The Association Between H63D Mutations in HFE and Amyotrophic Lateral Sclerosis in a Dutch Population FREE

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease that causes progressive muscle weakness.^{1 - 2} Amyotrophic lateral sclerosis occurs in families, but most cases are sporadic (SALS). The 2 types are clinically and histologically indistinguishable.³ The pathogenesis of SALS is unknown, but convincing evidence has shown that several distinct molecular mechanisms are involved.^{4 - 5} There is a growing body of evidence that genes affect susceptibility for and clinical phenotype of SALS.

In patients with ALS, elevated iron levels were demonstrated in the central nervous system, which could imply a change in iron exposure of motor neurons.⁶ Iron overload is present in hereditary hemochromatosis (HH), an autosomal recessive disorder that is predominantly caused by a homozygous state for the C282Y mutation in the HFE gene in populations with European ancestry.⁷ In C282Y heterozygotes, HH can also develop if another mutation, H63D, is present. Individuals who are heterozygous for C282Y or homozygous for H63D are not affected with HH but often have subclinical elevated iron levels.^{8 - 10} They may also have an increased risk of developing neurodegenerative disorders.^{11 - 13}

Patients with H63D mutations in HFE demonstrated an increased risk of developing SALS in several populations.^{14 - 16} Most of these studies^{17 - 18} have, however, used either small groups or inappropriate control subjects. Increasing the sample size and examining multiple populations can aid in estimating population-wide effects of genetic risk factors.

To determine whether the association between HFEH63D and C282Y mutations and risk of ALS is present throughout populations, we investigated a large Dutch population for HFE mutations and pooled these results with data from previous studies. Because ALS is a heterogeneous disease, we also studied the effect on clinical phenotype (age at onset, bulbar or spinal onset, and survival) to determine whether a particular subset of ALS is associated with HFE mutations.

Between January 1, 2000, and December 31, 2004, 289 patients newly diagnosed as having SALS at the University Medical Center Utrecht, a tertiary referral clinic in the Netherlands, were recruited. Diagnosis was made according to the El Escorial criteria after exclusion of other conditions. Patients with possible, probable, or definite SALS were included. All patients were white. Demographic features, age at onset, site at onset of disease, and survival were recorded. Onset of disease was defined as onset of first weakness, dysarthria, or dysphagia. Survival, as a measure of rate of progression, was defined as the interval between age at onset and age at death from any cause, tracheostomy, or persistent (24 hours a day) ventilatory assistance. The study protocol was approved by the institutional ethical committee of the University Medical Center Utrecht.

Controls were included from 2 prospective studies in the Netherlands described elsewhere.^{19 - 20} Briefly, from the Rotterdam Cohort Study, a population-based study of 7983 individuals 55 years and older, a random sample of 4275 individuals was genotyped for HFE mutations. The other sample was taken from a Dutch contribution to the European Prospective Investigation into Cancer and Nutrition. From this cohort of 17 357 women who attended the regional population-based breast cancer screening program, 1611 women were randomly genotyped for HFE mutations.

After extraction of genomic DNA from whole blood of patients with ALS, mutation analyses for the C282Y and H63D alleles were performed by an allelic discrimination assay (TaqMan) on an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, Calif). Genotyping results were available for 289 patients and 5748 controls for the C282Y allele and 288 patients and 5777 controls for the H63D allele.

An association between HFE mutations and risk of developing ALS was evaluated with logistic regression analysis, adjusting for the potential confounders age (at onset) and sex. To determine whether HFE mutations are associated with the clinical phenotype, their effect was studied by multivariate regression, adjusting for possible confounders. The influence of HFE mutations on clinical phenotype was analyzed by (1) a linear regression model with age at onset as the outcome variable, adjusting for sex and site at onset; (2) a Cox regression model with survival as the outcome variable, adjusting for age at onset, sex, and site at onset; and (3) a logistic regression model with site at onset as the outcome variable, adjusting for age at onset and sex. Analyses were performed for the C282Y and H63D genotypes combined. The wild-type genotype was used as the reference value. Because 2 loci (C282Y and H63D) were studied, we considered a more conservative P<.025 as statistically significant. A Mantel-Haenszel common odds ratio (OR) estimate was computed to pool the association between HFE mutations and risk of developing ALS with the associations described in 3 previous studies.^{14 - 16}

Table 1 gives the characteristics of the patients and controls. The control population had a higher median age and consisted of more women. We corrected for these confounders in all our analyses.

The frequencies for both the C282Y and H63D alleles were in Hardy-Weinberg equilibrium in the control population. Table 2 summarizes and compares the genotype distributions of patients and controls. Homozygous mutations at H63D were independently associated with an increased risk of developing ALS (OR, 2.2; 95% confidence interval [CI], 1.1-4.1; P = .02). Other genotypes were not significantly different between patients and controls. Comparing genotypes of the patient group with each control group separately provided similar results (data not shown). When our data were pooled with data from all previous HFE association studies performed in various geographical regions, an increased risk was observed for H63D mutation carriers (OR, 1.7; 95% CI, 1.1-2.5), homozygotes (OR, 2.7; 95% CI, 1.7-4.4), and heterozygotes (OR, 1.5; 95% CI, 1.0-2.1) (Figure).

