Introduction

Alzheimer’s disease (AD) is an age-related neurodegenerative disease, with symptoms of progressive cognitive decline and memory impairment. In 2019, fifty million people worldwide have been affected by dementia according to the World Alzheimer Report, and the number is predicted to reach 152 million by 2050, of which AD accounts for 50–75% [1]. Currently, there is no effective cure for AD. There remain urgent needs for therapeutic methods to slow the progress of the diseases.

AD is pathologically characterized by intercellular β-amyloid (Aβ) plaques and intracellular neurofibrillary tangles (NFTs) [2]. NFTs are composed of straight and paired-helical filaments (PHFs), both of which are composed predominantly of insoluble polymers of abnormally hyperphosphorylated microtubule-associated protein Tau [3]. In AD brains, Tau is aberrantly hyperphosphorylated, carrying a threefold to fourfold of phosphates [4]. The abnormal Tau phosphorylation seems to be related to the alteration of several kinases, including glycogen synthase kinase 3 beta (GSK3β), cyclin-dependent kinase 5 (CDK5), cAMP-dependent protein kinase (PKA), and microtubule-affinity-regulating kinase (MARK) [5]. GSK3β and CDK5 are the kinases primarily responsible for phosphorylation of Tau. GSK3β-phosphorylated sites of Tau include the Thr181, Thr231, Ser396, Ser404, and Ser202 [6]. GSK3β (Ser9) is a well-known Akt phosphorylation site, and the Ser9 phosphorylation reduces the kinase activation of GSK3β, whereas phosphorylation of GSK3β on Tyr216 leads to activation [7, 8]. The binding of p35 to Cdk5, and phosphorylation on Tyr15, can activate the Cdk5 kinase which in turn phosphorylates Tau [9, 10]. Therefore, GSK3β and CDK5 inactivation prevent Tau phosphorylation and are crucial to the development of AD. Besides Aβ plaques and NFTs in AD, neuroinflammation has been increasingly implicated as a major contributing factor to AD [11, 12], meanwhile inducing more Tau phosphorylation [13]. The use of existing drugs to target neuroinflammation has been proved effective in AD mouse models and patients; therefore, anti-inflammatory may represent a potential therapeutic strategy for AD [14,15,16,17,18].

In recent years, more and more traditional Chinese medicine (TCM) have been reported to ameliorate the symptoms of AD, including Kai-Xin-San (KXS), which was first presented by Sun Simiao in the book of Bei Ji Qian Jin Yao Fang during the Tang dynasty, consisting of Polygala tenuifolia Willd (PR), Panax ginseng C.A. Mey (GR), Poria cocos (Schw.) Wolf (PO), and Acorus tatarinowii Schott (AT) [19]. As a classic formula, KXS has been applied in amnesia treatment for thousands of years [20] and displays profound therapeutic effect for dementia and depression in clinical trials[21]. Some researchers have found that KXS exerts cardioprotective effects in a myocardial infarction [22], and could ameliorate chronic fatigue syndrome by affecting the levels of inflammatory factors [23]. KXS has been reported to ameliorate cognitive dysfunction in AD animal model by increasing cholinergic and glutamatergic neurotransmission and promoting Aβ42 degradation [24,25,26,27,28]. However, as TCM formulas contain multiple chemical components, the specific pharmacological mechanisms through which KXS exert their effects against AD are still difficult to illustrate.

The network pharmacology is well suited for analyzing the multi-targeted agents, so network pharmacology methods may be appropriate for identifying the complex mechanisms of KXS. In the present study, we screened 30 targets of KXS associated with AD and predicted the mechanisms of KXS in the treatment of AD by network pharmacology. Then, we validate the predictions of bioinformatics in senescence-accelerated mouse prone 8 (SAMP8) mice, which was originally generated from AKR/J mice in the 1970s in the laboratory of Professor Takeda at Kyoto University in Japan [29]. SAMP8 mice is particularly well-suited to study the “transitional switch” between aging and AD as it exhibits spontaneous cognitive decline and Tau protein activation [30] found in aged individuals [31,32,33] and, to a greater extent, in patients with AD[34]. And most neurodegenerative alterations of SAMP8 mice emerged from 7 months old [35]. In this study, we chose 7-month-old SAMP8 mice as AD mouse model to be examined whether KXS could be inhibit neuronal apoptosis and improve learning and memory functions after 3-month (5.4 g/kg once per day, oral gavage) KXS treatment. Subsequently, the potential underlying action mechanisms of KXS on AD predicted by the network pharmacology analyses were experimentally validated in SAMP8 mice. We expect that our results can elucidate the regulatory mechanism of KXS therapeutic effects in AD and highlight the potential of KXS as a drug for ameliorating AD-like pathology and cognitive impairment.

