The toxic effect of cyclophosphamide on the proliferative cell population of hair follicles plucked from the human scalp was examined by the in vivo nuclear aberration assay. Patients participating in an independent clinical trial received oral low dose cyclophosphamide, intravenous high dose cyclophosphamide or oral placebo treatment. The percent of cells with nuclear aberrations (indicating apoptosis, a special form of cell death) and the percent of mitotic cells, in the hair matrix, were calculated for each patient before treatment and at several time points following cyclophosphamide or placebo treatment. The mean percentages of nuclear aberrations in both the treated Low dose and High dose cyclophosphamide patients were significantly higher than those for the pre-treatment and Placebo patients. The nuclear aberrations in hair follicle cells increased from pre-treatment (and Placebo) to treated Low dose and finally to treated High dose patients. The average percentage for pre-treatment samples from all patients was 0.06 +/- 0.03 SE. For 1 week and 1 month samples from Low dose patients it was 0.35 +/- 0.08 SE, and for combined 2,3 and 4 day samples from High dose patients it was 1.08 +/- 0.12 SE. Cyclophosphamide also had a significant effect on mitosis. A decrease in mitotic activity was observed at 1 month following the initial low dose cyclophosphamide treatment and at 24 +/- 2 h following each of the first two high dose cyclophosphamide treatments. The observed increase in nuclear aberrations following low dose as well as high dose cyclophosphamide suggests that it is feasible to use the nuclear aberration assay for in vivo human genotoxicity testing, using proliferating hair follicle cells.