Down-regulation of Iron Regulatory Protein 1 Activities and Expression in Superoxide Dismutase 1 Knock-out Mice Is Not Associated with Alterations in Iron Metabolism*

Iron and oxygen (O 2 ) are intimately associated in many well characterized patho-physiological processes. These include oxidation of the [4Fe-4S] cluster of mitochondrial aconitase and inactivation of this Krebs cycle enzyme by the superoxide anion (O 2 .), a product of the one-electron reduction of O 2 . In contrast to the apparent toxicity of this reaction, the biological consequences of O 2 .-mediated inactivation of the cytosolic counterpart of mitochondrial aconitase, commonly known as iron regulatory protein 1 (IRP1), are not clear. Apart from its ability to convert citrate to iso -citrate, IRP1 in its apo-form binds to iron-responsive elements in the untrans-lated regions of mRNAs coding for proteins involved in iron metabolism, to regulate their synthesis and thus control the cellular homeostasis of this metal. Here, we show that

. -mediated inactivation of the cytosolic counterpart of mitochondrial aconitase, commonly known as iron regulatory protein 1 (IRP1), are not clear. Apart from its ability to convert citrate to iso-citrate, IRP1 in its apoform binds to iron-responsive elements in the untranslated regions of mRNAs coding for proteins involved in iron metabolism, to regulate their synthesis and thus control the cellular homeostasis of this metal. Here, we show that in superoxide dismutase 1 (SOD1) knock-out mice, lacking Cu,Zn-SOD, an enzyme that acts to reduce the concentration of O 2 . mainly in cytosol, not only is aconitase activity of IRP1 inhibited but the level of IRP1 is also strongly decreased. Despite such an evident alteration in IRP1 status, SOD1-deficient mice display a normal iron metabolism phenotype. Our findings clearly show that under conditions of O 2 . -mediated oxidative stress, IRP1 is not essential for the maintenance of iron metabolism in mammals.
Aerobic organisms use molecular oxygen (O 2 ) in the generation of energy. The use of O 2 as the terminal electron acceptor in oxidative phosphorylation involves biochemical reactions based on the interaction of O 2 with iron atoms acting as functional co-factors in the catalytic centers of a large number of proteins. However, the ability of iron to react with partially reduced forms of O 2 such as the superoxide anion (O 2 . ) and hydrogen peroxide (H 2 O 2 ) may lead to the formation of the hydroxyl radical ( ⅐ OH), a highly destructive oxidant (1). Since no specific molecule has evolved to fulfil the role of ⅐ OH scavenger, living organisms have developed different ways to pre-vent its formation. These include both O 2 .
(2) and H 2 O 2 (3,4) scavenging enzymes as well as the stringent control of iron metabolism (5). The role of O 2 . in oxygen toxicity is underscored by the ubiquity and diversity of superoxide dismutases (SODs) 1 that catalyze its conversion to H 2 O 2 and O 2 (2). Accordingly, deficiency in various forms of SOD promotes oxidative damage in a wide range of organisms (2). A long-standing explanation of the toxicity of O 2 . proposes that it is the rate-limiting reducing factor for the pre-existing pool of free iron active in the generation of ⅐ OH via a Fenton reaction (6). Alternatively, recent evidence coming from studies on bacteria (7) and yeast (8,9) shows that O 2 . is responsible for elevating free iron levels.
A likely mechanism is based on the O 2 . -mediated release of iron from certain [4Fe-4S]-containing proteins, of which mitochondrial aconitase is particularly susceptible to oxidative inactivation by O 2 . (10,11). In contrast to prokaryotes and yeast, higher eukaryotes possess an additional aconitase, a cytosolic counterpart of the mitochondrial enzyme playing a major role in the regulation of cellular iron metabolism (5). This bifunctional protein, commonly known as iron regulatory protein 1 (IRP1), apart from its ability to convert citrate to iso-citrate, also exhibits trans-regulatory activity.  (14).
The question addressed in this study was whether deficiency in the Cu,Zn-SOD in mammals, supposed to increase the O 2 .
steady-state level, modulates IRP1 activities, and consequently, influences iron metabolism. As an experimental model, we used mice lacking Cu,Zn-SOD activity due to disrup-tion of the SOD1 gene (15). As expected, we observed a substantial decrease in the aconitase activity of IRP1 in homozygous knock-out mice (SOD1 Ϫ/Ϫ ) as compared with wild-type (WT) littermates. Surprisingly, this inhibition of IRP1 enzymatic activity was associated with a marked down-regulation of IRP1 protein expression. Despite such a major alteration in IRP1 status, iron metabolism appeared to be largely unaffected in SOD1-deficient mice.

