Astrocyte Resilience to Oxidative Stress Induced by Insulin-like Growth Factor I (IGF-I) Involves Preserved AKT (Protein Kinase B) Activity*

Disruption of insulin-like growth factor I (IGF-I) signaling is a key step in the development of cancer or neurodegeneration. For example, interference of the prosurvival IGF-I/AKT/FOXO3 pathway by redox activation of the stress kinases p38 and JNK is instrumental in neuronal death by oxidative stress. However, in astrocytes, IGF-I retains its protective action against oxidative stress. The molecular mechanisms underlying this cell-specific protection remain obscure but may be relevant to unveil new ways to combat IGF-I/insulin resistance. Here, we describe that, in astrocytes exposed to oxidative stress by hydrogen peroxide (H2O2), p38 activation did not inhibit AKT (protein kinase B) activation by IGF-I, which is in contrast to our previous observations in neurons. Rather, stimulation of AKT by IGF-I was significantly higher and more sustained in astrocytes than in neurons either under normal or oxidative conditions. This may be explained by phosphorylation of the phosphatase PTEN at the plasma membrane in response to IGF-I, inducing its cytosolic translocation and preserving in this way AKT activity. Stimulation of AKT by IGF-I, mimicked also by a constitutively active AKT mutant, reduced oxidative stress levels and cell death in H2O2-exposed astrocytes, boosting their neuroprotective action in co-cultured neurons. These results indicate that armoring of AKT activation by IGF-I is crucial to preserve its cytoprotective effect in astrocytes and may form part of the brain defense mechanism against oxidative stress injury.

IGF-I 2 is a pleiotropic growth factor with important prosurvival effects in neurons (1). One of the main downstream targets of IGF-I is the Ser/Thr kinase AKT (2), which mediates cell survival and proliferation (3). Upon its activation, the IGF-I receptor recruits and phosphorylates IRS docking proteins (4), which allows translocation of phosphatidylinositol 3-kinase (PI3K) to the plasma membrane where it catalyzes the formation of the lipid second messenger phosphatidylinositol 3,4,5trisphosphate (PIP 3 ) (5). AKT is recruited to the membrane by interaction with these messengers so that it can be fully activated by PDK1 and mTORC2 kinases (3,6). This pathway is switched off (to prevent uncontrolled proliferation) through activation of the lipid and protein phosphatase PTEN, which catalyzes PIP 3 dephosphorylation (4).
Cell metabolism generates potentially harmful reactive oxygen species (ROS), which at moderate levels can act as second messengers (7). However, chronic and/or abrupt increases of ROS (uncontainable by detoxification through cellular defenses) generates oxidative stress, a pathological cellular condition that can interfere with redox-sensitive signaling pathways (7). Neurons are particularly vulnerable to oxidative stress because of their low ROS detoxifying capacity (8).
We previously described that oxidative stress interferes with the IGF-I/PI3K/AKT pathway by redox activation of p38 kinase to induce neuronal death (9). IGF-I signaling impairment by oxidative stress has been recently confirmed by others in neurons (10,11) and others cell types (12). However, a substantial body of evidence in vertebrates and invertebrates also indicates that IGF-I signaling may reduce cell defenses to oxidative stress (13)(14)(15) that, in turn, would affect neuron survival. These results question the neuroprotective role of IGF-I in response to brain oxidative insults. However, IGF-I has shown antioxidative activity in the majority of cell types present in the mammalian brain (16 -18) and clinical benefits in brain pathologies associated with oxidative stress (17,19,20). Furthermore, our group has recently showed that IGF-I cooperates with other trophic signals produced by astrocytes, essential contributors to neuronal homeostasis (21), to protect neurons against oxidative insults (16). Collectively, these results suggest that the antioxidative functions of IGF-I could be cell-and context-dependent and could play a neuroprotective role during brain oxidative insults (16).
