Defective T Cell Differentiation in the Absence of Jnk1

See allHide authors and affiliations

Science  11 Dec 1998:
Vol. 282, Issue 5396, pp. 2092-2095
DOI: 10.1126/science.282.5396.2092


The c-Jun NH2-terminal kinase (JNK) signaling pathway has been implicated in the immune response that is mediated by the activation and differentiation of CD4 helper T (TH) cells into TH1 and TH2 effector cells. JNK activity observed in wild-type activated TH cells was severely reduced in TH cells fromJnk1 –/– mice. TheJnk1 –/– T cells hyperproliferated, exhibited decreased activation-induced cell death, and preferentially differentiated to TH2 cells. The enhanced production of TH2 cytokines by Jnk1 –/– cells was associated with increased nuclear accumulation of the transcription factor NFATc. Thus, the JNK1 signaling pathway plays a key role in T cell receptor–initiated TH cell proliferation, apoptosis, and differentiation.

When activated by antigen-presenting cells (APCs), TH cells undergo clonal proliferation and produce interleukin 2 (IL-2). The activated TH cells may then become TH1 or TH2 effector cells (1), which mediate inflammatory or humoral responses, respectively. The polarization of TH cell differentiation is, at least in part, determined by the cytokine environment (1). IL-12, produced by activated APCs, induces TH1 development of naı̈ve TH cells. IL-4, made by T cells, is required for TH2 differentiation. Thus, early production of IL-4 or IL-12 determines TH cell lineage commitment and the type of immune response that occurs. Although most attention has focused on the effect of polarizing cytokines on TH cell differentiation, signals from the T cell receptor (TCR)–CD3 complex and from the costimulatory factor CD28 may also affect cytokine production by mechanisms not yet understood (2).

JNK, also known as stress-activated protein kinase, phosphorylates the transcription factor c-Jun and increases AP-1 transcription activity (3, 4). Other substrates include JunD, ATF2, ATFa, Elk-1, Sap-1, and NFAT4 (3, 4). Signals from both the TCR-CD3 complex and CD28 are required for JNK and AP-1 activation in T cells, and these signals may be integrated in such a way as to mediate T cell activation and the induction of IL-2 transcription (5). Although JNK is implicated in IL-2 gene transcription, JNK may also act to stabilize IL-2 mRNA (6). AP-1 has also been reported to be important for the regulation of TH1 and TH2 cytokine genes (7, 8).

To understand the role or roles of JNK in TH cell activation and differentiation, we generated Jnk1-deficient mice through homologous recombination in embryonic stem cells (9) (Fig. 1A). Targeted disruption of the Jnk1 gene resulted in a null allele, as confirmed by mRNA (10) and protein expression analysis of embryonic fibroblast (Fig. 1B) and T cell (11) extracts.Jnk1-deficient mice were fertile and of normal size. Lymphocyte development appeared normal, with typical ratios of T cells to B cells, CD4 to CD8, and naı̈ve to memory T cells in the periphery (10). The absence of apparent developmental defects of Jnk1 –/– lymphocytes might be the result of redundancy, because Jnk1 and Jnk2are coexpressed in lymphoid tissues (4). Therefore, we tested whether JNK1 and JNK2 are activated similarly during the course of TH cell activation.

Figure 1

Generation ofJnk1-deficient mice. (A) Schematic representation of the genomic Jnk1 locus, the targeting vector, and the mutated Jnk1 locus. Restriction enzyme sites (B, Bam HI; N, Not I; P, Pst I; S, Spe I) and the probe for Southern blot analysis are indicated. (B) Protein immunoblot analysis confirmed the absence of expression of the JNK1 protein. Extracts of embryo fibroblasts were examined by immunoblot analysis with a monoclonal antibody to JNK1 and JNK2. The 55-kD and 46-kD forms of JNK1 and JNK2 are indicated. (C) Abrogation of JNK activity in activated Jnk1-deficient CD4 T cells. Purified splenic CD4 T cells from wild-type (WT) orJnk1–/– (KO) mice were stimulated with anti-CD3 in the presence or absence of anti-CD28. Cell extracts were prepared at different times and JNK activity was measured with the use of c-Jun as substrate. The data shown are representative of two independent experiments with the same results.

