[1] 17w The potential LXRβ agonist stigmasterol protects against hypoxia/ reoxygenation injury by modulating mitophagy in primary hippocampal neurons

[1] 292w Background: Neuronal excitotoxicity induces a plethora of downstream signaling pathways, resulting in the calcium overload-induced excitotoxic cell death, a well-known phenomenon in cerebrovascular and neurodegenerative disorders. The naturally occurring phytosterol, stigmasterol (ST) is known for its potential role in cholesterol homeostasis and neuronal development. However, the ability of ST to protect against the induced excitotoxicity in hippocampal neurons has not been investigated yet. Purpose: The present study aimed to investigate whether ST could protect against hypoxia/reoxygenation (H/R)induced excitotoxicity in hippocampal neurons. Methods: After H/R, neurons were initially subjected to trypan blue exclusion assay for the assessment of cell viability. Live staining using fluorescence dyes namely JC-1 (5,5',6,6'-tetrachloro-1,1',3,3'-tetraethylbenzimidazolyl-carbocyanine iodide), DCFDA (2 ′ ,7 ′ -dichlorofluorescein diacetate) and FM1-43 (N-(3triethylammoniumpropyl)-4-(4-(dibutylamino)styryl) were used to measure MMP, ROS and synaptic vesicle pool size. Immunostaining was performed to analyze the expression levels of vesicular glutamate transporter 1 (VGLUT1), N-methyl-D-acetate receptor subunit 2B (GluN2B), LC3BII, p62, and PTEN induced protein kinase 1 (PINK1) in neuron after H/R. Western blotting was carried out to measure the protein expression of GluN2B. The molecular dynamics simulation was employed to elucidate the LXRβ agonistic conformation of ST.Result: Pre-incubation of neuronal cultures with ST (20 μM) protected against excitotoxicity, and attenuated reactive oxygen species (ROS) generation, double-stranded DNA break, and mitochondrial membrane potential (MMP) loss. ST treatment also resulted in the downregulation of the expressions of VGLUT1 and GluN2B and the reduction of the size of recyclable synaptic vesicle (SV) pool. Like LXRβ agonist GW3695, ST suppressed the expression of GluN2B. Furthermore, ST induced mitophagy through upregulating the expressions of LC3BII, p62, and PINK1. The molecular simulation study showed that ST interacted with the ligand binding domain of liver X receptor β (LXRβ), a known binding receptor of ST, through multiple hydrogen bonding.

[1] 192w Phytosterols have shown various neuroprotective functions in the central nervous system (CNS) disorders like Alzheimer's disease (AD), multiple sclerosis, amyotrophic lateral sclerosis/Parkinsonism-dementia complex (Vanmierlo et al., 2013). Mammals are unable to synthesize phytosterols and therefore, can only obtain them from their diet. Stigmasterol (ST) is the most common phytosterol and is isolated from various medicinal herbs like Butea monosperma (Lam.) Taub. (Fabaceae), Edgeworthia gardneri (Wall.) Meisn. (Thimelaceae), Sideritis lasiantha Juss. Ex. Pers. (Lamiaceae), Parkia speciosa Hassk. (Fabaceae), and Calotropis gigantea (L.) Dryand. (Apocynaceae) (Kaur et al., 2011), and plant-based food materials such as banana, cabbage, peanut, soybean, sunflower, and corn oils (Awad and Fink, 2000) as well as from marine algae (Luo et al., 2015). Unlike cholesterol, phytosterols, such as ST, can cross the blood-brain barrier (BBB) and accumulate in the brain (Jansen et al., 2006). It has been reported that ST strongly reduced amyloid β (Aβ) levels in human neuroblastoma SH-SY5Y cells, which were stably transfected with predominant amyloid precursor protein (APP) isoform 695. Likewise, ST decreases the cerebral activity of amyloidogenic enzymes in mice (Burg et al., 2013). This evidence suggests that ST might be beneficial for ameliorating neurodegenerative diseases.

