Novel Role of the IGF-1 Receptor in Endothelial Function and Repair: Studies in Endothelium-Targeted IGF-1 Receptor Transgenic Mice

To overcome the limitations of random and multiple copy insertion sites seen in standard transgenics, we used the hypoxanthine phosphoribosyl transferase (Hprt) targeting system (Genoway; see ref. 16) to generate genetically modified embryonic stem (ES) cells. This approach uses homologous recombination to target a single copy of a transgene (in this case, the human IGF-1r), driven by a promoter (in this case, the Tie2 promoter), into the Hprt locus on the X chromosome. The model was developed with E15Tg2a (E14) cells derived from the strain 129P2/OiaHsd (12901a). In E14 cells, 35 kb of the Hprt gene encompassing the 5′ untranslated region up to intron 2 is deleted. The Hprt gene encodes a constitutively expressed housekeeping enzyme involved in the synthesis of purines from the degradation products of nucleotide bases (salvage pathway). Cells normally synthesize purines by the salvage and de novo pathways. In Hprt-deleted cell lines, only the de novo pathway is functional, enabling the cells to grow in classical medium. However, in the presence of the aminopterin drug, the de novo pathway is blocked. As a result, Hprt-deleted cells die in hypoxanthine-aminopterin-thymidine (HAT) media (containing HAT substrates). The targeted insertion of a transgenic cassette in E14 ES cells with a functional Hprt gene rescues these cells, which can then be selected using HAT media to identify ES cells showing the correct targeting event. ES cells with the correct insertion can be selected by virtue of their expression of Hprt and their ability to grow in HAT medium. Numerous studies have shown that the functional properties of the Hprt locus protect transgenic constructs inserted in this region against gene silencing and positional or methylation effects. Furthermore, tissue-specific promoters including Tie2 (17) inserted into the Hprt locus maintain their expression properties.

Endothelium-specific transgene expression was achieved using the mouse Tie2 promoter and intronic enhancer as previously described (11,18). To reduce the size of the transgene, the minimal enhancer sequence (core enhancer) was inserted into the final construct. This sequence has been shown to confer a uniform and high level of expression of LacZ reporter gene in endothelial cells in vivo (17). The targeting vector was obtained by inserting human IGF-1R cDNA (a gift from Dr. R. Baserga, Jefferson University) (19) into the pHHNS plasmid (11,18), comprising the murine Tie2 promoter, LacZ cDNA, and SV40 polyA signal and a 10-kb intronic enhancer from the murine Tie2. The final transgenic vector (Fig. 1A), consisting of Tie2 promoter, human IGF-1R cDNA, SV40 polyadenylation site, and core enhancer, was then inserted into an Hprt targeting vector using standard cloning. The final targeting vector had the following features: 1) homology arms isogenic with E14 ES cells favoring homologous recombination; 2) symmetrical homology arms (5′ short arms: 3.8 kb, 3′ long arm: 3.7 kb); 3) transgenic cassette expressing the human IGF-1R cDNA under control of the short form of the Tie2 promoter; and 4) wild-type Hprt sequences to reconstitute the Hprt gene in the E14 ES cells. In ES cells, this process was highly successful, with three clones selected for blastocyst injection. The 5′ and 3′ targeting events were unambiguously confirmed by Southern blot analysis. These three ES clones were expanded and recombinant ES cells injected into C57BL/6J-derived blastocysts that were then implanted into the uteri of recipient females.

Seven highly chimeric males generated by blastocyst injection of ES clones were mated with two wild-type C57BL/6J female mice to examine whether the recombinant ES cells contributed to the germ layer. To assess whether ES cells have contributed to the germ layer of chimeras, mouse coat color markers were used. The ES cells used to develop the model were originally derived from a 129 strain of mice that have an agouti coat color. This marker is dominant over the black coat of C57BL/6 mice. Therefore, mating of the chimeras with C57BL/6J mice should yield agouti-colored pups when the ES cells have contributed to the germ layer. Eight out of 20 agouti F1 females were genotyped by Southern blot. Southern blot validated the correct heterozygous status of eight tested F1 females by detecting the 7.9-kb–sized AvrII fragment of the C57BL/6J Hprt wild-type allele and the 9.8-kb–sized AvrII fragment of the reconstituted Hprt allele (Fig. 1B).

