TLR7/9 versus TLR3/MDA5 signaling during virus infections and diabetes

Review on the cellular and molecular sources of IFN-I during virus infections with emphasis on plasmacytoid dendritic cells (pDC) and dsRNA sensors.

Abstract

IFN-I are pleiotropic cytokines that impact innate and adaptive immune responses. In this article, we discuss TLR7/9 versus TLR3/MDA5 signaling in antiviral responses and diabetes. pDCs are thought to have a critical role in antiviral defense because of their ability to rapidly secrete large amounts of IFN-I through TLR7/9 signaling. A recent study demonstrates that although pDCs are a source of IFN-I in vivo, their overall contribution to viral containment is limited and time-dependent, such that additional cellular sources of IFN-I are required to fully control viral infections. dsRNA sensors, such as TLR3 and MDA5, provide another important trigger for antiviral IFN-I responses, which can be exploited to enhance immune responses to vaccines. In the absence of infection, IFN-I production by pDCs or from signaling through dsRNA sensors has been implicated in the pathogenesis of autoimmune diseases such as diabetes. However, recent data demonstrate that IFN-I production via TLR3 and MDA5 is critical to counter diabetes caused by a virus with preferential tropism for pancreatic β-cells. This highlights the complexity of the host antiviral response and how multiple cellular and molecular components balance protective versus pathological responses.

Introduction

The majority of cells in the body is equipped with the machinery to produce and respond to IFN-I (i.e., IFN-α, IFN-β). IFN-I were first described over 50 years ago as factors that interfere with virus replication [1–3]. Since then, it has become clear that in addition to having potent antiviral properties, IFN-I possess immunomodulatory activities [4]. IFN-I promote T cell responses, including CD8⁺ T cell effector functions, survival and memory [5–8], and Th1 polarization of CD4⁺ T cells [9]. IFN-I enhance NK cell activation [10] and support the differentiation and maturation of DCs, allowing them to prime naïve T cells effectively [11–13]. The differentiation of B cells into antibody-secreting plasma cells [14–16] and the proliferation of dormant hematopoietic stem cells [17] are also induced by IFN-I. Thus, IFN-I confer resistance to viruses and promote multiple immune functions. This review focuses on cellular and molecular sources for IFN-I during virus infections and diabetes with emphasis on pDCs and the dsRNA sensors, TLR3 and MDA5.

pDCs SPECIALIZE IN THE SECRETION OF IFN-I

pDCs are bone marrow-derived cells that specialize in the secretion of IFN-I. They were first described over 30 years ago as natural IPCs that activate NK cells upon exposure to viruses [18, 19]. The murine equivalent to human IPC was described in 2001 [20–22]. pDCs detect RNA and DNA viruses through two endosomal sensors, TLR7 and TLR9, respectively, which induce secretion of IFN-I through the MyD88-IRF7 signaling pathway [23, 24]. pDCs can produce IFN-I independently of IFNAR feedback signaling [25]; however, they do respond to IFN-I, generating an autocrine circuit through IFNAR, which augments IFN-I secretion and induces activation and migration [26, 27].

Upon engagement with nucleic acid ligands, TLR7 and TLR9 signal through the cytosolic adaptor MyD88. To induce IFN-I, MyD88 associates with IRAK1/4, TRAF3, IKKα, and osteopontin, leading to the phosphorylation and activation of IRF7 [23, 24, 28]. Phosphorylated IRF7 translocates to the nucleus and induces the transcriptional activation of IFN-I genes [29, 30]. IRF7 is positively regulated by the PI3K-mammalian target of rapamycin-p70S6K pathway [31] and repressed at the translational level by eukaryotic initiation factor 4E-binding proteins [32]. To induce inflammatory cytokines, MyD88 associates with IRAK1/4, triggering the TRAF6-TGF-β-activated kinase 1 pathway, which activates NF-κB and MAPK. MyD88 also recruits IRF5, which cooperates with NF-κB to induce inflammatory cytokines and chemokines [33].

The endosomal compartmentalization of TLR7/9 and their ligands determines whether IFN-I or inflammatory cytokines are produced. Condensed or multimeric RNA/DNA aggregates in early endosomes, eliciting the production of IFN-I, whereas monomeric forms reach late endosomes, resulting in cytokine secretion and pDC maturation [34, 35]. To induce IFN-I, it is necessary for TLR7 and TLR9 to translocate from the ER to a specialized lysosome-related organelle. This process requires adaptor protein 3, as well as Slc15a4, BLOC-1, and BLOC-2 [36, 37]. TLR9 activation also requires proteolytic cleavage in the endosome as well as Ly49Q and granulin [38–42].