Materials and Methods

Network Pharmacology Analysis

Chemical compound information was obtained from the Encyclopedia of Traditional Chinese Medicine (ETCM) database for all the 4 medicinal herbs (ETCM, http://www.nrc.ac.cn:9090/ETCM/) [26]. Target genes of TCM ingredients, herbs, and formulas were collected according to the chemical fingerprint similarity between TCM ingredients and known drugs. High correlation target of a drug-compound-target is defined by a QED score greater than 0.8 [36]. The DisGeNET database, AlzPlatform database, and GeneCards database were screened to collect the known AD-related targets. Cytoscape (version 3.6.1) was used to construct the active components and AD target (C-T) network of KXS. We constituted protein-protein interaction (PPI) networks by using the STRING network database platform to further elucidate the interactions of the common target genes in the treatment of AD with KXS. Gene ontology (GO) enrichment analysis was performed for identifying BP (biological process) of the potential targets of KXS for AD. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the predicted targets can provide a more systematic and comprehensive understanding of the action mechanisms of the KXS on AD. In GO and KEGG analysis, p < 0.05 was defined as significantly enriched. Finally, significant pathways and its corresponding targets were imported into Cytoscape v3.6.1 to build a network diagram of the target-pathway(T-P).

Preparation of KXS

All herbs were purchased from Anhui Yiyuantang Sinopharm Co., Ltd. (Anhui, China) and authenticated by Prof. Dong-Yu Min (Affiliated Hospital of Liaoning University of Traditional Chinese Medicine, Shenyang, China) as shown in Supplementary Table 1. The ratio of PR, GR, PO, and AT was 2:3:3:2. The four herbal materials of KXS were soaked in cold water for 2 h before being boiled for 40 min. The first decoction was obtained. Then, the four herbal materials were boiled a second time for 40 min to obtain the second decoction. Finally, the first and second decoctions were mixed, filtered through gauze, and concentrated followed brought to volume to obtain a final extract (0.54 g/mL). For chemical identification of KXS, HPLC chromatographic analysis was conducted. The HPLC analysis procedure and characteristic chromatogram of KXS are shown in Supplementary Fig. 1.

Animals and Treatments

A total of 12 male SAMP8 mice (weight, 28–32 g; age, 7-months old) were used in the present study. The mice were housed in a specific pathogen-free (SPF) environment with a 12-h light/12-h dark cycle, allowed free access to water and standard food. All animals were treated according to the guidelines for the Care and Use of Laboratory Animals and the experimental procedures were approved by the Ethics Committee for Animal Use of China Medical University. The animals were randomly divided into vehicle group and KXS treatment group (6 mice/group). The KXS group was treated with KXS at a dosage of 5.4 g/kg body weight by oral gavage once daily for 3 months, and the vehicle group was treated with an equivalent volume of saline, and the anti-AD effect of KXS at such dosage has been proved in several previous studies [37, 38]. The dosage of KXS for the mice was based on clinical dose for human (weigh 70 kg) and then converted to mice dose according to the guidelines of FDA calculator based on the surface area of human and mice [39].

Behavioral Testing

After 3 months of KXS administration, mouse behavior was tested using Morris water maze (MWM) and the nest building tests, as described previously [40]. In general, mice were pretrained for 2 consecutive days in a circular water maze with a visible platform for MWM. The mice were expected to find the visible platform within 1 min; otherwise, mice were guided to find the platform and stayed on the platform for 15 s. After 2 days of training, the visible platform was hidden underneath 1 cm of water. Then, the mice were put into the water maze to find the hidden platform. This navigation test lasted for 5 days. The latency time and path length to find the hidden platform were recorded by the video-tracking SMART system version 3.0.06 (Panlab, Harvard Apparatus, MA, USA). In the probe trial on day 8, the platform was removed, and mice were allowed to swim for 1 min. The number of times that mice passed through the location of the platform was recorded.

In the nest building test, mice were singly housed. Eight pieces of paper in square shapes (5 \(\times\) 5 cm) were placed in the cage. The nest was photographed for 7 consecutive days and scored as follows: 1—no nesting, 2—nesting partly, 3—majority shredded with no identifiable nesting site, 4—an identifiable nesting site but with a flat part within the nest, and 5—perfect nest.

