In vitro bioactivity profile-driven discovery of drug synergies in hepatocellular carcinoma
TL;DR
A data-driven workflow to integrate two orthogonal single-agent resources for 244 drugs, including cell viability profiles across 11 liver cancer cell lines and bioactivity signatures across 1,925 biochemical and cell-based assays is developed.
Abstract
Combination therapy is a central strategy to overcome drug resistance in hepatocellular carcinoma (HCC), yet systematic identification of synergistic combinations is constrained by the combinatorial search space. We developed a data-driven workflow to integrate two orthogonal single-agent resources for 244 drugs, including cell viability profiles across 11 liver cancer cell lines and bioactivity signatures across 1,925 biochemical and cell-based assays. We calculated the correlation of the bioactivity profiles between each drug pair yielding 821 high-confidence combinations involving 190 unique drugs, from which eight anchors and a 125-drug library were selected by frequency-guided prioritization. Fixed-concentration screening of 992 anchor-library combinations in the Hep 3B2.1-7 cell line yielded 89 combinations with potency enhancement. Panobinostat showed the largest potency gains when combined with briciclib or thiocolchicine (69.5-fold and 81.1-fold lower IC₅₀, respectively). Matrix-based screening of these two representative combinations across 11 cell lines revealed pronounced concentration- and context-dependence. At optimal dose pairs, strong synergy was observed in Hep 3B2.1-7 (Zero Interaction Potency (ZIP) = 40.80 ± 4.22 with 58.53 ± 6.54% inhibition), whereas antagonism was observed in some cellular backgrounds. This label-free integration strategy reduces the experimental burden for hypothesis-driven combination discovery and provides context-specific preclinical leads for follow-up evaluation in HCC.