Mol. Nutr. Food Res. 2014, 00, 1–13 DOI 10.1002/mnfr.

201400539 1

RESEARCH ARTICLE

High-fat diet induces endothelial dysfunction through

a down-regulation of the endothelial
AMPK–PI3K–Akt–eNOS pathway
Concha F. Garcı́a-Prieto1 , Francisco Hernández-Nuño1 , Danila Del Rio1 , Gema Ruiz-Hurtado2,3 ,
Isabel Aránguez2,4 , Mariano Ruiz-Gayo1 , Beatriz Somoza1 and Marı́a S. Fernández-Alfonso2
1
Departamento de Ciencias Farmacéuticas y de la Salud, Facultad de Farmacia, Universidad CEU-San Pablo,
Madrid, Spain
2
Instituto Pluridisciplinar and Departamento de Farmacologı́a, Facultad de Farmacia, Universidad Complutense,
Madrid, Spain
3
Unidad de Hipertensión, imas12, Hospital 12 de Octubre, Madrid, Spain
4
Departamento de Bioquı́mica, Facultad de Farmacia, Universidad Complutense, Madrid, Spain

Scope: Activation of endothelial adenosine monophosphate-activated protein kinase (AMPK) Received: August 4, 2014
contributes to increase nitric oxide (NO) availability. The aim of this study was to assess if Revised: November 8, 2014
high-fat diet (HFD)-induced endothelial dysfunction is linked to AMPK deregulation. Accepted: November 13, 2014
Methods and results: Twelve-week-old Sprague Dawley male rats were assigned either to control
(10 kcal % from fat) or to HFD (45 kcal % from fat) for 8 wk. HFD rats segregated in obesity-prone
(OP) or obesity-resistant (OR) rats according to body weight. HFD triggered an impaired glucose
management together with impaired endothelium-dependent relaxation, reduced endothelial
AMPK activity and lower NO availability in aortic rings of OP and OR cohorts. Relaxation
evoked by AMPK activator, 5-aminoimidazole-4-carboxamide-1-␤-d-ribofuranoside (AICAR)
was reduced in both OP and OR rings, which exhibited lower p-AMPK␣-Thr172 /AMPK␣ ratios
that negatively correlated with plasma non-esterified fatty acids (NEFA) and triglycerides (TG).
Inhibition of PI3K (wortmannin, 10−7 M) or Akt (triciribine, 10−5 M) reduced relaxation to
AICAR only in the control group (p < 0.001). Akt (p-Akt-Ser473 ) and eNOS phosphorylation
(p-eNOS-Ser1177 ) were significantly reduced in OP and OR (p < 0.01).
Conclusion: Endothelial dysfunction caused by HFD is related to a dysfunctional endothelial
AMPK–PI3K–Akt–eNOS pathway correlating with the increase of plasma NEFA, TG, and an
impaired glucose management.

Keywords:
Adenosine monophosphate-activated protein kinase / Endothelial function / High-
fat diet / Obesity-prone / Obesity-resistant

 Additional supporting information may be found in the online version of this article at
the publisher’s web-site

1 Introduction
Correspondence: Dr. Marı́a S. Fernández-Alfonso, Instituto
The AMP-activated protein kinase (AMPK) is a ubiquitously
Pluridisciplinar, Paseo Juan XXIII, 1. 28040, Madrid, Spain
E-mail: [email protected] distributed Ser/Thr kinase. It is a sensor of cellular energy sta-
Fax: +34-913943264 tus that is activated to restore energy balance after depletion

Abbreviations: +E, with endothelium; –E, endothelium free;

ACh, acetylcholine; AICAR, 5-aminoimidazole-4-carboxamide-1- lial nitric oxide synthase; GTT, intraperitoneal glucose toler-
␤-d-ribofuranoside; AMPK, adenosine monophosphate-activated ance test; HFD, high-fat diet; LKB1, liver kinase B1; NEFA, non-
protein kinase; ANOVA, analysis of variance; AUC, area under esterified fatty acids; NO, nitric oxide; OLETF, Otsuka Long Evans
the concentration-response curves; BW, body weight; CaMKK␤, Tokushima Fatty; OP, obesity-prone; OR, obesity-resistant; Phe,
calcium/calmodulin-dependent protein kinase kinase-␤; DIO, diet- phenylephrine; SD, Sprague Dawley; SNP, sodium nitroprusside;
induced obesity; EMax, maximum response; eNOS, endothe- SOD, superoxide dismutase; TG, triglycerides


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
2 C. F. Garcı́a-Prieto et al. Mol. Nutr. Food Res. 2014, 00, 1–13

of energy stores [1, 2]. AMPK is the downstream component dark-light cycles (12h/12h from 8:00 am to 8:00 pm) and tem-
of a protein kinase cascade that is activated by the rise in cel- perature (22°C) with standard food and water ad libitum. Af-
lular AMP:ATP ratio and thus switches off ATP-consuming ter 1 wk, animals were housed individually, divided into two
anabolic pathways via phosphorylation or regulation of en- groups with a similar mean body weight (BW), and assigned
zymes relevant for energy storage [1, 2]. either to a control diet (D12450B, 10% of energy from fat, 70%
AMPK is expressed in the vasculature in both endothelial from carbohydrates, and 20% from protein; 3.78 kcal/g; n =
[3] and smooth muscle cells [4]. In cultured endothelial cells, 10) or to a high-fat diet (HFD, D12451, 45% of energy from fat,
activation of AMPK has been shown to stimulate endothelial 35% from carbohydrates, and 20% from protein; 4.65 kcal/g;
nitric oxide synthase (eNOS) phosphorylation at Ser1177 and n = 20) (Test Diet Limited BCM IPS Ltd., London, UK). Af-
Ser633 and NO availability [5, 6] through the Akt–eNOS [7] ter 4 wk of diet, HFD rats were subdivided in obesity-prone
and/or PI3K–Akt–eNOS [8] pathways. Whole vessel studies (OP) and obesity-resistant (OR) animals according to BW on
suggest that AMPK modulates vascular function through the basis of BW increase. HFD rats exhibiting BW increase
endothelium-dependent and independent mechanisms. In in the range of the control group BW (50.32 ± 4.78 g) were
conduit arteries, AMPK increases eNOS phosphorylation at assigned to the obesity-resistant group (n = 10, OR). HFD
Ser1177 [9] and enhances endothelium-dependent vasodilata- rats over these limits were considered as obesity-prone indi-
tion [10,11], although the pathway linking both events has not viduals (n = 10; OP). Animals had free access to food during
been characterized yet. Similar findings have been observed 4 additional weeks, during which BW and food intake were
in resistance vessels of the rat [12] and humans [13]. AMPK monitored twice a week. On day 49th of dietary treatment, rats
also regulates myosin-light chain kinase activity by decreas- were fasted from 8 am until 2 pm for intraperitoneal glucose
ing the sensitivity of vascular smooth muscle to intracellular tolerance test. The last day, rats were weighed and killed by de-
Ca++ , thus reducing vascular tone [14]. Direct pharma- capitation at 9 am. Blood was collected in chilled EDTA-coated
cological activation of vascular AMPK by metformin [15] polypropylene tubes. To minimize alterations in AMPK ac-
or 5-aminoimidazole-4-carboxamide-1-␤-d-ribofuranoside, tivity or phosphorylation, 24-h fasting was omitted to avoid
AICAR [15,16] as well as by hypoxia [17] leads to a decrease in lipolysis and any eventual reduction of mesenteric perivascu-
cholesterol synthesis [18] and to an increase in fatty acid oxida- lar adipose tissue amount [23]. Biochemical values represent
tion [19,20]. All these data strongly suggest a protective role of therefore postprandial concentrations. Plasma samples were
AMPK on vascular function. Accordingly, genetic models of frozen for biochemical determinations. Adipose tissues were
metabolic syndrome, such as Zucker diabetic fatty and Otsuka weighed and normalized by tibia length. Thoracic aortas were
Long Evans Tokushima Fatty (OLETF) rats, exhibit endothe- used for both vascular function and western blot studies. The
lial dysfunction associated to AMPK dysregulation [21, 22]. investigation conforms to the Guide for the Care and Use of
The substantial interest in the AMPK cascade as a possi- Laboratory Animals published by the US National Institute of
ble therapeutic target for the treatment of vascular complica- Health (NIH publication No. 85-23, revised in 2011) and was
tions in metabolic disorders warrants the characterization of approved by the Ethics Committee of Universidad CEU-San
the pathway, linking AMPK activation to NO release and en- Pablo (BFU2011-25303).
dothelial relaxation in whole vessels ex vivo. Moreover, there
are no studies assessing the impact of diet-induced obesity
(DIO), which better resembles human obesity, on endothe-
2.2 Plasma measurements
lial AMPK activity and vascular function. The hypothesis of
this study is that obesity induced by energy dense diets leads
Insulin was determined by using of a specific EIA kit
to a reduced AMPK activity contributing to a diminished NO
(Mercodia, Uppsala, Sweden; 1.8% intra-assay variation and
availability and endothelial dysfunction. The aim of this study
3.8% inter-assay variation) for rat insulin. Glucose was mea-
is to first characterize ex vivo the signaling pathway involved
sured by a spectrophotometric method (Glucose Trinder
in AMPK-induced NO increase, as well as the contribution of
Method, Roche, Barcelona, Spain). Triglycerides (TG) and
endothelial AMPK to relaxation in whole aortas ex vivo. Sec-
non-esterified fatty acids (NEFA) were determined using GPO
ond, we analyze if a high-fat diet (HFD) leads to alterations
(Biolabo, Maizy, France) and ACS-ACOD (Wako, Germany)
in endothelial AMPK–NO pathway that might correlate with
methods, respectively.
a reduced endothelial function.

