TL;DR
This study identifies 13 new genetic risk loci for Restless Legs Syndrome (RLS) by performing a large-scale meta-analysis of genome-wide association studies (GWASs). These findings strongly support a neurodevelopmental origin for RLS and highlight specific genes and pathways involved in axon guidance, synapse formation, and neuronal specification, paving the way for future functional research and potential new therapeutic targets.
What Problem Does This Paper Address?
Restless Legs Syndrome (RLS) is a common, chronic neurological disorder affecting up to 10% of individuals of European ancestry, yet its underlying genetic basis and pathophysiology remain poorly understood. While current treatments exist, they can lead to severe side effects, underscoring the urgent need for a clearer understanding of the disorder's molecular mechanisms to develop better therapeutic options. Previous GWASs had identified six risk loci, with MEIS1 being the most significant, but these findings were insufficient to fully elucidate the complex nature of RLS. This paper aims to expand on this by conducting a much larger meta-analysis to discover novel genetic risk factors and pathways associated with RLS.
Methods
The researchers employed a multi-stage approach:
- Discovery Meta-Analysis: They combined summary statistics from three large GWAS datasets (EU-RLS GENE, INTERVAL, and 23andMe) using a fixed-effect inverse-variance meta-analysis. This approach pools data from multiple studies to increase statistical power and identify common genetic variants associated with RLS.
- Replication: Significant genome-wide signals from the discovery stage were then tested for replication in an independent GWAS dataset. This crucial step helps to confirm the findings and reduce the likelihood of false positives.
- Joint Analysis: A final analysis combined the discovery and replication datasets to provide robust estimates of association for all identified loci.
- Gene Annotation, Pathway, and Gene-Set Enrichment Analyses: To understand the biological implications of the identified genetic variants, the researchers performed various bioinformatic analyses. Notably, they developed a novel algorithm called BI-ENRICH to identify enriched pathways, which builds upon biclustering techniques used in gene-expression analysis. They also utilized DEPICT for gene prioritization and enrichment.
- Genetic Correlation Analysis: They assessed the genetic overlap between RLS and other traits (neurological, psychiatric, sleep-related, and educational) using LD-score regression to understand shared genetic underpinnings.
Datasets
- Discovery Datasets:
- EU-RLS GENE: 6,228 cases and 10,992 controls.
- INTERVAL: 3,065 cases and 24,923 controls.
- 23andMe: 5,833 cases and 59,810 controls.
- Total Discovery: 15,126 cases and 95,725 controls.
- Replication Dataset:
- 23andMe (independent population): 30,770 cases and 286,913 controls.
- Total for Joint Analysis: 45,896 cases and 382,638 controls.
The paper notes that a subset of samples from EU-RLS GENE and INTERVAL had been used in previous GWASs. The 23andMe dataset used a single research question for phenotype ascertainment.
Key Insights and Ideas
- Neurodevelopmental Basis of RLS: The study provides strong evidence that RLS is fundamentally a neurodevelopmental disorder. The identified genes and pathways are heavily involved in processes like neurogenesis, axon guidance, and synapse formation, suggesting that disruptions in early brain development or ongoing neural circuit maintenance contribute significantly to RLS pathogenesis.
- Novel Therapeutic Avenues: The identification of specific genes like CRBN (encoding cereblon) opens up exciting possibilities for drug repurposing. Since cereblon is a target of thalidomide, and thalidomide has been shown to modulate Meis2 (another identified gene), this suggests thalidomide or its analogues could be potential therapeutic candidates for RLS, even in women beyond reproductive age.
- BI-ENRICH Algorithm: The development and application of the BI-ENRICH algorithm represent a methodological contribution, offering a novel way to identify enriched pathways, particularly when dealing with complex phenotypes and potentially limited sample sizes for certain analyses.
- Genetic Correlation with Psychiatric Traits: The finding of positive genetic correlations with depressive symptoms and neuroticism, while low to moderate, aligns with epidemiological observations and suggests shared genetic influences between RLS and certain psychiatric conditions.
Key Results
- 13 New Risk Loci: The meta-analysis identified and replicated 13 novel genetic risk loci for RLS, in addition to confirming the six previously known loci, bringing the total to 19.
- MEIS1 Remains Strongest Locus: The MEIS1 locus was reconfirmed as the strongest genetic risk factor for RLS, with an odds ratio of 1.92.
- Pathway Enrichment: Bioinformatic analyses highlighted pathways related to neurogenesis, axon guidance (e.g., SEMA6D), synapse formation (NTNG1), and neuronal specification (e.g., HOXB cluster family, MYT1). DNA-damage repair pathways were also found to be relevant.
- SNP Heritability: The identified 20 independent association signals explained 11.7% of the observed variance in RLS, which is 60% of the total SNP heritability (estimated at 19.6%).
- Polygenic Risk Score: Individuals in the highest polygenic risk score group (>4.4) had a significantly increased risk of RLS (OR 17.6), though the overlap between cases and controls precluded its use for individual risk prediction.
- Genetic Correlations: Significant genetic correlations were found with depressive symptoms (positive) and neuroticism (positive), and a positive correlation with the number of children. Negative genetic correlations were observed with educational attainment.
Limitations and Future Work
- Common Variants Focus: The study focused on common variants (minor allele frequency ≥1%). Rarer variants and structural variations, which could have larger effect sizes, were not assessed and may explain a portion of the heritability. Future whole-genome sequencing efforts are needed to investigate these.
- Phenotype Heterogeneity: The study acknowledges potential between-study heterogeneity, partly due to variations in phenotype ascertainment (e.g., clinical interviews vs. self-report questionnaires). The 23andMe and INTERVAL datasets, in particular, lacked full clinical validation. More detailed clinical assessments in future studies are recommended to dissect the specific roles of genetic variants and their interactions with different RLS phenotypes.
- Functional Studies Needed: While the study identifies candidate genes and pathways, extensive functional studies are required to elucidate the precise molecular mechanisms underlying RLS pathogenesis and to validate potential therapeutic targets.
- Drug Repurposing Trials: The potential for repurposing thalidomide for RLS is promising but requires carefully designed clinical trials to confirm efficacy and safety.
- Tissue Enrichment Limitations: Significance in tissue enrichment analyses was sometimes hampered by a lack of appropriate samples from specific anatomical regions or developmental stages in the input datasets.
Summary updated