Q.Y. Chen1, Y. Yin2, L. Li1, Y.J. Zhang1, W. He1, Y. Shi3
1. Department of Neurology, The Third Affiliated Hospital of Qiqihar Medical University, Qiqihar 161000, P.R. China; 2. Department of Teaching & Research, The Third Affiliated Hospital of Qiqihar Medical University, Qiqihar 161000, P.R. China; 3. College of Medical Technology, Qiqihar Medical University, Qiqihar 161006, P.R. China
Corresponding Author: Yan Shi College of Medical Technology, Qiqihar Medical University, No. 333, Bukui Street, Jianhua District, Qiqihar 161006, Heilongjiang Province, P.R. China E-mail: email@example.com, ORCID:0000-0001-6254-3201, Tel: +86-0452-2663346
J Prev Alz Dis 2021;
Published online November 17, 2021, http://dx.doi.org/10.14283/jpad.2021.60
BACKGROUND: Alzheimer’s disease (AD) is a major cause of dementia, which is a growing global health problem and has a huge impact on individuals and society. As the modifying role of geniposidic acid (GPA) has been suggested in AD, this study sets out to determine if and how GPA treatment affects AD progression in mice.
METHODS: Potential downstream target genes of GPA during AD were identified by bioinformatics analysis, revealing GAP43 as a primary candidate protein. Then, mPrP-APPswe/PS1De9 AD transgenic mice were treated with GPA via intragastric administration. This allowed for gain- and loss-of-function assays of candidate proteins being carried out with or without GPA treatment, after which behavioral tests could be conducted for mice. Cortical neuron apoptosis was measured by TUNEL staining, Amyloid β-protein (Aβ) expression in cerebral cortex by Thioflavin-s staining, and Aβ, IL-1β, IL-6, IL-4 and TNF-α levels in cerebral cortex by ELISA. GAP43 expression in cerebral cortex of mice was detected by immunohistochemistry. Primary cortical neurons of embryonic mice were isolated and induced by Aβ1-42 to construct AD cell model. Cell viability was assessed by CCK-8, and axon growth by immunofluorescence.
RESULTS: GPA administration significantly improved the cognitive impairment, reducing Aβ accumulation and neuronal apoptosis in AD mice, and alleviated inflammation and axonal injury of Aβ1-42-induced neurons. GAP43 was shown experimentally to be the target of GPA in AD. Silencing of GAP43 repressed the neuroprotective effect of GPA treatment on AD mice. GPA elevated GAP43 expression via PI3K/AKT pathway activation and ultimately improved nerve injury in AD mice.
CONCLUSION: GPA activates a PI3K/AKT/GAP43 regulatory axis to alleviate AD progression in mice.
Key words: Alzheimer’s disease, geniposidic acid, GAP43; PI3K/AKT pathway, neuroinflammation, neuron, axon, nerve injury.
A s the principal cause of dementia, Alzheimer’s disease (AD) is a growing global health problem which has a huge impact at an individual and societal level (1, 2). This disease is caused by amyloid plaques and neurofibrillary tangles that accumulate in the brain and result in gradual cognitive decline (3). As a neurodegenerative and prominent protein conformational disease, AD presents itself as memory loss and progressive neurocognitive dysfunction (4). Despite several treatments have been developed to alleviate mild symptoms, no drugs are currently in use which improve cognition or prevent AD progression (5). Therefore, furthering our understanding of molecular mechanism of AD is of great clinical significance and could reveal novel therapeutic targets.
Geniposidic acid (GPA), one of the main active components of Gardenia jasminoides J. Ellis (Rubiaceae), has several beneficial physiological effects such as blood pressure regulation, cancer prevention, repairing soft tissue injury, osteoporosis treatment, anti-inflammation, anti-thrombosis, and anti-platelet aggregation (6). Moreover, it has been shown that GPA can relieve the spatial learning, memory deficits, and neuroinflammation in AD-model mice (7). Interestingly, a prior study has reported that geniposide can upregulate growth-associated protein 43 (GAP43) which leads to a reduction in fluoxetine-suppressed neurite outgrowth in Neuro2a neuroblastoma cells (8). However, a direct relationship between GPA and GAP43 has not been fully explored. As an axonal membrane protein which is abundantly expressed in neuronal growth cones, GAP43 plays a critical role in the stabilization of synapse structure, axonal regeneration, and neural growth (9). Importantly, upregulation of GAP43 has been identified to be involved in the attenuation of behavioral deficits, modification of synaptic structure, and acceleration of neurite outgrowth in AD (10). Intriguingly, geniposide has also been shown to activate PI3K/AKT signaling, reducing neonatal mouse brain injury after hypoxic-ischemia treatment (11). In addition, activation of the PI3K/AKT pathway is thought to lead to the upregulation of GAP43 during hair cell protection in the neonatal murine cochlea (12). Furthermore, activation of the PI3K/AKT pathway might be capable of subduing cognitive deficits in AD (13).
Given these findings we hypothesized that the suppressive role GPA treatment plays in AD is via a PI3K/AKT/GAP43 regulatory axis. Therefore, we set out to determine whether GPA treatment affects Aβ accumulation, neuronal apoptosis, neuroinflammation and axonal injury via the PI3K/AKT/GAP43 regulatory axis.
Materials and methods
CTD database (http://ctdbase.org/) and SymMap database (https://www.symmap.org/) were used to predict the targets of GPA. Gene expression dataset GSE28146 related to AD in mice species was obtained from Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/). GPL570 was the platform file of the gene expression dataset. The samples of the gene expression dataset were grouped, including 8 control samples and 7 AD samples. Differential analysis was conducted using R language «limma» package (http://www.bioconductor.org/packages/release/bioc/html/limma.html) to screen the differentially expressed genes (DEGs) with |logFC| > 1 and p < 0. 05 as the screening criteria. The top 50 DEGs with the minimum p value were selected to draw a heat map using «pheatmap» package (https://cran.r-project.org/web/packages/pheatmap/index.html) of R language. The drug-target network was visualized by Cytoscape 3.5.1 software.
GPA (purity: HPLC ≥ 98%, wkq16052101) was purchased from Sichuan Victory Biological Technology Co., Ltd. (Chengdu, China) and dissolved into distilled water at a final concentration of 2.5 mg/mL, 5 mg/mL and 7.5 mg/mL. The GPA administration time of mice was 10:00-12:00 a.m.
