We have recently demonstrated that the combination of the alkylating agent nitrogen mustard (NM) and etoposide (VP-16) is capable of eliminating, ex vivo, leukemic cells contaminating PBSC collections and this is associated with a significant recovery of primitive and committed hematopoietic progenitor cells. Based on these data a pilot study on autologous transplantation of NM/VP-16 purged PBSC for high-risk leukemic patients was recently initiated. Twelve patients (seven females and five males) with a median age of 46 years (range 18-57) have been treated. Two patients had acute myeloblastic leukemia (AML) resistant to conventional induction treatment, four patients had secondary AML in I complete remission (CR), one patient was in II CR after failing a previous autologous BM transplantation, while two additional AML individuals were in I CR achieved after three or more cycles of induction treatment. Two patients with high-risk acute lymphoblastic leukemia (ALL) in I CR and one patient with mantle cell lymphoma and leukemic dissemination were also included. Eight patients showed karyotypic abnormalities associated with a poor clinical outcome. The mobilizing regimens included cytosine arabinoside and mitoxantrone with (n = 6) or without fludarabine (n = 3) followed by subcutaneous administration of G-CSF (5 microg/kg/day until the completion of PBSC collection) and G-CSF alone (n = 3) (15 microg/kg/day). A median of two aphereses (range 1-3) allowed the collection of 7.2 x 10(8) TNC/kg (range 3.4-11.5), 5 x 10(6) CD34+ cells/kg (range 2.1-15.3) and 9.2 x 10(4) CFU-GM/kg (0.3-236). PBSC were treated with a constant dose of 20 microg of VP-16/ml and a median individual-adjusted dose (survival < or = 5% of steady-state BM CFU-GM) of NM of 0.7 microg/ml (range 0.25-1.25). Eleven patients were reinfused after busulfan (16 mg/kg) and Cy (120 mg/kg) conditioning with a median residual dose of 0.3 x 10(4) CFU-GM/kg (0-11.5). The median time to neutrophil engraftment (>0.5 x 10(9)/l) for evaluable patients was 25 days (range 12-59); the median time to platelet transfusion independence (>20 and >50 x 10(9)/l) was 40 days (18-95) and 69 days (29-235), respectively. Hospital discharge occurred at a median of 25 days (18-58) after stem cell reinfusion. Four individuals are alive in CR (n = 3) or with residual nodal disease (n = 1 lymphoma patient) with a follow-up of 32, 26, 3 and 14 months, respectively. Seven patients died due to disease progression or relapse (n = 5) or extrahematological transplant toxicity (n = 2). Our data suggest that pharmacological purging of leukapheresis collections of leukemic patients at high-risk of relapse is feasible and ex vivo treated cells reconstitute autologous hematopoiesis.