Liver and adipose expression associated SNPs are enriched for association to type 2 diabetes

Abstract

Genome-wide association studies (GWAS) have demonstrated the ability to identify the strongest causal common variants in complex human diseases. However, to date, the massive data generated from GWAS have not been maximally explored to identify true associations that fail to meet the stringent level of association required to achieve genome-wide significance. Genetics of gene expression (GGE) studies have shown promise towards identifying DNA variations associated with disease and providing a path to functionally characterize findings from GWAS. Here, we present the first empiric study to systematically characterize the set of single nucleotide polymorphisms associated with expression (eSNPs) in liver, subcutaneous fat, and omental fat tissues, demonstrating these eSNPs are significantly more enriched for SNPs that associate with type 2 diabetes (T2D) in three large-scale GWAS than a matched set of randomly selected SNPs. This enrichment for T2D association increases as we restrict to eSNPs that correspond to genes comprising gene networks constructed from adipose gene expression data isolated from a mouse population segregating a T2D phenotype. Finally, by restricting to eSNPs corresponding to genes comprising an adipose subnetwork strongly predicted as causal for T2D, we dramatically increased the enrichment for SNPs associated with T2D and were able to identify a functionally related set of diabetes susceptibility genes. We identified and validated malic enzyme 1 (Me1) as a key regulator of this T2D subnetwork in mouse and provided support for the association of this gene to T2D in humans. This integration of eSNPs and networks provides a novel approach to identify disease susceptibility networks rather than the single SNPs or genes traditionally identified through GWAS, thereby extracting additional value from the wealth of data currently being generated by GWAS.

Figure 1. Diagram depicting the process of filtering SNPs using eSNPs and disease associated networks.

Figure 2. eSNP sets enriched for T2D associated SNPs in three GWAS.

The Y axis shows the proportion of SNPs with P_T2D< = 0.05. The P_T2D are from DIAGRAM, WTCCC, and DGI from left to right. In each GWAS cohort, from left to right, the 1^st bar shows the observed proportion of all studied SNPs; the 2^nd bar shows the proportion of all eSNPs, the 3^rd bar shows the proportion of the expanded eSNPs; the 4^th bar shows the proportion of the expanded adipose network eSNPs; and the 5^th bar shows the proportion of the expanded T2D adipose causal subnetwork eSNPs. In DIAGRAM study, the second bar is significantly higher than the first (p = 2.05×10⁻⁹), the third bar is higher than the second (p = 1.33×10⁻⁹), the fourth is higher than the third (p<10⁻¹⁶), and the fifth is higher than the fourth (p = 2.97×10⁻⁴). In WTCCC, the second bar is higher than the first (p = 1.09×10⁻²), the third is higher than the second (p = 1.22×10⁻⁶), but the fourth and fifth bars are not significantly higher than the third (p = 0.35) or fourth (p = 0.15), respectively. In the DGI study, the second bar is higher than the first (p = 8.0×10⁻⁴), the third is higher than the second, (p = 2.19×10⁻⁷), the fourth is higher than the third (p = 1.10×10⁻¹²), and the fifth is higher than the fourth (p<10⁻¹⁶).

Figure 3. Adipose T2D causal subnetwork and human supporting evidence.

A) The adipose coexpression network is comprised of 9,900 gene expression traits. The purple subnetwork comprised of 159 genes is highlighted as the subnetwork most enriched for genes supported as causal for T2D. B) LOD score plots for plasma insulin (solid black), Me1 adipose expression (solid red), Anxa2 adipose expression (solid blue), Bcl2l10 adipose expression (solid cyan), 2310046806Rik adipose expression (solid black), Car12 adipose expression (solid green), Paqr9 adipose expression (solid magenta), C730029A08Rik adipose expression (dashed red), Poclce2 adipose expression (dashed blue), and adipose expression traits linking to this region in trans (grey), all measured in the B6×BTBR cross. C) The T2D adipose causal subnetwork is enriched for genes supported as having a causal relationship with plasma insulin levels in the B6×BTBR cross (blue nodes). The white nodes represent genes in the T2D adipose causal subnetwork not supported as causal for insulin traits in the B6×BTBR cross. D) OGTT curves for Me1−/− (Male n = 19; Female n = 14) and wild-type control (Male n = 25; Female n = 16) mice (p = 3.16×10⁻⁴ for male OGTT AUC and p = 1.84×10⁻³ for female OGTT AUC; overall sex adjusted difference p = 7.30×10⁻⁸). E) The T2D adipose causal subnetwork is enriched for genes in the Me1−/− adipose gene expression signature (orange nodes). The white nodes represent genes in the purple subnetwork not in the Me1−/− adipose gene signature.