Global chromatin landscapes identify candidate noncoding modifiers of cardiac rhythm
Samadrita Bhattacharyya, … , Ralf Kittler, Nikhil V. Munshi
Samadrita Bhattacharyya, … , Ralf Kittler, Nikhil V. Munshi
Published December 1, 2022
Citation Information: J Clin Invest. 2023;133(3):e153635. https://doi.org/10.1172/JCI153635.
View: Text | PDF
Research Article Cardiology Article has an altmetric score of 45

Global chromatin landscapes identify candidate noncoding modifiers of cardiac rhythm

Abstract

Comprehensive cis-regulatory landscapes are essential for accurate enhancer prediction and disease variant mapping. Although cis-regulatory element (CRE) resources exist for most tissues and organs, many rare — yet functionally important — cell types remain overlooked. Despite representing only a small fraction of the heart’s cellular biomass, the cardiac conduction system (CCS) unfailingly coordinates every life-sustaining heartbeat. To globally profile the mouse CCS cis-regulatory landscape, we genetically tagged CCS component–specific nuclei for comprehensive assay for transposase-accessible chromatin–sequencing (ATAC-Seq) analysis. Thus, we established a global CCS-enriched CRE database, referred to as CCS-ATAC, as a key resource for studying CCS-wide and component-specific regulatory functions. Using transcription factor (TF) motifs to construct CCS component–specific gene regulatory networks (GRNs), we identified and independently confirmed several specific TF sub-networks. Highlighting the functional importance of CCS-ATAC, we also validated numerous CCS-enriched enhancer elements and suggested gene targets based on CCS single–cell RNA-Seq data. Furthermore, we leveraged CCS-ATAC to improve annotation of existing human variants related to cardiac rhythm and nominated a potential enhancer-target pair that was dysregulated by a specific SNP. Collectively, our results established a CCS-regulatory compendium, identified novel CCS enhancer elements, and illuminated potential functional associations between human genomic variants and CCS component–specific CREs.

Authors

Samadrita Bhattacharyya, Rahul K. Kollipara, Gabriela Orquera-Tornakian, Sean Goetsch, Minzhe Zhang, Cameron Perry, Boxun Li, John M. Shelton, Minoti Bhakta, Jialei Duan, Yang Xie, Guanghua Xiao, Bret M. Evers, Gary C. Hon, Ralf Kittler, Nikhil V. Munshi

×

Figure 5

Using CCS-ATAC to improve human GWAS SNP annotation.

Options: View larger image (or click on image) Download as PowerPoint
Using CCS-ATAC to improve human GWAS SNP annotation. (A) Pipeline for id...
(A) Pipeline for identification of syntenic cardiac rhythm-related SNPs by LiftOver to mouse reference genome (mm10) and overlap with CCS enhancers (Figure 4A). (B) Bar graph indicating the probability (in log10) of a cardiac rhythm–related SNP landing in the mouse genome (mm10), CM enhancers, and CCS enhancers. Statistical significance (P value < 0.001) by χ2 test of trends for the pairwise comparisons are shown. Fold enrichment for SNP probability relative to CM enhancers is labeled. (C) Workflow for comparing GWAS SNPs associated with individual EKG traits and CCS component-specific enhancers. (D) Bar graph demonstrating percentage of specific EKG trait SNPs landing within CCS specific enhancers. P values indicate statistical significance by χ2 test for pairwise comparisons with the CM enhancer subset. For QRS interval, AVN and VCS data sets were combined (green and red striped bar). (E) Genome browser view showing location of HR lead SNP rs867400 (mm10: chr10:121498835-121498836), which lies intergenic to the Tbk1 and Rassf3 genes, in relation to CCS-ATAC open regions. Reference and minor SNP alleles are indicated. (F) SAN-enriched genes within ± 500 kb of rs867400 are indicated by solid box. (G) UMAP plot of Rassf3 gene overlaid upon SAN scRNA-Seq atlas. SAN (blue) and transitional (brown) cells are indicated. (H) Bar graph represents RLUs of enhancer containing rs867400 relative to empty luciferase in primary mouse SAN cells. Error bars illustrate SEM of luciferase expression between 2 independent experiments. Ordinary 1-way ANOVA test was used to calculate P values. (I) Genome browser view showing location of QRS lead SNP rs12764182 (mm10: chr14:22666305-22666306), which lies within an intron of the Lrmda gene, in relation to CCS-ATAC open regions. Reference and minor SNP alleles are indicated. (J) AVCS-enriched genes within ± 500 kb of rs12764182 are indicated by solid box. (K) UMAP plot of Vcl gene overlaid upon AVCS scRNA-Seq atlas. Cells comprising compact AVN (red) and AVB (green) are indicated. (L) Bar graph represents RLUs of enhancer containing rs12764182 relative to empty luciferase in primary mouse AVCS cells. Error bars illustrate SEM of luciferase expression between 2 independent experiments. Ordinary 1-way ANOVA was used to calculate P values. (M) Genome browser view showing location of Q-T lead SNP rs2074238 (mm10: chr7:143122498-143122499), which lies within an intron of the Kcnq1 gene, in relation to CCS-ATAC open regions. Reference and minor SNP alleles are indicated. (N) AVB-enriched genes within ± 500 kb of rs2074238 are indicated by solid box. (O) UMAP plot of Kcnq1 gene overlaid upon AVCS scRNA-Seq atlas. AVB cells (green) are indicated. (P) Bar graph represents RLUs of enhancer containing rs2074238 relative to empty luciferase in primary mouse AVCS cells. Error bars illustrate SEM of luciferase expression between 2 independent experiments. Ordinary 1-way ANOVA test was used to calculate P values. (Q) Genome browser view showing location of PR lead SNP rs3807989 (mm10: chr6:17325447-17325448), which lies within an intron of the Cav1 gene, in relation to CCS-ATAC open regions. Reference and minor SNP alleles are indicated. (R) AVN-enriched genes within ± 500 kb of rs3807989 are indicated by solid box. (S) UMAP plot of Cav1 gene overlaid upon AVCS scRNA-Seq atlas. AVN cells (red) are indicated. (T) Sequence and evolutionary conservation surrounding rs3807989 is shown with SNP location highlighted in red. Matching SCRT1/2 consensus binding site logo is shown beneath for comparison. (U) Bar graph represents RLUs of reference and minor allelic variant for PR SNP relative to empty luciferase in primary mouse AVCS cells. Error bars illustrate SEM of luciferase expression between 2 independent experiments. Ordinary 1-way ANOVA test was used to calculate P values. (V) Bar graphs showing genomic localization by ChIP-qPCR fold-enrichment for Scrt1 compared with IgG control at the Cav1 locus. Error bars illustrate SEM of target gene expression among 3 independent experiments. Tubb3 served as a negative control. *P < 0.05; **P < 0.01; ***P < 0.005. For each SNP, lifted over mm10 coordinates are used to generate the ATAC tracks. Blue dotted lines below coordinates indicate SNP location. HR, heart rate; n.s., not significant; PWM, position weight matrix.