EP291: Disease characterization in sodium-potassium ATPases by reverse genetics in humans

Abstract

Most genetic disorders are discovered via forward genetics: a shared phenotype is recognized in a group of patients, allowing causative gene(s) to be determined. This can lead to excessively narrow definitions of disease phenotypes due to selection bias, as patients with symptoms outside the “classic” presentation are less likely to be diagnosed. We hypothesized that reverse human genetics, ie, studying variants in a gene without phenotype-based recruitment, could more accurately characterize new diseases. We established the first such approach for ATP1A1, the ubiquitously expressed housekeeping isoform of sodium-potassium ATPase. ATP1A1-associated disease is an emerging entity with several reports of heterozygous variants in individuals with differing neurological phenotypes: peripheral neuropathy, childhood epilepsy-renal hypomagnesemia, or spastic paraplegia. Gain-of-function ion leakage has been proposed as a mechanism for the more severe phenotype in the patients with epilepsy, versus loss-of-function in patients with neuropathy. However, only 14 families in total have been reported with possible ATP1A1 pathogenic variants and the true spectrum of disease remains unknown. Through genotype-based recruitment combined with detailed functional characterization of variants, we broadened the phenotypic spectrum to include incomplete penetrance and identified distinct molecular categories of ATP1A1 variants. We collected variants identified on next-generation sequencing of both clinical and research samples. We recruited patients with heterozygous ATP1A1 variants of uncertain significance found on testing for any indication. In addition, we filtered rare variants found in 7,000 “healthy” volunteers in a population genome sequencing study and prioritized 2 for study based on predicted functional impact. We re-contacted one individual from this cohort for targeted phenotyping at the NIH Clinical Center under an IRB-approved study protocol (ClinicalTrials.gov identifier NCT00369421). We developed a comprehensive panel of 3 assays measuring the ability of ATP1A1 to support survival in transfected cells and its total enzymatic activity and turnover rate per molecule in Xenopus oocytes. In the latter experimental system, we directly measured enzyme activity as steady-state electrical currents in two-electrode voltage clamp. Using sodium transient current measurements, we quantified sodium ion substrate binding by ATP1A1 as well as the number of ATP1A1 molecules reaching the cell surface. We identified 7 rare missense variants with significant impacts on enzyme function, as well as 1 nonsense variant predicted to lead to nonsense-mediated decay. Of 9 missense variants identified in clinically ascertained people, 6 had functional impact and 3 did not. Phenotypes of patients with functionally significant ATP1A1 variants included not only neurological deficits but also previously unreported features such as hypothyroidism (two patients), skin changes, and unusual facial features. Of the 2 variants selected from the population sequencing cohort, 1 was a missense variant with clear loss-of-function impact in multiple assays yet was present in an adult without neurological symptoms. This individual declined research phenotyping. The other was a nonsense variant identified in an asymptomatic adult. This individual underwent detailed phenotyping including nerve conduction/electromyography and was confirmed not to have subclinical neuropathy, magnesium wasting, or spasticity. Molecular characterization of the variants revealed three general patterns of impact on enzyme function. First, some variants decreased protein abundance at the cell membrane with a corresponding loss of enzyme function. Second, other variants resulted in ATP1A1 that reached the cell membrane but had decreased enzyme activity with or without ion leak currents. Third, some variants changed enzyme properties by reducing the voltage dependence of its activity, possibly suggesting gain-of-function at negative voltages. Genotype-based collection of variants for functional characterization was an effective strategy to study the spectrum of ultra-rare ATP1A1 disease. We identified 8 putative pathogenic variants, expanding the total known pool of such variants by more than half. Connecting functional impacts of variants to the phenotypes of the individuals carrying them revealed incomplete penetrance into adulthood in two individuals as well as possible non-neurological disease features for further study. Detailed molecular characterization allowed us to propose distinct classes of variants based on their effects on enzyme function, which could form the basis for future genotype-phenotype correlation.