TAM Receptor–Dependent Regulation of SOCS3 and MAPKs Contributes to Proinflammatory Cytokine Downregulation following Chronic NOD2 Stimulation of Human Macrophages

This article is featured in In This Issue, p.1387

Human nucleotide-binding oligomerization domain 2 (NOD2), an intracellular sensor of bacteria-derived peptidoglycan, confers the greatest genetic risk of developing Crohn’s disease, a disease of chronic intestinal inflammation (1). When peripheral monocytes enter mucosal sites, such as the intestinal lamina propria, they are continuously exposed to bacterial products (2, 3), including the NOD2 ligand muramyl dipeptide (MDP), which is the minimal peptidoglycan component specifically activating NOD2 (1). Acute NOD2 stimulation activates NF-κB and MAPK pathways, resulting in cytokine secretion (4–8). However, ongoing NOD2 stimulation significantly downregulates cytokine secretion upon restimulation through pattern recognition receptors (PRR) (3, 5, 6, 9, 10). This downregulation is impaired in individuals with Crohn’s disease–associated NOD2 polymorphisms (5, 9). Cytokine secretion in intestinal macrophages is similarly attenuated upon PRR stimulation (11), which is important for intestinal immune homeostasis. Moreover, chronic MDP treatment of mice in vivo attenuates subsequent experimental colitis (6).

Mechanisms contributing to cytokine downregulation after chronic NOD2 stimulation in human myeloid-derived cells include the upregulation of the intracellular inhibitors IL-1R–associated kinase (IRAK)-M (5) and IFN regulatory factor 4 (6, 12), the NF-κB1–dependent upregulation of the transcriptional repressor ATF3 (3), and the secretion of the inhibitory mediators IL-10 and TGF-β (9). Each of these mechanisms contributes only partially to cytokine downregulation and is operational to varying degrees in different individuals [e.g., IRAK-M (5, 13)]. Given the dramatic alterations in macrophage functions and importance of downregulating PRR-initiated pathways upon chronic microbial stimulation, we hypothesized that additional critical mechanisms mediating these changes have yet to be identified.

The Tyro3, Axl and Mer (TAM) receptors inhibit TLR3, TLR4, and TLR9 signaling outcomes (14). As such, Mer^−/− mice demonstrate increased TNF-dependent death upon acute LPS injection (15). TAM-deficient mice develop a lymphoproliferative disorder and various autoimmune disorders, including arthritis, pemphigus vulgaris, and lupus (16). Moreover, administration of the TAM ligands protein S and growth arrest–specific 6 (Gas6) can decrease inflammation in immune-mediated diseases in mice (17). In vitro, TAM^−/− mouse macrophages (18) and dendritic cells (DC) (19) demonstrate increased TLR-induced proinflammatory cytokines. This is associated with suppressor of cytokine signaling (SOCS)1 and SOCS3 mRNA upregulation (19). Whether the upregulated SOCS1 and SOCS3 account for the inhibitory effects of TAM has not been reported. Moreover, whether TAM plays a role in the dramatic downregulation of cytokines observed after chronic PRR stimulation, including of NOD2, is not known. Furthermore, inflammatory responses and pathways in human cells can be markedly different from those in mouse cells (20); TAM regulation in human myeloid cells has been relatively unexplored. The goal of this study was to determine the role and mechanisms wherein TAM regulates signaling and cytokines during acute and chronic NOD2 stimulation in human myeloid-derived cells.

We found that TAM was required for the downregulation of proinflammatory cytokines observed after chronic, but not acute, NOD2 and TLR stimulation in human myeloid-derived cells. Consistently, TAM was expressed at low levels on human monocyte-derived macrophages (MDM), and were each significantly upregulated after NOD2 stimulation; this upregulation required autocrine IL-10 and TGF-β. SOCS3, in turn, was upregulated in a TAM-dependent manner, and complementation of the reduced SOCS3 expression under TAM-deficient conditions was sufficient to restore the TAM-mediated proinflammatory cytokine downregulation under chronic NOD2 stimulation conditions. Interestingly, although TAM knockdown partially reversed the downregulation of proinflammatory cytokines observed after chronic NOD2 stimulation, the downregulation of anti-inflammatory cytokines was not reversed. At least one mechanism for this differential cytokine regulation was an increased dependency on MAPK signaling for promoter binding of activating transcription factors to and secretion of anti-inflammatory cytokines relative to proinflammatory cytokines. Finally, Axl and Mer were also required for cytokine downregulation after chronic NOD2 stimulation in vivo and under homeostatic conditions in intestinal tissues in mice.