In the present work, we describe two molecular adaptations present in astrocytes that preserve the activation of AKT by IGF-I during oxidative stress. These adaptations include 1) insensitivity of the IGF-I/PI3K/AKT pathway to This article has been withdrawn by the authors. The pAKT and AKT immunoblots in modulation by the kinase p38 and 2) phosphorylation of PTEN by IGF-I, which leads to its cytosolic translocation. Armoring of AKT activation by IGF-I in astrocytes contributes to normalize ROS levels and to prevent cell death during oxidative stress. Of note, the neuroprotective role of astrocytes is also enhanced by these adaptations. These results point out the importance of AKT activation for astrocyte survival during oxidative stress and reinforce the idea that modulation of astrocytes by IGF-I forms part of the brain responses to oxidative damage.
Astrocyte Culture and Transfections-Astrocyte cultures were prepared from postnatal day 3 rats as described (23). Cells were grown on Dulbecco's modified Eagle's medium/F-12 (DMEM/F-12) supplemented with 10% fetal calf serum. After 12 days, astrocytes were seeded at 2.5 ϫ 10 5 or 1.25 ϫ 10 5 cells/well in 6-and 12-well culture plates, respectively. The day of the experiment cells were starved for 3 h and treated with IGF-I (10 Ϫ7 M) and/or H 2 O 2 at doses of 50 -150 M. Inhibitory drugs were given 45 min before treatments. We used H 2 O 2 as an oxidant stimulus because it is an endogenously produced ROS that serves as a precursor to hydroxyl radicals and possesses redox signaling capacities (24). For transfection, astrocytes were seeded at 2.5 ϫ 10 5 or 1.25 ϫ 10 5 cells/well in 6-and 12-well culture plates, respectively, and after 16 h, constructs were mixed with FuGENE HD (Roche Applied Science) in a 1:3 ratio and added following the manufacturer's instructions. Alternatively, astrocytes were electroporated (2 ϫ 10 6 astrocytes with 2 g of DNA or shRNA) before seeding using an astrocyte Nucleofector kit (Lonza, Switzerland). After electroporation, cells were plated to obtain a final cell density on the day of the experiment similar to that obtained with the transfection method. All experiments were performed after 48 -72 h. The transfection efficiency was 20 -30% for FuGENE HD and 60 -80% for electroporation as assessed with a GFP vector. At least three independent experiments were done in duplicate wells.
Cerebellar Granule Neuron Culture and Transfections-Cerebellar granule neurons cultures were produced from P7 rats as described previously (25). In brief, cells were plated onto 6-or 12-well dishes coated with poly-L-lysine (1 g/ml) at a respective final density of 1.5 ϫ 10 6 /well or 0.45 ϫ 10 6 / well. Cells were incubated at 37°C in 5% CO 2 in Neurobasal (Gibco) medium supplemented with 10% B27 (Gibco), 5 mM glutamine, and 25 mM KCl. All experiments were carried out in 2-7-day-old cultures with neurons showing neurite extensions. Granule neurons were transfected with the reagent Neurofect (AMS Biotechnology, Milton, UK) 24 h after plating with a DNA:transfection agent ratio of 1:7. Alternatively, neurons were electroporated (2 ϫ 10 6 neurons with 2 g of DNA or shRNA) before seeding using a Nucleofector kit according to the manufacturer's instructions. Neurons were left untreated at least for 48 h. The transfection efficiency was 5-10% for Neurofect and 40 -50% for electroporation as assessed with a GFP vector. On the day of the experiment, medium was replaced with Neurobasal ϩ 25 mM KCl. One hour later neurons were treated with IGF-I (10 Ϫ7 M), H 2 O 2 (50 -200 M), and/or inhibitors as above.
Co-cultures of Neurons and Astrocytes-Co-cultures were produced as described previously (16). In brief, wild type astrocytes (1.25 ϫ 10 5 /well) were seeded on 12-well plates with coverslips using DMEM/F-12 plus 10% FBS and 16 h later transfected with appropriate constructs. After 48 -72 h, cerebellar granule neurons were isolated and plated onto astrocytes. We used forebrain astrocytes and cerebellar granule neurons because in our experience the forebrain and cerebellum yielded very high numbers of astrocytes and neurons, respectively (thus minimizing animal use). Nevertheless, in previous studies, we carried out co-cultures with neurons and astrocytes from the same region (forebrain), obtaining identical results (16). Culture medium was changed to DMEM/F-12 plus B27, 4 mM glutamine, and 25 mM KCl (the latter only in the case of neurons). Two days later, co-cultures were treated with 100 nM IGF-I Ϯ 50 -150 M H 2 O 2 . After 24 h, cells were fixed and immunostained.
