Abstract

Saturated free fatty acid-induced adipocyte inflammation plays a pivotal role in implementing insulin resistance and type 2 diabetes. Recent reports suggest A2A adenosine receptor (A2AAR) could be an attractive choice to counteract adipocyte inflammation and insulin resistance. Thus, an effective A2AAR agonist devoid of any toxicity is highly appealing. Here, we report that indirubin-3′-monoxime (I3M), a derivative of the bisindole alkaloid indirubin, efficiently binds and activates A2AAR which leads to the attenuation of lipid-induced adipocyte inflammation and insulin resistance. Using a combination of in silico virtual screening of potential anti-diabetic candidates and in vitro study on insulin-resistant model of 3T3-L1 adipocytes, we determined I3M through A2AAR activation markedly prevents lipid-induced impairment of the insulin signaling pathway in adipocytes without any toxic effects. While I3M restrains lipid-induced adipocyte inflammation by inhibiting NF-κB dependent pro-inflammatory cytokines expression, it also augments cAMP-mediated CREB activation and anti-inflammatory state in adipocytes. However, these attributes were compromised when cells were pretreated with the A2AAR antagonist, SCH 58261 or siRNA mediated knockdown of A2AAR. I3M, therefore, could be a valuable option to intervene adipocyte inflammation and thus showing promise for the management of insulin resistance and type 2 diabetes.

Introduction

Insulin resistance is an impairment of insulin-stimulated glucose disposal in insulin-responsive cells which is a major defect and early sign for future development and progression of type 2 diabetes (T2D) pathogenesis [1,2]. Over the last decade, mounting evidence has emerged demonstrating a close link between a state of chronic low-level inflammation in adipose tissue and obesity-induced insulin resistance and T2D [29]. Increasing accumulation of intra-abdominal adipose tissue in obese subjects is frequently associated with the enhanced rate of free fatty acids (FFAs) mobilization and higher levels of circulating FFAs [10,11] which trigger inflammatory pathways that compromise insulin sensitivity [29]. Consequently, targeting adipose tissue inflammation by genetic knockdown of inflammatory receptors or mediators has a beneficial effect on insulin sensitivity and glucose homeostasis [814].

Accumulating evidence highlights a critical role of adenosine signaling in the regulation of insulin synthesis from the pancreatic beta cells and also modulate the insulin responsiveness in adipose tissue, muscle, and liver that governs glucose homeostasis [15,16]. The purine nucleoside adenosine is an endogenous signaling molecule, and its physiological level is very low in the extracellular microenvironment. In response to cellular insult by metabolic stress, tissue injury and inflammation, increased accumulation of extracellular adenosine exerts a range of responses by binding with adenosine receptors (ARs) to succumb cellular homeostasis [1517]. Adenosine mediate its effects by binding to specific G-protein–coupled ARs which are widely distributed in metabolically active sites such as adipose tissue, liver, pancreas and various immune cells. Among the four different subtypes of AR, A1, A2A, A2B, and A3 [16], adenosine orchestrates its anti-inflammatory effect through the activation of A2AAR and A2BAR [1822] and therefore the signaling pathway of these receptors subtypes are more intensely studied to counter the pathophysiology of various inflammatory diseases including T2D.

Natural products, particularly phytochemicals, have been traditionally utilized for the management of various human diseases [23] and to develop various derivatives with reduced toxic side effects, improved pharmacokinetics and enhanced efficacy [24]. Indirubin, an active component of Indigo naturalis and 3,2′ bis-indole isomer of indigo, has been shown to be the main ingredient of the traditional Chinese herbal medicine, Danggui Longhui Wan, which is used to treat various leukemias and inflammatory diseases [2528]. Indirubin-3′-monoxime (I3M) is one of the most studied synthetic indirubin derivatives exhibiting higher potency and bioavailability in comparison to its parent compound [29]. Indirubin and its analogs act as potent inhibitors of cyclin-dependent kinases, glycogen synthase kinase-3β, JAKS/Src family kinase and nuclear factor-kappa B [2932] which confers their potential therapeutic efficacies as anti-cancer, anti-angiogenic, anti-viral, anti-aging, anti-inflammatory and anti-diabetic functions. However, the underlying mechanisms of many of these effects remain largely unexplained and that considerably delays their integration into the modern health-care system.

In searching for an effective A2AAR agonist, investigators primarily relied on the modification of adenosine [33] considering a time long notion that preserving sugar moiety is critical for receptor activation [34]. We found that I3M, a non-adenosine chemotype, is capable of binding and activating A2AAR signaling which effectively attenuates lipid-induced adipocyte inflammation and insulin resistance. Thus, our study revealed that I3M could be an effective therapeutic alternative to alleviate lipid-induced adipocyte inflammation and insulin resistance.

Material and methods

Homology modeling of A2AAR protein

The agonist conformation of A2AAR protein was selected in the present study and its crystallographic structure, bound to a G protein (PDB code - 5G53) [35] has been downloaded from the Protein Data Bank. The active site information of the structure has been reported previously. Missing structures were observed to be present in some regions of the crystal structure at the N-terminal (1–5 residues), C-terminal (313–314) and from residues ‘147–158’ and ‘212–223'. To build the final seven trans-membrane domain structure, a multi-template homology modeling approach was applied using modeler 9.17 [36]. The native structure of the selected A2AAR was conserved while the missing region has been modeled using another template of the A2AAR (PDB code: 5IU4). Although the missing residues are not from the active site region of the protein, the modeling of the residues was done for a complete structure of the receptor. Structure minimization has been done through Chimera [37] using 100 steepest descent and 10 conjugate gradient steps. Minimized structure of the modeled protein (modeling of only missing residues) has been considered as the A2AAR protein and used for further analysis.

In silico screening of putative A2AAR agonist

We employed various in silico screening approaches to find out putative A2AAR agonist molecules having drug-like properties and binding potential towards the receptor [38]. To do this, a dataset of molecules was created and listed in Supplementary Table S1 for the screening purpose to discover probable candidates of A2AAR agonist.

(a) Druglikeness property: The first screening approach was applied to check the druglike properties of the listed compounds. It has been shown that druglikeness of a molecule depends on certain properties known as the Lipinski's rule of five [39] such as molar mass, hydrogen bond donor, acceptor and logP should be satisfied. The oral bioavailability of the compound was found to be correlated with these properties. It plays a major role in screening of the compounds having properties of known drug molecules. According to this rule, the number of hydrogen bond donors and the partition coefficient (logP) should not be greater than 5. Other properties include the molar mass (≤500), the number of H-bond acceptors (≤10) and the molar refractivity (40–130). Poor absorption of the compounds is seen when these rules are violated. Molecular properties for Lipinski's rule of five predictions of the selected compounds were computed using the Sanjeevini server under SCFBio [40]. A soft screening approach was applied where molecules not satisfying at-least two rules were excluded from the dataset. To further minimize the set of ligand molecules, one more screening approach was applied after the Lipinski rule of five predictions, known as the Veber's rule [41]. This rule states that structural properties of the molecules like the rotatable bond count should be ≤10 and the polar surface area (PSA) ≤ 140. The properties of the compounds were calculated through the open-source virtual screening tool DruLiTo. Molecules not following at least one of this rule were removed and selected molecules were used for further analysis.

(b) QSAR analysis: To develop a correlation between the chemical properties of the molecules with their biological activity in terms of mathematical equations we utilized quantitative structure-activity relationship (QSAR) analysis [42]. The physicochemical properties of the molecules are considered as the independent variables and parameter related to the biological activity as the dependent variable. Through this approach, the effect of physicochemical properties on biological activity can be analyzed. In our study, we considered biological activity in terms of Lethal Dose (LD50) of the compounds. Filtering of potent compounds on the basis of its toxicological properties has been widely applied in screening approaches [43]. LD50 has been used as a toxicity measure to understand what concentration of the drug is toxic for the individuals. The compound prior to administration to humans undergoes a series of tests mostly in rodents. If a particular dose of the compound causes deaths of 50% of the tested populations, the dose is then termed as LD50 [44]. A higher concentration or LD50 value indicates a less toxic compound compared with a lower concentration or a smaller LD50 value. A literature search for the LD50 values was done for our dataset of molecules and observed that only a few molecules have experimental LD50 values. The values were converted to their negative logarithm (log 1/LD50), and used in the study. So, using the available values, a 2D-QSAR (using two dimensional descriptors), was applied to predict the LD50 values for the rest of the molecules in our dataset.

The Molecular Design Suite (VLife MDS 3.5, 2004) was used for performing 2D-QSAR studies. Different categories of descriptors like the molecular mass and volume, hydrogen bond donor and acceptor, XlogP, SlogP, rotatable bond count, molar refractivity (smr), polarizability, Polar Surface Area (PSA), Estate contributions, Estate Numbers, Estate topochemical based descriptors were calculated. Descriptors showing a high correlation or similar values in our dataset of molecules were removed and the rest of the descriptors were used as the independent variables for the generation of the equation. For generating the correlation equation between the dependent (LD50) and the independent variables (physicochemical descriptors), a regression-based approach; Multiple Linear Regression (MLR) was used. It attempts to calculate the correlation coefficients through straight-line fitting of the given data.

The regression equation generally takes the form of Y = b1 × x1 + b2 × x2 + b3 × x3 + c, Where, Y is the dependent variable, ‘b's are regression coefficients for the corresponding ‘x's (independent variable), and ‘c’ is a regression constant or intercept. This method of estimating the values of the biological activity is popularly known as Hansch model. The predicted model was evaluated by comparing the biological activity values of compounds in our dataset. Standard statistical parameters like r2 (coefficient of determination), internal predictive ability (q2), external predictive ability (pred_r2) and the F value has been considered. A higher value of correlation (r2 > 0.6) and a lower value of standard error like pred_r2_se and q2_se show good fitness of the model [45].

(c) ADMET Screening: ADMET (Absorption, Distribution, Metabolism, Excretion and Toxicity) properties play a crucial role in understanding the fate of the drug molecules inside the living organisms. For a drug to be bioactive it should be absorbed in the body and distributed properly to the target cells. In silico approaches are used to predict the bioactivity of the compounds by ADMET prediction. The ADMET profiles of the screened compounds were predicted through the admetSAR server [46]. A large number of ADMET properties: blood-brain barrier, intestinal absorption, Caco-2 permeability, P-glycoprotein substrate and inhibitor, substrate and inhibitor properties of different CYP450 isoforms (2C9, 2D6, 34A, 1A2, 2C9, 2D6, 2C19, 34A), human ether-a-go-go-related gene, AMES toxicity, carcinogens, biodegradation and acute oral toxicity has been predicted. Compounds shown to be toxic (AMES toxic, carcinogen) were removed from the dataset. To further screen out intestinal absorption property, P-glycoprotein (P-gp) substrate has been considered [47]. Compounds not supporting the P-gp substrate were excluded from the dataset.

(d) Molecular Docking: Molecular docking is widely used as in silico screening analysis in the identification of putative lead compounds forming stable interactions in the protein–ligand complex [48]. It has also been applied in identifying the native poses of ligands and retrieving information about the change in the functionality of a mutated receptor–ligand bound complex [49]. With the known structural information of the target protein and its reported active site, potential compounds which can bind to the protein could be screened out from a dataset of molecules. Protein-ligand docking was carried out through different approaches — (i) rigid docking keeping both the ligand and the protein rigid and (ii) inducing flexibility in the side chain atoms of the active site residues. The rigid docking was performed through the Autodock version 4.2.6 [50] using the Lamarckian genetic algorithm (LGA). Ligand binding to a receptor has been known to cause conformational changes to the receptor. Here, a docking approach was applied where partial flexibility is introduced to the receptor. The flexible docking method increases the conformational space of the ligand leading to the identification of an energetically favorable pose. The docking was performed using a Genetic Algorithm through AutoDockFR [51].

Ligand preparation: Ligand molecules, for the rigid docking simulations, were minimized using the Gabedit software [52] through the steepest descent method. The number of iterations for minimization used was10 000. Hydrogen atoms and charges were added to ligand molecules. A reference ligand was prepared for the flexible docking approach. The conformation of the agonist molecule present in the crystal structure of the receptor has been considered for this purpose. This leads to the generation of the protein binding site information.

