High-resolution spatial transcriptomics maps viral tropism and reveals spatially organized immune modules in viral encephalitis

Abstract

Viral encephalitis is a debilitating disease that most commonly affects vulnerable populations, including the very young and elderly. Despite its severity, few countermeasures exist due to the structural and immunological complexities of the central nervous system (CNS). To better understand the spatial dynamics of viral spread and the concurrent host immune responses, we used high-resolution spatial transcriptomics to profile reovirus infection in the neonatal mouse brain at 3, 5, and 7 days post-infection. We constructed a comprehensive spatiotemporal atlas of infection, which revealed viral dissemination from sites of cerebrospinal fluid circulation to neighboring brain parenchyma, with enriched infection in the thalamus and midbrain coincident with Slc17a6 (VGLUT2) excitatory neurons. Unbiased spatial gene co-expression network analysis uncovered rich gene lists linked to temporal waves of host immune responses, originating with interferon-stimulated gene modules, followed by myeloid cell infiltration and adaptive cytotoxic T-cell responses at times of peak disease. Furthermore, we identified spatial correlation between viral transcripts and several upregulated host snoRNA-related genes (e.g., Nop58 and Snhg1) with high-confidence, suggesting viral use of host ribosomal modification machinery. We also observed a strong anti-correlation of astrocyte markers with reovirus transcripts, suggesting glial cell disruption. This work provides a high-definition spatial framework to understand interactions between viral infection and host immunity in the brain.