Ferroportin1 in hepatocytes and macrophages is required for the efficient mobilization of body iron stores in mice

Authors

  • Zhuzhen Zhang,

    1. Group of Bio-Metal Metabolism, Key Laboratory of Nutrition and Metabolism, Shanghai Key Laboratory of Pediatric Gastroenterology and Nutrition, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School of the Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Fan Zhang,

    1. Group of Bio-Metal Metabolism, Key Laboratory of Nutrition and Metabolism, Shanghai Key Laboratory of Pediatric Gastroenterology and Nutrition, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School of the Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Xin Guo,

    1. Group of Bio-Metal Metabolism, Key Laboratory of Nutrition and Metabolism, Shanghai Key Laboratory of Pediatric Gastroenterology and Nutrition, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School of the Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Peng An,

    1. Group of Bio-Metal Metabolism, Key Laboratory of Nutrition and Metabolism, Shanghai Key Laboratory of Pediatric Gastroenterology and Nutrition, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School of the Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Yunlong Tao,

    1. Group of Bio-Metal Metabolism, Key Laboratory of Nutrition and Metabolism, Shanghai Key Laboratory of Pediatric Gastroenterology and Nutrition, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School of the Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Fudi Wang

    Corresponding author
    1. Group of Bio-Metal Metabolism, Key Laboratory of Nutrition and Metabolism, Shanghai Key Laboratory of Pediatric Gastroenterology and Nutrition, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Graduate School of the Chinese Academy of Sciences, Shanghai, China
    • Mineral Molecular Nutrition Laboratory, Institute for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 294 Taiyuan Road, Shanghai 200031, China===

    Search for more papers by this author
    • fax: +86-21-54920291


  • Potential conflict of interest: Nothing to report.

Abstract

The liver is a major site of iron storage where sequestered iron can be actively mobilized for utilization when needed elsewhere in the body. Currently, hepatocyte iron efflux mechanisms and their relationships to macrophage iron recycling during the control of whole-body iron homeostasis are unclear. We hypothesized that the iron exporter, ferroportin1 (Fpn1), is critical for both iron mobilization from hepatocytes and iron recycling from macrophages. To test this, we generated hepatocyte-specific Fpn1 deletion mice (Fpn1Alb/Alb) and mice that lacked Fpn1 in both hepatocytes and macrophages (Fpn1Alb/Alb;LysM/LysM). When fed a standard diet, Fpn1Alb/Alb mice showed mild hepatocyte iron retention. However, red blood cell (RBC) counts and hemoglobin (Hb) levels were normal, indicating intact erythropoiesis. When fed an iron-deficient diet, Fpn1Alb/Alb mice showed impaired liver iron mobilization and anemia, with much lower RBC and Hb levels than Fpn1flox/flox mice on the same diet. Using a strategy where mice were preloaded with differing amounts of dietary iron before iron deprivation, we determined that erythropoiesis in Fpn1Alb/Alb and Fpn1flox/flox mice depended on the balance between storage iron and iron demands. On a standard diet, Fpn1Alb/Alb;LysM/LysM mice displayed substantial iron retention in hepatocytes and macrophages, yet maintained intact erythropoiesis, implying a compensatory role for intestinal iron absorption. In contrast, when Fpn1Alb/Alb;LysM/LysM mice were fed an iron-deficient diet, they developed severe iron-deficiency anemia, regardless of their iron storage status. Thus, Fpn1 is critical for both hepatocyte iron mobilization and macrophage iron recycling during conditions of dietary iron deficiency. Conclusion: Our data reveal new insights into the relationships between Fpn1-mediated iron mobilization, iron storage, and intestinal iron absorption and how these processes interact to maintain systemic iron homeostasis. (HEPATOLOGY 2012;56:961–971)

Iron is essential for a wide spectrum of biologic functions, including oxygen transport and cellular respiration. Iron deficiency can cause cellular growth arrest and death, whereas iron excess can lead to oxidative stress that is detrimental to cell membranes, proteins, and nucleic acids.1 Because both iron deficiency and iron overload are harmful to cells and tissues, iron balance is meticulously regulated to meet the body's iron requirements and to avoid the toxicity associated with iron excess.2, 3

The liver has first-pass access to dietary nutrients and can readily take up an amount of circulating iron that exceeds immediate metabolic needs to minimize the potential toxicity of free iron.4 In human or animal models with hereditary hemochromatosis, acquired iron overload,4, 5 or congenital atransferrinemia,6, 7 iron is primarily sequestrated in hepatocytes and is stored in the form of ferritin. When iron is needed elsewhere in the body, the sequestered iron can be actively mobilized into circulation.8

The mechanism of iron export from hepatocytes is not known, but evidence suggests that the iron exporter, ferroportin1 (Fpn1), may be involved.9 Fpn1 is the only mammalian nonheme iron exporter identified to date and, consistent with its function in iron efflux, is mainly expressed in cell types that play critical roles in iron metabolism, such as placental syncytiotrophoblast cells, duodenal mature enterocytes, reticuloendothelial macrophages, and hepatocytes.10, 11

