Elucidation of the Transport Mechanism of Puerarin and Gastrodin and Their Interaction on the Absorption in a Caco-2 Cell Monolayer Model

Puerarin (PUR) and gastrodin (GAS) are often used in combined way for treating diseases caused by microcirculation disorders. The current study aimed to investigate the absorption and transportation mechanism of PUR and GAS and their interaction via Caco-2 monolayer cell model. In this work, the concentration in Caco-2 cell of PUR and GAS was determined by HPLC method. The bidirectional transport of PUR and GAS and the inhibition of drug efflux including verapamil and cyclosporine on the transport of these two components were studied. The mutual influence between PUR and GAS, especially the effect of the latter on the former of the bidirectional transport were also investigated. The transport of 50 μg·mL−1 PUR in Caco-2 cells has no obvious directionality. While the transport of 100 and 200 μg·mL−1 PUR presents a strong directionality, and this directionality can be inhibited by verapamil and cyclosporine. When PUR and GAS were used in combination, GAS could increase the absorption of PUR while PUR had no obvious influence on GAS. Therefore, the compatibility of PUR and GAS is reasonable, and GAS can promote the transmembrane transport of PUR, the effect of which is similar to that of verapamil.


Introduction
Puerarin ( Figure 1A, PUR), chemically known as 8-β-D-glucopyranose-4 ,7-dihydroxy isoflavone, is an important bioactive isoflavone glycoside, and it was isolated from several leguminous plants of the genus Pueraria, including Pueraria tuberosa (Willd.), Pueraria lobata (Willd.) Ohwi (Gegen in Chinese), and Pueraria thomsonii Benth [1][2][3]. PUR has the effects of dilating coronary arteries and cerebral blood vessels, improving microcirculation, anti-platelet aggregation, lowering blood pressure, and anti-oxidation [4]. It is mainly used to treat hypertension, coronary heart disease, angina pectoris, myocardial infarction, cerebral ischemia, migraine, sudden deafness and diabetes [5][6][7]. However, some of its physical properties, such as poor water solubility, poor permeability, and low oral bioavailability, lead to formulation of PUR is only for injection [8][9][10]. However, there are many reports of clinical adverse reactions caused by PUR injection, and the most serious is the hemolytic reaction, which is easily to cause the death of the patient when found and treated in an emergency. Therefore, there is widespread interest in how to improve the oral absorption of PUR. Gastrodin ( Figure 1B, GAS) is an active component of Gastrodia elata Blume, chemically known as 4-hydroxybenzyl alcohol-4-O-β-D-glucopyranoside. Gas has a variety of pharmacological effects such as sedation and sleeping, promoting intelligence, protecting neurons, lowering blood pressure, antioxidant, and improving microcirculation, etc. It is extensively used to treat dizziness, migraine, high blood pressure, stroke and epilepsy [11][12][13][14][15]. In the treatment of cardiovascular and cerebrovascular diseases, PUR and GAS are often used in combination. In addition, many Chinese medicine prescriptions usually contain these two components [16,17]. In our previous study [18][19][20], the physicochemical properties (including solubility and oil-water partition coefficient), anti-oxidation effect on microcirculation (including DPPH free radical scavenging, anticoagulation and antiplatelet aggregation in vivo and in vitro) and pharmacokinetics in rats were analyzed to explore the compatibility rationality and interaction between PUR and GAS. The results showed that (1) GAS can increase the solubility of PUR, (2) anticoagulation and anti-platelet aggregation effects of PUR and GAS can be enhanced when used in combination in a certain range of dosage, and (3) the combined use of GAS can promote the absorption, decrease the elimination rate, and prolong the mean residence time of PUR in rats. However, the mechanism of GAS promoting the absorption of PUR is not interpreted. Thus, it is worth studying and exploring the promotion mechanism of PUR and GAS.
To our knowledge, there have been some studies on the intestinal absorption mechanism of PUR. With regard to GAS, for its good oral bioavailability, few studies were concerned to the absorption of GAS. Liang [21] found that PUR is absorbed by passive diffusion without energy consumption, and there is no obvious directionality in transportation. Moreover, Zhang [22] proposed that the reason for the poor uptake exhibited by PUR alone is mainly due to its mode of transport, which is a passive diffusion of the cell monolayer. Cheng et al. [23] found that adding verapamil could enhance the absorption of PUR, so it is believed that the transcellular transport of PUR was also mediated by P-glycoprotein (P-gp) in addition to passive diffusion.
The Caco-2 cell line was derived from human colon adenocarcinoma and its morphology and function are similar to those of human intestinal epithelial cells, and it contains enzymes related to the brush border epithelium of small intestine [24]. Caco-2 cell model has been widely used in the study of drug uptake, efflux, transmembrane transport and other absorption mechanisms at home and abroad in the past decade because of its good correlation, reproducibility and applicability with oral drug absorption in intestinal tract [25][26][27][28][29].
Thus, in this study, based on the clinical application and combination of PUR and GAS, Caco-2 cells was used to study the absorption and transportation mechanism of PUR and GAS and their mutual influence, so as to lay an experimental foundation for their feasibility of the compatibility.

