TRIM38 Negatively Regulates TLR3/4-Mediated Innate Immune and Inflammatory Responses by Two Sequential and Distinct Mechanisms

This article is featured in In This Issue, p.4041

Toll-like receptors are evolutionarily conserved pattern recognition receptors that recognize a set of pathogen-associated molecular patterns and play critical roles in host defense against certain microbes. So far, 13 TLRs (TLR1–TLR13) have been identified in humans and mice (1). TLRs contain an extracellular domain, a transmembrane domain, and a conserved cytoplasmic Toll/IL-1R (TIR) domain. The TIR domain is responsible for mediating homotypic protein–protein interactions (2), and it acts as a platform to recruit downstream TIR domain–containing adaptor proteins and other signaling molecules upon ligand stimulation. These signaling cascades ultimately activate transcription factors, such as NF-κB and IFN regulatory factor (IRF3), which act alone or collaboratively to induce expression of proinflammatory cytokines, such as TNF-α and IL-1β, and/or type I IFNs (3). TNF-α and IL-1β are two important proinflammatory cytokines that trigger signaling pathways to further activate NF-κB, leading to expression of hundreds of chemokines and cytokines and induction of efficient inflammatory response (4). In addition, via JAK-STAT signaling, type I IFNs can further amplify the immune responses by induction of hundreds of cytokines or antimicrobial effector proteins (5).

Among the TLRs, TLR3 recognizes viral dsRNA, as well as its synthetic analog polyinosinic-polycytidylic acid [poly(I:C)], and it recruits the critical adaptor TIR domain–containing adapter-inducing IFN-β (TRIF; also called TICAM-1) for downstream signaling. TRIF, in turn, recruits the TNFR-associated factor (TRAF)2/6–inhibitor of NF-κB kinase complex to activate NF-κB and TANK-binding kinase-1 (TBK1) to activate IRF3, leading to the subsequent expression of proinflammatory cytokines, such as TNF-α and IL-1β, as well as type I IFNs (6). TRIF is indispensable for TLR3-mediated activation of NF-κB and IRF3, because Trif-deficient cells are defective in poly(I:C)-induced production of inflammatory cytokines and type I IFNs (7). Most other TLRs signal through the MyD88–TRAF6–inhibitor of NF-κB kinase complex to activate NF-κB. TLR4, which recognizes LPS of Gram-negative bacteria, is the only receptor that signals through TRIF-independent and MyD88-dependent pathways to activate NF-κB– and TRIF-dependent pathways to activate NF-κB and IRF3 (7). Trif-deficient cells are defective in LPS-induced activation of IRF3, and double knockout of Trif and Myd88 results in complete abolishment of LPS-induced activation of NF-κB (7). TLR4-mediated signaling is important for the host to elicit innate and adaptive immune responses to defend against the infection of Gram-negative bacteria. Tlr4^−/− mice are much more susceptible to Gram-negative bacterium infection than are wild-type (WT) mice (8).

Tripartite motif (TRIM) family proteins are E3 ubiquitin ligases featuring conserved RING, B-BOX, and cocoiled structures (9). Previous studies demonstrated critical regulatory roles of TRIM family members in innate immune and inflammatory responses (9, 10). Recently, we showed that TRIM38 negatively regulates TNF-α– and IL-1β–triggered NF-κB activation by mediating lysosomal-dependent degradation of TGF-β–activated kinase 1 (TAK1)-binding protein (TAB)2/3, two critical components of the TAK1 kinase complex (11). TRIM38 also was shown to inhibit TLR3/4-mediated activation of NF-κB and IRF3 by mediating ubiquitin-proteasomal degradation of Traf6 and NAK-associated protein 1, (Nap1), respectively, in a mouse Raw264.7 cell line (12, 13). Additionally, one group reported that TRIM38 inhibits TLR3-mediated signaling by mediating ubiquitin-proteasomal degradation of TRIF in the HEK293-TLR3 cell line (14). Because these studies were performed primarily in various cell lines, and TLRs are predominantly expressed and function in immune cells (15), the functions and mechanisms of endogenous TRIM38 in primary immune cells and in vivo are unknown.

