A genome-wide association study identifies the GPM6A locus associated with age at onset in ALS

Abstract

Amyotrophic lateral sclerosis (ALS) exhibits considerable clinical variability, such as differences in age at onset (AAO). Multiple factors, including genetic factors, may underlie this variability; however, the specific determinants remain unclear. To identify genes affecting AAO, we have conducted a genome-wide association study in Japanese patients with ALS (discovery cohort: n = 1808; replication cohort: n = 207). Here, we show that the minor A allele of rs113161727 at the ADAM29-GPM6A locus is associated with a younger AAO in the discovery cohort (effect, -4.27 years; p = 4.60 × 10^-8); this finding has been confirmed in the replication cohort (p = 0.0068) and meta-analysis (p = 1.08 × 10^-9). Among 65 ALS patients with a SOD1 mutation, the AAO has been found to be 10.2 years younger in those with the A allele than in those without it (p = 0.002). This variant correlates with GPM6A upregulation in iPSC-derived motor neurons, suggesting GPM6A as a candidate AAO modifier. Overall, our study highlights the impact of genetic modifiers on ALS heterogeneity and provides a potential target for delaying disease onset.

Fig. 1. Genome-wide association analysis for age at onset in Japanese patients with amyotrophic lateral sclerosis (ALS).

a Manhattan plot for age at onset in patients with ALS in the discovery cohort (n = 1808). The horizontal gray line represents the genome-wide significance level (α = 5 × 10⁻⁸). Loci significantly associated with the age at ALS onset are highlighted in red. One locus (4q34.2) showed a genome-wide significance. b Manhattan plot of the meta-analysis of the age at onset in Japanese patients with ALS. We combined the results from the discovery (n = 1808) and replication (n = 207) cohorts in a meta-analysis (total n = 2015 independent patients). The meta-analysis confirmed the genome-wide significance of the single-nucleotide polymorphisms (SNPs) at the 4q34.2 locus, with the leading SNP being rs113161727. c Regional association plots for the 4q34.2 locus identified in the genome-wide meta-analysis (n = 2015). The vertical axis represents −log₁₀ (p-value) for assessing the association between each SNP and age at onset. Colors indicate the linkage disequilibrium (r²) between each sentinel SNP and neighboring SNPs based on the JPT population of the 1000 Genomes Project phase 3. JPT Japanese people in Tokyo, Japan.

Fig. 2. Distribution of age at onset (AAO) and the relationship between AAO and rs113161727 in Japanese patients with amyotrophic lateral sclerosis (ALS).

a Distribution of AAO in the discovery cohort (n = 1808). b Distribution of AAO in the discovery cohort according to the rs113161727 genotype. Patients with the AG or AA genotype of rs113161727 showed a younger AAO distribution than those with the GG genotype. AAO < 40 years was 11.8% in patients with the AG or AA genotype, compared to 5.3% in those with the GG genotype. The AAO ≥ 70 years was 17.7% in patients with the AG or AA genotype, compared with 29.7% in those with the GG genotype. c Relationship between AAO and rs113161727 in the discovery cohort (n = 1808). d Relationship between AAO and rs113161727 in patients with ALS with a SOD1 mutation (n = 65). The bottom and top of the box indicate the interquartile ranges (25th and 75th percentiles), and the line represents the median. Whiskers under and over the box correspond to a 1.5× interquartile range, and circles indicate outliers.

Fig. 3. Forest plots of the association between rs113161727 and the site of onset.

Forest plot shows odds ratios (ORs) with 95% confidence intervals (CIs) for the A allele of rs113161727 in relation to the site of onset. The analyses were conducted using the discovery cohort, which included 1658 patients with available clinical data. N (+) indicates the number of patients with the onset site, and N (−) indicates the number of patients without the onset site. Black squares represent the OR, and dashed horizontal error bars indicate the 95% CI. The vertical gray line indicates OR = 1. The A allele of rs113161727 significantly increased the risk of lower limb onset (p = 0.005) and tended to decrease the risk of bulbar onset (p = 0.048) but did not affect upper limb onset. OR odds ratio, 95% CI 95% confidence interval.

Fig. 4. Relative expression levels of genes surrounding rs113161727 in induced pluripotent stem cell (iPSC)-derived motor neurons from patients with amyotrophic lateral sclerosis (ALS) with each genotype of rs113161727.

The expression levels of genes (GLRA3, ADAM29, GPM6A, WDR17, and SPATA4) surrounding rs113161727 in iPSC-derived motor neurons from 20 patients with ALS were examined using real-time quantitative reverse transcription-polymerase chain reaction (RT-qPCR). The expression levels of GPM6A mRNA were significantly higher in iPSC-derived motor neurons with the AG genotype of rs113161727 (n = 10) than in those with the GG genotype (n = 10, p = 0.0039). In contrast, expression levels of the other four genes (GLRA3, ADAM29, WDR17, and SPATA4) were not affected by the rs113161727 genotype. The mRNA expression levels of each genotype were compared using the Mann–Whitney U-test. The mRNA levels of each gene were normalized to the levels of ACTB. RT-qPCR analysis included technical replicates conducted in three wells for each plate. The Cq values from the three technical replicates were averaged to generate a single raw Cq value per assay. Each black dot represents an individual sample. The bottom and top of the box indicate the interquartile range (25th and 75th percentiles), and the line represents the median. The whiskers under and over the box correspond to the minimum and maximum values. Asterisk (*) indicates p < 0.01.