VIRUSES
Viruses enter the host cell to hijack the machinery necessary for viral replication. Viral PAMPs can trigger innate immune responses either on the host cell surface, within the endosome, or in the cytoplasm.
Emerging evidence also suggests that innate immune detection might occur in the nucleus. Inflammasome sensors are activated in response to both DNA and RNA viruses, ensuring efficient immunosurveillance and elimination of different families of viruses. In this section, we discuss the molecular mechanisms regulating virus-induced activation of the inflammasome.
DNA Viruses
DNA viruses replicate in either the cytoplasm or the nucleus, and the majority activate the inflammasome through innate immune recognition of viral DNA (Table 4). Cytoplasmic DNA is sensed by the inflammasome sensor AIM2; however, not all DNA viruses activate AIM2 (discussed further below).
Murine cytomegalovirus (MCMV) and vaccinia virus trigger activation of the AIM2 inflammasome in mouse bone marrowderived macrophages (BMDMs) or bone marrow-derived dendritic cells (BMDCs) (31, 162, 164), suggesting that their viral DNA activates the AIM2 inflammasome (Fig. 3). In addition, mice lacking AIM2 and infected with MCMV have lower levels of circulating IL-18 and higher viral titers after 36 h than those of infected WT mice (164).
FIG 3 Viral activators of the inflammasome. DNA and RNA viruses carry viral PAMPs and induce physiological aberrations and host cell damage that can trigger activation of the inflammasome. DNA from double-stranded DNA (dsDNA) viruses can enter the cytoplasm and the nucleus to engage activation of the AIM2 inflammasome and the putative IFI16 inflammasome, respectively. Both DNA and RNA viruses can cause physiological aberrations in the form of cathepsin B release, production of ROS, and potassium efflux, which are signaling cues leading to activation of the NLRP3 inflammasome. RNA and viral proteins of single-stranded RNA (ssRNA) viruses and replication intermediaries of a single-stranded DNA (ssDNA) virus have been shown to trigger activation of the NLRP3 inflammasome. Host DNA release from virus-induced damage can lead to activation of the AIM2 inflammasome. There is some evidence to suggest that an ssRNA virus might activate the NLRP1 inflammasome. Double-stranded RNA (dsRNA) from a dsRNA virus has been shown to activate a putative NLRP9b inflammasome in mouse enterocytes.
Similar to its murine counterpart, human cytomegalovirus (HCMV) also activates the AIM2 inflammasome in human THP-1 monocytic cells (193). Unlike the case in bacterium-induced activation of the AIM2 inflammasome, type I IFN signaling and IFN-inducible proteins are not required to facilitate sensing of MCMV by AIM2 (100, 162).
Interestingly, HCMV expresses the tegument protein pUL83, which interacts with and inhibits AIM2 (194) and the related ALR member IFI16 (195–197), indicating the importance of DNA inflammasome sensors for detection of this viral pathogen. Other DNA viruses, including hepatitis B virus (HBV), human papillomavirus 16 (HPV16), Epstein-Barr virus (EBV), and modified vaccinia virus Ankara (MVA), can activate the AIM2 inflammasome (198–201) (Table 4). HBV infection induces AIM2-mediated caspase-1 activation or IL-1 and IL-18 secretion in human glomerular mesangial cells, hepatocytes, and PBMCs (198, 202, 203). Similar observations have been made in human keratinocytes infected with HPV (199) and human THP-1 monocytes infected with EBV (200).
Transfected EBV DNA induces IL-1 secretion in the human tumor cell line HK1 via the AIM2 inflammasome (204). Despite these observations, the molecular mechanism underpinning activation of the AIM2 inflammasome by DNA viruses has not completely emerged. The putative activator of AIM2 in these cases is presumably the genomic DNA and/or replication intermediaries of these viruses. However, biochemical and imaging studies are required to confirm whether viral DNA binds to AIM2 in the cytoplasm and/or the nucleus during viral infection. Furthermore, not all DNA viruses activate the AIM2 inflammasome (Table 4). The DNA viruses adenovirus, herpes simplex virus 1 (HSV-1), MVA, varicella-zoster virus (VZV), and human bocavirus 1 all preferentially activate the NLRP3 inflammasome over AIM2 (164, 205–211). The main activator of the NLRP3 inflammasome induced by DNA viruses is unknown. However, adenovirus infection leads to disruption of lysosomal membranes, which results in the release of cathepsin B and activation of NLRP3 (210, 212, 213). In addition, adenovirus infection induces the production of ROS in THP-1 cells, which also contributes to NLRP3 inflammasome activation (208, 210, 213).
