Neutrophil extracellular trap‐microparticle complexes trigger neutrophil recruitment via high‐mobility group protein 1 (HMGB1)‐toll‐like receptors(TLR2)/TLR4 signalling

Abstract

Background and Purpose

Recent data suggest that neutrophil extracellular traps (NETs) form aggregates with microparticles (MPs) upon activation of neutrophils although the functional role of NET‐MP complexes remain elusive. The objective of this study was to examine the role of NET‐MP aggregates in leukocyte recruitment in vivo.

Experimental Approach

PMA stimulation of murine bone marrow neutrophils generated NET‐MP complexes and pretreatment with caspase and calpain inhibitors resulted in the formation of NETs depleted of MPs. Leukocyte–endothelium interactions were studied by using intravital microscopy of the mouse cremaster microcirculation.

Key Results

Intrascrotal injection of NET‐MP aggregates dose‐dependently increased leukocyte recruitment. In contrast, leukocyte responses were markedly reduced after administration of NETs depleted of MPs. Neutrophil depletion abolished intravascular and extravascular leukocytes in response to challenge with NET‐MP complexes. Electron microscopy revealed that NET‐associated MPs express HMGB1. Notably, immunoneutralization of HMGB1 markedly decreased NET‐MP complex‐induced neutrophil accumulation. Moreover, inhibition of TLR2 and TLR4 significantly reduced neutrophil recruitment in response to NET‐MP aggregates.

Conclusions and Implications

These data show that NET‐MP complexes are potent inducers of neutrophil recruitment, which is dependent on HMGB1 expressed on MPs and mediated via TLR2 and TLR4. Blocking MP binding to NETs or downstream inhibition of the HMGB1‐TLR2/TLR4 axis might provide useful targets to attenuating NET‐dependent tissue damage in acute inflammation.

Abbreviations

HMGB1

high‐mobility group protein 1

LFA‐1

lymphocyte function antigen‐1

Mac‐1

membrane‐activated complex‐1

MPs

microparticles

NETs

neutrophil extracellular traps

PAD

peptidylarginine deiminase

PSGL‐1

P‐selectin glycoprotein ligand‐1

RAGE

receptor for advanced glycation end products

What is already known

• Neutrophil extracellular traps can form aggregates with microparticles.

What this study adds

• Neutrophil extracellular trap‐microparticle complexes are potent inducers of neutrophil recruitment.

What is the clinical significance

• Targeting formation of neutrophil extracellular trap‐microparticle complexes could be useful to treat inflammatory diseases.

1. INTRODUCTION

Extracellular traps are expelled from activated neutrophils as part of the host defence against invading microbes (Brinkmann et al., 2004). However, excessive formation of neutrophil extracellular traps (NETs) are known to cause tissue damage in both infectious and non‐infectious diseases (Luo et al., 2014; Merza et al., 2015). Although there is no direct evidence showing that NETs can provoke neutrophil recruitment, indirect evidence such as inhibition of NET formation or NET disintegration has been shown to reduce inflammatory cell recruitment in models of inflammation (Luo et al., 2014; Merza et al., 2015). Leukocyte accumulation is a multistep process (rolling–adhesion–emigration) mediated by specific adhesion molecules expressed on endothelial cells and leukocytes (Butcher, 1991). Several studies have documented that leukocyte rolling is mainly supported by the selectin family (P‐, E‐, and L‐selectin) of adhesion molecules, which interact with their glycoprotein counter‐ligands (Carlos & Harlan, 1994; Vestweber & Blanks, 1999). P‐selectin glycoprotein ligand‐1 (PSGL‐1) is one of the best characterized selectin counter‐receptors, which preferentially binds to P‐selectin but can also bind to E‐selectin with low affinity (Yang et al., 1999). Inhibition of P‐selectin or PSGL‐1 effectively decreases leukocyte infiltration in different models of inflammation, such as reperfusion injury (Riaz et al., 2002) and septic lung damage (Asaduzzaman, Rahman, Jeppsson, & Thorlacius, 2009). Some reports suggest that subsets of integrins can support rolling under special conditions (Johnston, Issekutz, & Kubes, 1996). Firm adhesion of leukocytes to endothelial cells is mainly mediated by β₂‐integrins, such as lymphocyte function antigen‐1 (LFA‐1; CD11a/CD18), membrane‐activated complex‐1 (Mac‐1; CD11b/CD18), and p150, 95 (CD11c/CD18). The literature is rather complex and partly contradictory with respect to the role of β₂‐integrins in leukocyte adhesion and the relative importance of specific β₂‐integrins appears to vary depending on the type of inflammatory stimulus and experimental model (Issekutz, Chuluyan, & Lopes, 1995; Riaz et al., 2002; Rutter et al., 1994; Thorlacius et al., 2000). Taken together, the importance of selectins and β₂‐integrins for NET‐induced leukocyte recruitment is not known and deserves further studies.

NETs are decorated with nuclear, cytoplasmatic, and granular proteins, which exert important biological functions (Brinkmann et al., 2004). For example, NET‐associated histones trigger trypsin activation in acinar cells in the pancreas and activation of the coagulation system (Merza et al., 2015; Noubouossie et al., 2017). Moreover, myeloperoxidase (MPO) and cathepsin G exposed on NETs exert antimicrobial actions (Parker, Albrett, Kettle, & Winterbourn, 2012; Urban et al., 2009). Activated neutrophils also shed microparticles (MPs), that is, sphere‐shaped intact vesicles released from cell membranes with a size less than 1 μm (Hess, Sadallah, Hefti, Landmann, & Schifferli, 1999). The surface of MP is coated with molecules resembling the expression on the mother cells. For example, both activated neutrophils and neutrophil‐derived MPs express high levels of Mac‐1 (Wang et al., 2018). Understanding of the role of neutrophil‐derived MPs in inflammation is limited and partly controversial. One study reported that MPs released into circulation by activated neutrophils exert inhibitory functions on neutrophil adhesion and extravasation (Dalli et al., 2008). On the other hand, in response to external stimuli, neutrophil‐derived MPs generate pro‐inflammatory compounds, such as ROS and LTB₄ (Dalli et al., 2013). Recent evidence has shown that MPs form complexes with NETs via histone–phosphatidylserine interactions (Wang et al., 2018). These NET‐MP aggregates are powerful inducers of thrombin generation via the intrinsic pathway of coagulation (Wang et al., 2018). However, it is not known whether MPs associated with NETs contribute to pro‐inflammatory effects of NETs.

