Abstract

   Evodia rutaecarpa (Evodia) is a Chinese herbal medicine with analgesic
   and anti-neurodegenerative properties. However, whether Evodia
   compounds can be applied for the comorbid pain of Alzheimer's disease
   (AD) and the underlying mechanisms remain unclear. Herein, 137 common
   targets of Evodia between AD and pain were predicted from drug and
   disease target databases. Subsequently, protein-protein interaction
   (PPI) network, protein function module construction, and bioinformatics
   analyses were used to analyze the potential relationship among targets,
   pathways, and diseases. Evodia could simultaneously treat AD comorbid
   pain through multi-target, multi-component, and multi-pathway
   mechanisms, and inflammation was an important common phenotype of AD
   and pain. The relationship between important transcription factors such
   as RELA, NF-κB1, SP1, STAT3, and JUN on IL-17, TNF, and MAPK signaling
   pathways might be potential mechanisms of Evodia. Additionally, 10
   candidate compounds were predicted, and evodiamine might be the
   effective active ingredient of Evodia in treating AD or pain. In
   summary, this study provided a reference for subsequent research and a
   novel understanding and direction for the clinical use of evodiamine to
   treat AD patients with comorbid pain.

   Keywords: Evodia rutaecarpa, Alzheimer's disease, Pain, Evodiamine,
   Network pharmacology

1. Introduction

   Alzheimer's disease (AD) is the most common neurodegenerative disease
   in the elderly. In 2019, about 13.14 million people were diagnosed with
   AD in China, accounting for 20% of the total number of AD cases
   worldwide [[31]1]. Current investigations found that almost 50% of AD
   patients suffer from pain [[32][2], [33][3], [34][4]], and more than
   75% experience pain during AD progression [[35]5]. However, pain
   management in AD patients remains challenging due to the complex link
   between AD and pain [[36]6,[37]7]. Mild and moderate AD patients can
   still have pain. Nevertheless, patients lose almost all ability to
   express the pain in late AD stages, which makes the clinical diagnosis
   of pain more difficult [[38]8]. Treatment is another clinical issue for
   AD patients besides evaluation and diagnosis. Non-steroidal
   anti-inflammatory drugs (NSAIDs), such as acetaminophen, are the
   first-line medicine for AD-related pain [[39]9,[40]10]. Except for the
   compliance and overdose risk of AD patients, NSAIDs might not present
   enough efficacy for chronic pain or neuralgia [[41]11]. Other
   analgesics, such as opioids, still lack reliable clinical evidence and
   might increase adverse events when combined with psychotropic drugs to
   treat AD [[42]12,[43]13]. Meanwhile, the incidence of pain
   significantly increases with age, and pain level is positively
   correlated with the severity of cognitive impairment [[44]14].
   Therefore, the positive vicious cycle in AD patients with comorbid pain
   must be addressed.

   Neuroinflammation is an important bridge linking pain and cognitive
   decline by regulating microglia function (e.g., pro-inflammatory
   factors, phagocytosis), synaptogenesis, and dysfunction in several
   brain areas (e.g., locus coeruleus, frontal cortex) [[45]15,[46]16].
   Evodia rutaecarpa (Evodia) has been used to treat pain or AD patients
   in traditional Chinese medicine (TCM) [[47]17,[48]18]. Evodia and its
   compounds (e.g., evodiamine) can be anti-inflammation and
   anti-neurodegeneration candidate agents [[49]18,[50]19]. In a rat model
   of neuropathic pain, evodiamine administration significantly inhibited
   the elevation of inflammatory cytokines, ameliorated mitochondrial
   dysfunction, and maintained oxidative stress levels [[51]20].
   Rutaecarpine, an active compound of Evodia, can attenuate the
   pathological changes of AD-like mice, alleviate cognitive and memory
   impairments and enhance the plasticity of neural protrusions [[52]21].
   Despite the TCM emphasis on the comprehensive method to regulate
   homeostasis through the “sovereign (Jun), minister (Chen), assistant
   (Zuo) and guide (Shi)” herb medicines in the formula, further research
   is still necessary to determine the active compounds for reducing
   toxicity and side effects [[53]22]. In recent years, network
   pharmacology has been used to explore potential targets among several
   diseases through a comprehensive network analysis to reveal the
   relationship among known TCM herbal targets, genes, and pathways from
   multiple dimensions [[54]23,[55]24]. Thus, network pharmacology is a
   convenient and feasible method to undermine potential active compounds
   and common targets of Evodia in AD patients with comorbid pain.

