Evasion and disruption of innate immune signalling by hepatitis C and West Nile viruses

Summary

Signalling pathways leading to type I interferon production are the first line of defence employed by the host to combat viruses, and represent a barrier that an invading virus must overcome in order to establish infection. In this review we highlight the ability of two members of the Flaviviridae, a globally distributed family of RNA viruses that represent a significant public health concern, to disrupt and evade these defences. Hepatitis C virus is a hepatotropic virus, infecting greater than 170 million people worldwide, while West Nile virus is a neurotropic virus that causes encephalitis in humans and horses. While these viruses cause distinct disease phenotypes, the ability of pathogenic strains to modulate the innate immune response is a key factor in influencing disease outcome. Both viruses have evolved unique strategies to target various aspects of type I interferon induction and signalling in order to prevent viral clearance and to promote virus replication.

Background

The Flaviviridae are a diverse, globally distributed family of viruses that cause a wide range of diseases in human and other animals. The Flaviviridae are made up of three major genra: Pestivirus, Hepacivirus and Flavivirus. Pestiviruses are a major problem for the agricultural industry due to infections among cattle and pigs, causing reproductive losses and severe diseases among livestock (Gubler, 2007). On the other hand, Hepaciviruses, such as hepatitis C virus (HCV), and Flaviviruses, such as West Nile virus (WNV), represent significant public health concerns for humans. HCV and WNV are the leading causes of virus-induced liver disease and mosquito-borne encephalitis respectively in the United States (Lauer and Walker, 2001; CDC, 2008).

HCV is a hepatotropic virus, infecting greater than 170 million people worldwide. Within the United States, it is estimated that approximately 2% of adults are infected with HCV, while worldwide infection rates range from less than 1% to as high as 20% (Lauer and Walker, 2001). HCV is a blood-borne pathogen, and risk factors for transmission include blood transfusions from infected donors, reuse of needles and syringes by healthcare workers or by intravenous drug users, tattoos and high-risk sexual behaviour (Lauer and Walker, 2001; Maheshwari et al., 2008). A majority of HCV infections become chronic, which in the worst cases result in cirrhosis of the liver, hepatocellular carcinoma and liver failure requiring transplant (Lauer and Walker, 2001). The incidence of new HCV infections in the US peaked in the 1980s and has since been on the decline (Maheshwari et al., 2008). However, persistent HCV infections from this period are contributing to an increasing incidence of HCV-related liver failure. Sequence analysis divides HCV into six genotypes varying in geographic distribution and response to the standard combined therapy of pegylated interferon (IFN)-α and ribavirin (Simmonds, 2004). Genotype 1 viruses are the most prevalent in North America, as well as the most resistant to treatment with only a 50% sustained virologic response (Feld and Hoofnagle, 2005). Currently, there is no approved vaccine or virus-specific therapy for HCV infection.

WNV, on the other hand, is a mosquito-borne, neurotropic virus that causes encephalitis among humans and horses (Gubler, 2007). WNV was originally isolated in the West Nile region of Uganda in 1937 from the blood of a human showing fever-like symptoms (Smithburn et al., 1940). WNV is endemic within areas of Europe, Asia, the Middle East, Africa and, more recently, North America, and is divided into two distinct lineages based on geographical distribution and severity of disease outcome in humans (Lanciotti et al., 1999; Lanciotti et al., 2002). Lineage I viruses are represented by emergent strains associated with outbreaks of encephalitis and meningitis in Europe, the Middle East and North America (Lanciotti et al., 1999). In contrast, lineage II viruses are less virulent in humans and largely represent non-emergent strains geographically isolated within Southern Africa and the island countries of Madagascar and Cyprus (Smithburn et al., 1940; Gubler, 2007). Lineage I virus infections in humans generally causes a febrile illness characterized by flu-like symptoms, muscle aches, joint pain, fatigue, maculopapular or roseolar rash, and nausea (Hayes et al., 2005; Davis et al., 2006). Occasionally, severe infections occur causing encephalitis, meningitis, meningioencephalitis and mortality (Lanciotti et al., 1999; Hayes et al., 2005). Among clinically diagnosed cases in the US, rates of neuroinvasion are between 30–50%, and of these cases, there is 5–10% mortality rate (CDC, 2008). Although the more severe symptoms are associated with infections in young children, the elderly, transplant patients or immunocompromized individuals, in recent years healthy young adults have been afflicted with neurological disease (Emig and Apple, 2004; Debiasi et al., 2005; Fischer, 2008). This indicates an increase in virulence that can occur independently of immune senescence or immune deficiencies. There is no vaccine or specific treatment options for WNV infection.

