Based on Network Pharmacology to Explore the Mechanisms of Radix Astragali Combined with Prepared Radix Rehmanniain for Treating Osteoporosis


 Background: With the improvement of people's living standards, the aging population in China has gradually increased, and the treatment of osteoporosis (OP) has become a major problem afflicting the medical field. Radix Astragali(RA) and Prepared Radix Rehmannia(PRR) are commonly used Chinese herbal medicines. The combination of them can achieve a synergistic effect in the treatment of osteoporosis. However, its mechanism of action remains uncertain.Objective: This study aims to investigate the possible molecular mechanism of RA combined with PRR in the treatment of OP using an integrated strategy of network pharmacology and experimental validation.Methods: The active ingredients of RA combined with PRR were searched and screened by TCMSP database, and the targets of active ingredients were predicted and supplemented by TCMSP and SwissTargetPrediction databases. The target genes related to OP diseases were searched in GeneCards and OMIM comprehensive databases. The intersection targets of drugs and diseases were imported into String database to obtain the interaction information of intersection target genes. Cytoscape3.7.1 software was used to construct the protein interaction network diagram and "drug component-target-disease" network diagram. Then DAVID database was used for GO gene enrichment analysis and KEGG metabolic pathway analysis. Finally, in vivo experiments were also performed to validate the findings of network pharmacology.Results: A total of 98 active components of RA combined with PRR were finally retrieved and integrated into the TCMSP database, 1700 target genes related to OP were obtained through the disease database, 149 gene targets were obtained by taking the intersection of disease genes, and drug targets, 122 core targets and 514 interaction relationships were obtained after protein interaction network and topology analysis. GO analysis and KEGG pathway enrichment analysis showed that RA combined with PRR intervention for OP mainly through multiple pathways such as PI3K-Akt, MAPK, TNF, Rap1, and Toll-like receptors. In addition, in vivo experiments confirmed that RA combined with PRR could significantly increase bone mineral density, reduce bone spacing, improve bone tissue structure, and improve osteoporosis in ovariectomized rats.Conclusion: In this study, the network pharmacological approach was used to reveal the potential targets and key signal pathways of RA combined with PRR in treating osteoporosis. This study was also verified by animal experiments, which provided a reliable basis for clinical application.


Introduction
Osteoporosis (OP) is a metabolic bone disease known as "silent disease." It is mainly a systemic metabolic disease prone to fracture due to the loss and reduction of bone mass, the destruction of bone microstructure, and the increase of bone fragility [1]. With the aging of the world population, the incidence of OP has been on the rise. OP has brought a huge negative impact on patients' health, leading to a sharp decline in the quality of life of middle-aged and older adults. OP is mainly due to the reduction of estrogen levels after menopause, resulting in insu cient estrogen secretion in the body, so the probability of women suffering from this disease in China is higher than that in men [2]. Its incidence is closely related to age and gender, and reports have shown that the probability of osteoporosis in the elderly over 60 years of age in China is 36%, including 23% in males and 49% in females, which indicates that osteoporosis has become an important public health problem [3]. At present, for patients with osteoporosis, western medicine treatment is mainly based on diet and exercise combined with bone health supplements, bone resorption, and osteoclast inhibitors, and sex hormone supplements at the same time [4]. However, these drugs have signi cant limitations in treating osteoporosis, such as high price, signi cant side effects, and low tolerance, so it is di cult to achieve the expected curative effect of patients. Traditional Chinese medicine has a long history of treating osteoporosis. It has the advantages of minor toxic and side effects, moderate price, syndrome differentiation, and overall treatment according to the patient's condition and speci c treatment [5]. The theory of traditional Chinese medicine believes that the etiology and pathogenesis of osteoporosis are related to the de ciency of the spleen and kidney [6]. The treatment of osteoporosis is mainly based on the main principles of tonifying the kidney and strengthening the bone, invigorating the spleen, and nourishing the blood. In the process of traditional medical treatment, the compatibility of drug pairs is often carried out to achieve the purpose of adapting to complex conditions, enhancing the e cacy of drugs, and reducing the toxic and side effects of drugs. RA combined with PRR signi cantly improves osteoporosis, which is better than using a single medicine [7].
RA is a traditional Chinese medicine that was rst recorded in "Sheng Nong's herbal classic." It is the dried root of Astragalus membranaceus (Fisch.) Bge. var. mongholicus (Bge.) Hsiao or Astragalus membranaceus (Fisch.) Bge. It has the effects of invigorating the spleen to bene t the lung, generating uid and nourishing blood, astringeting sores, and generating muscles. It is often used to treat clinical diseases such as spleen and lung qi de ciency and blood de ciency. Modern clinical pharmacological studies show that RA has the functions of protecting the kidney, anti-aging, anti-in ammatory, enhancing immunity, preventing and treating osteoporosis, antioxidant stress, and so on[8]. PRR is a processed product of Rehmannia glutinosa Libosch. It has the effects of nourishing blood and Yin, bene ting essence and lling marrow. Clinically, it is often used to treat liver and kidney yin de ciency, blood de ciency, waist and knee weakness, etc. According to the results of modern pharmacology and clinical research, PPR has the effects of preventing and treating osteoporosis, regulating immunity, anti-aging, and anti-anxiety [9].
Traditional Chinese medicine has the characteristics of multiple components, multiple targets, and multiple pathways, which makes it di cult to clarify the material basis and mechanism of action of traditional Chinese medicine. The lack of a scienti c evaluation system greatly limits the development of Chinese medicine. The holistic and systematic characteristics of network pharmacology are similar to the holistic view of traditional Chinese medicine and the principles of syndrome differentiation and treatment. It embodies the comprehensive effect of traditional Chinese medicine with multiple targets, multiple components, and multiple pathways. It has been widely used to reveal the mechanism of action of traditional Chinese medicine. In this paper, network pharmacology research methods were used to construct a multi-level network diagram to explore the potential mechanism of RA combined with PRR in treating osteoporosis and further provide a new method and reference for clinical treatment of osteoporosis. The system ow chart is shown in Fig. 1.

