Over the past few years, we and a number of other groups have conducted laboratory experiments and clinical trials of human recombinant interleukin-2 (rIL-2) alone or in combination with autologous activated lymphocytes expressing in vitro tumoricidal activity in order to define toxicity and indicate its potential efficacy in patients with high-grade glioma. Because high rIL-2 concentrations can be attained with considerably less toxicity than with a systemic approach, all of the clinical trials, to date, have chosen a direct route; injecting lymphokine and cells into tumor tissue, the cystic cavity remaining after tumor excision, and/or neural parenchyma surrounding the site of tumor excision. While the rIL-2 therapies, as they have been applied in animal glioma models and patients, are safe, cerebral edema around the site of treatment has been a consistent finding. We have also seen, however, that steroid medications used by patients to control their cerebral edema may depress the anti-tumor activity of rIL-2 by depressing the capacity of lymphocytes to develop normal LAK activity. Although none of the immunotherapies involving rIL-2 have produced cures, the fact that sustained clinical responses have been reported, suggests that such therapies may slow a recurrence of tumor at the site of treatment. Efforts to improve outcome from rIL-2 - based immunotherapies for malignant glioma are continuing with manipulation of rIL-2 dosing and scheduling and also with combinations of rIL-2 and other recombinant cytokines.