Objectives Behçet’s disease (BD) is a chronic systemic vasculitis. Thrombosis is a frequent and life-threatening complication. The pathogenesis of BD is poorly understood and evidence supporting a role for primed neutrophils in BD-associated thrombotic risk is scant. To respond to inflammatory insults, neutrophils release web-like structures, known as neutrophil extracellular traps (NETs), which are prothrombotic. We evaluated the role of NETs and markers of NETs in BD.
Methods Blood samples were collected from patients with BD, according to the International Study Group Criteria for Behçet's disease, and healthy donors (HD). NET components, including cell-free DNA (CfDNA) and neutrophil enzymes myeloperoxidase (MPO), were assessed in serum or in purified neutrophils from patients with BD and HD.
Results Patients with active BD had elevated serum cfDNA levels and MPO-DNA complexes compared with patients with inactive BD and to HD. In addition, levels of cfDNA and MPO-DNA complexes were significantly higher in patients with BD with vascular involvement compared with those without vascular symptoms. Purified neutrophils from patients with BD exhibited spontaneous NETosis compared with HD. Thrombin generation in BD plasma was significantly increased and positively correlated with the levels of MPO-DNA complexes and cfDNA. Importantly, DNAse treatment significantly decreased thrombin generation in BD plasma but not in HD plasma. In addition, biopsy materials obtained from patients with BD showed NETs production in areas of vasculitic inflammation and thrombosis.
Conclusions Our data show that NETs and markers of NETS levels are elevated in patients with BD and contribute to the procoagulant state. Targeting NETs may represent a potential therapeutic target for the reduction or prevention of BD-associated thrombotic risk.
- neutrophils extracellular traps
Statistics from Altmetric.com
What is already known about this subject?
Behçet’s disease (BD) is a chronic systemic vasculitis with life-threatening complications associated with thrombosis.
What does this study add?
High levels markers of neutrophil extracellular traps (NETs) are present in Behçet’s patients and may contribute to the procoagulant state.
Neutrophils from Behçet’s patients are prone to NETosis.
NETs are present in inflamed vessels of Behcet’s patients in areas of vasculitic inflammation and thrombosis.
How might this impact on clinical practice or future developments?
NETs may represent a biological marker of BD severity and targeting NETosis may provide a therapeutic option in the treatment of this disease and its associated complications.
Behçet’s disease (BD) is a chronic systemic vasculitis characterised by mucocutaneous, ocular, gastrointestinal and cerebral recurrent lesions. This immune-inflammatory disorder involves different vessel types and sizes of the vascular tree and is often complicated by recurrent thrombosis, particularly in the venous compartment.1 However, the exact pathogenic mechanism underlying the thrombotic tendency is still unclear.2 Diagnosis of BD is primarily based on clinical manifestations and new criteria of the International Team for the revision of the international criteria for Behçet disease (ICBD) are now used in numerous studies.3
The general pathophysiology of BD is poorly understood, although dysregulation of adaptive and innate immunity in response of infectious triggers has been described.4 Chronic inflammation present in BD indicates an increased oxidative condition that induces platelet, leucocyte and endothelial cell activation through the release of proinflammatory cytokines and chemokines. Neutrophils of patients with BD are hyperactivated and exhibit increased phagocytosis and superoxide production,5 which potentially contribute to clot formation by fibrinogen oxidation.6
Under inflammatory or infectious conditions, neutrophils are able to generate neutrophil extracellular traps (NETs) via a distinct process of cell death termed NETosis.7 NETs consist of extruded cell-free DNA (CfDNA) decorated with histones and granular components that include antimicrobial peptides and proteases. The molecular pathways leading to NETosis include calcium mobilisation, generation of reactive oxygen species (ROS), nuclear delobulation involving enzymatic activities of myeloperoxidase (MPO) and neutrophil elastase and chromatin modification via the citrullination of histones by the peptidyl arginine deiminase (PAD4).8 Numerous studies have implicated NETs in the aetiology of autoinflammatory or autoimmune conditions such as systemic as autoimmune vasculitis.9 10 Furthermore, NETs are now recognised as a key trigger of thrombus initiation and progression in pathological conditions such as, deep vein thrombosis in mice and humans11 but also in arterial diseases such as stroke and myocardial infarction.12 13 In this study, we hypothesise that NETs might be the missing link in thrombosis-associated BD.
