CD4⁺CD25⁺ Regulatory T Cells Attenuate the Phosphatidylinositol 3-Kinase/Akt Pathway in Antigen-Primed Immature CD8⁺ CTLs during Functional Maturation ¹

Cytotoxic T lymphocytes play a key role in immune responses by directly killing target cells (TCs), ³ such as virus-infected cells, tumors, intracellular parasites, and allogeneic grafts, in an Ag-specific manner (1, 2, 3, 4, 5). However, naive immature CD8⁺ CTLs do not exhibit these cytotoxic abilities (6, 7) until they have passed through several key developmental stages, such as Ag priming, differentiation, and expansion into functional CTLs (8). Functional CTL differentiation is triggered by Ag recognition through TCR and requires activating signals mediated through the TCR/CD3 complex and costimulatory molecules as well as cognate interaction with stimulators/APCs via other adhesion molecules (7, 9, 10). With the exception of CTL generation during certain viral infections (11, 12, 13), Ag priming of immature CTLs requires CD4⁺ Th cells (14, 15, 16). IL-2 is thought to be the most important cytokine required for functional CTL maturation (17, 18) and supports CTL expansion and the production of perforin and other agents required for cytotoxicity (19, 20, 21). Despite the importance of IL-2 in functional CTL differentiation, excess IL-2 promotes the development of lymphokine-activated killer (LAK) cells, which lack Ag specificity. Because LAK cells are capable of killing self TCs in vitro (22, 23), they are thought to be generally undesirable as in vivo effectors.

IL-2 secreted from CD4⁺ T cells binds to the IL-2R, a heterotrimeric molecule, on Ag-primed immature CTLs (24). Signals from the IL-2R promote many important cellular events, including the survival, proliferation, and effector functions mediated by several pathways, such as the JAK/STAT and PI3K/Akt pathways (25). STAT5 is activated by JAK1 and JAK3, which are associated with IL-2R, after IL-2 binding (26). Activated STAT5 moves to the nucleus to serve as a transcription factor and directly activates the target genes (26). It is also known that IL-2R signaling activates the PI3K/Akt pathway (27). Activated Akt is thought to play an important role in cell cycle progression and survival (28, 29). Recently, it has been reported that activated Akt promotes increased cell size of resting T cells, suggesting that the PI3K/Akt pathway controls blast formation of T cells (30).

CD4⁺CD25⁺ immune regulatory T (Tr) (3) cells were originally identified as negative regulators of tissue-specific autoimmunity (31, 32, 33). However, it has also been reported that Tr cells may contribute to the regulation of immune responses, including virus infections (34, 35, 36), parasite infections (37, 38), and graft rejection (39, 40). Despite these extensive investigations, the precise nature of the immunological role of Tr cells remains unclear (41).

We have studied the effect of Tr cells on functional CTL maturation and effector functions. Our data show that Tr cells suppress functional differentiation of CTLs. However, they had no effect on CTL effector functions. When Tr cells were added within the first 2 days of MLC, functional maturation of CTLs was inhibited. However, after 3 days of MLC, suppression of mature CTL generation by Tr cells was not evident. Additionally, our data show that Tr cells inhibit Akt activation, but have no effect on STAT5 activation. The inhibition of Akt was accompanied by the arrest of increased cell size in Ag-primed immature CTLs. Conversely, enhancement of functional CTL differentiation characterized by nonspecific cytotoxicity was induced by the depletion of Tr cells from responders of MLC. Therefore, Tr cells act directly on immature CTLs and regulate their Ag-specific and functional maturation, at least in part, by down-regulation of the PI3K/Akt pathway.

BALB/c (H-2^d) and C57BL/6 (H-2^b) mice were purchased from CLEA Japan. Mice were housed under specific pathogen-free conditions in our Laboratory Animal Research Centre and were handled according to Guidelines for the Care and Use of Laboratory Animals, Dokkyo University School of Medicine (protocol 0047).