Purified CD4 T cells from wild-type or knockout mice were stimulated by antibodies to CD3 (anti-CD3) with or without anti-CD28 (12,13), and JNK activity was measured using c-Jun as the substrate. During the first 48 hours, induced JNK activity was greatly reduced in the Jnk1 –/– TH cells; moreover, anti-CD28 could not enhance kinase activity (Fig. 1C). Essentially no JNK activity was detected in Jnk1 –/–TH cells stimulated for only 5 min, despite the same JNK2 protein expression (11). Thus, JNK1 appeared to account for most of the JNK activity in newly activated T cells. After 60 hours of stimulation, JNK activity in the Jnk1 –/– cells was similar to that in wild-type cells, and in each case this activity was presumably derived from JNK2. In fact, JNK2 represents most JNK activity in TH1 effector cells (14).

To investigate the role or roles of JNK1 in TH cell activation and IL-2 production, we stimulated T cells with concanavalin A (Con A), anti-CD3, or anti-CD3 plus anti-CD28 (12, 13). Relative to wild-type cells, Jnk1 –/– spleen cells produced the same amount of IL-2 (Fig. 2A) and CD4 T cells produced the same amount of IL-2 mRNA (10) 24 hours after stimulation, despite the lack of JNK activation (Fig. 1C), similar to Jnk2- and c-Jun-deficient T cells (14, 15). Although JNK may therefore not be required for IL-2 expression, it is also possible that JNK1 and JNK2 are redundant for IL-2 regulation. Despite normal IL-2 production, Jnk1 –/– splenocytes and CD4 T cells (10) displayed enhanced proliferation (12,13) (Fig. 2B). In addition, Jnk1-deficient TH cells had a moderate reduction of activation-induced cell death (from 46 to 27%) (13), suggesting that JNK1 may be involved in the regulation of T cell apoptosis. Decreased apoptosis by Jnk1 –/– T cells could therefore contribute to the increased proliferation of these cells.

Figure 2

Increased proliferation and TH2 cytokines inJnk1 –/– T cells. (A) Normal IL-2 production by activated Jnk1 –/– spleen T cells. Spleen cells from 2-month-old wild-type (WT) or Jnk1–/– (KO) mice were stimulated as indicated for 24 hours, and the amount of IL-2 in the cell supernatant was measured by ELISA. A representative of three independent experiments using different mice is shown. (B) Increased proliferation byJnk1-deficient T cells. Wild-type orJnk1–/– splenocytes were stimulated as in (A) for 3 days, and [3H]thymidine was included during the last 8 hours to measure cellular proliferation. (C) Splenic CD4 T cells from 6- to 8-week-old wild-type (WT) or Jnk1 –/– (KO) mice were stimulated as shown (12); 4 days later, amounts of TH cytokines were measured by ELISA. The data shown are representative of 10 independent experiments, each using different mice.

To test whether JNK1 is involved in TH cell differentiation, we measured cytokine production by purified CD4 T cells after stimulation for 4 days by immobilized anti-CD3 with or without anti-CD28 (12, 13). Production of interferon-γ (IFN-γ) TH1 cytokine by Jnk1–/– TH cells appeared normal in response to anti-CD3, but there was a consistent lack of enhancement by anti-CD28 (Fig. 2C). Wild-type cells produced small amounts of IL-4, IL-5, and IL-10 TH2 cytokines in response to immobilized anti-CD3, whereas theJnk1 –/– TH cells secreted remarkably large amounts of these TH2 cytokines (Fig. 2C). Addition of anti-CD28 enhanced TH2 cytokine production in both wild-type and Jnk1–/– THcells (Fig. 2C). These observations indicate thatJnk1 –/– CD4 T cells were hyperresponsive to anti-CD3 and produced TH2 cytokines even in the absence of costimulation.

Differentiating Jnk1–/– THcells produced more IL-4 than did wild-type cells. To assess whetherJnk1–/– TH cells become different effector cells (that is, polarized toward TH2), we isolated wild-type andJnk1–/– CD44loCD45RBhinaı̈ve CD4 T cells and cultured them in vitro under neutral conditions, using immobilized anti-CD3, IL-2, and irradiated APCs derived from the wild-type mice (16). After restimulation with anti-CD3, wild-type effector cells made significant amounts of IFN-γ (361 U/ml in one representative experiment) but no detectable IL-4, whereas Jnk1–/– TH cells made much more IL-4 (33 U/ml) and much less IFN-γ (53 U/ml). Consistent with these observations, larger amounts of IL-4 mRNA were detected (17) (Fig. 3A). Thus,Jnk1 –/– CD4 T cells differentiated preferentially into TH2 cells, whereas the wild-type cells of the same genetic background became TH1 cells.

