441. SCREENING OF SCHIZOPHRENIA BIOMARKERS BASED ON PROTEOMICS AND THEIR APPLICATION IN EARLY DIAGNOSIS

Abstract Background Schizophrenia is a debilitating mental disorder, and identifying biomarkers can shift diagnosis from subjective symptom-based assessment towards objective detection. In the Clinical High Risk of Psychosis (CHR) population, early detection is crucial as one-third may develop severe illness, and there is limited research in China on the correlation between proteins and diseases based on proteomics studies. Our early findings indicate that Perceived Spatial Separation-induced Prepulse Inhibition (PSS-PPI) aids in early psychosis detection. Aims & Objectives To screen differentially expressed proteins based on proteomics, validate the selected target proteins as biomarkers for schizophrenia, integrate these biomarkers with clinical indicators to establish a diagnostic model，facilitate their application in the early detection of schizophrenia. Method Data were collected from individuals with First Episode Schizophrenia (FES), Clinical High Risk (CHR) individuals, and Healthy Controls (HC). Cross-sectional clinical data, cognitive function assessments, Prepulse Inhibition (PPI) measures, and peripheral blood protein levels were analyzed. CHR participants were followed until psychosis onset or for three years. Proteomics identified differentially expressed proteins across groups, assessing their potential to distinguish disease states and conducting enrichment analysis to explore disease mechanisms. Target proteins were selected based on proteomic results and validated using ELISA to confirm differences in protein levels across groups. Correlations between target proteins, PPI indices, and cognitive function were analyzed to construct a diagnostic model. Additionally, the predictive value of target proteins for psychosis transition in the CHR group was explored. Results Proteomic technology identified several differentially expressed proteins, with enrichment analysis aligning with schizophrenia mechanisms. Dimensionality reduction revealed an ideal protein expression distribution among FES, CHR, and HC groups. ELISA assays on seven target proteins (APOE, C1R, C4BPB, C4, IGLC2, IGLV7-46, MARCO) across 58 FES, 47 CHR, and 59 HC cases showed significant differences in protein levels. ANCOVA, adjusting for age, highlighted six proteins with significant expression differences between FES and HC groups: APOE, C1R, C4, IGLC2, IGLV7-46, and MARCO. Four proteins differed between CHR and HC groups, and five between FES and CHR groups. Pearson correlation analysis showed negative correlations between IgLC2 levels and PSSPPI at 120ms and 60ms in the healthy population. C4, IGLV7-46, and C4BP levels showed weak correlations with PSSPPI at 120ms across all groups. The MCCB total score correlated with APOE and IGLV7-46 levels in the healthy population and positively with C1R, C4, MARCO, IgLC2, and IGLV7-46 levels across all groups. Logistic regression models for FES versus HC and CHR versus HC showed high AUC values, which increased with PPI measures. The FES versus CHR model showed no improvement with PPI. Random forest models yielded similar results. In CHR, MARCO levels significantly differed between converted and non-converted groups, achieving an AUC of 0.917 when combined with gender and cognitive indicators in the transition prediction model. Discussion & Conclusions The findings provide potential biomarkers for the early detection of schizophrenia and lay the groundwork for the development of future prevention and treatment strategies. Future research should further validate the efficacy of these biomarkers in independent cohorts and explore their connection to the pathogenesis of schizophrenia.