All animal experiments were performed following ethical guidelines and were approved by the local authorities (G1713/18) of the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, Recklinghausen, Germany). Four-week-old male C57BL/6JHsd wild-type mice were received from Envigo, Netherlands. Mice were randomly divided into two groups and received a normal laboratory diet (Cat. 1329 FORTI, Ssnif) or a high-fat Western-style diet (Cat. E15721-347, Ssnif) for 6 weeks. Afterward, mice underwent sham or stroke surgery (transient middle cerebral artery occlusion, tMCAO) and were sacrificed three hours after the operation. The animals had ad libitum access to food and water. All experiments were performed and reported according to ARRIVE guidelines (12).

Brain injury was induced by tMCAO in C57BL/6JHsd mice (aged 11 weeks) anesthetized with 1.5% isoflurane in 100% oxygen. Mice were injected with the analgesic buprenorphine (0.1 mg/kg body weight, s.c.) 30 min before the surgery. An eye ointment (Bepanthen) was applied to avoid any harm to mouse eyes during the surgical procedures. A small incision was made between the ear and the eye to expose the temporal bone and a laser Doppler flow probe was attached to the skull above the core of the middle cerebral artery (MCA) territory. Mice were then placed in a supine position on a feedback-controlled heat pad and the midline neck region was exposed with a small incision. The common carotid artery (CCA) and left external carotid arteries were identified and ligated. A 2 mm silicon-coated filament (Cat. 702234PK5Re; Doccol) was inserted into the internal carotid artery to occlude the MCA. Brain ischemia was validated by a stable reduction of blood flow to ≤ 20% of baseline that was observed on the laser Doppler flow device. After 60 min of occlusion, the filament was removed for the reestablishment of blood flow. Mice were then injected with the anti-inflammatory drug carprofen (4-5 mg/kg body weight, s.c.), wounds were carefully sutured and mice returned to their cages with free access to food and water. Sham-operated mice underwent the same surgical protocol except the filament was inserted into the CCA and immediately removed. The exclusion criteria for experimental mice were as follows: inadequate ischemia (reduction of blood flow does not reach the ≤ 20% of baseline threshold), weight loss >20% of baseline body weight during the study, and spontaneous animal death.

Mice were deeply anesthetized with Ketamine/Xylazine (100 mg/kg/10 mg/kg, i.p.), and blood was collected via cardiac puncture and added to EDTA-containing tubes. Mice were then perfused with PBS and spleen and BM were collected in cold PBS. Single-cell suspensions were prepared by mincing the spleens in PBS and filtering through 70 µm cell filters. Spleen and blood samples were treated with erythrocytes lysis buffer and washed in PBS. The tibia was harvested through an incision a few millimeters below the knee joint and removed from muscle tissue. Both ends of the tibia were cut with sharp sterile scissors, and the bone marrow was flushed out with a 30-gauge needle and filtered through 30 µm cell filters. Single-cell suspensions were kept on ice and total cells per organ were calculated using an automated cell counter (Nexcelom Bioscience). From each sample, 5x10⁵ cells were stained with fluorochrome-conjugated antibodies against different cell surface markers. Antibodies used were CD45 (30-F11), CD11b (M1/70), Ly6G (1A8), CD68 (FA-11), Ly6C (HK1.4), CD62L (MEL-14) and CD162 (4RA10) were purchased from Biolegend, Germany. Cells were washed with cold PBS and suspended in a flow cytometry buffer. Samples were analyzed using MacsQuant Analyzer 16 (Miltenyi) and FlowJo software (BD Biosciences).

The EDTA blood was centrifuged at 3000 g and the supernatant was collected for a second centrifuge at 8000 g for 10 minutes. The quantification of NETs was performed as previously described capture ELISA, which is based on citrullinated histone H3 associated with DNA (13). Antihistone H3 antibody (5 µg/ml; ab5103, Abcam) was coated overnight at 4°C onto 96-well plates followed by 5% BSA blocking for 2 h. Wells were three times washed with 300 µl washing buffer followed by the addition of 50 µl plasma and 80 µl incubation buffer (including peroxidase-labeled anti-DNA antibody) for 2 h at 300 rpm (Cell Death ELISAPLUS, Cat. 11774425001, Roche). Then, the wells were washed three times with 300 µl washing buffer and 100 µl peroxidase substrate was added to the wells for 30 min in the dark. Afterward, 100 µl ABTS peroxidase stop solution was added to the wells, and absorbance was measured at 405 nm and was subtracted by absorbance at 490 nm (Abs 405 nm− Abs 490 nm). The absorbance values were considered in direct proportion to the amounts of soluble NETs and were presented as a relative increase to control.

