MORPHOLOGIC ASPECTS: THE SIDEROBLASTS
Sideroblasts are erythroblasts containing aggregates of non–heme iron that appear as one or more Prussian blue–positive granules on light microscopy.30 The morphology of these cells in normal and abnormal states is discussed in detail in Chap. 31. In normal marrow, virtually every erythroblast has siderosomes, iron-containing organelles that are demonstrable by transmission electron microscopy. Light microscopy of Prussian blue–stained marrow aspirates or biopsy sections is a relatively insensitive method to identify these structures. One can usually identify approximately 25 to 35 percent of erythroblasts with one to three very fine Prussian blue–stained granules in the cytoplasm of a well-prepared marrow sample. Pathologic sideroblasts may be of two types: erythroblasts with a marked increase in the number and size of siderotic granules in the cytoplasm, compared to normal erythroblasts, or ringed sidero-blasts. Ringed sideroblasts are the hallmark of the sideroblastic anemias. In contrast to the normal cytoplasmic location of siderotic granules, the pathologic sideroblasts in the sideroblastic anemias have large amounts of iron deposited as dust- or plaque-like ferruginous micelles between the cristae of mitochondria (Fig. 59–1).31 The iron-loaded mitochondria are distorted and swollen, their cristae are indistinct, and identification of mitochondria may itself be difficult. In humans, the mitochondria of the erythroblast are distributed perinuclearly,23 which accounts for the distinctive “ringed” sideroblast identified by Prussian blue staining when mitochondrial iron overload is present (Fig. 59–1). The morphologic features that characterize pathologic sideroblasts in various disorders have been summarized.32
Marrow films. A. Normal marrow stained with Prussian blue. Note several erythroblasts without apparent siderotic (blue-stained) granules. The arrow indicates erythroblast with several very small cytoplasmic blue-stained granules. It is very difficult to see siderosomes in most erythroblasts in normal marrow because they are often below the resolution of the light microscope. B. Sideroblastic anemia. Note the florid increase in Prussian blue–staining granules in the erythroblasts, most with circumnuclear locations. These are classic examples of ringed sideroblast, which are by definition pathologic changes in the red cell precursors. In some cases, cytoplasmic iron granules are also increased in size and number, also a pathologic change. (Reproduced with permission from Lichtman’s Atlas of Hematology, www.accessmedicine.com.)
The pathogenesis of most of the sideroblastic anemias is not well understood.33,34 It is not clear whether the basic mechanism by which abnormal accumulations of intramitochondrial iron occurs is the same in inherited and acquired forms of the disease. However, it seems appropriate, given the present state of knowledge, to discuss both forms together. The pathogenesis of the disorder may be viewed from two standpoints: the underlying biochemical lesions and the mechanism(s) of the anemia itself.
Biochemical Lesions and Genetics
In the search for the biochemical lesions responsible for the development of sideroblastic anemia, attention has been focused upon an intramitochondrial defect in heme synthesis and on possible disturbances in pyridoxine metabolism.
Defects in Heme Synthesis
The role of defects in heme biosynthesis have occupied central stage since the early studies of Garby and colleagues,35 who postulated that such a defect might exist; they demonstrated that the level of free erythrocyte protoporphyrin was decreased. Subsequently, a variety of abnormalities of the levels of precursors and of their rate of incorporation into heme was documented (Chap. 58).36,37,38,39,40,41 However, the findings have not all been consistent, as levels of free erythrocyte protoporphyrin have often been increased,42,43 not diminished. The role of mitochondria in the etiology of sideroblastic anemia gained further credence when mutations of the mitochondrial genome were found in patients with Pearson syndrome.15,16,17,18,19
Hereditary Sideroblastic Anemias
Shortly after the identification of erythroid-specific ALA synthase (ALAS2, the first enzyme in heme synthesis; Fig. 59–2), it became apparent that most patients with hereditary X-linked sideroblastic anemias (XLSAs) had mutations in the ALAS2 gene.44,45,46 However, a proportion of patients with congenital sideroblastic anemia had autosomal recessive inheritance. At least some such patients have a defect in the gene encoding the erythroid-specific mitochondrial carrier protein, SLC25A38.47 This transporter is important for the biosynthesis of heme in eukaryotes and it was proposed that this protein may be translocating glycine into mitochondria (Fig. 59–2).47 Hence, SLC25A38 defects would be expected to generate a phenotype identical to that seen in patients with defects in ALAS2. One can speculate that, in erythroid cells, a common control mechanism exists that regulates acquisition of the two substrates for heme synthesis (iron and glycine).
