Akt is a neutral amplifier for Th cell differentiation.

Both CD28 and its relative, inducible costimulator (ICOS), have a binding motif for phosphatidylinositol 3-kinase (PI3K) in their cytoplasmic tail, and the binding of PI3K leads to activation of a serine/threonine kinase, Akt. The role of Akt in cytokine production and helper T (Th) cell differentiation remains obscure. In this study, we found that enforced expression of the constitutively active form (E40K) of Akt rendered CD4(+) T cells activated. Wild-type of Akt and E40K promoted Th1 cell differentiation in C57BL/6-derived and Th1-polarized BALB/c-derived CD4(+) T cells, while both promoted Th2 cell differentiation in BALB/c-derived and Th2-polarized C57BL/6 CD4(+) T cells. E40K also facilitated Th1 differentiation in CD4(+) T cells from IL-4-deficient mice with the BALB/c background. E40K up-regulated expression of NF-AT and c-Myb, which may be related to the augmentation of cytokine production by E40K. These findings indicate that the mechanism by which Akt augments cytokine production via CD28 and ICOS is Th cell type-specific and reflects the intracellular status affected by the cytokine milieu. We conclude that Akt is a neutral amplifier of T cell activation and Th differentiation.

ground, and costimulation (2). Among them, costimulatory molecules are thought to regulate Th differentiation by altering signal strength, providing Th-specific signals and/or biased expression of themselves on Th cells (2)(3)(4)(5). T cell costimulatory molecule CD28 plays a crucial role in the immune system in conjunction with T cell receptor signals, exerting positive effects on proliferation, cytokine production, and an anti-apoptotic effect in T cells, while suppressing anergy induction (2). Current models suggest that greater CD28 signal can achieve Th2 differentiation (6,7). An inducible costimulator, ICOS, also called H4 (8,9) or AILIM (10), is a member of the CD28 family and has roles similar to those of CD28 except for IL-2 production (11)(12)(13). Accumulating evidence indicates a predominant role of ICOS in Th2-type immune responses, as exemplified by its promoting effect on Th2 cell differentiation, concomitant cytokine secretion, airway hypersensitivity reaction, and Igclass switching (14 -17). However, ICOS has recently been reported to be functionally involved in Th1-type immune responses as well and to regulate the severity of murine experimental autoimmune encephalomyelitis (18).
Both CD28 and ICOS have a YXXM motif for recruitment of phosphatidylinositol 3-kinase (PI3K) in their cytoplasmic tail, whereas a YXNX motif for the Src homology 2 domain of Grb2 family members and a PXXP motif for the Src homology 3 domain of various molecules are present only in CD28. The YXNX motif seems to be responsible for IL-2 production (19). Cross-linking of CD28 or ICOS with TCR/CD3 leads to activation of PI3K (4,20), and PI3K in turn phosphorylates the D-3 position of the inositol ring of phosphoinositides resulting in production of phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol 3,4,5-trisphosphate, which recruit pleckstrin homology domain-containing molecules including serine-threonine kinases, Akt/protein kinase B, and PDK-1 (21). Akt, upon subsequent phosphorylation at Thr-308 and Ser-473 by PDK-1, becomes activated and plays a critical role in cell survival and cell cycle regulation through a wide variety of downstream molecules (21)(22)(23)(24).
The relevance of the PI3K/Akt pathway to cytokine production remains controversial. Several lines of evidence have demonstrated CD28 to promote IL-2 production independently of its association with PI3K (25,26). In contrast, PI3K inhibitors have been found to block cytokine production mediated by CD28 in primary cells (27,28). Kane et al. (29) have demonstrated that Akt can provide the costimulatory signal for activation of the IL-2 promoter, which is indistinguishable from the CD28 costimulatory signal. They further showed that enforced Akt expression in primary CD4 ϩ T cells from CD28deficient mice restored Th1, but not Th2, cytokine production upon restimulation. Previously, we showed Akt phosphorylation to be enhanced by engagement of CD28 and even more strongly by that of ICOS. Since several reports including ours have shown that the engagement of ICOS has a preferential effect on Th2 development, Akt may be involved in Th2 development as well under certain circumstances. Furthermore, the question of whether Akt is differentially required for CD28 and ICOS costimulation remains to be answered.
