The iron exporter ferroportin/Slc40a1 is essential for iron homeostasis

The iron exporter ferroportin/Slc40a1 is essential for iron homeostasis

Adriana Donovan,1,2 Christine A. Lima,1,3 Jack L. Pinkus,Geraldine S. Pinkus,Leonard I. Zon,1,2,3 Sylvie Robine,and Nancy C. Andrews1,2,3,*

Equipe de Morphogenèse et Signalisation Cellulaires, Institut Curie, Paris, 75248, France *Correspondence: [email protected]


Ferroportin (SLC40A1) is an iron transporter postulated to play roles in intestinal iron absorption and cellular iron release. Hepcidin, a regulatory peptide, binds to ferroportin and causes it to be internalized and degraded. If ferroportin is the major cellular iron exporter, ineffective hepcidin function could explain manifestations of human hemochromatosis disorders. To investigate this, we inactivated the murine ferroportin (Fpn) gene globally and selectively. Embryonic lethality of Fpnnull/null animals indicated that ferroportin is essential early in development. Rescue of embryonic lethality through selective inactivation of ferroportin in the embryo proper suggested that ferroportin has an important function in the extraembryonic visceral endoderm. Ferroportin-deficient animals accumulated iron in enterocytes, macrophages, and hepatocytes, consistent with a key role for ferroportin in those cell types. Intestine-specific inactivation of ferroportin confirmed that it is critical for intestinal iron absorption. These observations define the major sites of ferroportin activity and give insight into hemochromatosis.

Introduction skin and mucosal cells. Homeostasis functions to meet tissue

iron needs while averting the toxicity of excess iron. Hereditary hemochromatosis is a disorder of iron homeostasis Iron homeostasis can be described in a simple manner by (Hentze et al., 2004). Homozygous mutations in any of four considering the actions of six cell types—extraembryonic vis-genes (HFE, TFR2, hepcidin, and HJV) result in similar pheno-ceral endodermal and placental cells that transfer iron from types of varying severity (Camaschella et al., 2000; Feder etmother to fetus, absorptive enterocytes, erythroid precursors, al., 1996; Papanikolaou et al., 2004; Roetto et al., 2003). Pa- recycling macrophages, and hepatocytes. The circulating pool tients have pathological iron deposition in parenchymal cells of iron bound to transferrin is determined by the activities of of the liver, heart, pancreas, and other tissues. Paradoxically, these cells. These can be further categorized as epithelial cells macrophages retain less iron than in normal individuals. Under-that transfer iron across their apical and basolateral memstanding the pathogenesis of hemochromatosis requires in-branes (preplacental and placental cells, enterocytes), cells sight into normal iron homeostasis and its perturbations in that store and release iron as needed (macrophages and hepadisease. tocytes), and cells that utilize iron but do not release it before

Mammalian iron homeostasis is maintained through complex they senesce (erythroid precursors). regulation of tissues that transport, store, and utilize iron. Initial Iron cannot pass through cellular membranes unassisted. To iron stores are established through maternoembryonic and ma-date, two transmembrane transporters have been identified— ternofetal transfer. After birth, dietary iron is absorbed through SLC11A2 (also known as DMT1, Nramp2, and DCT1) and ferenterocytes lining the duodenum. Iron circulates in the blood- roportin (also known as SLC40A1, IREG1, and MTP1) (Abboud stream bound to transferrin and is delivered to sites of utiliza- and Haile, 2000; Donovan et al., 2000; Fleming et al., 1997; tion and storage. Erythroid precursors are the primary consum- Gunshin et al., 1997; McKie et al., 2000 ). SLC11A2 acts as an ers, using iron to produce hemoglobin. Iron is recycled by iron importer, transporting iron into cells. Ferroportin acts as tissue macrophages that phagocytose senescent erythrocytes an iron exporter, transporting iron out of cells. Accordingly, in and disassemble hemoglobin to return the metal to the circula-polarized epithelial cells, SLC11A2 is found on the apical tion. Macrophage recycling is quantitatively important because membrane, and ferroportin is found on the basolateral memthe amount of iron supplied in this way is approximately 20- brane (Canonne-Hergaux et al., 1999; Donovan et al., 2000). fold greater than the amount absorbed through the intestine on SLC11A2 is also found in endosomes in erythroid precursors a daily basis (Hentze et al., 2004 ). Iron in excess of basal needs (Canonne-Hergaux et al., 2001; Fleming et al., 1998; Su et al., is stored in hepatocytes and macrophages. Unlike other met- 1998), which use the transferrin cycle for iron uptake. Ferropor als, iron cannot be eliminated from the body through the liver tin is not expressed in erythroid cells but is present in the plaor kidneys. Iron losses result from bleeding and exfoliation of centa, intestine, reticuloendothelial macrophages, and hepato- cytes (Abboud and Haile, 2000; Donovan et al., 2000; McKie

Table 1. Analysis of offspring from intercrosses of Fpn1null/+ animals et al., 2000).

Fpn+/+ Fpnnull/+ Fpnnull/null

Age Resorbed Total

Studies of animals carrying loss-of-function mutations in

Slc11a2 suggested that it was the major if not the only trans-E7.5 5 7 1 3 16 E8.52 4 1 5 12

membrane importer of nonheme iron (Fleming et al., 1997,

E9.57 9 0 6 22

1998 ). In contrast, evidence for a major role for ferroportin in

>E10.510 21 0 9 40 mammalian iron export has been indirect. Zebrafish embryos

Total 24 41 2 23 90 carrying mutations in a ferroportin ortholog have a defect in

% 26.7 45.6 2.2 25.6

iron transfer from yolk sac to embryo (Donovan et al., 2000). Ferroportin has been shown to have iron export function when Postnatal 65 85 0 150 expressed in Xenopus oocytes (Donovan et al., 2000; McKie et We sacrificed and dissected pregnant mothers at different times postfertilization. al., 2000 ). The ferroportin mRNA contains an iron regulatory At each time point, we counted the total number of embryonic sacs present in

the uterus. Subsequently, each sac was dissected to reveal the embryo. Sacs

element in its 5# untranslated region (Abboud and Haile, 2000;

that were empty were categorized as resorbed. Each embryo was developmentally

Donovan et al., 2000; Liu et al., 2002; McKie et al., 2000). Reg-

staged according to Theiler (1989). The total numbers of embryos of each ulation of gene expression in intestine and macrophages is genotype (Fpn+/+, Fpnnull/+, and Fpnnull/null) and the total number of resorbed consistent with expectations for a major iron exporter (Knutson embryos were counted for each developmental stage. The genotypes of animals

from live births are summarized at the bottom of the table.

et al., 2003; McKie et al., 2000; Muckenthaler et al., 2003). In addition, perturbation of the ferroportin promoter region in mice causes transitory polycythemia (pcm/+) and iron deficiency anemia (pcm/pcm) (Mok et al., 2004 ). These observa

cile this—that macrophage iron retention leads to irontions are all suggestive, but none directly address the roles of restricted erythropoiesis, and a compensatory augmentation of ferroportin in normal iron homeostasis.

