Minor ginsenoside F1 improves memory in APP/PS1 mice
Ginseng has been shown to produce a cognitive improvement effect. The key molecular components in ginseng that produce pharmacological effects are ginsenosides. Previous studies reported a memory improvement effect of a few major ginsenosides. However, the identity of specific minor ginsenosides mediating such function remains unknown. Here, we report that a minor ginsenoside F1 improves memory function in APPswe/PSEN1dE9 (APP/PS1) double-transgenic Alzheimer’s disease (AD) model mice. After 8-wk oral administration of F1 jelly, we observed that spatial working memory, but not context-dependent fear memory, was restored in AD mice. To search for a possible underlying molecular and cellular mechanism, we investigated the effect of F1 on Aβ plaque. We observed F1 administration reduced the Aβ plaque area and density in the cortex, but not in the hippocampus of AD mice. Next, we tested for the effect of F1 on the expression level of key molecules involved in learning and memory. Results from Western blot assay revealed that an abnormally reduced level of a phosphorylated form of CREB in the hippocampus of AD mice was restored to a normal level by F1 administration. Moreover, in the same animals, BDNF level was augmented in the cortex. Our results, therefore, suggest that minor ginsenoside F1 constitutes a promising target to develop therapeutic agents for AD.
KeywordsAlzheimer’s disease APP/PS1 mice Ginsenoside F1 Amyloid-beta plaque pCREB BDNF
Brain-derived neurotrophic factor
Contextual fear conditioning
cAMP response element-binding protein
Chimeric mouse/ human amyloid precursor protein
Phosphorylated form of CREB
Mutant human presenilin 1
Alzheimer’s disease (AD) is a neurodegenerative disease characterized by a loss of neurons and severe memory impairment. Because AD has become a major social problem globally in an accelerating aging society , there has been a dramatic increase in the need for effective treatments that restore or improve memory function. Rescue of the functional deficit is the most important clinical endpoint for patients with approaching the onset of overt dementia . Therefore, identifying a natural compound that is effective to improve memory function in AD is urgently needed.
Ginseng has been utilized as a natural medicinal herb for thousands of years in the East. Recent researches have reported a wide range of therapeutic effects of ginseng, including tumor suppression [3, 4], anti-aging , anti-oxidation , and cognitive improvement [7, 8, 9, 10]. The key molecular components in ginseng that produce such pharmacological effects are ginsenosides. Ginsenosides are natural steroid glycosides, which are abundant in the root of ginseng . Ginsenosides are classified into major and minor ginsenosides, which are produced by the deglycosylation of major ginsenosides [12, 13, 14].
After oral administration, major ginsenosides are converted into minor ginsenoside forms by hydrolyzation of the 6- and 20-glucoside bond by intestinal microflora and then absorbed into the body . Non-metabolized major ginsenosides have a low absorption rate in the body and are rapidly eliminated from it [15, 16]. The metabolic rate of intestinal microflora is very low. Moreover, since the composition of intestinal bacteria varies from individual to individual, the pharmacological effects of taking major ginsenosides vary widely from person to person . In contrast, minor ginsenosides are absorbed in the intestine and exert actual pharmacological effects . Therefore, it is critical to identify a single minor ginsenoside that produces a therapeutic effect. However, due to the technical difficulty to obtain a sufficient amount of minor ginsenosides from ginseng for research purposes, most of the extant literature has focused on major ginsenosides for their pharmacological effects. Therefore, a minor ginsenoside with cognitive improvement function in an AD model remains undetermined.
Because major ginsenoside Rg1, the precursor of F1, has been reported to reduce amyloid-beta (Aβ) plaque, modulate neurite outgrowth, and improve cognitive function [18, 19, 20], we hypothesized that F1 constitutes a promising candidate. Recently, our research group developed a novel system that enabled mass production of minor ginsenoside F1 from Rg1 . Facilitated by this technical innovation, in the present study, we investigated whether F1 has a therapeutic effect on AD by using AD model mice.