Material and methods
Patients over 18 years old of age diagnosed with BD according to ICBD from two departments of internal medicine in Paris were included. Patients treated with biologics or with corticosteroids >10 mg/day were excluded. Patients were divided into active BD (ABD) or inactive BD (IBD). Patients with IBD were defined by the absence of clinical manifestation (except unmodified aphthous ulcers or arthralgia) and increased inflammatory parameters. Patients with ABD were defined by the presence of clinical manifestation of BD and increased inflammatory parameters in the absence of infection and requiring a treatment modification, blood samples were collected before treatment modification.
Patients were also divided into angio BD or non-angio BD: patients were defined as angio BD if they had a history of arterial aneurysm or occlusion and/or venous thrombosis.
Demographic, clinical and analytical parameters were prospectively collected.
We provided instructional support to foster knowledge of research concepts and terminology. If the patients asked, we provided them information about the course of the study. Patients will be informed of the publication of the study and an understandable explanation will be given.
Blood samples from healthy donors (HD) or Behcet’s patients were collected by venipuncture into vacutainers containing 3.2% sodium citrate. Platelet-poor plasma (PPP) was prepared. Sera were obtained from blood collected in dry tubes and centrifugation at 2500 g for 12 min.
Blood was collected by venipuncture into ethylenediaminetetraacetic acid (EDTA)-coated vacutainers. Neutrophils were isolated by a single-step centrifugation of whole blood onto Polymorphprep as per the manufacturer’s recommendation less than 3 hours after venipuncture. Briefly, whole blood was layered onto Polymorphprep and centrifuged at 450 g for 40 min. The granulocyte layer was resuspended in Roswell Park Memorial Institute (RPMI) media with EDTA 2 mM and centrifuged at 500 g for 5 min. Cells were resuspended in RPMI with fetal bovine serum (FBS) 3% at a final concentration of 105/mL. Viability, assessed using tryptan blue, was >99% and purity, assessed by flow cytometry, was >94%.
cfDNA in serum was quantified using Quant-iT PicoGreen DNA quantification kit (Invitrogen) in accordance with the manufacturer’s instructions. Briefly, pure dsDNA standards and samples were incubated for 5 min with 100 µL of Quant-iT PicoGreen reagent. Fluorescence signals were measured in a microplate fluorescence reader (FLUOstar Optima, BMG labtech) with filter settings of 485 nm (excitation) and 538 nm (emission).
Myeloperoxydase (MPO)-DNA complexes
MPO-DNA complexes were identified using a capture ELISA.14 As the capturing antibody, 5 µg/mL anti-MPO monoclonal antibody (Abcam) was coated to 96-well microtiter plates overnight at 4°C. After blocking in incubation buffer, 40 µL of patient serum was added per well in combination with the peroxidase-labelled anti-DNA monoclonal antibody (component no. 2 of cell death detection ELISA kit; Roche, Cat. No:11774425001) following the manufacturer’s instructions. After 2 hours of incubation at room temperature (RT) on a shaking device, the samples were washed phosphate-buffered saline (PBS) and the 3.3’, 5.5’-tetramethylbenzidine was added. The absorbance at 450 nm wavelength was measured using a microplate fluorescence reader (FLUOstar Optima, BMG labtech) after 15 min incubation at 37°C in the dark.
ROS generation by flow cytometry
Neutrophils at 5.106/mL were incubated with an ROS probe (probeCM-H2DCFA) (Molecular probes) and then stimulated, or not, with 25 nM of phorbol 12-myristate 13-acetate (PMA) for 20 min. Reaction was stopped by adding PBS with 2% bovine serum albumin (BSA) and EDTA 2 mM. Fluorescence was measured by flow cytometry (Acurri C6, BD).
