High level expression of Nramp1^G169 in RAW264.7 cell transfectants: analysis of intracellular iron transport

Abstract

Nramp1 (natural resistance-associated macrophage protein) was positionally cloned as the defective biallelic locus in inbred mouse strains associated with uncontrolled proliferation of obligate intracellular macrophage pathogens. The causative defect was described as G169D within membrane spanning domain 4 of a transporter. The biochemical activity of Nramp1 is implied from sequence conservation with Nramp2. Nramp2 encodes a divalent cation transporter and is the carrier of a defect in models of microcytic anaemia, associated with impaired intestinal iron uptake. Iron sequestration has been proposed as an antimicrobial mechanism. Therefore, such an activity for Nramp1 is consistent with model systems. Here we showed that Nramp1 directs iron transport within the macrophage. We describe stable, high-level Nramp1^G169 allele-derived polypeptide expression in Balb/c Nramp1^D169 RAW264.7 cells. Transfectants express levels, comparable to those in Nramp1^G169-resistant macrophages, of a 90–100×10³ MW Nramp1 polypeptide. Expression of the Nramp1 polypeptide correlates with lower cellular iron loads and a reduced chelatable iron pool following challenge with iron: nitrilotriacetate. Pulse chase experiments support an enhanced iron flux in expressing cells. These data are supported using the fluorescent iron probe calcein. In Nramp1^G169-expressing cells we observed an increased iron flux into the cytoplasm from a calcein-inaccessible cellular location. These data suggest Nramp1, in resting macrophage cells, mobilizes iron, from an intracellular vesicle, which is destined for cell secretion. We propose that under these conditions Nramp1 plays a role in a salvage pathway of iron recycling.

INTRODUCTION

Natural resistance to infection by a number of unrelated, but obligate intracellular macrophage pathogens is controlled by Ity/Lsh/Bcg on mouse chromosome 1, reviewed in refs 1–3. Two major alleles are distinguished, designated as resistant or susceptible. Resistance is inherited in a dominant manner. Recently, a candidate gene was identified,⁴ full-length sequence subsequently characterized,⁵ and termed Nramp1 (natural resistance-associated macrophage protein). A non-conserved amino acid substitution, Gly-169-Asp, correlated with resistance or susceptibility to infection in all mouse strains analysed.⁶ Following a series of definitive experiments, using mouse gene knockout⁷ and transgenic⁸ techniques, Nramp1 was confirmed as Ity/Lsh/Bcg.

Nramp1 encodes a 548 amino acid polytopic integral membrane protein^4,5 with 10 or 12 putative membrane-spanning helical domains. Nramp1 encodes a transporter molecule, of unknown function, based on the identification of a conserved transport signature within one of the intracellular loop domains.⁴ The amino acid sequence contains motifs for N-linked glycosylation, three N-terminally and one C-terminally located protein kinase C phosphorylation consensus sites. The amino-terminus is Pro and Ser rich and has been suggested as encoding a protein–protein interaction domain.⁵ Described within the putative cytosolic N-and C-terminal domains are consensus endosomal localization sequences.⁹ Other data provided support that Nramp1 localiszes, not at the plasma membrane, but within an intracellular membrane,^2,10 co-localizing with pathogens in infected cells. The intimate localization of the Nramp1 protein and intracellular pathogens, whose growth rates are modulated, suggests that transport activity either introduces some toxin into the microenvironment of the phagolysosomal vesicle or eliminates some essential nutrient for pathogen growth.

Advancement in our understanding of Nramp1 function in the macrophage, where in mouse it is exclusively expressed, has come from work on sequence-related Nramp genes, termed Nramp2, and originally identified by low stringency hybridization.¹¹Nramp2 was re-isolated by expression cloning from a rat duodenal cDNA library.¹² The isolated gene, DCT1, supported iron uptake and DCT1 was considered the rat Nramp2-related gene. Secondly, positional-candidate gene cloning revealed Nramp2 as the defective gene in the mk¹³ and Belgrade¹⁴ mouse and rat models, respectively, of microcytic anaemia. Nramp2, but not Nramp1, complemented the yeast divalent cation transporter SMF.¹⁵ Other work has provided data regarding the direction of iron transport supporting a role for iron influx into the phagolysosomal vesicle.¹⁶ In this study we describe a high level, in vitro expression system to study Nramp1 function and report on the transport properties of Nramp1 using a number of iron transport assays. These data support a model for transport out of the lumen of the vesicle into the cytoplasmic compartment of the resting macrophage.