Fpn1 is expressed on the hepatocyte surface that faces the sinusoids around portal veins.10, 12 Hepcidin, a short peptide mainly secreted by the liver, is able to bind Fpn1 and induce its internalization and degradation, thereby decreasing iron efflux.13 In a primary hepatocyte culture system, both endogenously secreted and synthetic hepcidin were shown to efficiently decrease hepatocyte Fpn1 expression and increase cellular iron levels. In contrast, hepcidin-deficient mice displayed increased Fpn1 expression in hepatocytes.14 Fpn1 expression was also found to be increased in iron-loaded liver hepatocytes and Kupffer cells.15 Furthermore, human autosomal dominant FPN1 mutations that lead to a loss of function of FPN1 cause iron deposition primarily in macrophages, but also in hepatocytes.16, 17 In mice, Fpn1 deficiency leads to iron retention in enterocytes, macrophages, and, to a lesser extent, in hepatocytes,18 indicating a role for Fpn1 in hepatocyte iron release.

The degree to which Fpn1 controls hepatocyte iron export remains unclear. For example, flatiron mice, which have a missense mutation in Fpn1 that alters its localization and iron export activity, display excessive iron deposition in macrophages, but not hepatocytes; this pattern is also observed in some patients with Fpn1 missense mutations.19, 20 Furthermore, Fpn1 appears to have robust expression in macrophages and duodenum enterocytes, but lower expression in hepatocytes.10, 21 These data suggest that the role of Fpn1 in hepatocytes might not be as important as it is in duodenal enterocytes or macrophages. Moreover, several studies have shown that hepatocytes can secrete the iron storage protein, ferritin, a process that can be enhanced by iron stimulation.22, 23 Hepatocytes also express high levels of the heme transporter, feline leukemia virus subgroup C receptor.24 Therefore, Fpn1-independent pathways might regulate the efflux of iron from hepatocytes.

To elucidate the role of Fpn1 in hepatocyte iron metabolism and its contributions to whole-body iron homeostasis, we developed mouse models with Fpn1 deletion in hepatocytes or in both hepatocytes and macrophages. Mice with different iron storage levels and experiencing varying iron demand were obtained by using mice of different ages or by perturbing iron status by controlled iron administration and phlebotomy. Subsequently, mice were fed either a standard diet or an iron-deficient diet to investigate the roles of Fpn1 in both hepatocyte iron export and macrophage iron recycling and to elucidate the relationships between Fpn1-mediated iron mobilization, iron storage, and intestinal iron absorption.

Abbreviations

Alb, albumin; FAC, ferric ammonium citrate; Fpn1, ferroportin1; Hb, hemoglobin; HFE, hemochromatosis protein; HIS, high iron storage; ICP, inductively coupled plasma; IHC, immunohistochemistry; LFH, low iron storage, fast growth rate, and high iron demands; MIS, moderate iron storage; NSH, normal iron storage, slow growth rate, and high iron demands; NSL, normal iron storage, slow growth rate, and low iron demands; RBC, red blood cell SEM, standard error of the mean; UIBC, unsaturated iron binding capacity.

Materials and Methods

Animals.

Albumin (Alb)-Cre (Alb-Cre) mice25 maintained on a 129/SvEvTac background were mated with Fpn1flox/flox animals to generate Fpn1Alb/Alb mice. Genotyping was performed as previously described.18 The Alb-Cre recombinase was genotyped with forward primer 5′- GCAAACATACGCAAGGGATT-3′ and reverse primer 5′- AGGCAAATTTTGGTGTACGG-3′, resulting in a 350-base-pair band. Fpn1Alb/Alb mice were mated with Fpn1LysM/LysM mice26 to generate Fpn1Alb/Alb;LysM/LysM mice. Mice were fed a standard rodent laboratory diet (232 mg iron/kg) from SLRC Laboratory Animal Co. Ltd. (Shanghai, China). All experimental protocols were approved by the Institutional Animal Care and Use Committee of the Chinese Academy of Sciences (Shanghai, China).

Animal Treatment.

Age- and sex-matched mice were used in separate experiments. To induce anemia, 3-week-old mice (with low iron storage, but at fast growth phase and with high iron demands; referred to as LFH mice) were fed an iron-deficient diet for 5 weeks; 2-month-old mice (with normal iron storage, and at slow growth phase and with low iron demands; referred to as NSL mice) were fed an iron-deficient diet for 8 weeks; 0.5 mL of blood was extracted by retro-orbital puncture from 2-month-old mice, the procedure was repeated 1 week later, and these animals (with normal iron storage, but at slow growth phase and with high iron demands; referred to as NSH mice) were fed an iron-deficient diet for 4 weeks from the day of the first phlebotomy. For iron preloading, 3-week-old mice were fed an iron-rich diet for 1 week (with moderate iron storage; referred to as MIS mice) and then were fed an iron-deficient diet for 5 weeks. Three-week-old mice were fed an iron-rich diet for 4 weeks (with high iron storage; referred to as HIS mice) and then were fed an iron-deficient diet for 8 weeks. Both the iron-deficient diet (0.9 mg iron/kg) and iron-rich diet (8.3 g of carbonyl iron per kg) were egg-white–based AIN-76A-diets (Research Diets, Inc., New Brunswick, NJ).

