Ginsenoside Rb₁ and digoxin (internal standard, IS) were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). Collagenase type II, bovine serum albumin (BSA), gelatin, phloretin, and D (+)-glucose were purchased from Sigma (Sigma-Aldrich Co. LLC., United States). Fetal bovine serum (FBS) was purchase from GIBCO (Invitrogen Corporation, United States). Dulbecco’s modified Eagle’s medium (DMEM-F12) was purchased from HyClone (Thermo Scientific, United States). Rabbit anti-GLUT1, rabbit anti-GLUT3, and goat anti-rabbit IgG (H+L) conjugated with the appropriate horseradish peroxidase were purchased from Abcam. All other chemicals and reagents were of analytical or LC–MS grade as appropriate.

Male Sprague-Dawley rats (220–250 g) were supplied by Center of Experimental Animals, Shanghai First People’s Hospital Affiliated Shanghai Jiao Tong University. Animals were housed under a constant environment condition (temperature, 25 ± 1°C; humidity, 50 ± 5%) with free access to standard laboratory diet and water. Sprague-Dawley neonate rats of 10 days old were used. All the experiments were performed in accordance with China’s Guidelines for Care and Use of Laboratory Animals.

Isolation and culture of rat brain microvessel endothelial cells were operated according to previous studies with some modifications (Liu et al., 2013). The isolated rBMEC was seeded in cultured flasks coated with 2% gelatin and cultured in culture medium. The culture medium consisted of DMEM/F12 supplemented with 20% FBS, 2 mM L-glutamine, heparin (50 mg/L), 100 IU/mL penicillin, 100 μg/mL streptomycin, and NaHCO₃ (0.5 g/L). On the first day after seeding, the culture medium was replaced, after which the culture medium was replaced every 2 days. At 5–6 days, the rBMEC was transferred to 6-well plates coated with gelatin and cultured in an incubator with a saturated humidity at 37°C in 5% CO₂ and 95% air. Transport studies were performed when cells reached 100% confluency. The endothelial phenotype of the rBMEC was identified by staining for the endothelial cell marker vWF using a confocal laser scanning microscope (Leica TCS SP8).

When rBMEC was grown to 100% confluency in gelatin-coated 6-well plates, uptake experiments were performed. Briefly, the cultured cell monolayer was washed three times with 1 mL of pH 7.4 Hanks’ balanced salt solution (HBSS). The cultured rBMEC was pre-incubated for 30 min in 1 mL HBSS at 37 or 4°C. After pre-incubation, the HBSS solution was removed by suction, and HBSS (1 mL) containing 60 μg/mL Rb₁ was added to initiate uptake. Then the cells were incubated at 37 or 4°C at a set time (5, 10, 15, 30, 60, 120, and 240 min). After the end of the incubation, the incubation solution was aspirated and the monolayers were carefully washed three times with 1 mL ice-cold HBSS to terminate the uptake. Then, 0.5 mL ultrapure water was added to each incubated well, and the material in the wells was frozen (at -80°C)/thawed (at room temperature) repeatedly four times to lyse cells. After the solubilized, the monolayer supernatants were collected and 10 μL was injected into LC–MS/MS to determine Rb₁ concentration. The protein contents of the cell monolayers were measured by the method of Bradford using BSA as the standard (Bradford, 1976). Uptake was expressed as the cell-to-medium ratio (μL/mg protein or μL/mg protein/min) and termed the volume of distribution (V_d), obtained by dividing the uptake amount by the concentration of substrate in the transport medium. V_d is derived from the ratio of ng/mg protein to ng/μL buffer.

In order to assess the kinetic parameters, Rb₁ uptake data (7.5 to 960 μg/mL for 120 min) were analyzed using Michaelis-Menten plots based on the following equation: V =Vmax⁡×SKm⁡×S+Pdiff×S, where V is the initial uptake rate of substrate (nmol/mg protein/min), V_max is the maximum uptake rate (nmol/mg protein/min), S is the concentration of Rb₁ in the medium (μM), K_m is the Michaelis-Menten constant (μM), and P_diff is the non-saturable uptake clearance (μL/mg protein/min), respectively (Kitamura et al., 2014). The uptake data were fitted to the above equation by nonlinear least-squares regression analysis with Graphpad Prism 5.0 software.

