Tissue macrophages play an essential role in iron recycling through the phagocytosis of senescent RBCs (red blood cells). Following haem catabolism by HO1 (haem oxygenase 1), they recycle iron back into the plasma through the iron exporter Fpn (ferroportin). We previously described a cellular model of EP (erythrophagocytosis), based on primary cultures of mouse BMDMs (bone-marrow-derived macrophages) and aged murine RBCs, and showed that EP induces changes in the expression profiles of Fpn and HO1. In the present paper, we demonstrate that haem derived from human or murine RBCs or from an exogenous source of haem led to marked transcriptional activation of the Fpn and HO1 genes. Iron released from haem catabolism subsequently stimulated the Fpn mRNA and protein expression associated with localization of the transporter at the cell surface, which probably promotes the export of iron into the plasma. These findings highlight a dual mechanism of Fpn regulation in BMDMs, characterized by early induction of the gene transcription predominantly mediated by haem, followed by iron-mediated post-transcriptional regulation of the exporter.
Most of the iron in the body is associated with the Hb (haemoglobin) of circulating RBCs (red blood cells). Phagocytosis of senescent RBCs by tissue macrophages [namely EP (erythrophagocytosis)] ensures the efficient recycling of iron and provides the bone marrow with the iron required for erythropoiesis . EP takes place mainly in the spleen, the bone-marrow-derived macrophages and the Küpffer cells of the liver. After being recognized and bound to the macrophage, the RBC is engulfed in a phagosome, and the subsequent maturation of the phagosome leads to degradation of the RBC Hb and the catabolism of haem. This latter process leads to the release of iron, CO and bilirubin, through the action of an enzymatic complex containing an NADPH–cytochrome c reductase, HO1 (haem oxygenase 1) and biliverdin reductase. HO1 is the rate-limiting enzyme of haem catabolism in macrophage cells [2–4]. Following its release from haem, iron may be stored in macrophages associated with ferritin molecules, or exported back into the plasma to meet the body's iron needs . The egress of iron from the macrophages is controlled by Fpn (ferroportin) [6–9], the sole iron exporter to have been identified in mammals.
Fpn is a major player in iron homoeostasis, and the regulation of its expression is complex and has been shown to rely on transcriptional, post-transcriptional and post-translational mechanisms in response to various stimuli in different cell types. At the protein level, two distinct mechanisms have been extensively described in recent years. First of all, iron was shown to play a key role in Fpn protein expression. As with ferritin mRNAs, Fpn mRNAs present an IRE (iron-responsive element) motif in the 5′-UTR (5′-untranslated region) [7,8]. In the absence of iron, the binding of IRP1 (iron-regulatory protein 1) or IRP2 to 5′-IRE inhibits the translation of mRNAs . In contrast, the entry of iron into cell induces a change in IRP1 conformation (by acquisition of an iron–sulfur cluster), oxidation of IRP2 (followed by its degradation in the proteasome), and leads to the translation of 5′-IRE-containing mRNAs . The IRE present in Fpn mRNAs has been shown to be functional in various cell types, including the human monocytic cell line U937  and the mouse macrophage cell line RAW264.7 . Fpn protein expression in macrophages is dramatically increased by iron [13,14]. Iron-mediated overexpression of Fpn is also associated with an increased level of the iron exporter at the cell surface . In addition to this post-transcriptional positive regulation, a negative regulatory mechanism of Fpn protein expression was revealed by the discovery of hepcidin, the major regulator of iron homoeostasis [15–17]. Hepcidin is a small 25-amino-acid peptide produced mainly by the liver. Hepcidin has been shown to bind to Fpn in epithelial cells  and to induce the internalization and subsequent degradation of the exporter in the lysosomal compartment both in these cells [18,19] and in primary mouse macrophages [13,20].
