Introduction

The iron storage polypeptide apoferritin is a 24-subunit heteropolymer composed of heavy chain (H-ferritin; 21 kDa) and light chain (L-ferritin; 19 kDa) peptide subunits that are encoded by genes on chromosomes 11q13 and 19q13. 1, respectively.1 Expression of both the L- and H-ferritin genes (FTL and FTH) is ubiquitous and is regulated at transcriptional level by a series of proinflammatory cytokines and growth factors.2 Expression of FTL and FTH is also regulated by the availability of intracellular iron through a coordinated mechanism that acts at the level of mRNA translation. This ensures that the apoferritin heteropolymer is only synthesised when intracellular iron is abundant.

Iron regulates apoferritin synthesis by modulating the interaction between the cytoplasmic iron regulatory peptides IRP-1 and IRP-2 and cis-acting stem-loop motifs in the 5′ untranslated regions (5′UTRs) of the (mRNA)FTL and (mRNA)FTH termed the iron responsive elements (IREs).3, 4 When cytoplasmic iron is scarce, the IRPs bind the IREs and mRNA translation is repressed. However, when iron is abundant, a conformational change in IRP-1 prevents its interaction with the IREs and degradation of IRP-2 is increased.5 In the absence of the negative regulatory influence of the IRPs, the (mRNA)FTL and (mRNA)FTH are translated constitutively and apoferritin is synthesised.3

Hereditary hyperferritinaemia cataract syndrome (HHCS; OMIM 600886) is an autosomal dominant disorder of ferritin synthesis characterised by early onset cataracts and increased serum concentration of L-ferritin.6, 7 Affected individuals show dysregulated translation of (mRNA)FTL that is independent of the availability of iron.8 This results in increased concentrations of apoferritin polymers that are rich in L-ferritin in the tissues and serum.9 The cataracts comprise crystalline inclusions within the lens stroma that contain immunoreactive L-ferritin.10, 11 These may arise from overexpression of L-ferritin8 and failure of efficient secretion of this L-ferritin load by the lens epithelium.12

Individuals with HHCS show point mutations or short deletions in the region of the FTL 5′UTR corresponding to the (mRNA)FTL IRE.13, 14 Nucleotide substitutions that disrupt the secondary structure of IRE stem-loop motifs reduce the affinity of IRP binding in vitro.15 It has therefore been proposed that HHCS arises through disruption of the L-ferritin RP-IRE interaction in vivo leading to failure of suppression of constitutive translation of (mRNA)FTL.

Although the pathogenesis of HHCS is now understood, the clinical features of this disorder have been examined systematically in only a single previous European case series.16 In order to characterise further the clinical phenotype of HHCS and to examine the relationship of disease severity to specific mutations in FTL, we have analysed the clinical features of seven kindreds with HHCS from the United Kingdom.

Methods

Clinical assessment of patients

The seven probands described in this study were referred for investigation of elevated serum L-ferritin concentration between 1997 and 2002. A detailed history of iron status, the age of onset and treatment of cataracts and the presence of comorbid disorders was sought in all cases. In all, 42 additional affected family members were identified by a history of symptomatic cataract in childhood. From this group, 24 individuals and a further 10 family members with no history of cataract were assessed clinically by one of the authors and were venesected for measurement of serum L-ferritin concentration and determination of FTL genotype. Cataract morphology was determined in at least one affected individual from each kindred by slit-lamp examination by a specialist ophthalmologist. Electron microscopy of lens tissue obtained from a member of kindred 7 during cataract extraction was performed as described previously.10

The serum L-ferritin concentration was measured using a standard solid-phase chemiluminescent immunometric assay (DPC Immulite 2000) that is sensitive to L-ferritin but not H-ferritin. Evidence of iron deficiency was sought by measuring the haemoglobin concentration and mean red cell volume. Biochemical evidence of either iron deficiency or overload was determined by measuring the serum transferrin saturation.

