Presence of the iron exporter ferroportin at the plasma membrane of macrophages is enhanced by iron loading and down-regulated by hepcidin

Ferroportin (aka IREG1 [iron-regulated transporter 1], MTP1 [metal transporter protein-1], Slc40a1) is the sole iron exporter identified in mammals^1-3  that participates in iron release from both enterocytes of the duodenum and tissue macrophages. Ferroportin is highly expressed in absorptive duodenal enterocytes where it presents a strong basolateral subcellular localization^1-3  and in tissue macrophages in liver (Kupffer cells), spleen, and bone marrow.^1,4,5  Several lines of evidence highlight the importance of this protein in iron homeostasis. Inactivation of the ferroportin gene at the adult stage in mice leads to iron accumulation in enterocytes, Kupffer cells, and splenic macrophages.⁶  Ferroportin mutations in human patients with type 4 hemochromatosis induce predominant macrophage iron overload.⁷  Finally, ferroportin overexpressed in the macrophage cell line J774 stimulates iron release after erythrophagocytosis.⁸ 

Recently, ferroportin has been shown to be the molecular target of hepcidin.⁹  Hepcidin is a major systemic regulator of intestinal iron absorption and iron recycling from macrophages.¹⁰  In epithelial cells, hepcidin was shown to act on the efflux of iron through a direct interaction with ferroportin at the cell surface, leading to internalization and degradation of the iron exporter.⁹  Moreover, recent studies have shown that some hemochromatosis-associated ferroportin mutations are unresponsive to hepcidin-mediated internalization.^11-13  Of interest, in hepcidin-deficient mice,¹⁴  ferroportin is strongly up-regulated in both enterocytes and macrophages.¹⁵  Hepcidin is also assumed to regulate iron efflux from macrophages¹⁰  because it has been shown that treatment of J774 macrophages with hepcidin decreases ferroportin expression and significantly reduces iron efflux after erythrophagocytosis.⁸  However, previous studies have reported an intracellular and vesicular distribution of ferroportin in macrophages,^1,4,5,16  not easily compatible with a direct effect of hepcidin on ferroportin. In this work, we used a high-affinity antibody to specify the subcellular localization of endogenous ferroportin in primary culture of murine bone marrow–derived macrophages (BMDMs). We particularly looked at ferroportin expression and localization after iron treatment or erythrophagocytosis. Finally, the effects of a synthetic human hepcidin peptide have been studied on both expression and subcellular localization of the endogenous ferroportin expressed in macrophages.

Bone marrow–derived macrophages were cultured as previously descrided.^17,18  Briefly, bone marrow cells were isolated from femurs of 6- to 8-week-old mice (DBA2 strain) and seeded onto 10-cm diameter Petri dishes for protein extraction or onto glass coverslips in 24-well tissue culture plates for immunofluorescence studies. The culture medium was RPMI-glutamax (Gibco, Invitrogen, Cergy Pontoise, France) supplemented with 10% heat-inactivated fetal calf serum (FCS; low endotoxin content; Gibco), 10% L-cell–conditioned medium (LCCM; source of colony-stimulating factor 1 [CSF-1]), 2 mM l-glutamine, 50 U/mL penicillin, and 50 mg/mL streptomycin. At 4 days after seeding, the adherent cells were rinsed twice with Hanks balance salt solution (HBSS), and the medium was renewed each day until day 7. The mouse monocyte-macrophage cell line J774 was obtained from the American Type Cell Culture (ATCC, Manassas, VA) and was cultured in media and under conditions recommended by ATCC.

To increase or decrease cellular iron concentration, cells were incubated, either with Fe-nitrilotriacetate solution (Fe-NTA; FeCl₃ 100 μM-NTA 400 μM) or with desferrioxamine mesylate (DFO; 100 μM), for indicated periods of time (8 hours or 16 hours). Synthetic human hepcidin peptide (25 amino acids) obtained from Peptides International (Louisville, KY) was diluted in sterile water to obtain a 0.1 mM stock solution, divided into aliquots, and stored at -20°C until use. In all experiments hepcidin was diluted in cell culture media to reach a final concentration of 700 nM. To test the effect of hepcidin on endogenous ferroportin protein expression, the peptide was added alone or simultaneously with Fe-NTA. To test the effect of hepcidin on the cellular distribution of ferroportin, cells were pretreated with Fe-NTA for 16 hours, and hepcidin was then added to the iron-rich culture media for the indicated period of time. As control, sterile water was used instead of Fe-NTA, DFO, or hepcidin.

