Effect of endothelium-specific insulin resistance on endothelial function in vivo

Abstract

Objective: Insulin resistance is an independent risk factor for the development of cardiovascular atherosclerosis. A key step in the development of atherosclerosis is endothelial dysfunction, manifest by a reduction in bioactivity of nitric oxide (NO). Insulin resistance is associated with endothelial dysfunction; however, the mechanistic relationship between these abnormalities and the role of impaired endothelial insulin signaling versus global insulin resistance remains unclear.

Research design and methods: To examine the effects of insulin resistance specific to the endothelium, we generated a transgenic mouse with endothelium-targeted overexpression of a dominant-negative mutant human insulin receptor (ESMIRO). This receptor has a mutation (Ala-Thr(1134)) in its tyrosine kinase domain that disrupts insulin signaling. Humans with the Thr(1134) mutation are insulin resistant. We performed metabolic and vascular characterization of this model.

Results: ESMIRO mice had preserved glucose homeostasis and were normotensive. They had significant endothelial dysfunction as evidenced by blunted aortic vasorelaxant responses to acetylcholine (ACh) and calcium ionophore. Furthermore, the vascular action of insulin was lost in ESMIRO mice, and insulin-induced endothelial NO synthase (eNOS) phosphorylation was blunted. Despite this phenotype, ESMIRO mice demonstrate similar levels of eNOS mRNA and protein expression to wild type. ACh-induced relaxation was normalized by the superoxide dismutase mimetic, Mn(III)tetrakis(1-methyl-4-pyridyl) porphyrin pentachloride. Endothelial cells of ESMIRO mice showed increased superoxide generation and increased mRNA expression of the NADPH oxidase isoforms Nox2 and Nox4.

Conclusions: Selective endothelial insulin resistance is sufficient to induce a reduction in NO bioavailability and endothelial dysfunction that is secondary to increased generation of reactive oxygen species. This arises independent of a significant metabolic phenotype.

FIG. 1.

A: Schematic of the Tie2/mutant human insulin receptor transgene. Forward (F) and reverse (R) oligonucleotide primers specific for the human insulin receptor are represented. B: PCR of genomic DNA from tail lysates using the primers demonstrates incorporation of the transgene into genomic DNA in both lines of ESMIRO mice. C: Transgene expression in different organs from ESMIRO mice. ESMIRO mice demonstrated increased transgene expression in lung (Lu) and kidney (Ki) compared with liver (Li), spleen (SP), and heart (He). −ve, negative control; con, positive control (construct); RT-ve, reverse transcriptase negative. D: Immunohistochemistry of sections of thoracic aorta from ESMIRO and wild-type (WT) mice. Staining with an antibody specific for the human insulin receptor demonstrated transgene protein expression in ESMIRO but not wild-type endothelium (magnification ×400). The transgene colocalized in the endothelium with von Willebrand factor (data not shown). BP, base pair. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (Please see http://dx.doi.org/10.2337/db07-1111 for a high-quality digital representation of this figure.)

FIG. 2.

The action of insulin on the aorta of ESMIRO and wild-type (WT) mice. A: Wild-type (▪) and ESMIRO (▵) mice had similar aortic vasoconstrictor responses to phenylepherine. B: Incubation with insulin (100 mU/ml) for 2 h significantly blunted this response in wild-type mice but had no effect in ESMIRO mice (n = 8). C and D: Immunoblot of eNOS and phospho-eNOS protein expression in wild-type and ESMIRO aortae (n = 5). There was no difference in eNOS protein expression between the two groups (corrected for β-actin). However, wild-type phospho-eNOS expression was significantly greater than that of ESMIRO mice at baseline. Furthermore, a significant increment in eNOS phosphorylation at Ser¹¹⁷⁷ was seen in wild-type mice after insulin injection. No such increment was seen in ESMIRO mice. *P < 0.05 wild type vs. ESMIRO; #P < 0.05 wild type vs. wild type + insulin.

FIG. 3.

Aortic relaxation responses. ESMIRO mice (▵) demonstrated significantly impaired relaxation responses to ACh (A) and calcium ionophore A23187 (B) compared with wild-type (WT) littermates (▪). C: Relaxation responses to SNP were similar in ESMIRO and wild-type mice. n = 8, *P < 0.05.

FIG. 4.

Assessment of aortic superoxide production in ESMIRO and wild-type (WT) mice. ESMIRO mice demonstrate impaired relaxation responses to ACh at baseline. (see Fig. 3). A: Importantly, ESMIRO ACh responses (▵) were normalized after incubation with the SOD mimetic, 10 μmol/l MnTMPyP, providing physiological evidence that increased superoxide production in ESMIRO aortae contributes to the impaired relaxation responses seen. B: Lucigenin-enhanced chemiluminescence was performed on aortic homogenates. ESMIRO aortae demonstrated increased superoxide production compared with wild type. In both groups, superoxide production was completely inhibited by the ROS scavenger tiron or the flavoprotein inhibitor DPI. There was no significant inhibition with the NOS inhibitor l-NAME. These data suggest a role for NADPH oxidase as the major source of the increased superoxide in the ESMIRO aorta. *P < 0.001 with inhibitor compared with wild-type baseline; #P < 0.001 with inhibitor compared with ESMIRO baseline. ILU, integrated light units.

FIG. 5.

Assessment of CMEC superoxide production. A: In situ ROS generation in CMECs was assessed through measurement of peak fluorescence of CMECs exposed to 2 μmol/l DHE (5 min). CMECs from ESMIRO mice showed significantly greater fluorescence than wild type (WT); n = 6 mice, *P < 0.05. B: Using CMEC homogenates, 5 μmol/l lucigenin-enhanced chemiluminescence was repeated. Similar to the results observed in whole aortae, this suggested increased NADPH-dependent superoxide production in ESMIRO CMECs at baseline. Superoxide levels were reduced by tiron and DPI but not by l-NAME. *P < 0.001 with inhibitor compared with wild-type baseline; #P < 0.001 with inhibitor compared with ESMIRO baseline. ILU, integrated light units.

FIG. 6.

Real-time RT-PCR examination of Nox2 and Nox4 mRNA expression in CMECs and whole aortae (relative to β-actin). A and C: ESMIRO mice demonstrate significantly increased Nox2 and Nox4 mRNA expression in CMECs. B and D: A similar pattern of expression is noted in aortae. CMECs, n = 6; aortae, n = 5. *P < 0.05. WT, wild type.