Megan Leask, James Boocock, Hyon Choi, Tayaza Fadason, Justin O’Sullivan, Lisa Stamp, Nicola Dalbeth, Julia Horsfield, Tony R Merriman
Affiliation(s):
1. University Of Otago, Dunedin, NZ.
2. University Of California, Los Angeles, La, California, USA
3. Harvard Medical School And Massachusetts General Hospital, Boston, Ma USA
4. University Of Auckland, Auckland, NZ
5. University Of Otago, Christchurch, NZ
6. University Of Alabama, Birmignham, Birmingham, Alabama
At the strongly urate-associated locus SLC2A9 there are additional signals for gout and height in Europeans, suggesting that these traits could share common biological pathways. These association signals are complex, with multiple independent and overlapping signals for each of the traits. Additionally, the associated variants are non-coding, and therefore the causal mechanisms are unclear. Here we identify population-specific variation at SLC2A9 that associates with multiple traits and functions via regulation of gene expression.
Non-coding regulatory regions at SLC2A9 were resequenced in 447 individuals of Polynesian ancestry with hyperuricaemia or gout to identify Polynesian-specific variants that might provide insights into molecular control of urate and other traits at SLC2A9. Polynesian-specific variants that associated with gout in the Resequencing cohort were genotyped using Sequenom MassARRAY technology over a second cohort of gout cases and controls to identify associations with urate, gout, kidney function and height. Using CODES3D, which leverages HiC and gene expression (eQTL) data (GTEx v8), we linked the European trait-associated non-coding variants to protein-coding loci and gene expression. Non-coding regions harbouring Polynesian-specific variants were tested for regulatory function using our functional pipeline, including cell-based luciferase assays and zebrafish enhancer assays.
Three West Polynesian-specific non-coding variants in complete linkage disequilibrium that individually mark three regulatory regions (RR1, RR2 and RR3) associated with gout (p < 0.05) in the Resequencing cohort. One of these variants was genotyped in the larger replication cohort of Māori and Pacific people (n=1565). We found it lowers the risk of gout 5-fold (OR = 0.20, p = 1.9 x 10-5) and has a very large effect on height (β = -4.4 cm, p = 1.01 x 10-6) but not on serum urate (β = -0.013, p = 0.282). The European GWAS signal for height and second maximal GWAS signal for gout (gout vs. HU controls) at SLC2A9 overlap. We identified putative causal genes SLC2A9, ZNF518B and WDR1 in cis and TENS3 in trans for which the lead European height/gout variant is a spatially supported eQTL. RR1, RR2 and RR3 overlap multiple regulatory factors (DHS, H3K27ac and transcription factor binding sites) and spatially connect to SLC2A9, ZNF518B and WDR1. Finally, we tested one of these regulatory regions (RR2) harbouring a West Polynesian-specific gout- and height-associated variant, and show that it is a very strong enhancer that drives GFP expression in the brain in the developing zebrafish. Our analyses highlight the pleiotropic nature of the associations at the SLC2A9 locus. Outside of the SLC2A9 gene per se, but within the urate/gout associated region, there are Polynesian-specific variants that control height and risk of gout but not urate levels.