In this report, to address the above questions, we have explored the role of Akt in T cell activation costimulated by CD28 and ICOS, and in Th cell differentiation, by examining the effect of enforced expression of Akt in primary CD4 ϩ T cells under various Th-differentiating conditions.

MATERIALS AND METHODS
Animals-BALB/c and C57BL/6 mice were obtained from Japan SLC (Hamamatsu, Japan). IL-4 knock-out mice with a BALB/c background were purchased from Jackson Laboratories (Bar Harbor, ME) (30). All mice were bred and used in accordance with the guidelines of the Institute of Laboratory Animals, Tokyo Women's Medical University.
Construction of Expression Vectors-Akt cDNA was amplified by reverse transcriptase-PCR from total RNA of BALB/c mouse spleen cells by using specific primer pair 5Ј-GGCCACGGATACCATGAACG-3Ј with a BamHI linker and 5Ј-GGCCTCAGGCTGTGCCACTG-3Ј with an EcoRI linker. After double digestion with BamHI and EcoRI, the fragment was ligated into BamHI/EcoRI sites of the bicistronic retroviral vector pMX-IRES-enhanced green fluorescent protein (EGFP) (a gift from Dr. Y. Kamogawa, University of Tokyo, Tokyo, Japan) (33), and the nucleotide sequence of Akt was verified. To create the constitutively active and kinase-inactive forms of Akt (34), mutagenesis was conducted with the GeneEditor TM in vitro site-directed mutagenesis system (Promega, Madison, WI) according to manufacturer's instructions. The primers used for mutagenesis were as follows: 5Ј-ACAGGC-CGCTACTATGCCATGATGATCCTCAAGAAGGAGGTCATC-3Ј for the kinase-inactive form (K179M) and 5Ј-GGCACCTTTATTGGCTACAAG-AAACGGCCTCAGGATGTGGATCAG-3Ј for the active form (E40K) of Akt. GATA-3 cDNA (35), a gift from Dr. D. Engel (Northwestern University), was inserted at the EcoRI site of pMX-IRES-EGFP, and the retroviral expression vector for T-bet (pMSCV-T-bet) (36) was a generous gift from Dr. L. Glimcher (Harvard University).
Retrovirus-mediated Gene Transfer and in Vitro Th Cell Differentiation-Splenic CD4 ϩ T cells and APCs were prepared as described previously (13). To obtain naive T cells, splenic CD4 ϩ T cells were stained with anti-CD44 and anti-CD62L Abs and enriched by cell sorting with an Epics ALTRA flow cytometer (Coulter Immunology, Hialeah, FL). The purity of the CD44 low CD62L high cells obtained was Ͼ99%. To overexpress the Akt constructs in mouse primary CD4 ϩ T cells, retrovirusmediated gene transfer was conducted as described previously (37). Briefly, plasmid expression vectors were transfected into packaging cells for retrovirus, Plat-E (a gift from Dr. T. Kitamura, University of Tokyo) (38) by the calcium phosphate method. The culture supernatant containing recombinant retrovirus was collected at 1 or 2 days after transfection. To achieve infection, 2 ϫ 10 6 /ml CD4 ϩ T cells (0.5 ϫ 10 6 /ml for naive CD4 ϩ T cells) that had been preactivated with anti-CD3 mAb (2 g/ml) plus syngenic APCs 1 day before were incubated with viruscontaining supernatant from Plat-E supplemented with anti-CD3 (1 g/ml), polybrene (10 g/ml, Sigma), and human rIL-2 (50 units/ml, a gift from Shionogi Co., Osaka, Japan), and centrifuged for 1 h at 1900 rpm and incubated for 4 h at 37°C. Then, an equal volume of fresh Dulbecco's modified Eagle's medium was added. The following day, the cells were reinfected as above. One day later, the infected cells were expanded/rested in the presence of rIL-2 (100 units/ml) for an additional 2 days. To polarize CD4 ϩ T cells in vitro, 2.5 ng/ml of murine rIL-12 (Genzyme Techne, Framingham, MA) and 3 g/ml of anti-IL-4 mAb (11B11, ATCC, Rockville, MD) were added to the culture for Th1 cells, and 2 ng/ml of murine rIL-4 (BD Biosciences) and 3 g/ml of anti-IL-12 mAb (C17.8, Genzyme Techne) for Th2 cells, during the period of priming/infection. When required, EGFP ϩ (retrovirus-infected) cells were enriched by cell sorting. The purity of harvested EGFP ϩ cells exceeded 93%.
