Pharmacodynamic Material Basis and Potential Mechanism Study of Spatholobi Caulis in Reversing Osteoporosis

Objective To elucidate the mechanism of Spatholobi Caulis (SC) in treating osteoporosis (OP) integrated zebrafish model and bioinformatics. Methods Skeleton staining coupled with image quantification was performed to evaluate the effects of SC on skeleton mineralization area (SSA) and total optical density (TOD). Zebrafish locomotor activity was monitored using the EthoVision XT. Bioactive compounds of SC and their corresponding protein targets were acquired from Traditional Chinese Medicine Systems Pharmacology (TCMSP) database. Potential therapeutic targets for OP were summarized through retrieving 5 databases, and then, the overlapping genes between SC and OP were acquired. The core genes were selected by CytoHubba. Subsequently, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and Gene Ontology (GO) functional analysis of the intersection target genes were carried out by R software. Finally, the molecular docking simulation was manipulated between the ingredients and the hub genes. Results Compared with the model group, SC significantly increased the SSA and TOD at 10 mg/mL and improved the locomotor activity in a dose-dependent manner (p < 0.001). 33 components of SC were associated with 72 OP-related genes including 10 core genes (MAPK1, VEGFA, MMP9, AKT1, AR, IL6, CALM3, TP53, EGFR, and CAT). Advanced Glycation End Product (AGE) Receptor for AGE (RAGE) signaling pathway was screened out as the principal pathway of SC in anti-OP. The bioactive components (Aloe-emodin, Emodin, Formononetin, Licochalcone A, Luteolin, and Lopac-I-3766) have excellent affinity to core genes (MAPK1, VEGFA, MMP9, AKT1, and IL6). Conclusion SC had the hierarchical network characteristics of “multicomponents/multitargets/multifunctions/multipathways” in reversing OP, but AGE-RAGE signaling pathway may be the main regulatory mechanism.


Introduction
Osteoporosis (OP) is a common systemic bone disease with a signifcant impact on human health and characterized by low bone quality, low bone density, and bone microstructure destruction, which easily leads to decreased bone strength and increased fracture risk [1][2][3]. OP has been considered a silent disease, for it is usually asymptomatic and cannot be detected until the frst fracture occurs [4][5][6]. According to statistics released by international osteoporosis foundation, OP worldwide causes more than 8.9 million fractures annually, and this number is projected to be tripled by 2050 [7,8]. Te increasing incidence of OP places a heavy social and economic burden on the health care system [9][10][11]. Despite the fact that substantial pharmacological options have been approved for the management of OP, the potential long-term adverse efects of anti-OP medications and poor adherence to the treatment urgently await more efective strategies to satisfy the unmet medical needs [12][13][14][15]. Alternative natural medicine may play an emerging role in against OP by complementarily overcoming the limitations of the current marketed medications [16][17][18].
As a new interdisciplinary developed in recent years, network pharmacology provides a "multiway," "multitarget" method for drug analysis and is similar with the thoughts of traditional Chinese medicine (TCM) compound prescription, which concurs with the characteristics of the overall compatibility of TCM and comprehensive and multipathway treatment of diseases [35]. Recently, numerous studies illustrated that network pharmacology can reach high performance on prediction the mechanism of Chinese medicine, indicating that it might become a powerful tool to systematically explore the new application of TCM [36,37]. In this study, prednisolone-(PNSL-) induced zebrafsh OP model was employed to assess the therapeutic efect of SC. Subsequently, we adopted network pharmacology to explore the underlying mechanism of SC's antiosteoporotic efects. Te major bioactive components of SC and corresponding targets were obtained, and "drug-active components-diseasetargets-signaling pathways" network was constructed. Ultimately, the potential mechanism of SC in the treatment of OP was revealed via a combination of Gene Ontology (GO) function analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis.

Experimental Animal.
Zebrafsh larvae were purchased from Nanjing Yishu Lihua Biotechnology Co., Ltd and cultured in blank E3 medium (containing 0.33 mM CaCl 2 , 0.33 mM MgSO 4, 5 mM NaCl, and 0.17 mM KCl). Zebrafsh embryos were cultured in Intelligent light incubator with a 14/10 h light/dark cycle [38,39]. Animal experiments were carried out in accordance with the Guidelines for Animal Experimentation of Jiangsu University (Zhenjiang, China), and the protocol was approved by the Animal Ethics Committee of this institution.

