c9RAN translation: a potential therapeutic target for the treatment of amyotrophic lateral sclerosis and frontotemporal dementia

Abstract

A hexanucleotide (GGGGCC) repeat expansion within a non-coding region of the C9ORF72 gene is the most common mutation associated with both frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS). Elucidating how these expanded repeats (GGGGCC_exp) cause ‘c9FTD/ALS' has since become an important goal of the FTD/ALS field. GGGGCC_exp transcripts aggregate into discrete nuclear structures, termed RNA foci. This phenomenon, observed in various repeat expansion disorders, is associated with RNA-binding protein sequestration. Of note, recent findings show that GGGGCC_exp transcripts also succumb to an alternative fate: repeat-associated non-ATG translation (RAN translation). This unconventional mode of translation, which occurs in the absence of an initiating codon, results in the production of polyGA, polyGP and polyGR peptides. Antibodies generated against these peptides detect high molecular weight, insoluble material in brain homogenates, as well as neuronal inclusions throughout the central nervous system of c9FTD/ALS cases. Given that both foci formation and RAN translation in c9FTD/ALS require the synthesis of GGGGCC_exp RNA, therapeutic strategies that target these transcripts and result in their neutralization or degradation could effectively block these two potential pathogenic mechanisms and provide a much needed treatment for c9FTD/ALS.

Keywords::

• amyotrophic lateral sclerosis
• C9ORF72
• expanded hexanucleotide repeat
• foci
• frontotemporal dementia
• neurodegeneration
• repeat-associated non-ATG translation
• RNA

RNA-mediated toxicity has emerged as an important pathological mechanism in several neurological disorders caused by genomic expansions of microsatellite repeats, including myotonic dystrophy (DM), spinocerebellar ataxias (SCA), and fragile X-associated tremor ataxia syndrome (FXTAS) [1]. Recently added to this list are frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS) caused by an expansion of the GGGGCC repeat in the C9ORF72 gene from fewer than 30 copies to hundreds or thousands of copies [2,3].

The pathology of ‘c9FTD/ALS,' and its clinical presentations, conform to those of classical FTD and ALS, which are thought to represent a continuum of disease. FTD results from the degeneration of frontal and temporal lobes and encompasses a group of disorders distinguished clinically by abnormalities in behavior, language and personality. ALS is characterized by the degeneration of motor neurons, leading to muscle atrophy and paralysis. Some FTD patients develop motor neuron dysfunction resembling ALS, and ALS patients often develop abnormalities characteristic of FTD. Disease onset is variable among C9ORF72 mutation carriers, as is clinical presentation, and there is evidence of incomplete disease penetrance. Determining whether repeat length influences these features, and whether genetic or environmental modifiers are involved, will certainly offer insight into mechanisms of disease. Neuropathologically, TDP-43-positive neuronal and glial inclusions typical of both disorders are observed in c9FTD/ALS patients. In addition, TDP-43-negative, p62-positive inclusions in the cerebellum and hippocampus, as well as neuronal inclusions composed of aberrant peptides produced via unconventional translation of GGGGCC_exp RNA, are a specific pathology in patients with the C9ORF72 mutation [4,5].

As with many microsatellite expansion disorders, the hexanucleotide repeat in C9ORF72 is present within a non-coding region. Efforts to elucidate how such expansions cause disease have led to the discovery of various mechanisms involving the production of repeat-containing RNA. The accumulation of these transcripts, often observed as nuclear RNA foci, induces cellular toxicity by activating downstream signaling pathways and disrupting the function of RNA-binding proteins [1]. For example, RNA transcripts of an expanded CTG repeat in DMPK fold into hairpins that bind and inactivate muscleblind-like 1 protein (MBNL1) in DM1 [6-8]. MBNL1 sequestration results in the mis-splicing of a subset of pre-mRNAs, such as the muscle-specific chloride ion channel, which accounts for characteristic features of DM1 [9,10]. Transcripts of the GGGGCC_exp repeat in C9ORF72 similarly accumulate as nuclear RNA foci in c9FTD/ALS [2], implicating RNA-binding protein sequestration in disease pathogenesis. A second toxic mechanism stems from the susceptibility of such transcripts to repeat-associated non-ATG translation (RAN translation). This atypical form of translation, which occurs in the absence of an initiating codon, was first described by the Ranum group, who found that RAN translation occurs in all reading frames across long CAG repeats to produce homopolymeric proteins (polyQ, polyS, or polyA) [11]. While RAN translation remains poorly understood, it appears to depend on repeat-length and hairpin formation, at least for CAG_exp RNA [11]. It may involve an internal ribosome entry site (IRES)-like mechanism; hairpins can recruit initiation factors and ribosomal subunits to IRESs, and certain viruses use an IRES to initiate translation at non-AUG sites [12].

