The MAPK-Activated Kinase MK2 Attenuates Dendritic Cell–Mediated Th1 Differentiation and Autoimmune Encephalomyelitis

Dendritic cells (DCs) are essential mediators between innate and adaptive immunity. They elicit cytokine-driven immune responses upon invasion of pathogens. A switch of DCs from tolerance maintenance to immune stimulation, referred to as maturation, may be initiated by pathogen- (1) or damage-associated (2) molecular patterns, proinflammatory cytokines (3), or CD40/CD40L signaling (4). Pathogen-associated molecular pattern recognition, similar to binding of LPS to TLR4 on DCs, must be stringently regulated, as excessive expression of signaling components as well as proinflammatory cytokines can have devastating effects on the host, resulting in chronic inflammatory diseases or autoimmune disorders. Therefore, it is essential for negative regulators to act on multiple levels within the TLR signaling cascade involving IFN regulatory factors, NF-κB transcription factors, and the MAPK pathways (5).

ERK1/2 and p38 MAPKs are critical for both pro- and anti-inflammatory immune responses. They direct the production of cytokines, which are essential for the differentiation of naive CD4⁺ T cells into Th1, Th17, or regulatory T cell subsets (6–10). The mitogen- and stress-activated kinases and MAPK-activated protein kinases (MKs), downstream substrates of ERK1/2 and p38, streamline TLR-driven immune responses toward specific T cell effector functions (11). In this context, MK2 contributes to inflammation due to its essential role in promoting the expression of TNF-α, IL-1β, and IL-6 in macrophages (12) and sustained NF-κB activation (13). However, recent observations indicate an additional anti-inflammatory function of MK2 by regulating IL-10–mediated STAT3 activation in macrophages (14, 15).

In the early phases of their maturation process, DCs acquire a proinflammatory mode of action, which is characterized by the secretion of TNF-α, IL-1α/β, and IL-12, molecules critical for the regulation of adaptive immune responses (16–18). IL-12 is released during ∼24 h following exposure of a DC to LPS. During this phase DCs trigger Th1 responses and as a consequence initiate CD8⁺ CTL-dominated immune responses in vitro (19, 20) and in vivo (21, 22). In contrast to their effector cell priming capacity, DCs subsequently assume an anti-inflammatory mode of action (19, 22). This phase is characterized by enhanced activity of the tryptophan-metabolizing enzyme IDO and a high secretion level of IL-10, which support an immune regulatory DC phenotype mediated by the autocrine IL-10/STAT3 signaling cascade (23). Both IDO and IL-10 expressed by DCs contribute to the priming of regulatory T cells (24, 25).

In accordance with the time-dependent phenotypic changes of maturing DCs, we show that LPS exposure causes accumulation of MK2 protein in DCs up to approximately 1 d following stimulation. At this time, anti-inflammatory signaling molecules of the IL-10/STAT3 pathway are strongly induced. Based on this observation we sought to determine a potential regulatory function of MK2 in DCs. We found MK2-mediated regulation of effector T cell functions to be primarily executed via negative feedback signaling on p38 and positive cross-regulation of ERK1/2 activity. Down-modulation of p38 predominantly impairs the autocrine IL-1α/p38 axis, which terminates IL-1 and IL-12 secretion and enhances IL-10/STAT3 signaling. As a consequence, MK2 inhibits DC-mediated differentiation of naive CD4⁺ T cells toward Th1 effector mechanisms as shown in vivo by severe experimental autoimmune encephalomyelitis (EAE) in DC-specific MK2 knockout mice compared with wild-type (WT) littermates. Hence, we reveal MK2 as a key regulator of inflammatory mechanisms active in DCs.