[2] 114w ST also has neuromodulatory effects. For example, in mice, ST supplementation significantly increased the phosphorylation levels of the extracellular signal-regulated kinase (ERK) and cAMP response elementbinding protein (CREB) in the hippocampus and attenuated scopolamine-induced memory impairments (Park et al., 2012). ST also activated glutamatergic neurotransmission and inhibits acetylcholinesterase activity (Khabazian et al., 2002;Sultana and Khalid, 2010). Previously, our laboratory showed that ST upregulated cell division cycle 42 (Cdc42)-actin-related protein 2/3 (Arp2/3) and ERK1/2-CREB signaling, and thus, induced mushroom-type spines and the formation of the glutamatergic excitatory synapse (Haque et al., 2018). In addition, our laboratory also showed the migration stimulation effect of ST in neurosphere migration assays of brain neurons (Haque and Moon, 2018a).

[3] 254w In a transcriptomic and bioinformatic study undertaken to identify genes targeted by ST (Haque and Moon, 2018b), we observed that ST induced many of the immediate early genes and that the major proportion of these upregulated genes were related to CNS development (neurite outgrowth or synaptic transmission). Furthermore, studies on the regulation of gene expression and cytoarchitectural modulation, such as neuritogenesis and synaptogenesis by ST, strongly suggest that ST may exert neuroprotection. Indeed, it has also been reported that ST protected against glutamate-induced excitotoxicity in the HT-22 mouse hippocampal neuronal cell line (Lee et al., 2014). However, the molecular mechanism underlying the neuroprotection by ST is not clearly understood. Interestingly, our previous findings based on transcriptomic study (Haque and Moon, 2018b), as well as the recent report of Sabeva et al. (2011) inferred that ST differentially influences the expressions of ATP-binding cassette (ABC) transporters. These expressions are controlled by liver X receptors (LXRs) that are known to regulate glutamate toxicity and mitophagy in primary hippocampal neurons, in which LXR activation has been reported to inhibit GluN2B selectively and to induce PNIK1-mediated mitophagy (Baez-Becerra et al., 2018), and thus, to protect against cerebral ischemia-mediated brain damage. With this evidence, we speculate that ST might function as an LXR agonist and protect neurons from excitotoxicity by modulating NMDA receptor signaling and promoting mitophagy. To address this hypothesis, we performed in silico molecular docking analysis of ST to LXR binding and a series of cell biological experiments after exposing primary rat hippocampal neurons to hypoxia/reoxygenation (H/R) injury.

[1] 163w To determine whether ST protects mitochondrial function, hippocampal cultures (DIV 9) were exposed to H/R shock and ROS and MMP levels were measured after 96 h. Typical images of cultures after DCFDA staining are shown in Fig. 2A-a We next investigated the effect of ST on H/R-induced MPP reduction. JC-1 staining of treatment-naïve HR controls [(-ST) H/R] resulted in mostly green (i.e., low MMP) fluorescent mitochondria, whereas STtreated cultures [(+ST) H/R] were mostly red ones (Fig. 2B-a). The overall field intensities were measured using Image J software, and the intensity ratios of red to green fluorescence are shown in Fig. 2B-b. H/R reduced the red/green ratio significantly (from 1.28 to 1.14; p < 0.05) in treatment-naïve (-ST) H/R cultures [Fig. 2B-b, (-ST) H/R], whereas the ratio was maintained at 1.27 in (+ST) H/R cultures, and this suppression of MMP loss was significant (p < 0.05). These results indicate that ST suppressed H/R-induced ROS production and MMP in H/R shock and protects mitochondrial functions.

[2] 95w To investigate the effects of ST on VGLUT1 expression, we conducted ICC on DIV 13 cultures (Fig. 4A-a). The density of VGLUT1 puncta were significantly (p < 0.001) increased (by 55%, from 11 ± 0.75 to 17 ± 1.1) in treatment-naïve (-ST) H/R cultures as compared with normoxic control ones [Fig. 4A-b, (-ST) no H/R]. However, this increase was significantly (p < 0.001) suppressed down by 29% (from 17 ± 1.1 to 12 ± 0.79) in ST pretreated cultures [Fig. 4A-b, (+ST) H/R], suggesting that ST protects neurons from damage by suppressing H/R-induced VGLUT1 upregulation.

