CDDO-Me Distinctly Regulates Regional Specific Astroglial Responses to Status Epilepticus via ERK1/2-Nrf2, PTEN-PI3K-AKT and NFκB Signaling Pathways

2-Cyano-3,12-dioxo-oleana-1,9(11)-dien-28-oic acid methyl ester (CDDO-Me) is a triterpenoid analogue of oleanolic acid. CDDO-Me shows anti-inflammatory and neuroprotective effects. Furthermore, CDDO-Me has antioxidant properties, since it activates nuclear factor-erythroid 2-related factor 2 (Nrf2), which is a key player of redox homeostasis. In the present study, we evaluated whether CDDO-Me affects astroglial responses to status epilepticus (SE, a prolonged seizure activity) in the rat hippocampus in order to understand the underlying mechanisms of reactive astrogliosis and astroglial apoptosis. Under physiological conditions, CDDO-Me increased Nrf2 expression in the hippocampus without altering activities (phosphorylations) of phosphatase and tensin homolog deleted on chromosome 10 (PTEN), phosphatidylinositol-3-kinase (PI3K), and AKT. CDDO-Me did not affect seizure activity in response to pilocarpine. However, CDDO-Me ameliorated reduced astroglial Nrf2 expression in the CA1 region and the molecular layer of the dentate gyrus (ML), and attenuated reactive astrogliosis and ML astroglial apoptosis following SE. In CA1 astrocytes, CDDO-Me inhibited the PI3K/AKT pathway by activating PTEN. In contrast, CDDO-ME resulted in extracellular signal-related kinases 1/2 (ERK1/2)-mediated Nrf2 upregulation in ML astrocytes. Furthermore, CDDO-Me decreased nuclear factor-κB (NFκB) phosphorylation in both CA1 and ML astrocytes. Therefore, our findings suggest that CDDO-Me may attenuate SE-induced reactive astrogliosis and astroglial apoptosis via regulation of ERK1/2-Nrf2, PTEN-PI3K-AKT, and NFκB signaling pathways.


Introduction
Status epilepticus (SE) is an emergency neurological disorder showing prolonged seizure activities without spontaneous termination. SE has a 20-40% mortality rate and a 30% rate of neurological deficits [1]. SE leads to neuronal injury, alterations of neuronal networks, brain edema, and neuroinflammation, which cause various neurological complications, such as epilepsy and cognitive impairments [2][3][4]. SE also rapidly evokes reactive astrogliosis and astroglial apoptosis in regional specific manners independent of hemodynamics [5][6][7][8][9][10]. Reactive astrogliosis is a pathological hallmark of brain in epilepsy patients and various epilepsy models [11,12]. Although reactive astrogliosis is one of the healing processes after brain insults, it interferes with the functions of residual neuronal circuits [13]. In addition, reactive astrocyte and newly generated astrocytes after astroglial death show distinct properties, as compared to intact astrocytes, which are involved in acquisition of the physiological characteristics of the epileptic hippocampus [14,15]. This is because astrocytes maintain the brain microenvironment and homeostasis by regulating extracellular glutamate/ion concentration, brain-blood barrier, and energy metabolism [16,17]. Indeed, reactive astrogliosis and astroglial loss contribute to seizure generation via changes in responsiveness to glutamate, characteristics of Ca 2+ oscillations, and K + buffering [18][19][20][21][22]. Thus, the prevention of these astroglial responses to SE is one of the therapeutic strategies to inhibit abnormal neuronal synchronization and discharges, and the secondary undesirable post-SE complications [20,23].
Here, we demonstrate that CDDO-Me effectively attenuated reactive astrogliosis in the stratum radiatum of the CA1 region (referred to as CA1 astrocytes) and astroglial apoptosis in the molecular layer of the dentate gyrus (referred to as ML astrocyte below) following SE. These effects of CDDO-Me were relevant to the regulations of phosphatase and tensin homolog deleted on the chromosome 10 (PTEN)/PI3K/AKT pathway (in CA1 astrocytes), ERK1/2-mediated Nrf2 upregulation (in ML astrocytes), and p65 RelA NFκB phosphorylations (in both CA1 and ML astrocytes). Thus, these findings suggest that CDDO-Me may inhibit the regional-specific astroglial responses to SE by regulating diverse signaling molecules.

