Characterizing Neutrophil Subtypes in Cancer Using scRNA Sequencing Demonstrates the Importance of IL1β/CXCR2 Axis in Generation of Metastasis-specific Neutrophils

Abstract

Neutrophils are a highly heterogeneous cellular population. However, a thorough examination of the different transcriptional neutrophil states between health and malignancy has not been performed. We utilized single-cell RNA sequencing of human and murine datasets, both publicly available and independently generated, to identify neutrophil transcriptomic subtypes and developmental lineages in health and malignancy. Datasets of lung, breast, and colorectal cancer were integrated to establish and validate neutrophil gene signatures. Pseudotime analysis was used to identify genes driving neutrophil development from health to cancer. Finally, ligand-receptor interactions and signaling pathways between neutrophils and other immune cell populations in primary colorectal cancer and metastatic colorectal cancer were investigated. We define two main neutrophil subtypes in primary tumors: an activated subtype sharing the transcriptomic signatures of healthy neutrophils; and a tumor-specific subtype. This signature is conserved in murine and human cancer, across different tumor types. In colorectal cancer metastases, neutrophils are more heterogeneous, exhibiting additional transcriptomic subtypes. Pseudotime analysis implicates IL1β/CXCL8/CXCR2 axis in the progression of neutrophils from health to cancer and metastasis, with effects on T-cell effector function. Functional analysis of neutrophil-tumoroid cocultures and T-cell proliferation assays using orthotopic metastatic mouse models lacking Cxcr2 in neutrophils support our transcriptional analysis. We propose that the emergence of metastatic-specific neutrophil subtypes is driven by the IL1β/CXCL8/CXCR2 axis, with the evolution of different transcriptomic signals that impair T-cell function at the metastatic site. Thus, a better understanding of neutrophil transcriptomic programming could optimize immunotherapeutic interventions into early and late interventions, targeting different neutrophil states.

Significance: We identify two recurring neutrophil populations and demonstrate their staged evolution from health to malignancy through the IL1β/CXCL8/CXCR2 axis, allowing for immunotherapeutic neutrophil-targeting approaches to counteract immunosuppressive subtypes that emerge in metastasis.

FIGURE 1

Characterization of neutrophil signatures and lineages in health and PTs. A, Uniform Manifold Approximation and Projection for Dimension (UMAP) plot of healthy neutrophils grouped by tissue type. BM: healthy bone marrow, SP: healthy spleen, L: healthy lung, BL: healthy blood. B, UMAP plots of healthy and tumor-derived neutrophils grouped by tissue type and Seurat clusters (0–15). BM: healthy bone marrow, SP: healthy spleen, L: healthy lung, BL: healthy blood, L_AC: lung adenocarcinoma, KPN: colorectal cancer with Kras, Trp53 and Notch mutations. C, UMAP plot of neutrophils in mouse breast cancer model. WT: healthy breast tissue, PyMT: polyomavirus middle-T oncoprotein tumor. D and E, Scoring of H_enriched and T_enriched neutrophil signatures. F, UMAP plot of neutrophils in mouse colorectal cancer model. All neutrophils are tumor-derived. G and H, H_enriched and T_enriched signatures in colorectal cancer (CRC). I, UMAP plot of human NSCLC neutrophils. Blood: blood-derived, tumor: tumor-derived. J and K, H_enriched and T_enriched signatures NSCLC. L, UMAP plot of human breast carcinoma (BC) neutrophils. Most neutrophils are tumor-derived. M and N, H_enriched and T_enriched signatures in breast carcinoma. O–R, Unsupervised pseudotime analysis of neutrophils in mouse and human datasets. Lineages in the individual datasets are numbered. S–V,IL1β/IL1β is differentially expressed at the end of tumor-specific lineages. W–Z, Estimated smoothers for IL1β/IL1β expression over pseudotime across the different lineages.

FIGURE 2

Characterization of neutrophils in metastasis. A, UMAP of neutrophils in CRCLM. B, Coexpression of H_enriched and T_enriched signatures. One cluster is not enriched for either signature (blue arrow). C and D, Unsupervised pseudotime analysis and estimated smoothers for TXNIP expression over the different numbered pseudotime lineages. E, Coexpression of TXNIP and CXCR2. F and G, Expression and estimated smoothers for CXCL8 over pseudotime. H, Coexpression of CXCL8 and IL1β. I, IHC staining of TXNIP in a patient CRCLM sample at 4x (left) and 10x (right). Scale bars = 50 µm. J and K, IHC staining of ITGAM (Neutrophils) and CD3 (T cells) in a patient CRCLM sample at 4x, scale bars = 50 µm. Dashed squares indicate regions where immune cells cluster. L, Differentially expressed genes in metastasis-specific neutrophil cluster. M, H_enriched, T_enriched, and M_enriched gene signatures in mouse bulk-RNA-seq neutrophil dataset. L, Healthy liver tissue, PT: Primary tumor, LMET: Liver metastasis. N, Averaged expression of the individual genes of the three signatures. O, UMAP plot of integrated human neutrophils from PT and metastatic (M) datasets of different cancers. P, Coexpression of CXCR2 with IL1β (top) and TXNIP (bottom). Q, Differential gene expression between neutrophils in malignancy compared with PT. R and S, GO and KEGG analysis of M_CRC neutrophils.