Target preparation: The minimized structure of the target protein was used in both type of docking simulations with the selected ligands. Polar and non-polar hydrogen atoms were added to the target protein through Autodock with the addition of Gasteiger charges. The non-polar hydrogen atoms were merged and the protein was saved in the docking format. The prepared protein was considered as the target in all the docking approaches. Fifteen residues (Val84, Leu85, Thr88, Gln89, Phe168, Glu169, Met177, Asn181, Trp246, His250, Asn253, Met270, Ile274, Ser277 and His278) of the A2A adenosine receptor were reported to be in the active site. For docking of the screened molecules into the binding pocket of the receptor, partial change in the active residues was introduced. The side-chain atoms of these residues were made flexible, while the backbone atoms were kept as they were. The ligands were docked to this flexible receptor. Various parameters are incorporated by the algorithms employed for the docking which are given in Supplementary Table S4. The rest of the parameters were set as default in both software used. Ligplot [53] analysis of the docked complexes was executed to find the nonbonded interactions between ligand and the receptor. The selected ligands underwent experimental validation.

Reagents and antibodies

All tissue culture materials were obtained from Gibco, Thermo-Scientific, Grand Island, NY. We purchased phospho-Akt (Thr-308; Cat. No. #sc-16646-R), phospho-Akt (Ser-473; Cat. No. #sc-7985-R), Akt (Cat. No. #sc-8312), IRS-1 (Cat. No. #sc-7200), Glut-4 (Cat. No. #sc-7938) antibodies from Santa Cruz Biotechnology, Santa Cruz, CA. Phospho-IκBα (Ser-32; Cat. No. #2859), phospho-CREB (Ser-133, Cat. No. #9198), phospho-p44/42 MAPK (Erk1/2) (Thr-202/Tyr-204; D13.14.4E; Cat. No. #4370), p44/42 MAPK (Erk1/2) (137F5; Cat. No. #4695), phospho-p38 MAPK (Thr-180/Tyr-182; D3F9; Cat. No. #4511) and p38 MAPK (D13E1; Cat. No. #8690) and Phospho-PKC (pan, βII Ser660; Cat. No. #9371T) antibodies were procured from Cell Signaling Technology, Danvers, MA. Phospho-NF-κBp65 (Ser-281; Cat. No. #SAB4301496), Phosphotyrosine (Cat.No. #P5872), Horseradish peroxidase-conjugated anti-rabbit (Cat. No. #A9169) and anti-mouse (Cat. No. #A9044) secondary antibodies were purchased from Sigma–Aldrich, St. Louis, MO. We procured β-actin monoclonal antibody (AC-15) (Cat. No. #AM4302), Alexa Fluor 488 conjugated goat anti-rabbit IgG (H + L) secondary antibody (Cat. No. #A11034), Recombinant Protein G Agarose (Cat. No. #15920010), Lipofectamine 2000 Transfection Reagent (Cat. No. #11668027) and NP40 lysis buffer (Cat. No. #FNN0021) from Invitrogen, Thermo-Scientific, Grand Island, NY. Millicell EZ SLIDES (Cat. No. #PEZGS0896) was obtained from EMD-Millipore, Darmstadt, Germany. Halt Protease and Phosphatase Inhibitor Cocktail (Cat. No. #78440) was purchased from Thermo-Scientific, Grand Island, NY. ClarityTM Western enhanced chemiluminescence (ECL) Substrate, Precision Plus Western C Pack (Cat. No. #161-0385), iScript Reverse Transcription Supermix (Cat. No. #170-8841) and iTaq™ Universal SYBR® Green Supermix (Cat. No. # 1725121) were obtained from Bio-Rad Laboratories, Hercules, CA. Indirubin-3′-monoxime (Cat. No. #I0404), Cytisine (Cat. No. #C2899), 8-Cyclopentyl-1,3-dipropylxanthine (Cat. No. #C101), SCH 58261 (Cat. No. #S4568), Alloxazine (Cat. No. #A28651), VUF 5574 (Cat. No. #V5888) and Adenosine deaminase (Cat. No. #90751) were procured from Sigma–Aldrich, St. Louis, MO. Glucose Uptake Cell-Based Assay Kit (Cat. No. #600470) and Glycerol Cell-Based Assay Kit (Cat. No. #10011725) were procured from Cayman, Ann Arbor, MI. Cyclic AMP XP Assay Kit (Cat. No. #4339) was purchased from Cell Signaling Technology, Danvers, MA. We procured Legend Max Mouse IL-6 (Cat. No. #431307) and IL-10 (Cat. No. #431417) ELISA Kit from BioLegend, San Diego, CA. We purchased Steady-Glow Luciferase Assay System (Cat. No. #E2510) from Promega, Madison WI; RNeasy Lipid Tissue Mini Kit (Cat. No. #74804) from Qiagen, Hilden, Germany; Vectashield anti-fade mounting medium containing DAPI (Cat. No. #H-1200) from Vector Laboratories, Burlingame, CA; κB luciferase plasmid (Cat. No. #BB-V0035) from BioBharati Life Science, Kolkata, India; and EZcountTM LDH Cell Assay Kit, Homogeneous (Cat. No. #CCK058) from HiMedia Laboratories, Mumbai, India. pGL2B −1538/+64 and CMV500 A-CREB were a kind gift from Stephen Smale (Addgene plasmid #24942) and Charles Vinson (Addgene plasmid # 33371), respectively. Control siRNA (Cat. No. #sc-37007), Adenosine A2AAR siRNA (Cat. No. #sc-39851) and CGS 21680 Hydrochloride (Cat. No. #sc-211062) were purchased from Santa Cruz Biotechnology, Santa Cruz, CA. A2AAR cDNA ORF Clone (Cat. No. #HG12307-NH) was purchased from Sino Biological Inc., Beijing, China. Different gene-specific primers were procured from Integrated DNA Technologies, and Imperial Life Science (P) Limited, India. All other chemicals and reagents used were purchased from Sigma Chemical Co., St. Louis MO, U.S.A.

Cell culture and treatments

Mouse 3T3-L1 preadipocytes (Cat. No. #SP-L1-F) were procured from the ZenBio, NC, U.S.A. and cultured in Preadipocyte Medium (Cat. No. #PM-1-L1, ZenBio, NC, U.S.A.) supplemented with 1% Penicillin–Streptomycin solution (100 µg/ml) in a humidified 5% CO2 environment at 37°C. Confluent 3T3-L1 preadipocytes were differentiated using Differentiation Medium (Cat. No. #DM-2-L1, ZenBio, NC, U.S.A.). Rat L6 myoblasts were obtained from the National Center for Cell Science (NCCS), Pune, India and cultured in a similar manner as described by us previously [9]. We performed dose kinetics study of Akt and CREB phosphorylation at various concentrations of indirubin-3′-monoxime. In case of incubations with inhibitor or activator, cells were pretreated for 1 h with indirubin-3′-monoxime (10 µM), CYT (100 µM), adenosine deaminase (0.00138 U/ml), SCH 58261 (300 nM), 8-Cyclopentyl-1,3-dipropylxanthine (100 nM), Alloxazine (1.34 µM), VUF 5574 (100 nM), CGS 21680 (1 µM). Upon termination of incubations, cells were washed twice with ice-cold Dulbecco's phosphate-buffered saline (DPBS) and harvested with trypsin–EDTA solution. Harvested cell pellets were resuspended in NP40 lysis buffer supplemented with Halt protease and phosphatase inhibitor cocktail, vigorously vortexed in every 10 min for 30 min, centrifuged for 10 min at 13 000 rpm at 4°C and the supernatant was collected. The protein concentration of the supernatant was determined by following the method of Lowry et.al. [54].

Development of A2AAR stable clone

For the establishment of Chinese hamster ovary (CHO) cell line stably expressing human A2AAR, CHO cells were transfected with pCMV3-His-ADORA2A using Lipofectamine 2000 according to the manufacturer's instructions. After hygromycin B selection at 500 µg/ml for 3 weeks, stable transfectants were obtained and single clonal cell line (CHO/ADORA2A) was isolated by limiting dilution. Expression of the A2AAR receptor was verified in RT-qPCR analysis (Supplementary Fig. S2C). A2AAR overexpressing clonal cells were used for the determination of EC50 value (the concentration that produces a half-maximum response) of indirubin-3′-monoxime by performing a cAMP assay.

Radioligand-binding assay

Radioligand binding experiments were performed following the previously described procedures [55,56]. The human A2AAR or A2BAR were individually transfected into CHO cells and studied in membranes prepared from these cells [55]. The radioligand [³H]NECA (10 nM) was used for A2AAR. Due to the lack of a useful A2BAR radioligand, the relative affinity for this receptor subtype was determined in adenylyl cyclase experiments as described previously [55].

Glucose uptake assay

Glucose uptake assay was performed using a glucose uptake cell-based assay kit (Cayman, U.S.A.) following manufacturer's instruction. Briefly, 3T3-L1 adipocytes were serum-starved overnight in Kreb's Ringer Phosphate (KRP) buffer supplemented with 0.2% bovine serum albumin (BSA). Cells were pretreated with different compounds for 1 h followed by palmitate (0.75 mM) incubation for 6 h and 30 min before the termination of incubations; cells were treated with insulin (100 nM). Fluorescent labeled glucose analog 2-NBDG was added to each of the incubations for 5 min before termination of the experiment. Cells were then lysed and fluorescent intensity was measured by Varioskan LUX Multimode Microplate Reader (Thermo Scientific, Finland).

Cell viability assay

Cell viability was assessed using the lactate dehydrogenase (LDH) release and MTT assays. LDH release was measured using EZcountTM LDH Cell Assay Kit following the manufacturer's instructions. Briefly, confluent 3T3-L1 adipocytes were either subjected to lysis to measure the maximum LDH release or treated without or with different concentrations of I3M (0, 5, 10, 50 µM) for 6 h followed by the addition of LDH reagent to each well and incubated for 10 min at room temperature. Stop solution then added to terminate the incubations and fluorescent intensity (ex/em 560/590 nm) was measured by Varioskan LUX Multimode Microplate Reader (Thermo Scientific, Finland). LDH release was calculated as Experimental-Background control/ Max.LDH control-Background control × 100.

MTT assay was performed following the method described previously [57]. Briefly, 3T3-L1 adipocytes were incubated with varied concentration of I3M or CYT for 24 h followed by the addition of MTT and incubated for 4 h. On termination of incubations, formazan crystals formed in cells were dissolved in acidic isopropanol and incubated further for 30 min at 37°C. Cytotoxicity was measured spectrophotometrically at 570 nm with Varioskan LUX Multimode Microplate Reader (Thermo Scientific, Finland). Absorbance values were blanked against acidic isopropanol and the absorbance of cells exposed to medium only (without any treatment) were taken as 100% cell viability (control).

Immunoblotting

Immunoblot analysis was performed following our previously described method [9]. Briefly, cell lysates (40 µg of protein) were subjected to either 10% or 12.5% SDS–PAGE and transferred on to Immbilon­P PVDF membranes (Millipore, Bedford, MA) with the help of Wet/Tank Blotting System (Bio-Rad Laboratories, Hercules, CA). Membranes were first blocked with 5% BSA in TBS (Tris-buffered saline) buffer for 1 h followed by the overnight incubation with primary antibodies (1 : 500 or 1 : 1000 dilutions) in a rotating shaker at 4°C. The membranes were then washed three times with TBST (TBS containing 0.1% Tween 20) buffer for 10 min interval and incubated with peroxidise conjugated goat anti-rabbit or goat anti-mouse secondary antibodies (1 : 20 000 dilution) for 2 h at room temperature. Membranes were then washed three times with TBST for 10 min interval and subjected to ClarityTM Western ECL Substrate incubation for 5 min at room temperature. Protein bands were visualized and quantified in Chemidoc XRS+ System (Bio-Rad Laboratories, U.S.A.) using Image Lab Software.

Coimmunoprecipitation

Coimmunoprecipitation study was performed according to our earlier published method [9]. Briefly, 200 µg of protein from cell lysate was incubated with 2 µg of the anti-IRS antibody for overnight at 4°C in a shaking platform followed by the incubation with 50 µl of Protein-G Agarose for 1 h under rotation at 4°C. The samples were then centrifuged at 5000 rpm at 4°C for immune-complex precipitation. Pelleted immune-complex was washed thoroughly, boiled in 4× SDS sample buffer, vortexed and then centrifuged at 13 000 rpm for 10 min. The supernatant was isolated, run on 10% SDS–PAGE gel and transferred on to PVDF membrane. This was followed by immunoblotting with anti-phosphotyrosine or anti-IRS1 antibodies (1 : 1000 dilution). The blots were subjected to ClarityTM Western ECL Substrate and protein bands were observed and quantified in Chemidoc XRS+ System (Bio-Rad Laboratories, U.S.A.) using Image Lab Software.