Stimulation, Cytokine Assay, Intracellular Cytokine Staining, and Western Blot-1 ϫ 10 5 CD4 ϩ T cells were cultured for 24 h in a 96-well flat-bottomed culture plate serially coated with anti-CD3 (0.07 g/ml) and anti-CD28, anti-ICOS, or control HIg (3 g/ml), as described previously (13). The IL-2 concentrations in culture supernatants were determined as proliferation of IL-2-dependent CTLL-2 cells in a bioassay. IL-4 and IFN-␥ in culture supernatants were quantitated by sandwich ELISA according to the manufacturer's instructions (BD Biosciences). Results are expressed as the mean Ϯ S.D. of triplicate cultures, and the data were analyzed for variance by a one-way analysis of variance followed by Tukey-Kramer's test for multiple comparisons. Results were considered significant if p Ͻ 0.01. Intracellular cytokine staining was performed as described previously (39). For surface staining, CD4 ϩ T cells collected after expansion/resting with rIL-2 for 2 days were stained with PE-labeled mAbs or combinations of biotinylated mAbs and PE-streptavidin and then analyzed with the Epics XL flow cytometer (Coulter Immunology). Transcription factors were analyzed by extracting of nuclear protein with a reagent NE-PER TM (Pierce) according to manufacturer's instruction and performing Western blot analysis as described previously (13,40).

Enforced Expression and Kinase
Activity of Akt Constructs-To investigate the role of Akt in CD4 ϩ T cells, we prepared expression constructs for three types of Akt using a bicistronic retrovirus vector and achieved retrovirus-mediated transduction of the Akt constructs into mouse primary CD4 ϩ T cells. The constitutively active form (E40K) and the wild-type (WT) of Akt were highly phosphorylated in BALB/c CD4 ϩ T cells before and after restimulation, although the phosphorylation level was slightly higher in the E40K-infected cells than in WT-infected cells (Fig. 1A). On the other hand, the kinaseinactive form (K179M) of Akt did not show enhanced phosphorylation. These observations indicate the level of phosphorylation to be proportional to the kinase activity of the Akt construct itself. The protein expression levels of Akt and EGFP in WT and E40K were also much higher than those of K179M (Figs. 1 and 2). These differences in expression level do not appear to reflect the difference in virus titer produced but do reflect the kinase activy itself because they were observed even when the infectivity with K179M was higher than that with E40K (data not shown). In addition, as a consequence of the up-regulation of kinase activity, Bcl-X L expression, which is involved in anti-apoptosis through Akt (41), was up-regulated in WT-and E40K-infected CD4 ϩ T cells from both C57BL/6 and BALB/c mice (Fig. 1B). By contrast, the activity of mitogenactivated protein kinase, ERK, was not affected by enforced expression of Akt (Fig. 1A). Consequently, the effects of Akt are limited and specific to particular downstream molecules of the Akt pathway examined.
Cellular Phenotypes Induced by the Constitutively Active Form of Akt-Following Akt infection, we gated EGFP ϩ cells to monitor infected cells and examined the phenotypes of cells with the active or the inactive form of Akt. In the cells transfected with E40K, but not K179M, increases in the number and size of cell aggregates were evident ( Fig. 2A), suggesting upregulation/activation of integrins. E40K up-regulated several activation/adhesion molecules, CD69, LFA-1, and to a lesser extent, CD44, as well as increasing cell size (forward scatter (FS)) of both BALB/c and C57BL/6 CD4 ϩ T cells, although these changes were more marked in BALB/c cells (Fig. 2, B and C, and data not shown). CD69 expression in E40K-transfected activated CD4 ϩ T cells was consistently higher than in K179Mand empty vector-transfected cells, and a high level of CD69 expression was sustained in rested CD4 ϩ T cells (Fig. 2C). In contrast, the expression levels of CD3 and CD4 did not differ between these transfectants. Consequently, these findings indicate that the Akt signal is involved in T cell activation and may account for the costimulation mediated by CD28 and ICOS.
