Background
Astrocytes represent the most abundant cell type in the central nervous system (CNS), and they are key players for both physiological neuronal functions and pathological processes. They have long been recognized to provide energetic, metabolic and structural support to neurons through the secretion of nutrients and the maintenance of the blood–brain barrier integrity [
1]. More recently, they have been shown to exert a critical role in the extracellular milieu homeostasis by controlling the concentration of ions, mediators and neurotransmitters, thereby contributing to bidirectional neuron-glia communication [
2].
Astrocytes are capable of undergoing rapid changes in their phenotype as a result of alterations of CNS homeostasis. This process, known as “astrocyte activation” or astrogliosis, is a well-established feature of several pathological situations [
3], in particular within neuroinflammatory processes, in which an intense glial fibrillary acidic protein staining can be observed [
4,
5]. The way astrocyte activation contributes to an either neurodegenerative or neuroprotective role is still an open question [
6].
In this scenario, one important role that astrocytes might play is related to oxidative stress, a condition that has been ascribed to the reduction of antioxidant defenses and/or to the increase in reactive oxygen species (ROS) production [
7]. Indeed, neurons are considered particularly vulnerable to oxidative stress because of their high oxygen rate [
8], and oxidative stress has been associated with neurodegeneration during aging, as well as in several disorders, including Parkinson’s and Alzheimer’s diseases [
9]. Glial cell activation might also occur in response to oxidative stress and involves a complex range of responses that can either accelerate or delay the neurotoxic effects [
10]. Within this framework, growing attention has been paid to the iron content, since it is known to favor free radical generation by reacting with the H
2O
2 physiologically produced by the cellular metabolism [
11,
12].
Several therapeutic strategies addressing oxidative stress toxicity (e.g., the exogenous antioxidant supply) and iron accumulation (e.g., metal chelation therapy) have been attempted, and some proved to be effective (see, for instance, [
9,
11,
12]). While molecules with a direct protective effect on neurons have been searched for, less attention has been paid to astrocytes, even though they are an important alternative target, being a cellular link between oxidative stress toxicity and neuroinflammation [
13], as well as key players in non-cell autonomous mechanisms of neurodegeneration [
14]. Moreover, astrocytic activity is reported to strongly condition iron homeostasis in the CNS [
15]. In fact, they control iron flow through the blood–brain barrier and iron distribution to neurons. Since most of the extracellular iron in the CNS is not bound to transferrin, astrocytes also play an important role in buffering local iron changes, thereby preventing oxidative stress in neurons (see [
16] and references therein).
In order to investigate the molecular basis and the effects of the astrocytic activation process under conditions of oxidative stress, proper cellular models and reliable in vitro assays are required. Many studies exploit the use of continuous cell lines that, although easier to handle, substantially differ from primary cultures in terms of resistance to oxidative stress [
17]. Primary glial cultures are a better option, although proper care must be taken in order to reproduce the process of activation that occurs in vivo. In fact, the presence of a heterogeneous mixture of cell types in the culture may strongly affect the outcome of the experiments, especially as regards the contribution of microglia to inflammatory conditions [
18,
19]. Secondary cultures of pure astrocytes are often required to keep experimental variables under control [
18].
In this work, we set up a cellular assay in order to follow iron-induced cell stress in highly standardized pure secondary cultures of astrocytes obtained from rat cortex. From this standpoint, we investigated the molecular aspects that determine the resistance of astrocytes to oxidative stress with particular regards to the effects of proinflammatory activation. We also performed a transcriptomic analysis in order to identify the candidate genes involved in the mechanisms of protection.
Methods
Materials
Cell culture media and reagents were from BioWhittaker-Lonza (Basel, Switzerland). Other chemicals, if not otherwise stated, were from Sigma-Aldrich (St Louis, MO, USA). Culture flasks and multiwell plates were from Nalge Nunc (Rochester, NY, USA). Petri dishes were from Falcon BD (Franklin Lakes, NJ, USA).