MATERIALS AND METHODS

RAW264.7 cell culture and cDNA transfection

Balb/c-derived, Nramp1^D169 RAW264.7 peritoneal macrophages were maintained and transfected as previously described.¹⁷ Plasmid DNA for transfection was prepared using the Qiagen Endo-free maxi-plasmid DNA extraction kit (Qiagen Ltd, Crawley, UK). For transfection, 10 μg of DNA was mixed with 0·5 ml (5×10⁶) cells and incubated at room temperature for 5 min in a 0·4-cm electroporation cuvette (Bio-Rad Laboratories, Hemel Hempstead, UK). Cells were pulsed with 300 V (750 V/cm) at 975 μF, immediately placed in culture and incubated overnight. The following day G418 was added to 1 mg/ml and G418 was replaced 2–3 times for a 2-week period. Clonal lines were prepared from the drug-resistant culture by limiting dilution into microwell plates and colonies selected at 0·5 mg/ml G418. To evaluate clones expressing the Nramp1 transgene, Western blotting was performed on several drug-resistant lines.

Nramp1 plasmid DNA constructions

The full-length Nramp1 insert from pBABE λ8·1⁵ was removed by Bam HI digestion. The 5′ BamHI site is located within the polylinker sequence and the 3′ site within the 3′UTR at position 2198–2203 bp of GENBANK Acc. No X75355 upstream of the start of the polyadenylated tail at 2266 bp. The 2203 bp fragment was ligated into the BamHI site of pHβA-1-neo¹⁸ and appropriate orientation clones selected with restriction endonuclease digestion using Hin dIII. Sense orientation products generate a diagnostic fragment of approximately 649 bp from cleavage at position 649 bp of the λ8·1 sequence and upstream of the Nramp1 cDNA insert.

Western blotting using anti-Nramp1 immunoreactive serum

Anti-Nramp1 antibody was prepared against recombinant fusion protein using the pGEX expression vector system corresponding to amino acids 1–82 of the Nramp1 sequence as described before.¹⁹ Routinely, 20 μg of total cell protein was loaded on a single track for sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE). Immunoblotting was performed by standard procedures onto Immobilon membrane (Millipore UK, Watford, UK). Detection was by horseradish peroxidase (HRP) conjugated goat anti-rabbit secondary antibody. Blots were visualized by chemiluminescence (ECL, Amersham Life Science Ltd, Little Chalfont, UK).

Monitoring the iron status of RAW264.7 cells in situ

Cell-associated total iron levels were measured using the ferrozine assay as described previously.¹⁹

⁵⁹Fe studies in macrophages

Macrophages were harvested, counted and plated onto 6-well Nunc tissue culture plates (Life Technologies, Paisley, UK), 1–2×10⁶ per well, and incubated overnight to adhere. Next day ⁵⁹Fe (2–5μCi per well) was added as a chelate with nitrilotriacetate (NTA). Unlabelled iron stimulated the uptake of the radio-labelled iron²⁰ and was added to some experiments. Labelling was for 3–4 hr after which the cells were analysed directly or washed and incubated in full media containing unlabelled iron for the period of the chase. Cell extracts were prepared to monitor chelatable iron and total iron.²¹ Briefly, cells were washed in situ in PBS, harvested by scraping, and pelleted by brief centrifugation. Cells were lysed in 0·6 ml of Hanks buffered saline solution containing Triton-X100 (0·25%) and incubated at room temperature for 5 mins. Insoluble material was pelleted by brief centrifugation and 2,2 Bipyridyl (BIP) to 400 μm and ascorbate to 2 mm was added from stock solutions to 0·5 ml of the supernatant. The mixture was incubated for 2 mins to allow chelation of the redox-active iron pool. 0·5 ml of benzoyl alcohol was added to allow partitioning of the BIP radio-iron chelate into the organic phase. Phases were separated by brief centrifugation. The radioactive component of the organic and aqueous phase was quantitated by liquid scintillation counting of 100 μl from each phase.

