Cellular uptake of allicin in the hCMEC/D3 human brain endothelial cells: exploring blood-brain barrier penetration in an in vitro model

Reagents

Allicin (purity >98%) and DCFDA cellular ROS detection assay kit ab113851 were purchased from Abcam (USA). MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide was obtained from Invitrogen (Eugene, OR, USA). Lucifer yellow CH dipotassium salt was acquired from Sigma (St. Louis, MO, USA).

Cell culture

The hCMEC/D3, which is the brain microvascular endothelial cell line of the human blood–brain barrier, was purchased from Merck Millipore, MO, USA. The cells were grown in the endothelial basal medium-2 (EBM-2) containing EGM-2 MV SingleQuots® supplements and growth factors on a collagen-coated T-75 flask. Cell culture medium, supplements, and growth factors were bought from Lonza (Walkersville, MD, USA). The collagen Type I, Rat tail (extracted from rat tail tendons) was obtained from EMD Millipore (Billerica, MA, USA). The cultured cells were constantly incubated at 37 °C in a humidified incubator with 5% CO₂.

Cell viability assay

Cell viability of hCMEC/D3 cells was examined using an MTT assay. The yellow dye MTT is reduced to purple formazan form by the action of mitochondrial succinate dehydrogenase enzymes in living cells. The quantity of formazan is proportional to the number of viable cells. The hCMEC/D3 cells were seeded onto collagen-coated 96-well plates at a density of 1 × 10⁴ cells/well for 24 h. Subsequently, the culture medium was aspirated, and the cells were treated with 100 µl of allicin at concentrations ranging from 0 to 10 µg/ml for duration of 3 and 24 h. After allicin removal, 10 µl of MTT reagent (5 mg/ml) and 100 µl of phosphate-buffered saline (PBS) were added to each well and incubated for 4 h. After MTT removal, 50 µl of dimethyl sulfoxide (DMSO) was introduced to solubilize the formazan crystals. The optical density (OD) was measured to determine the cell viability at a wavelength of 540 nm using the Multiskan Go microplate spectrophotometer (Thermo Scientific, Finland). All experiments were conducted in triplicate.

ROS detection assay

Intracellular reactive oxygen species (ROS) formation was determined using the DCFDA cellular ROS detection assay kit following the manufacturer’s protocols. Briefly, after the hCMEC/D3 cells at a density of 1.5 × 10⁴ cells/well were allowed to adhere in a dark, clear bottom 96-well microplate with collagen-coated for 24 h. Following a single wash with PBS, the cells were stained by adding 100 µl of 15 µM fluorogenic dye DCFDA (2′,7′-dichlorofluorescin diacetate) into each well and incubated in darkness for 45 min at 37 °C. Post-DCFDA removal, the cells were subsequently washed with PBS, followed by incubation with 100 µl/well of various concentrations of allicin (ranging from 0 to10 µg/ml) and 50 µM tert-Butyl hydrogen peroxide (TBHP: positive control) for 3 h. The fluorescence intensity was measured to quantify the generated cellular ROS using a fluorescence microplate reader at excitation and emission wavelengths of 485 and 535 nm, respectively, employing a multi-mode plate reader (CLARIOstar Plus, Ortenberg, Germany). The experiments were performed in triplicate.

