Brain iron homeostasis.

The anatomical and cellular distribution of non-haem iron, ferritin, transferrin, and the transferrin receptor have been studied in postmortem human brain and these studies, together with data on the uptake and transport of labeled iron, by the rat brain, have been used to elucidate the role of iron and other metal ions in certain neurological disorders. High levels of non-haem iron, mainly in the form of ferritin, are found in the extrapyramidal system, associated predominantly with glial cells. In contrast to non-haem iron, the density of transferrin receptors is highest in cortical and brainstem structures and appears to relate to the iron requirement of neurones for mitochondrial respiratory activity. Transferrin is synthesized within the brain by oligodendrocytes and the choroid plexus, and is present in neurones, consistent with receptor mediated uptake. The uptake of iron into the brain appears to be by a two-stage process involving initial deposition of iron in the brain capillary endothelium by serum transferrin, and subsequent transfer of iron to brain-derived transferrin and transport within the brain to sites with a high transferrin receptor density. A second, as yet unidentified mechanism, may be involved in the transfer of iron from neurones possessing transferrin receptors to sites of storage in glial cells in the extrapyramidal system. The distribution of iron and the transferrin receptor may be of relevance to iron-induced free radical formation and selective neuronal vulnerability in neurodegenerative disorders.     

DISTRIBUTION OF SERUM PROTEINS AND BETA-TRACE PROTEIN WITHIN THE NERVOUS SYSTEM

Abstract— The concentration of beta-trace protein, a low molecular weight water-soluble protein, was significantly higher in cerebral and cerebellar white matter than in grey matter. A similar distribution was found for transferrin. The concentrations of gamma-trace protein and pre-albumin were almost constant in cerebral and cerebellar white and grey matter. A different distribution was shown for albumin, beta_lc/beta_lA globulins, and the immunoglobulins G, A and M, with the highest concentrations mostly encountered in the highly vascularized basal ganglia and grey matter, and the lowest concentrations in white matter. Thus, different parameters, hitherto unknown determine the distribution within the central nervous system of various proteins-those which originate from serum, and beta-trace protein which originates mainly from the central nervous system.
The amounts of the different proteins were higher in the choroid plexus than in brain tissue, with the exception of gamma-trace protein.
Foetal brains contained increasing concentrations of beta-trace protein and of transferrin with age.
Femoral nerve contained lower concentrations of beta-trace protein and gamma-trace protein, and higher concentrations of the other proteins, than the central nervous system.     

Characterization, Distribution, and Protein Kinase C-Mediated Regulation of [35S]Glutathione Binding Sites in Mouse and Human Spinal Cord

Abstract: We have characterized a high-affinity [³⁵S]-glutathione ([³⁵S]GSH) binding site in mouse and human spinal cord. [³⁵S]GSH binding sites in mouse and human spinal cord were observed largely within the gray matter in both the dorsal and ventral horns of spinal cord at cervical, thoracic, and lumbosacral segments. High-affinity [³⁵S]GSH binding was saturable, showing a B _max of 72 fmol/mg of protein and a K _D of 3.0 n M for mouse spinal cord and a B _max of 52 fmol/mg of protein and a K _D of 1.6 n M for human spinal cord. [³⁵S]GSH binding was displaceable by GSH, l -cysteine, and S -hexyl-GSH, but not by glutamate, glycine, or NMDA. These [³⁵S]GSH binding sites exhibited kinetic and saturation characteristics similar to GSH binding sites in rat brain astrocytes. To determine whether [³⁵S]GSH binding sites could be regulated by protein kinase C, we exposed human spinal cord sections to phorbol 12,13-diacetate for 1 h before ligand binding. Phorbol ester treatment increased [³⁵S]GSH binding by ∼60%, an effect that could be blocked by exposure of spinal cord sections to 1-(5-isoquinolinylsulfonyl)-2-methylpiperazine, a general protein kinase inhibitor. [³⁵S]GSH binding sites in the spinal cord of both species exhibited many of the characteristics of a receptor including saturable binding, high affinity, ligand specificity, and modulation by kinase activity. These data suggest that GSH is a neurotransmitter in the CNS.     

