A hallmark of type 2 diabetes is impaired insulin receptor (IR) signaling that results in dysregulation of glucose homeostasis. Understanding the molecular origins and progression of diabetes and developing therapeutics depend on experimental models of hyperglycemia, hyperinsulinemia, and insulin resistance. We present a novel monoclonal antibody, IRAB-B, that is a specific, potent IR antagonist that creates rapid and long-lasting insulin resistance. IRAB-B binds to the IR with nanomolar affinity and in the presence of insulin efficiently blocks receptor phosphorylation within minutes and is sustained for at least 3 days in vitro. We further confirm that IRAB-B antagonizes downstream signaling and metabolic function. In mice, a single dose of IRAB-B induces rapid onset of hyperglycemia within 6 h, and severe hyperglycemia persists for 2 weeks. IRAB-B hyperglycemia is normalized in mice treated with exendin-4, suggesting that this model can be effectively treated with a GLP-1 receptor agonist. Finally, a comparison of IRAB-B with the IR antagonist S961 shows distinct antagonism in vitro and in vivo. IRAB-B appears to be a powerful tool to generate both acute and chronic insulin resistance in mammalian models to elucidate diabetic pathogenesis and evaluate therapeutics.
Introduction
Impaired insulin signaling leads to physiological alterations in glucose homeostasis in diabetes (1–3). Insulin binds to the insulin receptor (IR) and induces autophosphorylation (4). This initiates a cascade that results in a multitude of metabolic functions, such as glucose uptake, glycogen synthesis, and inhibition of hepatic glucose output (5,6). Dysregulation of this pathway contributes to insulin resistance, leading to hyperglycemia and hyperinsulinemia (7–9). Treatment of hyperglycemia and hyperinsulinemia depends on reliable in vitro and in vivo models to better understand the pathophysiology associated with insulin resistance and evaluate potential therapies.
In humans, insulin resistance is associated with several alterations in normal physiology, such as adiposity, cellular stress, and inflammation (10,11). Multiple strategies with various mechanisms have been used to induce insulin resistance. Treating cells with cytokines, high insulin, glucose, and/or lipids and hypoxia exposure lead to decreased IR signaling within minutes to days (12–16). In vivo, insulin resistance has been induced genetically and environmentally, as with the liver IR knockout mouse and the diet-induced obesity mouse (17–25). However, these models are costly, take weeks to generate, and may be subject to adaptive responses. Therefore, less costly, inducible in vitro and in vivo mammalian models that incorporate both rapid-onset and chronic hyperglycemia, hyperinsulinemia, or insulin resistance would be useful.
A potential strategy is to directly block the IR. One molecule, S961, a biosynthetic peptide with mixed agonist/antagonist properties, directly competes with insulin (26,27). In vivo, S961 induces insulin resistance at high concentrations and has been used to generate rodent models of diabetes (28–30). Although S961 can induce insulin resistance, it acts as an agonist at low concentrations (26). This will produce altered signaling depending on pharmacokinetics in vivo and is a clear limitation of this tool. Therefore, pure IR antagonists are preferred.
We present a novel, potent allosteric monoclonal antibody (mAb) IR antagonist, IRAB-B. We show that IRAB-B selectively binds to the IR with nanomolar affinity and in the presence of insulin, decreases IR phosphorylation in vitro, impairs signaling, and prevents glucose uptake. In vivo, IRAB-B rapidly induces severe hyperglycemia and hyperinsulinemia in mice, with prolonged duration from a single administration. Furthermore, we show that hyperglycemia induced by IRAB-B is effectively reversed by a GLP-1 receptor agonist. Finally, we compare the effects of IRAB-B with S961 and show that they display differential activity. Overall, IRAB-B is a powerful tool to block cellular insulin signaling and induce hyperglycemia and insulin resistance in vivo. This mAb could lend valuable insight into the underlying biology of diabetes pathogenesis and provide a system to screen therapeutics in multiple mammalian models.
Research Design and Methods
Materials
Insulin (Humulin R, R002-8215-01; Eli Lilly) and S961 (051-86; Phoenix Peptide) were diluted in sterile PBS. Antibodies (Abs) were sourced as follows: β-actin (13E5) rabbit mAb (4970; Cell Signaling), rabbit (polyclonal) anti-IR (pY972) phosphospecific Ab (44800G; Invitrogen), rabbit (polyclonal) anti-insulin/IGF-I receptor (pY1158) phosphospecific Ab (44-802G; Invitrogen), phospho-IGF-I receptor-β (Tyr1135/1136)/IR-β (Tyr1150/1151) (19H7) rabbit mAb (3024; Cell Signaling), phospho-IR-β (Tyr1345) (14A4) rabbit mAb (3026; Cell Signaling), rabbit (polyclonal) anti-IR (pY1334) phosphospecific Ab (44-809G; Invitrogen), Akt (pan) C67E7 rabbit mAb (4691; Cell Signaling), phospho-Akt (Ser473) (D9E) XP (4060; Cell Signaling), and IR-β (4B8) rabbit mAb (3025; Cell Signaling). Secondary donkey anti-rabbit IgG (711-006-152; Jackson ImmunoResearch) was ruthenium labeled with MSD GOLD SULFO-TAG NHS-Ester (R91AO-2; Meso Scale Discovery [MSD]). Insulin Signaling Panel (Total Protein) Kit and Insulin Signaling (Phospho Protein) Kit were purchased from MSD (K15152C-3 and K15151C-3).
Identification of Anti-IR Abs
Fab phage display panning was conducted to identify IR-binding Abs. The long isoform of the human IR (11081-H08H; Sino Biological) was biotinylated and used as an antigen for panning with a panel of human Fab libraries. After three rounds of panning through ELISA screen, successful binders were then expressed as human Abs. Monomeric Abs were screened against HuH7 cells (JCRB0403; JCRB Cell Bank) with and without the presence of human insulin (Sigma-I1507, I2643). Abs that bound to cells were sorted into different epitope bins through MSD using Abs 83-14 (ab44914; Abcam), 83-7 (ab36550; Abcam), and MA-20 (MA1-16812; Pierce Antibodies). Abs were confirmed for binding to rat L6 skeletal muscle cells (ATCC CRL-1458) as well as for the absence of binding to IGF-I receptor (10164-H08H; Sino Biological).
