Long term expression of bicistronic vector driven by the FGF-1 IRES in mouse muscle

Background: Electrotransfer of plasmid DNA into skeletal muscle is a promising strategy for the delivery of therapeutic molecules targeting various muscular diseases, cancer and lower-limb ischemia. Internal Ribosome Entry Sites (IRESs) allow co-expression of proteins of interest from a single transcriptional unit. IRESs are RNA elements that have been found in viral RNAs as well as a variety of cellular mRNAs with long 5' untranslated regions. While the encephalomyocarditis virus (EMCV) IRES is often used in expression vectors, we have shown that the FGF-1 IRES is equally active to drive short term transgene expression in mouse muscle. To compare the ability of the FGF-1 IRES to drive long term expression against the EMCV and FGF-2 IRESs, we performed analyses of expression kinetics using bicistronic vectors that express the bioluminescent renilla and firefly luciferase reporter genes. Long term expression of bicistronic vectors was also compared to that of monocistronic vectors. Bioluminescence was quantified ex vivo using a luminometer and in vivo using a CCD camera that monitors luminescence within live animals. Results: Our data demonstrate that the efficiency of the FGF-1 IRES is comparable to that of the EMCV IRES for long term expression of bicistronic transgenes in mouse muscle, whereas the FGF-2 IRES has a very poor activity. Interestingly, we show that despite the global decrease of vector expression over time, the ratio of firefly to renilla luciferase remains stable with bicistronic vectors containing the FGF-1 or FGF-2 IRES and is slightly affected with the EMCV IRES, whereas it is clearly unstable for mixed monocistronic vectors. In addition, long term expression more drastically decreases with monocistronic vectors, and is different for single or mixed vector injection. Conclusion: These data validate the use of bicistronic vectors rather than mixed monocistronic vectors for long term expression, and support the use of the FGF-1 IRES. The use of a cellular IRES over one of viral origin is of particular interest in the goal of eliminating viral sequences from transgenic vectors. In addition, the FGF-1 IRES, compared to the EMCV IRES, has a more stable activity, is shorter in length and more flexible in terms of downstream cloning of second cistrons. Finally, the FGF-1 IRES is very attractive to develop multicistronic expression cassettes for gene transfer in mouse muscle.