Flaviviridae is a family of enveloped, positive-strand RNA viruses, many of which are spread by arthropod vectors (mainly ticks and mosquitoes). Within this viral family, the orthoflavivirus genus (hereafter referred to as flavivirus for simplicity) comprises arboviruses with a significant impact on public health (Smith, 2017; Pierson and Diamond, 2020), including Yellow fever virus (YFV), Dengue virus (DENV), Zika virus (ZIKV), Japanese encephalitis virus (JEV), and West Nile virus (WNV). These viruses are transmitted by the bite of mosquitoes, mainly of the genera Aedes and Culex (Figure 1). Once introduced in the skin, flaviviruses are capable of disseminating throughout the body and replicating in numerous organs. Flaviviruses can be classified according to the type of infection they cause in humans, which can be visceral (such as YFV, DENV, or ZIKV) and/or neurotropic (such as WNV, JEV, and ZIKV). In both cases, infections are often asymptomatic, but can cause severe and sometimes fatal symptoms, including hemorrhages, encephalitis, myelitis, paralysis or congenital malformations.

Given the recent and spectacular geographical expansion of these mosquitoes, in particular Aedes albopictus and Aedes aegypti, arboviruses have become a global health problem. Indeed, whereas they were restricted to tropical and sub-tropical regions until recently, flaviviruses now represent a threat also in temperate regions (Smith, 2017; Pierson and Diamond, 2020). Among them, Dengue is the most common and important arthropod-borne viral disease in humans. Its global incidence has grown dramatically in recent decades and all four dengue virus serotypes (DENV1-4) now threaten about half of the world population (Bhatt et al., 2013; Messina et al., 2019). Although DENV infection is often asymptomatic, it can cause a spectrum of illnesses, ranging from flu-like symptoms (Dengue fever) to the more severe and sometimes fatal Dengue hemorrhagic fever, which can progress to Dengue shock syndrome (Rigau-Pérez et al., 1998). According to the World Health Organization (WHO), the number of reported Dengue fever cases has increased from 500,000 cases in 2000 to more than 5 million in 2019 (source WHO, Dengue and severe dengue. Available at: https://www.who.int/news-room/fact-sheets/detail/dengue-and-severe-dengue. Accessed August, 2023).

ZIKV has also emerged as a major threat to human health. Indeed, while only sporadic infections were described in Africa and Asia from the 1960s until the 2000s, no major epidemics occurred until 2007, the year of the first outbreak of ZIKV-related disease on the Yap Island in Micronesia (Duffy et al., 2009). In the following years, two far greater epidemics occurred, the first in French Polynesia in 2013–2014 and the second which started in Brazil in 2015 and then spread to the rest of South America as well as in Central America in 2015–2016 (Zanluca et al., 2015; Faria et al., 2016). The epidemic had important impacts on the health of populations, coinciding with cases of Guillain-Barré syndrome in some adults and an unexpected epidemic of newborns with microcephaly and other neurological deficiencies, following in utero exposure to the virus (Mlakar et al., 2016). In February 2016, the WHO declared Zika-related microcephaly a Public Health Emergency of International Concern. This alert has now been lifted, but the virus remains under close surveillance by health authorities (Musso et al., 2019).

Unlike DENV and ZIKV, whose amplifying hosts are primates (including humans), WNV mainly infects birds, and humans are only incidental hosts (Campbell et al., 2002; McVey et al., 2015). Nevertheless, this emerging virus has caused major human epidemics over the last 20 years, including in New York in 1999, in Israel in 2000 and in Greece in 2010 (Jia et al., 1999; Bin et al., 2001; Hayes, 2001; Danis et al., 2011). The global emergence of WNV is perfectly illustrated by its recent rapid expansion in the United States. Indeed, since its sudden emergence in New York in 1999, the virus has quickly spread across the United States, resulting in over 56,000 reported cases and more than 2,700 deaths [source CDC, Historic Data (1999–2022). Available at: https://www.cdc.gov/westnile/statsmaps/historic-data.html. Accessed August, 2023]. It is currently considered one of the zoonotic diseases of greatest concern for the US population (Ronca et al., 2021). WNV belongs to the Japanese encephalitis virus (JEV) serocomplex (Chan et al., 2022). Both WNV and JEV are neurotropic viruses that can cause severe neurological symptoms, including encephalitis, meningitis or acute flaccid paralysis. However, while WNV circulates throughout most of the world, JEV is predominantly found in Asia, where it represents the leading cause of viral encephalitis (Le Flohic et al., 2013; Caldwell et al., 2022).

