Identification of Antituberculars with Favorable Potency and Pharmacokinetics through Structure-Based and...

Area of Science:

• Medicinal Chemistry
• Computational Drug Discovery
• Infectious Diseases

Background:

• Drug discovery for tuberculosis (TB) faces challenges in optimizing multiple molecular properties simultaneously.
• Identifying novel therapeutics requires efficient hit discovery pipelines that consider both efficacy and pharmacokinetic profiles.

Purpose of the Study:

• To develop and apply a hybrid structure- and ligand-based computational pipeline for identifying novel inhibitors of β-ketoacyl synthase KasA.
• To address multiple criteria critical for developing new therapeutics against Mycobacterium tuberculosis infection during the hit discovery phase.

Main Methods:

• A hybrid pipeline combining structure- and ligand-based approaches was developed.
• Bayesian optimization-guided docking and machine learning ensemble models were employed for compound nomination.
• Models predicted in vitro antibacterial efficacy (minimum inhibitory concentration, MIC) and mouse pharmacokinetic (PK) plasma exposure.

Main Results:

• The pipeline was applied to the Enamine HTS library (2.1 million molecules), selecting 93 diverse compounds.
• Experimental validation demonstrated high success rates: 41% for PK and 19% for MIC.
• Twelve compounds met hit-like criteria for both MIC and PK, serving as promising starting points for drug discovery.

Conclusions:

• The developed computational pipeline is effective for nominating drug candidates with desirable antibacterial and pharmacokinetic properties.
• This approach accelerates the hit discovery phase for novel anti-tuberculosis therapeutics.
• The identified hit compounds provide a strong foundation for further drug development programs targeting Mycobacterium tuberculosis.