Integrated Network Pharmacology and Proteomic Analyses of Targets and Mechanisms of Jianpi Tianjing Decoction in Treating Vascular Dementia

Background Vascular dementia (VD), associated with cerebrovascular injury, is characterized by severe cognitive impairment. Jianpi Tianjing Decoction (JTD) has been widely used to treat VD. However, its molecular targets and mechanisms of action in this treatment remain unclear. This study integrated network pharmacology and proteomics to identify targets and mechanisms of JTD in the treatment of VD and to provide new insights and goals for clinical treatments. Methods Systematic network pharmacology was used to identify active chemical compositions, potential targets, and mechanisms of JTD in VD treatment. Then, a mouse model of VD was induced via transient bilateral common carotid artery occlusion to verify the identified targets and mechanisms of JTD against VD using 4D label-free quantitative proteomics. Results By screening active chemical compositions and potential targets in relevant databases, 187 active chemical compositions and 416 disease-related compound targets were identified. In vivo experiments showed that JTD improved learning and memory in mice. Proteomics also identified 112 differentially expressed proteins in the model and sham groups and the JTD and model groups. Integrating the network pharmacology and proteomics results revealed that JTD may regulate expressions of cytochrome c oxidase subunit 7C, metabotropic glutamate receptor 2, Slc30a1 zinc transporter 1, and apolipoprotein A-IV in VD mice and that their mechanisms involve biological processes like oxidative phosphorylation, regulation of neuron death, glutamate secretion, cellular ion homeostasis, and lipoprotein metabolism. Conclusions JTD may suppress VD development via multiple components, targets, and pathways. It may thus serve as a complementary treatment option for patients with VD.