Male OT-I (C57BL/6-Tg(TcraTcrb)1100Mjb/Crl) or OT-II (C57BL/6-Tg(TcraTcrb)425Cbn/Crl) transgenic mice, purchased from the Research Institute for Laboratory Animal Breeding, University of Vienna (Himberg, Austria), were housed at the animal care unit of the Department of Pharmacology, Medical University of Vienna (Vienna, Austria). Conditional MK2 knockout mice were bred at the Division for Laboratory Animal Science and Genetics of the Medical University of Vienna (Himberg, Austria) and obtained by crossing Mapkapk2^{tm1a(EUCOMM)Hmgu}/H mice (provided by the European Mouse Mutant Archive) with B6.129S4-Gt(ROSA)26Sor^tm1(FLP1)Dym/RainJ (obtained from The Jackson Laboratory) for Flp deletion and further crossing the floxed strain with mice expressing the Cre recombinase under the control of the CD11c promoter (B6.Cg^{Tg(Itgax-Cre)1-1Reiz}/J, obtained from The Jackson Laboratory). Conditional MK2 knockout mice are hereafter referred to as CD11cCre-Mk2^flox/flox or MK2^ΔDC mice. For genotyping, DNA was extracted from lysed (buffer including proteinase K, Invitrogen) tail tip biopsies and subjected to direct PCR using innuTaq DNA polymerase (Analytik Jena). Specific PCRs were performed with the following primers: MK2 primer, forward, 5′-GACATGTGGTCCTTGGGTGTCATCATG-3′, reverse, 5′-GAGATGGCAAGGCCGTGATTGGAATAG-3′; FLP deletion primer, forward, 5′-CCCTTTCTTGTCCTTGAAGTGGTTCC-3′, reverse, 5′-CTCTGCGCGCCTTCGGACAGTCCTGG-3′; FLPe primer, forward, 5′-CACTGATATTGTAAGTAGTTTGC-3′, reverse, 5′-CTAGTGCGAAGTAGTGATCAGG-3′; CD11c Cre primer, forward, 5′-ACTTGGCAGCTGTCTCCA AG-3′, reverse, 5′-GCGAACATCTTCAGGTTCTG-3′. All in vivo mouse experiments were approved by the Institutional Review Board of the Medical University of Vienna. Bone marrow–derived DCs (BMDCs) were generated from MK2^−/− (12) and WT C57BL/6 mice as previously described (22). Splenic DCs were isolated from spleens of C57BL/6 mice digested with 10 mg/ml collagenase D (Roche) and 10 μg/ml DNAse I (Roche). CD11c⁺ cells were selected from single-cell suspensions by MACS (Miltenyi Biotec). For in vivo applications and in vitro assays murine DCs were pulsed with 1 μg/ml MHC class I peptide OVA_257–264 (SIINFEKL, H2-K^b) and/or the MHC class II peptide OVA_323–399 (both from Bachem) for 1 h prior to maturation. Then, peptide-pulsed DCs were stimulated with 100 ng/ml LPS (Escherichia coli strain O111:B4, Calbiochem), 2.5 μg/ml R848 (Santa Cruz Biotechnology), or 2 μg/ml poly(I:C) (Sigma-Aldrich) for 4 h. The use of human peripheral blood from healthy adult volunteers was approved by the Institutional Review Board of the St. Anna Children’s Cancer Research Institute and conducted according to the Declaration of Helsinki. Monocytes and T cells were isolated from peripheral blood and DCs differentiated in vitro from monocytes in the presence of IL-4 and GM-CSF as described previously (26). For cytokine profiling and in vitro proliferation assays, human DCs were transfected with 100 pmol/10⁶ DCs MK2-specific (small interfering RNA [siRNA] pool: CGAAUGGGCCAGUAUGAAU, GUUAUACACCGUACUAUGU, GGCAUCAACGGCAAAGUUU, CCACCAGCCACAACUCUUU) or nontargeting control siRNA (all reagents from Dharmacon) 16 h prior to the stimulation with 100 ng/ml LPS. For MK2 or p38 inhibition, DCs were treated with 30 μM MK2-I3 or 10 μM SB203580 (both from Sigma-Aldrich) 30 min prior to LPS stimulation for 4 h, and after washing and counting they were further phenotyped or cocultured with CD4⁺ OT-II cells. IL-1α blocking experiments were performed with 200 ng/ml anti-mouse IL-1α or Armenian hamster IgG isotype control Abs (eBioscience).

Gene expression profiling was performed as previously described using human DCs (Gene Expression Omnibus no. GSE11327, http://www.ncbi.nlm.nih.gov/geo/) (16). Protein expression and phosphorylation were analyzed in 10⁶ DCs lysed in RIPA buffer. Cell lysates were separated by electrophoresis using 10% acrylamide gels and then transferred onto nitrocellulose membranes (Whatman). Proteins were probed with the following Abs: MK2, MK2 phosphorylated at Thr³³⁴, p38, p38 phosphorylated at Thr¹⁸⁰/Tyr¹⁸², p44/42 (ERK1/2), p44/42 (ERK1/2) phosphorylated at Thr²⁰²/Tyr²⁰⁴, STAT3, STAT3 phosphorylated at Tyr⁷⁰⁵, heat shock protein (HSP)27, HSP27 phosphorylated at Ser⁸² (all from Cell Signaling Technology), and GAPDH (Ambion) followed by peroxidase- or DyLight 800–conjugated anti-rabbit or anti-mouse IgG (Pierce). Densitometric semiquantitative analysis of protein phosphorylation on blots was performed using ImageJ (National Institutes of Health, Bethesda, MD). Quantification of phosphorylated protein was performed by p-p38–, p-ERK1/2–, and p-STAT3–specific ELISA assays (eBioscience).

