Intestinal Absorption of Ergostane and Lanostane Triterpenoids from Antrodia cinnamomea Using Caco-2 Cell Monolayer Model

Abstract Antrodia cinnamomea is a precious medicinal mushroom. It exhibits promising therapeutic effects on cancer, intoxication, hypertension, hepatitis, and inflammation. Its major bioactive constituents are ergostane and lanostane triterpenoids. In this study, we used intestinal Caco-2 cell monolayer model to reveal the intestinal absorption property of 14 representative triterpenoids from A. cinnamomea. The bidirectional transport through the monolayer at different time points was monitored by a fully validated LC/MS/MS method. In the case of pure compounds, ergostanes 5 (25R-antcin H), 6 (25S-antcin H) and 10 (25R-antcin B) could readily pass through the Caco-2 cell layer, whereas lanostanes 13 (dehydroeburicoic acid) and 14 (eburicoic acid) could hardly pass through. When the cells were treated with A. cinnamomea extract, antcins A, B, C, H and K (1–6 and 9–11) were absorbed via passive transcellular diffusion, and showed high PAB and PBA values (> 2.5 × 10−5 cm/s). Meanwhile, the lanostanes dehydrosulphurenic acid (8), 15α-acetyldehydrosulphurenic acid (12), 13 and 14 exhibited poor permeability. Transport features of these compounds were consistent with their pharmacokinetic behaviors in rats. This study could also be helpful in predicting the intestinal absorption of A. cinnamomea in human. Graphical Abstract Electronic supplementary material The online version of this article (doi:10.1007/s13659-015-0072-4) contains supplementary material, which is available to authorized users.

Recently, we have reported the metabolism and pharmacokinetics of triterpenoids of A. cinnamomea in rats [11]. Interestingly, we found that the lanostanes and ergostanes showed remarkably different pharmacokinetic patterns, which were closely related with their chemical structures. High-polarity ergostanes (antcins H (5/6) and K (1/2), with two or three hydroxyl groups, respectively) were the major plasma-exposed components. The low-polarity ergostanes antcins B (9/10) and C (3/4), containing zero or one hydroxyl group, respectively, could also get into circulation, though exhibiting much lower plasma concentrations. The D 7,9(11) lanostanes remained in the plasma at a low concentration for a relatively long time. The D 8 lanostanes, however, could not get into circulation. The pharmacokinetic patterns of different types of triterpenoids may be correlated with their intestinal absorption.
No interfering peaks were observed in the chromatogram of blank HBSS medium samples, indicating good specificity of the detection method. The linearity of each calibration curve was determined by plotting the analyte/internal standard peak area ratio (Y) against the nominal concentration of analytes (X). The calibration curves showed good linearity with correlation coefficients (r 2 ) of above 0.99. The regression equations and dynamic ranges were listed in Table S1. Accuracy and precision of the method was assessed by testing QC samples at three different concentrations in the same day for five times and on three consecutive days. The intra-and inter-day precisions ranged from 0.87 to 15.66 and 0.56 to 13.41 %, respectively (Table S2). The accuracy was between 86.46 and 114.83 % for LQC, MQC, and HQC, except that the accuracy of 10 in HQC was 83.22 % and 12 in LQC was 84.72 %.
The extraction recoveries and the matrix effects of analytes were determined as we had previously reported [11]. The method recoveries were determined by comparing the measured concentration of un-extracted QC samples (analytes were diluted in methanol and added to freeze-dried HBSS matrix) to the nominal concentration. The extraction and method recoveries at three different QC concentrations were between 85.14-119.73 and 85.73-119.94 % (Table S3). The matrix effects were not significant for most analytes, with ion suppression ranged from -11.94 to 9.01 % (Table S3). Stability was determined at high and low concentrations after 3-day storage at -80°C. All the analytes showed variations between -18.33 and 10.38 %, indicating the samples were stable within 3 days (Table S4) (Fig. S1).

Leakage Marker Fluorescein
The integrity of cell monolayers was also assessed by the transport of the leakage marker sodium fluorescein. On day 21, the Caco-2 monolayers were rinsed twice with prewarmed HBSS and were then incubated with the same solution at 37°C for 30 min, on a shaker at 50 rpm. The fluorescein was pre-dissolved in 1 mM NaOH/H 2 O (1:10, v/v), and then diluted to 1 mM with an appropriate volume of HBSS. The sodium fluorescein and blank HBSS were added to the AP (0.5 mL) and BL side (1.5 mL), respectively. At 30, 60, 90, and 120 min, each 0.3 mL of the solution from BL side was collected, and replaced with an equal volume of HBSS. The sodium fluorescein intensity was determined with a Hitachi F-4500 fluorometer (excitation wavelength 490 nm, emission 525 nm). As shown in Table S5, the P app value of sodium fluorescein for AP to BL side was less than 1.0 9 10 -6 cm/s, indicating that integrity and permeability of the cell monolayer was desirable [14].

