Non-specific binding of compounds in in vitro metabolism assays: a comparison of microsomal and hepatocyte binding in different species and an assessment of the accuracy of prediction models.

Non-specific binding in in vitro metabolism systems leads to an underestimation of the true intrinsic metabolic clearance of compounds being studied. Therefore in vitro binding needs to be accounted for when extrapolating in vitro data to predict the in vivo metabolic clearance of a compound. While techniques exist for experimentally determining the fraction of a compound unbound in in vitro metabolism systems, early in drug discovery programmes computational approaches are often used to estimate the binding in the in vitro system.Experimental fraction unbound data (n = 60) were generated in liver microsomes (fumic) from five commonly used pre-clinical species (rat, mouse, dog, minipig, monkey) and humans. Unbound fraction in incubations with mouse, rat or human hepatocytes was determined for the same 60 compounds. These data were analysed to determine the relationship between experimentally determined binding in the different matrices and across different species. In hepatocytes there was a good correlation between fraction unbound in human and rat (r2=0.86) or mouse (r2=0.82) hepatocytes. Similar correlations were observed between binding in human liver microsomes and microsomes from rat, mouse, dog, Göttingen minipig or monkey liver microsomes (r2 of >0.89, n = 51 - 52 measurements in different species). Physicochemical parameters (logP, pKa and logD) were predicted for all evaluated compounds. In addition, logP and/or logD were measured for a subset of compounds.Binding to human hepatocytes predicted using 5 different methods was compared to the measured data for a set of 59 compounds. The best methods evaluated used measured microsomal binding in human liver microsomes to predict hepatocyte binding. The collated physicochemical data were used to predict the human fumic using four different in silico models for a set of 53-60 compounds. The correlation (r2) and root mean square error between predicted and observed microsomal binding was 0.69 & 0.20, 0.47 & 0.23, 0.56 & 0.21 and 0.54 & 0.26 for the Turner-Simcyp, Austin, Hallifax-Houston and Poulin models, respectively. These analyses were extended to include measured literature values for binding in human liver microsomes for a larger set of compounds (n=697). For the larger dataset of compounds, microsomal binding was well predicted for neutral compounds (r2=0.67 - 0.70) using the Poulin, Austin, or Turner-Simcyp methods but not for acidic or basic compounds (r2<0.5) using any of the models. While the lipophilicity-based models can be used, the in vitro binding should be measured for compounds where more certainty is needed, using appropriately calibrated assays and possibly established weak, moderate, and strong binders as reference compounds to allow comparison across databases.

Physiologically-Based Pharmacokinetic Models of CYP2D6 Substrate and Inhibitors Nebivolol, Cinacalcet and Mirabegron to Simulate Drug-Drug Interactions.

BACKGROUND AND OBJECTIVES: Index substrates and inhibitors to investigate the role of the polymorphic enzyme, cytochrome P450 (CYP) 2D6, in the metabolism of new compounds have been proposed by regulatory agencies. This work describes the development and verification of physiologically-based pharmacokinetic (PBPK) models for the CYP2D6-sensitive substrate, nebivolol and the index CYP2D6 inhibitors, mirabegron and cinacalcet. METHODS: PBPK models for nebivolol, mirabegron and cinacalcet were developed using in vitro and clinical data. The performance of the PBPK models was verified by comparing the simulated results against reported human systemic exposure and clinical drug-drug interactions (DDIs) studies. RESULTS: The exposure of nebivolol, cinacalcet and mirabegron predicted by the PBPK models was verified against pharmacokinetic data from 13, 3 and 9 clinical studies, respectively. For nebivolol, the predicted mean maximum plasma concentration (Cmax) and area under the plasma concentration-time (AUC) values in CYP2D6 extensive metaboliser subjects were within 0.9- to 1.49-fold of the observed values. In poor metaboliser CYP2D6 subjects, the predicted Cmax and AUC values were within 0.41- to 0.81-fold of observed values. For cinacalcet, the predicted Cmax and AUC values were within 0.97- to 1.32-fold of the observed data. For mirabegron, the predicted AUC values across all the studies investigated were within 0.71- to 1.88-fold of observed values. The PBPK model-predicted DDIs were in good agreement (within 2-fold) with observed DDIs in all verification studies (n = 8) assessed. The overall precision was 1.26 and 1.21 for Cmax and the AUC ratio, respectively. CONCLUSIONS: The developed PBPK models can be used to assess the DDI potential liability of new chemical entities that are substrates or inhibitors of CYP2D6.

Understanding Interindividual Variability in the Drug Interaction of a Highly Extracted CYP1A2 Substrate Tizanidine: Application of a Permeability-Limited Multicompartment Liver Model in a Population Based Physiologically Based Pharmacokinetic Framework.

