1532 The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
Ferroportin1 is required for normal
iron cycling in zebrafish
Paula G. Fraenkel,1,2 David Traver,1 Adriana Donovan,1 David Zahrieh,3 and Leonard I. Zon1
1Division of Hematology/Oncology, Children’s Hospital, Karp Research Laboratories, Boston, Massachusetts, USA.
2Division of Hematology/Oncology, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.
3Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, Boston, Massachusetts, USA.
Missense mutations in ferroportin1 (fpn1), an intestinal and macrophage iron exporter, have been identi-
fied between transmembrane helices 3 and 4 in the zebrafish anemia mutant weissherbst (wehTp85c–/–) and in
patients with type 4 hemochromatosis. To explore the effects of fpn1 mutation on blood development and
iron homeostasis in the adult zebrafish, wehTp85c–/– zebrafish were rescued by injection with iron dextran
and studied in comparison with injected and uninjected WT zebrafish and heterozygotes. Although iron
deposition was observed in all iron-injected fish, only wehTp85c–/– zebrafish exhibited iron accumulation in
the intestinal epithelium compatible with a block in iron export. Iron injections initially reversed the ane-
mia. However, 8 months after iron injections were discontinued, wehTp85c–/– zebrafish developed hypochromic
anemia and impaired erythroid maturation despite the persistence of iron-loaded macrophages and elevated
hepatic nonheme iron stores. Quantitative real-time RT-PCR revealed a significant decrease in mean hepatic
transcript levels of the secreted iron-regulator hepcidin and increased intestinal expression of fpn1 in anemic
wehTp85c–/– adults. Injection of iron dextran into WT or mutant zebrafish embryos, however, resulted in sig-
nificant increases in hepcidin expression 18 hours after injection, demonstrating that hepcidin expression in
zebrafish is iron responsive and independent of fpn1’s function as an iron exporter.
According to a current model of iron transport (1), duodenal
enterocytes and placental syncytiotrophoblasts are the principle
cells responsible for iron acquisition in mammals. At the apical
surface of the duodenal enterocyte, the divalent metal transporter 1
(DMT1) mediates uptake of iron from the intestinal lumen into
the enterocyte (2, 3). Absorbed iron may be stored bound to ferritin
or exported across the enterocyte’s basolateral membrane to the
circulation via ferroportin1 (fpn1) (4). In the plasma, iron binds
transferrin and is subsequently imported into erythroid precur-
sors and other cells via transferrin-receptor–mediated endocytosis.
Erythrocytes have no known means of iron export but rather incor-
porate iron into hemoglobin. Macrophages serve as iron scaven-
gers, phagocytosing senescent erythrocytes, degrading hemoglobin,
and storing iron incorporated in ferritin.
Hepcidin, a small cysteine-rich peptide with antimicrobial prop-
erties (5, 6), has recently been identified as a key regulator of iron
absorption and utilization. Produced in the liver, hepcidin cir-
culates in plasma and is excreted in the urine. Hepcidin protein
levels appear to be largely regulated by transcriptional control
(1). Hepcidin null mice develop iron overload (7) while animals
that overexpress hepcidin develop severe hypochromic anemia (8).
Hepatic expression of the hepcidin gene decreases in response to
iron deficiency, hypoxia, and anemia (9) while expression increases
in response to iron overload (10) or inflammation (9, 11).
The zebrafish, Danio rerio, provides an excellent system for the
identification and analysis of genes involved in iron metabolism and
erythroid development. Analogous to higher vertebrates, zebrafish
exhibit multilineage hematopoiesis, resulting in erythroid, mono-
cyte, granulocyte, and thrombocyte lineages, and undergo hemoglo-
bin switching (12). Large-scale genetic screens for embryonic hypo-
chromic anemia have identified recessive mutations with defects in
heme synthesis (13), globin production (12), and iron acquisition
via DMT1 (14) and transferrin receptor (15). Positional cloning of
the mutation responsible for the hypochromic zebrafish mutant
weissherbst (weh) resulted in the identification of the basolateral iron
exporter, fpn1, (4) also known as IREG1 (16) and MTP1 (17), a con-
served protein with 10 putative transmembrane domains and a func-
tional iron response element (18) in the 5′ untranslated region.
Two alleles of the weh phenotype have been described previous-
ly (19). Both have a homozygous recessive pattern of inheritance.
The wehTh238 allele has a premature stop mutation and is thought
to be a null allele (4). The wehTp85c allele is of particular inter-
est because it encodes a missense mutation (Leu 167 to Phe) in
the same conserved region of fpn1 (4) as several of the missense
mutations identified in patients with type 4 hemochromatosis
(20–26). Both wehTp85c and wehTh238 homozygotes develop severe
anemia by 48 hours after fertilization and die between 10 and 14
days of age. A single intravenous or intramuscular iron dextran
injection allowed the homozygote embryos to survive (4) for sev-
eral weeks but not to reach maturity.
Here we demonstrate that administration of multiple iron injec-
tions enabled the wehTp85c–/– zebrafish to survive to adulthood,
facilitating evaluation of the effects of the fpn1 mutation on blood
development and iron homeostasis. We provide evidence that
wehTp85c–/– adults have impaired iron export, both from enterocytes
and macrophages, and that normal fpn1 function is not required
for iron-responsive regulation of hepcidin in the zebrafish.
Nonstandard abbreviations used: CH, cellular hemoglobin; DAB, diaminobenzi-
dine; DMT1, divalent metal transporter 1; FSC, forward scatter; fpn1, ferroportin1;
HFE, hemochromatosis gene product; IRE, iron regulatory element; MCV, mean
corpuscular volume; pcm, polycythemia; SSC, side scatter; wehTp85c–/–, zebrafish anemia
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J. Clin. Invest. 115:1532–1541 (2005).
The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
The effects of a missense mutation in fpn1 have not been evaluated
previously in an adult model organism. In order to study the effects
of fpn1 deficiency on erythroid development and iron homeostasis,
we rescued wehTp85c homozygotes to adulthood by administering
iron dextran injections early in development. Without iron injec-
tions, 0 of 529 homozygous mutants were alive at 3 weeks. Treat-
ment with a single iron injection at 72 hours rescued the anemia
phenotype and allowed 47 of 87 (54%) of the mutants (wehTp85c–/–)
to survive to 1 month, which was similar to the survival rate of iron-
injected (37 of 61; 61%) and uninjected (47 of 89; 53%) controls
(wehTp85c+/– and wehTp85c+/+). Without additional iron injections, the
mutant animals did not continue to mature. After a total of 4 injec-
tions of iron dextran were administered to the wehTp85c–/– zebrafish
(day 3, week 5, week 8, and week 16), the animals reached adult size.
This regimen was adopted for subsequent experiments.
As fpn1 has been shown to be expressed in the yolk syncytial layer
that separates the embryo from the iron-rich yolk, we postulated
that the weh mutation could have a maternal-dominant effect,
impairing export of iron to the embryo from maternally derived
stores. WehTp85c homozygous females crossed with wehTp85c+/+ males
produced only apparently normal, nonanemic embryos (data not
shown). This indicates that iron transfer into the zebrafish embryo
does not depend on maternal transcripts of fpn1. Either fpn1 tran-
scripts derived from the zygote are sufficient, or iron is transferred
into the embryo via an fpn1-independent mechanism.
