A Study of Intestinal Absorption of Bicyclol in Rats: Active Efflux Transport and Metabolism as Causes of Its Poor Bioavailability

Purpose. To determine the possible mechanism of poor bioavailability of bicyclol, and clarify the respective contribution of P-glycoprotein (P-gp) and Cytochrome 3A (CYP3A). Methods. Rat in situ single-pass intestinal perfusion and Caco-2 cell monolayer model with selective inhibitors of CYP3A and P-gp were employed. Results. In rat intestinal perfusion, bicyclol (50µM) appearance in mesenteric blood (P blood) was increased 3, 12 and 16-fold after addition of inhibitors of P-gp (LSN335984), CYP3A (troleandomycin, TAO) or P-gp and CYP3A (Cyclosporin A, CsA), respectively, whereas permeability of midazolam (CYP3A substrate only) was unchanged after addition of LSN335984 and increased 5 fold after addition of TAO. Moreover, the cumulative amount of bicyclol in mesenteric blood was increased at concentration range 10-100µM of bicyclol in perfusate. The basolateral to apical permeability value of bicyclol in Caco-2 monolayer was significantly deceased by LSN335984 and CsA. Conclusions. The poor bioavailability of bicyclol is mostly due to P-gp mediated efflux and metabolism by CYP3A in intestine, with CYP3A making more contribution than P-gp.


INTRODUCTION
Bicyclol(4,4′-dimethoxy-5,6,5′,6′-Bis(dimethylen e-dioxy)-2-hydroxymethyl-2′-methoxy-carbonylbi phenyl; Fig 1 ), a new synthetic anti-hepatitis drug, has been used in patients with chronic hepatitis B to improve liver function and partially inhibit hepatitis B virus replication in China (1).A variety of pharmacological studies showed that bicyclol markedly protected against experimental liver injury induced by certain toxins such as acetaminophen, concanavalin A, carbon tetrachloride, lipopolysaccharide, D-galactosamine and dimethylnitrosamine (2)(3)(4)(5).Pharmacokinetic data indicated the poor bioavailability (9%) of bicyclol in rats.In addition, an earlier study demonstrated that bicyclol was mainly metabolized by CYP 3A in rat and human liver microsomes (6).Biotransformation of drugs on passage across the enterocyte during oral absorption has been shown to significantly contribute to the first-pass effect of a number of compounds (7).It has been further suggested that the intestinal metabolism may be increased if the compound is also a substrate of an apically localized efflux transporter, such as P-gp (8-10).In small intestine, P-gp has been localized to the apical membrane of intestinal epithelial cells, consistent with its role in effluxing compounds back into the intestinal lumen (11).

CYP3A4
is the major oxidative drug-metabolizing enzyme found in the intestine and is localized in the endoplasmic reticulum of enterocytes (12,13).The extensive overlap in substrate specificities, tissue, localization, and coinducibility of P-gp and CYP3A4 has led to the hypothesis that these two proteins work together to protect the body from absorption of harmful xenobiotics, including drugs ( 14).Therefore, it is more likely that the co-effect of P-gp and CYP3A4 attenuates the absorption of compounds and leads to low bioavailability.
In the present study, in situ single-pass perfusion and Caco-2 models were used to determine the possible mechanism of the low bioavailability of bicyclol.

Chemicals
Bicyclol and dimethyl dicarboxylate biphenyl (DDB) were kindly provided by Beijing Union Pharmaceutical Plant (Purity >99%).LSN335984 was a gift of Eli Lilly Co, USA.TAO was purchased from Sigma-Aldrich chemical (St.Louis, MO, USA).CsA was a product of Norvatis Co. (Beijing,China).Acetonitril (HPLC grade) was obtained from Fisher (Fisher Scientific, United Kingdom).Caco-2 cell were obtained from Cell Center, Institute of Basic Medical Research.Minimum Essential medium, fetal bovine serum, 0.05% trypsin ethylenediaminetetraacetic acid, penicillin-streptomycin were obtained from Gibco Co.Other chemicals were of analytical grade and obtained from the local market.

