IL-10 produces a dual effect on OGD-induced neuronal apoptosis of cultured cortical neurons via the NF-κB pathway

As a classic immunoregulatory cytokine, interleukin-10 (IL-10) can provide in vivo and in vitro neuroprotection respectively during cerebral ischemia and after the oxygen-glucose deprivation (OGD)-induced injury. However, its role in cortical neuronal survival at different post-ischemic phases remains unclear. The current study found that IL-10 had distinct effects on the neuronal apoptosis at different OGD stages: at an early stage after OGD, IL-10 promoted the OGD-induced neuronal apoptosis in the cultured primary cortical neurons by activating p65 subunit, which up-regulated Bax expression and down-regulated Bcl-xL expression; at a late OGD stage, however, it attenuated the OGD-induced neuronal apoptosis by activating c-Rel, which up-regulated Bcl-xL expression and down-regulated Bax expression. The early-stage pro-apoptosis and late-stage anti-apoptosis were both partly abolished by PDTC, an NF-κB inhibitor, and promoted by PMA, an NF-κB activator. The optimal anti-apoptotic effect appeared when the cultured neurons were treated with IL-10 at 9-24 h after OGD. Taken together, our findings suggest that IL-10 exerts a dual effect on the survival of the cultured neurons by activating the NF-κB pathway at different stages after OGD injury and that PMA treatment at a late stage can facilitate the IL-10-conferred neuroprotection against OGD-induced neuronal injury.


INTRODUCTION
As a major cause of mortality and disability worldwide, stroke often ensues from impaired blood supply to the brain [1]. After cerebral ischemia, the subsequent inflammatory activation and release of inflammatory cytokines can result in neuronal apoptosis, which ultimately induces severe neurologic deterioration [1][2][3].
As a prototypical anti-inflammatory cytokine, interleukin-10 (IL-10) can suppress pro-inflammatory signals and immune responses [4,5]. It has been demonstrated to attenuate the neuronal damage after cerebral infarction [4][5][6]. Studies with IL-10-deficient mice have also reported a larger infarct volume and higher pro-inflammatory cytokine release after permanent middle cerebral artery occlusion (pMCAO) [4,7]. This anti-inflammatory effect has also been observed in the molecular line of research. In vitro studies have found that the treatment with IL-10 can produce a concentration-dependent protection for the cultured primary cortical neurons that are exposed to oxygenglucose deprivation (OGD), in which the neuronal apoptosis can be reduced by up-regulating Bcl-2 expression and down-regulating Bax expression [8,9]. However, it remains sparsely understood at which postischemic stage IL-10 produces its neuronal protective effect.
In neurons, IL-10 can activate the nuclear factor kappa B (NF-κB) [10,11]. The latter, as a ubiquitous transcriptional factor, has been documented to be involved in the processes of inflammation and apoptosis and can be activated during the development of cerebral ischemia [12], but findings regarding its role in ischemic injury are complex and conflicting. Some studies have reported that NF-κB can promote cell apoptosis by up-regulating the expression of proapoptotic factors including Bax [13][14][15][16]; whereas, others report it as a survival factor for neurons in cerebral ischemia by up-regulating the Bcl-xL expression and down-regulating the Bax expression in neurons [10,11,[17][18][19]. Still more, NF-κB has been reported to perform a dual role in cerebral ischemia, which shows that hypoxia-ischemia induced two peaks of cerebral NF-κB activity: an early NF-κB activation contributing to neuronal damage by down-regulating Bcl-2 and Bcl-xL expression and a late NF-κB activation producing a neuronal protection by increasing Bcl-2 and Bcl-xL expression [12]. But it remains unclear why NF-κB has a dual effect on the brain damage at different post-ischemic stages.
In fact, NF-κB is composed of five subunits of the Rel family, including p65 (RelA), c-Rel, RelB, NF-κB1 (p50), and NF-κB2 (p52). Among these five subunits, p65 and c-Rel are closely related to the neuronal survival [20,21] and act as opposite regulators of neuronal vulnerability to ischemia. The activation of p65 contributes to the ischemic neuronal damage, while activated c-Rel promotes neuronal resistance to hypoxia [21]. In vivo studies have documented that p65 is mainly activated in neurons at an early stage after the middle cerebral artery occlusion (MCAO), which contributes to the post-ischemic neurodegenerative process [22]. This finding has been confirmed by in vitro studies, which evidence that p65, but not c-Rel, is highly activated in neurons 3 hours after OGD [21]. Therefore, the mediation of the expression of p65 subunit and c-Rel subunit may be involved in the observed dual effect of NF-κB on cerebral ischemia.
Taken together, given the findings and facts: the activation of NF-κB by IL-10 and the dual role of NF-κB activation in neuronal damage after hypoxiaischemia, it would be intriguing and rewarding to speculate whether IL-10 can produce a dual effect on OGD-injured cultured cortical neurons at different postischemic stages and whether p65 subunit and c-Rel subunit are involved in its hypothesized effects.
In the present work, we exposed cortical neurons to OGD injury to build a cerebral ischemic model in vitro and treated them with IL-10 and an NF-κB inhibitor (PDTC) and an activator (PMA) at different stages. We found that IL-10 produced a dual effect on the apoptosis of cortical neurons at different OGD stages by mediating the p65 pathway at an early stage and c-Rel pathway at a late stage, and that IL-10 administration showed an optimal neuronal protective effect at a middle-to-late stage after OGD. These findings provide new insight into the role of IL-10 in cerebral ischemia and its potential therapeutic value.

