Background
The type I interferon (IFN) response is the first line of defense against virus infections. IFNs are induced and secreted in infected cells upon pathogen recognition. Secreted type I IFN binds to the IFN-α/β receptors (IFNARs) in both an autocrine and a paracrine manner, activating the Janus kinase (JAK)/signal transducer and activator of transcription (STAT) pathway, which induces the expression of hundreds of interferon-stimulated genes (ISGs) [
1]. ISG responses vary according to specific pathogens, the cell type infected, and the immune status of the host, as well as the immune evasion capabilities of the pathogen. Given that neurons are largely nonrenewable, neurotropic viruses need to be cleared from the CNS without causing excessive immune-mediated damage. We have shown that the local IFN response in the brain is capable of protecting mice from low doses of flavivirus [
2,
3]. Astrocytes are a major source of IFNs, the secretion of which protects neighboring astrocytes as well as neurons very rapidly following infection [
4]. Recent studies have also shown that viral tropism within the CNS is shaped by the immune response within specific brain regions [
2,
5]. However, how innate antiviral responses in specific cell types and regions of the brain shape immune protection is not clear.
Viperin (virus inhibitory protein, endoplasmic reticulum-associated, interferon-inducible) is an ISG with broad-spectrum antiviral activity [
6]. Viperin expression is low at the basal state. However, upon viral recognition, it is highly induced in an IFN-dependent as well as in an IFN-independent manner [
7‐
9]. It has broad-range antiviral activity against influenza A virus [
10,
11], human immunodeficiency virus [
12], chikungunya virus [
13], Sindbis virus [
14], and respiratory syncytial virus [
15]. It is also active against members of the
Flaviviridae family such as hepatitis C virus (HCV), Zika virus (ZIKV), dengue virus (DENV), West Nile virus (WNV), and tick-borne encephalitis virus (TBEV) [
16‐
21]. Despite its broad antiviral activity in vitro, the antiviral activity of viperin in vivo has only been demonstrated in WNV, respiratory syncytial virus, and Chikungunya virus infection [
13,
15,
18]. One possible reason for this is that most viruses are able to block the IFN response in order to minimize ISG expression and the restriction of viral infection [
22‐
26]. Another explanation could be redundancy among ISGs, where multiple ISGs act in a coordinated manner to mount an antiviral response. Interestingly, although low basal levels of viperin are found in most cell types, we have found high levels of viperin in the brain, specifically in primary astrocyte cultures in vitro [
4]. They also show a rapid response to viral infection by upregulation of viperin expression. This could mean that viperin plays an important role in the antiviral defense of specific cell types. We have identified viperin as a potent inhibitor of TBEV [
19], Langat virus (LGTV), [
27] and ZIKV [
28]. Since these are neurotropic viruses, they represent useful models to investigate the antiviral capacity of viperin in the brain.
Here we have used LGTV as a model for tick-borne flavivirus infection in vivo. We found that viperin specifically restricted replication in vivo in the olfactory bulb and in the cerebrum, but not in the cerebellum. In primary cells isolated from the cerebrum, both neurons and astrocytes relied on viperin for their antiviral response. However, viperin failed to restrict highly pathogenic TBEV replication in the granule cell neurons from the cerebellum. Furthermore, we found that viperin was required for the IFN-mediated antiviral responses against flaviviruses in cortical neurons. In summary, our findings show that viperin acts as an inhibitor of flaviviruses in the brain with a region-specific role in the olfactory bulb and cerebrum and cell-specific role in cortical neurons.
Discussion
Viperin is a key player in the IFN-mediated antiviral response and inhibits a broad spectrum of viruses at various stages of their life cycles [
41,
42]. Previously, we have shown that ectopic expression of viperin strongly inhibits TBEV and ZIKV replication by inhibiting viral genome synthesis by degrading viral NS3 [
19,
28]. Here we used LGTV, as a surrogate model virus to determine the protective antiviral role of viperin in neurotropic flavivirus infection in vivo. We show that the antiviral effects of viperin are restricted to the specific brain regions, olfactory bulb and cerebrum, and this is probably mediated by their resident neurons. Finally, we also show that IFN-induced ISGs fail to compensate for the loss of viperin in the cortical neurons against various neurotropic flaviviruses.
Few studies have sought to define the function of viperin in vivo. For WNV, higher viral burdens were observed in the spleen, kidney, and brain in the absence of viperin, which was evident as viperin
−/− mice were more susceptible to infection [
18]. Although both WT and viperin
−/− mice survived LGTV infection, we detected higher early viral burden in the spleen, before the olfactory bulb and then cerebrum in viperin
−/− mice. Thus, the spleen could serve as an organ for early viral amplification. Viral replication and subsequent spread within the brain is an important concern in neurotropic viral infection as it results in subsequent morbidity and mortality. Our results show that viperin reduces LGTV dissemination into the brain and limits the subsequent replication in the cerebrum. Previous studies using vesicular stomatitis virus (VSV) have shown that astrocytes within the olfactory bulb produce an IFN response, which restricts VSV infection. It induces long-distance IFN signaling within the brain, which activates antiviral ISG expression in distinct brain parts that prevents virus spread within the brain [
43‐
45]. We found that although LGTV elicits an IFN response in the olfactory bulb of WT and viperin
−/− mice, LGTV disseminates into the cerebrum but not to the cerebellum in the absence of viperin. In addition, particularly cortical neurons rely on viperin to restrict viral growth in the presence of IFN, whereas granular cell neurons in the cerebellum can mediate an antiviral response in the absence of viperin. Our results indicate that the long-distance IFN signaling and antiviral response is blunted against LGTV in the absence of viperin.