Unlike Fpn regulation at the protein level by intracellular (iron) or systemic (hepcidin) regulators, the mechanisms involved in the transcriptional regulation of the Fpn gene remain to be elucidated. In cultured primary macrophages, activation by lipopolysaccharides and IFNγ (interferon γ) leads to a marked decrease in the Fpn mRNA . Fpn mRNAs also increase in freshly isolated human alveolar macrophages exposed for 24 h to ferric ammonium citrate , as well as in murine J774 macrophages loaded for 20 h with Fe-NTA (iron-nitrilotriacetate) . We  and others  have shown that the exporter mRNAs are increased during the first few hours of EP in various sources of macrophages.
The purpose of the present study was to investigate the molecular mechanisms involved in Fpn gene regulation during the first few hours of EP in BMDMs (bone-marrow-derived macrophages).
BMDMs were cultured as described previously . Briefly, bone marrow cells were isolated from femurs of 6–8-week-old mice (DBA/2 strain), and seeded on to 6- or 10-cm diameter Petri dishes for RNA or protein extraction respectively or on to glass coverslips in 24-well tissue culture plates for immunofluorescence studies. The culture medium was RPMI 1640/GlutaMAX™ (Invitrogen) supplemented with 10% (v/v) heat-inactivated fetal calf serum (Invitrogen; low endotoxin content), 10% L929-cell-conditioned medium [source of CSF-1 (colony-stimulating factor 1)], 2 mM L-glutamine, 50 units/ml penicillin and 50 mg/ml streptomycin. At 4 days after seeding, the adherent cells were rinsed twice in HBSS (Hanks balanced salt solution), and the medium was changed daily until day 7.
SnPPIX [tin-PPIX (protoporphyrin IX)] was obtained from Porphyrin Products (Frontier Scientific), and haemin and PPIX were from Sigma–Aldrich. Freshly prepared 1 mM stock solutions of porphyrins with 10 mM arginine were made by dilution in 0.2 M KOH/100% (v/v) ethanol (1:1, v/v). NS (Normosang®; human haemin; Orphan Europe) was obtained from the Centre Français des Porphyries, Hôpital Louis Mourier (Colombes, France).
Mouse and human blood samples (obtained by retro-orbital puncture and intravenous sampling respectively) were collected in heparin-precoated tubes and washed with PBS. The RBCs were then artificially aged by incubating with 2.5 mM calcium and 0.125–0.5 μM Ca2+ ionophore A23187 (Calbiochem) as previously described . BMDMs were incubated with aged RBCs [(1–3)×107 cells/ml; 1 ml/well in 24-well plates, 5 ml in 60 mm or 10 ml in 100 mm Petri dishes] for 1 h at 37 °C in a 5% CO2 incubator and then washed twice with HBSS. Cells were then incubated for 5 min in hypo-osmotic solution (140 mM NH4Cl and 17 mM Tris/HCl, pH 7.6) to lyse any non-ingested RBCs, and fresh medium was added after specified periods of time.
Macrophages were incubated with haem- or PPIX-arginate at a final concentration of 10 μM for 2 h. The medium was then discarded, the BMDMs were washed twice with PBS, and fresh medium was added for 2 or 6 h before RNA extraction or immunofluorescence study respectively. For the treatment with NS, the cells were incubated with this reagent for the periods of time indicated (4 or 8 h). To block HO1 activity, BMDMs were pretreated with 50 μM SnPPIX for 2 h, before adding 10 μM NS for 2 or 6 h before RNA or protein extraction respectively. We have previously shown that H-ferritin protein levels increased with Fe-NTA, reflecting increased intracellular iron concentrations . Therefore, to increase the cellular iron concentration, cells were incubated with Fe-NTA (100 μM FeCl3 and 400 μM NTA), for the indicated periods of time (4 or 8 h). To chelate iron from the cells, BMDMs were incubated with SIH (salicylaldehyde isonicotinoyl hydrazone) for the indicated periods of time at 37 °C. SIH was prepared as a 100 mM stock solution in DMSO (Sigma–Aldrich) and used at a final concentration of 10–50 μM. DMSO or 0.2 M KOH/100% ethanol was used as controls for the studies with SIH or porphyrins respectively. For the studies using NS, a solution corresponding to the excipient was used as a control. To block transcription, cells were treated with 1 μg/ml actinomycin D (Sigma–Aldrich), in the presence or absence of either RBCs or NS. Cell viability was assessed using the Guava Count technology according to the manufacturer's instructions and was shown to be >95% for all the treatments used.
