Iron incorporation into ferritins: evidence for the transfer of monomeric Fe(III) between ferritin molecules and for the formation of an unusual mineral in the ferritin of Escherichia coli

Iron that has been oxidized by H-chain ferritin can be transferred into other ferritin molecules before it is incorporated into mature ferrihydrite iron cores. Iron(III) dimers are formed at the ferroxidase centres of ferritin H chains at an early stage of Fe(II) oxidation. Mössbauer spectroscopic data now show that the iron is transferred as monomeric species arising from dimer dissociation and that it binds to the iron core of the acceptor ferritin. Human H-chain ferritin variants containing altered threefold channels can act as acceptors, as can the ferritin of Escherichia coli (Ec-FTN). A human H-chain ferritin variant with a substituted tyrosine (rHuHF-Y34F) can act as a donor of Fe(III). Since an Fe(III)-tyrosinate (first identified in bullfrog H-chain ferritin) is absent from variant rHuHF-Y34F, the Fe(III) transferred is not derived from this tyrosinate complex. Mössbauer parameters of the small iron cores formed within Ec-FTN are significantly different from those of mammalian ferritins. Analysis of the spectra suggests that they are derived from both ferrihydrite and non-ferrihydrite components. This provides further evidence that the ferritin protein shell can influence the structure of its iron core.

The role of the L-chain in ferritin iron incorporation. Studies of homo and heteropolymers

Mammalian ferritins are 24-meric proteins composed of variable proportions of H and L-subunits. The L-chain, in contrast to the H-chain, lacks detectable ferroxidase activity, and its role in ferritin iron incorporation is unclear. In this study, apoferritins were subjected to iron loading with large iron increments to favour spontaneous iron hydrolysis. The homopolymers of the wild-type H-chain, and of a mutant H-chain with an inactivated ferroxidase centre, formed massive protein aggregates, while the L-chain homopolymers remained mostly soluble. The difference between H and L-ferritins was not related to the rate of iron oxidation or to the presence of preformed iron cores. Heteropolymers were constructed in vitro by co-renaturing different proportions of the H-chain with the L-chain or mutant H-chain with an inactivated ferroxidase centre. After loading with high iron increments, protein aggregation of the heteropolymers was reduced when the L-chain content was above 70 to 80%, either in combination with the wild-type H-chain or with the inactivated mutant H-chain. Under acidic conditions (pH 5.5, 1000 Fe atoms per molecule) the heteropolymers with about 20% H and 80% L-chains incorporated three to fourfold more iron into soluble 24-mers than the homopolymers. The data indicate that ferritins with more than 18 L-chains per molecule have the capacity to lower non-specific iron hydrolysis in bulk solution. This property is possibly due to a specific attraction of the incoming oxidized iron into the cavity and may be related to an effect of the L-chain on the cavity microenvironment. It is concluded that under high iron increments the ferritins with high L:H-chain ratios are the most efficient in incorporating iron, and this goes some way to explain why iron storage tissues contain L-rich isoferritins.

Identification of the EPR-active iron-nitrosyl complexes in mammalian ferritins

This study was undertaken to characterize the nitric oxide complexes of mammalian ferritin and their EPR properties to gain a better understanding of the interaction of NO with non-heme iron proteins within the cell. Measurements were made with horse spleen apo- and holoferritins, with chemically modified proteins, and with recombinant human H-chain apoferritin and its site-directed mutants. Three types of EPR signals (A, B, and C) have been identified and attributed to iron-nitrosyl complexes at imidazole groups of histidine, thiol groups of cysteine, and carboxylate groups of aspartate and glutamate, respectively. The C-type axial spectrum has features at g perpendicular' = 4 and g parallel' = 2 characteristic of a paramagnetic Fe(3+)-NO- complex with total spin S = 3/2 and probably arises from nonspecific binding to carboxylate groups on the protein. The S = 1/2 axial B-type signal g perpendicular' = 2.033 and g parallel' = 2.014) is formed at Cys-130 (human H-chain sequence numbering). His-128 and possibly His-118 are sites of formation of the rhombic S = 1/2 A-type complex (gx' = 2.055, gy' = 2.033, and gz' = 2.015); the former residue perhaps plays a role in the conformational stability of the protein as well as in iron binding. The data reveal that the residues Cys-130 and His-128 in the vicinity of 3-fold channels leading to the interior of the protein shell are important in iron-nitrosyl complex formation in mammalian ferritins.