Animal treatment and grouping
Experimental male mPrP-APPswe/PS1De9 AD transgenic mice (AD mice, male, aged 6 months) were purchased from Beijing HFK Biotechnology Co., Ltd. (Beijing, China). The animals were kept at ambient temperature of 23 ± 1℃ with a 12 h light/dark cycle and relative humidity of 55 ± 5%, and free access to food and water. Before the experiment, mice were left to adapt to this environment for 2 to 3 days. AD mice were randomly arranged into treatment group and control group. AD mice were treated with GPA (25 mg/kg, 50 mg/kg, and 75 mg/kg) or normal saline (NS) via intragastric administration for three months. At the same time, 50 μg/kg LY294002 (S1105, Selleck, Houston, TX, USA) was injected 30 min after GPA administration or short hairpin RNA (sh)-negative control (NC)/sh-GAP43 adenovirus (1 × 1011 plaque-forming units, 10 μL, Shanghai Geneland Biotech Co., Ltd., Shanghai, China) was injected every two weeks into the lateral ventricle (The injection site relative to bregma position: 1 mm in the positive position, 1.5 mm in the right position, and 3.5 mm in the depth). The drug was injected into the ipsilateral ventricle with a Hamilton syringe (Hamilton Company, Renault, NV, USA). The needle was left for 5 min after injection and then slowly withdrawn for more than 5 min. After the needle was removed, the burr hole was sealed with bone wax. The 90 AD mice were randomly assigned into 9 groups with 10 mice in each group: AD + NS group (AD mice were fed with NS), AD + GPA 25 mg/kg group (AD mice were given with GPA at 25 mg/kg by gavage), AD + GPA 50 mg/kg group (AD mice were given with GPA at 50 mg/kg by gavage), AD + GPA 75 mg/kg group (AD mice were given with GPA at 75 mg/kg by gavage), AD + GPA + sh-NC group (AD mice were given with GPA at 75 mg/kg by gavage, and sh-NC adenovirus was injected into lateral ventricle), AD + GPA + sh-GAP43 group (AD mice were given with GPA 75 mg/kg by gavage, and sh-GAP43 adenovirus was injected into lateral ventricle), AD + GPA + overexpression (oe)-NC (GPA at 75 mg/kg was given to AD mice and oe-NC adenovirus was injected into lateral ventricle), AD + GPA + LY294002 + oe-NC (GPA at 75 mg/kg was given to AD mice, and oe-NC adenovirus and PI3K inhibitor LY294002 (14) were injected into the lateral ventricle at the same time), AD + GPA + LY294002 + oe-GAP43 (GPA at 75 mg/kg was given to AD mice, and oe-GAP43 adenovirus and PI3K inhibitor were injected into lateral ventricle at the same time). Next, 10 C57BL/6 mice (WT mice) of the same age were fed with NS as control: WT + NS (WT mice were fed with NS) group. Morris water maze (MWM) test and new object recognition test were carried out as described below. After behavioral experiment, mice were euthanized and brain tissue was prepared into paraffin section. The experiment was approved by the animal ethics committee of Qiqihar Medical University.
Animal behavioral experiment
MWM test was performed 90 to 95 days after administration. The MWM instrument (Chinese Academy of Medical Sciences, Beijing, China, DMS-2) consisted of a circular container (100 cm in diameter and 40 cm in height), a recording and analysis system, and a digital camera (TOTA Group Limited, Japan). The instrument was located in a low light test room. The water was made opaque by adding food grade white colorant and kept at 20 ± 2℃. During the training, the cylindrical escape platform (10 cm in diameter) was located in the center of the north and South quadrant of the pool, with a depth of about 0.5 cm. The space training test was conducted for 5 consecutive days, and the mice were put into the pool at the planned starting position every day. The mice were trained to find the platform within 60 s. If the mouse did not find the platform within 60s, the mouse was guided to the platform, and the mouse was allowed to stand on the platform for 20 s after reaching the platform. This was repeated four times a day. The swimming track of each mouse was recorded by a camera. The exploration experiment was carried out within 24 h after the 5-day positioning navigation test, and the platform was removed. The mice swam in the swimming pool for 60 s. The swimming trajectory of the mice within 60 s was recorded and analyzed. Within 24 h after the 5-day WMW experiment, the spatial exploration test of memory recovery ability was carried out, and the platform of shadow was removed. The mice were placed in a water tank with their faces facing the wall in a randomly selected quadrant. The crossing times of the platform location within 60 s were recorded. After that, in order to eliminate the potential impact of motor ability and visual obstacles on the experimental results, the visual platform test was conducted after the spatial exploration test. All the mice were placed in the same quadrant, and the mice were able to see the platform. The time required for mice to find the visible platform was recorded.
On the first day of the new object recognition test, the mice were gently placed in the test box for 5 min. On the next day, mice were gently placed in the test box with 2 identical objects. On the third day, the mice were gently placed in the test box, but one of the objects had been replaced by a new one. When the mice directly touched the object with their mouth, forehead or nose, the exploration time was recorded. The discovery index was that the time spent on a new object was divided by the cumulative time spent on two objects.
Thioflavin-s (Th-S) staining
Thioflavin was a fluorescent dye which was usually adopted to stain senile plaques. Firstly, the brain tissues of mice were fixed with paraformaldehyde, dehydrated, and embedded in paraffin. The paraffin-embedded tissues were cut with a slicer and fixed. The slices were dehydrated, hydrated in distilled water, and stained with Mayer’s hematoxylin for 5 min. The slides were exposed to Th-S solution (1% in distilled water) for 5 min. The slices were immersed in 70% alcohol for 5 min and sealed by glycerin gelatin. The slices were observed with an optical microscope (Olympus BX 41 microscope, 40 × magnification), and the number of senile plaques was calculated by Image Pro Plus software.
Enzyme-linked immunosorbent assay (ELISA)
The levels of interleukin (IL)-1β (Beyotime, Shanghai, China, PI301), IL-6 (Beyotime, PI326), IL-4 (Beyotime, PI612), TNF-α (Beyotime, PT512), and Aβ 1-42 (Mosake Biological Technology Co., Ltd., Wuhan, China, 69-21411) in mouse cortical homogenate were detected according to the instructions of ELISA kit. The antibody was diluted to a protein content of 1-10 μg/mL through 0.05 M carbonate coated buffer solution (PH = 9). The 0.1 mL diluent was added into the reaction well of each polystyrene plate which was kept at 4℃ overnight. The next day the plate was washed with washing buffer for 3 times. The 0.1 mL diluted sample to be tested was added into the coated reaction well and incubated at 37℃ for 1 h. The 0.1 mL fresh diluted enzyme-labeled antibody was added to each reaction well. The 0.1 mL tetramethylbenzidine substrate solution was added into each reaction well for 10-30 min of incubation at 37℃, followed by addition of 0.05 mL of 2 M sulfuric acid into each reaction well. At 450 nm, the optical density (OD) value of each well was measured using a microplate reader and corrected to blank control wells. Each experiment was repeated three times.
TdT-mediated dUTP-biotin nick end-labeling (TUNEL) staining
A TUNEL apoptosis detection kit (Millipore Corp., Billerica, MA, USA) was used to detect apoptosis in brain tissues according to the instructions provided. Paraffin mouse brain tissue sections were dewaxed in xylene for 5-10 min, with fresh xylene for 5-10 min, with anhydrous ethanol for 5-10 min, with 90% ethanol for 2 min, and with 70% ethanol for 2 min, and washed with distilled water for 2 min. The sections were treated with 20 μg/mL protease K (ST532, Beyotime) at 20-37℃ for 15-30 min. The sections were incubated in 3% hydrogen peroxide solution prepared by phosphate buffer saline (PBS) for 20 min at room temperature, and then incubated with biotin-labeled solution at 37℃ for 60 min. The sections were incubated with labeling reaction termination solution at room temperature for 10 min. The sections were incubated in 50 μL Streptavidin-horseradish peroxidase (HRP) solution at room temperature for 30 min. The 0.2-0.5 mL diaminobenzidine (DAB) chromogenic solution was added for incubation at room temperature for 5-30 min, followed by sealing. The sections were observed and photographed under an inverted microscope. Ten visual fields in each group were randomly selected and the number of positive cells and total cells in each were counted. The cells with brownish yellow nuclei were noted as apoptotic positive cells, while those with blue nuclei were classed as normal healthy cells. The apoptosis rate was expressed as the percentage of brown-yellow cells.