Cell culture
The animal use procedures were approved by the Institutional Animal Care and Use Committee of the San Raffaele Scientific Institute. Primary cultures of cortical astrocytes were prepared from 2- to 3-day-old Sprague–Dawley rats (Charles River Italia, Calco, LC, Italy) according to [
19]. Briefly, after dissection, cortices were cut into small sections with a razor blade. The pieces were collected and washed twice in Hank’s Balanced Salt Solution supplemented with 10 mM Hepes/Na pH 7.4, 12 mM MgSO
4, 50 U/ml penicillin and 50 μg/ml streptomycin (Gibco, Grand Island, NY, USA). The tissue was then incubated, in two subsequent steps, with 2.5 mg/ml trypsin type IX in the presence of 1 mg/ml deoxyribonuclease (Calbiochem, La Jolla, CA, USA) for 10 min at 37°C and mechanically dissociated. The supernatant obtained was diluted 1:1 in medium containing 10% donor horse serum (PAA Laboratories GmbH, Pasching, Austria). After centrifugation (100 g for 10 min), cells were plated in Minimum Essential Eagle Medium supplemented with 10% donor horse serum, 33 mM glucose, 2 mM glutamax (Gibco), 50 U/ml penicillin and 50 μg/ml streptomycin. Cells were maintained in 75 cm
2 flasks (about 1 per pup) at 37°C in a 5% CO
2 humidified incubator. In order to remove microglia and oligodendrocyte progenitors and obtain pure secondary cultures of astrocytes (>99.8%), flasks were shaken at 200 rpm for 24 h at 37°C at days 2 and 6 after dissection in Minimum Essential Medium with Hank's salts supplemented with 10% donor horse serum, 33 mM glucose, 2 mM glutamax and 10 mM Hepes/Na, pH 7.4. Contamination with non-astrocytic cells was assessed by immunocytochemistry (GFAP for astrocytes, IBA1 for microglia and A2B5 for oligodendrocyte precursors) according to [
20]. After reaching confluence, astrocytes were trypsinized and replated in Minimum Essential Medium Eagle supplemented with 10% donor horse serum (BioWhittaker-Lonza or, in comparative experiments, Invitrogen, Carlsbad, CA, USA), 33 mM glucose, 2 mM glutamax, 50 U/ml penicillin and 50 μg/ml streptomycin onto polylysine-coated glass coverslips or plastic multiwells. Cultures were maintained at 37°C in a 5% CO
2 humidified incubator, and experiments were performed within 3 days after re-plating.
Dye loading
Dye loading was performed in Krebs-Ringer-Hepes buffer (5 mM KCl, 125 mM NaCl, 2 mM CaCl2, 1.2 mM MgSO4, 1.2 mM KH2PO4, 6 mM glucose and 20 mM Hepes, pH 7.4). The fluorescent dyes (from Molecular Probes, Invitrogen, when not specified) were administered as follows: (1) fura-2 acetoxymethyl ester (Calbiochem), 4 μM, 40 min at 37°C; (2) Sytox Blue, 5 μM, kept in the bath during the experiments; (3) 5-(and-6)-chloromethyl-2’,7’-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H2DCFDA), 0.25 μM, 30 min at 37°C; (4) tetramethyl rhodamine methyl ester (TMRM), 25 nM, 15 min at room temperature and maintained in the bath during the experiments. After dye loading, cells were washed once with fresh Krebs-Ringer-Hepes and analyzed in the same buffer. Both single cell and monolayer culture experiments were performed at room temperature.
Videomicroscopy setup
The videoimaging setup for single cell studies is based on an Axioskope 2 microscope (Zeiss, Oberkochen, Germany) and a Polychrome IV (Till Photonics GmbH, Martinsried, Germany) light source. In the epifluorescence path, fura-2 was excited at 355 nm to monitor Fe2+ variations (as quenching of the fluorescence signal). The excitation wavelength of 355 nm was adopted instead of the theoretical 360-nm isosbestic wavelength because it turned out to be insensitive to Ca2+ variations in our optical configuration. Fluorescence images were collected by a cooled CCD videocamera (PCO Computer Optics GmbH, Kelheim, Germany). The Vision software (Till Photonics) was used to control the acquisition protocol and to perform data analysis.
