243. Dysconnectivity in schizophrenia: evidence at the molecular level via a network analysis of postmortem tissues

Abstract

Abstract Background Schizophrenia pathophysiology has consistently been associated with disruptions in the glutamatergic pathway, postsynaptic density (PSD) proteins, and monoaminergic neurotransmission, especially the dopaminergic one. Proper functionality of each neurotransmitter system is crucial for brain health, but their interaction plays an equally critical role in brain development, regulating synaptogenesis and synaptic maturation, which are pivotal processes for neural network stability and cognitive function. Although increasing evidence shows disrupted crosstalk between glutamatergic and dopaminergic pathways, neurotransmitters’ interactions have been poorly investigated in schizophrenia. Aims & Objectives To address this gap, we performed a network analysis based on actual quantified levels of 38 molecules, included in the glutamatergic, dopaminergic, noradrenergic, and serotonergic systems and detected in the dorsolateral prefrontal cortex (DLPFC) and hippocampus of patients affected by schizophrenia compared to controls. Method A total of 80 tissue samples from the DLPFC and hippocampus were collected from postmortem brains of 40 unique subjects, equally divided in non-psychiatric controls and schizophrenia patients. Thirty-eight molecules, including monoamines, D- and L-amino acids, synaptic receptors, transporters, and enzymes, were quantified by HPLC and Western blot. Individual levels of each molecule were compared between groups using the Mann-Whitney test. Molecular networks were estimated by entropy-derived measures, namely conditional mutual information, following dimensionality reduction through autoencoder. This approach allowed to detect both linear and non-linear associations, capturing complex relationships that traditional correlation-based methods might not reveal. Robustness was ensured by retaining only stable connections via bootstrapping procedures. Permutation-based comparisons were provided at three hierarchical levels of network organization. In all the phases of the statistical framework, p-values were adjusted for multiple comparisons. Results Although no significant differences were detected in levels of molecules after the direct comparison between groups, network analyses showed an overall reorganization of the connectivity patterns in schizophrenia patients. We found a reduced overall connectivity in the hippocampus (p=0.02) but not DLPFC (p=0.99) of schizophrenia patients compared to controls. In the DLPFC, altered node strength (the connectivity degree of each molecule) was observed in dopamine (p=0.03), GLT-1 (p=0.045), and GluA4 (p=0.03). In the hippocampus, decreases were detected in the node strength of dopamine (p=0.008), 5-HIAA (p=0.004), CaMKII (p=0.004), DOPAC (p=0.02), NR1 (p=0.02), and Synapsin I (p=0.004). In both the brain regions, comparisons at the most basic level of network organization, such as individual connections between pairs of molecules, highlighted impairments within the glutamatergic pathway, especially between pre- and postsynaptic elements. Other impairments were identified in the connectivity between the glutamatergic and monoaminergic pathways, particularly the dopaminergic one. Discussion & Conclusions We demonstrated a reorganization of molecular networks in schizophrenia, with alterations in the micro-architecture of molecular couplings. Disturbances were observed both within and between neurotransmitter systems, particularly between glutamatergic and dopaminergic pathways, which may represent neurobiological mechanisms underlying psychotic disorders. These findings strengthen the importance of investigating functional interactions among molecules, as such analyses may reveal significant alterations in schizophrenia that cannot be detected through simple comparisons of molecular levels.