|Home | About | Journals | Submit | Contact Us | Français|
Impaired insulin signaling via phosphatidylinositol 3-kinase/Akt to endothelial nitric oxide synthase (eNOS) in the vasculature has been postulated to lead to arterial dysfunction and hypertension in obesity and other insulin resistant states. To investigate this, we compared insulin signaling in the vasculature, endothelial function, and systemic blood pressure in mice fed a high-fat (HF) diet to mice with genetic ablation of insulin receptors in all vascular tissues (TTr-IR−/−) or mice with genetic ablation of Akt1 (Akt1−/−). HF mice developed obesity, impaired glucose tolerance, and elevated free fatty acids that was associated with endothelial dysfunction and hypertension. Basal and insulin-mediated phosphorylation of extracellular signal-regulated kinase 1/2 and Akt in the vasculature was preserved, but basal and insulin-stimulated eNOS phosphorylation was abolished in vessels from HF versus lean mice. In contrast, basal vascular eNOS phosphorylation, endothelial function, and blood pressure were normal despite absent insulin-mediated eNOS phosphorylation in TTr-IR−/− mice and absent insulin-mediated eNOS phosphorylation via Akt1 in Akt1−/− mice. In cultured endothelial cells, 6 hours of incubation with palmitate attenuated basal and insulin-stimulated eNOS phosphorylation and NO production despite normal activation of extracellular signal-regulated kinase and Akt. Moreover, incubation of isolated arteries with palmitate impaired endothelium-dependent but not vascular smooth muscle function. Collectively, these results indicate that lower arterial eNOS phosphorylation, hypertension, and vascular dysfunction following HF feeding do not result from defective upstream signaling via Akt, but from free fatty acid–mediated impairment of eNOS phosphorylation.
Stimulation of insulin receptors in the vasculature leads to increased activation of the phosphatidylinositol 3-kinase (PI3K)/Akt pathway and the mitogen-activated protein (MAP) kinase (eg, extracellular signaling-regulated kinase [ERK]1/2) pathway.1-4 Insulin receptor (IR)-mediated stimulation of PI3K/Akt leads to endothelial nitric oxide (NO) synthase (eNOS) phosphorylation, NO production, and vasorelaxation.4-7 Insulin-mediated activation of ERK1/2 leads to endothelin (ET)-1 production, inhibition of eNOS phosphorylation, and subsequent vasocontraction.1,4,8
Evidence from several experimental models of insulin resistance reveals impaired insulin-mediated PI3K/Akt-dependent signaling in the vasculature, whereas ERK1/2 pathways are preserved or even augmented.2,8,9 Collectively, these observations have led to the hypothesis that an imbalance in vascular insulin-mediated signaling can precipitate cardiovascular complications including endothelial dysfunction and hypertension.1,4 Recently, it was shown that insulin-mediated Akt phosphorylation was preserved, but NO-mediated vasorelaxation was blunted, in arteries from obese, glucose intolerant versus lean mice, suggesting that impaired insulin-mediated signaling to eNOS via Akt might not be required to reduce NO bioavailability to an extent that produces endothelial dysfunction.10
Cardiovascular consequences of decreased insulin-mediated signaling have been evaluated in mice homozygous for IR deletion in the endothelium (VENIRKO mice)3 and in heterozygous mice with germline IR deletion (heterozygous IRKO mice).11 Vascular eNOS and ET-1 mRNA were lower in VENIRKO versus wild-type (WT) mice, blood pressure was similar between groups, but eNOS protein and phosphorylation levels were not reported. In contrast, eNOS mRNA was unchanged, blood pressure was elevated, and insulin-mediated eNOS phosphorylation in the vasculature and indices of NO bioavailability were reduced in heterozygous IRKO versus WT animals. Heterozygous IRKO mice are glucose-intolerant and hyperinsulinemic, which may secondarily influence vascular function.11,12 Thus, the specific contribution from impaired vascular insulin-mediated signaling to the pathogenesis of arterial dysfunction and hypertension is unclear.