We also examined the extent to which a mutation in HFE influences clinical phenotype. Table 3 gives the age at onset and survival of patients with ALS together with HFE genotypes. Heterozygosity at H63D was associated with a higher age at onset (mean difference, 5.4 years; 95% CI, 2.2-8.5; P = .001). In contrast, in the control group, H63D homozygotes, heterozygotes, and mutation carriers were similar in age to wild types (data not shown). Presence of a C282Y or H63D mutation did not affect survival (Table 3) or site at onset (data not shown).

In this study of 289 patients and 5886 controls, we detected a positive association between homozygosity for the H63D mutation and ALS, suggesting that HFE is a contributing factor in the development of ALS in the Dutch population. Moreover, we found heterozygosity for the H63DHFE mutation to be associated with a higher age at onset, possibly indicating that H63D is a risk factor for a later-onset form of ALS.

Our large control group was taken from prospective population-based studies^{19 - 20} that reflect the general Dutch population and made genotyping of a new control sample redundant. The control group differed from the patient population with regard to age and sex, but we adjusted for these confounders in our analyses. Moreover, no significant differences in HFE mutation frequencies have been reported for different age and sex groups.^{21 - 22} Furthermore, all patients were white, and observed genotype frequencies in the control population were similar to those reported for non-Hispanic white individuals in previous population-based studies and in Hardy-Weinberg equilibrium.^{8 ,21} In addition, comparison of genotypes of the patient group with each control group separately gave similar results.

Our findings agree with those of a previous study¹⁴ of 121 patients and 133 controls, which demonstrated an increased risk of developing ALS when an H63D mutation was present. This association was significant for H63D heterozygotes. A more recent study,¹⁶ which included 379 patients and 400 controls, showed an increased risk of developing ALS for H63D homozygotes and heterozygotes in 2 populations. In a smaller population of 51 patients and 47 controls, no difference was found in the presence of HFE mutations between ALS patients and controls.¹⁵ We pooled these results and showed an association for H63D homozygotes, heterozygotes, and carriers, supporting a genetic association.

Recommendations for performing genetic association studies have been published previously. By increasing the sample size, pooling data of individual studies in a meta-analysis aids in estimating population-wide effects of genetic associations.^{17 - 18} Moreover, a single significant association should be independently replicated, preferably at least twice. Therefore, the present study adds insight to conclusions from previous studies.

In our study, only H63D homozygotes demonstrated significance, whereas previous studies^{14 ,16} also showed an association with H63D heterozygotes. A difference in genetic background in the Dutch population could account for the somewhat weaker association with H63D (in heterozygotes and carriers) found in our study. Nevertheless, our meta-analysis clearly shows an association between ALS and H63D homozygotes and heterozygotes.

Several possible mechanisms could explain the observed relationship between H63D mutations and the development of ALS. Increased oxidative stress caused by excessive iron could play a role. However, the C282Y mutation, rather than H63D, is shown to have a greater effect on iron concentrations in serum and deposition in liver.⁸ An overall increase in iron supplies, therefore, is not a plausible biological mechanism in ALS. In addition, no indications were found of relevant neurological involvement in HFE-linked HH.⁷ Additional roles of HFE in other tissues still require elucidation, and H63D mutations could lead to unique conformational changes in the HFE protein that exert an effect mainly on local iron concentration at the motor neuron level. In particular, it has been proposed that H63D mutations predominantly affect the binding of HFE to the transferrin receptor, which plays a role in neuronal iron uptake.^{12 - 13 ,23} Studies in patients with Alzheimer disease support a role for the transferrin receptor in neurodegeneration. Alternatively, H63D is in linkage disequilibrium with other genetic variants that may initiate pathological cellular processes.

In conclusion, the findings suggest a role for HFE mutations in the development of ALS, although caution should be used in estimating the size of the effect. Further independent HFE genotype association studies are needed in different geographical regions. Moreover, serum iron values could provide further clues about the possible role for disorders of iron metabolism in patients with ALS.

Correspondence: L. H. van den Berg, MD, PhD, Department of Neurology, Room G03.228, University Medical Center Utrecht, PO Box 85500, 3508 GA Utrecht, the Netherlands (l.h.vandenberg@umcutrecht.nl).

Accepted for Publication: August 2, 2006.

Author Contributions:Study concept and design: Sutedja, Wokke, and Van den Berg. Acquisition of data: Sutedja, Sinke, Van Vught, Van Duijn, Njajou, Van der Schouw, and Van den Berg. Analysis and interpretation of data: Sutedja, Sinke, Van der Linden, Van der Schouw, Veldink, and Van den Berg. Drafting of the manuscript: Sutedja, Van der Linden, Veldink, and Van den Berg. Critical revision of the manuscript for important intellectual content: Sinke, Van Vught, Van der Linden, Wokke, Van Duijn, Njajou, Van der Schouw, and Veldink. Statistical analysis: Sutedja, Van der Linden, Njajou, Van der Schouw, and Veldink. Obtained funding: Van den Berg. Administrative, technical, and material support: Sutedja, Sinke, and Van Vught. Study supervision: Sinke, Wokke, Van Duijn, Veldink, and Van den Berg.

Financial Disclosure: None reported.

Funding/Support: This study was supported by a grant from the Netherlands Organization for Health Research and Development.