TGF-β Converts Apoptotic Stimuli into the Signal for Th9 Differentiation

Naturally arising regulatory T cells (nTregs) develop in the thymus and are characterized by constitutive expression of CD25 and a transcription factor FoxP3 (1–3). FoxP3 plays critical roles in the development, survival, and functions of nTregs (2, 4–6) as depicted by severe autoimmune disorders caused by mutation in the foxp3 gene both in humans and mice (7–9). nTregs represent 5–10% of the CD4⁺ T cell population in the periphery, and relative increase or decrease of regulatory T cells (Tregs) is often associated with immune regulation disorders (1). Thus, mechanisms of maintenance of the balance between nTregs and non-Tregs (conventional T cells) could play a significant role in the regulation of immunity against self-Ags and non–self-Ags.

We demonstrated previously that nTregs survive and expand when stimulated with immobilized anti-CD3 and anti-CD28 Abs (by coating onto plastic plates) with the added presence of IL-2, whereas non-Treg T cells undergo apoptosis (10). Unlike classical activation-induced cell death, this form of apoptosis was p53 dependent and requires engagement of CD28, and was hence named p53-induced CD28-dependent T cell apoptosis (PICA). Unlike conventional T cells, nTregs are resistant to PICA. When stimulated under the same conditions, Foxp3⁺ Tregs expanded more robustly than that seen with a more commonly used bead-based stimulation method, and they expanded >7000-fold within 10 d. The data suggested that PICA might have a role in immune regulation by controlling the balance between nTregs and conventional T cells. The data also provided a potential explanation for previous observations on p53-deficient mice that exhibit earlier onset and exacerbated disease state in experimental autoimmune arthritis and other autoimmune disease models (11–13).

To determine the mechanism by which nTregs withstand PICA, we analyzed the role of TGF-β. TGF-β is a pleiotropic cytokine that is involved in various T cell responses, including promotion of Foxp3⁺ inducible Treg (iTreg) induction and mediation of suppressive functions of Tregs, and is expressed by nTregs on the cell surface upon TCR activation (14–18). In this study, we demonstrate that TGF-β signaling is required for survival of nTregs against PICA and that TGF-β can render conventional T cells resistant to PICA without induction of Foxp3 expression. Strikingly, conventional T cells treated with TGF-β not only survived PICA, but differentiated to IL-9–producing T cells (T_H9) and addition of exogenous IL-6 convert conventional T cells into IL-17 producing T cells (T_H17). The data show TGF-β as a key determinant of the fate of T cells when they receive PICA-inducing stimuli.

C57BL/6 and CD4dnTgfbr2 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). All mice were maintained under specific pathogen-free conditions. All procedures were approved and monitored by Institutional Animal Care and Use Committee of Loyola University Chicago.

Fluorochrome-conjugated Abs specific for Foxp3 (FJK-16s) and IL-17A (ebio17B7) were from eBioscience (San Diego, CA). Anti-CD4 (GK1.5) and anti-IL-9 (RM9A4) were from BioLegend (San Diego, CA). Annexin V, 7-aminoactinomycin D, anti-CD25 (7D4), anti-Fas (Jo2), and anti-FasL (MFL3) were from BD Biosciences (San Jose, CA). Cell surface staining was performed on ice (30 min, unless stated otherwise) with appropriately conditioned Abs. For Foxp3 staining, cells were fixed and permeabilized using eBioscience FOXP3 Staining Buffer Set as described by the manufacturer’s protocol. For intracellular cytokine staining, cells were harvested then restimulated with 50 ng/ml PMA and 1 μM ionomycin in the presence of monensin for 4 h. Cells were then fixed and permeabilized for staining with anti–IL-17 or anti–IL-9 Abs. Data were collected by a FACS Canto flow cytometer (BD Biosciences) or an Accuri C6 flow cytometer (Accuri Cytometers, Ann Arbor, MI) and analyzed using FlowJo software (Tree Star, Ashland, OR).

