Integrative drug repositioning identifies FDA-approved inhibitors of HSP90AA1 with therapeutic potential in colorectal cancer.

Colorectal cancer (CRC) is a leading cause of cancer mortality, with treatment efficacy limited by tumor heterogeneity and chemotherapy resistance. Drug repositioning can accelerate discovery using FDA-approved agents. Here, we introduce a multi-omics repositioning framework integrating bulk and single-cell data with network pharmacology, structural modeling, and validation and apply it for the first time in CRC. Bulk RNA-seq identified 3480 upregulated and 3816 downregulated genes in TCGA-COADREAD, and 602 upregulated and 744 downregulated in GSE103512.

Single-cell RNA-seq revealed 1227 upregulated and 728 downregulated in GSE231559, and 1628 upregulated and 850 downregulated in GSE144735. Integration with cMap and ASGARD yielded 20 candidates; four FDA-approved drugs (rifapentine, danazol, tolbutamide, and triamcinolone) had not been linked to CRC. DepMap PRISM data supported inhibitory activity. Docking and 200 ns molecular dynamics showed stable binding of all four in the HSP90α ATP-binding pocket, with hydrogen bonding and favorable MM/GBSA free energies (-30.59 kcal/mol for danazol to -37.37 kcal/mol for rifapentine; reference NVP-AUY922 -47.99 kcal/mol).

Independent triplicate simulations of the reference and top-ranked candidate complexes confirmed the statistical reliability and reproducibility of these structural and energetic descriptors (reference ΔGtotal = -47.90 ± 0.08 kcal/mol). Direct binding of all four compounds to HSP90α was further confirmed by surface plasmon resonance. In vitro, the drugs suppressed proliferation and induced apoptosis in HCT116, LoVo, and HCT-15 cells, with IC₅₀ values of 2.60-13.15 μM.

These results identify unreported HSP90AA1 inhibitors in CRC and establish a generalizable repositioning framework.