In light of the uncertainty concerning the cardiovascular consequence(s) of insulin resistance in the vasculature, we compared insulin-mediated signal transduction to eNOS in the vasculature, endothelial function, and blood pressure between mice that consumed high-fat (HF) and standard chow (CON). Our findings indicated that endothelial dysfunction and hypertension associated with diet-induced obesity could not be explained by defective upstream signaling via Akt to eNOS in the vasculature. To validate these observations, we used 2 genetic models of “insulin resistance” that lack systemic metabolic disturbances associated with obesity and type 2 diabetes. First, IR-null mice with transgenic reexpression of the IR in brain, liver, and pancreatic β-cells (TTr-IR−/− mice)13 were used to test the consequence of absent insulin-mediated signaling to eNOS on vascular function and blood pressure. Second, Akt1-null mice (Akt1−/− mice)14,15 were used to evaluate the contribution of insulin-mediated signaling via Akt1 to eNOS on vascular function and blood pressure. Akt1−/− mice were chosen because this isoform was recently shown to be most important in regulating eNOS, based on observations that basal and vascular endothelial cell growth factor (VEGF)-stimulated NO production were reduced in endothelial cells isolated from mice with deletion of Akt1 but not Akt2.16
Findings from these genetic models suggested that disruption of insulin or Akt1 signaling to eNOS in the vasculature does not precipitate endothelial dysfunction or hypertension. Because the cardiovascular complications observed in mice with diet-induced obesity might not result from impaired vascular insulin-mediated signal transduction, we sought to determine the role of elevated circulating free fatty acids (FFAs).17,18 When bovine aortic endothelial cells (BAECs) were incubated with the saturated FFA palmitate, both basal and insulin-mediated eNOS phosphorylation and NO production were impaired despite normal activation of Akt and ERK. Furthermore, palmitate incubation precipitated endothelium-dependent dysfunction in isolated vessels. Collectively, these results indicate that reduced arterial eNOS phosphorylation and vascular complications do not result from defective upstream signaling via Akt but might be secondary to FFA-mediated impairment of eNOS phosphorylation.
All protocols were approved by the Institutional Animal Care and Use Committee. Ten-week-old C57Bl6 mice that consumed standard (CON) or HF chow for 10 to 12 weeks were used for experiments on metabolic characterization, insulin-mediated signal transduction in the vasculature, blood pressure, vascular function, detection of vascular oxidant load, and mRNA expression. Similar experiments were completed in TTr-IR−/−13,19,20 mice, Akt1−/−14 mice, and their WT littermates. Detailed procedures are provided in the expanded Materials and Methods section in the online data supplement at http://circres.ahajournals.org.
Data are presented as means±SEM. Significance was accepted at P<0.05. Comparison of 1 time point between 2 groups was made using an unpaired t test. Comparison of multiple time points between groups was made using a 1-way or 2-way repeated measures ANOVA. Tukey post hoc tests were performed when significant main effects were obtained.
Body weight increased (41% versus 18%; Figure 1A), gonadal fat pad mass was greater (0.91±0.25 versus 0.35±0.08 g), and fasting glucose (122±9 versus 107±7 mg/dL), area under the glucose tolerance test (GTT) curve (42397±3064 versus 28178±1507; Figure 1B), insulin (Figure 1C), and FFAs were higher (Figure 1D) in HF versus CON mice, respectively. Dual energy X-ray absorption indicated increased body and fat mass and decreased lean mass in HF versus CON mice, respectively (Figure I in the online data supplement). Dihydroethidium staining21 revealed no differences in superoxide anion (O2−) production (relative fluorescence intensity [RFI] units) in the endothelium (1438±182 and 1255±124) or vascular smooth muscle (904±89 and 814±143) of HF and CON mice, respectively (n=5 per group). Vascular NADPH oxidase activity was similar between HF and CON mice (not shown). The fold change in eNOS (0.99±0.19) and ET-1 mRNA (1.41±0.14; P=0.08) in HF relative to WT mice was not different (n=5 per group).
