Hypothesis-Generating Single-Cell Analysis of the Endometrial Cancer Immune Microenvironment With Complementary Multiomics Evidence.

BACKGROUND

Endometrial cancer (EC) has a heterogeneous immune microenvironment, and response-associated biology remains incompletely characterized. We used the retained EC single-cell dataset to describe immune-cell states and complemented it with TCGA-UCEC, independent gut microbiome cohorts, and RL95-2 proof-of-concept experiments. All responder versus nonresponder scRNA-seq comparisons are descriptive and hypothesis generating because only one sample was available per clinical response group.

METHODS

Single-cell RNA sequencing (scRNA-seq) data from EC patients undergoing anti-PD-1 immunotherapy (GEO: GSE251923; one responder, GSM7990051, and one nonresponder, GSM7990052; 17,546 cells retained after quality control) were analyzed using Scanpy, Leiden clustering, and over-representation analysis with GO biological process, KEGG, and MSigDB Hallmark gene sets using Enrichr with Benjamini-Hochberg correction. Bulk RNA-seq from TCGA-UCEC (554 tumor and 35 normal samples) was used for complementary transcriptomic analysis. Gut microbiome 16S rRNA datasets (SRA: PRJNA897842 and PRJNA833670) were analyzed as independent cohorts, and RL95-2 qRT-PCR experiments were interpreted as proof-of-concept.

RESULTS

Ten major cell populations were identified in the EC immune microenvironment. In this single responder/nonresponder sample pair, the nonresponder sample showed patterns consistent with higher M2-associated macrophage marker expression, increased T-cell checkpoint/exhaustion marker expression, and lower cytotoxic activity scores. These observations are descriptive and cannot establish patient-level significance. TCGA-UCEC analyses and the independent microbiome and RL95-2 experiments provide complementary, hypothesis-generating evidence; no unverified classifier-performance estimate is reported.

CONCLUSIONS

This study provides a hypothesis-generating description of immune patterns in one EC responder and one nonresponder sample, complemented by TCGA-UCEC, independent microbiome datasets, and RL95-2 proof-of-concept experiments. The findings nominate candidate immune, hormonal, and microbiome-associated pathways for further investigation but do not establish patient-level response differences, causality, or clinically validated biomarkers.