Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The intestinal epithelium presents a barrier between the lumen of the intestine and the systemic circulation. This barrier has conventionally been considered to be of a physical nature in which most compounds are absorbed by a transcellular route, restricting drug absorption to compounds capable of partitioning across the apical membrane into the cytosol and out of the basolateral membrane. Recently, biochemical barriers have been implicated in the restriction of drug absorption. In particular, the drug efflux pump P-glycoprotein (Pgp) and the drug-metabolizing enzyme CYP3A4 have been shown to limit the intestinal absorption of some drugs (Watkins, 1997; Wacher et al., 1998).

Pgp is a drug efflux pump that has been shown to transport a large variety of structurally diverse compounds out of the cell interior (Gatmaitan and Arias, 1993; Schinkel, 1997). It is constitutively expressed on the apical membrane of intestinal epithelial cells (Cordon-Cardo et al., 1990), where it functions to secrete drugs from the cell interior back into the intestinal lumen. This active efflux runs countercurrent to the absorptive transport of drugs, thereby restricting the extent of oral absorption. Indeed, studies with Pgp-deficient mice and studies in which Pgp inhibitors were coadministered with drugs showed improved oral absorption of drugs that are Pgp substrates (Leu and Huang, 1995; Kim et al., 1998; Salphati and Benet, 1998a). Moreover, in human clinical studies, variability in cyclosporin absorption was strongly correlated to the level of Pgp expression (Fricker et al., 1996; Lown et al., 1997). Collectively, these studies strongly support a role of Pgp in the restriction of oral absorption of some drugs.

CYP3A4, the most prevalent cytochrome P-450 in the human intestine (Watkins et al., 1987; Paine et al., 1997), presents a metabolic barrier to drug absorption. Metabolism by intestinal CYP3A4 has been proposed to restrict the oral bioavailability of several drugs (Hebert et al., 1992; Paine et al., 1996; Lown et al., 1997). The coadministration of the potent CYP3A4 inhibitor ketoconazole led to increased bioavailability for a number of drugs in a manner consistent with inhibition of intestinal CYP3A4 (Gomez et al., 1995; Floren et al., 1997; Salphati and Benet, 1998a; Zhang et al., 1998). Similarly, increases in felodipine absorption when given with grapefruit juice were correlated with down-regulation of CYP3A4 in the small intestine, but not in the liver, indicating a significant role of intestinal CYP3A4 in the presystemic metabolism of felodipine (Lown et al., 1997b).

Although Pgp and CYP3A4 may play separate roles in restricting absorption, it has been proposed that the two proteins act synergistically to increase presystemic drug metabolism (Gan et al., 1996; Watkins, 1997; Wacher et al., 1998; Ito et al., 1999). According to these models, increased intestinal residence time resulting from Pgp efflux and prevention of product inhibition by the removal of primary metabolites from the cell interior result in more extensive intestinal metabolism by CYP3A4. Support for these models largely comes from circumstantial evidence citing overlapping substrate specificity for the two proteins (Wacher et al., 1995), similarities in the gene regulation and tissue distribution for CYP3A4 and Pgp (Wacher et al., 1995; Schuetz et al., 1996; Watkins, 1997; Salphati and Benet, 1998b), and the close intracellular spatial proximity of CYP3A4 to the apical membrane where Pgp is expressed (Watkins, 1997). However, direct evidence to support synergism between the two proteins has been elusive due to the lack of potent inhibitors specific to only CYP3A4 or Pgp that can be used in in vivo studies and the failure to identify an appropriate in vitro model expressing both CYP3A4 and Pgp.