FTL sequencing and HFE genotype analysis

Total genomic DNA was extracted from venous blood obtained from each proband and from available family members with informed consent for genotypic analysis. The FTL 5′UTR was amplified by PCR using the oligonucleotide primers AGAAGCCGCCCTAGCCACG (forward) and GAGCTAACCACAAAAACGGTGC (reverse) using experimental conditions described previously.17 Sequencing was performed in forward and reverse directions using the PCR primers on a model 3700 DNA analyser using an ABI PRISM Big Dye V2 terminator ready reaction kit (Applied Biosystems, Foster City, CA, USA). HFE genotype was determined using the PCR-SSCP method described previously.18

Molecular modelling

A predicted mRNA secondary structure of wild-type and mutant IREs was determined using the mfold web server available at http://www.bioinfo.rpi.edu/~zukerm/.19

Results

Clinical characteristics

The clinical characteristics of the seven probands are shown in Table 1. All the subjects were of Caucasian origin except the proband from kindred 6 who was of Thai extraction. Six probands were female and the median age of diagnosis of HHCS was 37 years. In all the probands, HHCS was identified after the incidental discovery of an elevated serum L-ferritin concentration (median minimum recorded L-ferritin 1420 μg/l; range 740–1960). The serum L-ferritin concentration varied over time within each individual, but there was no discernable trend for the L-ferritin concentration to increase with advancing age (Figure 1a). With the exception of one proband, serum L-ferritin was measured as part of the investigation of suspected anaemia or anaemia ‘screening’ during pregnancy. Although all the probands had increased serum L-ferritin concentration, there were no clinical features of iron overload and none showed elevated serum transferrin saturation (Table 1). Three probands instead showed reduced mean red cell volume and reduced serum transferrin saturation indicating iron deficiency.

Table 1 Clinical features and genotype of seven probands and 42 affected family members with HHCS
Full size table
Figure 1
figure 1

(a) Pedigree of kindred 5 showing age at diagnosis and serum L-ferritin concentration (μg/l). The proband is shown arrowed. There was complete segregation within the kindred between elevated serum L-ferritin concentration and a history of cataract. (b) Ophthalmalogical findings in individuals (i) IV3 and (ii) IV1 showing sparse ‘breadcrumb’ opacities. (iii) Individual III2 shows ‘sunflower-like’ cataract morphology with lens opacities extending to the lens periphery (images courtesy of Andre Ismail and Peter Hodgkins, Southampton Eye Unit). (iv) The electromicrographic appearance (× 50 000 magnification) of lens tissue obtained at cataract extraction from an individual from kindred 2. The lens opacities appear as square-shaped crystalline inclusions in the lens stroma.

Full size image

Among the 42 family members who gave a history of childhood cataract, all 24 of the subjects who were available for testing showed elevated serum L-ferritin concentration (Table 1). In contrast, the 10 available family members with no history of cataract showed serum L-ferritin concentration within the normal range. Among the affected individuals, there was wide variation both within and between the kindreds in the age at which cataract was formally diagnosed and treated and in the serum L-ferritin concentration (Figure 1a; Table 1). Other than visual impairment from cataracts, there were no other common clinical features shared by affected individuals.

In order to determine whether there was a relationship between inheritance of HHCS and Hereditary Haemochromatosis (HH; OMIM 235200), the haemochromatosis gene (HFE) was studied to identify whether the C282Y or H63D substitutions were present in the HHCS probands. The probands from kindreds 5 and 2 were heterozygous for the HFE C282Y (undetermined origin) and H63D (paternal origin) substitutions, respectively. These individuals had serum L-ferritin concentrations that were similar to those of other family members with HHCS who did not show these HFE substitutions. The probands from kindreds 1 and 6 had previously received erroneous diagnoses of HH before genotypic diagnosis of this disorder was available. One other individual from each kindred had previously undergone liver biopsy that showed no increase in iron staining. Two members of kindred 6 underwent unnecessary therapeutic venesection.