Measurement of Serum Iron and Hematological Parameters.

Measurement of serum iron and hematological parameters were performed as previously described.26

Tissue Nonheme Iron Assay and Hepatocyte Total Iron Analysis.

Quantitative measurement of tissue nonheme iron was performed as previously described.27 For hepatocyte total iron analysis, hepatocytes (1 × 107) were boiled in 70% HNO3 (Tama Chemicals Co., Ltd., Kanagawa, Japan) under a microwave digestion system. Samples were analyzed using an Agilent (7700X; Agilent Technologies, Santa Clara, CA) inductively coupled plasma (ICP) mass spectrometer.

Immunohistochemistry and Iron Staining.

Immunohistochemical (IHC) detection of Fpn118, 26 and nonheme iron staining was performed as previously described.18

Hepatocyte Isolation and Culture.

Primary hepatocytes were isolated by collagenase digestion. Live hepatocytes were separated by Percoll (Sigma-Aldrich, St. Louis, MO) centrifugation and then were seeded on collagen-coated six-well plates. For hepatocyte iron-release experiments, 10 μM of ferric ammonium citrate (FAC) was applied to cell-culture medium for 24 hours, hepatocytes were washed, fresh culture medium was added, and then cells were harvested at different time points for analysis.

Hepatocyte Membrane Protein Isolation.

The protocol for hepatocyte membrane protein isolation was provided by Dr. Mitchell Knutson.28

Western Blotting.

Western blotting for Fpn1 and H-ferritin was performed as previously described.26

Statistical Analysis.

Data are presented as means ± standard error of the mean (SEM). Statistical analyses between various groups were performed by the two-tailed Student's t test, with a P value of less than 0.05 considered statistically significant.

Results

Alb-Cre Recombinase-Mediated Disruption of Fpn1 Results in Hepatocyte Cellular Iron Retention.

To gain insight into the role of Fpn1 in hepatocyte iron homeostasis, we generated mice that lacked Fpn1 in hepatocytes. Real-time polymerase chain reaction results suggested that Fpn1 expression was reduced by approximately 99% in isolated primary hepatocytes (Supporting Fig. 1A), and this was also validated by western blotting results from isolated membranes of primary hepatocytes (Supporting Fig. 1B).

To investigate the role of Fpn1 in hepatocyte iron export, we determined tissue nonheme iron content as well as serum and blood parameters in mice of varied ages. Fpn1Alb/Alb mice exhibited significantly higher nonheme iron levels in livers and equivalent nonheme iron levels in spleens, compared with Fpn1flox/flox mice (Fig. 1A,B). Perls' Prussian Blue iron staining of tissue sections also supported the results. Fpn1Alb/Alb mice showed pronounced iron accumulation in hepatocytes, mainly around the portal vein, whereas no iron staining was observed in hepatocytes of Fpn1flox/flox mice (Supporting Fig. 2A). In addition, no differences were found in iron-staining results for spleens. This was further confirmed by H-ferritin western blotting results (Supporting Fig. 2B). Furthermore, Fpn1Alb/Alb mice revealed a trend toward slightly decreased serum iron concentrations and increased unsaturated iron-binding capacity (UIBC), compared to their littermate controls (Fig. 1E,F), but these differences reached statistical significance only in 3-week-old mice. However, no differences were found in red blood cell (RBC) counts, hemoglobin (Hb) content (Fig. 1C,D), or growth rates (data not shown). Taken together, these results indicated that Fpn1 was, at least partially, responsible for hepatocyte iron release into the circulation in healthy iron status. However, iron levels in the circulation of Fpn1Alb/Alb mice seemed to be high enough to meet the body's needs; thus, Fpn1 deletion in hepatocytes had no severe effect on RBC and Hb levels when mice were fed a standard diet.

Figure 1.

Disruption of Fpn1 in mouse hepatocytes results in iron accumulation in the liver. (A) Liver nonheme iron concentrations, (B) spleen nonheme iron concentrations, (C) RBC value, (D) Hb value, (E) serum-iron concentrations, and (F) UIBC levels were measured in sex-matched 3-week-old (n = 6 for each group; 3 male and 3 female), 2-month-old (n = 10 for each group; 5 male and 5 female), and 10-month-old (n = 6 for each group; 3 male and 3 female) Fpn1flox/flox and Fpn1Alb/Alb mice. Results are presented as mean ± SEM. *P < 0.05; **P < 0.01.

The Deletion of Fpn1 in Hepatocytes Impairs Iron Release Ability.

To confirm the role of Fpn1 in hepatocyte iron export, we isolated primary hepatocytes from Fpn1flox/flox and Fpn1Alb/Alb mice. More accumulated iron was found in Fpn1Alb/Alb hepatocytes, with amounts nearly twice that of Fpn1flox/flox hepatocytes (Fig. 2A). For iron-release assays, both Fpn1flox/flox and Fpn1Alb/Alb hepatocytes acquired iron equally well; this was also validated in mice that were fed an iron-rich diet for either 1 or 4 weeks (Supporting Fig. 3A). However, Fpn1Alb/Alb hepatocytes failed to divest themselves of iron effectively, as indicated by a continued high level of H-ferritin after a 12-hour wash, whereas Fpn1flox/flox hepatocytes efficiently released overloaded iron, as indicated by decreased H-ferritin levels over the 12-hour wash period (Fig. 2B). These data demonstrate that Fpn1 is critical for nonheme iron efflux from hepatocytes.