Erythrophagocytosis assay was performed using human red blood cells (RBCs) artificially aged as previously described.¹⁷  Briefly, washed RBCs (1 × 10⁸/mL) were incubated for 16 hours at 30°C in HEPES buffer (10 mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid], 140 mM NaCl, bovine serum albumin [BSA] 0.1%, pH 7.4) containing calcium (2.5 mM) and the Ca²⁺ ionophore A23187 (0.5 μM; Streptomyces chartreusis; Calbiochem, La Jolla, CA). Treated RBCs were washed and incubated (3 × 10⁷ cells/mL) with cultured BMDMs for 1 hour at 37°C in a 5% CO₂ incubator. Macrophages were washed twice with HBSS and incubated for 5 minutes in hypotonic solution (140 mM NH₄Cl, 17 mM Tris [tris(hydroxymethyl)aminomethane]–HCl pH 7.6) to lyse noningested RBCs. Cells were then maintained in culture for 7 hours prior to analysis.

For the production of the rabbit polyclonal anti–mouse ferroportin antibody, a fusion protein containing a 80-amino acid sequence length (residues 224 to 304) of mouse ferroportin protein fused in-frame to glutathione-S-transferase (GST) was constructed, affinity purified on GST columns, and injected into New Zealand White rabbits. Antiserum was then affinity purified against the same ferroportin peptide segment fused to dihydrofolate reductase using a preparative immunoblot procedure.¹⁹  Specificity of the rabbit polyclonal anti–mouse ferroportin antibody has already been described.^5,6,15  The monoclonal anti-CD11b (Mac-1) antibody developed by Timothy A. Springer and the monoclonal anti-Lamp1 (lysosome-associated membrane protein-1) antibody developed by J. Thomas August were obtained from the Developmental Studies Hybridoma bank, developed under the auspices of the National Institute of Child and Human Development (NICHD) and maintained by the University of Iowa, Department of Biological Sciences, Iowa City, IA. The mouse monoclonal anti–β-actin was purchased from Sigma (Saint Quentin Fallavier, France). Rabbit polyclonal anti–H-ferritin antibody was a kind gift from P. Santambrogio and S. Levi (San Raffalle Scientific Institute, Milan, Italy).

Cells were washed with cold phosphate-buffered saline (PBS), scraped into PBS-EDTA (ethylene-diaminetetraacetic acid; 2 mM) using a rubber policeman and centrifuged at 500g in a refrigerated microcentrifuge for 5 minutes. Cell pellets were then homogenized in 250 μL lysis buffer (10 mM Tris-HCl, pH 7; 1 mM MgCl₂) supplemented with protease inhibitor cocktail EDTA free (PIs; Roche Diagnostics, Meylan, France) by 20 passages through a 25-gauge needle (5/8-inch). The lysate was centrifuged at 2000g in a microcentrifuge at 4°C for 10 minutes to eliminate nuclei and unbroken cells. Postnuclear (nuclear depleted) supernatant fractions (PNSs) were either stored at -80°C for subsequent Western blot analysis or ultracentrifuged at 150 000g to separate crude membrane fractions from cytosolic proteins. After ultracentrifugation, supernatants corresponding to cytosolic extracts were collected, and membrane pellets were resuspended in TNE buffer (100 mM NaCl; 10 mM Tris-HCl, pH 7.0; 10 mM EDTA) containing 30% glycerol and PIs. All protein extracts (PNS, membrane, or cytosolic fractions) were stored at -80°C until use. Protein concentrations of all samples were determined by the Bradford assay (Bio-Rad, Hemel Hempstead, United Kingdom).

Protein extracts (10 μg) were solubilized in 1X Laemmli buffer and incubated for 30 minutes at room temperature (RT), prior to analysis by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and electrotransfer (2 hours at 110 mA; Novex Western transfer apparatus; Invitrogen, Cergy Pontoise, France) on polyvinylidene fluoride (PVDF) membrane. To control for loading and transfer, the membranes were stained by Red Ponceau after transfer and subsequently preincubated with blocking solution (7% skim milk in TBST [0.15% Tween20, in Tris-buffered saline]) for 16 hours at 4°C, and incubated with primary antibodies (antiferroportin, 1/300; anti–H-ferritin, 1/1000 or anti–β-actin, 1/5000) for 1 hour at RT. After 6 washes (5 minutes per wash) with TBST, blots were incubated with donkey peroxidase-labeled anti–rabbit or sheep peroxidase-labeled anti–mouse immunoglobulin (1/5000; Amersham, Marnes-la-coquette, France) for 1 hour at RT and revealed by ECL chemiluminescence (Amersham).

