A comparative evaluation of models to predict human intestinal metabolism from nonclinical data

Beyond CL and VSS: A comprehensive approach to human pharmacokinetic predictions

This article presents a comprehensive examination of processes related to the prediction of human pharmacokinetics (PK), a crucial task of clinical drug candidate selection. By systematically incorporating in vitro absorption, distribution, metabolism and excretion (ADME) and in vivo PK data with expert judgement, the study achieves high-quality human PK predictions for 40 orally administered compounds from Boehringer Ingelheim's new chemical entity (NCE) portfolio. Overall, the article provides a detailed evaluation of and guidance for a structured process to predict full concentration-time profiles beyond single-parameter predictions, using state-of-the-art methodologies. Furthermore, it discusses future challenges and improvements, and aims to provide valuable insights for scientists working in drug metabolism and PK (DMPK) or PK/pharmacodynamics (PK/PD) modelling.

Quantitative prediction of CYP3A induction-mediated drug-drug interactions in clinical practice

There have been no reports on the quantitative prediction of CYP3A induction-mediated decreases in AUC and Cmax for drug candidates identified as a "victims" of CYP3A induction. Our previous study separately evaluated the fold-induction of hepatic and intestinal CYP3A by known inducers using clinical induction data and revealed that we were able to quantitatively predict the AUC ratio (AUCR) of a few CYP3A substrates in the presence and absence of CYP3A inducers. In the present study, we investigate the predictability of AUCR and also Cmax ratio (CmaxR) in additional 54 clinical studies. The fraction metabolized by CYP3A (fm), the intestinal bioavailability (Fg), and the hepatic intrinsic clearance (CLint) of substrates were determined by the in vitro experiments as well as clinical data used for calculating AUCR and CmaxR. The result showed that 65-69% and 65-67% of predictions were within 2-fold of observed AUCR and CmaxR, respectively. A simulation using multiple parameter combinations suggested that the variability of fm and Fg within a certain range might have a minimal impact on the calculation output. These findings suggest that clinical AUCR and CmaxR of CYP3A substrates can be quantitatively predicted from the preclinical stage.

Clinical Trial Data-Driven Risk Assessment of Drug-Drug Interactions: A Rapid and Accurate Decision-Making Tool

BACKGROUND

In clinical practice, the vast array of potential drug combinations necessitates swift and accurate assessments of pharmacokinetic drug-drug interactions (DDIs), along with recommendations for adjustments. Current methodologies for clinical DDI evaluations primarily rely on basic extrapolations from clinical trial data. However, these methods are limited in accuracy owing to their lack of a comprehensive consideration of various critical factors, including the inhibitory potency, dosage, and type of the inhibitor, as well as the metabolic fraction and intestinal availability of the substrate.

OBJECTIVE

This study aims to propose an efficient and accurate clinical pharmacokinetic-mediated DDI assessment tool, which comprehensively considers the effects of inhibitory potency and dosage of inhibitors, intestinal availability and fraction metabolized of substrates on DDI outcomes.

METHODS

This study focuses on DDIs caused by cytochrome P450 3A4 enzyme inhibition, utilizing extensive clinical trial data to establish a methodology to calculate the metabolic fraction and intestinal availability for substrates, as well as the concentration and inhibitory potency for inhibitors ( K i ork inact / K I ). These parameters were then used to predict the outcomes of DDIs involving 33 substrates and 20 inhibitors. We also defined the risk index for substrates and the potency index for inhibitors to establish a clinical DDI risk scale. The training set for parameter calculation consisted of 73 clinical trials. The validation set comprised 89 clinical DDI trials involving 53 drugs. which was used to evaluate the reliability of in vivo values of K i andk inact / K I , the accuracy of DDI predictions, and the false-negative rate of risk scale.

RESULTS

First, the reliability of the in vivo K i andk inact / K I values calculated in this study was assessed using a basic static model. Compared with values obtained from other methods, this study values showed a lower geometric mean fold error and root mean square error. Additionally, incorporating these values into the physiologically based pharmacokinetic-DDI model facilitated a good fitting of the C-t curves when the substrate's metabolic enzymes are inhibited. Second, area under the curve ratio predictions of studied drugs were within a 1.5 × margin of error in 81% of cases compared with clinical observations in the validation set. Last, the clinical DDI risk scale developed in this study predicted the actual risks in the validation set with only a 5.6% incidence of serious false negatives.

CONCLUSIONS

This study offers a rapid and accurate approach for assessing the risk of pharmacokinetic-mediated DDIs in clinical practice, providing a foundation for rational combination drug use and dosage adjustments.

