The Insulin-Like Growth Factor-1 Receptor Is a Negative Regulator of Nitric Oxide Bioavailability and Insulin Sensitivity in the Endothelium

We used four different gene-modified mice to examine the role of IGF-1R in EC insulin sensitivity and NO production. All mice were bred onto a C57BL/6J background to at least 10 generations in a conventional animal facility with a 12-h light/dark cycle. Male mice aged 3 months were used. All experiments were conducted in accordance with accepted standards of humane animal care under U.K. Home Office Project license no. 40/2988.

Mice with whole-body haploinsufficiency of IGF-1R (16) were obtained from The European Mutant Mouse Archive (Munich, Germany; http://www.emmanet.org) and maintained in the heterozygous state on a C57BL/6J background by mating IGF-1R^+/− males with 9–12-week-old C57BL/6J female wild-type mice.

To examine the effect of changing IGF-1R:IR stoichiometry on endothelial responses, we bred male IGF-1R^+/− mice with female IR^+/− mice to generate mice with whole-body haploinsufficiency of IR and IGF-1R (IR^+/−/IGF-1R^+/−).

To generate mice with endothelium-specific deficiency of IGF-1R, we used mice carrying a floxed IGF-1R allele (IGF-1R^Lox), as previously described (17), and mice expressing the cre recombinase under control of the Tie2 promoter (Tie-2 cre; Jackson Laboratories, Bar Harbor, ME). A mating program of male Tie-2-cre mice (to avoid the possibility of inheritance of a null allele in the second generation, which is not tissue specific) with female IGF-1R^Lox was used. Male heterozygous and homozygous Tie-2-cre/IGF-1R^Lox mice were compared with age-matched littermate controls.

Blood was sampled from tail vein. Glucose, insulin, and IGF-1 tolerance tests were performed by blood sampling after an intraperitoneal injection of glucose (1 mg/g; Sigma-Aldrich, Dorset, U.K.), human recombinant insulin (0.75 units/kg, Actrapid; Novo Nordisk, Bagsvaerd, Denmark), or human recombinant IGF-1 (0.75 μg/g; GroPep, Adelaide, Australia), as previously described (18). Glucose concentrations were determined in whole blood by a portable meter (Roche Diagnostics, West Sussex, U.K.). Plasma insulin concentrations were determined by enzyme-linked immunoassay (Ultrasensitive mouse ELISA; CrystalChem, Downers Grove, IL). Plasma IGF-1 levels were determined by enzyme-linked immunoassay (19).

Vasomotor function was assessed in aortic rings as previously described (6–8). Rings were mounted in an organ bath containing Krebs Henseleit buffer and equilibrated at a resting tension of 3 g for 45 min before the experiments. A cumulative dose response to the constrictor phenylephrine (PE) (1 nmol/L to 10 μmol/L) was first performed. The vascular effect of IGF-1 was assessed by constructing PE dose–response curves before and 2 h after preincubation with IGF-1 (100 nmol/L) (19). Vasorelaxation in response to insulin was assessed by constructing PE dose–response curves before and after 2 h preincubation with insulin (100 mU/mL). Relaxation responses to cumulative addition of acetylcholine (1 nmol/L–10 μmol/L) and sodium nitroprusside (0.1 nmol/L–1 μmol/L) were performed and responses expressed as percent decrement in preconstricted tension. Basal NO bioavailability in response to isometric tension was measured by recording the increase in tension elicited by l-N^G-nitro-l-arginine methyl ester (l-NMMA; 0.1 mmol/L) in aortic segments maximally preconstricted with PE.

Endothelial NO synthase (eNOS) activity was determined by conversion of [¹⁴C]l-arginine to [¹⁴C]l-citrulline as we previously described (19,20). Aortic rings (5 mm in length) or ECs (1 × 10⁶) were incubated at 37°C for 20 min in HEPES buffer, pH 7.4 (in mmol/L): 10 HEPES, 145 NaCl, 5 KCl, 1 MgSO₄, 10 glucose, 1.5 CaCl₂, containing 0.25% BSA. 0.5 μCi/mL l-[¹⁴C]arginine was then added for 5 min and tissues were stimulated with insulin (100 nmol/L) or IGF-1 (100 nmol/L) for 30 min before the reaction was stopped with cold PBS, containing 5 mmol/L l-arginine and 4 mmol/L EDTA, after which, tissue was denatured in 95% ethanol. After evaporation, the soluble cellular components were dissolved in 20 mmol/L HEPES-Na⁺ (pH 5.5) and applied to a well equilibrated DOWEX (Na⁺ form) column. The l-[¹⁴C]citrulline content of the eluate was quantified by liquid scintillation counting and normalized against the weight of tissue used or total cellular protein.