Tissue Preparation

After behavior tests, mice were sacrificed, and the brains were removed immediately and placed on ice. Each brain was divided into two hemispheres along the midline. One-half was snap frozen for biochemical analysis, and the other was fixed in 4% paraformaldehyde for morphological analysis.

Western Blot Analysis

Mouse brain tissues were lysed with ice-cold RIPA buffer containing 1% protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA; 8340) and 2% phosphatase inhibitors (Abcam, Cambridge, MA, USA; ab201112). The protein concentrations were measured by Bradford assay (Beyotime Institute of Biotechnology; P0010). Equal amounts of protein were separated on 10% SDS-PAGE gel and transferred onto PVDF membranes (Millipore, Billerica, MA, USA). Membranes were then incubated in 5% skim milk for 1 h and then incubated overnight at 4°C with the primary antibodies (Supplementary Table 2). Membranes were washed with TBS-T 3 times for 10 min each and then were incubated with horseradish peroxidase–labeled secondary antibodies (HRP 1:10000, Thermo Fisher Scientific, Waltham, MA, USA) for 1.5 h at room temperature. After washing 3 times, bands were detected using ECL (EMD Millipore). Protein band intensity was quantified using ImageJ software.

TNF-α ELISA

Mouse tumor necrosis factor (TNF-α) level was determined by ELISA kit (CUSABIO, China). Brain tissues were homogenized by sonication in 9 vol. of cold medium (Elabscience, China) and centrifuged for 5 min at 5000 g, at 4°C. The concentration of TNF-α in 100 μL samples was determined according to the manufacturer’s protocol, and the sample values were then obtained from the standard curve. Every sample was measured in duplicate. The OD value was measured at 450 nm in a microplate reader (Cytation5; BioTek, Winooski, VT, USA).

Immunohistochemistry

Tissue blocks from SAMP8 mouse brains were embedded in paraffin and coronally sectioned into 5-µm-thick slices, mounted onto gelatin-coated slides. Deparaffinization was performed with 2-h incubation at 60°C, followed by antigen retrieval in citric acid buffer. Immunohistochemistry was performed using a staining kit (MXB, Fuzhou, China). Briefly, the sections were quenched using 3% H2O2-methanol for 10 min. After washing and blocking with normal goat serum for 30 min, primary antibodies were applied (Supplementary Table 2) and sections were incubated overnight at 4°C. On the second day, after washing, biotinylated goat anti-rabbit IgG were applied and incubated for 2 h at room temperature followed by amplification with streptavidin peroxidase for 30 min. After incubation, the sections were developed in diaminobenzidine (Sigma-Aldrich, St. Louis, MO, USA) for coloration, and then immersed in distilled water to halt the reaction. The sections were then counterstained with hematoxylin and differentiated in hydrochloric acid alcohol. After dehydration, clearing, and mounting with neutral gum, the sections were observed under a light microscope (ECHO, San Diego, CA, USA). Blinded staining was performed and evaluated by two independent investigators.

Nissl Staining

Deparaffinization was performed with 2-h incubation at 60°C. According to the kit instructions, the procedure was conducted. The sections were treated with 0.1% cresyl violet at 56°C for 1 h. After differentiation for 2 min with Nissl differentiation solution, the stained sections were dehydrated with a graded series of ethanol solutions. Following clearing with xylene, the slices were mounted with neutral balsam and examined with a microscope. Blinded staining was performed by two independent investigators. Blinded staining was performed and evaluated by two independent investigators.

Statistical Analysis

All results were obtained from three independent experiments and presented as mean ± SEM. Differences between groups were evaluated by the unpaired two-tailed Student’s t-test. p < 0.05 was considered statistically significant.

Results

Network Pharmacology Predicts Potential Pharmacological Mechanisms for KXS Treatment of AD

In the present study, a total of 168 active compounds (including 76 in GR, 29 in PO, 25 in PR, 7 in AT and 2 compounds overlapped between GR and PO) of KXS were obtained from the ETCM online databases, corresponding to 863 prediction targets. These targets were further analyzed through the DisGeNet, AlzPlatform, and GeneCard databases to check if they were relevant with AD. Finally, 30 common targets of KXS in the treatment of AD were obtained. To facilitate visualization and further explanation of the target prediction results, we constructed the C-T network, which embodied 150 nodes (120 candidate compounds and 30 potential targets) and 364 compound-target interactions (Fig. 1a, Supplementary Table 3). The C-T network revealed that most compounds exerted multi-target effects, representing multifarious therapeutic effects. We further extracted the 30 significant targets to construct the PPI containing 30 nodes and 68 edges based on the STRING database, and the PPI enrichment p-value of these hub genes was 1.0 \(\times\) 10−16, indicating that they were at least partially biologically connected (Fig. 1b).