IRs were absent and IGF1 mRNA was present in the vasculature of TTr-IR−/− mice (Figure 2A; Online Figure II, A). Body weight (Figure 2B), gonadal fat pad mass (0.35±0.08 and 0.22±0.04 g, P=0.07), fasting glucose (Figure 2C), FFAs (2.0±0.1 and 2.3±0.1 mmol/L), and area under the GTT curve (17486±370 and 17646±1310; Figure 2C) were similar between WT and TTr-IR−/− mice, respectively. Insulin was greater in both the fasting condition (ie, minute 0) and 30 minutes following IP glucose during the GTT in TTr-IR−/− mice (Figure 2D). Vascular dihydroethidium staining was similar between TTr-IR−/− and WT mice in the endothelium (1334±97 and 1538±88) and vascular smooth muscle (876±26 versus 1025±45; n=5 per group), respectively. The fold change in eNOS (0.98±0.28) and ET-1 (1.06±0.16) mRNA relative to WT mice was not different between groups (n=5 per group).
Because metabolic disturbances and vascular complications (see below) were observed in HF but not TTr-IR−/− mice, we placed TTr-IR−/− mice on HF chow (HF-TTr-IR−/− mice). However, HF-TTr-IR−/− mice did not develop systemic metabolic disturbances (Online Figure III, A through D) or arterial dysfunction (data not shown). As such, no further experiments were performed on these animals. Akt1−/− mice have a metabolic profile similar to their WT littermates.14,15
Compared to vehicle, insulin increased the ratio of phosphorylated (p)-ERK1/2 to total ERK similarly in aortae from CON and HF mice (Figure 3A). Insulin increased the ratio of p-Akt S473 and p-Akt T308 to total Akt in CON and HF animals, but insulin-mediated Akt phosphorylation at both residues was modestly yet significantly blunted in arteries from HF versus CON mice (Figure 3B). Furthermore, insulin increased the ratio of p-eNOS S1177 to total eNOS in arteries from CON but not HF mice (Figure 3C). Importantly, basal levels of p-eNOS S1177 were virtually nonexistent in arteries from HF versus CON mice, and insulin was not capable of increasing p-eNOS at this site. HF feeding did not alter vascular eNOS at T495, S617, or S635, or the upstream kinases AMP-activated protein kinase (AMPK) or protein kinase (PK)A (Online Table I).
Compared to vehicle stimulation, insulin increased the ratio of p-ERK1/2 to total ERK (Figure 4A), p-Akt S473 or p-Akt T308 to total Akt (Figure 4B), and p-eNOS S1177 to total eNOS in aortae from WT mice (Figure 4C). As expected, these effects were not observed in arteries from TTr-IR−/− animals.
Western blot experiments confirmed the absence of Akt1 protein in the vasculature of Akt1−/− mice (Online Figure II, B). In aortae from WT and Akt1−/− mice, insulin equivalently increased the ratio of p-ERK1/2 to total ERK1/2 (Figure 5A), p-Akt S473 or p-Akt T308 to total Akt (Figure 5B), and p-eNOS S1177 to total eNOS (Figure 5C). To explore whether insulin-mediated Akt phosphorylation in vessels from Akt1−/− mice was mediated via the Akt2 isoform, Akt2 was immunoprecipitated from vascular homogenates obtained from Akt1−/− and WT mice after stimulation with insulin or vehicle. Compared to vehicle, insulin increased the ratio of p-Akt S473 and p-Akt T308 to total Akt2 in both groups (Figure 5D). These results confirm insulin-mediated p-Akt in Akt1−/− mice is mediated via the Akt2 isoform. Additionally, we sought to confirm the compartment wherein eNOS phosphorylation was occurring. Results in Online Figure IV show that insulin-stimulation evokes a robust increase in p-eNOS to total eNOS in intact arteries that is abolished in vessels denuded of endothelium. The pattern of insulin-mediated signaling for ERK, Akt, and eNOS among the 3 groups (ie, HF, TTr-IR−/−, and Akt1−/− mice and their respective controls) was similar between aorta and homogenates of iliac/femoral arteries.
