Sir, We read with great interest the article published in Brain by Tezenas du Montcel et al. (2014) reporting modifying factors of age-at-onset in 1255 individuals affected with spinocerebellar ataxia (SCA) types 1, 2, 3, 6 and 7, recruited through the European Consortium on Spinocerebellar Ataxias (EUROSCA). We would like to report our findings on genetic modifiers and age-at-onset in 802 Chinese Han patients (429 males, 373 females) with Machado-Joseph disease (MJD/SCA3), on behalf of the Chinese Clinical Research Cooperative Group for Spinocerebellar Ataxias (CCRCG-SCA). Our study was approved by the ethics committee of Xiangya Hospital, Central South University. Blood samples were collected after written informed consent from all subjects. All MJD patients were genotyped for ATXN3 and nine other polyQ-related genes (ATXN1, ATXN2, CACNA1A, ATXN7, TBP, ATN1, KCNN3, RAI1 and HTT). None of the MJD patients was a ‘homozygote’ (i.e. bearing two expanded alleles). We used a similar statistical strategy as described by the EUROSCA consortium (Tezenas du Montcel et al., 2014). For some polyQ-related genes (ATXN1, TBP, ATN1 and HTT), the sizes of the shorter and the longer alleles were considered (separately) as continuous variables; given the skewed distribution of allele size in the other genes (ATXN2, ATXN3, ATXN7, CACNA1A and RAI1) these were classified as short, short-medium, medium or long (Table 1). The interaction between both alleles, their mean length and the size difference between alleles at each locus were also considered as dependent variables. Classification of polyQ-related normal alleles according to the number of CAG repeats Normal alleles of ATXN3, ATXN2, CACNA1A, ATXN7 and RAI1 were classified as short, medium, short intermediate or intermediate similar strategy according to reference of Tezenas du Montcel et al. (2014) (normal alleles of ATXN1, TBP, HTT, ATN1 and KCNN3 were considered as quantitative variables due to their almost normal distribution). Classification of polyQ-related normal alleles according to the number of CAG repeats Normal alleles of ATXN3, ATXN2, CACNA1A, ATXN7 and RAI1 were classified as short, medium, short intermediate or intermediate similar strategy according to reference of Tezenas du Montcel et al. (2014) (normal alleles of ATXN1, TBP, HTT, ATN1 and KCNN3 were considered as quantitative variables due to their almost normal distribution). The logarithmically transformed age-at-onset was treated as a dependent variable with homoscedasticity. Linear regression analysis (linear and quadratic effects) was performed to estimate the effect on age-at-onset of the expanded allele at ATXN3. Different models were then used, taking into account the expanded allele of ATXN3 and both alleles at the other nine polyQ-related genes. The determinant coefficient (R2) is the percentage explaining age-at-onset variance. All analyses were performed using the SPSS Statistics package (v.19.0). All P-values are two-tailed; a P-value ≤0.05 was considered significant. The length of the longer and shorter normal alleles for each polyQ-related locus was distributed similarly to previous reports (Tang et al., 2000; Tezenas du Montcel et al., 2014) (Fig. 1 and Table 2). Age-at-onset of patients with MJD was 34.8 ± 10.2 years (range 12–63); males (34.7 ± 10.3) and females (35.2 ± 9.8) had similar age-at-onset, as did paternal and maternal transmissions (533 parents). Mean repeat size in expanded alleles did not differ with gender of the patients and that of their transmitting parents. Type 2 (81.9%; mean age-at-onset 33.8 ± 7.3 years), type 3 (11.2%; 52.3 ± 3.7 years) and type 1 (5.7%; 16.6 ± 1.9 years) were the most common subtypes (Fig. 2 and Supplementary Table 1). Histograms of normal alleles repeat-length for ATXN3, ATXN1, ATXN2, CACNA1A, ATXN7, TBP, ATN1, HTT, KCNN3 and RAI1, in the CCRCG-SCA MJD cohort. Frequency of clinical subtypes in Chinese Han MJD subjects. Description of the CAG repeat length of polyQ-related genes Description of the CAG repeat length of polyQ-related genes Log age-at-onset was determined by size of the expanded ATXN3 allele, ranging 51–86 repeats (74.7 ± 3.6). Size of the expanded ATXN3 related to MJD subtype: type 1 tended to have larger expansions (mean, 80.3 ± 2.9; range, 70–86 repeats), type 3 had smaller ones (mean, 70.6 ± 3.3; range, 52–77), while type 2 