A mouse model for studying viscerotropic disease caused by yellow fever virus infection

Mosquito-borne yellow fever virus (YFV) causes highly lethal, viscerotropic disease in humans and non-human primates. Despite the availability of efficacious live-attenuated vaccine strains, 17D-204 and 17DD, derived by serial passage of pathogenic YFV strain Asibi, YFV continues to pose a significant threat to human health. Neither the disease caused by wild-type YFV, nor the molecular determinants of vaccine attenuation and immunogenicity, have been well characterized, in large part due to the lack of a small animal model for viscerotropic YFV infection. Here, we describe a small animal model for wild-type YFV that manifests clinical disease representative of that seen in primates without adaptation of the virus to the host, which was required for the current hamster YF model. Investigation of the role of type I interferon (IFN-alpha/beta) in protection of mice from viscerotropic YFV infection revealed that mice deficient in the IFN-alpha/beta receptor (A129) or the STAT1 signaling molecule (STAT129) were highly susceptible to infection and disease, succumbing within 6-7 days. Importantly, these animals developed viscerotropic disease reminiscent of human YF, instead of the encephalitic signs typically observed in mice. Rapid viremic dissemination and extensive replication in visceral organs, spleen and liver, was associated with severe pathologies in these tissues and dramatically elevated MCP-1 and IL-6 levels, suggestive of a cytokine storm. In striking contrast, infection of A129 and STAT129 mice with the 17D-204 vaccine virus was subclinical, similar to immunization in humans. Although, like wild-type YFV, 17D-204 virus amplified within regional lymph nodes and seeded a serum viremia in A129 mice, infection of visceral organs was rarely established and rapidly cleared, possibly by type II IFN-dependent mechanisms. The ability to establish systemic infection and cause viscerotropic disease in A129 mice correlated with infectivity for A129-derived, but not WT129-derived, macrophages and dendritic cells in vitro, suggesting a role for these cells in YFV pathogenesis. We conclude that the ability of wild-type YFV to evade and/or disable components of the IFN-alpha/beta response may be primate-specific such that infection of mice with a functional IFN-alpha/beta antiviral response is attenuated. Consequently, subcutaneous YFV infection of A129 mice represents a biologically relevant model for studying viscerotropic infection and disease development following wild-type virus inoculation, as well as mechanisms of 17D-204 vaccine attenuation, without a requirement for adaptation of the virus.

Non-pathogenic Sindbis virus causes hemorrhagic fever in the absence of alpha/beta and gamma interferons

The role of interferon-gamma (IFNgamma) in antiviral innate immune responses during acute alphavirus infection is not well defined. We examined the contribution of IFNgamma to the protection of adult mice from Sindbis virus (SB)-induced disease by comparing subcutaneous infection of mice lacking receptors for either IFNalpha/beta (A129), IFNgamma (G129) or both (AG129) to normal mice (WT129). While neither G129 nor WT129 mice exhibited clinical signs of disease, infection of A129 or AG129 mice was fatal with AG129 mice succumbing more rapidly. Furthermore, AG129 mice developed signs of viral hemorrhagic fever (VHF), including extensive hepatocellular damage, inflammatory infiltrates in multiple organs and vascular leakage, which were significantly delayed and/or partially ameliorated during fatal A129 infections. We conclude that: (i) IFNalpha/beta is the primary mediator of innate immunity to SB infection, however; (ii) IFNgamma is directly antiviral in vivo, acting before the adaptive immune response appears and; (iii) development of VHF may involve viral suppression of both IFNalpha/beta and IFNgamma responses.

