Tumor-associated neutrophil precursors impair homologous DNA repair and promote sensitivity to PARP inhibition

Abstract

Tumor evolution is one of the major mechanisms responsible for acquiring therapy-resistant and more aggressive cancer clones. Whether the tumor microenvironment through immune-mediated mechanisms might promote the development of more aggressive cancer types is crucial for the identification of additional therapeutic opportunities. Here, we identify a subset of tumor-associated neutrophils, defined as tumor-associated neutrophil precursors (PreNeu). These PreNeu are enriched in highly proliferative hormone-dependent breast cancers and impair DNA repair capacity. Mechanistically, succinate secreted by tumor-associated PreNeu inhibits homologous recombination, promoting error-prone DNA repair through non-homologous end-joining regulated by PARP-1. Consequently, breast cancer cells acquire genomic instability promoting tumor editing and progression. Selective inhibition of these pathways induces increased tumor cell killing in vitro and in vivo. Tumor-associated PreNeu score correlates with copy number alterations in highly proliferative hormone-dependent tumors from breast cancer patients. Treatment with PARP-1 inhibitors counteract the pro-tumoral effect of these neutrophils and synergize with endocrine therapy.

Fig. 1. Tumor-associated PreNeu predominantly infiltrate highly proliferative ER⁺ breast cancers.

a UMAP plots derived from PhenoGraph cluster analyses of concatenated low proliferative (n = 3) and highly proliferative (n = 3) breast cancer biopsies gated on CD45⁺ cells. Cluster 1: CD11b⁺, CD33⁺, CD66b⁺, CD15⁺, CD163⁻, LOX-1⁻; Cluster 2: CD11b⁺, CD33⁺, CD66b⁺, CD15⁺, CD163⁻, LOX-1⁺; Cluster 3: CD11b⁻, CD33⁻, CD66b⁺, CD15⁻, CD163⁻, Lox-1⁺; Cluster 4: CD11b⁻, CD33⁻, CD66b⁻, CD15⁻, CD163⁻, LOX-1⁻. b Bar graph displaying the frequency of the four clusters in normal tissues (NT, n = 3), low proliferative ER⁺ breast cancer biopsies (Low proliferative ER⁺ BC; n = 48) and highly proliferative breast cancer biopsies (Highly proliferative ER⁺ BC; n = 15). c UMAP plot derived from BD Rhapsody scRNA-seq cluster analysis of the neutrophil compartment in a highly proliferative ER+ tumor (n = 1). d Metascape analysis of enriched gene pathways of cluster C neutrophil compared to clusters A and B. Saturation of red color is proportional to p value significance. e Upregulated genes (FDR < 0.01; log2FoldChange > 1) in PreNeu compared to LOX-1⁺ neutrophils sorted from highly proliferative ER⁺ breast cancer biopsies resulting from single-cell differential expression analysis (Smartseq-2). Significantly upregulated genes are organized into a protein-protein interaction network generated using String Database. The dimension of the nodes is proportional to the highest statistical significance (represented as −10xlog10FDR). Saturation of red color is proportional to increased gene expression in PreNeu compared to LOX-1⁺ neutrophils (represented as log2FoldChange). Yellow color indicates the manually selected genes. Highlighted in yellow are genes belonging to the magnified subnetwork shown on the right of the figure panel. f UMAP plot derived from BD Rhapsody scRNA-seq data displaying the enrichment score of the PreNeu signature derived from scSMART-seq analysis based on significantly upregulated genes. g Pseudotime trajectory embedded in the UMAP plot with cluster C as the starting point of the trajectory (green line). h Bar graph showing the percentage of Ki67 positive cells in the different populations of neutrophils (from n = 3 patients). Data are represented as mean ± SEM. Statistical analyses (one-way analyses of variance, ANOVA, multiple comparison test). i Bar graph displaying the percentage of cluster 3 cells within CD45⁺ cells in normal tissues (NT, n = 3), low proliferative (LP, n = 48), or highly proliferative (HP, n = 15) ER⁺ breast cancer biopsies. Each symbol represents an individual patient. Data are represented as mean ± SEM. Statistical analyses (one-way analyses of variance, ANOVA, multiple comparison test). j Cluster 3 relative abundance on CD45+ cells stratified based on tumor grade (n = 9 for grade 1, n = 40 for grade 2, n = 13 for grade 3). Data are represented in violin plots with median (bold line) and quartile (regular lines). Statistical analyses (Brown-Forsythe and Welch ANOVA tests). k cluster 3 to cluster 2 ratio in NT (n = 3), LP (n = 48), and HP (n = 15) ER⁺ breast cancer biopsies. Each symbol represents an individual patient. Data are represented as mean ± SEM. Statistical analyses (one-way analyses of variance, ANOVA, multiple comparison test). l Bar graph displays the percentage of cluster 2 within CD45⁺ cells in NT (n = 3), LP (n = 48), and HP (n = 15) ER⁺ breast cancer biopsies. Each symbol represents an individual patient. Data are represented as mean ± SEM. Statistical analyses (one-way analyses of variance, ANOVA, multiple comparison test). Source data are provided as a Source data file.