Human ferritin H-chains can be obtained in non-assembled stable forms which have ferroxidase activity

We found conditions to obtain the Leu-169-->Arg mutant of human ferritin H chain in a stable and non-assembled state. The protein obtained is an oligomer of subunits with a high degree of structured conformation, and when concentrated it re-assembles into ferritin cages. Functional studies showed that (i) it promotes iron oxidation like the assembled ferritin, but at slower rate, (ii) it is readily precipitated by the oxidised iron unless apotransferrin or L-chain ferritin are added to sequester Fe(III). The results confirm that ferroxidase activity is located within the H-chain, and indicate that the cages of the fully assembled ferritins are important not only in maintaining iron in a soluble form, but also in eliciting the activity of the ferroxidase centres.

Ferroxidase kinetics of human liver apoferritin, recombinant H-chain apoferritin, and site-directed mutants

A detailed study of the kinetics of iron(II) oxidation by molecular oxygen in natural and recombinant human apoferritins has been carried out using electrode oximetry to better understand the ferroxidase activity of the protein shell. A comparative study of recombinant L-chain ferritin (rLF), recombinant H-chain ferritin (rHF), and variants has shown that (1) rLF lacks a ferroxidase activity, confirming the results of previous studies; (2) the ferroxidase site of rHF involves Glu-62 and His-65, presumably as Fe2+ ligands, since mutation of these residues abolishes most of the oxidase activity, in agreement with previous studies; and (3) mutation of both the putative ferroxidase and nucleation site ligands in rHF renders the protein totally incapable of catalyzing the oxidation of Fe2+ whereas mutation of nucleation site ligands alone (Glu-61, Glu-64, and Glu-67) decreases the activity only slightly. Analysis of the kinetics of rHF and natural human liver ferritin (HLF) (4% H-chain, 96% L-chain) gave the following apparent parameters at pH 7: Km,O2 = 6 +/- 2 microM, Km,Fe = 80 +/- 10 microM, and kcat = 201 +/- 14 min-1 for rHF and Km,O2 = 60 +/- 12 microM, Km,Fe = 50 +/- 10 microM, and kcat = 31.2 +/- 0.6 min-1 for HLF. Furthermore, Zn2+ was shown to be a noncompetitive inhibitor of Fe2+ oxidation in rHF but a mixed inhibitor in HLF. These different forms of Zn2+ inhibition in the two proteins and the higher activity of HLF than expected, based on its H-chain composition as well as differences in their enzyme kinetic parameters, suggest that H- and L-chains cooperate in modulating the ferroxidase activity of the apoferritin even though the L-subunit lacks a ferroxidase site itself.

Production and characterization of recombinant heteropolymers of human ferritin H and L chains

Vertebrate ferritins are iron storage proteins composed by 24 subunits of one or more types. The recombinant homopolymers of human ferritin H- and L-type chains differ in iron uptake and in physical stability, but the properties of heteropolymers with various proportions of H- and L-type chains cannot be predicted. Present study shows that unfolded human ferritin H- and L- type chains renature under similar conditions to form homopolymers indistinguishable from the native ones and that, when mixed, the unfolded H and L chains renature to form heteropolymers with restricted heterogeneity and with the expected H:L ratios. Seven of these ferritins with different H:L ratios were analyzed; electrophoretic mobility, immunological reactivity, and stability to guanidine denaturation varied as predicted, based on the homopolymers. In contrast, the rate of iron uptake, monitored by the variation of absorbance at 310 nm, increased in the ferritins that ranged in H chain content from 0 to 35%; further increments in H chains had no additional effect. This finding indicates that, under the present conditions, only a limited number of H chains are needed for the maximum rate of ferritin iron uptake. Variations of L- and H-type chains in vivo may thus have biological relevance.