Many viruses can promote pDC activation and secretion of IFN-I via the TLR7/9-MyD88-IRF7 pathway, such as MCMV, LCMV, HSV, VSV, HIV, HTLV-1, MHV, HCV, NDV, influenza, SeV, CVB, pneumovirus, and EBV [43–62]. Although TLR7/9 signaling promotes pDC activation and IFN-I secretion, TLR7 and TLR9 are also expressed in other immune cells, including monocytes, classical DCs, and B cell subsets [63, 64]. Thus, it is important to clarify to what extent pDCs contribute to innate and adaptive antiviral responses elicited by viruses sensed by the TLR7/9 pathway. TLR8 also uses the MyD88 signaling pathway and was initially shown to sense ssRNA in human monocytes, whereas the murine homologue was thought to be nonfunctional [65]. However, it was reported recently that viral DNA from vaccinia virus promotes pDC activation in a TLR8-dependent manner, suggesting the involvement of TLR8 in detecting ssRNA in mice [66].

IN VIVO ABLATION OF pDCs

To determine the contribution of pDCs in host-pathogen interactions, pDCs have been depleted in animal models with mAb. In early studies, pDCs were eliminated with a mAb specific for the antigen Gr-1 [20, 67], which is shared by Ly6C and Ly6G. Ly6C is expressed on pDCs but also on plasma cells [68], memory T cells [69], and inflammatory monocytes [70], whereas Ly6G is expressed on granulocytes. Thus, anti-Gr-1 mAb can deplete multiple cell types in addition to pDCs.

Several other pDC-depleting mAbs have been developed, including 120G8 [71], mouse pDC antigen 1 [47], and 927 [72]. These mAbs recognize BST2, which is expressed selectively on mouse pDCs and plasma cells in naïve mice but is up-regulated on most cell types following stimulation with IFN-I and IFN-γ [72]. Thus, the activation-induced, nonspecific expression of BST2 has important implications for using anti-BST2 mAb to deplete pDCs, as they may also deplete additional cell types during an immune response.

Given the limitations of pDC-depleting mAbs, mice lacking pDCs would be extremely valuable. Mice carrying a hypomorphic mutation in the Ikaros gene (Ik^L/L) lack pDCs [73]; however, Ikaros is a lymphoid cell-specific transcription factor that influences the development of other immune cell types as well [74–77]. Within the past few years, mice have been generated that more specifically lack pDCs, either constitutively or by inducible depletion [78, 79]. pDCs are absent in mice lacking the basic helix–loop–helix transcription factor E2-2/Tcf4, which is essential for pDC development [78]. In BDCA2-DTR Tg mice, which selectively express the DTR on pDCs, administration of DT results in the transient but specific depletion of pDCs [79]. Thus, E2-2-deficient mice and BDCA2-DTR Tg mice will be useful for assessing the contribution of pDCs to innate and adaptive immune responses to pathogens.

pDCs ARE AN EARLY BUT LIMITED SOURCE OF IFN-I DURING VIRUS INFECTIONS

In a recent study, BDCA2-DTR Tg mice were challenged with representative RNA and DNA viruses, VSV and MCMV, respectively [79]. It was shown that pDCs provide an early but limited source of IFN-I during VSV and MCMV infections (Fig. 1). In the case of VSV infection, IFN-I produced by pDCs limits viral replication and perpetuates the survival and accumulation of antigen-specific CD8⁺ T cells. During MCMV infection, pDC-derived IFN-I can partially control viral burden, which reduces the expansion of MCMV-specific Ly49H⁺ NK cells. The ability of pDCs to contain MCMV replication is highly dependent on the initial viral inoculum. At low viral doses, pDCs are effective at restraining viral replication; however, at higher doses, their contribution to antiviral responses appears to be negligible. The restricted capacity of pDCs to augment antiviral responses when higher doses of virus are administered may be a result of their limited numbers at sites of viral replication. Although pDCs are only capable of controlling viral replication at low-to-intermediate doses, many natural infections occur in vivo by transmission of low numbers of virions. Therefore, in physiological settings, pDCs might be sufficient for antiviral responses. Furthermore, pDC-derived IFN-I might augment the expression of IFN-induced genes, including viral sensors, which promote an antiviral state [1–3], and further amplify IFN-I responses.