The wehTp85c homozygotes and controls were monitored as they
aged. At 6 months of age, 2 months after the last iron injection,
wehTp85c–/– did not differ in appearance from injected or uninjected
controls. By 12 months of age, however, the homozygote animals
had developed pallor (Figure 1A), and the peripheral blood was
hypochromic on a blood smear (Figure 1B). To assess the severity of
anemia, we measured mean corpuscular volume (MCV) (Figure 1C)
and cellular hemoglobin (CH) (Figure 1D) from the peripheral
blood of individual zebrafish. MCV is a measure of erythrocyte size
while CH indicates the amount of hemoglobin per erythrocyte. At 6
months of age, there was no significant difference in MCV and CH
among the cohorts. The MCV was 81.8 ± 9.82 fl for wehTp85c–/– zebra-
fish versus 99.2 ± 6.65 for iron-injected WT zebrafish (P = 0.26). CH
was 26.6 ± 3.4 pg for wehTp85c–/– zebrafish versus 29.9 ± 1.74 pg for
iron-injected WT zebrafish (P = 0.50). When the zebrafish reached
12 months of age, however, we observed a reduction in MCV in iron-
injected wehTp85c–/– zebrafish compared with iron-injected WT zebra-
fish (69.6 ± 1.43 fl vs. 95.4 ± 1.36 fl, respectively; P < 0.0001) or unin-
jected WT zebrafish (89.9 ± 2.93 fl; P = 0.0008). Similarly, the amount
of hemoglobin per erythrocyte (CH) was reduced significantly from
32.2 ± 2.53 pg in iron-injected WT zebrafish to 23.4 ± 0.78 pg in iron-
injected wehTp85c–/– zebrafish (P < 0.0001). The decrease observed in
both erythrocyte size and hemoglobin content are consistent with
anemia due to impaired hemoglobin production.
To examine the effect of fpn1 deficiency on the adult site of
zebrafish hematopoiesis, kidneys from individual 12-month-old
wehTp85c–/– zebrafish and controls were dissected. Flow cytometry was
used to separate the kidney marrow cells into 4 populations based
on differences in forward scatter (FSC) and side scatter (SSC) in
controls (Figure 2A) and wehTp85c–/– zebrafish (Figure 2B). The mean
percentages of cells in these populations were compared (Figure 2C).
This revealed a 3-fold reduction in the mean percentage of mature
erythrocytes in wehTp85c–/– zebrafish (11.8 ± 1.10 in wehTp85c–/– vs.
36.8 ± 1.93 in WT plus iron; P < 0.0001). Also observed was a doubling
The phenotype of the weh mutation in adult zebrafish. (A) Photographs of an iron-injected WT zebrafish (left) compared with a pale, iron-injected
homozygote (wehTp85c–/–) (right) at 1 year of age. (B) Erythrocytes from peripheral blood of iron-injected wehTp85c–/– (Mut + Fe; right) at 1 year of
age exhibit pale cytoplasm and decondensed nuclei compared with WT zebrafish (WT no Fe; left) stained with Wright-Giemsa. Magnification
×100. Scale bars: 10 microns. (C and D) Erythroid indices were obtained with an ADVIA 120 automated analyzer at 6 months (gray bars, n = 3–5)
and 12 months (white bars, n = 4–6) for each cohort: uninjected wehTp85c+/+ zebrafish (WT no Fe), iron-injected wehTp85c+/+ zebrafish (WT + Fe),
uninjected wehTp85c+/– zebrafish (Het no Fe), iron-injected wehTp85c+/– zebrafish (Het + Fe), and iron-injected wehTp85c–/– zebrafish (Mut + Fe). MCV
(C) is a measure of erythrocyte size, while CH (D) quantitates the amount of hemoglobin per erythrocyte. Data shown are means ± SE. ND, not
done. *P < 0.0001 compared with 12-month-old iron-injected WT zebrafish.
1534 The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
in the mean percentage of the lymphoid/erythroblast population
that includes both lymphoid cells and small erythroblasts (32.1 ± 2.84
in wehTp85c–/– vs. 15.3 ± 1.96 in WT plus iron; P = 0.006). Enumera-
tion of manual differentials (Table 1) from cytospin preparations of
kidney marrow confirmed a significant increase in the percentage of
erythroblasts (Figure 2D) in wehTp85c–/– zebrafish compared with that
in iron-injected WT zebrafish (43.8 ± 6.34 vs. 9.25 ± 0.75; P = 0.02)
without a significant difference in the percentage of lymphocytes
(Table 1). These data indicate that wehTp85c–/– zebrafish undergo a
partial arrest in late erythropoiesis that results in a shift in the mar-
row population toward erythroblasts.
While fpn1 has been proposed as the principle enterocyte iron
exporter, it has not been demonstrated previously that a missense
mutation in fpn1 impairs intestinal iron
export. At 6 months, all the iron-rescued
wehTp85c–/– zebrafish evaluated (n = 4)
had strong staining of the intestinal villi
consistent with the hypothesized block
in iron export caused by fpn1 deficiency
(Figure 3A). This was not observed in any
of the other cohorts at 6 months (Table 2).
At 12 months, again, only wehTp85c–/–
homozygotes exhibited nonheme iron
staining in the intestinal villi. Although
the staining was less intense (Figure 3B),
it was apparent with diaminobenzidine
(DAB) (Figure 3C) in 3 of the 4 homo-
zygotes, consistent with impaired iron
export from the enterocytes.
In the vertebrate marrow, macrophages
store iron for release to developing erythroid
cells. While it is expected that iron dextran
injection would load the reticuloendothelial
system with iron, the delayed onset of ane-
mia in iron-injected wehTp85c–/– zebrafish rais-
es the possibility that kidney marrow iron
stores were exhausted by 12 months of age.
At 6 months, increased kidney iron stores
were evident upon Perls’ staining in the
macrophages of iron-injected fish, regardless
of genotype (Figure 4A). Increased kidney
macrophage iron stores were again demon-
strated in iron-injected wehTp85c–/– zebrafish
at 12 months (Figure 4, B and C), the same
wehTp85c–/– zebrafish shown to be anemic
by erythroid indices (Figure 1, C and D).
Due to the small size of the zebrafish kidney, it was not feasible to quan-
titate kidney nonheme iron levels; however, enumeration of iron-laden
macrophages per high-powered field approximates kidney iron stores.
Evaluation of uninjected heterozygotes at 12 months revealed kidney
macrophage iron stores similar to those of uninjected WT zebrafish
(0.75 ± 0.48 vs. 4.75 ± 4.75; P = 0.43). At 6 months and 12 months,
the mean numbers of iron-laden kidney macrophages per high-
powered field in wehTp85c–/– zebrafish were 38.0 ± 8.0 and 43.5 ± 7.12,
respectively (Figure 4D). These values were not significantly different
from those for iron-injected WT zebrafish at 6 months (48.3 ± 27.7;
P = 0.79) or 12 months (34.8 ± 11.9; P = 0.55). These data indicate that
the onset of anemia in wehTp85c–/– zebrafish was not associated with
depletion of kidney macrophage iron stores.