Figure 2.

Fpn1-deficient hepatocytes exhibit cellular iron retention and impaired iron release. (A) Total iron from primary hepatocytes was measured by ICP mass spectrometer and quantified as micrograms of iron per milligram of protein. (B) Primary hepatocytes from 2-month-old Fpn1flox/flox and Fpn1Alb/Alb mice were loaded with iron by incubation with FAC (10 μM) overnight. After removing the medium, cells were washed and then incubated in fresh medium without iron supplementation for 12 hours. H-ferritin levels, an indicator of cellular iron stores, were measured by western blotting analysis, the results were confirmed by three independent experiments, and H-ferritin/actin ratios were also quantitated by densitometry. Results are presented as mean ± SEM. *P < 0.05; **P < 0.01.

Fpn1 Is Required for Iron Mobilization From Hepatocytes During Anemia Conditions.

Fpn1Alb/Alb mice looked overtly healthy. They showed elevated iron retention, specifically in hepatocytes, but with only slight changes in serum-iron parameters. We therefore predicted that Fpn1 might be necessary for hepatocyte iron mobilization under iron-deficient conditions. To address this possibility, LFH mice (3 weeks old), NSL mice (2 months old), and NSH mice (2 months old, with phlebotomy) models were used for further testing.

After the administration of an iron-deficient diet, although Fpn1Alb/Alb mice had iron stored in hepatocytes, the stored iron was not mobilized normally. Conversely, Fpn1flox/flox mice efficiently mobilized iron from hepatocytes and macrophages. Consequently, iron levels in livers of Fpn1Alb/Alb NSL and NSH mice were approximately 4-fold higher than those of Fpn1flox/flox mice (Fig. 3A). Because weanling mice had a lower baseline iron storage status and were in a growth phase (with livers also growing quickly), the iron-deficient-diet–treated LFH mice displayed the lowest liver-iron content (Fig. 3A). Not surprisingly, the difference between liver-iron content in Fpn1flox/flox LFH and Fpn1Alb/Alb LFH mice was also minimized (Fig. 3A). Of note, healthy splenic iron mobilization occurred in Fpn1Alb/Alb mice and their splenic iron mobilization capacity was intact (Fig. 3B). Once iron absorption was restricted by the iron-deficient diet treatment, Fpn1Alb/Alb mice showed more severe anemia than Fpn1floxflox mice (Fig. 3C-F). This finding could have been the result of defective hepatocyte iron mobilization. Furthermore, we found that LFH and NSH mice suffered a more severe anemia than NSL mice when fed an iron-deficient diet, regardless of whether they were on the Fpn1floxflox or Fpn1Alb/Alb background (Fig. 3C,D).

Figure 3.

Fpn1 is responsible for iron release from hepatocytes in mice during iron-deficient conditions. (A) Liver nonheme iron concentrations, (B) spleen nonheme iron concentrations, (C) RBC value, (D) Hb value, (E) serum-iron concentrations, and (F) UIBC levels were measured in sex-matched LFH (3-week-old mice; n = 6 for each group; 3 male and 3 female), NSL (2-month-old adult mice; n = 6 for each group; 3 male and 3 female), and NSH (2-month-old adult phlebotomized mice; n = 6 for each group; 3 male and 3 female) Fpn1flox/flox and Fpn1Alb/Alb mice after consuming an iron-deficient diet for 5, 8, or 4 weeks, respectively. Fpn1Alb/Alb mice displayed impaired iron release from liver hepatocytes and more severe anemia, as compared with Fpn1flox/flox mice of the same group. LFH and NSH mice showed more severe anemia, compared to NSL mice. Results are presented as mean ± SEM. *P < 0.05; **P < 0.01.

These observations support our hypothesis that Fpn1 is responsible for iron mobilization from hepatocytes under iron-deficient conditions and also reiterate the importance of iron storage and iron mobilization for erythropoiesis.

Hepatocyte Iron Mobilization Mediated by Fpn1 Is Essential During Iron-Deficient Conditions.

To further elucidate the role of Fpn1 in iron mobilization, and the importance of hepatocyte iron storage, we pre-loaded iron stores in mice by feeding an iron-rich diet for either 1 or 4 weeks to generate MIS or HIS mice, respectively.

Both Fpn1flox/flox and Fpn1Alb/Alb mice acquired dietary iron equally well, (Supporting Fig. 3). Iron was mobilized efficiently in Fpn1flox/flox MIS mice, but was greatly impaired in Fpn1Alb/Alb MIS mice after iron-deficiency treatment, and the liver-iron content in Fpn1Alb/Alb MIS mice was approximately 13-fold higher than that of Fpn1flox/flox MIS mice (Fig. 4A). Because iron mobilization from hepatocytes was impaired in Fpn1Alb/Alb MIS mice, more iron was mobilized from macrophages, as indicated by the reduced spleen-iron content (Fig. 4B). However, the moderate iron storage in macrophages and other cells could not compensate for the whole-body iron needs of these mice. Thus, iron-deficient-diet–treated Fpn1Alb/Alb MIS mice showed anemia with decreased RBC and Hb, whereas Fpn1flox/flox MIS mice of the same treatment appeared to be healthy with regular RBC and Hb (Fig. 4C,D). Importantly, Fpn1Alb/Alb MIS mice showed significantly lower serum-iron levels and higher UIBC levels than that of Fpn1flox/flox MIS mice during these iron-deficient conditions (Fig. 4E,F).

