Invited reviewNMDA receptors, cognition and schizophrenia - Testing the validity of the NMDA receptor hypofunction hypothesis

Cognitive dysfunction is core to schizophrenia, and remains poorly treated by existing therapies. A prominent hypothesis suggests that many symptoms arise from N-methyl-d-aspartate receptor (NMDAR) hypofunction. Subsequently, there has emerged a widespread use of many preclinical and clinical NMDAR antagonist models in the search for novel treatments. Clinically, ketamine is broadly purported to induce cognitive symptoms similar to those of schizophrenia. Preclinically, acute, subchronic and neonatal NMDAR antagonist administration models are all utilised in this context, as well as NMDAR transgenic animals. In this review, key strengths and weaknesses of each of these approaches are described with regard to their ability to recapitulate the deficits seen in patients. Given the breadth of literature and vogue for research in this topic, instances of NMDAR antagonist effects in the desired domains can readily be found preclinically. However, it is surprisingly difficult to identify any single aspect of cognitive function that possesses complete translational integrity. That is, there does not seem to be an NMDAR antagonist regimen proven to engage NMDARs equivalently in humans and animals that reliably produces the same cognitive effects in each species. This is likely due to the diverse range of techniques and models used by preclinical researchers, a paucity of research describing pharmacokinetic-pharmacodynamic relationships of NMDAR antagonist regimens, little capability to measure target engagement, and the lack of harmonized procedures between preclinical and clinical studies. Realizing the potential of the NMDAR hypofunction hypothesis to model cognitive impairment in schizophrenia will require some of these issues to be addressed.This article is part of a Special Issue entitled 'Schizophrenia'.

► Critical examination of NMDA receptor antagonist modelling in schizophrenia. ► Models can exhibit relevant cognitive deficits, but clinical comparisons are weak. ► Translational validity challenged by lack of justification of model choice. ► Vital role of cross-species target engagement measures and homologous test constructs.