A simple model to solve a complex drug toxicity problem | JoVE Visualize

Area of Science:

Pharmacology and Toxicology
Medicinal Chemistry
Drug Discovery and Development

Background:

Current drug toxicity models (e.g., therapeutic index, rule of five) have limitations in predicting toxicity accurately.
Low productivity in drug discovery (<20%) and patient benefit highlight the need for improved toxicity prediction.
Predicting drug toxicity remains a significant challenge throughout drug development stages.

Purpose of the Study:

To develop a novel, simple drug toxicity model named Drug Toxicity Index (DTI) for improved prediction of drug safety.
To redefine drug toxicity as a function of pharmacokinetic (PK), pharmacodynamic (PD), and physicochemical parameters.
To assess DTI's performance against established toxicity assessment rules and its utility in various drug discovery and clinical settings.

Main Methods:

Developed DTI using 711 oral drugs, defining toxicity via scaled biphasic and exponential functions of PK, PD, and physicochemical parameters.
Estimated toxicity contributions from IC50, free Cmax, and logD values, scaled by molar dose and oral bioavailability.
Validated DTI against the US-FDA's Liver Toxicity Knowledge Base (LTKB) and compared its performance with existing methods.

Main Results:

DTI demonstrated superior performance over standard rules, LEI, and exposure-based TIs in identifying safe and toxic drugs.
Achieved high accuracy in classifying drugs from the LTKB, with AUC values ranging from 0.91 to 0.64 across WHO ATC categories.
Identified drugs with DTI above WHO ATC category averages as consistently toxic, and those below as relatively safe.

Conclusions:

DTI provides a robust and effective method for predicting drug toxicity, outperforming existing models.
The model aids in understanding specific toxicity contributions (PK, PD, physicochemical) and identifying potentially harmful drugs.
Proposed decision trees for DTI application in preclinical and clinical settings, potentially reducing drug discovery failures and aiding therapeutic drug monitoring.