Network pharmacology of bioactives from Sorghum bicolor with targets related to diabetes mellitus

Background Sorghum bicolor (SB) is rich in protective phytoconstituents with health benefits and regarded as a promising source of natural anti-diabetic substance. However, its comprehensive bioactive compound(s) and mechanism(s) against type-2 diabetes mellitus (T2DM) have not been exposed. Hence, we implemented network pharmacology to identify its key compounds and mechanism(s) against T2DM. Methods Compounds in SB were explored through GC-MS and screened by Lipinski’s rule. Genes associated with the selected compounds or T2DM were extracted from public databases, and the overlapping genes between SB-compound related genes and T2DM target genes were identified using Venn diagram. Then, the networking between selected compounds and overlapping genes was constructed, visualized, and analyzed by RStudio. Finally, affinity between compounds and genes was evaluated via molecular docking. Results GC-MS analysis of SB detected a total of 20 compounds which were accepted by the Lipinski’s rule. A total number of 16 compounds-related genes and T2DM-related genes (4,763) were identified, and 81 overlapping genes between them were selected. Gene set enrichment analysis exhibited that the mechanisms of SB against T2DM were associated with 12 signaling pathways, and the key mechanism might be to control blood glucose level by activating PPAR signaling pathway. Furthermore, the highest affinities were noted between four main compounds and six genes (FABP3-Propyleneglyco monoleate, FABP4-25-Oxo-27-norcholesterol, NR1H3-Campesterol, PPARA-β-sitosterol, PPARD-β-sitosterol, and PPARG-β-sitosterol). Conclusion Our study overall suggests that the four key compounds detected in SB might ameliorate T2DM severity by activating the PPAR signaling pathway.


Introduction
Type 2 diabetes mellitus (T2DM) is a metabolic disease triggered by the complex interaction between genetic and/or environmental associations. Patients with T2DM have either insulin resistance (sugar transfer dysregulation into cells) or lack of optimal insulin secretion to sustain natural glucose levels [1]. T2DM can affect at all ages, 1 in 3 adults has prediabetes, causes serious physiological complications, and brings diabetes diagnosis economic difficulties [2][3][4]. The typical signs of T2DM patients include recurrent urination, increased thirst, exhaustion, increased appetite, and indistinct vision. The main etiology for T2DM is persistent hyperglycemia which provokes mitochondrial dysfunction that stimulates to abundant reactive oxygen species (ROS) formation in several tissues and pancreatic β cells [5,6]. Such ROS accumulation in pancreatic β cells causes irreparable mitochondrial damage, resulting inhibition of insulin synthesis, and thus leading to diabetes progression by failing to produce enough level of insulin [7,8]. However, the anti-diabetic drug pathways include stimulating the insulin synthesis, inhibiting the production of endogenous glucose, and blocking carbohydrate absorption from intestine [9,10]. Currently, six classes of oral antidiabetic drugs, including Metformin, Glimepiride, Repaglinide, Pioglitazone, Sitagliptin, and Acarbose [11] are available which are reported to have serious side effects such as anorexia, nausea, dyspeptic episodes, and diarrhea [12,13]. Hence, the quest for potential drugs has now become more concentrated. In this regard, a rich source of phytoconstituents with health benefits may be the prospective candidate for T2DM intervention.
For the past few years, a report has shown that wholegrains including sorghum (Sorghum bicolor (L.) Moench) are the potential alternatives to ameliorate T2DM symptoms [14]. Sorghum bicolor (SB) contains a number of secondary metabolites which are reported to be effective in preventing various metabolic diseases, such as cancers, T2DM, obesity, and hyperglycemia [15,16]. A recent research demonstrated that oral treatment of SB (0.5% and 1%) noticeably reduced the low density lipoprotein cholesterol, triglycerides, and glucose level via PPARG in mice fed a high-fat diet, conversely, expression level of PPARG elevated to 1% [17]. Some reports expound that sorghum extracts at 1.0 mg/ml activates PPAR binding in mouse macrophage cell line; subsequently, PPAR agonists have been emerged as candidates against metabolic dysfunctions with T2DM [18,19]. Lecka-Czernik B reported that Aleglitazar, being developed by Roche Holding, is a dual agonist for PPARD and PPARA for the promising multiple treatment of hyperglycemia and dyslipidemia with T2DM patients [20]. It is evident that a dual PPAR agonist treatment is more potent than a single agonist treatment. A human study on sub-health condition with prediabetes confirmed that SB administration significantly decreased the level of glucose by 35% [21]. Although literature survey revealed that diabetes and obesity complications can be controlled upon consumption of SB [22,23]; however, its mechanism(s) against T2DM has not been explored to date. Therefore, the research on bioactive compound(s) and pathways of SB against T2DM should be justified in scientific testing, in order to maintain its pharmacological complements for T2DM.
Network pharmacology is a structured methodical mode, which can investigate the interactive networking elements such as compounds, genes, proteins, and diseases [24,25]. Network pharmacology can decode the mechanism of compounds with a multifunctional point of view, which highlights the interaction of diverse factors, instead of "one target, one compound" [26]. The network pharmacology is therefore a valuable approach to identify potential lead compounds (from natural sources) with specific mechanism of action for the prevention of various disease, and mostly to elucidate the synergistic impact of bioactive compounds [27]. Zhang B. et al. reported that the rapid development of bioinformatics, systems biology, and poly-pharmacology contributes greatly to network-based drug discovery, which is regarded as a costeffective drug development method [28]. A report explicates that the network pharmacology is used as a powerful tools to identify the mechanisms of actions between traditional Ge-Gen-Qin-Lian decoction (GGQLD) formula and target genes [29]. In our study, network pharmacology was used to evaluate the bioactive compounds and mechanism(s) of SB against T2DM. Firstly, GC-MS analysis of SB was conducted to identify the bioactive compounds and their "drug-likeness" property was screened by Lipinski's rule [30]. Secondly, genes related to selected compounds or T2DM were identified using public databases, and overlapping genes between genes related to SB and T2DM target genes were also identified. Thirdly, genes related to a hub signaling were selected by analyzing gene set enrichment analysis. Finally, the selected genes were implemented for molecular docking analysis to find the most potent candidates of SB against T2DM.

