Regulation of Cu/Zn-Superoxide Dismutase Expression via the Phosphatidylinositol 3 Kinase/Akt Pathway and Nuclear Factor-κB

Introduction

Reactive oxygen species (ROS) are generated during aerobic metabolism or in response to external toxins and produce oxidative damage to cellular macromolecules (Klein and Ackerman, 2003). Adaptive responses adjust the abundance and activity of antioxidant systems to cellular needs (Fridovich, 1999). One such system is constituted by the superoxide dismutase (SOD) enzymes, which convert superoxide anion into molecular oxygen and hydrogen peroxide (H₂O₂) (Fridovich, 1997; Zelko et al., 2002).

Cu/Zn-SOD, the most abundant and ubiquitous isoform, may have great physiological significance and therapeutic potential in neurodegenerative diseases because the CNS is particularly sensitive to oxidant injury. For instance, an inverse correlation has been reported between Cu/Zn-SOD activity and neuronal death after acute brain injury (Takeuchi et al., 2000). Moreover, the level of Cu/Zn-SOD mRNA was reduced significantly in the substantia nigra of marmosets treated with the parkinsonian toxin MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine) and in patients with Parkinson's disease (Kunikowska and Jenner, 2003). In addition, Cu/Zn-SOD transgenic mice were protected against the neurotoxic effects of 6-hydroxydopamine (6-OHDA) (Asanuma et al., 1998). Transcription of sod1, the gene coding for Cu/Zn-SOD, is regulated tightly in response to diverse stimuli, including stress, proinflammatory cytokines, and growth factors (Kong et al., 1993; Yoo et al., 1999; Chang et al., 2002; Park and Rho, 2002). However, despite the detailed characterization of the regulatory promoter region, little is known about the intracellular signaling cascades that govern sod1 expression.

The phosphatidylinositol 3 kinase (PI3K) and its best characterized effector, the Ser/Thr protein kinase Akt/protein kinase B (PKB), exert antioxidant effects in central and peripheral neurons (Brunet et al., 2001; Franke et al., 2003). Recently, we have reported the protective effect of active Akt1 against peptides of β-amyloid protein (Martin et al., 2001), against the Parkinson-inducing toxins 1-methyl-4-phenylpyridinium (MPP⁺) (Salinas et al., 2001) and 6-OHDA (Salinas et al., 2003), and against apoptotic concentrations of H₂O₂ (Martin et al., 2002). In these cases, expression of a constitutively active version of Akt1 prevented the increase in ROS that follows treatment of PC12 cells with these neurotoxins. Despite these reports, a direct regulatory role of this pathway on the canonical antioxidant defenses such as Cu/Zn-SOD has not been demonstrated clearly.

The survival effect of PI3K and Akt generally is attributed to Akt-mediated phosphorylation of proapoptotic proteins but also to transcription-dependent mechanisms (Kuruvilla et al., 2000; Brunet et al., 2001). Several transcription factors have been identified as substrates of Akt that might mediate Akt actions on gene expression of antioxidant enzymes. Akt directly phosphorylates the Forkhead transcription factors of the FOXO class (Brunet et al., 1999) and may participate in the activation of cAMP response element binding protein (CREB) (Du and Montminy, 1998) and nuclear factor-κB (NF-κB) (Kane et al., 1999).

In this study we report that the PI3K/Akt axis upregulates the expression of Cu/Zn-SOD via the activation of NF-κB. These data contribute to reveal the mechanism whereby PI3K/Akt provide protection against oxidative injury and suggest a new therapeutic strategy to increase the antioxidant capacity of neurons via the activation of this survival pathway.

Materials and Methods

Cell culture and reagents. Rat pheochromocytoma PC12 cells [gift of Dr. H. Kleinman, National Institute of Dental and Craniofacial Research (NIDCR), National Institutes of Health (NIH), Bethesda, MD] were grown in DMEM supplemented with 7.5% heat-inactivated fetal bovine serum, 7.5% heat-inactivated horse serum, and 80 μg/ml gentamycin. PC12 cells stably expressing myr-Akt1-ER^* were grown in medium supplemented with serum treated with charcoal-coated dextran (Sigma-Aldrich, Madrid, Spain). Stable transfections with the expression vectors pCEFL(X^-)EGFP, pCEFL(X^-)myr-EGFP-Akt1, and pcDNA3-myr-Akt1-ER^* were made with SuperFect transfection reagent (Qiagen, Valencia, CA) according to the manufacturer's instructions. Transfected cells were selected in 0.5 mg/ml G418. Other reagents were nerve growth factor (NGF) (Promega, Madison, WI), H₂O₂, glutathione, LY294002, and 4-hydroxytamoxifen (4-HT; Sigma-Aldrich).