Virus and host factors contribute to the major differences found between WNV and HCV infections, including tissue tropism (multiple versus targeted), infection outcome (acute versus persistent) and host range (broad versus limited) respectively. While HCV and WNV cause distinct disease outcomes, there is strong correlation between disease severity and the ability of pathogenic strains to modulate the innate immune response. In this review, we highlight the major pathways involved in triggering innate immune defences during HCV and WNV infection and the countermeasures employed by these viruses to modulate these responses.

Type I IFN signalling pathways

Type I IFN and the actions of innate immune genes are the first line of defence employed by the host to combat virus infection and are responsible for deterring virus replication and spread. Triggering type I IFN induction during infection begins with host recognition of an invading pathogen through unique biological structures called pathogen-associated molecular patterns (PAMP). In mammals, the major pattern recognition receptor (PRR) pathways include the retinoic acid inducible gene-I (RIG-I)-like receptor (RLR) and toll-like receptor (TLR) signalling pathways. In the case of RNA virus infections, viral nucleic acid structures and motifs, recognized by RLRs or TLRs, are the major stimulators of type I IFN induction.

RIG-I like receptors

The RLRs are cytoplasmic sensors of viral RNA, consisting of three members: RIG-I, melanoma differentiation antigen 5 (MDA5) and laboratory of genetics and physiology-2 (LGP2) (Kang et al., 2002; Yoneyama et al., 2004; 2005; Sumpter et al., 2005; Saito and Gale, 2008). The molecular structures of RIG-I and MDA5 are similar, consisting of two tandem caspase association and recruitment domains (CARD) in the N-terminus and an RNA helicase domain in the C-terminus (Yoneyama et al., 2005). In addition, RIG-I has a repressor domain that interacts with the CARD domains to maintain the receptor in a closed and non-active conformation in the absence of infection (Saito et al., 2006). LGP2 contains an RNA helicase domain but lacks the CARD domains and may function as a regulator of RLR signalling (Saito et al., 2006; Venkataraman et al., 2007). RIG-I and MDA5 recognize unique structures and sequence motifs of viral RNA (Kato et al., 2008; Saito et al., 2008). Single-stranded (ss)RNA containing a 5′-triphosphate, short double-stranded (ds)RNA and uridine- or adenosine-rich viral RNA motifs have been identified as RIG-I ligands (Hornung et al., 2006; Saito et al., 2008; Takahasi et al., 2008). The minimum RNA length for RIG-I recognition is 21 nucleotides, while MDA5 is believed to recognize long dsRNA (> 5 kb) (Marques et al., 2006; Kato et al., 2008). RIG-I and MDA5 become activated after binding RNA ligand and in turn bind to IFN-β promoter stimulator-1 (IPS-1; also known as Cardif, MAVS and VISA) (Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005). IPS-1 is a CARD protein and an essential adaptor of RLR signalling, wherein it functions to bind RIG-I or MDA5 through CARD–CARD interaction that promotes the activity of a poorly defined macromolecular complex consisting of signalling components that result in the activation of the transcription factors interferon regulatory factor-3 (IRF-3) and nuclear factor-κB (NF-κB). RLR signalling through IPS-1 leads to assembly of the IFN-β enhanceosome complex within the nucleus, consisting of IRF-3, NF-κB, CBP/p300 and ATF-2/c-Jun, and triggering induction of IFN-β (Johnson and Gale., 2005).

