J. Exp. Med.
© The Rockefeller University Press
0022-1007/97/09/941/13 $2.00
Volume 186, Number 6, September 15, 1997 941-953

Impaired CD28-mediated Interleukin 2 Production and Proliferation in Stress Kinase SAPK/ERK1 Kinase (SEK1)/Mitogen-activated Protein Kinase Kinase 4 (MKK4)-deficient T Lymphocytes

By Hiroshi Nishina,^* Martin Bachmann, Antonio J. Oliveira-dos-Santos,^* Ivona Kozieradzki,^* Klaus D. Fischer,^§ Bernhard Odermatt,^¶ Andrew Wakeham,^* Arda Shahinian,^* Hiroaki Takimoto,^* Alan Bernstein, Tak W. Mak,^* James R. Woodgett, Pamela S. Ohashi, and Josef M. Penninger^*

From the ^* Amgen Institute and Ontario Cancer Institute, Department of Medical Biophysics and Immunology, University of Toronto, M5G 2C1 Toronto, Ontario, Canada;  Ontario Cancer Institute, Department of Medical Biophysics and Immunology, University of Toronto, Toronto, Ontario, Canada; ^§ Institute for Radiation and Cell Research, University of Wuerzburg, D-97078 Wuerzburg, Germany;  Samuel Lunenfeld Research Institute and Department of Medical Genetics, University of Toronto, Mount Sinai Hospital, Toronto, Ontario, Canada; and ^¶ Institute for Experimental Immunology, University of Zürich, 8091, Zürich, Switzerland

Summary

Materials and Methods

Results

Discussion

Footnotes

Acknowledgements

References

Summary

Distinct and evolutionarily conserved signal transduction cascades mediate survival or death in response to developmental and environmental cues. Multiple stimuli for differentiation and cell growth activate the mitogen-activated protein kinases (MAPKs)¹, also known as the extracellular signal-regulated kinases ERK1 and ERK2 (1), which translocate to the nucleus and regulate the activity of transcription factors (5). MAPKs are activated by the phosphorylation of a threonine and a tyrosine residue mediated by the dual specificity MAPK kinases MAPK/ERK kinase (MEK)1 and MEK2, which relay Ras and Raf signal transduction to MAPK activation (6).

A second signaling cascade exists in all cells that leads to the activation of stress-activated protein kinases (SAPKs) or Jun NH₂ terminal kinase (JNKs; 9,10). The SAPK signaling cascade is parallel and independent from MAPK activation (11, 12). SAPKs/JNKs are activated in response to a variety of cellular stresses such as changes in osmolarity and metabolism, DNA damage, heat shock, ischemia, inflammatory cytokines, or ceramide (13). Activated SAPKs/JNKs phosphorylate c-Jun, which leads to activation of the transcriptional complex AP-1 (19). SAPKs/JNKs are activated by the phosphorylation of tyrosine and threonine residues, which is catalyzed by the dual specificity kinase SAPK/ ERK kinase (SEK)1 (also known as MAPK kinase [MKK4] and JNK kinase; 20-22). In addition to SEK1, a novel SAPK activator (SEK2 or MKK7) has been genetically identified but has not been cloned yet (23).

It has been proposed from transfection studies with dominant negative signaling mutants that the SEK1 SAPK/ JNK c-Jun signaling cascade is a common intracellular pathway required for the induction of apoptosis in response to many types of cellular stresses (16, 24). However, recent genetic evidence suggests that SEK1 and SEK1-mediated SAPK activation have no role in the induction of cell death in lymphocytes, but rather protect T cells from CD95 (FAS) and CD3-mediated apotosis (23). The SAPK/ JNK signaling cascade is also triggered by certain growth stimulating factors and phorbol esters (9, 14, 29, 30). In B cells, SEK1 and SAPK are activated in response to CD40 cross-linking (31, 32) and by the human STE20 homologue germinal center kinase (GCK) (33). The prominent expression of GCK in germinal centers (34) suggested that the GCK/SAPK pathway might be important for B cell differentiation or activation. Moreover, biochemical studies in T cells indicated that SAPKs/JNKs are involved in the integration of TCR-CD3 and CD28 costimulatory signals required for proliferation and IL-2 production (29, 35). Failure to activate SAPKs/JNKs in T cells may result in clonal anergy (36, 37).

