Simulated Block of Brain Na+ Channels by Ranolazine Predicts Selective Binding to Inactivated Conformations

Abstract

The antianginal drug ranolazine inhibits the increased persistent Na+ current associated with several NaV1.1 congenital mutations causing epilepsy and migraine syndromes. This inhibition is consistent with fast-inactivated state binding. In contrast, recently published studies of ranolazine block of the cardiac NaV1.5 and skeletal muscle NaV1.4 have found evidence for pre-open and/or open state binding. In the present study, computer modeling to identify the channel states necessary and sufficient to reproduce the experimentally-observed block of brain Na+ channels by ranolazine. A minimal Markov model of a brain Na+ channel was developed that reproduces all voltage-dependent gating behaviors including activation (V½= −21.8mV), fast-inactivation (V½= −62.6mV), slow-inactivation (V½= −67.4mV), persistent current (0.23% at −10mV) and single channel mean open time (2.6msec at 0mV). In addition to three closed states, this model includes one state each for pre-open (PO), open (O), fast-inactivated (FI) and slow-inactivated (SI) states. The binding rates of ranolazine to HEK293 cells stably expressing the hNaV1.2 isoform were measured (KON= 1M−1msec−1 and KOFF= 5e−5msec−1) and were used to guide simulated binding to PO, O, FI or SI channel conformations. Only simulated binding to the FI state delayed the recovery from fast inactivation (second phase tau: control 656msec, 30uM ranolazine 729msec), which is a common feature of ranolazine block of all Na+ channel isoforms. Simulations of evoked neuronal action potential firing with 10uM ranolazine binding either to the FI or SI state demonstrated equal effectiveness for the reduction of excitability (#evoked APs during 4sec: control 174, FI binding 54, SI binding 53). No effect on AP# or frequency was observed with simulated binding to either PO or O states. Our data suggest that ranolazine binds to the inactivated states of brain Na+ channels.