C57BL/6 WT and OT-II mice were immunized s.c. close to the inguinal lymph node (LN) with 5 × 10⁶ SIINFEKL peptide or OVA_323–399 peptide-pulsed, LPS-stimulated DCs with or without 10⁶ OT-II cells, which were purified by negative depletion using CD4⁺ MACS (Miltenyi Biotec). On day 3 T cells were isolated from in vitro cultures or LNs and analyzed directly for T-bet expression by intracellular staining. Further cells were restimulated with anti-CD3/CD28 (BD Pharmingen) and PMA/ionomycin (Sigma-Aldrich) treatment together with GolgiPlug or GolgiStop (BD Biosciences) and analyzed for intracellular cytokines. The following Abs were used for transcription factor, cytokine, and surface marker staining: anti-mouse CD4 PerCP (clone RM4-5, BD Pharmingen), CD25 PE-Cy7 (clone PC61.5), T-bet PerCP-Cy5.5 (clone 4B10), IL-2 eFluor 450 (clone JES6-5H4), IL-10 Alexa Fluor 647 (clone JES5-16E3), and IFN-γ PE (clone XMG1.2, all from eBioscience). T cell and DC supernatants were analyzed using the FlowCytomix system (eBioscience) following the manufacturer’s protocol. FACS acquisition was performed on an LSR II flow cytometer (BD Biosciences). Further analysis was performed using FlowJo software Version 6.7.1 (Tree Star).

Murine OT-II or OT-I splenocytes and human peripheral blood leukocytes from Elutra products were enriched for CD4⁺ or CD8⁺ T cells by negative depletion using MACS (Miltenyi Biotec). Both mouse and human T cells were labeled with a proliferation tracker (CFSE, Sigma-Aldrich) at a final concentration of 7 μM. Mouse T cells (50,000/200 μl) were cocultured with SIINFEKL or OVA_323–399 peptide-loaded LPS-stimulated DCs (25,000/200 μl) and analyzed after 3 d. Human T cells (50,000/200 μl) were cocultured with allogeneic LPS-stimulated DCs (25,000/200 μl) and analyzed after 6 d for CFSE. The absolute number of proliferating T cells was assessed by CFSE dilution using the Trucount system (BD Biosciences) and the following Abs: anti-mouse CD4 PerCP-Cy5.5 (clone RM4-5), CD8a allophycocyanin–eFluor 780 (clone 53-6.7), CD25 PE-Cy7 (clone PC61.5, all from eBioscience), anti-human CD8 allophycocyanin-Cy7 (clone SK-1, BD Biosciences), and anti-human CD25 Alexa Fluor 647 (clone MEM181, AbD Serotec).

For in vivo experiments CD4⁺ OT-II transplanted WT mice were immunized with a mixture of LPS (10 μg/mouse)/OVA_323–399 peptide (2 μg/mouse)/MK2-I3 (20 μg/mouse)/anti–IL-1α (400 ng/mouse). Control mice were injected with vehicle (DMSO, Sigma-Aldrich) and Armenian hamster IgG isotype control Abs (eBioscience). One hundred thousand cells from DC/OT-II splenocyte cocultures or draining LN cells from immunized mice were restimulated with 1 μg/ml OVA_323–399 peptide, lymphocytic choriomeningitis virus (LCMV) glycoprotein peptide 61–80 (AnaSpec), or 1 μl anti-CD3/CD28 Dynabeads (Life Technologies)/100,000 cells in 96-well MultiScreen_HTS IP filter plates (MSIPS4W10, Millipore) using CTL est serum-free medium. IL-17 and IFN-γ ELISPOTs were performed according to the manufacturer’s protocol (Mabtech) using a 5-bromo-4-chloro-3-indolyl phosphate/NBT liquid substrate system (Sigma-Aldrich).

Killing of target cells was performed according to a previously described protocol (27). Briefly, mice were immunized s.c. close to the inguinal LN with OVA_257–264 (SIINFEKL) and OVA_323–399 peptide-pulsed, LPS-stimulated DCs. Syngeneic target cells were prepared by combining splenocytes loaded with 2.5 μM CFSE and 1 μg/ml of OVA_257–264 with splenocytes loaded with 0.25 μM CFSE and 1 μg/ml mTRP_181–188 (VYDFFVWL; H2-K^b, derived from murine tyrosine-related protein-2; Bachem) control peptide. On day 4, 10⁷ target cells were administered by tail vein injection. Six hours later, draining LNs were analyzed for CFSE⁺ target cells. Reduction of OVA_257–264 peptide-pulsed cells (%) = [1 − (% CFSE^high/% CFSE^low)] × 100 and was calculated and expressed as killing (%) in relationship to PBS-injected control mice.

Active EAE was induced in MK2^ΔDC mice and their respective WT littermate controls. At day 0 immunization was performed by s.c. injection of 75 μg myelin oligodendrocyte glycoprotein (MOG)_35–55 peptide (Charité, Berlin, Germany) in 75 μl H₂O emulsified in 75 μl CFA, which was enriched with 10 mg/ml Mycobacterium tuberculosis (H37Ra, Difco/BD Pharmingen). Pertussis toxin (200 ng) from Bordetella pertussis (List/Quadratech) was administered i.p. at days 0 and 2 after immunization. For boosting, immunization was repeated according to the same scheme on days 22 and 24, respectively. Clinical signs of EAE were assessed by the following classical EAE disease scores: 0, no disease; 1, tail weakness; 2, paraparesis; 3, paraplegia; 4, paraplegia with forelimb weakness; 5, moribund or dead animals. For determination of Th1/Th17 contribution to EAE, mice were immunized i.p. with 1 μg/20 g body weight rat anti-mouse IFN-γ (clone XMG1.2) or rat IgG1 κ isotype control Ab (both from BioLegend) on day 6. At the day of euthanasia, splenocytes and lymphocytes (draining inguinal LNs) were harvested and restimulated in vitro with 5 μg/ml MOG_35–55 peptide, 1 μg/ml LCMV control peptide, or 1 μl anti-CD3/CD28 Dynabeads (Life Technologies)/100,000 cells for ELISPOT analysis of cytokine secretion as described above.