The colon cancer cell line Caco-2 (Pinto et al., 1983) has been used extensively as a model to study drug transport across the intestinal epithelium and the role of Pgp in restricting drug absorption (for a review, see Artursson et al., 1996; Bailey et al., 1996; Hidalgo and Li, 1996). However, under conventional growth conditions, Caco-2 cells do not express significant levels of CYP3A4 (Prueksaritanont et al., 1996). Recently, Schmiedlin-Ren et al. (1997) reported that Caco-2 cells grown in the presence of 1,25-di-OH vitamin D₃ (di-OH vit D₃) expressed high levels of CYP3A4 activity. In this study, we applied this culture technique to study the effects of Pgp and CYP3A4 on indinavir metabolism during drug transport. The results presented demonstrate the use of this culture system for studying Pgp-CYP3A4 interactions and provide evidence to support a synergistic role of Pgp efflux in increasing the extent of intestinal metabolism of drugs that are substrates for both proteins.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Engelbreth-Holme-Swarm (EHS) extracellular matrix was purchased from Promega Corp. (Madison, WI). Testosterone, 6-OH-testosterone, di-OH vit D₃, and cyclosporin A (CsA) were purchased from Sigma Chemical Co. (St. Louis, MO). FBS, glutamine, trypsin-EDTA solution, penicillin-streptomycin solution, nonessential amino acids, Hanks' balanced salt solution (HBSS), and HEPES buffer were purchased from Life Technologies (Grand Island, NY). Dulbecco's modified Eagle's medium with pyruvate and 4.5 g/l glucose were prepared by Mediatech (Herndon, VA) and obtained from Fisher Scientific (Pittsburgh, PA). The di-OH vit D₃ stock solutions were prepared at 0.1 mg/ml in ethanol and stored at 70°C. Indinavir, [¹⁴C]indinavir, and indinavir metabolite standards were prepared at Merck and Co. (West Point, PA, and Rahway, NJ).

Cell Culture. Caco-2 cells were obtained from American Type Culture Collection (Rockville, MD) and were used at passage 21 to 29. The cells were routinely maintained in Dulbecco's modified Eagle's medium supplemented with glutamine, nonessential amino acids, penicillin-streptomycin, and 5% FCS at 37°C in a humidified 5% CO₂/95% air environment. The di-OH vit D₃ induction of CYP3A4 in mixed Caco-2 cell cultures was performed as described by Schmiedlin-Ren et al. (1997). For comparisons between di-OH vit D₃-treated and untreated Caco-2 cell cultures, Caco-2 cells were plated at 2 × 10⁵ cells/cm² onto culture flasks or 24-mm-diameter polycarbonate filters (Costar Transwell, 0.2 µm pore size; Corning Glassworks, Corning, NY) coated with 2 µg/cm² EHS cell attachment matrix (extracellular matrix from EHS cells). After 5 days in culture, di-OH vit D₃ treatment was initiated by adding medium containing 0.1 µg/ml di-OH vit D₃. The di-OH vit D₃ was added to the medium immediately before feeding the cells, and the culture medium was changed every 2 to 3 days. The di-OH vit D₃ treatment was continued for 2 to 3 weeks before the cells were used for transport and metabolism studies.

Kinetics of M6 Formation. The di-OH vit D₃-treated and untreated Caco-2 cells grown in flasks were dissociated with trypsin-EDTA, resuspended in Dulbecco's modified Eagle's medium with 5% FCS, and washed by centrifugation (500g for 5 min), followed by washing two more times in HEPES-buffered HBSS, pH 7.4. The final cell pellets were resuspended at 5 million cells/ml in HEPES-buffered HBSS, and 0.5 ml/well was added in a 24-well tissue culture plate. Indinavir (10 mM stock solution in water) was then added to the cell suspension at final concentrations of 0.2 to 50 µM, and the mixtures were incubated in a tissue culture incubator at 37°C. After 4 h, viability of the cells was confirmed by trypan blue exclusion, 0.4 ml of the cell suspensions were transferred to glass tubes, and two volumes of acetonitrile were added to terminate metabolism and precipitate the cellular protein. An internal standard structurally related to indinavir was added to the mixture, precipitated protein was removed by centrifugation, and the samples were dried and reconstituted in 10% acetonitrile in water.

Transport and Metabolism of Indinavir. Before transport studies, culture medium was removed, and the filter-grown Caco-2 cells were preequilibrated in HEPES-buffered HBSS, pH 7.4. The HBSS was then replaced with 2 ml of fresh HBSS on the receiver side and 2 ml of HBSS containing indinavir on the donor side. In cases in which Pgp was inhibited with CsA, both the receiver and donor solutions also contained 5 µM CsA.