Preparation of Cell Samples
Transfer 30 µL of cell sample, 10 µL of p-hydroxy phenylethyl alcohol (IS) and 60 µL of methanol to an EP tube in turn, and vortex for 2 min. Then centrifuged at 12,000× g·min −1 for 10 min. Store the supernatant at 4 • C for subsequent analysis.

Method Validation
The specificity of the method was assessed by comparing the chromatograms of blank plasma, blank plasma spiked with internal standard/analyte, and cell samples. Calibration curves were established based on the peak area ratio versus nominal concentration. Intraday and inter-day precision and accuracy are expressed by evaluating the measurement results of QC samples at 0.1, 5, and 100 µg·mL −1 .

Specificity
The retention time for PUR, GAS, and p-hydroxy phenylethyl alcohol (IS) were 16.5, 8.6, and 14.8 min, respectively, and there were little interferential substances with the analytes and IS in the HBSS solution and cells. Representative chromatogram of analytes and IS in Caco-2 cell was shown in Figure 2.

Linearity
Due to the wide range of concentration span, two standard curves of the analytes are made. For PUR, the low linear rang is 0.025-2.5 μg·mL −1 , and the high linear rang is 2.5-100 μg·mL −1 . For GAS, the low linear rang is 0.1-5 μg·mL −1 , and the high linear rang is 5-100 μg·mL −1 . The calibration curves of PUR and GAS exhibited good linearity, and the regression equations with correlation coefficients and linear range were listed in Table 1. Table 1. Regression data and lower limit of quantitations (LLOQs) of low and high concentration of puerarin (PUR) and gastrodin (GAS) in Caco-2 cell model.

Analyte
Linear

Precision
Transfer 30 μL of each working solutions, and the samples were prepared in the same way as the cell samples at concentration of 0.1, 5, 100 μg·mL −1 . Each of the six samples were processed in parallel. Precisions were expressed by the relative standard deviation (RSD, %). The intra-day and inter-day precision R.S.D. values of PUR and GAS were all less than 10%.

Accuracy
The accuracy (%), i.e., relative recovery, compare the measured concentration with the actual concentration to obtain the method recovery rate, the difference value is the

Linearity
Due to the wide range of concentration span, two standard curves of the analytes are made. For PUR, the low linear rang is 0.025-2.5 µg·mL −1 , and the high linear rang is 2.5-100 µg·mL −1 . For GAS, the low linear rang is 0.1-5 µg·mL −1 , and the high linear rang is 5-100 µg·mL −1 . The calibration curves of PUR and GAS exhibited good linearity, and the regression equations with correlation coefficients and linear range were listed in Table 1. Table 1. Regression data and lower limit of quantitations (LLOQs) of low and high concentration of puerarin (PUR) and gastrodin (GAS) in Caco-2 cell model.

Analyte
Linear

Precision
Transfer 30 µL of each working solutions, and the samples were prepared in the same way as the cell samples at concentration of 0.1, 5, 100 µg·mL −1 . Each of the six samples were processed in parallel. Precisions were expressed by the relative standard deviation (RSD, %). The intra-day and inter-day precision R.S.D. values of PUR and GAS were all less than 10%.