Recent studies have defined novel signalling pathways and a variety of innate immune signalling partners of IPS-1 that contribute to the induction of IFN and innate defences during virus infection. These include, but are not limited to, DDX3, an RNA helicase that links IRF-3 activation to the TBK1 protein kinase (Schroder et al., 2008), TRIM22, a signalling adaptor protein that recruits ubiquitin ligase to modify the activity of the IPS-1 signalling complex (Gack et al., 2007), and the TRAF proteins, which participate in IPS-1-mediated NF-κB activation (Guo and Cheng, 2007). Additionally, the recent discovery and characterization of STING, an endoplasmic reticulum signalling adaptor protein, provides evidence that virus infection triggers signalling cross-talk between endoplasmic reticulum and mitochondrial components of innate antiviral immunity (Ishikawa and Barber, 2008). Thus, the overall antiviral processes that are triggered in the host cell during acute virus infection are complex and highly integrated. While the virus-specific signalling events of this complex pathway integration have not been defined, it is likely that innate defence signalling is tailored to specific viral pathogens in order to elicit a host response geared to suppress the specific virus challenge.

TLR signalling

The TLR family of innate immune signalling proteins consist of more than 10 different members, of which three recognize unique ribonucleic acid structures and motifs (Akira et al., 2006). TLR3, TLR7 and TLR8 are important for host defence against RNA viruses. TLR3 recognizes dsRNA ligands, whereas TLR7 and TLR8 recognize uridine and guanosine-rich ssRNA ligands (Alexopoulou et al., 2001; Diebold et al., 2004; Heil et al., 2004). TLR3 is expressed either on the cell surface or within the endosomal compartment (Matsumoto et al., 2003). Upon binding dsRNA TLR3 recruits the adaptor molecule TIR domain containing adaptor-inducing IFN-β (TRIF) and activates the transcription factors IRF-3 and NF-κB to trigger IFN-β transcription (Akira et al., 2006). TLR7 and TLR8 are expressed within the endosomal compartment (Heil et al., 2004). Upon binding ssRNA ligands, the adaptor molecule myeloid differentiation factor 88 (MyD88) is recruited to the TLR and the transcription factors IRF-7 and NF-κB are activated to trigger pro-inflammatory cytokines and type I IFN induction (Akira et al., 2006).

Type I IFN signalling

Induction of an antiviral state within a cell results from activation of JAK/STAT signalling leading to expression of interferon-stimulated gene (ISG) products (Fig. 1). Secreted type I IFN (IFN-α and IFN-β) bind in an autocrine or paracrine manner to cells through the IFN-alpha receptors 1 and 2 to initiate the JAK/STAT signalling pathway. Once IFN is bound to its cognate receptors, the receptor-associated Janus kinases (Jaks) Jak1 and tyrosine kinase 2 (Tyk2) are activated and phosphorylate key tyrosine residues on the cytoplasmic subunit of the receptor to allow for the recruitment and phosphorylation of the signal transducer and activator of transcription 1 and 2 (STAT1 and STAT2). STAT1 and STAT2 heterodimerize and associate with interferon regulatory factor 9 to form the ISGF3 transcription factor complex. This complex translocates to the nucleus and binds to specific regions of DNA known as interferon-stimulated response elements, and turns on transcription of ISGs. Amplification of the type I IFN response occurs through a positive feedback loop involving IFN-induced expression of IRF-7 followed by IRF-7-dependent transcription of various IFN-α species (Sen, 2001; Horvath, 2004).

Figure 1. Mechanisms of HCV and WNV innate immune evasion. 1, sequestering WNV RNA from RLR detection; 2, WNV NS1 blockade of TLR3 signalling; HCV NS3/4A protease cleavage of TRIF; 3, HCV NS5A binding/interference of MyD88; 4, HCV NS3/4A protease cleavage of IPS-1; 5, WNV blockade of Tyk2 activation, HCV core-induced SOCS-3 expression and subsequent blockade of STAT1 phosphorylation.