IRAB-B Binding Affinity Studies
IRAB-B binding to IR-A and IR-B extracellular constructs was tested by surface plasmon resonance (SPR) with a ProteOn XPR36 Protein Interaction Array system (Bio-Rad). Goat anti-human IgG (Fc) was covalently coupled to the surface of a GLC Sensor Chip (Bio-Rad) based on manufacturer instructions for amine coupling chemistry. Approximately 5,500 response units (RU) of goat anti-human IgG (Fc) antibody (109-005-098; Jackson ImmunoResearch) were immobilized. The sensor chip surface was also coated with 500–700 RU of a nonspecific IgG (IgG1 Λ, 15029-1MG; Sigma-Aldrich). Experiments were performed at 25°C in running buffer (PBS [pH 7.4], 0.005% P20 surfactant, 3 mmol/L EDTA). Four-fold serial dilutions of human IR-A or IR-B (400–1.56 nmol/L) were prepared in running buffer. Forty to 60 RU of the mAbs were captured on each channel of the sensor chip followed by injection (association phase) of antigen and then followed by buffer flow (dissociation phase). Reference spots (goat anti-human Fc) were used for reference subtraction. To determine whether insulin was able to bind to the receptor in the presence of IRAB-B, IRAB-B was captured on the anti-Fc channels of a GLC Sensor Chip. IR-A and/or IR-B were injected over the anti-Fc surface followed by insulin dilutions (2,000–3.2 nmol/L at fivefold dilutions) injected over the IR/mAb complex surface (association phase) followed by running buffer flow (dissociation phase). Interspots with nonspecific IgG were used for reference correction. Double referencing of the data was performed by subtracting the curves generated by buffer injection from the reference-subtracted curves for analyte injections. Data were processed on instrument software.
Cell Culture
HuH7 cells were plated at 50,000 cells/well (100 μL) in 96-well plates for IR phosphorylation assays or Falcon 6-well plates at 4.8 × 105 cells/well (2 mL) for Western blot assays in DMEM + GlutaMAX with 10% heat-inactivated FBS and incubated at 37°C in 5% CO2 for 18–24 h before use. 3T3-L1 fibroblasts were maintained in DMEM (11965-092; Gibco) containing 10% cosmic calf serum (SH30413.02; HyClone) and 5% CO2 at 37°C. Two days after reaching confluence, differentiation was induced by incubating cells for 48 h in DMEM containing 10% FBS, 0.5 mmol/L isobutylmethylxanthine (Sigma-I7018), 0.25 mmol/L dexamethasone (Sigma-D4902), and 1 μg/mL insulin (Sigma-I5523). After 2 days, the isobutylmethylxanthine and dexamethasone were removed, and insulin was maintained for 2 additional days. Insulin was removed, and cells were allowed to complete differentiation and used at 9–12 days postdifferentiation.
Cell-Based Assays
Insulin, IRAB-B, and S961 were all diluted in culture media. All treatments were performed at 37°C. Reactions were stopped by aspirating cell media and washing cells with ice-cold PBS. MSD Lysis Buffer (with protease and phosphatase inhibitors) was added. Plates were shaken for 30 min at room temperature and frozen at −80°C until analysis.
MSD Analyses for Insulin Signaling
Cell lysates were thawed on ice immediately before analysis. MSD assays were performed using the MSD Insulin Signaling Panel Kits (K15152C-3 and K15151C-3). MSD plates were blocked for 1 h at room temperature with vigorous shaking. Cell lysates were added to appropriate wells and the plates shaken at room temperature for 1 h. Site-specific Ab dilutions were prepared with 1/750 dilution of each IR phosphotyrosine Ab and 1/750 dilution of ruthenium-labeled secondary anti-rabbit IgG Ab. Total IR and pan phosphotyrosine Abs from MSD Insulin Signaling Panel Kits were diluted to 1/50. Ab dilutions were added to appropriate wells and shaken at room temperature for 1 h. Plates were washed three times between each step. MSD Read Buffer was added to wells and read on an MSD Plate Reader.
Western Blot Analysis
Cleared lysates were diluted with NuPAGE LDS Sample Buffer (NP0007; Invitrogen) and NuPAGE Sample Reducing Agent (NP0004; Invitrogen). Samples were heated to 98°C and loaded onto gels and resolved with NuPAGE MOPS SDS Running Buffer at 150 V for 60 min. Gels were transferred to polyvinylidene fluoride membrane by using iBlot Gel Transfer Stacks, PVDF, regular size (IB401001; Novex), using an iBlot Gel Transfer system. Membranes were blocked for 1 h at room temperature in Odyssey Blocking Buffer in PBS (927-40000; LI-COR) and washed three times with PBS. Primary Abs were diluted to 1/1,000 in 1% Odyssey Blocking Buffer in Tris-buffered saline with Tween (TBST) added for 1 h at room temperature. Membranes were washed three times with TBST. Goat anti-rabbit IRDye 800CW secondary Ab (926-32211, LI-COR) was diluted to 1/5,000 in 1% (volume for volume) Odyssey Blocking Buffer in TBST and added to membranes for 1 h at room temperature. Membranes were washed with TBST three times and imaged with an Odyssey infrared scanner.
