Abstract

Background

   In this study, we used the network pharmacology approach to explore the
   potential disease targets of the Eucommia ulmoides Oliver
   (EUO)-Tribulus terrestris L. (TT) drug pair in the treatment of
   hypertension-associated neurovascular lesions and IS via the
   ferroptosis pathway.

Methods

   We used the Traditional Chinese Medicine Systems Pharmacology Database
   and Analysis Platform to search for the key active compounds and
   targets of the drug pair. Based on the GeneCards database, the relevant
   targets for the drug pair were obtained. Then, we performed the
   molecular docking of the screened core active ingredients and proteins
   using the DAVID database and the R AutoDock Vina software. Based on the
   [31]GSE22255 dataset, these screened target proteins were used to build
   random forest (RF) and support vector machine (SVM) models. Finally, a
   new IS nomogram prediction model was constructed and evaluated.

Results

   There were 36 active compounds in the EUO-TT drug pair. CHRM1, NR3C1,
   ADRB2, and OPRD1 proteins of the neuroactive ligand-receptor
   interaction pathway interacted with the proteins related to the
   ferroptosis pathway. Molecular docking experiments identified 12 active
   ingredients of the drug pair that may tightly bind to those target
   proteins. We constructed a visual IS nomogram prediction model using
   four genes (CHRM1, NR3C1, ADRB2, and OPRD1). The calibration curve,
   DCA, and clinical impact curves all indicated that the nomogram model
   is clinically applicable and diagnostically capable. CHRM1, NR3C1,
   ADRB2, and OPRD1, the target genes of the four effective components of
   the EUO-TT drug pair, were considered as risk markers for IS.

Conclusions

   The active ingredients of EUO-TT drug pair may act on proteins
   associated with the neuroactive ligand-receptor interaction pathway to
   regulate ferroptosis in vascular neurons cells, ultimately affecting
   the onset and progression of hypertension.

   Keywords: stroke, hypertension, Eucommia ulmoides Oliver, Tribulus
   terrestris L, ferroptosis, network pharmacology Taubert D

Introduction

   Ischemic stroke (IS), one of the neurovascular disorders, is a common
   complication of hypertension, the second most common cause of death
   worldwide, and has become a major public health problem. Stroke affects
   approximately 15 million people worldwide each year, resulting in 5
   million deaths and 5 million disabilities ([32]1). A study based on the
   Global Burden of Disease, Injuries and Risk Factors (GDB2010) showed
   that stroke mortality has increased by 26% since 1990 ([33]2). A
   prospective observational study of 61 populations worldwide (~1 million
   people aged 40–89 years) found a positive association between stroke
   risk and blood pressure levels ([34]3). The earliest appearances are
   changes in arterial elastic function and an increase in pressure wave
   conduction velocity. The continuous increase in blood pressure
   progressively damages the integrity of the vascular endothelium
   resulting in structural changes to the vessel wall. The higher the
   blood pressure level, the more severe the damage. There is a strong
   causal relationship between blood pressure levels and stroke morbidity
   and mortality. Although the effect of western medicine in the treatment
   of hypertension is precise, it is mainly based on the control of blood
   pressure. This treatment regimen has several disadvantages, such as
   high dependence on medicines, numerous side effects, and heavy economic
   burden due to long-term use. Chinese medicine has unique advantages in
   the prevention and treatment of hypertension and neurovascular related
   diseases since they target multiple pathways and are less toxic.

   EUO is a living fossil plant. Recent studies have found more than a
   hundred compounds that can be extracted from EUO. Monomeric compounds
   and extracts of EUO have a wide range of pharmacological effects in the
   treatment of several diseases, including hypertension, hyperlipidemia,
   diabetes, and osteoporosis ([35]4, [36]5). The active ingredients of
   EUO that are anti-hypertensive are divided into four major groups:
   lignans, phenylpropanoids, flavonoids, and cyclic enol ether terpenes.
   Among these, lignans are the most studied class of chemical compounds
   which may regulate nitric oxide levels, renin-angiotensin system, and
   direct arterial diastole ([37]6). The lignans in EUO exert
   anti-hypertensive effects in spontaneously hypertensive rats (SHR),
   presumably by inhibiting aldose reductase (AR), inhibiting the Ang
   II-induced inflammation and oxidative stress signaling pathways, and
   preventing cardiovascular remodeling ([38]7, [39]8). Moreover, lignans
   can protect against renal damage caused by hypertension ([40]9). Other
   components of EUO, like phenylpropanoid caffeic acid, lowers the blood
   pressure by inducing nitric oxide synthase and promoting nitric oxide
   synthesis ([41]10, [42]11). Further, quercetin, a flavonoid, has the
   effect of endothelium-dependent vasodilatation ([43]11, [44]12). Thus,
   EUO contains a variety of anti-hypertensive components that may protect
   against hypertension-induced ischemic stroke.

   Another Chinese medicine, TT, has been shown by modern pharmacological
   studies to be anti-hypertensive, diuretic, hypolipidemic, and
   anti-atherosclerotic. It also reduces left ventricular remodeling and
   improves cardiac function in the early post-myocardial infarction
   period ([45]13–[46]15). Phillips et al. studied the effects of the
   methanolic and aqueous extracts of TT on blood pressure and perfused
   mesenteric vascular bed in rats, and showed that TT resulted in a
   dose-related reduction in blood pressure of SHR ([47]13). Further,
   Sharifi et al. found that TT lowers the blood pressure in hypertensive
   rat models by inhibiting the angiotensin-converting enzyme ([48]16).
   Active components isolated from TT also exert anti-inflammatory effects
   by inhibiting inflammatory mediators and cytokines, including IL-6,
   IL-10, and TNF-α ([49]17). Inflammation is one of the several
   pathogeneses of hypertension, and inflammatory factors such as IL-6 are
   involved in the development and progression of hypertension
   ([50]18–[51]20). Thus, the components of Tribulus terrestris L have a
   potential key role in the prevention and treatment of hypertension
   through several pathways.

