| Abstract |
NRF2 stabilizes redox potential through genes for glutathione and thioredoxin antioxidant systems. Whether blockade of glutathione and thioredoxin is useful in eliminating cancer stem cells remain unknown. We used xenografts derived from colorectal carcinoma patients to investigate the pharmacological inhibition of glutathione and thioredoxin systems. Higher expression of five glutathione S-transferase isoforms (GSTA1, A2, M4, O2, and P1) was observed in xenograft-derived spheroids than in fibroblasts. Piperlongumine (2.5-10 μmol/L) and auranofin (0.25-4 μmol/L) were used to inhibit glutathione S-transferase π and thioredoxin reductase, respectively. Piperlongumine or auranofin alone up-regulated the expression of NRF2 target genes, but not TP53 targets. While piperlongumine showed modest cancer-specific cell killing (IC50 difference between cancer spheroids and fibroblasts: P = 0.052), auranofin appeared more toxic to fibroblasts (IC50 difference between cancer spheroids and fibroblasts: P = 0.002). The synergism of dual inhibition was evaluated by determining the Combination Index, based on the number of surviving cells with combination treatments. Molar ratios indicated synergism in cancer spheroids, but not in fibroblasts: (auranofin:piperlongumine) = 2:5, 1:5, 1:10, and 1:20. Cancer-specific cell killing was achieved at the following drug concentrations (auranofin:piperlongumine): 0.25:2.5 μmol/L, 0.5:2.5 μmol/L, or 0.25:5 μmol/L. The dual inhibition successfully decreased CD44v9 surface presentation and delayed tumor emergence in nude mouse. However, a small subpopulation persistently survived and accumulated phosphorylated histone H2A. Such "persisters" still retained lesser but significant tumorigenicity. Thus, dual inhibition of glutathione S-transferase π and thioredoxin reductase could be a feasible option for decreasing the tumor mass and CD44v9-positive fraction by disrupting redox regulation.
|