Introduction
Vascular dementia (VD) is a cognitive dysfunction associated with cerebrovascular injury and characterized by severe impairment of cognitive functions, including attention, memory, verbal fuency, and executive function [1].
Epidemiological studies indicate that VD is the second leading cause of dementia and that in addition to economic burdens, it negatively impacts patient health, productivity, and daily activities [2]. Current studies identify VD as a cognitive dysfunction with multifactorial pathogenesis, probably related to atherosclerosis, lipid metabolism, oxidative stress, the infammatory response, calcium overload, excitotoxicity, and/or hemostatic activation [3][4][5][6]. Drugs that may improve VD symptoms include choline esterase inhibitors (donepezil, galantamine, and rivastigmine) and N-methyl-D-aspartate receptor antagonists (memantine) [7]. Studies have shown that donepezil and galantamine treatments modestly improve cognition but have no efect on activities of daily living [8,9]. In a randomized controlled trial, incidences of adverse events from donepezil (10 mg, 5 mg) and placebo were 16.3%, 10.1%, and 8.8%, respectively [10]. Another clinical trial showed that compared with a placebo, there were more deaths in the donepezil group [11]. In a trial testing galantamine treatment for VD, the incidence of adverse events from the drug and placebo were 13% and 6%, respectively [12]. Rivastigmine has minimal efects on cognitive symptoms [13]. Memantine produces small benefts in patients with mild to moderate vascular dementia, and current data are insufcient to support the widespread use of memantine in vascular dementia [14]. Tese drugs also have limited efcacy and adverse efects that include nausea, vomiting, diarrhea, dizziness, headache, and hypertension [14,15]. Terefore, there is an urgent need to develop complementary and alternative VD therapies.
Chinese herbal formula (CHF) has multiple targets and few side efects, playing an active role in VD treatment [16]. CHF showed fewer adverse efects, lower costs, and improved suitability for long-term use compared with currently prescribed drugs [17]. For example, clinical trials have confrmed that the Shenmayizhi formula combined with ginkgo extract tablets efectively improves cognitive function in mild-to-moderate VD without adverse efects, and clinical outcomes from Dingzhi Yicong granules are superior to those with piracetam in patients with VD [18,19]. Our team developed Jianpi Tianjing Decoction (JTD) based on years of clinical experience and guided by traditional Chinese medicine (TCM) theory. JTD consists of seven Chinese herbal medicines (CHM), including Panax ginseng C.A. Meyer, Gastrodia elata, Atractylodes macrocephala, Morinda ofcinalis Radix, Acorus tatarinowii Schott, Rhizoma coptidis, and Semen cuscutae. Previous studies have shown that Gastrodia elata ameliorates vessel elasticity and prevents atherosclerosis [20]. Panax ginseng C.A. Meyer extract attenuates neuronal injury and cognitive defcits in a VD rat model by upregulating the apoptosis regulator Bcl-2 and downregulating the apoptosis regulator BAX (Bax) protein expression [21]. Rhizoma coptidis improves the rat neurological function after acute brain injury by increasing the hippocampal brain-derived neurotrophic factor expression [22]. Our preliminary study confrmed that JTD signifcantly improves cognitive function and quality of life in patients with mild cognitive dysfunction [23,24]. Animal studies have also identifed potential mechanisms of JTD for treating VD, including reducing oxidative stress damage, maintaining hippocampal mitochondrial membrane potential and adenosine triphosphate (ATP) levels, and improving mitochondrial dysfunction [25,26]. However, CHF composition is so intricate that it is difcult to fully clarify its mechanisms through traditional research methods. Terefore, it is necessary to focus on the potential system-level mechanisms of JTD in VD treatment.
Network pharmacology is a novel method for studying CHM and CHF that combines systematic network analysis and pharmacology to identify interactions among compounds, genes, proteins, and diseases [27]. Tian et al. successfully predicted 28 potentially active Shenzhi Jiannao prescription ingredients and 21 VD therapy targets. Tey found that the potential targets of these 28 active ingredients mainly involve neuroactive ligand-receptor interactions, calcium, apoptosis, and cholinergic synaptic signaling pathways [28]. Trough network pharmacology analysis, Shi et al. discovered that the fve core compounds in Yizhi Tongmai decoction exert antiVD efects [29]. Proteomics, an important tool for exploring drug targets and molecular mechanisms, is now also widely applied in many life sciences [30]. It can be used to analyze diferentially expressed proteins (DEPs) to explore CHM molecular mechanisms of action. Yang et al. identifed 245 Fugui Wenyang Decoction (FGWYD) genes and 145 VD genes via network pharmacology, showing that the Nrf2/HO-1 pathway plays an important role in the FGWYD treatment of VD [31]. Tat group also used proteomics to verify the neuroprotective mechanistic role of the Nrf2/HO-1 pathway in the FGWYD treatment of VD. An integrated network pharmacology and proteomics analysis can provide a more comprehensive description of CHF molecular mechanisms. Hence, this study integrated network pharmacology and proteomics to analyze the molecular targets and mechanisms of action of JTD in the treatment of VD. First, network pharmacology was performed to predict the target proteins and pathways related to JTD in the treatment of VD. Second, mass spectrometry (MS) analysis was used to identify diferentially expressed proteins after VD model mice were treated with JTD. Finally, we revealed the targets and mechanisms of JTD by combining the network pharmacology and proteomic results. Cuscutae. Next, active chemical compositions were screened using the Swiss ADME database (http://www.swissadme.ch/ (accessed October 9, 2022)) and selected based on oral bioavailability "> 30%," gastrointestinal absorption level "high," and "Yes" for atleast three of Lipinski, Ghose, Veber,

Experimental
Drugs. JTD granules were composed of Panax ginseng C.A. Meyer 9 g, Gastrodia elata 9 g, Atractylodes macrocephala 10 g, Morindae Ofcinalis Radix 6 g, Acorus tatarinowii Schott 6 g, Rhizoma coptidis 3 g, and Semen Cuscutae 12 g. Te JTD granules are prepared in accordance with the previous methods [32,33]. Te collected CHM was washed with water to remove any dust or foreign particles present on them and shade-dried for one week at room temperature to avoid excessive loss of volatile components. After drying, the CHM was ground to prepare the crude powder. Te above crude powder was subjected to extraction using a hydroalcoholic (30 : 70, water: ethanol) solvent to obtain the CHM granules [34].  [25,26] and that the typical daily intragastric dose in mice is 10 mg/kg [35]. Tus, we dissolve JTD granules in an appropriate amount of normal saline to achieve a fnal solution concentration of 2.016 g/ml.

