TL;DR
This study presents the largest genome-wide association study (GWAS) to date for varicose veins (VVs), analyzing over 810,000 individuals. It identified 49 significant genetic signals at 46 loci, implicating genes involved in extracellular matrix, inflammation, and vascular development, and demonstrated the utility of a polygenic risk score for predicting disease severity.
What Problem Does This Paper Address?
Varicose veins (VVs) are a highly prevalent condition, affecting up to a third of Western populations, and can lead to serious complications like venous ulceration, incurring significant healthcare costs. Despite a known familial component and heritability of around 17%, the underlying genetic architecture of VVs has remained largely elusive. Previous genetic studies have been limited in sample size and replication, leaving a gap in our understanding of the genetic factors contributing to VV development and severity. This paper aims to address this by conducting a large-scale, two-stage GWAS to identify genetic loci associated with VVs and to explore the biological pathways involved.
Methods
The study employed a two-stage genome-wide association study (GWAS) design.
- Discovery GWAS: A large-scale GWAS was conducted using data from the UK Biobank cohort (401,656 individuals). This stage identified initial genetic associations with VVs.
- Replication GWAS: Top genetic signals from the discovery phase were then tested for association in an independent cohort from 23andMe (408,969 individuals) to confirm their validity.
- In Silico Annotation: Tools like FUMA (Functional Mapping and Annotation of GWAS) were used to annotate the identified genetic variants, predict their functional consequences (e.g., missense variants), and identify potential regulatory elements.
- Fine-mapping: Functionally informed fine-mapping techniques were applied to narrow down the list of likely causal variants within the associated loci.
- Gene Mapping: Multiple approaches were used to map genes to the identified loci, including positional mapping, eQTL (expression quantitative trait loci) mapping, gene-based association analysis (MAGMA), and summary-based Mendelian randomization (SMR).
- Pathway and Gene Set Enrichment Analysis: MAGMA and XGR (eXploring Genomic Relations) were used to identify enriched biological pathways and gene sets associated with the identified genes, providing insights into the biological mechanisms underlying VVs.
- Polygenic Risk Score (PRS) Analysis: A PRS was derived using summary statistics from an independent cohort (FinnGen) and then tested for its predictive utility in the UK Biobank cohort, specifically examining its correlation with VV surgery as a proxy for disease severity.
- Phenome-wide VV Associations: Genetic correlation analysis (LD score regression) was used to explore associations between VVs and other phenotypes.
Datasets
- UK Biobank: This large population-based cohort served as the discovery dataset for the GWAS. It included approximately 401,656 individuals, with detailed phenotypic data and genome-wide genotyping.
- 23andMe Research Participants: This cohort was used for replication of the GWAS findings, comprising approximately 408,969 individuals with self-reported VV status.
- FinnGen: This Finnish biobank resource was used to derive the polygenic risk score (PRS) for VVs. It included approximately 17,027 VV cases and 190,028 controls.
The study did not introduce new benchmark datasets but rather leveraged existing large-scale biobanks for its analyses.
Key Insights and Ideas
- Unveiling the Genetic Architecture: The study significantly expands our knowledge of the genetic underpinnings of VVs by identifying a substantial number of novel genetic loci, suggesting that VVs are a polygenic trait influenced by many genes.
- Biological Plausibility of Loci: The identified genes and pathways are highly biologically relevant to vascular health, including extracellular matrix regulation, inflammation, and vascular smooth muscle cell function, providing strong candidates for further research into VV pathogenesis.
- PIEZO1 as a Key Player: The identification of PIEZO1 as a significant locus is particularly interesting. This gene encodes a mechanosensitive ion channel crucial for sensing mechanical stress in blood vessels, directly linking mechanical forces to VV development.
- Inflammation and Immune Response: The enrichment of inflammatory pathways highlights the role of immune system dysregulation in VV development, suggesting potential therapeutic targets related to modulating inflammation.
- PRS for Disease Stratification: The successful development and validation of a PRS for VVs demonstrate its potential for predicting disease risk and, importantly, for stratifying patients based on disease severity (e.g., correlation with surgery), paving the way for personalized medicine approaches.
- Genetic Overlap with Other Conditions: The finding of a genetic correlation between VVs and systemic lupus erythematosus (SLE) suggests shared underlying biological mechanisms, opening avenues for cross-disease research.
Key Results
- 49 Significant Associations: The study identified 49 independent genetic variants at 46 distinct susceptibility loci associated with varicose veins.
- 135,514 Cases and 675,111 Controls: The combined sample size across the discovery and replication cohorts was substantial, providing high statistical power.
- 237 Genes Mapped: A total of 237 genes were mapped to the identified susceptibility loci, with several showing strong biological plausibility.
- Heritability Estimate: The estimated SNP heritability for VVs was around 8.04% in the UK Biobank and 5.40% in the 23andMe dataset.
- PRS Predictive Utility: Individuals in the top PRS decile had a markedly increased odds of VV diagnosis (OR 4.57) compared to the bottom decile, and PRS was significantly higher in VV cases who underwent surgery.
- Enriched Pathways: Pathway analysis revealed significant enrichment in extracellular matrix biology, inflammation, (lymph)angiogenesis, vascular smooth muscle cell migration, and apoptosis.
Limitations and Future Work
- Phenotyping Differences: The study acknowledges that the phenotyping for the replication GWAS in 23andMe was less stringent (self-report only) compared to the UK Biobank (which included diagnostic and surgical codes), potentially impacting the replication of sub-threshold signals.
- In Silico Analyses: The in silico annotation of SNPs was primarily based on summary statistics from the discovery GWAS, and access to full summary statistics for the replication cohort was not available.
- Functional Validation: While the study identified biologically plausible genes, functional validation in diseased venous tissue is required to definitively link candidate variants and genes to VVs.
- eQTL Data Limitations: The eQTL analysis was limited by the lack of access to venous tissue eQTL data, relying instead on GTEx tibial artery tissue data as a surrogate.
- Future Directions: The authors suggest that further investigation into the identified genes and pathways could lead to novel therapeutic targets. The development of a PRS also opens doors for improved prognostication and personalized management strategies for VVs patients, potentially guiding early intervention or identifying those at higher risk of recurrence.
Summary updated