Back

A multi-tissue single-cell tumor microenvironment atlas reveals myeloid-derived cell states with significant impact on clinical outcome

Maklouf, G. R.; Guimaraes, G. R.; Teixeira, C. E.; Pretti, M. A. M.; de Oliveira Santos, L.; Tessarollo, N. G.; Toledo, N. E.; Falchetti, M.; Dimas, M. M.; Serain, A. F.; Bastos, N. C.; de Macedo, F. C.; Rodrigues, F. R.; da Silva, J. L.; Lummertz-Rocha, E.; Chaves, C. B. P.; de Melo, A. C.; Moraes-Vieira, P. M. M.; Mori, M. A.; Boroni, M.

2023-01-04 cancer biology
10.1101/2023.01.04.522727 bioRxiv
Show abstract

Tumor-associated myeloid-derived cells (MDCs) significantly impact cancer prognosis and treatment response due to their remarkable plasticity and tumorigenic behaviors. We integrated single-cell RNA-Sequencing datasets from seven different cancers, resulting in a comprehensive collection of 29 MDC subpopulations in the tumor microenvironment (TME). Distinguishing resident-tissue from monocyte-derived macrophages, we discovered a resident-tissue-like subpopulation within monocyte-derived macrophages. Additionally, hypoxia-driven macrophages emerged as a prominent TME component. Deconvolution of these profiles revealed five subpopulations as independent prognostic markers across various cancer types. Validation in large cohorts confirmed the FOLR2-expressing macrophage association with poor clinical outcomes in ovarian and triple-negative breast cancer. Moreover, the marker TREM2, commonly used to define immunosuppressive tumor-associated macrophages, cannot solely predict cancer prognosis, as different polarization states of macrophages express this marker in a context-dependent manner. This comprehensive MDC atlas offers valuable insights and a foundation for novel analyses, advancing strategies for treating solid cancers.

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.