Murine mutants in the study of systemic iron metabolism and its disorders: an update on recent advances

Abstract

Many past and recent advances in the field of iron metabolism have relied upon the use of mouse models of disease. These models have arisen spontaneously in breeder colonies or have been engineered for global or conditional ablation or overexpression of select genes. Full phenotypic characterization of these models typically involves maintenance on iron-loaded or – deficient diets, treatment with oxidative or hemolytic agents, breeding to other mutant lines or other stresses. In this review, we focus on systemic iron biology and the contributions that mouse model-based studies have made to the field. We have divided the field into three broad areas of research: dietary iron absorption, regulation of hepcidin expression and cellular iron metabolism. For each area, we begin with an overview of the current understanding of key molecular and cellular determinants then discuss recent advances. Finally, we conclude with brief comments on prospects for future study.

1.1 Introduction

Inherited mutant or transgenic mouse models of disease have proven invaluable to the biomedical sciences; they offer insights into the pathophysiology of disease that cell culture experiments or in vitro biochemical assays typically cannot provide. The field of iron metabolism is no exception, where they have been actively sought out, created and studied for more than 50 years. The first models were spontaneous or induced anemia mutants that provided for physiological study and eventually, as the knowledge of genomes evolved, permitted phenotype-driven gene discovery through positional cloning. As methods of manipulating animal genomes became more advanced, prospectively engineered, gene-targeted mutants became more common. Currently, the standard is targeted global, developmental stage- and tissue-specific ablation or overexpression of selected genes that are often combined with nutritional or other environmental stresses to reveal non-equilibrium phenotypes.

In this review, we have divided systemic iron biology into several broad areas of research: dietary iron absorption, regulation of hepcidin expression, and intracellular iron metabolism. In each case, we will first present a brief overview of the relevant molecular and cellular biology with particular reference to the mouse and other animal models that helped to define their relevance[1]. We will then describe more recent findings and conclude with comments on the outlook for future work in the field.

1.2 Dietary iron absorption

Mammals regulate systemic iron stores entirely at the level of intestinal absorption; iron elimination from the body is not regulated and is entirely dependent upon physiological and non-physiological blood and epithelial cell loss. While dietary iron exists both as heme and non-heme iron, the absorption of the latter is much better understood and all animal models of aberrant dietary iron absorption appear to involve non-heme iron absorption. Absorption of dietary non-heme iron involves the liberation of elemental iron from digested food and its maintenance in a soluble form, which is accomplished in part by gastric acid[2]. Ferric iron (Fe³⁺) is then reduced to ferrous iron (Fe²⁺) by ferric reductases present on the apical surface of duodenal enterocytes. Deletion of one putative intestinal ferrireductase, Dcytb, has limited effects on iron absorption, even in iron deficient animals[3–5], suggesting that Dcytb is either not essential for iron absorption or other factors such as ascorbic acid can compensate for Dcytb deficiency in mice. The identification of the mutations in divalent metal transporter 1 (Dmt1, Slc11a2) found in the microcytic anemia (mk) mouse and the Belgrade (b) rat defined Dmt1 as the apical intestinal iron transporter. As Dmt1 is an Fe²⁺/H⁺ symporter[6], the acidic microenvironment of the proximal duodenum also promotes transmembrane iron transport. Iron trafficked across the apical membrane may be stored in ferritin, or trafficked across the enterocyte basolateral membrane by the Fe²⁺ transporter ferroportin (Fpn1, Slc40a1). Fpn1 was originally cloned as the gene responsible for the zebrafish weissherbst (wei) iron deficiency anemia phenotype[7]. The systemic iron regulatory hormone hepcidin (discussed in detail below) regulates iron efflux across the basolateral membrane of the enterocyte by binding to and causing the degradation of Fpn1. Efficient transport of iron across the basolateral membrane also requires reoxidation of Fe²⁺ to Fe³⁺ by the homologous multicopper oxidases hephaestin, mutated in the sex-linked anemia (sla) mouse, and ceruloplasmin[8–10].

