DL-methamphetamine (purity > 95%, supplied by Public Security Bureau of Beijing Municipality) was dissolved in a 0.9% saline solution. The ER stress inhibitors salubrinal (SB, SML-0951) and sodium 4-phenylbutyric acid (PBA, ab141253) were from Sigma (United States) and Abcam (Cambridge, MA, United States), respectively. The NADPH oxidase inhibitor apocynin (ab120615) and ROS scavenger N-tert-butyl-α-phenylnitrone (NBP, RQ75L-QN) were obtained from Abcam and Tokyo Chemical Industry (Tokyo, Japan), respectively. Antibodies were obtained from the following sources: CD31 (ab28364), claudin5 (ab15106), Bip (ab21685), p-IRE1 (ab48187), IRE1 (ab37073), CHOP (ab11419), and cytochrome c (ab13575) were from Abcam; occludin (13409-1-AP), bcl-2 (12789-1-AP), and bax (50599-2-1g) were from Proteintech (Chicago, IL, United States); ATF6 (BS6476), GAPDH (AP0066), and COX IV (BS2186) were from Bioworld (St. Louis, MO, United States); p-PERK (sc-32577) and PERK (sc-9477) were from Santa Cruz (Dallas, TX, United States).

The bEnd.3 cells (Wan et al., 2014; Hun Lee et al., 2015) were obtained from ATCC and cultured in ECM (ScienCell, San Diego, CA, United States), which contained 10% fetal bovine serum (Gibco, Carlsbad, CA, United States) and 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA, United States) under a humidified cell culture incubator at 37°C and with 5% (v/v) CO₂. The bEnd.3 cells were then exposed to 250 μM, 500 μM, and 1 mM METH for 24, 48, and 72 h, and prepared for the following experiments (Fernandes et al., 2015).

Three-month-old male C57BL/6J mice, weighing 18–25 g, were obtained from Beijing Vital River Laboratory Animal Technology Co. Ltd., (Beijing, China). All experiments were performed according to the Guideline for the Care and Use of Laboratory Animals. The experiments were approved by the Local Committee on Animal Care, Use and Protection of Hebei Medical University. Mice were randomly grouped into four experimental groups: Saline, METH, METH+PBA, and PBA. METH (5 mg/kg body weight, 3-h intervals) was injected intraperitoneally four times a day for 1 day. PBA (50 mg/kg body weight) was intraperitoneally injected 30 min before METH injection. One day after the first injection, evaluation of BBB permeability was performed and the animals were killed for tissue collection.

The bEnd.3 cells were seeded on coverslips and fixed in 4% paraformaldehyde for 20 min at 25°C. After blocking solution was treated, the cells were incubated with primary antibodies (anti-CD31 1:50, anti-cytochrome c 1:200) overnight at 4°C. Cells were clearly washed with PBS and incubated with the second antibody for 40 min at 37°C. After washing in PBS for 30 min; nuclear staining was performed using 4′,6-diamidino-2-phenylindole (DAPI). Cells were detected by an SP8 Leica confocal microscope (Leica Biosystems, Wetzlar, Germany).

The measurement of cell viability was carried out by MTT assay. Briefly, bEnd.3 cells were seeded at a density of 1 × 10⁴ cells per well. Then the cells were treated with METH. A volume of 10 μl MTT (5 mg/ml) was added to each well. After an incubation for 4 h, 150 μl of 10% SDS were added and treated for 7–17 h at 37°C. Absorbance was measured at OD = 570 nm by a full-length microplate reader (Thermo Fisher, Waltham, MA, United States).

Flow cytometry was carried out to measure cell apoptosis, which was described previously (Yin et al., 2015). The cells were harvested and washed twice with PBS. Thereafter, cells were labeled with annexin V-FITC and propidium iodide (PI) (BD Biosciences). Fluorescent signals were measured with a flow cytometer (FACSCalibur, BD Biosciences).

To determine relative levels of cell apoptosis, the S7165 ApopTag^®Red In Situ Apoptosis Detection Kit (Millipore, Billerica, MA, United States) was used. Briefly, the cells were seeded on slides. After permeabilizing and fixing for a second time, the cells were equilibrated and labeled. Then, DAPI was used to visualize cell nuclei. Cells were detected by an SP2 Leica confocal microscope. The numbers of TUNEL-positive cells were counted in five randomly selected fields (400× magnification) and the percentages calculated against total number of DAPI-stained cells. Two independent observers who were blinded to the experimental conditions performed the counts and calculated the average number of TUNEL-positive cells. Data were collected from more than three independent experiments performed in triplicate.

