Neutrophil extracellular traps form predominantly during the organizing stage of human venous thromboembolism development

Abstract

Background

A growing health problem, venous thromboembolism (VTE), including pulmonary embolism (PE) and deep vein thrombosis (DVT), requires refined diagnostic and therapeutic approaches. Neutrophils contribute to thrombus initiation and development in experimental DVT. Recent animal studies recognized neutrophil extracellular traps (NETs) as an important scaffold supporting thrombus stability. However, the hypothesis that human venous thrombi involve NETs has not undergone rigorous testing.

Objective

To explore the cellular composition and the presence of NETs within human venous thrombi at different stages of development.

Patients and Methods

We examined sixteen thrombi obtained from 11 patients during surgery or at autopsy using histomorphological, immunohistochemical and immunofluorescence analyses.

Results

We classified thrombus regions as unorganized, organizing, and organized according to their morphological characteristics. We then evaluated them focusing on neutrophil and platelet deposition as well as micro-vascularization of the thrombus body. We observed evidence of NET accumulation, including the presence of citrullinated histone H3 (H3Cit)-positive cells. NETs, defined as extracellular diffuse H3Cit areas associated with myeloperoxidase and DNA, localized predominantly during the phase of organization in human venous thrombi.

Conclusions

NETs are present in organizing thrombi in patients with VTE. They are associated with thrombus maturation in humans. Dissolution of NETs might thus facilitate thrombolysis. This finding provides new insights into the clinical development and pathology of thrombosis and provides new perspectives for therapeutic advances.

Introduction

Venous thromboembolism (VTE), which includes both deep vein thrombosis (DVT) and pulmonary embolism (PE), is the most common cardiovascular disease after myocardial infarction (MI) and ischemic stroke [1]. In animals, neutrophils contribute importantly to the pathogenesis of venous thrombosis [2–5]. Neutrophil-related pro-coagulant mechanisms might contribute to large-vessel thrombosis in various types of cardiovascular diseases [6]. Our group has previously reported that neutrophil extracellular traps (NETs) provide a scaffold for experimental thrombus formation by binding erythrocytes and platelets [7]. NETs consist of released nuclear DNA containing histones and neutrophil granule proteins, such as elastase, myeloperoxidase (MPO), and cathepsin G [8, 9]. NETs localize in experimental DVT in baboons [7] and extracellular chromatin comprises part of the structure of venous thrombi in mice [10, 11]. Treatment with deoxyribonuclease I (DNase I), an enzyme that degrades DNA, protects mice from flow restriction-induced DVT [10, 11]. Our recent experimental study indicates that neutrophil chromatin decondensation mediated by peptidylarginine deiminase 4 (PAD4), an enzyme required for NET formation (NETosis) [12], participates crucially in venous thrombosis in mice [13], thus emphasizing the impact of NETs on DVT. Increased levels of circulating nucleosomes and markers of neutrophil activation associate with a 3-fold higher risk of DVT in humans [14]. An elevated plasma DNA concentration correlates with biomarkers of DVT [15] and with a prothrombotic state in patients with severe atherosclerosis [16].

Evolution of thrombus generally proceeds through three distinct phases [17]. Initially, thrombi are unorganized. If formed under conditions of continued blood flow, roughly alternating layers of platelets and red cells generate laminae, denoted lines of Zahn; if formed in a static environment, the thrombus has a more uniform appearance. After a few days, the thrombus begins to organize, with inflammatory cells (initially neutrophils and later lymphocytes and monocytes) infiltrating the edges of the thrombus in association with fibroblasts, collagen deposition, and neovascularization. Over weeks or more, an organized thrombus assumes a more-ordered fibrotic appearance with associated hemosiderin-laden macrophages, but an otherwise sparse inflammatory infiltrate. The times suggested are approximate minimums, and should not be taken as absolutes; the progress of thrombus organization can vary significantly. Numerous factors, including host immune status, degree of residual blood flow at the thrombus site, and medication use, can all affect the process to varying degrees. Additionally, at any given time, a single thrombus will often display multiple stages of development; typically, the periphery adjacent to a vessel wall exhibits the greatest organization, with the center of the thrombus showing the least.