Based on these considerations, we wanted to examine the effect of NET‐MP complexes on the leukocyte extravasation process in detail and determine the adhesive mechanisms mediating leukocyte–endothelium interactions and recruitment in response to NET‐MP aggregates.

2. METHODS

2.1. Animals

All animal care and experimental procedures complied with the legislation on the protection of animals and were approved by the Regional Ethical Committee for Animal Experimentation at Lund University, Sweden. All effort was taken to minimize the number of animals used and their suffering. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny, Browne, Cuthill, Emerson, & Altman, 2010; McGrath, Drummond, McLachlan, Kilkenny, & Wainwright, 2010) and with the recommendations made by the British Journal of Pharmacology. Male C57BL/6 mice (Taconic Bioscience, Lille Skensved, Denmark; RRID:IMSR_JAX:000664) 8–9 weeks old (23–27 g) were kept in a pathogen‐free facility on a 12–12 hr light–dark cycle and had free access to food (R3 breeding food for mice, Lantmännen, Kemstad, Sweden) and tap water. Mice were housed for 1 week before use in the experiments. Animals were kept at a maximum of seven mice per cage with environment enrichment, such as a little house and toys. Animals were anaesthetized by intraperitoneal administration of 75‐mg ketamine hydrochloride (Hoffman‐La Roche, Basel, Switzerland) and 25‐mg xylazine (Janssen Pharmaceutica, Beerse, Belgium) per kg. Mice were randomly assigned to in vitro and in vivo experiments. Data collection and evaluation of all experiments were performed without the knowledge of the treatment.

2.2. NET formation

Neutrophils isolated from the bone marrow were incubated (4 × 10⁶ cells·ml⁻¹) with 500‐nM PMA or 500 ng·ml⁻¹ CXCL2 for 3 hr at 37°C in RPMI 1640. In indicated experiments, cells were co‐incubated with caspase (50 μM, Z‐VAD‐FMK) and calpain (25 μM, PD150606) inhibitors, which have been shown to generate NETs without MPs (Wang et al., 2018). Supernatants were discharged, and fresh media were added to isolate NETs. Residual neutrophils and NETs were removed through extensive pipetting. The mixture was centrifuged at 200× g (5 min) to remove cellular components, and NET‐containing supernatants were collected. NET‐containing supernatants were further centrifuged at 19,000× g (15 min) to collect NETs. NETs co‐incubated with PBS or DNase (50 μg·ml⁻¹) for 30 min at 37°C, and then the media with NETs were retrieved for further use. The content of NETs was quantified by using a fluorogenic assay for double‐stranded DNA (Quant‐IT PicoGreen dsDNA kit; Invitrogen, Eugene, USA). S100A9 was analysed in NETs by using a commercially available ELISA kit (Hycult Biotech Inc., Wayne, USA). S100A9 was diluted in a specific diluent provided by the ELISA kit manufacturer to make standard curves according to the manufacturer's instructions. Myeloperoxidase (MPO) levels in NETs were determined spectrophotometrically as the MPO‐catalysed change in absorbance in the redox reaction of H₂O₂ (450 nm, with a reference filter 540 nm). Values were expressed as MPO units.

2.3. Cremaster muscle preparation

Intravital microscopy of the cremaster muscle microcirculation was performed as previously described (Ley et al., 1995; Mansson, Zhang, Jeppsson, Johnell, & Thorlacius, 2000). Briefly, a midline incision of the skin and fascia was made over the ventral aspect of the right scrotum, and the incised tissues were retracted to expose the cremaster muscle sack. Then, the cremaster muscle was separated from the epididymis, and the testis was pushed back to the side of the preparation. The preparation was performed on a transparent pedestal allowing transillumination and microscopic observations of the cremaster muscle microcirculation after a 10‐min equilibration time. Intrascrotal injection of NETs (0.5–4.5 μg) and TNF‐α (0.5 μg, R&D Systems Europe, Ltd., Abingdon, Oxon, UK) diluted in 0.15‐ml PBS was performed 3 hr prior to microscopic observation. In order to delineate the role of the selectins and integrins in leukocyte–endothelium interactions, monoclonal antibodies directed against E‐selectin (2 mg·kg⁻¹, clone 10E9.6, rat IgG, BD Bioscience, San Diego, CA, USA; RRID:AB_2186705), P‐selectin (2 mg·kg⁻¹, clone RB40.34, rat IgG, BD Biosciences; RRID:AB_395023), PSGL‐1 (2 mg·kg⁻¹, clone 4RA10, rat IgG, BD Biosciences; RRID:AB_647340), Mac‐1 (4 mg·kg⁻¹, clone M1/7, rat IgG, BD Biosciences; RRID:AB_394771), and LFA‐1 (4 mg·kg⁻¹, clone M17/4, rat IgG, BD Biosciences; RRID:AB_394793) as well as a control antibody (clone R3‐34, rat IgG₁, BD Biosciences; RRID:AB_479672) were given intravenously immediately prior to intrascrotal administration of NETs. In separate experiments, monoclonal antibodies against HMGB1 (0.4 mg·kg⁻¹, clone DPH 1.1, mouse IgG, HMGBiotech, Milano, Italy), TLR2 (1 mg·kg⁻¹, clone T 2.5, mouse IgG, Biolegend, London, UK; RRID:AB_604142), and TLR4 (2 mg·kg⁻¹, clone MTS510, rat IgG, Biolegend; RRID:AB_2205136) were administered intravenously immediately before intrascrotal administration of NETs. All animals were randomized for treatment.

