A phase I trial of simultaneously administered recombinant interleukin-2 (rIL-2) and recombinant human IL-4 (rHuIL-4) was conducted to evaluate the toxicity and the clinical and immunologic effects of this cytokine combination. Thirty-nine eligible patients with refractory malignancy were treated at eight different dose levels (1A to 3B): 1-3 of rIL-2 [3.0, 12.0, and 48.0 x 10(6) IU/m(2) i.v. three times weekly (TIW)] and A-C of rHuIL-4 (40, 120, and 400 mu g/m(2) s.c. TIW). The toxicity of these two cytokines was moderate and was comparable with that seen with rIL-2 alone. The maximal tolerated dose (MTD) of the combination was not reached because of lack of sufficient rHuIL-4 but is at least 48.0 x 10(6) IU/m(2) of rIL-2 and 120 mu g/m(2) of rHuIL-4. Two patients with melanoma had partial responses. The immunologic effects included increases in absolute lymphocyte numbers, and the CD3- /CD56+/ CD2+, total CD56+, CD8+, and CD16c+ lymphocyte subsets with increasing rIL-2 dose levels, but not with rHuIL-4. This increase in natural killer (NK) cells in the peripheral blood was accompanied by an increase over baseline in NK lytic activity against K562 targets; however, concomitant increases in lymphokine-activated killer (LAK) activity (Daudi targets) were not seen. The CD3+, CD4+, and CD3+/CD25+/HLA-Dr+ T-cell subsets also increased, and these increases were related to both increasing rIL-2 and rRuIL-4 doses. Finally, in four of six patients, serial tumor biopsies demonstrated increases in major histocompatibility complex (MHC) class I or II antigen expression on tumor cells or increasing T-cell infiltrates during cytokine therapy or both. This trial demonstrated that rIL-2 and rHuIL-4 can be administered simultaneously with acceptable toxicity. The immunologic findings demonstrated the expected rIL-2-associated increases of CD56+ and CD16c+ lymphocytes and NK activity, and interestingly, no development of LAK activity. These findings suggest regulatory effects of rHuIL-4 on rIL-2-related effects in vivo.