Several genetic diseases caused by deficiencies in enzymes required throughout the body can be treated by enzyme replacement therapy at costs of upward of hundreds of thousands of dollars per year per patient. In the long run, such costs are prohibitive. Gene therapy is therefore an appealing option in that its aim is to treat patients after only a single or very few interventions. Clinical-grade viral vectors can efficiently deliver therapeutic expression cassettes, but such viruses can still be expensive to prepare. By contrast, clinical-grade nonviral nucleic acids are relatively inexpensive to produce but difficult to deliver into cells. In this issue of Molecular Therapy, two reports from the same group describe improvements to two limiting factors of nonviral gene therapy using DNA transposons delivered to the liver.1,2 The authors substantially increased efficiency of delivery of transgenes to the target organ and the amount of therapeutic protein exported per cell following transgene delivery. In the first paper, the authors show improved delivery of transgenic expression cassettes in piggyBac transposons as well as significantly increased output per expression cassette. First they developed an in silico bioinformatic strategy to identify transcriptional enhancers—which they refer to as CRMs (cis-regulatory modules)—that can direct elevated levels of gene expression in the liver.1 Although it is well known that there is a >1,000-fold range in gene expression, with about a dozen genes expressed at extremely high levels in mouse liver,3 transcriptional control elements can reside far away from the promoters they regulate.4 The strategy identified eight different transcription factors and their DNA-binding sites. The authors settled for using the minimal promoter of the liver-specific transthyretin gene (TTR), to which short assemblies of CRMs were joined to form hepatocyte-specific CRMs (HS CRMs). Fourteen HS CRMs were tested, and most of them raised FIX gene expression between 10- and 100-fold when delivered as plasmids to the liver using hydrodynamic infusion.5,6 These same assemblies are active in adeno-associated viral vectors.7 Although this rational design of vectors has significantly increased the prospects for successful DNA-based, liver-directed gene therapy, there is room for improvement. Linking enhancers without regard for the constraints of the double-helical geometry of DNA is likely to diminish the strength of the interactions. The authors mention that altering the spacing of the CRMs relative to the promoter would probably augment gene expression, but such alterations can be guided by taking into consideration the geometry of the double helix, in which the relative rotational positioning of enhancers is critical to optimal performance.8 Thus, even the next steps of fine-tuning are open to rational design. In the second study, the group reports the effects of coupling of the HS-CRM transcriptional drivers to a “codon-optimized” version of the hyperactive FIXR338L(Padua) gene and insertion of the therapeutic expression cassette into an improved piggyBac (PB) transposon system.2 PB, like the Sleeping Beauty transposon system, is a mobile element that can move an expression cassette from a donor plasmid into a chromosome of a recipient cell via a “cut-and-paste” mechanism. Figure 1 shows the essence of transposon-mediated gene therapy in which a complete transposon system is delivered to the liver (shown here via venous infusion), followed by transposition to chromosomes that will support sustained gene expression. The cognate transposase enzyme directs the excision (“cut”) and insertion (“paste”) into a new DNA sequence. In this study, the complete transposon system led to levels of FIX gene expression following nonviral delivery to mouse liver that were up to 10-fold greater than FIX levels in normal mice and more than 100-fold greater than levels from earlier expression constructs. Improvements to the piggyBac transposon system included use of a hyperactive transposase and truncation of the inverted terminal repeat sequences (inverted blue arrows in Figures 1 and ​22) of the transposon that are recognized by the transposases for mobilization. Figure 2 illustrates several aspects of gene expression from transposons described by these authors2 and others.9 Whereas the HS CRMs or transcriptional regulatory motifs (TRMs) are designed to provide sustained expression for up to a lifetime, the promoter for the transposase gene should have transient activity to provide a single round of transposition. Generally a cytomegalovirus promoter that is active only for a few days in hepatocytes following hydrodynamic infusion is used to direct transcription of the transposase gene so as to limit further transposition.10 Figure 1 Transposon-mediated gene therapy. A plasmid with a transposon (inverted blue arrows) containing a liver-specific promoter (green “LSP” arrowhead) and a therapeutic gene is infused with another plasmid containing a transposase gene (Txp) ... Figure 2 Improved transposons and expression cassettes for transposon-mediated gene therapy to the liver. Expression of transgenes from plasmids in the absence of both a transposase source (–Txp) and immune response inhibitors (IRIs) is very short-lived ... As with the first paper, there is room for improvement by tweaking the components in their liver-specific, FIX expression cassette. A surprise in the results was the effect of codon optimization (CO) of the FIX coding sequence (FIXCO). Generally, CO refers to improving translational rates by replacing certain codons for which there are minimal levels of charged transfer RNAs that consequently slow the elongation of the nascent polypeptide chain as the ribosome waits for a charged transfer RNA. However, codon optimization in these studies seemingly had the opposite effect: translational efficiency—the ratio of FIXCO protein to FIXCO messenger RNA (mRNA)—actually dropped by about fivefold rather than increasing (comparison of protein synthesis in panels a and b of their Figure 3 to the respective mRNA levels in panel 3d). It is likely that “codon optimization” increased the stability of the FIXCO mRNA nearly 30-fold. As a result, there was a net increase in gene expression of about sixfold despite the fivefold decrease in translational efficiency. Because initiation of translation is often the rate-limiting step in translational efficiency of an mRNA, CO may not even be possible per se, even though the optimization step leads to alterations in mRNA secondary structure that renders the mRNA far less susceptible to RNase degradation.11 Thus, further tweaking (and de-tweaking) the FIXCO sequence may retain the stabilization properties while maintaining, or even improving, translational efficiency. The overall improvement of gene expression from a transposon vector is highly significant for five reasons. First, because the efficiency of delivery of plasmids by hydrodynamic infusion drops considerably in larger animals,12 delivery of highly expressing therapeutic cassettes is vital to compensate for the fewer transgenes that direct production of the desired product. Second, the concerns of insertional mutagenesis, which increase dramatically with the number of mutations (insertions) per genome,13 effectively limit the number of integrations of therapeutic cassettes to one per genome. Third, for some systemic diseases, correction of diffuse, global pathology is difficult to achieve unless persistent activities of the therapeutic enzyme of 10- to 100-fold greater than normal are attained in the circulation.14,15 Fourth, the use of liver-specific transcriptional elements should restrict expression of transgenes to hepatocytes; this could significantly reduce immune responses that obstruct gene therapy.16 Fifth, the more impact per vector, the less expensive the treatments will be. In summary, the improvements described in these two papers demonstrate that sustained supranormal levels of many therapeutic proteins should be achievable using transposons—either piggyBac or Sleeping Beauty—to support affordable treatment of certain genetic diseases. The major challenge now is scaling up the hydrodynamic infusion procedure for effective delivery to liver in large animals for transgene expression over a lifetime.12