EXPERIMENTAL PROCEDURES
Reagents-The sources of the main reagents used in this study are indicated below. All others chemicals were obtained from Sigma.
Mice-A breeder pair of mice (strain B6;129S7-Sod1 tm1Leb ) heterozygous for a SOD1 tm1Leb targeted mutation (15) and their progeny were provided by The Jackson Laboratory (Bar Harbor, ME). Males and females heterozygous for the non-functional SOD1 allele (SOD1 Ϫ/ϩ ) were intercrossed, and their progeny was kept at 24 -25°C with a light-dark cycle of 12 h. Mice received a standard laboratory diet (Labofeed, Kcynia, Poland) and water ad libitum. Genotyping using DNA isolated from mouse tails was performed by PCR analysis according to the protocol provided by The Jackson Laboratory. Mice homozygous for the non-functional SOD1 allele (SOD1 Ϫ/Ϫ ) producing no SOD1 protein (15) and control mice homozygous for the wild-type SOD1 allele (SOD1 ϩ/ϩ ) were used in the study at the age of 2 months. All procedures were approved by the Local Ethical Commission (permission number 67/2001).
Preparation of Mouse Liver Extracts-Livers were excised immediately after death from mice killed by cervical dislocation, quick-frozen, and stored in liquid nitrogen. Liver samples (50 mg) were thawed and homogenized, and erythroid cells were lysed by suspension in a hypotonic solution. After a brief centrifugation, the pellets were washed in phosphate-buffered saline, resuspended in 100 mM HEPES (pH 7.4), 0.25 M sucrose, supplemented with protease inhibitor mixture (Roche Diagnostics). Digitonin was then added to 0.01%, and the mixture was held at 4°C for 15 min. After centrifugation at 1800 ϫ g for 10 min, supernatants were collected and recentrifuged at 150,000 ϫ g for 1 h at 4°C. Portions of the fresh cytosolic extracts were immediately used for IRP1 aconitase activity determination. The remaining extracts were aliquoted and kept at Ϫ80°C until required for the electrophoretic mobility shift assay and for Western blot analysis of IRP1 and Ft levels. For IRP2 detection by Western blotting, total hepatic extracts were prepared using 50-mg samples of liver. The lysis of erythroid cells was performed as for the cytosolic extracts preparation, and then pellets were homogenized in 20 mM HEPES (pH 7.6) containing 25 mM KCl, 0.5% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture. After centrifugation of homogenates at 12,000 ϫ g for 30 min at 4°C, supernatants were collected and used immediately for IRP2 detection by Western blotting. The whole procedure of lysate preparation was performed anaerobically under a flux of argon to avoid IRP2 oxidation.
Determination of Aconitase Activity-Aconitase activity in liver cytosolic extracts was measured spectrophotometrically by following the disappearance of cis-aconitate at 240 nm at 37°C, as described previously (16).
Electrophoretic Mobility Shift Assay-IRP1-IRE interactions were examined as described previously (17) by incubating 2 g of the cytosolic protein extracts with a molar excess of [ 32 P]CTP-labeled H-ferritin IRE probe. In parallel experiments, cytosolic extracts were treated with 2-mercaptoethanol at a final concentration of 2% before the addition of the IRE probe to produce maximal IRE-binding activity (18). IREprotein complexes were then separated by electrophoresis on 6% nondenaturing polyacrylamide gels. The signals representing the IRE-IRP1 complexes were quantified with a Molecular Imager using Quantity One software (Bio-Rad).