Schematic of iron uptake from transferrin and its delivery to the hemoglobin (Hgb) molecule. Extracellular differic transferrin is bound by the membrane-bound transferrin receptor (TfR) and internalized via receptor-mediated endocytosis into an endosome. Iron is released from transferrin by a decrease in pH (~pH 5.5), reduced by STEAP 3 (six-transmembrane epithelial antigen of prostate 3-ferric reductase), following which the metal is transported through the endosomal membrane by DMT 1. In erythroid cells, more than 90 percent of iron must enter mitochondria wherein ferrochelatase (FECH), the enzyme that inserts Fe2+ into protoporphyrin IX (Proto IX), resides on the inner leaflet of the inner mitochondrial membrane. The transport of coproporphyrinogen (Copro’gen) into mitochondria is not fully understood. Neither mechanisms nor the regulation of the transport of heme from mitochondria to globin polypeptides are known; however, it has been proposed that a carrier protein, heme binding protein 1 (gene: HEBP1), is involved in this process. CPO, coproporphyrinogen oxidase; NAD(P), nicotinamide adenine dinucleotide phosphate; NAD(P)H reduced form of nicotinamide adenine dinucleotide phosphate; PPO, protoporphyrinogen oxidase. (Reproduced with permission from Anderson GJ, McLaren G: Iron Physiology and Pathophysiology in Humans, New York, NY: Humana Press; 2012.)
Hereditary sideroblastic anemia with spinocerebellar degeneration with ataxia is a rare X-linked syndrome that appears to be distinct from the other forms of sideroblastic anemia.48,49,50,51 It is caused by mutation of ATP-binding cassette (ABCB7).48,52,53
Heteroplasmic point mutations in subunit 1 of the mitochondrial cytochrome oxidase have been documented in some patients with sideroblastic anemia.54,55,56
Rare autosomal forms of inherited sideroblastic anemia have also been reported,57,58 including those with a deficiency of uroporphyrinogen decarboxylase59,60 and ferrochelatase (FECH)36,41,61,62,63 enzymes, both necessary for the synthesis of heme (Chap. 58). The other reported defects of ferrochelatase could result from the inhibitory effect of mitochondrial iron overload on enzyme activity.41 A defect in coproporphyrinogen oxidase (CPO) could not be confirmed by direct measurement.64
An unusual phenotype with of inherited sideroblastic anemia, developmental delay with variable neurologic defects and B-cell lymphopenia with hypogammaglobulinemia was reported of yet unknown etiology.65
A role for pyridoxine has been supported by the demonstration that pyridoxine deficiency in animals is a prototype of sideroblastic anemia.31 Sideroblastic anemia can be induced by drugs that reduce the level of pyridoxal phosphate in blood, which decreases the ALAS2 activity in normoblasts.22,36,40 Moreover, certain sideroblastic disorders, although not a result of pyridoxine deficiency, are nonetheless responsive to pharmacologic doses of pyridoxine.44,66,67,68 Pyridoxal phosphate is a necessary coenzyme for the initial reaction of protoporphyrin synthesis, the condensation of glycine and succinyl coenzyme A to form ALA, a reaction mediated by ALA synthetase (Chap. 58). Furthermore, pyridoxal phosphate is a factor in the enzymatic conversion of serine to glycine (Chap. 41). This reaction generates a form of folate coenzyme necessary for the formation of thymidylate, an important step in DNA synthesis. Pyridoxal 5′-phosphate, the active form of the coenzyme, must itself be enzymatically synthesized from pyridoxine. Deficiencies in its biosynthesis have also been invoked as the possible cause of certain sideroblastic anemias,27,69 but direct measurements of pyridoxal kinase failed to confirm that the postulated lesion was present.70
Other Metabolic Defects and Acquired Associations with Sideroblastic Anemia of Uncertain Significance
Increased levels of uroporphyrinogen 1 synthase are commonly encountered in patients with sideroblastic anemias.39 Alcohol, a common cause of secondary sideroblastic anemia, inhibits heme synthesis at several steps.38 Dramatically altered activity ratios of a wide diversity of enzymes have been described,71,72 for example, elevated arginase activity.