Marrow differentials obtained from the kidneys of adult zebrafish at 1 year of age. From each
animal, 1 × 105 kidney marrow cells were analyzed by FSC and SSC, according to a previously
defined method (48). Representative flow cytometry plots are shown from an iron-injected WT
(A) and a wehTp85c–/– (B) zebrafish. Four major populations were delineated: erythroid (red
ellipse), lymphoid/erythroblast (yellow ellipse), myeloid (lilac ellipse), and the most immature
precursor cells (blue ellipse). Shown next to each ellipse is the percentage of kidney marrow
cells in each gate. (C) Mean percentages of kidney marrow cells in each of the major cell popu-
lations; n = 3–5 per cohort. *P = 0.006; **P < 0.0001 compared with iron-injected WT zebrafish.
(D) Kidney marrow cytospins stained with Wright-Giemsa for iron-injected WT (left) compared
with iron-injected wehTp85c–/– zebrafish (right). Open arrows indicate erythrocytes, while black
arrow indicates 1 of 5 erythroblasts shown in the wehTp85c–/– zebrafish photomicrograph. Mag-
nification, ×100. Scale bar: 10 microns.
Manual differentials of kidney marrow
WT no Fe
WT + Fe
Het no Fe
He + Fe
Mut + Fe
% Myeloid and lymphoid blasts
8.75 ± 5.25
7.50 ± 2.00
3.00 ± 0.50
8.33 ± 0.60
5.67 ± 1.74
% Myelomonocytes % Lymphocytes
44.2 ± 6.75
44.0 ± 10.0
27.8 ± 3.25
38.3 ± 2.52
32.5 ± 4.36
% Thrombocytes % Erythroblasts
1.25 ± 0.75
1.00 ± 0.50
1.00 ± 1.00
2.83 ± 0.93
0.33 ± 0.33
% Mature erythrocytes
23.5 ± 4.50
30.8 ± 12.2
38.8 ± 3.25
26.3 ± 3.48
9.33 ± 0.44B
9.00 ± 4.00
7.25 ± 0.25
13.2 ± 5.25
8.00 ± 0.50
8.33 ± 0.60
13.8 ± 4.25
9.25 ± 0.75
15.2 ± 4.25
15.7 ± 5.50
43.8 ± 6.34A
Manual differentials obtained from cytospins of kidney marrow from individual 12-month-old zebrafish; n = 2–3 per cohort. Slides were stained with Wright-
Giemsa, and 200 cells were counted per zebrafish. Data shown are percentages ± SE. ANOVA reached statistical significance only for proerythrocyte and
mature erythrocyte categories. AP < 0.05 compared with each of the other cohorts. BP < 0.05 compared with uninjected WT, uninjected heterozygotes, and
The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
Liver iron stores were also increased among the iron-injected
fish. At 6 months (Figure 5A), nonheme iron deposition was evi-
dent on Perls’ staining both diffusely in hepatocytes and localized
to Kupffer cells in the iron-injected fish (Table 2). At 12 months,
histologic evidence of iron accumulation was still apparent in
some of the wehTp85c–/– zebrafish (Figure 5B), primarily in Kupffer
cells (Figure 5C), and diffuse staining of hepatocytes was absent.
In contrast, iron-injected WT zebrafish continued to exhibit dif-
fuse hepatocyte iron staining (Figure 5B) with less Kupffer cell iron
staining (Figure 5C). These findings suggest that fpn1 deficiency
impaired iron export from Kupffer cells to a greater extent than it
impaired iron export from hepatocytes.
Quantitation of nonheme iron stores using the bathophenanth-
roline sulfonate chromagenic assay confirmed that iron-injected
wehTp85c–/– zebrafish had increased hepatic iron stores. Because the
assay required pooling of large numbers of fish, it was initially per-
formed on 3 cohorts at 8 months of age (n = 1 pool per cohort), prior
to the onset of anemia. Hepatic iron levels at 8 months were elevated
in iron-injected WT (508 μg/g) and wehTp85c–/– zebrafish (250 μg/g)
compared with levels in uninjected WT zebrafish (40.0 μg/g). At 12
months of age, hepatic iron stores were assayed in all cohorts; n = 2–4
pools per cohort (Figure 5D). Hepatic iron levels in wehTp85c–/– zebra-
fish were significantly greater than those in uninjected WT zebrafish
(111 ± 29.8 μg/g vs. 27.4 ± 8.44 μg/g; P = 0.02) but not significantly dif-
ferent from iron levels in iron-injected WT zebrafish (226 ± 69.8 μg/g;
P = 0.20). These results indicate that the anemia in wehTp85c–/– zebra-
fish was not associated with exhaustion of liver iron stores.
While anemia would be expected to suppress expression of the
iron regulator hepcidin and increase fpn1 expression, accumulation
of iron in wehTp85c–/– zebrafish could lead to the opposite effect. To
quantitate changes in gene expression, RNA was extracted from the
livers and intestines of zebrafish from mutant and
control cohorts at 10–12 months of age and used
to generate cDNA templates for multiplex real-
time PCR amplifying β-actin and the gene of inter-
est. Transcript abundance, normalized to β-actin
expression, is expressed as a fold increase over a
calibrator sample. Although the mean hepatic
transcript levels for transferrin (Figure 6A) and fpn1
(Figure 6B) were not significantly different among
the mutant and control cohorts, wehTp85c–/– zebra-
fish displayed significantly increased expression
of intestinal fpn1 (Figure 6C). The mean intesti-
nal transcript level of fpn1 in weh was 8.34 ± 2.43,
compared with 1.36 ± 0.46 in iron-injected WT
zebrafish (P = 0.01). In situ hybridizations for
fpn1 performed on paraffin-embedded sections of
zebrafish intestine at 1 year (Figure 7) confirmed
that the majority of wehTp85c–/– zebrafish (3/4) had
increased intestinal expression of fpn1 that was not
apparent in iron-injected WT zebrafish.
Although upregulation of fpn1 in the wehTp85c–/–
zebrafish intestine may be related to the local
effect of iron accumulation in the enterocyte,
downregulation of hepcidin would be consistent with
a systemic response to anemia. In wehTp85c–/– zebra-
fish, the mean hepcidin transcript level (Figure 6D)
was lower than in every other cohort (0.039 ± 0.013
vs. 3.71 ± 1.17 in iron-injected WT; P < 0.0001).
Although hepcidin expression was highly variable
among both injected and uninjected controls, there was no sig-
nificant difference in mean hepcidin transcript levels among the
control cohorts, suggesting that the acute effects of iron injection
had resolved months after the last treatment.
Zebrafish embryos from 48–66 hours after fertilization exhibit low
levels of endogenous hepcidin transcripts (Figure 6E), facilitating the
evaluation of factors affecting hepcidin expression. To test the hypoth-
esis that hepcidin is an iron-responsive gene in the zebrafish, embryos
at 48 hours after fertilization were injected with iron dextran. The
multiplex PCR assay was used to quantitate transcript abundance at
specific time points for injected embryos and age-matched controls.
An initial small increase in hepcidin transcript levels was observed 2
hours after iron injection (1.37 ± 0.37 vs. 0.003 ± 0.001 in age-matched,
Histology of intestinal villi in uninjected WT (left), WT injected with iron (center), and
wehTp85c–/– zebrafish injected with iron (right). Perls’ stain performed at 6 months of age
(A) and 12 months of age (B) revealed nonheme iron accumulation in the intestinal villi
of wehTp85c–/– zebrafish but not in controls. Magnification, ×40. Scale bar: 100 microns.
(C) DAB-enhanced Perls’ staining for nonheme iron at 12 months for the same speci-
mens as shown in B.