Cells were fixed with 100% methanol at -20°C for 15 minutes, washed with PBS, and then permeabilized with Triton X100 (0.1% in PBS) for 10 minutes. After 2 PBS washes, cells were incubated in a blocking solution (BSA 1% and 10% heat-inactivated goat serum in PBS) for 45 minutes at RT. Incubation with primary antibodies was then performed in a humid chamber at RT for 1 hour using the following dilution in blocking solution: rabbit antiferroportin, 1/50 to 1/100; rat anti-CD11b, 1/200; and rat anti-Lamp1, 1/350. After 3 washes with PBS/0.5% BSA, cells were incubated for 1 hour at RT with Alexa 488–conjugated goat antirabbit (green) and/or Alexa 588–conjugated goat antirat (red; MolecularProbes, Invitrogen) diluted at 1/200 in blocking solution. Slide coverslips were then washed 3 times in PBS/0.5% BSA, 1 time in PBS, mounted with antifading mounting reagent (Prolong Antifade kit P-7481; Molecular Probes), and processed for immunofluorescence. Cells were visualized using either an epifluorescence microscope LEICA DM-IRB (Leica Imaging Systems, Cambridge, United Kingdom) with a 40 ×/0.6 numeric aperture (NA; Figure 4Bi,iv) or 100 ×/1.4 NA (Figures 2B, 4Bii,v, 5C) oil immersion objective or a Zeiss confocal fluorescent microscope LSM510 MET (Carl Zeiss, Jena, Germany) with a 63 ×/1.4 NA oil immersion objective (Figures 3, 4Biii,vi, 6). Images were acquired using either ARCHIMED-PRO (Microvision Instruments, Evry, France) or Zeiss LSM Image Browser software.

By Western blotting, ferroportin was detected as a protein species around 65 kDa in quiescent BMDMs (Figure 1A). Previous studies have suggested that ferroportin protein levels can be modulated by cellular iron content.^20,21  Consistent with previous observations in J774 macrophage cell line,²²  a dramatic increase of ferroportin protein expression was observed in BMDMs following iron treatment with Fe-NTA for 16 hours (Figure 1A). In contrast, treatment of the cells with the iron chelator DFO lead to a net decrease of ferroportin detection. We also analyzed H-ferritin expression on the same samples. As expected, H-ferritin protein levels increased with Fe-NTA and decreased with DFO treatments, respectively (Figure 1A), confirming the changes in intracellular iron concentrations. As a control, the expression of β-actin was not affected by iron treatments and confirmed the similar loading of the samples.

We next investigated the subcellular localization of the iron exporter in quiescent, iron replete, or iron-depleted cells. A negative control was obtained by omitting the primary antiferroportin antibody in the immunofluorescence procedure (Figure 1Bi). By classic immunofluorescence, in untreated cells, we observed that ferroportin was concentrated in vesicles (punctuate staining) mainly localized intracellularly (Figure 1Bii,v). Of interest, ferroportin vesicles presented a certain degree of heterogeneity in size with small and big (up to ≈ 0.4 μm) entities distributed throughout the cell. Despite no clear evidence of cell surface localization, numerous ferroportin-positive vesicles (especially the larger ones), were observed at the periphery of macrophage (Figure 1Bv, arrows). Consistent with our Western analysis, a clear enhancement of ferroportin fluorescence labeling (compare Figure Biii and vi with Bii and v) was observed in Fe-NTA–treated cells. By contrast, the intensity of ferroportin staining was reduced in DFO-treated cells (compare Figure Biv with Bii and v). Of interest, in iron-overloaded cells that express high amounts of ferroportin protein, the ferroportin staining was strongly concentrated at the cell surface of macrophages.

Because the physiologic role of ferroportin in macrophages essentially consists in the release of iron after erythrophagocytosis (EP),^6,8  we also evaluated the expression and the localization of the iron exporter after red blood cell ingestion by BMDMs. For this purpose we used a cellular model developed in our laboratory using in vitro artificially aged RBCs that are efficiently recognized and engulfed by BMDMs.¹⁷  This model presents the advantage of mimicking as closely as possible the natural process of RBC clearance by macrophages.¹⁷  We and others have shown that ferroportin protein levels increase transiently after EP with maximal induction observed around 8 to 10 hours after EP.^17,22  Therefore, by Western blotting, we analyzed the expression of ferroportin 8 hours after EP. As previously observed,^17,22  both ferroportin and ferritin protein levels were significantly increased after EP (Figure 2A). We were also able to confirm the up-regulation of ferroportin expression by immunofluorescence staining (Figure 2B). In addition, in cells that present elevated levels of ferroportin after EP, a cell-surface staining of the iron exporter was clearly observed (Figure 2B, arrows). These results indicate that a physiologic source of cellular iron such as the one generated by heme catabolism leads to increased ferroportin cell-surface expression in macrophages, similar to the effect induced by Fe-NTA.