intestinal absorption overcomes haploinsufficiency of ferropor-Formally, it was possible that hemochromatosis could result tin in enterocytes. In support of this hypothesis, some patients from aberrant regulation of iron import, iron export, or both. An have been reported to have anemia early in the course ofimportant clue came from the discovery of hepcidin, a circulattheir disease.

ing peptide that has profound effects on iron absorption and The purposes of this study were to determine which cell distribution (Nicolas et al., 2001, 2002). Hereditary hemochro types are dependent upon ferroportin function and to directly matosis is associated with inappropriately low hepcidin levels, test the hypothesis that ferroportin disease results from heteroeither resulting from mutations in hepcidin itself or from mutazygous loss-of-function mutations in ferroportin. We took adtions in genes that must, in some way, control hepcidin expresvantage of strong similarities between murine and human iron sion (Bridle et al., 2003; Kawabata et al., 2005; Muckenthaler metabolism (Andrews, 2000 ). We inactivated the murine ferro et al., 2003; Nemeth et al., 2005; Nicolas et al., 2003; Papani portin gene globally and in selected tissues, through gene kolaou et al., 2004; Roetto et al., 2003).

targeting and homologous recombination in pluripotential em-Recently, insight into hepcidin function came from the obserbryonic stem cells. Our studies show that ferroportin is revation that hepcidin binds to ferroportin and targets it for dequired for both prenatal and postnatal iron homeostasis. Specistruction. The consequent hypothesis is that hemochromatosis fically, we find that ferroportin is required for iron transfer in results from unimpeded iron transport through ferroportin. Clinthe extraembryonic visceral endoderm and for iron export from ical findings could be explained by a greater transfer of iron

enterocytes, macrophages, and hepatocytes.

from absorptive intestinal epithelial cells to the bloodstream and accelerated macrophage iron release. If this is correct, fer-


roportin must be the primary conduit for iron export from both types of cells. The experiments described in this report test

Targeted disruption of ferroportin results

that interpretation and provide confirmation that it is correct.

in embryonic lethality

The role of ferroportin in vivo has been confounded by the To generate a ferroportin (Fpn) knockout allele, we replaced finding that mutations in ferroportin occur in a subset of human exon 6 and part of exon 7 (amino acids 172–338) with a neopatients with iron overload. Individuals heterozygous for a varimycin resistance cassette using homologous recombination in ety of missense mutations in ferroportin develop macrophageJ1 embryonic stem (ES) cells (see Figure S1A with the Supple-predominant iron overload that may progress to iron-induced mental Data available with this article online). This knockout organ damage (Montosi et al., 2001; Njajou et al., 2001; Piet

strategy deletes sequence that encodes three of the ten pre rangelo, 2004). This has been referred to as autosomal domi

dicted transmembrane segments of the ferroportin protein. We nant hemochromatosis or, more accurately, “ferroportin dis-

identified correctly targeted clones by Southern blot analysis. ease” (Pietrangelo, 2004). If it is true that ferroportin plays

We injected targeted ES cells into blastocysts to obtain chi-important roles in both enterocytes and macrophages, there is

meric animals, which were bred with 129/SvEvTac females to no simple loss-of-function or gain-of-function model to explain

produce F1 offspring carrying the modified ferroportin allele. this disorder. Gain-of-function mutations might lead to the in-We confirmed germline transmission of the targeted allele by creased intestinal iron absorption observed in patients with au-Southern blot analysis (Figure S1B). Animals carrying the tosomal dominant ferroportin disease but would not account targeted allele will be referred to as Fpnnull/+ or Fpnnull/null. for macrophage iron loading. Loss-of-function mutations We interbred Fpnnull/+ animals in an attempt to produce would explain the large amount of iron observed in patient ma-Fpnnull/null offspring. However, we obtained no Fpnnull/null anicrophages but would not readily account for increased intesti-mals among 150 pups genotyped, indicating that Fpnnull/null nal iron absorption. A hypothesis has been set forth to recon- animals died prenatally (Table 1). We determined that most

Figure 1. Embryonic expression of ferroportin protein Immunohistochemical staining for ferroportin in a wild-type E7.0 embryo (A). The brown staining indicates the expression of ferroportin in the extraembryonic visceral endoderm (arrow). At higher power ([A], inset), it is evident that these cells are a polarized epithelium with ferroportin expressed on the basolateral surface, as it is in duodenal enterocytes of the intestine. A littermate of the embryo in (A) that appears significantly smaller and has no detectable ferroportin protein by immunohistochemistry (B). The arrow points to the extraembryonic visceral endoderm.

Fpnnull/null animals die by E7.5. However, we were able to identify one Fpnnull/null embryo at E7.5 and one at E8.5 (Table 1). We reproduced these results in animals carrying a targeted ferroportin allele from which the neomycin cassette had been removed (data not shown).

Role for ferroportin in extraembryonic visceral endoderm

We hypothesized that Fpnnull/null embryos died from a defect in iron transfer from the mother. The extraembryonic visceral endoderm (exVE) functions in maternoembryonic nutrient transport prior to placenta formation (Bielinska et al., 1999). We examined the expression of ferroportin in early embryos by immunohistochemistry. In wild-type embryos, ferroportin was strongly expressed at the basolateral membrane of polarized epithelial cells in the exVE (Figure 1 A). Analogous studies of littermates from intercrosses of Fpnnull/+ animals identified em bryos with no ferroportin expression in the exVE (Figure 1B). These animals were significantly smaller than their ferroportinexpressing littermates and were presumed to be Fpnnull/null embryos. Considering the known function of ferroportin in cellular iron export, it appears that Fpnnull/null embryos cannot transfer iron from the exVE into the embryo proper, leading to a defect in embryonic growth and consequent death.

Fpnnull/+ animals have a mild disruption of iron homeostasis


Although animals died early in development, Fpnnull/+ animals were viable. If simple loss-of-function mutations in ferroportin cause autosomal dominant iron overload in the human disease, then Fpnnull/+ mice should be a model of that disorder. To test this hypothesis, we analyzed parameters of iron homeostasis in Fpnnull/+ animals as compared to wild-type littermates. At age 3 months, Fpnnull/+ animals were indistinguishable from controls in all measurements. However, by age 6 months, we detected evidence of disrupted iron homeostasis in Fpnnull/+ animals. Fpnnull/+ mice were not anemic, but both reticulocytes and mature erythrocytes had decreased cellular hemoglobin and decreased cell volume, indicative of iron-restricted erythropoiesis. The erythrocyte cellular hemoglobin content was 15.0 ± 0.2 pg in Fpn+/+ males (n = 9) and 14.1 ±

0.2 pg in Fpnnull/+ males (n = 5). The difference was highly significant (p value % 0.00001). The erythrocyte mean cell volume was 51.4 ± 0.9 fL in Fpn+/+ males (n = 9) and 48.7 ± 1.4 fL in Fpnnull/+ males (n = 5) (p value % 0.01). Six-month-old Fpnnull/+ male mice also had significantly less nonheme iron in the liver and showed a trend toward lower spleen iron when compared to wild-type littermates (liver: Fpnnull/+ 158.8 ± 64.9 fg/g, Fpn+/+ 243.3 ± 45.8 fg/g, p value % 0.04; spleen: Fpnnull/+ 1656.8 ± 411.8 fg/g, Fpn+/+ 2085.3 ± 449.1 fg/g, p value = 0.1).

