Endothelial Sphingosine-1-Phosphate Receptor 4 Regulates Blood-Brain Barrier Permeability and Promotes a Homeostatic Endothelial Phenotype

Growth, purification, and treatment media

Full growth medium (MCDB_complete) contained MCDB131 glutamine-free (Invitrogen) supplemented with 20% FBS #7524, 1% glutamin solution 200 mm, 1% Na-heparin solution 1800 units/ml, 1% penicillin/streptomycin (all purchased from Sigma-Aldrich), 1% NaHCO₃ solution 100 mg/ml (ROTH) and 1% ECGS solution (lyophilized porcine brain powder, 5 mg/ml in buffer A). Purification medium (MCDB_puromycin) contained 5 μg puromycin/ml MCDB_complete. Treatment medium (MCDB_treatment) contained MCDB131 glutamine-free (Invitrogen) supplemented with 1% fatty-acid-free BSA, 1% glutamine solution 200 mm and 1% penicillin/streptomycin (all obtained from Sigma-Aldrich).

Isolation of mouse brain microvessels (MBMVs)

MBMVs were isolated as described previously (Gurnik et al., 2016). Briefly, mice were deeply anesthetized (180 mg/kg ketamine and 16 mg/kg xylazine) and transcardially perfused with sterile PBS for 10 min. The isolated brains were stored on ice in MV buffer (MVB: 15 mm HEPES, 5 mm glucose, 147 mm NaCl, 4 mm KCl, 3 mm CaCl_2, 1.2 mm MgCl₂, and 0.5% BSA, pH 7.4). Olfactory lobes, brain stem and cerebellum were removed and the remaining cortex/midbrain segment was rolled on autoclaved Whatman paper (Schleicher & Schuell) to remove the leptomeninges. In case of MBMV isolation post-transient middle cerebral artery occlusion (tMCAO; Fig. 4), ipsilateral and contralateral stroke hemispheres were separated. For MBMV isolation under homeostatic conditions (Figs. 1) both hemispheres were used. Brains were homogenized in MVB with a 0.025-mm clearance glass-Teflon Dounce homogenizer (Wheaton Sciences) propelled by an electric overhead stirrer (VOS14, VWR International) set to 2000 rpm. Myelin fat was removed via density gradient centrifugation in 25% BSA (35 min at 2000 × g), followed by pellet resuspension and filtration through a 100-μm cell strainer (Falcon). For MV recovery, the flow-through was filtered through a 40-μm cell strainer (Falcon). MBMVs were harvested from the nylon meshes for RNA sequencing and qPCR in the respective buffers.

Isolation and purification of primary MBMECs and porcine BMECs (PBMECs)

MBMECs were isolated in a similar fashion to MBMVs as described previously (Gurnik et al., 2016). In short, after the removal of olfactory lobes, brain stem, cerebellum and leptomeninges, brains were stored on ice in buffer A (15 mm HEPES, 153 mm NaCl, 5.6 mm KCl, 1.7 mm CaCl₂, 1.2 mm MgCl₂, and 1% BSA, pH 7.4). Multiple mouse brains were pooled and subsequently homogenized at 2000 rpm with the help of a Dounce homogenizer as described above. The homogenate was digested with 0.75% Collagenase Type II (Worthington) in buffer A (1:1:1 v/v/v, ∼3000 U Collagenase II per brain) on a thermo rocker for 1 h at 37°C, 600 rpm. Myelin was removed via density gradient centrifugation with 25% BSA for 35 min at 2000 × g and the remaining cell and vessel pellet was digested in buffer A with collagenase/dispase (Roche; 0.1 U collagenase and 0.8 U dispase per mouse brain) and DNase I (Worthington, 2U DNase per brain) in buffer A on a thermo rocker for 20 min at 37°C, 600 rpm. The cell suspension was centrifuged [400 × g, 4 min, room temperature (TR)] and the pellet was resuspended in MCDB_complete medium. PBMEC isolation was done analogous to MBMEC isolation, however pig brain leptomeninges had to be removed manually with surgical pincers. Homogenization was performed on ice at 2000 rpm with a 55 ml Dounce homogenizer (Wheaton Sciences, 0.025-mm clearance) driven by an electric overhead stirrer. Subsequent enzymatic digestions and BSA removal were performed analogous to the MBMEC isolation protocol, however enzyme concentrations were adapted to improve the yield. Collagenase Type II (Worthington) digestion of porcine brain homogenate was performed with 0.75% collagenase Type II solution in buffer A (2:1:2 v/v/v, ∼370 U Collagenase II per 1 g of porcine brain tissue). Collagenase/dispase (Roche) digestion was done with 0.01 U collagenase/g, 0.08 U dispase/g, and 0.3 U DNase I (Worthington)/g tissue. After the second enzymatic digestion, the porcine cell and vessel pellet was resuspended in MCDB_complete medium. Seeding of murine and porcine cell and MV suspensions in MCDB_complete medium was done in six-well plates (two wells per five mouse brains) or T75 cell culture flasks (four flasks per pig brain) precoated with Collagen Type I 150 μg/cm², respectively. BMEC extravasation from MVs and attachment to plate or flask surfaces was allowed for 12 h in MCDB_complete before the medium was changed to MCDB_puromycin for 48 h to remove nonendothelial cell types. MBMECs were seeded on CellZScope inserts immediately after puromycin treatment. PBMECs were stored deep frozen in liquid nitrogen and seeded on inserts immediately after re-thawing.