2.4. Intravital microscopy

Observations of the cremaster microcirculation were made using an Olympus microscope (BX50WI, Olympus Optical Co. GmbH, Hamburg, Germany) equipped with water immersion lenses (40/NA 0.75 and 63/NA 0.90). The microscopic image was recorded in a computer using a charge‐coupled device video camera (FK 6990 Cohu, Pieper GmbH, Berlin, Germany) for subsequent off‐line analysis. Analyses of leukocyte flux and leukocyte–endothelium interactions (rolling and adhesion) were made in venules (inner diameter between 20 and 40 μm) with stable resting blood flow. Rolling leukocyte flux was determined at indicated time points by counting the number of rolling leukocytes per 20 s passing a reference point in the microvessel and expressed as cells·min⁻¹. Leukocyte rolling velocity was determined by calculating the velocity of 10 leukocytes rolling along the endothelial cell lining and is given as μm·s⁻¹. Leukocyte adhesion in venules (stationary for 20 s) was counted in 100‐μm‐long vascular segments and expressed as number of adherent cells·mm⁻². Leukocyte emigration was quantified by counting the number of extravascular leukocytes within an extravascular area of 100 × 70 μm immediately adjacent to the venules. Leukocyte emigration was expressed as number of extravascular cells·mm⁻². Diameters (d) were measured in micrometer perpendicularly to the vessel path. Microvascular haemodynamics were determined after injection of 0.1‐ml 5% FITC–dextran (MW 150,000, Sigma‐Aldrich, Stockholm, Sweden) for contrast enhancement by intravascular staining of plasma. The cremaster muscle microvasculature was visualized by a 100‐W mercury lamp and filter sets for blue (450‐ to 490‐nm excitation and >520‐nm emission wavelength) and green (530‐ to 560‐nm excitation; >580‐nm emission) light epi‐illumination. Flow velocity (v) was analysed by the computer‐assisted image analysis system using the line shift method. Venular wall shear rate was calculated based on the Newtonian definition: wall shear rate = 8 (red blood cell velocity/venular diameter; House & Lipowsky, 1987). All quantitative analysis of microhaemodynamic parameters in the cremaster microcirculation was performed by means of the computer‐assisted image analysis system CapImage (Zeintl, Heidelberg, Germany).

2.5. Neutrophil depletion

An antibody directed against Ly6G (20 mg·kg⁻¹, clone 1A8, rat IgG, BioXcell, West Lebanon, NH, USA; RRID:AB_1107721), which is known to effectively deplete neutrophils of mice (Daley, Thomay, Connolly, Reichner, & Albina, 2008), or a control antibody (20 mg·kg⁻¹, rat IgG) was administered intraperitoneally 24 hr before intrascrotal challenge with NETs.

2.6. Flow cytometry

Blood was harvested from the inferior vena cava and stained with phycoerythrin/Cy7‐conjugated anti‐mouse Ly‐6G (clone 1A8, BD Biosciences; RRID:AB_1727562), allophycocyanin‐conjugated anti‐mouse Ly‐6C (clone AL‐21, BD Biosciences; RRID:AB_1727554), FITC‐conjugated anti‐mouse CD19 (clone MB19‐1, Thermo Fisher Scientific; RRID:AB_464967), and FITC‐conjugated anti‐mouse CD3 (clone 145‐2C11, BD Biosciences; RRID:AB_394595) antibodies. Cells were further fixed, and erythrocytes were lysed for flow cytometry. In separate experiments, NET (100 ng·ml⁻¹) with and without MPs isolated from PMA‐stimulated neutrophils as described above were co‐incubated with freshly isolated bone marrow neutrophils. CXCL2 (100 ng·ml⁻¹) was used as a positive control. After 20‐min incubation, flow cytometry was used to quantify CD11b expression on neutrophils. Flow cytometric analysis was performed according to standard settings on a CytoFLEX (Beckman Coulter, Bromma, Sweden) and analysed with CytExpert 2.0 software (Beckman Coulter). A viable gate was used to exclude dead cells and fragmental components. Neutrophils were defined as Ly‐6G⁺/Ly‐6C⁺ cells.

2.7. NET‐induced activation of endothelial cells

The polyoma‐transformed murine endothelioma cell line eEnd.2 (RRID:CVCL_6274) was cultured in DMEM supplemented with 10% fetal calf serum, l‐glutamine, penicillin, and streptomycin as described previously (Williams et al., 1989). NET with and without MPs isolated from PMA‐stimulated neutrophils as described above were co‐incubated with subconfluent endothelial cells for 0–6 hr. Total RNA was isolated (RNeasy Mini Kit, Qiagen, West Sussex, UK) and treated with RNase‐free DNase (DNase I; Amersham Pharmacia Biotech, Sollentuna, Sweden). RNA concentrations were determined by measuring the absorbance at 260 nm. Each cDNA was synthesized by reverse transcription from 10 μg of total RNA by using the StrataScript First‐Strand Synthesis System and random hexamer primers (Stratagene, AH diagnostics, Stockholm, Sweden). Real‐time PCR was performed using a Brilliant SYBRgreen QPCR master mix and MX 3000P detection system (Stratagene). The primer sequences of ICAM‐1 and β‐actin were as follows: ICAM‐1 (forward) 5′‐AGCACCTCCCCACCTACTTT‐3′, ICAM‐1 (reverse) 5′‐AGCTTGCACGACCCTTCTAA‐3′ and GAPDH (forward) 5′‐CATGTTCGTCATGGGGTGAACCA‐3′, GAPDH (reverse) 5′‐AGTGATGGCATGGACTGTGGTCAT‐3′. Standard PCR curves were generated for each PCR product to establish linearity of the RT‐PCR reaction. PCR amplifications were performed in a total volume of 50 μl, containing 25 μl of SYBRgreen PCR 2× master mix, 2 μl of 0.15 μM each primer, 0.75 μl of reference dye, and one 1 μl of cDNA as a template adjusted up to 50 μl with water. PCR reactions were started with 10 min denaturing temperature of 95°C, followed by a total of 40 cycles (95°C for 30 s and 55°C for 1 min) and 1 min of elongation at 72°C. The relative differences in expression between groups were expressed by using cycling time values. Cycling time values for ICAM‐1 were first normalized with that of β‐actin in the same sample, and then relative differences between groups were expressed as percentage of control.

