Deciphering neutrophil dynamics in the focal lesion tumor microenvironment to overcome immunosuppression in multiple myeloma

Abstract

Understanding the roles of myeloid cells in the tumor microenvironment (TME) has emerged as a promising strategy to identify novel targets to counteract the immunosuppressive barriers protecting multiple myeloma (MM). Neutrophils are a new cancer research focus due to their potential to reduce the efficacy of immune-based therapies. This study aimed to deepen understanding of neutrophil function in MM by analyzing freshly isolated myeloid cells from paired focal lesions (FLs) and bone marrow using single-cell RNA sequencing, immunofluorescence imaging, and functional assays. We describe 3 distinct CXCR2+ mature neutrophil subsets: TREM1+CD10+, RETN+LCN2+, and TNFAIP3+CXCL8+, each exhibiting unique phenotypes within the TME. Notably, the TREM1+CD10+ subset was highly prevalent, particularly in FLs, demonstrating potent immunosuppressive effects on T cells. This subset's gene signature was correlated with shorter overall survival (OS) in a large data set of patients with MM, underscoring its clinical significance. Targeted inhibition of neutrophil activity through CXCR2 blockade, alone or combined with standard anti-MM therapies, significantly reduced tumor burden, improving OS in preclinical MM models. These insights into neutrophil-mediated immunosuppression in MM provide valuable knowledge regarding mechanisms driving immune evasion, and reveal new therapeutic approaches to enhance the efficacy of MM treatment.

Graphical abstract

Figure 1.

CD11b⁺ myeloid cells from BM and FLs from patients with MM, and HBM. (A) Samples from patients with NDMM (n = 6), RRMM (n = 7) (BM and FLs), and from HBM (n = 3) are shown in the UMAP representation of 105 192 CD11b⁺ cells captured with 10× V3 5ʹ scRNA-seq (see “Materials and methods” for complete details). Nine clusters observed were: (1) TREM⁺MME⁺ mature neutrophils, (2) S100A8⁺CD14⁺ macrophages (Mϕ), (3) S100A8/9⁺MMP9⁺ immature neutrophils, (4) LTF⁺CAMP⁺ immature neutrophils, (5) S100A8⁺LTF⁺ immature neutrophils, (6) TNFAIP3⁺CXCL8⁺ mature neutrophils, (7) RETN⁺LCN2⁺ mature neutrophils, (8) Mki67⁺DEFA3⁺ preneutrophils, and (9) HLA-DR⁺CD16⁺ macrophages. Bottom panels represent cell contributions from HBM (left), NDMM (middle), and RRMM (right). (B) UMAP representation of CD11b⁺ cells observed. The contribution of BM-derived cells (n = 59 861 cells) is shown in blue. The contribution of FL-derived cells (n = 45 331 cells) is shown in red. (C) Scaled bar graph representation of each cell type contribution (y-axis) from healthy donors and patients with MM (x-axis; n = 16). Color convention as per panel A. Top panel, cell type distribution from BM cells; bottom panel, cell type distribution from FLs (1 sample without cells). (D) Scaled normalized heat map of top 10 transcripts associated with each cluster. Sample downsized to 5000 cells per cluster. Blue represents downregulated genes, and red represents upregulated (log₂ fold-change) genes in cell types shown at top.

Figure 2.

Neutrophils associated with FLs acquire a protumor phenotype. (A) RNA velocity of the reanalyzed neutrophil data set excluding macrophages. UMAP embeddings from 84 616 cells. Color convention as per Figure 1A. The visual representation of neutrophil maturation from HBM (left), NDMM (middle), and RRMM (right). Scaled bar graphs below HBM and patient-specific UMAP represent scaled contribution of immature (gray) and mature (blue) neutrophils. (B) UMAP representation of neutrophils observed. The contribution of BM cells (n = 49 016 cells) is shown in blue. The contribution of FL cells (n = 35 600 cells) is shown in red. (C) Clusters scored against mature neutrophil signature (see “Materials and methods” for details). High scores represent transcripts within clusters most like mature neutrophils. (D) Highly variable features associated with neutrophil maturity in single cell data. Data are organized by early to late neutrophil development. Blue indicates BM cells, and red indicates FL cells. (E) Differentially expressed transcripts in neutrophils between BM (left, blue) and FLs (right, red). Features were selected to emphasize proinflammatory or immunosuppressive phenotypes. (F) Paired gene ontogeny analysis of mature neutrophils from FLs vs neutrophils from HBM.