Accuracy
The accuracy (%), i.e., relative recovery, compare the measured concentration with the actual concentration to obtain the method recovery rate, the difference value is the relative recovery rate. The recovery rates of PUR for low, medium and high (0.1, 5, 100 µg·mL −1 ) concentrations were 104.44%, 105.01%, and 100.17%, respectively; the recoveries for low, medium, and high concentrations of GAS were 108.71%, 107.16%, and 99.87%, respectively.

Transmembrane Resistance
The EVOM cell potentiometer was used to measure the transmembrane resistance of Caco-2 cells, and the transepithelial electrical resistance (TEER) was calculated according to Equation (1).
Before the transport experiment, R 1 (Ω) is the measured value of the cell group, R 0 (Ω) is the measured value of the blank group (without inoculated cells); in the transport experiment, R 1 (Ω) is the measured value of the administered cell group, R 0 (Ω) is the measured value of the blank cell group (before administration), and A (cm 2 ) is the area of the polycarbonate membrane per hole (The area of Transwell membrane used in this experiment is 1.12 cm 2 ).

Apparent Permeability Coefficient
The calculation of the apparent permeability coefficients (P app ) of the Caco-2 cell model refers to the data processing reported by Artursson [30] in 1991 through the P app of the Caco-2 cell monolayer, see Equation (2). The larger the P app value, the higher the permeability.
∆Q is the accumulative transport amount of drug (µg); ∆Q/ ∆t is the drug transport amount in unit time in the receiving pool, that is, the transport rate (v) (µg·mL −1 ); A is the same as the meaning in the Equation (1); C 0 is the initial mass concentration of the drug in the supply pool (µg·mL −1 ).
The drug efflux ratio (ER) is the ratio of P app(BL→AP) to P app(AP→BL) to the apparent permeability coefficient of the drug transported from the AP side to the BL side.
P app(BL→AP) is the P app of drug transported from the BL side to the AP side, and P app(AP→BL) is the P app of drug transported from the AP side to the BL side. It is generally believed that when P app(BL→AP) is close to P app(AP→BL) , the drug is absorbed by passive diffusion. When P app(BL→AP) /P app(AP→BL) > 1.5, it is indicated that the drug may involve in active transport mechanism.
Since the HBSS blank solution must be supplemented after each sampling, which is equivalent to a dilution effect on the penetration of the drug, the cumulative transport amount (∆Q) (µg) of the drug can be corrected by the following formula: C n is the permeability concentration of the nth sample (µg·mL −1 ); V R is the volume of the receiving cell (mL); V S is the sampling volume at each time point (mL, 0.2 mL in this experiment); ∑ n−1 i=0 C i is the sum of the concentration of samples taken from 0 to n − 1 time points (µg·mL −1 ). SPSS statistic 17.0 (SPSS Corporation, Chicago, IL, USA) was used for statistical analysis of the data, and the results were expressed as Mean ± SD.

Bidirectional Transport of Different Concentrations of PUR
When the concentration of PUR is 50 µg·mL −1 , the bidirectional transport amount increases linearly with time, indicating that within this concentration range, the absorption mode of PUR is just passive diffusion (Figure 3). From the AP side to BL side, the transport amount of PUR is concentration-dependent within 180 min, that is, the greater the concentration, the more transported. When the concentration is 100 or 200 µg·mL −1 , PUR is transported faster within 90 min, and the transport rate tends to be slow after that, indicating that the absorption of PUR is not simply passive absorption within this concentration range, and is protein-mediated active efflux transport mechanism ( Figure 3A). From the BL side to AP side, when the concentration of PUR is in the range of 100-200 µg·mL −1 , the cumulative transport amount of PUR decreases with the concentration increases, indicating that PUR transport is saturated from the BL side to AP side, which also suggests that there may be an active transport mechanism in the absorption of PUR within this concentration range ( Figure 3B). Moreover, the ER of 50, 100 and 200 µg·mL −1 PUR groups were 1.442 ± 0.073, 3.531 ± 0.129, 2.654 ± 0.693, respectively ( Table 2). The ratio of the former group was <1.5, and the ratio of the latter two groups were both >1.5. This also suggested that PUR is passive diffusion at 50 µg·mL −1 , and there may be an active transport mechanism in the concentration range of 100-200 µg·mL −1 .