[3] 50w Next, we measured the pool size of recyclable SVs by staining live neurons with FM1-43 (Fig. 4B-a), and found that the mean fluorescence intensity of SVs puncta was significant (p < 0.001) increased by 1.15 folds in treatment-naïve (-ST) cultures after H/R compared to normoxic control ones [Fig. 4B-b, (-ST

[4] 137w Excessive release of glutamate from synaptic endings and its spillover into the spine neck activates extra synaptic GluN2B receptors leading to excitotoxicity. We performed ICC and double-stained the cultures with antibodies against GluN2B and MAP2. Typical fluorescence images were shown in Fig. 4C-a Western blotting of proteins isolated from hippocampal cultures (Fig. 4C-c). The GluN2B band signal increased by 20% in treatmentnaïve (-ST) H/R cultures (Fig. 4C-d), whereas ST pretreatment suppressed this increase by 18% to the level of control cultures [Fig. 4C-d As shown in Fig. 4D-a, and b, ST pretreatment [(+ST) H/R] successfully attenuated (from 61% to 34%, p < 0.05) hypoxia-induced GluN2B overexpression as LXRβ agonist (1μM, GW3965 H/R) did (from 61% to 21%, p < 0.05). We further confirmed the LXRβ agonistic conformation of ST by molecular dynamics simulation in the subsequent study.

[5] 192w To evaluate the effects of ST on autophagy, we examined the expression of LC3IIB, a marker for autophagy, by WB and found LC3IIB expression in treatment-naïve (-ST) H/R cultures were comparable to that of normoxic control [(-ST) no H/R] (Figs. 5A-a emerge the ST role in the clearance of abnormal cellular materials. In order to determine whether autophagosome formation was chaperondependent or not, we examined the expression of heat-shock cognate 70 (HSC70), which has been shown to be the only chaperone that directly binds to the substrates of chaperon-mediated autophagy (CMA) through the KFERQ pentapeptide binding domain (Chiang et al., 1989;Kaushik and Cuervo, 2012). We found ST, as well as treatment-naïve (-ST) H/R culture had no effect on the expression of Hsc70 in both WB and ICC analyses after H/R compared to normoxic [(-ST) no H/R] cultures (Figs. 5A-a, 5B-c and 5C-a, -b). In contrast ST significantly (p < 0.05) increased the expression of PINK1 after H/R compared to treatment-naïve (-ST) H/R cultures by 28% and normoxic [(-ST) no H/R] culture (p < 0.05) by 27% (Figs. 5A-a, 5B-d). These results indicate that ST induces mitophagy in hippocampal neurons exposed to H/R.

[6] 129w In order to investigate the binding between ST and the ligand binding domain (LBD) domain of LXRβ, molecular docking simulation was performed using Glide XP docking. The docking study initially suggested that ST well fitted to the LXRβ binding site (Table S1) and that it adopts a binding orientation similar to that of desmosterol (a known LXRβ agonist, Fig. 6A). Detailed analysis revealed ST interacted with amino acids around the LBD that included two hydrogen bonds and several hydrophobic interactions. ST was found to hydrogen bond with ASN239 and PHE329 via the hydroxyl group at position C-3 of the steroid skeleton. In addition, the residues PHE243, PHE271, THR272, LEU274, ILE309, MET312, PHE329, LEU345, and TRP 457 interacted with ST by pi-alkyl bonding (a type of van der Waals interaction).

[7] 145w The binding energy of ST to LXRβ was calculated and compared with that of desmosterol (Table S2). Interestingly, the binding energy of ST was greater than that of desmosterol (-89.76 vs.-86.35 kcal/mol). The conformation of ST in the LXRβ binding site was further validated by a 100 ns molecular dynamics simulation. In this simulation, the free conformation of LXRβ was included to check for conformational changes in LXRβ induced by ST binding. The RMSD and RMSF values of the protein backbones (Cα atoms) of the apo and holo systems were calculated individually and plotted by comparing with the initial protein structure of the simulation. Both analyses revealed that ST binding remained stable and achieved equilibrium after a simulation time of 50 ns (Fig. 6B). Moreover, no significant local changes were found after ST binding, which also indicated stable binding between ST and LXRβ (Fig. 6C).