Nonheme iron staining
WT no Fe
WT + Fe
Het no Fe
Het + Fe
Mut + Fe
6 mo 12 mo
Summary of intestine and liver nonheme iron–staining data at 6 months
and 12 months of age. Shown in the numerator is the number of individ-
uals in each cohort with increased nonheme iron staining in enterocytes
(left) or hepatocytes (center) or more than 10 iron-laden Kupffer cells
per high-powered field (right). The denominator is the total number of
1536 The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
uninjected; P = 0.01) followed by a subsequent larger increase 18
hours after iron injection (13.0 ± 5.44 vs. 0.08 ± 0.06 in age-matched,
uninjected; P = 0.03). To assess the effects of fpn1 function on hepci-
din expression, transcript levels were measured in weh homozygotes
18 hours after iron dextran injection. Hepcidin expression increased
significantly (Figure 6F) in the iron-treated mutants (14.4 ± 1.70
vs. 0.48 ± 0.17 in age-matched, uninjected mutants; P = 0.002)
and was not significantly different from the level observed in iron-
injected WT siblings (27.9 ± 7.76). This indicates that induction of
hepcidin expression does not require normal fpn1 function.
Kidney marrow iron stores. Iron-laden macrophages demonstrated by Perls’ staining in uninjected WT (left), WT treated with iron (center), and wehTp85c–/–
zebrafish treated with iron (right) that were sacrificed at ages 6 months (A) and 12 months (B). Magnification, ×40. Scale bar: 100 microns. Iron-laden
macrophages were located in the hematopoietic tissue that surrounds the pale staining renal tubules. (C) Higher magnification of area marked by
rectangle in B. Magnification, ×160. Scale bar: 10 microns. (D) Average number of iron-laden macrophages per high-powered field (hpf) in each of
the cohorts at 6 months (gray bars; n = 2–3) and 12 months (white bars; n = 4) of age. Zero indicates that no iron-laden macrophages were identified.
*P = 0.029 compared with uninjected heterozygotes at 12 months; **P = 0.013 compared with uninjected heterozygotes at 12 months; ***P = 0.001
compared with uninjected heterozygotes at 12 months and P = 0.004 compared with uninjected WT zebrafish at 12 months.
Liver iron stores. Perls’ staining demonstrating accumulation of nonheme iron in the liver of uninjected WT (left), iron-injected WT (center), iron-
injected wehTp85c–/– zebrafish (right) at 6 months of age (A) and 12 months of age (B). Magnification, ×40. Scale bar: 100 microns. (C) Higher mag-
nification of area marked by rectangle in B. Magnification, ×160. Scale bar: 10 microns. (D) Nonheme iron levels were quantified (μg iron/g liver)
using the bathophenanthroline sulfonate chromagenic assay. Due to the small size of the organs, livers were dissected from individual zebrafish
at 12 months of age and pooled according to cohort; n = 2–4 pooled samples per cohort. *P = 0.02 compared with uninjected WT zebrafish.
The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
The zebrafish mutant weh is the first known animal model of fpn1
deficiency. Previous work demonstrated embryonic anemia in weh
zebrafish attributed to a defect in iron transfer across the yolk syn-
cytial layer (4). We have now demonstrated nonheme iron accumu-
lation in the intestinal epithelium of wehTp85c–/– adults, consistent
with a block in intestinal iron export. This resembles the pattern
of iron accumulation in the sex-linked anemia (sla) mouse (27, 28),
which has a defect in intestinal iron export due to a mutation in the
basolateral ferroxidase, hephaestin (29).
We propose that injecting wehTp85c–/– homozygotes with iron
dextran rescues the animals by bypassing the block in intestinal
iron absorption. Although most iron is tightly bound to dextran
and would be phagocytosed by macrophages, approximately 2% is
labile and can immediately bind transferrin (30) without further
processing by the reticuloendothelial system. The increase in hepci-
din expression observed in both WT and mutant embryos 18 hours
after iron dextran injection supports the hypothesis that serum
iron levels rise following the injection. Transferrin-bound iron
would be readily delivered to the erythroid compartment, driving
Quantification of transcript levels by multiplex real-time PCR. RNA from liver (A, B, and D) or intestine (C) of 10- to 12-month-old adult zebrafish
or pools of embryos (E and F) was used to generate cDNA templates for multiplex reactions amplifying the gene of interest and β-actin. Transcript
abundance, normalized to β-actin expression, is expressed as a fold increase over a calibrator sample. (A–D) Expression of transferrin (A), fpn1
(B and C), or hepcidin (D) in adult WT, heterozygote, or mutant zebrafish treated with iron 6–8 months previously (+Fe) or never treated (no Fe).
The calibrator sample was an uninjected WT adult; n = 6–9 individuals per cohort. *P = 0.01 compared with iron-injected WT zebrafish. **P < 0.0001
compared with iron-injected WT zebrafish. (E and F) At 48 hours after fertilization, embryos were anesthetized and injected with iron dextran (+)
or anesthetized without iron injection (–). (E) Eighteen-hour time course of hepcidin transcript abundance following iron dextran injection. The
calibrator sample consisted of a pool of embryos 2 hours after iron injection; n = 2 pools of embryos per cohort. ‡P < 0.05 compared with uninjected
age-matched embryos. (F) Induction of hepcidin expression 18 hours after iron dextran injection of wehTp85c–/– or WT sibling embryos. The calibrator
sample was an uninjected pool of embryos; n = 3 pools of embryos per cohort. ‡‡P < 0.01 compared with uninjected age-matched embryos.
Radioactive in situ hybridization for fpn1. (A) Iron-injected wehTp85c–/–
zebrafish with increased fpn1 expression compared with (B) iron-inject-
ed WT. Shown are intestines from zebrafish sacrificed at 1 year of age,
fixed in 4% paraformaldehyde, embedded in paraffin, and hybridized
with 35S-labeled antisense (left) or sense (right) probe for fpn1. Magni-
fication, ×4. Scale bar: 1 mm.
1538 The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
hemoglobin synthesis and erythroid maturation. Over time, aging
erythrocytes are phagocytosed by macrophages. The defect in fpn1
function impairs iron export, resulting in retention of iron in the
reticuloendothelial system and the onset of anemia. The anemia
is associated with decreased transcript levels of the iron regulator
hepcidin in wehTp85c–/– zebrafish, while transferrin and hepatic fpn1
expression are not significantly different from that in controls.
Our analysis of fpn1-deficiency in wehTp85c–/– zebrafish supports the
critical role of fpn1 in releasing iron to developing erythroid cells and
is consistent with findings in the polycythemia (pcm) mouse (31), which
develops pcm due to a hypermorphic allele of fpn1 that lacks iron/
iron regulatory element (IRE)–mediated translational regulation due
to a deletion in the promoter. At 7 weeks of age, pcm heterozygous
mice with low hepcidin expression exhibit increased hepatic and duo-
denal fpn1 expression, expanded erythroid differentiation (31), and
resulting polycythemia. On the other hand, our data illustrate that
impaired fpn1 function in wehTp85c–/– zebrafish is associated with
hypochromic anemia and impaired erythroid maturation.
