# Advancing Torsades de pointes risk prediction: unveiling the role of drug metabolites through molecular docking

## TL;DR

This study improves prediction of a dangerous heart rhythm disorder by analyzing how drugs and their metabolites interact with a key heart protein.

## Contribution

The study introduces a new computational framework combining binding energy, geometry, and physicochemical data to assess drug metabolite risk for arrhythmia.

## Key findings

- Non-toxic metabolite fexofenadine showed higher binding affinity than its toxic parent drug terfenadine, but its safety is due to its zwitterionic nature.
- Desmethylastemizole maintained high affinity with a relaxed fit, explaining its sustained potency compared to astemizole.
- Quetiapine acted as a steric blocker with a larger interface area, while its metabolite norquetiapine had a smaller, more specific binding site.

## Abstract

This study explores the risk of Torsades de Pointes (TdP) arrhythmia, focusing on the interactions of parent drugs and their metabolites with the Human ether-à-go-go-related gene (hERG) channel, which is crucial in cardiac electrical activity and TdP risk assessment. Using a dual-strategy molecular docking approach with AutoDock Vina and PatchDock, we analyzed clinically relevant ligand pairs: astemizole/desmethylastemizole, terfenadine/fexofenadine, and quetiapine/norquetiapine. Quantitative analysis revealed that high binding affinity does not always correlate with toxicity. For instance, the non-cardiotoxic metabolite fexofenadine exhibited a higher binding affinity (−9.3 kcal/mol) compared to its toxic parent terfenadine (−8.9 kcal/mol), but its safety is explained by physicochemical constraints (zwitterionic nature). Conversely, desmethylastemizole maintained high affinity (−9.2 kcal/mol) with a geometrically “relaxed” fit (Atomic Contact Energy: −338.36), rationalizing its sustained potency. Geometric analysis further distinguished quetiapine as a “steric blocker” (Contact Area: ~588 Å2) causing forced occlusion, whereas its metabolite norquetiapine acted as a specific ligand with a significantly smaller interface area (~417 Å2). These findings highlight the importance of focusing not only on the parent drug but also on metabolites for TdP risk assessment in new drug development. We advocate for an integrated computational framework combining binding energy, geometric complementarity, and physicochemical profiling to enhance the accuracy of early cardiac safety screenings.