Our study showed that the uptake of Rb₁ by rBMEC was dependent on time, concentration, and temperature. A steady-state uptake occurred at 120 min. In the present study, the uptake at 120 min was selected to measure the effect of test agents on the transport characteristics of Rb₁. Rb₁ concentration was set to be 60 μg/mL. The rBMEC was seeded out in 6-well plates, allowed to adhere, and grown to confluency. To study the transport mechanisms, uptake of Rb₁ was carried out in the presence of a metabolic energy inhibitor and a series of established transporter inhibitors. The impact of these inhibitors on Rb₁ uptake by rBMEC was assessed at the steady state at 37°C. NaN₃ (500 μM), probenecid (100 μM), CsA (25, 50, and 100 μM), or phloretin (5 μM) was added to inhibit different transport systems. In the inhibition study, to further assess the effect of phloretin on Rb₁ uptake into rBMEC, the cells were incubated at 37°C at a set time (5, 15, 30, 60, and 120 min). Uptake was measured after incubation with Rb₁ in the presence of phloretin at the concentration of 5 μM.

In the competition experiments, Rb₁ uptake (30–960 μg/mL) was measured in the absence and presence of D-glucose (100 mM) which is a principal physiological substrate of GLUT1. Briefly, cells were grown in 6-well plates for 7 days before their medium was removed, pre-incubated with warm glucose-free Hanks’ solution for 30 min. Subsequently, samples and controls were treated as described transport studies, but this time in addition with 100 mM D-glucose to the various concentrations of Rb₁ in glucose-free Hanks’ solution at 37°C for 120 min.

The cytotoxic effects of the test agents used in this study were assessed on confluent monolayers of cells in 96-well plates using a CCK-8 assay. Cells were subjected to a 200 μL/well of each drug in Hanks’ solution at the concentrations used in the experiments, followed by incubation for another 4 h. Then, 10 μL Cell Counting Kit-8 (CCK8, Dojindo) solution was added to each well. Plates were incubated for additional 4 h. Absorbance readings at 450 nm were obtained using a spectrophotometer (Thermo Varioskan) and expressed as percentage viability compared to control untreated cells.

Rats were randomly divided into two groups in 36 of each group. One group served as the control and another group were treated with phloretin. Dose of phloretin was set to be 5 mg/kg. Rb₁ (10 mg/kg, i.v.) was administered to the experimental rats after i.v. administration of phloretin or saline for consecutive 1 week. Six rats from each group were selected and the blood samples were collected into heparinized Eppendorf tubes via the abdominal aorta at 0.5, 2, and 6 h after Rb₁ administration. Then, brain samples were immediately collected. The plasma samples were obtained by centrifuging at 1000 × g for 10 min. Plasma and brain samples were frozen at -80°C until analysis.

HPLC–MS/MS was composed by a Shimadzu LC-20A chromatographic system and an API 4000 mass spectrometer equipped with electrospray ionization (ESI) source system. MS/MS detection was performed on an API 4000 mass spectrometer using multiple reaction monitoring (MRM) mode by monitoring the fragmentation of m/z 1107.6 → 179.0 for Rb₁ and m/z 779.4 → 345.2 for digoxin (IS). Chromatographic separations were carried out on a Shim-pack XD-ODS column (2.0 mm × 30 mm, 2.2 μm) with a Shim-pack GVD-ODS (2.0 mm × 5 mm, 4.6 μm) guard column (Shimadzu, Japan) at a flow rate of 0.28 mL/min using 10 mM acetic acid in water (phase A) and methanol (phase B) as mobile phase. A linear step gradient elution was performed as followed: phase B was increased from 45 to 90% within the first 3 min, and decreased to 45% within the next 3 min (total gradient time: 6 min). A 10 μL sample was injected into the system with the auto-sample conditioned at 4°C and column temperature maintained at 40°C.