Western Blotting-Fifty micrograms of liver cytosolic (for IRP1 and Ft subunits detection) or total (for IRP2 detection) extracts were re-solved on 8 (for IRP1 and IRP2 detection) or 15% (for Ft subunit detection) SDS-polyacrylamide gels and transferred to Hybond-ECL nitrocellulose membranes (Amersham Biosciences). The membranes were initially blocked by gentle agitation in Tween 20-Tris-buffered saline buffer containing 5% fat-free dry milk at room temperature for 1 h followed by overnight incubation at 4°C with a chicken polyclonal antibody raised against purified human recombinant IRP1 (Agro-Bio, La Ferté Saint-Aubin, France), antibody raised against IRP2 peptide, YQKAGKLSPLKVQPKKLP (Neosystems, Strasbourg, France), or rabbit antisera raised against recombinant mouse H-and L-Ft (kindly provided by Dr. P. Santambrogio, Milan, Italy). Membranes were then washed and incubated with peroxidase-conjugated rabbit anti-chicken or goat anti-rabbit secondary antibodies for 1 h at room temperature. Immunoreactive bands were detected using the enhanced chemiluminescence Western blotting detection system ECL Plus (Amersham Life Sciences).
FIG. 1. IRP1 activities and protein level are decreased in the livers of SOD1 ؊/؊ mice. A, aconitase activity of IRP1 was determined spectrophotometrically in hepatic cytosolic extracts (devoided of mitochondrial aconitase as checked by Western blot using an antibody raised against purified beef heart mitochondrial aconitase, kindly provided by Dr. R. B. Franklin, University of Maryland, Baltimore, MD, data not shown) by measuring the disappearance of cis-aconitate at 240 nm as described previously (16). Aconitase activity values are the means Ϯ S.D. from six mice of both SOD1 Ϫ/Ϫ and SOD1 ϩ/ϩ genotypes. B, left-hand panel, 2 g of cytosolic liver protein extracts were analyzed for IRE binding activity of IRP1 as described under "Experimental Procedures." Data shown are representative of EMSA analyses performed using liver cytosolic extracts obtained from six WT and SOD1 Ϫ/Ϫ mice. B, right-hand panel, radioactivity associated with IRE-IRP1 complexes without or with 2% 2-mercaptoethanol (2-ME) was quantified with a Molecular Imager using Quantity One software (Bio-Rad) and plotted in arbitrary units. Results are expressed as mean Ϯ S.D. for six mice of both SOD1 Ϫ/Ϫ and SOD1 ϩ/ϩ genotypes. C, left-hand panel, IRP1 levels were analyzed by Western blotting as described under "Experimental Procedures." The analyses were performed using liver cytosolic extracts obtained from six mice of each genotype, and representative results are shown. Blots were also reprobed with monoclonal anti-␤actin antibody from mouse as a loading control C, right-hand panel, the intensity of the IRP1 bands was quantified with a molecular Imager using Quantity One software (Bio-Rad) and is plotted in arbitrary units to present protein level. Results are expressed as mean Ϯ S.D. for six mice of both SOD1 Ϫ/Ϫ and SOD1 ϩ/ϩ genotypes.

IRP1 Activities and Expression in SOD1 Knock-out Mice
Reverse Transcription (RT)-PCR Analysis of IRP1 mRNA Abundance-For RT-PCR analysis, total cellular RNA was extracted from liver samples (20 mg) using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Two micrograms of total RNA were used for reverse-transcription with superscript II RNase H reverse transcriptase (Invitrogen) at 45°C for 1 h. Of this reaction, 2 l were then taken for PCR amplification with Platinum Pfx DNA polymerase (Invitrogen) using the respective pairs of oligonucleotide primers indicated in Table I. PCR products were analyzed on 2% agarose gels run with Tris-borate-EDTA buffer and stained with ethidium bromide. The bands were quantified with a Molecular Imager using Quantity One software (Bio-Rad) and normalized to those for a glyceraldehyde-3-phosphate dehydrogenase housekeeping gene control.
Real-time Quantitative RT-PCR-IRP1, H-Ft, L-Ft, and TfR hepatic mRNAs were measured by real-time quantitative RT-PCR using a Light Cycler (Roche Diagnostics). The same cDNA templates and respective primers indicated in Table I were used. The detection of amplification products was carried out using SYBR Green (Roche Applied Science) as described previously (19). To confirm amplification specificity, the PCR products from each primer pair were subjected to melting curve analysis and subsequent agarose gel electrophoresis. For data analysis, Light Cycler 3.5 Software was used. Quantification was performed relative to glyceraldehyde-3-phosphate dehydrogenase control.