Evidence of H- and L-chains have co-operative roles in the iron-uptake mechanism of human ferritin

The ability to incorporate iron in vitro was studied in homopolymers of human ferritin L-chain, human ferritin H-chain and its variants and in homopolymer mixtures. The H-chain variants carried amino acid substitutions in the ferroxidase centre and/or in carboxy residues on the cavity surface. Iron incorporation was examined by gel electrophoresis of the reaction products by staining for iron and protein. It was found that inactivation of the ferroxidase centre combined with the substitution of four carboxy groups on the cavity abolished the ability of H-chain ferritin to incorporate iron. Competition experiments with limited amounts of iron showed that, at neutral pH, L-chain ferritin is more efficient in forming iron cores than the H-chain variants altered at the ferroxidase activity or in the cavity. Competition experiments at pH 5.5 demonstrated that L-chain apoferritin is able to incorporate iron only when in the presence of H-chain variants with ferroxidase activity. The results indicate that L-chain apoferritin has a higher capacity than the H-chain apoferritin to induce iron-core nucleation, whereas H-chain ferritin is superior in promoting Fe(II) oxidation. The finding of cooperative roles of the H- and L-chains in ferritin iron uptake provides a clue to understanding the biological function of isoferritins.

Structure, function, and evolution of ferritins

The ferritins of animals and plants and the bacterioferritins (BFRs) have a common iron-storage function in spite of differences in cytological location and biosynthetic regulation. The plant ferritins and BFRs are more similar to the H chains of mammals than to mammalian L chains, with respect to primary structure and conservation of ferroxidase center residues. Hence they probably arose from a common H-type ancestor. The recent discovery in E. coli of a second type of iron-storage protein (FTN) resembling ferritin H chains raises the question of what the relative roles of these two proteins are in this organism. Mammalian L ferritins lack ferroxidase centers and form a distinct group. Comparison of the three-dimensional structures of mammalian and invertebrate ferritins, as well as computer modeling of plant ferritins and of BFR, indicate a well conserved molecular framework. The characterisation of numerous ferritin homopolymer variants has allowed the identification of some of the residues involved in iron uptake and an investigation of some of the functional differences between mammalian H and L chains.

Functional roles of the ferritin receptors of human liver, hepatoma, lymphoid and erythroid cells

Ferritin receptors are present on the membranes of many normal and malignant cells. The binding specificity of these receptors for H and L subunits was examined using recombinant human ferritin homopolymers. At least two different types of ferritin receptors were found, one derived from normal rat, pig, and human liver which shows similar binding of H- and L-ferritin. The second receptor type, specific for the H-chain ferritin, has been identified on membranes of hepatic and other transformed cells, and of normal lymphoblasts and erythroid precursors. These two receptor types may have different metabolic functions: the hepatic receptor acting as a scavenger for circulating ferritin and possibly for iron exchange between hepatocytes and macrophages; the H-ferritin receptor having a regulatory role which is not directly related to iron metabolism. The expression of the H-ferritin receptor is closely related to the activation and proliferation state of the cells. Addition of H-ferritin to the culture medium of cells expressing the H-ferritin receptor resulted in inhibition of cell proliferation and of colony formation.

Effect of cell proliferation on H-ferritin receptor expression in human T lymphoid (MOLT-4) cells

We have previously demonstrated the presence of receptors specific for human recombinant H ferritin (HrHF) on human T lymphoid cells (MOLT-4), and changes in receptor number and binding affinity with growth and cell cycling have now been examined. Specific binding of HrHF was maximal in MOLT-4 cells harvested during exponential growth with the cells in the DNA synthesis phase of the cell cycle. Specific binding decreased progressively to the plateau phase of growth with the cells in the resting phase of the cell cycle. Scatchard analysis of the competitive binding data for HrHF demonstrated that this decrease in binding was associated with a reduction in receptor number, from 42,140 per cell to 10,306 per cell. Receptor binding affinity increased only minimally over this period, from 7.1 x 10(7) L/mol to 14.9 x 10(7) L/mol. These results indicate that growth- and cell cycle-induced changes in H-ferritin receptor expression are primarily associated with changes in receptor number rather than receptor binding affinity. The present study demonstrates that the expression of this receptor is associated with the proliferative status of the cell and suggests that the H-ferritin receptor may mediate the putative regulatory role of H-ferritin.

