Lymph nodes (LN) draining progressively growing tumors contain tumor-sensitized but not fully functional preeffector lymphocytes. These cells could acquire therapeutic efficacy and be expanded upon sequential culture with anti-CD3 mAb for 2 days followed by incubation in IL-2 for 3 days. Using the weakly immunogenic MCA 106 and MCA 205 murine sarcomas, we have further defined conditions of this anti-CD3/IL-2 activation with which preeffector cells differentiated into immune effector cells. In vitro activation and expansion of effector cells required sequential but independent stimulation with anti-CD3 and IL-2 because the simultaneous presence of both anti-CD3 and IL-2 at either stage did not enhance the efficacy of activation. Generation of effector cells by this two-stage activation was critically dependent on the optimal concentrations of anti-CD3 (1.0 microgram/ml) and IL-2 (2-10 U/ml). However, these conditions were not optimal for inducing the greatest cellular proliferation. In adoptive immunotherapy experiments, although the transfer of anti-CD3/IL-2-activated cells alone could mediate the regression of established metastases, the concomitant administration of IL-2 enhanced the in vivo activity of these cells. More importantly, tumor regression mediated by the anti-CD3/IL-2-activated cells was found to be immunologically specific. The specificity was determined by the tumor that stimulated the preeffector cell response. In spite of their in vivo antitumor effects, the anti-CD3/IL-2-activated tumor-draining LN cells did not exhibit detectable in vitro cytotoxicity against the tumor target in the 4-h 51Cr-release assay. In mice bearing progressive tumor, draining LN contained most preeffector cells. Some preeffector cells were also detected in the spleen whereas mesenteric LN did not demonstrate any reactivity. In kinetics studies, sensitization of preeffector cells in the draining LN occurred between 4 to 6 days after tumor inoculation. As the tumor progressed, the presence of preeffector cells declined gradually suggesting a tumor-induced suppression. These results define the conditions whereby tumor-draining LN cells could be stimulated, in the absence of tumor Ag, to develop into specific therapeutic effector cells. Our findings also raise the possibility of using similar approaches for isolating immune effector cells from cancer patients for adoptive immunotherapy.