A novel protein tyrosine phosphatase 1B inhibitor with therapeutic potential for insulin resistance

Abstract

Background and Purpose

Insulin‐sensitizing drugs are currently limited, and identifying new candidates is a challenge. Protein tyrosine phosphatase 1B (PTP1B) negatively regulates insulin signalling, and its inhibition is anticipated to improve insulin resistance. Here, the pharmacological properties of CX08005, a novel PTP1B inhibitor, were investigated.

Experimental Approach

Recombinant hPTP1B protein was used to study enzyme activity and mode of inhibition. Docking simulation explored the interactions between CX08005 and PTP1B. Insulin sensitivity was evaluated by glucose tolerance test (GTT) in diet‐induced obese (DIO) and KKAy mice; glucose‐stimulated insulin secretion (GSIS), homeostasis model assessment of insulin resistance index (HOMA‐IR) and whole‐body insulin sensitivity (IS_WB) were also determined. A hyperinsulinaemic–euglycaemic clamp was performed to evaluate insulin‐stimulated glucose disposal in both whole‐body and insulin‐sensitive tissues. Furthermore, CX08005's effects on muscle, fat and liver cells were determined in vitro.

Key Results

CX08005 competitively inhibited PTP1B by binding to the catalytic P‐loop through hydrogen bonds. In DIO mice, CX08005 ameliorated glucose intolerance dose‐dependently (50–200 mg·kg⁻¹·day⁻¹) and decreased the HOMA‐IR. In KKAy mice, CX08005 (50 mg·kg⁻¹·day⁻¹) improved glucose intolerance, GSIS, IS_WB and hyperglycaemia. CX08005 also enhanced insulin‐stimulated glucose disposal, increased glucose infusion rate and glucose uptake in muscle and fat in DIO mice (hyperinsulinaemic–euglycaemic clamp test). CX08005 enhanced insulin‐induced glucose uptake in 3T3‐L1 adipocytes and C2C12 myotubes, and increased phosphorylation of IRβ/IRS1 and downstream molecules in hepatocytes in a dose‐ and insulin‐dependent manner respectively.

Conclusions and Implications

Our results strongly suggest that CX08005 directly enhances insulin action in vitro and in vivo through competitive inhibition of PTP1B.

Abbreviations

FPG

fasting plasma glucose

GSIS

glucose‐stimulated insulin secretion

GTT

glucose tolerance test

HOMA‐IR

homeostasis model assessment of insulin resistance index

IS_WB

whole‐body insulin sensitivity

MetS

metabolic syndrome

PTP1B

protein tyrosine phosphatase 1B

T2DM

type 2 diabetes

Tables of Links

TARGETS
Catalytic receptors a	Enzymes b
Insulin receptor (InsR)	Akt (PKB)
IRS1	GSK‐3β

LIGANDS
Dexamethasone	Morin
IBMX	Rosiglitazone
Insulin

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (^a,bAlexander et al., 2015a, 2015b).

Introduction

Insulin resistance is a risk factor for type 2 diabetes (T2DM) and one of the characteristics of metabolic syndrome (MetS). Insulin resistance contributes to the pathogenesis of glucose intolerance, arterial hypertension and dyslipidaemia and accelerates the development of T2DM, non‐alcoholic fatty liver disease and cardiovascular diseases in MetS (Cornier et al., 2008). The increased prevalence of MetS has become a public health concern worldwide (Grundy, 2008). Many etiological factors contribute to MetS. Nevertheless, insulin resistance appears to have a critical role in the pathophysiology of the MetS (Taylor, 2012; Mukherjee et al., 2013; Weiss et al., 2013). Insulin resistance is due to defective post‐receptor signalling of the insulin receptor, affecting muscle and adipose tissue glucose uptake, glycogen synthesis and endogenous glucose production. Insulin resistance is not only a powerful predictor of disease progression but also a therapeutic target. Currently, insulin‐sensitizing medicines are limited, and the identification of a new drug candidate is a challenge (Ye, 2011).

