Discussion
In last years, research on CF has been very successful in developing personalised therapies based on the correction of the underlying basic defects (De Boeck & Amaral, 2016; Fajac & Wainwright, 2017; Farinha & Matos, 2016; Paranjape & Mogayzel, 2018; I. Pranke et al., 2019; Roomans, 2014). This has led to the discovery of drugs that act on specific CFTR protein mutants either as potentiators or correctors (Boyle et al., 2014; Chaudary, 2018; Clancy, 2018; Gees et al., 2018; Gentzsch & Mall, 2018; Guimbellot et al., 2017; Phuan et al., 2018; Van Goor et al., 2009; Wu et al., 2019; Yeh et al., 2017). However, CF individuals with mutations that disrupt other basic mechanisms, and in particular CFTR RNA processing, cannot benefit of these personalized therapies. In this paper, to provide a useful rescuing approach acting at the RNA level, we focussed on a group of splicing mutations that induce exon skipping. We show here that a strategy based on modified U1 snRNAs, named Exon Specific U1s (ExSpeU1), when applied to ten representative splicing mutations, can correct the resulting exon skipping defects with recovery of the normal CFTR protein. With ten mutations and five exons analysed, our study is the first to show a therapeutic effect of an exon rescue strategy on several splicing mutations and various exons in a single gene. Indeed, in CF a different approach based on Anti Sense Oligonucleotides (ASO) has been tested only in the 2789+5G>A splicing mutation (Igreja et al. 2016). In general, this alternative ASO based approach for rescuing exon skipping defects is based on the identification of appropriate splicing regulatory elements and thus more time consuming and expensive and not easy testable on several splicing defects. Similarly, a pharmacological approach to correct splicing requires high-throughput screening platforms and may lack of specificity (Berg et al., 2019; Giuliano et al., 2018; Liang et al., 2017; Merkert et al., 2019; Pereira et al., 2019). The positive effect we observed here on different CFTR mutations, along with those previously reported for other diseases in several cellular and mouse models (Dario Balestra et al., 2020; Dal Mas, Fortugno, et al., 2015; Donadon et al., 2018; Tajnik et al., 2016), indicates a general applicability and translatability of the ExSpeU1 approach in rescuing exon skipping defects. Even if our results show a clear rescue potential of the ExSpeU1 approach in CF, the therapeutic window is restricted to those splicing defects that do not disrupt the invariant AG and GT dinucleotides of the splice site. These more severe cases cause a complete disruption of the splicing process and thus cannot benefit of the ExSpeU1-mediated rescue approach. In CF, we estimate that approximately 40 splicing mutations out of 352 CF-causing variants listed in CFTR2 database could benefit from the ExSpeU1 approach. The specificity and easy tailoring of the ExSpeU1 strategy is mainly related to its engineered 5’ tail that targets the ExSpeU1 on desired intronic sequences. As we identified for each CFTR exon a panel of active ExSpeU1s (with the notable exception of exon 13), several molecules could be then tested for better efficacy and safety. An important concern for ExSpeU1s, as well as for other splicing correction strategies with ASO or chemical compounds, is their potential off-target effects. Previous studies in cellular and animal models with one ExSpeU1 active on Spinal Muscular Atrophy identified a very limited number of off-targets by RNA-Seq analysis (Dal Mas, Rogalska, et al., 2015; Donadon et al., 2019). Indeed, the availability of cellular systems that overexpress the ExSpeU1s will allow the identification of the most promising and safe molecules active on the different CFTR exons. In CFTR exons 5 ,9 and 18, the ExSpeU1s tend to reduce their exon rescue efficacy when binding to more downstream intronic positions (Fig.1) and this phenomenon has been also observed previously in other cases (Alanis et al., 2012; Dal Mas, Fortugno, et al., 2015; Tajnik et al., 2016). Thus, the detection of only one ExSpeU1s active on exon 13 mutants was unexpected. We reasoned that the peculiar context of this exon could be involved and indeed we have identified in intron 13 an ISS that, through the formation of a stem-loop RNA secondary structure in the proximity of the 5’ss, has a negative effect on splicing. Similar stem-loop structures are known to modulate the 5’ss accessibility in other exon skipping defects (Buratti & Baralle, 2004). Thus, we speculate that the overlap between the ISS and the binding region of U1ex13-11 might explain why only this one was active on exon 13 splicing mutants. It is possible that part of the U1ex13-11 efficacy, which is not shared by the adjacent ExSpeU1s, relies on its ISS binding position with consequent opening of the unfavourable RNA secondary structure.
The short ExSpeU1 coding gene (~ 650 bp) can be easily accommodated into Adeno Associated Viruses or Lentiviral Vectors forin vivo gene therapy (Rogalska et al. , 2016; Donadonet al. , 2018) or ex vivo cellular delivery (Dal Mas, Fortugno, et al., 2015; Dal Mas, Rogalska, et al., 2015; Nizzardo et al., 2015). Alternative delivery approaches that could be tested include the de novo chemical RNA synthesis (Ohkubo et al., 2013). The availability of patient-derived cellular models with the splicing mutations in form of nasal epithelial cells (I. M. Pranke et al., 2017) or organoids (Boj et al., 2017; Dekkers et al., 2013, 2016; Sato et al., 2009, 2011) will allow a more direct assessment of the ExSpeU1s strategy on the CFTR functionality and the evaluation of the ExSpeU1 efficacy and safety profile in patient-derived target cells.