Fluorescence plate reader measurements
The measurements of the fluorescence intensity for pure astrocytic monolayers were based on a plate reader (Mithras LB 940, Berthold Technologies or Wallac Victor3TM 1420 multilabel counter, Perkin Elmer) provided with a flash lamp and sensitive PMTs coupled with high transmission band pass filters. Briefly, treated or untreated astrocytes plated on 24-well plates were loaded with fura-2 or CM-H2DCFDA, as previously described. Fluorescence was measured every 60 s up to 90 min.
Cellular treatments
Fe2+ was prepared as a solution of ferrous-ammonium sulfate, freshly dissolved in water and kept in ice until use. Acute iron overload was performed by incubating fura2-loaded astrocytes with 20 μM pyrithione, an iron ionophore, and Fe2+ 1 μM for 3 min to allow iron entry (monitored by fura2 quenching). Cells were then washed twice with Krebs-Ringer-Hepes solution to remove extracellular Fe2+, and the fluorescence signal was monitored over time.
In some experiments, a pre-treatment with drugs or pharmacologic agents was needed: (1) L-buthionine-sulfoximine (BSO), an inhibitor of glutathione (GSH) synthesis, 1 mM was added to the cellular medium 24 h before the experiment; (2) Mito-TEMPO [(2-(2,2,6,6-tetramethylpiperidin-1-oxyl-4-ylamino)-2-oxoethyl) triphenylphosphonium chloride monohydrate; ENZO Life Science, Farmingdale, NY, USA] 200 μM, as well as Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) 1 mM were added to cells 2 h before the experiment; (3) the inhibitor of GSH reductase, 2AAPA (R,R’-2-acetylamino-3-[4-(2-acetylamino-2-carboxyethyl-sulfanylthiocarbonylamino)-phenylthiocarbamoylsulfanyl] propionic acid hydrate S,S’-[1,4-Phenylenebis(iminocarbonothioyl)]bis[N-acetyl-L-cysteine], 25 μM, was administered at the beginning of the experiment (after acute iron overload) and left until the end.
Astrocyte activation was induced by treating cells with recombinant rat interleukin-1β (IL-1β) and tumor necrosis factor α (TNFα) from R&D Systems (Minneapolis, MN, USA). Stimuli were administered directly to the culture medium 24 h before the experiment as follows: 10 ng/ml IL-1β and 30 ng/ml TNFα. Lipopolysaccharide (LPS) (10 ng/ml) was added directly to the astrocyte culture medium 24 h before the experiment and used as a negative control for activation. Conditioned medium obtained from resting or activated (with 10 ng/ml LPS) microglia was incubated on astrocytes in substitution for their culture medium 24 h before the experiment.
GSH measurement
Reduced glutathione in pure astrocytic cultures was measured by using QuantiChromTM Glutathione Assay Kit (DIGIT-250) from BioAssay Systems (Hayward, CA, USA). Treated or untreated astrocytes were washed and then lysed for 15 min at 4°C with lysis buffer (150 μl for two 35 × 10 petri dishes) containing phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM sodium phosphate dibasic, 2 mM potassium phosphate monobasic, pH 7.4) supplemented with 2% Nonidet P-40, 10 mM EDTA/Na and a cocktail of protease inhibitors (chymostatin, leupeptin, antipain and pepstatin dissolved 1000x in dimethylsulfoxide and used 10 μg/ml each). Lysates were centrifuged for 15 min at 15,000 g at 4°C, and 120 μl of the supernatant was mixed with 120 μl of provided Reagent A (for deproteination and detection). After vortexing, the sample/reagent A mixture was centrifuged for 5 min at 15,000 g. Eventually, 200 μl of the supernatant was added with 100 μl of provided reagent B, i.e., 5,5’-dithiobis-2-nitrobenzoic acid, for colorimetric reaction and transferred into the wells of a 96-well plate. After an incubation of 25 min at room temperature, absorbance of the samples was read at 412 nm. To confirm the obtained data, we also employed a Glutathione Fluorometric Assay Kit from Biovision (Milpitas, CA, USA) according to the manufacturer’s instructions.