2.3 Intraperitoneal glucose tolerance test (GTT)

2 Materials and methods
During the last week of dietary treatment, rats were fasted
2.1 Animals and dietary treatments during 6 h before glucose load (i.p. bolus of 1 g/kg at time
0). Blood glucose level was measured immediately at 0, 15,
Twelve-week-old Sprague Dawley (SD) male rats (Harlan, 30, 60, and 120 min after injection. At the indicated times,
Barcelona, Spain; 250–300 g) were housed under controlled blood samples were drawn from tail vein of conscious rats


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
Mol. Nutr. Food Res. 2014, 00, 1–13 3

for glucose determination with Accu-Check Aviva glucometer Technology), p-PI3K p85 (Tyr458 ) and PI3K p85 (1:1000
(Roche Diagnostics, Mannheim, Germany). final dilution; Cell Signaling Technology), p-Akt (Ser473 )
and Akt (1:1000 final dilution; Cell Signaling Technol-
ogy), p-eNOS (Ser1177 ) (1:800 final dilution; Cell Signaling
2.4 Vascular reactivity in the thoracic aorta artery Technology), eNOS (1:800 final dilution; BD Transduction
Laboratories, Lexington, UK), iNOS (1:5000 final dilution;
Vascular reactivity was performed in an organ bath setting as BD-Transduction Laboratories), Mn-superoxide dismutase
previously described [24]. Briefly, aortic rings of 2 mm length (Mn-SOD) and Cu/Zn-SOD (1:1000 final dilution; Santa
were given an optimal resting tension of 1.5 g, which was Cruz Biotechnology), extracellular SOD (EC-SOD, 1:1000
readjusted every 15 min during a 60-min equilibration period. final dilution; Enzo Life Sciences, USA), p47phox and p22phox
Isometric tension was recorded in a Power Lab system (ADIn- (1:500 final dilution; Santa Cruz Biotechnology) and catalase
struments, Oxford, UK). Before starting the experiment, rings (1:2000 final dilution, Sigma-Aldrich, Spain) were applied
were contracted with 75 mM KCl to assess their contractil- overnight at 4°C. After washing, appropriate secondary anti-
ity. Endothelial integrity was analyzed by addition of acetyl- bodies (anti-rabbit or anti-mouse IgG-peroxidase conjugated)
choline (ACh, 10−9 to 10−4 M) and endothelium-independent were applied for 1 h at a dilution of 1:5000. Blots were washed,
response was evaluated in presence of sodium nitroprus- incubated in commercial enhanced chemiluminescence
side (SNP, 10−10 to 10−5 M) to segments pre-contracted with reagents (ECL Prime, Amersham Bioscence, UK) and bands
phenylephrine (Phe, 10−7 M). Segments with more than 60% were detected by ChemiDoc XRS+ Imaging System (Bio-Rad,
relaxation to 10−4 M ACh were considered with endothelium California, USA). To prove equal loadings of samples, blots
(+E) and segments with less than 10% relaxation to 10−4 M were re-incubated with ␤-actin antibody (1:5000 final dilution;
ACh were considered as endothelium-free (–E). NG -nitro-L- Sigma-Aldrich, Missouri, USA). Blots were quantified using
arginine methyl ester (L-NAME, 10−4 M) and indomethacin Image Lab 3.0 software (Bio-Rad, USA). Values for p-LKB1,
(3·10−6 M) were preincubated for 20 min. p-AMPK␣, p-PI3K p85, p-Akt, and p-eNOS were normalized
Protocol 1: Basal vascular AMPK activity was estimated with LKB1, AMPK␣, PI3K p85, Akt, and eNOS, respectively.
by comparing the relaxation to ACh (10−9 to 10−4 M) in ab-
sence/presence of Compound C (10−5 M, 25 min), an AMPK
inhibitor, in aortic rings +E.
2.6 Data analyses
Protocol 2: In another set of experiments, +E and –E aor-
tic rings were pre-contracted with Phe (10−7 M), and then
All values are given as mean ± SEM and n denotes the
treated with AICAR (10−5 to 8·10−3 M), an AMPK activa-
number of animals used in each experiment. Statistical
tor. This nucleoside is taken up by the cell and accumu-
significance was analyzed by using either one-way or two-way
lates in the cytoplasm as the monophosphorylated derivative,
analysis of variance (ANOVA) followed by Bonferroni or
5 -aminoimidazole-4-carboxamide-ribonucleoside, an AMP
Newman–Keuls post hoc test for comparison between
analogue that activates AMPK without disturbing cellular
groups. A value of p < 0.05 was considered statistically
adenine nucleotide ratio [25]. In some experiments, the PI3K
significant. KCl responses are expressed as absolute values.
inhibitor wortmannin (10−7 M), the Akt inhibitor triciribine
In vascular reactivity experiments, contractions are expressed
(10−5 M), or the nitric oxide synthases (NOS) inhibitor L-
as the percentage of contraction produced by 75 mM KCl.
NAME (10−4 M) were added into the bath 25 min before Phe
Relaxations are expressed as the percentage of a previous
administration.
Phe contraction. The maximum response (EMax values)
was calculated by nonlinear regression analyses of each
individual concentration-response curve. Area under the
2.5 Western blot analysis
concentration-response curves (AUC) were calculated from
the individual concentration-response curve plots (GraphPad
Aortas from control, OP, and OR rats were isolated in PSS so-
Software, California, USA).
lution [24] and cleaned of blood and perivascular fat. In some
segments, endothelium was removed by gentle rubbing.
Western blotting was performed as previously described [26].
Briefly, 30–40 ␮g protein samples were separated by SDS- 2.7 Chemicals
PAGE gels. Primary antibodies anti-phospho-liver kinase B1
[p-LKB1 (Ser428 )] and LKB1 (1:1000 final dilution; Cell Signal- ACh was dissolved in saline and Phe and SNP in 0.01% ascor-
ing Technology, Massachusetts, USA), calcium/calmodulin- bic acid/saline (Sigma-Aldrich, USA). Indomethacin was pre-
dependent protein kinase kinase-␤ (CaMKK␤, 1:1000 final pared in ethanol and both L-NAME (Sigma-Aldrich, USA)
dilution; Santa Cruz Biotechnology, Germany), p-AMPK␣ and AICAR (Toronto Research Chemicals, Canada) in water.
(Thr172 ) and AMPK␣ (1:1000 final dilution; Cell Signaling Compound C, wortmannin, and triciribine were dissolved in
Technology), PI3K␣-110 (1:800 final dilution; Cell Signaling DMSO and were provided by Sigma-Aldrich (USA).


