Cell type-specific function of TRAF2 and TRAF3 in regulating type I IFN induction

TRAF3 is known as a central mediator of type I interferon (IFN) induction by various pattern recognition receptors, but the in vivo function of TRAF3 in host defense against viral infection is poorly defined due to the lack of a viable mouse model. Here we show that mice carrying conditional deletion of TRAF3 in myeloid cells or dendritic cells do not have a significant defect in host defense against vesicular stomatitis virus (VSV) infection. However, whole-body inducible deletion of TRAF3 renders mice more sensitive to VSV infection. Consistently, TRAF3 was essential for type I IFN induction in mouse embryonic fibroblasts (MEFs) but not in macrophages. In dendritic cells, TRAF3 was required for type I IFN induction by TLR ligands but not by viruses. We further show that the IFN-regulating function is not unique to TRAF3, since TRAF2 is an essential mediator of type I IFN induction in several cell types, including macrophages, DCs, and MEFs. These findings suggest that both TRAF2 and TRAF3 play a crucial role in type I IFN induction, but their functions are cell type- and stimulus-specific.


Background
Type I interferon (IFN) plays a crucial role in innate immunity against viral infections [1,2]. Induction of type I IFNs is typically mediated by pattern-recognition receptors (PRRs), including toll-like receptors (TLRs), RIG-I-like receptors (RLRs), NOD-like receptors, and the cytosolic GAMP synthase (cGAS), which detect various molecular patterns associated with viral genomes or replication products [2,3]. In response to ligand stimulation, the PRRs transduce signals via specific signaling adaptors. TIR-domain-containing adapter-inducing interferon-β (TRIF) mediates type I IFN induction by TLR3 and TLR4 upon stimulation by double-stranded RNA and lipopolysaccharide (LPS), respectively. Mitochondrial antiviral-signaling protein (MAVS, also known as VISA and IPS-1) serves as an adaptor of RLRs, whereas stimulator of interferon gene (STING) mediates type I IFN induction by cGAS and other cytoplasmic DNA sensors [2,4]. A common downstream target of the different PRR pathways is the protein kinase TBK1, which upon activation phosphorylates and activates the transcription factor interferon regulatory factor 3 (IRF3) [4,5]. Phosphorylated IRF3 forms a dimer and translocates to the nucleus to participate in the induction of type I IFN genes.
Activation of TBK1 and IRF3 by the PRRs also requires members of the tumor necrosis factor receptor (TNFR)-associated factors (TRAFs) family [3,6]. TRAFs are adaptor proteins or E3 ubiquitin ligases that transduce signals from TNFR superfamily as well as other immune receptors, including the PRRs [7]. While the typical functions of TRAFs include activation of NF-kB and MAP kinase signaling pathways, TRAF3 has been shown to mediate activation of IRF3 and implicated as a common signaling adaptor for type I IFN induction [2,8,9]. However, whether TRAF3 is a specific TRAF member that mediates type I IFN induction is in debate, since in some systems other TRAF members are also involved in the type I IFN induction [10,11]. The in vivo function of TRAF3 in mediating antiviral host defense is also poorly defined, since current studies have been largely relied on in vitro systems using immortalized mouse embryonic fibroblasts (MEFs) or other cell lines.
To study the physiological function of TRAF3 in antiviral innate immunity, we employed TRAF3 conditional knockout (KO) mice carrying TRAF3 deficiencies in different immune cell types or with inducible deletion of TRAF3 in adult mice. We obtained genetic evidence that whole-body inducible deletion of TRAF3 in adult mice attenuated their host defense against vesicular stomatitis virus (VSV) infection. Surprisingly, however, TRAF3 was dispensable in both myeloid cells and dendritic cells for this innate immune function. Using primary cells, we further demonstrated that the function of TRAF3 in mediating type I IFN induction is cell type-and stimulusspecific. Moreover, this innate immune function is not unique for TRAF3, since TRAF2 is equally important for type I IFN induction. The results provide new insight into the physiological function of TRAF proteins in mediating antiviral innate immunity.

