MAVS signaling of long-lived brain-resident myeloid cells is needed during viral encephalitis to adjust the transcriptome of CNS infiltrating CD8+ T cells

Abstract

Neurotropic viruses like vesicular stomatitis virus (VSV) can infect the central nervous system (CNS) through the olfactory route. Following intranasal instillation, VSV moves along the axons of olfactory sensory neurons to the olfactory bulb. While within the olfactory bulb the spread of the virus is controlled by microglia activation and the recruitment of peripheral leukocytes, some of the underlying mechanisms remain unknown. To investigate these mechanisms, we used mice with conditional deletions of the mitochondrial antiviral-signaling protein (MAVS), an adaptor for RIG-I-like receptor (RLR) signaling. By selectively deleting MAVS in neurons, astrocytes, or long-lived myeloid cells, we discovered that RLR signaling specifically within brain-resident myeloid cells is crucial for protection against the virus. Infected mice with a MAVS deletion in these myeloid cells showed normal myeloid cell and leukocyte infiltration into the brain. However, the P2RY12⁺ microglia showed aberrant expression of genes involved in antigen cross-presentation. Furthermore, flow cytometry experiments revealed diminished MHC class I expression on MAVS deficient microglia. Moreover, CNS infiltrating CD8⁺ T cells had dysfunctional transcriptional profiles. Therefore, our findings indicate that during viral CNS infection, MAVS signaling in brain-resident myeloid cells, presumably microglia, is essential for antigen cross-presentation and the relicensing of protective, infiltrating CD8⁺ T cells.

Keywords: Brain infection; CD8+ T cells; CNS; Cross-presentation; MAVS signaling; Microglia; RLR signaling.

Fig. 1

MAVS signaling is essential for protection against intranasal VSV infection. C57BL/6 and Mavs^−/− mice were (A) i.v. and (B) i.n. infected with 10³ PFU of VSV and survival was monitored for 30 days (N = 2, n = 9 per genotype, combined data) Log-rank (Mantel Cox) Test. * < 0.05, ** < 0.01, ***P < 0.001, ****P < 0.0001. C Virus titers were determined by plaque assay from homogenates of the OB prepared at the indicated days after infection (N = 2, n = 5–7 per genotype, combined data). Two-tailed Mann–Whitney test * < 0.05, ** < 0.01, ***P < 0.001, ****P < 0.0001. D Mavs and Ddx58 (RIG-I) expression was determined by RT-qPCR in the indicated tissues from untreated C57BL/6 animals (N = 2, n ≥ 5, combined data). E Mavs expression was determined by RT-qPCR in OB samples from C57BL/6 mice 4, 6, and 8 days after i.n. infection (N = 2, n = 6). F Head-shielded mice were lethally irradiated with 10 Gy and the next day were i.v. reconstituted with the 1 × 10⁷ bone marrow cells of the indicating genotype. Following 8 weeks of recovery, mice were intranasally instilled with 10³ PFU of VSV and survival was monitored for 30 days (N = number of experiments, n = number of biological replicates, N = 2, n = 4–9 per genotype, combined data) Log-rank (Mantel Cox) Test. * < 0.05, ** < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 2

MAVS signaling of microglia is required to control VSV encephalitis. C57BL/6 (WT), Mavs^−/−, GFAP-Cre^±MAVS^fl/fl, Syn1-Cre^±MAVS^fl/fl, CX3CR1-Cre^ER−/−MAVS^fl/fl mice and CX3CR1-Cre^ER±MAVS^fl/fl mice were i.n. VSV infected with 10³ PFU (A-D). A Survival of infected mice was monitored for 30 days (N = 2, n = 9-11 per genotype, combined data). Log-rank (Mantel Cox) Test. * < 0.05, ** < 0.01, ***P < 0.001, ****P < 0.0001. The color code shown in the legend for single genotypes of mice applies to A-D. B The OB and spinal cord of infected mice were prepared 6 and 8 days post infection, respectively, and viral load was determined from lysates by a plaque assay (N = 2, n ≥ 5 per genotype, combined data). The dataset of day 6 post infection OB plaques for C57BL/6 and Mavs^−/− mice is the same as shown in Fig. 1C. C The olfactory bulb of infected mice was prepared 4 and 6 dpi and IFN-β was determined from lysates by an ELISA method (N = 2, n ≥ 5 per genotype, combined data). B – C Two-tailed Mann–Whitney test * < 0.05, ** < 0.01, ***P < 0.001, ****P < 0.0001. D The OB of infected mice was prepared 6 days post infection and a flow cytometry-based bead cytokine array was performed from the lysates. The heat map shows the concentration of the indicated cytokines as the mean value from mice of different genotypes (N = 2, n = 5 per genotype, combined data). E WT mice were intranasally infected with 10³ PFU of VSV-eGFP and 6 days later histological analysis was performed from the olfactory bulb, VSV-eGFP (green) and iba1⁺ (red) within the glomerular cell layer of the OB (representative data from n = 3). F Percentage of GFP⁺ iba⁺ on total iba⁺ cells from (E) (n = 3)

