MEKK3 Overexpression Contributes to the Hyperresponsiveness of IL-12–Overproducing Cells and CD4⁺ T Conventional Cells in Nonobese Diabetic Mice

Type 1 diabetes (T1D) is a chronic progressive autoimmune disease characterized by mononuclear cell infiltration, dominated by IL-12–dependent Th1 cells, of the pancreatic islets, with subsequent destruction of insulin-producing β cells (1, 2). NOD mice recapitulate many aspects of the pathogenesis of T1D in humans and are therefore frequently used in studies addressing the cellular and molecular mechanisms of T1D (3, 4). Several groups have demonstrated that compared with cells from reference strains of mice, including C57BL/6 (B6), NOD macrophages and bone marrow-derived dendritic cells (BMDCs) exhibit elevated IL-12 production in response to various inflammatory stimuli such as LPS (5–12). IL-12 is a heterodimeric 70-kDa (p70) cytokine composed of two disulfide-linked glycosylated chains of 40 kDa (p40) and 35 kDa (p35), encoded by two distinct genes (13, 14). This cytokine is mainly produced by macrophages and DC and plays a pivotal role in the regulation of cell-mediated immunity. IL-12 induces Th1 differentiation and has an important role in maintaining the balance between Th1 and Th2 responses in vivo (13, 14). Besides elevated IL-12 production, Th1 polarization in NOD mice might also benefit from the hyperresponsiveness of CD4⁺ T conventional (Tconv) cells. NOD CD4⁺CD25⁻ Tconv cells show higher rates of proliferation upon TCR engagement than do their B6 counterparts (15). The origin of the hyperresponsiveness is unknown. However, it has been revealed that this difference is independent of autoimmune inflammation, does not map to the idd3 region, and is not due to the overproduction of IL-21 in NOD mice (15).

There is an NF-κB binding site in the IL-12p40 promoter, which is critical for the expression of IL-12 (16, 17). It has been consistently found that the increased IL-12 secretion is a direct result of enhanced nuclear translocation and transcriptional activity of NF-κB (6–8). Under normal conditions, NF-κB is sequestered in an inactive state by IκBs in the cytoplasm. Inflammatory stimuli induce the oligomerization of high-affinity receptors. Adaptor molecules are then recruited to the cytosolic region of the receptors. The combination of adaptors is different from one receptor to another. These adaptors trigger the recruitment and the oligomerization of the IκB kinase (IKK) complex and MAPK kinase kinase (MAK3K or MEKK). The IKK complex contains two homologous catalytic subunits, IKKα and IKKβ, and the regulatory subunit IKKγ (18, 19). It remains controversial how IKK is activated. Certain MAP3Ks, such as TGF-β–activated protein kinase 1 (TAK1) and MEKK3, seem to be essential for the phosphorylation of IKKα/β and the consequent conformational changes resulting in kinase activation (18–20). Once activated, IKK phosphorylates IκBs, triggering their ubiquitination and subsequent degradation by the 26S proteosome. Degradation of IκBs unmasks the nuclear translocation signals of NF-κB. This allows NF-κB to translocate into the nucleus where it activates transcription of specific target genes, including IL-12 (18, 19). The enhanced NF-κB activation in NOD IL-12–overproducing cells has been shown to result from deviated IKK activation (6–8).

In addition to the IKK/NF-κB pathway, inflammatory stimuli also activate the MAPK pathways (21–23). There are three major groups of MAPKs in mammalian cells: ERK, JNK, and p38. Activation of MAPKs is typically mediated by sequential protein phosphorylation, namely, MAK3K → MAPK kinase (MAP2K) → MAPK. Activated MAPKs or their downstream effector kinases phosphorylate transcription factors, histones, and coactivators, which aid the induction of gene transcription by NF-κB. Additionally, p38 activates members of the MAPK-activated protein kinase (MK) family, especially MK-2, that rapidly stabilize mRNA transcripts in the cytoplasm (21–23). Controversy remains about the role of ERK or JNK in IL-12 production, whereas p38 activity has been shown to be essential for the expression of IL-12, although the underlying mechanisms remain elusive (24–27). Importantly, it remains to be determined whether deviated MAPKs activity synergizes with NF-κB to mediate elevated IL-12 production in NOD IL-12–overproducing cells.