Tizanidine, a centrally acting skeletal muscle relaxant, is predominantly metabolized by CYP1A2 and undergoes extensive hepatic first-pass metabolism after oral administration. As a highly extracted drug, the systemic exposure to tizanidine exhibits considerable interindividual variability and is altered substantially when coadministered with CYP1A2 inhibitors or inducers. The aim of the current study was to compare the performance of a permeability-limited multicompartment liver (PerMCL) model, which operates as an approximation of the dispersion model, and the well stirred model (WSM) for predicting tizanidine drug-drug interactions (DDIs). Physiologically based pharmacokinetic models were developed for tizanidine, incorporating the PerMCL model and the WSM, respectively, to simulate the interaction of tizanidine with a range of CYP1A2 inhibitors and inducers. Whereas the WSM showed a tendency to underpredict the fold change of tizanidine area under the plasma concentration-time curve (AUC ratio) in the presence of perpetrators, the use of PerMCL model increased precision (absolute average-fold error: 1.32-1.42 versus 1.58) and decreased bias (average-fold error: 0.97-1.25 versus 0.63) for the predictions of mean AUC ratios as compared with the WSM. The PerMCL model captured the observed range of individual AUC ratios of tizanidine as well as the correlation between individual AUC ratios and CYP1A2 activities without interactions, whereas the WSM was not able to capture these. The results demonstrate the advantage of using the PerMCL model over the WSM in predicting the magnitude and interindividual variability of DDIs for a highly extracted sensitive substrate tizanidine. SIGNIFICANCE STATEMENT: This study demonstrates the advantages of the PerMCL model, which operates as an approximation of the dispersion model, in mitigating the tendency of the WSM to underpredict the magnitude and variability of DDIs of a highly extracted CYP1A2 substrate tizanidine when it is administered with CYP1A2 inhibitors or inducers. The physiologically based pharmacokinetic modeling approach described herein is valuable to the understanding of drug interactions of highly extracted substrates and the source of its interindividual variability.

A Mechanistic Absorption and Disposition Model of Ritonavir to Predict Exposure and Drug-Drug Interaction Potential of CYP3A4/5 and CYP2D6 Substrates.

BACKGROUND AND OBJECTIVES: Due to health authority warnings and the recommended limited use of ketoconazole as a model inhibitor of cytochrome P450 (CYP) 3A4 in clinical drug-drug interaction (DDI) studies, there is a need to search for alternatives. Ritonavir is a strong inhibitor for CYP3A4/5-mediated DDIs and has been proposed as a suitable alternative to ketoconazole. It can also be used as a weak inhibitor for CYP2D6-mediated DDIs. Most of the currently available physiologically based pharmacokinetic (PBPK) inhibitor models developed for predicting DDIs use first-order absorption models, which do not mechanistically capture the effect of formulations on the systemic exposure of the inhibitor. Thus, the main purpose of the current study was to verify the predictive performance of a mechanistic absorption and disposition model of ritonavir when it was applied to the inhibition of CYP2D6 and CYP3A4/5 by ritonavir. METHODS: A PBPK model that incorporates formulation characteristics and enzyme kinetic parameters for post-absorptive pharmacokinetic processes of ritonavir was constructed. Key absorption-related parameters in the model were determined using mechanistic modelling of in vitro biopharmaceutics experiments. The model was verified for systemic exposure and DDI risk assessment using clinical observations from 13 and 18 studies, respectively. RESULTS: Maximal inhibition of hepatic (3.53% of the activity remaining) and gut (5.16% of the activity remaining) CYP3A4 activity was observed when ritonavir was orally administered in doses of 100 mg or higher. The PBPK model accurately described the concentrations of ritonavir in the different simulated studies. The prediction accuracy for maximum concentration (Cmax) and area under the plasma concentration versus time curve (AUC) were assessed. The bias (average fold error, AFE) for the prediction of Cmax and AUC was 0.92 and 1.06, respectively, and the precision (absolute average fold error, AAFE) was 1.29 and 1.23, respectively. The PBPK model predictions for all Cmax and AUC ratios when ritonavir was used as an inhibitor of CYP metabolism fell within twofold of the clinical observations. The prediction accuracy for Cmax and AUC ratios had a bias (AFE) of 0.85 and 0.99, respectively, and a precision (AAFE) of 1.21 and 1.33, respectively. CONCLUSIONS: The current model, which incorporates formulation characteristics and mechanistic disposition parameters, can be used to assess the DDI potential of CYP3A4/5 and CYP2D6 substrates administered with a twice-daily dose of 100 mg of ritonavir for 14 days.

Prediction of CYP-mediated DDIs involving inhibition: Approaches to address the requirements for system qualification of the Simcyp Simulator.

