**4. Discussion**

Aberrant cell cycle entry in post-mitotic neurons leads to neuronal death in various neurological diseases [35,36]. Similarly, SE increases the expression of cell cycle regulatory molecules such as cyclin D1 and CDK4, and evokes programmed necrosis in CA1 neurons through dysfunction of mitochondrial fission [8,10,16–18]. Imbalance of mitochondrial fusion/fission rate evokes cell degeneration: Dysfunction of mitochondrial fission (aberrant mitochondrial elongation) results in improper segregation of mitochondria and a decrease in ATP levels, which abrogates mitochondrial transports in dendrites or axons, and subsequently induces ATP deficiency in peripheral sites. Impaired mitochondrial fission also inhibits respiratory function in mitochondria that triggers excessive ROS production. Therefore, dysregulation of mitochondrial fission leads to cell degeneration [37–39]. Excessive mitochondrial fission (impaired mitochondrial fusion) also results in cell death. This is because fragmented mitochondria are not able to produce ATP, which impairs the detoxification of excess ROS and extrusion of intracellular Ca2+, and in turn increases mitochondrial ROS and susceptibility to apoptosis [15,34,40,41]. Indeed, abnormal mitochondrial fission induces PV cell apoptosis following SE [10,25]. Since CDDO-Me inhibits cyclin D1 and induces cell cycle arrest [19,20], it is likely that CDDO-Me would a ffect aberrant mitochondrial dynamics induced by SE. In the present study, we found that CDDO-Me reduced mitochondrial length in CA1 neurons and PV cells without altering mitochondrial LONP1 expression in control animals. Furthermore, CDDO-Me e ffectively ameliorated SE-induced CA1 neuronal death, but not PV cell loss, accompanied by abrogating abnormal mitochondrial elongation. Since CDDO-Me induces mitochondrial fission in cancer cells [22,31,32] and WY14643 (an enhancer of mitochondrial fission) attenuates SE-induced CA1 neuronal death by rescuing aberrant mitochondrial elongation [8,25], our findings indicate that CDDO-Me-induced mitochondrial fission may have a selective neuroprotective e ffect against SE-induced CA1 necrosis, but not PV cell apoptosis.

In the present study, CDDO-Me increased DRP1-S616, but not S637, phosphorylation without changing the expression of other molecular components of mitochondrial dynamics, such as OPA1 and MFN1/2. Thus, CDDO-Me increased the DRP1-S616/S637 phosphorylation ratio. In our previous studies [8,16,17], SE decreased DRP1 expression and the DRP1-S616/S637 phosphorylation ratio, accompanied by increased mitochondrial length and sphere formation in CA1 neurons. Furthermore, WY 14643 increases DRP1-S616 phosphorylation and the DRP1-S616/S637 phosphorylation ratio [8]. Therefore, it is likely that the decreased DRP1-S616/S637 phosphorylation ratio may lead to aberrant mitochondrial elongation following SE. Consistent with these previous reports, the present study reveals that SE diminished the DRP1-S616/S637 phosphorylation ratio, concomitant with abnormal mitochondrial elongation, which were abolished by CDDO-Me. These findings indicate that CDDO-Me may facilitate mitochondrial DRP1 localization via the increased DRP1-S616/S637 phosphorylation ratio, although we could not confirm the subcellular localization of DRP1.

DRP1-S616 phosphorylation is regulated by the ERK1/2 and JNK signal pathways, which facilitate mitochondrial fission [27,29,30]. However, DRP1-S637 phosphorylation by PKA leads to detached DRP1 from mitochondria, thus inhibiting mitochondrial fission [26,28]. Indeed, DRP1-S637 dephosphorylation by PP2B accelerates mitochondrial fission [34]. Consistent with previous studies demonstrating that CDDO-Me regulates ERK1/2 [20,31] and JNK phosphorylation [32,33], the present data also reveal that CDDO-Me activated ERK1/2 and JNK under physiological conditions, and mitigated SE-induced reductions in ERK1/2 and JNK phosphorylation. However, CDDO-Me did not influence PKA and PP phosphorylation under physiological- and post-SE conditions. Since ERK1/2 inhibitor deteriorated SE-induced CA1 neuronal damage concomitant with mitochondrial elongation [18], our findings indicate that CDDO-Me may facilitate DRP1-mediated mitochondrial fission by activating ERK1/2 and JNK, which may attenuate CA1 neuronal degenerations following SE.