Akt Promotes Th1 Cell Differentiation of C57BL/6 CD4 ϩ T Cells-In the next experiments, we explored the effects of Akt on cytokine production by CD4 ϩ T cells prepared from C57BL/6 mice. CD4 ϩ T cells infected with Akt constructs were enriched by cell sorting and assessed for cytokine production upon restimulation. In accordance with a previous report by another group (29), transfection with WT and E40K markedly enhanced IFN-␥ production in response to anti-CD3 plus anti-CD28 and anti-CD3 plus anti-ICOS, as compared with an empty control vector, whereas K179M had no effect (Fig. 3A). Furthermore, E40K significantly augmented IFN-␥ production upon restimulation even with anti-CD3 Ab alone, corroborating that the effects of costimulation could be achieved by Akt alone in the absence of anti-CD28 (29). IL-2 was also increased in these cells, showing the same pattern as IFN-␥. However, the effect on IL-4 production was very small or negligible.
There are two possible interpretations of the results. One is that Akt simply transactivated Th1 cytokine production, and the other is that Akt promoted Th1 differentiation resulting in elevated IFN-␥ production upon restimulation. To test these possibilities, we conducted intracellular cytokine staining after anti-CD3 plus anti-CD28 stimulation. In WT-and E40K-infected CD4 ϩ T cells, the percentage of IFN-␥-producing cells was elevated, and the ratio of IFN-␥ ϩ to IL-4 ϩ cells was much higher as compared with the control vector (Fig. 3, B and C,  top). These changes were more marked in the EGFP ϩ (retrovirus-infected) cells than in the EGFP Ϫ cells (Fig. 3C, bottom). The EGFP Ϫ cells also appeared to be affected to a lesser extent by transfected Akt through the environmental cytokines produced in the same wells in a paracrine manner. In contrast, K179M-infected cells did not show any change in numbers of IFN-␥-producing cells. The difference in the percentage of IL-4-producing cells was essentially negligible in all transfectants. In addition, ICOS and CD28 surface expression levels were each consistent among transfectants (data not shown). Consequently, these results indicate that Akt activity facilitates Th1 cell differentiation of C57BL/6 CD4 ϩ T cells.
Akt Promotes Th2 Cell Differentiation of BALB/c CD4 ϩ T Cells-We further examined whether the above findings obtained with Th1-prone C57BL/6 mice were applicable to Th2prone BALB/c mice. Unexpectedly, infection of BALB/c CD4 ϩ T cells with WT and E40K Akt constructs enhanced IL-4 production, as compared with the control, in response to anti-CD3 plus anti-CD28 (Fig. 4A). To a lesser extent, WT and E40K increased IL-4 production in response to anti-CD3 plus anti-ICOS. E40K resulted in costimulation of IL-4 production upon restimulation with anti-CD3 Ab alone. In contrast, IFN-␥ production in WT and E40K cells was reduced by about half in response to both types of stimulation. IL-2 production was also reduced in WT to the same extent as IFN-␥. K179M had no effect on cytokine production. Intracellular cytokine staining yielded essentially the same results in which the percentages of IL-4-producing cells were increased and the ratios of IFN-␥ ϩ to IL-4 ϩ cells were much lower in EGFP ϩ WT-and E40K-transfected cells than in EGFP ϩ K179M-transfected cells, although variation was observed (Fig. 4, B and C). These changes in EGFP Ϫ cells were less marked than those in EGFP ϩ cells (Fig.  4C, bottom). In addition, ICOS expression of E40K was much higher than that of K179M, although CD28 expression was slightly decreased (Fig. 4D). Moreover, expressions of the Th2specific transcription factors GATA-3 (42) and c-Maf (43) were induced in WT and E40K cells and correlated well with the amount of IL-4 produced by each transfectant (Fig. 4E). Taken together, these findings indicate enforced expression of Akt in BALB/c CD4 ϩ T cells to direct their Th2 differentiation.
