Association of a Locus in the CAMTA1 Gene With Survival in Patients With Sporadic Amyotrophic Lateral Sclerosis

Abstract

Importance: Amyotrophic lateral sclerosis (ALS) is a devastating adult-onset neurodegenerative disorder with a poor prognosis and a median survival of 3 years. However, a significant proportion of patients survive more than 10 years from symptom onset.

Objective: To identify gene variants influencing survival in ALS.

Design, setting, and participants: This genome-wide association study (GWAS) analyzed survival in data sets from several European countries and the United States that were collected by the Italian Consortium for the Genetics of ALS and the International Consortium on Amyotrophic Lateral Sclerosis Genetics. The study population included 4256 patients with ALS (3125 [73.4%] deceased) with genotype data extended to 7 174 392 variants by imputation analysis. Samples of DNA were collected from January 1, 1993, to December 31, 2009, and analyzed from March 1, 2014, to February 28, 2015.

Main outcomes and measures: Cox proportional hazards regression under an additive model with adjustment for age at onset, sex, and the first 4 principal components of ancestry, followed by meta-analysis, were used to analyze data. Survival distributions for the most associated genetic variants were assessed by Kaplan-Meier analysis.

Results: Among the 4256 patients included in the analysis (2589 male [60.8%] and 1667 female [39.2%]; mean [SD] age at onset, 59 [12] years), the following 2 novel loci were significantly associated with ALS survival: at 10q23 (rs139550538; P = 1.87 × 10-9) and in the CAMTA1 gene at 1p36 (rs2412208, P = 3.53 × 10-8). At locus 10q23, the adjusted hazard ratio for patients with the rs139550538 AA or AT genotype was 1.61 (95% CI, 1.38-1.89; P = 1.87 × 10-9), corresponding to an 8-month reduction in survival compared with TT carriers. For rs2412208 CAMTA1, the adjusted hazard ratio for patients with the GG or GT genotype was 1.17 (95% CI, 1.11-1.24; P = 3.53 × 10-8), corresponding to a 4-month reduction in survival compared with TT carriers.

Conclusions and relevance: This GWAS robustly identified 2 loci at genome-wide levels of significance that influence survival in patients with ALS. Because ALS is a rare disease and prevention is not feasible, treatment that modifies survival is the most realistic strategy. Therefore, identification of modifier genes that might influence ALS survival could improve the understanding of the biology of the disease and suggest biological targets for pharmaceutical intervention. In addition, genetic risk scores for survival could be used as an adjunct to clinical trials to account for the genetic contribution to survival.

Figure 1. Genome-wide Association Study of Survival in Patients With Sporadic Amyotrophic Lateral Sclerosis (ALS)

A, Manhattan plot of the combined (METAL software) Cox proportional hazards regression analysis. The threshold for genome-wide significance after correction for multiple testing was set at P = 5 × 10⁻⁸ (horizontal blue line). Loci significantly associated with ALS survival are highlighted in red and labeled according to the corresponding genes. At locus 10q23, the most associated single-nucleotide polymorphism (SNP), rs139550538 (P = 1.87 × 10⁻⁹), was moderately rare with a minor allele frequency (MAF) of 0.03, whereas at the 1p36 locus, the 4 SNPs significantly associated (rs2412208 [P = 3.53 × 10⁻⁸], rs4584415 [P = 3.68 × 10⁻⁸], rs35447019 [P = 3.86 × 10⁻⁸], and rs4409676 [P = 4.48 × 10⁻⁸]) were common (MAF > 0.26). B and C, Regional linkage disequilibrium (LD) plots of the 2 regions significantly associated with ALS survival. At the 1p36 locus, 4 SNPs passed the genome-wide significant threshold, followed by 87 tagged proxies suggestively associated (P < 10⁻⁴). All the associated SNPs mapped within introns 3 to 4 of the CAMTA1 gene. At the 10q23 locus, the top-ranked SNP, rs139550538, intronic to the IDE gene, was in weak (r² < 0.4) LD with the tagged proxies that were located in the neighbor gene, KIF11.

Figure 2. Survival Distribution Across the Genotypes of the Top-Ranked Single-Nucleotide Polymorphisms (SNPs)

Kaplan-Meier curves plot survival distribution. A, IDE rs139550538 distribution of genotypes under a dominant model. Survival in the 226 AA/AT carriers was compared with that in 4030 TT carriers, showing that the presence of at least 1 A allele is associated with a median survival of 30.7 months compared with 36.7 months in TT homozygotes. B, Variant CAMTA1 rs2412208 genotypes under an additive genetic model show 265 GG and 1644 GT carriers with a median survival of 36.0 and 36.8 months, respectively, compared with 40.8 months in 2347 TT carriers. C, Variant CAMTA1 rs2412208 genotypes under a dominant model show survival in 1909 GG/GT and 2347 TT carriers; TT homozygotes have a life span extended by more than 4 months. Kaplan-Meier curves report patients’ survival up to 10 years, plotted in SPSS software.

Figure 3. Forest Plot of CAMTA1 rs2412208 Hazard Ratio (HR) Estimates

Hazard ratio estimates are measured under an additive genetic model across the 7 genome-wide association study (GWAS) data sets (described in detail in eTable 1 in the Supplement). The HR in each cohort was estimated in a multivariate log-additive genetic model using the pacoxph program adjusted by age at onset, sex, and population stratification, whereas the summary HR was calculated by fixed-effects meta-analysis using R library rmeta. Genotype raw data included in each study and combined in the meta-analysis were collected by the Medical Center Utrecht, Utrecht, the Netherlands (UMC-1 and -2), the Beaumont Hospital Dublin, Ireland (Ireland), Massachusetts General Hospital, Boston (MGH), the National Institutes of Health, Bethesda, Maryland (NIH-IT), the Italian Consortium for the Genetics of ALS (SLAGEN), and the UK National MND DNA and Biobank Study (UK). Dark blue boxes indicate single studies and are proportional to the sample sizes; bars indicate 95% CI. Evry indicates University of Evry and Paris, Evry, France, and Hospital de la Salpetriere, Paris, France; KCL, King’s College London. ^aIncludes samples from Umeå and Leuven.