T Lymphocyte Activation Signals for Interleukin-2 Production Involve Activation of MKK6-p38 and MKK7-SAPK/JNK Signaling Pathways Sensitive to Cyclosporin A*

p38/CSBP, a subgroup member of mitogen-activated protein kinase (MAPK) superfamily molecules, is known to be activated by proinflammatory cytokines and environmental stresses. We report here that p38 is specifically activated by signals that lead to interleukin-2 (IL-2) production in T lymphocytes. A p38 activator MKK6 was also markedly activated by the same stimulation. Pretreatment of cells with SB203580, a specific inhibitor of p38, as well as expression of a dominant-negative mutant of MKK6, suppressed the transcriptional activation of the IL-2 promoter. We also demonstrated that MKK7, a recently described MAPK kinase family member, plays a major role in the activation of stress-activated protein kinase (SAPK)/c-Jun NH2-terminal kinase (JNK) in T lymphocytes. Moreover, a dominant-negative mutant of MKK7 abrogated the transcriptional activation of the distal nuclear factor of activated T cells response element in the IL-2 promoter. Cyclosporin A, a potent immunosuppressant, inhibited activation of both p38 and SAPK/JNK pathways but not the MAPK/extracellular signal-regulated kinase (ERK) pathway. Our results indicate that both MKK6 to p38 and MKK7 to SAPK/JNK signaling pathways are activated in a cyclosporin A-sensitive manner and contribute to IL-2 gene expression in T lymphocytes.

IL-2 production, which is a critical step for T lymphocyte activation and proliferation, requires two coordinate signals; the primary stimulus is the interaction of antigen with the T cell receptor (TCR)⅐CD3 complex (21), and the secondary stimulus, called costimulation, is generated by the interaction of CD28 auxiliary receptors on T cells with its ligand on antigenpresenting cells (22). Various studies indicate that activation signals for IL-2 production can be bypassed by phorbol esters (such as TPA) and Ca 2ϩ ionophore (such as A23187) (23)(24)(25).
Here we report that either treatment with TPA and Ca 2ϩ ionophore or simultaneous activation of the TCR and CD28 results in synergistic activation of MKK6 (26 -28) to the p38 pathway in T lymphocytes in a cyclosporin A (29)-sensitive manner. We also demonstrate that newly identified MKK7 (30 -34) plays a major role as an upstream activator of SAPK/ JNK in T cells. Several lines of evidence for involvement of p38 and SAPK/JNK in IL-2 gene expression are presented.

EXPERIMENTAL PROCEDURES
Activation of Cells-Jurkat cells were grown in RPMI 1640 with 10% fetal calf serum, 50 units/ml penicillin, 200 g/ml kanamycin, and 5 ϫ 10 Ϫ5 M ␤-mercaptoethanol. Monoclonal antibodies (mAbs) to human CD3 (NU-T3) and human CD28 (KOLT-2) were purchased from Nichirei. PD98059, SB203580, and CsA were purchased from New England Biolabs, Calbiochem,and Wako, respectively. mAb NU-T3 (5 g) was immobilized on a 35-mm dish by incubation for 12 h at 4°C. mAb KOLT-2 (5 g) was preincubated with 5 g of rabbit anti-mouse antibodies for 12 h at 4°C to induce cross-linking and then was added to Jurkat cells (2 ϫ 10 6 ), or cells were stimulated with various reagents such as TPA (Sigma) and Ca 2ϩ ionophore A23187 (Wako). After a 24-h treatment of Jurkat cells with various stimuli, IL-2 production was measured using ELISA (Endogen) according to the manufacturer's instructions.
Immune Complex Kinase Assay-Immunoprecipitation and immune complex kinase assays were performed as described previously (34,35). Anti-p38 antibody was purchased from Santa Cruz. To assay for p38 activity, p38 was immunoprecipitated from cell lysates in the presence of 0.1% SDS and assayed for kinase activity using ATF-2 as a substrate. Anti-SEK1 (36), anti-MKK6 (28,29), and anti-MKK7 (34) antibodies were previously produced and characterized.