Isolation and culture of primary cortical neurons
Primary cortical neurons were prepared from the cortex of 19-day embryo mice. In short, the embryonic mouse cortex was dissected in cold PBS. The tissues were collected and washed in PBS, then treated with 0.25% trypsin at 37℃ for 15 min. Finally, fetal bovine serum (FBS, Gibco, Carlsbad, California, USA) was added at the final concentration of 10% to stop the trypsinization. After, the neurons were precipitated by 10-min centrifugation at 800 rpm and resuspended with the neural basal medium (Gibco) containing 10% FBS in a cell incubator at 37℃ and 5% CO2. The next day, the basic medium was discarded and renewed with fresh medium in the presence of serum. On the third day of culture, the proliferation of glial cells was inhibited by cytarabine at final concentration of 10 μmoL/L, and glial cells were aspirated after 24 h. The cells cultured in vitro for 7-21 days were used for the experiment.
Identification of primary cortical neurons
On the third and seventh day of seeding, the morphological changes of the cells were observed using an optical microscope. The purity of neurons was identified by Neuron Specific Enolase (NSE) immunofluorescence staining. The cells were fixed with 4% paraformaldehyde at room temperature for 10 min. After the paraformaldehyde was removed, the cells were permeabilized for 15 min by treating with PBS supplemented with 0.12% Triton X100, and blocked with 5% bovine serum albumin (BSA) at room temperature for 1 h. Then cells were further cultured with specific rabbit anti-NSE (ab180943, 1:100, Abcam, Cambridge, UK) in a wet box at 4℃ overnight, and incubated with fluoresceinisothiocyanat (FITC)-labeled rabbit anti secondary antibody (ab6717, 1:100, Abcam) in the dark at room temperature for 1 h, followed by the addition of 4’,6-Diamidino-2-Phenylindole at room temperature in the dark for 5 min. Cells were sealed with anti-quench sealing tablet. NSE positive cells were observed under a fluorescence microscope. The percentage of NSE positive cells in the total number of cells in the visual field was the purity of neurons (Figure S1).
The core plasmid (pLKO.1) and auxiliary plasmid (psPAX2, pMD2.G) of target gene silencing sequence were applied to package the silencing lentivirus. The core plasmid (pHAGE-CMV-MCS-IzsGreen) and auxiliary plasmid (psPAX2, pMD2.G) of target gene cDNA sequence were adopted to package the overexpression lentivirus. The lentivirus was purchased from Shanghai Sangon Biotechnology Co. Ltd. (Shanghai, China). The primer sequence and plasmid were constructed by Shanghai Sangon Biotechnology Co., Ltd. The packaging virus and the target vector were co-transfected into 293T cells by Lipo2000 (11668-019, Invitrogen, Carlsbad, CA, USA). The supernatant was collected after 48 h of cell culture. The supernatant after filtration and centrifugation contained virus particles. Virus titer was detected. Viruses in exponential phase of growth were harvested and were arranged into sh-NC group (silencing lentivirus control group), sh-GAP43-1 group (silencing GAP43 lentivirus 1 group), and sh-GAP43-2 group (silencing GAP43 lentivirus 2 group) according to the different transfectants. The medium was renewed 8-h post transfection. After 96 h of transfection, the transfection efficiency was observed under an inverted fluorescence microscope. Silencing lentiviral transfection and silencing sequences are shown in Table S1.
Establishment of AD cell model in vitro
For AD cell model construction, 1 mg of Aβ1-42 (Anaspec, San Jose, CA, USA) lyophilized powder was dissolved in hexafluoroisopropanol (HFIP) (Sigma-Aldrich, St Louis, MO, USA) at the final concentration of 1 mg/mL, followed by water bath and ultrasonic wave treatment for 10 min. Aβ1-42 was stored at room temperature in dark for 5-24 h. The 0.1 mg powder was obtained by nitrogen drying and stored at -80℃. Then, the powder was added with dimethyl sulfoxide (DMSO) to the final concentration of 1 mg/mL, and diluted with PBS to the working concentration of 100 μM/L, and cultured in a constant temperature incubator at 37℃ for more than 3 h, allowing a Aβ1-42 oligomer to form. The activated Aβ1-42 (20 μM/L) was added to cortical neuron culture medium for 24 h to construct AD cell model.
Cell counting kit (CCK)-8 assay
A CCK-8 (K1018, APExBIO, Boston, MA, USA) kit was used to evaluate the cell viability of cell model which was treated with GPA or Aβ1-42. The cells were placed in 96 well plates (100 μL/well; 1 × 104 cells/well). Firstly, cells were treated with 20 μM/L GPA for 2 h, and then treated with activated Aβ1-42 respectively for 3 h, 6 h, 12 h, and 24 h. Six duplicated wells were set up. At the same time, the control group was added with the same volume of DMSO as Aβ1-42. After treatment, 10 μL CCK8 solution was added to each well and incubated at 37℃ for 2 h.
Next, the absorbance at 450 nm was measured using a microplate reader. The cell viability (%) was calculated by the formula [(As – Ab)/(Ac – Ab)] × 100%. As represents the absorbance value of supernatant from exposed or false exposed dishes, Ac represents the absorbance of the well containing the supernatant in the normal control, and Ab represents the absorbance value of the culture well containing 10% CCK-8 solution.
The culture plate was washed twice with PBS at 37℃ and the cells were fixed in 4% paraformaldehyde (pH 7.4) for 15 min. After blocking in 3% normal Donkey Serum (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) and permeating in 0.1% Triton X-100, cells were cultured with mouse anti-class III β-tubulin monoclonal antibody (1: 200, Beyotime) overnight in the dark room at 4℃. Afterwards, Cy3 donkey anti-mouse Immunoglobulin G (IgG) (1:400, Jackson ImmunoResearch Laboratories) and Dylight488 donkey anti-rabbit IgG (1:400, Jackson ImmunoResearch Laboratories) were incubated with cells. Then, the cells were stained with Hoechst 33342 (Sigma-Aldrich). The samples were observed under a confocal microscope (LSM 710; Carl Zeiss, Oberkochen, Germany), and the length of axon growth was measured by ZEN2009 software (Carl Zeiss). For each group and experiment, three fields of vision were observed in each cover glass. Each experiment was repeated three times.
