Methylphenidate-triggered ROS generation promotes caveolae-mediated transcytosis via Rac1 signaling and c-Src-dependent caveolin-1 phosphorylation in human brain endothelial cells

Abstract

Methylphenidate (MPH) is an amphetamine-like stimulant commonly prescribed for attention deficit hyperactivity disorder. Despite its widespread use, the cellular/molecular effects of MPH remain elusive. Here, we report a novel direct role of MPH on the regulation of macromolecular flux through human brain endothelial cells (ECs). MPH significantly increased caveolae-mediated transcytosis of horseradish peroxidase through ECs without affecting paracellular permeability. Using FRET-based live cell imaging, together with pharmacological inhibitors and lentiviral-mediated shRNA knockdown, we demonstrate that MPH promoted ROS generation via activation of Rac1-dependent NADPH oxidase (NOX) and c-Src activation at the plasma membrane. c-Src in turn was shown to mediate the phosphorylation of caveolin-1 (Cav1) on Tyr¹⁴ leading to enhanced caveolae formation and transendothelial transport. Accordingly, the inhibition of Cav1 phosphorylation by overexpression of a phosphodefective Cav1^Y14F mutant or knocking down Cav1 expression abrogated MPH-induced transcytosis. In addition, both vitamin C and inhibition of NOX blocked MPH-triggered vesicular transport. This study, therefore, identifies Rac1/NOX/c-Src-dependent signaling in MPH-induced increase in transendothelial permeability of brain endothelial cell monolayers via caveolae-mediated transcytosis.

Electronic supplementary material

The online version of this article (doi:10.1007/s00018-016-2301-3) contains supplementary material, which is available to authorized users.

Introduction

Methylphenidate (MPH) is the drug of choice for the treatment of attention deficit hyperactivity disorder (ADHD), a common neuropsychiatric disorder in children that can persist into adulthood [1]. Despite the well-known therapeutic effects of MPH on ADHD, several reports state a non-consensual diagnosis among physicians claiming that many children are being misdiagnosed with this disorder [2]. In addition, the consumption of stimulants to sustain attention, augment memory, and enhance intellectual capacity is increasing globally. In fact, the use of MPH for cognitive enhancement is a subject that has received much attention in recent years [1, 3]. Thus, it is crucial to better understand the overall impact of MPH on the central nervous system.

Endothelial cells (ECs) are the primary cellular component of the blood–brain barrier (BBB), a dynamic and highly specialized structure that plays a key role in brain homeostasis and protection [4]. Continuous non-fenestrated ECs with well-developed tight junctions (TJs) are responsible for maintaining high electrical resistance and thereby preventing the diffusion of small molecules between adjacent ECs of the BBB. In addition, the existence of selective transporters and a low rate of fluid-phase endocytosis, or (macro)pinocytosis, greatly limit the transport of large molecules through the brain endothelium [4]. An increase in transcellular transport is usually associated with brain injuries, such as traumatic brain injury and stroke [5]. Vesicular transport across brain ECs is primarily mediated by caveolae, which represent >95 % of EC vesicles [6]. Despite the deleterious role of caveolae in vascular and neurological diseases [7, 8], these vesicles can also be used for drug delivery [6, 9]. Caveolin-1 (Cav1) is the main structural component of caveolae which is thought to form an oligomeric coating on the cytoplasmic surface [10]. Cav1 also acts as an important scaffold protein that interacts with and modulates the activity of numerous signaling molecules [6]. Cav1 phosphorylation on tyrosine 14 by the Src family kinases (SFKs) is essential to initiate plasmalemmal vesicle formation, fission, and transendothelial vesicular transport [6, 11].

Vascular endothelial cell dysfunction has been described as a critical event in the development and progression of brain pathologies, such as Alzheimer’s disease [12]. In addition, the use of MPH seems to be associated with cardiovascular risks [1], increased brain–blood flow [13], and capillary wall structural changes, such as thickening of basement membrane and increased density of pinocytotic vesicles [14]. Nevertheless, the impact of MPH on brain ECs is not known. Herein, we report for the first time that MPH increases the permeability of human brain endothelial cell monolayers by stimulating caveolae-dependent vesicular transport. We reveal that MPH activates the GTPase Rac1 which promotes the assembly of NADPH oxidase (NOX) and increased production of reactive oxygen species (ROS). NOX-induced ROS generation activates c-Src which then phosphorylates Cav1, promoting transcytosis of macromolecules via caveolae-mediated transcellular transport in brain ECs.

Materials and methods

Cell cultures

Human samples were obtained from discarded temporal lobe tissue during operative treatment of epilepsy (outside epileptogenic foci) after informed consent and institutional review board ethical approval at the Neurosurgery Service, Coimbra Hospital and University Centre, Portugal. Primary cultures of human brain microvascular endothelial cells (HBMVECs) were isolated as previously described [15]. HBMVECs were maintained in Dulbecco’s modified Eagle’s medium/nutrient mixture F-12 (DMEM/F-12, Biochrom AG, Berlin, Germany) media containing 10 % fetal bovine serum (FBS, GIBCO, Rockville, MD, USA), endothelial cell growth supplement (ECGS, BD Biosciences, Franklin Lakes, NJ, USA), heparin (1 mg/mL, Biochrom AG), and 1 % antibiotic–antimycotic (Sigma-Aldrich, St. Louis, MO, USA). After reaching confluence, ECGS and heparin were removed.

The human cerebral microvascular endothelial cell line (hCMEC/D3) shows similar properties to primary HBMVECs and is routinely used as an in vitro model of the BBB [16]. This cell line was cultured in EBM-2 medium (Lonza, Walkersville, MD, USA) supplemented with 1 ng/mL basic fibroblast growth factor (bFGF, Sigma-Aldrich), chemically defined lipid concentrate (1:100; Invitrogen), 1.4 µM hydrocortisone (Sigma-Aldrich), 5 µg/mL acid ascorbic (Sigma-Aldrich), 1 % Penicillin–Streptomycin (Gibco, Paisley, UK), 5 % FBS (Invitrogen, Inchinnan Business Park, UK), and 10 mM HEPES (Lonza).

