Membrane Transport of Andrographolide in Artificial Membrane and Rat Small Intestine

MATERIALS AND METHODS

Materials: Andrographolide was extracted from A. paniculata dried powder and identified by TLC, HPLC, proton NMR and mass spectrometer. All solvents were purchased from Merck (Darmstadt, Germany). Acetaminophen, chlorpheniramine maleate, norfloxacin, amoxycillin and probenecid were purchased from Sigma-Aldrich (Germany). Aspirin and cyclosporine were from Fluka (Switzerland) and Novatris (Thailand), respectively.

Transport study of andrographolide using the artificial membrane: A side-by-side diffusion cell (Crown Glass Company, USA) with a diffusion area of 0.936 cm² and each diffusion cell volume of 3 mL was used for the transport study of andrographolide across the artificial membrane. A cellulose nitrate membrane filter with 0.2 μm pore size (Sartorius, Germany) was immersed into octanol for one day prior to the transport study and used as the artificial lipid membrane. The filter-disc membrane was then mounted in the diffusion cell and equilibrated with either the simulated intestinal fluid (pH 7, 0.05 mole L^-1 phosphate buffer solution) or the simulated gastric fluid (pH 2, 0.01 mol L^-1 HCl, 2.5 g sodium lauryl sulphate, 2 g NaCl) for 30 min. In the experiments, the effects of different medium pHs (simulated pH 2 gastric and pH 7 intestinal fluids), different concentrations of andrographolide and co-administering drugs on membrane transport of andrographolide were examined. Andrographolide was dissolved at a concentration of 16.0, 68.0, 154.7 or 253.0 μg mL^-1 in simulated intestinal fluid or at a concentration of 13.3, 33.3, 66.7 or 128.0 μg mL^-1 in simulated gastric fluid. As commonly administered drugs together with andrographolide in usual pharmacotherapy, acetaminophen (12 mg mL^-1), chlorpheniramine maleate (0.016 mg mL^-1), amoxycillin (2 mg mL^-1), aspirin (1.2 mg mL^-1) and norfloxacin (1.6 mg mL^-1) were employed, where the concentration of co-administered drug was determined according to their oral dosages. Andrographolide solution (0.2 mg mL^-1, 1 mL), above coadministering drug solutions (1 mL) and simulated intestinal or gastric fluid (1 mL) were mixed in the donor cell. To receiver cell, 3 mL of simulated intestinal or gastric fluid was placed. Sampling (2.5 mL) was made from receiver cell periodically for 120 min and an equal volume of fresh simulated fluid was re-supplied each time.

Transport study of andrographolide using rat everted intestine: Male Sprague-Dawley rats (250-350 g) were fasted one night prior to the experiment. The protocol of animal study was reviewed and approved by the Animal Ethics Committee of Khon Kaen University based on the Ethic of Animal Experimentation of National Research Council of Thailand. Rats were anaesthetized with ether and killed by decapitation. Proximal and distal parts of small intestine were isolated and the luminal contents were flushed with ice-cold normal saline. The intestine was everted and two everted sacs of approximately 10 cm long were made from each proximal and distal rejoin of the intestine. One everted sac was used as a control and the other one was used for test compounds to avoid inter-individual differences in P-glycoprotein or MRP2 function. The everted sac was incubated in 8 mL of pH 7.4, Dulbecco’s phosphate buffer solution (D-PBS), which was saturated with a 95/5% O₂/CO₂ gas in advance at 37°C for 20 min. Andrographolide was dissolved in D-PBS at a concentration of 0.1 mmol L^-1 and the solution (0.5 mL) was introduced into the inside (serosal side) of the everted sac. Sampling (0.5 mL) of mucosal side medium was made periodically for 120 min and equal volume (0.5 mL) of fresh D-PBS was re-supplied each time.

