Background/Purpose: Although millions of disease-associated variants have been discovered using GWAS, most of these occur in non-coding regions and their mechanistic link to disease is unclear. Further, separating causal from neutral variants within haplotypes remains challenging. Analyses of expression quantitative trait loci (eQTLs) in healthy tissues only explain some GWAS hits due to the different mechanisms driving variant discovery in these two approaches. Quantitative trait loci associated with epigenomic phenotypes such as chromatin accessibility (caQTL) and histone modifications (hQTL) may reveal genetically mediated regulatory mechanisms not detected by traditional eQTL analysis. Moreover, latent disease-specific mechanisms encoded in the epigenome may be elicited by stimulating healthy tissues.

Methods: We gathered expression data using RNA-seq, epigenomic data using ATAC-seq and CUTNRUN (for H3K27-acetylation and CTCF-binding), from B cells in 81 genotyped, healthy individuals from both a resting and immune-stimulated condition. Stimulation consisted of a 48-hour treatment with BCR crosslinking, CD40 co-stimulation and TLR7 agonist.

Results: We identified strong effects of stimulation across these “omic” layers, including thousands of differentially expressed genes and tens of thousands of differentially activated epigenomic modifications triggered by stimulation. Pairing this omics data with polymorphism data from our sample, we identified over 400,000 candidate QTLs, with over 1/3 of these exclusive to the stimulated state. We used the DegCre method to connect differentially expressed genes (DEG) with candidate regulatory elements (CRE) from our epigenomic data. In the context of our stimulation approach, DegCre identifies DEG-CRE associations based on the strength of coordinated changes across omic layers induced by stimulation. Using this approach, we identified over 24,000 significant associations between DEGs and CREs, with many genes apparently regulated by multiple CREs. Among omic layers, differential chromatin accessibility induced by stimulation had the most numerous associations with DEGs, with 20,018 distinct regions, followed by 3,276 regions of differential H3K27 acetylation and 858 regions of differential CTCF binding. The spatial distribution of CREs from each omic layer was mostly non-overlapping. We then examined 8,595 Systemic Lupus Erythematosus (SLE) GWAS variants and found that 1,670 of these matched QTLs that we identified in our data, with 201 matching QTLs exclusively identified in the stimulated B cell state. An additional 536 SLE-associated variants, that did not match any of our QTLs, resided within CREs that we identified with DegCre, supporting the idea that non-coding GWAS variants that do not manifest as QTLs may still be functionally annotated to better identify candidate regulatory elements for experimental validation.

Conclusion: Collectively, our results demonstrate the utility of a multi-omic, stimulation-augmented approach for annotating non-coding GWAS variants and identifying regulatory epigenomic modifications.

« Back to ACR Convergence 2024

ACR Meeting Abstracts - https://acrabstracts.org/abstract/activating-b-cells-facilitates-the-discovery-of-latent-disease-associated-variants/