Figure 3.

Mature neutrophil representation at sites of FLs. (A) The mature neutrophils were reclustered and represented using highly variable features. UMAP is plotted in the second and third dimension for visual optimization. Bottom panels represent cell contributions from HBM (left), NDMM (middle), and RRMM (right). (B) UMAP representation of reclustered mature neutrophils observed. In blue, the contribution of cells derived from BM (n = 11 935 cells). In red, the contribution of cells from sites of FLs (n = 23 913 cells). (C) Dot plot depicting the top 8 transcripts from each mature neutrophil cluster. The radius of each circle represents the percentage of cells expressing feature within cluster, and color represents average log₂ fold-change. (D) Schematic of CXCR2⁺ neutrophil developmental trajectories in patients with MM. (E) The fluorescence-activated cell sorter gating strategy involved an initial selection of CD11b⁺ neutrophils and C5AR1⁺ cells, followed by further selection based on CXCR2⁺ and CXCL8⁺ expression. Subsequently, the 3 CXCR2⁺ neutrophil clusters were identified using CD10 and TNFAIP3 gates. The analysis quantified the frequency of these 3 neutrophil subsets sampled from the BM of 3 individual patients. (F) Confocal immunofluorescence images of CXCR2⁺ neutrophils with markers associated with subset features identified in scRNA-seq (right) and CXCR2^– neutrophils with decreased marker expression (right). The data presented in panels E-F are representative of findings from 3 different patients.

Figure 4.

Visualization of heterogeneity of CXCR2⁺ TAN infiltration in MM BM and FLs. Vectra Polaris staining was applied to formalin-fixed, paraffin-embedded sections from 12 patient samples (BM, n = 12; paired FL, n = 12). Representative Vectra images from 3 individual patient samples ([A-C, top rows] original magnification ×10; [A-C, bottom rows] original magnification ×30). Differential marker and cell distribution of CD138⁺ tumor cells and CD16⁺ cells. (A) The absence of CXCR2⁺ cells in BM with <1% malignant PCs (CD138⁺). (B) Accumulation of CXCR2⁺ neutrophils along with an increase of CD138⁺ cells in BM. (C) Extensive intratumoral infiltration of CXCR2⁺ neutrophils with increased CD138⁺ cells. (D-F) Representative visualizations of the 3 mature TAN populations: (1) CXCR2⁺CD10⁺TNFAIP3⁺ cells (D, arrow), (2) CXCR2⁺CD10⁺TNFAIP3^– cells (E, arrow), and (3) CXCR2⁺CD10^–TNFAIP3^– cells (F, arrow). These data highlight the TAN heterogeneity and diversity within neutrophil populations in MM, emphasizing distinct phenotypic and functional subsets that may contribute variably to the TME.

Figure 5.

