Prediction of the oral absorption of low-permeability drugs using small intestine-like 2/4/A1 cell monolayers

Evaluation and suggested improvements of the Biopharmaceutics Classification System (BCS)

This review has evaluated the Biopharmaceutics Classification System (BCS) and improvements have been proposed. The BCS has a very strict solubility/dissolution limit, a generous Pe-limit (≥ 14-times higher rate constant limit for dissolution than for permeation), and is stricter for drugs with a long half-life (t1/2). Available human in-vivo, in-vitro, and in-silico Pe-methods cannot classify Pe for moderately to highly permeable substances sufficiently well, and in-vitro data often underpredict the in-vivo dissolution potential and rate. Good in-vivo dissolution and absorption can be expected for most high Pe drug products. It has not been possible to find a highly permeable product with a Dose number (Do) < 385 (< 2400 in the fed state) that is clearly incompletely absorbed, and near complete uptake has been shown for a drug product with a Do of 660000. The potential implication of these findings is that many true BCS Class I drug products are incorrectly classified. This could be a reason for the limited use of this system. On this basis, it has been suggested that: the limit for high for solubility/dissolution is decreased (to >40 and >95% dissolved within 30min and 3 h, respectively); the limit for high Pe is increased (to >Pe of metoprolol); accurate Pe-models or in-vivo fraction absorbed data are used; solubility/dissolution tests are performed using real or validated simulated gastrointestinal fluids; in-vitro/in-vivo dissolution relationships are established; the t½ is considered; and the rate-limiting step for in-vivo absorption is determined. A major change could be to reduce the BCS into two classes: permeation-rate (Class I) or dissolution-rate (Class II) limited absorption. It is believed that this could give a better balance and increase the number of biowaivers.

Ionization-specific analysis of human intestinal absorption

This study presents a mechanistic QSAR analysis of human intestinal absorption of drugs and drug-like compounds using a data set of 567 %HIA values. Experimental data represent passive diffusion across intestinal membranes, and are considered to be reasonably free of carrier-mediated transport or other unwanted effects. A nonlinear model was developed relating %HIA to physicochemical properties of drugs (lipophilicity, ionization, hydrogen bonding, and molecular size). The model describes ion-specific intestinal permeability of drugs by both transcellular and paracellular routes, and also accounts for unstirred water layer effects. The obtained model was validated on two external data sets consisting of in vivo human jejunal permeability coefficients (P(eff)) and absorption rate constants (K(a)). Validation results demonstrate good predictive power of the model (RMSE = 0.35-0.45 log units for log K(a) and log P(eff)). High prediction accuracy together with clear physicochemical interpretation (log P, pK(a)) makes this model particularly suitable for use in property-based drug design.

Claudin-2-dependent changes in noncharged solute flux are mediated by the extracellular domains and require attachment to the PDZ-scaffold

Paracellular transport through the tight junction shows selectivity for both ionic charge and solute size. It is known that charged residues on the extracellular loops of claudins control charge selectivity. It is also known that inducible expression of claudin-2, but not claudin-4, will selectively increase the permeability for polyethylene glycol (PEG) molecules which are <0.4 A in radius, but it is not known whether permeability is controlled by the same regions of claudins which control charge selectivity. Using inducible expression of chimeras of claudin-2 and claudin-4 in monolayers of MDCK II cells we show that the extracellular loops alone are responsible for controlling the permeability for noncharged PEGs as well as for charge selectivity. Further, the cytoplasmic C-terminal PDZ-binding motif is required for wild-type claudin-2 to control permeability, suggesting a requirement for attachment to the PDZ scaffold in order to form pores. These observations support a model where the loops form pores controlling permeability for both charged and noncharged solutes which are smaller than 0.4 A. They leave unanswered why both claudin-2 and -4 can influence electrical properties while only -2 can selectively increase permeability for small PEGs.