Determination of intracellular haem content
After incubating the macrophages with RBCs or NS for 50 or 90 min respectively, the intracellular haem content was determined by the method of Motterlini et al. . Briefly, cells were washed with PBS and centrifuged at 200 g for 5 min, and the pellet was then solubilized by adding 500 μl of concentrated formic acid. The haem concentration of the formic acid solution was determined spectrophotometrically at 400 nm.
Real-time quantitative PCR analysis
Total RNA was extracted from cells using the RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. Single-stranded cDNA was synthesized using SuperScript™ RNase H− reverse transcriptase (Invitrogen). Real-time quantification of transcripts was performed in 20 μl in a CHROMO IV detector (MJ Research) by using SYBR Green PCR master mix (Sigma SYBR Green JumpStart™ Taq ReadyMix™ for quantitative PCR), 10 pmol of forward and reverse primers, and 5 μl of reverse transcriptase reaction mixture. Sequences of the primers for Fpn, HO1 and S14 ribosomal protein were as previously described . The analyses were performed using the Opticon Monitor software version 2.03 (MJ Research). Amplification and dissociation curves of increasing amounts of BMDM cDNA were used to validate our assay: the dissociation curves displayed a single peak, ruling out the possibility of the presence of primer dimers or parasitic (unspecific) products. The results for each gene were normalized using the Ct (threshold cycle value) of S14 ribosomal protein. The relative quantification was calculated using a comparative Ct method with the following arithmetic formula: (1+EX)−ΔCtX/(1+ES14)−ΔCtS14, were EX is the efficiency of amplification of the gene of interest, ES14 is the efficiency of amplification of the S14 target and ΔCt is the mean Ct stimulated condition–mean Ct control condition. The S.D. of the difference was calculated from the S.D. values of each gene (s1) and S14 (s2) value using the following formula: s =Figures 1–6, the results are the means±S.D. of data obtained for three to six independent experiments.
Up-regulation of Fpn and HO1 mRNAs after phagocytosis of hRBCs
Haem is rapidly degraded following phagocytosis of RBCs by macrophages
Effect of haem on the expression of Fpn and HO1 mRNAs
Contribution of iron to the induction by haem of Fpn and HO1 mRNAs and proteins
Haem stimulates the transcription of the Fpn and HO1 genes
Iron release from haem by HO1 activity is necessary for the induction of Fpn protein
Statistical significance was evaluated using the non-parametric Kruskal–Wallis test (to compare two, three or more unpaired groups) and the Mann–Whitney test (to compare the medians of two unpaired groups). GraphPad Prism software was used for statistical evaluation.
The specificity of the rabbit polyclonal anti-mouse Fpn antibody has been described previously [24,25]. Rabbit polyclonal anti-mouse HO1 was purchased from StressGen Biotechnologies. Rabbit polyclonal anti-mouse H-ferritin antibody was a gift from P. Santambrogio and S. Levi (DIBIT San Raffaele Scientific Institute, Milan, Italy). The mouse monoclonal anti-β-actin was purchased from Sigma–Aldrich.