   Recent studies have found that treatment of ferroptosis may be an
   important means of preventing and improving the prognosis of patients
   with IS ([52]21, [53]22). Ferroptosis is a novel form of cell death
   discovered by Dolma et al. ([54]23). The main feature of ferroptosis is
   the iron ion-dependent accumulation of large amounts of lipid peroxides
   during cell death ([55]24). P53, System Xc-, glutathione peroxidase 4
   (GPX4), and Fe are the core components of the ferroptosis pathway.
   Moreover, ferroptosis-related genes are mainly involved in the
   oxidative stress response ([56]25). Oxidative stress and associated
   oxidative damage may be the main cause of vascular injury and may be
   involved in hypertension, where increased utilization and/or decreased
   inactivation of reactive oxygen species (ROS) is key to
   vasoconstriction dysfunction ([57]26–[58]28). GPX4 is a highly active
   antioxidant enzyme in the basal body that scavenges ROS and prevents
   oxidative stress. Hence, its decline, either by using a GPX4 inhibitor
   or knockdown of the GPX4 gene, is associated with the development of
   pre-eclampsia and a significant increase in intracellular ROS and lipid
   peroxidation, thereby inducing ferroptosis ([59]29–[60]31). Recent
   research has established a clear correlation between ferroptosis and
   hypertensive brain injury, indicating that increased blood pressure
   results in a large drop in GPX4 and GSH levels in brain tissue,
   resulting in iron overload. Excessive iron levels in the brain enhance
   oxidative stress and lipid peroxidation, finally resulting in brain
   damage ([61]32). Further, hypertensive patients have decreased activity
   and expression of GPX ([62]33). Inhibition of endothelial ferroptosis
   may help prevent endothelial cell apoptosis and calcification due to
   hypertension ([63]34, [64]35). Thus, ferroptosis-related pathways may
   play a role in hypertension related vascular endothelial damage, but
   there is a lack of direct research to prove this.

   In previous study, the ferroptosis of neuronal cells in hypertension is
   important for neurovascular diseases progression ([65]22). This may be
   a critical mechanism for neurological mediation of the progression of
   hypertension. According to traditional Chinese medicine theory, the
   EUO-TT medication combination may help increase cerebrovascular health
   in hypertension patients, however further research is needed to confirm
   this idea. We hypothesize that the EUO-TT drug pair may aid in
   inhibiting ferroptosis in the hypertensive milieu, hence reducing
   hypertensive damage to the brain's endothelial cells. Treatment with
   the EUO-TT medication combination may help prevent and improve the
   prognosis of people with IS.

   In this study, we used a network pharmacology approach to construct a
   “compound-target-disease” network through bioinformatics analysis to
   unravel the pharmacological basis and mechanism of action of the drug
   pair of EUO-TT in the treatment of hypertension-associated
   neurovascular lesions through the ferroptosis pathway. We believe that
   the findings of our study will provide a pharmacological basis for the
   clinical treatment of hypertension using Chinese medicine.

Materials and Methods

Acquisition of Effective Compounds and Action Targets of EUO-TT

   In this study, we searched for the chemical composition of EUO and TT
   through the pharmacology platform of the Chinese medicine network. We
   extracted the main active compounds of each medicine with the screening
   conditions of oral bioavailability ≥30% and drug-likeness ≥0.18
   ([66]36).

Prediction of Hypertension Disease Target

   The hypertension-related targets were searched by MOL ID numbers
   (expressed as protein name) of compounds using the Traditional Chinese
   Medicine Systems Pharmacology Database and Analysis Platform (TCMSP).
   Further, the gene names corresponding to each target were searched by
   the Uniport database ([67]37). Based on the GeneCards database, the
   relevant genes were searched by the keyword “hypertension” and
   intersected with the key active compounds of the pathway to obtain the
   relevant targets of the EUO-TT drug pair for the treatment of
   hypertension ([68]38).

Protein-Protein Interaction (PPI) Network Construction and Key Target
Screening

   After importing the intersecting targets into the STRING database and
   restricting the study species to humans, we obtained the PPI network of
   EUO-TT drug pair acting on neuroactive ligand-receptor interaction
   pathway and ferroptosis pathway ([69]39). The topological analysis of
   the PPI network was performed with the help of “Network Analyzer” in
   Cytoscape 3.7.1 software. The targets with degrees greater than the
   average of the degree of freedom, betweenness, and closeness were
   selected as key targets of the neuroactive ligand-receptor interaction
   pathway and ferroptosis pathway ([70]39).

Construction and Analysis of the Neuroactive Ligand-Receptor Interaction
Pathway-Ferroptosis Pathway Action Target Network

   The active compounds and their targets in the neuroactive
   ligand-receptor interaction pathway were imported into Cytoscape 3.7.1
   software, and the “STRING database” function was used to analyze the
   neuroactive ligand-receptor interaction pathway-ferroptosis pathway
   target network. The key compounds of EUO-Tribulus terrestris L drug
   pair mediating neuroactive ligand-receptor interaction pathway to
   coordinate ferroptosis pathway for the treatment of hypertension were
   screened based on compound-target connectivity.

Gene Ontology (GO) Functional Enrichment Analysis and Kyoto Encyclopedia of
Genes and Genomes (KEGG) Pathway Analysis

   In order to further understand the core target gene functions and the
   main pathways of action of EUO-Tribulus terrestris L in the treatment
   of hypertension, the drug-disease related targets were entered into the
   DAVID database ([71]40). The species “homo sapiens” was selected, and a
   threshold value of adjust p < 0.05 was set for GO function enrichment
   and KEGG pathway analysis, of which results were visualized by R
   programming software ([72]41, [73]42).