The biological samples (100 μL) were placed in a 1.5 mL Eppendorf tube, and mixed with 10 μL IS solution (500 ng/mL) for 3 min by vortexing. The mixture was extracted with methanol (0.9 mL) by vortexing, and then centrifuged at 14,000 rpm for 5 min. The supernatant (0.8 mL) was transferred to a new 1.5 mL Eppendorf tube and evaporated to dryness under vacuum. The dried residue was reconstituted with 100 μL methanol, vortex-mixed for 30 s, and centrifuged at 14,000 rpm for 5 min. Finally, 10 μL of the supernatant liquid was immediately subjected to HPLC–MS/MS analysis. The system control and data analysis were performed by AB Sciex Analyst software (the software version: Analyst 1.5.1). Retention time for Rb₁ was t_R = 3.80 ± 0.01 min and for IS digoxin t_R = 3.13 ± 0.01 min. The lowest limits of quantitation of Rb₁ in cell lysate, plasma, and brain homogenate were 1, 2.5, and 1 ng/mL, respectively. The recoveries were higher than 90%; the relative standard deviations (SD) of intra-day and inter-day were lower than 10%.

Total proteins were extracted and 60 μg of each protein was separated on SDS-10% polyacrylamide gels. Proteins on the gel were transferred to PVDF membrane (Merck-Millipore). Blots were blocked with 5% skimmed milk dissolved in PBS containing 0.1% Tween-20 (PBST) for 2 h at room temperature. The membrane was incubated with primary antibody at 4°C overnight. Washing of the blots with PBST three times was followed by incubation with horseradish peroxidase-conjugated secondary antibody for 2 h at room temperature. After washing, the chemiluminescence signal was monitored using the ChemiDoc XRS+ system (Bio-Rad, Hercules, CA, United States). Densitometric semi-quantitative analysis of protein bands was carried out using Image J.

All data were expressed as mean ± SD. Statistical significance of difference among groups was evaluated by one-way of analysis of variance (ANOVA) and the Student’s t-test. A P-value of <0.05 indicated a statistical significance.

Uptake of Rb₁ by rBMEC was evaluated at various intervals (5, 10, 15, 30, 60, 120, and 240 min) to obtain the optimal time for uptake studies. Figure 1 shows the time course of the uptake of Rb₁ by rBMEC at a concentration of 60 μg/mL at 37 and at 4°C. Rb₁ uptake increased linearly with time until 120 min, and a plateau of accumulation was observed at 120–240 min. Accordingly, uptake at 120 min was used for subsequent kinetic and inhibition studies. However, in contrast, the uptake rate at 4°C was slow and the intracellular amounts of Rb₁ were about two to threefold greater at 37°C than those at 4°C. This indicated transporters in rBMEC membranes were involved in the Rb₁ transport.

To study the mechanism of Rb₁ transport, uptake of Rb₁ by rBMEC was examined at various concentrations (7.5–960 μg/mL) at steady state and the kinetics parameters (K_m and V_max) were investigated by fitting the data to the Michaelis-Menten equation. As shown in Figure 2A, initial uptake of Rb₁ showed concentration dependency. Kinetics analysis provided a K_m value of 119.2 ± 15.6 μM and a V_max of 0.0034 ± 0.0003 nmol/(mg protein ⋅ min). The Eadie–Hofsee plot for saturation uptake, fitted by subtraction of nonsaturable uptake from total uptake, approximately appeared a straight line (R² = 0.9066), indicating involvement of a simple passive membrane diffusion (Figure 2B).

To clarify whether the accumulation of Rb₁ in rBMEC was not only related to a simple passive diffusion, but to facilitated diffusion, we performed the effects of various inhibitors of transporters that facilitate the uptake of Rb₁ into rBMEC. As shown in Figure 3, while no statistically significant differences were observed with inhibitors of the ABCB1 (P-gp, CsA) (Sun et al., 2006), organic anion-transporting polypeptides (OATPs, probenecid) (Jørgensen et al., 2007), and metabolic energy (NaN₃) (Kitamura et al., 2014), a significant decrease (^∗∗P < 0.01 versus the control) of Rb₁ uptake into rBMEC was seen in the presence of the inhibitor phloretin that inhibits glucose transporters (GLUT1) (Carruthers et al., 2009). This result indicated that Rb₁ uptake crossing rBMEC membranes was independent of ATP and was actively transported by GLUT1-mediated facilitated diffusion.