OGD induces peaks of p65 and c-Rel in the cultured primary cortical neurons at different OGD stages
As the literature indicates that hypoxia-ischemia can induce 2 peaks of NF-κB activity at 3 to 6 h and 24 h in the cerebral cortex of P7 rats after unilateral carotid artery occlusion and hypoxia [12] and that p65, rather than c-Rel, is rapidly activated in neurons after MCAO [22] and OGD [21], it is important to determine the levels of p65 and c-Rel at different time points after OGD. To elucidate the changes of p65 and c-Rel after OGD injury, their levels were measured at 0 h, 3 h, 6 h, 12 h, 24 h and 36 h after OGD. The experimental groups were designed as follows: Control group, OGD/R0h group, and OGD/R3h group, OGD/R6h group, OGD/R12h group, OGD/R24h group, OGD/R36h. The analyses showed that compared with the Control group, the OGD/R0h group reported a higher expression of p65 (1.26±0.06 vs. 0.31±0.05, p<0.001) ( Figure 1A, 1B); compared with the OGD/R0h group, OGD/R6h group, OGD/R12h group, OGD/R24h group, and OGD/R36h group, the OGD/R3h group reported a higher expression of p65 ( Figure 1A, 1C). These results showed that the expression peak of p65 appeared at 3 h after OGD injury and that of c-Rel activity did at 24 h after OGD injury.

IL-10 increases OGD-induced apoptosis in cultured primary cortical neurons via p65 at an early OGD stage
Further experiments were conducted to elucidate the mechanism of IL-10-induced early pro-apoptotic effect on neurons after OGD injury. The neurons were simultaneously treated with IL-10 and PDTC (a specific inhibitor of NF-κB) or PMA (an activator of NF-κB) immediately after OGD to further investigate whether p65 pathway is involved in the early pro-apoptotic effect of IL-10. Three hours after drug treatment, the medium was changed to a neuronal medium which contained no IL-10, PDTC and PMA. The groups were designated as follows: Control group, OGD group, OGD+IL-10/R0-3h group and OGD+IL-10/R0-3h+PDTC/R0-3h group, OGD+IL-10/R0-3h+PMA/R0-3h group. After three hours of drug treatment, western blotting was employed to evaluate the protein expression of p65. Forty-eight hours after OGD, flow cytometry assay and western blotting were performed to evaluate apoptosis.
Western blot results also indicated that neurons exposed to the OGD treatment showed a higher expression of Bax but a lower expression of Bcl-xL when the OGD group was compared with the Control group (Bax, 0.65±0.06 vs. 0.21±0.03, p<0.01; Bcl-xL, 1.14±0.07 vs. 1.50±0.08, p<0.01) ( Figure 3B, 3D, 3E). When compared with the OGD group, the OGD+IL-10/R0-3h group displayed a significant increase in Bax expression and a marked decrease in Bcl-xL expression (Bax, These results demonstrate that IL-10 activates p65 expression and accelerates OGD-induced apoptosis in neurons by up-regulating the expression of proapoptotic protein Bax and down-regulating the expression of the anti-apoptotic protein Bcl-xL at an early stage after OGD injury. These effects can be partially blocked by NF-κB inhibitor PDTC but facilitated by NF-κB promoter PMA.

