Integrated phenotypic and transcriptomic characterization of desmin-related cardiomyopathy in hiPSC-derived cardiomyocytes and machine learning-based classification of disease features

Abstract

Desmin-related diseases are characterized by skeletal muscle weakness, cardiomyopathy, and respiratory dysfunction due to mutations in the desmin gene (DES), which encodes a protein essential for muscle cell integrity. This study investigates the effects of a pathogenic desmin mutation (DES^E439K) in human cardiomyocytes derived from human induced pluripotent stem cells (hiPSCs) obtained from a patient carrying the DES^E439K mutation, compared to cardiomyocytes derived from hiPSCs of three healthy donors. To further validate our findings a genome edited cell line has been obtained following the insertion of the mutation in a control hiPSC line. Using advanced technologies, including transcriptomics and phenotypic machine learning algorithms, we analyzed how this mutation disrupts cellular function and contributes to disease phenotypes. Our findings reveal that cardiomyocytes carrying DES^E439K exhibit cytoplasmic protein aggregation, mitochondrial and sarcomere defects, and contractile dysfunctions, highlighting key phenotypic defects in desmin-related cardiomyopathy. Finaly, we developed a machine learning prediction model to classify cellular phenotypes, which can be used for translational research, including drug candidate screening. This research opens new avenues for understanding the molecular mechanisms of desmin-related cardiomyopathies and fosters the development of novel therapeutic strategies.

Keywords: Cardiomyocyte; Desminopathy; Dilated Cardiomyopathy; Disease Modeling; Machine Learning; Myofibrillar Myopathy.