Network pharmacology and UPLC-Q-TOF / MS studies on the anti-arthritic mechanism of Pterocephalus hookeri

Purpose: To investigate the mechanism underlying the anti-arthritic properties of Pterocephalus hookeri used for treatment of rheumatoid arthritis (RA). Methods: Aqueous methanol extract of P. hookeri was analyzed using UPLC-Q-TOF/MS, a Waters Acquity UPLCR BEH C18 column (2.1 × 100 mm, 1.7 μm) and gradient elution with acetonitrile-formic acid-water. Targets and related pathways were predicted by PharmMapper database and Molecule Annotation System, respectively. The network was built with Cytoscape software. Results: Forty compounds were identified, comprising 17 iridoid glycosides, 7 phenolic acids, 13 triterpenes, and 3 other compounds. A total of 38 targets and 44 pathways associated with RA were obtained. These involved mainly MAPK signaling pathway, adherens junction, and colorectal cancer. Conclusion: These results from network pharmacology suggest that P. hookeri exerts therapeutic effect on RA via multiple components, multiple targets and multiple pathways.


INTRODUCTION
Rheumatoid arthritis (RA) is a systemic autoimmune disease which occurs in three stages: chronic, progressive, and aggressive arthritis, eventually leading to joint deformity and loss of function.Thus, RA has a high frequency of disability [1].In Tibetan medicine, RA is called "zhen bu" disease [2], and available data show that the treatment efficacy of Tibetan medicine for RA is up to 94.6 % [3,4].
Previous studies have confirmed that the major chemical components of P. hookeri include iridoid glycosides, triterpenoid saponins, and phenolic acids [5,6].A survey of ancient literature revealed that P. hookeri was frequently used in the treatment of RA [7].Some recent studies have shown that aqueous and ethanolic extracts of P. hookeri have anti-inflammatory, anti-RA, and analgesic effects [8,9].Although there are several studies on P. hookeri, its mechanism of action on RA has not received much attention.

Recently,
network-based analyses have emerged as powerful tools for elucidating the multiple and active components of extracts, as well as their mechanisms of action [10].A new approach is provided by network pharmacology for the study of activities of multiple components and pharmacological mechanisms in traditional medicine.
The purpose of the research was to use network pharmacology and UPLC-Q-TOF/MS to unravel the active ingredients of P. hookeri, their targets, and the mechanism involved in the anti-RA effect of the plant.

Sample preparation
Powered P. hookeri (2.0 g) was accurately weighed and added to 50 mL of 70 % aqueous methanol.The mixture was subjected to ultrasonic extraction for 30 min, and filtered.The residue was washed with a small amount of 70 % aqueous methanol, and the combined filtrate was concentrated and dissolved in the same solvent.The concentrated solution was transferred to a 10 mL bottle, and 70 % methanol was added to the mark.The supernatant was centrifuged at 14000 rpm for 15 min and filtered through a 0.22 m microporous membrane.

UPLC chromatography
UPLC was performed in a 100 mm  2.

Mass spectrometry
Waters SYNAPT G2HDMS system of ion source was used for electrospray ionization (ESI).Scanning was done at positive (ESI+) and negative (ESI-) ion modes, with nitrogen as atomization and conical gas.The source temperature and cone gas flow rate were 100 C and 40 L/h, respectively.Desolvation temperature and gas flow rate were 350 C and 800 L/h, respectively.Other MS conditions used were sampling cone voltage of 40V, extraction cone voltage of 4V, capillary voltages of 3.0 kV (ESI+) and 2.5 kV (ESI-), scan time and inter scan time (0.3 s and 0.02 s, respectively), and mass-to-charge ratio, m/z of 50 -1700.Leucineenkephalin (0.5 g/mL) at a flow velocity of 5 L/min, was used for calibration of mass number (m/z 556.2771 for ESI+, and m/z 554.2615 for ESI-).

Prediction and screening of targets
ChemBio Office 2014 software was used to draw the structures of the compounds.These were converted to 3D forms with ChemBio3D ultra software, and stored in mol2 format.Then, in order to predict the potential target, chemical components were imported into the PharmMapper website (http://lilab.ecust.edu.cn/pharmmapper/) for potential target prediction analysis.The first 10 targets of each compound were selected for follow-up study.

Prediction and screening of pathways
The targets obtained were introduced into the Bio database (http://bioinfo.capitalbio.com/mas3/) and then screened for pathways that met the criterion of p < 0.01.

Construction of network
Cytoscape software was used to construct compound-target-pathway networks with chemical constituents, predicted targets and pathways.