YFV is found in tropical and subtropical areas of Africa and South America (Chen and Wilson, 2020; Tuells et al., 2022). In these regions, yellow fever (YF) remains a widespread threat despite a very efficient, safe and affordable live-attenuated vaccine (Chen and Wilson, 2020; Tuells et al., 2022). The clinical spectrum of YF ranges from mild non-specific viral symptoms to a severe clinical course culminating in liver failure, renal failure, cardiovascular instability, which can be fatal. Most cases of YF are reported in Africa, and a modeling study based on African data sources estimated that in 2013, yellow fever caused 84,000–170,000 severe cases and 29,000–60,000 deaths (Garske et al., 2014).

In addition to these main human pathogens, many other flaviviruses transmitted by mosquitoes are responsible for sporadic cases or local epidemics, including the Usutu virus (USUV), Murray Valley encephalitis virus (MVEV), St. Louis encephalitis virus (SLEV) or Spondweni virus (SPOV) (Figure 1).

The factors that govern the severity of infections are still not completely understood, but both viral and host factors exist (Fernandez-Garcia et al., 2009; Pierson and Diamond, 2020; van Leur et al., 2021). Among host factors, antiviral innate immunity is likely to play a central role, since it is the first line of defense against viral infections and, depending on its capacity to control viral replication in the early stages of infection, it allows or not viruses to disseminate within the whole body (Fernandez-Garcia et al., 2009; van Leur et al., 2021). As a result, flaviviruses have developed multiple strategies to interfere with innate immunity, and in particular with the interferon (IFN) response, the armed wing of antiviral defenses.

The first descriptions of flavivirus-encoded proteins that inhibit the IFN response date from the early 2000s (Muñoz-Jordán et al., 2003, 2005; Lin et al., 2004; Guo et al., 2005; Jones et al., 2005; Liu et al., 2005), and subsequent research aimed at elucidating the multiple mechanisms involved has been intense. This topic has already been the subject of excellent reviews, including (Gack and Diamond, 2016; Cumberworth et al., 2017; Miorin et al., 2017). In this review, we aim to provide an update on the state of the art, given the constant progression of this field. Specifically, we summarize the state of knowledge on how infected cells trigger the IFN response following flavivirus infection and how these viruses manage to inhibit every single step of this response, focusing on mosquito-borne pathogenic viruses. We detail common and virus-specific strategies that allow flaviviruses to escape detection, block signaling pathways leading to IFN synthesis, and also prevent IFN signaling, in order to escape its powerful antiviral activity and thus be able to propagate.

Flaviviruses are small enveloped viruses, whose genome consists of a 10 to 11 kb positive-sense single-stranded RNA, which contains a single open reading frame (ORF), flanked by 5′ and 3′ non-coding regions. This ORF encodes a single polyprotein which is co- and post-translationally processed by host and viral proteases into 10 proteins: 3 structural proteins, the capsid (Cap), a precursor of the membrane protein (prM) and the envelope glycoprotein (Env), as well as 7 non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5). The Cap proteins constitute the icosahedral capsid which contains the viral genome, while the prM and Env proteins are incorporated into the membrane of the virions. The NS1 protein is a non-structural glycoprotein, important for the replication of flaviviruses. It exists as dimers, which are associated to membranes and as hexamers, which are secreted by the infected cells (Rastogi et al., 2016). NS2B is a transmembrane protein and serves as a cofactor of the NS3 protein, the viral protease, which is involved in the proteolytic cleavage of the polyprotein. While NS3 carries a serine protease activity in its N-terminal domain, its C-terminal domain possesses 5′-RNA triphosphatase (RTPase), nucleoside triphosphatase (NTPase), and helicase activities. NS5 is the viral RNA-dependent RNA polymerase. It also contains a methyltransferase activity in its N-terminal part, necessary for the methylation of the 5′ cap of the viral RNA (Ray et al., 2006). NS2A, NS4A, and NS4B are transmembrane proteins anchored to the endoplasmic reticulum (ER). Together with the other NS proteins, they participate in viral RNA replication and also in the evasion of host immune response.

