Cav-1 Protein Levels in Serum and Infarcted Brain Correlate with Hemorrhagic Volume in a Mouse Model of Thromboembolic Stroke, Independently of rt-PA Administration

Thrombolytic therapy with recombinant tissue plasminogen activator (rt-PA) is currently the only FDA-approved drug for acute ischemic stroke. However, its administration is still limited due to the associated increased risk of hemorrhagic transformation (HT). rt-PA may exacerbate blood-brain barrier (BBB) injury by several mechanisms that have not been fully elucidated. Caveolin-1 (Cav-1), a major structural protein of caveolae, has been linked to the endothelial barrier function. The effects of rt-PA on Cav-1 expression remain largely unknown. Here, Cav-1 protein expression after ischemic conditions, with or without rt-PA administration, was analyzed in a murine thromboembolic middle cerebral artery occlusion (MCAO) and in brain microvascular endothelial bEnd.3 cells subjected to oxygen/glucose deprivation (OGD). Our results show that Cav-1 is overexpressed in endothelial cells of infarcted area and in bEnd.3 cell line after ischemia but there is disagreement regarding rt-PA effects on Cav-1 expression between both experimental models. Delayed rt-PA administration significantly reduced Cav-1 total levels from 24 to 72 h after reoxygenation and increased pCav-1/Cav-1 at 72 h in the bEnd.3 cells while it did not modify Cav-1 immunoreactivity in the infarcted area at 24 h post-MCAO. Importantly, tissue Cav-1 positively correlated with Cav-1 serum levels at 24 h post-MCAO and negatively correlated with the volume of hemorrhage after infarction, the latter supporting a protective role of Cav-1 in cerebral ischemia. In addition, the negative association between baseline serum Cav-1 levels and hemorrhagic volume points to a potential usefulness of baseline serum Cav-1 levels to predict hemorrhagic volume, independently of rt-PA administration.


Introduction
Acute ischemic stroke is the most frequent cause of permanent disability in adults worldwide [1]. Additionally, a poor prognosis and increased mortality after stroke is linked to different risk factor as diabetes, affecting also millions of people in the world [2,3]. However, less than 10% of stroke patients receive recombinant tissue plasminogen activator (rt-PA), which is still the only FDA-and EMA-approved fibrinolytic drug with a level I-A of evidence for the treatment of acute ischemic stroke [4,5]. This is due, among other reasons, to the increased associated risk of hemorrhagic transformation (HT) which occurs as a result of severe blood-brain barrier (BBB) disruption during reperfusion [6]. rt-PA itself may exacerbate BBB injury by several mechanisms, such as augmented neurovascular cells toxicity [7], elevated free radicals generation [8], and the activation of matrix metalloproteinase 9 (MMP-9) [9]. However, the molecular mechanisms underlying rt-PA effects on the BBB disruption remained to be fully understood. In this context, the maintenance of BBB homeostasis represents an interesting target not only for neurovascular protection but also for the development of thrombolytic adjuvant therapies aimed at decreasing the rt-PA-associated HT risk.
Caveolin-1 (Cav-1), a major structural protein of caveolae, has been linked to the endothelial barrier function [10], mainly by regulating endocytosis and vesicular trafficking [11]. Nevertheless, its role in cerebral ischemic injury and BBB dysfunction still needs to be clarified. For instance, some works suggest that Cav-1 may have a protective function after ischemic conditions, inhibiting MMP-9 activity and regulating post-ischemic angiogenesis [12,13] whereas other studies propose a Cav-1 detrimental role that impairs the endothelial tight junction proteins and increases BBB permeability [14,15]. Moreover, it has been suggested that phosphorylation of Cav-1 (pCav-1) may be associated with early BBB breakdown after brain injury [16].
Therefore, in the present study, we aim to investigate the effects of delayed rt-PA administration on Cav-1 protein expression in a murine thromboembolic-reperfusion middle cerebral artery occlusion (MCAO) model, both at brain and serum level, as well as in an oxygen and glucose deprivation (OGD)-exposed bEnd.3 cell line. The correlation between brain and serum Cav-1 protein levels and the volume of infarction, hemorrhage, and edema were also evaluated in the murine model.

