N1 cassette-lacking NMDA receptors mediate the antidepressant activity of ketamine

Ketamine is a fast-acting antidepressant, but how it works is unclear. A key question is whether NMDA receptors (NMDARs), which ketamine blocks, are responsible. Alternative splicing of the GluN1 subunit produces two versions: GluN1a (without exon 5) and GluN1b (with exon 5). In mice, ketamine blocked long-term potentiation (LTP) in the hippocampus only in GluN1a mice, not GluN1b mice. Although ketamine inhibited synaptic NMDARs in both types, GluN1a receptors remained blocked during neuronal firing while GluN1b receptors escaped blockade. Ketamine produced an antidepressant effect in GluN1a mice but not in GluN1b mice. GluN1a-containing NMDARs are selectively responsible for ketamine's antidepressant effect.

Study at a glance

Abstract

Ketamine has emerged as a rapid-acting and robust antidepressant1,2. However, the mechanism of its antidepressant action remains enigmatic. The core issue that has yet to be resolved is whether NMDA receptors (NMDARs), which are subject to open channel blockade by ketamine3,4, mediate the antidepressant effect. NMDARs naturally undergo alternative splicing of the obligatory GluN1 subunit5, producing receptor diversity in the brain that has not been considered in the actions of ketamine. Here we discover that alternative splicing of Grin1 exon 5, which leads to exclusion (GluN1a) or inclusion (GluN1b) of the N1 cassette, located in the N-terminal domain of GluN1 distant from the pore, unexpectedly dictates the level and dynamics of NMDAR blockade by ketamine and gates its antidepressant activity. We find that ketamine prevents NMDAR-dependent long-term potentiation (LTP) in the CA1 region of the hippocampus in mice engineered to exclude Grin1 exon 5 (GluN1a mice), but ketamine has no effect on LTP in mice engineered to include this exon (GluN1b mice). Ketamine inhibits synaptic NMDARs in CA1 pyramidal neurons in both GluN1a and GluN1b mice, with the level of steady-state blockade marginally greater in GluN1a- than in GluN1b-containing NMDARs. However, the rate of relief of ketamine blockade upon membrane depolarization is markedly slower in GluN1a than in GluN1b neurons such that GluN1a-containing receptors remain blocked during bursting activity, whereas those containing GluN1b escape the ketamine blockade. Furthermore, ketamine treatment, either via systemic administration or local infusion into the hippocampus, induces an antidepressant effect in GluN1a mice but has no effect in GluN1b mice. Collectively, we identify GluN1a-containing NMDARs, which are persistently blocked by ketamine during neuronal firing activity, to be selectively responsible for the antidepressant effect.