RFC1-Related Disease

Gene Discovery and Molecular Genetics

CANVAS was first defined as a clinical entity in 2011 after deep and careful phenotyping identified a somatosensory abnormality as a feature of many individuals with the combination of CA with bilateral vestibulopathy.^6,23 Greater than a third of affected individuals with a clinical diagnosis of CANVAS identified in these early studies had an affected family member, most of whom were siblings, suggesting a potential recessive mode of inheritance.^6,24 Despite the presumed genetic basis for the condition, the challenges expanded repeats present for traditional molecular biology tools such as PCR and cloning meant the repeat expansion in RFC1 was only recently identified as the cause of CANVAS. This discovery was made by 2 groups after the identification of a linkage region on chromosome 4p14 in multiple families with a clinical diagnosis of CANVAS.^2,3 Using different approaches to analyze whole-genome sequencing (WGS) data, both groups subsequently identified the nonreference AAGGG_(n) expansion in RFC1 within the linkage region. One team identified the expansion by visually inspecting the aligned WGS read pairs within the linkage region, which revealed a region of reduced read depth in all patients with CANVAS corresponding to the repeat in RFC1. By contrast, the second team used a bioinformatics-based approach to identify repeat expansions genome-wide resulting in the identification of the AAGGG_(n) repeat within the linkage interval. Clinically, CANVAS is the triad of (1) cerebellar impairment, (2) bilateral vestibular hypofunction, and (3) somatosensory abnormalities.²³ The clinical diagnosis of CANVAS is made using detailed oculomotor, electrophysiologic, and imaging criteria (Table 3), which were developed after careful phenotyping of affected individuals.²⁵ This deep phenotyping and development of diagnostic criteria proved an important step in the eventual identification of the causative gene. After the identification of the pathogenic AAGGG_(n) allele, genetic analysis identified a core ancestral haplotype that most likely arose in Europe 25,000 years ago, suggesting a single founder event for the pathogenic repeat expansion.^2,3

The RFC1 repeat locus is highly polymorphic in nature with considerable variability in both repeat motif and size in the population (Figure 1). In the first study to examine the RFC1 locus in unaffected control populations, 4 different repeat conformations were observed: the reference AAAAG₍₁₁₎, nonpathogenic expansions of AAAAG_(n) and AAAGG_(n) motifs, and the pathogenic expansion of AAGGG_(n).² Nonpathogenic expanded alleles account for approximately 20% of alleles in unaffected individuals.² These nonpathogenic expanded alleles range in size from 40 to 1,000 repeats for the AAAGG_(n) motif and from 15 to 200 repeats for the AAAAG_(n) motif. Initial studies suggested expansions of AAGGG_(n) exceeding ∼400 repeats are pathogenic, while the nonpathogenic expansions of both AAAAG_(n) and AAAGG_(n) motifs are typically smaller in size.² Three additional motifs have subsequently been described in the general population: AAGAG_(n), AGAGG_(n),²⁶ and most recently, AACGG_(n).²⁷ However, no individuals homozygous for these conformations have been reported, so the potential pathogenicity of these motifs remains uncertain. In addition to the AAGGG_(n) allele, 2 alternate pathogenic allele configurations have been identified in Asia Pacific cohorts. An apparently Māori-specific allele configuration of AAAGG_(10-25)AAGGG_(n)AAAGG_(4-6) was identified in individuals affected by CANVAS with New Zealand and Cook Island Māori heritage; affected individuals were phenotypically indistinguishable from AAGGG_(n)-mediated disease.²⁸ In addition, a novel pathogenic ACAGG_(n) motif was also identified in both Asia Pacific and Japanese cohorts.^29,30 Additional phenotypic features beyond the classical CANVAS triad have been reported in affected individuals with expansions of this ACAGG_(n) motif, including fasciculations and elevated serum creatine kinase levels.²⁹ Notably, alleles with both the ACAGG_(n) and AAAGG_(10-25)AAGGG_(n)AAAGG_(4-6) configurations share the same core haplotype as the AAGGG_(n) motif.^28,29 While the results of the 4 haplotype studies performed to date are suggestive,^2,3,28,29 additional analyses across additional populations are required to confirm the theory of a single founder event for the pathogenic expansion.³

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Polymorphism at the RFC1 Locus

Schematic representation of the repeat configurations at the RFC1 repeat locus showing the nonpathogenic alleles (top), pathogenic alleles (middle), and alleles of uncertain pathogenicity (bottom). RFC1 = replication factor C subunit 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Visually Enhanced Vestibulo-Ocular Reflex (VVOR)

Left trace displays normal VVOR with gain ∼1 and few low amplitude corrective saccades. Right trace is that of an abnormal VVOR with a reduced gain and multiple salvos of high-amplitude corrective saccades. Red = head velocity; black = eye velocity.

