Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2008 May;19(5):2014-25.
doi: 10.1091/mbc.e07-08-0811. Epub 2008 Feb 20.

Neuronal Death by Oxidative Stress Involves Activation of FOXO3 Through a Two-Arm Pathway That Activates Stress Kinases and Attenuates Insulin-Like Growth Factor I Signaling

Affiliations
Free PMC article

Neuronal Death by Oxidative Stress Involves Activation of FOXO3 Through a Two-Arm Pathway That Activates Stress Kinases and Attenuates Insulin-Like Growth Factor I Signaling

David Dávila et al. Mol Biol Cell. .
Free PMC article

Abstract

Oxidative stress kills neurons by stimulating FOXO3, a transcription factor whose activity is inhibited by insulin-like growth factor I (IGF-I), a wide-spectrum neurotrophic signal. Because recent evidence has shown that oxidative stress blocks neuroprotection by IGF-I, we examined whether attenuation of IGF-I signaling is linked to neuronal death by oxidative stress, as both events may contribute to neurodegeneration. We observed that in neurons, activation of FOXO3 by a burst of oxidative stress elicited by 50 muM hydrogen peroxide (H(2)O(2)) recruited a two-pronged pathway. A first, rapid arm attenuated AKT inhibition of FOXO3 through p38 MAPK-mediated blockade of IGF-I stimulation of AKT. A second delayed arm involved activation of FOXO3 by Jun-kinase 2 (JNK2). Notably, blockade of IGF-I signaling through p38 MAPK was necessary for JNK2 to activate FOXO3, unveiling a competitive regulatory interplay between the two arms onto FOXO3 activity. Therefore, an abrupt rise in oxidative stress activates p38 MAPK to tilt the balance in a competitive AKT/JNK2 regulation of FOXO3 toward its activation, eventually leading to neuronal death. In view of previous observations linking attenuation of IGF-I signaling to other causes of neuronal death, these findings suggest that blockade of trophic input is a common step in neuronal death.

