Systems genetics analyses of human adipose tissue gene expression identify cis and trans regulatory networks for cardio-metabolic traits. M. Civelek1, Y. Wu3, C. Pan2, A. He4, C. Tilford4, N. K. Saleem5, C. Fuchsberger6, A. Locke6, H. M. Stringham6, A. U. Jackson6, N. Narisu7, P. S. Chines7, Y. Zhao8, P. S. Gargalovic4, J. Kuusisto5, P. Pajukanta2, K. Hao9, X. Yang8, T. G. Kirchgessner4, F. S. Collins7, M. Boehnke6, M. Laakso5, K. L. Mohlke3, A. J. Lusis1,2 1) Department of Medicine, University of California, Los Angeles, CA; 2) Department of Human Genetics, University of California, Los Angeles, CA; 3) Department of Genetics, University of North Carolina, Chapel Hill, NC; 4) Bristol-Myers Squibb, Pennington, NJ; 5) Department of Medicine, University of Eastern Finland and Kuopio University Hospital, Kuopio, Finland; 6) Department of Biostatistics and Center for Statistical Genetics, University of Michigan, Ann Arbor, MI; 7) National Institutes of Health, Bethesda, MD; 8) Department of Integrative Biology and Physiology, University of California, Los Angeles, CA; 9) Department of Genetics and Genomic Sciences, Mount Sinai School of Medicine, New York, NY.
Genome-wide association studies (GWAS) have identified numerous loci that are associated with complex traits related to Metabolic Syndrome (MetSyn). However, the molecular mechanisms by which these loci affect MetSyn are usually not known. We used transcript abundance as an intermediate trait to map expression quantitative trait loci (eQTL). We tested the association of 621,695 common single nucleotide polymorphisms (SNPs) and the expression of 18,279 genes in subcutaneous adipose tissue of 1,381 Finnish males who are part of the METabolic Syndrome In Men (METSIM) study. We identified 7,941 genes for which local SNPs (<1 Mb) showed significant association at 1% FDR (local eQTLs). We focused on SNPs that are associated with 170 cardio-metabolic traits and diseases in GWAS. Of the 1,542 SNPs, 621 had been directly genotyped, representing 121 independent loci (r2<0.3). Using conditional analysis, we determined that 49 of the 121 loci were also local eQTLs for 54 genes. For example, rs8077889 has been associated with triglyceride levels in GWAS and is associated with the expression of two nearby genes, DUSP3 (P=1.5x10-12) and MPP3 (P=3.0x10-26). Only DUSP3 is significantly correlated with triglyceride levels (r=0.11, P=2.3x10-6) whereas MPP3 is not (r=-0.06, P=2.7x10-2), prioritizing DUSP3 as the causal gene. The large number of subjects profiled in METSIM also allowed the study of distant acting loci (>1Mb). KLF14 locus has previously been shown to regulate 10 adipose genes. We replicated these associations and identified an additional 20 genes that are trans regulated by this locus (P=5x10-8). These genes are C20orf194, CIB2, COX20, DYNLT1, EIF4E3, GALNT11, MAGED2, MSRA, NEO1, NGRN, PDCL3, PLIN5, SIRT3, SYNC, TECR, TUBB, UBE2Q2, UNC13B, ZNF219, ZNF226. Twelve of the genes, for which an expression probe was available, replicated in another eQTL study in omental and subcutaneous adipose tissue in 556 and 741 people (P=7.8x10-3-1.8x10-9). This locus was not a trans regulator of gene expression in liver (N=566 people), peripheral blood (N=5,311), monocytes and macrophages (N=745) suggesting an adipose-specific mechanism of regulation. All of the additional 20 genes had transcript levels significantly correlated with metabolic traits (|r|=0.33-0.11; P=6.6x10-37-9x10-4). Our studies extend the mechanistic understanding of the KLF14 locus and highlight the power of systems genetics approaches for dissecting complex traits to identify causal genes and pathways.
You may contact the first author (during and after the meeting) at