Transport experiments
Caco-2 cells (cultured for 21 days ) were rinsed for three times with HBSS and equilibrated at 37℃ for one hour before transport experiment (16,17).Subsequently, 40µl of HBSS containing bicyclol (50µM) was applied in the apical or basolateral side, respectively.A 50µl of aliquot was taken from the receiver side at different time intervals (30, 60, 90, 120, 180, and 240 min) and an equal volume of blank HBSS was immediately supplemented to maintain a constant volume.

Inhibition studies
To examine the influence of LSN335984 and CsA on the transport of bicyclol, CsA (10µM) and LSN335984 (2µM) were loaded into both sides of monolayer, simultaneously or 30 minutes before addition of bicyclol (50µM).Samples (50µl each) were taken at different time intervals (30, 60, 90, 120, 180, and 240 min) from the receiver side and the same volume of solution was supplemented after each sampling.

Animals
Male Sprague-Dawley rats (280-350g) were obtained from the Animal Center of Chinese Academy of Medical Sciences.The animals were maintained on a 12-h light/dark cycle at the IMM Animal Care Facility and acclimatized for at least 3 days before the study.Rats were fasted overnight with free access to water and were anesthetized with 10% chloral hydrate at a dose of 350mg/kg by intraperitoneal injection.All animals were kept at 37°C during the surgery and throughout in situ intestinal perfusion.Fresh heparinized donor blood (13U/ml) was collected from donor rats immediately before surgery by cardiac puncture (18).All animal experiments were in compliance with the guidelines of China for animal care, which was conformed to the internationally accepted principles in the care and use of experimental animals.

Surgical procedures
The surgical procedures used in single-pass rat jejunum experiment were similar to those described elsewhere, with slight modifications (19).Briefly, the rat was heparinized (90U/kg) via right jugular cannula ten minutes before mesenteric cannulation and the infusion rate of donor blood was 0.4ml/min based on the outlet rate of mesenteric vein.An approximately 7 to 15cm long segment of jejunum was chosen based on the mesenteric vein to be cannulated, and perfusate was infused into the intestinal segment at 0.3mL/min.

Single-pass intestinal perfusion
The perfusate (Hank's balanced salt solution) with pH adjusted to 7.4 at 37℃ was prepared.Phenol Red (50µM) was added to serve as a nonabsorbable marker for water flux.The perfusion buffer with bicyclol (10-100µM) or midazolam (50µM) in the absence or presence of 5µM LSN335984, 20µM TAO or 10µM CsA was infused into the jejunum individually.The blood from mesenteric vein and perfusate were collected at five minutes interval.The plasma separated by centrifugation and perfusate were immediately frozen at -80℃ until analysis.

Sample preparation
A 50µl of aliquots from Caco-2 cell culture or 40µl of perfusate was mixed with 70µl of acetonitrile and 10µl of DDB (10µg/ml) was added as internal standard.After centrifugation at 12,000rpm for 5min twice, 20µl or 1µl of each sample was taken for LC or LC-MS analysis, respectively.
Plasma sample (150µl) with 10µl of DDB (5µg/ml) was mixed with 1mL of dichloromethane.A 2µl of extraction was used for LC/MS determination.For the analysis of midazolam, perfusate sample (40µl) at each time point was mixed with 10µl of omeprazole (25µg/ml) as internal standard, and then mixed with 70µl of acetonitrile.After centrifugating at 12,000rpm for 5min twice, 2µl of extraction was analyzed by HPLC.Plasma sample (150µl) with 10µl of acetaminophen (20µg/ml) served as internal standard was mixed with 1ml of diethyl ether, and then extracted with the above method.

LC/MS analysis of bicyclol and midazolam
Bicyclol and midazolam in plasma and perfusate were analyzed by LC/MSD Trap XCT mass spectrometer equipped with a series 1100 HPLC system consisting of a quaternary pump, automatic solvent degasser, autosampler and an automatic thermostatic column compartment.Analysis of bicyclol was performed as follows: Briefly, the sample was loaded onto the analytical column (Agilent C18 column, 2.1×100 mm, 3.5 µm).The mobile phase consisted of water and methanol (45:55,v/v), the column temperature was maintained at 30°C, the flow rate was 0.2 ml/min, and the total run time was 9 min.Using selective ion monitoring, signals for [M+Na] + ions of bicyclol (m/z 413), and the internal standard DDB (m/z 441) were obtained.The lower limit of quantitation was 3.3 ng/ml in rat plasma and intra-and inter-day accuracy ranged from 102-107% and 102-110%, respectively.The absolute recovery values of bicyclol at plasma concentrations of 6.7, 66.7 and 333.3 ng/ml were 85.3±4.9, 94.6±5.9 and 94.6±2.6%,respectively.
Midazolam in plasma was analyzed with the same system.The mobile phase was consisted of water and methanol (30:70,v/v) containing ammonium acetate (5mM).Using selective ion monitoring, signals for [M+H] + ions of midazolam (m/z 326) and the internal standard acetaminophen (m/z 152) were obtained.