Few reports have sought NETs in thrombi from humans. Extensive histone citrullination was observed in the thrombus of a patient with microscopic polyangiitis complicated by pneumonia and DVT [18]. Also, coronary thrombi removed from patients with acute MI contain NETs [19], as do intraluminal thrombi from abdominal aortic aneurysm patients [20]. Histone proteins H2A/H2B with DNA are found in human arterial thrombi removed from acutely (age 1–3 days) thrombosed grafts and in chronic (more than 1 year) aortic hematoma samples [21].

NETs furnish an important new link between innate immunity, inflammation, hemostasis and thrombosis [2, 3, 22, 23]. Their recognition provides an opportunity to develop biomarker(s) based on NET components for human VTE diagnosis and to design new therapeutic approaches to improve the efficacy of thrombolysis. The current study aimed to establish the presence of NETs during the different stages of thrombus development in humans with VTE.

Materials and Methods

Sample acquisition

Specimens were identified through a search of the electronic records of the Department of Pathology and Laboratory Medicine at Brigham and Women’s Hospital by a cardiovascular pathologist who had full access to the origin of the samples and selected the specimens based on the prespecified inclusion criteria below. Search parameters included any one of the text strings containing “deep vein thrombosis,” “deep vein thrombus,” “pulmonary embolus,” or “pulmonary embolism,” being identified as an embolectomy specimen during accessioning, or being identified as a leg amputation and containing the text “thrombus” or “thrombosis.” Reports for cases meeting these criteria were manually reviewed to identify cases where deep vein thrombi, inferior vena cava (IVC) filter clots, and/or pulmonary emboli were sampled and processed as formalin-fixed paraffin-embedded (FFPE) specimens. Those specimens meeting such criteria were then retrieved from the archives of the Department, and de-identified sections were generated for study. These procedures were reviewed and approved by the Human Research Committee (HRC) of Partners HealthCare, and conform to current national and international standards for ethical conduct of human subjects research.

Histology

FFPE tissue sections cut at 4 μm thickness were stained with either hematoxylin and eosin (H&E) or Masson’s trichrome. A cardiovascular pathologist categorized the specimens using histomorphologic features as unorganized (lack of collagen deposition, interspersed regions of fibrin/platelets and red blood cells), organizing (poorly organized collagen and/or neovasculature, and/or granulation tissue-like appearance), or organized (well organized collagen with intermixed hemosiderin-laden macrophages, with or without a well-formed neovasculature) [17].

Immunohistochemistry

Following deparaffinization, microwave antigen retrieval used citrate buffer (pH 6.7) for 8 min, followed by the incubation of slides in heated buffer for an additional 20 min. After endogenous peroxidase activity blocking in methanol/H₂O₂ (100:1) for 30 min at room temperature, single immunohistochemical staining used an anti-H3Cit antibody (ab5103, 1:250 dilution, Abcam, Cambridge, MA, USA). Histofine Simple Stain MAX PO (MULTI) (414152F, ready to use, Nichirei Corporation, Tokyo, Japan) served as a secondary antibody. Diaminobenzidine (DAB) substrate kit (415192F, Nichirei) enabled visualization of staining by H3Cit antibody. Finally, sections were counterstained with hematoxylin. Isotype-matched IgG was used instead of primary antibody as a negative control of the staining. Because of differences observed in the amount of H3Cit staining, we performed a blinded semi-quantitative analysis using an H3Cit scoring system based on the proportion of H3Cit-positive cells across an entire sample. The regions with a majority of H3Cit-positive staining among the entire population of cells (observed with counterstaining) were defined as positive (+, or high H3Cit score). Those cases where H3Cit-positive signals were observed in less than half of the region’s cells were classified as (±), and areas without or with very few H3Cit-positive cells were scored as negative (−). All specimens were characterized blindly by a co-investigator who did not have knowledge of the nature of the specimens at the time of scoring.

