Bone marrow transplantation (BMT) can produce prolonged clinical remission in some patients with hematologic malignancies. Unfortunately, disease relapse may occur despite BMT. Studies in animal models and clinical experience have provided evidence that immunologic factors play an important role in preventing relapse post-BMT. To stimulate immunologic activity in patients post-BMT, we administered prolonged uninterrupted continuous infusions of low-dose recombinant interleukin-2 (rIL-2). Thirteen marrow recipients (seven autologous BMT, six CD6 T-depleted allogeneic BMT) received rIL-2 at a dose of 2 x 10(5) U/m2/d for a scheduled period of 90 days. rIL-2 was administered through a Hickman catheter with a portable pump beginning a median of 85 days after BMT. Toxicity was minimal and all treatment could be undertaken in the outpatient setting. No patient developed any signs of graft-versus-host disease, hypotension, or pulmonary capillary leak syndrome. Treatment did not affect the absolute neutrophil count or hemoglobin level, but eosinophils increased substantially in most patients. Platelet counts decreased by 20% in 10 of 13 individuals within 2 weeks, but stabilized thereafter. Despite the low dose of rIL-2 administered, significant immunologic changes were noted. Specifically, all 13 patients experienced a marked increase (fivefold to 40-fold) in natural killer (NK) cell number. Phenotypic characterization showed that the majority of NK cells were CD56bright+ CD16+ CD3-. In contrast, a minor increase in T-cell number was noted in only 4 of 13 patients. Low-dose rIL-2 treatment resulted in augmentation of in vitro cytotoxicity against K562 and COLO tumor targets. This cytotoxic activity could be dramatically enhanced by incubation with additional rIL-2 in vitro. The immunologic effects of rIL-2 treatment were similar in both autologous and allogeneic marrow recipients. Our data suggest that prolonged infusion of rIL-2 at low doses is safe and can selectively increase NK cell number and activity after BMT. Further studies to assess the impact these changes may have on disease relapse post-BMT will be undertaken.