Tumor cell lines EL4 (derived from C57BL/6; H-2^b), A20.2J (derived from BALB/c; H-2^d), and YAC-1 (derived from A/Sn; H-2^s) were cultured in complete RPMI 1640 (RPMI 1640 supplemented with 5% heat-inactivated FCS, 10 mM HEPES, 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, 0.1 mg/ml streptomycin, and 50 μM 2-ME). B6d, a C57BL/6 origin anti-H-2^d CTL line, was established by biweekly Ag stimulation. The CTL line was cultured in the complete RPMI 1640 further supplemented with 2.5% supernatant from Con A-stimulated rat spleen cells as a source of T cell growth factors and heat-inactivated FCS (at a final concentration of 10%). CD4⁺CD25⁺ Tr cells were purified from spleen and lymph nodes by treatment with a combination of FITC-conjugated anti-CD25 mAb (clone 7D4; BD Pharmingen) and anti-FITC MicroBeads (Miltenyi Biotec), followed by positive selection with the autoMACS system (Miltenyi Biotec). Before positive selection of Tr cells, CD4⁺ cells were enriched by the depletion of B220⁺ cells and CD8⁺ cells by plate-coated anti-B220 mAb (RA3-3A1/6.1) and by the combination of anti-CD8 mAb (83-12-5) plus complement. The purity of Tr cells was >78%. In some experiments, freshly purified Tr cells were stimulated with plate-coated anti-CD3 mAb (clone 2C11; BD Pharmingen) in the presence of human rIL-2 (100 U/ml) (Invitrogen Life Technologies) for 48 h, and they were used as preactivated Tr cells. All Tr cells used in this study were syngeneic to the responders or effectors. For cell depletion, two different systems were used: 1) for CD4⁺ and/or CD8⁺ cell depletion from effector cells in the cell-mediated cytotoxicity assay, cells were treated with anti-CD4 mAb (GK1.5) or anti-CD8 mAb, which were prepared from hybridoma culture supernatants, and then treated with complement (Cedarlane Laboratories); 2) for depletion of Tr cells, or CD4⁺ and MHC class II-positive cells from responders in MLC, responder cells were treated with FITC-conjugated anti-CD25 mAb, or PE-conjugated anti-CD4 mAb and PE-conjugated anti-MHC class II mAb, followed by treatment with anti-FITC or anti-PE MicroBeads (Miltenyi Biotec), respectively. Cells were then depleted using the autoMACS system (Miltenyi Biotec). T cell-depleted spleen cells were used as APCs. Spleen cells were stained with PE-conjugated anti-Thy-1 mAb. T cells were then depleted using the autoMACS system, as described above. Contaminations of Tr, CD4⁺, MHC class II-positive cells, and T cells after depletions were <0.14, 0.8, 1.1, and 1.2%, respectively. CD4⁺ and MHC class II-positive cell-depleted responders were used as purified CD8⁺ T cell responders (CD8⁺ T cells >82%). In preliminary experiments, it was confirmed that the mAb and complement treatment lysed almost 100% of thymocytes.

To induce functional mature CTL from naive CD8⁺ T cells, conventional MLC were used. Briefly, spleen cells from BALB/c mice (C57BL/6 mice for Fig. 2,A) were cocultured with irradiated spleen cells from C57BL/6 mice (BALB/c mice for Fig. 2 A) as stimulators for 5 days. To examine the direct effect of Tr cells on CD8⁺ CTL differentiation, purified CD8⁺ T cells or CD4⁺ T cell-depleted responder cells were used as responders. MLC without CD4⁺ T cells were supplemented with murine rIL-2 (rmIL-2) (Genzyme) at a final concentration of 2 U/ml; the concentration was determined by preliminary experiments to induce high Ag-specific cytotoxicity. Typically, 2 × 10⁶ responders and 0.8 × 10⁶ stimulators were mixed in a well of 48-well plate. Cell numbers were proportionally adjusted when MLC were conducted using 96-well plates. A PI3K inhibitor, wortmannin, was purchased from Sigma-Aldrich.

