Synergistic approach utilizing bioinformatics, machine learning, and traditional screening for the identification of novel CSK inhibitors targeting hepatocellular carcinoma

  • J Comput Aided Mol Des. 2025 Nov 8;39(2):119. doi: 10.1007/s10822-025-00703-3.
Yang Lu  #  1 Bizhi Li  #  2 Xiaoli Zheng  3 Lei Xu  4 Linghui Zeng  3 Chong Zhang  3 Jiankang Zhang  5
Affiliations
  • 1. Key Laboratory of Novel Targets and Drug Study for Neural Repair of Zhejiang Province, School of Medicine, Hangzhou City University, Hangzhou, 310015, China. [email protected].
  • 2. Hangzhou Institute of Innovative Medicine, College of Pharmaceutical Sciences, Zhejiang University, Hangzhou, 310058, China.
  • 3. Key Laboratory of Novel Targets and Drug Study for Neural Repair of Zhejiang Province, School of Medicine, Hangzhou City University, Hangzhou, 310015, China.
  • 4. Institute of Bioinformatics and Medical Engineering, School of Electrical and Information Engineering, Jiangsu University of Technology, Changzhou, 213001, China.
  • 5. Key Laboratory of Novel Targets and Drug Study for Neural Repair of Zhejiang Province, School of Medicine, Hangzhou City University, Hangzhou, 310015, China. [email protected].
  • # Contributed equally.
Abstract

The overexpression or activation of C-terminal Src kinase (CSK) has been recognized as a pivotal factor in the progression of hepatocellular carcinoma (HCC), positioning CSK as a promising therapeutic target. Despite this potential, no CSK-specific inhibitors have been developed for HCC treatment to date. Addressing this gap, our study established a robust virtual screening protocol that integrates energy-based screening techniques with machine learning methodologies. Through this systematic approach, we identified a novel compound, 6, exhibiting potent CSK inhibitory activity, as evidenced by an IC50 value of 675 nM in a homogeneous time-resolved fluorescence (HTRF) bioassay. Notably, this compound demonstrated significant growth inhibition in Huh-7 and Huh-6 cell lines, along with the suppression of clone formation. To elucidate the underlying mechanism, we conducted molecular dynamics simulations, which revealed critical binding interactions between compound 6 and CSK. Specifically, residues Phe333 and Met269 were found to play essential roles in mediating these interactions, providing valuable insights into the compound's mode of action.

Keywords
Bioinformatics; CSK; HCC; Machine learning; Molecular dynamic; Virtual screening.