Abstract
Reovirus is undergoing clinical testing as an oncolytic therapy for breast cancer. Given that reovirus naturally evolved to thrive in enteric environments, we sought to better understand how breast tumor microenvironments impinge on reovirus infection. Reovirus was treated with extracellular extracts generated from polyomavirus middle T-antigen-derived mouse breast tumors. Unexpectedly, these breast tumor extracellular extracts inactivated reovirus, reducing infectivity of reovirus particles by 100-fold. Mechanistically, inactivation was attributed to proteolytic cleavage of the viral cell attachment protein σ1, which diminished virus binding to sialic acid (SA)-low tumor cells. Among various specific protease class inhibitors and metal ions, EDTA and ZnCl2 effectively modulated σ1 cleavage, indicating that breast tumor-associated zinc-dependent metalloproteases are responsible for reovirus inactivation. Moreover, media from MCF7, MB468, MD-MB-231, and HS578T breast cancer cell lines recapitulated σ1 cleavage and reovirus inactivation, suggesting that inactivation of reovirus is shared among mouse and human breast cancers and that breast cancer cells by themselves can be a source of reovirus-inactivating proteases. Binding assays and quantification of SA levels on a panel of cancer cells showed that truncated σ1 reduced virus binding to cells with low surface SA. To overcome this restriction, we generated a reovirus mutant with a mutation (T249I) in σ1 that prevents σ1 cleavage and inactivation by breast tumor-associated proteases. The mutant reovirus showed similar replication kinetics in tumorigenic cells, toxicity equivalent to that of wild-type reovirus in a severely compromised mouse model, and increased tumor titers. Overall, the data show that tumor microenvironments have the potential to reduce infectivity of reovirus.IMPORTANCE We demonstrate that metalloproteases in breast tumor microenvironments can inactivate reovirus. Our findings expose that tumor microenvironment proteases could have a negative impact on proteinaceous cancer therapies, such as reovirus, and that modification of such therapies to circumvent inactivation by tumor metalloproteases merits consideration.