Mice were anesthetized by s.c. injection of ketamine/xylazine and perfused intracardially with PBS. Brain and spinal cords were isolated and fixed in 4% buffered formalin. Fixed tissues were dissected and embedded in paraffin before sectioning. Sections were stained with Klüver–Barrera (KLB) and H&E using standard procedures. Anti-CD3 (AbD Serotec) was used for immunohistochemistry of T cell infiltration. At least three cross-sections from each animal were used for histological evaluation. The inflammatory index represents the total number of mononuclear infiltrates per cross-section. For evaluation of demyelinated area, total and demyelinated area of each cross-section in the KLB myelin staining was measured. For immunohistological evaluation, all positive cells were counted. Image J (National Institutes of Health) was used for the above-mentioned histological analysis.

Two-tailed paired and unpaired Student t tests as well as a Mann–Whitney U and two-way ANOVA analysis (to analyze two groups over time) were used to determine exploratory p values using GraphPad Prism version S.02 for Windows (GraphPad Software, San Diego, CA). All data are given as mean ± SEM.

LPS is known to induce a variety of immunologically active genes enabling DCs to crosstalk via membrane-bound molecules or soluble cytokines in a proinflammatory but also in an anti-inflammatory mode. As recently described, stimulation of human monocyte-derived DCs with LPS together with IFN-γ indeed strongly induced proinflammatory genes (16). In the same setup, we observed a robust induction of IL1, IL12, TNF, and IL6 in human DCs over a time period of 6–48 h following LPS exposure. Moreover, starting 12 h after LPS stimulation, DCs acquired anti-inflammatory properties as indicated, for example, by the upregulation of IL10 or STAT3 (Fig. 1A). A similar transcription profile hinted at MK2 being involved as a critical factor in DC-mediated anti-inflammatory mechanisms, because MK2 showed peak expression at 24 h after LPS/IFN-γ stimulation, similar to known immune regulatory molecules. LPS, as a TLR4 agonist, and not IFN-γ appeared to be the dominant signal for the upregulation of the MK2 protein as well as for the induction of the p38/MK2 signaling pathway in human and murine DCs (Fig. 1B, 1C). Among other TLR ligands, R848 (TLR7/8) but not polyinosinic:polycytidylic acid [poly(I:C)] (TLR3) induced MK2 phosphorylation (Fig. 1C). Because TLR3-mediated signal transduction is transmitted by means of Toll/IL-1 receptor domain–containing adaptor molecules independent of MyD88, we found MK2 to be predominantly activated in DCs via the MyD88-dependent pathway of TLR signaling.

Immediate MK2 phosphorylation after 30 min of LPS and R848 stimulation suggests its essential role in early TNF-α translation and therefore supports the previously described contribution of MK2 to inflammation (28). Nevertheless, upregulation of MK2 protein expression during the late phase of DC maturation indicated additional functions of MK2, which we further investigated.

Cytokines released from DCs play an essential role in the differentiation of CD4⁺ T cells and inflammation. Therefore, we initially explored MK2-regulated cytokines in murine BMDCs and their role in MAPK signaling (Fig. 2). Although cell viability and surface marker expression on LPS-activated DCs (LPS-DCs) were only slightly affected (data not shown), cytokine secretion was significantly different in MK2-deficient as compared with control DCs. As previously described for MK2^−/− splenocytes or macrophages, also MK2^−/− secreted significantly lower levels of TNF-α as compared with WT LPS-DCs (Fig. 2A) (12). In contrast, MK2^−/− LPS-DCs increased IL-1α secretion up to 3-fold within 6–24 h after LPS activation. As signaling via p38 also drives the expression of Th17 supporting cytokines (10), we analyzed IL-6 and IL-1β, which, along with IL-1α, were secreted in significantly higher amounts by MK2^−/− compared with WT LPS-DCs (Fig. 2A). Concurrently, IL-23 secretion remained unaffected. By blocking MK2 activity using the MK2-I3 inhibitor, we again observed decreased TNF-α and increased IL-1α and IL-1β secretion from LPS- and R848-activated BMDCs, whereas IL-6 was only enhanced upon LPS stimulation (Fig. 2B). However, in response to poly(I:C) stimulation, inhibition of MK2 only slightly affected cytokine secretion. IL-1β production was actually reduced, indicating a different function of MK2 in TLR3-mediated signaling, which is in line with our observations of MK2 induction in DCs mainly through MyD88-dependent TLR signaling. The specificity of MK2 blockade with MK2-I3 was shown by the inhibition of Hsp27 phosphorylation in human monocyte-derived DCs, which constitutes a direct downstream target of MK2 (Supplemental Fig. 1A).

