K.-C. Cheng

Use of intrinsic clearance for prediction of human hepatic clearance

Importance of the field: The use of intrinsic metabolic stability/clearance and other in vitro pharmacokinetic data for the selection of drug candidates for clinical evaluation during discovery lead optimization has become one of the primary focuses of research organizations involved in new drug discovery. Using intrinsic clearance determined from human liver microsomal preparations and/or hepatocyte to predict human clearance has become more acceptable. Areas covered in this review: This review focuses on the current methods for determining intrinsic clearance and scaling to predict human hepatic clearance, and novel physiologically-based models for improvement of human hepatic clearance prediction. Published microsomal metabolic stability data and in-house hepatocyte clearance data were compared with published in vivo human hepatic clearance data. Various scaling models and the effect of protein binding were examined. What the reader will gain: Use of a novel microfluidic model and other physiologically-based models are presented. Microsomal metabolic clearance requires correction for protein binding and in vitro microsomal binding in order to better predict in vivo hepatic clearance of compounds that are mainly eliminated by hepatic metabolism. Take home message: Metabolic clearance obtained using hepatocytes may work well in combination with the well-stirred model. Novel models incorporating flow and protein binding in the system may be the most complete models for prediction of human in vivo metabolism.

Application and interpretation of hPXR screening data : Validation of reporter signal requirements for prediction of clinically relevant CYP3A4 inducers

A human pregnane X receptor (PXR) reporter-gene assay was established and validated using 19 therapeutic agents known to be clinical CYP3A4 inducers, 5 clinical non-inducers, and 6 known inducers in human hepatocytes. The extent of CYP3A4 induction (measured as RIF ratio in comparison to rifampicin) and EC50 was obtained from the dose-response curve. All of the clinical inducers (19/19) and human hepatocyte inducers (6/6) showed positive responses in the PXR assay. One out of five clinical non-inducers, pioglitazone, also showed a positive response. An additional series of 18 commonly used drugs with no reports of clinical induction was also evaluated as putative negative controls. Sixteen of these were negative (89%), whereas two of these, flutamide and haloperidol showed 16-fold (RIF ratio 0.79) and 10-fold (RIF ratio 0.48) maximal induction, respectively in the reporter-gene system. Flutamide and haloperidol were further demonstrated to cause CYP3A4 induction in human cryopreserved hepatocytes based on testosterone 6beta-hydroxylation activity. The induction potential index calculated based on the maximum RIF ratio, EC50, and in vivo maximum plasma concentration was used to predict the likelihood of CYP3A4 induction in humans. When the induction potential index is greater than 0.08, the compound is likely to cause induction in humans. A high-throughput screening strategy was developed based on the validation results at 1microM and 10microM for the same set of drugs. A RIF ratio of 0.4 was set as more practical screening cut-off to minimize the possibility of generating false positives. Thus, a tiered approach was implemented to use the human PXR reporter-gene assay from early lead optimization to late lead characterization in drug discovery.

Development of In Vitro Pharmacokinetic Screens Using Caco-2, Human Hepatocyte, and Caco-2/Human Hepatocyte Hybrid Systems for the Prediction of Oral Bioavailability in Humans

In this study, in vitro systems were used to build 2 pharmacokinetic models that predict human oral bioavailability: the Caco-2/hepatocyte combination model and the Caco-2/hepatocyte hybrid model. Data obtained in vitro on Caco-2 cell permeability and hepatocyte clearance are routinely used to predict the fraction of absorption after oral administration and the extent of first-pass metabolism, respectively. In the Caco-2/hepatocyte combination model, results from a Caco-2 cell permeability assay and a hepatocyte clearance assay were combined to project oral bioavailability. Comparison of oral bioavailabilities predicted by the combination model and reported oral bioavailabilities in humans for 30 marketed compounds resulted in a modest correlation (r 2 = 0.66). The Caco-2/hepatocyte hybrid model, as previously reported, joins the Caco-2 and hepatocyte clearance systems into 1 assay. Improvements to the previous model were made by incorporating an elimination phase into the Caco-2/hepatocyte hybrid model. In the new hybrid model, the compound was added to a Caco-2-containing donor compartment and allowed to permeate for 2 h to a hepatocyte-containing receiver compartment. Subsequently, to mimic an elimination phase, the donor compartment was removed, and permeated compound was incubated with hepatocytes alone for an additional 3 h. The area under the concentration versus time curve (AUC) was determined for each of the same 30 marketed compounds assessed by the combination model. A linear regression analysis comparing the in vitro AUCs and reported oral bioavailabilities in humans showed a reasonable correlation (r 2 = 0.73). This study demonstrates that the Caco-2/hepatocyte hybrid model is more favorable and further proves the potential and feasibility of using in vitro screenings for the prediction of in vivo pharmacokinetics in humans. (Journal of Biomolecular Screening 2007:1084-1091)