Reported sizes for the pathogenic AAGGG_(n) alleles are typically larger, ranging from ∼400 to 2,000 repeats. However, expansions of ∼100–160 repeat units have recently been reported as likely pathogenic in 4 individuals with a clinically defined movement disorder (see further).¹⁹ Of interest, and in contrast to other repeat expansion conditions such as Friedreich ataxia and CAG expansion disorders, no significant correlation between the size of the repeat expansion and disease severity or age at onset has been reported to date.¹³ This lack of correlation together with the high frequency of nonpathogenic expanded alleles supports the hypothesis that the repeat motif rather than repeat length may be the driver of pathogenicity. This hypothesis is supported by findings in other pentanucleotide repeat expansion disorders, including spinocerebellar ataxia (SCA) 31, SCA37, and familial adult myoclonic epilepsy, where only expansions of a nonreference motif are pathogenic.^31,-,33 In addition, interruptions of an ATCCT motif in SCA10 are strongly associated with epileptic seizures, providing additional evidence that the repeat motif rather than length is driving pathogenicity.³⁴ Similar to RFC1, nonpathogenic alleles at the SCA37 locus are highly polymorphic, suggesting there may be a common mechanism driving pathogenicity in pentanucleotide repeat disorders.³² Further studies are therefore needed to accurately elucidate the pathogenic expansion threshold and the contribution of various motifs to pathogenesis.

Neuropathology

Although the pathomechanisms of RFC1 expansions remain unknown, it is well established that neuronal loss occurs in ataxia. Current knowledge of the neuropathology of RFC1 expansion is limited and derives from postmortem examinations of 5 individuals with a clinical diagnosis of CANVAS and confirmed biallelic AAGGG expansions.^2,9,21,35,36 Postmortem histopathologic examination of brain tissue from these individuals revealed marked loss of cerebellar Purkinje cells, most severe in the vermis. Purkinje cell depletion was accompanied by Bergmann gliosis and empty baskets.^21,35,36 In 2 individuals with CANVAS and an enriched phenotype including parkinsonism, marked neuronal depletion was observed in the substantia nigra and locus coeruleus.^21,35 Sural nerve biopsies from individuals with RFC1 expansions diagnosed with CANVAS or chronic idiopathic axonal polyneuropathy demonstrated depletion of both large and small myelinated fibers with no evidence of active axonal regeneration.^2,12,35

Major Phenotypic Characteristics of RFC1-Related Disease

Since the discovery of expansions in RFC1, there has been considerable interest in defining the potential spectrum of associated phenotypes other than CANVAS.^{10,-,12,16,17,19,27,37,-,39} While this is a necessary and exciting step in the process of any evolving disease discovery, cautiousness and implementation of the same processes that were used to identify CANVAS remain important. Where the technology exists for objective measurement of clinical characteristics, these should be used in case identification. This particularly concerns the objective identification of bilateral vestibular hypofunction and somatosensory loss. Unfortunately, several studies have neglected to perform the relevant investigations and have instead relied on bedside assessments or incompletely described methodology.^4,10,-,13,16 Key examples of this are the clinical vs video head impulse test (HIT) and the bedside vs electrophysiologic assessment of somatosensory function.^40,41 The exception here may be an absence of cerebellar atrophy on MRI in the early stage of cerebellar disease where bedside examination findings may be more sensitive than MRI findings.^40,41 Oculomotor abnormalities related to cerebellar impairment are generally present in cases of neurodegenerative CA and measurable by various oculographic modalities, even when structural changes in imaging are absent, and so, may function as an objective metric of cerebellar dysfunction. So far as the bedside examination of cerebellar function is concerned, assiduous examination for the presence of the clinical signs of cerebellar impairment is of utmost importance, understanding that those signs that do not alter in the presence of vestibular and/or somatosensory pathology are of greatest value.^3,42 We did not aim to present an exhaustive exposition of phenotypic features reported to date but rather to focus on those we believe are key to recognizing potential RFC1-related disease based on the more robust evidence in the literature.