The observation of strongly increased p38 phosphorylation in MK2 inhibitor-treated LPS-activated BMDCs (Fig. 2C, 2D) along with enhanced levels of secreted IL-1α (Fig. 2B) prompted us to also investigate the effect of MK2 on IL-1α–induced p38 activation. Because TLRs and the IL-1R belong to the same receptor superfamily and share common downstream signaling adaptor molecules, including MyD88 (29), we suspected a similar mechanism of MK2 intervention. In fact, we observed increased p38 signaling in murine DCs in response to activation via IL-1α upon blockade of MK2 activity (Fig. 2E), as was the case for activation via LPS/TLR4. To confirm these findings, we blocked IL-1α binding to its receptor by means of an IL-1α–specific Ab. Upon IL-1α blockade, MK2-inhibited DCs exhibited impaired secondary IL-1α/p38 signaling up to 6 h after LPS activation (Fig. 2C). Based on these observations it became evident that p38 is negatively regulated by MK2 upon ligand binding to both TLR4/7/8 and IL-1R, indicating a potential negative feedback regulation of p38 by MK2.

MAPK signaling is thought to regulate Th1 immunity (6–10), which prompted us to investigate the influence of MK2 on Th1 characteristics. In the first 6 h of LPS-driven DC maturation, MK2 promoted the secretion of the Th1-polarizing cytokine IL-12 in murine BMDCs (Fig. 3A). However, 48 h after LPS activation, we detected significantly enhanced IL-12 levels in supernatants of DCs treated with the MK2 inhibitor MK2-I3 (Fig. 3B), which indicated a prolonged LPS-induced IL-12 response in the absence of MK2 activity. In line with these observations, we found a strong decline in IL-10 secretion from murine MK2^−/− compared with WT DCs in response to LPS (Fig. 3F). IL-10 was also decreased when we silenced MK2 in human LPS-DCs (Supplemental Fig. 2). Furthermore, also at 48 h after LPS or R848 activation, IL-10 was reduced in MK2-I3–treated DCs (Fig. 3G). As observed for IL-1α and IL-12, poly(I:C) activation did not affect IL-10 expression (Fig. 3G). This observation indicates MK2 being involved in switching off IL-12 secretion while simultaneously driving IL-10 production in a time-dependent manner to prevent an excessive inflammatory Th1-polarizing immune response.

We also investigated the role of secondary IL-1α signaling in IL-12 and IL-10 production and the effect of MK2 in this context. Although no pronounced effect could be observed in WT LPS-DCs, IL-12 secretion from MK2-deficient BMDCs was diminished upon IL-1α blockade, whereas IL-10 levels were restored (Fig. 3C, 3H). These findings further underline the role of IL-1α in MK2-mediated immune regulatory effects in DCs.

Having determined the modulating effect of MK2 on pro- and anti-inflammatory cytokines as well as p38 induction, we were further interested in unraveling the upstream impact of MK2 on other signaling cascades, including ERK1/2 and the IL-10/STAT3 signaling pathway, implicated in mediating important anti-inflammatory characteristics. Inhibition of MK2 led to a strong decline in LPS-induced ERK1/2 signaling after 6 h, a signal transduction mechanism that has been described to down-modulate Th1 immunity (7) (Fig. 3D, 3E). Based on these observations, we consider MK2 to act as a molecular switch in MAPK signaling by cross-regulating p38 and ERK1/2 activity and, as a consequence, IL-12–mediated Th1 cell differentiation. Additionally, STAT3 activity was down-modulated upon MK2 deficiency, which, as for IL-10, could be restored upon blocking of IL-1α binding to DCs (Fig. 3I, 3J). Notably, the blockade of IL-1α also reduced STAT3 signaling in the presence of MK2 (Fig. 3I, 3J), pointing to a dual function of autocrine IL-1α. Therefore, depending on MK2 activity, IL-1α either promotes or blocks IL-10/STAT3–mediated anti-inflammatory mechanisms.

Owing to the particular expression kinetics of MK2 in both human and murine DCs showing peak expression 24 h after LPS encounter (Fig. 1A, 1B), we investigated the T cell priming potential of murine BMDCs in the presence or absence of MK2 activity after different maturation times. DCs in response to TLR4 and TLR7/8 stimulation were analyzed after 6 and 24 h for their potential to differentiate CD4⁺ cells toward Th1 cells (Fig. 4A–C, Supplemental Fig. 3). OT-II splenocytes were strongly, almost 10-fold, enriched for OVA-specific IFN-γ–secreting T cells in in vitro cultures with MK2-I3–treated LPS-DCs independent of the DC maturation time (Fig. 4A). Using only the control peptide, thus without boosting the cultures, we also detected IFN-γ–producing T cells, but to a smaller extent and again enriched in cultures with MK2-deficient DCs. Furthermore, we found enhanced IL-12–mediated Th1 immunity in in vitro cocultures containing MK2 inhibitor-treated LPS-DCs as shown by upregulated IL-12Rβ2 expression on CD4⁺ OT-II cells and accumulation of IL-12 in the supernatants (Fig. 4B, 4C). This observation was in accordance with previous data demonstrating more sustained IL-12 secretion by LPS-DCs in the absence of MK2 activity (Fig. 3A, 3B), pointing out how delayed Th1-attenuating effects of MK2 in DCs become clear in the setting of T cell priming following DC maturation.