Flavivirus triggering of type I IFN

Hepatitis C virus

The positive-sense ssRNA genome of HCV is approximately 9.6 kb in length with conserved 5′- and 3′-untranslated regions (UTR). The 5′ end of the viral genome contains a triphosphate group and an internal ribosomal entry site consisting of several unique stem-loop structures allowing for cap-independent translation (Moradpour et al., 2007). The 3′-UTR contains unique stem-loop structures that are required for initiating minus-strand synthesis and believed to be important for RNA stability.

It is well established that HCV infection activates innate immune signalling through recognition of HCV RNA by RIG-I. Initial studies by Sumpter et al. (2005) defined a human liver carcinoma cell line (Huh7.5) that showed enhanced permissiveness for HCV RNA replication. This observation did not appear to be an HCV-specific enhancement to viral RNA replication, but rather a defect in innate immune control. Genetic and biochemical studies traced the defect to a single-amino-acid mutation (T55I) in the N-terminal CARD domain of RIG-I. This mutation turns RIG-I into a dominant-negative protein, thus abolishing its signalling. IFN-β induction by HCV RNA in Huh7.5 cells could be rescued by ectopic expression of a wild-type form of RIG-I (Sumpter et al., 2005). In order to identify the region of the HCV genome recognized by RIG-I, Saito et al. (2008) performed an extensive mapping study and found that RIG-I efficiently recognized an unstructured polyU/UC-rich region within the 3′-UTR. Furthermore, the replication intermediate of this region consists of a polyA/AG, which also activated RIG-I signalling, suggesting that poly adenosine and poly uridine regions are substrates of RIG-I. Genomics analysis of several additional RIG-I-dependent RNA viruses identified adenosine- and uridine-rich regions and that corresponded to potent RIG-I substrates. In this same study, the 5′-triphosphate was found to be essential, but not sufficient, for RIG-I binding to viral ssRNA. The RIG-I dependency of the HCV PAMP RNA was confirmed through studies that showed loss of signalling when the RNA was transfected into RIG-I^−/− cells. On the other hand, no effect on the potency of the HCV PAMP RNA to trigger IFN-β induction was observed in MDA5^−/−, MyD88^−/− and TRIF^−/− cells, demonstrating that MDA5 and TLR signalling were not required for triggering type I IFN in response to the HCV PAMP RNA (Saito et al., 2008). The importance of RIG-I in detecting HCV is underscored by the fact that transient IRF-3 activation can be observed early during infection of wild-type cells, but in clonal cells defective in RIG-I signalling, IRF-3 activation is abolished and the virus replicates to higher titers at earlier times during infection (Lohmann et al., 2003; Cheng et al., 2006; Loo et al., 2006; Dansako et al., 2007).

While these results have demonstrated that RIG-I is the major PRR for HCV detection, other studies have shown that TLRs may contribute to detection of HCV. Dolganiuc et al. (2004) have shown that purified HCV core and NS3 proteins can be detected by human monocytes in a TLR2-dependent manner leading to pro-inflammatory cytokine production. This detection was subsequently shown to involve TLR1 and TLR6 (Chang et al. 2007). However, the in vivo implications of this finding are not well understood. TLR3 may also detect HCV RNA as the virus encodes a mechanism to evade TLR3 signalling (discussed below), but HCV recognition via TLR3 has not yet been demonstrated.