Zinc Inhibition of t-[3H]Butylbicycloorthobenzoate Binding to the GABAA Receptor Complex

Abstract: The effect of Zn²⁺ on t -[³H]butylbicycloorthobenzoate ([³H]TBOB) binding to the GABA_A receptor complex was studied autoradiographically in rat brain. Zn²⁺ inhibited [³H]TBOB binding in a dose-dependent manner at physiological concentrations. Saturation analysis revealed noncompetitive inhibition in various brain regions. The inhibitory effect of Zn²⁺ had regional heterogeneity; regions showing the greatest inhibition of [³H]TBOB binding were cortical laminae I–III, most areas of hippocampus, striatum, septum, and cerebellar cortex. Regions with relatively less inhibition of [³H]TBOB binding included cortical laminae V–VI, thalamus, superior colliculus, inferior colliculus, and central gray matter. The effect of Zn²⁺ and those of other GABA_A ligands, such as benzodiazepines, bicuculline, isoguvacine, and picrotoxin, on [³H]TBOB binding seemed to be additive. Ni²⁺, Cd²⁺, and Cu²⁺ also inhibited [³H]TBOB binding with a regional heterogeneity similar to that produced by Zn²⁺. These results are consistent with Zn²⁺ acting at the previously detected recognition site on the GABA_A receptor complex, distinct from the picrotoxin, GABA, and benzodiazepine sites. The regional heterogeneity of the Zn²⁺ effect may reflect differential regional distribution of GABA_A receptor subtypes among brain regions. Other divalent cations probably act at the Zn²⁺ binding site.     

Botulinum A Like Type B and Tetanus Toxins Fulfils Criteria for Being a Zinc-Dependent Protease

Abstract: Iron is abnormally accumulated in the substantia nigra pars compacta of patients with Parkinson's disease (PD). Because neuronal and glial iron uptake seems to be mediated by the binding of ferrotransferrin to a specific high-affinity receptor on the cell surface, the number of transferrin receptors could be altered in this disease. The regional distribution of specific binding sites for human ¹²⁵I-diferric transferrin has been studied in the mesencephalon, on cryostat-cut sections from autopsy brains of control subjects and parkinsonian patients by in vitro autoradiography. Densities of binding sites were highest in the central gray substance (˜10 fmol/mg of tissue equivalent), intermediate in the catecholaminergic cell group A8, superior colliculus, and ventral tegmental area, and almost nonexistent in the substantia nigra. The density of ¹²⁵I-transferrin binding sites was not significantly different between parkinsonian and control brains in any region analyzed. These results show that in the mesencephalon the regional density of transferrin binding sites is lowest in the dopaminergic cell groups, which are the most vulnerable to PD, and suggest that iron does not accumulate through an increased density of transferrin receptors at the level of the substantia nigra.     

Immunolabeling of Central Serotonin 5-HT1Dβ Receptors in the Rat, Mouse, and Guinea Pig with a Specific Anti-Peptide Antiserum

Abstract: A synthetic peptide (25 amino acids) corresponding to a specific portion of the third intracytoplasmic loop of the rat serotonin 5-HT_1B/1Dβ receptor was coupled to keyhole limpet hemocyanin and injected monthly into rabbits. Anti-peptide antibodies were detected by enzyme-linked immunosorbent assay and characterized by immunoprecipitation of the 5-HT_1B/1Dβ receptor in CHAPS-solubilized extracts from rat striatal membranes. Up to 60% of solubilized striatal serotonin- O -carboxymethylglycyl[¹²⁵I]iodotyrosinamide ([¹²⁵I]GTI; a selective 5-HT_1B/1D radioligand) binding sites were immunoprecipitated and subsequently pharmacologically identified as 5-HT_1B receptors. The remaining 40% of [¹²⁵I]GTI binding sites were shown to be 5-HT_1D receptors. In addition, these antibodies were successfully used in immunofluorescence experiments to detect the 5-HT_1B/1Dβ, but not the 5-HT_1D/1Dα, receptor in transiently transfected LLC-PK1 cells. Immunoautoradiographic experiments performed with brain sections from the rat, mouse, and guinea pig showed that the substantia nigra and globus pallidus contained the highest densities of 5-HT_1Dβ receptor-like immunoreactivity. Comparison of the regional distribution of immunolabeling with that of the specific binding of [¹²⁵I]GTI in the brain of these species further confirmed that the anti-peptide antibodies selectively recognized only the 5-HT_1Dβ component of [¹²⁵I]GTI specific receptor binding sites.     

