An endoplasmic reticulum stress-related four-gene tissue diagnostic signature associated with endometriosis and the immune microenvironment

Yiwen Wang, Yanbo Zhong, Xiao Wang, Ling Xu
In: Letters in Drug Design & Discovery · 2026 · pp. 100662 · doi:10.1016/j.lddd.2026.100662 · W7225459674
OA: hybrid CC0
🔓 Open OA copy View on OpenAlex View at publisher
⚙ AI-generated summary by qwen3.7-flash, 2026-11-09 ⓘ

An endoplasmic reticulum stress-related four-gene signature comprising PDIA4, ITPR1, CCL2, and PTGS2 effectively distinguished endometriosis from controls in training and internal validation sets but showed limited external transferability and correlated with immune infiltration.

One-sentence paraphrase of the abstract; not a substitute for reading it. No clinical advice. How this works

Abstract

Background Endoplasmic reticulum stress (ERS) has been implicated in endometrial lesion progression and immune microenvironment remodeling, yet ERS-related molecular signatures for endometriosis (EM) diagnosis remain poorly defined. Methods Three GEO datasets (GSE11691, GSE7305, and GSE120103) were batch-corrected and integrated; GSE25628 was reserved for external dataset-level assessment. The resulting 74 tissue samples were divided at the patient level into a training set (N = 57) and an internal validation set (N = 17). Of 972 differentially expressed genes (DEGs), 47 overlapped with 357 unique ERS-related genes. GO, KEGG, and GSEA enrichment analyses were performed. Random forest and LASSO regression were used to screen candidate genes, followed by VIF assessment and multivariable logistic regression in the training set. Model performance was evaluated using ROC analysis, apparent calibration, decision curve analysis, and bootstrap optimism correction. The fixed training model was directly applied to the internal validation set without re-estimating its coefficients. A GSE120103-only sensitivity analysis evaluated differential-expression consistency. ssGSEA and CIBERSORT were used to estimate immune-cell abundance and assess gene–immune correlations. Results Forty-seven ERS-related DEGs were enriched in ER protein processing, apoptosis, TNF, and IL-17 signaling pathways. Four core genes (PDIA4, ITPR1, CCL2, and PTGS2) constituted the final signature. In the training set, the AUC was 0.985 (95% CI: 0.955–1.000), and the optimism-corrected AUC after 1000 bootstrap resamples was 0.966. Direct application of the fixed model to the internal validation set yielded an AUC of 0.914 (95% CI: 0.777–1.000). The fixed model showed limited discrimination in GSE25628 (AUC < 0.6), whereas the individual genes showed low-to-moderate discrimination in this external dataset. In the GSE120103-only analysis, PDIA4, ITPR1, and PTGS2 retained significant differential expression, whereas CCL2 did not meet the prespecified threshold. CCL2 correlated positively with inferred M2 macrophage abundance, whereas PDIA4 correlated negatively with inferred mast-cell abundance. Conclusion This ERS-based four-gene signature showed preliminary discrimination in the training and internal validation sets but limited external transferability and was associated with altered immune infiltration in EM. Further validation in larger independent cohorts and clinical tissue samples is required.

My notes (saved in your browser only)

Citation neighborhood

Papers in the corpus that this work cites (lower rings, blue) and that cite this one (upper rings, green). Dot size scales with the paper's in-corpus citation count — bigger dot = more influential within the endo/adeno field. Click a dot to open that paper. [ expand to 2 hops ] — adds papers reached through this work's immediate citers/citees. Heavier; up to 60 extra dots.

References (40)

Source provenance

openalex
last seen: 2026-11-10T06:07:28.844288+00:00
License: CC0 · commercial use OK