Cytotoxicity of CTLs from MLC was measured by ⁵¹Cr release assay. Briefly, TCs were labeled with 50 μCi of Na₂⁵¹CrO₄ (PerkinElmer) for 1 h at 37°C and then washed three times. ⁵¹Cr-labeled TCs (2500/well) were coincubated with effector cells from MLC for 4 h at 37°C in 96-well U-bottom plates. Culture supernatants were harvested using a Skatron harvesting system (Molecular Devices). Cytotoxic activities were expressed as percentage of specific ⁵¹Cr release from TC and were calculated as follows: percentage of specific ⁵¹Cr release = (Exp − Spt)/(Tot − Spt) × 100. Exp, Spt, and Tot indicate experimental release, spontaneous release, and total incorporation, respectively. Instead of an adjustment of effector cell number, numbers of input responders were adjusted in some experiments, in which cell dilution, instead of E/T ratio, was indicated. To adjust input cell number in the cytotoxicity assay, effector cells harvested from control MLC and from MLC with additional Tr cells were resuspended in equal volumes of medium. Effector cell suspensions were further serially diluted. The additional Tr cells were not removed from effector cells for the cytotoxicity assay. In all cytotoxicity assays used in this study, spontaneous ⁵¹Cr release was <18% of total incorporation.

For in vitro cell proliferation assays, B6d cells were labeled with 0.5 μM CFSE (Molecular Probes). CFSE-labeled CTLs (2 × 10⁵/well) were then stimulated with irradiated (3000 rad) BALB/c spleen cells (4 × 10⁵/well) in the presence of rmIL-2 at a final concentration of 2 U/ml in 48-well plates. After 2 days of stimulation, cells were harvested and CFSE intensity was assessed by flow cytometry.

B6d cells (5 × 10⁴/well) were cocultured with BALB/c spleen cells (1 × 10⁵/well) as stimulators, T cell-depleted B6 spleen cells (1 × 10⁵/well) as APC, and either preactivated or freshly prepared Tr cells (1 × 10⁴ or 5 × 10⁴/well) in 96-well U-bottom plates. Four hours after culture, supernatant were harvested and IFN-γ production was assessed by ELISA. The ELISA kits for IFN-γ were purchased from BioSource International.

All dye-conjugated and biotin-conjugated mAb used in this study were purchased from BD Pharmingen, except for PE-Akt1 mAb (Santa Cruz Biotechnology) and PE anti-granzyme B mAb (Caltag Laboratories) (42). Anti-phospho-Akt (p-Akt) (Ser⁴⁷³) mAb was purchased from Cell Signaling Technology. Anti-p-STAT5 mAb, anti-STAT5 mAb, anti-p-ZAP70, and anti-ZAP70 mAb were purchased from BD Pharmingen. Isotype controls, i.e., purified mouse IgG1, IgG2a, and IgG2b, were purchased from e-Bioscience. Culture supernatant from anti-CD16/CD32 (FcγR) mAb-producing hybridoma, clone 2.4G2, was prepared in our laboratories and was used for blocking nonspecific binding of dye-conjugated mAbs to FcγR. For intracellular staining, cells stained with the desired anti-surface molecule mAbs were fixed with PBS containing 4% paraformaldehyde for 5 min at 37°C. For ZAP70 and p-ZAP70 staining, cells were fixed with PBS containing 1% formaldehyde, followed by 80% EtOH. Fixed cells were washed with saponin-PBS (PBS containing 0.1% saponin, 0.1% BSA, 0.1% NaN₃, and 0.01 M HEPES). After washing, cells were resuspended in saponin-PBS and stained with Abs for intracellular molecules, followed by washing with saponin-PBS. Flow cytometry data acquisition and analysis were performed using FACSCalibur and CellQuest programs (BD Biosciences).

The χ² test and Student’s t test were used to determine the statistical significance. Values of p <0.05 were considered statistically significant.