Protein extracts from macrophage cell cultures
Protein extracts from BMDMs were prepared as previously described . Briefly, cells were washed with cold PBS, scraped into PBS–EDTA (2 mM) and centrifuged at 200 g for 5 min. The cell pellets were then homogenized in 150 μl of lysis buffer (10 mM Tris/HCl, pH 7, and 1 mM MgCl2) supplemented with EDTA-free protease inhibitor cocktail (Roche Diagnostics). The lysate was centrifuged at 900 g for 10 min, to eliminate nuclei and unbroken cells. The protein extracts were then ultracentrifuged at 75000 rev./min (Beckman TLA100 rotor) for 45 min to separate the crude membrane fractions from the cytosolic proteins. Supernatants corresponding to cytosolic extracts were then collected and membrane pellets were resuspended in TNE buffer (100 mM NaCl, 10 mM Tris/HCl, pH 7.0, and 10 mM EDTA) containing 30% (v/v) glycerol and protease inhibitors. All protein extracts (from postnuclear supernatant, membrane or cytosolic fractions) were stored at −80 °C until use. The protein concentrations of all the samples were determined by the Bradford assay (Bio-Rad Laboratories).
Western blot analysis
Crude membrane proteins (10 μg) solubilized in 1× Laemmli buffer were incubated for 30 min at room temperature (20 °C) for HO1 and Fpn detections. Cytosolic proteins (10 μg) were boiled for 5 min at 90 °C for H-ferritin detections. Samples were then analysed by SDS/PAGE and electrotransferred (2 h at 110 mA; Novex Western transfer apparatus; Invitrogen) on to a PVDF membrane. To control for loading and transfer, the membranes were stained with Ponceau Red after transfer, and subsequently pre-incubated with blocking solution [either 7% (w/v) non-fat dried skimmed milk powder in TBST (0.15% Tween 20, in Tris-buffered saline), for Fpn, HO1 and H-ferritin, or 1% BSA in TBST, for β-actin] for 1 h at room temperature. The membranes were then incubated with primary antibodies as follows: anti-Fpn, 1:300 (16 h at 4 °C); anti-HO1, 1:10000, anti-H-ferritin, 1:1000 (1 h at room temperature); and anti-β-actin, 1:5000 (30 min at room temperature). After washing with TBST, the blots were incubated with either donkey peroxidase-labelled anti-rabbit immunoglobulin (1:3000) (Nordic Immunologic) for 1 h at room temperature or sheep peroxidase-labelled anti-mouse immunoglobulin (1:5000) (GE Healthcare) for 30 min at room temperature and revealed by ECL® (GE Healthcare).
The cells were fixed with 100% (v/v) methanol at −20 °C for 15 min, washed with PBS and then permeabilized with Triton X-100 (0.1% in PBS) for 10 min. After two PBS washes, the cells were incubated in a blocking solution [1% BSA and 10% (v/v) heat-inactivated goat serum in PBS] for 45 min at room temperature. They were then incubated with primary rabbit anti-Fpn antibodies in a humid chamber at room temperature for 1 h by using a 1:50 dilution in blocking solution. After three washes with PBS/0.5% BSA, cells were incubated for 1 h at room temperature with Alexa Fluor® 488-conjugated goat anti-rabbit antibody (Molecular Probes) diluted at 1:200 in blocking solution. Coverslips were then washed three times with PBS/0.5% BSA, washed once with PBS, mounted with an antifading mounting reagent (Prolong Antifade kit P-7481; Molecular Probes) and processed for immunofluorescence. Cells were visualized using an epifluorescence microscope, Leica DM-IRM, with a ×100 oil immersion objective. Images were acquired using ARCHIMED-PRO (Microvision Instruments).