Transepithelial electrical resistance was used to determine the integrity of brain endothelial monolayers using the Millicell Voltohmmeter (Millipore) with STX01 chopstick electrodes. In brief, bEnd.3 cells were inoculated in transwell inserts of 24-well plates at a density of 5 × 10⁴/well. Prior to use, the machine was calibrated, then the longer electrode was placed so as to touch the bottom of the dish while the shorter electrode was prevented from reaching the bottom of the insert. The readings were corrected by transwell inserts with no cells (subtracted from each experimental measurement), then divided by filter size.

Protein extracts of bEnd.3 cells and brain tissue were loaded on 8–15% polyacrylamide minigels (Bio-Rad Laboratories, Hercules, CA, United States), and proteins were separated using SDS-PAGE. Further, protein was electroblotted to PVDF membranes and blocked using 5% non-fat dry milk at 37°C for 1 h. The membranes were incubated with 5% non-fat dry milk containing the appropriate primary antibody overnight at 4°C. After primary incubation, blots were incubated in the secondary antibody for 1 h at room temperature. An imager (LI-COR, Lincoln, NE, United States) was employed to detect the emitted light of the blot.

Total RNA was derived from bEnd.3 cells with an RNeasy Mini Kit (SABiosciences, Qiagen), and 1 μg of RNA was taken out for reverse-transcription using the RT² First Strand Kit (SABiosciences, Qiagen). Expression of 84 UPR pathway-related genes in METH-treated cells was compared to controls using RT² Profiler^TM PCR Array Mouse UPR (PAMM-089A). Heat maps and scattered plots were created with the SA Biosciences RT² Profiler PCR Array Data Analysis Web Portal. If the fold change was greater than 2 and p < 0.05, the genes were considered significantly different across conditions and described in Figure 2.

Short interfering RNA (siRNA) targeting CHOP (si-CHOP) (GGA AGC AAC GCA TGA AGGA) was obtained from Ribo Life Science Co. (Suzhou, Jiangsu province, China). Transfection of siRNA into bEnd.3 cells was conducted. Briefly, bEnd.3 cells were incubated at 60% confluence in 24-well plates. We used the protocol that included the Lipofectamine^®2000 transfection reagent (Invitrogen), in which 20 pmol si-CHOP or scrambled siRNA was added to 50 μL of serum-free medium. In another tube, 1 μL of Lipofectamine^®2000 was added to 50 μL of ECM medium and incubated for 5 min at 25°C. Both tubes were mixed and incubated for 20 min at 25°C. Thereafter, the cells were incubated with 100 μL of this mixture.

To determine the level of ROS, the oxidative conversion of 2,7-dichlorofluorescein diacetate (DCFH-DA) to the fluorescent compound dichlorofluorescein (DCF) was measured. Cells were cultured in 24-well plates at 5 × 10⁴ cells per well or in 96-well plates at 1 × 10⁴ cells per well. The medium was treated with aspiration and replaced by serum-free medium. DCFH-DA (Beyotime Company, Shanghai, China) stock solution (10 mM) was diluted to a final concentration of 10 μM using serum-free medium (V/V = 1:999), and then added to the plates for 30 min. DCF fluorescence was measured at 488-nm excitation and 520-nm emission using an SP8 Leica confocal microscope or a microplate reader (Infinite 200, Tecan, Mannedorf, Switzerland).

The change of MMP in the bEnd.3 cells was monitored with the MMP Detection Kit with JC-1 (Beyotime Company). Briefly, bEnd.3 cells were seeded in 6-well plates with 2.5 × 10⁵ cells per well. After drug treatments, the cells were harvested and treated with a JC-1 working solution (5 μg/mL) for 20 min at 37°C in a 5% CO₂ incubator. After washing in cold JC-1 staining buffer, the fluorescence intensity of the cells was monitored at excitation/emission wavelengths of 514/529 nm (JC-1 monomers) and 585/590 nm (JC-1 aggregates) using a microplate reader (Infinite 200, Tecan, Mannedorf, Switzerland). The MMP of bEnd.3 cells was calculated as the fluorescence ratio (JC-1 aggregates: monomers).

Collecting the cells with different treatments and mitochondria of cells was isolated with a Cell Mitochondria Isolation Kit (Beyotime Company). Suspending the cells with 1 mM of phenylmethanesulfonyl fluoride in mitochondrial isolation buffer, after that the cells were homogenized and then centrifuged at 4°C for 10 min, and the resulting supernatant was subjected to centrifugations with the speed of 11,000 × g for another 10 min to separate the cytosolic (supernatant) from the mitochondrial fractions (pellets).