Visualisation and quantification of NETs
Isolated neutrophils (105 cells) were seeded on glass coverslips and allowed to adhere for 30 min before stimulation with PMA (25 nM) for 3 hours at 37°C and 5% CO2 and then fixed with paraformaldehyde for 30 min. Neutrophils were then stained for NETs markers. Briefly, after 1-hour blocking (PBS with 3% goat serum, 1% BSA), NETs were detected using an mouse anti-MPO primary antibody (Abcam, Cambridge, Massachusetts, USA), diluted 1/100 in blocking buffer for 2 hours at 37°C. Slides were then incubated with 1/200 Alexa Fluor 555 goat antimouse antibody,) for 2 hours at 37°C. DNA was stained with 4′,6-diamidino-2-phenylindole (DAPI) (Fluoroshield Mounting Medium with DAPI, Abcam). NETs were visualised by using a nanozoomer imaging fluorescence microscope (Hamamatsu) and blindly quantified using image J software. Images were evaluated for MPO and DNA costaining; nuclear phenotypes and NETs were counted and expressed as percentage of the total number of cells in the fields.
Thrombin generation assay
Thrombin generation was continuously measured in platelet-free plasma (PFP) using the thrombogram method as previously described.15 Tissue factor 0.5pM (TF; Diagnostica Stago) was used as trigger. In some experiments, PFP was incubated for 2 hours at 37°C with DNAse (Pulmozyne, Roche) 100 µg/mL or vehicle (HEPES buffer) before the assay.
Four micrometre paraffin-embedded human aorta sections were incubated overnight with the primary anti-MPO antibody (dilution 1:100, Abcam) or anticitrullinated histone H4 (dilution 1:500, Millipore). Primary antibodies were visualised with an Alexa-647 labelled antibody against rabbit immunoglobulins (Interchim, France). Slides were prepared with mounting medium supplemented with DAPI (Prolong Gold antifade reagent, Abcam). Immunofluorescence staining was visualised using the nanozoomer.
Continuous variables are presented with the median and extreme values or with the mean±SEM Categorical variables are presented with counts and proportions. Statistical comparisons were performed by using the Mann-Whitney U test for quantitative unpaired data, two-way analysis of variance or Kruskal-Wallis for multiple comparisons, Spearman correlation test for correlations. All statistical tests were two tailed with a significance level of 0.05. Statistical significance was evaluated using GraphPad Prism V.5.00 for Mac (GraphPad Software, San Diego, California, USA).
Clinical characteristics of patients and healthy controls
Seventy-three patients were included in the study. Clinical characteristics and laboratory findings of patients are shown in table 1. Twenty-eight patients (38.4%) were classified as active. Active patients were younger than inactive ones (35 years vs 42 years, respectively, p=0.04) and had greater biological signs of inflammation (C reactive protein (CRP) 34 vs 5.25 mg/L, respectively, p=0.001 and leucocytes count at 10.4 10ˆ9/L vs 7.29 10ˆ9/L p=0.017). Forty-two (57.5%) received colchicine, 42 (57.5%) glucocorticoids (<10 mg/day) and 29 (39.7%) disease-modifying antirheumatic drugs (DMARDs). Fifteen HDs were recruited, of which 8 (53.3%) were men and the mean age was 36 years old.
NETs component levels are increased in serum of Behcet’s patients
We assess cfDNA levels, indirect markers of NETs and circulating MPO-DNA complex in serum of patients and HD. In serum, the levels of cfDNA and circulating MPO-DNA complexes were higher in patients than in HD (1669±122.6 vs 951.2±63.9 ng/mL, p=0.0001 and 2.321±0.187 vs 0.316±0.098 OD p=0.0001, respectively). Interestingly, the levels of cfDNA and circulating MPO-DNA complexes were significantly higher in active patients than in inactive (figure 1A,B). In addition, levels of cfDNA and MPO-DNA complexes were significantly higher in angio-BD patients compare to non-angio BD (figure 1C,D).
The levels of cfDNA and circulating MPO-DNA complexes were positively correlated (r2=0.576, p<0.001) but were not correlated with age, CRP or leucocytes count (cfDNA and CRP, p=0.6, cfDNA and leucocytes, p=0.1, MPO-DNA complexes and CRP, p=0.4, MPO-DNA complexes and leucocytes, p=0.9). Levels of cfDNA and circulating MPO-DNA complexes were not different between groups based on clinical characteristics (sex, skin, ocular neurological joint and gut involvement), treatments (colchicine, glucocorticoids, immunosuppressive and anticoagulant) received (see online supplementary figure S1A,B) and ICBD score (see online supplementary figure S1C,D).