Fig. 1
figure 1

Network pharmacological prediction for KXS treatment of AD. a Compound-disease-target networks of KXS against AD. The rectangle nodes represent ingredients of herbs (orange for PO, purple for GR, green for AT, pink for PR, and red for Multidrugs-GR and PO), and the yellow ellipse nodes represent targets. b Protein-protein interactions identified by STRING software. c GO enrichment analysis of the intersecting targets’ biological process. d KEGG pathway enrichment analysis. e Target-pathway networks for KXS treatment of AD. The blue nodes represent the related pathways, and the ellipse nodes represent potential targets. The color depth and size of nodes are in proportion to their degree

Full size image

The candidate targets in the PPI network were elucidated by performing GO and KEGG pathway enrichment analysis. The top 20 items enriched in the BP category are shown in Fig. 1c. The results of GO enrichment analysis could be classified into some functional modules related with AD-associated pathological processes, including anti-inflammatory, anti-apoptotic, and antioxidant effects. Next, we performed KEGG pathway enrichment analysis to examine the pathways for putative therapeutic targets of KXS for AD. The top 20 pathways were determined (Fig. 1d), mainly including pathways in non-alcoholic fatty liver disease (NAFLD), PI3K-AKT signaling pathway, AD, and other dysregulation of predominantly inflammatory signaling pathways. All the targets interacting with the active ingredients were mapped onto the 85 KEGG pathways, and the T-P network was generated. As shown in Fig. 1e, the AKT target shows the highest number of pathway connection, followed by interleukin-6 (IL-6), interleukin-1 beta (IL-1β), caspase 3, Insulin (INS), GSK3β, Toll-like receptor 4 (TLR4), and type 1 insulin-like growth factor receptor (IGF1R). The details are provided in Supplementary Table 4. Moreover, molecular docking results indicated that Kaempferol, Onjisaponin A, Ginsenoside Ii, Chikusetsusapon, Ginsenoside F1, and Cis-9, Cis-12-Linoleic-Acid exhibited good affinity for AKT, IL-6, IL-1β, caspase 3, and TLR4, respectively. The details are provided in Supplementary Fig. 2.

KXS Attenuates Tau Hyperphosphorylation by Suppressing GSK3β and CDK5 Protein Kinase Activation in SAMP8 Mice

Based on the results of network pharmacology, GSK3β is AD-associated targets of KXS and the kinase responsible for Tau hyperphosphorylation [41]. Tau hyperphosphorylation is a predominant pathological hallmark of AD, and it contributes to AD development by causing synaptic impairments, neuronal dysfunction, and NFTs formation [42]. Therefore, we employed immunohistochemistry and Western blot to determine whether KXS treatment could regulate Tau phosphorylation through GSK3β signaling pathway. We investigated the Tau phosphorylation level at multiple sites by Western blot. We observed decreased Tau phosphorylation levels at Ser396, Thr231, Ser404, Thr181, and Ser214 in the KXS group (Fig. 2a). The staining results further suggested that KXS treatment significantly inhibited Tau hyperphosphorylation at both Ser404 and Thr231 sites in cerebral cortex, CA1, CA3, and DG regions (Fig. 2b and c). The phosphorylated Tau-positive cells were shown as brown granules, mainly expressed in the cytoplasm mostly with a perinuclear pattern.

Fig. 2
figure 2

KXS abrogates Tau hyperphosphorylation in SAMP8 mice. a The levels of Tau phosphorylation at Ser396, Thr231, Ser404, Thr181, and Ser214 proteins were examined by Western blot assays. Tau phosphorylation at Ser404 (b) and Thr231 (c) immunohistochemical staining in cerebral cortex, CA1, CA3, and DG regions with the vehicle (Veh) group and KXS group. Right panels of each group show the representative images in high magnification. Arrows indicate the phosphorylation Tau positive staining. Values were the mean ± SEM (n = 5–6). *p < 0.05; **p < 0.01 versus the vehicle group. Scale bars: 60 μm. Scale bar is 30 μm in the high magnification of right panels

