Carlo Irace and Antonella Scorziello contributed equally to this work.
Divergent modulation of iron regulatory proteins and ferritin biosynthesis by hypoxia/reoxygenation in neurones and glial cells
Article first published online: 31 AUG 2005
Journal of Neurochemistry
Volume 95, Issue 5, pages 1321–1331, December 2005
How to Cite
Irace, C., Scorziello, A., Maffettone, C., Pignataro, G., Matrone, C., Adornetto, A., Santamaria, R., Lucio Annunziato and Colonna, A. (2005), Divergent modulation of iron regulatory proteins and ferritin biosynthesis by hypoxia/reoxygenation in neurones and glial cells. Journal of Neurochemistry, 95: 1321–1331. doi: 10.1111/j.1471-4159.2005.03449.x
- Issue published online: 7 SEP 2005
- Article first published online: 31 AUG 2005
- Received March 11, 2005; revised manuscript received May 16, 2005; accepted July 20, 2005.
- iron metabolism;
- iron regulatory proteins;
- neuronal vulnerability;
- oxidative stress
Ferritin, the main iron storage protein, exerts a cytoprotective effect against the iron-catalyzed production of reactive oxygen species, but its role in brain injury caused by hypoxia/reoxygenation is unclear. Ferritin expression is regulated mainly at post-transcriptional level by iron regulatory proteins (IRP1 and IRP2) that bind specific RNA sequences (IREs) in the 5′untranslated region of ferritin mRNA. Here, we show that hypoxia decreases IRP1 binding activity in glial cells and enhances it in cortical neurons. These effects were reversed by reoxygenation in both cell types. In glial cells there was an early increase of ferritin synthesis during hypoxia and reoxygenation. Conversely, in cortical neurons, ferritin synthesis increased during the late phase of reoxygenation. Steady-state analysis of ferritin mRNA levels suggested that ferritin synthesis is regulated mainly post-transcriptionally by IRPs in glioma cells, both transcriptionally and post-transcriptionally in type-1 astrocytes, and mainly at transcriptional level in an IRP-independent way in neurons. The different regulation of ferritin expression may account for the different vulnerability of neurons and glial cells to the injury elicited by oxygen and glucose deprivation (OGD)/reoxygenation. The greater vulnerability of cortical neurons to hypoxia-reoxygenation was strongly attenuated by the exogenous administration of ferritin during OGD/reoxygenation, suggesting the possible cytoprotective role exerted by this iron-segregating protein.
bovine serum albumin
Dulbecco's modified Eagle's medium
Earle's balanced salt solution
electrophoretic mobility shift assay
ferric ammonium citrate
fetal bovine serum
fetal calf serum
iron responsive element
iron regulatory protein
oxygen and glucose deprivation
reactive oxygen species
sodium dodecyl sulfate
Neuronal and glial cells require iron for DNA synthesis and mitochondrial respiration, and as a co-factor for enzymes involved in neurotransmitter synthesis and axon myelination (Chen et al. 1995; Connor and Menzies 1996). In addition, free iron promotes the generation of reactive oxygen species via Haber Weiss-Fenton reactions, thereby leading to lipid, protein and DNA damage (Halliwell and Gutteridge 1990). The maintenance of iron homeostasis is mainly regulated by the transferrin receptor that transports iron into the cell (Fishman et al. 1987), and by ferritin that sequesters this metal (Harrison and Arosio 1996). The level of ferritin, which is constituted by a heavy (H) and a light (L) subunit, is mainly regulated post-transcriptionally by interaction between the iron regulatory proteins IRP1 and IRP2, and a sequence located in the 5′ untranslated region of ferritin mRNA (iron responsive element, IRE). IRP1, the cytosolic counterpart of mitochondrial aconitase (Kennedy et al. 1992), is a bifunctional protein that, through [4Fe-4S] cluster assembly/disassembly, switches from the aconitase to the IRP1 form in response to the intracellular iron level (Guo et al. 1994). IRP2 is homologous to IRP1, lacks the [4Fe-4S] cluster, and its activity increases in iron-depleted cells by protein stabilization. In particular, iron regulates IRP1 by affecting its RNA-binding affinity and IRP2 by inducing its degradation (Guo et al. 1995; Henderson and Kühn 1995). When intracellular iron content decreases, the binding of IRPs to the IRE cis-element in ferritin mRNA is activated, thereby blocking protein translation. Conversely, when intracellular iron concentrations rise, IRP1 is no longer able to bind IRE, IRP2 is degraded and ferritin mRNA is efficiently translated (Eisenstein 2000; Theil 2000; Pantopoulos 2004). The RNA binding activity of IRPs is also regulated by oxidative stress (Pantopoulos et al. 1997), nitric oxide (Drapier et al. 1993), phosphorylation (Schalinske and Eisenstein 1996; Brown et al. 1998), hypoxia and hypoxia/reoxygenation (Hanson and Leibold 1998; Toth et al. 1999; Hanson et al. 1999; Meyron-Holtz et al. 2004).
Dysregulation of iron homeostasis coupled to oxidative stress occurs in several neurodegenerative disorders (Thompson et al. 2001; Ke and Ming Qian 2003) as well as in ischemic brain injury (Kondo et al. 1995; Dorrepaal et al. 1996). Furthermore, induction of ferritin immunoreactivity and activation of ferritin gene transcription has been shown in the brain during ischaemia/reperfusion (Ishimaru et al. 1996; Chi et al. 2000).
Given the close relationship between oxygen and iron, and the susceptibility of cerebral cells to iron-induced oxidative stress, we investigated the intracellular mechanisms that control ferritin synthesis and ferritin's role as an iron-segregating protein in the cell's defense against hypoxia/reoxygenation-induced injury. We thus examined IRP RNA-binding activity, ferritin expression and biosynthesis in cortical neurons and glial cells exposed to oxygen and glucose deprivation (OGD), and to OGD followed by reoxygenation. We demonstrate that IRP activity and ferritin biosynthesis during OGD and reoxygenation are differently regulated in these two phenotypically distinct cerebral cells. This finding may have significant implications for brain cell survival during anoxic conditions.