We first examined whether Tr cells attenuated CTL effector functions. We used a redirectional stimulation system (43, 44), as shown in Fig. 1,A (inner panel). An anti-H-2^d-specific conventional CD8⁺ CTL line, B6d, which was established from C57BL/6 spleen cells by repeated Ag stimulation, was used as effector cells. The B6d cells, preactivated Tr cells, which were pretreated with anti-CD3 mAb, from B6 mice, and ⁵¹Cr-labeled FcR-expressing BALB/c origin B lymphoma, A20.2J, were cocultured for 4 h to assess the effect of Tr cells on lethal hit delivery by B6d. As shown in Fig. 1,A, preactivated Tr cells had no impact on mature CTL Ag-specific cytotoxicity. It is important to note that identical conditions of pretreatment with anti-CD3 mAb successfully activated T cell in parallel control redirectional stimulation, and that this activation was inhibited by the addition of FcR-blocking mAb. Recently, it has been reported that Tr cells are self-reactive (45, 46, 47). Therefore, we used syngeneic APCs instead of a redirectional stimulation system to assess the effect of Tr cells on IFN-γ production by CTLs. As shown in Fig. 1,B, the addition of preactivated Tr cells did not affect the production of IFN-γ by CTLs. Freshly isolated Tr cells also did not show any suppressive effects on CTL functions (data not shown). Because functional CTL assays (cell-mediated cytotoxicity assay and IFN-γ production) and MLC were performed for 4 h and for 5 days, respectively, it was possible that 4 h of the CTL assay was too short a period for Tr cells to display any regulatory effects on CTLs. To investigate this possibility, CFSE-labeled resting B6d cells (3 wk after the last stimulation) were restimulated in the presence of Tr cells and syngeneic APCs. The cultures were supplemented with rmIL-2 at a final concentration of 2 U/ml. Two days after the stimulation, CFSE intensity was measured to assess cell division. In line with data from the short-term CTL function assessments, Tr cells did not affect mature CTL proliferation (Fig. 1 C). Taken together, these data indicate that Tr cells do not inhibit mature CTL proliferation and effector function.

Preactivated Tr cells, when used in cytotoxicity assays, abrogated Ag-specific CTL generation from primary MLC (Fig. 2,A), suggesting that Tr cells played a role in the regulation of the transition from immature to mature CTLs. The generation of functionally mature CTLs in MLC takes ∼4–5 days (6, 7). To determine the stage of CTL maturation that is sensitive to Tr cell regulation, preactivated Tr cells were added to MLC on the indicated days after beginning the cultures (Fig. 2,B). To avoid the confounding effect of Tr cells on CD4⁺ Th cells in CTL differentiation and to evaluate the direct effect of Tr cells on CD8⁺ CTL differentiation, we used CD4⁺ T cell-depleted responders and exogenously added rmIL-2 to the MLC. Five days after the MLC, cytotoxicity generated from MLC was tested. It was found that if Tr cells were added to MLC on day 0, CTL generation was abrogated (Fig. 2,B). The complete inhibition of CTL generation was also observed when preactivated Tr cells were added to the culture within 2 days (Fig. 2,B). In contrast, if preactivated Tr cells were added after 3 days of culture, Tr cells did not exhibit any inhibitory effect on CTL maturation (Fig. 2 B). These results indicate that at least one of the target events for Tr cells is in the early functional CTL maturation after Ag stimulation.

To investigate the mechanisms of Tr cell regulation, we examined day-to-day changes in expression of CD44, a surface marker for early T cell activation (48), and cell size of CD8⁺ T cells during functional CTL maturation in MLC. After 2 days of culture, expression of CD44 by CD8⁺ T cells in control cultures was similar to that of cultures in which Tr cells were added on day 0 (Fig. 3,A). However, the cell size of CD8⁺ T cells expressing high levels of CD44 was reduced by the addition of Tr cells. Supporting data for this result came from experiments in which preactivated Tr cells were added to cultures after 2 days of Ag stimulation (Fig. 3,B). One day after the addition of preactivated Tr cells, CD44 expression on CD8⁺ T cells was comparable to that of cells in control cultures. Similar to the results shown in Fig. 3,A, activating CD8⁺ T cells cultured with the Tr cells were smaller than those in the control cultures. Furthermore, 5 days after stimulation, the proportion of CD44^high large-sized CD8⁺ T cells, in the presence of Tr cells, was still reduced compared with that in the control cultures (Fig. 3 B). These findings suggest that Tr cells control cell size, expansion, and/or survival of Ag-primed immature CTLs at day 3 stage after Ag priming. Significantly, CD44 expression and cell size were both increased in CD8⁺ T cells after 2 days of culture, indicating that activation signals through TCR were already in effect. Thus, Tr cells regulate further processes of activation and maturation in CTL differentiation located after TCR stimulation rather than immediately downstream of TCR signaling.