Glucose Uptake Assay
Differentiated 3T3-L1 fibroblasts were washed with PBS (21-040-CM; Corning) and serum starved for 16 h in DMEM + 0.2% (weight for volume) BSA (Sigma-A9205). Cells were washed once with PBS and once with Krebs-Ringer phosphate HEPES buffer (KRPH) (20 mmol/L HEPES, 10 mmol/L NaPO4, 0.9 mmol/L MgSO4, 0.9 mmol/L CaCl2, 136 mmol/L NaCl, 4.7 mmol/L KCl [pH 7.4]) + 0.2% (weight for volume) BSA. Cells were incubated in KRPH + 0.2% BSA for 1 h at 37°C. Cytochalasin B 1 μmol/L(Sigma-C67620) was added to select wells, and cells were incubated for 1 h at 37°C. IRAB-B was added for 15 min followed by insulin for 15 min at 37°C and then a 10× solution of 2-deoxyglucose (Sigma-D8375) and [3H]-2-deoxyglucose (NET549001MC; PerkinElmer) in KRPH + 0.2% BSA for a final concentration of 100 μmol/L 2-deoxyglucose, and 0.5 μCi/mL [3H]-2-deoxyglucose was added for an additional 15 min. Cells were washed five times with ice-cold PBS and lysed with 0.05% SDS/PBS shaking for 30 min at room temperature. An aliquot of the lysate was added to Microscint-20 (6013621; PerkinElmer) scintillation fluid, and 3H was counted on the PerkinElmer TopCount NXT. Total protein was determined by Pierce BCA Protein Assay Kit (23225; Thermo Fisher Scientific). Counts per minute were normalized to total protein and cytochalasin B control.
In Vivo Mouse Studies
All procedures using experimental animals were approved by the Institutional Animal Care and Use Committee at Janssen Pharmaceutical Companies of Johnson & Johnson. All studies were performed in C57BL/6N mice (Taconic Biosciences, Hudson, NY) housed two per cage in a temperature-controlled room with a 12-h light/dark cycle and an acclimation period of 2 weeks. The mice were allowed ad libitum access to water and fed with 5001 Laboratory Rodent Diet (LabDiet). All mice are euthanized under CO2. Dosing solutions were prepared in sterile PBS (Mediatech, Manassas, VA). Blood glucose was measured with a OneTouch glucometer (LifeScan, Milpitas, CA). Plasma insulin was measured by ELISA (Mouse/Rat Insulin Kit K152BZC; MSD). For the exendin-4 experiment, mice were injected with 5 mg/kg IRAB-B s.c., and then on days 3 and 5 postinjection, food was removed 4 h before the acute treatment with exendin-4. Four-hour fasted blood glucose levels after intraperitoneal injection of exendin-4 or saline control was measured, and no agent or nutrients were given orally.
Results
IRAB-B Allosteric Binding to IR Isoforms
IRAB-B was identified through phage panning against the human IR. The ability of IRAB-B to bind to both the short (IR-A) and long (IR-B) isoforms was tested using SPR (31,32). Binding studies showed that IRAB-B bound to each construct with similar low nanomolar affinity (Fig. 1A and B and Table 1). We conducted binding experiments to test simultaneous binding of insulin and IRAB-B. Results showed that insulin binds to the receptor in the presence of IRAB-B, indicating that IRAB-B and insulin binding sites are independent (Fig. 1C). In addition, the SPR assay showed IRAB-B binding to rat IR with a Kd of 14.7 nmol/L (data not shown). This value was expected given the IR sequence conservation (>90% human vs. mouse vs. rat) among mammalian species. Furthermore, to determine whether IRAB-B bound with the IGF-I receptor, we performed an ELISA-based screen. Binding of IRAB-B to the IGF-I receptor was equivalent to that of background controls (data not shown).
IRAB-B binds with high affinity to IR-A and IR-B isoforms. IRAB-B was captured onto an anti-human/anti-mouse Fc surface. A and B: This was followed by the titration of IR short isoform (IR-A) or IR long isoform (IR-B) at the indicated concentrations. The association and dissociation phases of the receptor binding were monitored at 4 and 30 min, respectively. The binding profiles were globally fit to a 1:1 Langmuir model (black lines). C: SPR showing insulin binding to the receptor in the presence of IRAB-B. IRAB-B was captured on an anti-human/anti-mouse Fc surface (red box) followed by the capture of the receptor through IRAB-B (blue box). Finally, insulin was titrated to test for binding to the receptor (green box). The sensorgram shows the sequential injection of IRAB-B, the IR, and the titration of insulin.
IRAB-B binds with high affinity to IR-A and IR-B isoforms. IRAB-B was captured onto an anti-human/anti-mouse Fc surface. A and B: This was followed by the titration of IR short isoform (IR-A) or IR long isoform (IR-B) at the indicated concentrations. The association and dissociation phases of the receptor binding were monitored at 4 and 30 min, respectively. The binding profiles were globally fit to a 1:1 Langmuir model (black lines). C: SPR showing insulin binding to the receptor in the presence of IRAB-B. IRAB-B was captured on an anti-human/anti-mouse Fc surface (red box) followed by the capture of the receptor through IRAB-B (blue box). Finally, insulin was titrated to test for binding to the receptor (green box). The sensorgram shows the sequential injection of IRAB-B, the IR, and the titration of insulin.
IRAB-B binding affinity to IR isoforms
Receptor isoform . | Average kon (× 105 1/ms) . | Average koff (× 10-5 1/s) . | Average Kd (nmol/L) . |
---|---|---|---|
IR-A | 2.98 (2.96–2.99) | 76.6 (58.7–94.4) | 2.57 (1.98–3.16) |
IR-B | 2.51 (2.42–2.60) | 74.8 (57.9–91.6) | 2.98 (2.39–3.52) |
Receptor isoform . | Average kon (× 105 1/ms) . | Average koff (× 10-5 1/s) . | Average Kd (nmol/L) . |
---|---|---|---|
IR-A | 2.98 (2.96–2.99) | 76.6 (58.7–94.4) | 2.57 (1.98–3.16) |
IR-B | 2.51 (2.42–2.60) | 74.8 (57.9–91.6) | 2.98 (2.39–3.52) |
The binding results were reported in the format of kon (on rate), koff (off rate), and Kd. Kinetics analysis of the mAb binding to both IR isoforms was performed using a 1:1 Langmuir binding model with group fitting option.