LC/MS Analysis of Indinavir and Its Metabolites. Indinavir and its metabolites were separated on a 5-µm Betasil C-18 reversed phase column (50 × 3 mm) (Keystone analytical) using a linear gradient from 30 to 80% solvent B over the first 2 min, holding at 80% B for 0.5 min, and returning to 30% B over 1 min, in which solvent A is 5 mM ammonium acetate, pH 4.5, and solvent B is 70% acetonitrile. Indinavir and its metabolites were detected by LC/MS with a Sciex API 150 using APCI at capillary temperature of 425. Ions were monitored at m/z of 523, 529, 614, 630, and 613 for M6, M5, indinavir, addition of oxygen to indinavir, and the internal standard, respectively.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of CYP3A4 and Indinavir Metabolism in Caco-2 Cells. We applied the Caco-2 cell model to study the influence of Pgp efflux on the CYP3A4-mediated metabolism of indinavir. When grown under conventional culture conditions, Caco-2 cells express many of the barrier properties of the intestinal epithelium, including expression of the drug efflux pump Pgp, but fail to express significant CYP3A4 activity. CYP3A activity as determined by 6-OH-testosterone production was low in the Caco-2 filters grown in the absence of di-OH vit D₃ and was below the limits of detection in most cases (Table 1). To induce expression of CYP3A4, Caco-2 cells were grown on the extracellular matrix from EHS cells (EHS matrix) in the presence of di-OH vit D₃. Caco-2 cell filters treated with di-OH vit D₃ for 2 weeks showed a dramatic increase in 6-OH-testosterone production that was inhibited by the CYP3A4 inhibitor 1 µM ketoconazole. Over different passages of cells (passages 21-28), the level of CYP3A activity in Caco-2 cells treated for 2 to 3 weeks varied as much as 4-fold with an average value of 17.4 ± 11.3 pmol/min 6-OH-testosterone produced per 4.7-cm² filter on incubation with 50 µM testosterone.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   LC/MS chromatograms of metabolites formed from indinavir incubated with di-OH vit D3-treated (solid line) and untreated (dotted line) Caco-2 cell suspensions. Chromatograms represent total ion current and extracted ions for m/z = 523.3, 539.3, 613.3, 614.3, and 630.3, corresponding to M6, M5, the internal standard, indinavir, and hydroxylation of indinavir.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Vectorial transport of indinavir in Caco-2 cells. A, transport of indinavir in the apical-to-basolateral direction in the absence () and presence () of 10 µM CsA and in the basolateral-to-apical direction in the absence () and presence () of CsA. B, comparison of indinavir transport across di-OH vit D3-treated (solid bars) and untreated (hatched bars) Caco-2 cell monolayers in the apical-to-basolateral (A to B) and the basolateral-to-apical (B to A) direction in the presence and absence of CsA (n = 3 for each determination).

Pgp Activity in Caco-2 Cells. Directional transport of indinavir (Fig. 2A) in di-OH vit D₃-treated Caco-2 cells showed 5- to 20-fold higher transport in the basolateral-to-apical direction than in the apical-to-basolateral direction. In the presence of the potent Pgp inhibitor CsA (5 µM), basolateral-to-apical transport of indinavir decreased and apical-to-basolateral transport increased to the point where no significant directional transport was observed. Comparison of untreated Caco-2 cells and Caco-2 cells treated with di-OH vit D₃ showed no significant difference between directional transport of indinavir, indicating that di-OH vit D₃ treatment neither induced nor suppressed Pgp activity (Fig. 2B).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Distribution of M6 in the apical (solid bar), basolateral (hatched bar), and intracellular (open bar) compartments showing apical secretion of M6. Indinavir (5 µM) was applied to the apical (A to B) or the basolateral (B to A) compartment in the presence or absence of CsA (5 µM; n = 3 for each condition except n = 2 for A to B + CsA).