To determine the role of SEK1 in B cell function and CD28-mediated costimulation, we reconstructed T (23) and B cell development in SEK1 gene-deficient chimeras using recombination-activating gene (RAG)2 blastocyst complementation. We show that SEK1 is important for CD28-mediated costimulation for T cell proliferation and IL-2 production. B lymphocyte development was partially impaired. However, peripheral B cells displayed normal responses to IL-4 and to IgM and CD40 cross-linking, and exhibited normal IgG class switching after vesicular stomatitis virus (VSV) infections. Moreover, we show that CD28, but not SEK1, is crucial for VSV-specific germinal center formation. Interestingly, using the same activation regimen, i.e., PMA plus Ca²⁺ ionophore which mimics TCR-CD3- and CD28-mediated signal transduction (29), SAPK activation was observed in peripheral T cells, but not in thymocytes, from SEK1^/ mice. These data provide the first genetic evidence that SEK1-regulated stress signal transduction has a role in CD28 costimulation for IL-2 production and proliferation. These results also show that signaling pathways for SAPK activation are developmentally regulated in T cells.

Materials and Methods

Mice. The generation of embryonic stem (ES) cells homozygous for the SEK1 mutation, SEK1^/ somatic chimeras using RAG2^/ blastocyst complementation (23, 38), and CD28^/ mice (39) have been previously described. Since E14 ES cells are derived from a 129/J mouse background, age-matched 129/J mice were used as wild-type controls. T and B cells from SEK1^/ RAG2^/ mice were tested for the SEK1 mutation using PCR (sense primer: 5-ACAGCAAATTTTGGAAACAGC-3; antisense primer: 5-CTCCCCTACCCGGTAGAATTC-3). All data presented throughout this study were obtained from two independently derived SEK1^/ ES cell clones (No. 1-6 and No. 1-21), and all results were comparable between them. If not otherwise stated, all mice used for experiments were between 6 and 10 wk old. Mice were kept under pathogen-free conditions in accordance with guidelines of the Canadian Medical Research Council.

Immunocytometry. Single cell suspensions from thymocytes, spleen cells, mesenteric lymph node cells, and bone marrow cells from SEK1^/RAG2^/ chimeric, SEK1^+/+RAG2^/ chimeric, RAG2^/, and 129/J mice were prepared as described (40), resuspended in immunofluorescence-staining buffer (PBS, 4% FCS, 0.1% NaN₃) and incubated with appropriate Abs. The following mAbs were used: anti-CD4 (FITC-, or PE-labeled), anti-CD8 (FITC-labeled, PE-labeled, or biotinylated), anti-TCR/ (FITC-, or PE-labeled), anti-CD3- (FITC-labeled), anti-B220 (FITC-labeled, PE-labeled, or biotinylated), anti-CD43 (FITC-labeled), anti-CD25/IL-2R- (biotinylated), anti-H2K^b (FITC-labeled), anti-CD44 (PE-labeled), anti-FAS (PE-labeled, or biotinylated), anti-intercellular cell adhesion molecule 1 (ICAM-1; biotinylated); anti-CD23 (PE-labeled), anti-CD28 (PE-labeled), anti-CTLA-4 (PE-labeled), anti-CD69 (FITC-labeled), anti-CD40L (gp39; biotinylated) (all above Abs were from PharMingen, San Diego, CA); anti-surface (s)IgM (clone B67; FITC-labeled, gift of C. Paige, Ontario Cancer Institute, Toronto, Canada), anti-sIgD (PE-labeled; gift of C. Paige), and anti-CD40 (FITC- labeled; Serotec, Toronto, Canada). All staining combinations were as indicated in the figure and table legends. Biotinylated Abs were visualized using Streptavidin-RED670 (Life Technologies, Burlington, Canada). Samples were analyzed using a FACScan^® (Becton Dickinson, Mountain View, CA).

Cell Sorting. Bone marrow cells were isolated from RAG2^/, SEK1^/RAG2^/ chimeric, SEK1^+/+RAG2^/ chimeric, and 129/J control mice and double stained for CD43 and B220 expression using anti-CD43-FITC and anti-B220-PE. CD43⁺B220⁺ and B220⁺CD43 bone marrow B cell polulations (Fig. 2) were sorted using a FACS^® power sorter (FACS^® Vantage). In all experiments, postsorting purity of CD43⁺B220⁺ and B220⁺CD43 populations was >98%. Sorted cells were analyzed for the SEK1 mutation using PCR (see above).