Protein tyrosine phosphatase 1B (PTP1B), a member of the protein tyrosine phosphatases (PTPs) family, catalyses the dephosphorylation of the insulin receptor (InsR) that controls insulin signalling activity (Koren and Fantus, 2007). PTP1B knockout mice exhibit increased sensitivity to insulin and resistance to diet‐induced T2DM (Elchebly et al., 1999; Klaman et al., 2000), suggesting that PTP1B is the main negative regulator of insulin signalling. The ability of PTP1B to directly regulate InsR activation and signalling in liver and muscle has been substantiated by the generation of liver‐ and muscle‐specific PTP1B‐deficient mice (Delibegovic et al., 2007; Delibegovic et al., 2009). Interestingly, white adipose tissue‐specific PTP1B‐deficient mice exhibit increased adiposity, suggesting that PTP1B can promote white adipose tissue (WAT) lipogenesis by an alternative means (Owen et al., 2012). PTP1B inhibitors hold promise as a class of insulin‐sensitizing agents, and a variety of small‐molecule inhibitors have been reported (Zhang and Zhang, 2007; Takada et al., 2012; Thareja et al., 2012; Tang et al., 2013; Maeda et al., 2014). A few compounds have been tested in early‐stage clinical trials, but none of them have progressed into a large clinical trial or registration (Popov, 2011; Feldhammer et al., 2013). The discovery of a potential drug as a PTP1B inhibitor is very difficult because the study is limited by the number of inhibitors, their cell permeability and poor bioavailability. Here, we identified CX08005, a novel synthetic compound developed in our lab, as an effective and competitive PTP1B inhibitor. The compound enhanced insulin sensitivity in diet‐induced obese (DIO) mice and KKAy mice. The mechanism involved in this effect appeared to be the enhancement of insulin‐induced tyrosine phosphorylation of the insulin receptor.

Methods

Recombinant PTP1B

Recombinant human His‐PTP1B was expressed in BL21 (DE3) Escherichia coli containing the expression plasmid that encodes human PTP1B (1–321) with a C‐terminal six‐histidine tag. The plasmid was kindly provided by Anthony C. Bishop, Amherst College, USA. The PTP1B protein was purified using Ni‐NTA agarose (Qiagen, Hildon, Germany).

Enzyme‐inhibiting activity and inhibition mode

The reagent 4‐nitrophenyl phosphate disodium salt (pNPP) was used as a substrate for the measurement of PTP1B activity. CX08005 was pre‐incubated with the recombinant human PTP1B at room temperature for 5 min. The assay was performed in a final volume of 100 μL containing 50 mM HEPES, 5 mM DTT, 150 mM NaCl, 2 mM EDTA and 2 mM pNPP (pH 7.0), incubated at 30°C for 10 min, stopped by addition of 50 μL 3 M NaOH. Dephosphorylation of pNPP by PTP1B generates the product p‐nitrophenol, which was monitored at 405 nm. A similar system without PTP1B protein was used as a blank control. The effects of different dosages of CX08005 were measured, and the IC₅₀ value was calculated with GraphPad Prism 5.0. The absorbance min^‐1 at 405 nm due to the p‐nitrophenol produced by dephosphorylation was measured as an index of the PTP1B enzyme activity. The 1/[S] and 1/v _i values were obtained from the substrate concentration ([S]·μM⁻¹) and initial reaction rate (v _i·μM⁻¹·min⁻¹) in the presence of 0, 2.5, 10, 25, 37.5 and 50 μM CX08005 and were plotted on the x‐ and y‐axis respectively. The inhibition mode was estimated from the intersection characteristics of the approximate lines obtained (Lineweaver–Burk plot method).

To determine the potential oxidative properties of this compound, the recombinant human PTP1B was pre‐incubated with CX08005 (10 μM) at room temperature for 15 min with different dosages of DTT (0–100 mM). The enzyme assay was performed as above. A control experiment was carried out adding just DMSO. The inhibition rate, expressed as %, was defined as the ratio between changes in PTP1B activity measured in the presence of the inhibitor and the PTP1B activity of the control.

Molecular docking

The docking simulation was performed using the Ligandfit protocol in Discovery Studio 4.0. The default parameters were used. The crystal structure of PTP1B (PDB ID: 1NNY for human PTP1B) was defined as the receptor, and the active site of PTP1B was regarded as the binding site. Compound CX08005 was docked in a flexible manner.

In vivo experiments

Animals and diets

Animals were obtained from the Animal Center of Institute of Laboratory Animal Sciences, CAMS and PUMC. C57BL/6 mice (male, average weight ~14 g; 4 weeks old) were fed high‐fat diet (50% fat, 36% carbohydrate and 14% protein in energy) for DIO. Age‐matched male C57BL/6 mice fed with the standard chow diet (1022, Beijing HFK Bioscience Co. Ltd., China, containing 12% fat, 62% carbohydrate and 26% protein in energy) were used as normal control (Con). KKAy mice (male, average weight ~35 g; 11 weeks old) were fed a chow diet (1K65, Beijing HFK Bioscience Co. Ltd., China).

Validity of animal species or model selection

C57BL/6 mice developed insulin resistance and obesity after feeding with HFD. The model mice displayed insulin resistance and glucose intolerance, which were considered as the characteristics of prediabetes (Gao et al., 2004; Cong et al., 2008).

KKAy is a typical T2DM mouse model with hyperglycaemia and hyperinsulinaemia.