Animal
Modeling, Grouping, and Intervention. Te 40 mice were randomly divided into three groups: sham surgery (n � 10), model (n � 15), and JTD (n � 15). Te transient bilateral common carotid artery occlusion (BCCAO) surgery was performed as previously described with minor modifcations [36]. Te mice in each group were anesthetized by intraperitoneal injection of 0.3% pentobarbital sodium solution (0.25 ml/10 g). Mice in the model and JTD groups had a midline cervical incision. After exposure, both the right and left common carotid arteries were isolated from the adjacent vagus nerve, and silk was passed below each carotid artery for closure. Te bilateral carotid arteries were locked by silk strings for 10 min and then released for 10 min, and this was repeated three times. Te strings were then removed and the incision sutured. In the sham group, the same neck region was surgically opened to isolate the vagus nerve and then sutured without a transaction. To prevent wound infection, each mouse received an intramuscular injection of penicillin at a rate of 5,000 units per day for three days. Seven days after surgery, mice in the JTD group were treated with daily intragastric 10 ml/kg/d JTD solution for 28 days. Sham and model groups were treated on the same schedule with saline.

Morris Water Maze Test.
After the 28 treatment days, learning and memory were assessed by the Morris water maze test. Tis test uses a circular pool (100 cm in diameter) with a circular escape platform (6 cm in diameter, 1.0 cm below the water's surface) and an image acquisition system. Te pool is divided into four quadrants, with the circular escape platform in the third quadrant. Powdered milk is added to make the water opaque. Te mice were released from the four quadrants, respectively, and given 90 s (max) to fnd the platform. If the mice could not fnd the platform in 90 s, they were guided onto the platform and allowed to Evidence-Based Complementary and Alternative Medicine remain for 30 s. Training occurred on four consecutive days. At the end of this training period, the mice were randomly released into the frst, second, or fourth quadrant, and their time to reach the platform, or escape latency (EL), was recorded. Testing lasted fve days. On day 6, the circular escape platform was removed, and each mouse was placed in the frst quadrant. Duration spent in the third quadrant and number of times crossing the platform (TCP) during 1 min were recorded.

Sample Preparation and Protein Extraction.
Following Morris water maze testing, all mice were sacrifced, and the extracted brains were immediately placed in liquid nitrogen and stored at −80°C. Tree samples from each group were randomly assigned to subsequent analysis. SDT bufer (P0015F, Beyotime, 4% SDS, 100 mM Tris-HCl, pH � 7.6) was added to samples to extract proteins. Te supernatant was quantifed with the BCA Protein Assay Kit (P0012, Beyotime). Proteins were then digested using the flter-aided sample preparation procedure [37]. Te C18 column (IonOpticks, Australia; 25 cm × 75 μm, 1.6 μm C18 beads) was used to desalt the peptide segment. In addition, the study protocol is shown in Figure 1.

Targets of Disease-Related Compounds.
A total of 187 active chemical compositions in JTD and 854 potential targets for its herbal ingredients were screened from various databases and published literature. A total of 4,709 VD genes were obtained from the GeneCards, OMIM, and DrugBank databases. Te 416 disease-related compound targets were obtained by a Venn diagram (Figure 2(a)). Among those disease-related compound targets, Panax ginseng C.A. Meyer had 322 potential targets, Gastrodia elata had 143 potential targets, Atractylodes macrocephala had 84 potential targets, Morindae Ofcinalis Radix had 88 potential targets, Acorus tatarinowii Schott had 149 potential targets, Rhizoma coptidis had 269 potential targets, and Semen Cuscutae had 139 potential targets (Figure 2(b)).