Cell viability assay
Cell viability was quantified by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Briefly, after treatment, pure cultures of astrocytes plated on 24-well plates were incubated for 1 h at 37°C with 0.5 mg/ml MTT in culture medium. After removing the extracellular solution, formazan, the MTT metabolic product, was dissolved in dimethylsulfoxide, and the absorbance was read at 570 nm.
DNA and siRNA transfection
DNA and siRNA transfections were performed using Lipofectamine 2000 reagent (Life Technologies). Briefly, DNA or siRNA (1.8 μg/ml and 200 pmol/ml, respectively), as well as lipofectamine 2000 (4 μl/ml), was separately incubated with Optimem Reduced Serum Media (Gibco) for 5 min at room temperature. Subsequently, the Lipofectamine mix was added to DNA or siRNA mix and left for 30 min at room temperature. Astrocytes plated on 35 × 10-mm petri dishes or 24 multiwell plates were washed once with Optimem and incubated with the transfection solution for 90 min at 37°C. After two washes with Optimem, astrocyte-conditioned medium was eventually put on the cells. Twenty-four hours of expression was required before performing the experiments. SiRNAs used (forward and reverse) were: (1) cgaccuacgugaacaaucutt and agauuguucacguaggucgcg for the gene encoding mitochondrial superoxide dismutase (SOD2); (2) agcuaguguagaaauaauatt and uauuauuucuacacuagcutt plus cgaguuacagugucuuaautt and auuaagacacuguaacucggg for the nuclear factor (erythroid-derived 2)-like 2 gene (NFE2L2) encoding Nrf2.
Western blotting
Cells were washed twice with phosphate-buffered saline and lysed for 15 min at 4°C with 100 μl/petri dish (35 × 10 mm) of lysis buffer (phosphate-buffered saline supplemented with 2% Nonidet P-40, 0.2% sodium dodecyl sulfate, 10 mM EDTA/Na and the cocktail of protease inhibitors).
Lysates were centrifuged for 15 min at 15,000 g at 4°C, and the supernatants were collected and their total protein content analyzed by the MicroBCA reagent (ThermoFisher Scientific, Pierce, Waltham, MA, USA). About 20–40 μg of proteins was separated by standard sodium dodecyl sulfate polyacrylamide gel electrophoresis and then electrically transferred onto a nitrocellulose membrane in blotting buffer [2.5 mM tris(hydroxymethyl)aminomethane, 19.2 mM glycine, 20% methanol]. The nitrocellulose filter was stained with Ponceau S (0.2% in 3% trichloroacetic acid) and de-stained with double-distilled water for protein visualization.
Typically, membranes were blocked in washing solution (10 mM tris(hydroxymethyl)aminomethane/HCl, 150 mM NaCl, 0.1% Tween-20, pH 7.6) containing 5% non-fat dry milk overnight at 4°C. Primary antibodies were diluted in blocking solution and incubated for 2 h at room temperature. Subsequently, membranes were washed three times for 5 min with washing solution and incubated 1 h with goat anti-rabbit or anti-mouse horseradish peroxidase-conjugated secondary antibody (ThermoFisher Scientific, Pierce) diluted in blocking solution at room temperature. After three washing steps, protein bands were detected on auto-radiographic films (GE Healthcare, Piscataway, NJ, USA) by incubation with chemiluminescent solutions (Pico or Super Signal West Femto chemiluminescent kit, ThermoFisher Scientific, Pierce). The following primary antibodies were used: (1) rabbit polyclonal anti-SOD2 antibody from Merck Millipore (Billerica, MA, USA); (2) mouse monoclonal anti-catalase antibody from Sigma Aldrich; (3) rabbit polyclonal anti-peroxiredoxin I, III and V antibodies from AbFrontier (Korea); (4) goat polyclonal anti-peroxiredoxin VI antibody from R&D Systems; (5) rabbit polyclonal anti-SOD1 antibody from Santa Cruz Biotechnology (Santa Cruz, CA, USA).