Figure 4.

Hepatocyte iron storage and Fpn1-mediated iron mobilization were essential for meeting whole-body iron demands during iron-deficient conditions. (A) Liver nonheme iron concentrations, (B) spleen nonheme iron concentrations, (C) RBC value, (D) Hb value, (E) serum-iron concentrations, and (F) UIBC levels were measured in sex-matched MIS (3-week-old mice fed an iron-rich diet for 1 week, with moderate iron storage; n = 6 for each group; 3 male and 3 female), HIS (3-week-old mice fed an iron-rich diet for 4 weeks, with high iron storage; n = 6 for each group; 3 male and 3 female) Fpn1flox/flox and Fpn1Alb/Alb mice after consuming an iron-deficient diet for 5 or 8 weeks, respectively. Fpn1Alb/Alb MIS mice displayed greatly impaired iron release from hepatocytes and anemia phenotype, whereas Fpn1flox/flox MIS mice appeared healthy with regular RBC and Hb levels. Fpn1Alb/Alb HIS mice did not show an anemic phenotype, because sufficient iron was mobilized from macrophages, which was indicated by low splenic iron content (B). Results are presented as mean ± SEM. *P < 0.05; **P < 0.01.

In HIS mice, both Fpn1flox/flox HIS and Fpn1Alb/Alb HIS mice showed healthy phenotypes without anemia after iron-deficient treatment (Fig. 4C,D), and no significant difference in liver iron content or splenic iron content was found between them (Fig. 4A,B). However, Fpn1Alb/Alb HIS mice showed significantly lower serum-iron levels and higher UIBC levels than Fpn1flox/flox HIS mice during iron-deficient conditions (Fig. 4E,F).

These results indicated that Fpn1 was critical for iron mobilization from hepatocytes into the circulation to meet body iron needs, and suggested that iron mobilization from hepatocytes was especially important in mice with moderate iron storage under iron-deficient conditions. When iron storage was high, iron mobilized from macrophages and other tissues was enough for body demands; thus, iron mobilization from hepatocytes seemed less critical under these conditions.

The Contributions of Hepatocyte and Macrophage Fpn1 for Whole-Body Iron Homeostasis at Normal Iron Status.

Our previous study showed that mice with Fpn1 deletion in macrophages (Fpn1LysM/LysM) exhibited only a mild anemia.26 We therefore postulated that intestinal iron absorption or iron mobilization from other tissues might compensate for the iron-deficient phenotype caused by an iron-recycling defect in macrophages. In our current study, we found that Fpn1 deletion in hepatocytes had little effect in mice when fed a standard diet. To further examine the contributions of hepatocyte and macrophage Fpn1 in regulating whole-body iron homeostasis, we developed mice with Fpn1 deletion in both hepatocytes and macrophages. Fpn1Alb/Alb;LysM/LysM mice appeared overtly healthy, whereas they were maintained on a standard diet. Furthermore, iron loading was observed in both hepatocytes and macrophages of Fpn1Alb/Alb;LysM/LysM mice (Supporting Fig. 2A), liver nonheme iron concentrations of Fpn1Alb/Alb;LysM/LysM mice were approximately 3-fold higher than that of Fpn1flox/flox mice (Fig. 5A). The nonheme iron content in the spleens of Fpn1Alb/Alb;LysM/LysM mice was similar to that previously found in Fpn1LysM/LysM mice.26 Moreover, Fpn1Alb/Alb;LysM/LysM mice had significantly reduced serum-iron levels and increased UIBC levels, compared with Fpn1flox/flox mice (Fig. 5E,F). Despite impaired iron release and iron recycling from both hepatocytes and macrophages, little decrease was found in either RBC or Hb levels (Fig. 5C,D). These data, combined with data from our previous study, showed that Fpn1 was critical for mediating iron release from hepatocytes and macrophages.26 If there were no other major iron-export pathways, the mild phenotype of Fpn1Alb/Alb;LysM/LysM mice could be explained by increased intestinal iron absorption from a standard diet replete with iron.

Figure 5.

Mice with Fpn1 deletion in both hepatocytes and macrophages showed elevated iron retention at healthy status. (A) Liver nonheme iron concentrations, (B) spleen nonheme iron concentrations, (C) RBC value, (D) Hb value, (E) serum-iron concentrations, and (F) UIBC levels were measured in sex-matched 3-week-old (n = 6 for each group; 3 male and 3 female) and 2-month-old (n = 6 for each group; 3 male and 3 female) Fpn1flox/flox and Fpn1Alb/Alb;LysM/LysM mice. Fpn1Alb/Alb;LysM/LysM mice showed increased iron retention in livers, significantly reduced serum-iron levels, and increased UIBC levels, compared with Fpn1Alb/Alb mice. RBC and Hb levels in Fpn1Alb/Alb;LysM/LysM mice appeared healthy. Control Fpn1flox/flox mice were the same as those presented in Fig. 1. Results are presented as mean ± SEM. *P < 0.05; **P < 0.01.

