Design , synthesis and anti-diabetic activity of some novel xanthone derivatives targeting α-glucosidase

Twenty eight xanthone derivatives were designed and docked into the Nterminal catalytic domain of maltase-glucoamylase (ntMGAM) by considering miglitol as standard drug. Most of the molecules showed excellent docking scores and docking interaction as compared to the binding cavity of the standard molecule. The five best scoring ligands were synthesized and characterized by a number of analytical and spectroscopic techniques. The molecules were screened for the in vivo anti-diabetic activity in streptozotocininduced diabetic animal model in Wistar rats. Compound P4 showed the most prominent inhibition among others. The synthesized compounds reported significant (p<0.01) effect in lowering blood glucose levels compared to miglitol as a standard -glucosidase inhibitor. Article Info Received: 1 December 2015 Accepted: 12 February 2016 Available Online: 16 March 2016 DOI: 10.3329/bjp.v11i2.25851 Cite this article: Bairy PS, Das A, Nainwal LM, Mohanta TK, Kumawat MK, Mohapatra PK, Parida P. Design, synthesis and anti-diabetic activity of some novel xanthone derivatives targeting -glucosidase. Bangladesh J Pharmacol. 2016; 11: 308-18. Design, synthesis and anti-diabetic activity of some novel xanthone derivatives targeting α-glucosidase Partha Sarathi Bairy1, Aparoop Das1, Lalit Mohan Nainwal1, Tapan Kumar Mohanta2, Mukesh Kumar Kumawat3, Pradyumna Kishore Mohapatra3 and Pratap Parida4 Department of Pharmaceutical Sciences, Dibrugarh University, Dibrugarh, Assam 786 004, India; Free Major of Natural Science, School of Basic Studies, Yeungnam University, 280, Daehak Gyeongsan, Gyeongsanbuk, 712749, Republic of Korea ; Anand College of Pharmacy, Keetham, Agra, Uttar Pradesh 282 007, India; Regional Medical Research Centre, NE Region, Indian Council of Medical Research, Dibrugarh, 786 001, Assam, India. anti-bacterial, antifungal, antiviral (Diderot et al., 2006), antimycobacterial (Pickert et al., 1998), antidepressant (Galt et al., 1989), anti-diabetic (Liu et al., 2006) and monoamine oxidase inhibitors. Materials and Methods Design of new molecules A xanthone nucleus with different substitutions is proved to have a diverse class of pharmacological profile. Previously established -glucosidase inhibitor like acarbose, miglitol were observed to have polyhydroxy groups in their structure and new molecules are designed considering this fact. For this study, xanthones with several hydroxy groups and alkoxy groups at different position are designed and drawn using ChemAxon, a freeware developed by Advanced Chemistry Development, Inc. and converting 2D chemical structure of compound to 3D structures (Nainwal et al., 2014).


Introduction
Glucosidase catalyzes especially hydrolyses the carbohydrates to free glucose unit in blood in the final step of carbohydrate metabolism.Glucosidase causes hydrolysis of and glycosidic linkages of carbohydrates, thus they are -glucosidase and -glucosidase (Heightman et al., 1999).Among them -glucosidase (EC 3.2.1.20)draws considerable interest to the pharmaceutical research community because it helps to increase postprandial blood glucose level (Park et al., 2008).Inhibition of the enzyme is a useful chemotherapy for controlling diabetes and obesity.Due to its catalytic role, it also targeted in the treatment of other carbohydrate mediated diseases, including cancer (Humphries et al., 1986), viral infections (Mehta et al., 1998;Karpas et al., 1988), and hepatitis (Zitzmann et al., 1999).
So, the research continues to find other problem solving alternatives.Several fields of research upon xanthone nucleus going on and recent studies indicate that mangiferin, a xanthone C-glycoside, serve as potentglucosidase inhibitors (Liu et al., 2006).
Xanthones are secondary metabolites found in plants, fungi and lichens and now in the center of research interest by the past two decades because of its diverse class of pharmacological profile (Cardona et al., 1990).Xanthones are reported for biological activities like antitumor, anti-oxidant, anti-inflammatory, anti-allergy, anti-bacterial, antifungal, antiviral (Diderot et al., 2006), antimycobacterial (Pickert et al., 1998), antidepressant (Galt et al., 1989), anti-diabetic (Liu et al., 2006) and monoamine oxidase inhibitors.