Ginsenoside F1 rescues memory impairment in 14-month-old APP/PS1 mice
To test the cognitive improvement effect of F1 in AD, we utilized APPswe/PSEN1dE9 (APP/PS1) double-transgenic AD model mice. The APP/PS1 transgenic mouse expresses chimeric mouse/human amyloid precursor protein (Mo/HuAPP695swe) and a mutant human presenilin 1 (PS1-dE9), which accumulate amyloid beta burden in CNS from 6 to 7 months of age [22, 23, 24, 25]. A deficit of learning and memory [26, 27, 28] has been reported for these AD model mice.
To examine the specificity of the memory improvement effect of F1, we next performed contextual fear conditioning (CFC), which comprised a hippocampus-dependent associative emotional memory . It was reported in AD patients that associated fear-conditioned memory is impaired . The same groups of mice used for the Y-maze test were trained for CFC, and 24-h later tested for long-term memory recall. Consistent with previous reports [34, 35], CFC memory was impaired in APP/PS1 mice. However, the F1 administration did not improve such memory impairment compared to control mice (Fig. 1c). Taken together, our behavioral results suggest that F1 improves working memory function rather than hippocampal-dependent long term memory in AD model mice.
Ginsenoside F1 reduces Aβ plaques in the cortex of AD mice
Ginsenoside F1 rescues the expression level of pCREB in the hippocampus and increases the expression level of BDNF in the cortex of APP/PS1 mice
An abnormal decrease of the expression level of the phosphorylated form of CREB (pCREB) [40, 41, 42, 43] and BDNF [44, 45, 46] has been implicated in a deficit of memory function in AD patients and model mice. Besides, the previous study reported that recovery of cognitive deficit in AD model mice is accompanied by a reduction of Aβ and increase of pCREB and BDNF level . Thus, we investigated whether F1 exerts any effect on the expression level of pCREB and BDNF. We performed Western blot analysis to determine the expression level of pCREB and BDNF in the hippocampus and cortex of 8-month-old APP/PS1 mice following 8 wk. of F1 administration, as done previously.
In this study, we report for the first time that the administration of minor ginsenoside F1 rescues memory impairment in APP/PS1 double transgenic mice which are known as Alzheimer’s disease model mice. Results from the Y-maze test showed spatial working memory was recovered by F1 in AD mice. To find out the possible underlying mechanism, we examined an effect of F1 on Aβ plaque in the retrosplenial cortex of AD mice. We observed a significant reduction of the ThS positive plaques, but not the 6E10 positive plaques. These results suggest that F1 inhibits the formation of dense Aβ plaques or elicits their disaggregation without changing the total quantity of Aβ plaques in the cortex. Given that the working memory is dependent on the function of the cortex, the reduction of dense Aβ plaques in the cortex may explain the rescue of spatial working memory by F1 in AD mice. In addition to the Aβ plaques, western blot results in this study show F1 increases the level of BDNF above normal levels in the cortex. Previous studies show that the BDNF level is directly correlated with AD severity  and the increase of BDNF level is effective to improve cognitive function in AD [48, 49]. Therefore, the increase of BDNF expression in the cortex is also a possible mechanism explaining the improvement of working memory by F1. In the hippocampus, although the number of aggregated forms of Aβ detected with ThS was slightly reduced (P = 0.0596), we did not observe any significant effect of F1 on Aβ plaques. Considering the different effect of F1 on Aβ plaques in the cortex and hippocampus, a plausible explanation for the difference of F1 effect on Y-maze versus contextual fear conditioning might be the specific reduction of Aβ plaques only in the cortex. Western blot results showed F1 rescues the pCREB level in the hippocampus to the normal level, but it did not affect the expression level of BDNF. Because contextual fear memory was not improved by F1, it is likely that the restoration of the pCREB level in the hippocampus may not be sufficient to rescue the deficit of hippocampus-dependent memory in AD model mice.