Cell Assays-Viability of astrocytes or neurons was assessed in different ways to assure reproducibility of results using 12-well plates. In the first assay, cells were stained after respective treatments with propidium iodide (2 g/ml). Propidium iodide-positive cells were counted in a Leica CTR 6000 fluorescence microscope (Wetzlar, Germany). In the second assay, cell cultures were transfected with a GFP-pCMV vector and the different constructs under evaluation in a 1:5 ratio. GFP ϩ cells were scored before treatment to determine baseline survival (time 0) and at different times thereafter. In these two assays, cells were counted in 12 different and random fields per well at 40ϫ magnification. In the third assay, the amount of lactate dehydrogenase released from damaged cells into the culture medium was used to quantify cell death. Lactate dehydrogenase levels were measured at various times with a commercial kit (Roche Diagnostics). In co-cultures, cellular viability was assessed in immunostained coverslips by counting the total number of viable ␤-III-tubulin ϩ (neuronal marker) and GFAP ϩ (astrocyte marker) cells, respectively, in 12 different and random fields per coverslip at 40ϫ magnification. Cellular viability was determined by plasma membrane integrity and absence of nuclear alterations detected with DAPI. The percentage of viable neurons and astrocytes in each treatment was expressed relative to vehicle treatments. All assays were done in triplicate dishes in at least three independent experiments.
Immunoassays-For immunostaining, cell cultures were washed with PBS and fixed with 4% paraformaldehyde in 0.1 M phosphate buffer for 15 min. Cells were treated with 4% horse serum in 0.1 M phosphate buffer for 1 h and washed three times for 10 min with 0.2% Triton and 0.3% BSA in 0.1 M phosphate buffer. Cells were incubated with primary antibody overnight, rinsed with phosphate buffer-Tween, and incubated with secondary fluorescent antibodies (Invitrogen) at 37°C for 1 h. Western blotting was performed as described (25). Neurons were washed once with ice-cold PBS and lysed with PIK buffer (1% Nonidet P-40, 150 mM NaCl, 20 mM Tris, pH 7.4, 10% glycerol, 1 mM CaCl 2 , 1 mM MgCl 2 , 400 M sodium vanadate, 0.2 mM PMSF, 1 g/ml leupeptin, 1 g/ml aprotinin, 0.1% phosphatase inhibitor mixtures I and II (Sigma-Aldrich)). To normalize for protein load, membranes were reblotted (Re-Blot, Chemicon) and incubated with an appropriate control antibody (see "Results"). Levels of the protein under study were expressed relative to protein load. Different exposures of each blot were collected to ensure linearity and to match control levels for quantification. Densitometric analysis was performed using Quantity One program (Bio-Rad). A representative blot is shown from a total of at least three independent experiments.
Luciferase Assays-Cells were transfected with a reporter construct bearing six canonical FOXO binding sites (6xDBE-luciferase) and co-transfected with different constructs as indicated in each experiment. Transfections were performed in triplicate dishes. Luciferase counts were normalized using TK-Renilla luciferase. At given times, cells were lysed in passive lysis buffer, and luciferase activity was analyzed using a luminometer and the Dual-Luciferase assay kit according to the manufacturer's instructions (Promega). Background luminescence was sub-tracted. Luciferase activity was expressed as -fold change of control levels.
Subcellular Fractionation-Membrane and cytosolic fractions were obtained from wild type or transfected astrocytes seeded on 6-well plates (1.25 ϫ 10 5 /well) as described (26). The quality of the fractionations was determined by assaying for the presence of the cytosolic protein Cu,Zn-SOD and the membrane protein R-RAS.