Randomization

Within an experiment, animals were randomized by arbitrarily placing them in treatment groups. For the experiments using DIO mice, animals were subdivided in groups such that each group showed identical levels of insulin sensitivity and body weight. These ‘identical’ groups were then randomized to their treatment. Similarly for the KKAy mice, animals were subdivided in groups that showed equal and identical levels of fasting blood glucose, insulin sensitivity and body weight. These ‘identical’ groups were then randomized to their treatment.

Blinding

The operator was blind to the treatment drugs. The positive drug and different concentrations of CX08005 were numbered by the designer.

Housing and husbandry

Mice were grouped, housed (four mice per cage) in grommet cages at room temperature 21–23°C with a humidity of 40–60%, 12 h light/dark cycle, ad libitum access to water and chow diet. Mice are housed in specific pathogen‐free zones. Cages were lined with sterilized corncob bedding material and changed daily. Mice were given ~7 days to acclimatize to the housing conditions before the start of the experiments.

Ethical statement

Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny et al., 2010; McGrath and Lilley, 2015). All animal experiments were performed in accordance with the guidelines established by the National Institutes of Health for the Care and Use of Laboratory animals and were approved by the Animal Care Committee of the Peking Union Medical College and Chinese Academy of Medical Sciences (Beijing, China). Mice were not deprived of food or water at any time. Efforts were made to minimize animal suffering. Group numbers varied depending on availability of mice (n = 8).

Study in DIO mice

After 8 weeks of acclimatization, DIO mice were randomly divided into five groups: model control group (DIO), rosiglitazone‐treated group (Rosi, 10 mg·kg⁻¹ body weight · day⁻¹), and three CX08005‐treated groups (50, 100 and 200 mg·kg⁻¹·day⁻¹, respectively). Insulin sensitivity was determined with an i.p. glucose tolerance test (GTT) and the homeostasis model assessment of insulin resistance index (HOMA‐IR) in mice treated with CX08005 for 14 days. For i.p. GTT, glucose (1 g·kg⁻¹ body weight) was i.p. injected, and blood glucose was determined in tail vein blood at 0, 30, 60 and 120 min after glucose injection. Glucose was measured using a glucose‐oxidase assay kit (Sigma). The area under the blood glucose–time curve between 0 and 120 min was calculated. For HOMA‐IR assessment, tail vein blood was collected after 6 h of fasting. Serum insulin was measured with elisa kits (ALPCO). The HOMA‐IR was calculated according to the following formula: HOMA‐IR = (FPI × FPG)/22.5 (Wallace et al., 2004), where FPI is fasting plasma insulin concentration (mU·L⁻¹) and FPG is fasting plasma glucose (mM).

Study in KKAy mice

After 1 week of acclimatization to the environment, the mice were randomly divided into three groups: the model control group (KKAy), the rosiglitazone‐treated group (Rosi; at 10 mg·kg⁻¹ body weight · day⁻¹) and CX08005‐treated groups (50 mg·kg⁻¹·day⁻¹). Age‐matched male C57BL/6 mice were used as the normal control. After the administration of CX08005 for 16 days, an oral GTT was performed, and blood insulin level was determined. The area under the glucose–time curve (AUC_glucose), the area under the insulin–time curve (AUC_insulin), the HOMA‐IR and the whole‐body insulin sensitivity (IS_WB) index (Matsuda and DeFronzo, 1999) were calculated respectively. IS_WB was calculated as follows: IS_WB =  $\frac{10,000}{\sqrt{\left(\mathrm{FPG}\times \mathrm{FPI}\right)\times \left(\bar{\mathrm{G}}\times \bar{\mathrm{I}}\right)}}$, FPG, fasting plasma glucose; FPI, fasting plasma insulin. To determine fasting blood glucose levels, animals were deprived of food for 4 h.

Hyperinsulinaemic–euglycaemic clamp test

The hyperinsulinaemic–euglycaemic clamp was performed in DIO mice after 12 days of CX08005 treatment. The test was conducted according to a protocol established in our lab (Ye et al., 2008). Mice were deprived of food for 4 h, then anaesthetized with 80 mg·kg⁻¹ sodium pentobarbital (i.p.) and treated with 1 U·kg⁻¹ heparin (i.v.). Insulin was infused with a programmable syringe pump (Cole Parmer, Vernon Hills, ILL, USA), and glucose was perfused with a low‐flow, high‐accuracy pump (IPC, Ismatec, Switzerland). Human fast‐acting insulin was perfused continuously at a rate of 60 pmol·kg⁻¹·min⁻¹. For regulating the blood glucose to the basal level (950 ± 50 mg·L^{− 1}), 10% glucose was infused at various rates according to the value of blood glucose obtained, immediately, at 10 min intervals using a glucose analyzer (Biosen 5030, EKF Diagnostic, Germany). When the blood glucose became steady for at least 20 min, the glucose infusion rate was measured three times for a mean value. Then, the mice were injected with 2‐NBDG (250 μg per mice) and killed with exsanguination after 40 min. The epididymal adipose tissue and gastrocnemius were taken immediately, homogenized in 0.9% sodium chloride, and analysed for 2‐NBDG. The fluorescence intensity and protein concentration in these tissues were determined, and the concentration of 2‐NBDG was calculated according to the standard curve of the fluorescence intensity/2‐NBDG concentrations.

