Molecular and topographic mapping of antipsychotic effects: a meta-analysis of postsynaptic density proteins in animal models with translational implications

Background: While antipsychotics primarily target dopamine D2 receptor, the putative synaptic mechanisms underlying their therapeutic effects remain unclear. Postsynaptic density (PSD) at glutamatergic synapses represents a dynamic protein network involved in synaptic plasticity and neurotransmission, whose dysfunction has been implicated in the pathophysiology of schizophrenia and bipolar disorder. This study aims to explore, for the first time, with a quantitative meta-analytical approach, how antipsychotic treatments affect PSD molecules across various brain regions in preclinical settings, which may overcome the heterogeneity of human studies. Methods: We systematically reviewed and meta-analyzed peer-reviewed preclinical studies reporting quantitative effects of typical and atypical antipsychotics on PSD-related outcomes. Statistical analyses used random-effects models, with heterogeneity assessed via τ², I², and Cochran's Q. Separate meta-analyses were conducted for PSD protein type, brain region, animal type, and treatment paradigms. Meta-regressions assessed the impact of cofounding variables. Results: We included 81 studies (n = 2542; rodents=2510; monkeys=26), resulting in 226 meta-analyses. Both typical and atypical antipsychotics modulated PSD molecules albeit with different degree due to the drug receptor profile, dose, specific protein detected, brain region, and administration regimen. Haloperidol, amisulpride, and aripiprazole significantly increased PSD protein expression in the striatum. Acute olanzapine increased PSD molecules levels in striatal regions, reducing Arc expression in frontal cortex (SMD [95%CIs]=-2.15 [-2.7;-1.61], p < 0.0001); chronic treatment had opposite effects. Clozapine increased Homer1a levels in dorsal (1.13 [0.41;1.86], p = 0.002) and ventral striatum (1.48 [0.72;2.24], p = 0.0001), and cingulate cortex (2.12 [1.12;3.12], p < 0.0001), as well as increased NMDAR subunits levels in frontal cortex.