Exploring the Mechanisms Underlying the Therapeutic Effect of Xixin Decoction on Alzheimer's Disease Based on Network Pharmacology and Molecular Docking

Background and objective: Alzheimer’s disease (AD) has been a worldwide problem, not only the treatment but also the prevention. As a commonly used Chinese Herbal Formula, Xixin Decoction (XXD) has signi�cant therapeutic effect on AD but without clear mechanism. This study was aimed to predict the main active compounds and targets of XXD in the treatment of AD and to explore the potential mechanism by using network pharmacology and molecular docking. Methods: The compounds of XXD were searched in the TCMSP and the TCMID database, and the active compounds were screened based on the ADME model and SwissADME platform. SwissTargetPrediction platform was used to search for the primary candidate targets of XXD. The common targets related to AD obtained by two databases (GeneCards and DisGeNET) were determined as candidate proteins involved in AD. To acquire the related targets of XXD in the treatment of AD, the target proteins related to AD were intersected with the predicted targets of XXD. Then these overlapping targets were imported into the STRING database to build PPI network including hub targets; Cytoscape 3.7.2 software was used to construct the topology analysis for the herb-compound-target network diagram while one of it’s plug-in called CytoNCA was used to calculate degree value to screen the main active compounds of XXD. GO and KEGG pathway enrichment analyses were conducted to explore the core mechanism of action and biological pathways associated with the decoction via Metascape platform. We used AutoDock Vina and PyMOL 2.4.0 softwares for molecular docking of hub targets and main compounds. Results: We determined 114 active compounds which meet the conditions of ADME screening, 973 drug targets, and 973 disease targets. However, intersection analysis screened out 208 shared targets. PPI network identi�ed 9 hub targets, including TP53, PIK3CA, MAPK1, MAPK3, STAT3, AKT1, etc. The 10 main active compounds play a major role in treatment of AD by XXD. Hub targets were found to be enriched in 10 KEGG pathways, involving the Pathways in cancer, AGE-RAGE signaling pathway in diabetic complications, Alzheimer's disease, Neuroactive ligand-receptor interaction, Dopaminergic synapse, Serotonergic synapse and MAPK signaling pathway. The docking results indicated that the 8 hub targets exhibit good binding activity with the 9 main active compounds of XXD. Conclusions: We found the advantages of multi-compounds-multi-targets-multi-pathways regulation to reveal the mechanism of XXD for treating AD based on network pharmacology and molecular docking. Our study provided a theorical basis for further clinical application and experimental research of XXD for anti-AD in the future.