eNOS, peNOS, IR, and IGF-1R protein expression was quantified in aorta and ECs using Western blotting (19). Immunoprecipitation of protein for the quantification of hybrid receptors was achieved by incubating equal amounts of protein lysates (100 µg) and 50 µL protein A DynaBeads (Invitrogen, Paisley, U.K.), precoated with indicated anti-rabbit antibodies for 20 min. After three rounds of washing with PBS–0.01% Tween 20, the beads were resuspended in 2× Laemmli loading buffer supplied with the NOVEX gel electrophoresis system. Aortic protein was separated by SDS-PAGE and blots were probed with mouse anti–IGF-1R-α (for hybrid expression), IR-β (C-19), and IGF-1R-β (C-20) antibodies for receptor expression (Santa Cruz Biotechnologies, Santa Cruz, CA).

Primary cells were lysed in extraction buffer containing (in mmol/L, unless otherwise specified) 50 HEPES, 120 NaCl, 1 MgCl₂, 1 CaCl₂, 10 NaP₂O₇, 20 NaF, 1 EDTA, 10% glycerol, 1% NP40, 2 sodium orthovanadate, 0.5 μg/mL leupeptin, 0.2 phenylmethylsulfonyl fluoride, and 0.5 μg/mL aprotinin. Cell extracts were sonicated in an ice bath and centrifuged for 15 min, before protein measurements were carried out by biocinochinic acid assay (Pierce; Thermo Scientific, Rockford, IL) using the supernatant. Equal amounts of cellular protein were resolved on SDS-polyacrylamide gels (Invitrogen) and transferred to polyvinylidine difluoride membranes. Immunoblotting was carried out with indicated primary antibodies. Blots were incubated with appropriate peroxidize-conjugated secondary antibodies and developed with enhanced chemiluminescence (Millipore, Billerica, MA). Fifty micrograms of total cell lysate was used for immunoprecipitation with indicated antibodies: IRβ (C19; Santa Cruz Biotechnology), phophotyrosine (pY4G10; Millipore), and protein A Dynabeads (Invitrogen), for 20 min at room temperature. Immune complexes were collected using magnetic field, washed extensively with PBS–0.02% Tween-20, solubilized in Laemmli sample buffer, and analyzed as above. Immunoblots were scanned on a Kodak (Hemel Hempstead, U.K.) camera-scanner Image Station 2000R and bands were quantified using Kodak ID Image Analysis software.

mRNA was isolated using a commercial kit (Roche, Indianapolis, IN), and the levels of IGF-1R and IR mRNA quantified using real-time quantitative PCR (e.g., 6–8).

Primary ECs were isolated from lungs by immunoselection with CD146 antibody–coated magnetic beads (Miltenyi Biotech, Surrey, U.K.), as previously reported (21). Sheep anti–rat IgG magnetic beads were coated with monoclonal rat anti-CD31 antibody (BD Biosciences, Oxford, U.K.) according to the manufacturer’s instructions. Lungs were harvested, washed, finely minced, and digested in Hanks’ balanced salt solution containing 0.18 units/mL collagenase (10 mg/mL; Roche) for 45 min at 37°C. The digested tissue was filtered through a 70-μm cell strainer and centrifuged at 1,000 rpm for 10 min. The cell pellet was washed with DMEM/10% FCS, centrifuged, resuspended in 1 mL DMEM/10% FCS, and incubated with 1 × 10⁶ CD31 antibody–coated beads at 4°C for 30 min. Bead-bound cells were separated from non–bead-bound cells using a magnet. Bead-bound (CD31 positive) and non–bead-bound cells were resuspended in 2 mL EGM-2-MV (Cambrex, East Rutherford, NJ), supplemented with human epidermal growth factor, hydrocortisone, vascular endothelial growth factor (VEGF), human fibroblast growth factor-basic, R3-IGF-1, ascorbic acid, gentamicin, amphotericin-B, and 5% FCS, and plated out. Only the EC population tested positive for a range of endothelial markers, including eNOS, Tie2, and CD102 protein, measured using immunoblotting.