Axl/Mer^−/− mice on a C57BL/6 background (generously provided by Dr. Carla Rothlin) have been described previously (21). Age-matched wild-type (WT) and knockout mice at age 8–14 wk were cohoused for a minimum of 4 wk. Mice were maintained on autoclaved food in a specific pathogen-free facility. Experiments were performed in agreement with the Yale University Institutional Animal Care and Use Committee and National Institutes of Health guidelines.

Informed consent was obtained as approved by the Yale University Institutional Review Board. Monocytes were purified from healthy individuals and differentiated to MDM as in Hedl et al. (5).

Total RNA was isolated, reverse transcribed, and quantitative PCR was performed as described previously (4) with normalization to GAPDH. Primers are available upon request.

For tolerance induction in vitro, human MDM (0.5 × 10⁶) were pretreated with 100 μg/ml MDP (Bachem, King of Prussia, PA) for 48 h prior to extensive wash and then retreated for 24 h with MDP for assessment of cytokine secretion. In some cases cells were also transfected with small interfering RNA (siRNA) or vectors of interest 24 h into the pretreatment period (see below). In other cases neutralizing anti–IL-10 or anti–TGF-β (R&D Systems, Minneapolis, MN) was used. TNF, IL-6, IL-8, IL-10 (BD Biosciences, San Jose, CA), IL-12 (eBioscience), or IL-1Ra (R&D Systems) secretion was assessed by ELISA. Mice were treated with i.p. MDP at the indicated times and doses. Mouse serum was examined for MIP2 (R&D Systems), TNF, IL-6, IL-12p40, and IL-10 (eBioscience).

Western blot analysis was performed as in Hedl et al. (5) using anti-SOCS3 (Santa Cruz Biotechnology, Santa Cruz, CA) or anti-GAPDH (EMD Millipore, Billerica, MA) Abs.

Phosphoprotein induction was determined by flow cytometry using Alexa Fluor 647–, PE-, or Alexa Fluor 488–labeled Abs against phospho-ERK, phospho-p38, or phospho-JNK (Cell Signaling Technology, Danvers, MA). TAM surface expression was determined with PE-labeled anti-Tyro3, Alexa Fluor 488–labeled anti-Axl, and allophycocyanin-labeled anti-Mer Abs (R&D Systems).

Primary human MDM were transfected with 200 nM scrambled or ON-TARGETplus SMARTpool siRNA against Tyro3, Axl, Mer, Gas6, protein S, SOCS3, c-Fos, c-Jun, musculoaponeurotic fibrosarcoma oncogene homolog K (MAFK), and PU.1 (Dharmacon, Lafayette, CO) (four pooled siRNAs for each gene), or pReceiver-M2-SOCS3 (GenoCopoeia, Rockville, MD), pMCL-MKK1 (R4F) (constitutively active ERK kinase) (22), pSRα-3HA-JNKK2-JNK1-WT (constitutively active JNK) (23) (generously provided by Dr. Ben Turk), pCDNA3-Flag MKK6(glu) (constitutively active p38 kinase) (24) (Addgene plasmid 13518), or empty vector using Amaxa Nucleofector technology (Amaxa, San Diego, CA).

Intestinal lamina propria cells were isolated from uninvolved colonic resection specimens from six noninflammatory bowel disease patients undergoing surgery for diverticular disease or colon cancer as per Hedl et al. (25).

Chromatin immunoprecipitation (ChIP) analysis was performed according to Zheng and Abraham (3) with 5 × 10⁶ MDM. Primers were designed to amplify genomic sequences at the cytokine gene promoter region (available upon request). Abs used include c-Fos (sc253x), c-Jun (sc44x) (Santa Cruz Biotechnology), MAFK (ab50322; Abcam, Cambridge, U.K.), and PU.1 (PA5-17505; Thermo Scientific, Rockford, IL).