Calcein iron imaging

RAW264.7 cells and transfectants were loaded with the iron-sensitive fluorescent dye, calcein, as described²¹ with slight modifications. Briefly, cells were scraped from plates and washed twice in warm (37°) HEPES-buffered Dulbecco's modified Eagle's minimal essential medium (HEPES/DMEM, pH 7·3). Cells were resuspended at 3×10⁶ cells/ml in HEPES/DMEM with calcein 0·125 μm (acetoxymethyl ester, Molecular Probes Europe B.V., Leiden, the Netherlands). Post-loading (10min, 37°), cells were washed twice with HEPES/DMEM and stored on ice in HEPES/DMEM supplemented with 1% (w/v) bovine serum albumin (BSA) at 5×10⁶ cells/ml. Calcein fluorescence (exitation wavelength 488 nm, emission wavelength 512 nm) was determined in a fluorescence spectrophotometer (Hitachi F-2000; Hitachi Scientific Instruments, Finchhampstead, UK) at 37° in a 3-ml fluorescence cuvette with stirring. For each experiment a 1-ml aliquot of suspended cells (5×10⁶) was pelleted and resuspended in 2 ml warm (37°) HBS (Hepes-buffered saline solution; 20 mm HEPES-Tris, 150 mm NaCl, pH 7·3). Cells were equilibrated to establish a stable baseline of fluorescence prior to iron challenge. Iron challenge was achieved with the addition of concentrations of Fe³⁺:NTA (1:1) iron chelates and monitored by the decrease in calcein fluorescence (quench).²² Quench was iron mediated as specific ferric/ferrous iron chelators diethylenetriaminepentaacetic acid (DTPA, Sigma–Aldrich Co. Ltd, Poole, UK) and salicylaldehyde isonicotinyl hydrazone (SIH, kindly synthesized by R. Broadbridge of this department), re-established full calcein fluorescence (not shown). Where appropriate, cells were solubilized with 1·5% (w/v) octyl glucoside (Molecular Probes)/HBS.

RESULTS

Nramp1 expression in RAW264.7 cells

Several molecular sizes for products of the Nramp1 gene have been proposed including proteins of 90–100 MW×10³, 45 MW×10³ and 65 MW×10³.^9,19,23 Cells derived from susceptible mouse strains were reported to lack expression of an extensively glycosylated 90–100 MW×10³ polypeptide.²³ We report here on experiments using a characterized antibody raised to the amino-terminal 82 amino acids of the Nramp1 polypeptide which is reactive towards two polypeptide species of 45 MW×10³ and 90–100 MW×10³ transiently expressed from the Nramp1 cDNA in COS-1 cells.¹⁹ In RAW264.7 cells, expression of a precursor 45 MW×10³ polypeptide is detected, but as in agreement with others,²³ using macrophages from Balb/c mice or other susceptible strains there is no evidence that they express the mature 90–100 MW×10³ protein found in Nramp1^G169 allele macrophages nor is it inducible with lipopolysaccharide (LPS) or interferon-γ (IFN-γ) (Fig. 1). This antibody also detects an immunoreactive band of around 50 MW×10³ inducible by IFN-γ and/or LPS which we believe is not a component of the Nramp1 gene as it is also found in non-macrophage cells, NIH3T3 cells, which do not express the Nramp1 mRNA (not shown). As before Nramp1 aggregates following standard boiling treatment for SDS–PAGE.¹⁹

Figure 1.

Nramp1 expression in Nramp1^D169 RAW264.7 macrophage cells and 90–100 MW×10³ Nramp1^G169 expression in stable RAW264.7 transfectant clones. RAW264.7 parental cell extracts (20 μg) were analysed for Nramp1 polypeptide expression by Western blotting. Extracts were prepared from untreated cells or following stimulation with: IFN-γ (10 U/ml), LPS (100 ng/ml) or both for 20 hr with or without boiling as indicated (a). (b) 20 μg of total cell extracts prepared from G418 drug resistance Nramp1 transfectant clones were analysed for Nramp1 expression without the standard boiling treatment and compared with an equivalent loading of Nramp1^G169 N11 cells. Of the sense orientation transfectant clones, R32, R37, R38 and R39 exhibited high molecular polypeptides displaying immunoreactivity to the anti-Nramp1 antisera of 90–100 MW×10³ and 65 MW×10³. Sense orientation clones R33 and R34 failed to exhibit immunoreactivty despite being resistant to G418.