The effect of commonly administered drugs on andrographolide efflux transport was also examined in everted rat intestine using acetaminophen (1 mM), chlorpheniramine maleate (1 mM), amoxycillin (1 mM), aspirin (1 mM) and norfloxacin (1 mM). In addition, the participation of ATP-dependent efflux transporters in membrane transport of andrographolide was examined by using cyclosporine as a P-glycoprotein and MRP2 inhibitor and probenecid as an MRP2 inhibitor in proximal and distal everted intestine. These compounds were added into the mucosal side of everted sac at a concentration of 1 mM and the efflux transport from serosal to mucosal surfaces of andrographolide was determined periodically.

Analytical method of andrographolide: To each aqueous sample (500 μL) containing andrographolide, 10 μL of naproxen sodium solution (100 μg mL^-1) as an internal standard and 490 μL of acetonitrile were added. The mixture solution was analysed by HPLC (A Hewlett Packard LC-1100 liquid chromatography instrument, equipped with an autosampler system) on a C18 reversed phase column (a Hypersil^® ODS column, 250x4.0 mm, 5 μm particle). Mobile phase used was a mixture of acetonitrile and 10 mM phosphate buffer solution (50:50 %v/v) adjusted to pH 2.0 with phosphoric acid. Andrographolide was monitored at a wavelength of 230 nm.

Data analysis: The transport rate of andrographolide across artificial membrane and rat everted intestine was calculated from the slope of the transported amount plotted against the time. The transport rate constant of andrographolide was estimated by dividing the transport rate by a product of the exposed area of the membrane (0.936 cm²) and the initial concentration of andrographolide in donor cell.

Results were expressed as the mean±the standard deviation. The analysis of variance (ANOVA) was applied to evaluate the significance of difference observed. p-values less than 0.05 were considered significant.

RESULTS AND DISCUSSION

Transport of andrographolide across the artificial membrane: The time course of andrographolide transport across the artificial membrane from pH 7 simulated intestinal fluids and pH 2 simulated gastric fluids are shown in Fig. 1 and 2, respectively. The transported amount of andrographolide increased with time in a zero-order fashion and the transport rate of andrographolide increased almost in parallel with increase in the initial concentrations in the simulated fluids as follows: 0.0940± 0.0017 μ g min^{- 1} at 16.0 μ g mL^{- 1} to 1.7570± 0.0058 μ g min^{- 1} at 253.0 μ g mL^{- 1} in simulated intestinal fluid and 0.0774± 0.0018 μ g min^{- 1} at 13.3 μ g mL^{- 1} to 1.0525± 0.1134 μ g min^{- 1} at 128.0 μ g mL^{- 1} in simulated gastric fluid. That is, the transport rate constants of andrographolide, estimated by dividing the transport rate by a product of initial concentration of andrographolide and exposure area of artificial membrane, were comparable between different pHs of medium (simulated gastric and intestinal fluids) and among different concentrations of andrographolide. Collectively, it was confirmed that the transport of andrographolide across artificial lipid membrane follows passive diffusion and the transport rate constant of andrographolide is constant among different pHs of medium, because andrographolide is an uncharged neutral compound.

Fig. 1:	Transport profiles of andrographolide from the simulated intestinal fluid across the artificial membrane. The concentration of andrographolide in donor cell: () 16.00 μ g mL- 1; () 68.00 μ g mL- 1; () 154.7 μ g mL- 1 and () 253.0 μ g mL- 1. Each value represents the mean± SD (n = 3)

Fig. 2:	Transport profiles of andrographolide from simulated gastric fluid across artificial membrane. The concentration of andrographolide in donor cell: () 13.33 μ g mL- 1; () 33.33 μ g mL- 1; () 66.67 μ g mL- 1 and () 128.0 μ g mL- 1. Each value represents the mean± SD (n = 3)

In clinical pharmacotherapy andrographolide, or extracts of A. paniculata, is commonly used in combination with other drugs such as acetaminophen, amoxycillin, aspirin, chlorpheniramine maleate and norfloxacin, as already described in the introduction. The possibility of drug interactions of andrographolide with above co-administering drugs was examined. In artificial lipophilic membrane, physicochemical drug interactions such as chelation, ion pair formation, micelle formation, hydrogen bonding and stacking in the aqueous medium or lipid layer could be detected, since the transport rate of andrographolide, or apparent lipophilicity, would change if above-mentioned drug interactions occurred. Results are summarized in Table 1. The transport rate and the transport rate constant of andrographolide were not affected by above co-administered drugs, suggesting that direct physicochemical interaction will not occur between andrographolide and commonly co-administered drugs such as acetaminophen, amoxycillin, aspirin, chlorpheniramine maleate and norfloxacin.