Significance was assessed using a two-tailed Student t test. A Bonferroni correction was applied for multiple comparisons. A p value <0.05 was considered significant.

To elucidate the role of TAM in cytokine downregulation after both acute and chronic NOD2 stimulation in human MDM, we used siRNA to reduce the expression of each TAM, alone and in combination. We found selective, specific downregulation of the targeted TAM (Supplemental Fig 1A). Knockdown of each TAM alone or in combination did not alter cytokine secretion upon acute NOD2 stimulation (Fig. 1A, 1B). We previously found that pretreating MDM with 100 μg/ml MDP for 48 h is optimal for downregulating cytokine secretion upon restimulation through NOD2 or other PRR (3, 5, 9); 100 μg/ml MDP approximates muramic acid stool levels (26). Upon knockdown of each TAM after initiation of chronic NOD2 stimulation, and then restimulating through NOD2, there was significant, partial reversal in the downregulation of the proinflammatory cytokines TNF, IL-6, IL-8, and IL-12p40 (Fig. 1B). This partial reversal was further enhanced when TAM was knocked down in combination (Fig. 1B). Reduction in TAM expression, either alone or in combination, at the time chronic NOD2 stimulation was initiated resulted in similar partial reversal of proinflammatory cytokine downregulation (data not shown). In contrast to proinflammatory cytokine outcomes, the downregulation of the anti-inflammatory cytokine IL-10 was not reversed (Fig. 1B), highlighting distinct regulation of pro- and anti-inflammatory cytokines downstream of TAM during chronic NOD2 stimulation.

In distinction to TLR studies in mouse DC and macrophages (18, 19), TAM did not inhibit acute NOD2 signaling in primary human MDM (Fig. 1B). We therefore questioned whether this was due to differences between NOD2 and other TLRs, or between distinct myeloid-derived cell subsets. This may also be due to species differences. However, TAM did not inhibit cytokine secretion upon acute stimulation of various TLRs in human MDM (Supplemental Fig. 1B, 1C). Moreover, neither acute NOD2- nor TLR4-induced cytokine secretion was inhibited by TAM in human monocyte-derived DC (MDDC) (Supplemental Fig. 1D). In contrast, and similar to NOD2 (Fig. 1B, Supplemental Fig. 1D), TAM was required for optimal proinflammatory cytokine downregulation after chronic TLR4 stimulation in both human MDM (Supplemental Fig. 1B) and MDDC (Supplemental Fig. 1D). Taken together, whereas TAM does not regulate cytokines during acute PRR stimulation, it is required for optimal proinflammatory cytokine downregulation after chronic PRR stimulation in human MDM and MDDC.

The TAM ligands protein S and Gas6 can mediate the suppressive regulation of acute TLR stimulation (19). With knockdown of these respective ligands (Supplemental Fig. 1E), we found that protein S, but not Gas6, was necessary for optimal proinflammatory cytokine downregulation during chronic NOD2 stimulation (Fig. 1C). Similar to TAM knockdown results, upon protein S knockdown, the anti-inflammatory cytokine IL-10 was unaffected with chronic NOD2 stimulation (Fig. 1C), and both pro- and anti-inflammatory cytokines were unaffected with acute NOD2 stimulation (Fig. 1C). Therefore, protein S, but not Gas6, is required for the proinflammatory cytokine downregulation observed after chronic NOD2 stimulation in human MDM.

To determine whether Axl and Mer are required for chronic NOD2-induced cytokine downregulation in vivo, we used Axl/Mer^−/− mice. Previous studies have observed that a short period of priming with MDP allows for improved detection of serum cytokines (3, 27). Therefore, utilizing previously optimized MDP treatment conditions (3), mice were injected with i.p. MDP for acute (12 h priming) or chronic (48 h) NOD2 stimulation, and both conditions were followed by a 4-h period of MDP retreatment. Axl/Mer deficiency did not affect serum cytokines after acute MDP treatment (Fig. 2A). However, compared with WT mice, Axl/Mer^−/− mice were impaired in the downregulation of serum proinflammatory cytokines after chronic NOD2 stimulation (Fig. 2A). Similar to the in vitro human cell outcomes, the downregulation of the anti-inflammatory cytokine IL-10 was not impaired after chronic NOD2 stimulation in Axl/Mer^−/− mice (Fig. 2A). Therefore, consistent with our in vitro findings in primary human MDM, Axl and Mer are required for the downregulation in systemic cytokine levels after chronic MDP treatment in vivo in mice.