The decreased hepcidin expression demonstrated in anemic wehTp85c–/–
adults and the increased levels in WT zebrafish and mutants fol-
lowing acute iron dextran injection provide the first evidence for a
conserved role for hepcidin in the zebrafish as a regulator of iron
homeostasis. Others have identified increases in hepcidin expression
in response to bacterial infection in zebrafish (11), suggesting that
hepcidin in zebrafish, as in other organisms, is part of the inflamma-
tory response. The low hepcidin expression in anemic wehTp85c–/– adults
is consistent with the response to anemia observed in the hypotrans-
ferrinemic mouse, which also exhibits anemia despite adequate iron
stores (32). Similarly, the Belgrade b/b rat, which has hypochromic
anemia due to a mutation in DMT1, exhibits undetectable levels of
hepcidin mRNA accompanied by an increase in intestinal but not
hepatic fpn1 expression (33). Investigators have noted that adminis-
tration of recombinant hepcidin peptide to rats results in a decline
in duodenal but not hepatic fpn1 expression (33). We observed differ-
ences in hepatic and intestinal fpn1 expression that may reflect the
effects of decreased hepcidin: mean intestinal fpn1 transcript levels
were increased in anemic wehTp85c–/– adults, but mean hepatic fpn1
mRNA levels were not significantly different from those of controls.
In addition to having effects on transcriptional regulation, hep-
cidin appears to be a critical homeostatic regulator affecting both
iron uptake and iron export. Exposing cultured human intestinal
epithelial cells to hepcidin (34) resulted in decreased apical iron
uptake and diminished mRNA and protein expression of DMT1
(3′-IRE-containing isoform). In addition, treatment of tissue cul-
ture cells expressing fpn1-GFP with hepcidin resulted in internal-
ization and degradation of fpn1 as well as decreased iron export
(35). Further supporting a homeostatic loop involving fpn1 and
hepcidin, pcm heterozygote mice with increased fpn1 protein
expression at 7 weeks developed increased hepcidin expression and
a decline in fpn1 protein to normal levels by 12 weeks (31), result-
ing in resolution of the polycythemia phenotype. The profoundly
low hepcidin expression in wehTp85c–/– zebrafish can also be seen as a
homeostatic response to anemia. Unlike the pcm mouse, however,
wehTp85c–/– zebrafish possess a missense mutation in fpn1, which
results in a failure to correct the phenotype despite the compensa-
tory change in hepcidin expression.
Given the homeostatic relationship of hepcidin and fpn1, it is
conceivable that the wehTp85c mutation might restrict hepcidin
expression. Mice deficient in the hereditary hemochromatosis
gene product (HFE) exhibit inappropriately low levels of hepcidin
expression when fed a high-iron diet (36) or injected with iron dex-
tran (37). These findings imply that HFE function is required for
proper regulation of hepcidin expression in response to iron loading.
In contrast, wehTp85c–/– embryos injected with iron dextran manifest
an appropriate increase in hepcidin expression, indicating that the
L167F mutation in fpn1 does not impair iron-induced regulation
of hepcidin expression.
While the majority of hereditary hemochromatosis cases are
attributed to mutations in the atypical MHC class I–protein HFE
(38), it is now recognized that hereditary hemochromatosis can be
associated with defects in hepcidin or hemojuvelin (type 2), transfer-
rin receptor 2 (type 3), or fpn1 (type 4) (39). The Tp85c allele of weh
has a mutation (L167F) (4) at a conserved leucine in a loop between
the third and fourth predicted transmembrane helices of fpn1 (22).
A cluster of human mutations (Supplemental Table 1; available
online with this article; doi:10.1172/JCI23780DS1) in patients with
type 4 hemochromatosis or elevated serum ferritin has now been
identified at conserved residues in this loop (20–26), suggesting
that this may be a functional domain, and at other locations (24,
40–42) of fpn1. Although hepcidin expression has been found to
be low in patients with hemochromatosis attributed to mutations
in HFE (43), transferrin receptor 2 (44), hepcidin (45), or hemojuvelin
(HFE2) (46), urinary protein levels have been found to be increased
(47) in 2 hemochromatosis patients heterozygous for the Val162del
mutation, localized between transmembrane helices 3 and 4 of fpn1.
This is consistent with our observation that iron-responsive hepcidin
expression is not ablated by the zebrafish L167F mutation. Taken
together, these findings suggest that regulation of hepcidin expres-
sion is independent of fpn1 function as an iron transporter.
While a gain of function mutation has not been ruled out in
patients with type 4 hemochromatosis, the disease manifests with
distinctive clinical features that are consistent with impaired iron
export from the reticuloendothelial system. These features include
accumulation of iron specifically in macrophages and Kupffer
cells, absence of hepatic fibrosis, and a tendency among those who
have the disease to develop anemia (39, 40) following venesection.
Patients with type 4 hemochromatosis, unlike those with types 1,
2, or 3, usually have high serum ferritin levels that appear to be
disproportionate to their serum iron levels and transferrin satura-
tions (22, 24). Although patients with type 4 hemochromatosis
are heterozygous for fpn1 missense mutations, we have not identi-
fied anemia or iron accumulation in wehTp85c heterozygotes. This
raises the possibility that the effect of the zebrafish mutation is
less severe than that of the human mutation or that there may be
secondary mutations in fpn1 or other modifying genes that con-
tribute to iron overload. The presence or absence of a haploinsuf-
ficiency phenotype in zebrafish varies depending on the nature
of the mutation: a mild anemia phenotype has been detected in
zebrafish that are heterozygous for defects in solute carrier family
4 but not in the interacting erythrocyte membrane protein band
4.1 (48). The absence of a haploinsufficiency phenotype in wehTp85c
zebrafish may be due to partial function of the mutated fpn1 or to
compensation by an alternative exporter.
In summary, wehTp85c–/– zebrafish have impaired function of
fpn1, which results in defective intestinal and macrophage iron
export. We have identified hepcidin as an iron-responsive gene in
the zebrafish that does not require normal fpn1 export function
for its expression. Our studies demonstrate the use of zebrafish to
study the interaction of genes involved in hemochromatosis and
establish a foundation for further genetic analysis.
The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
Zebrafish strains, maintenance, and determination of genotype. Zebrafish were
maintained as described (49). Ethical approval for these experiments was
obtained from the Institutional Animal Care and Use Committees of
Children’s Hospital and Beth Israel Deaconess Medical Center. WehTp85c
were maintained on the AB background. Homozygote and heterozygote
progeny were generated by crosses of heterozygotes with heterozygotes or
iron-rescued homozygotes. WehTp85c+/+ (WT) controls were siblings of the
homozygotes generated by heterozygote incrosses. Additional WT zebrafish
were generated in AB WT incrosses maintained in parallel. The embryos were
assessed for anemia at 48 hours after fertilization. Embryos in the injected
cohorts were first microinjected with 100 mg/ml iron dextran (Sigma-
Aldrich) at 48–72 hours after fertilization, as described (4). Additional injec-
tions of iron dextran were performed i.m. at 3 weeks (4 nl), 8 weeks (200 μl),
and 16 weeks (200 μl). Fish were of similar sizes across the cohorts, and dose
was not adjusted for weight. Both injected and uninjected cohorts were fed
an iron-rich diet of brine shrimp. Genotypes were confirmed by amplifying
with the following primers: 5′-AATGGTCATCTCCATTGCTAATATCGC-3′
and 5′-AAATGTACAATGGCCAAAAAACGAATT -3′. The PCR product
was purified using the QIAquick spin column (QIAGEN). The presence or
absence of the mutation was determined by direct automated sequencing
with the primer 5′-GCAATGAACTATAAATGG-3′.