/

/

/

/

/

/

/

Table 1. T and B Cell Subpopulations in SEK1/ Chimeric Mice Percentages ± SEM of positive cells per total cells Cell subset 129/J SEK1+/+ SEK1/ RAG2/ Thymus   CD4+CD8+ 81.8 ± 2.2     84 ± 0.9 57.7 ± 4.2 0   CD4+CD8 11.6 ± 0.8  7.7 ± 0.4 27.2 ± 2.8 0   CD4CD8+  3.2 ± 0.4  2.1 ± 0.7  7.5 ± 1.1 0   CD4CD8  3.5 ± 0.6  6.1 ± 1.0 4.8 ± 0.4 100 Lymph node   CD4+CD8 48.2 ± 3.2 44.2 ± 4.3 55.1 ± 3.1 0   CD4CD8+ 19.0 ± 2.8 20.2 ± 3.7 22.1 ± 1.8 0   B220+slgM+ 30.5 ± 1.6 26.6 ± 3.5 12.0 ± 3.0 0 Spleen   CD4+CD8 24.0 ± 5.3 22.6 ± 5.3 33.2 ± 2.8 0   CD4CD8+ 11.3 ± 2.0  8.9 ± 2.7 13.5 ± 1.5 0   B220+slgM+ 33.5 ± 5.5 32.6 ± 3.5 16.7 ± 1.6 0 Bone marrow   B220+CD43+ 11.9 ± 1.4 10.6 ± 2.5 13.4 ± 0.4 16.5 ± 3.5   B220+CD43 30.8 ± 2.5 34.5 ± 4.5 11.6 ± 1.6 <1   B220+CD25+ 30.5 ± 2.1 32.0 ± 3.3 10.2 ± 1.5 <1   B220+CD25 10.5 ± 0.9 15.1 ± 2.7 12.6 ± 1.2 14.6 ± 4.1 Cells from thymi, mesenteric lymph nodes, spleens, and bone marrow (one femur) from 129/J (n = 6), SEK1+/+ chimeric (n = 3), SEK1/ chimeric (n = 6), and RAG2/ (n = 5) mice were stained with the indicated Abs and populations determined using a FACScan®. Total cell numbers (× 106 ± SEM) were: 129J: thymus, 80 ± 6.3; LN, 24 ± 2.2; spleen, 21± 3.4; BM, 7.8 ± 0.7; SEK1+/+: thymus, 85 ± 10.1; LN, 22 ± 4.6; spleen, 32 ± 3.14; BM, 8.6 ± 2.5; SEK1/: thymus, 14 ± 4.1; LN, 29 ± 4.2; spleen 27 ± 5.3; BM, 10.6 ± 0.4; RAG2/: thymus, 1.1 ± 0.2; LN, 0.7± 0.2; spleen, 8.3 ± 1.6; BM 9.2 ± 0.9 (reference 23). Bold numbers indicate, statistically significant differences between 129/J or SEK1+/+ and SEK1/ subpopulations (Student's t test P <0.05).

B and T Cell Stimulation Assays. Lymph node T cells were negatively enriched from lymph nodes of SEK1^/RAG2^/ chimeric and SEK1^+/+RAG2^/ chimeric mice using affinity columns (R&D Sys. Inc., Minneapolis, MN) to avoid receptor cross-linking during the purification process. Purified (>95%) T cells (10⁴) and freshly isolated thymocytes were placed into round-bottom 96-well plates (Costar, Fisher Scientific, Unionville, Canada) in freshly prepared IMDM (10% FCS, 10⁵ M  mercaptoethanol) and activated with PMA (12.5 ng/ml) plus Ca²⁺ ionophore A23617 (100 ng/ml), plate-bound anti-CD3- (clone 145-2C11, hamster IgG; PharMingen), soluble anti-CD3- (clone 145-2C11), and soluble anti-CD28 (clone 37.51, hamster IgG; gift of Dr. J. Allison, University of California, Berkeley, CA). PMA/Ca²⁺ ionophore and mAbs were added at optimal concentrations determined in pilot studies. For CD3 cross-linking, plates were coated overnight (4°C) with 10 µg/well of rabbit anti-hamster IgG (Jackson Labs., West Grove, PA), and subsequently with anti-CD3- (37°C for 2 h, clone 145-2C11).