In Vivo Thromboembolic Stroke Model Performance
Adult male Swiss mice (Jackson labs, Bar Harbor, Me) with a mean weight of 30 g were used for this study. All procedures were performed in accordance with the European Communities Council Directive (86/609/EEC) and approved by the Ethics Committee on Animal Welfare of University Complutense (PROEX No. 016/18) and are reported according to ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. Animals were housed individually under standard conditions of temperature and humidity and at a 12 h light/dark cycle (lights on at 8 h) with free access to food and water.
The surgical procedure for in situ thromboembolic model was carried out during light cycle (9h-13h) as previously described [17,18]. Briefly, mice were anaesthetized in a chamber ventilated with 2.5% isoflurane and then maintained at 1.5-2% isoflurane in a 30/70% mixture of O 2 /air. Body temperature was maintained at 36.5-37 °C using a feedback-controlled heating blanket. Mouse alphathrombin (2 UI) was injected into the MCA to induce its occlusion by a clot. A clot was defined as stable when laser Doppler flowmetry displayed a drastic fall of brain perfusion (mean reduction of 70-80%) that remained stable during 60 min. For reperfusion, rt-PA (10 mg/kg) was intravenously administered 3 h after thrombin injection. We considered that reperfusion was effective when blood flow was recovered (in the range of 60 to 100% of basal values) and remained stable within the first 60 min after rt-PA injection.
A total of 32 animals were assigned arbitrarily to three groups: (1) middle cerebral artery occlusion (MCAO) (n= 12), in which the middle cerebral artery (MCA) was permanently occluded and vehicle was intravenously administered 3 h after thrombin injection; (2) MCAO + rt-PA (n= 11), in which artery recanalization was achieved administering rt-PA at 3 h after thrombin injection; and (3) sham (n = 9), in which the MCA was surgically exposed but not occluded. In vivo data analysis was performed by a person other than the experimenter and sample size estimation was based in previous studies. Mice with spontaneous reperfusion (without rt-PA administration) (n=3), with extraparenchymal hemorrhages (n=2), or with striatal lesions (n=1) were excluded from further analysis. No spontaneous mortality was found after MCAO and this was unaffected by the rt-PA administration.

Blood Sample Collection
Blood tail samples were collected before (t = pre-MCAO) and after the experimental procedure (3 h post-thrombin injection, previously to rt-PA or vehicle administration (t=0), and at 3 and 24 h after rt-PA or vehicle injection) in the different experimental groups (Fig. 1a). Samples were kept at room temperature for 1 h and at 4 °C overnight, allowing coagulation. Samples were then centrifuged at 1500g and 4 °C, and the obtained serum was kept at 80 °C until its analysis.

Measurement of Serum Cav-1 Levels
Serum Cav-1 concentration was measured at pre-MCAO, 0, 3, and 24 h using an ELISA kit (SEA214Mu, Cloud Clone Corp. Houston, USA) in accordance with the manufacturer's instructions. In brief, 100 µl of blank, standards, and serum samples (diluted 1/20) were added to wells and incubated for 2 h at 37 °C. After removing the liquid of each well without washing, 100 µl of detection reagent A working solution was added to each well and incubated for 1 h at 37 °C. After washing, wells were sequentially incubated with 100 µl of detection reagent B working solution for 30 min and 90 µl of substrate solution for 20 min at 37 °C. Finally, 50 µl of stop solution was added to each well and absorbance was immediately measured at 450 nm using the SpectraMax 340PC384 Microplate Reader (Molecular Devices). Data were divided by the corresponding pre-MCAO level and represented as a percentage to reduce initial variability. The time point of 0 h was taken as the baseline levels.

Tissue Collection
Twenty-four hours after MCAO (Fig. 1a), mice were sacrificed by an overdose of sodium pentobarbital and were transcardially perfused with 0.1 M phosphate buffer (pH 7.4) followed by a solution of 4% paraformaldehyde in PBS. Brains were post-fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) solution overnight and then placed in 30% sucrose in PBS for 3 days at 4 °C until they sank. Brains were then frozen in isopentane and stored at −80 °C for further analysis.