Transendothelial electrical resistance (TEER) measurements

One-micrometer pore diameter 24-well PET transwell inserts (Greiner Bio-One) were coated with 5 μg/cm² bovine fibronectin (bfn; Sigma-Aldrich) for 90 min before MBMEC/PBMEC seeding. The primary endothelial cells (passage 0) were trypsinized (MBMECs) or thawed (PBMECs), counted and seeded in the bfn-coated transwell inserts (100,000 MBMECs/cm² or 65,000 PBMECs/cm², respectively) in 270 μl MCDB_complete medium per insert. TEER was measured and recorded with the aid of a CellZScope device (NanoAnalytics). TEER and Ccl values were obtained from continuous impedance measurements (every 10 min) as described previously (Czupalla et al., 2014). TEER measurements were started immediately after BMEC seeding and the cells remained in MCDB_complete medium until barrier formation reached a plateau phase.

Pharmacological treatment

Before pharmacological treatment, all medium in the basal CellZScope compartment and 80% of the medium in the apical compartment was carefully removed (to maintain BMEC monolayer integrity) and exchanged with the appropriate treatment medium. Pharmacological agents were dissolved in fatty acid-free MCDB_treatment medium and the appropriate vehicle (or combination of different vehicles) for each agent (or combination of agents) was used to prepare the corresponding control medium. S1P (Tocris) was dissolved in PBS containing 1% fatty acid-free BSA. S1P4 agonist CYM50260 and S1P4 antagonist CYM50358 (both Tocris) were dissolved in DMSO and used at final DMSO concentrations of 1:1000, respectively. Pertussis toxin (Calbiochem) was diluted in PBS and used at a final concentration of 100 ng/ml. The final concentrations of all pharmacological agents used for each experiment are indicated in the corresponding figure legends.

Quantification of TEER measurement data

To compensate slight interinsert TEER differences during barrier formation, TEER values in representative TEER graphs and quantification figures are always shown as percentage of the TEER value measured before treatment (set to 100% for each individual insert). To obtain the electrical resistance values of the cell monolayer, the resistance value of an empty transwell insert was subtracted from each insert's TEER value semi-automatically by NanoAnalytics CellZScope software V.2.2.4. TEER values of all individual inserts of the different treatment groups were then normalized to the mean of the corresponding vehicle control values (of the same biological replicate) at each respective time point.