Figure 3

Jnk1–/–TH precursor cells preferentially differentiate to the TH2 lineage, and differentiated Jnk1–/–TH1 and TH2 effector cells make increased amounts of TH2 cytokines. (A) Spleen and lymph node naı̈ve CD4 T cells of wild-type or Jnk1–/– mice were differentiated under neutral conditions (16) and then restimulated with anti-CD3 for 24 hours. The cytokine mRNAs produced after restimulation were analyzed by competitive RT-PCR and cytokines by ELISA. (B) Wild-type (WT) or knockout (KO) splenic CD4 T cells were differentiated and restimulated in vitro as in (A), except that IL-12 and anti–IL-4 were added for TH1 differentiation, and IL-4 and anti–IFN-γ for TH2 differentiation. (C) Age- and sex-matched wild-type (WT) or knockout (KO) animals were immunized with KLH peptide precipitated in alum; 9 days later, the draining lymph node cells were treated with or without KLH peptide for 4 days. The amount of IL-5 produced was determined by ELISA.

To test the possibility that the hyperproduction of TH2 cytokines by Jnk1 –/–TH cells was caused by an intrinsic deficiency in their ability to become TH1 cells, we differentiated the CD4 T cells under the above conditions with the addition of IL-12 and anti–IL-4 to promote TH1 differentiation, or IL-4 and anti–IFN-γ for TH2 development (16). When restimulated with anti-CD3 for 24 hours, theJnk1–/– TH1 effector cells secreted amounts of IFN-γ similar to wild-type cells (Fig. 3B), which indicated that the Jnk1 –/– THcells could differentiate to TH1 cells and produce TH1 cytokines. However, theJnk1 –/– TH1 population also made IL-4, IL-5, and IL-10 after anti-CD3 restimulation (Fig. 3B), suggesting a failure to down-regulate these cytokines by the IL-12– and IFN-γ–rich environment. This TH2 cytokine production by the Jnk1 –/– TH1 population was independent of IL-4, because anti–IL-4 was present during the differentiation process. The Jnk1 –/– Jnk1 −/− TH2 effector cells also secreted greatly increased amounts of TH2 cytokines: Relative to the same number of wild-type cells, theJnk1 −/− TH2 effector cells produced almost 10 times as much IL-4, five times as much IL-5, and moderately increased amounts of IL-10 (Fig. 3B). The same results were obtained when purified naı̈ve CD4 T cells were differentiated (10). The increased amount of TH2 cytokines inJnk1 –/– TH cells was also not the result of differential survival of TH2 versus TH1 cells in Jnk1 –/– mice, as their patterns of apoptosis were similar (10).

To test whether the exaggerated TH2 cytokine production by the polyclonally activated Jnk1–/– TH cells in vitro reflects a feature of the in vivo antigen-specific immune response, we immunized age- and sex-matched mice with keyhole limpet hemocyanin (KLH) precipitated in alum, an adjuvant promoting TH2 responses. When the draining lymph node cells were restimulated with KLH in vitro, relative to wild-type cells, Jnk1–/– cells produced at least four times as much IL-5 (Fig. 3C). This finding is indicative of an enhanced TH2 response, which correlates well with the earlier in vitro results with anti-CD3 activation (Fig. 3B); as expected, neither wild-type nor knockout T cells made IFN-γ under these immunization conditions.

These results showed that Jnk1 –/–TH effector cells, in vitro or ex vivo, made elevated TH2 cytokines. We also found that 24 and 48 hours after treatment with anti-CD3, relative to wild-type cells,Jnk1–/– CD4 TH precursor cells made more IL-4 mRNA and slightly more IL-5 mRNA (10). Because IL-4 is necessary and sufficient to generate TH2 cells, the preference of naı̈ve Jnk1–/– TH cells to differentiate to the TH2 lineage and make IL-5 and IL-10. under neutral conditions (Fig. 3A) could be caused by increased IL-4 production in the early activation phase.