Semi-quantitative RT-PCR and real-time quantitative PCR

Total RNA was extracted from the cells of different incubations using RNeasy Lipid Tissue Mini Kit (Qiagen, Germany) according to the manufacturer's instruction. RNA was treated with DNase I and reverse transcribed using the iScript Reverse Transcription Supermix. We used 2X PCR Master Mix for semi-quantitative RT-PCR in BioRad C-1000 Thermal Cycler and iTaq™ Universal SYBR® Green Supermix to perform real-time quantitative PCR in ABI-7500 system using gene-specific primers. The following cycling conditions were used for real-time qPCR: 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, 55°C for 30 s, and 72°C for 30 s. After the final extension, a melting curve analysis was performed to ensure the specificity of the products. The fold changes in expression were determined using 2ΔΔCt and the expression of target genes were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression. Primer sequences used for PCR analysis are listed in Supplementary Table S5.

Immunofluorescence analysis

L6 cells were cultured on sterile Millicell EZ-SLIDES (EMD-Millipore, Germany) and treated with different conditions. On termination of incubations, cells were fixed with 4% paraformaldehyde for 10 min followed by blocking with 2% BSA in PBS for 1 h at room temperature. Cells were then incubated with anti-Glut4 antibody (1 : 50 dilution) in 2% BSA in PBS overnight at 4°C in a rotating platform. After washing with ice-cold PBS, cells were incubated with AlexaFluor 488-conjugated goat anti-rabbit secondary antibody (1 : 200 dilution) for 1 h at room temperature. Cells were then washed thrice with ice-cold PBS and mounted in Vectashield anti-fade mounting medium containing DAPI (Vector Laboratories, U.S.A.). Cellular images were taken in an inverted fluorescent microscope (Leica DMi8, Germany) using LAS X software.

Enzyme-linked immunosorbent assay (ELISA)

To assess the protein levels of IL-6 and IL-10 in the control and treated cell culture supernatants, ELISA were performed using LEGEND MAX™ mouse IL-6 and IL-10 ELISA kits according to the manufacturer's instructions.

Cyclic AMP assay

The cytosolic cAMP level was measured using the Cyclic AMP XP Assay Kit (Cell Signaling Technology, U.S.A.) in accordance with the manufacturer's protocol. Briefly, CHO cells stably expressing A2AAR were treated without or with IBMX (0.5 mM) for 30 min followed by the 15 min incubation of varied concentrations of indirubin-3′-monoxime. On termination of incubations, cells were lysed with 100 µl of lysis buffer and 50 µl of cell lysate was incubated with 50 µl of HRP-linked cAMP solution in 1 : 1 ratio for 3 h at room temperature on a horizontal orbital plate shaker. After incubation, the plate content was discarded and wells were properly washed with 1× wash buffer. An amount of 100 µl of TMB substrate was added to the wells and incubated until color develops. Upon color development, 100 µl stop solution was added and absorbance is measured at 450 nm in Varioskan LUX Multimode Microplate Reader (Thermo Scientific, Finland). The percentage of activity was calculated. % activity = 100 × [(A − Abasal)/(Amax − Abasal)], where A is the sample absorbance, Amax is the absorbance at maximum stimulation, and Abasal is the absorbance at basal level (without indirubin-3′-monoxime). The % activity was plotted versus the concentrations of indirubin-3′-monoxime and the dose-response curve fitted to a non-linear regression model using GraphPad Prism 5 software (San Diego, CA) for the determination of EC50 value of indirubin-3′-monoxime.

Glycerol release assay

Lipolysis was determined in 3T3-L1 adipocytes by measuring free glycerol level in the control and treated cell media using Glycerol Cell-Based Assay Kit following the manufacturer's instructions.

Luciferase reporter assay

3T3-L1 adipocytes (2 × 105 cells/well) were transfected with either κB-luciferase or IL-10 promoter-luciferase expression plasmid (0.25 mg/well) using Lipofectamine 2000 Transfection Reagent (Invitrogen, U.S.A.) following manufacturer's protocol. Briefly, 7.5 µl of Lipofectamine 2000 reagent and 6 µl of 0.25 mg κB-luciferase or IL-10 promoter-luciferase plasmid were added separately into 100 µl of Opti-MEM medium. After 5 min incubation, both solutions were mixed and incubated for 30 min. The transfection mixture was added to the cells containing 0.8 ml of 2% FBS containing Dulbecco's modified Eagle's medium (DMEM) without antibiotics. After incubation at 37°C for 6 h, the culture medium was changed to DMEM containing 10% FBS. After 48 h of transfection, cells were washed with DMEM and used for different incubations. On termination of incubations, cells were lysed and luciferase activity was measured using Steady-Glo Luciferase Assay System (Promega, U.S.A.) with the help of Varioskan LUX Multimode Microplate Reader (Thermo Scientific, Finland).

RNA interference study

3T3-L1 adipocytes were transfected with Control siRNA or A2AAR siRNA using Lipofectamine 2000 Transfection Reagent (Invitrogen, U.S.A.) following the manufacturer's protocol. The transfection mixture prepared in Opti-MEM was added to the cells and incubated for 6 h at 37°C. After the addition of 20% FBS in culture media, cells were kept for an additional 18 h. Media replaced with fresh culture medium containing 10% FBS and incubated for 48 h. After 48 h of transfection, knockdown efficiency was analyzed by RT-qPCR. Control or A2AAR siRNA transfected cells were incubated without or with indirubin-3′-monoxime (10 µM).

Statistical analysis

All data were derived from at least three independent experiments and statistical analyzes were conducted using Sigma Plot 10.0 software. Data were analyzed by one-way analysis of variance (ANOVA), where the P value indicated significance, means were compared by a post hoc multiple range test. All values were means ± SEM. A level of P < 0.05 was considered significant.

Results

In silico studies for screening of potential A2A adenosine receptor (A2AAR) agonists

To find out effective A2AAR agonists of non-adenosine structures, we selected 142 potential anti-diabetic compounds of different categories like flavonoid, alkaloids, terpenes and sulfonylurea from the literature (Supplementary Table S1) and investigated their binding affinity with the agonist conformation of A2AAR using various in silico approaches. These molecules were initially selected for the screening processes against the chosen target A2AAR and were included in our dataset. We first screened the druglikeness property of these compounds using Lipinski's rules which showed that 121 out of 142 molecules screened were validated by this test. Further screening with Veber's rule ruled out five more molecules, which were then excluded from the dataset. Therefore, 116 molecules favoring both Lipinski's and Veber's rules were selected for the next screening process. Out of 116 molecules, the experimental LD50 values of only 16 compounds were available which was used for the QSAR model generation. During the division of the molecules into the training and the test set, compounds behaving as outliers were removed. Two molecules were found to be not fitting the data and hence excluded from the QSAR model generation. From the 14 selected compounds with known LD50 values (Supplementary Table S2), we used nine molecules in training and five in the test set to analyze the fitness of our QSAR model. Out of 115 descriptors that were calculated for the compounds, only four descriptors were found to correlate with the toxicity values of the molecules. The equation generated is given as: 
formula
Where, x1, x2, x3 and x4 are the descriptors mentioned below:

x1 = Total number of nitrogen connected with one single and one double bond (SdsNcount)

x2 = Electrotopological state indices for number of –CH group connected with three single bonds (SsssCHE –index),

x3 = Electrotopological state indices for number of a carbon atom connected with one double and two single bonds (SdssCE –index),

x4 = Total number of nitrogen connected with three single bonds (SsssNE –index)

From the equation, it is observed that the three variables SdsNcount, (SsssCHE –index) and SsssNE -index show a positive correlation with the biological activity, while the variable SdssCE –index is correlating negatively. Both internal and external cross-validation was done for the model. The model had a variance of 98% (r2 = 0.99) with an internal validation of 65% (q2) and an external validation of 60% (pred_r2). The statistical parameters are given in Supplementary Table S3 and showed good accuracy between the actual and the predicted values. The observed and the predicted pLD50 values of the training and test set have been given in Supplementary Table S2. After observing the accuracy of the model, it was used for the prediction of the rest of the molecules in the screened dataset (100 molecules). Ranking of the molecules was done based on their pLD50 values. The compounds with lower pLD50 values have been selected for the analysis of their bioactive properties. The ADMET profiles of the 35 selected molecules from the QSAR studies were predicted using admetSAR. Through ADMET screening, the compounds that were predicted to be more bioactive and nontoxic were screened. However, one molecule from the dataset was predicted to be AMES positive and hence removed. Screening the compounds for their bioavailability and intestinal absorption, the P-glycoprotein (P-gp) substrate property has been considered and six molecules were found not to be substrates of P-gp, indicating that they will not bind to the transport protein and get efflux. Therefore, the six molecules showing good absorption, distribution and metabolic properties were selected based on their ADMET profiles.

When these selected six putative ligand molecules were docked to the target A2AAR, several conformations were obtained. The best binding pose for each putative ligand was analyzed. The Autodock results were evaluated based on the binding energy and the number of hydrogen bonds formed between the target A2AAR and the molecules in the docked complex. To find the optimal ligand–receptor conformation, two docking methods were applied. Along with the rigid docking, an amount of flexibility to the side chain atoms of the active site residues has been induced to get a better orientation of the ligands within the active pocket. Comparison of the docked complexes to the rigid docking showed that the putative ligands remained confined to the active site pocket with a slight variation in their poses. The docking parameters are provided in Supplementary Table S4 and the results of these two dockings are summarized in Table 1. Phe168 and Glu169 have been reported as important residues of A2AAR for its ligand binding interactions [58]. Binding of the screened molecules to these residues have shown that the molecules have the potential to alter the activity of A2AAR. Among these six putative ligand molecules, indirubin-3′-monoxime (I3M), has been found to be having the highest number of intermolecular interactions with the A2AAR suggesting a stable interaction between the receptor and the I3M molecule (Table 1). Further in both the docking scenarios, I3M along with CYT has shown the lowest binding energy (Table 1). Ligplot analysis of the A2AAR with I3M or CYT docked complexes (Figure 1A and Supplementary Fig. S1A) showed the formation of hydrogen bonding at Phe168 or Glu169 residues of A2AAR and also have lower energy in both the docking methods. The comparative analysis has suggested that I3M and CYT are forming stable complexes after an effective docking into the binding cavity of A2AAR, however, it could be interesting to note that His278 residue of A2AAR, critical for receptor activation, interacted only with I3M (Figure 1A and Supplementary Fig. S1A). The selected compounds specifically I3M shows optimal binding energy, hydrogen bonding and nonbonded interaction with the active site residues, and thus have a potential binding affinity towards the A2AAR.

In silico analysis of A2AAR-indirubin-3′-monoxime (I3M) interaction and its effect on insulin sensitivity and cellular viability.

Figure 1.
In silico analysis of A2AAR-indirubin-3′-monoxime (I3M) interaction and its effect on insulin sensitivity and cellular viability.

(A) Images representing the chemical structure of indirubin-3′-monoxime (left), and its interactions with A2AAR using ligplot analysis (right). (B) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes in response to Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in presence or absence of indirubin-3′-monoxime (I3M, 10 µM) or cytisine (CYT, 10 µM). (C) Determination of viable cells (%) in response to indicated concentrations of I3M or CYT treated 3T3-L1 adipocytes. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001, ** P < 0.01, * P < 0.05.

Figure 1.
In silico analysis of A2AAR-indirubin-3′-monoxime (I3M) interaction and its effect on insulin sensitivity and cellular viability.

(A) Images representing the chemical structure of indirubin-3′-monoxime (left), and its interactions with A2AAR using ligplot analysis (right). (B) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes in response to Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in presence or absence of indirubin-3′-monoxime (I3M, 10 µM) or cytisine (CYT, 10 µM). (C) Determination of viable cells (%) in response to indicated concentrations of I3M or CYT treated 3T3-L1 adipocytes. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001, ** P < 0.01, * P < 0.05.

