Neutrophil Extracellular Traps in Ischemic Stroke Thrombi Are Associated Wth Poor Clinical Outcome

Abstract

Background

Acute ischemic stroke is a neurologic emergency associated with severe disability and death. There is growing evidence that neutrophil extracellular traps (NETs) contribute to the pathogenesis of acute stroke. By mechanical removal of the occluding thrombus from the patient's vasculature, endovascular thrombectomy enables the collection of thrombus material for immunohistologic analysis. The aim of our study was to strengthen the association of NET content in ischemic thrombi with clinical outcome and guide future therapeutics.

Methods

We performed an immunohistologic analysis of thrombi from 101 patients with acute ischemic stroke, focusing on the association between NET content and clinical and interventional indicators.

Results

NETs were present in every patient with acute ischemic stroke. Their abundance in thrombi was associated with interventional markers of thrombus stability. NET‐rich thrombi were associated with unsuccessful recanalization (modified Thrombolysis in Cerebral Infarction <2B) and longer procedure time, and NET abundance in acute ischemic stroke thrombi was associated with outcomes evaluated by patients’ postassessment National Institutes of Health Stroke Scale and modified Rankin Scale scores.

Conclusion

These findings suggest that NET content is critically important to thrombus stability and clinical outcome in acute stroke. They should open new perspectives for innovative immunotherapy strategies based on neutrophil modulation.

Nonstandard Abbreviations and Acronyms

AIS

acute ischemic stroke

mRS

modified Rankin Scale

mTICI

modified Thrombolysis in Cerebral Infarction

NE

neutrophil elastase

NET

neutrophil extracellular trap

NIHSS

National Institutes of Health Stroke Scale

Clinical Perspectives

• What Is New?

• Our findings suggest that neutrophil extracellular trap content is critically important to thrombus stability and clinical outcome in acute stroke, and could be the indicator of a greater inflammatory response favoring the development of ischemic lesions.

• What Are the Clinical Implications?

• These results highlight a potential target area for future therapeutic advancements that promote greater deformation in neutrophil extracellular trap–abundant clots.

After several successful thrombectomy trials in 2015,^{1, 2, 3, 4, 5} thrombectomy has dramatically changed acute ischemic stroke (AIS) care. Nevertheless, substantial reperfusion is not achieved in 12%–56% of cases.⁶ While the reasons for these poor revascularization rates are multifaceted, the nature of the thrombus itself undoubtedly plays a crucial role.⁷ By mechanical removal of the occluding thrombus from the patient's vasculature, endovascular thrombectomy enables the collection of thrombus material for immunohistologic analysis. Better understanding of thrombus composition may help to overcome the current limitations of both pharmacologic and mechanical revascularization therapies.

Neutrophils are the earliest peripheral myeloid cells recruited after cerebral ischemia and can be detected in microvessels within an hour.^{8, 9} We recently reported that a hyperactivation state of circulating neutrophils during AIS is correlated with the clinical severity of acute stroke disease progression.¹⁰ Further activation of neutrophils in inflamed tissues might lead to NETosis, a unique type of cell death, resulting in the formation of neutrophil extracellular traps (NETs)–extracellularly placed nuclear and mitochondrial DNA scaffolds studded with cytotoxic histones and proteases.¹¹ Beyond antimicrobial defense, growing evidence suggests that NETs contribute to the pathogenesis of numerous diseases, including central nervous system diseases, through their excessive formation or impaired removal, which appear toxic to the host.¹² Recent data demonstrate that NETs are the major component of thrombi¹³ and participate in thrombolysis resistance with tissue plasminogen activator,^{14, 15} evidence of NETs’ relevance in thrombus stabilization. However, contradictory results have been reported regarding the potential relationship between NET content in thrombi and clinical outcome. While Ducroux et al reported that quantitative measurement of thrombus NET content (by measuring neutrophil elastase antigen in supernatants of thrombi) was not associated with clinical outcome in AIS,¹⁵ Novotny et al demonstrated the association of increased thrombi NET content with worse neurologic status in patients with stroke.¹⁶ These data highlight the great need to extend clinical immunohistologic analyses of thrombi from patients with AIS to strengthen evidence of NETs’ association with clinical outcome and to guide future therapeutics.

In our study, we performed an immunohistologic analysis of thrombi from 101 patients with AIS, with special emphasis on the role of NETs in regards to clinical and procedural indicators.

Methods

The authors declare that all supporting data are available within the article.