/

/

/

CD40 Cross-linking. For CD40-mediated upregulation of ICAM-1 and CD23 (42), purified B cells were activated with anti-CD40 (2 µg/ml; Serotec) in the absence or presence of IL-4 (50 U/ml) in IMDM (10% FCS, 37°C). After 24 h of activation, cells were harvested and triple stained with Abs reactive against B220 (PE), sIgM (FITC), and ICAM-1 (biotin) or CD23 (biotin). Biotinylated Abs were visualized using Streptavidin-RED670 and staining of cells was analyzed using a FACScan^®.

Detection of Ig-subclasses. Sera were collected from 6-wk-old individual SEK1^/RAG2^/ and SEK1^+/+RAG2^/ chimeric mice. The concentrations of Ig subclasses were determined by ELISA with isotype-specific, alkaline phosphatase-conjugated Abs (Southern Biotechnology Assoc. Birmingham, AL). Serum Ig concentrations were determined by fivefold serial dilutions and calculated according to standard charts as described previously (39).

VSV Infections and Detection of VSV-neutralizing Abs. Mice were immunized with VSV-Indiana (2 × 10⁶ PFU, intravenously). After 4, 8, and 12 d, sera were collected and neutralizing IgM and IgG Ab titers determined as described (43). In brief, 1:2 dilutions of 40-fold prediluted serum were incubated with VSV for 90 min. The presence of remaining infectious virus was determined by incubating the VSV serum samples with fibroblasts for another 24 h. Serum dilutions that reduced the number of viral plaques by 50% were taken as specific titers. IgG titers were determined after preincubation of sera with 2 mercaptoethanol, a procedure that eliminates IgM (43).

Germinal Center Formation and Immunohistochemistry. To determine formation of germinal centers, 6 wk old SEK1^/RAG2^/ and SEK1^+/+RAG2^/ chimeric mice and CD28^/ mice were infected with VSV-Indiana as described above. Spleens from VSV-infected animals were harvested 12 d after the intial infection, frozen in liquid nitrogen, and processed for cryosections. Cryostat sections (5 µm) were fixed in acetone (10 min). Sections were incubated with PNA (diluted 1:200) and bound PNA was detected by rabbit anti-PNA Abs (diluted 1:300; DAKO, Glostrup, Denmark). CD4 was detected by the rat mAb YTS191. Binding of primary Abs was detected by alkaline phosphatase-labeled goat Abs to rabbit or rat Ig (1:80 dilution; Jackson Labs.) followed by alkaline phosphatase-labeled donkey Abs against goat Ig (1:80 dilution; DAKO). Alkaline phosphatase was visualized using Napthol AS-BI phosphatase and New Fuchsin as a substrate, which yields a red precipitate. Endogenous alkaline phosphatase activity was blocked by levamisole (44).

SAPK/JNK Activities. Thymocytes and purified lymph node T cells (5 × 10⁶) were activated with PMA (50 ng/ml) and the Ca²⁺ ionophore A23617 (1 µg/ml) as previously described (23, 29). Cells were lysed in ice-cold lysis buffer (10 mM NaCl, 20 mM Pipes, pH 7.0, 0.5% NP-40, 5 mM EDTA, 0.05  mercaptoethanol, 100 µM Na₃VO₄, 50 mM NaF, 20 µg/ml leupeptin, and 1 mM benzamidine). Cleared lysates were adjusted to equal protein concentrations (BioRad Protein Assay; Bio Rad Labs., Hercules, CA). SAPKs/JNKs were immunoprecipitated (1 h, 4°C) using polyclonal rabbit anti-SAPK/JNK Abs reactive against all SAPK/JNK isoforms (10). Immune complexes were harvested on protein A-Sepharose beads. For kinase assays, immune complexes were washed three times with PBS-Triton buffer (150 mM NaCl, Na₂HPO₄, 4 mM NaH₂PO₄, 0.1% Triton X-100, 100 mM Na₃VO₄, 50 mM NaF, 20 µg/ml leupeptin, and 1 mM benzamidine). SAPK/JNK kinase assays were performed in 20 µl of kinase buffer (10 mM MgCl₂, 50 mM Tris-Cl, pH 7.5, 1 mM EGTA, pH 7.5) in the presence of 1.2 µCi [³²P]-ATP and 5 µg glutathione-S-transferase-c-Jun as in vitro substrate (30°C, 30 min). The reaction was stopped by the addition of 2× SDS sample buffer. Phosphoproteins were separated by SDS-PAGE and visualized by autoradiography as described (10). The levels of expression of SAPK/JNKs in thymocytes and lymph node cells were determined by immunoblotting using goat anti-JNK1 and rabbit anti-JNK2 polyclonal Abs (both from Santa Cruz Biotechnology Inc., Santa Cruz, CA; 23).