Fpn and HO1 mRNA induction after EP of hRBCs by BMDMs
Using murine RBCs, we have previously shown that Fpn and HO1 mRNAs are maximally induced after 4 h of EP in BMDMs . Here, we performed similar experiments using hRBCs (human RBCs) treated with the calcium ionophore A23187 and calcium in order to mimic eryptosis, the process of RBC aging . After EP, the expression of Fpn and HO1 was followed by using quantitative RT–PCR (reverse transcription–PCR) (Figure 1A). Using different concentrations of artificially aged hRBCs, we reproduced our previous observations  and then extended them by showing that both HO1 and Fpn mRNA expressions increase in parallel with the number of hRBCs used in our EP assay. To find out whether these mRNA changes depended on iron release during EP, we performed the EP assay and subsequently kinetic studies in the presence of the rapidly permeant iron chelator SIH (Figure 1B). SIH had no effect on HO1 mRNA, and produced modest but not significant inhibition of Fpn induction.
Early increase in Fpn and HO1 mRNA levels in the presence of haem in BMDMs
Haem is known to enhance HO1-gene transcription in various cell types, including macrophages . We therefore wondered whether haem itself, released following the degradation of RBCs, could be directly involved in regulating the expression of HO1 and Fpn during EP. To test this hypothesis, we first evaluated the rate of haem catabolism following EP in BMDMs (Figure 2). For this purpose, BMDMs were incubated with the same number of artificially aged murine (Figure 2A) or human (Figure 2B) RBCs for 50 min to induce EP. The intracellular haem content over time after EP displayed a similar kinetic profile for both types of RBCs, with an increase during the first 30 min following EP and a progressive decrease thereafter (Figure 2).
We then studied the effect of different sources of haem (NS; haem-arginate) on the level of Fpn and HO1 mRNAs (Figure 3). NS treatment induced a rapid and marked accumulation of intracellular haem (increased more than 30-fold in response to the 50 μM concentration) (Figure 3A). Interestingly, Fpn and HO1 mRNAs were also markedly increased by NS in a dose-dependent manner (Figure 3A; Supplementary Data 1 at http://www.BiochemJ.org/bj/411/bj4110123add.htm). Furthermore, incubating BMDMs with PPIX-arginate, an iron-free haem precursor, did not significantly increase the level of Fpn and HO1 mRNAs when compared with incubating with the same dose of haem-arginate (Figure 3B). This observation clearly indicates the importance of the iron atom present in the protoporphyrin ring in Fpn and HO1 mRNA-mediated induction.
Effect of iron chelation in the induction of Fpn and HO1 mRNAs and proteins by haem in BMDMs
To discriminate between the respective roles of haem and of iron released from NS, we carried out NS treatment in the presence of SIH (Figure 4A). Results showed that induction of HO1 mRNAs after 4 h of NS was not significantly modified by iron chelation. On the other hand, the induction of Fpn by NS was modestly but significantly blocked by SIH (Figure 4A).
Fpn, HO1 and H-ferritin proteins are indirectly induced by haem in BMDMs
We then tested whether induction of Fpn mRNA by a source of haem is associated with increased protein synthesis. BMDMs were treated with NS, haem-arginate or PPIX-arginate and then processed for immunofluorescence (Figure 4B). As a positive control, macrophages were treated with Fe-NTA, which was previously shown to increase Fpn expression . We observed that, like Fe-NTA, both NS and haem-arginate dramatically increased the protein expression of the transporter relative to untreated cells. The iron in haem was strictly necessary for this induction, since PPIX-arginate did not induce any change in Fpn expression. As observed with Fe-NTA, Fpn induced by haem strongly localized to the plasma membrane of macrophages (Figure 4B). We next evaluated the effect of the iron chelator SIH on the induction of Fpn and HO1 proteins by Western blotting after exposure of NS for 8 h (Figure 4C). As a control of intracellular iron changes, H-ferritin expression was also studied. NS strongly induced the expression of Fpn, HO1 and H-ferritin proteins. Interestingly, the induction of both Fpn and H-ferritin by NS (Figure 4C) was markedly inhibited in the presence of SIH. In contrast, induction of the HO1 protein by NS was not affected by iron chelation. Similar observations were made after 8 h of EP (Supplementary Data 2 at http://www.BiochemJ.org/bj/411/bj4110123add.htm).