RNA was extracted from treated or untreated cells plated on 35 × 10-mm Petri dishes with TRIzol (Invitrogen) and phenol/chlorophorm/isoamyl alcohol (25:24:1 v/v), following the manufacturer’s instruction. Briefly, cells were lysed in 1 ml of TRIzol, to which 200 μl of phenol/chlorophorm/isoamyl alcohol was added. After centrifugation (12,000 g, 15 min.), the upper aqueous phase was transferred in a new tube, and RNA was precipitated through addition of an equivalent amount of isopropanol. Samples were centrifuged (12,000 g, 10 min) and washed with 70% ethanol. RNA pellets were air-dried for 5 min, resuspended in 20 μl of RNase-free water and stored at −80°C.
Reverse transcription (RT) was carried out with random hexamers as primers, using the Superscript III Retrotranscription Kit (Invitrogen) following the manufacturer’s instructions. RT was performed at 50°C for 50 min, then incubating samples were stopped at 85°C for 5 min. Single-strand cDNA was obtained by digesting complementary RNA strands with provided RNase H for 20 min at 37°C.
Quantitative polymerase chain reaction (qPCR) was performed on a LightCycler 480 machine (Roche Diagnostics, Basel, Switzerland), with proprietary SybrGreen mix (LightCycler 480 Master Mix, Roche), following the manufacturer’s instructions. Both forward and reverse primers were used at a 0.5 μM concentration. RT-derived cDNA was typically diluted 1:4 before use. A PCR program was performed with 10 min of the denaturation step at 95°C and 35 to 45 cycles of amplification. Each cycle consisted of a denaturation step (95°C, 10 s), annealing step (60°C, 25 s) and elongation step (72°C, 15 s). After amplification, a melting step was performed (95°C for 30 s, 60°C for 1 min). Determination of crossing points and melting peaks was performed with LightCycler 480 Software (version 1.5.0.39, Roche). Primers used (forward and reverse) were: (1) gacctacgtgaacaatctgaacg and cttgatagcctccagcaactct for the gene encoding SOD2; (2) gcagagacattcccatttgtagat and cttaaatcagtcatggccgtct for the gene encoding Nrf2; (3) tcaccattaagctgggcg and ttcttcccggtccagtcata for the gene encoding frataxin; (4) gtatgaacagcgatgatgcact and gaagaccagagcagattttcaatag for the gene encoding interleukin-6 (used as a positive control for activation) and (5) gaagaagaaattagagaagcgttcc and gtagtttacctgaccatccccat for calmodulin 2 (used as an internal reference for normalization).
Transcriptomic analysis
After the treatments, astrocytes were subjected to RNA extraction using Trizol (Invitrogen), according to the manufacturer’s instructions. The gene expression profiling was determined using RatRef-12 Expression Beadchips (Illumina Inc., San Diego, CA, USA). Each beadchip investigates for more than 21,000 transcripts selected primarily from the NCBI RefSeq database (Release 16) and in a minor part from the UniGene database. An amount of 500 ng of total RNA was reverse transcribed into complementary RNA and biotin-UTP labeled using the Illumina TotalPrep RNA Amplification Kit (Applied Biosystems) according to the manufacturer’s protocol; 750 ng of complementary RNA was then hybridized to the BeadChip array and stained with streptavidin-Cy3. All procedures were performed following the manufacturer’s instructions. BeadChips were imaged using the Illumina BeadArray Reader, a two-channel 0.8-μm-resolution confocal laser scanner and Illumina BeadScan software. Illumina GenomeStudio v.2011.1 software was used to elaborate the fluorescence signal to a value whose intensity corresponded to the quantity of the respective transcript in the original sample. The same software was used to assess the quality controls, which included the biological specimen controls, hybridization controls, signal generation controls and negative controls. The samples belonging to three different experimental conditions (i.e., resting astrocytes, astrocytes stimulated with cytokines and astrocytes stimulated with microglia-conditioned medium) were tested in technical duplicates, and a scatter plot with a correlation coefficient for each couple of replicates was calculated, leading to a mean value of 0.99. The fold change based on raw data was calculated for the comparisons.
Data analysis was performed using the Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems Inc., Redwood City, CA, USA). lllumina probes, filtered for p < 0.05 and a threshold of 1.5 points of fold change between different conditions (in upregulation or downregulation) were uploaded into the software together with their differential expression p-values measured by t-test. Each probe was mapped to its own gene object in the Ingenuity Pathways Knowledge Database.
