Preconditioning-Activated AKT Controls Neuronal Tolerance to Ischemia through the MDM2–p53 Pathway

We previously described the impact of MDM2–p53 interaction on neuronal susceptibility to ischemia [34] and IT [33]. Here, we explored the potential role of AKT on IPC-mediated neuroprotection as a candidate to be involved in the MDM2–p53 pathway. First, we confirmed that ischemia promoted AKT activation in neurons, as evidenced by AKT phosphorylation at Ser⁴⁷³. As shown in Figure 1A, short (20 min) OGD significantly induced p(Ser⁴⁷³)AKT and MDM2 expression, whereas AKT protein levels remained unaltered (Figure S1A). However, AKT phosphorylation was not observed when neurons were subjected to prolonged OGD (90 min). Moreover, MDM2 protein levels were lower, which is consistent with the higher expression of p53 protein as shown in Figure 1A.

The time-dependent upregulation of Mdm2 expression after OGD (Figure 1B) confirms that subacute ischemia may be important to induce mechanisms that prevent p53 stabilization after OGD, as previously described [33]. We used short OGD (20 min) followed by 2 h of reoxygenation as a model of IPC (Figure S1B) [33]; thus, we analyzed neuronal extracts collected at 4 h of reoxygenation after OGD (OGD/R) or after OGD preceded by the IPC protocol (IPC + OGD/R). In parallel, neurons were incubated in normoxia (Nx) or preconditioning (IPC) settings (Figure S1B).

As shown in Figure 1C and Figure S1C, IPC induced the early activation of AKT, as revealed by phosphorylation at Ser⁴⁷³ [35], followed by MDM2 protein stabilization and phosphorylation at Ser¹⁶⁶. IPC also prevented p53 stabilization induced by OGD/R (Figure 1C). Interestingly, immunofluorescence images shown in Figure 1D revealed that IPC promoted AKT phosphorylation at Ser⁴⁷³ in neurons, which predominantly accumulated in the nucleus, and decreased p53 stabilization after OGD/R (IPC + OGD/R), when compared with non-preconditioned neurons (OGD/R). Consequently, IPC prevented neuronal apoptosis and caspase-3 activation caused by OGD/R, as measured by flow cytometry (Figure S1D) and fluorimetry assays (Figure S1E), respectively. To confirm the role of p53 in IPC-mediated neuroprotection, we used neurons expressing (wild-type; wt) or not (knockout; ko) p53 protein. Our results show that neurons lacking p53 (Figure S1F) were more resistant to OGD-induced apoptosis than p53 wt neurons. Moreover, apoptosis levels in p53KO neurons were similar to those observed in preconditioned (IPC + OGD/R) wt neurons (Figure S1G), thus confirming the key role of p53 destabilization in IPC-mediated neuroprotection [33]. Our results show that IPC induced neuroprotection against an ischemic insult through a mechanism that involves phosphorylation of AKT at Ser⁴⁷³, MDM2 stabilization and phosphorylation at Ser¹⁶⁶, and p53 destabilization.

The PI3K/AKT signaling pathway is involved in neuronal IT both in vitro [36] and in vivo [37]. However, the role of IPC-mediated activation of PI3K/AKT pathway in the regulation of the MDM2–p53 complex remains unexplored. In order to clarify this, neurons were incubated with the irreversible and specific inhibitor of the PI3K/AKT pathway, wortmannin [19]. As shown in Figure 2A, wortmannin abrogated IPC-enhanced (Ser⁴⁷³)AKT phosphorylation and p53 destabilization, as shown in Figure 1D. These results suggest a direct link between AKT activation and inhibition of p53-mediated neuronal apoptosis (Figure S1D) and caspase-3 activation (Figure S1E) induced after OGD/R. The main regulator of p53 stabilization, MDM2, is a target of AKT [26], which phosphorylates MDM2 at Ser¹⁶⁶ and Ser¹⁸⁶ [26]. PI3K inhibition with wortmannin prevented the phosphorylation of both (Ser⁴⁷³)AKT and (Ser¹⁶⁶)MDM2 induced by IPC (Figure 2B). The specific AKT-mediated phosphorylation of MDM2 at Ser¹⁶⁶ induced by IPC was confirmed by using a small interfering RNA (siRNA) specifically designed against AKT1 protein (siAkt), highly expressed in cortical neurons, and whose activity is essential for neuronal survival after ischemia [38]. As shown in Figure 3, siAkt reduced total AKT and p(Ser⁴⁷³)AKT protein levels at day 3 after transfection, both in HEK-293T cells (Figure 3A) and in cortical neurons (Figure 3B). Moreover, AKT knockdown (siAkt) prevented (Ser¹⁶⁶)MDM2 phosphorylation (Figure 3B). These results demonstrate that the IPC-activated PI3K/AKT signaling pathway promotes MDM2 phosphorylation at Ser¹⁶⁶, which may be responsible for MDM2 stabilization and consequent p53 destabilization after ischemic injury.