The following validated oligonucleotides were synthesized from Sigma-Aldrich (SASI_Hs01_00126194 HUMAN NM_000875). Human umbilical vein endothelial cells (HUVECs) were grown in endothelial growth medium. Cells with 70% confluency were transfected with 10 nmol/L siRNA using the NTER transfection reagent (Sigma-Aldrich) according to the manufacturer's protocol, optimized for use in HUVECs by the manufacturer. In brief, a mixture of target and scrambled siRNA (10 nmol/L), diluted in siRNA buffer and NTER transfection reagent, was preincubated at 37°C for 20 min. Growth medium was then removed and replaced with culture medium containing the target or scrambled siRNA mixture and further incubated at 37°C for 4 h. Four hours post-transfection, the medium was replaced with 0.2% serum containing medium, and HUVECs were incubated overnight. The cells were then stimulated with 100 nmol/L insulin for 15 min at 37°C, after which the cells were harvested for protein analysis and Western blotting to analyze phosphorylation levels of eNOS. Assessment of eNOS activity was carried out on transfected and serum-starved HUVECs, as described above. After the overnight serum starvation, the medium was replaced with HEPES buffer, pH 7.4, for 30 min for equilibration.

Insulin (100 nmol/L)-mediated NO release was measured as previously described (22). In brief, pulmonary endothelial cells (PECs) cultured in 96-well plates (90% confluent) in SBS (in mmol/L: 135 NaCl, 5 KCl, 1.2 MgCl₂, 8 glucose, 10 HEPES, and 1.5 CaCl₂; 290 mOsm osmolarity; pH 7.4, adjusted using 4 mol/L NaOH) were loaded with DAF (DAF-FM acetate [Invitrogen, D-23844], stock 5 mmol/L in DMSO; dilute DAF-FM acetate in SBS 2:1,000 to obtain 10 μmol/L DAF) in SBS by incubating at room temperature for 45 min. DAF was then removed and cells were washed in SBS twice. Cells were then left in SBS for 20 min at room temperature. SBS was replaced again and the plate was loaded into Flex station (FlexStation; Molecular Devices, Sunnyvale, CA). Fluorescence emission at 520 nm was measured with 495-nm excitation using the multiwell fluorescence plate reader system.

Results are expressed as mean (SEM). Comparisons within groups were made using paired Student t tests and between groups using unpaired Student t tests or repeated-measures ANOVA, as appropriate; where repeated t tests were performed, a Bonferroni correction was applied. P < 0.05 was considered statistically significant.

IGF-1R^+/− mice had similar fasting glucose concentrations (5.4 mmol/L [0.3] vs. 4.9 mmol/L [0.1]), and as previously reported, were less efficient at dealing with a glucose bolus (Fig. 1A), with exaggerated postbolus hyperglycemia compared with wild-type mice. Despite this glucose intolerance, IGF-1R^+/− mice had an exaggerated response to insulin, indicative of enhanced whole-body insulin sensitivity (Fig. 1B). IGF-1R^+/− mice were resistant to IGF-1–mediated glucose lowering (Fig. 1C). Fasting plasma insulin and IGF-1 were similar in IGF-1R^+/− and wild-type mice (Fig. 1D and E). There was no difference in vascular responses of IGF-1R^+/− and wild-type mice to acetylcholine or sodium nitroprusside (data not shown). The vasorelaxant response to IGF-1 in IGF-1R^+/− mice was significantly blunted (Fig. 2A and B), and IGF-1–mediated activation of eNOS, measured using the conversion of l-arginine to l-citrulline, was also blunted (Fig. 2C), demonstrating that IGF-1–mediated vascular activation of eNOS is exquisitely sensitive to a reduction in IGF-1R. Consistent with enhanced basal NO bioavailability, the constrictor response to the nonspecific NO synthase inhibitor l-NMMA was increased in IGF-1R^+/− mice (Fig. 2D–F), and constrictor responses to PE were blunted in IGF-1R^+/− mice (Fig. 2G and H). To probe the effect of haploinsufficiency of the IGF-1R in ECs, we studied responses in PECs. As expected, PECs from IGF-1R^+/− mice had substantially lower expression of IGF-1R (Fig. 2I). We examined the effect of IGF-1R haploinsufficiency on insulin sensitivity in PECs using a number of approaches. In PECs from IGF-1R^+/− there was a significantly higher level of tyrosine-phosphorylated IR (Fig. 3A). Consistent with enhanced insulin sensitivity of PECs from IGF-1R^+/−, insulin-stimulated NO release in PECs from IGF-1R^+/− mice was enhanced (Fig. 3B). PECs from IGF-1R^+/− mice were also significantly more sensitive to insulin-mediated serine phosphorylation of eNOS (Fig. 3C). In aortas from IGF-1R^+/− mice, there was, as expected, significantly less IGF-1R protein expression than wild-type mice. There was no difference in IR, but a significant reduction in hybrid receptors, in keeping with an increase in the relative level of free IR homodimers (Fig. 3D–F).

