The Anti-RA Activities of a Couplet Medicinals, Gastrodia Elata and Radix Aconitic Lateralis preparata, Explored by Untargeted Metabolomics and Network Pharmacology.


 Background

The couplet Medicinals, Gastrodia Elata and Radix Aconitic Lateralis preparata (GERA), is an established formula extensively used in Chinese medicine for treating rheumatoid arthritis (RA). However, the anti-RA mechanisms of GERA are still unclear. This paper aims to explore the anti-RA mechanisms of GERA by a combined strategy of untargeted metabolomics and network pharmacology.
Methods

Three water extracts were prepared as propotions: Gastrodia Elata compared with Radix Aconitic Lateralis (w/w): 1:1, 3:2, or 2:3. The untargeted metabolomics were executed with UPLC-MS. The metabolites were annotated and identified by Human metabolome database (HMDB) and Lipid maps database. Finally, key genes and pathways related to anti-RA activities of GERA were mined by network pharmacology.
Results

The untargeted metabolomics profile displayed that four differentially expressed metabolites were involved in isoflavonoid biosynthesis and biosynthesis of unsaturated fatty acids (p < 0.05). Among these differential metabolites, the essential ingredients of GERA were linoleic acid, daidzein, and daidzin. The principal targets of anti-RA activities of GERA were IL-6, TNF, VEGFA, TP53, CASP3 and PTGS2. Thirty anti-RA targets of GERA were majorly belonged to pathways response for anti-inflammation, endothelial function and apoptosis, suggesting the fundamental process of RA treatment.
Conclusion

The anti-RA activities of GERA were based on the inhibition of inflammation and regulation of endothelial function and apoptosis.


Conclusion
The anti-RA activities of GERA were based on the inhibition of in ammation and regulation of endothelial function and apoptosis.

Background
The couplet medicinals is one of the famous features of Chinese medicine which displays marked synengy effects compared with individual herbs. A couplet medicinals, combined by Gastrodia elata and Radix aconitic lateralis preparata (GERA), has been extensively used in Chinese medicine. 1159 of GERA were observed from the formulae of anicent Chinese medicine [1] . GERA is an establised couplet Medicinals for the treatment of rheumatoid arthritis (RA) [2] . A GERA extract made by Chen et al. [3] displayed signi cant anti-RA activity, with the cure rate 85.30%. Our former studies indicated that GERA could e ciently relieve RA, joint swelling and pain syndrome [4] . On the other hand, network pharmacology has become a novel branch of systems biology, particularly used in the identi cation of e cient targets and their pathways of Chinese formulas [5][6] . Twenty active compounds of a couplet medicinals, Gastrodia elata and Ligusticum chuanxiong Hort, were annoted to 48 molecular targets which are associated with migraine headaches by network pharmacology [7][8] . The couplet of Gastrodia elata and Ramulus Uncariae demonstrated multi-target and multi-pathway regulation for the therapy of cerebral ischemia [9] . However, the anti-RA mechanisms of GERA is waiting for further exploration. Therefore we applied the untargeted metabolomics for identifying differentially extracted metabolites of GERA solutions, and mined the hidden KEGG pathways involved in RA therapy by network pharmacological approach.

Sample preparation
Both herbs used in this study were collected in their genuine area. The Gastrodia elata was collected from Dafang county, Guizhou province, China. The Radix aconitic lateralis preparata was collected from Jiangyou county, Sichuan province, China. The two herbs were identi ed by Professor Yun Deng, Chengdu University of Traditional Chinese Medicine. Three kinds of water solutions were prepared as follow propotions: Gastrodia Elata compared with Radix Aconitic Lateralis preparata (w/w): 1:1, 3:2 or 2:3. Different ratios of medicinal herbs can extract different compounds and, consequently, show distinguish pharmaceutical activities. Ten times of distilled water was added to the raw herbs (w/w), soaked at room temperature for 30 min, boiling for 30 min, ltered and stored the boiled solution. Then added eight times of distilled water into the dregs of decoction (w/w), boiling for 30 min, ltered and mixed the solution to the stored one, and enriched the mixed solution to 1:1.1( raw herb vs extracted solution, w/v). Finally, stored at 4 °C for follow-up analysis.

