This study includes preclinical and clinical data and considers modern techniques for acute vessel recanalization in both the animal experiments and the clinical investigations. First, mouse and rat brains were subjected to different types of focal cerebral ischemia in order to analyze cytoskeletal alterations using immunofluorescence microscopy, Western blotting, electron microscopy, and atomic force microscopy. Second, NF-L and MAP2 levels were measured in blood samples from patients suffering from different types of focal cerebral ischemia.
Animal experiments included the filament-based model in 64 male C57BL/6 J mice and the thromboembolic model in 12 male Wistar rats, both leading to right-sided focal cerebral ischemia due to an occlusion of the middle cerebral artery (MCAO; details are given below). Animals were provided by Charles River Laboratories (Sulzfeld, Germany) with a mean body weight of 25 g (mice) and 290 g (rats). Both models were used to induce either permanent ischemia (pMCAO) or 60 min of transient ischemia (tMCAO) with observation periods of 4 and 24 h in mice and rats as well as 72 h in mice. While the middle cerebral artery of 4h, 24h, and 72h pMCAO animals remained occluded for the given observation times, 4h, 24h, and 72h tMCAO animals underwent either mechanical (mice) or pharmacological recanalization (rats) after an ischemic period of 60 min (details are given below). The different models are abbreviated as follows: “h” refers to the observation time, while “t/p” refers to the applied model (transient/permanent MCAO). Sufficient MCAO-related cerebral infarction was evaluated according to the standardized scoring system by Menzies et al. , whereas animals had to reach a minimum of 2 (decreased grip of the contralateral forelimb while tail pulled) as pre-defined study inclusion criterion.
Ischemia Induction in Animals
MCAO surgery was performed under deep anesthesia using 2–2.5% isoflurane (Baxter, Unterschleißheim, Germany) in a mixture of 70% N2O/30% O2. In mice, MCAO was induced according to Longa et al. , with minor modifications as described before in detail . In short, a standardized silicon-coated 6-0 filament (Doccol Corporation, Redlands, CA, USA) was inserted into the internal carotid artery and carefully moved forward to the origin of the middle cerebral artery until bending was observed or resistance felt. In rats, MCAO surgery was performed according to Zhang et al. , with minor modifications as described earlier . In brief, a weight-adapted blood clot was prepared and later injected in the distal section of the internal carotid artery using a PE-50 catheter and a small bolus of saline. Every procedure during anesthesia was performed using a rectal probe and a thermostatically controlled warming pad adjusting the body temperature to 37°C. In order to mimic recanalizing approaches such as intravenous thrombolysis and mechanical thrombectomy as routinely applied in the clinical setting, transient MCAO (tMCAO) was accomplished by either filament removal in mice or intravenous administration of recombinant tissue plasminogen activator (rtPA, Actilyse®, Boehringer Ingelheim, Ingelheim am Rhein, Germany) in rats 60 min after ischemia induction, whereas animals with permanent ischemia (pMCAO) remained untreated.
Animals were sacrificed and transcardially perfused with saline and 4% paraformaldehyde (PFA; Serva, Heidelberg, Germany) in phosphate-buffered saline (PBS) at pre-defined observation periods of 4, 24, and 72 h. Brains were removed from the skull and post-fixed in 4% PFA for 24 h, followed by equilibration in 30% phosphate-buffered sucrose. Forebrains were then serially cut into coronal 30-μm-thick sections with a freezing microtome (Leica SM 2000R, Leica Microsystems, Wetzlar, Germany). All brain sections were stored at 4°C in 0.1 M Tris-buffered saline, pH 7.4 (TBS), containing 0.2% sodium azide. For multiple immunofluorescence labeling, sections were blocked with 5% normal donkey serum and 0.3% Triton X-100 in TBS for 1 h and then incubated overnight with primary antibodies (Table 1) diluted in the blocking solution. The next day, sections were incubated with mixtures of appropriate secondary antibodies (Table 1) in TBS containing 2% bovine serum albumin for 1 h at room temperature. Every change of incubation medium was preluded and followed by thorough rinsing in TBS. Finally, sections were mounted onto fluorescence-free microscope slides and cover-slipped with fluorescence mounting medium (Dako North America, Inc., Carpinteria, CA, USA). Omitting primary antibodies served as control, which resulted in the absence of staining. Microscopy and image acquisition of qualitative MAP2/NF-L data were performed with the Biorevo BZ-9000 microscope (Keyence, Neu-Isenburg, Germany).
