Dual ENPP1/ATM depletion blunts DNA damage repair boosting radioimmune efficacy to abrogate triple-negative breast cancer

Abstract

The ATP-hydrolytic ectoenzyme ENPP1 has been implicated in the metastasis and recurrence in triple-negative breast cancer (TNBC), primarily by contributing to tumor cell survival and treatment resistance. However, the precise mechanisms remain unclear. In a model of local recurrence (LR), circulating tumor cells (CTC) engrafting in the post-resection tumor bed developed a radioresistant phenotype linked to an ENPP1⁺-gene signature which was also identified in TNBC patients, suggesting ENPP1´s role in genome integrity. Blockade of ENPP1 using a permeable ENPP1 inhibitor (AVA-NP-695) reduced radioresistance, mechanistically attributed to decreased homologous recombination (HR) resulting in persistent DNA damage, as evidenced by enhanced tail moment and sustained γH2AX formation. This impaired DNA damage repair (DDR) sensitized tumor cells to ionizing radiation (IR). Notably, several DDR inhibitors (i) (including PARPi and ATMi) showed the highest synergy score in a targeted pharmacological screening. In vivo, dual ENPP1/ATM inhibition heightened radiosensitivity, compromised tumor cell survival and enhanced STING-TBK1 signaling by preventing ENPP1-mediated cGAMP hydrolysis. This resulted in robust innate and long-lasting adaptive antitumor immune memory responses, leading to significant tumor regression. Remarkably, combined treatment post-IR reduced spontaneous metastasis and local recurrence, and induced abscopal effects that impacted distant tumor spread in orthotopic tumor models. Thus, these findings position ENPP1 as a critical link between genome integrity and immunosuppression, offering promising translational opportunities for treating local or distant dissemination in TNBC.

Fig. 1

CTC acquire an ENPP1⁺ signature displaying genome integrity and stemness traits. a Schematic of the isolation of engrafting “CTC-in” at the RTB after the intracardiac inoculation in previously orthotopically tumor resected mice (generated by BioRender). b Selected hierarchical functional GO categories based on their biological relevance related to the observed phenotype, obtained by transcriptomic analysis of gene signatures from two different independently isolated CTC-in cells derived from each ANV5 (CTC_A_B5coh) and 4T1 (CTC_T_B5coh) cell lines, compared to their respective parental cells using RNA-seq. Number of coherent genes with B > 5 appears in parenthesis for each cell line. c Hierarchical cluster of transcriptomic upregulated genes related to “Response to radiation” (Left panel) and Stemness (Right panel) GO categories obtained in two different independently isolated CTC-in cells derived from each ANV5 and 4T1 parental cell lines. d Validation by RT-qPCR of commonly regulated genes expressed in different CTC-in derivatives from ANV5 (700, 803, and 054) and 4T1 (1589, 1592, 1593 and 1600) isolated at the RTB compared to their respective parental ANV5 and 4T1 cell lines