A rabbit monoclonal anti-CD11b (ab52478, 1:250 dilution, Abcam) served to identify leukocytes and mouse monoclonal anti-CD42b (ab74449, ready to use, Abcam) to visualize platelets. Also, the combination of anti-CD11b and mouse monoclonal anti-MPO (ab24467, 1:250 dilution, Abcam) antibodies was used to specify the origin of CD11b-positive cells. Histofine Simple Stain MAX PO (MULTI) was used as a secondary antibody for anti-CD11b, the first primary agent. Alkaline Phosphatase (AP)-goat anti-mouse IgG (H+L) (No. 626522, 1:500 dilution, Invitrogen, Carlsbad, CA, USA) was applied as a secondary antibody for anti-MPO, the second primary agent. DAB substrate kit (SK-4100, Vector Laboratories, Burlingame, CA, USA) was used for visualization of double immunohistochemical staining. Finally, sections were counterstained with Nuclear Fast Red (Sigma-Aldrich, St. Louis, MO, USA). Images were captured using Axiovision software and a Zeiss Axioplan microscope coupled with a Zeiss AxioCam HRc camera. Individual images were manually aligned to generate composite images of entire thrombus sections.

Immunofluorescence analysis

Antigen retrieval was performed after deparaffinization. The slides were briefly boiled in sodium citrate buffer (10 mM, pH 6.0) in a microwave and then incubated in heated buffer for an additional 15 min. The slides were washed twice in PBS and then permeabilized with 0.1% Triton X-100 for 10 min on ice. After blocking with 3% bovine serum albumin for 1 h at 37°C or 10% normal goat serum for 2 h at room temperature, slides were incubated overnight in antibody dilution buffer (0.3% BSA 0.05% Tween-20) containing one of the following: rabbit polyclonal anti-citrullinated histone H3 (ab5103, 0.5 μg/mL, Abcam), mouse monoclonal anti-myeloperoxidase (MPO) (ab25989, 0.67 μg/mL, Abcam), mouse monoclonal anti-peptidylarginine deiminase 4 (PAD4) (ab128086, 4 μg/mL, Abcam), sheep polyclonal anti-von Willebrand factor (VWF) (ab11713, IgG fraction 1:250 dilution, Abcam), rabbit polyclonal anti-citrulline (panCit) (ab6464, whole antiserum, 1:1500 dilution, Abcam) or mouse monoclonal anti-CD31 (ab9498, 2 μg/mL, Abcam). For each staining, an IgG used at the same concentrations as the primary antibody served as the negative control. After thorough washing with PBS, slides were incubated in antibody dilution buffer containing 1.5 μg/mL Alexa488-conjugated rabbit IgG, 1.5 μg/mL Alexa568-conjugated sheep IgG, or 1.5 μg/mL Alexa555-conjugated mouse IgG (all from Invitrogen) for 2 h at room temperature. After thorough washing, DNA was counterstained using 1 μg/mL Hoechst 33342 (Invitrogen) and cover-slipped using Fluorogel mounting medium. Images were acquired using a Zeiss Axiovert widefield fluorescence microscope in conjunction with a Zeiss MRm monochromatic CCD camera and Axiovision software.

Results

Clinical specimen characteristics

This analysis is based on sixteen thrombi collected from eleven individuals (mean age 52.0 ± 17.9 years; three men and eight women, Table 1.) Six thrombus samples derived from three autopsy cases (all within 24 hours of death), and ten from eight surgical cases. Clinical diagnoses included: chronic thromboembolic pulmonary hypertension (n=4), recurrent venous thromboembolism (n=2), cancer (1 case of metastatic endometrial adenocarcinoma and 1 of lung adenocarcinoma) and one case each of hernia, aortic valve replacement and status post lung transplant. Six thrombus specimens were obtained from three patients with PE and DVT; from another two patients, PE and IVC clot samples (n=4) were collected. Four thrombi derived from patients with PE as a main diagnosis and two from patients with either DVT or IVC clot.

Table 1.

