We recently reported the hemochromatosis-like phenotype observed in our Usf2 knockout mice. In these mice, as in murine models of hemochromatosis and patients with hereditary hemochromatosis, iron accumulates in parenchymal cells (in particular, liver and pancreas), whereas the reticuloendothelial system is spared from this iron loading. We suggested that this phenotypic trait could be attributed to the absence, in the Usf2 knockout mice, of a secreted liver-specific peptide, hepcidin. We conjectured that the reverse situation, namely overexpression of hepcidin, might result in phenotypic traits of iron deficiency. This question was addressed by generating transgenic mice expressing hepcidin under the control of the liver-specific transthyretin promoter. We found that the majority of the transgenic mice were born with a pale skin and died within a few hours after birth. These transgenic animals had decreased body iron levels and presented severe microcytic hypochromic anemia. So far, three mosaic transgenic animals have survived. They were unequivocally identified by physical features, including reduced body size, pallor, hairless and crumpled skin. These pleiotropic effects were found to be associated with erythrocyte abnormalities, with marked anisocytosis, poikylocytosis and hypochromia, which are features characteristic of iron-deficiency anemia. These results strongly support the proposed role of hepcidin as a putative iron-regulatory hormone. The animal models devoid of hepcidin (the Usf2 knockout mice) or overexpressing the peptide (the transgenic mice presented in this paper) represent valuable tools for investigating iron homeostasis in vivo and for deciphering the molecular mechanisms of hepcidin action.

ron is an essential element that is required for growth and survival. However, in excess, the metal is toxic to the cell. To deal with this problem, elaborate cellular mechanisms have evolved in virtually all organisms to meet the iron needs of the body. In adult mammals, iron homeostasis depends upon regulated absorption of iron by the enterocyte, a highly specialized cell of the duodenum that coordinates dietary iron uptake and transport into the body. In the fetus, the mechanisms involved in placental maternofetal iron transport are also highly efficient and tightly regulated. Indeed, although fetal serum iron concentrations generally exceed those in the maternal circulation, only rare newborns exhibit evidence of iron toxicity. Although considerable progress has been made these last years in identifying key components of the iron transport molecular pathway (for review, see refs. 1–3), important questions are still unanswered. In particular, the molecular nature of the signals that programs the enterocyte cells of the duodenum and the syncytiotrophoblast cells of the placenta to adjust iron absorption and iron transport still remains to be identified.

In this context, we recently identified a putative key component of iron homeostasis that could act as a central player in the programming pathway from body iron stores to iron absorptive cells (4). It concerns a recently characterized mammalian pep-tide exhibiting antimicrobial activity, hepcidin (5, 6). The expression of the gene encoding hepcidin is restricted nearly to the liver (5–7). Whereas only one copy of the gene exists in humans, two duplicated hepcidin genes have been reported in mice (4, 7). The human gene encodes a precursor protein of 84 aa, including a putative 24-aa leader peptide. The human circulating form of hepcidin, described by two research groups in blood (6) and in urine (5), consists of the C-terminal 25 aa of the protein. Like many antimicrobial peptides, the mature hepcidin peptide seems to be highly structured with four intrachain disulfide bonds (5, 6). The link between hepcidin and iron metabolism stemmed from our finding that hepcidin gene expression was totally inhibited in mice exhibiting iron overload consequent to target disruption of the Usf 2 (Upstream Stimulatory Factor 2) gene (4). In these Usf 2 knockout mice, the development of iron overload was strikingly similar to that observed in human hereditary hemochromatosis and in Hfe^–/–mice, the mouse model for hemochromatosis, with preferential loading of iron in parenchymal cells, increased circulating iron and decreased reticuloendothelial (RE) iron (8, 9). Interestingly, whereas our data show that a complete defect in hepcidin expression was linked to progressive tissue iron overload, Pigeon et al. (7) demonstrated that hepcidin gene expression was strongly up-regulated in the liver of iron-overloaded mice. This latter result highlighted the role of hepcidin as a putative iron regulator that could be induced to decrease dietary iron absorption.

pressing hepcidin died perinatally with severe iron-deficiency anemia.

