In vivo retargeting of adenovirus type 5 to αvβ6 integrin results in reduced hepatotoxicity and improved tumor uptake following systemic delivery

A key impediment to successful cancer therapy with adenoviral vectors is the inefficient transduction of malignant tissue in vivo. Compounding this problem is the lack of cancer-specific targets, coupled with a shortage of corresponding high-efficiency ligands, permitting selective retargeting. The epithelial cell-specific integrin αvβ6 represents an attractive target for directed therapy since it is generally not expressed on normal epithelium but is upregulated in numerous carcinomas, where it plays a role in tumor progression. We previously have characterized a high-affinity, αvβ6-selective peptide (A20FMDV2) derived from VP1 of foot-and-mouth disease virus. We generated recombinant adenovirus type 5 (Ad5) fiber knob, incorporating A20FMDV2 in the HI loop, for which we validated the selectivity of binding and functional inhibition of αvβ6. The corresponding αvβ6-retargeted virus Ad5-EGFP_A20 exhibited up to 50-fold increases in coxsackievirus-and-adenovirus-receptor-independent transduction and up to 480-fold-increased cytotoxicity on a panel of αvβ6-positive human carcinoma lines compared with Ad5-EGFPWT. Using an αvβ6- positive (DX3-β6) xenograft model, we observed a ∼2-fold enhancement in tumor uptake over Ad5-EGFP_WT following systemic delivery. Furthermore, ∼5-fold-fewer Ad5-EGFP_A20 genomes were detected in the liver (P = 0.0002), correlating with reduced serum transaminase levels and E1A expression. Warfarin pretreatment, to deplete coagulation factors, did not improve tumor uptake significantly with either virus but did significantly reduce liver sequestration and hepatic toxicity. The ability of Ad5-EGFP _A20 to improve delivery to αvβ6, combined with its reduced hepatic tropism and toxicity, highlights its potential as a prototype virus for future clinical investigation.