NF-κB is also a key player in the expansion and survival of T cells. TCR engagement leads to rapid and robust activation of NF-κB. TCR ligation-induced NF-κB activation also depends on the IKK complex. Even though the mechanism by which IKK is recruited and activated upon TCR ligation is quite different from that in innate immunity, MAP3Ks, especially TAK1 and MEKK3, seem to be involved (18, 28–30). Inhibition of NF-κB activity with various strategies results in defective T cell expansion and increased cell death in activated T cells (31–33). It is unknown whether the signaling defects in NOD IL-12 overproducing cells also occur in NOD CD4⁺ Tconv cells, which might result in enhanced IKK/NF-κB activation and consequently mediate the higher rate of proliferation upon TCR ligation.

In this study, we report that the protein levels of MEKK3 are upregulated in NOD IL-12–overproducing cells and CD4⁺ Tconv cells, which contributes to the deviated activation of IKK and MAPKs and the consequent hyperresponsiveness.

Female C57BL/6 mice and female NOD/LtJ mice were purchased from Institutes of Experimental Animals, Academy of Chinese Medical Sciences (Beijing, China). All mice were maintained under specific pathogen-free conditions. All experiments were performed in accordance with institutional guidelines for animal care. Unless otherwise specified, mice were used between 3 and 4 wk of age and were therefore nondiabetic. Diabetes incidence was monitored by weekly measurement of venous blood glucose concentrations in nonfasting mice using Glucometer Elite strips (Bayer, Pittsburgh, PA). Mice with two consecutive blood glucose concentrations ≥11.3 mmol/l were considered diabetic. T1D develops in >80% of NOD/LtJ female mice between 12 and 16 wk of age in our colony.

All signaling inhibitors were obtained from CalBiochem (San Diego, CA) and were used at the concentration of 10 μM with 30 min pretreatment unless otherwise specified. Thioglycolate and LPS were purchased from Sigma-Aldrich (St. Louis, MO), and LPS was always used at the concentration of 100 ng/ml. FBS was from HyClone Laboratories (Logan, UT). M-CSF was from Cetus (Emeryville, CA). GM-CSF and IL-1β were from PeproTech (Rocky Hill, NJ). IL-4 and TNF-α were from R&D Systems (Minneapolis, MN). CpG (1826), TRIzol reagent, Moloney murine leukemia virus reverse transcriptase, oligo(dT) primer, and CFSE were from Invitrogen (Carlsbad, CA). FITC-anti–CD8, PE-anti–CD4, and Abs against CD3 and CD28 were from eBioscience (San Diego, CA). Abs against phospho-ERK, ERK, phospho-JNK, JNK, phospho-p38, phospho-IKKα/β, and phospho-MK-2 were from Cell Signaling Technology (Beverly, MA). Abs against actin, p38, IKKα/β, TAK1, apoptosis signal-regulating kinase 1 (ASK1), heat shock protein 90 (HSP90), IκBα, and MAPK phosphatase-1 (MKP-1) were from Santa Cruz Biotechnology (Santa Cruz, CA). Ab against MEKK3 was purchased from BD Biosciences (Franklin Lakes, NJ). An Ab against phospho-MEKK3 was generated as previously described (34). IL-2, CD4⁺ T cell isolation kit II, CD25 microbeads, CD8 microbeads, and MS columns were obtained from Miltenyi Biotec (Bergisch Gladbach, Germany). ATPlite assay kit was purchased from PerkinElmer (Waltham, MA).

Bone marrow-derived macrophages (BMMs) and BMDCs were obtained by culturing the nonadherent bone marrow cells in RPMI 1640 medium containing 15% (v/v) FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μM 2-ME with 100 ng/ml M-CSF or 10 ng/ml GM-CSF and 10 ng/ml IL-4 for 7 d, as described previously (5–11).

Peritoneal macrophages were obtained from peritoneal exudate cells by peritoneal lavage 4 d after i.p. injection of 2 ml thioglycolate, as described previously (5–7).

IL-12p70 or TNF-α levels in the supernatants were determined by using kits from eBioscience according to the manufacturer’s protocols.

Cells were stimulated with 100 ng/ml LPS, 10 ng/ml TNF-α, 10 ng/ml IL-1β, 1 μg/ml CpG, or Abs against CD3 (precoated, 5 μg/ml) and CD28 (2 μg/ml) before they were subjected to immunoblotting. Immunoblotting analysis was done as previously described (35).

RNA was extracted by using TRIzol reagent. cDNA was derived from 1 μg total RNA by reverse transcription using Moloney murine leukemia virus reverse transcriptase and oligo(dT) primer. PCR was performed on 25 ng cDNA with specific primers of IL-12p40 (17), MEKK3 (36), and GAPDH (37). All PCR reactions were carried out by a standard protocol for 25 (GAPDH) or 30 (IL-12p40 or MEKK3) cycles. Aliquots of PCR reactions were separated on a 1% (w/v) agarose gel and visualized with ultraviolet light after ethidium bromide staining.

To isolate CD4⁺CD25⁻ splenic cells, CD4⁺ T cells were first isolated (negative selection) from single-cell suspensions of spleens by using a CD4⁺ T cell isolation kit II. The cells were then incubated with CD25 microbeads. All CD25⁺ cells were depleted with an MS column.

To isolate CD4⁺ single-positive (SP) thymocytes and CD4⁺CD8⁺ double-positive (DP) thymocytes, single-cell suspensions of thymi were magnetically labeled with CD8 microbeads followed by separation with an MS column. Both the negative fraction and the positive fraction were collected and stained with FITC-anti–CD8 and PE-anti–CD4. CD4⁺ SP thymocytes and CD4⁺CD8⁺ DP thymocytes were then sorted using a FACSVantage (BD Biosciences).

In each case, purity of the cell separation was verified by immunostaining and FACS analysis. Cell purity was verified to be at least 95%.

Cells were labeled by incubation at 10⁶/ml in RPMI 1640 with 0.1 μM CFSE at 37°C for 20 min, washed, and resuspended in complete culture medium as stated above. Cells (1 × 10⁵/well) were then seeded in 96-well plates. Stimulation was effected by Abs against CD3 (precoated, 5 μg/ml) and CD28 (2 μg/ml). Proliferation was assessed by ATPlite assay or by flow cytometric analysis of CFSE dilution.

RNA interference was carried out with Accell small interfering RNA (siRNA; Dharmacon, Lafayette, CO) using the Accell siRNA delivery protocol. Briefly, NOD CD4⁺CD25⁻ splenic cells (5 × 10⁵/well) were seeded in 96-well plates. The growth media was removed and cells were transfected with 1 μM MEKK3 Accell siRNA or Accell Non-Targeting siRNA in 100 μl Accell siRNA Delivery Media (Dharmacon) supplemented with 100 U/ml IL-2. NOD BMMs (5 × 10⁴ per well) were seeded into 96-well plates on day 4 of macrophage development. After overnight incubation, the growth media was removed and cells were transfected as stated above in the presence of 100 ng/ml M-CSF.

BMMs from 3-wk-old B6 and NOD mice were activated with LPS for various periods of time. Immunoblotting analysis revealed that LPS induced rapid and robust phosphorylation of p38 (P-p38), JNK (P-JNK), and ERK (P-ERK) within the tripeptide motif (Thr-X-Tyr), which is required for the activation of these kinases (21–23), in both B6 and NOD BMMs (Fig. 1A, 1B). The initial activation of p38, JNK, and ERK with LPS treatment for 15–30 min was similar between B6 and NOD BMMs, but p38 activation was much more persistent over time in NOD BMMs (Fig. 1A, 1B). NOD BMMs also exhibited slightly higher levels of JNK and ERK activation with LPS treatment for >60 min (Fig. 1B). Prolonged p38 activation was not only observed in BMMs from 3-wk-old NOD mice when the autoimmune process has barely started, but it was also observed in BMMs from diabetic NOD mice (Fig. 1C). More persistent p38 activation in response to LPS was also observed in BMDCs (Fig. 1D) and peritoneal macrophages (Fig. 1E) from 3-wk-old NOD mice. Taken together, these data suggest that more persistent p38 activation occurs in all the NOD IL-12–overproducing cells in response to LPS.

Similar to B6 BMMs, NOD BMMs showed no detectable basal IL-12 production (data not shown). However, NOD BMMs produced more IL-12 with LPS treatment for 24 h compared with their B6 counterparts, which became more evident with time (Fig. 2A). To test whether aberrant p38 activation contributes to elevated IL-12 production, NOD BMMs were pretreated with inhibitors against MAPKs before the cells were stimulated with LPS. ELISA assay revealed that ERK inhibitor U0126 (38) and JNK inhibitor SP600125 (39) had no significant effect on elevated IL-12 production, whereas p38 inhibitor SB203580 (40) significantly suppressed IL-12 overproduction (Fig. 2B). Additionally, another p38 inhibitor, SB239063 (41), also suppressed elevated IL-12 production in a dose-dependent manner (Fig. 2C). Decreased IL-12 production in the presence of p38 inhibitors was not due to decreased viability (Supplemental Fig. 1). Consistently, even though basal IL-12 production was not detectable in both B6 and NOD peritoneal macrophages (data not shown), NOD peritoneal macrophages exhibited elevated IL-12 production in response to LPS, which was significantly inhibited by SB239063 (Supplemental Fig. 2). Thus, deviated p38 activation indeed contributes to IL-12 overproduction.

p38 activation leads to its translocation into the nucleus, where it phosphorylates and activates several transcription factors, including NF-κB. Moreover, MK-2 is the main target of p38 that is involved in posttranscriptional regulation of cytokine expression (21–23). The role of NF-κB in the expression of IL-12 has been well established (16, 17). However, it remains unclear whether MK-2 contributes to IL-12 production. LPS induced more persistent MK-2 activation (as indicated by MK-2 phosphorylation [P-MK-2]) in NOD BMMs compared with their B6 counterparts (Fig. 2D). SB203580 blocked LPS-induced MK-2 activation in both B6 BMMs and NOD BMMs (Fig. 2E). Thus, deviated MK-2 activation in NOD BMMs was mediated by aberrant p38 activation. To explore the role of posttranscriptional mechanisms in the expression of IL-12, we measured IL-12p40 mRNA in transcription-arrested B6 BMMs (actinomycin D) in the absence or presence of LPS with an RT-PCR assay (42). When B6 BMMs were stimulated with LPS for 3 h followed by wash-out and a fresh change of medium, the elevated levels of IL-12p40 mRNA declined rapidly in the presence of actinomycin D. However, when LPS was added to fresh medium, the decrease in IL-12p40 mRNA was reversed even in the presence of actinomycin D. Thus, LPS stabilized IL-12p40 mRNA. The stabilizing effect was abrogated by SB203580 or MK-2 inhibitor CMPD1 (43) (Fig. 2F). Additionally, ELISA assay revealed that CMPD1 suppressed elevated IL-12 production in NOD BMMs in a dose-dependent manner (Fig. 2G). Under the same conditions, CMPD1 showed marginal effects on the viability of NOD BMMs (Supplemental Fig. 1). Taken together, these data suggest that deviated p38 activation contributes to elevated IL-12 production via, at least partially, MK-2–mediated stabilization of IL-12p40 mRNA.

Recent studies have demonstrated a critical role of MKP-1 in the negative control of MAPK-regulated inflammatory reactions (44, 45). Because LPS induced more persistent p38 activation in NOD IL-12–overproducing cells, it was of interest to investigate whether LPS-induced expression of MKP-1 is defective in these cells. Immunoblotting assay revealed that LPS-induced expression of MKP-1 occurred as soon as 30 min after the treatment and became evident 60 min after the treatment in both B6 and NOD BMMs (Fig. 3A, 3B). The induction of MKP-1 was similar in B6 and NOD BMMs with LPS treatment for no longer than 2 h (Fig. 3A). With LPS treatment for longer times, the protein levels of MKP-1 in NOD BMMs were even significantly higher than those in B6 BMMs (Fig. 3B). Thus, more persistent p38 activation in NOD IL-12–overproducing cells is not due to defective MKP-1 induction.