Fig. 2. PreNeu impair homologous DNA repair, which leads to genomic instability.

a Experimental scheme created using BioRender. Garda, c. (2025) https://BioRender.com/yj15tp2. Briefly, ER⁺ MCF-7 breast cancer cells were irradiated and co-cultured with conditioned media (cm) from PreNeu for different time points (0 h, 2 h, 6 h, and 24 h) or control (Ø). b Quantification and representative images of the neutral comet assay of MCF-7 cells treated as indicated in a; black for Ø and red for cm-PreNeu). Each dot represents an individual comet from a pool of three biological replicates. Aggregated data from three independent experiments are reported as mean ± SEM (n = 10 for time point). Statistical analyses (two-tailed unpaired Student t test). c Heatmap showing fold change in protein levels using western blot of the reported proteins in MCF-7 cells treated as described in a. Values were normalized to the relative housekeeping gene followed by normalization on t = 0 for both untreated and treated MCF-7 cells upon irradiation (n = 3 biological replicates for time point, except n = 4 biological replicates for pATM/ATM). Statistical analysis (two-tailed unpaired t test). d Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/8prt8a8, Immunofluorescence images (left panel: DAPI blue, TIP60 green, P-53BP1 red; right panel: DAPI blue, p-ATM green, γH2AX red), and quantification of MCF-7 cells treated with either the condition media of the PreNeu or LOX-1⁺ neutrophil, or control (Ø). Scale bar 5 µm. (P53BP1: n = 91 for Ø and cm-PreNeu, n = 98 for cm-LOX-1⁺ neutrophils; TIP60: n = 141 for Ø, n = 149 for cm-PreNeu, n = 139 for cm-LOX-1⁺ neutrophils. γH2AX: n = 34 for Ø, n = 40 for cm-PreNeu, n = 31 for cm-LOX-1⁺ neutrophils. p-ATM: n = 31 for Ø, n = 50 for cm-PreNeu, n = 33 for cm-LOX-1⁺ neutrophils. Aggregated data from three biological replicates per condition. Data are reported as mean ± SEM. Statistical analysis (two-tailed unpaired t test). e Correlation analyses of PreNeu score with copy number alterations (R² = 0.32, p = 9.7e-05 linear regression). RNA sequencing and WES data were collected from 235 breast cancer cases treated with pre-operative chemotherapy. Data are reported with the regression line and confidence interval of 95%. Statistical analyses (simple linear regression). f Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/6n3mu0w. Female NRG mice were surgically implanted with estrogen pellets three weeks before the injection of MCF-7 Luciferase⁺ cells in the mammary fat pad (FP-MCF-7 Luc⁺). When the bioluminescence-based total flux reached more than 10⁶ counts, mice were randomly enrolled in the following groups: untreated (Ø; n = 8) or treated with PreNeu (30 × 10³, twice a week, i.p.; +PreNeu; n = 5). The mice were sacrificed after 4 months. g Immunofluorescence images (DAPI blue, RAD51 green, γH2AX red) and quantification of untreated vs PreNeu-treated MCF-7 tumors (n = 20 for each group). Scale bar 5 µm. Aggregated data from at least three tumors per condition. Data are reported as mean ± SEM. Statistical analyses (two-tailed unpaired Student t test). h Quantification and representative images of neutral comet assay using cells collected from MCF-7 tumors described in f (n = 20 for each group). Scale bar 5 µm. Data are reported as mean ± SEM. Statistical analyses (two-tailed unpaired Student t test). i Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/014yk4s. Tumors collected from mice treated as described in (f) were processed as single-cell suspensions and EpCaM+ epithelial cells purified to extract DNA. Whole-genome sequencing was performed. Hierarchical clustering of untreated (n = 2) and PreNeu-treated (n = 4) mice based on standardized whole-genome copy number calls (euclidean distance, average linkage), is shown on the left. The violin plot indicates the proportion of the cancer genome affected by aneuploidy. Source data are provided as a Source data file.