IRAB-B Antagonizes Insulin-Induced IR Phosphorylation In Vitro
IRAB-B was tested in a dose-titration experiment against a fixed, physiologically relevant concentration of 1 nmol/L insulin in HuH7 cells. A human IgG isotype mAb was used as a control. IRAB-B antagonized insulin-induced IR phosphorylation in a concentration-dependent manner with an IC50 of 6.5 nmol/L (Fig. 2A). Next, full insulin dose-response curves were performed in the presence of 3, 10, or 30 nmol/L of IRAB-B (Fig. 2B). IRAB-B decreased insulin potency in a dose-dependent manner, with clear shifts in the insulin dose-response curve at each IRAB-B concentration tested. IRAB-B treatment of 30 nmol/L resulted in a 10-fold half-maximal effective concentration (EC50) increase from 0.2 nmol/L of insulin alone (or isotype) to 2.25 nmol/L (Supplementary Table 1). Of note, IRAB-B did not affect insulin-induced maximum phosphorylation levels on the IR. Because this analysis measured the phosphotyrosines collectively, we subsequently evaluated each of the six individual IR phosphotyrosines by using a previously described method (33). Results indicated that individual residues are affected by IRAB-B without any apparent bias (Supplementary Fig. 1A and B).
IRAB-B decreases insulin-induced IR phosphorylation in vitro in a dose-dependent manner. A: IR tyrosine phosphorylation levels were measured in HuH7 cells treated with increasing concentrations of IRAB-B for 30 min followed by the addition of 1 nmol/L insulin for 5 min. Percent phospho-IR (%pIR) signal/total IR levels are normalized to isotype-treated cells at equivalent concentrations. B: Insulin dose-response curves of IR tyrosine phosphorylation in the presence of IRAB-B. Cells were treated for 30 min with 30 nmol/L isotype or IRAB-B at 3, 10, or 30 nmol/L IRAB-B followed by increasing concentrations of insulin for 5 min. The %pIR signal/total IR levels are normalized to the insulin dose-response curve. All data were obtained using MSD analysis and are mean ± SEM (n = 3). IC50 was calculated as log(inhibitor) vs. response – variable slope (four parameters), and EC50 values were calculated as log(agonist) vs. response – variable slope (four parameters) in GraphPad Prism 6 software.
IRAB-B decreases insulin-induced IR phosphorylation in vitro in a dose-dependent manner. A: IR tyrosine phosphorylation levels were measured in HuH7 cells treated with increasing concentrations of IRAB-B for 30 min followed by the addition of 1 nmol/L insulin for 5 min. Percent phospho-IR (%pIR) signal/total IR levels are normalized to isotype-treated cells at equivalent concentrations. B: Insulin dose-response curves of IR tyrosine phosphorylation in the presence of IRAB-B. Cells were treated for 30 min with 30 nmol/L isotype or IRAB-B at 3, 10, or 30 nmol/L IRAB-B followed by increasing concentrations of insulin for 5 min. The %pIR signal/total IR levels are normalized to the insulin dose-response curve. All data were obtained using MSD analysis and are mean ± SEM (n = 3). IC50 was calculated as log(inhibitor) vs. response – variable slope (four parameters), and EC50 values were calculated as log(agonist) vs. response – variable slope (four parameters) in GraphPad Prism 6 software.
Next, a time course experiment in HuH7 cells treated with insulin alone or in the presence of IRAB-B was performed. Insulin rapidly induced IR phosphorylation to a maximum level at approximately 5 min that gradually decreased over 120 min (Fig. 3A). Pretreatment with IRAB-B prevented insulin-induced IR tyrosine phosphorylation. Because IRAB-B binds with high affinity, we hypothesized that IRAB-B would antagonize receptor activation over extended durations. Cells were pretreated with IRAB-B for 3 h or 30 min or cotreated with an insulin dose titration for 5 min. IRAB-B induced a robust, yet equal 10-fold decrease of insulin potency at each duration (Fig. 3B).
IRAB-B prevents insulin-induced IR phosphorylation over time. A: Insulin time course of cells stimulated with 1 nmol/L insulin that were treated 30 min before each time point with 30 nmol/L IgG isotype or 30 nmol/L IRAB-B. Phospho-IR (pIR) signal/total IR levels were normalized to percent maximum of insulin curve. B: Insulin dose-response curves in cells treated with IRAB-B. Cells were treated with 30 nmol/L IRAB-B for 3 h, 30 min, or 0 min (cotreatment with insulin) followed by 5-min insulin stimulation. Data are presented as log(agonist) vs. response – variable slope (four parameters) calculated in GraphPad Prism 6 software, with percent pIR (%pIR) signal/total IR levels normalized to 30 nmol/L for 3 h IgG isotype dose-response curve. C: HuH7 cells were treated with 30 nmol/L IRAB-B or IgG isotype for an extended duration of 24, 48, and 72 h followed by 0, 1, 10, or 100 nmol/L insulin stimulation for 5 min. pIR signal/total IR levels were normalized to percent maximum. D: Total IR levels were calculated for each condition and normalized to percent maximum. All data were obtained using MSD analysis and are mean ± SEM (n = 3). Unpaired two-tailed t test was used for comparisons with IgG isotype controls. *P < 0.05; **P < 0.01; ***P < 0.001.