Medium was changed every 2 days until the cells reached confluence. HBMVECs and hCMEC/D3 (passage 26–35) were seeded on culture plates coated with collagen type I (R&D Systems, Inc., Minneapolis, USA) and maintained at 37 °C with 5 % CO₂. ECs were either left untreated (control) or treated for different time points, as specified in the figure legends.

Plasmids

pUSE-Src-YF-UniRapR-mCerulean-myc (hereafter termed c-Src CFP; Plasmid 45381) and pFRET-HSP33 (plasmid 12260) were obtained from Addgene. Cells transfection was performed using 1–2 µg of each plasmid with jetPRIME^® (Polyplus transfection, Illkirch-Graffenstaden, France) according to the manufacturer’s protocol. c-Src Mission shRNA clone TRCN0000023597, Rac1 Mission shRNA clone TRCN0000055188, Cdc42 Mission shRNA clone TRCN0000071686, RhoA Mission shRNA clone TRCN0000068198, and Cav1 Mission shRNA clone TRCN0000008002 were from Sigma-Aldrich.

Lentivirus production

The production was similar to that described previously [17]. Briefly, HEK293T cells were seeded in 90 mm culture dishes and, at 80 % confluence, were co-transfected overnight with virus-producing plasmids using jetPRIME^® (Polyplus transfection) according to the manufacturer’s instructions. Transfection ratios were as follows: 6 µg of shRNA plasmids to 3 µg of psPAX2 to 3 µg of VSVG (2:1:1). The next day, normal growth media were replaced by transfection media, and cells were cultivated for an additional 48 h. Then, media with viral particles were collected, centrifuged at 1500 rpm for 5 min, and the supernatant was collected into new tubes.

TUNEL assay

The protocol was performed as previously described [18]. Briefly, HBMVECs and hCMEC/D3 were fixed with 4 % paraformaldehyde (PFA; Sigma-Aldrich), permeabilized in 0.25 % Triton X-100 for 30 min at room temperature (RT), and incubated with terminal deoxynucleotidyl transferase buffer for 1 h at 37 °C in a humidified chamber. Incubation with fluorescein (1:100; Vector Laboratories, Burlingame, CA, USA) was performed for 1 h, followed by nuclei counterstaining with 5 µg/mL Hoechst 33342 (Sigma-Aldrich) for 5 min. The slides were mounted in Dako fluorescent medium (Dako North America Inc., Carpinteria, CA, USA) and fluorescent images for cell counts were recorded using an Axiovert 200 M fluorescence microscope (Carl Zeiss, Oberkochen, Germany).

Evaluation of endothelial cell monolayer integrity

Cell monolayer integrity was determined as previously described [19]. Briefly, HBMVECs were grown on collagen type I-coated 12-mm Transwell filters (Costar, Corning, NY, USA) and sodium fluorescein (376 Da Na-F; 10 μg/mL), fluorescein isothiocyanate (4 kDa FITC; 1 mg/mL), or rhodamine B isothiocyanate dextran (70 kDa RITC; 1 mg/mL; all from Sigma-Aldrich) were added to the apical side. Samples (50 μL) were collected from the basal chamber at 30 min intervals for 120 and 240 min before and after treatments, respectively. Fluorescence was measured and plotted against time. Permeability was determined from linear slope changes before and after the addition of the compounds.

Transendothelial electrical resistance (TEER) of monocultures was measured using an STX-2 electrode coupled to an EVOM resistance meter (World Precision Instruments, Hertfordshire, UK). TEER readings of cell-free inserts were subtracted from the values obtained with cells, and results were expressed as % of control.

Immunocytochemistry

For zonula occludens (ZO-1) and β-catenin identification, HBMVECs were fixed with 4 % PFA (Sigma-Aldrich) for 20 min at RT, whereas for claudin-5 staining, cells were fixed with methanol for 15 min at −20 °C. Afterwards, cells were permeabilized with 0.2 or 0.1 % Triton X-100 (Sigma-Aldrich) and blocked with 3 % bovine serum albumin (BSA; Sigma-Aldrich) for 1 h at RT or with 5 % BSA (Sigma-Aldrich) for 45 min at RT for β-catenin and ZO-1 or claudin-5, respectively. Incubation with primary antibodies was performed as follows: rabbit ZO-1 (1:200, Invitrogen) and rabbit β-catenin (1:100, Invitrogen) for 1 h at 37 °C; and mouse claudin-5 (1:50, Invitrogen) overnight at 4 °C. Cells were incubated with secondary antibodies (Alexa Fluor 488 or Alexa Fluor 594, 1:200, Invitrogen) for 1 h at RT in the dark, and nuclei were stained with 4 μg/mL Hoechst 33342 (Sigma-Aldrich) for 5 min at RT. Finally, cells were mounted in Dako fluorescence medium (Dako North America), and images were recorded using an LSM 710 Meta Confocal microscope (Carl Zeiss, Oberkochen, Germany).

The analysis of claudin-5 immunofluorescence images was performed using the FIJI Software version 2.0 as previously described [20]. A total of three independent cultures were used, and from each coverslip, six images were blindly captured and analyzed.

Horseradish peroxidase transport

HBMVECs and hCMEC/D3 were seeded on collagen type I-coated collagen 12-mm Transwell filters (Costar) and grown to confluence. Horseradish peroxide (10 mg/mL HRP, Sigma-Aldrich) was added to the apical chamber, and after 3 h of incubation, the top and the bottom of the Transwells were washed and new medium was added. Samples were taken for 2 h at 30 min intervals. Absorbance was measured in a microplate reader (Biotek, Synergy HT, Winooski, USA), HRP activity was plotted against time, and transport rates determined by linear regression.