Isolation of Mouse Peritoneal Cells and LIP Assay-Mice were killed by cervical dislocation, and the peritoneal cavities were washed with Hanks' balanced salt solution, pH 7.4. The lavage fluid from individual mice was centrifuged at 400 ϫ g for 10 min. Sedimented cells were washed twice with Hanks' balanced salt solution, counted, checked for their viability by trypan blue exclusion, and processed for labile iron pool (LIP) measurement using the fluorescent probe calcein as described previously (20).
Non-heme Iron Determination-Quantification of the non-heme iron content liver fragments (100 mg) was performed using a BIO-CHEMTEST (Polish Chemical Reagents, Gliwice, Poland) as described previously (21).
Total Iron Determination-Liver samples (0.5 g) were mineralized in a mixture of 5 ml of HNO 3 and 1 ml of H 2 O 2 in a microwave oven (Ethos 900, Milestone). Iron content was estimated by flame (air-acetylene) atomic absorption spectrometer (PerkinElmer Life Sciences 1100B) (248.3 nm). The external standards were prepared using a 9972 Titrisol iron standard (Merck).
Hematological Analysis of Mice-Heparinized blood was obtained by direct heart puncture immediately after death. Blood cell counts and erythrocyte parameters were determined using an automated SYMEX F820 Analyzer.
Protein Determination-The protein content of extracts was determined spectrophotometrically at 595 nm by the method of Bradford using a Bio-Rad protein assay kit with bovine serum albumin as the standard.
Statistical Analysis-For statistical evaluation, one-way analysis of variance was applied. For estimation of the significance of difference between arithmetic means, the Scheffe test was used. All calculations were done using Statgraphics Plus 6.0 software.

RESULTS AND DISCUSSION
It is generally believed that the generation of extremely toxic ⅐ OH through the Fenton reaction is the main point at which the pathologic interaction between oxygen and iron occurs. In mammals, the participation of iron in ⅐ OH formation is limited by numerous proteins implicated in iron trafficking, storage, and redox state regulation (22). However, the iron-related toxicity of reactive oxygen species, i.e. O 2 . and H 2 O 2 , is not limited to their direct redox reactions with iron leading to ⅐ OH formation but may also involve disturbance of regulatory mechanisms underlying the homeostasis of this metal (23). A primary response to extracellular H 2 O 2 is the induction of IRP1 RNA binding activity (24,25), which results in a coordinated pattern of down-and up-regulation of Ft and TfR expression, respectively, and subsequent rearrangement of cellular iron (26). The superoxide anion, like intracellular H 2 O 2 (24,25), has the ability to mediate partial disassembly of the IRP1 [4Fe-4S] cluster, which proceeds by its oxidation, the subsequent release of one iron atom, and the generation of a so-called "null" [3Fe-4S] form of IRP1, active neither as an aconitase nor as a trans-regulator (27). It is important to note that evidence of the generation of a non-functional IRP1 form by O 2 . comes from experiments in which mammalian cell lysates were exposed to a bolus flux of O 2 . produced enzymatically (12,28 4). C, IRP1 mRNA abundance in livers was measured by real-time RT-PCR as described under "Experimental Procedures." Each column represents the mean (Ϯ S.D. for SOD1 Ϫ/Ϫ mice) of two amplification reactions, performed on a single cDNA sample reverse-transcribed from RNA prepared from three mice. The level of IRP1 mRNA level in SOD1 ϩ/ϩ mice was assigned as the 100% value.
FIG. 3. SOD1 ؉/؉ and SOD1 ؊/؊ mice show similar hepatic IRP2 levels. IRP2 levels were assessed in total liver extracts prepared anaerobically by Western blotting as described under "Experimental Procedures." To assist identification of IRP2 bands, cytosolic extracts from control, ferric ammonium citrate (FAC)-, and isonicotinoyl salicylaldehyde hydrazone (SIH)-treated mouse RAW 264.7 macrophages were electrophoresed on the same gel. Iron-dependent regulation of IRP2 levels in RAW 264.7 cells is as described previously in response to fluctuations in iron level (48). The position of the molecular mass marker is indicated. Data are representative for independent Western blot analyses performed using liver extracts from three mice of each genotype.
sophila (14). Interestingly, in contrast to the aforementioned in vitro studies, increased O 2 . in Drosophila-null mutants for SOD1 resulted not only in inhibition of IRP1 aconitase activity but also in activation of IRP1 IRE binding (14).
