Weekly Cardiology Research Analysis
This week’s cardiology literature emphasizes major advances in human genetics, developmental biology, and AI-driven mechanistic discovery. Large sequencing and GWAS efforts clarified rare and common variant architecture for atrial fibrillation and coronary anatomy (CXCL12), enabling new mechanistic targets and improved genomic prediction. Complementary translational work using explainable machine learning identified an actionable off-target pathway for escitalopram that attenuates cardiomyocyte
Summary
This week’s cardiology literature emphasizes major advances in human genetics, developmental biology, and AI-driven mechanistic discovery. Large sequencing and GWAS efforts clarified rare and common variant architecture for atrial fibrillation and coronary anatomy (CXCL12), enabling new mechanistic targets and improved genomic prediction. Complementary translational work using explainable machine learning identified an actionable off-target pathway for escitalopram that attenuates cardiomyocyte hypertrophy, highlighting repurposing opportunities.
Selected Articles
1. Sequencing in over 50,000 cases identifies coding and structural variation underlying atrial fibrillation risk.
Meta-analysis of genome and exome sequencing across >52,000 AF cases and 277,000+ controls identified novel rare coding genes (e.g., MYBPC3, LMNA, PKP2, FAM189A2, KDM5B) and structural variants (CTNNA3 deletions, GATA4 duplications) associated with AF. Findings were replicated in independent cohorts and CRISPR knockout of KDM5B in atrial cardiomyocytes produced electrophysiologic changes, linking rare variants to atrial physiology.
Impact: Defines the rare-variant landscape of AF with replication and functional CRISPR validation, strengthening causal links between rare coding/structural variants and atrial electrophysiology—critical for genetic testing, mechanistic research, and novel target discovery.
Clinical Implications: Supports expansion of AF genetic testing panels to include validated rare coding/structural variants; mechanistic leads (e.g., KDM5B) may inform novel antiarrhythmic targets and refine risk counseling for patients/families.
Key Findings
- Rare coding variants in MYBPC3, LMNA, PKP2, FAM189A2, and KDM5B associated with AF by burden testing.
- Rare structural variants (CTNNA3 deletions, GATA4 duplications) confer AF risk.
- CRISPR KDM5B knockout in atrial cardiomyocytes shortens action potential duration and dysregulates atrial homeostasis genes.
- Replicated across MyCode, deCODE, and UK Biobank cohorts.
2. CXCL12 drives natural variation in coronary artery anatomy across diverse populations.
A cross-ancestry GWAS of coronary dominance using >60,000 angiograms identified 10 loci with the strongest signal near CXCL12. CXCL12 is expressed in human fetal hearts at the time dominance is established and mouse Cxcl12 reduction alters coronary dominance and septal artery trajectories, linking a developmental chemokine pathway to coronary anatomy.
Impact: Establishes CXCL12 as a causal developmental regulator of coronary artery patterning using convergent human genetics, fetal expression data, and mouse perturbation—opening mechanistic insight with potential implications for revascularization planning and novel therapeutic concepts.
Clinical Implications: Understanding genetic determinants of coronary dominance may refine preprocedural planning and ischemia risk stratification; long-term, developmental pathway modulation could inform advanced revascularization concepts.
Key Findings
- GWAS of >60,000 angiograms identified 10 loci for coronary dominance with moderate heritability.
- Strongest association localized near CXCL12 across European- and African-ancestry cohorts.
- CXCL12 is expressed in human fetal hearts during dominance establishment.
- Reducing Cxcl12 in mice altered coronary dominance and septal artery development.
3. Logic-based machine learning predicts how escitalopram attenuates cardiomyocyte hypertrophy.
Authors developed LogiRx, an explainable logic-based machine learning method to predict drug-induced pathways and validated that escitalopram attenuates cardiomyocyte hypertrophy via an off-target serotonin receptor/PI3Kγ pathway. Findings were supported by in vitro neonatal cardiomyocytes, in vivo mouse models, and observational human database analyses, suggesting repurposing opportunities.
Impact: Demonstrates an explainable AI pipeline that yields experimentally testable mechanistic hypotheses and identifies a real-world repurposing candidate (escitalopram) with translational validation across models—accelerating mechanism-driven drug repurposing for cardiac remodeling.
Clinical Implications: If validated in prospective trials, escitalopram or agents targeting the serotonin receptor/PI3Kγ axis could be repurposed to limit pathological hypertrophy; also exemplifies how explainable AI can prioritize translational candidates.
Key Findings
- LogiRx, a logic-based mechanistic ML framework, predicts drug-induced pathways.
- Escitalopram attenuates cardiomyocyte hypertrophy via an off-target serotonin receptor/PI3Kγ pathway predicted by LogiRx.
- Validation in neonatal cardiomyocytes, mouse hypertrophy/fibrosis models, and human database analyses supports translational relevance.