Passive oral drug absorption can be predicted more reliably by experimental than computational models--fact or myth

This study assessed the prediction power of experimental and computational models that are widely used to predict human passive intestinal absorption. The models evaluated included two cell lines, three artificial membrane models, in vivo rat experiments, and seven previously described computational quantitative structure property relationship models based on human absorption values. The data sets used in the assessment of the models were carefully chosen from the literature, and different models were compared using the same compounds to ensure objective results. Three of the computational models were found to be significantly more reliable in predicting human passive intestinal absorption than the artificial membrane models that have been developed for the prediction of passive intestinal absorption. Two of these computational models were found to be as reliable as the Caco-2 and the 2/4/A1 cell lines and, furthermore, one of the models was able to predict the absorption of a set of 65 drugs nearly as well as absorption studies in rats. The unexpectedly good prediction power of the simple computational models with high throughput makes them ideal tools in the early screening of drug candidates, whereas laborious cell culture models and animal studies can be useful in the later phases when detailed information about the transport mechanisms is needed.

The density of small tight junction pores varies among cell types and is increased by expression of claudin-2

Epithelial tight junctions contain size- and charge-selective pores that control the paracellular movement of charged and noncharged solutes. Claudins influence the charge selectivity and electrical resistance of junctions, but there is no direct evidence describing pore composition or whether pore size or density differs among cell types. To characterize paracellular pores independent of influences from charge selectivity, we profiled the ;apparent permeabilities' (P(app)) of a continuous series of noncharged polyethylene glycols (PEGs) across monolayers of five different epithelial cell lines and porcine ileum. We also characterized P(app) of high and low electrical resistance MDCK cell monolayers expressing heterologous claudins. P(app) profiling confirms that the paracellular barrier to noncharged solutes can be modeled as two distinct pathways: high-capacity small pores and a size-independent pathway allowing flux of larger solutes. All cell lines and ileum share a pore aperture of radius 4 A. Using P(app) of a PEG of radius 3.5 A to report the relative pore number provides the novel insight that pore density along the junction varies among cell types and is not necessarily related to electrical resistance. Expression of claudin-2 results in a selective increase in pore number but not size and has no effect on the permeability of PEGs that are larger than the pores; however, neither knockdown of claudin-2 nor overexpression of several other claudins altered either the number of small pores or their size. We speculate that permeability of all small solutes is proportional to pore number but that small electrolytes are subject to further selectivity by the profile of claudins expressed, explaining the dissociation between the P(app) for noncharged solutes and electrical resistance. Although claudins are likely to be components of the small pores, other factors might regulate pore number.

Prediction of human pharmacokinetics--gastrointestinal absorption

Permeability (P(e)) and solubility/dissolution are two major determinants of gastrointestinal (GI) drug absorption. Good prediction of these is crucial for predicting doses, exposures and potential interactions, and for selecting appropriate candidate drugs. The main objective was to evaluate screening methods for prediction of GI P(e), solubility/dissolution and fraction absorbed (f(a)) in humans. The most accurate P(e) models for prediction of f(a) of passively transported and highly soluble compounds appear to be the 2/4/A1 rat small intestinal cell model (in-vitro and in-silico), a newly developed artificial-membrane method, and a semi-empirical approach based on in-vitro membrane affinity to immobilized lipid bilayers, effective molecular weight and physiological GI variables. The predictability of in-vitro Caco-2, in-situ perfusion and other artificial membrane methods seems comparably low. The P(e) and f(a) in humans for compounds that undergo mainly active transport were predicted poorly by all models investigated. However, the rat in-situ perfusion model appears useful for prediction of active uptake potential (complete active uptake is generally well predicted), and Caco-2 cells are useful for studying bidirectional active transport, respectively. Human intestinal in-vitro P(e), which correlates well with f(a) for passively transported compounds, could possibly also have potential to improve/enable predictions of f(a) for actively transported substances. Molecular descriptor data could give an indication of the passive absorption potential. The 'maximum absorbable dose' and 'dose number' approaches, and solubility/dissolution data obtained in aqueous media, appear to underestimate in-vivo dissolution to a considerable extent. Predictions of in-vivo dissolution should preferably be done from in-vitro dissolution data obtained using either real or validated simulated GI fluids.

