Toxicology and Applied Pharmacology 202 (2005) 199 – 211
`
`www.elsevier.com/locate/ytaap
`
`Review
`
`Iron metabolism and toxicity
`
`G. Papanikolaoua, K. Pantopoulosb,c,*
`
`aFirst Department of Internal Medicine, National and Kapodistrian University of Athens, School of Medicine, Laikon General Hospital, Athens 11527, Greece
`bLady Davis Institute for Medical Research, Sir Mortimer B. Davis Jewish General Hospital, Montreal, Quebec, Canada H3T 1E2
`cDepartment of Medicine, McGill University, Canada
`
`Received 19 April 2004; accepted 24 June 2004
`Available online 11 August 2004
`
`Abstract
`
`Iron is an essential nutrient with limited bioavailability. When present in excess, iron poses a threat to cells and tissues, and therefore iron
`homeostasis has to be tightly controlled. Iron’s toxicity is largely based on its ability to catalyze the generation of radicals, which attack and
`damage cellular macromolecules and promote cell death and tissue injury. This is lucidly illustrated in diseases of iron overload, such as
`hereditary hemochromatosis or transfusional siderosis, where excessive iron accumulation results in tissue damage and organ failure.
`Pathological iron accumulation in the liver has also been linked to the development of hepatocellular cancer. Here we provide a background
`on the biology and toxicity of iron and the basic concepts of iron homeostasis at the cellular and systemic level. In addition, we provide an
`overview of the various disorders of iron overload, which are directly linked to iron’s toxicity. Finally, we discuss the potential role of iron in
`malignant transformation and cancer.
`D 2004 Elsevier Inc. All rights reserved.
`
`Keywords: Iron; Toxicity; Homeostasis; Heme
`
`Chemical properties and biological functions of iron
`
`Heme
`
`Iron is a component of several metaloproteins and plays a
`crucial role in vital biochemical activities, such as oxygen
`sensing and transport, electron transfer, and catalysis (Aisen
`et al., 2001). Iron is thus indispensable for
`life. The
`biological functions of iron are based on its chemical
`properties, e.g., its capacity to form a variety of coordination
`complexes with organic ligands in a dynamic and flexible
`mode, and its favorable redox potential to switch between
`the ferrous, Fe(II), and ferric, Fe(III), states (+772 mV at
`neutral pH). The bioavailability of iron is generally limited,
`because under aerobic conditions, Fe(II) is readily oxidized
`in solution to Fe(III), which is virtually insoluble at
`18 M].
`physiological pH [Kfree Fe(III) = 10
`
`* Corresponding author. Lady Davis Institute for Medical Research, Sir
`Mortimer B. Davis Jewish General Hospital, 3755 Cote-Ste-Catherine
`Road, Montreal, Quebec, Canada H3T 1E2. Fax: +1 514 340 7502.
`E-mail address: kostas.pantopoulos@mcgill.ca (K. Pantopoulos).
`
`0041-008X/$ - see front matter D 2004 Elsevier Inc. All rights reserved.
`doi:10.1016/j.taap.2004.06.021
`
`A significant fraction of cellular iron is associated with
`proteins in the form of heme, a common prosthetic group
`composed of protoporphyrin IX and a Fe(II) ion. The
`tetrapyrrol porphyrin ring is synthesized from the universal
`precursor 5-aminolevulinic acid (ALA) by a series of
`reactions in the cytosol and in mitochondria. In metazoans,
`ALA is generated by the condensation of glycine and
`succinyl-CoA, which is catalyzed by the 5-aminolevulinic
`acid synthase (ALAS) (Ryter and Tyrrell, 2000). Mammals
`express two ALAS isoforms: a housekeeping (ALAS1)
`isoform and an erythroid-specific (ALAS2) isoform (Ponka,
`1997). The insertion of Fe(II) into protoporphyrin IX,
`catalyzed by ferrochelatase in the mitochondria, defines
`the terminal step of the heme biosynthetic pathway. Heme is
`then exported to the cytosol for incorporation into hemo-
`proteins. Heme degradation is catalyzed by the microsomal
`heme oxygenase 1 (HO-1) and its homologues HO-2 and
`HO-3 (Ryter and Tyrrell, 2000). The liberated Fe(II) is
`reutilized. The reaction also generates carbon monoxide
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1066 - Page 1
`
`

`

`200
`
`G. Papanikolaou, K. Pantopoulos / Toxicology and Applied Pharmacology 202 (2005) 199–211
`
`compatible with the constrains of the cellular environment,
`is a fundamental feature for many biochemical reactions and
`renders iron to an essential mineral and nutrient. However,
`this very property turns iron into a potential biohazard,
`because under aerobic conditions, iron can readily catalyze
`the generation of noxious radicals. Iron’s toxicity is largely
`based on Fenton and Haber–Weiss chemistry (Fig. 1A),
`where catalytic amounts of iron are sufficient
`to yield
`S
`S
`) and hydro-
`) from superoxide (O2
`hydroxyl radicals (OH
`gen peroxide (H2O2), collectively known as breactive
`oxygen intermediatesQ (ROIs) (Halliwell and Gutteridge,
`1990). Importantly, ROIs are inevitable byproducts of
`aerobic respiration and emerge by incomplete reduction of
`dioxygen in mitochondria. ROIs can also be generated
`during enzymatic reactions in other subcellular compart-
`ments, such as in peroxisomes, the endoplasmic reticulum,
`or the cytoplasm.