Sample sizes used for immunofluorescence microscopy are as follows: mice, 4h-t (n=5), 4h-p (n=6), 24h-t (n=6), 24h-p (n=6), 72h-t (n=6), and 72h-p (n=2); rats, 4h-t (n=3), 4h-p (n=4); 24h-t (n=3), and 24h-p (n=2).
Fluorescence Intensity Measurements
For quantification of fluorescence intensities, NF-L- and MAP2-labeled brain sections were scanned with an Axio Scan.Z1 slide scanner (Carl Zeiss Microscopy GmbH, Jena, Germany), and files were analyzed using the netScope Viewer Pro software (Net-Base Software GmbH, Freiburg i. Br., Germany). Two different quantifications were performed:
(a) Corresponding to data shown in Fig. 2: Images of striatal and cortical regions in mice after 4h-t, 4h-p, 24h-t, 24h-p, 72h-t, and 72h-p were captured at 10× magnification. Due to the slightly variable distribution of the ischemic lesion after MCAO , sections and regions of interest (ROIs) were selected based on the ischemia-induced decrease of MAP2-related immunofluorescence intensity. They were then mirrored to the contralateral hemisphere which served as control, thus capturing 4 ROIs per animal. In one 72h-p mouse brain, poor tissue integrity impeded quantification of the cortex, leading to a sample size of 72h-p: cortex, n=1, and striatum n=2, only.
(b) Corresponding to data shown in Fig. 3: To assess possible region-dependent alterations of immunofluorescence intensities, 9 ROIs were placed from non-affected medial to affected lateral regions throughout the cortex of the ischemic hemisphere of 24h-t and 24h-p mice. Thereby, ROIs 4 and 5 were strictly placed next to the ischemic border. Thus, ROIs 1–4 were always placed in non-ischemic and ROIs 5–9 in ischemic areas. On the contralateral hemisphere, 3 ROIs were mirrored to corresponding regions and served as control. To capture the immunofluorescence intensity throughout all cortical layers, each ROI had a field dimension of 200μm width and a height which was adapted to the height of the cortex at each point. Per animal and ROI, the mean fluorescence intensity of each ROI was measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA) to calculate the n-fold change compared to the mean value of the 3 contralateral ROIs, which served as control.
Mice were sacrificed and perfused with saline, only. Brains were removed from the skull and manually dissected. Next, stroke-affected areas, being demarked by the ischemia-associated edema, as well as the respective contralateral areas were dissected, snap-frozen in liquid nitrogen, and stored at −80°C. Each sample was homogenized and lysed by ultrasonification in 60 mM Tris-HCl, pH 6.8, containing 2% sodium dodecyl sulfate (SDS), 10% sucrose, and a protease inhibitor cocktail (Cell Signaling, Leiden, The Netherlands) on ice. Next, probes were centrifuged at 13,000 rpm and 4°C for 10 min, followed by protein concentration measurements using the BCA kit (Thermo Fisher, Waltham, MA, USA). Proteins were denatured in sample buffer (250 mM Tris-HCl, pH 6.8, containing 4% SDS, 10% glycerol, and 2% β-mercaptoethanol) at 95°C for 5 min. Then, proteins were separated using a 12.5% SDS-PAGE and transferred to nitrocellulose membranes (Th.Geyer, Renningen, Germany). Membranes were blocked with 5% dry milk in TBS (50 mM Tris-HCl, 150 mM NaCl, pH 7.5) for 30 min and incubated with primary antibodies (Table 1) at 4°C overnight. After thorough washing in buffer (6 g/l Tris, 8.8 g/l NaCl, 3 ml/l Tween 20), horseradish peroxidase–conjugated secondary antibodies (Table 1) were added for 1 h, and membranes were developed with the ECL kit (Thermo Fisher). After image acquisition, membranes were stripped with stripping buffer (15 g/l glycine, 1 g/l SDS, 10 ml/l Tween 20, pH 2.2) and reused to detect β-actin as housekeeping protein for reference. The relative protein concentration of MAP2 and NF-L was calculated from the respective β-actin-related chemiluminescence intensity. Sample sizes for Western blot analyses were 4h-t (n=5), 4h-p (n=6), 24h-t (n=6), and 24h-p (n=5).