In vivo function of TRAF3 in host defense against viral infection
Germline deletion of TRAF3 causes postnatal lethality, which has hampered studies to examine in vivo functions of TRAF3 [12]. To study the in vivo role of TRAF3 in regulating antiviral innate immunity, we generated myeloid cell-conditional Traf3 KO (Traf3 fl/fl Lyz2-Cre; hereafter called Traf3 MKO ) mice by crossing Traf3-flox mice [13] with Lyz2-Cre [14]. We then challenged these mice with VSV, a frequently used model for studying host defense against RNA virus infections [15]. VSV is known to induce type I IFN induction via the cytoplasmic RNA-responsive PRR RIG-I [15]. Surprisingly, TRAF3 deletion in myeloid cells did not sensitize mice to VSV infection, since the Traf3 MKO and wild-type control mice displayed comparable survival rate following intravenous (i.v.) infection with VSV (Fig. 1a). Since dendritic   (Fig. 1b). These results suggest that TRAF3 is dispensable in innate immune cells for host defense against VSV infections.
To further investigate the antiviral innate immune function of TRAF3, we employed an inducible KO system allowing TRAF3 deletion in different cell types in adult mice. We crossed the Traf3-flox mice with transgenic mice expressing a tamoxifen-inducible Cre recombinase, ER-Cre [16]. We then injected the Traf3 fl/fl ER Cre/+ and control Traf3 fl/fl mice with tamoxifen to generate the Traf3 inducible KO (Traf3 IKO ) and wild-type control mice. Immunoblot analysis using splenocytes and lymph node cells revealed efficient deletion of TRAF3 in the Traf3 IKO mice (Fig. 1c). Importantly, the whole-body deletion of TRAF3 in adult mice significantly attenuated the host defense against VSV infection (Fig. 1d). These results suggest that TRAF3 plays an important role in mediating antiviral immunity, but it may have cell typespecific functions.

TRAF3 is required for type I IFN induction in MEFs but not in macrophages
To better understand the cell type-specific functions of TRAF3, we prepared primary MEFs as well as innate immune cells from TRAF3-deficient or wild-type control mice. In agreement with previous studies [8,9], TRAF3 deficiency in primary MEFs severely attenuated the Ifna and Ifnb induction by the RNA viruses VSV and Sendai virus (SeV) (Fig. 2a, b). In addition, the TRAF3-deficient MEFs were also defective in type I IFN induction by transfected polyI:C known to mimic RNA viruses and stimulate the RIG-I signaling pathway (Fig. 2c). Consistently, parallel immunoblot assays revealed reduced phosphorylation of TBK1 and its target transcription factor IRF3 in TRAF3-deficient MEFs stimulated by VSV (Fig. 2d). Thus, TRAF3 is required for type I IFN induction by the RIG-I pathway in primary MEFs.
A recent study suggests that TRAF3 negatively regulates DNA virus-induced type I IFN induction with a mechanism that involves STING activation by the TRAF3-controlled kinase NIK [17]. Consistent with this report, we found that the TRAF3-deficient MEFs had greatly enhanced type I IFN induction by a DNA virus, herpes simplex virus 1 (HSV-1) (Fig. 2e). Surprisingly, however, the TRAF3 deficiency did not promote, but rather significantly attenuated, the induction of type I IFN induction by non-viral STING agonists, cGAMP ( Fig. 2f ), ISD (Fig. 2g) or DMXAA (Fig. 2h). Consistently, while TRAF3 deficiency promoted the activation of STING/TBK1/IRF3 signaling pathways by HSV-1 (Fig. 2i), it attenuated these signaling events induced by cGAMP (Fig. 2j). These results suggest that TRAF3 may play different roles in regulating STING activation and upstream signaling steps in the DNA-sensing pathway.
To examine the role of TRAF3 in innate immune cells, we prepared primary bone barrow derived macrophages (BMDMs) from Traf3 MKO and wild-type control mice. In contrast to the results obtained with MEFs, TRAF3 deletion in BMDMs had no effect on the induction of Ifna and Ifnb genes by VSV or SeV (Fig. 2k, l). On the other hand, as seen in MEFs, the TRAF3 deficiency in macrophages promoted type I IFN induction by the DNA virus HSV (Fig. 2m). Furthermore, TRAF3 was completely dispensable for the induction of Ifna and Ifnb by the TLR ligands LPS and polyI:C (Fig. 2n, o). Consistently, loss of TRAF3 in macrophages did not affect LPS-stimulated phosphorylation of TBK1 or its homolog IKKi (Fig. 2p). These results suggest cell type-specific functions of TRAF3 in mediating type I IFN induction.

