Rapid endothelial cytoskeletal reorganization enables early blood–brain barrier disruption and long-term ischaemic reperfusion brain injury

The mechanism and long-term consequences of early blood–brain barrier (BBB) disruption after cerebral ischaemic/reperfusion (I/R) injury are poorly understood. Here we discover that I/R induces subtle BBB leakage within 30–60 min, likely independent of gelatinase B/MMP-9 activities. The early BBB disruption is caused by the activation of ROCK/MLC signalling, persistent actin polymerization and the disassembly of junctional proteins within microvascular endothelial cells (ECs). Furthermore, the EC alterations facilitate subsequent infiltration of peripheral immune cells, including MMP-9-producing neutrophils/macrophages, resulting in late-onset, irreversible BBB damage. Inactivation of actin depolymerizing factor (ADF) causes sustained actin polymerization in ECs, whereas EC-targeted overexpression of constitutively active mutant ADF reduces actin polymerization and junctional protein disassembly, attenuates both early- and late-onset BBB impairment, and improves long-term histological and neurological outcomes. Thus, we identify a previously unexplored role for early BBB disruption in stroke outcomes, whereby BBB rupture may be a cause rather than a consequence of parenchymal cell injury.


Supplementary Figure 6 | Generation and characterization of EC-targeted ADF-or ADFm-overexpressing Tg mice.
(a) Illustration of the targeting strategy used for generating EC-specific ADF-or ADFmoverexpressing mice. cDNA of full-length WT human ADF or its constitutively active mutant (ADFm) bearing single amino acid substitution (S3A) was targeted to the ROSA26 locus downstream of a stop codon flanked by two loxP sites, driven by the CAG promoter (Tg-ADF stop or Tg-ADFm stop ). Tg-ADF stop or Tg-ADFm stop mice were crossed with Tek-Cre mice, in which the Cre recombinase expression is driven by the Tek (endothelial-specific receptor tyrosine kinase) promoter and thus restricted to ECs.
In the presence of Cre recombinase, the stop codon is excised and a 5' HA-tagged ADF or ADFm protein is expressed specifically in ECs (Tg-ADF or Tg-ADFm). (b) A typical genotyping gel showing analysis of PCR products with agarose gel electrophoresis. The predicted size of the WT or mutant allele is 510 bp and 364 bp, respectively. (c) Overexpression of ADF or ADFm in ECs was confirmed by Western blotting for the HA tag in brain protein extracts. HA was expressed in Tg-ADF and Tg-ADFm animals but not in uncrossed Tg-ADF stop or Tg-ADFm stop animals. (d) Double-label immunostaining for the endothelial marker CD31 (red) and the HA tag (green) in the cerebral cortex of uncrossed Tg-ADFm stop or crossed Tg-ADFm mice. ADFm was not expressed in uncrossed Tg-ADFm stop brains. In Tg-ADFm brains, expression of ADFm was predominantly in microvessels, as shown by colocalization of HA and CD31 in yellow. Scale bar: 500 µm. (e) Cerebral microvasculature was examined in the cortex and striatum of WT, Tg-ADFm stop , and Tg-ADFm brains by perfusion with FITC-lectin (green). Scale bar: 200 µm. (f) The vascular surface area, length, and branch points were quantified on lectin-labeled images. n=6 mice per group. No significant difference was observed in microvessel distribution and anatomy between WT and Tg animals.

Supplementary Figure 7 | ADFm expression preserves BBB integrity after tFCI.
Uncrossed Tg-ADFm stop mice or Cre-recombined Tg-ADFm mice were subjected to 1 h of tFCI and 1 h (a) or 24 h (b) of reperfusion. (a) Representative microscopic images (magnification: x200) demonstrate the extravasation of Alexa 555-dextran (3 kDa, red) into ipsilateral cortical and striatal parenchyma at 1 h after tFCI, which resulted in positive staining of non-vascular cells. Sections were counterstained with DAPI (blue) for nuclear labeling. Scale bar: 100 µm. EC-targeted expression of ADFm markedly reduced the extravasation of Alexa 555-dextran in both the cortex and striatum. (b) Representative images show the leakage of endogenous plasma IgG (green) and the injected Evans blue dye (red) into ipsilateral cortical and striatal parenchyma at 24 h after tFCI. Scale bar: 100 µm. EC-targeted expression of ADFm attenuated the extravasation of both plasma IgG and Evans blue. (c,d) Fluorescence intensity of Alexa 555-dextran and IgG immunostaining was measured, respectively, and expressed relative to the ADFm stop sham group. n=6 mice per group. *p≤0.05, **p≤0.01 versus ADFm stop .