Full size image

GSK3β, CDK5, and P35/25 are key participants of Tau hyperphosphorylation. Thus, we further evaluated these kinase expression levels by Western blot. The significantly decreased p-GSK3β (Tyr216)/GSK3β ratio and increased p-GSK3β (Ser9)/GSK3β ratio were observed in KXS group (Fig. 3), which were consistent with the network pharmacology results. In addition to GSK3β, p-CDK5 protein expression in SAMP8 mice brains was also significantly repressed when compared to vehicle group (Fig. 3). In contrast, the p35 and p25 expression levels were not significantly different between the vehicle and KXS groups (Fig. 3). Furthermore, p-Akt (Ser473) phosphorylates GSK3β to inhibit its activity. And we have detected the expression of p-AKT(Ser473), AKT in brains from both groups. The significantly increased p-AKT(S473)/AKT ratio was observed in KXS group. The results indicate that KXS inactivates GSK3β by an increase in p-Akt (Ser473). All the results indicate that KXS may inhibit Tau hyperphosphorylation by suppressing the activation of GSK3β and CDK5.

Fig. 3
figure 3

KXS represses Tau hyperphosphorylation by suppressing the AKT/GSK3β signaling and CDK5 protein kinases. a Western blot analysis of p-AKT (Ser473), AKT, p-GSK3β (Ser9), p-GSK3β (Tyr216), GSK3β, p-CDK5 (Tyr15), CDK5, p35, and p25 expression in the brains of vehicle group and KXS group mice. b Quantified protein expression levels of above protein. Values were the mean ± SEM (n = 5–6). *p < 0.05; **p < 0.01 versus the vehicle group

Full size image

KXS Inhibits Neuroinflammation in the Brain of SAMP8 Mice

The network pharmacology analysis indicated that the therapeutic mechanism of KXS for AD involves neuroinflammation. To investigate whether KXS prevents inflammatory response, we evaluated the astrocytic and microglial activities by detecting the expression of anti-GFAP and anti-Iba1 antibodies in the hippocampus and cortex of SAMP8 mice. Arrows indicate the GFAP or Iba1-positive cells. We observed that the expression of GFAP and Iba1 following KXS treatment significantly reduced compared to the vehicle mice (Fig. 4a and b). We also measured GFAP and Iba1 levels using Western blot, and the results were consistent with immunochemistry (Fig. 4c).

Fig. 4
figure 4

KXS inhibits neuroinflammation in the brain of SAMP8 mice. a, b Astrocytes were labeled with anti-GFAP antibodies; microglia were labeled with anti-Iba1 antibodies. Right panels of each group show the representative images in high magnification. Arrows point to GFAP or Iba1 positive cells. c Western blot analysis of GFAP and Iba1, and quantification, in the SAMP8 mice brain. d Representative immunohistochemistry images showing the labeling of NLRP3 within the cerebral cortex and CA1, CA3, and DG region areas of the hippocampus from the brain slices of vehicle group and KXS group mice. Right panels of each group show the representative images in high magnification. The arrow indicates the NLRP3 positive cells. e Western blot analysis revealed the protein expression of NLRP3, IL‐1β, and IL-6 in the brains of vehicle group and KXS group mice. f Inflammatory factors TNF-α detected by ELISA. Values were the mean ± SEM (n = 5–6). *p < 0.05; **p < 0.01 versus the vehicle group. Scale bars: 60 μm. Scale bar is 30 μm in the high magnification of right panels

Full size image

Considering the pathogenic role of inflammation in the development of AD, the expression levels of several important inflammatory factors were further evaluated by immunohistochemistry, Western blot, and ELISA, including nucleotide-binding oligomerization domain-like receptor family, pyrin domain containing 3 (NLRP3), IL-1β, IL-6, and TNF-α. The results showed KXS treatment significantly decreased NLRP3 inflammasome protein expression (Fig. 4d and e). The assembled NLRP3 inflammasome is responsible for the cleavage of IL-1β into secreted form [43]. Therefore, in the present study, we found that KXS repressed the secretion of IL-1β (Fig. 4e). Moreover, IL-6 and TNF-α were downregulated in the KXS group compared to the vehicle group (Fig. 4e and f). These results suggest that KXS effectively attenuates microglia and astrocyte activation, and effectively suppress the inflammatory cytokine secretion.

KXS Inhibits TLR4/MyD88/NF-κB Signaling Pathway in SAMP8 Mice Brain

Based on the network pharmacology results, TLR4 and its downstream proteins, myeloid differentiation factor 88 (MyD88) and nuclear factor-kappa B (NF-κB), were selected to further explore the effect of KXS on the neuroinflammation pathway. Our results demonstrated that KXS significantly decreased the expression of TLR4 (Fig. 5). As the known downstream effector of TLR4, MyD88 and NF-κB are critical for TLR4-mediated inflammation; therefore, we determined their expression level by Western blot. As shown in Fig. 5, the protein expression of MyD88 was significantly decreased by KXS treatment. Besides, we also observed a significant reduction in p-NF-κB (Ser536) and NF-κB expression in KXS treatment group (Fig. 5). Taken together, the results indicate that KXS attenuates neuroinflammation in SAMP8 mice by inhibiting the TLR4/MyD88/NF-κB signaling pathway.