2.8. Transmission electron microscope and scanning electron microscope

Neutrophils stimulated with PMA as described above were fixed in 2.5% glutaraldehyde in 0.15 mol·L⁻¹ sodium cacodylate, pH 7.4 (cacodylate buffer) for 30 min at room temperature. Specimens were washed with cacodylate buffer and dehydrated with an ascending ethanol series from 50% (vol/vol) to absolute ethanol (10 min per step). Then, specimens were subjected to critical point drying in carbon dioxide, with absolute ethanol as intermediate solvent, mounted on aluminium holders, and finally sputtered with 20‐nm palladium/gold. Specimens were examined in a Jeol/FEI XL 30 FEG scanning electron microscope at the Core Facility for Integrated Microscopy at Panum Institute (University of Copenhagen, Denmark). The location of individual target molecules was analysed at high resolution by ultrathin sectioning and transmission immunoelectron microscopy. Specimens on coverslips were embedded in Epon 812 and sectioned into 50‐nm‐thick ultrathin sections with a diamond knife in an ultramicrotome. For immunohistochemistry, sections were incubated overnight at 4°C with primary antibodies against Mac‐1, citrullinated histone 3 and HMGB1 (Abcam, Cambridge, UK; RRID:AB_306832; RRID:AB_304752; RRID:AB_1566303). Controls without primary antibodies were included. The grids then were incubated with species‐specific, gold‐conjugated secondary antibodies (Electron Microscopy Sciences, Fort Washington, MD, USA). Gold‐labelled annexin V were also used. Finally, the sections were postfixed in 2% glutaraldehyde and post‐stained with 2% uranyl acetate and lead citrate. Specimens were observed in a Jeol/FEI CM100 transmission electron microscope operated at 80‐kV accelerating voltage at the Core Facility for Integrated Microscopy at Panum Institute. The antibody‐based procedures used in this study comply with the recommendations made by the British Journal of Pharmacology.

2.9. Data and statistical analysis

The data and statistical analysis comply with the recommendations made of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018). For all animal experiments, at least five randomized mice were included in each group. For neutrophil and endothelial cells activation experiment in vitro, five repeat experiments were carried out. Data are presented as means ± SEM, and n indicates the number of animals per group. Statistical evaluations were performed using Mann–Whitney rank sum test for comparing two groups. P < .05 was considered significant. Statistical analysis was performed by using GraphPad Prism 7 software (GraphPad Software, La Jolla, CA, USA; RRID:SCR_002798).

2.10. Materials

TNF‐α and Z‐VAD‐FMK was supplied by R&D Systems Europe, Ltd., (Abingdon, Oxon, UK). PD150606 and PMA were supplied by Sigma‐Aldrich (Stockholm, Sweden). The chemokine CXCL2 (MIP‐2) was supplied by PeproTech (Rocky Hill, USA).

2.11. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander, Fabbro et al., 2017; Alexander, Kelly et al., 2017).

3. RESULTS

3.1. NET‐induced leukocyte recruitment

Surgical preparation of tissues for intravital microscopy is known to induce leukocyte rolling (Thorlacius, Raud, Xie, Hedqvist, & Lindbom, 1995; Yamaki, Lindbom, Thorlacius, Hedqvist, & Raud, 1998). Thus, as expected, local challenge with NETs generated by PMA stimulation did not significantly increase the number of rolling leukocytes (Figure 1a,b). In contrast, leukocyte rolling velocity was significantly decreased after challenge with NETs. In fact, NET reduced leukocyte rolling velocity by more than 66% (Figure 1a,b). Notably, intrascrotal injection of NETs triggered a dose‐dependent increase in firm leukocyte adhesion (Figure 1a,d). For example, NET (1.5 μg) enhanced leukocyte adhesion by more than 13‐fold (Figure 1d). We also studied the number of extravascular leukocytes and found that 1.5‐ and 4.5‐μg NET increased leukocyte emigration by eightfold more (Figure 1a,e). In addition, we examined if NETs generated by stimulation of CXCL2 induced similar effects. Indeed, CXCL2 generated NETs triggered neutrophil responses similar to the NETs generated by PMA (Figure 1). Based on these results, we selected 1.5‐μg NET induced by PMA stimulation for further studies. We next compared the effect of NETs with 0.5 μg of TNF‐α, which is known to cause a powerful leukocyte response in the cremaster muscle (Thorlacius, Lindbom, & Raud, 1997). Similar to NETs, TNF‐α stimulation had no effect on venular leukocyte rolling flux (Figure 1b). Interestingly, we observed that 0.5 μg of TNF‐α resulted in similar decreases in leukocyte rolling velocity (Figure 1c) and increases in leukocyte adhesion and emigration (Figure 2d,e) as compared to 1.5‐μg NET stimulation.

Figure 1.