Correlation between PAMPA permeability and cellular activities of hepatitis C virus protease inhibitors

Parallel artificial membrane permeability assay (PAMPA) and Caco-2 cells have been frequently used for the evaluation of in vitro permeability of new chemical entities. In this study we evaluated the correlation between permeability, assessed by both methods, and the cellular potency of 34 novel hepatitis C virus (HCV) protease inhibitors. Two types of assays were used to determine the potency of HCV protease inhibitors: a cell-free assay that evaluates the intrinsic affinity (K(i)) between the protease and the inhibitor and a cell-based replicon assay that determines the inhibitors IC90. When the K(i)/IC90 ratios were compared with the PAMPA permeability and the Caco-2 permeability by linear regression analysis, a reasonable correlation was found between the K(i)/IC90 ratio and PAMPA permeability (r2=0.76) but not with Caco-2 permeability (r2=0.29). Correlations were also assessed between K(i)/IC90 ratios and the following physico-chemical properties: logP (r2=0.41), logD (r2=0.58), clogP (r2=0.13), and mlogP (r2=0.30). These results suggest that passive permeability may play a role in the uptake and cellular activity of these HCV protease inhibitors, and that PAMPA was more predictive of cellular activity than physico-chemical properties or Caco-2 permeability.

The introduction of P4 substituted 1-methylcyclohexyl groups into Boceprevir®: A change in direction in the search for a second generation HCV NS3 protease inhibitor

In the search for a second generation HCV protease inhibitor, molecular modeling studies of the X-ray crystal structure of Boceprevir1 bound to the NS3 protein suggest that expansion into the S4 pocket could provide additional hydrophobic Van der Waals interactions. Effective replacement of the P4 tert-butyl with a cyclohexylmethyl ligand led to inhibitor 2 with improved enzyme and replicon activities. Subsequent modeling and SAR studies led to the pyridine 38 and sulfone analogues 52 and 53 with vastly improved PK parameters in monkeys, forming a new foundation for further exploration.

Evaluation of the binding orientations of testosterone in the active site of homology models for CYP2C11 and CYP2C13.

Cytochromes P-450 2C11 and 2C13 are the major CYPs in rat liver microsomes. Despite a high degree of sequence identity, these two isozymes display different positional and regio-specific metabolism of steroid hormones, such as testosterone. CYP2C11 converts testosterone to 2alpha-hydroxyl and 16alpha-hydroxyl metabolites, while CYP2C13 produces primarily the 6beta-hydroxyl metabolite. Using a human CYP2C9 crystal structure as the template, homology models were generated for CYP2C11 and CYP2C13. Despite similar volume of the binding pockets for CYP2C11 and CYP2C13, the models for these two CYPs showed a substantial difference in the shape of the substrate-binding sites. Substrate docking using rigid and induced-fit methods showed that testosterone fits into the substrate-binding sites of both CYP2C11 and CYP2C13 without the need of added constraints. These docking exercises appear to support testosterone binding in both CYP2C11 and CYP2C13. A constrained docking using energy minimization is required to position testosterone for more precise positional and regio-specificity in supporting the observed metabolism. These results demonstrate the complexity of using modeling for understanding the binding of substrate to CYPs, and suggest that, as a complement to the metabolism data, modeling and docking may yield reliable structural information for the molecular interaction between the substrate and the CYPs.