Fig. 5
figure 5

KXS inhibits TLR4/MyD88/NF-κB signaling pathway in SAMP8 mice brain. a Western blot analysis of TLR4, MyD88, NF-κB subunit p65, and NF-κB subunit p65 phosphorylation on ser536. b Quantitative analysis of TLR4, MyD88, NF-κB, and p-NF-κB protein expression. Values were the mean ± SEM (n = 5–6). *p < 0.05; **p < 0.01 versus the vehicle group

Full size image

KXS Attenuates Apoptosis in SAMP8 Mice

Our network pharmacology analysis suggested that KXS could regulate the expression of the target protein caspase 3, which is the critical effector of apoptosis. To verify whether KXS participates in the regulation of apoptosis, several apoptosis-related proteins were determined by Western blot analysis. A significant decrease in the cleaved-caspase 3 and the cleaved-caspase 1 level was observed in the KXS-treated group (Fig. 6). Besides, KXS enhanced B cell lymphoma 2 (BCL2) expression and suppressed BCL2-associated X protein (BAX) expression, resulting in upregulation of BCL2/BAX ratio. Nissl stains were performed to assess the neuroprotective functions of KXS on the mouse brain. As shown in Supplementary Fig. 3, compared with KXS-treated SAMP8 mice, the vehicle group exhibited significant nuclear breakdown and less intact Nissl substance in the cortex and hippocampus. All the results suggest that KXS attenuates neuronal apoptosis by decreasing the BCL2/BAX ratio, inhibiting caspase 1 and caspase 3 activity in SAMP8 mice.

Fig. 6
figure 6

Protective effect of KXS against apoptosis in SAMP8 mice. a Western blot analysis of caspase 3, cleaved caspase 3 (c-caspase 3), caspase 1, cleaved caspase 1 (c-caspase 1), BCL2, and BAX. b Quantitative analysis of caspase 3, cleaved caspase 3, c-caspase 3/ caspase 3 ratio, caspase 1, cleaved caspase 1, c-caspase 1/caspase 1 ratio, BCL2, BAX, and BCL2/BAX ratio. Values were the mean ± SEM (n = 5–6). *p < 0.05; **p < 0.01 versus the vehicle group

Full size image

KXS Ameliorates Memory Deficits in Aged SAMP8 Mice

To investigate whether KXS treatment improved cognitive function in SAMP8 mice, MWM and nest building tests were performed. In the visible-platform tests of MWM, there were no significant differences between vehicle and KXS-treated groups (Fig. 7a), indicating that KXS administration did not influence the motility or vision of the mice. However, KXS-treated SAMP8 mice exhibited significantly shorter escape latency and travel distance to the platform than the vehicle group on days 5, 6, and 7, as shown in Fig. 7a and b. And both groups had similar swimming speed (Supplementary Fig. 4), which suggests that the differences in latency time between groups are not due to the variation in motoric abilities. All the results indicating that KXS could markedly improve the spatial memory of SAMP8 mice. In the probe trial, KXS-treated mice exhibited no significant differences compared to the vehicle mice, with the number of platform location crossings of KXS group showing an increasing trend (Fig. 7c and d).

Fig. 7
figure 7

KXS treatment ameliorates the spatial learning and reference memory deficiency of SAMP8 mice. a Escape latencies in visible and hidden platform during MWM task training in both the vehicle group and KXS group. b Movement track of both groups in in the hidden platform trail. c Number of crossing the platform location of each group in the probe test. d Movement track of both groups in the probe trail. e Representative images were obtained from 0 to 7days in the nesting behavior test. f Nest-building scores for each group. Values were the mean ± SEM (n = 6). *p < 0.05; **p < 0.01 versus the vehicle group

Full size image

We further assessed the effects of KXS on general health and hippocampal function with the nest building test. Treatment with KXS resulted in significant improvements in nesting scores, when compared to the vehicle group (Fig. 7e). Figure 7f shows the representative results in nest building test. Taken together, the results suggested that KXS improves cognitive impairment in SAMP8 mice.