It has been demonstrated that NOD IL-12–overproducing cells also exhibited aberrant IKK/NF-κB activation (6–8). Because deviated p38 activation in NOD IL-12–overproducing cells does not result from defective MKP-1 induction, it is possible that the origin of the deviated activation may come from an upstream abnormality. To address this matter, B6 and NOD BMMs were activated with TNF-α, IL-1β, or CpG for various periods of time. Immunoblotting analysis revealed that compared with their B6 counterparts, NOD BMMs exhibited enhanced and more persistent p38 activation in response to all of these inflammatory stimuli (Fig. 3C, 3D). These data further suggest that the deviated activation may be caused by alterations at an upstream signaling complex.

To further track the origin of the deviated activation of these signaling pathways, CD4⁺CD25⁻ splenic cells were purified from 3-wk-old B6 and NOD mice. The phenotype of these Tconv cells was first explored. These NOD CD4⁺CD25⁻ splenic cells exhibited similar expression levels of CD62L and CD69 (CD62L⁺CD69⁻), but enhanced expression levels of CD44 compared with the B6 counterparts (Supplemental Fig. 3A). Because a portion of CD4⁺CD25⁻ splenic cells isolated from 8-wk-old NOD mice showed even stronger CD44 expression (Supplemental Fig. 3B) and CD4⁺ SP thymocytes isolated from 3-wk-old NOD mice also exhibited enhanced levels of CD44 expression compared with their B6 counterparts (Supplemental Fig. 3C), the enhanced CD44 expression in CD4⁺CD25⁻ splenic cells isolated from 3-wk-old NOD mice should result from some intrinsic defects, rather than from enhanced T cell activation. These data suggest that most of these Tconv cells isolated from 3-wk-old B6 and NOD mice are naive T cells. These CD4⁺ Tconv cells were then stimulated with Abs against CD3 and CD28. Immunoblotting assay revealed hyperactivation of IKK (as indicated by IKKα/β phosphorylation [P-IKKα/β]) and all the three major groups of MAPKs in NOD CD4⁺CD25⁻ splenic cells, compared with their B6 counterparts, upon TCR ligation (Fig. 4A). The hyperactivation appeared quite early in differentiation, as purified CD4⁺ SP thymocytes from NOD mice also exhibited deviated activation of IKK and MAPKs (Fig. 4B). Consistently, NOD CD4⁺ Tconv cells exhibited augmented proliferation upon TCR ligation as measured by a CFSE dilution assay (Fig. 4C) and an ATPlite assay (Fig. 4D). To test whether deviated activation of IKK and MAPKs contributes to NOD CD4⁺ Tconv cell proliferation, NOD CD4⁺CD25⁻ splenic cells were pretreated with inhibitors against IKK or MAPKs prior to TCR ligation. Proliferation of NOD CD4⁺ Tconv cells was abrogated by IKK inhibitor Bay 11-7082 (46), SP600125, and U0126 (Fig. 4E). Taken together, these data suggest that the hyperactivation of IKK and MAPKs mediates the higher rate of proliferation of NOD CD4⁺ Tconv cells.

Our previous data suggest that the deviated activation of the IKK/NF-κB pathway and the MAPK pathways may result from defects in certain common upstream signaling components. Even though the mechanism by which IKK is recruited and activated upon TCR ligation is distinct from that in innate immune cells, MAP3Ks, especially TAK1 and MEKK3, seem to be involved in both processes (18, 28–30). Thus, it is highly possible that the defects occur at the level of MAP3K, especially TAK1 and MEKK3. To test this scenario, the expression of TAK1 and MEKK3, as well as another MAP3K, ASK1, was evaluated in both B6 and NOD BMMs. An immunoblotting assay revealed that the protein levels of TAK1 and ASK1 were similar in B6 and NOD BMMs (Fig. 5A). However, NOD BMMs exhibited significant higher levels of MEKK3 protein compared with their B6 counterparts (Fig. 5A). Furthermore, the overexpression of MEKK3 was not associated with aberrant expression of HSP90 (Fig. 5A), which has been demonstrated to mediate the stabilization of MEKK3 protein (47). The overexpression of MEKK3 might result from enhanced transcription because NOD BMMs exhibited higher levels of MEKK3 mRNA than did their B6 counterparts (Fig. 5A). MEKK3 overexpression was also seen in NOD BMDCs and peritoneal macrophages (Fig. 5B). As these IL-12–overproducing cells were isolated from 3-wk-old mice, when the autoimmune process has barely started, MEKK3 overexpression is unlikely the indirect result of ongoing autoimmunity.