Metabolism and Transport of Indinavir in di-OH vit D₃-Treated Caco-2 Cells. Directional transport studies were performed to evaluate the influence of Pgp efflux on metabolism of indinavir. Indinavir transport was determined after dosing to the apical or the basolateral side in the presence or absence of the Pgp inhibitor CsA. M6 retained in the cells and released into the apical and basolateral media were quantified by LC/MS. In the absence of CsA, M6 was almost exclusively secreted into the apical compartment (Fig. 3). Inhibition of Pgp by CsA resulted in M6 being released into both the basolateral and the apical compartments and some M6 being retained in the cells, suggesting that M6 is actively effluxed by Pgp. Similarly, M5 and OH-indinavir also showed selective efflux into the apical compartment (Fig. 4). In contrast to the apical secretion of indinavir metabolites, 6-OH-testosterone did not show apical secretion (Table 1), suggesting that Pgp efflux cannot be generalized to all CYP3A-generated metabolites.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Extracted ion chromatograms of the apical (solid line) and the basolateral (dotted line) solutions showing secretion of M6 (m/z = 523.3), M5 (m/z = 539.3), and OH- indinavir (m/z = 630) into the apical compartment 4 h after the addition of 1.5 µM indinavir (m/z = 614.3) to the basolateral compartment.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Formation of M6 relative to indinavir transport in di-OH vit D3-treated Caco-2 cells. A, total M6 formation during indinavir transport across Caco-2 cells. B, cumulative transport of indinavir into the receiver compartment after 4 h. C, ratio of M6 formation to indinavir transport (n = 3 for each condition except n = 2 for A to B + CsA).

Concentration Dependence of M6 Formation during Indinavir Transport. Indinavir metabolism and drug transport were evaluated as a function of indinavir concentration in the donor compartment. Results for metabolite formation relative to drug transport in the apical-to-basolateral direction and the basolateral-to-apical direction are shown in Fig. 6. The kinetic parameters for M6 formation were determined from a nonlinear fit to Michaelis-Menton kinetics and confirmed by Scatchard analysis yielding a K_m value of 8.1 µM and a V_max value of 1.54 × 10⁴ nmol · min¹ · mg protein¹ for Caco-2 cell filters at 0.16 mg cellular protein · cm². The kinetics for Pgp efflux determined from the difference between apical-to-basolateral and basolateral-to-apical transport yielded a K_m value of 140 µM and a V_max value of 2.01 × 10¹ nmol · min¹ · mg protein¹, and the intrinsic permeability coefficient (Papp_i) reflecting the passive permeability of indinavir across Caco-2 cells was determined to be 3.87 × 10⁴ cm¹ · min¹. Using these values, it is possible to simulate the effects of extracellular indinavir concentration (C_o) and transport direction on the ratio of metabolite to drug transport assuming that the kinetics of metabolism relative to the intracellular drug concentration is independent of Pgp activity as illustrated in Fig. 7. Under steady-state transport conditions, the rate of drug entering the cell equals the rate of drug being removed from the cell interior via metabolism (V_met), active efflux (V_Pgp), and passive transport. Thus:

where P_ap and P_bl are the passive permeability coefficients of the apical and basolateral membranes, respectively; C_o is the drug concentration in the donor compartment; and C_i is the drug concentration inside the cell. If we assume that the permeability coefficients for the apical and basolateral membranes are equal (P_ap = P_bl = P_{ap, bl}), then the intracellular concentration (C_i) can be estimated:
The permeability across the apical (P_ap) and basolateral membranes (P_bl) are related to the intrinsic permeability (Papp_i):
Assuming that P_ap and P_bl are equal, then P_{ap, bl} = 2 Papp_i = 7.74 × 10⁴ · cm¹ · min¹. The concentration inside the cell (C_i) can then be determined as a function of extracellular indinavir concentration using the K_m (relative to extracellular indinavir) and V_max values to determine the rate of Pgp efflux (V_Pgp) and metabolism (V_met). Transport of drug into the receiver compartment is equal to P_bl C_i for transport in the apical-to-basolateral direction and P_ap C_i + V_Pgp for transport in the basolateral-to-apical direction. For the simulated curves in Fig. 6, the rate of metabolism as a function of C_i was derived using the V_max value determined for indinavir metabolism during the transport experiment and a K_m value (relative to intracellular drug concentration) of 2 µM (Chiba et al., 1997). The simulated curves for the ratio of metabolite formation to drug transport are indicated by the solid and broken lines in Fig. 6.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Ratio of M6 formation to indinavir transport in the apical-to-basolateral direction () and the basolateral-to-apical direction () at indinavir concentrations from 0.3 to 82 µM. The solid and dashed lines represent simulated curves for the ratio of M6 formation to indinavir transport in the apical-to-basolateral direction and the basolateral-to-apical direction as a function of indinavir concentration. The simulated curves were derived as described in the text using a value for the Km for indinavir metabolism relative to the intracellular drug concentration of 2 µM (value observed with human intestinal microsomes) and values for Vmax, passive membrane permeability, and Pgp efflux obtained with the Caco-2 cell filters (n = 3 for each determination).