Behcet’s neutrophils are primed to NETosis
We hypothesised that the increased level of NETs in BD could be secondary to an increased release by neutrophils. Neutrophils isolated from Behçet patients produced more ROS than those isolated from HD in resting conditions (1618±241.9 vs 622.1±120.2 MFI, respectively) and after stimulation with PMA (9037±2079 vs 2367±894 MFI, respectively) (see online supplementary figure S2-A). We then assessed the ability of purified neutrophils from patients and HD to undergo NETosis. As shown in figure 2, neutrophils isolated from active patients produced more NETs than those from inactive and HD’s patients (51.21±3.3 vs 29.35±2.6 vs 7.2%±1.2%, respectively) in resting conditions. After PMA activation BD-derived neutrophils produced more NETs than HD’s (41.8±4.8 vs 18%±3.8%, respectively, p=0.0023) (see online supplementary Figure S2B). Patient’s or HD’s neutrophils treated with DNAse show no NETs formation. These results indicate that BD neutrophils are in a preactivated state and are prone to NETosis.
Hypercoagulability in Behcet’s patients is associated with increased levels of NETs component
To investigate the role of markers of NETs in the hypercoagulable state in Behcet’s patients, we performed thrombin generation in PPP. As shown in figure 3, the peak and the velocity of thrombin generation and the endogenous thrombin potential (ETP)) in plasma from active Behçet’s patients were higher than in Inactive and HD plasma (341±12.5 vs 239.6±16.35 vs 154.7±12.61 nM, 105.6±11.1 vs 56.0±7.4 vs 31.25±3.9 nM/min and 2293±106.4 vs 1867±117.6 vs 1408±92.6 nM/Min, respectively). However, the other parameters (lag time and time to peak) were not statistically different between patients and HD (see online supplementary table S1). These data indicate that the propagation phase of the coagulation is enhanced in BD plasma.
Interestingly, the peak and velocity were positively correlated with cfDNA (r2=0.80 p=0.02 and r2=0.85 p=0.006, respectively) (figure 4A,B), and with circulating MPO-DNA complexes (figure 4C,D) (r2=0.9 p=0.001 and r2=0.85 p=0.006, respectively) levels in serum. To assess the effect of cfDNA in thrombin generation, plasma from HD and BD was incubated with or without DNAse. In the presence of DNAse, the peak of thrombin and the velocity and the ETP were significantly reduced in inactive and active plasma patients whereas DNAse had no effect on thrombin generation parameters in HD plasma (figure 5 and see online supplementary table S2), suggesting that the elevated levels of markers of NETs in BD contribute to the procoagulant state. As a control, we performed thrombin generation in HD’s plasma with NETs supplemented or not with DNAse (see online supplementary figure S3) and confirmed that markers of NETs enhanced the thrombin generation.
NETs are identified in inflammatory aorta from Behçet’s patients
To obtain evidence of in vivo NETs formation and confirmation of a putative pathogenic role on vessels, we analysed aorta samples obtained from Behçet’s patients with oral ulcers. Histological evaluation confirmed the presence of both vasculitis and microthrombosis on aorta sections. Vascular pathology was revealed by perivascular neutrophilic inflammation with fibrinoid necrosis and thrombosis at the intima layer of the aorta. Patients were characterised as follows: patient 1, a 39-year-old woman with an history of oral and genital ulcers, developed a pseudo-aneurysm of the subrenal abdominal aorta, patient 2, a 55-year-old man, had an inflammatory syndrome, oral and genital ulcers and a positive pathergy test with a thrombus in the right ventricle, patient 3, a 58-year-old man with an history of oral and genital ulcers, had a pseudoaneurysm of the abdominal aorta with aortitis. All these patients fulfilled ICB criteria and respond to immunosuppressive therapies after surgery. By immunofluorescence microscopy, we quantified infiltrating NETing neutrophils (positive staining for MPO, citrullinated histone 4 and DNA) in the aorta. Immunofluorescence studies suggested the presence of NETs in areas presenting with vasculitis or thrombotic lesions and in areas of neutrophilic infiltration. These results demonstrate that infiltrating NETing neutrophils are seen in areas of vasculitis and microthrombi in tissues from patients with BD, suggesting a potential role of NET formation in the pathogenesis of the disease (figure 6).
