Common non-coding SNPs near the Hepatocyte Nuclear Factor-4 Alpha gene are associated with type 2 diabetes

K.L. Mohlke,1 K. Silander,1 E.C. Peck,1 L.J. Scott,2 P.E. Hollstein,1 A.D. Skol,2 C. Li,2 P.S. Chines,1 T.T. Valle,3 R.N. Bergman,4 J. Tuomilehto,3 R.M. Watanabe,4 M. Boehnke,2 F.S. Collins1

1) NHGRI, Bethesda, MD; 2) U. Michigan, Ann Arbor, MI; 3) National Public Health Institute, Helsinki, Finland; 4) U. Southern California, Los Angeles, CA

The Finland-United States Investigation of NIDDM genetics (FUSION) study aims to positionally clone variants that predispose to type 2 diabetes (T2D). A genome scan using 495 Finnish affected sibling pair families yielded a LOD score of 2.32 at 69 cM on chromosome 20q12-13.3; linkage in this region has been observed in 7 other populations. To fine-map the 13 cM (7.4 Mb) 1-LOD support interval, we are genotyping SNPs on case and control DNA pools. Among the first 291 SNPs, we selected 32 to confirm by individually genotyping 793 T2D cases (one from each family) and 413 controls. The most strongly associated SNP has a frequency of 0.201 in cases and 0.156 in controls (OR 1.36, p-value .007) and is located 1 kb downstream of an alternative transcription site P2 for the Hepatocyte Nuclear Factor-4 Alpha (HNF4A) gene. We genotyped 39 additional nearby SNPs, including 19 selected from a haplotype map (kindly provided by P. Deloukas, Sanger Institute). The strongest association is observed with the original SNP and one located 5.3 kb away that is in complete linkage disequilibrium (LD). These SNPs are located in a block of LD that extends >143 kb upstream of HNF4A. We observed greater linkage evidence in families where the genotyped sibling has at least one risk allele compared to families where the genotyped sibling has no risk alleles (MLS=1.93 vs 0.17). Mutations in HNF4A are known to cause maturity-onset diabetes of the young (MODY), a monogenic dominantly-inherited form of diabetes, but no previously identified variants within HNF4A have been associated with T2D. Significantly, a SNP in complete LD with the 2 reported here was detected independently and found to be associated with T2D (p-value .003) in Ashkenazim (L. Love-Gregory and M.A. Permutt, personal communication). These data suggest that defects in unknown HNF4A regulatory elements near the P2 promoter may increase susceptibility to T2D.