[8] 132w Hydrogen bonding was also calculated, and results showed that the total number of hydrogen bonds changes from the start of the simulation time. However, ST maintained significant hydrogen bonding with the active site residues, over the simulation (Fig. 6D). Although molecular docking revealed hydrogen bonding by the ASN239 and PHE329 residues, only ASN239 interacted during the simulation. Moreover, it formed several H-bonding with other residues including GLU281, ARG319, THR316, SER278, ASN239, and SER278 during the simulation. High H-bond occupancies were found for the GLU281 and ARG319 residues at 20.21% and 21.04%, respectively. However, ASN239 had a H-bond occupancy of only 7.60% through the simulation. Since hydrogen bonding interactions with GLU281 and ARG319 are necessary for sterol binding to the LXRβ, these results confirm the agonistic conformation of ST of the LXRβ activation.

[1] 226w Metabolic regulations play crucial roles in neuronal function, and the hippocampus is one of the brain regions most affected by metabolic dysregulations like hypoxia-ischemia. Excitotoxicity is the leading cause of neuronal death in hypoxia/ischemia, and pyramidal neurons in the hippocampus are highly vulnerable to hypoxic damage due to high densities of glutamate receptors (Wall et al., 2014). Excessive glutamate release during brain ischemia is a well-known phenomenon that causes over stimulation of glutamate receptors, especially GluN2B containing NMDA receptors (NMDARs). Elevated intracellular Ca 2+ level resulting from NMDAR over stimulation, in turn, causes a variety of death pathways, which include protease activation, NO release, and mitochondrial dysfunction (Puyal et al., 2013). In this study, we observed that H/R surges glutamate like ischemic injury and ST pretreatment attenuated this excitotoxicity by both pre-and postsynaptic neuronal activities as evidenced by the decrease in expression of VGLUT1, ready to release synaptic vesicle pool size as well as extrasynaptic glutamate receptor GluN2B. Evidence suggests that the activation of LXRβ leads to the downregulation of GluN2B, and thus attenuates glutamate excitotoxicity (Baez-Becerra et al., 2018). In line with this, we also observed similar suppression of GluN2B expression following LXRβ agonist application, which is comparable to the effect of ST, suggesting its function as an LXRβ agonist. The in silico data showing LXRβ agonistic conformation of ST also supported this cellular effect.

[2] 141w Moreover, under H/R conditions, the prolonged activation of extrasynaptic glutamate receptors by excessive glutamate release can increase ROS production and reduce MMPs (Wahl et al., 2009). In fact, MMP loss provides an early indication of glutamate-induced damage leading to neuronal death (Wahl et al., 2009). Subsequently, the present study also demonstrated H/R induced ROS production and MMP loss, and ST pre treatment counter balanced this MMP loss and ROS productions. Endogenous antioxidant systems are critically important in limiting reoxygenation-induced cellular damage as the reoxygenation that follows after hypoxia usually induces bursts of ROS, which not only cause the oxidative damage central to the pathophysiology of H/R injury but also activate signaling mechanisms that synergize the death signaling to neuron. ROS contributes not only to the injury of macromolecules such as lipids and proteins, but also to the transductions of apoptotic signals.

[3] 67w The main source of ROS production is dysfunctional mitochondria which activate the multiple intrinsic pathways leading to inter nucleosomal fragmentation of DNA double-strand due to rapid phosphorylation of the histone H2A variant H2AX (Yang et al., 2017). The present study shows by ICC a significant increase in H2AX immunopuncta in treatment-naïve (-ST) H/R cultures, which was effectively suppressed by ST pretreatment, indicating ST prevented ROS-induced DNA fragmentation.