   Therefore, we hypothesized that Evodia or its compounds might be used
   to develop treatments for AD patients with comorbid pain. Thus, we used
   network pharmacology to predict the common targets of Evodia for AD and
   pain. Then, we conducted bioinformatic analyses to determine the
   potential targets and candidate compounds, which supported our
   hypothesis and might facilitate future studies.

2. Methods

2.1. Collection of potential drug and disease targets

   Different disease databases can be used for different research
   purposes, including studying the molecular basis of human diseases and
   their comorbidities and analyzing the properties of disease genes.
   First, we collected targets from multiple databases: Human Phenotype
   Ontology (HPO, [56]http://www.human-phenotype-ontology.org) [[57]25],
   DisGeNET ([58]https://www.disgenet.org) [[59]26], National Center for
   Biotechnology Information (NCBI,
   [60]https://www.ncbi.nlm.nih.gov/gene), and Pharmacogenomics
   Knowledgebase (PharmGKB, [61]https://www.pharmgkb.org) [[62]27].
   Subsequently, “Alzheimer's disease” and “pain” were used as search
   terms for AD and pain-related genes. Search results were collected from
   each database, duplicate values were removed, and the final results
   were the pathogenic targets of AD and pain.

   The Traditional Chinese Medicine Systems Pharmacology and Analysis
   Platform (TCMSP, [63]https://old.tcmsp-e.com/tcmsp.php) [[64]28]
   contains the chemical composition, the action target, and related
   pharmacokinetic properties of natural compounds. We used “wuzhuyu"
   (Evodia pinyin name) as the keyword to obtain relevant drug targets and
   selected “Homo sapiens” as the species in the UniProt
   ([65]https://www.uniprot.org) [[66]29] database to standardize the
   retrieved target names into gene names. Next, we used Venny 2.1.0
   ([67]http://bioinfogp.cnb.csic.es/tools/Venny/index.html) to analyze
   the common targets of AD and pain, as well as the common targets of
   Evodia in AD and pain treatment. Then, gene names and
   inter-relationships of drug-treated diseases were entered into
   Cytoscape 3.9.1 ([68]https://cytoscape.org) [[69]30] to obtain the
   visual action network of Evodia in treating AD and pain.

2.2. Protein-protein interaction (PPI) network construction and network
topology analysis

   The PPI network comprises individual protein interactions,
   participating in signal transmission, gene expression regulation,
   energy, substance metabolism, cell cycle regulation, and other life
   processes. The STRING 11.5 platform ([70]https://cn.string-db.org
   [[71]31]) was used to determine protein-protein interactions between
   common targets. The species was selected as Homo sapiens, with the
   “highest confidence” set to 0.9 and the “Clustering Options” to
   “k-means clustering,” and the “number of clusters” to 3. Subsequently,
   the results of the PPI network analysis were imported into Cytoscape
   3.9.1 software in the TSV format, and the Network Analyzer [[72]32]
   plug-in was used for topology analysis to obtain the node degree and
   edge betweenness of each target gene to highlight their importance and
   eliminate targets with node degree of 0.