Patients and Thrombi Processing

From February 2019 through October 2020, 101 thrombi were collected from patients with AIS who had undergone endovascular thrombectomy at the neuroradiology department of Pitié‐Salpêtrière Hospital. We included consecutive adult patients, aged >18 years, with AIS meeting the following criteria: (1) clinical indication for mechanical recanalization of an anterior circulation large‐vessel occlusion according to in‐house standard operating procedures and (2) sufficient thrombus material for extensive histologic analyses. We used second‐generation devices (stent retrievers or large‐bore aspiration catheters) for mechanical recanalization in all cases, as previously described.¹⁷ We assessed the following clinical and interventional variables: time between the onset of clinical symptoms and angiographic reperfusion, and time between the first angiographic series and the final angiographic assessment after mechanical recanalization (recanalization time).

We used the modified Thrombolysis in Cerebral Infarction (mTICI) score¹⁸ to evaluate recanalization status. mTICI scores ≥2b are accompanied by faster times to reperfusion and successful reperfusion (≥50%–99% reperfusion), while lower mTICI scores are defined as unsuccessful recanalization.¹⁸ Patient information, including stroke causes, was collected with the international Trial of Org 10172 in Acute Stroke Treatment classification,¹⁹ the National Institutes of Health Stroke Scale (NIHSS) score before endovascular thrombectomy and at discharge, recanalization status, and the modified Rankin Scale (mRS) at discharge. The Ethics Committee of “Ile de France I” approved the study. After retrieval, thrombi were immediately fixed with 4% paraformaldehyde at room temperature, embedded in paraffin within 48 hours.

Immunohistochemical Staining of Thrombi

The paraffin‐embedded thrombi were cut into 5‐µm‐thick sections by using a vibrating tissue slicer (RM2245, Leica Biosystems). Quantification of NET formation required (1) the presence of filamentary structured extracellular DNA stained with DAPI; (2) this DNA originated from cells stained positively for neutrophil elastase a neutrophil marker; and (3) the filamentary structures were decorated with a human histone H3. Section deparaffinization by immersion in xylene and rehydration in decreasing concentrations of ethanol (100°/80°/50°) were performed. After heat‐induced antigen retrieval (citrate buffer, pH 6.12) for 1.5 hours at 50°C and blocking with 5% BSA for 1 hour, the thrombus sections were incubated with a rabbit anti‐human histone H3 (citrulline R2+8+17, citrullinated histone 3) polyclonal antibody (Abcam, Ab5103), and a mouse anti‐human NE polyclonal antibody (R&D, MAB91671), at 4°C overnight. Subsequently, the sections were incubated with Alexa Fluor 488‐conjugated donkey anti‐rabbit IgG antibody (Abcam, Ab150129) and Alexa Fluor 598‐donkey anti‐mouse IgG antibody (Thermo Fisher, No. R37115) or biotinylated secondary anti‐rabbit IgG antibody (Dako) for 1 hour. DNA molecules were stained with DAPI.

Samples were analyzed with an Olympus BX61 Upright Microscope and cellSens software. We counted NETs in at least 4 fields of view with a 40× objective (444.6×277.8 µm) dependent on the size of the thrombus, extrapolating the results to NETs/mm².

Statistical Analysis

The statistical analysis used Prism 9.2.0. (GraphPad Software). All tests were 2‐tailed, with a significance level of α=0.05. Statistical significance was tested with the Mann–Whitney U‐test. Linear partial correlation analysis, adjusted for covariates, identified correlations. A receiver operating characteristic analysis was used to assess the predictive value of NET concentrations for recanalization and clinical outcomes.

Results

The Table presents the characteristics of our patient cohort. The proportion of causes of cerebral infarction observed in patients appears consistent with recent data.²⁰

Table  . Characteristics of Patients With Acute Ischemic Stroke Participating in the Study.

	Patients (N=101)
Sex, male, n (%)	43 (42.5)
Age, y, mean (min–max]	71.7 [19.5–99.4]
NIHSS, median (IQR)	16.4 [1–28]
Causes/TOAST, n (%)*
Thromboembolism	48 (49.5%)
Cardioembolism	22 (22.7%)
Lacunar stroke	0 (0%)
Other specific cause	5 (5.2%)
Undetermined cause	22 (22.7)
Thrombolysis therapy before index intervention (n) [%]	34 (33.6)
Time from symptom onset to thrombectomy, min, mean (min–max)	273.7 (96–1474)

IQR indicates interquartile range; NIHSS, National Institute of Health Stroke Scale; and TOAST, Trial of Org 10172 in Acute Stroke Treatment.

^* Data on causes were available for only 97 patients.