RFC1 Pathogenic Mechanisms

The pathogenic mechanism(s) underlying biallelic RFC1 expansions is currently unknown. There are multiple known mechanisms of repeat expansion pathogenesis including loss-of-function or gain-of-function, RNA-binding protein sequestration, and repeat-associated non-AUG (RAN) translation.⁵⁴ RAN translation is a process whereby protein synthesis is initiated at the site of a repeat expansion in the absence of an AUG start codon. These RAN peptides can accumulate and form toxic aggregates.⁵⁴ RAN translation has been associated with more than 10 repeat expansion disorders, but no evidence of RAN proteins has been observed in individuals with biallelic RFC1 expansions to date.² Unlike other recessive repeat expansion disorders such as Friedreich ataxia, the RFC1 expansion does not appear to have a loss-of-function effect.^2,8 In both peripheral tissues and postmortem brain samples, no reduction in RFC1 mRNA or protein was observed.² In other conditions caused by intronic repeat expansions, such as C9orf72 amyotrophic lateral sclerosis/frontotemporal dementia, altered RNA splicing results in intron retention and the accumulation of RAN proteins and RNA foci, both of which affect downstream pathways.^54,55 To date, no abnormal splicing or intron retention has been consistently observed in individuals with biallelic RFC1 expansions, although a slight increase in the retention of intron 2 in pre-mRNA was reported in patient-derived lymphoblasts and muscle biopsies.² While suggestive that abnormal pre-mRNA processing may play a role in the pathogenic mechanisms of RFC1 expansions, these studies were small (n = 2 and 7, respectively); therefore, additional studies with larger sample sizes are required to confirm these data.

RFC1 encodes the large subunit of replication factor C, a pentameric DNA polymerase accessory protein involved in DNA replication and repair.¹ The RFC complex functions as a DNA-dependent ATPase that catalyzes the opening of the DNA-clamp protein proliferating cell nuclear antigen (PCNA), allowing it to encircle DNA and act as a scaffold to recruit DNA polymerases and other proteins involved in DNA replication and repair.¹ RFC1 has been associated with various DNA damage response pathways including DNA damage recognition, recruitment of DNA repair machinery, base and nucleotide excision repair, mismatch repair, and promotion of cell survival.^56,-,58 Of interest pathogenic variants in multiple genes involved in DNA repair pathways including tyrosyl-DNA phosphodiesterase 1, polynucleotide kinase 3′-phosphatase, X-ray repair cross-complementing 1, aprataxin, ataxia telangiectasia mutated, meiotic recombination 11 homolog A, and PCNA result in ataxia (Figure 3).^59,-,63 This suggests that the cerebellum is particularly vulnerable to DNA damage.⁵⁹ However, fibroblasts with biallelic RFC1 expansions do not show evidence of an impaired response to DNA damage.² Furthermore, in a cohort of 100 White individuals with genetically confirmed biallelic RFC1 expansions, the incidence of neoplasm was not increased when compared with that in 100 randomly chosen age-matched and sex-matched individuals with ataxia, contrary to what might be expected if RFC1 function is impaired by the homozygous repeat expansion.¹³

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 DNA Repair Pathway Genes Associated With Ataxia

Pathogenic variants in multiple genes involved in single-strand break repair and double-strand break repair result in ataxia.^59,63 Key proteins in the repair pathways are shown; those directly linked with ataxia are in bold and highlighted in green. TDP1 is associated with spinocerebellar ataxia with axonal neuropathy-1 (SCAN1); XRCC1 is associated with autosomal recessive spinocerebellar ataxia-26 (SCAR26); PNKP is associated with ataxia-oculomotor apraxia-4 (AOA4); APTX is associated with ataxia-oculomotor apraxia-1 (AOA1); PCNA is associated with ataxia-telangiectasia–like disorder-2 (ATLD2); RFC1 is associated with cerebellar ataxia, neuropathy, and vestibular areflexia syndrome (CANVAS); MRE11 is associated with ataxia-telangiectasia–like disorder-1 (ATLD1); ATM is associated with ataxia-telangiectasia. nt = nucleotide.