From the western blot result, we failed to see the significant changes of pCREB in the cortex and BDNF in the hippocampus. In our study, western blot analyses were done using 8-months-old mice while behaviors were tested in much older mice. Given the age-dependent progress of AD pathology , the relatively young age condition may explain why we failed to see the significant changes of p-CREB in the cortex and BDNF in the hippocampus. Given the mutually positive effects on the expression of BDNF and CREB , it is expected that increase of BDNF may in turn cause the same an increase of pCREB level in the cortex after F1 treatment. Similarly, recovery of pCREB level may also lead to an increase of BDNF in the hippocampus. However, these are not what we found. One possible explanation for such discrepancy is that the mutually positive regulation mechanism does not work properly in the AD brain and F1 produces its effect on pCREB and BDNF through other molecular pathways [52, 53, 54, 55].
Previous studies reported Rg1, the precursor of F1, rescues cognitive function in AD mice regardless of behavior tasks including Morris water maze, radial arm maze, Y-maze, contextual fear conditioning [7, 19, 56, 57]. Moreover, the Rg1 administration showed a reduction of Aβ plaques in the brain and recovery of memory-related genes like pCREB or BDNF in the hippocampus [19, 56, 57]. In this study, we observed the specific effect of F1 in the Y-maze task which is mainly dependent on the cortex. Consistently, F1 reduced Aβ plaques and increased BDNF expression levels in the cortex but not in the hippocampus. These results suggest that the therapeutic effect of F1 on AD is specific to the cortex, compared to the global effects of Rg1. F1 may have an advantage over Rg1 in terms of the delivery method. In most research, Rg1 was administered via intraperitoneal injection because of the low absorption rate in the intestine of major ginsenosides, a limitation of using Rg1 as a drug or health supplement [15, 16]. However, we show here that F1 can be delivered by oral administration to restore memory impairment with reducing Aβ plaque and increasing pCREB and BDNF expression in AD mice, and thus is a more promising candidate to treat AD.
Considering that synthetic compounds often cause unwanted side effects in many cases, the identification of a natural compound with cognitive improvement function is invaluable to develop medicinal drugs or health supplement foods not only for aged people but also for AD patients. Our results provide evidence that minor ginsenoside F1 improves memory function in AD model mice. Therefore, F1 is a promising target to develop therapeutic agents for memory improvement.
Materials and methods
APPswe/PSEN1dE9 double-transgenic AD mice with a B6 × C3 background and B6 × C3 wild type mice were purchased from the Jackson Laboratory (MMRRC Stock No. 034829-JAX). Heterozygous males were bred with wild type females. Offspring were genotyped by using a standard PCR protocol detecting PSEN1 transgene. Mice that did not express the transgene were used as wild type controls. Mice were housed on a 12-h light/dark cycle at a constant temperature (21–23 °C) and humidity (40–60%). Food and water were available ad libitum. All procedures and protocols were approved by the Animal Ethics Committee at the Korea Advanced Institute of Science and Technology. All experiments were performed in accordance with the guideline of Institutional Animal Care and Use Committee.
Preparation of ginsenoside F1
Ginsenoside F1 (> 95% pure) was prepared using an enzymatic method from Panax ginseng extract as previously reported  and isolated using Recycling Preparative HPLC (Japan Analytical Industry Co., Ltd.) with JAIGEL-ODS-AP column (10 mm, 500 × 20 mm id, Japan Analytical Industry Co., Ltd.).
To test the effect of F1 on AD, F1 was orally administrated via gelatin-based jelly at a dose level of 20 mg/kg/day. Gelatin-based jelly was prepared as previously described [58, 59]. For a 1-d dose of jelly, 0.6 mg of ginsenoside F1 was dissolved in 0.45 ml of 20% Splenda solution. F1 solution was further mixed with 1.35 ml of 14% gelatin, 20% Splenda solution, and 0.15 ml chocolate-flavoring in a 24-well plate. A piece of jelly (~ 1.9 mg) was provided, and complete intake of jelly was confirmed daily.
