Neuroprotective Potential of Raloxifene via G-Protein-Coupled Estrogen Receptors in Aβ-Oligomer-Induced Neuronal Injury

Amyloid-β (Aβ) is one of the causes of Alzheimer’s disease (AD), damaging nerve membranes and inducing neurotoxicity. AD is more prevalent in female patients than in male patients, and women are more susceptible to developing AD due to the decline in estrogen levels around menopause. Raloxifene, a selective estrogen receptor modulator, exhibits protective effects by activating the transmembrane G-protein-coupled estrogen receptor (GPER). Additionally, raloxifene prevents mild cognitive impairment and restores cognition. However, the influence of raloxifene via GPER on highly toxic Aβ-oligomers (Aβo)-induced neurotoxicity remains uncertain. In this study, we investigated the GPER-mediated neuroprotective effects of raloxifene against the neurotoxicity caused by Aβo-induced cytotoxicity. The impact of raloxifene on Aβo-induced cell damage was evaluated using measures such as cell viability, production of reactive oxygen species (ROS) and mitochondrial ROS, peroxidation of cell-membrane phospholipids, and changes in intracellular calcium ion concentration ([Ca2+]i) levels. Raloxifene hindered Aβo-induced oxidative stress and reduced excessive [Ca2+]i, resulting in improved cell viability. Furthermore, these effects of raloxifene were inhibited with pretreatment with a GPER antagonist. Our findings suggest that raloxifene safeguards against Aβo-induced neurotoxicity by modifying oxidative parameters and maintaining [Ca2+]i homeostasis. Raloxifene may prove effective in preventing and inhibiting the progression of AD.


Introduction
The prevalence of dementia is expected to rise as life expectancy increases and the elderly population grows due to advancements in medical technology. It is estimated to reach 175.6 million people worldwide by 2050, up from 130.8 million [1], making it a major medical and social concern. Alzheimer's disease (AD) is the most common form of dementia, affecting over 4% of adults aged 60 and above [2].
AD is a disorder characterized by initial short-term memory loss and disorientation, progressing to psychiatric symptoms such as wandering, irritability, and higher brain dysfunction, eventually leading to bedridden patients within approximately 10 years [3]. The cause of AD remains unknown, but it is pathologically characterized by the formation of senile plaques, which are aggregates of amyloid-β (Aβ), and neurofibrillary tangles composed of tau protein [4]. sulfate, amphotericin B, and fetal bovine serum (FBS) were purchased from Thermo Fisher Scientific K.K. (Waltham, MA, USA). 17β-estradiol and G-15 were procured from Cayman Chemical Company (Ann Arbor, MI, USA), and raloxifene hydrochloride was obtained from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). The other chemicals used in this experiment were commercially available and of the purest grade. Their structures are shown below (Figure 1). Special-grade products were used for other reagents.

Separation and Collection of Aβ Molecular Species with HPLC
Aβo, which is the most toxic form of Aβ, was isolated and collected. Aβ1-42 peptides were dissolved in 10 mM NaOH and sonicated for 2 min; then, a 10-fold concentration of phosphate buffer was added to bring the concentration to 500 μM. The Aβ1-42 solution was passed through a Millex ® -LG filter (0.20 μm; Millipore Ireland BV, Dublin, Ireland) to remove high-molecular-weight substances and incubated at 37 °C for 1 h. After incubation, the Aβ1-42 solutions were centrifuged at 15,000 rpm for 5 min. The supernatant was collected, and the protein concentration was measured using Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The supernatant was diluted to 50 μM with phosphate buffer, and size exclusion chromatography was used to confirm the presence of Aβo. A Superdex 75 increase 10/300 GL column was employed, and the supernatant was fractionated at a flow rate of 0.8 mL/min. The peak of Aβo appeared at 10 min, confirming that Aβo occupied the majority of the Aβ solution ( Figure  2). Previous reports have identified this peak as Aβo [20,21].   and Aβ  The aggregation kinetics of Aβ 1-42 and Aβ 1-40 were measured using a SensoLyte Thioflavin T β-Amyloid (1-42) Aggregation Kit and a SensoLyte Thioflavin T β-Amyloid (1-40) Aggregation Kit (AnaSpec, Inc., Fremont, CA, USA). Thioflavin T (ThT) emits little fluorescence when present free in solution, but the dye emits very strong fluorescence when bound to beta-sheet-rich amyloid structures. To measure the aggregation kinetics of Aβ  or Aβ 1-40 in a 96-well black microplate, 10 µL ThT (2 mM), and 5 µL of raloxifene and estradiol dissolved in dimethyl sulfoxide (DMSO) were added to each well, and 85 µL of Aβ solution was mixed. The final concentration of Aβ 1-42 peptide was 25 µM; those of raloxifene were 1, 5, 10, 20, and 50 µM; and those of estradiol were 1, 5, 10, and 20 µM. ThT fluorescence intensity was monitored at 37 • C for 2 h (Aβ 1-42 ) or 6 h (Aβ 1-40 ) at 15 min intervals with an excitation wavelength of 440 nm and an emission wavelength of 484 nm using a SpectraMax i3 microplate reader (Molecular Devices, LLC., Sunnyvale, CA, USA). The experiments were performed in triplicate.