As previously reported, mice homozygous for IGF-1R deletion were a knockdown model rather than complete deletion (17). “Holoinsufficient” mice, therefore, have some residual IGF-1R in the endothelium (Supplementary Fig. 1). EC-specific IGF-1R–deficient mice had blunting of IGF-1R–mediated aortic relaxation (Supplementary Fig. 2). Unlike mice with global IGF-1R deficiency, mice with haploinsufficiency and holoinsufficiency of the IGF-1R in the endothelium had similar responses in insulin and glucose tolerance tests as wild-type mice (Fig. 4A–D), and similar fasting glucose and insulin concentrations (Fig. 4E and F). Despite this, and consistent with our findings in whole-body IGF-1R^+/− mice, endothelium-specific IGF-1R^+/− and IGF-1R^−/− mice had blunted constrictor responses to PE and an exaggerated constrictor response to l-NMMA, indicative of increased basal NO bioavailability (Fig. 5A–F). When comparing wild-type, endothelium-specific IGF-1R^+/− and IGF-1R^−/− mice, there was a clear gene dose effect on these vascular responses (Figs. 5G and H). Consistent with our findings in mice with whole-body haploinsufficiency of the IGF-1R hybrid receptor, expression was significantly lower in PEC from endothelium-specific IGF-1R–deficient mice (Fig. 5I).

We have previously demonstrated that IR^+/− mice, while maintaining glucose competence, have significant blunting of insulin-mediated vasorelaxation and eNOS activation (6,20). To test our hypothesis that changing IR:IGF-1R stoichiometry would impact favorably on insulin-mediated NO production in vivo, we generated mice with whole-body haploinsufficiency of the IR and IGF-1R. IGF-1R^+/−/IR^+/− and IR^+/− mice had similar fasting glucose concentrations and responses to a glucose bolus (Fig. 6A). IGF-1R^+/−/IR^+/− mice had higher fasting glucose concentrations than IGF-1R^+/− and wild-type mice (Supplementary Fig. 3A). In insulin tolerance tests, IGF-1R^+/−/IR^+/− and IR^+/− mice had similar responses (Fig. 6B and Supplementary Fig. 3B). IGF-1R^+/−/IR^+/− mice had increased insulin-stimulated eNOS activation in aortic rings (Fig. 6C), restored aortic relaxation in response to insulin (Fig. 6D–F), and increased NO generation in PEC in response to insulin (Fig. 6G and H), and reduced numbers of hybrid receptors (Fig. 6I).

To examine the specific effects of reducing IGF-1R expression on basal serine phosphorylation and activation of eNOS in response to insulin, we used siRNA-mediated knockdown of IGF-1R. siRNA-treated HUVEC had a significant reduction in IGF-1R but no change in IR (Fig. 7A–C), assessed using RT-PCR. This reduction in IGF-1R led to a significant decrease in hybrid receptors (Fig. 7D). In keeping with our in vivo findings, knocking down IGF-1R led to an increase in basal and insulin-stimulated serine-phosphorylated eNOS (Fig. 7E), and increased insulin-mediated eNOS activation (Fig. 7F).