Design of new molecules
A xanthone nucleus with different substitutions is proved to have a diverse class of pharmacological profile.Previously established -glucosidase inhibitor like acarbose, miglitol were observed to have polyhydroxy groups in their structure and new molecules are designed considering this fact.For this study, xanthones with several hydroxy groups and alkoxy groups at different position are designed and drawn using ChemAxon, a freeware developed by Advanced Chemistry Development, Inc. and converting 2D chemical structure of compound to 3D structures (Nainwal et al., 2014).

Molecular property
All the drawn structures were screened for in silico biological data prediction using Molinspiration online filter and Drulito software.They screened the designed molecules according to 'Lipinski's rule of five' predicting molecular properties like molecular weight, total polar surface area (TPSA), LogP, hydrogen bond donor (HBD), hydrogen bond acceptor (HBA) and number of roratable bonds.

Drug likeness and bioavailability
Virtual screening of the compounds was performed before molecular docking simulation studies.All the compounds were screened for predictive molecular physicochemical properties such as absorption, distribution, metabolism, excretion and toxicity (ADME/ Tox).This aids the screening of the compounds for potential drug like properties.Drug intensity and kinetics of drug contact to various tissues is greatly influenced by ADME/Tox, which in turn reflects the efficiency and pharmacological activity of the compound as a drug.The mutagenicity, tumourogenicity, irritating and reproductive effects as well as drug likeness and drug scores of the compounds were predicted by using OSIRIS property explorer (http:// www.organic-chemistry.org/prog/peo/).

Protein retrieval and preparation
-Glucosidase crystal structure was obtained from RCSB (Research Collaboratory for Structural Bioinformatics), Protein Databank (PDB, http://www.pdb.org).The PDB ID of the selected protein was found to be 3L4W (Sim et al., 2010) and downloaded as PDB text file format.All the heteroatom including co-crystallized ligand was removed by Molegro Molecular Viewer to make protein free and suitable for further docking studies.Now the protein comprises of only one chain containing 863 amino acid residues and 13493 numbers of atoms and 13693 numbers of bonds.Additions of nonpolar hydrogens were done for better interaction.Kollman united atom charges were applied to the protein.Appropriate ionization and tautomeric states of numerous amino acid residues such as Arg, His, Asp, Glu and Ser were managed by adding H-atoms to the protein at pH 7.0.

Molecular docking study
Molecular docking is generally used to detect the protein-ligand orientation and interaction.AutoDock Tools package version 2.4 was utilized to create the docking input files.The grid region was surrounded by the active site for binding.So, grid region was selected on the basis of amino acid residues representing the binding site of miglitol as the standard drug obtained from PDB with ID-3L4W and considered as the best active region for the favorable interaction.The grid box was set at 80 × 80 × 80 A˚ for x, y and z axis and covered 12 amino acid residues (TYR299, ASP327, ILE364, TRP406, TRP441, ASP443, MET444, ARG526, TRP539, ASP542, PHE575, HIS600) of active site.The Lamarckian Genetic Algorithm (LGA), a local search algorithm was utilized for ligands conformers searching.During the docking process, a maximum of 10 conformers were considered for each compound.This method was applied for each designed compound and after completion the conformer with lowest binding energy was chosen.
The conformational similarity by visualizing the binding site and its energy (Kcal/mol), and the docked amino acid residues forming hydrogen bonds and other parameters like intermolecular energy (Kcal/mol) and inhibition constant ( M) were analyzed by AutoDock tool.Ten best poses were generated for each ligand and scored using AutoDock 4.2 scoring functions (Morris et al., 1998).Based on the docked energy all the ligands were ranked.The ligand interacting residues with the target protein were analyzed using AutoDock tools, PyMOL (Konc et al., 2011) and LigPlot (Madeswaran et al., 2011).

Structural investigation
All chemicals used in the work were used without further purification.The intermediate was taken in an open capillary on the Veego-MPI melting point apparatus and the melting point of the synthesized compounds was determined.The progress of the reactions was monitored on silica gel-G TLC plate using various solvent combinations.The spots were detected with iodine vapors and observed under UV-light.The UVvisible spectra of the synthesized compounds were recorded on UV-visible spectrophotometer (Shimadzu UV-1800).Infrared spectra were recorded on an FT-IR Perkin-Elmer spectrometer.The 1 H and 13 C NMR spectra were recorded at 400 MHz and 100 MHz, respectively, on a Bruker Avance-II 400 NMR spectrometer using DMSO-d6 as solvent with tetramethylsilane (TMS) as an internal standard.Mass spectra were obtained on a Waters Q-TOF MICROMA SS LC mass spectrometer (Silverstein and Webster, 1963).

