We prospectively recruited 166 patients who underwent MT for acute ischemic stroke with large vessel occlusion (LVO) at Hannover Medical School during two time periods between March 2018 and August 2019 (n = 92) (cohort 1) and between June 2020 and May 2021 (n = 74) (cohort 2). Additional criteria required for inclusion was the successful retrieval of a blood sample prior to MT. In cohort 1, samples were taken before thrombectomy by a venous blood sampling. Samples of cohort 1 were acquired during a former study by our group, as reported previously [13]. In cohort 2, blood was taken via the arterial groin puncture used for thrombectomy, immediately before the procedure. This was implemented to further streamline study procedures. We excluded all patients with known malignoma, immunodeficiency or ongoing medication with immunomodulatory drugs. Stroke thrombi were retrieved during MT. After seven days, venous blood samples were collected for both cohorts. An overview of the recruitment process, as well as available material for the different analyses, can be found in supplemental figure SFig. 1.

Clinical data were gathered at inclusion and seven days after initial treatment, using a standardized electronic case report form. The collected data include demographic data, basic clinical parameters, vascular risk factors evaluated via the Essen Stroke Risk Score (ESRS) [15] and stroke severity at baseline determined by the National Institutes of Health Stroke Scale (NIHSS) obtained in the emergency department. Additionally, the 7 day follow-up included the occurrence of stroke associated infections (SAI) according to Centers for Disease Control and Prevention’s (CDC) Criteria [16]. The SAI status of patients who received antibiotic therapy without meeting the CDC criteria was labeled as indeterminate. Cerebral reperfusion was evaluated according to the modified Treatment in Cerebral Infarction (mTICI) score, graded by board certified neuroradiologists. An mTICI score of 2c or 3 in the anterior circulation or of 2b or higher in the posterior circulation was considered as sufficient reperfusion [17]. Both status of infection and mTICI were rated by two observers separately, before systematically agreeing on a consensus. The investigators grading the status of infection and the mTICI score were blinded to the patients’ biomarker data. Thrombus origin was classified at discharge according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria [18].

Finally, for cohort 2 we scored the patients’ outcome with the modified Rankin Scale (mRS) [19] after the end of rehabilitation, based on the documentation of the respective inpatient rehabilitation centers. A mRS score of 2 or lower was considered as a favorable functional outcome.

The stroke thrombi were analyzed in batches of 10 thrombi at a time. After retrieval, the thrombi were immediately fixed in 4% buffered formalin and later embedded in paraffin. For staining, 2 µm thick sections were cut from each thrombus and subjected to immunofluorescence staining as described previously [20]. Briefly, NETs were stained using as primary antibodies: a mouse monoclonal antibody (IgG2a) against DNA/histone 1 (MAB3864; Sigma Aldrich, Millipore 0.55 mg/mL diluted 1:100, Billerica, MA, USA) and a rabbit antibody (IgG) against anti-human myeloperoxidase (A039829-2 Agilent, Santa Clara, CA, USA, 3.3 mg, 1:300). These were incubated overnight at 4 °C. The respective isotype control was included in each staining batch of analyzed thrombi. Therefore, murine IgG2a (from murine myeloma, M5409-1 mg conc. 0.2 mg/mL, 1:36,4 Sigma Aldrich, Munich, Germany) and rabbit IgG (from rabbit serum, Sigma Aldrich, Munich, Germany, I5006, 1.16 mg/mL, 1:96,7) was incubated under the same conditions as the primary antibodies.

As secondary antibodies goat anti-mouse Alexa 488Plus 2 mg/ml (# A32723, Invitrogen, Carlsbad, CA, USA) and goat anti-rabbit Alexa 633 2 mg/ml (# A21070, Invitrogen, Carlsbad, CA, USA, 2 mg, Waltham, MA, USA), both diluted 1:500 in blocking buffer, were used. Finally, all samples were processed using the TrueVIEW autofluorescence quenching kit (# SP-8400–15, Vector Laboratories, Newark, California, United States) following the manufacturer’s instructions and counterstained using Hoechst 33,342 (1:1000, stock 50 mg/mL, Sigma Aldrich, Munich, Germany).

