UCSF

Adenovirus (Ad) is the most commonly used vector in gene therapy trials worldwide. Therefore, understanding the interaction between the virus and the cell surface and how this interaction impacts cell infection is of great importance to both the analysis of current trials using Ad vectors and the design of next generation Ad vectors. Cancer is in particular an important target of Ad-based therapeutics. Therefore, we have measured the ability of a replication incompetent subgroup C adenovirus (Ad5-GFP) to infect a panel of cancer cell lines. Infection across this panel was highly variable. Coxsackie and Adenovirus Receptor (CAR) is the known cellular receptor for subgroup C adenoviruses so we hypothesized that varying levels of CAR on these cell lines would explain the varying infections. However, neither CAR mRNA levels, as measured by both Affymetrix array and QT-PCR, nor CAR protein levels, as measured by both FACS and western, correlated with infection. One cell line, MDA MB 435, is CAR negative by all criteria that we have measured, but is one of the most infectible cell lines on the panel. Additionally, MCF7 cells and WM278 cells have minimal surface CAR but are infectible. Ad5 binds to these cell lines via a high affinity (0.16 nM) interaction. Surprisingly, the infection of CAR-negative cells is fiber-independent, as determined by competition experiments using soluble fiber. Because the penton base of the Ad virion is known to interact with RGD-binding integrins, we examined the ability of an RGD peptide to block the binding of Ad5 to CAR-negative cells. We found that Ad5 is using an RGD-binding integrin as a primary receptor for binding and infection. We then utilized blocking antibodies to determine which integrins are involved. We found that a blocking antibody to integrin αvβ5 blocks Ad5 from binding to CAR-negative cells lines. We conclude that integrin αvβ5 is an alternate attachment receptor for Ad5, representing a previously unidentified entry pathway for Ad5 that is both CAR and fiber-independent.