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
4 C. F. Garcı́a-Prieto et al. Mol. Nutr. Food Res. 2014, 00, 1–13

Table 1. Effect of HFD on tissue weights and plasma parameters

Control OP OR

Lumbar adipose tissue (g/cm) 0.53 ± 0.07 0.90 ± 0.05*** 0.66 ± 0.05##
Mesenteric adipose tissue (g/cm) 0.85 ± 0.08 1.23 ± 0.08** 0.91 ± 0.04##
Subcutaneous adipose tissue (g/cm) 0.45 ± 0.05 0.63 ± 0.05* 0.49 ± 0.04#
Liver (g/cm) 2.97 ± 0.12 2.95 ± 0.10 2.91 ± 0.05
Plasma glucose (mg/dL) 96.70 ± 5.31 98.29 ± 4.90 100.9 ± 8.41
Plasma TG (mg/dL) 54.38 ± 12.32 165.4 ± 30.61** 119.4 ± 13.81*
Plasma NEFA (mg/dL) 9.83 ± 1.39 19.14 ± 2.14** 16.24 ± 1.10**
Plasma insulin (ng/mL) 2.75 ± 0.69 3.44 ± 0.58 2.12 ± 0.34

Tissue weights are expressed in g per cm tibia length, which was not different between groups.
*
p < 0.05, ** p < 0.01, and *** p < 0.001 versus control group;
#
p < 0.05 and ## p < 0.01 versus OP group (one-way ANOVA, Newman–Keuls post hoc test). Data are expressed as mean ± SEM (n = 7–9).

3 Results However, endothelial relaxation to ACh was completely abol-

ished by L-NAME (10−4 M) in controls, but was only partially
3.1 Body weight, adiposity, and plasma parameters reduced in both OP and OR groups (Fig. 1C). The difference
between AUC in presence or absence of L-NAME, which in-
After 8 wk of HFD, BW was similar between OR and control directly indicates basal NO release, was significantly lower
rats, whereas it was significantly increased in OP rats (Sup- after HFD independently of the group (AUCcontrol = 275.12
porting Information Fig. 1A). As shown in Table 1, HFD sig- ± 3.21; AUCOP = 175.25 ± 23.13, p < 0.01; AUCOR = 165.35
nificantly increased lumbar, mesenteric and subcutaneous ± 9.52, p < 0.001), suggesting a reduction in NO availabil-
adipose tissue weight in the OP, but not in the OR group. ity. Concentration-response relaxation to ACh in presence of
These changes were not attributable to differences in food L-NAME (10−4 M) plus indomethacin (3·10−6 M) (Supporting
intake since this was similar between groups (Supporting In- Information Fig. 4) was completely abolished in control, but
formation Fig. 1B). Plasma NEFA and TG were significantly only partially reduced in both OP and OR groups. This result
higher in both OP and OR groups compared to control an- suggests the contribution of other vasodilatory mechanisms
imals (Table 1). No differences were observed in glucose or (probably of a hyperpolarizing factor) to compensate reduced
insulin levels. However, GTT shows a worse management of NO bioavailability in response to HFD.
glucose in OP and OR than in control rats (Supporting Infor-
mation Fig. 2) (AUCcontrol = 15171 ± 341.5; AUCOP = 18954
± 799, p < 0.001; AUCOR = 18042 ± 515.8, p < 0.01).
3.3 Basal AMPK was reduced in both OP and OR
groups

3.2 Endothelial function is reduced in both OP and In order to explore whether the reduction in NO availability
OR groups due to a lower NO availability was linked to AMPK activity, we characterized relaxation to
ACh (10−9 to 10−4 M) in presence of the inhibitor of AMPK,
In intact arteries, ACh (10−9 to 10−4 M) elicited a Compound C (10−5 M). We observed that response to ACh
concentration-dependent relaxation which was significantly was significantly reduced in all three groups (Fig. 2A). Quan-
reduced in both OP and OR animals compared to the con- titative AUC analysis in presence or absence of Compound
trol group (Fig. 1A). We detected a negative correlation be- C, which indirectly reflects basal AMPK activity, revealed that
tween EMax of ACh and plasma NEFA, plasma TG or AUC of difference in AUC was significantly lower in both OP and OR
GTT in all control, OP and OR animals (Supporting Informa- than in the control group (Fig. 2B).
tion Fig. 3). In contrast, endothelium-independent relaxation
elicited by SNP (10−10 to 10−5 M, Fig. 1B) as well as contrac-
tion to 75 mM KCl (control = 1.91 ± 0.06 g; OP = 2.12 ±
0.10 g; OR = 1.98 ± 0.08 g) were similar in all experimental 3.4 HFD induces endothelial but not muscular
groups, excluding alterations in contractile machinery of the AMPK down-regulation
media.
The relative contribution of endothelial factors involved Relaxation to the AMPK activator, AICAR, was characterized
in ACh-evoked relaxation was then analyzed separately (Fig. in both +E and –E arteries in order to evaluate the partial
1C). Pre-incubation with indomethacin (3·10−6 M) had no contribution of endothelial and muscular AMPK. In +E seg-
significant effect on ACh-induced relaxation in none of the ments, relaxation to AICAR (10−5 to 8·10−3 M) was signifi-
three groups (Fig. 1C), excluding prostanoids contribution. cantly higher in control arteries (EMaxcontrol = 76.49 ± 3.65%)


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
 Mol. Nutr. Food Res. 2014, 00, 1–13 5

A B
0 0
Tension (% Phe 10-7 M)

Tension (% Phe 10-7 M)

-20 -20

-40 -40

-60 -60
control control
OP
***
### OP
-80 -80
OR OR
-100 -100
-9 -8 -7 -6 -5 -4 -9 -8 -7 -6 -5
ACh, log [M] SNP, log [M]

C
control OP OR ACh
20 ***
###
ACh+L-NAME 10-4M
Tension (% Phe 10-7 M)

0 ACh+indo 3·10-6M

-20
*** ***
###
###
-40

-60

-80

-100
-9 -8 -7 -6 -5 -4 -9 -8 -7 -6 -5 -4 -9 -8 -7 -6 -5 -4
ACh, log [M] ACh, log [M] ACh, log [M]