Akt-mediated Th Differentiation Is Modified by the Cytokine Milieu-APCs, such as dendritic cells, have recently been reported to regulate Th cell polarization (44), and thus it was conceivable that the above results were attributable to APCs derived from a particular strain of mice and not to T cells. Alternatively, a small number of already activated (polarized) T cells remaining in the cell preparations above may have been the determinant of the above results. To address these questions, we purified C57BL/6 (B6) and BALB/c naive CD62L high CD44 low CD4 ϩ T cells by cell sorting (Fig. 5A, top) and stimulated them in the presence of C57BL/6 or BALB/c APCs (irradiated T-depleted splenocytes) during Akt infection. The results again showed that E40K up-regulated IFN-␥ in C57BL/6 naive CD4 ϩ T cells and IL-4 in BALB/c naive CD4 ϩ T cells as compared with the empty vector control (Vec), irrespective of the APC type (Fig. 5A, bottom). Thus, the results ruled out the above possibilities and indicated that enforced expression of Akt in naive CD4 ϩ T cells promoted Th differentiation. Since the effects of Akt on cytokine production and Th differentiation were opposite in C57BL/6 and BALB/c mice, the question arose as to whether the dependence of Akt on the genetic background was absolute or could be modified by extracellular cytokines. Thus, we conducted in vitro polarization experiments during Akt infection to answer this question. Under Th1-polarizing conditions in Th2-prone BALB/c CD4 ϩ T cells, E40K markedly up-regulated IFN-␥-producing cells as compared with the control vector, whereas GATA-3 consistently increased IL-4 producing cells (Fig. 5B, top). Contrary to this, under Th2-polarizing conditions in Th1-prone C57BL/6 CD4 ϩ T cells, E40K up-regulated IL-4-producing cells as compared with K179M and vector alone, while T-bet, a Th1-specific transcription factor (36), reversed the deviation and consistently up-regulated IFN-␥ producing cells (Fig. 5B, bottom). Furthermore, to test the cytokine requirement for Akt-mediated differentiation, we introduced Akt into CD4 ϩ T cells from IL-4-deficient mice with the BALB/c genetic background. We found that IFN-␥-producing cells increased among WT and E40K cells as compared with K179M cells, whereas IL-5 (another Th2 cytokine)-producing cells were unchanged (Fig. 5C). In contrast, WT and E40K in normal BALB/c mice up-regulated IL-5-producinng cells, which was consistent with IL-4 production (Figs. 4 and 5C). We therefore concluded that Akt per se is essentially neutral in regard to Th1/Th2 balance and that its promoting activity on Th differentiation requires extracellular cytokines produced by Akt in accordance with Th skewness. Akt appears to indirectly drive T cells to further Th polarization, especially when concomitant costimulation is provided.
Transcription Factors Induced by the Akt Signal-To identify the transcription factors that are located downstream of Akt and are responsible for the above phenomena, Akt-infected CD4 ϩ T cells were restimulated with plate-bound anti-CD3 Ab, and nuclear extracts were assessed for induction of the tran- scription factors. Expression of the NF-B component c-Rel in E40K-transfected cells was slightly higher than in K179M cells (Fig. 6). c-Myb expression, which has been reported to be up-regulated by Akt (45), was augmented in E40K transfectants, especially in BALB/c, as compared with K179M cells. NF-ATc1 (NF-AT2), which has been found to be involved in Th2 differ- entiation (46), was up-regulated more markedly in BALB/c E40K cells that were more deviated to Th2 direction (Fig. 4). NF-ATc2 (NF-AT1) was augmented in E40K cells of both C57BL/6 and BALB/c mice, although it is noteworthy that nuclear NF-ATc2 expression, especially of the hypophosphorylated (lower molecular weight) form, was reduced following TCR stimulation. Collectively, these findings indicate that Akt directly and/or indirectly up-regulates the expression of multiple transcription factors, some of which may be responsible for Akt-mediated T cell activation, cytokine production, and Th cell differentiation.

DISCUSSION
This study was conducted to clarify the role of PI3K/Akt signaling in T cell activation and Th cell differentiation. The results showed that enforced expression of Akt activated T and purified by cell-sorting. The purity of naive T cells obtained was Ͼ99%. These naive CD4 ϩ T cells were primed with anti-CD3 Ab in the presence of APCs (irradiated T-depleted splenocytes) derived from C57BL/6 (B6) or BALB/c mice during Akt infection. The infected T cells were restimulated with anti-CD3 (0.1 g/ml) plus anti-CD28 (1 g/ml) and assessed for cytokine production by ELISA. *, p Ͻ 0.01 as compared with an empty vector. B, BALB/c-derived CD4 ϩ T cells were infected with the retrovirus for Akt or GATA-3 in the absence or presence of supplements promoting Th1 polarization, then restimulated with anti-CD3 plus anti-CD28 and subjected to intracellular cytokine staining. The percentages positive for cytokine-producing cells among EGFP ϩ cells are presented as graphs (top). C57BL/6-derived CD4 ϩ T cells were infected with the retrovirus for Akt or T-bet with or without supplements promoting Th2 polarization. The infected CD4 ϩ T cells were subjected to intracellular cytokine staining. The results are presented as graphs (bottom). C, CD4 ϩ T cells from IL-4 deficient mice with a BALB/c genetic background and normal BALB/c mice were infected with Akt retroviruses and restimulated for intracellular cytokine staining (top). The above results are presented as graphs (bottom). These figures are representative of two independent experiments with similar results.