1.2.1 The role of duodenal acidification in iron absorption

While the role of an acidic microenvironment in iron absorption has been known for decades, only recently have genetic models been established to directly assess the physiological relevance of gastric acid secretion to intestinal iron absorption. For example, mice with a targeted deletion of gastrin, which stimulates acid production by the gastric parietal cell, develop a hypochromic, microcytic anemia, but only when fed an iron-deficient diet[11]. Intriguingly, in these mice, hepcidin levels do not decrease as markedly as in wild-type mice on iron-deficient diets. This effect may be related to the finding that parietal cells themselves express hepcidin and that hepdicin may also directly or indirectly regulate gastric acid production[12]. Furthermore, based on phenotypic characterization of the ethylnitrosourea-induced mouse mutant sublytic, it appears that the parietal cell acid secretion is largely accomplished by the H⁺/K⁺ ATPase Atp4a[13]. These animals have increased gastric pH, decreased serum and liver iron levels, appropriately decreased liver hepcidin levels, and an age-dependent hypochromic, microcytic anemia that is corrected by acidification of drinking water. That sublytic animals show aberrant ⁵⁹Fe absorption on an iron-deficient, but not an iron-replete diet, further emphasizes the notion introduced by gastrin-deficient animals that duodenal pH may only be critical in iron-limited conditions.

1.2.2 Post-translational modulation of intestinal iron transporters

Since its description as the apical intestinal iron transporter, it has been known that Dmt1 mRNA is strongly upregulated in response to iron deficiency[6]. Dmt1 mRNA stability may also be regulated by an iron-responsive element in the 3’ untranslated region[14]. In vitro studies indicated that Dmt1 protein degradation is regulated in a ubiquitin-dependent manner mediated by the E3 ubiquitin ligase Nedd4-2 whose activity is facilitated by the adapter proteins Ndfip1 and Ndfip2[15]. Ndfip-null mice have a complex phenotype that includes an inflammation-dependent microcytic anemia; on a Rag-null immunodeficient background these animals overexpress intestinal Dmt1 and develop a mild iron overload phenotype.[16] More recently it has been reported that the enterocyte basolateral transporter Fpn1 is also regulated by Nedd4-2 and Ndfip in a hepcidin-independent manner[17], suggesting that the iron overload phenotype seen in immunodeficient Ndfip-null mice may also be related to inappropriate persistence of Fpn1 expression. Expression of Dmt1 and Fpn1 are also dependent upon iron regulatory proteins (Irp) 1 and 2, RNA-binding proteins that regulate protein expression by binding to RNA motifs known as iron-responsive elements (IREs). Mice with combined intestinal Irp1 and 2 deficiency exhibit aberrant intestinal villus architecture and die postnatally from intestinal malabsorption and dehydration[18]. At the molecular level, these mice also display aberrant expression levels of known Irp targets including Tfr1, ferritin H and L chains, Fpn1 and surprisingly +IRE and −IRE forms of Dmt1. Increased liver hepcidin levels may reflect decreased erythropoietic inhibition or increased inflammatory stimulation of unclear etiology, given that serum or tissue iron levels are not perturbed. Notably the increased Fpn1 protein levels in spite of increased hepcidin levels indicate that Irp activity is essential for Fpn1 regulation.

1.2.3 Regulation of intestinal iron absorption by ferritin

Ferritin is a multimeric protein composed of H (Fth) and L (Ftl) chains capable of storing up to 4500 Fe³⁺ ions per multimer. The early embryonic lethality of Fth deficiency in mice[19] prompted the interrogation of ferritin function using a conditional deletion strategy. Fth deletion induced in adulthood leads to decreased tissue iron stores when fed a diet with normal iron content and severe liver damage when fed a high-iron diet[20]. Intestine-specific deletion of the Fth chain indicates that intestinal ferritin is essential for appropriate dietary iron absorption [21]. These mice display increased serum and tissue iron levels and appropriately increased hepcidin levels. The observed increase in intestinal ⁵⁹Fe absorption and Fpn1 expression may reflect inactivation of intestinal Irp activity by increased intracellular iron levels.

1.2.4 Regulation of intestinal iron absorption by hypoxia inducible factors

Increased iron absorption is one of the metabolic adaptations to hypoxia. Hypoxia-induced transcriptional programs are mediated primarily by the transcription factors Hif1 and Hif2 and regulated by the von Hippel-Lindau factor (VHL or Vhlh) E3 ubiquitin ligase and oxygen-dependent prolyl hydroxylases. Hif1 and Hif2 share a common β subunit, Arnt (aryl hydrocarbon nuclear translocator), and each has a labile α subunit that is degraded in a Vhlh- and proteasomal-dependent manner in iron- and oxygen-replete conditions. With the use of germline and tissue-specific conditional deletions of genes in this pathway, it has been possible to dissect those changes that are a consequence of a direct effect on intestinal epithelial cells and those that are secondary to primary responses in other organs.