Identification of Endothelin Receptor Subtype (ETB) in Human Cerebral Cortex Using Subtype-Selective Ligands

Abstract: Specific endothelin (ET) binding sites were characterized in membranes prepared from human cerebral cortices using binding assay and cross-linking analysis. The presence of immunoreactive (IR) ET-1 was studied by radioimmunoassay. Saturation binding experiments revealed that the K _D and B _max for ¹²⁵I-ET-1 and ¹²⁵l-ET-3 to membranes from gray matter were 25 ± 6 pM and 115 ± 15 fmol/mg of protein and 24 ± 5 p M and 108 ± 13 fmol/mg of protein, respectively. Similar results were obtained for white matter. In the presence of 10 n M sarafotoxin-6c, which is selective for ET_B receptors, ¹²⁵I-ET-1 and ¹²⁵l-ET-3 binding was totally abolished. However, in the presence of 1 μ M BQ123, which is selective for ET_Areceptors, both bindings were not affected. These results suggest that the human cerebral cortex contains only ET_Breceptors. Cross-linking of ¹²⁵I-ET-1 and ¹²⁵l-ET-3 to membranes with disuccinimidyl suberate resulted in the labeling of two bands of 48 and 31 kDa. Concentrations of IR-ET-1 in the gray and white matter were 7.0 ± 3.2 and 2.5 ± 1.7 fmol/g wet weight, respectively. The demonstration of high-affinity ET_B receptors and the presence of IRET-1 suggest that the peptide may act as a neurotransmitter or neuromodulator in the human cerebral cortex.     

Tim‐2 is the receptor for H‐ferritin on oligodendrocytes

Oligodendrocytes stain more strongly for iron than any other cell in the CNS, and they require iron for the production of myelin. For most cell types transferrin is the major iron delivery protein, yet neither transferrin receptor protein nor mRNA are detectable in mature oligodendrocytes. Thus an alternative iron delivery mechanism must exist. Given the significant long term consequences of developmental iron deficiency and the iron requirements for normal myelination, identification of the iron delivery mechanism for oligodendrocytes is important. Previously we have reported that oligodendrocytes bind H‐ferritin and that H‐ferritin binds to white matter tracts in vivo. Recently, T cell immunoglobulin and mucin domain‐containing protein‐2 (Tim‐2) was shown to bind and internalize H‐ferritin. In the present study we show that Tim‐2 is expressed on oligodendrocytes both in vivo and in vitro. Further, the onset of saturable H‐ferritin binding in CG4 oligodendrocyte cell line is accompanied by Tim‐2 expression. Application of a blocking antibody to the extracellular domain of Tim‐2 significantly reduces H‐ferritin binding to the differentiated CG4 cells and primary oligodendrocytes. Tim‐2 expression on CG4 cells is responsive to iron; decreasing with iron loading and increasing with iron chelation. Taken together, these data provide compelling evidence that Tim‐2 is the H‐ferritin receptor on oligodendrocytes suggesting it is the primary mechanism for iron acquisition by these cells.     

Pharmacological and Regional Characterization of [3H]LY278584 Binding Sites in Human Brain

Abstract: Binding of [³H]LY278584, which has been previously shown to label 5-hydroxytryptamine₃ (5-HT₃) receptors in rat cortex, was studied in human brain. Saturation experiments revealed a homogeneous population of saturable binding sites in amygdala ( K _D= 3.08 ± 0.67 n M, B _max= 11.86 ± 1.87 fmol/mg of protein) as well as in hippocampus, caudate, and putamen. Specific binding was also high in nucleus accumbens and entorhinal cortex. Specific binding was negligible in neocortical areas. Kinetic studies conducted in human hippocampus revealed a K _on of 0.025 ± 0.009 n M ⁻¹ min⁻¹ and a K _off of 0.010 ± 0.002 min⁻¹. The kinetics of [³H]LY278584 binding were similar in the caudate. Pharmacological characterization of [³H]LY278584 specific binding in caudate and amygdala indicated the compound was binding to 5-HT₃ receptors. We conclude that 5-HT₃ receptors labeled by [³H]LY278584 are present in both limbic and striatal areas in human brain, suggesting that 5-HT₃ receptor antagonists may be able to influence the dopamine system in humans, similarly to their effects in rodent studies.     