Supplementary Figure 8 | Absence of Evans blue extravasation into the brain parenchyma after sham surgery.
Representative images of coronal brain sections demonstrate that sham operation did not cause leakage of the Evans blue dye into brain parenchyma in either WT or Tg-ADFm animals after 24 h. Scale bar: 2 mm.

Supplementary Figure 9 | ADF stop or ADFm stop mice exhibit comparable infarct volumes after tFCI compared to WT mice.
Brain infarct volume was measured in TTC-stained slices at 48 h after tFCI in WT and uncrossed Tg-ADF stop or Tg-ADFm stop mice (without crossing with the TEK-Cre mice). No significant difference in infarct volume was observed between any two groups. n=6 mice per group.

Supplementary Figure 10 | Labeling of brain microvessels with CD31.
Shown are representative images of CD31 + (green) microvessels taken under 40× and 300x magnifications, respectively. Square: the region enlarged in the high-power image. Scale bar: 10 µm. CD31 immunofluorescence shows a relatively diffused pattern under 40×, whereas a punctate pattern was observed in high-power images take under 300×.

Supplementary Figure 11 | Overexpression of ADFm in ECs does not alter the expression of junctional proteins in whole-cell lysates from brain microvessel extracts.
tFCI was induced for 1 h in WT and Tg-ADFm mice followed by 1 h of reperfusion. Whole cell lysates were prepared from brain microvessel extracts and probed with immunoblotting for occludin, VE-cadherin, ZO-1, and subfraction markers β-actin, CD31, or α-tubulin (see Fig. 7f). Quantification of blots (normalized to WT contralateral) is presented. tFCI does not induce significant loss of junctional proteins at 1 h of reperfusion, nor does endothelial ADFm overexpression alter the level of these proteins in whole-cell microvessel extracts. n=6 mice per group.

Supplementary Figure 12 | EC-targeted ADFm overexpression reduces ZO-1 degradation after tFCI.
tFCI was induced for 1 h in WT and Tg-ADFm mice followed by 24 h of reperfusion. Double-label immunostaining of the TJ protein ZO-1 (green) and endothelial marker CD31 (red) in the ischemic cortex is shown. Scale bar: 50 µm. In contrast to the distribution pattern of the basal lamina protein laminin, which was observed mainly in the outer layer of CD31 + microvessels (see Fig. 8b), ZO-1 was distributed predominantly inside CD31 + microvessels, as shown by colocalization with CD31 in yellow. tFCI caused degradation of ZO-1 protein, demonstrated by the loss of green signal and only a red color in merged images. ADFm overexpression significantly preserved ZO-1 from tFCI-induced degradation. Rectangle: the region enlarged in high power images. Arrowhead: loss of ZO-1 immunostaining.

Supplementary Figure 13 | ADFm overexpression does not alter MMP-9 production in blood neutrophils.
tFCI was induced for 1 h in WT and Tg-ADFm animals followed by 24 h of reperfusion. (a) Double-label immunostaining for MMP-9 (green) and neutrophil marker MPO (red) in ipsilateral cortex of WT brains. MMP-9 was expressed in MPO + neutrophils (arrowhead) and microvessels (arrow, see Fig. 8c). Scale bar: 50 µm in upper panels and 15 µm in lower panels. (b) MMP-9 levels in blood neutrophils were measured by ELISA in WT and Tg-ADFm mice at 24 h after tFCI or sham operation. Endothelial ADFm overexpression did not change neutrophil MMP-9 production. n=6 mice per group.