To investigate the expression of MEKK3 in T cells, DP thymocytes, CD4⁺ SP thymocytes, and CD4⁺CD25⁻ splenic cells were isolated and purified from 3-wk-old B6 and NOD mice. Immunoblotting assay revealed that MEKK3 was barely detectable in both B6 and NOD DP thymocytes (Fig. 5C). The protein levels of MEKK3 were upregulated during the development of T cells (Fig. 5C). Compared with their B6 counterparts, NOD CD4⁺ SP thymocytes as well as CD4⁺CD25⁻ splenic cells exhibited higher protein levels of MEKK3 (Fig. 5C), suggesting that MEKK3 overexpression is independent of autoimmune inflammation. The overexpression of MEKK3 in NOD T cells seemed to be unique since these cells showed similar protein levels of TAK1, ASK1, and HSP90 to their B6 counterparts (Fig. 5D).

MEKK3 overexpression seems to lead to enhanced activation since NOD BMMs exhibited higher levels of MEKK3 phosphorylation (P-MEKK3) than did B6 BMMs (Fig. 5E). Additionally, treatment with LPS for 24 h (1440 min) showed marginal effect on the protein levels of MEKK3 in both B6 and NOD BMMs (Fig. 5E), further suggesting that MEKK3 overexpression is independent of autoimmune inflammation. Similar MEKK3 hyperactivation was also observed in NOD BMDCs and peritoneal macrophages (Fig. 5E). Consistently, NOD CD4⁺CD25⁻ splenic cells showed more enhanced MEKK3 activation than did their B6 counterparts upon TCR ligation (Fig. 5F).

To further test whether MEKK3 overexpression mediates the hyperresponsiveness of NOD IL-12–overproducing cells and CD4⁺ Tconv cells, we used MEKK3 Accell siRNA to knock down MEKK3 overexpression. Immunoblotting assay revealed that MEKK3 siRNA indeed led to decreased MEKK3 protein level in NOD BMMs, which was associated with decreased MEKK3 activation in response to LPS (Fig. 6A). Knock-down of MEKK3 in NOD BMMs showed little effect on the initial activation of IKK, but dampened the maintenance of activation, as demonstrated by IKKα/β phosphorylation (Fig. 6A), IκBα phosphorylation (P-IκBα; Supplemental Fig. 4A), and IκBα degradation and resynthesis (Fig. 6A). Similar reversal of the persistent activation of p38, but not the initial activation, by MEKK3 siRNA was also observed (Fig. 6A). Under the same conditions, MEKK3 siRNA showed marginal effects on the activation of JNK and ERK (Fig. 6A). The reversal of deviated IKK and p38 activation was associated with significantly reduced IL-12 production in response to LPS. However, the same treatment showed marginal effect on LPS-induced TNF-α production (Fig. 6B). The basal levels of IL-12 and TNF-α production in NOD BMMs with MEKK3 siRNA treatment remained undetectable (data not shown). Taken together, these data suggest that hyperresponsiveness of IL-12–producing cells in NOD mice may be due to MEKK3 overexpression.