Liver-on-chips for drug discovery and development

Recent FDA modernization act 2.0 has led to increasing industrial R&D investment in advanced in vitro 3D models such as organoids, spheroids, organ-on-chips, 3D bioprinting, and in silico approaches. Liver-related advanced in vitro models remain the prime area of interest, as liver plays a central role in drug clearance of compounds. Growing evidence indicates the importance of recapitulating the overall liver microenvironment to enhance hepatocyte maturity and culture longevity using liver-on-chips (LoC) in vitro. Hence, pharmaceutical industries have started exploring LoC assays in the two of the most challenging areas: accurate in vitro-in vivo extrapolation (IVIVE) of hepatic drug clearance and drug-induced liver injury. We examine the joint efforts of commercial chip manufacturers and pharmaceutical companies to present an up-to-date overview of the adoption of LoC technology in the drug discovery. Further, several roadblocks are identified to the rapid adoption of LoC assays in the current drug development framework. Finally, we discuss some of the underexplored application areas of LoC models, where conventional 2D hepatic models are deemed unsuitable. These include clearance prediction of metabolically stable compounds, immune-mediated drug-induced liver injury (DILI) predictions, bioavailability prediction with gut-liver systems, hepatic clearance prediction of drugs given during pregnancy, and dose adjustment studies in disease conditions. We conclude the review by discussing the importance of PBPK modeling with LoC, digital twins, and AI/ML integration with LoC.

Predicting the in vivo developmental toxicity of fenarimol from in vitro toxicity data using PBTK modelling-facilitated reverse dosimetry approach

In vitro methods are widely used in modern toxicological testing; however, the data cannot be directly employed for risk assessment. In vivo toxicity of chemicals can be predicted from in vitro data using physiologically based toxicokinetic (PBTK) modelling-facilitated reverse dosimetry (PBTK-RD). In this study, a minimal-PBTK model was constructed to predict the in-vivo kinetic profile of fenarimol (FNL) in rats and humans. The model was verified by comparing the observed and predicted pharmacokinetics of FNL for rats (calibrator) and further applied to humans. Using the PBTK-RD approach, the reported in vitro developmental toxicity data for FNL was translated to in vivo dose-response data to predict the assay equivalent oral dose in rats and humans. The predicted assay equivalent rat oral dose (36.46 mg/kg) was comparable to the literature reported in vivo BMD10 value (22.8 mg/kg). The model was also employed to derive the chemical-specific adjustment factor (CSAF) for interspecies toxicokinetics variability of FNL. Further, Monte Carlo simulations were performed to predict the population variability in the plasma concentration of FNL and to derive CSAF for intersubject human kinetic differences. The comparison of CSAF values for interspecies and intersubject toxicokinetic variability with their respective default values revealed that the applied uncertainty factors were adequately protective.

Novel Psilocin Prodrugs with Altered Pharmacological Properties as Candidate Therapies for Treatment-Resistant Anxiety Disorders

The psychedelic prodrug psilocybin has shown therapeutic benefits for the treatment of numerous psychiatric conditions. Despite positive clinical end points targeting depression and anxiety, concerns regarding the duration of the psychedelic experience produced by psilocybin, associated with enduring systemic exposure to the active metabolite psilocin, pose a barrier to its therapeutic application. Our objective was to create a novel prodrug of psilocin with similar therapeutic benefits but a reduced duration of psychedelic effects compared with psilocybin. Here, we report the synthesis and functional screening of 28 new chemical entities. Our strategy was to introduce a diversity of cleavable groups at the 4-hydroxy position of the core indole moiety to modulate metabolic processing. We identified several novel prodrugs of psilocin with altered pharmacokinetic profiles and reduced pharmacological exposure compared with psilocybin. These candidate prodrugs have the potential to maintain the long-term benefits of psilocybin therapy while attenuating the duration of psychedelic effects.

Physiologically based modelling of dermal absorption and kinetics of consumer-relevant chemicals: A case study with exposure to bisphenol A from thermal paper

Bisphenol A (BPA) is one of the best studied industrial chemicals in terms of exposure, toxicity, and toxicokinetics. This renders it an ideal candidate to exploit the recent advancements in physiologically based pharmacokinetic (PBPK) modelling to support risk assessment of BPA specifically, and of other consumer-relevant hazardous chemicals in general. Using the exposure from thermal paper as a case scenario, this study employed the multi-phase multi-layer mechanistic dermal absorption (MPML MechDermA) model available in the Simcyp® Simulator to simulate the dermal toxicokinetics of BPA at local and systemic levels. Sensitivity analysis helped to identify physicochemical and physiological factors influencing the systemic exposure to BPA. The iterative modelling process was as follows: (i) development of compound files for BPA and its conjugates, (ii) setting-up of a PBPK model for intravenous administration, (iii) extension for oral administration, and (iv) extension for exposure via skin (i.e., hand) contact. A toxicokinetic study involving hand contact to BPA-containing paper was used for model refinement. Cumulative urinary excretion of total BPA had to be employed for dose reconstruction. PBPK model performance was verified using the observed serum BPA concentrations. The predicted distribution across the skin compartments revealed a depot of BPA in the stratum corneum (SC). These findings shed light on the role of the SC to act as temporary reservoir for lipophilic chemicals prior to systemic absorption, which inter alia is relevant for the interpretation of human biomonitoring data and for establishing the relationship between external and internal measures of exposure.