CXCR2⁺CD10⁺ neutrophils are inflammatory and immunosuppressive. (A) The percentage of CXCR2⁺CD10⁺ neutrophils in BM and FLs from patients with MM (MM BM and MM FLs, respectively) and HBM. CD138^–CD3^–CD56^–CSF1R^–CD11b⁺CD10⁺CXCR2⁺C5AR1⁺ were characterized as mature neutrophils. CXCR2⁺CD10⁺ neutrophils accumulated in MM BM and MM FL samples. (B) Confocal microscopy and Wright Giemsa staining of CXCR2⁺ and CXCR2^– neutrophils from MM microenvironment; CXCR2⁺ neutrophils were mature with increased nuclear segmentation (lobulation), and CXCR2^– neutrophils were immature cells with nuclear hypolobulation. CXCR2 was marked by fluorescein isothiocyanate anti-CXCR2 (green), while the nucleus was stained with NucSpot 750/780 (red). These imaging results further emphasize the high purity of the CXCR2⁺ neutrophil sorting, underscoring the precision in identifying distinct neutrophil populations within the MM microenvironment. (C) CD11b⁺C5AR1⁺CXCR2^– (immature) and CD11b⁺C5AR1⁺CXCR2⁺ (mature) neutrophils were isolated from MM BM and MM FLs. Sorted neutrophils were cultured in complete media for 24, 48, and 72 hours. Cells were harvested and stained for Annexin V at each time point. CXCR2⁺ cells had significantly shorter survival compared with CXCR2^– cells. (D) CD138^–CD3^–CD56^–CSF1R^–CD11b⁺CD10⁺CXCR2⁺C5AR1⁺ neutrophils isolated from HBM (n = 3), MM BM (n = 3), and MM FLs (n = 3). Sorted neutrophils were cultured in complete media for 24 hours, and supernatants were collected (see “Materials and methods” for details). Multiplex enzyme-linked immunosorbent assay protein measurement was performed on collected supernatants. Several markers of inflammation were elevated in MM FLs compared with MM BM and HBM. CXCL10 was elevated in HBM. (E-F) HBM, MM BM, and MM FL neutrophils were cocultured (1:4 ratio of neutrophils to T cells) with healthy carboxyfluorescein succinimidyl ester–stained T cells activated with CD28/CD3 antibodies in the presence of IL-2. On day 4, T-cell activation was analyzed using CFSE dilution by flow cytometry. Proliferation index, with a value of 1.0 representing baseline (no proliferation), serves as the reference point. (E) Representative flow cytometry histograms of T-cell proliferation. (F) FL neutrophils showed significantly higher immunosuppression of T-cell proliferation compared with BM neutrophils from the same (paired) patient with MM. (G) Kaplan-Meier curves of OS (left) and PFS (right) for patients with MM from the MMRF CoMMpass data set stratified by the “immature neutrophil” signature. (H) OS (left) and PFS (right) curves for patients with MM from the same data set stratified by the “mature neutrophil” signature. (I) T-cell activation and exhaustion signatures were calculated from the MMRF CoMMpass data set, and compared based on the ratio of “high mature neutrophil” and “low mature neutrophil” signatures. An exhausted T-cell signature was significantly associated with the “high mature neutrophil” signature. One-way analysis of variance with Tukey multiple comparison tests was used to test significant differences between groups in panels A,C. One-way analysis of variance with Kruskal-Wallis multiple comparisons test was used to compare means between groups in panel F; statistics were derived from BM and FLs from 3 HBM and 3 patients with MM. Experiments were performed in triplicate in panels D,F. ∗P < .05; ∗∗P < .01 (in all experiments). Data are presented as mean ± standard deviation.

Figure 6.

CXCR2 blockade promotes MM control in murine MM models. Vk12653 murine MM cell lines were injected into recipients for different in vivo tumor models (see “Materials and methods” for full details). We used the following models: the NDMM model (tumor cell injection without irradiation, n = 40 from 2 experiments), the RRMM model (tumor cell injection followed by total body lethal irradiation, 1000 cGy, and syngeneic [autologous] stem cell transplant, n = 40 from 2 experiments), the CD8⁺ T-cell depletion mouse model (tumor cell injection without irradiation, n = 40 from 1 experiment), and the NSG mouse model (tumor cell injection without irradiation, n = 30 from 1 experiment). Progression of MM was confirmed by weekly serum protein electrophoresis measurement of the monoclonal protein (M-spike) to albumin ratio in the NDMM model. In the RRMM model, the mice were classified as MM-relapsed or MM-remission based on the M-spike to albumin ratio. Recipient BM was harvested and stained with appropriate antibodies. Results were analyzed via flow cytometry. (A) Quantification of frequency and number of CXCR2⁺ neutrophil clusters from the samples of CBM (no tumor injection), MTB, and HTB, the latter 2 based on serum protein electrophoresis measurements. Cells gated from CD11b⁺Ly6G⁺Ly6C^– population. (B) Quantitation of M-spike development, and survival curve in the NDMM model. (C) UMAP representation of immune cells (excluding macrophages and neutrophils) observed in various experimental arms in the NDMM model. Cell type distribution highlighted below UMAP to show changes in cell kinetics upon treatment. (D) Concatenated mature neutrophils show distinct developmental stages observed in humans with little variance among experimental conditions: EIN represent $\overline{\mathrm{x}}$ = 22.39%, σ = 2.06%, LIN represent $\overline{\mathrm{x}}$ = 22.87%, σ = 1.36%, EMN represent $\overline{\mathrm{x}}$ = 28.23%, σ = 0.66%, LMN represent $\overline{\mathrm{x}}$ = 26.5%, σ = 2.96%. Neutrophils were scored with the same signatures as Figure 2. (E) Quantitation of M-spike development, and survival curve in T-cell depletion mouse model. (F) Quantitation of M-spike development, and survival curve in NSG mouse model. (G) Quantitation of M-spike development, survival curve, and CD8⁺ T-cell cytokine expression of IFN-γ and TNF-α. Data represent mean ± standard error of the mean. The 2-way analysis of variance was used to analyze statistical significance among tumor growth in different groups. Survival data are presented as percent survival (log-rank Mantel-Cox test). ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. CBM, control bone marrow; EIN, early immature neutrophils; EMN, early mature neutrophils; HTB, high tumor burden bone marrow; LIN, late immature neutrophils; LMN, late mature neutrophils; MTB, moderate tumor burden bone marrow; ns, not significant; PBS, phosphate buffered saline; TNF-α, tumor necrosis factor-α.