[4] 81w The present study also showed that ST induces mitophagy in H/R exposed hippocampal cells. Samokhvalov et al. showed that activation of autophagy has an important neuroprotective role after hypoxic injury (Samokhvalov et al., 2008). In particular, it has been reported that for survival, cells trigger a robust increase in mitophagy via a PINK1-dependent Parkin pathway (Park et al., 2017;Wang et al., 2011). Furthermore, ubiquitin-binding adaptor p62 recruits cargo into autophagosomes by binding to LC3 after ROS production (Chao et al., 2019).

[5] 200w In the present study, we found ST also significantly induced the autophagy marker protein LC3II and adjunct adaptor protein p62, which suggested the neuroprotective effects of ST are due to the enhancement of autophagy. Interestingly, ST did not increase Hsc70 but induced PINK1 expression, which stimulates mitophagy. These results indicate that the type of autophagy induced by ST in H/R is not CMA but a mitophagy. Recently, Sun et al. (2019) reported that ST exerts neuroprotection against ischemic/reperfusion (I/R) injury by inactivation of autophagy (Sun et al., 2019), which is apparently in contradiction with our results. The reason for this discrepancy is not readily explained. Neuronal autophagy is known to have contrasting roles in chronic neurodegenerative and acute neural disorders (Puyal et al., 2012). Primary neuronal cultures are different from brain neurons in that they are immature (in our case, DIV 9, which is early maturation state) and that the influence of the glia components present in vivo is minimal in culture. As neurons are highly dependent on healthy mitochondria, mitophagy may be more important in H/R in vitro and I/R in vivo experiments. Therefore, caution should be taken when comparing data regarding autophagy in the brain and isolated neurons.

[6] 252w How does ST exert its neuroprotective effects at the molecular level? A recent report by Báez-Becerra et al. showed that activation of LXRs regulates glutamate toxicity and mitophagy in primary hippocampal neurons, in which this activation selectively inhibits GluN2B and induces PNIK1 mediated mitophagy (Baez-Becerra et al., 2018). Furthermore, LXRs are known to provide protection against cerebral ischemia-mediated brain damage (Sironi et al., 2008). Our present molecular dynamic simulation studies concur with these results, by indicating that ST exerts its functions by interacting with LXRs. This possibility stimulated us to investigate in silico interaction between ST and LXRβ. To our surprise, ST formed a similar binding orientation to the known agonist desmosterol to LXRβ. More precisely, docking simulation revealed that ST formed multiple hydrogen bonds with the GLU281 and ARG319 residues of LXRβ in a similar orientation to the other endogenous steroids-nuclear receptor complexes (Spencer et al., 2001). It was shown in a previous study that hydrogen bonding between endogenous steroids and LXRβ at GLU281 on helix 3 and ARG319 on helix 5 is essential for ligand binding (Brzozowski et al., 1997). Our computational studies revealed the hydroxyl group in position C-3 of the steroid skeleton of ST formed strong hydrogen bonds with GLU281 and ARG319 of LBD, which indicated the agonistic binding orientation of ST was similar to that found in other endogenous steroid-nuclear receptor complexes. Hence forth, our study confirms that ST acts as LXRβ agonist and protects neuron from the excitotoxicity by modulating NMDA receptor function and promoting mitophagy.

[1] 202w The present study demonstrates that ST protected against H/Rinduced neuronal injury, suppressed ROS generation, prevents doublestranded DNA break and preserved MMP. The ST-mediated attenuation of excitotoxicity was attributed, at least in part, to the repressed expression of VGLUT1 and the reduction of the size of vesicle pool at the presynaptic terminals, and downregulation of GluN2B at the postsynaptic zones. Moreover, overexpression of LC3BII, p62, and PINK1 by ST against H/R-induced oxidative stress underscored a potential neuroprotective mechanism that involves mitophagy. With these evidences, Fig. 5. Immunoblot analysis on the effect of ST pretreatment on the expressions of LC3II and PINK1 after H/R. Vehicle or ST was added to hippocampal cultures plated at 1.2 × 10 5 cells/cm 2 . Proteins were isolated and immunoblotted following H/R. (A) Representative immunoblot bands showing LC3II, p62, HSC70, PINK1, and actin expressions. (B) Relative intensities and their test statistics as measured using Image J software and normalized versus actin. (C) Representative immunofluorescence images of tubulin (green)/HSC70 (red) staining. Bars represent means ± SEMs (n = 3; 30 microscopic fields per group for immunofluorescence images). ### p < 0.001 compared with the normoxia control; * p < 0.05; ** p < 0.01 compared with hypoxia control (ANOVA).