Table 1
Docking results of two dockings methods
Ligands ADFR Autodock 
Binding energy (kcal/mol) Total interactions* Binding energy (kcal/mol) Total interactions* 
Indirubin-3-monoxime −8.0 11 −6.6 11 
Cytisine −8.1 −5.7 
Vanillin −4.7 −6.1 
Vanillylacetone −7.1 −4.8 
Chrysin −7.0 −7.8 10 
d-Mannitol −1.4 −0.9 
Ligands ADFR Autodock 
Binding energy (kcal/mol) Total interactions* Binding energy (kcal/mol) Total interactions* 
Indirubin-3-monoxime −8.0 11 −6.6 11 
Cytisine −8.1 −5.7 
Vanillin −4.7 −6.1 
Vanillylacetone −7.1 −4.8 
Chrysin −7.0 −7.8 10 
d-Mannitol −1.4 −0.9 
*

Total interactions include both the hydrogen bonding and the hydrophobic interactions.

Indirubin-3′-monoxime (I3M) prevent lipid-induced insulin resistance through the activation of A2AAR

Impairment of insulin signaling cascade by palmitate, a saturated free fatty acid, is known to be associated with the reduction in glucose transport resulting in a state of insulin unresponsiveness or insulin resistance [1,2,59]. To observe whether impairment of insulin signaling by palmitate could be prevented by I3M, or CYT, we incubated these compounds individually with 3T3-L1 adipocytes followed by the treatment with palmitate in absence or presence of insulin. Insulin-stimulated 2-NBDG uptake by 3T3-L1 adipocytes was significantly inhibited by palmitate, however, such inhibition was considerably reduced by I3M (Figure 1B). Interestingly, I3M effects are more pronounced in comparison with CYT (Figure 1B). To test the cytotoxicity of these compounds, we performed MTT assay on 3T3-L1 adipocytes. The result showed that CYT has the potent cytotoxic effect, whereas I3M does not have any toxic effect on the tested concentrations (Figure 1C). We also performed the LDH release assay that affirms nontoxic nature of I3M (Supplementary Fig. S1B). Based on these results, we selected I3M for further study. A dose-dependent stimulation of glucose uptake (Figure 2A), Akt activation (Supplementary Fig. S1C) and CREB phosphorylation (Supplementary Fig. S1D) were observed in 3T3-L1 adipocytes in response to I3M, suggesting the efficacy of I3M on insulin signaling inducement and A2AAR activation in adipocytes. Interestingly, the beneficial effect of I3M on insulin-stimulated glucose uptake was significantly prevented by A2AAR antagonist, SCH 58261 (Figure 2B). Moreover, I3M effect on insulin sensitivity was not inhibited by adenosine deaminase, thereby ruling out the possibility of extracellular adenosine accumulation in response to I3M which could have activated A2AAR (Figure 2B).

Activation of A2AAR signaling by indirubin-3′-monoxime (I3M) prevents lipid-induced insulin resistance.

Figure 2.
Activation of A2AAR signaling by indirubin-3′-monoxime (I3M) prevents lipid-induced insulin resistance.

(A) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes in response to Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in absence or presence of various concentrations of I3M (2, 5, 10 µM). (B) Effect of adenosine deaminase (ADA, 0.00138 U/ml), or SCH 58261 (SCH, 300 nM) on the I3M (10 µM) mediated attenuation of lipid-induced impairment of 2-NBDG uptake in 3T3-L1 adipocytes. (C) Western blot (upper) and its quantification (lower) showing pY-IRS1 and pAkt (S473 and T308) abundance in 3T3-L1 adipocytes in response to insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) without or with SCH 58261 (SCH, 300 nM). IRS-1 and Akt were used as loading controls. (D) Representative immunofluorescence images (upper) and its quantification (lower) showing Glut4 abundance and localization in indicated incubations of L6 cells. DAPI used for nuclear counterstaining. Scale bar, 20 mm. Corrected total cell surface fluorescence (CTCF) was calculated using the following formula with the help of image J software. CTCF = Integrated Density − (Area of selected cell × Mean fluorescence of background readings). All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001, ** P < 0.01, * P < 0.05, ns = nonsignificant.

Figure 2.
Activation of A2AAR signaling by indirubin-3′-monoxime (I3M) prevents lipid-induced insulin resistance.

(A) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes in response to Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in absence or presence of various concentrations of I3M (2, 5, 10 µM). (B) Effect of adenosine deaminase (ADA, 0.00138 U/ml), or SCH 58261 (SCH, 300 nM) on the I3M (10 µM) mediated attenuation of lipid-induced impairment of 2-NBDG uptake in 3T3-L1 adipocytes. (C) Western blot (upper) and its quantification (lower) showing pY-IRS1 and pAkt (S473 and T308) abundance in 3T3-L1 adipocytes in response to insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) without or with SCH 58261 (SCH, 300 nM). IRS-1 and Akt were used as loading controls. (D) Representative immunofluorescence images (upper) and its quantification (lower) showing Glut4 abundance and localization in indicated incubations of L6 cells. DAPI used for nuclear counterstaining. Scale bar, 20 mm. Corrected total cell surface fluorescence (CTCF) was calculated using the following formula with the help of image J software. CTCF = Integrated Density − (Area of selected cell × Mean fluorescence of background readings). All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001, ** P < 0.01, * P < 0.05, ns = nonsignificant.

Activation of the insulin signaling cascade is initiated when insulin binds to its receptors on target cells. Insulin binding activates insulin receptor (InR) tyrosine kinase which autophosphorylates InR and also phosphorylates its downstream molecule insulin receptor substrate-1 (IRS-1) on its tyrosine residues. Phosphorylation of IRS-1 recruits and activates different downstream signaling molecules including protein kinase B/Akt and that facilitates glucose transporter-4 (Glut4) migration from the cytosol to plasma membrane leading to cellular glucose uptake [1,2]. I3M incubation considerably prevents palmitate-induced impairment of IRS-1 and Akt phosphorylation, however, pretreatment with SCH 58261 notably abrogated IRS-1 and Akt phosphorylation at T308 and S473 (Figure 2C) suggesting the involvement of A2AAR in I3M mediated insulin signaling pathway stimulation. To further examine the effect of I3M on the stimulation of insulin sensitivity and participation of A2AAR, we investigated Glut-4 migration, an important marker for insulin action, in L6 myotubes. Insulin affected Glut-4 migration from the cytosol to the membrane was prevented by palmitate which was markedly mitigated by I3M incubation. However, the activity of I3M was diminished in cells when pretreated with SCH 58261 (Figure 2D). Our results suggest that I3M prevent lipid-induced insulin resistance possibly via the mediation of A2AAR activation.

I3M directly binds with and stimulates A2AAR signaling

It is well known that stimulation of A2AAR, a Gαs coupled receptor, caused activation of MAPK thereby promoting cellular proliferation, and adenylyl cyclase dependent cAMP production which confers resolution of inflammation [1722]. To explore the efficacy of I3M in A2AAR receptor activation, we observed a noticeable stimulation of ERK1/2 phosphorylation in response to I3M treatment without any significant change in p38 activation (Figure 3A). Blockage of A2AAR signaling by SCH 58261 notably prevents I3M stimulated induction of ERK1/2 phosphorylation (Figure 3A). Activation of A2AAR caused an increased production of cAMP which by activating protein kinase A (PKA) stimulates cAMP-responsive element-binding protein (CREB) phosphorylation and thus regulates its downstream genes expression [17]. As illustrated in Figure 3B, increased abundance of phosphorylated CREB in 3T3-L1 adipocytes in response to I3M incubation was notably reduced by SCH 58261. PKA is known to regulate the activation of hormone-sensitive lipase (HSL) and perilipin by phosphorylation and that induces lipolysis [60]. However, we did not observe any significant changes of adipocytes lipolysis in response to indicated concentrations of I3M (Supplementary Fig. S2). To explore the involvement of specific AR in I3M mediated effect, we incubated 3T3-L1 cells with different AR subtype specific inhibitors in presence of I3M. A significant attenuation of I3M mediated CREB phosphorylation was observed in the presence of A2AAR antagonist SCH 58261 and A1AR antagonist 8-Cyclopentyl-1,3-dipropylxanthine, whereas A2BAR antagonist Alloxazine and A3AR antagonist VUF 5574 did not produce any effect (Figure 3C). It was also interesting to observe that only A2AAR specific antagonist SCH 58261 markedly prevents I3M induced glucose uptake in 3T3-L1 adipocytes (Figure 3D). These results suggest that I3M specifically interacts with and activates A2AAR signaling.

Indirubin-3′-monoxime (I3M) stimulates A2AAR signaling.

Figure 3.
Indirubin-3′-monoxime (I3M) stimulates A2AAR signaling.

(A) Western blots (upper) and its quantification (lower) showing pERK1/2 (T202/Y204) and pp38 (T180/Y182) abundance in 3T3-L1 adipocytes incubated without or with I3M (10 µM) in absence or presence of SCH 58261 (SCH, 300 nM). ERK1/2 and p38 were used as loading controls. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05 vs Con, # P < 0.05 vs I3M. (B) Western blots (upper) and its quantification (lower) showing pCREB (S133) abundance in 3T3-L1 adipocytes incubated without or with I3M (10 µM) in absence or presence of SCH 58261 (SCH, 300 nM). β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, * P < 0.05 vs Con, # P < 0.05 vs I3M. (C) Western blots (left) and its quantification (right) showing pCREB (S133) abundance in 3T3-L1 adipocytes incubated without or with I3M (10 µM) in absence or presence of SCH 58261 (SCH, 300 nM) or Alloxazine (ALLO, 1.34 µM) or VUF 5574 (VUF, 100 nM) or 8-Cyclopentyl-1,3-dipropylxanthine (CPX, 100 nM). β-actin was used as a loading control. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, ** P < 0.01 vs Con; ## P < 0.01, # P < 0.05 vs I3M. (D) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes in response to Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) without or with SCH 58261 (SCH, 300 nM) or Alloxazine (ALLO, 1.34 µM) or VUF 5574 (VUF, 100 nM) or 8-Cyclopentyl-1,3-dipropylxanthine (CPX, 100 nM). All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, ** P < 0.01; ## P < 0.01.

Figure 3.
Indirubin-3′-monoxime (I3M) stimulates A2AAR signaling.

(A) Western blots (upper) and its quantification (lower) showing pERK1/2 (T202/Y204) and pp38 (T180/Y182) abundance in 3T3-L1 adipocytes incubated without or with I3M (10 µM) in absence or presence of SCH 58261 (SCH, 300 nM). ERK1/2 and p38 were used as loading controls. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05 vs Con, # P < 0.05 vs I3M. (B) Western blots (upper) and its quantification (lower) showing pCREB (S133) abundance in 3T3-L1 adipocytes incubated without or with I3M (10 µM) in absence or presence of SCH 58261 (SCH, 300 nM). β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, * P < 0.05 vs Con, # P < 0.05 vs I3M. (C) Western blots (left) and its quantification (right) showing pCREB (S133) abundance in 3T3-L1 adipocytes incubated without or with I3M (10 µM) in absence or presence of SCH 58261 (SCH, 300 nM) or Alloxazine (ALLO, 1.34 µM) or VUF 5574 (VUF, 100 nM) or 8-Cyclopentyl-1,3-dipropylxanthine (CPX, 100 nM). β-actin was used as a loading control. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, ** P < 0.01 vs Con; ## P < 0.01, # P < 0.05 vs I3M. (D) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes in response to Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) without or with SCH 58261 (SCH, 300 nM) or Alloxazine (ALLO, 1.34 µM) or VUF 5574 (VUF, 100 nM) or 8-Cyclopentyl-1,3-dipropylxanthine (CPX, 100 nM). All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, ** P < 0.01; ## P < 0.01.

To secure direct evidence of A2AAR involvement in I3M effect, we examined the efficacy of I3M on CREB activation, inflammatory mediators’ expressions and insulin signaling in A2AAR silenced 3T3-L1 cells (Supplementary Fig. S3A). I3M-induced CREB phosphorylation was considerably attenuated in A2AAR knockdown cells (Figure 4A). I3M incubation markedly up-regulates the anti-inflammatory (IL-10 and TGF-β) gene expressions and prevents FFA-induced pro-inflammatory mediators (MCP-1 and iNOS) gene expressions in 3T3-L1 adipocytes, however, such effects were significantly waived in A2AAR silenced cells (Figure 4B). We also observed that I3M notably attenuates FFA-induced impairment of Akt activation and insulin-stimulated glucose uptake which were significantly compromised in A2AAR silenced cells (Figure 4C,D). All these results indicate the participation of A2AAR in I3M mediated effect, however, it would be interesting to note that I3M alone was unable to activate Akt phosphorylation and glucose uptake (Figure 4C,D) that void the possibility of the direct effect of I3M on insulin signaling cascade.