NET‐induced leukocyte recruitment in vivo. (a) Intravital images of the cremaster microvasculature in response to vehicle, TNF‐α, CXCL2, NET 1.5, NET 0.5, 1.5, and 4.5. Leukocyte (b) rolling flux, (c) rolling velocity, (d) adhesion, and (e) emigration in mouse cremaster muscle 3 hr after intrascrotal challenge with vehicle, TNF‐α, CXCL2 NET 1.5, NET 0.5, 1.5, and 4.5. NETs were generated by PMA, and CXCL2 NETs 1.5 were generated by CXCL2 stimulation of neutrophils as described in Section 2. Data are mean ± SEM and n = 5 different animals for each group. *P < .05, significantly different from vehicle

Figure 2.

Neutrophil depletion. (a) Circulating neutrophils (Ly‐6G⁺/Ly‐6C⁺), leukocyte (b) rolling flux, (c) rolling velocity, (d) adhesion, and (e) emigration were determined in animals treated intraperitoneally with an antibody (Ab) directed against Ly‐6G (clone 1A8) or a control antibody 24 hr before intrascrotal challenge with NETs 1.5. Data are mean ± SEM and n = 5 different animals for each group. *P < .05, significantly different from control antibody

3.2. Neutrophils are the dominating leukocyte responding to NET challenge

We next wanted to identify which leukocyte subpopulation was the one responding to challenge with NETs. Thus, we used an antibody against Ly‐6G on neutrophils, which decreased the number of circulating neutrophils by 99% (Figure 2a) while having no effect on other leukocyte subsets, such as monocytes, T‐, and B‐cells (not shown). Notably, we observed that depletion of neutrophils abolished NET‐induced leukocyte rolling (Figure 2b,c), adhesion (Figure 2d), and emigration (Figure 2e), suggesting that neutrophils were the dominating leukocyte subtype responding to NET challenge.

3.3. NET‐induced neutrophil–endothelial cell interactions

To study the molecular mechanisms of NET‐induced neutrophil recruitment, we administered functional blocking antibodies before intrascrotal injection of NETs. Pretreatment with an antibody directed against P‐selectin reduced NET‐induced neutrophil rolling by 99%, firm adhesion by 95%, and emigration by 91% (Figure 3a–d). Moreover, immunoneutralization of PSGL‐1 decreased NETs‐evoked neutrophil rolling, adhesion, and emigration by 99%, 99%, and 90%, respectively (Figure 3a–d). In contrast, administration of an antibody against E‐selectin had no effect on NET‐induced neutrophil–endothelium interactions or extravascular accumulation (Figure 3a–d). We next studied the role of β₂‐integrins in NET‐triggered neutrophil accumulation. First, we observed that immunoneutralization of Mac‐1 and LFA‐1 had no effect on NET‐induced neutrophil rolling flux or rolling velocity (Figure 3a,b). However, administration of the antibody against Mac‐1 reduced the number of firmly adherent and emigrated neutrophil by 96% and 99%, respectively, in NET‐exposed tissues (Figure 3c,d). In addition, blocking LFA‐1 function had no effect on neutrophil rolling flux or rolling velocity (Figure 3a,b) but decreased NET‐evoked neutrophil adhesion by 98% and extravasation by 85% (Figure 3c,d). Administration of TNF‐α and NETs reduced microvascular blood flow and wall shear rate illustrated by the significant reduction in wall shear rate (Table 1). Table 1 shows that all treatments that reduced neutrophil–endothelial cell interactions also restored blood flow and wall shear rates towards baseline levels.

Figure 3.

Leukocyte (a) rolling flux, (b) rolling velocity, (c) adhesion, and (d) emigration in mouse cremaster muscle 3 hr after intrascrotal challenge with NETs 1.5. Animals were pretreated with an anti‐E‐selectin, anti‐P‐selectin, anti‐PSGL‐1, anti‐Mac‐1, anti‐LFA‐1, and a control antibody (Ab). Control animals received intrascrotal injection of PBS. Data are mean ± SEM and n = 5 different animals for each group. *P < .05, significantly different from vehicle and ^# P < .05, significantly different from control Ab

Table 1.

Microhaemodynamics

Microhemodynamics	Diameter (μm)	Flow velocity (mm·s−1)	Shear rate (S−1)
Vehicle	26.9 ± 0.9	1.3 ± 0.0	390.5 ± 16.5
TNF‐α	27.4 ± 1.4	1.1 ± 0.1*	327.6 ± 9.4*
NETs 0.5	27.4 ± 0.7	1.2 ± 0.0	343.0 ± 7.0*
NETs 1.5	28.8 ± 0.9	1.1 ± 0.0*	302.2 ± 13.4*
NETs 4.5	28.4 ± 1.5	1.1 ± 0.0*	309.0 ± 22.6*
NETs 1.5 + control Ab	25.5 ± 1.1	1.1 ± 0.0	339.5 ± 10.0
NETs 1.5 + anti‐P‐selectin Ab	27.7 ± 0.5	1.3 ± 0.0#	375.3 ± 7.7
NETs 1.5 + anti‐PSGL‐1 Ab	26.5 ± 0.9	1.3 ± 0.0#	396.5 ± 12.0#
NETs 1.5 + anti‐MAC‐1 Ab	25.4 ± 0.7	1.2 ± 0.0#	388.9 ± 8.6#
NETs 1.5 + anti‐LFA‐1 Ab	25.2 ± 1.1	1.2 ± 0.0#	387.8 ± 13.6
NETs 1.5 + anti‐HMGB1 Ab	26.0 ± 0.7	1.3 ± 0.0#	397.9 ± 6.5#
NETs 1.5 + Cas/Cal	23.6 ± 1.4	1.2 ± 0.0#	416.3 ± 19.2
NETs 1.5 + HMGB1 + Cas/Cal	26.1 ± 0.7	1.3 ± 0.0#	390.1 ± 6.6#
NETs 1.5 + DNase	26.0 ± 1.2	1.2 ± 0.0#	374.0 ± 7.9
NETs 1.5 + anti‐Ly6G Ab	27.5 ± 1.1	1.3 ± 0.0#	379.8 ± 7.3#
NETs 1.5 + anti‐TLR2 Ab	28.0 ± 0.5	1.2 ± 0.0#	363.4 ± 4.6
NETs 1.5 + anti‐TLR4 Ab	29.3 ± 1.4	1.2 ± 0.0#	332.9 ± 12.6

Diameter, flow velocity, and shear rate in cremaster muscle venules quantified in control animals and animals treated intrascrotally with NETs. Animals were pretreated with an anti‐P‐selectin, anti‐PSGL‐1, anti‐Mac‐1, anti‐LFA‐1, anti‐HMGB1, anti‐Ly6G, anti‐TLR2, anti‐TLR4, a control antibody (Ab), and DNase. Data are mean ± SEM from n = 5 experiments.