Plant preparation, extraction
The Sorghum bicolor (SB) was purchased from Chuncheon local market, Korea. The collected SB was dried and powdered using electric blender. Approximately 500 g of SB powder was soaked in 500 mL of 100% methanol (Daejung, Korea) for 3 days and repeated for 3 times to collect extraction. The solvent extract was collected, filtered, and evaporated using a vacuum evaporator (IKA-RV8, Japan). The evaporated sample was dried under a boiling water bath (IKA-HB10, Japan) at 40˚C to obtain yield.

GC-MS analysis
Agilent 7890A was used to carry out GC-MS analysis. GC was equipped with a DB-5 (30m×0.25mm×0.25μm) capillary column. Initially, the instrument was maintained at a temperature of 100˚C for 2.1 minutes. The temperature was risen to 300˚C at the rate of 25˚C/min and maintained for 20 minutes. Injection port temperature and helium flow rate were ensured as 250˚C and 1.5 ml/min, respectively. The ionization voltage was 70 eV. The samples injected in split mode at 5:1. MS scan range was set at 35-550 (m/z). The fragmentation patterns of mass spectra were compared with those stored in the using W8N05ST Library MS database. The percentage of each compound was calculated from the relative peak area of each compound in the chromatogram. The concept of integration used the ChemStation integrater algorithms.

Network construction of interacted overlapping genes
Through STRING (https://string-db.org/) analysis, the overlapping genes were closely correlated, and the signaling pathways of overlapping genes were analyzed by RStudio bubble chart. Networking between bioactive compounds and genes of SB against T2DM identified a hub signaling pathway.

Preparation for molecular docking of ligand molecules
The ligand molecules were converted.sdf from PubChem into.pdb format using Pymol, and the ligand molecules were converted into.pdbqt format through Autodock.

Ligand-protein docking
The ligand molecules were docked with target proteins utilizing autodock4 by setting-up 4 energy range and 8 exhaustiveness as default to obtain 10 different poses of ligand molecules [31]. The 2D binding interactions was identified through LigPlot+ v.2.2 (https://www.ebi.ac. uk/thornton-srv/software/LigPlus/). After docking, ligands of lowest binding energy were selected to visualize the ligand-protein interaction in Pymol.