Development and application of a physiologically based pharmacokinetic model for entrectinib in rats and scale-up to humans: Route-dependent gut wall metabolism

Entrectinib (Rozlytrek®) is an oral antineoplastic agent approved by the U.S. Food and Drug Administration in 2019 for the treatment of c-ros oncogene 1 (ROS1)-positive non-small cell lung cancer and neurotrophic tyrosine receptor kinase (NTRK) fusion-positive solid tumors. Although there have been a few studies on the pharmacokinetics of entrectinib, the relative contributions of several kinetic factors determining the oral bioavailability and systemic exposure of entrectinib are still worthy of investigation. Experimental data on the intestinal absorption and disposition of entrectinib in rats were acquired from studies on in vitro protein binding/tissue S9 metabolism, in situ intestinal perfusion, and in vivo dose-escalation/hepatic extraction. Using these datasets, an in-house whole-body physiologically based pharmacokinetic (PBPK) model incorporating the Q_Gut model concepts and segregated blood flow in the gut was constructed and optimized with respect to drug-specific parameters. The established rat PBPK model was further extrapolated to humans through relevant physiological scale-up and parameter optimization processes. The optimized rat and human PBPK models adequately captured the impact of route-dependent gut metabolism on the systemic exposure to entrectinib and closely mirrored various preclinical and clinical observations. Our proposed PBPK model could be useful in optimizing dosage regimens and predicting drug interaction potential in various clinical conditions, after partial modification and validation.

Extrapolation of physiologically based pharmacokinetic model for tacrolimus from renal to liver transplant patients

Physiologically based pharmacokinetic (PBPK) modeling is useful for evaluating differences in drug exposure among special populations, but it has not yet been employed to evaluate the absorption process of tacrolimus. In this study, we developed a minimal PBPK model with a compartmental absorption and transit model for renal transplant patients using available data in the literature and clinical data from our hospital. The effective permeability value of tacrolimus absorption and parameters for the single adjusting compartment were optimized via sensitivity analyses, generating a PBPK model of tacrolimus for renal transplant patients with good predictability. Next, we extrapolated the pharmacokinetics of tacrolimus for liver transplant patients by changing the population demographic parameters of the model. When the physiological parameters of a population with normal liver function were changed to those of a population with impaired hepatic function (Child-Pugh class A) in the constructed renal transplant PBPK model, the predicted tacrolimus concentrations were consistent with the observed concentrations in liver transplant patients. In conclusion, the constructed tacrolimus PBPK model for renal transplant patients could predict the pharmacokinetics in liver transplant patients by slightly reducing the hepatic function, even at three weeks post-transplantation.

Reduced physiologically-based pharmacokinetic model of dabigatran etexilate-dabigatran and its application for prediction of intestinal P-gp-mediated drug-drug interactions

BACKGROUND

Dabigatran etexilate (DABE) has been suggested as a clinical probe for intestinal P-glycoprotein (P-gp)-mediated drug-drug interaction (DDI) studies and, as an alternative to digoxin. Clinical DDI data with various P-gp inhibitors demonstrated a dose-dependent inhibition of P-gp with DABE. The aims of this study were to develop a joint DABE (prodrug)-dabigatran reduced physiologically-based-pharmacokinetic (PBPK) model and to evaluate its ability to predict differences in P-gp DDI magnitude between a microdose and a therapeutic dose of DABE.

METHODS

A joint DABE-dabigatran PBPK model was developed with a mechanistic intestinal model accounting for the regional P-gp distribution in the gastrointestinal tract. Model input parameters were estimated using DABE and dabigatran pharmacokinetic (PK) clinical data obtained after administration of DABE alone or with a strong P-gp inhibitor, itraconazole, and over a wide range of DABE doses (from 375 µg to 400 mg). Subsequently, the model was used to predict extent of DDI with additional P-gp inhibitors and with different DABE doses.