Direct observation of the iron binding sites in a ferritin

X-Ray analysis of the ferritin of Escherichia coli (Ec-FTN) and of Ec-FTN crystals soaked in (NH₄)₂Fe(SO₄)₂ has revealed the presence of three iron-binding sites per subunit. Two of these form a di-iron site in the centre of the subunit as has been proposed for the ‘ferroxidase centres’ of human ferritin H chains. This di-iron site, lying within the 4-alpha-helix bundle, resemble those of ribonucleotide reductase, methane monoxygenase and haemerythrin. The third iron is bound by ligands unique to Ec-FTN on the inner surface of the protein shell. It is speculated that this state may represent the nucleation centre of a novel type of Fe(III) cluster, recently observed in Ec-FTN.     

Binding Characteristics of a Potent AMPA Receptor Antagonist [3H]Ro 48-8587 in Rat Brain

Abstract: A new AMPA receptor antagonist, Ro 48-8587, was characterized pharmacologically in vitro. It is highly potent and selective for AMPA receptors as shown by its effects on [³H]AMPA, [³H]kainate, and [³H]MK-801 binding to rat brain membranes and on AMPA- or NMDA-induced depolarization in rat cortical wedges. [³H]Ro 48-8587 bound with a high affinity ( K _D = 3 n M ) to a single population of binding sites with a B _max of 1 pmol/mg of protein in rat whole brain membranes. [³H]Ro 48-8587 binding to rat whole brain membranes was inhibited by several compounds with the following rank order of potency: Ro 48-8587 > 6-nitro-7-sulphamoylbenzo[ f ]quinoxaline-2,3-dione (NBQX) > YM 90K > 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) > quisqualate > AMPA > glutamate > kainate > NMDA. The distribution and abundance of specific binding sites (∼95% of total) in sections of rat CNS, revealed by quantitative receptor radioautography and image analysis, indicated a very discrete localization. Highest binding values were observed in cortical layers (binding in layers 1 and 2 > binding in layers 3–6), hippocampal formation, striatum, dorsal septum, reticular thalamic nucleus, cerebellar molecular layer, and spinal cord dorsal horn. At 1 n M , the values for specific binding were highest in the cortical layers 1 and 2 and lowest in the brainstem (∼2.6 and 0.4 pmol/mg of protein, respectively). Ro 48-8587 is a potent and selective AMPA receptor antagonist with improved binding characteristics (higher affinity, selectivity, and specific binding) compared with those previously reported.     

CHARACTERIZATION OF THE BINDING OF [3H]MUSCIMOL, A POTENT γ-AMINOBUTYRIC ACID AGONIST, TO RAT BRAIN SYNAPTOSOMAL MEMBRANES USING A FILTRATION ASSAY

Abstract— The binding of [³H]muscimol, a potent GABA agonist, to crude synaptic membranes prepared from rat brain was studied using a filtration method to isolate membrane-bound ligand. Specific binding was found to be saturable and occurred to two binding sites of K _d5 5 and 30 n m . Binding was Na⁺-independent and enhanced by both freezing and Triton treatment. Regional and subcellular distribution studies and pharmacological characterization of specific [³H]muscimol binding are consistent with binding to the synaptic GABA receptor.     

Guinea pig hippocampal 5-HT1E receptors: a tool for selective drug development

Recent studies have indicated that the serotonin [5-hydroxytryptamine (5-HT)] 1E receptor, originally discovered in human brain tissue, is not expressed in rat or mouse brain. Thus, there have been few reports on 5-HT_1E receptor drug development. However, expression of 5-HT_1E receptor mRNA has been shown in guinea pig brain. To establish this species as an animal model for 5-HT_1E drug development, we identified brain regions that exhibit 5-carboxyamidotryptamine, ritanserin, and LY344864 – insensitive [³H]5-HT binding (characteristic of the 5-HT_1E receptor). In hippocampal homogenates, where 5-HT_1E receptor density was sufficiently high for radioligand binding analysis, 100 nM 5-carboxyamidotryptamine, 30 nM ritanserin, and 100 nM LY344864 were used to mask [³H]5-HT binding at non-5-HT_1E receptors. The K _d of [³H]5-HT was 5.7 ± 0.7 nM and is indistinguishable from the cloned receptor K _d of 6.5 ± 0.6 nM. The affinities of 16 drugs for the cloned and hippocampal-expressed guinea pig 5-HT_1E receptors are essentially identical ( R ² = 0.97). These findings indicate that using these conditions autoradiographical distribution and signal transduction studies of the 5-HT_1E receptor in guinea pig brain are feasible. Using the guinea pig as an animal model should provide important insights into possible functions of this receptor and the therapeutic potential of selective human 5-HT_1E drugs.     