Molecular Docking

   As described in previous studies, we molecularly docked the screened
   core active ingredient to the core protein to verify the interaction
   strength ([74]43). The active ingredient structures were obtained from
   the PubChem database and the protein structures were obtained from the
   Protein Data Bank (PDB) ([75]44, [76]45). The 3D structures of the
   active ingredients (mol2 format) were downloaded, and operations such
   as hydrogenation, charge addition, detection of ligand roots, and
   search and definition of rotatable bonds were performed with AutoDock.
   Molecular docking was performed using AutoDock Vina software, and the
   results were optimized for mapping and further analysis with the help
   of Pymol software ([77]46, [78]47). R software was used to analyze and
   thermogram small molecule bound amino acids.

Data Acquisition and Pre-processing

   The [79]GSE22255 dataset is an ischaemic stroke-related microarray
   dataset obtained from the GEO database ([80]48). The dataset consisted
   of 20 IS patients and 20 age-matched controls. Analysis of variance was
   done using the “limma” package of R software (version 4.0.5) and the
   remove Barch effect function was used to remove batch effects. Gene
   expression was used for normalization analysis and p-values less than
   0.05 were defined as significant differences.

Building Predictive Models

   In short, the predictive model was constructed with reference to
   previous studies ([81]49–[82]53). With CHRM1, NR3C1, ADRB2, and OPRD1
   as dependent variables and the presence or absence of IS as the
   dependent variable, the “pacman” package of R software was used to
   build random forest (RF) and support vector machine (SVM) models. A
   receiver operating characteristic (ROC) curve, reverse cumulative
   distribution of |residual|, and boxplots of |residual| were plotted to
   select the most appropriate model. Then, using the “mlbench” and
   “caret” programmes, the cross curve was plotted to identify the
   relevance of each element. Following that, the “rms” function was used
   to build the nomogram prediction model. Additionally, calibration
   curves, DCA curves, and clinical effect curves were developed to
   validate the model's predictive ability ([83]54, [84]55).

Results

Extraction of Effective Compounds and Action Targets of EUO-TT Drug Pair

   The TCMSP analysis platform was used to search the main active
   ingredients of the EUO-TT drug pair. The active ingredients of each
   drug were screened and the targets of each compound were obtained
   according to the active ingredient “MOL ID”. Among them, 25 active
   compounds were screened from EUO, 12 from TT, and one duplicate
   compound from both EUO and TT (MOL000422: kaempferol). We found 36
   active compounds after de-weighting. Further, 30 compounds were found
   to correspond to 114 genes, and the gene names of each target were
   obtained from Uniprot ([85]Table 1).

Table 1.

   Detailed information on active compounds from Eucommia ulmoides Oliver
   and Tribulus terrestris L.
   Drug MOL ID Molecule name OB Drug-likeness
   Eucommia ulmoides Oliver MOL000073 ent-Epicatechin 48.96 0.24
   Eucommia ulmoides Oliver MOL000098 quercetin 46.43 0.28
   Eucommia ulmoides Oliver MOL000211 Mairin 55.38 0.78
   Eucommia ulmoides Oliver MOL000358 beta-sitosterol 36.91 0.75
   Eucommia ulmoides Oliver MOL000422 kaempferol 41.88 0.24
   Eucommia ulmoides Oliver MOL000443 Erythraline 49.18 0.55
   Eucommia ulmoides Oliver MOL002058 40957-99-1 57.20 0.62
   Eucommia ulmoides Oliver MOL004367 olivil 62.23 0.41
   Eucommia ulmoides Oliver MOL005922 Acanthoside B 43.35 0.77
   Eucommia ulmoides Oliver MOL006709 AIDS214634 92.43 0.55
   Eucommia ulmoides Oliver MOL007059
   3-beta-Hydroxymethyllenetanshiquinone 32.16 0.41
   Eucommia ulmoides Oliver MOL007563 Yangambin 57.53 0.81
   Eucommia ulmoides Oliver MOL009007 Eucommin A 30.51 0.85
   Eucommia ulmoides Oliver MOL009009 (+)-medioresinol 87.19 0.62
   Eucommia ulmoides Oliver MOL009015 (–)-Tabernemontanine 58.67 0.61
   Eucommia ulmoides Oliver MOL009027 Cyclopamine 55.42 0.82
   Eucommia ulmoides Oliver MOL009029 Dehydrodiconiferyl alcohol
   4,gamma'-di-O-beta-D-glucopyanoside_qt 51.44 0.40
   Eucommia ulmoides Oliver MOL009030 Dehydrodieugenol 30.10 0.24
   Eucommia ulmoides Oliver MOL009031 Epiquinidine 68.22 0.40
   Eucommia ulmoides Oliver MOL009038 GBGB 45.58 0.83
   Eucommia ulmoides Oliver MOL009042 Helenalin 77.01 0.19
   Eucommia ulmoides Oliver MOL009047 (+)-Eudesmin 33.29 0.62
   Eucommia ulmoides Oliver MOL009053
   4-[(2S,3R)-5-[(E)-3-hydroxyprop-1-enyl]-7-methoxy-3-methylol-2,3-dihydr
   obenzofuran-2-yl]-2-methoxy-phenol 50.76 0.39
   Eucommia ulmoides Oliver MOL009055 hirsutin_qt 49.81 0.37
   Eucommia ulmoides Oliver MOL009057 liriodendrin_qt 53.14 0.80
   Tribulus terrestris L. MOL000354 isorhamnetin 49.60 0.31
   Tribulus terrestris L. MOL000359 sitosterol 36.91 0.75
   Tribulus terrestris L. MOL000422 kaempferol 41.88 0.24
   Tribulus terrestris L. MOL000483 Moupinamide 118.35 0.26
   Tribulus terrestris L. MOL008559
   (2aR,2'R,4R,6aR,6bS,8aS,8bR,11aS,12aR,12bR)-4-((S)-2-(2,6-dimethylpheny
   l)propoxy)-5',5',6a,8a-tetramethyl-8-methylenedocosahydro-1H-spiro[pent
   aleno[2,1-a]phenanthrene-10,2'-pyran] 59.49 0.28
   Tribulus terrestris L. MOL008563
   (3R,8S,9S,10R,13R,14R,17S)-17-((2S,5R)-5-ethyl-6-methylheptan-2-yl)-3-h
   ydroxy-10,13-dimethyl-3,4,8,9,10,11,12,13,14,15,16,17-dodecahydro-1H-cy
   clopenta[a]phenanthren-7(2H)-one 40.93 0.79
   Tribulus terrestris L. MOL008567
   (3R,7R,8S,9S,10S,13R,14S,17R)-17-((2R,5S)-5-ethyl-6-methylheptan-2-yl)-
   3,10-dimethyl-2,3,4,7,8,9,10,11,12,13,14,15,16,17-tetradecahydro-1H-cyc
   lopenta[a]phenanthren-7-ol 34.21 0.76
   Tribulus terrestris L. MOL008568
   (Z)-3-(3,4-dihydroxyphenyl)-N-[2-(4-hydroxyphenyl)ethyl]acrylamide
   113.25 0.24
   Tribulus terrestris L. MOL008569 β-sitosterol-β-D-glucopyranoside 32.41
   0.71
   Tribulus terrestris L. MOL008588 terrestriamide 114.09 0.29
   Tribulus terrestris L. MOL008590
   (2aR,2'S,4R,4'R,5'S,6aS,6bS,8aS,8bR,9S,11aR,12aR,12bR)-4,4'-dihydroxy-5
   ',6a,8a,9-tetramethylicosahydro-1H-spiro[pentaleno[2,1-a]phenanthrene-1
   0,2'-pyran]-8(2H)-one 58.74 0.76
   Tribulus terrestris L. MOL008593
   (2aR,5S,6aS,6bS,8aS,8bS,11aS,12aR,12bR)-10-isopentyl-6a,8a,9-trimethyl-
   2,2a,3,4,5,6,6a,6b,7,8,8a,8b,11a,12,12a,12b-hexadecahydro-1H-naphtho[2'
   ,1':4,5]indeno[2,1-b]furan-5-ol 39.21 0.84
   [86]Open in a new tab