RESULTS

The reduced DABE-dabigatran PBPK model successfully described plasma concentrations of both prodrug and metabolite following administration of DABE at different dose levels and when co-administered with itraconazole. The model was able to capture the dose dependency in P-gp mediated DDI. Predicted magnitude of itraconazole P-gp DDI was higher at the microdose (predicted vs. observed median fold-increase in AUC_+inh/AUC_control (min-max) = 5.88 (4.29-7.93) vs. 6.92 (4.96-9.66) ) compared to the therapeutic dose (predicted median fold-increase in AUC_+inh/AUC_control = 3.48 (2.37-4.84) ). In addition, the reduced DABE-dabigatran PBPK model predicted successfully the extent of DDI with verapamil and clarithromycin as P-gp inhibitors. Model-based simulations of dose staggering predicted the maximum inhibition of P-gp when DABE microdose was concomitantly administered with itraconazole solution; simulations also highlighted dosing intervals required to minimise the DDI risk depending on the DABE dose administered (microdose vs. therapeutic).

CONCLUSIONS

This study provides a modelling framework for the evaluation of P-gp inhibitory potential of new molecular entities using DABE as a clinical probe. Simulations of dose staggering and regional differences in the extent of intestinal P-gp inhibition for DABE microdose and therapeutic dose provide model-based guidance for design of prospective clinical P-gp DDI studies.

A New Intestinal Model for Analysis of Drug Absorption and Interactions Considering Physiological Translocation of Contents

Precise prediction of drug absorption is key to the success of new drug development and efficacious pharmacotherapy. In this study, we developed a new absorption model, the advanced translocation model (ATOM), by extending our previous model, the translocation model. ATOM reproduces the translocation of a substance in the intestinal lumen using a partial differential equation with variable dispersion and convection terms to describe natural flow and micromixing within the intestine under not only fasted but also fed conditions. In comparison with ATOM, it was suggested that a conventional absorption model, advanced compartmental absorption and transit model, tends to underestimate micromixing in the upper intestine, and it is difficult to adequately describe movements under the fasted and fed conditions. ATOM explains the observed nonlinear absorption of midazolam successfully, with a minimal number of scaling factors. Furthermore, ATOM considers the apical and basolateral membrane permeabilities of enterocytes separately and assumes compartmentation of the lamina propria, including blood vessels, to consider intestinal blood flow appropriately. ATOM estimates changes in the intestinal availability caused by drug interaction associated with inhibition of CYP3A and P-glycoprotein in the intestine. Additionally, ATOM can estimate the drug absorption in the fed state considering delayed intestinal drug flow. Therefore, ATOM is a useful tool for the analysis of local pharmacokinetics in the gastrointestinal tract, especially for the estimation of nonlinear drug absorption, which may involve various interactions with intestinal contents or other drugs. SIGNIFICANCE STATEMENT: The newly developed advanced translocation model precisely explains various movements of intestinal contents under fasted and fed conditions, which cannot be adequately described by the current physiological pharmacokinetic models.

Current EU regulatory requirements for the assessment of chemicals and cosmetic products: challenges and opportunities for introducing new approach methodologies

The EU Directive 2010/63/EU   on the protection of animals used for scientific purposes and other EU regulations, such as REACH and the Cosmetic Products Regulation advocate for a change in the way toxicity testing is conducted. Whilst the Cosmetic Products Regulation bans animal testing altogether, REACH aims for a progressive shift from in vivo testing towards quantitative in vitro and computational approaches. Several endpoints can already be addressed using non-animal approaches including skin corrosion and irritation, serious eye damage and irritation, skin sensitisation, and mutagenicity and genotoxicity. However, for systemic effects such as acute toxicity, repeated dose toxicity and reproductive and developmental toxicity, evaluation of chemicals under REACH still heavily relies on animal tests. Here we summarise current EU regulatory requirements for the human health assessment of chemicals under REACH and the Cosmetic Products Regulation, considering the more critical endpoints and identifying the main challenges in introducing alternative methods into regulatory testing practice. This supports a recent initiative taken by the International Cooperation on Alternative Test Methods (ICATM) to summarise current regulatory requirements specific for the assessment of chemicals and cosmetic products for several human health-related endpoints, with the aim of comparing different jurisdictions and coordinating the promotion and ultimately the implementation of non-animal approaches worldwide. Recent initiatives undertaken at European level to promote the 3Rs and the use of alternative methods in current regulatory practice are also discussed.

