Mitochondrial DNA in the tumour microenvironment activates neutrophils and is associated with worse outcomes in patients with advanced epithelial ovarian cancer

Abstract

Background: Advanced cancer causes necrosis and releases damage-associated molecular patterns (DAMPs). Mitochondrial DAMPs activate neutrophils, including generation of neutrophil extracellular traps (NETs), which are injurious, thrombogenic, and implicated in metastasis. We hypothesised that extracellular mitochondrial DNA (mtDNA) in ascites from patients with epithelial ovarian cancer (EOC) would correlate with worse outcomes.

Methods: Banked ascites supernatants from patients with newly diagnosed advanced EOC were analysed for mtDNA, neutrophil elastase, and activation of healthy donor neutrophils and platelets. TCGA was mined for expression of SELP and ELANE.

Results: The highest quartile of ascites mtDNA correlated with reduced progression-free survival (PFS) and a higher likelihood of disease progression within 12-months following primary surgery (n = 68, log-rank, p = 0.0178). NETs were detected in resected tumours. Ascites supernatants chemoattracted neutrophils, induced NETs, and activated platelets. Ascites exposure rendered neutrophils suppressive, based on abrogation of ex vivo stimulated T cell proliferation. Increased SELP mRNA expression correlated with worse overall survival (n = 302, Cox model, p = 0.02).

Conclusion: In this single-centre retrospective analysis, ascites mtDNA correlated with worse PFS in advanced EOC. Mitochondrial and other DAMPs in ascites may activate neutrophil and platelet responses that facilitate metastasis and obstruct anti-tumour immunity. These pathways are potential prognostic markers and therapeutic targets.

Fig. 1

High ascites mtDNA levels are associated with reduced progression-free survival following primary surgery. a–c High ascites levels of mtDNA, but not gDNA, are associated with reduced median PFS. Banked ascites supernatants from patients with newly diagnosed advanced EOC were analysed for mtDNA and gDNA. a Variability exists between ascites mtDNA (ng/μL) levels as measured by qPCR (n = 68). b High mtDNA levels as defined by the upper 75^th quartile (red line) are associated with reduced PFS as compared to the lowest 75^th quartile (black line) of mtDNA (n = 68, log-rank, p = 0.0178). c Ascites gDNA levels are not associated with PFS (n = 62, log-rank, p = 0.839)

Fig. 2

Ovarian cancer ascites is chemoattractive to neutrophils and induces NETs and suppressive neutrophils. a–h Resected tumours were collected from patients with newly diagnosed advanced EOC (n = 5). Tumours were evaluated for the presence of (a–d) intact neutrophils (PMNs, white arrows), identified by a hypersegmented nucleus (DNA, blue; histone H1, red), and cytoplasmic NE (green), and (e–h) NETs, identified by co-localised extracellular DNA, histone H1, and NE (white boxes). PMNs within tumours were rare. NETs were visualised in 4/5 tumours evaluated. i Ascites were collected from patients with newly diagnosed HGSOC (n = 9). >90% of the cells are CD45⁺ leukocytes and on average 15–20% are granulocytic (blue). j–m Ascites supernatants (ASC) were evaluated for pathways of PMN activation. PMNs were isolated from peripheral blood of healthy donors. j Donor PMN chemotaxis was measured in response to mtDAMPs (positive control) and ascites supernatants (n = 2). Sphingosine kinase inhibitor (SKI) was added to both mtDAMPs and ascites supernatants as a negative control to inhibit chemotaxis (**p < 0.01). k Donor PMNs were treated for 1 h with ascites supernatants and evaluated for NET generation (white boxes) by immunofluorescent confocal microscopy. Exposure of neutrophils to 3 of 4 ascites supernatants resulted in NETs. l Donor PMNs were treated for 1 h with media, 100 nM PMA (positive control), ascites supernatants (n = 4), DNase I-pre-treated ascites supernatants, or heat-inactivated ascites supernatants, and evaluated for degranulation by NE ELISA. Ascites-induced neutrophil degranulation, which was partially reversed with heat-inactivation (HI-ASC) or DNase I pre-treatment. m Autologous donor PMNs, CD4⁺, and CD8⁺ T cells were isolated and used in co-culture. Neither PMNs nor ascites supernatants alone suppress anti-CD3/CD28-stimulated T cell proliferation, however, the majority of ascites supernatants from patients with newly diagnosed HGSOC (n = 17/22) induced PMNs to suppress T cell proliferation. ASC-1 is an example of an ascites sample that induced the neutrophil suppressor phenotype, and ASC-2 is an example of an ascites sample that did not induce the neutrophil suppressor phenotype. Data are from ≥3 independent experiments

Fig. 3

Ovarian cancer ascites induces rapid platelet activation and aggregation that is partially abrogated by DNase and protease treatments. a–g Ascites were collected from patients with newly diagnosed advanced EOC and 500 g supernatants (ASC) were used. a Floating aggregates were collected from ascites before centrifugation and analysed by H&E. Abundant neutrophils (black box) and a sparse number of tumour cells (black arrow) embedded in fibrin deposits (pink filaments) were identified. b, c CD41⁺ PMP were measured in ascites supernatants by modified flow cytometry (n = 2). Variability exists between different patients (B, left) 16.58% and (C, left) 1.16% CD41a⁺. In parallel, samples were 0.1 µm-filtered to remove microparticles as a specificity control (b, c, right panels). d–g. Platelets were isolated from peripheral blood of healthy donors and murine cardiac puncture. d Donor platelets were exposed to Tyrode’s buffer with 1 mM CaCl₂ (negative control) or ascites supernatants (n = 7) in the presence of 1 mM CaCl₂ for 30 minutes prior to staining for flow cytometry. Representative density plots show increased number of P-selectin⁺ platelets (P11 gate) after exposure to ascites supernatants, which is quantified to the right (**, p < 0.01). E) Naïve murine platelets were exposed to the same ascites supernatants (n = 7), resulting in increased P-selectin expression and loss of CD42d from the surface of platelets within 15 minutes. The proportion of platelets in the upper left (UL; P-selectin^-CD42d⁺; resting) and lower right (LR; P-selectin⁺CD42d⁻; activated) quadrants of the density plots are quantified to the right (***p < 0.001). f, g) After treatment with ascites supernatants with or without DNase I (0.05% w/v) and/or protease inhibitors (1:100), the proportions of F) P-selectin⁺ donor platelets and (g) CD42d⁺ murine platelets are quantified. Data are from ≥ 3 independent experiments. h High P-selectin mRNA expression in tumour is associated with reduced OS following primary surgery. TCGA provisional RNA Seq V2 RSEM data from cBioportal for ovarian serous cystadenocarcinoma (n = 302) were mined for expression of SELP. Quantiles of SELP expression were plotted against restricted mean survival (RMS) and SELP expression was associated with worse OS (Cox model, p = 0.02, HR = 1.14, 95% CI: 1.03–1.28)

Fig. 4

Model of DAMPs and neutrophil-platelet responses in the ascites of patients with advanced EOC. (1) A hallmark of advanced cancer is cellular necrosis, which releases DAMPs, and minor numbers of tumour cells into the ascites. (2) mtDNA, and likely other DAMPs, recruit and activate neutrophils, induce NETosis, and (3) activate platelets. (4) Platelet activation and aggregation with NETs and fibrin filaments trapfree floating tumour cells and enhance seeding to the serosa and local dissemination within the peritoneal cavity. In addition, neutrophils acquire a T cell suppressor phenotype after ascites exposure, an effect that is predicted to impair T cell immunity and obstruct immunotherapy