Mattivi, Connor L.

Introduction: Loss-of-function (LOF) and gain-of-function (GOF) pathogenic variants in the KCNH2 -encoded Kv11.1 potassium channel cause type 2 long QT syndrome(LQT2) and type 1 short QT syndrome (SQT1), respectively. Given the prevalence of KCNH2- mediated heart disease (1 in 8000 for LQT2,1 in 100,000 for SQT1) and the reduced penetrance/variable expressivity seen in LQTS and SQTS, large databases such as the Genome Aggregation Database (gnomAD, n=141,352 individuals) might harbor subjects with unidentified LQT2/SQT1-causative mutations. Objective: To determine which “ultra-rare” KCNH2 missense variants (MVs) in gnomAD are most likely to be LQT2- or SQT1-causative. Methods: A list of gnomAD-derived MVs in KCNH2 was compiled. The frequency of each MV in gnomAD was assessed. Each MV was analyzed by 8 in silico pathogenicity prediction tools. MVs that were seen once in gnomAD and predicted damaging by all 8 tools were considered possible LQT2/SQT1-causative MVs worthy of functional validation via whole-cell patch clamp. Results: Overall, 233/474 (49%) of gnomAD KCNH2 MVs were seen in just 1 of >140,000 subjects. Of these, 2 (0.4%) MVs (S320W, S379Y) were predicted damaging by 8 in silico tools. Both S320W and S379Y localize to the N-terminus of the Kv11.1 potassium channel. There was no significant change in peak or tail current density between heterozygously expressed KCNH2-S320W (n=13, p=NS) and KCNH2-WT (n=13). However, heterozygous KCNH2-S379Y (n=12) produced a marked increase in peak current density across the range -30 mV to +20 mV and shifted the V 1/2 of activation by -17 mV (p < 0.001). Also, tail current density of S379Y channels significantly increased across the range -20 mV to +20 mV and the V 1/2 of deactivation shifted by -26 mV (p < 0.001). Conclusion: With this stringent requirement of ultra-rarity, universal prediction of damaging impact, and functionally abnormal electrophysiological properties, one SQT1-causative pathogenic variant in gnomAD was identified consistent with the predicted frequency of SQTS. The search continues among the remaining 32 ultra-rare MVs in KCNH2 where at least 75% of the in silico tools predicted deleterious impact of the MV to find the 15-20 subjects in gnomAD’s population who ought to have an LQT2-causative MV.

Introduction: Loss-of-function (LOF) variants in the KCNH2 -encoded Kv11.1 potassium channel cause type 2 long QT syndrome (LQT2). To date, hundreds of KCNH2 missense variants (MVs) have been published as “disease-causative”. However, it has been suggested that 10% of rare LQTS variants in the literature may be "false positives”. Objective: To determine previously published KCNH2 variants that are likely false positives and warrant demotion to at least likely benign status. Methods: A list of all case-derived LQT2-associated MVs from six large published compendia was compiled. The frequency of each MV within the Genome Aggregation Database (gnomAD, n = 141,352 individuals) was recorded. Eight in silico variant assessment tools were used to grade each MV. Variants (n=8) absent in gnomAD but predicted “benign” by all 8 tools were considered potential false positives, despite being ultra-rare, and were characterized functionally using whole-cell patch clamp. Results: Overall, 339 KCNH2 MVs were identified among LQTS cases. Of these, 13 (4%) were seen at a minor allele frequency (MAF) > 0.0004 in gnomAD. This MAF cut-off is derived from the frequency of R176W-KCNH2, a well-established albeit weakly penetrant (i.e. < 20%) LQT2-causative MV. Variants seen at MAF > R176W-KCNH2 are unlikely to be LQT2-contributing. However, 253 MVs (75%) were absent in gnomAD. Of these, 8 (4%) MVs (I96V, G187S, A203T, P241L, H254Q, G314S, P935S, and P963T) were predicted by all 8 in silico tools to be benign and had never been characterized functionally. Patch clamp studies did not demonstrate any LOF perturbation for these 8 MVs. Conclusion: This study offers compelling evidence for the demotion from LQT2-causative status to likely benign status for 21 (6%) of the 339 previously published LQT2 MVs based either on i) gnomAD MAF unacceptably high for a 1:8000 disorder (LQT2) for 13 MVs or ii) ultra-rarity (never seen in gnomAD) but in silico variant assessment tools predicting benign impact of the amino acid substitution and in vitro functional validation studies devoid of LOF for 8 additional MVs. This meticulous “pruning” exercise must be conducted for all published variants previously implicated as the monogenic cause for LQTS in particular and all genetic heart diseases in general.

Background: Missense variants in the MYH7 -encoded MYH7 (beta myosin heavy chain 7) represent a leading cause of hypertrophic cardiomyopathy (HCM). MYH7 -specific American College of Medical Genetics and Genomics (ACMG) variant classification guidelines were released recently but have yet to be assessed independently. We set out to assess the performance of the MYH7 -specific ACMG guidelines and determine if the addition of phenotype-enhanced criteria (PE-ACMG) using the HCM Genotype Predictor Score can further reduce the burden of variants of uncertain significance (VUS). Methods: Re-assessment was performed on 70 MYH7 -variants in 121 unique patients from Mayo Clinic, and an independent cohort of 54 variants in 70 patients from Royal Prince Alfred Hospital (Australia). Qualifying variants were re-adjudicated using both standard ACMG and MYH7 -ACMG guidelines, and HCM Genotype Predictor Score was used to provide a validated measure of strength of clinical phenotype to be incorporated into the MYH7 -ACMG framework. Results: Among Mayo Clinic identified variants, 11/70 (16%) were classified as pathogenic (P), 10/70 (14%) as likely pathogenic, and 49/70 (70%) as a VUS. A similar distribution was seen in the Australian patients (12/54 [22%] P, 12/54 [22%] likely pathogenic, and 30/54 [56%] VUS; P =not significant). Application of the MYH7 -ACMG resulted in a nonsignificant reduction of the VUS burden in both cohorts from 49/70 to 39/70 (56%; P =0.1; Mayo Clinic) and from 30/54 to 20/54 (37%; P =0.1; Australia). Using the combined PE-MYH7-ACMG framework, the VUS decreased significantly from 49 to 27 ( P <0.001, Mayo Clinic) and from 30 to 16 ( P <0.001; Australia). Conclusions: Use of the MYH7 -specific guidelines alone failed to significantly decrease VUS burden in 2 independent cohorts. However, a significant reduction in VUS burden was observed after the addition of phenotypic criteria. Using a patient’s strength of sarcomeric HCM phenotype for variant adjudication can increase significantly the clinical utility of genetic testing for patients with HCM.