Compositions of P. hookeri
Data on the ESI+ and ESI-modes are shown in Figure 1.It was found that ESI-mode provided a neater MS fragmentation and sharper peak shapes than ESI+.However, it was easier to find the quasi-molecular ion peaks of compounds with combination of positive and negative ions.
Using relative retention times, exact masses, MS fragments, standards, and references, 40 compounds were identified or elucidated with the information shown in Figure 1, Figure 2, Table 1.
Compounds of iridoid glycosides have variety of activities, such as anti-inflammatory [8], anti-RA [2], and the anti-tumor [24] H] -ion at m/z 179.0350 showed that the compound structure contained a quinic acid and caffeic acid.Comparing with standard, peak 4 was unambiguously confirmed as chlorogenic acid.Peaks 2 and 6 had uniform quasi-molecular ion at m/z 353, which indicates that these two compounds were isomeric with 4. It was easy to distinguish peaks 2 and 6 using literature reports and their relative retention times [14].Peaks 15, 16, and 18 were had the same quasi-molecular and fragment ions, indicating that these three compounds were isomeric.On the basis of comparisons with standards, and literature reports, the three peaks were confirmed as 3,4dicaffeoylquinic acid, 3,5-dicaffeoylquinic acid, and 4,5-dicaffeoylquinic acid.The derivation of the fragmentation pathway of 3,5-dicaffeoylquinic acid is shown in Figure 4B.

Network pharmacology
A total of 38 targets were obtained by importing 40 chemical compositions predicted to be absorbable into the PharmMapper database for directional docking (Table 2).These targets were then imported into the Molecule Annotation System, which resulted 57 pathways regulated by P. hookeri with significant differences (p < 0.05).Forty-four of these pathways that met the criterion of P < 0.01 (Table 3).
Cytoscape software was used to construct a pharmacology network of P. hookeri to generate the correlations of chemical components, targets, Trop J Pharm Res, June 2018; 17 (6): 1098           and pathways (Figure 6).Though the network diagram, a preliminary mechanism of the anti-RA of P. hookeri was obtained.
Table 3 show that the MAPK signaling pathway was involved in the largest number of targets.This pathway affects the release of inflammatory cytokines, and inhibits abnormal synaptic proliferation, thereby inhibition of RA bone erosion [25].In addition, some cancer-related pathways were found, including those related to colorectal, prostate, and pancreatic cancers.This is consistent with a previous report [24].

DISCUSSION
The technique of UPLC-Q-TOF/MS has the preponderances of fast analysis, low detection limit and strong qualitative ability.It has been widely used in the analysis of the chemical composition of medicinal plants, and has become an important means of identifying various compounds.In the MS data analysis process, it is extremely important to identify the correct quasi-molecular ion.This is related to the accuracy of the results.In the ESI+ mode, the quasi-molecular ion [M+Na] + is easily formed, while the ESI-mode mainly forms the [M-H] quasi-molecular ion, and the difference between  It is evident from the results obtained in this study that CA2, MAPK10 and PNP were associated with most of the compounds.Therefore, these factors can be considered to be the main targets.The CA2 gene encodes carbonic anhydrase 2, which is crucial for bone resorption and osteoclast differentiation, and the pathway involved is nitrogen metabolism.There is an imbalance in nitrogen metabolism in patients with RA.However, most patients with RA are in positive nitrogen balance when nitrogen intake is adequate [26].Purine nucleoside phosphorylase is encoded by PNP gene, and it catalyzes the phosphorolytic cleavage of the N-glycosidic linkage in purine (deoxy) ribonucleosides, with the liberation of free purine base and pentose-1-phosphate.The pathway involved is purine metabolism.Purine metabolism has a highly synergistic effect on immune cell function in RA [27].The MAPK10 gene encodes mitogen-activated protein kinase 10, which is associated with inflammation and immune function [28].The pathways related to RA include MAPK, VEGF, TGF-beta signaling pathway, and focal adhesion.Results from network pharmacological prediction showed that one compound can act on one or more different targets, and one target can also act on different pathways, suggesting that anti-RA property of P. hookeri involves multiple components, multiple targets, and multiple pathways.
The network pharmacology method used in this study is an innovative methodology due to the establishment of multilayer networks of diseasetarget-drug for forecasting drug targets in an integral manner, and for boosting efficient drug discovery [29].This method represents a breakthrough, when compared with the traditional herbal medicine research methodology of gene-target-disease, and initiates a new system of multiple genes-multiple targetscomplex diseases.This is the first study on the mechanism involved in the anti-RA effect of P. hookeri, and the first to use the novel technique of network pharmacology.Using this method, the main targets and pathways were successfully predicted, thereby providing a foundation for further research.This method is important for the study of complex drugs and should be applied in future studies.

CONCLUSION
A UPLC-Q-TOF/MS method has been successfully developed for the rapid analysis of chemical components of P. hookeri.A total of 40 compounds have been identified.Network pharmacology method enables the prediction of the therapeutic effect of P. hookeri on RA via a mechanism involving multiple components, multiple targets and multiple pathways.

Figure 6 :
Figure 6: The "component-target-pathway" network of P. hookeri [M+Na]+ and [M-H] -is 24 Da.It can be judged that the ion is a quasi-molecular ion.In the present study, UPLC-Q-TOF/MS technique was used to analyze the chemical components of P. hookeri.The results of this study provide a basis for research on the pharmacodynamics and quality control of P. hookeri.

Table 2 :
Information on potential targets from compounds of P. hookeri