had medium-size repeats (mean, 75.1 ± 3.1; range, 51–83) (Supplementary Table 1). R2 was 0.468 using a linear model, and 0.569 with a quadratic model increasing 10.1% compared to the linear one (Fig. 3). The combined effect of normal and expanded ATXN3 alleles was also tested: the quadratic model increased the percentage of age-at-onset variance explained (by additional 1.4% and 2.0%) with normal ATXN3 allele <19 repeats (short normal, R2 = 58.3%) and those ranging 26–40 (long normal, R2 = 58.9%), suggesting longer and shorter normal alleles explained more variability in age-at-onset than short-medium ones (Supplementary Table 2). Modulation of age-at-onset in MJD patients due to the ATXN3 expanded allele. Solid line: the regression curve of CAG repeats and logAO (LOA); dotted line: 95% confidence interval of LOA. Model parameters, left: LOA = 3.398 (±0.071, P < 0.001) −0.025 (±0.001, P < 0.001) Exp, R2 = 46.8%; right: LOA = −3.573 (±0.510, P < 0.001) +0.167 (±0.014, P < 0.001) Exp − 1.312 × 10 −3 (±0.095 × 10 −3, P < 0.001) Exp2, R2 = 56.9% (Exp: the CAG repeats of the expanded allele in ATXN3). AO = age-at-onset. Regression models incorporating the expanded allele size at ATXN3 with normal allele size at other nine polyQ-related genes were used to estimate potential modifying effects. The shorter allele at ATXN1 interacted with the expanded allele at ATXN3, contributing to 0.4% variation in age-at-onset (P = 0.043, R2 = 57.2%), tending to delay it. Likewise, the shorter KCNN3 allele delayed age-at-onset, contributing to 0.4% variation of age-at-onset (0.004 ± 0.002, P = 0.034, R2 = 57.2%). The difference between repeats number on both alleles at KCNN3 and HTT, however, interacted with the ATXN3 expansion decreasing age-at-onset: the greater the difference, the earlier the onset (KCNN3: P = 0.025, R2 = 57.5%; HTT: P = 0.016, R2 = 57.5%) (Supplementary Table 3). Patients with a long normal ATXN2 allele (27–32) had an earlier age-at-onset than those with shorter ones (0.031 ± 0.020 earlier log age-at-onset, P = 0.016, R2 = 80.1%) (Table 3). A tendency was observed in CACNA1A, ATXN7 and RAI1 for their effect on the quadratic model being more remarkable, due to increased R2: long normal homozygous CACNA1A alleles (9–17) contributed to 4.8% age-at-onset variation (R2 = 61.7%) (Supplementary Table 4); short ATXN7 alleles (<10) and long normal ones (12–17) contributed to 2.5% and 2.3% age-at-onset variation (shorter allele: R2 = 59.4%; longer allele: R2 = 59.2%) (Supplementary Table 5); long normal RAI1 allele (13–14) contributed to 7.3% age-at-onset variation (R2 = 64.2%) (Supplementary Table 6). Effects of CAG repeats of ATXN2 normal alleles on the log age at onset ATXN2 normal alleles were analyzed as qualitative variables. A1= the repeat length of shorter allele, A2= the repeat length of longer allele. The classification of A1 allele and A2 allele are indicated in the methods section and in Table 1. The ATXN2 genotypes were divided into four classes: 1 at least one short allele, 2 homozygous medium 22 CAG alleles, 3 at least one intermediate allele, and 4 at least one short intermediate allele with or without a medium 22 CAG allele. Effects of CAG repeats of ATXN2 normal alleles on the log age at onset ATXN2 normal alleles were analyzed as qualitative variables. A1= the repeat length of shorter allele, A2= the repeat length of longer allele. The classification of A1 allele and A2 allele are indicated in the methods section and in Table 1. The ATXN2 genotypes were divided into four classes: 1 at least one short allele, 2 homozygous medium 22 CAG alleles, 3 at least one intermediate allele, and 4 at least one short intermediate allele with or without a medium 22 CAG allele. To our knowledge, this cohort of Chinese Han patients with MJD is the largest ever described with the same origin. Among the five subtypes of MJD, type 2 was the most frequent in mainland China, followed by type 3, type 1, and then the rare types 4 and 5, as in other populations (Riess et al., 2008; Costa Mdo and Paulson, 2012). This percentage of MJD type 2 patients, however, may be overrated, as it did not reflect time after onset, and type 2 may change into type 1 or 3 some years later in the course of the disease. Mean age-at-onset was earlier than in the EUROSCA study, which may be related to a larger mean ATXN3 expansion size in our