Taken together, our previous data and the results presented in this paper strongly suggest that hepcidin is a secreted factor acting as a signaling molecule required to regulate iron homeostasis. Any pathophysiological situations leading to variation of circulating hepcidin are predicted to produce either iron deficiency or iron overload.

Materials and Methods Animals. All animals used in these experiments were cared for in accordance with criteria outlined in the European Convention for the Protection of Vertebrate Animals. Animals were maintained in a temperature-and light-controlled environment and were given free access to tap water and food (standard laboratory mouse chow, AO3, Usine d’Alimentation Rationnelle, Epinay-S/Orge, France).

Where indicated, iron-dextran was injected s.c. at the dose of 0.5–1 mg of iron per g of body weight (Sigma).

Generation of Transgenic Mice. The Paris and Houston knockout Usf 2 mice have been described (10, 11). Transgenic mice had a mixed genetic background that included contributions from C57BL/6, 129/Sv, and B6D2 strains as described (12). Full-length cDNA of the murine HEPC1 cDNA was amplified by using primers 5′-GGGGGATATCAGGCCTCTGCACAGCAGAACAGAAGG-3′ and 5′-GGGGGATATCAGGCCTCTATGTTTTGCAACAGATACC-3′. The HEPC1 PCR fragment was introduced between the transthyretin (TTR) sequences (consisting of the 3 kb of the mouse TTR regulatory regions 5′ to the cap site, the first exon, first intron, and most of the second exon) and the SV40 small-T poly(A) signal sequence (13). The construct was checked by DNA sequencing. The 4.7-kbp TTR-HEPC1 transgene was separated from the plasmid sequence by digestion with HindIII and used for pronuclear microinjection.

Genotyping of Transgenic Mice by PCR and Southern Blotting. Southern blotting was done as described (4). BamHI-digested DNA was electrophoresed, transferred to a nylon membrane (Hybond-N+, Amersham Pharmacia), and hybridized to a probe corresponding to the 1.7-kbp BglII-HindIII fragment of the TTR plasmid (13). The probe was labeled with [³²P]dCTP by random priming with a commercially available kit (Megaprime DNA labeling systems, Amersham Pharmacia). The 5.3-kbp-labeled fragment corresponds to endogenous TTR gene, and the 4.7kbp-labeled fragment corresponds to the transgene.

For PCR, genomic DNA (0.5–1 fg) was used in 50-fl reactions that included three primers. The TTR-HEPC1 transgene was amplified by using the forward primer 5′-GAGTCAGGAAGTATGTGAGGG-3′ (annealing in intron 1 of TTR) and the reverse primer 5′-AACAGATACCACACTGGGAA-3′ (annealing in HEPC1 cDNA). The endogenous TTR gene was amplified by using the same forward primer 5′-GAGTCAGGAAGTATGTGAGGG-3′ and the reverse primer 5’CTTTCCGGTGATAGACTCTGG-3′, both annealing in intron 1 of TTR. The PCR was performed as follows: 34 cycles (each cycle consisting of 40 s at 94°C, 40sat53°C, and 40 s at 72°C), with an initial denaturation step at 94°C for 4 min and a final elongation step at 72°C for 5 min, in 20 mM Tris·HCl (pH 8.4)/50 mM KCl/0.05% W-1/2 mM MgCl2/0.2 mM each dNTP/0.2 fM each primer, and 4 units of Taq polymerase (GIBCO/BRL). The 705-bp-specific product corresponds to the TTR-HEPC1 trans-gene and the 364-bp product corresponds to the endogenous TTR gene. The reaction was analyzed on 1.5–2% agarose gel containing ethidium bromide. This PCR method for mouse genotyping was found to give the same results as the Southern blot method.

MEDICAL SCIENCES

KO mice exhibit marked iron overloading in the liver and pancreas, iron did not accumulate in these tissues from Houston Usf 2^–/–KO mice. In the spleen, the normal positive reaction with the Perls’ Prussian blue staining observed between the red pulp cells was also clearly present in the Houston KO mice. In contrast, as previously demonstrated, the spleen reticuloendothelial system from Paris Usf 2^–/–KO mice is spared from this iron loading (4).