Plasmids-pME-MKK6, an expression plasmid for hemagglutinin (HA)-tagged MKK6 under a control of the SR␣ promoter, has previously been characterized (28). A dominant-negative MKK6 (MKK6(AA)) was constructed by replacing Ser-207 and Thr-211 with alanine residues. The mutation was made by site-directed mutagenesis (Quick Change TM , Stratagene) and was confirmed by DNA sequencing. An HA-tagged kinase-negative MKK7 (MKK7(KL)), constructed by replacing ATP-binding lysine with leucine residue, has previously been reported (34). A luciferase reporter plasmid carrying the human IL-2 enhancer/promoter element (encompassing nucleotides Ϫ326 to ϩ47), IL-2-Luc, was kindly provided by Dr. M. Iwashima (Mitsubishi Chemical Corporation Yokohama Research Center, Japan). Another luciferase reporter plasmid with three tandem repeats of the NFAT site, NFAT-Luc (37), was kindly provided by Dr. G. Crabtree (Stanford University, Stanford, CA).
Transfection and Luciferase Assay-Transient transfection of Jurkat cells (5 ϫ 10 6 ) was performed using SuperFect TM (Qiagen) with luciferase reporter plasmids (either 1 g of IL-2-Luc or 0.5 g of NFAT-Luc) along with various amounts of expression plasmids. The total amount of DNA transfected was maintained at 5 g with pcDNA 3.1ϩ (Invitrogen). After 24 h of incubation at 37°C, cells were stimulated for 8 h with indicated reagents and then lysed for luciferase assay. Luciferase in cell lysates was measured in duplicate by a luminometer (LB9507, Berthold), using a luciferase assay system (Promega). The protein concentration was determined with a protein assay kit (Bio-Rad) and used for normalization of the luciferase assays. Expression of HA-tagged proteins was analyzed by immunoblotting with the 12CA5 mouse mAb against the HA epitope.

RESULTS AND DISCUSSION
Activation of p38 and MKK6 by Costimulatory Signals-In a human T cell line, Jurkat, the activity of p38 markedly increased in response to signals generated by TPA and Ca 2ϩ ionophore, whereas each stimulus alone resulted in little or no activation (Fig. 1A, shaded bar). Similar to its effect on IL-2 production (37), CsA inhibited the activation of p38 (Fig. 1A, shaded bar). SAPK/JNK was also activated in a similar manner as reported previously (Fig. 1A, black bar (9)). In contrast, incubation with TPA was sufficient for efficient activation of MAPK/ERK, which was neither enhanced by Ca 2ϩ ionophore nor inhibited by CsA (Fig. 1A, white bar). Interestingly, treatment with TPA and Ca 2ϩ ionophore did not induce the activation of p38 in non-lymphoid cells such as an epidermoid carcinoma cell line, KB (data not shown). We next examined whether MKK6, an upstream activator of p38 (27)(28)(29)38), is activated in parallel with p38 activation during T cell activation. Whereas TPA alone had little effect on MKK6 activity, it induced marked activation of MKK6 in combination with Ca 2ϩ ionophore (Fig. 1B). Furthermore, CsA inhibited the activation of MKK6 as observed in the p38 activation (Fig. 1B). Although treatment with Ca 2ϩ ionophore alone induced slight activation of MKK6 in the Jurkat cells used here (Fig. 1B), this is not the case with another Jurkat line, TAg Jurkat (data not shown). The reason for this inconsistency remains to be determined. Whereas incubation of Jurkat cells with anti-CD3 or anti-CD28 mAb alone had little effect on either p38 activity or MKK6 activity, simultaneous incubation with both mAbs resulted in synergistic activation of both MKK6 and p38 and IL-2 produc-tion (Fig. 1C). Activation of MKK6 and p38 induced by simultaneous stimulation with anti-CD3 and anti-CD28 mAbs was also inhibited by CsA (data not shown).
