Introduction
Frontotemporal dementia (FTD) is characterised by early behavioural change and progressive erosion of social cognition associated with frontotemporal lobar degeneration [
1]. A substantial number of cases of FTD have a familial basis [
2], and an expanded hexanucleotide (GGGGCC) repeat insertion in a noncoding promoter region of ORF 72 of chromosome 9 (
C9ORF72) was recently identified as an important cause of FTD and motor neuron disease [
3,
4]. Recent reports of
C9ORF72 mutations suggest these are a common cause of FTD and motor neuron disease, representing approximately one-third of all cases due to genetic mutations [
5,
6], of comparable frequency to mutations in progranulin (
GRN) and micro-tubule protein tau (
MAPT) as a cause of autosomal dominant FTD [
6].
Clinically,
C9ORF72 expansions have been associated with a behavioural dysexecutive phenotype but also with notable early features including psychosis and anxiety as well as impaired episodic memory [
6,
7]. Individual cross-sectional magnetic resonance imaging (MRI) has revealed a highly variable imaging phenotype with involvement of frontal, temporal and parietal cortices, and limited previous longitudinal data have suggested similar rates of whole brain atrophy in
C9ORF72 and
MAPT mutation cases [
6]. Group-level cross-sectional imaging studies have confirmed this distributed pattern of atrophy as well as emphasising additional cerebellar and subcortical involvement [
6,
8,
9]. Both imaging and clinical studies suggest that the neurodegenerative process associated with the
C9ORF72 expansion is rather diffuse [
10]. Whilst early symptoms are most in keeping with frontal lobe dysfunction, parietal dysfunction becomes more apparent as the disease progresses [
6]. These clinical features suggest that the disease may propagate along a rostrocaudal gradient, perhaps spreading via a distributed brain network. Understanding the clinical and radiological evolution of
C9ORF72- associated FTD is an important issue. Detailed studies of longitudinal imaging profiles and neuropsychological changes associated with
C9ORF72 mutations remain limited: longitudinal studies may enable evaluation of candidate biomarkers both for diagnosis and future clinical trials of disease-modifying agents. More fundamentally, the concept of network-led degeneration is gaining currency as an important general theme in neurodegeneration [
11] and
C9ORF72-associated FTD, as a novel genetic proteinopathy, may offer fresh insights into the mechanisms of neurodegenerative disease propagation.
Here we present longitudinal data on a cohort of patients with FTD associated with C9ORF72 expansions. We detail profiles of neuropsychological progression, rates of whole brain, cerebellar and subcortical atrophy and anatomical profiles of disease progression using nonlinear fluid registration of serial MRI.
Discussion and Conclusion
We have described longitudinal neuropsychological and neuroimaging features in a cohort of patients with FTD due to a
C9ORF72 expansion. Considering the mutation group as a whole, impaired executive function and also episodic memory were early and prominent neuropsychological features. Over follow-up intervals of some 18 months, there was a significant decline in general intellect and a further decline in visual memory, naming and parietal skills, whereas nondominant parietal (visuoperceptual) functions remained relatively intact. Mean brain atrophy and ventricular expansion rates were increased compared with healthy controls and broadly in line with rates of change in previous longitudinal imaging studies of FTD [
29,
30], although there was substantial diversity across the group. Increased rates of ventricular expansion were consistently observed and may be a candidate biomarker of disease evolution associated with the
C9ORF72 mutation. Cerebellar atrophy was also a relatively prominent feature in the
C9ORF72 mutation group, with an approximately 10-fold increase in mean atrophy rate compared with controls. Notably, no specific cortical region appeared disproportionately affected; however, subcortical structures including the thalamus and globus pallidus showed mean rates of atrophy around three times greater than controls. Unlike certain other genetic variants of FTD, notably
GRN [
31], hemispheric atrophy remained largely symmetrical. In further contrast to previous neuroimaging findings in association with mutations of
GRN (asymmetric fronto-temporo-parietal atrophy) and
MAPT (antero-medial temporal lobe atrophy) [
9], individual atrophy profiles in this
C9ORF72 mutation cohort were highly variable (Figure
2) - some patients showing chiefly frontal volume loss, whilst others showed relatively more posterior volume loss. Cerebellar atrophy was a relatively consistent feature in individual cases here, although whether this is truly a signature of
C9ORF72-associated FTD requires substantiation in larger patient cohorts from different centres.
