CDDO-Me Abrogates Aberrant Mitochondrial Elongation in Clasmatodendritic Degeneration by Regulating NF-κB-PDI-Mediated S-Nitrosylation of DRP1

Clasmatodendrosis is a kind of astroglial degeneration pattern which facilitates excessive autophagy. Although abnormal mitochondrial elongation is relevant to this astroglial degeneration, the underlying mechanisms of aberrant mitochondrial dynamics are still incompletely understood. Protein disulfide isomerase (PDI) is an oxidoreductase in the endoplasmic reticulum (ER). Since PDI expression is downregulated in clasmatodendritic astrocytes, PDI may be involved in aberrant mitochondrial elongation in clasmatodendritic astrocytes. In the present study, 26% of CA1 astrocytes showed clasmatodendritic degeneration in chronic epilepsy rats. 2-cyano-3,12-dioxo-oleana-1,9(11)-dien-28-oic acid methyl ester (CDDO-Me; bardoxolone methyl or RTA 402) and SN50 (a nuclear factor-κB (NF-κB) inhibitor) ameliorated the fraction of clasmatodendritic astrocytes to 6.8 and 8.1% in CA1 astrocytes, accompanied by the decreases in lysosomal-associated membrane protein 1 (LAMP1) expression and microtubule-associated protein 1A/1B light-chain 3 (LC3)-II/LC3-I ratio, indicating the reduced autophagy flux. Furthermore, CDDO-Me and SN50 reduced NF-κB S529 fluorescent intensity to 0.6- and 0.57-fold of vehicle-treated animal level, respectively. CDDO-Me and SN50 facilitated mitochondrial fission in CA1 astrocytes, independent of dynamin-related protein 1 (DRP1) S616 phosphorylation. In chronic epilepsy rats, total PDI protein, S-nitrosylated PDI (SNO-PDI), and SNO-DRP1 levels were 0.35-, 0.34- and 0.45-fold of control level, respectively, in the CA1 region and increased CDDO-Me and SN50. Furthermore, PDI knockdown resulted in mitochondrial elongation in intact CA1 astrocytes under physiological condition, while it did not evoke clasmatodendrosis. Therefore, our findings suggest that NF-κB-mediated PDI inhibition may play an important role in clasmatodendrosis via aberrant mitochondrial elongation.


Introduction
In the brain, astrocytes maintain a state of balance in the acid-base equilibrium, energy metabolism, and neuronal activity [1]. Therefore, the dysfunction of astrocytes is one of the causes of neurological diseases, including epilepsy [2,3]. Vacuolized degeneration represents an irreversible fatal change in astrocytes [4], which was first reported by Alzheimer in 1910 and termed "clasmatodendrosis" by Cajal [5]. Clasmatodendrosis is evoked by failure of bioenergetics and acidosis coupled to impaired mitochondrial functions under pathophysiological conditions, such as Alzheimer disease, brain ischemia, and epileptic seizures [3,[6][7][8][9][10][11]. In previous studies, we have reported that clasmatodendritic astrocytes show the upregulations in lysosomal-associated membrane protein 1 (LAMP1), beclin-1, and microtubule-associated protein 1A/1B light-chain 3 (LC3) expression, which indicates the activation of autophagic processes [12,13]. Basically, autophagy is the process to remove aberrant components for cell survival [14]. However, excessive autophagy evokes type II programmed cell death, known as autophagic cell death [15]. Indeed, clasmatodendritic astrocytes include condensed chromatin, lysosomes, and large membrane-bound osmiophilic cytoplasmic inclusions [16,17]. Therefore, we have reported for the first time that clasmatodendrosis is one of the astroglial degenerations which is relevant to excessive autophagy [12,13].
Protein disulfide isomerase (PDI) is an oxidoreductase regulating protein folding in the endoplasmic reticulum (ER) [21]. Interestingly, PDI plays the role of a nitric oxide (NO) donor to DRP1 and exerts its S616 phosphorylation, which facilitates mitochondrial fission (fragmentation) in CA1 neurons [22]. Furthermore, PDI expression is downregulated in clasmatodendritic astrocytes compared to intact astrocytes [23]. Therefore, it is likely that downregulation of PDI may be involved in aberrant mitochondrial elongation in clasmatodendritic astrocytes.
Here, we demonstrate that CDDO-Me and SN50 (a NF-κB inhibitor) ameliorated clasmatodendrosis in CA1 astrocytes in the hippocampus of chronic epilepsy rats accompanied by the reduced NF-κB S529 phosphorylation. CDDO-Me and SN50 facilitated mitochondrial fission in CA1 astrocytes, independent of DRP1 S616 phosphorylation. Both CDDO-Me and SN50 increased total PDI protein and S-nitrosylated (SNO)-PDI levels in CA1 astrocytes of chronic epilepsy rats. PDI knockdown resulted in mitochondrial elongation in intact CA1 astrocytes under physiological conditions, while it did not evoke clasmatodendrosis. Therefore, our findings suggest that NF-κB-mediated PDI inhibition may regulate clasmatodendritic degeneration via aberrant mitochondrial elongation.