Observation of Zebrafsh Locomotor
Activity. After 4-day drug deliveries, zebrafsh locomotor activity was monitored by EthoVision behavior system [42,43]. Te instrument parameters are set as follows: "prior to the start of tracking, the software needed to be calibrated; the video sampling rate was 25 frames per second (fps), based on the design recommendations [44]; frst, under the Trial List, one trial was selected. For the Arena Settings, each well/arena was calibrated based on the diameter of the well [45]. Te diameter of the wells within the well plates used in this manuscript (96-well plate) is 6.54 mm [46]. For the Detection Settings, dynamic subtraction was selected, and the dark contrast and subject contour were adjusted to optimize tracking efciency [47,48]. Within the Analysis Profles, the selected dependent variables were distance moved, velocity, and time spent moving [49]. Tese endpoints were based on the larvae's center-point activity [50]. Te results were then exported to Excel and statistical analysis software suites." According to the methods described above, moving distance (MD), moving speed (AS), travel frequency (TF), and hotspot are selected as the anti-OP drug efciency evaluation indexes in this model.

Screening of Active Compounds and Corresponding Target
Genes in SC. Traditional Chinese Medicine Systems Pharmacology database (TCMSP) was utilized to search the bioactive components of SC, with the screening conditions of "oral bioavailability (OB) ≥20% and drug like (DL) ≥0.1." Meanwhile, the corresponding target genes of the above components were obtained [56,57].

Overlapping Targets of Drug Potential Genes and OP-Related
Genes. R language [63] was applied to obtain the overlapping targets of drug potential genes and OP-related genes. Tese common target genes were used for further analysis.
2.9. GO and KEGG Analysis. R language and Bioconductor platform were used to perform GO enrichment analysis and KEGG pathway analysis. Te frst 10 functional categories of biological process (BP), cellular component (CC), and molecular function (MF) were screened out to construct the histogram and bubble diagram. Ten, the target proteins were analyzed using the KEGG enrichment analysis, and the top 30 signaling pathways were chosen to draw the histogram and bubble diagram.

Network Visualization of "Active Ingredients-Potential
Targets-Signaling Pathways". Cytoscape software was adopted to draw a ternary network including bioactive ingredients-target genes-signaling pathways. Te above bioactive components and corresponding genes were imported into the software, and the ternary network "drugactive components-potential therapeutic targets of diseaserelevant signaling pathways" was drawn [67,68].
2.11. Molecular Docking Analysis. Te plug-in CytoHubba of Cytoscape software was used to screen top 10 hub genes. Te core genes were selected to fnd the related drug components in the compound regulatory network as small molecular ligands. Molecular docking simulation was performed as previously described. Te specifc process is as follows: "the 2D structure information of drug chemical components was downloaded from PubChem (https://www.ncbi.nlm.nih. gov/) platform and converted into the 3D structure by ChemBio3D software, and the energy optimization of MM2 was carried out to complete the preparation of small molecule ligands. Te 3D structure of the candidate target proteins was downloaded from PDB (https://www.rcsb.org/) database, and then, the protein receptors were prepared after the water molecules, and ligands were removed by PyMOL2.4.0 software. AutoDockTools software was used to read the receptor fles, which were converted to PDBQT format after hydrotreating ion modifcation. Te ligand fles were also converted to PDBQT format for saving and then converted into the 2D structure to draw the active pockets. Finally, AutoDock vina software will be used for molecular docking, and the lowest free energy model is selected for visual analysis." 2.12. Statistical Analysis. Te GraphPad Prism 5 software [69][70][71]was used for statistical analysis, and the data was expressed as the mean ± SD. Te diferences were performed by one-way analysis of variance, and p < 0.05 was considered signifcant diference [39,72,73].