The EDTA blood was centrifuged at 3000 g and the supernatant was collected for a second centrifuge at 8000 g for 10 minutes. Finally, plasma samples from HFD and ND-treated stroke mice were analyzed for total cholesterol levels using a bioanalyzer (ADVIA^® 2400; Siemens, Erlangen, Germany).

Plasma cytokines were measured using a bead-based Legendplex immunoassay according to the manufacturer’s instructions (Cat. 740622, Biolegend). Briefly, 12.5 µl of mouse plasma was diluted with an assay buffer. All diluted samples, standards, and controls were added to the V-shaped 96-well plate. Supplied antibodies-specific beads were vortexed and added to samples in the wells and incubated for 2 h at room temperature with continuous shaking. Afterward, samples and technical controls were centrifuged and washed three times and biotinylated detection antibodies were added to the samples for 1 h at room temperature. Without washing, streptavidin-phycoerythrin is subsequently added, which will bind to the biotinylated detection antibodies providing fluorescent signal intensities in proportion to the amount of bound analytes. Samples were washed and acquired on a flow cytometer (BD FACSAria). Different cytokine amounts were quantified against the acquired standard curve using LEGENDplex software.

Mice were sacrificed and perfused with 20 ml PBS followed by perfusion with 20 ml 4% PFA. Brains were carefully removed from the skull bone and transferred to 4% PFA for 12 h at 4°C. The next day, PFA was replaced by 30% sucrose solution and samples were incubated for 2 days at 4°C. Samples were tap-dried with tissue paper and quickly frozen in cold isopentane. Brains were transferred to −80°C until further processing. Cryosections of 20 μm thickness were prepared at 500 µm intervals from the neocortex and stained with cresyl violet (Cat. C5042, Sigma). All sections were scanned using 600 dpi and analyzed using the ImageJ software (NIH). Using a scale of 23.62 pixels/mm, the area of non-stained infarct tissue was measured and integrated into the total brain. An edema correction for brain infarct volume was performed using the following formula: (ischemic area) = (direct lesion volume) − [(ipsilateral hemisphere) − (contralateral hemisphere)]. The lesion volume per hemisphere was presented as mm³.

Mice were deeply anesthetized and intravenously injected with fluorochrome-conjugated 5 µg anti-Ly6G (1A8, Cat.127648, Biolegend) and 7.5 µg anti-CD31 (MEC13.3, Cat.102528, Biolegend) antibodies. After 30 minutes, mice were sacrificed and transcardially perfused with 20 ml PBS and 20 ml of 4% PFA. Samples were incubated in 4% PFA at 4°C for 12 h. The next day, brains were washed in PBS and dehydrated with serial dilutions of tetrahydrofuran (THF) 30-60-80-100% 12 h each at RT with constant shaking at 70 rpm. The last 100% THF incubation was repeated twice and refractory index matching was achieved by immersion of the samples in ethyl cinnamate (ECI, Cat.112372, Sigma). Optically transparent samples were imaged via LSFM (Ultramicroscope BLAZE, Miltenyi Biotec) with a 40% light sheet width and a sheet thickness of 4 µm. The images were acquired with a zoom factor of 6.64x (Objective LVBT 4x, zoom 1.66x) and a step size of 2 µm for a total volume of 4.5 mm³ per ipsilateral hemisphere. The acquired 3D images were converted via the Imaris file converter and processed using Imaris software version 10.1.0. (Bitplane, Switzerland). Ly6G⁺ surfaces were considered as neutrophil-positive areas. Imaris surface function was used to create the respective areas and neutrophil surfaces per mm³ of brain volumes were quantified and presented.

Data were analyzed using GraphPad Prism version 9.0. All datasets were tested for normal distribution using the Shapiro–Wilk normality test. The comparisons between the two groups were analyzed using the two-tailed Mann–Whitney U test. For statistics between more than two groups, ordinary one-way ANOVA with Tukey’s multiple comparisons tests was used for normally distributed data. The Kruskal-Wallis test was performed for comparisons of more than two groups followed by the Dunn’s multiple comparisons test for not normally distributed data. Differences with p-values ≤ 0.05 were considered to be statistically significant.

