Three new prodrugs for suicide gene therapy using carboxypeptidase G2 elicit bystander efficacy in two xenograft models.

Three new prodrugs, [prodrug 1: 4-[bis(2-iodoethyl)amino]-phenyloxycarbonyl-L-glutamic acid; prodrug 2: 3-fluoro-4-[bis(2-chlorethyl)amino]benzoyl-L-glutamic acid; and prodrug 3: 3,5-difluoro-4-[bis(2-iodoethyl)amino]benzoyl-L-glutamic acid] have been assessed for use with a mutant of carboxypeptidase G2 (CPG2, glutamate carboxypeptidase, EC 3.4.17.11,) engineered to be tethered to the outer tumor cell surface (stCPG2(Q)3) as the activating enzyme in suicide gene therapy systems. All three of the prodrugs produce much greater cytotoxicity differentials between stCPG2(Q)3- and control beta-galactosidase (beta-gal)-expressing breast carcinoma MDA MB 361 and colon carcinoma WiDr cells (70- to 450-fold) than was previously observed (19- to 27-fold) with 4-[(2-chloroethyl)(2-mesyloxyethyl)amino]benzoyl-L-glutamic acid (CMDA). Prodrug 1 is the most effective antitumor agent in xenografts in mice inoculated with 100% stCPG2(Q)3-expressing MDA MB 361 cells, whereas prodrugs 2 and 3 are most effective when the percentage of stCPG2(Q)3-expressing cells is 50% or 10%. In nude mice bearing xenografts arising from inocula of 100% stCPG2(Q)3-expressing WiDr cells, prodrug 2 is the most effective antitumor agent. All three of the prodrugs produced histological evidence of substantial bystander cell killing in WiDr xenografts in which only 10% or 50% of the cells inoculated were expressing stCPG2(Q)3. We conclude that all three of the prodrugs are more effective therapeutically with stCPG2(Q)3 than is the previously described prodrug CMDA and, also, that the optimal choice of prodrug varies among different tumor types and that prodrugs, optimized for their bystander effect, are effective when only low percentages of cells in a tumor express CPG2.