Mice were sacrificed and transcardially perfused with saline, followed by perfusion with 4% PFA and 0.5% glutaraldehyde in PBS. After removal from the skull, brains were post-fixed in the same fixative for 24 h. Then, brains were cut into 60-μm-thick coronal sections using a vibratome (Leica Microsystems, Wetzlar, Germany) in cooled PBS. Sections were then transferred into PBS and stained with 0.5% osmium tetroxide (EMS, Hatfield, PA, USA) for 30 min, followed by dehydration in graded ethanol and another staining step with 1% uranyl acetate (Serva) in 70% ethanol for 1 h. Sections were further dehydrated in ethanol and propylene oxide (Sigma Aldrich, Steinheim, Germany) and then incubated in Durcupan (Sigma Aldrich). After embedding between coated microscope slides and cover slips, the Durcupan-saturated sections were polymerized at 56°C for 48 h. Areas showing stroke-associated edema were identified by light microscopy and transferred onto blocks of resin for a second polymerization step. The blocks were trimmed and cut into 55-nm-thick ultra-thin sections using an ultra-microtome (Leica Microsystems). Finally, the sections were transferred on formvar-coated grids and stained with lead citrate for 6 min. Ultrastructural analysis was performed using a Zeiss SIGMA electron microscope (Carl Zeiss Microscopy GmbH, Oberkochen, Germany).
In ischemia-affected cortical regions, the neurofilament density was analyzed in cross sections of myelinated axons. For this purpose, detectable neurofilaments were identified by diameter and morphology and counted, and the area of the cross-sectioned axon was measured to calculate the neurofilament density per μm2. Measurements thereby included 5 different fields of view from ischemia-affected cortical areas per animal, which were calculated as mean values and compared with contralateral unaffected regions. Of note, as a proper distinction of neurofilaments is impeded in tangentially cross-sectioned axons, the analysis was restricted to proper cross sections, while cytoplasmic areas showing organelles such as mitochondria were also excluded. Sample sizes for electron microscopy were: 4h-t and 24h-t (n=5, each).
Atomic Force Microscopy
To demonstrate alterations of mechanical properties within ischemia-affected regions, atomic force microscopy (AFM) was performed in a mouse 24 h after pMCAO (24h-p). To enable the detection of ischemia-affected regions, fluorescein isothiocyanate–conjugated albumin (FITC-albumin, 0.2 mg dissolved in 0.1ml saline, Sigma, Taufkirchen, Germany) was intravenously applied to demarcate areas of ischemia-associated blood-brain barrier (BBB) breakdown 23 h after ischemia induction . After a circulation time of 1h, the brain was removed from the skull and immediately cut into 350-μm-thick slices using a vibratome (HM 650V, ThermoFisher Scientific, Walldorf, Germany) in artificial cerebrospinal fluid (ACSF, containing 2.5 mM KCl, 260 mM D-Glucose, 26 mM NaHCO3, 1.25 mM NaH2PO4, 2 mM Na-pyruvate, 3 mM myo-inositol, 0.4 mM ascorbic acid, 1 mM MgCl2, 2 mM CaCl2, 20 mM Hepes, pH 7.4). Subsequently, slices were glued (Histroacryl, Braun, Melsungen, Germany) onto microscope slides followed by measurement of the elastic strength (Young’s modulus) in the ischemia-affected and contralateral cortex. The AFM used is a NanoWizard 4 with 300μm HybridStage (JPK, Berlin, Germany) combined with an Axio Zoom.V16 (Zeiss, Oberkochen, Germany) for fluorescence imaging . The AFM image and optical image are recorded in a common reference system using the AFM software calibration routine. A CONT (Nanoworld, Neuchâtel, Switzerland) contact mode cantilever (spring constant 213mN/m) was modified with a 6-μm-diameter polystyrene bead to increase contact area. Force ramps were recorded with the following parameters: maximum force 7.5 nN, z-speed 20 μm/s, z-length 30 μm, 2048 Hz capture rate, and 10 μm data point spacing (Suppl. Figure 1 A & B). AFM data was first analyzed with the JPK data processing software to calculate the Young’s modulus using a Hertz fit to the smoothed and baseline corrected force-indentation curves (Suppl. Figure 1C). Data was post-processed with a custom written Matlab program (MathWorks, Natick, MA, USA) to calculate local averages and construct the overlay of AFM data with the whole-slice overview shown in Fig. 6. In the ischemia-affected cortex, the elastic strength of the tissue was recorded over a distance of approximately 1100 μm starting from the area of evident FITC-albumin extravasation across the border zone of detectable FITC-albumin extravasation, which delineates the infarcted tissue [52,53,54]. Mean values of 11 positions were calculated, with each position measuring 100×100 μm and consisting of 100 single measurements. In the contralateral cortex, 3 different positions (total 300 data points) served as control. During the whole procedure, the coronal brain slices were kept submerged in ACSF to prevent dehydration. Sample size is 24h-p (n=1).