Inducer-specific function of TRAF3 in DCs
DCs also serve as an important innate immune cell type mediating host defense against viral infections. We thus examined the role of TRAF3 in regulating type I IFN induction in primary bone marrow derived DCs (BMDCs). Compared to the wild-type BMDCs, the TRAF3-deficient BMDCs had a significant reduction in type I IFN gene induction by both LPS and polyI:C ( Fig. 3a, b). Consistently, the TRAF3-deficient DCs also had impaired induction of TBK1 phosphorylation by LPS (Fig. 3c). Interestingly, however, TRAF3 was completely dispensable for induction of Ifna and Ifnb by the RNA viruses VSV and SeV (Fig. 3d, e). Similarly, TRAF3 deletion in DCs had no significant effect on type I IFN induction by the DNA virus HSV (Fig. 3f ). To further confirm these results, we also directly stimulated the RIG-I and STING pathways using transfected polyI:C and cGAMP, respectively. Once again, induction of type I IFN gene expression by these inducers did not require TRAF3 in DCs (Fig. 3g, h). Thus, the function of TRAF3 in mediating type I IFN induction is dependent on inducers as well as cell types.

TRAF2 is an essential mediator of type I IFN induction in multiple cell types
TRAF3 has been frequently cited in the literature as a central mediator of antiviral innate immunity; however, it is unclear whether this function is unique for TRAF3 or also for other TRAF members. To address this question, we examined the role of TRAF2 by generating primary cells from TRAF2-conditional KO and wild-type control mice. Interestingly, like TRAF3, TRAF2 was essential for type I IFN induction by the RNA virus SeV in primary MEFs (Fig. 4a). This finding was intriguing, since it suggests that deletion of either TRAF2 or TRAF3 in MEFs blocks type I IFN induction, thus suggesting non-redundant functions of TRAF2 and TRAF3. Remarkably, TRAF2 was also required for type I IFN induction by the DNA virus HSV-1 (Fig. 4b), which was in sharp contrast to the negative role of TRAF3 in mediating the DNA-sensing pathway (Fig. 2e). These results suggest that TRAF2 has both similar and different functions from TRAF3 in the regulation of type I IFN induction.
We next examined the role of TRAF2 in regulating type I IFN induction by TRIF-dependent TLRs, TLR3 and TLR4, using macrophages. BMDMs prepared for the Traf2 MKO mice had a severe defect in LPS-stimulated Ifna and Ifnb expression and also a significant reduction in polyI:C-stimulated Ifnb expression (Fig. 4c, d). Consistently, the TRAF2 deficiency attenuated, although did not completely block, LPS-stimulated TBK1 phosphorylation (Fig. 4e). TRAF2 deficiency did not appreciablely alter the activation of IKKi or other signaling factors, suggesting the involvement of TRAF2 in regulating the TBK1 Traf3 . p Immunoblot analysis of the indicated proteins in whole-cell lysates of WT or Traf3 MKO BMDMs stimulated with LPS for the indicated time periods. All qRT-PCR data are presented as fold relative to the amount of an internal control, Actb. Data are representative of three to four independent experiments, and statistical analyses are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. ns nonsignificant signaling axis. Parallel studies using the RNA virus SeV and the DNA virus HSV-1 did not reveal significant differences in Ifna and Ifnb induction between the Traf2 MKO and wild-type control macrophages (Fig. 4f, g).
To investigate the role of TRAF2 in DCs, we generated DC-conditional TRAF2 KO (Traf2 DC-KO ) mice. BMDCs prepared from the Traf2 DC-KO mice displayed a severe defect in type I IFN induction by LPS and polyI:C (Fig. 5a, b). In contrast to the TLR pathways, induction of type I IFNs by the RNA viruses, SeV and VSV, and the DNA virus HSV-1 was not affected by the TRAF2 deficiency in DCs (Fig. 5c-e). Collectively, these results suggest that like TRAF3, TRAF2 has an essential role in mediating type I IFN induction in an inducer-specific manner.

