Rectifying RNA splicing errors in hereditary neurodegenerative disease

The normal pattern of human gene expression requires precise coordination between the transcriptional and RNA processing machineries (1). During and following transcription, nascent RNAs may be modified by 5′-end capping, splicing, and 3′-end cleavage/polyadenylation to generate the vast repertoire of gene products required for proper gene expression, from embryonic to late adult life. RNA splicing is a particularly critical step in this process because it generates the multitude of coding and noncoding RNAs required for specific cell and tissue functions. However, splicing is susceptible to processing errors as a result of regulation by numerous cis-acting sequences and transacting factors, and 5′ splice sites may be particularly vulnerable (2). Thus, it is not surprising that a large number of hereditary mutations result in abnormal splicing and disease (3–5). One striking example is the sensory and autonomic neuropathy familial dysautonomia [FD, also known as Riley-Day syndrome or hereditary sensory and autonomic neuropathy type III (MIM 223900)], which is remarkably common in the Ashkenazi Jewish population with a carrier frequency of ∼1 in 30 (6). FD results from a homozygous mutation in intron 20 (IVS20 + 6T > C) of the inhibitor of kappa light polypeptide gene enhancer in B cells, kinase complex-associated protein (IKBKAP) gene, which encodes the IκB kinase complex-associated protein, that is predicted to disrupt interactions between IKBKAP pre-mRNA and U1 small nuclear ribonucleoprotein particle (snRNP), resulting in exon 20 skipping, frameshifting, and enhanced RNA turnover via the nonsense-mediated decay pathway (7) (Fig. 1A). Interestingly, the degree of exon 20 skipping induced by this FD mutation varies between tissues, with the highest level in the central and peripheral nervous systems, likely accounting for the neurodegenerative phenotype associated with this disorder. In PNAS, Yoshida et al. (8) use a bichromatic (red and green fluorescent proteins) splicing reporter assay to identify a small-molecule compound, which they term RECTAS (rectifier of aberrant splicing) that enhances exon 20 splicing leading to elevated IKBKAP protein levels. Importantly, RECTAS appears to possess several advantages over previously reported small-molecule therapeutics for FD.

Fig. 1.

(A) IKBKAP gene structure (introns, gray lines; exons, blue numbered boxes) with the intron 20 5′ splice site sequence indicated together with the FD disease-associated U→C mutation at position +6. In unaffected controls, exons 19, 20, and 21 are spliced (dotted lines) whereas exon 20 is skipped in FD cells. Skipping of exon 20 introduces a frameshift with a premature termination codon (PTC, red stop sign) in exon 21, resulting in nonsense-mediated decay of IKBKAP RNA (gray arrow). Also illustrated is the bichromatic minigene reporter assay with GFP (green) expression indicating exon 20 splicing; RFP (red) is generated by exon 20 skipping. Using this reporter system, Yoshida et al. (8) screened 638 small molecules from chemical libraries and identified the RECTAS compound. (B) Small-molecule compounds that have been reported to influence IKBKAP exon 20 splicing with the effective concentrations indicated. Also listed are the potential mechanisms of action including kinetin-induced up-regulation of LUC7L (U1 snRNP-associated protein) and down-regulation of SNRPA (U1 snRNP core protein U1A) (19), EGCG-induced down-regulation of hnRNP A2B1, and a possible RECTAS-induced suppression (gray bar) of hnRNP H1 and H2 ESS binding activity (8). (C) Schematic of tRNA structure (Left) with anticodon region (orange) and the ncm⁵ modified wobble U (red) in the anticodon stem-loop region of tRNA^Val(UAC) (Right) is included as an example.