To generate mice with endothelium-specific deficiency of the IGF-1R, we used mice carrying a floxed IGF-1R allele (IGF-1RLox) as previously described (15) and mice expressing the cre-recombinase under control of the Tie2 promoter (Tie2 cre; The Jackson Laboratory, Bar Harbor, ME). Male heterozygous Tie2-cre/IGF-1RLox mice were compared with age-matched littermate controls.

hIGFREO mice and mice with endothelium-specific haploinsufficiency of the IGF-1R (ECIGF-1R^+/−) were bred onto a C57BL/6J background to at least 12 generations. Mice were housed in a conventional animal facility with a 12-h light/dark cycle. Genotyping was performed using ear notch DNA (Fig. 1C). Male mice aged 3–5 months were used in all experiments, which were conducted in accordance with accepted standards of humane animal care under United Kingdom Home Office Project license number 40/3523.

Glucose, insulin, and IGF-1 tolerance tests were performed by blood sampling after an intraperitoneal injection of glucose, human recombinant insulin, or IGF-1 as previously described (15,20,21). Glucose concentrations were determined in whole blood using a portable meter. Plasma insulin and IGF-1 levels were determined by enzyme-linked immunoassay (20,21).

Vasomotor function was assessed ex vivo in aortic rings as previously described (21).

After activation, endothelial NO synthase (eNOS) produces NO and l-citrulline in a stoichiometric reaction. The effect of insulin/IGF-1 on eNOS activity in endothelial cells was determined by conversion of [¹⁴C]-l-arginine to [¹⁴C]-l-citrulline (15,20).

Protein expression was quantified in endothelial cells (before and after 10-min exposure to 150 nmol/L insulin or IGF-1) using Western blotting with β-actin as control (15,20,21). Immunoprecipitation of protein for the quantification of hybrid receptors and tyrosine-phosphorylated IR was performed as previously reported (15), incubating equal amounts of protein lysates (100 µg) and 50 µL of protein A-DynaBeads (Invitrogen) 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 (Invitrogen). Aortic protein was separated by SDS-PAGE, and blots were probed with a mouse anti–IGF-1R-α (for hybrid expression), IR-β (C-19), and IGF-1R-β (C-20), antibodies for receptor expression (Santa Cruz Biotechnology), Akt, phosphorylated Akt (pAkt) (21) (Cell Signaling Technology, Beverley, MA), and neuronal NO synthase (nNOS; Transduction Laboratories, Franklin Lakes, NJ). Tyrosine-phosphorylated IR expression was examined in pulmonary endothelial cells (PECs) as we previously reported (15).

Primary cells were lysed in extraction buffer containing (in mmol/L unless otherwise noted): 50 HEPES, 120 NaCl, 1 MgCl₂, 1 CaCl₂, 10 NaP₂O₇, 20 NaF, 1 EDTA, 10% glycerol, 1% Nonidet P-40, 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 BCA assay (Pierce) using the supernatant. Equal amounts of cellular protein were resolved on NuPage 4–12% Bis-Tris gels (Invitrogen) and transferred to polyvinylidene difluoride membranes. Immunoblotting was carried out with indicated primary antibodies. Blots were incubated with appropriate peroxidase-conjugated secondary antibodies and developed with enhanced chemiluminescence (Millipore). Fifty micrograms of total cell lysate was used for immunoprecipitation with indicated antibodies: IR-β (C19; Santa Cruz Biotechnology) 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 camera-scanner Image Station 2000R, and bands were quantified using Kodak ID Image Analysis software (Kodak).

mRNA was isolated using a commercial kit (Roche) and the levels of IGF-1R, IR, nNOS, and eNOS mRNA quantified using real-time quantitative PCR (9,15); a range of housekeeping genes were screened for validity and stability prior to measurements (data not shown). Nonspecific primers and those specific for human and mouse IGF-1R were designed and validated in human umbilical vein endothelial cells and mouse pulmonary endothelial cells (PECs) (Supplementary Table 1).

Primary PECs were isolated from murine lung as we previously described (15).

Insulin- and IGF-1–mediated NO release was measured using 4-amino-5-methylamino-2-difluorofluorescein (DAF-FM) fluorescence (Invitrogen) as we recently described (15).