Plasmids. Constructs expressing the EGFP-Akt1 fusion proteins were generated as previously described (Salinas et al., 2001). For generation of pcDNA3-myr-Akt-ER^*, the BamHI/EcoRI fragment of pBluescript KS⁺MER (Littlewood et al., 1995) was subcloned in the same sites of pcDNA3.myr-Akt1 (Salinas et al., 2000). This fragment comprises a mutant hormone-binding domain of the estrogen receptor (ER^*) that does not bind β-estradiol but retains responsiveness to the synthetic steroid 4-HT. Then we put myr-Akt1 and ER^* in the same reading frame and removed the stop codon of myrAkt1 by using a PCR strategy (details available on request). Expression vectors for anchored membrane forms of myristoylated Akt1 and farnesylated PI3K included the following: pcDNA3-myr-EGFP-Akt1 (Salinas et al., 2000), pSG5gag-Akt1, and pSG5p110α-PI3K-CAAX (gift of Dr. J. Downward, Imperial Cancer Research Fund, London, UK). From pSG5p110α-PI3K-CAAX the EcoRI fragment containing the coding sequence for PI3K-CAAX was subcloned into vector pTG6600 under the control of the CMV promoter; pCMV-m1, expressing the muscarinic m1 receptor, was a gift of Dr. S. Gutkind (NIDCR, NIH, Bethesda, MD); NF-κB-HIV contains the NF-κB site of the HIV enhancer/promoter (Montaner et al., 1999). Other expression vectors were double-point mutant IκBα(S32A/S36A) (Traenckner et al., 1995) and pGal4-p65 and pGal4-LUC (Montaner et al., 1999; Madrid et al., 2000). Deletion mutants from the 5′ flanking region of sod1 were generated by following the same strategy described in Minc et al. (1999). Briefly, PCR fragments of different lengths were amplified from the 5′-flanking region of the human sod1 gene by using Taq Gold DNA polymerase (Promega). For amplification of fragments -1499, -952, -750, and -552, human genomic DNA was used as template, and a common reverse primer was used (5′-CCGAAAGCTTGAGACTACGACGCAAACCCAG-3′) with a HindIII site to facilitate subsequent cloning. Forward oligonucleotides were designed according to the size of the future fragment: -1499 bp, 5′-CCGACTCGAGCCCTTGGCAAGTTTACAATG-3′; -952 bp, 5′-CCGACTCGAGGTGGTCCCAGGTACTTGGGGA-3′; -750 bp, 5′-CCGACTCGAGTATTTCCCTTGAAAGGTAAG-3′; -552 bp, 5′-CCGACTCGAGACCGAATTCTGCCAACCAAA-3′. A XhoI site was introduced in all forward primers to allow directional cloning into the reporter vector pGL3basic (Promega). Amplifications were performed by using 20 ng, with a thermal profile of (1) 96°C, 3 min; (2) 96°C, 2 min; (3) 65°C, 2 min; (4) 72°C, 3 min; (5) 40× to 2; (6) 72°C, 5 min. The amplified fragments were introduced in pGL3basic between the XhoI and HindIII sites. PCR fragments -355, -157, -71, and -29 were amplified by using plasmid pGL3basic-sod1-1499 as template and the reverse primer 5′-ATAGGATCTCTG-GCATGCGAGAATCTCACG-3′, in which an SphI is introduced to allow subsequent cloning. Forward oligonucleotides included the following: -355 bp, 5′-CCGACTCGAGTGGCCAAACTCAGTCATAAC-3′; -157 bp, 5′-CCGACTCGAGACGCGCCCCTTGCCCCGCCC-3′; -71 bp, 5′-CCGAACTCGAGATTGGTTTGGGGCCAGAGTG-3′; and -29 bp, 5′-CCGACTCGAGTATAAAGTAGTCGCGGAGAC-3′. The amplified fragments were subcloned in pGL3basic between sites XhoI and SphI. NF-κB-sod1 and NF-κB-mut were obtained with the following oligonucleotides: 5′-CTAGCAAAGGTAAGTCCCGGATCCAAAGGTAAGTCCCGGATCAAGGTAAGCCCGGC-3′ and 5′-CTAGCAAAGGTAAGTATTGTCGACAAAGGTAAGTTACGGATCAAGGTAAGTATTGGC-3′ (underlined sites, NF-κB). These oligonucleotides were subcloned into NheI and XhoI sites of the minimal sod1 promoter pGL3basic-sod1-29.

Luciferase assays. PC12 cells were seeded in 24-well plates (75,000 cells per well), cultured for 16 hr, and transfected with a DNA mixture consisting of 100 ng of the appropriate luciferase construct and 400 ng of the effector plasmid or empty vector, using Lipofectamine reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's recommendations. After 24 hr from transfection the cells were serum-starved for 16 hr, lysed, and assayed for luciferase activity with the Luciferase Assay System (Promega) according to the manufacturer's instructions; relative light units were measured in a BG1 Optocomp I, GEM Biomedical luminometer (Sparks, NV).

Analysis of mRNA levels by reverse transcriptase-PCR. Total cellular RNA was extracted with Trizol reagent (Invitrogen). Equal amounts (1 μg) of RNA from each treatment were reverse-transcribed (75 min, 42°C), using5Uof avian myeloblastosis virus reverse transcriptase (RT) in the presence of 20 U of RNasin (Promega). cDNA amplification was performed in 25 μl of PCR buffer [containing (in mm): 10 Tris-HCl, 50 KCl, 5 MgCl₂, plus 0.1% Triton X-100, pH 9.0] containing 2.5 mm digoxigenin-dUTP, 0.6 U of Taq DNA polymerase (Promega), and 30 pmol of synthetic gene-specific primers for Cu²⁺/Zn²⁺-SOD (forward, 5′-TTCGAGCAGAAGGCAAGCGGTGAA-3′; reverse, 5′-AATCCCAATCACACCACAAGCCAA-3′). To ensure that equal amounts cDNA were added to the PCR, we amplified the β-actin housekeeping gene by using synthetic oligonucleotides (forward, 5′-TGTTTGAGACCTTCAACACC-3′; reverse, 5′-CGCTCATTGCCGATAGTGAT-3′). After an initial denaturation step (4 min at 94°C) the amplification of each cDNA was performed at the optimal number of cycles within linear range (data not shown), 19 cycles, using a thermal profile of 1 min at 94°C (denaturation), 1 min at 58°C (annealing), and 1 min at 72°C (elongation). The amplified PCR products were resolved in 1.8% agarose gel electrophoresis and transferred to nylon membranes (Biodyne, Pall Corporation, New York, NY). Chemiluminescence detection was performed with a digoxigenin luminescence detection kit (Roche Molecular Biochemicals, Basel, Switzerland).

