Neutrophil "plucking" on megakaryocytes drives platelet production and boosts cardiovascular disease

Abstract

Intravascular neutrophils and platelets collaborate in maintaining host integrity, but their interaction can also trigger thrombotic complications. We report here that cooperation between neutrophil and platelet lineages extends to the earliest stages of platelet formation by megakaryocytes in the bone marrow. Using intravital microscopy, we show that neutrophils "plucked" intravascular megakaryocyte extensions, termed proplatelets, to control platelet production. Following CXCR4-CXCL12-dependent migration towards perisinusoidal megakaryocytes, plucking neutrophils actively pulled on proplatelets and triggered myosin light chain and extracellular-signal-regulated kinase activation through reactive oxygen species. By these mechanisms, neutrophils accelerate proplatelet growth and facilitate continuous release of platelets in steady state. Following myocardial infarction, plucking neutrophils drove excessive release of young, reticulated platelets and boosted the risk of recurrent ischemia. Ablation of neutrophil plucking normalized thrombopoiesis and reduced recurrent thrombosis after myocardial infarction and thrombus burden in venous thrombosis. We establish neutrophil plucking as a target to reduce thromboischemic events.

Keywords: bone marrow; mechanotransduction; neutrophils; thromboinflammation; thrombopoiesis; thrombosis.

Figure 1

Neutrophil plucking on PPL-forming MKs (A) Two-photon imaging of thrombopoiesis inside BM in dual reporter mice (Pf4-cre(+)/ Confetti^−/−/Lyz2-eGFP). Representative images from 4 independent experiments of MK (multicolor here shown in red) – neutrophil (green) interactions within the BM and a 3D image reconstruction of the magnified area are shown. Scale bars represent 40 μm or 10 μm (magnified area), respectively. (B) MK characterization by sphericity index. Representative images from 4 independent experiments of low and high sphericity MKs are shown (bar represents 20 μm). Frequency distributions of MKs’ sphericity indices within the BM are shown (n = 4 animals per group, 29 MKs were observed and analyzed over the time frame of 1 h). (C and D) In vivo video analyses quantifying MK-neutrophil interactions and interaction times by sphericity index. (C) Frequency distribution of MK-neutrophil (i.e. individual cells) interactions and (D) interaction times over 1 h are shown (n = 4 animals per group, 7 videos per group). (E) Scheme of MK-neutrophil interactions during thrombopoiesis. (F) Analysis of neutrophil interaction times at the PPL/budding region or Non-PPL/budding region of MKs, symbols indicate individual neutrophils (analysis from 4 animals). (G) Image series and 4D reconstruction of neutrophil plucking on PPL-forming MK around the PPL budding site is shown. White arrows indicate plucking neutrophil. Scale bars represent 40 μm (overview image) or 5 μm (magnified area), respectively, timeline is indicated. (H) Brightfield microscopy of MK-neutrophil co-culture is shown. Scale bar represents 20 μm. (I) Neutrophil plucking during PPL elongation is visualized by time lapse imaging. Arrows indicate plucking neutrophil. Scale bar represents 5 μm. Bars represent mean ± SEM; symbols indicate individual animals; p values are indicated, ^∗<0.05, ^∗∗<0.01, ^∗∗∗<0.001, ^∗∗∗∗<0.0001, n.s. not significant. P values were determined using unpaired Student’s t test (F), one-way (C, D) ANOVA multigroup test. Please see also Figure S1.

