The C9orf72 hexanucleotide repeat expansion in FTD and ALS

Abstract

Two new studies provide strong evidence for the link between mutations in the C9orf72 gene and familial frontotemporal dementia or amyotrophic lateral sclerosis. One of the papers presents some unique associations between clinical features and C9orf72 mutation, and raises questions regarding the specificity of some previously reported pathological findings.

The clinical overlap between amyotrophic lateral sclerosis (ALS), or motor neuron disease (MND), and frontotemporal dementia (FTD) has been known for many years. The long-sought mutation in FTD and ALS cases linked to chromosome 9p was reported in two Neuron papers in September 2011.^1,2 The mutation comprises expansion of a unique hexanucleotide repeat in a noncoding region of the chromosome 9 open reading frame 72 (C9orf72) gene, which encodes a protein of unknown function. Two new publications^3,4 join several recently published studies that describe the clinical and pathological phenotype of frontotemporal lobar degeneration (FTLD) with or without ALS, or ALS alone, associated with C9orf72 repeat expansion.

The syndromes resulting from the overlapping diseases are referred to clinically as FTD–ALS or FTD–MND, and are designated pathologically as FTLD with MND (FTLD–MND). Most cases are linked pathologically by inclusions composed of aggregates of abnormal TAR DNA-binding protein 43 (TDP 43). The distribution of TDP pathology varies, usually with specific clinical and genetic associations.^5,6 Patients with TDP type A may have granulin (GRN) mutations, are equally likely to present with behavioural variant FTD (bvFTD) or primary progressive aphasia (PPA), rarely have ALS, and have TDP 43-positive cytoplasmic and intranuclear inclusions, predominantly in the upper cortex. Patients with TDP type B sometimes have linkage to chromosome 9p, most often present with bvFTD with or without ALS, and have TDP 43-positive cytoplasmic inclusions in all layers of the cortex. Patients with TDP type C most often present with PPA (but up to one-third present with bvFTD), do not usually have ALS, and have TDP 43-positive long neurites in the upper cortex.⁶

The two most recent studies to shed light on the link between the above disorders and C9orf72 mutations were conducted in the Netherlands and the UK.^3,4 In the Netherlands study, Simon-Sanchez et al. screened 353 patients with bvFTD or PPA, with or without ALS, and 522 controls.³ Patients with FTLD-tau pathology with or without microtubule-associated protein tau (MAPT) mutations, and patients with GRN mutations, were assumed not to harbour C9orf72 mutations and were, therefore, excluded. In the UK study, Snowden et al. screened for C9orf72 mutations in 398 patients with bvFTD or PPA (progressive nonfluent aphasia or semantic dementia), 55 of whom also had MND.⁴

Simon-Sanchez et al. found C9orf72 mutations in 42 patients (12%) in the FTD cohort; among these patients, 34 had bvFTD, eight had PPA, and seven had ALS. C9orf72 repeat expansion was also detected in three of the 522 controls.³ On the basis of these results, 53.8% of cases of familial FTD in the Netherlands are now explained: 17.8% have C9orf72 mutations, 22.1% have MAPT mutations, and 13.9% have GRN mutations.³ Snowden et al. found C9orf72 mutations in 32 patients (8%) in their cohort, which included 19 FTD cases, nine FTD–MND cases, three progressive nonfluent aphasia cases, and one semantic dementia case.⁴

In both studies, bvFTD predominated over PPA in mutation carriers with cognitive impairment.^3,4 In the UK study, 38% of mutation carriers presented with psychosis.⁴ With a few exceptions, most other studies have noted delusions only occasionally or not at all in patients with C9orf72 mutations. Whether this feature is unique to the UK cohort, or whether psychiatric signs, delusions or psychosis will be found in more patients with C9orf72 mutations if more specifically sought, remains to be seen.

The C9orf72 mutation seems to be inherited in an autosomal dominant manner, but reduced penetrance is evident, in that some obligate carriers do not develop ALS or dementia.^3,7 In both of the new studies, the C9orf72 mutation was more common in patients with ALS and in those with a family history of ALS or FTD than in individuals without these characteristics,^3,4 similar to another recent report.⁷ Should screening for the C9orf72 mutation be considered, targeting of individuals with ALS plus dementia who have a family history of ALS or FTD would, therefore, be a reasonable approach.

All previous studies of patients with C9orf72 mutations or chromosome 9p linkage have reported FTLD-TDP or ALS pathology.^8,9 Some studies found a predominance of TDP type B pathology, the type most often associated with ALS, whereas others reported a heterogeneity of TDP types including some apparently type A, the type most often associated with GRN mutations. Conversely, in the UK study, a brain autopsy on one patient with a C9orf72 mutation revealed FTLD-tau with corticobasal degeneration (CBD),⁴ representing the only report of a diagnosis other than FTLD-TDP or ALS with this mutation. This finding, if correct, calls into question the specificity of the C9orf72 mutation for FTLD-TDP. Chromosome 9p linkage and a subsequent finding of C9orf72 hexanucleotide repeat expansion has previously only been reported in cases of ALS, with or without dementia. However, exclusion of patients with MAPT and GRN mutations or FTLD-tau pathology, as in the Netherlands study,³ might cause ‘outliers’ as seen in the UK study to be missed.

Both studies report the results of immunostaining using an antibody to the C9orf72 protein. The two groups found diffuse and granular cytoplasmic labelling of neurons in areas affected by pathology in FTLD-TDP and ALS, but this feature was present in all cases, including controls, and was considered to be physiological.^3,4 C9orf72-positive hippocampal CA4 synaptic arborizations extending into areas CA2–3 were also observed. The Netherlands group found this labelling in cases of FTD with and without the repeat expansion.³ By contrast, the UK group found this labelling to be strong in healthy controls, variably preserved in cases of MND, but decreased or absent in some patients with FTLD and not markedly different in those with C9orf72 mutations.⁴

Both groups found p62-positive TDP 43-negative inclusions in cerebellar granular neurons,^3,4 as has previously been reported.^8,9 These inclusions were not labelled with the C9orf72-specific antibody; the identity and significance of the inclusion protein, therefore, remain unknown. Although previously thought to be specific for C9orf72 cases, these inclusions were not found by Snowden et al. in the C9orf72 mutation carrier with FTLD-tau (CBD) pathology, and were detected in three cases of FTLD-TDP seemingly without a C9orf72 mutation.⁴ If the absence of a repeat expansion in these three cases proves to be true, the inclusions may not be specific to cases with C9orf72 repeat expansion, as has been previously reported.^8,9

RNA misprocessing might be the pathophysiological link between C9orf72 mutation and FTD–ALS, as intranuclear RNA foci have been observed in cases with C9orf72 repeat expansion.¹ Simon-Sanchez et al. were unable to confirm this finding,³ possibly owing to use of a different RNA probe from that employed in previous studies. Other disorders with repeat expansion are associated with RNA misprocessing, which has provided researchers with testable potential therapeutic targets.

These two new reports confirm that C9orf72 repeat expansion is common in familial cases of FTD–ALS.^3,4 Unique findings from the UK study, such as the association of the mutation with psychosis, remain to be confirmed by others.⁴ As more is learned about the clinical and pathological characteristics of this genetic FTLD subtype, it is hoped that the threads linking ALS and FTD will be better understood, ultimately enabling the design of a targeted therapy for these disorders.