Physiologically-based pharmacokinetic (PBPK) modeling is being increasingly used in drug development to avoid unnecessary clinical drug-drug interaction (DDI) studies and inform drug labels. Thus, regulatory agencies are recommending, or indeed requesting, more rigorous demonstration of the prediction accuracy of PBPK platforms in the area of their intended use. We describe a framework for qualification of the Simcyp Simulator with respect to competitive and mechanism-based inhibition (MBI) of CYP1A2, CYP2D6, CYP2C8, CYP2C9, CYP2C19, and CYP3A4/5. Initially, a DDI matrix, consisting of a range of weak, moderate, and strong inhibitors and substrates with varying fraction metabolized by specific CYP enzymes that were susceptible to different degrees of inhibition, were identified. Simulations were run with 123 clinical DDI studies involving competitive inhibition and 78 clinical DDI studies involving MBI. For competitive inhibition, the overall prediction accuracy was good with an average fold error (AFE) of 0.91 and 0.92 for changes in the maximum plasma concentration (Cmax ) and area under the plasma concentration (AUC) time profile, respectively, as a consequence of the DDI. For MBI, an AFE of 1.03 was determined for both Cmax and AUC. The prediction accuracy was generally comparable across all CYP enzymes, irrespective of the isozyme and mechanism of inhibition. These findings provide confidence in application of the Simcyp Simulator (V19 R1) for assessment of the DDI potential of drugs in development either as inhibitors or victim drugs of CYP-mediated interactions. The approach described herein and the identified DDI matrix can be used to qualify subsequent versions of the platform.

Liquid Biopsy Enables Quantification of the Abundance and Interindividual Variability of Hepatic Enzymes and Transporters.

Variability in individual capacity for hepatic elimination of therapeutic drugs is well recognized and is associated with variable expression and activity of liver enzymes and transporters. Although genotyping offers some degree of stratification, there is often large variability within the same genotype. Direct measurement of protein expression is impractical due to limited access to tissue biopsies. Hence, determination of variability in hepatic drug metabolism and disposition using liquid biopsy (blood samples) is an attractive proposition during drug development and in clinical practice. This study used a multi-"omic" strategy to establish a liquid biopsy technology intended to assess hepatic capacity for metabolism and disposition in individual patients. Plasma exosomal analysis (n = 29) revealed expression of 533 pharmacologically relevant genes at the RNA level, with 147 genes showing evidence of expression at the protein level in matching liver tissue. Correction of exosomal RNA expression using a novel shedding factor improved correlation against liver protein expression for 97 liver-enriched genes. Strong correlation was demonstrated for 12 key drug-metabolizing enzymes and 4 drug transporters. The developed test allowed reliable patient stratification, and in silico trials demonstrated utility in adjusting drug dose to achieve similar drug exposure between patients with variable hepatic elimination. Accordingly, this approach can be applied in characterization of volunteers prior to enrollment in clinical trials and for patient stratification in clinical practice to achieve more precise individual dosing.

Biopharmaceutic In Vitro In Vivo Extrapolation (IVIV_E) Informed Physiologically-Based Pharmacokinetic Model of Ritonavir Norvir Tablet Absorption in Humans Under Fasted and Fed State Conditions.

Ritonavir is a well-known CYP3A4 and CYP2D6 enzyme inhibitor, frequently used to assess the drug-drug interaction (DDI) liability of susceptible drugs. It is also used as a pharmacokinetic booster to increase exposure to CYP3A4 substrates. This study aimed to develop a mechanistic absorption and disposition model to describe exposure to ritonavir following oral dosing of the commercial amorphous solid dispersion tablet, Norvir, under fasted and fed conditions. A mechanistic description of ritonavir absorption from Norvir tablets may help to improve the design of DDI studies. Key parameters of amorphous ritonavir including free base solubility (solubility of the unbound, un-ionized species), bile micelle partition coefficients, formulation wetting/disintegration, and in vivo precipitation parameters were either obtained from the literature or estimated by modeling in vitro biopharmaceutic experiments. Based on variety of in vitro evidence, a main assumption of the model is that ritonavir does not form a crystalline precipitate while resident in the gastrointestinal tract. In the model, if simulated luminal concentration exceeds the amorphous solubility limit, then precipitation to an amorphous form is immediate. Simulated and observed Cmax and AUC0-t parameters were well captured (within 1.5-fold) for both fasted and fed states in healthy volunteers. By accounting for luminal fluid viscosity differences in the different prandial states (affecting drug diffusivity) as well as the effect of drug free fraction on gut wall permeation rates, it was possible to explain the negative food effect observed for Norvir tablets in humans. In summary, a biopharmaceutic in vitro in vivo extrapolation approach provides confidence in (verification of) key input parameters of the physiologically-based pharmacokinetic ritonavir model which resulted in successful simulation of observed plasma profiles.

Prediction of drug clearance by glucuronidation from in vitro data: use of combined cytochrome P450 and UDP-glucuronosyltransferase cofactors in alamethicin-activated human liver microsomes.