Determination of Brain Edema, Infarct Volume, and Volume of Hemorrhage
One coronal section of 30 µm thickness every 400 µm was stained with cresyl violet (Nissl immunostaining) and diaminobenzidine (DAB) to measure edema and infarct and hemorrhagic volumes as previously reported [17]. Briefly, the ratio of the entire area of the ipsilateral hemisphere to that of the contralateral one was considered edema. The infarct area was delineated and determined (in mm 2 ) by counting the number of pixels within the outline. The infarct volume (in mm 3 ) was calculated as the sum of the orthogonal projections of each damaged area over the section thickness. In order to exclude the brain swelling effects, infarct volume was corrected by the edema and data were expressed as a percentage of the hemisphere. All noticeable hemorrhages, both petechial and parenchymal ones, were quantified by stereology using Cast Grid software (Visiopharm, Denmark). The volume of extravasated red cells was calculated by Cavalieri applying the following formula: (volume = a(p)·d· P) where a(p) is the area associated to the dot, d the distance between two consecutive sections, and P the counted dots inside the hemorrhage.

Immunohistochemistry
Serial coronal sections of cryopreserved brain (15-μm-thick) were obtained in a cryostat (CM1950, Leica) at −23 °C, at coordinates between Bregma 2.4 mm and −4.2 mm. The bEnd.3 cells were subjected to 2.5 h of OGD and subsequently reoxygenated with/ without rt-PA. Total protein was extracted at 0, 3, 24, and 72 h for western blot analysis. An immunofluorescence analysis was performed at 72 h after treatments slices were mounted onto SuperFrost/Plus slides (Menzek-Gläser, Braunschweig, Germany) and stored at −80 °C until immunohistochemistry staining.
Four double labelling of Cav-1 and extravasated IgG, frozen sections were dried and permeabilized with TBS 0.5% Triton X-100 for 10 min and blocked with TBS-T (TBS 0.1% Triton X-100) 1% bovine serum albumin (BSA) for 30 min. Sections were then incubated with a rabbit anticaveolin-1 antibody (1:200, sc-894, Santa Cruz Biotechnology) for 3 h at room temperature, washed 3 times with TBS-T, and incubated 2 h more with Alexa Fluor® 488 goat anti-rabbit and 594 goat anti-mouse IgGs (1:750 and 1:100 respectively, Invitrogen). Finally, samples were washed, stained with DAPI, and mounted with Dako fluorescent mounting medium (Dako North America Inc., USA). No immunostaining was observed in control slides without the primary or secondary antibodies. Mouse IgG staining was used to identify the infarcted zone where the BBB leakage occurs [19].
Additionally, a similar immunohistochemistry protocol was used to confirm the expression of Cav-1 in brain endothelial cells. Sections were simultaneously incubated with the rabbit anti-caveolin-1 antibody and the rat anti-PECAM-1 antibody (1:50, sc-18916, Santa Cruz Biotechnology) for overnight at 4°C and the Alexa Fluor® 488 goat anti-rabbit and 594 goat anti-rat (1:500, A-11007, Invitrogen) for 2 h at room temperature, respectively.
Microphotographs were taken with an Olympus DP70 digital camera (Japan) attached to a BX41 Olympus microscope. Image-J image analysis software was used to assess greyscale intensity levels. An average of Cav-1 intensity was measured in the ipsilateral hemisphere (infarcted zone) and in the contralateral hemisphere (CTR) between Bregma 2.4 mm and −4.2 mm for all regions. In the sham group, anatomically equivalent brain areas in the ipsilateral and contralateral hemispheres were analyzed.

Cell Culture
Immortalized mouse brain endothelial cell line (bEnd.3) purchased from ATCC (CRL-2299), was seeded in 60 mm Petri dishes (Corning, USA) for western blot analysis, or on the top of a transwell insert (0.3 cm 2 surface area, 0.4 µm pore size, PET membrane, BD Falcon), for metabolic activity and transcellular permeability analysis, as previously described [20]. bEnd.3 cells were grown as a monolayer in DMEM high glucose (HG) medium with 1% glutamine (Gibco, USA), 10% fetal bovine serum (Gibco, USA), and 1% Penicillin/Streptomycin (HyClone Laboratories, USA). All bEnd.3 cells used for these experiments were cultured between 25 and 30 passages, which have been shown to maintain excellent BBB characteristics in vitro [21].