Table 1:	The transport rate constants of andrographolide across artificial lipid membrane in the absence (control) and presence of commonly co-administered drugs
Each value represents the mean± SD (n = 5)

Transport of andrographolide across rat everted intestine: The possibility of drug interactions of andrographolide with above co-administering drugs was then examined in rat intestine. In biomembranes, it is expected that pharmacokinetic drug interactions such as metabolism-mediated and transporter-mediated drug interactions could be detected in addition to physicochemical drug interactions. In the present study, the proximal and distal regions of small intestine were employed, since the expression and function of P-glycoprotein and MRP2 are known to show regional differences (Terao et al., 1996; Yumoto et al., 1999; Chan et al., 2004). In Fig. 3, the efflux transported amounts of andrographolide across everted intestine in the absence and presence of co-administered drugs are summarized. As shown in the Fig. 3, the values in the absence of co-administered drugs were scattered among different intestines (experiments), possibly due to the inter-individual variation of membrane permeability. However, examined co-administered drugs such as acetaminophen, amoxycillin, aspirin, chlorpheniramine maleate and norfloxacin showed no significant effect on andrographolide transport in each intestine. These results suggested that pharmacokinetic drug interaction, as well as physicochemical drug interaction, between andrographolide and above commonly co-administered drugs will not occur in the intestine in vivo. Acetaminophene and norfloxacin are reportedly absorbed mostly by passive diffusion (Goon and Klassen, 1990; Bermejo et al., 2004; Schaiquevich et al., 2004). Amoxycillin and aspirin are absorbed by both passive diffusion and active transport mediated by oligopeptide carrier l (Lennernas et al., 2002) and monocarboxylic transporter (Kattherine, 2002), respectively.

Fig. 3:

Transported amounts of andrographolide across everted proximal (a) and distal (b) intestine in the absence and presence of commonly co-administered drugs during a 120 min incubation. The concentration of andrographolide was 0.1 mM. Open column: absence; Closed column, presence of co-administered drugs: 1, acetaminophen (1 mM); 2, chlorpheniramine (1 mM); 3, amoxicillin (1 mM); 4, aspirin (1 mM); 5, norfloxacin (1 mM). Each value represents the mean± SD (n = 3)

In contrast, the intestinal transport mechanism of chlorpheniramine is obscure. For example, Kandimalla and Donovan (2005) recently reported that chlorpheniramine is a P-gp substrate, however, in other papers, modulators of P-glycoprotein and cytochrome P450 such as cyclosporin A and ketoconazole did not modify the intestinal absorption of this compound in vitro and in situ (Hiep et al., 2001).

Though chlorpheniramine showed no significant effect on andrographolide efflux transport in the present study, the participation of efflux transporters such as P-glycoprotein and MRP2 in andrographolide transport was examined using cyclosporine and probenecid as inhibitors. Andrographolide is a diterpene lactone compound and some diterpene compounds such as taxol (or paclitaxel) and taxotere (docetaxel) are known to be substrates of P-glycoprotein and/or MRP2 (Dantzig et al., 2001; Huisman et al., 2005; Vaclavikova et al., 2006).