Following the bite of an infected female mosquito, the skin is the first organ to be infected. Flaviviruses can replicate in fibroblasts, keratinocytes, and also in resident immune cells of the skin (Garcia et al., 2017). If flaviviruses can infect so many different cell types, it is because they are able to use many receptors to enter cells, including phosphatidylserine receptors and C-type lectins (Laureti et al., 2018). Following the interaction of envelope glycoproteins with their receptors, most flaviviruses penetrate their target cells by clathrin-dependent endocytosis (Figure 2). Once in late endosomes, the drop in pH induces a change in the conformation of the Env proteins, allowing the fusion of the viral and endosomal membranes (Mukhopadhyay et al., 2005) (Figure 2). Following its uncoating in the cytoplasm, the viral RNA is translated by ribosomes into polyproteins, which will then be cleaved by cellular and viral proteases (Rice et al., 1985) (NS2B/3).

Viruses replicate in invaginations of the ER membrane (Welsch et al., 2009; Den Boon and Ahlquist, 2010; Gillespie et al., 2010) (Figure 2). Within these virus-induced compartments, non-structural proteins form replication complexes, in which viral RNA is replicated. Viral assembly takes place at the ER membrane, close to sites of viral RNA replication (Welsch et al., 2009; Gillespie et al., 2010). These new immature virions will bud in the lumen of the endoplasmic reticulum then transit to the Golgi complex. It is in this compartment that furin cleaves the pr domain of the prM protein, allowing the release of mature infectious virions from the cell (Mukhopadhyay et al., 2005) (Figure 2).

The body’s first line of defense against viral infections is provided by the IFN response. IFNs are a family of antiviral cytokines that have been discovered more than 60 years ago (Isaacs and Lindenmann, 1957). They have no intrinsic antiviral properties, but act through the induction of hundreds of interferon-stimulated genes (ISGs), whose products confer the cells a so-called antiviral state. IFNs are typically divided into three classes: Type I (mainly IFN-α and -β), type II (IFN-γ), and type III (IFN-λs) (Schneider et al., 2014). While the direct antiviral effects of type II IFN are limited, type I and III IFNs induce a potent antiviral state within target cells, but with a different spectrum of action (Schneider et al., 2014). Indeed, while almost all nucleated cells are able to respond to IFN type I, the response to IFN type III is limited to epithelial cells and certain immune cells (Sommereyns et al., 2008). For this reason, we will only discuss the relationship between flaviviruses and IFN type I in this review.

This antiviral defense mechanism proceeds in two waves, namely the induction of type I IFN, triggered by viral infection, and the type I IFN signaling pathway, which leads to the synthesis of ISGs. These two steps of the IFN response are detailed below.