Figure 1. (A) Cumulative concentration-response curves to ACh (10−9 to 10−4 M). ***p < 0.001 OP versus control group; ### p < 0.001 OR
versus control group (two-way ANOVA, Bonferroni post hoc test; n ࣙ 7). (B) Endothelium-independent relaxation to SNP (10−9 to 10−4
M) (n ࣙ 5). (C) Cumulative concentration-response curves to ACh with or without L-NAME (10−4 M) and indomethacin (indo, 3·10−6 M).
***p < 0.001 compared to their corresponding matched ACh curves; ### p < 0.001 compared to their corresponding indomethacin treatment
(two-way ANOVA, Bonferroni post hoc test; n ࣙ 4). All data are means ± SEM.

compared with the OP (EMaxOP = 57.47 ± 4.22%) or the OR endothelium-denuded segments, suggesting that HFD selec-
(EMaxOR = 50.25 ± 5.17%; p < 0.01) groups (Fig. 3A). En- tively altered endothelial AMPK.
dothelial removal significantly reduced AICAR-induced re- To further confirm this result, p-AMPK␣ (Thr172 )/AMPK␣
laxation in the control group, suggesting a significant con- ratios were quantified in +E (Fig. 3B) and –E arteries
tribution of endothelial AMPK to AICAR-induced relaxation. (Fig. 3C). We observed a significant reduction of p-AMPK␣
In both OP and OR animals, however, no differences were (Thr172 )/AMPK-␣ in intact arteries from OP and OR animals
observed in AICAR-induced relaxation between intact and compared to controls (Fig. 3B), whereas no difference was

A
control OP OR
0
Tension (% Phe 10-7 M)

-20

-40

-60
* *** Figure 2. (A) Cumulative concent-
ACh ***
-80
-5
ration-response curves to ACh (10−9 to
ACh+compound C 10 M
-100 10−4 M) in absence/presence of AMPK
-9 -8 -7 -6 -5 -4 -9 -8 -7 -6 -5 -4 -9 -8 -7 -6 -5 -4
inhibitor, compound C (10−5 M). *p
ACh, log [M] ACh, log [M] ACh, log [M]
< 0.05 and ***p < 0.001 compared
to their corresponding matched ACh
B curves (two-way ANOVA, Bonferroni
*
150 post hoc test; n ࣙ 5). (B) Differences
Difference between AUC

* control
between areas under concentration-
OP
100 OR response curves to ACh in pres-
ence/absence of Compound C. *p <
0.05 compared to control animals
50
(one-way ANOVA, Newman–Keuls
post hoc test; n ࣙ 5). All data are
0 means ± SEM.


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
 6 C. F. Garcı́a-Prieto et al. Mol. Nutr. Food Res. 2014, 00, 1–13

A control OP OR
0
Tension (% Phe 10 -7 M)

-20

-40
Figure 3. (A) Cumulative
concentration-response curves
-60
*** to AMPK activator (AICAR, 10−5
-80 +E to 8·10−3 M) in aorta with (+E)
-E and without (–E) endothelium.
-100 ***p < 0.001 compared to
-5 -4 -3 -2 -5 -4 -3 -2 -5 -4 -3 -2
AICAR, log [M] AICAR, log [M] AICAR, log [M] their corresponding matched
+E curves (two-way ANOVA,
Bonferroni post hoc test; n ࣙ 5).
B C
+E -E Data are means ± SEM. (B, C)
control OP OR control OP OR Representative immunoblots
p-AMPKα 62 KDa p-AMPKα 62 KDa of p-AMPK␣ (Thr172 ) in +E (B)
AMPKα 62 KDa AMPKα 62 KDa and −E aortas (C) and den-
β-acn 42 KDa β-acn 42 KDa
sitometric analysis expressed
150 150 as percentage of p-AMPK␣
% p-AMPKa/AMPKa

% p-AMPKa/AMPKa

control control
OP OP (Thr172 )/AMPK␣ in the control
100 OR 100 OR group ± SEM. **p < 0.01 and
*** ** ***p < 0.001 compared to con-
50 50 trol animals (one-way ANOVA,
Newman–Keuls post hoc test;
0 0 n ࣙ 5).

observed in –E arteries between groups (Fig. 3C). A negative duced only in the OR group (Fig. 5B), whereas neither p-PI3K
correlation was found between p-AMPK␣ (Thr172 )/AMPK␣ p85 (Tyr458 ) nor PI3K p85 (a regulatory subunit of PI3K) were
and plasma NEFA or TG levels in intact aorta from control, different between groups (Fig. 5C).
OP and OR animals (Supporting Information Fig. 5A and Relaxation to AICAR (10−5 to 8·10−3 M) in presence of
5B). triciribine (10−5 M) was reduced in the control but not in
the OP nor OR groups (Fig. 6A). A significant reduction of
p-Akt (Ser473 )/Akt was observed in intact arteries from OP
3.5 HFD induces a down-regulation of and OR animals compared to control (Fig. 6B). We detected
AMPK-kinases, LKB1, and CaMKK␤ a negative correlation between p-Akt (Ser473 )/Akt and AUC
of GTT in all control, OP and OR animals (Supporting
To determine the molecular mechanism leading to p-AMPK␣ Information Fig. 5C).
L-NAME (10−4 M) significantly reduced relaxation to
down-regulation, protein levels of upstream AMPK-kinases
were determined. Both LKB1 and p-LKB1 (Ser428 ) expression AICAR (10−5 to 8·10−3 M) in control and OP aortic segments,
was significantly reduced in OP and OR compared to control but not in the OR group (Fig. 7A). Relaxation and difference
rings (Fig. 4A). A similar result was observed for CaMKK␤ in AUC between curves in presence and absence of L-NAME
(Fig. 4B). were significantly lower between control and OP or OR group,
respectively (Fig. 7B). These results indicated a lower NO re-
lease in HFD animals. In both OP and OR groups, p-eNOS
(Ser1177 )/eNOS level was significantly reduced compared to
3.6 HFD induces a down-regulation of endothelial the control group (Fig. 7C). To further analyze the effect of
PI3K–Akt–eNOS pathway L-NAME in the OP group, a possible induction of iNOS was
assessed. The expression of iNOS was significantly higher in
To characterize the pathway connecting endothelial AMPK the OP compared to both control and OR groups (Fig. 7D),
to NO production, we designed a set of experiments in +E suggesting an up-regulation of iNOS only in OP animals.
arteries in which AICAR-induced relaxation was sequentially
analyzed with inhibitors of PI3K (wortmannin), Akt (tricirib-
ine) and eNOS (L-NAME). Moreover, the goal of these exper-
iments was to establish whether endothelial AMPK–PI3K– 3.7 HFD induces a down-regulation of mitochondrial
Akt–eNOS pathway is altered in DIO. SOD
Wortmannin (10−7 M) significantly reduced relaxation to
AICAR (10−5 to 8·10−3 M) in the control but not in the OP The expression of the mitochondrial SOD isoform, Mn-SOD,
nor OR groups (Fig. 5A). Western blot analysis revealed that was significantly reduced in both OP and OR rings (Fig. 8A).
PI3K␣-110 (a catalytic subunit of PI3K) was significantly re- However, no differences between groups were observed for


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
 Mol. Nutr. Food Res. 2014, 00, 1–13 7

control OP OR
A p-LKB1 54 KDa
LKB1 54 KDa
β-acn 42 KDa

150 control 150 control

OP OP

% p-LKB1/ b -actin
% LKB1/b -actin

OR OR
100 100
* * * *
50 50

Figure 4. Western blot analysis

0 0 of upstream AMPK-kinases LKB1
and CaMKK␤. (A) Representative
immunoblots of LKB1 and p-LKB1
B control OP OR
(Ser428 ) in +E aorta and densitometric
CaMKKβ 66 KDa analysis expressed as percentage of
β-acn 42 KDa LKB1/␤-actin and p-LKB1(Ser428 )/␤-
actin, respectively, in the control
group ± SEM. *p < 0.05 compared to
150 control
OP the control group (one-way ANOVA,
% CaMKKb /b -actin

Newman–Keuls post hoc test; n = 5).

