Immune disease risk variants regulate gene expression dynamics during CD4+ T cell activation

Abstract

During activation, T cells undergo extensive changes in gene expression which shape the properties of cells to exert their effector function. Therefore, understanding the genetic regulation of gene expression during T cell activation provides essential insights into how genetic variants influence the response to infections and immune diseases. We generated a single-cell map of expression quantitative trait loci (eQTL) across a T cell activation time-course. We profiled 655,349 CD4+ naive and memory T cells, capturing transcriptional states of unstimulated cells and three time points of cell activation in 119 healthy individuals. We identified 38 cell clusters, including stable clusters such as central and effector memory T cells and transient clusters that were only present at individual time points of activation, such as interferon-responding cells. We mapped eQTLs using a T cell activation trajectory and identified 6,407 eQTL genes, of which a third (2,265 genes) were dynamically regulated during T cell activation. We integrated this information with GWAS variants for immune-mediated diseases and observed 127 colocalizations, with significant enrichment in dynamic eQTLs. Immune disease loci colocalized with genes that are involved in the regulation of T cell activation, and genes with similar functions tended to be perturbed in the same direction by disease risk alleles. Our results emphasize the importance of mapping context-specific gene expression regulation, provide insights into the mechanisms of genetic susceptibility of immune diseases, and help prioritize new therapeutic targets.

Dynamics, signals and influencing factors of CD4 T cell activation revealed by single cell RNA-seq

“…There are 6 T cell subsets mainly from resting CD4 T cells, namely T N ( CCR7 hi , SELL hi ), T CM ( CCR7 + , SELL + , ANXA1 + ), T EM ( CCR7 low , ANXA1 + , TIMP1 + ), CD4 T EMRA ( ANXA1 + , CST7 + , GZMK + , GZMA + ), Treg ( FOXP3 + , CTLA4 + ) and IFN signaling associated gene (ISAG) highly expressed T cells (ISAG hi T) ( ISG15 hi , CCR7 + , SELL + ) (Fig 1, D and E). These T cell subsets are essentially consistent with recent studies on peripheral CD4 T cells ( 4, 20, 26 ). There are 5 T cell subsets mainly from stimulated CD4 T cells, four of which belong to CD4 T EFF ( IL2RA + , NME1 + and MIR155HG + , FABP5 + ) (Fig 1D, E).…”

“…CD4 T N differentiates into Th2 with IL-4 polarization, while CD4 T M could not differentiate into Th2 phenotype. Soskic, et al identified lots of genetic variants that regulate gene expression dynamics during CD4 T cell activation ( 26 ). However, many factors affecting T cell activation have not been explored.…”

Modeling tissue co-regulation to estimate tissue-specific contributions to disease

“…However, genome-wide results from TCSC may be used as a prior for locus-based methods (analogous to GWAS fine-mapping with functional priors 97 ). Finally, we did not apply TCSC to single-cell RNA-seq (scRNA-seq) data, which represents a promising new direction as scRNA-seq sample sizes increase [98][99][100]32 ; we caution that scRNA-seq data may require new eQTL modeling approaches 98 . Despite these limitations, TCSC is a powerful and generalizable approach for modeling tissue co-regulation to estimate tissuespecific contributions to disease.…”

maxATAC: genome-scale transcription-factor binding prediction from ATAC-seq with deep neural networks

“…We observed increased expression of all three genes (RTEl1, RTEL1-TNFRSF6B , TNFRSF6B) associated with rs6062490, when comparing the single pair of homozygous-risk AD to homozygous non-risk control, but we observed inconsistent trends from the two donor pairs in which the AD patients were homozygous risk and controls were heterozygous non-risk. These trends, in addition to support for SLC2A4RG as an eGene in CD4 + T cells [53–55], nominate potential molecular mechanisms, specific TFs that might alter the expression of genes important to T cell function (MYB and FOXP1) in AD patients. Interestingly, for the regions shown in Fig.…”

“…6E ). LIME1 is a transmembrane protein that controls effector T cell migration to sites of inflammation[52], while SLC2A4RG is a known eGene, (associated with an eQTL) in activated CD4 + T cells [53], regulatory T cells [54] and multiple other T cell subsets [55]. Thus, genes at both risk loci have ties to T cell biology, T-cell eQTLs and AD.…”