Co-Induction of CYP3A12 and 3A26 in Dog Liver Slices by Xenobiotics: Species Difference Between Human and Dog CYP3A Induction

The induction of dog CYP3A12 and CYP3A26 mRNA levels was evaluated in liver slices after treatment with 22 xenobiotics. Eleven of the 22 xenobiotics increased 3A12 mRNA by more than four-fold, while nine did the same for 3A26 mRNA. A four-fold increase in the mRNA level was used as the cut-off for indication of induction based on the noise level of the real time-PCR. A good correlation was found between the mRNA levels for 3A12 and 3A26 after treatment with compounds, suggesting that these two CYPs may be co-induced. Induction of CYP3A4 in human hepatocytes was evaluated after treatment with the same 22 compounds. Thirteen out of the 22 compounds increased the 3A4 mRNA levels by more than four-fold. When the mRNA levels of 3A4 and 3A12 were compared after treatment with compounds, no correlation was found. The regulation of CYP3A expression has been demonstrated to be controlled by pregnane X receptor (PXR). Upon examination of the sequence homology and the three-dimensional structures of human PXR and a dog PXR model, only two different amino acids (met323/val and arg410/lys) were found in the ligand-binding domain. This finding suggests that these two amino acids may play a role in the binding specificity of ligands.

Permeability evaluation of peptidic HCV protease inhibitors in Caco-2 cells-correlation with in vivo absorption predicted in humans

The permeability of six peptidic hepatitis C virus (HCV) protease inhibitors, with molecular weights ranging from 500 to 780, was examined in the Caco-2 cell system. All six compounds permeated the cells transcellularly; paracellular permeability, evaluated in the Caco-2 cell system by reducing the calcium concentration in the media to increase the pore size of the tight junctions, most likely contributes only minimally to the oral absorption of the compounds. All six compounds were shown to be efflux substrates displaying concentration-dependent saturation of efflux. The efflux could be blocked by cyclosporine A, a specific P-glycoprotein (P-gp) inhibitor, suggesting that P-gp may be the responsible transporter. Oral absorption in rats was calculated using in vivo oral bioavailability and hepatic extraction ratios. Human oral absorption was projected to be similar to that of rats, as reported previously by comparing rat and human absorption values for 23 marketed drugs. Upon comparison of human oral absorption predicted by Caco-2 permeability and by rat pharmacokinetics, we show a better correlation with Caco-2 permeability obtained at higher compound concentrations, where efflux is saturated, than at lower concentrations. The higher concentrations are likely reflecting the lumen concentrations after in vivo oral dosing. The results presented in this study demonstrate that, when tested at relevant compound concentrations, Caco-2 permeability is useful for predicting the oral absorption of peptidic compounds.

High-Throughput Screening Using Caco-2 Cell and PAMPA Systems

This chapter will focus on the two most commonly used high-throughput screening methods detecting cellular/bio-membrane permeability in the pharmaceutical industry: Caco-2 cells and the parallel artificial membrane permeability assay (PAMPA). Both assays have advantages and disadvantages, and it is essential to understand these limitations. Since it is well recognized that human intestinal absorption cannot be precisely predicted by a single screening assay, it is important to utilize various in vitro and in vivo preclinical studies during lead optimization in drug discovery.

Rat PXR reporter-gene activity correlates with the induction of CYP3A in rat precision-cut liver slices.

Recent studies have suggested that both constitutive androstane receptor (CAR) and pregnane X-receptor (PXR) are involved in the induction of rat liver microsomal cytochrome P-450 (CYP) 2B and 3A through a mechanism called cross-talk. In this study we intend to determine if a PXR-reporter gene assay could be used for the prediction of CYP3A and/or CYP2B induction in rats. The induction of rat CYP2B and CYP3A by nineteen structurally diverse compounds was evaluated by using rat precision-cut liver slices and a rat PXR reporter-gene system. Induction of CYP2B and CYP3A mRNAs in rat liver slices was quantified by real-time polymerase chain reaction. Rat PXR activation was measured by induction of luciferase activity in rat PXR reporter-gene system. Linear regression analysis of the fold of induction of mRNA in liver slices and the fold of luciferase activity in rat PXR reporter-gene system shows that a reasonable correlation (r2 = 0.6) exists between the CYP3A induction and the rat PXR activation. A much lower correlation was observed between CYP2B induction and the rat PXR activation (r2 = 0.1). The results from this study suggest that the PXR may play a major role in the induction of rat CYP3A, but not CYP2B. Therefore, the PXR-reporter gene assay may be useful in a high-throughput screening to predict CYP3A induction in rats.