In order to investigate whether the changes in Rb₁ uptake by rBMEC in presence of the tested agents resulted from cell damage caused by the agents used, cells were incubated in Hanks’ solution containing of 500 μM NaN₃, 100 μM probenecid, 25, 50, and 100 μM CsA, or 5 μM phloretin at 37°C for 4 h. Proliferation of rBMEC exposure to the test agents used was measured by CCK-8 assay. Compared with the control, no statistically significant differences existed (P > 0.05 versus the control, Figure 4). Thus, we concluded that the agents used did not have cytotoxicity at the concentration used, and that changes in the uptake of Rb₁ did not result from cell damage. Phloretin, which significantly inhibited the uptake of Rb₁ by rBMEC, was selected for further study.

In order to further investigate the effect of phloretin on Rb₁ uptake into rBMEC, the time course of Rb₁ uptake in the absence or presence of 5 μM phloretin was evaluated at 37°C. As shown in Figure 5A, the cell-to-medium (C/M) ratio of Rb₁ was increased with time, but it was found that phloretin significantly decreased the C/M ratio of Rb₁, especially from min 30 to 120. GLUT1 and GLUT3, encoded by the SLC2A1 and SLC2A3 gene, respectively, are 2 of 14 facilitative GLUT transporter members and are mainly expressed in CNS. To see whether the decreased C/M ratio of Rb₁ was correlated with the expression level of GLUT1 or GLUT3 in rBMEC, western blot assay was performed (Figure 5B). Statistical analysis showed that the expression levels of GLUT1 at 30, 60, and 120 min were significantly lower than that of control (Figure 5C, ^∗P < 0.05 and ^∗∗P < 0.01 versus the control). However, compared with the control, there was no significant difference in the expression level of GLUT3. These data indicated that the decreased C/M ratio of Rb₁ was correlative well to the down-regulated function and level of GLUT1 in rBMEC.

While GLUT1 is able to transport galactose, mannose, and glucosamine, its principal physiological substrate is glucose. Under normal physiological conditions, the brain is absolutely dependent on glucose as a fuel source. GLUT1, as the major GLUT isoform expressed in brain endothelial cells, plays an important role in glucose transport across the BBB (Yeh et al., 2008). In the initial experiments, we analyzed the uptake characteristics of Rb₁ by rBMEC. When we added the inhibitor phloretin of GLUT1, we observed significantly reduced C/M ratio of Rb₁ uptake into rBMEC. Likewise, in Lineweaver–Burk plot of the inhibitory effects on the uptake of Rb₁, the plots of Rb₁ uptake in the absence and presence of D-glucose intersected at the ordinate axis (Figure 6), indicating that D-glucose competitively inhibited Rb₁ uptake with a K_i value of 885 μM.

Except for the time point of 0.5 h after a single i.v. administration of Rb₁ to the normal rats given phloretin for consecutive 1 week, plasma concentrations of Rb₁ in the phloretin-treated normal rats at 2 h and 6 h were no statistically significant differences than those in the untreated normal rats. Whereas cerebral concentrations of Rb₁ in the phloretin-treated normal rats were significantly lower than those in the untreated normal rats (Table 1). As a result, the brain-to-plasma concentration ratio of Rb₁ in the phloretin-treated normal rats appeared to be significantly lower than that in the untreated normal rats, indicating that phloretin inhibits the transport of Rb₁ across the BBB. To see whether the decreased brain–plasma concentration ratio of Rb₁ was also related to the expression level change of GLUT1 or GLUT3 in the phloretin-treated normal rats, western blot assay was performed. The western blot results further showed that the expression levels of GLUT1 in the brain of normal rats given phloretin for 1 week were significantly decreased than those in the untreated normal rats (P < 0.05, Figure 7). Although phloretin also binds to the GLUT3, it potently inhibits the GLUT1-type glucose transporter (Martin et al., 2003). These results demonstrated that the decreased brain–plasma concentration ratio of Rb₁ was correlated well with the decreased function and level of GLUT1 in the phloretin-treated normal rats, conforming the involvement of GLUT1 in Rb₁ transport across the BBB into brain.