Murine Models of Hepatitis A Virus Infection

Abstract

Mechanistic analyses of hepatitis A virus (HAV)-induced pathogenesis have long been thwarted by the lack of tractable small animal models that recapitulate disease observed in humans. Several approaches have shown success, including infection of chimeric mice with human liver cells. Other recent studies show that HAV can replicate to high titer in mice lacking expression of the type I interferon (IFN) receptor (IFN-α/β receptor) or mitochondrial antiviral signaling (MAVS) protein. Mice deficient in the IFN receptor show critical features of type A hepatitis in humans when challenged with human HAV, including histological evidence of liver damage, leukocyte infiltration, and the release of liver enzymes into blood. Acute pathogenesis is caused by MAVS-dependent signaling that leads to intrinsic apoptosis of hepatocytes.

Figure 1.

Fluorescent in situ hybridization (FISH) detection of intrahepatic hepatitis A virus (HAV) RNA in two different murine models supporting HAV infection. (A) Liver of an infected Alb-urokinase-type plasminogen activator (uPA)/severe combined immunodeficiency (SCID) beige mouse containing 10⁸ genome equivalents (GE)/g liver tissue as determined by RT-qPCR. Red fluorescence indicates the presence of HAV RNA. (B) Similar image of a liver section from an infected Ifnar1^−/−Ifngr1^−/− (DKO) mouse with 10⁷ HAV GE/g liver tissue.

Figure 2.

Buoyant densities of hepatitis A virus (HAV) virions in HAV-infected Alb-urokinase-type plasminogen activator (uPA)/SCID beige mice with chimeric human livers. Virus present in (A) serum, (B) bile, and (C) feces of an infected Alb-uPA/SCID mouse were centrifuged to equilibrium in an isopycnic gradient as described by Feng et al. (2013). Twenty samples were collected from the top of the gradient tube and density determined using a refractometer. HAV RNA was assayed in all fractions by RT-qPCR. Only low-density virus (1.06–1.10 g/cm³) consistent with quasi-enveloped HAV (Feng et al. 2013) was identified in the serum of infected mice, whereas only nonenveloped virus particles were found in mouse bile and feces.

Figure 3.

Hepatitis A virus (HAV) challenge of Ifnar1^−/−, Ifngr1^−/−, and Ifnar1^−/−/Ifngr1^−/− double knockout (DKO) and wild-type (WT) B6 mice. WT mice or mice deficient in the type I interferon (IFN) receptor (Ifnar1^−/−), type II IFN receptor (Ifngr1^−/−), or both receptors (DKO) were inoculated intravenous (i.v.) with 10⁸ genome equivalents (GE) of 4th mouse passage HAV. RT-qPCR quantitation of HAV RNA in (A) fecal samples, (B) liver, and (C) serum alanine aminotransferase (ALT) activity were assessed at multiple time points. WT and Ifngr1^−/− mice showed no evidence of infection or increased ALT at any time. In contrast, both Ifnar1^−/− and DKO mice shed virus in feces for 2 months. Ifnar1^−/− mice showed greater fecal shedding for a lengthier period compared with DKO mice. Viral RNA in the liver peaked during the first month of infection, declined by 2 months and was maintained at ∼3 × 10⁵ GE/μg for 5 months. ALT levels in both Ifnar1^−/− and DKO mice peaked during the first 2 weeks of infection, then rapidly declined to a homeostatic level that was maintained for 5 months. Increases in serum ALT activity correlated with persistence of viral RNA in the liver. There was a transient period when DKO mice showed significantly greater ALT activity than Ifnar1^−/− mice, suggesting a possible role for IFN-γ in liver pathogenesis. (Modified from data in Hirai-Yuki et al. 2016a.)

Figure 4.

Hepatitis A virus (HAV) inoculum size and infection kinetic in Ifnar1^−/− and Ifnar1^−/−Ifngr1^−/− (DKO) mice. Ifnar1^−/− and Ifnar1^−/−Ifngr1^−/− mice were inoculated intravenously with 4th mouse passage virus ranging from 10⁸ to 10² genome equivalents (GE) as shown at the top. Lack of the type I interferon (IFN) receptor determines permissiveness and these two knockouts appear similar in terms of susceptibility to infection. Higher titer inocula result in earlier increases in serum alanine aminotransferase (ALT). The minimal infectious dose is between 10⁴ and 10⁵ GE, which is equivalent to 50–500 infectious virus particles given the particle/infectivity ratio of HAV.

