Nothing to declare.
New evidence for, and challenges in, linking small CGG repeat expansion FMR1 alleles with Parkinson's disease
Article first published online: 25 DEC 2012
© 2012 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Volume 84, Issue 4, pages 382–385, October 2013
How to Cite
New evidence for, and challenges in, linking small CGG repeat expansion FMR1 alleles with Parkinson's disease., , , , , , , .
- Issue published online: 11 SEP 2013
- Article first published online: 25 DEC 2012
- Accepted manuscript online: 30 NOV 2012 12:09PM EST
- Manuscript Revised: 19 NOV 2012
- Manuscript Accepted: 19 NOV 2012
- Manuscript Received: 20 SEP 2012
- National Institutes of Child Health and Human Development. Grant Number: 36071
- NHMRC project. Grant Number: 330400
- CGG repeat;
- FMR1 gene;
- fragile X;
- grey zone alleles;
- Parkinson's disease;
- screening results
We recently reported a significant increase in the frequency of carriers of grey zone (GZ) alleles of FMR1 gene in Australian males with Parkinson's disease (PD) from Victoria and Tasmania. Here, we report data comparing an independent sample of 817 PD patients from Queensland to 1078 consecutive Australian male newborns from Victoria. We confirmed the earlier finding by observing a significant excess of GZ alleles in PD (4.8%) compared to controls (1.5%). Although both studies provided evidence in support of an association between GZ-carrier status and increased risk for parkinsonism, the existing evidence in the literature from screening studies remains equivocal and we discuss the need for alternative approaches to resolve the issue.
Fragile X-associated tremor/ataxia syndrome (FXTAS)  is a progressive late-onset neurodegenerative disorder associated with the pre-mutation (PM) allele of the fragile X mental retardation (FMR1) gene, which contains small expansions of CGG repeat in the promoter region ranging 55–200 . Characteristic features of FXTAS are intention tremor, ataxia and cognitive decline [1, 3-5], with parkinsonian symptoms occurring in nearly half of the affected individuals . The underlying neuropathology comprises widespread cerebellar and cerebral white matter loss [1, 3, 5, 6], associated with the presence of ubiquitin-positive intranuclear inclusions throughout the brain . These manifestations have been associated with increased blood FMR1 mRNA levels , suggesting RNA-mediated toxicity [1, 3, 6]. The FMR1 alleles with CGG repeat expansions ranging from 45 to 54 CGGs and named ‘grey zone’ (GZ)  are also associated with elevated FMR1 mRNA levels . However, since the threshold for this increase is 40 CGGs , we have considered 40–54 range as a more appropriate definition of GZ alleles. The correlation between repeat length and FMR1 mRNA levels continues throughout the GZ and PM ranges , which strongly suggest that the same pathophysiology may lead to FXTAS or other neurological involvements in GZ, as well as PM carriers. However, evidence for this is still lacking apart from rare reports concerning female carriers [9, 10].
The reason for difficulties in establishing possible links between neurological involvement and GZ alleles through fragile X families is that, unlike the PM alleles, they do not expand sufficiently rapidly to reach the >200 CGG (‘full mutation’) range  over two generations. This implies that these carriers cannot be ascertained through clinical admissions of their grandchildren, where full mutation leads to severe developmental disorder, Fragile X Syndrome . Hence, screening for these alleles among patients diagnosed with tremor or/and ataxia disorders has been an alternative practical option.
We previously screened an unselected sample of 228 Australian male Parkinson's disease (PD) patients (from Tasmanian and Victorian states) for GZ (40–54 CGGs) alleles . The results, compared with the frequencies in a sample of 578 Guthrie spots from consecutive Tasmanian male newborns, showed an increase in GZ carriers (6.8% vs 3.3% in controls), with odds ratio (OR) = 2.36, and 95% confidence intervals (CI: 1.20–4.63).
Here, we present the results of an independent case–control study that compares the frequency of GZ allele carriers in a cohort of 817 males with PD from the Queensland Parkinson's Disease Project (QPP)  to data from a new sample of newborn controls, which support our earlier finding by showing a significant (p = 0.00013) excess of these alleles in patients with parkinsonism.
Materials and methods
Data collection and analysis from PD patients were approved by Human Ethics Committees of the Griffith University in Brisbane, and anonymous screening of a control sample of the Victorian newborns by the Human Ethics Committee at the Royal Children's Hospital in Melbourne. All PD participants gave informed consent for their involvement in the study. The cases and controls were white Caucasians of European origin, residing in Australia, with a small admixture (∼5%) of individuals of the Asian origin.
The PD sample comprised 817 male patients aged 71 ± 10 (SD) years, with an average age of disease onset of 59 ± 12 (SD) years. This sample is a subset of a larger cohort of over 3000 community dwelling individuals recruited since 2005 to participate in the QPP . All patients were clinically diagnosed with idiopathic PD by movement disorder neurologists using the UK Brain Bank criteria . Normal population-based data were obtained from a sample of 1094 consecutive newborn Guthrie blood spots collected, in an anonymous fashion, from the Australian males born in the State of Victoria in the years 2007/2008.
CGG repeat sizing
CGG repeat size was measured on DNA extracted from whole blood/dry blood spots. Testing of the whole PD sample and a proportion of the Victorian newborn sample (for cross-checking and validation) was performed in the MIND Institute, UC Davis, using polymerase chain reaction (PCR) as previously described [15, 16]. Testing of the Victorian newborn spots was conducted at the Victorian Clinical Genetics Services, Royal Children's Hospital in Melbourne, using PCR conditions as described in . All assays have been fully validated by internal and external quality assessment to provide precision of +/− one repeat in the range 40–60 CGG repeats.