`ROIs are also produced by the membrane-bound
`NADPH oxidase complex (Hampton et al., 1998), a
`multisubunit enzyme primarily expressed in phagocytic
`neutrophils and macrophages, but also in other cell types.
`NADPH oxidase is an important tool for the antimicrobial
`defense of the organism. The enzyme complex assembles
`upon infection and generates high levels of superoxide in a
`brespiratory burstQ, which is enzymatically and spontane-
`ously dismutated to hydrogen peroxide. The reaction
`products give rise to more potent oxidants such as
`
`
`) and hypochlorite (OCl
`), which
`peroxynitrite (ONOO
`
`(CO), which may be involved in signaling pathways, and
`biliverdin, which is further enzymatically converted to the
`antioxidant bilirubin.
`The most abundant mammalian hemoproteins, hemoglo-
`bin and myoglobin, serve as oxygen carriers in the erythroid
`tissue and in the muscle, respectively. Oxygen binding is
`mediated by the heme moieties. Other hemoproteins include
`various cytochromes and enzymes, such as oxygenases,
`peroxidases, nitric oxide (NO) synthases, or guanylate
`cyclase. The heme moiety may also function in electron
`transfer reactions (e.g., in cytochromes a, b, and c), as a
`substrate activator (e.g., in cytochrome oxidase, cytochrome
`P450, catalase) or as an NO sensor (in guanylate cyclase).
`
`Non-heme iron
`
`The most prevalent forms of non-heme iron in metaloQ
`proteins are iron–sulfur clusters, such as 2Fe–2S, 3Fe–4S,
`or 4Fe–4S (Beinert et al., 1997). They play diverse
`functional roles, ranging from electron transfer (e.g., the
`Rieske proteins in complex III of the respiratory chain),
`transcriptional regulation (the bacterial SoxR and FNR
`transcription factors), structural stabilization (bacterial
`endonuclease III) to catalysis (e.g., aconitase, an enzyme
`of the citric acid cycle). Other forms of catalytically active,
`protein-associated iron may include iron-oxo clusters (e.g.,
`in ribonucleotide reductase, an enzyme required for DNA
`synthesis) or mononuclear iron centers (e.g.,
`in cyclo-
`oxygenase and lipoxygenase, enzymes involved in inflam-
`matory responses).
`It should also be noted that non-heme iron has a central
`function in a recently discovered mechanism for oxygen
`sensing, via the hypoxia-inducible factor
`(HIF). This
`controls the transcription of a wide array of genes involved
`in erythropoiesis, angiogenesis, cell proliferation/survival,
`glycolysis, and iron metabolism in response to oxygen
`availability (Bruick, 2003). The stability of the HIF-1a
`subunit is directly regulated by oxygen. In normoxia, HIF-
`1a is degraded by the proteasome following hydroxylation
`at P402 and P564 within two functionally independent
`degradation domains. This modification is required for the
`recognition of HIF-1a by the von Hippel–Lindau protein
`(pVHL), which is a component of an E3 ubiquitin ligase
`complex (Ivan et al., 2001; Jaakkola et al., 2001). The
`hydroxylation of P402 and P564 is catalyzed by HIF prolyl-
`4-hydroxylases, which are oxygen- and iron-dependent
`enzymes, members of the broader family of 2-oxogluta-
`rate-dependent oxygenases (Bruick and McKnight, 2001;
`Epstein et al., 2001). Thus, the HIF prolyl-4-hydroxylases
`essentially function as boxygen sensorsQ.