Because increased IL-12 secretion from LPS-DCs was IL-1α–dependent, we next analyzed the role of IL-1α in Th1 differentiation. In cocultures of OT-II cells with MK2^−/− LPS-DCs, IFN-γ–producing CD4⁺ T cells were detected to a higher level compared with WT LPS-DCs (Fig. 4D), which was abrogated upon IL-1α blockade, proving again involvement of IL-1α in MK2-mediated Th1 control.

Based on the Th1-modifying properties of MK2, we further investigated the effect of MK2 signaling on CD8⁺ cytotoxic T cells. DCs deficient for MK2 showed a strong stimulatory potential for CD8⁺ T cells (Fig. 4E, 4F). Murine CD8⁺ OT-I cells proliferated up to 100-fold stronger in priming cultures with MK2^−/− LPS-DCs compared with WT LPS-DCs (Fig. 4E). Proliferation of CD8⁺ T cells was also increased in in vitro cocultures of human allogeneic T cells with MK2-silenced in comparison with control-silenced human monocyte-derived LPS-DCs (Fig. 4F, Supplemental Fig. 1C).

Next, we investigated the role of MK2 expressed in BMDCs in an in vivo setting of T cell activation. We therefore immunized OT-II mice with OVA-pulsed BMDCs isolated from WT or MK2^−/− mice and activated with LPS for 4 h prior to injection. In mice receiving MK2^−/− LPS-DCs, we observed a 2-fold increase in the mean percentage of T-bet⁺CD4⁺ OT-II cells in the LN (Fig. 5A, 5B). Additionally, we observed CD4⁺ T cell enrichment correlating with a strong increase in IL-2–expressing cells in the CD4⁺ subsets, from 42 to 57% IL-2⁺IL-10⁻ cells, when MK2 was absent from LPS-DCs. Furthermore, IL-2 secreted from cultured LN cells of MK2^−/− LPS-DC–injected mice was also strongly enriched from 5 up to 15 ng/ml (Supplemental Fig. 4).

The strong MK2-dependent induction of CD8⁺ T cells in vitro prompted us to further test the potential of MK2-deficient LPS-DCs to reinforce T cell–mediated cytotoxicity in vivo. We immunized WT mice with OVA peptide (SIINFEKL)–loaded MK2^−/− or WT BMDCs activated for 4 h with LPS prior to injection (Fig. 5C). In vivo T cell priming by WT LPS-DCs led to a 14% reduction of SIINFEKL-pulsed target cells injected 4 d later, within a time frame of 6 h (Fig. 5D). Immunization with MK2^−/− LPS-DCs had a yet stronger effect, resulting in a 22% reduction of target cells. Control mice injected with PBS showed a decline in target cells of 8%.

To characterize MK2 in a physiological context, we investigated whether MK2 expressed in tissue-derived endogenous DCs also regulates effector T cell differentiation. We did indeed observe a 2-fold enrichment of OVA-specific IFN-γ–producing cells in cocultures of OT-II cells with MK2-I3–treated compared with untreated CD11c⁺ LPS-DCs isolated from the spleen of C57BL/6 mice (Fig. 6A). IFN-γ–producing cells were detected at lower numbers in control peptide–stimulated DC/OT-II cocultures compared with specific OVA peptide boosting. Nevertheless, IFN-γ–secreting cells were strongly increased as well in the presence of MK2-I3 without OVA restimulation, demonstrating the strong Th1-skewing capacity of DCs lacking MK2 activity during the priming phase of CD4⁺ T cells (upon first encounter of an Ag). Cell viability and surface expression of MHC and costimulatory molecules were similar in MK2-I3–treated and untreated splenic LPS-DCs (data not shown).

Based on the enhanced Th1 priming potential of MK2-deficient splenic DCs, we investigated the function of MK2 in tissue-resident DCs in vivo by directly immunizing WT mice intradermally with LPS and OVA peptide for activation of skin-resident DCs after transfusing CD4⁺ OT-II cells to enhance the induced OVA-specific immune response (Fig. 6B, 6C). Injection of LPS and OVA along with MK2-I3 into WT mice significantly enriched IFN-γ–secreting cells in draining LNs as shown by ELISPOT following OVA peptide and anti-CD3 restimulation (Fig. 6C). Interestingly, LNs showed no response to control peptide restimulation (data not shown).