Immunoblotting. The primary antibodies used were anti-Cu²⁺/Zn²⁺-SOD (StressGen, Victoria, British Columbia, Canada), anti-p65, anti-Sp1, anti-IκBα, and anti-Akt1C20 (Santa Cruz Biotechnology, Santa Cruz, CA); anti-phospho-Thr ³⁰⁸Akt1 (New England Biolabs, Beverly, MA); and anti-protein-disulfide-isomerase (gift from Dr. J. G. Castaño, Chief Advisor of Scientific Investigations, Institute of Biomedical Investigations, Autónoma University of Madrid). Cells were washed once with cold PBS and lysed on ice with 200 μl of lysis buffer [containing (in mm): 137 NaCl, 20 Tris-HCl, pH 7.5, 1 phenylmethylsulfonyl fluoride, 20 NaF, 1 sodium pyrophosphate, and 1 Na₃VO₄ plus 1% Nonidet P-40, 10% glycerol, 1 μg/ml leupeptin]. Precleared cell lysates were resolved in SDS-PAGE and transferred to Immobilon membranes (Millipore, Billerica, MA). These membranes were analyzed by using the primary antibodies indicated above and peroxidase-conjugated secondary antibodies. Proteins were detected by enhanced chemiluminescence (Pierce Biotechnology, Rockford, IL). Densitometric analyses of representative immunoblots were performed with NIH Image software (Bethesda, MD).

Flow cytometry. Intracellular ROS were detected with hydroethidine (HEt; Molecular Probes, Leiden, The Netherlands) (bandpass, 575/24 nm filter), using a FACScan flow cytometer (BD Biosciences, Madrid, Spain). Cells were detached from the plates, washed with PBS, incubated with 2 μm HEt for 1 hr at 37°C, and analyzed immediately. Baseline incorporation of the probe was determined in cells incubated in 2 μm HEt for 1 hr at 4°C. For Annexin-V-phycoerythrin (Annexin-V-PE) staining the cells were detached mechanically after treatments and washed twice with Annexin-V binding buffer [containing (in mm): 10 HEPES-NaOH, pH 7.4, 150 NaCl, 5 KCl, 1 MgCl₂, 1.8 CaCl₂]. Cells were resuspended in 100 μl of a 1:20 dilution of Annexin-V-PE (Bender Med-Systems Diagnostics, Vienna, Austria) in Annexin-V binding buffer containing 2 μm 7-aminoactinomycin D (7-AAD) and were incubated for 15 min at room temperature. Fluorescence was measured with bandpass 620/22 and 575/24 nm filters. For immunostaining the cells were fixed with 4% paraformaldehyde for 1 hr at 4°C, washed with PBS, permeabilized with 0.2% Triton X-100 in PBS buffer for 5 min at 4°C, and incubated for 1 hr with anti-Cu/Zn-SOD (1:100) diluted in PBS and for 1 hr with Alexa 465-conjugated secondary antibody (1:500; Molecular Probes) in PBS.

Confocal microscopy. Cells were seeded on glass-bottom 35 mm plates (Willco, Amsterdam, The Netherlands) and incubated with 0.5 mm H₂O₂ for 3 hr and afterward with 2 μm HEtfor 1 hr at 37°C. Confocal microscopy was performed via a Leica SP2 system (Nussloch, Germany). For HEt we used an excitation wavelength of 488 nm; fluorescence was detected at wavelengths between 557 and 740 nm. For Hoechst 33258 we used an excitation wavelength of 351 nm; fluorescence was detected at wavelengths between 386 and 512 nm.

Preparation of nuclear and cytosolic extracts. Cytosolic and nuclear fractions were prepared as previously described (Minc et al., 1999). Briefly, 5 × 10⁶ cells were washed three times with cold PBS and harvested by centrifugation at 1100 rpm for 10 min. The cell pellet was resuspended carefully in three pellet volumes of cold buffer A [containing (in mm): 20 HEPES, pH 7.0, 0.15 EDTA, 0.015 EGTA, 10 KCl, 1 phenylmethylsulfonyl fluoride, 20 NaF, 1 sodium pyrophosphate, and 1 Na₃VO₄ plus 1% Nonidet P-40, 1 μg/ml leupeptin]. Then the homogenate was centrifuged at 500 × g for 5 min. The supernatant corresponding to the cytosolic fraction was resolved in SDS-PAGE and immunoblotted with anti-p65, anti-IκBα, or anti-PDI antibodies. The nuclear pellet was resuspended in five pellet volumes of cold buffer B [containing (in mm): 10 HEPES, pH 8.0, 0.1 EDTA, 1 phenylmethylsulfonyl fluoride, 20 NaF, 1 sodium pyrophosphate, and 1 Na₃VO₄ plus 25% glycerol, 0.1 m NaCl, 1 μg/ml leupeptin]. After centrifugation in the same conditions indicated above, the nuclei were resuspended in two pellet volumes of hypertonic cold buffer C [containing (in mm): 10 HEPES, pH 8.0, 0.1 EDTA, 1 phenylmethylsulfonyl fluoride, 20 NaF, 1 sodium pyrophosphate, and 1 Na₃VO₄ plus 25% glycerol, 0.4 m NaCl, 1 μg/ml leupeptin] and were incubated for 30 min at 4°C in a rotating wheel. Nuclear debris was removed by centrifugation at 900 × g for 20 min at 4°C; the supernatant was resolved in SDS-PAGE and immunoblotted with anti-p65 and anti-Sp1 antibodies.