Discussion
The results presented in the present study provide genetic evidence that TRAF3 mediates antiviral host defense in mice. Our data support the previous reports that TRAF3 plays a crucial role in regulating type I IFN induction [8,9]. However, by employing different types of primary cells as well as Traf3 conditional KO mice, we demonstrated that this function of TRAF3 was cell typeand stimulus-specific (Table 1). Surprisingly, deletion of TRAF3 in either myeloid cells or dendritic cells had no effect on host defense against VSV infection, but wholebody inducible deletion of TRAF3 impaired this innate immune function. Consistent with these in vivo results, cell culture studies revealed that TRAF3 was dispensable for type I IFN induction by both TLR ligands and viruses in macrophages. Furthermore, TRAF3 deletion in DCs had no effect on virus-induced type I IFN induction, although it inhibited IFN induction by TLR3 and TLR4 ligands. In contrast, TRAF3 was essential for IFN induction by RNA viruses in MEFs. These results suggest that TRAF3 is dispensable for antiviral responses in innate immune cells but may play an essential role in other cell types, such as fibroblasts.
Our finding that TRAF3 is dispensable in myeloid cells for innate immunity against VSV infection is in agreement with a previous study that myeloid cell-specific deletion of TRAF3 has no effect on LPS-induced IFNβ production in vivo [18]. However, in contrast to our present finding that TRAF3 is dispensable for LPS-stimulated type IFN expression in macrophages, this prior . Data are representative of three independent experiments, and statistical analyses are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001. ns nonsignificant study suggests that TRAF3 is required for in vitro type I IFN induction by LPS in macrophages. The reason for this discrepancy is unclear, but it could be due to the use of two different Traf3-flox mouse strains or differences in experimental conditions. Nevertheless, we found that TRAF3 deletion in macrophages also did not affect type I IFN induction by another TLR ligand, polyI:C. Future studies will further study the role of TRAF3 and other TRAF members in TLR-stimulated type I IFN induction.
Although TRAF3 has been implicated as a specific antiviral signaling adaptor, our data suggest that TRAF2 is equally important for mediating type I IFN induction. In fact, TRAF2 displayed an essential role in mediating type I IFN induction in several cell types, including MEFs, macrophages, and DCs (Table 2). In macrophages and DCs, TRAF2 appeared to be more important for the TRIF-dependent TLR pathways, since TRAF2 deletion in these cells inhibited type I IFN induction by TLR3 and TLR4 ligands but not by RNA or DNA viruses. The molecular mechanism underlying the cell type-and inducer-specific functions of TRAF2 and TRAF3 is currently unclear, but it may involve functional redundancy with other TRAF members. However, it is remarkable that TRAF2 and TRAF3 both had non-redundant roles in type I IFN induction in the same cell types, including MEFs and DCs, since deletion of either TRAF2 or TRAF3 impaired the induction of type I IFNs by the same types of inducers. This finding suggests that TRAF2 and TRAF3 may function cooperatively, as seen in the regulation of the noncanonical NF-kB kinase NIK [19], or at different steps of the IFN induction pathways.
A recent study suggests that TRAF3 negatively regulates type I IFN induction by DNA viruses, such as HSV-1, through controlling the noncanonical NF-κB inducing kinase NIK [17]. Accumulation of NIK due to TRAF3 deficiency promotes STING activation and induction of type I IFNs [17]. In support of this report, we found that TRAF3 deletion in MEFs and macrophages promotes HSV-1-stimulated expression of Ifna and Ifnb genes. However, the mechanism by which TRAF3 regulates STING pathway appears to be complex. First, while TRAF3 deficiency promoted type I IFN induction by HSV-1, it inhibited type I IFN induction by the STING stimulators cGAMP, ISD, and DMXAA. These results suggest that TRAF3 may have dual functions in DNA-stimulated type I IFN production. It is possible that TRAF3 may positively regulate STING activation of TBK1 or IRF3 and negatively regulates an upstream step in the DNA-sensing pathway. Second, TRAF2 deficiency, which also causes NIK accumulation [20], did not promote type I IFN induction by HSV-1, which argues for the involvement of a NIKindependent mechanism. In fact, deletion of TRAF2 in MEFs largely blocked the induction of type I IFN genes by HSV-1. These results, along with the finding that TRAF3 deficiency blocks cGAMP-induced type I IFN induction, further suggest a positive role for TRAF2 and TRAF3 in regulating STING-mediated signaling. HSV-1 has evolved multiple strategies to evade host antiviral responses. Some HSV-1-encoded gene products have been identified inhibitors targeting upstream steps of the STING/TBK1/IRF3 pathway. For example, the HSV-1 tegument protein UL41 inhibits type I IFN induction by selectively degrading the mRNA of cGAS, a key DNA sensor mediating cGAMP synthesis and STING activation by DNA viruses [21]. Another HSV-1 tegument protein, UL37, deamidates mouse cGAS, impairing its ability to catalyze cGAMP synthesis [22]. Moreover, an HSV-1 virulence factor, γ134.5, inhibits STING activation by interfering the translocation of STING from endoplasmic reticulum to Golgi apparatus [23]. Future studies will examine whether TRAF3 plays