Given that expression of RAN translation products is sufficient to cause apoptosis in cultured cells, their accumulation may contribute to the neurodegeneration in repeat diseases, such as DM1 and SCA8 [11]. Researchers also recently demonstrated that the CGG repeat expansion in FXTAS triggers RAN translational initiation within the 5′ UTR of FMR1 mRNA to generate a cryptic polyG-containing protein [13]. As mentioned above, expanded GGGGCC repeats undergo RAN translation in c9FTD/ALS, and this results in the synthesis of polyGA, polyGP and polyGR peptides [4,5]. Antibodies generated against these peptides specifically detect high molecular weight, insoluble material in c9FTD/ALS brain homogenates, as well as neuronal inclusions in c9FTD/ALS brain sections [4,5]. Moreover, evidence that the mutant C9ORF72 allele is bidirectionally transcribed [5] introduces the possibility that up to five RAN translated products may be expressed in c9FTD/ALS.

While the exact mechanisms of toxicity at play in c9FTD/ALS have yet to be fully elucidated, the need for treatment is no less crucial. Taking cues from investigations of other neurodegenerative diseases, one can surmise that the accumulation of aggregation-prone peptides resulting from RAN translation could be harmful to neurons. The inclusions may themselves be neurotoxic or they may sequester proteins causing loss of function, overwhelm protein degradation systems, or displace cytoplasmic organelles. Furthermore, RNA foci in c9FTD/ALS are expected to sequester various RNA-binding proteins and cause the misregulation of numerous downstream RNA targets [14-16]. For example, the RNA-binding proteins hnRNPA2/B1 and hnRNPA1, mutations in which cause ALS and related disorders [17], have been found to bind GGGGCC_exp transcripts [16].

Given that both foci formation and RAN translation in c9FTD/ALS require the synthesis of GGGGCC_exp transcripts, strategies that target these transcripts and result in their neutralization or degradation could effectively block an early culprit in these putative mechanisms of disease (Figure 1). No study examining the protective effect of targeting GGGGCC_exp RNA has yet been published; however, investigations have no doubt been underway since the discovery that mutant C9ORF72 causes c9FTD/ALS. Based on advances made in the development of therapeutic strategies for other repeat expansion diseases marked by RNA toxicity, initial progress is expected to be relatively rapid. Antisense oligonucleotides (ASOs), small interfering RNAs (siRNAs), and small molecules are but some of the approaches used to target pathogenic RNAs.

Figure 1. Potential mechanisms of disease in c9FTD/ALS. Expansion of the GGGGCC hexanucleotide repeat within intron 1 of the C9ORF72 gene may cause c9FTD/ALS via RNA-dependent mechanisms. RNA transcripts resulting from transcription of GGGGCC_exp aggregate into nuclear RNA foci that can sequester RNA-binding proteins and cause their loss of function. GGGGCC_exp RNA transcripts are also susceptible to repeat-associated non-ATG translation, producing polyGA, polyGP and polyGR peptides, which form neuronal inclusions in c9FTD/ALS. Therapeutic strategies that target GGGGCC_exp RNA transcripts may block both these putative pathogenic mechanisms. Moreover, while the mechanisms regulating RAN translation are poorly understood, and the means by which GGGGCC_exp transcripts make their way to the cytoplasm are not yet known, targeting these events may also provide an approach to block RAN translation.

Through base pairing to complementary nucleotide sequences, ASOs modulate the function of the targeted RNA via steric block or cleavage mechanisms. Steric blocking ASOs can modify pre-mRNA splicing by masking splicing enhancer or repressor sequences, inhibiting translation, or blocking the interaction of target RNA with RNA-binding proteins. For example, a steric blocking ASO, termed CAG25, was shown to disrupt the interaction between CUG_exp RNA and MBNL1 following intramuscular injection in a mouse model of DM1 [18]. Following release of MBNL1 from sequestration, splicing defects and defective chloride channel conductance were reversed [18]. Of note, while CAG25 was not designed to induce target degradation, levels of CUG_exp RNA were nonetheless decreased. This was postulated to result from its accelerated clearance given that, despite the overall decrease in CUG_exp RNA levels, the amount of this transcript was increased in the cytoplasm where it would be accessible to degradation systems. Small molecules have likewise shown promise in improving DM1-associated features, including disrupting nuclear foci formation, blocking the CUG_exp RNA-MBNL1 interaction, and attenuating translational and pre-mRNA splicing defects [19,20]. Given their permeability and tight binding affinities, small ligands offer an attractive therapeutic option.