Data Analysis
The permeability coefficient (P app ) of bicyclol in Caco-2 cell was calculated as previously described (20).The concentrations obtained from the perfusate were corrected by the changes in water flux at each time interval using (Eq.1): The permeability of bicyclol and midazolam across rat intestine was calculated based on the appearance of drug in blood (P blood ) using the following equations ( 19): Where r is the radius of the perfused intestinal lumen, l is the length of the segment , ΔM B /Δt is the rate of drug appearance in mesenteric blood and <C> is the logarithmic mean concentration of drug in lumen.Statistically significant difference between two groups was evaluated by Student's t test.The prior level of significance was set at p < 0.05.

Caco-2 cell monolayer model studies
The absorptive P app of bicyclol in Caco-2 cell was 1.6±0.3×10 - cm/s, which was comparable to that from well-absorbed compounds (21).The basolateral to apical/ apical to basolateral (B-A/A-B) transport ratio of bicyclol was 1.94, indicating a significant efflux transport existing in Caco-2 cell.LSN335984 and CsA, the selective inhibitors of P-gp, were used to further determine how each inhibitor affected the transepithelial transport of bicyclol.As a result, the transport ratios of bicyclol were decreased to 1.00 and 1.19, respectively, after addition of LSN335984 or CsA (Table 1).

Absorption and metabolism of bicyclol
The profiles of bicyclol (50 µM) appearance in mesenteric vein in the absence or presence of inhibitors of P-gp (LSN335984), CYP3A (TAO) or co-inhibitor of P-gp and CYP3A (CsA) are shown in Figure 2, and the relevant data are presented in Table 2. P blood of bicyclol at a concentration of 50 µM was 0.04±0.01×10 - cm/s and increased to 0.20±0.06×10 - cm/s, 0.55±0.18×10 - cm/s and 0.69±0.39×10 - cm/s in the presence of LSN335984, TAO or CsA, respectively (Table 2).Therefore, the appearance of bicyclol in mesenteric blood (P blood ) was increased by 3-fold in the presence of LSN335984, and elevated more significantly with TAO or CsA (Table 2 and Fig. 2).The data suggested that the co-effect of P-gp and CYP3A was significantly limiting the absorption of bicyclol across the intestine.Table 2 also presents the P blood and cumulative amount in blood data as a function of the perfusion concentration.When the concentration was increased from 10 µM to 50 µM, there was a nonlinear increase (50%) in the appearance of the cumulative amount of bicyclol in mesenteric blood, indicating that P-gp was effectively pumping the drug back into the intestinal lumen and limiting the drug absorption.However, when the bicyclol concentration was further increased to 100 µM, the unchanged amount of bicyclol in mesenteric blood was increased by 8.8 fold, suggesting that there might be a saturation of intestinal efflux and/or metabolism of bicyclol at the highest dose (Table 2, Fig. 3).

Absorption and metabolism of midazolam
Midazolam, a CYP3A substrate not transported by P-gp, was used as a negative control in the present study.The P blood of midazolam at steady state was 0.05±0.03×10 - cm/s and unchanged by addition of LSN335984 (0.06±0.03×10 -4 cm/s).However, it was significantly increased by 5 fold by addition of TAO (Table 3).