Discussion

Previous studies have shown that KXS, a classic formula for amnesia, ameliorated the pathological changes and improved cognitive function in animal models of AD [44, 45]. However, the bioactive ingredients of KXS and the drug targets are still not clear, which is unfavorable to KXS modern development and clinical utility. In the present study, we used network pharmacology to identify the bioactive ingredients and action mechanism of KXS in the treatment of AD. Since network pharmacology was based on computational analysis only, further verification is needed to validate the therapeutic mechanism of KXS. Therefore, a series of experiments were carried out to validate the predicted molecular mechanisms obtained from network pharmacology analysis. Network pharmacological analysis revealed a total of 168 active compounds and 863 targets in the KXS formula. We further identified 30 targets closely correlated with AD through Venn diagram. Based on the aforementioned GO and KEGG pathway enrichment analysis, an integrated AD pathway can be separated into four representative therapeutic modules, namely, Tau hyperphosphorylated module, inflammation module, cell apoptosis module, and oxidative stress module. Among relevant protein targets with the 85 related pathways, AKT, IL-6, IL-1β, caspase 3, INS, GSK3β, TLR4, and IGF1R were the key target nodes with the top degree, indicating that they may be the most likely targets for AD therapy of KXS. Previous research has found that KXS could improve learning and memory impairment in AD animal model by inhibiting oxidative stress [44, 46]. To investigate the mechanisms of the regulatory function of KXS during AD pathogenesis, in addition to antioxidant activities, we also wondered whether KXS influences the other modules.

Network pharmacology analysis suggested AKT and GSK3β were the important therapeutic target of KXS. GSK3β is the key kinases for Tau phosphorylation. The activation of GSK3β could be inhibited by phosphorylation at Ser9 or promoted by phosphorylation at Tyr216. Overactivity or overexpression of GSK3β increases the phosphorylation of most serine and threonine residues of Tau, which leads to Tau aggregation and neuronal dysfunction in AD [47,48,49,50]. In line with the predicted results of network pharmacology, we detected a remarkably decreased level of GSK3β (Tyr216) and increased level GSK3β (Ser9) of in KXS-treated group, which indicated a significant inhibition of GSK3β activation. The best studied mechanism by which GSK3β is regulated is Akt-dependent serine phosphorylation [51]. AKT phosphorylates GSK3β at the S9 residue, which inactivates GSK3β [52]. And the phosphorylation of Akt (Ser473) is the primary phosphorylated form after the activation of Akt[53]. As a result, we examined Ser473 phosphorylation of Akt in this study. KXS significantly increased p-AKT(S473) expression, which indicates that KXS prevent tau phosphorylation via Akt/GSK3β signaling because Akt activation inhibits tau phosphorylation through GSK3β inactivation. In addition, we also evaluated p-CDK5(Tyr15), CDK5, p35, and p25 levels after KXS treatment. CDK5 appears to be key factor in Tau phosphorylation, and the activation of CDK5 is significantly increased in AD [53, 54]. Studies show that the CDK5 can be activated by phosphorylation at Tyr15 residues [54, 55] and the binding of p25 (the truncated form of p35) [56]. Our results showed that KXS remarkably reduced the level of p-CDK5 at Tyr15 sites, while the levels of CDK5, p35, and p25 were not changed. Consistently, the phosphorylation of Tau at several AD-related sites (Ser396, Thr231, Ser404, Thr181, and Ser214) was significantly reduced in both the hippocampus and cortex of SAMP8 mice by KXS treatment. Taken together, all these imply that KXS exerts inhibited function on Tau hyperphosphorylation in SAMP8 mice via predicted targets AKT/GSK3β and CDK5.

