- Top of page
- Patients and methods
Two brothers with X-linked ataxia (XLA) were found to have hypochromic red cells and increased erythrocyte protoporphyrin despite normal iron stores. The mother was unaffected by ataxia and had normal iron stores but showed evidence of some red cell hypochromia with heavy basophilic stippling that stained positive for iron. Bone marrow biopsy confirmed the presence of ring sideroblasts in one of the brothers. The absence of mutations in the ALAS2 gene and the predominance of zinc over free protoporphyrin led to a search using a combination of DNA and cDNA analysis for the presence of mutations in the ABC7 gene. ABC7 encodes a mitochondrial half-type ATP Binding Cassette transporter involved in iron homeostasis. The published cDNA sequence was used to search databases for the genomic sequence of which 12 exons spanning 23·4 kb were mapped leaving the most 5′ nucleotides unaccounted for. The identified exons and their exon–intron boundaries were amplified from DNA while the most 5′ sequence including the initiation codon was amplified from cDNA of peripheral blood cells. Direct sequencing revealed hemizygosity in the brothers and heterozygosity in the mother for a GC transversion at position 1299 of the published cDNA. This predicts a V411L substitution at the beginning of the last of six putative transmembrane regions of the protein. Restriction enzyme digestion confirmed the presence of this mutation in the three family members but could not detect it in 200 normal alleles. An uncle affected by ataxia also carried this mutation. This study supports the recently hypothesized involvement of the ABC7 gene in XLSA/A and highlights a protein structure region of importance to this syndrome.
The sideroblastic anaemias (SA) comprise a number of different conditions with varied haematological features and clinical outcomes. They are characterized by the presence in erythroblasts of iron-loaded mitochondria visualized by Perl's staining as a perinuclear ring of Prussian blue granules. Ineffective erythropoiesis is a typical associated abnormality and may predominate causing anaemias of varying severity. Some patients require regular blood transfusions for survival while others may have such a need offset only by supplementation of the diet with pyridoxine or thiamine. The red cells show some degree of hypochromia and may contain basophilic, iron-staining inclusions known as Pappenheimer bodies. Iron overload as a result of increased iron absorption, treatment with blood transfusions or inappropriate treatment with iron is an important cause of morbidity and mortality (Bottomley, 1998). When the sideroblastic anaemia is part of a clinical syndrome such as Pearson's syndrome, other aspects of pathology may be the focus of attention for its clinical management (Pearson et al, 1979).
The genetic causes of ringed sideroblasts show various patterns of inheritance indicating heterogeneity. Until recently, mutations in the X-linked, erythroid-specific ALA synthase gene (ALAS2) and large deletions of mitochondrial DNA were the only known molecular causes (reviewed in May & Fitzsimons, 1994). More recently, homozygosity mapping has identified two further loci at 4p16.1 and 1q23.3 involved in thiamine-responsive forms of SA with autosomal recessive inheritance (Borgna-Pignatti et al, 1989; Inoue et al, 1998; Strom et al, 1998; Labay et al, 1999). Somatic mutations in cytochrome oxidase and other mitochondrial genes have been implicated in refractory anaemia with ring sideroblasts, an acquired form of SA (Gatterman et al, 1996; Gatterman, 1999). Autosomal recessive inheritance has been observed in SA unresponsive to thiamine and autosomal dominant inheritance has also been observed (Bottomley, 1998).
In addition, a form of X-linked SA distinct from those caused by abnormalities in ALAS2 has been described. This form of sideroblastic anaemia is rare and is associated with spinocerebellar ataxia (XLSA/A; OMIM number 301310) (Pagon et al, 1985). Haematological analysis of the affected hemizygotes revealed microcytic, hypochromic red cells with raised erythrocyte protoporphyrin levels. Patients presented with a mild anaemia and spinocerebellar ataxia from an early age. The red cells from the obligatory carrier women were affected to a variable extent. Some but not all carriers showed raised protoporphyrin levels and some had ringed sideroblasts in the bone marrow. No female carrier, however, has been reported as being affected by the ataxia and the inheritance is described as X-linked recessive. Using linkage analysis, the condition was mapped to Xq13 (Raskind et al, 1991), a region distinct from that associated with pyridoxine-responsive XLSA.
The human ABC7 gene maps to Xq13 and encodes a ‘half-type’ ATP Binding Cassette (ABC) transporter, containing a single transmembrane domain consisting of several putative transmembrane regions and a single ATP binding domain. It is located in the inner mitochondrial membrane (Savary et al, 1997; Csere et al, 1998; Shimada et al, 1998) as a homodimer or as a heterodimer with another ‘half-type’ transporter. The yeast protein Atm1p, essential for iron homeostasis, shares 48·9% residue similarity with ABC7 (Kispal et al, 1997; Csere et al, 1998; Shimada et al, 1998) and ATM1 gene inactivation causes yeast cells to show changes reminiscent of changes that occur in erythroblasts in sideroblastic anaemia. They have a 30-fold increase in mitochondrial iron, are deficient in the holoforms but not the apoforms of haem-containing proteins, have increased glutathione, and show poor growth on certain substrates (Leighton & Schatz, 1995; Kispal et al, 1997). That this phenotype can be reversed by addition of a plasmid-born copy of ABC7 (Kispal et al, 1997; Allikmets et al, 1999) demonstrates the functional equivalence of these proteins.