The Contributions of Hepatocyte and Macrophage Fpn1 for Whole-Body Iron Homeostasis at Iron-Deficient Status.

To further investigate the importance of intestinal iron absorption and elucidate the contributions of Fpn1 in hepatocyte and macrophage iron mobilization, LFH and NSL mice models were selected. After iron-deficiency treatment, relative to Fpn1flox/flox mice, iron mobilization from both hepatocytes and macrophages was greatly impaired. The liver-iron concentration in Fpn1Alb/Alb;LysM/LysM LFH mice was approximately 13-fold higher than that of Fpn1flox/flox LFH mice (Fig. 6A). The corresponding values for splenic iron content were 45-fold greater (Fig. 6B). The liver-iron concentration in Fpn1Alb/Alb;LysM/LysM NSL mice was approximately 18-fold higher than that of Fpn1flox/flox NSL mice under iron-deficient conditions (Fig. 6A). The splenic iron concentration in Fpn1Alb/Alb;LysM/LysM NSL mice was approximately 7-fold higher than that of Fpn1flox/flox NSL mice (Fig. 6B), but the Fpn1Alb/Alb;LysM/LysM NSL mice had significantly enlarged spleens, compared with Fpn1flox/flox NSL mice (Supporting Fig. 4D). Thus, the total splenic iron content in Fpn1Alb/Alb;LysM/LysM NSL mice was approximately 21-fold higher than that of Fpn1flox/flox NSL mice.

Figure 6.

Severe iron-deficiency anemia in Fpn1Alb/Alb;LysM/LysM mice during iron-deficient conditions, regardless of iron-store status, indicating that Fpn1 is critical for iron mobilization from both hepatocytes and macrophages. (A) Liver nonheme iron concentrations, (B) spleen nonheme iron concentrations, (C) RBC value, (D) Hb value, (E) serum-iron concentrations, and (F) UIBC levels were measured in sex-matched LFH (3-week-old mice; n = 6 for each group; 3 male and 3 female) and NSL (2-month-old mice; n = 6 for each group; 3 male and 3 female) Fpn1flox/flox and Fpn1Alb/Alb;LysM/LysM mice after consuming an iron-deficient diet for 5 or 8 weeks, respectively. Fpn1Alb/Alb;LysM/LysM mice showed iron-release defects from both hepatocytes and macrophages, whereas Fpn1flox/flox mice released liver and spleen iron efficiently. Consequently, Fpn1Alb/Alb;LysM/LysM displayed severe iron-deficiency anemia. Control Fpn1flox/flox mice were the same as those presented in Fig. 3. Results are presented as mean ± SEM. *P < 0.05; **P < 0.01.

When iron absorption was restricted, as a result of cumulative defects in iron absorption, iron mobilization, and iron recycling, Fpn1Alb/Alb;LysM/LysM LFH and NSL mice both suffered severe anemia, regardless of their iron storage status, indicated by strongly decreased RBC, Hb, and serum-iron levels (Fig. 6C-F). The anemic phenotype was also confirmed by blood-smear results (Supporting Fig. 4A). Consequently, the growth rate of Fpn1Alb/Alb;LysM/LysM LFH mice was slower, they were significantly smaller than Fpn1flox/flox LFH mice, and they displayed hair loss (Supporting Fig. 4B,C).

Together, these results, first, demonstrate that Fpn1 plays a critical role in iron mobilization from both hepatocytes and macrophages. Second, iron mobilization from both hepatocytes and macrophages is critical under iron-deficient conditions. Third, intestinal iron absorption plays an important role in compensating for iron deficiency caused by iron-mobilization defects in both hepatocytes and macrophages when dietary iron is sufficient.

The Importance of Intestinal Fpn1 for Whole-Body Iron Homeostasis.

To further validate that iron absorption compensated for the iron mobilization and recycling defect in Fpn1Alb/Alb;LysM/LysM mice, we examined Fpn1 expression in the duodenum of Fpn1flox/flox and Fpn1Alb/Alb;LysM/LysM mice by IHC. When fed a standard diet, Fpn1Alb/Alb;LysM/LysM mice showed higher intestinal Fpn1 expression than Fpn1flox/flox mice, especially at 3 weeks of age. After iron-deficient diet treatment, Fpn1 expression was increased in both Fpn1flox/flox and Fpn1Alb/Alb;LysM/LysM mice, compared to untreated mice of the same age and genotype. However, Fpn1Alb/Alb;LysM/LysM mice exhibited higher Fpn1 expression than Fpn1flox/flox mice (Fig. 7). These findings indicated that Fpn1 expression in the duodenum could be regulated by iron-demand status, and suggested that higher Fpn1 expression in the duodenum could compensate for iron mobilization and recycling defects under conditions of abundant dietary iron (Fig. 8).