Separation and Collection of Aβ Molecular Species with HPLC
Aβo, which is the most toxic form of Aβ, was isolated and collected. Aβ 1-42 peptides were dissolved in 10 mM NaOH and sonicated for 2 min; then, a 10-fold concentration of phosphate buffer was added to bring the concentration to 500 µM. The Aβ 1-42 solution was passed through a Millex ® -LG filter (0.20 µm; Millipore Ireland BV, Dublin, Ireland) to remove high-molecular-weight substances and incubated at 37 • C for 1 h. After incubation, the Aβ 1-42 solutions were centrifuged at 15,000 rpm for 5 min. The supernatant was collected, and the protein concentration was measured using Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The supernatant was diluted to 50 µM with phosphate buffer, and size exclusion chromatography was used to confirm the presence of Aβo. A Superdex 75 increase 10/300 GL column was employed, and the supernatant was fractionated at a flow rate of 0.8 mL/min. The peak of Aβo appeared at 10 min, confirming that Aβo occupied the majority of the Aβ solution ( Figure 2). Previous reports have identified this peak as Aβo [20,21].

SH-SY5Y Cell Culture and the Reagent Treatment Method
SH-SY5Y cells were cultured in DMEM/Ham's F-12 medium containing 10% FBS. Differentiation was then induced by culturing for 7 days in medium containing ATRA at a final concentration of 10 µM. Raloxifene and estradiol were dissolved in DMSO to achieve a final DMSO concentration of 0.1%. In the subsequent experiments, cells treated with 0.1% DMSO alone were used as the control. The viability of treated SH-SY5Y cells was evaluated using MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) Cell Proliferation kit I (Nacalai Tesque Inc., Kyoto, Japan). Differentiated SH-SY5Y cells were seeded at a concentration of 1.0 × 10 6 cells/mL onto 96-well collagen-coated plates and incubated at 37 • C for 24 h. First, SH-SY5Y cells were exposed to 0.5 to 10 µM Aβo for 3 h to determine the optimal concentration of Aβo for cytotoxicity.
Next, to examine the protective effect of raloxifene and estradiol against Aβo-induced cytotoxicity, SH-SY5Y cells were treated with Aβo + raloxifene (0.5, 1, and 5 µM) or Aβo + estradiol (0.5, 1, and 5 µM) for 3 h. The effects of GPER and estrogen receptor (ER) on the protective effects of raloxifene and estradiol were confirmed using G-15 (GPER antagonist) and fulvestrant (ER antagonist), respectively. SH-SY5Y cells pretreated with G-15 for 30 min and fulvestrant for 30 min were then treated with Aβo, Aβo + raloxifene, or Aβo + estradiol for 3 h. After incubation, the MTT assay was performed, and the absorbance was measured at 570 nm using SpectraMax i3 (Molecular Devices). The viability of SH-SY5Y cells incubated with 10% FBS-containing medium was showed as 100% and the viability of cells exposed to 1% saponin was expressed as 0%.

Detection of Cell Viability and Cytotoxicity Using Calcein-AM/Ethidium
Homodimer-1 (Live/Dead) Cell Assay Cell viability was measured by simultaneously assessing live and dead cells using calcein-AM and ethidium homodimer-1 (Thermo Fisher Scientific K.K.). Live cells were observed with calcein-AM and dead cells with ethidium homodimer-1. Differentiated SH-SY5Y cells were seeded at a concentration of 1.0 × 10 6 cells/mL onto 96-well collagencoated plates and incubated at 37 • C for 24 h. Cells were then exposed to Aβo or treated with Aβo + 5 µM raloxifene or Aβo + 5 µM estradiol for 3 h. The treated cells were stained with 2 µM calcein-AM and 10 µM ethidium homodimer-1. Calcein-AM is hydrolyzed by ubiquitous intracellular esterases, resulting in green fluorescence that is proportional to the number of viable cells. Ethidium homodimer-1 penetrates only into cells with damaged membranes, binds to nucleic acids, and exhibits red fluorescence in proportion to the number of dead cells. Red fluorescence intensity was measured at excitation (Ex) of 495 nm and emission (Em) of 645 nm, while the green fluorescence intensity was measured at Ex of 495 nm and Em of 530 nm using SpectraMax i3 (Molecular Devices). Additionally, the cytotoxicity of individual cells was assessed by observing with a fluorescence microscope (BZX800; Keyence Co., Osaka, Japan). Red fluorescence intensity is also represented as gray-scale images.
2.6. Detection of Oxidative Stress 2.6.1. Reactive Oxygen Species (ROS) Oxidative stress was measured using 5-(and-6)-chloromethyl-2',7'-dichlorodihydrofluorescein diacetate (CM-H2DCFDA; Thermo Fisher Scientific Inc.). CM-H2DCFDA is taken up intracellularly and converted to 2 ,7 -dichlorofluorescein (DCF) by esterase. DCF is then oxidized by intracellular ROS and exhibits green fluorescence. Differentiated SH-SY5Y cells were seeded in 96-well collagen-coated plates at a concentration of 1.0 × 10 6 cells/mL and incubated at 37 • C for 24 h. After incubation, the cells were pretreated with 5 µM G-15 for 10 min and then treated with Aβo, Aβo + 5 µM raloxifene, or Aβo + 5 µM estradiol for 3 h. ROS production was measured using the SpectraMax i3 microplate reader at Ex of 480 nm and Em of 530 nm. Following fluorescence measurements, the cellular protein content in the plate was determined using Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad Laboratories, Inc., Hercules, CA, USA), and the results were represented as fluorescence intensity per unit protein. Furthermore, the oxidative stress state was observed with a fluorescence microscope (BZX800; Keyence Co.).