Introduction
Alzheimer's disease (AD) is a neurodegenerative disease, which is the commonest cause of dementia.The primary clinical features of dementia are de ciency of activity of daily living, abnormal behavior and cognitive impairment.The disease's growing high incidence rate and its pathogenesis which has not yet been fully elucidated are global public health concerns for not only individuals and caregivers but also the whole society.Current surveys estimate that there are over 50 million people living with dementia worldwide at present [1].The number of new AD cases all around the word is estimated to reach 100 million by 2050 [2].Therefore, the prevention and treatment of AD are a worldwide problem to be solved.The pathological features of AD are neurodegeneration, neuron damage and apoptosis, formation of senile plaques involving the amyloid-β(Aβ) accumulation, neuro brillary tangles involving hyperphosphorylation of the tau protein, as well as neuroin ammation, abnormal insulin signaling pathway and brain insulin resistance [3].
However, studies that target on Aβ and the tau protein to treat AD didn't show satisfactory results.Current drugs, such as Donepezil and Memantine, are not effective enough to meet clinical needs.Therefore, it is necessary to study the treatment of AD from the perspective of multi-target therapy.
Traditional Chinese medicine (TCM) has been used in treating AD which can be traced back to the traditional Chinese medicine book "TSO Chuan" for thousands of years and previous dynasties Chinese medicine physicians have accumulated a tremendous amount of signi cant clinical evidences.The advantages of TCM in the treatment of AD are as follows:(1)TCM treatment improves patients' cognitive function and make it last longer; (2)TCM aims to some physical symptoms that are di cult to solve with western medicine, such as behavioral and psychological symptoms of dementia(BPSD),loss of appetite, insomnia, constipation;(3)TCM has less side effects than western medicine on liver, kidney, stomach and other organs.(4) Multiple targets of TCM act on diseases through various pathways.
XXD comes from "Syndrome Differentiation Record" written in Qing Dynasty.It has the effect of opening depression, expelling phlegm and invigorating the stomach, consisting of Renshen(Panax ginseng), Fushen(Poria cum Radix Pini), Suanzaoren(Ziziphus jujuba Mill), Banxia(Pinellia ternate), tangerine peel(Citrus reticulata Blanco), Medicated Leaven (Massa Medicata Fermentata), licorice(Glycyrrhiza uralensis Fisch), aconite(Aconitum carmichaeli Debx), Acorus calamus(Acorus tatarinowii).The formula has been mainly used to treat dementia.A large number of experiments have found that XXD can prevent SAD(sporadic Alzheimer's disease) pathological progress by inhibiting hyperphosphorylation at the key sites of tau proteins [4], enhancing O-GlcNAc glycosylation of tau proteins [5], improving the expressions of synaptic functional proteins and receptors [6], repairing mitochondrial function and increasing the survival rate of Aβ 1−42 induced cells [7].No report has systematically analyzed the constituents of XXD and its potential mechanism of action in treating dementia so far.
As an interdisciplinary product of traditional pharmacology, molecular biology, bioinformatics and other disciplines, network pharmacology has broken through the traditional concept of "single target-single pathway" drug research.Its holistic and systematic characteristics coincide with TCM holistic view, the principle of syndrome differentiation and treatment, and the compatibility of prescriptions.It provides a new idea and method for elucidating the speci c mechanism of XXD in the treatment of AD.The purpose of this study is to explore the main material basis, core target and potential mechanism of XXD's action in treating AD using the network pharmacology method and molecular docking.The study design owchart was shown in Fig. 1, the active compounds and intersection genes for AD were collected at rst.Then, the protein-protein interaction (PPI) and Pathway-target-compound network were constructed.In addition, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were performed.Furtherly, the potential main active compounds and core targets were validated by the molecular docking simulation.

Screening of the active compounds of XXD
Traditional Chinese Medicine System Pharmacology Database (TCMSP) is a systems pharmacology platform of Chinese herbal medicines that captures the relationships between drugs, targets and diseases [8].It includes chemicals, targets and drug-target networks, and associated drug-target-disease networks, as well as pharmacokinetic properties for natural compounds involving oral bioavailability, drug-likeness, intestinal epithelial permeability, blood-brain-barrier, aqueous solubility and etc. Traditional Chinese Medicine Integrated Database (TCMID) is a comprehensive database to provide information and bridge the gap between traditional Chinese medicine and modern life sciences, having collected information on all respects of TCM including formulae, herbs and herbal ingredients [9].It has also contained information for drugs, diseases which are deeply studied by modern pharmacology and biomedical sciences.We searched TCMSP and TCMID to collect the pharmacologically active compounds of XXD.ADME (absorption, distribution, metabolism, and excretion) screening model is an effective model to screen the main active ingredients of herbs in the TCMSP database by using pharmacokinetic parameters such as oral bioavailability (OB), druglikeness (DL) and blood brain barrier (BBB).OB is de ned as the rate of active ingredient being absorbed from a drug product and becoming available at the site of action and DL is a qualitative concept describe the optimize pharmacokinetics and molecular properties [10].In this study, active compounds were screened according to ADME features with OB ≥ 30%, DL ≥ 0.18 and BBB≥-0.3.113 compounds were obtained after removing the overlapping compounds among the 9 herbs of XXD and they were used to the subsequent network pharmacology study.

Prediction Of Targets Of Active Compounds
SwissTargetPrediction database can predict compound targets according to 2D and 3D structures of known compounds [11].The targets of XXD's active compounds were predicted from this database and the probability value ≥ 0 served as the target screening standard."Common name" in the information table refers to gene symbol.

Collection Of Targets For Ad
We screened AD-related genes by using DisGeNET [12] and GeneCards [13].DisGeNET is a comprehensive discovery platform containing publicly available collections of genes and variants associated to human diseases.GeneCards is a searchable, integrative database that provides predicted human genes and user-friendly information, including genomic, transcriptomic, proteomic, genetic, clinical and functional information.We searched AD-related targets from these two databases using "Alzheimer's Disease" as keyword.In GeneCards database, the higher the score is, the closer the target is related to the disease.As a rule of thumb, if there are too many targets, the targets with score greater than the median score is set as the potential target of AD.After combining the two databases' targets and deleting the repeated targets, we got the AD target.

Target Mapping
To acquire the candidate targets responsible for AD therapy, the target proteins related to AD were intersected with the predicted targets of XXD.For overlapping targets, numbers were visualized with a Venn diagram plotted by http://www.bioinformatics.com.cn, an online platform for data analysis and visualization.

Protein-protein Interaction (Ppi) Analysis
The overlapping targets were then imported into STRING [14], a database that predict interactions including direct (physical) and indirect (functional) associations, to build protein-protein interaction(PPI) network which would be used for the subsequent topological properties analysis to identify hub targets in treatment of AD, and the targets were selected with the highest con dence score (low: < 0.4; medium:0.4to 0.7; high: > 0.7; highest :> 0.9).

Go And Kegg Pathway Enrichment Analysis
Metascape [15] is a web-based portal designed to provide a comprehensive gene list annotation and analysis resource for experimental biologists.It combines functional enrichment, interactome analysis, gene annotation, and membership search and facilitates comparative analyses of datasets across multiple independent and orthogonal experiments.The candidate targets' gene symbols of XXD responsible for AD therapy were imported into Metascape platform, setting species as "Homo sapiens", and then the Gene Ontology (GO)biological process enrichment and Kyoto Encyclopedia of Genes and Genomes(KEGG) enrichment were analyzed and enriched [16].The smaller modi ed sher exact p-value, the more enriched.
Network Construction Cytoscape 3.7.2software was used to construct the following networks and topology analyses: (1) The herb-compoundtarget network of XXD; (2) The network of shared targets between XXD's active compounds targets and AD targets; (3) The network of shared targets and their corresponding compounds;(4) The pathway-target-compound network of XXD for treating AD.A plug-in of Cytoscape called CytoNCA was used to calculate parameter to screen the main active compounds of XXD and the hub targets of shared targets.Degree [17], betweenness centrality(BC),closeness centrality(CC) and Local average connectivity-based method(LAC)[18] are the common topology parameters for evaluating the central properties of nodes in a network.The main active compounds were screened through the degree value size sorting.The screening parameter settings of hub targets of the XXD-AD target interaction network are as follow: The rst screening threshold was Degree ≥ twofold median, the second threshold was Degree ≥ median, BC ≥ median, CC ≥ median and LAC ≥ median.The higher the values are, the more signi cant the nodes are.The screened hub target genes are used for further molecular docking study.

Molecular Docking Technology
We investigated the possible interaction activity between hub genes and main active compounds of XXD with molecular docking.Meanwhlie, the hub targets and main active compounds were further screened by the docking a nity values or reported by AutoDock Vina.There were 72 pairs delivered into the docking simulation (Table 4).The scoring function of AutoDock and AutoDock Vina stipulates that when the docking a nity values < 0, the ligand and the receptor molecule can bind stably, and the greater the absolute value, the stronger the binding stability between the compounds and the active site of the hub targets.According to the docking results, most binding complexes possessed high binding a nity values with an average of − 9 kcal/mol.Besides, there is at least one hydrogen bond formation between ligand of every target and residues of active compounds.And most hydrogen bond interactions were considered as strong interaction with the average distance of 3.4, also recorded in Table 4. 9 pairs docking models with hydrogen bond and their a nity are shown in Fig. 9.Over all, molecular docking indicated that all XXD's active compounds could easily enter and bind the active pocket of the hub targets well.

Results
Active compounds of XXD XXD compounds were retrieved from the TCMSP and TCMID database and after the screening with OB ≥ 30%, DL ≥ 0.18 and BBB≥-0.3,114 active compounds were obtained after deleting repetitions.The details besides common active compounds of different herbs are shown in the Supplementary le, Table S1.

Potential Targets
We gained a total of 973 XXD target genes from SwissTargetPrediction and 973 AD targets from DisGeNET and GeneCards, respectively showed in Table S2 and Table S3.All the targets' names were put into UniProtKB to be converted to gene symbol.Then, we obtained 208 shared targets by further integrating the potential XXD-target genes and AD-related genes(Fig.2and Table S4).

Ppi Analysis Of Shared Targets
The PPI network of shared targets containing 174 nodes constructed by Cytoscape 3.7.2software, after selected with the highest con dence score 0.9 in the STRING database, depicts the interactions between them identi ed, as shown in Fig. 3A.The hub targets were screened by plug-in CytoNCA through calculating four topological features by two steps for each node in the network.The rst screening threshold was Degree ≥ 14 median, which resulted in 28 nodes and 161 edges.Then these 28 nodes were screened with the second threshold of Degree ≥ 22.5, BC ≥ 0.04, CC ≥ 0.44 and LAC ≥ 5.7, ultimately remained nine hub targets including TP53, PIK3CA, MAPK1, MAPK3, STAT3, AKT1, NFKB1, EGFR and HSP90AA1(Table 1 and Fig. 3B ).

Networks Construction
The information of 9 herbs,114 active compounds and 970 XXD targets were imported into Cytoscape3.7.2 software and the herb-compound-target network of XXD was constructed(Fig.4); The network of 208 shared targets, their corresponding compounds and herbs are shown in Fig. 5 with details in Table S5, while using Network Analyzer to calculate degree and sorting 10 main active compounds shown in Table 2 and Fig. 6;

Go Analysis
The gene ontology (GO) enrichment analysis was consisted of BP (biological process), CC (cellular component), and MF (molecular function).We used GO to further elucidate the signi cance of overlapping genes targets.As shown in the results of the enrichment, the top 10 GO items are selected based on counts of hit genes and P values (Fig. 7A).For biological processes, the targets were mainly enriched in cellular response to nitrogen compound (GO:1901699, GO:0010942), positive regulation of kinase activity (GO:0033674, GO:0051347, GO:0043408), cellular response to organonitrogen compound (GO:0071417) and transmembrane receptor protein tyrosine kinase sign (GO:0007169).
Additionally, the rest functional annotations are associated with response to inorganic substance (GO:0010035) and behavior (GO:0099537, GO:0099536).XXD effected AD by regulating three principal CC, namely, dendrite, membrane raft and postsynapse, which might re ect that most ingredients of XXD were targeted to neural cells.Protein kinase activity, kinase binding, endopeptidase activity and protein domain speci c binding were ranked as top four MF.

Kegg Pathway Enrichment Analysis
We conducted KEGG pathway enrichment of 208 shared targets to determine the potential therapeutic mechanism of the XXD for AD.Based on the log10(P value), the bubble plot of the most signi cant 20 KEGG pathways was shown in Fig. 7B.We sorted top 10 pathways including Pathways in cancer (hsa05200), AGE-RAGE signaling pathway in diabetic complications (hsa04933), Alzheimer's disease (hsa05010), Toxoplasmosis (hsa05145), Neuroactive ligand-receptor interaction (hsa04080), Dopaminergic synapse (hsa04728), Serotonergic synapse (hsa04726), MAPK signaling pathway (hsa04010), Viral carcinogenesis (hsa05203) and Transcriptional misregulation in cancer(hsa05202) with details in Table 3.The pathway maps of the Alzheimer's disease and MAPK signaling pathway were illustrated in Fig. 7(C, D).
Then we constructed a pathway-target-compound network based on the shared targets enriched in each pathway (Fig. 8).

Discussion
AD is a complicated neurodegenerative disease characterized by a progressive loss of cognitive function resulted from many etiological factors that causes and pathogenesis are not clear.So, combination treatments rather than monotherapy are warranted for the development of effective approaches in treating AD.The unique advantages of TCM such as naturally sourced, multitargeted, individual mode and a holistic concept are obvious in treating complicated diseases especially those with a poor e cacy to western medicine alone.In the theory of TCM, AD is mainly caused by old and weak with de ciency of healthy Qi as a result of endogenesis of turbid phlegm, blinding the clear ori ces.The basic pathogenesis is de ciency of origin and excess of sign.XXD is a classical formula which is used to resolve phlegm, induce resuscitation and nourish the heart to calm the mind.XXD has been used to treat dementia clinically for a long history from hundreds of years ago in China.But it is undeniable that the complexity of compounds and compatibility in the use of TCM, as well as the dynamic process of the development of chronic and complex diseases, bring great challenges to the research of TCM.With the continuous development of bioinformatics, TCM network pharmacology has become a topic on hot spots to reveal above mentioned complex biological processes from the perspective of integrated multi-component networks [22].There are few studies on the mechanism of XXD in the prevention and treatment of AD.To reveal the bene cial effects of XXD on AD, we used network pharmacology approach to explore the putative active compounds and potential mechanisms.
In our study, we constructed the following network to reveal the potential targets and pathways of XXD in AD treatment: compound-target network, PPI network of shared targets-corresponding compounds and compound-hub target -pathway network.After integrating and consolidating information from diverse sources of available databases, 10 active compounds of XXD acted on 208 different targets associated with AD.According to the shared targets-corresponding compounds network,10 compounds and 208 shared targets are highly connected and they can be de ned as vital compounds in XXD.In XXD, according to the TCM theory in formula prescribing, the Jun or "monarch herb" is Pinellia ternate (Banxia), exerting the effects of resolving phlegm.The Chen or "minister herbs" consist of Panax ginseng (Renshen) and Aconitum carmichaeli Debx(Fuzi), with the function of tonifying Qi and Warming and activating Yang Qi, respectively.In addition, Glycyrrhiza uralensis Fisch (Gancao), playing the role as Zuoshi or "adjuvant and guide herb", helps tonifying Qi and harmonizing various herbs.In our study, three of main active compounds including Gomisin B, baicalein and Karanjin, were identi ed from the "monarch herbs", the "minister herbs" and the"adjuvant and guide herb", respectively.This prescription feature might provide the pharmacological evidences for XXD to be used as multi-target regulation in treating AD.
Moreover, according to the pathway-target-compound network, components in XXD such as licochalcone a, Gomisin B,7-Methoxy-2-methyl iso avone, baicalein, Glypallichalcone, zizyphusine,7-Acetoxy-2-methyliso avone, Karanjin, Girinimbin and 6-prenylated eriodictyol interacted with a large number of targets, indicating the important roles in the anti-AD system.Some of them have been reported to show biological activities against AD in the past.For example, schisantherin B(Gomisin B) identi ed in Renshen was able to protect against learning and memory impairments and remit the toxicity caused by excessive activation of hyperphosphorylated tau in the dementia mouse model induced by Aβ 1−42 through regulating glial glutamate transporter type 1 (GLT-1), which has an effect on the level of p-tau protein [23].
Baicalein in Banxia was proved to show therapeutic potential for AD through reducing oxidative stress, antiin ammatory properties, inhibiting aggregation of amyloid proteins, stimulating neurogenesis and differentiation action and anti-apoptosis effects [24].A research showed that karanjin in Fuzi possesses learning and memory improvement [25].
According to GO enrichment analysis, BP terms enriched by target genes were mainly concentrated in neurodegeneration-related progressions such as positive regulation of kinase activity, positive regulation of cell death and synaptic signaling.Through KEGG enrichment analysis, we found that the mechanisms of XXD in treatment of AD mainly include Alzheimer's disease pathway, neuroactive ligand-receptor interaction pathway, serotonergic synapse and MAPK signaling pathway.
The Alzheimer's disease pathway with a high degree, containing insulin signaling pathway with hub targets such as PIK3CA and AKT, is thought to be tightly associated with the effect of XXD on AD.Previous research indicated that insulin signaling pathway depicts a strong correlation of energy utilization, mitochondrial function, oxidative stress, synaptic plasticity and cognitive function and the occurrence and development of AD is partially attributed to the impairment of insulin signaling pathway [26].
Some receptor genes control the neuroactive ligand-receptor interaction pathway, which regulates learning and memory ability through enhancing cholinergic function [27], and may participate in cognitive and emotional regulation [28].The abnormal function of GABA A (γ-aminobutyric acid A)receptor, one of the important receptors of neuroactive ligandreceptor interaction pathway, is closely associated with AD [29].
Serotonergic synapse is one of the primary pathological factors in neuropsychiatric symptoms [30,31].Previous researches provided su cient evidence that serotonergic synapse pathway are likely to play a role in memory dysfunction or AD [32,33].
Mitogen-Activated Protein Kinase (MAPK) pathways, including the Extracellular Signal-Regulated Kinase (ERK), c-Jun Nterminal kinase (JNK) and p38 pathways, play the critical roles in regulating cellular functions such as cell apoptosis, synapse plasticity, neural cell survival and neuroin ammation.All the MAPKs are activated in related neurons in patients with AD, suggesting that MAPK pathways are involved in AD pathophysiology and pathogenesis [34].Furthermore, inhibition of these pathways depressed multiple pathological features including Aβ production, tau hyperphosphorylation and synaptic loss, and also made the degeneration of cognitive function in AD animal models [35,36].Among the potential active ingredients, baicalien was demonstrated to alleviate neurotoxicity in Aβ 25−35 -induced PC12 cells via the increased expression of MAPK pathway, suggesting its neuroprotective effect [37].
We performed molecular docking simulation between 8 hub target genes and 9 main active compounds to verify the importance of hub targets.The results showed that all of the target-compound pairs possess high scores of docking a nity, especially, TP53, PIK3CA, AKT1.Previous studies have shown that these genes regulate multiple functions contributing to AD progression.As a downstream effector of DAPK-1 and a Ca 2+ /calmodulin (CaM)-dependent serine/threonine protein kinase pathways ,TP53 plays an important role in cell apoptosis involving transcriptional induction of proapoptotic genes which consequently trigger mitochondrial pathways.TP53 has been suggested as potential therapeutic targets of AD because of its correlation with AD-related neurodegeneration in the brain [38].Akt1, a cell survival kinase, mediate Akt1 signaling through its oxidative modi cation, occurring in the APP/PS1 transgenic mouse model of AD.The loss of Akt1 kinase lead to potential synaptic dysfunction in neurodegeneration including AD [39].PIK3CA, a catalytic subunit of PI3K, plays a critical role in PI3K/AKT signaling pathways and affects the regulation of the pathogenesis of AD via inhibition of tyrosine phosphatases to protective against Aβ toxicity[40].

Conclusion
AD treatments have been an intractable problem up till now, but TCM may be a potential alternative treatment to a certain extent.XXD has been reported to play a protective effect through various mechanisms in AD in the past.For the rst time, this study systematically illustrated the mechanism of action of XXD anti-AD using network pharmacology.
Besides, this potential mechanism is veri ed and visualized through molecular docking.In the present study, 10 main compounds of XXD anti-AD are identi ed.Moreover, TP53, PIK3CA, MAPK1, MAPK3, STAT3, AKT1, etc., are the key genes enriched to pathways (e.g apoptosis, synapse plasticity and neuroin ammation) for XXD to exert its anti-apoptosis and synapse protect effect on AD.Our research not only contributed to deepening the understanding of multi-compoundsmulti-targets-multi-pathways regulation of XXD against AD, but also provided a theorical basis for further experiments to unravel the above possibility.

Declarations Figures
The  for the PPI network through a topological method.In the third image, the bigger size represents higher degree value.

study design owchart Figure 2
Figure 2

Table 2
Information of 10 main active compounds in XXD