Enrichment Analysis of the Targets of Active Compounds in the EUO-TT Drug
Pair

   The potential target genes of the drugs were imported into the DAVID
   6.8 database, the list of target genes was entered, the species was
   limited to “Homo sapiens”, and the threshold was set to P < 0.05 for GO
   functional analysis and KEGG pathway enrichment analysis. The pathways
   and functional information with the top P-values in each type of
   process and pathways were selected, where smaller P-values and longer
   bar lengths indicated the strongest enrichment significance ([87]Figure
   1A). The top-ranked biological processes (BP), cellular components
   (CC), and molecular functions (MF) included: response to ammonium ion,
   cellular response to drug, response to xenobiotic stimulus, integral
   component of synaptic membrane, intrinsic component of postsynaptic
   membrane, integral component of postsynaptic membrane, adrenergic
   receptor activity, catecholamine binding, and G protein-coupled amine
   receptor activity. The top-ranked KEGG included: Folate biosynthesis,
   Calcium signaling pathway, AGE-RAGE signaling pathway in diabetic
   complications, and Neuroactive ligand-receptor interaction. Among them,
   the target genes of the EUO-TT drug pair were found to be significantly
   enriched in the neuroactive ligand-receptor interaction pathway.
   Therefore, we speculated that this drug pair may regulate the
   ferroptosis pathway by modulating the neuroactive ligand-receptor
   interaction pathway.

Figure 1.

   [88]Figure 1
   [89]Open in a new tab

   Enrichment analysis and protein-protein interaction (PPI) network of
   the active compound targets of EUO-TT L. GO functional analysis and
   KEGG enrichment analysis of the active compound targets of EUO-TT L.
   drug pair (A); PPI network analysis of the target genes of the
   neuroactive ligand-receptor interaction pathway, and the ferroptosis
   pathway mediated by the drug pair (B).

Potential Mechanism of Neuroactive Ligand-Receptor Interaction Pathway
Mediated by the Active Compounds of EUO-TT Drug Pair for the Regulation of
Ferroptosis

   Since the target genes of the drug pair were significantly enriched in
   the neuroactive ligand-receptor interaction pathway, we completed the
   intersection analysis of the drug-neuroactive ligand-receptor
   interaction pathway, and 26 intersection targets were identified
   ([90]Figure 2A). However, the drug-ferroptosis intersection analysis
   identified only two intersection targets ([91]Figure 2B). We,
   therefore, hypothesized that the neuroactive ligand-receptor
   interaction pathway may be the direct targeting pathway of EUO-TT drug
   pair, while the ferroptosis pathway may be the indirect pathway. The
   STRING database was used for the PPI analysis of the 26 intersecting
   targets of the drug-neuroactive ligand-receptor interaction pathway and
   ferroptosis pathway proteins (from the KEGG database). The study
   species was set as “Homo sapiens” to obtain the PPI network of the drug
   pair ([92]Figure 1B). We used the default parameters on the web page,
   and the file was saved in TSV format and imported into Cytoscape
   software for topological analysis of the PPI network. The nodes in
   [93]Figure 1B were selected out in [94]Figure 2C. Combined with the
   information on the degrees of freedom, degree, and centrality, the
   neuroactive ligand-receptor interaction pathway - ferroptosis pathway
   regulatory network was constructed ([95]Figure 2C). The degree sizes of
   the proteins in the ferroptosis pathway were indicated by the size of
   the squares, and several proteins were found to be critical for the
   pathway, including GPX4, FTH1, TP53, SLC11A2, HMOX1, GCLM, STEAP3,
   SLC11A2, FTL, ATG7, and ATG5. Further, CHRM1, NR3C1, ADRB2, and OPRD1
   in the neuroactive ligand-receptor interaction pathway were also found
   to interact with proteins related to the ferroptosis pathway. Among
   them, CHRM1 interacted with SLC3A2 and FTH1, NR3C1 interacted with
   TP53, ADRB2 interacted with PCBP2, and OPRD1 interacted with PCBP2.
   Therefore, CHRM1, NR3C1, ADRB2, OPRD1, FTH1, TP53, and PCBP2 were
   defined as the key proteins of the EUO-TT drug pair mediating
   ferroptosis through the neuroactive ligand-receptor interaction
   pathway.