The construction of the in vitro BBB model

An in vitro model of BBB was generated utilizing a Transwell system that consists of the upper (apical or AP) and lower (basolateral or BL) chambers, designed to simulate the blood and brain side of the BBB, respectively. The two chambers of the Transwell insert are separated by a porous membrane (0.4 µm). Initially, the 24-well inserts (Millicell, Germany) were coated with collagen 1:20 in Dulbecco’s phosphate-buffered saline (DPBS) at 200 µl/well and incubated for 1 h at 37 °C and 5% CO₂. After collagen removal, the hCMEC/D3 cells were seeded onto the collagen-coated membrane of the insert at a density of 1 × 10⁴ cells/well. The cell-free insert with collagen coating served as a control insert. Subsequently, the apical and basolateral chambers of the BBB model were filled with the EBM-2 medium at 200 µl and 1,000 µl, respectively. The cells were cultured for approximately 21 days at 37 °C and 5% CO₂. The culture medium in both AP and BL chambers was carefully changed every 3 days from 1X to 0.5X and 0.25X during the culture period, where the value of X refers to the concentration of the growth factors and supplements that were added into the culture medium. On day 21, the integrity of the in vitro BBB model was verified by measuring trans-endothelial electrical resistance (TEER) across the cell monolayers using a Millicell ERS voltohmmeter (Millipore). The integrity of the cell monolayers was also further confirmed by monitoring the passage of Lucifer yellow (LY), a paracellular marker, across the in vitro BBB model. For the LY permeability assay, all media was removed from the AP and BL chambers of the inserts. Then, 200 µl of LY (20 µM) and 1,000 µl of Hanks’ Balanced Salt Solution (HBSS) were added into AP and BL chambers, respectively. After incubation for 1 h at 37 °C and 5% CO₂, the samples were collected. The concentration of LY from the BL chamber of the inserts was measured using a fluorescence microplate reader with excitation at 485 nm and emission at 530 nm. The %LY rejection of the in vitro BBB model was calculated using the following formula (Nkabinde et al., 2012). ${\displaystyle \text{\%}LY\text{rejection}=\left[1-\left(L{Y}_{\mathrm{BL}}/L{Y}_{\mathrm{I}}\right)\right]\times 100}$ Where LY_BL is the concentration of Lucifer yellow passing into the BL chamber and LY_I is the initial concentration of Lucifer yellow.

Determination of allicin in the in vitro BBB model by HPLC analysis

According to the cell viability assay and ROS detection results, the non-toxicity concentrations of allicin (0.5, 1, 2, and 5 µg/ml) were selected. After verifying the in vitro BBB model by TEER measurement and Lucifer yellow assay, the allicin test was carried out by adding 200 µl of the selected concentrations of allicin and 1,000 µl of HBSS into the AP and BL chambers of the inserts, respectively. Simultaneously, a cell-free control insert with collagen-coated was also treated with 200 µl of allicin 5 µg/ml. After incubation for 3 h at 37 °C and 5% CO2, the samples were collected. The LY assay and TEER measurement were performed again, respectively to confirm the intactness of the in vitro BBB model at the end of the allicin test. The concentrations of allicin that can cross the brain endothelial layer from AP to BL chamber were analyzed by high-performance liquid chromatography (HPLC) analysis using Aligent HPLC 1260 (Aligent Technologies, Santa Clara, CA, USA). The HPLC system consisted of the following components: a pump, an injector, a Hypersil ODS column (250 × 4.0 mm, 5 µm particle size), and a UV detector (254 nm). Isocratic mode operation at a flow rate of 0.5 ml/min was employed with mobile phases composed deionized H₂O and methanol in a ratio of 50:50.

Cellular uptake experiments

In explore the allicin uptake capability of hCMEC/D3 cells, cellular uptake experiments were performed. The experiments were first consisted of two conditions: (1) with cells and (2) without cells. In the first condition (with cells), the hCMEC/D3 cells were seeded in the collagen-coated wells of the 24-well plate at a density of 1 × 10⁴ cells/well. Simultaneously, one well was also seeded with hCMEC/D3 cells at a density of 5  × 10⁴ cells/well to investigate whether an increases in the cell number influences allicin uptake. The second condition (without cells) was filled with a cell-free medium. All wells were filled with 500 µl of culture medium and incubated for 24 h at 37 °C and 5% CO₂. Afterwards, the culture medium was removed from the plate. Then, 250 µl of the non-toxicity concentrations of allicin, which are 0.5, 1, 2, and 5 µg/ml, were added into the wells with cells (1 × 10⁴ cells/well) and those without cells. The well with a cell density of 5 × 10⁴ cells/well was treated with only 250 µl of 5 µg/ml allicin. After 3 hour-incubation at 37 °C and 5% CO₂, the supernatants were collected and subjected to HPLC to determine the allicin concentration.

To further assess whether the amount of allicin taken up by hCMEC/D3 cells exhibited proportionality to the cell number, the cells were seeded in the collagen-coated wells of the 24-well plate at a density of 0, 1 × 10⁴, 2.5 × 10⁴, 5 × 10⁴, and 7.5 × 10⁴ cells/well. After allowing the cells were allowed to attach for 24 h at 37 °C and 5% CO₂, the culture medium was replaced with 250 µl of 5 µg/ml allicin, followed by an incubation period at 37 °C and 5% CO₂ for 3 h. Then, the allicin concentration in the collected supernatants was determined by HPLC analysis.