Conclusions
The findings of the present study suggest that in addition to TRAF3, TRAF2 is a crucial mediator of type I IFN induction and that the function of TRAF2 and TRAF3 is cell type-and stimulus-specific.

Cell preparation and stimulation
Bone marrows were prepared from the femurs of adult mice and cultured in a M-CSF conditional medium for BMDM differentiation. BMDCs were generated by cultivating bone marrow cells in RPMI medium supplemented with GM-CSF (10 ng/ml) for 7 days, and the differentiated BMDCs were stained with Pacific blueconjugated anti-CD11c and purified by flow cytometric cell sorting as described [26]. To prepare primary MEFs, we bred heterozygous mice for obtaining the KO and WT embryos from the same pregnant female mice. The MEFs, BMDMs, and BMDCs were starved overnight in medium supplemented with 0.5% FCS before being stimulated with LPS (1 μg/ml for immunoblot experiments and 100 ng/ml for cytokine induction experiments), poly(I:C) (20 μg/ml), DMXAA (50 μg/ml), Lipofectamine-transfected poly(I:C) (20 μg/ml), cGAMP (10 μg/ml) or ISD (1 μg/ml) for the indicated times. Total and subcellular extracts were prepared for immunoblot assays, and total RNA was prepared for QPCR assays.

Immunoblot
Whole-cell lysates or subcellular extracts were prepared and subjected to immunoblot assays as described [27,28]. The samples were resolved by 8.25% SDS-PAGE. After electrophoresis, separated proteins were transferred onto polyvinylidene difluoride membrane (Millipore). For IB assays, the polyvinylidene difluoride membrane was blocked with 5% non-fat milk. After incubation with specific primary antibody, horseradish peroxidaseconjugated secondary antibody was applied. The positive immune reactive signal was detected by ECL (Amersham Biosciences).

Viral infection
For gene induction and signaling analysis, MEFs were seeded into 12-well plates (5 × 10 5 cells per well) and infected with VSV-AV1 mutant, HSV-1 (KOS) or Sev strains in serum-free medium for 1 h. The cells were washed once and cultured in growth medium for the indicated time periods and then collected for immunoblot or QPCR assays. For mouse infection, the indicated age-matched (6-8 weeks' old) Traf3 conditional KO or WT control mice were housed in a biosafety level 2 facility and infected i.v. with VSV (2 × 10 7 PFU per mouse in 200 μl). The infected mice were monitored for survival up to 14 days.

Statistical analysis
Prism software was used for two-tailed unpaired t-tests and data are presented as mean ± SEM. Log-rank and Gehan-Wilcoxon tests were performed for survival curves. p-values < 0.05 and 0.01 are considered significant and very significant, respectively.