Receptor-Mediated Endocytosis of a Manganese Complex of Transferrin into Neuroblastoma (SHSY5Y) Cells in Culture

Abstract: Exposure to manganese compounds often occurs as the result of industrial production or mining. Although manganese appears in traces in animal and human tissue and is essential to certain biological processes, it is also toxic. In humans and animals, toxicity is mainly associated with the nervous system. The mechanism underlying behavioral and biochemical alterations observed after manganese intoxication is not fully understood. We have shown that the manganese present in serum after exposure to manganese oxide is bound to transferrin as trivalent manganic ion. In this study of manganese uptake and storage we used a clone of human neuroblastoma cells (SHSY5Y). These cells differentiate and express catechol-aminergic properties. Saturation binding analysis of the transferrin-manganese complex to the cells revealed a single class of binding sites, with an apparent K _D of 13 ± 1 n M and a density of 11, 000 ± 2, 000 binding sites per cell. The complex was internalized in a temperature-dependent way and reached saturation after 2 h when ∼2% of the added manganese had been internalized. About 80% of the internalized manganese was found in ferritin after 24 h of exposure. The results demonstrate that the transferrin receptor on SHSY5Y cells can bind and internalize a manganese-transferrin complex as efficiently as an iron-transferrin complex, although a saturation of the manganese uptake was achieved.     

[3H]LY303870, a Novel Nonpeptide Radioligand for the NK-1 Receptor

Abstract: We synthesized a potent and selective antagonist radioligand for the neurokinin (NK)-1 receptor and characterized its binding to guinea pig striatal membranes. ( R ) - N - [2 - [Acetyl[³H₃][(2 - methoxyphenyl) - methyl]amino] - 1 - (1 H - indol - 3 - ylmethyl)ethyl][1,4' - bipiperidine]-1'-acetamide ([³H]LY303870) binds to a single class of sites with an equilibrium K _D of 0.22 n M and a B _max of 723 fmol/mg of protein. Unlabeled LY303870 potently inhibited the binding with an IC₅₀ of 0.56 n M , whereas the less active ( S )-enantiomer (LY306155) was substantially less potent. The nonpeptide NK-1 antagonists (±)-CP96,345 and (±)-RP 67580 had IC₅₀ values of 0.74 and 49 n M , respectively. Substance P (SP) was also a potent inhibitor with with an IC₅₀ of 3.1 n M . The inhibition by SP could be separated into two components: a high-affinity component with a K _i of 0.53 n M and a lower-affinity component with a K _i of 155 n M . Addition of 100 µ M guanylyl 5'-imidodiphosphate [Gpp(NH)p] in the incubation increased the relative amount of the low-affinity agonist state of the receptor. Consistent with the antagonist properties of LY303870, the dissociation rate of [³H]LY303870 was not changed by the presence of 100 µ M Gpp(NH)p. The distribution of [³H]LY303870 binding sites in the guinea pig brain closely matched the distribution of NK-1 receptors labeled by [³H]SP. Therefore, [³H]LY303870 is a potent and selective antagonist radioligand for NK-1 receptors in guinea pig brain. In addition, regulation of NK-1 agonist affinity by guanine nucleotides is similar to that seen for monoaminergic receptors.     

GANGLIOSIDE ABNORMALITIES IN MULTIPLE SCLEROSIS

Abstract— Gangliosides were isolated from plaque tissue and normal appearing white matter of multiple sclerosis (MS) brain. All four plaques showed decreased ganglioside concn relative to normal human white matter on a wet wt basis, but significant elevation in terms of dry wt. The wet wt and dry wt concn of MS white matter gangliosides showed smaller but statistically significant decreases below normal. Thin-layer patterns of the plaques showed several departures from normal white matter, including decrease of G₄ and G₅, and complete loss of G₇ (sialosylgalactosylceramide). Most of the plaques had significant elevation of G_2A and G_3A along with increases of the slower-migrating polysialogangliosides. An additional ganglioside was present between G₂ and G_2A which was not seen in normal white matter. The TLC pattern of MS white matter gangliosides was essentially normal. The evidence for a general decrease of acidic lipids within normal appearing white matter is discussed.     