Femoral artery wire injury was performed as we previously reported (10). Mice were anesthetized at 5 days after wire injury and 50 mL of 5% Evans blue dye injected into the vena cava. The mice were perfused/fixed with formaldehyde before femoral arteries (injured and uninjured) were harvested. The vessels were opened longitudinally. The areas stained and unstained in blue were measured in the injured area 5 mm from the proximal suture and the percentage areas calculated using ImagePro Plus 6.0 software (Media Cybernetics, Bethesda, MD).

PECs were suspended in supplemented basal medium. Approximately 4 × 10⁴ cells in 500 μL MV2 medium (PromoCell, Heidelberg, Germany) were placed in the upper compartment of a Modified Boyden chamber with a polycarbonate membrane containing 8 μm pores. The lower compartment contained 750 μL MV2 and 50 ng/ml vascular endothelial growth factor (VEGF165); negative control wells contained basal MV2 medium (vehicle) alone. Wells were set up in triplicate and incubated in 5% CO₂ at 37°C for 24 h. Membranes were fixed in 70% ethanol at −20°C for 1 h before mechanical removal of cells adherent to the upper surface using a cotton swab. Hematoxylin and eosin stains were then applied (30 s each) to membranes that were then washed in water and mounted on glass slides. Migrant PECs were counted in 10 high-power fields (×200 magnification, area of 720 × 530 μm). Results expressed as mean cell number per high-power field.

Endothelial progenitor cells (EPC) were isolated from bone marrow and spleen as we previously reported (10).

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; when repeated t tests were performed, a Bonferroni correction was applied. A P value <0.05 was considered statistically significant.

We generated a novel hIGFREO mouse (Figs. 1–3). hIGFREO mice were born with the same frequency as wild-type mice. There was no difference in behavior, grooming, growth (Fig. 2A), or organ weight (Fig. 2B). We determined the levels of human and native (mouse) IGF-1R mRNA expression in organs with different proportions of endothelial cells (lung, aorta, and spleen). Human IGF-1R mRNA expression was approximately threefold higher in lung than in aorta from hIGFREO; P < 0.01 (Fig. 3A). In spleen, there was very little IGF-1R, commensurate with proportionately fewer endothelial cells. Human IGF-1R expression in hIGFREO mice tracked the pattern for native IGF-1R seen in wild-type mice (Fig. 3B). There was no difference in expression of mouse IGF-1R between hIGFREO and wild type (Fig. 3C). There was no hIGF-1R mRNA detectable in endothelial cells from wild-type mice or nonendothelial cells from hIGFREO mice (Fig. 3D); nonendothelial cells did not express the endothelial-specific marker vascular-endothelial-cadherin in contrast to endothelial cells (Fig. 3E). Western blot analysis confirmed that total IGF-1R protein levels were significantly increased in hIGFREO endothelial cells compared with wild type (Fig. 3F).

hIGFREO had similar responses to insulin, glucose, and IGF-1 tolerance tests as wild-type mice (Fig. 4A–C). Fasting glucose (4.4 [0.3] versus 4.5 [0.3] mmol/L; P = NS), insulin (hIGFREO 2.0 [0.3] versus wild-type 2.1 [0.3] ng/ml; P = NS), and IGF-1 (hIGFREO 474 [31] versus wild-type 471 [16] ng/ml; P = NS) were similar in hIGFREO and wild-type mice (Fig. 4D and E).

There were no differences in endothelial cell or lung eNOS mRNA (Fig. 5A). In keeping with our previous report (15), insulin-resistant hybrid receptors were increased in endothelial cells from hIGFREO mice (Fig. 5B); there was no difference in IR-β or tyrosine-phosphorylated IR-β expression between hIGFREO and wild-type littermates (Supplementary Fig. 1). Basal serine 1177 eNOS phosphorylation (Fig. 5C), eNOS activation (Fig. 5D), and aortic ring relaxation responses (Fig. 5E and F) in response to IGF-1 were no different when comparing hIGFREO mice and wild-type mice. Insulin-mediated aortic relaxation in hIGFREO mice was no different than wild-type mice (data not shown). IGF-1–stimulated 1177 serine phosphorylation of eNOS was similar in endothelial cells from hIGFREO and wild-type mice (Supplementary Figs. 2 and 3).