Protective Efect of SC on the Zebrafsh OP Model Induced by PNSL.
At the 3DAF, zebrafsh were incubated with PNSL for three days in 6 well plates to establish a rapid OP model. Te SSA and TOD, which represented osteoblast diferentiation, were used to determine the amount of bone mineralization of zebrafsh larvae. SSA and TOD of the zebrafsh larvae were signifcantly lower in the PNSL-treated groups than in the DMSO with values of 51423. 33 and 19405.03, respectively, indicating that PNSL reduced bone mineralization and inhibited osteogenic diferentiation in zebrafsh larvae ( Figure 1). When PNSL-induced zebrafsh was treated with 0.1, 1, and 10 mg/mL SC, higher mineralization of the vertebrate column was discovered in a dose-dependent pattern. SSA and TOD were signifcantly increased to 295764.70 and 109542.50 at 10 mg/mL, respectively, as shown in Figure 1. In conclusion, the study indicated that SC could reverse the bone loss of zebrafsh induced by PNSL.

Protective Efect of SC on Zebrafsh Locomotor Activity.
Te behavior analyzer EthoVision was used to track the movement of zebrafsh. As shown in Figures 2(a)-2(c), the Moving speed (MS), moving distance (MD), and travel frequency (TF) of PNSL group were signifcantly lower than those of the CON, while for SC treatment group, these parameters were close to the DMSO. Te results showed that diferent concentrations of SC increased MD by 109.46%, 79.91%, and 47.27%, respectively, with partial signifcance (p < 0.001 − 0.01). TF for SC-H, SC-M, and SC-L was 371.43%, 254.54%, and 159.62%, respectively. Also, the hot plot revealed zebrafsh activity degree (Figure 2  Evidence-Based Complementary and Alternative Medicine

Main Components of SC and Treatment Targets of OP.
68 active components in SC were obtained through retrieving the TCMSP database, and after screening by "OB ≥20% and DL ≥0.1," 33 chemical bioactive components of SC were obtained (Table 1). 609 corresponding genes were collected by UniProt gene annotation simultaneously. Trough PharmGkb, TTD, DrugBank, GeneCards, and OMIM, 3150 OP-related targets were collected (Figure 3(a)). By matching the target of SC with OP-related targets, 72 common genes were derived, which were potential targets of SC in treating OP (Table 2, Figure 3(b)).

Construction of PPI Network and Acquisition of Core
Genes. Te 72 obtained intersecting genes were imported into the STRING database to obtain the interaction relationships between them and save the data as a fle in TSV format ( Figure 4). Secondly, we visualized the PPI network by importing it into the Cytoscape software and identifed 10 core genes using plug in-CytoHubba of Cytoscape software, namely, MAPK1, VEGFA, MMP9, AKT1, AR, IL6, CALM3, TP53, EGFR, and CAT ( Figure 5).

GO and KEGG Enrichment Analysis.
In order to further elucidate the potential relationship between common genes and the mechanism by which the SC might remedy OP, R language was applied to perform GO and KEGG analysis. A total of 2049 GO terms were collected, of which 1892 are biological process (BP) entries, 106 molecular function (MF) entries, and 51 cellular component (CC) entries. As shown in Figure 6, the top 10 BP terms were principally relevant to the response to steroid hormone, response to metal ion, cellular response to chemical stress, response to drug, gland development, regulation of DNA-binding transcription factor activity, response to oxygen levels, reactive oxygen species metabolic process, muscle cell proliferation, and rhythmic process. Te MF analysis indicated that these gene targets were mainly related to nuclear receptor activity, ligandactivated transcription factor activity, cytokine receptor binding, steroid hormone receptor activity, DNA-binding transcription factor binding, and RNA polymerase IIspecifc DNA-binding transcription factor binding. Te CC entries were mainly related to membrane microdomain, membrane raft, membrane region, caveolae, and plasma membrane raft.
Moreover, a total of 139 signaling pathways were yielded through KEGG enrichment analysis, including Bladder cancer, Lipid and atherosclerosis, Kaposi sarcomaassociated herpesvirus infection, AGE-RAGE signaling pathway, IL-17 signaling pathway, Hepatitis C, Human cytomegalovirus infection, Estrogen signaling pathway, Breast cancer, and TNF signaling pathway. As shown in Figure 7, we visualized the top 30 fltered pathways using bubble plot and bar diagram. Among these signaling pathways, AGE-RAGE signaling pathway was chosen to further research ( Figure 8); in the picture, red nodes mean Evidence-Based Complementary and Alternative Medicine     Figure 4: Protein-protein interaction network (PPI). Te PPI network from STRING was further analyzed using Cytoscape software (the line between two nodes indicates the interaction; the darker the color of the node, the more the relationship between them).  the genes in the road map present in the SC network in this study. Finally, a ternary network of "active ingredients-potential targets-signaling pathways" was drawn successfully (Figure 9).    Figure 8: Te main nodes of AGE-RAGE signaling pathway. Red nodes mean the genes in the road map present in the SC network in this study. Figure 9: Te network of "drug-active ingredients-target genes -pathways -disease." Yellow triangle represents drug; hexagon node represents the ingredients of SC. Blue square node represents potential treatment targets; red V node represents potential OP-related pathways; light blue node represents OP.   It is generally considered that the docking afnity is stronger when the binding energy is less than-5.0 kcal/mol, and the docking activity is extremely strong when the binding energy is less than-7.0 kcal/mol." From the docking results (Tables 3 and  4), it was found that AKT1-luteolin, IL6-formononetin, IL6lopac-I-3766, MAPK1-emodin, MMP9-luteolin, and VEGFA-lopac-I-3766 had the lower binding energies. It is speculated that SC mainly exerted anti-OP efect through above molecular docking process. Eventually, we chose above protein targets and active component with the lowest docking afnity for visualization ( Figure 10).