Figure 2

Neutrophil plucking accelerates PPL growth and release during thrombopoiesis (A–E) Analysis of MK-neutrophil interaction in Gr-1-treated neutropenic and control-antibody-treated dual reporter mice (Pf4-cre(+)/Confetti^−/−/ Lyz2-eGFP) by two-photon microscopy. (A) Scheme of imaging setup in neutropenic mice. (B) Localization of the interstitial, perisinusoidal, and intravascular BM compartment to further characterize MK-neutrophil interactions. (C) Image sequence showing PPL formation under neutropenic conditions. White arrows indicate PPL-plucking neutrophil around the budding site. Asterisks indicate PPLs, dashed lines show vasculature, yellow box indicates magnified area. Scale bar represents 5 μm (magnified area) and 20 μm, timeline (min) is indicated. (D) Quantification of interaction frequencies within BM compartments (control: n = 4 animals, ND: n = 3 animals, symbols indicate single MKs). (E) Analysis of PPL kinetics. Relative PPL length of individual PPLs (control: 3 PPLs; ND: n = 2PPLs), PPL growth speed, and release time (control: n = 4 animals, ND: n = 3 animals) are shown. (F–H) Neutrophil rescue by adoptive transfer in neutropenic Mrp8-cre(+)/iDTR mice. (F) Treatment scheme is shown. (G) Representative image series of 3 independent experiments and 3D reconstruction of PPL release in mice before and after neutrophil transfer, scale bars represent 15 μm. Analysis of PPL kinetics. Relative length of individual PPLs is shown before and after neutrophil transfer (before: n = 5 PPLs; after: n = 3 PPLs). (I) Summarizing scheme of adoptive transfer experiments is shown. (H) Quantification of PPL growth speed before and after neutrophil transfer (n = 3 animals per group). Bars represent mean ± SEM, symbols indicate individual animals or MKs; p values are indicated, ^∗<0.05, ^∗∗<0.01, ^∗∗∗<0.001, n.s. not significant. P values were determined using unpaired (E) or paired (H) Student’s t test and two-way (D) ANOVA multigroup test. Please see also Figures S1 and S2.

Figure 3

Neutrophil-expressed CXCR4 regulates neutrophil plucking and platelet production (A and B) Quantification of platelet and reticulated platelet counts in two murine models of neutropenia. (A) Platelet and reticulated platelet counts in a model of Gr-1-antibody-induced neutropenia are shown (n = 7 animals per group). (B) Platelet and reticulated platelet counts in a model of diphtheria-toxin-induced neutropenia in Mrp8-Cre/Rosa26iDTR mice (n = 4 animals per group) are shown. (C–E) In vitro thrombopoiesis in a co-culture of primary-fetal-liver-cell-derived MKs and BM-derived neutrophils. (C) Quantification of platelet particles (PPs) within supernatants by flow cytometry 6 h after addition of neutrophils, sphingosine-1-phosphate (10 nM), neutrophils separated by trans-well, pan CD3⁺ T cells, CD19⁺ B cells, monocytes, or neutrophil supernatants following PMA stimulation (no neutrophil: n = 6, neutrophil: n = 6, S1P: n = 5, trans-well: n = 3, CD3⁺ T cells: n = 5, CD19⁺ B cells: n = 4, monocytes: n = 4, neutrophil supernatant: n = 3). (D) Analysis of PP production within co-culture supernatants after neutrophil treatment with actin polymerization inhibitor (cytochalasin D)- or myosin inhibitor (blebbistatin)-treated neutrophils. PP release was determined by flow cytometry 6 h after co-culturing (n = 3). (E) Quantification of PPs within supernatants by flow cytometry 6 h after addition of indicated receptor-deficient neutrophils or inhibitor (no neutrophil: n = 5; neutrophil: n = 5; Selplg: n = 5; Itgb2^−/−: n = 5; Cxcr4^−/−: n = 3; AMD3100 (0.5 μg/ml): n = 3; Itgb1^−/−: n = 3; Kindlin3^−/−: n = 3). (F) Adoptive transfer experiments of Cxcr4^fl/fl-expressing or Cxcr4^Δ/Δ neutrophils in C57BL/6 mice. Representative 3D reconstruction pictures from 3 independent experiments of whole-mount-stained bones are shown (right, scale bar represents 20 μm). (G) Minimal MK-neutrophil distance between MKs and neutrophils (n = 3 animals per group). (H–J) In vivo analysis of thrombopoesis in Mrp8-cre/Cxcr4^fl/fl mice. (H) Time lapse analysis thrombopoiesis in vivo. White arrows indicate neutrophil plucking on PPL budding sites. Asterisks indicate PPLs. Timeline (min) is indicated and dashed lines show vasculature, yellow box indicates magnified area. Scale bar represents 5 μm (magnified area) or 20 μm. (I) In vivo imaging analysis in dual-antibody-labeled Mrp8-cre/cxcr4 mice. Analysis of interaction frequencies within different compartments (Mrp8-cre[−]: n = 4 animals; Mrp8-cre[+]: n = 3 animals), (J) Analysis of PPL kinetics. Relative length of individual PPLs (Mrp8-cre[−]: n = 3 PPLs; Mrp8-cre[+]: n = 1 PPL), PPL growth speed, and release time (Mrp8-cre[−]: n = 3 animals; Mrp8-cre[+] n = 3 animals) are shown. (K and L) Characterization of Mrp8-cre/Cxcr4 mice. Quantification of (K) platelet and reticulated platelet counts (Mrp8-cre[−]: n = 5 animals; Mrp8-cre[+] n = 4 animals) as well as (L) platelet life span (n = 3 animals per group) are shown. Bars represent mean ± SEM; symbols indicate individual animals or MKs; p values are indicated, ^∗<0.05, ^∗∗<0.01, ^∗∗∗<0.001, ^∗∗∗∗<0.0001, n.s. not significant. P values were determined with unpaired Student’s t test (A, B, J (right), and K), or one-way (C, D, and E) or two-way (G, I, and L) ANOVA multigroup test. Please see also Figure S3.