Recent studies reported that Akt (also designated protein kinase B), which is activated by PI3K, promotes T cell blast formation, proliferation, and survival (27, 28, 29, 30). We therefore examined the effect of Tr cells on Akt activation in CD8⁺ T cells during the course of CTL maturation. Two days after Ag stimulation, p-Akt was observed in only a few activating CD8⁺ T cells with large cell size (Fig. 4,A). Levels of p-Akt in activating CD8⁺ T cells were markedly elevated 3 days after Ag stimulation. However, when Tr cells were added to the culture 2 days after Ag stimulation, p-Akt levels in activating CD8⁺ T cells were markedly reduced (Fig. 4,A). Similar results were observed in experiments using another anti-p-Akt Ab, which recognizes another phosphorylation site of p-Akt (data not shown). The reduction of p-Akt levels was not due to decreased Akt expression, because a similar level of Akt in CD8⁺ T cells was observed in cultures with or without Tr cells (Fig. 4,B). It is well known that Akt is activated by TCR, CD28, and IL-2R stimuli (27). To investigate whether Tr cells inhibited the PI3K/Akt pathway located downstream of TCR signaling, we assessed levels of p-ZAP70 in CD8⁺ T cells during functional maturation. As shown in Fig. 4,C, after 2 days of culture, levels of p-ZAP70 in activating large-sized CD8⁺ T cells in control cultures were similar to those in a culture with Tr cells, which were added on day 0. This result supports data shown in Fig. 3. Thus, in the presence of Tr cells, the pathway immediately downstream of TCR signaling was intact in CD8⁺ immature CTLs 2 days after stimulation. Next, we analyzed the effect of Tr cells on IL-2R signaling in CD8⁺ immature CTLs during functional maturation. It is well documented that IL-2 plays a key role in T cell growth, including CTL expansion during differentiation (17, 18). We examined whether the status of active STAT5, a mandatory molecule of the STAT family in IL-2R signaling (26, 49, 50), was affected by Tr cells during functional CTL maturation. Because p-STAT5 was observed in Ag-primed immature CTLs 2 days after stimulation (Fig. 4,C), preactivated Tr cells were cocultured with naive CD8⁺ CTLs at the beginning of culture. In contrast to p-Akt, Tr cells did not affect the level of p-STAT5 (Fig. 4 C). Taken together, these findings suggest that Tr cells principally regulate cell expansion by modulating the PI3K/Akt pathway during functional CTL maturation.

Akt is activated by PI3K through both TCR and IL-2R signaling (27). Therefore, we assessed the effect of Tr cells on IL-2R expression by activating CD8⁺ immature CTLs. If preactivated Tr cells were added after 2 days of culture, the proportion of IL-2Rα-expressing cells decreased in comparison with control cultures (Fig. 5,A). Significantly, the proportion of activating immature CTLs (large-sized cells) among CD8⁺ CTLs cultured with additional Tr cells (12.9 + 21.0%) was equal to that in the control culture (19.7 + 14.5%). These data suggest the possibility that p-Akt levels are reduced by Tr cells due to the inhibition of IL-2Rα expression in immature CTLs. To test this possibility, wortmannin, a PI3K-specific inhibitor, was added to MLC, in which purified CD8⁺ T cells were used as responders, after 2 days of culture. IL-2Rα expression was determined 24 h later. As shown in Fig. 5 B, wortmannin suppressed IL-2Rα expression as well as p-Akt levels in immature CTLs, suggesting that increments in IL-2Rα expression in immature CTLs from the day 2 stage to day 3 stage during CTL maturation depend on the PI3K pathway. These results suggest that Tr cells inhibit IL-2Rα expression in Ag-primed immature CTLs as a result of down-regulation of the PI3K/Akt pathway that potentiates IL-2Rα expression.