All the probands had cataracts recognised in their first decade and with the exception of one individual, all had undergone cataract extraction (median age 25 years; range 22–42). The evolution of cataracts was studied in detail in kindred 5. Individuals IV1 and IV3 (aged 4 years and 6 weeks respectively) showed sparse ‘breadcrumb-like’ lens opacities, mostly in the posterior cortex (Figure 1b). Individual III2 at age 15 years showed a ‘sunflower-like’ cataract with opacities extending radially to the lens periphery (Figure 1b). Affected adults from the other kindreds showed similar cataract morphology. Electron microscopy of the lens stroma from a member of kindred 2 who underwent cataract extraction at 48 years old showed that the lens opacities comprised crystalline inclusions (Figure 1b). We have previously shown that these contain immunoreactive L-ferritin.10

FTL mutations and molecular modelling

All the probands were heterozygous for one of six nucleotide substitutions within the FTL 5′UTR (Table 1). There was complete segregation between the FTL genotype and disease phenotype in the 24 affected and 10 unaffected family members from the seven kindreds who were available for testing.

All the observed mutations in FTL were predicted to result in nucleotide substitutions in the IRE of the (mRNA)FTL. The 39C → T and 40A → G mutations were predicted to alter the nucleotide sequence of the IRE apical loop without disrupting Watson–Crick pairing within the IRE stem (Figure 2). The remaining substitutions lay either in the upper stem (36C → A) or in the region of the upper unpaired cytosine bulge of the IRE motif (33C → T, 32G → C and 32G → T). These substitutions were predicted to induce alternative energetically favourable Watson–Crick pairings within the IRE stems disrupting the secondary structure of the apical loops (Figure 2).

Figure 2
figure 2

The predicted secondary structures of the (a) wild-type and (b–g) variant (mRNA)FTL IREs observed in the study kindreds with nucleotide substitutions indicated by asterixes. The wild-type IRE is a stem-loop motif comprising an apical loop (AL; nucleotides 39–44) flanked by palindromic sequences that associate through Watson–Crick pairing to form a stem structure. The stem contains two unpaired cytosine residues that form the cytosine bulges (CB; nucleotides 27 and 33). The substitutions C39T (b) and A40G (c) change the primary nucleotide sequence of the apical loop without disrupting the Watson–Crick pairing within the stem. In contrast, the substitutions C36A, G32C, G32 T and C33T (d–g) result in alternative energetically favourable Watson–Crick pairings within the upper stem that induce altered secondary structure of the apical loop. Nucleotides are numbered from the transcription start site.

Full size image

Discussion

HHCS is a recently described disorder6, 7 that has now been reported from European,20, 21 North American22 and Australian13 centres. This description of a case series from the United Kingdom highlights the wide geographical distribution of HHCS. We show that the only consistent clinical findings in HHCS are hyperferritinaemia and childhood cataract but that there is wide variation in phenotype. Our study therefore provides independent conformation of the findings of the previous European case series of HHCS that comprised 62 affected individuals.16

L-ferritin is a widely used marker of body iron load because its rate of synthesis is closely regulated by the availability of iron in the tissues through the IRP-IRE system.2, 3 Therefore, increased serum L-ferritin concentration usually indicates iron overload. It has been proposed that HHCS arises from selective failure of suppression of L-ferritin synthesis and that L-ferritin-rich homopolymers accumulate in the tissues and serum irrespective of the availability of iron.8 The observation that our probands showed increased serum L-ferritin, but no iron overload supports this model of pathogenesis of HHCS.

Instead of iron overload, three probands, who had not previously undergone therapeutic venesection, showed iron deficiency at presentation. Similar findings have been noted in the probands of other kindreds with HHCS16 and this raises the possibility that overexpression of L-ferritin may have a more subtle effect on iron metabolism in this disorder. However, serum L-ferritin is commonly measured only in individuals with clinically suspected iron deficiency suggesting that this finding may reflect selection bias. Family members with HHCS in this, and other series16 showed no consistent evidence of iron deficiency so this abnormality is unlikely to be part of the HHCS phenotype. Selection bias may also explain the high frequency of young female subjects within our probands. However, female bias was also evident among the affected family members in our series. Previous short series from other centres have shown no sex bias in HHCS16, 23 and the reasons for the female preponderance among subjects from our series remain unclear.