Evidence that a salt bridge in the light chain contributes to the physical stability difference between heavy and light human ferritins

Human ferritin, a multimeric iron storage protein, is composed by various proportions of two subunit types: the H- and L-chains. The biological functions of these two genic products have not been clarified, although differences in reactivity with iron have been shown. Starting from the hypothesis that the high stability typical of ferritin is an important property which may be relevant for its iron storage function, we studied ferritin homopolymers of H- and L-chains in different denaturing conditions. In addition we analyzed 13 H-chain variants with alterations in regions conserved within mammalian H-chains. In all the denaturation experiments H-chain ferritin showed lower stability than L-chain ferritin. The difference was greater in guanidine HCl denaturation experiments, where the end products are fully unfolded peptides, than in acidic denaturation experiments, where the end products are peptides with properties analogous to "molten globule." The study on H-chain variants showed: (i) ferritin stability was not affected by alterations of regions exposed to the inner or outer surface of the shell and not involved in intra- or inter-chain interactions; (ii) stability was reduced by alterations of sequences involved in inter-subunit interactions such as the deletion of the N-terminal extension or substitutions along the hydrophobic and hydrophilic channels; (iii) stability was increased by the substitution of 2 amino acids inside the four-helix bundle with those of the homologous L-chain. One of the residues is involved in a salt bridge in the L-chain, and we concluded that the stability difference between H- and L-ferritins is to a large extent due to the stabilizing effect of this salt bridge on the L-subunit fold.

Binding and suppressive activity of human recombinant ferritins on erythroid cells

We studied the relation between ferritin cellular binding and suppressive activity of recombinant H- and L-ferritin on human erythroid cells at different proliferation/differentiation phases. L-ferritin failed to show any suppressive activity or detectable binding to erythroblasts at any stage of maturation. In contrast, H-ferritin demonstrated binding to erythroblasts derived from peripheral BFU-E cells which increased steadily between 7-14 days of culture up to 15,000 molecules per cell. Reticulocytes and erythrocytes failed to bind either L- or H-ferritin. H-ferritin suppressed BFU-E colony formation and reduced K562 cell proliferation at nanomolar concentrations. This suggests that the expression of H-ferritin binding sites is modulated by cellular proliferation and differentiation, that cells expressing H-ferritin binding sites are sensitive to ferritin suppressive activity and that a causal relation exists between ferritin cellular binding and suppressive activity.

Characterization of the ferritin receptors of human T lymphoid (MOLT-4) cells

We have previously demonstrated that distinct binding sites exist for human recombinant H ferritin (HrHF) and human liver ferritin (HLF) on human T lymphoid cells (MOLT-4). This study demonstrates that these binding sites have the characteristics of receptors specific for HrHF, and the binding characteristics and internalization of HrHF to MOLT-4 cells have now been examined. Iodinated HrHF was displaced by an excess of unlabeled HrHF. Heavy ferritin was the major subunit bound with only a small amount of light-ferritin binding, consistent with our immunofluorescence studies. Scatchard plot analysis of the competitive binding data for HrHF revealed an association constant of 6.3 to 6.7 x 10(7) L/mol with approximately 6000 to 15,000 receptor sites per MOLT-4 cell. Internalization of HrHF was demonstrated with pronase. Chloroquine substantially reduced the uptake of HrHF. Release of internalized HrHF was not observed when cells were rewarmed to 37 degrees C. These results indicate that HrHF is internalized by a mechanism consistent with receptor-mediated endocytosis, with possible involvement of the lysosome. The internalized HrHF remains associated with the cell. Although lymphoid cell growth and differentiation were not examined in this study, the presence of the demonstrated receptors may indicate a regulatory role for heavy ferritin in such cells.

Influence of site-directed modifications on the formation of iron cores in ferritin

The structure and crystal chemical properties of iron cores of reconstituted recombinant human ferritins and their site-directed variants have been studied by transmission electron microscopy and electron diffraction. The kinetics of Fe uptake have been compared spectrophotometrically. Recombinant L and H-chain ferritins, and recombinant H-chain variants incorporating modifications in the threefold (Asp131----His or Glu134----Ala) and fourfold (Leu169----Arg) channels, at the partially buried ferroxidase sites (Glu62,His65----Lys,Gly), a putative nucleation site on the inner surface (Glu61,Glu64,Glu67----Ala), and both the ferroxidase and nucleation sites (Glu62,His65----Lys,Gly and Glu61,Glu64,Glu67----Ala), were investigated. An additional H-chain variant, incorporating substitution of the last ten C-terminal residues for those of the L-chain protein, was also studied. Most of the proteins assimilated iron to give discrete electron-dense cores of the Fe(III) hydrated oxide, ferrihydrite (Fe2O3.nH2O). No differences were observed for variants modified in the three- or fourfold channels compared with the unmodified H-chain ferritin. The recombinant L-chain ferritin and H-chain variant depleted of the ferroxidase site, however, showed markedly reduced uptake kinetics and comprised cores of increased diameter and regularity. Depletion of the inner surface Glu residues, whilst maintaining the ferroxidase site, resulted in a partially reduced rate of Fe uptake and iron cores of wider particle size distribution. Modification of both ferroxidase and inner surface Glu residues resulted in complete inhibition of iron uptake and deposition. No cores were observed by electron microscopy although negative staining showed that the protein shell was intact. The general requirement of an appropriate spatial charge density across the cavity surface rather than specific amino acid residues could explain how, in spite of an almost complete lack of identity between the amino acid sequences of bacterioferritin and mammalian ferritins, ferrihydrite is deposited within the cavity of both proteins under similar reconstitution conditions.