Western blot analysis
Radio-Immunoprecipitation assay cell lysis buffer containing phenylmethylsulfonyl fluoride (P0013B, Beyotime) was added to lyse cells. The supernatant was harvested and the total protein concentration of each sample was measured with a bicinchoninic acid kit as per the manufacturers protocol (P0011, Beyotime). The adjusted protein concentration was 1 μg/μL. The volume of sample in each tube was set at 100 μL. The sample was boiled at 100℃ for 10 min to denature the protein which was stored at – 80℃ for use. According to the size of the target protein, 8%-12% sodium dodecyl sulfate gel was prepared, and the proteins were added to the lanes in equal amounts for electrophoresis. The protein on the gel was transferred to a polyvinylidene fluoride membrane (1620177, Bio-Rad Laboratories, Hercules, CA, USA). The membrane was sealed with 5% skimmed milk or 5% BSA at room temperature for 1 h. The membrane was probed with primary rabbit antibodies (Abcam) to PI3K (ab154598, 1:1000,) and GAP43 (ab75810, 1:1000) and primary rabbit antibodies [Cell Signaling Technologies (CST), Beverly, MA, USA] to AKT (#9272, 1:1000), phosphorylation (p)-AKT (Serine473, #9271, 1:1000), and β-actin (#4970, 1:5000) overnight at 4℃. The membrane was re-probed with horseradish peroxidase (HRP)-labeled goat anti-rabbit IgG (ab6721, 1:5000, Abcam) secondary antibody at room temperature for 1 h. The membrane was immersed in electrogenerated chemiluminescence reaction solution (1705062, Bio-Rad Laboratories) at room temperature for 1 min. The liquid was removed, and the membrane was covered with plastic wrap. Strip exposure imaging was performed on the Image Quant LAS 4000C gel imager (Amersham Biosciences/GE Healthcare, Piscataway, NJ, USA). β-actin was used as the internal reference. The ratio of gray value of target band to internal reference band was used as the relative expression level of protein. Each experiment was repeated three times.
Reverse transcription quantitative polymerase chain reaction (RT-qPCR)
Total RNA was extracted from cells and mouse cerebral cortex using Trizol (16096020, Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). Then the total RNA was reversely transcribed into complementary (cDNA) by using PrimeScript RT Kit (Takara Biotechnology Ltd., Dalian, China). RT-qPCR experiment was carried out with a RT-qPCR kit (Q511-02, NanJing Vazyme Biotech Co., Ltd, Nanjing, China) according to the instructions. PCR amplification was performed on a Bio-rad quantitative real-time PCR system CFX96. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was utilized as the internal reference. The primer sequence was designed and provided by Shanghai Sangon Biotechnology Co. Ltd. The 2-ΔΔCt method was the multiple ratio relationship between the experimental group and the control group. The primer sequences are described in Table S2. Each experiment was repeated three times.
Immunohistochemistry (IHC) staining
The brain tissue sections of mice were baked at 60℃ for 20 min, placed in xylene solution successively, and then soaked for 15 min after xylene solution replacement. Mouse brain tissue sections were dehydrated with 100%, 95%, 90%, 85% and 80% ethanol. After that, 3% H2O2 was added to each section, which was soaked at room temperature for 10 min to block the endogenous peroxidase activity. The sections were added to citric acid buffer, heated in a microwave oven for 3 min, and cultured with antigen repair solution at room temperature for 10 min. The sections were incubated with normal goat serum blocking solution (Shanghai Sangon Biotechnology Co., Ltd.) at room temperature for 20 min. Then, the sections were incubated with primary rabbit anti-GAP43 antibody (ab75810, 1:500, Abcam) for a night in the dark room. The next day, the sections were incubated with goat anti-rabbit IgG secondary antibody (ab6721, 1: 1000, Abcam) for 30 min, and then added with streptavidin-biotin complex (vector company, USA) in a 37℃ incubator for 30 min. Diaminobenzidine chromogenic Kit (P0203, Beyotime) was utilized to add one drop of chromogenic reagent to the specimen for coloring for 6 min. The sections were stained in hematoxylin for 30 s. The sections were put into 70%, 80%, 90%, 95% ethanol and anhydrous ethanol for 2 min each time. Finally, the sections were cleared twice in xylene for 5 min and sealed with neutral resin. The sections were observed under an upright microscope (BX63, Olympus Optical Co., Ltd, Tokyo, Japan). Each experiment was repeated three times.
SPSS 21.0 (IBM Corp. Armonk, NY, USA) was adopted for statistical analysis. The measurement data is presented as the mean ± standard deviation throughout this study. Unpaired t-tests were implemented for comparison between the two groups. One-way analysis of variance (ANOVA) was applied for comparison among multiple groups, while repeated measurement ANOVA was utilized to compare the data at different time points, followed by Tukey’s post-hoc test. p < 0. 05 was considered to be statistically significant difference.
GPA alleviates cognitive impairment in AD mice
In order to explore the role of GPA in the cognitive dysfunction of AD mice, the behavior of AD mice was initially assessed. MWM test results showed that after 5 days of positioning navigation test, the time for mice to find the hidden platform reduced gradually. Compared with the WT group, the escape latency of the AD group was significantly longer, while the escape latency of the AD + GPA group was significantly reduced (Figure 1A). On the last day of the MWM test, the hidden platform was removed for exploratory testing. The results showed that compared with the WT group, the times crossing the platform was significantly reduced in the AD group. In contrast to the AD group, mice in the AD + GPA group crossed the platform more regularly (Figure 1B).
A, The escape latency of mice on different days detected by the positioning navigation test. B, The times that mice crossed the platform in each group. C, Swimming time of mice in each group in the target quadrant, D, Exploration index of mice on day 1 in new object recognition test. E, Exploration index of mice on day 2 in new object recognition test. * p < 0.05 vs. the WT group. # p < 0.05 vs. the AD group, n=10.
Visual platform test results showed that, compared with the WT group, the time it took for mice to find the visual platform was significantly increased in the AD group, while this time was dramatically decreased in the GPA treated AD + GPA group (Figure 1C). The results of the new object recognition test demonstrated that there was no significant difference in the exploration index of all groups when the mice were placed in a box with two identical objects (Figure 1D). When one of the objects was replaced by a new one on the next day, the exploration index in the AD group was strikingly lower than that of the WT group. Compared with the AD group, the exploratory index of the AD + GPA group was clearly elevated (Figure 1E). In conclusion, GPA treatment significantly improved the cognitive impairment of AD mice in a dose dependent manner.
GPA alleviates Aβ accumulation, neuronal apoptosis, neuroinflammation and axonal injury in AD mice
We then turned our attention to examine the mechanisms underpinning GPA-mediated alleviation of cognitive impairment, more specifically effects on Aβ deposition, neuroinflammation, and axonal injury in AD mice. ELISA analysis revealed that the level of Aβ in the AD group was notably higher than that in the WT group. Compared with the AD group, the level of Aβ in the AD + GPA group was reduced, with a clear inverse correlation between GPA concentration and Aβ levels (Figure 2A). In addition, the use of Th-S staining further demonstrated that GPA triggered reduction of the accumulation of Aβ in cerebral cortex (Figure 2B). TUNEL analysis showed, that compared with the WT group, the apoptosis rate was higher in the AD group, an effect that was reduced when mice were given GPA, where again a dose dependent response was observed. (Figure 2C). As reflected by ELISA analysis, the expression of IL-6, TNF-α and IL-1β was strikingly enhanced and the expression of IL-4 was reduced in the AD group, while an opposite trend was observed in the AD + GPA group (Figure 2D). Next, primary cortical neurons of mice were induced by Aβ1-42 to provide an in vitro cell model of AD, followed by identification of primary cortical neurons, shown in Supplementary result (Figure S1A-C). Then, as CCK-8 assay demonstrated, when compared with the DMSO group, cell viability of the Aβ group was markedly decreased. In contrast to the Aβ group, cell viability of the GPA + Aβ group was noticeably increased (Figure 2E). Immunofluorescence microscopy revealed that axon length was observably shorter in the Aβ group than in the DMSO group, which was longer in the GPA + Aβ group (Figure 2F).