The demonstration of hyperferritinaemia as a heritable trait in two of our kindreds led to erroneous clinical diagnoses of HH and consequently, unnecessary investigation and treatment. Similar misdiagnoses have arisen in other reported individuals with HHCS.24, 13 This observation emphasises that HH should only be considered as an explanation for familial hyperferritinaemia if there is recessive inheritance, biochemical evidence of iron overload and an abnormal HFE genotype. Hyperferritinaemia may be inherited as an dominant trait in kindreds with hereditary haemochromatosis type 4 (OMIM 606069) and classical HH may also follow pseudo-dominant inheritance in some kindreds because the population frequency of the responsible polymorphisms is high. In both cases, hyperferritinaemia is associated with iron overload and affected individuals do not have premature cataracts and so can be distinguished from HHCS. Ironically, two probands in our series, who had not received diagnoses of HH, were heterozygous for either the HFE C282Y or H63D substitutions. Heterozygosity for these polymorphisms was not associated with an iron overload phenotype in a large population study25 and inheritance of these alleles in our HHCS probands did not alter their HHCS phenotype.

The affected individuals in our series showed the same highly distinctive ‘sunflower-like’ cataract morphology that closely resembled previous descriptions of cataracts in HHCS.6, 7, 23 Our observations also suggest that cataracts increase in density through early life even though serum L-ferritin concentrations appear to reach steady state very rapidly. Progression of cataracts has previously been noted in a single longitudinal ophthalmic study of children with HHCS.23

All our kindreds showed point mutations in the FTL 5′UTR that segregated with the HHCS phenotype. These mutations were all predicted to cause nucleotide substitutions in the IRE of the (mRNA)FTL and have now been recognised in other individuals with HHCS.6, 14, 16, 21, 22, 26, 27 Mutagenesis studies of (mRNA)FTL IREs in vitro suggest that the primary nucleotide sequence in the apical loop of the motif is an important determinant of IRP binding.15, 28 Our findings support this hypothesis since two of the observed nucleotide substitutions lay within the apical loop of the IRE motif. We have also shown that substitutions in the upper stem of the IRE are predicted to change the secondary structures of the IRE apical loops. In our series, and in other reports,13, 16 mutations were common around the upper unpaired cytosine bulge (residues +32 and +33) indicating the importance of this region in maintaining IRE structure.28 These findings support the proposal that the overexpression of FTL in HHCS arises from impaired IRP-IRE binding.9 Genotypic diagnosis of HHCS is straightforward since the observed mutations are closely grouped to a narrow functional region of FTL and may be a useful diagnostic approach, particularly in the rare HHCS kindreds with a mild cataract phenotype.16, 29

The poor correlation within this series between the site of the nucleotide substitution in FTL and severity of the HHCS phenotype arose largely because of the wide variation in phenotype within kindreds. This may indicate true phenotypic variability in individuals with the same HHCS genotype. More likely, it reflects difficulties in quantifying the clinical severity of HHCS caused by physiological variation in serum L-ferritin concentration and differences between referral centres in cataract detection and management. This was particularly evident in the wide variation in the timing of surgical treatment of cataract in the affected family members even though they were all symptomatic of cataract in childhood. Determination of FTL genotype, serum ferritin concentration and severity of cataract in other affected family members is therefore unlikely to enable to accurate prediction of the rate of progression of cataract in newly diagnosed individuals with HHCS.

HHCS represents a valuable model of a highly penetrant but variably expressive human genetic disorder resulting from abnormal mRNA translation. It is conspicuous that no patients from our series presented through ophthalmology or clinical genetics sources. Therefore, it is likely that HHCS is under-recognised in these specialties. HHCS should be considered in all individuals with autosomal dominant cataract and should be included in the differential diagnosis of familial hyperferritinaemia. HHCS should be carefully distinguished from HH and other iron overload disorders.