[1] 53w ST (~95% purity) and LXRβ agonist GW3965 hydrochloride (~98% purity) were purchased from Sigma-Aldrich (St. Louis, MO, USA), dissolved in ethanol (EtOH), and stored in foil-wrapped vials at -20 • C. All chemicals, reagents, and media used in primary cultures of hippocampal neurons were purchased from Invitrogen (Carlsbad, CA, USA) unless otherwise stated.

[2] 88w Neurons were exposed to H/R, as described by Mohibbullah et al. (2015). Briefly, after 9 days, neurons on coverslips in culture plates were transferred to a hypoxic incubation chamber (Modular Incubator Chamber MIC-101; Billups-Rothenberg Inc., Del Mar, CA, USA) containing 94% N 2 , 5% CO 2 , and 1% O 2 and incubated for 3 h at 37 • C. To re-oxygenate hypoxic cells, culture plates were placed in an incubator containing 95% air and 5% CO 2 at 37 • C and incubated for 96 h.

[3] 95w Neuronal viability was assessed using a trypan blue exclusion assay after 96 h following hypoxic shock of 9 days old cultures. Briefly, neurons on coverslips were stained with 0.4% trypan blue for 15 min at room temperature (RT) and then washed with Dulbecco's phosphatebuffered saline (D-PBS; Invitrogen). Since, trypan blue dye itself cannot permeate viable cells but can enter cells with compromised membranes and stains such dead cells dark blue in phase-contrast images. Cell viabilities are expressed as percentages of trypan blueimpermeable cells (live neurons), and results were normalized versus trypan blue-stained non-H/R exposed controls.

[4] 131w To evaluate the excitability of hippocampal neurons in culture after H/R, cycling synaptic vesicles (SVs) in the nerve terminals were stained with the styryl dye N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino) styryl) pyridium dibromide (FM1-43; Molecular Probes). To label SVs, ST or vehicle-treated hippocampal neurons were incubated for 3 min in depolarizing solution (50 mM KCl, 54 mM NaCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 20 mM HEPES; pH 7.3) containing 10 µM of FM1-43 (the depolarizing solution induced vesicle exocytosis and facilitated FM1-43 uptake). Neurons were then washed (3 × 5 min) in FM1-43 free extracellular solution (without Ca 2+ /3 mM Mg 2+ ) to wash off excess dye. Stained SVs were visualized at 568 nm and the fluorescence intensities of FM1-43 stained SV puncta were quantified using Image J.

[5] 136w Neurons on coverslips were rinsed briefly with D-PBS and fixed by sequential treatment in paraformaldehyde and methanol (Moon et al., 2007) on DIV 13 (96 h post-hypoxia). The following antibodies were used for ICC: primary antibodies of the tubulin α-subunit (mouse monoclonal 12G10, 1:1,000 dilution; Developmental Studies Hybridoma Bank, University of Iowa, IO, USA); anti-microtubule associated protein 2 (MAP2, 1:500, mouse monoclonal, Sigma-Aldrich), GluN2B (1:1,000, rabbit polyclonal) andanti-phospho-H2AX antibody (1:500; Millipore, Billerica, MA, USA); and anti-heat shock conjugated 70 (HSC70, 1:100, rabbit polyclonal, Proteintech, Chicago, IL, USA). The secondary antibodies were Alexa Fluor 488-conjugated goat anti-mouse IgG [1:1,000] and Alexa Fluor 568-conjugated donkey anti-rabbit IgG [1:1,000], Molecular Probes). Fixed neurons were incubated with primary antibodies overnight at 4 • C, then treated with secondary antibodies, and mounted on slides as previously described (Moon et al., 2007).

