Elongated neutrophil-derived structures are blood-borne microparticles formed by rolling neutrophils during sepsis

Abstract

Rolling neutrophils form tethers with submicron diameters. Here, we report that these tethers detach, forming elongated neutrophil-derived structures (ENDS) in the vessel lumen. We studied ENDS formation in mice and humans in vitro and in vivo. ENDS do not contain mitochondria, endoplasmic reticulum, or DNA, but are enriched for S100A8, S100A9, and 57 other proteins. Within hours of formation, ENDS round up, and some of them begin to present phosphatidylserine on their surface (detected by annexin-5 binding) and release S100A8-S100A9 complex, a damage-associated molecular pattern protein that is a known biomarker of neutrophilic inflammation. ENDS appear in blood plasma of mice upon induction of septic shock. Compared with healthy donors, ENDS are 10-100-fold elevated in blood plasma of septic patients. Unlike neutrophil-derived extracellular vesicles, most ENDS are negative for the tetraspanins CD9, CD63, and CD81. We conclude that ENDS are a new class of bloodborne submicron particles with a formation mechanism linked to neutrophil rolling on the vessel wall.

Graphical abstract

Figure 1.

Rolling neutrophils form ENDS. (A) Consecutive frames of rolling neutrophils in cremaster venules of mice. Before intravital microscopy (IVM), the mice were intrascrotally injected with TNF-α to stimulate neutrophil interaction with the vasculature. Neutrophils were labeled via i.v. injection of Ly6G-AF647 antibody. Blood flow is from left to right. Asterisks indicate ENDS-forming neutrophils, and arrows point to ENDS. Dotted lines indicate the venule wall. Scale bar, 8 µm. (B) Analysis of neutrophil rolling speed and WSS in the venules shown in A. ENDS-forming neutrophils (indicated with full symbols) had the highest rolling velocity in both examples. Red lines indicate median values. (C) ENDS (indicated by arrows) of various lengths observed in cremaster venules. Mice were pretreated as described above. Scale bars, 6 µm. (D) Human neutrophil labeled with CD16-AF647 forms ENDS while rolling on E-selectin substrate at 40 dyn/cm² WSS. Arrows indicate ENDS. Scale bar, 10 µm. (E) Image segmentation was developed to quantify ENDS formation in records of neutrophils rolling on E-selectin substrate. Cell body is color coded with cyan, tethers that are still attached to the cell are magenta, and ENDS that are not attached to the cell are yellow. Scale bar, 20 µm. (F) Human or mouse neutrophils rolled on P-selectin (blue curve) or E-selectin (red curve) substrate at increasing WSS. Surface area covered by ENDS to quantify ENDS formation. One representative example of three experiments is shown. (G) Rolling mouse neutrophils were allowed to form ENDS in a flow chamber. Only the upstream half was coated with P-selectin; the downstream half was uncoated. Neutrophils and sliding ENDS detached as they reached the noncoated half of the chamber. Upper panel shows a snapshot of rolling neutrophils and sliding ENDS. Lower panel shows kymograph analyses that were done in the positions indicated by the white lines on the upper panel. Scale bar, 50 µm. One representative example of two independent experiments is shown.

Figure S1.

Rolling neutrophils form ENDS. (A–E) Mouse neutrophils were labeled with Ly6G-AF647. Blood flow was from left to right. (A) Consecutive frames show ENDS (arrows) sliding along the venular wall and then detaching. Scale bar, 8 µm. (B) Examples of mouse ENDS with bead-like structures in vivo. Scale bars, 6 µm. (C) Consecutive frames show ENDS flowing in arteriolar blood stream. Scale bar, 8 µm. (D) ENDS detected with confocal microscopy in isolated blood plasma. Blood was collected at the end of the imaging session via a femoral artery cannula. Scale bar, 8 µm. (E) After rolling of unlabeled neutrophils, the flow chamber was stained with CD16-AF647 antibody and was scanned with bright-field and fluorescent imaging. The upper panel shows the record of the entire chamber; on the upper bright-field panel, the rectangles are imaging artifact; the lower panel shows the CD16-AF647 signal to the left behind ENDS. Rectangles indicate the magnified areas. Scale bars represent 1 mm (top image) and 0.5 mm (bottom images). The spherical objects within the magnified areas (indicated by arrows) are neutrophils that remained in the flow chamber after washing. (F) Length of human ENDS formed during rolling on E-selectin under a WSS ramp.

Figure 2.

Structure and composition of human ENDS. (A) SEM and confocal image of ENDS formed on P-selectin substrate by human neutrophils surface labeled with CD16-AF647 antibody (magenta). Scale bar, 1 µm. (B) ENDS formed on E-selectin labeled with CD16-AF647 antibody were analyzed with STORM. Image shows a section of ENDS. Dots indicate the localization of the labeling antibody detected with STORM. Certainty of localization ≤20 nm. Scale bar, 100 nm. (C) The frequency histogram shows the distribution of CD16-AF647 labeling antibody across the width of ENDS measured as in B. Data were collected from two independent experiments (black and gray curves). The distance between the vertical red lines indicates mean ENDS diameter determined as the full width at half maximum (FWHM) value. Bin size is 5 nm. (D) To test the presence of organelles in ENDS, neutrophils were labeled for their surface (CD16-AF647 antibody, magenta), cytoplasm (CellTracker Orange CMRA, yellow), ER (ER Tracker Green), mitochondria (Mito Tracker Green), or nucleus/DNA (Hoechst, cyan) and rolled on E-selectin substrate. Only the cytoplasm tracer was detectable in ENDS formed by these neutrophils. Scale bars, 10 µm. Each of the experiments shown on this figure were repeated at least twice.

Figure S2.