In vitro permeability tracer assay

Permeability tracer assays of pharmacologically treated MBMEC monolayers in transwell inserts were conducted 24 h posttreatment, a time point identified as significant by prior TEER analyses. In vitro permeability tracer assays were performed with a multi-size fluorescent dextran tracer mix as described previously (Czupalla et al., 2014). Briefly, different-size dextran tracers [TexasRed-3-kDa dextran (Invitrogen, Thermo Fisher Scientific), tetramethylrhodamine isothiocyanate-20-kDa dextran (Sigma-Aldrich), and fluorescein isothiocyanate-70-kDa dextran (Sigma-Aldrich)] were dissolved in (phenol red-free) DMEM medium (Invitrogen) at a concentration of 2 mm each (= 2× tracer mix); 24 h posttreatment, TEER measurements were stopped, and inserts were transferred to a 24-well plate containing 1-ml tracer-free DMEM/well. Half of the top compartment supernatant (135 μl) was removed from each insert and exchanged with 2× tracer mix (final top compartment tracer concentration: 1 mm each). A total of 300 μl tracer-free DMEM was collected as a control for the subtraction of DMEM background fluorescence. For the next 4 h, every hour, 100-μl medium was collected from the bottom compartment of each well. At 4 h, medium was also collected from the top compartment as normalization control. Fluorescence intensity of each sample was assessed with the help of a 384-well fluorescence Tecan infinite M200PRO plate reader and i-control 2.0 software (both Tecan) at excitation/emission wavelengths appropriate for all used tracers (595/625 nm for TR, 550/580 nm for TMR, and 490/525 nm for FITC, respectively). Background fluorescence of tracer-free DMEM was subtracted from all measured relative fluorescence unit (RFU) values and the relative fluorescence intensity of bottom chamber supernatant was normalized to the corresponding insert's top compartment fluorescence at 4 h. Inserts with compromised monolayer integrity were excluded from analysis. When comparing different pharmacological treatments and timepoints, the mean fluorescence intensity of the corresponding vehicle control condition at 1 h was set to 100% and the relative fluorescence of all other conditions and timepoints was calculated accordingly.

MBMEC transfection with S1P4 and ctrl siRNA

Primary pure MBMECs (48-h puromycin treatment) were grown to 70% confluency in collagen Type I (150 μg/cm²)-coated six-well plates and used for transfection with S1P4 siRNA and nontarget control siRNA, respectively. MCDB_complete medium was exchanged with MCDB131 (serum-free and antibiotic-free) 2 h before transfection. S1P4 knock-down was performed with RNAi-Mix consisting of four different QIAGEN Flexitube siRNAs (QIAGEN) targeting S1P4 (Mm_Edg6_2, FlexiTube siRNA #SI00989842; Mm_Edg6_3 FlexiTube siRNA#SI00989849; Mm_Edg6_4 FlexiTube siRNA #SI00989856 and Mm_S1pr4_1, FlexiTube siRNA #SI04918536, all 250 nm final). RNAi-Mix and nontarget control siRNA (QIAGEN, siRNA #1022076, 1 μm final) were used for transfection with RNAiMax lipofectamine (Thermo Fisher Scientific) according to the manufacturer's protocol. Briefly, siRNA oligos (a mix of 4 × 50 pmol S1P4 siRNAs or 200 pmol Ctrl siRNA, respectively) were prediluted in 100 μl MCDB131 and coincubated with lipofectamine solution (10 μl RNAiMax lipofectamine prediluted in 90 μl MCDB131) for 15 min at RT before the lipofectamine-siRNA-mix was added to the MBMECs. Six hours posttransfection, MCDB131 was exchanged with MCDB_complete, and cells were allowed to recover in full medium overnight. The next morning, MBMECs were trypsinized and cells were seeded on 24-well transwell inserts precoated with 5 μg/cm² bfn (100,000 MBMECs/cm² in 270 μl MCDB_complete). RNA isolation was performed 48 h posttransfection to ensure that S1P4 knock-down and related gene expression changes were still detectable at the time of harvesting. MBMEC barrier formation on CellZScope inserts was monitored for 96 h posttransfection (when all inserts had reached a TEER value plateau phase eventually).

