Inflammatory-dependent Sting activation induces antiviral autophagy to limit zika virus in the Drosophila brain

ABSTRACT

Incidences of congenital syndrome associated with maternal zika virus (ZIKV) infection during pregnancy are well documented; however, the cellular and molecular mechanisms by which ZIKV infection causes these devastating fetal pathologies are still under active investigation. ZIKV is a member of the flavivirus family and is mainly transmitted to human hosts through Aedes mosquito vectors. However, in vivo models for the neurological tropism of the virus and the arthropod vector have been lacking. A recent study published in Cell Host & Microbe from Dr. Sara Cherry’s lab investigates both of these key aspects of the ZIKV infectious life cycle. Liu et al. demonstrate how inflammatory activated Sting/dSTING-dependent antiviral macroautophagy/autophagy is sufficient to restrict ZIKV infection in the Drosophila melanogaster brain. Additionally, this study provides further evidence for the ancestral function of autophagy in protecting host cells from viral invaders.

Abbreviations: AGO2: Argonaute 2; ATG: autophagy-related; Dcr-2: Dicer-2; DptA/Dipt: Diptericin A; Drs: Drosomycin; DCV: Drosophila C virus; IMD: immune-deficiency; qRT-PCR: quantitative real-time PCR; Rel/NF-κB: Relish; RNAi: RNA interference; ZIKV: zika virus

In recent years, zika virus (ZIKV) infection has emerged as a significant public health threat, particularly among pregnant women. Maternal ZIKV infection is associated with poor infant outcomes, including miscarriage, stillbirth, intrauterine growth restriction and microcephaly [1]. The stage of pregnancy at which the mother acquires ZIKV infection also appears to be important, with earlier gestational stages associated with more detrimental congenital pathology [2]. In otherwise healthy adults, ZIKV infection has also been implicated in the development of the autoimmune disorder Guillan-Barré [3]. Not surprisingly, the severity of disease associated with ZIKV infection has prompted investigation into vaccine development [3,4]

ZIKV is a single-stranded positive-sense RNA virus and member of the Flaviviridae family, which includes West Nile, yellow fever, dengue and hepatitis C viruses. Recent work identified the induction of host autophagy in response to ZIKV infection [5–7], indicating that autophagy may function in a proviral capacity to promote replication of this virus. Lennemann and Coyne have demonstrated that ZIKV (and other flaviviruses) target and cleave the reticulophagy receptor RETREG1/FAM134B through NS2B3 virally encoded proteases [8]. A study by Cao and colleagues suggested that inhibition of autophagy might limit vertical transmission of ZIKV in a pregnant mouse model [9]. Studies on ZIKV pathogenesis have focused on human placental trophoblasts [10], mice [1,9,11–13] and non-human primates [14]. Yet, much remains unknown regarding the mechanisms associated with ZIKV infection in neurons and in insect hosts. The relationship between viruses and host autophagy remains provocative, with the pathway exerting dual roles in both promoting and suppressing infection; this is likely dependent on the virus, cell type, stage and context of the infection. Here we focus on the antiviral role of autophagy (often referred to as xenophagy or virophagy). Autophagy may function in an antiviral capacity by limiting viral replication [15–17] and/or interacting with innate immune components [18,19].

Autophagy is an ancient mechanism that is ubiquitous across eukaryotes and essential for cell survival during starvation and other stress conditions. Many of the 42 autophagy-related (ATG) components that have thus far been identified in fungi have homologs or analogous counterparts that are highly conserved in more complex eukaryotes. In fact, the emergence of autophagy appears to be fundamental in the development of eukaryotic life [20]. A recent study by Liu and colleagues report the development of a fly model to examine innate immune responses to ZIKV infection in Drosophila melanogaster [21]; Drosophila is a powerful and well-established model organism for the study of innate immunity.

To establish a Drosophila model of ZIKV pathogenesis, Liu et al. monitored ZIKV infection by quantitative real-time PCR (qRT-PCR), and were able to observe increased levels of ZIKV. A time course of infection established that ZIKV propagated to the greatest extent in the fly brain with evidence established by multiple assays, including qRT-PCR, TCID50 and confocal microscopy. The authors then investigated antiviral pathways that could potentially limit ZIKV infection. The RNA interference (RNAi) pathway limits virus replication in insects [22], and Liu and colleagues first examined whether components of RNAi could be involved in restricting ZIKV infection in Drosophila. Using mutant flies deficient in the key RNAi components AGO2 (Argonaute 2) and Dcr-2 (Dicer-2), the authors did not find evidence for antiviral RNAi in ZIKV replication.

Liu et al. next explored whether the antimicrobial Toll or IMD (immune deficiency) inflammatory Rel/NF-κB (Relish) pathways were activated in response to ZIKV infection. The Toll and IMD pathways are the 2 major arms of Drosophila innate immune defense [23]. The authors noted the upregulation of the IMD pathway target gene DptA/Dipt (Diptericin A) but not the Toll target Drs (Drosomycin) in whole flies and heads by qRT-PCR, supporting the involvement of the IMD but not the Toll arm of the innate immune system. Activation of DptA is dependent on Rel, and a loss-of-function allele no longer induces DptA in a ZIKV-dependent manner. ZIKV infection is upregulated in the heads of mutant Rel flies when compared to sibling controls as measured by both qRT-PCR and TCID50 assays, providing further evidence that the IMD arm of the Rel/NF-κB pathway is an important suppressor of ZIKV. Moreover, ZIKV challenge results in increased lethality in Rel mutants. Next, the authors investigated whether neurons or glia (the 2 major cell types in the brain) accounted for the IMD-mediated signaling and viral repression in response to ZIKV infection. Depletion of Rel in a neuronal- or glial-specific manner increases ZIKV infection in these cell types, supporting their importance in ZIKV pathogenesis. The data presented support a role for IMD pathway-induced Rel/NF-κB-mediated restriction of ZIKV in the fly brain.