Suppression of indirubin-3′-monoxime (I3M)-induced improvement of insulin sensitivity in A2AAR knockdown adipocytes.

Figure 4.
Suppression of indirubin-3′-monoxime (I3M)-induced improvement of insulin sensitivity in A2AAR knockdown adipocytes.

(A) Western blots (left) and its quantification (right) showing pCREB (S133) abundance in control siRNA (siCon) and A2AAR siRNA (siA2AAR) transfected 3T3-L1 adipocytes incubated without or with I3M (10 µM). β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, ** P < 0.01 vs control siRNA transfected cells control cells, ## P < 0.01 vs control siRNA transfected cells treated with I3M. (B) RT-PCR analysis (left) and its quantification (right) showing pro-inflammatory (MCP-1 and iNOS) and anti-inflammatory (IL-10 and TGF-β) markers gene expressions in 3T3-L1 adipocytes treated without (Con) or FFA (0.75 mM) in absence or presence of I3M (10 µM). β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05 vs control siRNA transfected cells treated with FFA, # P < 0.05 vs control siRNA transfected cells treated with FFA + I3M. (C) Western blots (left) and its quantification (right) showing pAkt (S473) abundance in control siRNA (siCon) and A2AAR siRNA (siA2AAR) transfected 3T3-L1 adipocytes incubated without (Con) or with I3M (10 µM) or Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) or Ins + FFA + I3M. β-actin was used as a loading control. (D) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes transfected with control siRNA (siCon) and A2AAR siRNA (siA2AAR) in response to I3M (10 µM) or Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) or Ins + FFA + I3M. Each value is the mean ± SEM of three independent experiments, * P < 0.05 vs control siRNA transfected cells treated with Ins + FFA, # P < 0.05 vs control siRNA transfected cells treated with Ins + FFA + I3M.

Figure 4.
Suppression of indirubin-3′-monoxime (I3M)-induced improvement of insulin sensitivity in A2AAR knockdown adipocytes.

(A) Western blots (left) and its quantification (right) showing pCREB (S133) abundance in control siRNA (siCon) and A2AAR siRNA (siA2AAR) transfected 3T3-L1 adipocytes incubated without or with I3M (10 µM). β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, ** P < 0.01 vs control siRNA transfected cells control cells, ## P < 0.01 vs control siRNA transfected cells treated with I3M. (B) RT-PCR analysis (left) and its quantification (right) showing pro-inflammatory (MCP-1 and iNOS) and anti-inflammatory (IL-10 and TGF-β) markers gene expressions in 3T3-L1 adipocytes treated without (Con) or FFA (0.75 mM) in absence or presence of I3M (10 µM). β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05 vs control siRNA transfected cells treated with FFA, # P < 0.05 vs control siRNA transfected cells treated with FFA + I3M. (C) Western blots (left) and its quantification (right) showing pAkt (S473) abundance in control siRNA (siCon) and A2AAR siRNA (siA2AAR) transfected 3T3-L1 adipocytes incubated without (Con) or with I3M (10 µM) or Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) or Ins + FFA + I3M. β-actin was used as a loading control. (D) Analysis of 2-NBDG uptake by 3T3-L1 adipocytes transfected with control siRNA (siCon) and A2AAR siRNA (siA2AAR) in response to I3M (10 µM) or Insulin (Ins, 100 nM) or Ins + FFA (palmitate, 0.75 mM) or Ins + FFA + I3M. Each value is the mean ± SEM of three independent experiments, * P < 0.05 vs control siRNA transfected cells treated with Ins + FFA, # P < 0.05 vs control siRNA transfected cells treated with Ins + FFA + I3M.

To confirm the A2AAR ligand nature of I3M, a radioligand binding assay was performed. The radioligand competition curve of I3M competing for [³H]NECA (10 nM) to bind with A2AAR showed the Ki value of 0.83 µM where nonspecific binding amounts ∼10% of total binding (Supplementary Fig. S3B). Radioligand binding assay displayed a potent high-affinity binding of I3M with A2AAR (Ki: 0.52 µM) (Table 2). To determine the concentration of I3M that gives a half-maximal response (EC50), we investigated cAMP activity at varying concentrations of I3M in the CHO cells stably overexpressing A2AAR (Supplementary Fig. S3C,D) and found the EC50 value of I3M is 0.12 µM (Figure 5A). To substantiate the A2AAR agonistic nature of I3M, A2AAR stably expressed CHO cells were incubated with I3M followed by the treatment of increasing concentrations of SCH 58261. Stimulation of CREB phosphorylation in response to I3M was significantly attenuated with increasing concentrations of SCH 58261 (Figure 5B). These results support the notion that I3M is a potent agonist of A2AAR. Furthermore, to investigate whether a selective A2AAR agonist could able to mimic the effects of I3M, we observed that treatment with A2AAR selective agonist CGS 21680 noticeably up-regulated CREB phosphorylation (Supplementary Fig. S4A) and insulin-stimulated glucose uptake (Supplementary Fig. S4B) in 3T3-L1 adipocytes.

Indirubin-3′-monoxime (I3M) directly binds with and activates A2AAR signaling.

Figure 5.
Indirubin-3′-monoxime (I3M) directly binds with and activates A2AAR signaling.

(A) A2AAR stably expressed CHO cells were incubated with the indicated concentrations of I3M followed by the determination of EC50 value of I3M by measuring percentage activity of cAMP assay. % activity = 100 × [(A − Abasal)/(Amax − Abasal)], where A is the sample absorbance, Amax is the absorbance at maximum stimulation, and Abasal is the absorbance at basal level (without I3M). (B) Western blots (left) and its quantification (right) showing pCREB (S133) abundance in 3T3-L1 adipocytes incubated with I3M (10 µM) for 1 h followed by 4 h of incubation of indicated concentrations of SCH 58261. β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05 vs Con; ## P < 0.01, # P < 0.05 vs I3M.

Figure 5.
Indirubin-3′-monoxime (I3M) directly binds with and activates A2AAR signaling.

(A) A2AAR stably expressed CHO cells were incubated with the indicated concentrations of I3M followed by the determination of EC50 value of I3M by measuring percentage activity of cAMP assay. % activity = 100 × [(A − Abasal)/(Amax − Abasal)], where A is the sample absorbance, Amax is the absorbance at maximum stimulation, and Abasal is the absorbance at basal level (without I3M). (B) Western blots (left) and its quantification (right) showing pCREB (S133) abundance in 3T3-L1 adipocytes incubated with I3M (10 µM) for 1 h followed by 4 h of incubation of indicated concentrations of SCH 58261. β-actin was used as a loading control. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05 vs Con; ## P < 0.01, # P < 0.05 vs I3M.

Table 2
Competition of putative agonist (I3M) or agonist (CGS 21680) for [3H]NECA binding at A2AAR and the effect of I3M or CGS 21680 on adenylyl cyclase activity/cAMP level via A2BAR
Compounds A2AAR A2BAR 
Ki (µM) 95% confidence limits EC50/IC50 (µM) 95% confidence limits 
Indirubin-3′-monoxime 0.52 0.28 0.966 >100 
CGS 21680 0.0276 0.0186 0.0409 >100 
Compounds A2AAR A2BAR 
Ki (µM) 95% confidence limits EC50/IC50 (µM) 95% confidence limits 
Indirubin-3′-monoxime 0.52 0.28 0.966 >100 
CGS 21680 0.0276 0.0186 0.0409 >100 

I3M promotes the anti-inflammatory state in adipocytes through the activation of A2AAR signaling

We next explored the induction of A2AAR signaling by I3M on the stimulation of anti-inflammatory state of adipocytes in the in vitro model of lipid-induced insulin resistance. I3M incubation markedly prevents palmitate-induced suppression of CREB phosphorylation in adipocytes which was strikingly attenuated by SCH 58261 (Figure 6A). Investigating the I3M-mediated induction of A2AAR signaling in the transactivation potential of CREB, we found that incubation of palmitate significantly attenuates IL-10 luciferase reporter activity in 3T3-L1 adipocytes which was considerably rescued by I3M. However, the I3M effect was diminished when cells were pretreated with SCH 58261 (Figure 6B). We then explored the efficacy of I3M-mediated A2AAR signaling on gene expression of various anti-inflammatory cytokines. I3M distinctly up-regulated IL-10, IL-13, IL-4 and TGF-β gene expression, which were considerably attenuated by SCH 58261 incubation (Figure 6C). Moreover, I3M-induced IL-10 protein secretion from 3T3-L1 adipocytes was significantly inhibited by SCH 58261 (Supplementary Fig. S5). To examine the direct effect of I3M, we observed dose-dependent CREB activation and anti-inflammatory (IL-10 and TGF-β) gene expressions in 3T3-L1 adipocytes in response to I3M incubations (Figure 6D,E). Furthermore, to clarify the role of CREB as a mediator of I3M anti-inflammatory actions, we suppressed CREB activity in adipocytes by transfecting A-CREB, a dominant-negative mutant of CREB [61], and investigated its role on anti-inflammatory gene expressions. Interestingly, I3M effected induction of IL-10 and TGF-β gene expressions were markedly attenuated in A-CREB transfected adipocytes (Figure 7A). It is now well established that adipocytes inflammatory state play a critical role in regulating systemic insulin sensitivity [711]. We, therefore, envisioned that I3M mediated alteration of inflammatory factors expressions in adipocytes could able to prevent the FFA-induced insulin resistance. To address this issue, we incubated L6 myotubes with the conditional media isolated from I3M incubated adipocytes and investigated the insulin-stimulated muscle glucose uptake. Intriguingly, I3M incubated adipocytes conditional media notably mitigate the FFA-induced impairment of insulin-stimulated glucose uptake in L6 myotubes (Figure 7B). Collectively, all these results suggest that I3M strongly promotes an anti-inflammatory state in adipocytes through the activation of the A2AAR-cAMP-CREB pathway which could be sufficient to improve FFA-induced insulin resistance.

Indirubin-3′-monoxime (I3M) promotes adipocytes anti-inflammatory state by stimulating A2AAR signaling cascade.

Figure 6.
Indirubin-3′-monoxime (I3M) promotes adipocytes anti-inflammatory state by stimulating A2AAR signaling cascade.

(A) Western blots (left) and its quantification (right) showing abundance of pCREB (S133) level in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) without or with SCH 58261 (SCH, 300 nM). β-actin was used as loading control. (B) 3T3-L1 adipocytes transfected with IL-10 promoter-luciferase plasmid (pGL2B −1538/+64) were incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). On termination of incubations, cells were lysed and luciferase activity was measured by the multimode reader. (C) Real-time quantitative PCR analysis showing fold change of IL-10, IL-13, IL-4 and TGF-β mRNA level in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). GAPDH served as an internal control for normalization. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001, ** P < 0.01, * P < 0.05. (D and E) Western blots (left) and its quantifications (right) showing pCREB (S133) level (D) and RT-PCR analysis (left) and its quantification (right) showing anti-inflammatory cytokines (IL-10 and TGF-β) gene expressions (E) in 3T3-L1 adipocytes in response to different concentrations of I3M. β-actin was serves as loading controls. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001 Vs Con, ### P < 0.001 Vs Con.

Figure 6.
Indirubin-3′-monoxime (I3M) promotes adipocytes anti-inflammatory state by stimulating A2AAR signaling cascade.

(A) Western blots (left) and its quantification (right) showing abundance of pCREB (S133) level in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) without or with SCH 58261 (SCH, 300 nM). β-actin was used as loading control. (B) 3T3-L1 adipocytes transfected with IL-10 promoter-luciferase plasmid (pGL2B −1538/+64) were incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). On termination of incubations, cells were lysed and luciferase activity was measured by the multimode reader. (C) Real-time quantitative PCR analysis showing fold change of IL-10, IL-13, IL-4 and TGF-β mRNA level in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). GAPDH served as an internal control for normalization. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001, ** P < 0.01, * P < 0.05. (D and E) Western blots (left) and its quantifications (right) showing pCREB (S133) level (D) and RT-PCR analysis (left) and its quantification (right) showing anti-inflammatory cytokines (IL-10 and TGF-β) gene expressions (E) in 3T3-L1 adipocytes in response to different concentrations of I3M. β-actin was serves as loading controls. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001 Vs Con, ### P < 0.001 Vs Con.