RNA isolation and reverse transcription

MBMECs/PBMECs were harvested directly from transwell inserts in 350 μl RLTplus lysis buffer (three inserts/condition/350 μl RLTplus). RNA was isolated with a RNeasy Plus Microkit (QIAGEN) according to the manufacturer's protocol. Genomic DNA contaminants were digested on microkit columns during RNA isolation. RNA concentrations were measured via QuBit 4 fluorometer (Thermo Fisher Scientific). For RNA sequencing, RNA was frozen immediately after isolation and stored at −80°C until its transfer to the sequencing facility. For qPCR analysis, cDNA was generated from 200 ng RNA/sample using a RevertAid cDNA synthesis kit (Thermo Fisher Scientific) according to the manufacturer's protocol. cDNA was stored frozen at −20°C.

RNA sequencing analysis

RNA was transferred deep frozen (−80°C) to the mRNA sequencing facility at Max Planck Institute for Heart and Lung Research in Bad Nauheim, Germany. RNA quality was verified via LabChip Gx Touch 24 (PerkinElmer). mRNA library preparation with equal amounts of RNA from each sample (100 ng) was achieved with VAHTS Stranded mRNA-seq Library preparation kit (Vazyme) following the manufacturer's protocol. Sequencing was performed with the help of a NextSeq500 Sequencer with 75-bp single-end setup (Illumina). FastQC software (Babraham Bioinformatics) was used to verify signal quality, adaptor content and duplication rate. After adapter removal and data trimming by Trimmomatic version 0.38, reads of a length between 30 and 150 bp were accepted for further analysis. Trimmed reads were aligned to the murine mm10 genome (Ensembl) with Star 2.7.3a. The maximum ratio of mismatches to mapped length accepted was 10% and reads matching multiple genes or different genomic regions were excluded. The number of reads aligning to genes was counted with featureCounts 1.6.5 tool from the Subread package (Liao et al., 2014). Only reads mapping at least partially inside exons were admitted and aggregated per gene. Reads overlapping multiple genes or aligning to multiple regions were excluded.

TaqMan and SYBR qRT-PCR analysis

TaqMan gene expression assay was used for the comparison of S1PR expression levels between MBMVs and primary MBMECs (Fig. 1B,C). All other qPCR assays were performed with SYBR Green qRT-PCR. Taqman qPCR was run with an AB7500 fast system (Applied Biosystems) qPCR device. Taqman probes [CD31: Mm01242576, S1pr1: Mm02619656, S1pr2: AJVI4VC (customized), S1pr3: Mm02620181 and S1pr4: Mm00468695] were acquired from Life Technologies. Relative mRNA expression levels were calculated using the 2^-ΔΔCT method. For SYBR qRT-PCR, ABsolute SYBR GREEN fluorescein mix (Thermo Fisher Scientific) was used and RT-PCR was run with a CFX96 Real-Time C1000 thermal cycler. Data were analyzed and relative mRNA expression levels were calculated with Bio-Rad CFX manager Software 3.1 based on the 2^-ΔΔCT method. The amplification efficiency of murine and porcine qPCR primers was confirmed via relative qPCR standard curve method before experimental use. Murine qPCR primer sequences are listed in Table 1, porcine qPCR primer sequences in Table 2.

Liquid chromatography–m spectrometry/mass spectrometry analysis of S1P and S1P4 modulator concentrations in supernatant

After completion of MBMEC barrier formation, MCDB_complete medium in both apical and basal compartments of a transwell setup was replaced with MCDB_treatment medium containing 38 ng/ml (=100 nm) S1P. A total of 200-μl supernatant from each compartment were collected 2 h, 6 h and 24 h posttreatment. S1P concentrations in top and bottom supernatants were measured via LC-MS/MS; 200-μl supernatant sample was mixed with 200-μl extraction buffer (citric acid 30 mm, disodium hydrogen phosphate 40 mm) and 20-μl internal standard solution containing sphingosine-d7, sphinganine-d7 (200 ng/ml each), S1P-d7, and sphinganine-1-phosphate-d7 (400 ng/ml each), all from Avanti Polar Lipids. The mixture was extracted once with 1000 μl methanol/chloroform/hydrochloric acid (15:83:2, v/v/v). After evaporation of the organic phase, reconstitution was performed in 100 μl tetrahydrofuran/water (9:1, v/v) with 0.2% formic acid and 10 mm ammonium formate. LC-MS/MS analysis applying gradient conditions were performed using an Agilent 1100 LC-system (Agilent Technologies) coupled to an API4000 triple quadrupole MS (Sciex) with a Turbo V Ion Source operating in positive ESI MRM mode. A Luna C8 column (150 × 2 mm ID, 3-μm particle size, 100-Å pore size; Phenomenex) was used. The mobile phases consisted of water with 0.2% formic acid and 2 mm ammonium formate (mobile phase A) and acetonitrile/isopropanol/acetone (50:30:20, v/v/v) with 0.2% formic acid (mobile phase B). Acetonitrile with 0.1% formic acid was infused postcolumn using an isocratic pump at a flow rate of 0.15 ml/min. The internal standard method (isotope dilution mass spectrometry) was employed. Data acquisition and quantification was performed with Analyst software V1.6 and MultiQuant Software V 3.0 (both Sciex).