BD is a model of human vasculitis that primarily affects young adults. Although the pathogenesis of BD is largely unknown, it is accepted that BD is at the crossroads between autoimmune and autoinflammatory diseases.4 Besides adaptive immune system dysregulation, innate immunity seems to be involved in the pathophysiology of the disease.5 6 Recurrent thrombosis events occur more frequently in male patients with active disease, involving both arterial and venous vessels of all sizes, but deep and superficial vein thrombosis of the lower extremities are the most common vascular manifestations of the disease.3 4
So far, there is no clear evidence of thrombosis mechanisms in Behçet patients and there is no specific tool or plasma/serum biomarker to identify and quantify the severity of BD. Patients with BD present with increased circulating leukocytes as well as high levels of cytokines.5 Despite the poorly defined aetiology of BD, neutrophil hyperfunction has been proposed to be the basis of BD since 1975 when increased chemotaxis of BD neutrophils was first demonstrated in vitro.16 17 Then, BD neutrophils were found to be activated using flow cytometry in BD-active cases.18 Similarly, infiltrate of neutrophils were evidenced by histopathology in skin, synovial, intestinal and central nervous system biopsies.19 20 More recently, a study has shown increased ROS production by neutrophils.6 All these data support the critical role of neutrophils in BD pathophysiology. Based on clinical outcomes, the European League Against Rheumatism recommendations for the management of BD has suggested that thrombosis should be treated with immunosuppression3 rather than anticoagulation, which suggests a continuous cross-talk between inflammation and coagulation. However, the causal link between activated leucocytes and thrombosis has not yet been clearly established in BD.
In the present study, we show that neutrophils from patients with BD are prone to undergo NETosis in vitro even without stimulation. Consistent with previous studies, we observed that ROS levels in BD neutrophils are higher compared with HD. Importantly, we show that markers of NETs are significantly increased in the serum of active BD compared with inactive BD and to HD, supporting our hypothesis that activated neutrophils and NETs might play a role in BD pathophysiology.
Besides the fact that NETs levels and markers of NETs are increased in BD in the active stage, we showed that they are higher in angio-Behçet patients than in non-angio Behçet suggesting an association with the potential to contribute to the vascular disease.
Espinosa et al found that BD is associated with endothelial cell injury accompanied by increased thrombin generation.21 Presence of thrombophilic factors in BD does not seem to be the basis of this clinical and biological hypercoagulable state. As NETs are now well known to play a key role in thrombosis,22 we aimed to investigate the role of NETs in BD in an in vitro model of thrombin generation. NETs can activate coagulation via the intrinsic and extrinsic pathways, and increase polyphosphate-triggered thrombin generation in plasma.6 Histones have been shown to increase thrombin generation in a platelet-dependent and platelet-independent manner23 and addition of NETs in a PFP has been shown to enhance thrombin generation.23 In the context of BD, platelets have been shown to be sensitive to adenosine diphosphate (ADP)-dependent activation. Like in patients with sepsis, it is likely that the platelets–thrombin–NETs axis promotes intravascular coagulation and vascular dysfunction in BD. Our report is in agreement with a previous study by Mejia et al, that thrombin generation, more specifically the velocity, in BD is increased.24 Other procoagulant parameters were also measured by using calibrated automated thrombogram (CAT) and rotational thromboelastometry (ROTEM) which are two global haemostasis assays. Both ROTEM and CAT tests revealed that patients with BD have a procoagulant state and reported that patients with BD have increased levels of fibrinogen, plasminogen activator inhibitor-1 (PAI-1) and thrombin–antithrombin complex (TAT) arguing for a global prothrombotic state of this pathology.2 Notably, Yurdakul et al showed that D-dimer levels are increased in the acute phase of thrombosis in patients with BD.25 In addition, our results show that the velocity and peak of thrombin generation are positively correlated with markers of NETs in sera and that disruption of NETs with DNAse significantly reduces thrombin generation in BD but not in HD. These data suggest that increased levels of soluble NETs components in BD’s sera represent an important factor promoting efficient thrombin generation in an in vitro model.