Supplementary Figure 14 | ADFm overexpression reduces neutrophil transmigration across the endothelium after OGD.
(a) Illustration of the in vitro neutrophil migration assay. Active neutrophils were extracted from the blood of mice at 24 h after tFCI. Neutrophils were labeled with fluorochromes and plated on top of the HBMEC monolayer pre-exposed to 1-h OGD in the migration inserts. Cells were allowed to migrate for 6 h and non-migratory cells were removed. Cells in the bottom chamber were lysed and fluorescence intensity was measured. (b,c) Migration of neutrophils obtained from CCR2 +/+ , CCR2 -/-, MMP-9 +/+ or MMP-9 -/mice across the HBMEC monolayer was assessed in the presence of absence of 5 ng/mL CCL2 in the bottom chamber. Data were expressed relative to CCR2 +/+ or MMP-9 +/+ control with no CCL2. Neutrophil migration was dramatically promoted by CCL2, but hampered in CCR2 -/or MMP-9 -/neutrophils. Data represent 4 independent experiments. *p≤0.05, **p≤0.01 versus non-CCL2 control. # p≤0.05, ## p≤0.01 versus CCR2 +/+ (b) or MMP-9 +/+ (c). (d) HBMECs were infected with Lenti, Lenti-ADFm, Lenti-MLCsc, or Lenti-MLCt, and subjected to 1 h of OGD. The transmigration of MMP-9 +/+ neutrophils across the endothelial monolayer was quantified after 6 h of co-culture, and expressed relative to non-transfected non-OGD controls. ADFm overexpression or MLC knockdown in ECs inhibited OGD-induced neutrophil transmigration. Data represent 4 independent experiments. *p≤0.05 versus Lenti or MLCsc.

Supplementary Figure 15 | ADFm overexpression reduces the expression of inflammatory markers after tFCI.
A panel of inflammatory markers was examined using the quantitative inflammation array in microvessel extracts from WT and Tg-ADFm brains at 24 h after tFCI. The levels of inflammatory markers were expressed relative to the WT sham group. ADFm overexpression significantly reduced the expression of several markers, including TNFα, TNFR1, IL-1β, leptin, and FasL. In contrast, no significant changes were observed in IL-12p70, G-CSF, TIMP1, GM-CSF and CXCL3. n=4 mice per group. *p≤0.05 versus WT.  b A coronal plane is specified based on its distance from the frontal pole in millimeters.

Supplementary Methods
As mentioned in the main text, Stroke Therapy Academic Industry Roundtable (STAIR) guidelines 1 were strictly followed throughout the experiments. For example, animals were randomly selected for experiments with a lottery-drawing box. We verified that blood pH, gases, glucose levels, and cerebral blood flow were not altered by transgene expression. Furthermore, surgeries and all outcome assessments were performed by investigators blinded to mouse genotype and experimental group assignments.

Generation of Tg-ADFm stop or Tg-ADF stop mice for endothelial cell-targeted overexpression of actin depolymerizing factor mutant (ADFm) or ADF
Tg-ADFm stop or Tg-ADF stop transgenic mice were generated for conditional overexpression of ADFm and ADF, respectively. Briefly, a targeting vector was constructed to contain WT human ADF or its constitutively active mutant (S3A, ADFm) downstream of a loxP-flanked stop sequence (neomycin resistance gene and a trimer of the SV40 polyadenylation sequence). This construct was inserted into the Gt(ROSA)26Sor locus via electroporation of embryonic stem (ES) cells, and the correctly targeted ES cells were subjected to microinjection in C57BL/6J blastocysts. The ES cell screening and microinjection procedures were performed at the Shanghai Research Center for Model Organisms through a service contract. Mutant mice progeny were backcrossed to the C57/B6 background for at least six generations before use to minimize the potential influence of genetic heterogeneity on the susceptibility of animals to cerebral ischemia. To obtain EC-specific ADF-or ADFm-overexpressing mice, homozygous Tg-ADF stop or Tg-ADFm stop mice were crossed with Tek-Cre mice 2 , in which the Cre recombinase expression is driven by the Tek (endothelial-specific receptor tyrosine kinase) promoter and thus restricted to ECs. In the presence of Cre recombinase, the stop codon is excised and a 5' HA-tagged ADF or ADFm protein is expressed specifically in ECs (Tg-ADF or Tg-ADFm). Overexpression of ADF/ADFm was validated by immunohistochemistry and Western blotting for HA (see Supplementary Fig. 6).