Similarly, HSV-1 infection induces expression of NLRP3, cleavage of caspase-1, and maturation of IL-1 in human foreskin fibroblasts and human THP-1 cells (205, 208). Secretion of IL-1 in response to HSV-1 was also found to be independent of AIM2 in mouse thioglycolate-elicited macrophages (164). Together these examples highlight that although DNA viruses possess DNA, some preferentially activate the NLRP3 inflammasome over AIM2. The reasons underlying why certain DNA viruses do not engage activation of the AIM2 inflammasome are unknown. Given that AIM2 binds DNA in a sequence independent manner (169), it is unlikely that DNA sequences from these viruses are not recognized by AIM2. A more plausible explanation is that these viruses encode inhibitory proteins similar to the host-derived AIM2-inhibiting PYD-containing proteins that can directly bind and inhibit AIM2 (169, 214). Viral DNA binding proteins may also prevent access of AIM2 to the DNA and thereby prevent inflammasome activation. Alternatively, it is possible that rapid shuttling of viral DNA into the host nucleus evades detection by AIM2 in the cytoplasm. In the latter case, other nuclear DNA sensors might be able to compensate the innate immune recognition of the virus. Indeed, the DNA of HSV-1 has been shown to activate IFI16 in the nucleus (Fig. 3) (205, 215). It is also possible that, as shown for human monocytes, the cGAS-STING axis may drive NLRP3 inflammasome activation in response to cytosolic DNA (168). This would support findings of past studies suggesting that DNA viruses activate the NLRP3 inflammasome rather than the AIM2 inflammasome in human monocyte-derived cells but cannot explain why AIM2 is not activated in other cell types (168, 205, 208, 210, 211). Taken together, the data show that DNA viruses can activate either the AIM2 or NLRP3 inflammasome (Fig. 3); however, the precise activators and regulatory mechanisms leading to activation of these inflammasomes are still not entirely clear.
RNA Viruses
RNA viruses can harbor either a single- or double-stranded RNA genome. An increasing number of RNA viruses have been found to activate the inflammasome (Table 5), including those which pose major threats to public health. Below, we focus our discussions primarily on influenza virus and human immunodeficiency virus (HIV) and on how infection by each of these viruses leads to activation of the inflammasome (Fig. 3). Influenza A virus. Influenza A virus is an important human pathogen that has caused multiple pandemics in human history. The virus is subtyped based on its surface glycoproteins hemagglutinin (H) and neuraminidase (N) (216). There is strong evidence to demonstrate that multiple subtypes of influenza A virus trigger activation of the NLRP3 inflammasome in macrophages and dendritic cells (217–224). Several mechanisms have been proposed to explain how influenza virus infection leads to NLRP3 activation. Both RNA and protein components of influenza virus trigger activation of the NLRP3 inflammasome (Fig. 3) (219–221, 223, 225). Transfection of RNA derived from the A/H7N9 strain activates the NLRP3 inflammasome in BMDCs (219, 225), implying that viral RNA is sufficient to activate the NLRP3 inflammasome. However, the viral peptide PB1-F2, derived from strains A/H7N9 and A/H1N1, can also activate NLRP3 in BMDMs (220, 221).
Further studies have identified additional mechanisms by which influenza virus can activate the NLRP3 inflammasome. For instance, influenza virus strain A/H3N2 encodes an ion channel, M2, causing ion imbalances during infection of mouse BMDMs and BMDCs and thereby activates NLRP3 (223). Mice lacking NLRP3 and administered the viral peptide PB1-F2 of strain A/H7N9 produce less IL-1 in the lungs than that in WT mice (221), suggesting that this viral peptide mediates activation of the NLRP3 inflammasome in vivo. Together these observations show that it is possible that both viral RNA and peptides are the PAMPs responsible for activating the NLRP3 inflammasome during influenza virus infection. Studies have also suggested that an RNA-protein complex comprised of the influenza A viral RNA and viral proteins NP and PB1 is recognized by the innate immune sensor ZBP1 (also known as DAI or DLM-1) (226–228). Recognition of this RNA-protein complex by ZBP1 drives activation of the NLRP3 inflammasome and multiple forms of programmed cell death in a manner dependent on the kinases RIPK1 and RIPK3 and on caspase-8 (226). How the ZBP1–RIPK1–RIPK3– caspase-8 scaffold mediates activation of the NLRP3 inflammasome is unknown. Nevertheless, these studies suggest that activation of the NLRP3 inflammasome by influenza virus might require an upstream sensor that recognizes viral PAMPs. Together these studies indicate that in the context of infection of a host, multiple pathogen detection pathways and mechanisms of inflammasome activation are likely to be utilized.
