Single-cell transcriptome analysis yields new insights into the pathogenic mechanisms and possible genetic etiology of cardiomyopathies

Abstract

Abstract Background Inherited cardiomyopathies (CM) represent a clinically heterogeneous group of primary cardiac muscle disorders with a strong genetic underpinning. Recent rapid genomic advances have led to the identification of numerous disease-causing genes for both non-syndromic (nsCM) and syndromic (sCM) cardiomyopathies. This has greatly facilitated molecular genetic testing, thus enabling accurate disease diagnosis needed for the practice of precision medicine and the optimization of patient outcome. However, many cardiomyopathies remain unexplained with the known genes and dominant genetic model of disease. Purpose To reassess the genetic features of known CM genes as a strategy to recover novel candidate CM genes. Methods Known hypertrophic CM (HCM), dilated CM (DCM) and pediatric CM genes were curated from the literature and from 23 commercial CM diagnostic panels. They were classified as non-syndromic and syndromic, and further annotated using two constraint metrics, the missense Z score and pLI score obtained from the GnomAD database. Publicly available mouse (n=6) and human (n=3) single-cell RNA (scRNA) datasets were downloaded and cardiomyocyte specific differentially expressed genes (DEGs) (fold change >0.1; adjusted P<0.0001) were recovered. Genes identified as DEGs in at least 4 mouse or 3 human scRNA datasets were recovered as nsCM candidate genes. Results Analyses of 9 scRNA datasets showed the majority of known nsCM genes are cardiomyocyte specific (Fig. 1a-b). nsCM and sCM genes have distinct expression and genetic profile. nsCM genes are associated with higher heart expression and lower loss intolerance (Fig. 1c). In contrast, syndromic CM genes mostly showed lower expression with high loss intolerance, consistent with their higher clinical impact. However, interestingly some of the nsCM genes (MYLK2, TMPO and KLF10) show low or even no detectable expression in mouse and human cardiac cells. Using the scRNA data, we assessed cellular expression of genes in the 23 commercial CM diagnostic panels. This analysis showed some of the CM genes with high cardiomyocytes expression have low coverage on the current commercial CM panels (Fig. 1d). Using human and mouse scRNA data, we recovered 224 mouse and 157 human nsCM candidate genes. MTUS2 (microtubule-associated tumor suppressor candidate 2) was identified as a strong nsCM candidate gene supported by evidence from both mouse and human studies (Fig. 1e-f). Conclusions Our analysis showed many of the nsCM genes have differential cardiomyocyte expression with low loss intolerance, while the reverse was observed for many sCM genes. We propose increasing commercial panel coverage of cardiomyocytes-specific expressed genes may help increase disease diagnostic yield. Additionally, novel candidate genes uncovered trained on cardiomyocyte expression profile may help accelerate elucidation of unsolved cardiomyopathy cases. Funding Acknowledgement Type of funding sources: None.