Recombinant human tumor necrosis factor-alpha (TNF) has had only modest success in clinical trials, in part because of dose limiting toxicities. To more fully realize TNF's clinical activity it will be necessary to develop more potent TNF analogues possessing fewer untoward effects. Several studies have found that TNF cytotoxicity depends on TNF internalization and processing by the cell. We compared the processing of TNF by TNF-sensitive and resistant cancer cells of breast and prostate origins. Sensitive cells were found to process internalized TNF to 15 kD and 5.5 kD species, while the resistant cell lines produced multiple lower molecular weight TNF products. We also evaluated the toxicity of conditioned medium from sensitive cells to determine if media containing the degradation products was toxic to resistant cells. Hydrophobic interaction column HPLC was employed to isolate and purify the degradation products for evaluation of their cytotoxic activity. Conditioned media and the HPLC fractions of conditioned media enriched for the 5.5 kD TNF fragment were toxic to both TNF-sensitive and -resistant tumor cells. Phase I of this project would isolate, sequence and synthesize the 5.5 kD degradation product and assay its antineoplastic activity against fresh human tumors of various histology's. Phase II of this project would evaluate the toxicity profile and anti-tumor action of this peptide in vivo in murine models if Phase I results are encouraging. The ultimate goal would be to develop a new chemotherapeutic agent.