Evans Blue and sodium-fluorescein (NaFluo) were used to evaluate large and small solute permeability of the BBB, respectively. The procedures were performed in accordance with previous methods (Ramirez et al., 2009; Martins et al., 2011). Briefly, mice were injected intraperitoneally with 2% NaFluo (2.5 ml/kg, Sigma) or into the tail vein with 2% Evans Blue (4 ml/kg, Sigma). The tracers were circulated for 30 min and 2 h, respectively. The mice were then transcardially perfused with saline until colorless perfusate was observed. The animals were then sacrificed under euthanasia and the brains were immediately collected. Tissue was homogenized in 3 ml of 50% trichloroacetic acid. The homogenate was centrifuged at 25°C for 10 min, 5 M NaOH was used to neutralize the supernatant. Measurements of NaFluo and Evans Blue fluorescence were measured at the wavelengths of excitation/emission at 440/525 nm and 550/620 nm, respectively, by a microplate reader (Infinite 200, Tecan, Mannedorf, Switzerland).

The data are presented as the means ± SD. One-way ANOVA and Student’s t-test were performed for statistical analysis as appropriate. A p-value < 0.05 indicated statistical significance. All analysis were conducted using SPSS 16.0 (Chicago, IL, United States).

First, in order to select the appropriate concentration of METH for the subsequent experiments, we used 250 μM, 500 μM, and 1 mM of METH to treat bEnd.3 cells; then, MTT assay was performed to observe cell viability. The results showed that 250 and 500 μM of METH had no significant influence on bEnd.3 cell viability, but 1 mM of METH significantly inhibited bEnd.3 cell viability (Figure 1A). Thus, we chose 1 mM of METH for subsequent study. Next, we detected cell viability using the MTT assay after bEnd.3 cells were exposed to METH (1 mM) for 24, 48, and 72 h. Exposure of cells to METH showed a significantly decrease of cell viability in a time-dependent manner (Figure 1B). Flow cytometry and TUNEL assays were used to detect apoptosis of bEnd.3 cells after METH treatment. As shown in Figure 1C, the apoptotic ratio significantly and gradually increased after METH treatment for 24, 48, and 72 h. Moreover, the TUNEL assay showed that compared with the control group (6.97 ± 1.31%), 1 mM METH caused a significant increase in apoptosis rate at 24 h (22.31 ± 2.50%, p < 0.001) (Figure 1D). These results suggest that METH could reduce cell viability and induce the apoptosis of bEnd.3 cells.

Transepithelial electrical resistance measurement supplies an efficient and easy evaluation for assessing of BBB integrity using an in vitro model. In our study, we found that 1 mM of METH reduced TEER after 24 h, and that the reduction was gradually enhanced after 48 and 72 h of METH treatment (Figure 1E). Furthermore, our results revealed that METH treatment gradually reduced the expression of tight junction (TJ) proteins claudin5 and occludin (Figure 1F). The results suggest that the integrity of BBB endothelial bEnd.3 cells could be destroyed by METH, partly via down-regulation of tight junction proteins. In summary, METH treatment reduced cell viability through induction of apoptosis and appeared to induce damage to the bEnd.3 cells that would diminish the tightness (and thus the strength) of the BBB.

It is widely accepted that long time ER stress promotes cell apoptosis. To evaluate whether bEnd.3 cells exposure of METH led to ER stress induction, we identified the expression of numerous critical genes (84 genes) of the UPR pathway using PCR arrays. Our results showed that ER stress affects the expression of many critical genes, such as ATF4, ATF6, Ddit3, Eif2a, Hspa5, and Xbp1, was increased > two-fold after METH treatment for 24 h (Figures 2A–C). Further, we selected several typical proteins, including Bip/GRP78, p-IRE1, p-PERK, and ATF6, in the UPR pathway to verify the results of RT-PCR arrays and further determine whether the ER stress was sustained or transient. Briefly, bEnd.3 cells were exposed to METH (1 mM) for 0.5, 1, 3, 6, 12, and 24 h, after which the expression of ER stress related-proteins was assessed. As is presented in Figure 2D, exposure of bEnd.3 cells to METH significantly promoted the expression of ER stress sensor proteins in a time-dependent manner, such as p-IRE1, p-PERK, and ATF6. Therefore, these results suggest METH-induced prolonged ER stress in bEnd.3 cells.