Choose of sample size
The number of animals required for the in vivo studies was determined by power analysis based on our experience with the murine MCAO model. To detect a 30% decrease in infarct volume or neurological deficits with 80% power at an α value of 0.05 (two-tailed), approximately 6-8 mice per group was needed. For blood-brain transfer coefficient and immunohistochemistry, 4 samples were required for 80% power (β=0.8, α=0.05) to detect a 30% change after tFCI. For Western blotting and ELISA analysis, 4-5 samples were required for 80% power (β=0.8, α=0.05) to detect a 30% change. Brains from 2 mice needed to be pooled together to prepare 1 sample of single cell suspension for flow cytometry analysis. Four samples (8 mouse brains) are required for 80% power (β=0.8, α=0.05) to detect a 30% change after tFCI.

Evaluation of mouse cerebrovascular anatomy
Cerebrovascular anatomy was quantitatively evaluated in Tg mice and their WT littermates as we described before 3 , to assess the potential impact of variations in the anatomy of the cerebral circulation on susceptibility to ischemic injury. Briefly, mice were sacrificed by CO 2 overdose, and transcardial perfusion fixation was performed immediately via the left ventricle with heparinized saline (10 units per mL) followed by warm formalin. Evans blue (EB; Sigma-Aldrich; 2% in normal saline, with sonication) was mixed with gelatin (20% in water) in equal volumes and kept warm to prevent solidification. The mixture was then injected through a cannula into the ascending aorta.
Brains were harvested and stored in formalin. Ventral and dorsal photographs of the brain were taken with a dissecting microscope to visualize the middle cerebral artery (MCA) and posterior cerebral artery (PCA) territories. The MCA territory was determined by the localization of anastomoses, and distances between the midline and the anastomoses were measured at coronal planes 2, 4, and 6 mm from the frontal pole. The plasticity of the posterior communicating artery (PcomA) was graded on a qualitative scale of 0-3: 0, no anastomosis between PCA and superior cerebellar artery (SCA); 1, anastomoses between PCA and SCA in capillary phase; 2, small truncal PcomA; 3, truncal PcomA.

Vascular labeling and three-dimensional analysis of vascular density
Brain microvessels were labeled by perfusion with lectin, as we described previously 4 . Briefly, mice were transcardially perfused with FITC-conjugated tomato lectin (Sigma-Aldrich) at a dose of 100 µg µL -1 . Coronal brain sections were prepared and imaged as described in Methods. Six sections, 0.5-mm apart, were analyzed for each brain, and six regions of interest (ROIs; 233 × 233 μm 2 ) in the cerebral cortex or striatum were selected from each section. The ROIs were scanned at 512 × 512 pixels in the x-y direction, and 1-μm-step-size optical sections along the z-axis were acquired with a 40× objective lens. Three-dimensional reconstruction was performed using an image analysis software package (3D Doctor 3.5, Able software, USA). The vascular surface area (mm 2 ) and the total vascular length (mm) per volume of tissue (mm 3 ) were calculated by the software, and the number of vascular branch points was counted in the three-dimensional images by a blinded investigator.

