Genetic analysis reveals the shared genetic architecture between breast cancer and atrial fibrillation

Abstract

Background: Epidemiological studies have observed an association between atrial fibrillation (AF) and breast cancer (BC). However, the underlying mechanisms linking these two conditions remain unclear. This study aims to systematically explore the genetic association between AF and BC.

Methods: We utilized the largest available genome-wide association study (GWAS) datasets for European individuals, including summary data for AF (N = 1,030,836) and BC (N = 247,173). Multiple approaches were employed to systematically investigate the genetic relationship between AF and BC from the perspectives of pleiotropy and causality.

Results: Global genetic analysis using LDSC and HDL revealed a genetic correlation between AF and BC (rg = 0.0435, P = 0.039). Mixer predicted genetic overlap between non-MHC regions of the two conditions (n = 125, rg = 0.05). Local genetic analyses using LAVA and GWAS-PW identified 22 regions with potential genetic sharing. Cross-trait meta-analysis by CPASSOC identified one novel pleiotropic SNP and 14 pleiotropic SNPs, which were subsequently annotated. Eight of these SNPs passed Bayesian colocalization tests, including one novel pleiotropic SNP. Further fine-mapping analysis identified a set of causal SNPs for each significant SNP. TWAS analyses using JTI and FOCUS models jointly identified 10 pleiotropic genes. Phenome-wide association study (PheWAS) of novel pleiotropic SNPs identified two eQTLs (PELO, ITGA1). Gene-based PheWAS results showed strong associations with BMI, height, and educational attainment. PCGA methods combining GTEx V8 tissue data and single-cell RNA data identified 16 co-enriched tissue types (including cardiovascular, reproductive, and digestive systems) and 5 cell types (including macrophages and smooth muscle cells). Finally, univariable and multivariable bidirectional Mendelian randomization analyses excluded a causal relationship between AF and BC.

Conclusion: This study systematically investigated the shared genetic overlap between AF and BC. Several pleiotropic SNPs and genes were identified, and co-enriched tissue and cell types were revealed. The findings highlight common mechanisms from a genetic perspective rather than a causal relationship. This study provides new insights into the AF-BC association and suggests potential experimental targets and directions for future research. Additionally, the results underscore the importance of monitoring the potential risk of one disease in patients diagnosed with the other.

Keywords: atrial fibrillation; breast cancer; causal inference; genetic correlation; pleiotropic gene; shared genetic etiology.

FIGURE 1

Flowchart of this study. (A) Heritability analysis using LDSC for two phenotypes: AF and BC. (B) Genetic correlation analysis using LDSC, comparing the genetic correlation between the two diseases. (C) Genetic overlap estimated by MiXeR for AF and BC. (D) PPI network analysis showing key pleiotropic genes, including DNMT3A, related to both diseases. (E) Colocalization analysis of GWAS for both diseases, showing shared genetic variants. (F) Basic assumptions in MR analysis, including assumptions related to IVs, exposure, and outcome. (G) Flowchart summarizing the genetic and causal analysis steps.

FIGURE 2

(A) represents the results of the LAVA analysis, with the X-axis representing genetic correlation and the Y-axis representing the -log10 (P _FDR ) values. The dashed line represents -log10 (0.05). (B, C) respectively represent the results for AF and BC from the GWAS-PW analysis, with the X-axis showing the maximum absolute value of the Z-score, and the Y-axis displaying data for PPA 3.

FIGURE 3

(A) A Venn diagram represents the data on the potential pleiotropic genetic overlap between AF and BC from the MiXer results. (B) QQ plots represent the results stratified by P-values for BC and AF. (C) The negative log-likelihood plot shows a minimum model score of about 15, a maximum score of about 110, and the best model score close to 0, indicating a good fit of the model.

FIGURE 4

Co-enriched tissues and cell types between atrial fibrillation and breast cancer. Red font indicates co-enrichment. (A): Represents the tissue types. (B): Represents the cell types.

FIGURE 5

(A) Results of univariate and multivariate Mendelian randomization analyses between AF and BC. This includes 7 methods of univariate MR analysis and 4 methods of multivariate MR analysis. (B) Scatter plot of univariate MR analysis from AF to BC. (C) Funnel plot of univariate MR analysis from AF to BC. (D) Leave-one-out (LOO) plot of univariate MR analysis from AF to BC. (E) Scatter plot of univariate MR analysis from BC to AF. (F) Funnel plot of univariate MR analysis from BC to AF. (G) LOO plot of univariate MR analysis from BC to AF.