Figure 2.

   [96]Figure 2
   [97]Open in a new tab

   Network analysis of the active compounds of EUO-TT L. mediating the
   neuroactive ligand-receptor interaction pathway regulating ferroptosis.
   Intersection analysis of drug-neuroactive ligand-receptor interaction
   pathway (A); Intersection analysis of drug-ferroptosis pathway (B); PPI
   network of neuroactive ligand-receptor interaction pathway -
   ferroptosis pathway (C).

Active Ingredient-Pathway Target Network Construction and Analysis

   The key proteins of the drug pair mediating ferroptosis were further
   used for enrichment analysis. The top-ranked BP, CC, and MF included:
   core promoter binding, G protein-coupled neurotransmitter receptor
   activity, neuropeptide binding, membrane raft, intrinsic component of
   the presynaptic membrane, integral component of the presynaptic
   membrane, G protein-coupled receptor signaling pathway coupled to
   cyclic nucleotide second messenger, adenylate cyclase-modulating G
   protein-coupled receptor signaling pathway, and phospholipase
   C-activating G protein-coupled receptor signaling pathway. The
   top-ranked KEGG included: Calcium signaling pathway, Ferroptosis, and
   Neuroactive ligand-receptor interaction ([98]Figure 3A). Hence, the key
   proteins may be associated with cell membrane channels, cell signaling,
   ferroptosis, and the G-protein pathway. Further, Sankey plots were used
   to illustrate the regulatory network of drug-active ingredient-key
   targets of neuroactive ligand-receptor interaction pathway ([99]Figure
   3B). The “four target proteins with the highest values” refers to the
   neuroactive ligand-receptor interaction with the four most closely
   associated proteins in the ferroptosis pathway in the PPI network.
   Those four key target proteins of the neuroactive ligand-receptor
   interaction pathway may be regulated by 15 active ingredients of the
   EUO-TT drug pair.

Figure 3.

   [100]Figure 3
   [101]Open in a new tab

   Network construction and pathway analysis of the active
   ingredient-pathway targets. GO and KEGG enrichment analysis of the key
   proteins CHRM1, NR3C1, ADRB2, OPRD1, FTH1, TP53, and PCBP2 (A); Sankey
   plots were used to demonstrate the drug-active ingredients-key
   neuroactive ligand-receptor interaction pathway targets regulatory
   network (B).

Molecular Docking Verification Experiment

   The 15 active ingredients and four target proteins with the highest
   values were screened according to the “drug-component-pathway-target”
   network, and molecular docking experiments were performed. Ultimately,
   12 active ingredients were validated for potential target binding to
   these four proteins ([102]Table 2). The results showed that the binding
   energy of docking of all proteins and the active ingredients were less
   than −5 kJ/mol, which indicates a tight binding between the ligand and
   receptor. The strongest protein-active ingredient bindings were as
   follows: ADRB2 and cyclopamine (−17.1 kJ/mol), CHRM1 and
   beta-sitosterol (−13.7 kJ/mol), and NR3C1 and mairin (−12.3 kJ/mol).
   OPRD1 bound to tabernemontanine (−9.4 kJ/mol), moupinamide (-8.1
   kJ/mol), epiquinidine (−9 kJ/mol), 3-Hydroxymethylenetanshinquinone (−9
   kJ/mol), and (–)- Tabernemontanine (−9.5 kJ/mol). In the molecular
   docking validation of ADRB2, CHRM1, NR3C1, and OPRD1 ([103]Figures 4,
   [104]5), valine, serine, and phenylalanine were found to be the key
   amino acids for the action of the active ingredients on ADRB2
   ([105]Figure 4A). Erythraline, quercetin, cyclopamine, moupinamide,
   terrestriamide, and beta-sitosterol had highly similar binding sites on
   ADRB2 ([106]Figure 4B). Tyrosine was found to be the key amino acid for
   the action of the active ingredient on CHRM1 ([107]Figure 4C).
   Kaempferol, beta-sitosterol, epiquinidine, (–)-tabememontanine,
   erythraline, and mouplinamide had highly similar binding sites on CHRM1
   ([108]Figure 4D). Lysine, arginine, and phenylalanine were found to be
   the key amino acids for the action of the active ingredient on NR3C1
   ([109]Figure 5A). Erythraline, moupinamide, beta-sitosterol, and mairin
   had highly similar binding sites on NR3C1 ([110]Figure 5B). PHE was
   found to be the key amino acid for the action of the active ingredient
   on OPRD1 ([111]Figures 5C,D).

Table 2.