OGD Exposure
To mimic acute ischemia-like conditions in vitro, bEnd.3 cells were exposed to OGD for 2.5 h as we described previously [20,22]. In brief, after overnight starvation in DMEM HG with 1% fetal bovine, bEnd.3 monolayers were subjected to OGD. The medium was replaced with glucosefree DMEM without FBS (Gibco, USA) previously perfused with N 2 to purge the oxygen. The cells were then placed into a 37 °C humidified hypoxic chamber with a constant N 2 flow of 1 l/min and 0.15 bar pressure for 2.5 h. Regarding the control (CTR) group, the same procedure was carried out with the difference that the glucose-free medium was supplemented with glucose (5.5 mM) and incubated at 37 °C with 5% of CO 2 . At the end of the OGD period, the media were removed and replaced with DMEM HG medium containing 10% FBS and with or without rt-PA at a concentration of 20 μg/ml and cultures were returned to the normoxic incubator. As reported in previous publications, we used 20 μg/ml of rt-PA, based on the finding that such a concentration can be observed in blood [23].
MTT assay was performed as follows. The medium of the transwells was aspirated, 100 µl of fresh medium and 10 µl of MTT (5mg/ml) (Sigma) were added to each transwell, and cells were incubated at 37 °C for 2 h. The medium was then carefully removed, and formazan crystals were lysed in 100 µl of DMSO by gently shaking the plate. Absorbance was measured at 570 nm using the SpectraMax 340PC384 Microplate Reader (Molecular Devices). MTT results were expressed as a percentage of the value in the CTR group.
For the analysis of transcellular permeability, inserts were transferred into new wells containing 0.75 ml of fresh serum-free medium and the medium of the luminal chamber was replaced with 0.15 ml of medium containing 0.35 mg/ml of FITC-BSA. After 1 h, the abluminal medium was sampled (duplicates of 200 µl) and fluorescence was measured on a Cytation™ 5 Cell Imaging Multi-Mode Reader (Biotek) at excitation and emission wavelengths of 485 and 520 nm. Changes in permeability were calculated relative to inserts without cells (blank inserts), which served as a reference for maximum permeability. The following formula was used: permeability (% of max) = ((FITC reading of experimental insert-average FITC reading of the vehicle group)/(FITC reading of the blank insert-average FITC reading of the vehicle group)) × 100.

Western Blot Analysis
At 0, 3, 24, and 72 h post-reoxygenation, cells were collected and protein was isolated using lysis buffer (Cell Signaling, The Netherlands) (Fig. 1b). The protein concentration was measured using the BCA method (Thermo Fisher Scientific, USA). Protein samples (10 μg) were loaded and separated by electrophoresis on 4-15% Criterion™ TGX Stain-Free™ Precast Gels (Bio-Rad) at 120 V for 80-90 min. Proteins were then transferred to PVDF membranes at 30 V overnight at 4 °C. After 1 h of blocking with Tris buffered saline with 0.1% Tween-20 (TBST) 5% BSA (EMD Millipore, USA), membranes were incubated with anti-phospho-caveolin-1 (Tyr14) (1:1000, Cell Signaling, The Netherlands), anticaveolin-1 (sc-894) (1:40000, Santa Cruz Biotechnology, USA), and anti-rabbit HRP-conjugated (1:10000, Cell Signaling, The Netherlands) antibodies in TBST 3% BSA. Stripping was performed to reprove the membranes. Protein bands were revealed using Immobilon Western Chemiluminescent HRP Substrate (EMD Millipore, USA). Quantification of the results was performed using Alpha Ease FC software (Alpha Innotech, USA) to measure integrated density of bands after background subtraction. Normalized expression of pCav-1 was obtained by comparing to total expression of Cav-1. Total expression of protein in the same lane was used as a loading control. These expressions were obtained after exposition to 5 min with UV light in order to activate the trihalo compounds present in the used criterion stain-free gels according to the previously described [22].

Statistics
SPSS software (IBM SPSS Statistics 22) was used to perform the statistical analysis. Shapiro-Wilk test was performed to assess the normality of the data. Cav-1 immunoreactivity, metabolic activity, transcellular permeability, and Cav-1 and pCav-1 in vitro expression were compared by one-way ANOVA followed by a Bonferroni post hoc analysis when required. Analyses of Cav-1 serum levels were conducted with a linear mixed model, which corresponded to two between-group factors, the GROUP (sham, MCAO, and MCAO + rt-PA) and the TIME (0, 3, and 24 h post-treatment). Bonferroni post hoc contrast was used when required. Correlations between variables were estimated using the Spearman test. The significance level (alpha) for all tests was set at .05.