We considered the possibility that more time was needed for the iron-loading phenotype to develop. At age 12 months, Fpnnull/+ mice continued to have significantly lower erythrocyte cellular hemoglobin and mean cell volume as compared to wild-type littermates (data not shown). In addition, liver iron content remained lower in Fpnnull/+ mice (116.2 ± 26.8 fg/g, n = 20) than in Fpn+/+ littermates (178.3 ± 33.1 fg/g, n = 14) (p value % 0.00001). However, spleen iron content in Fpnnull/+ animals was greater (2259.24 ± 392.9 fg/g, n = 20) than in Fpn+/+ animals (1797.16 ± 289.6 fg/g, n = 14) (p value % 0.0005).

Most nonheme iron in the spleen is contained in macro-phages, while nonheme iron in the liver is contained in both hepatocytes and macrophages. At 1 year of age, increased splenic iron suggests that iron is being retained in macro-phages in Fpnnull/+ animals to a greater degree than in wild-type controls. This is similar to the human iron overload disease due to ferroportin mutations. However, in contrast to human patients, hepatic iron stores were diminished. Thus, our data suggest that Fpnnull/+ mice are not a faithful model of the human disease.

Conditional disruption of ferroportin

In order to study the role of ferroportin in adult mouse tissues, we retargeted the ferroportin locus and introduced LoxP sites into introns flanking exons 6 and 7 to generate a “floxed” allele (Figure S1A). We refer to mice that are homozygous for the floxed allele as Fpnflox/flox. In the presence of Cre recombinase, sequences between the LoxP sites can be excised. This condi-tional knockout strategy deletes sequence that encodes six of the ten predicted transmembrane segments of the ferroportin protein.

Ferroportin is required for iron homeostasis after birth

To circumvent embryonic lethality due to loss of ferroportin in tissues involved in maternoembryonic iron transfer, we bred Fpnflox/flox animals to Meox2-Cre mice (Tallquist and Soriano, 2000 ). Meox2-Cre;Fpnflox/flox mice should express Cre recombinase and inactivate ferroportin in all tissues except exVE and placenta. Meox2-Cre;Fpnflox/flox mice were born alive and initially appeared identical to their wild-type littermates. However, within the first few days of life, they were noted to be pale and runted. By P10–P12, the difference between wild-type and mutant animals was pronounced (Figure 2A). We observed variability in both the onset and severity of the phenotype, likely due to incomplete Cre-mediated excision in some animals. Accordingly, genotyping of animals with severe phenotypes demonstrated complete or near-complete excision of the floxed allele in all tissues and organs examined, whereas excision was incomplete in animals with milder phenotypes (data not shown).

Meox2-cre;Fpnflox/flox mice were anemic, explaining their pallor. Blood smears revealed hypochromia, anisocytosis, poi kilocytosis, and reticulocytosis (Figures 2B and 2C), similar to other mouse mutants with severe iron deficiency anemia. Measurement of erythrocyte parameters revealed decreased total hemoglobin, mean cell volume, and mean cell hemoglobin in both reticulocytes and mature erythrocytes (Table S1). The red cell distribution width and reticulocyte count were increased. These data strongly support the conclusion that the anemia in Meox2-Cre;Fpnflox/flox mice resulted from iron deficiency.

We examined tissues suspected to require ferroportin function including duodenum, liver, and spleen. Because ferroportin functions as an iron exporter, loss of its activity should result in cellular iron retention, demonstrable by histochemical stains that detect nonheme iron. We compared tissues from Meox2Cre;Fpnflox/floxmice and wild-type littermates. Wild-type animals had no stainable nonheme iron in duodenal enterocytes (Figure 3A). In contrast, Meox2-Cre;Fpnflox/flox mice had abun dant enterocyte iron (Figure 3B), suggesting that iron taken up from the diet or plasma cannot be exported across the basolateral membrane into the circulation. Considering the substantial requirement for efficient dietary iron absorption to support growth in the first 3 weeks of life, this intestinal defect might, by itself, explain the development of iron deficiency anemia in the mutant animals.

We also examined tissue sections from the liver and spleen.

Figure 2. Conditional deletion of ferroportin by Meox2-Cre causes iron defi

Wild-type liver samples showed no detectable iron in hepato

ciency anemia cytes or Kupffer cells at age 12 days. In contrast, Meox2

The 10-day-old Meox2-cre;Fpn1flox/flox pup ([A], bottom) is small and pale comCre;Fpnflox/flox mice showed pronounced iron accumulation in pared to a wild-type littermate ([A], top). Peripheral blood smear of a 12-day-old Kupffer cells, the resident macrophages of the liver, and in he-wild-type animal stained with Wright-Giemsa (B). Peripheral blood smear of a

12-day-old Meox2-Cre;Fpn1flox/flox animal stained with Wright-Giemsa (C).

patocytes (Figures 3C and 3D). We also observed considerable iron accumulation in splenic macrophages of Meox2-Cre; Fpnflox/flox mice but not controls (Figures 3E and 3F). These

observations support the hypothesis that ferroportin is impor- as compared to their wild-type littermates (Figure 4), consistent tant for export of iron from both macrophages and hepato-with the histochemical staining. At 18–22 days of age, Meox2cytes. Cre;Fpn mutant animals continued to have higher spleen iron

We measured nonheme tissue iron content. Meox2-Cre;Fpn (4.4-fold), but their liver iron content was 1.7-fold lower than mutant animals sacrificed at 10–12 days of age had 2.3-and their wild-type littermates (Figure 4). This pattern is most likely 1.7-fold more nonheme iron in liver and spleen, respectively, explained by the overall iron deficiency of these animals. The

Figure 3. Meox2-Cre;Fpn mutant mice accumulate iron in enterocytes, macrophages, and hepatocytes

Perls Prussian blue stain for iron in crosssections of duodenal villi of 12-day-old mice (A and B). Blue staining shows iron accumulated in enterocytes. DAB-enhanced Perls stain for iron in liver sections from 12-day-old mice (C and D). Brown staining shows iron accumulated in cells. The asterix (*) marks a Kupffer cell, and the arrow points to hepatocytes that are loaded with iron. DAB-enhanced Perls stain for iron in spleen sections of 12-day-old mice (E and F). The genotypes of the mice are the following: Fpnflox/flox (A, C, and E) and Meox2-Cre;