In agreement with this normal iron metabolism phenotype in the Houston Usf 2-null mice, we found normal amounts of both HEPC1 and HEPC2 transcripts in the liver of these mice by Northern blot and reverse transcription (RT)-PCR (data not shown).

Taken together, these data, which demonstrate that the Houston Usf2-null mice have a normal iron-related phenotype with normal amounts of liver hepcidin, are strongly in favor of the hypothesis that the lack of USF2 protein is not responsible for the increased iron accumulation observed in the Paris Usf 2 KO, and that the absence of hepcidin gene expression could indeed explain the expression of the hemochromatosis-like phenotype in these mice. This supposition led us to conjecture that the reverse situation, namely an increase in hepcidin gene expression, could result in pathologies associated with iron deprivation. To investigate this hypothesis, we created transgenic mice over-expressing hepcidin.

Generation of TTR-HEPC1 Transgenic Mice. Because of the highly liver-specific pattern of hepcidin gene expression (5–7), we used a liver-specific promoter to direct selective expression of hepcidin into hepatocytes. Although there is only one copy of the hepcidin gene in the human genome, two highly related copies, HEPC1 and HEPC2, have been reported in the mouse. So far, however, there is no data assessing differential functional roles for HEPC1 and HEPC2 in mice. In this study, the HEPC1 peptide was chosen rather than HEPC2, because it is the peptide most closely related to that found in humans. Indeed, as shown in Fig. 2A, there is a 76% sequence identity of the human 25-residue peptide with HEPC1 and only a 58% identity with HEPC2.

The transgenic construct was made by introducing a murine PCR-made HEPC1 cDNA fragment between the -3 kb mouse TTR promoter region and the SV40 small-T poly(A) signal (13). The construction is schematized in Fig. 2B. After standard microinjection of the linearized construct, a total of 14 independent transgenic F0 was obtained. Fig. 2C shows the Southern blot of the BamHI-digested genomic DNA from these different F0 mice. The probe revealed a band at 4.7 kbp corresponding to the transgene (Tg) and one at 5.3 kbp corresponding to the transthyretin endogenous gene (End). This latter band was used as the internal reference corresponding to two allele copies. The transgenic F0 mice could be separated into two groups according to their phenotype. The first one corresponded to transgenic mice that were born with a pale skin and died within a few hours after birth (Fig. 2C, newborn mice transgenic for hepcidin— THnb—including THnb 2, 3, 4, 10, 25, and 29). In this case, a transgene band at the expected size and representing more than two copies was always detected. The second group was characterized by the same genomic pattern—i.e., at least two copies of the transgene—and corresponded to three transgenic mouse founders (Fig. 2C, group 2, TH5, TH35, and TH61). These animals all had an unequivocal phenotype with small size, apathy, poor coat, and crumpled skin. This phenotype appeared rapidly after birth, except for TH61, whose phenotype was only clearly established after 2 weeks. The health of TH5 and TH35 founders deteriorated significantly after 2–3 weeks of life. Because hepcidin was hypothesized to impair iron absorption, we sought to rescue these animals by directly administrating iron by means of s.c. injection of iron-dextran. This iron therapy caused

Severe Microcytic Anemia in Transgenic Mice Overexpressing Hepcidin. If hepcidin compromises iron transport, then erythropoiesis should be altered because most iron is directed to hemoglobin synthesis. Therefore, to directly assess the consequences of exogenous hepcidin on erythropoiesis, we performed WrightGiemsa-stained blood smears on newborn transgenic F0 mice and on TH5. Fig. 3 shows that, in comparison to control erythrocytes, erythrocytes from both THnb10 and TH5 exhibited evidence of poikylocytosis, hypochromia, and microcytosis. Most red cell centers were flattened with the hemoglobin being extruded to the periphery, leading to the typical ring aspect of hypochromic red cells. The same erythrocyte abnormalities were found in F1 progenies from TH5, TH35, and TH61. Because of the early lethality and, as a result, the very small amount of plasma available, it was difficult to characterize further the anemia by blood parameters (especially ferritin, which is a good indicator and whose level drops substantially during iron deficiency). On the other hand, total iron deficiency was evaluated in F1 progeny from TH5 by directly measuring iron content in the whole embryo at embryonic day 15.5 (E15.5). The results indicated that there was 4-fold less total iron in a transgenic fetus as compared with a control fetus (5.3 ± 0.4 fg of iron per gram wet tissue in transgenic F1 progeny from TH5 [n = 5] vs. 19.7 ±

