The Mechanism of Catalpol in Alzheimer’s Disease by Modulating Mitochondrial Function Through miR-124-Mediated STIM2

Alzheimer’s disease (AD) is a dementia-related disease with cognitive deterioration and memory impairment. Catalpol was reported to relieve impairments in learning and memory. The present study assessed the functional mechanism of catalpol in AD via miR-124/STIM2-mediated mitochondrial function. Primary hippocampal neurons were isolated and cultured. AD cell model was induced by Aβ 1−42 and treated with catalpol. APP/PS1 mouse model was established and treated with catalpol and miR-124 agomir. Aβ 1−42 induced mitochondrial damage and reactive oxygen species (ROS) generation in AD cell model. Catalpol alleviated mitochondrial damage and reduced ROS generation in hippocampal neurons. miR-124 was highly expressed in AD cell model and catalpol inhibited miR-124 expression. Catalpol alleviated Aβ 1−42 induced mitochondrial damage and ROS generation in hippocampal neurons by inhibiting miR-124 expression. miR-124 overexpression after catalpol treatment promoted mitochondrial damage and ROS generation in hippocampal neurons. miR-124 targeted STIM2. Silencing STIM2 after catalpol treatment promoted mitochondrial damage and ROS generation in hippocampal neurons. Catalpol slowed AD progression via the miR-124/STIM2 axis in vivo. The results of the present study indicated that catalpol alleviated mitochondrial damage and ROS generation and thus attenuated AD by regulating miR-124-mediated STIM2.


Introduction
Alzheimer's disease (AD) is a brain disease often occurring in the elderly and causes dementia and frailty . AD accounts for 50%-70% of dementia cases among people above 60 years old (Sochocka et al., 2017;Lashley et al., 2018). The incidence rate of AD is 1% among people aged 65 years and 8% among people aged 85 years or above (Reitz, 2015). Given the fact that AD has caused heavy economic and social burdens on patients, novel therapeutic strategies for AD are in urgent need.
Catalpol is an iridoid glucoside widely distributed in many plants (Bai et  miR-124, a neurosystem-speci c microRNA (miRNA), plays a regulatory role in the progression of AD . miR-124 also regulates mitochondrial activity in amyotrophic lateral sclerosis (Yardeni et al., 2018). Catalpol could regulate miR-124 expression in neurons after ischemia . Stromal interaction molecule 2 (STIM2) is also involved in the development of AD (O'Day, 2020). The downregulation of STIM2 protein causes disrupted store-operated Ca 2+ entry pathway in AD spines, consequently leading to memory loss . Therefore, this study investigated the potential mechanism of catalpol in attenuating AD via modulating the miR-124/STIM2-mediated mitochondrial function.

Ethics statement
This study was approved by the Ethics Committee of The Seventh A liated Hospital, Sun Yat-sen University. All experiment procedures were approved by the Laboratory Animal Committee in accordance with the guidelines of International Association for the Study of Pain.
Primary culture of primary hippocampal neurons in mice As previously mentioned (Taylor et al., 2014), mixed cortex and hippocampal neurons were isolated from fetal mice of embryonic day 14-16. The isolated hippocampal neurons were cultured with Dulbecco's modi ed Eagle's medium (DMEM; Gibco, Grand Island, NY, USA ) containing 10% FBS and 1% penicillinstreptomycin (Gibco) in an incubator with 5% CO 2 at 37°C.

Cell treatment and grouping
According to the reference , AD cell model was induced using Aβ 1−42 (5 µM). STIM2 was silenced using recombinant lentivirus eukaryotic expressing plasmid (sh-NC, sh-STIM2) and packaged into lentivirus. Above lentivirus was prepared in large quantities and puri ed. Corresponding lentivirus was transfected into hippocampal neurons at logarithmic phase when con uence reached 30%.
After 48 h of transfection, each well was added with 1 µg/mL puromycin to screen stably-transfected cells. miR-124 mimic or its control (GenePharma, Shanghai, China) was transfected into hippocampal neurons using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) as per the instructions. Catalpol treatment was conducted based on an earlier research , and the concentration of catalpol used was 500 µM. Cells were assigned into 9 groups: control group, AD group, AD + DMSO group, AD + catalpol group, AD + DMSO + mimic NC group, AD + catalpol + mimic NC group, AD + catalpol + miR-124 mimic group, AD + catalpol + sh-NC group, and AD + catalpol + sh-STIM2 group.