Figure 1. Cellular sources of IFN-I during viral infections.

pDCs detect RNA and DNA viruses through endosomal TLR7 and TLR9, respectively, leading to the secretion of IFN-I. IFN-I production by pDCs is early and transient, suggesting a need for additional cellular sources that provide a more broad, extended production of IFN-I to resolve viral infections. Other cellular sources that are critical in the production or responsive to IFN-I during virus infections include monocytes, B cells, DCs, macrophages, and stromal cells, which can produce IFN-I via TLRs (TLR2, TLR3, TLR7, TLR9) and RLRs (RIG-I, MDA5, and LGP2), and cytoplasmic DNA sensors, which are not discussed in this review (DNA-dependent activator of IRFs; IFN-γ-inducible protein IFI16 or p204; stimulator of IFN genes; HMGB1 and HMGB2; DHX36) [79–86]. Cytoplasmic helicases (DDX1-DDX21-DHX36 complex) have been identified, which sense dsRNA and activate IFN-I responses in the cytosol of myeloid DCs [87]. Macrophage subsets include subcapsular sinus macrophages, splenic marginal zone macrophages, metallophilic macrophages, microglia, and tissue macrophages, such as those found in the liver and lung (i.e., Kupffer and alveolar).

DIVERSITY AND SPECIFICITY OF dsRNA SENSORS

TLR7 and TLR9 are only two of multiple pathogen recognition sensors that detect viral nucleic acids. Viral dsRNA is sensed by TLR3 and the RLRs: RIG-I, MDA5, and LGP2 [23, 89–93]. TLR3 is mainly expressed in hematopoietic cells, particularly DCs and macrophages, but also in some stromal cells. TLR3 detects dsRNA, which gains access to the endosomal compartment by phagocytosis of virus-infected cells or apoptotic cell debris, internalization of antibodies bound to viruses, or autophagy [48, 58, 94]. TLR3 transmits signals through the TRIF pathway, which results in the phosphorylation and nuclear translocation of IRF3 and the transcription and secretion of IFN-β [95, 96]. TLR3-TRIF signaling also activates NF-κB and the transcription of inflammatory cytokine genes. The specificity of TLR3 for dsRNA allows recognition of RNA viruses such as EMCV, influenza A virus, CVB, WNV, DV, reovirus, and rhinovirus [97–106]. Moreover, TLR3 detects the DNA viruses MCMV and HSV-1, most likely through recognition of RNA intermediates, which may be generated during viral replication [107–111].

RIG-I and MDA5 are two cytosolic helicases induced by IFN-I in most cell types. RIG-I and MDA5 detect dsRNA intermediates that accumulate in the cytosol during viral replication [23, 91–93] and interact with the adaptor molecule IPS-1 (also known as mitochondrial antiviral-signaling protein, virus-induced signaling adapter, or Cardif) [112–115]. IPS-1 is localized on the mitochondria and recruits TRAF3, which activates TRAF family member-associated NF-κB-binding kinase 1 and IKKε, leading to the phosphorylation and nuclear translocation of IRF3 and IRF7 and production of IFN-β and IFN-α [116–118]. Additionally, IPS-1 associates with FADD protein and receptor-interacting protein-1, which activate caspase-8 and caspase-10, resulting in NF-κB activation and production of inflammatory cytokines [112, 119, 120]. IPS-1 is also located on peroxisomes and facilitates rapid antiviral responses through IRF1 [121].

RIG-I and MDA5 detect distinct dsRNA forms that differ in structure, length, and 5′ cap structures [122–128]. The distinct ligand preferences of the MDA5 and RIG-I receptors confer specific recognition of disparate viruses. RIG-I has been shown to detect paramyxoviruses (SeV, NDV, respiratory syncytial virus, and Measles virus); orthomyxovirus (influenza A and B viruses); rhabdovirus (VSV and Rabies virus); flavivirus (Japanese encephalitis virus, HCV, WNV, and DV); filovirus (Ebola virus); reovirus; and metapneumovirus [106, 123, 128–144]. Recent studies have shown that RIG-I also recognizes DNA viruses by detecting RNA intermediates generated through the RNA polymerase III-mediated transcription of dsDNA [145, 146]. MDA5 detects picornaviruses such as EMCV, CVB, Mengo virus, and Theiler virus, as well as murine norovirus 1 [132, 146–149]. MDA5 is also involved in the recognition of WNV, DV, reovirus, SeV, MHV, Measles virus, and LCMV [106, 134, 138, 139, 150–153].