Transmission electron microscopy analysis

HBMVECs were manipulated as mentioned for HRP transport assay and then washed with 0.01 M phosphate-buffered saline (PBS: 137 mM sodium chloride, 2.7 mM potassium chloride, 4.3 mM disodium hydrogen phosphate, and 1.47 mM monopotassium dihydrogen phosphate, pH 7.4) followed by fixation in 2 % PFA and 2 % glutaraldehyde in 0.05 M cacodylate buffer for 30 min. For the diaminobenzide (DAB) reaction, cells were incubated with 0.05 % DAB and 0.015 % hydrogen peroxide (H₂O₂) in the dark for 10 min at RT, washed with PBS, and incubated with 1.5 % potassium ferricyanide and 1 % osmium tetroxide (all from Sigma-Aldrich) for 1 h in the dark at 4 °C. Afterwards, cells were dehydrated in a graded ethanol series (50–100 %), and impregnated using an epoxy embedding kit (Fluka Analytical, USA). Ultrathin sections (70 nm) were mounted on copper grids with no post-staining and observations carried out on an FEI-Tecnai G2 Spirit Bio Twin at 100 kV.

Western blot analysis

Experiments were performed as previously described [18]. Primary antibodies were as follows: rabbit anti-caveolin-1 (1:200, Santa Cruz Biotechnology Inc.), rabbit anti-p-Cav1 Tyr¹⁴ (1:50, Santa Cruz Biotechnology Inc.), mouse anti-Rac1 (1:500, Cytoskeleton, Inc., Denver, CO, USA), mouse anti-Cdc42 (1:500, Cytoskeleton, Inc), mouse anti-RhoA (1:500, Cytoskeleton, Inc.), rabbit anti-p-c-Src Tyr⁴¹⁶ (1:500, Cell Signaling Technology, USA), and rabbit anti-c-Src (1:500, Abcam, Cambridge, UK). Then, membranes were then probed with alkaline phosphatase-conjugated secondary antibodies; anti-rabbit (1:20,000) and anti-mouse (1:10,000) (Amersham GE Healthcare Life Science, USA). Immunoblots were stripped and reprobed with an antibody against glyceraldehyde 3-phosphate dehydrogenase (GAPDH, 1:500; Abcam, Cambridge, UK) to ensure equal sample loading. Bands were visualized using the enhanced chemifluorescence (ECF) reagent assay (Amersham) on the Typhoon FLA 9000 (GE Healthcare Bioscience AB, Uppsala, Sweden), and quantification was performed using the ImageJ 1.44o software.

Fluorescence resonance energy transfer (FRET) and image analysis

The immortalized hCMEC/D3 cell line was plated on glass-bottom culture dishes (µ-Dish 35 mm, iBidi). As previously described [17], imaging was performed using an inverted epifluorescence microscope (DMI6000B, Leica Microsystems). The excitation light source was a mercury metal halide bulb with an integrated light attenuator (EL6000, Leica Microsystems). High-speed low vibration external excitation and emission filter wheels (equipped with CFP/YFP excitation and emission filters) were mounted on the microscope (Fast Filter Wheels, Leica Microsystems). A 440–520 nm dichroic mirror (CG1, Leica Microsystems) and a PlanApo 63 × 1.3 N.A glycerol immersion objective were used. Images were acquired with 4 × 4 binning using a digital CMOS camera (ORCA-Flash4.0, Hamamatsu Photonics). All modules were controlled by the LAS AF software (Leica Microsystems). At each time-point, CFP and FRET images were sequentially acquired using different combinations of filters (YFP excitation plus CFP emission and CFP excitation plus YFP emission, respectively). Images were exported as 32 bit tiff files and processed using the FIJI software. Background was dynamically subtracted from all slices from both channels using a routine macro, and images were filtered using the second momentum of a mean filter in FIJI. Segmentation was achieved on a pixel-by-pixel basis. After background subtraction and filtering, ratiometric images (CFP/FRET or FRET/CFP) were generated in intensity modulated display mode using the FRET images as intensity modulators.

RhoA/Rac1/Cdc42 pull-down assay

Rho GTPase activation was assessed by a pull-down assay kit (Cytoskeleton, Inc., Denver, USA) according to the manufacturer’s protocol. In brief, hCMEC/D3 lysates were collected and GTP-bound RhoA or Rac1/Cdc42 was captured using pull-down assays with immobilized Rhotekin-RBD or PAK-PBD, respectively. The levels of activated small GTPases, as well as total amount of GTPase pulled down, were evaluated by western blot analysis using specific antibodies (1:500).

Reactive oxygen species detection

HBMVECs were cultured in coated black 96-well plates and after treatments, 2′,7′-dichlorodihydrofluorescein diacetate (5 µM H₂DCFDA; Molecular Probes, Eugene, Oregon, USA) or dihydroethidium (2 µM DHE; Molecular Probes) was added to the cells during 1 h at 37 °C in the dark. The fluorescence intensity was measured (DHE, ex/em: 485/590 nm; H₂DCFDA, ex/em: 485/528 nm) and divided by the amount of protein.

Statistical analysis

Evaluation of EC monolayer integrity, HRP transport, claudin-5 immunofluorescence images, and western blot analysis were performed by a person blinded to treatments. Results are expressed as mean ± standard error of the mean (SEM). Data were analyzed using one-way ANOVA followed by Dunnett’s or Bonferroni’s post hoc test, or Mann–Whitney test using GraphPad Prism 5.0 (GraphPad Software, San Diego, CA, USA). The level of significance was P < 0.05, and the “n” represents the total number of experiments obtained from at least three independent cell cultures.

Results

MPH increases human brain endothelial cell permeability via caveolae-dependent transcytosis

To investigate the impact of MPH on the integrity of the brain endothelial barrier, we measured the macromolecular flux of fluorescent dyes across confluent HBMVECs. For these studies, 100 μM of MPH was used, since this concentration is within the range of the recommended daily dosage prescribed for ADHD [21]. In addition, studies with animal models have suggested that due to an active accumulation process, brain concentrations of MPH are substantially higher than those found in the plasma [22] and are usually between 23.3 and 242 μM [23, 24].