It was reported that both TCR stimulation and IL-2 per se could induce IL-2Rα expression (51). Therefore, we next assessed whether high doses of IL-2 reversed the down-regulation of IL-2Rα and the inhibition of CTL maturation induced by Tr cells. High doses of IL-2 restored suppressed CTL maturation by Tr cells (Fig. 6,A). In the absence of Tr cells, 2.5 U/ml rmIL-2 was successful in inducing Ag-specific CTLs in CD4⁺ T cell-depleted MLC. However, a 4-fold increase in rmIL-2 concentration (10 U/ml) was required for the induction of functional CTL in the presence of Tr cells. This was accompanied by restoration of both the p-Akt and IL-2Rα levels in CTLs (Fig. 6,B). Furthermore, Ag-nonspecific CTL activities were induced in MLC with Tr cells in the presence of 20 U/ml rmIL-2. These observations indicate that IL-2R signals can overcome the suppression of functional CTL maturation by Tr cells. Significantly, Tr cells did not affect mature CTL proliferation in the presence of 2 U/ml rmIL-2 (Fig. 1,C). Supporting data shown in Fig. 6 demonstrate that failure of Tr cells to inhibit CTL proliferation was not due to supplemented rmIL-2.

Highly Ag-specific CTLs were generated in normal MLC, in which normal spleen cells containing Tr cell were used as responders. We therefore investigated whether Tr cells regulate CTL development under normal physiological circumstances. To clarify this point, Tr cells were depleted from normal BALB/c spleen cells and the Tr cell-depleted cells were cocultured with irradiated spleen cells derived from C57BL/6 mice. After 5 days in culture, cells were harvested and cytotoxicity was assessed. As expected, effector cells harvested from control MLC killed allogeneic TCs (C57BL/6 origin EL4 cells), but not syngeneic (BALB/c origin A20.2J cells) or third party TCs (A/Sn origin YAC-1 cells), which are susceptible to NK cell-mediated cytotoxicity, in an Ag-specific manner (Fig. 7,A). Interestingly, three different modes of cell-mediated cytotoxicity, namely Ag-specific, Ag-nonspecific (anti-self), and NK activity, were observed in effector cells induced from MLC with Tr cell-depleted responder cells. Furthermore, effector cells induced from MLC with Tr cell-depleted responder cells killed EL4 cells much more effectively than did those from control MLC. However, the enhanced cytotoxicity was accompanied by undesirable Ag-nonspecific or autocytotoxicity and NK activity (Fig. 7,A). Further studies revealed that CD8⁺ CTLs generated from Tr cell-depleted responders developed nonspecific cytotoxicity, or lost high Ag specificity. The cytotoxicity against EL4 TCs induced from Tr cell-depleted MLC was abrogated by pretreatment with anti-CD8 mAb and complement, indicating that CD8⁺ T cells were the effector cells responsible for the Ag-specific lysis of EL4 TCs (Fig. 7,B). CD8⁺ T cell depletion partially reduced nonspecific TC (A20.2J) lysis observed in Tr cell-depleted MLC, suggesting that some of the CD8⁺ CTLs generated in MLC without Tr cells lysed syngeneic TCs (Fig. 7,B). Importantly, both CD4⁺ and CD8⁺ T cell-depleted effectors retained little nonspecific cytotoxicity. This might be due to NK cells, because significant NK activity, which was resistant to CD4⁺ and CD8⁺ T cell depletion, was observed in Tr cell-depleted MLC (Fig. 7,B). Generally, the proportion of CD8⁺ cells in effector cells obtained from Tr cell-depleted MLC increased in comparison control MLC (data not shown). To determine whether the enhancement of CD8⁺ CTL cytotoxicity was due to the acceleration of expansion and/or enhancement of the magnitude of CTL, the E:T cell ratio was recalculated as the CD8⁺ cell/TC ratio based on results from flow cytometric analysis. In the same ranges of CD8⁺ cell/TC ratio, the cytotoxicity of effector cells from Tr cell-depleted MLC was higher than that from control MLC, suggesting that depletion of Tr cells from responders enhanced not only the expansion, but also the magnitude of cytotoxicity of CD8⁺ CTLs (Fig. 7,C). This was further confirmed by analysis of expression of granzyme B, one of the effector molecules for cytotoxicity (52). CTLs differentiated in Tr cell-depleted MLC expressed higher granzyme B than did those in normal control MLC (Fig. 7 D). The depletion of Tr cells also enhanced and prolonged p-Akt levels in Ag-primed CTLs during differentiation (data not shown). Taken together, these data suggest that the possible role of Tr cells in functional CTL maturation is to fine-tune Ag specificity and generate optimal CTL immune responses.