Bidirectional Transport of GAS
This study only investigated the bidirectional transport of GAS at a concentration of 100 µg·mL −1 in the Caco-2 cell model ( Table 2 and Figure 4). The ER of 100 µg·mL −1 GAS was 1.191, which was less than 1.5. In Caco-2 cells, the trans-cell transport of GAS has no obvious directionality, and the transport rate is almost constant.

Effects of Verapamil on Bidirectional Transport of PUR
It can be seen from Figure 5A that when 100 µg·mL −1 PUR was combined with 100 µmol·L −1 verapamil, the ∆Q (BL→AP) of PUR was reduced, while the ∆Q (AP→BL) was basically unchanged, that is, the addition of verapamil can reduce the efflux of PUR and finally promote the absorption of PUR. In addition, the absorption of PUR increased (P app(AP→BL) increased from 1.285 to 1.413), the efflux decreased significantly (P app(BL→AP) decreased from 4.539 × 10 −6 cm·s −1 to 3.004 × 10 −6 cm·s −1 ), and the ER dropped from 3.531 to 2.126 (a decrease of 39.79%) ( Table 3).

Effects of Verapamil on Transport of GAS from AP to BL Side
For GAS, after combined with 100 µmol·L −1 verapamil, the ∆Q (AP→BL) increased, and there was no significant difference in the apparent permeability coefficient between the two groups ( Table 3 and Figure 5B).

Effects of Cyclosporin on Bidirectional Transport of PUR
It can be seen from Figure 5C, when 100 µg·mL −1 PUR is used in combination with 10 µmol·L −1 cyclosporine, the bidirectional cumulative transport amount of the Caco-2 cell monolayer model varies with time. The ∆Q (AP→BL) of PUR increased, while the ∆Q (BL→AP) was basically constant in the concentration range of 50-200 µg·mL −1 . That is, after adding cyclosporin, it appears to increase the absorption of PUR. In additional, the P app(AP→BL) increased significantly from 1.285 × 10 −6 cm·s −1 to 4.759 × 10 −6 cm·s −1 , while P app(BL→AP) decreased from 4.539 × 10 −6 cm·s −1 to 4.014 × 10 −6 cm·s −1 , a decrease of 76.11% (Tables 2 and 3).

Effects of Cyclosporin on Transport of GAS from AP to BL Side
For GAS, when combined with 100 µmol·L −1 cyclosporine, there is no significant difference in the ∆Q (AP→BL) and P app(AP→BL) ( Table 3 and Figure 5D).

The Interaction between PUR and GAS on Bidirectional Transport in the Caco-2 Cell Model
When 100 µg·mL −1 PUR is combined with 100 µg·mL −1 GAS, the biodirectional ∆Q were shown in Figure 6. In addition, the ∆Q (BL→AP) of PUR decreased, while there was no obvious change with the ∆Q (AP→BL) ( Figure 6A). In addition, the P app(AP→BL) of PUR increased (from 1.285 to 1.425, the P app(BL→AP) decreased significantly from 4.539 × 10 −6 cm·s −1 to 3.108 × 10 −6 cm·s −1 , and the ER decreased from 3.531 to 2.181, a decrease of 38.22% (Table 4).  As for Gas, when it was combined with PUR, the ∆Q (BL→AP) of GAS was significant increased, while the ∆Q (AP→BL) was basically unchanged ( Figure 6B). Compared with GAS alone, there was no significant difference in P app value and ER in the combination group. In addition, the ER of 100 µg·mL −1 GAS is 1.191, which is less than 1.5 (Table 4).
In summary, the transport of 50 µg·mL −1 PUR in Caco-2 cells has no obvious directionality, while the transport of 100 and 200 µg·mL −1 PUR presents a strong directionality, and this directionality can be inhibited by verapamil and cyclosporine. The transport of 100 µg·mL −1 GAS in Caco-2 cells has no obvious directionality, and the intestinal absorption could not be altered by verapamil and cyclosporine. When PUR and GAS were used in combination, GAS could increase the absorption of PUR while PUR had no obvious influence on GAS.