SE Induction and Electroencephalogram (EEG) Analysis
Two days after surgery, rats were injected intraperitoneally (i.p.) with LiCl (127 mg/kg). Next day (three days after surgery), rats were given atropine methylbromide (5 mg/kg, i.p.) to block the peripheral effect of pilocarpine. Twenty minutes after atropine methylbromide injection, rats were administered pilocarpine (30 mg/kg, i.p., i.p.). SE induction was stopped 2 h after pilocarpine injection by treatment of diazepam (10 mg/kg, i.p.). As needed, diazepam (10 mg/kg, i.p.) was repeatedly administered. To evaluate the effect of CDDO-Me on seizure activity induced by pilocarpine, EEG signals were measured with a DAM 80 differential amplifier (0.1-3000 Hz bandpass; World Precision Instruments, Sarasota, FL, USA). EEG was acquired (400 Hz) during a 2-h recording session and analyzed using LabChart Pro v7 (AD Instruments, New South Wales, Australia). The time point starting paroxysmal discharges that showed 4-10 Hz with 2 times higher amplitude than the basal level and lasted more than 3 s was defined as the time of seizure on-set. Spectrograms were automatically made by a Hanning sliding window with a 50% overlap [28,29].

Tissue Processing
Three days after SE induction, animals were perfused transcardially with phosphate-buffered saline (PBS, pH 7.4) followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4) under urethane anesthesia (1.5 g/kg i.p.). The brains were removed and post-fixed with the same fixative for 4 h. Subsequently, brain tissues were cryoprotected by PB containing 30% sucrose at 4 • C for 2 days. Consecutive coronal sections (30 µm) were made with a cryo-microtome and collected in six-well plates containing PBS. After decapitation under urethane anesthesia, the hippocampus was rapidly dissected out and homogenized in lysis buffer for Western blots. The protein concentration was calibrated using a Micro BCA Protein Assay Kit (Pierce Chemical, Dallas, TX, USA).

Immunohistochemistry and TUNEL Staining
Tissue sections were incubated with 10% normal goat serum (Vector, Burlingame, CA, USA) and subsequently with a mixture of primary antibodies in PBS containing 0.3% Triton X-100 at room temperature, overnight (Table 1). After washing, tissues were reacted with a fluorescein isothiocyanate (FITC) or Cy3-conjugated secondary antibodies (Vector, Burlingame, CA, USA) for 1 h at room temperature. For negative control, sections were reacted with pre-immune serum instead of primary antibody. An AxioImage M2 microscope and AxioVision Rel. 4.8 software (Carl Zeiss Korea, Seoul, Korea) were used for image capture and analysis of fluorescent intensity. Fluorescent intensity was measured and represented as the number of a 256-gray scale (15 sections per each animal). The intensities were corrected by subtracting the average values of background noise obtained from 5 image inputs. To analyze the astroglial apoptosis, terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining was applied according to the manufacturer's instructions (Upstate, Lake Placid, NY, USA) prior to immunofluorescent study. Areas of interest in the dentate gyrus (1 × 10 5 µm 2 , 15 sections per each animal) were selected, and cell count of TUNEL-positive cells was performed using AxioVision Rel. 4.8 software.

Western Blot
After electrophoresis and transfer, membranes were blocked overnight at 4 • C with 2% bovine serum albumin (BSA) in Tris-buffered saline (TBS; in mM 10 Tris, 150 NaCl, pH 7.5, and 0.05% Tween 20) and subsequently incubated in primary antibodies (Table 1). After washing, membranes were incubated for 1 h at room temperature in a solution containing horseradish peroxidase (HRP)-conjugated secondary antibodies. Immunoblots were quantified by membrane scanning in an ImageQuant LAS4000 system (GE Healthcare Korea, Seoul, South Korea). Optical densities of proteins were measured by the protein/β-actin ratio. The ratio of phosphoprotein to total protein was described as the phosphorylation level.

Quantification of Data and Statistical Analysis
The number (n) of each experimental group used for the evaluation was seven. The data obtained from each group were analyzed. After evaluating the values on normality using Shapiro-Wilk W-test, all results were analyzed by the two-tailed Student t-test, repeated one-way ANOVA, or one-way ANOVA to determine statistical significance. The Newman-Keuls test was used for post hoc comparisons. A p-value below 0.05 was considered statistically significant.

