Introduction

Machado–Joseph disease (MJD) (MIM 109150), also known as spinocerebellar ataxia type 3 (SCA3), is an autosomal dominant neurodegenerative disorder of late onset that involves the cerebellar, ocular motor, pyramidal, extrapyramidal and peripheral motor systems. Mean age at onset is 40.2 years, with extremes of 7 and 70 years, and mean survival time being 20 years (Coutinho 1992). The clinical heterogeneity presented by MJD patients led to their classification into three clinical types, which can be occasionally present in a single family, with cerebellar ataxia and ophthalmoplegia being common to all of them (Coutinho and Andrade 1978).

The MJD locus was mapped to 14q32.1 (Takiyama et al. 1993). An expansion of a triplet repeat with a CAG motif at exon 10 of the ATXN3 gene was identified as the causative mutation for this disorder (Kawaguchi et al. 1994; Ichikawa et al. 2001). The wild-type alleles present between 12 and 44 CAG units, whereas the expanded alleles contain between 61 and 87 CAG repeats (Maciel et al. 2001). Molecular testing by direct detection of the mutation is available for MJD, allowing predictive testing and prenatal (Sequeiros et al. 1998) and preimplantation genetic diagnosis (Drüsedau et al. 2004). As the estimated penetrance of the gene is almost complete (~98%) (Sequeiros 1993), a result of “carrier” implies that the individual will become affected. Each affected individual (assumed to be heterozygous) has an a priori risk of 50% of transmitting the disease allele to each offspring. Individuals at risk mainly base their future expectations upon this a priori risk. However, MJD penetrance displays an age-dependent pattern, implying that the probability of being a mutation carrier diminishes with age of the asymptomatic individuals (Sequeiros 1996). Thus, the a posterior risk or residual risk status (RRS) may differ considerably depending on the age of the individual.

SCAs are considered rare disorders, as their prevalence has been estimated to be ~3 in 100,000. As a member of the SCA group, MJD is considered the most common form of SCA worldwide (Cagnoli et al. 2006). The Azores, a group of Portuguese Islands to where the first descriptions of MJD trace (Nakano et al. 1972; Woods and Schaumburg 1972), remain the most important cluster of this disease. To date, 32 extended MJD families with origins in Flores, S. Miguel, Terceira and Graciosa islands, were identified (MJD prevalence by January 2008 ~1:3,472). In Flores Island, although the prevalence values have been progressively decreasing in the last 10 years, ranging from 1:103 (Lima et al. 1998) to 1:239 (by January 2008), MJD still reaches its highest worldwide value, constituting a public health problem.

Several mechanisms may explain the relatively high representation of MJD worldwide. The hypothesis of a higher fitness in Portuguese MJD families has been investigated, but no differences were found between cases and controls (Lima 1996; Abade et al. 2000; Santos et al. 2000; Lima et al. 2001). An alternative hypothesis, which may also explain the high representation of the disease, could be the presence of segregation ratio distortion (SRD) in favour of the expanded alleles. Results concerning SRD in previous studies, based on different methodologies (Ikeuchi et al. 1996; Riess et al. 1997; Takiyama et al. 1997; Iughetti et al. 1998; Grewal et al. 1999), are not consensual and therefore do not elucidate this possibility.

The availability of an extended genealogical database for the Azorean MJD families, associated with the thorough follow-up of patients, as well as previous work on SRD using normal families (Bettencourt et al. 2008), provided the background to conduct our study. The aim of this study was to analyse the segregation patterns of ATXN3 alleles to determine whether SRD in favour of expanded alleles could constitute one of the mechanisms responsible for the maintenance of the high levels of MJD prevalence.

Subjects and methods

Risk assessment

Age at onset for a total of 176 patients from the Azores was used to determine the probability of detectable gene expression in this population. Given the fact that a higher mean age at onset was described for patients from Flores Island (Coutinho 1992; Bettencourt et al. 2005), data for 78 patients with origin in this island was used separately to calculate the probability of heterozygosity (RRS) for at-risk individuals from Flores. For risk assessment purposes, the MJD gene (whose penetrance has been estimated to be 98%) was considered fully penetrant. Given the age-dependent penetrance, Bayesian analyses were used, according to Young (2007), to calculate the RRS for individuals who are still asymptomatic at different ages.