Figures

Figure 1.
Figure 1.
Neuronal death after exposure to hydrogen peroxide. (A) Time schedule of experiments. Cells were transfected 24 h after plating. Experiments were conducted in 3–7-d-old cultures, when neurons showed well developed neurites. (B) Addition of 50–100 μM H2O2 to the cultures generated reactive oxygen species (ROS; ***p < 0.001 vs. 0 dose) in a dose-dependent manner as determined by superoxide levels (RLUs arbitrary units), as well as a significant increase in the cellular NAD(P)/NAD(P)H ratio 15 min later (C), indicating a change in the redox status of the cell (**p < 0.01; n = 3). (D) Six hours after addition of 50 μM H2O2 to the cultures, many cells were apoptotic, as determined with a pan-active caspase fluorescent indicator (FITC-VAD-FMK) regardless of the presence of IGF-I. Histograms: quantitation of fluorescent-labeled cells (n = 3, p < 0.001 vs. controls). DAPI stain was used to mark all the cells in the cultures. (E) Neuronal death elicited by exposure to 50 μM H2O2 was also determined by percentage of dead neurons counting either propidium iodide-stained cells (PI+) 6 h after insult, or (F) measuring LDH released to the medium 12 h after adding H2O2 (100% cell death corresponds to LDH levels in H2O2-treated cultures after 12 h). IGF-I (100 nM) protected neurons only in the absence of H2O2 (***p < 0.001 and *p < 0.05 vs. controls; n = 3). (G) Transcriptional activity of Bim EL is increased by H2O2 in IGF-I–treated neuronal cultures. Blots: protein levels of Bim EL were increased by H2O2 (10 μM) even though IGF-I was present. Actin levels are shown in lower blots. A representative experiment is shown (*p < 0.05 vs. IGF-I alone; n = 3 for both types of experiments).
Figure 2.
Figure 2.
Modulation of AKT/FOXO by H2O2 in cerebellar neurons. (A) In the presence of H2O2, IGF-I–induced phosphoAKT (pAKT, 15 min after IGF-I) was significantly blocked. Levels of total AKT were measured as loading control. **p < 0.01 versus all other groups; n = 5. (B) Nuclear translocation of pAKT after IGF-I was blocked by H2O2. NeuN levels were determined as nuclear fraction control (n = 3). (C) In response to IGF-I levels of FOXO3 in the nucleus 4 h later were reduced, but this effect was blocked by H2O2. NeuN was used as a marker of the nuclear fraction; n = 3. (D) The opposite effect is observed with cytoplasmic levels of FOXO3, which were increased 4 h after IGF-I, and addition of H2O2 counteracted this effect. β-actin was measured as a marker of the cytosolic fraction (n = 4). (E) Transcriptional activity of FOXO was significantly attenuated by IGF-I after 8 h, but addition of H2O2 greatly enhanced it; n = 3. **p < 0.01 versus IGF-I–treated cells. (F) Neurons cotransfected with GFP and dominant negative (DN)-FOXO3 were protected against the deleterious effects of H2O2 compared with GFP-transfected neurons (controls), were a 40% drop in the number of GFP+ cells was observed 6 h after treatment with H2O2, regardless of the presence or absence of IGF-1. ***p < 0.001, and **p < 0.01 versus controls (n = 3).
Figure 3.
Figure 3.
Reactive oxygen species inhibit IGF-I signaling in neurons. (A) Conjoint stimulation of cerebellar neurons with IGF-I and H2O2 (50 μM) resulted in reduction of Tyr-phosphorylation of IRS-1 15 min later, whereas Ser-phosphorylation of IRS-1 was enhanced 1 h later. Representative blots are shown. Histograms: densitometric quantification of blots. *p < 0.05; n = 3. (B) The presence of H2O2 in IGF-I–treated neuronal cultures also led to disrupted interaction of IRS-1 with PI3K, as less p85PI3K coimmunoprecipitated with it. Association of IRS-1 to p85PI3K 15 min after adding IGF-I was determined by reciprocal coimmunoprecipitation. Representative blots are shown. Histograms: densitometric quantification of blots. *p < 0.05 (n = 2 for each). (C) In the presence of 7.5 μM SB239063, a P38α,β MAPK inhibitor, H2O2 no longer induces the phosphorylation of this kinase, and Tyr-phosphorylation of IRS-1, and phosphoAKT (pAKT) in response to IGF-I is preserved (n = 3). Representative blots are shown. (D) In the presence of 7.5 μM SB239063, FOXO3 phosphorylation at the AKT-sensitive Thr32 residue by IGF-I is preserved. Representative blots are shown. Histograms: densitometry of blots. **p < 0.01; n = 3. (E) Neuronal cultures transfected with a DN-p38β MAPK before exposure to IGF-I+ H2O2 showed markedly reduced FOXO transcriptional activity in response to H2O2. Cultures transfected with a DN-p38α MAPK showed also, albeit smaller, a reduction in FOXO activity. *p < 0.05 and ***p < 0.001 versus controls; n = 3. (F) The P38α,β MAPK inhibitor SB239063 blocked H2O2-induced cell death in IGF-I–treated cultures, as assessed by the number of PI+ cells 6 h later. *p < 0.05 versus controls; n = 3 for each assay.
Figure 4.
Figure 4.