Mitochondrial ROS Production
Mitochondrial ROS contribute significantly to intracellular ROS. Intracellular mitochondrial ROS were detected using a Mitochondrial ROS detection kit (701600; Cayman Chemical Company, Ann Arbor, MI, USA). Differentiated SH-SY5Y cells were seeded in 96-well collagen-coated plates at a density of 1.0 × 10 6 cells/mL and incubated at 37 • C for 24 h. After incubation, the cells were pretreated with 5 µM G-15 for 30 min and then treated with Aβo, Aβo + 5 µM raloxifene, or Aβo + estradiol for 3 h and 24 h. Following fluorescence measurements, the cellular protein content in the plate was determined using Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad Laboratories, Inc.), and the results were represented as fluorescence intensity per unit protein. Fluorescence intensity was measured at Ex of 500 nm and Em of 580 nm using Spectra Max i3 (Molecular Devices).

Measurement of Phospholipid Peroxidation in Cell Membranes
Lipid peroxidation in cell membranes was measured using diphenyl-1-pyrenyl phosphine (DPPP) (Dojin Chemical Laboratory, Kumamoto, Japan), a compound with high selectivity for hydroperoxides. DPPP does not fluoresce on its own but reacts quantitatively with hydroperoxides, emitting strong fluorescence. Differentiated SH-SY5Y cells were seeded at a density of 1.0 × 10 6 cells/mL in 96-well collagen-coated plates and incubated at 37 • C for 24 h. After incubation, the cells were pretreated with 5 µM G-15 for 30 min and then treated with Aβo, Aβo + 5 µM raloxifene, or Aβo + estradiol for 30 min and 3 h. Subsequently, 50 µM DPPP was added to the cells and incubated at 37 • C for 10 min. Fluorescence measurements were made using Spectra Max i3 (Molecular Devices) at Ex of 352 nm and Em of 380 nm. Following fluorescence measurements, the cellular protein content on the plate was determined using Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad Laboratories, Inc.), and the results were represented as fluorescence intensity per unit protein. Additionally, the phospholipid peroxidation state was observed using a fluorescence microscope (BZX800; Keyence Co.). Fluorescence intensity values are shown as gray-scale images. Next, the effects of raloxifene and estradiol on [Ca 2+ ] i changes following Aβo exposure were examined by pretreating the cells with raloxifene and estradiol for 10 min. Similarly, to investigate whether the effects of Aβo on [Ca 2+ ] i are mediated by Ca channels or NMDA receptors, the cells were pretreated with 10 µM nicardipine (an L-type voltage-gated Ca channel blocker), or 10 µM MK801 (an NMDA receptor blocker) for 10 min, and changes in [Ca 2+ ] i following Aβo exposure were measured.

Measurement of [Ca 2+ ] i Changes
In addition, 5 µM G-15, a GPER blocker, was pretreated 10 min before the addition of raloxifene and estradiol. The changes in [Ca 2+ ] i were measured using fluorescence at Ex of 485 nm and Em of 525 nm every 5 s for 300 s at 37 • C using FlexStation 3 (Molecular Devices). The results were expressed as 100% of the starting fluorescence intensity.

Statistical Analysis
Each measurement was performed three times, and the experimental results are presented as means ± standard errors of the mean (SEMs). The effects of raloxifene and estradiol were compared with Aβo using Tukey's or Dunnett's post hoc test after analysis of variance (ANOVA). Values with p < 0.05 were considered statistically significant.

Effects of Raloxifene and Estradiol on Aβ 1-42 and Aβ 1-40 Aggregation
A ThT fluorescence assay was used to compare the effects of raloxifene and estradiol on the aggregation kinetics of Aβ 1-42 and Aβ 1-40 peptides.
The aggregation kinetics of the Aβ 1-42 peptide was monitored for 120 min. Since Aβ readily aggregates to form Aβo, aggregation of Aβ 1-42 with 0.1% DMSO alone increased exponentially without delay, reaching fluorescence intensity after 120 min that was approximately four times greater than that observed at the beginning.
Next, to confirm whether the protective effect of raloxifene against Aβo-induced cytotoxicity is mediated by GPER, the same experiments were performed with 5 µM G-15 pretreatment for 30 min, followed by 5 µM raloxifene, 5 µM estradiol, and 5 µM Aβo. G-15 pretreatment resulted in a decrease in viability compared with the G-15-untreated group for both raloxifene and estradiol (p < 0.0001 vs. G-15-untreated for each cell). No effect on viability was observed with G-15 pretreatment for 5 µM Aβo exposure ( Figure 4C).
To determine whether the protective effect of raloxifene against Aβo-induced cytotoxicity is mediated by the ER, the same experiments were performed with 5 µM fulvestrant pretreatment for 30 min, followed by 5 µM raloxifene, 5 µM estradiol, and 5 µM Aβo treatment. Fulvestrant pretreatment did not significantly reduce survival compared with the untreated group for both raloxifene and estradiol ( Figure 4D). These experiments suggest that upon exposure to Aβo for 3 h, the protective effect of raloxifene and estradiol against Aβo-induced cytotoxicity is not mediated by ER stimulation but rather by GPER stimulation.