There are conflicting studies regarding the role of viperin in interferon production. One study showed that viperin enhances TLR-7 and TLR-9 signaling and facilitates the nuclear translocation of IRF-7 to induce type I IFN response in plasmacytoid dendritic cells (pDCs) [
39], while another study in macrophage cells showed that viperin acts as a negative regulator of an interferon response by interacting with interferon-beta promoter stimulator 1 (IPS-1) [
46]. We found that viperin was involved in the systemic IFN response after LGTV infection; similar findings were observed in WNV infection [
18]. In contrast, viperin was not involved in the IFN induction in the olfactory bulb. Most likely this reflects the cell-type-specific role of viperin in regulating the IFN response, since viperin is needed for the TLR-7/9 mediated induction in pDCs, and pDCs are known for their capacity to produce large amounts of type I IFN [
47,
48], whereas IFN induction in the olfactory bulb is strongly dependent on IPS-1 [
2].
Recent findings suggest differential immune responses within distinct regions of the brain that can modulate virus pathogenesis [
2,
38,
49]. In our system, we observed a region-specific role of viperin following ic LGTV infection with higher viral replication in the olfactory bulb and cerebrum of viperin
−/− mice. This was consistent with the observations that viperin inhibited LGTV in the olfactory bulb and cerebrum after peripheral viral replication. In WNV and VSV infection, an antiviral effect of IFIT2 has been observed in the olfactory bulb, cortex, cerebellum, and brain stem [
50,
51]. After WNV infection in Ifi27la
−/− mice, regional restriction was observed in the brain stem and cerebellum in the context of peripheral viral infection, while only in the brain stem post-ic infection [
5], while viperin was required to control WNV replication in cortex and white matter [
18], demonstrating distinct roles of ISGs, in different brain regions. This phenomenon might be due to the different subsets of neurons in different regions, as cortical neurons rely on viperin expression to limit TBEV replication while granule cell neurons isolated from cerebellum do not. Both epigenetic- as well as microRNA-mediated mechanisms might be involved in the regulation of ISGs in different subsets of neurons [
38]. However, further studies are required to understand the complexity and heterogeneity of the ISG response within different brain regions. We speculate that cell-type- and tissue-specific differential expression of these ISGs within different brain regions and the availability of cofactors might influence their antiviral action.
Astrocytes are important producers of IFNs during neurotropic viral infections [
43,
52]. We have previously shown that IFN signaling controls flavivirus infection and that viperin is highly upregulated in astrocytes [
4]. Here we could demonstrate the importance of viperin in astrocytes, and that it acts in concert with other ISGs to restrict TBEV infection. One such candidate ISG is TRIM79α which is also highly upregulated in astrocytes [
4], and antivirally active against TBEV [
53]. Other ISGs like ISG15, Oas1b, Ifit2, PKR, and RNase L were also upregulated in astrocytes after TBEV infection and might also contribute to the antiviral effects against TBEV in the absence of viperin [
4].
Neurons are the main target of TBEV infection in the CNS. We have previously shown that both WT and in the absence of IPS-1 or IFNAR signaling, neurons are the main target of LGTV infection [
2,
3]. Cortical neurons induce type I IFN after infection, respond to IFN treatment, and upregulate viperin and other ISGs (Mx1 and Ifit1). However, here we show that they rely on viperin expression to reduce TBEV and ZIKV replication, indicating that neither Mx1 nor Ifit1 is targeting TBEV replication. We and others have found that endogenous viperin expression fails to control WNV in cortical neurons [
18]. However, we now show that viperin is needed for the early IFN-mediated antiviral responses against WNV in these neurons. A possible explanation could be that WNV efficiently inhibits viperin induction in cortical neurons and thereby avoid its antiviral effect. WNV interferes with both upregulation of IFN through RIG-I-IPS-1 and cGAS-STING signaling pathways as well as inhibit IFNAR-mediated signal transduction (reviewed in [
54]). The different impact of viperin on the different flaviviruses could also be explained by differences between the flaviviruses.
Acknowledgements
We thank Gerhard Dobler from the Bundeswehr Institute of Microbiology (Munich, Germany) for providing the virus strains LGTV strain TP21, TBEV strain Hypr, and ZIKV strain MR766 and Sirkka Vene (Folkhälosinstitutet, Stockholm, Sweden) for providing JEV and WNV. Viperin-/- mice were a kind gift from Peter Cresswell, Department of Immunobiology, Yale University School of Medicine.