Discussion
OP is a metabolic disease caused by various factors, such as age, endocrine, viral infection, and a variety of cytokines. Given that the multitarget and multipathway pharmacological properties of TCM make it more suitable for the treatment of OP, SC is a widely used traditional medicine for promoting blood circulation to remove blood stasis and nourishing blood. Modern pharmacological studies revealed that SC exerted anti-OP through inhibition of osteoclastogenesis and stimulation of chondrogenesis. However, the underlying mechanism remains unclear. Te aim is to clarify the potential mechanisms of SC for OP treatment integrated zebrafsh model and network pharmacology.
Zebrafsh is an ideal animal model in vivo for studying bone deformations for the high similarity of the structure and genetics of bone with human beings [74]. In the present study, we established the zebrafsh OP model induced by PNSL to evaluate the efects of SC on the bone formation and examined the bone mineralization by Alizarin red staining, which is a vital dye staining, and the biological mineralization process of zebrafsh can be observed directly. Our results indicated that PNSL exposure inhibited osteogenic diferentiation and bone mineralization in zebrafsh larvae (p < 0.05; Figures 1(a) and 1(b)). Compared with DMSO group, treatment with PNSL at 25 μmol/L caused an apparent decrease in the SSA and TOD with values of 51423. 33 and 19405.03, respectively, which corresponded to the ventral view of Alizarin Red stained zebrafsh skull, as shown in Figure 1(c). After the intervention of SC, TOD and SSA were increased to 295764.70 and 109542.50 at 10 mg/mL, respectively (Figures 1(a) and 1(b)). Besides, SC increased MD by 131.50%, 85.34%, and 69.60%, with partial signifcance (p < 0.001 − 0.01). TF for SC-H, SC-M, and SC-L was 371.43%, 254.54%, and 159.62% (Figures 2(a)-2(c)). Also, the hot plot revealed zebrafsh activity degree (Figure 2(d)). Tese results demonstrated that SC played a positive efect on the function of osteoblastic diferentiation and mineralization in zebrafsh larvae.
Behavioral changes of zebrafsh have been linked to chemical exposure [75,76]. Te behavior analyzer Etho-Vision XT made it possible to examine numerous motor events and facilitated the quantitative analysis of behavior [77]. Te behavioral change of zebrafsh is an essential indicator to assess the anti-OP efect of SC. Te MD and TF in PNSL group further supported that the construction of  Tese results with signifcant diferences demonstrated that SC could improve dyskinesia of zebrafsh to some extent. By evaluating the pharmacokinetic characteristics of the chemical ingredients in SC, we identifed 33 bioactive components and determined 72 potential therapeutic targets of OP. Tese common targets of SC were related to the regulation of diverse biological activities, including response to steroid hormone, response to metal ion, cellular response to chemical stress, response to drug, regulation of DNAbinding transcription factor activity, response to oxygen levels, and muscle cell proliferation. Tese results were consistent with those of the previously reported pathological processes of OP [78][79][80]. Pathway enrichment analysis indicated that the targets of SC were enriched in diverse pathways implicated in OP pathology, such as AGE-RAGE, lipid and atherosclerosis, IL-17, Estrogen, and TNF signaling pathway.