OR
100 (B) Representative immunoblots of
CaMKK␤ in +E aorta and densitometric
** ** analysis expressed as percentage of
50 CaMKK␤/␤-actin in the control group
± SEM. **p < 0.01 compared to
the control group (one-way ANOVA,
0 Newman–Keuls post hoc test; n = 5).

the cytosolic Cu/Zn-SOD and extracellular SOD (EC-SOD) that HFD-induced endothelial impairment is associated to
(Fig. 8B and C). a down-regulation of the AMPK–PI3K–Akt–eNOS pathway
To determine if elevated superoxide levels could con- in endothelial cells probably related to high lipid levels and
tribute to reduce NO bioavailability, expression of the impaired glucose management.
NADPH oxidase subunits, p47phox and p22phox , was assessed. Here we first show that AICAR-induced relaxation in
No differences were observed between groups for both p47phox aorta of control SD rats is mediated by stimulation of both
(control = 100 ± 13.72%, OP = 99.86 ± 12.94% and OR = endothelial and muscular AMPK. Relaxation to AICAR is
86.39 ± 19.94%) and p22phox (control = 100 ± 12.46%, OP = more efficient in intact aortic rings and it seems exclusively
99.88 ± 29.97% and OR = 113.10 ± 14.40%). These results mediated by NO, suggesting a prominent contribution for
suggest that the reduction in NO bioavailability is not due to the endothelial AMPK–NO pathway to vascular relaxation.
NADPH oxidase activation. Our results are in agreement with previous studies show-
Protein levels of catalase, another important antioxidant, ing that AMPK modulates vascular function in conduit arter-
were similar between groups (control = 100 ± 4.99%, OP = ies through both endothelium-dependent and independent
88.49 ± 3.49% and OR = 101.30 ± 29.08%). mechanisms [9–11], although an endothelium-derived con-
tractile factor has been also suggested to be involved in AICAR
response [10].
4 Discussion This study also demonstrates that the endothelial AMPK–
PI3K–Akt–eNOS pathway accounts for an adequate endothe-
The present study shows for the first time that AMPK-induced lial function in whole aorta. Although previous studies show
NO increase is mediated by activation of the PI3K–Akt– that AMPK activation increases NO availability through eNOS
eNOS pathway in whole vessels ex vivo. HFD specifically phosphorylation at Ser1177 [5,6,9], the cascade downstream of
impairs endothelial AMPK-mediated relaxation through the AMPK had not been properly identified in whole vessels. To
down-regulation of this pathway, as well as AMPK-kinases rule out a potential nonspecific activity of compound C [27]
LKB1 and CaMKK␤. This down-regulation is independent of or AICAR [28], functional results were confirmed by deter-
BW gain and correlates with NEFA and TG levels, as well mining protein expression of p-AMPK␣ (Thr172 ), PI3K-␣110,
as with GTT profiles in both OP and OR rats. We propose p-Akt (Ser473 ) and p-eNOS (Ser1177 ) ex vivo. Our results are in


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
 8 C. F. Garcı́a-Prieto et al. Mol. Nutr. Food Res. 2014, 00, 1–13

A
control OP OR
0
Tension (% Phe 10 -7 M)

-20

-40

-60 ***
AICAR
-80
AICAR + wortmannin 10-7 M
-100
-5 -4 -3 -2 -5 -4 -3 -2 -5 -4 -3 -2
AICAR, log [M] AICAR, log [M] AICAR, log [M]

B C
control OP OR
control OP OR
p-PI3K p85 85 KDa
PI3Kα-110 110 KDa
PI3K p85 85 KDa
β-acn 42 KDa
β-acn 42 KDa

150 control % p-PI3K p85/PI3K p85 150 control

% PI3Ka-110/b-actin

OP OP
OR OR
100 # 100
*
50 50

0 0

Figure 5. (A) Cumulative concentration-response curves to AICAR (10−5 to 8·10−3 M) in aorta +E in presence/absence of PI3K inhibitor
(wortmannin, 10−7 M). ***p < 0.001 compared to their corresponding matched AICAR curves (two-way ANOVA, Bonferroni post hoc test; n
ࣙ 4). Data are means ± SEM. (B) Representative immunoblot of catalytic subunit of PI3K (PI3K␣-110) in +E aorta and densitometric analysis
expressed as percentage of PI3K␣-110/␤-actin in the control group ± SEM. *p < 0.05 compared to control group; # p < 0.05 compared to
OP group (one-way ANOVA, Newman–Keuls post hoc test; n = 6). (C) Representative immunoblot of regulatory subunit of PI3K (PI3K p85)
in +E aorta and densitometric analysis expressed as percentage of p-PI3K p85 (Tyr458 )/PI3K p85 in the control group ± SEM (n = 6).

accordance to previous studies in cultured endothelial cells the AMPK-kinases, LKB1 and CaMKK␤. Although we have
showing that AMPK activation stimulates eNOS phosphory- not determined ATP levels, it might be speculated that ATP
lation [5, 6] through Akt [7] or PI3K–Akt [8]. would be increased in an environment of energy surplus. In
Since AMPK functions as a fuel sensor that is activated by this context, Thors et al. [31] have demonstrated in cultured
energy depletion, it is conceivable to hypothesize that HFD- endothelial cells that AMPK is activated by both LKB1 and
induced obesity leads to AMPK down-regulation. Here we CaMKK under conditions of low ATP, and by CaMKK alone
show that DIO selectively alters endothelial-dependent relax- when intracellular ATP remains unchanged [31].
ation to AICAR in both OP and OR groups, due to a decrease Although dysregulation of vascular AMPK has been ob-
of endothelial p-AMPK␣ (Thr172 ) levels that leads to a de- served in genetic models of metabolic syndrome, such as
sensitization of PI3K␣-110, p-Akt (Ser473 ), p-eNOS (Ser1177 ), Zucker diabetic fatty [21] and OLETF rats [22], this is the
and ultimately to the impairment of endothelial function. first study demonstrating endothelial AMPK dysfunction in
PI3K␣-110 plays an important role by regulating cell prolif- response to HFD and suggesting that surplus energy input
eration and survival in heart [29] and endothelial cells [30]. leads to endothelial damage through the down-regulation of
Although an eventual decrease in PI3K␣-110 expression in the AMPK–PI3K–Akt–eNOS cascade.
OP animals under long-term HFD cannot be excluded, our The precise role played by AMPK in endothelial cell
current data only show a down-regulation of this protein in metabolism remains poorly understood. AMPK phosphory-
the OR group. This finding suggest that OR animals might lates acetyl-CoA-carboxylase and thereby increases fatty acid
show an increase in endothelial cells apoptosis and vascular oxidation [19, 20]. Moreover, it is becoming clear that en-
hypertrophy, that could contribute to relaxation impairment dothelial AMPK plays a central role in regulating both mi-
in this group. The molecular mechanism leading to lower tochondrial function and biogenesis, thus playing a pivotal
p-AMPK␣ levels seems to involve a decrease in the activity of role in oxidative stress [32]. Several studies have shown that