2.6 fg in control fetuses [n = 19] P < 0.0001). The classical hematological features of iron deficiency anemia associated with the severe iron deficit in transgenic mice strongly support the idea of hepcidin acting as a negative regulator of iron transport.

Developmental Expression of Hepcidin Genes in Wild-Type Mice. To understand better the reasons for the early postnatal lethality of most of the transgenic animals, we sought to determine the level of murine hepcidin gene expression during liver development in wild-type C57BL/6 mice. As shown in the Northern blot of Fig. 4, from E15.5 to postnatal day 56 (P56), there was no detectable expression of hepcidin genes apart from a strong transient induction at birth (B) and at P1 and P2. Hepcidin expression only reached a high level in adult liver (P56). This pattern of expression was different to that of transthyretin. Indeed, as shown in the lower part of Fig. 4, transthyretin expression was

25 and 29). Very interestingly, we noted in these transgenic mice that the normal hepcidin mRNA induction observed at birth (see the nontransgenic littermates) was totally inhibited.

Transgenic TH5, TH35, and TH61 progenies also expressed the hepcidin transgene at a high level, whereas the endogenous hepcidin gene again was found to be silent in these animals (Fig. 5B). In the nontransgenic pups from the same litter, the endogenous hepcidin gene was, as expected, expressed 1 day after birth in TH5 male offspring, not expressed at day 13 in TH35 female offspring and, more interestingly, not expressed at day 1 in TH61 female offspring. The difference between TH5 and TH61 offspring is that the mother was wild-type in the first case and transgenic with microcytic anemia in the second.

Discussion

The present study provides substantial evidence for a central role for hepcidin in iron homeostasis. We demonstrate that the iron-overload phenotype previously described in Usf2/hepcidindeficient mice is not related to USF2 deficiency and is thus most likely associated with the lack of hepcidin. Indeed, in another Usf2^–/–model (the Houston KO), iron metabolism and hepcidin gene expression are normal. In addition, we show that, when constitutively expressed, hepcidin impairs maternofetal iron transport leading to severe iron deficiency anemia in neonates. This latter observation relies on transgenic mice expressing hepcidin under the control of the TTR promoter. This promoter is known to direct liver-specific expression, but most importantly, this expression starts during liver development. We report here that the endogenous TTR gene is active at E15.5; data in the literature have reported TTR promoter activity as early as E9.5 (15).

The Severity of the Anemia in Hepcidin Transgenic Mice Is Associated with the Fetal Expression of the Transgene. The majority of the transgenic animals we have generated die after birth, presumably from severe anemia. This anemia was revealed by abnormalities of erythrocyte morphology (Fig. 3), and iron deficiency was directly assessed by measuring iron level in transgenic fetuses. Attempts to determine whether some changes occurred in tissue distribution between control and transgenic fetuses were hampered because of the small amount of material available and the very low concentration of iron in fetal fetuses. Today, only three phenotypically affected founders are still alive: the male TH5, and the two females, TH35 and TH61. In fact, Southern blot analysis comparing genomic DNA from parents and progeny showed that TH5, TH35, and TH61 are mosaics (data not shown). This result might explain why these parents survive, whereas their progeny die around birth, and also the low transmission rate of the transgene.