In this study, 10 core genes (MAPK1, VEGFA, MMP9, AKT1, AR, IL6, CALM3, TP53, EGFR, and CAT) were identifed using plug-in CytoHubba in Cytoscape. Subsequently, we found that fve of the top 10 hub genes (MAPK1, VEGFA, MMP9, AKT1, and IL6) were enriched in AGE-RAGE signaling pathway. MAPK1, generally known as ERK2, is a member of mitogen-activated protein kinase family and acts as an essential regulator of cell proliferation, diferentiation, infammation, and bone metabolism. It has been demonstrated the activation of MAPK1 could promote osteoblastogenesis and bone mineralization in zebrafsh larvae. VEGFA is a signal factor promoting neovascularization and increase vascular permeability. A recent research reported that VEGFA was a latent marker of endothelial dysfunction in postmenopausal osteoporosis. Moreover, through directly targeting VEGFA, MIR-16-5P mitigated the symptom of postmenopausal women with osteoporosis. MMP9 is a member of matrix metalloproteinases family, responsible for extracellular matrix degradation and cleavage of its structural components. A previous study illustrated that MMP9 exerted a crucial role in pathogenesis of osteoporosis, and the inhibitory efect on bone resorption was emerged by inhibiting the expression of MMP9. AKT1, also called Protein Kinase B, is a crucial signal transducer of PI3K/AKT signaling pathway. Te phosphorylation of AKT1 modulates the expression of multiple downstream efectors including mTOR-C1 and FOXO3 proteins and then mitigates OP induced by iron overload. Based on the above analysis results, it is also speculated that SC against OP may play a role through the above process.
GO and KEGG enrichment analysis results showed that the therapeutic targets of SC for diseases mainly enriched in Bladder cancer, Lipid and atherosclerosis, AGE-RAGE signaling pathway, Kaposi sarcoma-associated herpesvirus infection, IL-17 signaling pathway, Hepatitis C, Human cytomegalovirus infection, Estrogen signaling pathway, TNF signaling pathway, and breast cancer. Studies have shown that all of these play a crucial role in the progression of OP. We searched for OP in the KEGG pathway database and found that there are mainly two pathways directly related to OP and related diseases, including AGE-RAGE signaling pathway and Estrogen signaling pathway. Subsequently, we found that fve of the top 10 hub genes (MAPK1, VEGFA, MMP9, AKT1, and IL6) were enriched in AGE-RAGE signaling pathway. Te AGE-RAGE signaling pathway plays a crucial role in bone remodeling process. In general, AGEs not only induce osteoclastogenesis by upregulation of RANKL mRNA, but they also afect osteoblasts by suppressing cell growth, promoting apoptosis, and downregulating diferentiation, which impairs mineralization. It was reported that AGEs upregulated the expression of RAGE in human MSCs and then AGEs interacted with RAGE and increased the expression of TGF-b mRNA, which suppressed bone mineralization and destroyed bone remodeling [33]. In this study, fve of the top 10 hub genes were enriched in the AGE-RAGE signaling pathway, so we speculated that the underlying mechanism of SC on OP might be mainly the inhibition of AGE-RAGE signaling pathway.

Conclusion
Compared with model group, SC signifcantly increased the SSA and TOD at 10 mg/mL and improved the locomotor activity in a dose-dependent manner (p < 0.001). 33 components of SC were associated with 72 OP-related genes including 10 core genes. AGE-RAGE signaling pathway was screened out as the principal pathway of SC in anti-OP. Te bioactive components (Aloe-emodin, Emodin, Formononetin, Licochalcone A, Luteolin, and Lopac-I-3766) have excellent afnity to core genes (MAPK1, VEGFA, MMP9, AKT1, and IL6). Conclusion. SC had the hierarchical network characteristics of "multicomponents/multitargets/ multifunctions/multipathways" in reversing OP, but AGE-RAGE signaling pathway may be the main regulatory mechanism. We have demonstrated the anti-OP efect of SC and revealed its underlying mechanism by adopting zebrafsh model and network pharmacology. SC mainly regulated AGE-RAGE signaling pathway to exert anti-OP therapeutic efect.

Data Availability
Te original contributions presented in the study are included in the article andsupplementary materials. Further inquiries can be directed to the corresponding authors.

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