Screening of Active Ingredients and Targets
The Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, http://tcmspw.com/tcmsp.php) contains 499 herbal medicines and more than 29,000 compound components of each herbal medicine. In this study, the effective ingredients of the "RA combined with PRR" drug pair were retrieved and collected through the pharmacological system database and traditional Chinese medicine analysis platform. The active ingredients that cannot be retrieved in the TCMSP database were supplemented by the SwissTargetPrediction database (http://www.swisstargetprediction.ch/) with "human" as the research species. Finally, all active ingredients were checked and integrated with RA and PRR components.

Acquisition of osteoporosis targets
GeneCards database (https://www.genecards.org/) is a database that includes the human genome, transcriptome, and proteome. The OMIM database (http://www.omim.org/search/advanced/geneMap) is a database on human genes and genetic disorders. In the GeneCards and OMIM comprehensive databases, "osteoporosis" was used as the search term to search the target genes related to osteoporosis diseases, remove duplicate genes, integrate all targets, and establish an osteoporosis disease gene database.

Drawing of Venn diagram
The target protein names searched in the TCMSP database were converted to gene ID names using the UniProt database (https://www.uniprot.org/), and the race was set to "Human." Drew a Venn diagram after obtaining the key targets for the treatment of osteoporosis.

Construction of "active ingredient-target" network
The targets of the intersection of drugs and osteoporosis were imported into the String database (https://string-db.org/) to obtain the interactive information of the target genes, and the species was set to "Homosapiens." Then the active ingredients were imported into Cytoscape 3.7.1 software to construct a "drug ingredient-target" network diagram.
GO gene analysis and KEGG pathway enrichment analysis DAVID database (https://david.ncifcrf.gov/) is a bioinformatics database, which most commonly performs gene function enrichment analysis. This study used the DAVID database to input the key targets of drugs and diseases, followed by Go gene enrichment analysis and KEGG metabolic pathway analysis.
After integrating and ranking the genes, they were visualized and processed with the ImaGP online mapping website.

Experimental animals
Sixty-six healthy SPF SD female rats were purchased from Chengdu Dashuo Laboratory Animal Co., Ltd., with a bodyweight of about 180 ~ 200g and certi cate number of SCXK (Chuan) 2020-030. The rats were adaptively fed for seven days in the animal room of the School of Pharmacy, Shaanxi University of Traditional Chinese Medicine before the experiment, randomly divided into six cages, with free diet and water, room temperature of about 25℃, the humidity of about 50%, and kept ventilated in the room. The

Preparation of Experimental Drugs
The decoctions of RA and PRR were prepared by the water decocting method and added 10 times of water to 200g RA and PRR to decoct for 50min and then ltered. Added 8 times of water to decoct for 40 minutes and combined the obtained ltrate (2g/ml crude drug). Since the positive drug alendronate is a tablet, 6 tablets need to be ground into a powder with a mortar, then dissolved in 666ml of pure water, and refrigerated at 4°C for later use.