There were no differences in aortic ring relaxation responses to acetylcholine or SNP between mice (Fig. 6A and B). In contrast to this, hIGFREO mice had enhanced constrictor responses to phenylephrine (Fig. 6C), and consistent with reduced basal NO bioavailability, a blunted constrictor response to N^G-monomethyl-l-arginine (l-NMMA) (Fig. 6D–F).

Trained, semirestrained hIGFREO mice had systolic blood pressure of 106.7 (3.0) mmHg compared with wild-type littermates 105.8 (4.5) mmHg; P = 0.8 (Fig. 7A).

Insulin-mediated NO release (Fig. 7B and C), eNOS activation (Fig. 7D), and eNOS serine 1177 phosphorylation (Fig. 7E) were blunted in hIGFREO endothelial cells compared with wild type.

Recent studies provide compelling evidence for a role for nNOS-derived NO in basal vascular tone (22). We therefore measured nNOS mRNA and protein expression in endothelial cells from wild-type and hIGFREO mice. There was no difference nNOS mRNA expression (data not shown), whereas protein expression was reduced approximately twofold in hIGFREO mice (Supplementary Fig. 4).

Endothelium regeneration after wire injury (Fig. 8A) was significantly enhanced in hIGFREO mice compared with wild-type mice (percent recovered area wild type 57% [2%] versus hIGFREO 47% [5%]; P < 0.05). Consistent with enhanced in situ endothelium regeneration, migration to control and VEGF was substantially enhanced in endothelial cells from hIGFREO mice (Fig. 8B). To further examine the role of the IGF-1R in endothelial regeneration after denuding arterial injury, we performed complementary wire injury experiments in ECIGF-1R^+/− mice and assessed regeneration. Consistent with our findings in hIGFREO mice, ECIGF-1R^+/− mice had blunted endothelium regeneration when compared with wild-type littermates (percent recovered at 5 days, ECIGF-1R^+/− 26% (1%) versus wild-type 42% (1%); P < 0.01) (Fig. 8C). We examined the expression of hIGF-1R in spleen and bone marrow-derived EPC, and hIGF-1R was expressed in EPC from hIGFREO but not wild type (Supplementary Fig. 5).

In this study, we describe the phenotype of a novel transgenic mouse with endothelium-targeted expression of the human IGF-1R generated to investigate the role of the IGF-1R in regulating endothelial cell function. The major findings of this study are as follows: 1) endothelial-specific IGF-1R overexpression leads to an increase in total IGF-1R mRNA and protein in hIGFREO mice compared with wild-type littermates; 2) endothelial IGF-1R overexpression is associated with an increase in vasoconstrictor responsiveness to phenylephrine, reduced basal NO release, and reduced nNOS expression; 3) the IGF-1R is a potent negative regulator of insulin-mediated NO release; and 4) despite this reduced NO bioavailability, endothelial IGF-1R overexpression leads to enhanced endothelium regeneration after denuding wire injury.

We recently demonstrated that global haploinsufficiency of the IGF-1R enhances whole-body insulin sensitivity and insulin-mediated NO production (15) and knocking down IGF-1R in human endothelial cells enhanced basal and insulin-stimulated serine-phosphorylated eNOS. In mice with endothelium-specific knockdown of the IGF-1R, we confirmed a gene-dose response effect on basal NO bioavailability. Moreover, by crossing IGF-1R haploinsufficient mice with IR-haploinsufficient mice (IR^+/−), we were able to restore insulin-mediated NO release in IR^+/− mice. We also showed that reducing IGF-1R density altered IGF-1R–IR stoichiometry in favor of IR holoreceptors with a relative reduction in insulin-resistant hybrid receptors.