The samples were imaged using a Leica TCS SP5 AOBS confocal inverted-base fluorescence microscope with an HCX PL APO × 40 0.75–1.25 oil immersion objective. The settings for each batch were adjusted to their respective isotype controls. Up to 10 images were taken per thrombus, if the size of the thrombus was sufficient. If the thrombus was too small, the maximum amount of images possible was taken instead. The imaged areas were chosen randomly without overlap, from the entirety of the stained sample. The raw integrated density was then separately measured using ImageJ software (ver. 1.53e) for the three different color channels (blue = DNA [counterstaining], green = DNA/histone-1-complexes and red = MPO). The raw integrated density of both antibody signals was then set in relation to the raw integrated density of the counterstaining, to adjust for thrombus cell count. This ratio of signal intensity is therefore represented as percentage. Every batch was imaged within 24 h after the completion of the staining process. An overview of the thrombus analysis process, including the pattern used to select the imaged areas, can be seen in Fig. 1.

a The figure contains a flowchart, describing the process of NET marker measurement in the thrombi. 10 images were taken randomly and without overlap from the entirety of the thrombus. The pattern used to ensure no overlap can be seen below. After imaging the color channels were split and their intensity measured separately. The final value was then determined by dividing the NET marker intensity by the DNA counterstaining. b Scatterplots of the analyzed thrombus markers DNA-histone 1 complexes (green) and MPO (red) are presented. The integrated density in both channels as well as for the DNA-counterstaining (blue) was measured up to 10 pictures of each thrombus. The ratio was calculated and is presented as percentage. Each dot presents the mean of each individual patient. The bars show median and interquartile range. c 10 analyzed images of two thrombi (green: DNA-histone 1 complexes, red: MPO, blue: DNA counterstaining, bar length: 100 µm) are presented to show the distribution of NET markers in the thrombus. Image series A is from a thrombus with relative intensities in the first (DNA-histone 1 complexes) and second (MPO) quartile, while image series B shows a thrombus with relative intensities in the fourth quartile (both markers). Created with biorender.com, Microsoft Powerpoint, Adobe Illustrator, Adobe Photoshop and Image J

Myeloperoxidase (MPO)-histone complexes were measured using a previously established sandwich enzyme-linked immunosorbent assay (ELISA), as reported earlier by de Buhr et. al. [21]. For this we used components of the Cell Death Detection ELISAPLUS Kit (Roche11774425001). The additional antibodies used were a rabbit anti-human MPO antibody (Merck Millipore #07–496-I, 1 mg/mL; 1:200 diluted in 1% PBS-BSA (bovine serum albumin, Roth CP84.2 or 1ETA.2)) and goat anti-rabbit IgG HRP conjugated (Merck Millipore #12–348, 1:5000 diluted in PBS). For this analysis we used 100 µl of EDTA buffered plasma per sample at baseline and after seven days.

Cell-free DNA was measured in 25 µl citrate buffered plasma at baseline and after seven days using a Quant-iT™ PicoGreen® assay (Invitrogen, Carlsbad, California, USA, P11496) as described previously [21].

DNase activity was measured in 25 µl serum at baseline and after seven days using a DNase I Activity Assay Kit (BioVision, Milpitas, California, USA, Fluorometric, K429-100), according to the manufacturer’s instructions.

The cytokine analysis was only conducted for cohort 2 since this has already been reported for cohort 1 in a previous work by our group [22]. The second group was analyzed using a MERCK Milliplex® Immunology Multiplex Assay (Merck Millipore #HCYTA-60 K-38C) according to the manufacturer’ s instructions. We analyzed EDTA buffered blood plasma samples taken at both baseline and after seven days.

Statistical analyses were conducted using IBM SPSS Statistics Version 28 and 29. The correlation between thrombus composition and NET markers in blood was evaluated using the Spearman rank test. The same test was used to identify correlations between the various NET markers and cytokines. To evaluate potential differences in blood markers due to different approaches of sample collection (i.e., venous vs. arterial puncture), in both cohorts’ blood markers were compared using the Mann–Whitney-U-test. An evaluation of the NET markers and the clinical parameters and outcomes was conducted using Mann–Whitney-U-test and Kruskal–Wallis-H-test for continuous data. For categorical data the Chi square test or Fisher’s exact test were used, as appropriate. Binary logistic regression analysis was performed on the association of MPO-histone complexes and mTICI score adjusting for intravenous thrombolysis and stroke etiology. These confounders were identified by testing for significant differences between outcome groups using the above-mentioned statistical tests and were also theoretically reasoned confounders in the causal path between exposure and outcome.