Figure 4

Neutrophil-derived reactive oxygen species drive thrombopoiesis (A) PP release from in vitro MK cultures following co-culture with neutrophils in the presence or absence of NADPH inhibitor apocynin, superoxide dismutase (SOD) (50 U/ml), and catalase (50 μg/ml; n = 3), Cyba^mt/mt or Cybb^−/− neutrophils (n = 4 per group). (B) Visualization of ROS in co-cultured MKs. ROS (DCFDA, green) mean fluorescent intensity in MKs with or without neutrophil interactions (Ly6G positive) was quantified. Symbols indicate individual MKs. Representative images of 3 independent experiments are shown, scale bar represents 10 μm. (C) Representative immunoblots of 3 independent experiments from co-cultured MKs for p-MLC or p-ERK following the indicated treatment are shown. Phospho-signals were quantified and normalized to total MLC and ERK protein (n = 3). (D) MK-neutrophil in vitro co-culture in the presence or absence of neutrophils. Immunofluorescence staining for p-MLC (green), CD41 (red; MKs), Ly6G (white; neutrophil), and DAPI (blue; nuclei). Representative confocal microscopy images of 4 independent experiments are shown (n = 4). (E) Relative (compared to MK w/o neutrophil condition) single-cell p-MLC signal mean fluorescent intensity (MFI) following 3D reconstruction (symbols represent individual MKs). Scheme for p-MLC signal distribution analysis using 3D rendering technique to quantify MFI within the MK body versus PPL. (F) p-MLC distribution within MKs that were cultured in the presence or absence of neutrophils is shown (symbols indicate individual MKs). (G) Scheme of PPL-plucking mechanism by neutrophils during thrombopoiesis. Bars represent mean ± SEM; p values are indicated, ^∗<0.05, ^∗∗<0.01, ^∗∗∗<0.001, ^∗∗∗∗<0.0001, n.s. not significant. P values were determined using unpaired (B and E) or paired (F) Student’s t test, (C) one-sided t test, and one-way (A) ANOVA multigroup test. Please see also Table S1 and Figure S4.

