Background. Previous studies in high and low expressors has demonstrated that a variant in the GNPAT gene (D519G, Rs11558492, chromosome 1, exon 11) has been associated with severe iron overload in C282Y homozygotes for hemochromatosis. In this study, a GNPATvariant was assessed prospectively in patients referred for HFE testing over a range of serum ferritin levels.
Material and methods. Consecutive patients sent for HFE testing were studied for the GNPAT variant using a TaqMan kit assay (Life Technologies, Burlington, ON). Serum ferritin and iron removed by phlebotomy was compared in C282Y homozygotes with and without the GNPATvariant. The frequency of the GNPATvariant in referred patients was compared to a control population of voluntary blood donors without HFE mutations.
Results. There were 533 patients that had GNPATanalysis. The allele frequency for the GNPATvariant in C282Y homozygotes (n = 75) was 0.226 and in wild type control patients (n = 458) was 0.213 (p = .07). Forty-eight percent (of the C282Y homozygotes were heterozygous (n = 28) or homozygous (n = 8) for the GNPATvariant. The mean (log)ferritin and iron removed did not significantly differ between C282Y homozygous with GNPAT homozygotes, GNPAT heterozygotes, and without the GNPATvariant (p = 0.84).
Conclusions. C282Y homozygotes referred for HFE testing commonly have a GNPATvariant. This GNPATvariant does not appear be a co-modifying gene affecting expression of HFE related hemochromatosis in this population. The GNPATvariant does not predict the severity of iron overload.
Since a genetic test (HFE) has been available for the diagnosis of hemochromatosis since 1996, many studies have reported a wide range of clinical expression in C282Y homzoygotes. The cause of this variability has not been clearly established and environmental and genetic factors have been considered.1 The presence of co-modiftying genes remains a possibility and this has been studied using candidate gene analysis.2 genome wide association studies (GWAS),3 and exome sequencing.4 A recent study using exome sequencing and sequence kernel association testing suggested that a variant of the GNPAT gene may be associated with more severe iron overload.4 Gene silencing of the GNPAT gene led to a reduction in intracellular hepcidin. If this GNPAT mutation could predict severity of iron overload it could be a useful diagnostic test to predict follow up care and prognosis. In this study, GNPAT mutations were studied in a prospective population of male and female C282Y homozygotes with a broad spectrum of iron overload as assessed by serum ferritin and iron removed by phlebotomy.
Material and MethodsThis study was approved by the Human Ethics Committee of Western University.
Consecutive patients sent for HFE testing were studied for the GNPAT variant Serum ferritin at the time of presentation in untreated C282Y homozygotes was compared with and without the GNPAT variant. The frequency of the GNPAT variant in referred C282Y homozygotes was compared to a control population of voluntary blood donors without HFE mutations. Not all of the blood donors had serum ferritin available. Peripheral blood was taken from patients and extracted to an aqueous solution or salt precipitate. Prior to testing, salt precipitate samples were rehydrated in 300 µL TE Buffer (Teknova). Rehydrated and aqueous samples were tested for the HFE and GNPAT variant using a TaqMan genotyping assay (Life Technologies Burlington, ON), prepared using a clinically validated HFE run protocol and the Roche Lightcycler 480. Each run was analyzed using the endpoint genotyping setting of the Lightcycler 480 software (release 1.5.1.61). Sanger sequencing of GNPAT exon 11 was conducted on three samples, homozygous, heterozygous, and wildtype, that served as D519G standards throughout the study. Sequence was run on the 3730 DNA analyzer (Applied Bio-systems) using custom primers (Primer 3, SNPCheck) and the results were analyzed using Mutation surveyor (version 4.0.7). The allele frequency of GNPAT mutations in the C282Y homozygotes and a control population without HFE mutations were compared using the Fisher exact test. Ferritin was converted to Log (Ferritin) for comparative analysis because of the lack of a Gaussian distribution. Log (Serum ferritin) was compared across groups by ANOVA. Iron removed by phlebotomy in grams of iron was the number of 500 mL phlebotomies to bring the initial serum ferritin down to 50 µg/L x 0.25. The effects of a GNPAT mutation on log (ferritin) and iron removed (g) were studied in a multiple regression model with age, gender and C282Y homozygosity as the other variables.
