Development of an Endoplasmic Reticulum Stress-Related Diagnostic Signature in Polycystic Ovary Syndrome.

Abstract

Polycystic ovary syndrome (PCOS) is a prevalent endocrine and metabolic disorder in premenopausal women. This investigation was to elucidate the underlying mechanism of endoplasmic reticulum stress (ERS) activation in granulosa cells, which has been implicated in the etiology of PCOS. Differentially expressed genes (DEGs) between PCOS and control groups were integrated with ERS gene lists from databases to identify DE-ERS genes, and functional analyses were performed. Univariate regression analysis and the LASSO method were used to select diagnostic factors, followed by establishing a DE-ERS gene-based diagnostic model. A nomogram model was further generated to predict the risk of PCOS. The correlation between ERS gene expression and immune cell proportion was assessed. A total of 14 DE-ERS genes associated with "protein processing in endoplasmic reticulum", "ferroptosis", and "glycerophospholipid metabolism" were selected as PCOS-related factors. An eight-DE-ERS genes-based diagnostic model was developed and displayed satisfactory performance in the training (Area under curve (AUC) = 0.983) and validation datasets (AUC = 0.802). High risk of PCOS can be accurately predicted, which might contribute to clinical decision-making. Moreover, EDEM1 expression was significantly positively correlated with naive B cell infiltration, while PDIA6 was negatively correlated with neutrophil proportion (P < 0.001). We identified eight novel molecules and developed an ERS gene-based diagnostic model in PCOS, which might provide novel insight for finding biomarkers and treatment methods.