LC-MS/MS analysis of CYM50260 and CYM50358 was conducted with apical and basal supernatant of MBMECs in a transwell setup collected at 1, 8, and 24 h after basolateral treatment with either CYM50260 or CYM50358, respectively. Here, 20 µl of supernatant were mixed with 300 µl 2.5% ammonia in water and 20 µl internal standard containing 125 ng/µl 2.6-dibromo-3-[2,4-dichlorophenoxy)ethoxy-pyridine (Cayman Chemical). The mixture was extracted twice with 500 µl ethyl acetate. The extracts were combined and evaporated to dryness under a gentle stream of nitrogen at 45°C. The samples were reconstituted in 200 µl acetonitrile and injected on a Luna C18 150 × 2.0 mm ID, 3-µm particle size column (Phenomenex) using a 1200 Infinity Series LC-system (Agilent Technologies). The mobile phase consisted of water +0.1% formic acid (mobile phase A) and acetonitrile (mobile phase B). An API5500 QTRAP MS (Sciex) in positive ESI MRM mode was used for data acquisition and quantification using Analyst software V1.7.1 and MultiQuant Software V 3.0.2 (both Sciex).

Generation of luminal and abluminal bovine BMEC (BBMEC) membrane fragments

Luminal and abluminal BBMEC membrane fragments were generated as described previously (Devraj et al., 2011). Briefly, MVs were isolated from 10 bovine cortices and digested with collagenase Type IA (1800 units/g) to remove basement membrane, pericytes, and glial fragments. After MV homogenization, luminal and abluminal membrane fractions were separated by discontinuous ficoll gradient centrifugation (as described previously by Sánchez del Pino et al., 1995). BBMEC membrane fragments were stored deep frozen at −80°C and used for Western blot analysis immediately after thawing.

Immunoblotting of BBMEC membrane fragments, MBMECs, and mouse brain homogenate

Total sample protein concentrations were assessed with a BCA assay kit (Thermo Fisher Scientific #23227) according to the manufacturer's protocol. Following sonication, the solubilization of MBMECs and mouse brain homogenate was performed with Laemmli buffer [70 mm Tris-HCl, pH 6.8, 0.6% SDS, 2.5% (w/v) glycerol, 1.25% β-mercaptoethanol, and 0.01% BPB] for 4 min at 94°C. BBMEC membrane fraction samples were solubilized in SB buffer (2.3 m urea, 1.5% SDS, 50 mm Tris, pH 6.8, 25 mm TCEP, and 0.01% BPB) for 1 h at 30°C as described previously (Devraj et al., 2020). Equal amounts of protein were loaded into 10% Tris-HCl bis-acrylamide gels, protein concentrations are mentioned in each blot's corresponding figure legend. For BBMEC membrane immunoblotting, SDS-PAGE was run at a constant 80 V for 3 h and for MBMEC blotting, SDS-PAGE was run at a constant 60 V for 3.5–5 h. For mouse brain homogenate immunoblotting, SDS-PAGE was run at 100 V for 20 min initially, then voltage was increased to 140 V for about 1 h. All SDS-PAGE runs were conducted at room temperature. Proteins were subsequently either transferred to nitrocellulose membranes (Fig. 3B,F) of PVDF membranes (Figs. 3C, 4D). Nitrocellulose blotting was performed at 4°C overnight (36 V constant for 16 h). For blotting on PVDF, membranes were briefly activated with 100% methanol and protein was transferred (70 mA per gel for 1.4 h) at room temperature. Membrane blocking was done with 1× ROTI blocking solution (Carl ROTH) for nitrocellulose membranes or 2.5% dry milk in TBST for PVDF membranes, respectively, and antibody stainings were conducted as described previously (Simpson et al., 2001). All used antibodies are listed in Table 3.