Detection of Cell Viability and Cytotoxicity Using the Calcein-AM and Ethidium
Homodimer-1 (Live/Dead) Cell Assay Figure 5 shows the results of the calcein-AM/ethidium homodimer-1 staining of SH-SY5Y cells treated with 5 µM Aβo for 3 h in the presence of raloxifene and estradiol. Cells exposed to Aβo exhibited a significant increase in cytotoxicity compared with the control (n = 10, p < 0.0001 vs. control, Tukey's). Raloxifene treatment significantly suppressed Aβoinduced cytotoxicity (p = 0.0001 vs. 5 µM Aβo). Estradiol treatment showed a tendency to suppress Aβo-induced cytotoxicity (p = 0.0705 vs. 5 µM Aβo), although the difference was not significant ( Figure 5A).
When the calcein-AM/ethidium homodimer-1 staining of SH-SY5Y cells was evaluated with fluorescence microscopy, increased red fluorescence indicative of dead cells was observed in cells exposed to Aβo ( Figure 5G,K). In raloxifene-treated cells, a decrease in red fluorescence associated with Aβo exposure was observed ( Figure 5H,I,L,M).

Effects of Raloxifene and Estradiol on Aβo-Induced Oxidative Stress
Aβo induces oxidative stress, and increased oxidative stress contributes to cell-membrane damage and cell death, suggesting that oxidative stress plays an important role in the pathogenesis of AD [22]. Next, the protective effects of raloxifene and estradiol against Aβo-induced oxidative stress were investigated.
When the calcein-AM/ethidium homodimer-1 staining of SH-SY5Y cells was evaluated with fluorescence microscopy, increased red fluorescence indicative of dead cells was observed in cells exposed to Aβo ( Figure 5G,K). In raloxifene-treated cells, a decrease in red fluorescence associated with Aβo exposure was observed ( Figure 5H,I,L,M).

Effects of Raloxifene and Estradiol on Aβo-Induced Oxidative Stress
Aβo induces oxidative stress, and increased oxidative stress contributes to cell-membrane damage and cell death, suggesting that oxidative stress plays an important role in the pathogenesis of AD [22]. Next, the protective effects of raloxifene and estradiol against Aβo-induced oxidative stress were investigated. To further confirm whether raloxifene's inhibitory effect on oxidative stress induced by Aβo involves GPER, we conducted identical experiments with a 30 min pretreatment of 5 µM G-15, followed by 5 µM raloxifene, 5 µM estradiol, and 5 µM Aβo. Pretreatment with G-15 resulted in increased ROS production in both raloxifene + 5 µM Aβo (p = 0.0351) and estradiol + 5 µM Aβo (p = 0.0004 vs. G-15-untreated in each cell) compared with the G-15-untreated group. No effect on ROS production was observed with G-15 pretreatment for 5 µM Aβo exposure ( Figure 6A). G-15-untreated group. No effect on ROS production was observed with G-15 pretreatment for 5 μM Aβo exposure ( Figure 6A).

Mitochondrial ROS
SH-SY5Y cells exposed to 5 µM Aβo demonstrated a significant increase in mitochondrial ROS production at 3 h (n = 8, p = 0.0024, Tukey) and 24 h of exposure (n = 8, p = 0.0010, Tukey's) compared with the control, confirming the oxidative stress effect of Aβo on mitochondria. At 3 h of exposure, a significant decrease in mitochondrial ROS levels was observed in cells treated with raloxifene (p < 0.0001 vs. 5 µM Aβo) and estradiol (p < 0.0001 vs. 5 µM Aβo) compared with those stimulated with Aβo alone (Figure 7A). At 24 h of exposure, cells treated with raloxifene (p = 0.0454 vs. 5 µM Aβo) exhibited a significant decrease in mitochondrial ROS compared with cells stimulated with Aβo alone, whereas cells treated with estradiol (p = 0.2059 vs. 5 µM Aβo) showed no significant decrease in mitochondrial ROS compared with cells stimulated with Aβo alone ( Figure 7B).