The present report using a comprehensive in vivo and in vitro approach describes a novel mechanism by which the bioavailability of the endothelium-derived signaling radical NO is regulated. We have demonstrated using mice with whole-body haploinsufficiency of IGF-1R and endothelium-specific deletion of IGF-1R that NO bioavailability is negatively regulated by IGF-1R in the endothelium.

To examine the therapeutic potential of manipulating IGF-1R numbers, we generated mice with whole-body haploinsufficiency of the IR and IGF-1R. We have previously demonstrated substantial abnormalities of EC function in IR^+/− mice (6,7,20). Reducing IGF-1R numbers in IR^+/− mice restored insulin-mediated vasorelaxation, enhanced insulin-stimulated eNOS activation, and enhanced insulin-stimulated NO release in ECs. In vitro experiments in human ECs confirmed that IGF-1R deletion enhances basal and insulin-stimulated serine phosphorylation of eNOS and eNOS activity in response to insulin. These data demonstrate for the first time that IGF-1R is a key negative regulator of insulin sensitivity and NO bioavailability in the endothelium.

A well defined series of steps occur in the arterial wall during the development of atherosclerosis (23). Early in the development of atherosclerosis, before the onset of morphological changes, there are important alterations in EC phenotype, a situation described as endothelial dysfunction. A key feature of endothelial dysfunction is a reduction in NO bioavailability (24). In the endothelium, NO is generated by l-arginine to l-citrulline conversion principally by the NO synthase isoform eNOS (for review see Ref. 25). NO can influence multiple steps in the atherosclerotic process, including vascular smooth muscle cell phenotype (26), vascular inflammation (27), platelet aggregation (28), EC migration, proliferation and survival (29), and progenitor cell mobilization (30). In addition, NO has physiologically important antioxidant actions (31).

eNOS activity is regulated at multiple different levels (for review see 25), including the availability of its substrate or cofactors, protein–protein interactions, posttranslational modifications, and serine phosphorylation. Serine phosphorylation of eNOS accelerates the rate of electron flow through the enzyme, increases its calcium sensitivity, and augments NO biosynthesis and release (32).

Insulin and IGF-1, via activation of their receptors, have been shown to increase levels of serine-phosphorylated eNOS in vivo (19), and in vitro (33), and to stimulate NO production (34). In humans, insulin sensitivity closely correlates with NO bioavailability in resistance vessels (35,36). In a novel murine model of endothelium-specific insulin resistance, we demonstrated a significant reduction in insulin-mediated eNOS phosphorylation and NO bioavailability (8). In other preclinical models, eNOS phosphorylation status has been shown to play an important and favorable role in a range of pathological situations reminiscent of human vascular disease such as stroke (37) and lower limb ischemia (38).

In this mouse, previously shown to be protected against oxidative stress (16), and age-related inflammatory degenerative brain disease (39), we demonstrated an intriguing phenotype. Using a range of complementary approaches we showed that despite whole-body glucose intolerance, haploinsufficiency of IGF-1R leads to the favorable vascular phenotype of reduced vascoconstriction to PE, increased basal NO production, and increased EC insulin sensitivity leading to augmented insulin-mediated NO generation.

IGF-1R^+/−mice are glucose intolerant after a glucose bolus but have a greater fall in plasma glucose in response to a systemic bolus of insulin. The present report and studies in mice with pancreatic β cell–specific deletion of IGF-1R help explain this apparently divergent metabolic phenotype. Mice with pancreatic β cell–specific deletion of IGF-1R are thought to have impaired insulin secretion/glucose sensing (40), accounting for the abnormal glucose tolerance. Our dataset provide an explanation for the increased insulin sensitivity, i.e., removal of the negative regulatory effect of IGF-1R on signaling through IR.

To separate the systemic effects of IGF-1R from the EC effects, we generated mice with different levels of IGF-1R using the Tie-2 promoter to direct gene expression principally to the endothelium. Mice with Tie-2–directed deletion of IGF-1R had glucose and insulin tolerance tests similar to wild-type mice, but reduced aortic constriction to PE and increased basal NO. Consistent with a role for IGF-1R numbers as the principal mechanism underlying this observation, we demonstrated a gene dose effect of IGF-1R on PE-induced constriction and basal NO production.