   Virtual docking of representative ingredients and proteins.
                                    ADRB2 CHRM1 NR3C1 OPRD1
   Terrestriamide                    −10
   Tabernemontanine                  −9.7        −8.2  −9.4
   Quercetin                        −10.5
   Moupinamide                       −9.9  −9.7  −7.8  −8.1
   Mairin                                       −12.3
   kaempferol                              −10
   Erythraline                      −12.2 −11.4  −9.4
   Epiquinidine                      −9.5  −11          −9
   Cyclopamine                      −17.1
   beta-sitosterol                  −13.2 −13.7 −10.6
   3-Hydroxymethylenetanshinquinone  −8.6  −9.2         −9
   (–)-Tabernemontanine              −9.6 −11.3        −9.5
   [112]Open in a new tab

   Binging energy/(kcal/mol).

Figure 4.

   [113]Figure 4
   [114]Open in a new tab

   Visualization of molecular docking validation of ADRB2 and CHRM1. Amino
   acid composition of ADRB2 binding site (A); Bulk molecular docking map
   of ADRB2 (B); Amino acid composition of CHRM1 binding site (C); Bulk
   molecular docking map of CHRM1 (D).

Figure 5.

   [115]Figure 5
   [116]Open in a new tab

   Visualization of NR3C1 and OPRD1 molecular docking validation. Amino
   acid composition of NR3C1 binding site (A); Bulk molecular docking map
   of NR3C1 (B); Amino acid composition of OPRD1 binding site (C); Bulk
   molecular docking map of OPRD1 (D).

Construction of IS Prediction Models Based on the Expression Profiles of
CHRM1, NR3C1, ADRB2, and OPRD1

   Person correlation analysis suggested that in the control group, OPRD1,
   ADRB2, and NR3C1 were positively correlated with each other, while
   CHRM1 was found to be negatively correlated with NR3C1 and ADRB2
   ([117]Figure 6A). And in the IS group, CHRM1 was found to be
   significantly and positively correlated with NR3C1 and OPRD1
   ([118]Figure 6A). The positions of the four genes CHRM1, NR3C1, ADRB2,
   and OPRD1 on the chromosomes are shown in a circular diagram
   ([119]Figure 6B). To evaluate the predictive power of CHRM1, NR3C1,
   ADRB2, and OPRD1 for IS, RF and SVM models were constructed separately.
   The boxplot and cumulative distribution characteristics of the
   “residuals” indicate that the RF model has a lower residual
   distribution than the SVM model ([120]Figures 6C,D). The variation of
   error in the random forest model with the number of “trees” included in
   the model is shown in [121]Figure 6E. Based on the RF model, the
   importance of each feature was ranked from the largest to the smallest
   as OPRD1, CHRM1, NR3C1, and ADRB2 ([122]Figure 6F).

Figure 6.

   [123]Figure 6
   [124]Open in a new tab

   Expression characterization and predictive power analysis of CHRM1,
   NR3C1, ADRB2, and OPRD1. Correlation analysis of CHRM1, NR3C1, ADRB2,
   and OPRD1 in the [125]GSE22255 dataset in the control and IS groups,
   with red indicating positive correlation and blue indicating negative
   correlation (A); The positions of the four genes CHRM1, NR3C1, ADRB2,
   and OPRD1 in the chromosome are shown in a circular diagram (B);
   Boxplot of “residuals” in RF and SVM models (C); Cumulative
   distribution characteristics of “residuals” in RF and SVM models (D);
   The variation of error in the random forest model with the number of
   “trees” included in the model (E); The genes are ranked according to
   their importance (F); ROC curves reflecting the predictive power of RF
   and SVM models (G).

   The ROC curve analysis revealed that the AUC value of the RF model (AUC
   = 1.0) was higher than that of the SVM model (AUC = 0.845) ([126]Figure
   6G).

The Construction of a Nomogram Prediction Model, as Well as Its Risk
Enrichment Characteristics

   To create a visual model for IS prediction, we created a nomogram
   ([127]Figure 7A) using the four identified genes (CHRM1, NR3C1, ADRB2,
   and OPRD1). According to the calibration curve, the IS positivity rate
   predicted by this nomogram was generally consistent with the actual
   positivity rate ([128]Figure 7B). The DCA and clinical impact curves
   show that the nomogram model has good clinical application and
   diagnostic ability ([129]Figures 7C,D). Based on the IS prediction
   model, all samples in the [130]GSE22255 dataset were scored and divided
   into high-risk and low-risk groups based on their scores. Differential
   analysis was performed between the high-risk and low-risk groups. KEGG
   and GO enrichment were then performed and KEGG enrichment analysis
   revealed that yersinia infection, B cell receptor signaling pathway and
   autophagy (animal) were significantly associated with the risk score of
   this nomogram ([131]Figure 7E). GO enrichment analysis revealed that
   the risk score of this nomogram was significantly associated with the
   enrichment of dendritic cell differentiation, aging, focal adhesion,
   overall composition of organelle membranes, protein
   serine/threonine/tyrosine kinase activity and manganese ion binding
   pathways ([132]Figure 7F). Among them, dendritic cell differentiation
   and protein serine/threonine/tyrosine kinase activity and manganese ion
   binding were the most enriched ([133]Figure 7G). Therefore, CHRM1,
   NR3C1, ADRB2, and OPRD1, the target genes of the four EUO-TT drug pair
   active ingredients, are considered to be risk markers for IS.

Figure 7.

   [134]Figure 7
   [135]Open in a new tab

   Construction of nomogram prediction model and enrichment analysis based
   on this model. Construction an IS nomogram model based on the selected
   genes CHRM1, NR3C1, ADRB2, and OPRD1 (A); Calibration curve showing the
   diagnostic power of the nomogram model (B); DCA shows that the nomogram
   model has a good clinical application (C); Clinical impact curves show
   high diagnostic power of the nomogram model (D); KEGG (E) and GO (F)
   enrichment analysis of the genes included in this prediction model
   (E,F); The most significant GO enrichment pathway associated with this
   predictive model (G).