A, ELISA to detect the level of Aβ in the brain of mice in each group. B, Th-S staining and quantitative map of the brain. C, TUNEL staining to measure the apoptosis rate in each group. D, The expression of inflammatory factors in cerebral cortex determined by ELISA. E, CCK-8 to assess cell viability in each group. F, The growth of axons observed by immunofluorescence staining. * p < 0.05 vs. the WT/DMSO group. # p < 0.05 vs. the AD/Aβ group. Each cell experiment was repeated three times
In conclusion, GPA, in an apparently dose-related manner, reduces Aβ accumulation and neuronal apoptosis in AD mouse brain, while also alleviating inflammation and axonal injury. It is highly likely that these processes underpin the improved cognitive function following GPA treatment, thus indicating that GPA may play a neuroprotective role in AD.
GAP43 is a downstream target of GPA
In order to further explore the target of GPA, the differential analysis of AD cohort GSE28146 was analyzed. In total, 770 DEGs were screened in GSE28146, among which 379 significantly upregulated and 391 significantly downregulated genes downregulated (Figure 3A). The top 50 genes with significant difference were identified and used to plot a heat map (Figure 3B). The 19 GPA targets that were predicted by CTD database and symmap database were intersected with AD-related genes that were obtained by GeneCards database and DEGs, which obtained candidate genes GAP43 and IL-1β (Figure 3C). The drug-target regulatory network was obtained by Cytoscape 3.5.1 software, as shown in Figure 3D, in which genes labeled with yellow were the common target. GAP43 was significantly downregulated in AD microarray dataset GSE28146 (Figure 3E). Overall, this bioinformatics analysis indicates that GAP43 is a probable downstream target of GPA in AD.
A, Volcano map of DEGs in AD in GSE28146. Red dots indicated upregulated genes, and green dots indicated downregulated genes. B, Heat map of expression of candidate genes. The color scale from blue to red indicates level of gene expression from low to high. C, Venn diagram of intersection of DEGs, AD-related genes and drug targets, D, The drug-target regulatory network. E, Expression box diagram of GAP43 in AD in GSE28146.
GAP43 silencing inhibits GPA-mediated improvements in cognitive function
In order to further verify whether GPA alleviated cognitive impairment and nerve injury in AD mice through GAP43, the mRNA level of GAP43 in the cerebral cortex of each group was measured by RT-qPCR. RT-qPCR results found that GAP43 expression in the AD group was substantially lower than that of the WT group. Compared with the AD group, GAP43 expression and positive rate of GAP43 in the AD + GPA group were significantly enhanced as GPA concentration (Figure 4A, B). In order to explore whether GAP43 is a target of GPA, sh-GAP43 adenovirus were injected into the lateral ventricle of AD mice to achieve a local silencing of GAP43 expression during administration period of GPA. Western blot results indicated that GAP43 expression in the AD + GPA + sh-NC group was substantially higher than that in the AD + sh-NC group. In contrast to the AD + GPA + sh-NC group, GAP43 expression was obviously lower in the AD + GPA + sh-GAP43 group, confirming a GAP43 knockdown (Figure 4C). The result of MWM tests demonstrated that the escape latency of mice in the AD + GPA + sh-NC group was significantly shorter than that of the AD + sh-NC group. Additionally, compared with the AD + GPA + sh-NC group, the escape latency of mice in the AD + GPA + sh-GAP43 group was noticeably decreased (Figure 4D). Also, on the last day of the MWM test, the hidden platform was removed for exploratory test. These results showed that, in comparison with the AD + sh-NC group, the times of crossing platform in the AD + GPA + sh-NC group was increased, while the AD + GPA + sh-GAP43 group displayed the opposite results (Figure 4E). The swimming time of each group in the target quadrant is shown in Figure 4F. Compared with the AD + sh-NC group, the swimming time of mice in the AD + GPA + sh-NC group was strikingly increased. Compared with the AD + GPA + sh-NC group, the swimming time of mice was remarkably reduced in the AD + GPA + sh-GAP43 group. The result of new object recognition test showed that the exploratory index of the AD + GPA + sh-NC group was prominently higher than that of AD + sh-NC group when one of the objects was replaced by a new one. However, in comparison with the AD + GPA + sh-NC group, the exploration index of the AD + GPA + sh-GAP43 group was observably decreased (Figure 4G-H). In summary, silencing of GAP43 reduced GPA-mediated improvements in cognitive function in AD mice.
A, mRNA expression of GAP43 in cerebral cortex of mice detected by RT-qPCR. B, GAP43 expression in cerebral cortex of mice determined by IHC staining. * p < 0.05 vs. the WT group. # p < 0.05 vs. the AD group. C, The silence efficiency of GAP43 measured by Western blot. D, Escape latency of mice in different days in the positioning and navigation test. E, The times of mice crossing the platform in each group. F, Swimming time of mice in target quadrant. G, Exploration index of mice on day 1 in new object recognition test. H, Exploration index of mice on day 2 of new object recognition test. * p < 0.05 vs. the AD + sh-NC group. # p < 0.05 vs. the AD + GPA + sh-NC group.
GAP43 silencing reduces GPA-mediated changes in Aβ deposition, neuronal apoptosis, neuroinflammation, and axonal injury
We subsequently investigated whether silencing of GAP43 could affect the function of GPA in Aβ deposition, neuronal apoptosis, neuroinflammation and axonal injury in AD mice. Based on the result of ELISAs, compared with the AD + sh-NC group, Aβ level in the AD + GPA + sh-NC group was dramatically decreased. However, the AD + GPA + sh-GAP43 group had opposite result (Figure 5A). The results of Th-S staining were identical to that of the ELISAs (Figure 5B). TUNEL staining results displayed that the apoptosis rate was markedly reduced in the AD + GPA + sh-NC group in comparison with the AD + sh-NC group, which was reverse in the AD + GPA + sh-GAP43 group relative to the AD + GPA + sh-NC group (Figure 5C). ELISA results showed that lower IL-6, TNF-α and IL-1β expression and higher IL-4 expression in the AD + GPA + sh-NC group than in the AD + sh-NC group but higher IL-6, TNF-α and IL-1β expression and lower IL-4 expression in the AD + GPA + sh-GAP43 group than in the AD + GPA + sh-NC group (Figure 5D). GAP43 was silenced in Aβ1-42-induced cells treated with GPA. The silence efficiency is shown in Figure 5E. sh-GAP43-1 (sh-GAP43 group) with better silence efficiency was selected for the next experiment. As shown in CCK-8 results, higher cell viability was noted in the GPA + Aβ + sh-NC group than in the Aβ + sh-NC group, which was reverse in the GPA + Aβ + sh-GAP43 group when compared to the GPA + Aβ + sh-NC group (Figure 5F). Immunofluorescence staining highlighted that the axon length of Aβ + GPA + sh-NC was longer than that of the Aβ + sh-NC group. Compared with the Aβ + GPA + sh-NC group, the shorter axon length was observed in the Aβ + GPA + sh-GAP43 group (Figure 5G). These results suggested that GAP43 knockdown inhibits the alleviating effects GPA treatment has on Aβ deposition, neuroinflammation, neuronal apoptosis and axonal injury in AD mice.