Untargeted Metabolomics Analysis Integrated With Uplc-ms
The stored solutions of Gastrodia Elata (Chinese name, Tian-Ma)and Radix Aconitic Lateralis (Chinese name, Fu-Zi) were respectively freeze-dried into powder (100 mg), grounded with liquid nitrogen, homogenated and resuspended with pre-chilled 80% methanol and 0.1% formic acid by vortexing brie y. Three formulae were then composed based on the relative qualities(w/w) of Tian-Ma (T) and Fu-Zi (F) powder: 1:1 (TF11), 3:2 (TF32), or 2:3 (TF32). The samples were incubated on ice for 5 min, centrifuged at 15000 rpm, 4 °C for 5 min. The supernatant was diluted to the nal concentration containing 60% methanol in LC-MS grade water. The samples were subsequently transferred to a fresh tube after 0.22 µm ltering, centrifuged at 15000 rpm, 4 °C for 10 min. Finally, the ltrate was injected into the LC-MS system for metabolite analysis.
LC-MS analyses were performed using a Vanquish UHPLC system (Thermo Fisher) coupled with an Orbitrap Q Exactive series mass spectrometer (Thermo Fisher). Samples were injected into a Hyperil Gold column (100 × 2.1 mm, 1.9 µm) using a 16-min linear gradient at a ow rate of 0.2 mL /min. The eluents for the positive polarity mode were eluent A (0.1% formic acid) and eluent B (Methanol). The eluents for the negative polarity mode were eluent A (5 mM ammonium acetate, pH 9.0) and eluent B (Methanol). The solvent gradients were set as follows: 98%A, 2% B, 1.5 min; 100% B, 12.0 min; 100% B, 14.0 min; 98%A, 2% B, 14.1 min; 98%A, 2% B, 16 min. Q Exactive mass spectrometer was operated in positive and negative polarity mode. The ESI optical source was set as follows: spray voltage of 3.2 kV, the capillary temperature of 320 °C, sheath gas ow rate of 35 arb, and aux gas ow rate of 10 arb.
The raw data les generated by UHPLC-MS were processed using Compound Discoverer 3.0 (CD3.0, Thermo Fisher) to perform peak alignment, peak picking, and quantitation for each metabolite. The main parameters were set as follows: retention time tolerance, 0.2 minutes; actual mass tolerance, 5 ppm; signal intensity tolerance, 30%; signal/noise ratio, 3; and minimum intensity, 100000. After that, the peak intensities were normalized to the total spectral intensity. The normalized data were used to predict the molecular formula based on additive ions, molecular ion peaks, and fragment ions. Then the peaks were matched with the mzCloud (https://www.mzcloud.org/) and ChemSpider (http://www.chemspider.com/) databases to obtain accurate qualitative and relative quantitative results.
For clustering heat maps, the data were normalized using z-scores of the intensity areas of differential metabolites and were plotted by the Pheatmap package in R language. The functions of these metabolites and metabolic pathways were studied using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The metabolic pathway enrichment of differential metabolites was performed. When the ratio was satis ed by x, n > y N, the metabolic pathway was considered as enrichment, when P-value of metabolic pathway < 0.05, the metabolic pathways was considered as statistically signi cant enrichment.

