One aim of the genetic modification of tumor cells is the generation of immunogenic variants that can be used for the induction of immune responses against tumors. We engineered the human colorectal carcinoma cell line SW480 by means of plasmid transfection to secrete interleukin (IL)-2. Transfection of SW480 cells resulted in stable IL-2 secretion at 5-30 ng/ml per 10(5) cells in 24 h and, unexpectedly, in CD54 expression on the cell surface. SW480 variants expressing IL-2 and CD54 were tested for their capacity to induce T lymphocyte activation in vitro in comparison to untransfected and CD54 transfected cells. The cytolytic effector function of a class I MHC restricted CD8+, peptide antigen specific T cell clone was augmented following expression of CD54. IL-2 secreting SW480 variants did not further increase antigen-dependent cytolysis. Primary activation of resting T lymphocytes was assessed following allogeneic stimulation. When compared with unmodified SW480 cells, CD54 expressing variants did not initiate T cell proliferation. In contrast, IL-2 secreting SW480 cells strongly promoted primary T cell proliferation. Similarly, exogenous IL-2 and SW480 cells induced T cell proliferation which was not only due to IL-2 but was dependent on tumor cells. However, following the initial wave of cell growth in response to IL-2 secreting SW480 cells T lymphocytes could not be restimulated with SW480 or IL-2 secreting variants and could not be further expanded. T cells initially activated by IL-2 secreting SW480 cells exhibited cytolytic activity towards SW480 cells. This reactivity, however, was transient and completely blocked by K562 cells, suggesting MHC-unrestricted, nonspecific cytotoxicity. We conclude that endogenous IL-2 secretion by the colorectal carcinoma cell line SW480 does not result in the activation of MHC restricted specific T lymphocytes but predominantly induces lymphokine-activated killer cells. Considering that tumor cell vaccines are aimed at inducing tumor-specific immune responses, our in vitro observation would rather argue against the in vivo application of such a tumor cell modification in colorectal cancer.