Currently used vectors in human gene therapy suffer from a number of limitations with respect to safety and reproducibility. The ideal vector has to be free of these effects and should allowlong-term expression of a transgene in the absence of selection. Recently a novel non-viral episomal expression system which fulfils these criteria, in principle, was reported. This vector, pEPI, replicates episomally in CHO, HeLa, K562, as well as in primary human fibroblast-like cells, and is mitotically stable in the absence of selection for more than 100 generations. Here we show the long term expression of pEPI in a range of cells including cystic fibrosis airway (IB3) cells to be stable for at least 10 weeks. We demonstrate, however, that antibiotic selection pressure, is necessary for the first weeks following transfection. FACS sorting of CHO and IB3 cells, show that expression and maintenance in cells without initial antibiotic selection diminished rapidly, whereas constructs under selection pressure for an initial period of 2 weeks were mitotically stable and expressed transgene for at least 10 weeks, even following the removal of selection pressure. However, initial in vivo results showed no persistance of vector and limited expression of the reporter gene when using viral promoters. We describe the developments we have made to overcome these limitations and the results of the application of an S/MAR plasmid carrying a human promoter and chromosomal destabilising elements to drive the expression and episomal maintenance of these vectors in vivo. Currently used vectors in human gene therapy suffer from a number of limitations with respect to safety and reproducibility. The ideal vector has to be free of these effects and should allowlong-term expression of a transgene in the absence of selection. Recently a novel non-viral episomal expression system which fulfils these criteria, in principle, was reported. This vector, pEPI, replicates episomally in CHO, HeLa, K562, as well as in primary human fibroblast-like cells, and is mitotically stable in the absence of selection for more than 100 generations. Here we show the long term expression of pEPI in a range of cells including cystic fibrosis airway (IB3) cells to be stable for at least 10 weeks. We demonstrate, however, that antibiotic selection pressure, is necessary for the first weeks following transfection. FACS sorting of CHO and IB3 cells, show that expression and maintenance in cells without initial antibiotic selection diminished rapidly, whereas constructs under selection pressure for an initial period of 2 weeks were mitotically stable and expressed transgene for at least 10 weeks, even following the removal of selection pressure. However, initial in vivo results showed no persistance of vector and limited expression of the reporter gene when using viral promoters. We describe the developments we have made to overcome these limitations and the results of the application of an S/MAR plasmid carrying a human promoter and chromosomal destabilising elements to drive the expression and episomal maintenance of these vectors in vivo.