Endosomal TLRs, and in particular TLR3, participate in the detection of flaviviruses in certain cell types, as well as in the physiopathology of infection, but their involvement in the induction of the IFN response is relatively modest. Indeed, while the ability of TLR3 to trigger type I IFN synthesis has been demonstrated in certain cell models following infection by DENV (Tsai et al., 2009; Nasirudeen et al., 2011), the results obtained with other flaviviruses, in particular ZIKV or WNV, are less clear (Fredericksen and Gale, 2006; Hamel et al., 2015). It has been shown in a mouse model that the absence of TLR3 does not influence the replication of WNV in peripheral tissues, nor the IFN-α/β response, but leads to a higher mortality of mice, probably because of a protective role of TLR3 in the central nervous system (Daffis et al., 2008). A study has recently shown that, in human neural progenitor cells, TLR3 is involved in the synthesis of pro-inflammatory cytokines following ZIKV infection, but not in the synthesis of type I IFN (Plociennikowska et al., 2021). Furthermore, the authors showed that TLR3 activation inhibits RIG-I-induced type I IFN production (Plociennikowska et al., 2021). TLR3 therefore seems to be involved in the proinflammatory response and the physiopathology of flavivirus infections, but not or only modestly in the IFN response. Flaviviruses can also be sensed by TLR7 within endosomes, but this only occurs in plasmacytoid dendritic cells (pDCs) (Wang et al., 2006; Sun et al., 2009; Assil et al., 2019).

Unlike TLRs, numerous studies have demonstrated the predominant role of RIG-I and MDA-5 in the recognition of flavivirus RNAs and IFN response upon infection in most cells (Kato et al., 2006; Fredericksen et al., 2008; Loo et al., 2008). These two PRRs recognize double-stranded viral RNA in the cytoplasm and are partially redundant. In the case of flavivirus infections, they act in a coordinated manner, with an IFN response in two stages, first involving RIG-I, then MDA5 (Fredericksen and Gale, 2006; Fredericksen et al., 2008; Loo et al., 2008).

Besides the RNA sensors TLR3, RIG-I, and MDA5, the cGAS-STING axis, triggered by the presence of DNA in the cytoplasm, has also been shown to be involved in flavivirus infections (Schoggins et al., 2014). For instance, DENV infection causes mitochondrial membrane disruptions, leading to the leakage of mitochondrial DNA into the cytoplasm, thereby triggering the cGAS-STING pathway (Aguirre et al., 2017; Sun B. et al., 2017). As a consequence, cGAS^−/− mice are more susceptible to flavivirus infection (Schoggins et al., 2014; Zheng et al., 2018).

Once the synthesis of IFN-induced expression of ISGs is engaged, cells become virtually uninfectable. This is the reason why viruses engage in a race against time to replicate before this IFN-induced antiviral state hinders their dissemination (Katze et al., 2002). To ensure victory, viruses have developed various strategies to prevent the synthesis of IFN, to limit the induction of ISGs or to inhibit the activities of antiviral mediators. In this context, flaviviruses master the art and manner of inhibiting every single step of the innate immune response, in order to delay, if not prevent, the establishment of an antiviral response (Katze et al., 2002).

All tested flaviviruses proved capable of blocking the JAK/STAT signaling pathway, in order to prevent the synthesis of ISGs, in particular DENV (Muñoz-Jordán et al., 2003), ZIKV (Kumar et al., 2016), WNV (Guo et al., 2005), JEV (Lin et al., 2004), and YFV (Laurent-Rolle et al., 2014). Again, many non-structural proteins have been implicated in this inhibition, but NS5 appears to be the most effective viral antagonist. However, while all flaviviruses seem to inhibit type I IFN signaling through NS5, the mechanisms involved are virus-specific (Best, 2017).

The NS5 protein of DENV was shown to target STAT2 and to induce its degradation by the proteasome (Jones et al., 2005; Ashour et al., 2009; Mazzon et al., 2009), by recruiting the E3 ubiquitin ligase UBR4 (Morrison et al., 2013) (Figure 4).

ZIKV NS5 also binds to STAT2 and induces its proteosomal degradation, but, unlike DENV, it does not require UBR4, suggesting that another cellular protein may be involved (Grant et al., 2016; Kumar et al., 2016) (Figure 4). Whether it depends on other cellular proteins or not remains to be addressed. However and interestingly, it was recently reported that ZIKV sfRNAs potentiated the inhibitory effect of NS5, by binding to the viral protein and stabilizing it (Slonchak et al., 2022).

Surprisingly, the NS5 protein encoded by SPOV, a virus that is phylogenetically close to ZIKV, does not interact with STAT2 nor interfere with its phosphorylation (Grant et al., 2016). Given that SPOV NS5 is nonetheless capable of inhibiting type I IFN signaling, these observations suggest that it may act downstream of the pathway (Grant et al., 2016).