MEKK3 siRNA also specifically inhibited the overexpression of MEKK3 and the consequent MEKK3 hyperactivation in NOD CD4⁺CD25⁻ splenic cells (Fig. 6C). MEKK3 knock-down not only abolished the hyperactivation of IKK, as demonstrated by IKKα/β phosphorylation (Fig. 6C), IκBα phosphorylation (Supplemental Fig. 4B), and IκBα degradation (Fig. 6C), but it also reversed the hyperactivation of all the three major groups of MAPKs upon TCR ligation (Fig. 6C). However, MEKK3 knock-down showed little effect on STAT3 activity, as demonstrated by STAT3 phosphorylation (P-STAT3; Fig. 6C), suggesting that the regulation of IKK and MAPKs activity by MEKK3 in CD4⁺ Tconv cells was specific. The impaired IKK and MAPKs hyperactivation was associated with significantly decreased proliferation of NOD CD4⁺CD25⁻ splenic cells upon TCR ligation as measured by ATPlite assay and CFSE dilution assay (Fig. 6D, 6E). Taken together, these data suggest that hyperresponsiveness of CD4⁺ Tconv cells in NOD mice results from MEKK3 overexpression.

NOD mice are very susceptible to autoimmune diseases that extend beyond diabetes. Some of this susceptibility stems from defective T cell tolerance induction in the thymus (48, 49), but it also results from, at least in part, deficient control of autoreactive T cells by regulatory populations (15), in particular regulatory T (Treg) cells. Recent progress has revealed that the defect lies in an overreactivity of CD4⁺ Tconv cells, not in the Treg cell side of the balance (15). The root of the hyperresponsiveness of NOD CD4⁺ Tconv cells is unknown. In contrast, elevated IL-12 production in NOD mice has been long demonstrated to contribute to aberrant T cell differentiation (5–12). The increased IL-12 secretion has been disclosed as a direct result of enhanced activation of the IKK/NF-κB pathway (6–8). Because NF-κB is also a key player in the expansion and survival of T cells (18, 28–30), it is possible that the signaling defects that lead to deviated IKK/NF-κB activation in NOD IL-12–overproducing cells might also occur in NOD CD4⁺ Tconv cells, which results in a higher rate of proliferation. Our work indicates that this is indeed the case. The signaling defect(s) could be attributed to, at least partially, the overexpression of a single MAP3K, namely MEKK3.

MEKK3 overexpression seems to result in preferential, more persistent activation of IKK and p38 in NOD IL-12–overproducing cells. The underlying mechanism is of interest. MEKK3 has been demonstrated to contribute to the initial activation of IKK and MAPKs in response to inflammatory stimuli by collaborating with TAK1 (20, 50). Moreover, MEKK3 plays a pivotal role in the maintenance of IKK and MAPKs (in particular p38) activation in a TAK1-independent manner (20, 50). Therefore, MEKK3 activation is essential but not adequate for the initial activation of IKK and MAPKs in response to inflammatory stimuli (20, 50). TAK1 is not upregulated in NOD IL-12–overproducing cells, which might be the speed-limiting factor that hinders overexpressed and hyperactivated MEKK3 from contributing much to the initial activation of IKK and MAPKs. In contrast, the maintenance of IKK and MAPKs (in particular p38) activation depends on MEKK3 only (20, 50). Thus, overexpressed and hyperactivated MEKK3 leads to more persistent activation of IKK and MAPKs (in particular p38). Our data suggest that deviated p38 activity synergizes with the IKK/NF-κB pathway to mediate elevated IL-12 production with a mechanism involving MK-2–mediated stabilization of IL-12p40 mRNA. Thus, these results provide a novel mechanism by which p38 contributes to IL-12 production. As for CD4⁺ Tconv cells, MEKK3 overexpression leads to significant hyperactivation of IKK and all the three major groups of MAPKs. Hyperactivation of MAPKs in NOD CD4⁺ Tconv cells synergizes with the IKK/NF-κB pathway to mediate higher rates of proliferation. However, JNK and p38 are also key players in activation-induced cell death of CD4⁺ T cells (21–23). They might simultaneously neutralize the prosurvival effect of IKK/NF-κB hyperactivation in NOD CD4⁺ Tconv cells. Indeed, activation-induced cell death was not reduced in NOD CD4⁺ Tconv cells (data not shown). MEKK3 knock-down reversed the deviated activation of both the MAPK pathways and the IKK/NF-κB pathway. Consequently, MEKK3 knock-down led to reduced IL-12 production in NOD IL-12–overproducing cells and decreased proliferation of NOD CD4⁺ Tconv cells.