IRAB-B prevents insulin-induced IR phosphorylation over time. A: Insulin time course of cells stimulated with 1 nmol/L insulin that were treated 30 min before each time point with 30 nmol/L IgG isotype or 30 nmol/L IRAB-B. Phospho-IR (pIR) signal/total IR levels were normalized to percent maximum of insulin curve. B: Insulin dose-response curves in cells treated with IRAB-B. Cells were treated with 30 nmol/L IRAB-B for 3 h, 30 min, or 0 min (cotreatment with insulin) followed by 5-min insulin stimulation. Data are presented as log(agonist) vs. response – variable slope (four parameters) calculated in GraphPad Prism 6 software, with percent pIR (%pIR) signal/total IR levels normalized to 30 nmol/L for 3 h IgG isotype dose-response curve. C: HuH7 cells were treated with 30 nmol/L IRAB-B or IgG isotype for an extended duration of 24, 48, and 72 h followed by 0, 1, 10, or 100 nmol/L insulin stimulation for 5 min. pIR signal/total IR levels were normalized to percent maximum. D: Total IR levels were calculated for each condition and normalized to percent maximum. All data were obtained using MSD analysis and are mean ± SEM (n = 3). Unpaired two-tailed t test was used for comparisons with IgG isotype controls. *P < 0.05; **P < 0.01; ***P < 0.001.
We next tested whether this mAb has prolonged effects in vitro. Long-term IRAB-B treatment in HuH7 cells was performed for 24, 48, and 72 h followed by stimulation with increasing concentrations of insulin (Fig. 3C). IRAB-B maintained antagonist activity for the duration of this experiment and significantly decreased (P < 0.05) basal phosphorylation (0 nmol/L insulin) of the IR. To test whether these effects were due to prolonged IR antagonism or altered cellular receptor levels, we measured total IR and found that levels were consistent in each condition (Fig. 3D). IRAB-B antagonized the IR for at least 72 h after a single treatment.
Taken together, these data show that IRAB-B is a potent IR antagonist that can efficiently and robustly block IR activation in vitro. IRAB-B prevents IR phosphorylation dose dependently with rapid and sustained effects.
IRAB-B Affects Downstream Insulin Signaling Cascade and Functional Glucose Uptake In Vitro
To confirm that IRAB-B blocks IR signaling, we tested the effects on downstream targets and function. IRAB-B treatment resulted in decreased basal and insulin-induced levels of IRS-1 tyrosine phosphorylation and Akt phosphorylation on Ser473 in HuH7 cells (Fig. 4A and B). Finally, we found that 10 nmol/L IRAB-B treatment significantly (P < 0.05) decreases insulin-induced glucose uptake in 3T3 cells (Fig. 4C). These results confirm IRAB-B activity on insulin signaling in an additional cell model.
IRAB-B antagonizes downstream insulin signaling and glucose uptake in vitro. A: Insulin dose-response curves of IRS-1 tyrosine phosphorylation were measured in HuH7 cells treated with 30 nmol/L IgG isotype or IRAB-B (30 min) followed by insulin stimulation for an additional 5 min. Data were obtained using MSD analysis normalized to the IgG isotype curve and are mean ± SEM (n = 3). B: Western blots from HuH7 cells that were treated for 30 min with 10 nmol/L of IgG isotype or IRAB-B. Cells were unstimulated or stimulated with 3 nmol/L insulin for an additional 15 min. Western blots were performed against Akt Ser473 phosphorylation and a β-actin loading control. C: Glucose uptake was measured in 3T3 cells treated for 15 min with increasing concentrations of IRAB-B and stimulated with insulin for 15 min. Data are mean ± SEM (n = 3). Statistical comparisons were performed with GraphPad Prism 6 software. One-way ANOVA and Dunnett test were used for multiple comparisons with untreated controls. *P < 0.05. CPM, counts per minute.
IRAB-B antagonizes downstream insulin signaling and glucose uptake in vitro. A: Insulin dose-response curves of IRS-1 tyrosine phosphorylation were measured in HuH7 cells treated with 30 nmol/L IgG isotype or IRAB-B (30 min) followed by insulin stimulation for an additional 5 min. Data were obtained using MSD analysis normalized to the IgG isotype curve and are mean ± SEM (n = 3). B: Western blots from HuH7 cells that were treated for 30 min with 10 nmol/L of IgG isotype or IRAB-B. Cells were unstimulated or stimulated with 3 nmol/L insulin for an additional 15 min. Western blots were performed against Akt Ser473 phosphorylation and a β-actin loading control. C: Glucose uptake was measured in 3T3 cells treated for 15 min with increasing concentrations of IRAB-B and stimulated with insulin for 15 min. Data are mean ± SEM (n = 3). Statistical comparisons were performed with GraphPad Prism 6 software. One-way ANOVA and Dunnett test were used for multiple comparisons with untreated controls. *P < 0.05. CPM, counts per minute.
IRAB-B Is an IR Antagonist In Vivo
Having observed the potent antagonism of IRAB-B on IR signaling in vitro, we tested whether IRAB-B could induce insulin resistance in vivo. Normal C57 mice were subcutaneously dosed with 5, 10, or 25 mg/kg IRAB-B, and changes in fed and fasted blood glucose levels were monitored. IRAB-B treatment resulted in elevated fed blood glucose levels as rapidly as 4 h postdose and in severe hyperglycemia (>400 mg/dL) at 6 h postdose in the 25 mg/kg group (Fig. 5A). No change in fed blood glucose was observed for lower IRAB-B doses at early time points. Next, mice were fasted overnight and challenged with an oral glucose tolerance test (OGTT) (2 g/kg) 24 h after dosing with IRAB-B. Fasted blood glucose levels were increased upon IRAB-B treatment, and administration of an OGTT resulted in blood glucose levels >500 mg/dL within 30 min that remained >400 mg/dL over 120 min for each IRAB-B dosage cohort (Fig. 5B). To test the duration of IRAB-B hyperglycemic effects, fed blood glucose was measured out to 3 weeks after the initial IRAB-B dose (Fig. 5C). After 1 week, the 10 and 25 mg/kg IRAB-B treatment resulted in fed blood glucose levels of ∼525 and 600 mg/dL, respectively. Of note, 2 weeks after a single dose of 25 mg/kg IRAB-B, blood glucose levels remained elevated at ∼575 mg/dL. Fed and fasting blood glucose levels normalized 22 days postinjection with IRAB-B (Fig. 5C and data not shown). Despite severe hyperglycemia, IRAB-B–treated mice did not exhibit adverse reactions throughout the entire study. However, IRAB-B–treated mice exhibited significant decreases in body weight after a single administration of 10 mg/kg IRAB-B at 1 week postadministration. The 25 mg/kg cohort maintained a significant weight decrease out to 2 weeks postadministration (Fig. 5D and E). Body weight for all IRAB-B–treated cohorts normalized 3 weeks after initial IRAB-B treatment. To better understand pharmacokinetic properties of IRAB-B, serum levels were measured. IRAB-B serum levels were less than isotype levels on day 4 and showed even more pronounced differences compared with isotype at days 6 and 8 likely due to changes in IRAB-B tissue engagement (Supplementary Fig. 2).
