Abstract 4127919: Disease-associated variants in <i>DSP</i> -encoded desmoplakin are common, yet low penetrant alleles associated with development of myocarditis and cardiomyopathy at a population level

Intro: DSP- encoded desmoplakin is implicated in a distinct form of arrhythmic cardiomyopathy (ACM), often involving the LV and characterized by myocardial inflammation. Conventional ACM diagnostic criteria has poor sensitivity for DSP -associated ACM, thus highlighting the need to identify the full spectrum of phenotypic risk associated with disease-associated DSP variants. Objective: To analyze the genotypic and phenotypic spectrum of DSP -mediated disease at a large population level. Methods: UK Biobank participants with exome sequencing (ES) were included. Variants were filtered by gene evidence category, yielding 3 groups: predicted deleterious (pDel); subset with predicted loss of function (LOF); and ultra-filtered subset by ClinVar 2* pathogenic/likely pathogenic (P/LP). Phenotypic penetrance was analyzed with DSP -neg individuals as control. Variant location analysis assessed LOFs by susceptibility to nonsense mediated decay region and missenses by amino acid sequence to identify mutational hotspots where P/LPs localize. Results: Out of 200,580 with ES, 1407 DSP carriers had pDel, 168 with LOF, and 44 with ClinVar 2* P/LP. DSP carriers had higher burden of myocarditis, cardiomyopathy (CM), and heart failure. A progressive enrichment in myocarditis and CM was observed by more stringent variant filtering in DSP carriers compared to control. A higher proportion of DSP carriers had myocarditis – 0.28% (4) of the pDels, 1.8% (3) of LOFs, and 4.5% (2) compared to control 0.07% (p&lt;0.05). A parallel trend noted the proportion of DSP carriers with CM growing from 1.4% (19) to 5.4% (9) and 6.8% (3) in the pDel, LOF, and ClinVar 2* P/LPs respectively compared to 0.6% in the control group (p&lt;0.01). Location analysis of missense variants revealed that P/LPs localized in higher proportions (74%) to key domains compared to non-P/LPs (50%; p&lt;0.001), highlighting prognostic value of predicting pathogenicity of missense DSP variants by mutation hotspots. Conclusions: While DSP variants are relatively common in the population, variant presence is associated with low penetrance. Enhancing clinical penetrance with increasingly stringent filtering of DSP variants emphasizes the importance of variant evaluation in clinical risk assessment. This first-in-kind population study of DSP provides insights into novel risk prediction factors, allowing us to predict pathogenicity by type and location of mutation and using variant filtering strategy to better assess phenotypic penetrance.