CDDO-Me and SN50 Restore the Reduced PDI Expression in CA1 Astrocytes
PDI is a chaperone in ER and also presents in cytoplasm and cell surface [21]. Although PDI acts as thiol-disulfide exchanger, it also plays a role as a transporter of NO residue [33]. In particular, we have reported that PDI regulates S-nitrosylation of DRP1 and accelerates mitochondrial fission [22] since S-nitrosylation induces DRP1 dimerization, which directly increases its activity [34]. Considering that PDI expression is downregulated in clasmatodendritic astrocytes compared to intact astrocytes [23], it is likely

CDDO-Me and SN50 Restore the Reduced PDI Expression in CA1 Astrocytes
PDI is a chaperone in ER and also presents in cytoplasm and cell surface [21]. Although PDI acts as thiol-disulfide exchanger, it also plays a role as a transporter of NO residue [33]. In particular, we have reported that PDI regulates S-nitrosylation of DRP1 and accelerates mitochondrial fission [22] since S-nitrosylation induces DRP1 dimerization, which directly increases its activity [34]. Considering that PDI expression is downregulated in clasmatodendritic astrocytes compared to intact astrocytes [23], it is likely that CDDO-Me and SN50 may restore the downregulated PDI expression and recover aberrant mitochondrial elongation in clasmatodendritic astrocytes.

PDI Knockdown Leads to Mitochondrial Elongation in CA1 Astrocytes under Physiological Condition
To elucidate whether the downregulated PDI protein level abrogates DRP1-medaited mitochondrial fission in CA1 astrocytes, we applied PDI siRNA to normal rats. Compared to control siRNA, PDI siRNA decreased PDI protein level in the CA1 region (t(12) = 15.2, p < 0.001 vs. control siRNA, Student t-test, n = 7; Figure 8A,B and Supplementary Figure S2)