Intestinal immune cells are chronically exposed to bacteria and their products. Consistently, orally administered peptidoglycan (2) and MDP (3) can be visualized in the intestinal lamina propria. Furthermore, intestinal macrophages do not secrete cytokines in response to PRR stimulation (11), consistent with the downregulation of cytokines after chronic PRR stimulation. Given the importance of Axl and Mer in downregulating cytokines after chronic PRR stimulation, we asked whether intestinal tissues from Axl/Mer^−/− mice have dysregulated cytokine expression under baseline conditions; Axl/Mer^−/− mice have been previously shown to have increased susceptibility to experimental colitis (28). Both ileal and colonic tissues from Axl/Mer^−/− mice demonstrated increased proinflammatory and anti-inflammatory cytokines relative to WT mice (Fig. 2B). Therefore, Axl and Mer are essential for cytokine homeostasis in the intestinal environment where there is ongoing exposure to microbial products.

Given the selective role for TAM in downregulating cytokines upon chronic PRR stimulation, we questioned whether TAM receptors are upregulated with chronic PRR stimulation on human MDM. Surface expression of each TAM progressively increased during 24 h NOD2 stimulation and increased even further upon restimulation of 48 h MDP-pretreated MDM (Fig. 3A). We further found that expression of each TAM was increased on human intestinal compared with peripheral myeloid-derived cells (Fig. 3B), consistent with the chronic PRR stimulation conditions in the intestine. Therefore, consistent with the role for TAM in downregulating cytokines specifically after chronic PRR stimulation, TAM expression is increased after chronic NOD2 stimulation on MDM and on intestinal myeloid-derived cells compared with peripheral MDM.

We next sought to define the mechanisms upregulating TAM expression upon chronic NOD2 stimulation in human MDM. TGF-β has been reported to contribute to Axl upregulation during Langerhans cell differentiation (29). Moreover, IL-10 can contribute to Mer upregulation (30). We have previously shown that the early secretion of IL-10 and TGF-β upon chronic NOD2 stimulation is required for subsequent pro- and anti-inflammatory cytokine downregulation (9). Through blockade of NOD2-induced autocrine IL-10 and TGF-β, we found that each of these cytokines was required for upregulation in TAM expression upon chronic NOD2 stimulation in human MDM (Fig. 4).

To more closely examine the mechanisms through which TAM downregulates proinflammatory cytokines upon chronic NOD2 stimulation in human MDM, we considered the SOCS family of proteins. A prior study in mouse BMDC identified that TAM stimulation upregulates SOCS1 and SOCS3 expression (19). However, whether these proteins were mediating downstream TAM outcomes was not specifically examined. Moreover, whether these or other SOCS family proteins regulate TAM effects in human myeloid cells and downstream of NOD2, in particular under conditions of chronic NOD2 stimulation, is not known. We found that SOCS1, SOCS2, SOCS3, SOCS6, SOCS7, and CIS were each upregulated with acute NOD2 stimulation, and to varying degrees even further upregulated after chronic NOD2 stimulation (Fig. 5A). However, only SOCS1 and SOCS3 upregulation was TAM-dependent (Fig. 5A). We therefore questioned whether one or both of these SOCS proteins were required for TAM-mediated inhibition of cytokines after chronic NOD2 stimulation. Reduction in SOCS1 expression (Supplemental Fig. 2A) did not reverse chronic NOD2-induced cytokine downregulation in human MDM (Supplemental Fig. 2B). We confirmed SOCS1 knockdown efficacy by ensuring that LPS-induced cytokines were inhibited by SOCS1 (Supplemental Fig. 2C) as per previous reports (31).