It is well established that IR and IGF-1R can heterodimerize to form insulin-resistant hybrid receptors (10–14). The formation of hybrid receptors is determined by the molar proportion of receptors. In ECs, the specific binding of insulin is significantly lower than IGF-1, indicative of a lower number of homodimeric IR than IGF-1R (41); studies suggest that IGF-1R outnumber IR tenfold (34). Crucially, for high-affinity binding of insulin to the IR and activation to occur, binding to both subunits is necessary, whereas IGF-1 requires only one subunit (13). This provides one explanation as to why hybrids have a low affinity for insulin but bind IGF-1 with the same high affinity as IGF-1R homodimers. In vitro studies of a range of cell types (42–45) have suggested that reduction of IGF-1R density enhances insulin sensitivity. In cells from humans with polymorphisms that lead to a reduction in IGF-1R density, a similar pattern of enhanced insulin sensitivity has been demonstrated (46).

Insulin-resistant humans have a significant reduction in cell surface IR (14) and reduced insulin-mediated, NO–dependent vasorelaxation (47). We have previously used mice with whole-body haploinsufficiency of the IR as a model of non-obese whole-body insulin resistance. These mice, similar to insulin-resistant humans, have reduced insulin-mediated eNOS phosphorylation, reduced insulin-mediated eNOS activation, and reduced insulin-mediated vasorelaxation (6,7). More recently, consistent with reduced NO bioavailability, we have shown that these mice have reduced endothelial progenitor cell mobilization in response to VEGF and delayed endothelium regeneration after femoral artery wire injury (20). To examine whether reducing IGF-1R in IR^+/− mice can restore insulin-mediated NO production, we generated IGF-1R^+/−/IR^+/− mice. IGF-1R^+/−/IR^+/− mice had enhanced insulin-mediated NO release in PEC and enhanced insulin-mediated vasorelaxation and eNOS activation in aortas compared with IR^+/−. In complementary studies in HUVECs, we knocked down IGF-1R. HUVEC deficiency in IGF-1R, in keeping with our findings in vivo, had increased basal serine-phosphorylated eNOS and enhanced insulin-mediated serine eNOS phosphorylation and activation.

An increase in hybrid receptors is a hallmark of type 2 diabetes and obesity (14,48,49). Elegant studies from Sesti et al. provide important insights into how aspects of the type 2 diabetes phenotype may regulate the formation of hybrids. In a study of patients with primary, nongenetically determined hyperinsulinaemia (i.e., patients with insulinoma), Federici et al. (50) demonstrated that insulin-mediated downregulation of the IR, by shifting IR:IGF-1R stoichiometry in favor of IGF-1R led to a higher proportion of hybrid receptors in skeletal muscle in these patients. This suggests that hybrid receptor formation in patients with type 2 diabetes may be driven by changes in insulin concentration.

The role of hyperglycemia was also examined by this group (51). In diabetic rats rendered normoglycemic, Federici et al. showed that reducing glucose reduced hybrid receptor numbers. Moreover, the glucosamine pathway, at least in part, may mediate glucose-induced formation of hybrid receptors. These studies demonstrate that by driving the formation of hybrid receptors, different components of the type 2 diabetes phenotype may initiate a vicious cycle, leading to worsening of insulin sensitivity.

In summary we have demonstrated in complementary in vivo and in vitro models that IGF-1R is an important negative regulator of insulin sensitivity in the endothelium. Deletion of IGF-1R leads to a potentially favorable increase in NO bioavailability and, in a well characterized model of human insulin resistance, rescues the adverse phenotype. Our dataset raises the exciting possibility that IGF-1R may be a novel target for the treatment of insulin resistance–related vascular disease.

This work was supported by the British Heart Foundation and Medical Research Council UK.

No potential conflicts of interest relevant to this article were reported.

A.A., H.I., H.V., P.S., J.S., S.G., and N.Y. performed experiments and analyzed data. A.R., R.M.C., M.G., J.S., S.G., and M.K. assisted with experiments. S.X. supplied IGF-1R floxed mouse, designed experiments, and critically reviewed the manuscript. M.T.K. wrote the manuscript. P.J.G., K.M.C., D.J.B., and S.B.W. contributed to discussion and reviewed and edited the manuscript. K.M.C., D.J.B., S.B.W., and M.T.K. obtained funding and designed experiments.