Results

Impaired Proliferation and IL-2 Production of SEK1^/ Peripheral T Cells.

Recent biochemical studies implied that the SAPK/JNK signaling pathway is operating in T cells, and that cell proliferation and IL-2 production induced by CD28 costimulation may be mediated via SAPK/JNK (29, 36, 37). SEK1^/RAG2^/ chimeric mice have a smaller thymus, but normal numbers of peripheral T cells (Table 1; reference 23). To test the role of SEK1 in CD28 costimulation, lymph node T cells were cultured in anti-CD3- Ab-coated plates in the absence or presence of various concentrations of soluble anti-CD28 Abs. Whereas SEK1^/RAG2^/ and SEK1^+/+RAG2^/ T cells reponded in the same way to CD3- activation alone, CD28-mediated upregulation of proliferation and IL-2 production were significantly reduced in SEK1^/ T cells (Fig. 1, A and B). Reduced proliferation and IL-2 production were also observed in SEK1^/ T cells after stimulation with PMA/Ca²⁺ ionophore (Fig. 1, A and B), which mimic TCR-CD3- and CD28-mediated signal transduction (29, 45).

/

/

Since the proliferative response to plate-bound anti- CD3- Abs alone was still very vigorous (Fig. 1 A), we analyzed T cell activation in response to suboptimal concentrations of soluble anti-CD3- Abs. As shown in Fig. 1, C and D, proliferation and IL-2 production of SEK1^/RAG2^/ and SEK1^+/+RAG2^/ chimeric T cells were minimal after stimulation with soluble anti-CD3- alone. Although the addition of anti-CD28 greatly enhanced the proliferation and IL-2 production of SEK1^+/+RAG2^/ T cells, SEK1^/RAG2^/ T cells were significantly impaired in proliferation and IL-2 production (Fig. 1, C and D). It should be noted that freshly isolated T cells from SEK1^/ mice displayed upregulated expression of CD28, but normal surface expression of the TCR-/-CD3- complex, IL-2R- chain (CD25), CD69, and adhesion molecules such as ICAM-1 (not shown). These data provide the first genetic evidence that SEK1 plays an important role in T cell proliferation and IL-2 production in transmitting CD28 signals to downstream effector molecules.

To determine the effect of the SEK1 mutation on B cell development, single cell suspensions from spleen, lymph nodes, and bone marrow of SEK1^/RAG2^/ chimeric, SEK1^+/+RAG2^/ chimeric, RAG2^/, and control 129/J mice were stained with mAbs against B lineage-specific markers (Fig. 2, Table 1). The bone marrow of 129/J mice contained a relatively low number (12%) of B220⁺CD43⁺ pro-B cells and a larger population (30%) of B220⁺CD43 pre-B cell precursors, and mature B cells in peripheral lymphatic organs expressed both IgM (Fig. 2) and IgD (not shown) on the cell surface (46). By contrast, B cell differentiation in the bone marrow of RAG2^/ mice was blocked at the pro-B cell stage (B220⁺CD43⁺IgM) and RAG2^/ mice did not have any mature sIgM⁺ B cells (Fig. 2, Table 1; 47, 48). B-cell development and expression of sIgM and sIgD were restored in chimeras derived from injection with parental SEK1^+/+ ES cells. In contrast, the relative and total numbers of B220⁺CD43 bone marrow cells and B220⁺ sIgM⁺ peripheral B cells were significantly reduced in SEK1^/RAG2^/ chimeric mice (Fig. 2, Table 1). Peripheral B cells from SEK1^/ mice expressed normal levels of CD23, CD40, CD44, ICAM-1, CD95 (FAS), and H2K^b on the cell surface (not shown). The partial block in the development from B220⁺CD43⁺ pro-B cells to more mature B220⁺CD43 pre-B cells was also observed by alterations in IL-2R- chain (CD25) expression, an early B cell maturation marker that is expressed before sIgM expression (46). Although ~75% of 129/J or SEK1^+/+ chimeric B220⁺ bone marrow B cells expressed CD25 on the cell surface, expression of CD25 was significantly reduced in SEK1^/ bone marrow B cells (Table 1).