Pharmacokinetic Drug Interaction between Tofacitinib and Voriconazole in Rats

Fungal infections are prevalent in patients with immune diseases. Voriconazole, a triazole antifungal drug, inhibits the cytochromes CYP3A4 and CYP2C, and tofacitinib, a Janus kinase inhibitor for the treatment of rheumatoid arthritis, is metabolized by CYP3A4 and CYP2C19 in humans. Here, we investigated their interaction during simultaneous administration of both drugs to rats, either intravenously or orally. The area under the plasma concentration–time curve from time zero to time infinity (AUC) of tofacitinib was significantly greater, by 166% and 171%, respectively, and the time-averaged non-renal clearance (CL_NR) of tofacitinib was significantly slower (59.5%) than that for tofacitinib alone. An in vitro metabolism study showed non-competitive inhibition of tofacitinib metabolism in the liver and intestine by voriconazole. The concentration/apparent inhibition constant (K_i) ratios of voriconazole were greater than two, indicating that the inhibition of tofacitinib metabolism could be due to the inhibition of the CYP3A1/2 and CYP2C11 enzymes by voriconazole. The pharmacokinetics of voriconazole were not affected by the co-administration of tofacitinib. In conclusion, the significantly greater AUC and slower CL_NR of tofacitinib after intravenous and oral administration of both drugs were attributable to the non-competitive inhibition of tofacitinib metabolism via CYP3A1/2 and CYP2C11 by voriconazole in rats.

Interaction between Omeprazole and Gliclazide in Relation to CYP2C19 Phenotype

The antidiabetic drug gliclazide is partly metabolized by CYP2C19, the main enzyme involved in omeprazole metabolism. The aim of the study was to explore the interaction between omeprazole and gliclazide in relation to CYP2C19 phenotype using physiologically based pharmacokinetic (PBPK) modeling approach. Developed PBPK models were verified using in vivo pharmacokinetic profiles obtained from a clinical trial on omeprazole-gliclazide interaction in healthy volunteers, CYP2C19 normal/rapid/ultrarapid metabolizers (NM/RM/UM). In addition, the association of omeprazole cotreatment with gliclazide-induced hypoglycemia was explored in 267 patients with type 2 diabetes (T2D) from the GoDARTS cohort, Scotland. The PBPK simulations predicted 1.4–1.6-fold higher gliclazide area under the curve (AUC) after 5-day treatment with 20 mg omeprazole in all CYP2C19 phenotype groups except in poor metabolizers. The predicted gliclazide AUC increased 2.1 and 2.5-fold in intermediate metabolizers, and 2.6- and 3.8-fold in NM/RM/UM group, after simulated 20-day dosing with 40 mg omeprazole once and twice daily, respectively. The predicted results were corroborated by findings in patients with T2D which demonstrated 3.3-fold higher odds of severe gliclazide-induced hypoglycemia in NM/RM/UM patients concomitantly treated with omeprazole. Our results indicate that omeprazole may increase exposure to gliclazide and thus increase the risk of gliclazide-associated hypoglycemia in the majority of patients.

Factors and dosage formulations affecting the solubility and bioavailability of P-glycoprotein substrate drugs

Introduction: Expression of P-glycoprotein (P-gp) increases toward the distal small intestine, implying that the duodenum is the preferential absorption site for P-gp substrate drugs. Oral bioavailability of poorly soluble P-gp substrate drugs is low and varied but increases with high-fat meals that supply lipoidal components and bile in the duodenum.Areas covered: Absorption properties of P-gp substrate drugs along with factors and oral dosage formulations affecting their solubility and bioavailability were reviewed with PubMed literature searches. An overview is provided from the viewpoint of the 'spring-and-parachute approach' that generates supersaturation of poorly soluble P-gp substrate drugs.Expert opinion: The oral bioavailability of P-gp substrate drugs is difficult to predict because of their low solubility, preferential absorption sites, and overlapping substrate specificities with CYP3A4, along with the scattered intestinal P-gp expression/function. To attain high and steady oral bioavailability of poorly soluble P-gp substrate drugs, physicochemical modification of drugs to improve solubility, or oral dosage formulations that generate long-lasting supersaturation in the duodenum, is preferred. In particular, supersaturable lipid-based drug delivery systems that can increase passive diffusion and/or lymphatic absorption are effective and applicable to many poorly soluble P-gp substrate drugs.

Prediction of Drug Clearance from Enzyme and Transporter Kinetics

Accurate estimation of in vivo clearance in human is pivotal to determine the dose and dosing regimen for drug development. In vitro-in vivo extrapolation (IVIVE) has been performed to predict drug clearance using empirical and physiological scalars. Multiple in vitro systems and mathematical modeling techniques have been employed to estimate in vivo clearance. The models for predicting clearance have significantly improved and have evolved to become more complex by integrating multiple processes such as drug metabolism and transport as well as passive diffusion. This chapter covers the use of conventional as well as recently developed methods to predict metabolic and transporter-mediated clearance along with the advantages and disadvantages of using these methods and the associated experimental considerations. The general approaches to improve IVIVE by use of appropriate scalars, incorporation of extrahepatic metabolism and transport and application of physiologically based pharmacokinetic (PBPK) models with proteomics data are also discussed. The chapter also provides an overview of the advantages of using such dynamic mechanistic models over static models for clearance predictions to improve IVIVE.