`
`Toxicity of iron
`
`The efficiency of Fe(II) as an electron donor and of
`Fe(III) as an electron acceptor, with a redox potential
`
`Fig. 1. (A) Iron-catalyzed generation of the hydroxyl radical via the Fenton
`reaction; the net Haber–Weiss reaction is also indicated. (B) Iron-catalyzed
`generation of organic radicals. (C) Heme-catalyzed generation of oxygen
`radicals via oxoferryl intermediates. (D) Direct interaction of iron with
`oxygen.
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1066 - Page 2
`
`

`

`G. Papanikolaou, K. Pantopoulos / Toxicology and Applied Pharmacology 202 (2005) 199–211
`
`201
`
`amplify the bactericidal (and cytotoxic) capacity of phag-
`ocytic cells and constitute major toxic species in vivo
`(Ischiropoulos and Beckman, 2003). The former is gener-
`ated by the spontaneous reaction of superoxide with NO,
`while the latter is synthesized from hydrogen peroxide and
`chloride in a reaction catalyzed by myeloperoxidase.
`In this milieu, redox active iron catalyzes the generation
`of not only hydroxyl radicals, but also of organic reactive
`S
`S
`), alkoxyl (RO
`), thiyl (RS),
`species, such as peroxyl (ROO
`S
`) radicals (Fig. 1B). Interestingly,
`or thiyl-peroxyl (RSOO
`heme iron (either bfreeQ or within hemoproteins) may also
`catalyze the formation of radicals, mainly via formation of
`oxoferryl intermediates (Ryter and Tyrrell, 2000) (Fig. 1C).
`Finally, ferrous iron can also contribute as a reactant, rather
`than as a catalyst, to free radical generation by a direct
`interaction with oxygen, via ferryl (Fe2+–O) or perferryl
`(Fe2+–O2) iron intermediates (Fig. 1D). It has been proposed
`that when [O2]/[H2O2] N 100, these reactions may represent
`an important source for free radical generation in vivo
`(Huang, 2003).
`Free radicals are highly reactive species and may
`promote oxidation of proteins, peroxidation of membrane
`lipids, and modification of nucleic acids. Likewise, reactive
`nitrogen species, such as peroxynitrite, may lead to protein
`damage via nitration. An increase in the steady state levels
`of
`reactive oxygen (and nitrogen) species beyond the
`antioxidant capacity of the organism, called oxidative (and
`nitrosative) stress,
`is encountered in many pathological
`conditions, such as chronic inflammation, ischemia–reper-
`fusion injury, or neurodegeneration (Ischiropoulos and
`Beckman, 2003). Excess of redox active iron (as well as
`copper) aggravates oxidative (and nitrosative) stress and
`leads to accelerated tissue degeneration. This is evident in
`disorders of hereditary or secondary iron overload (see
`below).
`Under physiological conditions, extracellular iron is
`23 M at neutral
`exclusively bound to transferrin (Kd = 10
`pH), a monomeric glycoprotein serving as the plasma iron
`carrier, which maintains iron soluble and nontoxic, unable to
`engage in Fenton/Haber–Weiss reactions (Ponka et al.,
`1998). In healthy individuals, only 30% of circulating
`transferrin binds to iron. In pathological iron overload, iron
`gradually saturates the iron-binding capacity of transferrin
`and forms redox-active,
`low-molecular-weight chelates.
`Non-transferrin-bound iron eventually gets internalized into
`tissues by poorly defined mechanisms, resulting in cell
`damage and tissue injury.
`
`Cellular iron metabolism
`
`Iron-loaded transferrin binds to its specific receptor on
`the cell surface, the transferrin receptor 1 (TfR1) (Ponka et
`al., 1998). The complex undergoes endocytosis (Cheng et
`al., 2004) and iron is released from transferrin, following
`acidification of the endosome to pH ~5.5, and transported
`
`across the endosomal membrane by the divalent metal
`transporter DMT1. Internalized iron is utilized for metabolic
`purposes (incorporation into iron-containing proteins) and
`excess is detoxified by sequestration into ferritin, an iron-
`storage protein. Ferritin is composed of 24 subunits of H-
`and L-chains that assemble into a shell-like structure
`forming a cavity that accommodates up to 4500 Fe(III)
`ions (Harrison and Arosio, 1996).