MBMEC blots and BBMEC fragment blots were visualized with a Li-COR Odyssey FC Dual Mode Imaging System (Li-COR Biosciences) via use of the 700 nm and 800 m channels, respectively. Mouse brain homogenate blots were visualized after 1 min incubation with enhanced chemiluminescence solution (Pierce ECL Western, Thermo Scientific) and chemiluminescence was detected with light sensitive x-ray film (GE HealthcareHyperfilm ECL, GE Healthcare). To facilitate the detection of S1PR4 in brain homogenate despite low protein expression levels, SuperSignal West Femto Maximum Sensitivity Substrate (#34094, Thermo Scientific) was used. Relative protein expression of all blots was determined semi-quantitatively from band densitometry using LiCor ImageStudio version 3.1 software. Quantifications were conducted on raw images. To improve band visibility, brightness and contrast adjustments were subsequently applied uniformly to every pixel of the respective image using Affinity Designer software V.1.8.3.641 (Serif Ltd.). All adjustments were disclosed in the corresponding figure legends.

Animals

Porcine brains for PBMEC isolation were kindly provided by the trauma surgeons of the Goethe University Hospital neuropathology department, Frankfurt am Main (Germany) at the end of a myocardial surgical intervention so that no pigs had to be killed for the purpose of our experiments. C57BL/6J mice were bought from JAX (The Jackson Laboratory) via their European supplier Charles River. Global S1P4 knock-out (KO) mice were created as described previously (Golfier et al., 2010) and kindly provided to our group by Andreas Weigert and Bernhard Brüne (pathobiochemistry department, Goethe University Hospital). S1P4 KO and C57BL/6J mouse colonies were housed and bred at mfd Diagnostics and transferred to ZFE facility for organ collections and conduct of animal experiments. All animals were kept on a 12/12 h light/dark cycle with food and water ad libitum. Before experimental use, S1P4 KO mice were backcrossed to our C57BL/6J strain for 10 generations to minimize transcriptional variation based on genetic background differences. For tMCAO experiments, only male C57BL/6J mice (10 weeks of age) were used, whereas mice of both sexes (10–14 weeks) were used for MBMEC and MBMV isolation and in vivo S1P4 antagonist treatment. All animal experiments were conducted in accordance with the German Protection of Animals Act and were approved by the local animal welfare committee (RP Darmstadt, approvals number FU/1143, FU/1014, and FU/1202).

tMCAO

The transient occlusion of the right middle cerebral artery was conducted as described previously (Pfeilschifter et al., 2012). In short, male 10-week-old mice were anesthetized with 1.5% isoflurane under spontaneous respiration with buprenorphine analgesia (0.1 mg/kg). A midline cervical incision was done to expose the right carotid bifurcation and the right common and external carotid arteries were ligated. The middle cerebral artery was occluded for 60 min by advancing a standard silicone rubber-coated monofilament (6–0 medium, Doccol) up to the ostium of the middle cerebral artery. After 1 h, the filament was retracted and reperfusion was allowed. Sham-operated mice received the same surgical intervention, but the silicone filament was retracted immediately. Following the operation, mice were kept on heating mats (Beurer) with jelly food and water access on the cage floor. Animals were killed 24 h post-tMCAO.