To interpret the gene expression results of the different condition of stimulation in the context of biological processes, pathways and networks, IPA Core Analyses were conducted. Networks of these genes were assigned a score based on their connectivity: the score reflected the number of focus genes in the network and how relevant this network is to the original list of focus genes. The significance of the association between the data set of genes and the canonical pathways contained in the Ingenuity Pathways Knowledge Database was determined by adjusted p-values using the Benjamini-Hochberg correction for multiple testing, and a p-value < 0.05 after correction was considered statistically significant according to functional enrichment analysis. Besides, the “overlay” function was used to give a graphic visualization of which subset of genes inside a given canonical pathway was modulated in different conditions.
Data analysis
Data are presented as mean ± SEM of at least three independent experiments. Statistical significance was tested by using either paired or unpaired two-tailed Student’s t-test or one-way ANOVA followed by the Bonferroni (for all pairwise comparisons) post hoc test. Statistical analysis was performed using GraphPad Prism (GraphPad Software, San Diego, CA, USA).
Discussion
In this work, we evaluate the capability of astrocytes to cope with the oxidative stress induced by moderate and acute intracellular Fe2+ accumulation promoted by an iron ionophore. The reason behind this is that iron dysregulation and accumulation are common features of both ageing and neurodegenerative disorders and that Fe2+ favors the production of the highly toxic hydroxyl radicals by reacting with the H2O2 (within the framework of the self-regenerative Fenton reaction) that is physiologically produced during cellular metabolism. Under our experimental conditions, astrocytes appeared to stand the iron overload for variable periods of time, most likely up to exhaustion of their antioxidant potential. Afterward, an acute intracellular burst of oxidative stress was observed, and this is expected to reflect the reaction of Fe2+ with the intracellular H2O2, with consequent lipid peroxidation, loss of membrane integrity and, as a final event, cell death. The whole process occurred within minutes and appeared to depend not only on the cellular state, but also on the culture conditions. In fact, the presence of antioxidants in the culture medium was able to delay the entire process, including cell death.
Based on the work of Röhl and colleagues [
23], in which astrocytes pretreated with conditioned medium from LPS-activated microglia (MCM[+]) and were found to be resistant to extracellular H
2O
2 administration, we considered the possibility that activated astrocytes might also be protected against the oxidative stress mediated by intracellular iron loading. In fact, the recently reported increase in iron import observed in activated astrocytes [
28] could trigger oxidative stress if specific protective mechanisms are not induced.
The conversion of astrocytes from the resting to reactive state is well described in animal models of acute CNS injury and neurodegeneration [
29]. However, under in vivo conditions, it is difficult to identify the causative mechanisms of the astrogliosis process, since they result from the complex integration of the responses of the various cell types present in the CNS. We reproduced the process of reactive astrogliosis by exposing astrocytes to IL-1β and TNFα, the two main cytokines released from activated microglia [
30], as well as to MCM[+], a condition that is expected to better mimic the in vivo situation since LPS induces the release of a wide spectrum of proinflammatory molecules from microglia [
31]. From our data, it is clear that TNFα and IL-1β are able to reproduce, in broad outline, the more complex scenario that was supposed to result from the mix of factors released by LPS-activated microglia, and, in both models, activated astrocytes acquired a protective competence not only towards H
2O
2, but also iron overload, with prevention of both oxidative stress and cell death. More intriguing was the analysis of the determinants of this protective phenotype. We employed a transcriptomic approach with the aim to investigate the changes in a wide panel of genes involved in the oxidative stress response. First of all, the transcriptomic analysis rules out a possible reinforcement of the control of iron detoxification, since the analysis of the transcripts did not show an increased expression of proteins involved in iron extrusion and storage. The only clear evidence from both the biochemical and transcriptomic analysis was a great increase in SOD2 expression upon activation. This draws attention to the mitochondrial enzymes that, in astrocytes exposed to inflammatory cytokines, are expected to have a pivotal role. Indeed, the dysfunction of these organelles is generally considered one of the primary causes of cell death in neurodegenerative diseases [