View larger version (11K): [in this window] [in a new window]   Fig. 7.   Model depicting the role of Pgp, CYP3A4, and passive membrane transport on drug absorption. The model depicts transport from the apical-to-basolateral direction under sink conditions. Co and Ci represent the concentration of indinavir in the donor solution and inside the cell, and Pap and Pbl correspond to the rate of passive drug transport across the apical and basolateral membrane, respectively. Vmet represents the rate of indinavir metabolism.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

It has been proposed that Pgp acts in synergy with CYP3A4 in the intestine to restrict oral absorption of drugs. According to this model, Pgp can enhance the extent of intestinal CYP3A4 metabolism by both 1) removal of CYP3A4-generated metabolites from inside the cells, thus preventing further interaction of metabolites with CYP3A4, and 2) prolonging the time required for absorption through repeated cycles of absorption and efflux, thus increasing the exposure of a drug to CYP3A4 before absorption into the systemic circulation (Gan et al., 1996; Watkins, 1997; Wacher et al.; 1998; Ito et al., 1999). However, no direct evidence to support this model has been obtained. This is in part due to the difficulty in separating the individual contributions of CYP3A4 and Pgp in vivo.

An in vitro model capable of emulating the transport properties of the intestinal epithelium that expresses both Pgp and CYP3A4 would provide a useful tool for assessment of the contribution of both activities. Although Caco-2 cells have been widely used as a model for intestinal absorption and Pgp transport, the expression of CYP3A4 activity is generally too low to study metabolite formation as a function of drug transport. Recently, Schmiedlin-Ren et al. (1997) demonstrated that the maintenance of Caco-2 cells in culture media containing di-OH vit D₃ induces expression of significant levels of CYP3A activity. The expression of CYP3A activity is primarily ascribed to increased CYP3A4 expression, although smaller amounts of CYP3A5 and 3A7 were also induced. Consistent with their results, we saw significant oxidation of testosterone to 6--OH-testosterone in di-OH vit D₃-treated Caco-2 cells, whereas this activity was barely detectable in untreated cells. Pgp activity in di-OH vit D₃-treated Caco-2 cells was comparable to activity observed in untreated cells. Thus, this system provides an intestinal model in which both CYP3A4 and Pgp activities are expressed and can be easily manipulated under controlled conditions to assess their roles and interactions in intestinal metabolism.

In the studies detailed here, we evaluated the interactions between intestinal CYP3A4 and Pgp using the human immunodeficiency virus protease inhibitor indinavir as a model substrate. Indinavir is a useful model compound because it is a substrate for both Pgp (Kim et al., 1998; Lee et al., 1998; Washington et al., 1998) and CYP3A4 (Balani et al., 1996; Chiba et al., 1996, 1997; Lin et al., 1996). Oxidative metabolism of indinavir by CYP3A enzymes occurs with high affinity in liver microsomes and generates six metabolites. N-Depyridomethylation (M6 formation) is the most prevalent route of metabolism, with N-oxide formation and oxidation of the phenylmethyl and indan moieties occurring in smaller amounts. The profile for indinavir metabolism in di-OH vit D₃-treated Caco-2 cells was consistent with the pattern for CYP3A4-mediated oxidative metabolism, with M6 being the most prominent metabolite formed and secondary metabolism of M6 and oxidation of indinavir occurring at lower amounts.