Fig. 2

CTC-in cells display a radioresistant phenotype mediated by ENPP1. a Clonogenic assays with independently isolated CTC-in derived from ANV5 (700 and 803) and 4T1 (1589 and 1592) as well as CTC-in with silenced levels of ENPP1 using shRNAs targeting ENPP1, ERN1, TIMELESS, and STAT5a. The right inset depicts ENPP1 protein levels assessed by immunoblotting. An extra sum-of-squares F test was used for comparison. D₅₀ and RER (Radiation enhancement ratio at 2 Gy) values are included. Ns Not significant, ** p < 0.01, ***p < 0.0001. b Left panels: Clonogenic assay performed in ENPP1-overexpressing cells (Top, OE-ANV5 and bottom, OE-4T1) alone or incubated with ENPP1i (5 µM for OE-ANV5 and 10 µM for OE-4T1) at the indicated doses of IR. An extra sum-of-squares F test was used for comparison. ***p < 0.0001. D₅₀ and RER values are included. Right panels: Survival assay conducted with OE cells in similar conditions. One-way ANOVA was used for comparison of ENPP1i treatment with OE cells. c Survival assays using CTC-in derivatives (Left) or CTC-in cells with silenced levels of ENPP1 (Right). d Clonogenic assay of CTC-in with silenced levels of cGAS. Right inset depicts cGAS protein levels. Extra sum-of-squares F test was used for comparison, ***p < 0.0001. D₅₀ and RER values are included. e Immunoblot for PARP1, cleaved-PARP1 (c-PARP1), ATM, phosphorylated H2AX (γH2AX), and ENPP1 in CTC-in cells derived from 4T1 and ANV5 with silenced ENPP1 levels (shENPP1) and shControl cells (transduced with shRNA Control) assessed at the indicated time points post-IR (2 Gy). Dashed lines segregate different treatments within individual immunoblots to enhance visualization. f Immunoblot evaluation for the indicated proteins in OE-ANV5 and OE-4T1 cell lysates after treatment with IR (2 Gy) or in combination with ENPP1i (5 µM). Bands of interest from representative immunoblots from three independent experiments are shown. g Left panel: Outline of the Traffic Light Reporter (TLR) assay (Modified from). Three plasmids are co-transduced (in gray): the TLR, the I-SceI encoding nuclease (I-SceI-T2A-IFP) and the Donor template (Donor-T2A-BFP). Right panels: Quantification of TLR readout after applying a nuclease titration gating analysis in cells co-transduced with I-SceI and Donor-T2A-BFP. RAD51i was used as a positive control for decreased HR. ENPP1i significantly reduced the percentage of normalized homologous recombination (HR) events in the OE cells and in human MDA-MB-231 cells, which endogenously express high levels of ENPP1. Average of 3 independent experiments. Mean ± SEM are represented. One-way ANOVA was used for comparison of HR events in treated cells against Control (TLR/Donor/SceI)

Fig. 3

ENPP1 blockade synergizes with DDR inhibitors and boosts DNA damage post-IR. a Left panels: Heatmaps displaying the percentage of cell viability after 5 days of treatment with modulators of DNA integrity and ENPP1i from a targeted pharmacological screen. The x-axis features 11 kinase inhibitors that target components of the DNA damage response. Right panel: Synergy score. ATMi in combination with ENPP1i showed the highest synergistic effect (Combination Index close to 0). b Viability assay was conducted 5 days after incubation with ATMi or PARPi post-IR at increasing concentrations of ENPP1i showing a decrease in the IC₅₀. c Left panels: Quantification of the Comet assay. Right panels: Representative images of the Comet assay showing the tail moment of the indicated cells treated with IR (2 Gy), ATMi (5 µM) and ENPP1i (5 µM). Brackets point to the tail length. Kruskal-Wallis was used for comparison and Dunn’s post-hoc multiple comparisons test against the Control group. ***P < 0.001. Scale bar = 50 µm. d Left panels: Quantitative assessment of the number of positively labeled cells with anti-γH2AX antibody performed by an in-house developed macro based on ImageJ®. Cells were incubated with ENPP1i (5 µM) and ATMi (5 µM) for 24 h. n > 100 cells were examined over three biologically independent experiments. Median and interquartile range are represented. One-way ANOVA was used for comparisons, and Dunnett´s multiple comparisons test against the Control group. Right panels: Representative immunofluorescence images of nuclear γH2AX (red) and nuclei (blue) in cells subjected to the indicated treatments. Scale bar = 10 µm. e Immunoblot analysis of protein expression levels of γHA2X, PARP1, cleaved PARP1 (c-PARP), ENPP1, RAD51, GAPDH, and Tubulin from cell lysates extracted from a time course after treatment with IR (2 Gy), IR/ENPP1i (5 µM), IR/ATMi (5 µM), or the triple combination in ANV5-OE and OE-4T1 cells. Normalization in each immunoblot was performed relative to the control cells (C) at time 0. Cross-comparison between immunoblots should take into account variation among control samples on each membrane. Dashed lines segregate different treatments within individual immunoblots to enhance visualization

