Integrative Metaprogram Analysis Reveals Transcriptional Dysregulation of Oxidative Stress Response in Polycystic Ovary Syndrome

Objective: The molecular mechanisms underlying polycystic ovary syndrome (PCOS) pathogenesis were investigated using an integrative metaprogram analysis framework, with specific focus on oxidative stress-related pathways through combined single-cell and bulk RNA sequencing approaches. Methods: Three datasets were integrated: a single-cell RNA-seq dataset (GSE240688), a laboratory-generated bulk RNA-seq dataset, and a validation microarray dataset (GSE34526). Transcriptional programs in granulosa cells were identified by non-negative matrix factorization, while dysregulated genes were characterized by differential expression and weighted gene co-expression network analyses. An integrative multi-omics approach was employed to identify key regulatory genes at the intersection of these complementary analytical methods. Results: Ten distinct transcriptional metaprograms in granulosa cells were identified, with metaprogram 4 (MP4) significantly enriched in PCOS samples and associated with oxidative stress response pathways. 199 consistently dysregulated genes across independent patient cohorts were revealed by differential expression analysis. Through integrative analysis, glutathione peroxidase 3 (GPX3) was identified as the only gene present across all three analytical approaches. GPX3 was significantly upregulated in PCOS samples with robust discriminatory power (AUC>0.8) and was functionally associated with insulin signaling, glucose metabolism, and mitochondrial pathways, suggesting its role in connecting oxidative stress and metabolic dysfunction. Conclusions: Oxidative stress responses were revealed as central to PCOS pathophysiology by our comprehensive multi-omics approach, with GPX3 identified as a key regulatory node connecting metabolic and reproductive dysfunction. A molecular basis for the oxidative stress-mediated pathology in PCOS is established by these findings. RNA was extracted from granulosa cells of 9 PCOS patients and 9 controls. PCOS diagnosis followed Rotterdam 2003 criteria (at least two of: oligo/anovulation, hyperandrogenism, polycystic ovarian morphology). Controls were women undergoing IVF for tubal or male factor infertility. Controls were matched with PCOS patients by age, BMI, and thyroid function. Patients with other endocrinopathies, diabetes, or taking reproductive/metabolic medications within three months were excluded. Granulosa cells were collected during oocyte retrieval. Total RNA was extracted using TRIzol reagent. Libraries were prepared with TruSeq RNA Library Prep Kit v2 and sequenced on Illumina NovaSeq 6000. The study aimed to identify transcriptional differences in granulosa cells between PCOS and control subjects.