Electrophoretic mobility shift assays. The double-stranded wild-type oligonucleotide used as an NF-κB probe was 5′-CTAGCAAAGGTAAGTCCCGGATCCAAAGGTAAGTCCCGGATCAAGGTAAGTCCC-GGC-3′ and for the mutant version was 5′-CTAGCAAAGGTAAGTAT TGTCGACAAAGGTAAGTTACGGATCAAGGTAAGTATTGGC-3′. Then 250 ng/μl of each oligonucleotide was annealed by incubation in STE buffer [containing (in mm): 100 NaCl, 10 Tris-HCl, pH 8.0, and 1 EDTA] at 80°C for 2 min. The mixture was cooled down slowly to 4°C, with a thermal profile of 1°C/min in a thermal incubator. Annealed oligonucleotides were diluted to 25 ng/μl in STE buffer. 5′-End labeling was performed with T4 polynucleotide kinase (Promega), using 25 ng of double-stranded oligonucleotide and 25 μCi [γ ³²P]-ATP (3000 Ci/mmol; Amersham Biosciences, Buckinghamshire, UK). The labeled probes were purified in a G-25 spin column (Amersham Biosciences). The binding reaction mixture contained 5 μg of nuclear protein extract diluted in a buffer containing, at final concentration, 40 mm HEPES, pH 8.0, 50 mm KCl, 0.05% Nonidet P-40, 1% dithiothreitol, and 10 μg/ml poly(dI-dC) in a total volume of 20 μl. Antibody (2 μg) or unlabeled competitor probe (25 ng) was added, and the reaction was incubated for 1 hr at 25°C. Labeled DNA (0.25 ng) was added to the mixture and was submitted to an additional 20 min of incubation at 25°C. Samples were resolved at 4°C in a 5% nondenaturing polyacrylamide gel in 0.5× Trisborate/EDTA buffer. After electrophoresis the gel was dried and autoradiographed.

Statistics. Student's t test was used to assess differences between groups. A p value < 0.05 was considered significant. Unless indicated, all experiments were performed at least three times with similar results. The values in the graphs correspond to the mean of at least three samples. Error bars indicate SD.

Results

To establish the conditions for analysis of ROS production under oxidative stress, we maintained PC12 cells in low-serum medium (1% fetal calf serum plus 1% horse serum) for 16 hr and then submitted the cells to 0.5 mm H₂O₂ for 3 hr. During the last hour of incubation the medium was supplemented with 2 μm HEt, a fluorescent probe most sensitive to anion superoxide, which with oxidation converts into ethidium and emits orange fluorescence (Bindokas et al., 1996; Chan et al., 1998). As shown in Figure 1A, counterstaining of these cells with the nuclear dye Hoechst 33258 and analysis by confocal microscopy evidenced HEt accumulation mostly in the nuclei of the cells submitted to oxidative damage. Then we quantified the incorporation of HEt by flow cytometry (Fig. 1B). Untreated cells exhibited a weak HEt incorporation. In contrast, H₂O₂ induced a significant increase in HEt fluorescence, consistent with the intracellular generation of ROS such as superoxide anion.

Next we analyzed the rate of HEt incorporation in PC12 cells submitted to H₂O₂ alone or in combination with reduced glutathione (GSH). As shown in Figure 1C, incubation with 0.5 mm aH₂O₂ for >2 hr resulted in a three- to fourfold accumulation of HEt fluorescence. Moreover, the addition of 10 mm GSH 30 min before H₂O₂ treatment prevented the oxidation of HEt, as expected from its antioxidant properties.

To determine the role of the PI3K/Akt pathway in protection against ROS toxicity, we used PC12 cells stably transfected with expression vectors for enhanced green fluorescence protein (EGFP), as a control, or an Akt fusion protein, myr-EGFP-Akt1, containing EGFP at the N terminus of mouse Akt1 and a myristoylation signal that targets this chimera to the plasma membrane and confers constitutive activation (Salinas et al., 2001). Figure 2A shows the endogenous 59 kDa Akt protein and the ectopically expressed 85 kDa myr-EGFP-Akt1 fusion protein in EGFP- and myr-EGFP-Akt1-transfected cells, respectively. In addition, as shown in Figure 2B, when these cells were placed in low-serum medium for 16 hr and analyzed with antibodies that specifically recognize the phosphorylated, active form of Akt, we observed that only myr-EGFP-Akt1 retains activity (for a more detailed description of these constructs, see Salinas et al., 2001).

The presence of EGFP in the control cells or the EGFP moiety in the myr-EGFP-Akt1 cells allowed for analysis of single living cells by flow cytometry in combination with the orange fluorescence of HEt-stained cells. Cells maintained under low-serum conditions for 16 hr were submitted to several H₂O₂ concentrations for 3 hr and stained with HEt during the last hour. As shown in Figure 2C, in the EGFP cells H₂O₂ induced a three- to fourfold increase in the HEt fluorescence in a concentration range of 0.5-1 mm. Interestingly, the myr-EGFP-Akt1 cells exhibited a lower basal level of HEt fluorescence and were much more refractory to HEt incorporation in the presence of H₂O₂. These results indicate that activation of the PI3K/Akt pathway exerts antioxidant effects in PC12 cells.