Several research groups have investigated the role of the Hif pathway in regulating iron metabolism in intestinal cells. Mastrogiannaki et al. characterized mice with intestinal Hif1α or Hif2α deficiency [22]. Only in the latter did they observe significant alterations in iron metabolism, including decreased intestinal Dmt1 and Fpn1 mRNA levels that were associated with decreased serum and liver iron levels, increased serum erythropoietin and decreased hepatic hepcidin expression. Despite the decreased mRNA levels, duodenal Fpn1 protein expression was unchanged, likely reflecting equilibrium between decreased Hif2-dependent effects on Fpn1 mRNA and decreased hepcidin-dependent Fpn1 protein destabilization. An iron-deficient diet worsened hematologic parameters in intestinal Hif2α-deficient mice but not wild-type controls. Taylor et al. also found that intestinal Hif2α deficiency abrogated the increase in duodenal Fpn1 protein observed in mice placed on an iron-deficient diet [23]. Shah and colleagues, using Vhlh- and Hif1α-deficient mice, which lack Hif1 and have constitutively active Hif2, as well as Vhlh- and Arnt-deficient mice, which lack both Hif1 and HIf2 activity, reached largely similar conclusions [24]. The pivotal role of intestinal Hif2α has been further tested by treating intestinal Hif2α-deficient mice with the hemolytic agent phenylhydrazine [25]. Results of these studies indicate that Hif2α is essential for the increased intestinal Dmt1 mRNA and protein levels and increased serum iron levels but not the increased renal erythropoietin and intestinal Fpn1 levels seen in acute hemolysis. Also consistent with a cell-autonomous role for Hif2α in the regulation of intestinal iron absorption is the recent finding that enterocyte Hif2α deficiency attenuates tissue iron loading in hepcidin-deficient mice [26].

1.3 Regulation of hepcidin expression in hepatocytes

As described above, intestinal iron absorption, and consequently total body iron, is strongly influenced by hepcidin, a peptide hormone synthesized predominantly by hepatocytes in the liver that circulates in the plasma and induces internalization and lysosomal degradation of Fpn1 [27]. Hepcidin expression is stimulated by iron overload and inflammation and inhibited by anemia and hypoxia. Given the central role of hepcidin in modulating systemic iron stores, the regulation of hepcidin is one of the most active areas of research within the field of iron biology [28].

In many ways, our early understanding of the regulation of hepcidin in hepatocytes mirrored the identification of genes responsible for autosomal recessive (AR) forms of the human iron overload disorders collectively called hereditary hemochromatosis (HH). AR HH can be due to loss of function mutations in HFE, hemojuvelin (HJV), transferrin receptor 2 (TFR2) or hepcidin (HAMP) itself. In each case, the expression of hepcidin is inappropriately low for the degree of iron stores and the severity of the iron overload is roughly proportional to the extent of relative hepcidin deficiency. Cell-based assays demonstrated that there is a network of interactions between diferric transferrin (Fe₂TF), transferrin receptor 1 (TFRC or TFR1), TFR2 and HFE in which Fe₂TF competes with HFE for binding to TFR1. In this manner, HFE binding to TFR2 is sensitive to the concentration of Fe₂TF. HJV is a bone morphogenetic protein (BMP) co-receptor essential for hepcidin regulation in hepatocytes that appears to be functionally downstream of HFE and TFR2. Taken together, these data have led to the notion that loss of any of the AR HH proteins leads to a defect in “iron sensing” by the hepatocyte and that HH is a disorder in which hepcidin production is decoupled from serum Fe₂TF levels, which normally serve to induce its production [28]. Not surprisingly, soon after the description of each AR HH gene, a targeted loss of function mutation in the mouse for each of these proteins was shown to reproduce the HH phenotype. There are also mouse models of AR HH established for reasons other than validation of genes identified in patient studies—examples include mice deficient in β₂-microglobulin, a factor believed to be required for HFE localization, and recently developed mouse models which will be reviewed here. These recent models have focused on delineating the requirements for these proteins in iron metabolism in a cell-type specific manner, further characterizing the interrelationships of the AR HH proteins, and identifying or validating other proteins involved in the pathway.

1.3.1 Hereditary hemochromatosis is a primary liver disease dependent on the interactions of the AR HH proteins and TF

Although the predominant expression of hepcidin by the liver would suggest that primary, and not secondary, defects in hepatocyte hepcidin expression would underlie the pathogenesis of AR HH, the widespread expression of HJV, HFE, and to a much lesser extent TFR2, indicated that other tissues might contribute to hepatocellular hepcidin dysregulation. The creation of conditionally targeted alleles of all three proteins, however, has dispelled any of these hypotheses and has helped elaborate the mechanisms by which they each influence hepcidin gene transcription in the liver.