To investigate the immune regulatory role of MK2 in DCs in an endogenous in vivo context, we monitored the development of MOG_35–55-directed autoimmune encephalomyelitis in CD11cCre-Mk2^fl/fl conditional knockout mice as compared with WT littermate controls. The question of MK2 function in DCs in the course of autoimmune diseases is especially interesting, because MK2 has been proposed as a promising target for therapeutic intervention because it constitutes an important downstream mediator of p38 signaling (28).

Whereas the first round of immunization resulted in no significant differences in the onset of clinical symptoms (Fig. 7A, Supplemental Fig. 1B), strongly enhanced levels of IFN-γ–producing cells were detected in conditional MK2 knockout mice on day 21 upon MOG_35–55-specific restimulation (Fig. 7B). Subsequent boost immunization launched a massive autoimmune reaction as observed by neurologic symptoms, weight loss, as well as histopathological analysis (Fig. 7A, 7D), which proved to be largely Th1 driven, as symptoms were almost completely abrogated upon administration of an IFN-γ–blocking Ab (Fig. 7C). Most interestingly, this effect was not observed in WT littermate controls, which underlined the down-modulation of Th1 versus Th17 cells in the presence of MK2 in DCs. In conditional knockout mice, delayed onset of neurologic symptoms showed cytokine response to precede clinical manifestation of the progressing autoimmune reaction. Symptoms observed in this model were characterized by ascending caudocranial paralysis, which has been described to be associated with the classical mostly Th1-driven autoimmune encephalomyelitis phenotype in mice (30, 31). We thereby conclude MK2 expression in DCs to severely attenuate Th1-mediated immune reactions also in a physiologically relevant in vivo situation.

In summary, to our knowledge we provide first evidence that MK2 controls physiological Th1 effector mechanisms by modulating MAPK and JAK/STAT signaling in DCs depending on IL-1α secretion in response to TLR stimulation (Fig. 7). MK2-mediated control of the IL-10/STAT3 axis further promotes an anti-inflammatory DC phenotype that down-modulates IL-12–mediated Th1 immunity. Additionally, we found lack of MK2 in DCs to strongly drive Th1-mediated autoimmune encephalomyelitis, further confirming its role in a stringent control of Th1-polarized immune responses.

Previous studies have shown that down-modulation of MK2 results in reduced protein biosynthesis of several inflammatory cytokines in macrophages, suggesting that MK2 mediates proinflammatory properties (12, 32). Most importantly, MK2 has been shown to be directly involved in the stabilization of TNF-α mRNA via phosphorylation of the mRNA-AU–rich element binding protein tristetraprolin (33). Based on the observations presented in this study, we propose an unexpected regulatory function of MK2 in human and mouse DCs, stabilizing an immune regulatory phenotype in this specific cell type.

Under MK2-deficient conditions, we observed a shift toward a Th1-promoting phenotype in TLR-activated DCs both differentiated from murine bone marrow and human monocytes, or isolated directly from murine spleens. Upon MK2 knockout or inhibition, DCs clearly skewed T cell responses toward type 1 immunity in vitro and in vivo. These findings were underlined by our observations in a murine model of multiple sclerosis (EAE). In mice lacking MK2 specifically in CD11c⁺ DCs, clinical symptoms correlated strongly with the established classical Th1-mediated autoimmune phenotype (31) and could be abrogated upon administration of an IFN-γ–blocking Ab. Hence, down-regulating MK2 expression in DCs could be shown to contribute to onset and development of de-regulated autoimmune reactions in the mouse model.

Systemic MK2 knockout mice were reportedly found to show a higher susceptibility to Listeria monocytogenes infection and EAE, which suggested MK2 as a key molecule of the p38 signaling axis involved in bacterial inflammation, host defense, and autoimmunity (34, 35). These observations were primarily linked to a decreased capacity of macrophages to produce inflammatory cytokines, most prominently TNF-α (34). Increased pathogenic invasion in systemic MK2 knockout mice leads to a broad activation of macrophages as well as DCs including a variety of signals activating different TLRs as well as other pattern recognition receptors (36, 37). Such a plethora of activating signals results in complex interaction of multiple intracellular signaling pathways and affects DCs to drive T cell differentiation toward distinct Th cell phenotypes (38). For example, MK2-mediated signals in response to LPS/TLR4 were shown to be influenced by other signaling pathways induced by DC-specific ICAM3-grabbing nonintegrin expressed on DCs, which favors Th2 over Th1 differentiation (39). Furthermore, and of utmost importance for our observations, massive immune activation via DCs induces also subsequent immunosuppressive responses following secondary cytokine signaling, such as immune regulatory activities of bystander DCs (40), IL-10 secretion, and interaction with other cell subsets such as epithelial cells (41). By these means, DCs help to limit inflammatory processes and turn off excessive immune responses, thereby preventing damage to the host system. In line with our observations, locally distinct proinflammatory responses were induced in disease models propagating cutaneous or lung inflammation in systemic MK2-deficient mice in a tissue-dependent manner (13, 42). These heterogeneous responses may result from variable expression patterns in different cell types and cellular context dependency of MK2. Defects in immune regulation in tissue-specific DCs may therefore explain local accumulation of inflammatory factors in systemic MK2 knockout mice. Given the complexity of such immunological interactions, it becomes clear that DCs require delicate intrinsic feedback mechanisms, such as MK2 cross-regulation of MAPK signaling, to precisely balance pathogen-driven immune responses.

