Data for: Endothelial glycocalyx disruption as a driver of blood–brain barrier dysfunction
Data files
Sep 29, 2026 version files 48.61 KB
-
OHare_et_al_2026_Endothelial_Glycocalyx_BBB_Data_Set.xlsx
32.44 KB
-
README.md
16.17 KB
Abstract
This dataset contains the raw and processed quantitative data supporting the manuscript “Endothelial Glycocalyx Disruption as a Driver of Blood-Brain Barrier Dysfunction” (O’Hare et al., 2026). It combines in vivo cerebral microvascular measurements from young adult, aged, APP/PS1 Alzheimer’s disease, and mild traumatic brain injury mouse cohorts with in vitro measurements from human blood-brain barrier models following selective knockdown of the endothelial glycocalyx core proteins CD44 or syndecan-1. The deposited workbook includes quantitative measurements of glycocalyx composition, endothelial junctional organization, transporter expression, vascular inflammation, and barrier permeability. Raw, normalized, biological-replicate, and image-level values are retained where applicable, with missing values and analysis exclusions documented in the accompanying README. The dataset can support comparative studies of neurovascular dysfunction, independent statistical reanalysis, and benchmarking of vessel-segmentation and junction-fragmentation workflows.
Dataset DOI: 10.5061/dryad.rbnzs7htx
Description of the data and file structure
This Excel workbook contains the quantitative data supporting the in vivo and in vitro experiments described in the manuscript. The workbook contains 16 worksheets. Numeric outputs are stored as cell values and can be viewed without specialized software. The workbook contains no macros, charts, or embedded figures.
Files and variables
File: OHare_et_al_2026_Endothelial_Glycocalyx_BBB_Data_Set.xlsx
Description:
| Sheet name | Description |
|---|---|
| In Vivo HA % Area Coverage | Quantification of hyaluronic acid labeling in cerebral vessels from young, aged, APP/PS1 Alzheimer's disease, and traumatic brain injury mice. Data are organized by biological replicate and brain region (PFC, hippocampus, and midbrain). Columns include total vessel area coverage, HA area coverage, and HA coverage within the vessel mask. |
| In Vivo WGA % Area Coverage | Quantification of wheat germ agglutinin labeling as a broad measure of endothelial glycocalyx-associated carbohydrate coverage. Includes vessel area coverage, WGA area coverage, and WGA coverage within vessels for each cohort, brain region, and biological replicate. |
| In Vivo CD44 MFI | Raw and normalized mean fluorescence intensity of CD44 labeling in cerebral vessels across the four mouse cohorts and three brain regions. CD44 was evaluated by fluorescence intensity rather than vessel-restricted percent area coverage. |
| In Vivo ZO-1 % Area Coverage | Quantification of ZO-1 labeling within cerebral vessels. Includes vessel area coverage, ZO-1 area coverage, and ZO-1 coverage within vessels for each cohort, region, and biological replicate. |
| In Vivo ICAM-1 % Area Coverage | Quantification of vascular ICAM-1 labeling as a measure of endothelial inflammatory activation. Includes vessel area coverage, ICAM-1 area coverage, and ICAM-1 coverage within vessels. |
| Cell Information | Experimental metadata for hCMEC/D3 endothelial cells, primary HBMECs, human pericytes, and human astrocytes. Fields include vendor, lot number, passage range, culture medium, supplements, plating density, and culture duration. |
| CD44 KD Densitometry | Western blot densitometry used to validate CD44 knockdown. Includes CD44 band intensity, β-actin intensity, the CD44/β-actin ratio, and the ratio normalized to the matched control. Samples prefixed with "Primary" correspond to primary HBMEC validation experiments; samples without this prefix correspond to hCMEC/D3 cells. |
| Syndecan-1 KD Densitometry | Western blot densitometry used to validate syndecan-1 knockdown. Includes syndecan-1 band intensity, β-actin intensity, normalization to β-actin, and normalization to the matched control. "Primary" sample names identify HBMEC validation experiments. |
| Hyaluronic Acid Quantification | Confocal mean fluorescence intensity of HA in control and CD44-knockdown endothelial cells. The sheet contains five biological replicates, with three image fields per condition within each biological replicate. Raw MFI and MFI normalized to the control-field mean are provided. |
