Novel Cell-Type Probability Classifier Maps Pro-Tumor Neutrophil States

Neutrophils are best known as key defenders against infection, yet within tumors they can adopt contrasting roles that either promote or resist cancer progression. Their behavior varies widely, and determining which neutrophil states favor tumor growth has been a persistent technical problem. The difficulty arises because single‑cell RNA sequencing, the standard technology used to study cell states, reads gene transcripts one cell at a time—but neutrophils contain exceptionally low levels of RNA, making their transcriptional profiles hard to define.

To overcome this limitation, Mikaël Pittet and colleagues at Ludwig Lausanne developed a cell‑type probability classifier that can identify neutrophil states directly from raw sequencing data. By applying this computational tool to more than 190 human and murine tumors, including previously collected datasets, the team established a comprehensive view of tumor‑associated neutrophils (TANs). Their analysis revealed a consistent “aged” TAN population characterized by high expression of a protein called CCL3 across diverse tumor types.

“We found that tumors induce in neutrophils a genetic program that sets them on a trajectory of continuous maturation, culminating in a terminal ‘aged’ state characterized by high CCL3 expression,” said Pittet, senior author of the paper published in Cancer Cell. These CCL3^hi TANs occupy low‑oxygen regions of tumors and activate gene programs that enable them to adapt and sustain cancer cell survival.

The classifier’s ability to parse subtle transcriptional differences linked these CCL3^hi TANs to a conserved pro‑tumor gene expression profile previously associated with cancer progression. Functionally, the team showed that CCL3 activates its receptor CCR1 on the same neutrophils, reinforcing their aged state and supporting tumor growth. When mice lacked either CCL3 or CCR1 in their neutrophils, they no longer promoted tumor development, demonstrating that this signaling axis maintains the pro‑tumor TAN population.

“Our work establishes that tumors actively maintain pro‑tumor neutrophils through CCL3 and identifies these cells as a conserved and clinically relevant compartment of cells in the TME,” said Pittet.

By enabling accurate characterization of neutrophil subtypes, the cell‑type probability classifier addresses a key analytical gap in single‑cell studies. It also provides a framework for exploring other elusive immune cell populations whose gene expression patterns shape the tumor microenvironment and influence cancer outcomes.