Stable, high level expression of Nramp1 in RAW264.7 cells

To obtain transfectant lines in RAW264.7 cells we adopted the vector pHβA-1-neo, containing 4·6 kb of 5′ flanking DNA, the first intervening sequence, and 5′ untranslated sequence of the human β-actin promoter.¹⁸ This vector is competent at expressing Nramp1 polypeptides by transient transfection in COS-1 cells, but at a lower level than using the pCDNA3 plasmid (not shown). Analysis of drug-resistant clonal RAW264.7 lines from the transfection of the pHβA-1-neo plasmid containing the Nramp1 insert in the sense, but not the antisense orientation, revealed that this plasmid construct directed high-level expression of a high molecular weight polypeptide (Fig. 1). This polypeptide displays no immunoreactivity in parental RAW264.7 cells with the anti-Nramp1 rabbit polyclonal antibody (Fig. 1). The levels of protein, previously shown to be the mature and fully glycosylated Nramp1 polypeptide were comparable or slightly more than the Nramp1^G169 N11 cells¹⁹. However, like the other workers²⁴ we also show that the Nramp1 protein in the RAW264.7 background is of slightly reduced apparent molecular weight compared with Nramp1 from N11 cells. We propose this microheterogeneity is reflected at the post-translational level and possibly at the level of glycosylation, since treatment with PNGaseF produces an identically sized aglycosyl 45 MW×10³ protein in both cell types (not shown). In the sense orientation Nramp1 transfectant cells we also observe an intermediate band between the 90–100 MW×10³ protein and the precursor form of 45 MW×10³. This band could be related to the IFN-γ- and LPS-inducible product described previously in both bone marrow-derived and microglial macrophages.⁹

Levels of iron in parental and Nramp1-expressing RAW264.7 cells

Previously we described that ectopic expression of Nramp1 in COS-1 cells was associated with reduced iron loads.¹⁹ A similar assay applied to the described RAW264.7 Nramp1 transfectants demonstrated expression of the 90–100 MW×10³ Nramp1 polypeptide also was associated with reduced iron loads by about 29–43% and 27–49% in experiments 1 and 2, respectively, using the biochemical ferrozine assay and compared with R21 cells (Table 1). Statistical analysis by pooling groups of cells R21 and R22 compared with R32, R37, R38, R39 revealed that Nramp1-expressing cells were associated with significantly lower iron levels (P < 0·001)

Macrophage-associated iron following challenge with iron:NTA. RAW264.7 cell transfectants were seeded at 5×10⁶ cells per 90-mm plate and incubated overnight to adhere. Next day cells were treated with 500 μm Fe:NTA and harvested 20 hr later for iron determination by the ferrozine assay. Results are presented as total cellular iron/mg total protein as a percentage of the line 21. Shown are two independent experiments conducted on separate days for Nramp1-expressing lines 32, 37, 38, 39 and the non-expressing lines 21 and 22 as indicated. Levels of iron in the two groups were tested for differences with the Student's t-test and non-expressing lines (100±20) (mean±SD) were found to be significantly greater than expressing lines (66±10·4), P < 0·001

Chelatable iron in parental and Nramp1-expressing RAW264.7 cells

As Nramp1 is expressed on an intracellular membrane and therefore located at the interface between extracellular iron within the lumen of the endosomal vesicle, and its entry into the cytoplasm, we proposed that expression of the functional Nramp1 polypeptide could modify the rate and or magnitude of iron entry into the cytoplasm. Here is a so-called free or redox-active pool of iron that provides the necessary iron for biosynthetic pathways. The latter also constitutes a transit pool of iron from the import vesicle destined either for storage within the ferritin protein or secretion. To evaluate the influence of Nramp1 over this iron pool we developed an assay to monitor levels of chelatable iron in macrophages in response to challenge with ⁵⁹Fe. In these experiments we established that iron uptake and the magnitude of the chelatable iron pool was stimulated by the presence of extracellular iron over the range of unlabelled iron concentrations. No evidence of saturable iron uptake was observed. We established that the level of chelatable iron associated with antisense Nramp1 transfectant RAW264.7 cells was 1·2–2-fold greater than Nramp1 90–100 MW×10³ polypeptide expressing cells (Table 2) significant to at least the 5% level over all concentrations of iron tested.