Arterial blood pressure and heart rate were assessed in conscious mice for 30 seconds every 15 minutes over 72 hours starting 6 days following surgical implantation of a telemetry device in the abdomen. Blood pressure was greater in HF versus CON mice (Figure 6A), but heart rate (646±45 versus 601±18 bpm, respectively) was similar between groups. Mean arterial pressure (Figure 6B) and heart rate (608±21 and 636±12 bpm) were similar between TTr-IR−/− and WT mice, respectively. Likewise, arterial pressure (Figure 6C) and heart rate were similar between Akt1−/− (646±16 bpm) and WT mice (620±26).
Because basal vascular eNOS phosphorylation was compromised, and hypertension existed, in HF but not TTr-IR−/− or Akt1−/− mice, we sought to determine whether functional indices of stimulated and basal eNOS activity displayed a similar pattern. Acetylcholine-evoked (Figure 7A) and NG-monomethyl-l-arginine (L-NMMA)-evoked (Figure 7B) eNOS activity and insulin-mediated vasorelaxation (Online Figure V, A) were blunted, whereas phenylephrine-induced vasocontraction was greater (Online Figure VI, A) in femoral arteries from HF versus CON mice. Relaxation to sodium nitroprusside (Figure 7C) was similar between groups. Acetylcholine-evoked (Figure 7D) and L-NMMA–evoked (Figure 7E) eNOS activity, as well as vascular smooth muscle responses to sodium nitroprusside (Figure 7F), were similar among femoral arteries between TTr-IR−/− and WT mice. Insulin-mediated vasorelaxation was predictably absent in TTr-IR−/− versus WT mice (Online Figure V, B). Furthermore, vasocontraction was greater in arteries from TTr-IR−/− versus WT mice only at the highest doses of phenylephrine (Online Figure VI, B). Results from Akt1−/− and WT mice (Figure 7G through 7I; Online Figure VI, C) were similar to those obtained from TTr-IR−/− mice. Results using aortae (data not shown) were similar to those obtained from femoral arteries. Vessel characteristics for femoral arteries are shown in Online Table II.
Arterial dysfunction existed in HF mice, yet insulin-mediated signal transduction to eNOS via Akt1 in the vasculature was intact. Thus, we sought to determine whether a component of the systemic environment ie, elevated FFAs, was responsible for arterial dysfunction and lower basal and insulin-stimulated p-eNOS S1177 to total eNOS in the vasculature of HF versus CON mice. First, BAECs treated with 500 μmol/L palmitate for 6 hours had lower basal and insulin-stimulated p-eNOS S1177, yet signaling to Akt and ERK was intact (Figure 8A through 8C). Second, 6 hours of palmitate treatment abolished insulin-mediated NO production by BAECs (Figure 8D). Finally, 3 hours of palmitate treatment impaired endothelium-dependent (Figure 8E) but not endothelium-independent vasorelaxation (not shown). In parallel experiments, neither endothelium-dependent nor endothelium-independent vasorelaxation was impaired in arteries treated for 3 hours with vehicle (BSA) or volume (milliQ water; data not shown). Basal p-eNOS was lowered to the same degree by 3 hours (Figure 8F) and 6 hours (data not shown) of palmitate incubation.
We assessed the contribution from vascular insulin signaling to eNOS in regulating endothelial function and blood pressure in mice with diet-induced obesity and in mice with complete (ie, TTr-IR−/− mice) and selective (ie, Akt1−/− mice) “resistance” to insulin signaling. Five main findings were observed. First, in HF mice, insulin-stimulated phosphorylation of eNOS was abolished, hypertension existed, but insulin-mediated signaling in the vasculature to ERK1/2 and Akt was preserved. Second, in TTr-IR−/− mice, insulin-stimulated phosphorylation of ERK1/2, Akt, and eNOS in the vasculature was predictably absent, but arterial blood pressure was similar to WT littermates. Third, the absence of Akt1 expression did not influence insulin-mediated phosphorylation of ERK1/2, Akt, and eNOS in the vasculature or blood pressure. Fourth, basal eNOS phosphorylation was almost absent in the vasculature of HF mice, and these animals displayed endothelial dysfunction together with systemic disturbances associated with diet-induced obesity. In contrast, basal eNOS phosphorylation, vascular function, and metabolic characteristics were similar among TTr-IR−/− and Akt1−/− animals and their respective WT littermates. These results suggested that hypertension and vascular dysfunction in mice with diet-induced obesity might be precipitated by reduced basal vascular eNOS phosphorylation caused by a component(s) of the circulating metabolic environment rather than by defective signaling via Akt1 to eNOS. The component we focused on was the elevation in FFAs observed in obese versus lean mice. In this regard, our fifth main finding was that basal and insulin-mediated eNOS phosphorylation and NO bioavailability are impaired, but signaling to Akt and ERK are intact, in endothelial cells incubated with the saturated FFA palmitate. Thus, elevated FFAs might be responsible for lowering vascular eNOS to a degree that precipitates cardiovascular complications in mice with diet-induced obesity.