In vivo S1P4 antagonist treatment and tracer permeability assay

As no published data about the in vivo dosage of the S1P4 antagonist used for this study was available, we based its dosage and the suggested time of application on reference values reported for FTY720 (Czech et al., 2009; Gao et al., 2017) and the TEER kinetics of our own in vitro experiments, respectively; 1 mg/kg body weight CYM50358 (Tocris) or the respective DMSO vehicle control was intraperitoneally injected into male and female 10- to 13-week-old wild-type (WT) and S1P4 KO mice twice (16 and 2 h before BBB permeability assessment, respectively). The animals appeared phenotypically normal and showed no signs of pain or neurologic impairment postinjection. In vivo BBB permeability tracer assay was performed as described previously (Vutukuri et al., 2018). In short, mice were deeply anesthetized (180 mg/kg ketamine and 16 mg/kg xylazine in 0.9% NaCl solution) and received a lateral tail vein injection with tetramethylrhodamine-conjugated 20-kDa dextran (TMR 20 kDa, Sigma-Aldrich) dissolved in PBS (2 mm TMR tracer, application volume 100 μl). Following 20 min of recirculation, 200-μl cardiac blood was collected from the atrium (normalization control) and mice were transcardially perfused with PBS for 5 min. Brain, lung and kidney were collected and homogenized (one kidney, one hemibrain free of olfactory lobes and cerebellum, or one lobe of the lung in 300-μl PBS, respectively). Following 15 min of centrifugation (10000 × g at 4°C), the fluorescence in organ supernatant and blood serum was measured with a 384-well fluorescence Tecan infinite M200PRO plate reader (Tecan) at 555/585 nm excitation/emission wavelength to assess TMR fluorescence intensity in RFUs. Sham animals (without tracer injection) were used to subtract organ supernatant autofluorescence values. Permeability index (ml/g) at the time points analyzed was calculated as the ratio of (tissue RFUs/g tissue weight) to (serum RFUs/ml serum).

Statistical analyses

Statistical analyses of TEER data, LC-MS/MS data, qPCR, and Western blotting expression data and tracer permeability data were performed using GraphPad Prism 5.0 software (GraphPad Software Inc.); p < 0.05 (*) was considered to be statistically significant, as further levels we denoted p < 0.01 (**) and p < 0.001 (***). Data in graphs are represented as mean ± SEM. The number of animals, cells, biological replicates, and statistical tests employed are described in each experiment's corresponding methodology illustration or figure legend, respectively. Statistical significance of mean S1PR expression level differences between MBMVs and MBMECs as well as between MBMECs from WT and S1P4 KO mice was assessed via unpaired two-tailed Student's t test. Statistical significance of TEER and tracer permeability quantifications as well as LC-MS/MS, qPCR, and Western blot analyses of MBMECs that underwent two different pharmacological treatments was assessed via paired two-tailed Student's t test. Statistical significance of TEER quantifications and qPCR analysis of S1P4 siRNA-treated and control siRNA-treated MBMECs was assessed via paired two-tailed Student's t test. To assess normal distribution of in vivo organ permeability index data, D'Agostino and Pearson omnibus normality test was used. Statistical significance of organ permeability index differences between S1P4 antagonist-treated and vehicle-treated mice was thus assessed via Mann–Whitney test for lung permeability indices and via unpaired two-tailed Student's t tests for brain and kidney permeability indices. For tMCAO-associated qPCR, RNA sequencing, and Western blot analyses, statistical significance between ipsilateral and contralateral hemispheres (of the same animals) was assessed via paired two-tailed Student's t tests, whereas statistical significance between hemispheres from tMCAO-operated and sham-operated animals was assessed via unpaired two-tailed Student's t tests. Statistical significance of TEER and tracer permeability quantifications as well as qPCR analyses from MBMECs that received multiple different pharmacological treatments was assessed via ANOVA. For the statistical analysis of RNA sequencing reads, differentially expressed genes were identified using pairwise comparisons with DESeq2 version 1.26.0 (Love et al., 2014). Genes were considered as significantly regulated if a logarithmic fold change of at least ± 0.585, a mean value higher than 5 and a p value lower than 0.05 were reached.

Graphics software

Methodology illustrations were generated with the help of BioRender. Figures were graphically designed, arranged, and labeled with Affinity Designer software V.1.8.3.641 (Serif Ltd.).