ROS Measurement-ROS generation was assessed by two different methods. Astrocytes cultured on coverslips were treated with the fluorogenic marker carboxy-H 2 DCFDA (Molecular Probes) for 30 min at 37°C with protection from the light. After incubation, cells were gently washed three times with warm DMEM and mounted. Pictures were taken at 40ϫ magnification using a Leica fluorescence microscope (Germany). A representative picture is shown. The percentage of astrocytes positive for the carboxy-H 2 DCFDA marker was quantified using ImageJ software. Using this ROS marker it is not possible to distinguish endogenous ROS from exogenously applied H 2 O 2 . Nevertheless, we compared this method with the oxidation of luminol (that detects superoxide anions (O 2 . )) that distinguishes H 2 O 2 from other ROS. To this end, we used a specific kit (Sigma). Half a million astrocytes were added to luminometer tubes containing the reagents for luminol oxidation and different concentrations of H 2 O 2 (0, 50, and 100 M) in a final volume of 200 l. Ten minutes later the chemiluminescence signal was determined. Results obtained with carboxy-H 2 DCFDA and with the kit based on luminol oxidation were similar (data not shown). All assays were done in triplicate dishes in at least three independent experiments. PTEN Lipid Phosphatase Activity-Phosphate released from substrates was measured with a Malachite Green Phosphatase Assay (Echelon). After treatments, astrocyte cultures were washed with ice-cold PBS and lysed with PIK buffer (see above). 250 g of total protein was immunoprecipitated with a monoclonal antimouse PTEN antibody. Lipid phosphatase PTEN activity was measured in 50 l of PTEN activity buffer (100 mM Tris-HCl, pH 8.0, 10 mM DTT) containing water-soluble D(ϩ)-sn-1,2-di-O-octanoylglyceryl,3-O-phospho-linked phosphatidylinositol phosphate. Samples were incubated for 30 min at 22°C with gentle shaking before measuring absorbance at 620 nm. Inorganic phosphate release was quantified by comparison with a standard curve of KH 2 PO 4 in distilled H 2 O.
Statistical Analysis-Data are expressed as mean Ϯ S.E. Differences between groups were analyzed by one-way analysis of variance followed by a Newman-Keuls or t test. A p value Ͻ0.05 was considered significant.

Results
Activation of AKT by IGF-I in Astrocytes Is Not Affected by Oxidative Stress-Previous evidence showed that oxidative stress elicited by H 2 O 2 interfered with IGF-I signaling and with its cytoprotective effect on neurons (9). Specifically, H 2 O 2 (50 -100 M) decreased activation of the prosurvival kinase AKT induced by 100 nM IGF-I in a dose-dependent manner (Fig. 1, A  and B). However, in astrocytes, the cytoprotective effect of IGF-I was preserved during oxidative stress in sharp contrast to neurons (16). Different doses of H 2 O 2 (50 -150 M), able to induce oxidative stress in astrocytes (16) addition but also significantly enhanced it (Fig. 1C). Thus, we next elucidated the molecular mechanisms underlying differential responses of neurons and astrocytes. AKT Activation by IGF-I in Astrocytes Is Not Affected by p38 Redox Activation-In neurons, interference by H 2 O 2 of IGF-Iinduced activation of AKT was mediated by the stress kinase p38 (9). Thus, redox activation of p38 prevented the phosphorylation of the docking protein IRS-1 (Tyr-612) induced by IGF-I, leading to the down-regulation of the PI3K/AKT pathway (9). Pretreatment with the p38 inhibitor SB239063 (7.5 M) prevented its activation by H 2 O 2 in neurons, restoring IRS-1 phosphorylation and AKT activation by IGF-I ( Fig. 2A). We tested whether this molecular mechanism could interfere with IGF-I signaling in astrocytes. We first observed that H 2 O 2 did not prevent IRS-1 phosphorylation after IGF-I treatment in astrocytes (Fig. 2B), which is in clear contrast to that observed in neurons (9) (Fig. 2C). Nevertheless, H 2 O 2 induced p38 activation in astrocytes just as in neurons (Fig. 2D). Furthermore, we also observed in astrocytes exposed to H 2 O 2 (50 -150 M) that pretreatment with SB239063 prevented p38 activation without up-regulating phosphorylation of IRS-1 or AKT by IGF-I (Fig. 2D). On the contrary, SB239063 significantly downregulated phosphorylation of IRS-1 in astrocytes exposed to the highest dose of H 2 O 2 (150 M) (Fig. 2D). These results indicate that activation of p38 by H 2 O 2 does not prevent AKT activation by IGF-I in astrocytes.