Activation of MKK7 during T Cell Activation-One of the MAPKKs, SEK1 (also termed as MKK4 or JNKK) has previously been reported to function as a direct activator of SAPK/ JNK (19,39,40). In addition to SEK1, we have previously identified an activity to phosphorylate and activate SAPK distinct from SEK1 (36). This novel kinase has been identified by cDNA cloning and named MKK7 (34). We thus examined whether SEK1 or MKK7 plays a major role in SAPK/JNK activation during T cell activation. Whereas incubation of Jurkat cells with either anti-CD3 or anti-CD28 mAb alone had a weak effect on MKK7 activity, simultaneous incubation with both mAbs resulted in strong synergistic activation of MKK7, which was inhibited in the presence of CsA ( Fig. 2A). SEK1 activity was only slightly activated in response to TPA and Ca 2ϩ ionophore, whereas activation of MKK7 activity was more prominent to the same stimulation (Fig. 2B), which was also sensitive to CsA (data not shown). Note that both SEK1 and MKK7 activities markedly increased when Jurkat cells were exposed to hyperosmolar media as expected (Fig. 2B) (34, 36). were incubated at 37°C with no addition (none), with 10 ng/ml TPA (T), with 1 g/ml Ca 2ϩ ionophore A23187 (I), or with the combination of TPA and Ca 2ϩ ionophore (TϩI). A single group of cells was pretreated with 100 ng/ml CsA for 15 min before the TPA and Ca 2ϩ ionophore incubation, and remained in CsA for the entire experiment (TϩIϩCsA). At the end of a 20-min incubation period, a portion of cells (4 ϫ 10 6 ) was assayed for MAPK/ERK (white bar), SAPK/JNK (black bar), and p38/CSBP (shaded bar) activities (upper panel). The rest of cells were incubated for another 24 h, and then IL-2 production was measured by examining culture supernatants using ELISA (lower panel). This experiment is a representative of three. B, Jurkat cells (1 ϫ 10 7 ) were treated as above and at the end of a 20-min incubation period, MKK6 was immunoprecipitated from cell lysates and assayed for kinase activity toward p38/CSBP as described (28). This experiment is a representative of three. C, Jurkat cells (5 ϫ 10 6 ) were incubated for 20 min at 37°C with no addition (none), immobilized mAb to CD3 (CD3), cross-linked mAb to CD28 (CD28), or the combination (CD3ϩCD28). Cell lysates were prepared and assayed for p38/CSBP activity (black bar) and MKK6 activity (shaded bar) using immune complex kinase assays (left panel). IL-2 production was measured by examining supernatants of similarly treated Jurkat cells (after 24 h of treatment) using ELISA (right panel). This experiment was repeated twice with similar results.
Similar results were obtained in mouse thymocytes (Fig. 2C) and splenic T cells (data not shown). These results suggest that MKK7 acts as a major upstream activator of SAPK/JNK during T cell activation, as is the case with tumor necrosis factor-␣ and Fas signaling pathway (34,41). It has recently been reported that SAPK/JNK activation is not induced in thymocytes of SEK1-deficient mice in response to TPA and Ca 2ϩ ionophore (42,43). The activation pathway of SAPK/JNK might be developmentally regulated such that SEK1 but not MKK7 activates SAPK/JNK in the thymus. Functions of SEK1 and MKK7 in SAPK/JNK pathway in thymocytes remain to be investigated in further studies.
Inhibition of MAPK Superfamily Pathway Blocks the Transcriptional Activation of IL-2 Gene-Activation of MAPK/ERK alone was not sufficient to induce IL-2 production (Fig. 1A) (9). To evaluate the role of MAPK/ERK in IL-2 production directly, we employed PD98059, a specific inhibitor of classical MAPKK (also termed as MEK) (44,45). As shown in Fig. 3A, preincubation with PD98059 prevented IL-2 production induced by simultaneous stimulation with TPA and Ca 2ϩ ionophore in a dose-dependent manner. The concentration of PD98059, which inhibits IL-2 production by 50% in Jurkat cells (Ͻ5 M), is similar to that inhibiting activation of MAPKK/MEK in vitro and in vivo (44). We also found that IL-2 production induced by stimulation via CD3 and CD28 was almost completely inhibited by pretreatment of the cells with 20 M PD98059 (Fig. 3B). Furthermore, prevention of MAPK/ERK activation resulted in the suppression of IL-2 promoter-driven transcriptional activity (data not shown (46)). It is noteworthy that preteatment of the cells with 20 M PD98059 did not prevent the activation of SAPK/JNK or p38 (data not shown). These results clearly indicated that the activation of MAPK/ERK pathway is required for IL-2 production.