Figure 5.

Mitochondrial antiviral-signaling protein (MAVS)-associated apoptosis in hepatitis A virus (HAV)-infected mice. Immunohistochemical staining for cleaved caspase-3 was performed on liver sections from Ifnar1^−/− and Mavs^−/− mice 15 days after infection. Ifnar1^−/− liver (left) shows evidence of caspase activation (brown stain), magnified in the inset. Cells with caspase-3 staining also show surrounding inflammatory cell infiltrates. Neither cleaved caspase-3 nor cellular infiltrates are present in the Mavs^−/− livers, despite an almost 10-fold higher abundance of viral RNA. Scale bars, 12.5 µm (left, inset); 100 µm (right). (From Hirai-Yuki et al. 2016a; reprinted, with permission, from The American Association for the Advancement of Science © 2016.)

Figure 6.

Absence of mitochondrial antiviral-signaling protein (MAVS)-signaling results in sustained fecal shedding of hepatitis A virus (HAV). (A) Serum alanine aminotransferase (ALT) activity, and (B) fecal viral shedding were tracked over time in infected Ifnar1^−/− and Mavs^−/− mice. Infected Ifnar1^−/− mice show elevated ALT during the first 2 weeks followed by a gradual decline matched by reductions in fecal virus shedding. Mavs^−/− mice showed no evidence of increased ALT following infection, consistent with the absence of apoptosis and leukocyte infiltration of the liver, and sustained fecal shedding of virus. (From Hirai-Yuki et al. 2016a; reprinted, with permission, from The American Association for the Advancement of Science © 2016.)

Figure 7.

A myeloid-derived suppressor cell-like population is recruited to the liver during hepatitis A virus (HAV) infection. Leukocytes were identified in the livers of uninfected or HAV-infected Ifnar1^−/− mice. (A) At day 7 postinfection, flow cytometry identified a robust increase in probable MDSC cells (F4/80⁺CD11b⁺MHCII⁻) in the liver. Macrophage-like cells also showed an increase as a result of infection but were lower in number compared with other subsets of cells. There were also modest increases in T cells and a 10-fold increase in NK⁺ cells and NK⁺ T cells. The change in the small population of γ/δ T cells did not reach significance. (B) Immunohistochemical staining of CD3, CD4, and CD8 in an Ifnar1^−/− mouse 15 days postinfection, ALT = 528 IU/L. (C) Dual immunohistochemical staining of infected Ifnar1^−/− liver at day 14 postinfection. The magenta represents anti-CD4 staining and the brown represents anti-CD8 staining. The image reveals a mixed cellular infiltrate, with a large proportion of the infiltrating cells staining for neither CD8 nor CD4.

Figure 8.

Mitochondrial antiviral-signaling protein (MAVS)- and IRF3/7-dependent signaling limits hepatitis A virus (HAV) replication and triggers hepatocyte apoptosis. (A) HAV infection of hepatocytes in Ifnar1^−/− mice results in RIG-I/MDA recognition of cytosolic viral RNA and MAVS activation, leading to phosphorylation and dimerization of IRF3/IRF7 and interferon (IFN) production. Although IFN is inactive because of the absence of type I IFN receptors, activated IRF3 directly stimulates the transcription of proapoptotic IFN-stimulated genes (ISGs), and chemokines such as CCL5, that induce hepatocellular apoptosis and drive inflammation within the liver. Although much less effective than IFN-driven ISG responses in wild-type mice that completely restrict replication, IRF3-stimulated gene expression also suppresses viral replication over time in concert with the induction of adaptive immunity. Activated IRF3 may directly induce apoptosis or may induce the expression of proapoptotic factors that result in hepatocyte apoptosis. (B) In infected Mavs^−/− mice, IRF3 remains inactive, and there is no induction of apoptosis by either mechanism. Nuclear factor (NF)-κB is unlikely to have a critical role in disease pathogenesis, as it should be activated through MAVS in infected Irf3^−/−Irf7^−/− mice that (like Mavs^−/− mice) show no evidence of disease.