Ultrastructure and composition of ENDS. (A) SEM of a manually graded coverslip, CD16-AF647 antibody–labeled human neutrophil, or human ENDS that were centrifuged on the coverslip. Imaged with confocal microscopy and then with SEM. Left scale bar indicates 5 µm, and right scale bar indicates 3 µm. (B) Comparison of ENDS proteome to published proteomes of neutrophil-derived EVs, NETs, neutrophil-derived primary granules (PG), secondary granules (SG), tertiary granules (TG), ficolin granules (FG), and secretory vesicles (SV), and neutrophil plasma membrane (PM). ENDS proteome was also compared with neutrophil lysate (Neut) that was prepared from one of the donor samples that was used to makes ENDS. (C) Gene ontology analysis of ENDS and polymorphonuclear neutrophils proteome. (D) ENDS quantified with confocal microscopy in blood plasma of Ly6G-cre-mTmG mice before and 2 h after i.p. LPS injection. (E) Human ENDS were surface labeled with CD16-AF647 and CD66b-FITC antibodies. Top panel shows overlay of CD16-AF647 (magenta) and CD66b-FITC (green), and bottom panel shows CD66b-FITC alone. CD66b antibody-labeled ENDS, but to a weaker extent compared with CD16 antibody. Scale bar, 10 µm. (F) Human ENDS were surface labeled with CD16-AF647 and antibodies against TS (CD9-PE, CD63-PE, and CD81-PE). Top panel shows overlay of CD16-AF647 (magenta) and labeling for TSs (green). Bottom panel shows TS labeling alone. TSs were present only on some ENDS and in a patchy way. One example of two independent experiments are shown. Scale bar, 10 µm.

Figure 3.

ENDS degrade and release S100A8–S100A9 complex. ENDS were produced from human neutrophils on an E-selectin–coated substrate as shown in Fig. 1. (A) ENDS were surface labeled with CD16-AF647 antibody and imaged over 4 h. During that period, ENDS contracted as indicated by their decreasing size (area; black line) and increasing sphericity (red line). Values represent average ± SEM of three independent experiments. (B) Human ENDS were imaged over 4 h in the presence of fluorescently labeled annexin-5 (Ann5) in the incubation buffer. The red outlines indicate the ENDS contours determined based on surface labeling with CD16-AF647 antibody. Black pixels show the thresholded annexin-5 signal. After 4-h incubation, ∼50% of ENDS were annexin-5 positive. Average ± SEM of three independent experiments. (C) Fresh (0-h-old) and 7-h-old human ENDS were loaded with Fluo4-AM to measure intraluminal free Ca²⁺ before and after exposure to ionomycin (Iono). The red outlines indicate the ENDS contours indicated by CD16-AF647 surface labeling. Black pixels indicate the thresholded Fluo4 signal. Values represent average ± SEM of seven independent experiments. Statistical analysis was done using a paired t test; ****, P < 0.0001. (D) S100A8–S100A9 complex concentrations were measured in supernatants of ENDS produced with human neutrophils in a flow chamber. ENDS were incubated in HBSS buffer for 1, 7, or 24 h. Statistical analysis was performed using a ratio paired t test; **, P = 0.0096. Scale bars, 10 µm.

Figure 4.

Blood plasma contains elevated number of ENDS during sepsis. (A) Three mice each were injected i.p. with LPS (40 mg/kg) or PBS (not treated, nt). Retro-orbital blood was collected at 2 h, and plasma was prepared by centrifuging at 1,000 g for 5 min. Fluorescently labeled anti-Ly6G antibody was added, and samples were scanned with ImageStream. The images show examples of Ly6G antibody–stained ENDS. Scale bar, 10 µm. (B) Confocal images show examples of EGFP-labeled ENDS, detected in the blood plasma of Ly6G-cre-mT/mG mice harvested 2 h after i.p. LPS injection. Scale bar, 5 µm. (C and D) Count (C) and ratio (D) of neutrophil-derived (CD66b⁺ and CD16⁺) micro objects detected with ImageStream in healthy (black) and septic (gray) human plasma. Micro objects were labeled for TSs with PE-labeled CD9, CD63, and CD81 antibodies and for PS with annexin-5-eF450. Septic plasma samples contained significantly more micro objects with each marker combination (average and SEM are shown; two-tailed t test with Welch’s correction, from left to right: *, P = 0.0251; **, P = 0.029; *, P = 0.0184; **, P = 0.039). In septic samples the ratio of TS⁻/PS⁻ objects was increased at the expense of the TS⁺/PS⁻ objects (two-tailed t test: *, P = 0.0123; **, P = 0.003). (E) Aspect ratio (width/length ratio) of TS⁻ and PS⁻ neutrophil-derived micro objects in septic and healthy plasma. Objects with an aspect ratio <0.6 were considered ENDS, and objects with an aspect ratio >0.6 were considered to contain fragmented ENDS. (F) Examples of ENDS and fragmented ENDS detected with ImageStream in septic blood are shown. Each lane represents a different object, and each column represents a channel of a different marker. Scale bar is 10 µm. (G) Confocal images confirm the presence of CD16⁺ elongated structures in septic plasma labeled with CD16-AF647 antibody. Numbers indicate ENDS length (in micrometers). Scale bar, 5 µm. (H) ENDS content of healthy and septic blood plasma are shown. Red lines indicate average values. Statistical analysis: two-tailed T test with Welch’s correction, **, P = 0.0045. (I) ENDS count showed linear correlation with neutrophil counts. Statistical analysis was done using a correlation analysis and two-tailed test; **, P = 0.006. (J) In septic, but not healthy, samples, the plasma S100A8–S100A9 complex concentration correlated with fragmented ENDS count. Statistical analysis was done using a correlation analysis and two-tailed test; *, P = 0.011. n.s., not significant.