In this study, we investigated the role of Tr cells in CTL effector function and functional maturation. Using in vitro assays, we demonstrated that Tr cells failed to suppress the effector functions of mature CTLs, but effectively inhibited functional CTL maturation in MLC. It has been recently reported that Tr cells suppress memory/effector CD8⁺ CTL function in allograft rejection, which is predominantly mediated by CD8⁺ CTLs (53). Delay of allograft rejection by Tr cells was reported to be due to acceleration of memory/effector CD8⁺ CTL apoptosis (53). Together with our data, this suggests that Tr cells regulate long-term survival, but not proliferation and cytotoxicity of memory/effector CTLs. In contrast to our data, it was reported that Tr cells from HSV-1-infected mice suppressed memory/effector CD8⁺ T cell proliferation (34). It is possible that Tr cells induced by HSV-1 infection are different from those used in our system and the regulatory effects or machinery may depend on the type of Tr cells. Further studies will be required to clarify this issue. Piccirillo and Shevach (54) demonstrated that Tr cells abrogated IFN-γ production of CD8⁺ T cell. According to their observations, IFN-γ production of 5 × 10⁴ naive CD8⁺ cells was affected by 1 × 10⁴ Tr cells and 5 × 10⁴ Tr cells almost abrogated IFN-γ production of naive CD8⁺ cells. In contrast, we found that 5 × 10⁴ Tr cells did not exert any suppressive effect on IFN-γ production of 5 × 10⁴ mature CTLs. As discussed below, these findings further supported that Tr cells suppress immature CD8⁺ T cell functions, but not mature CTL functions.

Functional maturation of CTLs from naive CD8⁺ T cells involves several important stages, including Ag priming of naive CD8⁺ T cells, differentiation into functional CTL, and expansion (6, 7, 8). It is well documented that Tr cells suppress naive T cells, including CD8⁺ immature CTL functions (54). It is important to determine which stage of CTL functional maturation is susceptible to Tr cell regulation. Piccirillo and Shevach (54) reported that Tr cells inhibited naive CD8⁺ T cell functions within 24 h of stimulation. According to their report, Tr cells affected the very early events of T cell activation. However, we found that Tr cells inhibited functional CTL maturation if they were added to the MLC within 2 days of culture. This indicated that at least one of the target events for Tr cell regulation occurred at the late phase of T cell activation, because in immature CTLs, even when cocultured with Tr cells, ZAP70 was already phosphorylated within 2 days after stimulation. Together, these findings suggest that there may be at least two sites of action for Tr cell regulation in functional CTL maturation depending on the system examined: one is located within the activation-induced signaling occurring within 24 h, and the other is a signaling event induced ∼2 days after Ag priming.

We found that the addition of Tr cells 2 days after priming of naive CD8⁺ T cells decreased the cell size of activating CD8⁺ cells. Recently, it was reported that activated Akt, a downstream effector of PI3K, induced an increase in the cell size of T cells (30). Levels of p-Akt in activating immature CTLs were greatly reduced at day 3 after stimulation if Tr cells were added 2 days after Ag priming. The down-regulation of p-Akt by Tr cells may arrest increases in cell size at the transition from the day 2 stage to day 3 stage that leads to further maturation. However, Harriague and Bismuth (55) reported that PI3K activation was rapid after Ag recognition by TCR. In our system, p-Akt was observed in immature CTLs 1–2 days after Ag priming. The discrepancy in kinetics of the PI3K/Akt pathway is probably a reflection of the different experimental systems, i.e., the dependency of IL-2 and/or antigenic intensity. Harriague and Bismuth (55) used T cell hybridomas and resting human T cells stimulated with super-Ags. In contrast, we used naive CD8⁺ T cells and allogeneic stimulators to generate functional CTLs in our system, which is completely dependent on IL-2. The PI3K/Akt pathway is known to be associated with both TCR and IL-2R (27).