Materials and methods
Female Wistar rats (Charles River, Calco, Lecco, Italy) were housed in diurnal lighting conditions (12 h darkness and 12 h light), fasted overnight but allowed free access to water before the experiment. Experiments were performed according to international guidelines for animal research and the experimental protocol was approved by the Animal Care Committee of the University of Naples ‘Federico II’.
Primary cortical neuronal cultures
Cortical neurons were obtained from brains of 15–16-day-old Wistar rat embryos (Scorziello et al. 2001). Briefly, the rats were anaesthetized and then decapitated to minimize the animals' pain and distress. Dissection and dissociation were performed in Ca2+/Mg2+-free phosphate-buffered saline (PBS) containing glucose (30 mm). Tissues were incubated with papain for 10 min at 37°C and dissociated by trituration in Earle's balanced salt solution (EBSS) containing DNase, bovine serum albumin (BSA) and ovomucoid. Cells were plated at a concentration of 15 × 106 on 100 mm plastic Petri dishes pre-coated with poly-d-lysine (20 µg/mL) in Ham's F12 nutrient mixture/Eagle's minimal essential medium containing glucose (1 : 1), supplemented with 5% deactivated fetal calf serum (FCS) and 5% deactivated horse serum (HS), l-glutamine (2 mm), penicillin (50 U/mL) and streptomycin (50 µg/mL). Cytosine arabinoside (arabinoside-C) (10 µm) was added within 48 h of plating to prevent the growth of non-neuronal cells. Neurons were cultured at 37°C in a humidified 5% CO2 atmosphere with medium replenishment after 6 days, and used after 11 days of culture in all experiments.
Primary glial cell cultures
Cultures of cortical type-1 astrocytes were obtained as described elsewhere (McCarthy and de Vellis 1980). In brief, 2–3-day-old Wistar rats were decapitated, the brains removed under aseptic conditions and placed in PBS containing 100 U/mL penicillin and 100 µg/mL streptomycin. Under a stereomicroscope, cortices were dissected, the meninges carefully removed and the tissues cut into small fragments. They were then digested with trypsin at 37°C for 20 min and mechanically dissociated in Dulbecco's modified Eagle's medium (DMEM) containing 100 U/mL penicillin and 100 µg/mL streptomycin, 10% FCS and 2 mm glutamine. The medium was changed after 24 h and then twice each week. Once confluent, the cultures were shaken vigorously to remove non-adherent cells and subcultured 1 : 3. The cells were then mechanically purified and subcultured once again, 1 : 4, before the experiments were performed. This protocol produced cultures in which 95% of cells were positive to glial fibrillary acid protein.
Rat glioma cells (C6 line) were grown in DMEM containing 4.5 g/L glucose and supplemented with 10% fetal bovine serum (FBS), l-glutamine (2 mm), penicillin (100 U/mL) and streptomycin (100 µg/mL). For iron repletion-depletion experiments, cells were treated with 50 µg/mL ferric ammonium citrate (FAC) as a source of Fe ions or with 100 µm desferrioxamine (Desferal, Novartis, Origgio, Varese, Italy) for 18 h.
Combined oxygen and glucose deprivation and reoxygenation
Cortical neurons, type-1 astrocytes and C6 glioma cells were exposed to oxygen and glucose deprivation (OGD) for various times according to a previously reported protocol (Goldberg and Choi 1993). Briefly, the culture medium was replaced with deoxygenated (saturated for 20 min with 95% N2 and 5% CO2), glucose-free EBSS containing NaCl 116 mm, KCl 5.4 mm, MgSO4 0.8 mm, NaHCO3 26.2 mm, NaH2PO4 1 mm, CaCl2 1.8 mm, glycine 0.01 mm and 0.001 w/v phenol red. Cultures were then placed in an humidified 37°C incubator within an anaerobic chamber (Billups-Rothenberg, Inc., Del Mar, CA, USA) containing a gas mixture of 95% N2 and 5% CO2. The final oxygen concentration in the medium in these experimental conditions, measured by an oxygen-sensitive electrode (OxyLite 2000, Oxford Optronix, Oxford, UK), was 5 mm Hg. Reoxygenation (Reoxy) was achieved by replacing the OGD medium with oxygenated regular medium containing glucose, and returning cultures to normoxic conditions (37°C in a humidified 5% CO2 atmosphere) for various times.
Preparation of cellular extracts
After OGD and OGD/Reoxy exposure, cortical neurons, type-1 astrocytes and C6 glioma cells were washed and scraped off with PBS containing 1 mm EDTA. To obtain cytosolic extracts for electrophoretic mobility shift assay (EMSA) and ferritin western blot analysis, cells were treated with lysis buffer containing 10 mm HEPES, pH 7.5, 3 mm MgCl2, 40 mm KCl, 5% (v/v) glycerol, 1 mm dithiothreitol (DTT), 0.2% (v/v) Nonidet P-40 (NP-40) and protease inhibitor tablets (Roche, Mannheim, Germany) at 4°C. Cell debris and nuclei were pelleted by centrifugation at 15 000 g for 10 min at 4°C, and supernatant fluids were stored at −80°C. For western blot analysis of cytochrome c and caspase 3, cells were collected by scraping and low-speed centrifugation. Cell pellets were lysed at 4°C for 1 h in a buffer containing 10 mm KCl, 1.5 mm MgCl2, 20 mm HEPES, pH 7.5, 1 mm EGTA, 1 mm EDTA, 1 mm DTT, 0.1 mm phenylmethylsulfonyl fluoride and protease inhibitor tablets (Roche). The mitochondrial-containing fraction was obtained by centrifugation at 16 000 g for 25 min. Then, the supernatant fraction was centrifuged at 100 000 g for 30 min in a Beckman L8-70 ultracentrifuge to obtain mitochondria-free cytosolic extracts. The protein concentration was determined by the Bio-Rad protein assay according to the supplier's manual (Bio-Rad, Milan, Italy).