In this study, we investigated the effect of complete SOD1 deficiency in mammals on IRP1 activities and the consequences for iron metabolism regulation. For this purpose, we used mice lacking the Cu,Zn-SOD activity due to disruption of the SOD1 gene (15). Considering the high expression of both IRP1 (29) and SOD1 (30) in liver, and the key role of this organ in the regulation of mammalian iron homeostasis (31), we focused our analysis on hepatic IRP1 and measured most parameters of iron metabolism in this tissue. As expected, we found a substantial Ͼ80% decrease in the aconitase activity of IRP1 in SOD1 Ϫ/Ϫ mice as compared with their wild-type, SOD1 ϩ/ϩ siblings (Fig. 1A). These results are reminiscent of those showing inactivation of cytosolic aconitase in mammalian cell lines in response to O 2 . (10,12). We also observed that spontaneous IRE binding activity of IRP1 is decreased in SOD1 Ϫ/Ϫ as compared with WT mice (Fig. 1B) (32). Surprisingly, in SOD1 Ϫ/Ϫ mice, maximal IRP1 binding activity present in 2-mercaptoethanol-treated hepatic cytosolic extracts revealed almost the same decrease as for aconitase activity (Fig. 1B). Since assessment of total IRP1 IRE binding activity is considered an indirect measure of IRP1 level (18), this finding suggested that the content of IRP1 protein is greatly reduced in SOD1 Ϫ/Ϫ mice. Indeed, direct measurement of IRP1 in liver cytosolic extracts by Western blotting confirmed the results of electrophoretic mobility shift assay experiments and revealed a Ͼ80% decrease in IRP1 level in knockout SOD1 versus normal mice (Fig. 1C). We next determined whether this diminution in the amount of IRP1 at the protein level reflected a decrease in the level of IRP1 mRNA. Quantification of IRP1 mRNA content by both semiquantitative RT-PCR (Fig. 2, A and B) and real-time quantitative RT-PCR (Fig.  2C) consistently showed a reduction of IRP1 mRNA in knockout SOD1 mice of about 30 -40% relative to the WT. These analyses suggest that the substantial decrease in IRP1 protein content in SOD1-deficient mice may be only partially due to less efficient transcription of the IRP1 gene or decreased stability of IRP1 mRNA. Considering that degradation of several [Fe-S] cluster-containing enzymes is associated with alter-ations in their clusters (33,34), it is tempting to propose that down-regulation of IRP1 in SOD1 Ϫ/Ϫ mice may also be due to an increased susceptibility of this protein to proteolytic degradation after destabilization of its  (38), and subsequent increased sensitivity of phosphorylated apo-IRP1 to iron-dependent degradation (39). We also considered the possibility that aberrations in zinc metabolism in SOD1 knockout mice, reported previously in yeast SOD1 mutants (40), could be the cause of altered IRP1 IRE binding. In this respect, it is noteworthy that inhibition of IRP1 binding to IRE due to specific in vitro aggregation of purified IRP1 apo-protein in the presence of micromolar concentrations of zinc has been recently reported (41). Disruption of zinc homeostasis has been proposed to be one of several possible factors responsible for protein precipitation observed in several neurodegenerative disorders (42) including SOD1 precipitation in amyotrophic lateral sclerosis (43). Although we were unable to find a significant difference in total hepatic zinc content between SOD1 ϩ/ϩ and SOD1 Ϫ/Ϫ mice (data not shown), for several reasons, we cannot exclude that the level of cytosolic free zinc is elevated in SOD1-  -FT and rL-Ft), generous gift from Dr. P. Santambrogio) and purified liver ferritin were used as positive controls. The analyses were performed six times using liver cytosolic extracts obtained from separate mice, and representative results are shown. Blots were also reprobed with monoclonal anti-␤-actin antibody from mouse as a loading control. B, mRNA levels of liver Ft subunits were measured by real-time RT-PCR as described under "Experimental Procedures." Each column represents the mean of two amplification reactions (Ϯ S.D. for SOD1 Ϫ/Ϫ mice) performed on a single cDNA sample reverse-transcribed from RNA prepared from three mice. Levels of H-and L-Ft mRNAs in SOD1 ϩ/ϩ mice were assigned as the 100% values.