In this study, we performed Trim38 gene-knockout studies and showed in vivo evidence for a critical negative-regulatory role for TRIM38 in TLR3/4-mediated innate immune and inflammatory responses. Mechanically, TRIM38 mediates K48-linked polyubiquitination at K228 and proteasomal degradation of TRIF. Furthermore, we found that TRIM38 was highly induced by type I IFNs and negatively regulated TNF-α/IL-1β signaling in IFN-β–primed, but not unprimed, mouse immune cells. These findings reveal that TRIM38 plays important negative-regulatory roles in TLR3/4-mediated innate immune and inflammatory responses by two sequential and distinct mechanisms.

Rabbit mAb against ubiquitin (Abcam); mouse mAbs against Flag (Sigma), hemagglutinin (OriGene), β-actin (Sigma), p-IκBα (CST), p-IRF3 (CST), and p-TBK1 (Abcam); recombinant mouse TNF-α, IL-1β, IFN-α4, IFN-β, and IFN-γ (R&D Systems), and poly(I:C) (InvivoGen), LPS (Sigma), R848 (InvivoGen), PGN (InvivoGen), and cycloheximide (Sigma) were purchased from the indicated companies. Mouse anti-Trim38 and rabbit anti-Trim38 antisera were raised against recombinant Trim38 protein. Mouse anti-TRIF, mouse anti-IκBα, mouse anti-TAB2, and rabbit anti-TAB3 were described previously (11, 16–18).

Mammalian expression plasmids for Flag- or hemagglutinin-tagged murine Trim38 and its mutant, Trif and its mutants, Traf6, and Nap1 were constructed by standard molecular biology techniques.

Trim38 gene-knockout mice with a CL7/B6 background were purchased from KOMP. In Trim38-knockout mice, the entire fragment containing all exons of the Trim38 gene is replaced with a fragment containing the selection marker neomycin. The strategy for construction of the targeting vector is illustrated in Supplemental Fig. 1A. Genotyping by PCR was performed using the following pairs of primers: WT(F): 5′-ACTGGAGCGTGACATTGAGAAAAGC-3′ and WT(R): 5′-TAGTGTGAACCTTCAGGGGATCAGC-3′; and KO(F): 5′-TACCTCTGCCATTCTTAGCACTTG-3′ and KO(R): 5′-GGCGGCTAGTCTGTTCAGC-3′. Amplification of the WT allele with primers WT(F) and WT(R) results in a 109-bp fragment, whereas amplification of the disrupted allele with primers KO(F) and KO(R) results in a 450-bp fragment. The following primers were used for RT-PCR analysis of murine Trim38 mRNA: 5′-AAGAGCAGGATCAAGACATGGT-3′ (forward) and 5′-AGGAGTTGACCGTCATCTTCA-3′ (reverse).

All animal experiments were performed in accordance with the Wuhan University Animal Care and Use Committee guidelines.

Bone marrow–derived macrophages (BMDMs) and bone marrow–derived dendritic cells (BMDCs) were generated as described (19). Bone marrow cells (1 × 10⁷) were cultured in RPMI 1640 medium containing 10% FBS and 10 ng/ml recombinant murine M-CSF (PeproTech) or GM-CSF–containing conditional medium in a 100-mm dish for 5 or 9 d for generation of BMDMs or BMDCs, respectively.