`The transferrin/TfR1 pathway defines the major route for
`cellular iron uptake and certain cell
`types (for example
`erythroid cells) completely depend on it for the acquisition
`of iron. Thus, the targeted disruption of mouse TfR1 has
`been associated with early embryonic lethality due to
`defective erythropoiesis (Levy et al., 1999). Transferrin
`receptor 2 (TfR2), a close homologue of TfR1, primarily
`expressed in hepatocytes and hematopoietic cells (Kawabata
`et al., 1999), cannot compensate for TfR1 deficiency and
`/
`mice.
`rescue the embryonic lethal phenotype of TfR1
`/
`embryos is restricted within the
`The defect
`in TfR1
`erythroid and neuronal tissue (Levy et al., 1999), suggesting
`that alternative pathways for cellular iron acquisition in
`other cell types exist.
`For example, a new pathway for the delivery of iron into
`cells via enterobactin and the neutrophil-derived protein
`neutral gelatinase-associated lipocalin (NGAL) has recently
`been described (Goetz et al., 2002; Yang et al., 2002).
`Enterobactin belongs to a class of low-molecular-weight
`iron-chelating metabolites known as siderophores, which
`are synthesized by some bacteria and fungi to scavenge
`extracellular iron. It
`is not clear whether mammals can
`produce siderophores themselves, but
`the above data
`suggest that they can utilize siderophore-based mechanisms
`for iron acquisition.
`the mechanisms of intra-
`Not much is known about
`cellular iron transport to organelles, such as mitochondria.
`There is strong evidence for the existence of a transit pool
`of iron in the cytosol (Kakhlon and Cabantchik, 2002;
`Petrat et al., 2002), which presumably remains bound to
`low-molecular-weight chelates, such as citrate, ATP, AMP,
`or pyrophosphate. It
`is believed that
`the levels of this
`elusive chelatable or blabile iron poolQ (LIP) reflect the
`overall iron status of the cell. The LIP can be monitored
`in situ by fluorescent
`techniques (Kakhlon and Cabant-
`chik, 2002; Petrat et al., 2002). The employment of such
`techniques has revealed the dynamic nature of the LIP
`and uncovered that fluctuations in its levels translate into
`the onset of homeostatic responses to iron availability.
`Importantly, the LIP appears to define a source of redox-
`active iron, which significantly contributes to mechanisms
`of
`iron-mediated toxicity. Thus, expansion of
`the LIP
`upon repression of ferritin expression in human K562
`cells correlated with oxidative stress (Kakhlon et al.,
`2001). Conversely,
`the overexpression of H-ferritin in
`murine MEL cells resulted in reduction of
`the LIP
`concomitantly with the levels of ROIs (Epsztejn et al.,
`1999).
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1066 - Page 3
`
`

`

`202
`
`G. Papanikolaou, K. Pantopoulos / Toxicology and Applied Pharmacology 202 (2005) 199–211
`
`Regulation of cellular iron metabolism: the IRE–IRP system
`
`TfR1 and ferritin are crucial proteins for the control of
`cellular
`iron homeostasis (Hentze et al., 2004). Their
`expression is coordinately and reciprocally controlled in
`response to iron levels by a posttranscriptional mechanism,
`which involves mRNA–protein interactions (Hentze et al.,
`2004; Pantopoulos, 2004). The mRNAs encoding TfR1 (but
`not TfR2) and ferritin (both H- and L-chains) contain biron
`responsive elementsQ (IREs) in their 3V or 5V untranslated
`regions (UTRs), respectively. These are structural motifs of
`
`~30 nucleotides that fold and form a loop with a 5V-
`CAGUGN-3V consensus sequence (the underlined C and G
`interact by hydrogen bonding), and a moderately stable
`(DG c 7 kcal/mol) stem, interrupted by an unpaired C
`residue. They are evolutionary conserved in vertebrates and
`some insects and bacteria (Johansson and Theil, 2002).
