Hypertrophic cardiomyopathy: a genome wide association meta-analysis and polygenic risk score

This study identifies novel genetic loci associated with hypertrophic cardiomyopathy (HCM) and demonstrates that a newly developed polygenic risk score effectively predicts HCM diagnosis, imaging phenotypes, and adverse clinical outcomes, particularly in sarcomere-negative cases.

The Gist

Imagine your heart is a high-performance engine. In most people, this engine is built with a standard blueprint. But in people with Hypertrophic Cardiomyopathy (HCM), the engine walls get too thick, like a muscle that has been over-exercised, which can make it harder for the heart to pump blood.

For a long time, doctors thought this condition was caused by a single, broken instruction in the "instruction manual" (a gene) that builds the heart's muscle fibers. They found these broken instructions in about half of the patients. But for the other half, the manual looked perfect, yet the engine was still faulty. Scientists wondered: If the main instructions are fine, what else is going wrong?

This paper is like a massive detective story where researchers scoured the entire "instruction manual" (the genome) of thousands of people to find the hidden clues.

The Big Search (The Study)

The researchers gathered a team of 2,284 people with the thickened heart condition and compared them to 4,525 healthy people. They looked at millions of tiny genetic "spelling mistakes" (variations) across the entire human genome to see which ones appeared more often in the sick group.

To make sure they didn't miss anything, they combined their findings with a previous, even larger study. It's like taking two different maps of the same territory and overlaying them to find the most accurate landmarks.

What They Found: New Landmarks

1. Three New Clues: In their own group, they found three specific spots in the DNA that were linked to the disease. One of these spots, called PPP1R3A, was a brand-new discovery. Think of this as finding a new room in a house that no one knew existed before.
2. The Big Reveal: When they combined their data with the larger study, they found 70 different spots in the DNA linked to HCM. Four of these were completely new to science: NOS1AP, OBSCN, MYPN, and YWHAE.
   • Analogy: Imagine trying to find the source of a leak in a giant dam. Before, you only knew about a few cracks. Now, you've mapped out 70 potential weak spots, some of which were previously invisible.

The "Genetic Scorecard" (Polygenic Risk Score)

Since many of these spots are small, weak clues rather than one giant "broken" gene, the researchers created a Polygenic Risk Score (PRS).

• The Analogy: Imagine you are grading a student. Instead of looking for one failed test (a single broken gene), you look at their entire report card. Maybe they got a few B's and C's in different subjects. Alone, a B isn't a failure, but if you add up all the grades, you can predict if they are likely to struggle in the future.
• The Result: They tested this "scorecard" on over 400,000 people in the UK Biobank (a giant database of health records).
  • People in the top 5% of this score were three times more likely to be diagnosed with HCM than those in the bottom 95%.
  • Even in people who didn't have the disease yet, a higher score meant their heart walls were slightly thicker and their hearts pumped slightly more forcefully. It's like the score predicts the engine is "revving" a bit too high before it actually breaks down.

The Twist: Who Does This Help?

The researchers split the patients into two groups:

1. Sarcomere-Positive: People with a known, major broken gene (the "single broken instruction").
2. Sarcomere-Negative: People with no known broken gene (the "mystery cases").

The Surprise: The "scorecard" didn't predict much for the first group (because their broken gene was already the main problem). However, for the Sarcomere-Negative group, the score was very important.

• People in this group who had a high score were 2.7 times more likely to suffer a sudden cardiac event (like a heart stoppage) compared to those with a low score.
• Analogy: If you have a broken engine part (Sarcomere-Positive), you know you're in trouble. But if your engine looks perfect but your "scorecard" says you have a lot of tiny, hidden stressors (Sarcomere-Negative), you are actually at high risk of a sudden breakdown.

What This Means (and What It Doesn't)

• What they claim: They found new genetic spots linked to the disease and created a score that can predict who is at risk and who might have a sudden heart event, especially for those without a known genetic cause.
• What they don't claim: They explicitly state this is a research paper and should not be used to guide clinical practice yet. They haven't proven that changing your lifestyle or taking a specific drug based on this score will fix the problem. They also note that their study was mostly on people of European ancestry, so the results might not apply to everyone yet.

The Bottom Line

This study is like finding a new set of tools to look at the heart's blueprint. It shows that for many people, the disease isn't just about one big broken gene, but a combination of many tiny genetic factors. By adding up these tiny factors, doctors might one day be able to spot who is at risk for a sudden heart event, even if they don't have the "classic" broken gene. But for now, it's a discovery, not a cure.

Technical Summary

Technical Summary: Hypertrophic Cardiomyopathy: A Genome-Wide Association Meta-Analysis and Polygenic Risk Score

Problem Statement
Hypertrophic cardiomyopathy (HCM) is a heritable trait characterized by left ventricular wall thickening. While often caused by pathogenic variants in cardiac sarcomere genes, more than half of all cases remain genetically elusive. Previous research has suggested a polygenic influence on the HCM phenotype, particularly in patients lacking dominant sarcomere gene variants. However, existing genome-wide association studies (GWAS) have identified loci with varying degrees of replication, and there is a need to discover novel genetic loci and develop robust polygenic risk scores (PRS) to better predict disease incidence, imaging phenotypes, and clinical outcomes, specifically stratified by sarcomere genotype status.