Glucose uptake in cellular models

Insulin‐stimulated glucose uptake was examined in C2C12 myotubes and 3T3‐L1 adipocytes in vitro. Murine skeletal muscle precursor C2C12 cells (myoblasts) and murine fibroblasts 3T3‐L1 cells were obtained from Cell Resource Centre, Institute of Basic Medical Sciences, CAMS and PUMC. The cells were cultured in DMEM supplemented with 10% FBS, and antibiotics (50 U·mL⁻¹ penicillin and 50 μg·mL⁻¹ streptomycin). C2C12 cells were differentiated into skeletal muscle myotubes in DMEM supplemented with 2% horse serum. 3T3‐L1 fibroblasts were differentiated into adipocytes using a standard adipogenic protocol as follows: cells were cultured in proliferation media for 48 h to confluence, treated with adipogenic cocktail (0.5 mM IBMX, 1 μM dexamethasone and 10 μg·mL⁻¹ insulin) for 2 days, and followed by insulin treatment for 3–4 days. 3T3‐L1 adipocytes were used for experiments at day 8.

Then glucose uptake was measured with a fluorescent d‐glucose analogue (2‐NBDG) according to the published procedure (Jung et al., 2011). Cells plated in the 96‐well plate were incubated with 0–500 nM CX08005 for 5 h and 100 nM insulin for 30 min. After being washed with glucose‐free DMEM, 2‐NBDG was added at a final concentration of 100 μM in glucose‐free DMEM for another 30 min. The medium was then removed, and the cells were washed three times with cold PBS. The intensity of the fluorescence was measured at ex/em 485/535 nm using a microplate fluorometer.

Insulin signalling assays

The human hepatoma cell line (HepG2) was obtained from Cell Resource Centre, Institute of Basic Medical Sciences, CAMS and PUMC, and cultured in DMEM (GIBCO) supplemented with 10% FBS (GIBCO) in a 5% CO₂ incubator. Cells were deprived of serum for 24 h and then treated with insulin (10 nM) to activate the insulin signalling pathway. Western blotting was performed to determine the phosphorylation status of signalling molecules as previously described (Ma et al., 2011). The images were acquired using a gel image analysis system (Flurochem 5500; Alpha Innotech, San Leandro, CA, USA). The following antibodies were used: β‐actin, INSR (IRβ), IRS1, IRS1 (pY895), Akt, Akt (pS473), Foxo1, Foxo1 (pS256), GSK3β, GSK3β (pS9) and goat anti‐rabbit IgG‐HRP (Cell Signaling Technology, Beverly, MA, USA), IRβ (pY1162/1163) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and PTP1B (Millipore, Bedford, MA, USA).

Statistical analysis

All values are presented as means ± SEM. The analysis of pharmacological drug effects between the model (DIO or KKAy mice in vivo; vehicle group in vitro) and treatment groups was performed using one‐way ANOVA followed by a Tukey's post hoc test to determine differences versus vehicle. After ANOVA, post hoc tests were only performed if F achieved the necessary level of statistical significance (P < 0.05) and there was no significant variance in homogeneity. If these conditions were not satisfied, data were transformed to natural logarithm (ln), and the latter problem was resolved. The difference between control and model groups was analysed using Student's t‐test. Statistical significance was set at a value of P < 0.05. The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015).

Chemicals and reagents

General reagents were purchased from Sigma‐Aldrich (St. Louis, MO, USA). Cell culture media was purchased from GIBCO (Grand Island, NY, USA). Water was collected from a PALL Cascada™ Lab Water Purification System with the water outlet operating at 18.2 MΩ (Pall Corporation, New York, USA). Recombinant human insulin was from Eli Lilly & Co (Indianapolis, IN, USA). The fluorescent glucose analogue 2‐(N‐(7‐nitrobenz‐2‐oxa‐1,3‐diazol‐4‐yl) amino)‐2‐deoxyglucose (2‐NBDG) was purchased from Invitrogen Corporation (Carlsbad, CA, USA). Insulin ultrasensitive elisa kit was purchased from ALPCO (Salem, MA, USA). PTP1B inhibitor, CX08005 (2‐[[[2‐(tetradecyloxy) phenyl]amino]carbonyl] benzoic acid, purity >97%), was chemically synthesized and identified by the Institute of Materia Medica, Chinese Academy of Medical Sciences (CAMS) and Peking Union Medical College (PUMC). The chemical structure of CX08005 is shown in Figure 1A.

