Predicting Passive Permeability of Drug-like Molecules from Chemical Structure: Where Are We?

Intestinal absorption in human is routinely predicted in drug discovery using in vitro assays such as permeability in the Madin-Darby canine kidney cell line. In silico models trained on these data are used in drug discovery efforts to prioritize novel chemical targets for synthesis; however, their proprietary nature and the limited validation available, which is usually restricted to predicting in vitro permeability, are barriers to widespread adoption. Because of the categorical nature of the in vitro permeability assay, intrinsic assay variability, and the challenges often encountered when translating in vitro data to an in vivo drug property, validation based solely on in vitro data might not be a good characterization of the usefulness of the in silico tool. In this work, we analyze the performance of three different in silico models in predicting the in vitro and in vivo permeability of 300 marketed drugs and 86 discovery compounds. The models differ in their approach (mechanistic vs quantitative structure-activity relationship) and the degree of complexity; one of them is a linear equation based on seven simple physicochemical descriptors and is presented for the first time in this work. Results show that in silico models can be successfully used to complement the discovery toolbox for characterizing in vivo intestinal permeability, defined using fraction of dose absorbed in human (Fa) and human jejunal permeability (Peff). While the in vitro permeability models outperformed the in silico approach at predicting each of the in vivo end points explored, the gap in predictivity between the in vitro and the in vivo data was generally comparable to the gap between in silico and in vitro data. The in vitro and in silico approaches shared many of the same outliers, which can often be explained by the route of drug absorption (paracellular vs transcellular, active vs passive). Data suggest that the discovery process can greatly benefit from an early adoption of in silico models for predicting permeability as well as from a careful analysis of the in silico to in vivo disconnects.

Highlights From the American Association of Pharmaceutical Scientists/ International Transporter Consortium Joint Workshop on Drug Transporters in Absorption, Distribution, Metabolism, and Excretion: From the Bench to the Bedside - Clinical Pharmacology Considerations.

The American Association of Pharmaceutical Scientists/International Transporter Consortium Joint Workshop on Drug Transporters in absorption, distribution, metabolism, and excretion was held with the objective of discussing innovative advances in transporter pharmacology. Specific topics included (i) transporters at the blood-brain barrier (BBB); (ii) emerging transport proteins; (iii) recent advances in achieving hepatoselectivity and optimizing clearance for organic anion-transporting polypeptide (OATP) substrates; (iv) utility of animal models for transporter studies; and (v) clinical correlation of transporter polymorphisms. Here, we present state-of-the-art highlights from this workshop in these key areas of focus.

Ontogeny of Hepatic Drug Transporters as Quantified by LC-MS/MS Proteomics.

Protein expression of major hepatic uptake and efflux drug transporters in human pediatric (n = 69) and adult (n = 41) livers was quantified by liquid chromatography / tandem mass spectroscopy (LC-MS/MS). Transporter protein expression of OCT1, OATP1B3, P-gp, and MRP3 was age-dependent. Particularly, significant differences were observed in transporter expression (P < 0.05) between the following age groups: neonates vs. adults (OCT1, OATP1B3, P-gp), neonates or infants vs. adolescents and/or adults (OCT1, OATP1B3, and P-gp), infants vs. children (OATP1B3 and P-gp), and adolescents vs. adults (MRP3). OCT1 showed the largest increase, of almost 5-fold, in protein expression with age. Ontogenic expression of OATP1B1 was confounded by genotype and was revealed only in livers harboring SLCO1B1*1A/*1A. In livers >1 year, tissues harboring SLCO1B1*14/*1A showed 2.5-fold higher (P < 0.05) protein expression than SLCO1B1*15/*1A. Integration of these ontogeny data in physiologically based pharmacokinetic (PBPK) models will be a crucial step in predicting hepatic drug disposition in children.

Evaluation of a single-pass intestinal-perfusion method in rat for the prediction of absorption in man.