Although impaired insulin signaling in the vasculature is widely accepted to be associated with obesity, type 2 diabetes, and generalized insulin resistance,1 the specific contribution of vascular insulin-resistance to the pathogenesis of arterial dysfunction and hypertension has not been definitively proven. For example, insulin resistance in spontaneously hypertensive rats (SHR) was accompanied by impaired vascular PI3K-dependent NO production, enhanced mitogen-activated protein kinase (MAPK)-dependent ET-1 secretion, and endothelial dysfunction versus control (WKY) rats.8 In that report, mesenteric arterial function was restored in SHRs by ET-1A/B receptor blockade or by MAPK/ERK kinase inhibition. Whereas this study and others2,9 suggest that an imbalance in vascular insulin-mediated signaling via Akt to eNOS together with intact or exaggerated MAPK-dependent ET-1 production has the potential to disrupt vascular homeostasis, a recent publication challenged this concept. Specifically, insulin-stimulated Akt phosphorylation was normal in arteries from mice with diet-induced obesity and glucose intolerance, although eNOS structure was disrupted and endothelial dysfunction was present.10 These results suggest that a component of endothelial dysfunction might be independent of upstream modulation via Akt.
The contribution of insulin-mediated signal transduction in the vasculature to blood pressure regulation is also unclear. For instance, blood pressure measured via tail cuff was similar between VENIRKO and WT mice,3 whereas hypertension existed in IRS-1–deficient (IRS-1−/−) mice versus their respective controls.22 Because VENIRKO mice model nonselective insulin resistance and IRS-1−/− mice model PI3K/Akt selective “resistance” to IR-mediated signaling, these findings support the hypothesis that imbalanced downstream signaling from the IR disrupts cardiovascular homeostasis.1,4,23,24 In contrast, arterial pressures were higher in mice wherein both signaling pathways downstream from the IR were thought to be compromised (ie, heterozygous IRKO mice) in a balanced manner.11
In an attempt to clarify this issue, we compared insulin signaling in the vasculature, endothelial function, and systemic blood pressure in mice with diet-induced obesity to mice with genetic ablation of insulin receptors in all vascular tissues (TTr-IR−/−) or mice with genetic ablation of Akt1 (Akt1−/−). We hypothesized that HF feeding would selectively impair insulin-mediated signaling to eNOS via Akt in the vasculature, and this defect would lead to hypertension and endothelial dysfunction. Our results did not support this hypothesis. Instead, relative to arteries from lean mice, insulin-mediated Akt phosphorylation was preserved but insulin-stimulated eNOS phosphorylation at S1177 was abolished in vessels from obese animals. Furthermore, hypertension and endothelial dysfunction existed in obese versus lean mice. These data indicate the lack of insulin-mediated eNOS phosphorylation at S1177 in the vasculature, systemic hypertension, and vascular dysfunction that existed in HF mice did not result from deficient upstream activation of eNOS via Akt.