NOD macrophages, which reproducibly exhibit IL-12 overproduction irrespective of their sources, are the major source of the elevated IL-12 production in NOD mice (5–8, 12). Accordingly, both BMMs and peritoneal macrophages from NOD mice exhibit MEKK3 overexpression and hyperactivation. In contrast, DC are very heterogeneous. Even though NOD BMDCs produce more IL-12 than do their B6 counterparts in response to inflammatory stimuli (9–11), such a defect does not happen to other DC subsets from NOD mice (12). The in vivo equivalents of BMDCs are monocyte-derived inflammatory DC, which are absent in steady-state mice. During GM-CSF–dependent inflammation, a small subset of blood-derived monocytes is recruited to lymphoid tissues and differentiates into inflammatory DC (51). Therefore, BMDCs make no contribution to steady-state DC (51). Accordingly, immunoblotting assays revealed that MEKK3 expression and LPS-induced activation of MEKK3 and p38 were similar between CD11^high splenic DC from 3-wk-old B6 and NOD mice (Supplemental Fig. 5). These data are consistent with our finding that MEKK3 overexpression results in IL-12 overproduction. It remains unknown whether NOD inflammatory DC also exhibit MEKK3 overexpression, whether NOD inflammatory DC contribute to elevated IL-12 production in NOD mice during GM-CSF–dependent inflammation, and what the physiological relevance might be. Further studies are required to clarify these issues.

MEKK3 overexpression appeared to be intrinsic to the NOD background. Because these IL-12–overproducing cells and CD4⁺ Tconv cells were isolated from 3-wk-old mice, when the autoimmune process has barely started, MEKK3 overexpression is unlikely the indirect result of ongoing autoimmunity. Furthermore, treatment of IL-12–overproducing cells with LPS for 24 h showed no effect on the protein level of MEKK3, which further suggests that MEKK3 overexpression is independent of autoimmune inflammation. MEKK3 overexpression was also observed in CD4⁺ SP thymocytes from NOD mice. Thus, MEKK3 overexpression represents a cell-intrinsic feature in T cells of NOD mice rather than a secondary adaptation to particular homeostatic conditions in secondary lymphoid organs.

The mechanism underlying MEKK3 overexpression is unknown. Our results indicate that it was due to enhanced transcription, but not to enhanced protein stability. The murine mekk3 gene is on chromosome 11 locus. NOD congenic analysis has revealed that chromosome 11 locus contributes to diabetes susceptibility (52). Interestingly, it has been demonstrated that the underlying genetic element responsible for the IL-12 overexpression lies within chromosome 11 locus, even though the detailed locus remains to be identified (53). It is possible that the defective genetic element lies within the mekk3 gene, which directly leads to its overexpression. Another possibility is that the defective genetic element lies outside mekk3 gene, but nearby in chromosome 11 locus, which might alter the characteristic of certain transcription factors or other elements that are implicated in the regulation of mekk3 transcription. Future studies are required to address these issues.

With regard to the human context, it has been recently demonstrated that Tconv cells from the T1D patients have a lower threshold of activation and are more refractory to Treg cell inhibition than are those from age- and HLA-matched controls (54, 55). Based on these findings, it has been proposed that targeting Tconv cells should be pursued when considering immunomodulating strategies based on infusing Treg cells in T1D patients (55, 56). A prerequisite for such new approaches is a better understanding of the mechanisms by which Tconv cells attain overactivity. Among the genes known to be associated with T1D that could also alter Tconv cell responsiveness are the HLADRB1 0301 and 0401 alleles, as well as the PTPN22 1858T variant of the protein tyrosine phosphatase Lyp. However, Tconv cells derived from healthy subjects who carry these genes show no resistance to suppression (54). Thus, the root of the altered Tconv cell responsiveness in T1D patients remains unknown. Our work with NOD mice might provide some clue to resolve this puzzle.

We thank Ming Yu and Fengjun Xiao for technical assistance.

Disclosures The authors have no financial conflicts of interest.