Macrophage inflammatory protein (MIP)-1 beta abrogates the capacity of MIP-1 alpha to suppress myeloid progenitor cell growth

The effects of recombinant murine macrophage inflammatory protein (MIP)-1 beta and MIP-2 on the suppressive activity of MIP-1 alpha were tested using colony formation by human and murine bone marrow burst-forming unit-erythroid (BFU-E), colony-forming unit-granulocyte erythroid macrophage, megakaryocyte (CFU-GEMM), and colony-forming unit-granulocyte macrophage (CFU-GM) progenitor cells. MIP-1 beta, but not MIP-2, when added with MIP-1 alpha to cells, blocked the suppressive effects of MIP-1 alpha on both human and murine BFU-E, CFU-GEMM, and CFU-GM colony formation. Similar results were observed regardless of the early acting cytokines used: human rGM-CSF plus human rIL-3, and two recently described potent cytokines, a genetically engineered human rGM-CSF/IL-3 fusion protein and MGF, a c-kit ligand. The more potent the stimuli, the greater the suppressive activity noted. Pulse treatment of hu bone marrow cells with MIP-1 alpha at 4 degrees C for 1 h was as effective in inhibiting colony formation as continuous exposure of cells to MIP-1 alpha, and the pulsing effect with MIP-1 alpha could not be overcome by subsequent exposure of cells to MIP-1 beta. Also, pulse exposure of cells to MIP-1 beta blocked the activity of subsequently added MIP-1 alpha. For specificity, the action of a nonrelated myelosuppressive factor H-ferritin, was compared. MIP-1 alpha and H-ferritin were shown to act on similar target populations of early BFU-E, CFU-GEMM, and CFU-GM. MIP-1 beta did not block the suppressive activity of H-ferritin. Also, hemin and an inactive recombinant human H-ferritin mutein counteracted the suppressive effects of the wildtype H-ferritin molecule, but did not block the suppressive effects of MIP-1 alpha. These results show that MIP-1 beta's ability to block the action of MIP-1 alpha is specific. In addition, the results suggest that MIP-1 alpha and MIP-beta can, through rapid action, modulate early myeloid progenitor cell proliferation.

Macrophage inflammatory protein (MIP)-1 beta abrogates the capacity of MIP-1 alpha to suppress myeloid progenitor cell growth

The effects of recombinant murine macrophage inflammatory protein (MIP)-1 beta and MIP-2 on the suppressive activity of MIP-1 alpha were tested using colony formation by human and murine bone marrow burst-forming unit-erythroid (BFU-E), colony-forming unit-granulocyte erythroid macrophage, megakaryocyte (CFU-GEMM), and colony-forming unit-granulocyte macrophage (CFU-GM) progenitor cells. MIP-1 beta, but not MIP-2, when added with MIP-1 alpha to cells, blocked the suppressive effects of MIP-1 alpha on both human and murine BFU-E, CFU-GEMM, and CFU-GM colony formation. Similar results were observed regardless of the early acting cytokines used: human rGM-CSF plus human rIL-3, and two recently described potent cytokines, a genetically engineered human rGM-CSF/IL-3 fusion protein and MGF, a c-kit ligand. The more potent the stimuli, the greater the suppressive activity noted. Pulse treatment of hu bone marrow cells with MIP-1 alpha at 4 degrees C for 1 h was as effective in inhibiting colony formation as continuous exposure of cells to MIP-1 alpha, and the pulsing effect with MIP-1 alpha could not be overcome by subsequent exposure of cells to MIP-1 beta. Also, pulse exposure of cells to MIP-1 beta blocked the activity of subsequently added MIP-1 alpha. For specificity, the action of a nonrelated myelosuppressive factor H-ferritin, was compared. MIP-1 alpha and H-ferritin were shown to act on similar target populations of early BFU-E, CFU-GEMM, and CFU-GM. MIP-1 beta did not block the suppressive activity of H-ferritin. Also, hemin and an inactive recombinant human H-ferritin mutein counteracted the suppressive effects of the wildtype H-ferritin molecule, but did not block the suppressive effects of MIP-1 alpha. These results show that MIP-1 beta's ability to block the action of MIP-1 alpha is specific. In addition, the results suggest that MIP-1 alpha and MIP-beta can, through rapid action, modulate early myeloid progenitor cell proliferation.