Early restriction of alphavirus replication and dissemination contributes to age-dependent attenuation of systemic hyperinflammatory disease

Severity of alphavirus infection in humans tends to be strongly age-dependent and several studies using laboratory-adapted Sindbis virus (SB) AR339 strains have indicated that SB-induced disease in mice is similarly contingent upon host developmental status. In the current studies, the consensus wild-type SB, TR339, and in vivo imaging technology have been utilized to examine virus replication and disease manifestations in mice infected subcutaneously at 5 days of age (5D) vs 11D. Initial virulence studies with TR339 indicated that this age range is coincident with rapid transition from fatal to non-fatal outcome. Fatal infection of 5D mice is characterized by high-titre serum viraemia, extensive virus replication in skin, fibroblast connective tissue, muscle and brain, and hyperinflammatory cytokine induction. In contrast, 11D-infected mice experience more limited virus replication and tissue damage and develop mild, immune-mediated pathologies including encephalitis. These results further establish the linkage between hyperinflammatory cytokine induction and fatal outcome of infection. In vivo imaging using luciferase-expressing viruses and non-propagative replicons revealed that host development results in a restriction of virus replication within individual infected cells that is manifested as a delay in reduction of virus replication in the younger mice. Thus, an important contributing factor in age-dependent resistance to alphavirus infection is restriction of replication within first infected cells in peripheral tissues, which may augment other developmentally regulated attenuating effects, such as increasing neuronal resistance to virus infection and apoptotic death.

Heparan sulfate binding can contribute to the neurovirulence of neuroadapted and nonneuroadapted Sindbis viruses

Cell culture-adapted laboratory strains of Sindbis virus (SB) exhibit efficient initial attachment to cell surface heparan sulfate (HS) receptors. In contrast, non-cell-adapted strains, such as the SB consensus sequence virus TR339, interact weakly with HS and cell surfaces. Regardless of their HS binding phenotype, most SB strains do not cause fatal disease in adult mice, whether inoculated subcutaneously (s.c.) or intracranially (i.c.). However, laboratory strains of SB can be rendered neurovirulent for adult mice by introduction of a glutamine (Gln)-to-histidine (His) mutation at position 55 of the E2 envelope glycoprotein. In the current work, we have determined that E2 His 55-containing viruses require a second-site mutation (Glu to Lys) at E2 position 70 that confers efficient HS binding in order to exhibit virulence for adult mice and that virulence is correlated with very high infectivity for many cell types. Furthermore, introduction of E2 Lys 70 or certain other HS-binding mutations alone also increased morbidity and/or mortality over that of TR339 for older mice inoculated i.c. However, all viruses containing single HS-binding mutations were attenuated in s.c. inoculated suckling mice in comparison with TR339. These results suggest that HS binding may attenuate viral disease that is dependent on high-titer viremia; however, efficient cell attachment through HS binding can increase virulence, presumably through enhancing the replication of SB within specific host tissues such as the brain.

Sindbis virus translation is inhibited by a PKR/RNase L-independent effector induced by alpha/beta interferon priming of dendritic cells

The tropism of Sindbis virus (SB) for cells of the dendritic cell (DC) lineage and the virulence of SB in vivo are largely determined by the efficacy of alpha/beta interferon (IFN-alpha/beta)-mediated antiviral responses. These responses are essentially intact in the absence of PKR and/or RNase L (K. D. Ryman, L. J. White, R. E. Johnston, and W. B. Klimstra, Viral Immunol. 15:53-76, 2002). In the present studies, we investigated the nature of antiviral effects and identity of antiviral effectors primed by IFN-alpha/beta treatment of bone marrow-derived DCs (BMDCs) generated from mice deficient in PKR and RNase L (TD). IFN-alpha/beta priming exerted significant antiviral activity at very early stages of SB replication and most likely inhibited the initial translation of infecting genomes. The early effect targeted cap-dependent translation as protein synthesis from an SB-like and a simple RNA were inhibited by interferon treatment, but an encephalomyocarditis virus internal ribosome entry site-driven element exhibited no inhibition. Phosphorylation of the alpha subunit of eukaryotic translation initiation factor 2 was defective after virus infection of TD cells, suggesting other mechanisms of translation inhibition. To identify components of these alternative antiviral pathway(s), we have compared global gene regulation in BMDCs derived from normal 129 Sv/Ev, IFNAR1-/-, and TD mice following infection with SB or treatment with IFN-alpha/beta. Candidate effectors of alternative antiviral pathways were those genes induced by virus infection or IFN-alpha/beta treatment in 129 Sv/Ev and TD-derived BMDC but not in virus-infected or IFN-alpha/beta-treated IFNAR1-/- cells. Statistical analyses of gene array data identified 44 genes that met these criteria which are discussed.