In the present report, we demonstrate that increasing IGF-1R numbers specifically in the endothelium has no effect on the response to systemic insulin, IGF-1, or glucose. Despite having no effect on whole-body glucose homeostasis, IGF-1R overexpression led to reduced insulin-stimulated eNOS serine phosphorylation, eNOS activation, and NO generation. Overexpression of the IGF-1R also reduced basal NO bioavailability. Reduced NO has been shown to promote the development of atherosclerosis (23,24). Serine phosphorylation increases eNOS calcium sensitivity and augments NO biosynthesis and release (25); consistent with this, eNOS phosphorylation status has been shown to play an important and favorable role in a range of pathological situations reminiscent of human vascular disease (26,27). Our dataset demonstrates that a relative increase in IGF-1R in relation to IR, which is a hallmark of type 2 diabetes (28,29), has an unfavorable impact on NO bioavailability.

eNOS can be activated by a range of different stimuli. Fluid shear stress and growth factors such as insulin activate eNOS via phosphatidylinositol 3-kinase and Akt/protein kinase B activation in what is often described as a calcium-independent fashion (25). Secondly, different agonists (such as acetylcholine) activate eNOS by stimulating a rise in intracellular calcium, described as calcium-dependent eNOS activation (30). In the present report, we have shown that the IGF-1R selectively inhibits insulin-mediated eNOS activation with preservation of acetylcholine responses. Interestingly, hIGFREO mice had a relative reduction in endothelial cell expression of nNOS, possibly contributing to reduced basal NO bioavailability.

Endothelial cell-specific overexpression of the IGF-1R had no effect on growth, whole body, or organ weight. In contrast to a recent report (31), the endothelial cell insulin resistance demonstrated in the present report had no effect on whole-body glucose uptake. The mechanisms underlying the difference between mice with endothelium-specific deletion of IR substrate-2 described by Kubota et al. (31) and hIGFREO mice are unclear. It may be that insulin resistance in the endothelium due to perturbation of insulin signaling at its most proximal signaling node (i.e., the IR), as seen in this report, our previous study (11), and that of Vicent et al. (32), does not affect insulin sensitivity in skeletal muscle. These data highlight the complexity of communication between different cell types involved in insulin signaling.

It is now a well-accepted paradigm that cardiovascular risk factors lead to damage/death of resident endothelial cells, and endogenous repair mechanisms are activated to repair the damage (for review, see Ref. 33). Intriguingly and despite reduced NO bioavailability, overexpression of the human IGF-1R in the endothelium accelerated repair of the endothelium after wire injury. Recent data demonstrate that repair of denuded endothelium is principally a result of elongation, growth, and migration of resident endothelial cells (34). The present dataset is compatible with this paradigm, showing enhanced migration of endothelial cells from hIGFREO mice.

More recently, it has emerged that bone marrow or other tissue-derived EPC may contribute to endothelial repair (33). We have demonstrated that insulin-resistant humans (35) and mice (10) have impairment of EPC mobilization and function. Studies have shown that IGF-1 may have favorable effects on EPC function and numbers in humans (36) and mice (37). Our dataset demonstrating the expression of the human IGF-1R in EPC from hIGFREO mice but not wild type and the enhanced endothelium repair after injury raises the possibility that increasing IGF-1R expression may have an effect on EPC function, and this warrants further studies.

In the present dataset and our recent report (15), we have demonstrated that within a tight and physiologically relevant range, the IGF-1R plays a critical role in regulating basal NO bioavailability and insulin-mediated NO release. We have also shown a novel role for the IGF-1R in regulation of nNOS and that the effect of the IGF-1R appears to be specific to insulin-mediated NO release.

Despite leading to reduced basal and insulin-stimulated NO release, increasing IGF-1R enhances endothelial repair. This provides a compelling portfolio of evidence supporting an important role for IGF-1R within a (patho)physiological range in regulating NO bioavailability and vascular repair. Many human diseases are characterized by excessive or impaired tissue growth and vascularization. The present dataset and our previous report (15) establish the manipulation of IGF-1R: insulin receptor stoichiometry as a novel approach to regulating NO bioavailability and a potential treatment for disorders of excessive or impaired vascular growth and repair. Examining how hybrid formation is regulated and whether or not the human IGF-1R behaves differently to mouse IGF-1R in hybrid formation warrant future work.

This work was supported by the British Heart Foundation and Medical Research Council, U.K.

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

H.I., H.V., P.S., A.A., R.M.C., N.Y., M.G., J.S., S.G., A.S., S.T.R., T.S.F., and V.K.G. performed experiments and analyzed data. S.X., P.J.G., K.M.C., D.J.B., and S.B.W. contributed to discussion and reviewed and edited the manuscript. M.T.K. wrote the manuscript. K.M.C., D.J.B., S.B.W., and M.T.K. obtained funding and designed experiments. M.T.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.