Figure 1.

Structure and inhibitory activity of CX08005. (A) Chemical structure of CX08005 and (B) IC₅₀ values. Data were obtained by plotting the relative activity of PTP1B versus a standard curve of inhibitor concentration. (C) Lineweaver–Burk plot for inhibition of enzyme reaction by CX08005. At various concentrations of CX08005, the initial velocity was determined with various concentrations of p‐nitrophenyl phosphate. (D) The effect of DTT on the PTP1B inhibition elicited by CX08005. The recombinant human PTP1B was pre‐incubated with CX08005 (10 μM) at room temperature for 15 min with different concentrations of DTT (0–100 mM). Enzyme assay was performed with pNPP as a substrate. Control experiment was carried out adding DMSO.

Results

Inhibition of PTP1B by CX08005

CX08005 activity was determined in vitro using the recombinant hPTP1B. The chemical structure of CX08005 is shown in Figure 1A. CX08005 exhibited remarkable dose‐dependent inhibitory activity against PTP1B with an IC₅₀ of 7.81 × 10⁻⁷ M (Figure 1B). A typical example of the Lineweaver–Burk plot enzyme reaction is presented at each concentration of CX08005 (Figure 1C). The lines intersecting on the y‐axis demonstrate competitive inhibition of PTP1B by CX08005. Furthermore, the presence of DTT had no effect on the PTP1B inhibition elicited by CX08005, indicating that this compound is an intrinsic inhibitor of PTP1B without oxidative effects on this enzyme (Figure 1D).

Binding mode by molecular docking

Molecular docking was performed to investigate the physical interaction of CX08005 and PTP1B at the atomic level (Figure 2). The results suggest that CX08005 forms a hydrogen bond with PTP1B protein at Arg²²¹, Gly²²⁰; Ser²¹⁶, Ala²¹⁷, Gly²¹⁸, Cys²¹⁵; and Gln²⁶⁶ in the catalytic site.

Figure 2.

The interactions between CX08005 and PTP1B at the substrate‐binding site. The docking simulation was performed using the Ligandfit protocol in Discovery Studio 4.0. Default parameters were used. The crystal structure of PTP1B (PDB ID: 1NNY for human PTP1B) was defined as the receptor, and the active site of PTP1B was regarded as the binding site. Compound CX08005 was docked in a flexible manner.

Insulin‐sensitizing effects of CX08005 in DIO mice

CX08005 was tested in DIO mice for improvement of insulin sensitivity with the GTT. DIO mice exhibited significant impairment in glucose tolerance with a higher AUC value (increased by 25.2%) compared with the lean mice (Figure 3A and B). The impaired GTT was significantly improved by CX08005 in a dose‐dependent manner. The improvement was 16.0% (CX08005‐L), 21.7% (CX08005‐M) and 28.5% (CX08005‐H) respectively. CX08005 exhibited a similar activity to rosiglitazone (Rosi) in the improvement in GTT (Figure 3A and B). The HOMA‐IR was significantly improved by CX08005 (100 mg·kg⁻¹) in DIO mice (Figure 3C). CX08005 did not change the body weight significantly in this study (Figure 3D).

Figure 3.

Insulin‐sensitizing effects of CX08005 in male DIO mice. Animals were treated with CX08005 (50, 100 and 200 mg·kg⁻¹ in the low‐dose, middle‐dose and high‐dose group) or rosiglitazone (10 mg·kg⁻¹) for 14 days. (A) GTT, (B) AUC in GTT, (C) HOMA‐IR and (D) body weight. Data are expressed as mean ± SEM (n = 8). ^# P < 0.05 versus Con; *P < 0.05 versus DIO. IPGTT, i.p. GTT.

Insulin‐sensitizing effects of CX08005 in KKAy mice

CX08005 was tested in KKAy mice, which is a genetic model of T2DM. KKAy mice exhibited a higher level of glucose and insulin in response to the GTT. The two parameters were much higher in KKAy mice than in control mice in GTT (Figure 4A–D). After CX08005 treatment, GTT was improved significantly with 37.0% decrease in AUC_glucose (Figure 4B). Glucose‐stimulated insulin secretion was markedly suppressed by CX08005 with 38.7% reduction in AUC_insulin (Figure 4D). Notably, the elevated HOMA‐IR in KKAy group was reduced by 50.1% (Figure 4E). The decreased IS_WB was 2.6‐fold increased by CX08005 treatment (Figure 4F). Additionally, fasting blood glucose was reduced by CX08005 significantly when the assay was performed at days 7 and 12 respectively (Figure 4G).