We determined whether the downstream pro-apoptotic pathway effector CHOP was activated following METH treatment. It has reported that CHOP/GADD153 belongs to the bZip transcriptional factors family and is activated by p-IRE1, ATF6, and p-PERK pathways. We found that METH exposure caused a time-dependent increase in CHOP (Figure 3A), and that this increase was blocked by pre-treatment of the ER stress inhibitors SB (50 μM) or PBA (5 mM) 30 min before METH treatment (Figure 3B). These results suggest that METH induced downstream pro-apoptotic pathways through sustaining ER stress.

Subsequently, we further examined the role of CHOP in METH-induced endothelial cell dysfunction using si-CHOP to knockdown the expression of CHOP. The efficiency of si-CHOP was 55% in naïve cells. The transfection of si-CHOP was found to attenuate the level of CHOP increased by METH treatment (Figure 4A). Furthermore, si-CHOP could partially restore the reduction of cell viability after METH treatment (Figure 4B). The results of these experiments show that the knockdown of CHOP significantly inhibited METH-induced apoptosis (Figures 4C,D) and alleviated the disruption of tightness of bEnd.3 monolayers (Figures 4E,F). In summary, as the downstream effector of pro-apoptotic pathways, CHOP was involved in ER stress and played a key role in METH-induced bEnd.3 cell dysfunction.

Endoplasmic reticulum stress is triggered under oxidative conditions. ROS generation accelerates the production of misfolded proteins, and ultimately results in the ER stress (Tsai et al., 2012; Tsukiyama-Kohara, 2012). We thus aimed to determine whether METH subject to bEnd.3 cells could cause the production of ROS. DCF fluorescence assay was performed to assess the levels of intracellular ROS after bEnd.3 cells were treated with METH (1 mM). We found that METH treatment dramatically promoted the levels of ROS, and it was peaked at 3 h. The ROS levels decreased quickly, and then returned to normalcy after 24 h of treatment, indicating that the increase of ROS levels was rapid and transient (Figure 5A). To further investigate the involvement of mitochondrial dysfunction in METH-induced apoptosis of bEnd.3 cells, the Bcl2/Bax, MMP depolarization, and cytochrome c release were detected. As expected, cells supplemented with METH significantly decreased the ratio of Bcl2/Bax and the expression of MMP in a time-dependent manner (Figures 5B,C). Western blot analysis and immunofluorescence confirmed that treatment with METH prominently affected the release of cytochrome c, which led to a increase in cytosolic fraction, but a decrease in mitochondrial fraction (Figures 5D,E). Together, the data suggest that METH could induce the generation of ROS as well as mitochondrial dysfunction.

To explore the interaction between ROS generation and ER stress response, ROS inhibitors apocynin (a NADPH oxidase inhibitor, 200 μM) or NBP (a ROS scavenger, 100 μM) were subjected to bEnd.3 cells for 30 min before METH treatment. As is presented in Figure 6A, both inhibitors prominently ameliorated the METH-induced expression of p-PERK, ATF6, Bip and p-IRE1α. The results indicated the important role of ROS in the METH-mediated ER stress in bEnd.3 cells, and also indicate that the dysregulated expression of p-PERK, ATF6, Bip and p-IRE1α protein was mediated by METH-induced ER stress. Furthermore, CHOP silence inhibited cytochrome c release following METH treatment, thereby implicating CHOP in the METH-induced ER stress response induces the release of cytochrome c, which can activate the caspase family and accelerate cell apoptosis (Figure 6B).

To further identify whether PBA could reverse METH-induced BBB disruption, we used c57BL/6J mice and intraperitoneally injected METH (5 mg/kg body weight) four times/day for 1 day to generate an acute model of METH abuse (Gou et al., 2015). One group of mice was injected intraperitoneally with PBA (50 mg/kg) 30 min before METH injection, which is often used as an ER stress inhibitor in vivo (Jangra et al., 2016). Western blots of TJ proteins to analyze the BBB structure were used, as well as NaFluo (small molecular weight) and Evans Blue (large molecular weight) leakage to test BBB integrity. As illustrated in Figure 7A, the administration of METH decreased the expression of TJ proteins, including occludin and claudin5. The decrease of TJ proteins was partly reversed by PBA. Similarly, the permeability to both NaFluo and Evans Blue was also increased following METH exposure, a phenomenon which was partially inhibited by PBA (Figures 7B,C). Together, our results suggest that PBA could be a potential drug to reverse METH-induced disruption of BBB in vivo.