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
 Mol. Nutr. Food Res. 2014, 00, 1–13 9

A control OP OR
0
Tension (% Phe 10 -7 M)

-20 ***
-40

-60

AICAR
-80
AICAR + triciribine 10-5 M
-100
-5 -4 -3 -2 -5 -4 -3 -2 -5 -4 -3 -2
AICAR, log [M] AICAR, log [M] AICAR, log [M]

B control OP OR
p-Akt 60 KDa
Akt 60 KDa
β-acn 42 KDa

150 control
OP
% p-Akt/Akt

100 OR
* *
50

Figure 6. (A) Cumulative concentration-response curves to AICAR (10−5 to 8·10−3 M) in presence/absence of Akt inhibitor (triciribine, 10−5
M) in +E aorta. ***p < 0.001 compared to their corresponding matched AICAR curves (two-way ANOVA, Bonferroni post hoc test; n ࣙ 4).
Data are means ± SEM. (B) Representative immunoblots of p-Akt (Ser473 ) in +E aorta and densitometric analysis expressed as percentage
of p-Akt (Ser473 )/Akt in the control group ± SEM. *p < 0.05 compared to the control group (one-way ANOVA, Newman–Keuls post hoc test;
n = 6).

increased fatty acids availability reduces the activity of AMPK triggered by HFD would contribute to endothelial injury and
in heart and liver of rodents [33, 34]. At the vascular level, dysfunction in fatty acid enriched environments.
the decrease in AMPK activity in endothelial cells of obese In the present study, we used a modification of the rat DIO
OLETF rats parallels lipid accumulation, apoptosis and a re- model originally developed by Levin et al. [40]. This model en-
duction of NO, which are associated with the impairment ables to dissociate between factors related to HFD from obe-
of endothelial-dependent relaxation [22]. Inhibition of AMPK sity per se. SD rats fed a moderately HFD exhibit a bimodal
has been also observed in C57BL/6J mice fed a palmitate- pattern of BW gain similar to that observed in humans. Half
enriched HFD [35]. Moreover, long-term high levels of fatty of the rats gain weight rapidly compared with chow-fed rats
acids trigger ceramide-protein phosphatase 2A-dependent in- (OP), whereas the other half gain weight at a similar rate to
hibition of both AMPK and eNOS [35]. Bakker et al. [36] pro- that of the chow-fed animals (OR). Using this model, sev-
posed that increased availability of TG and long chain fatty eral studies demonstrated that OP, but not OR rats, exhibit
acyl-CoA in endothelial cells could cause oxidative stress by cardiometabolic complications, such as hypertension, hyperc-
affecting the mitochondrial respiratory chain. The dramatic holesterolemia, hyperinsulinemia, renin-angiotensin system
reduction in Mn-SOD, which is the only known scavenger activation and increased renal oxidative stress [41–44]. Here
of superoxide anion in the mitochondrial matrix, suggests we observe, however, that endothelial function is reduced
that superoxide levels would be increased locally in the mi- in OP as well as in OR groups probably due to the higher
tochondria, thus contributing to mitochondrial dysfunction. amount of fat in our diet (45% of calories from fat) compared
Although Mn-SOD is not the most expressed SOD isoform at to the original model fed with a diet containing 32% kcal
the vascular wall, it plays a critical role as first line of defense from fat. In fact, the impairment of endothelium-dependent
in protecting the mitochondria from superoxide anion pro- relaxation (EMax), as well as the reduction in p-AMPK␣
duced as by-product of the respiratory chain [37, 38]. In fact, (Thr172 ) expression, negatively correlates with the increase
reduced Mn-SOD activity in vivo correlates to both increased in NEFA and TG levels, which are similar in both OP and OR
oxidative damage and alteration in mitochondrial function groups. In this context, we suggest that the similarities be-
[39]. Our current results suggest that mitochondrial damage tween OP and OR are related to the HFD and rich fatty acids


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
10 C. F. Garcı́a-Prieto et al. Mol. Nutr. Food Res. 2014, 00, 1–13

A
control OP OR
0
Tension (% Phe 10 -7 M)

-20

-40 **
-60 ***
AICAR
-80
AICAR + L-NAME 10-4 M
-100
-5 -4 -3 -2 -5 -4 -3 -2 -5 -4 -3 -2
AICAR, log [M] AICAR, log [M] AICAR, log [M]

B C D
control OP OR
control OP OR
p-eNOS 140 KDa
iNOS 131 KDa
** eNOS 140 KDa
β-acn 42 KDa
50 * β-acn 42 KDa
Difference between AUC

control
40 150 250
OP
*
% p-eNOS/eNOS

OR 200

% iNOS/ b-actin
30
100
150
20
50
** ** 100
10 50

0 0 0
control OP OR control OP OR

Figure 7. (A) Cumulative concentration-response curves to AICAR (10−5 to 8·10−3 M) in presence/absence of L-NAME (10−4 M) in +E aorta.
**p < 0.01 and ***p < 0.001 compared to their corresponding matched AICAR curves (two-way ANOVA, Bonferroni post hoc test; n ࣙ
4). Data are means ± SEM. (B) Differences between areas under concentration-response curves to AICAR in presence/absence of L-NAME
(10−4 M) in +E aorta. *p < 0.05 and **p < 0.01 compared to control group (one-way ANOVA, Newman–Keuls post hoc test; n ࣙ 4). Data
are means ± SEM. (C) Representative immunoblots of p-eNOS (Ser1177 ) in +E aorta and densitometric analysis expressed as percentage
of p-eNOS (Ser1177 )/eNOS in the control group ± SEM. **p < 0.01 compared to control animals (one-way ANOVA, Newman–Keuls post
hoc test; n ࣙ 5). (D) Representative immunoblots of iNOS in +E aorta and densitometric analysis expressed as percentage of iNOS/␤-actin
in the control group ± SEM. *p < 0.05 compared to control animals (one-way ANOVA, Newman–Keuls post hoc test; n ࣙ 5).

environment but not to BW increase. In fact, the HFD used in In conclusion, this is the first study demonstrating
this study contains 20% lard, which is rich in palmitate [45]. In that HFD leads to endothelial impairment through the
addition, the poor management of glucose detected in our down-regulation of the endothelial AMPK–PI3K–Akt–eNOS
HFD rats could also contribute to the alterations observed, as cascade. This down-regulation is independent of BW gain and
the PI3K/Akt axis is integral to the signaling pathway linked is probably related to the rich content of the diet in palmitate
to the insulin receptor in endothelial cells [46]. Moreover, and to high blood lipid levels, although a pre-diabetic state
since HFD animals display a poor glucose management, a could also account for this dysfunction. We propose the en-
possible glycosylation of Ser1177 residue of eNOS cannot be dothelial AMPK–PI3K–Akt–eNOS pathway as a possible ther-
excluded as a cause of reduced p-eNOS (Ser1177 ) levels [47]. apeutic target for the treatment of vascular complications in
As suggested by the abolishment of AICAR-induced re- metabolic disorders related to high energy input. This path-
laxation in presence of L-NAME in the OR, but not in the way might be of special relevance since weight loss strategies,
OP group, an up-regulation of iNOS was observed in the such as diet, lifestyle, or behavioral therapy have proven rela-
aorta from this latter group. This is a mechanism probably tively ineffective in improving associated cardiovascular risk
related to BW increase and not to HFD. Indeed, increased factors.
levels of iNOS have been found in different DIO models [48]
and showed that HFD in mice increase basal vascular NO This work was supported by grants from Ministerio de
bioavailability in part derived from iNOS. In fact, KO mice Economı́a y Competitividad (BFU2012-35353, BFU2011-
for iNOS are protected against obesity-induced metabolic in- 25303), Grupos UCM (GR-921641), Fundación Universi-
sulin resistance despite a similar weight gain of wild type taria CEU-San Pablo, Fundación Mutua Madrileña, and
mice [48]. SESCAMET. CFG-P is recipient of a Ministerio de Educación,