Specific binding sites for H-ferritin on human lymphocytes: modulation during cellular proliferation and potential implication in cell growth control

Interactions between human recombinant H- and L-ferritins and human lymphocytes were studied in vitro by direct binding assays and by flow cytometry. L-ferritin did not cause detectable specific binding, whereas H-ferritin showed a specific and saturable binding that increased markedly in phytohemagglutinin (PHA)-stimulated cells. This ferritin bound up to 30% of CD4+ and CD8+ T-lymphocytes and most B cells, indicating that expression of ferritin binding sites is not related to cell lineage or function. Dual-color flow cytometry experiments showed that ferritin binding sites were present on cells expressing the proliferation markers HLA-DR, MLR3, interleukin 2 (IL-2), and transferrin receptors (Tf-R). In addition, after PHA induction, the time course of the expression of H-ferritin binding sites was similar to those of the above proliferation markers. Ferritin binding sites were observed in lymphocytes at all cell cycle phases, including the early S-phase. H-Ferritin at nanomolar and picomolar concentrations had an inhibitory effect on PHA-induced blastogenesis. We propose that H-ferritin binding sites behave like proliferation markers, with the unusual function of downregulating proliferation.

Iron detoxifying activity of ferritin. Effects of H and L human apoferritins on lipid peroxidation in vitro

Three recombinant human apoferritin variants were added to ferrous iron and the amount of lipid peroxidation produced by hydrogen peroxide was studied. The H-apoferritin had the strongest inhibitory effect on lipid peroxidation, probably due to its ferroxidase activity. The L-apoferritin inhibited lipid peroxidation slowly and only at neutral pH. The H-mutant 91, deleted of the last 22 C-terminal amino acids, and which is not able to form an iron core, had minimal effects on iron lipid peroxidation. It was concluded that both ferro-oxidase and iron mineralization activities are necessary for ferritin iron detoxifying action.

Immunocytochemical detection of ferritin in human bone marrow and peripheral blood cells using monoclonal antibodies specific for the H and L subunit

We have used the monoclonal antibodies 2A4 (specific for the H subunit of human ferritin) and LO3 (specific for the L subunit) for immunocytochemical detection of ferritin in bone marrow and peripheral blood cells from normal subjects and patients with various haematological disorders. Formalin-fixed slides were stained by the immunoalkaline phosphatase procedure (APAAP). In normal subjects, ferritin could be found only in bone marrow smears and appeared to be largely confined to erythroid precursors and reticuloendothelial cells. The more immature erythroid precursors contained higher concentrations of cellular ferritin. Although evaluation could be only semiquantitative, erythroblast ferritin appeared to be more reactive with the monoclonal 2A4 (15 +/- 7% positive erythroblasts) than with the monoclonal LO3 (6 +/- 5% positive erythroblasts), indicating that H-type ferritin was predominant, particularly in proerythroblasts and basophilic erythroblasts. By contrast, the ferritin present in reticuloendothelial cells appeared to be predominantly of L-type. Patients with iron deficiency showed low levels of positive erythroblast, whereas the reverse was true in patients with transfusional iron overload. Intense positivity for reticuloendothelial cell ferritin was found in patients with anaemia of chronic disease. In myelodysplastic syndromes and acute myeloid leukaemia (AML), ferritin positivity was generally very strong at any stage of erythroblast development, particularly with the monoclonal antibody 2A4. Perls-positive perinuclear granules of ring sideroblasts were not stained, confirming that mitochondrial iron deposition is not in the form of ferritin. In AML and myelodysplastic syndromes with excess of blasts, ferritin could be detected also in immature myeloid cells. These data indicate that: (a) in normal conditions ferritin is mainly expressed in red cell precursors and reticuloendothelial cells, and this is in keeping with the peculiar role of these cells in iron metabolism; (b) abnormal cell ferritin contents can be observed in both iron overload and malignancy.