Primary lung fibroblasts were isolated from ∼4–6-wk-old mice. Lungs were minced and digested in calcium and magnesium–free HBSS containing 10 μg/ml type II collagenase (Worthington) and 20 μg/ml DNase I (Sigma) for 3 h at 37°C with shaking. Cell suspensions were filtered through progressively smaller cell strainers (100 and 40 μm), centrifuged at 1500 rpm for 4 min, and plated in culture medium (1:1 [v/v] DMEM/Ham’s F-12 containing 10% FBS, 15 mM HEPES, 2 mM l-glutamine, 50 U/ml penicillin, and 50 μg/ml streptomycin). After 1 h, adherent fibroblasts were rinsed with HBSS and cultured in media.

Age- and sex-matched Trim38^+/+ and Trim38^−/− mice were injected i.p. with poly(I:C) (2 μg/g body weight) plus d-galactosamine (0.5 mg/g body weight) or with LPS (10 μg/g body weight). The survival of mice was monitored every 2 h.

Age- and sex-matched Trim38^+/+ and Trim38^−/− mice were given S. typhimurium orally (5 × 10⁵ PFU/mouse). The body weight and survival of the infected mice were monitored every day.

Blood from mice injected with poly(I:C) plus d-galactosamine or LPS was collected at the indicated times, and the serum concentration of TNF-α (BioLegend), IL-6 (BioLegend), IFN-α (PBL), and IFN-β (PBL) was measured using ELISA kits from the indicated manufacturers. BMDMs, BMDCs, and mouse lung fibroblasts (MLFs) were stimulated with various TLR ligands for 18 h before the culture medium was collected for measurement of TNF-α, IL-6, and IFN-β by ELISA.

Transfection and reporter assays were performed as previously described (20–22). HEK293 cells were seeded on 24-well plates and transfected on the following day by standard calcium phosphate precipitation. Where necessary, empty control plasmid was added to ensure that each transfection received the same amount of total DNA. To normalize for transfection efficiency, pRL-TK (Renilla luciferase) reporter plasmid (0.01 μg) was added to each transfection. Luciferase assays were performed using a dual-specific luciferase assay kit (Promega, Madison, WI). Firefly luciferase activities were normalized based on Renilla luciferase activities.

Coimmunoprecipitation, immunoblotting analysis, and ubiquitination assays were performed as previously described (23–25). For ubiquitination assays, the immunoprecipitates were re-extracted in lysis buffer containing 1% SDS and denatured by heating for 5 min. The supernatants were diluted with regular lysis buffer until the concentration of SDS was decreased to 0.1%, followed by reimmunoprecipitation with the indicated Abs. The immunoprecipitates were analyzed by immunoblotting with the ubiquitin Ab.

Quantitative real-time PCR (qPCR) assays were performed as previously described (26–29). Total RNA from mouse or human cells was isolated using TRIzol reagent (Invitrogen). After reverse transcription with an oligo(dT) primer using a RevertAid First Strand cDNA Synthesis Kit (Fermentas), aliquots of products were subjected to qPCR analysis to measure mRNA levels of the tested genes. Gapdh was used as a reference gene. Gene-specific primer sequences were described previously (11, 30). The following primers were used for qPCR analysis of mouse Trim38 mRNA: 5′-TGCTGGCGCTGTTATTGGG-3′ (forward) and 5′-GTTCACGCAATTTTGTCATGGT-3′ (reverse).

Differences between averages were analyzed by the Student t test. A p value < 0.05 was considered significant.

To investigate the functions of TRIM38 in vivo, we purchased Trim38 gene-knockout mice with a CL7/B6 background (KOMP). In Trim38-knockout mice, the entire fragment containing all exons of the Trim38 gene is replaced with a fragment containing the selection marker neomycin (Supplemental Fig. 1A). The deletion of Trim38 in Trim38^−/− mice or primary cells was confirmed by genotyping and qPCR (Supplemental Fig. 1B, 1C).