Discussion

   This study unraveled the mechanism of action of EUO-TT drug pair in the
   treatment of hypertension related endothelial injury through the
   ferroptosis pathway, providing a pharmacological basis for the clinical
   treatment of hypertension related cerebrovascular endothelial injury
   using Chinese medicine. CHRM1, NR3C1, ADRB2, and OPRD1, the target
   genes of the four active ingredients of the EUO-TT drug pair, were
   considered as risk markers for IS and a visual IS nomogram prediction
   model was constructed.

   Network pharmacology was used to screen the active compounds of EUO-TT
   L. drug pair and found that the target genes were mainly enriched in
   the neuroactive ligand-receptor interaction pathway in this research.
   By constructing the PPI network of neuroactive ligand-receptor
   interaction pathway-ferroptosis, we found that CHRM1, NR3C1, ADRB2,
   OPRD1, FTH1, TP53, and PCBP2 were the main target genes of the drug
   pair. Through the regulatory network of the drug's active ingredients
   and their key targets in the neuroactive ligand-receptor interaction
   pathway, we found 15 active ingredients of the drug pair that were
   involved in the regulation of the key target proteins. Further,
   molecular docking experiments verified that 12 active ingredients of
   the drug pair were tightly bound to the target proteins. The expression
   profiles of CHRM1, NR3C1, ADRB2, and OPRD1 were used to construct IS
   prediction models. The residual distribution of the RF model was found
   to be lower than that of the SVM model, and the ROC curve suggested a
   higher prediction accuracy of the RF model. The RF model rated the
   genes in ascending order of importance: OPRD1, CHRM1, NR3C1, and ADRB2.
   Following that, the nomogram prediction model was constructed using the
   four selected genes (CHRM1, NR3C1, ADRB2, and OPRD1). However, only
   CHRM1, NR3C1, and OPRD1 were eventually included in the prediction
   model, as inclusion of these three components was sufficient to
   reliably predict prognosis in the logistic model. The calibration
   curve, DCA, and clinical effect curves all indicated that this nomogram
   model is clinically applicable and diagnostically capable. The
   enrichment of dendritic cell differentiation and protein
   serine/threonine/tyrosine kinase activity and manganese ion binding
   pathways was found to be the most correlated with the prediction
   results based on nomogram risk scores.

   SLC3A2, FTH1, TP53, and PCBP2 are directly or indirectly associated
   with ferroptosis. SLC3A2 is a SystemXc- heavy chain subunit involved in
   the maintenance of a redox steady state ([136]56, [137]57). When
   SystemXc- is blocked, glutamate and cystine cannot be exchanged,
   resulting in intracellular glutamate accumulation, reduced glutathione
   synthesis, and reduced GPX4 activity, which ultimately leads to
   ferroptosis ([138]58). Further, the inhibition of SLC3A2, a ferroptosis
   suppressor gene, promotes ferroptosis in tumor and normal cells
   ([139]57, [140]59), and hence, its upregulation is associated with poor
   prognosis of cancer ([141]60). FTH1 (Ferritin heavy polypeptide 1) is
   an important component of ferritin. Ferritin degradation increases
   intracellular free iron content and can activate ferroptosis
   ([142]61–[143]63), while the overexpression of FTH1 can inhibit
   ferroptosis by impairing ferritinophagy ([144]64). For instance, in the
   head and neck squamous cell carcinoma, FTH1 suppressed ferroptosis and
   led to the poor prognosis of the carcinoma ([145]65). Further, FTH1
   expression is downregulated in cells sensitive to ferroptosis relative
   to cells resistant to ferroptosis ([146]66). p53 is an oncogene that
   regulates the cell cycle, induces apoptosis, and promotes DNA
   metabolism ([147]67, [148]68). P53 may also regulate ferroptosis
   ([149]69–[150]72). Jiang et al. found that P53 inhibited SystemXc-
   activity and downregulated the expression of solute carrier family 7
   member 11, which in turn induced cellular ferroptosis ([151]70). P53
   can also catalyse glutamate production by enhancing glutaminase 2
   activity, thereby reducing the entry of cystine into cells, reducing
   glutathione synthesis, and inducing ferroptosis ([152]73). In addition,
   P53 promotes cellular unsaturated fatty acid oxidation by enhancing the
   activation of spermidine/spermine N1-acetyltransferase 1, which also
   leads to cellular ferroptosis ([153]71). Thus, P53 is closely related
   to ferroptosis. PCBP2 is a multifunctional adaptor protein of the
   Poly-binding family which is involved in iron metabolism and is an iron
   chaperone of ferritin ([154]74). However, the specific binding sites of
   iron and ferritin on PCBP2 have not yet been identified ([155]75). Iron
   and ROS are promoters and mediators of ferroptosis ([156]76). Iron
   metabolic homeostasis and lipid peroxidation are crucial for the
   occurrence of cellular ferroptosis ([157]77). A study found that PCBP2
   binding to severely oxidized RNA inhibited apoptosis induced by
   oxidative stress ([158]78). Moreover, PCBP2 is one of the ferroptosis
   risk signature genes and can predict the prognosis of adrenocortical
   carcinoma in combination with the other five ferroptosis risk signature
   genes ([159]79).