Network Pharmacological Analysis
Based on the differential expressed metabolites involved KEGG pathways with P value less than 0.05, we used a Traditional Chinese medicine system pharmacology database and analysis platform (TCMSP) for obtaining these metabolites' targets. RA targets were obtained from the human gene database (https: //www.genecards.org). The intersection between GERA and RA targets was produced. Subsequently, we found critical genes related to GERA against RA by the Protein-Protein interaction networks database (String database). The Cytoscape 3.7.1 software was used in constructing a compound-target and target-pathway network. Finally, functional enrichment analysis of common targets was performed on the DAVID Bioinformatics Resources 6.8 platform database (https://david.ncifcrf.gov).

Statistical analysis
Statistical analyses were performed using the statistical software R (version R-3.4.3), Python (Python 2.7.6 version), and CentOS (CentOS release 6.6). While data were not normally distributed, standard transformations were attempted using of area normalization method. We applied univariate analysis (ttest) to calculate the statistical signi cance.

Results
Classi cation annotation of metabolites from GERA

Metabolites Analysis Of The Gera Extract
It is well known that different composition ratios of certain formulae may result in distinctive secondary metabolites. The scatter diagrams of PCA and PLS-DA analysis were displayed in Figs. 1 and 2. PLS-DA analysis is usually used to discriminate the metabolite contents and their distribution pattern; results shown in Fig. 2 indicated that the discriminate model recruited here was not over-tted.
As shown in Fig. 6, the compound-target network consisted of 57 nodes (one disease, one couplet medicinals, three pharmaceutical ingredients and 52 targets) and 120 edges. This network displayed that EIC (linoleic acid), daidzein and daidzin interacted with 2, 26 and 6 targets, respectively. Which preliminarily revealed the interrelationships between GERA compounds and anti-RA targets. On the other hand, PTGS2 was targeted by three active ingredients of GERA, playing an important role in in ammatory modulation.
Results of the target-pathway network analysis were shown in Fig. 7. Of which 30 targets and 12 pathways were obtained, with an average crosstalks of one target related to 6.67 pathways, and one pathway to 10.00 targets. Results displayed that the KEGG pathway, Kaposi sarcoma-associated herpesvirus infection (hsa05167), possess the most crosstalks with anti-RA targets. Pathways of hepatitis B (hsa05161), Epstein-Barr virus infection (hsa05169), human cytomegalovirus infection (hsa05163) are also frequently related with anti-RA targets. RA is a typical autoimmunological in ammation. Hence in ammatory regulating pathways, such as TNF signaling pathway (hsa04668) and IL-17 signaling pathway (hsa04657) mined from GERA, are the critical anti-RA process. It is note worthy that the enriched pathways, i. e., human cytomegalovirus infection, TNF signaling pathway, arachidonic acid metabolism and Jak/STAT signaling pathway, are all involved in the excess release of proin ammatory cytokines. Therefore, the therapeutical activities of GERA may majorly attribute to tune these proin ammatory pathways.

Discussion
Untargeted metabolomics effectively mined the anti-RA targets of GERA Tian-Ma pill as a classic anti-RA formula has been extensively applying in Chinese medicine since Ming Dynasty (600 years ago). Tian-Ma pill is composed of ten Chinese herbs and GERA, the couplet Medicinals make up Tian-Ma and Fu-Zi, is the key factor of this effective formula. Accumulate data indicated that different dosage of Tian-Ma and Fu-Zi displayed distinguish pharmaceutical activities. For instance, TF11 and TF32 are widely used for the treatment of arthromyodynia and RA, respectively. However, the underlying mechanisms and targets are open to be explained. Our results of untargeted metabolomics combined with UPLC-MS determined not only the differentially isolated metabolites from blood serum and tissues, but also the GERA extract [10][11] . Figures 3 and 4 showed differentially isolated metabolites between three propositions of GERA, while enrichment analysis of KEGG pathways displayed the overall differential metabolites were majorly involved in iso avonoid biosynthesis and biosynthesis of unsaturated fatty acids. Both iso avonoid and unsaturated fatty acids possess broad pharmacological e cacy against RA. Linoleic acid is the major n-6 polyunsaturated fatty acid, its desaturation and elongation gives rise to arachidonic acid and gamma-linolenic acid as a precursor of prostaglandin E 2 [12][13] . Hence, appropriate intake of linoleic acid has effective anti-RA activities [14][15][16] . Daidzein and daidzin can ameliorate RA symptoms and decrease RA occurrence [17][18][19][20] . Our results demonstrated that the content of linoleic acid in TF11 was higher than that of TF32, while the daidzin content of TF32 was higher than either TF11 or TF23 (Table 1). Which help to explain the quantitively anti-RA activities of GERA.