[6] 146w Hippocampal cells (1.2 × 10 5 cells/cm 2 , DIV13) were harvested after H/R treatment and lysed in ice-cold RIPA buffer [50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% (v/v) NP-40, 0.5% (w/v) sodium deoxycholate, 1% (w/v) sodium dodecylsulfate (SDS), and protease inhibitor cocktail (Thermo Scientific, Rockford, IL, USA)]. Protein concentrations were measured using the Bradford method (Bradford, 1976). Equal amounts of protein were separated by 8% SDS-PAGE and transferred to PVDF membranes (Haque and Moon, 2018a), which were incubated with primary antibodies: rabbit polyclonal anti-LC3II , anti-p62 (1:2,000; rabbit polyclonal, Sigma-Aldrich), anti-PINK1 (1:500, rabbit polyclonal, GeneTex, Irvine, CA, USA) and anti-actin (JLA20; 1:1,000, mouse monoclonal, Developmental Studies Hybridoma Bank). After rinsing with TTBS (0.05% Tween-20 in TBS), membranes were incubated with horse radish peroxidase-conjugated secondary antibodies (1:1,000; anti-mouse or -rabbit IgG; Amersham Biosciences, Chalfont, UK). Signals were detected using an ECL detection kit (Invitrogen).

[7] 50w One-way analysis of variance (ANOVA) with post hoc Duncan's multiple comparisons and Bonferroni's correction (SPSS software, version 16.0) was used to determine the significances of differences. Results are presented as the means ± standard errors of at least three independent experiments. Predetermined p values ≤ 0.05 were considered statistically significant.

[1] 146w The animal experiment was approved beforehand by the Institutional Animal Care and Use Committee of the College of Medicine, Dongguk University (approval certificate number IACUC-2016-001). On the 13 th day of pregnancy, time-pregnant Sprague-Dawley rats were housed in the animal care room at 22 • C under a 12 h light-dark cycle with access to sufficient food and water. At 19 th day of pregnancy rats were then euthanized with isofluorane and fetuses were collected. Primary hippocampal cell cultures were prepared as previously described (Haque and Moon, 2018b). The dissociated cells were seeded at a density of 3.0 × 10 4 cells/cm 2 onto poly-DL-lysine-coated (PDL, Sigma-Aldrich) 12-mm glass coverslips in 24-well culture plates for morphological and viability analysis, or 1.0 × 10 6 cells/cm 2 onto PDL-coated six-well culture plates for Western blotting in culture medium containing ST or vehicle (EtOH, final concentration < 0.5%).

[2] 90w MMP is an indicator of mitochondrial energy status and were assessed by 5',6,1',3, labeling (1 μg/ml; Molecular Probes Inc., Eugene, OR, USA). Live neurons on coverslips were washed with prewarmed fresh media, incubated with JC-1 for 30 min at 37 • C, and washed twice with pre-warmed fresh media. Fluorescence images were obtained at 488 nm (green) for JC-1 monomers and at 568 nm (red) for JC-1 aggregates. Ratios of red/green intensity were used to assess MMP levels using Image J (version 1.45; National Institutes of Health, Bethesda, MD, USA).

[3] 67w ROS levels were assessed using the fluorescent probe 2 ′ ,7'-dichlorofluorescein diacetate (DCFDA, Molecular Probes) on the day in vitro (DIV) 13. Briefly, the live cultures were exposed to DCFDA (10 µM) for 30 min under dark condition in the incubator, washed with pre-warmed PBS, and visualized under a fluorescence microscope. Images were obtained at 488 nm (green), and ROS positive neurons were quantified using Image J.

