CPP alters cross-frequency coupling between theta and gamma in CA1 in rats: Simulation and experiment

Uncompetitive NMDA receptor (NMDAR) antagonists like ketamine and PCP are used to produce pharmacological models for schizophrenia. Other NMDAR antagonists (e.g. CPP), with different molecular mechanisms of action, are used to explore memory and cognitive functions in animals but can also produce psychosis. CPP is a competitive antagonist which displaces glutamate from its binding site on NMDAR. Because competitive antagonism results in more glutamate availability, other glutamate receptors may show increased activation at the same synapses. We modeled this effect by increasing AMPAR conductance while setting NMDAR conductance to zero. We compared the modeling results with data obtained from the CA1 region in rats. Simulations were run in NEURON. Network consisted of 800 5-Compartment pyramidal cells (PYR), 200 1-C basket cell interneurons (BAS), and 200 1-C oriens lacunosum-moleculare interneurons (OLM). Cell classes were interconnected probabilistically at suitable densities with AMPA/NMDA and both fast and slow GABAA synapses. OLMs formed synapses on the distal dendrites of PYRs, while BASs synapsed on proximally PYRs and on other BASs. We compared models to look at blockage at synaptic sites on the different cell classes, considering differing affinities of inhibitors due to different isotypes. Experimental recordings were made from tetrode arrays implanted in the CA1 region of Long Evans rats trained to chase sugar pellets in a square recording environment. Recordings were made before and after 5 mg /kg of intraperitoneal CPP. CPP reduced theta and increased gamma power experimentally. We were able to replicate these changes with 2 models of competitive NMDAR inhibition: 1. at all synapse locations (OLM, BAS, and PYR) ; 2. effect of CPP only at OLM sites. Both CPP models also showed reduction in the strength of cross-frequency coupling (CFC) of 2 gamma frequencies, 50 and 80 Hz coupled with theta phase, similar to experiment. However, the relative strength of the peaks changed in different directions in the 2 models: the 50 Hz peak increased in Model 1, while the 80 Hz peak increased in Model 2. We are now analyzing the experimental results in greater detail to determine which of these model predictions is supported.