Transcriptional regulation of Fpn by haem
The effect of haem on Fpn and HO1 mRNA levels could reflect induction of transcription of these genes or stabilization of pre-existing mRNAs. To clarify this point, BMDMs stimulated with NS for 4 h were incubated with the transcriptional inhibitor actinomycin D for the same period of time (Figure 5A). Interestingly, the induction of Fpn and HO1 mRNAs by NS was inhibited in the presence of actinomycin D, supporting the hypothesis that the positive effect of haem on these two genes in BMDMs involves transcriptional activation. Accordingly, by blocking transcription, actinomycin D also prevented the NS-induced synthesis of Fpn and HO1 proteins (Figure 5B). On the other hand, actinomycin D had only a limited effect on the induction of H-ferritin by NS, confirming a strong effect of a post-transcriptional mechanism (iron-dependent) in the regulation of this gene.
HO1-dependent release of iron from haem causes induction of the Fpn protein in BMDMs
Our previous findings (Figure 4) suggest that iron released from haem after EP or NS treatment is responsible for inducing Fpn and H-ferritin proteins. HO1 is the enzyme responsible for catabolizing haem and releasing iron in phagocytic cells . We therefore investigated whether this enzyme was necessary for the Fpn and H-ferritin proteins to be induced by NS in BMDMs (Figure 6). To do this, cells were pre-incubated for 2 h with an inhibitor of HO1 activity, namely SnPPIX, before being exposed to NS for 2 h (Figure 6A) or 6 h (Figure 6B). As illustrated, SnPPIX did not significantly modify the high level of Fpn or HO1 mRNAs induced by NS (Figure 6A), indicating that inhibiting haem degradation does not affect the induction of either of the genes by NS (Figure 6A). In addition, the Fpn mRNAs are not sensitive to SIH when haem degradation is inhibited by SnPPIX, supporting the hypothesis that transcriptional regulation of Fpn is predominantly regulated by haem, and only slightly by iron. At the protein level, SnPPIX only partially prevented the accumulation of HO1 protein mediated by NS, whereas it completely prevented that of Fpn. H-ferritin protein expression (Figure 6B) induced by NS was also blocked by the induction of SnPPIX, indicating that no iron was released from haem.
In the present study, we used our cellular model to identify the respective roles of haem and iron in regulating the expression of Fpn during EP . We demonstrate for the first time that in primary cultures of BMDMs, haem derived from EP or from an exogenous source (such as NS) can stimulate the transcription of the Fpn gene. Our kinetic studies of intracellular haem accumulation during EP indicate that haem is released from damaged RBCs engulfed by BMDMs. The haem content then progressively decreases, probably reflecting its catabolism by HO1. Haem accumulation in BMDMs appears to be greater following the phagocytosis of hRBCs than that of murine ones, which probably reflects interspecies differences in terms of RBC size and Hb content. However, the general profile of haem cataboltism is similar in both species. It is noteworthy that changes in the intracellular haem content following EP in BMDMs parallel the previously observed changes in HO1 and Fpn gene expression profiles , suggesting that haem itself could have a modulatory effect. We then describe several findings that demonstrate the early transcriptional up-regulation of the Fpn gene by haem in BMDMs. First, haem-arginate or NS used as an exogenous source of haem induced Fpn mRNA accumulation. Secondly, pretreating the cells with actinomycin D (an inhibitor of RNA polymerase II) blocked the haem-mediated induction of Fpn and HO1 mRNAs. Finally, inhibiting HO1 activity by SnPPIX and/or intracellular chelation of iron partly prevented the accumulation of Fpn mRNA. Our observations after EP or haem treatment suggest that most of the Fpn mRNAs levels are stimulated by haem and the remainder by the release of iron from haem. This transcriptional effect mediated by haem is consistent with the recent demonstration of Hb-induced Fpn mRNA in haptoglobin-null spleen macrophages and in vitro in RAW264.7 macrophages .