Several animal models demonstrate that HFE plays a role in hepcidin regulation in the liver. Most directly, mice lacking Hfe only in hepatocytes have a hemochromatosis phenotype identical to animals lacking global expression of the molecule [29]. Conversely, overexpression of Hfe in a liver-specific manner is not only able to correct the hemochromatosis phenotype of Hfe-deficient mice, but in fact leads to elevated hepcidin expression and a microcytic, hypochromic anemia [30, 31]. Similarly, adenoviral expression of Hfe in the liver of wild-type or Hfe-deficient animals is able to upregulate hepcidin expression [32].

As described above, the prevailing model of HFE function holds that HFE acts as a sensor for the concentration of Fe₂TF. Analysis of mice with mutations in Tfr1 that promote the Hfe-Tfr1 interaction demonstrated that, when Hfe is sequestered by Tfr1, animals develop iron overload and have inappropriately low hepcidin expression. Conversely, animals with Hfe mutations that abrogate the ability of Hfe to bind to Tfr1 develop an iron deficiency related to hepcidin overexpression [30]. In contrast to cell culture work demonstrating that the cytoplasmic tail is crucial for Hfe function [33], mouse models overexpressing Hfe lacking the cytoplasmic domain exhibit upregulated hepcidin expression leading to decreased total body iron stores [31].

The role that TF plays in regulating hepatocyte hepcidin expression has also recently been investigated using the hypotransferrinemia (hpx) mouse model of transferrin deficiency [34]. Due to the essential role of TF in delivering iron for erythropoiesis, these animals suffer from severe hypochromic, microcytic, iron deficiency anemia. However, perhaps paradoxically, hpx animals have massive systemic iron overload and profound hepcidin insufficiency. Transfusion of hpx animals with wild-type red blood cells leads to increased hepcidin levels, suggesting that anemia and/or hypoxia suppress hepcidin expression. Furthermore, treatment of hpx mice with the chemotherapeutic agent doxorubicin, thereby ablating erythropoiesis, and then with TF also leads to increased hepcidin levels. This indicates that TF can stimulate hepcidin expression independently of its role in erythropoiesis [35]. Similar transfusion and marrow ablation experiments were also performed in hpx animals on an Hjv-deficient background. In this case, there was a very poor hepcidin response following TF supplementation, indicating that Hjv is essential for TF-dependent and -independent hepcidin expression [36].

While the phenotype of hepcidin deficiency and iron overload observed in mice with liver-specific Tfr2 deficiency suggests that the regulation of hepcidin expression by Tfr2 depends upon Tfr2 activity in the liver [37], the mechanism by which this regulation occurs is uncertain. A potential functional relationship between Hfe and Tfr2 has been explored using several different mouse models. Mice deficient in both Hfe and Tfr2 have been shown to have more severe iron loading than animals lacking either gene alone [38], and when normalized to iron stores, hepcidin expression was more profoundly suppressed in animals lacking both proteins. Moreover, induction of hepcidin by lipopolysaccharide (LPS), which models inflammation-induced hepcidin expression, is blunted in mice deficient in Hfe, Tfr2 or both proteins, and that this effect is most pronounced in the double mutant [39].

Germline targeted mutation of Hjv has been shown by several investigators to result in an HH phenotype [40, 41]. Interestingly, however, HJV mRNA is most highly expressed in skeletal muscle and only at lower levels in liver [42]. As a soluble form of HJV is found in circulation and had been shown to modulate BMP-mediated regulation of hepcidin expression, it had been hypothesized that soluble HJV produced in muscle was potentially relevant to hepatic hepcidin expression. However, liver- and skeletal muscle-specific Hjv deletion models show that this clearly is not the case. Loss of Hjv in the liver alone is sufficient to cause hepatic hepcidin deficiency and systemic iron overload, whereas, ablation of Hjv in skeletal muscle has no effect on iron loading or hepcidin expression [43, 44].

1.3.2 BMP6, ALK2 and ALK3 are the BMP isoform and BMP type I receptors most relevant to hepcidin regulation

In vitro studies had indicated that several different BMP isoforms stimulate hepcidin production. Studies in wild type mice fed iron-rich or -deficient diets demonstrated that Bmp6 protein levels correlated with iron levels [45], suggesting that this protein was most relevant to regulating hepcidin in vivo. This was confirmed by two research groups who demonstrated simultaneously that global deletion of Bmp6 in mice causes a severe HH phenotype [46, 47]. Furthermore, treatment of mice with a Bmp6-neutralizing antibody inhibits hepcidin expression and increases serum iron levels, while treatment of mice with Bmp6 increases hepcidin expression and reduces serum iron levels [46]. Although the tissue-specific targeted deletion studies have not yet been performed, several reports have indicated that the liver is the main source of Bmp6 [48–50]; one conflicting study, however, suggests that the duodenum is the predominant site of BMP6 expression [51].