^* P < .05, significantly different from vehicle.

^# P < .05, significantly different from control Ab. Cas/Cal, caspase and calpain inhibitors.

3.4. Role of NET‐associated MPs in neutrophil recruitment

We next asked if MPs attached to NETs play a role in NET‐induced inflammation. As shown recently (Wang et al., 2018), PMA stimulation of neutrophils triggers the formation of NETs containing numerous MPs (Figure 4a,c). Co‐incubation of neutrophils with calpain and caspase inhibitors resulted in the formation of NETs with markedly less MPs (Figure 4b,d). NETs generated in the presence of caspase and calpain inhibitors exhibited intact levels of MPO (Figure 4e) and S100A9 (Figure 4f), which are common proteins attached to NETs (Urban et al., 2009). It was found that NETs depleted of MPs by co‐incubation with inhibitors of caspase and calpain had a significantly lower capacity to trigger neutrophil recruitment (Figure 5a–e). For example, co‐incubation with caspase and calpain inhibitors decreased NET‐induced neutrophil adhesion by 65% (Figure 5d) and emigration by 77% (Figure 5e). NET‐induced rolling flux of neutrophils was independent of MP (Figure 5b), whereas leukocyte rolling velocity was higher after challenge with NETs without MPs compared to NETs with MPs (Figure 5c). Notably, disintegration of NETs by using DNase reduced NET‐induced neutrophil adhesion (Figure 5d) and emigration (Figure 5e) down to levels similar to those to NETs depleted of MPs (Figure 5d,e). In separate experiments, we observed that NETs increased Mac‐1 expression on isolated neutrophils (Figure 6a) and ICAM‐1 gene expression in endothelial cells (Figure 6b), indicating that NETs can activate both neutrophils and endothelial cells directly. Notably, NETs depleted of MPs were significantly less effective activators of neutrophils and endothelial cells (Figure 6a,b).

Figure 4.

Formation of NET‐MP complexes. NETs were generated from PMA‐stimulated bone marrow neutrophils in the presence of (a,c) vehicle or (b,d) caspase and calpain (Cas/Cal) inhibitors. Scanning electron microscopy showing MPs attached to neutrophil‐derived NETs. Scale bar = 2 μm. MPs are denoted in pink colour in the lower inserts. (e) MPO and (f) S100A9 levels in NETs were quantified as described in Section 2. Data are mean ± SEM and n = 5 different animals for each group

Figure 5.

NET‐MP complex‐induced leukocyte recruitment in vivo. (a) Intravital images of the cremaster microvasculature in response to vehicle, NETs co‐incubated with DNase, and NET co‐incubated with caspase and calpain inhibitors (Cas/Cal). Neutrophil (b) rolling flux, (c) rolling velocity, (d) adhesion, and (e) emigration in mouse cremaster muscle 3 hr after intrascrotal challenge with vehicle, NETs co‐incubated with DNase and NETs co‐incubated with caspase, and calpain inhibitors vehicle. Data are mean ± SEM and n = 5 different animals for each group. *P < .05, significantly different from vehicle

Figure 6.

NET‐induced activation of neutrophils and endothelial cells. (a) Mac‐1 expression on isolated bone marrow neutrophils. (b) Relative expression of ICAM‐1 in endothelial cells. Cells were incubated with NETs generated in the absence or presence of caspase and calpain inhibitors (Cas/Cal). Flow cytometry was used to quantify Mac‐1 levels, and quantitative RT‐PCR was used to determine the mRNA levels of ICAM‐1 and β‐actin as described in Section 2. CXCL2was used as a positive control in the flow cytometric analysis. Data are mean ± SEM and n = 5. (a) n = 5 different animals for each group, (b) n = 5 samples run in triplicate. *P < .05, significantly different from 0‐hr vehicle and ^# P < .05, significantly different from vehicle

3.5. NET‐induced neutrophil recruitment is mediated by MP‐derived HMGB1

By using transmission immunoelectron microscopy, we found that MPs associated with NETs not only expressed Mac‐1 (Figure 7a,b) but also expressed HMGB1 (Figure 7c,d). We therefore asked whether HMGB1 expressed on MPs plays an important role in NET‐evoked neutrophil recruitment. Immunoneutralization of HMGB1 had no effect on leukocyte rolling flux but increased leukocyte rolling velocity (not shown) along the venular endothelium exposed to NETs. Moreover, we found that inhibition of HMGB1 significantly attenuated NET‐induced neutrophil adhesion by 82% (Figure 7e) and emigration by 74% (Figure 7f). Notably, immunoneutralization of HMGB1 had no further inhibitory effect on neutrophil adhesion and extravasation triggered by NETs depleted of MPs (Figure 7e,f).

Figure 7.