Fpnflox/flox (B, D, and F).

liver continues to grow in size with increasing age, but iron-carrying an intestine-restricted villin-Cre transgene that is in-deficient animals cannot deposit additional iron in hepatocytes, ducible by tamoxifen (el Marjou et al., 2004). The resulting vileventually leading to a decrease in liver iron content. Cre-ERT2;Fpnflox/flox mice should be born with ferroportin alleles intact. However, after administration of tamoxifen, Cre Ferroportin is important in intestinal iron absorption recombinase causes tissue-specific inactivation of ferroportin. We administered tamoxifen to vil-Cre-ERT2;Fpnflox/flox and

We wanted to develop mice expressing ferroportin in all tissues Fpnflox/flox mice starting at age 1 week. We confirmed by both

except the intestine in order to distinguish the role of the pro-PCR and Southern blot analyses that inactivation of ferroportin

tein in the intestine from its roles in other cell types. Initially, we occurred exclusively in the intestine (Figure 5A and data not bred Fpnflox/flox animals to mice expressing Cre recombinase shown). Approximately 50% of the total DNA in a duodenal under the control of the villin promoter (el Marjou et al., 2004).

segment showed ferroportin deletion. Based on earlier charac-In adult animals, villin expression is highly restricted to the interization of the vil-Cre-ERT2 mouse model, it is likely that the testine and, to a lesser extent, the kidney. However, similar to observed 50% deletion in whole intestine corresponds toour first, global knockout model, no ferroportin-deleted mice 100% deletion in crypt cells and enterocytes, with no deletion were born. Because intestinal ferroportin should not be re

in nonepithelial cell types (el Marjou et al., 2004). In order to quired before birth, we suspected that the villin-Cre transgene

confirm that ferroportin was not expressed in vil-Cre-ERT2;

was also expressed in the exVE, resulting in early deletion of Fpnflox/flox mice, we performed immunohistochemistry with an ferroportin and embryonic death. Subsequent studies con-anti-ferroportin antibody. Ferroportin is clearly expressed on firmed that the villin promoter directs expression of Cre recom-the basolateral surface of enterocytes in the duodenum of

binase in exVE (el Marjou et al., 2004). Fpnflox/flox mice, but is absent in vil-Cre-ERT2;Fpnflox/flox mice To bypass this problem, we bred Fpnflox/flox mice to mice(Figures 5B and 5C).

Figure 4. Meox2-Cre;Fpn mutant mice have abnor

mal nonheme iron levels in liver and spleen Box plot depicting the measurement of nonheme iron levels (fg Fe/g wet weight) in liver and spleen of wild-type controls (pink boxes) and Meox2-Cre; Fpn mutants (white boxes). The genotypes of the Meox2-Cre;Fpn mutant mice were Meox2-Cre; Fpnflox/flox and Meox2-Cre;Fpnflox/null. The genotypes of the wild-type controls were Fpnflox/flox,

, Fpnflox/+and Fpnflox/null. We analyzed animals at two different ages: (1) 10-to 12-day-old mice (wild-type n = 6 and mutantn=7)and (2) 18-to 22-day-old mice (wild-type n = 13 and mutant n = 11). This data was collected from both male and female animals. p values were calculated using Student’s T-test. The box plot was drawn using Statview 5.0.1 (SAS Institute, Inc.) The middle bar of the box represents the median. The top of the box is the 75th percentile of the data set, and the bottom of the box is the 25th percentile. The top whisker represents the 90th percentile of the data, and the bottom whisker represents the 10th percentile of the data. The circles represent data points that lie outside of the 10th and 90th percentiles.

The vil-Cre-ERT2;Fpnflox/flox mice were visibly anemic as early Cre-ERT2;Fpnflox/flox mice had a mean hematocrit of 41.7% ±
as 8 days after the first tamoxifen injection. At 6–7 weeks of 4.5% (n = 3) (p value % 0.002). These results confirm that ferro
age, peripheral blood smears were consistent with severe iron portin is critical for intestinal absorption of iron.
deficiency anemia (Figures 6A and 6B). The average hematocrit
of vil-Cre-ERT2;Fpnflox/flox animals was 11.2% ± 1.6% (n = 9), Discussion
compared to 45.3% ± 5.9% (n = 11) in Fpnflox/flox controls (p
value % 0.00001). The average hemoglobin concentration of Our findings indicate that ferroportin is the major (if not only)
mutant mice was 2.8 ± 0.6 g/dl (n = 9), compared to 13.8 ± 1.8 basolateral iron exporter functioning in epithelial cells of the
g/dl (n = 11) in wild-type littermates (p value % 0.00001). The exVE and the intestine. When ferroportin is inactivated globally,
duodenal enterocytes of vil-Cre-ERT2;Fpnflox/flox animals there is an early failure in embryonic development, likely due
showed marked iron accumulation (Figures 6C and 6D). How- to iron insufficiency. When it is preserved in the exVE and pla
ever, there was no stainable iron in the liver or spleen of vil centa but inactivated in all other tissues, prenatal development
Cre-ERT2;Fpnflox/flox animals (data not shown). The vil-Cre proceeds normally, but iron deficiency ensues rapidly after
ERT2;Fpnflox/flox mice had 4.7-fold less nonheme iron in liver birth, when the intestine becomes the only route for iron entry.
and 5.5-fold less nonheme iron in spleen as compared to their Selective inactivation of ferroportin in postnatal intestine re
Fpnflox/flox littermates (liver: vil-Cre-ERT2;Fpnflox/flox 23.0 ± 3.9 sulted in severe iron deficiency that was corrected by paren
fg/g, Fpnflox/flox 106.9 ± 6.3 fg/g, p value % 0.0003; spleen: teral administration of iron, confirming that ferroportin function
vil-Cre-ERT2;Fpnflox/flox 27.0 ± 5.8 fg/g, Fpnflox/flox 148.3 ± 25.9 is necessary for dietary iron absorption in mice. While it re
fg/g, p value % 0.013). This severe iron deficiency is consis mains possible that there are alternative, minor pathways for
tent with an important role for ferroportin in intestinal iron ab enterocyte iron export, the severity of the phenotype suggests
sorption. When ferroportin is inactivated exclusively in the in- that none can substitute for the activity of ferroportin.
testine but intact in other tissues, iron is not retained in The requirement for ferroportin activity in the intestine has
macrophages and hepatocytes, consistent with an appropriate additional implications for iron homeostasis. The undifferenti
response to iron deficiency. ated crypt cells that develop into enterocytes, and perhaps the
If anemia and iron deficiency in vil-Cre-ERT2;Fpnflox/flox ani enterocytes themselves, acquire iron from the circulation (Be
mals result from a block in intestinal iron absorption, it should dard et al., 1976). The amount of plasma iron entering these
be possible to restore normal iron balance through parenteral cells is not insignificant, because the surface area of the villous
treatment with iron dextran. Accordingly, after administration epithelium is immense. If export of iron is interrupted, the flow
of iron dextran, the hematocrits of vil-Cre-ERT2;Fpnflox/flox of iron into this compartment becomes unidirectional. Entero
mice returned to normal within 1 week. The hematocrits of cytes have a limited lifespan—after several days, they senesce
Fpnflox/flox animals before and after treatment ranged from 46% and are sloughed into the gut lumen, resulting in loss of any
to 49%. Before treatment, the hematocrits of vil-Cre-ERT2; iron they contain. Thus, we would expect that animals lacking
Fpnflox/flox animals ranged from 9% to 11%. One week later, intestinal ferroportin function would not only have a defect in
untreated vil-Cre-ERT2;Fpnflox/flox mice still had a mean hema iron absorption—they would also have increased iron losses.
tocrit of 8.9% ± 2.2% (n = 3), while the iron dextran-treated vil- While it has been assumed that there is no regulated iron