Potential bioactive compounds from SB
A total of 20 compounds in SB were detected by the GC-MS analysis (Fig 1), and the name of compounds, retention time, peak area (%) are enlisted in Table 1. All 20 compounds were checked and accepted by Lipinski's rule (Molecular Weight � 500g/mol; Moriguchi octanolwater partition coefficient �4.15; Number of Nitrogen or Oxygen �10; Number of NH or OH �5), and all compounds have satisfactory "Abbott Bioavailability Score (> 0.1)" identified through SwissADME (Table 2).

Overlapping genes between SEA and STP linked to 20 compounds
Based on the SMILES, a total of 308 genes from SEA and 324 genes from STP linked to 20 compounds were extracted (S1 Table). The result of Venn diagram exhibited that 118 genes were overlapped between the two public databases (Fig 2).

PLOS ONE
The analysis of bioactive compounds in Sorghum bicolor against type 2 diabetes mellitus

Overlapping genes between T2DM-related genes and the 118 overlapping genes
A total of 4,736 genes related to T2DM were sorted by retrieving DisGeNET and OMIM databases (S2 Table). The result of Venn diagram unveiled that 81 overlapping genes was identified between 4,763 genes related to T2DM and the 118 overlapping genes (Fig 3), (S3 Table). As shown in S4 Table, a total of 81 overlapping genes linked to 16 compounds from aforementioned 20 compounds were identified, retrieving from both SEA and STP public databases, and no genes were found associated with other 4 compounds (Glycerin; Glyceraldehyde; 2,3-dihydro-3,5-dihydroxy-6-methyl-4H-pyran-2-one; and 2-propanol, 1-chloro-3-(1-methylethoxy)) in the two databases.

Signaling pathways and finding of a hub signaling of SB against T2DM
The result of KEGG pathway enrichment analysis demonstrated that 81 genes were related to 12 signaling pathways (False Discovery Rate < 0.05). The 12 signaling pathways were directly related to the progression of T2DM and indicated that these 12 signaling pathways might be the key pathways of SB against T2DM. The description of 12 signaling pathways is shown in Table 3. Additionally, a bubble chart suggested that PPAR signaling pathway might be a hub signaling pathway of SB against T2DM (Fig 6).  7). Also, each six proteins have strong molecular interactions on PPAR signaling pathway (Fig  8). The nominated compounds to dock with six proteins were selected via both SEA and STP. The molecular docking analysis was performed to evaluate the binding energy of these six genes against their related each gene, respectively, and the docking figures are depicted in Fig 9. Molecular docking score of A1-A6 on FABP3 protein (PDB ID: 5HZ9) is analyzed in the "Homo Sapiens" mode. Docking simulation of the affinity between A1-A6 and FABP3 protein in the "Homo sapiens" setting was analyzed. Based on the docking score, the order of the priority of binding energy is given: A5>A6>A2>A3>A1>A4. The six-binding energy of A1-FABP3, A2-FABP3, A3-FABP3, A4-FABP3, A5-FABP3, and A6-FABP3 demonstrated -6.6, -7.4, -6.7, -6.5, -8.4, and -7.8 kcal/mol, respectively. Interaction analysis of the best-docked compound namely "propyleneglycol monoleate" showed several hydrophobic bonds on FABP3

PLOS ONE
The analysis of bioactive compounds in Sorghum bicolor against type 2 diabetes mellitus

PLOS ONE
The analysis of bioactive compounds in Sorghum bicolor against type 2 diabetes mellitus

Discussion
Compounds-genes networking system unveiled that therapeutic effect of SB against T2DM was related to 16 compounds out of 20 compounds detected by GC-MS, including five steroid derivatives, four fatty acyls, two fatty acid esters, one organooxygen, one prenol lipid, one pyrimidine nucleoside, one azole, and one coumarin. The proportion of steroid derivatives to 16 compounds was close to 30%, implying that steroid derivatives was the most essential than any other sorts of compounds for the amelioration effect of SB on T2DM. It was reported that some steroid derivatives have strong effect on hypoglycemic activity based on lipophilic properties [32]. Noticeably, a report showed that β-sitosterol (steroid derivatives) controls the glycemic level through regulation of IR (insulin receptor) and GLUT4 (glucose transporter 4) proteins in adipocytes of high fat and sucrose treated type 2 diabetic rats, interestingly, the in vivo result was in line with in silico analysis [33]. In addition, researchers found that campesterol (steroid derivatives) decreases the LDL (Low Density Lipoprotein) level, associated with the occurrence and development of T2DM [34,35]. A patent revealed that propyleneglycol monoleate is an agent for treatment or amelioration of diabetes, obesity or arteriosclerosis [36]. Another report suggested that 25-Oxo-27-norcholesterol interrupts cholesterol oxidation, which is related to insulin resistance [37,38]. These results coincide