Construction of the JTD-VD PPI Network and Crucial
Targets. Te active chemical compositions and diseaserelated compound targets were imported into Cytoscape 3.9.1 to construct the herb-component-target network diagram ( Figure 3), which consists of 3,671 edges and 585 nodes. Te higher the degree value, the larger the node. Te top fve degrees among all active chemical compositions were quercetin, dauricine, kaempferol, deoxyharringtonine, and panaxacol. Next, the 416 disease-related compound targets were imported into the String database to build the JTD-VD PPI network ( Figure 4). Te top 30 Hub genes were selected and mapped using the CytoHubba plug-in of Cytoscape 3.9.1, and the top-ranked genes were RAC-alpha serine/threonine-protein kinase, cellular tumor antigen p53, CREB-binding protein, ethylene-responsive transcription factor ESR1, and cyclin-dependent kinase inhibitor 1 ( Figure 5).

GO and KEGG Pathways Enrichment Analysis of Potential
Targets. To investigate potential signaling pathways or biological processes (BPs), GO and KEGG pathways were analyzed for potential JTD targets. GO enrichment analysis showed that potential targets were involved in 2,766 GO terms: 3,414 in BP, 234 in cellular components (CC), and 395 in molecular functions (MF). Te top 20 GO enrichment analyses are shown in Figure 6(a). GO-BP analysis showed that potential targets focused mainly on the negative regulation of phosphorylation, the infammatory response, cellular calcium ion homeostasis, regulation of synapse organization, the regulation of the lipid catabolic process, and the cellular response to nitrogen compounds. GO-CC analysis showed that potential targets were primarily focused on the receptor complex, postsynaptic membrane, and presynaptic membrane. Additionally, GO-MF analysis showed that potential targets were concentrated mainly on protease binding, copper ion binding, and lipoprotein particle binding. KEGG enrichment analysis showed that potential targets were involved in 236 pathways ( Figure 6(b)). Tus, the mechanisms of action of JTD in the treatment of VD may be closely related to multiple pathways,  Evidence-Based Complementary and Alternative Medicine 5 including lipid and atherosclerosis, neurodegeneration in multiple diseases, calcium signaling, fuid shear stress and atherosclerosis, PI3K-Akt signaling, and MAPK signaling.  (Figure 8(a)). Cluster analysis showed that the up-or downregulated model group proteins showed back regulation in the sham and JTD groups (Figures 9(a) and 9(b)). Alpha-1-antitrypsin 1-3 (serpina1c), potassium voltagegated channel subfamily C member 2, histone H2A, and protein FAM234B showed upregulation in the model group and downregulation in the JTD group. Adar and Ighg showed downregulation in the model group and upregulation in the JTD group. Tere were also three overlapping proteins among the DEPs identifed by proteomics and the potential targets selected by the network pharmacology: metabotropic glutamate receptor 2 (Grm2), carbonic anhydrase 1, and glycolate oxidase 1.   Figure 10). GO analysis of the DEPs of the model and sham groups showed that negative regulation of endopeptidase activity, negative regulation of complement activation, the lectin pathway, positive regulation of the fatty acid biosynthetic process, and negative regulation of ATPase activity were the primary BPs. Tese DEPs are mainly in the extracellular space, the external plasma membrane, and the hemoglobin complex. Tey are also associated with receptor binding, ion exchange, and functions (Figure 11(a)).

General Condition of Experimental Animals and Morris
Next, GO enrichment analysis was performed on the DEPs of the JTD and model groups (Figure 11(b)). GO-BP analysis showed that these DEPs are signifcantly involved in plasma lipoprotein particle clearance, negative regulation of neurotransmitter secretion, the lipoprotein metabolism process, positive regulation of phagocytosis, defnitive hemopoiesis, cellular calcium ion homeostasis, and negative regulation of phosphorylation. GO-CC analysis showed that these DEPs are mainly located in the organelle membrane, receptor complex, and outer side of the plasma membrane. Te GO-MF analysis showed they are associated with MFs like cholesterol binding, lipoprotein binding, ion binding, and protease activity.
KEGG pathway enrichment analysis showed that the top fve signaling pathways were complement and coagulation cascades, African trypanosomiasis, tuberculosis, legionellosis, and systemic lupus erythematosus in the DEPs of the model and sham groups (Figure 12(a)). Similarly, the DEPs of the JTD and model groups are mainly involved in vitamin digestion and absorption, fat digestion and absorption, and pyruvate metabolism (Figure 12(b)).