| Heparan Sulfate Quantification | Confocal mean fluorescence intensity of heparan sulfate in control and syndecan-1-knockdown endothelial cells. The sheet contains five biological replicates, with three image fields per condition within each biological replicate. Raw and control-normalized MFI values are provided. |
| PERMEABILITY VALUES | Raw blood-brain barrier permeability coefficients and values normalized to matched controls for CD44-knockdown and syndecan-1-knockdown tri-culture models. Permeability was measured using a 3 kDa Texas Red dextran tracer and is reported in cm/s. |
| ZO-1 Densitometry | Western blot densitometry for ZO-1 in CD44-knockdown and syndecan-1-knockdown endothelial cells. Includes target intensity, β-actin intensity, target/β-actin ratio, and normalization to the matched control. |
| LRP-1 Densitometry | Western blot densitometry for LRP1 in CD44-knockdown and syndecan-1-knockdown endothelial cells. Includes target intensity, β-actin intensity, target/β-actin ratio, and normalization to the matched control. |
| Junctional Quantification | Image-level quantification of endothelial junctional organization for control, CD44-knockdown, and syndecan-1-knockdown cells. Variables include ZO-1/VE-cadherin Manders' colocalization coefficient, ZO-1 internalization ratio, average ZO-1 fragment area, and ZO-1 junctional ratio (junctional coverage). The sheet contains five biological replicates and three image fields per condition; the fragment-analysis exclusion is described below. |
| GLUT-1 Densitometry | Western blot densitometry for GLUT1 in CD44-knockdown and syndecan-1-knockdown endothelial cells. Includes GLUT1 intensity, β-actin intensity, normalization to β-actin, and normalization to the matched control. |
| VE-Cadherin Densitometry | Western blot densitometry for VE-cadherin in CD44-knockdown and syndecan-1-knockdown endothelial cells. Includes VE-cadherin intensity, β-actin intensity, normalization to β-actin, and normalization to the matched control. |
Variables
In vivo mouse cohorts
Young: Young adult healthy C57BL/6-background female mice, 10 to 14 weeks old, n = 4.
Old: Aged C57BL/6-background female mice, 80 to 85 weeks old, n = 4 overall, subject to any sheet-specific missing values.
AD: APP/PS1 double-transgenic female mice, 80 to 85 weeks old, n = 4.
TBI: Young female mice subjected to mild traumatic brain injury, n = 2. These data were treated descriptively and excluded from the inferential statistical models.
Brain regions
PFC: Prefrontal cortex.
Hipp: Hippocampus.
Mid: Midbrain.
In vitro experimental groups
Control: Brain endothelial cells expressing a non-targeting scrambled shRNA construct.
CD44_Knockdown: Endothelial cells with stable shRNA-mediated CD44 knockdown.
Syndecan-1_Knockdown: Endothelial cells with stable shRNA-mediated syndecan-1 knockdown.
The majority of experiments used hCMEC/D3 endothelial cells at passages 30 to 40. Primary HBMECs at passages 4 to 8 were used in parallel for knockdown validation and junctional image quantification. Primary human brain vascular pericytes at passages 4 to 8 and primary human astrocytes at passages 3 to 7 completed the tri-culture model. Endothelial cells were exposed to 12 dyn/cm² fluid shear stress for 24 hours, while pericytes and astrocytes remained under static conditions.
Description of variables
N: Biological replicate number. In imaging sheets containing repeated fields, numbered rows within an N block represent technical image replicates from the same biological replicate.
Vessel Area Coverage (%): Percentage of the analyzed image area classified as lectin-positive cerebral vasculature.
Marker Area Coverage (%): Percentage of the total analyzed image area positive for HA, WGA, ZO-1, or ICAM-1.
Marker Coverage In Vessel (%): Percentage of the vessel mask positive for the specified marker, calculated as marker-positive area within vessels relative to vessel area.
MFI: Mean fluorescence intensity, reported in arbitrary fluorescence units unless normalized.
Normalized MFI: Fluorescence intensity expressed relative to the applicable control reference. The precise normalization reference for each sheet is described below.
Target Intensity: Background-corrected Western blot band intensity for the indicated target protein.
B-Actin Intensity: Western blot band intensity for β-actin, used as the loading control.
Normalized to B-Actin: Target-protein band intensity divided by the corresponding β-actin band intensity.
Normalized to Control: Target/β-actin ratio or permeability coefficient divided by the matched control value, so the matched control equals 1.0.