Studies in multiple animal models have demonstrated that Hfe functions in BMP-mediated regulation of hepcidin expression. Liver-specific overexpression of Hfe leads to elevated hepcidin expression due to increased BMP signaling, decreased tisue iron levels and microcytic, hypochromic anemia [30, 31, 52]. Conversely, Hfe-deficient mice, although iron overloaded, do not have elevated levels of hepatic phosphorylated Smad 1/5/8 protein and other markers of BMP signaling [52, 53]. Nevertheless Bmp6 expression is appropriately increased in the iron loaded livers of Hfe-deficient mice, demonstrating that Hfe is not necessary for regulation of Bmp6 by iron. Treatment of transgenic mice overexpressing Hfe with a Bmp6 neutralizing antibody decreases hepcidin expression and increases iron stores [54]. Conversely, treatment of Hfe-deficient mice with Bmp6 increases hepcidin expression leading to reduced serum iron and TF saturation. Transgenic hepatic overexpression of Hfe in Hjv-deficient mice does not increase hepcidin expression or modulate Hjv-deficient iron overload [31]. In total, these studies indicate loss of Hfe and Hjv do not impair Bmp6 production by the hepatocyte in response to iron, but do, to a lesser and greater degree, respectively, inhibit BMPR/SMAD-dependent signaling and hepcidin transcription in response to Bmp6.

Although it was well established that HJV was a BMPR co-receptor [55, 56], it was not clear which of the type I or II BMPRs was required for HJV-dependent signaling in response to Bmp6 and iron. An initial clue came from the treatment of mice with the drug dorsomorphin, an inhibitor of the BMP type I receptors Alk2, Alk3 and Alk6, which was found to inhibit hepcidin expression [57]. Subsequently, it was shown that liver-specific deletions of Alk2 or Alk3 cause iron overload in mice, but loss of Alk3 caused a more pronounced iron loading due to more substantial downregulation of BMP signaling and hepcidin expression [58]. Although Alk6 is also expressed in hepatocytes, loss of this receptor does not appear to affect hepcidin expression.

1.3.3 Regulation of hepatocellular HJV expression and function by multiple mechanisms alters hepcidin production

Commensurate with the severe, juvenile onset HH phenotype seen in human patients with biallelic mutations in HJV, hepatocyte hepcidin expression appears to be exquisitely tuned to the levels of HJV protein expressed by the hepatocyte. At least three distinct post-transcriptional mechanisms of modulating HJV expression have been described.

Beutler and colleagues identified the mask mouse in a chemical mutagenesis screen for mouse phenotypic abnormalities. Homozygous mask mice have microcytic anemia and alopecia that spares the head (a “mask” of fur); the phenotype results from systemic iron deficiency due to impaired intestinal iron absorption secondary to inappropriately high levels of hepcidin production for the degree of iron stores [59]—the phenotype “opposite” that of AR HH. The mask mutation is a splicing mutation in the Tmprss6 gene, which encodes a transmembrane serine protease that is highly and selectively expressed in the liver [59]. Targeted disruption of Tmprss6 produces an identical phenotype [59, 60]. Erythropoietin administration to mask mice improves the anemia and decreases hepcidin levels, suggesting that Tmprss6 is not required for erythropoietin-dependent regulation of hepcidin expression [61]. Intercrossing Tmprss6-deficient animals with Hfe-deficient or liver-specific Hfe-transgenic mice demonstrated that Hfe does not modify the hepcidin overexpression or iron deficiency due to inactivation of Tmprss6 [62]. In contrast, coexisting Bmp6 deficiency can correct the Tmprss6 anemia and the Tmprss6- and Hjv-deficient animal has a phenotype indistinguishable from Hjv deficiency alone [63–65]. Taken together, these results demonstrate that Tmprss6 acts in iron-dependent BMP/SMAD signaling in hepatocytes, and support the idea that Tmprss6, like Hjv, acts downstream of Hfe in regulating hepcidin gene transcription. Furthermore, these results are compatible with the notion, as has been demonstrated in tissue culture, that Tmprss6 functions by proteolyzing Hjv from the hepatocyte cell surface [66].