PPI Network.
DEPs of the model and sham groups and the JTD and model groups were combined and deduplicated to obtain 112 common DEPs. Next, Cytoscape 3.9.1 was used to construct a PPI network diagram for these DEPs (Figure 13). Metascape analysis of this network included regulation of the fatty acid biosynthetic process, positive regulation of phagocytosis, response to inorganic substance, pyruvate metabolism, and aerobic electron transport chain (Figures 14(a) and 14(b)). Similarly, the DEPs of the JTD and model groups were imported into String to construct a PPI network ( Figure 15). Based on the CytoHubba plug-in of Cytoscape 3.9.1, 10 hub genes were screened, including actin-like protein 6A, tyrosine-protein kinase Mer, cysteine and glycine-rich protein 1, protein Evidence-Based Complementary and Alternative Medicine turtle homolog B, Apolipoprotein A-IV (Apoa4), radixin, disheveled-associated activator of morphogenesis 2, Serpi-na1c, and guanylate cyclase soluble subunit alpha-1. Grm2, cytochrome c oxidase subunit 7C (Cox7c), and Slc30a1 (also known as zinc transporter 1 (Znt1)) also participated in this PPI network.

Discussion
VD is a cluster of cognitive disorder syndromes caused by cerebrovascular lesions. Current risk factors for VD include advanced age, diabetes, hypertension, hyperlipidemia, atherosclerosis, and stroke [39]. In particular, VD risk nearly doubles post-stroke [40]. Cerebral hypoperfusion from cerebrovascular disorders may be a potential VD mechanism [41,42]. Neuropathology studies have reported that cerebral hypoperfusion results in reduced glucose and oxygen supplies, leading to cellular energy metabolism [43], ionic imbalance [44], excitotoxicity [45], oxidative stress [46], and neuroinfammation [47]. Tese mechanisms drive downstream structural changes, including blood-brain barrier dysfunction, white matter lesions, microinfarcts, and hippocampal atrophy, which may play a direct pathogenic role in VD [48,49]. VD accounts for ∼15-20% of dementia cases, and its incidence increases dramatically with age [50,51]. VD both afects patient quality of life and increases the risk of death [52]. Terefore, fnding efective treatments remains a research focus. JTD has long been used to efectively treat VD [23,24]. To further investigate the molecular mechanisms of JTD in VD treatment, this study combined network pharmacology and proteomics data to gain a global overview.

Regulation of Mitochondrial Dysfunction.
Mitochondrial dysfunction has also been reported to be a signifcant factor in VD [43]. Under anoxic conditions, the mitochondrial electron transport chain is disturbed, leading to increased reactive oxygen species (ROS) production [53]. Oxidative stress, which occurs when the ROS-antioxidant balance is disrupted, is increasingly understood to be involved in VD [54]. Under normal conditions, the brain depends on a constant energy supply from ATP through mitochondrial oxidative phosphorylation [55]. However, oxidative stress can cause mitochondrial dysfunction, triggering impaired cerebral energy metabolism and neuronal death [56,57]. Cyclooxygenases (COX) are the main enzyme in the mitochondrial electron transport chain, which uses oxygen in the generation of ATP via oxidative phosphorylation [58]. Te network pharmacology and proteomic analyses herein show that negative regulation of phosphorylation is another important BP. Proteomics identifed Cox7c to be involved in oxidative phosphorylation and metabolism pathways. Cox7c, a member of the cytochrome c oxidase complex responsible for mitochondrial respiration promotes ATP synthesis and reduces mitochondrial dysfunction [59]. Recent studies have revealed Cox7c to be a potential biomarker of pathogenesis in Alzheimer's disease [60]. However, the efects of Cox7c in VD require further clarifcation. Herein, Cox7c expression was upregulated in brain tissue of VD mice treated with JTD, indicating that this CHF may play a role in treating VD by attenuating mitochondrial dysfunction.