Fig. 4:

The transport profiles of andrographolide across everted proximal intestine in the absence and presence of cyclosporine A (a) or probenecid (b). Andrographolide was added into the serosal-side medium of a 10 cm long everted intestine at a concentration of 0.1 mM (0.5 mL). Cyclosporin A or probenecid was added to mucosal-side medium at a concentration of 1 mM (8 mL). Each value represents the mean± SD (n = 3)

Fig. 5:

The transport profiles of andrographolide across everted distal intestine in the absence and presence of cyclosporine A (a) or probenecid (b). Andrographolide was added into the serosal-side medium of a 10 cm long everted intestine at a concentration of 0.1 mM (0.5 mL). Cyclosporin A or probenecid was added to mucosal-side medium at a concentration of 1 mM (8 mL). Each value represents the mean± SD (n = 3)

Also, some diterpene compounds are known to inhibit P-glycoprotein (Ferreira et al., 2005a, b). P-glycoprotein is highly expressed in the distal intestine, but not in the proximal region (Terao et al., 1996; Yumoto et al., 1999; Takano et al., 2006) and MRP2 is highly expressed in the proximal intestine, but not in distal region (Chan et al., 2004). In the proximal intestine, cyclosporine and probenecid suppressed the efflux transport of andrographolide a little, but not significantly (Fig. 4). In contrast, in the distal intestine, the efflux transport of andrographolide was significantly suppressed by cyclosporine A, though probenecid showed no significant effect (Fig. 5). Thus, it was suggested that P-glycoprotein, but not MRP, is participated in the intestinal absorption of andrographolide and P-glycoprotein-mediated drug interactions may occur depending on the co-administered drugs and its concentrations. Further in vivo studies including multiple treatments with commonly co-administered drugs would be necessary to reveal, predict and prevent the P-glycoprotein or MRP2-mediated drug interaction of andrographolide, since, for example, it is reported that acetaminophen significantly increased the expression of both MRP-2 and P-glycoprotein (Ghanam et al., 2004).

At present, it is well recognized that natural products such as grapefruit juice, several components of grapefruit juice, St. John’s wort, curcuminoids, green tea extract, salvinal and flavonoid constituents also modify efflux transporters functions (Chearwae et al., 2006; Ferreira et al., 2005a and b; Netsch et al., 2005; Panchagnula et al., 2005; Singh 2005; Wang et al., 2005; Chang et al., 2004; Honda et al., 2004). As well, the intake of Andrographis paniculate extracts, or andrographolide, may cause efflux transporter-mediated drug interaction by modulating their functions, as demonstrated in the present study. The analysis of transport property, or drug interaction, of andrographolide would be important information for pharmacotherapy with andrographolide or A. paniculata (He et al., 2003; Imanidis et al., 1995; Jonker, 2002; Legen et al., 2003).

CONCLUSIONS

In the present study, the possible drug interaction of andrographolide with commonly used drugs such as acetaminophen, amoxycillin, aspirin, chlorpheniramine and norfloxacin were examined using artificial lipophilic membrane and everted rat intestine. In former experimental system, direct physicochemical drug interactions and in latter experimental system, pharmacokinetic drug interactions would be detected. Above co-administered drugs examined showed no significant effect on andrographolide transport in both artificial lipophilic membrane and everted rat intestine. Then, the participation of ATP-dependent efflux transporters such as P-glycoprotein and MRP2 in andrographolide transport was examined in rat intestine in vivo. As a result andrographolide, a diterpene compound, was found to be related to P-glycoprotein and caused P-glycoprotein-mediated drug interaction. It will be necessary to pay attention in clinical pharmacotherapy with andrographolide from the viewpoint of P-glycoprotein-mediated drug interaction. In addition, it was suggested that the combined experimental system of absorption simulation using artificial lipophilic membrane and the transport study in everted intestine are useful in predicting some types of drug interaction that may occur in the intestine in vivo, though there was no significant drug interaction with commonly used drugs such as acetaminophen, amoxycillin, aspirin, chlorpheniramine and norfloxacin in the present study.

ACKNOWLEDGMENTS

The authors would like to thank the Graduate School of Khon Kaen University and the Central for Research and Development of Herbal Health Product, Khon Kaen University for their financial support.