Recent work also suggests that Hjv function may be controlled by factors other than Tmprss6. Similar to AR HH models, transgenic knockdown of the deleted in colorectal cancer family member neogenin results in iron loading, decreased hepcidin expression and reduced BMP signaling [67]. Intriguingly, in this animal model loss of neogenin decreased Hjv protein expression in both muscle and liver tissue suggesting it may play a general role in maintenance of the protein on the cell membrane. Other work also indicates that the liver-specific microRNA miR-122 regulates expression of Hjv. Depletion of miR-122 by injection of specific anti-miR reagents produced systemic iron deficiency in mice along with increased mRNA expression of hepcidin, Hjv and Hfe; in vitro work demonstrated that miR-122 modulates gene expression by targeting the 3’ untranslated region of Hjv and Hfe mRNA [68].

1.3.4 Regulation of hepcidin by anemia and hypoxia

The regulation of hepcidin expression by anemia and hypoxia is a complicated area of study, largely because anemia and hypoxia necessarily coexist, making it difficult to segregate the effect of one from the other. Nonetheless, it does appear that liver hepcidin expression is suppressed by a bone marrow-derived factor likely elaborated by erythroid cells. This as yet undefined factor, termed “the erythroid regulator” of iron metabolism, appears to exert a particularly strong effect in the setting of ineffective erythropoiesis—a condition in which erythropoietic mass is expanded, but reduced numbers of mature, enucleated erythrocytes are released due to death of precursors in the bone marrow. Two such erythroid precursor-derived factors, growth differentiation factor 15 (GDF15) and twisted gastrulation (TWSG1), both of which inhibit BMP signaling, have been advanced as candidates the erythroid regulator. Both are elevated in anemias characterized by ineffective erythropoiesis and both inhibit hepcidin expression in vitro, but in neither case have animal mutant models validating this effect in vivo been reported [69].

As described above, hypoxia inducible factors mediate the systemic and cellular response to hypoxia. The role of liver-derived HIFs in the direct regulation of hepcidin expression has been investigated with tissue-specific mouse models of HIF deficiency. Hepatocyte-specific Hif1α deletion in mice partially blunts the decrease in hepcidin levels observed in wild-type mice on an iron-deficient diet, suggesting that Hif1α is involved, but not solely responsible, for downregulation of hepcidin expression under conditions of iron deficiency [70]. Analysis of gene expression in mice with hepatocyte-specific deficiencies in Vhlh, Hif1α and/or Hif2α indicated that hepatic expression of iron metabolism-related genes such as Dmt1 is dependent largely on Hif2 [71]. Mice with hepatocyte Vhlh deficiency exhibit increased Hif levels and polycythemia most likely secondary to their increased erythropoietin levels; a low mean cell hemoglobin level, along with decreased liver and spleen iron and ferritin levels, indicate a relative iron deficiency [70]. They also exhibit decreased hepcidin levels, suggesting that Vhlh is not essential for the regulation of hepcidin expression by iron deficiency, and increased duodenal Fpn1 expression.

1.4 Cellular iron metabolism

Most of the iron in the body is dedicated to erythropoiesis. The delivery of iron to erythroid precursors is largely if not wholly dependent upon TF. Fe₂₋TF binds to the ubiquitously expressed cell surface membrane protein TFR1 and the resulting complex is internalized via receptor-mediated endocytosis. Endosomal acidification lowers the affinity of TF for Fe³⁺, which can then be transported into the cytosol by DMT1 after reduction to Fe²⁺ by the ferrioxidase Steap3. Within erythroid and non-erythroid cells, iron is incorporated into heme, iron-sulfur clusters and iron-dependent enzymes. Once red blood cells senesce or become damaged, they are phagocytosed by macrophages in the spleen, bone marrow, and other tissues. Iron is liberated from scavenged heme by heme oxygenase 1, exported into plasma by Fpn1, oxidized by ceruloplasmin then bound to and distributed by TF.

1.4.1 Appropriate regulation of erythroid TFR1 expression is essential for erythropoiesis

Targeted deletion of Nme-1 and -2, encoding NdpkA and B, which are highly similar and ubiquitously expressed proteins that catalyze phosphate transfer between ribo- and deoxyribonucleotides, results in mice with reduced perinatal viability, growth retardation and anemia [72, 73]. Decreased Tfr1 expression in erythroid precursors and decreased blood cellular iron and heme levels, despite increased serum iron levels in the mutant mice, suggest that NdpkA/B are essential for appropriate expression of erythroid Tfr1. Aberrant erythroid Tfr1 expression also results from hematopoietic-specific deletion of the Stat5a/b locus encoding the cytokine-regulated transcription factors Stat5A/B[74, 75]. These mice also exhibit a perinatal mortality and anemia despite increased serum and liver iron levels. While decreased expression of anti-apoptotic proteins and apoptosis in fetal liver, a dominant site of fetal erythropoiesis, suggests that anemia results from apoptosis of erythroid precursors, erythroid cells from Stat5a/b-deficient mice have markedly decreased Tfr1 mRNA and protein levels, but also decreased expression and mRNA-binding activity of iron regulatory protein 2 (Irp2), a known post-translational mediator of Tfr1 expression.