FIG. 5. Hepatic TfR mRNA abundance is unaffected in SOD1 ؊/؊ mice. The liver TfR mRNA level was measured by real-time RT-PCR as described under "Experimental Procedures." Each column represents the mean of two amplification reactions (Ϯ S.D. for SOD1 Ϫ/Ϫ mice) performed on a single cDNA sample reverse-transcribed from RNA prepared from three mice. Levels of H-and L-Ft mRNAs in SOD1 ϩ/ϩ mice were assigned as the 100% values. deficient mice. First, Cu,Zn-SOD represents an abundant "sink" for zinc, and in mice lacking SOD1 protein, the intracellular environment may contain an excess of this metal. Second, under conditions of oxidative stress, which seem to occur in SOD1-deficient mice, the potential for metallothionein to scavenge cellular zinc is decreased (44).
Most studies on the regulation of IRP1 activities have been mainly focused on post-translational mechanisms underlying the insertion/extrusion of a [4Fe-4S] cluster into/from the IRP1 molecule (5,13). A few recent studies have indicated the possible impact of changes in intracellular IRP1 protein levels on the binding of IRP1 to IREs (45)(46)(47). Interestingly, nitric oxide (NO), the ability of which to induce IRP1 activity by removing its [4Fe-4S] cluster is now well established (13), has been shown to down-regulate IRP1 expression (45). There is a striking similarity between NO and O 2 . as regards the major mechanisms by which these two free radical molecules affect IRP1 activity, i.e. interaction with the [4Fe-4S] cluster and regulation of IRP1 protein abundance. It therefore seems reasonable to assume that a recent proposal concerning the bidirectional regulation of IRP1 by NO, which has been suggested to serve as a flexible adaptation of iron metabolism to nitrosative stress (45), may also be valid under conditions of O 2 . -mediated oxidative stress. Apart from IRP1, the post-transcriptional control of iron homeostasis in mammals also involves IRP2, a second IRP, which binds IREs with similar affinity to IRP1. IRP2 does not contain an [Fe-S] cluster and therefore has no aconitase activity. Its RNA binding activity is controlled through the regulation of protein stability (48). The two IRPs are regulated by the intracellular iron level, although in several mammalian cell lines, IRP2 has been shown to respond more promptly to changes in iron concentration than IRP1 (49). Accordingly, in mice receiving an iron-deficient diet, IRP2 has been reported to be more readily activated than IRP1 (29). Several other stimuli such as hypoxia (50) and NO (51) have been shown to differentially regulate the RNA binding activities of the two IRPs, suggesting that other regulatory mechanisms exist to counterbalance an excessive iron deregulation under stressful conditions. Moreover, it is noteworthy that in some tissues of IRP1 Ϫ/Ϫ mice, IRP2 levels have been found to be increased, suggesting that IRP2 can compensate for loss of IRP1 (29). These data strongly suggest that in SOD1-deficient mice with reduced IRP1 content in the liver, hepatic expression of IRP2 might be up-regulated. However, as shown in Fig. 3, the IRP2 level in the livers of SOD1 Ϫ/Ϫ mice assessed by Western blotting is the same as that found in WT littermates. This indicates that IRP2-mediated control of iron metabolism in SOD1-deficient mice is unaltered.
To study the consequences of decreased IRP1 expression on the phenotype of iron metabolism in SOD1 Ϫ/Ϫ mice, we first compared hepatic levels of both H-and L-Ft subunits, the synthesis of which is controlled by IRP1 at the post-transcriptional level (5). As shown in Fig. 4A, the amount of the two Ft peptides remains unaffected in the livers of SOD1-deficient mice. Similarly, these mice display unchanged levels of both Hand L-Ft mRNA (Fig. 4B). The lack of changes between two groups of mice is also visible as regards TfR mRNA abundance in liver (Fig. 5). The normal content of hepatic Ft subunits and TfR mRNA level in SOD1 Ϫ/Ϫ mice, despite the marked decrease in IRP1 level, may be explained by the fact that in SOD1 ϩ/ϩ mice, the amount of hepatic apo-IRP1 able to bind IRE, and thus to inhibit Ft mRNA translation, is very low, accounting for only 10 -15% of the total activable IRP1. In agreement with these data, apo-IRP1 has been shown to constitute a minor fraction of total IRP1 in animal tissues (29). Although reduction of the IRP1 content in SOD1 Ϫ/Ϫ mice is reflected by diminished spontaneous IRP1 IRE binding activity, this decrease appears to be of negligible importance for the