Splenic CD4⁺ T cells were purified by depletion of non-CD4⁺ T cells by using the panning method. Cells were labeled with anti-CD8 (3-155) Ab, washed, and allowed to adhere to plate-bound goat anti-mouse Ig. After 30 min, nonadherent cells were collected. This crude fraction of CD4⁺ T cells was then labeled with fluorochrome-conjugated anti-CD4 and anti-CD25 Abs and sorted into CD4⁺CD25⁻ cells /CD4⁺CD25⁺ cells fractions with a FACS ARIA cell sorter (BD Biosciences). Sorted cells were rested overnight at 4°C before being used for each experiment.

For plate-bound anti-CD3/anti-CD28 Abs stimulation, sorted CD4⁺CD25⁻ or CD4⁺CD25⁺ (1.5 × 10⁵) T cells were placed into 5 ml culture medium in 60-mm dishes that had been precoated overnight at room temperature with 2 ml of anti-CD3 (2C11; BioLegend) and anti-CD28 (37-51; BioLegend) Abs (5 μg/ml each) in 0.1 M borate buffer (pH 8.5). The culture medium was RPMI 1640 medium supplemented with 10% FCS (Atlanta Biologicals), β-mercaptoethanol (50 μM), glutamine, sodium pyruvate (1 mM), and nonessential amino acids (Invitrogen Life Technologies, Grand Island, NY) in the presence of recombinant IL-2 (10 ng/ml). To block TGF-β signaling, 5 μg/ml anti–TGF-β1, 2, 3 Ab (1D11; R&D Systems, Minneapolis, MN) or 10 μM SB431542 (Sigma-Aldrich, St. Louis, MO) were added into culture medium. Recombinant human TGF-β (2.5 ng/ml; R&D Systems) was used for an active form of TGF-β. To block IL-4 signaling, 10% 11B11 hybridoma culture supernatant, which contains anti–IL-4, was added to culture medium.

Cells were lysed directly in SDS sample buffer (2% SDS, 125 mM DTT, 10% glycerol, 62.5 mM Tris-HCl, pH 6.8). Cell lysates were boiled for 10 min, and an equal amount (based on cell count) was loaded onto SDS PAGE gels (8–15%). After gel electrophoresis, separated proteins were blotted onto polyvinylidene difluoride membranes. The membranes were probed with the following Abs. Anti–phospho-Akt (Ser473), phospho-Erk1/2, phospho-FoxO1 (Ser256), FoxO1, phospho-FoxO3a (Ser253), FoxO3a, and Bim were from Cell Signaling Technology (Denvers, MA). Anti-Akt Ab was from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Erk1/2 Ab was from Millipore (Billerica, MA). Anti–β-actin was from Sigma-Aldrich. The membranes were further probed with anti-rabbit, anti-goat, or anti-mouse HRP-conjugated Abs (Cell Signaling Technology). Signals were detected with the ECL system (GE Healthcare, Piscataway, NJ). Band intensity of scanned data from films was quantified using Image J software (National Institutes of Health).

Statistical significance was determined with two-tailed Student t tests.

Previous reports have shown that TGF-β is involved in apoptosis (19–24) and cell survival (25–27). Because TGF-β is differently expressed by nTregs and other T cells (28), we hypothesized that resistance to PICA by Tregs is mediated in part by TGF-β. Our hypothesis predicted that inhibition of the TGF-β signaling pathway will abrogate PICA resistance by Tregs, whereas addition of exogenous TGF-β will increase the frequency of live cells that survive PICA. Thus, we cultured purified CD4⁺CD25⁻ T cells under PICA-inducing conditions in the presence or absence of exogenous TGF-β. After 3 d of culturing, we harvested cells and assessed their survival. As observed previously, cells that were stimulated by plate-bound anti-CD3/anti-CD28 Abs underwent apoptotic cell death detected by an increase of annexin V⁺ cells (Fig. 1A). When exogenous TGF-β was added to the culture, the frequency of apoptotic and dead cells decreased substantially. This change with TGF-β was due to expansion of the number of live cells and not due to a decrease of annexin V⁺ cell numbers (Fig. 1B). When CD4⁺CD25⁻ T cells were stimulated with plate-bound anti-CD3/anti-CD28 Abs, the final live cell number after 3 d was about the same as the starting sample (1.2-fold increase). In contrast, the annexin V⁻ cell number increased by 2.8-fold when CD4⁺CD25⁻ T cells were stimulated with plate-bound anti-CD3/anti-CD28 Abs in the presence of TGF-β. These data show that TGF-β renders CD4⁺CD25⁻ T cells resistant to PICA and allows them to expand.

