Murine leukemia L1210 cells grown for 2-3 weeks in the presence of 1% serum without selenium supplementation [L·Se(−) cells] typically exhibited <10% of the glutathione peroxidase (GPX) and phospholipid hydroperoxide glutathione peroxidase (PHGPX) activity of selenium-satisfied controls [L·Se(+) cells]. Concomitant with diminished GPX and PHGPX activity was a 1.5- to 2.0-fold increase in catalase (CAT) activity, which reverted to control levels when L·Se(−) cells were given sufficient Se for full expression of selenoperoxidase activity. Selenium manipulation affected total glutathione content similarly, but had no effect on glutathione-S-transferase or superoxide dismutase activity. Long-term growth under Se-deficient conditions resulted in a progressive additional increase in CAT activity, which maximized after ca. 5 months. These cells [referred to as L′·Se(−)] attained CAT activity levels at least 100-times greater than those of Se-supplemented [L′·Se(+)] controls, whereas their glutathione content remained elevated by ∼70%. Supplying L′·Se(−) cells with Se resulted in a rapid elevation to full GPX activity; however, CAT failed to decline in this case, suggesting that a selection for stable CAT hyperexpressing variants had been accomplished. Quantitative immunoblot analysis indicated that the high CAT activity of L′·Se(−) cells is accounted for by an elevated level of enzyme protein. Induction of CAT and selection for CAT-rich phenotypes, as apparent for Se-starved L1210 cells, was not observed in human K562 counterparts, which lack GPX and express only a low level of PHGPX. L·Se(−) cells were found to be more sensitive to H2O2-induced killing than L·Se(+) controls, whereas L′·Se(−) cells were exceedingly more resistant to H2O2 than L′·Se(+) counterparts. By contrast, L·Se(−) and L′·Se(−) cells were both more sensitive to t-butyl hydroperoxide than Se(+) controls, consistent with CAT being unimportant in the detoxification of this peroxide compared with GPX. This appears to be the first reported evidence for CAT hyperexpression in response to selenium deprivation.

Resealed ghosts of human erythrocytes are sensitive to oxidative damage induced by xanthine oxidase acting on xanthine in the presence of iron. Damage was assessed in terms of lipid peroxidation and increased permeation of trapped markers, Na+ and glucose-6-P. Key findings are as follows. (a) Marker efflux from xanthine/xanthine oxidase/iron-treated ghosts accelerated after a lag, Na+ emerging far ahead of glucose-6-P. (b) Both effluxes and lipid peroxidation were stimulated by Fe(III) in a dose-dependent fashion and inhibited by chelating agents. (c) The antioxidant butylated hydroxytoluene effectively halted lipid peroxidation and net glucose-6-P efflux, but slowed Na+ efflux only partially. (d) Lipid peroxidation and marker release could be completely inhibited by superoxide dismutase or catalase, indicating that O2- and H2O2 are both required, possibly as precursors of OH. via the iron-catalyzed Haber-Weiss reaction (O2- + H2O2 leads to OH- + OH. + O2). (e) OH. scavengers, e.g. ethanol, mannitol, choline, had no protective effect against marker efflux and lipid peroxidation. Yet these agents did intercept OH. in the bulk medium, since they inhibited the degradation of 2-deoxyribose added as an extramembranous OH. probe. It is proposed that OH. produced on the membrane at iron binding sites reacts so rapidly with target molecules that scavengers cannot compete. (f) Desferrioxamine abolished all effects, including net egress of Na+. EDTA, while totally inhibitory toward lipid peroxidation and glucose-6-P release, diminished Na+ release partially, changing it to first order, approximately 3-fold faster than background. The latter response was totally inhibited by catalase, but only marginally by superoxide dismutase. This and other evidence suggests that different forms of membrane damage are responsible for enhanced permeation of the two markers; although glucose-6-P depends on lipid peroxidation, Na+ does not, certainly when EDTA is present.