Mode of transgene expression after fusion to early or late viral genes of a conditionally replicating adenovirus via an optimized internal ribosome entry site in vitro and in vivo

The expression of therapeutic genes by oncolytic viruses is a promising strategy to improve viral oncolysis, to augment gene transfer compared with a nonreplicating adenoviral vector, or to combine virotherapy and gene therapy. Both the mode of transgene expression and the locale of transgene insertion into the virus genome critically determine the efficacy of this approach. We report here on the properties of oncolytic adenoviruses which contain the luciferase cDNA fused via an optimized internal ribosome entry site (IRES) to the immediate early adenoviral gene E1A (AdΔE1AIL), the early gene E2B (AdΔE2BIL), or the late fiber gene (AdΔfiberIL). These viruses showed distinct kinetics of transgene expression and luciferase activity. Early after infection, luciferase activities were lower for these viruses, especially for AdΔE2BIL, compared with nonreplicating AdTL, which contained the luciferase gene expressed from the strong CMV promoter. However, 6 days after infection, luciferase activities were approximately four (AdΔE1AIL) to six (AdΔfiberIL) orders of magnitude higher than for AdTL, reflecting virus replication and efficient transgene expression. Similar results were obtained in vivo after intratumoral injection of AdΔE2BIL, AdΔfiberIL, and AdTL. AdΔfiberIL and the parental virus, Ad5-Δ24, resulted in similar cytotoxicity, but AdΔE2BIL and AdΔE1AIL were slightly attenuated. Disruption of the expression of neighboring viral genes by insertion of the transgene was minimal for AdΔE2BIL and AdΔfiberIL, but substantial for AdΔE1AIL. Our observations suggest that insertion of IRES-transgene cassettes into viral transcription units is an attractive strategy for the development of armed oncolytic adenoviruses with defined kinetics and strength of transgene expression.