`The IRE motifs provide a binding site for two biron
`regulatory proteinsQ, IRP1 and IRP2 (Fig. 2A), which are
`activated in iron-starved cells to bind with high affinity
`12 M)
`c 10
`to cognate IREs. The IRE–IRP
`(K d
`interactions stabilize the otherwise unstable TfR1 mRNA,
`
`Fig. 2. (A) Homeostatic responses to iron supply mediated by IRE–IRP interactions. Decreased iron supply activates binding of IRPs to IREs, resulting in
`stabilization of TfR1 mRNA and translational inhibition of the mRNAs encoding ferritin (H- and L-chains), erythroid-specific ALAS2, and mitochondrial
`aconitase. Conversely, IRPs do not bind to cognate IREs in iron-replete cells, permitting degradation of TfR1 mRNA and translation of ferritin, ALAS2, and
`mitochondrial aconitase mRNAs. (B) Posttranslational regulation of bifunctional IRP1 in response to iron, NO, and H2O2 via an iron–sulfur cluster switch. (C)
`Iron-dependent degradation of IRP2 by a mechanism involving 2-oxoglutarate-dependent oxygenases.
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1066 - Page 4
`
`

`

`G. Papanikolaou, K. Pantopoulos / Toxicology and Applied Pharmacology 202 (2005) 199–211
`
`203
`
`which contains five IRE copies in its long and complex 3V
`UTR and is the only thus far
`identified mRNA with
`multiple IREs. As a result, TfR1 expression levels increase
`and the iron-starved cells acquire more iron from trans-
`ferrin. On the other hand, the IRE–IRP interaction in the 5V
`UTR of ferritin mRNAs specifically inhibits their trans-
`lation and the iron-starved cells do not synthesize ferritin,
`which is under these conditions obsolete. Conversely, in
`iron-replete cells, IRP1 and IRP2 get inactivated, thereby
`permitting TfR1 mRNA degradation and ferritin mRNA
`translation. This response inhibits further iron uptake from
`transferrin and promotes the storage (and detoxification) of
`excess intracellular iron.
`The IRE–IRP system also operates beyond the control of
`cellular iron uptake and storage. This became evident with
`the identification and characterization of additional IRE-
`containing mRNAs, primarily encoding proteins of iron and
`energy homeostasis (Johansson and Theil, 2002). Thus, the
`mRNAs encoding the erythroid ALAS2 and mitochondrial
`aconitase contain a ferritin-type btranslationalQ IRE in their
`5V UTR.
`In addition,
`the mRNAs encoding the iron
`transporters DMT1 and ferroportin 1 contain an incom-
`pletely characterized IRE in their 3V or 5V UTR, respectively.
`Interestingly, a radiation-induced deletion of 58 bp within
`ferroportin 1 IRE has been associated with hypochromic
`microcytic anemia in homozygotic, and polycytemia in
`heterozygotic mice (Mok et al., 2004).
`
`Iron regulatory proteins
`
`The iron regulatory proteins, IRP1 and IRP2, function as
`intracellular iron sensors (Hentze et al., 2004; Pantopoulos,
`2004). They share significant homology and belong to the
`protein family of iron–sulfur cluster isomerases that also
`includes mitochondrial aconitase. However, only IRP1 (Fig.
`2B) is able to assemble an aconitase type 4Fe–4S cluster,
`which not only determines the mode of its function, but also
`serves as a major regulatory site. The cluster assembly only
`occurs in iron-replete cells. Under these conditions, IRP1 is
`catalytically active as a cytosolic aconitase but unable to
`bind to IRE sequences. Iron deficiency triggers a slow
`disassembly of the 4Fe–4S cluster and IRP1 acquires IRE-
`binding activity. The actual mechanism for cluster removal
`is still unclear. IRP1 responds to additional stimuli, such as
`NO or H2O2, which also activate IRE binding, albeit by
`different pathways (Pantopoulos et al., 1996). The H2O2-
`mediated activation of IRP1 very likely involves a signaling
`mechanism (Pantopoulos and Hentze, 1998) and warrants
`particular attention in light of the role of this ROI in iron’s
`toxicity (Fig. 1). This response leads to increased iron
`uptake by the TfR1 and reduced storage in ferritin, and may
`be physiologically relevant in the context of inflammation.