Role of HMGB1, TLR2, and TLR4 in NET‐induced neutrophil recruitment. Transmission electron microscopy showing NETs and MPs incubated with a gold‐labelled anti‐citrullinated histone H3 antibody (small gold particles, black arrowhead), (a, b) gold‐labelled anti‐Mac‐1 antibody (large gold particles, black arrow), and (c, d) gold‐labelled anti‐HMGB1 antibody (large gold particles, black arrow). Scale bar = 0.25 μm. Neutrophil (e, g) adhesion and (f, h) emigration in mouse cremaster muscle 3 hr after intrascrotal challenge with NETs 1.5. Animals were pretreated with an anti‐HMGB1, anti‐TLR2, anti‐TLR4 antibody, and a control antibody (Ab). Control animals received intrascrotal injection of PBS. Data are mean ± SEM and n = 5 different animals for each group. *P < .05, significantly different from control Ab. Cas/Cal, caspase and calpain inhibitors

3.6. TLR2 and TLR4 mediates NET‐provoked neutrophil recruitment

Finally, we examined the potential role of TLR2 and TLR4 in NET‐induced neutrophil accumulation. Pretreatment with an antibody directed against TLR2 reduced the number of adherent and extravasated neutrophil by 78% (Figure 7g) and 61% (Figure 7h), respectively, in response to NET challenge. In addition, immunoneutralization of TLR4 significantly decreased NET‐evoked neutrophil adhesion by 79% (Figure 7g) and emigration by 78% (Figure 7h).

4. DISCUSSION

NET formation is a key feature in severe inflammation and infection. Our study demonstrates that NET‐MP aggregates are powerful stimulators of neutrophil recruitment in vivo. The results show that MPs attached to NETs are critical for inducing neutrophil adhesion and extravasation in the microvasculature. HMGB1 expressed on MPs triggers this pro‐inflammatory effect of NET‐MP complexes via TLR2 and TLR4 signalling. Thus, these findings suggest that MPs attached to NETs play an important role in leukocyte recruitment and could be used as targets to treat clinical conditions dominated by acute inflammatory reactions.

Tissue accumulation of leukocytes is generally considered to be a multistep process, in which initial leukocyte rolling is followed by firm adhesion and transendothelial migration out into the extravascular space (Montefort et al., 1994; Vestweber & Blanks, 1999). This study shows that NETs are potent inducers of all steps of the extravasation process of leukocytes at sites of inflammation. This effect of NETs was not only specific for NETs generated by PMA but was also observed after challenge with NETs triggered by CXCL2 stimulation of neutrophils, indicating that NETs in general are potent stimulators of leukocyte recruitment. We observed that neutrophil depletion abolished leukocyte responses to NETs, suggesting that neutrophils were the main leukocyte subtype responding to NET challenge. Moreover, we found that NETs can activate both neutrophils and endothelial cells directly. Numerous previous studies have shown that disintegration of NETs by using DNase or inhibition of NET formation by inhibition of PAD4, an enzyme necessary for NET formation, reduce pathological accumulation of inflammatory cells in conditions, such as sepsis (Luo et al., 2014), pancreatitis (Merza et al., 2015), ischaemia–reperfusion injury (Ge et al., 2015), and inflammatory bowel (He et al., 2016) and rheumatic (Apel, Zychlinsky, & Kenny, 2018) diseases. Herein, we could show for the first time by using intravital microscopy that NETs per se trigger neutrophil rolling, firm adhesion, and emigration in the microvasculature in vivo. It should be noted that although NETs did not increase leukocyte rolling flux above levels of that triggered by the preparation of cremaster muscle tissue for microscopy, NETs markedly decreased leukocyte rolling velocity, which is mediated by up‐regulation of endothelial adhesion molecules (Steeber, Campbell, Basit, Ley, & Tedder, 1998). In fact, the potency of NETs in provoking neutrophil–endothelium interactions and emigration was similar to that of TNF‐α, a well‐known inducer of neutrophil accumulation (Thorlacius et al., 1997). Note that TNF‐α, similar to NETs, did not increase leukocyte rolling flux but also decreased leukocyte rolling velocity. In the present study, we found that immunoneutralization of P‐selectin and its major counter‐receptor PSGL‐1 not only inhibited NET‐induced neutrophil rolling by more than 99% but also concomitantly abolished neutrophil adhesion and extravasation, suggesting that P‐selectin/PSGL‐1 interactions play a key role in NET‐triggered tissue accumulation of neutrophils. In contrast, immunoneutralization of E‐selectin had no effect on NET‐induced neutropil responses in the cremaster muscle microvasculature. Our data are line with and extend on previous studies showing that P‐selectin and PSGL‐1 exert a central function in mediating neutrophil rolling in TNF‐α‐ and chemokine‐induced inflammation (Mansson et al., 2000; Wan, Wang, Liu, Schramm, & Thorlacius, 2003; Zhang, Liu, Wang, & Thorlacius, 2001) as well as in models of septic lung injury (Asaduzzaman et al., 2009), reperfusion injury (Riaz et al., 2002), colitis (Wan et al., 2002), and cholestatic liver damage (Dold et al., 2010). Herein, we also asked which mechanisms regulate NET‐provoked neutrophil firm adhesion and extravasation. We show that inhibition of Mac‐1 and LFA‐1 decreased NET‐induced firm adhesion of neutrophils on the venular endothelium in vivo, indicating that both LFA‐1 and Mac‐1 facilitate adhesive interactions in response to NET stimulation. Previous work have reported contradictory data on the individual role of Mac‐1 and LFA‐1 in specific models of inflammation (Issekutz et al., 1995; Li et al., 2004; Riaz et al., 2002; Rutter et al., 1994; Thorlacius et al., 2000). However, considering more recent investigations, it is possible that both of these adhesion molecules can cooperate for optimal recruitment of inflammatory cells. For example, one study showed that LFA‐1 may initiate first stable contacts and that Mac‐1 establishes more sustainable adhesion onto activated endothelial cells in inflamed organs (Ding et al., 1999). Herein, we found that inhibition of Mac‐1 or LFA‐1 decreased neutrophil emigration, suggesting that both these molecules are important for extravasation of neutrophils in NET‐provoked inflammation. It is notable that one previous study reported a role of integrins in facilitating leukocyte rolling in the cremaster muscle (Dunne, Ballantyne, Beaudet, & Ley, 2002). In contrast, we found that inhibition of Mac‐1 and LFA‐1 had no effect on NET‐evoked neutrophil rolling. Our findings are also supported by numerous studies showing that targeting Mac‐1 and LFA‐1 reduces leukocyte adhesion while having no effect on leukocyte rolling (Arfors et al., 1987; Becker, Garman, Whitcup, Planck, & Rosenbaum, 2001; Dold, Laschke, Lavasani, Menger, & Thorlacius, 2008; Mihaescu et al., 2007; Nolte et al., 1994; Riaz et al., 2002; von Andrian et al., 1991). In addition, considering the dominating role of P‐selectin and PSGL‐1, it is perhaps not surprising that β₂‐integrins appear to play no role in NET‐induced‐provoked neutrophil rolling. For therapeutic reasons, it should be mentioned that targeting leukocyte rolling might be a theoretically more attractive strategy knowing that the rolling interaction is related to endothelial cell activation, whereas rolling leukocytes remain inactivated until stimulated by a chemoattractant causing stationary adhesion. Thus, prevention of rolling results in the dislodgement of non‐activated leukocytes back into the circulation. In contrast, targeting firm adhesion and transmigration dislodges activated leukocytes back into the circulation where they can cause remote tissue damage in the liver or lung (Mercer‐Jones et al., 1997).