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
 Mol. Nutr. Food Res. 2014, 00, 1–13 11

A B C
control OP OR control OP OR control OP OR
Mn-SOD 25 KDa Cu/Zn-SOD 23 KDa EC-SOD 32 KDa
β-acn 42 KDa β-acn 42 KDa β-acn 42 KDa

150 150 150

% Cu/Zn-SOD/ b -actin

% EC-SOD/b -actin
% Mn-SOD/b -actin

100 100 100

50 50 50
***
***
0 0 0
control OP OR control OP OR control OP OR

Figure 8. Western blot analysis of superoxide dismutase (SOD) isoforms. (A) Representative immunoblots of Mn-SOD in +E aorta and
densitometric analysis expressed as percentage of Mn-SOD/␤-actin in the control group ± SEM. ***p < 0.001 compared to the control
group (one-way ANOVA, Newman–Keuls post hoc test; n = 5). (B) Representative immunoblots of Cu/Zn-SOD in +E aorta and densitometric
analysis expressed as percentage of Cu/Zn-SOD/␤-actin in the control group ± SEM (n = 4). (C) Representative immunoblots of extracellular
SOD (EC-SOD) in +E aorta and densitometric analysis expressed as percentage of EC-SOD/␤-actin in the control group ± SEM (n = 4).

Cultura y Deporte fellowship. FH-N and DDR are recipients of [9] Morrow, V.-A., Foufelle, F., Connell, J.-M., Petrie, J.-R. et al.,
a CEU-Universidad San Pablo fellowship. GR-H is supported by Direct activation of AMP-activated protein kinase stimu-
Juan de la Cierva Program. The funders had no role in study lates nitric-oxide synthesis in human aortic endothelial cells.
design, data collection and analysis, decision to publish or prepa- J. Biol. Chem. 2003, 278, 31629–31639.
ration of the manuscript. [10] Ford, R.-J., Rush, J.-W., Endothelium-dependent vasore-
laxation to the AMPK activator AICAR is enhanced in
The authors have declared no conflict of interest. aorta from hypertensive rats and is NO and EDCF depen-
dent. Am. J. Physiol. Heart. Circ. Physiol. 2011, 300, H64–
5 References H75.
[11] Lee, K.-Y., Choi, H.-C., Acetylcholine-induced AMP-activated
[1] Hardie D.-G., Carling, D., Carlson, M., The AMP- protein kinase activation attenuates vasoconstriction
activated/SNF1 protein kinase subfamily: metabolic sensors through an LKB1-dependent mechanism in rat aorta. Vascul.
of the eukaryotic cell? Annu. Rev. Biochem. 1998, 67, 821– Pharmacol. 2013, 59, 96–102.
855. [12] Bradley, E.-A., Eringa, E.-C., Stehouwer, C.-D.-A., Korstjens,
[2] Kahn, B.-B., Alquier, T., Carling, D., Hardie, D.-G., AMP- I. et al., Activation of AMP-activated protein kinase by 5-
activated protein kinase: ancient energy gauge provides aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside in
clues to modern understanding of metabolism. Cell. Metab. the muscle microcirculation increases nitric oxide synthe-
2005, 1, 15–25. sis and microvascular perfusion. Arterioscler. Thromb. Vasc.
[3] Fisslthaler, B., Fleming, I., Activation and signaling by the Biol. 2010, 30, 1137–1142.
AMP-activated protein kinase in endothelial cells. Circ. Res. [13] Bosselaar, M., Boon, H., van Loon, L.-J.-C., van den Broek, P.-
2009, 105, 114–127. H. et al., Intra-arterial AICA-riboside administration induces
[4] Rubin, L.-J., Magliola, L., Feng, X., Jones, A.-W. et al., NO-dependent vasodilation in vivo in human skeletal mus-
Metabolic activation of AMP kinase in vascular smooth mus- cle. Endocrinol. Metab. 2009, 297, E759–E766.
cle. J. Appl. Physiol. (1985). 2005, 581, 1163–1171. [14] Horman, S., Morel, N., Vertommen, D., Hussain, N. et al.,
[5] Chen, Z.-P., Mitchelhill, K.-I., Michell, B.-J., Stapleton, D. et al., AMP-activated protein kinase phosphorylates and desensi-
AMP-activated protein kinase phosphorylation of endothe- tizes smooth muscle myosin light chain kinase. J. Biol. Chem.
lial NO synthase. FEBS Lett. 1999, 443, 285–289. 2008, 283, 18505–18512.

[6] Chen, Z., Peng, I.-C., Sun, W., Su, M.-I. et al., AMP-activated [15] Xie, Z., Zhang, J., Wu, J., Viollet, B. et al., Upregulation of
protein kinase functionally phosphorylates endothelial nitric mitochondrial uncoupling protein-2 by the AMP-activated
oxide synthase Ser633. Circ. Res. 2009, 104, 496–505. protein kinase in endothelial cells attenuates oxidative stress
in diabetes. Diabetes. 2008, 57, 3222–3230.
[7] Levine, Y.-C., Li, G.-K., Michel, T., Agonist-modulated regu-
lation of AMP-activated protein kinase (AMPK) in endothe- [16] Goirand, F., Solar, M., Athea, Y., Viollet, B. et al., Activation
lial cells: evidence for an AMPK/Rac1/Akt/endothelial nitric- of AMP kinase alpha1 subunit induces aortic vasorelaxation
oxide synthase pathway. J. Biol. Chem. 2007, 282, 20351– in mice. J. Physiol. 2007, 581, 1163–1171.
20364. [17] Nagata, D., Mogi, M., Walsh, K., AMP-activated protein ki-
[8] Ning, W.-H., Zhao, K., Propionyl-L-carnitine induces eNOS nase (AMPK) signaling in endothelial cells is essential for
activation and nitric oxide synthesis in endothelial cells via angiogenesis in response to hypoxic stress. J. Biol. Chem.
PI3 and Akt kinases. Vascul. Pharmacol. 2013, 59, 76–82. 2003, 278, 31000–31006.