Figure 7.

Compensatory role of intestinal iron absorption in Fpn1Alb/Alb;LysM/LysM mice. Detection of duodenal Fpn1 expression in LFH (3-week-old, top panel) and NSL (2-month-old, middle panel) mice or NSL Fpn1flox/flox and Fpn1Alb/Alb;LysM/LysM mice fed an iron-deficient diet for 2 additional months (bottom panel). Duodenum sections were incubated with a rabbit anti-Fpn1 primary antibody and goat antirabbit immunoglobulin G/horseradish peroxidase secondary antibody. Diaminobenzidine staining was used to visualize Fpn1, and nuclei were counterstained with hematoxylin. Brown staining indicates the expression of Fpn1 on the basolateral surface of enterocytes. Fpn1Alb/Alb;LysM/LysM mice showed increased Fpn1 expression in duodenum enterocytes relative to Fpn1flox/flox mice. Microscope: Olympus BX61; Objective lens: UPlanApo 10×/0.40 (Tokyo, Japan). Camera: QImaging QICAM, Fast 1394 (Burnaby, British Columbia, Canada). Imaging software: Q Capture 2.90.1 (Quantitative Imaging, Surrey, British Columbia, Canada). Software used to manipulate images: Adobe Photoshop CS4 11.0.

Figure 8.

Model for the network among Fpn1-mediated iron mobilization from hepatocytes and macrophages, iron-storage status, and intestinal iron absorption. At healthy status, iron homeostasis in the body is balanced by iron absorption through duodenum enterocytes, iron recycling from macrophages, and iron stored and released from hepatocytes. Blocking of intestinal iron absorption results in iron mobilization from both hepatocytes and macrophages, with erythropoiesis activity dependent upon iron storage and iron-demand status. Blocking iron mobilization from hepatocytes and recycling from macrophages by Fpn1 deletion resulted in compensatory absorption of iron from duodenum enterocytes, and erythropoiesis activity depended on dietary iron content. Blocking of intestinal iron absorption and iron mobilization from both hepatocytes and macrophages (Fpn1 deletion in hepatocytes and macrophages) results in severe anemia, regardless of iron-store status; this is ascribed to the lack of circulating iron for erythropoiesis activity.

Discussion

To investigate the physiological role of Fpn1 in hepatocyte iron metabolism in vivo, we generated mice with a specific Fpn1 deletion in hepatocytes. Fpn1Alb/Alb mice showed mild hepatocyte iron retention and also displayed slightly lower serum iron and higher UIBC than their littermate controls when maintained on a standard diet (Fig. 1A,E,F). However, Fpn1Alb/Alb mice were overtly healthy, with no differences in RBC and Hb levels relative to their littermate controls (Fig. 1C,D). These results indicate that Fpn1 was responsible for hepatocyte iron export into the circulation at healthy status, but did not affect erythropoiesis. However, 3-week-old weanling Fpn1Alb/Alb mice, which were in a rapid growth phase and had increased iron needs, showed slightly decreased RBC and Hb levels, compared to their littermate controls (Fig. 1C,D). Moreover, sufficient iron was provided through the standard diet. Based on these findings, we hypothesized that Fpn1 might be more important during iron-deficient dietary conditions or other conditions requiring increased iron demands.

To further investigate our hypothesis, we used three mouse models (i.e., LFH, NSL, and NSH) with different iron storage and iron-demand status. All three types of mice were then fed an iron-deficient diet. As Fpn1 was deleted in hepatocytes of Fpn1Alb/Alb mice, defective iron mobilization from hepatocytes was observed (Fig. 3A). Consequently, Fpn1Alb/Alb mice suffered a more severe anemia, compared with Fpn1flox/flox mice (Fig. 3C,D). Collectively, these data provided evidence that Fpn1 was responsible for iron mobilization from hepatocytes, especially under iron-deficient conditions. Furthermore, we found that LFH and NSH mice developed more severe anemia than NSL mice at iron-deficient conditions, whether in Fpn1flox/flox or Fpn1Alb/Alb mice (Fig. 3C-F). These findings indicated that both iron storage and iron-demand status correlated with the anemic phenotype.

To further investigate the importance of iron storage and the role of Fpn1 in hepatocyte iron mobilization, we used an iron-rich diet to preload animals with different levels of iron stores (e.g., MIS and HIS), then restricted iron absorption by administering an iron-deficient diet. After iron-deficient diet treatment, the liver-iron content of Fpn1Alb/Alb MIS mice was approximately 13-fold greater than that of Fpn1flox/flox MIS mice, wherein Fpn1 was deleted in hepatocytes (Fig. 4A). As a result, Fpn1Alb/Alb MIS mice suffered obvious anemia. However, Fpn1flox/flox MIS mice appeared healthy because they efficiently mobilized preloaded iron for erythropoiesis (Fig. 4C,D). Furthermore, in iron-deficient-diet–treated HIS mice, neither Fpn1flox/flox nor Fpn1Alb/Alb mice developed anemia (Fig. 4A,C,D). This observation could be explained by the fact that iron stores in macrophages and other tissues were sufficient for the body's needs. This idea was supported by the observed decreases in splenic iron content in Fpn1Alb/Alb HIS mice relative to Fpn1flox/flox HIS mice (Fig. 4B). These results provided further evidence that Fpn1 was critical for iron mobilization from hepatocytes to meet the body's needs when the body had moderate iron stores, but when iron stores were extremely high, iron stored in macrophages and other cells was enough for the body's needs. Thus, Fpn1-mediated iron mobilization from hepatocytes was not essential under these conditions. These findings underscore the importance of moderate iron storage in meeting the body's iron needs.