`IRP1 by H2O2 can be
`Note that
`the activation of
`antagonized by hypochloric acid (Mu¨tze et al., 2003), the
`product of
`the myeloperoxidase reaction that emerges
`during the respiratory burst of phagocytic cells. This
`
`IRP1 is able to integrate multiple and
`indicates that
`opposing signals, a feature reminiscent of cytokine regu-
`latory networks.
`IRP2 (Fig. 2C) is regulated in response to iron and
`oxygen supply by distinct post-translational mechanisms
`(Hentze et al., 2004; Pantopoulos, 2004). It is synthesized
`de novo in iron-starved cells and remains stable under these
`conditions or in hypoxia. By contrast, IRP2 undergoes
`proteasomal degradation in iron-replete normoxic cells or in
`cells exposed to sodium nitroprusside, an iron-containing
`compound utilized for protein S-nitrosylation. A character-
`istic feature of IRP2 is the insertion of 73 amino acids
`within a region close to its N-terminus. This highly
`conserved 73 amino acids motif is encoded by a distinct
`exon and has been proposed to be crucial for the iron-
`dependent degradation of IRP2 via site-specific oxidation of
`cysteine residues (Iwai et al., 1995). More recent data
`showed that this region is dispensable for IRP2 turnover and
`the signal initiating IRP2 degradation involves the activity
`of 2-oxoglutarate-dependent oxygenases, by analogy to the
`degradation of HIF-1a (Hanson et al., 2003; Wang et al.,
`2004). Other recent data proposed that the 73 amino acids
`motif plays a role in IRP2 degradation in response to heme
`(Yamanaka et al., 2003) or sodium nitroprusside (Kim et al.,
`2004).
`The targeted disruption of IRP1 and IRP2 has provided
`some insights on the function of these proteins in vivo.
`/
`mice display a mild phenotype with abnormalities
`IRP1
`in iron metabolism restricted to the kidney and brown fat
`/
`mice
`(Meyron Holtz et al., 2004). By contrast, IRP2
`accumulate iron in the intestinal mucosa and the CNS, and
`develop a progressive neurodegenerative disorder (LaVaute
`et al., 2001). Tissue-specific and conditional knockout
`experiments are expected to shed more light on the function
`of IRP1 and IRP2 in the control of cellular and systemic iron
`metabolism.
`
`Body iron homeostasis
`
`Iron distribution in humans
`
`The human body contains approximately 3–5 g of iron
`(45–55 mg/kg of body weight in adult women and men,
`respectively), distributed as illustrated in Fig. 3. The
`majority of body iron (~60–70%)
`is utilized within
`hemoglobin in circulating red blood cells (Andrews,
`1999; Ponka, 1997). Other iron-rich organs are the liver
`and muscles. Approximately 20–30% of body iron is
`stored in hepatocytes and in reticuloendothelial macro-
`phages, to a large extent within ferritin and its degrada-
`tion product hemosiderin. The remaining body iron is
`primarily localized in myoglobin, cytochromes, and iron-
`containing enzymes. A healthy individual absorbs daily
`1–2 mg of
`iron from the diet, which compensates
`nonspecific iron losses by cell desquamation in the skin
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1066 - Page 5
`
`

`

`204
`
`G. Papanikolaou, K. Pantopoulos / Toxicology and Applied Pharmacology 202 (2005) 199–211
`
`Fig. 3. Iron distribution in the adult human body.
`
`In addition, menstruating women
`and the intestine.
`physiologically lose iron from the blood. Erythropoiesis
`requires approximately 30 mg iron/day, which is mainly
`provided by the recycling of iron via reticuloendothelial
`macrophages. These ingest senescent red blood cells and
`release iron to circulating transferrin. The pool of trans-
`ferrin-bound iron (~3 mg) is very dynamic and undergoes
`N10 times daily recycling.
`
`Iron absorption
`
`An average daily Western diet contains approximately 15
`mg of iron, from which only 1–2 mg is absorbed. Two thirds
`from absorbed iron derives from heme, while the remaining
`fraction is inorganic. Both heme and inorganic iron are
`absorbed in the apical membrane of duodenal enterocytes.