IL-10 suppresses OGD-induced apoptosis in cultured primary cortical neurons via c-Rel at a late OGD stage
In order to further explore the mechanism of IL-10induced anti-apoptotic effect on neurons at a late stage after OGD injury, IL-10, PDTC, and PMA were simultaneously added into OGD-injured neurons at 21 h after OGD to further investigate whether c-Rel pathway is involved in the anti-apoptotic effect of IL-10 at a late stage after OGD injury. Three hours after drug treatment, the medium was changed to a neuronal medium which contained no IL-10, PDTC and PMA. The groups were as follows: Control group, OGD group, OGD+IL-10/R21-24h group and OGD+IL-10/R21-24h+PDTC/R21-24h group, OGD+IL-10/R21-24h+PMA/R21-24h group. After three hours of drug treatment, western blot analysis was adopted to evaluate the protein expression of c-Rel. Forty-eight hours after OGD, flow cytometry assay and western blotting were performed to evaluate the apoptosis of each group.
Then the expression of Bax and Bcl-xL was detected by western blotting to evaluate the apoptosis level of each group. We found that the OGD treatment increased the expression of Bax, but decreased the expression of Bcl-xL when the OGD group was compared with the Control group (Bax, 2. These results indicate that the treatment with IL-10 at 9-24 h after OGD produces the strongest anti-apoptotic effect by inducing a lower expression of Bax and a higher expression of Bcl-xL in the cortical neurons.