It is well established that TGF-β can induce differentiation of naive CD4⁺ T cells into Foxp3⁺ iTregs (16, 17). Thus, the survival of CD4⁺CD25⁻ T cells observed with exogenous TGF-β may have been due to conversion of CD4⁺CD25⁻ T cells to Foxp3⁺ iTregs. To test this possibility, we stimulated sorted CD4⁺CD25⁻ T cells with plate-bound anti-CD3 plus either soluble or plate-bound anti-CD28 Abs with the culture medium conditioned for induction of iTregs (including IL-2) (29). After 3 d of stimulation, expression of Foxp3 by expanded cells was examined by flow cytometry. When stimulated by plate-bound anti-CD3 and soluble anti-CD28 Abs in the presence of TGF-β, a significant proportion of cells (37.3%) expressed Foxp3 (Fig. 1C). In contrast, only 5.3% of cells expanded with both the anti-CD3 and anti-CD28 Abs being plate-bound expressed Foxp3. The level of Foxp3⁺ cells from plate-bound anti–CD28-stimulated cells was comparable to those stimulated without TGF-β. The data show that resistance of CD4⁺CD25⁻ T cells against PICA by TGF-β is due to antiapoptotic responses of CD4⁺CD25⁻ T cells and is not caused by enhanced induction of iTregs.

The data presented above showed that TGF-β may also have a role in survival of nTregs against PICA because CD4⁺CD25⁺ nTregs, but not other T cell populations, express TGF-β on their cell surface (14). Thus, we determined whether TGF-β receptor signaling is required for survival of nTregs. To inhibit TGF-β receptor signaling in nTregs, we used a TGF-β superfamily type I receptor kinase inhibitor (SB431542) or TGF-β neutralizing Ab (anti–TGF-β–1, –2, and –3 Ab). Purified CD4⁺CD25⁺ nTregs were stimulated by plate-bound anti-CD3/anti-CD28 Abs in the presence of SB431542 or anti–TGF-β neutralizing Ab. After three days of stimulation, cells were harvested and analyzed by flow cytometry. CD4⁺CD25⁺ nTregs expanded ∼2-fold compared with the starting cell number. When SB431542 was added, cell growth was substantially blocked and the cell number decreased ∼5-fold. Similarly, when CD4⁺CD25⁺ Tregs were treated with anti–TGF-β Ab, live cell number decreased substantially compared with the starting number (Fig. 2A). Flow cytometric analysis showed that this decrease in cell numbers corresponds to an increase in annexin V⁺ apoptotic or dead cell frequency (Fig. 2B).

To confirm these results, we examined PICA resistance by CD4⁺CD25⁺ Tregs isolated from transgenic mice expressing a dominant-negative form of TGF-β receptor type II under the control of mouse CD4 promoter (CD4dnTgfbr2). These mice have a normal level of Foxp3⁺CD4⁺CD25⁺ nTregs (∼8 wk old), although TGF-β receptor signaling is substantially blocked in T cells (30). We isolated splenic CD4⁺CD25⁺ nTregs from CD4dnTgfbr2 mice or their wild type littermate control mice and stimulated them with plate-bound anti-CD3/anti-CD28 Abs in the presence of IL-2. After 3 d of culture, we harvested cells and assessed their survival (Fig. 2C). The cell number of wild type littermate nTregs increased from day 0, but numbers of CD4^dnTgfbr2 nTregs were less than 10% of the control and decreased compared with the starting sample cell number. As observed with the chemical inhibitor and blocking Ab, the frequency of annexin V⁺ cells was ∼2-fold higher in CD4^dnTgfbr2 T cell culture compared with that of the littermate control (Fig. 2D). Together, the data show that TGF-β is required for survival of nTregs against PICA. Because we did not add exogenous TGF-β to the culture, the data strongly suggest that CD4⁺CD25⁺ Tregs provide TGF-β in an autocrine manner to maintain nTregs resistance against PICA.