The evolution of cognitive deficits here suggests a distributed disease process implicating frontal, temporal and parietal cortices, particularly in the dominant hemisphere. Degeneration of a distributed subcortical network might reconcile this neuropsychological profile with the rather variable and diffuse profiles of brain atrophy observed here. Degeneration of thalamus, cerebellum and thalamic and frontal white matter tracts has been identified previously in cross-sectional imaging studies of
C9ORF72 expansions [
6,
8,
9]. In the present study, we provide further evidence that the pathophysiological mechanisms of
C9ORF72- associated FTD target subcortical networks: rates of thalamic and cerebellar atrophy and ventricular expansion were disproportionately increased relative to whole brain atrophy rates, consistent with involvement of subcortical structures and pathways [
32]. The involvement of the globus pallidus observed here is in line with the development of extrapyramidal symptoms in a substantial proportion of
C9ORF72 cases in other series [
33], although our patients did not manifest clear-cut features of Parkinsonism. The thalamus, globus pallidus and cerebellum together act as key hubs coordinating distributed cortico-subcortical circuits and the cognitive operations they mediate [
34,
35]. Early involvement of such hub regions and projections could facilitate diffusive spread of the molecular pathology responsible for the brain degeneration associated with
C9ORF72 expansions [
36‐
38] and might be anticipated to lead to rapid clinical evolution, although the very wide range of clinical disease durations among individual patients with a C9ORF72 mutation remains an important unsolved problem. Both the thalamus and cerebellum have been previously implicated in cross-sectional neuroimaging work in
C9ORF72-associated FTD [
6]. The increased prevalence of cerebellar p62 inclusions with
C9ORF72 expansions compared with other pathologically proven cases of FTD further supports the role of the cerebellum as an important anatomical nidus of
C9ORF72-associated pathology [
6,
39].
This study has a number of limitations. Case numbers here were relatively small and individual variation was substantial; larger (ideally, multicentre) longitudinal studies are required to establish the true range of cognitive and neuroimaging features associated with C9ORF72-associated FTD and to evaluate candidate biomarkers. The historical nature of the present cohort was a particular limitation on the systematic analysis of behavioural deficits; for example, the nature of the naming impairment here remains ill-defined, and this could in principle reflect primary word retrieval, semantic or mixed deficits. Inclusion of presymptomatic carriers in future studies may allow the earliest behavioural and neuroimaging markers of disease onset to be determined. The specificity of any candidate biomarkers will only be established by comparisons with other genetic and sporadic forms of FTD. We argue that future work should particularly target subcortical (including cerebellar) structures and cognitive functions in the C9ORF72 mutation group, incorporating neuroimaging modalities that capture white matter disintegration: although any synthesis must be preliminary, we interpret the present findings as further circumstantial evidence that a distributed cortico-subcortical network is integral to the phenotypic expression of C9ORF72-associated FTD.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
CJM contributed to the conception and design of this study, data collection, data analysis and drafting the manuscript. LED contributed to the data collection and data analysis. GRR contributed to data analysis and drafting of the manuscript. JB contributed to the genetic analysis of subjects. SC, MB, SF, KKL and TY contributed to data analysis. HG contributed to data collection. SM contributed to genetic analysis of subjects. JDR contributed to data collection. NCF contributed to conception and design of the study and review of the manuscript. JDW contributed to the conception and design of this study and drafting and critical revision of the manuscript. All authors read and approved the final manuscript.