Network Pharmacology Unraveled Crucial Anti-ra Targets Of Gera
Recently, a large amount of clinical cases and research data displayed the anti-RA effects of GERA. However, less work has addressed the overall targets of GERA against RA. Network pharmacology, an integrated platform widely applied to Chinese medicine, can effectively explain the multiple relationships between herbs, ingredients, diseases and their targets [21] . Results of network pharmacological analysis shown that linoleic acid, daidzein and daidzin were the key metabolites response for anti-RA activities of GERA. Figure 5 displayed that essential genes related to anti-RA ingredients of GERA. Both TNF-α and IL-6 are the major proin ammatory cytokines which modulate the pathologic RA progress [22][23][24] . COX-2 (PTGS2) is the rate-limiting enzyme of arachidonic acid metabolism, which controls the production of PGE 2 , the proin ammatory mediators of autoimmunity activities [25][26][27] . VEGF is an essential angiogenic marker for RA disorders [28][29][30] . TP53 and CASP3 regulate the apoptosis of rheumatoid synovial cells [31][32] . On the whole, results of network pharmacological analysis uncovered key anti-RA targets of GERA.
RA is a typical autoimmune disease induced by various risk factors, including Kaposi sarcomaassociated herpesvirus [33][34] , Epstein-barr virus [35][36] and human cytomegalovirus infection [37] . Combined analysis on target-pathway network, we believe that KEGG pathways shown in Fig. 7 may monitor the transcription of anti-RA genes involved in in ammation process, vascular endothelial function and apoptosis. First, the activation of TNF [38] and IL-17 signaling pathways [39] indicate the production of TNF-α, IL-6, IL-7 and COX-2, pro-in ammatory cytokines result in the pathological process of RA. Second, PI3K/AKT/mTOR/NF-κB signaling pathway can modulate Chondrocyte proliferation, synovial broblasts apoptosis and autophagy that are intensively related to RA [40][41] . Then, suppression of MAPK and FoxO signaling pathways are important mechanisms for ameliorating in ammation and provoking apoptosis in RA patients [42] . Finally, pathways responsible for reducing reactive oxygen species [43] are of vital importance of anti-RA treatment focused on excessive in ammation. These enriched pathways unravel a network characteristic of GERA against RA as a multi-component, multitarget and multi-pathway pattern.
The present work explained the pharmaceutical basis of GERA against RA by an untargeted metabolomics combined with network pharmacological strategy. However, the aetiology of RA includes increasing factors, such as gut microbiota and miRNA regulation. In future researches, we will determine the pharmaceutical activates of linoleic acid, daidzein and daidzin, key anti-RA ingredentis of GERA. The substantial metabolic process of GERA will also be detected by a metagenomic platform.  Precedence ordering plot of PLS-DA from three GERA proportions. Three formulae were composed on the proportion (w/w) of Tian-Ma (T) and Fu-Zi (F): 1:1 (TF11), 3:2 (TF32), or 2:3 (TF23). PC: the principal component. X-axis is the relative random group with the primary group. Y-axis is the score of R2 or Q2. R2 represents for tting ability. Q2 represents for forecasting ability. Left, positive ion mode. Right, negative ion mode.
Red represents up-regulated metabolites, blue represents down-regulated metabolites, and white represents non-signi cantly expressed metabolites. Left, positive ion mode. Right, negative ion mode.