Heme-induced neutrophil extracellular traps contribute to the pathogenesis of sickle cell disease

Abstract

Sickle cell disease (SCD) is characterized by recurring episodes of vascular occlusion in which neutrophil activation plays a major role. The disease is associated with chronic hemolysis with elevated cell-free hemoglobin and heme. The ensuing depletion of heme scavenger proteins leads to nonspecific heme uptake and heme-catalyzed generation of reactive oxygen species. Here, we have identified a novel role for heme in the induction of neutrophil extracellular trap (NET) formation in SCD. NETs are decondensed chromatin decorated by granular enzymes and are released by activated neutrophils. In humanized SCD mice, we have detected NETs in the lungs and soluble NET components in plasma. The presence of NETs was associated with hypothermia and death of these mice, which could be prevented and delayed, respectively, by dismantling NETs with DNase I treatment. We have identified heme as the plasma factor that stimulates neutrophils to release NETs in vitro and in vivo. Increasing or decreasing plasma heme concentrations can induce or prevent, respectively, in vivo NET formation, indicating that heme plays a crucial role in stimulating NET release in SCD. Our results thus suggest that NETs significantly contribute to SCD pathogenesis and can serve as a therapeutic target for treating SCD.

Figure 1

NETs are present within pulmonary vessels of TNF-α–stimulated SCD mice. (A) Representative images of NETs (DNA/red/Sytox orange) detected in the lungs of TNF-α–stimulated SCD mice (i) and the lack of NETs in TNF-α–stimulated SA mice (ii). CD31 (blue/Alexa Fluor 647) labeled the endothelial cells. Scale bar, 10 μm. For imaging, different areas along the dissection were scanned down to an approximate 70 to 80 μm depth using an Axio Examiner.D1 microscope (Zeiss) equipped with a Yokogawa CSU-X1 confocal scan head with a 4-stack laser system (405-nm, 488-nm, 561-nm, and 642-nm wavelengths). Images were obtained using a 20× water immersion objective and as 3-dimensional stacks using Slidebook software (Intelligent Imaging Innovations). (B) Representative histological images of NETs in the lungs of TNF-α–stimulated SCD mice (i,ii, arrows indicate NETs) and the lack of NETs in TNF-α–stimulated SA mice (iii,iv). Scale bar, 10 μm. Images were captured using a Zeiss Axio Observer Z1 microscope equipped with a Zeiss Axiocam HRc camera (color) and a 63× oil immersion objective. (C) Representative immunofluorescence images of the lung sections of TNF-α–stimulated SCD mice showing colocalization of extracellular DNA (green/SYTO 13) and NE (red/Alexa Fluor 568, i,ii) or H3Cit (red/Alexa Fluor 568, iii,iv). CD31 (blue/Alexa Fluor 647) labeled the endothelial cells. Scale bar, 10 μm. Images were captured using an Axio Examiner.D1 microscope equipped with a Yokogawa CSU-X1 confocal scan head with a 4-stack laser system (405-nm, 488-nm, 561-nm, and 642-nm wavelengths) and a 20× water immersion objective. Images were obtained using Slidebook software. (D) Quantification of NETs in the lungs of TNF-α–stimulated SA mice (gray bar, n = 3), SCD mice (white bar, n = 3), and PBS-treated SCD mice (green bar, n = 3). Results presented are the average values (mean ± SEM) of independent experiments with sex- and age-matched mice. *P < .05, ****P < .0001. (E) Quantification of NET biomarkers, plasma DNA (left), and plasma nucleosome (right) of TNF-α–treated SA mice (gray circle, n = 5), SCD mice (white circle, n = 5), and PBS-treated SCD mice (green circle, n = 5; Mann-Whitney test, median [IQR], *P < .05, **P < .01). (F) Quantification of plasma MPO activity of TNF-α–treated SA mice (gray circle, n = 4), SCD mice (white circle, n = 6), and PBS-treated SCD mice (green circle, n = 4; Mann-Whitney test, median [IQR], **P < .01).