Grouping and administration of experimental animals
Sixty-six SD female rats were randomly divided into 6 groups of 11 rats each. They were: RA group, PRR group, RA combined with PRR group, positive control group, model group, and sham operation group. On the 7th day after the completion of model establishment, intragastric administration was started. RA group, PRR group and RA combined with PRR group were infused with RA aqueous decoction and PRR aqueous decoction (5.4g·kg −1 ), RA combined with PRR aqueous decoction (RA 2.7g·kg −1 and PRR 2.7g·kg −1 , respectively), positive control group (alendronate 1.05mg·kg −1 ). The model group and sham operation group were intragastrically administered with an equal volume of pure water.

Preparation of Osteoporosis Model Rats
Ovariectomy castration modeling (back ovariectomy) was performed in all rats except the blank group. The rats were fasted but could not help but water 12 hours before modeling. After intraperitoneal injection of chloral hydrate anesthesia, the dorsal skin was performed. A longitudinal incision was made about 0.5-1.0cm beside the midline of the back to nd the ovaries and remove them. Penicillin was administered intramuscularly for 3 days postoperatively. The rats were given intragastric administration on the 7th day after the model was established, and the vaginal exfoliated cells were smeared and observed under a microscope on the 10th day.

Detection of Serum Markers
After blood sampling from the abdominal aorta, serum was obtained by centrifugation for serum bone metabolism markers β-CTX, BGP, MSTN, and PIIINP. The detection methods were determined according to the instructions of the ELISA kit. Use an automatic microplate reader to detect the absorbance at 450 nm, record the value, and then convert the content of the marker according to the standard curve.

Bone Mineral Density (BMD) Analysis
After 12 weeks of administration, animals in each group were weighed. Rats were anesthetized by intraperitoneal injection of sodium pentobarbital. After the scanning conditions were set, the lumbar vertebra and left tibia parameters were analyzed.

Pathological Changes in Rat Bone Tissue
The left femur of the rats was xed with 10% formalin and decalci ed with 10% EDTA. The specimens were embedded and cut into slices with a thickness of 4 µm. After staining with hematoxylin-eosin staining solution, it was observed under a microscope.

Screening results of active ingredients
In the TCMSP database, "RA or Huangqi, PRR or Shudi" was used as the search terms, respectively. The search conditions were: oral bioavailability (OB)≥30%, drug-like (DL)≥0.18. Finally, 21 main chemical components of RA and 77 chemical components of PRR were selected. The active ingredients that could not be retrieved in the TCMSP database supplemented their target information through the SwissTargetPrediction database. A total of 98 active ingredients were obtained by focusing and integrating the retrieved RA with all active ingredients of PRR, and the results are shown in Table 1 and  Table S1. Screening results of targets By searching the disease target genes in GeneCards, OMIM databases, 1108 and 592 disease targets of osteoporosis were obtained, respectively, integrated with ranking. A total of 149 intersections of RA combined with PRR drugs on genes and osteoporosis disease target genes were taken. The gene intersection was input in Microbial Information software to draw Venn diagrams of drug and disease targets, as shown in Fig. 2A.
Construction of the "active ingredient-core target" network The selected active ingredients of the drug and related targets of action were imported into Cytoscape3.7.1 software to construct a "drug-active component-target of action" network diagram. Among them, different colors are used to indicate that they come from different components. The green part in the gure is the target of RA combined with PRR, the purple part represents the active ingredients of RA, and the orange part represents the active ingredients of PRR. The area of the node indicates that the larger the degree value and the more critical the biological function of the node, as shown in Fig. 2B.