derepression of Ft synthesis or destabilization of TfR mRNA. As Ft has been shown to be the main determinant of LIP levels in mammalian cells (52), the unchanged content of Ft subunits provides an explanation for the almost identical LIP levels found in peritoneal cells of SOD1 Ϫ/Ϫ and SOD1 ϩ/ϩ mice (Fig.  6A). Since LIP levels are a sensitive indicator of intracellular iron homeostasis (53), these data therefore suggest that the cellular source of iron for metabolic processes is intact in SOD1 Ϫ/Ϫ mice. The measurement of hepatic non-heme and total iron content is a valuable tool for estimation of the overall iron status in mammals. It is affected in both iron deficiency and overload (54). Here, we provide evidence that the iron content in the livers of SOD1 Ϫ/Ϫ mice does not differ significantly from that found in control mice (Fig. 6, B and C). This suggests the absence of major abnormalities in systemic iron metabolism in SOD1 Ϫ/Ϫ mice. Blood hematological parameters can also indicate the status of the body iron metabolism. There-  6. Unaltered LIP level in peritoneal cells and liver iron content of SOD1 ؊/؊ mice. A, LIP levels cells were measured in peritoneal cells using a fluorescent probe, calcein, as described previously (20). The number of viable cells recovered from individual mice was between 1.5 and 2.5 ϫ 10 6 . Data shown are means Ϯ S.D., n ϭ 4. The values for hepatic non-heme (B) and total (C) iron content are expressed as the mean Ϯ S.D. and were obtained from 10 mice of each genotype. fore, we determined blood cell indices in the two groups of mice and found significant decrease in the erythrocytes number in SOD1 Ϫ/Ϫ mice as compared with their WT littermates (Table  II). Although these values are symptoms of a moderate anemia, it is unlikely that this is the result of iron deficiency since iron deficiency anemias are usually microcytic or normocytic (54) but never macrocytic as suggested by the increased mean corpuscular volume of erythrocytes observed in SOD1 Ϫ/Ϫ mice. Alternatively, it is possible that decreased erythrocyte number and hemoglobin level are associated with hemolysis that occurs in SOD1-deficient mice as a result of increased fluidity of the red blood cell membranes exposed to permanent O 2 . -mediated oxidative stress (55). Most features of iron metabolism analyzed in SOD1-deficient mice displaying reduced hepatic IRP1 level recall the unaltered iron phenotype described recently in mice with knock-out of the IRP1 gene (29). Similarly, murine pro-B lymphocytes, which do not express IRP1, are able to maintain normal post-transcriptional control of intracellular iron by using IRP2 exclusively (56). These examples of IRP1 deficiency suggest that in mammals, IRP1 is not essential for regulation of iron metabolism under the normal physiological conditions or upon O 2 . -mediated oxidative stress as reported in this study. It is relevant to note that the unaffected iron metabolism in SOD1-deficient mice is in keeping with the relatively normal development of these animals (15,30). It is noteworthy that misregulation of iron metabolism, e.g. that associated with IRP2 deficiency, is in general associated with severe neurodegenerative pathologies (57). On the other hand, there is a discrepancy between the intact iron metabolism in SOD1 Ϫ/Ϫ mice as shown here and the altered iron homeostasis reported in Cu,Zn-SOD-deficient bacteria (7) and yeast (8,9). In these former organisms, a raised level of free iron is correlated with increased sensitivity to oxidative stress associated with SOD1 deficiency (7).
It is now known that Cu,Zn-SOD possesses a number of other minor functions apart from O 2 . scavenging (58,59). Nevertheless, considering that O 2 . is a key detrimental molecule overproduced in SOD1-deficient organisms (2), our finding of IRP1 down-regulation in SOD1 Ϫ/Ϫ mice establishes a new regulation mechanism that may ensure the correct balance between oxygen-derived free radicals and iron. Moreover, we provide evidence that the regulation of IRP1 in liver in response to oxidative stress may discriminate between O 2 . -and H 2 O 2 -mediated stresses (60). Finally, it is tempting to speculate that mice lacking SOD1 and displaying a decreased potential for IRP1 to target IRE-containing mRNAs may be resistant to misregulation of iron metabolism by H 2 O 2 and NO. This proposal is consistent with the fact that H 2 O 2 is able to activate IRP1 but not IRP2 (5) and that IRP1 is a preferential target for NO-like chemical messengers (61).