Previously, we demonstrated that PICA requires expression of Bim and Fas/FasL, which are known molecules for apoptosis by T cells (10). Because TGF-β rescued CD4⁺CD25⁻ T cells from PICA, we determined whether the addition of exogenous TGF-β reduces expression of Bim or Fas ligand, or both, by CD4⁺CD25⁻ T cells when stimulated by plate-bound anti-CD3/anti-CD28 Abs (Fig. 3A). Unstimulated CD4⁺CD25⁻ T cells expressed two forms (L and EL isoforms) of Bim at a low level. When stimulated with anti-CD3/anti-CD28 Abs, CD4⁺CD25⁻ T cells expressed both forms of Bim at a level clearly higher than that seen in unstimulated T cells. Stimulated and TGF-β–treated T cells, however, showed a markedly reduced level of Bim protein expression, even lower than that in unstimulated T cells (EL isoform). In contrast to Bim expression, TGF-β treatment caused a mild reduction in expression of FasL by CD4⁺CD25⁻ T cells (Fig. 3B), whereas expression of Fas did not differ between TGF-β–treated or untreated samples (Supplemental Fig. 1A). The data clearly show that TGF-β suppresses expression of molecules required for apoptosis, particularly Bim, by CD4⁺CD25⁻ cells stimulated by PICA-inducing conditions.

We next determined whether TGF-β signaling is required for nTreg resistance against PICA for the same reason as conventional T cells. If TGF-β receptor signaling in nTregs acted to keep Tregs resistant to PICA, it was predicted that nTregs treated with TGF-β signaling inhibitor would express higher levels of Bim. To test this prediction, CD4⁺CD25⁺ Tregs were purified and stimulated with plate-bound anti-CD3/anti-CD28 Abs for 2 d with or without SB431542. Cells were harvested and tested for the expression of Bim, Fas, and FasL (cell surface; Fig. 3C, 3D, Supplemental Fig. 1B). Stimulated Tregs expressed a lower level of Bim protein to unstimulated cells and showed a stark contrast to Bim expression by CD4⁺CD25⁻ T cells as reported previously (10). In contrast, Tregs that were stimulated in the presence of TGF-β signaling inhibitor showed a substantial upregulation of both isoforms of Bim expression (Fig. 3C). The EL form is considered to have a major role in apoptosis by inducing release of apoptotic proteins Bax and Bak (31). Unlike Bim, Fas, and FasL, expression by stimulated CD4⁺CD25⁺ nTregs did not change with TGF-β treatment (Fig. 3D, Supplemental Fig. 1B). Taken together with the data from studies with CD4⁺CD25⁻ T cells, the data demonstrate that TGF-β suppresses Bim protein expression under PICA-inducing conditions and blocks apoptosis.

TGF-β is not only involved in iTreg differentiation but also for other helper T cell subset differentiations, such as T_H9 or T_H17 (32, 33). Because TGF-β rescued CD4⁺CD25⁻ T cells from PICA without inducing Foxp3⁺ Tregs, we determined whether cells survived PICA in the presence of TGF-β differentiated into other effector T cell subsets. To address this question, we stimulated purified CD4⁺CD25⁻ T cells with plate-bound anti-CD3 plus either soluble or plate-bound anti-CD28 Abs in the presence or absence of TGF-β. After 3 d of stimulation, cells expressing IL-9 or IL-17 were assessed by intracellular cytokine staining. CD4⁺CD25⁻ T cells stimulated by plate-bound anti-CD3 plus anti-CD28 without TGF-β did not express IL-9, but a significant portion of the cells stimulated by the same manner in the presence of TGF-β expressed IL-9 (14%; Fig. 4A). Culture supernatant from cells stimulated with plate-bound Abs and TGF-β showed a substantial increase in IL-9 compared with the samples from cells stimulated without TGF-β (Fig. 4B). Actual cell number producing IL-9 also increased significantly with TGF-β (Fig. 4C), showing that TGF-β induced differentiation or expansion of a group of CD4⁺CD25⁻ T cells into T_H9 under PICA-inducing conditions. In contrast, CD4⁺CD25⁻ T cells stimulated by plate-bound anti-CD3 plus soluble anti-CD28 express a significantly lower level of IL-9 with TGF-β (Fig. 4B). No increase in T_H17 was observed under either of these conditions (Fig. 4A, 4C).