Fig. 3. Succinate is a major soluble factor produced by PreNeu and drives homologous recombination deficiency.

a LC-MS/MS quantification of the most upregulated metabolites. b Succinate quantified in sorted neutrophil precursors (PreNeu; n = 1: pool of three independent highly proliferative ER+ breast cancer patients) and LOX-1⁺ neutrophils (n = 1: pool of three independent highly proliferative ER⁺ breast cancer patients), cord blood-derived PreNeu (n = 5), cord blood-derived LOX-1⁺ neutrophils (n = 6), PyMT-N tumor-sorted PreNeu (n = 6) and tumor-associated neutrophils (n = 3). Data are represented as mean ± SEM. Statistical analyses (two-tailed unpaired Student t test). c Succinate dehydrogenase (SDH) activity quantified by SDH Colorimetric Assay in cord-blood derived PreNeu, LOX-1⁺ neutrophils and PreNeu (n = 4 for neutrophils and PreNeu, n = 5 for LOX-1⁺ neutrophils). Data are represented as mean ± SEM, one-way analysis of variance (ANOVA multiple comparison test). d Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/8prt8a8. Briefly, MCF-7 cells were irradiated at 10 Gy, then cultured in hormone deprivation and treated or control (Ø) with cm-PreNeu or succinate (2 mM). e Immunofluorescence images (DAPI blue, TIP60 green, P-53BP1 red) and quantification of irradiated MCF-7 cells treated or control (Ø) with condition media of PreNeu (p-53BP1: n = 51 for Ø, n = 50 for both Succinate and cm-PreNeu; TIP60: n = 49 for Ø and cm-PreNeu, n = 48 for Succinate). Data are represented as mean ± SEM. Statistical analysis (two-tailed unpaired Student t test). f Immunofluorescence images (DAPI blue, p-ATM green, γH2AX red) and quantification of irradiated MCF-7 cells treated or control (Ø) with condition media of PreNeu (γH2AX: n = 51 for Ø, n = 50 for Succinate, n = 52 for cm-PreNeu. p-ATM: n = 55 for Ø, n = 53 for Succinate, n = 58 for cm-PreNeu). Aggregated data from three biological replicates per condition. Data are represented as mean ± SEM. Statistical analysis (two-way ANOVA). g Cartoon created using BioRender showing the mechanism of succinate release by PreNeu and uptake/transport by tumor cells. Pharmacological blocking of SCL13a5 prevents succinate transport in tumor cells. h Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/8w9k8zb. Briefly, MCF-7 cells were irradiated at 10 Gy and then cultured in hormone deprivation and treated with either SCL13a5 blocking (αSCL13a5), cm-PreNeu, combinatorial treatment, or control. i Immunofluorescence images (Upper panel: DAPI blue, TIP60 green, P-53BP1 red. Down panel: DAPI blue, pATM green, γH2AX red) and quantification of irradiated MCF-7 cells treated or control (Ø) with cm-PreNeu or cm-PreNeu and αSLC13a5 (P53BP1: n = 48 for Ø and αSLC13a5, n = 45 for cm-PreNeu, n = 51 for cm-PreNeu+ αSLC13a5. TIP60: n = 38 for Ø, n = 30 for αSLC13a5, n = 35 for cm-PreNeu, n = 27 for cm-PreNeu+ αSLC13a5. γH2AX: n = 40 for Ø and αSLC13a5, n = 34 for cm-PreNeu, n = 43 for cm-PreNeu+ αSLC13a5. P-ATM: n = 23 for Ø, n = 35 for αSLC13a5, n = 28 for cm-PreNeu, n = 31 for cm-PreNeu + αSLC13a5). Aggregated data from three biological replicates per condition. Data are represented as mean ± SEM. Statistical analysis (two-way ANOVA). j Quantification and representative images of neutral comet assays of MCF-7 cells treated or control (Ø) with cm-PreNeu or αSLC13a5 (n = 47 for Ø, n = 57 for αSLC13a5, n = 36 for cm-PreNeu, n = 77 for cm-PreNeu+ αSLC13a5). Aggregated data from three biological replicates per condition. Data are represented as mean ± SEM. Statistical analysis (two-way ANOVA). k Cartoon created in BioRender. Garda, c. (2025) https://BioRender.com/svwrqhv showing the mechanism of pharmacological inhibition of succinate release by MCT1 inhibitor and the experimental scheme. Briefly, conditioned media form bone marrow-derived PreNeu (^bmPreNeu) was generated in the presence of MCT1 inhibitor (cm-^bmPreNeu^MCT1i) or control (cm-^bmPreNeu^Ø) and used to pre-treat PyMT-N cells in vitro for 48 hours. Then, cancer cells were implanted in the fat pad (FP) of NRG female mice, and tumors were collected when palpable. l Quantification and representative images of neutral comet assays of tumors whose cancer cells were pre-treated before the experiment with cm-^bmPreNeu^MCT1i or control (n = 135 for Ø, n = 128 for cm-PreNeu^Ø, n = 186 for cm-PreNeu^MCT1i). Aggregated data from three biological replicates per condition. Data are represented as mean ± SEM. Statistical analysis (one-way ANOVA). m Immunofluorescence images (DAPI blue, RAD51 green, γH2AX red) and quantification of tumors from mice whose cancer cells were pre-treated before the experiment with cm-^bmPreNeu^MCT1i or control (γH2AX: n = 16 for Ø, n = 16 for cm-PreNeu^Ø and cm-PreNeu^MCT1i. RAD51: n = 31 for Ø, n = 21 for cm-PreNeu^Ø, n = 21 for cm-PreNeu^MCT1i). Aggregated data from three biological replicates per condition. Data are reported as mean ± SEM. Statistical analysis (one-way ANOVA). Source data are provided as a Source data file.