To identify factors that may be responsible for the exaggerated IL-4 production by Jnk1-deficient cells, we treated purified CD4 T cells with anti-CD3 for 48 hours and prepared nuclear extracts. Several transcription factors involved in IL-4 transcription or required for TH2 differentiation were assayed by immunoblot analysis (18, 19). There was no difference in the amounts of the transcription factors GATA-3 or Stat-6 in the nuclei of wild-type or Jnk1–/– cells, both of which were diminished by anti–IL-4 treatment in culture (Fig. 4A). Thus, GATA-3, like Stat-6, is probably responsive to IL-4 under these conditions. There was a moderate increase in the amount of JunB protein (Fig. 4A), a member of the Jun family that is increased in TH2 cells and important for IL-4 transcription (7). NFATc is essential for IL-4 production and TH2 differentiation (20). Anti-CD3–activated wild-type cells had little NFATc in the nucleus, whereas costimulation with anti-CD28 enhanced NFATc nuclear accumulation (Fig. 4B), in keeping with the finding that TH2 cytokine induction in wild-type TH cells requires costimulation (Fig. 2C). In contrast, anti-CD3 treatment alone led to an increase in nuclear NFATc in Jnk1–/– TH cells and a decrease in cytoplasmic NFATc (Fig. 4, A and B), consistent with the high TH2 cytokine production by CD3-activated Jnk1 –/– cells (Fig. 2C). The enhanced accumulation of nuclear NFATc inJnk1–/– TH cells was observed in cells 8, 24, and 48 hours after stimulation, but was not observed in nonactivated cells (10). NFATc accumulation was specific because the amount of nuclear NFATp, a proposed negative regulator of TH2 cytokine genes (21), was the same in wild-type and Jnk1–/– cells (Fig. 4A). Enhanced nuclear accumulation of NFATc inJnk1 –/– T cells was not blocked by anti–IL-4 (Fig. 4A); hence, increased IL-4 production and NFATc nuclear localization is intrinsic to T cell receptor signaling and is not secondary to IL-4 production. Because NFATc can bind to the IL-4 promoter and is required for IL-4 production and TH2 differentiation (20, 22), the greatly enhanced amount of nuclear NFATc could account for the increased IL-4 production in CD3-activated Jnk1-deficient mice.

Figure 4

Enhanced NFATc nuclear localization by anti-CD3–activated Jnk1-deficient TH cells. Wild-type or knockout CD4 T cells were stimulated (2 days) by anti-CD3 with or without anti–IL-4 (A) or by anti-CD3 with or without anti-CD28 (B). Transcription factors were examined by immunoblot analysis of cytoplasmic (NFATc) or nuclear (JunB, GATA-3, Stat-6, NFATp, and NFATc) extracts. The data shown are representative of three independent experiments.

The mechanism by which JNK1 negatively regulates NFATc nuclear accumulation remains to be resolved. The isoform NFAT4 is phosphorylated and negatively regulated by JNK, leading to nuclear exclusion (23). This regulation appears to be specific to the NFAT4 isoform; evidence for JNK regulation of NFATc was not reported (23). An indirect mechanism may therefore account for the altered regulation of NFATc inJnk1–/– TH cells. NFATc and NFATp can bind to the IL-4 promoter NFAT sites (22). BothJnk1 and NFATp knockout mice have enhanced T cell proliferation and TH2 cytokine production (21,24), precisely the opposite of the NFATc knockout. It is therefore possible that these two NFAT factors antagonize each other in the regulation of the IL-4 gene. The apparent similarity betweenNFATp–/– and Jnk1–/– phenotypes supports a functional linkage between JNK1 and NFAT.

Our results further reveal a novel mechanism by which TCR signaling negatively regulates TH2 cytokines through JNK1. Positive and negative regulation of JNK1 activity may affect the decision of TH cells to differentiate into TH1 or TH2 effectors, and therefore may affect the type of immune response that is initiated. The function of JNK1 demonstrated in this study is distinct from that of JNK2, which is required for IFN-γ production in TH1 cells (14). Moreover, the related p38 mitogen-activated protein kinase pathway is TH1 specific and drives IFN-γ transcription (25). Together, these pathways potentiate the TH1 response and provide a potential target for pharmaceutical intervention.

  • * These authors contributed equally to this report.

  • Present address: Lilly Research Laboratory, Eli Lilly and Co., Indianapolis, IN 46285, USA.

  • To whom correspondence should be addressed. E-mail: richard.flavell{at}


Stay Connected to Science

Navigate This Article