When available, F0 mice, as well as the F1 progeny from TH5, TH35, and TH61, were analyzed for transgenic hepcidin expression (Fig. 5). These transgenic animals, which were all visibly anemic, expressed high levels of hepcidin. For TH5 progeny, hepcidin expression was demonstrated to occur at least at E15.5, a time when pallor was already evident as well as alteration of erythrocyte morphology (data not shown). This result suggests that hepcidin can be correctly processed and secreted in fetal liver. In fact, although we did not detect hepcidin transcription between E15.5 and birth in wild-type mice liver, an earlier fetal expression is conceivable because it was reported by two groups (5, 7) in human liver fetuses between 18 and 24 weeks of gestation. The absence of hepcidin gene expression late in pregnancy is in good agreement with the known high rate of iron transfer from the mother tothe embryoduringthisperiod.Inthe postnatal period, hepcidin gene expression is only transiently detected around birth and then in young adults, i.e., after the period of rapid growth in which active iron absorption is crucial. The transient expression at birth probably reflects birth-associated stress. Indeed, the hepcidin gene was reported to be responsive to different stimuli; in particular, it was associated with inflammation (7).

In any case, the phenotypes of both hepcidin-hyperexpressing (this paper) and hepcidin-deficient mice (4) demonstrate that this peptide plays an essential role in regulating iron transport through both placental and intestinal barriers and is most likely involved at other sites as well, e.g., reticuloendothelial cells.

The anemia of human chronic diseases is particularly interesting because it occurs under conditions such as infection or inflammation, which involve decreased circulating iron, increased reticuloendothelial iron and decreased intestinal iron absorption, all features of iron disorders that could be explained by up-regulation of hepcidin. Fleming and Sly (21) already predicted such an inflammation-induced increase in hepcidin, and this increase is now under investigation with mouse models of inflammation. The putative role of hepcidin in inflammation is reinforced by the interesting finding by Pigeon et al. (7) that lipopolysaccharide, a classical inducer of acute-phase proteins involved in response to inflammation and infection, is able to induce hepcidin gene expression both in vitro and in vivo.

1. Andrews, N. C. (2000) Nat. Rev. Genet. 1, 208–217.
2. Murray, K. F. & Kowdley, K. V. (2001) Pediatrics 108, 960–964.
3. Roy, C. N. & Enns, C. A. (2000) Blood 96, 4020–4027.
   1. Nicolas, G., Bennoun, M., Devaux, I., Beaumont, C., Grandchamp, B., Kahn,
   2. A. & Vaulont, S. (2001) Proc. Natl. Acad. Sci. USA 98, 8780–8785.
4. Park, C. H., Valore, E. V., Waring, A. J. & Ganz, T. (2001) J. Biol. Chem. 276, 7806–7810.
5. Krause, A., Neitz, S., Magert, H. J., Schulz, A., Forssmann, W. G., Schulz-Knappe, P. & Adermann, K. (2000) FEBS Lett. 480, 147–150.
6. Pigeon, C., Ilyin, G., Courselaud, B., Leroyer, P., Turlin, B., Brissot, P. & Loreal, O. (2001) J. Biol. Chem. 276, 7811–7819.
7. Feder, J. N., Gnirke, A., Thomas, W., Tsuchihashi, Z., Ruddy, D. A., Basava, A., Dormishian, F., Domingo, R., Ellis, M. C., Fullan, A., et al. (1996) Nat. Genet. 13, 399–408.
8. Zhou, X. Y., Tomatsu, S., Fleming, R. E., Parkkila, S., Waheed, A., Jiang, J., Fei, Y., Brunt, E. M., Ruddy, D. A., Prass, C. E., et al. (1998) Proc. Natl. Acad. Sci. USA 95, 2492–2497.
9. Sirito, M., Lin, Q., Deng, J. M., Behringer, R. R. & Sawadogo, M. (1998) Proc. Natl. Acad. Sci. USA 95, 3758–3763.
10. Vallet, V. S., Henrion, A. A., Bucchini, D., Casado, M., Raymondjean, M., Kahn, A. & Vaulont, S. (1997) J. Biol. Chem. 272, 21944–21949.

W. F. & Andrews, N. C. (1997) Nat. Genet. 16, 383–386.

MEDICAL SCIENCES