We next examined SOCS3 and confirmed that SOCS3 protein expression was upregulated following chronic NOD2 treatment and that this protein upregulation was TAM-dependent (Fig. 5B). We then effectively reduced SOCS3 expression (Fig. 5C). Similar to TAM knockdown results (Fig. 1B), acute NOD2-induced cytokines were not altered, whereas the downregulation in proinflammatory cytokines observed after chronic NOD2 stimulation was impaired (Fig. 5D). Also similar to TAM, SOCS3 knockdown did not affect anti-inflammatory cytokine secretion upon chronic NOD2 stimulation (Fig. 5D). Given the TAM-dependent upregulation of SOCS3 and the similar regulation by TAM and SOCS3 of proinflammatory, but not anti-inflammatory, cytokines with chronic NOD2 stimulation, we questioned whether restoring SOCS3 expression under conditions of TAM deficiency could restore the downregulation of proinflammatory cytokines. We restored SOCS3 expression under chronic NOD2-stimulated, TAM-knockdown conditions to physiological levels (Fig. 5E). Despite the absence of TAM, the reconstituted SOCS3 could restore the downregulation of proinflammatory cytokines observed after chronic NOD2 stimulation (Fig. 5F). Taken together, upon chronic NOD2 stimulation, SOCS3 is upregulated in a TAM-dependent manner, which is required for the TAM-mediated inhibition of proinflammatory cytokines.

We observed a consistently distinct regulation of proinflammatory and anti-inflammatory cytokines with knockdown of TAM, protein S, and SOCS3 during chronic NOD2 stimulation. We therefore questioned what mechanisms might be contributing to this differential proinflammatory versus anti-inflammatory cytokine regulation. The role of TAM in regulating anti-inflammatory cytokines upon TLR stimulation in mice has not been reported (15, 19). We previously found that anti-inflammatory cytokines are highly dependent on the strength of MAPK signaling during acute PRR stimulation, and in a manner distinct to proinflammatory cytokines (32). We therefore investigated how MAPK activation was regulated during chronic NOD2 stimulation and how TAM contributes to this regulation. In contrast to the robust activation of ERK, p38, and JNK with acute NOD2 stimulation, these pathways were dramatically attenuated upon NOD2 restimulation of chronic NOD2-stimulated human MDM (Fig. 6A). MAPK activation was only partially restored with TAM knockdown under chronic NOD2 stimulation conditions, such that activation was only 13–20% the levels observed with acute NOD2 stimulation (Fig. 6A). We observed a similar partial restoration with SOCS3 knockdown (Fig. 6A), which we had identified to be a major mediator of TAM-dependent effects (Fig. 5F). Therefore, MAPK activation was still significantly impaired in the absence of TAM or SOCS3. Of note, however, is that the low, but significant level of MAPK activation in chronic NOD2-stimulated, TAM-deficient MDM was contributing to the restoration of proinflammatory cytokine secretion observed as assessed by combined inhibition of ERK, p38, and JNK (Supplemental Fig. 3A); the cells were viable under these conditions (Supplemental Fig. 3B). We thus questioned whether this level of MAPK activation that was sufficient for partial restoration of proinflammatory cytokines was insufficient for restoration of anti-inflammatory cytokines. If this were the case, full reconstitution of MAPK activation under chronic NOD2 stimulation of TAM-deficient or SOCS3-deficient MDM would be expected to restore anti-inflammatory cytokines. We used constitutively active ERK, p38, and JNK to reconstitute activation levels to those observed upon acute NOD2 stimulation (Fig. 6A). With reconstitution of MAPK activation, levels of IL-10 secretion were upregulated during chronic NOD2 stimulation in the absence of either TAM or SOCS3 expression (Fig. 6B). We previously observed that the anti-inflammatory mediator IL-1Ra followed regulatory patterns similar to those of IL-10 (9, 32). We found that IL-1Ra downregulation after chronic NOD2 stimulation was similarly TAM-independent, but MAPK-dependent (Fig. 6B). The levels of proinflammatory cytokines that had already been significantly restored increased even further with restoration of MAPK activation (Fig. 6B). Therefore, TAM and SOCS3 contribute to decreasing MAPK activation after chronic NOD2 stimulation, which, in turn, attenuates proinflammatory cytokines. In contrast, additional TAM-independent mechanisms contributing to the attenuation in MAPK activation after chronic NOD2 stimulation are essential for downregulating anti-inflammatory cytokine secretion.