To investigate the temporal dynamics of allicin uptake by hCMEC/D3 cells, the cells were seeded in the collagen-coated 24-well plate at a density of 5 × 10⁴ cells/well and incubated for 24 h at 37 °C and 5% CO₂. Subsequently, both the well with cells and the cell-free well (without cells) were treated with 250 µl of 5 µg/ml allicin at 37 °C and 5% CO₂. Supernatants were collected at 0.5, 1, 1.5, 2, 2.5, and 3 h of incubation period. The allicin concentration in each collected supernatant was quantified by HPLC analysis.

In silico ADME profiling predictions

Physiochemical characterization of absorption, distribution, metabolism, and excretion (ADME) properties was conducted through in silico analysis using the pkCSM online tool (https://biosig.lab.uq.edu.au/pkcsm/; accessed May 10, 2022).

Statistical analysis

All experiments were performed at least in triplicate per treatment condition. The mean value is averaged of 3 individual experiments. The data presented in this study were tested for normality test and were normally distributed. All data were expressed as the mean ± standard error of the mean (SEM). Cell viability data and ROS detection data were subjected to multiple group comparisons using a one-way analysis of variance (one-way ANOVA) followed by Tukey’s post hoc test. The significant difference between the allicin concentration in the presence and absence of hCMEC/D3 cells was assessed using an independent t-test. Statistical analyses were performed using the SPSS version 26 software, and GraphPad Prism software, version 10.2.3 (403). A value of p<0.05 was considered statistically significant.

Effects of allicin on the cell viability and ROS formation in hCMEC/D3 cells

To investigate allicin’s ability to penetrate the in vitro BBB model, the non-toxic concentrations of allicin were firstly determined on the hCMEC/D3 cells using the MTT assay. Figure 1A illustrates the results obtained after incubation of allicin with hCMEC/D3 cells for 3 h, revealing a significant reduction in cell viability at a concentration of 10 µg/ml compared to the untreated control group. Furthermore, results at 24 h demonstrated that allicin at the concentrations of 5 µg/ml exhibited significant decreases in cell viability (Fig. 1B). Additionally, morphological changes of the cells throughout the experimental process were shown in Figs. 1C–1D. At the toxicity concentrations of allicin, the cells were found rounding, cytoplasmic shrinking and cell death.

The formation of intracellular ROS within hCMEC/D3 cells after 3 h of allicin exposure was also investigated since the elevated ROS production can potentially overwhelm the cellular antioxidant capacity and lead to cell damage, which would be affected the in vitro BBB integrity. As shown in Fig. 2, the results revealed that allicin at a concentration of 10 µg/ml significantly increased the ROS level compared to the untreated control group. These findings suggest that allicin may be involved in inducing ROS-mediated damage in hCMEC/D3 cells. Moreover, these observations align with the trends observed in the cell viability assessments. Therefore, the concentrations of allicin (0.5, 1, 2, and 5 µg/ml), which were not significantly toxic to the cells, and did not significantly increase ROS generation, were selected to further investigate the ability of allicin to cross the in vitro BBB model.

The ability of allicin to cross the in vitro BBB model

The integrity of in vitro BBB models was verified on day 21 by TEER measurement and LY permeability assay. The %LY rejections before allicin testing were 97.61 ± 0.57% and those after the tests were 95.04 ± 2.26%. The non-toxicity concentrations of allicin (ranging from 0.5 to 5 µg/ml) were then employed to evaluate the ability of allicin to cross the in vitro BBB models over a 3-hour period. Surprisingly, the results from HPLC analysis of the samples from both AP and BL chambers of the in vitro BBB model revealed the absence of the allicin across all tested concentrations of 0.5, 1, 2, and 5 µg/ml (Figs. 3A–3H). The outcome was contrary to expectations. It was anticipated that if allicin failed to penetrate the in vitro BBB model, the allicin peak would be present in the AP samples due to a higher allicin concentration compared to the BL samples. However, this inconsistency led us to hypothesize that allicin degradation might occur during the experimental process. To test this hypothesis, allicin was added in the cell-free insert as a control. As illustrated in Figs. 3I–3J and Fig. S1, the HPLC results from the cell-free control insert tested with allicin 5 µg/ml demonstrated that the allicin peak was detected in both AP and BL samples, and with a retention time of approximately 8.8 min, as indicated by arrows. These findings revealed that the disappearance of allicin occurred only in the presence of hCMEC/D3 cells in the in vitro BBB model. Therefore, we hypothesized that allicin could be taken up by hCMEC/D3 cells.