Detection of Cell-Membrane Phospholipid Peroxidation Capacity
Aβo is believed to directly bind to membrane lipids, damaging the phospholipid bilayer structure and entering cells [23]. In this study, we examined changes in plasmamembrane phospholipid peroxidation induced by Aβo exposure.
As shown in Figure 8A,B, there was a significant increase in plasma-membrane phospholipid peroxidation in cells exposed to 5 µM Aβo compared with the control at 30 min (n = 10, p = 0.0003, Tukey's) and 3 h (n = 10, p = 0.0009, Tukey's) of exposure. At 30 min of exposure, SH-SY5Y cells treated with raloxifene or estradiol exhibited significant inhibition of Aβo-induced plasma-membrane phospholipid peroxidation ( Figure 8A). Moreover, at 3 h of exposure, SH-SY5Y cells treated with raloxifene exhibited significant inhibition of Aβoinduced plasma-membrane phospholipid peroxidation, and those treated with estradiol tended to inhibit Aβo-induced plasma-membrane phospholipid peroxidation ( Figure 8B). To further confirm whether raloxifene's inhibitory effect on increased plasma-membrane phospholipid peroxidation by Aβo involves GPER, we pretreated with 5 µM G-15 for 10 min, followed by 5 µM raloxifene + Aβo and 5 µM estradiol + Aβo for 30 min. G-15 pretreatment had no significant effect on plasma-membrane phospholipid peroxidation compared with the G-15-untreated group for both raloxifene and estradiol at 30 min and 3 h of exposure to Aβo. No effect on plasma-membrane phospholipid peroxidation was observed with G-15 pretreatment for 5 µM Aβo exposure.
(n = 10, p = 0.0003, Tukey's) and 3 h (n = 10, p = 0.0009, Tukey's) of exposure. At 30 min of exposure, SH-SY5Y cells treated with raloxifene or estradiol exhibited significant inhibition of Aβo-induced plasma-membrane phospholipid peroxidation ( Figure 8A). Moreover, at 3 h of exposure, SH-SY5Y cells treated with raloxifene exhibited significant inhibition of Aβo-induced plasma-membrane phospholipid peroxidation, and those treated with estradiol tended to inhibit Aβo-induced plasma-membrane phospholipid peroxidation ( Figure 8B). To further confirm whether raloxifene's inhibitory effect on increased plasma-membrane phospholipid peroxidation by Aβo involves GPER, we pretreated with 5 μM G-15 for 10 min, followed by 5 μM raloxifene + Aβo and 5 μM estradiol + Aβo for 30 min. G-15 pretreatment had no significant effect on plasma-membrane phospholipid peroxidation compared with the G-15-untreated group for both raloxifene and estradiol at 30 min and 3 h of exposure to Aβo. No effect on plasma-membrane phospholipid peroxidation was observed with G-15 pretreatment for 5 μM Aβo exposure.
Fluorescence microscopy images are shown in Figure 8G-J. DPPP reacted with hydroperoxides and emitted intense fluorescence, which was enhanced by Aβo exposure ( Figure 8H) but reduced by raloxifene and estradiol treatments ( Figure 8I,J). The fluorescence intensity values are also presented as gray-scale images.

Changes in [Ca 2+ ] i Levels following Treatment with Raloxifene and Estradiol
As depicted in Figure 9A, there was a rapid increase in [Ca 2+ ] i immediately after exposure to 5 µM Aβo, reaching a peak of 25.5% at 60 s and then maintaining a nearly constant increased level. When loaded with Ca 2+ -free buffer, no increase in [Ca 2+ ] i was observed after exposure to 5 µM Aβo. However, when pretreated with 10 µM Nicardipine and 10 µM MK801, [Ca 2+ ] i increased by only 10% immediately after the addition of 5 µM Aβo. ]i in SH-SY5Y cells after the addition of Aβo was measured using fluorescence intensity for 300 sec. SH-SY5Y cells were exposed to 5 μM Aβo in Ca 2+ -containing or Ca 2+ -less buffer. Furthermore, SH-SY5Y cells were exposed to 5 μM Aβo after calcium-channel antagonist (10 μM nicardipine) or an NMDA receptor blocker (10 μM MK801) pretreatment for 10 min. (B) SH-SY5Y cells were supplemented with raloxifene or estradiol after G-15 pretreatment or no pretreatment. (C) SH-SY5Y cells treated with raloxifene were exposed to 5 μM Aβo after G-15 pretreatment or no pretreatment. (D) SH-SY5Y cells treated with estradiol were exposed to 5 μM Aβo after G-15 pretreatment or no pretreatment. The time of administration is indicated by an arrow. Fluorescence intensity was expressed as 100% of the value at the onset.

