Antiretroviral therapy restores the homeostatic state of microglia in SIV-infected rhesus macaques

Abstract

Microglia and macrophages are essential for homeostatic maintenance and innate immune response in the brain. They are the first line of defense against infections such as HIV/SIV in the brain. However, they are susceptible to infection and function as viral reservoirs even under effective viral suppression. While current antiretroviral regimens successfully suppress viremia and improve quality of life and lifespan, neurologic complications persist and are in part attributed to activated microglia. We sought to test the hypothesis that brain microglia return to a more homeostatic-like state when viremia is suppressed by combination antiretroviral therapy. Using the SIV-rhesus macaque model, we combined single-cell RNA sequencing, bioinformatics, and pathway analysis to compare gene expression profiles of brain myeloid cells under 4 conditions: uninfected, SIV infected, SIV infected with cART suppression, and SIV encephalitis (SIVE). Our study reveals greater myeloid diversity and an elevated proinflammatory state are associated with untreated SIV infection compared with uninfected animals. The development of encephalitis and suppression of viremia both reduced myeloid diversity. However, they had converse effects on the activation state of microglia and inflammation. Notably, suggestive of a restoration of a homeostatic state in microglia, gene expression and activation of pathways related to inflammation and immune response in cART-suppressed monkeys were most similar to that in uninfected monkeys. Untreated SIV infection shared characteristics, especially in brain macrophages to SIVE, with SIVE showing dramatic inflammation. In support of our hypothesis, our study demonstrates that cART indeed restores this key component of the brain's homeostatic state. Summary: ScRNA-seq of rhesus monkey microglia reveals clusters of cells in activated states in the setting of SIV infection, which is primarily reversed by suppressing viremia with combination antiretroviral therapy.

Keywords: HIV; brain; macrophage; neuroHIV; scRNA-seq.

FIGURE 1

UMAP visualizations of scRNA‐seq data. (A) Showing all cells following QA/QC, normalization, and filtering with the annotated gene list. Cells from the monkeys in the different conditions are designated by different colors. (B) as in (A), but with cells from the different graph‐based clusters designated by different colors. (C) As in (A), but cluster 7 (CTL and NK cells) and cluster 9 (endothelial cells) removed and colored by the expression of myeloid genes AIF1 and CSF1R

FIGURE 2

Different conditions are made up of different clusters of cells. (A) UMAP visualization of brain myeloid cell sc‐RNAseq (excluding clusters 7 and 9), split to show cells from the different conditions, colored by graph‐based clusters. (B) Pie charts with areas proportional to the percentage of cells in each condition from the different graph‐based clusters of brain myeloid cell sc‐RNAseq

FIGURE 3

Expression of myeloid cell genes graph‐based clusters. In the bubble plot, the size of the circles represents the percentage of cells in the cluster expressing the indicated gene, and the color represents the level of expression

FIGURE 4

Relatedness of up‐regulated genes. The circos plot shows the up‐regulated genes within each cluster that are shared with other clusters (dark orange) and those that are unique to that cluster (light orange). The purple chords connect the clusters sharing up‐regulated genes

FIGURE 5

Identification of SIV‐infected cells. (A) Dot plot of cells without SIV transcripts (not infected) and those with transcripts from the SIV genome (SIV infected), separated by, and colored by, condition. The 2 infected cells in SIV untreated are indicated by arrows. (B) As in (A), except separated by graph‐based cluster. (C) Dot plot overlayed upon violin plot showing cells expressing SIV RNA from the gag‐pol region (indicated full‐length transcript), separated by, and colored by, graph‐based cluster

FIGURE 6

Cluster analysis of canonical IPA pathways. Euclidean clustering of activation Z‐scores for canonical IPA pathways in each graph‐based cell cluster is represented in the heatmap. Dendrograms illustrate the linkage relationships and distances between rows (pathways) and columns (graph‐based cell clusters)

FIGURE 7

IPA reveals upstream regulators linked to infection status. Euclidean clustering of activation Z‐scores in each graph‐based cell cluster are represented in the heatmaps for upstream regulator classes (A) general regulation (transcriptional regulation), (B) neuroinflammation (cytokine), and (C) neuroprotection, cell survival, and signal transduction (growth factor). Dendrograms illustrate the linkage relationships and distances between rows (upstream regulator) and columns (graph‐based cell clusters)