General procedure
Salicylic acid derivative and polyhydroxy phenols were used to synthesize hydroxyxanthones (intermediate) and further alkylation done by various alkyl bromide in the presence of acetone (Scheme 1).For the synthesis of hydroxyxanthones Eaton's reagent (Eaton and Carlson, 1973) was poured in the mixture of salicylic acid derivative (60 mmol) and polyhydroxy phenol (60 mmol), stirred at 70°C for 30 min.The mixture cooled, stirred with cold water, keeping temperature 0-4°C for 2.5 hours.The resulting solid collected by filtration, washed with water until pH 6 and dried at 60°C (Varache-Lembege et al., 2008).To the intermediate (2 mmol) and alkyl bromide (3 mmol) in acetone (55-60 mL) was added potassium carbonate (2.5 mmol).Mixture was refluxed under stirring for 2-4 hours, cooled, filtered and it concentrated.Further recrystallization was done and the product collected as yellow solid (Liu et al., 2006).

Evaluation of anti-diabetic activity
Synthesized compounds are evaluated for in vivo antidiabetic activity in diabetic rats using streptozotocin as diabetic inducing agent (Abeeleh et al., 2009).

Animals
Adult male Wistar rats (150-200 g) were used to study the anti-diabetic activity.Animals were housed in standard laboratory conditions (temperature 22 ± 2°C and humidity 45 ± 5% with 12 hours day: 12 hours night cycle).All animals received standard laboratory diet and water ad libitum.

Acute toxicity studies
Acute oral toxicity study was performed according to OECD-423 guidelines (acute toxic class method).Adult female Wistar rats (n = 5; 120-200 g) were selected by random sampling for acute toxicity study.The animals were kept fasting overnight and provided water ad libitum.The synthesized drugs were administrated orally at 5 mg/kg body weight orally and observed for 14 days.If mortality was observed in two out of three animals, then the dose administrated was assigned as toxic dose.If mortality was observed in one animal, then the same dose repeated again to confirm the toxic dose.If mortality was not observed, the procedure was repeated for higher doses such as 50, 100 and 1500 mg/ kg body weight.Animals were observed individually after dosing for first 30 min, 1 hour, 2 hours and daily thereafter, till 14 days.Any toxicity sign of gross changes in skin and fur, eyes and mucous membranes, circulatory, respiratory, autonomic and central nervous systems, and behavior pattern was reported (Kumudhavalli and Jaykar, 2012).

Induction of diabetes
Type 2 diabetes mellitus was induced by injecting freshly prepared streptozotocin (50 mg/kg; i.p.) in cold citrate buffer (0.1M, pH 4.5) (Ramachandan et al., 2013) in overnight fasting experimental rats.Confirmation of diabetes was decided to measure blood glucose level after 72 hours of injecting streptozotocin.Each time blood from tail vein of experimental rats is collected and blood glucose was measured with glucometer during whole study.Animals were kept in laboratory condition for 7 day to stabilize diabetes and animals showing blood glucose >250 mg/dL taken for activity assessment of synthesized drugs.

Study design and grouping of animals
Animals were randomly divided into 13 groups (n = 5) for the whole study. 1 st group treated as normal control and administered 0.3% carboxymethyl cellulose (CMC). 2 nd group remains diabetic control and no drug administered throughout the study period.Standard drug (miglitol) at a dose level of 25 mg/kg (p.o.) body weight was administered to 3 rd group.Others groups receive synthesized drugs of two dose level i.e. 100 mg/ kg and 250 mg/kg body weight.All the drugs were suspended on freshly prepared CMC (0.3% w/v) before administered.The treatments were administered orally to the animals daily for 14 days.Blood samples were collected from tail vein for determination of blood glucose level on 0, 5 th , 10th and 15 th day (Selvan et al., 2008).The blood glucose levels were measured by one touch glucometer (AccuSure) throughout the two weeks of treatment.

Body weight measurement
Animals were weighed on 0, 5 th , 10 th and 15 th day after treatment to detect any change in their body weights.

Blood sample collection
On 15 th day blood samples were collected from retroorbital puncture under mild anesthesia and stored with 4%w/v sodium citrate for plasma separation.The animals were sacrificed through cervical dislocation after subjecting them to mild ether anesthesia.Blood was carefully collected from each rat from retro-orbital puncture under mild anesthesia in microfuge tube and then centrifuged at 2,500 rpm for 15 min.Plasma, thus obtained, was collected carefully in individual microfuge tube and stored at -20°C.Plasma was separated immediately and used for the analysis of total cholesterol (TC) and triglyceride (TG).