IGF-I Induces Phosphorylation of PTEN to Facilitate AKT Activation in Astrocytes-By quantification of phospho-AKT (Ser-473) protein levels, we observed that IGF-I-mediated activation of AKT was significantly higher and more sustained in astrocytes than in neurons. This effect was present in both normal and oxidative stress conditions (data not shown). Screening of pathways related with AKT activation showed that IGF-I induced (5 min after its addition) phosphorylation of the phosphatase PTEN at three specific residues (Ser-380, Thr-382, and Thr-383) of its C-terminal tail in astrocytes (Fig. 3A). This increase in PTEN phosphorylation was not affected by co-treatment with H 2 O 2 (50 -150 M), and it was not present in neurons treated with IGF-I where we observed the opposite effect, a decrease of PTEN phosphorylation (Fig. 3, A and B). These residues form part of a serine/threonine cluster located at the PTEN tail whose phosphorylation can decrease its phosphatase activity (27,28) or induce its cytosolic translocation, preventing PIP 3 dephosphorylation at the membrane (28,29). However, whereas H 2 O 2 (100 M) reduced the phosphatase activity of PTEN, IGF-I did not produce any significant effect (Fig. 3C). Nevertheless, we detected that PTEN phosphorylation mainly occurred at the membrane of astrocytes treated with IGF-I alone or in combination with H 2 O 2 (Fig. 3D). A time course analysis showed that PTEN translocates to the cytosol after its phosphorylation 15 min after IGF-I treatment (Fig. 3E). This translocation was enhanced by co-administration of H 2 O 2 (Fig.  3E). Phosphorylation of these residues is essential for the translocation of PTEN from the membrane to the cytosol after addition of IGF-I ϩ H 2 O 2 because levels of mutant PTENA4 (where Ser-380, Thr-382 and -383, and Ser-385, also a member of the serine/threonine cluster, were mutated to alanine) were not affected by treatment with IGF-I and H 2 O 2 (Fig. 3E). Confirming the specificity of these events in astrocytes, we did not observe PTEN phosphorylation at the membrane of neurons treated with IGF-I and H 2 O 2 (50 M) (data not shown). Finally, we analyzed whether PTEN phosphorylation was involved in the activation of AKT by IGF-I. Astrocytes were transfected with PTENA4 prior to treatment with IGF-I and H 2 O 2 (100 M). We observed that PTENA4 expression prevented the activation of AKT induced by IGF-I both in normal and under oxidant conditions (Fig. 3F). These results indicate that PTEN phosphorylation by IGF-I contributes to preserve AKT activity in astrocytes during oxidative stress conditions.
AKT Activation by IGF-I Prevents FOXO3-mediated Cell Death in Astrocytes during Oxidative Stress-AKT phosphorylates and inactivates the transcription factors of the FOXO family (30,31), which are key players in the cell survival/death outcome. In particular, FOXO3 is necessary for oxidative stress-mediated apoptosis in neurons (9, 32), and we found a similar role in astrocytes (16). Indeed, in astrocytes transfected with DN-FOXO3, cell death after exposure to H 2 O 2 (100 M) was significantly prevented (Fig. 4A). Our group and others have demonstrated that IGF-I induces phosphorylation of FOXO3 by AKT and its transcriptional inactivation (9,33). In neurons, H 2 O 2 (50 M) counteracts this pathway, allowing FOXO3 activation and promoting neuronal death (9). By contrast, in astrocytes, we observed that H 2 O 2 (100 M) or higher concentrations (data not shown) did not prevent IGF-I effects on FOXO3, keeping its phosphorylation by AKT (Fig. 4B) and therefore its transcriptional inac-  (Fig. 4C). Confirming our previous evidence (16), IGF-I-mediated inactivation of FOXO3 protected astrocytes from H 2 O 2 treatment (Fig. 4D).