1.4.2 Erythroid hemoglobinization requires mitochondrial iron import, iron incorporation into heme precursors and erythroid heme export

Mitochondrial iron import, an essential step in heme biosynthesis, is mediated by mitoferrin (Slc25a37) [76]. Whereas germline mitoferrin deficiency leads to embryonic lethality, induced deletion of mitoferrin in adult hematopoietic tissues leads to severe anemia due to defective erythroblast maturation [77]. Incorporation of iron into protoporphyrin IX, the last step in heme biosynthesis, is catalyzed by ferrochelatase and ferrochelatase-deficient mice develop mild hypochromic, microcytic anemia and increased protoporphyrin levels without alterations in erythroid iron, serum iron or hepcidin levels [78]. Intriguingly, the phenotype of mice deficient in the feline leukemia virus subgroup C receptor (Flvcr) suggests that heme export is essential for erythroid precursor viability. While total deletion of Flvcr in mice leads to embryonic lethality, multiple congenital abnormalities and fetal erythropoiesis halted prior to hemoglobinization [79], mice with postnatal Flvcr deficiency exhibit severe hyperchromic, microcytic anemia and a block in erythroid maturation. Bone marrow transplantation studies indicate that hematopoietic Flvcr deficiency is sufficient for the phenotype observed in globally deleted mice, while in vitro experiments using macrophages derived from mutant mice demonstrate a defect in heme export.

1.4.3 Red cell iron reclamation is essential for iron homeostasis

Given that a majority of total body iron is found in red cells, the reclamation of red cell iron has a profound effect on iron homeostasis. Central to such reclamation is heme oxygenase-1 (Hmox1) which catalyzes the oxidation of heme to carbon monoxide, iron and biliverdin. Hmox1 deficiency leads to increased perinatal lethality and, in adult mice, anemia, hypoferremia and hepatic and renal iron loading [80]. Recent findings suggest that Hmox1 haploinsufficiency in mice leads to retention of erythroblasts in erythroblastic islands of the spleen [81] and that Hmox1 deficiency leads to decreased viability of erythrophagocytic macrophages, most likely due to heme toxicity [82].

Once iron is liberated from heme, it must be exported from phagocytic compartments; this step is thought to be mediated, in part, by Nramp1, a homolog of Dmt1 expressed in the membranes of late endosomes and phagolysosomes in phagocytic cells. Nramp1-deficient mice demonstrate increased transferrin saturation, spleen iron levels and duodenal Fpn1 and Dmt1 expression and decreased liver iron and hepcidin levels [83]. Upon treatment with the hemolytic agent phenylhydrazine, deficient mice have a profound decrease in Tf saturation and hematocrit, followed by increased tissue iron levels, splenomegaly and reticulocytosis relative to wild-type controls. These results, in conjunction with in vivo studies on ⁵⁹Fe-labeled damaged red cells, demonstrate that Nramp1 is essential for efficient hemoglobin iron recycling.

Once iron is effluxed from phagolysosomal compartments within reticuloendothelial macrophages, Fpn1 mediates its export out of the cell into the plasma for redistribution by Tf. Mice with macrophage-specific Fpn1 deletion develop iron accumulation in liver Kupffer cells and splenic red pulp macrophages, mild anemia and decreased serum iron levels and Tf saturation; iron dextran administration worsens liver and spleen iron accumulation and corrects the anemia, while an iron-deficient diet has the opposite effect [84]. Liu et al. also recently demonstrated that macrophage maturation is perturbed in mice deficient in heme-regulated eIF2α kinase (HRI), a factor that inhibits globin synthesis in heme-limited erythroid precursors; mutant mice display impaired erythrophagocytosis in vitro and in vivo in conditions of chronic hemolytic anemia[85].