Chelatable iron in Nramp1-expressing and -non-expressing RAW264.7 cell lines

Cells were plated out as described in the legend to Table 1 and challenged with ^59Fe (2·3 μm) and with the indicated doses of unlabelled iron, both as chelates with NTA, for 3 hr. After 3 hr cells were washed with phosphate-buffered saline (PBS), cells harvested and chelatable iron determined as described in the Materials and methods. Determinations were made for triplicate cultures and results presented as means ± standard deviation for the Nramp1-expressing transfectant line R37 and the antisense control R21. Statistical analysis using the Student's t-test revealed iron levels in R21 cells were significantly greater than R37 cells over all concentrations of iron at least to the 5% level.

The measurement of cytosolic redox-active iron following detergent lysis of cells is likely to be over-estimated owing to the release of iron from solubilized vesicles. Therefore we investigated the dynamics of the ⁵⁹Fe pool first by loading cells with the low-molecular-weight iron:NTA chelate followed by pulsing with unlabelled iron. As before all expressing cell lines R32, R37, R38 and R39 exhibited significantly lower levels of chelatable iron compared with R21 at 0 hr (P < 0·05). The dynamics of this free, chelatable iron pool were determined at 45-min intervals after the chase for lines R21 and R37. Using this protocol we observed a steady drop in the percentage of the chelatable iron pool by 50% in expressing line R37 over a 3-hr chase period. Iron levels at 3 hr post-chase were significantly lower than at 0 hr (P < 0·01) for R37. In contrast, over this same time period changes in the iron levels for R21 were not statistically significantly different between values before and after cold chase (Table 3).

Chelatable iron levels in transfectant lines and changes following chase with unlabelled iron in Nramp1-expressing (R37) and -non-expressing (R21) RAW264.7 cell lines. Cells were plated out, 2×10⁶ per 90-mm dish, the previous day and incubated for 3 hr for loading with ⁵⁹Fe at 2·3 μm final and 50 μm of non-readiactive iron, both as chelates with NTA. After 3 hr, cells were washed and chelatable iron determined according to the Materials and methods or chased with 50 μm non-radioactive iron for indicated times. Chelatable iron and total cellular iron was measured in triplicate as described in the Materials and methods at indicated time intervals. The chase experiment shown is a representative for lines 37 and 21. Statistical analysis was performed by use of the Student's t-test. Differences at the 5% level between line 21 at 0 hr, are indicated a, between line 37 at 0, b (0·1% level) and values for line 21 between 0 and 3 hr were not significantly different, c. Results are presented for chelatable iron as the percentage of total iron uptake (chelatable/chelatable+nonchelatable)×100

Calcein imaging of iron transport

To complement the detergent lysis assay for determination of cellular free-iron measurements another approach was adopted to evaluate flux of iron from the endosomal vesicle into the cytoplasm. This assay utilized the fluorescent indicator dye calcein, which upon iron binding exhibits fluorescence quenching. Challenge of calcein-loaded cells with the Fe:NTA resulted in a greater reduction in the quenching of fluorescence in Nramp1-expressing cell R37 compared with the parental or antisense line R21 (Fig. 2). The magnitude of quench was greater following challenge with 100 μm iron than 10 μm, 25 vs. 10 fluorescence unit decrease. That these effects were mediated by iron was shown by the restoration of calcein fluorescence by addition of specific iron chelators (not shown).

Figure 2.

Determination of endosomal iron efflux using the fluorescent dye calcein. RAW264.7 parental cells, sense (R37) or antisense orientation (R21). Nramp1 cDNA transfectants, as indicated, were loaded with calcein as described in the Materials and methods. Prior to the addition of iron chelate a fluorescent baseline was established of which the final 2 min are shown on the traces. Fe: NTA at (a) 10 μm or (b) 100 μm was added at the indicated time and the resulting decrease in calcein fluorescence monitored.