Figure 4.

Insulin‐sensitizing effects of CX08005 in male KKAy mice. Animals were treated with CX08005 (50 mg·kg⁻¹) or rosiglitazone (10 mg·kg⁻¹) for 16 days. (A) oral GTT (OGTT), (B) AUC in OGTT, (C) blood insulin levels after the glucose loading, (D) area under the insulin curve in OGTT, (E) HOMA‐IR, (F) whole‐body insulin sensitivity index and (G) fasting blood glucose. Data are expressed as mean ± SEM (n = 8). ^# P < 0.05 versus Con; *P < 0.05 versus KKAy.

CX08005 enhances insulin‐stimulated glucose uptake

Hyperinsulinaemic–euglycaemic clamp test in DIO mice

To access the effect of CX08005 on tissue‐specific insulin sensitivity, we performed a hyperinsulinaemic–euglycaemic clamp test in DIO mice. The glucose infusion rate was 83.6% lower in DIO mice than in lean mice (Figure 5A). After a 12 day treatment with CX08005, the glucose infusion rate was increased significantly in DIO mice. The increase was dose‐dependent at 69.8% (L), 188.2% (M) and 339.4% (H). At the high dosage, CX08005 exhibited comparable activity to rosiglitazone (Figure 5A). Tissue‐specific insulin sensitivity was determined in epididymal adipose tissue and gastrocnemius with 2‐NBDG infusion. Insulin sensitivity was 43.2% (muscle) and 55.4% (adipose) lower in DIO mice (Figure 5B and C). After CX08005 treatment, the insulin sensitivity was increased by 65.7% and 148.4% in those tissues of DIO mice (Figure 5B and C). CX08005 exhibited a comparable activity to that of rosiglitazone in this assay. These data suggest that CX08005 enhances the effects of insulin in muscle and fat tissues of DIO mice.

Figure 5.

CX08005 elevates insulin‐stimulated glucose uptake. (A) Glucose infusion rate (GIR) during the steady state in hyperinsulinaemic–euglycaemic clamp test in DIO mice. Animals were treated with CX08005 at three dosages (50, 100 and 200 mg·kg⁻¹ for low‐dose, middle‐dose and high‐dose group) for 12 days. Rosiglitazone (10 mg·kg⁻¹) was used as a positive control. The mice were fasted for 4 h and infused with insulin at 60 pmol·kg⁻¹·min⁻¹ and 10% glucose at the different rates to maintain the blood glucose at 950 ± 50 mg·L⁻¹. (B) 2‐NBDG uptake by skeletal muscle. (C) 2‐NBDG uptake by adipose tissue. Data are expressed as mean ± SEM (n = 8). ^# P < 0.05 versus Con; *P < 0.05 versus DIO. (D, E) C2C12 myotubes and 3T3L1 adipocytes were treated with CX08005 at the concentrations for 5 h. Insulin treatment was 30 min in serum‐free DMEM. Then 2‐NBDG was added at a final concentration of 100 μM for 30 min. Then, 2‐NBDG uptake was measured using a microplate fluorometer. Data are expressed as mean ± SEM (n = 8). *P < 0.05 versus insulin.

Enhanced insulin action in cells

The effects of CX08005 on insulin sensitivity were tested in cellular models in vitro. Insulin‐induced glucose uptake was determined in C2C12 myotubes and 3T3‐L1 adipocytes. The glucose uptake was induced with 100 nM insulin. CX08005 was tested at concentrations from 10 to 500 nM. CX08005 enhanced the action of insulin in both cellular models (Figure 5D and E). The maximum increase was observed with 500 nM CX08005 in 3T3L1 adipocytes with a 66.0% increase in glucose uptake (Figure 5D). These data suggest that CX08005 may directly improve insulin's action.

Enhanced insulin signalling by CX08005

Insulin action is dependent on activation of insulin receptor and the downstream signalling molecules in the insulin signalling pathway. Phosphorylation of the signalling molecules is an indicator of pathway activation. Phosphorylation of the insulin receptor and IRS1 was examined in the pathway to determine the effect of CX08005. Insulin‐stimulated phosphorylation of IRβ (Tyr^1162/1163) was increased by CX08005 treatment in the cell cultures (Figure 6A and B). Phosphorylation of IRS1 (Tyr⁸⁹⁵) was also enhanced (Figure 6A and C). Consistent with InsR/IRS1 activation, the insulin‐stimulated phosphorylation of several downstream molecules was also increased, including Akt (Ser⁴⁷³), Foxo1 (Ser²⁵³) and GSK‐3β (Ser⁹) (Figure 6A, D–F). Remarkably, the phosphorylation of GSK‐3β (Ser⁹) in insulin‐free group was also enhanced by CX08005 (Figure 6A and G). Moreover, the CX08005 activity was dependent on dosages at concentrations between 10 and 500 nM (Figure 6B–G). The protein levels of PTP1B were not altered by CX08005. These results suggest that CX08005 directly enhanced the insulin signalling activity without changing the protein level of PTP1B.