Fig. 4

ENPP1i/DDRi post-IR eradicates local control and impacts disseminated disease. a Left panel: Tumor volume kinetics after orthotopic implantation of OE-ANV5 cells treated with FD (6.2 Gy × 4) alone or in combination with ENPP1i (6 mg/kg daily, BID), ATMi (5 mg/kg daily), or the triple combination (n = 8 mice/group). Right panel: Waterfall plot at the day of sacrifice. Kruskal-Wallis test was applied. Mean ± SEM are represented. **P < 0.01; *** P < 0.0001. b Tumor volume kinetics of orthotopic tumors in Control and dual-treated post-IR treated animals (n = 5 per group) which did not develop tumors were rechallenged by orthotopically implanting OE-ANV5 cells 2 months after treatment interruption. One-way ANOVA was performed. ****P < 0.00001. c Left panel: Similar experiment as in a using 4T1 cells. Treatments included ENPP1i (12 mg/Kg BID) and ATMi (5 mg/Kg daily). Right panel: Waterfall plot at the day of sacrifice. Kruskal-Wallis test was applied. Mean ± SEM are represented. ***P < 0.0001. d Top panels: Quantification of the metastatic surface (left) and the number of pulmonary nodules in histological section of mice treated performed by an in-house developed macro based on ImageJ®. Median and inter-quartile range are represented. Mann-Whitney U test was used for comparison. *P < 0.05; **P < 0.01; ***P < 0.0001. Bottom panels: Representative H/E images of lung lobules in Control and treated-mice with the triple combination. Scale bar = 5 mm. e Left panel: Quantification of Caspase-3 immunostaining in tumor sections of treated animals (n = 5/ group) for 4 days. Right panels: Representative images. Scale bar = 50 µm. f Quantification of hematological parameters in blood samples extracted from naïve animals compared to Control and triple-treated animals for 2 weeks of the indicated cell subpopulations. One-way ANOVA was used for comparison. MPV Mean Platelet Volume. g Quantification of plasma levels of the indicated biochemical markers. ALT Alanine aminotransferase, AST aspartate aminotransferase, HDL High density lipoprotein, BILT Bilirubin, LDL-C LDL-cholesterol. h Left panel: Schematic outline of LF assay. Middle panel: LF-free survival after tumor resection from OE-implanted cells. Mice (n = 15/group) were treated with ATMi (5 mg/Kg/day), ENPP1i (6 mg/Kg BID) or the combination from the day of tumor resection or treated with vehicle (control). Right panel: tumor volume at the day of tumor-resection (Day 0). Log-rank test was used in Kaplan-Meier curves. ** P < 0.001. i Left panel: LF-free survival after surgical resection of tumors derived from OE-ANV5 cells orthotopically implanted as in h. Mice (15 mice/group) were treated with FD (4 × 6.2 Gy) IR alone, on two consecutive days after surgery with an implanted catheter, in combination with ATMi (5 mg/kg daily) or with the dual ENPP1i (6 mg/Kg by oral gavage BID), and ATMi. Right panel: tumor volume at the day of surgery in each group. No differences in tumor margins between groups were detected. Log-rank test was used in Kaplan-Meier curves. **P < 0.01; ***P < 0.0001. j Experimental outline shows the orthotopic tumor cell inoculation in the irradiated mammary gland whereas the contralateral mammary gland was not irradiated. Animals were treated with ENPP1i and ATMi. k Left panels: Orthotopic tumor growth after double simultaneous inoculation of OE-ANV5 cells in opposite inguinal mammary glands in 3 groups of mice (8 mice/group). Treatments include IR only with fractionated dose (FD) performed in one flank, systemic ENPP1i/ATMi treatment, and triple treatment. Tumor volumes were monitored over time in both flanks. Right panels: Tumor volumes of each tumor at the IR flank and the non-IR contralateral flank. Kruskal-Wallis test was used for comparison. *** P < 0.0001