In addition to diet-induced obesity produced by HF feeding, we used 2 genetic models to test our overall hypothesis. First, we reasoned that if insulin-mediated signal transduction was absent, ie, using TTr-IR−/− mice, then there likely would be no net effect on blood pressure. Our findings support this hypothesis and confirm a previous study wherein similar blood pressures existed between VENIRKO and WT mice3 but contrast with results indicating that heterozygous IRKO mice are hypertensive relative to their WT littermates.11 In this latter study, vascular insulin-mediated eNOS phosphorylation was attenuated in aortae from heterozygous IRKO versus WT mice. Although these data suggest the heterozygous IRKO phenotype might have resulted from impaired insulin signaling to eNOS, upstream kinases responsible for activating this enzyme were not evaluated. As such, it is not possible to know whether an imbalance in insulin-mediated signal transduction existed in vessels from heterozygous IRKO versus WT mice.
Mice with targeted deletion of Akt1 were used as a second approach to test our overall hypothesis. We reasoned that if selective resistance to insulin-stimulated eNOS phosphorylation via Akt in the vasculature contributed to hypertension, then blood pressure should be elevated in Akt1−/− versus WT mice. Mice with specific deletion of Akt1 were chosen for several reasons. First, of the 3 major Akt isoforms (Akt1/PKBα, Akt2/PKBβ, and Akt3/PKBγ), only Akt1 and Akt2 are present in aortae and femoral arteries of mice.16 Second, evidence from mouse lung endothelial cells indicates that Akt1 is the isoform predominantly responsible for basal and VEGF-stimulated NO production.16 Finally, because Akt1 mice do not possess a diabetic phenotype,14,15 contributions from circulating metabolic factors to endothelial dysfunction or hypertension do not exist. Contrary to our hypothesis, blood pressures were not different between Akt1−/− and WT mice. Surprisingly, insulin-mediated Akt and eNOS phosphorylation were intact in both groups. Because Akt1 protein was verified to be absent in the same vessels wherein these results were obtained, we explored the possibility and confirmed experimentally that insulin-mediated eNOS S1177 phosphorylation in the vasculature can occur via the Akt2 isoform.
Whereas we observed robust insulin-mediated eNOS S1177 phosphorylation in vessels from Akt1−/− and WT mice, Ackah et al showed that VEGF-stimulated NO production was impaired in mouse lung endothelial cells obtained from Akt1−/− versus WT mice.16 Our studies suggest that in endothelial cells and the vasculature, insulin and VEGF differ importantly in their ability to phosphorylate eNOS S1177 via Akt2. In our study and that of Ackah et al, the presence of Akt2 in endothelial cells of Akt1−/− mice was confirmed. Thus, whereas VEGF does not stimulate eNOS phosphorylation in the absence of Akt1 via residual Akt2 in mouse lung endothelial cells, we clearly show that insulin-mediated activation of Akt2 is sufficient to activate eNOS in the vasculature. We do not know the molecular basis for these differences, but spatial differences in the subcellular localization of Akt1 versus Akt2 and components of the VEGF or insulin signaling pathways might play a role. We also considered the possibility that differences arose from the fact that we studied signaling in whole artery homogenates versus cultured endothelial cells, because vascular compartments other than the endothelium might have contributed to the eNOS phosphorylation we observed. We believe this is unlikely, however, because insulin-mediated eNOS S1177 phosphorylation in intact vessels was abolished in arteries denuded of their endothelium (Online Figure IV). Thus, the discrepancy between studies likely resides in differences in the signal transduction mechanisms for VEGF versus insulin.