Analysis of peripheral blood and kidney marrow. The peripheral blood was col-
lected from individual adult zebrafish as previously described (12, 13) and
washed in 40 ml 0.9 × PBS-1% BSA. After centrifugation, the blood was resus-
pended in 200 μl of 0.9 × PBS-1% BSA and analyzed on an automated cell
counter (ADVIA 120; Bayer Diagnostics) for determination of MCV and CH.
MCV readings obtained from the ADVIA were validated by determining
packed erythrocyte cell volume for WT zebrafish using classical techniques.
Blood was pooled from cohorts of WT adult zebrafish (n = 3 cohorts, 9 fish
per cohort). Pooled blood was washed, as described above, and resuspend-
ed in 60 μl of 0.9 × PBS-1% BSA. The blood was then centrifuged in 9-μl
microcapillary tubes for 2 minutes in a StatSpin multipurpose centrifuge to
determine the spun hematocrit. From the same samples, the cell density was
determined by enumeration with a hemocytometer. The packed erythrocyte
cell volume was calculated as follows: (spun hematocrit × 0.01 × 109 fl/μl)/
number of cells per μl. Aliquots of the pooled blood samples were also dilut-
ed 10-fold in 0.9 × PBS-1% BSA for evaluation of MCV on the ADVIA. We
determined that the 2 techniques did not result in significantly different
MCVs for WT zebrafish: 97.0 ± 3.69 fl (classical method) compared with
92.9 ± 2.99 fl obtained with the ADVIA 120 (P = 0.43). Data reported in our
analyses were obtained on the ADVIA 120.
Kidney marrow was dissected and gently suspended in 0.5 ml ice-cold
PBS-5% fetal calf serum containing heparin (5 U/ml). The samples were
filtered through a 40 micron filter. Propidium iodide was added to a con-
centration of 1 μg/ml to exclude dead cells. As previously described (48),
1 × 105 cells were analyzed for each fish by flow cytometry, using differ-
ences in FSC and SSC to obtain automated differentials. Cell populations
were analyzed with the program FlowJo 4.6 (Treestar Inc.). Cytospins of
peripheral blood and kidney marrow were also obtained from 1 × 105 kid-
ney cells using a Shandon Cytospin 3 centrifuge (Thermo Electron Corp.).
Cytospin slides were processed through Wright and Giemsa stains (Har-
leco) for morphological analysis and cell counts.
In situ hybridization and histological analysis. Zebrafish tissues were fixed in
4% paraformaldehyde-1× PBS (pH 7.4) at 4°C for 24 hours, embedded in
paraffin, and sectioned at a thickness of 4 microns. As riboprobes 0.5–1 kb
in size are optimal for in situ hybridization on fixed sections, cDNA tem-
plates for the riboprobes were generated by amplifying smaller pieces
within the open reading frame of fpn1. Sp6 and T7 were incorporated into
the forward and reverse primers, respectively, for subsequent run-off. The
PCR reaction conditions were as follows: 25 ng template (cloned full-length
cDNA provided by A. Donovan), 0.2 mM dNTP, 0.2 μM forward primer,
0.2 μM reverse primer, 1 × Advantage 2 Polymerase Mix (Invitrogen Corp.).
Cycling was as follows: 94°C for 2 minutes, followed by 35 cycles at 94°C for
20 seconds, 55°C for 30 seconds, and 68°C for 3 minutes, followed by 68°C
for 10 minutes. Primers were as follows: Fpn1F3Sp6 (5′-CAAGCTTGATT-
CAATTTGCTCCGATCAT-3′) to generate a 500 bp fragment (nucleotides
763–1193). The size of the PCR products was verified by electrophoresis.
The PCR reactions were purified with a QIAquick spin column and brought
to a final concentration of 1 μg/μl. S35-labeled riboprobes were prepared
with an in vitro transcription kit (Roche Diagnostics) using T7 or Sp6 to
run off the antisense or sense probe, respectively. The efficiency of transcrip-
tion and incorporation of S35-UTP into the riboprobes was evaluated by 5%
acrylamide–urea gel electrophoresis and scintillation counting.
Radioactive in situ hybridization was performed on paraffin-embedded
sections mounted on slides. After deparaffinization, slides were fixed in 4%
paraformaldehyde for 10 minutes and digested with proteinase K (10 μg/ml)
for 10 minutes at 37°C. Sense and antisense riboprobes were diluted in
hybridization buffer to 2 × 106 counts per minute per slide. Hybridization
was performed overnight at 60°C. After hybridization, slides were washed
at 65°C for 2 hours in 0.1 × SSC. Slides were dipped in Kodak NTB2 emul-
sion, exposed at 4°C for 4 weeks, developed, and counterstained with H&E.
Corresponding sections from the same animals were stained either with
H&E or Perls’ stain. For Perls’ stain (50, 51), sections were deparaffinized,
incubated in 5% potassium ferrocyanide/5% HCl for 20 minutes, washed in
distilled water, then counterstained with nuclear fast red. DAB-enhanced
iron staining was performed as follows: following deparaffinization, the
slides were incubated with 1% potassium ferrocyanide/0.12 N HCl, washed
with PBS, quenched in 0.3% H2O2 in methanol for 20 minutes, washed in
PBS, and incubated in DAB/H2O2 (DAB kit no. 2020; Zymed Laboratories)
for 7 minutes. No counterstain was performed on DAB-stained sections.
Nonheme iron quantification. Male zebrafish were sacrificed to avoid poten-
tial contamination of samples from iron-rich oocytes. Livers were dissect-
ed, pooled, and weighed in groups of 6 (at 8 months) or 3 (at 12 months)
and lysed for 48 hours at 60°C in 250 μl of 3 M hydrochloric acid/0.61 M
trichloroacetic acid in iron-free water. One ml of chromagen solution
(0.01% bathophenanthroline sulfonate, 0.1% thioglycolic acid, saturated
sodium acetate) was added to 50 μl of lysate. OD was measured at 535 nm
according to the method previously described (52). Iron was quantified in
comparison to a standard curve generated from iron standards (no. 565-11;
Sigma-Aldrich). The values were expressed as μg iron/g of wet tissue.
Quantitative analysis of gene expression. Individual wehTp85c mutant and control
zebrafish were dissected to obtain liver and intestine. Dissected organs were
immediately placed in RNAlater (Ambion) to stabilize transcripts. To ana-
lyze the effects of iron injection on embryonic gene expression, 48-hour-old
embryos were injected with iron dextran or anesthetized. At specified time
points after treatment, embryos were pooled in groups of 18–20, euthanized
with tricaine, and placed in RNAlater. Each organ or embryo pool was frozen
with liquid nitrogen and ground with a mortar and pestle. RNA was then
extracted from the ground tissue using the RNeasy lysis and purification
procedure (QIAGEN). RNA yield was quantified by absorption at 260 nm.
To obtain cDNA, a reverse transcription was performed on 100 ng of RNA
(for organs) or 1 pool (for embryos) using Superscript II reverse transcriptase
(Invitrogen Corp.) in a 20-μl reaction. The reaction was then diluted to 100 μl.