Measurement of Adenosine triphosphate (ATP) synthesis
A total of 1 × 10 4 cells were harvested and incubated with 200 µL cell lysis buffer according to the ATP determination kit (Beyotime). Cells were aspirated and centrifuged at 1200 × g for 5 min. Next, 100 µL supernatants were moved to prepared wells for 5 min. ATP luminescence was detected using a microplate reader (Synergy 2 Multi-Mode microplate reader, BioTek, Winooski, VT).

Measurement of ROS production
The process of DCFH-DA and DHR 123 immuno uorescence was as follows. Hippocampal neurons were collected from each group and seeded into 96-microwell plates (237105, Thermo Fisher Scienti c, Rockford, IL, USA) at the density of 3 × 10 4 per well. Total cellular ROS was measured using DCFH-DA (Cell Biolabs, San Diego, CA, USA). After 10 min of incubation with DCFH-DA (10 mM) at 37°C, DCF uorescence was detected using a uorescence microplate reader (In nite F 200, Tecan Japan, Kanagawa, Japan). Based on manufacturer's instructions, the excitation and emission wavelength was 485 nm and 535 nm, respectively. Mitochondrial ROS generation was determined using dihydrorhodamine 123 (DHR 123) assay. Fluorescence microscope was used for photographing and observation.
Mitochondrial superoxide in the hippocampus was determined (Jia et al., 2016). The hippocampus slices (300 µM) of each group were incubated for 30 min using MitoSOX (5 µM, Invitrogen) to determine mitochondrial ROS level in vivo. The uorescence of oxidative dye in the slices was detected with a laser confocal microscope (TCS-SP2, Leica, Wetzlar, Germany). The excitation and emission wavelength was 498 nm and 522 nm, respectively. The ROS uorescence values were analyzed using the Leica SP2 software.
Reverse transcription-quantitative PCR (RT-qPCR) Total RNA was extracted using TRIzol reagent (Invitrogen) according to the instructions. Reverse transcription of RNA and antisense miRNA to cDNA was performed using PrimeScript RT kit (RR037A, Takara, Tokyo, Japan) with 10 µL reaction system based on the instructions. Reaction solution was used for uorescence qPCR according to the introductions of SYBR®Premix ExTaqTM kit (RR820A, TaKaRa) using real-time uorescent quantitative PCR system (ABI 7500, ABI, Foster City, CA, USA). With U6 as an internal control, relative expression of each target gene was calculated using the 2 −ΔΔCt method. Each experiment was repeated 3 times. ΔΔCt = ΔCt test group -ΔCt control group and ΔCt = Ct target gene -Ct internal control. Related primers were designed by Sangon Biotech (Shanghai, China) ( Table 1). Table 1 Primer sequences Name of primer Sequences Notes: miR, microRNA; F, forward; R, reverse

Western blot
Tissues or cells were collected with trypsin digestion and then lysed with enhanced radioimmunoprecipitation assay (RIPA) buffer (Boster, Wuhan, China) containing protease inhibitor. Protein concentration was detected using the bicinchoninic acid (BCA) kit (Boster). Protein samples were separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto the polyvinylidene uoride (PVDF) membranes. After 2 h of blocking with 5% bovine serum albumin (BSA) to block the non-speci c binding, diluted primary antibodies STIM2 (ab181258; 1:1000, Abcam, Cambridge, UK) and GAPDH (ab9485; 1:2500, Abcam) were added and incubated with the membranes at 4°C overnight. Then, the membranes were washed and added with HRP labeled goat anti-rabbit (ab6721; 1:2000, Abcam) and incubated for 1 h, followed by incubation of transfer lm with ECL working solution at room temperature for 1 min. Excess ECL reagent was removed and the membranes were sealed with fresh-keeping lm. X-ray lm was placed in the cassette for 5-10 min-exposure before imaging and xation. The bands of each group in Western blot images were analyzed by Image J software. GAPDH was internal control. Each experiment was repeated 3 times.

Dual-luciferase reporter assay
The 3'UTR dual-luciferase reporter gene vector of STIM2 wild type and mutant-type plasmid containing miR-124 binding sites were constructed. The STIM2-WT, STIM2-MUT, reporter plasmid miR-124 mimic plasmid and negative control plasmid were cotransfected into 293T cells, respectively. Cells were lysed 24 h after transfection and centrifuged at 12000 rpm for 1 min. Supernatants were collected. Luciferase activities were measured using Dual-Luciferase® Reporter Assay System (E1910, Promega, Madison, WI, USA). Each cell sample was added with 100 µL Fire y luciferase working solution to detect Fire y luciferase, and added with 100 µL Renilla luciferase working solution to detect Renilla luciferase. The Fire y luciferase and Renilla luciferase were taken as relative luciferase activity. Each experiment was repeated 3 times.

