9-beta-D-Arabinofuranosylguanine (araG), an analog of deoxyguanosine which is not degraded by purine nucleoside phosphorylase, has been previously shown in in vitro studies by our laboratory to be effective in purging malignant T cells from human bone marrow (1). We now describe studies in a murine model of T-cell acute lymphoblastic leukemia (ALL) in which we tested whether bone marrow, contaminated with malignant T cells and purged ex vivo with araG, could reconstitute both the lymphoid and myeloerythroid lineages in the absence of leukemic relapse. The model utilized 6C3HED tumor cells, derived from a Thy 1.2+ malignant murine T-cell line, which were shown to cause lethal leukemia in C3H/HeN mice. Intravenous injection of 10(6) 6C3HED cells resulted in 100% mortality within 18 days, with autopsy revealing tumor infiltration of multiple organs. 100% of non-leukemia bearing lethally irradiated C3H/HeN mice transplanted with syngeneic bone marrow, treated ex vivo with 100 microM of araG for 18 hours, survived > 365 days post-transplant with full lymphohematopoietic reconstitution. Evidence of araG's ability to purge bone marrow of malignant tumor cells without causing significant toxicity to normal marrow derived hematopoietic progenitor cells was documented in experiments in which 75% of lethally irradiated mice transplanted with syngeneic bone marrow, contaminated with 6C3HED tumor cells and treated ex vivo with 100 microM araG for 18 hours, survived for > 250 to > 400 days. Death in 25% of mice was secondary to infection which developed before marrow reconstitution occurred. Reconstitution of the lymphoid, myeloid, and erythroid lineages with donor cells in surviving mice was documented. The data presented indicate that araG may effectively purge bone marrow of malignant T cells without irreversible toxicity to hematopoietic stem cells. This purging regimen is recommended for consideration for clinical trials in patients with T-cell malignancies undergoing autologous bone marrow transplantation and may also be a viable option for T-cell depletion as a strategy to prevent graft-versus-host disease.