Unraveling genotype-phenotype relationships in hereditary hemochromatosis through integrated biobank data analysis.

BACKGROUND

Type I hereditary hemochromatosis (HH), caused by pathogenic HFE variants, is among the most common autosomal recessive disorders in Northern Europe. HH genotype–phenotype associations have been difficult to predict due to variable variant penetrance and expressivity. In this study, population-based biobank data were used to conduct a large-scale analysis of symptoms associated with different HH genotypes and their potential genetic modifiers.

METHODS

Linked genotypic and electronic health records from the Estonian Biobank (n = 211,994) and UK Biobank (n = 405,931) were used to investigate HH genotype–phenotype associations. Pathogenic variants of all HH types in the Estonian Biobank sample were identified. Clinical markers derived from laboratory measurements and diagnosis codes were compared between HFE pathogenic variant carriers and controls [chi-squared test, phenome-wide association study (PheWAS)]. Carrier subgroups were defined according to the presence of p.C282Y, p.H63D, p.S65C, and all pairwise combination genotypes. Kruskal–Wallis and Mann–Whitney U tests were used to identify ICD-10 codes and genotype groups with significantly different mean ages at first diagnosis. Genome-wide association studies (GWAS) based on ceruloplasmin and ferritin concentrations were conducted to identify potential genetic modifiers. Fine-mapping was used to identify putatively causal single nucleotide polymorphisms (credible sets) in the Estonian Biobank GWAS results. Fixed-effects meta-analyses of combined GWAS results from both biobanks were performed.

RESULTS

We conducted the largest PheWAS with data from HFE variant carriers to date and catalogued pathogenic HH variants in the Estonian population. p.S65C homozygotes and p.C282Y/p.S65C heterozygotes in the Estonian Biobank sample were diagnosed with urogenital conditions (ICD-10 codes N42 and N50.8) at significantly higher rates than were controls. On average, N42 was first diagnosed 11 years earlier in p.S65C compound heterozygotes and homozygotes than in controls. We also identified a novel GWAS hit, CP rs61733458, with a significant impact on the ceruloplasmin level and ties to iron metabolism–related diseases.

CONCLUSIONS

Our results suggest that the p.S65C variant is more clinically relevant than previously thought and should be integrated more into HH testing guidelines. CP variant rs61733458 is a potential genetic modifier associated with iron metabolism–related diseases.