Animals and grouping
A total of 30 healthy male APP/PS1 transgenic AD mice aged 3 months (average weight, 28 ± 2 g) and 6 C57BL/6J WT mice (average weight, 26 ± 4 g) of the same background and growth phase were purchased from JKboit (Nanjing, China).

Morris water maze (MWM) test
As previously mentioned (Pei et al., 2015), MWM test was performed to examine the abilities of spatial learning and memory of mice in this study. Brie y, the water maze consisted of a round bathtub (120 cm in diameter and 60 cm in height) and a circular platform (6 cm in diameter) which was submerged 1 cm below the water surface. The bathtub was located in an environment with rich visual hints outside the maze. Mice were given 20 min to adapt to the test environment before experiment. Invisible platform training was performed for 6 days consecutively and each included 4 trials. During each trial, mice were released from the walls of the box and given 60 s to search and stand on the invisible platform. If mice failed to reach the platform in the given time, they would be guided to the platform manually. Orientation navigation experiment was conducted 24 h after training. In this experiment, the platform was removed and performance of each mouse within 60 s was recorded. The escape latency before reaching the platform, the time spent in each quadrant and the times of platform crossing were recorded.

Hematoxylin and eosin (HE) staining
The tissues were embedded in para n, xed on the microtome and sectioned at 4 µm. The slices were placed in water at 46°C and spread, and then dried in a slide drier for 2 h at 72°C. Next, the slices were cooled for 10 min, cleared with xylene I for 10 min and xylene II for 10 min, dehydrated with ethanol I for 5 min and ethanol II for 5 min, followed by 90% ethanol for 2 min, 80% ethanol for 2 min, 70% ethanol for 2 min. After washed with owing water for 5 min, the slices were stained with hematoxylin for 5-10 min, washed for 5 min and treated with hydrochloric ethanol for 2-3 s. After washed with owing water for 5 min, the slices were rinsed with lithium carbonate for 5 min to turn back to blue, washed for 10 min and stained with eosin for 2 min. After washed for 5 min, the slices were treated with 8-% ethanol for 2 min, 90% ethanol for 2 min, absolute ethanol for 2 min, and xylene for 2 min. Finally, the slices were swabbed, sealed with neutral resin and observed under a light microscope.

Statistical analysis
All data were analyzed using SPSS 21.0 statistical software (IBM Corp., Armonk, NY, USA). The normality and homogeneity of variance tests were conducted rst. Measurement data were expressed as means ± standard deviation (SD). The data conformed to normal distribution between two groups were compared by independent sample t test, while comparison among multiple groups was made by one-way analysis of variance (ANOVA), followed by Tukey's multiple comparisons test. Comparison among multiple groups at multiple time points was made by repeated measures ANOVA, followed by Bonferroni test. P < 0.05 was considered statistically signi cant.

Catalpol inhibited Aβ 1−42 -induced injury in AD cell model by mediating mitochondrial function
Recent study has showed that catalpol exerts neuroprotective effects in AD . To investigate whether catalpol participates in AD, AD cell model was induced by Aβ 1−42 and mitochondrial functions were determined after treating AD cell model with catalpol. RT-qPCR showed that mtDNA level was decreased in AD group and increased in AD + catalpol group (Fig. 1A). ΔΨm measurement result showed that ΔΨm in AD group was in green, suggesting decreased ΔΨm and mitochondrial function; on the other hand, ΔΨm in AD + catalpol group was in red, suggesting increased ΔΨm and enhanced mitochondrial function (Fig. 1B). ATP measurement observed decreased ATP generation in AD group, which suggested injured mitochondria; while ATP generation in AD + catalpol group was increased (Fig. 1C). The measurement of ROS generation found that total ROS and mitochondrial ROS generation in AD group were signi cantly increased while total ROS and mitochondrial ROS generation in AD + catalpol group were decreased (Fig. 1D-E). . RT-qPCR result showed that miR-124 expression was increased in AD group but decreased in AD + catalpol group (Fig. 1F). Above results indicated that miR-124 was highly expressed in AD cell model and catalpol inhibited miR-124 expression.