Electrophoretic mobility-shift assay (EMSA)
Plasmid pSPT-fer containing the sequence corresponding to the IRE of the H-chain of human ferritin, linearized at the Bam HI site, was transcribed in vitro as previously described (Festa et al. 2000a). For RNA-protein band-shift analysis, cytosolic extracts (5 µg) were incubated for 30 min at room temperature (20–22°C) with 0.2 ng in vitro-transcribed 32P-labelled IRE RNA. The reaction was performed in buffer containing 10 mm HEPES, pH 7.5, 3 mm MgCl2, 40 mm KCl, 5% (v/v) glycerol, 1 mm DTT and 0.07% (v/v) NP-40, in a final volume of 20 µL. To recover total IRP1 binding activity, cytosolic extracts were pre-incubated for 10 min with 2-mercaptoethanol at a 2% (v/v) final concentration, before the addition of 32P-labelled IRE RNA. IRP2 activity was differentiated by super-shift assay, incubating cytosolic extracts with a goat polyclonal anti-IRP2 antibody (Santa Cruz Biotechnology Inc, Santa Cruz, CA, USA) for 45 min at room temperature before electrophoresis. Unbound RNA was digested for 10 min with 1 U RNase T1 (Roche), and non-specific RNA–protein interactions were displaced by the addition of 5 mg/mL heparin for 10 min. RNA–protein complexes were separated on 6% non-denaturing polyacrylamide gel for 2 h at 200 V. After electrophoresis, the gel was dried and autoradiographed at −80°C. The IRP–IRE complexes were quantified with a GS-700 imaging densitometer and/or with a GS-505 molecular imager system (Bio-Rad). The results are expressed as the percentage of IRP binding activity versus 2-mercaptoethanol-treated samples.
Western blot analysis
Samples containing 50 µg of proteins were denatured, separated on a 12% (for ferritin) or 15% (for cytochrome c and caspase 3) SDS-polyacrylamide gel and electro-transferred onto a nitrocellulose membrane (Amersham Biosciences, Little Chalfont, UK) using a Bio-Rad Transblot. Proteins were visualized on the filters by reversible staining with Ponceau-S solution and destained in PBS. Membranes were blocked at 4°C in milk buffer [1 × PBS, 10% (w/v) non-fat dry milk, 0.1% (v/v) Triton X-100] and then incubated for 3 h at room temperature with 1 : 250 sheep polyclonal antibody to human ferritin (Biodesign International, Saco, ME, USA), 1 : 1000 mouse monoclonal anti-cytochrome c antibody (Biosource International, Camarillo, CA, USA) or 1 : 1500 rabbit polyclonal anti-caspase 3/CPP32 antibody (Transduction Laboratories, Lexington, KY, USA). Subsequently, the membranes were incubated for 90 min at room temperature with donkey peroxidase-conjugated anti-sheep IgG for ferritin determination (Biodesign International) and with sheep peroxidase-conjugated anti-mouse and anti-rabbit IgG for cytochrome c and caspase 3/CPP32 determination (Amersham Biosciences), respectively. The resulting complexes were visualized using chemoluminescence western blotting detection reagents (ECL, Supersignal West Pico, Pierce, Rockford, IL, USA). The optical density of the bands was determined by a GS-700 imaging densitometer (Bio-Rad). Normalization of results was ensured by incubating the nitrocellulose membrane in parallel with the β-actin antibody.
Metabolic labelling with 35S-methionine/cysteine and immunoprecipitation
Cells were pre-incubated for 30 min at 37°C in a methionine-cysteine-free medium. The medium was removed and the cells were labelled with 80 µCi/mL Promix L-[35S]-label (Amersham Biosciences) in the appropriate methionine-cysteine-free hypoxia medium for 3 h during normoxic, OGD or OGD/Reoxy conditions. Cells were washed three times with PBS and lysed in buffer containing 50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 1% (v/v) Triton X-100 and protease inhibitor tablets (Roche). Aliquots of cytosolic lysates from neurons containing 500 µg of proteins, or from C6 cells containing 850 µg of proteins, were cleared with protein G PLUS-Agarose (Santa Cruz). Then, 5 µg anti-ferritin antibody, previously conjugated with 30 µL protein G PLUS-Agarose, were added to the lysates and incubated overnight at 4°C. The protein-agarose beads were pelleted, washed three times with cold lysis buffer and boiled with SDS loading buffer. Immunoprecipitated proteins were resolved using 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). After electrophoresis, the gel was treated with Amplify (Amersham Biosciences) for 30 min, fixed, dried, and then visualized by autoradiography. Normalization of the incorporation of 35S-methionine/cysteine into total proteins was made by measuring trichloroacetic precipitable radioactive counts or protein content.
RNA extraction and northern blot analysis
After OGD and OGD/Reoxy treatments, total cellular RNA was isolated from cells by the TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) as indicated in the manufacturer's instructions. For northern blots, 10 µg total RNA were fractionated on a 1.5% agarose denaturing formaldehyde gel and then transferred to Hybond-N filters (Amersham Biosciences). The hybridization was performed for 18 h at 65°C in 0.5 m sodium phosphate buffer, pH 7.2, 1 mm EDTA, pH 8.0, 7% (w/v) SDS. The filters were washed in 0.05 m sodium phosphate buffer, pH 7.2, 1% (w/v) SDS at 65°C, and autoradiographed at −80°C. A cDNA fragment corresponding to human cDNA for H-ferritin was 32P-radiolabelled using the random priming method (TaKaRa Biomedicals, Otsu, Shiga, Japan). The β-actin probe was used to standardize the amounts of mRNA in each lane.
