Cardiac Cell Type-Specific Gene Regulatory Programs and Disease Risk Association

Abstract

BackgroundCis-regulatory elements such as enhancers and promoters are crucial for directing gene expression in the human heart. Dysregulation of these elements can result in many cardiovascular diseases that are major leading causes of morbidity and mortality worldwide. In addition, genetic variants associated with cardiovascular disease risk are enriched within cis-regulatory elements. However, the location and activity of these cis-regulatory elements in individual cardiac cell types remains to be fully defined.MethodsWe performed single nucleus ATAC-seq and single nucleus RNA-seq to define a comprehensive catalogue of candidate cis-regulatory elements (cCREs) and gene expression patterns for the distinct cell types comprising each chamber of four non-failing human hearts. We used this catalogue to computationally deconvolute dynamic enhancers in failing hearts and to assign cardiovascular disease risk variants to cCREs in individual cardiac cell types. Finally, we applied reporter assays, genome editing and electrophysiogical measurements in in vitro differentiated human cardiomyocytes to validate the molecular mechanisms of cardiovascular disease risk variants.ResultsWe defined >287,000 candidate cis-regulatory elements (cCREs) in human hearts at single-cell resolution, which notably revealed gene regulatory programs controlling specific cell types in a cardiac region/structure-dependent manner and during heart failure. We further report enrichment of cardiovascular disease risk variants in cCREs of distinct cardiac cell types, including a strong enrichment of atrial fibrillation variants in cardiomyocyte cCREs, and reveal 38 candidate causal atrial fibrillation variants localized to cardiomyocyte cCREs. Two such risk variants residing within a cardiomyocyte-specific cCRE at the KCNH2/HERG locus resulted in reduced enhancer activity compared to the non-risk allele. Finally, we found that deletion of the cCRE containing these variants decreased KCNH2 expression and prolonged action potential repolarization in an enhancer dosage-dependent manner.ConclusionsThis comprehensive atlas of human cardiac cCREs provides the foundation for not only illuminating cell type-specific gene regulatory programs controlling human hearts during health and disease, but also interpreting genetic risk loci for a wide spectrum of cardiovascular diseases.

A cell atlas of chromatin accessibility across 25 adult human tissues

“…Together, these findings are in line with recent characterizations of unexpected cardiogenic gene programs in cardiac fibroblasts (Furtado et al, 2014). We finally compared subtype-specific cCREs with chromatin profiles from in vitro cultured fibroblast biosamples and cardiac fibroblasts from sci-ATAC-seq (Hocker et al, 2020). While all fibroblast subtypes from the current study showed similarity to in vitro fibroblasts based on core fibroblast cCRE signatures, only the fibroblast subtype Fib.3 matched previously reported cardiac fibroblasts based on subtype-specific fibroblast cCRE signatures ( Figure 5F), suggesting that fibroblast subtypespecific signatures are environment dependent and may be lost during in vitro culturing.…”

“…We thus optimized nuclear isolation methods and buffer conditions for each tissue type (Table S2, see Methods). Subsequently, we generated sci-ATACseq datasets using a semi-automated workflow (Hocker et al, 2020;Li et al, 2020;Preissl et al, 2018) and sequenced resulting libraries to 7,651 raw sequence reads per nucleus on average, with a median read duplication rate of 44% (Table S3). Open chromatin fragments from these libraries were computationally assigned to individual nuclei using nucleus-specific DNA barcodes.…”

“…Genetic variants associated with complex diseases and traits from GWAS predominantly reside in non-coding regions of the genome (Claussnitzer et al, 2020) and are enriched in cCREs in a tissue and cell type-specific fashion (Corces et al, 2020;Cusanovich et al, 2018;Domcke et al, 2020;Hocker et al, 2020;Maurano et al, 2012;Song et al, 2020;Song et al, 2019). To examine the genome-wide enrichment of disease and trait associated variants within cCREs annotated in each of the 54 human cell types characterized in the current study, we performed cell typestratified linkage disequilibrium score regression (LDSC) analysis using GWAS summary statistics for 56 phenotypes including diseases and non-disease traits ( Figure 6A-B, Table S9, See Methods).…”

“…The resulting data can be used to deconvolute cell types from mixed cell populations and to dissect cell type-specific transcriptomic and epigenomic states in primary tissues. While these tools have been applied to mammalian tissues including murine biosamples (Cusanovich et al, 2018;Lareau et al, 2019;Li et al, 2020;Preissl et al, 2018;Sinnamon et al, 2019), human fetal tissues (Domcke et al, 2020), and individual adult human organ systems (Chiou et al, 2019;Corces et al, 2020;Hocker et al, 2020;, we still lack comprehensive maps of cCREs in the cell types comprising primary tissues of the adult human body.…”

“…In the present study we used a modified single-cell combinatorial indexing ATAC-seq (sci-ATACseq) protocol optimized for flash frozen primary tissues (Hocker et al, 2020;Preissl et al, 2018) to profile chromatin accessibility in 25 adult human tissue types from multiple donors. We profiled 472,373 nuclei from these tissues, grouped them into 54 cell types based on similarity in chromatin landscapes, and identified a union of 756,414 open chromatin regions and candidate CREs (cCREs) from the resulting maps.…”