1.4.4 Cellular iron homeostasis requires iron regulatory protein function

Iron regulatory proteins 1 and 2 (Irp1 and Irp2) are cytosolic RNA-binding proteins that regulate the stability and translation of multiple mRNAs encoding proteins required for iron transport, iron/sulfur cluster assembly and heme biosynthesis. The essential nature of Irp activity has been demonstrated by the early embryonic lethality of mice deficient in both Irp1 and Irp2 [86]. Using conditional mutants, Galy et al. demonstrated that mice with hepatocyte-specific Irp1 and Irp2 deletions evelop mitochondrial iron deficiency, impaired iron/sulfur cluster assembly and heme biosynthesis and fatal liver failure [87]. The essential role of Irps in iron homeostasis is further highlighted by the finding by Moroishi et al. that F box and leucine-rich repeat protein 5 (Fbxl5), shown to mediate iron-dependent degradation of Irp1 and Irp2 in vitro [88, 89], is essential for appropriate regulation of Irp2; the embryonic lethality and iron overload of Fbxl5-deficient mice is largely rescued by concomitant Fbxl5 and Irp2 deficiency, while hepatocyte-specific deletion of Fbxl5 resulted in decreased Bmp6 and hepcidin signaling and fatal acute liver toxicity when mice were fed an iron-rich diet [90].

1.5 Prospects for the future

The identification of novel genes with roles in iron metabolism is a significant driving force in the field of iron biology. However, many of the current models have not been investigated to the fullest. While commonly performed, measurements of iron levels in blood, serum, tissue or other sites paint a static picture of a dynamic process; while more tedious to execute, radioisotope studies provide essential information regarding metal influx into, efflux from and distribution within a particular compartment. While the lack of embryonic lethality in a mouse model indicates that a particular gene is not essential for early development, a potential role for such genes in early development is typically not addressed, nor is the phenotype of young adult mice always compared to the phenotype of aged mice. Furthermore, testing the effect of acute and chronic stresses in mouse models can also yield invaluable information regarding the temporal role of a particular gene. Structure-function studies, while not as high profile as the establishment of novel mouse models, provide essential information regarding mechanism of function, although the field of iron metabolism has been hampered by the fact that cell lines that recapitulate in vivo phenotypes are not always available. Despite these shortcomings, many of which reflect the challenges inherent to all research studies and environments, we have seen in recent years tremendous advances in our understanding of iron biology. The establishment of new murine models of aberrant iron metabolism and the study of previously established models has been key to this greater understanding.

Highlights.

1. Mouse models have proven invaluable to the study of iron metabolism.

2. Recent models target dietary absorption, hepcidin regulation and cellular metabolism.

3. Optimal use of models involves a combination of genetic and environmental approaches.

Figure 1. Recent advances in animal models of aberrant iron metabolism.

Models of iron metabolism in the enterocyte, hepatocyte, macrophage and erythroid precursor are shown with labels indicating those gene products for which models of deficiency have been recently reported. For simplicity, not all sites of gene/protein function are indicated nor are all factors essential for iron metabolism shown. Red and green fonts indicate that a factor acts in a stimulatory or inhibitory manner respectively. For a more detailed description of gene function, refer to text. In the gastrointestinal tract, gastrin and Atp4a stimulate gastric acidification, thereby aiding food digestion and maintenance of Fe³⁺ solubility. After reduction of Fe³⁺ to Fe²⁺, iron is transported into the enterocyte by Dmt1, the levels of which are respectively increased and decreased by activity of Hif2α and Ndfip1. Within the enterocyte, ferritin stores iron and indirectly regulates the rate of gastrointestinal iron absorption, while iron regulatory proteins Irp1 and Irp2 influence levels of Dmt1 and Fpn1. Hif2α activity leads to increased basolateral membrane levels of Fpn1, which effluxes iron into plasma for oxidation and binding to Tf. In the hepatocyte, diferric Tf and Bmp6 stimulate hepcidin expression via a pathway dependent upon Hfe, Tfr2, Alk2, Alk3 and Hjv. Expression levels of Hjv are increased by activity of neogenin and miR-122 and decreased by activity of Tmprss6. Expression of hepcidin is also influenced by Hif1α, Hif2α, Vhl, Irp2 and Fbxl5. In the macrophage, senescent red blood cells are catabolized. Nramp1 mediates iron efflux from phagolysosomes, while HRI is required for macrophage maturation. Fpn1 effluxes reclaimed iron into the plasma for binding to Tf; Hif2α activity stimulates Fpn1 expression levels. In the erythroid precursor, Fe is obtained by receptor-mediated endocytosis of diferric Tf complexed to Tfr1, the levels of which are regulated by Stat5a/b and Nme1/2. Iron is imported into mitochondria by mitoferrin while Fech catalyzes the incorporation of iron into heme precursors. Flvcr maintains heme at appropriate levels within the erythroid precursor.