Predicting Volume of Distribution in Humans: Performance of In Silico Methods for a Large Set of Structurally Diverse Clinical Compounds

Murad, Neha; Pasikanti, Kishore Kumar; Madej, Benjamin D.; Minnich, Amanda; Crouch, Sabrinia; Polli, Joseph W.; Weber, Andrew

doi:10.1124/dmd.120.000202

Cited by 33 publications

(9 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This result is aligned with recent publications , and may be explained by additional variability and complexity of oral PK studies, where compounds’ solid form, applied dose, and formulation play a major role in final outcomes. Compared to recently developed computational models based on comparable data sets, our ML models showed similar, ,,,, improved, , or decreased performance. For example, if focusing on the drug-design stage results (i.e., in the absence of in vitro data), Obrezanova et al showed RMSLE values of 0.35, 0.31, and 0.57 for predictions of rat CLp, Vss, and oral F, respectively .…”

Section: Discussionmentioning

confidence: 65%

Computational Predictions of Nonclinical Pharmacokinetics at the Drug Design Stage

Stoyanova

Katzberger

Komissarov

et al. 2023

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Although computational predictions of pharmacokinetics (PK) are desirable at the drug design stage, existing approaches are often limited by prediction accuracy and human interpretability. Using a discovery data set of mouse and rat PK studies at Roche (9,685 unique compounds), we performed a proof-of-concept study to predict key PK properties from chemical structure alone, including plasma clearance (CLp), volume of distribution at steady-state (Vss), and oral bioavailability (F). Ten machine learning (ML) models were evaluated, including Single-Task, Multitask, and transfer learning approaches (i.e., pretraining with in vitro data). In addition to prediction accuracy, we emphasized human interpretability of outcomes, especially the quantification of uncertainty, applicability domains, and explanations of predictions in terms of molecular features. Results show that intravenous (IV) PK properties (CLp and Vss) can be predicted with good precision (average absolute fold error, AAFE of 1.96− 2.84 depending on data split) and low bias (average fold error, AFE of 0.98−1.36), with AutoGluon, Gaussian Process Regressor (GP), and ChemProp displaying the best performance. Driven by higher complexity of oral PK studies, predictions of F were more challenging, with the best AAFE values of 2.35−2.60 and higher overprediction bias (AFE of 1.45−1.62). Multi-Task approaches and pretraining of ChemProp neural networks with in vitro data showed similar precision to Single-Task models but helped reduce the bias and increase correlations between observations and predictions. A combination of GP-computed prediction variance, molecular clustering, and dimensionality-reduction provided valuable quantitative insights into prediction uncertainty and applicability domains. SHAPley Additive exPlanations (SHAPs) highlighted molecular features contributing to prediction outcomes of Vss, providing explanations that could aid drug design. Combined results show that computational predictions of PK are feasible at the drug design stage, with several ML technologies converging to successfully leverage historical PK data sets. Further studies are needed to unlock the full potential of this approach, especially with respect to data set sizes and quality, transfer learning between in vitro and in vivo data sets, model-independent quantification of uncertainty, and explainability of predictions.

show abstract

Section: Discussionmentioning

confidence: 65%

Computational Predictions of Nonclinical Pharmacokinetics at the Drug Design Stage

Stoyanova

Katzberger

Komissarov

et al. 2023

J. Chem. Inf. Model.

View full text Add to dashboard Cite

show abstract

“…Systemic clearance ( CL ) and volume of distribution ( V ss ) are critical PK parameters. CL describes how fast the drug would be eliminated from the blood; V ss describes the relationship between drug concentration measured in plasma or blood to the amount of drug in the body at equilibrium ( Murad et al, 2021 ). CL together with V ss , can be used to estimate elimination half-life and mean residence time, which define the concentration–time profile of a drug if it is administered via IV bolus.…”

Section: Discussionmentioning

confidence: 99%

“…The extrapolated volume of distribution (0.101 L) was close to the observed value (0.0924 L) in the rat PBPK model. Meanwhile, we also attempted to apply methods by Poulin and Thei, Rodgers, and Lukacova to fit V ss ( Poulin and Theil, 2002 ; Rodgers and Rowland, 2007 ; Murad et al, 2021 ). Results showed that all the predicted V ss values of ST-246 from the three methods dramatically overestimated the observed values.…”

Section: Discussionmentioning

confidence: 99%

A PBPK Model of Ternary Cyclodextrin Complex of ST-246 Was Built to Achieve a Reasonable IV Infusion Regimen for the Treatment of Human Severe Smallpox

Zhang¹,

Fu²,

Wang³

et al. 2022

Front. Pharmacol.