Modulation of tight junctions does not predict oral absorption of hydrophilic compounds: Use of caco-2 and calu-3 cells

Permeability estimates using Caco-2 cells do not accurately predict the absorption of hydrophilic drugs that are primarily absorbed via the paracellular pathway. The objective of this study was to investigate whether modulation of tight junctions would help differentiation of paracellularly absorbed compounds. Tight junctions in Caco-2 cell monolayers were manipulated using calcium depletion approaches to decrease the transepithelial electrical resistance (TEER) of the monolayers, and permeability of hydrophilic compounds were measured under these conditions. Permeability of these compounds were also measured in Calu-3 cells, which have tighter junctions than Caco-2 cells. Calcium depletion loosened the tight junctions of Caco-2 cells to varying levels as measured by the decrease in TEER values of the monolayers. While the absolute permeability of all the model compounds increased as the tight junctions were loosened, the ratios of their permeability relative to mannitol permeability were similar. The permeability of these compounds in the tighter Calu-3 cells were also found to be similar to each other. Altering the tight junctions of Caco-2 cells to obtain leakier cell monolayers, or using a cell line with tighter junctions like Calu-3 cells, did not improve differentiation between well absorbed and poorly absorbed hydrophilic drugs. Mere manipulation of the tight junctions to increase or decrease transepithelial electrical resistance does not appear to be a viable approach to predict human absorption for hydrophilic compounds that are primarily absorbed via the paracellular pathway.

Predicting Human Drug Pharmacokinetics from In Vitro Permeability Using an Absorption–Disposition Model

The purpose of this research is to simulate the in vivo performance of drugs with a wide range of solubility and permeability characteristics formulated as oral dosage forms. The absorption-disposition model was developed using a number of physiological parameters as well as in vitro permeability data generated with Caco-2 cells, 2/4/A1 cells, and hexadecane membranes. A total of 13 drugs with varying solubility and permeability properties were examined using the absorption-disposition model to predict their pharmacokinetic profile. The correlation of predicted and experimentally determined AUC and Cmax, as measures of the pharmacokinetic profile, were >0.96 for all permeation techniques examined. The predictive ability of the model is influenced by the type of permeation method employed; 2/4/A1 cell data yielded the highest degree of accuracy in predicting Cmax and AUC values. The absorption-disposition model developed in this work accurately predicts the in vivo performance of a wide range of orally administered drugs with 8 of 9 drugs examined falling within 80-125% of the experimental value of AUC when using 2/4/A1 cells.

A Cheminformatic Toolkit for Mining Biomedical Knowledge

PURPOSE

Cheminformatics can be broadly defined to encompass any activity related to the application of information technology to the study of properties, effects and uses of chemical agents. One of the most important current challenges in cheminformatics is to allow researchers to search databases of biomedical knowledge, using chemical structures as input.

MATERIALS AND METHODS

An important step towards this goal was the establishment of PubChem, an open, centralized database of small molecules accessible through the World Wide Web. While PubChem is primarily intended to serve as a repository for high throughput screening data from federally-funded screening centers and academic research laboratories, the major impact of PubChem could also reside in its ability to serve as a chemical gateway to biomedical databases such as PubMed.

CONCLUSION

This article will review cheminformatic tools that can be applied to facilitate annotation of PubChem through links to the scientific literature; to integrate PubChem with transcriptomic, proteomic, and metabolomic datasets; to incorporate results of numerical simulations of physiological systems into PubChem annotation; and ultimately, to translate data of chemical genomics screening efforts into information that will benefit biomedical researchers and physician scientists across all therapeutic areas.

In vitro behavior of a phosphate ester prodrug of amprenavir in human intestinal fluids and in the Caco-2 system: illustration of intraluminal supersaturation

As a result of their improved aqueous solubility, the development of phosphate ester prodrugs is an interesting approach to increase intestinal absorption of poorly water-soluble drugs. Absorption of a drug from its phosphate ester prodrug is based on intestinal dephosphorylation of the prodrug which may result in intraluminal supersaturation of the parent drug, followed by an increased absorptive flux across the intestinal mucosa. In this study, we evaluated the behavior of fosamprenavir, a phosphate ester prodrug of amprenavir, in the Caco-2 system and in aspirated human intestinal fluids (HIF), both showing phosphatase activity. Starting from a solution of fosamprenavir in HIF, a supersaturated solution of amprenavir was generated and maintained during a time period sufficient for absorption. Moreover, supersaturation of amprenavir resulted in an enhanced flux across Caco-2 monolayers. To our knowledge, this is the first illustration of supersaturation in real intestinal media. Next, we showed an inhibitory effect of inorganic phosphate on the dephosphorylation of fosamprenavir, both in the Caco-2 model and in HIF. As a consequence, phosphate-buffered media, including fasted state simulated intestinal fluid (FaSSIF), are incompatible with the study of phosphate ester prodrugs and should be replaced with media containing a biorelevant phosphate concentration (0.4-1 mM) and another buffering compound such as 2-morpholinoethanesulfonic acid (MES).