2.3. Enrichment analysis of the intersection of drug-disease targets

   To understand the biological significance behind the targets, we used
   the Database for Annotation, Visualization, and Integrated Discovery
   (DAVID, [73]https://david.ncifcrf.gov) [[74]33,[75]34] to perform Gene
   Ontology (GO) functional analysis of the intersection of drug-disease
   targets. GO consists of three ontologies: process (BP), cellular
   component (CC), and molecular function (MF). We also performed The
   Kyoto Encyclopedia of Genes and Genomes (KEGG, [76]https://www.kegg.jp)
   [[77]35] pathway enrichment analysis of these targets at the molecular
   level using the online bioinformatics tools
   ([78]http://www.bioinformatics.com.cn). Then, we imported the
   intersection of these drug-disease targets into Metascape
   ([79]http://Metascape.org) [[80]36], and GO function and KEGG pathway
   enrichment analyses were performed again. The top 10 GO and top 20 KEGG
   pathway results were compared with DAVID and KEGG databases,
   respectively. At this time, the relationship between the top 20 targets
   with node degree in the PPI network and the intersection of KEGG
   pathways was input into Cytoscape 3.9.1 to obtain the network analysis
   diagram of the significant targets acting on the significant pathways.

2.4. Identification of the functional clusters of 107 intersection targets

   Metascape is a platform that can cluster similar proteins and construct
   functional modules using the Molecular Complex Detection (MCODE)
   algorithm [[81]37]. With 107 targets imported, MCODE analysis was used
   to identify protein relationships between molecular complexes and
   cluster them into a gene cluster where each target has the same or a
   similar gene function. Additionally, transcription factors that
   regulate other targets were further predicted, and information was
   obtained on target interactions.

2.5. Creation of the network diagram

   Next, we used a Bioinformatics Analysis Tool for the Molecular
   Mechanism of Traditional Chinese Medicine (BATMAN-TCM,
   [82]http://bionet.ncpsb.org/batman-tcm/) [[83]38] to analyze Evodia. We
   used “wuzhuyu” as the search term and set the “Score cutoff” to 20, the
   “Adjusted P-value” to 0.05, and the species to “Homo sapiens.” The
   multi-component, multi-target, multi-pathway, and multi-disease network
   diagram was obtained for Evodia. The screening condition for disease
   enrichment results was set to p < 0.05, and the top ten results were
   ranked from large to small by target number. After the pain-related
   components of AD were obtained, they were entered into the TCMSP
   database for bioactivity prediction. The screening threshold conditions
   were: oral bioavailability (OB) ≥ 30%, similarity (DL) ≥ 0.18, and
   blood-brain barrier (BBB) ≥ 0.3. The relationship between eligible
   active ingredients and target diseases was plotted. Finally, GO and
   KEGG analyses were performed again for the targets of the component of
   Evodia.

3. Result

3.1. Common targets of Evodia for AD and pain

   After retrieving and deleting duplicate genes from multiple disease
   databases, a total of 3507 AD-related targets were obtained, among
   which 200 were related to the drug target of Evodia. There were 2277
   targets associated with pain, 157 of which were associated with Evodia.
   According to the Venn diagram intersection, 1101 common targets of AD
   and pain ([84]Fig. 1A) and 137 overlapping targets of Evodia and
   AD-pain ([85]Fig. 1B) were detected. Then, the visualization network of
   “AD targets”- “common targets of AD & Pain”- “Pain targets” indicated
   the target genes involved in AD and Pain ([86]Fig. 1C).

Fig. 1.

   [87]Fig. 1
   [88]Open in a new tab

   The plot of intersection and interaction between drug and disease
   targets. A. Purple represents targets of AD, yellow represents targets
   of pain, and yellow-green represents targets of AD comorbid pain. B.
   Purple represents targets associated with Evodia in AD, yellow
   represents targets associated with Evodia in pain, and yellow-green
   represents targets associated with Evodia in comorbidity. C. Cytoscape
   3.9.1 clearly showed the relationship of the targets between Evodia and
   AD comorbid pain. (For interpretation of the references to color in