DISCUSSION
In this study, we demonstrated that p65 was rapidly activated and peaked at 3 h after OGD injury while c-Rel was activated at a late stage and peaked at 24 h after OGD injury. The study also found IL-10 increased neuronal apoptosis if it was administrated at 0-3h after OGD but prevented neuronal apoptosis if it was used at 9-12 h or 21-24 h after OGD. Of interest, IL-10 administration at 0-3 h after OGD significantly activated p65 and promoted apoptosis by enhancing Bax expression and reducing Bcl-xL expression in cortical neurons, which were partly abolished by NF-κB inhibitor PDTC but facilitated by NF-κB promoter PMA; IL-10 administration at 21-24 h after OGD, however, activated c-Rel and attenuated the apoptosis by decreasing Bax expression and increasing Bcl-xL expression in neurons, which were also partly abolished by PDTC but facilitated by PMA. Moreover, the IL-10 treatment at 9-24 h after OGD results in the strongest anti-apoptotic effect by inducing a lower expression of Bax and a higher expression of Bcl-xL in cortical neurons.
Ischemic brain damage has been documented to induce 2 peaks of NF-κB activity in the brain after ischemia [12]. Other studies report that 2 hours after MCAO, activated p65 is mainly found in neurons [22,23]. Another study also demonstrates that 3 hours after OGD, p65 is highly activated but c-Rel is not changed in neurons [21]. Adding to the existing findings, the current study provides direct evidence to support that in the cultured cortical neurons, p65 subunit, but not c-Rel subunit, is activated immediately after OGD injury, which peaked at 3 h after OGD while c-Rel subunit is significantly activated at a late stage, rather than an early stage, after OGD injury, which peaked at 24 h after OGD. These results show that p65 and c-Rel can be activated and peak at different stages in OGD-injured neurons.
As a traditional protective factor in cerebral ischemia, IL-10 has been found to suppress local inflammatory reactions and decrease neuronal damages in vivo and in vitro [4-6, 8, 9]. After the ischemic stroke, it can reduce the infarct volume by 40% in the IL-10 transgenic mice when compared with that of the wild type mice [5]. Our previous studies also show that IL-10 promotes neuronal survival and neurite outgrowth after OGD via the JAK1/STAT3 and PI3K/Akt pathway [8,9]. However, different from the traditional view, the current study found that IL-10 increased neuronal apoptosis if administrated at 0-3 h after OGD but reduced the neuronal apoptosis if administered at 9-12 h or 21-24 h after OGD. These results indicate that IL-10 has a dual effect on neuronal survival after OGD injury.
p65 activity contributes to ischemic neuronal damage, while c-Rel promotes neuronal resistance to hypoxia [21]. In the current study, we demonstrated that p65 was mainly activated at an early OGD phase and c-Rel at a late phase. According to the facts above, we further speculated that p65 and c-Rel were involved in the observed dual effect of IL-10 on neuronal survival at different OGD stages. To verify this speculation, we conducted experiments at an early stage and at a late stage, respectively.
At the early OGD stage, we found that IL-10 increased the expression and nuclear translocation of p65 in OGDinjured neurons. Moreover, it promoted OGD-induced apoptosis in neurons by up-regulating the expression of pro-apoptotic protein Bax and down-regulating the antiapoptotic protein Bcl-xL expression. Of note, NF-κB inhibitor PDTC decreased the neuronal apoptosis by preventing IL-10-increased Bax expression and IL-10reduced Bcl-xL expression. NF-κB activator PMA increased the neuron apoptosis by promoting IL-10increased Bax and IL-10-reduced Bcl-xL expression. These results indicate that p65 is involved in the proapoptotic effect of IL-10 at an early OGD stage, which is in line with the previous findings. The existing literature documents that neurotoxic stimuli, such as ischemia [21,23], glutamate [26], β-amyloid [27,28], or 1-methyl-4phenylpyridinium (MPP + ) [29,30], can induce p65 activation and the transcription of a panel of pro-apoptotic genes in neurons. In cerebral ischemia, p65 activity is associated with an imbalanced expression of pro-apoptotic target genes, such as an increased expression of Bim, Noxa and Bax genes [21,23]. Bax, as a pro-apoptotic protein, plays an important role in the apoptosis of neurons during the cerebral ischemia [31] and Bcl-xL, as an anti-apoptotic protein, is crucial for decreasing neuronal apoptosis in cerebral ischemia [32], which are further confirmed in the current study. Taken together, the results from the current study reveal that IL-10 remarkably enhances the apoptosis of neurons at an early OGD stage by activating p65, which in turn up-regulates Bax expression and down-regulates Bcl-xL expression and that the inhibition of the early p65 activity by PDTC reduces the IL-10-induced neuronal damage after OGD injury.