IRAB-B induces hyperglycemia and weight loss in vivo. Blood glucose was measured in normal C57 mice treated with IRAB-B or isotype. Body weight and fed blood glucose were measured 1 h before subcutaneous injection with a weight-adjusted dose volume of PBS, IgG isotype, or IRAB-B at 5, 10, and 25 mg/kg. A: Changes in blood glucose were monitored at 2, 4, 6, and 8 h postdose. B: One day after treatment, a fasted (16-h) OGTT (2 g/kg) was performed. Glucose excursions were measured at 0, 30, 60, and 120 min post–glucose challenge. C–E: Blood glucose and body weight were followed for an additional 3 weeks (days 7, 14, and 21). F and G: Body weight and fed blood glucose were measured 1 h before subcutaneous injection with a weight-adjusted dose volume of IgG isotype or IRAB-B at 5 mg/kg. On days 3 and 5 postdose, exendin-4 was evaluated for efficacy to reverse IRAB-B–induced hyperglycemia. A 4-h food removal preceded intraperitoneal administration of saline or exendin-4 (1 μg/kg). Blood glucose was measured at 0, 30, 60, 120, 180, and 240 min posttreatment. The mice were grouped based on average glucose (1°) and body weight (2°), and data are mean ± SEM (n = 8) for each experimental group. Statistical comparisons were performed using GraphPad Prism 6 software. One-way ANOVA and Dunnett test were used for multiple comparisons with vehicle (five groups). *P < 0.05; ***P < 0.001; ****P < 0.0001.
IRAB-B induces hyperglycemia and weight loss in vivo. Blood glucose was measured in normal C57 mice treated with IRAB-B or isotype. Body weight and fed blood glucose were measured 1 h before subcutaneous injection with a weight-adjusted dose volume of PBS, IgG isotype, or IRAB-B at 5, 10, and 25 mg/kg. A: Changes in blood glucose were monitored at 2, 4, 6, and 8 h postdose. B: One day after treatment, a fasted (16-h) OGTT (2 g/kg) was performed. Glucose excursions were measured at 0, 30, 60, and 120 min post–glucose challenge. C–E: Blood glucose and body weight were followed for an additional 3 weeks (days 7, 14, and 21). F and G: Body weight and fed blood glucose were measured 1 h before subcutaneous injection with a weight-adjusted dose volume of IgG isotype or IRAB-B at 5 mg/kg. On days 3 and 5 postdose, exendin-4 was evaluated for efficacy to reverse IRAB-B–induced hyperglycemia. A 4-h food removal preceded intraperitoneal administration of saline or exendin-4 (1 μg/kg). Blood glucose was measured at 0, 30, 60, 120, 180, and 240 min posttreatment. The mice were grouped based on average glucose (1°) and body weight (2°), and data are mean ± SEM (n = 8) for each experimental group. Statistical comparisons were performed using GraphPad Prism 6 software. One-way ANOVA and Dunnett test were used for multiple comparisons with vehicle (five groups). *P < 0.05; ***P < 0.001; ****P < 0.0001.
To assess whether the hyperglycemia induced by IRAB-B can be ameliorated by a clinically validated antihyperglycemic therapy, we tested a GLP-1 receptor agonist, exendin-4, on lowering blood glucose in IRAB-B–treated mice. Mice were dosed with IRAB-B or isotype at 5 mg/kg (single dose). On days 3 and 5 after dosing, 4-h fasting blood glucose was measured, and exendin-4 or saline were administered and measured over 240 min (Fig. 5F and G). IRAB-B–treated mice had elevated fasting blood glucose on both day 3 and day 5 compared with isotype. On day 3, exendin-4 treatment exhibited a slight decrease in IRAB-B–induced hyperglycemia. However, exendin-4 clearly lowered blood glucose to baseline levels in IRAB-B–treated mice within 120 min on day 5, demonstrating that the IRAB-B model of hyperglycemia can be therapeutically treated.
IRAB-B and S961 Display Distinct IR Antagonism
Because IRAB-B binds in the presence of insulin, whereas S961 has been reported to compete with insulin, we hypothesized that these molecules may exhibit different pharmacology (26,27,30). We compared insulin-induced changes in IR and Akt phosphorylation in HuH7 cells treated with IRAB-B or S961 (Fig. 6A). IRAB-B blocked insulin-induced phosphorylation in a concentration-dependent manner, with 100 nmol/L IRAB-B completely ablating the effects of 1 nmol/L insulin, whereas S961 fully blocked insulin-induced phosphorylation at all S961 concentrations tested. To further elucidate differences between the antagonists, cells were treated with fixed, approximate equieffective antagonist concentrations followed by an insulin titration (Fig. 6B). IRAB-B and S961 antagonize insulin-induced phosphorylation but at different potencies across the insulin titration. Because the phosphorylation patterns were different, we next evaluated the treatment duration of these antagonists. Cells were treated with IRAB-B or S961 for increasing durations followed by an insulin challenge (Fig. 6C). IRAB-B–treated cells displayed consistent phosphorylation levels independent of treatment duration. In contrast, S961 treatment appeared to decrease IR and Akt phosphorylation with increasing duration, which may be due to the mixed agonist/antagonist properties or stability of this peptide. Thus, IRAB-B and S961 showed different effects on insulin-induced IR and Akt phosphorylation with treatment time.