Fig. 4. Breast tumor cells challenged with the secretome of PreNeu rely on PARP to survive.

a Quantification and representative images of western blots for PARP1, Ku70, Ku80, and DNA-PKcs levels in MCF-7 cells treated with or without conditioned media obtained from human cord blood-derived neutrophil precursors (cm-PreNeu; n = 4 for all the molecules), except n = 7 for PARP1). Data are reported as mean ± SEM. Statistical analyses (unpaired Student t test). b Quantification and representative image of western blot for PARP1 in MCF-7 cells silenced or not for the succinate receptor (SUCNR1) and treated or not with cm-PreNeu (n = 4 for each group). Data are reported as mean ± SEM. Statistical analysis (one-way ANOVA). c Quantification and representative image of western blot for PARP1 in MCF-7 cells silenced or not for the succinate receptor (SUCNR1) and treated or not with succinate (n = 3 for each group). Data are reported as mean ± SEM. Statistical analysis (unpaired t test). d Graph showing fold change in mRNA level of PARP1 in MCF-7 cells and LNCaP cells. Cells were previously irradiated at 10 Gy and then treated or control (Ø) with cm-PreNeu or succinate (2 mM) for 6 h (n = 3 for each group). Data are reported as mean ± SEM. Statistical analysis (unpaired t test). e Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/uueypah. PARP1-luciferase MCF-7 cells were transfected with pGL3 plasmid, treated or control (Ø) with cm-PreNeu in the presence or absence of αSLC13a5 for 6 hours. Cells were analyzed for luciferase activity using Dual-Glo Luciferase Assay System (Promega, E2920) following manufacturer’s instructions. Expression of PARP1 (Renilla) was normalized to the expression of pGL3 (Firefly) to calculate luciferase activity. The values were further normalized to the condition treated with cm-PreNeu. (n = 7 for each group, each dot is the mean of three independent technical replicates from two independent experiments). Data are reported as mean ± SEM. Statistical analysis (unpaired t test). f Cell proliferation of MCF-7 cells treated or control (Ø) with cm-PreNeu and challenge with PARPi (1 µM) after 24 h. g Annexin V-positive intensity in MCF-7 cells treated or control (Ø) with cm-PreNeu and challenge with PARPi (1 µM) after 24 h. f, g Aggregated data from three independent experiments are reported as mean ± SEM. Statistical analyses (two-way ANOVA and Ordinary one-way ANOVA multiple comparison test, respectively). h–j Tumor growth from the h E0771, i PyMT-N, and j PyMT-M models in female mice treated with PARPi or vehicle (Ø) when tumors were palpable as indicated in the graph (for the E0771 model n = 4 for each group, for the PyMT-N model n = 4 for each group, for the PyMT-M model n = 5 for Ø and n = 4 for PARPi groups). Data are reported as mean ± SEM. Statistical analysis (two-way ANOVA). k Representative histology images and their quantification. H&E and Cleaved-Caspase 3 immunohistochemical staining (Cleaved-Caspase 3 brown; nuclei, blue) of representative E0771, PyMT-N, and PyMT-M tumors from mice at the endpoint, treated with or without PARPi (n = 8 for E0771 + Ø, n = 9 for E0771 + PARPi, n = 30 for both PyMT-N and PyMT-M models; mean of three sections per mouse, ≥3 fields per section). Magnification ×20. Scale bar 200 µm. Quantification of Cleaved-Caspase 3 was reported as a percentage of the total cells within the tumor tissues. Data are represented as mean ± SEM. Statistical analyses (Unpaired Student t test). Source data are provided as a Source data file.