In the case of YFV again, NS5 was found to interact with STAT2, but only when cells have been previously stimulated with type I IFN (Laurent-Rolle et al., 2014). Indeed, this interaction was found to require the IFN-induced phosphorylation of STAT1 and the TRIM23-induced K63-linked polyubiquitination of NS5 (Laurent-Rolle et al., 2014) (Figure 4). As for ZIKV, this interaction is UBR4-independent.

JEV NS5 has also been shown to inhibit the JAK/STAT pathway, presumably by preventing Tyk2 tyrosine phosphorylation (Lin et al., 2004, 2006; Yang et al., 2013) (Figure 4).

Finally, WNV NS5 does not target STAT2 nor Tyk2, but the type I IFN receptor subunit IFNAR1 (Laurent-Rolle et al., 2010; Lubick et al., 2015), thus explaining why IFNAR1 is depleted in WNV-infected cells (Evans et al., 2011). Specifically, NS5 inhibits the expression of IFNAR1 by recruiting prolidase (PEPD), a cellular peptidase (Lubick et al., 2015). According to this study, PEPD would be involved in IFNAR1 biosynthesis, by facilitating its trafficking through the ER-to-golgi network (Lubick et al., 2015) (Figure 4).

NS5 seems therefore to play a predominant role for flaviviruses in order to counteract the antiviral effect of type I IFN. However, it is striking to note such a convergence of NS5 function in several flaviviruses, but with such a heterogeneity of mechanisms. Besides NS5, other flavivirus-encoded proteins could also participate in the inhibition of type I IFN signaling.

This is the case of NS4B which, in several flaviviruses, including DENV, YFV, ZIKV, and WNV, interferes with the phosphorylation of STAT1 and its nuclear translocation (Liu et al., 2005; Muñoz-Jordán et al., 2005; Fanunza et al., 2021b) (Figure 4). The involvement of the NS2B/3 protease complex of ZIKV has also been proposed, since its ectopic expression leads to the degradation of JAK1 (Wu et al., 2017). Finally, when overexpressed in HEK293T cells, ZIKV NS2A induces the degradation of STAT1 and STAT2 in a proteasome-dependent manner, suggesting an anti-IFN activity of this protein (Fanunza et al., 2021a). However, once again, these observations will have to be confirmed, particularly in the context of infection.

As discussed above, many studies demonstrate that flaviviruses possess an important arsenal, consisting of sfRNAs, non-structural and possibly even structural proteins to prevent type I IFN synthesis and signaling. Importantly, the ability of flaviviruses to interfere with the IFN response has not only been observed in vitro, but also in infected patients. For example, RNA sequencing-based transcriptional profiling studies performed on cells isolated from ZIKV-infected patients revealed that the transcription of type I IFNs and ISGs was unaffected in infected patients, thus demonstrating the capacity of ZIKV to inhibit type I IFN production and type I IFN signaling in vivo (Sun X. et al., 2017; Carlin et al., 2018; Hu et al., 2019).

A large number of studies have also shown that the ability to efficiently antagonize the IFN response confers an important selective advantage to flaviviruses. There is therefore a correlation between the IFN resistance of a given virus and its replication efficiency, pathogenicity and/or epidemiological fitness (Keller et al., 2006; Manokaran et al., 2015; Xia et al., 2018; Castro-Jiménez et al., 2022).