Previous research shows that diverse neurotoxic settings, such as S100B in models of NMDA-mediated excitotoxicity [33], agonists at mGlu5 receptors against β-amyloid [27] and MPP + toxicity [29] or IL-1β against glutamate-mediated cell death [20], can activate the c-Rel expression, which in turn up-regulates the expression of anti-apoptotic genes, facilitating the neuroprotection. Moreover, further evidence shows that the inhibition of c-Rel can decrease Bcl-xL expression against hypoxia-induced cell death in the hippocampus [34]. Also, in cortical neurons, Leptin exerts neuroprotection through c-Rel-dependent transcription of Bcl-xL against OGD-induced apoptosis [35]. Nevertheless, little is known about the effect of c-Rel activation on Bax expression. In this study, we found that IL-10 increased c-Rel expression and c-Rel nuclear translocation in OGD-injured neurons at a late OGD stage. Moreover, IL-10 reduced OGD-induced apoptosis in neurons by down-regulating the expression of pro-apoptotic protein Bax and up-regulating the antiapoptotic protein Bcl-xL expression. Besides, PDTC increased neuronal apoptosis by preventing IL-10decreased Bax expression and IL-10-increaced Bcl-xL expression, and PMA decreased neuron apoptosis by promoting IL-10-decreased Bax expression and IL-10increaced Bcl-xL expression. Taken together, these findings demonstrate that IL-10 remarkably reduces neuronal apoptosis at a late OGD stage by activating c-Rel, which in turn up-regulates Bcl-xL expression and down-regulates Bax expression and that the activation of a late c-Rel activity by PMA facilitates the IL-10conferred neuroprotection of OGD-injured neurons.
We have found that IL-10 promotes neuronal survival at 9-12 h and 21--24 h after OGD injury. We wondered whether IL-10 administration at 9-24 h induced a stronger neuronal protection and found that the apoptotic rate at 9-24h markedly declined when compared with that at 9-12 h and 21-24 h. The IL-10/R9-24h group also showed a lower expression of Bax and a higher expression of Bcl-xL than the IL-10/R9-12h or IL-10/R21-24h group. Taken together, the current study demonstrates that for OGD-injured neurons, the administration of IL-10 at a middle-to-late OGD stage can produce the strongest neuronal protective effect. Therefore, IL-10 treatment should be withheld until a middle-to-late stage after ischemic insult in order to preserve protective NF-κB-dependent mechanisms that occur in a later ischemic phase. Clinical applications of IL-10 should be restricted to cases of well-documented acute cerebral infarction, in which detailed records of timing of the insult are available. Our study suggests that with proper timing, IL-10 treatment may be a very effective strategy to combat brain damage after stroke.
On the basis of the above analysis and discussion, the current study concludes that IL-10 treatment can exert an early pro-apoptotic and a late anti-apoptotic effect on the OGD-injured neurons respectively by activating p65 subunit at an early OGD phase and c-Rel subunit at a mid-to-late OGD stage. The optimal neuroprotection appears when IL-10 is administered at 9-24 h after OGD injury. Different from the traditional view of the neuroprotective effect of IL-10, these findings broaden our understanding of the role of IL-10 in cerebral ischemia and provide empirical evidence for its therapeutic value. In treating cerebral ischemia, an early administration of IL-10 can produce adverse effect and a later application in combination with NF-κB activator PMA is recommended. Meanwhile, there are some limitations in our study: we only testified the dual effect of IL-10 on neuronal survival in vitro. Future studies should focus on its effects in vivo. Nevertheless, the study extends the understanding of the neuroprotective effect of IL-10 and suggests its potential therapeutic strategies against cerebral ischemia.