Figure 2

DNase I treatment significantly reduces NETs in TNF-α–stimulated SCD mice, protects SCD mice from NET-associated hypothermia, and prolongs their survival by reducing acute lung injury. (A) Quantification of NETs in the lungs of TNF-α–stimulated SCD mice pretreated with vehicle (white bar, n = 3) or DNase I (red bar, n = 7). Results presented are the average values of independent experiments with sex- and age-matched mice (mean ± SEM, ***P < .001). (B) Quantification of NET biomarker, plasma DNA, of TNF-α–stimulated SCD mice pretreated with vehicle (white circle, n = 6) or pretreated with DNase I (red circle, n = 10) (Mann-Whitney test, median [IQR]). (C) Reduction in rectal temperature of TNF-α–stimulated SA mice (gray circle, n = 10), PBS-treated SCD mice (green circle, n = 3), and TNF-α–stimulated SCD mice (white circle, n = 10; mean ± SEM, *P < .05, ****P < .0001). (D) Reduction in rectal temperature of TNF-α–stimulated SCD mice, pretreated with vehicle (white circle, n = 6) or DNase I (red circle, n = 11; mean ± SEM, *P < .05). (E) Relationship between decrease in body temperature and number of NETs presented in the lungs of SA and SCD mice (r = 0.79, P < .0001, different color circles represent different treatments as stated in Figure 2C-D). (F) Kaplan-Meier survival curves after TNF-α treatment and surgical trauma in vehicle-treated (black line, n = 5) and DNase I–treated (red line, n = 7) groups (log-rank test, *P < .05). (G) Representative histology of lungs from vehicle-infused (arrows indicate the fibrillar appearance of alveolar walls) and DNase I–infused SCD mice after TNF-α administration and surgical procedure. Scale bar, 10 μm. Images were captured using Zeiss Axio Observer Z1 microscope equipped with Zeiss Axiocam HRc camera (color) and a 40× oil immersion objective. (H) (Top) Total protein (left) and IgM concentration (right) in BAL fluid of vehicle- (white circle, n = 7) and DNase I–treated (red circle, n = 7) SCD mice and vehicle-treated SA mice (gray circle, n = 4) after TNF-α administration and surgical procedure (mean ± SEM, *P < .05, **P < .01). (Bottom) MPO activity per gram of tissue in lung homogenates (left) and in cell-free BAL fluid (right) of vehicle- (white circle, n = 7) and DNase I–treated (red circle, n = 7) SCD mice and vehicle-treated SA mice (gray circle, n = 4) after TNF-α administration and surgical procedure (mean ± SEM, * P < .05, **P < .01).

Figure 3

Plasma from TNF-α–stimulated SCD mice and crisis SCD patients induce NET formation in TNF-α–primed neutrophils. (A) Representative immunofluorescence images of NETs released by TNF-α–primed wild-type BM neutrophils. DNA was stained with SYTO13 (cell-permeable nuclear acid dye, green) and Sytox orange (cell-impermeable nuclear acid dye, red). Primed wild-type neutrophils were stimulated with plasma of either TNF-α–treated SA mice (left) or TNF-α–treated SCD mice (right). Scale bar, 10 μm. (B) Representative immunofluorescence images of DNA (green/SYTO13), H3Cit (red/Alexa Fluor 568, i), or NE (red/Alexa Fluor 568, ii) of NETs generated by TNF-α–primed wild-type neutrophils that were stimulated with plasma of TNF-α–treated SCD mice. Insets show NETs stained with DNA dye (green/SYTO13) and isotype control antibodies for H3Cit and NE. Scale bar, 10 μm. (C) Quantification of NETs released by TNF-α–primed wild-type BM neutrophils stimulated with plasma of TNF-α–treated SA mice (black bar) and TNF-α–treated SCD mice (white bar). Results are average values from 4 independent experiments (mean ± SEM, **P < .01). (D) Quantification of NET biomarkers, plasma DNA (left), and nucleosome (right) of non-SCD individuals (gray circle, n = 5), steady-state SCD patients (yellow circle, n = 7), and crisis SCD patients (red circle, n = 10; Mann-Whitney test, median [IQR], *P < .05). (E) Quantification of plasma MPO activity of non-SCD individuals (gray circle, n = 4), steady-state SCD patients (yellow circle, n = 7), and crisis SCD patients (red circle, n = 10; Mann-Whitney test, median [IQR], *P < .05, **P < .01). (F) Representative immunofluorescence images of DNA (green/SYTO13), H3Cit (red/Alexa Fluor 568, i), or NE (red/Alexa Fluor 568, ii) of NETs generated by TNF-α–primed human neutrophils that were stimulated with plasma of crisis SCD patient. Insets show NETs stained with DNA dye (green/SYTO13) and isotype control antibodies for H3Cit and NE. Scale bar, 10 μm. (G) Quantification of NETs released by TNF-α–primed human neutrophils stimulated with plasma of non-SCD individuals (gray circle, n = 3), steady-state SCD patients (yellow circle, n = 7), and crisis SCD patients (red circle, n = 8; mean ± SEM, *P < .05, **P < .01).

Figure 4

Heme stimulates NET formation in neutrophils in vitro. (A) Representative images of plasma of TNF-α–stimulated SCD mice and SA mice. (B) Quantification of plasma heme concentrations of TNF-α–stimulated SA mice (black circle, n = 8), TNF-α–stimulated SCD mice (white circle, n = 5), and PBS-treated SCD mice (gray circle, n = 6; mean ± SEM, ***P < .001). (C) Quantification of NETs released by TNF-α–primed neutrophils stimulated with 100 nM phorbol myristate acetate (PMA) (white bar), 10 μM heme (blue bar), and 20 μM heme (red bar). The shaded bars represent DMSO treatment of each condition. Results are the average values of 3 independent experiments (mean ± SEM, *P < .05). (D) Representative immunofluorescence images of NETs released by TNF-α–primed neutrophils stimulated with (i) 100 nM PMA, (ii) 20 μM heme, (iii) DMSO, and (iv) 20 μM heme; the slide was then treated with DNase I. (E) Representative immunofluorescence images of DNA (green/SYTO13), H3Cit (red/Alexa Fluor 568, i,ii), or NE (red/Alexa Fluor 568, iii,iv) of NETs generated by TNF-α–primed wild-type neutrophils that were stimulated with 20 μM heme. Scale bar, 10 μm.