In the present study, we first demonstrated that MPH (100 μM) did not cause EC death when applied to either the primary cultures (Supplementary Fig. 1a, b) or cell line (Supplementary Fig. 1c). Nevertheless, it increased 376 Da Na-F and 4 kDa FITC-dextran flux across HBMVECs (Fig. 1a, b) without changing the permeability to 70 kDa rhodamine B (Fig. 1c). It should be noted that larger tracers are excluded from the paracellular pathway if intercellular junctions are intact [25, 26]. In addition, despite the MPH-induced increase in endothelial permeability to small molecular weight tracers, there were no alterations in TEER values (Fig. 1d) suggesting that hyperpermeability promoted by MPH was not due to alterations in the paracellular permeability pathway. Accordingly, no changes were observed in the expression and organization of inter-endothelial junctions, specifically in both tight and adherens junction proteins claudin-5 (Fig. 1e), ZO-1, and β-catenin, respectively (Fig. 1f). To support our hypothesis, we further investigated the impact of MPH on tight junction organization by morphometric analysis of claudin-5 staining. Our results show that MPH did not alter the membrane/cytoplasm ratio of claudin-5 localization compared to control (Fig. 1e).

Fig. 1.

MPH increases the permeability of human brain ECs. Macromolecular flux across HBMVECs was assessed using a 376 Da Na-F, b 4 kDa FITC-dextran, or c 70 kDa rhodamine B-dextran at different time points after MPH exposure (100 μM). d–f MPH altered neither d transendothelial electrical resistance (TEER) nor e, f tight and adherens junction protein expression/organization. e Representative confocal images of claudin-5 immunofluorescence and plot intensity profile that enables the calculation of the ratio between membrane and cytoplasm pixel intensity. Scale bar 20 μm. f Representative confocal images of ZO-1 and β-catenin immunoreactivity. Nuclei were stained with Hoechst 33342. Scale bar 50 μm. The results are expressed as mean ± SEM, n = 4–12. ***P < 0.001 vs. control (dashed line)

We next explored whether MPH-induced increase in macromolecular flux was due to transcytosis. We observed endothelial membrane invaginations by electron microscopy (Fig. 2a) suggesting that MPH may have increased unidirectional and non-junctional HRP flux across HBMVECs via a vesicular pathway (Fig. 2b). To further identify the type of vesicles formed upon MPH exposure, HBMVECs were treated with either caveolae or clathrin disrupting agents. We used methyl‐β‐cyclodextrin (5 mM M-β-C) or hypertonic sucrose (0.4 M) to inhibit caveloae- or clathrin-mediated endocytosis, respectively [27]. Our results showed that only the depletion of caveolar plasmalemmal vesicles blocked MPH-induced transcytosis in HBMVECs (Fig. 2c). Given that Cav1 is the major protein constituent of caveolae, we then demonstrated that Cav1 knockdown in hCMEC/D3 cells (validated in Fig. 2d) abolished MPH-induced HRP transport (Fig. 2e). In addition, MPH increased Cav1 phosphorylation at Tyr¹⁴ (p-Cav1^Y14), which is required for caveolae-mediated endocytosis [11, 28] in a time-dependent manner in both HBMVECs (Fig. 2f) and hCMEC/D3 cells (Fig. 2g). The phosphorylation kinetics triggered by MPH has never been shown, and our results are in accordance with other studies that investigated different stimuli, including the psychostimulant methamphetamine [29–31]. In addition, MPH was not able to induce Cav1 phosphorylation in ECs expressing the phosphodefective Cav1 mutant (Cav1^Y14F; Fig. 2g). To specifically address the role of p-Cav1 in MPH-induced vesicular transport of HRP, we used hCMEC/D3 cells overexpressing either phosphomimicking (Cav1^Y14D) or phosphodefective (Cav1^Y14F) Cav1 mutants (Fig. 2h). While MPH-induced increase in HRP transport was blocked by Cav1^Y14F overexpression (Fig. 2h), Cav1^Y14D by itself was sufficient to increase HRP transport, mimicking the MPH effect. Taken together, our data show that MPH promotes transcytosis across human ECs through a caveolae-dependent process.

Fig. 2.

MPH promotes caveolae-mediated transcytosis in human brain ECs. a Representative transmission electron micrographs of HBMVECs under control conditions (CTR) or 3 h after MPH exposure (100 µM) showing the formation of caveolar structures (outlined in red). Scale bar 2 μm. b MPH enhanced vesicular transport of HRP (44 kDa) in HBMVECs, n = 16–18. c Disruption of caveolae with methyl-β-cyclodextrin (5 mM M-β-C) blocked the effect of MPH in HBMVECs, whereas blockade of clathrin-coated vesicles formation with sucrose (0.4 M) had no effect, n = 7–18. d Quantification of Cav1 protein levels in hCMEC/D3 cells after shRNA-mediated knockdown or infection with empty vector (pLKO). Top representative western blot image of Cav1 (22 kDa) and GAPDH (37 kDa) is shown, n = 4–6. e In the absence of Cav1 (shCav1), MPH was not able to promote HRP transport in hCMEC/D3 cells, n = 3–12. f, g MPH increased the phosphorylation of Cav1 Tyr¹⁴ (p-Cav1^Y14) in both f HBMVECs and g hCMEC/D3 cells. Expression of phosphodefective Cav1^Y14F abolished MPH-induced phosphorylation (analyzed after 120 min). Top representative western blot images of p-Cav1^Y14 and Cav1 (22 kDa) are shown, n = 12–14. h Phosphomimicking Cav1^Y14D mutant is sufficient to increase HRP vesicular transport, similar to that observed with MPH, whereas phosphodefective Cav1^Y14F mutant blocked the effect of MPH, n = 11–33. The results are expressed as mean ± SEM. **P < 0.01, ***P < 0.001 vs. control (CTR or dashed line) or pLKO; ⁺⁺⁺ P < 0.001 vs. MPH

ROS production via Rac1/NOX activation is a key mediator of MPH-induced vesicular transport