Mice were 12-months-old when F1 treatment began, and all the mice were male for behavioral tests and immunohistochemistry of amyloid beta plaque. AD mice and wild type mice were separated into three groups: F1-treated AD; vehicle-treated AD; and non-treated wild type mice. After 8-wk administration of F1, behavioral tests and immunohistochemistry test were performed. F1 was administered to six-month-old male and female mice for 8 wk. for Western blot test.
Contextual fear conditioning
For contextual fear conditioning (CFC), mice were handled for 5 min on 3 d prior to conditioning. On conditioning day, mice were placed in a fear conditioning chamber (Coulbourn Instruments) with a metal grid floor. Mice were allowed to explore the context for 150 s, and 2 s of 0.5 mA electrical foot shock was delivered twice (120 s inter-stimulus-interval). Mice were left in the conditioning chamber for an additional 30 s and placed back in their home cage. For the contextual fear memory test, mice were placed back into the same context 24 h after conditioning. Behavior of mice was recorded for 5 min, and mice were returned to their home cage. Freezing was automatically scored using FreezeFrame3.0 software (Coulbourn Instruments).
Brain sample preparation
Mice were anesthetized with 2.5% avertin by intraperitoneal injection. Mice were perfused with phosphate-buffered saline (PBS) and then fixed with cold 4% paraformaldehyde (PFA). After perfusion, brain samples were stored in 4% PFA overnight for post-fixation. Fixed brain samples were immersed in 30% sucrose in filtered PBS until they sank to the bottom of the vial at 4 °C for dehydration. Dehydrated brains were fixed on a disk with OCT compound at − 20 °C. 40-μm thickness sections of hippocampal tissue were collected using Cryostat (Leica CM1850, Leica Biosystems).
Immunohistochemistry and thioflavin S staining
To visualize amyloid beta plaque in brain sections, amyloid beta was stained by 6E10 and thioflavin S. After three times of PBS washing, brain sections were blocked with blocking solution (0.1% BSA, 0.2% Triton X-100, 2% goat serum in PBS). Brain sections were then incubated with rabbit anti-6E10 monoclonal antibody (BioLegend, 803,015, 1:2000) overnight at room temperature. Next, Alexa fluor-594 conjugated goat anti-rabbit antibody (Molecular Probes, A-11037, 1:1000) was used as a secondary antibody. For thioflavin S staining, brain sections were incubated for 10 min in 0.0008% ThS dissolved in 50% ethanol. Sections were washed with 50% ethanol and PBS twice each. The sections were mounted with VECTASHIELD Antifade Mounting Media with DAPI (H-1200-10, Vector Laboratories) on glass slides. Images were taken on a slide scanner (ZEISS Axio Scan.Z1, Carl ZEISS). To analyze the number of amyloid beta plaque and plaque area, the hippocampus region and the retrosplenial cortex region of coronal brain sections (bregma − 1.6 to − 2.4 mm) was analyzed by using the Image-J program (NIH). Plaques less than 10 μm in diameter were not scored.
