Genetic and Metabolic Determinants of Atrial Fibrillation in a General Population Sample: The CHRIS Study

Abstract

Atrial fibrillation (AF) is a supraventricular arrhythmia deriving from uncoordinated electrical activation with considerable associated morbidity and mortality. To expand the limited understanding of AF biological mechanisms, we performed two screenings, investigating the genetic and metabolic determinants of AF in the Cooperative Health Research in South Tyrol study. We found 110 AF cases out of 10,509 general population individuals. A genome-wide association scan (GWAS) identified two novel loci (p-value < 5 × 10^-8) around SNPs rs745582874, next to gene PBX1, and rs768476991, within gene PCCA, with genotype calling confirmed by Sanger sequencing. Risk alleles at both SNPs were enriched in a family detected through familial aggregation analysis of the phenotype, and both rare alleles co-segregated with AF. The metabolic screening of 175 metabolites, in a subset of individuals, revealed a 41% lower concentration of lysophosphatidylcholine lysoPC a C20:3 in AF cases compared to controls (p-adj = 0.005). The genetic findings, combined with previous evidence, indicate that the two identified GWAS loci may be considered novel genetic rare determinants for AF. Considering additionally the association of lysoPC a C20:3 with AF by metabolic screening, our results demonstrate the valuable contribution of the combined genomic and metabolomic approach in studying AF in large-scale population studies.

Keywords: Cooperative Health Research in South Tyrol; GWAS; atrial fibrillation; familial aggregation; metabolomics; rare alleles.

Figure 1

Manhattan plot. Red line indicates genome-wide significance level of 5 × 10⁻⁸. X: Chromosome X (not included in the GWAS). See Supplementary Figure S1 for the QQ plot.

Figure 2

Schematic representation of genomic localization of 4 significant SNPs. The red circles indicate the confirmed SNPs by Sanger sequencing.

Figure 3

Minimal pedigree containing investigated AF cases. For clarity, the pedigree starts with the second generation and displays only the successors of persons with AF; a full pedigree spanning up to 5 generations is provided in Figure S2. Gray symbols indicate known, non-participating, and therefore not phenotyped individuals without further information. The three AF cases are highlighted by black symbols, whereas white symbols refer to unaffected study participants. Age groups (in decades) are provided in the second line below each individual: 20s, age 18 to 29 years; 30s, age 30 to 39 years; etc.; 70+, 70 years or older. The third line presents information on the 10 s ECG P wave, which is either zero (P = 0) or not zero (P > 0). The last line provides detailed genotype information for the loci of interest. Homozygous carriers of the reference allele of variant rs745582874 on chromosome 1 are shown as “aa”, and heterozygous carriers of this variant are marked by “aA”. The same type of information is encoded for rs768476991 on chromosome 13 using the letters “bb” and “bB”, respectively. Genotype information refers to the HRC-imputed data set unless followed by an additional “(S)”, which denotes confirmation by Sanger sequencing. A hyphen (“-“) is displayed for missing genotype information. Persons with a question mark (“?”) inside their symbol have an incomplete phenotype characterized by missing information on drug intake.

Figure 4

Left: Volcano plot representing the results from the targeted metabolomics analysis. coef_AF: log2 difference of metabolite concentrations between AF cases and controls. Significant metabolites are highlighted in red. Right: lysophosphatidylcholine a C20:3 concentration (adjusted for age, sex, BMI, and AF therapy) in controls (CTRL) and AF cases.

Figure 5

Lysophosphatidylcholine a C20:3 concentration (adjusted for age, sex, BMI and AF therapy) in nine family members and other CHRIS participants.