MDA5 Is Critical to Host Defense during Infection with Murine Coronavirus

Abstract

Infection with the murine coronavirus mouse hepatitis virus (MHV) activates the pattern recognition receptors melanoma differentiation-associated gene 5 (MDA5) and Toll-like receptor 7 (TLR7) to induce transcription of type I interferon. Type I interferon is crucial for control of viral replication and spread in the natural host, but the specific contributions of MDA5 signaling to this pathway as well as to pathogenesis and subsequent immune responses are largely unknown. In this study, we use MHV infection of the liver as a model to demonstrate that MDA5 signaling is critically important for controlling MHV-induced pathology and regulation of the immune response. Mice deficient in MDA5 expression (MDA5(-/-) mice) experienced more severe disease following MHV infection, with reduced survival, increased spread of virus to additional sites of infection, and more extensive liver damage than did wild-type mice. Although type I interferon transcription decreased in MDA5(-/-) mice, the interferon-stimulated gene response remained intact. Cytokine production by innate and adaptive immune cells was largely intact in MDA5(-/-) mice, but perforin induction by natural killer cells and levels of interferon gamma, interleukin-6 (IL-6), and tumor necrosis factor alpha (TNF-α) in serum were elevated. These data suggest that MDA5 signaling reduces the severity of MHV-induced disease, at least in part by reducing the intensity of the proinflammatory cytokine response.

Importance: Multicellular organisms employ a wide range of sensors to detect viruses and other pathogens. One such sensor, MDA5, detects viral RNA and triggers induction of type I interferons, chemical messengers that induce inflammation and help regulate the immune responses. In this study, we sought to determine the role of MDA5 during infection with mouse hepatitis virus, a murine coronavirus used to model viral hepatitis as well as other human diseases. We found that mice lacking the MDA5 sensor were more susceptible to infection than were mice with MDA5 and experienced decreased survival. Viral replication in the liver was similar in mice with and without MDA5, but liver damage was increased in MDA5(-/-) mice, suggesting that the immune response is causing the damage. Production of several proinflammatory cytokines was elevated in MDA5(-/-) mice, suggesting that MDA5 may be responsible for keeping pathological inflammatory responses in check.

FIG 1

MDA5^−/− mice have heightened susceptibility to MHV infection. Survival was monitored after i.p. inoculation of 500 PFU MHV. Results were statistically significant by the Mantel-Cox test (A). Viral titers in the liver, spleen, brain, and spinal cord were determined by plaque assay 3, 5, and 7 days after i.p. inoculation with 500 PFU MHV (B) and at 5 days postinoculation in the lungs, heart, and kidney (C). For panels B and C, statistical significance was determined by a two-part test using a conditional t test and proportion test. Bars represent the means. Data for all panels were pooled from two independent experiments.

FIG 2

MDA5^−/− mice have increased pathology in the liver. Three, 5, and 7 days following i.p. inoculation with 500 PFU MHV, livers were removed, fixed, sectioned, and stained with hematoxylin and eosin. Four or five nonoverlapping fields of view, each 9 by 10⁶ μm², were chosen at random, the number of inflammatory foci was determined, and the mean was calculated (A). Statistical significance was determined by a two-part test using a conditional t test and proportion test. Samples in which foci had coalesced into a continuous confluence were scored as too numerous to count (TNTC). Representative fields are shown in panel B. Scale bar, 500 μm. Data are from one experiment.

FIG 3

Interferon, but not interferon-stimulated gene, expression is reduced in MDA5^−/− mice. On days 3, 5, and 7 following i.p. inoculation with 500 PFU MHV, RNA was isolated from the liver. Gene expressions for type I interferons (A) and ISGs (B) were quantified by qRT-PCR and expressed as fold changes from values for mock-infected mice. Statistical significance was determined by an unpaired, two-tailed t test. Outliers were eliminated by a ROUT test (Q = 1%). Bars represent the means. Data were pooled from two independent experiments, with the exception of viperin mRNA quantification, which was performed once.