Evaluation of In Vitro Models for Assessment of Human Intestinal Metabolism in Drug Discovery

Although intestinal metabolism plays an important role in drug disposition, early predictions of human outcomes are challenging, in part because of limitations of available in vitro models. To address this, we have evaluated three in vitro models of human intestine (microsomes, permeabilized enterocytes, and cryopreserved intestinal mucosal epithelium) as tools to assess intestinal metabolism and estimate the fraction escaping gut metabolism (f g) in drug discovery. The models were tested with a chemically diverse set of 32 compounds, including substrates for oxidoreductive, hydrolytic, and conjugative enzymes. Liquid chromatography-high-resolution mass spectrometry was used to quantify substrate disappearance [intrinsic clearance (CLint)] and qualify metabolite formation (quantitative-qualitative bioanalysis). Fraction unbound in the incubation (f u,inc) was determined by rapid equilibrium dialysis. Measured in vitro results (CLint and f u,inc) were supplemented with literature data [passive Caco-2 apical to basolateral permeability, enterocyte blood flow, and intestinal surface area (A)] and combined using a midazolam-calibrated Q gut model to predict human f g values. All three models showed reliable CYP and UDP-glucuronosyltransferase activities, but enterocytes and mucosa may offer advantages for low-clearance compounds and alternative pathways (e.g., sulfation, hydrolases, and flavin-containing monooxigenases). Early predictions of human f g values were acceptable for the high-f g compounds (arbitrarily f g > 0.7). However, predictions of low- and moderate-f g values (arbitrarily f g < 0.7) remain challenging, indicating that further evaluation is needed (e.g., saturation effects and impact of transporters) but not immediate compound avoidance. Results suggest that tested models offer an additional value in drug discovery, especially for drug design and chemotype evaluation. SIGNIFICANCE STATEMENT: We found that cellular models of the human gut (permeabilized enterocytes and cryopreserved intestinal mucosa) offer an alternative to and potential advantage over intestinal microsomes in studies of drug metabolism, particularly for low-clearance compounds and alternative pathways (e.g., sulfation, hydrolases, and flavin-containing monooxigenases). The predictivity of human fraction escaping gut metabolism for common CYP and UDP-glucuronosyltransferase substrates based on the Q gut model is still limited, however, and appropriate further evaluation is recommended.

Direct Comparison of Total Clearance Prediction: Computational Machine Learning Model versus Bottom-Up Approach Using In Vitro Assay

The in vitro-in vivo extrapolation (IVIVE) approach for predicting total plasma clearance (CL_tot) has been widely used to rank order compounds early in discovery. More recently, a computational machine learning approach utilizing physicochemical descriptors and fingerprints calculated from chemical structure information has emerged, enabling virtual predictions even earlier in discovery. Previously, this approach focused more on in vitro intrinsic clearance (CL_int) prediction. Herein, we directly compare these two approaches for predicting CL_tot in rats. A structurally diverse set of 1114 compounds with known in vivo CL_tot, in vitro CL_int, and plasma protein binding was used as the basis for this evaluation. The machine learning models were assessed by validation approaches using the time- and cluster-split training and test sets, and five-fold cross validation. Assessed by five-fold validation, the random forest regression (RF) and radial basis function (RBF) models demonstrated better prediction performance in eight attempted machine learning models. The CL_tot values predicted by the RF and RBF models were within two-fold of the observed values for 67.7 and 71.9% of cluster-split test set compounds, respectively, while the predictivity was worse in the time-split dataset. The predictivity of both models tended to be improved by incorporating in vitro parameters, unbound fraction in plasma (f_u,p), and CL_int. CL_tot prediction utilizing in vitro CL_int and the well-stirred model, correcting for the fraction unbound in blood, was substantially worse compared to machine learning approaches for the same cluster-split test set. The reason that CL_tot is underestimated by IVIVE is not fully explained by considering the calculated microsomal unbound fraction (cf_u,mic), extended clearance classification system (ECCS), and omitting high clearance compounds in excess of hepatic blood flow. The analysis suggests that in silico machine learning models may have the power to reduce reliance on or replace in vitro and in vivo studies for chemical structure optimization in early drug discovery.