Two-dimensional laser speckle imaging
Cortical cerebral blood flow (CBF) was monitored using the laser speckle technique as we described previously 5 . Briefly, a charge-coupled device camera (PeriCam PSI System; Perimed Inc., Ardmore, PA, USA) was placed above the head, and the intact skull surface was illuminated by a laser diode (785 nm) to allow laser penetration through the brain in a diffuse manner. Speckle contrast, defined as the ratio of the standard deviation of pixel intensity to the mean pixel intensity, represents the speckle visibility relative to the velocity of the light-scattering particles (blood) and was therefore used to measure cortical blood flow. The speckle contrast was then converted to correlation time values, which were inversely proportional to the mean blood flow velocity. Two-dimensional microcirculation images were obtained 15 min before tFCI and continued throughout the ischemic period until 15 min after the onset of reperfusion. The area of the ischemic core (0-20% residual CBF) or the penumbra (20-30% residual CBF) region 6, 7 was measured from laser speckle images.

Neurobehavioral tests
Neurobehavioral tests were carried out 1 d before and 1-28 d after MCAO. Sensorimotor deficits were evaluated by the rotarod, cylinder and corner tests. Longterm cognitive deficits were evaluated by the Morris water maze test.

Rotarod test
The rotarod test was performed to assess post-stroke motor functions as we described previously 6 . Briefly, animals were placed on a rotating drum with a speed accelerating from 2.5 to 25 rpm within 5 min. The time at which the mouse fell off the drum (latency to fall) was recorded. The test began 1 d before surgery and consisted of 2 trials. On the day of surgery, 5 trials were performed on each mouse and the mean of trial # 3, 4, and 5 was used as the pre-surgery baseline value. After surgery, mice were tested for 5 trials per day with intervals of 15 min, and the data for trial #3-5 were expressed as the mean latency to fall on each testing day.

Cylinder test
The cylinder test was performed to assess forepaw use asymmetry, as we described previously 6 . The mouse was placed in a transparent cylinder (diameter: 9 cm; height: 15 cm), and videotaped for 5 min. A mirror was placed behind the cylinder at an angle which allowed the rater to record all forepaw movements. Videotapes were analyzed in slow motion, and forepaw (left/right/both) use during the first contact against the cylinder wall after rearing and during lateral exploration was recorded. Preference of the non-impaired forepaw (left) was calculated as a relative proportion of right forepaw contacts: (left-right)/(left+right+both)×100%. Uninjured mice typically show no preference for either forepaw, whereas injured mice have increased left forepaw preference depending on the severity of the injury.

Corner test
The corner test was performed as we described previously 8 . The injured animal turns preferentially toward the non-impaired (left) side. Performance was expressed as the number of left turns out of 10 trials for each test.

Morris water maze test
The Morris water maze test was carried out 23-28 d after tFCI to evaluate long-term cognitive functions, as we described previously 9 . Briefly, a platform (diameter: 11 cm) was submerged in a pool (diameter: 109 cm) of opaque water. Mice were placed into the pool from one of the four locations and allowed 90 s to locate the hidden platform (learning phase of the test). The time at which the animal found the platform (escape latency) was recorded for each trial. At the end of each trial, the mouse was placed on the platform or allowed to stay on the platform for 30 s with prominent spatial cues displayed around the room. Four trials were performed on each day for 5 consecutive days. After the last day of the hidden platform test, a single, 60-s probe trial was performed in which the platform was removed. The time spent in the target quadrant where the platform was previously located was recorded (memory phase of the test). Swim speed was also recorded to assess locomotor function.

Measurement of albumin leakage
BBB permeability and leakage of plasma albumin was determined by measuring EB extravasation as we described previously 10 . Briefly, 2.5% EB (5 mL kg -1 ) was injected into the femoral vein. Three hours later, animals were transcardially perfused with cold saline to remove intravascular EB. Coronal brains sections (2 mm thick) were cut and EB extravasation was visualized. Alternatively, EB fluorescence was examined on 30µm coronal sections using confocal microscopy, as described in Methods. To quantify EB extravasation, sections were carefully weighed and soaked in methanamide for 48 h at 37.0°C, and subsequently centrifuged for 30 min at 14,000 rpm. The absorption of the supernatant was measured at 632 nm with a spectrophotometer. EB concentration in the tissue was quantified using a standard curve and expressed as ng per mg of protein.