To evaluate the biological significance of p38 activation in T lymphocytes, we examined the effect of SB203580, a highly specific inhibitor for p38 (47,48), on the IL-2 promoter-driven transcriptional activity. As shown in Fig. 4A, pretreatment of the Jurkat cells with SB203580 resulted in the reduction of IL-2 promoter-driven luciferase (IL-2-Luc) activity in a dosedependent manner. IL-2 production induced by simultaneous stimulation with anti-CD3 and anti-CD28 mAbs was also inhibited by SB203580 by 50% (data not shown). We further investigated the effect of SB203580 on the transcriptional activation driven by the distal NFAT site in the IL-2 promoter, which is essential for full transcriptional activation of the IL-2 gene (for review see Ref. 49). NFAT-driven luciferase (NFAT-Luc) (37) activity was increased after stimulation with TPA and Ca 2ϩ ionophore as previously shown (50), and pretreatment of the cells with SB203580 led to marked reduction of luciferase activity (Fig. 4B). The same concentration of SB203580 used in this study does not inhibit the activity of SAPK/JNK or MAPK/ERK (Refs. 47 and 48 and data not shown). Furthermore, high concentrations of SB203580 (up to 10 M) had little effect on SV40 promoter-driven luciferase activity (data not shown). The specificity of SB203580 indicates that the prevention of the IL-2 gene expression resulted from the blockade of p38 activity. To further elucidate whether the p38 signaling pathway functions in the IL-2 gene expression, we examined the effect of a dominant-negative form of MKK6 (27), referred to here as MKK6(AA). The expression of MKK6(AA) reduced the NFAT-Luc activity (Fig. 4C) and the IL-2-Luc activity (data not shown) induced by TPA and Ca 2ϩ ionophore (Fig. 4C) and by anti-CD3 and anti-CD28 stimulation (data not shown), whereas the expression of wild type MKK6 slightly increased the NFAT-Luc activity (Fig. 4C). These data clearly indicate that p38 is significantly involved in the signaling pathway leading to IL-2 gene expression.
It was noted that neither SB203580 nor MKK6(AA) showed complete inhibition of the NFAT-or IL-2-Luc activity. It is possible that recently identified molecules, SAPK3 and SAPK4 (51, 52), which are resistant to SB203580 treatment, might contribute to the residual transcriptional activity in the pres-

FIG. 2. Activation of MKK7 during T cell activation.
A, Jurkat cells (5 ϫ 10 6 ) were incubated for 15 min at 37°C with no addition (none), immobilized mAb to CD3 (CD3), cross-linked mAb to CD28 (CD28), or the combination (CD3ϩCD28). A single group of cells was pretreated with 100 ng/ml CsA for 15 min before the CD3ϩCD28 treatment and remained in CsA for the entire incubation period (CD3ϩCD28ϩCsA). At the end of a 15-min incubation period, MKK7 was immunoprecipitated from cell lysates and assayed for SAPK activating activity as described previously (34). B, Jurkat cells (1 ϫ 10 7 ) were left untreated (none), incubated with the combination of TPA and Ca 2ϩ ionophore (TϩI) for 15 min, or exposed to 0.7 M NaCl for 50 min (NaCl). At the end of incubation periods, SEK1 and MKK7 were immunoprecipitated from cell lysates and assayed for SAPK activating activity. C, thymocytes (1 ϫ 10 7 ) were isolated from 8-week-old BALB/c mice and incubated with (TϩI) or without (none) TPA and Ca 2ϩ ionophore. At the end of a 15-min incubation period, SEK1 and MKK7 were immunoprecipitated and assayed for SAPK activating activity. This experiment is a representative of two.
