Introduction: Although it is clear that there is a poor correlation between hypoxemia, severity of airflow obstruction, and pulmonary artery (PA) pressures, alternative mechanisms underpinning pulmonary vascular disease in COPD have not yet been well-elucidated. PA enlargement (PAE), measured by CT as the PA/A ratio of the diameter of the PA to that of the aorta (A) > 1, correlates with pulmonary hypertension gauged by right heart catheterization and is associated with worse clinical outcomes in subjects with COPD. Genome-wide association studies previously identified multiple associations with PAE, but these studies lacked comprehensive coverage of genetic variants and the biological functions of the associated variants are unknown. To identify novel genetic determinants for PAE and characterize the functions of PAE-associated variants, we performed a whole-genome sequencing (WGS) analysis and integrated the results with publicly available epigenomic data.

Methods: A total of 5,131 subjects with available WGS and PA/A data in the COPDGene non-Hispanic Whites, COPDGene African Americans, and ECLIPSE studies were analyzed. PA/A ratio (> or ≤ 1) was tested for genetic associations using variants with MAF>0.01% in the TOPMed Freeze8 WGS data. Analyses were conducted on the ENCORE cloud computing platform using the Scalable and Accurate Implementation of Generalized mixed model (SAIGE) approach and adjusting for age, sex, pack-years of smoking, and genetic ancestry. Separate analyses in each study population were followed by a fixed-effect meta-analysis. We performed single-variant, VEGAS (Versatile Gene-based Association Study), and ingenuity pathway analyses. We then quantified the enrichment of PAE-WGS regions in DNaseI peaks from ENCODE and Roadmap cell types using the GARFIELD (GWAS analysis of regulatory of functional information enrichment with LD correction) program.

Results: We identified two loci associated with PAE at genome-wide significance: One previously reported (15q31 near IREB2) and one new (12p12 near LMO3) risk loci. Additional novel loci approaching genome-wide significance included regions near FREM2, CHEK2, and ZNF516. Top association results appeared to be enriched for several cell types including BMP4-derived mesendoderm cells and fibroblasts. Gene-based and pathway analyses highlighted potentially relevant biological mechanisms, including immune response and cytoskeleton remodeling (FDR 10%).

Conclusion: This study leveraged imaging phenotyping, genomic and epigenomic resources, and advanced bioinformatics tools and provided insights into new putative genetic factors, biological pathways, and cell types that may influence pulmonary vascular remodeling in COPD. Additional analyses are needed to replicate these findings and understand their consequences on downstream processes.