HIS mice fed an iron-deficient diet had significantly lower serum iron and higher UIBC than did untreated healthy mice of the same age, regardless of Fpn1flox/flox or Fpn1Alb/Alb genotype (compare Figs. 4E,F and 1E,F). Because both Fpn1flox/flox HIS and Fpn1Alb/Alb HIS mice had healthy RBC and Hb levels, we postulated that the lower serum-iron levels were sufficient to meet the body's needs. However, mice absorbed more iron than needed and stored it in hepatocytes or macrophages when fed a standard diet. This conclusion was supported by the observed increases in tissue-iron content during age-related growth. These findings revealed a mechanism by which the body absorbed more iron than it needed and then stored it to facilitate the maintenance of healthy status in the case of future iron deficiency.

In looking at splenic iron levels, we found that Fpn1Alb/Alb mice had slightly lower splenic iron stores than did Fpn1flox/flox mice under both healthy and iron-deficient dietary conditions (Figs. 1B, 3B, and 4B). These results indicated that the body could mobilize iron from the macrophage iron pool to meet demands when iron mobilization from hepatocytes was impaired. To gain more detailed insight into the role of Fpn1 in iron mobilization, we generated mice in which both hepatocytes and macrophages lacked Fpn1.

Fpn1Alb/Alb;LysM/LysM mice appeared normal with only minor reductions in RBC and Hb levels, although they displayed more liver-iron accumulation, reduced serum iron, and increased UIBC when they were fed a standard diet (Fig. 5A,E,F). Our previous study had shown that Fpn1LysM/LysM mice displayed only mild anemia, which highlighted the important contribution of intestinal iron absorption in compensating for macrophage iron-recycling defects caused by Fpn1 loss or iron mobilization from other cells.26 The importance of iron absorption had been demonstrated by Donovan et al.18 In this study, once the iron exporter, Fpn1, was no longer expressed in the intestine, mice suffered severe anemia, which could be corrected by iron-dextran injections to boost iron stores.18 To elucidate the role of intestinal iron absorption and iron mobilization, we treated Fpn1Alb/Alb;LysM/LysM mice with an iron-deficient diet. Because iron mobilization from both hepatocytes and macrophages was seriously impaired (Fig. 6A,B), Fpn1Alb/Alb;LysM/LysM mice developed severe anemia (Fig. 6C,D). These results indicate that, first, Fpn1 is critical for iron mobilization from both hepatocytes and macrophages, and, second, iron absorption could compensate for iron demands when iron mobilization was impaired from both hepatocytes and macrophages, provided mice were fed a standard diet with replete iron. This conclusion was confirmed by increased intestinal Fpn1 expression in Fpn1Alb/Alb;LysM/LysM mice (Fig. 7). Third, iron mobilization from both hepatocytes and macrophages through Fpn1 was essential for meeting the body's iron needs when mice were suffering from dietary iron restriction (Fig. 8).

In conclusion, these findings indicate the important contributions of iron absorption, iron storage, and iron mobilization (i.e., Fpn1 mediated) in maintaining whole-body iron homeostasis. Our study might also provide clues for the treatment of some iron-metabolism disorders. For example, loss of function in hemochromatosis protein (HFE), a protein associated with hereditary hemochromatosis, led to an approximately 50% reduction in hepcidin levels,29 which resulted in higher iron absorption through increased Fpn1 expression in the intestine. If HFE-mutant patients had low dietary iron availability, they might display lower iron absorption. Moreover, iron loss through pregnancy and menstruation is thought to contribute to the morbidity of hereditary hemochromatosis.30 Together, these factors might explain the extremely variable clinical symptoms in patients with HFE mutations, because some HFE-mutation patients remain without overt clinical symptoms throughout their lives.31 However, in some juvenile hemochromatosis cases, patients have no detectable hepcidin expression.32 This complete loss of iron regulation by FPN1 in the intestine may have led to very high iron absorption rates, resulting in early disease onset in these patients.33 Although phlebotomy treatment could mobilize iron stores, these patients would quickly reaccumulate dietary iron through high iron-absorption efficiency. Research on controlling the amount of dietary iron and iron mobilization in these patients, finding antagonists for FPN1, or otherwise disturbing intestinal iron-absorption status would be helpful in controlling these diseases.

Finally, the relationship among Fpn1-mediated iron mobilization, iron absorption, and iron storage might provide clues for the clinical treatment of iron-disordered diseases and reconciling them would be beneficial for iron homeostasis in healthy populations under different physiological conditions.

Acknowledgements

The authors thank the other members of the Wang Laboratory for their helpful comments and Dr. Gregory J. Anderson for his helpful discussion.

Ancillary