Neuronal Glutamate Excitotoxicity Regulation.
A main cause of VD-induced cognitive dysfunction is excitotoxicity. Glutamate excitotoxicity has been hypothesized to be excessively activated by excitatory glutamate receptors, causing neuronal dysfunction or death [45]. Grm2 encodes metabotropic glutamate subtype receptor 2, known as mGluR2. Past studies have shown that GRM2 may be involved in regulating neural apoptosis, or cell death caused by hypoxia and ischemia [61]. Herein, Grm2 was an overlap protein between potential JTD targets and DEPs, which is mainly involved in the regulation of neuronal death, glutamate receptor activity, and glutamate secretion. Terefore, JTD may regulate the GRM2 expression and modulate these processes, reducing brain tissue damage and improving cognitive function.

Cellular Ion Homeostasis Regulation.
Cerebral ischemia enhances synaptic activity, leading to increased zinc release. However, high intracellular zinc levels may become toxic to neurons and neuroglia [62], rapidly leading to cell death 24  Evidence-Based Complementary and Alternative Medicine [63]. Tis may be another contributor to VD. Among the 65 DEPs, Slc30a1 was the only protein that regulates cellular zinc ions and calcium ion homeostatic processes. Slc30a1/ Znt1 is well-known as a crucial regulator of zinc absorption and transport [64]. Znt1 reduces glial and neuronal zinc levels, protecting these cells from zinc toxicity and reducing their deaths [65,66]. Te proteomics data herein also confrmed that Znt1 levels are signifcantly increased in VD mice treated with JTD, further confrming the neuroprotective efects of JTD.

Atherosclerosis Regulation.
Herein, the KEGG pathway enrichment analysis revealed that ApoA4 participates in several signaling pathways, including lipid and atherosclerosis, fat digestion and absorption, and cholesterol metabolism. Carotid artery stenosis and atherosclerosis are risk represents GO or KEGG terms, whose size is proportional to the number of genes under that term, and nodes of the same color belong to the same cluster.
factors for cognitive impairment [67]. It is well accepted that ApoA4 is a major component of high-density lipoprotein and chylomicrons, which have antiatherosclerotic efects [68]. Exogenous administration of ApoA4 reduces the incidence of acute rupture of arterial plaques in the apolipoprotein E knockout mouse model, confrming its ability to stabilize plaque [69]. In addition, previous research has confrmed that ApoA4 is involved in lipid uptake and metabolism [70] and antiatherosclerosis [71] and inhibits thrombosis [72]. Tis is consistent with our results showing that, compared with the model group, ApoA4 expression was upregulated in the JTD group, further confrming its important role in the anti-atherosclerotic efects of JTD.

Conclusions
Integrated network pharmacology and proteomics analysis revealed that Cox7c, Grm2, Slc30a1, and ApoA4 are critical targets of JTD in VD treatment. In vivo mechanisms may be involved in attenuating mitochondrial dysfunction, reducing excitotoxicity, maintaining cellular ion homeostasis and antiatherosclerosis.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Ethical Approval
Animals were raised and handled at the Zhejiang Chinese Medical University Laboratory Animal Research Center. Te use of the experimental animals involved in this research followed the relevant national requirements for medical laboratory animals and was approved by the Animal Experimentation Ethics Committee of Zhejiang Chinese Medical University (No. IACUC-20210906-12).

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
JL, YH 1-, and JX wrote and revised the manuscript. JL and JG designed and performed the experiments. JL, JX, MF, MS, and WL searched the databases for relevant results. YH 1 , JG, YW, and YH 4 reviewed drafts of the paper. YH 1 helped in coordinate funding. All authors have read and approved the fnal manuscript.