While the advent of mass cytometry and single-cell RNA sequencing has undoubtedly enabled high-resolution profiling of neutrophil heterogeneity across developmental and maturation stages, the functional significance of this diversity — particularly in the context of cancer — remains incompletely understood. In a recent Cell Research paper, Liu et al. comprehensively dissected the developmental stages of human neutrophils within the bone marrow and uncovered that human myelocytes and metamyelocytes possess potent immunosuppressive properties and play a critical role in cancer progression.

Neutrophil heterogeneity is increasingly recognized as a critical component of the tumor immune landscape. This heterogeneity is observed in peripheral blood (PB) and tumor-infiltrating neutrophils (TINs), where the enrichment of specific neutrophil subpopulations often correlates with tumor progression, immune evasion, and adverse clinical outcomes.1,2,3,4

Several studies reported that specific neutrophil subsets are key determinants of their immune-regulatory functions, influencing tumor progression both in mouse and human pancreatic, liver, and lung cancers.5,6,7,8,9,10 More interestingly, distinct neutrophil maturation stages have been demonstrated to impact their functions within the tumor environment (TME) and confer pro-tumoral activities.5,6,7,8 Clinically, this complexity presents a major challenge: without clear functional annotation or actionable targets, this heterogeneity risks becoming more of a conceptual burden than a therapeutic asset.

In their recent study, Liu et al.11 performed extensive transcriptomic analyses of different neutrophil developmental stages from both non-tumor bone marrow (BM) donors and BM and PB from a humanized cancer model, revealing the biological identity and therapeutic potential of myelocyte (MC) & metamyelocyte (MM)-stage neutrophils in a broad range of human cancers (Fig. 1).

Fig. 1: Immunosuppressive MC & MM-stage neutrophils originate in the human BM and expand in PB and tumor in cancer patients.
figure 1

MC & MM neutrophils express high levels of cell surface marker Galectin 3 and CD63, and their ability to suppress T cell proliferation indirectly promotes tumor growth. FLT3L overexpression in humanized mouse models of cancer shifts MC & MM neutrophils toward a monocyte-like phenotype and suppresses tumor growth by relieving immunosuppression.

Full size image

To examine their development in humans, BM neutrophils from non-tumor donors were separated into five key developmental stages using well-established gating strategies: promyelocyte (PM), MC, MM, band cell (BD), and segmented cell (SC), and subjected to bulk RNA-seq to generate stage-specific gene signatures. Bioinformatics analyses confirmed a continuous differentiation trajectory from PM to BD & SC stages, with MC and MM cells occupying a distinct, intermediate stage enriched for immunosuppressive gene expression within the BM. Further analyses highlighted that MC & MM neutrophils overlap with the intermediate developmental stages identified as N2 & N35 and Pre Neu & Immature Neu subsets4 from previous studies, and possess T cell-suppressive functions similar to PMN-MDSCs,1,8 already within the human BM. Of note, these immunosuppressive features were specific to human neutrophils, as healthy mouse BM neutrophils lacked suppressive function at all stages. However, mouse TINs did show suppressive activity, indicating that in mice, immunosuppression is acquired in the TME rather than being developmentally programmed in the BM as observed in humans.

Human MC & MM neutrophils sorted from non-tumor donors exhibited the strongest T cell suppression in vitro. In contrast, earlier (PM) and more mature (BD & SC) neutrophils lacked this immunosuppressive function. Further, MC & MM neutrophils survive longer than other neutrophil subsets, resist apoptosis more effectively, and exhibit low Ki67 positivity, consistent with their transitional status within the BM.

To explore the relevance of MC & MM neutrophils in tumor growth, the team employed a humanized immune system (HIS) mouse model using NOD/ShiLtJGpt-Prkdcem26Cd52Il2rgem26Cd22/Gpt (NCG) mice and transferred either MC & MM or BD & SC neutrophils sorted from healthy human BM into A375 melanoma-bearing mice. While BD and SC neutrophils had no effect on tumor growth, mice receiving MC and MM neutrophils developed significantly greater tumors.

Interestingly, flow cytometry analyses of clinical samples unveiled that MC & MM neutrophils are the predominant populations in TINs and PB in gastric cancer patients. This observation was further extended by applying the MC & MM neutrophil scores to publicly available single-cell RNA sequencing (scRNA-seq) datasets from 17 additional human cancer types. MC & MM neutrophils were consistently the most abundant subset among TINs, suggesting a conserved immunosuppressive program, and underscoring a broad and unique aspect of this study compared to existing literature.

To overcome the limitations of standard HIS mice in reconstituting peripheral human neutrophils, the team further developed a novel humanized mouse model by knocking out the Gfi1 gene — a key regulator of neutrophil development — in NCG mice. These NCG-Gfi1−/− HIS mice reconstituted with human neutrophils showed severe mouse neutropenia, but a robust peripheral and intratumor human neutrophil. Using this model, the authors verified the accumulation of MC & MM neutrophils in both BM and tumor from tumor-bearing NCG-Gfi1−/− HIS mice, closely recapitulating the distribution patterns observed in cancer patients.

scRNA-seq dataset analyses of neutrophils from non-tumor human BM donors and BM and tumors from tumor-bearing NCG-Gfi1−/− HIS mice allowed the identification of eight genes consistently upregulated in MC & MM neutrophils: four intracellular and four cell surface markers. Among these, CD63 and Galectin-3 emerged as particularly promising tumor-associated neutrophil surface markers, as their expression in tumor tissues correlated with worse patient survival across 32 cancer types, and was nearly absent in healthy donors, though abundant in cancer patients, especially within the MC & MM compartment.

From a therapeutic perspective, the authors demonstrated that MC & MM neutrophils exhibit the ability to trans-differentiate into non-immunosuppressive monocytic cells. MC & MM neutrophil exposure to FLT3L — a cytokine associated with monocytic differentiation — downregulated neutrophil markers and upregulated monocyte markers, such as CD14 and CD68 both in vitro and in vivo. Moreover, FLT3L overexpression in NCG-Gfi1−/− HIS reduced MC & MM neutrophil numbers, suppressed tumor growth, and enhanced anti-PD-L1 immunotherapy. However, the translational potential of the proposed Flt3L-based strategy warrants a more cautious interpretation. While conceptually innovative, this approach is supported by data from a single humanized mouse model, and its relevance in a clinical context remains uncertain.

In conclusion, the discovery that human, but not mouse, MC & MM-stage neutrophils dominate both the BM and TINs across a broad spectrum of cancers is particularly notable. It underscores the need for humanized models — like the NCG-Gfi1−/− HIS mouse system employed here — that better recapitulate human myelopoiesis and immune-tumor dynamics. This model not only enables the detailed study of human TINs in vivo but also opens new avenues for testing therapeutic strategies in a setting that closely mimics human physiology.