Catalpol attenuated Aβ 1−42 -induced injury in AD cell model via miR-124-mediated mitochondrial function
To explore whether catalpol affects AD by inhibiting miR-124 expression, miR-124 was overexpressed in AD cell model treated with catalpol. RT-qPCR result showed that miR-124 expression was remarkably downregulated in AD + catalpol + mimic NC group but upregulated in AD + catalpol + miR-124 mimic group (Fig. 2A). The measurement mtDNA level was increased in AD + catalpol + mimic NC group but decreased in AD + catalpol + miR-124 mimic group (Fig. 2B). The ΔΨm in AD + catalpol + mimic NC was increased and mitochondrial function was enhanced, while ΔΨm in AD + catalpol + miR-124 mimic group was decreased and mitochondrial function was suppressed (Fig. 2C). The mitochondrial ATP generation was increased in AD + catalpol + mimic NC group but decreased in AD + catalpol + miR-124 mimic group (Fig. 2D). The total ROS and mitochondrial ROS generation were decreased in AD + catalpol + mimic NC group but increased in AD + catalpol + miR-124 mimic group (Fig. 2E-F). These results suggested that catalpol treatment on hippocampal neurons alleviated Aβ 1−42 -induced mitochondrial damage and reduced ROS generation by inhibiting miR-124, and miR-124 overexpression after catalpol treatment intensi ed mitochondrial damage in hippocampal neurons and ROS generation.

miR-124 targeted STIM2
A recent study has showed that STIM2 affects the progress of AD (Deng et al., 2020). Western blot showed that STIM2 was poorly expressed in AD cell model (Fig. 3A). The binding sites of miR-124 and STIM2 were predicted on bioinformatics website Targetscan (Fig. 3B). Dual-luciferase reporter assay showed that luciferase activity in co-transfection group of miR-124 and STIM2-WT was downregulated while no obvious difference was observed in other co-transfection groups of STIM2-MUT (P > 0.05, Fig. 3C). To clarify the effect of miR-124, miR-124 expression was detected by RT-qPCR after overexpressing miR-124 in AD cells. The result suggested that miR-124 expression was obviously increased in cells treated with miR-124 mimic (Fig. 3D). Western blot showed that STIM2 expression was downregulated in cells treated with miR-124 mimic (Fig. 3E). These ndings revealed that miR-124 targeted STIM2 in AD cell model.

Catalpol mediated AD development and progression through miR-124/STIM2-mediated mitochondrial function in vivo
To study whether the catalpol/miR-124/STIM2 axis participates in AD, APP/PS1 transgenic AD mouse model was established. RT-qPCR and Western blot results showed that miR-124 was upregulated while STIM2 was downregulated in APP/PS1-treated mice, and miR-124 was downregulated while STIM2 was upregulated in APP/PS1 + catalpol-treated mice; miR-124 was upregulated while STIM2 was downregulated in APP/PS1 + catalpol + miR-124 agomir-treated mice (Fig. 5A-B). MWM test showed that the escape latency of APP/PS1-treated mice was prolonged and the time spent in the quadrant of the platform and the number of crossings the platform were reduced; on the other hand, the escape latency of APP/PS1 + catalpol-treated mice was shortened and the time spent in the quadrant of the platform and the number of crossings the platform was increased; the escape latency of APP/PS1 + catalpol + miR-124 agomir-treated mice was prolonged and the time spent in the quadrant of the platform and the number of crossings the platform was reduced (Fig. 5C-E). HE staining revealed that the hippocampal tissue injury was aggravated in APP/PS1-treated mice; the hippocampal tissue injury of mice treated with APP/PS1 + catalpol was alleviated; while that of mice treated with APP/PS1 + catalpol + miR-124 agomir was exacerbated (Fig. 5F). As for ROS expression, in APP/PS1-treated mice, ROS expression was increased while ROS expression was decreased in mice treated with APP/PS1 + catalpol; ROS expression was elevated in mice treated with APP/PS1 + catalpol + miR-124 agomir (Fig. 5G). Collectively, catalpol slowed AD progression through miR-124/STIM2-mediated mitochondrial function.