excretion because neither the liver nor the kidney efficiently releases iron from the body, our results suggest that iron loss could be regulated at a different level. In iron deficiency, increased ferroportin expression would not only ensure increased intestinal absorption but also allow for maximal recovery of iron that has entered the intestinal epithelium from the plasma. On the other hand, decreased ferroportin expression in iron overload could augment normal iron losses by increasing the amount of iron contained in exfoliated enterocytes. Thus, regulation of ferroportin might be considered a novel mechanism of regulating iron “excretion.” This need not occur through a mechanism that regulates the production of ferroportin. Indeed, hepcidin can exert control of ferroportin by modu lating protein levels (Nemeth et al., 2004). This homeostatic mechanism allows regulation of intestinal iron absorption in response to iron availability and needs elsewhere in the body.

We have shown that ferroportin also plays important roles in nonepithelial cells. Global inactivation of ferroportin resulted in other perturbations of cellular iron distribution. Macrophages of the liver and spleen showed marked iron retention, likely because they had phagocytosed effete erythrocytes but had no mechanism for releasing recovered iron. Hepatocytes also appeared to retain the iron that they had acquired prenatally. These observations suggest that, as in enterocytes, ferroportin is the major iron exporter functioning in iron-recycling macrophages and hepatocytes. Iron retention by those cells undoubtedly compounds the effects of inadequate iron absorption, further depriving erythroid precursors of iron for hemoglobin production and thus exacerbating the anemia.

Few other cell types need to have an iron export function. However, iron transfer across the blood-brain barrier does require a cellular export step. We have not directly addressed the role of ferroportin in the brain in this study. However, in a separate study, we bred with mice expressing Cre recombinase un der the control of a nestin promoter (Tronche et al., 1999) and did not note any gross behavioral abnormalities that might be attributed to cerebral or cerebellar dysfunction (data not shown).

Mice heterozygous for ferroportin mutations have a subtle phenotype that is not measurable before 6 months of age. Similar to human patients with autosomal dominant ferroportin disease, they accumulated iron in tissue macrophages and showed mild evidence of iron-restricted erythropoiesis. However, even at 1 year of age, they did not appear to have increased liver iron that would have been indicative of total body iron overload.

There are several possible explanations for the differences between Fpnnull/+ mice and human patients. First, it might be that clinical iron overload requires more time, and 1 year is not long enough for it to become manifest. This seems unlikely, because liver iron content was consistently decreased in Fpnnull/+ mice. Alternatively, it might be that mice and humans

Figure 5. Conditional deletion of ferroportin in enterocytes by vil-Cre-ERT2 respond differently to heterozygous loss-of-function of this

PCR analysis of tissue DNA from a vil-Cre-ERT2;Fpnflox/flox mouse that was injected with tamoxifen (A). Tail (T), liver (L), spleen (S), kidney (K), lung (Lu), pan-gene. We favor a third hypothesis—that human ferroportin mucreas (P), heart (H), intestine (I). The marker is a 100 bp ladder. The floxed allele tations are not simple loss-of-function mutations and therefore is 522 bp, and the full deletion allele is 398 bp. Immunohistochemistry for ferro

not directly analogous to the situation in Fpnnull/+ mice. Mutaportin protein in a wild-type animal (B) and a vil-Cre-ERT2;Fpnflox/flox mouse in

tions may not affect iron transport, per se, but rather perturb

jected with tamoxifen (C). Brown staining represents ferroportin expression in

critical protein-protein interactions that might be cell type spe

duodenal enterocytes. Higher magnification images show basolateral localization

of ferroportin in the wild-type animals ([B], inset) and absence of ferroportin ex- cific. Accordingly, it is interesting to note that no frameshift or

pression in the vil-Cre-ERT2;Fpnflox/flox mouse ([C], inset). termination codon mutations have been found among the many patients analyzed to date.

Figure 6. Ferroportin is critical for intestinal iron ab

sorption Wright-Giemsa-stained peripheral blood smears of tamoxifen-injected Fpnflox/flox (A) and vil-Cre-ERT2; Fpnflox/flox (B) animals. Perls stain for nonheme iron in the duodenum of tamoxifen injected Fpnflox/flox (C) and vil-Cre-ERT2;Fpnflox/flox (D) animals. Blue staining represents nonheme iron accumulation in duodenal enterocytes.

These studies also explain why hepcidin deficiency results bred with 129/SvEvTac females to generate F1 offspring carrying the modified ferroportin alleles. Germline transmission of the modified ferroportin

in clinical hemochromatosis. Hepcidin acts as a modulator of

alleles was determined by Southern blot analysis of tail DNA samples. For

ferroportin function by triggering its degradation (Nemeth et al.,

the conditional knockout model, we removed the neomycin resistance cas

2004 ). Normally, increased iron availability stimulates hepcidin

sette from the targeted allele by mating to transgenic mice expressing Cre expression, probably triggering a homeostatic response in recombinase under control of the E2A promoter (Lakso et al., 1996). The which hepcidin binds to ferroportin to cause its degradation. E2A-Cre transgenic line had a 129/SvEvTac background.