Sideroblastic anemia has been found in a patient with apparent antibody-mediated red cell aplasia.73 There are alterations in red cell antigen patterns frequently with an increase of i and a loss of A1 antigens (Chap. 136).74 Similar findings occur in certain hereditary and acquired refractory anemias with cellular marrows but without ringed sideroblasts.72 Such dyscrasias are also characterized by ineffective erythropoiesis and, except for the lack of ringed sideroblasts, may in some instances be virtually indistinguishable from their sideroblastic counterparts.75
Pathogenesis of Ring Sideroblast Formation
Iron accumulation within mitochondria is an unusual pathologic phenomenon occurring only in erythroblasts of patients with sideroblastic anemias and, to a much lesser degree, in cardiomyocytes of patients with Friedreich ataxia.76,77 Mitochondrial iron accumulation has not been demonstrated in patients with either primary or secondary iron overload. The pathophysiology of ring sideroblast formation in patients with ALAS2 defects and those resulting from inhibitors of porphyrin biosynthesis (see Table 59–1) is likely because of the unique aspects of the regulation of iron metabolism and heme synthesis in erythroid cells (Chap. 58).78 These differences can account for the accumulation of non–heme iron in erythroid mitochondria of sideroblastic anemia patients. In hemoglobin-synthesizing cells, iron is specifically targeted toward mitochondria that avidly take up iron even when the synthesis of protoporphyrin IX is suppressed (Chap. 58).79,80,81,82 In contrast, nonerythroid cells store iron in excess of metabolic needs within ferritin.83 Hence, erythroid-specific mechanisms and controls are involved in the transport of iron into mitochondria in erythroid cells, but the nature of these processes, including the role of mitoferrin 1 (Chap. 42); an inner mitochondrial membrane protein, which presumably provides Fe2+ to FECH,84 is poorly understood. The transferrin-bound iron is used for hemoglobin synthesis78,82 with a high degree of efficiency and is targeted into erythroid mitochondria (Chap. 52), and because no intermediate for cytoplasmic iron transport has ever been identified in erythroid cells, the following hypothesis of intracellular iron trafficking in developing red cells has been proposed (see Fig. 59-2 from Chap. 52). This model postulates that iron released from transferrin in the endosome is passed directly from protein to protein until it reaches FECH, which incorporates Fe2+ into protoporphyrin IX85 in the mitochondrion. Such a transfer bypasses the cytosol, as the movement of iron between proteins could be mediated by a direct interaction of the endosome with the mitochondrion.78,86 The results of supporting experiments revealed that (1) iron, delivered to mitochondria via the transferrin–transferrin receptor (TfR) pathway, is unavailable to cytoplasmic chelators87,88; (2) transferrin-containing endosomes move to and contact mitochondria in erythroid cells; and (3) endosomal movement is required for iron delivery to mitochondria (see Fig. 59–2).87 These studies also revealed that cytoplasmic iron not bound to transferrin is inefficiently used for heme biosynthesis and that the endosome–mitochondrion interaction increases chelatable mitochondrial iron.87
An important distinction between erythroid and nonerythroid cells is the presence of a feedback mechanism in which “uncommitted” heme inhibits iron acquisition from transferrin.89,90,91,92 Although it is still unresolved whether heme inhibits transferrin endocytosis89,90 or iron release from transferrin,92 the lack of heme plays an important role in mitochondrial iron accumulation. Additionally, non–heme iron, which accumulates in erythroid mitochondria, cannot be released from the organelle unless it is inserted into heme.82 This finding suggests that mitochondria can release iron only when the metal is in a proper chemical form, in this case, inserted into protoporphyrin IX. These considerations provide framework to the pathogenesis of mitochondrial iron accumulation in erythroblasts of patients with sideroblastic anemia caused by ALAS2 defects, as well as those caused by agents inhibiting porphyrin biosynthesis (see Table 59–1).