Manders’ Coefficient (ZO-1/VE-CAD): Colocalization coefficient quantifying the overlap of ZO-1 and VE-cadherin fluorescence at endothelial junctions.
ZO-1 Internalization Ratio: Cytoplasmic-to-junctional ZO-1 signal ratio. Higher values indicate greater redistribution of ZO-1 away from cell-cell junctions.
ZO-1 Fragment Area: Average size of discrete ZO-1-positive junctional fragments. Smaller values indicate greater junctional fragmentation.
ZO-1 Junctional Coverage (%): Percentage of the analyzed junctional region occupied by ZO-1 signal above background. Lower values indicate reduced junctional continuity.
Raw Permeability (cm/s): Blood-brain barrier permeability coefficient calculated from 3 kDa Texas Red dextran transport and corrected for cell-free membrane permeability.
PFC: Prefrontal cortex.
Hipp: Hippocampus.
Mid: Midbrain.
HA: Hyaluronic acid.
HS: Heparan sulfate.
WGA: Wheat germ agglutinin.
ZO-1: Zonula occludens-1.
ICAM-1: Intercellular adhesion molecule 1.
SDC-1: Syndecan-1.
LRP1 or LRP-1: Low-density lipoprotein receptor-related protein 1.
GLUT1 or GLUT-1: Glucose transporter 1.
VE-cadherin or VE-CAD: Vascular endothelial cadherin.
KD: Knockdown.
BBB: Blood-brain barrier.
GCX: Endothelial glycocalyx.
Missing values, exclusions, and blank cells
No numeric value should be inferred from a blank cell. ROUT or experimental exclusions are noted in the notes section where applicable.
The TBI cohort contains two biological replicates, so the N = 3 and N = 4 blocks are intentionally blank for TBI samples in the in vivo worksheets.
The In Vivo CD44 MFI sheet contains no value for the Old N = 3 midbrain sample due to sample damage during processing.
In the Junctional Quantification sheet, CD44-knockdown biological replicate N = 5 contains Manders’ coefficient and internalization measurements but blank average ZO-1 fragment-area and junctional-ratio values. These blanks correspond to one CD44-knockdown sample excluded from fragment analysis because of weak ZO-1 signal, while its Manders’ coefficient and internalization measurements were retained.
The manuscript reports ROUT-based exclusion of one CD44-knockdown replicate from the ZO-1 and LRP1 Western blot analyses. The deposited workbook retains the complete raw paired datasets. Control_2 and CD44_Knockdown_2 were excluded from ZO-1 and LRP-1 quantification based on ROUTS analysis.
The syndecan-1 knockdown-validation sheet contains six hCMEC/D3 paired experiments, while the manuscript reports n = 5 for the summarized knockdown validation. The replicate not included in the manuscript is n=3 due to a failed knockdown in this replicate.
Code/software
General viewing
Microsoft Excel, LibreOffice Calc, or Google Sheets can be used to open the .xlsx workbook. No proprietary software is required to view the deposited numeric values.
Image and fluorescence quantification
ImageJ/Fiji was used for fluorescence measurements, region-of-interest analysis, and Western blot densitometry.
Ilastik pixel classification and a Fiji macro were used for vessel-restricted in vivo image segmentation and quantification.
A custom Python analysis pipeline was used for ZO-1 fragment morphometry and related junctional measurements.
Companion image-analysis dataset
Companion data and image-analysis methods are available in the private peer-review Dryad dataset for Peck et al., “Fragment Morphometry Analysis and Same-Color-Channel Separation Enable Objective Quantification Across BBB Models” (2026). The associated methods manuscript is available as a bioRxiv preprint (doi: 10.64898/2026.06.26.734824).
The companion workbook provides per-image pipeline outputs demonstrating how the measurements reported here were generated. In vitro ZO-1 outputs include filtered fragment count, mean fragment area, total junctional area, junctional ratio, adaptive threshold, and background intensity. In vivo outputs include vessel area, marker coverage, background thresholds, region-of-interest dimensions, and acquisition metadata for ZO-1, ICAM-1, and eNOS within tomato-lectin-defined vessels. Its README documents the workflows and links to version 1.0 source code for the ZO-1 fragment-morphometry and ilastik/Fiji vessel-colocalization pipelines.
