Integrative pathway genomics of lung function and airflow obstruction

Sina A Gharib, Daan W Loth, María Soler Artigas, Timothy P Birkland, Jemma B Wilk, Louise V Wain, Jennifer A Brody, Ma'en Obeidat, Dana B Hancock, Wenbo Tang, Rajesh Rawal, H Marike Boezen, Medea Imboden, Jennifer E Huffman, Lies Lahousse, Alexessander C Alves, Ani Manichaikul, Jennie Hui, Alanna C Morrison, Adaikalavan Ramasamy, Albert Vernon Smith, Vilmundur Gudnason, Ida Surakka, Veronique Vitart, David M Evans, David P Strachan, Ian J Deary, Albert Hofman, Sven Gläser, James F Wilson, Kari E North, Jing Hua Zhao, Susan R Heckbert, Deborah L Jarvis, Nicole Probst-Hensch, Holger Schulz, R Graham Barr, Marjo-Riitta Jarvelin, George T O'Connor, Mika Kähönen, Patricia A Cassano, Pirro G Hysi, Josée Dupuis, Caroline Hayward, Bruce M Psaty, Ian P Hall, William C Parks, Martin D Tobin, Stephanie J London

Chronic respiratory disorders are important contributors to the global burden of disease. Genome-wide association studies (GWASs) of lung function measures have identified several trait-associated loci, but explain only a modest portion of the phenotypic variability. We postulated that integrating pathway-based methods with GWASs of pulmonary function and airflow obstruction would identify a broader repertoire of genes and processes influencing these traits. We performed two independent GWASs of lung function and applied gene set enrichment analysis to one of the studies and validated the results using the second GWAS. We identified 131 significantly enriched gene sets associated with lung function and clustered them into larger biological modules involved in diverse processes including development, immunity, cell signaling, proliferation, and arachidonic acid. We found that enrichment of gene sets was not driven by GWAS-significant variants or loci, but instead by those with less stringent association P-values. Next, we applied pathway enrichment analysis to a meta-analyzed GWAS of airflow obstruction. We identified several biologic modules that functionally overlapped with those associated with pulmonary function. However, differences were also noted, including enrichment of extracellular matrix (ECM) processes specifically in the airflow obstruction study. Network analysis of the ECM module implicated a candidate gene, matrix metalloproteinase 10 (MMP10), as a putative disease target. We used a knock-out mouse model to functionally validate MMP10's role in influencing lung's susceptibility to cigarette smoke-induced emphysema. By integrating pathway analysis with population-based genomics, we unraveled biologic processes underlying pulmonary function traits and identified a candidate gene for obstructive lung disease.