Biochemical estimation
Serum marker such as serum glutamic oxaloacetic transaminase (SGOT), serum glutamic pyruvic transaminase (SGPT) and alkaline phosphatase (ALP) were measured (Cole et al., 2005;Silverstein and Webster, 1963).Lipid profile like triglyceride and total cholesterol were also measured using test kits.

Statistical analysis
All the data were expressed as mean ± SEM.Statistical analysis was carried out by one way ANOVA (Analysis of Variance) followed by Dunnett's t-test with the level of significance at p<0.01 and p<0.05.

Drug likeness studies
The compounds were derived from the xanthone nucleus with various substituents (Table I).Pharmacokinetics properties and toxicity studies of all the derived compounds were studied.Molecular properties of the designed ligands such as logP, molecular weight, HBA (hydrogen bond acceptor), HBD (hydrogen bond donor), nRB (number of rotatable bond), TPSA (total polar surface area) were calculated by Molinspiration property calculator and are reported in Table II.Most of the compounds showed good predictive bioavailability and pharmacokinetics hardly violating the Lipinski rule.The ligands were further studied for different property such as mutagenicity, tumourogenicity, irritating and reproductive effects as well as drug likeness and drug scores and were predicted.

Docking study
Prior to the docking simulation, the authors were applied the re-docking process to validate the docking protocol for its reproducibility.The co-crystal migiltol docked into the binding pocket of human maltase-glucoamylase.The structural superimposition of the crystal ligand along with the crystal ligand was perfomed by AutoDock 4.2.The docked pose of Migiltol achieved by AutoDock and the co-crystal conformation results an RMSD of 1.33 Å (Figure 1).The RMSD of <2 Å were considered as success wherever the RMSD between 2 and 3 Å were believed to be partially successful.It was found that all the interactions and interacting residues in the docked pose were identical with the co-crystal ligand.Lowest RMSD value, similar interactions and binding poses with same interacting residues between the docked ligand and the crystal ligand validates our docking protocol to be optimal.The molecules were targeted to the binding site of the standard compound.
The docking results predicted that migiltol had less binding affinity as compared to the most of the derived compounds.
Docking results were predicted that the standard ligand had a -8.71 Kcal/mol binding energy.The hydrogen bond forming amino acid residues of the target protein was found to be Asp327, Trp406, Asp443 and Met444.Excluding hydrogen interaction, the ligand was also formed nine hydrophobic bonds with the amino acids (Figure 2).The ligand p22 was predicted to have better binding energy as compared to the standard drug.It formed a -12.99 Kcal/mol binding energy.It only formed seven hydrophobic bonds with the target protein to the same binding pattern as of the standard (Figure 3).Docking interaction of the ligand p2 was formed -10.2 Kcal/mol binding energy.Two hydrogen bond interactions were identified with the amino acid residue Trp406 and Arg526.It also formed twelve hydrophobic interactions with the protein.The ligand p5 was formed a -9.99 Kcal/mol binding energy.It was fitted to the binding pocket of the standard molecule forming ten hydrophobic interactions.The ligand p10 was inhibited the protein having binding energy -9.93 Kcal/mol.The hydrogen bond interactions were identified with the helix of protein.Two hydrogen bonds were formed by the ligand with the residue Tyr299 and Arg526.The number of hydrophobic interacting residues was found to be ten.The ligand p4 was formed -9.89 Kcal/mol binding interaction energy.It was inhibited the target protein with eight hydrophobic bonds.

Chemistry
Five best scoring (lesser binding energy) compounds (P2, P4, P5, P10 and P22) were selected for synthesis.The synthesis of selected novel alkoxyxanthones was achieved in two step reaction (Scheme 1).In the first step, salicylic acid derivative (1) and poly-hydroxy phenols (2) were reacted together to yield hydroxyxanthones (3) in the presence of acid.Hydroxyxanthones (3) was further reacted with alkyl bromide in the presence potassium carbonate, acetone and reflux for 4 hours at 55ºC to yield targeted alkoxyxanthones (4).

Anti-diabetic activity evaluation
In vivo acute toxicity studies The synthesized compounds showed no serious toxicity up to dose level 1,500 mg/kg body weight in experimental rats.

Blood glucose level
Before starting the treatments, blood glucose level of all the animals was within normal range.After 72 hours of streptozotocin treatment, the blood glucose level was significantly changed more than 240 mg/dL.There was significant decrease (p<0.01) in blood glucose level of the animals after treatment with drugs on 5 th day, 10 th day and 15 th day (Table III).