Discussion
In this experiment, we demonstrated that raloxifene provided protection against Aβo-induced neuronal cytotoxicity in SH-SY5Y cells. Aβ is produced as a monomer via sequential two-step cleavage of intracellularly produced APP by β-secretase1 and γ-secretase. The extracellularly secreted Aβ monomer then aggregates to form oligomers, which exhibit neuronal cytotoxicity [24]. For this study, we utilized high-molecular-weight Aβo, which is considered the most toxic form of Aβo [25].
Estrogen is a steroid hormone primarily produced in the granulosa cells of the ovary. As an estrogen receptor (ER) agonist, estrogen promotes proliferation in mammary and endometrial cells, maintains bone cell health, and affects lipid metabolism. Estradiol replacement therapy was introduced in the 1960s to counter the decrease in BMD observed in postmenopausal women due to a rapid decline in estradiol levels. However, concerns  2+ ] i in SH-SY5Y cells after the addition of Aβo was measured using fluorescence intensity for 300 sec. SH-SY5Y cells were exposed to 5 µM Aβo in Ca 2+ -containing or Ca 2+ -less buffer. Furthermore, SH-SY5Y cells were exposed to 5 µM Aβo after calcium-channel antagonist (10 µM nicardipine) or an NMDA receptor blocker (10 µM MK801) pretreatment for 10 min. (B) SH-SY5Y cells were supplemented with raloxifene or estradiol after G-15 pretreatment or no pretreatment. (C) SH-SY5Y cells treated with raloxifene were exposed to 5 µM Aβo after G-15 pretreatment or no pretreatment. (D) SH-SY5Y cells treated with estradiol were exposed to 5 µM Aβo after G-15 pretreatment or no pretreatment. The time of administration is indicated by an arrow. Fluorescence intensity was expressed as 100% of the value at the onset.
Upon exposure to raloxifene and estradiol, [Ca 2+ ] i temporarily increased by approximately 110% immediately after exposure but quickly decreased and maintained a constant [Ca 2+ ] i thereafter. When pretreated with G-15, there was no transient increase in [Ca 2+ ] i immediately after raloxifene and estradiol addition ( Figure 9B). Furthermore, Figure 9C,