Indirubin-3′-monoxime (I3M)-induced CREB activation promotes anti-inflammatory gene expression and improves insulin sensitivity in 3T3-L1 adipocytes.

Figure 7.
Indirubin-3′-monoxime (I3M)-induced CREB activation promotes anti-inflammatory gene expression and improves insulin sensitivity in 3T3-L1 adipocytes.

(A) 3T3-L1 adipocytes transfected with scrambled plasmid or CMV500 A-CREB (dominant-negative inhibitor of CREB) followed by the incubations without (Con) or with I3M (10 µM). On termination of incubations, cells were subjected for RT-PCR analysis of IL-10 and TGF-β gene expressions. β-actin was used as a loading control. * P < 0.05 vs scrambled transfected control cells, # P < 0.05 vs scrambled transfected I3M treated cells. (B) Effect of conditional media, isolated from the control or I3M (10 µM) treated 3T3-L1 adipocytes, on 2-NBDG uptake in FFA-induced insulin-resistant L6 myotubes. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05.

Figure 7.
Indirubin-3′-monoxime (I3M)-induced CREB activation promotes anti-inflammatory gene expression and improves insulin sensitivity in 3T3-L1 adipocytes.

(A) 3T3-L1 adipocytes transfected with scrambled plasmid or CMV500 A-CREB (dominant-negative inhibitor of CREB) followed by the incubations without (Con) or with I3M (10 µM). On termination of incubations, cells were subjected for RT-PCR analysis of IL-10 and TGF-β gene expressions. β-actin was used as a loading control. * P < 0.05 vs scrambled transfected control cells, # P < 0.05 vs scrambled transfected I3M treated cells. (B) Effect of conditional media, isolated from the control or I3M (10 µM) treated 3T3-L1 adipocytes, on 2-NBDG uptake in FFA-induced insulin-resistant L6 myotubes. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, ** P < 0.01, * P < 0.05.

Activation of A2AAR by I3M attenuates lipid-induced adipocyte inflammation

We and several other investigators have shown that FFA-induced activation of nuclear factor-κB (NF-κB) stimulates chronic low-grade inflammation in adipose tissue causing obesity-induced insulin resistance [29]. Interestingly, palmitate-induced NF-κBp65 and IκBα phosphorylation in 3T3-L1 adipocytes was substantially impeded by I3M. However, the beneficial effect of I3M was compromised when cells were treated with SCH 58261 (Figure 8A). Examining the direct effect of I3M on the inflammatory markers in 3T3-L1 adipocytes, we did not notice any significant change of NF-κB activation and its target genes (MCP-1 and iNOS) expression (Supplementary Fig. S6A,B) in response to I3M incubation. To have a better insight of I3M effect and the contribution of A2AAR in attenuating palmitate stimulation of NF-κB transactivation potential, we investigated κB luciferase reporter activity in 3T3-L1 adipocytes. Palmitate-enhanced κB luciferase reporter activity was significantly reduced in I3M treated cells, however, I3M effect was reversed when cells were pretreated with SCH 58261 (Figure 8B) indicating that I3M down-regulates NF-κB transactivation competence by activation of A2AAR. Since FFA-induced activation of PKCs, particularly PKCθ and PKCε, are known to be involved in the negative action on insulin signaling [62], we examined the effect of I3M on PKC activation and observed that FFA-induced stimulation of PKC phosphorylation was strikingly prevented by I3M, however, such effect was considerably hindered by SCH 58261 (Figure 8C). We then examined the expression of NF-κB-regulated genes and found I3M incubation remarkably prevents palmitate-induced up-regulation of MCP-1, IL-6, IL-1β and TNF-α cytokines gene expression which was significantly inhibited by SCH 58261 (Figure 8D). Moreover, significant inhibition of FFA-induced IL-6 protein secretion was observed when 3T3-L1 adipocytes were pretreated with I3M, which was markedly abrogated by SCH 58261 (Supplementary Fig. S7). All these results suggest I3M via A2AAR activation alleviate lipid-induced adipocyte inflammation.

Activation of A2AAR by indirubin-3′-monoxime (I3M) attenuates lipid-induced adipocyte inflammation.

Figure 8.
Activation of A2AAR by indirubin-3′-monoxime (I3M) attenuates lipid-induced adipocyte inflammation.

(A) Western blots (left) and its quantification (right) showing pNF-κB (S281) and pIκB-α (S32) abundance in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). β-actin was used as loading control. (B) 3T3-L1 adipocytes transfected with κB luciferase plasmid were incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). On termination of incubations, cells were lysed and luciferase activity was measured by the multimode reader. (C) Western blot (left) and its quantification (right) showing abundance of pan-phospho-PKC (βII Ser660) level in 3T3-L1 adipocytes incubated without (Con) or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 mM) without or with SCH 58261 (SCH, 300 nM). β-actin was used as loading control. (D) Real-time quantitative PCR analysis showing fold change of MCP-1, TNF-α, IL-6 and IL-1β mRNA level in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (300 nM). GAPDH served as an internal control for normalization. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001; ** P < 0.01; * P < 0.05.

Figure 8.
Activation of A2AAR by indirubin-3′-monoxime (I3M) attenuates lipid-induced adipocyte inflammation.

(A) Western blots (left) and its quantification (right) showing pNF-κB (S281) and pIκB-α (S32) abundance in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). β-actin was used as loading control. (B) 3T3-L1 adipocytes transfected with κB luciferase plasmid were incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (SCH, 300 nM). On termination of incubations, cells were lysed and luciferase activity was measured by the multimode reader. (C) Western blot (left) and its quantification (right) showing abundance of pan-phospho-PKC (βII Ser660) level in 3T3-L1 adipocytes incubated without (Con) or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 mM) without or with SCH 58261 (SCH, 300 nM). β-actin was used as loading control. (D) Real-time quantitative PCR analysis showing fold change of MCP-1, TNF-α, IL-6 and IL-1β mRNA level in 3T3-L1 adipocytes incubated without or with FFA (palmitate, 0.75 mM) in absence or presence of I3M (10 µM) or I3M (10 µM) + SCH 58261 (300 nM). GAPDH served as an internal control for normalization. All experiments were performed in triplicate. Each value is the mean ± SEM of three independent experiments, *** P < 0.001; ** P < 0.01; * P < 0.05.

Discussion

After almost a century of intense research, ARs have now become the center of attention and are considered a potential therapeutic target for cancer, diabetes, cardiovascular disease and immune, inflammatory and neurodegenerative disorders. In normal physiological conditions, the extracellular level of the purine nucleoside adenosine is rather low which can rapidly increase in various pathological conditions such as metabolic stress, tissue injury and inflammation [1522]. Extracellular adenosine mediates a range of responses to restore tissue homeostasis through ligation of G-protein coupled cell-surface ARs. These receptors are widely distributed in metabolically active organs such as the liver, pancreas and adipose tissue as well as in the immune system and categorized into four subtypes that include A1, A2A, A2B and A3 [1517]. Several studies have indicated that AR, particularly A2AAR signaling stimulates anti-inflammatory effect in adipose tissue and its resident immune cells that promote cellular insulin responsiveness [1522]. A recent study showed that activation of A2AAR increased brown adipocytes and ‘browning’ of white adipocytes which augmented thermogenesis and inhibited insulin resistance [63]. Experiments on mouse islets indicate that, besides insulin responsiveness, adenosine also stimulates insulin secretion through activation of A2AAR receptors which can be reversed by A2AAR antagonist SCH 58261 [64]. Several studies have documented the association of adipose tissue inflammation for the promotion of insulin resistance and type 2 diabetes (T2D) [29]. Despite considerable progress in research and development of various anti-diabetic agents, currently, no anti-diabetic drugs have been proven clinically effective for T2D therapy. All these studies attest A2AAR could be an imperative target for the development of therapeutics against T2D which raises the demand for selective A2AAR agonist.

The main approach to develop novel A2AAR agonists lies on the modification of adenosine itself. Many attempts have been made to develop various adenosine derivatives with most useful modifications in position 2 or N6 of the adenine ring and 3′, 4′ and 5′-position of the ribose that provides better binding affinity and activity in comparison with adenosine as selective agonists for four different ARs [34]. The recent breakthrough in crystallography of the human A2AAR bound to its agonist adenosine revealed that the major interactive regions of the adenine scaffold are the hydrogen bonding with both Glu169 and Asn253, and the π-stacking and hydrophobic interactions with Phe168 and Ile274 [65]. The ribose moiety of adenosine plays a critical role in receptor activation, by forming the hydrogen bonds with Ser277 and His278 of A2AAR along with conformational changes of receptor due to the positional shift of Val84 and Trp246 [58,65]. However, non-adenosine compounds with different chemotypes exhibited selective and potent agonistic activity against ARs [66,67]. All these studies open an opportunity to discover new ligand chemotypes for selective AR agonist.

In our attempt to find potential A2AAR agonists, we selected 142 different molecules based on two different aspects; (i) it should not have structural similarity with adenosine and therefore be devoid of a ribose moiety and (ii) should have an anti-diabetic effect published in the literature. Screening of these compounds with the active conformation of A2AAR through in silico approaches yielded six molecules with lower toxic values (through 2D-QSAR), satisfying the requisite druglike, Veber's and the ADMET properties. Two different docking approaches have been applied to analyze potential binding affinity of these molecules to the binding pocket of the A2AAR. Applying flexibility to the active site residues has strengthened the results. Ligplot analysis of the docked compounds has depicted the presence of hydrogen bonds and hydrophobic interactions with the active site residues. It is known that two important residues Phe168 and Glu169 of A2AAR play a critical role in ligand binding and His278 residue is critical for receptor activation [58]. This is corroborated from our results as indirubin-3′monoxime (I3M) form H-bonds with these important residues. It has been observed that a novel ligand chemotype I3M has better interaction pattern with the A2AAR.

Investigating the efficacy of these molecules for preventing insulin resistance, we utilized a well established in vitro model of insulin resistance by incubating 3T3-L1 adipocytes with a saturated free fatty acid palmitate. Palmitate incubation significantly impairs insulin signaling represented by attenuated insulin-stimulated glucose uptake, compromised insulin signaling pathway molecules activation and reduced Glut4 migration from the cytosol to the plasma membrane. Based on the efficiency of I3M, and CYT in reversing palmitate-induced impairment of insulin-stimulated glucose uptake and their cytotoxicity effect; we selected I3M for further investigation. I3M effectively up-regulates the activation of ERK1/2 and CREB, however, such effect was compromised in A2AAR knockdown cells indicating the possible involvement of A2AAR signaling in mediating I3M effect. The radioligand binding assay confirms direct high-affinity binding of I3M to A2AAR. Investigating the cAMP level in response to increasing concentrations of I3M and the effect of increasing concentrations of A2AAR antagonist on I3M mediated induction of CREB activation in the A2AAR stable clone of CHO cells revealed the A2AAR agonist nature of I3M. Utilizing various antagonists of AR subtypes, we have shown the selective involvement of A2AAR in mediating I3M effect.

Using A2AAR deficient mice combined with a pharmacological approach, it has been shown that adenosine through A2AAR activation inhibits LPS induced macrophage inflammation by suppressing TNF-α, IL-6 and IL-12 expression and augmenting IL-10 production [68]. In this regard, it is of note that our results demonstrated that I3M strongly attenuates palmitate-induced NF-κB activation, κB promoter-reporter induction, and gene expression of various pro-inflammatory cytokines, however, such effects were compromised by pharmacological inhibition of A2AAR. Furthermore, I3M induced A2AAR-cAMP-CREB signaling also leads to the up-regulation of IL-10 and other anti-inflammatory cytokines gene expression which were attenuated in the presence of A2AAR selective antagonist. All these results support the agonistic nature of I3M adding a novel scaffold of selective non-adenosine based A2AAR agonists.

Taken together, our results demonstrate that I3M could be a selective novel agonist of A2AAR which efficiently attenuates lipid-induced adipocyte inflammation and insulin resistance. Thus, I3M has therapeutic potential for the prevention and/or treatment of type 2 diabetes.