Discussion
The major findings in the present study are that CDDO-Me and SN50 attenuated clasmatodendrosis in CA1 astrocytes by restoring PDI-mediated mitochondrial fission, suggesting that NF-κB-mediated PDI downregulation may lead to clasmatodendrosis in CA1 astrocytes (Figure 9).
The major findings in the present study are that CDDO-Me and SN50 attenuated clasmatodendrosis in CA1 astrocytes by restoring PDI-mediated mitochondrial fission, suggesting that NF-κB-mediated PDI downregulation may lead to clasmatodendrosis in CA1 astrocytes (Figure 9).
Clasmatodendrosis shows autophagic phenomena in response to epileptic seizures through NF-κB S529 phosphorylation [12,13], which is closely relevant to the synchronous epileptiform discharges [3,11]. Under stressful condition, the cell organelles enveloped by double membrane vesicles (autophagosomes) and delivered to the lysosomes for digestion and the subsequent recycle of amino acid into the cell machinery [35]. However, dysregulation of this autophagic process leads to non-apoptotic programmed cell death [14,36]. Although NF-κB phosphorylation is involved in the regulation of autophagic process [37], little data are available to explain the downstream effectors of NF-κB-mediated clasmatodendrosis.
Clasmatodendrosis shows autophagic phenomena in response to epileptic seizures through NF-κB S529 phosphorylation [12,13], which is closely relevant to the synchronous epileptiform discharges [3,11]. Under stressful condition, the cell organelles enveloped by double membrane vesicles (autophagosomes) and delivered to the lysosomes for digestion and the subsequent recycle of amino acid into the cell machinery [35]. However, dysregulation of this autophagic process leads to non-apoptotic programmed cell death [14,36]. Although NF-κB phosphorylation is involved in the regulation of autophagic process [37], little data are available to explain the downstream effectors of NF-κB-mediated clasmatodendrosis.
PDI is responsible for modulating disulfide bond formation [42]. Under ER stress, PDI removes misfolded proteins to maintain the ER homeostasis [43,44]. Thus, upregulation of PDI expression is an adaptive response to protect cells from ER stress [45,46]. However, PDI also induces mitochondrial membrane permeabilization leading to apoptosis [47]. Consistent with a previous study [23], the present data demonstrate that PDI expression was downregulated in clasmatodendritic CA1 astrocytes. Since upregulated PDI expression results in acquisition of tolerance against detrimental stress in astrocytes [46] and excessive unfolded protein response (UPR) induced by accumulating misfolding and aggregation of proteins leads to apoptosis or autophagy [48,49], it is not surprising that PDI downreg-ulation may be involved in clasmatodendrosis in CA1 astrocytes. In the present study, interestingly, CDDO-Me and SN50 increased PDI expression in CA1 astrocytes, concomitant with the abrogation of NF-κB S529 phosphorylation. These findings indicate that the NF-κB S529 phosphorylation may trigger clasmatodendrosis by reducing PDI expression.
On the other hand, PDI deletion mitigates neuroinflammation after traumatic brain injury (TBI) in mice, which is associated with the decreased NF-κB phosphorylation [50]. However, the roles of PDI in NF-κB activity are still controversial. PDI ablation inactivates NF-κB phosphorylation [51]. In contrast, over-expression of PDI suppresses NF-κB activity [52]. Furthermore, tumor necrosis factor-α (TNF-α)-stimulated NF-κB signaling is unaffected by PDI knockdown [21]. Considering that the TNF-α neutralization attenuates clasmatodendritic astrocytes accompanied by reduced p65/RelA-Ser529 phosphorylation [13], it is likely that NF-κB may be an upstream regulator for PDI rather than its downstream effector for activity at least during clasmatodendritic process.
In addition to the phosphorylation, S-nitrosylation of the cysteine (C) 644 site enhances DRP1 GTPase activity and its oligomerization in association with excessive mitochondrial fission [34]. S-nitrosylation is a post-translational modification induced by the covalent binding of NO to a cysteine thiol group of the protein. This modification affects a variety of cellular processes, protein function and protein-protein interactions [62]. Although S-nitrosylation inhibits PDI function [63,64], PDI acts as NO donor of DRP1 to exert S616 phosphorylation, which facilitates mitochondrial fission (fragmentation) [22]. In the present study, PDI expression is downregulated in clasmatodendritic astrocytes compared to intact astrocytes. Furthermore, CDDO-Me and SN50 restored the decreased PDI expression in clasmatodendritic astrocytes and increased SNO-PDI level in the CA1 region without total DRP1 protein level. Therefore, our findings indicate that the reduction in PDI-mediated S-nitrosylation of DRP1 may result in aberrant mitochondrial elongation in CA1 astrocytes during clasmatodendritic process. Indeed, PDI knockdown led to mitochondrial elongation in CA1 astrocytes under physiological condition. Taken together, our findings suggest that NF-κB S529 phosphorylation may decrease PDI expression, which would abrogate appropriate mitochondrial fragmentation by reducing S-nitrosylation of DRP1 and consequently evoke clasmatodendrosis.
In previous studies, we reported that clasmatodendrosis is Tdt-mediated dUTP Nick-End Labeling (TUNEL)-negative coagulative necrosis in astrocytes [3,9]. Later, we found that vacuoles in clasmatodendritic astrocytes contain LC3 and LAMP1 signals. Since LC3 is required for the autophagosome formation and LAMP1 is a marker for lysosomal biogenesis, amounts, and morphology [68][69][70][71], we reported that large vacuoles in clasmatodendritic astrocytes are autolysosomes [12,13]. LAMP1 is the predominant lysosomal membrane protein to maintain the integrity of the lysosomal membrane and the clearance of autophagosomes [68,69]. Cytoplasmic form LC3 (LC3-I) is diffusely observed in cell bodies, which is modified to LC3-II and concentrated in autophagosomes that exhibits granular puncta and a different mobility in electrophoresis [14,37,70,71]. Thus, an increase in LC3-II/LC3-I ratio is indicative of activation of autophagy process [37,70,71]. In the present study, clasmatodendritic CA1 astrocytes showed the increased LAMP1 and LC3 expression, which were ameliorated by CDDO-Me and SN50. Furthermore, LC3-positive puncta were apparently detected in vacuolized astrocytes of vehicle-treated epilepsy rats. Western blot data also revealed that LAMP1, LC3-I and LC3-II densities were elevated in the epileptic hippocampus concomitant with the increased LC3-II/LC3-I ratio, which were attenuated by CDDO-Me and SN50. Compatible with the present data, Sakai et al. [70] report that clasmatodendrosis is relevant to UPS-mediated autophagy. Qin et al. [71] also demonstrate that ischemia-injured astrocytes contain numerous multimembrane vesicles, described as typical for autophagosomes, which eventually fused with lysosomes in the cytoplasm, indicating that the autophagic/lysosomal pathway activation contributes to the decreased viability of astrocytes. Therefore, our findings suggest that excessive autophagy is involved in the pathogenesis of clasmatodendritic degeneration in the epileptic hippocampus, although it is unclear whether aberrant autophagy is the main cause of clasmatodendrosis or directly leads to astroglial degeneration.

Experimental Animals and Chemicals
Male Sprague-Dawley (SD) rats (7 weeks old) were used in the present study. Animals were cared under standard condition (22 ± 2 • C, 55 ± 5% and a 12:12 light/dark cycle with lights). Animals were provided with a commercial diet and water ad libitum.