High-throughputinvitroprofiling assays: lessons learnt from experiences at Novartis

This article reviews the use of a selection of in vitro assays implemented at Novartis and intends to address exposure and safety in early drug discovery. The authors' own experience, based on a large number of 'real' drug discovery compounds, is described to reflect on what has worked, where improvement is needed and how to best use the data for decision making. Possible strategies are discussed, and guidelines are provided on how to organise assays, extract value and contribute knowledge from the data.

Computational models to predict aqueous drug solubility, permeability and intestinal absorption

In the last decade, poor intestinal absorption of candidate drugs intended for oral administration has been identified as a major bottleneck in drug development. Poor intestinal absorption can often be related to poor aqueous solubility and/or poor permeability across the intestinal wall. Other factors, such as poor stability and the metabolism of the compounds, can also decrease the amount of compound absorbed. In an effort to design compounds with enhanced absorption profile, theoretical predictions of solubility and permeability, among other factors, have gained increased interest, and a large number of papers have been published. In this review, the databases and techniques used for the development of in silico absorption models will be discussed. The focus is on aqueous drug solubility, which has become a major problem in drug development.

Voltage Dependence of Transepithelial Guanidine Permeation Across Caco-2 Epithelia Allows Determination of the Paracellular Flux Component

PURPOSE

The aim of this study was to investigate transepithelial ionic permeation via the paracellular pathway of human Caco-2 epithelial monolayers and its contribution to absorption of the base guanidine.

METHODS

Confluent monolayers of Caco-2 epithelial cells were mounted in Ussing chambers and the transepithelial conductance and electrical potential difference (p.d.) determined after NaCl dilution or medium Na substitution (bi-ionic conditions). Guanidine absorption (Ja-b) was measured +/- transepithelial potential gradients using bi-ionic p.d.'s.

RESULTS

Basal NaCl replacement with mannitol gives a transepithelial dilution p.d. of 28.0 +/- 3.1 mV basal solution electropositive (PCl/PNa = 0.34). Bi-ionic p.d.'s (basal replacements) indicate a cation selectivity of NH4+ > K+ approximately Cs+ > Na+ > Li+ > tetraethylammonium+ > N-methyl-D: -glucamine+ approximately choline+. Transepithelial conductances show good correspondence with bi-ionic potential data. Guanidine Ja-b was markedly sensitive to imposed transepithelial potential difference. The ratio of guanidine to mannitol permeability (measured simultaneously) increased from 3.6 in the absence of an imposed p.d. to 13.8 (basolateral negative p.d.).

CONCLUSIONS

Hydrated monovalent ions preferentially permeate the paracellular pathway (Eisenman sequence 2 or 3). Guanidine may access the paracellular pathway because absorptive flux is sensitive to the transepithelial potential difference. An alternative method to assess paracellular-mediated flux of charged organic molecules is suggested.

Drug permeability in 16HBE14o- airway cell layers correlates with absorption from the isolated perfused rat lung

The permeability of the lung is critical in determining the disposition of inhaled drugs and the respiratory epithelium provides the main physical barrier to drug absorption. The 16HBE14o- human bronchial epithelial cell line has been developed recently as a model of the airway epithelium. In this study, the transport of 10 low molecular weight compounds was measured in the 16HBE14o- cell layers, with apical to basolateral (absorptive) apparent permeability coefficients (P(app)) ranging from 0.4 x 10(-6)cms(-1) for Tyr-D-Arg-Phe-Phe-NH(2) to 25.2x10(-6)cms(-1) for metoprolol. Permeability in 16HBE14o- cells was found to correlate with previously reported P(app) in Caco-2 cells and absorption rates in the isolated perfused rat lung (k(a,lung)) and the rat lung in vivo (k(a,in vivo)). Log linear relationships were established between P(app) in 16HBE14o- cells and P(app) in Caco-2 cells (r(2)=0.82), k(a,lung) (r(2)=0.78) and k(a,in vivo) (r(2)=0.68). The findings suggest that permeability in 16HBE14o- cells may be useful to predict the permeability of compounds in the lung, although no advantage of using the organ-specific cell line 16HBE14o- compared to Caco-2 cells was found in this study.