FIG. 3. Prevention of IL-2 production by PD98059. A, Jurkat cells (2 ϫ 10 6 ) were preincubated for 60 min at 37°C with different concentrations of PD98059 (0 -30 M) followed by incubation with 10 ng/ml TPA and 1 g/ml Ca 2ϩ ionophore A23187. After 24 h of stimulation, IL-2 production was examined with supernatants of the cells by ELISA. This experiment is a representative of three. B, Jurkat cells (2 ϫ 10 6 ) were preincubated for 60 min at 37°C with (ϩPD98059 (20 M)) or without (control) 20 M PD98059 followed by incubation with immobilized mAbs to CD3 and CD28. After 24 h of stimulation, IL-2 production was examined with supernatants of Jurkat cells by ELISA. This experiment is a representative of two. Although not shown, PD98059 did not affect the cell growth or survival. ence of SB203580. This possibility seems unlikely, however, since a dominant-negative mutant of MKK6, which can also function upstream of SAPK3 and SAPK4, had the same effect as SB203580 as shown in Fig. 4C. One possibility is that p38 and SAPK/JNK exhibit redundant functions in vivo. To test this possibility, the roles of MKK7 and SAPK/JNK in the signaling pathway leading to IL-2 gene expression were examined. We employed a kinase-negative mutant of MKK7 (referred to here as MKK7(KL)), which can inhibit the activation of SAPK/JNK induced by tumor necrosis factor-␣ or Fas (34,41). In Jurkat cells, the expression of MKK7(KL) resulted in the inhibition of the NFAT-Luc activity induced by TPA and Ca 2ϩ ionophore (Fig. 4D) and by anti-CD3 and anti-CD28 stimulation (data not shown) in a dose-dependent manner. In this case, too, MKK7(KL) showed only partial inhibition of the NFAT-Luc activity. We thus evaluated the effect of SB203580 on the NFAT-Luc activity in MKK7(KL)-transfected Jurkat cells. Interestingly, a blockade of both signaling pathways further reduced the NFAT-Luc activity to near basal level (Fig.  4E). It is therefore likely that the MKK6 to p38 and the MKK7 to SAPK/JNK cascades comprise redundant signaling pathways leading to the IL-2 gene expression. This conclusion is further supported by the fact that p38 and SAPK/JNK have a similar specificity of substrate recognition in vitro, although MAPK/ERK displays a distinct substrate selectivity (26,53,54). Further studies on downstream targets of p38 and SAPK/ JNK in T lymphocytes should elucidate mechanisms by which the p38 signaling pathway affects IL-2 production.
It has been widely accepted that CsA binds to cytosolic protein, cyclophilin, and that the CsA-cyclophilin complex blocks the IL-2 production by inhibiting the activity of Ca 2ϩ calmodulindependent protein phosphatase, calcineurin (55,56). We found that activation of MKK7 and MKK6, but not MAPKK/MEK, was inhibited by CsA (Fig. 1B, Fig. 2, and data not shown). It is possible that calcineurin positively regulates the activity of MKK7 and MKK6 through a mechanism yet to be determined. Alternatively, CsA-cyclophilin complex may directly inactivate an upstream activator of these MAPKKs. Nevertheless, current studies show that the signaling pathways mediated by SAPK/ JNK and p38/CSBP could be alternative targets of CsA. Precise mechanisms of the inhibitory effect of CsA on SAPK/JNK and p38/CSBP are now under investigation. or stimulated (stimuli: ϩ) with 10 ng/ml TPA and 1 g/ml Ca 2ϩ ionophore A23187. After 8 h of incubation, cells were lysed, and the soluble extracts were assayed for luciferase activity. These experiments are representative of two. C and D, Jurkat cells were transiently transfected with the NFAT-Luc plasmid along with the indicated amounts of wild type MKK6, MKK6(AA) (C), or MKK7(KL) (D). Transfected cells were cultured for 24 h followed by incubation with (TϩI) or without (none) 10 ng/ml TPA and 1 g/ml Ca 2ϩ ionophore A23187 for a further 8 h. The cells were then lysed, and the soluble extracts were assayed for luciferase activity. Four independent experiments were performed, and data are presented as means Ϯ S.E. Expression of each construct was analyzed by immunoblot with anti-HA mAb (inset in C and D). E, Jurkat cells were transiently transfected with the NFAT-Luc plasmid with or without 2 g of MKK7(KL) and cultured for 24 h followed by incubation with or without 5 M SB203580 for another 1 h. The cells were then incubated with or without 10 ng/ml TPA and 1 g/ml Ca 2ϩ ionophore A23187. After 8 h of incubation, cells were lysed, and the soluble extracts were assayed for the luciferase activity. This experiment is a representative of two.