Transcriptional profiling was then used to explore Rel-dependent genes involved in suppressing virus infection with the model virus Drosophila C (DCV); this analysis identified Sting/dSTING as a Rel-dependent gene transcriptionally activated in response to virus infection. TMEM173/STING (transmembrane protein 173) is a highly conserved pathogen cyclic dinucleotide sensor that activates antimicrobial innate immune signaling pathways in mammals [21,24]. Interestingly, prior to this report, an antiviral role for Sting in Drosophila had not yet been described, and TMEM173/STING had not yet been implicated in antiviral autophagy. The authors then investigated whether Sting was induced in response to ZIKV challenge by qRT-PCR analysis and found that expression is significantly upregulated in fly heads. Furthermore, Liu et al. determined that Sting induction is reduced in Rel-depleted fly mutants, supporting the idea that Sting activation is Rel-dependent. In Sting^EP/EP mutant flies with a P-element insertion allele of Sting (EY06491), the ZIKV viral loads are higher in heads and whole flies (but not bodies alone). Additional analysis demonstrated that neuronal-specific depletion of Sting leads to increased ZIKV replication. Taken together, these data indicate that ZIKV is restricted through Rel/NF-κB-induced Sting.

Recent reports indicate a role for TMEM173/STING in antibacterial autophagy [25,26]. Liu and colleagues next explored whether ZIKV-dependent activation of Sting induced autophagy in flies. By western blotting, the authors observed upregulated Atg8-II in the heads of flies infected with ZIKV, supporting a role for enhanced autophagy with ZIKV infection in the brain. Atg8 associates with the expanding phagophore and the autophagosome, and thus serves as a marker of autophagy induction. Atg8 exists in 2 forms—a non-lipidated soluble species and a lipidated phosphatidylethanolamine (PE)-conjugated form referred to as Atg8-II (or Atg8–PE). Monitoring the lipidation status of Atg8 by western blot is an assay for autophagy induction. Additionally, the authors confirmed their findings by evaluating the localization of Atg8 in both uninfected and ZIKV-infected fly brains using mCherry-Atg8 through confocal microscopy. In ZIKV-infected flies, mCherry-Atg8 relocalizes to distinct puncta, indicating autophagy induction; whereas the mCherry-Atg8 signal in uninfected flies remains diffuse. These data indicate ZIKV induces autophagy in the fly brain in response to infection.

Liu et al. then examined whether autophagy is the mechanism whereby ZIKV is restricted in the Drosophila brain. Using heat shock-inducible silencing of genes encoding core autophagy components, Atg5 and Atg7, the authors found that ZIKV levels increase in whole flies and heads when assayed by qRT-PCR, and titers increase in Atg5 mutants. When the Drosophila autophagy cargo receptor ref(2)p/SQSTM1 is also depleted, ZIKV infection is bolstered. Additionally, when neuronal- or glial-specific loss of Atg5 occurs, ZIKV replication increases. Furthermore, ZIKV infection is enhanced in Atg5 null flies, leading to significant lethality over an extended time course of infection. Taken together, these data reveal that autophagy is a key mechanism for restricting ZIKV pathogenesis and promoting survival.

The authors propose a model in which ZIKV infection is restricted in fly brains through Rel-dependent Sting transcriptional induction. To test this, Rel fly mutants were infected with ZIKV and assayed for autophagy induction by monitoring Atg8 lipidation. As expected Atg8-II was upregulated in WT but not Rel mutant flies, supporting the idea that auto-phagy induction is dependent on Rel. Liu and colleagues then tested the requirement for Sting in autophagy activation. Atg8-II accumulates in the heads of ZIKV-infected Sting heterozygous flies; however, Atg8 lipidation is abolished in Sting mutants. Flies fed the autophagy-inducing agent rapamycin demonstrate increased protection from ZIKV. Altogether, these data support a model for the restriction of ZIKV replication in fly brains through autophagy, which functions downstream of Sting and Rel through the IMD arm of the Rel/NF-κB pathway.

This work demonstrates how a single virus can exhibit diverse tissue tropism throughout its pathogenesis, which is accompanied by differential effects on autophagy-dependent viral replication. Autophagy is activated in response to ZIKV infection (reviewed in ref. 27). In mammalian cell types (including fetal stem cells, epithelial cells and fibroblasts), autophagy functions in a proviral capacity to enhance ZIKV replication (reviewed in ref. 21); whereas here Liu and colleagues demonstrate a clear antiviral role with regard to ZIKV in the Drosophila brain. The role of autophagy in virus infection has remained under debate. This leaves us with the following question: What are the diverse intracellular factors that mediate whether autophagy functions in either a proviral or antiviral capacity for ZIKV in each of these different cell types? Additionally, it is also possible that there may be unique virus-host factor interactions that are dependent on the specific ZIKV isolate under study.

One possible explanation comes from the conservation of TMEM173/STING and its evolutionary significance in antimicrobial autophagy. The TMEM173/STING domain necessary for innate immune activation of IRF3 arose during vertebrate evolution [21,24]; it is therefore absent in the Drosophila ortholog Sting. Whether autophagy first evolved as a response to starvation, pathogens or overall stress management is unknown. However, Liu et al. provide further evidence here [21] for the role of autophagy as an ancient antimicrobial pathway that predates the evolution of interferon-mediated immunity.

Funding Statement

This work was supported by the National Institute of General Medical Sciences [GM053396].

Acknowledgments

The authors apologize to those whose work was not included here due to space limitations, and thank Dr. Sara Cherry for helpful comments.

Disclosure statement

No potential conflict of interest was reported by the authors.