   Similar to previous studies, we found that SLC3A2, FTH1, TP53, and
   PCBP2 were the ferroptosis pathway-associated proteins that interacted
   with CHRM1, NR3C1, ADRB2, and OPRD1. The NR3C1 protein regulates the
   genetic features involved in the P53 signaling pathway ([160]80).
   Further, ADRB2 increases the protein levels of c-myc in pancreatic
   ductal adenocarcinoma cells through the recruitment of PCBP2 ([161]81).
   Similarly, in this study, we found that NR3C1 interacted with TP53
   andADRB2 interacted with PCBP2. In addition, this is the first study to
   show that CHRM1 interacted with SLC3A2 and FTH1, and OPRD1 interacted
   with PCBP2. These results indicated that CHRM1, NR3C1, ADRB2, and OPRD1
   may influence the ferroptosis pathway by interacting with the key
   proteins related to ferroptosis. CHRM1, NR3C1, ADRB2, and OPRD1 were
   the target proteins related to the neuroactive ligand-receptor
   interaction pathway. Therefore, for the first time, we show that the
   neuroactive ligand-receptor interaction pathway may influence the
   process of ferroptosis through the interaction of its related proteins
   CHRM1, NR3C1, ADRB2, and OPRD1 with the key proteins of ferroptosis. We
   further showed that ADRB2 bound most tightly to cyclopamine; CHRM1 to
   beta-sitosterol; NR3C1 to mairin; OPRD1 to tabernemontanine,
   moupinamide, epiquinidine; and OPRD1 to tabernemontanine, moupinamide,
   epiquinidine, 3-Hydroxymethylenetanshinquinone, and
   (-)-Tabernemontanine. Beta-sitosterol has anti-inflammatory,
   anti-tumor, antioxidant, anti-hyperlipidemic, and antihypertensive
   effects ([162]82–[163]84), and moupinamide is also anti-inflammatory
   ([164]85). Inflammatory processes are involved in the development of
   hypertension ([165]86, [166]87). Therefore, it can be hypothesized that
   moupinamide has potential anti-hypertensive effects. Tanshinone was
   also found to have a protective effect against cardiac hypertrophy in
   SHRs ([167]88). Tanshinone slowed the metabolism of the
   antihypertensive drug cloxacin by inhibiting CYP3A4/CYP2C9 activity
   ([168]89). Further, the regulation of KV current by tanshinone reduced
   hypoxic pulmonary hypertension and inhibited hypoxia-induced pulmonary
   artery wall remodeling ([169]90).

   In this study, CHRM1, NR3C1, ADRB2, and OPRD1, the target genes of the
   four EUO-TT L. drug pair active ingredients, were considered as risk
   markers for IS. CHRM1 was found to be associated with muscarinic
   receptor-mediated anti-inflammatory mechanisms ([170]91). Following IS,
   there is a decrease in B lymphocyte production, which may be related
   with aberrant NR3C1 expression ([171]92). ADRB2, an apoptosis-related
   gene, was found to enhance the activity of BDNF/TrkB and cAMP/PKA
   signaling pathways leading to cellular ischemic injury ([172]93). In
   addition, polymorphisms in the ADRB2 gene have been found to
   potentially contribute to a high risk of stroke ([173]94). Common
   variants of OPRD1 have also been found to be associated with
   neurodegenerative diseases ([174]95).

   In the present study, the active ingredients of EUO-TT L., namely
   cyclopamine, beta-sitosterol, mairin, tabernemontanine, moupinamide,
   epiquinidine, and 3-Hydroxymethylenetanshinquinone, were found for the
   first time to target and bind neuroactive ligand-receptor interaction
   pathway-related proteins, which in turn regulated the process of
   cellular ferroptosis and ultimately affected the onset and progression
   of hypertension. There is a lack of previous studies confirming the
   association of the alkaloids tabernemontanine and epiquinidine, and the
   terpenoids cyclopamine, and mairin of EUO with hypertension,
   ferroptosis, and the neuroactive ligand-receptor interaction pathway.
   Hence, these pathways can be a new therapeutic target for the
   prevention and treatment of hypertension.

   However, this study also has its shortcomings. The network pharmacology
   is based on the bioinformatics analysis of the database of herbal
   targets and disease genes, and thus lacks experimental validation.
   Hence, further experimental validation is needed to explore the
   mechanism of action of the drugs or their active ingredients. In
   addition, this study did not consider the selection of dosage in drug
   combination and the possible chemical reactions and changes of active
   ingredients in the process of concoction and decoction of herbal
   medicines. Furthermore, we also acknowledge that the current research
   in TCM for the treatment of endothelial injury in hypertension-related
   cerebrovascular disease is still limited, and therefore we suggest to
   provide more confirmation and discussion in the future.

Conclusions

   The active ingredients of EUO-TT L. drug pair can act on proteins
   related to the neuroactive ligand-receptor interaction pathway, which
   may affect the process of cellular ferroptosis and potentially affect
   the progression of hypertension. CHRM1, NR3C1, ADRB2, and OPRD1, the
   target genes of the four EUO-TT L. drug pair active ingredients, were
   considered as risk markers for IS.

Data Availability Statement

   The original contributions presented in the study are included in the
   article/supplementary material, further inquiries can be directed to
   the corresponding author.

Author Contributions

   QZ: software, validation, formal analysis, data curation, and
   writing-original draft. JY: conceptualization, methodology,
   supervision, funding acquisition, and writing-review and editing. CY:
   methodology, conceptualization, project administration, and funding
   acquisition. XY: data curation and writing-review and editing. YC:
   software and data curation. All authors contributed to the article and
   approved the submitted version.

Funding

   This work was supported by the National Natural Science Foundation of
   China (No. 81804061), Shandong Province Taishan Scholar Construction
   Project Funds No. 2018-35, and Ji'nan Science and Technology Project
   (No. 201805078).

Conflict of Interest

   The authors declare that the research was conducted in the absence of
   any commercial or financial relationships that could be construed as a
   potential conflict of interest.

Publisher's Note

   All claims expressed in this article are solely those of the authors
   and do not necessarily represent those of their affiliated
   organizations, or those of the publisher, the editors and the
   reviewers. Any product that may be evaluated in this article, or claim
   that may be made by its manufacturer, is not guaranteed or endorsed by
   the publisher.

Acknowledgments