Blood Sampling in Patients Suffering from Focal Cerebral Ischemia
In a prospective, non-interventional study, a total of 81 patients, hospitalized at the stroke unit of the Department of Neurology, Leipzig University, were included. Blood samples were collected at two pre-defined time points: 12–24 h (“day 1”) and 3–5 days (“day 3”) after hospital admission. Main inclusion criteria were (a) ischemic stroke, defined by a sudden onset of a focal neurological deficit independent of the vascular territory with evidence of a cerebral infarction either on computed tomography or magnetic resonance tomography, and (b) transient ischemic attack (TIA) with a naturally lacking evidence for a cerebral infarction on radiological examination . Main exclusion criteria were any kind of known other cerebral pathologies such as neurodegenerative or inflammatory disorders as well as intracerebral hemorrhage. Patients were categorized for (a) “TIA,” (b) “stroke,” and (c) “stroke with intervention,” which represent patients who underwent a therapy aiming to re-establish cerebral blood flow (i.e., intravenous thrombolysis and/or mechanical thrombectomy). Furthermore, imaging data were used to categorize for (a) “lacunar infarcts” (diameter ≤1.5 cm) and (b) “larger infarcts” (diameter >1.5 cm). For descriptive analyses, additional data were recorded concerning the severity of the neurological deficit (National Institute of Health Stroke Scale, NIHSS) at hospital admission as well as the individual infarct etiology and medical history.
Enzyme-Linked Immunosorbent Assay (ELISA)
Human blood samples were allowed to coagulate for 1 h at room temperature and then centrifuged at 3500 × g for 10 min. Thereafter, the supernatant was aliquoted and stored at −80°C until further processing. Serum protein levels were measured using commercial ELISA kits from Abbexa (Cambridge Science Park, Cambridge, United Kingdom; NF-L, Cat.# abx258398; MAP2, Cat.# abx358608). Analyses were performed according to the instructions given by the manufacturer. The optical density (OD) was measured at 450nm using a Mithras LB940 microplate reader (Berthold Technologies, Bad Wildbad, Germany), and then protein concentrations were calculated by extrapolation of the linear portion of the standard curve. The following sample sizes were achieved: TIA, MAP2 (n=12) and NF-L (n=12); stroke, MAP2 (n=26) and NF-L (n=27); stroke with intervention, MAP2 (n=21) and NF-L (n=20); lacunar infarct, MAP2 (n=17) and NF-L (n=16); and larger infarct, MAP2 (n=29) and NF-L (n=30).
Data analysis was performed using Graph Pad Prism 5.01v (GraphPad Software Inc., La Jolla, CA, USA) and the SPSS software package version 25 (IBM SPSS Statistics for Windows, IBM Corp., Armonk, NY, USA). The Grubbs’ or ROUT test was used to check for statistical outliers, and the Kolmogorov-Smirnov test (SigmaStat; v3.10, San Jose, CA, USA) was used to check for normal distribution of the data. After confirmation of a normal distribution, ANOVA followed by Bonferroni’s multiple comparison post hoc test was used to check for statistically significant differences between three or more groups. Non-normally distributed data of two dependent groups were analyzed with the Wilcoxon signed-rank test, while two independent groups were analyzed using the non-parametric Mann-Whitney U test. Non-normally distributed data of three or more independent groups were analyzed using the Kruskal-Wallis test followed by Dunn’s multiple comparison post hoc test. For analysis of the human serum samples, the potentially confounding factors age and hypertension were ruled out by additional calculations (not shown). In general, with the significance level α=0.05, p<0.05, indicated statistically significant differences.