In additional flow cytometry experiments, we confirmed that the antioxidant effects elicited by the PI3K/Akt pathway correlated with protection against ROS-induced apoptosis. We analyzed exposure of phosphatidylserine on the outer plasma membrane leaflet as an early hallmark of apoptosis. This alteration was analyzed with Annexin-V-PE. To discriminate apoptosis from necrosis, we stained the cells simultaneously with the red fluorescent vital dye 7-AAD, which enters cells lacking membrane integrity and, therefore, allows for detection of necrotic cells (Hasper et al., 2000). EGFP and myr-EGFP-Akt1 cells were maintained under low-serum conditions for 16 hr and then were submitted to 0.5 or 1.5 mm H₂O₂ for 6 hr and finally were stained for 15 min with Annexin-V-PE and 7-AAD. As shown in Figure 2, D and E, in the absence of H₂O₂ most cells were Annexin-V-PE- and 7-AAD-negative, indicating that they were viable. After 0.5 mm H₂O₂ treatment ∼40% of the EGFP cells were Annexin-V-PE-positive and 7-AAD-negative, indicating that this population was undergoing apoptotic cell death. In contrast, in the myr-EGFPAkt1 cells this percentage was only ∼20%. On the other hand, after 1.5 mm H₂O₂ treatment almost the entire populations of EGFP and myr-EGFP-Akt1 cells were stained with both 7-AAD and Annexin-V-PE. Therefore, these results indicate that the PI3K/Akt pathway attenuates ROS-induced apoptosis, but not necrosis.

Because the oxidant fluorescent probe used in this study, HEt, is most sensitive for anion superoxide, we analyzed whether Akt exerts its antioxidant effect via the activation of Cu/Zn-SOD. As shown Figure 3, A and B, the basal messenger RNA and protein levels of Cu/Zn-SOD were elevated significantly in myr-EGFPAkt1 cells, suggesting that Akt upregulates Cu/Zn-SOD expression. Moreover, as shown in Figure 3C, flow cytometry analysis of EGFP and myr-EGFP-Akt1 cells immunostained with anti-Cu/Zn-SOD antibodies indicated that the levels of this enzyme were increased two- to threefold in cells expressing active Akt1.

To confirm these observations further, we established a PC12 cell line transfected with an expression vector for a conditionally active chimera of Akt, myr-Akt1-ER^*. In this protein the myristoylated Akt was fused at its C-terminal end with a mutated version of the hormone-binding domain from the estrogen receptor (see Materials and Methods). This fusion protein is expressed in an inactive form and becomes activated in the presence of 4-HT. Figure 4A shows the expression of endogenous 59 kDa Akt and the ectopically expressed 110 kDa fusion protein, and Figure 4B shows that the addition of 1 μm 4-HT resulted in the activation of this chimera as determined with activation-specific anti-phospho-Akt1 antibodies. Interestingly, as shown in Figure 4, C and D, when PC12 cells stably transfected with this construct were used to assess the levels of Cu/Zn-SOD, we observed that the addition of 4-HT for 16 hr induced the expression of both mRNA and protein of this antioxidant enzyme in the myr-Akt1-ER^* cells and did not have a significant effect on control vector-transfected PC12 cells.

In a previous study we reported that H₂O₂ induces a transient activation of Akt but, in turn, results in inactivation and caspase 3-dependent and -independent degradation of Akt (Martin et al., 2002). For that reason we analyzed the kinetics of Akt phosphorylation/dephosphorylation and proteolytic degradation together with the changes in Cu/Zn-SOD protein levels (Fig. 5). Control EGFP-transfected cells exhibited a rise in phospho-Akt1 within minutes after 0.5 mm H₂O₂ stimulation, but no phosphorylated/active Akt was detected after 2 hr. In myr-EGFP-Akt1-transfected cells a drop in phospho-Akt1 also was detected, but, in contrast to control cells, a significant amount of phospho-EGFP-Akt1 remained even after 4 hr. Moreover, control EGFP cells exhibited a dramatic loss of total Akt protein after 2 hr of incubation with H₂O₂, and Akt was no longer detectable after 4 hr. In myr-EGFPAkt1-transfected cells a drop in Akt1 levels also was detected, but again a significant amount of Akt still was found after 4 hr. In agreement with our previous study, these results suggest that membrane-targeted active Akt is more resistant to H₂O₂-induced dephosphorylation and degradation than the endogenous wild type (Martin et al., 2002). In this experimental frame we analyzed the changes in Cu/Zn-SOD protein levels. As shown in Figure 5, in EGFP cells H₂O₂ induced a slight increase in Cu/Zn-SOD after 1 hr, as expected (Yoo et al., 1999). However, the amount of Cu/Zn-SOD dropped to undetectable levels after 4 hr. The drop in Cu/Zn-SOD protein levels was prevented by preincubation with the nonselective caspase inhibitor z-Val-Ala-Aspfluoromethylketone (zVAD-fmk) (data not shown). In contrast, in myr-EGFP-Akt1 cells H₂O₂ did not increase Cu/Zn-SOD levels over the basal ones, which are already high, suggesting that the ectopic expression of active Akt saturated stimulation of Cu/Zn-SOD expression. Moreover, Cu/Zn-SOD levels dropped after long-term incubation with H₂O₂, but, in contrast to control cells, a significant amount still could be detected after 4 hr. Therefore, although these experiments do not demonstrate a causal relationship between regulation of Akt and induction of Cu/Zn-SOD by H₂O₂, they provide a mechanism that may be used, at least in part, by ROS to upregulate Cu/Zn-SOD via activation of Akt (see Discussion).