IL-4 has a pivotal role in the generation of T_H9 (34). The addition of anti–IL-4 Ab abrogated induction of T_H9 by TGF-β and plate-bound anti-CD3/anti-CD28 Abs (Fig. 4A). Whereas IL-4–producing cells were not detectable by cytokine staining after 3 d of stimulation (data not shown), culture supernatants from cells stimulated with plate-bound anti-CD3/anti-CD28 Abs contained a clearly detectable level of IL-4 either in the presence or absence of TGF-β (Fig. 4B). TGF-β abrogated IL-4 production from cells stimulated with plate-bound anti-CD3 and soluble anti-CD28, whereas no decrease of IL-4 was observed for cells stimulated with plate-bound anti-CD3/anti-CD28 Abs (Fig. 4B).

T cells from BALB/c mice showed the same responses when stimulated by plate-bound anti-CD3 and anti-CD28 Abs (Supplemental Fig. 2A). TGF-β rescued CD4⁺CD25⁻ T cells from PICA and induced T_H9 differentiation. A difference was found when T cells were stimulated by soluble anti-CD28 Ab. Unlike T cells from C57BL6 mice, a substantial amount of BALB/c mouse T cells developed into T_H9 after simulation by soluble anti-CD28 Abs in the presence of TGF-β. This result is likely due to a high level of IL-4 production with soluble anti-CD28 Ab stimulation (Supplemental Fig. 2B). Whereas IL-4 expression by C57BL6 T cells was abrogated by TGF-β when anti-CD28 Ab was provided in a soluble form (Fig. 3B), TGF-β increased IL-4 production by BALB/c T cells stimulated under the same conditions. The data are in agreement with those observed with C57BL6 mouse T cells and show the significance of IL-4 in T_H9 generation by plate-bound anti-CD3/anti-CD28 Abs plus TGF-β. Together, the data suggest that T cells stimulated with plate-bound anti-CD3/anti-CD28 Abs differentiate into T_H9 in part because of the presence of autocrine IL-4.

In contrast to the effect on IL-4, TGF-β suppressed production of IFN-γ regardless of how anti-CD28 Abs were provided (Supplemental Fig. 3). No differentiation of IFN-γ⁺ cells were observed from cells that resisted PICA by TGF-β addition. TGF-β also suppressed expression of IFN-γ by BALB/c T cells (Supplemental Fig. 2A, 2B).

IL-6 has a critical role in regulating the balance between T_H17 and Tregs and induces T_H17 along with TGF-β (35). Because T cells do not produce IL-6, we tested whether exogenous IL-6 changes the fate of CD4⁺CD25⁻ T cells under PICA-inducing conditions. When CD4⁺CD25⁻ T cells were stimulated in the presence of TGF-β and IL-6, the frequency of IL-17⁺ cells showed a modest increase over the TGF-β–only control groups (Fig. 4A). The increase was higher for the plate-bound anti-CD28 Ab stimulation than for soluble anti-CD28 stimulation (3.7% versus 1.4%). In addition, we observed a significant increase in the amount of IL-17 detected in the culture supernatant for cells stimulated with plate-bound anti-CD28 Ab than with soluble anti-CD28 controls (Fig. 4B). The addition of IL-6 increased the total cell number and IL-17⁺ cells (Fig. 4C, Supplemental Fig. 4). Therefore, a marked increase in IL-17 production by plate-bound anti-CD28 Ab-simulated T cells may be due to an increase in total live cell numbers in the presence of exogenous IL-6. The data show that TGF-β promotes differentiation and expansion of T_H17 in the presence of IL-6 when T cells are stimulated by plate-bound anti-CD3 and anti-CD28 Abs. IL-6 also increased the secreted IL-9 by T cells stimulated with plate-bound anti-CD28 in the presence of TGF-β, although IL-9⁺ cells were at the level comparable to cells stimulated without TGF-β, suggesting an increase in the level of IL-9 production per individual cell (Fig 4B, 4C).