Body weight
Statistical analysis revealed that there was significant (p<0.01)difference in body weight of experimental animals, among the group at 5 th day, 10 th day and 15 th day when compared with normal control group.But there are no significant changes in body weight between the groups when compared with diabetic control group.Effects of drugs over body weight and the changes in body weight after two weeks of treatment were given in the Table IV.

Effect on SGOT, SGPT, ALP, TC and TG level
Activity of the enzymes SGOT, SGPT and ALP were measured from the plasma of the experimental animals.The level of TC and TG content was also measured from the plasma collected from the rats after 14 days treatment using biochemical kits and results are represented in Table V.
Elevation of biomarker enzymes such as SGOT, SGPT ALP were observed in diabetic group which indicates hepatic damage.On treatment with synthesized compounds reduced the levels of the elevated marker enzymes i.e.SGOT, SGPT, ALP and restored these close to normal values which indicates recovery of insulin secretion.So it can be clearly understood that there is positive effect which is statistically significant (p<0.01).

Discussion
Type 2 diabetes mellitus is major health problem.In this manuscript we present an approach to control blood glucose levels in individuals with type 2 diabetes by targeting maltase-glucoamylase and intestinal glucosidases using some novel xanthone alpha-glucosidase inhibitors.One of the intestinal glucosidases targeted the N-terminal catalytic domain of maltase-glucoamylase (ntMGAM) which is responsible for the hydrolysis of terminal starch products into glucose (Sim et al., 2010).Hence to slow down the glucose level we targeted the ntMGAM.Previously acarbose and miglitol were found to be potent inhibitors alpha-glucosidase having polyhydroxy groups in their structure (Asano et al., 2003).Hence novel molecules are designed to xanthone based on the fact.Computational docking study was done to the binding pocket of miglitol with the crystal structure of maltase-glucoamylase.We redocked to verify the interaction of the cocrystal ligand (miglitol) with the novel xanthone molecules.We observed excellent correlation on the docking interaction we synthesized the most effective compounds.In vivo acute toxicity study showed the synthesized compounds showed no serious toxicity up to dose level 1500 mg/kg body weight in experimental rats.The molecules were screened for the in vivo antidiabetic activity in streptozotocin-induced diabetic animal model.After 72 hours of STZ treatment the blood glucose level was significantly changed more than 240 mg/dL.Statistical analysis by one way revealed that there was significant decrease (p<0.01) in blood glucose level of the animals after treat -ment with drugs on 5th day, 10th day and 15th day.
The designed ligands were predicted for bioavailability and others molecular property to check their potential as future drug candidate and they hardly violating those filters.When they docked in the same binding pocket with same amino acid residues of targeted protein most of the derived ligands predicted better binding energy than miglitol.Among 28 designed compounds the compound p22 showed to have the best binding energy (-12.99Kcal/mol) however the compound p26 showed the least binding energy (-7.17Kcal/mol).Based on these in silico results further proceded for the laboratory experiment.Top five compounds were synthesized in laboratory scale, purified and crystallized properly to characterize them by UV, FTIR, NMR and mass analysis.The analytical and spectral data of the compounds analyzed and are found in compliance with the structure of the synthesized compounds.
The compounds were further screened for in vivo antidiabetic activity using streptozotocin-induced diabetic model in Wistar rats.Effective dose was selected after oral acute toxicity study.All the study group animals were subjected to 15 days anti-diabetic treatment protocol and blood glucose level were measured on 0 th , 5 th , 10 th and 15 th day collecting blood from tail vein of study subjects.Study shows a significant reduction in blood glucose as well as it was also noted that the total cholesterol and triglycerides level was increased in diabetic control groups and there was a notable change after 15 days treatment with synthesized compounds.On treatment with synthesized compounds total cholesterol and triglyceride level decreased significantly (p<0.01) as compared with diabetic control group.

Conclusion
Xanthone derivatives proved to act as lead molecules towards the development of potential -glucosidase inhibitors.These compounds showed excellent correlation between docking results, synthetic data and in-vivo anti-diabetic activity.

Figure 1 :
Figure 1: The re-docking pose of miglitol crystal ligand.The crystal ligand is marked in red color

Figure 3 :
Figure 2: The docked pose of standard miglitol

Table III Effect on blood glucose level of synthesized compounds
All the values are given as mean ± SEM; n = 5; Diabetic control vs all group ( a p<0.01, b p< 0.05), Normal control vs all group ( c p<0.01, d p<0.05)

Table V Effect on SGOT, SGPT, ALP, TC and TG levels
Diabetic control vs all group ( a p<0.01, b p< 0.05), Normal control vs all group ( c p<0.01, d p<0.05)