Back

Site-dependent Treg cell transcriptional reprograming in a metastatic colorectal cancer model holds prognostic significance

Aristin Revilla, S.; Verheem, A.; Frederiks, C. L.; Kim, Y.; Chalkiadakis, T.; Viergever, B.; Gyorffy, B.; Mocholi, E.; Kranenburg, O.; Prekovic, S.; Coffer, P. J.

2025-03-15 immunology
10.1101/2025.03.13.643028 bioRxiv
Show abstract

In colorectal cancer (CRC), tumor-infiltrating regulatory T (Treg) cells suppress anti-tumor immunity, promoting immune evasion and tumor progression. Effective therapies require selectively targeting tumor-infiltrating Treg (TI-Treg) cells while preserving systemic Treg cells, necessitating insight into their adaptations within the tumor microenvironment. Here, CRC-organoids were implanted in the liver of Foxp3eGFP mice to investigate location-specific phenotypic differences in TI-Treg cells. Tumor tissue exhibited an increased proportion of Treg cells and a decrease of effector CD4 and CD8 T cells compared to matched healthy tissue. RNA sequencing of Treg cells isolated from the spleen, primary liver tumor transplant, or metastases identified gene expression profiles previously associated with CRC-related Treg cells in patients. Location-specific differences included elevated expression of WNT-pathway genes in peritoneal TI-Treg cells compared to liver counterparts. Higher expression of genes upregulated in liver TI-Treg cells correlated with poor CRC prognosis. Splenic Treg cells from tumor-bearing mice displayed distinct transcriptional profiles from both their healthy counterparts and TI-Treg cells, suggesting they represent a distinct CD4+ population. Taken together, these findings highlight TI-Treg cells heterogeneity across different tumor sites and the distinct nature of splenic Treg cells in tumor-bearing hosts.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.