Of considerable interest, in 2014, Fleming and his coworkers93 demonstrated that mutations in an enhancer element in ALAS2 intron 1, which contains a GATA-binding site, cause a clinical phenotype similar to patients with XLSA resulting from mutations in the ALAS2 coding sequence itself.
A distinct form of XLSA, that associated with ataxia (XLSA/A), was described in several families with putative mutations mapped to chromosome region Xq13.50 In contrast to ALAS2-linked disease, the XLSA/A syndrome is associated with elevated erythrocyte protoporphyrin IX levels. It was demonstrated that mutations of the ABCB7 gene is responsible for XLSA/A,48 and this was confirmed by other reports.52,94 The ABCB7 protein is thought to transfer iron-sulfur (Fe-S) clusters from mitochondria to the cytosol (Chap. 42).76,95,96 How the disruption of Fe-S cluster export might impede heme biosynthesis is, however, not clear, but the accumulation of erythrocyte zinc–protoporphyrin IX is found in XLSA/A.48,50,52 Additionally, mouse erythrocytes with mutated (E433K) ABCB7 have an increase in zinc–protoporphyrin IX-to-heme ratios.97 Because the formation of zinc–protoporphyrin IX requires FECH, ABCB7 mutations cannot interfere with the activity of this enzyme. Instead, the loss of function of ABCB7 may, by a yet-to-be-defined mechanism, diminish the availability of reduced iron (the only substrate of iron for FECH) required for the assembly of heme from protoporphyrin IX. In XLSA, as in ALAS2-associated sideroblastic anemia, decreased levels of heme likely contribute to the pathogenesis of ring sideroblast formation.
Another type of hereditary hypochromic anemia was described in shiraz (sir) zebra fish mutants.98 These mutants have a deficiency of glutaredoxin 5 (GLRX5) encoded by a gene (GLRX5) whose product is required for Fe-S cluster assembly. This study demonstrated that the loss of the Fe-S cluster in the iron-regulatory protein 1 (IRP1) blocked ALAS2 translation by binding to the iron-responsive element (IRE) located in the 5′-untranslated region of ALAS2 mRNA. Subsequently, a case of GLRX5 deficiency in an anemic male with iron overload and a low number of ringed sideroblasts was reported.99 As in zebra fish with shiraz mutants, ferritin levels were low and TfR levels were high in the patient’s cells; this can be explained by increased IRP1 binding to IREs in mRNAs of these two proteins. However, erythroblasts from zebra fish shiraz mutants were not found to contain iron-loaded mitochondria.
