The fate of the toxic metabolite of dapsone, dapsone hydroxylamine, has been studied in the human red cell. Twice-washed red cells were incubated at 37-degrees with dapsone hydroxylamine: at 3 and 5 min, 27.0 +/- 2.2 and 33.2 +/- 2.7% of the haemoglobin had been converted to methaemoglobin, leading to a maximum at 45 min (45 +/- 1.8%). HPLC analysis revealed that parent amine was produced from dapsone hydroxylamine during methaemoglobin formation in the red cells. At 3 min, conversion of dapsone hydroxylamine to dapsone reached 7.0 +/- 3.9% leading to a maximum at 30 min (18.1 +/- 3.7%). There was a linear relationship between hydroxylamine-dependent methaemoglobin formation and conversion of hydroxylamine to dapsone (r = 0.97). At 4-degrees, methemoglobin and dapsone formation was greatly retarded, and did not exceed 10%. Co-incubation of diethyl dithiocarbamate (DDC) with dapsone hydroxylamine and red cells led to a marked increase in methaemoglobin formation (61.4 +/- 3.4%) compared with hydroxylamine and red cells alone (45.0 +/- 1.8%, P < 0.001) at 44 min, and conversion of dapsone hydroxylamine to dapsone was almost doubled at 45 min (35.7 +/- 5.3%) compared with hydroxylamine and red cells (18.1 +/- 2.5%). A linear relationship between met-haemoglobin formation and dapsone formation (r = 0.96) was also shown to occur in the presence of DDC. Incubation of red cells with DDC and dapsone hydroxylamine caused a significantly greater reduction in glutathione levels (98.3 +/- 1.6%) compared with red cells and dapsone hydroxylamine alone (84.8 +/- 2.7%) at 5 min (P < 0.001), although there was no significant difference between the groups at 15 min (96.9 +/- 2.6 vs 98.1 +/- 2.2%). Intra-erythrocytic glutathione was then depleted by 75 +/- 3.4%, by pretreatment with diethyl maleate (6 mM), and these cells in the presence of the hydroxylamine showed a significant fall in both methaemoglobin generation (29.7 +/- 1.2 vs 35.0 +/- 1.7%) and parent amine formation (11.1 +/- 0.2 vs 16.5 +/- 1.1 %) compared with untreated red cells at 45 min. It is possible that a cycle exists between hepatic oxidation of dapsone to its hydroxylamine and reduction to the amine within the red cell, which may lead to re-oxidation by hepatic cytochrome P450. This process may contribute to the persistence of the drug in vivo.