The analysis of subcellular localization of ferroportin in bone marrow–derived macrophages by immunofluorescence and confocal microscopy confirms the vesicular distribution of the protein in untreated cells (Figure 3Ai). However, some positive ferroportin vesicles (arrows) are found at the plasma membrane of macrophages, where they colocalize with CD11b/Mac1, a specific phagocyte marker expressed at the cell surface (Figure 3Aii). In addition, the distribution of ferroportin at the plasma membrane was also clearly observed on XZ sections of double-positive cells (Figure 3Aiii-x, arrows). On iron treatment, ferroportin expression increased with a strong staining at plasma membranes that colocalized with CD11b/Mac1 (Figure 3B). Indeed, serial sections clearly illustrated the presence of ferroportin at the plasma membrane, with a strong labeling that tends to delineate almost all the cell surface.

Similar results were observed when using the macrophage cell line J774 (Figure 4). In untreated J774 cells, ferroportin was poorly detected (Figure 4A). However, ferroportin was strongly up-regulated in iron-treated J774 cells and detected as a major protein species around 65 kDa. In this macrophage cell line, immunofluorescence staining confirmed the up-regulation of ferroportin following iron treatment (Figure 4B; compare fluorescence intensity between Biv-vi and Bi-iii) and the plasma membrane localization of the protein, as described for BMDMs.

In epithelial cells expressing green fluorescent protein (GFP)–ferroportin fusion protein, hepcidin was shown to decrease ferroportin expression through direct protein-protein interaction, leading to internalization and degradation of the protein in lysosomes.⁹  We investigated whether hepcidin could have similar effects on endogenous protein expressed in primary macrophages (Figures 5 and 6). First of all, we looked at the effect of synthetic human hepcidin (25 aas) on ferroportin protein levels in BMDMs (Figure 5A). When added in the culture media at the final concentration of 700 nM during 3 hours, hepcidin dramatically reduced the level of ferroportin expression (Figure 5A). A decrease in ferroportin expression was also observed by immunofluorescence in hepcidintreated cells (not shown). When added in parallel with Fe-NTA during 6 hours (Figure 5B), hepcidin abolished the up-regulation of ferroportin protein expression induced by iron overload. Together these results suggest a rapid and particularly efficient mechanism of action of hepcidin on ferroportin macrophage expression.

To explore the possibility that in macrophages hepcidin acts through internalization of cell-surface ferroportin, BMDMs were pretreated for 16 hours with Fe-NTA to increase ferroportin at the cell surface, and hepcidin (700 nM) or sterile water (as a control) was then added to the culture medium for 1 or 3 hours before cell fixation and immunofluorescence procedure. In hepcidin-untreated cells, Fe-NTA treatment leads to a marked expression of the ferroportin at the cell surface of BMDMs as reported previously (Figures 1 and 3). In deep contrast, when these iron-overloaded cells were treated with hepcidin, the ferroportin staining mostly disappeared from the cell surface and was mainly observed in large intracellular vesicles after 1 hour and markedly decreased after 3 hours (Figure 5C). At the subcellular level (Figure 6), coimmunodetection of ferroportin and Lamp1, a late endosomal/lysosomal marker, reveals that hepcidin treatment induced a rapid (at 1 hour, Figure 6; or even at shorter times, not shown) internalization of ferroportin protein in intracellular vesicles that partly colocalized with Lamp1. After 3 hours with hepcidin, the ferroportin staining was decreased and essentially seen in small intracellular vesicles. Together, these observations tend to indicate that hepcidin is able to induce rapid internalization and degradation of ferroportin in macrophages.

In this article, we report that in quiescent macrophages, ferroportin seems to be expressed in vesicular endomembranes that could be seen inside the cell or at the plasma membrane, likely suggesting a vesicular trafficking of the protein between the cytosol and the cell surface. The presence of the iron exporter at the cell surface of macrophages is in agreement with a possible interaction with extracellular hepcidin. Such interaction has been shown in epithelial cells overexpressing a ferroportin-GFP fusion protein.⁹  To our knowledge, this is the first report that clearly describes the presence of the iron exporter ferroportin at the plasma membrane of macrophages.