IRAB-B and S961 display distinct antagonism of IR signaling in vitro. Insulin-induced IR Tyr1162/1163 and Akt Ser473 phosphorylation levels were measured by Western blot in HuH7 cells treated with IRAB-B or S961. A: Cells were treated with increasing concentrations (1, 10, or 100 nmol/L) of IRAB-B or S961 for 30 min and stimulated with 1 nmol/L insulin for 15 min. B: Cells were treated with 100 nmol/L IRAB-B or 1 nmol/L S961 for 30 min followed by a 15-min stimulation with increasing concentrations (1, 10, or 100 nmol/L) of insulin. C: The effect of IRAB-B and S961 on insulin-induced IR and Akt phosphorylation was tested over time. A 15-min 10 nmol/L insulin stimulation was performed on HuH7 cells that were treated with 100 nmol/L IRAB-B or 1 nmol/L S961 for 6 h, 3 h, 60 min, 30 min, or 0 min (cotreated with insulin). Basal phosphorylation levels and IgG isotype–treated cells were used as controls. p, phosphorylated.
IRAB-B and S961 display distinct antagonism of IR signaling in vitro. Insulin-induced IR Tyr1162/1163 and Akt Ser473 phosphorylation levels were measured by Western blot in HuH7 cells treated with IRAB-B or S961. A: Cells were treated with increasing concentrations (1, 10, or 100 nmol/L) of IRAB-B or S961 for 30 min and stimulated with 1 nmol/L insulin for 15 min. B: Cells were treated with 100 nmol/L IRAB-B or 1 nmol/L S961 for 30 min followed by a 15-min stimulation with increasing concentrations (1, 10, or 100 nmol/L) of insulin. C: The effect of IRAB-B and S961 on insulin-induced IR and Akt phosphorylation was tested over time. A 15-min 10 nmol/L insulin stimulation was performed on HuH7 cells that were treated with 100 nmol/L IRAB-B or 1 nmol/L S961 for 6 h, 3 h, 60 min, 30 min, or 0 min (cotreated with insulin). Basal phosphorylation levels and IgG isotype–treated cells were used as controls. p, phosphorylated.
Having observed distinct antagonism of insulin signaling in vitro, we compared the impact of IRAB-B and S961 on glucose and insulin levels in normal C57 mice. Because the IRAB-B mAb and S961 peptide display different pharmacokinetic properties, measurements were taken 24 h postdose for IRAB-B and 1 h postdose for S961 to achieve similar degrees of hyperglycemia. Twenty-four hours after treatment with 10 mg/kg IRAB-B and 1 h after treatment with 1.0 mg/kg S961 resulted in similar blood glucose levels >450 mg/dL (Fig. 7A). Insulin levels for IRAB-B–treated mice were significantly higher (>50 ng/dL) compared with the S961 cohort (∼30 ng/dL) (Fig. 7B). Thus, IRAB-B induces more severe hyperinsulinemia than S961.
Comparison of IRAB-B and S961 by blood glucose and insulin levels in vivo. Food was removed for 4 h each morning before IRAB-B (10 mg/kg s.c.) or the IgG isotype (10 mg/kg s.c.) were administered 24 h before blood collection, and PBS or S961 (0.25, 0.5, and 1 mg/kg s.c.) were administered 1 h before sampling. A and B: Blood glucose and plasma insulin levels were then measured. The mice were grouped based on average glucose (1°) and body weight (2°), and data are mean ± SEM (n = 8) for each experimental group except the 1 mg/kg S961 cohort, where n = 7. Statistical comparisons were performed with one-way ANOVA and Dunnett test for S961 vs. vehicle (four groups) or unpaired t test for IRAB-B vs. IgG isotype (two groups) using GraphPad Prism 6 software. Statistical comparisons between 10 mg/kg IRAB-B and 1.0 mg/kg S961 (two groups) were performed using two-tailed unpaired t test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Comparison of IRAB-B and S961 by blood glucose and insulin levels in vivo. Food was removed for 4 h each morning before IRAB-B (10 mg/kg s.c.) or the IgG isotype (10 mg/kg s.c.) were administered 24 h before blood collection, and PBS or S961 (0.25, 0.5, and 1 mg/kg s.c.) were administered 1 h before sampling. A and B: Blood glucose and plasma insulin levels were then measured. The mice were grouped based on average glucose (1°) and body weight (2°), and data are mean ± SEM (n = 8) for each experimental group except the 1 mg/kg S961 cohort, where n = 7. Statistical comparisons were performed with one-way ANOVA and Dunnett test for S961 vs. vehicle (four groups) or unpaired t test for IRAB-B vs. IgG isotype (two groups) using GraphPad Prism 6 software. Statistical comparisons between 10 mg/kg IRAB-B and 1.0 mg/kg S961 (two groups) were performed using two-tailed unpaired t test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Discussion
We present a specific allosteric mAb IR antagonist, IRAB-B, that generates inducible models of acute and prolonged insulin resistance in mammalian systems. IRAB-B blocked insulin signaling within minutes in vitro and induced severe hyperglycemia as rapidly as 6 h in vivo. Of note, a single dose of IRAB-B also exhibited prolonged antagonist effects because significant decreases in insulin-induced IR phosphorylation were observed for at least 3 days in cells and severe hyperglycemia >575 mg/mL was measured in mice 2 weeks postinjection. We show that IRAB-B–induced hyperglycemia can be therapeutically treated by using the GLP-1 receptor agonist exendin-4. Finally, we compared IRAB-B with another routinely used IR antagonist, S961, to demonstrate differential pharmacology of IR signaling and the advantage of IRAB-B in generating a more-robust model of insulin resistance.
