Using phage-displayed random enzyme libraries, our goal is to develop tumor-targeted effector molecules that convert prodrugs to toxins for cancer therapy. The rationale for this approach is based on published experiments with single-chain antibody (scFv) targeted enzymes that specifically activate prodrugs of potent cytotoxic agents. These prodrugs have significantly less cytotoxicity than the toxins released by cleavage with the targeted enzyme. We hypothesize that phage enzymes with the targeting moiety built directly into the protein scaffold when panned against tumor cells isolated from patients following surgical resection will contain tumor selective effector molecules. The benefits of panning enzyme libraries on fresh human tumor cells include the presence of a vast number of possible targets, targets will be in their native configuration, subtraction of the pools of ligands binding to normal tissue, and the ability to rapidly test these novel targeting agents for efficacy using a growing portfolio of prodrugs.
The Specific Aims are 1) Generate diverse randomized loop libraries of enzymes on phage 2) Surgically sample tumor tissue, isolate small numbers of fresh tumor cells, and identify phage enzymes that have bound selectively to tumor cells using a novel method of PCR-based phage display, 3) Determine whether candidate enzymes bind selectively to tumor tissue specimens and not to normal tissues using a novel fluorometric substrate, and 4) Characterize targeted enzyme activation of cytotoxic prodrugs in vitro and in vivo in animal models. Successful completion of these experiments will result in demonstration of novel targeted effector molecules that bind selectively to tumors and activate cytotoxic prodrugs.