The Rho family of small GTPases regulates many aspects of actin cytoskeletal dynamics, including vesicle trafficking [32, 33]. Thus, we further focused on the impact of MPH on Rho GTPase activity in brain ECs by studying the dynamic activity of Rho family members (RhoA, Rac1, and Cdc42) using live cell imaging with FRET-based Rho biosensors in hCMEC/D3 transfected cells. MPH induced robust Rac1 activation at membrane ruffles (Fig. 3a), as detected by a significant increase in the FRET/CFP ratio of the Raichu Rac1 FRET probe [34] (Fig. 3a, b, red circles). On the contrary, there was a decrease in RhoA activity over time (Fig. 3a, b, green circles) and no alteration of Cdc42 activity (Fig. 3a, b, blue circles). These observations were confirmed by pull-down assay in hCMEC/D3 cells (Supplementary Fig. 2a–c). Moreover, a causal link between Rac1 activation and MPH-induced EC permeability was further demonstrated in that knockdown of Rac1 in hCMEC/D3 cells with shRNA (validated in Fig. 3c) prevented the increase in MPH-triggered vesicular transport of HRP (Fig. 3f). However, knockdown of RhoA or Cdc42 (validated in Fig. 3d, e) had no effect on the MPH-induced response (Fig. 3g, h).

Fig. 3.

Impact of MPH on Rac1, RhoA, and Cdc42 GTPase activity and their involvement in vesicular transport. a Representative FRET/CFP ratio images of hCMEC/D3 cells expressing Rac1, RhoA, and Cdc42 biosensors after MPH (100 μM) exposure at the indicated time points and coded according to a pseudocolor scale. Scale bar 40 μm. b MPH increased Rac1 activity (red circles), decreased RhoA (green circles), and had no effect on Cdc42 (blue circles) activity. The mean FRET/CFP emission ratios were normalized with a reference value acquired from unstimulated cells. Symbols represent the mean ± SEM, n = 4–7 cells. c–e Quantification of c Rac1, d RhoA, and e Cdc42 protein levels in hCMEC/D3 cells after shRNA-mediated knockdown or infection with empty vector (pLKO or pSicoR-DsRed). Top representative western blot images of the Rho GTPases (22 kDa) and GAPDH (37 kDa), n = 4–6. f–h hCMEC/D3 cells infected with lentiviral shRNA clone to f Rac1, g RhoA, and h Cdc42 indicate that Rac1 is required for MPH-increased caveolar transport. pLKO and pScicoR.DsRed were the respective empty vectors. The results are expressed as mean ± SEM, n = 6–8. ***P < 0.001 vs. pLKO, pSicoR.DsRed, or control (dashed line); ⁺⁺⁺ P < 0.001 vs. MPH + pLKO

Previous reports have demonstrated that MPH increases oxidative stress in the brain [35, 36], and oxidant signaling in ECs has been linked to caveolae-mediated transcytosis [28, 37]. Thus, we assessed cellular redox responses to MPH in live ECs using a redox-sensitive FRET biosensor [heat shock protein 33 fluorescence resonance energy transfer (HSP-FRET)] [17]. We observed that MPH increased ROS generation in hCMEC/D3 cells, as detected by an increase in the CFP/FRET ratio using the HSP sensor (Fig. 4a, b, green circle). Interestingly, the antioxidant vitamin C (80 μM VitC) attenuated ROS production triggered by MPH (Fig. 4a, b blue circles). Since it is known that the NOX complex plays a key role in vascular ROS production [38], we further showed that a specific NOX inhibitor, VAS2870 (5 μM VAS) [39] blocked ROS generation triggered by MPH (Fig. 4a, b, black circles). In addition, taking into consideration that the small GTPase Rac1 mediates ligand-dependent ROS generation by NOX assembly in ECs [40], we further evaluated the potential involvement of Rac1 activation in oxidative events associated with MPH exposure by conducting FRET studies with the HSP sensor in ECs after knocking down Rac1. We observed that MPH-induced ROS production was largely attenuated upon shRNA-mediated Rac1 knockdown (Fig. 4a, b, orange circles). As proof of concept, we also showed that ROS production was reestablished in cells transfected with the empty vector (transduction control) of Rac1 (pLKO; Fig. 4a, b, yellow circles), and hydrogen peroxide (H₂O₂) was used as a positive control (Fig. 4a, b, red circles). Taken together, these data indicate that MPH increases ROS generation in brain ECs via the Rac1/NOX signaling pathway.

Fig. 4.

MPH increases the production of intracellular ROS by activation of Rac1/NOX complex culminating in enhanced caveolae-mediated transcytosis in brain ECs. a, b hCMEC/D3 cells expressing ROS FRET probe (HSP) were treated with MPH (100 μM; green circles) alone and in the presence of the antioxidant vitamin C (80 µM VitC; blue circles) or NOX inhibitor VAS2870 (5 μM VAS; black circles). To specifically identify the role of Rac1 on MPH-induced ROS production, hCMEC/D3 cells were infected with lentivirus carrying pLKO (empty vector; yellow circles) or Rac1 shRNA (orange circles, and then transfected with HSP probe. H₂O₂ (500 μM; red circles) was used as a positive control. a Representative ratio images of hCMEC/D3 cells show time-lapse CFP/FRET color-coded according to the pseudocolor ramp. Scale bar 40 μm. b CFP/FRET emission ratios of the chimera over time were normalized with a reference value acquired from unstimulated cells. Symbols represent the mean ± SEM, n = 5–18 cells. c MPH increased the production of ROS (DCF dye), namely, d superoxide anion (DHE dye), n = 16–48. e Increase in HRP transport mediated by MPH in HBMVECs was prevented by the antioxidant VitC or NOX inhibitor VAS. H₂O₂ (500 µM) was used as a positive control, n = 5–18. Results are expressed as mean ± SEM. *P < 0.05, ***P < 0.001 vs. control (dashed line), ⁺⁺⁺ P < 0.001 vs. MPH

Similar to what was demonstrated with the hCMEC/D3 cell line (Fig. 4a, b), we observed that VitC also prevented MPH-induced ROS formation in HBMVECs (Fig. 4c). The use of DHE, which is specifically oxidized by superoxide anion, also allowed us to identify an increase in superoxide anion content after MPH exposure (Fig. 4d). As expected, H₂O₂ (positive control) robustly increased ROS generation (Fig. 4c, d). We then explored the possibility that ROS participated in MPH-induced increase of vesicular transport. In fact, both VitC and VAS blocked HRP transport promoted by MPH (Fig. 4e). These results demonstrate that MPH-induced Rac1/NOX-dependent ROS generation culminates in vesicular transport in human brain ECs.