Fig. 5

ENPP1i / DDRi post-IR boosts anti-tumor immune responses. a Immunoblot analysis of protein expression levels of STING, TBK1, phospho-TBK1 and Vinculin from cells lysates of OE-ANV5 and OE-4T1 cells treated with IR (2 Gy), ENPP1i (5 µM) and ATMi (5 µM) extracted for the indicated time post-IR. Data are representative of three independent experiments. b Expression levels assessed by RT-qPCR of the indicated genes in OE-4T1 and OE-ANV5 cells treated as indicated. c Quantification by flow cytometry of tumor-infiltrating immune subpopulations derived from orthotopic tumors at day 4 post-IR from tumor cell inoculation treated with ENPP1 and ATMi daily, showing the reverted immunosuppression. Mean and SD are represented. Kruskal-Wallis test was used for comparison. d Quantification of the multispectral analysis performed for the indicated subpopulations of myeloid M1 and M2 macrophages (top) and (bottom) lymphoid CD4⁺, CD8⁺ T cells and CD4 regulatory T cells. Mean and SD are represented. Representative images from multispectral fluorescence analysis in the indicated groups, showing the spatial resolution of different immune myeloid subpopulations (right), and lymphoid (bottom) infiltrating the tumor core in the group treated with FD or FD plus ENPP1i/ATMi (Triple). Scale bar = 50 µm. Kruskal-Wallis test was used in all panels, except for CD8⁺T cells where Student´s t-test was used for comparison. e Left panel: Tumor volume kinetics after orthotopic inoculation of OE-ANV5 tumor cells. Mice (n = 8 mice/group) were treated when tumors reached a 70 mm³ volume, with vehicle (Control), FD (6.2 Gy × 4) and triple combination of FD/ ENPP1i (6 mg/Kg BID)/ ATMi (5 mg/Kg). In other groups of mice, in addition to the triple combination, mice were treated with anti-CD8, anti-NK1.1, anti-CD4 depleting antibodies (200 µg of antibodies, three times per week) and STING inhibitor (i) (C-176, 5 mg/Kg i.p. daily). Kruskal-Wallis test was used for comparison of depleting antibodies against triple treatment. f Quantification of the indicated immune subpopulations in the blood for each group of mice at the day of sacrifice. *P < 0.05; **P < 0.01; ***P < 0.0001

Fig. 6

Identified ENPP1⁺ gene signature is sustained in human breast cancer tumors. a Uniform Manifold Approximation and Projection (UMAP) of scRNA-seq data of 31 human breast cancer tumors showing the different cell compartments (Left panel) and the ENPP1 expression levels (Right panel). b UMAP of the tumor cell compartment showing the upregulated expression of the indicated genes that overlap with expression of ENPP1 in tumor cells. c Dot plot for the gene features (CD3E, CD8A and ENPP1) in each sample patient from EGAD00001006608 scRNA-seq dataset. Point size reflects the percentage of gene expression (pct.exp) for the corresponding feature in each sample. The color scale indicates the average scaled gene expression (avg.exp.scaled) of each feature. d Schematic representation of the acquisition of ENPP1⁺ phenotype leading to radioresistance. Top left: During primary tumor growth, tumor cells precondition local and distant sites. After resection, remaining preconditioned cells along with wound repair events, create a host niche for the engraftment of residual cells and/or CTC that acquire a gene transcriptomic signature characterized by enhanced genome integrity and stem-like features. In this preconditioned environment, engrafted cells with the acquired ENPP1⁺-transcriptomic signature exhibit a resistant phenotype mediated by changes in dePARylation and phopho-ATM kinetics, thereby favoring HR-mediated DNA damage repair. This mechanism endows cells ´ability to overcome IR-mediated genotoxic stress. Concurrent inhibition of ENPP1 and ATM post-IR impairs DNA repair and boosts immunocompetency by eliciting STING activation in both tumor and non-tumor cells demonstrating a therapeutic susceptibility (Figure generated by BioRender)