Evangelia Bolli, Pratyaksha Wirapati, Mehdi Hicham, Yuxuan Xie, Marie Siwicki, Florent Duval, Anne-Gaëlle Goubet, Máté Kiss, Béatrice Zitti, Thomas Zwahlen, Sheri A. C. McDowell, Ruben Bill, Simona Angerani, Camilla Engblom, Seth Anderson, Aiping Jiang, Oliver Hartley, D. Sykes, Maja Jankovic, Nadine Fournier, Matthias Gunzer, David Tarussio, Stéphanie Tissot, Peter M. Sadow, William C. Faquin, Moshe Sade-Feldman, Ralph Weissleder, Sara I. Pai, François Mercier, Robert Manguso, Mikaël J. Pittet
Tumor-associated neutrophils (TANs) are abundant across cancers, yet their phenotypic diversity and functional states remain poorly defined. Here, we introduce a cell-type probability classifier that recovers low-transcript neutrophils from scRNAseq datasets, enabling pan-cancer analyses of TAN heterogeneity. Across >190 human and murine tumors, we identify a conserved differentiation trajectory that culminates in a terminal CCL3 hi state. This state exhibits pro-tumor transcriptional programs, including those involved in hypoxic adaptation and senescence. Consistently, CCL3 hi TANs are enriched in hypoxic tumor niches in both humans and mice. Through mechanistic perturbations of neutrophil-derived CCL3 in mice, we show that it sustains TAN survival in hypoxic tumor regions via CCR1-dependent signaling. These findings establish CCL3 as a conserved marker and functional driver of pro-tumor neutrophils in growing tumors, and provide a scalable framework for dissecting neutrophil biology across cancer types.