Molecular Testing for RFC1 Expansions

Based on the more robust elements of the RFC1 literature to date, we suggest testing for RFC1-related disease where the following phenotypes are accurately identified: CANVAS, late-onset pure CA, a sensory neuronopathy, combined CA and sensory neuronopathy or other peripheral sensory impairment, combined CA and a peripheral sensory motor neuropathy, any of these phenotypes in combination with autonomic dysfunction, parkinsonism, and/or a chronic cough. Where a family history of any of these phenotypes exists, the index of suspicion is raised. To date, it appears that RFC1-related disease generally has a very late onset and families with many affected individuals are relatively uncommon; hence, it may be difficult to discriminate between a dominant and recessive pattern, meaning that family history is an indicator but not exclusionary. The reader is directed to additional published guidance regarding clinical pathways for genetic testing of CANVAS and the genetic ataxias more broadly.^24,64,65

Molecular diagnosis of an RFC1 expansion requires a combination of a flanking PCR to test for the absence of a normal allele and repeat-primed PCR (RP-PCR) to identify the presence of expanded alleles. Individuals with biallelic RFC1 expansions fail to generate a flanking PCR product but generate a characteristic sawtooth pattern by RP-PCR using a primer specific for the pathogenic repeat motif (Figure 4). Given the motif-specific nature of RP-PCR, assays targeting the AAGGG_(n) allele will fail to detect the alternate pathogenic repeat conformations such as the ACAGG_(n) allele. Accurate sizing of an expanded allele is challenging as it can range in size up to several thousand repeats, but size estimates can be generated by Southern blotting or long-read sequencing. Indeed, long-read sequencing is increasingly being used for the detection and size estimation of repeat expansions and has the potential to replace Southern blotting as the gold standard for sizing of expanded repeats.⁶⁶ In addition, despite the intronic location, RFC1 expansions have been identified in both WGS and whole-exome sequencing data.^3,38 Although not yet routinely incorporated into diagnostic pipelines, it is likely that in the near future, short-read diagnostic genome data will be interrogated using bioinformatic tools to identify expanded repeats,⁵ with subsequent orthogonal testing to validate the result and estimate the size of pathogenic alleles at specific loci performed, if appropriate. Currently, the availability of diagnostic testing for RFC1 expansion is very limited. Interrogation of the GTR resource (NCBI, 2-2022) identified a single diagnostic provider of RFC1 testing using flanking PCR and RP-PCR testing for the AAGGG_(n) pathogenic motif, although the authors were aware of multiple providers currently developing accredited tests.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4 Comparison of Methods to Identify Repeat Expansions

Graphical representation of the different diagnostic methods for detecting repeat expansions. The samples analyzed for the flanking and repeat-primed PCR are taken from a patient with biallelic RFC1 expansions. ctrl = control; rpts = repeats.

Conclusions and Future Directions

Since the discovery in 1991 of the first expansion repeats, causing fragile X syndrome and spinal bulbar muscular atrophy, at least 50 repeat-mediated disorders have been described. Remarkably, more than half of these repeat expansions have been identified in the past 10 years, predominantly due to the advent of new sequencing technologies and associated bioinformatic analyses. Many of these expansions cause heterogenous neurogenetic conditions that have proven difficult to both diagnose clinically and confirm with molecular genetic testing. For many of the recent discoveries, such as the repeats underlying familial cortical myoclonic tremor and CANVAS, the careful collection and deep phenotyping of affected family members has proven crucial to identification of the causal repeat expansion. The ability to achieve a molecular genetic diagnosis provides the opportunity to understand the natural history of a condition, and starting with CANVAS, we are just beginning a fascinating journey to understand RFC1-mediated disease. Additional deep phenotyping leveraging available objective methods will be essential in determining the true spectrum of phenotypes directly caused by RFC1 expansions. This process is also a necessary step in minimizing misdiagnoses, unnecessary or missed opportunities for prudent gene testing, and suboptimal management. Finally, we anticipate these studies will help to reveal the underlying disease mechanisms of pathogenic RFC1 expansions, which currently remain elusive. In an era of emerging therapies for repeat expansion disorders,⁶⁷ understanding the mechanism will allow the identification of potential therapeutic targets and ultimately, the realization of safe and effective treatments.