Evaluation of Generic Methods to Predict Human Pharmacokinetics Using Physiologically Based Pharmacokinetic Model for Early Drug Discovery of Tyrosine Kinase Inhibitors

Identification of SARS-CoV-2 Main Protease Inhibitors Using Chemical Similarity Analysis Combined with Machine Learning

SARS-CoV-2 Main Protease (Mpro) is an enzyme that cleaves viral polyproteins translated from the viral genome, which is critical for viral replication. Mpro is a target for anti-SARS-CoV-2 drug development. Herein, we performed a large-scale virtual screening by comparing multiple structural descriptors of reference molecules with reported anti-coronavirus activity against a library with >17 million compounds. Further filtering, performed by applying two machine learning algorithms, identified eighteen computational hits as anti-SARS-CoV-2 compounds with high structural diversity and drug-like properties. The activities of twelve compounds on Mpro's enzymatic activity were evaluated by fluorescence resonance energy transfer (FRET) assays. Compound 13 (ZINC13878776) significantly inhibited SARS-CoV-2 Mpro activity and was employed as a reference for an experimentally hit expansion. The structural analogues 13a (ZINC4248385), 13b (ZNC13523222), and 13c (ZINC4248365) were tested as Mpro inhibitors, reducing the enzymatic activity of recombinant Mpro with potency as follows: 13c > 13 > 13b > 13a. Then, their anti-SARS-CoV-2 activities were evaluated in plaque reduction assays using Vero CCL81 cells. Subtoxic concentrations of compounds 13a, 13c, and 13b displayed in vitro antiviral activity with IC50 in the mid micromolar range. Compounds 13a-c could become lead compounds for the development of new Mpro inhibitors with improved activity against anti-SARS-CoV-2.

Prediction of human pharmacokinetics from chemical structure: combining mechanistic modeling with machine learning

Pharmacokinetics (PK) is the result of a complex interplay between compound properties and physiology, and a detailed characterization of a molecule's PK during preclinical research is key to understanding the relationship between applied dose, exposure, and pharmacological effect. Predictions of human PK based on the chemical structure of a compound are highly desirable to avoid advancing compounds with unfavorable properties early on and to reduce animal testing, but data to train such models are scarce. To address this problem, we combine well-established physiologically based pharmacokinetic models with Deep Learning models for molecular property prediction into a hybrid model to predict PK parameters for small molecules directly from chemical structure. Our model predicts exposure after oral and intravenous administration with fold change errors of 1.87 and 1.86, respectively, in healthy subjects and 2.32 and 2.23, respectively, in patients with various diseases. Unlike pure Deep Learning models, the hybrid model can predict endpoints on which it was not trained. We validate this extrapolation capability by predicting full concentration-time profiles for compounds with published PK data. Our model enables early selection and prioritization of the most promising drug candidates, which can lead to a reduction in animal testing during drug discovery and development.

Mechanism-Based Pharmacokinetic Model for the Deglycosylation Kinetics of 20(S)-Ginsenosides Rh2

Aim: The 20(S)-ginsenoside Rh2 (Rh2) is being developed as a new antitumor drug. However, to date, little is known about the kinetics of its deglycosylation metabolite (protopanoxadiol) (PPD) following Rh2 administration. The aim of this work was to 1) simultaneously characterise the pharmacokinetics of Rh2 and PPD following intravenous and oral Rh2 administration, 2) develop and validate a mechanism-based pharmacokinetic model to describe the deglycosylation kinetics and 3) predict the percentage of Rh2 entering the systemic circulation in PPD form.

Methods: Plasma samples were collected from rats after the I.V. or P.O. administration of Rh2. The plasma Rh2 and PPD concentrations were determined using HPLC-MS. The transformation from Rh2 to PPD, its absorption, and elimination were integrated into the mechanism based pharmacokinetic model to describe the pharmacokinetics of Rh2 and PPD simultaneously at 10 mg/kg. The concentration data collected following a 20 mg/kg dose of Rh2 was used for model validation.

Results: Following Rh2 administration, PPD exhibited high exposure and atypical double peaks. The model described the abnormal kinetics well and was further validated using external data. A total of 11% of the administered Rh2 was predicted to be transformed into PPD and enter the systemic circulation after I.V. administration, and a total of 20% of Rh2 was predicted to be absorbed into the systemic circulation in PPD form after P.O. administration of Rh2.

Conclusion: The developed model provides a useful tool to quantitatively study the deglycosylation kinetics of Rh2 and thus, provides a valuable resource for future pharmacokinetic studies of glycosides with similar deglycosylation metabolism.

Predicting the Drug-Drug Interaction Mediated by CYP3A4 Inhibition: Method Development and Performance Evaluation

The prediction of drug-drug interactions (DDIs) plays critical roles for the estimation of DDI risk caused by inhibition of CYP3A4. The aim of this paper is to develop a physiologically based pharmacokinetic (PBPK)-DDI model for prediction of the DDI co-administrated with ketoconazole in humans and evaluate the predictive performance of the model. The pharmacokinetic and biopharmaceutical properties of 35 approved drugs, as victims, were collected for the development of a PBPK model, which were linked to the PBPK model of ketoconazole for the DDI prediction. The PBPK model of victims and ketoconazole were validated by matching actual in vivo pharmacokinetic data. The predicted results of DDI were compared with actual data to evaluate the predictive performance. The percentage of predicted ratio of AUC (AUCR), C_max (C_maxR), and T_max (T_maxR) was 75%, 69%, and 91%, respectively, which were within the twofold threshold (range, 0.5-2.0×) of the observed values. Only 3% of the predicted AUCRs are obviously underestimated. After integration of the reported fraction of metabolism (f_m) into the PBPK-DDI model for limited four cases, the model-predicted AUCRs were improved from the twofold range of the observed AUCRs to the 90% confidence interval. The developed method could reasonably predict drug-drug interaction with a low risk of underestimation. The present accuracy of the prediction was improved compared with that of static mechanistic models. The evaluation of predictive performance increases the confidence using the model to evaluate the risk of DDIs co-administrated with ketoconazole before the in vivo DDI study.

Improving the Accuracy of Predicted Human Pharmacokinetics: Lessons Learned from the AstraZeneca Drug Pipeline Over Two Decades

During drug discovery and prior to the first human dose of a novel candidate drug, the pharmacokinetic (PK) behavior of the drug in humans is predicted from preclinical data. This helps to inform the likelihood of achieving therapeutic exposures in early clinical development. Once clinical data are available, the observed human PK are compared with predictions, providing an opportunity to assess and refine prediction methods. Application of best practice in experimental data generation and predictive methodologies, and a focus on robust mechanistic understanding of the candidate drug disposition properties before nomination to clinical development, have led to maximizing the probability of successful PK predictions so that 83% of AstraZeneca drug development projects progress in the clinic with no PK issues; and 71% of key PK parameter predictions [64% of area under the curve (AUC) predictions; 78% of maximum concentration (C_max) predictions; and 70% of half-life predictions] are accurate to within twofold. Here, we discuss methods to predict human PK used by AstraZeneca, how these predictions are assessed and what can be learned from evaluating the predictions for 116 candidate drugs.