Figure 6.

CXCR2 blockade promotes MM control in murine MM models. Vk12653 murine MM cell lines were injected into recipients for different in vivo tumor models (see “Materials and methods” for full details). We used the following models: the NDMM model (tumor cell injection without irradiation, n = 40 from 2 experiments), the RRMM model (tumor cell injection followed by total body lethal irradiation, 1000 cGy, and syngeneic [autologous] stem cell transplant, n = 40 from 2 experiments), the CD8⁺ T-cell depletion mouse model (tumor cell injection without irradiation, n = 40 from 1 experiment), and the NSG mouse model (tumor cell injection without irradiation, n = 30 from 1 experiment). Progression of MM was confirmed by weekly serum protein electrophoresis measurement of the monoclonal protein (M-spike) to albumin ratio in the NDMM model. In the RRMM model, the mice were classified as MM-relapsed or MM-remission based on the M-spike to albumin ratio. Recipient BM was harvested and stained with appropriate antibodies. Results were analyzed via flow cytometry. (A) Quantification of frequency and number of CXCR2⁺ neutrophil clusters from the samples of CBM (no tumor injection), MTB, and HTB, the latter 2 based on serum protein electrophoresis measurements. Cells gated from CD11b⁺Ly6G⁺Ly6C^– population. (B) Quantitation of M-spike development, and survival curve in the NDMM model. (C) UMAP representation of immune cells (excluding macrophages and neutrophils) observed in various experimental arms in the NDMM model. Cell type distribution highlighted below UMAP to show changes in cell kinetics upon treatment. (D) Concatenated mature neutrophils show distinct developmental stages observed in humans with little variance among experimental conditions: EIN represent $\overline{\mathrm{x}}$ = 22.39%, σ = 2.06%, LIN represent $\overline{\mathrm{x}}$ = 22.87%, σ = 1.36%, EMN represent $\overline{\mathrm{x}}$ = 28.23%, σ = 0.66%, LMN represent $\overline{\mathrm{x}}$ = 26.5%, σ = 2.96%. Neutrophils were scored with the same signatures as Figure 2. (E) Quantitation of M-spike development, and survival curve in T-cell depletion mouse model. (F) Quantitation of M-spike development, and survival curve in NSG mouse model. (G) Quantitation of M-spike development, survival curve, and CD8⁺ T-cell cytokine expression of IFN-γ and TNF-α. Data represent mean ± standard error of the mean. The 2-way analysis of variance was used to analyze statistical significance among tumor growth in different groups. Survival data are presented as percent survival (log-rank Mantel-Cox test). ∗P < .05; ∗∗P < .01; ∗∗∗P < .001. CBM, control bone marrow; EIN, early immature neutrophils; EMN, early mature neutrophils; HTB, high tumor burden bone marrow; LIN, late immature neutrophils; LMN, late mature neutrophils; MTB, moderate tumor burden bone marrow; ns, not significant; PBS, phosphate buffered saline; TNF-α, tumor necrosis factor-α.