[4] 317w The 3-dimensional structure of human LXR ligand-binding domain (LBD) was retrieved from the protein data bank (PDB ID: 1P8D) and prepared using bond orders, hydrogen and charges. The structure of its LBD was refined by removing water molecules and optimizing the protein at neutral pH. In addition, some thiol and hydroxyl groups, asparagine, and glutamine amide groups, imidazole rings of histidines, and protonation states of histidines and aspartic and glutamic acids were adjusted. Minimization was performed by applying the OPLS 3 force field and adjusting maximum heavy atom RMSD to 0.30 Å. In order to perform Glide XP (Extra precision) docking, the three-dimensional coordinates of ST were downloaded from PubChem databases and prepared using the ligand preparation wizard in Maestro 11.1 with an OPLS 3 force field. Their ionization states were generated at pH 7.0 ± 2.0 using Epik 2.2 in Schrödinger Suite 2017-1. The active site of the protein was fixed for docking simulation by generating a grid box at the reference ligand binding of the protein. Grid generation parameters were kept at default with a box size of 18 Å × 18 Å × 18 Å, and the OPLS 3 force field was used for post-minimization. The charge cut-off and the van der Waals scaling factor were set at 0.25 and 1.00, respectively. Extra precision (XP) flexible docking was performed using the Glide module of Schrödinger-Maestro v 9.4. All ligands were treated flexibly using a partial charge and van der Waals factor of 0.15 and 0.80, respectively. Minimization was performed to the docked complex after docking using the OPLS 2005 force field. Additionally, Prime MM-GBSA approach was used to calculate binding free energy, which combines OPLSAA molecular mechanics energies (EMM), an SGB solvation model for polar solvation (GSGB), and a non-polar solvation term (GNP) composed of non-polar solvent accessible surface area and van der Waals interactions. The total free energy of binding is given by:

[5] 178w After, the docked complex and the ligand-free protein were subjected to molecular dynamic simulations using YASARA Dynamic software. Prior to simulation, all structures were cleaned and the hydrogen bonding network was optimized. For each simulation system, a cubic simulation cell with periodic boundary condition was generated, and then all atoms were parameterized with the AMBER14 force field. The transferable intermolecular potential 3 points (TIP3P) water model was used to create the solvation system, and density was maintained at 0.997 g/l. Using the simulated annealing method, initial energy minimization of each simulation system was performed over 5000 cycles using the steepest gradient approach. Molecular dynamics simulations were performed using the PME method to describe long-range electrostatic interactions at a cut-off distance of 8 Å under physiologic conditions (298 K, pH 7.4, 0.9% NaCl). A multiple time-step algorithm and a simulation time step interval of 2.50 fs were chosen (Krieger and Vriend, 2015). At a constant pressure and Berendsen thermostat, molecular dynamic simulations were performed for 100 ns long, and MD trajectories were saved every 25 ps for further analysis.

[6] 95w Phase-contrast and epifluorescence microscopy were performed using a Leica Research Microscope DM IRE2 equipped with I3S, N2.1S, and Y5 filter systems (Leica Microsystems AG, Wetzlar, Germany), a high-resolution CoolSNAP TM CCD camera (Photometrics Inc., Tucson, AZ, USA), and Leica FW4000 software. Digital images were processed using Adobe Photoshop 7.0. Quantitative analysis of the staining intensities of FM1-43 labeled recycling SVs, and VGLUT1 and GluN2B puncta on the dendritic shaft were performed by measuring fluorescence intensities of individual puncta using Image J (version 1.45) (National Institute of Health (NIH), Bethesda, MD, USA) and puncta analyzer software.

[7] 78w A trypan blue exclusion assay was used to assess the viabilities of primary hippocampal cells. Typical images of trypan blue staining after treatment with ST 20 µM are shown in Fig. 1A. ST was administered at concentrations from 2.50 µM to 225 µM and was found to increase cell viability at a concentration ranging from 2.50-150 µM with an optimum concentration of 20 µM (Fig. 1B). Therefore, we carried out further experiments at 20 µM concentration of ST.

[8] 75w To investigate the protection of ST against double-stranded DNA break, hippocampal cultures were exposed to H/R on DIV 9 and doublestained on DIV 13 cultures with anti-phospho γ-H2AX and -tubulin antibodies (Fig. 3A). The number of γ-H2AX puncta per nucleus was significantly (p < 0.001) increased by 100%, in treatment-naïve (-ST) H/ R cultures compared to normoxic control ones (from 4.32 ± 1.37 [(-ST) no H/R]), to 8.39 ± 2.36 [(-ST) H/R]) [Fig. 3B (-ST