For each real-time PCR reaction, 10 μl were used as template. Reaction cock-
tail total volume was 50 μl and included 0.4-μM forward primer, 0.4-μM
reverse primer, 0.2-μM probe, and 1 × QuantiTect Probe PCR Master Mix
(QIAGEN). Reactions were performed in duplicate. Cycling parameters were
1540 The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
as follows: 50°C for 2 minutes and 95°C for 15 minutes, followed by 40
cycles at 94°C for 15 seconds and 60°C for 1 minute. Detection and analysis
were performed on an ABI Prism 7700 (Applied Biosystems).
Each reaction was designed to quantitate expression of fpn1, transferrin, or
hepcidin, normalized to zebrafish β-actin, according to a previously described
method (53). Transcript abundance was expressed as a fold increase over
calibrator sample. To design the primers used in the real-time PCR assays,
sequences for the zebrafish genes of interest (fpn1, transferrin, and hepci-
din) as compiled in Genbank were blasted against the zebrafish genome
assembly to identify predicted intron-exon boundaries. The β-actin primer
probe set was as previously described (53). The genes of interest were ampli-
fied with the following primers, each spanning a predicted intron-exon
boundary: fpn1, 5′- GGCCAGCACAGCTATGTC-3′, 5′-GCCAGAATGTT-
GGTCAACTG-3′; transferrin, 5′- TTACATGGGAGGGTCCTAATGAG-3′,
5′- GGACACAACTGCTCGAGAAGAA-3′; hepcidin, 5′- CCTGGCTGCT-
GTCGTCAT-3′, 5′- TGGTTCTCCTGCAGTTCTTCAC-3′; and β-actin,
5′- AGGTCATCACCATCGGCAAT-3′, 5′- GATGTCCACGTCGC-
ACTTCAT-3′. Probes were as follows: fpn1, 5′-FAM-TGCATTCATA-
TCTGCCAATTTGCTCCGA-TAMRA-3′; transferrin, 5′-FAM-
CTGACACAGCCCTCTCGACAGG-TAMRA-3′; hepcidin, 5′-FAM-
CCGTTCCCTTCATACAGCAGGTACAGG-TAMRA-3′; and β-actin, 5′-VIC-
The 100 bp amplicons for each gene were cloned and sequenced to con-
firm amplification of the appropriate genes. Serial dilutions of the cloned
amplicons were tested in the assay. Cycle thresholds were plotted versus
natural log of template concentration, yielding linear results over a range
of template concentrations from 1.4 × 10–17 to 1.4 × 10–12 M. The coeffi-
cients of variation determined by linear regression were –0.9837 (β-actin),
–0.9887 (fpn1), –0.9801 (hepcidin), and –0.9974 (transferrin). In each assay of
adult organs, cDNA from the same uninjected WT zebrafish was used as
a calibrator template. For pooled embryos, the calibrator template was 1
of the age-matched embryo control pools. The reactions failed to amplify
nontemplate controls, genomic DNA, or RNA preparations that were not
treated with reverse transcriptase.
Biostatistical analysis. Heterogeneity among age-matched cohorts was ana-
lyzed by ANOVA. Results reported from the ANOVA analysis are means ± SE.
Tests for heterogeneity among age-matched cohorts used the natural
logarithm for the assessment of nonheme iron levels and fpn1, transferrin,
and hepcidin transcript levels. All estimates and standard errors presented
have been transformed back to the original units. When the global P value
obtained from the ANOVA analysis was statistically significant, pairwise
comparisons between the age-matched cohorts were performed. All P val-
ues were obtained from 2-sided Student’s t tests. We did not adjust for
multiple comparisons. P values less than or equal to 0.05 were deemed sta-
tistically significant. ANOVA and Student’s t tests were performed using
InStat 3.0 (GraphPad Software).
We thank Victoria Petkova of the Beth Israel Deaconess Medi-
cal Center Real-Time PCR Core for assistance with the real-time
PCR assay and Yu Yang of the Harvard Dana-Farber Cancer
Institute In Situ Hybridization Core Laboratory for assistance
with in situ hybridizations. We are grateful to Tenora Archibald,
Children’s Hospital, for assistance with histopathology. We
also thank Kim Dooley and Rebecca Wingert for critical review
of the manuscript and Mark Fleming for fruitful discussions.
This work was supported by NIH grants DK-61685-03 (to P.G.
Fraenkel) and DK-053298-07 (to L.I. Zon), the Howard Hughes
Medical Institute grant (to L.I. Zon), and the American Society
of Hematology Scholar Award (P.G. Fraenkel).
Received for publication October 29, 2004, and accepted in revised
form March 29, 2005.
Address correspondence to: L.I. Zon, Division of Hematology/
Oncology, Children’s Hospital, Karp Research Laboratories, 1 Black-
fan Circle, Boston, Massachusetts 02115, USA. Phone: (617) 919-
2069; Fax: (617) 730-0222; E-mail: email@example.com.
1. Hentze, M.W., Muckenthaler, M.U., and Andrews,
N.C. 2004. Balancing acts: molecular control of
mammalian iron metabolism. Cell. 117:285–297.
2. Andrews, N.C. 1999. The iron transporter DMT1.
Int. J. Biochem. Cell Biol. 31:991–994.
3. Andrews, N.C., Fleming, M.D., and Levy, J.E. 1999.
Molecular insights into mechanisms of iron trans-
port. Curr. Opin. Hematol. 6:61–64.
4. Donovan, A., et al. 2000. Positional cloning of
zebrafish ferroportin1 identifies a conserved ver-
tebrate iron exporter. Nature. 403:776–781.
5. Park, C.H., Valore, E.V., Waring, A.J., and Ganz, T.
2001. Hepcidin, a urinary antimicrobial peptide syn-
thesized in the liver. J. Biol. Chem. 276:7806–7810.
6. Krause, A., et al. 2000. LEAP-1, a novel highly disul-
fide-bonded human peptide, exhibits antimicrobial
activity. FEBS Lett. 480:147–150.
7. Nicolas, G., et al. 2001. Lack of hepcidin gene
expression and severe tissue iron overload in
upstream stimulatory factor 2 (USF2) knockout
mice. Proc. Natl. Acad. Sci. U. S. A. 98:8780–8785.
8. Nicolas, G., et al. 2002. Severe iron deficiency ane-
mia in transgenic mice expressing liver hepcidin.
Proc. Natl. Acad. Sci. U. S. A. 99:4596–4601.
9. Nicolas, G., et al. 2002. The gene encoding the
iron regulatory peptide hepcidin is regulated by
anemia, hypoxia, and inflammation. J. Clin. Invest.
10. Pigeon, C., et al. 2001. A new mouse liver-specific
gene, encoding a protein homologous to human
antimicrobial peptide hepcidin, is overexpressed
during iron overload. J. Biol. Chem. 276:7811–7819.
11. Shike, H., Shimizu, C., Lauth, X., and Burns, J.C.
2004. Organization and expression analysis of the
zebrafish hepcidin gene, an antimicrobial peptide
gene conserved among vertebrates. Dev. Comp.
12. Brownlie, A., et al. 2003. Characterization of
embryonic globin genes of the zebrafish. Dev. Biol.
13. Brownlie, A., et al. 1998. Positional cloning of the
zebrafish sauternes gene: a model for congenital
sideroblastic anaemia. Nat. Genet. 20:244–250.
14. Donovan, A., et al. 2002. The zebrafish mutant
gene chardonnay (cdy) encodes divalent metal
transporter 1 (DMT1). Blood. 100:4655–4659.