DISSCUSION
This is a comprehensive study that employed both in vivo and in vitro models to further characterize the absorption and metabolism properties of bicyclol.The single-pass rat intestinal perfusion model with mesenteric blood sampling provided an isolated in vivo system to study the respective role of intestinal efflux transporter-P-gp and drug metabolizing enzyme-CYP3A (22).On the other hand, Caco-2 cell model was used to study the transport of bicyclol in vitro.The AP-to-BL and BL-to-AP transport of bicyclol (50µM) across Caco-2 cell monolayers in the absence (control) or presence of 2µM LSN335984 (P-gp inhibitor) or 10µM CsA (CYP3A and P-gp inhibitor) was examined.Data are expressed as mean± SD, and n = 3.The statistically significant difference is between control and other experimental group.*P< 0.05.
Data from Caco-2 cell and rat intestinal perfusion suggested that bicyclol is a well absorbed compound comparable to other well-absorbed markers (21).However, the high permeability did not lead to high-level of bicyclol in the mesenteric blood ( 23), indicating that p-gp mediated efflux and drug metabolizing enzyme may both contribute to the poor bioavailability of bicyclol.It was found that the permeability of bicyclol (P app A-B = 1.6±0.3×10 - cm/s) in Caco-2 cell was similar to those of well-absorbed compounds such as propranolol (1.13×10 -5 cm/s), and testosterone (3.3×10 -5 cm/s) (24).The result from rat perfusion was consistent with in vitro study (25).The ratio (B-A/A-B) of bicyclol in Caco-2 cell was dramatically decreased in the presence of P-gp inhibitor LSN335984 or CsA, indicating that P-gp was involved in the transport of bicyclol in Caco-2 cell.Take together, the results suggested that bicyclol was a well-absorbed compound and P-gp mediated active efflux was likely to be one of reasons for its poor bioavailability.
In addition to P-gp active efflux, extensive metabolism in rat intestine may be another important mechanism responsible for poor bioavailability of bicyclol.For the purpose of observing the co-effect of CYP3A and P-gp in intestine, a suitable concentration of bicyclol was needed to apply to the inhibition study.The 8.2-fold increase in P blood of bicyclol with 2-fold increase of bicyclol in perfusate (50→100µM) indicated that P-gp was not effectively pumping the drug back into the intestinal lumen at the 100 µM concentration.In addition, a two-fold increase in the perfusion dose (50→100µM) resulted in a significant increase (8.7 fold) in the appearance of the cumulative amount of bicyclol in mesenteric blood, suggesting that there might be a saturation of intestinal efflux and metabolism of bicyclol at the highest dose.Therefore, a 50 µM concentration of bicyclol was chosen in the further studies.Cummins et al (21) reported that the absorptive permeability of substrate of CYP3A and P-gp in mesenteric vein was increased significantly by GG918 (inhibitor of P-gp), while the permeability of midazolam (substrate of CYP3A only) was unchanged under similar conditions.In addition, another earlier study by Johnson et al showed that the appearance of substrate of CYP3A and P-gp in the mesenteric blood was increased dramatically in the presence of TAO or CsA, compared with control group (18) .In our present study, P blood of bicyclol was increased by 3,12 or16-fold in the presence of LSN335984, TAO or CsA, while that of midazolam was almost unchanged with P-gp inhibitor.Moreover, the cumulative amount of bicyclol or midazolam in mesenteric blood showed the same trend as P blood .Therefore, our results are consistent with previous studies and suggest that bicyclol is a dual substrate of CYP3A and P-gp.Moreover, the significantly higher cumulative amount of bicyclol in blood (3-fold) after addition of TAO, compared with values after LSN335984, indicates that CYP3A related metabolism is more important than P-gp mediated efflux in intestine for the low bioavailability of bicyclol.
In summary, the poor bioavailability of bicyclol is believed to be due to P-gp mediated active efflux and extensive metabolism by CYP3A in rat intestine.Compared with P-gp, CYP3A related metabolism appears to be more important for the low bioavailability of bicyclol.

Figure 2 .
Figure 2.Cumulative amount of bicyclol appearing in mesenteric vein in the absence (control) or presence of LSN335984 (P-gp inhibitor), TAO (CYP3A inhibitor) or CsA (CYP3A and P-gp inhibitor) in rat jejunum perfusion study (n = 3-5).The concentration of bicyclol in the perfusate is 50 µM.Data are expressed as mean± SD.

Table 1 .
Effects of LSN33598 and CsA on the transepithelial transport of bicyclol.

Table 2 .
Permeabilities and cumulative amount of bicyclol in mesenteric vein across rat jejunum.

Table 3 .
Permeabilities and cumulative amount of midazolam in mesenteric vein across rat jejunum.