Cell culture-based models for intestinal permeability: a critique

The model systems that are currently used to determine the intestinal permeability characteristics of discovery compounds often represent a combination of high-throughout, but less predictive, in silico and in vitro models and low-throughput, but more predictive, in vivo models. Cell-based permeability models have been integrated into the discovery paradigm for some time and represent the "method of choice" across the industry. Here, in addition to an objective analysis of the utility of cell culture models for permeability screening, anticipated future trends in the field of cell culture models are discussed.

Orally active antiviral tripeptide glycyl-prolyl-glycinamide is activated by CD26 (dipeptidyl peptidase IV) before transport across the intestinal epithelium

The tripeptide amide glycyl-prolyl-glycinamide (GPG-amide) is a new antiretroviral drug candidate, but its absorption mechanism is unknown. In this investigation, the transport and metabolism of GPG-amide were studied in a model of the human intestinal epithelium, Caco-2 cell monolayers. The results show that when the tripeptide amide came into contact with the apical enterocyte membrane, it was degraded by CD26 (dipeptidyl peptidase IV) to glycylproline and the antiretrovirally active metabolite glycinamide. Glycinamide retained antiretroviral activity in vitro after transport through the Caco-2 cell monolayers. The transport of glycinamide across Caco-2 cell monolayers occurred via passive diffusion with an apparent permeability coefficient of about 2 x 10(-6) cm s(-1), which suggests that it is absorbed by the oral route in sufficient amounts to be considered for oral administration. In conclusion, the tripeptide GPG-amide acts as a prodrug that is activated by CD26 to release the orally active antiretroviral compound glycinamide.

Which in vitro Screens Guide the Prediction of Oral Absorption and Volume of Distribution?

The development of medium to high-throughput in vitro screening of ADME (Absorption, Distribution, Metabolism, Excretion) properties has been the reply to higher demands on drug metabolism scientists to cope with progress in chemistry and biology. Two areas will be discussed here, namely screens for oral absorption and for volume of distribution. The prediction of these human pharmacokinetic parameters can be based on proper combination of simple physicochemical measurements. In the future in vitro screens most likely will be combined with in silico assessments of various ADME properties leading to the concept of in combo screening in drug discovery.

Determination of diffusion coefficients of peptides and prediction of permeability through a porous membrane

The diffusion coefficient (D) of peptide and protein drugs needs to be determined to examine the permeability through biological barriers and to optimize delivery systems. In this study, the D values of fluorescein isothiocyanate (FITC)-labelled dextrans (FDs) and peptides were determined and the permeability through a porous membrane was discussed. The observed D values of FDs and peptides, except in the case of insulin, were similar to those calculated based on a relationship previously reported between the molecular weight and D of lower-molecular-weight compounds, although the molecular weight range was completely different. The observed D value of insulin was between the calculated values for the insulin monomer and hexamer. The permeability of poly-lysine and insulin through the membrane was determined and the observed values were compared with predicted values by using the relationship between molecular weight and D and an equation based on the Renkin function. The observed permeability of insulin through the membrane was between that of the predicted permeability for the insulin monomer and hexamer. For the permeation of insulin, the determination of D was useful for estimating the permeability because of the irregular relationship between molecular weight and D. The methodology used in this study will be useful for a more quantitative evaluation of the absorption of peptide and protein drugs applied to mucous membranes.

Assessment and modulation of acamprosate intestinal absorption: comparative studies using in situ, in vitro (CACO-2 cell monolayers) and in vivo models

The purpose of this study was to explore the intestinal absorption mechanism of acamprosate and to attempt to improve the bioavailability (BA) of the drug through modulation of its intestinal absorption using two enhancers (polysorbate 80 and sodium caprate) based on in situ, in vitro and in vivo models and comparing the results obtained. Intestinal transport of the drug, in the absence and in presence of polysorbate 80 (0.06, 0.28 and 9.6 mM) or sodium caprate (13 and 16 mM) was measured by using an in situ rat gut technique and Caco-2 cell monolayers. Additionally, the effect of sodium caprate on drug oral bioavailability, measured as urinary recovery, was quantified by performing in vivo experiments with the rat as animal model. Only sodium caprate was able to increase the absorption rate constant (ka) of acamprosate in the mid-intestine of the rats from 0.29 +/- 0.07 h-1 in the absence of the promoter to 0.51 +/- 0.19 h-1 in the presence of C10 16 mM, along with the apparent permeability (Papp) obtained in Caco-2 cells (around two-fold). However, the drug bioavailability in rats (around 20%) did not improve in the presence of any of the concentrations tested (13, 16 and 50 mM). It is concluded that acamprosate absorption likely occurs via paracellular pathway and can be enhanced by sodium caprate in situ and in vitro but not in vivo-thus suggesting that although in situ and in vitro studies could be useful in early screening to select a potential promoter, in vivo studies in animal models are necessary to confirm the utility of the enhancer and to determine the influence of physiological variables.

