Computational Pipeline for Identification of Disease-Causing Variants in Genes of the Cardiac Sarcomere

PROJECT SUMMARY / ABSTRACT
Contractile force in the heart is generated by densely packed subcellular structures known as sarcomeres.
Mutations in sarcomeric genes have been repeatedly linked to potentially lethal conditions known as
cardiomyopathies, including hypertrophic (HCM) and dilated (DCM) forms. Optimal clinical management of
HCM/DCM requires identification of at-risk individuals before they experience symptoms. Genetic testing can
be useful, but results are not always definitive enough to support clinical decision making. This is because
genetic variants found in patients are often unique to their family. These so-called variants of unknown
significance (VUS) could be pathogenic or benign. Unfortunately, testing each VUS experimentally is
prohibitively expensive, and generic pathogenicity algorithms are proving unreliable for prediction of
cardiomyopathies. The goal of this proposal is to create an accurate and scalable computational method for
classifying sarcomeric variants of unknown significance so that more HCM/DCM families can benefit from
genetic screening and early diagnosis. Our long-term approach to solving this critical shortcoming is to create a
computational pipeline to predict pathogenicity of novel sarcomeric gene variants, providing cardiologists with a
biophysical basis for performing risk stratification in HCM/DCM patient families. Our work during the last
funding period was focused specifically on mutations to the protein tropomyosin (Tpm), leading to important
milestones in genotype-phenotype predictive modeling. For this renewal, we aim to expand these capabilities
to include characterization of highly prevalent VUS in Tpm’s binding partners, troponin I (TnI) and troponin T
(TnT). This will widen the impact of our work, encompassing families with VUS in TPM1, TNNI3, or TNNT2.
Breakthroughs documented in our recent published work on the actin/Tpm/troponin regulatory complex have
allowed us to construct increasingly precise maps of the binding interactions among these proteins at an
atomic level, including the specific amino acid sidechains upon which binding and regulatory function depend.
Our hypothesis is that these refined structural interaction maps will allow us to make more accurate predictions
of thin filament VUS pathogenicity using our computational pipeline. We will test this hypothesis in three aims.
Aim 1 experiments will extend our preliminary tests of the interacting pairs hypothesis through the study of 18
additional mutations scattered strategically across our three proteins of interest. In Aim 2, twelve mutations in
TPM1, TNNT2, and TNNI3 that are known to produce clinical disease in humans will be analyzed in our dual
computational/experimental pipeline. These real-world cases will make it possible to define what constitutes a
meaningful mutation-induced change in muscle function. Having established model accuracy (Aim 1) and
thresholds of pathogenicity (Aim 2), in Aim 3 we will perform a computational screen of 200+ VUS from the
ClinVar database and validate twelve of these HCM/DCM pathogenicity predictions in engineered heart
tissues. This project will continue our successful efforts to achieve sarcomeric genotype-phenotype predictions.

Grant Number: 5R01HL136590-08
NIH Institute/Center: NIH
Principal Investigator: STUART CAMPBELL