In contrast to steric blocking ASOs, RNAse H-dependent ASOs and siRNAs are designed to eliminate the toxic RNA. A robust decrease in CUG_exp RNA was observed in skeletal muscle of DM1 transgenic mice after subcutaneous injection or injection into the tibialis anterior muscle of RNAse H-dependent ASOs, and this was accompanied by a decrease in nuclear RNA foci [21,22]. Also observed were release of MBNL1, correction of myotonia, and rescue of misregulated splicing with sustained effects for up to 1 year after treatment was discontinued [21]. Similarly, delivery of siRNAs targeting CUG repeats to skeletal muscle in DM1 transgenic mice led to a robust downregulation of CUG_exp transcripts, dispersal of RNA foci, and restoration of MBNL1 splicing regulation [23].

Expert opinion

The above-mentioned studies provide compelling evidence that targeting pathogenic RNA is a promising therapeutic strategy for the treatment of diseases in which RNA toxicity plays a role. Nonetheless, these strategies are still being refined and barriers exist that must be overcome if such approaches are to be utilized for treating neurodegenerative diseases. Existing challenges include improving the delivery of therapeutics to the central nervous system, as well as gaining a better understanding of potential off-target effects (e.g., effects resulting from the binding of the therapeutic to alternative transcripts that contain the repeat sequence being targeted). There are also several issues specific to c9FTD/ALS drug development that must be resolved. For instance, the secondary structure of GGGGCC_exp must be taken into account. Studies suggest that GGGGCC repeats can form G-quadruplexes, and secondary structure prediction modeling indicates that the repeats may also assume hairpins [4,15,24]. A G-quadruplex structure may negatively influence the ability of ASOs to bind the transcript. Small molecules, however, may be used to target such structures. It should also be kept in mind that conclusive proof that RNA-mediated toxicity underlies neurodegeneration in c9FTD/ALS has yet to be shown. While it may be a correct assumption that such is the case, a better understanding of disease mechanisms is required for designing the most effective therapeutic strategies. For example, while the means by which GGGGCC_exp transcripts make their way to the cytoplasm and are loaded onto ribosomes are not yet known, targeting these events could well be an efficient therapeutic approach to block RAN translation. Also to consider is the possibility that therapeutics that prevent the binding of GGGGCC_exp transcripts with RNA-binding proteins but increase cytoplasmic GGGGCC_exp RNA (akin to the effect of GAC25 on CUG_exp) may lessen one toxic cascade but enhance another. Of course, this may be avoided altogether by choosing approaches that eliminate toxic RNA. Yet, even then, challenges may arise; while ASOs and siRNAs are expected to preferentially cause the degradation of RNA from the expanded allele, there is likely to be some level of degradation of mRNA from the wild-type allele. Given that several groups have shown that mRNA levels of at least one C9ORF72 variant are decreased in c9FTD/ALS [2,3,25], a further decrease caused by a purportedly protective intervention could contribute to a loss of C9ORF72 function. The effects of RNA-targeting therapies on the wild-type allele should therefore be carefully monitored. In addition, the potential contribution of C9ORF72 loss-of-function as a neurotoxic mechanism in c9FTD/ALS should be explored, and the functions of this almost completely uncharacterized protein should be determined. It is also important to note that the co-occurrence of nuclear RNA foci, expression of RAN translated peptides, and accumulation of TDP-43 inclusions could suggest the existence of a complex pathway in which all are interconnected, or the involvement of parallel pathological processes. If the latter is true, targeting only aberrant RNA species would not alleviate TDP-43-mediated toxicity. As such, combinatorial therapeutic strategies may be needed for c9FTD/ALS. Finally, at present, a major limiting factor in c9FTD/ALS drug discovery efforts is the lack of robust experimental models that recapitulate key features of disease. Such models are necessary to gain a better understanding of disease mechanisms and will be crucial for testing efficacy of possible therapeutics. Conversely, the development of small molecules, ASOs and the like that target GGGGCC_exp RNA may not only lead to the development of potential therapies, but would also serve as valuable tools to decipher the many questions still unanswered.

Declaration of interest

This work was supported by the National Institutes of Health/National Institute on Aging [R01 AG026251 (LP)]; National Institutes of Health/National Institute of Neurological Disorders and Stroke [R21 NS074121-01 (TFG), R01 NS063964 (LP); R01 NS077402 (LP); R21 NS084528-01 (LP)]; National Institute of Environmental Health Services [R01 ES20395 (LP)]; and Amyotrophic Lateral Sclerosis Association (LP).