Considerations for a Successful RNA Trans-splicing Repair of Genetic Disorders

To the Editor: Spliceosome-mediated RNA trans-splicing (SMaRT) has emerged as a novel technology for the repair of mutations in monogenic disorders. SMaRT can be used to replace 5′, 3′, or internal sequences in an exon-wise manner. For this purpose, RNA trans-splicing molecules (RTMs) are engineered, which comprise (i) the wild-type coding region to be replaced, (ii) a splicing domain that includes essential splicing elements such as branch point and polypyrimidine tract, (iii) a binding domain (BD) hybridizing to a selected target region to bring RTM and target transcript into close proximity, and (iv) a spacer sequence for steric reasons. The replacement of exonic portions provides several advantages relative to other gene therapeutic approaches, such as complementary DNA (cDNA) therapy. First, for very large genes, the sequence to be replaced can be reduced to a size that facilitates viral packaging and transduction. Second, the increase of wild-type alleles and the concomitant decrease of mutated alleles leads to an enhanced shift in their quantitative proportions, making this technology interesting for the correction of dominant mutations. Third, as trans-splicing (TS) takes place at the messenger RNA (mRNA) level, target gene expression remains under the control of the respective endogenous promoter, and problems arising from transgene overexpression can be excluded. The first publications on SMaRT date back to the 1990s, when minigene systems were used in cotransfection experiments. Accordingly, a selected target region for RTM binding is cloned into an expression vector and transfected, together with a rationally designed RTM, into cell lines that do not express the gene of interest.1,2 This approach facilitates the monitoring of correct TS but does not reflect a real disease setting, wherein the amount of the target molecule is often significantly reduced due to missense or nonsense mutations. Successful TS in an authentic, endogenous environment requires a highly functional and specific RTM to provide a therapeutically relevant level of correction and to exclude unspecific splicing events due to the relative underrepresentation of the target molecule. Currently, RNA TS has evolved into an elegant tool to reprogram endogenous pre-mRNAs. By specifically replacing a portion from the gene of interest, a reprogrammed, mature wild-type mRNA is generated, encoding a functional protein.3 With increasing experience, it turned out that the composition and characteristics of the BD sequence play a crucial role in the TS process. Therefore, during the past years, the main focus has been on the identification and design of highly functional BDs. Due to the diversity of primary and secondary structure characteristics of target regions (mostly introns), no reproducible principles for BD design applicable for a majority of target regions have emerged. Therefore, we developed a fluorescence-based RTM screening system, which we have successfully applied for several genes underlying the heterogeneous skin blistering disease epidermolysis bullosa.4–7 In this letter, we want to share our expertise and provide insight into a method that facilitates the identification of highly functional BDs for any gene of interest. The choice of using 3′ TS, 5′ TS, or internal exon replacement (IER) depends mainly on the type and number of mutations known in a gene of interest. Whereas 5′ and 3′ TS are used to replace large upstream or downstream regions of a pre-mRNA, thereby covering all potential mutations in this region, IER is used to replace one or more central exons.8 The latter is especially suitable for genes with mutational hotspot mutations because the size of the respective RTM can be minimized, facilitating cloning and transfections. Another method of TS, without targeting endogenous transcripts, was recently shown by Koo et al.9 In their study, the cDNA (>11 kb) of the Duchenne muscular dystrophy gene was transported into skeletal muscles of mdx mice using three independent adeno-associated virus vectors carrying sequential parts of the coding sequence of the Duchenne muscular dystrophy gene . TS between the three vectors led to expression of the full-length protein. Although this method is an elegant way of full-length cDNA replacement, it does not provide the advantages of corrective TS. However, it shows the wide applicability of TS, and general aspects of RTM design will also be valid for this approach. For the design of RTMs that target endogenous pre-mRNAs, there is not much scope regarding the composition of the sequence to be replaced, the spacer, and the splicing domain. However, the BD is very variable, and even slight variations can change the efficiency of the TS process significantly. Therefore, the characteristics of the targeted intron also have to be considered because longer introns provide the possibility of obtaining a library with high BD diversity. Finally, the inclusion of potentially existing isoforms in the TS process, or the exclusion of isoforms, has to be borne in mind. To select optimal BDs, applicable for the replacement of any gene and gene portion of interest, we developed a fluorescence-based screening system. For each targeted pre-mRNA, an artificial target and an RTM library is needed. RTM backbones are cloned, comprising a spacer, splicing domains, and respective parts (5′, 3′, or a central part) of a fluorescent reporter molecule (e.g., green fluorescent protein), which mimics the endogenous Considerations for a Successful RNA Trans-splicing Repair of Genetic Disorders LetteR to the eDitoR