`The pathway for heme uptake by the duodenal epithelium is
`still incompletely characterized, while the mechanism for
`inorganic iron transport has been widely explored in the last
`few years. The low pH of the gastric effluent dissolves
`
`ingested inorganic iron and facilitates its enzymatic reduc-
`tion to the ferrous form by the brushborder ferrireductase
`Dcytb (McKie et al., 2001). Ferrous iron is transported
`across the apical membrane by DMT1 (also found in
`literature as DCT1 or Nramp2), a divalent metal transporter
`of broad specificity, including manganese, copper, cobalt,
`zinc, cadmium, and lead cations (Fleming et al., 1997;
`/
`Gunshin et al., 1997). Animals defective in DMT1 (mk
`mice and Belgrade rats) suffer from a severe hypochromic
`microcytic anemia due to impaired iron absorption and
`recycling.
`Inside absorptive enterocytes, heme iron is enzymatically
`liberated by heme oxygenase and follows the fate of
`inorganic iron: It is either stored in ferritin or transported
`across the basolateral membrane to plasma transferrin.
`Basolateral
`iron transport
`is mediated by ferroportin 1
`(Donovan et al., 2000), also found in literature as IREG1
`(McKie et al., 2000) or MTP1 (Abboud and Haile, 2000),
`which is thought to function in concert with the membrane
`ferroxidase hephaestin. The importance of hephaestin in
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1066 - Page 6
`
`

`

`G. Papanikolaou, K. Pantopoulos / Toxicology and Applied Pharmacology 202 (2005) 199–211
`
`205
`
`iron absorption has been denoted by the observation that
`sex-linked anemia (sla) mice with a defect in hephaestin
`display normal dietary iron absorption in enterocytes but
`impaired export of
`iron from the enterocytes to the
`circulation (Vulpe et al., 1999). Consequently, sla mice
`develop microcytic anemia. The export of
`iron from
`enterocytes and its binding to transferrin may also be
`facilitated by the soluble plasma ferroxidase ceruloplasmin
`(Hellman and Gitlin, 2002), which shares significant
`homology with hephaestin.
`
`Maintenance of body iron homeostasis
`
`Mammals do not possess any physiological pathway for
`iron excretion. Thus, body iron homeostasis is regulated at
`the level of iron absorption. Misregulated iron absorption
`leads to iron deficiency or overload. It is believed that three
`regulatory cues contribute to the maintenance of homeo-
`stasis (Andrews, 1999; Finch, 1994). The first is called
`bdietary regulatorQ. It has long been known that after the
`ingestion of a dietary iron bolus, absorptive enterocytes are
`resistant
`in acquiring additional
`iron for several days
`(Stewart et al., 1950). This phenomenon, also described as
`bmucosal blockQ, probably results from the accumulation of
`intracellular iron. High intracellular iron may suppress the
`expression of DMT1 in an IRE–IRP-mediated manner
`(Frazer et al., 2003) because DMT1 contains an IRE in
`the 3V UTR of its mRNA.
`A second signal, called bstores regulatorQ, controls iron
`uptake in response to body iron stores. It is well established
`that in iron-deficient conditions, iron absorption is signifi-
`cantly stimulated by two- to threefold. When iron stores are
`replenished, iron absorption returns to basal levels. It has
`been hypothesized that this type of regulation requires the
`programming of precursor crypt cells (Roy and Enns, 2000)
`in the duodenal epithelium after sensing plasma transferrin
`saturation.
`A third signal, called berythropoietic regulatorQ, modu-
`lates iron absorption in response to erythropoiesis. Because
`most of the body’s iron is utilized by the bone marrow for
`hemoglobinization of red blood cells, it is not surprising that
`this signal has a dominant function in the control of iron
`homeostasis. In other words, the erythropoietic regulator has
`a greater capacity to increase iron absorption compared to the
`stores regulator (Andrews, 1999; Finch, 1994). Moreover, it
`increases iron absorption independently of body iron stores.
`A reason for the pathological iron accumulation encountered
`in disorders with ineffective erythropoiesis (such as thalas-
`semia syndromes, congenital dyserythropoetic anemias,
`sideroblastic anemias, or atransferrinemia) is the increased
`iron absorption under conditions where iron stores are
`replete. The nature of the erythropoietic regulator has
`remained for long time elusive. Recently, a dual function
`compatible with both a bstoresQ and berythropoieticQ
`regulator has been proposed for hepcidin, a small circulating
`peptide derived from the liver (Ganz, 2003).