In the present study we have shown that MK2 is important for p38-mediated IL-10 secretion from LPS-DCs, as it has recently been described for macrophages, where IL-10 expression essentially requires the presence of MK2 (14). Our inhibitor and gene depletion studies blocking p38 or MK2 revealed that the p38/MK2 signaling route in response to LPS and autocrine IL-1α predominantly induces IL-10, corroborating the association between p38 and increased IL-10 secretion and the resulting induction of immune suppression (43, 44). Furthermore, we observed that MK2 deficiency reduces STAT3 activity depending on the IL-1α/p38/MK2 axis. This is in line with reports on autocrine feedback signaling of IL-10 via the JAK/STAT3 signaling cascade that stabilizes a regulatory DC phenotype by down-modulating IL-12 secretion and consequently attenuating Th1 responses (23, 45). IL-10 has also been reported to be involved in various other aspects of DC–T cell interaction, such as inhibiting costimulatory molecule expression (46), down-modulating MHC class II expression (47, 48), or overall hindering DC maturation (49, 50). Blockade of MK2 activity does not affect expression of CD80/86 (data not shown) but triggers IL-12 and IL-1α production, and hence, as shown by others and us, Th1-mediated responses (6, 16, 51, 52). Nevertheless, in our study enhanced IL-12 secretion became evident only in the presence of T cells and was corroborated by increased levels of IL-12Rβ2 and IFN-γ in T cells, pointing to the involvement of additional T cell–associated signals. To support this notion, others and we observed that CD40L/CD40 signal transduction requires the p38/MK2 pathway, as shown by increased phosphorylation of MK2 in DCs upon CD40L stimulation (53). With regard to IL-10 induction, it nevertheless remains unclear whether MK2 expression in DCs drives regulatory T cell generation. Finally, the strong induction of IL-1α, which we observed under MK2-deficient conditions, hints at IL-1α functionally replacing TNF-α, whose translation has been shown to be dependent on functional MK2 (33). This notion arises from the various overlapping and synergistic effects of IL-1α and TNF-α that have been described (54).

Even though we also detected enhanced levels of IL-6, which along with IL-1 forms an important mediator of Th17 polarization, the implications of MK2 in Th17 responses remain to be resolved. Along with reduced ERK1/2 phosphorylation we found increased IL-1β secretion in the absence of MK2 activity. In line with these observations, Brereton et al. (8) reported on the importance of ERK1/2 in LPS-DCs for triggering IL-1β–mediated Th17 responses. In our study we demonstrate that lack of MK2 in DCs plays only a minor role for Th17 functionality in an in vivo context of autoimmune disease (Fig. 7). Additionally, multiple studies have shown that increased activity of the ERK1/2 signaling pathway in DCs attenuates IL-12 and Th1 immunity while strongly inducing IL-10 secretion and therefore serving as a molecular switch in immune regulatory processes (7, 52, 55, 56). Interestingly, we observed both reduced ERK1/2 and enhanced p38 activation in DCs upon MK2 inhibition. This suggests MK2 acting in a cross-regulatory manner between these two MAPK signaling routes, which are known to mediate various opposite effects, especially regarding cytokine induction (57).

Taken together, MK2 may be considered a central molecular switch with regard to DC plasticity by fine-tuning an initiated immune response (Fig. 8). Three dominant molecular mechanisms explain the enhanced stimulatory potential of DCs upon MK2 inhibition: 1) increased LPS/p38 and IL-1α/p38 signaling, 2) impaired autocrine IL-10/STAT3 signaling, and 3) disrupted ERK1/2 signaling, which overall boost IL-12 secreted from LPS-activated DCs. Hence, MK2 appears to play an essential homeostatic role in limiting the extent and duration of a stimulatory immune reaction. In addition to mouse models, we also provide evidence for an analogous function of MK2 in human DCs. Therefore MK2-driven anti-inflammatory functions should be considered in autoimmune, tumor, or transplant models with regard to clinical intervention. Especially in the context of inflammatory diseases such as rheumatoid arthritis, for the treatment of which MK2 inhibitors are being tested, monitoring of potential immune dysregulation will be of the utmost importance (28, 58).

We thank S. Strehl, S. Piechot, R. Panzer, and A. Heitger for critically reviewing the manuscript, as well as D. Printz and E. Casanova for conceptual support.

A.M.D. and T.F. declare competing financial interests with regard to the patent: “Kinase inhibitor treated cellular immune therapeutics.” The other authors have no financial conflicts of interest.