DISCUSSION

It has been established that the product of the Nramp1 locus plays a major role in the immune response leading to the restriction of intramacrophage pathogen proliferation in mice. The biochemical basis for this growth restriction is not clear. Nor is the basis by which Nramp1 exerts effects over macrophage differentiation and pro-inflammatory immune responses, although reports suggest a role in the stabilization of macrophage pro-inflammatory gene mRNA transcripts.²⁵ The functional analysis of Nramp1 and other members of this gene family has implications for delineating mechanisms of susceptiblity to human disease and understanding disease resistance. Recent genetic analysis of human NRAMP1 revealed evidence for a role not only in susceptibility to infectious disease,^26,27 but in autoimmune diseases.^28,29 The susceptible allelic variant of murine Nramp1, G169D, has not, so far, been identified in humans. Functional polymorphisms have been described in NRAMP1 including a promoter dinucleotide repeat polymorphism. An allelic variant of this polymorphism has been identified that influences transcriptional responses. These data provide evidence that over-expression of NRAMP1 may contribute to autoimmune diseases.^1–3 Evidence supporting a role for Nramp1 in intracellular iron transport is compelling and is implied by the strong sequence conservation with Nramp2. Furthermore, Nramp1 was reported to stimulate ⁵⁵Fe uptake when expressed in Xenopus oocytes¹² and to modulate iron levels in phagolysosomes,¹⁶ although all studies do not necessarily provide an association between Nramp1 and iron transport, since Nramp1, unlike Nramp2 could not complement the yeast SMF genes. Gomes & Appelberg have shown that excess iron hampers Nramp1-encoded function and strongly argue a role of the Nramp1 product that depletes the phagosome of iron.³⁰ Recently Hackman and colleagues³¹ reported Nramp1 influences acidification of phagolysosomal vesicles containing live mycobacteria. In these studies it was suggested that recruitment of functional Nramp1 to the phagolysosomal membrane represents an obligatory step in the maturation of this organelle and that Nramp1 in the membrane of the phagolysosome may act as a fusogen to facilitate V-ATPase recruitment and subsequent acidifcation.³¹ Here we have developed a model system to study Nramp1 protein structure and function in order to examine another role for Nramp1, that of intracellular iron transport.

The results of this study substantiate the association between Nramp1 and a role in iron transport. Assays using calcein and the chelation of ⁵⁹Fe provide support for increased flux of iron into the cell which can be stored, used or secreted. Current data favour release of the acquired iron from the resting macrophage. These studies also confirm the G169D mutation abolishes 90–100 MW×10³ Nramp1 polypeptide expression in macrophages from susceptible strains of mice. We do however, detect expression of a 45 MW×10³ Nramp1 precursor polypeptide derived from the D169 allelic variant, which has not been reported previously, suggesting the G169D mutation inhibits glycosylation by some unknown mechanism, although this polypeptide exhibits a vesicular perinuclear distribution like the functional polypeptide¹⁰. The effect of the G169D mutation on glycosylation is suprising in view of the distance in the primary sequence between the mutation site at 169 and glycosylation sites at 321 and 335 amino acids. It is not known if this polypeptide is active or exhibits partial function; however, there are no reports that Nramp1 susceptible mice suffer from anaemia, but this may be as a consequence of genetic redundancy.

The current work provided support for Nramp1 acting to mobilize iron from the endosomal vesicle following uptake of low molecular iron chelates. Whether this is analogous to the process of iron salvage from effete red blood cells or apoptotic cells in general under non-inflammatory conditions requires further study as does the fate of the iron following its translocation from the endosomal vesicle to the cytoplasm. Although the observation of reduced total cellular iron- load estimations favour secretion of the iron. Thus, Nramp1-deficient cells may have a block in normal iron flux resulting in the accumulation of iron within the endosomal vesicle and it is this pool that may be accessible to intracellular pathogens resulting in their enhanced proliferative rates.

Another area for future study is the influence of cytokine stimulation on macrophage iron homeostasis, as during chronic infection a condition of the anaemia of chronic disease or anaemia of inflammation is described in which a redistribution of iron from the circulation to the tissues takes place. It is not apparent if Nramp1 plays a role in this process, although the reticuloendothelial cells are pivotal, nor is there any evidence to support altered Nramp1 transport activity following cytokine stimulation, but cytokine stimulation does modulate Nramp1 prevalence. We do have preliminary data supporting a change in the phosphorylation status of Nramp1 in response to cytokine stimulation (Atkinson & Barton unpublished data); the functional consequences are however, unknown. We believe the transfection system described in this work provides a suitable system in which to explore the role of phosphorylation over the control of Nramp1 function, be it modulatory at the level of transport activity or regulation of cellular location, and to evaluate the influences of cytokines on both mobilization, transport and secretion of iron.