Prediction of the fraction of dose absorbed from the intestine (Fa) in man is essential in the early drug discovery stage. In-vitro assays in Caco-2 and MDCK cells are routinely used for that purpose, and their predictive value has been reported. However, in-situ techniques might provide a more accurate estimation of Fa. In this study, we evaluated a single-pass intestinal-perfusion (SPIP) method in the rat for its use in the prediction of absorption in man and compared it with a previous report using cell-based assays. Effective permeability coefficients (Peff) were determined in rats for 14 compounds, and ranged from 0.043x 10(-4) cm s(-1) to 1.67 x 10(-4) cm s(-1). These values strongly correlated (r2 = 0.88) with reported Peff values for man. In addition, the Spearman rank correlation coefficient calculated for in-situ-derived Peff and absorption in man was 0.92 while for the previously tested in-vitro Caco-2 and MDCK systems vs absorption in man, the correlation coefficients were 0.61 and 0.59, respectively. SPIP provided a better prediction of human absorption than the cell-based assays. This method, although time consuming, could be used as a secondary test for studying the mechanisms governing the absorption of new compounds, and for predicting more accurately the fraction absorbed in man.

Active secretion and enterocytic drug metabolism barriers to drug absorption.

Intestinal phase I metabolism and active extrusion of absorbed drug have only recently been recognized as major determinants of oral drug bioavailability. Both CYP3A4, the major phase I drug metabolizing enzyme in humans, and the multidrug efflux pump, P-glycoprotein (P-gp), are present at high levels in the villus enterocytes of the small intestine, the primary site of absorption for orally administered drugs. Moreover, these proteins are induced by many of the same compounds and demonstrate a broad overlap in substrate and inhibitor specificities, suggesting that they act as a concerted barrier to drug absorption. Clinical studies have demonstrated that inhibition of CYP3A4-mediated intestinal metabolism can significantly improve the oral bioavailability of a wide range of drugs. Intestinal P-gp is a major route of elimination for both orally and intravenously administered anticancer drugs in animal models, and experiments with the Caco-2 cell line have provided strong evidence that inhibition of intestinal P-gp is another means by which oral drug bioavailability could be enhanced.

Metabolism of digoxin and digoxigenin digitoxosides in rat liver microsomes: involvement of cytochrome P4503A.

1. The sequential metabolism of digoxin (Dg3) to digoxigenin bis-digitoxoside (Dg2), digoxigenin mono-digitoxoside (Dg1) and digoxigenin (Dg0) was investigated in rat liver microsomes. 2. Kinetic studies produced results consistent with a single enzyme mechanism describing the successive oxidative cleavages. Formation of Dg2 was catalysed with mean (+/-SD) Km and Vmax of 125 +/- 22 microM and 362 +/- 37 pmol/min/mg protein, respectively. The corresponding values for the formation of Dg1 were 61 +/- 5 microM and 7 +/-1 pmol/min/mg protein. Dg0 formation was catalysed with the apparent values of 30 +/- 9 microM and 310 +/- 30 pmol/min/mg protein. 3. Chemical inhibition of cytochrome P450 (CYP) 3A subfamily with ketoconazole and triacetyoleandomycin decreased the formation of Dg2 and Dg1 by up to 90%. Antibodies specific to rat CYP3A2 lowered the rate of oxidative cleavage of Dg3 and Dg2 by up to 85%. Inhibition of CYP2E1, CYP2C subfamily and CYP1A2 by chemical and immuno-inhibition did not affect initial rates of metabolism of Dg3 and Dg2. In contrast, Dg1 metabolism was not affected by triacetyloleandomycin as well as by antibodies to CYP3A2, CYP2C11, CYP2E1, CYP2B1/2B2 and CYP1A2. It was however inhibited by >80% by gestodene and 17alpha-ethynylestradiol (selective inhibitors of human CYP3A). 4. Collectively, these data support the involvement of CYP3A in the cleavage of Dg3 and Dg2 in rat liver microsomes. The enzyme-metabolizing Dg1 remains to be identified.

Effects of ketoconazole on digoxin absorption and disposition in rat.