FIG. 5. Akt-mediated Th differentiation is modified by the cytokine milieu. A, splenic CD4 ϩ T cells prepared from C57BL/6 (B6) and BALB/c mice were stained with anti-CD62L and anti-CD44 Abs FIG. 6. Transcription factors induced by the Akt signal. C57BL/6-and BALB/c-derived CD4 ϩ T cells were infected with the retrovirus for Akt, E40K, or K179M and restimulated for 24 h with immoblilized anti-CD3 Ab (0.1 g/ml). The nuclear extracts were examined for induction of transcription factors by Western blot analysis with the specifc Abs indicated. The protein amount in each band was determined by densitometry with normalization against actin as a control, and the relative value is shown below each lane. cells, as exemplified by increases in cell size and cell aggregation, as well as up-regulation of surface activation markers (Fig. 2, B and C). However, Akt-transfected T cells failed to produce detectable cytokine levels in the absence of restimulation (data not shown), suggesting that other signal(s) act synergistically with Akt to induce cytokine secretion. These findings indicate that the PI3K/Akt pathway potently elicits or sustains, and is responsible for, T cell activation through TCR and costimulators, such as CD28 and ICOS. At the same time, Akt (WT and E40K but not K179M) increased expression of Akt protein per se and EGFP intensity, the genes of which are encoded by a retrovirus vector containing the long terminal repeat promoter, thus indicating a positive feedback mechanism. Consequently, we hypothesize that Akt overexpression activates downstream signaling molecules that result in activation of particular transcriptional factors responsible for elevated Akt/EGFP expression as well as for the activation status of T cells. In fact, we found that several transcription factors were induced in E40K-transfected cells (Fig. 6). c-Myb has been reported to be augmented by Akt (45) and to be capable of transactivating the long terminal repeat promoter of retrovirus (47). Thus it is conceivable that c-Myb is responsible for the up-regulation of Akt/EGFP.
The E40K-infected cells showing the highest Akt phosphorylation level were the best producers (or suppressors) of cytokines, indicating a good correlation between Akt activity and the amount of cytokine produced (Figs. 1, 3, and 4). Kane et al. (29) recently reported the preferential role played by Akt in Th1 cytokine, but not Th2 cytokine, production elicited by CD28. In contrast, we demonstrated Akt to up-regulate specific cytokine productions, according to Th cell type. Akt further facilitated Th1 differentiation in Th1-prone/conditioned CD4 ϩ T cells, whereas it further facilitated Th2 differentiation in Th2-prone/conditioned CD4 ϩ T cells. Moreover, Akt augmented IFN-␥-producing CD4 ϩ T cells without affecting IL-5-producing cells in IL-4-deficient mice despite their BALB/c background. We therefore conclude that Akt does not act on the genetic background directly and that Akt-mediated promotion of Th cell development requires the production of extracellular cytokines. Akt activation simply enhances cytokine production, thereby facilitating Th differentiation, depending on the intracellular status, which is primarily based on murine genetic background but can be affected by the extracellular cytokine milieu. Another possible explanation for Akt-mediated augmentation of cytokine production is the cell survival-promoting effect of Akt (41). We tested this possibility by examining cell recovery and the percentage of EGFP ϩ cells in E40K-or K179M-transfected T cell populations after the period of priming/infection. The absolute number of viable EGFP ϩ cells and the percentage of EGFP ϩ cells in E40K-transfected CD4 ϩ T cells were almost equal to, and sometimes even lower than, those in K179M-transfected CD4 ϩ T cells after primary stimulation (data not shown). Moreover, the percentage of EGFP ϩ cells in E40K-transfected T cells decreased by about half after restimulation, while that in K179M-transfected T cells was not changed (data not shown). Therefore, augmentation of cytokine production by Akt was not merely a consequence of a survivalpromoting effect of Akt but probably attributable to a direct or indirect transactivational effect of Akt on cytokine genes. In addition, we undertook this study to identify differences between CD28 and ICOS costimulation on the effects of Akt, because in our previous study engagement of ICOS elicited higher Akt phosphorylation than CD28 did (13). However, no differences in outcome were observed after costimulation via these two receptors, suggesting that Akt is not required for differential CD28 and ICOS costimulation.