The uptake of allicin by hCMEC/D3 cells

Cellular uptake experiments were performed to indirectly ascertain the potential uptake of allicin by hCMEC/D3 cells within the in vitro BBB model. Firstly, the non-toxicity concentrations of allicin (0.5–5 µg/ml) were applied to hCMEC/D3 cells at a density of 1 × 10⁴ cells/well and their cell-free wells for 3 h. The results showed that the allicin at concentrations of 0.5, 1, and 2 µg/ml were significantly reduced in the presence of hCMEC/D3 cells when compared to its cell-free wells. However, there was no statistical difference observed in the concentration of allicin at 5 µg/ml in the presence or absence of hCMEC/D3 cells (Fig. 4A). Thus, we postulated that the cell count utilized in this experiment might not be sufficient to uptake allicin at 5 µg/ml. The cell number used in both cellular uptake experiments and the in vitro BBB model remained the same at 1 × 10⁴ cells, although the cell culture duration of the in vitro BBB model (21 days) was longer than the cellular uptake experiments (1 day). Hence, the number of cells that we initially used was presumably much lower than that in the in vitro BBB model. The number of hCMEC/D3 cells was increased to test the hypothesis that the greater number of cells, the more uptake of allicin. As shown in Fig. 4B, the HPLC results also showed that the concentration of allicin at 5 µg/ml was significantly reduced in the presence of hCMEC/D3 cells at a density of 5 × 10⁴ cells/well compared to its cell-free well, suggesting that allicin was taken up by the hCMEC/D3 cells. As shown in Fig. 4C, the data further confirmed that the allicin uptake in hCMEC/D3 cells was proportional to the cell number as the concentration of allicin (5 µg/ml) was reduced following the increasing cell number. Allicin was completely taken up by hCMEC/D3 cells at a density of 5 × 10⁴ cells/well. To explore whether the allicin uptake of hCMEC/D3 cells was increased through the incubation period, the concentration of allicin was quantified every 30 min until the end of experiments at 3 h. The results showed that allicin was completely uptaken by hCMEC/D3 cells (5 × 10⁴ cells/well) at 0.5 h, while the concentration of allicin constantly remained the same throughout the experiments in the absence of hCMEC/D3 cells (Fig. 4D). This data indicated that allicin was rapidly taken up by the cells. Taken together, the results indicated cellular uptake of allicin by hCMEC/D3 cells.

Computational pharmacokinetic profile of allicin and its metabolite (GSSA)

The absorption, distribution, metabolism, and excretion parameters of both allicin and S-Allylmercaptoglutathione (GSSA) was presented in Table 1, with assessments conducted through the implementation of the pkCSM online software (Pires, Blundell & Ascher, 2015). In consideration of allicin molecular, the topological polar surface area (TPSA) was found to be below 140 Ų, facilitating cellular permeation. Furthermore, it is noteworthy that for molecules to cross the BBB, TPSA is determined to be less than 90 Ų (Samuel et al., 2019).

The TPSA value of allicin was 62.082 Ų, signifying its potential capability to cross the BBB. Conversely, this permeability profile was not observed for GSSA, which demonstrated a TPSA value of 146.749 Ų. Indeed, these findings involve to the model assessing blood–brain barrier permeability. Moreover, a logBB value exceeding 0.3 implies that the substance may penetrate to BBB easily. The blood–brain permeability-surface area product (logPS) serves as an alternative model that excludes systemic confounding factors. Due to the logPS value of GSSA being below -3, it was determined that the compound could not penetrate the central nervous system (CNS).

Allicin exhibits high bioavailability, with absorption occurring predominantly in the small intestine, as evidenced by its intestinal absorption values exceeding 90% and Caco-2 permeability value was above 1.30. Subsequently, allicin demonstrates a decreased volume of distribution at 0.71 L/kg, underscoring its preferential distribution within the plasma compartment as opposed to various tissues. In the context of metabolism parameters, it was observed that allicin and GSSA cannot serve as substrates for cytochromes 2D6 (CYP2D6) and 3A4 (CYP3A4). These predicting data suggest that allicin is not metabolized by the liver. Consequently, it is reasonable to infer that alternative metabolic pathways may facilitate the transportation of allicin to the brain. Both allicin and GSSA exhibited a lack of inhibitory effects on CYP2D6 and CYP3A4, indicating a lack of interference with the biotransformation processes mediated by these specific cytochrome P450 enzymes. Regarding their in-silico excretion profile, allicin display a faster clearance rate compared to GSSA (0.714 and 0.251 mL/min/kg).