To analyze whether the observed block in B cell maturation was due to the SEK1 mutation and not due to low chimerism and contribution of RAG2^/ cells to pro-B cells, we FACS^® sorted B220⁺CD43⁺ and B220⁺CD43 bone marrow cells and analyzed the genotype of sorted cells by PCR (Fig. 3). Both B220⁺CD43⁺ pro-B cells and the more mature B220⁺CD43 B cell populations in SEK1^/ chimeras contained mutant, but not wild-type SEK1 alleles indicating that both populations were derived from SEK1^/ ES cells. These data imply that SEK1-mediated signaling plays a role at the transition from B220⁺CD25CD43⁺ pro-B cells to B220⁺CD25⁺CD43 pre-B cells in the bone marrow.

/

/

/

Previously it has been shown that CD40 signaling in B cells leads to the induction of SAPK/JNK activity (31, 32). To determine the requirement of SEK1 for B cell activation, we measured proliferation of B cells in response to various stimuli. SEK1^/RAG2^/ B cells responded normally to LPS, IL-4, anti-CD40, IL-4 plus anti-CD40, and Igµ cross-linking (Fig. 4 A). Moreover, SEK1^/RAG2^/ B cells upregulated ICAM-1 and CD23 upon activation with anti-CD40 in the absence or presence of IL-4 (not shown; 42). The basal serum levels for the Ig subclasses IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA were also comparable between SEK1^/RAG2^/ and SEK1^+/+RAG2^/ chimeric mice (Fig. 4 B).

/

/

/

/

/

VSV infections are exclusively controlled by neutralizing Abs (49). All neutralizing Abs are directed against the VSV glycoprotein which is present in a highly repetitive form in the viral envelope. Due to this repetitiveness, neutralizing IgM Abs are induced in complete absence of T cell help (49). However, the isotype switch from IgM to IgG is strictly T cell dependent (50). Recently, it has also been shown that the production of VSV-neutralizing IgG Abs is decreased in CD28^/ mice (39). Since, SEK1^/ T cells had reduced proliferation and IL-2 production in response to CD28 costimulation (Fig. 1), but SEK1^/ B cells could be activated normally and produced normal levels of Ig subclasses (Fig. 4), we examined T cell help and IgG class switching in SEK1^/RAG2^/ and SEK1^+/+RAG2^/ mice after infection with VSV (Table 2). Neutralizing serum IgM was assessed 4 d and neutralizing serum IgG levels were measured 4, 8, and 12 d after VSV infection. VSV infection induced rapid, T cell-independent IgM production, followed by a T helper cell and CD28 costimulation-dependent IgG response (Table 2). Surprisingly, both early IgM production and IgG class switching were comparable between SEK1^/RAG2^/ and SEK1^+/+ RAG2^/ mice (Table 2). Moreover, SEK1^/ mice survived for more than 4 wk after infection, indicating that the B cell response was protective.

Table 2. Neutralizing Anti-VSV Response in SEK1/ and CD28/ Mice Titers of neutralizing activities ( log 2)* Day 4 Day8 Day12 Genotype IgM IgG IgG IgG SEK1+/+  9 <1 9 11 10 <1 9 11 SEK1/  9 <1 7 10 10 <1 8 11 CD28/  9 <1 2  2  9 <1 2  3 *  Sera were isolated from mice after intravenous infection with VSV (2 × 106 PFU). Neutralizing IgM and IgG titers were determined as described in Materials and Methods. Titers represent twofold dilution steps of sera starting with 1:40.