Because TAM knockdown during chronic NOD2 stimulation did not alter anti-inflammatory cytokine levels, we hypothesized that there was insufficient binding of activating transcription factors to the IL-10 cytokine promoter in chronic NOD2-stimulated MDM, regardless of whether TAM was present or absent. We further hypothesized that binding of MAPK pathway–associated transcription factors, such as c-Fos and c-Jun, to the IL-10 promoter could be restored with increased strength of MAPK signaling, thereby leading to the restoration of IL-10 secretion. We examined the binding of c-Fos and c-Jun to cytokine promoters utilizing ChIP. Both transcription factors bound to pro- and anti-inflammatory cytokine promoters with acute NOD2 stimulation, and binding decreased after chronic NOD2 stimulation (Fig. 7), consistent with the attenuated cytokine secretion under these conditions (Figs. 1, 5, 6B). Also consistent with the cytokine secretion patterns during chronic NOD2 stimulation, upon TAM knockdown, c-Fos and c-Jun binding to proinflammatory cytokine promoters increased, whereas their binding to the IL-10 promoter did not (Fig. 7). These transcription factors were critical for the restoration of proinflammatory cytokine secretion upon TAM knockdown during chronic NOD2 stimulation as determined by knockdown of c-Fos and c-Jun (Fig. 8, Supplemental Fig. 3C). Finally, reconstitution of MAPK signaling in chronic NOD2-stimulated, TAM-deficient MDM (as per Fig. 6A) resulted in effective binding of c-Fos and c-Jun to the IL-10 promoter (Fig. 7), consistent with the IL-10 secretion observed under these conditions (Fig. 6B). Binding of c-Fos and c-Jun to proinflammatory cytokine promoters under these conditions was also enhanced (Fig. 7). To identify additional transcription factors binding to IL-10 regulatory elements that might be modulated downstream of MAPKs, we examined ENCODE and identified MAFK and PU.1 as candidates. MAFK can heterodimerize with other transcription factors, including c-Fos (33). However, a role for MAFK downstream of PRR signaling or in IL-10 regulation has not been reported. PU.1 can serve as an organizer for activation of enhancers in macrophages (34). We found that both MAFK and PU.1 were regulated in a manner similar to that observed with c-Fos and c-Jun (Fig. 7, Supplemental Fig. 3C); these transcription factors were similarly critical for the restored cytokine secretion in chronic NOD2-stimulated, TAM-deficient MDM (Fig. 8). Taken together, TAM-mediated attenuation of MAPK activation upon chronic NOD2 stimulation results in decreased binding of activating transcription factors to proinflammatory cytokine promoters, which, in turn, decreases proinflammatory cytokine secretion. In contrast, TAM-independent mechanisms regulate the attenuated MAPK-dependent binding of activating transcription factors to the promoter of the anti-inflammatory cytokine IL-10.

In this study, we found that TAM contributes to the downregulation of proinflammatory cytokines observed after chronic, but not acute, PRR stimulation in primary human MDM and MDDC. The surface expression of each TAM progressively increased with chronic NOD2 stimulation, and consistently TAM expression was upregulated on human intestinal lamina propria myeloid-derived cells relative to peripheral MDM. The importance of TAM in mediating intestinal homeostasis has recently been demonstrated through Axl/Mer^−/− mice, which show more severe DSS-induced colitis relative to WT mice (28). We found that Axl/Mer^−/− mice have dysregulated cytokines in intestinal tissues under baseline conditions, as well as defects in downregulating cytokines systemically with chronic NOD2 stimulation. Chronic NOD2 stimulation of human MDM led to IL-10– and TGF-β–dependent TAM expression upregulation, which, in turn, upregulated SOCS3 expression; SOCS3 was required for the TAM-mediated outcomes observed. TAM contributed selectively to proinflammatory cytokine downregulation during chronic NOD2 stimulation in human MDM, in part through attenuating MAPK activation and binding of downstream transcription factors to proinflammatory cytokine promoters. However, TAM-independent pathways also contributed to attenuation of MAPK activation, which were ultimately critical in regulating the binding of activating transcription factors to the IL-10 promoter and the secretion of IL-10. Such TAM-independent pathways likely include the IRAK-1 inhibitory molecule IRAK-M, which contributes to downregulation of IL-10 secretion with chronic NOD2 stimulation (3, 5). Therefore, we now elucidate a clear role for TAM-mediated mechanisms in downregulating proinflammatory cytokines upon chronic PRR stimulation of primary human MDM (Supplemental Fig. 4).