Primary Acquired Sideroblastic Anemia (Refractory Anemia with Ring Sideroblasts—Myelodysplastic Syndrome)
The pathophysiology of acquired idiopathic sideroblastic anemias associated with myelodysplastic syndrome is distinct from the above discussed XLSAs. In these patients there is no evidence for a decrease in the formation protoporphyrin IX levels; instead, the amount of protoporphyrin IX is moderately increased.43 Impaired iron reduction could cause intramitochondrial iron accumulation in patients with myelodysplastic syndromes. The ferric reductase, STEAP 3 (six-transmembrane epithelial antigen of prostate 3-ferric reductase), is involved in the reduction of Fe3+ to Fe2+ in endosomes.100 Based on the model of the direct interorganellar transfer of iron (see Fig. 59–2), it can be assumed that there is only one reduction step during the path of iron from endosomes to FECH. However, the efficient insertion of ferrous ions into protoporphyrin IX may still require a reducing environment in mitochondria that would be provided by an uninterrupted respiratory chain. This proposal is compatible with the fact that sideroblastic anemia accompanying Pearson marrow-pancreas syndrome101 is caused by deletions of mitochondrial DNA genes, products of these are involved in electron transport.102 Indeed, there are at least some myelodysplasia-associated sideroblastic anemia patients described caused by acquired mutations in cytochrome oxidase, encoded by mitochondrial DNA.54,55,103,104,105 However, a rigorous study failed to find cytochrome oxidase mutations in 10 patients with myelodysplasia-associated sideroblastic anemia.106 Alternatively, these is some evidence that ABCB7 (see the above discussion on XLSA/A) could be a possible candidate gene for the formation of ringed sideroblasts in refractory anemia with ring sideroblasts.107
Mitochondrial Myopathy and Sideroblastic Anemia
There are some similarities and some dissimilarities between Pearson marrow-pancreas syndrome and patients with mitochondrial myopathy and sideroblastic anemia (MLASA).57,108,109 In both cases, there are defects in the mitochondrial electron transport chain, likely creating an environment that retards iron access to FECH in the reduced form. Both disorders are hereditary, but Pearson syndrome is caused by large deletions of mitochondrial DNA, whereas MLASA results from a homozygous missense mutation in the genomic DNA of pseudouridine synthase 1, encoded by PUS1 gene.108 It has been proposed that deficient pseudouridylation of mitochondrial transfer RNAs (tRNAs) explains the pathogenesis of MLASA type of sideroblastic anemia.108
Mitochondrial ferritin is a ferritin isoform with ferroxidase activity that is expressed only in mitochondria (Chap. 42). This protein is encoded by an intronless nuclear gene and can store iron within a shell of homopolymers.110,111,112 Although the function and regulation of expression of this protein is not fully understood, the induction of mitochondrial ferritin causes the transfer of iron from cytosolic ferritin to mitochondrial ferritin.113 The mitochondrial ferritin has a very low expression in all tissues except testis.110,112 Although mitochondrial ferritin is not expressed in normal erythroblasts, it is expressed in ring sideroblasts of patients with sideroblastic anemias,114 caused by ALAS2 defects, as well as those associated with myelodysplastic syndromes. In both, iron is sequestered within mitochondrial ferritin.114 Because mitochondrial ferritin has ferroxidase activity, it likely protects mitochondria by converting the toxic ferrous iron to ferric iron that is stored. Further research is needed to explain the mechanism of mitochondrial ferritin induction in erythroblasts of patients with sideroblastic anemias, both hereditary and acquired. Whether the mitochondrial ferritin also accumulates iron in ring sideroblasts of patients with XLSA/A has not yet been studied.
The dominant factor that determines anemia is ineffective erythropoiesis (intramedullary apoptosis); the rate of red cell destruction is usually near-normal or only moderately accelerated to levels for which a normally functioning marrow can compensate.115 The half-time of disappearance of intravenously injected tracer doses of radioactive iron is usually is rapid (25 to 50 minutes; normal mean: 90 to 100 minutes) but in some patients may be normal. The plasma iron turnover tends to be increased (1.5 to 5.9 mg/dL of blood per day; normal: approximately 0.30 to 0.70 mg/dL per day), but incorporation of radioactive iron into heme and its delivery to the blood as newly synthesized hemoglobin are depressed (15 to 30 percent of tracer dose; normal: 70 to 90 percent). Red cell survival ranges between 40 and 120 days, indicating some cases have moderate or only very slightly shortened red cell life-span, whereas in other cases red cell survival is normal. As in other kinds of anemia characterized by ineffective erythropoiesis, the fecal stercobilin excreted per day may be greater than can be accounted for by the daily catabolism of circulating hemoglobin.