Patient clinical data and thrombus samples included in the study

No. of patient		Age	Gender	Diagnosis	Tissue
Patient 1	surgical	36	F	CTEPH	PE
					IVC clot
Patient 2		47	F	CTEPH	PE
Patient 3		72	F	metastatic endometrial adenocarcinoma	DVT
Patient 4		69	F	s/p lung transplant for UIP	PE
					IVC clot
Patient 5		16	M	CTEPH	PE
Patient 6		60	M	s/p hernia repair	PE
Patient 7		44	F	CTEPH	PE
Patient 8		53	M	idiopathic recurrent VTE	IVC clot
Patient 9	autopsy	74	F	s/p aortic valve replacement	PE
					DVT
Patient 10		38	F	obesity, recurrent VTE	DVT
					PE
Patient 11		63	F	lung adenocarcinoma	PE
					DVT

Abbreviations: CTEPH, chronic thromboembolic pulmonary hypertension; s/p, status post; UIP, usual interstitial pneumonia; VTE, venous thromboembolism

According to the morphologic criteria described above, based on the results of hemotoxylin and eosin and of trichrome staining, thrombus regions fell into three major stages of thrombus development [17]: unorganized (16 regions from 11 patients), organizing (13 regions from 11 patients) and organized (8 regions from 7 patients). We then used immunohistochemical analysis to determine the distribution of platelets and leukocytes in the human thrombi during each stage of maturation, as well as the presence of H3Cit as a marker of NETosis. At least two regions corresponding to different stages of thrombus organization were identified in 15 out of 16 specimen samples.

Unorganized thrombi harbor rare H3Cit-positive cells, while these cells and NETs abound in the organizing stage of thrombus development

Fig. 1 shows a representative thrombus from a PE patient with both unorganized (square box 2) and organizing (square box 1) regions. We observed platelet- and fibrin/erythrocyte-rich zones, defined by the presence of lines of Zahn (white arrows, Fig. 1A). The organizing part of a thrombus mainly contained leukocytes positive for CD11b (brown, Fig. 1A). The majority of CD11b-positive cells (brown) were identified as neutrophils by double staining with MPO (blue), showing that most CD11b-expressing cells are also MPO positive (dark blue). Prominent H3Cit-positive staining (brown, Fig. 1A, last panel) localized only within the organizing, but not the unorganized, part of thrombus. This observation suggests that NETosis occurs in human thrombi during the organizing stage of maturation.

Fig. 1.

Evidence of prominent NETosis observed in the organizing, but not unorganized, portions of a pulmonary embolism thrombus. Representative composite images from analysis of a pulmonary embolism (PE) sample obtained from autopsy. Consecutive sections of the thrombus were stained. (A) Trichrome stain revealed organizing (square 1) and unorganized (square 2) regions of the thrombus and the presence of lines of Zahn (white arrows). Double immunohistochemical staining for CD11b (brown, leukocytes) and CD42b (blue, platelets) showed both platelets and leukocytes within the organizing stage of the thrombus (oval 1), whereas the unorganized part contained mostly platelets (square 2). H3Cit-positive cells (brown) were observed mostly in the CD11b/MPO-double positive (dark blue) region of the organizing area. Higher magnification images from the organizing part of a thrombus (indicated as oval 1) are shown in B and C. (B) Punctiform (intracellular) and diffuse (extracellular) H3Cit staining (brown) was observed together with MPO (dark blue) within platelet/neutrophil-rich zones. Negative control staining was performed using isotype-matched IgG. Images in A and B are composites of photographs of stained sections of the entire thrombus (A) and its organizing portion (B). (C) Upper panel confirmed the close localization of the intra- (green arrow) and extracellular (green asterisk) H3Cit with MPO (single channels and merge are shown). Platelet islands stained for VWF (red), and diffuse VWF (white asterisk) was present in areas containing H3Cit. A positive staining pattern for PAD4 (red) was seen in regions with neutrophil infiltration. Citrullinated proteins were detected using anti-panCit antibody (green). IgG isotype control served as a negative control. Sections were counterstained for DNA as indicated (Hoechst 33342, blue). Scale bars, 500 μm (A), 100 μm (B), 50 μm (C).