Discussion
In this experiment, we demonstrated that raloxifene provided protection against Aβo-induced neuronal cytotoxicity in SH-SY5Y cells. Aβ is produced as a monomer via sequential two-step cleavage of intracellularly produced APP by β-secretase1 and γ-secretase. The extracellularly secreted Aβ monomer then aggregates to form oligomers, which exhibit neuronal cytotoxicity [24]. For this study, we utilized high-molecular-weight Aβo, which is considered the most toxic form of Aβo [25].
Estrogen is a steroid hormone primarily produced in the granulosa cells of the ovary. As an estrogen receptor (ER) agonist, estrogen promotes proliferation in mammary and endometrial cells, maintains bone cell health, and affects lipid metabolism. Estradiol replacement therapy was introduced in the 1960s to counter the decrease in BMD observed in postmenopausal women due to a rapid decline in estradiol levels. However, concerns were raised about the increased risk of estrogen-dependent tumors, such as breast and uterine cancer [26].
Raloxifene is an ER agonist that acts as an agonist on bone tissue, similar to estrogen, while acting as an antagonist on mammary gland cells and endometrial tissue. This selective action minimizes the risk of estrogen-dependent tumors. Hence, it is categorized as an SERM and is recommended for treating osteoporosis in the early postmenopausal period [27]. Both raloxifene and estradiol are fat-soluble drugs capable of crossing the bloodbrain barrier via simple diffusion, allowing them to enter the central nervous system. Recent studies have reported several beneficial effects of raloxifene on the central nervous system, such as marked inhibition of brain tissue damage progression in mice with traumatic brain injury [28].
A study investigating the effects of raloxifene on AD indicated a trend toward suppressing cognitive decline when taken for less than a year [29]. However, a large clinical trial conducted in 2015 in patients with mild cognitive impairment (MCI) and AD demonstrated that raloxifene was not effective in preventing the onset of AD [18]. Nevertheless, recent studies have shown that Aβ aggregation already accumulates in the brain tissues of patients with AD in the stage when cognitive decline becomes apparent [19]. This highlights the importance of preclinical therapeutic intervention before Aβ accumulation in brain tissue. Therefore, no clear evidence has been presented regarding the preventive effect of raloxifene on the onset of AD.
Recent research has revealed expression of the estrogen receptor, GPER (a member of the seven transmembrane receptor superfamily), in various cells and tissues, including cell membranes and intracellular organelles such as mitochondria and the endoplasmic reticulum, in addition to the classical ER nuclear receptor. Previous studies have shown that GPER stimulation with estradiol is involved in promoting cell proliferation and differentiation via the activation of the Notch signaling pathway and the Motogen-activated protein kinase (MAPK) pathway [30,31]. Studies on mouse cardiomyocytes have reported that treatment with G-1, a GPER stimulator, suppresses the increase in [Ca 2+ ] i induced by L-type calcium channel stimulation and exhibits cardioprotective effects [32]. However, the involvement of GPER and the impact of raloxifene on changes in [Ca 2+ ] i in the central nervous system remain unknown.
In our ThT assay experiments, raloxifene demonstrated concentration-dependent inhibition of Aβ 1-42 aggregation, while estradiol did not inhibit Aβ 1-42 aggregation (Figure 3). Aβ  undergoes folding at the Glu22-Asp23 site, forming a hairpin structure that alters intramolecular polarity and promotes oligomer formation [33]. Previous reports have already indicated the inhibitory effect of raloxifene on Aβ 1-42 aggregation, suggesting its direct interaction with the N-terminal and intermediate domains of the Aβ 1-42 peptide, leading to the destabilization of preformed Aβ 1-42 fibrils [15]. Conversely, estradiol did not exhibit any impact on Aβ 1-42 aggregation, implying that it does not bind to or act on Aβ 1-42 peptides as observed with raloxifene. This disparity is likely attributed to the presence of an independent phenolic hydroxyl group in raloxifene, which is believed to inhibit Aβ aggregation [34]. As noted, intracellularly produced APP is cleaved by Aβ monomers via two-step enzymatic reactions and secreted extracellularly. Therefore, it presumably suppresses the extracellular aggregation reaction to Aβo without the effect of estrogen receptors, which are present on the cell membrane or in intracellular organelles.
In clinical practice, the pharmacokinetic concentration of raloxifene is 1.635 ng/mL or 3.45 nM for a single 120 mg dose. The concentration of raloxifene used in this experiment (0.5, 1.0, and 5.0 µM) is 100-1000 times higher than that, which poses a consideration for future studies. However, it is worth noting that even at one-fifth of the concentration of Aβ, raloxifene inhibited aggregation when the concentration of Aβ was 25 µM ( Figure 3A). The average Aβ 1-42 concentration in cerebrospinal fluid is 792 pg/mL or 0.175 nM [35]. Comparatively, one-fifth of this raloxifene concentration amounts to 0.035 nM. Considering that the concentration in the brain is approximately 1/100 of that in the blood and that the total protein concentration in the cerebrospinal fluid is 1/200 of that in blood, the dose required for raloxifene to bind to Aβ and inhibit aggregation in the brain is not noticeably distant from the dose used in clinical practice. However, the accumulation of Aβ in the brain is presumably involved in the pathogenesis of patients with AD, and the concentration of Aβ (Aβ1-40 + Aβ1-42) in the brain of cognitively normal older adults is 130-600 nM [36], which is higher than that in the cerebrospinal fluid, depending on the brain region. The amount of orally ingested raloxifene entering the brain parenchyma has still not been established. Therefore, in the future, it will be necessary to estimate the appropriate dosage by evaluating the in vitro ability of orally ingested raloxifene to enter brain tissue.
In our experiment, both raloxifene and estradiol inhibited the decrease in cell viability induced by Aβo exposure (Figure 4B) and demonstrated inhibition of neuronal cytotoxicity ( Figure 5A,H,I). Nuclear ER is expressed in neurons like other cells. Raloxifene, like estradiol, acts as an agonist in neurons [37] and triggers the production of growth factors such as BDNF (brain-derived neurotrophic factor), leading to increased cell viability. However, when conducting similar experiments with pretreatment using fulvestrant, a selective ER blocker, no effect on cell viability was observed compared with non-treating with fulvestrant ( Figure 4D). Conversely, pretreatment with G-15, a GPER blocker, prevented the increased viability of Aβo-exposed cells seen with raloxifene and estradiol treatment ( Figure 4C). These results suggest that raloxifene and estradiol exert their protective effects against short-term neuronal injury caused by Aβo exposure via a mechanism mediated by GPER stimulation.