The nature of the ferroportin-positive vesicles is still not known and needs to be determined. Of interest, Abboud and Haile¹  have suggested the colocalization of ferroportin with hemosiderin, a degraded form of ferritin with high iron storage capacities, found in membrane-bound compartments. Previous studies have shown that iron loading and erythrophagocytosis increase ferroportin expression.^17,22  Indeed on iron treatment with Fe-NTA, we do observe a net up-regulation of macrophage ferroportin at the protein level. In addition, in iron-overloaded cells, ferroportin staining was concentrated at the surface of macrophages. This could be viewed as a cellular response to remove the excess of iron and to limit iron-mediated oxidative stress in the cells. Such protein targeting to the plasma membrane has been previously documented for other ion transport proteins. Indeed, cellular exposure to high copper concentrations can lead to the rapid translocation of Menkes disease protein (MNK) from trans-Golgi network to the plasma membrane.²³  However, it still remains to be determined whether iron treatment directly influences the presence of ferroportin at the plasma membrane (eg, by increasing its stability or by specific targeting of the protein to that site), or whether our observations simply reflect the major change in expression level (more proteins expressed, more proteins seen at that site). Of interest, following physiologic up-regulation of ferroportin observed after erythrophagocytosis, the iron exporter was also enriched at the cell surface of the macrophages. Together our observations favor a cell-surface export of iron during macrophage heme iron recycling. In transfected human kidney epithelial cells, iron export through ferroportin was shown to be dependent on cell-surface expression of the protein in an in vitro functional study of human ferroportin mutations²⁴  and its hemochromatosis-associated mutations. It would be interesting to reproduce such studies in macrophages.

Both transcriptional and posttranscriptional mechanisms have been implicated in the regulation of ferroportin induced by changes in cellular iron status.^20,22  Indeed, in J774 macrophages, ferroportin mRNA levels have been shown to decrease following iron depletion or to increase in response to iron overload, a regulation likely occurring at the level of gene transcription.²²  In addition, the 5′ noncoding region of the ferroportin mRNA contains an iron-responsive element (IRE) as the ferritin mRNAs.^1-3  When cellular iron levels are low, iron regulatory proteins (IRPs) bind to 5′-IREs and inhibit the translation of the target mRNAs.²⁵  On the contrary, when cellular iron levels are high, IRP activity decreases, leading to an increased rate of translation. Previous studies have suggested that iron could posttranscriptionally regulate ferroportin in macrophages.²²  Moreover, during EP, iron released from heme could also participate in ferroportin and ferritin gene regulation through modulation of IRP activities. Indeed iron chelation has been shown to suppress ferroportin induction after erythrophagocytosis, predominantly at the protein level.²²  It is not yet clear to what extent transcriptional and posttranscriptional mechanisms contribute to ferroportin up-regulation in our BMDMs after iron treatment and EP.

Our observations also clearly implicate hepcidin in the regulation of endogenous ferroportin expressed in BMDMs. Indeed, we demonstrated that the addition of hepcidin in macrophage culture medium leads to a profound decrease of ferroportin protein expression, even in the face of an induced ferroportin expression by iron. Similarly, Knutson et al⁸  have shown by Western blot that hepcidin induces a rapid drop in ferroportin protein levels in the macrophage cell line J774. Finally, here we show for the first time that the addition of hepcidin to macrophages expressing a high amount of endogenous ferroportin induces a rapid and dramatic change in the distribution of the iron exporter from the plasma membrane to intracellular vesicles that partially colocalize with Lamp1. All together, these observations confirm that endogenous macrophage ferroportin is a specific target of hepcidin and strongly suggest that, as previously observed for epithelial cells,⁹  hepcidin action passes through internalization and degradation of the protein in macrophage lysosomes. Such a rapid posttranscriptional regulation of ferroportin in macrophages that are the major sources of body iron through the process of erythrophagocytosis and heme iron recycling may likely explain cellular iron accumulation in phagocytic cells and the rapid drop in serum iron concentrations observed in inflammation.^26,27  Hepcidin expression is strongly induced in inflammatory conditions, and pharmacokinetics studies in mice have recently shown that a single injection of hepcidin causes a rapid fall of serum iron, consistent with a blockade of iron export from macrophages.²⁸  It still remains to be determined what are the molecular mechanisms and signaling pathway involved in the rapid hepcidin-induced internalization of ferroportin in macrophages, as well as in epithelial cells.⁹ 

We thank Cécile Pouzet (IFR 07, Bichat, Paris) for technical assistance in confocal microscopy analysis; Philippe Gros (McGill University, Montreal, Canada), Sophie Vaulont, and Lydie Viatte (Institut Cochin, INSERM 567, Paris, France) for helpful discussion and critical reading of the manuscript.