We also found that high dose IL-2 restored functional CTL maturation in the presence of Tr cells, accompanied by normal levels of p-Akt and of IL-2Rα expression. It is possible that the Tr cell regulation is due to the consumption of IL-2 during CTL maturation. However, it should be noted that if Tr cells were cultured with immature CTLs separately in the same well using Cell Culture Insert (BD Falcon), the Tr cells were unable to express the suppressive effect (data not shown). Therefore, this possibility is unlikely to explain how high doses of IL-2 block Tr regulation of CTLs. However, Piccirillo and Shevach (54) reported that IL-2 failed to restore IL-2Rα expression inhibited by Tr cells. The discrepancy may be explained by kinetics and mechanisms of IL-2Rα expression, because we tested IL-2Rα expression on CD8⁺ T cells 3 days after stimulation, whereas they assessed it 24 h after stimulation. Rapid IL-2Rα mRNA expression induced by TCR stimulation reportedly does not require de novo protein synthesis (56). It was also reported that IL-2 per se promoted IL-2Rα gene expression via the STAT5 pathway (51). Therefore, rapid and late phases of IL-2Rα expression might be regulated differently. Furthermore, we found that the addition of PI3K inhibitors at the day 2 stage of MLC inhibited IL-2Rα expression on CD8⁺ T cells by the next day, suggesting that late-phase IL-2Rα expression on CD8⁺ immature CTLs during functional CTL maturation depends on the PI3K signaling pathway. These findings support the possibility that Tr cells inhibit IL-2Rα expression in Ag-primed immature CTLs as a result of down-regulation of the PI3K/Akt pathway, which potentiates late-phase IL-2Rα expression triggered by the STAT5 pathway.

Elimination of Tr cells from responders in MLC revealed that enhancement of CTL generation was accompanied by nonspecific cytotoxicity. The enhanced CTL activity was due to accelerated expansion of effector CTLs and increased magnitude of the cytolytic capacity of CTLs. In this system, naive CD4⁺ Th cells were cocultured with stimulators. Therefore, in the absence of negative regulation by Tr cells, CD4⁺ Th cells produced massive amounts of IL-2 (data not shown). It is well accepted that IL-2 induces CTL agents such as perforin and granzyme B, as well as cell proliferation (19, 20). Furthermore, excess IL-2 promotes CTLs with less Ag specificity, namely LAK cells, which have the potential to lyse TCs, including self tissues (22, 23). One of the most important properties of Tr cells is suppression of IL-2 production from Th cells (57). Thus, Tr cells may regulate IL-2 production from Th cells to inhibit generation of LAK activity.

The depletion of Tr cells in vivo has been reported to enhance and induce antitumor immunity exerted by CTLs (58). Together with our data, this finding suggests that Tr cells might inhibit expansion of tumor Ag-primed immature CTLs. One could also speculate that Tr cells may play a role in preventing damage caused by undesirable immune responses during normal immune responses in vivo. Indeed, it has been accepted that CD8⁺ CTLs contribute to autoimmune diseases (5). Some cases of CD8⁺ CTL-mediated autoimmune diseases may be due to insufficient Tr cell function during normal immune responses.

It is important to determine which molecules trigger the PI3K/Akt pathway regulated by Tr cells, i.e., TCR, costimulatory molecules, or IL-2R. In our system, high dose IL-2 blocked Tr cell regulation. Therefore, it is possible that down-regulation of p-Akt by Tr cells is a consequence of the PI3K/Akt pathway activated by IL-2R rather than by TCR. However, more extensive research is required to resolve this point.

In conclusion, our present study demonstrates a possible role for Tr cells in optimal CTL responses. Our data show that Tr cells directly regulate functional CTL maturation through PI3K/Akt and IL-2 pathways, and regulate CTL Ag specificity through IL-2 production by Th cells. These findings will be significant in determining the role of Tr cells in CTL-mediated immune responses during transplantation, tumor rejection, and autoimmune diseases.

We thank Yoshie Nitta for secretarial assistance. We also thank the Laboratory Animal Research Center, Dokkyo University School of Medicine, for allowing us to use the facilities.

The authors have no financial conflict of interest.