The activation of AKT has been involved in nuclear translocation of MDM2 in tumor cells [26]. Considering the relevance of nuclear MDM2 stabilization for neuronal survival after ischemia [34] and, more specifically, its neuroprotective role in IPC [33], we decided to further investigate the relevance of PI3K/AKT signaling pathway in the regulation of subcellular localization of MDM2 protein. Thus, neurons or HEK-293T cells were transfected with human MDM2-tagged protein (MDM2-GFP). Representative blots of transfected HEK-293T cells and images from neurons ectopically expressing human MDM2 protein after four different experimental conditions (Nx, IPC, OGD/R, and IPC + OGD/R) are shown in Figure 4A and Figure S1H, respectively. Ectopic expression of MDM2-GFP confirmed that IPC promotes MDM2 nuclear accumulation compared with non-preconditioned ischemic (OGD/R) or normoxic (Nx) neurons (Figure 4A,B), as revealed by the quantification of nuclear/cytosolic fluorescence ratio (Figure S2B) and nuclear fluorescence intensity of MDM2-GFP (Figure S2C).

Moreover, confocal immunofluorescence analysis revealed that IPC induced the colocalization of p(Ser⁴⁷³)AKT and endogenous MDM2 within the nucleus, whereas this effect was not observed under OGD/R condition (Figure 5A,B). The maximal intensity of endogenous nuclear MDM2 was higher in preconditioned (89.1% IPC and 90.3% IPC + OGD/R) than in non-preconditioned neurons (43.7% Nx and 52.6% OGD/R). Values for p(Ser⁴⁷³)AKT protein were also higher in preconditioned neurons (96.5% in IPC and 82.7% in IPC + OGD/R) compared to non-preconditioned neurons (39.8% in Nx and 54.9% in OGD/R) (Figure S2C).

Furthermore, neuron treatment with wortmannin or siAkt before the IPC or OGD/R protocol impaired MDM2 nuclear translocation after OGD/R, when compared with control neurons (Figure 5A and Figure S2D). Indeed, both treatments (wortmannin or siAkt) prevented the IPC-increased mean percentage of nuclear MDM2 maximal fluorescence intensity (39.4% and 45%) and p(Ser⁴⁷³)AKT (15.3% and 20%) after OGD/R, respectively (Figure 5B). Thus, inactivation of the AKT signaling pathway results in MDM2 protein accumulation in the cytoplasm of preconditioned neurons after ischemia. These results highlight the relevance of IPC-enhanced nuclear translocation and stabilization of MDM2 mediated by the activation of the PI3K/AKT pathway, which might confer protection against ischemic injury.

Following demonstration of the role of IPC-enhanced activation of PI3K/AKT pathway in the nuclear stabilization of MDM2, we further investigated whether p(Ser⁴⁷³)AKT and MDM2 interacted within the nucleus (Figure 6A). MDM2 immunoprecipitation from nuclear protein extracts, followed by immunoblotting against MDM2 and p(Ser⁴⁷³)AKT, revealed that IPC promoted the interaction between p(Ser⁴⁷³)AKT and MDM2, after OGD/R, thus preventing OGD/R-induced nuclear p53 stabilization, as shown in the nucleus input (Figure 6A).

Lastly, we studied the detrimental effect of PI3K/AKT pathway disruption on neuronal apoptosis (Figure 6B). AKT inhibition counteracted the protective effect of IPC prior to OGD, which confirms the neuroprotective role of AKT–MDM2 in the context of IT. Our results, thus, demonstrate that IPC induces phosphorylation and activation of AKT, which promotes MDM2 phosphorylation at Ser¹⁶⁶ and nuclear translocation, where it interacts with p(Ser⁴⁷³)AKT. This mechanism may contribute to enhanced nuclear stabilization of MDM2, which plays an essential role in IPC-induced ischemic tolerance.

Neuronal cultures were prepared from C57Bl/6J or p53-null (Tp53^−/−, B6.129S2, The Jackson Laboratory) mouse embryo (14.5E) cortices. Neurons were seeded at 1.8 × 10⁵ cells/cm² in Neurobasal medium supplemented with 2% B27 and 2 mM glutamine (Invitrogen, Madrid, Spain) and incubated at 37 °C in a humidified 5% CO₂-containing atmosphere [55].