To determine the functions of TRIM38 in innate immune and inflammatory responses, we first prepared BMDMs from Trim38^+/+ and Trim38^−/− mice. We found that Trim38 deficiency potentiated poly(I:C) (a ligand for TLR3)-induced and LPS (a ligand for TLR4)-induced, but not R848 (a ligand for TLR7)-induced or PGN (a ligand for TLR2)-induced, transcription of Ifnb1, Ifna4, Isg54, Il1b, Tnfa, and Il6 in qPCR experiments (Fig. 1A). Consistently, poly(I:C)- and LPS-induced, but not R848- or PGN-induced, secretion of IFN-β, IL-6, and TNF-α was significantly increased in Trim38^−/− BMDMs (Fig. 1B). Similarly, poly(I:C)- and LPS-induced, but not R848- or PGN-induced, downstream gene expression was significantly increased in Trim38^−/− BMDCs (Fig. 2A, 2B) or MLFs (Supplemental Fig. 1D, 1E) in comparison with their WT counterparts. Because expression of type I IFNs (e.g., Ifnb1 and Ifna4) and their downstream effector (e.g., Isg54) requires IRF3 activation, whereas transcription of the proinflammatory cytokines Tnfa, Il1b, and Il6 is dependent on NF-κB activation, our results suggest that TRIM38 negatively regulates both TLR3/4-mediated IRF3 and NF-κB pathways, as well as downstream gene induction in immune cells or MLFs. This conclusion was supported by our observations that phosphorylation of TBK1 and IRF3 (hallmarks of IRF3 activation) as well as phosphorylation of IκBα (hallmarks of NF-κB activation) were markedly increased in Trim38^−/− BMDCs in comparison with their WT counterparts (Fig. 2C).

To examine whether TRIM38 regulates TLR3/4-mediated innate immune and inflammatory responses in vivo, we examined poly(I:C)- and LPS-induced production of type I IFNs and proinflammatory cytokines in the sera of mice. Age- and sex-matched WT and Trim38^−/− mice were injected i.p. with poly(I:C) or LPS. As shown in Fig. 3A, poly(I:C)- and LPS-induced production of IFN-β, IFN-α, TNF-α, and IL-6 was significantly enhanced in the sera of Trim38^−/− mice compared with Trim38^+/+ mice. These results suggest that Trim38 deficiency potentiates TLR3/4-mediated cytokine induction in vivo.

To further confirm that TRIM38 regulates TLR3/4-mediated innate immune and inflammatory responses in vivo, we examined poly(I:C)- and LPS-induced inflammatory death and S. typhimurium–induced septic shock of WT and Trim38-deficient mice. Consistent with the results of cytokine production in the sera, Trim38^−/− mice experienced early death onset and a higher percentage of lethality within 30 h compared with their WT counterparts after injection of poly(I:C) plus d-galactosamine (Fig. 3B) or LPS (Fig. 3C). After oral administration of S. typhimurium, the body weights of Trim38^−/− mice decreased more dramatically than those of WT mice within 6 d, after which time Trim38^−/− mice began to die (Fig. 3D). Consistently, Trim38^−/− mice showed an early death onset and a higher percentage of lethality within 15 d in comparison with their WT counterparts after administration of S. typhimurium (Fig. 3D). These results suggest that TRIM38 negatively regulates TLR3/4-mediated inflammatory and innate immune responses and bacteria-induced septic shock in vivo.

Previously, several signaling components, including human TAB2/3, TRAF6, NAP1, and TRIF, were shown to be associated with human TRIM38 in various cell lines or mammalian overexpression systems (11–14). We revisited these observations with proteins of murine origin using coimmunoprecipitation experiments. The results indicated that murine Trim38 interacted with Trif, Traf6, Nap1, and Tab2, but not with Traf3, in a mammalian overexpression system (Fig. 4A). Endogenous coimmunoprecipitation experiments indicated that Trim38 was constitutively associated with Trif and Tab2, and these associations were enhanced following poly(I:C) or LPS stimulation in BMDCs (Fig. 4B). In the same experiments, we failed to detect an endogenous association between Trim38 and Traf6 or Nap1 in BMDCs (Fig. 4B). Furthermore, cotransfection experiments indicated that Trim38 mediated downregulation of Trif and Tab2, but not Traf3, Traf6, or Nap1, although human TRIM38 could destabilize human NAP1 protein (Fig. 4C, 4D). The same conclusions were drawn when the experiments were performed under the condition of protein synthesis inhibition by cycloheximide (Supplemental Fig. 2). These results suggest that Trim38 may negatively regulate TLR3/4-mediated signaling by targeting Trif, but not Traf6 or Nap1, for degradation in mouse cells.