A, The Aβ level in the brain of each group was detected by ELISA. B, Th-s staining in brain tissue of mice in each group. C, TUNEL staining to evaluate the apoptosis rate of neurons in each group. D, The expression of inflammatory factors in cerebral cortex was detected by ELISA. * p < 0.05 vs. the AD + sh-NC group. # p < 0.05 vs. the AD + GPA + sh-NC group. E, The silencing efficiency of GAP43 (* p < 0.05 vs. the sh-NC group). F, CCK-8 assay for cell viability. G, The growth of axons observed by immunofluorescence staining. * p < 0.05 vs. the Aβ + sh-NC group. # p < 0.05 vs. the Aβ + GPA + sh-NC group. All cell experiments were repeated three times.
GPA alleviates cognitive impairment in AD mice by upregulating GAP43 through PI3K/AKT signaling
It has been reported in the literature that GPA can activate the PI3K/AKT pathway, and PI3K/AKT pathway can regulate the expression of GAP43 (15, 16). However, whether GPA can alleviate AD by activating PI3K/AKT/GAP43 regulatory axis remains unclear. Therefore, we speculated that GPA upregulates GAP43 expression through PI3K/AKT pathway to improve cognitive impairment and alleviate nerve injury in AD mice.
In order to verify this hypothesis, PI3K, AKT and p-AKT expression was examined by Western blot. This experiment revealed that, when compared with the WT group, PI3K and p-AKT expression was significantly decreased in the AD group, whereas the AD + GPA group displayed the opposite result (Figure 6A). PI3K inhibitor LY294002 was used to combine treatment with GAP43. Based on the result of Western blot, PI3K, GAP43 and p-AKT expression in the AD + GPA + LY294002 + oe-NC group was obviously reduced in contrast to the AD + GPA + oe-NC group. There was no significant difference in PI3K and p-AKT expression, but GAP43 expression was significantly increased in the AD + GPA + LY294002 + oe-GAP43 group (Figure 6B). These results indicate that activation of PI3K/AKT pathway upregulates GAP43 expression, and that GAP43 expression was markedly diminished following treatment with a PI3K inhibitor. MWM test result showed that the escape latency of mice was substantially shorter in the AD + GPA + oe-NC group than in the AD + GPA + LY294002 + oe-NC group, as well as shorter in the AD + GPA + LY294002 + oe-GAP43 group (Figure 6C). Exploratory test results described that the times of mice crossing platform were noticeably reduced in the AD + GPA + LY294002 + oe-NC group in comparison with the AD + GPA + oe-NC group, whilst the AD + GPA + LY294002 + oe-GAP43 group had reverse result relative to the AD + GPA + LY294002 + oe-NC group (Figure 6D). The swimming time in the target quadrant is exhibited in Figure 6E. Compared with the AD + GPA + oe-NC group, the swimming time of mice in the AD + GPA + LY294002 + oe-NC group was observably reduced. Compared with the AD + GPA + LY294002 + oe-NC group, the swimming time of mice in the AD + GPA + LY294002 + oe-GAP43 group was significantly enhanced. Based on the result of new object recognition test, the exploratory index of the AD + GPA + LY294002 + oe-NC group was significantly lower than that of the AD + GPA + oe-NC group when one of the objects was replaced by a new one. However, compared with the AD + GPA + LY294002 + oe-NC group, the exploration index of the AD + GPA + LY294002 + oe-GAP43 group was observably diminished (Figure 6F-G).
Collectively, these results indicate that GPA activates PI3K/AKT signaling which in turn elevates GAP43 expression, alleviating cognitive dysfunction in AD mice.
A, The PI3K, AKT and p-AKT expression in the brain of each group detected by Western blot. * p < 0.05 vs. the WT group. # p < 0.05 vs. the AD group. B, The PI3K, AKT, p-AKT and GAP43 expression in the brain of each group determined by Western blot. C, Escape latency of mice on different days in the positioning and navigation test. D, The Times of crossing the platform in each group. E, Swimming time of mice in target quadrant. F, Exploratory index of mice on day 1 of new object recognition test. G, Exploratory index of mice on day 2 of new object recognition test. * p < 0.05 vs. the AD + GPA + oe-NC group. # p < 0.05 vs. the AD + GPA + LY294002 + oe-NC group
GPA reduces Aβ levels, neuroinflammation, and neuronal damage in AD mice through activation of PI3K/AKT/GAP43 regulatory axis
The aim of this section of the study was to further verify the possible effects of GPA on via PI3K/AKT/GAP43 axis in the setting of Aβ deposition, neuroinflammation, neuronal apoptosis and axonal injury in AD mice. We adopted an ELISA approach, which showed that in comparison with the AD + GPA + oe-NC group, the Aβ level in the AD + GPA + LY294002 + oe-NC group was markedly augmented. However, compared with the AD + GPA + LY294002 + oe-NC group, the Aβ level in the AD + GPA + LY294002 + oe-GAP43 group was strikingly reduced (Figure 7A). The result of Th-S staining revealed a similar finding (Figure 7B). TUNEL staining indicated that apoptosis rate in the AD + GPA + LY294002 + oe-NC group was higher than that of the AD + GPA + oe-NC group, while the AD + GPA + LY294002 + oe-GAP43 group had opposite result relative to the AD + GPA + LY294002 + oe-NC group (Figure 7C). ELISA results documented lower IL-4 expression and higher IL-6, TNF-α, and IL-1β expression in the AD + GPA + LY294002 + oe-NC group than in the AD + GPA + oe-NC group, which was reverse in the AD + GPA + LY294002 + oe-GAP43 group relative to the AD + GPA + LY294002 + oe-NC group (Figure 7D). In a CCK-8 assay, when compared with the Aβ + GPA + LY294002 + oe-NC group, diminished cell viability was observed in the Aβ + GPA + LY294002 + oe-GAP43 group (Figure 7E). Western blot analysis demonstrated that in comparison with the DMSO group, p-AKT and GAP43 expression in the Aβ group was appreciably reduced; compared to the Aβ group, p-AKT and GAP43 expression was augmented in the Aβ + GPA + oe-NC group; p-AKT and GAP43 expression in the Aβ + GPA + LY294002 + oe-NC group was substantially lower than that of the AD + GPA + oe-NC group; in contrast to the AD + GPA + LY294002 + oe-NC group, p-AKT expression in the AD + GPA + LY294002 + oe-GAP43 group was not significantly different, but GAP43 expression was appreciably increased (Figure 7F). Immunofluorescence staining documented that relative to the Aβ + GPA + LY294002 + oe-NC group, the axon length of the Aβ + GPA + LY294002 + oe-GAP43 group was strikingly augmented (Figure 7G).