Originally, ICOS was reported to be more intimately involved in Th2-type immunity than in Th1 species. This preferential involvement of ICOS could be explained in two ways. The first involves the much higher expression of ICOS in Th2 cells than in Th1 cells (4,13). The mechanism is likely to be a direct effect of elevated expression of GATA-3 in Th2 cells, which is a critical transcription factor for Th2 cytokines, since enforced expression of GATA-3 in CD4 ϩ T cells from C57BL/6 and IL-4-deficient mice increased ICOS expression (48). The second involves signaling via ICOS mediated by cross-linking with specific Abs or by blockade with ICOS-Ig fusion protein during priming, which can directly control the appearance of Th2 cells (13,49). In Th2-prone BALB/c-activated CD4 ϩ T cells, Akt is more strongly elicited via ICOS than in Th1-prone C57BL/6 activated CD4 ϩ T cells (13). In this study, GATA-3 and c-Maf expressions, which are induced by IL-4, and expression of NF-ATc1, which is involved in Th2 differentiation (3,46), were markedly elevated in WT-and E40K-transfected cells from BALB/c mice (Figs. 4D and 6). Taken together, these observations raise the possibility that once CD4 ϩ T cells start to differentiate into Th2 cells, ICOS expression is gradually upregulated, thereby providing stronger activation of Akt than in Th1 cells and in turn leading to an increase in IL-4 production. Consequently, Akt could be a driving factor toward further Th2 cell maturation, through ICOS engagement, the expression of which was elevated in this lineage, in a positive feedback manner, although Akt per se is a neutral player in Th differentiation.
Absence or blockade of CD28 can lead to global immunosuppression and/or a selective limitation in the generation of normal Th responses. CD28-deficient mice show reduced basal levels of IgG1 and IgG2b, increased IgG2a, and inefficient help to B cells (50,51). In addition, IL-4 production and, to a lesser extent, IFN-␥ production have been shown to be regulated by CD28-mediated signals (52,53). Thus, CD28 costimulation was predicted to be required for maturation of Th2 cells. However, the relevance of CD28 to Th differentiation is limited, and CD28 is not essential (54). The CD28 requirement for Th differentiation varies according to the genetic background and the nature of the immunogen (54). Linsley's group (55) recently reported that TCR and CD28 signals up-regulated a similar set of transcripts, suggesting the major consequence of CD28 costimulation to be quantitative rather than qualitative. PI3K/ Akt activation is, in fact, elicited by TCR as well as CD28, and its activity is most potent when cross-linked with both receptors, thereby indicating that PI3K/Akt is one of convergence points for two signaling pathways. In the current study, the effect of Akt was found to be influenced by genetic background as well as the cytokine milieu. Consequently, Akt is not a critical determinant of Th cell development and concomitant cytokine production. Akt thus appears to provide quantitative, but not qualitative, effects on T cell function.
What are the downstream targets for Akt, via which it activates T cells and up-regulates cytokine secretion at the initial step in each Th cell type? A large number of molecules have been reported for substrates of Akt, in which consensus sequences for phosphorylation are RXRXXS/T (21,22). In this study, multiple transcription factors were found to be up-regulated by E40K of Akt. Among them, NF-AT members are well known to be crucial to the production of various cytokines (46,56), thus suggesting the possibility that Akt-mediated cytokine production and Th differentiation are achieved through these transcription factors. NF-ATc1 plays an important role in Th2 cytokine expression (46). On the other hand, NF-ATc2 was originally reported to down-regulate Th2 cytokines (57); however, a recent report has shown that NF-ATc2 is a positive regulator of IL-4 (58). Thus, NF-ATc1 and NF-ATc2 appear to be differentially regulated in T cells (58). Consistent with this, in the current study, the induction of expression of these two molecules by E40K and their translocations upon TCR restimulation were also different (Fig. 6). The precise mechanism by which Akt directly or indirectly activates/up-regulates NF-ATs and c-Myb expression remains unknown and is now under investigation. Screening of the other target molecules of Akt is also in progress.