Single-cell eQTL mapping identifies cell type–specific genetic control of autoimmune disease

The human immune system displays substantial variation between individuals, leading to differences in susceptibility to autoimmune disease. We present single-cell RNA sequencing (scRNA-seq) data from 1,267,758 peripheral blood mononuclear cells from 982 healthy human subjects. For 14 cell types, we identified 26,597 independent cis-expression quantitative trait loci (eQTLs) and 990 trans-eQTLs, with most showing cell type–specific effects on gene expression. We subsequently show how eQTLs have dynamic allelic effects in B cells that are transitioning from naïve to memory states and demonstrate how commonly segregating alleles lead to interindividual variation in immune function. Finally, using a Mendelian randomization approach, we identify the causal route by which 305 risk loci contribute to autoimmune disease at the cellular level. This work brings together genetic epidemiology with scRNA-seq to uncover drivers of interindividual variation in the immune system. Description INTRODUCTION The human immune system has evolved to maintain tissue homeostasis and target exogenous pathogens by regulating specialized cell populations. It displays substantial variation between individuals, defining how people vary in susceptibility to disease and respond to pathogens or cancer. RATIONALE Our knowledge of how genetic differences contribute to immune variation at the cellular level has been limited by two main challenges in the generation of data at single-cell resolution. One of these challenges is to sequence from many individuals and the other is to sequence a large number of cells from each individual. Addressing these challenges is necessary to dissect the genetic and molecular underpinnings of common, heterogeneous diseases. RESULTS We present the OneK1K cohort, which consists of single-cell RNA sequencing (scRNA-seq) data from 1.27 million peripheral blood mononuclear cells (PMBCs) collected from 982 donors. We developed a framework for the classification of individual cells, and by combining the scRNA-seq data with genotype data, we mapped the genetic effects on gene expression in each of 14 immune cell types and identified 26,597 independent cis–expression quantitative trait loci (eQTLs). We show that most of these have an allelic effect on gene expression that is cell type–specific. Our results replicated in two independent cohorts, one of which comprises individuals with a different ancestry to our discovery cohort. Over all loci, our discovery and replication cohorts have a concordance of allelic direction ranging from 72.2 to 98.1% across cell types. Using the top associated eQTL single-nucleotide polymorphism (eSNP) at each locus outside the major histocompatibility complex (MHC) region, we identified 990 trans-acting effects, most (63.6%) of which were cell type–specific. We show how eQTLs have dynamic allelic effects in B cells that are transitioning from naïve to memory states. Overall, we identified a set of 1988 eSNP–eGene (a gene with an eQTL) pairs expressed across the B cell maturation landscape, of which 333 have a statistically significant change in their allelic effect as B cells differentiate. Of these, 66% were only identified from the dynamic eQTL analysis and were not observed when testing for effects independently in cell types, highlighting the importance of investigating cell state–specific effects that underlie immune cell function. We investigated how eQTLs affect the expression variation of essential immune genes in specific cell types and provided experimental support for established hypotheses of cellular mechanisms in complex autoimmune diseases. Finally, we integrated genetic association data for seven common autoimmune diseases and identified significant enrichment of genetic effects operating in a cell type–specific manner. Through colocalization of single-cell eQTL and genome-wide association study (GWAS) loci, we found that 19% of cis-eQTLs share the same causal locus as a GWAS risk association. Using a Mendelian randomization approach, we uncovered the causal route by which 305 loci contribute to autoimmune disease through changes in gene expression in specific cell types and subsets. Of the shared causal loci, 38.4% are outside the MHC region and exhibit highly cell-specific effects. Highlighting multiple sclerosis, we identified the causal route underlying 57 risk loci. For example, we show that the loci at 3q12 causally acts through changes in EAF2 expression, but only in immature and naïve B (BIN) and memory B (BMem) cells, despite this gene being ubiquitously expressed in all cell types in our data. CONCLUSION This work brings together population genetics and scRNA-seq data to uncover drivers of interindividual variation in the immune system. Our results demonstrate how segregating genetic variation influences the expression of genes that encode proteins involved in critical immune regulatory and signaling pathways in a cell type–specific manner. Understanding the genetic underpinnings of immune system regulation will have broad implications in the treatment of autoimmune diseases and infections, transplantation, and cancers. Single-cell eQTL mapping and colocalization with autoimmune disease risk loci. scRNA-seq data from 1.27 million PBMCs were used to identify 26,597 cis-eQTLs (gray box). Dynamic eQTLs were uncovered as cells move from a naïve to a memory state (top right). Genetic variation between individuals influences immune regulation in a cell type–specific manner (middle right). In this study, 990 trans-eQTL effects (bottom right) and the causal effects for 305 autoimmune disease loci were identified (bottom left). Browsable results are available at www.onek1k.org. CD4NC, CD4 naïve and central memory T cells; CD8NC, CD8 naïve and central memory T cells; QC, quality control. Analyzing immune system gene expression Diseases involving the immune system are heritable, but it is unknown how genetic variation contributes to different diseases. To identify how implicated loci affect gene expression in immune cells from individuals from different populations, two groups performed single-cell RNA sequencing of immune cells, with each study investigating hundreds of individuals and more than 1 million immune cells (see the Perspective by Sumida and Hafler). These studies examined both proximal (cis) and distal (trans) genetic variants affecting gene expression in 14 different immune cell types. Perez et al. studied healthy individuals of both European and Asian descent, as well as individuals diagnosed with systemic lupus erythematosus. Yazar et al. performed a population-based study investigating how segregating alleles contribute to variation in immune function. Integrating these data with autoimmune disease cohorts identifies causal effects for more than 160 loci. Both studies observe how gene expression patterns are cell-type and context specific and can explain observed variation in immune cell function among individuals. Both studies also identify causal links between genome-wide analyses and expression quantitative trait loci, identifying potential mechanisms underlying autoimmune diseases. —LMZ Cell type and context affects gene expression affecting immune-mediated disease.