Alkaline Phosphatase Activity
Alkaline phosphatase activity was determined in the cytosol of Caco-2 cells. The cell monolayer was washed three times with PBS, and then collected into 1 mL homogenizing buffer (0.25 M sucrose in 10 mM Tris-HCl pH 7.4). Cells were centrifuged (5 min, 100009g) and re-suspended in 100 lL of homogenizing buffer before being disrupted by five freeze/ thaw cycles in liquid nitrogen and 37°C water bath. The cytosol was collected after centrifugation (20 min, 100009g, 4°C) and stored at -80°C before analysis. The total intracellular protein was determined by the Lowry method, and an AKPase kit was used to determine alkaline phosphatase activity [15]. Alkaline phosphatase activity was calculated as follows: DA/min represents variety of absorbance per minute, e = 8.5 (absorbency per mole), V t is the total volume of the mixture (mL), V s is volume of the sample (mL), and d is thickness of sample cell.   (14) Intestinal Absorption of Ergostane and Lanostane Triterpenoids 239 The absorbance of each well was measured at 520 nm by the Infinite Ò M200 Pro (Tecan, Mannedorf, Switzerland). Alkaline phosphatase activity was calculated using the above formula and expressed as U/g protein. U/g was defined as that 1 mg of phenol was generated when 1 g of cellular protein reacted chemically with substrate for 15 min at 37°C. The results of alkaline phosphatase activity of the Caco-2 cell monolayers were illustrated in Table S6. The activity increased from 309.8 U/g protein (day 8) to 444.8 U/g protein on day 15 and 533.4 U/ g protein on day 21, which was consistent with previous reports [15].

Bidirectional Transport Experiments of Single Triterpenoids
Considering the complex chemical composition of A. cinnamomea, the transport features of five single compounds (5, 6, 10, 13 and 14) were studied. These five compounds belong to high-polarity ergostanes (5 and 6), low-polarity ergostanes (10), D 7,9(11) lanostanes (13), and D 8 lanostanes (14), respectively, and represent the major types of A. cinnamomea compounds. In the Caco-2 cell monolayer model, permeability was described by P AB values (the apparent permeability coefficients from AP to BL side). The P AB values of well-transported and poorly transported compounds were [ 1 9 10 -5 and \ 1 9 10 -6 cm/s, respectively [16,17]. P AB values of the marker compounds propranolol (3.81 ± 0.34 9 10 -5 cm/s) and atenolol (7.76 ± 4.65 9 10 -7 cm/s) in our experiments were consistent with previous reports (Table S5). The P AB values of 5, 6 and 10 (as pure compounds) across Caco-2 cells were all greater than 1 9 10 -5 cm/s (Table 1), indicating good permeability. Moreover, the efflux ratios (P BA /P AB ) of 5, 6, and 10 were [ 2.0 (2.26, 2.05 and 2.07, respectively), indicating an involvement of apical efflux transporters. Their transportation could involve active transport [18][19][20]. 10 showed lower P AB and P BA values than 5/6. This might be related to its lower polarity (containing one hydroxyl group), or due to its metabolism. We detected two hydrogenated metabolites of 10 in the AP side (M1 and M2, metabolic studies in rats [11], and could be a factor to change its transport behavior. In contrast to ergostanes, lanostanes 13 and 14 could not be detected in all samples of different time points, indicating poor permeability through the cell monolayer. This could also result from poor water solubility.