FOXO3 activation by oxidative stress also depends on the activation of JNKs (9,32). This pathway was also present in astrocytes because transfection with a JNK-interfering protein, JIP-1, reduced FOXO3 activation by H 2 O 2 (Fig. 4E). Previously, we had described in neurons how the blockade of IGF-I signal-ing by oxidative stress was necessary for JNKs to activate FOXO3, unveiling a sequential activation of FOXO3 that depended first on down-regulation of its phosphorylation by AKT and second on its activation by JNKs (9). Bearing in mind these results, we examined in astrocytes whether AKT activation by IGF-I could prevent JNK activation of FOXO3 during oxidative stress. Initially, we observed lower levels of phospho- AKT astrocytes when IGF-I (100 nM) was co-administered (data not shown). This effect was prevented by pretreatment with the AKT inhibitor LY294002 (Fig. 4F). This result suggested a cross-talk between AKT and JNK pathways as described previously (34). Next, we analyzed whether this cross-talk could prevent FOXO3 activation by JNKs. Astrocytes were co-transfected with constitutively active MEKK (MEKK-CA), the kinase responsible for JNK activation, and AKT-CA or their control vectors. MEKK-CA expression significantly up-regulated FOXO3 transcriptional activity (Fig. 4G), whereas co-expression of AKT-CA prevented this activation (Fig. 4G). Overall, these results suggest that AKT activation by IGF-I could be crucial to prevent FOXO3 activation and cell death in astrocytes during oxidative stress. FOXO3 inactivation would depend on two AKT-mediated steps: phosphorylation of FOXO3 and inactivation of JNKs.
AKT Activation by IGF-I Reduces ROS Levels in Astrocytes-Activation of AKT has previously been involved in up-regulation of ROS detoxification mechanisms (35,36). We analyzed whether AKT activation by IGF-I or, alternatively, AKT-CA reduced ROS levels in astrocytes. Two different methods were used. Total ROS levels (endogenous ROS and those generated from exogenously applied H 2 O 2 ) were assessed with the fluorogenic marker carboxy-H 2 DCFDA. IGF-I significantly counteracted the ROS fluorescence signal induced by H 2 O 2 (Fig. 5, A  and B). We also observed that AKT inhibition by pretreatment with LY294002 neutralized the effects of IGF-I on total ROS levels (Fig. 5, A and B) (16). We now found that up-regulation of Cu,Zn-SOD by IGF-I and H 2 O 2 was abrogated by AKT inhibition with LY294002 (Fig. 5D). In contrast, AKT-CA expression in astrocytes mimicked the effect of IGF-I, up-regulating Cu,Zn-SOD levels both in normal and oxidant conditions (Fig.  5E). These results illustrate the antioxidative role of AKT in astrocytes.

AKT Activation by IGF-I Is Involved in Neuroprotection by
Astrocytes-In view of the cytoprotective and antioxidative effects mediated by AKT activation in astrocytes during oxidative stress, we analyzed whether this activation could also have an impact on neuroprotection. Neurons were cultured alone or with astrocytes in the presence of H 2 O 2 (50 M) and IGF-I (100 M). As expected (9,16), IGF-I did not prevent cell death by H 2 O 2 when neurons were cultured alone (Fig. 6A). However, when co-cultured with astrocytes, IGF-I significantly increased neuronal survival (Fig. 6A). This confirmed a neuroprotective effect mediated by IGF-I signaling in astrocytes. To test whether this effect depended on AKT activation, we transfected astrocytes with a PTENA4 mutant or a control vector to prevent this activation as demonstrated previously (Fig.  3G). We observed that PTENA4 expression significantly reduced survival of neurons in co-cultures treated with IGF-I and H 2 O 2 (50 M) compared with co-cultures where astrocytes were transfected with a control vector (Fig. 6B). To confirm that AKT activation was important for the neuroprotective effect of IGF-I, we transfected astrocytes with AKT-CA or a control vector. AKT-CA expression mimicked the effect of IGF-I (Fig. 6C).

