Twenty patients with disseminated melanoma were treated with interferon alfa-2a, given by intramuscular (IM) injection three times a week in escalating doses from 15 to 50 X 10(6) U/m2. Of 18 patients considered evaluable, two had complete remission and in two others the disease was stabilized. Laboratory tests 6 hours after injection of interferon alfa-2a indicated a marked lymphopenia and a reduction in natural killer (NK) cell activity. Sequential changes (measured before injection of interferon alfa-2a on days 3, 10, and 31) consisted of neutropenia, thrombocytopenia, and a slight increase in OKT4 positive T cells compared with OKT8 positive T cells. NK activity against the K562 target cells was increased in most patients during the first week of treatment, returning to near or below pretreatment levels thereafter. This response contrasted with a delayed increase against melanoma target cells in 10 patients. The latter correlated with an increase in mitogen-stimulated interleukin-2 (IL2) production, and may indicate that the cytotoxic activity resulted from lymphokine-activated killer (LAK) cells. Changes in cortisol levels may explain some effects on the immune system, such as depression of IL2 and immunoglobulin production in vitro, and the differences noted in clinical responses during the present study compared with those observed with interferon alfa-2b given by intravenous (IV) injection in 5-day cycles. These results suggest that interferon alfa-2a has antitumor activity in certain melanoma patients, in particular those with metastases to pulmonary or subcutaneous sites. Assays of IL2 production and LAK activity may assist in the selection of patients who respond to interferon alfa-2a and help to optimize treatment regimens.