LGP2 is another RLR that detects dsRNA [154–156]. LGP2 does not contain any signaling domains and was initially thought to negatively regulate MDA5 and RIG-I [156]. Accordingly, LGP2-deficient mice have more robust IFN-I responses following poly(I:C) stimulation and VSV infection compared with WT mice [155]. However, recent data have demonstrated that LGP2 may positively influence antiviral responses, as RLR-mediated IFN-I responses were impaired in mice lacking LGP2 or the LGP2 ATP-binding site [154].

DISTINCT ROLES OF TLR3 AND MDA5 IN poly(I:C)-STIMULATED IMMUNE RESPONSES

Recent studies demonstrate that IFN-I responses via dsRNA sensors can be exploited in vaccination to enhance adaptive immune responses to antigens. Poly(I:C) is a synthetic analog of viral dsRNA, which is mainly detected by TLR3 and MDA5 [98, 132, 147, 157, 158], although RIG-I can also detect short forms of poly(I:C) [126]. By stimulating hematopoietic and stromal cells through TLR3 and MDA5, poly(I:C) induces a rise in systemic IFN-I, which activates DCs, boosting their capacity to elicit antigen-specific CD4⁺ T cells [159–162]. Poly(I:C) also stimulates antigen-specific CD8⁺ T cell immunity. Initial studies showed that poly(I:C) triggers TLR3 in CD8α⁺ DCs, which then acquire the capacity to cross-prime CD8⁺ T cells [94]. Then, it was found that poly(I:C) requires TRIF and IPS-1 signaling pathways to boost antigen-specific CD8⁺ T cell and antigen-specific B cell responses [163, 164], suggesting that TLR3 and MDA5 impact T and B cell responses. Accordingly, recent bone marrow chimera experiments demonstrated that poly(I:C) stimulates TLR3 in hematopoietic cells, most likely DCs, promoting cross-priming of antigen-specific CD8⁺ T cells, i.e., primary responses [165]. Additionally, poly(I:C) stimulates MDA5 in stromal cells, inducing a rise in systemic IFN-I, which supports the survival of antigen-primed CD8⁺ T cells and the establishment of CD8⁺ T cell memory [165] (Fig. 2).

Figure 2. Poly(I:C) promotes T cell responses and NK cell activation through TLR3 and MDA5 signaling.

(A) TLR3 and MDA5 impact CD8⁺ and CD4⁺ T cell responses. Poly(I:C)-mediated stimulation of TLR3 is required in hematopoietic cells to promote cross-priming of antigen-specific CD8⁺ T cells, i.e., primary responses, whereas poly(I:C)-mediated stimulation of MDA5 is required in stromal cells to induce a systemic rise in IFN-I secretion that promotes the survival of antigen-primed CD8⁺ T cells and the establishment of CD8⁺ T cell memory. IFN-I, from hematopoietic and stromal cells, is necessary for the adjuvant function of poly(I:C) to induce effective CD4⁺ Th1 immunity. (B) Roles of TLR3 and MDA5 in poly(I:C)-mediated NK cell activation. TLR3 contributes to NK cell activation by promoting IFN-I and IL-12 release by hematopoietic cells, whereas MDA5 activates NK cells indirectly, by promoting stromal cell release of IFN-I.

Poly(I:C)-mediated triggering of TLR3 and MDA5 can also be used to promote NK cell activity against cancer cells. Studies in humans showed that poly(I:C) promotes NK activities by triggering TLR3 expressed in cultured NK cells [166–169]. In contrast, murine NK cells do not express TLR3, and poly(I:C)-mediated activation of NK cells is indirect. Initially, it was shown that poly(I:C) activates DCs through TLR3 and that DCs in turn activate NK cells [170]. Subsequently, it was demonstrated that poly(I:C) triggers the TRIF and IPS-1 pathways in CD8α⁺ DCs, which then acquire the ability to activate NK cells [171], implicating TLR3 and MDA5 in poly(I:C)-mediated NK cell activation. A recent study showed that NK cell activation is mostly dependent on MDA5, whereas TLR3 has a modest impact, only evident in the absence of MDA5 [172] (Fig. 2). TLR3 contributed to NK cell activation by promoting IFN-I and IL-12 release by hematopoietic cells. MDA5 acted indirectly by promoting stromal cell release of IFN-I, which activated NK cells. Thus, in mice, poly(I:C) stimulates NK cell responses by inducing the production of various cytokines in a manner dependent on the cell type and dsRNA sensor.