Generation of Chronic Epilepsy Rats
Rats were given LiCl (127 mg/kg, i.p.) 1 day before the pilocarpine administration (30 mg/kg, i.p.). Twenty minutes before pilocarpine treatment, rats were treated with atropine methylbromide (5 mg/kg i.p.). Two hours after status epilepticus (SE) on-set, animals were given with diazepam (Valium; Hoffman la Roche, Neuilly sur-Seine, France; 10 mg/kg, i.p.) and repeated, as needed. Control animals received saline. SE-experienced rats were video-monitored 8 h a day to classify chronic epilepsy rats.

Tissue Processing
Seven days after surgery (infusion), rats were perfused with 4% paraformaldehyde through ascending aorta under urethane anesthesia (1.5 g/kg i.p.). After cryoprotection with PB containing 30% sucrose, 30 µm thick sections were made using a cryostat. For Western blot, animals were decapitated under the same anesthesia (1.5 g/kg, i.p.). The brains were quickly dissected to 1 mm thickness using rodent brain matrix (World Precision Instruments, Sarasota, FL, United States). Thereafter, the stratum radiatum of the CA1 region of the dorsal hippocampus were rapidly obtained. The CA1 tissues were lysed and protein concentration determined using a Micro BCA Protein Assay Kit (Pierce Chemical, Rockford, IL, United States) [11,75].

Immunohistochemistry, Cell Counts, and Mitochondrial Morphometry
After blocking with 0.1% bovine serum albumin, sections were incubated with primary antibodies (Table 1) and Cy2-or Cy3-conjugated secondary antibodies. Immunofluorescence was observed using AxioScope microscope (Carl Zeiss Korea, Seoul, South Korea). A negative control test was performed with normal rabbit serum (#31883, ThermoFisher Korea, Seoul, South Korea), mouse IgG1 isotype control (#02-6100, ThermoFisher Korea, Seoul, South Korea), or mouse IgG2a isotype control (#02-6200, ThermoFisher Korea, Seoul, South Korea). Cell counts were performed using AxioVision Rel. 4.8 Software. The areas of interest (1 × 10 4 µm 2 ) in a section (10 sections per each animal) were selected from the CA1 region, and 30 areas/rat (400 µm 2 /area) in the stratum radiatum of CA1 region (15 sections from each animal, n = 7 in each group) were selected and fluorescent intensity was measured using AxioVision Rel. 4.8 software (Carl Zeiss Korea, Seoul, South Korea). In addition, 5 hippocampal sections from each animal were randomly selected and mitochondrial morphometry was performed in CA1 astrocytes from each slice (total 35 cells in each group, respectively) using ImageJ software. Thereafter, mitochondrial parameters (area-weighted form factor, form factor and cumulative area/perimeter ratio) were calculated [28,29]. The morphological analyses were conducted by investigators were blinded to experimental groups [11,76].

Western Blot
Sample proteins (10 µg) were loaded on a Bis-Tris sodium dodecyl sulfate-polyacrylamide gel. After electrophoresis and transfer, membranes were incubated in primary antibodies ( Table 1). The visualization and quantification of immunoband were conducted using ImageQuant LAS4000 system (GE Healthcare Korea, Seoul, South Korea). Density of immunobands was calibrated with the β-actin.

Measurement of SNO-PDI
The quantification of SNO-PDI performed using the S-nitrosylation Western Blot Kit (ThermoFisher). Lysates were labeled with iodoTMTzero reagents after MMT pretreatment. Thereafter, TMT-labeled proteins were precipitated by anti-TMT resin, eluted by TMT elusion buffer, and identified by Western blot. For controls, ascorbate was eliminated from each sample [22,74].

Data Analysis
The values on normality were determined by Shapiro-Wilk W-test. For comparisons of data, Student t-test or one-way ANOVA were performed. The post hoc test was performed with Bonferroni's test. A p-value less than 0.05 was considered to be significant.

Conclusions
The present data reveal that the increased NF-κB S529 phosphorylation reduced the PDI protein level in clasmatodendritic CA1 astrocytes within the hippocampus of chronic epilepsy rats. Furthermore, clasmatodendritic CA1 astrocytes showed the accumulation of elongated mitochondria. CDDO-Me and SN50 attenuated clasmatodendrosis in CA1 astrocytes accompanied by the reduced NF-κB S529 phosphorylation and facilitated mitochondrial fission in CA1 astrocytes by increasing total PDI protein and SNO-PDI levels. PDI knockdown led to mitochondrial elongation in intact CA1 astrocytes under physiological conditions. Therefore, our findings suggest that NF-κB-mediated PDI downregulation may trigger clasmatodendrosis by abrogating DRP1-mediated mitochondrial fission, which is mitigated by CDDO-Me ( Figure 9).

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.