The present study also reveals that SE increased LONP1 expression in CA1 neurons as well as PV cells. LONP1 is one of the quality control proteins in the mitochondria supporting cell viability via the degradation of misfolded and damaged proteins under oxidative, hypoxic and endoplasmic

reticulum-stress conditions [42,43]. Indeed, LONP1 expression is up-regulated under these stressful conditions [44,45]. Thus, LONP1 knockdown results in the disruption of mitochondrial function, reduced proliferation and the fragmented shape of the mitochondrial network [22,24,46]. With respect to these reports, it is likely that the up-regulation of LONP1 in PV cells and CA1 neurons may be an adaptive response to protect these neurons from SE. Considering this hypothesis and CDDO-Me as a LONP1 inhibitor [21,22], it is plausible that CDDO-Me would exacerbate degenerations of hilus interneurons and CA1 neurons by inhibiting LONP1 activity following SE. In the present study, however, CDDO-Me selectively alleviated CA1 neuronal damage without a ffecting the up-regulation of LONP1 expression induced by SE. In contrast, LONP1 knockdown aggravated neuronal loss of CA1 neurons. What makes these discrepancies in the e ffects of CDDO-Me and LONP1 siRNA on neuronal viability? In previous studies [21,22,24,46], CDDO-Me and LONP1 siRNA induced apoptosis of various cancer cells that have potent proliferative and di fferentiating abilities by inhibiting mitochondrial functionality. However, we applied CDDO-Me and LONP1 knockdown to post-mitotic neurons in the present study. Therefore, it is presumable that the di fferential proliferating ability of cancer cells and neurons would lead to the distinct e ffects of CDDO-Me and LONP1 siRNA on neuronal damage. Conversely, it is also considerable that LONP1 would not be a specific target of CDDO-Me. Indeed, the present data show that CDDO-Me a ffected ERK1/2 and JNK phosphorylation, while LONP1 siRNA did not influence the expression and phosphorylation of DRP1, ERK1/2, JNK, PKA and PPs under physiological- and post-SE conditions. In addition, co-treatment of CDDO-Me with LONP1 siRNA ameliorated only CA1 neuronal death, but not hilus neurons and dentate granule cells, following SE. Although LONP1 knockdown did not a ffect mitochondrial length in CA1 neurons, co-treatment of COOD-Me abrogated mitochondrial elongation induced by SE. Therefore, the present findings indicate that the neuroprotective e ffects of CDDO-Me may not be relevant to LONP1 inhibition. Since SE leads to programmed necrosis in CA1 neurons [8,16,17] and apoptosis in hilus interneurons, respectively [10], it cannot be excluded that CDDO-Me may selectively attenuate programmed necrosis rather than apoptosis. Taken together, the data from previous reports [8,25] and the present study identically sugges<sup>t</sup> that the repair of dysfunction of mitochondrial fission may selectively rescue SE-induced CA1 neuronal death.

The heterogeneous vulnerability of hippocampal neurons in response to various insults has been reported. Briefly, dentate granule cells are remarkably resistant to various insults when compared to CA1 neurons or hilus interneurons [8,10,16,17,47–49]. Consistent with these previous studies, the present data show that SE induced a massive neuronal loss of hilus interneurons and CA1 neurons, although the degeneration of dentate granule cells was negligible. However, LONP1 knockdown provoked the massive degeneration of dentate granule cells, and aggravated loss of CA1 neurons and hilus interneurons. These findings indicate that LONP1 may be one of the important housekeeping molecules for neuronal viability, regardless of the heterogeneous vulnerability in response to harmful stimuli. Thus, it is likely that the regulation of LONP1 may be a useful therapeutic strategy for prevention of neurodegeneration. Future studies are needed to elucidate the role of LONP1 in other neurological diseases and the underlying regulatory mechanisms for its expression and activity.
