Structure-activity study with bioreductive benzoquinone alkylating agents: effects on DT-diaphorase-mediated DNA crosslink and strand break formation in relation to mechanisms of cytotoxicity

Structure-activity studies were carried out with the model bioreductive alkylating agent benzoquinone mustard (BM) and its structural analogs. The specific objectives were: (1) to investigate the effects of functional group substitutions to the benzoquinone ring on DNA crosslink and strand break formation subsequent to reduction of the analogs by DT-diaphorase (DTD) in vitro, (2) to correlate DNA crosslink and strand break formation by the analogs with anaerobic reduction of the BM analogs by DTD and their redox cycling in vitro, and (3) to correlate DNA crosslink and strand break formation by the BM analogs with their cytotoxic effects in cancer cells.

DNA interstrand crosslink and single-strand break formation were assessed using agarose gel assays. To determine DNA interstrand crosslinks or single-strand breaks, linearized or supercoiled plasmid DNA, respectively, were incubated with purified human DTD and increasing concentrations of each BM analog. Subsequently, DNA was electrophoresed on an agarose gel and DNA crosslink and strand break formation were quantified using densitometry. The rates of reduction of the BM analogs by purified human DTD were measured in vitro under hypoxic conditions, and the redox cycling potential was determined under aerobic conditions using HPLC analysis. The cytotoxic activities of these agents in human tumor cell lines were measured by the MTT assay, with and without the DTD inhibitor, dicoumarol.

BM analogs with electron-donating groups (MeBM, MBM, m-MeBM), electron-withdrawing groups (CBM, FBM), sterically bulky groups (PBM, m-PBM, m-TBM) and positional isomers (MeBM, m-MeBM, PBM, m-PBM) were synthesized. After reduction by DTD, the BM analogs produced a concentration-dependent increase in DNA crosslink and DNA strand break formation. The E₁₀ (extent of DNA crosslink formation produced by 10 μM BM analog) for DNA crosslink formation displayed the rank order MeBM≈MBM>m-MeBM≈PBM≈BM>CBM>FBM>m-PBM≈m-TBM. For DNA strand break formation, the E₁₀ values (extent of DNA strand break formation produced by 10 μM BM analog) displayed the rank order MeBM>MBM>m-MeBM>PBM>BM≈CBM>FBM>m-PBM≈m-TBM. Importantly, the cytotoxic activity of the BM analogs in SK-Mel-28 human melanoma cells correlated positively with the E₁₀ values for DTD-mediated DNA crosslink formation (r _s=0.87, P<0.05) and DNA strand break formation (r _s=0.95, P<0.05). Similar correlations were observed in NCI-H661 human lung carcinoma cells. Furthermore, the D₁₀ values (concentration of BM analog that decreased the surviving cell fraction to 0.1) for cytotoxic activity of the BM analogs correlated with the maximum levels of DNA crosslinks formed with each BM analog, with r _s values of −0.85 (P<0.05) for the NCI-H661 cell line, and −0.81 (P<0.05) for the SK-MEL-28 cell line. The half-time of reduction (t_1/2) of the BM analogs by DTD did not correlate with DNA crosslink formation, DNA strand break formation, or cytotoxic potency of the analogs.

Functional groups on the benzoquinone ring affect the ability of BM to produce DNA crosslinks and strand breaks following reduction by DTD. Electron-donating groups increased DNA damage, whereas electron-withdrawing groups and sterically bulky groups at the C6 position had no effect or decreased the ability of the compounds to produce DNA damage compared to BM. Moreover, both DNA crosslink and strand break formation appear to have an important impact on the cytotoxicity of the BM analogs. These results may have significance for optimal use of BM-based antitumor agents and for rationalization of the development of novel therapeutic compounds that require bioactivation by DTD.