Figure 5

Myocardial infarction triggers thrombopoiesis by neutrophil plucking (A and B) Patients admitted because of acute ST-elevation myocardial infarction (STEMI) with symptom onset <12 h undergoing revascularization by percutaneous coronary intervention or control patients with stable coronary artery disease were recruited. (A) Platelet counts and reticulated platelet fraction (control patients: n = 10; STEMI patients: n = 10, symbols indicate individual patients) and (B) neutrophil-expressed CXCR4 was determined by flow cytometry (control patients: n = 5; STEMI patients: n = 7, symbols indicate individual patients). (C) Adoptive transfer experiments of neutrophils into I/R or sham-treated C57BL/6 mice. Representative 2D pictures from 3 independent experiments of whole-mount-stained bones are shown (scale bar represent 10 μm). Minimal MK-neutrophil distance between MKs and neutrophils was quantified (n = 3 animals per group). (D–G) In vivo multiphoton visualization of MK-neutrophil interactions 48 h after I/R or sham treatment in C57BL/6 mice. (D) Image series of PPL release in I/R and sham-treated mice is shown. Scale bars represent 20 μm (n = 3 animals per group), timeline (min) is indicated. (E) Quantification of MK-neutrophil interaction frequencies within different compartments (n = 3 animals per group). (F) Analysis of PPL growth speed and release time (n = 3 animals per group, each symbol indicates individual PPLs). (G) Frequency of PPL-forming MKs (n = 3 animals per group, each symbol indicates individual ROI area). (H) Platelet and reticulated platelet counts. (I) CXCR4 surface expression on peripheral neutrophils was determined in I/R and sham-treated animals by flow cytometry (n = 4 animals per group), a representative histogram blot of 4 independent experiments is shown. (J) Reactive oxygen species in peripheral neutrophils in I/R and sham-treated animals (n = 4 animals). Bars represent mean ± SEM, symbols indicate individual animals or MKs or ROIs; p values are indicated, ^∗<0.05, ^∗∗∗<0.01, ^∗∗∗∗<0.0001, n.s. not significant. P values were determined by unpaired Student’s t test (F, G, H, I, and J) or one-way (A and B) or two-way (C and E) ANOVA multigroup test. Please see also Table S2 and Figure S5.

Figure 6

Genetic ablation of neutrophil CXCR4 prevents secondary ischemic events and reduces thrombus burden in cardiovascular diseases (A and B) Fe-(III) chloride, mesenteric artery thrombosis in I/R and sham-treated animals. (A) Representative images of 7 (sham) or 8 (MI) independent experiments of platelet recruitment into forming thrombus are shown (scale bar represents 5 μm). (B) Quantification of reticulated platelet fraction within peripheral blood and forming thrombus (sham: n = 7 animals; MI: n = 8 animals; multiple thrombi per mouse were quantified). (C) Fe-(III) chloride induced carotid artery thrombosis in I/R and sham treated animals. Representative images of 8 (sham) or 9 (MI) independent experiments of formed thrombus are shown, dashed lines indicate vessel wall (scale bars represent 200 μm). Occlusive thrombus formation was analyzed (sham: n = 8 animals; MI: n = 9 animals). (D) Fe-(III) chloride-induced carotid artery thrombosis in Mrp8-cre /cxcr4^fl/fl mice 48 h after I/R and sham treatment. A representative image of formed thrombus is shown, dashed lines indicate vessel wall (scale bars represent 200 μm). (E) Occlusive thrombus formation was analyzed by video analysis (Mrp8-cre[−]/cxcr4^fl/fl n = 6 animals; (Mrp8-cre[+]/cxcr4^Δ/Δ n = 5 animals). (F) Vena cava thrombosis was induced in Mrp8-cre/cxcr4^fl/fl mice. Thrombus weights were determined 48 h after induction of venous thrombosis, representative images of 3 independent experiments are shown. (G) Thrombus frequency is shown. Bars represent mean ± SEM, symbols indicate individual animals; p values are indicated, ^∗<0.05, ^∗∗∗<0.001, n.s. not significant. P values were determined with unpaired Student’s t test (B and F) or Chi-square t test (C, E, and G). Please see also Figures S5 and S6.