Clinical trials are being performed using tumor genetically engineered to produce cytokines as a vaccine. The design of such a vaccine may be made more effective by further study using in-vitro as well as in-vivo models. We studied an in-vitro tumor 'vaccine' model in glioblastoma. We have demonstrated high efficiency transfection of the Interleukin-2 (IL-2) gene into glioblastoma cell lines using adenoviral vectors. Glioblastoma cell lines transduced with this vector could produce high levels of IL-2 for up to 2 weeks, long enough to elicit an antitumor immune response. We studied tumor/effector cell interactions using cytotoxicity assays coupled with flow cytometric analysis. Activation of CD8+ and expansion of CD3+/CD16+ effector cell subpopulations were observed, suggesting the generation of a specific anti-tumor response and the potential for systemic immunity. We demonstrated that glioblastoma produce immunosuppressive factors which reduce the antitumor response by peripheral blood effector cells. These immunosuppressive factors could be neutralized to improve antitumor response. A better understanding of tumor/effector cell interactions may improve the design of gene therapy trials.