Given the importance of NLRP3 in sensing infection, influenza virus encodes a nonstructural protein, NS1, that is capable of binding to and inhibiting activation of the NLRP3 inflammasome (229, 230). Specifically, mutating residues in the RNA- and TRIM25-binding domains of NS1 prevented immunoprecipitation of NS1 with NLRP3 and increased secretion of IL-1 relative to that with WT NS1 (229), indicating that these regions are important for mediating NLRP3 binding and inhibition. While the NLRP3 inflammasome responds to influenza virus infection, it remains unclear whether this pathway confers protection against or susceptibility to the virus in vivo. A study has shown that mice lacking NLRP3 and infected with influenza virus strain A/H1N1 (also known as A/PR/8/34) harbor higher viral titers in their lungs and have a higher mortality rate than those of WT mice (217, 218). Contrastingly, another study found that mice lacking NLRP3 and infected with the highly pathogenic influenza virus strain A/H7N9 have increased survival and develop less pulmonary inflammation than that in WT mice (219). These differing results may be explained, in part, by studies which report that temporal inhibition of NLRP3 by use of the small-molecule compound MCC950 influences the survival rate of mice infected with the influenza virus strains A/WSN/1933 (231) and A/PR/8/34 (232).
The latter study found that delayed treatment with MCC950 increases survival of mice, whereas treatment with MCC950 as early as 1 day after infection results in lethality (232). These results partially reconcile the earlier conflicting results and suggest that the NLRP3 inflammasome may contribute to antiviral host defense at the early stages of infection while driving overt inflammation and immunopathology at later stages. Conflicting results also exist regarding the role of AIM2 in influenza virus infection. Influenza A virus infection of lung epithelial cells can cause substantial damage, resulting in the release of host DNA that can activate the AIM2 inflammasome (233, 234). Whether AIM2 inflammasome activation benefits the host during influenza virus infection remains uncertain. While a study reported that mice lacking AIM2 and infected with influenza virus strain A/H1N1 have less lung injury and a higher survival rate than those of infected WT mice (233), another study found that mice lacking AIM2 are more susceptible than infected WT mice to infection with A/H1N1 (234). These conflicting results may be due to differences in the source of the A/H1N1 virus or the infection dose used or to subtle differences in mouse genetic background, microbiota, or the vivarium used to house these animals (235). However, the exciting finding that both influenza virus-derived PAMPs and the DAMPs generated as a result of infection can trigger activation of two inflammasome complexes, NLRP3 and AIM2, clearly highlights the functional overlap between inflammasome sensors in host defense against this clinically important virus. HIV. Since the discovery and initial isolation of HIV 30 years ago, approximately 75 million people have been infected, and more than 37 million currently live with the infection worldwide (236). Polymorphisms in the genes encoding NLRP3 and IL-1 are associated with susceptibility to HIV infection (237, 238). HIV infection triggers activation of the NLRP3 inflammasome in a variety of cell types, including human monocytes (237, 239–241), human and rodent microglia (242–244), and human podocytes (245). While the mechanism of NLRP3 inflammasome activation by HIV has yet to be fully defined, evidence suggests that ROS, cathepsin B, and virus-induced potassium efflux from the host cell might all contribute to activation of NLRP3 (240, 242, 245). Furthermore, transfection of HIV-derived single-stranded RNA (ssRNA) is sufficient to activate caspase-1 and induce secretion of IL-1 (240), implying that viral RNA may activate the NLRP3 inflammasome. There is also some evidence to suggest that the HIV proteins R and Tat can activate NLRP3 in human and rodent microglia, respectively (243, 244). However, further biochemical and mechanistic studies are required to more fully assess which viral PAMPs activate the NLRP3 inflammasome and whether direct or indirect sensing of PAMPs is involved. Further studies have suggested that HIV infection can trigger pyroptosis in a manner that is independent of the NLRP3 inflammasome in tissue-derived CD4 T cells but not in blood-derived CD4 T cells (246–249). In this scenario, the incomplete reverse transcripts of HIV are recognized by IFI16 in CD4 T cells of the lymphoid tissue, which subsequently drives the infected cells toward pyroptosis (246, 248, 249). The