Protein interaction network (PPI) analysis
The 149 intersection targets of the interaction between the target genes of active components of RA and PRR and the target genes of osteoporosis were input into string software for protein interaction analysis. Then, the downloaded TSV le was imported into Cytoscape 3.7.1 software, and the "network analysis" function was used for network topology analysis. A total of 122 nodes and 514 edges were obtained. The average betweenness centrality, closeness centrality, and degree values are 0.0161, 0.3562, and 8.4262, respectively. With the help of this software, the drug active ingredient disease target network and PPI network diagram were analyzed to construct the intersection target PPI network diagram (Fig. 2C). The top 10 of the degree value is used as the core target of the network using the CytoHubba plug-in, as shown in Fig. 2D.

GO gene analysis and KEGG pathway enrichment analysis
The genes were integrated and ranked, and a total of 812 entries with p<0.05 were obtained after GO gene analysis and screening, including 644 biological process (BP) enrichment results, 47 cell composition (CC) enrichment results, and 121 molecular function (MF) enrichment results. The top 20 BP, CC, and MF genes were used to draw the bubble diagram ( Fig. 3A-C). Through KEGG pathway enrichment analysis, a total of 114 signaling pathways with p<0.05 were obtained after screening. The results of enrichment analysis showed that RA combined with PRR mainly involved multiple signaling pathways such as PI3K-Akt, MAPK, TNF, Rap1, and Toll-like receptor in the treatment of osteoporosis, are shown in Fig. 3D.

Changes of blood biochemical indicators
Compared with the sham operation group, the BGP level of the model group was signi cantly reduced, while β-CTX, MSTN, and PIIINP were signi cantly increased (P<0.01). Compared with the model group, the BGP level of the RA and PRR groups was signi cantly increased (P<0.05), the levels of β-CTX, MSTN, and PIIINP were signi cantly reduced (P<0.05). It showed that RA combined with PRR could improve the abnormality of bone formation markers of osteoporosis (Tab 2 and Fig. 4).

Pathological changes of bone tissue in rats
The femurs of rats in the sham operation group were mesh-like and neatly arranged, and the bone microstructure was complete. Compared with the sham operation group, the reticular structure of the femur in the model group was destroyed. The bone cortex of the femur of the model group became thin, and the space in the bone marrow cavity increased, indicating that the model group had osteoporosis. Compared with the model group, the number of bones in the RA and PRR groups was signi cantly increased, the bone tissue structure was signi cantly improved, and only a few were unevenly arranged (Fig. 5).

Detection of bone mineral density in rats
Compared with the sham operation group, the BMD of the model group's left tibia and lumbar spine was signi cantly reduced (P<0.01). Compared with the model group, the BMD of rats in the RA, PRR, and RA combined with the PRR group were signi cantly higher (P<0.05). It was suggested that RA combined with PRR treatment could improve the symptoms of osteoporosis in ovariectomized rats (Tab 3 and Fig. 6).

