Treg and immune proliferative programs characterize anti-PD-1-responsive endometrial cancer: an integrative single-cell, spatial transcriptomic, and TCGA-UCEC reanalysis

The responses to programmed cell death protein 1 (PD-1)-based therapy in advanced endometrial cancer are heterogeneous. The cellular programs associated with such responses remain incompletely defined, particularly when the immune cell composition, spatial expression patterns, and population-level transcriptomic features are considered together.

Publicly available TCGA-UCEC and GSE251923 data were reanalyzed using bulk transcriptomics, single-cell RNA sequencing (scRNA-seq), and the Visium expression space platform. The GSE251923 dataset included one anti-PD-1 responder and one non-responder; hence, all comparisons between these two cases are descriptive. Prespecified immune- and proliferation-related signatures were evaluated across the datasets, and the Visium clusters were assigned broad marker-based expression compartments. Exploratory quantitative reverse transcription polymerase chain reaction (qRT-PCR) assays were used to examine selected transcripts in the Ishikawa and HEC-1A endometrial cancer cell lines under 1% O₂ hypoxia and in primary neutrophils exposed to endometrial-cancer ascites-conditioned medium. These assays were used as ancillary models of condition-associated transcription; neither hypoxia nor ascites exposure was treated as a direct model of clinical immune checkpoint inhibitor response.

After quality control, we obtained 16,725 single cells that were assigned to 14 populations. Compared with the non-responder sample, the responder sample contained higher proportions of CD8 T cells [+3.1 percentage points (pp)], regulatory T cells (Tregs; +3.3 pp), B cells (+6.7 pp), plasma cells (+5.4 pp), CD4 T cells (+5.9 pp), and proliferating cells (+7.7 pp); the non-responder sample dominantly contained epithelial/tumor and endothelial cells. In TCGA-UCEC (553 tumors and 35 normal tissues), a high Treg signature was associated with longer overall survival [log-rank p = 0.047; hazard ratio (HR) = 0.58; 95% confidence interval (CI): 0.34-1.00], whereas a high proliferation signature was associated with shorter overall survival (HR = 1.72; 95% CI: 1.01-2.94; p = 0.046). The responder Visium section showed descriptively higher Treg and proliferation scores with lower epithelial and extracellular matrix scores. Under hypoxia, the two endometrial cancer cell lines showed higher MKI67, FOXP3, CXCL13, PDCD1, and CTLA4 transcript levels; the ascites-conditioned medium increased PAD4 and ELANE expressions in primary neutrophils.

The integrated analyses helped identify Treg- and proliferation-related patterns that differed between two anti-PD-1-treated endometrial cancers and were directionally contextualized in TCGA-UCEC. The qRT-PCR assays provide ancillary evidence of condition-associated transcriptional changes but do not validate immune cell states, clinical response mechanisms, or functional neutrophil extracellular trap formation. These findings are hypothesis-generating and require confirmation in independent treatment-annotated cohorts.