Fig. 3

Normal myeloid cell density in mice with microglia-selective MAVS deletion. A Representative histology and quantification of distribution of iba1⁺ cells within the OB C57BL/6 (WT), Mavs^−/−, and CX3CR1-Cre^ER±MAVS^fl/fl mice intranasally infected with 10³ PFU of VSV at 6 days post infection in the glomerular and the granular cell layer (N = 2, n ≥ 6 per genotype). Two-tailed Mann–Whitney test * < 0.05, ** < 0.01, ***P < 0.001, ****P < 0.0001. B CX3CR1-Cre^ER±tdTomato^wt/st were i.n. instilled with PBS or with 10³ PFU of VSV, brains were extracted at 6 days post treatment and brain-resident immune cells were isolated and immunolabelled. Percentage of P2RY12⁺ tdTomato⁺ counts on total tdTomato⁺ cells in PBS and VSV treated mice. (N = 2, n = 5 per condition, representative data)

Fig. 4

Selective MAVS deletion in microglia impairs the expression of genes relevant for cross-presentation. C57BL/6 (WT) and CX3CR1-Cre^ER±MAVS^fl/fl mice were intranasally instilled with PBS or 10³ PFU of VSV and on 6 days post treatment the olfactory bulb was prepared, immune cells were isolated and CD45⁺CD11b^lowLy6C⁻P2RY12⁺ microglia were directly sorted into 350 µl of β-mercaptoethanol-containing RA1 buffer for RNA extraction and subsequent bulk sequencing. A Principal component analysis of RNA-seq data from uninfected (PBS) and infected microglia from C57BL/6 (WT) and CX3CR1-Cre^ER± MAVS^fl/fl mice. Each dot represents data from a single mouse. B Heat map of relative expression of differentially expressed genes that are relevant in adaptive immune response pathway with a log₂-fold change cut-off between microglia from C57BL/6 (WT) and CX3CR1-Cre^ER± MAVS^fl/fl mice treated with PBS or 10³ PFU of VSV. C Gene counts of Tap1, Tap2, and Ifngr gene from CD45⁺CD11b^lowLy6C⁻P2RY12⁺ cells from C57BL/6 (WT) and CX3CR1-Cre^ER± MAVS^fl/fl mice infected with 10³ PFU of VSV. D Flow cytometry histogram plots and percentage of MHC class I expression in total microglia counts within the OB of C57BL/6 (WT) and CX3CR1-Cre^ER±MAVS^fl/fl mice i.n. instilled with 10³ PFU of VSV at 6 dpi (n = 3 per genotype)

Fig. 5

MAVS signaling of long-lived CX3CR1+ myeloid cells defines the transcriptomic profile of CNS infiltrating CD8⁺ T cells during viral encephalitis. C57BL/6 (WT), Mavs^−/− and CX3CR1-Cre^ER± MAVS^fl/fl mice were intranasally infected with 10³ PFU of VSV and mice were prepared 6 days post infection. A Histological analysis of CD45⁺ and CD8α⁺ within the OB (representative data, N = 2, n ≥ 5 per genotype). Small insets were selected and magnified for better visualization from the glomerular layer of the OB where we observe the vast majority of immune infiltrating CD45⁺ and CD8a⁺ cells. B Quantification of CD8⁺ T cells within the glomerular layer of the OB from the experiment in Fig. 5A. (N = 2, n ≥ 5 per genotype, combined data). Two-tailed Mann–Whitney test * < 0.05, ** < 0.01, ***P < 0.001, ****P < 0.0001. C—E C57BL/6 (WT), CX3CR1-Cre^ER−/−MAVS^fl/fl and CX3CR1-Cre^ER±MAVS^fl/fl mice were intranasally instilled with PBS or 10³ PFU of VSV and on 6 days post treatment the OB and the cervical lymph nodes were prepared, immune cells were isolated and CD8β⁺ T cells were directly sorted into 350 µl of β-mercaptoethanol RA1 buffer for RNA extraction and subsequent bulk sequencing. C Principal component analysis of CD8⁺ T cells from the OB and the cervical lymph nodes of C57BL/6 (WT), CX3CR1-Cre^ER−/−MAVS^fl/fl and CX3CR1-Cre^ER±MAVS^fl/fl mice i.n. infected with 10³ PFU of VSV or with PBS instilled. Each dot represents a single mouse. D Comparative pathway analysis of top 10 differentially enriched pathways of CD8⁺ T cells from CX3CR1-Cre^ER−/−MAVS^fl/fl and CX3CR1-Cre^ER±MAVS^fl/fl mice i.n. infected with 10³ PFU of VSV. E Gene counts of Ifng and Cd44 gene from CD8⁺ T cells from C57BL/6 (WT), CX3CR1-Cre^ER−/−MAVS^fl/fl and CX3CR1-Cre^ER±MAVS^fl/fl mice infected with 10³ PFU of VSV. Each dot represents a single mouse