Polycystic ovary syndrome affects an estimated 8 to 13 percent of women worldwide, yet its underlying biology remains only partly understood. A new study published in Biological Procedures Online has combined large-scale gene mining with laboratory experiments to trace how immune dysregulation and a form of iron-dependent cell death called ferroptosis may jointly drive the disorder.
The research team began by mining public gene expression data. They downloaded five datasets from the Gene Expression Omnibus, using GSE34526 and GSE80432 as a training set and GSE10946, GSE48031 and GSE84958 as independent validation sets. After correcting for batch effects, they screened for differentially expressed genes using a false discovery rate threshold below 0.05 and a log2 fold change greater than 0.585. The analysis revealed 1,102 genes upregulated and 654 genes downregulated in PCOS samples. Cluster analysis showed that the molecular profiles of PCOS and control groups were markedly distinct, indicating a strong correlation between the dysregulation of these genes and the disease process.
Functional enrichment analysis of the differentially expressed genes pointed to two converging biological themes. The upregulated genes were dominated by immune-inflammatory functions, including positive regulation of cytokine production and leukocyte migration, and clustered in pathways such as phagosome and NOD-like receptor immune signaling. The downregulated genes were mainly involved in metabolic regulation and signal transduction, including insulin secretion in response to glucose stimulus and negative regulation of Wnt signaling, with enrichment in calcium signaling and fatty acid metabolism. Gene set enrichment analysis further showed that the immune response regulating signaling pathway and ferroptosis were both significantly enriched, suggesting that abnormal immune regulation and disordered ferroptosis processes may jointly participate in the pathogenesis of PCOS.
To narrow the field of candidate genes, the researchers applied weighted gene co-expression network analysis, intersecting PCOS-related modules with immune- and ferroptosis-related gene sets. This produced 37 module genes related to ferroptosis and 402 related to immunity, with 19 genes common to all three sets. Cross-referencing these with upregulated genes from the three validation datasets yielded twelve key candidate genes: RNF144B, C3AR1, GLIPR1, S100A8, NNMT, GBP2, DNAJC3, CD36, FTL, GPAM, PELI3 and VSIG4. Three machine learning approaches, LASSO regression, RandomForest and support vector machine-recursive feature elimination, were then used to evaluate diagnostic performance. In the training set, the RandomForest model showed the best discrimination with an area under the ROC curve of 0.75, while external validation revealed a degree of cohort-dependence, with SVM performing better in GSE84958 and RandomForest performing better in GSE48031 and GSE10946.
The experimental arm of the study focused on human ovarian granulosa cells, the KGN cell line, treated with 500 nanomolar dihydrotestosterone for 24 hours to mimic the hyperandrogenic environment characteristic of PCOS. Among the twelve candidate genes, NNMT showed the most striking response. Its mRNA expression rose significantly over 24 hours of DHT exposure, and Western blot confirmed that protein levels increased with both longer treatment duration and higher DHT concentrations.
To test whether NNMT could drive ferroptosis directly, the team constructed an overexpression vector and transduced it into KGN cells. The results were clear. Overexpressing NNMT significantly elevated malondialdehyde, a marker of lipid peroxidation, increased intracellular ferrous iron levels, and raised lipid reactive oxygen species measured by flow cytometry using the C11-BODIPY fluorescent probe. Western blot revealed that ferroptosis-inhibitory proteins SLC7A11 and GPX4 were significantly downregulated, while ACSL4, a ferroptosis-promoting protein, was markedly upregulated. Cell viability dropped and cytotoxicity, measured by lactate dehydrogenase release, increased. Treatment with the ferroptosis inhibitor Ferrostatin-1 reversed these effects, confirming that the cell death induced by NNMT overexpression was related to the ferroptosis pathway.
When the researchers knocked down NNMT using short-hairpin RNA vectors before exposing cells to DHT, cell viability was markedly restored compared to the scrambled control plus DHT group, and cytotoxicity was significantly reduced. Knockdown also inhibited DHT-induced malondialdehyde production, ferrous iron accumulation and lipid ROS increase, while reversing the downregulation of SLC7A11 and GPX4 and the upregulation of ACSL4.
To extend the findings beyond cell culture, the team established a PCOS mouse model by daily subcutaneous injection of dehydroepiandrosterone for 20 days in four-week-old female C57BL/6 mice. Ovarian tissue showed significantly increased numbers of cystic follicles and decreased corpora lutea, recapitulating classic PCOS pathology. Enzyme-linked immunosorbent assay of plasma revealed markedly elevated luteinizing hormone and anti-Mullerian hormone alongside significantly decreased follicle-stimulating hormone, reproducing the endocrine signature of the syndrome. At the molecular level, ovarian tissue from PCOS mice showed significant upregulation of NNMT and ACSL4 mRNA and downregulation of SLC7A11 and GPX4, with Western blot confirming the same pattern at the protein level.
As a core metabolic enzyme, NNMT consumes the methyl donor S-adenosylmethionine in large quantities, reshaping cellular state at both epigenetic and metabolic levels and creating a microenvironment that promotes oxidative stress. This disrupted metabolic background may undermine the cell’s core defense system against ferroptosis, enhancing the iron-ion-dependent Fenton reaction and leading to ferrous iron accumulation and elevated lipid peroxidation markers.
The authors note that their DHEA-induced mouse model, while reproducing cardinal features of PCOS, may not capture all metabolic dimensions of the syndrome, since previous work suggests that combining a high-fat diet with DHEA or letrozole produces a more metabolically complete model. Further studies will be needed to determine whether the NNMT and ferroptosis changes observed here hold in alternative models. Nevertheless, by systematically connecting immune-inflammatory activation, androgen-driven metabolic enzyme upregulation and iron-dependent cell death, the study offers a theoretical basis for therapeutic strategies targeting the NNMT-ferroptosis axis and adds twelve candidate genes with potential diagnostic value to the growing molecular map of polycystic ovary syndrome.
Subject of Research: Bioinformatics and experimental identification of immune and ferroptosis genes, including NNMT, involved in polycystic ovary syndrome
Article Title: Identification and Validation of Genes Involved in Immune and Ferroptosis in Polycystic Ovarian Syndrome Using Bioinformatics and Experimental Approach
Article References: Chen, Y., Xing, L., He, Z., Wang, X., Chen, Y., Yuan, X., & Zhao, R. (2026). Identification and Validation of Genes Involved in Immune and Ferroptosis in Polycystic Ovarian Syndrome Using Bioinformatics and Experimental Approach. Biological Procedures Online, 28(1), Article 43. https://doi.org/10.1186/s12575-026-00355-z
Image Credits: AI Generated
DOI: 10.1186/s12575-026-00355-z
Keywords: PCOS, ferroptosis, NNMT, immune dysregulation, ovarian granulosa cells, dihydrotestosterone, SLC7A11, GPX4, ACSL4, bioinformatics, machine learning, DHEA mouse model
News Source: Juliet Wilcox. (October 30, 2026). Scientists Link Immune Genes and Iron-Driven Cell Death to PCOS. Scienmag.