pDCs AND dsRNA SENSORS IN DIABETES

A major outcome of TLR7/9 and TLR3/MDA5 signaling is IFN-I production. Although IFN-I is critical for inducing an antiviral state and promoting protective immune responses, it has been proposed that excessive IFN-I production can result in autoimmunity [4, 173–176]. Exaggerated IFN-I has been associated with T1D, an autoimmune disease caused by T cell destruction of pancreatic β-cells [177–179]. One report demonstrated that Tg mice expressing IFN-I in insulin-producing β-cells develop islet inflammation and hypoinsulinemic diabetes, which can be prevented by neutralizing IFN-I [180]. Subsequently, it was shown that IFN-I accelerates T1D in NOD mice and breaks tolerance to β-cells in nondiabetes-prone mice [181]. A more recent study revealed that NOD mice have increased IFN-I levels and IFN-I-producing pDCs in pancreatic LNs prior to the onset of diabetes [182], shortly after the wave of β-cell apoptosis, which peaks at postnatal days 14–17 [183]. Blockade of IFNAR in NOD mice at 2–3 weeks of age delayed the onset and incidence of T1D. Moreover, mAb-mediated depletion of pDCs in young NOD mice also delayed the onset and incidence of T1D [184]. Although these studies were conducted in mice, the expansion of IFN-I-producing pDCs has been documented in human patients with T1D around the time of diagnosis [185]. Thus, pDC-derived IFN-I may serve a pathogenic role in the development of T1D.

poly(I:C) can hasten T1D in animal models [186, 187], suggesting that dsRNA sensors, such as TLR3 and RLRs, may also promote T1D in an IFN-I-dependent manner. In support of this hypothesis, genome-wide association studies have demonstrated that human polymorphisms in the RIG-I and MDA5 genes, which reduce IFN-I responses to dsRNA, are linked with resistance to T1D [188–192]. Interestingly, It has been reported that NK cells mediate islet inflammation in humans with T1D, infiltrate the islets of NOD mice, and are required for accelerated T1D driven by IFN-I [193–195]. A recent study found that the activating receptor NKp46, expressed almost exclusively by NK cells, is essential for the development of T1D in NOD mice and in the streptozotocin diabetes model [196]. Thus, IFN-I produced by pDCs (TLR7/9) or other cell types via dsRNA sensors may promote NK cell accumulation and activation in islets, which results in NK cell-mediated destruction of β-cells and T1D.

Paradoxically, other studies have found that treatment with IFN-I can be protective and reduce the incidence of T1D in NOD mice [197–199]. Furthermore, NOD mice have delayed onset or are protected completely from diabetes after exposure to certain viruses [200–202], suggesting that viral detection and IFN-I responses may promote T cell tolerance by enhancing T regulatory cell activity and up-regulation of inhibitory molecules and cytokines [202–205].

Certain viruses have tropism for the pancreas and can induce diabetes by directly destroying β-cells or by causing immune cell-mediated β-cell death [177]. In the context of these viral infections, IFN-I may be protective [177]. Acute LCMV infection in Tg mice expressing the LCMV glycoprotein under control of the RIP-GP causes diabetes [206]. In this model, pDCs counter infection and diabetes through IFN-I secretion [203]. Paradoxically, another study reported that hematopoietic-derived IFN-I was essential for CD8⁺ T cell-mediated destruction of β-cells and induction of T1D in LCMV-infected RIP-GP mice [207]. CVB and EMCV-D replicate in the pancreas of mice, causing extensive tissue damage and diabetes [208, 209]. Diabetes onset in CVB-infected mice has been attributed to bystander damage of β-cells, leading to the release of sequestered islet antigens, which results in the restimulation of resting, autoreactive T cells [210]. In contrast, diabetes induced by EMCV-D appears to be a consequence of direct β-cell destruction mediated by macrophages rather than T cells [211]. Clinical studies have also linked coxsackievirus and enterovirus infections to human T1D, but whether these viruses induce direct β-cell damage or trigger autoimmunity remains unclear [212, 213]. It has been shown that IFN-I induces an antiviral state in human islet cells and protects them from coxsackievirus replication [214]. Thus, IFN-I may be critical for β-cell survival and protection from virus-induced T1D in humans.