Our data demonstrate fundamental differences in T cell activation when CD28 is engaged by the plate-bound or soluble form of anti-CD28. To determine the underling mechanism that controls apoptosis or cell survival and differentiation, signaling processes involved in Bim expression were compared between soluble anti-CD28 and plate-bound anti-CD28 Ab-stimulated T cells. CD4⁺CD25⁻ T cells were purified from total splenocytes and stimulated with plate-bound anti-CD3 plus soluble or plate-bound anti-CD28 Abs in the presence or absence of TGF-β. After 1 d of stimulation, total cell lysates were prepared and analyzed by Western blot (Fig. 5).

We determined whether plate-bound and soluble anti-CD28 Ab stimulation differs in inducing the signaling process of the Akt/FoxO3a axis, because previous studies on cytokine deprivation-induced apoptosis of T cells showed that FoxO3a, a Forkhead transcription family member, induced expression of Bim, whereas Akt suppressed Bim expression via inhibitory phosphorylation of FoxO3a (36, 37). Expression of FoxO3a showed a substantial increase in plate-bound Ab stimulated T cells over unstimulated or soluble anti-CD28 stimulated samples. The addition of TGF-β, which renders T cells resistant to PICA, caused a marked decrease of FoxO3a expression by plate-bound anti-CD28 Ab stimulated samples, whereas no obvious change was observed for soluble anti-CD28 stimulated T cells. Inhibitory phosphorylation of FoxO3a at Ser253 was not substantially changed by TGF-β either in plate-bound or soluble anti-CD28 Ab stimulated samples.

Expression of Akt, a negative regulator of FoxO3a, increased after soluble and plate-bound anti-CD28 Ab stimulation. TGF-β did not cause significant changes in Akt protein levels. However, TGF-β upregulated the level of activating phosphorylation of Akt at residue 473 only in plate-bound anit-CD28 stimulated samples, suggesting that TGF-β inhibited FoxO3a expression in part by activation of Akt.

FoxO1 is another Forkhead transcription factor that is regulated by Akt (38). Expression and phosphorylation of FoxO1 was markedly induced by TGF-β in cells stimulated by plate-bound and soluble anti-CD28 Abs to a comparable extent. Therefore, FoxO1 expression is a downstream target of TGF-β but not linked to plate-bound anti-CD28 Ab stimulation or PICA. The data suggest that expression of FoxO3a is one of the unique downstream signaling events that differs between plate-bound and soluble anti-CD28 Ab stimulation and is potentially involved in PICA. ERK1/2 is known to downregulate FoxO3a via MDM-mediated degradation (39). A mild increase in the level of activated ERK was observed in plate-bound anti-CD28 Ab-stimulated samples compared with unstimulated T cells or T cells stimulated by soluble anti-CD28 Abs. However, TGF-β did not enhance ERK activation or expression. Thus, ERK activity did not correlate with the level of FoxO3a expression. The data show a correlative link between PICA and expression of FoxO3a, which is negatively regulated by TGF-β under PICA-inducing conditions.

In this study, we demonstrated that TGF-β signaling renders CD4⁺CD25⁻ T cells resistant to PICA and is required for survival and expansion by nTregs ex vivo when stimulated by plate-bound anti-CD3/anti-CD28 Abs. TGF-β rendered CD4⁺CD25⁻ T cells resistant to PICA and differentiated them to T_H9 or T_H17, depending on the presence of IL-4 and IL-6, respectively. These data suggest that TGF-β signaling has another role in controlling numbers of conventional and regulatory CD4⁺ T cells during Ag stimulation.

