Page 229 - 2020_01-Haematologica-web

Neutrophils and cancer-associated thrombosis
ciated thrombosis in ambulatory cancer patients.11-13 In addition, some patients have increased circulating levels of hematopoietic cytokines, such as granulocyte-colony stimulating factor (G-CSF).14 The coagulation cascade is activated by pathogens as part of the innate immune sys- tem to limit dissemination of infection.15 Recently, the term “immunothrombosis” was introduced to describe the contribution of immune cells to thrombus.16 Activated monocytes can trigger thrombosis by expressing tissue factor.17 Activated neutrophils release proteases, such as neutrophil elastase (NE), which enhance thrombosis by degrading the anticoagulant protein tissue factor pathway inhibitor.18 In addition, neutrophils release neutrophil extracellular traps (NET). NET are composed of extracel- lular chromatin components and neutrophil granule pro- teins that enhance thrombosis by capturing platelets and procoagulant extracellular vesicles.19-22 NET are present in both arterial and venous thrombi.19,23,24 NET can also obstruct smaller blood vessels in a coagulation-indepen- dent manner.25 Interestingly, two studies showed that neutrophils contribute to thrombosis in the mouse inferi- or vena cava (IVC) stenosis model, although this was not observed in a third study.20,26,27 In contrast, neutrophil depletion did not affect thrombosis in the IVC stasis model.28
There is a wide range of agonists that can induce NET formation.29 In neutrophils histone citrullination by pep- tidylarginine deiminases (PAD), including PAD4, is consid- ered a driver of chromatin decondensation and subsequent NET formation.30 PAD4 is also expressed by the human breast cancer cell line MCF7.31 Citrullinated histones, such as citrullinated histone H3 (H3Cit), are therefore widely used as a biomarker of NET formation. In mice, it has been proposed that PAD4 is required for NET formation.32 Indeed, PAD4-/- mice have smaller thrombi in the IVC stenosis model.33 However, a recent study found that inhi- bition of PAD did not affect human neutrophil NET forma- tion induced by a variety of pathogens,29 suggesting that certain forms of NET formation can occur without PAD. Interestingly, a recent study found an association between plasma levels of H3Cit and VTE in patients with pancreatic and lung cancer but not in those with other types of cancer, such as breast cancer.34 In another study plasma levels of nucleosomes and cell-free DNA (cfDNA) were higher in cancer patients than in healthy controls, but these are not NET-specific biomarkers.35
Neutrophilia was observed in mice bearing murine breast 4T1 tumors and human pancreatic BxPc-3 tumors.10,36-38 In addition, mice bearing 4T1 breast tumors had increased levels of circulating markers of neutrophil activation and NET, such as H3Cit and myeloperoxi- dase.37,38 Furthermore, tumor-bearing mice had more rapid thrombotic occlusion in a jugular vein Rose Bengal/laser- induced injury model.38 Interestingly, administration of DNase I to degrade cfDNA and NET did not affect throm- botic occlusion in control mice but provided protection from the enhanced venous thrombosis observed in tumor-bearing mice.38 These studies suggest that neu- trophils and NET contribute to venous thrombosis in a murine breast cancer model.
In the light of recent clinical data suggesting a role of NET in VTE in patients with pancreatic cancer,34 we investigated the contribution of neutrophils and NET to venous thrombosis in mice bearing human pancreatic BxPc-3 tumors.
Methods
Cells and the mouse tumor model
We used a human pancreatic cancer cell line BxPc-3 expressing the firefly luciferase reporter.10 BxPc-3 tumors were grown in the pancreas of Crl:NU-Foxn1nu male mice (nude mice) and monitored by measuring luciferase expression.10 We used mice with tumors weighing from 1.5 to 3.9 grams. All animal studies were approved by the University of North Carolina at Chapel Hill Animal Care and Use Committee, and complied with National Institutes of Health guidelines.
Measurement of blood cells
A Hemavet HV950FS (Drew Scientific, Miami Lakes, FL, USA) was used to count neutrophils.
Preparation of plasma and measurement of plasma biomarkers
Blood was collected39 and plasma was prepared by centrifuging the blood at 4500 x g for 15 min. CfDNA was quantified as described elsewhere.40 Mouse G-CSF and NE were measured using commercially available enzyme-linked immunosorbent assays (R&D systems, Minneapolis, MN, USA). H3Cit was meas- ured using an in-house enzyme-linked immunosorbent assay.41
Thrombosis model
We used the IVC stasis model of thrombosis.10
Western blot analysis of thrombus samples
Processing of thrombi and detection of primary antibodies is described in the Online Supplementary Methods. Membranes were probed with 2 mg/mL anti-Ly6G (BioXCell, West Lebanon, NH, USA), a 1,000-fold dilution of anti-β-actin (Abgent, San Diego, CA, USA), a 2,000-fold dilution of anti-PAD4 (Abcam), 1 mg/mL anti-H3Cit (Abcam) or 0.5 mg/mL anti-histone H3 (Abcam) pri- mary antibody.
Immunofluorescence
Analysis of thrombi by immunofluorescence is described in the Online Supplementary Methods. Areas of different fluorescent sig- nals were quantified using Image J software.
Scanning electron microscopy
Analysis of thrombi by scanning electron microscopy is described in the Online Supplementary Methods.
Neutrophil depletion
Neutrophils were depleted by intravenous administration of 5 mg/kg anti-Ly6G antibody (BioXCell) 24 h and 1 h before IVC sta- sis. A rat IgG (Sigma-Aldrich) was used as a control.
DNase I treatment
DNase I (50 U/mouse (Genentech, South San Francisco, CA, USA) or phosphate-buffered saline was intravenously adminis- tered to mice 1 h before and 24 h after IVC stasis.
Statistical analysis
Data are shown as mean ± standard error of the means for nor- mally distributed data or median ± interquartile range for non-nor- mally distributed data. The Shapiro-Wilk test was used to deter- mine normality. For the majority of the studies two-group com- parisons, the unpaired two-tailed Student t-test or the Mann- Whitney U-test was used depending on the data distribution. For the ultrasound data, two-way analysis of variance with the Sidak multiple comparison test was used. These statistical analyses were
haematologica | 2020; 105(1)
219