The regulatory effect of PI3K and Akt on expression of Cu/Zn-SOD also was analyzed by using a luciferase reporter construct carrying the luciferase gene under the control of 1499 bp from the 5′-flanking region of the human sod1 gene. As shown in Figure 6, cotransfection of this construct with membrane-targeted, active versions of PI3K (PI3K-CAAX) or Akt (myr-Akt1 and gag-Akt) resulted in a two- to threefold increase in luciferase activity over the level of cells cotransfected with a control vector carrying the muscarinic m1 receptor. Taken together, these results indicate that activation of the PI3K/Akt pathway increases the cellular antioxidant capacity at least in part via the upregulation of Cu/Zn-SOD expression.

There are several regulatory elements in the 5′ promoter region of human sod1 that might be targets of the PI3K/Akt pathway (Fig. 7A). To identify these targets, we generated several deletions of the 1499 bp sod1 promoter and subcloned them into the plasmid pGL3basic to govern luciferase expression. These luciferase reporter constructs were transfected with control vector or with an expression vector for membrane-anchored, active PI3K (PI3K-CAAX). As shown in Figure 7B, reporter constructs -1499 to -552 exhibited both basal and PI3K-inducible activity. In contrast, construct -355 exhibited basal activity, but no significant induction by PI3K. The smallest fragments -71 and -29 displayed low basal activity and little or no induction by PI3K. These results indicate the existence of a highly sensitive site for PI3K regulation between -355 and -552 bp of the human sod1 promoter. As shown in Figure 7A, within this region we identified a putative NF-κB-responsive element (analysis by TFsearch software, available at http://www.cbrc.jp/research/db/TFSEARCH.html).

Before the characterization of this site, we first verified the functional activation of NF-κB by the PI3K/Akt signaling pathway in our system in the context of NGF stimulation. Regulation of NF-κB-dependent transcription involves at least two mechanisms (Madrid et al., 2000): (1) modulation of the transactivating domain located in the C-terminal 120 amino acids of p65 and (2) release of transcription factor from the cytoplasmic inhibitor κB (IκB) and subsequent translocation to the nucleus. To analyze whether the PI3K/Akt pathway targets the transactivating domain (TD1) of p65-NF-κB, we used a fusion protein consisting of the p65-NF-κB TD1 and the DNA-binding domain of Gal4. PC12 cells were cotransfected with this expression vector (pGal4-p65) and a reporter plasmid controlling luciferase expression via four Gal4-responsive elements (pGal4-LUC). As shown in Figure 8A, in these cells NGF induced luciferase activity by more than twofold. Interestingly, preincubation with the PI3K inhibitor LY294002 significantly attenuated the NGF-induced luciferase expression. In additional experiments we analyzed the regulation of p65-NF-κB transactivation domain 1 in the background of EGFP and myr-EGFP-Akt1 cells. As shown in Figure 8A, when cells were cotransfected with pGal4-p65 and pGal4-LUC, we observed that myr-EGFP-Akt1 cells exhibited approximately fourfold more luciferase activity than EGFP cells. These results indicate that NGF regulates NF-κB activity in a PI3K/Akt-dependent manner by a mechanism directed, at least in part, to turn on the transactivation domain of p65-NF-κB. Because regulation of NF-κB also involves nuclear translocation, we compared the nuclear levels of p65-NF-κB after NGF stimulation in control EGFP cells and in myr-EGFP-Akt1 cells. As shown in Figure 8B, the basal levels of nuclear p65 were higher and those of cytosolic p65 were slightly lower in myr-EGFP-Akt1 cells as compared with control EGFP cells. Accordingly, basal IκBα levels were slightly lower in myr-EGFP-Akt1 cells. After 15 min of NGF stimulation EGFP cells exhibited a transient increase of nuclear p65-NF-κB, whereas the levels of the housekeeping transcription factor Sp1 were stable. This transient translocation is consistent with observations by other authors (Maggirwar et al., 1998). In contrast, NGF did not alter the distribution of p65-NF-κB in myr-EGFPAkt1 cells, suggesting either that Akt promotes close to maximal translocation or that Akt plays a substantial role as mediator of NGF-induced regulation of this transcription factor. Interestingly, although control EGFP cells exhibited higher basal IκBα levels than myr-EGFP-Akt1 cells, NGF did not alter these levels significantly under our experimental conditions, suggesting, in agreement with previous reports (Bui et al., 2001), that NGF induces at least in part the dissociation of p65-NF-κB from the NF-κB/IκB complex by a mechanism that does not necessarily require IκBα degradation (see Discussion).

Taken together, those results indicated that the PI3K/Akt pathway is necessary and sufficient for signaling to NF-κB and provided further support to analyze the regulation of sod1 via the newly identified NF-κB response element. Therefore, we used electrophoretic mobility shift assays (EMSAs) to determine whether the increase in nuclear p65-NF-κB was sufficient to promote binding of this transcription factor to the putative NF-κB sequence of the human sod1 promoter. Nuclear extracts from EGFP and myr-EGFP-Akt1 cells were incubated with a ³²P-labeled, double-stranded oligonucleotide comprising three tandem copies of the putative sod1 NF-κB sequence or a similar probe mutated to abolish NF-κB binding (see Materials and Methods). As shown in Figure 9A, among the several retarded complexes formed with the NF-κB probe, one of them was barely detectable in the mutated probe. In addition, the intensity of this retarded band was increased significantly in the myr-EGFP-Akt1 nuclear extracts as compared with the EGFP ones. Moreover, when an antibody against p65-NF-κB was included in the binding reaction, the intensity of this retarded band decreased dramatically, indicating the presence of p65-NF-κB in the complex. These results support the existence of a p65-NF-κB binding site in this region of the sod1 promoter and its upregulation by the PI3K/Akt pathway.