Our data show that TGF-β reduced expression of Bim and FoxO3a. Recent reports showed that TGF-β regulates expression of Bim in nonlymphoid cells and mitogen- and stress-activated protein kinase-1 (MSK1) had a critical role in the antiapoptotic function of TGF-β (40, 41). Currently, it is not known whether MSK1 plays any role in T cell activation or death but investigations to determine the role, if any, of MSK1 in PICA are ongoing. It should also be noted that reduction of FoxO3a expression by TGF-β in T cells has not been reported. The data presented here are correlative evidence, and whether the reduction of FoxO3a by TGF-β has a functional role in PICA is currently under investigation.

Although the underlying mechanism is not clear, the data also demonstrate that induction of FoxO3a by anti-CD28 Ab immobilized on the plastic surface, but not by soluble anti-CD28 Ab. This FoxO3a expression was reduced by TGF-β. A recent report showed that TGF-β causes inactivation of FoxO3a and reduction of Bim expression in a PI3K dependent manner in mesangial cells (42). In this study, it was shown that TGF-β caused activation of Akt and inactivating phosphorylation or degradation of FoxO3a. Our data also show that the addition of TGF-β causes reduction of FoxO3a and a mild but reproducible increase in Akt phosphorylation, suggesting that reduction of FoxO3a by TGF-β is mediated by activation of the PI3K/Akt pathway. Although CD28 engagement is known to provoke PI3K signaling and its downstream process involving Akt and mTOR (43), multiple residues in the cytoplasmic region of CD28 are known to have functional roles other than activation of PI3K (44, 45). Therefore, it is possible that differential signaling is provided to the PI3K/Akt signaling pathway from CD28 when stimulation is provided by plate-bound or by soluble anti-CD28 Ab.

TGF-β promotes nTreg cell survival during negative selection where Bim has a critical role (46). Although thymic negative selection does not require p53, the data suggest that TGF-β signaling can be antiapoptotic under certain conditions in connection to Bim expression. Although PICA is an ex vivo event established by the use of antireceptor Abs, our previous work showed that PICA can be induced by extended stimulation from allogeneic dendritic cells in vitro (10). Therefore, it will be interesting to see whether TGF-β rescues conventional T cells from PICA in vivo. PICA may be used by chronically infecting agents or tumor cells, or both, that establish their survival by expansion of nTregs. Conversely, because TGF-β is critical for the survival of nTregs against PICA, inhibition of TGF-β signaling could lead to the loss of nTregs and abrogation of suppression or tolerance.

Complex and intricate regulation of Bim by TGF-β potentially reflects what has been reported on the role of miR-25 (47). In both CD4⁺CD25⁻ and CD4⁺CD25⁺ T cells, Bim protein level is regulated negatively by TGF-β. Notably, recent reports showed that miR-25, which regulates Bim protein synthesis and promotes antiapoptotic responses, was much reduced in Tregs from patients with multiple sclerosis (48, 49). Loss of this miR-25 could lead to an increase in Bim protein expression by Tregs and their death, hence less effective maintenance of self-tolerance. We are currently investigating the potential role of this miR-25 in Bim expression in Tregs under PICA-inducing conditions.

Data presented in this study also showed that TGF-β promotes differentiation of CD4⁺CD25⁻ T cells that receive PICA-inducing stimuli. Currently, the molecular mechanism underlying this phenomenon is unknown. TGF-β may be simply providing signaling required for survival of T cells, and IL-4 provides differentiation signaling for T_H9. Similarly, TGF-β might allow T cells to survive PICA so that exogenous IL-6 can induce differentiation of surviving cells into T_H17. Alternatively, TGF-β is also providing signaling required for initiation and establishment of differentiation. In either case, the plasticity of T cell differentiation provided by TGF-β and PICA-inducing stimulation could have significant roles in determining the outcomes of in vivo immune responses. Further study is needed to elucidate the potential roles of PICA and TGF-β under physiologic and pathologic conditions.

We thank Dr. Phong Le, Dr. Pandelakis Koni, Dr. Kathleen Jaeger, and Dr. Shauna Marvin for critical reading of the manuscript, Dr. Yoichi Seki for suggestions, and Patricia Simms (Loyola FACS core facility) for cell sorting.

The authors have no financial conflicts of interest.