Oxidant signaling triggered by MPH activates c-Src

ROS have been described as regulators of c-Src activity via oxidization [41]. Thus, we asked whether MPH-mediated oxidant signaling could activate c-Src in human brain ECs. We first showed that MPH significantly increased c-Src phosphorylation (p-c-Src^Y416) in both HBMVECs (Fig. 5a) and hCMEC/D3 cells (Fig. 5b). Then, tyrosine kinase activity of c-Src was also evaluated at the plasma membrane using a specific c-Src FRET-based biosensor (KRas Src YPet) which does not recognize the activity of other SFKs [34]. From these studies, we conclude that MPH induced a rapid and sustained increase in c-Src activity in live hCMEC/D3 cells, as demonstrated by an increase in CFP/FRET emission ratio of the KRas Src YPet probe (Fig. 5c, d, green circles). This effect was blocked by SKI-1 (100 nM SKI, 1 h pretreatment; Fig. 5c, d, cyan circles). To further confirm that MPH specifically activates c-Src, we used the negative control KRas Src (RV) construct [34] and showed that MPH had no effect on FRET in ECs under the same conditions (Fig. 5c, d purple circles).

Fig. 5.

ROS generation by Rac1/NOX activation is required for MPH-mediated c-Src activation in human brain ECs. a, b MPH induced c-Src activation, as shown by the increased phosphorylation at Tyr⁴¹⁶ (p-c-Src) in both a HBMVECs and b hCMEC/D3 cells. Top representative western blot images of c-Src Tyr⁴¹⁶ (60 kDa) and total c-Src are shown. Results are expressed as mean ± SEM. **P < 0.01, ***P < 0.001 vs. control (CTR), n = 10–14. c hCMEC/D3 cells expressing c-Src FRET probe (KRas Src YPet) or its negative control (Kras Src (RV) YPet; purple circles) were exposed to MPH (100 μM; green circles) alone or in the presence of SKI-1 (100 nM SKI; cyan circles), vitamin C (80 µM VitC; blue circles) or VAS2870 (5 μM VAS; black circles). Other cells were infected with lentivirus carrying pLKO (control empty vector; yellow circles) or Rac1 shRNA (shRac1; orange circles). H₂O₂ (500 μM; red circles) was used as positive stimulus to trigger ROS production. Representative ratio images show time-lapse CFP/FRET color-coded according to the pseudocolor ramp. Scale bar 40 μm. d Time course of CFP/FRET emission ratios of c-Src activity in hCMEC/D3 cells under the same experimental conditions as mentioned previously. The results were normalized with a reference value acquired from unstimulated cells. Symbols represent the mean ± SEM, n = 5–18 cells

The antioxidant VitC (Fig. 5c, d, blue circles) or NOX inhibition with VAS (Fig. 5a, b, black circles) blocked MPH-mediated increase of c-Src activation. Since it was previously shown that Rac1 plays a key role in redox-sensitive signal transduction in ECs [38, 40], we assessed MPH-induced activation of c-Src following Rac1 knockdown in hCMEC/D3 cells. As shown in Fig. 5, Rac1 shRNA decreased Src activity (Fig. 5b, c, orange circles) compared to control cells (pLKO empty vector; Fig. 5b, c, yellow circles). ROS-mediated c-Src activation was confirmed by stimulating hCMEC/D3 cells with H₂O₂ (Fig. 5c, b, red circles). We conclude that MPH activates c-Src via Rac1/NOX-dependent ROS generation in human brain ECs.

MPH-mediated c-Src activation leads to Cav1 Tyr¹⁴ phosphorylation and transendothelial hyperpermeability

Cav1 phosphorylation has been implicated in the mechanism of oxidative stress-induced transcellular transport, and it is a well-known substrate for c-Src [25]. Thus, the role of c-Src in MPH-induced vesicular transport was investigated. We observed that MPH-triggered HRP hyperpermeability of HBMVECs was prevented by c-Src inhibition by SKI (100 nM; Fig. 6a) and that shRNA-mediated c-Src knockdown (validated in Fig. 6b) had the same effect in hCMEC/D3 cells (Fig. 6c). In addition, pharmacological inhibition or c-Src knockdown blocked MPH-induced Cav1 phosphorylation (at 120 min) in both HBMVECs (Fig. 6d) and hCMEC/D3 cells (Fig. 6e), respectively. To further clarify whether MPH-induced transcytosis was dependent on Cav1 phosphorylation by c-Src, hCMEC/D3 cells were co-transfected with a rapamycin (Rap)-inducible c-Src heterodimerization chimera (RapR-Src) [17], and/or with Cav1^Y14F. We observed that c-Src activation with Rap in RapR-Src-transfected ECs was sufficient to trigger HRP transport, whereas the overexpression of Cav1^Y14F prevented RapR-Src-induced HRP transport (Fig. 6f). c-Src and Cav1 interaction assessed by FRET in hCMEC/D3 cells transfected with Cav1-YFP and RapR-Src-CF further demonstrate that RapR-Src-CFP binding to Cav1-YFP was significantly increased by Rap activation of RapR-Src (Fig. 6g, h). The same FRET result was obtained with MPH co-administration, further indicating MPH-triggered Cav1 Tyr¹⁴ phosphorylation was mediated by activated c-Src. Overall, we conclude that MPH-induced vesicular transport in human ECs requires Cav1 Tyr¹⁴ phosphorylation by c-Src downstream of Rac1/NOX-mediated ROS generation.