NETs are known to be loaded with numerous different nuclear, cytoplasmatic, and granular proteins, which contribute to the biological effects of NETs (Brinkmann et al., 2004). We have recently demonstrated that MPs formed during neutrophil activation can bind to NETs and form NET‐MP complexes, which are important for thrombin generation in sepsis (Wang et al., 2018). In the present study, we wanted to define the role of NET‐MP aggregates in the regulation of neutrophil recruitment in acute inflammation. By using electron microscopy, we could confirm that co‐incubation of caspase and calpain inhibitors abolished concomitant generation MPs resulting in the formation of NETs without MPs. Notably, these NETs depleted of MPs were significantly less effective in triggering neutrophil accumulation. For example, NET‐induced neutrophil adhesion and emigration were attenuated by more than 65% and 77%, respectively, indicating that NETs constitute a functional assembly scaffold for MPs in acute inflammation. In fact, this reduction in neutrophil adhesion and emigration in response to NETs depleted of MPs was similar to that observed when disintegrating NETs by using DNase, indicating that MPs attached to NETs exert a dominating role in NET‐induced neutrophil recruitment. Thus, our findings demonstrate that NET‐MP complexes are potent inducers of neutrophil infiltration, and it is concluded that NET‐MP complexes might have broad implications on various neutrophil‐dependent diseases, such as sepsis, acute pancreatitis, and ischaemia–reperfusion injury. HMGB1 is a ubiquitously expressed nuclear protein stabilizing nucleosome formation and can be released to the extracellular space upon cell injury and function as damage‐associated molecular pattern molecule with pro‐inflammatory functions (Chen, Ward, Sama, & Wang, 2004). For example, HMGB1 has been shown to cause vascular barrier dysfunction and accumulation of neutrophils (Nawaz & Mohammad, 2015). Based on a recent publication showing that platelet‐derived MPs from patients with systemic sclerosis express high levels of HMGB1 (Maugeri et al., 2018), we asked, in our experiments, whether neutrophil‐derived MPs also contain HMGB1. By using transmission immunoelectron microscopy, we found that neutrophil‐derived MPs attached to NETs express HMGB1. Thus, these findings indicate that both platelet‐ and neutrophil‐derived MPs express HMGB1 which could be a general feature of MP formation, independent of stimulus and cell origin. Nonetheless, immunoneutralization of HMGB1 markedly reduced NET‐induced neutrophil adhesion and emigration in vivo, suggesting that HMGB1 plays an important role in the pro‐inflammatory effects of NET‐MP aggregates. The main receptors of HMGB1 constitute TLR2, TLR4, and RAGE (Park et al., 2006). Here, we found that inhibition of TLR2 or TLR4 greatly attenuated the number of adherent and extravasated neutrophils in response to NET‐MP aggregate challenge, suggesting that both TLR2 and TLR4 signalling mediated NET‐triggered neutrophil recruitment. This notion is in line with a previous study showing that both TLR2 and TLR4 are critical for HMGB1‐induced accumulation of stem cells along microvascular endothelial cells in vivo (Furlani et al., 2012). In this context, it should be mentioned that both endothelial cells and neutrophils express varying levels of TLR2 and TLR4 (Khakpour, Wilhelmsen, & Hellman, 2015; Sabroe, Jones, Usher, Whyte, & Dower, 2002), and future studies should address their relative importance in regulating neutrophil–endothelium interactions in response to NET‐MP complexes.

In conclusion, our study demonstrates that NET‐induced neutrophil recruitment is dependent on neutrophil‐derived MPs expressing HMGB1 and mediated via TLR2 and TLR4 signalling in vivo. Thus, blocking MP binding to NETs or downstream inhibition of the HMGB1‐TLR2/TLR4 axis might provide useful targets to attenuate NET‐dependent tissue damage in acute inflammation.

CONFLICT OF INTEREST

The authors have no financial conflicts of interest.

AUTHOR CONTRIBUTIONS

Y.W., F.D., A.H., and M.M. performed the experiments. H.T. designed the study. All authors analysed the data and contributed to the writing.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design & Analysis, Immunoblotting and Immunochemistry, and Animal Experimentation, and as recommended by funding agencies, publishers and other organisations engaged with supporting research.

Wang Y, Du F, Hawez A, Mörgelin M, Thorlacius H. Neutrophil extracellular trap‐microparticle complexes trigger neutrophil recruitment via high‐mobility group protein 1 (HMGB1)‐toll‐like receptors(TLR2)/TLR4 signalling. Br J Pharmacol. 2019;176:3350–3363. 10.1111/bph.14765