NETs in ischemic stroke thrombi were identified by costaining of citrullinated histone 3 (red) with neutrophil elastase (green) and DNA (blue) in immunofluorescent images. As Figure 1A shows, extracellular DNA colocalized with neutrophil elastase and citrullinated histone 3, indicating that NETs were present in the thrombi from patients with AIS. NETs were present in all stroke thrombi, as previously reported.^{14, 16} The abundance of NETs per unit of area varied greatly between patients, ranging from 5 to 162 NETs/mm² (Figure 1B), with a mean 61 NETs/mm², also as previously reported.¹⁶ NET content did not differ in men compared with women (men, 63.80±5.94; women, 64.59±6.44; P=0.94). We did not observe a significant difference in the number of NETs/mm² by either Trial of Org 10172 in Acute Stroke Treatment classification (Figure 1B) or age (Figure 1C). Notably, NET density was positively correlated with time to last known normal (P=0.0225).

Figure 1.

Presence of NETs in acute ischemic stroke thrombi associated with interventional parameters. A, Representative immunohistochemical illustration of NETs in ischemic stroke thrombi by staining for neutrophil elastase (green), citrullinated histone H3 (red), and DNA (blue). Arrows, nuclei; arrowheads, NET fibers (bars 20 µm). B, Quantification of NET abundance in AIS thrombi by TOAST classification (TOAST‐1: large artery atherosclerosis, TOAST‐2: cardioembolism, and TOAST‐5: stroke of undetermined causes). C, Correlation between NET content in AIS thrombi and patient age D, Quantification of NET content in AIS thrombi by mTICI score. E, ROC curve analysis performed on combined discovery and validation studies to assess the predictive value of NET content with mTICI score. F, Correlation between NET content and procedure time. All measurements came from AIS patients at inclusion. Values are means±SEM. Statistical significance as determined by the nonparametric Mann–Whitney U test is indicated. ^*Significantly different than TICI 3, P<0.05, ^**P<0.01, ^***P<0.001. AIS indicates acute ischemic stroke; AUC, area under the curve; mTICI, modified Thrombolysis in Cerebral Infarction; NET, neutrophil extracellular trap; ROC, receiver operating characteristic; SEM, standard error of the mean; TICI, Thrombolysis in Cerebral Infarction; and TOAST, Trial of Org 10172 in Acute Stroke Treatment.

We next considered whether the abundance of NETs in thrombi was associated with predefined interventional and clinical variables. NET content in thrombi was significantly lower in patients with mTICI 3 (complete reperfusion of the target downstream territory), compared with all other patient groups (Figure 1D). NET content in thrombi was also significantly lower in patients with mTICI 2B (50%–90% reperfusion) and 2C (90%–99% reperfusion) compared with patients with mTICI 2A (partial reperfusion <50%). To address in more detail the extent and character of NET association with TICI scores in patients with stroke, we performed a receiver operating characteristic analysis and defined an appropriate cutoff value (absence of recanalization, mTICI <2B; presence of recanalization, mTICI≥2B) (Figure 1E). NET content was also positively associated with the procedure duration (Figure 1F).

We then investigated whether NET abundance in AIS thrombi might be associated with stroke severity and clinical outcome on the basis of the NIHSS. Patients were dichotomized according to NIHSS score into 2 categories: minor to moderate (NIHSS ≤12) and severe stroke (NIHSS >12). NET content did not differ between severe and minor to moderate strokes, whether judged by initial NIHSS scores (Figure S1A, B). In stark contrast, when we assessed NIHSS at 24‐hour postinclusion or at discharge, NET content was higher in severe than in minor to moderate strokes (Figures S1B and S2A). In addition, higher NIHSS scores after discharge, delta NIHSS (NIHSS at discharge – initial NIHSS) (Figure 2C and Figure S1C) and mRS evaluated at hospital discharge (Figure 2D) were associated with greater NET concentrations. Finally, to further evaluate the ability of NET abundance to discriminate patients by outcome severity on the basis of last mRS, we plotted a receiver operating characteristic curve (Figure 2E). The receiver operating characteristic analysis of these concentrations yielded an optimal threshold (74.5 NETs/mm²) for further dichotomized analysis according to the Youden index with optimal sensitivity and specificity. We then performed an mRS shift analysis and observed a clear shift in all mRS groups toward worse outcomes in patients with higher NET concentrations in AIS thrombi (P<0.0001 for good clinical outcome, mRS score ≤4; Figure 2F). No significant association was found between NET content and hemorrhagic transformation (P=0.95).

Figure 2.