The prominent expression of the GCK in follicular germinal centers (34) and activation of SAPK through GCK (33) suggested that the GCK/SAPK pathway might be important for B cell differentiation within germinal centers. Moreover, mice lacking CD28 (51) or CD40 (52, 53) do not develop germinal centers. Since all of these receptors can activate SAPKs/JNKs (10, 29, 31, 32), we tested whether virus-specific germinal center formation was normal in SEK1^/ mice. Although VSV-specific germinal centers were completely absent in CD28^/ mice after challenge with VSV, SEK1^/RAG2^/ chimeric mice developed germinal centers with normal morphology (Fig. 5) and at normal frequency (Table 3). Germinal center B cells were positive for PNA expression (Fig. 5, A-C). CD4⁺ T cells were mainly present in the T area, but were also observed within germinal centers (Fig. 5, D-F). Moreover, a light zone containing strongly VSV-binding germinal center B cells could be distinguished from a dark zone containing sIg-negative B lymphocytes (Fig. 5, G-I). It should be noted that VSV-specific plasma cells were detectable in the spleens of CD28^/ mice (Fig. 5 J) and that CD28^/ mice could still produce, albeit at low levels, neutralizing IgG Abs (Table 2). These data show that SEK1^/RAG2^/ mice can mount biologically relevant responses against VSV and that SEK1 has no apparent role in CD28-dependent, virus-specific germinal center formation.

/

/

/

/

/

/

/

/

/

/

/

/

Table 3. Quantification of Germinal Centers in SEK1/ and CD28/ Mice No. of germinal centers* Genotype PNA+ VSV specific SEK1+/+ 53 ± 4.4 51 ± 5.0 SEK1/ 55 ± 2.0 49 ± 2.4 CD28/ 0 0 *  The mean germinal center numbers (± SD) per section were determined on at least eight different spleen section levels from two individual mice/group. PNA-positive and VSV-specific germinal centers were visualized as described in Materials and Methods.

Although SEK1^/ peripheral T cells displayed reduced proliferation and IL-2 production in response to CD28 in vitro, SEK1 expression was not an absolute requirement for T cell activation in vivo. Thus, similar to our data that shows that SEK1-dependent and SEK1-independent signaling pathways for SAPKs/JNKs activation exist in ES cells (23), it was possible that a SEK1-independent pathway for SAPK/JNK activation was operational in T cells. To test this hypothesis, we examined SAPKs/JNKs activation in peripheral T cells and thymocytes in response to PMA/Ca²⁺ ionophore (Fig. 6 A), which mimic TCR- CD3 and CD28-mediated signal transduction (29,45). SAPK activation, i.e., SAPK-mediated c-Jun phosphorylation, was observed in SEK1^+/+, but not in SEK1^/, thymocytes, indicating that SEK1 is the crucial genetic regulator of PMA/Ca²⁺ ionophore-triggered activation of SAPKs/JNKs in thymocytes (23; Fig. 6 A). Surprisingly, using the same PMA/Ca²⁺ ionophore activation regimen, SAPK activation was observed in peripheral lymph node T cells from both SEK1^/ and SEK1^+/+ mice (Fig. 6 A). The levels of SAPKs (JNK1 and JNK2) expression were comparable among SEK1^/RAG2^/ and SEK1^+/+RAG2^/ thymocytes and peripheral T cells (Fig. 6 B), suggesting that the observed differences in PMA/Ca²⁺ ionophore-mediated SAPK activation in SEK1^/ thymocytes versus SEK1^/ lymph node T cells were not the result of alterations in SAPK expression. Our data that SAPKs are activated in peripheral T cells, but not in thymocytes, from SEK1^/ chimeras in response to the same stimulus PMA/Ca²⁺ ionophore indicate that signaling pathways for SAPK/JNK activation are developmentally regulated.

/

/

IL-2 Production in SEK1^/ Thymocytes.

The results in SEK1^/ lymph node T cells indicated that SEK1 has a role in CD28-mediated costimulation for proliferation and IL-2 production in peripheral T cells and that lymph node T cells use a second signaling pathway for SAPK activation that is independent of SEK1. Since this second signaling pathway is not operational in SEK1^/ thymocytes, we tested proliferation and IL-2 production of SEK1^/ thymocytes in response to PMA/Ca²⁺ ionophore and CD3/ CD28 activation. Surprisingly, SEK1^/RAG2^/ thymocytes proliferated and made IL-2 after PMA/Ca²⁺ ionophore and CD3/CD28 stimulation, albeit at significantly lower levels compared to SEK1^+/+RAG2^/ thymocytes (Fig. 7). These data further confirm that SEK1 relays CD28 costimulatory signals to IL-2 production in T cells. However, these results in thymocytes also indicate that CD28 costimulation and PMA/Ca²⁺ ionophore-triggered signaling pathways exist that can mediate proliferation and IL-2 production independently of SAPK activation.