After 9–10 days in vitro (DIV), neurons were exposed to oxygen and glucose deprivation (OGD) by incubating cells at 37 °C for 90 min in an incubator equipped with an airlock and continuously gassed with 95% N₂/5% CO₂. The incubation medium (buffered Hanks’ solution without glucose: 5.26 mM KCl, 0.43 mM KH₂PO₄, 132.4 mM NaCl, 4.09 mM NaHCO₃, 0.33 mM Na₂HPO₄, 2 mM CaCl₂, and 20 mM HEPES, pH 7.4) was previously gassed with 95% N₂/5% CO₂ for 30 min. Under these conditions, oxygen concentrations in the incubation medium were 6.7 ± 0.5 μM as measured with a Clark-type oxygen electrode [56,57]. When indicated, the neurons were exposed to ischemic preconditioning (IPC; short OGD for 20 min followed by 2 h of reoxygenation) prior to a subsequent prolonged ischemia (OGD, 90 min) and 4 h of reoxygenation (IPC + OGD/R) (Figure S1B). In parallel, neurons were incubated in normoxia (Nx) at 37 °C in a humidified atmosphere of 95% air/5% CO₂ or ischemic preconditioning (IPC). When indicated, neurons were incubated 30 min before IPC in buffered Hanks’ solution (pH 7.4), in the absence or presence of wortmannin (100 nmol/L), as described previously [19].

Neurons (8 DIV) or HEK-293T cells were transfected with a plasmid vector expressing YFP tagged Mdm2 from MDM2 human promoter. MDM2p/Mdm2-YFP was a gift from Uri Alon & Galit Lahav (Addgene plasmid # 53962, Watertown, MA, USA) [58]. When required, an empty vector (pYFP) was used as control in the same conditions. Plasmid transfection was performed using Lipofectamine^® LTX (Invitrogen, Carlsbad, MA, USA), according to the manufacturer’s instructions. Cells were transfected with 1.5 µg/µL of the plasmid vectors and used after 24 h. AKT knockdown in 6 DIV neurons was achieved by transfection with small interfering double-stranded ribonucleotides (siRNA). Targeted sequences were as follows: 5′–CUCAAGUACUCAUUCCAGAtt–3′, antisense: 5′–UCUGGAAUGAGUACUUGAGgg–3′(mouse, s62216, corresponding to nucleotides 1006–1025, GenBank accession number NM_009652↗) [59]. As a negative control, we used Silencer™ Select Negative Control No. 1 siRNA (siControl). All siRNAs were purchased from Ambion^®, Invitrogen^®, Thermo Fischer Scientific (Offenbach, Germany). According to the degree of protein knockdown, the efficiency of transfection of siRNA was estimated to be 70–80% at 3 days post transfection. For silencing experiments, neurons were transfected with siRNA (10 nM) using Lipofectamine^® RNAiMAX (Invitrogen), following the manufacturer’s instructions. Neurons were further incubated in Neurobasal medium for 72 h before their use.

Neurons were carefully detached from the plates using 1 mM EDTA tetrasodium salt in PBS (pH 7.4) and were stained with annexin V/APC and 7-AAD, performed exactly as previously described [55].

Caspase-3 activity was assessed in cell lysates [33] and according to the manufacturer’s instructions using the Fluorimetric Assay kit CASP3F from SIGMA and read at emission at a wavelength of 405 nm. The method is based on the release of the fluorescent 7-amino-4-methyl coumarin (AMC) moiety. The AMC concentration is calculated using an AMC standard.