FIGURE 3

CD4⁺ T cells are transcriptomically altered in metastasis. A and B, UMAP plots of CD4⁺ and CD8⁺ T cells in PT and LM grouped by cell type and tumor type, respectively. C, Differential gene expression of CD4⁺ T cells in LM compared with PT. D and E, GO and KEGG analyses of metastatic CD4⁺ T cells. F and G, Global cell-cell communication network and the interaction strengths between neutrophils, CD4⁺, CD8⁺, and Tregs. H–J, The contribution of each L-R pair to the overall signaling pathway for CXCL, IL1, and TNF pathways. K–M, Visualization of the cell-cell communication patterns mediated by each significant L-R pair for the three pathways.

FIGURE 4

Signaling patterns in CRCLM. A and B, Outgoing and incoming signaling patterns in CRCLM. C, Cellular roles as dominant senders (sources) and receivers (targets) in CRCPT and LM. CRCPT dataset did not contain any Tregs or neutrophils. D, UMAP plot of neutrophil subtypes in CRCLM. E, Interaction strength of the global communication patterns between neutrophil, T-cell, and macrophage subtypes. F–I, Outgoing signal strengths from TXNIP+ neutrophils, SPP1+ macrophages, M1- and M2-like macrophages, respectively. J, The contribution of each L-R pair to the overall CXCL signaling pathway among macrophage, neutrophil, and T-cell subtypes. K–M, Visualization of the cell-cell communication patterns mediated by the most significant L-R pairs in CXCL pathway. N, Heat map showing the relative strengths of the significant outgoing and incoming signaling patterns in all communication pathways with dominant sender and receiver immune cell subtypes. The top colored bar plot represents the sum of column of values displayed in the heat map representing the different cell populations. The right gray bar plots represents the sum of row of values, representing the different signaling pathways.

FIGURE 5

Loss of neutrophil-specific CXCR2 does not alter suppression of T-cell proliferation in neutrophils from the metastatic niche in mice bearing tumors. Neutrophils from the liver and blood of littermate mice either Mrp8-Cre-Cxcr2^+/+ (Mrp8CreN) or Mrp8-Cre-Cxcr2^fl/fl (Mrp8CreYCxcr2Δneut) mice were harvested at clinical endpoint 32-33 days following intrasplenic injection of villinCreER Kras^G12D/+Trp53^fl/flRosa26^N1CD/+ (KPN) cells derived from digestion of organoids from a single organoid line. These were placed in coculture with T cells from WT mice with CD3/CD28 stimulating beads and proliferation of CD3⁺ CD4⁺ and CD8⁺ T cells, and analyzed through flow cytometry 40 hours later. A, Flow cytometry gating of T cells and neutrophils from coculture. Cells were gated on FSC-A/SSC-A (not shown), FCS-A/FSC-H for singlets, Live/Dead, CD3, and CD4/CD8 to define CD3⁺, CD3⁺/CD4⁺, and CD3⁺/CD8⁺ populations for analysis. Cell trace yellow allows for tracking of T-cell proliferation. The signal for each T-cell division becomes subsequently dimmer, and allows for calculation of the number of T-cell generations. Gating obtained from stimulated T cell only control (blue) and unstimulated T cell control (red). B, Proportion of CD3⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers (B–J, cocultures from 7 Mrp8CreN mice and 5 Mrp8CreYCxcr2Δneut mice). C, Expansion index of CD3⁺ T cells in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. D, Division index of CD3⁺ T cells in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. E, Proportion of CD4⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. F, Expansion index of CD4⁺ T cells in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. G, Division index of CD4⁺ T cells in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. H, Proportion of CD8⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. I, Expansion index of CD8⁺ T cells in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. J, Division index of CD8⁺ T cells in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from tumor-bearing livers. K, Proportion of CD3⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice (K–S, 7 Mrp8CreN mice and 5 Mrp8CreYCxcr2Δneut mice). L, Expansion index of CD3⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. M, Division index of CD3⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. N, Proportion of CD4⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. O, Expansion index of CD4⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. P, Division index of CD4⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. Q, Proportion of CD8⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. R, Expansion index of CD8⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. S, Division index of CD8⁺ T cells that proliferated in coculture with Mrp8CreN and Mrp8CreYCxcr2Δneut neutrophils from blood of tumor-bearing mice. *, P < 0.05 on unpaired t test. **, P < 0.01 on unpaired t test.