We can conclude that this is effective in controlling iron homeo-


stasis because we have shown that ferroportin gates iron re-Mice were maintained on standard mouse diet. Both the Fpn null and Fpn

lease at two key sites—in absorptive intestinal epithelial cells

floxed lines were maintained on the 129/SvEvTac background. For timed

and recycling macrophages. We propose that, in hemochro

matings of Fpnnull/+ animals, the morning a plug was identified was considmatosis, perturbations in iron homeostasis result from a failure

ered E0.5. Plugged females were then sacrificed at E6.5 and later. Embryos to control iron export through ferroportin, leading to increased were dissected from the uterus and either genotyped or used for immunoiron release from macrophages and enhanced intestinal ab-histochemistry. The Meox2-Cre mouse strain was obtained from the Jack-sorption. son Laboratory (Bar Harbor, ME) on a mixed B6.129S4 background

(Tallquist and Soriano, 2000). The vil-Cre-ERT2 transgenic line was obtained

on a mixed background of B6/D2 and BL6 (el Marjou et al., 2004) . The

Meox2-Cre and vil-Cre-ERT2 mice were mated to Fpnflox/+ animals to create Targeted mutagenesis of ferroportin Meox2-Cre; Fpnflox/+ and vil-Cre-ERT2; Fpnflox/+mice. Since Meox2-Cre can

Experimental Procedures

Both the global and conditional targeting constructs modified a 15.5 kb cause deletion in the germline, we also obtained Meox2-Cre;Fpnnull/+ mice KpnI fragment containing exons 4–8 of the mouse ferroportin genomic lo-from these crosses. Mice with the three genotypes described above were cus (Figure S1A). We used the pTKLNCL vector (originally from R. Morten- subsequently mated to Fpnflox/flox animals to create the animals that were son). For the traditional targeting construct, the XhoI/BamHI fragment con-studied in these experiments. Administration of tamoxifen to Fpnflox/flox and taining exon 6 and part of exon 7 was replaced by a 4.5 kb LoxP flanked vil-Cre-ERT2; Fpnflox/flox mice was performed once daily for 4 days starting cassette encoding a neomycin resistance gene and a cytosine deaminase between 7 and 10 days after birth. Tamoxifen (Sigma, St. Louis, MO) was gene. For the conditional targeting construct, the LoxP flanked cassette first dissolved in 100% ethanol (50 mg/ml). Injection solution was prepared was inserted into an XhoI site in intron 5. A third LoxP site was inserted into by diluting the tamoxifen to 8.3 mg/ml in sunflower seed oil (Sigma). Mice an SphI site in intron 7 (Figure S1A). Coding sequence for thymidine kinase were weighed and injected with an amount of tamoxifen equal to 0.075 was also present in the pTKLNCL vector outside of the homology regions. mg/g mouse. Mice were injected subcutaneously into the subscapular re-

The targeting constructs were introduced by electroporation into J1 embry-gion with a single dose of either 1 mg or 5 mg of Fe-dextran (Watson onic stem cells (129 background), and cells were selected for resistance to Pharma, Morristown, NJ). G418 and ganciclovir. DNA samples from clones resistant to both G418 and ganciclovir were examined for homologous recombination by Southern blot Southern blot and PCR genotyping analysis of HindIII digested DNA using a probe to exon 3 (Figures S1B and For Southern blot analysis, 10 fg of genomic DNA (Puregene Kit, Gentra S1C). For each targeted allele (total knockout and conditional knockout), a Systems, Minneapolis, MN) was digested overnight with either HindIII or single homologously recombined clone with a normal karyotype was in-BamHI restriction enzyme, fractionated on a 0.5% TAE agarose gel, and jected into C57BL/6J mouse blastocysts, and these were transferred into transferred to Hybond N+ membrane (Amersham Biosciences, Piscataway, the uteri of pseudopregnant females. High percentage male chimeras were NJ). The probe for Southern blots was labeled by incorporation of 32P-dCTP into a PCR product amplified with the following primers: forward (Bayer, Tarrytown, NY) with software specialized for mouse blood. Blood 5#-GCAATTTATTGGGTATAGCAG-3# and reverse 5#-TGCCACCACCAGTCC analysis was performed in the Clinical Core Laboratories at Children’s HosATAGAC-3#. We genotyped mice by PCR analysis. Fpnnull/+ mice and em-pital Boston. bryos from intercrosses were genotyped using two PCR reactions. The wild-type allele (355 bp) was identified by amplification with the forward Statistics primer 5#-CTACACGTGCTCTCTTGAGAT-3# (Xho5#) and the reverse primer We calculated all p values using Student’s t test in Microsoft Excel. 5#-GGTTAAACTGCTTCAAAGG-3# (Xho3#). The targeted allele (359 bp) was identified by amplification with the same forward primer (Xho5#) and a reverse primer, 5#-GCCAAGTTCTAATTCCATCAG-3#, derived from the Supplemental data targeting vector. The Fpnflox/+ and Fpnflox/flox mice were genotyped with one Supplemental Data include one figure and one table and can be found with PCR reaction that had one forward primer (Xho5#) and two reverse primers, this article online at and 5#-CCTCATATGTGAGTCAAAGTATAG-3#. The wild-type allele DC1/. generated a 355 bp band, and the floxed allele generated a 522 bp band. The band generated by full deletion with Cre recombinase was 398 bp. The Meox2-Cre mice were genotyped with primers that amplify Cre recombi-Acknowledgments nase (200 bp): CREF (5#-CGTATAGCCGAAATTGCCAG-3#) and Cre-2-1 (5#-CAAAACAGGTAGTTATTCGG-3#). The vil-Cre-ERT2 mice were geno-This work was supported by P01 HL32262 (N.C.A.) and K01 5 K01 typed with the primers villin Cre-F (5#-CAAGCCTGGCTCGACGGCC-3#) DK64924-02 (A.D.). N.C.A. is an investigator of the Howard Hughes Medical and villin Cre-R (5#-CGCGAACATCTTCAGGTTCT-3#), resulting in a 300 bp Institute. We thank Margaret Thompson and the Children’s Hospital Mental band. Retardation Research Center Gene Manipulation Facility (NIH grant

NIHP30-HD 18655) for ES cell transfections and blastocyst injections. We Immunohistochemistry thank Vonnie Lee for technical assistance in the early stages of this project, Mark Fleming for help with immunohistochemistry, and Cameron Trenor for

Immunohistochemistry of embryos was performed with a rabbit polyclonal assistance with characterization of the mouse ferroportin genomic locus.

anti-mouse ferroportin antibody elicited by injecting a fusion protein of glu-We also thank Heiner Westphal for E2A-Cre transgenic mice and François tathione S-transferase and 80 amino acids (residues 224–304) of mouse Canonne-Hergaux and Philippe Gros for providing anti-ferroportin antiseferroportin protein into a New Zealand White rabbit. Antiserum was affinity rum. The authors have no competing financial interests. The following are author contributions: A.D. and L.I.Z. enabled the knockout project by clonpurified against the same ferroportin peptide segment fused to dihydrofo

late reductase using a preparative immunoblot procedure, as previously deing and sequencing the mouse ferroportin genomic locus. A.D. and N.C.A.

scribed (Canonne-Hergaux et al., 1999 ). Specificity to mouse ferroportin conceived and designed the experiments. A.D. performed the experiments

was confirmed by Western blot on membrane preparations from Madinto create the ferroportin mutant lines. A.D. and C.A.L. analyzed the knock-