Aβo, formed via extracellular aggregation, is believed to induce toxicity by directly damaging neuronal cell membranes [38]. One of the pathological mechanisms involves Aβo directly impairing neuronal cell membranes, forming ion-channel-like pores that alter the membrane potential and result in sustained Ca 2+ influx into cells [39]. Our study also showed persistently high [Ca 2+ ] i levels due to Aβo exposure, but this response was suppressed by pretreatment with Nicardipine, an L-type Ca channel blocker, or MK801, a non-competitive NMDA receptor antagonist ( Figure 9A). MK801, like memantine, inhibits the NMDA receptor and exhibits a cell-protective effect [40]. In this experiment, MK801 selectively inhibited the ion-channel portion of NMDA receptors that opened upon Aβo exposure. In the presence of raloxifene and estradiol pretreatment, there was a suppressive effect on the Aβo-induced increase in [Ca 2+ ] i . These findings suggest that raloxifene or estradiol binds to the nicardipine-binding sites in the Ca 2+ channel and NMDA receptor, inhibiting the increase in [Ca 2+ ] i caused by Aβo. Previous studies using cultured rat cortical neurons have demonstrated that raloxifene inhibits the glutamate-induced increase in [Ca 2+ ] i by blocking voltage-gated calcium channels [41].
Inferred from the viability results ( Figure 4C,D) of this experiment, the suppressive effect of pretreatment with raloxifene or estradiol on the Aβo-induced increase in intracellular calcium concentration ([Ca 2+ ] i ) could be mediated by GPER on the neural membrane.
However, when G-15 pretreatment was administered before raloxifene or estradiol treatment, it further inhibited the Aβo-induced increase in [Ca 2+ ] i compared with the absence of G-15 treatment ( Figure 9C,D). GPER stimulation activates phospholipase C, leading to the formation of phosphatidylinositol 4,5 bisphosphate (PIP2), which constitutes cell-membrane phospholipids. As a result, inositol 1,4,5 trisphosphate (IP3) and diacylglycerol (DAG) are formed. These reactions occur at an extremely rapid rate, taking only a few milliseconds, and IP3 facilitates the release of Ca 2+ from the endoplasmic reticulum. As shown in Figure 9B, treatment with raloxifene or estradiol alone resulted in a transient increase in [Ca 2+ ] i 10 s after addition. As previously described, raloxifene or estradiol binds to the nicardipine binding site and the NMDA receptor action site, thereby suppressing the Aβo-induced increase in [Ca 2+ ] i . However, raloxifene or estradiol transiently increased [Ca 2+ ] i , and this transient increase was blocked by G-15 pretreatment. In summary, G-15 pretreatment in combination with raloxifene suppressed Aβo-induced increases in [Ca 2+ ] i more effectively than raloxifene pretreatment alone.
Oxidative stress induces biochemical changes in neurons and is implicated in the disease progression of many neurological disorders, particularly neurodegenerative diseases such as AD [20]. This experiment also observed various oxidative stress responses, including increased production of ROS due to Aβo exposure, elevated mitochondrial ROS production, and enhanced peroxidation capacity of cell-membrane phospholipids. Treatment with raloxifene and estradiol demonstrated suppression of the oxidative stress response induced by Aβo exposure, including ROS production and mitochondrial ROS production. However, G-15 pretreatment counteracted the inhibition of these oxidative stress responses (Figures 6 and 7). Persistently high levels of [Ca 2+ ] i have been shown to overload mitochondria with calcium and increase ROS production [42]. However, considering the effects of GPER treatment on Aβo-induced changes in [Ca 2+ ] i , it is unlikely that the inhibitory effects of raloxifene and estradiol on Aβo-induced ROS production can be solely explained by mechanisms mediated by membrane GPER stimulation. It has been reported that raloxifene or estradiol increase the expression of Glutathione-SH, an antioxidant in cells [43,44]. Based on these previous studies, the increased expression of antioxidants, including Glutathione-SH, by raloxifene or estradiol may be caused by the stimulation of GPER, which is expressed in intracellular organelles such as mitochondria and the endoplasmic reticulum. In this study, mitochondrial ROS levels increased after 24 h of Aβo exposure in cells treated with G-15 alone compared with those without G-15 treatment. It was speculated that G-15, which had penetrated the cells sufficiently after prolonged exposure, inhibited the basic GPER activity of intracellular organelles, thereby preventing optimum antioxidant function. However, the GPER activity of intracellular organelles is not clear and needs further research.
Aβo exposure also increased membrane phospholipid peroxidation, a type of oxidative stress response. In neuronal cell membranes damaged by Aβo exposure, free radicals initiate a chain reaction with the phospholipid portion of the cell membrane, generating lipid peroxide. This, in turn, leads to structural and functional changes in membrane component proteins, resulting in neuronal cytotoxicity [45]. In this study, when raloxifene or estradiol were administered simultaneously with Aβo, there was a significant inhibition of Aβo-induced phospholipid peroxidation in the raloxifene treatment, which differed significantly from the estradiol treatment. Furthermore, there was no significant difference in the inhibition of phospholipid peroxidation between G-15 pretreatment and no treatment. These findings suggest that raloxifene suppresses Aβo-induced plasma-membrane phospholipid peroxidation without GPER, thereby exerting its neuroprotective effect.
In addition to the previously mentioned ion-channel-like pore formation in neuronal cell membranes, it has been reported that Aβo binds to and disrupts phospholipids in the plasma membrane with its surfactant-like action, owing to its amphiphilic affinity. The N-termini of Aβo tend to interact with the hydrophilic phospholipid heads, while their C-termini penetrate hydrophobic sites [46]. Raloxifene, in addition to its inhibitory effect on Aβ aggregation, has also been reported to destabilize the formed oligomers [15]. This suggests that raloxifene binds to the Aβ oligomer, attenuating its harmful effects on cells, or partially degrades the Aβ-oligomer structure, thereby suppressing the destruction of plasma-membrane phospholipids.
In this experiment, raloxifene inhibited Aβo-induced increases in [Ca2+]i to an extent similar to that exerted by estradiol. Moreover, raloxifene exhibited superior potency to estradiol in inhibiting Aβ aggregation and Aβ-oligomer-induced oxidative stress and exhibited neuroprotective effects against Aβo exposure ( Figure 10). The mechanism by which raloxifene treatment significantly improved Aβo exposure-induced cell viability compared with estradiol treatment is attributed to its ability to inhibit Aβo aggregation and promote Aβo degradation [15].

Conclusions
Raloxifene demonstrated its ability to provide protection against Aβo-induced neuronal cell injury via its inhibitory effects on Aβ aggregation and GPER-mediated antioxidant mechanisms. These findings suggest that raloxifene could potentially play a beneficial role in preventing the onset and progression of AD.

Conclusions
Raloxifene demonstrated its ability to provide protection against Aβo-induced neuronal cell injury via its inhibitory effects on Aβ aggregation and GPER-mediated antioxidant mechanisms. These findings suggest that raloxifene could potentially play a beneficial role in preventing the onset and progression of AD.

Conflicts of Interest:
The authors declare no conflict of interest.