Cell viability assay
Cell viability was assessed by measuring the level of mitochondrial dehydrogenase activity using 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) as substrate (Hansen et al. 1989). The assay was based on the redox ability of living mitochondria to convert dissolved MTT into insoluble formazan. Briefly, after OGD and OGD/Reoxy, the medium was removed and the cells were incubated in 1 mL MTT solution (0.5 mg/mL) for 1 h in a humidified 5% CO2 incubator at 37°C. The incubation was stopped by removing the MTT solution and adding 1 mL dimethylsulfoxide (DMSO) to solubilize the formazan. The absorbance was monitored at 540 nm using a Perkin-Elmer LS 55 Luminescence Spectrometer (Perkin-Elmer Ltd, Beaconsfield, UK). The data are expressed as the percentage of cell viability to control cultures.
Lipid peroxidation assay
Lipid peroxidation products from cells were measured by the thiobarbituric acid colorimetric assay (Esterbauer and Cheeseman 1990). Briefly, after OGD and OGD/Reoxy, cells were washed and collected in PBS Ca2+/Mg2+-free medium containing 1 mm EDTA and 1.13 mm butylated hydroxytoluene (BHT). Cells were broken up by sonication. Trichloroacetic acid, 10% (w/v), was added to the cellular lysate and, after centrifugation at 1000 g for 10 min, the supernatant fluid was collected and incubated with 0.5% (w/v) thiobarbituric acid at 80°C for 30 min. After cooling, malondialdehyde (MDA) formation was recorded (A530 nm and A550 nm) in a Perkin Elmer LS-55 spectrofluorimeter. Samples were scaled for protein concentration determined by the Bio-Rad protein assay, and a standard curve of MDA was used to quantify the MDA levels formed during the experiments. The results are presented as percentage of MDA production versus a control obtained in untreated cultures.
Apoferritin treatment of cortical neurons
Primary cortical neurons cultured in Ham's F12 nutrient mixture/Eagle's minimal essential medium, as described above, were treated for 18 h with 0.3 mg/mL apoferritin from horse spleen (Sigma-Aldrich, St Louis, MO, USA), which is easily pinocytosed by the cells (Balla et al. 1992; Festa et al. 2000b). After treatment, cells were washed and the culture medium was replaced with deoxygenated glucose-free EBSS; the cells were then exposed to OGD and OGD/Reoxy. Lipid peroxidation products from cells were measured as above described. The intracellular ferritin content in the cytosolic extract was determined using a fluorimetric enzyme immunoassay system according to the supplier's manual (Enzymum test, Roche). The results are expressed as nanograms of ferritin per milligram of cell protein.
The densitometric data from EMSA, western blot, northern blot, biosynthesis bands and neuron apoferritin treatment are reported as means ± SEM. Statistical significance among the means was determined by the anova followed by the Newman–Keuls test. A p-value ≤ 0.05 was considered statistically significant.
IRP RNA-binding activity in cortical neurons and in glial cells after OGD
To determine the effects of OGD on IRP RNA-binding activity, we exposed cortical neurons from rat embryos, primary cultures of rat type-1 astrocytes and the C6 rat glioma cell line to normoxic and hypoxic conditions for 0–5 h, and then measured the IRP RNA-binding activity by RNA band-shift assay. The results are shown in Fig. 1. In cortical neurons, OGD significantly enhanced IRP1 RNA-binding activity. This effect appeared after only 30 min of OGD (22% increase vs. control), peaked after 3 h (98% of increase vs. control) and persisted up to 5 and 12 h (data not shown). To determine the total amount of IRP1 RNA-binding activity, β-mercaptoethanol was added to the binding reactions before the addition of 32P-labelled IRE. β-mercaptoethanol reveals ‘latent’ IRP1 RNA-binding activity, thus giving the total amount (100% of IRE-binding) of IRP1 activity (Hentze et al. 1989). Addition of β-mercaptoethanol to neuron extracts did not enhance further OGD-stimulated IRP1 RNA-binding activity (Fig. 1a, lanes 6–10), which suggests that OGD promoted maximal IRP1–RNA association without altering the IRP1 protein content.
Type-1 primary astrocytes responded differently to OGD, with a sudden decrease (42% decrease vs. control) in IRP1 RNA-binding activity after 30 min, and this reduction persisted up to 5 h. Addition of β-mercaptoethanol caused an increase in IRP1 RNA-binding activity in all samples (Fig. 1a, lanes 6–10), suggesting that the OGD-induced inactivation of IRP1 was not due to a variation in IRP1 protein levels, nor to modifications that irreversibly inactivated IRP1 RNA-binding.
Exposure of C6 glioma cells to OGD produced a similar rapid decrease (38% decrease vs. control) of IRP1 RNA-binding after 30 min, reaching a maximum decrease of 60% versus control at 3 h. This reduction also persisted for 5 h. Concomitant with the IRP1 RNA-binding decrease, there was a slight OGD-dependent increase in IRP2 RNA-binding activity. It should be noted that 2-mercaptoethanol treatment decreases IRP2 activity in both glial cells. This phenomenon could be explained by the sensibility of IRP2 to redox influence. In fact, routine in vitro treatment of cell extracts with 2% 2-mercaptoethanol, useful for maximally activating IRP1 binding activity, results in underestimation of IRP2 activity (Bouton et al. 1997).
To verify that the faster migrating band was indeed IRP2, we treated C6 cells for 18 h with 50 µg/mL iron salt ferric ammonium citrate (FAC) or with 100 µm of the potent iron chelator, desferrioxamine (DFX), and then exposed the cells to OGD. Under normoxic conditions, the IRP2–IRE complex was not present in either the control or the iron-treated cells but appeared in iron-depleted cells (Fig. 2, lanes 1–3). This is in accordance with iron regulation of IRP2 (Iwai et al. 1998). In cells exposed to OGD, the faster migrating band disappeared upon exposure to FAC (lane 5) and reappeared upon exposure to DFX (lane 6), which indicates that this band corresponded to the IRP2–IRE complex. These results were confirmed by super-shift assay using IRP2-specific antibody (Fig. 2, right panel).