15. Wingert, R.A., et al. 2004. The chianti zebrafish
mutant provides a model for erythroid-specific
disruption of transferrin receptor 1. Development.
16. McKie, A.T., et al. 2000. A novel duodenal iron-
regulated transporter, IREG1, implicated in the
basolateral transfer of iron to the circulation. Mol.
17. Abboud, S., and Haile, D.J. 2000. A novel mamma-
lian iron-regulated protein involved in intracellular
iron metabolism. J. Biol. Chem. 275:19906–19912.
18. Lymboussaki, A., et al. 2003. The role of the iron
responsive element in the control of ferropor-
tin1/IREG1/MTP1 gene expression. J. Hepatol.
19. Ransom, D.G., et al. 1996. Characterization of
zebrafish mutants with defects in embryonic
hematopoiesis. Development. 123:311–319.
20. Wallace, D.F., et al. 2002. Novel mutation in fer-
roportin1 is associated with autosomal dominant
hemochromatosis. Blood. 100:692–694.
21. Roetto, A., et al. 2002. A valine deletion of ferropor-
tin 1: a common mutation in hemochromastosis
type 4. Blood. 100:733–734.
22. Devalia, V., et al. 2002. Autosomal dominant
reticuloendothelial iron overload associated with
a 3-base pair deletion in the ferroportin 1 gene
(SLC11A3). Blood. 100:695–697.
23. Njajou, O.T., et al. 2001. A mutation in SLC11A3
is associated with autosomal dominant hemochro-
matosis. Nat. Genet. 28:213–214.
24. Hetet, G., Devaux, I., Soufir, N., Grandchamp, B., and
Beaumont, C. 2003. Molecular analyses of patients
with hyperferritinemia and normal serum iron val-
ues reveal both L ferritin IRE and 3 new ferroportin
(slc11A3) mutations. Blood. 102:1904–1910.
25. Arden, K.E., et al. 2003. A novel mutation in fer-
roportin1 is associated with haemochromatosis in
a Solomon Islands patient. Gut. 52:1215–1217.
26. Wallace, D.F., Clark, R.M., Harley, H.A., and Sub-
ramaniam, V.N. 2004. Autosomal dominant iron
overload due to a novel mutation of ferroportin1
associated with parenchymal iron loading and cir-
rhosis. J. Hepatol. 40:710–713.
27. Edwards, J.A., and Bannerman, R.M. 1970. Heredi-
tary defect of intestinal iron transport in mice with
sex-linked anemia. J. Clin. Invest. 49:1869–1871.
28. Manis, J. 1971. Intestinal iron-transport defect in
the mouse with sex-linked anemia. Am J. Physiol.
The Journal of Clinical Investigation http://www.jci.org Volume 115 Number 6 June 2005
29. Vulpe, C.D., et al. 1999. Hephaestin, a cerulo-
plasmin homologue implicated in intestinal iron
transport, is defective in the sla mouse. Nat. Genet.
30. Van Wyck, D., Anderson, J., and Johnson, K. 2004.
Labile iron in parenteral iron formulations: a
quantitative and comparative study. Nephrol. Dial.
31. Mok, H., et al. 2004. Disruption of ferroportin
1 regulation causes dynamic alterations in iron
homeostasis and erythropoiesis in polycythaemia
mice. Development. 131:1859–1868.
32. Weinstein, D.A., et al. 2002. Inappropriate expres-
sion of hepcidin is associated with iron refractory
anemia: implications for the anemia of chronic
disease. Blood. 100:3776–3781.
33. Yeh, K.Y., Yeh, M., and Glass, J. 2004. Hepcidin
regulation of ferroportin 1 expression in the liver
and intestine of the rat. Am. J. Physiol. Gastrointest.
Liver Physiol. 286:G385–G394.
34. Yamaji, S., Sharp, P., Ramesh, B., and Srai, S.K.
2004. Inhibition of iron transport across human
intestinal epithelial cells by hepcidin. Blood.
35. Nemeth, E., et al. 2004. Hepcidin regulates cellular
iron efflux by binding to ferroportin and inducing
its internalization. Science. 306:2090–2093.
36. Ahmad, K.A., et al. 2002. Decreased liver hepcidin
expression in the Hfe knockout mouse. Blood Cells
Mol. Dis. 29:361–366.
37. Muckenthaler, M., et al. 2003. Regulatory defects
in liver and intestine implicate abnormal hepcidin
and Cybrd1 expression in mouse hemochromato-
sis. Nat. Genet. 34:102–107.
38. Merryweather-Clarke, A.T., Pointon, J.J., Jouanolle,
A.M., Rochette, J., and Robson, K.J. 2000. Geogra-
phy of HFE C282Y and H63D mutations. Genet.
39. Pietrangelo, A. 2004. Hereditary hemochroma-
tosis–a new look at an old disease. N. Engl. J. Med.
40. Montosi, G., et al. 2001. Autosomal-dominant
hemochromatosis is associated with a mutation
in the ferroportin (SLC11A3) gene. J. Clin. Invest.
41. Gordeuk, V.R., et al. 2003. Iron overload in Africans
and African-Americans and a common mutation
in the SCL40A1 (ferroportin 1) gene. Blood Cells
Mol. Dis. 31:299–304.
42. Jouanolle, A.M., et al. 2003. Novel mutation in fer-
roportin 1 gene is associated with autosomal domi-
nant iron overload. J. Hepatol. 39:286–289.
43. Bridle, K.R., et al. 2003. Disrupted hepcidin regu-
lation in HFE-associated haemochromatosis and
the liver as a regulator of body iron homoeostasis.
44. Nemeth, E., Roetto, A., Garozzo, G., Ganz, T., and
Camaschella, C. 2005. Hepcidin is decreased in
TFR2 hemochromatosis. Blood. 105:1803–1806.
45. Roetto, A., et al. 2003. Mutant antimicrobial
peptide hepcidin is associated with severe juvenile
hemochromatosis. Nat. Genet. 33:21–22.
46. Papanikolaou, G., et al. 2004. Mutations in HFE2
cause iron overload in chromosome 1q-linked juve-
nile hemochromatosis. Nat. Genet. 36:77–82.
47. Papanikolaou, G., et al. 2005. Hepcidin in iron
overload disorders. Blood. doi:10.1182/blood-2004-
48. Traver, D., et al. 2003. Transplantation and in vivo
imaging of multilineage engraftment in zebrafish
bloodless mutants. Nat. Immunol. 4:1238–1246.
49. Westerfield, M. 1994. The zebrafish book: a guide for
the laboratory use of zebrafish (Brachydanio rerio). Uni-
versity of Oregon Press. Eugene, Oregon, USA.
50. Sheehan, D.C., and Hrapchak, B.B. 1980. Theory
and practice of histotechnology. Battelle Press. Colum-
bus, Ohio, USA. 217–218.
51. Luna, L. 1968. Manual of histologic staining methods of
the AFIP. McGraw-Hill. New York, New York, USA.
52. Cook, J.D. 1980. Methods in hematology. Churchill
Livingstone. New York, New York, USA. 105–108.
53. Trant, J.M., Gavasso, S., Ackers, J., Chung, B.C.,
and Place, A.R. 2001. Developmental expression
of cytochrome P450 aromatase genes (CYP19a and
CYP19b) in zebrafish fry (Danio rerio). J. Exp. Zool.