A novel immunosuppressant, succinyl acetone (4,6-dioxoheptanoic acid), was studied in dogs. Results with bolus intravenous injections at doses ranging from 50 to 1600 mg/kg showed dose-dependent alpha and beta half-lives, ranging from 30 to 80 min and 7 to 27 hr, respectively. Results suggested that continuous i.v. infusion was necessary to maintain constant plasma levels. Four dogs were given 9.2 Gy total-body irradiation and autologous marrow transplants along with continuous i.v. infusion of succinyl acetone at 50, 100, 200, or 400 mg/kg/day for 21 days, and all four had rapid, sustained hematopoietic engraftment. However, two of the four dogs receiving 200 and 400 mg succinyl acetone/kg/day, respectively, developed bilateral hind-limb ataxia, with histologically confirmed cerebellar lesions in the dog given the higher dose, thus establishing a potential dose-limiting neurotoxicity. Prevention of graft-versus-host disease was studied in recipients of allogeneic marrow. Dogs were given 9.2 Gy TBI, followed by hematopoietic grafts from unrelated DLA-nonidentical or DLA-haploidentical littermate dogs. Succinyl acetone was given as continuous infusion for 21 days after transplant at doses of 100-300 mg/kg/day. Starting succinyl acetone on the day of marrow infusion in four dogs failed to prevent rapid onset of acute GVHD, and dogs survived no longer than controls. Starting succinyl acetone 3 days before transplant delayed the onset of acute GVHD and prolonged survival significantly compared with that of dogs not given postgrafting immunosuppression (P = 0.008); survival was comparable to that in previously reported dogs given either methotrexate or cyclosporine as postgrafting immunosuppression (P = 0.88 and 0.99, respectively). Seven of the sixteen allogeneic recipients developed evidence of neurotoxicity during succinyl-acetone infusion. Neurological dysfunctions were manifested by hind-limb ataxia and posterior paresis. In conclusion, succinyl acetone significantly delayed the onset of GVHD and prolonged survival of DLA-nonidentical marrow graft recipients but did not induce graft-host tolerance and was associated with dose-limiting neurotoxicity.