Extended analysis of a genome-wide association study in primary sclerosing cholangitis detects multiple novel risk loci

Abstract

Background & aims: A limited number of genetic risk factors have been reported in primary sclerosing cholangitis (PSC). To discover further genetic susceptibility factors for PSC, we followed up on a second tier of single nucleotide polymorphisms (SNPs) from a genome-wide association study (GWAS).

Methods: We analyzed 45 SNPs in 1221 PSC cases and 3508 controls. The association results from the replication analysis and the original GWAS (715 PSC cases and 2962 controls) were combined in a meta-analysis comprising 1936 PSC cases and 6470 controls. We performed an analysis of bile microbial community composition in 39 PSC patients by 16S rRNA sequencing.

Results: Seventeen SNPs representing 12 distinct genetic loci achieved nominal significance (p(replication) <0.05) in the replication. The most robust novel association was detected at chromosome 1p36 (rs3748816; p(combined)=2.1 × 10(-8)) where the MMEL1 and TNFRSF14 genes represent potential disease genes. Eight additional novel loci showed suggestive evidence of association (p(repl) <0.05). FUT2 at chromosome 19q13 (rs602662; p(comb)=1.9 × 10(-6), rs281377; p(comb)=2.1 × 10(-6) and rs601338; p(comb)=2.7 × 10(-6)) is notable due to its implication in altered susceptibility to infectious agents. We found that FUT2 secretor status and genotype defined by rs601338 significantly influence biliary microbial community composition in PSC patients.

Conclusions: We identify multiple new PSC risk loci by extended analysis of a PSC GWAS. FUT2 genotype needs to be taken into account when assessing the influence of microbiota on biliary pathology in PSC.

Figure 1. Regional association plots for MMEL1/TNFRSF14, CLEC16A and FUT2

The association results for both the genotyped and imputed SNPs are represented by the −log₁₀ P-value plotted against the genomic position. The index SNP is marked out with a purple diamond while the colors of the remaining SNPs indicate the linkage disequilibrium with the index SNP. The recombination rates were derived from the HapMap project and are represented by the thin blue lines. The plots were generated using the LocusZoom software [10].

Figure 2. Biliary FUT2 Phenotyping

Lectin staining of the hilar liver biopsy specimens from PSC patients. Paraffin-embedded sections from individuals with AA (non-secretor) and GG (secretor) variants of the FUT2 rs601338 SNP were used to evaluate the expression of α(1,2)fucosylated glycans in bile duct epithelium. H antigen, detected with the α(1,2)fucose-specific lectin Ulex europaeus agglutinin-I (UEA-I) (brown staining), is expressed on the apical surface of the biliary epithelial layer of secretor variant, while it is absent on non-secretor epithelia (Original magnification X400).

Figure 3. Influence of FUT2 genotype on phyla abundances and alpha diversity

(A) Mean abundances (± SE) of the major phyla with respect to genotype at the nonsense SNP rs601338 (W(TGG)→*(TAG). (B) The Abundance based Coverage Estimator (ACE) as a measure of the approximated species richness [35] (C) Alpha diversity measured as the sum of the total branch length in each sample (i.e. Phylogenetic Diversity) [36]. All values are based on the normalized dataset (2,000 reads per individual).