Abbreviations

     
  • [³H]NECA

    [tritium] 5′-N-ethyl-carboxamide-adenosine

  •  
  • 2-NBDG

    2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose

  •  
  • ADMET

    absorption, distribution, metabolism, excretion and toxicity

  •  
  • ANOVA

    analysis of variance

  •  
  • AR

    adenosine receptors

  •  
  • BSA

    bovine serum albumin

  •  
  • cAMP

    cyclic adenosine monophosphate

  •  
  • cDNA

    complementary DNA

  •  
  • CHO

    Chinese hamster ovary

  •  
  • CMV3

    cytomegalovirus 3

  •  
  • CREB

    cAMP responsive element binding protein

  •  
  • CTCF

    corrected total cell surface fluorescence

  •  
  • CYT

    cytisine

  •  
  • DAPI

    4′,6-diamidino-2-phenylindole

  •  
  • DMEM

    Dulbecco's modified Eagle's medium

  •  
  • DNase I

    deoxyribonuclease I

  •  
  • DPBS

    Dulbecco's phosphate-buffered saline

  •  
  • EC50

    effective concentration 50

  •  
  • ECL

    enhanced chemiluminescence

  •  
  • EDTA

    ethylenediaminetetraacetic acid

  •  
  • ERK1/2

    extracellular signal-regulated kinases ½

  •  
  • FBS

    fetal bovine serum

  •  
  • FFA

    free fatty acid

  •  
  • GAPDH

    glyceraldehyde-3-phosphate dehydrogenase

  •  
  • Glut4

    glucose transporter 4

  •  
  • HRP

    horseradish peroxidase

  •  
  • I3M

    indirubin-3′-monoxime

  •  
  • IBMX

    3-isobutyl-1-methylxanthine

  •  
  • iNOS

    inducible nitric oxide synthase

  •  
  • InR

    insulin receptor

  •  
  • IRS-1

    insulin receptor substrate-1

  •  
  • IκBα

    inhibitor of κB alpha

  •  
  • JAK

    Janus kinase

  •  
  • KRP

    Kreb's Ringer phosphate

  •  
  • LD50

    lethal dose 50

  •  
  • LDH

    lactate dehydrogenase

  •  
  • MAPK

    mitogen-activated protein kinase

  •  
  • MCP-1

    monocyte chemoattractant protein-1

  •  
  • MFI

    mean fluorescence intensity

  •  
  • MTT

    3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium

  •  
  • NF-κB

    nuclear factor-kappa B

  •  
  • NP40

    Nonidet P-40

  •  
  • ORF

    open reading frame

  •  
  • PAGE

    polyacrylamide gel electrophoresis

  •  
  • PDB

    protein data bank

  •  
  • PKA

    protein kinase A

  •  
  • PKC

    protein kinase C

  •  
  • PVDF

    polyvinylidene difluoride

  •  
  • QSAR

    quantitative structure-activity relationship

  •  
  • RT-qPCR

    real-time reverse transcriptase-polymerase chain reaction

  •  
  • SDS

    sodium dodecyl sulfate

  •  
  • SEM

    standard error of the mean

  •  
  • siRNA

    small interfering RNA

  •  
  • T2D

    type 2 diabetes

  •  
  • TBST

    tris-buffered saline with 0.1% Tween-20

  •  
  • TGF-β

    transforming growth factor beta

  •  
  • TMB

    3,3′,5,5′-tetramethylbenzidin

Author Contribution

S.A.C. performed the experimental studies; analyzed the data and wrote the manuscript. N.B. performed the in silico studies, analyzed the data and wrote the manuscript. D.B. and L.A. performed some experiments and analyzed the data. A.S.D. helped in some experiments. R.Y. and K.-N. K. performed the radio-ligand binding assay, analyzed and interpreted the data. D.P. designed and supervised some experiments, analyzed data and wrote the manuscript. S.D.G. and A.N.J. conceived and designed the experiments, analyzed data, wrote the manuscript and supervised this study. All authors approved the final version.

Funding

This work was supported in part by the ‘Unit of Excellence' grant from the Department of Biotechnology (DBT), New Delhi [Grant number: BT/554/NE/U-Excel/2014]' and the Startup grant for Young Scientists from the Science and Engineering Research Board (SERB), New Delhi [Grant number: YSS/2014/000067] to S.D.G.