IRP RNA-binding activity during OGD followed by reoxygenation
Reoxygenation restored the hypoxic modulation of IRP RNA-binding activity in several cell types (Tacchini et al. 1997; Hanson and Leibold 1998; Schneider and Leibold 2003). To test whether reoxygenation restored IRP modulation by OGD in cortical neurons and glial cells, cells were exposed to OGD for 3 h and reoxygenated for 3 and 24 h, after which IRP RNA-binding activity was measured. Results are depicted in Fig. 3. In cortical neurons, reoxygenation gradually restored OGD-induced IRP1 activation, and baseline levels were reached within 24 h of treatment. Contrarily, in C6 glioma cells, IRP1-RNA binding activity increased after 3 h of reoxygenation and reached control levels after 24 h. IRP2 RNA-binding became undetectable after 24 h of reoxygenation, as occurred in cells cultured in normoxic conditions. Similarly, in astrocytes, IRP1 RNA-binding activity was completely restored after 24 h of reoxygenation, although there was no significant change after 3 h of reoxygenation.
Effects of OGD and OGD/reoxygenation on ferritin expression and biosynthesis
To evaluate the combined effect of OGD and OGD followed by reoxygenation on ferritin expression in cortical neuronal and glial cells, we determined ferritin protein levels by western blot analysis after 3 h of OGD and after OGD followed by 3 and 24 h of reoxygenation. Ferritin content was unchanged in cortical neurons after 3 h of OGD and after 3 h of reoxygenation. However, there was a significant 1.6-fold increase in ferritin content in the late phase of reoxygenation (24 h) (Fig. 4). On the other hand, ferritin levels significantly increased by about two and 1.8-fold in astrocytes and C6 glioma cells, respectively, after 3 h of OGD. In C6 cells, exposure to reoxygenation for 3 h induced a decrease in ferritin content that was, however, higher than control levels. After 24 h of reoxygenation, ferritin content decreased to control levels. Similarly, in astrocytes, ferritin content decreased during reoxygenation (33% decrease vs. OGD after 24 h reoxygenation).
We evaluated ferritin biosynthesis by immunoprecipitation of 35S-labelling in cortical neurons, type-1 astrocytes and C6 cells exposed to OGD and OGD/reoxygenation (Fig. 5). In type-1 astrocytes, the rate of 35S-methionine/cysteine incorporation into ferritin was enhanced about 2.6-fold during OGD and progressively increased about 4.6-fold during the subsequent reoxygenation phases. In C6 cells, ferritin biosynthesis increased about 2.6-fold during 3 h of OGD. The subsequent 3 h of reoxygenation increased ferritin biosynthesis about fivefold and after 24 h of reoxygenation, ferritin returned to control levels. Similar experiments with cortical neurons showed a consistent increase in ferritin biosynthesis only in the reoxygenation phase, i.e. a six and 11-fold increase after 3 and 24 h, respectively.
Analysis of H-ferritin mRNA levels
To assess whether the increased ferritin level observed during OGD and OGD/reoxygenation might result from transcriptional control, we analysed the steady-state levels of H-ferritin mRNA (Fig. 6). In type-1 astrocytes, the level of H-ferritin mRNA, analysed by northern blot, was slightly decreased after the OGD period and progressively increased during the reoxygenation phases (3 and 24 h). H-ferritin mRNA content remained unchanged in C6 cells exposed to OGD and OGD/reoxygenation. By contrast, in cortical neurons, northern blot analysis showed a significant increase in H-ferritin mRNA (about 2.3-fold) only in the late phase of reoxygenation (24 h).
Effects of OGD and OGD/reoxygenation on lipid peroxidation and on survival of cortical neurons and glial cells
To investigate the effects of OGD and OGD/reoxygenation on cell survival, we evaluated cell viability by MTT assay, cell membrane injury by MDA production, and the induction of apoptotic features by cytochrome c release and activated caspase 3 expression. The results are summarized in Table 1. Exposure of cortical neurons to OGD for 3 h induced impairment of mitochondrial oxidative capacity associated with increased MDA production. When the neurons were exposed to OGD/reoxygenation, enhanced mitochondrial activity impairment was associated with a significant increase in lipid peroxidation.
|Type of rat cells||Treatment||MTT assaya%||MDA production %||Cytochrome c release||Caspase 3 cleavage (CPP32)|
|Cortical neurons||Control cells||100 ± 3.9||100 ± 7.6||No release||No cleavage|
|OGD 3 h||63 ± 4.8*||133 ± 8.7*||Release||Cleavage|
|OGD 3 h/Reoxy 24 h||53 ± 4.8**||275 ± 7.1**||Release||Cleavage|
|Type-1 astrocytes||Control cells||100 ± 10||100 ± 10||No release||No cleavage|
|OGD 3 h||100 ± 10||100 ± 10||No release||No cleavage|
|OGD 3 h/Reoxy 24 h||115 ± 10||91 ± 10||No release||No cleavage|
|C6 glioma cells||Control cells||100 ± 6.8||100 ± 14||No release||No cleavage|
|OGD 3 h||95 ± 5.8||83 ± 28.9||No release||No cleavage|
|OGD 3 h/Reoxy 24 h||108 ± 4.8||108 ± 19.6||No release||No cleavage|
Exposure of cortical neurons to OGD and to subsequent reoxygenation caused the release of cytochrome c into cytosol and induced activation of caspase 3, two hallmarks of apoptosis. Conversely, OGD and OGD/reoxygenation did not significantly modify mitochondrial oxidative capacity or MDA production in either astrocytes or glioma cells. No apoptotic features were identified in either cell type. Figure 7 shows the results of western blot analysis for cytochrome c release and caspase 3 cleavage in all cell types analysed.
