DNA strand breakage, thymine glycol production, and hydroxyl radical generation induced by different samples of crystalline silica in vitro.

Five preparations of alpha-quartz [Min-U-Sil 5 (MQZ), MQZ pretreated with hydrofluoric acid (HFMQZ), Chinese standard alpha-quartz (CSQZ), and two German samples, DQ-12 and F600] and two preparations of the crystalline silica polymorphs, cristobalite and tridymite, previously characterized for surface area and surface charge, were evaluated for their relative activities in the following assays: (i) in vitro assays of short duration (< or = 15 min) for oxygen consumption and for generation of hydroxyl radicals (measured by electron spin resonance spin trapping), and (ii) in vitro assays of longer duration for DNA strand breakage (measured using linear DNA as a detector molecule) and for production of the oxidized DNA base, thymine glycol (measured by gas chromatography-mass spectrometry). Marked differences among the samples were found for their levels of oxygen consumption and of hydroxyl radicals' generation. All samples caused increased formation of thymine glycol, with wide variations in activity among samples. When normalized for equal surface area, the samples produced different levels of DNA strand breakage. Addition of hydrogen peroxide strongly accelerated DNA damage--more for cristobalite than for the alpha-quartz samples. DNA damage by quartz was enhanced by ferric chloride and inhibited by iron chelators. The order of relative activity of the samples varied with different types of in vitro assays and was not directly correlated to surface area. Electrophoretic mobility, as measured by zeta potential, was not significantly different among samples. The results suggest that the ability of different crystalline silica samples to generate a rapid burst of oxygen free radicals is distinct from their ability to induce DNA damage and DNA base oxidation over longer time periods. The relative activities of the samples in cellular assays (hemolysis of human erythrocytes; cytotoxicity and neoplastic transformation of BALB/3T3/A31-1-1 cells) were in turn markedly different from those listed above, suggesting a more critical role for surface area. The mechanisms of carcinogenesis by crystalline silica need to be further investigated in relation to the underlying physicochemical characteristics.