/

Discussion

SAPKs/JNKs are activated in response to many cellular stresses such as osmolarity changes, metabolic poisons, DNA-damaging agents, heat shock, ischemia/reperfusion injury, UV-, or -irradiation (9, 10, 13, 14, 18, 25). The dual specificity kinase SEK1 (JNK kinase/MKK4) has been identified as a potent and direct activator of SAPKs/JNKs in vitro and in cell lines in vivo (20). Although it has been shown genetically that a second SAPK activator, SEK2, exists (23), SEK1 is the only cloned kinase that can directly activate SAPKs/JNKs (11, 12).

In addition to the induction of SAPK/JNK activity by many types of cellular stresses (10), SAPKs/JNKs are activated in response to certain growth factors, heterotrimeric G proteins, phorbol esters, CD40 cross-linking, and CD28-mediated costimulation in T cells (9, 10, 14, 29, 54). Moreover, activation of SAPKs/JNKs leads to phosphorylation of c-Jun and activation of Jun/Fos heterodimeric AP-1 complexes, which are involved in the coordinate activation of IL-2 transcription (10, 19, 20, 55). In T cells, ligation of the TCR results in rapid activation of the Ras Raf MEK MAPK signaling cascade (11, 56). However, activation of the MAPK cascade is not sufficient for effective IL-2 production and cell proliferation for which T cells require a second signal (35). Recently, it has been shown that coordinate stimulation of the TCR-CD3 complex and the costimulatory receptor CD28 correlates with the activation of SAPKs/JNKs, phosphorylation of c-Jun, and induction of AP-1 activity (29). These biochemical data indicated that T cells use two distinct signaling cascades for antigen-specific activation, TCR-triggered MAPK activation and TCR-CD28-induced activation of SAPKs/JNKs. Importantly, it has been suggested that failure to activate SAPKs/JNKs in T cells might result in clonal anergy and the induction of immunological tolerance (36, 37).

Our demonstration of defective IL-2 production and proliferation in SEK1^/ T cells in response to CD28 costimulation and PMA/Ca²⁺ ionophore provides the first genetic evidence that the stress signaling kinase SEK1 is a downstream effector involved in TCR and CD28 coreceptor signaling. However, the impairments of proliferation and IL-2 production were not complete, and a strong activation signal via the TCR-CD3 complex alone triggered normal proliferation of SEK1^/ T cells. Thus, although SEK1 appears to be necessary for adequate IL-2 production and proliferation in T cells, another activator(s) can compensate for the SEK1 deficiency in peripheral T cells. This hypothesis is in line with our biochemical data on SEK1-independent activation of SAPKs/JNKs in lymph node T cells in response to PMA/Ca²⁺ ionophore stimulation. Interestingly, this second pathway for SAPK/JNK activation is only operational in peripheral T cells but not in thymocytes, indicating that signaling pathways for SAPK/JNK activation are developmentally regulated in T cells. It has also been shown that proliferation and IL-2 production are normal in c-Jun^/RAG2^/ T cells, suggesting that not only c-Jun, but also other Jun family members, i.e., JunD and JunB, may have a role in T cell activation (57). The exact role of distinct SAPK/JNK activators, SEK1 versus SEK2, and of different Jun family transcription factors in CD28-mediated IL-2 production and T cell activation needs to be determined.

SEK1^/RAG2^/ thymocytes failed to induce SAPK/ JNK in response to PMA/Ca²⁺ ionophore. Interestingly, SEK1^/RAG2^/ thymocytes still proliferated and produced IL-2 after PMA/Ca²⁺ ionophore and CD3/CD28 stimulation, albeit at significantly lower levels compared to SEK1^+/+RAG2^/ thymocytes. These data further confirmed that SEK1 relays CD28 costimulatory signals to IL-2 production in T cells. However, these results also indicate that, at least in thymocytes, CD28 and PMA/Ca²⁺ ionophore-triggered signaling pathways exist that can mediate proliferation and IL-2 production independently of SAPK activation. Besides activation of SEK1 and SAPKs/JNKs, additional downstream effectors for CD28 signaling have been identified including PI3K, PLC1, Raf-1, and Vav (see review in reference 58). In particular, it has been shown that Vav, Ras, and the Vav-associated tyrosine phosphoprotein SLP76 can cooperate to induce nuclear factor of activated T ce