Figure 6.

Enhanced insulin signalling by CX08005. (A) Effects of CX08005 on insulin signalling pathway. HepG2 was pretreated with CX08005 for 5 h followed by 10 nM insulin treatment for 5 min in serum‐free DMEM. The phosphorylation status was determined in IRβ (InsR), IRS1, Akt, Foxo1 and GSK3β in immunoblotting. (B) Quantification of IRβ phosphorylation. (C) Quantification of IRS1 phosphorylation. (D) Quantification of Akt phosphorylation. (E) Quantification of Foxo1 phosphorylation. (F) Quantification of GSK3β phosphorylation. (G) Quantification of GSK3β phosphorylation without insulin treatment. Signal density was quantified in each band and normalized to β‐actin signal. Relative protein expression levels were obtained from five independent experiments. Data are expressed as mean ± SEM (n = 5). *P < 0.05 versus 0 nM in insulin‐treated groups; ^# P < 0.05 versus 0 nM in insulin‐free groups.

Discussion and conclusions

CX08005 inhibits PTP1B through competitive competition. PTP1B inhibits insulin signalling by tyrosine dephosphorylation of the insulin receptor. In obesity, the enhanced PTP1B activity contributes to the pathogenesis of insulin resistance (Klaman et al., 2000; Johnson et al., 2002). Inhibition of PTP1B activity is a strategy for the treatment of insulin resistance (Koren and Fantus, 2007). In the current study, CX08005 was found to be a competitive chemical inhibitor of PTP1B. CX08005 exhibited marked dose‐dependent inhibition of PTP1B with an IC₅₀ of 7.81 × 10⁻⁷ M (Figure 1B and C). It was shown to be an intrinsic inhibitor of PTP1B without oxidative effects on this enzyme (Figure 1D). Docking simulation demonstrated the competitive mode of inhibition (Figure 2), in which CX08005 binds to PTP1B at the catalytic P‐loop through hydrogen bonds at residues Cys²¹⁵, Ser²¹⁶, Ala²¹⁷, Gly²¹⁸; and Gly²²⁰, Arg²²¹. PTP1B consists of 435 amino acid residues, in which the catalytic domain includes residues 30–278. The catalytic active site of PTP1B contains a signature motif from His²¹⁴ to Arg²²¹ (His‐Cys‐Ser‐Ala‐Gly‐Ile‐Gly‐Arg), a loop of eight amino acids that forms a rigid cradle‐like structure that co‐ordinates to the aryl phosphate moiety of the substrate (Thareja et al., 2012). This loop also contains a Cys²¹⁵ residue in the active site that is responsible for executing the nucleophilic attack on the substrate phosphate moiety (Zhang, 1998; Johnson et al., 2002). It might be speculated that CX08005 competes with the substrate in binding to the catalytic active site of PTP1B.

The bioavailability of CX08005 is excellent. Bioavailability is an important issue in the development of small‐molecule inhibitors of PTP1B. Most PTP1B inhibitors have a limited cell membrane permeability and low bioavailability in vivo (Zhang and Zhang, 2007). CX08005 displayed a good pharmacokinetic profile in Sprague–Dawley rats after intragastric administration with a plasma drug exposure C_max = 7425.9 ng·mL⁻¹, T _max = 6.67 ± 1.15 h and t _1/2 = 2.5 h at 100 mg·kg⁻¹ (detailed data not shown). We found that the bioavailability of CX08005 in DIO and KKAy mice in vivo was excellent. C57BL/6 mice fed with a high‐fat diet for 8 weeks developed insulin resistance (Gao et al., 2004; Tao et al., 2009). In this study, CX08005 exhibited strong efficacy in correcting this insulin resistance. A 14 day treatment with CX08005 effectively ameliorated glucose intolerance in a dose‐dependent manner (Figure 3A and B) and decreased HOMA‐IR values (Figure 3C). In hyperinsulinaemic–euglycaemic clamp test, CX08005 increased the glucose infusion rate and glucose uptake in muscle and fat of DIO mice (Figure 5). In KKAy mice, a genetic T2DM model, CX08005 (50 mg·kg⁻¹), improved glucose intolerance, elevated the IS_WB and lowered the FPG (Figure 4). CX08005 also reduced the glucose‐stimulated insulin secretion (Figure 4C and D). In 3T3‐L1 adipocytes and C2C12 muscle cells, CX08005 enhanced insulin‐induced glucose uptake (Figure 5). These data suggest that CX08005 directly enhances the action of insulin in vitro and in vivo to promote glucose uptake.