levels of IFI16 expression in tissue- and blood-derived CD4 T cells might provide an explanation for this observation; indeed, qPCR analysis revealed that tissue-derived CD4 T cells have three times the level of IFI16 transcripts in blood-derived CD4 T cells (246). Further studies revealed that the virological synapse—an intercellular junction between an infected cell and an uninfected cell that is generated by HIV—mediates viral transmission and pyroptosis between neighboring lymphoid CD4 T cells, ultimately accelerating the demise of the CD4 T cell population and the progression to AIDS (248–250). Although pyroptosis of CD4 T cells is detrimental to the host, there is some evidence that anti-HIV therapies are linked to inflammasome activation. For example, the nucleoside analog reverse transcriptase inhibitor abacavir, which is used as part of anti-HIV therapies, induces mitochondrial damage and subsequent activation of the NLRP3 inflammasome in human monocytes (251). Similarly, nelfinavir, an inhibitor of the HIV aspartyl protease, is used in the treatment of HIV infection and potentially activates the AIM2 inflammasome owing to the ability of the drug to rupture the nuclear envelope (252). In this way, activation of inflammasome pathways by anti-HIV therapies potentially contributes to their anti-HIV effects, although further research is needed to substantiate this hypothesis. Indeed, it has been suggested that inflammasome activation by abacavir contributes to severe delayed-type drug hypersensitivity (251), an off-target effect sometimes experienced by patients. The inflammasome pathway therefore may be beneficial in the detection of HIV infection but appears to also have detrimental outcomes in causing CD4 T cell pyroptosis and some of the side effects associated with anti-HIV therapies.
Other RNA viruses.
Other RNA viruses, including enterovirus type 71 (EV71), hepatitis C virus (HCV), Japanese encephalitis virus (JEV), respiratory syncytial virus (RSV), rhinovirus, encephalomyocarditis virus (EMCV), measles virus, parainfluenza virus type 3, vesicular stomatitis virus (VSV), Newcastle disease virus (NDV), rabies virus, Sendai virus, dengue virus, and Zika virus, have been shown to trigger activation of the NLRP3 inflammasome (Table 5). EV71 has been reported to be recognized by both the NLRP3 and AIM2 inflammasomes, depending on the cell types studied (253–255). The RNAdependent RNA polymerase (RdRp) 3D protein of EV71 has been suggested to directly bind and activate the NLRP3 inflammasome in human THP-1 cells, rhabdomyosarcoma cells, and HEK293T cells (254). In a similar manner, the RdRp NS5 protein produced by Zika virus directly binds and activates NLRP3 (256), implying that NLRP3 recognition of viral RdRp proteins may facilitate inflammasome activation in response to ssRNA viruses more generally. It will be interesting to assess whether RdRp proteins expressed by other ssRNA viral pathogens can likewise bind to and activate NLRP3. While EV71 infection also appears to induce activation of the AIM2 inflammasome in neuronal cells (255), the activator in this case is unknown. The EV71 proteases 2A and 3C specifically cleave NLRP3 and prevent it from forming an inflammasome complex in infected cells (253). Viral RNA and the viral ion channel proteins of RNA viruses have been suggested to activate the inflammasome. For example, the viral RNA of reovirus (182), the viral RNA and viral protein viroporin P7 of HCV (257–259), the viral RNA and viroporin SH of RSV (182, 226, 260), and the viroporin 2B proteins of EMCV, EV71, poliovirus, and rhinovirus (261, 262) all result in activation of the NLRP3 inflammasome. A study reported that the RNA of rotavirus triggers a poorly characterized NLR, NLRP9b, in intestinal organoids (Fig. 3) (263). The rotavirus RNA interacts with the host RNA helicase DHX9, which then binds NLRP9b and licenses activation of the inflammasome (263). Moreover, mice lacking NLRP9b and infected with rotavirus have an increased viral load in enterocytes, increased fecal antigen shedding, the occurrence of diarrhea, and more severe pathology than that in WT mice (263). The discovery of a new putative inflammasome sensor in enterocytes argues that many more sensors might have inflammasome-forming capacity and are expressed in a cell-type-specific manner (264). It is possible that distinct inflammasome complexes are formed in different cell types in response to different viral factors generated by the same virus. An exciting avenue of research would be to investigate how these viral factors are liberated and presented to inflammasome sensors in immune and nonimmune cells to provide an integrated antiviral response in the host.