Abstract Body (Do not enter title and authors here): Background/Aims: Molecular mechanisms that potentiate atherosclerotic peripheral artery disease (PAD) outside of classical risk factors remain poorly understand. In this analysis, we prioritize proteomic mediators of PAD by integrating population-based plasma proteomic and genomic data.
Methods/Results: We performed a proteome-wide screen for circulating proteins associated with incident PAD using data from 44,140 individuals in the UK Biobank that have undergone antibody-based plasma proteomic profiling on Olink across 2,911 unique proteins. We investigated for causal protein-PAD effects using Mendelian Randomization (MR) with protein quantitative trait loci within 1Mb of the protein-coding gene (cis-pQTLs) as instruments. Of the 1,009 proteins with significant observational associations and at least one independent pQTL that matched PAD-associated variants, 66 were identified as putative candidate causal proteins in the primary MR analyses (range of 0.28-2.41 odds of PAD per SD increase in protein level). Of the 66 putative causal protein-disease effects, 40 proteins were robust to sensitivity analyses, including weighted median estimation and tests of directionality, heterogeneity, and pleiotropy. Cross-platform validation using cis-pQTLs detected on the aptamer-based SomaScan assay identified 7 proteins with consistent causal effects on PAD.
Single and multivariant colocalization analysis identified 3 proteins that had shared genetic signals between cis-pQTLs and PAD-associated loci in European individuals: MMP3 PP.H4 = 0.89, MMP12 PP.H4 = 1.0, and ENTPD6 PP.H4 = 0.99. To identify whether the causal effects of the proteins on PAD were generalizable across populations, we used pQTLs from African individuals in the Jackson Heart Study and European participants in the UK Biobank to calculate trans-ethnic genetic correlations and perform trans-ethnic MR. Of the 3 genetic signals with evidence of colocalization, 2 protein-disease associations demonstrated generalizable causal effects on PAD in African and European ancestry groups: MMP3 (European OR: 1.19, p=4.5x10^-5; African OR: 1.11, p=3.2x10^-36) and ENTPD6 (European OR: 1.08, p=6.9x10^-6; African OR: 1.16, p= 1x10^-20).
Conclusions: Higher genetically predicted levels of ENTPD6 and MMP3 showed evidence of a causal effect on PAD across European and African ancestries. Further investigation of the mechanisms by which these proteins potentiate disease could lead to actionable insights and targets to prevent PAD.