Cerebrovascular glycocalyx damage and microcirculation impairment in patients with temporal lobe epilepsy

Temporal lobe epilepsy (TLE) is increasingly associated with blood-brain barrier dysfunction and microvascular alterations, yet the pathophysiological link is missing. An important barrier function is exerted by the glycocalyx, a gel-like layer coating the endothelium. To explore such associations, we used intraoperative videomicroscopy to quantify glycocalyx and microcirculation properties of the neocortex and hippocampus of 15 patients undergoing resective brain surgery as treatment for drug-resistant TLE, and 15 non-epileptic controls. Fluorescent lectin staining of neocortex and hippocampal tissue was used for blood vessel surface area quantification. Neocortical perfused boundary region, the thickness of the glycocalyx’ impaired layer, was higher in patients (2.64 ± 0.52 µm) compared to controls (1.31 ± 0.29 µm), P < 0.01, indicative of reduced glycocalyx integrity in patients. Moreover, erythrocyte flow velocity analysis revealed an impaired ability of TLE patients to (de-)recruit capillaries in response to changing metabolic demands (R2 = 0.75, P < 0.01), indicating failure of neurovascular coupling mechanisms. Blood vessel quantification comparison between intraoperative measurements and resected tissue showed strong correlation (R2 = 0.94, P < 0.01). This is the first report on in vivo assessment of glycocalyx and microcirculation properties in TLE patients, confirming the pivotal role of cerebrovascular changes. Further assessment of the cerebral microcirculation in relation to epileptogenesis might open avenues for new therapeutic targets for drug-resistant epilepsy.


Supplemental figure 2. Photomicrographic stills of an intraoperative SDF recording.
Images showing blood vessels as automatically detected by GlycoCheck software.Red lines indicate an identified blood vessel, green lines a valid blood vessel segment for measurement, yellow lines an invalid measurement.(A) example of an included video, showing correct blood vessel and segment detection.(B) example of a video showing an air bubble obstructing the SDF camera from adequately visualizing the microvasculature, which was therefore excluded from further analysis.
Of the videos that were excluded from further analysis after manual identification of insufficient quality, many still had several vessel segments there were adequately visualized.
Similarly, manually selected videos that were included for further analysis still had some incorrect vessel segments that should be discarded.By using automatic GlycoCheck vessel segment selection, the correct vessel segments are adequately selected for further analysis, while the incorrect vessel segments in each video are discarded.Therefore, we decided to use the data generated after automatic selection of correct vessel segments by GlycoCheck software.This automatic selection resulted in the loss of a portion of possible data (average valid vessel segments: controls cortical 1369/3000=45.6%,patients cortical 1044/3000=34.8%;patients hippocampal 1305/3000=43.5%).
Both neocortex and hippocampus tissue samples were collected from 10 patients (P2, P3, P5, P7, P8, P9, P10, P11, P12, P13).Tissue was obtained during surgical procedure; the interval between first tissue manipulation and actual removal of the tissue varied between 15 and 90 minutes, and the time interval between tissue removal and fixation varied between 5 and 30 minutes.These samples were fixed in buffered 4% formaldehyde at room temperature upon resection; after 48 hours of fixation, they were embedded in paraffin.For each patient, one neocortex and one hippocampus cross-section (5 µm in thickness) were deparaffinized, rehydrated using Milli-Q, and incubated first for one hour at room temperature with Ulex Europaeus Agglutinin I lectin (10 µg/mL, UEA I DL-1067; Vector Laboratories USA), then for 20 minutes at room temperature with DAPI (1:10,000; Roche Diagnostics, Switzerland).
After incubation, tissues were thoroughly washed five times with 200 µL TBS-T for five minutes each, and mounted using ProLong Diamond Antifade (P36961; ThermoFisher, USA).
The slides were dried at 37°C for five minutes, and stored at 4°C.Visualization was performed using a fluorescence microscope (Olympus BX51WI) equipped with a digital camera (EM-CCD; C9100; HAMAMATSU PHOTONICS Europe GmbH) using 10x magnification (OPlanSApo; 10x/0.40;/0.17/FN26.5).Red, green and blue filters were used in QuPath (version 0.3.2) to image lectin-stained vessels, lipofuscin and nuclei, respectively. 36Sensitivity and offset for all filters were zero.Images were scaled and analyzed using FIJI (ImageJ; software version 2.3.0/1.53q). 37To improve identification of blood vessels, the autofluorescence from lipofuscin was reduced by spectral separation of the red and green filter using the FIJI plugin Poisson NMF.
Total Blood vessel Stained Surface Area (BSSA) was quantified by automated annotation using the red channel in QuPath. 36Regions with folded tissue were excluded using the 'wand' tool.
with Full Resolution (0.81 µm/px), the Gaussian prefilter, and a smoothing sigma of 0.5 for automated quantification.We used a total of three thresholds throughout the data collection to accurately quantify blood vessels and their surface area, as no single appropriate threshold using the 'pixel classification' function for automated quantification could be created.BSSA was calculated for neocortex and hippocampus tissue samples by dividing total vessel area (µm 2 ) by total sample area (µm 2 ).

Table 3 .
Intra-operatively recorded parameters at time of measurement.