Discussion
Osteoporosis is a systemic bone disease. Due to osteopenia and deterioration of bone microstructure, the risk of fragility fractures increases [10]. The results of modern clinical medical research show that multiple factors cause the pathogenesis of osteoporosis, and the imbalance of bone remodeling is considered to be the main mechanism leading to osteoporosis. Growth hormone activity, musculoskeletal function, daily dietary calcium and vitamin D intake, genetics, and environment can affect bone remodeling and break the bone formation-resorption balance [11]. The myostatin signaling pathway increases bone resorption by activating the RANKL signaling pathway [12].
The results showed that the main active ingredients of RA combined with PRR in treating osteoporosis were mistletoe, kaempferol, isorhamnetin, verbascoside, diosgenin, and catechins. Kaempferol is a avonol compound with biological effects such as anti-oxidation, anti-tumor, and immune regulation. It can promote the proliferation and differentiation of osteoblasts and increase osteoblast alkaline phosphatase's activity to promote bone formation [13]. Studies have shown that high-dose kanamycin can treat osteoporosis by regulating the balance of Ca2+ metabolism, promoting collagen production, and reducing bone loss [14]. Mistletoe is a kind of polyhydroxy avonoids, which can effectively promote osteoblast differentiation in MC3T3-E1 cells and inhibit osteoclastogenesis in RAW264.7 cells [15]. At the same time, mistletoe can also promote the proliferation and bone differentiation of bone marrow stem cells [16].
PPI network topology analysis results showed that STAT3, Jun, SRC, AKT1, Mapk14, mapk1, TNF, IL6, and other proteins were the main targets of RA combined with PRR in the treatment of osteoporosis. JUN protein is a family of protein kinases in the mitogen-activated protein kinase (MAPK) signal transduction cascade, and c-Jun is also an essential substance for activating transcriptional activator protein 1 (AP-1) [17]. Studies have shown that the increase of JUN protein level can initiate more AP-1, AP-1 can affect the apoptotic process of cells, promote apoptosis, and accelerate the differentiation of osteoclast precursor cells into osteoclasts by regulating the expression of related genes [18]. In addition, it can also increase the activity of metalloproteinases to promote the degradation of collagen in human skin [19][20]. Src proteins are a class of non-receptor protein tyrosine kinases activated by various extracellular signaling molecules [21]. Many studies have shown that the high activation of Src protein is closely related to the occurrence and development of osteoporosis. Its inhibitors are expected to be a critical method for treating osteoporosis drugs [22]. AKT1 is a member of the protein kinase (B) family and is necessary for cell proliferation, cell growth, and differentiation in the body. The researchers found that mice with the AKT1 gene knocked out had skeletal muscle atrophy and impaired bone development. In recent years, TNF and IL6 factors have been considered the main factors regulating bone resorption under pathological conditions, and they have a close relationship with estrogen levels. Many experiments in vitro and in vivo have con rmed that IL-6 can elevate the formation of osteoclasts and increase bone resorption [23][24]. TNF indirectly activates mature osteoclasts by stimulating osteoblasts and inhibits the apoptosis of osteoclasts [25].
KEGG pathway enrichment analysis results showed that RA combined with PRR treatment of osteoporosis mainly involves PI3K-Akt, MAPK, TNF, Rap1, Toll-like receptors, HIF-1, Ras signaling pathways. PI3K/Akt pathway is closely related to osteogenesis and osteoclast pathways. The PI3K/Akt pathway can regulate the differentiation and apoptosis of osteoblasts and osteoclasts to maintain the balance of bone resorption and bone formation [26]. Decreasing the activity of caspases-9 through the PI3K-Akt pathway can weaken the absorptive capacity of osteoclasts and reduce the apoptosis caused by cell damage, thereby delaying the process of osteoporosis [27]. The MAPK pathway is directly involved in the regulation of bone metabolism. It has a complementary relationship with the PI3K/Akt pathway [28] and plays a vital role in promoting osteoblast growth and differentiation. MAPK signaling pathways mainly include extracellular regulated protein kinase (ERK) transduction pathway, Jun N-terminal kinase (JNK), ERK5/macro lament mitin-activated protein kinase transduction pathway, and p38 signaling pathway. The JNK pathway mainly affects bone formation [29]. It has been shown that fucoidan promotes the osteogenesis of human bone marrow stem cells by increasing phosphorylation-induced BMP2 expression and stimulating the activation of ERK, JNK, and p38 signaling pathways [30]. The p38 pathway and ERK signaling pathway can promote the expression of osteoprotegerin, which is the key to maintaining the balance of bone metabolism [31]. Tumor necrosis factor (TNF) is the main cytokine regulating osteoclast activity and bone resorption [32]. Researchers isolated the glycoprotein osteocalcin (OPG) from TNF factors through fetal mouse gene sequencing experiments, which can be used as a soluble factor to regulate bone mass [33].
In addition, this study established a rat model of osteoporosis by removing the rat's ovaries. The mechanism of RA combined with PRR in the treatment of osteoporosis was further explored through experimental methods such as dual-energy X-ray absorption assay, enzyme-linked immunosorbent assay, and HE staining method. Experimental results showed that RA combined with PRR could promote bone formation and increase BMD. The experimental results showed that the bone formation marker BGP was signi cantly increased compared with the model group, and the levels of β-CTX, MSTN, and PIIINP were signi cantly decreased. It showed that RA combined with PRR could increase bone formation markers to promote bone formation and have a therapeutic effect on osteoporosis. Pathological examination results showed that the bone tissue of the model group was disordered, and the bone mass was reduced. RA combined with PRR could signi cantly improve the microstructure of bone tissue in osteoporotic rats.

Conclusion
In summary, this article used network pharmacology technology combined with in vivo experimental veri cation to screen out the active ingredients, key targets, and signal pathways of RA combined with PRR in treating osteoporosis. This article analyzed the mechanism of RA combined with PRR in treating osteoporosis from the perspective of multiple targets, multiple pathways, and multiple components. It