In summary, GPA upregulates GAP43 expression, leading to a decline in Aβ levels, inflammation, and neuronal damage in AD mice in a process that requires PI3K/AKT signaling pathway.
A, The Aβ level in the brain of each group detected by ELISA. B, Th-S staining in brain tissue of mice in each group. C, TUNEL staining to measure the apoptosis rate of neurons in each group. D, The inflammatory factors in cerebral cortex determined with ELISA. * p < 0.05 vs. the AD + GPA + oe-NC group. # p < 0.05 vs. the AD + GPA + LY294002 + oe-NC group. E, Cell viability assessed using CCK-8 (* p < 0.05 vs. the Aβ + GPA + LY294002 + oe-NC group). F, Western blot to detect the related protein expression (*p < 0.05 vs. the DMSO group. # p < 0.05 vs. the Aβ group, & p < 0.05 vs. the Aβ + GPA + oe-NC group, ^ p < 0.05 vs. the Aβ + GPA + LY294002 + oe-NC group). G, The growth of axons observed by immunofluorescence staining (* p < 0.05 vs. the Aβ + GPA + LY294002 + oe-NC group). All cell experiments were repeated three times
Pathologically, AD characterized by amyloid plaques which lead to dementia in increasing sections of the elderly population (17). GPA has been documented to have therapeutic effects on AD (7). However, at the time of publication, the mechanism underlying these beneficial effects remains unclear and required further investigation. Therefore, we adopted in vitro and in vivo AD model systems to examine the mechanism by which GPA acts in AD, revealing that GPA treatment activated the PI3K/AKT pathway, leading to an increase in GAP43 expression. This in turn thus improves cognitive function by reducing Aβ accumulation, neuronal apoptosis, neuroinflammation, and axonal injury, thus exerting a neuroprotective effect on AD mice.
We initially found that GPA improved cognitive impairment, and alleviated Aβ accumulation, neuronal apoptosis, neuroinflammation, and axonal injury in AD mice in a dose dependent manner. Concurring with our results, research conducted by Zhou et al. uncovered that memory deficits, cognitive impairment, Aβ deposition, and neuroinflammation were repressed after GPA treatment in AD mice (7). In addition, GPA treatment reduces cellular apoptosis in mice with D-galactosamine and lipopolysaccharide-induced hepatic failure (18). Moreover, neuronal geniposide activity, is widely recognized as a derivative of GPA (6), has been reported to be associated with increased gene expression related to cell growth and repair, and inhibited apoptosis and inflammation of neurons (19). Geniposide has also been demonstrated to repress amyloid deposition and behavioral and cognitive impairments of AD mice (20). Therefore, it is highly likely that GPA has neuroprotective effects on AD.
Further mechanistic analysis revealed that GAP43 is a target of GPA in AD. A prior study identified that after geniposide treatment, Neuro2a neuroblastoma cells had elevation of GAP43 expression, supporting our finding (8). Importantly, our data clarified that silencing of GAP43 negated the alleviating role of GAP in cognitive impairment, Aβ accumulation, neuronal apoptosis, neuroinflammation, and axonal injury in AD mice. As a presynaptic protein, GAP43 is overexpressed during neuronal development and synaptogenesis where it plays a crucial role in the orchestration of learning and memory functions, axonal outgrowth, and synaptic plasticity (21). Interestingly, the downregulation of GAP43 has been detected in the brain tissues of AD mice (22). Additionally, GAP43 is also involved in the decline of neuronal apoptosis in rats with spinal cord injury (23). A previous work uncovered the finding that GAP43 upregulation is associated with relief and suppression of inflammation in mouse diabetic retinopathy (24). Activation of GAP43 has also been shown to lead to attenuation of cognitive impairment in rats with subthreshold convulsant discharge (25). Research conducted by Liu et al. demonstrated that GAP43 overexpression causes the promotion of axonal regeneration in the spinal cord of rats with spinal cord injury (26). Based on our results and the existing literature, GAP43 overexpression might participate in the neuroprotective effects of GPA on AD.
Another central finding of this project was that GPA activates the PI3K/AKT pathway to upregulate GAP43, thus alleviating cognitive impairment by reducing Aβ accumulation, neuronal apoptosis, neuroinflammation, and axonal injury in AD mice. Concordant with our finding, GPA has been shown to activate the PI3K/AKT pathway in human melanocytes (15). It has also been elucidated in another work that activation of the PI3K/AKT pathway was capable of elevating GAP43 expression during hair cell protection in the neonatal murine cochlea (12). Intriguingly, activation of the PI3K/AKT pathway could contribute to the repression of cognitive impairment in AD rat (13). In line with our results, activation of the PI3K/AKT pathway ameliorates learning and memory dysfunction, the histology structure of damaged neurons in hippocampal area, and neuronal apoptosis in AD mice (27).
Our data demonstrates that GPA relieves cognitive impairment in a process that leads to a reduction in Aβ accumulation, neuronal apoptosis, neuroinflammation, and axonal injury in AD mice. Mechanistically, we have revealed that GPA exerts a neuroprotective effect on AD in mice via activation of the PI3K/AKT/GAP43 regulatory axis (summarized in Figure 8). This study uncovers a mechanism by which GPA suppresses AD progression, imparting an improved understanding of the pathogenesis of AD and providing novel potential therapeutic targets for AD treatment. However, considering the limitations of our study, more extensive research is required to investigate the specific mechanism of GPA in AD, with eventual translation to clinical trials. Absorption of GPA is affected by different processing methods in rats following crude Gardeniae Fructus administration based on a comparative pharmacokinetics of GPA, Geniposide, Genipin, and Crocetin (28). Moreover, the toxicity of GPA has not been discovered yet (29). Future studies should be conducted to foresee the future investigations of the therapeutic interest of this compound.
Funding: This work is supported by Clinical Research Special Fund of Qiqihar Academy of Medical Sciences (QMSI2020L-11).
Acknowledgements: We would like to give our sincere appreciation to the reviewers for their helpful comments on this article.
Declaration of conflicting interests: The authors declare no conflict of interest.
Ethical Statement: The experiment was approved by the animal ethics committee of Qiqihar Medical University.
1. Scheltens P, Blennow K, Breteler MM, et al. Alzheimer’s disease. Lancet 2016;388:505-517, doi: 10.1016/S0140-6736(15)01124-1.
2. Lane CA, Hardy J, Schott JM. Alzheimer’s disease. Eur J Neurol 2018;25:59-70, doi: 10.1111/ene.13439.
3. Robinson N, Grabowski P, Rehman I. Alzheimer’s disease pathogenesis: Is there a role for folate? Mech Ageing Dev 2018;174:86-94, doi: 10.1016/j.mad.2017.10.001.
4. Tiwari S, Atluri V, Kaushik A, Yndart A, Nair M. Alzheimer’s disease: pathogenesis, diagnostics, and therapeutics. Int J Nanomedicine 2019;14:5541-5554, doi: 10.2147/IJN.S200490.
5. Joe E, Ringman JM. Cognitive symptoms of Alzheimer’s disease: clinical management and prevention. BMJ 2019;367:l6217, doi: 10.1136/bmj.l6217.