Discussion
Previous observations indicate that IGF-I exerts a protective action on astrocytes, contributing to the resilience of these glial cells against oxidative stress, and collaborates with other trophic signals produced by astrocytes to mediate neuroprotection (16). Results presented herein confirm and broaden these findings, identifying two molecular adaptations, absent in neurons, that allow astrocytes to maintain the cytoprotective and antioxidative effects of IGF-I during oxidative insults. Thus, astrocytes enhance the IGF-I receptor/PI3K/AKT pathway in response to oxidative stress in part by cytosolic translocation of PTEN. Intriguingly, this prosurvival pathway is not down-modulated by the redox activation of p38 as seen in neurons and other cell types. These two molecular adaptations preserve AKT activation by IGF-I in astrocytes, allowing the inactivation of the proapoptotic transcription factor FOXO3 and the expression of ROS-detoxifying enzymes such as Cu,Zn-SOD. Finally, we also observed that by preserving AKT activation the neuroprotective role of astrocytes is maintained during oxidative stress insults (these results are summarized in Fig. 7).
Our group previously described a p38-mediated interference of the IGF-I/PI3K/AKT pathway induced by oxidative stress in neurons (9). Numerous lines of evidence in models of diabetes, cardiovascular dysfunction, and obesity confirm the role of p38 as a possible mediator of IGFs/insulin resistance induced by oxidative stress (37)(38)(39)(40)(41)(42) and the potential benefits of its inhibition (43,44). p38 redox activation prevents IRS-1/2 phosphorylation by IGF-I or insulin, blocking in this way PI3K and AKT activation (45). Here, we confirmed that H 2 O 2 treatment stim- ulated p38 in astrocytes but in this case without down-modulating IRS-1 phosphorylation by IGF-I. Even in astrocytes exposed to high H 2 O 2 doses p38 activation seemed necessary to maintain IRS-1 phosphorylation, which would confirm its possible protective role in astrocytes (46). This specific response of astrocytes could be explained by the activation of specific p38 isoforms, which could present different, even opposite, functions (47,48), or also by the existence of different pools of p38, which could be located in different subcellular compartments as well as bound to different partners (49). Further research aimed to identify these specific p38 isoforms and partners could help understand the molecular mechanism of IGF-I/insulin resistance in other susceptible cells such as pancreatic ␤ cells or neurons. A specific regulation of p38 in astrocytes may help prevent IGF-I resistance in astrocytes during oxidative stress insults.
Our results showed that in astrocytes, but not in neurons, IGF-I induced PTEN phosphorylation in a specific serine/threonine cluster of amino acids (Ser-380, Thr-382, and Thr-383) at its C-terminal tail. This confirms previous work describing that this phosphorylation can interfere with the presence of PTEN at the plasma membrane, which in turn results in increased AKT activation (28,29,50), and that oxidative stress induced by H 2 O 2 enhances this effect. PTEN phosphorylation is induced by the prosurvival kinase casein kinase 2 (51-53), which could be a novel target of IGF-I-mediated signaling in astrocytes.    Translocation of PTEN from the membrane to the cytosol took place within 15 min after IGF-I treatment, whereas our previous results had shown a transient reduction of PTEN mRNA and protein levels 6 h after IGF-I addition (54), indicating that IGF-I could use both short and long term mechanisms to regulate PTEN activity. Collectively, these results suggest that PTEN inhibition may be a key mechanism to preserve IGF-Imediated AKT activation in astrocytes during oxidative stress conditions. We have also revealed that preservation of AKT activation by IGF-I provides important advantages for astrocyte survival during oxidative stress. First, AKT inactivates FOXO3, a transcription factor that is key to trigger cell death after oxidative stress both in neurons and astrocytes (9,16). Hence, AKT activation prevented FOXO3 activation by JNKs, the signaling pathway that stimulates its proapoptotic role in neurons during oxidative stress (9,32). These results highlight the importance of AKT activation by IGF-I in astrocytes to tilt the balance between inhibitory and excitatory signals of FOXO3, preventing its proapoptotic effects during oxidative stress. Second, AKT activity reduces ROS levels during oxidative stress. Previous observations already indicated antioxidative actions of IGF-I in astrocytes (16). We describe here that one of them, the up-regulation of the antioxidative enzyme Cu,Zn-SOD, depends specifically on AKT activation. Additional antioxidative mechanisms related to AKT activity could also participate in IGF-I cytoprotection, as for example up-regulation of the transcription factor Nrf2 (36).