To further investigate how Trim38 downregulates Trif, we determined the effects of NH₄Cl (inhibitor of lysosome-dependent degradation pathway) and MG132 (inhibitor of ubiquitin-proteasome–dependent degradation pathway) on Trim38-mediated downregulation of Trif in a mammalian overexpression system. The results indicated that MG132 inhibited Trim38-mediated downregulation of Trif, and NH₄Cl inhibited Trim38-mediated downregulation of Tab2 (Fig. 4E). In addition, Trim38(C31S), an E3 ligase inactive mutant of murine Trim38, was unable to downregulate Trif but retained its ability to downregulate Tab2 (Fig. 4F). These results suggest that Trim38 interacts with and mediates degradation of Trif by the polyubiquitination-proteasome pathways.

We also determined the effects of endogenous Trim38 on Trif levels in mouse immune cells. We found that the level of Trif was markedly upregulated in Trim38^−/− BMDCs in comparison with their WT counterparts, both before and after poly(I:C) or LPS stimulation (Fig. 4G, Supplemental Fig. 2G). The results also indicated that Trif was downregulated following poly(I:C) or LPS stimulation, and Trim38 deficiency reduced the degree of poly(I:C)- or LPS-induced downregulation of Trif in BMDCs (Fig. 4G, Supplemental Fig. 2G). In these experiments, the levels of Traf6, Nap1, and Tab2 in Trim38^−/− and Trim38^+/+ BMDCs were unchanged and comparable before and after poly(I:C) or LPS stimulation (Fig. 4G, Supplemental Fig. 2G). These results suggest that endogenous Trim38 mediates downregulation of Trif, but not Traf6, Nap1, or Tab2, in BMDCs.

Because Trim38 mediates degradation of Trif via the ubiquitin-proteasomal pathway in mouse cells, we examined whether Trim38 could promote K48-linked polyubiquitination of Trif. The results showed that WT murine Trim38, but not Trim38(C31S), promoted K48-linked polyubiquitination of Trif (Fig. 5A, 5B). Furthermore, Trim38 deficiency markedly dampened poly(I:C)- and LPS-induced polyubiquitination of Trif and nearly abolished K48-linked polyubiquitination of Trif (Fig. 5C). These data suggest that Trim38 is essential for K48-linked polyubiquitination of Trif in mice.

We next investigated which lysine residues in murine Trif are targeted by Trim38 for K48-linked polyubiquitination. We constructed various lysine point mutants of Trif and tested the effects of Trim38 on their activation of ISRE in reporter assays. As shown in Fig. 5D and 5E, Trim38 markedly inhibited activation of ISRE mediated by WT and all of the tested mutants of Trif, with the exception of Trif(K228R). Further experiments indicated that Trim38 mediated K48-linked polyubiquitination of WT Trif and Trif(K321R) but not Trif(K228R) (Fig. 5F). Taken together, these results suggest that murine Trim38 promotes K48-linked polyubiquitination of murine Trif at Lys²²⁸.