cohort. Contrasting with the EUROSCA report, there was no significant difference in mean age-at-onset between paternal and maternal transmissions. This may be due to diverse ethnic and geographical origin, but also to ours being a larger sample from a single population. Compared to the linear model to determine the effect of (CAG)n size on age-at-onset, an increase of 10.1% in R2 was obtained when using a quadratic fit, which better describes age-at-onset variance in MJD. The slope of the curve varied slightly through the range 51–68 repeats and showed a marked steepness thereafter, suggesting ATXN3 expansions over 68 repeats have a stronger effect. Additionally, a modulating effect of the normal ATXN3 allele, as previously observed in Brazilian and Chinese cohorts, but not by EUROSCA, was partially replicated in this study (de Castilhos et al., 2014; Tezenas du Montcel et al., 2014; Zhou et al., 2014). This may suggest ethnic differences, although we cannot exclude different methodologies and sample size or composition. Notably, the finding that age-at-onset was influenced by an intermediate ATXN2 allele, shown by EUROSCA and confirmed in our cohort, suggests that ATXN2 might be a significant genetic modifier for MJD. Further research is needed to replicate the effect we observed of the shorter ATXN1 and KCNN3 alleles increasing age-at-onset. A potential modifying effect was also found for CACNA1A, ATXN7 and RAI1, based on the better explanation of age-at-onset variance (increased R2) in different models. In conclusion, we confirmed the influence on age-at-onset in MJD of normal-size alleles at other polyQ-related loci. Although using a similar methodology and allele groups, the nomenclature used for each class of normal alleles (in ATXN3, ATXN2, CACNA1A, ATXN7 and RAI1) was different: we categorized them as ‘short’, ‘short-medium’, ‘medium’ or ‘long’ (corresponding to the classification of ‘short’, ‘medium’, ‘short intermediate’ or ‘intermediate’ used by the EUROSCA cohort). The term ‘intermediate’ might be confusing and is inappropriate, as it has been used with different meaning in most previous studies, e.g. in Huntington disease, this designation is reserved to large-normal unstable alleles that might expand into the incomplete or full-penetrance range (Sequeiros et al., 2012; Losekoot et al., 2013). Investigation of genetic modifiers of age-at-onset should not focus on single factors only, but also explore the interaction among polyQ-related genes (Bettencourt et al., 2014). A strategy combining high-throughput sequencing and big data mining to discover rare variants would be a powerful tool to identify further potential modifiers. Based on the comparison of the CCRCG-SCA cohort with other MJD populations, differences may be ethnically based, suggesting that potential therapies might also have to be genotype and population-oriented. Both population genetic analysis and functional evaluations should now be performed, to facilitate the development of therapeutic approaches to delay onset age and retard disease progression in MJD. Supplementary material is available at Brain online. The authors want to thank patients and their family members for their generous participation in this study. They are grateful to the technical support of European project on the spinocerebellar ataxias (EUROSCA) consortium. This study was supported by the National Basic Research Program (973 Program) (Nos. 2012CB944601, 2012CB517902 and 2011CB510002 to Hong Jiang), the National Natural Science Foundation of China (Nos. 81471156, 81271260 to Hong Jiang; No 31401135 to Rong Qiu; Nos. 81271262, 81571086 to Li Cao; No. 81130021 to Beisha Tang), National Department Public Benefit Research Foundation (No. 201302001 to Beisha Tang), Key Projects in the National Science & Technology Pillar Program during the Twelfth Five-year Plan Period (No. 2012BAI09B04 to Beisha Tang), Hunan Provincial Natural Science Foundation of China (No. 14JJ1008 to Hong Jiang), Clinical Research Funds of Xiangya Hospital (2014L03 to Hong Jiang), Fundamental Research Funds for the Central Universities of Central South University (No. 2014zzts078 to Zhao Chen) and the High-level medical personnel of Hunan province (‘225’ Project).
No takes yet. Share an insight, caveat, or question.
Zhao et al. (2016) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: