A Network Pharmacology Approach to Estimate Potential Targets of the Active Ingredients of Epimedium for Alleviating Mild Cognitive Impairment and Treating Alzheimer's Disease

Background The present study made use of a network pharmacological approach to evaluate the mechanisms and potential targets of the active ingredients of Epimedium for alleviating mild cognitive impairment (MCI) and treating Alzheimer's disease (AD). Methods The active ingredients of Epimedium were acquired from the Traditional Chinese Medicine System Pharmacology database, and potential targets were predicted using the TCMSP target module, SwissTargetPrediction, and PharmMapper database. Target proteins correlating with MCI and AD were downloaded from the GeneCards, DisGeNet, and OMIM databases. The common targets of Epimedium, MCI, and AD were identified using the Jvenn online tool, and a protein-protein interaction (PPI) network was constructed using the String database and Cytoscape. Finally, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the common targets was performed using DAVID, and molecular docking between active ingredients and target genes was modeled using AutoDock Vina. Results A total of 20 active ingredients were analyzed, and 337 compound-related targets were identified for Epimedium. Out of 236 proteins associated with MCI and AD, 54 overlapped with the targets of Epimedium. The top 30 interacting proteins in this set were ranked by topological analysis. GO and KEGG enrichment analysis suggested that the common targets participated in diverse biological processes and pathways, including cell proliferation and apoptosis, inflammatory response, signal transduction, and protein phosphorylation through cancer pathway, MAPK signaling pathway, PI3K-Akt signaling pathway, HIF-1 signaling pathway, sphingolipid signaling pathway, FoxO signaling pathway, and TNF signaling pathway. Molecular docking analysis suggested that the 20 active ingredients could bind to the top 5 protein targets. Conclusions The present study provides theoretical evidence for in-depth analysis of the mechanisms and molecular targets by which Epimedium protects against MCI, AD, and other neurodegenerative diseases and lays the foundation for pragmatic clinical applications and potential new drug development.


Introduction
e incidence of Alzheimer's disease (AD) is increasing significantly with the globally increasing average population age [1]. AD is an irreversible neurodegenerative disease and is the leading cause of dementia [2]. Forty-four million people worldwide currently suffer from dementia, and the number of afflicted people is expected to double by 2050 [3]. AD is thought to occur 20 years or more before symptoms appear, and then progressively deteriorate over time, resulting in the loss of physical function, disability, and death [4]. e pathogenesis of AD is closely related to epigenetic factors, including the environment, vascular disease, inflammation, oxidative stress, or other risk factors, whereas emblematic pathological manifestations include deposition of amyloid-protein (plaque) outside the brain neurons, and tau protein within the brain neurons [5]. ese abnormal changes lead to neuronal injury or destruction and a sequence of complicated symptoms. Early symptoms include memory loss, cognitive impairment, apathy, and depression. As the disease progresses, the patient's memory and cognitive impairment gradually worsen, resulting in communication disorders, poor judgment, muddled thoughts, behavioral changes, aphasia, and difficulty swallowing and walking, which drastically interfere with the patient's daily life and impose a burden on the family [6,7]. e progression and symptoms of AD are continuous and unified, and patients usually transition from normal cognition to mild cognitive impairment (MCI) caused by AD, which further develops into dementia [8]. Epidemiological studies have shown that approximately 10%-15% of people with amnesic-MCI morph into AD; in contrast, only 1% of healthy seniors develop AD [9].
MCI is a clinically recognized neurodegenerative disease that is defined by cognitive changes, and represents a transition stage between healthy aging and dementia [10].
is transition phase is considered to be an optimal treatment window for the protection of brain function, reduction of cognitive impairment, and delay of AD, essentially gaining more time for patients with relatively good cognitive health [11]. Currently, the universal treatments for AD are divided into medication-and non-medication-based therapies. Non-medication therapies can maintain or improve the patient's quality of daily life, but cannot prevent further injury and destruction of neurons. Medication approved by the United States Food and Drug Administration (FDA) for therapeutic use, including rivastigmine, galantamine, donepezil, memantine, and memantine combined with donepezil, can ameliorate the patients' mental state and cognitive symptoms, and slow down memory loss, but the clinical efficacy varies from person to person and the treatment window is limited [12]. MCI is identified as a precursor to AD; however, to date there is no explicit medication for the treatment of MCI and AD. Improving cognitive deficits before the onset of AD can reduce disease severity, and consequently, clinicians often prescribe cholinesterase inhibitors or memantine to alleviate symptoms [13,14]. Given this situation, the development of new drugs is essential for the effective treatment of MCI, AD, and other neurodegenerative diseases.
Modern pharmacological studies have demonstrated that Epimedium is a flavonoid compound with estrogen-like effects and simultaneously possesses anti-aging, antioxidant, anti-apoptotic, and anti-inflammatory properties [15,16]. Estrogen protects and regulates mitochondrial function in the brain and plays a significant role in regulating neuroplasticity and synapse formation, with prominent potential for the treatment of AD [17]. ese salient effects support the use of Epimedium for the prevention and treatment of various neurodegenerative diseases.
Epimedium, also known as Horny Goat Weed, is a perennial herbaceous plant that is widely applied in Asian countries, such as Japan, South Korea, and China [18,19]. Epimedium is principally used for curing impotence, infertility, amnesia, osteoporosis, functional diseases of the elderly, cardiovascular diseases, and primary ovarian insufficiency [20].
e Epimedium extract Icariin has been extensively studied in animal models and is able to reduce Aβ production and amyloid plaque deposition, as well as enhancing learning and memory capacity [21]. Icariin mediates the P13K/Akt pathway to inhibit oxidative stress of tau protein, reduce apoptosis, promote cell viability, and prevent oxidative damage in vitro [22]. Icariin, which is the dominant active ingredient of Epimedium, exerts protective effects on the nervous system. However, less research has been conducted on the other active ingredients of Epimedium.
Traditional Chinese Medicine (TCM) has the peculiarities of multiple ingredients, multiple targets, and multiple pathways, and using experimental methods to explain the potential mechanisms of a TCM can therefore be difficult. Network pharmacology is a new methodological system based on pharmacology and pharmacodynamics that has the potential to enable exploration of the efficacy of TCMs [23]. By analyzing complicated and multi-layered networks of the active ingredients, targets, and specific therapeutic effects of TCM using molecular biology and correlative databases, the synergistic interaction of multiple compounds and their molecular mechanisms can be elucidated. is method can improve our understanding of TCM and provides a new route for the development of new drugs based on active ingredients [24].
In the present study, we made use of network pharmacology to investigate the active ingredients, drug targets, and key pathways of Epimedium against MCI and AD. e prediction of interactions between proteins implicated in MCI and AD and the active ingredients of Epimedium will highlight potential targets for further molecular studies, which may uncover the beneficial components of Epimedium and lead to the development of an effective treatment for neurodegenerative diseases. e experimental workflow is shown in Figure 1.

Identification of Active Ingredients in Epimedium.
We searched for the active ingredients of Epimedium in the Traditional Chinese Medicine System Pharmacology (TCMSP, http://lsp.nwu.edu.cn/browse.php) database. e TCMSP contains comprehensive information about traditional Chinese herbs, including the chemical structure, oral bioavailability, drug-likeness, half-life, Caco-2 intestinal epithelial permeability, blood-brain barrier, and associated drug-target-disease network. We used pharmacokinetic properties, including absorption, distribution, metabolism and excretion (ADME), oral bioavailability (OB) ≥ 30%, and drug-likeness (DL) ≥ 0. 18%, to screen the active ingredients of Epimedium. Compounds with OB ≥ 30% have good absorption, and a DL ≥ 0. 18% is chemically conducive to drug development [25]. We then obtained the SDF structural formulae of active ingredients from the PubChem database (http://pubchem.ncbi.nlm.nih. gov/) for ulterior target prediction [26].

Target Prediction of Active Ingredients.
We used the TCMSP target module, SwissTargetPrediction (http://www. swisstargetprediction.ch/), and PharmMapper (http://www. lilab-ecust.cn/pharmmapper/) databases to predict the potential targets of Epimedium active ingredients. e TCMSP database provides the chemical composition of Chinese herbs, as well as all-sided drug targets of active ingredients, and their relation to diseases [25]. SwissTargetPrediction predicts the targets of active ingredients by comparing their structures to the two-and three-dimensional structures of known compounds [27]. PharmMapper is a freely accessible server that uses pharmacodynamic gene mapping methods to recognize candidate targets for drugs, natural products, or other newly found compounds with unidentified binding targets [28]. e resulting target protein names were ID mapped to their gene names using the UniProt database (http://www.UniProt.org/) [29].

Target Prediction of Disease.
GeneCards (https://www. genecards.org/), DisGeNet (http://www.disgenet.org/), and OMIM (https://omim.org/) databases are regularly updated online tools containing information about human genes and genetic diseases. GeneCards includes proteomic, genetic, genomic, transcriptomic, clinical, and other information, which can be applied to predict human gene functions [30]. DisGeNet conveniently integrates the existing databases of mutation sites and human disease-related genes with corresponding information from a mass of literature to construct a unified gene or mutation site disease-related database [31]. OMIM is an authoritative database of human genes and genetic diseases [32]. In the present study, "mild cognitive impairment and Alzheimer's disease" were used as the keywords to collect known therapeutic targets from these three databases for the species "Homo sapiens."

Intersection between Active Ingredients and Disease
Targets. We used the Jvenn online tool to generate a Venn diagram and identify common targets for Epimedium, MCI, and AD. Common targets were submitted to String (https:// string-db.org/) to generate a protein-protein interaction (PPT) network [33]. e obtained network was visualized using Cytoscape v3. 7. 2, and topological analysis was performed using the network analysis module.

GO/KEGG Pathway Enrichment
Analysis. GO is a constantly updated and internationally applicable gene function classification system that can be used to analyze functional enrichment of genes and proteins. KEGG pathway enrichment analysis provided a deeper explanation of different gene and protein functions by identifying enriched biochemical pathways in datasets. To gain insight into the gene functions and signaling pathways of the Epimedium targets, we used DAVID (https://david.ncifcrf.gov/), which is an online biological knowledge base and analysis tool, to identify enriched cellular components (CC), biological processes (BP), molecular function (MF), and KEGG pathways [34].

Validation of Putative Targets by Molecular Docking
Analysis. Molecular docking is a burgeoning approach to explore drug mechanisms, elevate the pharmacological effect of drugs to the molecular level, and provide a theoretical basis for treating diseases. In the present study, the top five protein targets of Epimedium were selected based on PPI network analysis and included serum albumin (ALB), RACalpha serine/threonine-protein kinase (AKT1), interleukin-6 (IL6), caspase-3 (CASP3), and tumor necrosis factor (TNF), to perform molecular docking with the 20 active ingredients. We retrieved the SDFs of the 20 compounds from Pubchem  [26]. Next, we used PDB (http://www.rcsb.org/) to obtain the structural data of binding sites on AKT1 (PDB ID: 4EKL), ALB (PDB ID: 6JE7), CASP3 (PDB ID: 2J32), IL6 (PDB ID: 1ALU), and TNF (PDB ID: 5MU8) [35]. We used PyMOL 2. 3. 4 software to perform dehydration and ligand removal on the binding sites. e five proteins were modified using AutoDock Tools for hydrogenation and charge balancing. We used the "grid option" tool to set the grid point spacing to 1, adjust the volume of binding pocket so that the pre-docked molecules can rotate within the box in their most extended state, set the center of pocket as the center of binding site, and convert the protein and ligand small molecules into pdbqt format. Finally, we used AutoDock Vina 1. 1. 2 to perform molecular docking of the five proteins with the 20 ligand small molecules. We selected the receptors and ligands with strong binding free energy and generated three-dimensional graphs in PyMOL to analyze their interactions.

Active Ingredients of Epimedium.
A total of 130 chemical ingredients of Epimedium were obtained from the TCMSP database (Supplementary Table 1). After screening these and confirming their SDF structural formulae in Pubchem, the list was narrowed to 20 active ingredients of Epimedium. e characteristics of the 20 active ingredients are shown in Table 1.

Potential Targets of Active Ingredients.
We searched the TCMSP database using the filtered list of Epimedium active ingredients and identified 204 compound-related targets. In addition, the SDF structural formulae of the active ingredients were used to search the SwissTarget and Pharm-Mapper databases, and a list of relevant targets was acquired. e targets were screened to include only those with a probability value ≥0.5 and a Normfit ≥0.8. Using these cutoffs, 108 and 84 compound-related targets were obtained from SwissTarget and PharmMapper, respectively (Supplementary Table 2). After removal of the repeated targets, a total of 337 compound-related targets were identified ( Figure 2).

Potential Targets of Disease.
e GeneCards, DisGeNet, and OMIM databases were used to identify reviewed therapeutic targets in MCI and AD. e scores in GeneCards and DisGeNet were used to define a set of high-confidence interactions between these diseases and potential targets, with the relevance score ≥20 in GeneCards and score-gda ≥0.1 in DisGeNet. As a result, 546 AD-related targets and 441 MCI-related targets were acquired from GeneCards, and 303 AD-related targets and 5 MCI-related targets were obtained from DisGeNet. In addition, 523 AD-related targets and 85 MCI-related targets were obtained from OMIM. After removing the duplicate targets, 619 potential targets remained for MCI and 1166 for AD, and 236 common targets for both diseases were identified (Supplementary Table 3).

Intersection of Active Ingredients and Disease Targets.
A total of 54 targets overlapped for the Epimedium active ingredients, MCI, and AD. e overlap between the filtered targets was plotted using a Venn diagram (Figure 3). We used String to analyze the interactions between the 54 shared targets.
e string analysis contained 54 nodes, 610 link edges, and 195 expected edges at the medium confidence. e average node degree was 22.6, the average local clustering coefficient was 0.737, and the PPI enrichment P value was <0.001. Cytoscape software was used to construct the PPI network. Different colors and circle sizes represent different biological information, with larger circles and darker colors (from orange to red) indicating stronger interactions (Figure 4(a)). In addition, we performed a topological analysis of the 54 common targets using a Cytoscape analysis module. e top 30 targets were ranked according to their degree values, represented by a bar graph (Figure 4(b)). Detailed topological analysis results and parameters are presented in Table 2.

GO/KEGG Function and Pathway Enrichment Analysis.
To elaborate on the pivotal mechanisms underpinning the neuroprotective activity of Epimedium against MCI and AD, GO enrichment analysis was performed on 54 common targets using DAVID. e GO terms with P values <0.01 were considered to be enriched (Supplementary Table 4). In the BP category, regulation of aging was enriched, associated with GO terms related to transcriptional regulation, regulation of cell proliferation and apoptosis, regulation of gene expression, inflammation, signal transduction, and protein phosphorylation ( Figure 5(a)). In the CC category, target proteins were primarily distributed in the plasma membrane, and some in the cytosol and mitochondria ( Figure 5(b)). In the MF category, target proteins were mainly involved in protein binding, with some involved in beta-amyloid binding ( Figure 5(c)).
To identify pathways that are enriched, and therefore likely to be involved in the neuroprotective effects of Epimedium, we performed KEGG enrichment analysis of the 54 potential targets using DAVID, with a P value cutoff of <0.01. We selected the top 20 enriched pathways, which were represented as bar and dot plots ( Figure 6).
In addition, we identified targets that were shared between multiple pathways, which are listed in Supplementary  Table 5. We constructed a target-pathway network, as shown in Figure 7(a). To understand the mutual effects of Epimedium, target, disease, and pathway, seven typical pathways were selected to set up an integral network, including the cancer pathway, mitogen-activated protein kinase (MAPK) signaling pathway, PI3K-Akt signaling pathway, HIF-1 signaling pathway, sphingolipid signaling pathway, FoxO signaling pathway, and TNF signaling pathway. us, an intricate network was formed to describe the effects of Epimedium active ingredients and potential targets of MCI and AD (Figure 7(b)).    Table 3. AutoDock Vina outputs the results in the form of an affinity energy, which is assessed by calculating the spatial effect, repulsion, hydrogen bonding, hydrophobic interactions, and molecular flexibility of the receptor-ligand complexes. e affinity energy is a prerequisite indicator to determine whether the ligand small molecule can effectively bind to the receptor, with lower energy values suggesting better binding capacity for the receptor and ligand. According to this scoring system, AKT1 and Yinyanghuo A, ALB and Yinyanghuo C, CASP3 and Yinyanghuo E, IL6 and Yinyanghuo C, and TNF and Yinyanghuo A were selected for further analysis (Figure 8).

Discussion
In terms of TCM theory, MCI and AD fall under the category of amnesia and dementia. e same treatment for different diseases (Yi Bing Tong Zhi) is a fundamental treatment principle of TCM, which has accumulated vast experience in treating dementia over hundreds of years of clinical practice. Epimedium is one of the cross-sectional herbs that is prescribed as part of Chinese medicine, owing to its anti-apoptotic and anti-oxidative properties, and it is frequently applied to cure central nervous system diseases like forgetfulness, cognitive impairment, and dementia. e mechanism by which Epimedium exerts its therapeutic function is still poorly defined due to the complexity of its constituents. To enable better understanding of the mechanisms behind the protective effects of Epimedium, in the present study we made use of network pharmacology to explore multiple molecular perspectives, and to provide a theoretical basis for the treatment of neurodegenerative diseases.
We identified 20 chemical compounds as the potential active ingredients of Epimedium, many of which have been widely reported to possess inhibitory effects on neurological aging. For instance, icariin was used in Tg2576 mice (containing the Swedish AD mutant gene), with a standard dose of 60 mg/kg. After three months of administration, the hippocampal nerve regeneration and working memory ability of the mice had improved [36]. Icariin also reduced the expression of amyloid β1-42 and prevented the expression of amyloid precursor protein and β-site APP cleaving enzyme 1 in AD animal models [37]. Another active ingredient, kaempferol, inhibited the oxidative stress response and neuroinflammation in transgenic AD drosophila and an ovariectomized rat model, as well as reversing the loss of climbing ability and improving memory functions [38]. Quercetin is a flavonoid compound that protects neurons from oxidative and excitotoxic effects by modulating cell death mechanisms [39]. It also reduces Aβ polymerization, APP, and BACE1 expression and reduces oxidative stress [40]. Our molecular docking results predicted probable interactions between several active ingredients of Epimedium and protein targets that are related to MCI and AD. ese predicted interactions imply that the active ingredients of Epimedium interact with proteins that have known functions in neurological diseases, which may explain its therapeutic effects. Based on these results, Epimedium and its active ingredients should be studied further, with a view to developing novel drugs for the treatment and prevention of MCI, AD, and other neurodegenerative diseases.  ACHE  IGF2  F2  PTEN  ICAM1  IFNG  SOD1  HMOX1  TGFB1  MMP2  ESR1  CRP  SERPINE1  PPARG  MPO  IL1B  CCL2  MMP9  IL10  NOS3  APP  EGF  VEGFA  MAPK1  TP53  TNF  CASP3  IL6  AKT1  ALB   0  1 0  2 0  3 0  4 0  5 0   20  21  23  24  24  25  27  27  28  28  28  28  29  30  30  31  31  32  32  33  34  35  36  37  39  39  In the present study, KEGG pathway and GO enrichment analysis was carried out for 54 common targets of Epimedium, MCI, and AD. Based on these analyses, we identified a subset of proteins, AKT1, MAPK1, CASP3, IL6, TNF, BCL2, TP53 and NOS3, which exert their effects via multiple different biochemical pathways. Such proteins are   Evidence-Based Complementary and Alternative Medicine likely to be "master regulators" and could be major mediators of neurodegenerative diseases. For example, AKT1 regulates cell survival, proliferation, and growth and metabolism mediated by serine and threonine phosphorylation of downstream substrates [41,42]. AKT positively regulates CREB1 protein activity and plays a crucial role in the regulation of NF-κB-dependent gene transcription. CREB1 phosphorylation induces the binding of auxiliary proteins that are also necessary for the transcription of prosurvival genes, such as BCL2 and MCL1 [43]. e MAPK/  Evidence-Based Complementary and Alternative Medicine 9 ERK cascade accommodates biological functions such as cell survival, growth, adhesion, and differentiation and phosphorylates transcription factors and regulates meiosis and mitosis in differentiated cells [44]. BCL2 is a pivotal apoptosis regulator in the apoptotic signaling pathway. BCL2 inhibits apoptosis in the outer mitochondrial membrane and inhibits the reduction of glutathione and redox state, controls membrane potential, and prevents the release of pro-apoptotic proteins such as cytochrome C and apoptosis-inducing factor AIF to suppress apoptosis [45,46]. Caspase-3 (CASP3) is an important pro-apoptotic protease that effects apoptosis mediators such as BCL2 and Bax. BCL2 is a direct substrate upstream of CASP3. BCL2 and CASP3 are both interrelated and mutually constrained and are jointly involved in the process of neuronal apoptosis [47,48]. Interleukin-6 (IL6) is a cellular inflammatory factor secreted by monocytes and is involved in neuroinflammatory processes along with IL10 and IL1β. IL6 exacerbates the expression of inflammatory factor IL10 in downstream signaling pathways and promotes the oxidative stress response, leading to apoptosis or atrophy of neuronal cells. Moreover, IL6 promotes β-amyloid deposition and affects the integrity of neuronal axons or dendritic membranes [49,50]. e MAPK signaling pathway is one of the essential signaling systems in living organisms and has been broadly   Evidence-Based Complementary and Alternative Medicine 11 studied in plants, insects, and mammals. MAPK signaling participates in a wide variety of physiopathological processes, such as cell growth, development, differentiation, and apoptosis, and is a regulator of inflammation and the stress response [51]. Neuroinflammation is the subject of recent attention as a latent element in the pathology of AD. Massive β-amyloid deposition and hyperphosphorylation of tau protein bring about cell autophagy, function loss, and apoptosis, and these pathological reactions are compactly associated with MAPK activation and signal conduction [52]. Pinocembrin, a natural flavonoid compound, inhibits RAGE expression both in vivo and in vitro, preventing P38MAPK and MK2 activation and downregulating the NF-κB and SAPK/JNK-c-Jun signaling pathways to reduce cerebral cortex neurodegeneration and ameliorate cognitive function [53,54]. Drugs that directly target the MAPK signaling pathway may prove to be an effective strategy for the treatment of AD. e PI3K-Akt signaling pathway participates in the processes of cell reproduction, differentiation, and apoptosis and has many effectors downstream. Akt is the central link in the PI3K-Akt signaling pathway, and it translocates from the cell membrane to the nucleus, where it affects many transcriptional regulators and factors after activation, such as NF-κB, CREB, mTOR, GSK-3β, caspase, and the BCL2 family [55]. e PI3K-Akt signaling pathway is involved in the energy metabolism of the aging process, as well as numerous physiological-pathological processes of age-related diseases. Polyphenolic anthocyanins are a natural compound with antioxidant and neuroprotective features that reduce the production of Aβ and ROS in in vitro AD models and APP/PS1 transgenic mouse models. Polyphenolic anthocyanins reduce oxidative stress to exert their neuroprotective effects, improve memory dysfunction, and delay neurodegenerative diseases through the regulation of the PI3K/Akt/Nrf2 pathway [56]. e HIF-1 signaling pathway is a physiological response to hypoxia that is mediated by hypoxia-inducible factors (HIFs) and activated or initiated in mammals to adapt to hypoxic environments, also allowing for the accommodation of hypoxia and increased cellular energy metabolism [57]. HIFs are involved in many disease processes, such as inflammation-induced cancer exacerbation, angiogenesis, malignant cell proliferation, and anti-apoptosis [58]. During aging, the supply of oxygen to the brain gradually decreases, and metabolism slows affecting APP expression and transport, leading to an imbalance between the production and clearance of Aβ and increasing the risk of AD [59]. Animal experiments suggest that HIF-1α upregulates BACE1 expression and catalyzes β-cleavage of APP, while BACE1 expression in the hippocampus and cerebral cortex of mice is reduced by the lack of HIF-1α. e above examples demonstrate that the HIF-1 signaling pathway is one of the most important pathways for regulating APP amyloidosis [60]. e sphingolipid signaling pathway plays crucial roles in aging, neurodegenerative diseases, and immune regulation [61]. Sphingosine, sphingosine-1-phosphate, ceramide, and ceramide-1-phosphate are functionally similar and have modulatory effects on proliferation, differentiation, maturation, aging, death, stress response, and inflammation in central nervous cells and are currently the strongest known biologically active sphingomyelin compounds [62]. For example, Aβ25-35 was applied to SH-SY5Y neuronal cells to create an AD model to induce SPHK1 inhibition and neuroamide accumulation, but the upregulation of sphingosine kinase prevented cell death and exerted a protective function [63]. Sphingolipids are pro-secretory molecules that regulate APP secretion and directly protect the nervous system of AD patients, which have excellent prospects for clinical application [64].
Forkhead box class O (FoxO) is a mammalian forkhead transcription factor that interferes with differentiated cell survival and stem cell proliferation and is a potential target for the control of neurological diseases [65].
e FoxO transcription factor interacts with numerous signaling pathways, such as Akt, SIRT1, IKK, and growth factor [66,67]. FoxO proteins are selectively expressed throughout the body and exert different biological functions in the nervous system [68]. e complex interactions between FoxO proteins and their signal transduction pathways are capable of promoting cellular autophagy, exerting antioxidant effects under oxidative stress, improving cell survival, and scavenging toxic proteins [69]. FoxO proteins hold promise as a new strategy for the treatment of neurodegenerative diseases.
Tumor necrosis factor (TNF) is a cytokine ligand that participates in systemic inflammation. Both TNF and its type 1 receptor (TNFRI) are involved in AD-associated cerebral neuroinflammation and regulate amyloid formation through secreted enzymes [70]. Aging, injury, or degenerative neurological conditions contribute to the production and release of TNFα, inducing destructive or toxic biological effects when it binds to specific receptors [71]. Developing drugs to modulate the production of TNFα, such as NF-κB, P38 MAPK, and TNF receptor antagonist inhibitors, may be potential candidates for the treatment of cognitive impairment, AD, and other neurodegenerative diseases.
As mentioned above, the targets identified in our study are consistent with previous attempts. Epimedium active ingredients have confirmed interacting partners with proteins that are important regulators of health and disease; these interactions could point future research towards novel therapeutic strategies for treating neurodegenerative diseases.

Conclusions
In summary, we made use of network pharmacology to reveal the interactions between the possible active ingredients of Epimedium and known therapeutic genes associated with MCI and AD. rough further screening, we verified the potential targets of Epimedium at the molecular level and predicted possible major regulatory pathways that are involved in its neuroprotective effects. e pathogenesis of MCI and AD is complicated, forming a multi-path and multi-target action mode with interlaced signal pathways. e active ingredients of Epimedium against MCI and AD may function by modulating these pathways, demonstrating the multi-component and multi-target capabilities of Chinese herbs. Our findings provide a hypothetical basis for further practical molecular studies of the interactions between Epimedium and these proteins. e combination of Chinese medicine and modern analytical methods will allow in-depth studies of Epimedium's efficacy at treating neurodegenerative diseases. Our findings could lay the foundation for pragmatic clinical application of Epimedium, and new drug development.

Data Availability
e data used in the study are available upon request to the corresponding author.

Conflicts of Interest
All authors have no conflicts of interest to declare.

Authors' Contributions
XWG, SNL, CC, YJW, and LWX proposed the concept and design of the study. XWG and SNL contributed to collecting and analyzing data. CC and YJW contributed to checking the accuracy of data. XWG and CC drafted the main manuscript text, while SNL, YJW, and LWX contributed to editing and providing critical revision. All the authors reviewed and approved the final manuscript.