Statistical analysis
GraphPad Prism was used for statistical analysis. Data normality was confirmed using the Shapiro-Wilk test, and data are presented as mean ± SEM. Normalized permeability values were tested against the theoretical control value of 1.0 using one-sample t-tests. Other analyses used unpaired t-tests, one-way ANOVA with Tukey post-hoc testing, or two-way ANOVA with Tukey post-hoc testing, as appropriate. ROUT identified specified outliers. The TBI cohort was shown descriptively and excluded from inferential statistical models. Statistical significance was defined as p < 0.05.
Experimental acquisition and instrumentation
In vivo confocal images were acquired using a Zeiss LSM 880 at 20x or 40x magnification with DAPI, 488 nm for vessel labeling, 555 nm for the protein of interest, and 647 nm for Texas Red dextran. Each Z-stack contained approximately 10 optical sections acquired at a 1 µm step size.
In vitro confocal images were acquired using a Zeiss LSM 800 at 63× magnification with DAPI, 555 nm, and 647 nm channels. Z-stacks were acquired using a 0.6 µm step size, and three image fields were analyzed per biological replicate.
Permeability fluorescence was measured using a SpectraMax i3x plate reader.
Access information
Other publicly accessible locations of the data:
None at the time of deposit.
Data were derived from the following external sources:
None. All deposited values were generated as part of the experiments described in the associated manuscript.
Data integrity and reuse
All numeric values are stored directly in the Excel workbook as cell values. The workbook does not contain macros, charts, or embedded images. Worksheets use replicate blocks and paired sample identifiers to preserve the relationship between raw and normalized values.
The data may be imported into R, Python, MATLAB, Prism, or other statistical software for reanalysis. Potential reuse applications include meta-analysis of blood-brain barrier dysfunction, benchmarking of vessel-segmentation and junction-fragmentation workflows, comparison of glycocalyx markers across neurological disease models, and integration with endothelial junctional or transporter-expression datasets.
Users should preserve the distinction between biological and technical image replicates, retain the paired-control structure of Western blot and permeability datasets, and treat the TBI cohort as descriptive because of its sample size.
In vivo measurements were generated from four cohorts of female C57BL/6-background mice: young adults (10 to 14 weeks; n = 4), aged mice (80 to 85 weeks; n = 4), age-matched APP/PS1 Alzheimer’s disease mice (n = 4), and mice subjected to mild traumatic brain injury (n = 2). The traumatic brain injury cohort was evaluated descriptively. Tissue from the prefrontal cortex, hippocampus, and midbrain was labeled with DyLight 488-conjugated tomato lectin to identify cerebral vessels and fluorescent probes or antibodies targeting hyaluronic acid, wheat germ agglutinin, CD44, ZO-1, or ICAM-1. Confocal Z-stacks contained approximately 10 optical sections collected at a 1 µm step size. Lectin-positive vessel masks were generated using a blinded ilastik and Fiji workflow, and marker coverage within vessels or mean fluorescence intensity was quantified as appropriate.
In vitro experiments used hCMEC/D3 endothelial cells at passages 30 to 40 and primary human brain microvascular endothelial cells at passages 4 to 8. Stable CD44 or syndecan-1 knockdown was produced using lentiviral shRNA, with a scrambled shRNA construct serving as the control. Endothelial cells were incorporated into tri-culture blood-brain barrier models with primary human brain vascular pericytes and astrocytes. Endothelial cells were exposed to 12 dyn/cm² fluid shear stress for 24 hours, while pericytes and astrocytes remained under static conditions. Glycocalyx composition and junctional organization were evaluated by confocal microscopy, and protein expression was measured by Western blot densitometry. In vitro confocal Z-stacks were acquired on a Zeiss LSM 800 at 63× magnification using DAPI, 555 nm, and 647 nm channels with a 0.6 µm step size and three image fields per biological replicate. ZO-1 colocalization, internalization, and fragmentation were quantified using Fiji and a custom Python pipeline.
Barrier permeability was measured using 0.25 mg/mL Texas Red-conjugated 3 kDa dextran over 60 minutes, with abluminal samples collected in triplicate and analyzed using a SpectraMax i3x plate reader. Western blot values were normalized first to β-actin and then to the matched control. Experiment-dependent statistical analyses were performed in GraphPad Prism using one-sample or unpaired t-tests, one-way or two-way ANOVA with Tukey post hoc testing, and the ROUT method for specified outlier exclusions. Complete normalization procedures, missing values, and exclusions are documented in the accompanying README.