Previous studies have shown that iron can increase the transcription of Fpn in macrophages [14,22]. We also observed that Fe-NTA triggered a slight increase in Fpn mRNA, whereas SIH reduced basal Fpn mRNA expression (results not shown). However, and this is consistent with our findings, iron chelation was shown to produce only partial inhibition of the increase in Fpn mRNAs in a different model of EP using J774 macrophages . This suggests that haem-dependent transcriptional induction of this gene predominates after RBC phagocytosis and that the iron released from haem is mildly involved in this regulation.
Haem is known to be a direct transcriptional inducer of the HO1 gene. Indeed, this gene has been shown to be transcriptionally regulated by haem in various types of cells, including hepatoma cells , rat glioma cells  and cultured alveolar macrophages . Regulation of the HO1 gene involves the transportation of haem into the nucleus by a carrier and its binding to the physiological repressor Bach1, which allows transcriptional activators to initiate transcription [31,32]. Bach1 forms a heterodimer with MafK and binds to the multiple MAREs (Maf recognition elements) of HO1 enhancers. Thereby Bach1 represses their activity in vivo, whereas haem interferes with this function of Bach1 by inhibiting its binding to the HO1 enhancers . It is tempting to speculate that similar regulatory mechanisms could be involved in the transcriptional regulation of Fpn by haem. Even though sequence analysis of the Fpn gene promoter did not reveal the presence of MAREs, possible direct transcriptional regulation of Fpn by haem cannot be ruled out. We have previously shown differences in the patterns of expression of the HO1 and Fpn genes in BMDMs during EP . Indeed, we observed that 2 h after EP, HO1 mRNA was already strongly induced, whereas Fpn mRNA level was not changed. Later, after EP, Fpn mRNA was strongly repressed, whereas expression of HO1 was maintained at a quite high level. These observations suggest that these two genes could be regulated by different mechanisms, both of which are mediated by haem. Various haem-regulated transcription factors may exist in mammalian cells , and further analysis is needed to identify the functional cis and trans regulatory elements involved in haem-mediated activation of Fpn transcription.
It is striking that our results indicate that an iron atom must be present in the protoporphyrin ring in order to observe the induction of Fpn and HO1 mRNAs. One can speculate that conformation changes induced by the presence of iron allow haem to interact with haem transporters or protein partners in order to mediate its transcriptional activity. We also noted that SnPPIX affects Fpn and HO1 mRNAs in BMDMs (Figure 6). SnPPIX acts as a competitive inhibitor of HO1, even though it also increases its expression . This effect is probably linked to the similarities in chemistry and structure of this haem analogue, which could bind to the same regulatory sites as haem itself , and one can speculate that SnPPIX could induce HO1 and Fpn mRNAs in a similar manner. This induction could also be the consequence of the potential increase in the cellular haem content, resulting from the inhibition of HO1 activity. Further investigation is needed to clarify this point.
Although the haem contained in RBCs or haem/NS appears to be the main transcriptional inducer of the Fpn gene in BMDMs, we demonstrate that the synthesis of the transporter at the protein level depends mainly on iron. In contrast, up-regulation of HO1 protein by haem was shown to be iron-independent. Our results illustrate the absolute requirement for the catabolism of haem by HO1, and the subsequent release of iron, to increase the level of the Fpn protein in BMDMs. The observation that an iron chelator (SIH) or an inhibitor of HO1 activity (SnPPIX) suppresses EP or haem induction of Fpn at the protein level shows that iron extracted from haem by the catabolic activity of HO1 is indeed responsible for up-regulating Fpn proteins. Like the mRNA that codes for the iron storage protein H-ferritin, Fpn mRNA contains an IRE in its 5′-UTR. Translation of proteins from mRNA-containing IREs in their 5′-UTRs is known to be controlled by a mechanism involving the iron sensors IRPs . Several observations suggest that there is a functional IRE in the 5′-UTR of Fpn mRNA. First of all, IRPs can bind to Fpn IRE . More importantly, when placed in front of a luciferase reporter gene, the Fpn IRE is able to control gene expression via an iron-dependent regulatory effect [11,12]. It is therefore likely that Fpn IRE confers translational control of the Fpn gene during EP. However, one cannot exclude the possibility that a post-transcriptional mechanism could exist that also affects mRNA stability.