Neurons were lysed in buffer containing 1% SDS, 2 mM EDTA, 150 mM NaCl, 12.5 mM Na₂HPO₄, and 1% Triton X-100 (NP40: 1% NP40, EDTA diK⁺ 5 mM, Tris pH8 20 mM, NaCl 135 mM, and 10% glycerol) supplemented with phosphatase inhibitors (1 mM Na₃VO₄ and 50 mM NaF) and protease inhibitors (100 mM phenylmethylsulfonyl fluoride, 50 µg/mL anti-papain, 50 µg/mL pepstatin, 50 µg/mL amastatin, 50 µg/mL leupeptin, 50 µg/mL bestatin, and 50 µg/mL soybean trypsin inhibitor), stored on ice for 30 min and boiled for 5 min. Aliquots of lysed extracts were subjected to SDS polyacrylamide gel (MiniProtean^®, Bio-Rad) and blotted with antibodies overnight at 4 °C. Antibodies used were anti-AKT (9272), anti-p(Ser⁴⁷³)AKT (9271), anti-cleaved caspase-3 (Asp175, 9661) (Cell Signaling, Danvers, MA, USA), anti-p53 (554157, BD Biosciences), anti-MDM2 (2A10, ab-16895), anti-p(Ser¹⁶⁶)MDM2 (ab131355), anti-GFP (ab290; also detects YFP) (Abcam, Cambridge, UK), anti-LAMIN B (sc-374015, Santa Cruz Biotechnology, Heidelberg, Germany), and anti-GAPDH (Ambion, Cambridge, UK) overnight at 4 °C. After incubation with horseradish peroxidase-conjugated goat anti-rabbit IgG (Pierce, Thermo Scientific) or goat anti-mouse IgG (Bio-Rad), membranes were immediately incubated with enhanced chemiluminescence SuperSignal West Dura (Pierce) for 5 min before exposure to Kodak XAR-5 film for 1 to 5 min and the autoradiograms were scanned. Band intensities were quantified using ImageJ 1.48v software, as described previously [60]. For the co-immunoprecipitation assay, neurons were lysed in ice-cold buffer containing 50 mM Tris (pH 7.5), 150 mM NaCl, 2 mM EDTA, 1% NP-40) supplemented with phosphatase inhibitors described above. After clearing debris by centrifugation, neuronal lysates (100 mg) were incubated with 1 mg of the antibody for 24 h at 4 °C followed by the addition of 10 mL of protein A-agarose (GE Healthcare Life Sciences) for 2 h at 4 °C. Immunoprecipitates were extensively washed with lysis buffer and resolved by SDS-PAGE and immunoblotted with indicated antibodies [61]. The relative protein abundances are shown in Figure S1. Full blots and gel scans are included in Figure S3.

Neurons were grown on glass coverslips and fixed with 4% (w/v, in PBS) paraformaldehyde for 30 min and immunostained with rabbit anti-phosphoAKT (Ser⁴⁷³; 9271; Cell Signaling, MA, USA), mouse anti-MDM2 (2A10, ab-16895), mouse anti-MAP2 (1:500; M#1406, Sigma-Aldrich, St. Louis, MO, USA) [55], mouse anti-p53 (1:200; 554157, BD Pharmingen, San Diego, CA, USA), and anti-GFP (1:1000; ab290; also validated to detect YFP). Immunolabeling was detected using secondary antibodies anti-rabbit IgG–Cy3 or anti-mouse IgG–Cy2 (1:500; Jackson ImmunoResearch. Cambridge, UK). Nuclei were stained with 4′,6-diamidino-2 phenylindole (DAPI; D9542, Sigma-Aldrich). Coverslips were washed, mounted in SlowFade light antifade reagent (Invitrogen) on glass slides, and examined using a microscope (Nikon Inverted microscope Eclipse Ti-E, (NY, USA) equipped with 40× objective, a pre-centered fiber illuminator Nikon Intensilight C-HGFI, and a black-and-white charge-coupled device digital camera Hamamatsu ORCAER or a scanning laser confocal microscope (“Spinning Disk” Roper Scientific Olympus IX81, Tokyo, Japan) with three lasers 405, 491, and 561 nm, equipped with 40×, 63×, and 100× PL Apo oil-immersion objective for high-resolution imaging and device digital camera Evolve Photometrics. All microscope settings were set to collect fluorescent images below saturation and were kept constant for all images taken in the experiment. Images were analyzed with the ImageJ 1.48v software (National Institutes of Health). The percentage of p(Ser⁴⁷³)AKT⁺ and p53⁺ neurons and the quantification of maximal protein fluorescence intensity of p(Ser⁴⁷³)AKT and p53 are shown in Figure S2A. In MDM2-GFP-transfected neurons, the nucleocytoplasmic distribution of MDM2-GFP was calculated as the ratio of the nuclear mean fluorescence to the cytoplasmic mean fluorescence of endogenous MDM2, measured in 24 neurons (six neurons per condition in four different neuronal cultures) (Figure S2B) [62]. To quantify the maximal nuclear fluorescence intensity of endogenous MDM2 staining and pSer⁴⁷³AKT, 40 neurons (10 neurons per condition in four different cultures) were measured (Figure S2C), as described previously [44]. In the representative cross-sectional intensity profiles shown in Figure 5B, the percentage of p(Ser⁴⁷³)AKT and MDM2 indicated below each condition was quantified as nuclear mean fluorescence. In all cases, nuclei were identified by DAPI staining. The maximal nuclear MDM2 fluorescence intensity in neurons treated with wortmannin or siAkt is shown in Figure S2D.

Experimental results were evaluated by one-way analysis of variance, followed by the Bonferroni post hoc test, used to compare values between multiple groups. The results are expressed as means ± SEM. Student’s t-test was used for comparisons between two groups of values. In all cases, p < 0.05 was considered significant (* p < 0.05 versus Nx; # p <0.05 versus OGD). Statistical analyses were performed using SPSS Statistics 24.0 for Macintosh (IBM).