Basal eNOS phosphorylation was virtually absent in vessels from HF versus control mice, whereas it was similar between TTr-IR−/− and Akt1−/− animals and their respective controls. Because phosphorylation of eNOS S1177 positively regulates eNOS activity,5,25-27 it is reasonable that NO bio-availability might have been compromised in vessels from HF mice to an extent that precipitated endothelial dysfunction and hypertension relative to TTr-IR−/− or Akt1−/− animals. This is supported by results from 3 protocols using isolated vessels. First, tension development in response to eNOS inhibition (a functional estimate of basal eNOS activity28) was less in aortae and femoral arteries from HF versus control mice. Second, phenylephrine-induced responses were greater in HF mice, suggesting that endogenous opposition to vasocontraction from NO might be compromised. Finally, acetylcholine-evoked vasorelaxation (a functional estimate of stimulated eNOS activity) was blunted in vessels from HF versus control mice. Because these end points were generally similar between TTr-IR−/− and Akt1−/− mice and their WT littermates, it is reasonable to speculate that minimal basal eNOS phosphorylation at S1177 observed in vessels from HF mice might have precipitated endothelial dysfunction and hypertension. Collectively, the lack of basal arterial eNOS phosphorylation at S1177, rather than deficient signaling to eNOS via Akt1 in the vasculature, appears to contribute importantly to hypertension that exists in mice with diet-induced obesity. Further evidence supporting this statement is that p-eNOS 617 (another eNOS phosphorylation target downstream from Akt) was similar in vessels from CON and HF mice. In this regard, if defective signaling via Akt to eNOS contributed to minimal p-eNOS S1177, then p-eNOS 617 likely would have been lower.
Findings from studies similar to ours implicate hyperglycemia, defective signaling to eNOS via AMPK, and elevated FFAs as important contributors to impaired arterial eNOS activity, function, and/or phosphorylation.10,29,30 Hyperglycemia in our HF mice was much less severe than values reported by Molnar et al, which might have contributed to increased oxidative stress and disruption of eNOS protein dimers in the vasculature.10 Whereas Wu et al29 demonstrated lower arterial p-AMPK T172 to total AMPK and impaired p-eNOS S1177 to total eNOS in arteries from fat-fed versus lean mice and in cultured cells treated with saturated versus unsaturated fatty acids, we found no difference in arterial p-AMPK between vessels from HF and CON animals. Du et al reported that reduced aortic eNOS activity in mice fed high-fat chow could be normalized by treatment with an antilipolytic agent that decreased fatty acid release from adipose cells.30 Moreover, Edirisinghe et al observed impaired endothelial function in aortic segments incubated with palmitate for 60 minutes.17 Because we observed a >3-fold elevation of FFAs in HF versus CON mice, we explored this mechanism further. When BAECs were incubated with palmitate, basal and insulin-stimulated eNOS phosphorylation was impaired, but signaling to ERK and Akt was intact. The palmitate-induced reduction of eNOS phosphorylation was sufficient to impair insulin-mediated NO production by BAECs and endothelium-dependent but not vascular smooth muscle function of isolated arteries. Results in BAECs and vessels exposed to palmitate mimic those we observed in arteries from HF mice. Because the source of fat in the HF diet primarily was from lard, it is possible that metabolites of saturated fatty acids (eg, the sphingolipid ceramide) might accumulate in response to HF feeding and lower arterial NO bioavailability.29 Indeed, preliminary data from our laboratory indicate vascular ceramide increases ≈2-fold (P<0.05) in mice exposed to HF feeding for 10 to 14 weeks. Although further work is warranted to elucidate the precise molecular signal(s) whereby HF feeding precipitates hypertension and vascular dysfunction, results from the present study indicate these cardiovascular complications occur via mechanisms that are independent of defective IR-mediated signaling to eNOS via Akt1 in the vasculature and might be related to the accumulation of FFAs and subsequent impairment of vascular NO production.
The assistance of Dr V. Zaha and Heather Theobald (RT-PCR analyses) and Drs. S. Sena (mouse colonies), S. Litwin (echocardiography), A. Rohrwasser (biotelemetry), W. Holland, and S. Summers (cell-based experiments) is greatly appreciated. Breeding pairs of TTr-IR−/− mice were generously provided by Dr D. Accili from (Columbia University, New York, NY). Student support was provided, in part, by the Western States Affiliate of the American Heart Association Summer Student Research Program, the American Physiological Society, and the University of Utah Undergraduate Research Opportunities Program.
Sources of Funding
This work was funded, in part, by the University of Utah College of Health, the University of Utah Research Foundation, American Heart Association Western States Affiliate Grant-In-Aid 06-55222Y, and NIH grants R15 HL 091493-01 (to J.D.S.) and R01 HL 070070 (to E.D.A.). E.D.A. is an Established Investigator of the American Heart Association.