Cav-1 Expression in the Infarcted Area and Correlation with Damage Parameters After Thromboembolic Stroke and Delayed rt-PA Administration
A previously described in situ mouse model of thromboembolic stroke and reperfusion [17,18] was used to analyze Cav-1 levels in the infarcted area and serum and their correlation with the parameters of infarct volume, edema, and hemorrhagic volume.
In MCAO animals, the thrombin injection caused a rapid fall in cerebral blood flow rate generating an infarction ( Fig. 2a and b, p=.023 vs. sham) in the cortex region. Delayed rt-PA administration (after 3 h of MCAO) was effective in recovering cerebral perfusion (data not shown) although it did not significantly affect the size of the lesion (p=.965 vs. MCAO). Both MCAO and MCAO + rt-PA animals showed hemorrhagic transformation of the ischemic lesion (p=.045 and p=.004 vs. sham, respectively; Fig. 2c). Importantly, significant increased edema values were only detected in the rt-PA-treated MCAO animals (p=.004 vs. sham) (Fig. 2d).
As shown in Fig. 3, Cav-1 immunoreactivity was significantly increased in the infarcted area of the MCAO group (p=.009 vs. sham) after 24 h post-thrombin injection. A similar Cav-1 increase was detected in the infarcted area of MCAO + rt-PA group (p<.002 vs. sham; p=.211 vs. MCAO). Cav-1 immunoreactivity in contralateral hemispheres of both MCAO and MCAO + rt-PA groups was similar to Cav-1 immunoreactivity in both hemispheres of the sham group (p=.932) ( Fig. 3a and b). Cav-1 protein was expressed in PECAM positive cells, showing that Cav-1 was specifically overexpressed in murine brain endothelial cells after 24 h post-occlusion (Fig. 3c).
The analysis of correlations demonstrated a significant negative correlation between Cav-1 immunoreactivity in the infarcted area and the hemorrhagic volume (p=−.900, p=.037) in the MCAO group at 24 h post-thrombin injection, with a similar trend in MCAO + rt-PA animals (p=−.700, p=.188) (Fig. 3d). There was not any significant association between Cav-1 immunoreactivity in infarcted area with the other damage parameters (infarct volume and edema; data not shown).

Serum Cav-1 Levels and Correlation with Tissue Cav-1 Expression and Damage Parameters After Thromboembolic Stroke and Delayed rt-PA Administration
Serum Cav-1 levels were similar in all groups (sham, MCAO, and MCAO + rt-PA), and no differences were either found at any of the times analyzed (0, 3, and 24 h after rt-PA or vehicle administration). The interaction between GROUP and TIME was not statistically significant either (Fig. 4a).
Baseline serum Cav-1 levels (3 h after ischemia induction and prior to rt-PA administration) and hemorrhagic volume at 24 h post-treatment were negatively correlated in MCAO animals (p=−.648, p=.043) (Fig. 4c). No significant correlations were found between serum Cav-1 levels and edema or infarct volume (data not shown).

Cav-1 and pCav-1 Expression in bEnd.3 Cells After OGD and Delayed rt-PA Treatment
Cav-1 expression was significantly increased by 2.5 h of OGD (0 h post-reoxygenation p=.006) and remained higher compared to CTR conditions from 3 to 72 h after reoxygenation (3 h p=.006; 24 h p<.001; 72 h p=.011). However, when rt-PA was added to OGD-exposed cells, Cav-1 expression significantly diminished to levels similar to CTR conditions from 24 to 72 h after rt-PA administration (24 h p=.002; 72 h p=.004 vs. OGD). In CTR conditions, rt-PA treatment did not modify Cav-1 expression at any of the analyzed time-points ( Fig. 5a and b).
The effects of rt-PA on the Cav-1 phosphorylation ratio (pCav-1/Cav-1) were also analyzed. As shown in Fig. 5c, this ratio was only significantly altered at 72 h. At this time, the ratio decreased in OGD compared to CTR condition (p=.043) and increased in OGD + rt-PA compared to OGD condition (p=.037).
The evaluation of metabolic activity and transcellular permeability was also performed at 72 h. Metabolic activity was significantly decreased after OGD (p=.002), and the addition of rt-PA further decreased it in cells subjected to OGD (p=.026 vs. OGD; p<.001 vs. CTR + rt-PA) (Fig. 5d). Transcellular permeability was significantly increased after OGD or after rt-PA treatment in CTR cells (p=.015 and .002 vs. CTR, respectively), and it was found that the effect was potentiated when the two treatments were carried out together (p=.005 vs OGD; p=.017 vs. CTR + rt-PA) (Fig. 5e).
The changes observed in total Cav-1 expression by western blotting were also confirmed by immunofluorescence in samples obtained at 72 h post-reoxygenation. It was also demonstrated that the significant increase of Cav-1 immunoreactivity induced in the bEnd.3 cells subjected to OGD (p<.001 vs. CTR) was not observed when cells were treated with rt-PA (p<.001 vs. OGD) (Fig. 6).