However, although flaviviruses can inhibit IFN-induced responses, type I IFN still restricts viral replication in vitro, but only when added prior to infection, since it is much less effective once the infection is established (Diamond et al., 2000; Anderson and Rahal, 2002; Crance et al., 2003; Lin et al., 2004; Ho et al., 2005; Samuel and Diamond, 2005). Type I IFN is also able to effectively inhibit flavivirus replication and spread in vivo. In this respect, mice lacking the type I IFN receptor (IFNAR^−/−) or key components of the IFN signaling pathway such as STAT1 and STAT2, show markedly enhanced lethality and viral replication when infected with DENV (Shresta et al., 2004; Ashour et al., 2010), ZIKV (Lazear et al., 2016; Tripathi et al., 2017), YFV (Meier et al., 2009; Erickson and Pfeiffer, 2013), WNV (Samuel and Diamond, 2005; Keller et al., 2006), USUV (Martín-Acebes et al., 2016) or MVEV (Lobigs et al., 2003). Interestingly, in these IFN response-deficient mice, increased infection is observed in normally resistant cell populations and tissues, suggesting that type I IFN acts in part to restrict viral tropism.

The capacity of type I IFN to restrict flavivirus infection has also been confirmed in therapeutic disease models, since pretreatment with IFN-α or inducers of IFN-α has been shown to attenuate viral dissemination and to improve clinical outcome in mice or hamsters infected by several viruses (Brooks and Phillpotts, 1999; Leyssen et al., 2003; Morrey et al., 2004; Julander et al., 2007; Chan et al., 2016).

However, despite these promising results in preclinical models, the therapeutic efficacy of type I IFN in patients has been disappointing. During a severe DENV epidemic in Cuba in 1981, IFN-α was administered to patients (Limonta Vidal et al., 1984). Although some clinical improvement was noted, no other trials have been reported and it is therefore difficult to draw conclusions from this single study. Another study reported the case of two patients with severe Japanese encephalitis, out of a group of four, who showed improvement in clinical signs and recovered from the infection after treatment with IFN-α, whereas the two patients who did not receive IFN died (Harinasuta et al., 1985). However, a randomized double-blind placebo-controlled trial performed on 112 Vietnamese children with suspected Japanese encephalitis concluded that IFN-α did not improve the outcome of patients (Solomon et al., 2003).

While the use of type I IFN in the treatment of flavivirus infections is therefore an unlikely prospect, other therapeutic strategies could exploit its antiviral potency. For example, an inhibitor of YFV NS4B was recently described as able to suppress viral replication as well as enhance IFN-β expression in infected cells (Gao et al., 2022). Therefore, compounds that counteract flavivirus-induced inhibition of the IFN response might be an interesting perspective.

Higher eukaryotes have developed sophisticated mechanisms to prevent the spread of viruses in the body. Type I IFN is the armed wing of these defenses. Extremely fast and efficient, the IFN response can abolish the replication of any virus, at least in theory. Indeed, most viruses have developed strategies to avoid detection, to inhibit IFN synthesis by infected cells or to prevent the establishment of an IFN-induced antiviral state in neighboring cells. As detailed in this review, flaviviruses are experts in terms of countermeasures, as they are able to interfere with every single step of the IFN response. The fact that all the proteins encoded by these viruses are involved in hijacking cellular defenses perfectly illustrates the incredible selection pressure that these defenses exert on viruses in general, and flaviviruses in particular. Strikingly, even the viral RNA is implicated, since flaviviruses have developed a particularly original mechanism involving non-coding subgenomic RNAs to prevent their detection. Our understanding of all these strategies has considerably progressed over the past decade, even if many gray areas remain. For instance, some degree of uncertainty remains for the many observations that were made by overexpressing individual flavivirus proteins, and which need to be confirmed under conditions of infection. Furthermore, while the involvement of non-structural viral proteins in the inhibition of the IFN response is very well documented, recent descriptions of a potential role for structural proteins will need to be confirmed by further studies (Airo et al., 2022; Sui et al., 2023). Finally, it is still not fully understood how the balance between the IFN response and viral antagonists influences the dissemination of flaviviruses in the body, and in particular how IFN manages to control viral replication in certain organs but not in others.

Elucidating the molecular mechanisms that allow flaviviruses to bypass cellular defenses in order to spread is an important issue, especially for the development of future vaccines and antivirals.

JZ: Writing – original draft, Writing – review & editing, Visualization. SN: Writing – original draft, Writing – review & editing, Conceptualization, Funding acquisition.