Finally, to analyze the functional regulation of the NF-κB binding site of sod1 by the PI3K/Akt pathway, we generated two luciferase reporter constructs containing three wild-type (NF-κB-sod1) or three mutated (NF-κB-mut) sequences of the 10 bp corresponding to the putative NF-κB regulatory element of human sod1 and linked them to the -29 luciferase sod1 reporter construct. As a positive control of NF-κB regulation, we also analyzed the promoter activity of a well established NF-κB regulatory sequence from human immunodeficiency virus (NF-κBHIV). As shown in Figure 9B, PC12 cells cotransfected with active PI3K-CAAX and either NF-κB-sod1 or NF-κB-HIV induced an approximately threefold increase in luciferase activity over the level of cells cotransfected with empty vector. In contrast, cells cotransfected with NF-κB-mut, p29-sod1, or pGL3basic yielded similar luciferase activity in the presence of PI3K-CAAX or empty vector. To demonstrate further a role of NF-κB as a mediator of PI3K regulation for this site, we used a dominant-negative mutant of the inhibitor of NF-κB, IκBα(S32A/S36A). This mutant contains two substitutions (S32A and S36A) that prevent its phosphorylation and subsequent proteasomal degradation, therefore exhibiting permanent sequestration of NF-κB (Traenckner et al., 1995). As shown in Figure 9C, IκBα(S32A/S36A) blocked the PI3K-CAAX-mediated increase both in NF-κB-HIV and in NF-κB-sod1 activities and did not have a significant inhibitory effect on the activities of NF-κB-mut, p29-sod1, and pGL3basic. Taken together, these results indicate that the PI3K/Akt pathway increases the cellular antioxidant defense at least in part via the upregulation of Cu/Zn-SOD expression at the level of the newly identified NF-κB regulatory site.

Discussion

In this study we present evidence that the PI3K/Akt survival pathway is involved in protection against oxidative damage by increasing the expression of Cu/Zn-SOD. Cells expressing active Akt1 kinase exhibited lower ROS levels in response to H₂O₂ and higher mRNA and protein levels of Cu/Zn-SOD. In addition to the regulation of Cu/Zn-SOD, we also observed mRNA and protein accumulation of other antioxidant enzymes, including Mn-SOD (data not shown) and heme oxygenase-1 (Salinas et al., 2003; Martin et al., 2004) in cells with an active PI3K/Akt pathway. Of note, several groups including ours have reported the activation of Akt in response to oxidative stress in a PI3K-dependent manner (Konishi et al., 1997; Shaw et al., 1998; Martin et al., 2002). According to these observations, it is tempting to consider the PI3K/Akt axis as a sensor for intracellular ROS levels that maintains ROS homeostasis via transcriptional activation of antioxidant enzymes.

Activation of Akt by H₂O₂ correlated with an increase in Cu/Zn-SOD levels. However, when cells were pretreated with the PI3K inhibitor LY294002, we did not detect a decrease in Cu/Zn-SOD expression (data not shown). Therefore, H₂O₂ induces sod1 via several overlapping mechanisms that may be Akt-dependent and -independent. For instance, Yoo et al. (1999) have reported the regulation of sod1 promoter by H₂O₂ at the level of hydrogen peroxide-responsive (PPRE) and heat shock (HSE) elements. These results illustrate (as shown in Fig. 7A) the complex transcriptional regulation of this gene that probably involves multiple positive and negative regulatory elements acting in concert.

Previous studies have reported the upregulation of Cu/Zn-SOD in response to growth factors and neurotrophins (Kong et al., 1993; Yoo et al., 1999; Chang et al., 2002; Park and Rho, 2002). Our results also suggest that sod1 is subjected to NGF control via PI3K/Akt/NF-κB signaling. Here we report for the first time the functional characterization of one NF-κB site in human sod1 and, most importantly, its regulation by the PI3K/Akt pathway. The identified sequence starts at position -486 and binds p65-NF-κB. Moreover, NGF activated the transactivation domain of p65-NF-κB in a PI3K/Akt-dependent manner and also produced a transient translocation of this transcription factor to the nucleus, therefore providing a mechanism to explain how Akt1 upregulates Cu/Zn-SOD expression. The relevance of NF-κB on the PI3K/Akt1-mediated regulation of Cu/Zn-SOD expression was demonstrated further by the fact that a dominant-negative IκBα mutant prevented the sod1 promoter activation by PI3K-CAAX. Interestingly, when we analyzed Cu/Zn-SOD protein levels after NGF stimulation, we could detect barely a small increase after 4 hr of treatment (data not shown). Although we do not have a conclusive explanation for this discrepancy, we suggest that regulation of sod1 expression is submitted to an exquisite homeostatic control via positive and negative regulatory elements, and, therefore, NGF-induced stimulation requires either a very strong/persistent activation of the PI3K/Akt pathway or cooperation with other factors.