Segregation analysis

Detailed clinical and demographic information from 32 extended Azorean MJD families (which comprise ~900 sibships) was compiled by our group. A sample of 330 offspring of MJD patients (62 sibships) belonging to 18 extended families originating from Flores Island was selected for this work. Sibship selection was made according to the following criteria: (a) complete sibship of offspring is known, (b) all offspring were born between 1892 and 1965 (to warrant that by December of 2007, those individuals would be >42 years old, which corresponds to the average MJD onset), (c) known disease status by clinical observation (affected/not affected at the time of the study or at the time at death) according to criteria defined by Coutinho and Andrade (1978), and (d) known age at December 2007 and/or age at death.

In attributing the probability of inheriting the expanded ATXN3 allele, a probability of 1 was given to individuals who presented the disease phenotype (116 individuals). For individuals who were still asymptomatic by December 2007 (168 individuals) or at time of death (162 individuals), the age-dependent RRS assessed above was attributed to the correspondent ages. The formula [1 − probability of inheriting the expanded allele] was used to calculate the probability of inheriting the normal ATXN3 allele. The sum of those probabilities was used to obtain the proportions of transmission of expanded and normal alleles, respectively.

The χ 2 goodness-of-fit test was used to compare proportions of transmission of expanded and normal alleles, considering the expected segregation ratio of 1:1. Statistical analyses were performed using SPSS 15.0 (SPSS Inc. 2006).

Results

Risk assessment

For the total Azorean series of MJD patients (176 individuals), mean age at onset was 40.69 ± 0.95 [standard error (SE)] years, with extremes of 5 and 70 years. For the Flores series, the mean age at onset was 42.53 ± 1.65 (SE) years, which was significantly higher (two-tailed Mann–Whitney test, Z = −2.011, P = 0.044) than that obtained when excluding Flores patients from the total series [39.23 ± 1.09 (SE)].

The probability of detectable gene expression (Table 1) older than 25 years was lower when departing from onset data of Flores series than that obtained when using onset data of the total Azorean series of patients or even of previously published data (Sequeiros 1996). This fact was translated into an increased RRS for later ages in the Flores series. Therefore, probability computations using series of patients with compatible origins are essential for proper assessment of RRS, which could be useful for family members who choose not to be tested or want to know more about their RRS in a pretest situation. Moreover, as RRS values differ depending on the series of patients used for calculations, segregation analysis was performed using RRS obtained from the Flores series.

Table 1 Age-dependent risk for clinically unaffected individuals with an a priori risk of 50% of developing Machado–Joseph disease
Full size table

Segregation analysis

Segregation analysis showed no significant deviation from the mendelian 1:1 assumption (Table 2). However, there was a tendency in favour of transmission of normal alleles (~53%), which was more marked in the case of maternal transmission (~56%).

Table 2 Frequency of transmission of normal and expanded alleles by Machado–Joseph disease patients
Full size table

Discussion

In MJD, age at onset can display important differences between series of patients, as previously reported (Coutinho 1992; Bettencourt et al. 2005) and confirmed by our results. Therefore, it is important to consider such variation in the assessment and attribution of RRS in the context of genetic counselling (GC).

Research previously developed concerning the thematic of segregation patterns of the ATXN3 gene by MJD patients produced conflicting results, not clarifying whether SRD could be one of the mechanisms responsible for the maintenance of high levels of MJD prevalence worldwide. In agreement with our results, Grewal et al. (1999), using a single-sperm analysis approach, found no SRD but observed a tendency in favour of the transmission of normal alleles (53.5% normal: 46.5% expanded). Takiyama et al. (1997), however, using the same approach as Grewal et al. (1999), reported SRD favouring the expanded alleles (61%). In family studies, Ikeuchi et al.(1996) and Iughetti et al.(1998) identified SRD favouring the expanded alleles in paternal transmissions (73% and 59%, respectively), whereas Riess et al. (1997) observed the same phenomenon but in maternal transmissions (72%). Issues related to criteria for family selection in these studies could at least partially explain the conflicting results. If families were recruited for linkage or other genetic studies, as suggested by Grewal et al. (1999), the tendency in favour of transmission of expanded alleles could be highly influenced by a sampling bias. Our work provides the analysis of complete sibships, comprising 330 transmissions, with a mean ~5 offspring per transmitter (minimum 1; maximum 14) and a range of 0–7 affected and 0–11 unaffected offspring. Moreover, to our knowledge, this is the first clinical study that has taken into consideration the residual risk status for asymptomatic individuals. Data obtained in this study does not support the existence of SRD in favour of expanded alleles. Furthermore, a tendency for a preferential transmission of normal alleles is shown, being in agreement with what was described for normal families of Azorean origin (Bettencourt et al. 2008) for whom SRD in favour of the smaller normal alleles is reported. Thus, our results do not support SRD as a mechanism on the basis of the high values of MJD prevalence.