FIG 4

Innate inflammatory cell recruitment is largely intact in MDA5^−/− mice, but activation of dendritic cells is elevated. On days 1, 2, and 3 following i.p. inoculation with 500 PFU MHV, mice were sacrificed, organs were harvested, and single-cell suspensions derived from the liver (A) or spleen (B) were immunophenotyped by staining with cell type-specific antibodies and analysis by flow cytometry. The total numbers of macrophages (CD3⁻ CD19⁻ NK1.1⁻ CD45⁺ Ly6-C⁻ Ly6-G⁻ CD11c⁻ CD11b⁺ F4/80⁺), neutrophils (CD3⁻ CD19⁻ NK1.1⁻ CD45⁺ Ly6-C⁻ Ly6-G⁺), natural killer cells (CD3⁻ CD19⁻ NK1.1⁺CD45⁺ Ly6-C⁻ Ly6-G⁻), inflammatory monocytes (CD3⁻ CD19⁻ NK1.1⁻ CD45⁺ CD11b⁺ Ly6-C⁺Ly6-G⁻), and total dendritic cells (CD3⁻ CD19⁻ NK1.1⁻ CD45⁺ Ly6-C⁻ Ly6-G⁻ CD11c⁺) are shown, normalized by the mass of the tissue (A). Surface expression of the activation markers CD80 and CD86 was assessed on splenic plasmacytoid (CD3⁻ CD19⁻ NK1.1⁻ CD45⁺ CD11c⁺ B220⁺PDCA-1⁺), lymphoid (CD3⁻ CD19⁻ NK1.1⁻ CD45⁺ CD11c⁺ CD11b⁻), and myeloid (CD3⁻ CD19⁻ NK1.1⁻ CD45⁺ CD11c⁺ CD11b⁻) dendritic cells, and the total number of CD80⁺ CD86⁺ DCs of each subtype is shown (B). Statistical significance was determined by an unpaired, two-tailed t test. Bars represent the means. Data were pooled from two independent experiments.

FIG 5

Cytokine production by natural killer and natural killer T cells in MDA5^−/− mice is similar to or greater than that observed in WT mice. WT and MDA5^−/− mice were sacrificed 5 days after i.p. inoculation with 500 PFU MHV, spleens and livers were harvested, and single-cell suspensions derived from the organs were incubated with brefeldin A for 3 h and immunophenotyped by staining with cell type-specific antibodies. Cells were permeabilized and stained with antibody specific to interferon gamma and perforin and then analyzed by flow cytometry. Natural killer (CD3⁻ NK1.1⁺) and natural killer T cells (CD3⁺ NK1.1⁺) were assessed. Statistical significance was determined by an unpaired, two-tailed t test. Bars represent the means. Data were pooled from two independent experiments.

FIG 6

T cell activation is independent of MDA5 signaling. Day 7 following i.p. inoculation with 500 PFU MHV, mice were sacrificed, spleens were harvested, and single-cell suspensions were derived from the spleens of wild-type and MDA5^−/− mice. Cells were immunophenotyped by staining with cell type-specific antibodies and analysis by flow cytometry. The percentage (A) and total number (B) of CD4 (CD3⁺ CD4⁺) and CD8 (CD3⁺ CD8⁺) T cells with elevated surface expression of CD44 and CD11a are shown. Expression of PD-1 (representative image in panel C) was assessed on CD44⁺ and CD11a⁺ CD4 and CD8 T cells (D). Statistical significance was determined by an unpaired, two-tailed t test. Bars represent the means. Data are representative of two independent experiments.

FIG 7

Interferon gamma production, but not perforin production, is elevated in T cells from MDA5^−/− mice. WT and MDA5^−/− mice were sacrificed 7 days after i.p. inoculation with 500 PFU MHV, spleens and livers were harvested, single-cell suspensions derived from the organs were incubated with immunodominant MHV peptides S598 (class I) and M133 (class II), and cells were immunophenotyped by staining with cell type-specific antibodies. Cells were permeabilized and stained with antibody specific to interferon gamma and perforin and then analyzed by flow cytometry (representative image in panel A). CD8 (CD3⁺ CD8⁺) T cells, cells from the spleen and liver, were assessed for expression of perforin (B). CD4 (CD3⁺ CD4⁺) and CD8 (CD3⁺ CD8⁺) T cells from the spleen (C) and liver (D) were also assessed for expression of interferon gamma. Statistical significance was determined by an unpaired, two-tailed t test. Bars represent the means. Data are representative of two independent experiments.

FIG 8

Elevated cytokines, but not T cells, may explain decreased survival of infected MDA5^−/− mice. MDA5^−/− mice were treated with CD4- and CD8-depleting antibodies or a negative control −2, 4, and 6 days after infection with 500 PFU MHV inoculated i.p. Survival was monitored (A). Data were pooled from two independent experiments. Results are nonsignificant by the Mantel-Cox test. Serum was collected from WT and MDA5^−/− mice with intact T cell compartments 3, 5, and 7 days after i.p. inoculation of 500 PFU MHV and analyzed for levels of interferon gamma, IL-6, and TNF-α by enzyme-linked immunosorbent assay (ELISA) (B). Statistical significance was determined by a two-part test using a conditional t test and proportion test. Data were pooled from four independent experiments (n of 3 to 6).