6. Gao Y, Chen ZY, Liang X, Xie C, Chen YF. Anti-atherosclerotic effect of geniposidic acid in a rabbit model and related cellular mechanisms. Pharm Biol 2015;53:280-285, doi: 10.3109/13880209.2014.916310.
7. Zhou Z, Hou J, Mo Y, et al. Geniposidic acid ameliorates spatial learning and memory deficits and alleviates neuroinflammation via inhibiting HMGB-1 and downregulating TLR4/2 signaling pathway in APP/PS1 mice. Eur J Pharmacol 2020;869:172857, doi: 10.1016/j.ejphar.2019.172857.
8. Chen MK, Peng CC, Maner RS, Zulkefli ND, Huang SM, Hsieh CL. Geniposide ameliorated fluoxetine-suppressed neurite outgrowth in Neuro2a neuroblastoma cells. Life Sci 2019;226:1-11, doi: 10.1016/j.lfs.2019.04.003.
9. Zhang F, Ying L, Jin J, et al. GAP43, a novel metastasis promoter in non-small cell lung cancer. J Transl Med 2018;16:310, doi: 10.1186/s12967-018-1682-5.
10. Long ZM, Zhao L, Jiang R, et al. Valproic Acid Modifies Synaptic Structure and Accelerates Neurite Outgrowth Via the Glycogen Synthase Kinase-3beta Signaling Pathway in an Alzheimer’s Disease Model. CNS Neurosci Ther 2015;21:887-897, doi: 10.1111/cns.12445.
11. Liu F, Wang Y, Yao W, Xue Y, Zhou J, Liu Z. Geniposide attenuates neonatal mouse brain injury after hypoxic-ischemia involving the activation of PI3K/Akt signaling pathway. J Chem Neuroanat 2019;102:101687, doi: 10.1016/j.jchemneu.2019.101687.
12. Hayashi Y, Yamamoto N, Nakagawa T, Ito J. Insulin-like growth factor 1 induces the transcription of Gap43 and Ntn1 during hair cell protection in the neonatal murine cochlea. Neurosci Lett 2014;560:7-11, doi: 10.1016/j.neulet.2013.11.062.
13. Kamel AS, Abdelkader NF, Abd El-Rahman SS, Emara M, Zaki HF, Khattab MM. Stimulation of ACE2/ANG(1-7)/Mas Axis by Diminazene Ameliorates Alzheimer’s Disease in the D-Galactose-Ovariectomized Rat Model: Role of PI3K/Akt Pathway. Mol Neurobiol 2018;55:8188-8202, doi: 10.1007/s12035-018-0966-3.
14. Biskind MS, Schreier H. On the significance of nutritional deficiency in diabetes. Exp Med Surg 1945;3:299-316.
15. Lu W, Zhao Y, Kong Y, et al. Geniposide prevents H2 O2 -induced oxidative damage in melanocytes by activating the PI3K-Akt signalling pathway. Clin Exp Dermatol 2018;43:667-674, doi: 10.1111/ced.13409.
16. Zhang L, Yue Y, Ouyang M, Liu H, Li Z. The Effects of IGF-1 on TNF-alpha-Treated DRG Neurons by Modulating ATF3 and GAP-43 Expression via PI3K/Akt/S6K Signaling Pathway. Neurochem Res 2017;42:1403-1421, doi: 10.1007/s11064-017-2192-1.
17. Gao LB, Yu XF, Chen Q, Zhou D. Alzheimer’s Disease therapeutics: current and future therapies. Minerva Med 2016;107:108-113.
18. Kim SJ, Kim KM, Park J, Kwak JH, Kim YS, Lee SM. Geniposidic acid protects against D-galactosamine and lipopolysaccharide-induced hepatic failure in mice. J Ethnopharmacol 2013;146:271-277, doi: 10.1016/j.jep.2012.12.042.
19. Liu W, Li G, Holscher C, Li L. Neuroprotective effects of geniposide on Alzheimer’s disease pathology. Rev Neurosci 2015;26:371-383, doi: 10.1515/revneuro-2015-0005.
20. Zhang Z, Wang X, Zhang D, Liu Y, Li L. Geniposide-mediated protection against amyloid deposition and behavioral impairment correlates with downregulation of mTOR signaling and enhanced autophagy in a mouse model of Alzheimer’s disease. Aging (Albany NY) 2019;11:536-548, doi: 10.18632/aging.101759.
21. Sandelius A, Portelius E, Kallen A, et al. Elevated CSF GAP-43 is Alzheimer’s disease specific and associated with tau and amyloid pathology. Alzheimers Dement 2019;15:55-64, doi: 10.1016/j.jalz.2018.08.006
22. Reutzel M, Grewal R, Silaidos C, et al. Effects of Long-Term Treatment with a Blend of Highly Purified Olive Secoiridoids on Cognition and Brain ATP Levels in Aged NMRI Mice. Oxid Med Cell Longev 2018;2018:4070935, doi: 10.1155/2018/4070935.
23. Ni Y, Gu J, Wu J, Xu L, Rui Y. MGMT-Mediated neuron Apoptosis in Injured Rat Spinal Cord. Tissue Cell 2020;62:101311, doi: 10.1016/j.tice.2019.101311.
24. Mohammad HMF, Sami MM, Makary S, Toraih EA, Mohamed AO, El-Ghaiesh SH. Neuroprotective effect of levetiracetam in mouse diabetic retinopathy: Effect on glucose transporter-1 and GAP43 expression. Life Sci 2019;232:116588, doi: 10.1016/j.lfs.2019.116588.
25. Wang MJ, Jiang L, Chen HS, Cheng L. Levetiracetam Protects Against Cognitive Impairment of Subthreshold Convulsant Discharge Model Rats by Activating Protein Kinase C (PKC)-Growth-Associated Protein 43 (GAP-43)-Calmodulin-Dependent Protein Kinase (CaMK) Signal Transduction Pathway. Med Sci Monit 2019;25:4627-4638, doi: 10.12659/MSM.913542.
26. Liu H, Xiong D, Pang R, et al. Effects of repetitive magnetic stimulation on motor function and GAP43 and 5-HT expression in rats with spinal cord injury. J Int Med Res 2020;48:300060520970765, doi: 10.1177/0300060520970765.
27. Yao Y, Wang Y, Kong L, Chen Y, Yang J. Osthole decreases tau protein phosphorylation via PI3K/AKT/GSK-3beta signaling pathway in Alzheimer’s disease. Life Sci 2019;217:16-24, doi: 10.1016/j.lfs.2018.11.038.
28. Yang X, Li J, Yang X, He J, Chang YX. Comparative Pharmacokinetics of Geniposidic Acid, Genipin-1-beta-Gentiobioside, Geniposide, Genipin, and Crocetin in Rats after Oral Administration of Crude Gardeniae Fructus and Its Three Processed Products Using LC-MS/MS. Evid Based Complement Alternat Med 2020;2020:1642761, doi: 10.1155/2020/1642761.
29. Khanal T, Kim HG, Choi JH, et al. Biotransformation of geniposide by human intestinal microflora on cytotoxicity against HepG2 cells. Toxicol Lett 2012;209:246-254, doi: 10.1016/j.toxlet.2011.12.017.