Our recent study suggested that human TRIM38 negatively regulates TNF-α– and IL-1β–triggered NF-κB activation in human cell lines by mediating lysosomal-dependent degradation of TAB2/3, two critical components of the TAK1 kinase complex (11). Interestingly, this study also suggested that TRIM38-mediated lysosomal degradation of TAB2/3 is independent of its E3 ligase activity. In our earlier experiments, we also found that murine Trim38 could interact with and downregulate Tab2 in a lysosomal-dependent manner in a mammalian overexpression system (Fig. 4C, 4E, 4F). Surprisingly, Trim38 deficiency had no marked effects on the level of Tab2 in BMDCs (Fig. 4G). These results were puzzling to us. To determine whether endogenous Trim38 plays a regulatory role in TNF-α– and IL-1β–triggered signaling in primary mouse immune cells, we examined expression of downstream genes induced by TNF-α and IL-1β in Trim38^+/+ and Trim38^−/− BMDMs. The results indicated that the expression levels of Tnfa, Il1b, and Il6 induced by TNF-α and IL-1β were similar between Trim38^+/+ and Trim38^−/− BMDMs (see later discussion), which are consistent with our observation that endogenous Trim38 did not regulate the level of Tab2 in BMDCs (Fig. 4G) and suggest that endogenous Trim38 does not play a significant role in TNF-α and IL-1β signaling in BMDMs under physiological conditions.

Previously, it was reported that Trim30α, a member of the TRIM family, inhibits TLR-mediated activation of NF-κB by targeting Tab2 for lysosomal degradation in mice. However, deficiency of Trim30α potentiates TLR-mediated activation of NF-κB in LPS-pretreated, but not untreated, immune cells as a result of its low expression level in untreated cells (31). To investigate whether Trim38 acts similarly, we analyzed the mRNA levels of Trim38 in mouse BMDMs and BMDCs, as well as human cell lines (THP1 and HCT116), with which our previous experiments were performed (11). Interestingly, the levels of TRIM38 mRNA in human cell lines (HCT116 and Thp1) were ∼30-fold of those in primary mouse immune cells (BMDMs and BMDCs) (Fig. 6A). This prompted us to investigate whether Trim38, similar to Trim30α, regulates TNF-α and IL-1β signaling only at high levels of expression in mouse cells. First, we determined whether transcription of Trim38 is induced by various cytokines involved in innate immune and inflammatory responses. The results indicated that IFN-β and IFN-α4, but not TNF-α, IL-1β, or IFN-γ, dramatically induced the expression of Trim38 at both the mRNA and protein levels in mouse BMDMs (Fig. 6B, 6C).

To determine whether higher levels of Trim38 regulate TNF-α and IL-1β signaling in mouse immune cells, we treated BMDMs and BMDCs with IFN-β for 12 h to increase the expression of Trim38 before TNF-α or IL-1β stimulation was performed. qPCR experiments indicated that the expression levels of Tnfa, Il1b, and Il6 induced by TNF-α and IL-1β were similar between Trim38^+/+ and Trim38^−/− BMDMs and BMDCs under normal conditions. However, transcription of Tnfa and Il1b was significantly downregulated in IFN-β–primed Trim38^+/+, but not Trim38^−/−, BMDMs and BMDCs (Fig. 6D, Supplemental Fig. 3A). In addition, Trim38 deficiency potentiated TNF-α– and IL-1β–induced expression of inflammatory cytokines, such as Tnfa and Il1b, in IFN-β–primed, but not unprimed, BMDMs and BMDCs (Fig. 6E, Supplemental Fig. 3B). Consistently, the expression levels of Tab2/3 in IFN-β–primed Trim38^−/− BMDCs were markedly higher than in their WT counterparts. Moreover, TNF-α and IL-1β treatment induced downregulation of Tab2/3 in Trim38^+/+, but not Trim38^−/−, BMDCs (Fig. 6F). Because Trim38 is robustly induced by IFN-β, we examined the effects of IFN-β treatment on the expression of endogenous Trif and Tab2 in BMDCs. Interestingly, IFN-β treatment markedly decreased the protein level of Tab2, but not Trif (Fig. 6G), suggesting that IFN-β–induced Trim38 is sufficient for mediating degradation of Tab2 but not Trif. This is explained by our observations that the E3 ligase activity of Trim38 is required for its ability to mediate Trif degradation. Collectively, these results suggest that the high levels of Trim38 induced by IFN-β inhibit TNF-α– and IL-1β–induced downstream gene expression by promoting degradation of Tab2/3 in mouse immune cells.

