Identification of Exon Specific U1s that correct splicing defects in CFTR exons 5, 9, 13, 16 and 18.
To evaluate the therapeutic potential of ExSpeU1 in CF we focussed on ten splicing defects. Mutations were selected based on their relative high frequency in CF, on exon coverage and on the splicing regulatory elements affected (Table 1). Among those analysed, according to CFTR2 database (www.cftr2.org), 2789+5G>A is the most frequent mutation (1027 alleles reported), followed by 3120G>A, 711+3A>G and 711+5G>A (85, 63 and 60 alleles, respectively) and 1898+3A>G (27 alleles) while 711+3A>C and 1863C>T (Y577Y) are rare mutations. As well, we analysed two polymorphic repeat regions in the 3′- splice site of human CFTR exon 9 affecting the polypyrimidine tract: TG12T5 and TG13T5 associated with atypical CF, and TG13T3 linked to classical CF (Cuppens et al. 1998; Pagani et al. 2000; Pagani et al. 2003; Groman et al. 2004; Du et al. 2014; Boussaroque et al. 2020). Altogether the mutations we have analysed cover five CFTR exons: 5, 9, 13, 16 and 18. Six mutations are located in the 5’ss (donor) consensus, (711+3A>C/G, 711+5G>A, 1898+3A>G, 2789+5G>A, 3120G>A), one affects an exonic regulatory element (1863C>T) and three variants involve the polypyrimidine tract (TG12T5, TG13T5 and TG13T3). In all mutations, rescue of exon skipping will result in the production of a normal coding CFTR mRNA either because the mutations are intronic or because they are synonymous. To test the splicing rescue activity of ExSpeU1s, we created minigenes with exons and corresponding flanking intronic regions that contain the target wild type (WT) and mutant sequences. CFTR exon 5 minigene experiments showed that, in basal conditions, the WT construct was nearly completely included whereas 711+3A>C and 711+5G>A induced severe exon 5 skipping and 711+3A>G retained some exon inclusion (~ 18%) (Fig.1 A, lanes 1, 2, 4 and 3, respectively). We tested four ExSpeU1s binding downstream the 5’ss (U1ex5-9, U1ex5-13, U1ex5-21 and U1ex5-27 bp) (Fig.1A). These ExSpeU1s induced complete rescue of the exon 5 splicing defects; the only exception was represented by U1ex5-21 and U1ex5-27 that partially recovered (∼ 52%) the 711+3A>C variant. For exon 9, we analysed four TG(m)T(n) variants: TG11T7, TG12T5, TG13T5 and TG13T3 along with four ExSpeU1s (U1ex9-3, U1ex9-12, U1ex9-26 and U1ex9-34). These mutations induce different degrees of exon skipping (Fig. 1B) with activation of a cryptic site inside the exon, as previously reported (Ayala et al., 2006; Buratti et al., 2001; Zuccato et al., 2004). In basal condition, TG11T7 showed ~ 98% of exon inclusion, TG12T5 ∼ 61%, TG13T5 ∼ 54% and TG13T3 ∼ 6% (Fig. 1B). ExSpeU1s co-transfection totally rescued aberrant splicing in TG13T5 and TG12T5. In TG13T3, splicing rescue was ∼ 72 - 84% for U1ex9-3, U1ex9-12, U1ex9-26, whereas downstream ExSpeU1s had a lower effect (U1ex9-34 ∼ 51%). We then evaluated two exon 13 variants, 1863C>T and 1898+3A>G that induce respectively ∼ 4% exon inclusion and complete skipping, respectively (Fig.1 C, lanes 2 and 3). We tested several ExSpeU1 but only one U1ex13-11 was working efficiently by rescuing both variants to WT levels of exon inclusion, suggesting that the CFTR exon 13 context has some peculiarity that restricts the ExSpeU1 splicing capacity. To evaluate 2789+5G>A variant in exon 16 we tested five ExSpeU1s (U1ex16(-3), U1ex16-7, U1ex16-10, U1ex16-12 and U1ex16-17) (Fig.1D). In this case, the mutant showed ∼53% exon inclusion and all ExSpeU1s were able to completely correct the splicing defect (Fig. 1 D). We next analysed the 3120G>A variant in exon 18. This mutation, located in the last base of the exon, induced complete exon 18 skipping (Fig. 1E, lane 2). The different ExSpeU1 tested were able to efficiently rescue the splicing defect: U1ex18-3 and U18-20 induced ∼ 85% of exon inclusion and U1ex18-10 and U18-13 ~100% (Fig. 1 E). Altogether our results have identified active ExSpeU1s able to correct with high efficiency ten splicing defects that affect CFTR exon 5, 9, 13, 16 and 18.