Effect of exogenous ferritin addition on lipid peroxidation in cortical neurons exposed to OGD followed by reoxygenation
To establish whether the greater vulnerability observed in cortical neurons compared with glial cells was related to the delayed increase in ferritin biosynthesis occurring during the OGD/reoxygenation phase, we exposed cortical neurons to apoferritin, which is easily pinocytosed by the cell (Balla et al. 1992; Festa et al. 2000b), via receptors that have been identified in the CNS (Hulet et al. 2000). Then, we evaluated the effect of ferritin on OGD/reoxygenation-induced lipid peroxidation. To determine whether exogenous apoferritin was actually introduced into cells by simple incubation, cortical neurons were pre-treated with apoferritin (0.3 mg/mL) for 18 h and exposed to OGD and OGD/reoxygenation; then, ferritin concentration was determined in cytosolic extracts by enzyme immunoassay. The ferritin content was increased in neurons after apoferritin treatment compared with control cultures (Fig. 8). Pre-incubation of neurons with exogenous apoferritin reduced the lipid peroxidation occurring after the OGD/reoxygenation phase by 57% (Fig. 8), thus suggesting that the increased intracellular content of the iron-scavenging ferritin is indeed related to antioxidant cellular defense (Santamaria et al. 2004).
We demonstrated that oxygen-glucose deprivation followed by reoxygenation affects IRP1 activity and ferritin biosynthesis differently in cortical neurons and glial cells. IRP1 binding activity was significantly decreased in primary cultures of type-1 astrocytes and C6 glioma cells during OGD. On the contrary, OGD remarkably increased IRP1 binding activity in primary cultures of cortical neurons. The experiments with the reducing agent 2-mercaptoethanol, which maximally activates ‘latent’ IRP1 RNA-binding activity, show that the effects were not caused by a change in IRP1 protein content during OGD. The evidence that hypoxia did not affect IRP1 protein content coincides with data obtained in human embryonic kidney cells (Hanson et al. 1999). In addition, IRP2 binding activity occurred in astrocytes and, at a higher level, in C6 glioma cells exposed to OGD, but not in cortical neurons. The presence of IRP2 in glia-derived cells could be due to the differential expression of IRP2 in different cerebral cell types (Leibold et al. 2001) or to the developmental status of the cells (Siddappa et al. 2002).
Moreover, in glial cells OGD had an opposite effect on the IRPs, i.e. it decreased IRP1 binding activity and increased IRP2 binding activity. The same divergent modulation occurs in epithelial HEK293 cells (Hanson et al. 1999; Schneider and Leibold 2003) and in some mammalian tissues (Meyron-Holtz et al. 2004). The increase in IRP2 binding activity during OGD could be due to the accumulation of IRP2 protein (Hanson et al. 1999; Schneider and Leibold 2003) as a consequence of an oxygen-dependent decrease in IRP2 ubiquitination (Hanson et al. 2003). Thus, it appears that oxygen tension exerts a relevant, opposite role in the modulation of IRP1 and IRP2 binding activity, even though the possibility cannot be excluded that IRPs may themselves be subjected to as yet unidentified regulatory mechanisms induced by OGD/reoxygenation. In C6 glioma cells, intracellular iron can also contribute to IRP1 and IRP2 hypoxia-induced regulation. It is generally agreed that iron overload and iron depletion exert a regulatory function through IRP1 [4Fe-4S] cluster assembly/disassembly and IRP2 degradation/stability (Pantopoulos 2004). Interestingly, our results in glioma cells show that when the increase or decrease of iron levels inhibit or stimulate IRP1 and IRP2 activity, respectively, the subsequent reduction of oxygen tension by OGD does not further affect IRP binding activity. This is compatible with the hypothesis that iron and oxygen regulate the Fe-S aconitase/IRP1 cluster switch from the same site.
During reoxygenation, the OGD-induced changes in IRP binding activity were reversed and progressively returned to normoxic levels in neuronal and glial cells. Restoration of IRP activity by normoxic conditions agrees with reports on non-excitable cells (Tacchini et al. 1997, 2002; Hanson and Leibold 1998; Schneider and Leibold 2003). The results obtained during reoxygenation can be ascribed to restoration of oxygen level, or to production of radical oxygen species (ROS) that elicit activation of IRP1 (Hanson and Leibold 1998).
During OGD and reoxygenation, the opposite and differential modulation of IRPs occurring in neurons and glial cells is accompanied by changes in ferritin expression. In fact, in C6 glial cells, the steady-state level and synthesis of ferritin increased early during OGD, in accordance with the decrease in IRP1 RNA-binding activity. The further increase in ferritin biosynthesis observed at 3 h of reoxygenation is not reflected in a corresponding increase in the protein steady-state levels at this time, suggesting that the ferritin half-life could be decreased during the early phase of reoxygenation. The prolonged reoxygenation did not cause a further increase in biosynthesis and steady-state levels of ferritin, in accordance with the recovery of IRP1 binding activity. In these cells, the transcriptional control of ferritin expression did not appear to operate as there were no changes in H-ferritin mRNA levels; therefore, ferritin expression was regulated by the more rapid post-transcriptional mechanism operated by IRPs.
In type-1 astrocytes, transcriptional and post-transcriptional mechanisms appeared to operate in a different way during OGD and OGD/reoxygenation in order to maintain high ferritin levels. In fact, during OGD, H-ferritin mRNA levels were decreased and the increased ferritin expression was IRP-dependent. During the early reoxygenation phase, ferritin expression seemed to be dependent on a coupled transcriptional and post-transcriptional control, since H-ferritin mRNA increased and IRP1 binding activity was low. Also, in this case the possibility cannot be excluded that ferritin half-life could be decreased during this phase. During the prolonged reoxygenation, in accordance with the increased level of H-ferritin mRNA, transcriptional control seemed to be the only operating mechanism to maintain high ferritin content, since IRP1 binding activity returned to basal levels.