ResultsThere were 533 patients that had GNPAT analysis. There were 75 C282Y homozygotes. Median age was 57 with a range from 25-77 years. The allele frequency for the GNPAT variant in C282Y homozygotes (n = 75) was 0.226 and in wild type control patients (n = 458) was 0.213 (p = 0.07, Fisher exact test). Forty eight percent of the C282Y homozygotes were heterozygous (n = 28) or homozygous (n = 8). The mean ferritin ± standard deviation did not significantly differ between C282Y homozygous who were GNPAT homozygotes (n = 8, 1450 μg/L ± 1529), GNPAT heterozygotes (n = 28, 1012 μg/L ± 1,626) and without the GNPAT variant (n = 39, 1,464 μg/L ± 1,169). Mean (log) serum ferritin did not differ between the 3 groups. Serum ferritin in GNPAT homozygotes, heterozygotes and wild type in 75 C282Y homozygotes is shown in figure 1. Not all patients tested for GNPAT had serum ferritin available. Iron removed by phlebotomy was similar between C282Y homozygotes with and without the GNPAT mutation (Figure 2) (p = 0.33). The multiple regression models demonstrated that male gender (p = 0.04) and C282Y homozygosity (p = 0.01) had significant effects on an elevation in log serum ferritin (Table 1). Male gender was related to iron removed (p = 0.05) There was no effect of the GNPAT mutation on serum ferritin or iron removed in the regression models. There were no significant differences in mean serum in a grouping of patients with a serum ferritin ≥ 1,000 µg/L and < 1,000 ≥ g/L (p = 0.49) (Table 3). The relationship between iron removed by phlebotomy and the presence of a GNPAT mutation is shown in figure 3.
Percentage distribution of GNPAT alleles in C282Y homozygotes and wildtype.
| N | TT(wild type) | CT(heterozygote) | CC (homozygote) | Allele Freq C | |
|---|---|---|---|---|---|
| C282Y homozygote | 75 | 39 (52%) | 28 (37%) | 8 (11%) | 0.226 |
| Wild type | 458 | 290 (63%) | 149 (33 %) | 19 (4%) | 0.213 |
There was no difference in GNPAT allele frequency between C282Y homozygotes and wild type patients (p = 0.07, Fisher exact test).
In this study, we have demonstrated that the D519G frequency did not differ significantly in a broad spectrum of C282Y homozygotes and a reference wild type population. This result agrees with the findings of Ryan, et al.,5 but differs from the findings of Besson, et al.6 and McLaren, et al.4 who found D519G enrichment in high expressing C282Y homozygotes.11 Our findings are also consistent with the observations of Bardou-Jacquet, et al., who re-evaluated the role of GNPAT in a large GWAS study of C282Y homozygotes.7 In a study of 83 healthy volunteers in Taiwan, an oral iron tolerance test showed an increase in serum iron and transferrin saturation in 19 participants with GNPAT variants. Fasting serum iron was also higher in subjects with GNPAT variants than in wild type subjects.8 The cause of these iron abnormalities is not clear but could be mediated by serum hepcidin since gene silencing of the GNPAT gene in cultured hepatocytes reduced intracellular hepcidin.4 Serum hepcidin could be of interest in these patients but hepcidin can fluctuate widely even in iron overloaded patients9 and the diagnostic role of serum hepcidin has not been clearly established.10
The patient selection differs in the current study from the original observations of McLaren, et al. In their study, only men were included and extreme expression (high and low) was assessed using body iron stores as assessed by liver iron concentration and/or quantitative phlebotomy. The advantages of not using serum ferritin as the marker of iron overload is that patients with false positive elevations in ferritin from alcohol use or inflammation would be excluded. We have been able to analyze a sub-group of patients with iron removed by phlebotomy and have not shown an effect of the GNPAT mutation. The study of low expressors by McLaren, et al.4 found no GNPAT mutations whereas our current study has demonstrated GNPAT mutations across the spectrum of serum ferritin and is consistent with other population studies of GNPAT allele frequency. There is a trend showing an increase in GNPAT mutation as iron removed increases but the presence of the mutation across all groups limits the value of GNPAT as a diagnostic test to predict severity of iron overload.
The role of GNPAT mutations in iron metabolism remains elusive8 but this study clearly demonstrates that the test will not be useful to predict severity of iron overload in C282Y homozygotes.