CDDO-Me Increases Nrf2 Expression and Attenuates Reactive CA1 Astrogliosis and Apoptosis of ML Astrocytes Following SE
Under physiological conditions, CDDO-Me increased Nrf2 expression in the whole hippocampus (p < 0.05 vs. vehicle, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 1A,B and Supplementary Figure S1). Consistent with our previous study [28], CDDO-Me did not affect the seizure latency and its severity in response to pilocarpine (repeated one-way ANOVA; n = 7; Figure 1C,D). Following SE, total Nrf2 expression was decreased in the whole hippocampus (p < 0.05 vs. vehicle, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 1A,B and Supplementary Figure S1). In addition, CA1 astrocytes showed hypertrophy and upregulated GFAP expression (reactive astrogliosis; p < 0.05 vs. control animals, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 2A,B). In addition, TUNEL staining showed SE-induced apoptosis of ML astrocytes (p < 0.05 vs. control animals, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 2A,C-E). As compared to control animals, Nrf2 expression was reduced in reactive CA1 astrocytes and remaining ML astrocytes (p < 0.05 vs. control animals, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 2A-C). CDDO-Me effectively ameliorated the reduced Nrf2 expression in CA1 and ML astrocytes, and attenuated reactive CA1 astrogliosis and ML astroglial apoptosis following SE (p < 0.05 vs. vehicle, one-way ANOVA followed by Newman-Keuls post hoc test and two-tailed Student t-test; n = 7, respectively; Figure 2A-E). These findings suggest that CDDO-Me may upregulate Nrf2 expression and prevent reactive CA1 astrogliosis as well as astroglial degeneration following SE.

CDDO-Me Ameliorates Reactive CA1 Astrogliosis by Regulating the PTEN/PI3K/AKT Pathway
Nrf2 expression and its activity are regulated by the PI3K/AKT signaling pathway [34,35]. Therefore, we investigated whether CDDO-Me mitigates astroglial responses by influencing this pathway. In the whole hippocampus, CDDO-Me did not influence expressions and phosphorylations of PI3K and AKT under physiological condition, which were unaffected by SE (one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 3A,B and Supplementary Figure S1). Following SE, CDDO-Me reduced phosphorylation levels of PI3K/AKT in the whole hippocampus without altering their expression levels (p < 0.05 vs. control animals and vehicle, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 3A,B and Supplementary Figure S1). These findings indicate that CDDO-Me may inhibit PI3K/AKT signaling cascades. Since PTEN negatively regulates the PI3K/AKT pathway [36], we also validated the effect of CDDO-Me on PTEN. SE did not affect PTEN expression and its phosphorylation (one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 3A,B and Supplementary Figure S1). However, CDDO-Me reduced PTEN phosphorylation without changing its expression level under the post-SE condition but not the physiological condition (p < 0.05 vs. control animals and vehicle, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 3A,B). Considering that PTEN phosphorylation represents its inactivation [37], these findings indicate that CDDO-Me may inhibit the PI3K/AKT pathway by increasing PTEN activity under the post-SE condition.
To further confirm the roles of AKT in Nrf2 expression and astroglial responses following SE, we applied 3CAI (an AKT inhibitor) and SC79 (an AKT activator) prior to SE induction. Consistent with our previous study [31], 3CAI attenuated reactive CA1 astrogliosis but deteriorated ML astroglial apoptosis following SE, which were reversed by SC79 (p < 0.05 vs. vehicle, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 4A-E). However, neither 3CAI and nor SC79 changed astroglial Nrf2 expression (one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 4A-C). These findings indicate that CDDO-Me-mediated PI3K/AKT inhibition may attenuate SE-induced reactive CA1 astrogliosis, although this pathway may not be involved in ML astroglial apoptosis and upregulation of Nrf2 expression.

CDDO-Me Attenuates Reactive CA1 Astrogliosis and Astroglial Apoptosis by Inhibiting NFκB Phosphorylation
NFκB phosphorylation plays an important role in the maintenance of its optimal activity [38]. In particular, NFκB S311 phosphorylation increases its transcriptional activity and anti-apoptotic function [39]. In contrast, S468 phosphorylation terminates NFκB-dependent gene expression upon assisting in binding of an E3 ubiquitin ligase complex to NFκB [40]. In addition, reactive CA1 astrogliosis and ML astroglial apoptosis are differently regulated by NFκB phosphorylation following SE: NFκB-S311 phosphorylation modulates reactive CA1 astrogliosis, while its S468 phosphorylation is involved in ML astroglial apoptosis [31,41]. Since CDDO-Me directly inhibits NFκB activity [24,[42][43][44], it is likely that CDDO-Me may attenuate reactive CA1 astrogliosis and ML astroglial apoptosis by inhibiting NFκB phosphorylation. In the present study, SE increased NFκB-S311 and S468 phosphorylation in reactive CA1 astrocytes but only-S468 phosphorylation in ML astrocytes, which were abrogated by CDDO-Me (p < 0.05 vs. control animals and vehicle, one-way ANOVA followed by Newman-Keuls post hoc test; n = 7, respectively; Figure 5A-D). Therefore, our findings suggest that CDDO-Me may ameliorate SE-induced astroglial responses by inhibiting NFκB phosphorylation.