Role of JNK in H2O2-induced neuronal death. (A) Addition of H2O2 to IGF-I–treated cerebellar neurons produced a delayed increase in JNK2 activity as determined by enhanced levels of phospho-JNK2 (pJNK2) between 1 and 2 h later; n = 3. (B) Densitometric quantification of the increase in pJNK2 at 2 h. Representative blot is shown below; n = 4. (C) Transfection of the JNK inhibitor JIP-1 to neurons was sufficient to reduce activation of FOXO by H2O2 regardless of IGF-I. Neuronal cultures were cotransfected with WTFOXO3- and JIP-1–expressing vectors and FOXO activity was measured in luciferase gene reporter assays. **p < 0.01 and ***p < 0.001; n = 4. (D) Neuronal death is substantially abrogated when JIP-1 is expressed in ±IGF-I–treated neurons before addition of H2O2. *p < 0.05, and ***p < 0.001 versus respective JIP-1–expressing cultures; n = 3. Cell death was assessed as number of GFP+ cells 6 h after adding H2O2 and expressed as percentage of GFP+ cells at time 0. (E) After addition of Sp600125 (20 μM), an inhibitor of JNKs, translocation of FOXO3 from the nucleus to the cytoplasm by IGF-I is preserved in the presence of H2O2 in the cultures. β-actin levels were measured as control of the cytoplasmic fraction. Histograms: densitometric quantification of blots. *p < 0.05; n = 3. (F) Sp600125 restored FOXO3 phosphorylation by IGF-I at the AKT-sensitive Thr32 residue, even in the presence of H2O2. Histograms: quantification of pThr32FOXO3 corrected for total FOXO3. (*p < 0.05; n = 5). Control experiments in the absence of IGF-I are shown in a separate representative blot in the left (*p < 0.05; n = 3). (G) Levels of pAKT remained reduced after treatment of IGF-I–treated cultures with H2O2 when Sp600125 was present. This correlated with intact activation of p38MAPK by H2O2. Inhibition of JNK by Sp600125 was corroborated by inhibition of phosphorylation of c-Jun (a downstream target of JNK). Histograms: quantification of pAkt corrected for total Akt (*p < 0.05 vs. respective control; n = 4). Representative blots are shown in all panels.
Figure 5.
Figure 5.
An interplay of AKT and JNK onto FOXO activity. (A) Expression of a constitutively active (CA) form of AKT resulted in abrogation of cell death by H2O2 even in the absence of IGF-I (right-hand bars). Cotransfection of GFP, WTFOXO3 and CA-AKT expressing vectors inhibits neuronal death after H2O2 compared with cultures cotransfected with control vector, GFP- and WTFOXO3-expressing vectors. Live neurons are expressed as percentage of GFP+ cells at time 0. **p < 0.01, n = 2. (B) Transcriptional activity of FOXO is robustly modulated by AKT. Cotransfection of CA-AKT with WTFOXO3 results in markedly reduced FOXO activity after H2O2 regardless of IGF-I. In contrast, cotransfection of CA-AKT with an AKT-insensitive mutant (M) FOXO3 results in preservation of enhanced FOXO activity after H2O2. **p < 0.01 and ***p < 0.001 versus indicated groups; n = 4. (C) Transcriptional activity of FOXO is enhanced by MEKK, which in turn counteracts the inhibitory action of IGF-I. *p < 0.05 and **p < 0.01 versus indicated groups; n = 3. (D) Antagonistic regulation of FOXO activity by MEKK and AKT, respectively. AKT was able to counteract the stimulatory effect of MEKK onto FOXO in a dose-dependent way. Neuronal cultures were cotransfected with various amounts of the respective vectors, and FOXO activity was determined with a luciferase gene-reporter system. ***p < 0.001 versus CA-AKT–cotransfected cultures; n = 3. (E) Expression of MEKK in neuronal cultures resulted in increased cell death as determined by the number of GFP+ cells. Cultures were cotransfected with GFP- and CA-MEKK– or GFP-expressing vectors only (controls) with or without IGF-I, and the number of cells was scored 24 h later. *p < 0.05 versus IGF-I–treated cultures; n = 3. (F) Partial abrogation of H2O2-induced FOXO activity by expression of CA-AKT (1 μg) was not modified by addition of the JNK inhibitor Sp600125. Note that inhibition of FOXO activity was similar by either inhibition of JNK or stimulation of AKT (p < 0.01 vs. H2O2-treated cultures (dotted bar); n = 3).
Figure 6.
Figure 6.
Activation of FOXO3 by H2O2 through JNK involves Ser phosphorylation. (A) Cerebellar neurons transfected with an HA-tagged MFOXO3 (where AKT-sensitive residues have been mutated) show dose-dependent increases in Ser-phosphorylation (pSer) levels after addition of H2O2. Cell extracts were immunoprecipitated with anti-HA antibodies 4 h after exposure to H2O2, and levels of pSer in immunoprecipitates were determined; n = 3. (B) Addition of the JNK inhibitor Sp600125 to HA.MFOXO3–transfected neuronal cultures abrogated PSer by 75 μM H2O2 measured 4 h later. Representative blots are shown; n = 3. Coexpression of the JNK inhibitor JIP-1 instead of adding the JNK inhibitor produced an identical effect (not shown). *p < 0.05.
Figure 7.
Figure 7.
Inhibition of IGF-I signaling is involved in neuronal death after H2O2. A competitive interplay between AKT and JNK in regulating FOXO3 activity dictates the neuronal response to an abrupt rise in ROS elicited by H2O2. Cytotoxic levels of H2O2 sequentially activate two independent pathways. A rapid route (1) includes activation of p38 MAPK to inhibit IGF-I signaling by interfering IGF-I receptor/IRS-1 interactions through phosphorylation of serine residues in IRS-1. This leads to abrogation of AKT inhibition of FOXO. A delayed route (2) recruits JNK2 to activate FOXO. Activation of the p38 MAPK pathway is essential because JNK is unable to activate FOXO if AKT is active. Conversely, once JNK activates FOXO, it becomes refractory to inactivation by AKT. The different inhibitory drugs and stimulatory and inhibitory DNA constructs used in this study are indicated.

Similar articles

See all similar articles

Cited by 31 articles

See all "Cited by" articles

Publication types

MeSH terms

Substances

Feedback