`
`The function of hepcidin
`
`Hepcidin shares similarity with many antimicrobial
`peptides and was initially isolated from plasma and urine
`ultrafiltrates on the basis of its antimicrobial properties
`(Krause et al., 2000; Park et al., 2001). Unlike other
`antimicrobial peptides, hepcidin is predominantly expressed
`in the liver. The gene encoding hepcidin (HAMP) resides in
`chromosome 19q13 and encodes an 84 amino acid
`precursor. This is processed to a mature, biologically active
`N-terminal peptide of 20–25 residues, upon enzymatic
`cleavage at a site comprising five arginines (Ganz, 2003).
`Overwhelming genetic data underscore the importance of
`hepcidin in the regulation of body iron homeostasis. First,
`/
`mice with disrupted hepcidin expression display
`USF2
`an iron overload phenotype, similar to that observed in
`hereditary hemochromatosis (Nicolas et al., 2001). More-
`over, transgenic mice overexpressing hepcidin from a liver-
`specific promoter display severe iron deficiency anemia
`(Nicolas et al., 2002a). A first link of hepcidin with human
`disease was established by the identification of hepcidin-
`inactivating mutations in a rare form of juvenile hemochro-
`matosis (Roetto et al., 2003). At the other end, increased
`expression of hepcidin clearly contributes to the patho-
`genesis of the anemia of chronic disease (or anemia of
`inflammation), a condition characterized by hypoferremia as
`a result of iron retention within macrophages (Weiss, 2002).
`The expression of hepcidin is induced by lipopolysaccharite
`(Roy et al., 2004) and the inflammatory cytokine IL-6
`(Nemeth et al., 2004), which is typical of a type II acute
`phase molecule. Anemia and hypoxia are negative regu-
`lators (Nicolas et al., 2002b). Hepcidin expression is
`increased in inflammation (Nemeth et al., 2004) and
`secondary iron overload (Adamsky et al., 2004) and is
`suppressed in hereditary hemochromatosis (Bridle et al.,
`2003; Gehrke et al., 2003).
`A model for hepcidin function derived from a growing
`body of experimental data can be summarized as follows
`(Fig. 4): Under physiological conditions, mammals produce
`and maintain relatively stable levels of hepcidin. Low
`hepcidin levels trigger increased iron absorption from the
`duodenum and iron release from reticuloendothelial macro-
`phages. By contrast, increased secretion of hepcidin leads to
`decreased iron absorption and iron retention in reticuloen-
`dothelial macrophages. Hepcidin levels reflect body iron
`stores and the iron demand for erythropoiesis. Thus,
`hepcidin expression is inhibited when body iron stores are
`depleted (e.g., in iron deficiency), or when iron is acutely
`required for erythropoiesis (e.g., in hypoxia and anemia).
`Likewise, hepcidin expression increases when iron stores
`are replenished or during the inflammatory response, where
`erythropoiesis is contained. Hence, hepcidin clearly fulfills
`functions of both berythroidQ as well as bstoresQ regulator.
`It is likely that liver hepatocytes relay signals derived
`from the bone marrow and the sites of
`iron storage
`(hepatocytes and macrophages) that in turn regulate the
`
`PGR2020-00009
`Pharmacosmos A/S v. American Regent, Inc.
`Petitioner Ex. 1066 - Page 7
`
`

`

`206
`
`G. Papanikolaou, K. Pantopoulos / Toxicology and Applied Pharmacology 202 (2005) 199–211
`
`Fig. 4. (A) The sequence of mature human hepcidin with cysteine disulfide bridges. Cleavage sites giving rise to three isoforms of 25, 22, or 20 amino acids,
`respectively, are indicated by arrows. (B) Model for the regulatory functions of hepcidin. A decrease in plasma hepcidin levels, as a result of reduction in body
`iron stores, requirement for erythropoiesis or hypoxia, promotes dietary iron absorption and iron release from macrophages. An increase in plasma hepcidin
`levels in response to iron loading or inflammation inhibits dietary iron absorption and iron release from macrophages.
`
`production of hepcidin. Because the liver and bone marrow
`the berythroidQ regulation must
`are not adjacent tissues,
`depend on a soluble plasma signal, which remains
`unknown. Hepcidin itself appears to be a final mediator of
`both erythroid and stores regulators and its levels are critical
`for intestinal iron absorption and iron release from macro-
`phages. It should be noted that the above model on hepcidin
`function is almost entirely based on genetic d