The surprising result that FD is caused by a C >T transition mutation in a weakly conserved position in IKBKAP intron 20 led to the finding that sequences both within and flanking exon 20 are inherently weak, including the presence of several exon splicing silencers (ESSs) (9). Several small-molecule compounds have been reported to promote IKBKAP exon 20 inclusion, possibly because of interference with factors that recognize ESSs (Fig. 1B). An early study indicated that the polyphenol (−)-epigallocatechin gallate (EGCG) reduces the expression level of heterogeneous nuclear ribonucleoprotein (hnRNP) A2B1, which often recognizes ESSs, and increases IKBKAP exon 20 splicing and protein levels (10). The plant cytokinin kinetin (6-furfurylaminopurine) enhances exon 20 splicing, although relatively high concentrations (>10 μM) are required (11). Indeed, RECTAS appears to be a much stronger activator of IKBKAP exon 20 splicing (∼25-fold more potent than kinetin). Yoshida et al. (8) use a variety of experimental strategies, including in vitro splicing and formation of active spliceosomal complexes on precursor RNAs in HeLa nuclear extracts, to argue that RECTAS directly activates IKBKAP exon 20 splicing by promoting U1 snRNP binding, although the recruitment mechanism has not been determined. Although microarray analysis indicates that RECTAS affects the splicing of multiple exons, this compound appears to be a relatively specific activator of IKBKAP exon 20 splicing.

What pathways are directly influenced by loss of IKBKAP in FD? IKBKAP has been implicated in multiple regulatory steps, including a recent report describing a neuron autonomous function in microtubule dynamics (12). Nevertheless, current studies have focused on its primary role as one of the six subunits of the Elongator complex involved in tRNA side-chain modifications (ncm⁵, 5-carbamoylmethyl; mcm⁵, 5-methoxycarbonylmethyl) on uridines at the wobble position (13) (Fig. 1C). These tRNA modifications are one of the numerous posttranscriptional changes that regulate the normal folding, stability, and anticodon function of this central regulator of gene expression (14). If IKBKAP functions in tRNA modification, then FD cells should show hypomodified wobble uridines; Yoshida et al. (8) isolate specific tRNAs and use liquid chromatography/mass spectrometry, as well as additional techniques, to confirm this prediction for four cytoplasmic tRNAs. Importantly, RECTAS treatment of FD fibroblasts rescues both wobble uridine hypomodification and slow-growth phenotypes.

Perhaps the most important question evoked by the Yoshida et al. study (8) is whether RECTAS is a viable therapeutic candidate for FD. Because disease-causing mutations in splice sites and splicing regulatory sequences are common, several strategies have been pursued to correct mis-splicing in disease (15). One particularly promising approach is to target the mis-spliced exon directly using modified antisense oligonucleotides (ASOs) to modulate splicing (16). This RNA-targeted strategy is currently being tested as a potential therapy for spinal muscular atrophy (SMA), an

Perhaps the most important question evoked by the Yoshida et al. study is whether RECTAS is a viable therapeutic candidate for FD.

autosomal recessive motor neuron disease caused by survival of motor neuron 1 (SMN1) deletion or loss-of-function mutations. The SMN protein is required for the normal biogenesis and recycling of spliceosomal snRNPs (17). Although the closely related gene SMN2 is linked to SMN1, the splicing of SMN2 exon 7 is impaired because of a C→T transition in exon 7, resulting in the production of a truncated SMNΔ7 protein that fails to fully restore SMN activity. By analogy to FD, SMN2 exon 7 splicing is also regulated by splicing silencers and targeting of a SMN2 intronic silencer by a modified ASO (ASO-10-27) effectively restores SMN protein levels and is currently in clinical trials for SMA (16, 18). Several features of RECTAS suggest that it might also be a strong therapeutic candidate for FD, including restoration of normal IKBKAP exon 20 splicing, plasma stability following oral administration, and the observation that effective doses can be achieved in the brain (8). Nevertheless, important mechanistic questions about the RECTAS mode of action must now be addressed in relevant animal models, including target versus off-target splicing effects, positive and negative impacts on transacting splicing factors, and whether potential neuron-specific functions of IKBKAP are also restored following treatment with this promising small molecule drug.