Given that EMCV replication generates dsRNA intermediates that trigger IFN-I responses through MDA5 [132, 147] and TLR3 [97], we recently examined the relative contributions of these sensors in the host response to EMCV-D and development of diabetes [215] (Fig. 3). We found that TLR3-deficient mice develop diabetes as a result of hematopoietic cells failing to mount an early IFN-I response that protects β-cells from virus-induced damage. The role of MDA5 in diabetes could not be assessed precisely, as MDA5 has a dominant role in protecting the heart against EMCV-D infection, and MDA5^−/− mice die from severe myocarditis before they can develop diabetes. However, MDA5^+/− mice, which survive EMCV-D infection, do develop transient hyperglycemia, suggesting a role for MDA5 in preventing virus-induced diabetes [215]. Altogether, these data indicate that IFN-I responses mediated by TLR3 and MDA5 during EMCV-D infection influence the pathogenesis of T1D. As LGP2 positively regulates RLR-mediated responses to picornaviruses, such as EMCV [154, 155], it will be worthwhile to investigate whether mice deficient in LGP2 are more susceptible to virus-induced diabetes. In summary, IFN-I responses may trigger the activation of autoreactive T cells in autoimmune T1D; however, during infection with a pancreatic-tropic virus, IFN-I responses reduce viral replication, thereby preventing β-cell damage and diabetes (Fig. 3).

Figure 3. Roles of MDA5 and TLR3 in virus-induced diabetes.

MDA5 has a dominant role in protecting the heart against EMCV-D infection. TLR3-deficient mice develop diabetes as a result of hematopoietic cells, failing to mount an early IFN-I response that protects β-cells from virus-induced damage. The role of MDA5 in virus-induced diabetes could not be assessed, as MDA5-deficient mice die from myocarditis within the first 5 days of infection. However, studies in MDA5^+/− mice, which survive EMCV-D infection, do develop transient hyperglycemia, suggesting at least a partial for MDA5 in preventing virus-induced diabetes. Thus, IFN-I responses mediated by MDA5 and TLR3 can reduce viral replication, preventing β-cell destruction and diabetes.

CONCLUDING REMARKS

Virtually all cells in the body are capable of producing IFN-I. As pDCs are more specialized than other cell types, it was thought that they might be the predominant source of IFN-I during viral infections. However, increasing evidence suggests that in vivo, pDC contribution to antiviral response is limited in magnitude and time and that multiple cellular sources contribute to protective or in the case of autoimmunity, detrimental responses mediated by IFN-I (Fig. 1). Recently, it was shown that subcapsular sinus macrophages produce IFN-I during peripheral VSV infection and prevent CNS invasion [216]. Splenic marginal zone and metallophilic macrophages are important for IFN-I production during systemic HSV-1 [217] and LCMV [218] infections. Tissue resident macrophages in the liver control viral replication in an IFN-I-dependent matter in LCMV-infected mice, although they were not reported to produce IFN-I themselves [219], whereas alveolar macrophages and microglia are dominant sources of IFN-I during respiratory or brain viral infections, respectively [51, 151]. Inflammatory monocytes secrete IFN-I in a TLR2-dependent manner and are necessary for early restriction of vaccinia virus replication [220]. Splenic DCs produce IFN-I in response to adenoviruses [221] and MCMV [222]. It has also been shown that splenic macrophages [223] and a subset of B cells [224] are major producers of IFN-I during Listeria infection and that splenic DCs secrete IFN-I in response to group B streptococcus [224]. Our studies indicate that stromal cells and DCs are critical sources of IFN-I during infection with EMCV [215]. Altogether, these findings indicate that several cellular sources of IFN-I exist in vivo and probably depend on the pathogen and site of infection (Fig. 1).

A multiplicity of viral sensors seems essential for effective IFN-I responses. Clearly, distinct sensors differ in their specificity for viruses and viral products. Furthermore, distinct sensors may differ in their tissue distribution, such that sensors required for responses to a given virus in one organ may vary in another. During EMCV-D infection, for example, MDA5 is critical in the heart, whereas MDA5 and TLR3 restrict virus-induced damage in the pancreas. To make matters more complex, organs and tissues might express different sensors at different time-points during infections. MDA5 is an IFN-induced protein, whereas TLR3 is constitutively expressed; therefore, TLR3-mediated responses might occur earlier than MDA5-mediated responses [215, 226].

Specificity, tissue distribution, and time-frame of action of viral sensors are all critical factors contributing to protective or detrimental immune responses. Understanding whether viral sensors have complementary and nonredundant or redundant roles in particular organs and cell types, whether they synergize or antagonize each other during host responses, will aid in the design of better vaccines and therapies for infections and autoimmune diseases.

AUTHORSHIP

M.S. contributed data and wrote the manuscript. S.A.M. and Y.W. contributed data and reviewed the manuscript. M.C. reviewed the manuscript.