An integrative omics-guided druggability analysis of VCX2 in hepatocellular carcinoma using Peruvian natural products

Hepatocellular carcinoma (HCC) is among the deadliest cancers, and current biomarkers offer limited diagnostic and therapeutic utility. Identifying novel druggable targets remains a critical challenge for improving HCC management.

We implemented an omics-guided computational pipeline integrating single-cell RNA sequencing (scRNA-seq), differential gene expression (DGE) analysis, UMAP clustering, and protein-protein interaction (PPI) network mapping to prioritize candidate genes. Structural characterization of the selected target was performed using AlphaFold-derived models followed by long-timescale molecular dynamics (MD) simulations. Virtual screening of the PeruNPDB (Peruvian Natural Products Database) was conducted using Glide docking, with further evaluation by MM-GBSA and MD-based interaction analyses.

Among prioritized genes (TMBIM4, RGS5, CEACAM7, and VCX2), the cancer/testis antigen VCX2 emerged as a promising candidate due to its aberrant expression and potential involvement in chromosomal instability. MD refinement yielded a stable and ligand-accessible VCX2 conformation. Virtual screening identified luteolin-5-O-glucoside from Equisetum arvense as the top ligand (Glide score: -3.949 ± 0.85 kcal/mol; ΔG_bind: -35.43 ± 1.12 kcal/mol). Despite a modest docking score, consistent with the shallow and polar binding site, MM-GBSA and MD simulations supported a thermodynamically favorable and dynamically stable interaction. Key hydrogen bonds with residues Glu68, Thr63, and Ala61 were maintained within a stabilized polar groove.

These findings support VCX2 as a potential molecular target in HCC and highlight luteolin-5-O-glucoside as a promising lead scaffold. This study provides a hypothesis-generating framework that integrates single-cell transcriptomics with structure-based druggability analysis, offering new avenues for targeted therapeutic development in HCC.