Interleukin-2 (IL-2) at high doses or at low doses in concert with lymphokine-activated killer (LAK) cells can produce regression of established pulmonary and hepatic metastases from a variety of tumors in mice. IL-2 appears to mediate its antitumor effect through the generation of LAK cells in vivo from endogenous lymphocytes and by the stimulation of host and transferred LAK cell proliferation in tissues. In this paper we have investigated different strategies for IL-2 administration to determine which regimen produced maximal in vivo proliferation and optimal immunotherapeutic efficacy of LAK cells. Tissue expansion of lymphoid cells was assessed using an assay of in vivo labeling of dividing cells by the thymidine analogue, 5-[125I]iododeoxyuridine. The therapeutic effect of the different IL-2 administration protocols was determined by evaluating their efficacy in the treatment of established, 3-day pulmonary metastases from sarcomas in mice. The selection of IL-2 injection regimens for evaluation was based upon pharmacokinetic studies of IL-2 in mice. A single i.v. or i.p. dose yielded high peak IL-2 levels that could be measured for only a few hours after injection, while IL-2 given i.p. thrice daily produced titers that were detectable throughout the study periods (greater than or equal to 6 units/ml of serum after 100,000 units of IL-2 i.p. thrice daily). Using the proliferation and therapy models, we tested the same cumulative daily doses of IL-2 administered by i.v. or i.p. once daily, or i.p. thrice daily regimens. The i.p. thrice daily protocol stimulated greater lymphoid cell proliferation in the lungs, for example, than did the other regimens. Similarly, 300,000 units of IL-2 divided i.p. thrice daily were more successful in reducing metastases (n = 16) than was the entire dose given i.v. once daily (n = 190; P less than 0.05) or i.p. once daily (n = 71; P less than 0.05). When compared to the i.p. or i.v. once daily protocols, the i.p. thrice daily regimen for IL-2 also produced greater proliferation of exogenous LAK cells, as well as a more effective therapeutic outcome when IL-2 was combined with transferred LAK cells. Thus, sustained, lower levels of IL-2 were more effective than brief, high peak titers for stimulation of proliferation and antitumor activity. We then evaluated the effect of duration of IL-2 treatment as well as the number of LAK cell injections in the two models.(ABSTRACT TRUNCATED AT 400 WORDS)