Absorption and Intestinal Metabolic Profile of Oleocanthal in Rats

Oleocanthal (OLC), a phenolic compound of extra virgin olive oil (EVOO), has emerged as a potential therapeutic agent against a variety of diseases due to its anti-inflammatory activity. The aim of the present study is to explore its in vivo intestinal absorption and metabolism. An in situ perfusion technique in rats was used, involving simultaneous sampling from the luminal perfusate and mesenteric blood. Samples were analysed by UHPLC-MS-MS for the presence of oleocanthal (OLC) and its metabolites. OLC was mostly metabolized by phase I metabolism, undergoing hydration, hydrogenation and hydroxylation. Phase II reactions (glucuronidation of hydrogenated OLC and hydrated metabolites) were observed in plasma samples. OLC was poorly absorbed in the intestine, as indicated by the low effective permeability coefficient (2.23 ± 3.16 × 10-5 cm/s) and apparent permeability coefficient (4.12 ± 2.33 × 10-6 cm/s) obtained relative to the values of the highly permeable reference compound levofloxacin (LEV). The extent of OLC absorption reflected by the area under the mesenteric blood-time curve normalized by the inlet concentration (AUC) was also lower than that of LEV (0.25 ± 0.04 vs. 0.64 ± 0.03, respectively). These results, together with the observed intestinal metabolism, suggest that OLC has a moderate-to-low oral absorption; but higher levels of OLC are expected to reach human plasma vs. rat plasma.

Prediction of active human dose: learnings from 20 years of Merck KGaA experience, illustrated by case studies

High-quality dose predictions based on a good understanding of target engagement is one of the main translational goals in drug development. Here, we systematically evaluate active human dose predictions for 15 Merck KGaA/EMD Serono assets spanning several modalities and therapeutic areas. Using case studies, we illustrate the value of adhering to the translational best practices of having an exposure-response relationship in an appropriate animal model; having validated, translatable pharmacodynamic (PD) biomarkers measurable in Phase I populations in the right tissue; having a deeper understanding of biology; and capturing uncertainties in predictions. Given the gap in publications on the subject, we believe that the learnings from this unique diverse data set, which are generic to the industry, will trigger actions to improve future predictions.

A novel Css-MRTpo approach to simulate oral plasma concentration-time profiles of the partial glucokinase activator PF-04937319 and its disproportionate N-demethylated metabolite in humans using chimeric mice with humanized livers

A Css-MRTpo superposition method was devised to predict (retrospectively) oral plasma concentration-time profiles of PF-04937319 and its MIST-related metabolite, M1, in humans using chimeric mice with humanized liver.Original PK data were taken from a published report in which PF-04937319 and M1 were given to chimeric mice orally and/or intravenously. Human CL and Vss were predicted by single-species allometry and MRTiv,pred were calculated as Vss,pred/CL,pred. MRTpo,human were assumed to be MRTiv,pred plus MAT or mean metabolite formation time (MFT). Human Css was calculated by dividing the corrected oral dose by Vss,pred.Chronological sampling time and measured plasma concentrations were corrected by MRTpo,human and Css,human, respectively, and transformed to the corresponding values in humans.The obtained concentration-time profile of PF-04937319 was superimposed well with the observed data after single and repeated oral administration to humans. The transformed plasma concentration of M1 was somewhat lower than the observed value, but a slow increase of the simulated metabolite reflected gradual increase of observed M1 on Day 1. Transformed M1 gave an almost-flat concentration-time profile on Day 14, which was consistent with the curve observed in humans. Application of this novel method to other MIST-related compounds is discussed.

Requirements to Establishing Confidence in Physiologically Based Pharmacokinetic (PBPK) Models and Overcoming Some of the Challenges to Meeting Them

When scientifically well-founded, the mechanistic basis of physiologically based pharmacokinetic (PBPK) models can help reduce the uncertainty and increase confidence in extrapolations outside the studied scenarios or studied populations. However, it is not always possible to establish mechanistically credible PBPK models. Requirements to establishing confidence in PBPK models, and challenges to meeting these requirements, are presented in this article. Parameter non-identifiability is the most challenging among the barriers to establishing confidence in PBPK models. Using case examples of small molecule drugs, this article examines the use of hypothesis testing to overcome parameter non-identifiability issues, with the objective of enhancing confidence in the mechanistic basis of PBPK models and thereby improving the quality of predictions that are meant for internal decisions and regulatory submissions. When the mechanistic basis of a PBPK model cannot be established, we propose the use of simpler models or evidence-based approaches.