Animals
Pregnant Sprague Dawley rats were provided by the Animal Center of Fujian Medical University (Fuzhou, China), housed in a controlled environment (23 ±1℃ and 50% ± 5% humidity) and allowed food and water ad libitum. The room was lighted between 8:00 a.m. and 8:00 p.m. All experimental animals were euthanized with isoflurane, which contained 3% induction and 1.5% maintenance in 30% O2 and 70% N2O. Experimental protocols followed the Guidelines of the National Institute of Health (NIH Publications No. 80-23, revised in 1996) and were approved by Institutional Animal Care and Use Committee of Fujian Medical University.

Primary cortical neuron cultures
Primary cortical neurons were cultured as described previously [36,37]. Samples of cerebral cortex from brains of Sprague-Dawley rat embryos (aged 16-18 days) were dissected and trypsinized cells were cultured in a neurobasal medium (Gibco, USA) with 2% B27 (Gibco, USA) supplement, 0.5mM glutamine, 50units/ml penicillin. After 24 hours of incubation in a chamber at 37 °C with 5% CO2, the neurobasal medium was refreshed. The purity of the neuronal cultures was confirmed by class III-β -Tubulin and Hoechst 33342 staining, which indicated about 90% of cultured neurons.

Oxygen-glucose deprivation/reoxygenation
The oxygen-glucose deprivation (OGD) model was established as described previously with minor modifications [38]. Cells were cultured for seven days before exposure to OGD. In brief, the neurobasal medium was replaced with a glucose-free DMEM (Gibco, USA) and then maintained in an incubator containing 5% CO2 and 95% N2 at 37°C to induce oxygen deprivation. Two hours later, OGD was terminated by replacing the glucose-free DMEM with the original medium and then cells were incubated in their original culture condition for reoxygenation. Control cells were not exposed to OGD conditions.

Drug treatment
IL-10 from rat recombinant was purchased from PeproTech. It was administered to the cultured cortical neurons after OGD, reaching a final concentration of 20 ng/ml. For neurons treated with IL-10, PDTC (Abcam, UK), a specific inhibitor of NF-κB [39,40], or PMA (Abcam, UK), an activator of NF-κB [41,42], was simultaneously administered, respectively reaching a final concentration of 20uM and of 20ng/ml. The OGD group was vehicle-treated in the right time after OGD.

Flow cytometry using Annexin V/PI staining
To detect the apoptosis of neurons quantitatively, we performed flow cytometry as described previously with modifications [43]. Fluorescein Annexin V-FITC/PI double labeling was performed with an Annexin V-FITC apoptosis detection kit (Beyotime, China). Briefly, neurons were seeded in culture flasks (25cm²). Forty-eight hours after OGD, the cells were stained with Annexin V-FITC and PI according to the instruction of the manufacturer. The apoptotic cells were determined with a flow cytometer (Beckton Dickinson, USA). The nuclear translocation of p65 and c-Rel was quantified using the Fiji/Image J [44,45].

Assessment of cell survival by MTT extraction assay
Cell survival was quantified by the tetrazoliumsaltextraction method as previously described [47]. Because viable cells with active mitochondria can cleave the tetrazolium ring into a visible dark-blue formazan reaction product, 3,(4,5-dimethyldiazol-2yl)2,5-diphenyl-tetrazolium bromide (MTT) was added to the culture medium at a final concentration of 5 mg/ml and incubated at 37°C for 4 hours and OD was measured at 570 nm using a microplate reader (Multiskan FC, Thermo, USA).

Immunofluorescence staining
Immunofluorescence staining was performed to locate the expression p65 and c-Rel as described previously [48,49]. It was applied to cells 3 hours after the drug treatments as described. Cells were washed in PBS at 37°C and fixed on 4% paraformaldehyde. Then corresponding antibodies were used to incubate neurons at 4℃ overnight. After three washes with PBS, the neurons were further incubated with secondary antibodies at room temperature for 2 hours. Nuclei were stained by hoechst33342 (5 μ g/ml, Sigma, USA) and visualized under a ZEISS LSM 780 (Carl Zeiss, Germany).

Statistical analysis
All data were presented as mean±SEM and analyzed with SPSS17.0 statistical software (IBM, USA). Each measurement was performed in replication of 3 independent trials. One-way analysis of variance (ANOVA) was performed for statistical significance among groups; Student-Newman-Keuls multiple comparison test was conducted when equal variances were assumed and Dunnett's T3 was applied when equal variances was not assumed; unpaired two-tailed Student's t-tests were employed to determine the significance of mean differences between two groups. P values less than 0.05 (two-sided) were considered as statistically significant.

ACKNOWLEDGMENTS
We wish to thank Professor Hongzhi Huang, Dr. Xiaodong Pan and Dr. Lin Fan for language revision. We also thank assistant researcher Yunmei Huang and Zuanfang Li from Fujian Academy of Integrative Medicine (Fujian University of Traditional Chinese Medicine) for technical guidance in immunefluorescence staining.