In Vitro Studies of Ferritin Iron Release and Neurotoxicity

Abstract: The increase in brain iron associated with several neurodegenerative diseases may lead to an increased production of free radicals via the Fenton reaction. Intracellular iron is usually tightly regulated, being bound by ferritin in an insoluble ferrihydrite core. The neurotoxin 6-hydroxydopamine (6-OHDA) releases iron from the ferritin core by reducing it to the ferrous form. Iron release induced by 6-OHDA and structurally related compounds and two other dopaminergic neurotoxins, 1-methyl-4-phenylpyridinium iodide (MPP⁺) and 1-trichloromethyl-1,2,3,4-tetrahydro-β-carboline (TaClo), were compared, to identify the structural characteristics important for such release. 1,2,4-Trihydroxybenzene (THB) was most effective in releasing ferritin-bound iron, followed by 6-OHDA, dopamine, catechol, and hydroquinone. Resorcinol, MPP⁺, and TaClo were ineffective. The ability to release iron was associated with a low oxidation potential. It is proposed that a low oxidation potential and an ortho -dihydroxyphenyl structure are important in the mechanism by which ferritin iron is mobilized. In the presence of ferritin, both 6-OHDA and THB strongly stimulated lipid peroxidation, an effect abolished by the addition of the iron chelator deferoxamine. These results suggest that ferritin iron release contributes to free radical-induced cell damage in vivo.     

[3H]SCH 58261, a Selective Adenosine A2A Receptor Antagonist, Is a Useful Ligand in Autoradiographic Studies

Abstract: We have characterized the new potent and selective nonxanthine adenosine A_2A receptor antagonist SCH 58261 as a new radioligand for receptor autoradiography. In autoradiographic studies using agonist radioligands for A_2A receptors ([³H]CGS 21680) or A₁ receptors ( N ⁶-[³H]cyclohexyladenosine), it was found that SCH 58261 is close to 800-fold selective for rat brain A_2A versus A₁ receptors ( K _i values of 1.2 n M versus 0.8 µ M ). Moreover, receptor autoradiography showed that [³H]SCH 58261, in concentrations below 2 n M , binds only to the dopamine-rich regions of the rat brain, with a K _D value of 1.4 (0.8–1.8) n M . The maximal number of binding sites was 310 fmol/mg of protein in the striatum. Below concentrations of 3 n M , the nonspecific binding was <15%. Three adenosine analogues displaced all specific binding of [³H]SCH 58261 with the following estimated K _i values (n M ): 2-hex-1-ynyl-5'- N -ethylcarboxamidoadenosine, 3.9 (1.8–8.4); CGS 21680, 130 (42–405); N ⁶-cyclohexyladenosine, 9,985 (3,169–31,462). The binding of low concentrations of SCH 58261 was not influenced by either GTP (100 µ M ) or Mg²⁺ (10 m M ). The present results show that in its tritium-labeled form, SCH 58261 appears to be a good radioligand for autoradiographic studies, because it does not suffer from some of the problems encountered with the currently used agonist radioligand [³H]CGS 21680.     

Secretion of Ferritin by Iron-laden Macrophages and Influence of Lipoproteins

Increasing evidence supports a role of cellular iron in the initiation and development of atherosclerosis. We and others reported earlier that iron-laden macrophages are associated with LDL oxidation, angiogenesis, nitric oxide production and apoptosis in atherosclerotic processes. Here we have further studied perturbed iron metabolism in macrophages, their interaction with lipoproteins and the origin of iron accumulation in human atheroma. In both early and advanced human atheroma lesions, hemoglobin and ferritin accumulation correlated with the macrophage-rich areas. Iron uptake into macrophages, via transferrin receptors or scavenger receptor-mediated erythrophagocytosis, increased cellular iron and accelerated ferritin synthesis at both mRNA and protein levels. The binding activity of iron regulatory proteins was enhanced by desferrioxamine (DFO) and decreased by hemin and iron compounds. Iron-laden macrophages exocytosed both iron and ferritin into the culture medium. Exposure to oxidized low-density lipoprotein (oxLDL, ≥50?μg/mL) resulted in <20% apoptosis of iron-laden human macrophages, but cells remained impermeable after a 24?h period and an increased excretion of ferritin could be observed by immunostaining techniques. Exposure to high-density lipoprotein (HDL) significantly decreased ferritin excretion from these cells. We conclude: (i) erythrophagocytosis and hemoglobin catabolism by macrophages contribute to ferritin accumulation in human atherosclerotic lesions and; (ii) iron uptake into macrophages leads to increased synthesis and secretion of ferritin; (iii) oxidized LDL and HDL have different effects on these processes.