View full text Add to dashboard Cite

ST-246 is an oral drug against pathogenic orthopoxvirus infections. An intravenous formulation is required for some critical patients. A ternary complex of ST-246/meglumine/hydroxypropyl-β-cyclodextrin with well-improved solubility was successfully developed in our institute. The aim of this study was to achieve a reasonable intravenous infusion regimen of this novel formulation by a robust PBPK model based on preclinical pharmacokinetic studies. The pharmacokinetics of ST-246 after intravenous injection at different doses in rats, dogs, and monkeys were conducted to obtain clearances. The clearance of humans was generated by using the allometric scaling approach. Tissue distribution of ST-246 was conducted in rats to obtain tissue partition coefficients (Kp). The PBPK model of the rat was first built using in vivo clearance and Kp combined with in vitro physicochemical properties, unbound fraction, and cyclodextrin effect parameters of ST-246. Then the PBPK model was transferred to a dog and monkey and validated simultaneously. Finally, pharmacokinetic profiles after IV infusion at different dosages utilizing the human PBPK model were compared to the observed oral PK profile of ST-246 at therapeutic dosage (600 mg). The mechanistic PBPK model described the animal PK behaviors of ST-246 via intravenous injection and infusion with fold errors within 1.2. It appeared that 6h-IV infusion at 5 mg/kg BID produced similar Cmax and AUC as oral administration at 600 mg. A PBPK model of ST-246 was built to achieve a reasonable regimen of IV infusion for the treatment of severe smallpox, which will facilitate the clinical translation of this novel formulation.

show abstract

“…A compound with Papp more than 8x10 -6 cm/s is considered to have a high Caco2-2 permeability value [56]. All ligands were not able to present high Caco-2 permeability.…”

Section: B Admet Predictionmentioning

confidence: 99%

The Significance of Serine of NTD-NPC1L1 for Cholesterol Binding with Compounds in Traditional Spices

Amelia¹,

Hidayat²,

Iryani³

et al. 2023

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

It has been proposed that Niemann-Pick C1 like-1 (NPC1L1) is an intermediate membrane protein that facilitates cholesterol absorption at enterocytes, yet it is still being unknown mechanism. The study aimed to evaluate the pharmacokinetics of anti-cholesterol in traditional spices and investigate the interactions between anti-cholesterol compounds and NPC1L1. This research analyzed binding pocket, Admet, drug-likeness, the interactions, and binding affinities by DoGSiteScorer and Depth, AdmetSAR tools, and MOE.2009 software, respectively. The interactions and binding affinities were determined by molecular docking between NPC1L1 and 18 ligands derived from spices such as Cinnamon, Bay leaf, coriander, Garlic, Red Onion, Tumeric, Indosenian Chilli Pepper. Inhibitors were docked with NPC1L1 (PDB ID: 3QNT), and a comparison was made between the results of Ezetimibe, a prescribed NPC1L1 inhibitor. The Lipinski Rule of Five aids in identifying drug-like compounds and those that are not. As an octanol-water partition coefficient log P not greater than 5, all ligand including Ezetimibe has a higher affinity for the aqueous phase. 11 out of 18 inhibitors were well absorbed and distributed by forecasting oral bioavailability. Quercetin, Curcumin, and 6-Gingerol from onion, turmeric, and ginger are the potential to inhibit cholesterol absorption in the small intestine. These three highest binding energy ligands (-14.0320 to -12.3998 kcal/mol) had high binding affinities as Ezetimibe (-15.5075 kcal/mol). High binding affinities of Ezetimibe and these ligands interact at almost in the exact locations of the N-terminal domain. Through Ser_102 of N-terminal domain NPC1L1 binding with the ligands, we suggest that traditional spices of three ligands could interfere with cholesterol absorption at the early stages.

show abstract

Predicting Volume of Distribution in Humans: Performance of In Silico Methods for a Large Set of Structurally Diverse Clinical Compounds

Cited by 33 publications

References 34 publications

Computational Predictions of Nonclinical Pharmacokinetics at the Drug Design Stage

Computational Predictions of Nonclinical Pharmacokinetics at the Drug Design Stage

A PBPK Model of Ternary Cyclodextrin Complex of ST-246 Was Built to Achieve a Reasonable IV Infusion Regimen for the Treatment of Human Severe Smallpox

The Significance of Serine of NTD-NPC1L1 for Cholesterol Binding with Compounds in Traditional Spices

Contact Info

Product

Resources

About