The network pharmacology analysis has identified KXS could modulate neuroinflammation through regulating IL-6, IL-1β, and TLR4. It is reported that the inflammatory processes during the pathogenesis of AD are mostly mediated by microglia and astrocytes [57]. Tau pathology is also regulated by neuroinflammations [58]. Therefore, we examined the activation of astrocytes and microglia by evaluating GFAP and Iba-1 level, which have been reported to be upregulated in the activated astrocytes and microglia in AD [59]. We found that KXS reduced significantly GFAP and Iba1 expression levels in the brain of SAMP8 mice, which indicated that KXS has a significant anti-inflammatory activity. Additionally, crosstalk between TLR4 and NLRP3 inflammasome promotes neuroinflammation in AD [60]. TLR4 is a transmembrane protein belonging to the pattern recognition receptor family [61], which is regarded as neuroinflammatory receptors detected in neurons, astrocytes, and microglia [62, 63]. MyD88 is an adaptor protein downstream of most TLRs and leads to the activation of NF-κB [64]. Once activated, the TLR4/MyD88/NF-κB signaling would initiate the NLRP3 inflammasome activation [65] and other cytokines (especially IL-6 and TNF-α) production [66, 67], which are considered as the key contributor to neuroinflammation during neurodegeneration. NLRP3 inflammasome is a multimeric protein that consists of NLRP3, adaptor protein ASC, and pro-caspase 1 [68, 69]. The activation of NLRP3 inflammasome leads to the cleavage of pro-caspase 1 into active caspase 1, which further cleaves pro-IL-1β into mature forms, and eventually drives the Tau pathology in AD brain [70, 71]. In our present study, we detected significantly reduced TLR4, MyD88, and NF-κB protein expression levels in KXS-treated mice. Consistent with the results, we also found KXS effectively suppressed activation of the NLRP3 inflammasome in SAMP8 mice. Pro-inflammatory elements such as IL-6, TNF-α, and IL-1β ultimately exacerbate neuroinflammation by stimulating the synthesis of Aβ and phosphorylation of Tau [72,73,74]. As expected, we also detected a decreased expression level of IL-6, TNF-α, and IL-1β in KXS-treated SAMP8 mice, which is consistent with the findings from other network pharmacology studies [75]. Our study indicates that KXS might exert inhibitory activities against neuroinflammation by suppressing NLRP3 inflammasome activation and pro-inflammatory cytokine (IL-6, TNF-α, and IL-1β) production through TLR4/MyD88/NF-kB signaling pathway. Our work complements and extends other studies on KXS using network pharmacology.

Hyperphosphorylated Tau and neuroinflammation contribute to the neuronal apoptosis in AD. In the present study, we demonstrated that KXS has remarkable inhibitory activity on Tau hyperphosphorylation and inflammation; joined with the network pharmacology analysis, we speculate that KXS may have anti-apoptosis activity. Previous studies have shown that caspase 1 activation, caspase 3 activation, and BCL2/BAX interference are directly involved in apoptosis [76, 77]. BCL2 exerts inhibitory effects on apoptosis while BAX exerts pro-apoptotic effects [78,79,80,81]. Caspase 3 is the key effector of apoptosis, the activation of caspase 3 would initiate irreversible apoptosis [82]. Caspase 1 activation induces the maturation and production of IL-1β and subsequently potentiates inflammatory responses [83, 84]. Proinflammatory cytokines such as IL-1β and TNF-α could also cause neuronal apoptosis via neurotoxic effect [85]. Our results showed that KXS may upregulate BCL2/BAX ratio, inhibit caspase 1 and caspase 3 activity, and alleviate apoptosis in the SAMP8 mice brain. Based on our finding, an overview of mechanism diagram for the neuroprotective effect of KXS against AD pathogenesis is shown in Fig. 8.

Fig. 8
figure 8

Diagram of the mechanism of action of KXS against AD pathogenesis

Full size image

In light of the inhibitory role of KXS on Tau hyperphosphorylation, neuroinflammation, and apoptosis, we also considered the cognitive function of the KXS-treated mice. We observed that KXS significantly improved cognitive decline of SAMP8 mice in MWM and nest building test. Furthermore, INS and IGF1R, predicted targets of KXS in AD therapy, are also associated with diabetes. Thus, KXS may be beneficial for slowing the progression of AD through improving insulin resistance which requires further investigation in the future. We also found Kaempferol, Chikusetsusapon, Ginsenoside F1, Ginsenoside Ii, Onjisaponin A, Cis-9, and Cis-12-Linoleic-Acid were the core genes targeted by KXS compounds. These results suggested that KXS exerts therapeutic efficacy through the synergistic effect of multi-compounds, multi-targets, and multi-pathways. It is noteworthy that the multiple compounds of KXS may provide a mutual enhancement effect on AD treatment, but this must be further tested using single or mixed compounds.

Conclusion

In this work, we combined network pharmacology approach and experimental validation in vivo to explore and verify the mechanisms of KXS’s potential in treating AD. We found that KXS upregulated AKT phosphorylation, suppressed GSK3β and CDK5 activation, and inhibited the TLR4/MyD88/NF-κB signaling pathway to attenuate Tau hyperphosphorylation and neuroinflammation, therefore suppressing neuronal apoptosis and improving the cognitive function of SAMP8 mice. Our present study clearly highlights the therapeutic value of KXS for attenuating the progression of AD. Furthermore, our study sheds novel lights on investigating the mechanisms of the TCM formula in AD therapy from the perspectives of bioinformatic methods. Our study is crucial for the future development of TCM modernization, including KXS, giving hopes for their further and broader clinical application.