West Nile virus

The positive-sense ssRNA WNV genome, approximately 11 kb in length, contains a 5′-methylated cap and no poly A tail. The use of an excellent murine pathogenesis model has defined key signalling components in triggering type I IFN induction during WNV infection (Samuel and Diamond, 2006). Initial studies defined a key role for IRF-3 and found that infection by both lineage I and lineage II WNV strains activated IRF-3 signalling (Fredericksen et al., 2004; Fredericksen and Gale, 2006; Keller et al., 2006). Infection of IRF-3^−/− or IRF-7^−/− mice with a pathogenic lineage I strain of WNV demonstrated that IRF-3 and IRF-7 were essential for protection (Daffis et al., 2007; 2008a). Analysis of IRF-3^−/− or IRF-7^−/− mice revealed that primary neuronal cells infected with WNV followed the canonical IRF-3 and IRF-7 pathways to induce IFN-β and IFN-α expression respectively (Daffis et al., 2007; 2008a). Intriguingly, IRF-7^−/−, but not IRF-3^−/−, BM-DCs, BM-Macs and mouse embryo fibroblasts infected with WNV blunted IFN gene expression, suggesting that IRF-3 and IRF-7 can induce IFN-β expression or that an undefined signalling component/pathway is involved in IFN-β induction during WNV infection (Daffis et al., 2007; 2008a). Most importantly, these results underscore the importance of key host transcription factors during WNV infection.

RLR, and to a lesser extent TLR, signalling pathways have been shown to play important roles in triggering IFN-β induction during WNV infection. Infection of RIG-I or MDA5 knockout mouse embryo fibroblasts with WNV showed that RIG-I and MDA5 were essential in triggering a robust type I IFN response (Fredericksen and Gale, 2006; Fredericksen et al., 2008). In these studies, RIG-I was found to be essential for triggering IFN-β induction and downstream ISG expression early during WNV infection, while MDA5 was important for conferring robust type I IFN induction and ISG expression profiles. Infection of IPS-1^−/− mice or primary cells with WNV has revealed the essential nature of the RLR signalling pathway in protection against WNV infection and triggering IFN-β (Fredericksen et al., 2008; M.S. Suthar and M. Gale, unpublished).

The role of TLRs, in particular TLR3, during WNV infection has been controversial. Studies by Wang et al. (2004) found that TLR3^−/− mice were more resistant to a lineage I WNV infection and that this was due, in part, to impaired cytokine production and reduction in blood–brain barrier permeability. A second study performed by Daffis et al. (2008b) found that TLR3 was important for protection against WNV infection. In this case TLR3^−/− mice showed a slight increase in virulence and enhanced tissue tropism.

Taken together, these studies imply that the RLR signalling pathway plays an essential role in triggering IFN production and ISG expression during WNV infection, and that TLR signalling may impart additional, RLR-independent defences that regulate immunity against WNV infection.

Modulation of innate immune signalling

Evasion of IFN-β induction

A major strategy employed by HCV to subvert the host innate immune response is by disrupting the RIG-I signalling pathway through HCV NS3/4A protease-mediated cleavage of IPS-1. The NS3/4A protein complex is essential for virus replication and polyprotein processing (Moradpour et al., 2007). The NS3 protein contains a helicase and protease domain, while the NS4A protein is a small, 54-amino-acid cofactor that intercalates into NS3-modulating protease specificity (Bartenschlager et al., 1995). The importance of NS3/4A in disrupting antiviral innate immunity was first identified in a study by Foy et al. (2003), which showed that ectopic expression of NS3/4A could block Sendai virus-induced phosphorylation, dimerization and nuclear translocation of IRF-3. This blockade was dependent on the protease activity of NS3/4A and could be reversed by treatment of cells with a pharmacological inhibitor of the protease function of NS3/4A or by mutating the active site within the NS3/4A protease domain (Foy et al., 2003). Subsequently, it was shown that NS3/4A acted downstream of RIG-I as IFN-β induction was blocked by coexpression of NS3/4A with a constitutively active form of RIG-I (N-RIG) (Foy et al., 2005), but upstream of the IRF-3-activating kinases TBK1 and IKKε, as the activities of these signalling molecules were not affected by coexpression with NS3/4A (Breiman et al., 2005). When IPS-1 was identified as the adaptor molecule bridging RIG-I to these downstream signalling components, it was also found to be the target of NS3/4A-mediated blockade. NS3/4A cleaves IPS-1 at amino acid position 508 immediately proximal to the mitochondrial membrane anchor, resulting in redistribution away from the mitochondria, and abolishing RLR-mediated signalling (Li et al., 2005a; Lin et al., 2006; Loo et al., 2006). Cleavage of IPS-1 during HCV infection occurs between 24 and 48 h post infection in tissue culture and correlates with accumulation of HCV RNA and protein, and IPS-1 cleavage has been demonstrated in liver tissues from chronically infected patients (Loo et al., 2006). Additional studies to define the minimal components of NS3/4A required for the blockade found that a single-chain protease, consisting of the NS3 protease domain covalently linked to NS4A, was sufficient for IPS-1 cleavage (Johnson et al., 2007). Interestingly, the closest relative to HCV, GB virus B, also mediates cleavage of IPS-1 by NS3/4A (Chen et al., 2007). As HCV and GB viruses cause persistent infection in humans, this suggests the importance of disrupting IPS-1 signalling through viral protease actions in establishing a persistent infection.

