Loss of wild-type p53 activity is thought to be a major predictor of failure to respond to radiotherapy and chemotherapy in various human cancers. This assumption is largely based on some cell-death studies in p53-knockout mice and on correlations of p53 status assessed by immunochemistry or single-strand conformational polymorphism (SSCP) analysis, and responses to therapy in human cancers in vivo. In principle, p53 may enhance chemosensitivity by promoting apoptosis via transcription-independent mechanisms as well as transcriptional activation of proapoptotic genes such as bax and transcriptional repression of antiapoptotic genes such as bcl-2. Drug-induced suicide mediated by the CD95/CD95 ligand system may also involve a p53-controlled pathway. Yet, p53 may decrease chemosensitivity by promoting p21-mediated and p21-independent growth arrest, DNA repair, and differentiation, and by enhancing the transcription of antiapoptotic genes such as bcl-x. Cell-culture work indicates that the effects of altering the p53 status on chemosensitivity depend very much on the cellular context. Disruption of p53 function in otherwise normal, nonneoplastic cells may enhance rather than decrease chemosensitivity. However, targeted p53 gene disruption in some cell types obtained from p53-knockout mice results in enhanced rather than decreased sensitivity, e.g., to irradiation. Transformed cells that have retained wild-type p53 function tend to acquire chemoresistance when p53 function is disabled, with few exceptions. Thus, preexisting molecular alterations or consecutive accumulation of molecular alterations after loss of p53 rather than the loss of wild-type p53 activity per se may confer chemoresistence to tumor cells. Moreover, p53 accumulation resulting from the increased half-life of mutant p53 proteins can act as a gain-of-function mutation, presumably as a consequence of multiple protein-protein interactions. Finally, significant tumor cell-type- and drug-specific patterns of modulation of chemosensitivity by p53 are beginning to emerge. Transfer of wild-type p53 genes into tumor cells commonly induces growth arrest but may render these cells relatively more resistant to most chemotherapeutic drugs. Therefore, careful experimental in vitro and in vivo studies are required before chemotherapy-supported p53 gene therapy for human cancer is introduced into clinical practice.