A possible metabolic linkage between hepatic thyroxine (T4) 5'-deiodination and the NADPH-glutathione (GSH) cycle was studied in rat liver. Supplementation of 1 mM NADPH to stocked liver homogenates in vitro produced 4 fold increase in 3, 5, 3'-triiodothyronine (T3) formation from T4, whereas the effect of 1 mM FMN, FDN, NAD, NADH, or GSH was relatively small. An exponential dose-response relation was obtained between NADPH and T3 generated. The dose-dependent increase in T3 formation on GSH was eliminated in the presence of 1 mM MADPH, and the additive effect of GSH to NADPH was not apparent in comparison with NADPH alone. Inhibition of T3 generation by graded doses of methylene blue was not affected by the presence of 5 mM GSH. Furthermore, metabolic changes in the hexose-monophosphate shunt were produced in male Wistar rats aged 5 w by treating them with fasting-refeeding (FF group), with the administration of insulin and glucose (IG group), with propylthiouracil (PTU group) and with T4 (T4 group). All these treatments significantly reduced hepatic T4 5'-deiodinase activity (P less than 0.01-0.001 vs control), while glucose-6-phosphate dehydrogenase (G6PD) and glutathione reductase (GSSG-R) activities were increased. Between generated T3 and G6PD or GSSG-R activity, an inverse correlation was noted (r = -0.802 and -0.933, P less than 0.001). No consistent relation was found between T4 5'-deiodinase activity and GSH or non-protein SH contents. The addition of 1 mM NADPH and GSH to the homogenates of FF, T4 and the control group stocked for 4 w at -20 degrees C, restored T4 5'-deiodinase activity from a level of 10% to 60% of the initial value, whereas the activity remained depressed in PTU (19%) and the IG group (37%). These results indicate that both GSH and NADPH are important cofactors of the T3 generating system, but NADPH is more rate-limiting and its effect appears to be rather direct, not mediated by GSH formation. It is possible that T4 5'-deiodinase may be one of the NADPH-dependent enzymes.