In addition to cleavage of IPS-1, HCV has been shown to evade type I IFN induction through a variety of alternative strategies. NS3/4A has been shown to cleave the TLR3 adaptor molecule TRIF, thus blocking TLR3-dependent signalling in vitro, although this has not yet been demonstrated in vivo (Li et al., 2005b). Others have shown that HCV NS3, NS3/4A, NS4B and NS5A inhibit MyD88-dependent TLR signalling in a murine macrophage cell line (Abe et al., 2007). This study found that NS5A directly binds MyD88 and inhibits the recruitment of interleukin-1 receptor-associated kinase 1 to MyD88, thus blocking MyD88-dependent signalling. The NS5A–MyD88 interaction was found to be mediated by the interferon sensitivity-determining region located between amino acids 204–280 within NS5A (Abe et al., 2007). While these findings are intriguing, the importance of the TLRs in innate immune detection of HCV remains unclear. It is likely that TLR signalling plays a role in IFN induction during chronic HCV infection, as hepatic ISGs are clearly induced in HCV patients and experimental chimeric mouse models of HCV infection even though IPS-1 is cleaved and inactivated by the viral NS3/4A protease (Loo et al., 2006; Smith et al., 2006; Walters et al., 2006; Walters and Katze, 2009). However, the source of the hepatic IFN that drives this response has not been defined, and further studies are needed in order to understand the nature of the IFN response in HCV patient liver.

While HCV specifically targets RLR or TLR signalling components, WNV has been shown to evade triggering IFN-β induction through two distinct pathways: (i) masking viral RNA from cellular PRRs and (ii) NS1-mediated inhibition of TLR3 signalling. Fredericksen et al. found that IRF-3 activation was delayed relative to the accumulation of viral protein production, suggesting that WNV was either actively antagonizing or evading detection by the RLRs (Fredericksen and Gale, 2006). Additional studies found that WNV infection failed to actively antagonize IFN-β induction, as transfection with poly (I : C) or coinfection with vesicular stomatitis virus resulted in normal kinetics of IFN-β induction. This suggested that WNV employs a passive evasion strategy early during infection, possibly by masking or sequestering viral RNA from recognition by the RLRs. Delaying type I IFN induction would allow for WNV to establish infection within the cell prior to triggering the innate immune defence programmes (Fig. 1).

Results from other studies suggest that WNV can inhibit TLR3 signalling by preventing IRF-3 activation. Wilson et al. (2008) found that ectopic expression of the WNV NS1 protein, which is normally secreted from infected cells and involved in RNA replication, inhibited the activation of IRF-3 and NF-κB and prevented IFN-β promoter activation. However, the exact mechanism by which NS1 inhibits TLR3 signalling is not completely understood. Nonetheless, these studies show that WNV employs multiple mechanisms, similar to HCV, to evade triggering type I IFN production.