We identified 39 GZ (40–54 CGG range) carriers in the sample of 817 PD males (4.8%), compared with 19 in the sample of 1094 Victorian male newborns (1.5%). This represented a highly significant excess regardless of variously defined lower bound (Table 1).
|Normal controls||QPP PD cohort|
|OR (95% CI) = 2.61 (1.44–4.76), p = 0.00031 (Fisher's exact)|
|OR (95% CI) = 2.57 (1.34–4.96), p = 0.001 (Fisher's exact)|
|OR (95% CI) = 4.06 (1.21–14.96), p = 0.007 (Fisher's exact)|
Demographic data on comparison between the identified carriers of the FMR1 GZ alleles and those carrying the normal alleles revealed no differences in the patients' age, disease duration, the presence of other risk factors for parkinsonism such as cigarette smoking or pesticide exposures (data not shown). While not reaching statistical significance, the GZ carriers claimed a higher frequency of first-degree relatives with PD (21%) compared to the non-GZ carriers (13.2%); OR (95%) CI = 1.70 (0.72–4.01).
The present data obtained from a large cohort of the Australian male PD patients from the state of Queensland were consistent with our earlier finding in a sample of PD patients from the states of Victoria and Tasmania  of an excess of GZ allele carriers in PD (6.8%) compared with the Tasmanian controls (3.3%). This provides supporting evidence for an association between GZ-carrier status and increased risk for parkinsonism. Considering that the effect of a ‘toxic’ FMR1 transcript on neurodegeneration sharply increases after the age of 70 , a somewhat lower frequency of carriers in the present sample compared with the earlier sample may be attributed to a slightly younger age of the former. The frequencies in both these cohorts are close to the estimate of 5.1% of GZ carriers (41–54 CGGs) in a sample of 137 Norwegian/German males with PD, compared with 1.6% in 310 ethnicity-matched controls . A significant excess of GZ alleles (11.5%) compared to collected controls was also reported in a single US study of women with this disorder , and in another (Chinese) cohort of 360 females, where 11 GZ carriers were identified compared with 1 GZ carrier among 295 gender and age-matched controls (p < 0.005) .
However, the existing evidence from screening studies is still controversial. No significant excess of GZ alleles (41–54 repeats) was reported in 414 male patients with parkinsonism , but this sample comprised patients from several geographical locations and ethnic groups, and the results were compared with previously published data from other populations. A subsequent study, which compared 903 Caucasian males with either early and late onset PD or essential tremor to earlier population norms from French Canadians, also reported no significant excess of GZ allele (45–54) carriers . One other study compared only 66 male patients (with non-specified parkinsonism) to 74 normal controls from the same population and, again, reported no association .
These negative studies have had some important limitations not only with respect to sample size, but also ethnic and geographic differences in case and control samples; and differences in case ascertainment and phenotype by utilizing more stringent or wider criteria for diagnosis or pre-selection. Moreover, the instability of frequency estimates is the inherent problem associated with underpowered studies seeking to test for association between low frequency alleles and an outcome such as parkinsonism. Modelling the cumulative allele frequency in controls as estimated from consecutive batches of 80–90 newborn genotypes screened in this study for CGG sizes shows that this frequency stabilizes to the actual population estimate once approximately 2000 alleles are examined (data not shown). This suggests that much larger samples than reported earlier (especially of controls) are required to make a confident assessment of the true effect size of the association between GZ alleles and PD.
A general dilemma concerning case–control studies is the choice of most appropriate control samples. Anonymous screening of newborn samples has the advantages of being population based and unselected; however, a disadvantage is in a large age gap between the patients affected with the late-onset condition and the newborn. This is combined with a lack of knowledge about possible role of GZ alleles in predisposing to or preventing other conditions which may impact an individual's life span. An additional bias is related to the fact that the newborn cohort contains individuals destined to develop parkinsonism. However, because of the low prevalence of parkinsonian disorders this is expected to be negligible; moreover, if GZ alleles are indeed related to an increase in risk of these disorders, this would actually bias results towards the null. Other types of control groups such as community dwelling older individuals, while better matched to the cases in many respects, may not be representative of the population norms.
We have recently applied a novel approach in which detailed clinical assessment and genetic testing were performed in males with parkinsonism who were either GZ allele or normal allele carriers . This study not only provided direct evidence for the role of GZ alleles in Parkinson's spectrum disorders by showing significant correlations between the size of CGG repeat within this range and severity of parkinsonism, but it also revealed that pathogenic changes in blood leucocytes and whole blood (such as mitochondrial dysfunction) were much more severe than those found in the sample of disease controls.
In summary, the latest epidemiological and molecular data provide strong evidence for the role of FMR1 GZ alleles in the development of Parkinson's spectrum disorders. Considering that approximately 1 in 100 male and 1 in 20 female newborns are carriers of GZ alleles, large-scale population-based studies are recommended to confirm this role and the possible link to disorders associated with tremor, ataxia and cognitive decline. However, this would require large cohorts with the follow up or retrospective data on medical history and survival relating to GZ-carrier status. A potentially more powerful yet less extensive approach is suggested by our previous study , which relied on comparisons of clinical and pathogenic changes at the cellular level between GZ carriers and non-carriers identified from among patients with Parkinson's spectrum disorders or the other relevant neurological syndromes.
This study was supported by the National Institutes of Child Health and Human Development Grant HD 36071, and NHMRC project Grant No 330400, to D. Z. L.
- 11The physical and behavioural phenotype. In: Hagerman RJ, Hagerman PJ, eds. Fragile X syndrome. Diagnosis, treatment and research, 3 edn. Baltimore, MD: John Hopkins University Press, 2002: 3–109..