Regulatory mechanisms mediating cytokine suppression during acute PRR stimulation can overlap or be distinct from those contributing to cytokine suppression with chronic PRR stimulation. A role for TAM in the dramatic downregulation of cytokines observed with chronic PRR stimulation has not been previously reported. Although dependency on the different TAM members can be cell- and organ-specific for apoptotic cell clearance (35), we found that each of the TAM receptors contributed to cytokine inhibition with chronic NOD2 stimulation to approximately equal degrees, and they cooperated with each other for optimal regulation (Fig. 1). Mechanistically, we identified a role for TAM in selectively regulating SOCS1 and SOCS3 expression, as was previously described in mouse BMDC (19). Prior studies had not examined which of the SOCS proteins was necessary for TAM effects. We clearly established that upon chronic NOD2 stimulation in human MDM, only SOCS3 recapitulates the regulation observed by TAM. Importantly, SOCS3 complementation was able to rescue the impaired downregulation of cytokines observed after chronic NOD2 stimulation in TAM-deficient MDM, in part through partial restoration of MAPK signaling. SOCS3 can lead to MAPK inhibition through inhibiting upstream pathways (36). SOCS3 also contributes to cytokine inhibition through additional mechanisms, including regulation of cytokine receptor signaling (36); autocrine cytokine signaling is critical to overall PRR-initiated cytokine secretion.

We identified a selective role for TAM in inhibiting proinflammatory, but not anti-inflammatory, cytokines upon chronic NOD2 stimulation in human MDM. Interestingly, IL-10 secretion is also not restored with IFN-γ–mediated reversal of tolerance (37), demonstrating a clear differential regulation of IL-10 relative to proinflammatory cytokines during chronic PRR-induced tolerance. Similar to the lack of IL-10 regulation after chronic NOD2 stimulation with TAM knockdown in human MDM, IL-10 downregulation was unaffected after chronic systemic MDP treatment in Axl/Mer^−/− mice in vivo. In contrast, IL-10 was upregulated in ileal and colonic tissues in Axl/Mer^−/− mice. One possibility for these differences in IL-10 regulation in intestinal tissues relative to the peripheral regulation assessed in mice may be due to the consequences of prolonged proinflammatory cytokine dysregulation in intestinal tissues in the context of persistent microbial exposure, thereby leading to a compensatory elevation of anti-inflammatory cytokines. The strength and kinetics of MAPK signaling can profoundly influence cellular outcomes and cytokine secretion (38, 39). We found that at least one mechanism for the distinct pro- and anti-inflammatory cytokine regulation is due to differential sensitivity to the strength of MAPK signaling, which led to differential efficacy in the binding of transactivating transcription factors, including c-Jun, c-Fos, MAFK, and PU.1, to proinflammatory and anti-inflammatory cytokine promoters. The role of transactivating transcription factor binding to cytokine promoters during chronic NOD2 stimulation had not been previously defined. Moreover, the role of MAFK in contributing to cytokine secretion with either acute or chronic PRR stimulation had not been well defined; we identify a clear requirement for this transcription factor downstream of PRR signaling. Taken together, we identify a role for TAM-mediated mechanisms in downregulating proinflammatory cytokines during the chronic PRR stimulation conditions observed in environments such as the intestinal lamina propria.

We thank Drs. Carla Rothlin and Ben Turk for reagents.

The authors have no financial conflicts of interest.