To clarify the composition of the human thrombus samples observed in this study, Fig. 1B presents larger images of the organizing region at higher magnification (areas indicated as oval 1 in Fig. 1A). We also performed immunofluorescence staining (Fig. 1C). We observed both intracellular (H3Cit-positive nuclei, green arrow in Fig. 1C) and diffuse extracellular H3Cit-positive staining patterns (green asterisk in Fig. 1C), suggesting ongoing NETosis and the prominent presence of NETs in this inflammatory stage of thrombus development.

We have previously shown that NETs bind VWF [7] and that VWF codistributes with NET biomarkers in thrombi from baboons and mice [7, 10]. Here we found platelets stained strongly positive for VWF. In addition, we observed regions with diffuse VWF staining (white asterisk) associated with H3Cit extracellular staining (Fig. 1C, lower left panel), indicating the interaction of VWF with NETs. Single channels showing VWF (diffuse VWF staining is indicated by white asterisk) and H3Cit expression pattern are presented in Fig. S1. We also stained thrombus samples for PAD4, the enzyme required for NETosis [12, 24, 25], and found PAD4-positive cells in corresponding areas of neutrophil infiltration and NETosis (Fig. 1C, lower middle panel). Also, panCit antibody, which recognizes citrulline modifications on proteins, visualized an intracellular and a diffuse extracellular staining pattern (Fig. 1C, lower middle panel), again indicating that protein citrullination occurs in human thrombi. IgG-isotype control staining showed no positive signal by either immunohistochemical (Fig. 1B, rightmost panel) or immunofluorescence (Fig. 1C, second row, rightmost right panel) analyses.

We also observed the presence of lines of Zahn (Fig. 2A) with platelet aggregates (Fig. 2B) and neutrophil infiltrates (Fig. 2B and 2C) in an organizing deep vein thrombus. Of interest, H3Cit-positive cells (Fig. 2D and 2E) originating from neutrophils (Fig. 2F) localized predominantly in the immediate vicinity of VWF-positive (Fig. 2H) platelet-rich zones (Fig. 2E and 2H), suggesting a possible involvement of NETs in the process of platelet recruitment during human thrombus organization. We further found PAD4 and panCit staining in these areas (Fig. 2G), whereas isotype-control staining displayed no specific signal (Fig. 2I). Thus, the intensity of H3Cit staining was varied throughout the different stages of thrombus development.

Fig. 2.

Platelet-rich zones are often surrounded by MPO and H3Cit-positive staining in the organizing deep vein thrombus. Representative images from a DVT sample collected during autopsy. (A–C) Thrombus was classified as organizing based on the presence of minimal collagen deposition, early neovascularization, and/or granulation-like tissue (A); CD42b-positive platelet aggregates (blue, B); and infiltrating CD11b-positive neutrophils (brown, B), co-stained with MPO (dark blue, C). (D) H3Cit-positive cells were detected in neutrophil-rich zones. Images A–D are composites of photographs of stained sections of the entire thrombus. The area indicated by the blue square is shown at a higher magnification in E and F and the area indicated by the orange square in G, H and I. Platelet-rich zones stained for VWF (red) and were surrounded by H3Cit-positive cells (H). PAD4 (red) and panCit (green) were seen primarily in neutrophil-rich areas surrounding platelet islands (G). Negative control staining was performed using isotype-matched non-specific IgG (I). Scale bars, 500 μm (A–D), 50 μm (E and F), 20 μm (G–I).

Organized thrombi contain multiple microvessels while NETs are rare at this stage

Organized thrombi contained microvessels, observed by trichrome staining (Fig. 3A, blue square is magnified in Fig. 3C) and confirmed by immunofluorescence analysis using an anti-CD31 antibody (Fig. 3B). The negative control staining using isotype-matched IgG showed no specific signal. A few microvessels were observed in only one case of an organizing thrombus surgically removed from a PE patient (Fig. S2).

Fig. 3.

Accumulation of microvessels within the organized region of a pulmonary embolism thrombus. (A,C) Trichrome stain indicated organized regions of a thrombus with micro-vascularization (area within blue square is magnified in C). (B) CD31-positive vascular endothelial cells were found in the area with newly formed blood vessels. Negative staining with non-specific IgG showed no specific pattern and nuclei are observed as distinct round shapes in this cell-rich region stained for DNA. Image A and C are composites of photographs of a stained section. Scale bars, 500 μm (A), 100 μm (B), 50 μm (C).