Discussion
Following brain insults, astrocytes show hypertrophy and proliferation in the affected region, which is termed reactive astrogliosis [47]. Reactive astrogliosis is one of the scar tissues that inhibit dendritic and axonal remodeling in neuronal circuits. Reactive astrocytes also influence neural viability and synaptogenesis after brain injury by releases of growth factors and trophic factors [48,49]. SE-induced reactive astrogliosis originates from distinct sources and different pathways in the hippocampus: gliogenesis in the dentate gyrus and in situ proliferation in the stratum radiatum of the CA1 region [5,14]. Thus, the mechanisms that regulate reactive astrogliosis are complex and remain elusive. Recently, we reported that PI3K/AKT inhibition attenuates reactive CA1 astrogliosis following SE [31]. In the present study, CDDO-Me abolished SE-induced reactive CA1 astrogliosis concomitant with reduced PI3K/AKT phosphorylation. These findings are consistent with previous reports demonstrating CDDO-Me-induced AKT inhibition [50][51][52][53]. Furthermore, our findings demonstrate that CDDO-Me-activated (dephosphorylated) PTEN, which leads to the constitutive PI3K/AKT inhibition [54]. The present data also reveal that 3CAI (an AKT inhibitor) also inhibited reactive CA1 astrogliosis, which were reversed by SC79 (an AKT activator). Similar to 3CAI, U0126 attenuated reactive CA1 astrogliosis [31]. In the present study, however, U0126 co-treatment did not enhance the inhibitory effect of CDDO-ME on reactive CA1 astrogliosis. Since the PI3K/AKT pathway plays an important role in regulating cell division and viability [55,56], these findings indicate that CDDO-Me may mitigate reactive CA1 astrogliosis through the PTEN/PI3K/AKT system rather than the ERK1/2 pathway.
On the other hand, neuronal death is one of the powerful signals that induces reactive astrogliosis [57,58]. Since CDDO-Me selectively attenuates CA1 pyramidal cell loss induced by SE [28], it cannot be excluded the possibility that abrogation of reactive CA1 astrogliosis would result from the neuroprotective effect of CDDO-Me. However, recent studies reveal that reactive astrogliosis is irrelevant to neuronal death [30,32,33]. Furthermore, AKT inhibition does not affect CA1 neuronal death induced by SE [30], and CDDO-Me did not affect the PI3K/AKT pathway under physiological condition in the present study. Therefore, our findings indicate that CDDO-Me may ameliorate reactive CA1 astrogliosis by regulating the PTEN/PI3K/AKT signaling pathway, independent of CA1 neuronal loss.

Conclusions
In conclusion, the present study reports a novel effect of CDDO-Me on reactive astrogliosis and astroglial apoptosis induced by SE, and demonstrates its underlying molecular mechanisms mediated by the PTEN/PI3K/AKT, ERK1/2, and NFκB signaling pathways. Briefly, CDDO-Me ameliorated reactive CA1 astrogliosis by enhancing PTEN-mediated PI3K/AKT inhibition. In addition, it attenuated ML astroglial apoptosis via an increase of ERK1/2-mediated Nrf2 upregulation. CDDO-Me also abated reactive astrogliosis and astroglial apoptosis by abolishing NFκB phosphorylation ( Figure 8). Therefore, our findings propose the availability of CDDO-Me and its derivates for various neurological diseases relating to reactive astrogliosis and astroglial degeneration. CDDO-Me attenuates reactive CA1 astrogliosis by activating PTEN, which inhibits PI3K/AKT activities. CDDO-Me also abates reactive CA1 astrogliosis by reducing NFκB S311 and S468 phosphorylation. In addition, CDDO-Me ameliorates ML astroglial apoptosis via upregulation of ERK1/2-mediated Nrf2 expression and inhibition of NFκB S468 phosphorylation. Therefore, CDDO-Me mitigates both reactive astrogliosis and astroglial degeneration induced by SE.

Conflicts of Interest:
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.