Bidirectional Transport Experiments of A.
cinnamomea Extract Based on the study of bidirectional transport of five represent single compounds, the Caco-2 cell monolayer was treated with A. cinnamomea extract, and the transportation of 14 triterpenoids (1-14) were studied. Compared to the single form, P BA values of 5, 6, and 10 in A. cinnamomea extract (containing 2.61, 3.66, 2.82 lM of 5, 6, and 10, respectively) decreased significantly by 50-60 % to 3.0-3.8 9 10 -5 cm/s (Fig. 3a). However, no significant difference was observed for the P AB values. Accordingly, efflux ratios of 5, 6, and 10 decreased to lower than 1.0 (0.88, 0.99, and 0.94, respectively), indicating their intestinal absorptions were dominated by passive diffusion [20,21] (Table 1). These data suggested that A. cinnamomea might contain transporter inhibitors (e.g. P-gp or MRP inhibitors) [19]. Moreover, though 25R/Santcin H (5/6) were epimers, they showed different transport capacities. The transport rate of 5 was higher than that of 6, both in single form and in the extract (Fig. 3b). As a result, although the content of 5 was lower than 6 in A. cinnamomea extract, its concentration was higher than 6 in the basolateral side (Fig. 3c).
The P app values of ergostanes 1-4, 7, 9 and 11 were summarized in Table 1 and Fig. 4. They showed favorable permeability (P app [ 2.5 9 10 -5 cm/s). Ergostanes have high polarity than lanostanes, and their log P values were between 3 and 4.7. This was consistent with previous reports [12,22], where the compounds with moderate log P values (3-4) exhibited higher permeability. Meanwhile, among different ergostanes, the permeability was not significantly correlated with log P values, though the polarity (number of hydroxyl groups) was inversely correlated with log P values. To all these ergostanes, their efflux ratio (B1.0) indicated that they were absorbed mainly by passive diffusion [20,21]. Moreover, as shown in Fig. 5, we found the transport rates of 25R ergostane epimers (2, 4, 5 and 10) from AP to BL side were significantly higher than 25S ergostane epimers (1, 3, 6 and 9) in A. cinnamomea extract (Fig. 5). Since that the absorption rates are closely related to physico-chemical properties (such as log P and hydrogen bonding capacity) of the drug molecule [12], the chemical environment of C-25 might be an important factor for transportation. In addition, the transport rate of 2 was higher than 100 % (112 %), probably due to the transformation of other compounds into 2 [11]. D 7,9(11) lanostanes (8,12,13) and D 8 lanostane (14) exhibited very low solubility and permeability, both in the   through the Caco-2 cell monolayer with P AB value of 3.74 ± 0.43 9 10 -5 cm/s, though its transport rate (38.8 % at 180 min) were lower than ergostane triterpenoids ([ 49.5 % at 180 min).

Conclusion
In this work, absorption properties of 14 ergostane and lanostane triterpenoids in A. cinnamomea were investigated using the Caco-2 cell monolayer model. The concentrations were determined by using a fully validated LC/MS/MS method. As pure compounds, ergostanes 25R/S-antcin H (5/6) and 25R-antcin B (10) exhibited high permeability, and their transportation might involve efflux transporters (efflux ratio [ 2.0). When the cell monolayer was treated with A. cinnamomea extract, all ergostanes (antcins A, B, C, H, K, log P 3.0-4.7) showed favorable permeability (P app [ 2.5 9 10 -5 cm/s), and they were mainly absorbed via passive diffusion (efflux ratio B 1.0). Transportation rate of 25R-ergostanes was higher than 25S-ergostanes. Meanwhile, lanostanes (log P [ 5.4) were barely permeable. The transport features of ergostane and lanostane triterpenoids through the Caco-2 cell monolayer were

Chemicals and Reagents
The fruiting bodies of A. cinnamomea were cultivated by Professor Yew-Min Tzeng at Chaoyang University of Technology, Taiwan. A voucher specimen was deposited at School of Pharmaceutical Sciences, Peking University, Beijing, China. Reference compounds 1-14 were isolated by the authors [6]. The purities were above 98 % as determined by HPLC/UV analysis. Their structures are given in Fig. 1. HPLC-grade acetonitrile, methanol and formic acid were purchased from Mallinckrodt Baker (Phillipsburg, NJ, USA). Ultra-pure water was prepared with a Milli-Q water purification system (Millipore, Billerica, MA, USA). The other reagents were of analytical grade.

Sample Preparation
Representative compounds 5, 6, 10, 13 and 14 were respectively dissolved in DMSO (9 lL) and then diluted with HBSS (3 mL) to the concentration at 10 lM as cell treatment solution. The final concentrations of DMSO were controlled below 0.3 % (v/v) to ensure safety to the cells [23].
The dried fruiting bodies of A. cinnamomea (30 g) were powdered and extracted with 500 mL ethanol for three times (2 h for each time) under reflux. The extracts were combined, concentrated in vacuo, dispersed in H 2 O and freeze-dried to obtain the A. cinnamomea extract (33.3 % yield). Before the bidirectional transport experiments, 250.13 lg A. cinnamomea extract was suspended in 0.03 mL DMSO and then diluted with HBSS to prepare the extract solution (25 lg/mL).