Acknowledgements

S.A.C., N.B., and D.B. acknowledge DBT for research fellowship. We thank Dr. R. Mukhopadhyay, Department of MBBT, Tezpur University, for kind help in extending his laboratory facilities and the National Center for Cell Science, Pune, India for providing L6 cell line. We deeply appreciate the help of Dr. Gaurangi Maitra, Tezpur University, for correcting grammatical errors. The expert technical assistance of Sonja Kachler is also greatly acknowledged. We are also thankful to the Head, Department of MBBT, Tezpur University, and the Head, CBME, Indian Institute of Technology Ropar for extending the facilities required for the present investigation. Financial support in the form of UGC-SAP-DRS-II, DST-FIST-I, DBT-Strengthening, DBT-BIF and DBT-Biotech Hub to the Department of MBBT, Tezpur University is acknowledged.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Goldstein
,
B.J.
(
2002
)
Insulin resistance as the core defect in type 2 diabetes mellitus
.
Am. J. Cardiol.
90
,
3G
10G
2
Czech
,
M.P.
(
2017
)
Insulin action and resistance in obesity and type 2 diabetes
.
Nat. Med.
23
,
804
814
3
Shoelson
,
S.E.
,
Lee
,
J.
and
Goldfine
,
A.B.
(
2006
)
Inflammation and insulin resistance
.
J. Clin. Invest.
116
,
1793
1801
4
Glass
,
C.
and
Olefsky
,
J.M.
(
2012
)
Inflammation and lipid signaling in the etiology of insulin resistance
.
Cell Metab.
15
,
635
642
5
Gregor
,
M.F.
and
Hotamisligil
,
G.S.
(
2011
)
Inflammatory mechanisms in obesity
.
Annu. Rev. Immunol.
29
,
415
445
6
Lackey
,
D.E.
and
Olefsky
,
J.M.
(
2016
)
Regulation of metabolism by the innate immune system
.
Nat. Rev. Endocrinol.
12
,
15
28
7
Reilly
,
S.M.
and
Saltiel
,
A.R.
(
2017
)
Adapting to obesity with adipose tissue inflammation
.
Nat. Rev. Endocrinol.
13
,
633
643
8
Shi
,
H.
,
Kokoeva
,
M.V.
,
Inouye
,
K.
,
Tzameli
,
I.
,
Yin
,
H.
and
Flier
,
J.S.
(
2006
)
TLR4 links innate immunity and fatty acid-induced insulin resistance
.
J. Clin. Invest.
116
,
3015
3025
9
Pal
,
D.
,
Dasgupta
,
S.
,
Kundu
,
R.
,
Maitra
,
S.
,
Das
,
G.
,
Mukhopadhyay
,
S.
et al.  (
2012
)
Fetuin-A acts as an endogenous ligand of TLR4 to promote lipid-induced insulin resistance
.
Nat. Med.
18
,
1279
1285
10
Despres
,
J.P.
and
Lemieux
,
I.
(
2006
)
Abdominal obesity and metabolic syndrome
.
Nature
444
,
881
887
11
Guilherme
,
A.
,
Virbasius
,
J.V.
,
Puri
,
V.
and
Czech
,
M.P.
(
2008
)
Adipocyte dysfunctions linking obesity to insulin resistance and type 2 diabetes
.
Nat. Rev. Mol. Cell Biol.
9
,
367
377
12
Yuan
,
M.
,
Konstantopoulos
,
N.
,
Lee
,
J.
,
Hansen
,
L.
,
Li
,
Z.W.
,
Karin
,
M.
et al.  (
2001
)
Reversal of obesity- and diet-induced insulin resistance with salicylates or targeted disruption of Ikkβ
.
Science
293
,
1673
1677
13
Spite
,
M.
,
Hellmann
,
J.
,
Tang
,
Y.
,
Mathis
,
S.P.
,
Kosuri
,
M.
,
Bhatnagar
,
A.
et al.  (
2011
)
Deficiency of the leukotriene B4 receptor, BLT-1, protects against systemic insulin resistance in diet-induced obesity
.
J. Immunol.
187
,
1942
1949
14
Donath
,
M.Y.
(
2014
)
Targeting inflammation in the treatment of type 2 diabetes: time to start
.
Nat. Rev. Drug Discov.
13
,
465
476
15
Koupenova
,
M.
and
Ravid
,
K.
(
2013
)
Adenosine, adenosine receptors and their role in glucose homeostasis and lipid metabolism
.
J. Cell. Physiol.
228
,
1703
1712
16
Antonioli
,
L.
,
Blandizzi
,
C.
,
Csóka
,
B.
,
Pacher
,
P.
and
Haskó
,
G.
(
2015
)
Adenosine signaling in diabetes mellitus - pathophysiology and therapeutic considerations
.
Nat. Rev. Endocrinol.
11
,
228
241
17
Haskó
,
G.
and
Cronstein
,
B.N.
(
2004
)
Adenosine: an endogenous regulator of innate immunity
.
Trends Immunol.
25
,
33
39
18
Csóka
,
B.
,
Törő
,
G.
,
Vindeirinho
,
J.
,
Varga
,
Z.V.
,
Koscsó
,
B.
,
Németh
,
Z.H.
et al.  (
2017
)
A2a adenosine receptors control pancreatic dysfunction in high-fat-diet induced obesity
.
FASEB J.
31
,
4985
4997
19
Csóka
,
B.
,
Selmeczy
,
Z.
,
Koscsó
,
B.
,
Németh
,
Z.H.
,
Pacher
,
P.
,
Murray
,
P.J.
et al.  (
2012
)
Adenosine promotes alternative macrophage activation via A2A and A2B receptors
.
FASEB J.
26
,
376
386
20
Haskó
,
G.
,
Szabó
,
C.
,
Németh
,
Z.H.
,
Kvetan
,
V.
,
Pastores
,
S.M.
and
Vizi
,
E.S.
(
1996
)
Adenosine receptor agonists differentially regulate IL-10, TNF-α, and nitric oxide production in RAW 264.7 macrophages and in endotoxemic mice
.
J. Immunol.
157
,
4634
4640
PMID:
[PubMed]
21
Haskó
,
G.
,
Kuhel
,
D.G.
,
Chen
,
J.F.
,
Schwarzschild
,
M.A.
,
Deitch
,
E.A.
,
Mabley
,
J.G.
et al.  (
2000
)
Adenosine inhibits IL-12 and TNF-α production via adenosine A2a receptor-dependent and independent mechanisms
.
FASEB J.
14
,
2065
2074
22
Csóka
,
B.
,
Koscsó
,
B.
,
Töro
,
G.
,
Kókai
,
E.
,
Virág
,
L.
,
Németh
,
Z.H.
et al.  (
2014
)
A2b adenosine receptors prevent insulin resistance by inhibiting adipose tissue inflammation via maintaining alternative macrophage activation
.
Diabetes
63
,
850
866
23
Newman
,
D.J.
and
Cragg
,
G.M.
(
2012
)
Natural products as sources of new drugs over the 30 years from 1981 to 2010
.
J. Nat. Prod.
75
,
311
335
24
Kinghorn
,
A.D.
,
Pan
,
L.
,
Fletcher
,
J.N.
and
Chai
,
H.
(
2011
)
The relevance of higher plants in lead compound discovery programs
.
J. Nat. Prod.
74
,
1539
1555
25
Eisenbrand
,
G.
,
Hippe
,
F.
,
Jakobs
,
S.
and
Muehlbeyer
,
S.
(
2004
)
Molecular mechanisms of indirubin and its derivates: novel anticancer molecules with their origin in traditional Chinese phytomedicine
.
J. Cancer Res. Clin. Oncol.
130
,
627
635
26
Kim
,
S.A.
,
Kim
,
Y.C.
,
Kim
,
S.W.
,
Lee
,
S.H.
,
Min
,
J.J.
,
Ahn
,
S.G.
et al.  (
2007
)
Antitumor activity of novel indirubin derivatives in rat tumor model
.
Clin. Cancer Res.
13
,
253
259
27
Zhang
,
A.
,
Ning
,
B.
,
Sun
,
N.
,
Wei
,
J.
and
Ju
,
X.
(
2015
)
Indirubin increases CD4+ CD25+ Foxp3+ regulatory T cells to prevent immune thrombocytopenia in mice
.
PLoS ONE
10
,
e0142634
28
Xie
,
X.J.
,
Di
,
T.T.
,
Wang
,
Y.
,
Wang
,
M.X.
,
Meng
,
Y.J.
,
Lin
,
Y.
et al.  (
2018
)
Indirubin ameliorates imiquimod-induced psoriasis-like skin lesions in mice by inhibiting inflammatory responses mediated by IL-17A-producing γδ T cells
.
Mol. Immunol.
101
,
386
395
29
Hoessel
,
R.
,
Leclerc
,
S.
,
Endicott
,
J.A.
,
Nobel
,
M.E.
,
Lawrie
,
A.
,
Tunnah
,
P.
et al.  (
1999
)
Indirubin, the active constituent of a Chinese antileukaemia medicine, inhibits cyclin-dependent kinases
.
Nat. Cell Biol.
1
,
60
67
30
Pergola
,
C.
,
Gaboriaud-Kolar
,
N.
,
Jestädt
,
N.
,
König
,
S.
,
Kritsanida
,
M.
,
Schaible
,
A.M.
et al.  (
2014
)
Indirubin core structure of glycogen synthase kinase-3 inhibitors as novel chemotype for intervention with 5-lipoxygenase
.
J. Med. Chem.
57
,
3715
3723
31
Nam
,
S.
,
Wen
,
W.
,
Schroeder
,
A.
,
Herrmann
,
A.
,
Yu
,
H.
,
Cheng
,
X.
et al.  (
2013
)
Dual inhibition of Janus and Src family kinases by novel indirubin derivative blocks constitutively-activated Stat3 signaling associated with apoptosis of human pancreatic cancer cells
.
Mol. Oncol.
7
,
369
378
32
Lai
,
J.L.
,
Liu
,
Y.H.
,
Liu
,
C.
,
Qi
,
M.P.
,
Liu
,
R.N.
,
Zhu
,
X.F.
et al.  (
2017
)
Indirubin inhibits LPS-induced inflammation via TLR4 abrogation mediated by the NFkB and MAPK signaling pathways
.
Inflammation
40
,
1
12
33
de Lera Ruiz
,
M.
,
Lim
,
Y.H.
and
Zheng
,
J.
(
2014
)
Adenosine A2A receptor as a drug discovery target
.
J. Med. Chem.
57
,
3623
3650
34
Jacobson
,
K.A.
and
Gao
,
Z.G.
(
2006
)
Adenosine receptors as therapeutic targets
.
Nat. Rev. Drug Discov.
5
,
247
264
35
Carpenter
,
B.
,
Nehmé
,
R.
,
Warne
,
T.
,
Leslie
,
A.G.
and
Tate
,
C.G.
(
2016
)
Structure of the adenosine A(2A) receptor bound to an engineered G protein
.
Nature
536
,
104
107
36
Webb
,
B.
and
Sali
,
A.
(
2016
)
Comparative protein structure modeling using MODELLER
.
Curr. Protoc. Bioinformatics
54
,
5.6.1
5.6.37
37
Pettersen
,
E.F.
,
Goddard
,
T.D.
,
Huang
,
C.C.
,
Couch
,
G.S.
,
Greenblatt
,
D.M.
,
Meng
,
E.C.
et al.  (
2004
)
UCSF Chimera - a visualization system for exploratory research and analysis
.
J. Comput. Chem.
25
,
1605
1612
38
Borah
,
P.K.
,
Chakraborty
,
S.
,
Jha
,
A.N.
,
Rajkhowa
,
S.
and
Duary
,
R.K.
(
2016
)
In silico approaches and proportional odds model towards identifying selective ADAM17 inhibitors from anti-inflammatory natural molecules
.
J. Mol. Graph. Model
70
,
129
139
39
Lipinski
,
C.A.
(
2004
)
Lead- and drug-like compounds: the rule-of-five revolution
.
Drug Discov. Today Technol.
1
,
337
341
40
Jayaram
,
B.
,
Singh
,
T.
,
Mukherjee
,
G.
,
Mathur
,
A.
,
Shekhar
,
S.
and
Shekhar
,
V.
(
2012
)
Sanjeevini: a freely accessible web-server for target directed lead molecule discovery
.
BMC Bioinformatics
13
,
S7
41
Veber
,
D.F.
,
Johnson
,
S.R.
,
Cheng
,
H.Y.
,
Smith
,
B.R.
,
Ward
,
K.W.
and
Kopple
,
K.D.
(
2002
)
Molecular properties that influence the oral bioavailability of drug candidates
.
J. Med. Chem.
45
,
2615
2623
42
Rajkhowa
,
S.
,
Jha
,
A.N.
and
Deka
,
R.C.
(
2015
)
Anti-tubercular drug development: computational strategies to identify potential compounds
.
J. Mol. Graph. Model.
62
,
56
68
43
Karthick
,
V.
,
Nagasundaram
,
N.
,
Doss
,
C.G.
,
Chakraborty
,
C.
,
Siva
,
R.
,
Lu
,
A.
et al.  (
2016
)
Virtual screening of the inhibitors targeting at the viral protein 40 of Ebola virus
.
Infect. Dis. Poverty
5
,
1
12
44
Zbinden
,
G.
and
Flury-Roversi
,
M.
(
1981
)
Significance of the LD50-test for the toxicological evaluation of chemical substances
.
Arch. Toxicol.
47
,
77
99
45
Alexander
,
D.L.
,
Tropsha
,
A.
and
Winkler
,
D.A.
(
2015
)
Beware of R(2): simple, unambiguous assessment of the prediction accuracy of QSAR and QSPR models
.
J. Chem. Inf. Model.
55
,
1316
1322
46
Cheng
,
F.
,
Li
,
W.
,
Zhou
,
Y.
,
Shen
,
J.
,
Wu
,
Z.
,
Liu
,
G.
et al.  (
2012
)
admetSAR: a comprehensive source and free tool for assessment of chemical ADMET properties
.
J. Chem. Inf. Model.
52
,
3099
3105
47
Li
,
A.P.
(
2001
)
Screening for human ADME/Tox drug properties in drug discovery
.
Drug Discov. Today
6
,
357
366
48
Das
,
S.
,
Bora
,
N.
,
Rohman
,
M.A.
,
Sharma
,
R.
,
Jha
,
A.N.
and
Singha Roy
,
A.
(
2018
)
Molecular recognition of bio-active flavonoids quercetin and rutin by bovine hemoglobin: an overview of the binding mechanism, thermodynamics and structural aspects through multi-spectroscopic and molecular dynamics simulation studies
.
Phys. Chem. Chem. Phy.
20
,
21668
21684
49
Bora
,
N.
and
Jha
,
A.N.
(
2019
)
An integrative approach using systems biology, mutational analysis with molecular dynamics simulation to challenge the functionality of a target protein
.
Chem. Biol. Drug Des.
93
,
1050
1060
50
Morris
,
G.M.
,
Goodsell
,
D.S.
,
Pique
,
M.E.
,
Lindstrom
,
W.
,
Huey
,
R.
,
Forli
,
S.
et al.  (
2014
)
AutoDock Version 4.2; Updated for Version 4.2.6
,
The Scripps Research Institute
,
SL
51
Ravindranath
,
P.A.
,
Forli
,
S.
,
Goodsell
,
D.S.
,
Olson
,
A.J.
and
Sanner
,
M.F.
(
2015
)
AutoDockFR: advances in protein-ligand docking with explicitly specified binding site flexibility
.
PLoS Comput. Biol.
11
,
e1004586
52
Allouche
,
A.R.
(
2011
)
Gabedit–a graphical user interface for computational chemistry softwares
.
J. Comput. Chem.
32
,
174
182
53
Wallace
,
A.C.
,
Laskowski
,
R.A.
and
Thornton
,
J.M.
(
1995
)
LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions
.
Protein Eng.
8
,
127
134
54
Lowry
,
O.H.
,
Rosebrough
,
N.J.
,
Farr
,
A.E.
and
Randall
,
R.J.
(
1951
)
Protein measurement with Folin phenol reagent
.
J. Biol. Chem.
193
,
265
275
PMID:
[PubMed]
55
Klotz
,
K.-N.
,
Hessling
,
J.
,
Hegler
,
J.
,
Owman
,
C.
,
Kull
,
B.
,
Fredholm
,
B.B.
et al.  (
1998
)
Comparative pharmacology of human adenosine receptor subtypes - Characterization of stably transfected receptors in CHO cells
.
Naunyn Schmiedebergs Arch. Pharmacol.
357
,
1
9
56
Klotz
,
K.-N.
,
Falgner
,
N.
,
Kachler
,
S.
,
Lambertucci
,
C.
,
Vittori
,
S.
,
Volpini
,
R.
et al.  (
2007
)
[3h]HEMADO—a novel highly potent and selective radiolabeled agonist for A3 adenosine receptors
.
Eur. J. Pharmacol.
556
,
14
18
57
Banerjee
,
D.
,
Sinha
,
A.
,
Saikia
,
S.
,
Gogoi
,
B.
,
Rathore
,
A.K.
,
Das
,
A.S.
et al.  (
2018
)
Inflammation-induced mTORC2-Akt-mTORC1 signaling promotes macrophage foam cell formation
.
Biochimie
151
,
139
149
58
Jaakola
,
V.P.
,
Lane
,
J.R.
,
Lin
,
J.Y.
,
Katritch
,
V.
,
Ijzerman
,
A.P.
and
Stevens
,
R.C.
(
2010
)
Ligand binding and subtype selectivity of the human A(2A) adenosine receptor: identification and characterization of essential amino acid residues
.
J. Biol. Chem.
285
,
13032
13044
59
Boden
,
G.
(
1997
)
Role of fatty acids in the pathogenesis of insulin resistance and NIDDM
.
Diabetes
46
,
3
10
60
Duncan
,
R.E.
,
Ahmadian
,
M.
,
Jaworski
,
K.
,
Sarkadi-Nagy
,
E.
and
Sul
,
H.S.
(
2007
)
Regulation of lipolysis in adipocytes
.
Ann. Rev. Nutr.
27
,
79
101
61
Ahn
,
S.
,
Olive
,
M.
,
Aggarwal
,
S.
,
Krylov
,
D.
,
Ginty
,
D.D.
and
Vinson
,
C.
(
1998
)
A dominant-negative inhibitor of CREB reveals that it is a general mediator of stimulus-dependent transcription of c-fos
.
Mol. Cell Biol.
18
,
967
977
62
Samuel
,
V.T.
,
Petersen
,
K.F.
and
Shulman
,
G.I.
(
2010
)
Lipid-induced insulin resistance: unravelling the mechanism
.
Lancet
375
,
2267
2277
63
Gnad
,
T.
,
Scheibler
,
S.
,
von Kügelgen
,
I.
,
Scheele
,
C.
,
Kilić
,
A.
,
Glöde
,
A.
et al.  (
2014
)
Adenosine activates brown adipose tissue and recruits beige adipocytes via A2A receptors
.
Nature
516
,
395
399
64
Ohtani
,
M.
,
Oka
,
T.
and
Ohura
,
K.
(
2013
)
Possible involvement of A2A and A3 receptors in modulation of insulin secretion and β-cell survival in mouse pancreatic islets
.
Gen. Comp. Endocrinol.
187
,
86
94
65
Lebon
,
G.
,
Warne
,
T.
,
Edwards
,
P.C.
,
Bennett
,
K.
,
Langmead
,
C.J.
,
Leslie
,
A.G.
et al.  (
2011
)
Agonist-bound adenosine A2A receptor structures reveal common features of GPCR activation
.
Nature
474
,
521
525
66
Beukers
,
M.W.
,
Chang
,
L.C.
,
von Frijtag Drabbe Künzel
,
J.K.
,
Mulder-Krieger
,
T.
,
Spanjersberg
,
R.F.
,
Brussee
,
J.
et al.  (
2004
)
New, non-adenosine, high-potency agonists for the human adenosine A2B receptor with an improved selectivity profile compared to the reference agonist N-ethylcarboxamidoadenosine
.
J. Med. Chem.
47
,
3707
3709
67
Bharate
,
S.B.
,
Singh
,
B.
,
Kachler
,
S.
,
Oliveira
,
A.
,
Kumar
,
V.
,
Bharate
,
S.S.
et al.  (
2016
)
Discovery of 7-(Prolinol-N-yl)-2-phenylamino-thiazolo[5,4-d]pyrimidines as novel non-nucleoside partial agonists for the A2A adenosine receptor: prediction from molecular modeling
.
J. Med. Chem.
59
,
5922
5928
68
Haskó
,
G.
and
Cronstein
,
B.
(
2013
)
Regulation of inflammation by adenosine
.
Front. Immunol.
4
,
85

Author notes

*

These authors contributed equally to this work.