Because IFN-β–induced Trim38 accumulation negatively regulates TNF-α and IL-1β signaling, we wondered whether Trim38 plays a feedback-regulatory role in the type I IFN–induced innate immune response. We found that mRNA levels of Cxcl10, Irf7, and Isg56 induced by IFN-α or IFN-β were comparable between Trim38^+/+ and Trim38^−/− BMDMs and BMDCs (Supplemental Fig. 4). These results suggest that Trim38 does not regulate signaling triggered by type I IFNs.

In this study, we investigated the functions of Trim38 in primary immune cells and in vivo using a mouse gene-knockout approach. We demonstrated that Trim38 deficiency potentiated poly(I:C)- and LPS-induced, but not R848- or PGN-induced, expression of type I IFNs (IFN-β and IFN-α4) and proinflammatory cytokines (TNF-α, IL-1β, and IL-6) in BMDMs, BMDCs, and MLFs. Trim38 deficiency also increased the serum cytokine levels induced by poly(I:C) and LPS and increased the susceptibility to body weight loss and death triggered by the administration of poly(I:C), LPS, and S. typhimurium. These results demonstrate that Trim38 negatively regulates TLR3/4-mediated cytokine production and innate immune and inflammatory responses in mouse immune cells in vivo.

Several experiments suggest that Trim38 negatively regulates TLR3/4-mediated signaling by targeting Trif for degradation via the ubiquitination-proteasomal pathway in mouse cells. First, endogenous coimmunoprecipitation experiments indicated that Trim38 was constitutively associated with Trif, and this association was enhanced following poly(I:C) or LPS stimulation of BMDCs. Second, Trim38, but not the enzyme-inactive mutant Trim38(C31S), mediated downregulation of Trif, which was blocked by the proteasomal inhibitor MG132. Third, WT Trim38, but not Trim38(C31S), promoted K48-linked polyubiquitination of Trif at Lys²²⁸, and Trim38 deficiency abolished K48-linked polyubiquitination of Trif in BMDCs. Fourth, the level of Trif in Trim38^−/− BMDCs was markedly upregulated compared with their WT counterparts before and after poly(I:C) or LPS stimulation.

Previously, it was reported that Trim38 inhibits TLR3/4-mediated signaling by promoting degradation of Traf6 and Nap1 in the mouse Raw264.7 cell line (12, 13). However, our studies, as well as reports by other investigators, suggest that Trim38 may not target Traf6 and Nap1 for regulation of TLR-mediated signaling in mice for the following reasons. First, we failed to detect an endogenous association between Trim38 and Traf6 or Nap1 in BMDCs both before and after poly(I:C) or LPS stimulation. Second, murine Trim38 could not cause degradation of Traf6 and Nap1 in cotransfection experiments, although human TRIM38 was able to mediate degradation of human NAP1 in similar experiments, as previously reported (12, 13). Third, the protein levels of Traf6 and Nap1 were comparable in Trim38^+/+ and Trim38^−/− BMDCs before and after poly(I:C) or LPS stimulation. Fourth, it was demonstrated that Traf6 is involved in signaling mediated by all TLRs. However, our studies suggest that Trim38 deficiency potentiated TLR3/4-mediated, but not TLR2- or TLR7-mediated, induction of cytokines. Finally, mouse gene-knockout studies demonstrated that Nap1 is not required for TLR3/4-mediated innate immune responses (32).