Discussion
AD is a common cause of dementia and a major cause of death, the diagnosis and treatment of which require careful and comprehensive medical evaluation (Alzheimer's, 2016). Catalpol has been found to play a neuroprotective role in neurodegenerative disorders such as AD and Parkinson's disease (Jiang et al., 2015). In this study, we investigated the mechanism of catalpol in AD in mice. The result showed that catalpol could mitigate AD progression in mice by regulating the miR-124/STIM2-mediated mitochondrial function.
Catalpol mitigated Aβ 1−42 -induced blood-brain barrier disruption in AD . In the present study, AD cell model was induced by Aβ 1−42 to investigate the role of catalpol in mitochondrial function in AD. The alteration of mtDNA has been reported to affect AD pathogenesis (Yan et al., 2013). The mtDNA level in AD cells was decreased while mtDNA level was increased after catalpol treatment. The decreased ATP generation has been reported in AD (Agrawal & Jha, 2020). The result showed that ATP generation was decreased in AD cells while ATP generation was increased after catalpol treatment. Mitochondrial ROS accumulation and loss of ΔΨm were found in AD (Lee et al., 2018). Our study found that total ROS and mitochondrial ROS generation in AD cells were signi cantly increased while total ROS and mitochondrial ROS generation were decreased after catalpol treatment. ΔΨm in AD cells was decreased while ΔΨm was increased after catalpol treatment. Catalpol reduces ROS level and enhanced mitochondrial ΔΨm in mice with liver injury, and reduces mitochondrial damage in ischemic injury ( These results suggested that catalpol slowed AD progression in vivo. Essentially, miR-124 agomir treatment in APP/PS1 transgenic AD mice treated with catalpol annulled the protection of catalpol in AD. The suppression of miR-124 in hippocampal neurons enhances cognitive performance (Dutta et al., 2013). Taken together, catalpol slowed AD progression via miR-124/STIM2-mediated mitochondrial function in vivo.
In conclusion, our study supported that catalpol alleviated AD by mitigating mitochondrial damage and ROS production through the miR-124/STIM2 axis. Since this study did not provide other potential mechanisms and downstream pathways involving catalpol in AD, further investigations are needed in the eld of AD treatment.  Catalpol inhibited Aβ1-42-induced injury in AD cell model by mediating mitochondrial function. Notes: A: mtDNA level in hippocampal neurons was detected by RT-qPCR; B: ΔΨm change was detected using JC-1 kit; C: ATP generation in hippocampal neurons was detected; D: total ROS generation in hippocampal neurons was detected by DCFH-DA uorescence; E: mitochondrial ROS generation in hippocampal neurons was detected by DHR123 uorescence; F: miR-124 expression was detected by RT-qPCR. * P < 0.05, compared with AD + DMSO + mimic NC group, # P < 0.05, compared with AD + catalpol + miR-124 mimic group. Data in the graphs were measurement data and expressed as mean ± SD. Unpaired t test was performed for pairwise comparison. The experiment was repeated 3 times.
Notes: A: miR-124 expression was detected by RT-qPCR; B: mtDNA level in hippocampal neurons was detected by RT-qPCR; C: ΔΨm change was detected using JC-1 kit; D: ATP generation in hippocampal neurons was detected; E: total ROS generation in hippocampal neurons was detected by DCFH-DA uorescence; F: mitochondrial ROS generation in hippocampal neurons was detected by DHR 123 uorescence. * P < 0.05, compared with AD + DMSO + mimic NC group, # P < 0.05, compared with AD + catalpol + miR-124 mimic group. Data in the graphs were measurement data and expressed as mean ± SD. Unpaired t test was performed for pairwise comparison. The experiment was repeated 3 times. the binding relation of miR-124 and STIM2 was predicted on bioinformatics website Targetscan; C: the binding relation of miR-124 and STIM2 was veri ed by dual-luciferase reporter assay; D: miR-124 expression was detected by RT-qPCR; E: STIM2 expression in each group was detected by Western blot. * P < 0.05, compared with control group and mimic NC group. Data in the graphs were measurement data and expressed as mean ± SD. Unpaired t test was performed for pairwise comparison. The experiment was repeated 3 times.

Figure 4
Catalpol inhibited Aβ1-42-induced injury in AD cell model through miR-124/STIM2-mediated mitochondrial function. Notes: A: STIM2 expression in AD cell model was detected by Western blot; B: mtDNA level in hippocampal neurons was detected by RT-qPCR; C: ΔΨm change was detected using JC-1 kit; D: ATP generation in hippocampal neurons was detected; E: total ROS generation in hippocampal neurons was detected by DCFH-DA uorescence; F: mitochondrial ROS generation in hippocampal neurons was detected by DHR123 uorescence. * P < 0.05, compared with AD + catalpol + sh-NC group.
Data in the graphs were measurement data and expressed as mean ± SD. Unpaired t test was performed for pairwise comparison. The experiment was repeated 3 times. mitochondrial ROS expression in hippocampal tissues was detected. * P < 0.05, compared with normal group, # P < 0.05, compared with APP/PS1 + DMSO group, & P < 0.05, compared with APP/PS1 + catalpol + agomir NC group. Measurement data were expressed as mean ± SD. Comparison among multiple groups were performed using one-way ANOVA, followed by Tukey's multiple comparisons test. Comparison among multiple groups at multiple time points was made by repeated measures ANOVA, followed by Bonferroni test. N = 6.