Glucuronidation via UDP-glucuronosyltransferase (UGT) is an increasingly important clearance pathway. In this study intrinsic clearance (CL(int)) values for buprenorphine, carvedilol, codeine, diclofenac, gemfibrozil, ketoprofen, midazolam, naloxone, raloxifene, and zidovudine were determined in pooled human liver microsomes using the substrate depletion approach. The in vitro clearance data indicated a varying contribution of glucuronidation to the clearance of the compounds studied, ranging from 6 to 79% for midazolam and gemfibrozil, respectively. The CL(int) was obtained using either individual or combined cofactors for cytochrome P450 (P450) and UGT enzymes with alamethicin activation and in the presence and absence of 2% bovine serum albumin (BSA). In the presence of combined P450 and UGT cofactors, CL(int) ranged from 2.8 to 688 microl/min/mg for zidovudine and buprenorphine, respectively; the clearance was approximately equal to the sum of the CL(int) values obtained in the presence of individual cofactors. The unbound intrinsic clearance (CL(int, u)) was scaled to provide an in vivo predicted CL(int); the data obtained in the presence of combined cofactors resulted in 5-fold underprediction on average. Addition of 2% BSA to the incubation with both P450 and UGT cofactors reduced the bias in the clearance prediction, with 8 of 10 compounds predicted within 2-fold of in vivo values with the exception of raloxifene and gemfibrozil. The current study indicates the applicability of combined cofactor conditions in the assessment of clearance for compounds with a differential contribution of P450 and UGT enzymes to their elimination. In addition, improved predictability of microsomal data is observed in the presence of BSA, in particular for UGT2B7 substrates.

Hepatocellular binding of drugs: correction for unbound fraction in hepatocyte incubations using microsomal binding or drug lipophilicity data.

Analogous to the fraction unbound in microsomes (fu(mic)), fraction unbound in hepatocyte incubations (fu(hep)) is an important parameter in the prediction of intrinsic clearance and potential drug-drug interactions. A recent study by Austin et al. (Drug Metab Dispos 33:419-425, 2005) proposed a linear 1:1 relationship between the extent of binding to microsomes at 1 mg/ml and to hepatocytes at 10(6) million cells/ml. The current study collates a fu(mic) and fu(hep) database for 39 drugs to examine the relationship between binding in microsomes and hepatocytes. A new nonlinear empirical equation is proposed as an alternative to the linear relationship to relate binding between the two systems. The nonlinear equation results in higher prediction accuracy and lower bias in comparison to the linear relationship, in particular for drugs with fu(hep) < 0.4. The proposed equation is further extended to allow direct prediction of fu(hep) from drug lipophilicity data by substituting the fu(mic) term by the Hallifax and Houston predictive equation (Drug Metab Dispos 34:724-726, 2006). The prediction accuracy of this approach is high for relatively hydrophilic drugs (logP/D < or = 2.5), whereas less accurate prediction of the fu(hep) is observed for lipophilic drugs (logP > 3), consistent with the limitations observed for microsomal binding predictive tools. In conclusion, the proposed nonlinear equations provide an accurate predictive tool to estimate fu(hep) for the in vitro-in vivo extrapolation of intrinsic clearance and inhibition parameters.

Drug lipophilicity and microsomal protein concentration as determinants in the prediction of the fraction unbound in microsomal incubations.

Two predictive tools have been proposed by Austin et al. (Drug Metab Dispos 30:1497-1503, 2002) and Hallifax and Houston (Drug Metab Dispos 34:724-726, 2006) to estimate the fraction unbound in the incubation (fu(inc)). The current study was undertaken to elucidate the relative utility of these prediction tools over a range of drug lipophilicity and microsomal protein concentration. The fu(inc) data set (n = 127) comprised 35 drugs determined experimentally in this study and 92 collated from Austin and Hallifax data. The observed fu(inc) values at three microsomal concentrations were compared with the estimates obtained using the Austin and Hallifax equations. In addition, the impact of variability in the logP on the fu(inc) predictions was assessed. The current analysis highlights the importance of accurate estimation of lipophilicity for the prediction of the fu(inc), regardless of the prediction equation used. Both equations represent useful tools for estimation of fu(inc) for low lipophilicity drugs (logP/D = 0-3), especially at low microsomal protein concentration. However, the accuracy of fu(inc) predictions of highly lipophilic drugs was poor for both equations, implying that fu(inc) should be experimentally confirmed for drugs with logP/D >or= 3, unless the microsomal protein concentration is as low as 0.1 mg/ml, in which case a cutoff of logP/D >or= 5 can be applied. A significant difference in the predictions by the two proposed tools was observed in the area of intermediate lipophilicity (logP/D = 2.5-5), where the Hallifax equation provided more accurate fu(inc) predictions on average, irrespective of the microsomal protein concentration investigated.