In addition, even though most of the effects of haem stimulation on Fpn protein are mediated by iron, the increase in Fpn mRNA levels mediated by haem itself may also be critical in this induction. Indeed, by blocking transcription, the up-regulation of the Fpn protein by haem is completely abolished. It is therefore tempting to speculate that Fpn mRNA levels are low or unstable in the cells. A combination of these two regulatory pathways (the haem-mediated transcriptional induction followed by the iron-mediated post-transcriptional stimulation) is probably responsible for the marked increase in Fpn proteins following haem treatment in BMDMs. We also show that, similarly to iron treatment and EP , haem increases the Fpn distribution to the cell surface of BMDMs. Such a localization of Fpn after NS treatment is consistent with a role of the exporter in the efflux of haem-derived iron.
In contrast with Fpn, the strong increase in ferritin protein after NS treatment is mildly decreased by the addition of actinomycin D. This indicates that H-ferritin induction by haemin is strongly related to post-transcriptional mechanisms, as previously observed [35,36]. In addition, our observation also suggests the existence of a transcriptional regulation of ferritin after haem treatment. Indeed, recently, haem was shown to regulate the transcription of H-ferritin through the haem reversible repressor Bach1 .
In conclusion, we show that haem is a novel effector of Fpn gene regulation in macrophages. Such a regulation reveals the existence of a balance between haem and iron in activating the expression of the Fpn gene in BMDMs during the early stages of EP (Figure 7). We propose that during the process of EP or after haem treatment, haem may stimulate both HO1 and Fpn mRNAs (step 1). The increased level of the HO1 protein then leads to haem catabolism and iron release (step 2), which in turn stimulates the synthesis of Fpn at the transcription and particularly at a post-transcription level (step 3). Iron export driven by Fpn could then contribute to restore the baseline situation (step 3). This regulation is of particular interest with regard to disorders such as haemolytic anaemia. Both extra- and intra-vascular haemolyses lead to excessive haem absorption and iron accumulation in macrophages. During extravascular haemolysis, these observations are the consequence of an accelerated EP process in tissue macrophages. In the context of intravascular haemolysis, lysis of RBCs in the circulation leads to an excess of Hb and haem, which are rapidly bound to haptoglobin and haemopexin respectively. Such complexes are then internalized by macrophages via receptor-mediated endocytosis [38–40]. Further characterizations of the dual haem/iron event involved in regulating the Fpn gene during EP will help to elucidate the molecular mechanisms involved in haemolytic processes.
Sequential regulation of Fpn expression in BMDMs after EP: early mRNA induction by haem, followed by iron-dependent protein expression
We are grateful to Dr Sophie Vaulont (Institut Cochin, Inserm 567, Paris, France) and to Dr Cécile Bouton (ICSN, CNRS, UPR 2301, Gif-sur-Yvette, France) for helpful discussion and a critical reading of this paper. We gratefully acknowledge Professor Jean Charles Deybach (Centre Français des Porphyries, Hôpital Louis Mourier, Colombes, France) for the gift of NS. We also thank Dr Prem Ponka (Lady Davis Institute of Medical Research, Montreal, QC, Canada) for the gift of SIH, and P. Santambrogio and S. Levi for providing the anti-ferritin antibody. This study was financially supported by the Inserm.
Hanks balanced salt solution
haem oxygenase 1
red blood cell
Maf recognition element
salicylaldehyde isonicotinoyl hydrazone