Previously, we reported that another E3 ubiquitin ligase, WWP2, also targets TRIF for K48-linked polyubiquitination and proteasomal degradation (16). However, WWP2 functions differently with TRIM38. First, Trim38 negatively regulates both TLR3- and TLR4-mediated signaling, whereas WWP2 specifically inhibits TLR3-mediated, but not TLR4-mediated, signaling. Second, WWP2 deficiency potentiates TLR3-mediated cytokine induction in BMDMs, but not in BMDCs, whereas Trim38 deficiency potentiates TLR3- and TLR4-mediated cytokine induction in both cell types. Third, Trim38 deficiency resulted in upregulation of Trif before and after poly(I:C) or LPS stimulation, whereas the level of Trif was comparable in cells from Wwp2^+/+ and Wwp2^−/− mice (Y. Yang and Y.Y. Wang, unpublished observations). Therefore, it is possible that Trim38 and WWP2 function in different cell types and distinct pathways.

Our experiments indicated that TRIM38 was highly expressed in the examined human cell lines (THP1 and HCT116) but was expressed at low levels in mouse BMDMs and BMDCs. Previous studies demonstrated that TRIM38 inhibits TNF-α– and IL-1β–induced signaling by targeting TAB2/3 for lysosomal-dependent degradation, and this effect is independent of the E3 ligase activity of TRIM38. Interestingly, Trim38 deficiency did not affect TNF-α– and IL-1β–induced signaling in BMDMs and BMDCs. One explanation is that Trim38 is expressed at low levels in BMDMs and BMDCs, which are below the threshold for Trim38 to exert its nonenzymatic activity to inhibit TNF-α and IL-1β signaling in primary mouse cells. Indeed, we found that expression of Trim38 was highly induced by type I IFNs (IFN-β and IFN-α4), but not by TNF-α, IL-1β, and IFN-γ, in primary mouse cells. Interestingly, TNF-α– and IL-1β–induced transcription of Tnfa, Il1b, and Il6 was significantly downregulated in IFN-β–primed Trim38^+/+, but not Trim38^−/−, BMDMs and BMDCs. In the meantime, Trim38 deficiency potentiated TNF-α– and IL-1β–induced expression of inflammatory cytokines in IFN-β–primed, but not unprimed, BMDMs and BMDCs. Consistently, the expression levels of Tab2/3 in IFN-β–primed Trim38^−/− BMDCs were markedly higher than in their WT counterparts. Moreover, TNF-α and IL-1β treatment induced downregulation of Tab2/3 in Trim38^+/+, but not Trim38^−/−, BMDCs. The simplest explanation for these data is that the high levels of Trim38 induced by IFN-β inhibit TNF-α and IL-1β signaling by promoting lysosomal-dependent degradation of Tab2/3 in mouse immune cells.

Based on our data, we propose a working model for the regulatory roles of Trim38 in innate immune and inflammatory responses. Upon ligand binding, TLR3/4 induces expression of type I IFNs and inflammatory cytokines, leading to the rapid onset of innate immune and inflammatory responses. Constitutively expressed Trim38, which is at a low level, provides initial negative regulation of TLR3/4-mediated signaling, whereas accumulated high levels of Trim38 induced by type I IFNs further inhibit TLR3/4-mediated responses in a feedback negative-regulatory manner during the late phase of infection. Trim38 negatively regulates TLR3/4 signaling by catalyzing the critical adapter protein Trif for K48-linked polyubiquitination and proteasomal degradation. During the late phase of infection, type I IFNs induce the accumulation of Trim38, which promotes lysosomal-dependent degradation of Tab2, independent of its E3 ligase activity, and negatively regulates TNF-α and IL-1β signaling. Thus, Trim38 negatively regulates TLR3/4-mediated innate immune and inflammatory responses by two sequential and distinct mechanisms. Our studies of Trim38 reveal an important mechanism for the host to initiate proper innate immune and inflammatory responses to defend against microbial invasion during the early phase of infection and then efficiently terminate cytokine production during the late phase of infection to prevent excessive and harmful inflammatory responses.

We thank Dr. Lin Guo for providing S. typhimurium.

The authors have no financial conflicts of interest.