Discussion
In the present study, the rationality of the compatibility of PUR and GAS was verified from the perspective of Caco-2 cell transport. It is found that GAS enhanced the P app(AP→BL) significantly, which indicates that GAS can improve the PUR's intestinal absorption. For intestinal absorption is an important index reflecting the process of drug absorption, so the poor absorption of PUR may closely relate to the intestinal efflux. P-gp and MRP, its subfamily MRP2, are both expressed in the Caco-2 monolayer cell model, and they can efflux drugs from the inside of the cell to the outside of the cell [31,32]. Verapamil is a P-gp inhibitor, while cyclosporine is both a P-gp inhibitor and an MRP2 inhibitor [33,34]. Therefore, we chose these two inhibitors to elucidate the factors of how GAS enhances the absorption of PUR. Results revealed that the transport of 50 µg·mL −1 PUR in Caco-2 cells has no obvious directionality, and the transport rate is basically constant, indicating that the transport of PUR within this concentration range is a passive transport process. The transmembrane transport of 100 and 200 µg·mL −1 PUR in Caco-2 cells present a strong directionality, with ER both greater than 1.5, and this tropism can be inhibited by verapamil and cyclosporin. So, it is speculated that PUR is not only a substrate of P-gp, but also a substrate of MRP2. Cyclosporin is more obvious than verapamil to promote transmembrane transport of PUR. Moreover, in addition to passive diffusion, there is also an active transport mode for PUR across the membrane. This result is consistent with the research of Cheng et al. [23]. In addition, it is well known that the efflux function of P-gp is saturated. When the binding sites between P-gp and drugs are almost full, the efflux function decreases significantly [35]. In our study, the ER of 100 and 200 µg·mL −1 PUR was 3.531 and 2.654, respectively, which was consistent with the previous research [35]. The transmembrane transport of 100 µg·mL −1 GAS in Caco-2 cells has no obvious directionality, and the transport profile could not be altered by verapamil and cyclosporine, suggesting that the transport of GAS is a pure passive process. However, as GAS has no oral malabsorption, thus we just investigated 100 µg·mL −1 GAS, did not explore the bidirectional transport of different concentrations of GAS in Caco-2 cells, and the P app(BL→AP) experiment of GAS was also omitted.
When GAS and PUR were used in combination, for PUR, the absorption index P app(AP→BL) increased from 1.285 × 10 −6 cm·s −1 to 1.425 × 10 −6 cm·s −1 , the efflux index P app(BL→AP) decreased significantly from 4.539 × 10 −6 cm·s −1 to 3.108 × 10 −6 cm·s −1 , and the ER decreased by 38.22%. This is consistent with our previous pharmacokinetic results in rats, which showed that the combined administration of GAS and PUR could in-crease the absorption, decrease the clearance rate, and prolong the mean retention time, and the bioavailability of the two components when compared with the single-administered group [18]. It is speculated that GAS can play a role similar to verapamil, that is, GAS is a P-gp inhibitor and can promote the transmembrane transport of PUR, which is similar to verapamil. However, whether P-gp and MRP2 have a key influence on the absorption of PUR remains to be further studied.
In this study, the P app values of PUR and GAS were 1.766 × 10 −6 cm·s −1 and 2.866 × 10 −6 cm·s −1 , respectively. The P app of PUR is smaller than that of GAS, which is consistent with the determination of the oil-water partition coefficient in our previous study [36]. Studies [37] have shown that, the permeability coefficient of drugs with complete absorption is high (P app > 1 × 10 −6 cm·s −1 ), and that of drugs with incomplete absorption is low (P app < 1 × 10 −7 cm·s −1 ) [38,39]. It can be inferred that the two components are well absorbed (not absorbed well) in vivo. Actually, previous studies have shown that PUR is poorly absorbed and GAS is well absorbed [18]. This is because there was a good correlation between oral absorption and P app for passive diffusion drugs. However, for active transport drugs, the P app obtained from Caco-2 cell experiments can only be used as a qualitative rather than quantitative indicator of in vivo absorption.

Conclusions
The absorption and transport mechanism of PUR and GAS and their interaction were studied and we try to clarify the rationality of their absorption mechanism and compatibility application from the cellular level. The research results show that the compatibility of PUR and GAS is reasonable, and GAS can promote the transmembrane transport of PUR, the effect of which is similar to that of verapamil, which may provide a certain theoretical basis for its combined application.