To test the effect of F1 on the expression level of pCREB and BDNF, 8-month-old AD mice and age-matched wild type mice were used: F1-treated AD (n = 3); vehicle-treated AD (n = 3); and age-matched non-treated wild type mice (n = 5). Mice were anesthetized with isoflurane, and brains were extracted. Hippocampus tissue and whole cortex tissue were collected by dissecting brain (bregma − 1 to − 3 mm), and lysed in 100 μl of ice-cold lysis buffer (50 mM HEPES pH 8.0, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 12 mM MgCl2, 20 mM EGTA, 10 mM NaPPi, 100 mM NaF, 10 mM Na-Orthovanadate, 1 mM DTT) containing a protease inhibitor cocktail (Roche, 11,836,153,001). Total protein concentrations were measured by Bradford assay. Because CREB and pCREB have almost the same protein size, they had to be blotted in separate gels. Two identical protein samples from the same tissue lysate were prepared and processed in parallel. Proteins (30 μg per lane) were resolved by SDS-PAGE, and transferred to PVDF membranes by using the Trans-Blot Turbo Blotting System (Bio-Rad). Membranes were blocked by 5% NFDM (nonfat dried milk) in TNTX buffer (50 mM Tris-HCl, pH 7.5, 200 mM NaCl, 0.2% Triton X-100) for 30 min at room temperature. We used 3% BSA instead of 5% NFDM for blocking membrane that was used to blot pCREB. After blocking, membranes were incubated with primary antibodies (anti-CREB antibody, Cell Signaling, 9197S, 1:1000; anti-phospho-CREB (Ser133) antibody, Millipore, 06–519, 1:2000; anti-BDNF antibody, Abcam, ab226843, 1:2000) in 3% BSA overnight at 4 °C. HRP-conjugated goat anti-rabbit IgG (Millipore, 12–348, 1:2000) was used as a secondary antibody. Signals were developed with ECL solution (GE Healthcare, RPN2232) and detected using ChemiDoc MP imaging system (Bio-Rad). The results were analyzed using ImageLab software (Bio-Rad). BDNF and GAPDH were stained on the same membrane after stripping the CREB or pCREB antibodies using Restore Western Blot Stripping Buffer (21,059, Thermo Fisher Scientific). As a loading control, anti-GAPDH antibody (Invitrogen, MA5–15738, 1:2000) was used. All Western blot data were normalized to GAPDH expression level for comparison.
GraphPad Prism 6 was utilized to obtain graphs and perform statistical analysis. To confirm normality of data, we conducted D’Agostino-Pearson or Shapiro-Wilk test. A two-tailed, unpaired student’s t-test was performed to analyze plaque density and area. One-way ANOVA followed by Tukey’s post-hoc was conducted to analyze the behavioral task and Western blot assay data. Error bars represent the s.e.m.
J-HH and S-CK designed and directed the study. JH and J-PO prepared animals and administrated F1 jelly. J-PO performed behavioral tasks. JH conducted immunohistochemistry and plaque counting. JH and MY prepared lysate samples and performed Western blot assay. C-HC and B-MJ prepared ginsenoside F1. JH, JPO, MY, C-HC, and J-HH drafted the manuscript. All authors read and approved the final manuscript.
This work was supported by an Intelligent Synthetic Biology Center of the Global Frontier Project funded by the MSI (2011–0031955). J.-P.O. was supported by the project of the Global Ph.D. Fellowship from the National Research Foundation of Korea (NRF-2014H1A2A1020081).
All procedures and protocols were approved by the Animal Ethics Committee at the Korea Advanced Institute of Science and Technology. All experiments were performed in accordance with the guideline of Institutional Animal Care and Use Committee.
Consent for publication
The authors declare that they have no competing interests.
- 2.US Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Center for Biologics Evaluation and Research (CBER). Early Alzheimer’s disease: developing drugs for treatment–guidance for industry 2018. https://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM596728.pdf. Accessed Feb 2018.
- 20.Li N, Zhou L, Li W, Liu Y, Wang J, He P. Protective effects of ginsenosides Rg1 and Rb1 on an Alzheimer's disease mouse model: a metabolomics study. J Chromatogr. 2015;985:54–61.Google Scholar
- 39.Paxinos G, Franklin BJ. The Mouse Brain in Stereotaxic Coordinates. 2nd ed. Academic Press;2001.Google Scholar
- 59.Zhang L. Voluntary oral administration of drugs in mice. Protocol Exchange. 2011;10.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.