Abstract 11153: Massively Parallel, High-Throughput Approach for Identifying Defective KCNH2 (hERG) Variants

Introduction: Individuals with congenital long QT syndrome (LQTS) are at high risk for the life-threatening arrhythmia torsades de pointes . Loss-of-function KCNH2 (aka hERG ) variants that reduce repolarizing current, I Kr , are responsible for >30% of LQTS cases with ~6% of sudden death in the young and ~1% in infants. Most KCNH2 variants are observed in some individual(s) making it difficult to obtain clinical evidence to assign pathogenicity; >80% of KCNH2 missense variants in ClinVar are listed as variants of uncertain significance. To help adjudicate variants lacking in compelling clinical data, we demonstrate a method to characterize KCNH2 variant dysfunctions. Methods: We combined deep mutational scanning with a HA tagged KCNH2 construct to enable massively parallel characterization of channel trafficking of missense variants in KCNH2 . We have applied this method to the first 160 residues of KCNH2 ; 3,123 missense variants in total, a known hotspot for disease causing variants. For a subset of 458 non-redundant single nucleotide missense variants, we compared trafficking data, and in silico classifiers, with patch clamp data collected on the SyncroPatch platform. Results: Trafficking data is better than in silico features at predicting peak tail I Kr values, with an area under the receiver operating characteristic curve of > 0.9 for trafficking data versus 0.8 for in silico feature. The Spearman correlation coefficients were 0.73 [0.67-0.78] compared to 0.51 [0.44-0.58] for trafficking and in silico predictions respectively. Our assay identified substitution sensitive regions, surrounded by more permissive residues. We also found substitution intolerant sites involved in networked hydrogen-bonding. Conclusion: A massively parallel trafficking assay can recapitulate most KCNH2 defects more effectively than in silico methods. Extended application of our method will enable the assessment of all missense variants in KCNH2 and other cardiac ion channels.