It has been proposed that one potential mechanism by which Pgp can enhance intestinal metabolism of drugs is by facilitating removal of the metabolites, thus preventing product inhibition due to competition for CYP3A4 (Watkins, 1997). The distribution of M6 formed during indinavir transport is consistent with this role. In the absence of a Pgp inhibitor, M6 was almost exclusively secreted into the apical compartment, and intracellular M6 was not detected. In the presence of the Pgp inhibitor CsA, both basolateral and intracellular M6 was detected, suggesting that the removal of M6 from the intracellular compartment is in part mediated by Pgp. This vectoral efflux of CYP3A-generated metabolites is consistent with findings previously reported for metabolites of CsA (Gan et al., 1996) and midazolam (Schmiedlin-Ren et al., 1997). However, Pgp-mediated efflux cannot be generalized to all CYP3A metabolites because 6-OH testosterone did not show selective efflux into the apical compartment and its distribution was not affected by CsA. It is also worth noting that although the results are consistent with a role of Pgp in the removal of M6 from inside the cell, accumulation of M6 within the cells when Pgp was inhibited did not affect the extent of indinavir metabolism in our study.

The second mechanism by which Pgp can increase the extent of metabolism by CYP3A4 is by decreasing the rate of absorption of a drug through repeated cycles of intracellular uptake and efflux, thereby increasing the exposure of a drug to CYP3A4 before absorption in the systemic circulation. The results presented here clearly support this view. Although the total metabolite formed during the transport studies was not significantly affected by Pgp transport, resistance to transport conferred by Pgp efflux increased the amount of metabolite formed when normalized by the amount of parent drug transported across the monolayer. In extrapolation of these findings to in vivo absorption, Pgp efflux would have the effect of decreasing the rate of absorption, thus increasing the residence time resulting in increased metabolism. A model to explain these results is illustrated in Fig. 7. The net effect of Pgp efflux is to decrease the intracellular concentration of drug. Assuming the drug is transported across the basolateral membrane via passive diffusion, the observed drug transport across the epithelial layer will be a linear function of the intracellular drug concentration. Thus, a 50% decrease in the intracellular drug concentration due to Pgp efflux will result in a 50% decrease in the rate of drug transport. However, metabolism by CYP3A4 is not a linear function of the intracellular concentration, but it is saturable. Consequently, changes in intracellular concentration are not proportionally reflected in the rate of drug metabolism.

One limitation to this model is the assumption that the intracellular concentration of drug is uniform throughout the cells. If a drug is subject to extensive protein and membrane binding, it is possible that the distribution of drug within the cell may form a gradient from high concentration (the donor side) to low concentration (receiver side). Because cytochrome P-450s in enterocytes are localized toward the apical portion of the cell (Watkins, 1997), heterogeneity in the drug concentration within the cell could affect the relative influence of drug concentration on metabolism and transport. This could explain the small discrepancy between the results we obtained with indinavir and the results reported previously for CsA (Gan et al., 1996). Although both indinavir and CsA show a higher ratio of metabolite formation to drug transport for apical-to-basolateral transport than for basolateral-to-apical transport, the rate of CsA metabolism was also affected by the direction of transport. As opposed to the results we obtained showing essentially no effect of the direction of transport on the rate of M6 formation, Gan et al. (1996) found 2-fold higher metabolism of CsA when CsA was transported in the apical-to-basolateral direction than in the basolateral-to-apical direction. If we assume that cytochrome P-450s enzymes are localized in the apical portion of Caco-2 cells similar to their distribution in normal intestinal cells, then a gradient of CsA concentration could result in a lower concentration of CsA at the site of metabolism when the drug is administered to the basolateral side than when it is administered to the apical side. This could then be reflected as decreased metabolism observed for cyclosporin when administered to the basolateral compartment as opposed to the apical compartment.

In summary, the results presented here demonstrate the use of di-OH vit D₃-treated Caco-2 cells for study of the interactions between CYP3A4 and support a role of Pgp activity in enhancement of CYP3A4 intestinal metabolism.