Intratumoral injection of recombinant human tumor necrosis factor (TNF) for inoperable pancreatic cancer has shown some efficacy in suppressing tumor growth or decreasing tumor markers. However, complete regression has not yet been achieved, possibly due to a lack of TNF receptors on tumor cells or an abundance of intracellular resistance factors. Recently, two distinct types of TNF receptors, R55 and R75, were identified, which are responsible for signaling of cytotoxicity and of proinflammation, respectively. In this study, a novel type of suicide gene therapy is proposed that is based on transfection of the R55 gene into human pancreatic cancer cells (AsPC-1 and PANC-1) and subsequent administration of TNF. The transfectants from both cell lines showed higher TNF susceptibility than their parental cells. In vivo tumor formation of an AsPC-1 clone (clone 10) inoculated in nude mice was substantially suppressed by administration of TNF. For practical use of this strategy, however, the adverse effects of TNF may become an obstacle. We previously produced mutein TNF 471, which had a higher affinity for R55, superior antitumor activity, and fewer adverse effects. This mutein TNF 471 manifested greater antitumor activity against clone 10. Because the R55 receptor is known to be involved in augmentation of cellular immunity by TNF, mutein TNF 471 is also expected to be highly potent in this function. In fact, the mutein TNF 471 induced higher splenic natural killer cell activity in nude mice inoculated with clone 10 than did native TNF. This property of augumenting cellular responses may be advantageous in the eradication of viable tumor cells left untransfected in practical gene therapy regimens in which 100% transfection of the R55 gene into tumors is not feasible. Thus, gene therapy combining transfection of the TNF-R55 gene with administration of mutein TNF 471 may provide a new modality for the treatment of pancreatic cancer.