Fig. 5. PreNeu enhance PARP inhibition in vivo and synergize with standard endocrine therapy.

a Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/uueypah. Female NRG mice were implanted with estrogen pellets 2 weeks before of the injection of 1 × 10⁶ FP-MCF7-Luc+ cells. Starting from this point, every week mice received ^cbPreNeu (i.p.) and the radiance of the tumors was monitored by luminescence. When the total flux of the tumor was >10⁶, mice were treated as indicated: untreated (Ø, n = 5); Tamoxifen (n = 5), Olaparib (n = 5), neutrophil precursors (PreNeu; n = 9); PreNeu + tamoxifen (n = 6); PreNeu + Olaparib (n = 14); PreNeu + Olaparib + Tamoxifen (n = 16). After 12 weeks of treatment, mice were sacrificed. b Graph showing tumor area (mm²) of mice belonging to the different groups: untreated (Ø, n = 5); Tamoxifen (n = 5), Olaparib (n = 5), PreNeu (n = 9); PreNeu + tamoxifen (n = 6); PreNeu + Olaparib (n = 14); PreNeu + Olaparib + Tamoxifen (n = 16). Synergism was calculated with the coefficient of drug interaction (CDI). CDI (cm-PreNeu + Olaparib + Tamoxifen) = 0.59. Statistical analyses (two-tailed unpaired Student t test). c Quantifications and representative images of immunohistochemistry analysis (H&E, Ki67, γH2AX) of mice treated in the different conditions listed in a. Aggregated data from one tumor per mouse, mean of three sections per mouse, ≥3 fields per section. Statistical analyses (unpaired Student t test). d Experimental scheme created in BioRender. Garda, c. (2025) https://BioRender.com/6n3mu0w. Female NRG mice were injected with either HCC1937 BRCA^wt or HCC1937 BRCA^mut cells in the mammary fat pad. Starting from when tumors were palpable, mice were injected either with olaparib or PreNeu or PreNeu + olaparib or vehicle (Ø). e, f Tumor growth from mice injected either with e HCC1937 BRCA^mut or f HCC1937 BRCA^wt and treated as described in d (n = 5 mice for each group). Data are represented as mean ± SEM. Statistical analysis (two-way ANOVA). g, h Quantifications and representative images of immunohistochemistry analysis (H&E, Ki67, γH2AX) of mice injected either with g HCC1937 BRCA^mut or h HCC1937 BRCA^wt and treated as described in d. g Ki67: n = 10 for each groups, for γH2AX n = 11 for PreNeu and Olaparib groups, n = 10 for Ø and PreNeu + Olaparib. f Ki67: n = 12 for PreNeu, n = 8 for Ø and PreNeu + Olaparib, n = 10 for Olaparib. γH2AX: n = 10 for PreNeu and Ø, n = 9 for Olaparib, n = 8 for PreNeu+Olaparib. Mean of three sections per mouse, ≥3 fields per section. Statistical analyses (one-way ANOVA). Source data are provided as a Source data file.