Abstracts

Rationale: The widespread use of genetic and genomic testing has led to an explosive growth in the number of ion channel variants associated with human diseases and in reference populations. In particular, there are more than 2,000 known epilepsy-associated variants in gene encoding voltage-gated ion channels. The >300 known KCNQ2 variants include subgroups associated with self-limiting and severe phenotypes. KCNQ2 encodes a voltage-gated potassium channel (KV7.2) that assembles with KV7.3 (encoded by KCNQ3) to form the channel responsible for neuronal M-current, which regulates neuronal excitability. The ever-increasing number of ion channel variants presents a formidable challenge for assessing their functional and pharmacological consequences. Methods: We developed a CHO-K1 cell line stably expressing KV7.3 that we transiently transfect with wild type (WT) or variant KV7.2 by high efficiency electroporation. The KV7.2 plasmids include the gene for green fluorescent protein enabling quantification of transfection efficiency by flow cytometry. Whole-cell currents were recorded by automated patch-clamp before and after application of the neuronal Kv7 channel activator ezogabine/retigabine (EZO, 10 µM). Following the recording of EZO-induced currents, the Kv7 channel blocker XE-991 (25 µM) was added and whole-cell currents recorded again, enabling offline subtraction of non-specific currents. Results: We investigated the functional and pharmacological properties of 70 KCNQ2 variants. The variants screened included subsets with known pathogenicity for self-limited epilepsy and epileptic encephalopahthy, variants of unknown significance, and gnomAD population variants with low probability of pathogenicity. The variants displayed a range of functional changes (relative to WT) including severe loss-of-function (>75% current reduction, e.g., E130K, R557S); mild current reduction (25-75% of WT, e.g., A43V, R333Q); depolarizing shifts in voltage dependence of activation (S113F, R207W), and slower activation and deactivation kinetics (S113F, R207W). Some variants exhibited WT-like current density (75-110% of WT, e.g., T826I) and normal voltage dependence of activation. For most of the variants we tested, EZO (10 μM) induced a hyperpolarizing shift in the voltage dependence of activation, which was often accompanied by an increase in peak current density. However some loss-of-function variants exhibited no response to EZO, whereas the drug completely inhibited current conducted by one variant (R144Q). Conclusions: Our results demonstrate that automated electrophysiology can be successfully applied for the functional and pharmacological investigation of KCNQ2 variants associated with epilepsy. We demonstrated a range of channel dysfunction including some variants with near normal activity. These functional data may help classify the pathogenicity of variants. We also show the capabilities of automated patch-clamp recording to determine the in vitro response of variants to retigabine and demonstrate that not all epilepsy-associated KCNQ2 variants encode mutant channels that can be rescued by this drug. Funding: NIH: U54-NS108874.