CX08005 activity is dependent on insulin. CX08005 does not activate insulin signalling itself. In the insulin‐treated HepG2 cells, CX08005 enhanced insulin‐stimulated tyrosine phosphorylation of IRβ/IRS1 in a dose‐dependent manner (Figure 6). The enhanced phosphorylation promoted the activity of insulin pathway as indicated by the increased phosphorylation of several downstream molecules, including Akt, Foxo1 and GSK3β, suggesting that CX08005 sensitized the insulin pathway through enhancing IRβ/IRS1 tyrosine phosphorylation. As well‐known, Foxo1 and GSK3β are involved in hepatic gluconeogenesis and glycogen synthesis, indicating that CX08005 could augment insulin's ability to suppress hepatic glucose output. This might explain the lower FPG in the CX08005‐treated KKAy mice. There are several new PTP1B inhibitors with good bioactivities such as morin (Paoli et al., 2013), safranal (Maeda et al., 2014) and pyrroloquinoline quinone (Takada et al., 2012). However, these inhibitors enhance glucose uptake in the absence of insulin inducing insulin‐mimic activity. Other PTP1B inhibitors such as JTT‐551 (Fukuda et al., 2010) do not have insulin‐mimic activity. Our data suggest that CX08005 does not have insulin‐mimic activity.

CX08005 shows an advantage over other PTP1B inhibitors. In cell culture, the effective concentration of CX08005 was less than 500 nM (Figures 5D and E and 6.), and had a low IC₅₀ value (7.81 × 10⁻⁷ M). These data suggest that CX08005 has better cell membrane permeability than JTT‐551 (10–30 μM) whose membrane permeability is low (Fukuda et al., 2010). In vivo, CX08005 exhibited an advantage over JTT‐551 in the improvement on hyperinsulinaemia, which was not significantly improved by JTT‐551 (Ito et al., 2014).

In terms of selectivity of CX08005 against PTP1B, we set up an experiment to check CX08005 against a panel of PTPs, including TCPTP, LAR, SHP1 and VHR. As shown in Table 1, as TCPTP shares ~72% catalytic domain sequence identity with PTP1B, the IC₅₀s of CX08005 were similar. Surprisingly, the activity of CX08005 on LAR, SHP1 and VHR was much lower than that on PTP1B. PTP1B and TCPTP are PTPs that share high sequence and structural homology yet play distinct physiological roles. While PTP1B plays a central role in metabolism and is an attractive drug target for obesity and T2DM, TCPTP is necessary for the control of inflammation. As TCPTP is involved in inflammation‐induced tumuorigenesis, it is necessary to exclude this as a side effect of CX08005 (Stuible et al., 2008). Human hepatic carcinoma HepG2 cells were treated with CX08005, and the cell viability was assessed by CCK‐8 kit. It was observed that there were no significant changes in cell viability up to 500 nM (detailed data not shown).

Table 1.

Selectivity of CX08005 against a panel of PTPs

	Inhibition (%) by CX08005 (10 μM)	IC50 (M)
PTP1B	100.8	7.81E‐07
TCPTP	101.3	4.75E‐07
LAR	51.0	—
SHP1	88.5	—
VHR	60.2	—

In summary, we identified CX08005 as a new PTP1B inhibitor that improved insulin sensitivity and insulin‐stimulated glucose disposal in DIO mice and KKAy mice. CX08005 enhanced insulin‐induced IRβ/IRS1 phosphorylation through competitive inhibition of PTP1B. The bioavailability of CX08005 in vivo was shown to be excellent.

Author contributions

X.Z., J.T., J.L., L.H., W.L. and L.Z. conducted the experiments and performed data analysis. F.Y., S.W. and J.Y. designed the study, made data interpretation and prepared the manuscript.

Conflict of interest

The authors declare no conflicts of interest.

Declaration of transparency and scientific rigour

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organizations engaged with supporting research.

Zhang, X. , Tian, J. , Li, J. , Huang, L. , Wu, S. , Liang, W. , Zhong, L. , Ye, J. , and Ye, F. (2016) A novel protein tyrosine phosphatase 1B inhibitor with therapeutic potential for insulin resistance. British Journal of Pharmacology, 173: 1939–1949. doi: 10.1111/bph.13483.