Back

Single-Cell RNA Sequencing and Inferred Protein Activity Analysis Reveal a Distinct Tumor Phenotype in Early-Onset Colorectal Cancer Patients

Cohen, D. A.; Obradovic, A.; Jovanovic, L.; Mitani, Y.; Li, T.; Lee, S.; De, D.; Bethancourt, C.-N.; Shin, A.; Dunbar, K.; Cambuli, F.; Lengner, C.; Li, N.; Rogers, C.; Ho, J.; Jakubikova, V.; Blystone, E.; Pumpalova, Y.; Dionigi, B.; Vickovic, S.; Cheng, C.-W.; Gabre, J.

2025-05-16 cancer biology
10.1101/2025.05.15.653154 bioRxiv
Show abstract

Colorectal cancer (CRC) diagnosed before age 50 years (early-onset CRC, EO-CRC) is rising at an alarming rate, yet its molecular and microenvironmental drivers remain poorly understood. EO-CRC is highly heterogeneous, and while subtle differences from late-onset CRC (LO-CRC) have been reported, their full extent remains unresolved due to the limited scope of previous studies. Here, we integrate public data with in-house clinical samples profiled by single-cell RNA sequencing (scRNA-seq) and multiplex immunofluorescence (mIF) to compare EO-CRC and LO-CRC. Additionally, we employ gene regulatory network-based protein activity inference (VIPER), enabling a more precise characterization of key regulatory proteins driving tumor-stroma interactions. Our analysis reveals that EO-CRC and LO-CRC have a largely similar immune composition, challenging previous reports of an "immune-cold" phenotype in EO-CRC. However, we identify distinct stromal differences, including a significant enrichment of fibroblasts in EO-CRC. Notably, we define a previously unrecognized epithelial subpopulation in EO-CRC, marked by high expression of toll-like receptor 4 (TLR4) and C-C chemokine receptor type 5 (CCR5)--key mediators of inflammation-driven tumor progression and fibroblast recruitment. These findings suggest that EO-CRC may be driven by a tumor-intrinsic inflammatory phenotype with enhanced stromagenesis, providing new insights into potential mechanisms underlying its increasing incidence in young adults.

Matching journals

The top 8 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.