The regulation of NF-κB by cytokine signaling, such as tumor necrosis factor-α (TNF-α), has been established in detail (Hoffmann et al., 2002). The current model indicates that, under non-stimulated conditions, p65-NF-κB is sequestered in the cytoplasm via association with protein inhibitors of κB (IκB). Cytokine stimulation promotes IκBα phosphorylation at serine residue 32 by specific kinases (IKK) and subsequent targeting for degradation by the ubiquitin-proteasome pathway. Therefore, this model provides a mechanism for translocation of p65-NF-κB to the nucleus and heterodimerization with other NF-κB members for transcriptional regulation. However, the regulation of NF-κB activity by NGF and by Akt remains controversial. For instance, Madrid et al. (2000) reported that Akt stimulates the transactivation potential but barely increases the nuclear translocation of p65-NF-κB in HEK 293T cells. Regarding NGF, Bui et al. (2001) reported the NGF-induced dissociation of the cytosolic p65/IκBα complex via phosphorylation of IκBα at tyrosine residue 42. Contrary to cytokine model, this mechanism results in nuclear translocation of p65-NF-κB without significant degradation of IκBα. On the other hand, Maggirwar et al. (1998) described that NGF-induced NF-κB contains heteromeric complexes of p65 and p50 members and that NGF induces degradation of IκBα, which is prevented by proteasome inhibitors. The discrepancies among these studies may be related in part to cell type specificity and methodological differences. For instance, the Maggirwar study reported the partial degradation of IκBα in cells isotopically labeled with ³⁵S-methionine. In agreement with this, we have observed a substantial decrease in IκBα levels in cells submitted to the protein synthesis inhibitor cycloheximide (data not shown). However, the studies that relied on direct immunoblot detection of total IκBα might have missed this observation because the expression of IκBα is upregulated by NF-κB as a negative feedback loop (Hoffmann et al., 2002), and, therefore, the change in IκBα levels is very transient in NGF-treated cells. The low rate of NGF-induced degradation of IκBα in comparison to that of TNF-α may argue against the relevance of such degradation in this scenario. In any case, our results in PC12 cells indicate that NGF induces the transactivating activity of p65-NF-κB in a PI3K-dependent manner and also induces its transient nuclear translocation by a mechanism that requires little or no contribution of IκBα degradation. In addition, the persistent Akt activation also results in a sustained increase of nuclear p65-NF-κB and a decrease of IκBα protein levels. Finally, our results are in agreement with all of the above mentioned studies in the observation that increased IκBα levels prevent activation of p65-NF-κB by Akt because the dominant-negative mutant IκBα(S32A/S36A) substantially attenuated Akt-mediated activation of the NF-KB regulatory site identified in the human sod1.

In the 5′-flanking region of human sod1 there are several putative binding sites for transcription factors that are regulated by phosphorylation and might be targets of PI3K and Akt. These include a CCAAT/enhancer binding protein (c/EBP), HSE, XRE, ARE NF-κB, and PPRE. Although some of these sites have been studied in detail (Minc et al., 1999; Yoo et al., 1999; Chang et al., 2002; Park and Rho, 2002), the regulation by NF-κB has remained unexplored. However, this site remains particularly attractive to study the regulation of Cu/Zn-SOD because NF-κB family members are controlled by cell redox status and are involved in cell survival (Mattson and Camandola, 2001; Karin and Lin, 2002; Wang et al., 2002; Storz and Toker, 2003). Therefore, the upregulation of the antioxidant enzyme Cu/Zn-SOD by the PI3K/Akt axis provides a connection between a transcription factor sensitive to redox regulation, such as NF-κB, and a survival pathway, represented by the PI3K and Akt.

Contrary to the observations reported in this study, Honda and Honda (2002) have reported in Caenorhabditis elegans that the absence of PI3K/Akt signaling results in activation of the Forkhead transcription factor homolog DAF-16 and causes the dauer phenotype, characterized by low metabolic rate and increased resistance to oxidative stress. From consistent studies in mammalian cells, Kops et al. (2002) have suggested that Akt and Forkhead transcription factors of the FOXO family might exert complementary functions in proliferating and quiescent cells. In agreement with this hypothesis, we suggest that the PI3K/Akt axis inversely regulates the subcellular distribution of NF-κB and FOXO transcription factors to adapt the level of antioxidant enzymes to specific metabolic situations. In the presence of neurotrophins, growth factors, or oxidative stress the activation of Akt results in the direct phosphorylation of FOXO factors, causing their displacement from the nucleus into the cytoplasm. Simultaneously, Akt induces translocation of NF-κB to the nucleus and activation of antioxidant genes. On the contrary, in quiescent cells, where Akt is not active, most of NF-κB is sequestered into the cytoplasm while nonphosphorylated active FOXO transcription factors localize to the nucleus and guard the proper expression of antioxidant genes. Together with Mn-SOD (Kops et al., 2002), Cu/Zn-SOD might be included among the antioxidant enzymes subjected to dual regulation by the pair NF-κB/FOXO because, in addition to the control exerted by PI3K/Akt/NF-κB, we have identified putative sites of regulation by Forkhead transcription factors in the human sod1 gene (starting at positions -1376, -1231, and -822) that might be targets of FOXO members under quiescent conditions.

Oxidative stress has been implicated in the neuronal cell death of several neurodegenerative diseases such as Alzheimer's and Parkinson's diseases. Strategies aimed to prevent the increase in ROS during the course of these disorders may attenuate or stop neuronal death and therefore may have significant clinical utility. This study suggests that activation of the PI3K/Akt/NF-κB pathway might reinforce the antioxidant cell capacity and provide protection against oxidant damage via the upregulation of antioxidant genes such as sod1.