Presence of NETs in ischemic stroke thrombi associated with clinical outcome. A, Quantification of NETs in AIS thrombi according to NIHSS score (NIHSS ≤12: minor to moderate stroke, NIHSS >12: severe stroke at discharge. B and C, Correlation between NET content in AIS thrombi and clinical outcome based on the NIHSS at 24‐hour postinclusion (B) and NIHSS at discharge (C). D, Correlation between NET content in ischemic stroke thrombi and clinical outcome assessed by mRS at discharge. E, ROC curve analysis and definition of an appropriate NET cutoff value in ischemic stroke thrombi for further dichotomized analysis according to the Youden index. F, Clinical effect of NET content in thrombi on mRS shift at discharge in patients with ischemic stroke. AIS indicates acute ischemic stroke; AUC, area under the curve; mRS, modified Rankin scale; NET, neutrophil extracellular trap; NIHSS, National Institutes of Health Stroke Scale; and ROC, receiver operating characteristic.

Interestingly, we also investigated the relationship between NET density in thrombi and the ratio of blood neutrophils to blood lymphocytes calculated before thrombectomy. Ratio of blood neutrophils to blood lymphocytes has recently been reported as a potential novel biomarker of baseline inflammatory process and could predict postthrombolysis early neurologic deterioration in patients with AIS.²¹ We observed a positive correlation between this ratio and the density of NETs (P=0.0008).

Discussion

By mechanically removing the occluding thrombus, usually en bloc, from the patient's vasculature, endovascular thrombectomy enables the collection of thrombus material for immunohistologic analysis.²² Few studies, however, have analyzed the immune cell composition of AIS thrombi, with discordant results regarding NET content and clinical outcomes.^{14, 15, 16} We believe that the use of immunostaining allows for a more accurate quantification of NET density compared with the quantification of markers such as NE using ELISA methods and could explain the discrepancy between precedent studies.

In line with a recent study,¹⁶ we report here the association between the presence of NETs in AIS thrombi and worse neurologic outcome. Specifically, greater procedure time, less complete mTICI reperfusion grade, and worse NIHSS and mRS scores at discharge are associated with higher NET concentrations. In accordance with previous studies,^{14, 15, 16} NETs were present in all thrombi. Interestingly, we observed a correlation between NET content and time to last known normal but no difference in NET content regarding the causes of stroke. Clot formation and platelet activation trigger an inflammatory response and promote the infiltration of leukocytes such as neutrophils.²³ We therefore believe that neutrophils and NETs accumulate in the thrombi over time regardless of the causes. A recent study demonstrated a common structural feature of thrombi across various causes with an outer shell that show a decreased susceptibility to tissue plasminogen activator thrombolysis, and Ducroux et al¹⁵ observed that NETs were preferentially found in the outer layers of the thrombi. Therefore, we postulate that neutrophils are not intrinsic to the thrombus and that this outer shell develops secondarily to the thrombus formation and contains NETs that participate in the resistance to thrombolysis. The lack of difference in NET content according to the Trial of Org 10172 in Acute Stroke Treatment classification might also be related to neutrophils that accumulate over time secondarily in the circulation and release NETs regardless of causes and the site of thrombus formation.

The higher NET content in ischemic thrombi may reflect a greater inflammatory response. Studies conducted in mice models of stroke have highlighted the role of neutrophils in stroke. In particular, intercellular adhesion molecule‐1 knockout mice²⁴ display reduced ischemic lesions after permanent or transient middle cerebral artery occlusion, and depletion of granulocytes is associated with a better preservation of microcirculation after focal ischemia, highlighting the prominent role of neutrophils in the pathophysiology of stroke. Furthermore, NETs have been reported to give rise to the disruption of blood‐brain barrier integrity and the increase of its permeability, microcirculatory disturbances, vascular leakage, thrombosis, release of proinflammatory cytokines, oxidative stress, neuronal injury, and death as well as neuroinflammation in ischemic stroke. Netting neutrophils have the capacity to actively participate in these cellular and molecular cascades, leading to inflammation and cell death by releasing metalloproteinases, proteases, cytokines, extracellular histones, DNA, and ROS.²⁵ Interestingly, we demonstrate here a link between NET content in thrombi and the neutrophil/lymphocyte ratio, which has been recently reported as a potential novel biomarker of baseline inflammatory process and could predict postthrombectomy early neurologic deterioration in patients with AIS.²¹ We therefore believe that beyond the negative effect on thrombectomy by modifying the mechanical properties of thrombi, NET density could be the indicator of a greater inflammatory response favoring the development of ischemic lesions.