Darby canine kidney (MDCK) cells stably expressing ferroportin carrying a out models. S.R. developed the villin-Cre mouse strains. G.S.P. and J.L.P.

single missense mutation fused to green fluorescent protein at the C termiperformed the immunohistochemistry experiments. A.D. and N.C.A. ana

nus (data not shown; F. Canonne-Hergaux, submitted). Embryos in their lyzed the data and wrote the manuscript.

deciduum were fixed in 10% buffered formalin (Fisher Scientific, Fair Lawn, NJ) and embedded in paraffin. Sections were deparaffinized and heat treated for 30 min in 1.0 mM EDTA (pH 8.0) in an HS80 steamer (Black and Decker). Endogenous peroxidase activity was quenched with 3% aqueous Received: November 9, 2004

hydrogen peroxide. Slides were then incubated in 3% normal swine serum Revised: December 23, 2004

in 0.05 M Tris (pH 7.6) followed by rabbit anti-ferroportin antibody (1:200 Accepted: January 6, 2005

dilution) for 16 hr at room temperature. Slides were then incubated for 40 min with a horseradish peroxidase-labeled polymer conjugated to goat anti-Published: March 15, 2005

rabbit immunoglobulin (EnVision+, Dako, Carpinteria, CA). Following antibody localization with 3.3#-diaminobenzidine tetrahydrochloride (DAB+, References Dako), the stain was intensified by using a DAB enhancing solution (Zymed, Abboud, S., and Haile, D.J. (2000). A novel mammalian iron-regulated pro-South San Francisco, CA). Slides were counterstained with methyl green, tein involved in intracellular iron metabolism. J. Biol. Chem. 275, 19906– dehydrated, and covered in Polymount (Poly Scientific, Bay Shore, NY).


Immunohistochemistry on intestine sections was performed with a previously described polyclonal antibody against the C terminus of ferroportin Andrews, N.C. (2000). Iron homeostasis: insights from genetics and animal (Donovan et al., 2000). The same procedure described above was followed, models. Nat. Rev. Genet. 1, 208–217. except that the antibody was incubated at a 1:400 dilution and for 1 hr.

Bedard, Y.C., Pinkerton, P.H., and Simon, G.T. (1976). Uptake of circulating iron by the duodenum of normal mice and mice with altered iron stores,

Tissue iron staining and quantitative tissue iron measurements

including sex-linked anemia: high resolution radioautographic study. Lab.

Intestine, liver, and spleen samples were fixed in 10% buffered formalin and Invest. 34, 611–615.

embedded in paraffin. Deparaffinized tissue sections were stained with the Perls Prussian blue stain for nonheme iron in the Children’s Hospital Boston Bielinska, M., Narita, N., and Wilson, D.B. (1999). Distinct roles for visceral Pathology Laboratory. A subset of tissue sections was stained with a DAB-endoderm during embryonic mouse development. Int. J. Dev. Biol. 43, enhanced Perls iron stain (Dragatsis et al., 1998). Deparaffinized tissue sec- 183–205. tions were incubated for 30 min in 1% potassium ferrocyanide in 0.12N

Bridle, K.R., Frazer, D.M., Wilkins, S.J., Dixon, J.L., Purdie, D.M., Crawford, HCl. Endogenous peroxidase activity was quenched for 20 min at room D.H., Subramaniam, V.N., Powell, L.W., Anderson, G.J., and Ram, G.A.

temperature in 0.3% H2O2 in methanol. Sections were rinsed in PBS and

(2003). Disrupted hepcidin regulation in HFE-associated haemochromatosis incubated for 6 min in DAB/ H2O2 (DAB Plus Substrate Kit, Zymed). Quanti

and the liver as a regulator of body iron homoeostasis. Lancet 361, 669– tative measurement of nonheme iron was performed as described (Levy et

673. al., 1999 ). Data is expressed as fg Fe/g wet weight of tissue.

Camaschella, C., Roetto, A., Cali, A., De Gobbi, M., Garozzo, G., Carella, M., Majorano, N., Totaro, A., and Gasparini, P. (2000). The gene TFR2 is

Blood cell analysis

mutated in a new type of haemochromatosis mapping to 7q22. Nat. Genet.

Mice were bled retroorbitally under anesthesia with 125–240 fg/kg avertin

25, 14–15.

(Sigma). Blood smears were stained with Wright-Giemsa stain (Fisher Diagnostics, Middletown, VA). Blood for complete blood counts was collected Canonne-Hergaux, F., Gruenheid, S., Ponka, P., and Gros, P. (1999). Cellular into 600 fl EDTA anticoagulated microtainer tubes (Becton Dickenson, and subcellular localization of the Nramp2 iron transporter in the intestinal Franklin Lakes, NJ) and analyzed on an ADVIA 120 hematology analyzer brush border and regulation by dietary iron. Blood 93, 4406–4417.

Canonne-Hergaux, F., Zhang, A.S., Ponka, P., and Gros, P. (2001). Charac-dynamic alterations in iron homeostasis and erythropoiesis in polycythaeterization of the iron transporter DMT1 (NRAMP2/DCT1) in red blood cells mia mice. Development 131, 1859–1868. of normal and anemic mk/mk mice. Blood 98, 3823–3830.

Montosi, G., Donovan, A., Totaro, A., Garuti, C., Pignatti, E., Cassanelli, Donovan, A., Brownlie, A., Zhou, Y., Shepard, J., Pratt, S.J., Moynihan, J., S., Trenor, C.C., Gasparini, P., Andrews, N.C., and Pietrangelo, A. (2001). Paw, B.H., Drejer, A., Barut, B., Zapata, A., et al. (2000). Positional cloning Autosomal-dominant hemochromatosis is associated with a mutation in the of zebrafish ferroportin1 identifies a conserved vertebrate iron exporter. Na-ferroportin (SLC11A3) gene. J. Clin. Invest. 108, 619–623. ture 403, 776–781.

Muckenthaler, M., Roy, C.N., Custodio, A.O., Minana, B., DeGraaf, J., Dragatsis, I., Efstratiadis, A., and Zeitlin, S. (1998). Mouse mutant embryos Montross, L.K., Andrews, N.C., and Hentze, M.W. (2003). Regulatory de-lacking huntingtin are rescued from lethality by wild-type extraembryonic fects in liver and intestine implicate abnormal hepcidin and Cybrd1 exprestissues. Development 125, 1529–1539. sion in mouse hemochromatosis. Nat. Genet. 34, 102–107.

el Marjou, F., Janssen, K.P., Chang, B.H., Li, M., Hindie, V., Chan, L., Lou-Nemeth, E., Tuttle, M.S., Powelson, J., Vaughn, M.B., Donovan, A., Ward, vard, D., Chambon, P., Metzger, D., and Robine, S. (2004). Tissue-specific D.M., Ganz, T., and Kaplan, J. (2004). Hepcidin regulates cellular iron efflux and inducible Cre-mediated recombination in the gut epithelium. Genesis by binding to ferroportin and inducing its internalization. Science 306,

39, 186–193. 2090–2093.