Conversely, in cortical neurons, ferritin expression increased only during the late reoxygenation phase when the IRP1 activity returned to basal level. However, it should be noted that after 3 h of reoxygenation, ferritin synthesis was increased whereas protein and mRNA steady-state levels were unchanged, and IRP1 binding activity, although in a declining phase, was still elevated. This apparent discrepancy between increased ferritin synthesis, sustained IRP-binding activity and no change in mRNA content could be due to the activation of protein synthesis machinery that more efficiently translates the ferritin mRNAs present in the cell. In addition, it is possible that the increase in ferritin synthesis cannot be detected by the western blot analysis because of a decreased ferritin half-life during the early reoxygenation phase. The significant increase in ferritin levels at 24 h of reoxygenation seemed to be mainly regulated at transcriptional level as northern blot analysis demonstrated a marked increase in H-ferritin mRNA at this time. This latter result is in agreement with a report showing a significant induction of ferritin gene transcription following cerebral ischaemia/reperfusion (Chi et al. 2000). The transcriptional regulation of H-ferritin observed during reoxygenation, which implicates an increase in the relative amount of this subunit, could reflect the need for an increase in iron storage, and is consistent with the concept that cells are responding to the stress by increasing H-ferritin (Theil 1987).
Since neurons and glia have a different susceptibility to oxidative stress caused by hypoxia/reoxygenation (Rosenblum 1997), the different regulation of ferritin expression may help to explain the diverse vulnerability of neurons and glial cells to OGD/reoxygenation injury. In fact, exposure of neurons to OGD/reoxygenation determined impairment of mitochondrial activity, marked lipid peroxidation and apoptosis. Under the same experimental conditions, glial cells did not show any appreciable reduction in cell viability, lipid peroxidation or apoptosis. The response of neurons can be explained by the fact that increased IRP1 binding activity during OGD determines reduced ferritin synthesis and induction of transferrin receptor expression, which could expand the intracellular labile iron pool (Tacchini et al. 1997) and enhance oxidative damage that the delayed ferritin synthesis during reoxygenation is unable to counteract. Conversely, in glial cells, the down-regulation of IRP1 binding activity during OGD could facilitate iron sequestration into ferritin and prevent the oxidative damage due to the formation of highly reactive oxygen species. Moreover, the results obtained with exogenously-added apoferritin to cortical neurons demonstrated that ferritin decreased OGD/reoxygenation-induced lipid peroxidation, thus suggesting that ferritin acts as a cytoprotective agent by limiting the ability of intracellular iron to generate ROS. Of course, the cytoprotective role exerted by exogenous ferritin is appreciable only during the reoxygenation phase, when the oxygen availability promotes iron-induced ROS production. Therefore, the earlier ferritin synthesis increase observed in glial cells may represent an adaptive response to oxidative stress generated by OGD/reoxygenation, explaining their reduced vulnerability to anoxic insults.
It is conceivable that as well as being directly regulated by IRPs, ferritin can be indirectly modulated by other factors activated during hypoxia/reoxygenation. In fact, the hypoxia-inducible factor-1 (HIF-1) that induces the expression of several genes, including those involved in iron metabolism, i.e. transferrin, transferrin receptor and heme oxygenase-1 (Semenza 2000), may contribute to the regulation of iron homeostasis. Indeed, under the oxidative stress of hypoxia/reperfusion, heme oxygenase-1 is induced and cleaves the heme porphyrin ring, releasing Fe2+ that, in turn, may induce ferritin expression.
In conclusion, our results demonstrate that IRP RNA-binding activity and ferritin expression are oppositely regulated in cortical and glial cells exposed to an experimental condition that mimics brain ischaemia. The changes observed may be viewed as an adaptive response of the cell to the iron homeostasis dysregulation induced by hypoxia/reoxygenation.
In the light of data on cross-talk between glial and neuronal cells (Bezzi et al. 2001) and our results, it is feasible that when neurons and glial cells are exposed to hypoxia/reoxygenation, glial cells by the early increase in ferritin, which is able to sequester iron, might indirectly protect neurons from the metal-induced injury.
This work was supported by grants from COFIN 2004 and Regione Campania L.R. 5/02 2003 to AC, and FIRB 2002 and COFIN 2002 to LA. We are indebted to Jean Gilder for text editing.
- 1992) Ferritin: a cytoprotective antioxidant strategem of endothelium. J. Biol. Chem. 267, 18 148–18 153. , , , , , , and (
- 2001) Neuron-astrocyte cross-talk during synaptic transmission: physiological and neuropathological implications. Prog. Brain Res. 132, 255–265. , , and (
- 1997) Redox modulation of iron regulatory proteins by peroxynitrite. J. Biol. Chem. 272, 19 969–19 975. , and (
- 1998) Novel role of phosphorylation in Fe-S cluster stability revealed by phosphomimetic mutation at Ser-138 of iron regulatory protein 1. Proc. Natl Acad. Sci. USA 95, 15 235–15 240. , , , , , and (
- 1995) Abnormal rat brain monoamine metabolism in iron deficiency anemia. J. Nutr. Biochem. 6, 486–493. , and (
- 2000) Differential regulation of H- and L-ferritin messenger RNA subunits, ferritin protein and iron following focal cerebral ischemia-reperfusion. Neuroscience 100, 475–484. , , , and (
- 1996) Relationship of iron to oligodendrocytes and myelination. Glia 17, 83–93. and (
- 1996) Nonprotein-bound iron in postasphyxial reperfusion injury of the newborn. Pediatrics 98, 883–889. , , , , and (