Discussion
rt-PA has been proved to be an effective thrombolytic therapy after acute ischemic stroke. However, it has been associated with increased BBB permeability and, therefore, with an increased risk of HT after delayed treatment. On the other hand, after an ischemic insult, Cav-1 has been related to BBB dysfunction but the previously published data are still controversial. In this context, possible interactions between rt-PA and Cav-1 in in vivo ischemic conditions are largely unknown. To the best of our knowledge, this is the first study to analyze brain and serum Cav-1 levels in a mouse model of thromboembolic stroke with a delayed rt-PA administration. Our results show that tissue Cav-1 protein expression at 24 h post-MCAO (1) increases in endothelial cells of the infarcted area, (2) positively correlates with Cav-1 serum levels at 24 h, (3) negatively correlates with the volume of hemorrhage after infarction, and (4) is not modified by The thromboembolic stroke mouse model with delayed rt-PA administration has been used to analyze Cav-1 expression in similar conditions to the human clinical situation. MCAO animals showed infarct and hemorrhagic volumes that were similar to those that have been previously published [17,18]. Additionally, although no significant rt-PA effect on infarct size lesion was found, probably due to irreversible tissue damage at the time of rt-PA administration, increased edema values in the rt-PA-treated MCAO animals would be in agreement with the concept that delayed administration of rt-PA exacerbates the disruption of the BBB [24].
Our results show that Cav-1 immunoreactivity is significantly and specifically increased in endothelial cells from the infarcted area at 24 h post-MCAO. Several studies carried out in different in vivo rat ischemic models (embolic, tMCAO, and photothrombotic [13,25,26]) also show increase of Cav-1 expression in endothelial cells of the ischemic hemisphere at 24 h and 48 h. According to the controversial role of Cav-1 in cerebral ischemia, the increased Cav-1 immunoreactivity could account for (1) the reported increase in density of caveolae in vascular segments showing BBB breakdown described in previous ultrastructural studies [27] or, as has been recently reported, (2) a protective role on the neurovascular unit [28]. The negative correlation between brain Cav-1 protein expression and hemorrhagic volume at 24 h post-MCAO detected in the present study could point towards a protective function of this protein in ischemic conditions.
With regard to rt-PA effects, our results show that Cav-1 expression in the infarcted area at 24 h does not differ between rt-PA treated and non-treated MCAO groups, and both groups have a similar trend to negatively correlate with the volume of hemorrhage at 24 h post-MCAO. To date only one study has previously analyzed t-PA effects on Cav-1 immunoreactivity in the infarcted area of MCAO animals, showing an enhancement of Cav-1 expression in the surviving endothelial cells [26]. The apparent discrepancies could be due to the different MCAO model applied and the time of t-PA administration (1 and 3 h later than in our study). Additionally, a direct relationship between Cav-1 expression and the cerebral edema in the focal ischemic brain has been described using Cav-1-deficient (Cav-1 −/− ) mice [25] but any study has published this association after MCAO with rt-PA treatment in Cav-1-non-deficient mice. In contrast with the expected anti-edema effects of Cav-1, our results showed that Cav-1 immunoreactivity and brain edema values were not correlated neither MCAO nor MCAO + rt-PA group. These current findings need to be confirmed in new studies since differences were observed regarding features of We know of no other studies exploring serum Cav-1 levels in ischemic conditions in animal models. Though Cav-1 plays a significant role in the pathogenesis of other relevant diseases such as cancer, a low number of animal studies and clinical trials have examined serum Cav-1 levels. Thus, the molecular function of the circulating concentrations of Cav-1 has remained uncertain. Importantly, our results demonstrate a positive correlation between Cav-1 expression in the infarcted area and serum after 24 h post-MCAO. These findings suggest that serum Cav-1 levels could reflect changes occurring at tissue level. The activation of MMPs, unleashed by the ischemic process, could be contributing to the release of Cav-1 by proteolysis as has been described for other cell signaling factors [29]. Moreover, it is notorious that baseline serum Cav-1 levels (3 h after ischemia induction) negatively correlated with hemorrhagic volume at 24 h. In accordance with this, a clinical report by our research group demonstrated that low serum Cav-1 levels are an independent predictor of HT after rt-PA administration [30]. In addition, Zhang et al. described that ischemic stroke patients with low serum Cav-1 levels have a 3-fold increased risk of cerebral microbleeds (CMBs) compared with patients with high Cav-1 level [31]. These data suggest a protective effect of Cav-1 in ischemic brain damage, and we could hypothesize that a lower Cav-1 response will be more prone to a worse outcome, independently of rt-PA administration. Therefore, an early analysis of serum Cav-1 levels could contribute to a better prediction of the development of HT.