The prediction of human oral absorption for diffusion rate-limited drugs based on heuristic method and support vector machine

Support vector machine (SVM), as a novel machine learning technique, was used for the prediction of the human oral absorption for a large and diverse data set using the five descriptors calculated from the molecular structure alone. The molecular descriptors were selected by heuristic method (HM) implemented in CODESSA. At the same time, in order to show the influence of different molecular descriptors on absorption and to well understand the absorption mechanism, HM was used to build several multivariable linear models using different numbers of molecular descriptors. Both the linear and non-linear model can give satisfactory prediction results: the square of correlation coefficient R(2) was 0.78 and 0.86 for the training set, and 0.70 and 0.73 for the test set respectively. In addition, this paper provides a new and effective method for predicting the absorption of the drugs from their structures and gives some insight into structural features related to the absorption of the drugs.

Caco-2 replace or refine?

Multiple screening techniques have been developed to gain simplicity and rapidness in prediction of human intestinal permeability. The most extensively used method for years has been the Caco-2 cell monolayers. Are the less time- and resource-consuming artificial membranes and computational-based predictions on their way to replace the Caco-2 cells? The importance of mechanistic approaches and correct interpretation of the data using the techniques available is compared and discussed in this article.:

Drug permeation in biomembranes

In the past decades, it has become increasingly apparent that in addition to therapeutic effect, drugs need to exhibit favourable absorption, distribution, metabolism and excretion (ADME) characteristics to produce a desirable response in vivo. As the recent progress in drug discovery technology enables rapid synthesis of vast numbers of potential drug candidates, robust methods are required for the effective screening of compounds synthesized within such programs, so that compounds with poor pharmacokinetic properties can be rejected at an early stage of drug development. Furthermore, a viable in silico method would save resources by enabling virtual screening of drug candidates already prior to synthesis. This review gives a general overview of the approaches aimed at predicting biological permeation, one of the cornerstones behind the ADME behaviour of drugs. The most important experimental and computational models are reviewed. Physicochemical factors underlying the permeation process are discussed.

An improved cell culture model based on 2/4/A1 cell monolayers for studies of intestinal drug transport: characterization of transport routes

PURPOSE

To improve the viability of the 2/4/A1 cell culture model and to investigate different routes of drug transport in this cell line.

METHODS

Two approaches were taken to decrease apoptosis. First, rat intestinal 2/4/A1 cells were transfected to overexpress the antiapoptotic protein Bcl-2. Second. normal 2/4/A1 cells were cultivated under conditions that stimulate differentiation and limit apoptosis. The monolayer integrity was investigated by transepithelial electrical resistance, permeability, and microscopy. The expression of drug transporters was investigated by RT-PCR, and transport function was assessed using specific markers.

RESULTS

Normal 2/4/A1 cells died by apoptosis at 39 degrees C. Bcl-2-expressing 2/4/A1 cells were viable but adopted a morphology of less-differentiated epithelial cells. Optimization of the culture conditions for 2/4/A1 cells inhibited cell death. The integrity was comparable to that of the human jejunum (50 omega x cm2), making this approach preferable to Bcl-2 overexpression. Transcriptional analysis showed that some (e.g., MDRI). but not all (e.g., PepT1), transporters were found in 2/4/A1 cells. Studies using substrates for PepT1, P-gp. MRP2, and BCRP showed that none of the transporters were functional in 2/4/A1.

CONCLUSIONS

The improved culture procedure will facilitate the use of 2/4/A1 cells. 2/4/A1 lack several transporters, which makes them a promising alternative to Caco-2 cells and artificial membranes in studies of passive drug transport.