Cell-type specific modulation of NMDA receptors triggers antidepressant actions.

Both the NMDA receptor (NMDAR) positive allosteric modulator (PAM), and antagonist, can exert rapid antidepressant effects as shown in several animal and human studies. However, how this bidirectional modulation of NMDARs causes similar antidepressant effects remains unknown. Notably, the initial cellular trigger, specific cell-type(s), and subunit(s) of NMDARs mediating the antidepressant-like effects of a PAM or an antagonist have not been identified. Here, we used electrophysiology, microdialysis, and NMR spectroscopy to evaluate the effect of a NMDAR PAM (rapastinel) or NMDAR antagonist, ketamine on NMDAR function and disinhibition-mediated glutamate release. Further, we used cell-type specific knockdown (KD), pharmacological, and behavioral approaches to dissect the cell-type specific role of GluN2B, GluN2A, and dopamine receptor subunits in the actions of NMDAR PAM vs. antagonists. We demonstrate that rapastinel directly enhances NMDAR activity on principal glutamatergic neurons in medial prefrontal cortex (mPFC) without any effect on glutamate efflux, while ketamine blocks NMDAR on GABA interneurons to cause glutamate efflux and indirect activation of excitatory synapses. Behavioral studies using cell-type-specific KD in mPFC demonstrate that NMDAR-GluN2B KD on Camk2a- but not Gad1-expressing neurons blocks the antidepressant effects of rapastinel. In contrast, GluN2B KD on Gad1- but not Camk2a-expressing neurons blocks the actions of ketamine. The results also demonstrate that Drd1-expressing pyramidal neurons in mPFC mediate the rapid antidepressant actions of ketamine and rapastinel. Together, these results demonstrate unique initial cellular triggers as well as converging effects on Drd1-pyramidal cell signaling that underlie the antidepressant actions of NMDAR-positive modulation vs. NMDAR blockade.