Thus, ABC7 became an additional candidate gene in the study of SA and, in particular, XLSA/A. Allikmets et al (1999) reported the finding of a missense mutation in the fifth putative transmembrane region of the protein and recently, Bekri et al (2000) reported the finding of a separate mutation C-terminal to the sixth putative transmembrane region. We wish to report the finding of a third missense mutation, which lies between the two already reported, in an unrelated family with four male members affected by XLSA/A (Hellier et al, 2001). These findings confirm the involvement of this gene in this syndrome, and demonstrate a key role for this part of the protein in erythroblast iron metabolism.
- Top of page
- Patients and methods
The genomic sequence and structure of ABC7 was determined in members of a family with the biochemical and haematological features of sideroblastic anaemia and the clinical symptoms of spinocerebellar ataxia. While this manuscript was in preparation, a second mutation in ABC7 was reported (Bekri et al, 2000). When we performed RT-PCR on the first 520nt of the gene, Bekri et al (2000) characterized this portion of the gene and mapped its coding sequence to four exons, predicting 16 exons altogether. In our family we demonstrated a third mutation (cDNA: G1299C predicting V411L) in exon 9 of the ABC7 transporter gene, further strengthening the link between ABC7 and XLSA/A.
A change from valine to leucine is rather conservative and the predicted effect on protein structure slight. However, that this is the only mutation found in all affected family members tested and that it is absent in over 200 normal alleles strongly support its importance in XLSA/A in this family. Furthermore, the predicted amino acid substitutions arising from previously reported mutations I400M (Allikmets et al, 1999) and E433K (Bekri et al, 2000) and that reported here at residue 411 highlight the importance of this region of the protein in iron homeostasis and ataxia.
The precise transport function of ABC7 is not certain, but growing evidence suggests a role in the transport out of mitochondria of iron–sulphur clusters destined for the cytosol (Kispal et al, 1999; Bekri et al, 2000; Lill & Kispal, 2000). The reported mutations in ABC7 are missense causing amino acid substitutions that lie towards the C-terminal end of the transmembrane domain responsible for binding and transport of the substrate. It seems probable that these are the functions with which they interfere rather than ATP binding or hydrolysis.
Some significance may be sought from the severity of the disease. The haematological and biochemical values of the patients of this study are compared with those of the subjects of the two previous reports in Table I. In the patients described anaemia was absent or mild, the decrease in mean cell volume (MCV) was slight and red cell protoporphyrin levels were only slightly increased. In the family most recently published (Bekri et al, 2000), the amino acid change is much less conservative and is of a basic lysine residue for the acidic glutamate. The MCV and haemoglobin were markedly reduced and erythrocyte protoporphyrin greatly elevated. In the first reported family with this disorder (Pagon et al, 1985; Allikmets et al, 1999), the amino acid substitution is more conservative and the red cell changes more moderate, albeit more pronounced than those of the patients studied here. A difference in severity of ataxia between the families is, however, not obvious (Table III) and our findings confirm the conclusion of Allikmets et al (1999) that only a slight structural change to this part of the protein is required for this clinical effect. For the time being, therefore, mutations in ABC7 should be considered in any unexplained X-linked ataxias, even in the absence of haematological changes.
Table III. A comparison of the neurological assessments of families described with mutations in the human ABC7 gene and XLSA/A.
| ||Age of onset (years)||Age of walking (years)||Intention tremor/ dysmetria|| Dysarthria|| Nystagmus|| Strabismus||Deep tendon reflexes||Plantar responses|
|Bekri et al (2000)||< 1||6||No data||+||+||No data|
|Pagon et al (1985)||0·5||9/1||+||+ –||1/5||3/5||/|
|This study||< 2||11/3/5||+||+||+||+|
Attempting to explain the erythroid phenotype highlights a paucity of direct information about key regulatory phenomena, so we can only speculate. The presence of raised total erythrocyte protophorphyrin in affected individuals, despite increased iron within the mitochondrion, points to defective incorporation of ferrous iron into protoporphyrin IX by ferrochelatase, the final step of the haem synthesis pathway. Increased levels of zinc protoporphyrin in the red cells of our patients, however, demonstrate that ferrochelatase is active. This poses the question of why disruption in the export of Fe–S clusters or their precursors should prevent incorporation of iron into haem. Ferrochelatase accepts ferrous but not ferric iron as a substrate and Lange et al (1999) showed that reduced iron for haem synthesis, in yeast, is generated and delivered to ferrochelatase as required and is not preformed. A physical link between the import of ferrous iron and the export of Fe–S would be an attractive explanation for the deleterious effect on the former of deficiency in the latter.
A reduction in cytosolic iron sulphur clusters mimics a low iron state in the cytosol by converting the cytoplasmic enzyme aconitase to the iron regulatory protein IRP1 (Haile et al, 1992). In the erythroblast this should lead to a decreased synthesis of ferritin (Hentze et al, 1987), increased production of transferrin receptors (Müllner et al, 1989) and continued uptake of available iron. Whether or not this is sufficient to produce the ring sideroblasts remains to be seen; however, ALAS2 mRNA carries a 5′ iron responsive element functional in erythroleukaemia cell lines (Cox et al, 1991; Dandekar et al, 1991) and its translation may therefore be decreased as a result of negative regulation by IRP1. Alternatively, ALAS1 activity could be enhanced by haem feedback regulation sufficient to contribute in these patients to the raised protoporphyrin in the reticulocytes and red cells. The earlier stages of erythroid differentiation, however, are more dependent on high levels of ALAS2 activity which are unlikely to be replaced by ALAS1 for their haem supply. ALAS activity rather than the ferrochelatase step may therefore be limiting at these earlier stages of differentiation and its decreased activity able to contribute to the decrease in overall haem production, mitochondrial iron loading and the anaemia.