Evaluating the clinical importance of bacterial degradation of therapeutic agents in the lower intestine of adults using adult fecal material

PURPOSE

Optimize adult fecal material composition for evaluating the clinical importance of bacterial degradation of therapeutic agents in the lower intestine (distal small intestine, D-SI and proximal colon, P-COL). Evaluate the usefulness of optimized fecal material in the evaluation of bacterial degradation of five model highly permeable drugs: two nitroreductase substrates (nitrendipine and nimodipine), three drugs for which published data indicate no impact of bacterial degradation on in vivo performance (levodopa, budesonide and rivaroxaban) and one prodrug (sulfasalazine, an azoreductase substrate) from which a locally acting on the mucosa of the lower intestine drug is derived (mesalamine).

METHODS

30 min and 95 min were used as point estimates of maximum bacterial degradation half-lives for bacterial degradation in D-SI or in P-COL, respectively, to be clinically important, i.e. for at least 20% reduction in absorption from D-SI or P-COL to occur. Optimization of fecal material was based on recently reported degradation profiles of metronidazole (a nitroreductase substrate) and olsalazine (an azoreductase substrate) in the lower intestine of healthy adults which are clinically important. Model compounds were tested in optimized fecal materials and data were evaluated vs. existing in vivo data in adults.

RESULTS

Simulated ileal bacteria (SIB) consisted of 5.5% (w/v) stools in normal saline and simulated colonic bacteria (SCoB) consisted of 8.3% (w/v) stools in normal saline. For all model compounds, data in SIB and SCoB were in line with available information in adults. [Degradation half-life in SIB/Degradation half-life in SCoB] ≈ [Stool content in SCoB/Stool content in SIB] ≈ 1.5, i.e. bacterial degradation in SIB could be predicted from bacterial degradation in SCoB.

CONCLUSION

Data in SCoB only are useful for evaluating whether bacterial degradation in P-COL and in D-SI is likely to be clinically important for orally administered, highly permeable drugs or prodrugs which act locally after bacterial degradation. The usefulness of this approach in cases where enzymes other than nitroreductases or azoreductases are involved requires further confirmation.

Improved Predictions of Drug-Drug Interactions Mediated by Time-Dependent Inhibition of CYP3A

Time-dependent inactivation (TDI) of cytochrome P450s (CYPs) is a leading cause of clinical drug-drug interactions (DDIs). Current methods tend to overpredict DDIs. In this study, a numerical approach was used to model complex CYP3A TDI in human-liver microsomes. The inhibitors evaluated included troleandomycin (TAO), erythromycin (ERY), verapamil (VER), and diltiazem (DTZ) along with the primary metabolites N-demethyl erythromycin (NDE), norverapamil (NV), and N-desmethyl diltiazem (NDD). The complexities incorporated into the models included multiple-binding kinetics, quasi-irreversible inactivation, sequential metabolism, inhibitor depletion, and membrane partitioning. The resulting inactivation parameters were incorporated into static in vitro-in vivo correlation (IVIVC) models to predict clinical DDIs. For 77 clinically observed DDIs, with a hepatic-CYP3A-synthesis-rate constant of 0.000 146 min-1, the average fold difference between the observed and predicted DDIs was 3.17 for the standard replot method and 1.45 for the numerical method. Similar results were obtained using a synthesis-rate constant of 0.000 32 min-1. These results suggest that numerical methods can successfully model complex in vitro TDI kinetics and that the resulting DDI predictions are more accurate than those obtained with the standard replot approach.

A comparative evaluation of models to predict human intestinal metabolism from nonclinical data

Extensive gut metabolism is often associated with the risk of low and variable bioavailability. The prediction of the fraction of drug escaping gut wall metabolism as well as transporter-mediated secretion (F_g ) has been challenged by the lack of appropriate preclinical models. The purpose of this study is to compare the performance of models that are widely employed in the pharmaceutical industry today to estimate F_g and, based on the outcome, to provide recommendations for the prediction of human F_g during drug discovery and early drug development. The use of in vitro intrinsic clearance from human liver microsomes (HLM) in three mechanistic models - the ADAM, Q_gut and Competing Rates - was evaluated for drugs whose metabolism is dominated by CYP450s, assuming that the effect of transporters is negligible. The utility of rat as a model for human F_g was also explored. The ADAM, Q_gut and Competing Rates models had comparable prediction success (70%, 74%, 69%, respectively) and bias (AFE = 1.26, 0.74 and 0.81, respectively). However, the ADAM model showed better accuracy compared with the Q_gut and Competing Rates models (RMSE =0.20 vs 0.30 and 0.25, respectively). Rat is not a good model (prediction success =32%, RMSE =0.48 and AFE = 0.44) as it seems systematically to under-predict human F_g . Hence, we would recommend the use of rat to identify the need for F_g assessment, followed by the use of HLM in simple models to predict human F_g . © 2017 Merck KGaA. Biopharmaceutics & Drug Disposition Published by John Wiley & Sons, Ltd.