Modulation of type I IFN signalling

HCV and WNV target distinct aspects of the JAK/STAT signalling pathway to modulate the host response (Fig. 1). HCV is able to antagonize type I IFN signalling by inducing the expression of a negative regulator of JAK/STAT signalling. It was found that expression of the HCV core protein in the human hepatocellular liver carcinoma cell line, HepG2, induced the expression of suppressor of cytokine signalling (SOCS)-3 (Bode et al., 2003). SOCS-3 expression has been observed in liver biopsies of HCV-infected patients undergoing IFN therapy (Walsh et al., 2006). While SOCS-3 inhibits cytokine signalling by preventing the downstream activation of STAT proteins, the mechanism by which HCV core induces SOCS-3 expression has yet to be determined.

The relative importance of these activities for immune evasion and establishment of persistent HCV infection is not completely understood. A robust symptomatic acute infection is associated with greater likelihood of spontaneous clearance and potentially a result of a robust innate immune response (Maheshwari et al., 2008). Among persistently HCV-infected patients, responders and non-responders to IFN show different ISG expression profiles (Chen et al., 2005). This may represent variability in the induction of SOCS-3 expression by different HCV core sequences, variability in effective cleavage of IPS-1 by NS3/4A, or other variations in host defence regulation.

Several groups have reported that WNV infection modulates JAK/STAT signalling as a means of preventing expression of innate immune effector genes (Guo et al., 2005; Liu et al., 2005; Munoz-Jordan et al., 2005; Keller et al., 2006). Inhibition of JAK/STAT signalling is believed to be a feature unique to pathogenic lineage I isolates while non-pathogenic lineage 2 isolates are unable to inhibit this pathway (Keller et al., 2006). Furthermore, blockade of the JAK/STAT signalling pathway is believed to occur upstream of the receptor-associated Tyk2 protein kinase but not Jak1 protein kinase and inhibition correlates with the accumulation of viral proteins.

The WNV factors underlying regulation of JAK/STAT signalling are still not well defined. Several groups have implicated the WNV non-structural proteins in attenuating type I IFN signalling (Liu et al., 2005; Munoz-Jordan et al., 2005; Evans and Seeger, 2007). However, many of these studies were based on the use of replicon cell lines lacking the viral structural proteins or by ectopic expression of viral proteins independent of the context of virus infection. Evans et al. through the use of a WNV replicon cell line, identified a single mutation within NS4B that was essential for inhibiting type I IFN signalling. However, mutational analysis of this determinant in the context of WNV infection found that this same mutation did not influence IFN signalling, suggesting the presence of additional viral determinants that regulate JAK/STAT signalling (Evans and Seeger, 2007). Therefore, further studies are warranted to define the viral protein and/or host factors responsible for imposing a blockade on JAK/STAT signalling.

Conclusion

While significant advancements have been made in understanding the virus–host interactions that control HCV and WNV infection, further studies are required to understand the complete mechanisms of viral evasion of innate immune defences. Several questions remain in defining the host pathways and viral and factors that regulate early RLR recognition and type I IFN signalling and are critical for our understanding of host defence against these important pathogens. What are the relative contributions of the IPS-1-dependent and independent programmes to host defence? What is the cross-talk between these defence programmes? How does STING contribute to antiviral actions against HCV and the Flaviviruses? Addressing these questions will facilitate discoveries that could provide a platform for therapeutic and vaccine design to combat WNV emergence and to enhance current therapy for HCV infection.

Acknowledgements

We thank Drs Brian P. Doehle and Takeshi Saito for critical reading and discussions. The Gale laboratory is supported by funds from the State of Washington, National Institutes of Health (NIH; Grants AI1057568, AI060389, DA024563 and AI40035), The Burroughs Wellcome Fund, National Institute of Allergy and Infectious Diseases (NIAID) training Grant T32AI007411 and a gift from Mr and Mrs R. Batcheldor. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or NIAID.