Digoxin, a cardiac glycoside, is a substrate of the multidrug transporter P-glycoprotein (Pgp), and in rats has also been identified as a substrate for cytochrome P450 3A (CYP3A). Ketoconazole, an antifungal agent, was shown to inhibit Pgp in a multidrug-resistant cell line, and is known to be a potent inhibitor of CYP3A. Here, we determined the effects of ketoconazole on digoxin absorption and disposition in rats. Digoxin was administered intravenously or orally with or without a concomitant oral dose of ketoconazole. When given intravenously, digoxin AUC increased from 93 +/- 22 to 486 +/- 26 microg x h/l with ketoconazole administration. Similarly, ketoconazole raised the AUC of orally administered digoxin from 63 +/- 17 to 411 +/- 50 microg x h/l. Concomitant ketoconazole administration prolonged digoxin elimination, yielding a nonlinear pharmacokinetic profile. Using time-averaged values, digoxin bioavailability increased from 0.68 +/- 0.18 to 0.84 +/- 0.10, while mean absorption time was reduced from 1.1 +/- 0.4 to 0.3 +/- 0.1 h. Thus, in rats, ketoconazole increases digoxin plasma concentrations, rate of absorption and bioavailability. Although the effects of ketoconazole on AUC could be explained by inhibition of both CYP3A and Pgp, which cannot be differentiated in this study, the decreased mean absorption time can only be explained by inhibition of Pgp in the intestine.

Modulation of P-glycoprotein expression by cytochrome P450 3A inducers in male and female rat livers.

A strong overlap between P-glycoprotein (Pgp) and cytochrome P450 3A (CYP3A) substrates and modulators has been reported. To test the hypothesis that CYP3A and Pgp are coordinately regulated, we examined the effects of known inducers of CYP3A (triacetyloleandomycin, rifampicin, dexamethasone, pregnenolone 16alpha-carbonitrile) on Pgp expression in rat liver. We also investigated the gender-specific expression of Pgp and compared its response to dexamethasone between male and female rats. In male rats, western blot analyses showed that rifampicin and dexamethasone caused 50% and 5-fold increases in Pgp levels, respectively. RNase protection assays using gene-specific probes for the three Pgp isoforms revealed a 3-fold increase in mdr2 mRNA levels after dexamethasone administration and a 2-fold increase following rifampicin treatment. Triacetyloleandomycin and pregnenolone 16alpha-carbonitrile had no effect on Pgp expression and mRNA levels. We also observed that the basal level of Pgp was 40% lower in male rats than in females and that mdr2 mRNA levels in male rats were one-half those in females. As opposed to the results in male rats, dexamethasone reduced Pgp expression by approximately 60% and caused a 30% decrease in mdr2 mRNA levels in female rats. Mdr1a was not affected and mdr1b was not detected in female or male rats. We conclude that, at the dosage regimen used, CYP3A and Pgp responses to CYP3A inducers are regulated independently in rat liver. In addition, this study shows that Pgp expression and regulation are gender specific.

Variability in the disposition of chlorzoxazone.

Chlorzoxazone is 6-hydroxylated by cytochrome P450 2E1 (CYP 2E1), which bioactivates many toxic and carcinogenic molecules. Seventeen volunteers of varying age, ethnicity, and gender received a 250 mg tablet of chlorzoxazone and their blood and urine were sampled frequently for 8 h. V/F = 42 +/- 21 L and CL/F = 412 +/- 120 mL min-1. Comparison of these values with a study by other investigators using a suspension dosage form suggested that relative Ftablet approximately 0.7. The fraction excreted in the urine as 6-hydroxychlorzoxazone (fe,6-OH) was 0.39 +/- 0.20 and that portion of the total CL accounted for by CYP 2E1-mediated metabolism (CL6-OH) was 163 +/- 95 mL min-1. Thus, while V/F and CL/F varied by factors of less than five, fe,6-OH varied 16-fold and CL6-OH varied 28-fold. These results suggested that there was considerable inter-individual variability in the metabolism of chlorzoxazone to 6-hydroxychlorzoxazone. This variability will significantly affect the construction of physiologically based pharmacokinetic models that use the 6-hydroxylation of chlorzoxazone as a marker for an individual's CYP 2E1 phenotype.