Toll-like receptor 3 has a protective role against West Nile virus infection

Abstract

Protection against West Nile virus (WNV) infection requires rapid viral sensing and the generation of an interferon (IFN) response. Mice lacking IFN regulatory factor 3 (IRF-3) show increased vulnerability to WNV infection with enhanced viral replication and blunted IFN-stimulated gene (ISG) responses. IRF-3 functions downstream of several viral sensors, including Toll-like receptor 3 (TLR3), RIG-I, and MDA5. Cell culture studies suggest that host recognizes WNV in part, through the cytoplasmic helicase RIG-I and to a lesser extent, MDA5, both of which activate ISG expression through IRF-3. However, the role of TLR3 in vivo in recognizing viral RNA and activating antiviral defense pathways has remained controversial. We show here that an absence of TLR3 enhances WNV mortality in mice and increases viral burden in the brain. Compared to congenic wild-type controls, TLR3(-/-) mice showed relatively modest changes in peripheral viral loads. Consistent with this, little difference in multistep viral growth kinetics or IFN-alpha/beta induction was observed between wild-type and TLR3(-/-) fibroblasts, macrophages, and dendritic cells. In contrast, a deficiency of TLR3 was associated with enhanced viral replication in primary cortical neuron cultures and greater WNV infection in central nervous system neurons after intracranial inoculation. Taken together, our data suggest that TLR3 serves a protective role against WNV in part, by restricting replication in neurons.

FIG. 1.

Survival analysis of wild-type and TLR3^−/− C57BL/6 mice. (A and B) Eight- to twelve-week-old mice (n = 10 to 20) were inoculated with either 10² PFU of C6/36- or Vero-derived WNV via the subcutaneous route (footpad injection) (A) or with 10³ PFU of C6/36- or Vero-derived WNV via the intraperitoneal route (B) and monitored for mortality for 21 days. Survival curves for TLR3^−/− and wild-type mice were statistically significant independently of the virus stock or the route of infection (P < 0.0001).

FIG. 2.

Virological analysis of wild-type and TLR3^−/− C57BL/6 mice. (A to F) Viral burden in peripheral and CNS tissues of wild-type and TLR3^−/− C57BL/6 mice infected with 10² PFU of C6/36-derived viruses administered subcutaneously. WNV RNA in serum (A) and draining lymph node (B) and infectious virus in the spleen (C), kidney (D), brain (E), and spinal cord (F) were determined from samples harvested on days 2, 4, 6, 8, and 10 by qRT-PCR (A and B) or viral plaque assay (C to F). The data are shown as viral RNA equivalents or PFU per gram of tissue for 10 to 12 mice per time point. For viral load data, the solid line represents the median PFU per gram at the indicated time point, and the dotted line represents the limit of sensitivity of the assay. Asterisks indicate values that are statistically significant (*, P < 0.05; **, P < 0.005; ***, P < 0.0001) compared to wild-type mice.

FIG. 3.

WNV-specific antibody responses and CD8⁺ T-cell activation in wild-type and TLR3^−/− mice. (A and B) Wild-type and TLR3^−/− mice were infected with WNV, and serum was collected at the indicated time points. The development of specific IgM (A) or IgG (B) against WNV was determined by ELISA using purified WNV E protein. The data are the average of at least five mice per time point. (C) Wild-type and TLR3^−/− mice were infected with WNV, and splenocytes were harvested at day 7 after infection. The percentage of IFN-γ-producing CD8⁺ T cells after ex vivo restimulation with a WNV-specific NS4B peptide or PMA and ionomycin was determined by flow cytometry. n.d, values that were below the limit of detection of the assay; ns, differences that were not statistically different. (D) Trafficking of inflammatory cells into the CNS after WNV infection. Wild-type and TLR3^−/− mice were infected with 10² PFU of insect cell-derived WNV, and brains were harvested after extensive perfusion at day 9 after infection. CNS leukocytes were isolated after by Percoll centrifugation and analyzed by flow cytometry. The absolute number of specific inflammatory cells (CD4⁺ T, CD8⁺ T, CD45⁺, and CD11b⁺ cells) in the brains of wild-type and TLR3^−/− mice after WNV infection were calculated and reflect an average of five mice per group.

FIG. 4.

Effect of TLR3 on BBB permeability upon WNV infection. Wild-type and TLR3^−/− mice were infected with 10² PFU of insect cell-derived WNV via subcutaneous injection and administered a 1% Evans blue solution at the indicated time points. The levels of Evans blue in whole brains were quantified by measuring the absorbance at 620 nm by spectrophotometry after tissue homogenization and precipitation. The data are the average of five mice per time point.

FIG. 5.

IFN induction in draining lymph node and serum of wild-type and TLR3^−/− mice infected with WNV. (A and B) Mice were inoculated with 10² PFU of WNV by footpad injection and sacrificed at the indicated times. Total RNA from the draining lymph was analyzed for IFN-α (A) and IFN-β (B) mRNA expression by qRT-PCR. The data are normalized to 18S rRNA and are expressed as the relative fold increase over normalized RNA from uninfected controls. Average values are from 5 to 12 mice per time point, and error bars indicate the standard deviations. (C) IFN activity was determined from serum collected on days 1 to 4 after infection by an EMCV protection bioassay in L929 cells. Serum was added to L929 cells and, after subsequent infection with EMCV, the cytopathic effect was measured. The data reflect the average of serum samples harvested from 5 to 10 mice per time point, and the data are expressed as IU of IFN-α/ml. The specificity of the assay was confirmed with an anti-IFN-α/β receptor neutralizing MAb (data not shown).

FIG. 6.

Effect of TLR3 on WNV infection and IFN induction in primary fibroblasts, dendritic cells, and macrophages. MEFs (A), BM-DC (D), and BM-Mφ (G) generated from wild-type or TLR3^−/− mice were infected with WNV, and virus production was evaluated at the indicated times postinfection by plaque assay. IFN-α and β levels were quantified by a capture ELISA from the supernatants of WNV-infected MEFs (B and C) and BM-DC (E and F). Values are an average of quadruplicate samples generated from at least three independent experiments.

FIG. 7.

Role of TLR3 on WNV infection and IFN production in primary cortical neurons. (A) Primary cortical neurons generated from wild-type or TLR3^−/− mice were infected at an MOI of 0.001, and virus production was evaluated at the indicated times by plaque assay. (B and C) The production of IFN-α (B) and IFN-β (C) proteins by WNV-infected cortical neurons was analyzed by ELISA. Values are an average of triplicate samples generated from three independent experiments. Asterisks indicate values that are statistically significant (*, P < 0.05; **, P < 0.005; ***, P < 0.0001).

FIG. 8.

TLR3 controls viral replication in the CNS after i.c. WNV infection. Wild-type or TLR3^−/− mice were inoculated with 10¹ PFU of WNV by i.c. inoculation. Brains (A) and spinal cord (B)s were harvested at the indicated time points, and the virus titers were determined as described in Fig. 2 (*, P < 0.05; ***, P < 0.0001).

FIG. 9.

TLR3 controls WNV replication in the CNS by restricting infection of neurons. (A to C) Brains sections from wild-type and TLR3^−/− mice that were infected i.c. with WNV were costained for WNV antigen (red), the nuclear stain TO-PRO3 (blue) and the neuronal marker MAP-2 (green) (A), the astrocyte-specific marker GFAP (green) (B), or the microglial/macrophage cell-specific marker CD11b (green) (C). White arrows indicate double-positive cells. The data are representative of sections from at least five wild-type or TLR3^−/− mice.