Fig. 6.

MPH-induced HRP vesicular transport is dependent on c-Src activation. a Inhibition of Src pathway using SKI-1 (100 nM SKI) prevented MPH-induced increase in transcytosis in HBMVECs, n = 10–18. b Quantification of c-Src protein levels in hCMEC/D3 cells after shRNA-mediated knockdown or infection with empty vector (pLKO). Top representative western blot images of c-Src (60 kDa), and GAPDH (37 kDa) are shown, n = 5–9. c Knockdown of c-Src (shc-Src) prevented MPH-induced vesicular transport compared with cells infected with pLKO (empty vector), n = 4–5. d, e Increase in Cav1 Tyr¹⁴ phosphorylation (p-Cav1) triggered by MPH (100 µM for 120 min) was prevented by d SKI (100 nM) in HBMVECs and by e c-Src knockdown in hCMEC/D3 cells. Top representative western blot images of p-Cav1 (22 kDa) and Cav1 (22 kDa) are shown, n = 10–15. f hCMEC/D3 cells were transfected with Cav1^Y14F and/or RapR-Src, and vesicular transport was assessed. Rap (200 nM) used to activate c-Src construct increased HRP transport, n = 9–12. g Quantification and h representative ratio images of Cav1-YFP and RapR-Src-CFP interaction assessed by FRET in hCMEC/D3 cells, n = 12–15. Scale bar 40 μm. Results are expressed as mean ± SEM. ***P < 0.001 vs. control (CTR or dashed line) or pLKO; ⁺⁺⁺ P < 0.001 vs. MPH, pLKO + MPH or Rap

Discussion

Methylphenidate is the drug of choice for the treatment of ADHD and its use has increased significantly over the last few years. Several studies have shown beneficial effects of MPH in ADHD [1], but nevertheless, its use among children is non-consensual [1, 3]. In addition, there is currently a concern regarding MPH misuse for cognitive enhancement. Thus, it crucial to better clarify the central effects of MPH. In the present study, we investigated whether MPH affects human brain ECs, the main component of the BBB, using a representative therapeutic dose range of MPH [21]. We demonstrate, for the first time, that MPH increases brain EC permeability by stimulating vesicular transport and unveil the intracellular signaling pathway responsible for this effect (Fig. 7).

Fig. 7.

Schematic illustrating the mechanism by which MPH promotes caveolae-mediated transcellular transport in human brain endothelial cells. Plus and minus signs represent stimulation and inhibition, respectively. Cav1 caveolin-1, MPH methylphenidate, M - β-C methyl-β-cyclodextrin, NOX nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, ROS reactive oxygen species, VitC vitamin C

It was previously demonstrated that MPH augments the number of vesicles in ECs [14] and that methamphetamine, a powerful psychostimulant drug of abuse, increases both paracellular and transcellular transports in ECs [18, 19]. Paralleling these findings, we showed that MPH increased transendothelial flux without altering monolayer electrical resistance or the expression/distribution of intercellular junctional proteins in human brain ECs. Instead, MPH promoted vesicular transport across ECs via a caveolae-dependent mechanism. We also revealed that this psychostimulant increased Cav1 phosphorylation. Induction of Cav1 expression has been identified in a number of brain disorders, such as Alzheimer’s disease [8] and intracerebral hemorrhage [7]. Moreover, caveolae-dependent processes are involved in the transport of macromolecules [37], virus [42], and fungal pathogens across ECs [32]. Caveolae have also been implicated in the internalization of TJ transmembrane proteins leading to brain endothelial barrier disruption during CNS inflammation [27], although recycling of TJs also allows for faster re-establishment of barrier function without new protein synthesis [43]. In addition, Cav1 reduction was associated with the deletion of TJ-associated proteins and a consequent increase of barrier permeability [44, 45]. In the present study, we identified neither alterations in the expression/organization of inter-endothelial junctions nor in the expression of Cav1 protein. Nevertheless, several studies have suggested that Cav1 phosphorylation plays a crucial role in the regulation of caveolae formation and transcytosis in pulmonary endothelial cells [6, 28, 37]. Very recently, it was shown that endocytosis and trafficking of caveolae are associated with a Cav1 Tyr¹⁴ phosphorylation-dependent conformational change in rat lung microvascular endothelial cells [11]. Accordingly, in the present study, we demonstrated for the first time that shRNA-induced depletion of Cav1 or the use of Cav1^Y14F mutant prevented increase in brain endothelial transcytosis induced by MPH. On the contrary, overexpression of Cav1^Y14D by itself enhanced vesicular transport, supporting the hypothesis that phosphorylation of Cav1 at Tyr¹⁴ promotes caveolae formation upon MPH stimulation. These data were further corroborated by immunoelectron microscopy, wherein numerous caveolae-like invaginations on the plasma membrane and in the cytosol of ECs were observed after exposure to MPH.

We also investigated the mechanism by which MPH increased caveolae vesicular transport. It has been shown that young rats treated with MPH exhibit oxidative damage as evidenced by an increase in both lipid peroxidation and protein carbonyl adducts in the brain [36]. In addition, MPH seems to interfere with important brain antioxidant defenses [46]. Whereas ROS generation by ECs at low levels signal important physiological activities, such as cell growth and differentiation [38], excessive generation of ROS overwhelms the intracellular antioxidant defense systems leading to an imbalance in redox homeostasis, oxidative stress, and endothelial dysfunction [38]. In fact, we found that intracellular ROS generation in human ECs, namely superoxide anion, increased upon exposure to MPH. Likewise, acute MPH administration in young rats increased the amount of brain superoxide [35]. We also identified NOX as the source of ROS production triggered by MPH. Our findings are consistent with the previous studies showing a crucial role for NOX-generated ROS in BBB disruption [38, 47]. Collectively, these observations led us to hypothesize that an antioxidant strategy could have a beneficial effect on ECs exposed to MPH. Indeed, the antioxidant VitC was able to prevent both MPH-induced ROS generation and transcytosis. Accordingly, others have reported that VitC could prevent microvascular endothelial barrier dysfunction during septic insult by blocking NOX-dependent ROS generation [48].