To investigate if the higher plasma cholesterol levels were linked to early innate immune activation after stroke, we induced hypercholesterolemia in mice with HFD for six weeks before the induction of stroke or sham operation ( Figure 1A ). As expected, HFD mice showed significantly increased body weight and higher plasma cholesterol amounts compared to ND mice and these levels remained unchanged at 3 h post-stroke compared to sham controls ( Figures 1B, C ). In addition, HFD mice had a higher liver-body weight ratio compared to ND mice ( Supplementary Figure 1A ).

Next, we analyzed the dynamics of neutrophils in systemic lymphoid tissues in HFD and ND mice after the induction of stroke or sham operation using flow cytometry ( Supplementary Figure 1B ). Our data showed that HFD increased the number of neutrophils in blood and spleen but their numbers were reduced in tibial bone marrow after stroke compared to sham controls ( Figures 1D–F ). In ND mice, the total number of neutrophils was only increased in the spleen after stroke compared to sham and bone marrow remained unaffected. In addition, a strong increase in the frequencies of blood and splenic neutrophils along with their decreased frequencies in tibial bone marrow after stroke in HFD mice indicated an accelerated neutrophil release to the circulation ( Supplementary Figures 1C–E ). However, stroke only increases the frequencies of splenic neutrophils compared to sham-operated ND mice. These data suggest that HFD-induced hypercholesteremia promotes post-stroke neutrophil activation and possibly increases their release from bone marrow.

To determine the activation pattern of neutrophils in HFD and ND mice, we compared the expression of different surface activation and adhesion receptors in blood, spleen, and tibial bone marrow after the induction of stroke or sham operation. The results showed that circulating neutrophils in HFD mice were characterized by a significantly higher expression of Ly6G ( Figure 2A ) and adhesion receptor CD162 ( Figure 2B ) compared to ND mice. However, stroke induction only increased the expression of CD162 and not Ly6G in splenic neutrophils in HFD mice compared to sham controls ( Supplementary Figures 1F, G ). In addition, we analyzed the activation pattern of MΦ after stroke in the blood and spleen of HFD and ND mice. Interestingly, our data showed that stroke in HFD mice upregulates the expression of a surrogate marker of activation Ly6C and the co-stimulatory receptor CD68 on blood MΦ ( Figures 2C, D ). However, no increase in the expression of these receptors was observed in ND mice after stroke compared to the sham group. Similarly, stroke increased the expression of CD68 but not Ly6C on splenic MΦ in HFD mice ( Supplementary Figures 1H, I ). Moreover, we found higher levels of plasma citH3-DNA nucleosomes as a measure of NET−formation in stroke mice with HFD compared to ND mice ( Figure 2E ). These results highlight the modulated responses of neutrophils and MΦ in systemic lymphoid tissues in response to HFD-induced hypercholesteremia and the induction of stroke.

Next, we analyzed the circulating amounts of proinflammatory cytokines in HFD and ND mice after stroke. Interestingly, stroke mice with HFD showed a significant upregulation of plasma TNF-α, IL-6 and MCP-1 amounts compared to sham controls within 3 h after stroke onset ( Figures 3A–C ). However, this rapid increase was not observed in stroke mice with ND. HFD sham mice have also increased plasma levels of TNF-α compared to ND mice. Furthermore, we found that stroke mice with HFD or ND show similar reduction and reperfusion of cerebral blood flow during laser Doppler flow (LDF) recordings ( Figure 3D ). In both groups, LDF reproducibly decreased to ~15-20% of baseline values during MCAO, followed by the reconstitution to ~75% within 15 minutes after reperfusion. Histological analysis of brain tissues of stroke mice showed large infarct volumes in HFD mice compared to ND controls after 3 h of ischemia-reperfusion injury ( Figure 3E ).

To further elucidate the underlying pathophysiology of increased brain infarcts in HFD mice after stroke, we implemented our three-dimensional light-sheet fluorescence microscopy (3D-LSFM) approach to analyze the extent of activated neutrophils migration to ischemic hemispheres ( Supplementary Figure J ). Our imaging data showed an augmented accumulation of neutrophils in the ipsilateral brain microvasculature and parenchymal tissue in HFD mice compared to ND mice ( Figure 3F ). Furthermore, the automated quantitative surface volume analysis showed a significant increase in Ly6G surface volumes in the ischemic brain hemispheres of HFD mice compared to ND mice ( Figure 3G ).