32], a view that finds confirmation in the high sensitivity of mitochondria to ROS injury. Many antioxidant therapies have been proposed to protect mitochondria, and thus to prevent neuronal damage; however, this therapeutic strategy has turned out to be too simplistic, given the complexity of mitochondrial ROS metabolism. In fact, the process of astrocyte activation is expected to increase the mitochondrial resistance against oxidative stress by acting on a multiplicity of targets, for instance, by controlling the synthesis of SOD2, mitochondrial peroxiredoxins and glutathione. Within this complex framework, however, it is difficult to put all of them into perspective. If we consider the role SOD2 expression can play in the presence of iron overload, we can infer that removal of superoxide from mitochondria limits the self-regeneration (Haber-Weiss reaction) of the substrate for the Fenton reaction, as this molecule is able to reduce Fe
3+ to the reactive Fe
2+. However, superoxide dismutation also causes an elevation of H
2O
2, a condition that is extremely harmful in the presence of Fe
2+ since it leads to the production of the hydroxyl radicals. Activated astrocytes appear to better deal with this problem, since they contain higher levels of not only SOD2, but also reduced glutathione with respect to resting astrocytes. Therefore, activated astrocytes appear to neutralize the harmful potential of iron by detoxifying mitochondria from superoxide as well as H
2O
2. Within this general view, the involvement of SOD2 appears to be less critical than that of glutathione. In fact, silencing of SOD2 did not revert the protective phenotype, while changes in the level of glutathione significantly affected the protective competence of astrocytes, suggesting that H
2O
2 detoxification is more critical than the disposal of excess superoxide.
Astrocytes are known to respond to oxidative stress with changes in gene expression, and the transcription factor Nrf2 was reported to play a central role in raising astrocytic antioxidant defenses [
33]. In particular, a new mode of action of natural antioxidant compounds has recently emerged in which the induction of the Nrf2 pathway promotes the transcription of downstream genes involved in the protection against both oxidative stress—among others, glutathione—and inflammation [
34]. However, when we exposed astrocytes to proinflammatory conditions, the Nrf2 pathway did not appear to be activated. It should be noted in this respect that the Nrf2 response against oxidative stress is activated mainly by modulation of the interaction of Nrf2 with Keap1, i.e., the repressor of its translocation to the nucleus, and not by the simple upregulation of the Nrf2 gene [
35]. Overall, although activated astrocytes are reported to show a moderate upregulation of Nrf2 expression, the expression of the Nrf2 target genes was almost unaffected in our experimental conditions, leading us to conclude that the Nrf2 pathway does not contribute to the protective phenotype by increasing glutathione levels. In any event, it is clear that even if the glutathione level is important against oxidative stress, it cannot fully account for the differences observed between quiescent and activated astrocytes. In fact, even when depleted of this antioxidant molecule, activated astrocytes were more resistant to iron-mediated toxicity and cell death, thereby suggesting that other molecules potentiate the defense system. In addition to glutathione peroxidases and peroxiredoxins, recognized as key enzymes for H
2O
2 detoxification in astrocytes [
36,
37], our transcriptomic approach revealed new potential candidates for oxidative stress control. Of particular interest are TXNRD1, involved in the peroxiredoxin-mediated detoxification of H
2O
2 via regeneration of thioredoxins [
38], and EPHX2, a hydrolase that might help to detoxify reactive epoxydes formed during oxidative stress [
39]. Moreover, a downregulation of enzymes that produce H
2O
2, such as the aldehyde oxidase 1, which was identified in our analysis, might also contribute to decreasing the oxidative load of astrocytes. Further work is required to better define the role of these genes, and, in particular, it must be taken into account that the resistance of activated astrocytes might rely not only on changes in the expression of specific mRNAs, but also on changes in the activity of enzymes, an issue that could not be addressed by a simple transcriptomic analysis.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
RM, IP, AC and IV prepared the cells and performed the experiments. GG and FMB performed the Illumina transcriptomic analysis. RM, FC and DZ analyzed the data. RM, FC, FG and DZ conceived and designed the experimental plan and wrote the manuscript. All authors have read and approved the final version of the manuscript.