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
12 C. F. Garcı́a-Prieto et al. Mol. Nutr. Food Res. 2014, 00, 1–13

[18] Fisslthaler, B., Fleming, I., Keserü, B., Walsh, K. et al., Fluid [32] Colombo, S.-L., Moncada, S., AMPKalpha1 regulates the an-
shear stress and NO decrease the activity of the hydroxy- tioxidant status of vascular endothelial cells. Biochem. J.
methylglutaryl coenzyme A reductase in endothelial cells 2009, 421, 163–169.
via the AMP-activated protein kinase and FoxO1. Circ. Res. [33] Ko, H.-J., Zhang, Z., Jung, D.-Y., Jun, J.-Y. et al., Nu-
2007, 100, e12–e21. trient stress activates inflammation and reduces glucose
[19] Dagher, Z., Ruderman, N., Tornheim, K., Ido, Y., The effect metabolism by suppressing AMP-activated proteinkinase in
of AMP-activated protein kinase and its activator AICAR the heart. Diabetes. 2009, 58, 2536–2546.
on the metabolism of human umbilical vein endothelial [34] Muse, E.-D., Obici, S., Bhanot, S., Monia, B.-P. et al., Role
cells. Biochem. Biophys. Res. Commun. 1999, 265, 112– of resistin in diet-induced hepatic insulin resistance. J. Clin.
115. Invest. 2004, 114, 232–239.
[20] Dagher, Z., Ruderman, N., Tornheim, K., Ido, Y., Acute regula- [35] Wu, Y., Song, P., Xu, J., Zhang, M. et al., Activation of protein
tion of fatty acid oxidation and amp-activated protein kinase phosphatase 2A by palmitate inhibits AMP-activated protein
in human umbilical vein endothelial cells. Circ. Res. 2001, kinase. J. Biol. Chem. 2007, 282, 9777–9788.
88, 1276–1282.
[36] Bakker, S.-J., Ijzerman, R.-G., Teerlink, T., Westerhoff, H.-V.
[21] Blume, C., Benz, P.-M., Walter, U., Ha, J. et al., AMP-activated et al., Cytosolic triglycerides and oxidative stress in central
protein kinase impairs endothelial actin cytoskeleton assem- obesity: the missing link between excessive atherosclerosis,
bly by phosphorylating vasodilator-stimulated phosphopro- endothelial dysfunction, and beta-cell failure? Atherosclero-
tein. J. Biol. Chem. 2007, 282, 4601–4612. sis. 2000, 148, 17–21.
[22] Lee, W.-J., Lee, I.-K., Kim, H.-S., Kim, Y.-M. et al., Alpha-lipoic [37] Madamanchi, N.-R., Moon, S.-K., Hakim, Z.-S., Clark, S. et al.,
acid prevents endothelial dysfunction in obese rats via acti- Differential activation of mitogenic signalling pathways in
vation of AMP-activate protein kinase. Arterioscler. Thromb. aortic smooth muscle cells deficient in superoxide dismu-
Vascul. Biol. 2005, 25, 2488–2494. tase isoforms. Aterioscler. Thromb. Vasc. Biol. 2005, 25, 950–
[23] Gálvez, B., de Castro, J., Herold, D., Dubrovska, G. 956.
et al., Perivascular adipose tissue and mesenteric vascu- [38] Fridovich, I., Fundamental aspects of reactive oxygen
lar function in spontaneously hypertensive rats. Arterioscler. species, or what’s the matter with oxygen? Ann. N. Y. Acad.
Thromb. Vasc. Biol. 2006, 26, 1297–1302. Sci. 1999, 893, 13–18.
[24] Steireif, C., Garcı́a-Prieto, C.-F., Ruiz-Hurtado, G., Pulido- [39] Williams, M.-D., Van Remmen, H., Conrad, C.-C., Huang, T.-T.
Olmo, H. et al., Dissecting the genetic predisposition to albu- et al., Increased oxidative damage is correlated to altered
minuria and endothelial dysfunction in a genetic rat model. mitochondrial function in heterozygous manganese super-
J. Hypertens. 2013, 31, 2203–2212. oxide dismutase knockout mice. J. Biol. Chem. 1998, 273,
[25] Corton, J.-M., Gillespie, J.-G., Hawley, S.-A., Hardie, D.-G., 28510–28515.
5-Aminoimidazole-4-carboxamide ribonucleoside: a specific [40] Levin, B.-E., Triscari, J., Sullivan, A.-C., Relationship between
method for activating AMP-activated protein kinase in intact sympathetic activity and diet-induced obesity in two rat
cells? Eur. J. Biochem. 1995, 229, 558–565. strains. Am. J. Physiol. 1983, 245, R364-R371.
[26] Somoza, B., Guzmán, R., Cano, V., Merino, B. et al., Induc- [41] Dobrian, A.-D., Davies, M.-J., Prewitt, R.-L., Lauterio,
tion of cardiac uncoupling protein-2 expression and adeno- T.-J., Development of hypertension in a rat model
sine 5 -monophosphate-activated protein kinase phospho- of diet-induced obesity. Hypertension. 2000, 35, 1009–
rylation during early states of diet-induced obesity in mice. 1015.
Endocrinology. 2007, 148, 924–931.
[42] Dobrian, A.-D., Davies, M.-J., Schriver, S.-D., Lauterio, T.-J.
[27] Bain, J., Plater, L., Elliott, M., Shpiro, N. et al., The selectivity et al. Oxidative stress in a rat model of obesity-induced hy-
of protein kinase inhibitors: a further update. Biochem. J. pertension. Hypertension. 2001, 37, 554–560.
2007, 408, 297–315.
[43] Boustany, C.-M., Bharadwaj, K., Daugherty, A., Brown, D.-
[28] Rutter, G.-A., Da Silva Xavier, G., Leclerc, I., Roles of 5 AMP- R. et al., Activation of the systemic and adipose renin-
activated protein kinase in mammalian glucose homeosta- angiotensin system in rats with diet-induced obesity and
sis. Biochem. J. 2003, 375, 1–16. hypertension. Am. J. Physiol. Regul. Integr. Comp. Physiol.
[29] Trivedi, P., Yang., Barouch L.-A., Decreased p110alpha cat- 2004, 287, R943–R949.
alytic activity accompanies increased myocyte apoptosis [44] Boustany, C.-M., Brown, D.-R., Randall, D.-C., Cassis, L.-
and cardiac hypertrophy in leptin deficient ob/ob mice. Cell A., AT1-receptor antagonism reverses the blood pressure
Cycle. 2008, 7, 560–565. elevation associated with diet-induced obesity. Am. J.
[30] Lelievre, E., Bourbon, P.-M., Duan, L.-J., Nussbaum, R.-L. Physio.l Regul. Integr. Comp. Physiol. 2005, 289, R181–
et al., Deficiency in the p110alpha subunit of PI3K results R186.
in diminished Tie2 expression and Tie2(-/-)-like vascular de- [45] Rohman, A., Triyana, K., Sismindari, Erwanto, Y., Differenti-
fects in mice. Blood. 2005, 105, 3935–3938. ation of lard and other animal fats based on triacylglycerols
[31] Thors, B., Halldórsson, H., Thorgeirsson, G., eNOS activa- composition and principal component analysis. Int. Food
tion mediated by AMPK after stimulation of endothelial cells Res. J. 2012, 19, 475–479.
with histamine or thrombin is dependent on LKB1. Biochim. [46] Zeng, G., Nystrom, F.-H., Ravichandran, L.-V., Cong, L.-N.
Biophys. Acta. 2011, 1813, 322–331. et al., Roles for insulin receptor, PI3-kinase, and Akt in


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com
Mol. Nutr. Food Res. 2014, 00, 1–13 13

insulin-signaling pathways related to production of nitric ox- dysfunction. Proc. Natl. Acad. Sci. USA. 2005, 102, 11870–
ide in human vascular endothelial cells. Circulation. 2000, 11875.
101, 1539–1545. [48] Noronha, B.-T., Li, J.-M., Wheatcroft, S.-B., Shah, A.-M. et al.,
[47] Musicki, B., Kramer, M.-F., Becker, R.-E., Burnett, A.-L., In- Inducible nitric oxide synthase has divergent effects on vas-
activation of phosphorylated endothelial nitric oxide syn- cular and metabolic function in obesity. Diabetes. 2005, 54,
thase (Ser-1177) by O-GlcNAc in diabetes-associated erectile 1082–1089.


C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com