Feder, J.N., Gnirke, A., Thomas, W., Tsuchihashi, Z., Ruddy, D.A., Basava, Nemeth, E., Roetto, A., Garozzo, G., Ganz, T., and Camaschella, C. (2005). A., Dormishian, F., Domingo, R., Ellis, M.C., Fullan, A., et al. (1996). A novel Hepcidin is decreased in TFR2-hemochromatosis. Blood 105, 1803–1806. MHC class I-like gene is mutated in patients with hereditary haemochro-Published online October 14, 2004. 10.1182/blood-2004-08-3042

matosis. Nat. Genet. 13, 399–408.

Nicolas, G., Bennoun, M., Devaux, I., Beaumont, C., Grandchamp, B., Fleming, M.D., Trenor, C.C., Su, M.A., Foernzler, D., Beier, D.R., Dietrich,

Kahn, A., and Vaulont, S. (2001). Lack of hepcidin gene expression and W.F., and Andrews, N.C. (1997). Microcytic anemia mice have a mutation in

severe tissue iron overload in upstream stimulatory factor 2 (USF2) knockout mice. Proc. Natl. Acad. Sci. USA 98, 8780–8785.

Nramp2, a candidate iron transporter gene. Nat. Genet. 16, 383–386.

Nicolas, G., Bennoun, M., Porteu, A., Mativet, S., Beaumont, C.,

Fleming, M.D., Romano, M.A., Su, M.A., Garrick, L.M., Garrick, M.D., and Grandchamp, B., Sirito, M., Sawadogo, M., Kahn, A., and Vaulont, S. (2002).

Andrews, N.C. (1998). Nramp2 is mutated in the anemic Belgrade (b) rat: Severe iron deficiency anemia in transgenic mice expressing liver hepcidin.

evidence of a role for Nramp2 in endosomal iron transport. Proc. Natl. Acad.

Proc. Natl. Acad. Sci. USA 99, 4596–4601.

Sci. USA 95, 1148–1153.

Nicolas, G., Viatte, L., Lou, D.Q., Bennoun, M., Beaumont, C., Kahn, A.,

Gunshin, H., Mackenzie, B., Berger, U.V., Gunshin, Y., Romero, M.F., Boron, Andrews, N.C., and Vaulont, S. (2003). Constitutive hepcidin expression W.F., Nussberger, S., Gollan, J.L., and Hediger, M.A. (1997). Cloning and prevents iron overload in a mouse model of hemochromatosis. Nat. Genet.

characterization of a mammalian proton-coupled metal-ion transporter. Na

34, 97–101.

ture 388, 482–488. Njajou, O.T., Vaessen, N., Joosse, M., Berghuis, B., van Dongen, J.W.,

Hentze, M.W., Muckenthaler, M.U., and Andrews, N.C. (2004). Balancing Breuning, M.H., Snijders, P.J., Rutten, W.P., Sandkuijl, L.A., Oostra, B.A., et acts: molecular control of mammalian iron metabolism. Cell 117, 285–297.

al. (2001). A mutation in SLC11A3 is associated with autosomal dominant Kawabata, H., Fleming, R.E., Gui, D., Moon, S.Y., Saitoh, T., O’Kelly, J., hemochromatosis. Nat. Genet. 28, 213–214. Umehara, Y., Wano, Y., Said, J.W., and Koeffler, H.P. (2005). Expression of

Papanikolaou, G., Samuels, M.E., Ludwig, E.H., MacDonald, M.L.,

hepcidin is down-regulated in TfR2 mutant mice manifesting a phenotype Franchini, P.L., Dube, M.P., Andres, L., MacFarlane, J., Sakellaropoulos, N.,

of hereditary hemochromatosis. Blood 105, 376–381.

Politou, M., et al. (2004). Mutations in HFE2 cause iron overload in chromo-Knutson, M.D., Vafa, M.R., Haile, D.J., and Wessling-Resnick, M. (2003).

some 1q-linked juvenile hemochromatosis. Nat. Genet. 36, 77–82. Iron loading and erythrophagocytosis increase ferroportin 1 (FPN1) expres-

Pietrangelo, A. (2004). The ferroportin disease. Blood Cells Mol. Dis. 32,sion in J774 macrophages. Blood 102, 4191–4197.


Lakso, M., Pichel, J.G., Gorman, J.R., Sauer, B., Okamoto, Y., Lee, E., Alt, Roetto, A., Papanikolaou, G., Politou, M., Alberti, F., Girelli, D., Christakis, J.,

F.W., and Westphal, H. (1996). Efficient in vivo manipulation of mouse geno-

Loukopoulos, D., and Camaschella, C. (2003). Mutant antimicrobial peptide mic sequences at the zygote stage. Proc. Natl. Acad. Sci. USA 93, 5860–

hepcidin is associated with severe juvenile hemochromatosis. Nat. Genet. 5865.

33, 21–22.

Levy, J.E., Montross, L.K., Cohen, D.E., Fleming, M.D., and Andrews, N.C.

Su, M.A., Trenor, C.C., Fleming, J.C., Fleming, M.D., and Andrews, N.C. (1999). The C282Y mutation causing hereditary hemochromatosis does not

(1998). The G185R mutation disrupts function of iron transporter Nramp2. produce a null allele. Blood 94, 9–11.

Blood 92, 2157–2163.

Liu, X.B., Hill, P., and Haile, D.J. (2002). Role of the ferroportin iron-respon-Tallquist, M.D., and Soriano, P. (2000). Epiblast-restricted Cre expression in sive element in iron and nitric oxide dependent gene regulation. Blood Cells MORE mice: a tool to distinguish embryonic vs. extra-embryonic gene func-Mol. Dis. 29, 315–326. tion. Genesis 26, 113–115.

McKie, A.T., Marciani, P., Rolfs, A., Brennan, K., Wehr, K., Barrow, D., Miret, Theiler, K. (1989). The House Mouse: Atlas of Embryonic Development (New S., Bomford, A., Peters, T.J., Farzaneh, F., et al. (2000). A novel duodenal York: Springer Verlag). iron-regulated transporter, IREG1, implicated in the basolateral transfer of

Tronche, F., Kellendonk, C., Kretz, O., Gass, P., Anlag, K., Orban, P.C., Bock, iron to the circulation. Mol. Cell 5, 299–309.

R., Klein, R., and Schutz, G. (1999). Disruption of the glucocorticoid recep-Mok, H., Jelinek, J., Pai, S., Cattanach, B.M., Prchal, J.T., Youssoufian, H., tor gene in the nervous system results in reduced anxiety. Nat. Genet. 23, and Schumacher, A. (2004). Disruption of ferroportin 1 regulation causes 99–103.