We observed areas of collagen-rich connective tissue in organized regions of thrombi (Fig. 4A, blue square box indicates a representative region), which were leukocyte deficient (Fig. 4B–C and Fig. 4F) with only a few neutrophils found in the thrombus body (Fig. 4C). Both immunohistochemical (Fig. 4D and 4E) and immunofluorescence (Fig. 4F and 4G) analyses showed the lack of prominent H3Cit-positive staining, with many completely negative sections, indicating that NETosis does not characterize mature regions of organized thrombi. Control staining with isotype-matched IgG showed no positive signal (Fig. 4H).

Fig. 4.

H3Cit is rare in the organized regions of thrombi. Analyses of a surgically removed IVC clot. (A) Collagen-rich connective tissue (blue) was present within the organized part of the thrombus and is indicated with a blue square. (B–C) Cell-specific immunohistochemical analyses revealed few platelets and neutrophils within the entire thrombus body and their absence in the organized part of thrombus. (D) Immunohistochemical staining for H3Cit antibody (brown) displayed a negative reactivity (area indicated as a blue square). Images A – D are composites of photographs of stained sections of the entire thrombus. (E–H) High magnification images of the organized part from the area indicated with a blue square in A–D. No staining was detected using anti-H3Cit (E–G), MPO (F) and VWF (G) antibodies in this thrombus. Negative control using IgG isotype showed a non-specific staining pattern (H). Scale bars, 500 μm (A–D), 20 μm (E).

Quantification confirms that excessive H3Cit staining was observed only in the organizing stage of thrombus development

Semi-quantitative analysis of H3Cit-positive cells was performed in a blinded fashion. We developed an H3Cit scoring system based on the frequency of positive staining for H3Cit (Material and Methods). This scoring system classified the specimens with predominant H3Cit-positive cells as H3Cit score (+), H3Cit score (±) for specimens with less than half H3Cit-positive cells, and negative H3Cit score (−) for specimens with none or very minor positive staining for anti-H3Cit antibody.

Only organizing thrombi received a high H3Cit (+) score (Fig. 5A and 5B). Thrombus samples from PE and DVT patients revealed a similar distribution of H3Cit scores, whereas no IVC clot samples had a high H3Cit score (Fig. 5A). Of interest, an H3Cit (+) score was seen in 54% of all organizing thrombi and in the remaining 46% a score of (±) (Fig. 5B). No H3Cit (−) cases were identified in regions of early organization, whereas 56% of unorganized and 50% of organized thrombi had a (−) H3Cit score (Fig. 5B). These data showed that NETosis occurs almost exclusively within the organizing stage of thrombus development.

Fig. 5.

Summary of clinical, histomorphological and immunological analyses performed for all thrombus samples. (A) Table summarizes clinical data from patients included in this study as well as results of histopathological validation. H3Cit score is presented for each of the cases. H3Cit scoring system: (+, highlighted in blue) strongly positive with majority of H3Cit-positive cells throughout the entire thrombus body; (±) focally positive with a few H3Cit-positive cells in the population; (−) negative with mostly H3Cit-negative cells. (B) Graph shows the distribution of H3Cit score (percentage and number (n) of corresponding thrombus regions) between the different stages of thrombus organization.

Discussion

This study evaluated the composition of venous thromboemboli obtained either surgically or from autopsy specimens. The results show accumulation of NETs during the organizing stage of pathological venous thrombus development and thus support the participation of NETs in the organization of human thrombus.

Histomorphological and immunological assays using cell marker-specific antibodies showed differences in platelet and neutrophil deposition as well as micro-vascularization within the unorganized, organizing and organized stages of thrombus development. These findings suggest that immunohistological analysis could become a useful adjunct for the pathological description of human thrombi.

The process of leukocyte infiltration and monocyte recruitment during thrombus formation and the impact of NETs have received previous attention in experimental DVT [10, 11, 13, 26]. This study shows evidence of ongoing NETosis in the organizing stage of human thrombus development including both intra- and extracellular H3Cit closely associated with MPO, as well as extracellular DNA in H3Cit-rich areas. Thrombus composition did not differ between patients with PE and DVT. Thus, NETs appear to form as part of the inflammatory response and organization in human thrombus development.