Interestingly, several cases of patients with BD with intracardiac thrombi revealed that anticoagulation treatments fail to prevent recurrent thrombosis but immunosuppressive medications are much more effective.26 In light of our results, we suggest that NETs markers have the potential to contribute to the vascular lesions in patients with BD and targeting NETs may be worthy of interest in patients with BD. Consistent with this hypothesis, and in accordance with histological studies of thrombi in other diseases, our immunopathology analysis showed the presence of NET markers such as MPO and citrullinated histones in vascular lesions and intracardiac thrombi. Indeed, the prothrombotic role of NETs has been initially documented in the venous compartment such as in the deep vein thrombosis disease. However, there is now mounting evidence showing that NETs-associated thrombi are also located in arterial diseases. For example, Mangold et al reported in 111 patients with acute coronary syndrome the presence of NETs in coronary thrombectomies and thrombus NETs burden correlated positively with infarct size.27 Similarly, two different studies reported the presence of NETs in thrombi retrieved during endovascular therapy in patients with acute ischaemic stroke which contribute to the thrombolysis resistance.12 13
NETs have been also implicated in the pathogenesis of many inflammatory diseases. For example, previous studies have shown neutrophils from antineutrophil cytoplasmic antibody-associated vasculitis (AAV) patients release more NETs in vitro. In addition, it has been shown that these patients have not only elevated levels of NETs9 in the circulation but NETs are also present in skin lesions10 and thrombi from AAV patients.28 Similarly, neutrophils from patients with rheumatoid arthritis (RA), were shown to undergo NETosis contributing to the inflammatory response in synovial fibroblasts through IL17A and TNFα.29 30 NETs have been also reported to be a source of autoantigens in patients with RA.31 Neutrophils in gout are associated with the formation of proinflammatory NETs through both autophagy and IL-1β.32 Interestingly, these cytokines are increased in BD and IL17, TNFα and IL-1β play an important role in Behçet-associated inflammation.4 In a recent study, NETs from patients BD have been shown to contribute to the endothelial dysfunction by decreasing cell proliferation and increasing cell apoptosis suggesting a role of NETs beyond thrombosis.33 Interestingly, treatments widely used in BD such as colchicine, anti-TNFα or anti IL-6 were recently found to inhibit in vitro NETs extrusion and further abridged the endothelial dysfunction and the activation of immune cells, thus influencing the global activity of the vascular system. Potential therapeutic targets that can reduce circulating NETs include: ROS scavenger as N-acetylcysteine, PAD inhibitors as Cl-amidine and DNAse have already been used with proven efficacy in systemic lupus erythematous34 35 and RA.36 37
One limitation of this study pertains to the use of samples from some patients who were already receiving colchicine or DMARDs. However, we did not observe any correlation between the use of these medications and the ability of patient’s neutrophils to form NETs. The second limitation is the use of an in vitro model of thrombin generation that may not reflect the complex mechanism of clot formation and vascular injury that may be involved in BD.
Taken together, our results suggest that NETs and markers of NETs are increased in patients with BD and may have the potential to contribute to the vascular manifestations. Further studies are needed to explore the different mechanisms by which NETs collaborate with other cell types and lead to endothelial dysfunction. Identifying the role of NETosis in BD pathogenesis could provide new potential targets for the treatment of this disease and its associated complications.
Our thanks to Louis‐Marie Bobay and David S. Paul for revising the English version of this manuscript.
Handling editor Josef S Smolen
Contributors ALJ, MJP, M-CB, DS and YB were involved in drafting the article. ALJ, YB and DS had full access to all data in the analysis.Study conception and design: ALJ, DS and YB. Acquisition of data: ALJ, RM, SL, NL, AD, AC, PF, FD, TP, PC, DS and YB. Analysis and interpretation of data: ALJ, RM, SL, NL, AD, AC, MJ-P, M-CB, NA, DS and YB.
Funding This work was supported by FDF no 0075823 and DHU FIRE 012 to YB.
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval The study was approved by the CPP (Comité de Protection des Personnes). The study was done in accordance to the ethical guidelines of the 1975 Declaration of Helsinki.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as online supplementary information.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.