We recently reported that the harmful neutrophil senescent subset, correlated with the clinical severity of acute stroke disease progression, expands in AIS.¹⁰ As senescent neutrophils in the circulation have also been reported to migrate rapidly to the inflammatory site²⁶ and produce NETs at a higher rate than the total circulating pool,²⁷ we speculate that the expansion of the neutrophil senescent subset observed in patients with AIS might result in enhanced migration of these hyperactive cells in the brain, explaining at least in part the high NET content in AIS thrombi. Taken together, these data suggest that the higher density of NETs in patients with a poor prognosis could therefore be a sign of a greater inflammatory response, allowing more rapid recruitment of neutrophils, especially senescent neutrophils, to the thrombus and the ischemic and penumbral zone, leading to a greater extent of lesions. This would imply a thrombus containing more NETs and therefore presenting mechanical properties favoring its stability, making thrombolysis and endovascular thrombectomy less effective, and the production of several cytokines and other inflammatory factors by neutrophils. Parallel analyses of thrombi and blood samples would allow investigation of the relationship between functional parameters of circulating neutrophils, intravascular NETosis, and NET abundance in AIS thrombi. This could allow the identification of blood biomarkers of NET abundance in AIS thrombi and pave the way to adapt patient management.

The cellular mechanisms of NET‐mediated thrombosis still need full elucidation. NETs have thrombogenic effects in vitro and have been observed in thrombi in other diseases.^{28, 29} The DNA in NETs can activate clotting factor XII, initiating the contact coagulation pathway³⁰; in turn, activated factor XIIa can activate neutrophils and enhance NET formation.³¹ Some studies hypothesize that NETs help clots form and resist lysis by acting as a scaffold.³² Platelets populate thrombi, and we know that activated platelets can express P‐selectin, binding P‐selectin glycoprotein ligand‐1 receptor on neutrophils and producing further integrin activation and NET formation.³³ Finally, NETs can activate complement.²⁸

The strength of our findings is tempered by some limitations. First, this was a single‐center study. Furthermore, neurologic outcome was evaluated at patient discharge. The mRS has been found to be particularly valuable when evaluated at 90 days after stroke onset,^{34, 35} and a recent study has highlighted that while mRS tends to be stable during the first month, changes in mRS between 30 and 90 days are frequent.³⁶ However, long‐term follow‐up is difficult to obtain, and discharge mRS scores are often employed as a surrogate for long‐term outcomes. Finally, although we observed an association between the richness of NETs in AIS thrombi and worse neurologic outcome, this NET richness can be determined only after thrombectomy, giving access to the thrombus, and is therefore available only late in the management and only for selected patients.

Conclusions

In conclusion, our study confirms the association between thrombus NET content and both poor clinical outcomes and markers of thrombus stability. These results highlight a potential target area for future therapeutic advancements that promote greater deformation in NET‐abundant clots.

Author Contributions

Carole Elbim and Arnaud Lapostolle conceived the study and designed the experiments. Carole Elbim supervised the project. Arnaud Lapostolle, Chloé Loyer, and Thomas Chaigneau performed experiments. Arnaud Lapostolle and Carole Elbim analyzed data. Arnaud Lapostolle, Chloé Loyer, Mahmoud Elhorany, Thomas Chaigneau, Franck Bielle, Sonia Alamowitch, Frédéric Clarençon, and Carole Elbim interpreted results of experiments. Mahmoud Elhorany and Frédéric Clarençon were involved in patient recruitment and characterization. Arnaud Lapostolle and Carole Elbim prepared the figures. Arnaud Lapostolle and Carole Elbim drafted the manuscript. Arnaud Lapostolle, Chloé Loyer, Mahmoud Elhorany, Thomas Chaigneau, Franck Bielle, Sonia Alamowitch, Frédéric Clarençon, and Carole Elbim edited, reviewed, and approved the final version of the manuscript.

Sources of Funding

Arnaud Lapostolle received a grant from the Société Française NeuroVasculaire for his work in this project.

Disclosures

None.

Supplementary Material

Figure S1: (A, B) Quantification of NETs in AIS thrombi according to initial NIHSS score (A) and NIHSS at 24 h post‐inclusion (B) (NHISS ≤12: minor‐to‐moderate stroke, NHISS >12: severe stroke) (C) Correlation between NET content in AIS thrombi and clinical outcome based on the initial NIHSS at discharge (D) Correlation between NET content in IS thrombi and clinical outcome based on delta NIHSS (NIHSS at discharge initial NIHSS)