transient induction of Akt, even at low dose, which may be necessary for T cell activation and proliferation regardless of differentiation outcome. Subsequent activation of PTEN by MEK1 then inhibits Akt in concert with the negative feedback provided by indirect Akt inhibition of mTORC2 (31).
Akt regulates FoxO1 nuclear localization and binding to PTEN promoter
To further elucidate the role of Akt in the regulation of FoxO1, we treated CD4+ T cells with an Akt inhibitor (Akti1/2) simultaneously with activation by a high dose of stimulation. The efficacy of the Akt inhibitor was confirmed by a reduction in pS6 expression (Fig. 4A). Under high TCR stimulation alone, we observed sustained phosphorylation of FoxO1 (Fig. 4B) and cytoplasmic localization of FoxO1 at 12 h (Fig. 4C). We also observed loss of FoxO1 binding to the 5 kb element upstream of the PTEN promoter (Fig. 4D) and loss of PTEN mRNA expression (Fig. 4E). Addition of the Akt inhibitor, alternatively, resulted in only transient phosphorylation of FoxO1 (Fig. 4B), which remained in the nucleus at 12 h (Fig. 4C) and was accompanied by sustained binding of FoxO1 upstream of the PTEN promoter and significantly higher PTEN mRNA expression. These results show that Akt is an important regulator of PTEN expression, acting through the transcription factor FoxO1.
In conclusion, we have demonstrated that differential regulation of PTEN determines CD4+ T cell fate. In Th differentiation, stimulated by high TCR strength, multiple mechanisms suppress PTEN activity by lowering protein levels and enzymatic activity (Fig. 4F). Low TCR signal strength, resulting in Treg induction, also involves early downregulation of PTEN but requires that PTEN activity is re-established at longer times because inhibition of PTEN under these conditions reduces Treg and increases Th induction. Additionally, we identify a critical feedback loop driven by differential Akt/mTOR signaling that regulates PTEN transcription via the FoxO1 transcription factor. By combining biochemical and computational modeling approaches, our work defines how differential regulation of PTEN can produce alternate CD4+ T cell fate outcomes.
We thank Leonard Harris for help with the Boolean2BNGL package.
This work was supported by National Institutes of Health Grant P41 GM103712 and National Science Foundation Grant 0926181 (to J.R.F. and N.M.-Z.), National Institutes of Health Grants GM080398 and AI095730 (to L.P.K.), and by National Institutes of Health Postdoctoral Training Grant T32 AI089443 (to W.F.H.) and Predoctoral Training Grant T32 EB009403 (to R.P.S.).
The online version of the article contains supplemental material.
The authors have no financial conflicts of interest.