Abstract 4146208: Blood protein mediators of association between smoking and abdominal aortic aneurysm: exploration integrating human proteomic and genomic data

Background: Cigarette smoking has been identified as an important risk factor for the development of abdominal aortic aneurysm (AAA) and is known to influence levels of blood proteins. However, protein mediators of the smoking-AAA link remain unexplored. Here, we conducted a two-stage Mendelian randomization (MR) study to comprehensively pinpoint possible protein mediators by leveraging large-scale proteomic and genomic data. Methods: We first assessed the associations between genetically predicted smoking traits (lifetime smoking index, smoking initiation and cigarettes per day) and AAA risk (β total ). We then performed a protein-wide MR analysis to explore circulating proteins associated with smoking traits (β 1 ). Among identified proteins, we established their levels associated with AAA risk (β 2 ). According to directionality (β total vs. β 1 ×β 2 ), pathways were constructed, and corresponding mediation proportions were estimated by (β 1 × β 2 )/ β total ×100%. Genetic instruments for smoking traits were selected by filtering variants at P < 5×10 -8 and r 2 <0.001. Cis variants from encoding gene region ± 1Mb were used as instruments for protein. Summary level data on proteins and AAA were obtained from the deCODE study (N=35,559), Fenland study (N=10,708), and AAAgen consortium (37,214 cases), respectively. MR associations were estimated by the multiplicative random effects inverse variance method and summary-data-based MR method were applicable. We performed colocalization analysis for protein-AAA associations. Multiple testing was controlled with False Discovery Rate. Results: Genetically predicted smoking traits were associated with an increased genetic liability to AAA and genetically predicted levels of 48 proteins. Among 48 smoking-associated proteins, genetically predicted levels of 9 proteins were associated with genetic liability to AAA. Three protein pathways were established. Specifically, genetically predicted levels of WFDC2, MMP12, and SMPD1 mediated approximately 10.0 (1.4-18.6) %, 3.6 (1.7-5.5) %, and 2.2 (95% 0.7-3.7) % of the association between genetically predicted lifetime smoking index and genetic liability to AAA. Colocalization evidence was only observed for the SMPD1-AAA association (PH4=0.98). Conclusions: The study revealed a wide effect of cigarette smoking on blood proteins and suggested three potential pathways linking smoking and AAA, including MMP12 which has long been appreciated as important to AAA formation in model systems.