bEnd.3 murine cells were used in order to explore rt-PA effects on endothelial Cav-1 expression in a longer timecourse than in the in vivo model. This cell line is a widely accepted in vitro BBB model [21,32] allowing to elucidate time-dependent molecular mechanisms associated with BBB breakdown after OGD and delayed rt-PA administration [20]. Although this represented a more simplified approach, it required less time and cost to evaluate Cav-1 changes from 3 up to 72 h after OGD and delayed rt-PA administration. To date, only one in vitro study has analyzed rt-PA effects on Cav-1 expression in brain endothelial cells in the context of ischemia [15] and, as far as we know, our study is the first to analyze the effects of OGD and rt-PA on Cav-1 phosphorylation levels. Our results show that OGD induces a significant increase of Cav-1 expression in bEnd.3 cells from 3 up to 72 h after the ischemic insult. Supporting this, unchanged total Cav-1 protein expression has been reported after 2 h of OGD [33] in bEnd.3 cells but a similar Cav-1 increase has been described in human brain microvascular endothelial cells (HBMECs) after 24 h of OGD [34]. In addition, our in vitro data show a significant reduction in total Cav-1 protein expression following rt-PA treatment. In agreement with our results, Song et al. have shown that rt-PA decreased Cav-1 protein expression in OGD-treated bEnd.3 cells by promoting their secretion to the culture medium after 2 h of OGD and 6 h of reoxygenation [15]. Our study adds that the rt-PA-induced Cav-1 decrease in bEnd.3 cells occurred beyond 3 h of reoxygenation and was maintained until 72 h, time-point where pCav-1 levels were also significantly higher than in rt-PA non-treated cells. The phosphorylation of Cav-1 at tyrosine 14 modulates caveolae formation and detachment from the plasma membrane, key processes for caveolae transcytosis [35][36][37]. Although in a different endothelial model (pulmonary cells), phosphorylation of Cav-1 has been reported to contribute to endothelial barrier disruption [38]. Accordingly, the inhibition of Cav-1 phosphorylation abrogated transcytosis in human brain endothelial cells [39]. Since the significantly increased pCav-1/ Cav-1 ratio coincided with a higher transcellular permeability in the OGD-rt-PA-treated cells at 72 h, it is plausible that the rt-PA effects on bEnd.3 cells under ischemic conditions could be mediated by the phosphorylation of Cav-1.
In summary, the present report has evaluated the effects of delayed rt-PA administration on Cav-1 expression in an in vivo and in an in vitro model of cerebral ischemia. Taken together, our results show that Cav-1 is overexpressed in endothelial cells due to ischemia but there is disagreement regarding rt-PA effects on Cav-1 expression between both experimental models. While rt-PA does not modify Cav-1 expression in the in vivo model, it significantly reduces Cav-1 expression and increases pCav-1/Cav-1 ratio in bEnd.3 cells. We suggest that the evident differences between the monoculture of bEnd.3 endothelial cells and the neurovascular unit, a complex tissue composed of diverse cell types, may account for the Cav-1 expression differences detected between the in vitro and in vivo results when rt-PA was administrated. In this way, astrocyte-endothelial interaction is crucial for BBB homeostasis and the astrocyte-derived fatty acid-binding protein 7 (FABP7) has been described as an endogenous protective response to BBB disruption partly mediated through upregulation of endothelial Cav-1 following traumatic brain injury [40]. Further studies are needed to confirm the present results using cocultures including endothelial cells and astrocytes. Importantly, the results obtained in the in vivo model support a protective role of Cav-1 and point to a potential usefulness of baseline serum Cav-1 levels to predict hemorrhagic volume after ischemic stroke. Further studies analyzing pCav-1 expression in the in vivo thromboembolic model could add relevant data to accurately define Cav-1 role during ischemic stroke.