- 1993) Biosynthesis of nitric oxide activates iron regulatory factor in macrophages. EMBO J. 12, 3643–3649. , , , and (
- 2000) Iron regulatory proteins and the molecular control of mammalian iron metabolism. Annu. Rev. Nutr. 20, 627–662. (
- 1990) Determination of aldehydic lipid peroxidation products: malonaldehyde and 4-hydroxynonenal. Meth. Enzymol. 186, 407–421. and (
- 2000a) Oxalomalate, a competitive inhibitor of aconitase, modulates the RNA-binding activity of iron-regulatory proteins. Biochem. J. 348, 315–320. , , and (
- 2000b) Overexpression of H-ferritin and up-regulation of iron regulatory protein genes during differentiation of 3T3-L1 pre-adipocytes. J. Biol. Chem. 275, 36 708–36 712. , , , , and (
- 1987) Receptor mediated transcytosis of transferrin across the blood–brain barrier. J. Neurosci. Res. 18, 299–304. , , , and (
- 1993) Combined oxygen and glucose deprivation in cortical cell culture: calcium-dependent and calcium-independent mechanisms of neuronal injury. J. Neurosci. 13, 3510–3524. and (
- 1994) Iron regulates cytoplasmic levels of a novel iron-responsive element-binding protein without aconitase activity. J. Biol. Chem. 269, 24 252–24 260. , and (
- 1995) Iron regulates intracellular degradation of iron regulatory protein 2 by the proteasome. J. Biol. Chem. 270, 21 645–21 651. , , and (
- 1990) Role of free radicals and catalytic metal ions in human disease: an overview. Meth. Enzymol. 186, 1–85. and (
- 1989) Re-examination and further development of a precise and rapid dye method for measuring cell growth/cell killing. J. Immunol. Meth. 119, 203–210. , and (
- 1998) Regulation of iron regulatory protein 1 during hypoxia and hypoxia/reoxygenation. J. Biol. Chem. 273, 7588–7593. and (
- 1999) Hypoxia post-translationally activates iron-regulatory protein 2. J. Biol. Chem. 274, 5047–5052. , and (
- 2003) Oxygen and iron regulation of iron regulatory protein 2. J. Biol. Chem. 278, 40 337–40 342. , and (
- 1996) The ferritins: molecular properties, iron storage function and cellular regulation. Biochim. Biophys. Acta 1275, 161–203. and (
- 1995) Differential modulation of the RNA-binding protein IRP-1 and IRP-2 in response to iron. IRP-2 inactivation require translation of another protein. J. Biol. Chem. 270, 20 509–20 515. and (
- 1989) Oxidation-reduction and the molecular mechanism of a regulatory RNA–protein interaction. Science 244, 357–359. , , and (
- 2000) Oligodendrocyte progenitor cells internalize ferritin via clathrin-dependent receptor mediated endocytosis. J. Neurosci. Res. 61, 52–60. , , and (
- 1996) Activation of iron handling system within the gerbil hippocampus after cerebral ischemia. Brain Res. 726, 23–30. , , , , and (
- 1998) Iron-dependent oxidation, ubiquitination, and degradation of iron regulatory protein 2: implication for degradation of oxidized proteins. Proc. Natl Acad. Sci. USA 95, 4924–4928. , , , , , , , and (
- 2003) Iron misregulation in the brain: a primary cause of neurodegenerative disorders. Lancet Neurol. 2, 246–253. and (
- 1992) Purification and characterization of cytosolic aconitase from beef liver and its relationship to the iron-responsive element binding protein. Proc. Natl Acad. Sci. USA 89, 11 730–11 734. , , and (
- 1995) Regional differences in late-onset iron deposition, ferritin, transferrin, astrocyte proliferation and microglia activation after transient forebrain ischemia in rat brain. J. Cereb. Blood Flow Metab. 15, 216–226. , , , and (
- 2001) Immunolocalization of iron regulatory protein expression in the murine central nervous system. Histochem. Cell Biol. 115, 195–203. , and (
- 1980) Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. J. Cell Biol. 85, 890–902. and (
- 2004) Mammalian tissue oxygen levels modulate iron-regulatory protein activities in vivo. Science 306, 2087–2090. , and (
- 2004) Iron metabolism and the IRE/IRP regulatory system. An update. Ann. NY Acad. Sci. 1012, 1–13. (
- 1997) Differences in the regulation of iron regulatory protein-1 (IRP1) by extra- and intra-cellular oxidative stress. J. Biol. Chem. 272, 9802–9808. , , , , and (
- 1997) Histopathologic clues to the pathways of neuronal death following ischemia/hypoxia. J. Neurotrauma 14, 313–326. (
- 2004) Induction of ferritin expression by oxalomalate. Biochim. Biophys. Acta 1691, 151–159. , , , and (
- 1996) Phosphorylation and activation of both iron regulatory proteins 1 and 2 in HL-60 cells. J. Biol. Chem. 271, 7168–7176. and (
- 2003) Effects of iron regulatory protein regulation on iron homeostasis during hypoxia. Blood 102, 3404–3411. and (
- 2001) Differential vulnerability of cortical and cerebellar neurons in primary culture to oxygen glucose deprivation followed by reoxygenation. J. Neurosci. Res. 63, 20–26. , , , , , , , , and (
- 2000) Surviving ischemia: adaptive responses mediated by hypoxia-inducible factor 1. J. Clin. Invest. 106, 809–812. (
- 2002) Developmental changes in the expression of iron regulatory proteins and iron transport proteins in the perinatal rat brain. J. Neurosci. Res. 68, 761–775. , , , , and (
- 1997) Induction of ferritin synthesis in ischemic-reperfused rat liver: analysis of the molecular mechanisms. Gastroenterology 113, 946–953. , , and (
- 2002) Transferrin receptor gene expression and transferrin-bound iron uptake are increased during postischemic rat liver reperfusion. Hepatology 36, 103–111. , , and (
- 1987) Ferritin: structure, gene regulation, and cellular function in animals, plants, and microorganisms. Annu. Rev. Biochem. 56, 289–315. (
- 2000) Targeting mRNA to regulate iron and oxygen metabolism. Biochem. Pharmacol. 59, 87–93. (
- 2001) Iron and neurodegenerative disorders. Brain Res. Bull. 55, 155–164. , and (
- 1999) Hypoxia alters iron-regulatory protein-1 binding capacity and modulates cellular iron homeostasis in human hepatoma and erythroleukemia cells. J. Biol. Chem. 274, 4467–4473. , , , and (