Methodology
The study employed a multi-stage approach combining a discovery GWAS, meta-analysis, bioinformatic annotation, and PRS validation:

1. Discovery GWAS: A case-control GWAS was performed on 2,284 European ancestry HCM cases (collected from St. Bartholomew's Hospital, UCLH, and Complexo Hospitalario Universitario A Coruña) and 4,525 controls from the EXCEED cohort. Cases were defined by maximal left ventricular wall thickness (≥15mm in probands, ≥13mm in relatives) and excluded phenocopies. Controls were free of known HCM and cardiovascular conditions. Genotyping was performed on different arrays (Illumina GSA vs. Affymetrix UK Biobank Axiom), necessitating strict quality control, harmonization, and joint imputation using the Haplotype Reference Consortium (HRC) panel. Association analysis was conducted using REGENIE.
2. Meta-Analysis: The discovery GWAS summary statistics were combined via fixed-effects inverse-variance weighted meta-analysis with two previously published datasets from Tadros et al.: (i) a single-trait HCM GWAS (5,900 cases, 68,359 controls) and (ii) a multi-trait analysis (MTAG) incorporating 36,083 UK Biobank participants with cardiovascular magnetic resonance (CMR) data.
3. Bioinformatic Analysis: Discovered loci underwent comprehensive annotation including:
   • Variant Annotation: Fine-mapping to 99% credible sets and functional prediction (CADD, RegulomeDB).
   • Colocalization: Analysis of cis-eQTL signals in heart and arterial tissues using GTEx v8 data.
   • Chromatin Interaction: Hi-C data analysis to identify distal candidate genes via promoter interactions.
   • Gene Prioritization: A custom scoring system integrating PoPS (Polygenic Priority Score), OMIM/ClinGen annotations, and tissue-specific enrichment.
   • Pathway Enrichment: DEPICT and IMPC analyses to identify enriched gene sets and mouse knockout phenotypes.
   • Druggability: Assessment of gene-drug interactions using Open Targets and DGIdb.
4. Polygenic Risk Score (PRS) Validation:
   • Incidence and Phenotype: A weighted PRS was constructed from unique lead variants and tested in 411,213 unrelated UK Biobank (UKBB) individuals for association with HCM diagnosis (using Cox proportional hazards) and imaging phenotypes (LVEF, wall thickness) in 56,945 individuals without HCM.
   • Outcomes in HCM: The PRS was evaluated in 1,756 HCM cases for association with a composite endpoint (all-cause mortality, transplant, sudden cardiac death [SCD]) and SCD alone.
   • Stratification: All PRS analyses were stratified by sarcomere genotype status (positive vs. negative).

Key Contributions

• Novel Loci Discovery: Identified three loci in the discovery GWAS (BAG3, FHOD3, and the novel PPP1R3A).
• Meta-Analysis Expansion: Combined datasets to identify 70 unique genome-wide significant loci, including four novel loci (NOS1AP, OBSCN, MYPN, and YWHAE).
• Gene Prioritization: Developed a scoring framework that prioritized NOS1AP as a high-confidence candidate gene, supported by colocalization with eQTLs in cardiac tissues and regulatory potential.
• PRS Development and Validation: Created a new PRS that demonstrated significant predictive power for HCM diagnosis and specific imaging traits in the general population.
• Stratified Risk Prediction: Provided evidence that the PRS predicts adverse outcomes (specifically SCD) in sarcomere-negative HCM cases, a group where risk stratification is currently challenging.

Results

• Genetic Findings: The discovery GWAS identified BAG3, FHOD3, and PPP1R3A (novel). The meta-analyses confirmed these and added NOS1AP, OBSCN, MYPN, and YWHAE as novel loci. PPP1R3A showed nominal significance in the meta-analysis but with significant heterogeneity. NOS1AP was highlighted as a candidate gene due to colocalization with expression in the atrial appendage and left ventricle.
• Functional Insights: Bioinformatic analyses linked loci to pathways such as the Hippo signaling pathway and TEAD1 subnetworks. Druggability analysis identified 20 genes as drug targets, 14 of which are targets for cardiovascular drugs.
• PRS Performance in UKBB:
  • Diagnosis: The PRS was significantly associated with HCM diagnosis (Hazard Ratio [HR] 1.88 per SD increase; HR 3.19 for top 5% vs. bottom 95%).
  • Imaging: In individuals without HCM, higher PRS correlated with greater left ventricular wall thickness (β = 0.13 mm per SD) and higher left ventricular ejection fraction (LVEF) (β = 0.71% per SD).
• PRS Performance in HCM Cohort:
  • Imaging: A positive association was found between PRS and LVEF in sarcomere-positive cases.
  • Outcomes: While no significant association was found per SD increase, sarcomere-negative HCM cases in the top 20% of the PRS distribution had a significantly increased risk of SCD or equivalent (HR 2.72, 95% CI: 1.03–7.17).

Significance and Claims
The authors claim that this study reports novel genetic loci associated with HCM and demonstrates that a new PRS can predict the risk of developing HCM, associated imaging characteristics in the general population, and adverse outcomes in HCM patients. Specifically, the paper highlights the utility of the PRS in stratifying risk for sudden cardiac death in sarcomere-negative patients, a subgroup that often lacks clear genetic markers for risk assessment. The authors note that while the PRS shows promise, the study has limitations, including potential sample overlap in the meta-analysis, modest sample sizes for outcome analysis, and a restriction to European ancestry, necessitating replication in independent and diverse cohorts. The findings support the hypothesis that common genetic variation contributes to HCM susceptibility and outcomes, particularly in the absence of rare sarcomere variants.