Cell Culture
The human colon adenocarcinoma cell line Caco-2 was purchased from the Cell Resource Center, Peking Union Medical College (CRC/PUMC, China), and was cultured in a humidified atmosphere of 5 % CO 2 at 37°C. Caco-2 cells were cultured in DMEM with 10 % FBS (inactivation at 56°C for 30 min), 1 % NEAA, and 1 % penicillin and streptomycin solution. The culture medium was changed every other day during cell growth and differentiation. On achieving 80-90 % confluence, the cells were rinsed with 4 mL pre-warm PBS (pH 7.4) and split using 1 mL trypsin. The cells were then seeded on polycarbonate membranes of cell culture inserts at a density of 1.5 9 10 5 cells/mL and placed in 12 well cell culture clusters. The transepithelial electrical resistance (TEER) using a Millicells Ò ERS-2 (Millipore, USA) and accumulation of fluorescein were assessed to reflect the tightness of intercellular junctions [24]. Reference compounds propranolol and atenolol were used to validate the Caco-2 cell monolayer model. They represent paracellular flux and transcellular flux compounds, respectively [16,17]. Differentiation of Caco-2 cells was checked on days 8, 15, 21 by determining the alkaline phosphatase activity using the AKP assay kit, a brush border enzyme marker [15,25]. The Caco-2 cell monolayers were used for transport experiments on day 21 post-seeding with TEER values [500 XÁcm 2 .

Bidirectional Transport Experiments
Caco-2 cells were obtained from CRC/PUMC at passage 18 and all experiments were performed from passages 24-35. Before the transport experiments, the cell monolayer was washed three times with HBSS. Then, the plates were incubated in fresh HBSS for 30 min at 37°C. The experiments were conducted by adding the samples (pure compound solution (10 lM) and A. cinnamomea extract solution (25 lg/mL), respectively) to either the apical (AP, 0.5 mL) or basolateral side (BL, 1.5 mL), while the receiving chamber contained the corresponding volume of pre-warmed HBSS. Every experiment was repeated three times, and the plates were incubated in an orbital shaker at 37°C, 50 r/min. To assess the drug transport, at the incubation time of 30, 60, 90, 120, 150 and 180 min, a 0.3 mL aliquot was removed and was immediately replenished with an equal volume of HBSS [15,18]. The samples were freeze-dried and then dissolved in 0.3 mL of methanol. The solution was filtered through a 0.22-lm membrane for LC/ MS/MS analysis. The transport rate was calculated as the ratio of cumulative concentration in the receiver to the donor side 9 100 %. The apparent permeability coefficient was indicated by the absorption rate constant P app . It could be calculated as P BA (measured in BL to AP direction) or P AB (measured in AP to BL direction), using the equation of P app = (dQ/dt)/(A 9 C 0 ). dQ/dt is the rate at which the compound appears in the receiver chamber (lmol/s), A is the surface area of the filter membrane (1.12 cm 2 ) and C 0 is the initial concentration in the donor chamber (lmol/mL). Efflux ratio (ER) was calculated by the equation of ER = P BA /P AB , where P BA is the P app value measured in BL to AP direction, and P AB is the P app value measured in AP to BL direction.

Calibration Standards and Quality Control Solutions
Reference standard compounds were dissolved in DMSO to prepare individual stock solutions (70 lM for 1-4, 7 and 9; 140 lM for 8, 10, 11, 13 and 14; 350 lM for 5, 6 and 12). These stock solutions were mixed and then serially diluted using HBSS to obtain calibration standard solutions (20, 10, 5, 2.5, 1, 0.5, 0.2, and 0.1 lM for each compound). Quality control (QC) stock solutions were prepared at three concentration levels as high QC (HQC), middle QC (MQC), and low QC (LQC), based on linear ranges of the analytes. All solutions were sealed and stored at -20°C until use.

LC/MS/MS Analysis
The LC/MS/MS method we had previously established for the pharmacokinetics study of A. cinnamomea was used [11]. Data were processed by Xcalibur 2.0.7 software (ThermoFisher).

Statistical Analysis
Results were expressed as the mean ± SD, of which the mean value was the average of at least three replicates. Analysis of variance (ANOVA) was used to test the statistical significance of differences between groups. Statistical significance in the differences of the means was determined by Student's t test, with a significance level of P \ 0.05. The logarithm of octanol-water partition coefficient (log P) was calculated with Pallas 3.3 (CompuDrug International, Inc., USA).