Corbin and colleagues (34,35) identified another allosteric IR mAb antagonist, XMetD, which appears to be different from IRAB-B both in vitro and in vivo. In vitro, XMetD did not affect insulin-induced Akt phosphorylation at mAb levels of 13 nmol/L, whereas at least 10 nmol/L IRAB-B results in clear decreases in Akt phosphorylation (Figs. 4B and 6A) (34). In normal mice under fasting conditions, we show that a 10 mg/kg dose of IRAB-B induces blood glucose levels >500 mg/dL and insulin levels of 50 ng/mL (Fig. 7A and B). Corbin et al. (34) reported that 10 mg/kg XMetD results in no significant change in fasted blood glucose and in fasted insulin levels of 12 ng/mL. What is driving these differences between IRAB-B and XMetD is unclear but might be partly due to different epitope engagement because both mAbs display low nanomolar binding affinity (IRAB-B Kd ∼3 nmol/L and XMetD Kd 8 nmol/L). From our comparisons, IRAB-B is more potent than XMetD, thus providing an advantage of IRAB-B for generating more severe models of insulin resistance.
IRAB-B demonstrates both rapid and sustained antagonism of the IR, blocking insulin activation within minutes and lasting for days with a single treatment in vitro. We showed a 10-fold EC50 shift in the insulin concentration-response curve, and IRAB-B treatment resulted in significantly decreased IR phosphorylation up to 3 days after treatment without affecting cellular receptor levels (Supplementary Table 1 and Figs. 2B and 3C and D). Specific changes regarding the effect of IRAB-B on receptor turnover would require extensive investigation and would be of considerable interest for future studies. IRAB-B antagonism was further confirmed with decreased downstream signaling on IRS-1, Akt, and glucose uptake (Fig. 4). Of note, insulin resistance in humans has been associated with decreased IR, IRS-1, and Akt phosphorylation (36–39). Therefore, using IRAB-B to generate models of insulin resistance correlates with impaired signaling events relevant to those observed in humans. Overall, we propose that IRAB-B is a valuable tool that specifically antagonizes the IR to study immediate and prolonged effects of impaired insulin signaling events relevant to those observed in humans.
IRAB-B rapidly induces severe hyperglycemia and is sustained for >2 weeks, allowing for the study of both acute and chronic effects in vivo. Of note, only the 25 mg/kg dose of IRAB-B induced elevated blood glucose levels within the 8-h period, whereas hyperglycemia was observed with the 5 and 10 mg/kg doses with longer exposure times (Fig. 5C, F, and G). This may be due to a receptor occupancy threshold that must be exceeded to induce hyperglycemia. Although body composition was not measured, we show that IRAB-B–treated mice exhibited decreased body weight likely attributable to downregulation of multiple anabolic processes (Fig. 5D and E).
Effective treatment of hyperglycemia with the GLP-1 receptor agonist exendin-4 demonstrates that IRAB-B–treated mice are suitable for testing novel therapeutics. Exendin-4 increases insulin levels that may partially overcome IR antagonism induced by IRAB-B. Alternatively, GLP-1 receptor agonists have been reported to have insulin-independent effects, which may partially account for exendin-4 glucose lowering effects in the IRAB-B model (40–42). Of note, exendin-4 lowered blood glucose in day 5 IRAB-B–treated mice but not in day 3 IRAB-B–treated mice (Fig. 5F and G). The difference could be due to IRAB-B tissue engagement, biodistribution, or compensatory mechanisms, resulting in altered IR binding and glucose homeostasis between these time points (Supplementary Fig. 2). Because the IR is highly conserved, IRAB-B can likely be adapted to any mammalian system, including primates.
IRAB-B displays attractive properties in vitro and in vivo distinct from S961 (Fig. 7A and B). Whereas IRAB-B has the advantageous pharmacokinetic properties of a mAb with a single dose lasting weeks, S961 requires repeat administration to maintain hyperglycemia (28–30,43,44). Taking this with the mixed agonist/antagonist activity, animals will exhibit variable responses with S961, whereas IRAB-B is an absolute and potent antagonist. The pharmacological differences observed with IRAB-B and S961 may be due to different binding sites on the IR. S961 has been reported to compete with insulin for the orthosteric site, whereas IRAB-B binds in the presence of insulin. Future structural studies and IRAB-B epitope identification are needed to determine how allosteric binding to the extracellular domain negatively modulates intracellular kinase activity of the IR. Using tools like IRAB-B could help to elucidate the unknown structure-function relationship of IR activation.
In conclusion, we present the novel, potent IR antagonist IRAB-B that rapidly and chronically antagonizes the IR in vitro and in vivo. The IRAB-B model offers advantages over current reagents by specifically antagonizing IR signaling for the study of the effects of acute and chronic hyperglycemia, hyperinsulinemia, and insulin resistance. We propose that IRAB-B is a powerful tool that can be adapted to multiple mammalian models to provide insight into the pathogenesis of impaired insulin signaling and to test therapeutics that effectively lower blood glucose and insulin levels.
Article Information
Acknowledgments. The authors thank John Mabus, Nathaniel Wallace, Fuyong Du, Cuifen Hou, and Pamela Hornby (all of Janssen Pharmaceutical Companies of Johnson & Johnson) for expertise and insight.
Duality of Interest. For the duration of this study, all authors were employed by Janssen Pharmaceutical Companies of Johnson & Johnson. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. A.M.C., T.K., S.A.H., R.N., K.D., K.B., P.R.C., R.P., and S.J. performed the research. A.M.C., T.K., S.A.H., R.N., and K.D. analyzed the data. M.L.C., D.L.J., J.M.W., E.R.L., R.B.L., Y.L., and A.J.K. supervised the research. A.M.C. and A.J.K. wrote the manuscript. A.J.K. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Prior Presentation. This study was presented as a published late-breaking abstract poster presentation at the 76th Scientific Sessions of the American Diabetes Association, New Orleans, LA, 10–14 June 2016.