PLOS ONE
The analysis of bioactive compounds in Sorghum bicolor against type 2 diabetes mellitus with our findings of SB on T2DM, suggesting that the quantity of the four compounds in SB is greatly enough to exhibit anti-diabetic efficacy.
Compounds-genes networking also specified that the pharmacological effect of SB on T2DM was directly associated with 81 genes. The results of KEGG pathway enrichment analysis of 81 genes showed that 12 signaling pathways were directly linked to the occurrence and progression of T2DM, demonstrating that these signaling pathways might be the key pathways of SB against T2DM. The 12 signaling pathways with T2DM were succinctly discussed as follows.
PPAR signaling pathway: PPAR ligands are the potential therapeutic candidates against T2DM, also, alleviate metabolic syndrome including obesity and insulin resistance [39]. Furthermore, dual agonists approach with both PPARA agonists (such as fibrates) and PPARG agonists (such as thiazolidinediones) can have better metabolic efficacy and less side effects than its single administration [40]. cAMP signaling pathway: cAMP signaling pathway modulates glucose homeostasis with insulin and glucagon secretion, glucose uptake, gluconeogenesis, glycogen synthesis and breakdown of glucose [41]. HIF-1 signaling pathway: Inhibition of HIF-1 signaling caused by diabetes is associated with hypoxia and high degradation of HIF-1α  protein [42]. Sphingolipid signaling pathway: Sphingolipid is a significant class of signaling lipids, have been recognized as vital players in the progression and pathogenesis of insulin resistance and T2DM [43]. AMPK signaling pathway: The activation of AMPK enhances homeostasis of glycemic level, lipid concentration, and blood pressure in insulin-resistant rodents, which is considered as an important therapeutic target against T2DM [44]. VEGF signaling pathway: VEGF is related deeply to the development of T2DM. High content of VEGF is generally detected in plasma of T2DM patients [45]. T cell receptor activation: Overactivated T cells with T2DM patients might be an indication of losing the natural regulatory mechanism, thus, intervention of T cells by T2DM might down T cell receptor sensitization [46]. Fc epsilon RI signaling pathway: Fc epsilon RI-mediated signaling in mouse bone marrow is potentiated by insulin [47]. Estrogen signaling pathway: In a research of postmenopausal mice and human cells, the report demonstrated that estrogen is related to lower risk of T2DM by targeting particular cells in the pancreas and gut to improve tolerance to glucose [48]. Prolactin signaling pathway: A normal range of prolactin concentration is linked to a lower T2DM risk, which may play an inhibitory effect on the development of T2DM [49]. Adipocytokine signaling pathway: Adipocutokines leptin and adiponectin might be significant biomarkers for first prediction on T2DM, which is further associated with diabetic microvascular complications [50]. Collectively, this study suggests that 12 signaling pathways with 81 genes are linked to the development of T2DM. In addition, rich factor defines that the proportion of the DEGs number and the number of genes have been annotated in pathway [52]. In other words, the higher of the rich factor is, the higher the degree of enrichment. The rich factor of PPAR signaling

Conclusion
Sorghum bicolor (SB) is rich in beneficial phytoconstituents and seen as a possible source of natural antidiabetic agents. However, in this report, its comprehensive bioactive compounds and T2DM pathways were firstly investigated using the network pharmacology. The findings of this study indicate that the antidiabetic ability of the SB could be attributed to four main compounds (β-sitosterol, campesterol, propyleneglycol monoleate, and 25-Oxo-27-norcholesterol) that were strongly related to PPAR signaling pathway. Therefore, our study suggests that the four key compounds of SB might ameliorate T2DM by activating the PPAR signaling pathway.
Supporting information S1