NETs’ DNA and histones provide a scaffold for thrombi by binding RBCs and promoting platelet aggregation in vitro and in mice [6, 7]. The results of this study suggest that this concept could apply to human thrombogenesis as well. Murine venous thrombi contain H3Cit, a marker of NETs, co-distributing with VWF predominantly in the red blood cell (RBC)-rich portion of the thrombus [10], and VWF-platelet interaction contributes substantially to VTE development in animals [27, 28]. Plasma VWF secreted from activated vascular endothelium is induced by histone infusion, which also promotes DVT in mice [10]. This finding could result from the enhancement of thrombin generation by extracellular histones that occurs via a platelet-dependent mechanism [29, 30]. We observed that H3Cit-positive cells surrounded VWF-positive platelet islands in human DVT samples and that diffuse VWF-positive staining was associated with diffuse H3Cit patterns in the organizing regions of thrombi. This observation supports the notion that NETs together with VWF could enhance thrombus formation/stability in humans.

Of interest, we found that diffuse extracellular H3Cit-positive areas in the organizing part of thrombi also contained PAD4, an enzyme required for chromatin decondensation during NETosis. PAD4-deficient mice, which cannot produce NETs [12], exhibit protection from DVT [13] and MI [31]. Thus, these data show NETting neutrophils in human venous thrombosis and identify PAD4 as a potential target for future drug development.

Targeting NET components may also provide a foundation for molecular imaging to identify patients with thromboembolic disease that may respond optimally to thrombolytic therapy. Molecular imaging techniques using magnetic resonance and radionuclide scanning have been shown to distinguish fresh from chronic thrombus but have not yet reached widespread clinical use [32, 33].

Fibrin clots containing DNA and histones have prolonged resistance to fibrinolysis that is reversed by DNase [7, 34]. Our observations support targeting NETs, in addition to fibrin, as an approach to facilitate thrombolytic therapy in patients. DNase I has decreased the incidence of thrombosis in experimental DVT in mice [10, 11]. Furthermore, we have recently reported that DNase I also has an anti-inflammatory effect when administered to mice with experimental myocardial infarction [31]. The anticoagulant heparin can also dissociate NETs [7] and prevent histone interactions with platelets, thus protecting mice from histone-induced thrombocytopenia [35]. Aspirin, a known anti-thrombotic, can inhibit NETosis in vitro [36] as well as reduce the incidence of recurrent VTE [37, 38]. The combined approach of heparin or aspirin with DNase I or PAD4 inhibitors might enhance venous thrombus prevention. Co-administration of DNase I with a fibrinolytic might facilitate thrombolysis, particularly during the organizing stage of thrombus development.

In conclusion, this study demonstrates the presence of NETs during the process of thrombus organization in human VTE. NETs can provide a scaffold for human thrombus stability during its maturation and could become a diagnostic or therapeutic target to improve the effectiveness of thrombolytic therapy in patients with thromboembolic diseases.

Supplementary Material

Supp FigureS1-S2

Fig. S1. High magnification images from the organizing region (indicated as oval 1 in Figure 1A) of a pulmonary embolism thrombus documents the expression pattern for VWF and H3Cit. Left panel shows the bright VWF staining for platelet-rich zones and diffuse staining (white asterisk) was observed in areas with H3Cit-positive cells and H3Cit diffuse staining (single channel, middle panel). Right panel displays the merged image. Scale bars, 50 μm.

Fig. S2. Evidence of blood vessel formation within an organizing part of a pulmonary embolism thrombus. CD31-positive microvessels were observed only within one organizing region (region 1) and were not detected in any other organizing portions of thrombi. Left image is a composite of photographs of a stained section of the entire thrombus. Anti-CD31 antibody also stained platelet islands (region 2). Negative control staining was performed using non-specific IgG isotype control. Scale bars, 500 μm (trichrome stain), 50 μm (immunofluorescence analyses).