Astrocytes are coupled to neurons to provide ROS detoxification support during oxidative stress insults (21). Several works also suggest that the release by astrocytes of soluble and insoluble factors could contribute to their neuroprotective role (55,56). Supporting this idea, our group has recently described that IGF-I cooperates with stem cell factor secreted by astrocytes to protect neurons against oxidative stress (16). The present results suggest that preservation of AKT activity in astrocytes can be key for IGF-I neuroprotection. Expression of CA-AKT in astrocytes partially mimicked IGF-I neuroprotection in co-cultures with neurons exposed to H 2 O 2 , whereas expression of a PTENA4 mutant insensitive to IGF-I inhibition prevented AKT activation in astrocytes and reduced IGF-I neuroprotection. AKT-mediated neuroprotection probably includes its effect on ROS detoxification (57) and up-regulation of secreted factors such as stem cell factor, whose promoter can be regulated by transcription factors targeted by AKT (58,59). Furthermore, PTEN inhibition in astrocytes could be neuroprotective in pathologies associated with oxidative stress. Supporting this idea, PTEN inhibition has shown cytoprotective effects in animal models of oxidative stress-associated pathologies such as diabetes, obesity, and cardiac ischemia (60 -63). However, evidence about the neuroprotective role of PTEN inhibition in brain ischemia is contradictory (64, 65). PTEN . In astrocytes, IGF-I induces the stimulation of the prosurvival kinase AKT by the activation of a signaling cascade that includes IRS-1 phosphorylation, translocation of PI3K to the membrane (allowing PIP 3 generation and AKT recruitment), and activation of AKT by PDK1/ mTORC2 kinases. Oxidative stress (generated by an uncontrollable ROS increase) can induce redox activation of p38 kinase in several cells types. In neurons, p38 induces IGF-I resistance, preventing IRS-1 phosphorylation by IGF-I receptor IGF-IR, whereas in astrocytes, this response is not present. Furthermore, IGF-I induces only in astrocytes the inactivation of the phosphatase PTEN (by inducing its phosphorylation and cytosolic translocation), preserving in this way generation of PIP 3 and AKT activation in both normal and oxidative stress conditions. Armoring of AKT activation by IGF-I increases astrocyte survival during oxidative stress through inactivation of the proapoptotic FOXO3 transcription factor (by its phosphorylation by AKT and by inactivation of JNKs) and stimulation of ROS-detoxifying enzymes. Furthermore, AKT activation is also necessary to preserve the neuroprotective action of astrocytes.
displays multiple functions in the brain, including tumor suppression, axonal outgrowth, astrogliosis, and cognitive function regulation (66 -69), and its total inhibition could affect all of them in a nonspecific manner. Therefore, further research to develop more precise modulators of PTEN and appropriate timings of administration may help develop its possible neuroprotective role.
Overall, the results presented in this study reinforce the notion that IGF-I could display antioxidative actions in specific cellular contexts and types. We demonstrate that in astrocytes these actions depend, at least in part, on the preservation of AKT activation. This is achieved through molecular adaptations targeting p38, PTEN, and FOXO3 that result in increased resilience of astrocytes to IGF-I resistance induced by oxidative stress. Our results also suggest that activation of AKT in astrocytes by growth factors such as IGF-I, which is produced by astrocytes during stress situations (16,70), may be part of an endogenous brain defense mechanism against oxidative stress injury.
Author Contributions-D. D. conducted most of the experiments, analyzed the results, and conceived and wrote most of the paper. S. F. conducted experiments to determine PTEN phosphatase activity. I. T.-A. conceived the idea for the project and wrote the paper with D. D.