ROS signaling plays an important role in the control of endothelial permeability [38] by interfering with the dynamics of the actin cytoskeleton via Rho GTPases [39]. It is known that Rho-regulated cytoskeletal remodeling is essential for targeting vesicles to their correct location, enabling exocytosis [32, 33]. However, the mechanisms by which Rho GTPases exert their effects on intracellular trafficking are still largely not known. Herein, we observed that MPH had different effects on the activities of the endothelial Rho GTPases Rac1, RhoA, and Cdc42. In particular, while MPH induced the rapid activation of Rac1, it decreased the activity of RhoA and had no effect on Cdc42. Likewise, Rac-mediated ROS production in HeLa cells results in the downregulation of Rho activity, which is required for Rac-induced formation of membrane ruffles and integrin-mediated cell spreading [49]. Interestingly, others have demonstrated the involvement of Rac1 in the recruitment and assembly of the endothelial NOX complex [40]. Here, we showed that MPH activated Rac1 and that shRNA-mediated Rac1 knockdown prevented ROS generation and, consequently, the effect of MPH. Although the exact role of Rac1 in EC barrier function is not fully understood, Rac-dependent generation of ROS is known to cause barrier dysfunction [40]. Chen et al. [40] reported that Rac1 inhibition in human pulmonary artery ECs contributes to barrier protection via the inhibition of NOX and superoxide generation. Similarly, we demonstrated that Rac1 knockdown inhibits MPH-induced caveolae-mediated transcytosis. Furthermore, Rac1 activation was reported to be required for bacterial entry into human ECs [33]. Several pathogens use caveolae as a carrier vacuole to hijack endosomal trafficking in host cells and escape lysosomal degradation [50]. In light of these observations, it is plausible to speculate that MPH-induced caveolae formation and trafficking might ultimately promote the entrance of pathogens into the brain.

RhoA activation is known to be involved in functional changes of TJ proteins, such as claudin-5, and also reorganization of the actin cytoskeleton leading to disruption of endothelial cell–cell contacts and paracellular hyperpermeability [43, 51]. RhoA inhibition and dominant negative RhoA mutant prevented the loss of tight and adherens junctions, the decrease of transendothelial resistance, and stress fiber formation in human umbilical vein endothelial cells [52]. Moreover, RhoA inactivation caused the disassembly of actomyosin stress fibers and reorganized F-actin and phosphotyrosine-containing proteins to β-catenin-containing cell–cell junctions, a process that increased the size-selective permeability of endothelial monolayers [53]. In addition, the reduction of RhoA activity and elevation of Rac1 signaling are important steps in the control of endothelial permeability due to their involvement in the re-annealing of the intercellular junctions [54]. In our study, we did not observe alterations in the intercellular junctions, which may be justified by data indicating the inactivation of RhoA in response to MPH.

The activity of c-Src has been implicated in endothelial hyperpermeability [55], and its inhibition ameliorates vascular leakage and inflammation in rodent brains [56]. Here, we demonstrated that oxidant signaling is a key feature of c-Src activation in response to MPH in brain ECs, since the antioxidant VitC or NOX inhibition abrogated MPH-induced c-Src activation. Intracellular ROS may regulate the activity of c-Src via oxidation of two cysteine residues that control c-Src conformational changes necessary for its activation [41]. The c-Src and Rac1 hierarchy is generally context and stimuli specific in different cells. Our results show that MPH-mediated Rac1 activation is upstream of c-Src in human brain ECs. In response to oxidative stress, Tyr¹⁴ on Cav1 is the principal c-Src target [28, 37]. We showed that MPH leads to Cav1 phosphorylation by c-Src, and, consequently, to HRP transcytosis. Direct activation of c-Src via rapamycin increased vesicular transport and promoted the interaction between c-Src and Cav1 at the plasma membrane. In this respect, our results are consistent with other reports showing that Src-Cav1 interaction requires c-Src activity to promote caveolae-mediated pulmonary endothelial hyperpermeability and edema formation [37]. In addition, overexpression of Cav1^Y14F in brain ECs abrogated RapR-Src-mediated transcytosis, supporting the premise that c-Src-dependent phosphorylation of Cav1 on Tyr¹⁴ is necessary and sufficient to promote caveolae trafficking dynamics and endothelial hyperpermeability. Accordingly, Zimnicka and collaborators [11] have recently demonstrated a key role of Src-dependent Cav1 phosphorylation in promoting caveolae release from the plasma membrane via phosphorylation-dependent destabilization of Cav1 oligomers. Overall, our data indicate that MPH induces Rac1/NOX-dependent ROS generation and subsequent c-Src activation-dependent Cav1 Tyr¹⁴ phosphorylation promotes transcellular transport (transcytosis) in human brain ECs.

The present study shows for the first time that MPH has a direct effect on human brain ECs and provides new insights into the mechanism underlying MPH-induced BBB hyperpermeability. In this context, our data raise the question of whether MPH use enhances brain susceptibility to peripheral factors and the risk of neurological disease. However, while the cerebral vasculature provides a crucial protective role in maintaining brain homeostasis, the BBB also represents a substantial obstacle to the delivery of many neurotherapeutic drugs. Thus, our results also suggest that transient and controlled MPH exposure might constitute a potential strategy for enhancing drug delivery into the brain.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

ADHD

Attention deficit hyperactivity disorder

BBB

Blood–brain barrier

Cav1

Caveolin-1

ECs

Endothelial cells

HRP

Horseradish peroxidase

HBMVECs

Human brain microvascular endothelial cells

MPH

Methylphenidate

M-β-C

Methyl-β-cyclodextrin

NOX

Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase

ROS

Reactive oxygen species

TEER

Transendothelial electrical resistance

TJs

Tight junctions

VitC

Vitamin C

Compliance with ethical standards

Conflict of interest

The authors declare no conflict of interest.