Figure 5

Heme-mediated NET formation in TNF-α–primed neutrophils is dependent on ROS and heme-iron and can be blocked by hemopexin complexation. (A) Quantification of NETs released by TNF-α–primed neutrophils treated with DMSO (white bar), antioxidant NAC alone (10 mM, blue bar), heme alone (20 μM, black bar), heme (20 μM), and NAC (10 mM) (gray bar). Results are the average values of 3 independent experiments (mean ± SEM, **P < .01, ***P < .001). (B) Quantification of NETs released by TNF-α–primed neutrophils treated with free heme (black bar) or Hx complex (gray bar). Results are the average values of 3 independent experiments (mean ± SEM, *P < .05). (C) Quantification of NETs released by TNF-α–primed neutrophils treated with free heme (black bar) or HSA complex (gray bar). Results presented are the average values of 3 independent experiments (mean ± SEM). (D) Quantification of NETs released by TNF-α–primed neutrophils treated with free heme (black bar) or HbA (gray bar). Results are the average values of 4 independent experiments (mean ± SEM, **P < .01). (E) Representative immunofluorescence images of NETs released by TNF-α–primed neutrophils stimulated with 20 μM of (i) heme, (ii) zinc (II) protoporphyrin IX (Zn (II) PPIX), (iii) protoporphyrin (PPIX), and (iv) iron (III) mesoporphyrin IX (Fe (III) Meso). (F) Quantification of NETs released by TNF-α–primed neutrophils treated with DMSO (white bar), heme (black bar), Zn (II) PPIX (red bar), PPIX (blue bar), and Fe (III) Meso (gray bar). Results are the average values of 3 independent experiments (mean ± SEM, *P < .05, ***P < .001).

Figure 6

Heme stimulates NET formation in neutrophils in vivo. (A) Quantification of plasma heme concentrations of TNF-α–stimulated SA mice pretreated with vehicle (gray circle, n = 8) or heme (red circle, n = 11) (mean ± SEM, ****P < .0001). (B) Quantification of NETs in the lungs of TNF-α–stimulated SA mice, vehicle pretreated (gray bar, n = 4) or heme pretreated (red bar, n = 6). Results presented are the average values of independent experiments with sex- and age-matched mice (mean ± SEM, **P < .01). (C) Representative histological images of NETs in the lungs of heme pretreated-, TNF-α–stimulated SA mice (arrows indicate NETs). Scale bar, 10 μm. (D) Representative immunofluorescence images of the lung sections of heme pretreated-, TNF-α–stimulated SA mice showing colocalization of extracellular DNA (green/SYTO13) and NE (red/Alexa Fluor 568, i-ii) or H3Cit (red/Alexa Fluor 568, iii-iv). CD31 (blue/Alexa Fluor 647) labeled the endothelial cells. Scale bar, 10 μm. (E) Quantification of NET biomarkers, plasma DNA (left), and plasma nucleosome (right) of TNF-α–challenged SA mice, vehicle pretreated (gray circle, n = 7) or heme pretreated (red circle, n = 6) (Mann-Whitney test, median [IQR], **P < .01). (F) Quantification of plasma MPO activity of TNF-α–challenged SA mice, vehicle pretreated (gray circle, n = 4) or heme pretreated (red circle, n = 6) (Mann-Whitney test, median [IQR], *P < .05). (G) Decrease in rectal temperature of TNF-α–stimulated SA mice, pretreated with vehicle (gray circle, n = 12) or heme (red circle, n = 12) (mean ± SEM, **P < .01).

Figure 7

Hemopexin infusion decreases NETs in TNF-α–stimulated SCD mice and protects SCD mice from NET-associated hypothermia. (A) Quantification of plasma heme concentrations of TNF-α–stimulated SCD mice, vehicle pretreated (white circle, n = 13) or Hx pretreated (gray circle, n = 13) (mean ± SEM, **P < .01). (B) Quantification of NETs in the lungs of TNF-α–stimulated SCD mice, vehicle pretreated (white bar, n = 6) or Hx pretreated (gray bar, n = 10). Results presented are the average values of independent experiments with sex- and age-matched mice (mean ± SEM, **P < .01). (C) Quantification of NET biomarkers, plasma DNA (left) and plasma nucleosome (right), from TNF-α–challenged SCD mice, vehicle pretreated (white circle, n = 14) or Hx pretreated (gray circle, n = 10) (Mann-Whitney test, median [IQR], **P < .01). (D) Quantification of plasma MPO activity of TNF-α–challenged SCD mice, vehicle pretreated (white circle, n = 6) or Hx pretreated (gray circle, n = 6) (Mann-Whitney test, median [IQR], **P < .01). (E) Decrease in rectal temperature of TNF-α–stimulated SCD mice, pretreated with vehicle (white circle, n = 13) or Hx (gray circle, n = 13; mean ± SEM, *P < .05).