High consumption of simple sugars causes adverse cardiometabolic effects. We investigated the mechanisms underlying the metabolic and vascular effects of glucose or fructose intake and determined whether these effects are exclusively related to increased calorie consumption. Female Sprague-Dawley rats were supplemented with 20% wt/vol glucose or fructose for 2 mo, and plasma analytes and aortic response to vasodilator and vasoconstrictor agents were determined. Expression of molecules associated with lipid metabolism, insulin signaling, and vascular response were evaluated in hepatic and/or aortic tissues. Caloric intake was increased in both sugar-supplemented groups vs. control and in glucose- vs. fructose-supplemented rats. Hepatic lipogenesis was induced in both groups. Plasma triglycerides were increased only in the fructose group, together with decreased expression of carnitine palmitoyltransferase-1A and increased microsomal triglyceride transfer protein expression in the liver. Plasma adiponectin and peroxisome proliferator-activated receptor (PPAR)-α expression was increased only by glucose supplementation. Insulin signaling in liver and aorta was impaired in both sugar-supplemented groups, but the effect was more pronounced in the fructose group. Fructose supplementation attenuated aortic relaxation response to a nitric oxide (NO) donor, whereas glucose potentiated it. Phenylephrine-induced maximal contractions were reduced in the glucose group, which could be related to increased endothelial NO synthase (eNOS) phosphorylation and subsequent elevated basal NO in the glucose group. In conclusion, despite higher caloric intake in glucose-supplemented rats, fructose caused worse metabolic and vascular responses. This may be because of the elevated adiponectin level and the subsequent enhancement of PPARα and eNOS phosphorylation in glucose-supplemented rats.
NEW & NOTEWORTHY This is the first study comparing the effects of glucose and fructose consumption on metabolic factors and aortic function in female rats. Our results show that, although total caloric consumption was higher in glucose-supplemented rats, fructose ingestion had a greater impact in inducing metabolic and aortic dysfunction.
- insulin resistance
insulin resistance, obesity, and type 2 diabetes are metabolic disturbances leading to cardiovascular diseases (CVD). CVD are main causes of morbidity and mortality in diabetes, and both micro- and macrovascular complications are thought to play a major role in the development of CVD in prediabetic and diabetic patients (14, 30).
In humans, an excessive intake of added sugars has been linked to the development of metabolic disturbances (22, 39) and therefore to an increase in the risk for CVD mortality (59). Despite the evidence generated from epidemiological studies, the molecular mechanisms linking diabetes and CVD in the population with excessive sugar intake are not fully understood. At the vascular level, insulin resistance (32, 34) along with dyslipidemia, local inflammation, and a decrease in the synthesis of endothelium-derived relaxation factors such as nitric oxide (NO) may play a key role in the development of CVD (25).
The adverse cardiometabolic effects of simple sugars seem to be worse when they are ingested in liquid than in solid form because the level of food intake is not reduced enough to compensate the extra calories provided by beverages (53), leading to an increase in total caloric intake. Consumption of simple sugars in liquid form essentially occurs as sugar-sweetened beverages (SSB), which include sodas, colas, fruit punches, lemonade, and other fruit drinks with added sugars. The main compounds used by the food industry to sweeten these beverages are high-fructose corn syrup (in the United States) and sucrose (in Europe), both containing approximately equal amounts of fructose and glucose (48). At present, there is an intense debate in the scientific community about whether the adverse cardiovascular and metabolic effects of SSB are mostly attributable to specific effects of the simple sugars used as sweeteners or are merely the consequence of the increase in caloric intake and weight gain in the population consuming large quantities of SSB (10, 26, 49, 52).
To study the effects of simple sugar consumption in liquid form on glucose and lipid metabolism, the rat is a good model that has been used extensively by us (7, 41–43, 54–56). In previous studies, we showed that female rats supplemented for short periods of time (14 days) with liquid fructose displayed a more detrimental response than male rats. Specifically, we reported that fructose induced hypertriglyceridemia and fatty liver in both sexes, but only females showed glucose intolerance and hepatic insulin resistance (41, 56). Insulin resistance is a prominent feature of metabolic diseases such as obesity or type 2 diabetes mellitus, which is also a major risk factor for CVD. A relationship between insulin resistance and endothelial dysfunction has been proposed as a link between cardiovascular and metabolic diseases (27). Endothelial dysfunction is defined as a reduced endothelium-dependent vasodilation (EDV) to vasodilators, such as acetylcholine (ACh) and bradykinin (BK), or flow-mediated vasodilation. Thus, EDV is generally used as a reproducible parameter to investigate endothelial function under various pathological conditions such as diabetes, obesity, and dyslipidemia. Whereas past studies were performed largely on males, there is increasing awareness by the National Institutes of Health (NIH) that research should include females. Here, we sought to study the effects of 2 mo supplementation with liquid fructose or glucose (20% wt/vol) on the metabolic response and vascular reactivity in aorta, a large conduit artery, in female rats. To measure the vascular function, EDV was assessed by examining the aortic relaxation responses to ACh and BK (receptor-mediated NO-dependent vasodilators). Aortic response to a NO donor was also determined by measuring the relaxation responses to sodium nitroprusside (SNP). Furthermore, vasoconstrictor responses to phenylephrine (PE) were studied. Our aims were to investigate the molecular mechanisms underlying the metabolic and vascular effects of these simple sugars and to determine whether these effects are exclusively related to increased calorie consumption.
MATERIALS AND METHODS
Animals and experimental design.
Female Sprague-Dawley rats, aged 9–11 wk (Simonsen Laboratories, Gilroy, CA), were maintained with water and standard rodent chow food ad libitum at constant humidity and temperature, with a light-dark cycle of 12 h. After acclimation for 1 wk, the animals were randomly assigned to a control group, a glucose-supplemented group, and a fructose-supplemented group (14 rats/group). Sugars were supplied as a 20% (wt/vol) solution in drinking water for 8 wk. Body weight and food and drink intake were monitored throughout the experiment. After 8 wk, the rats were fasted for 12 h and euthanized using CO2 according to the recommendations from the 2013 AVMA Guidelines on Euthanasia (28) and the NIH Guidelines for the Care and Use of Laboratory Animals (8th edition; U.S. NIH 2011). All animal protocols were approved by the Animal Care Committee of the University of the Pacific and complied with the Guide for the Care and Use of Laboratory Animals (8th edition; U.S. NIH 2011) and with ARRIVE guidelines. The 8-wk duration of sugar supplementation was chosen to mimic a subchronic regime, since this period of ingestion in rats is roughly equivalent to 6 yr of consumption in humans (45).
Glucose, triglycerides, and cholesterol were measured in 12-h-fasted rats using an Accutrend Plus System glucometer and specific test strips (Roche Farma, Barcelona, Spain) with blood collected from the tail vein. Blood samples were obtained by intracardiac puncture and collected in tubes containing anticoagulant. Plasma was obtained by centrifugation at 10,000 g for 5 min at 4°C and stored at −80°C until used. Leptin (Invitrogen, Camarillo, CA), adiponectin (Adipogen, Liestal, Switzerland), and insulin (Spi Bio, Montigny Le Bretonneux, France) levels were determined in plasma samples by ELISA kits according to the manufacturer’s protocol.
Measurement of arterial tension.
The thoracic aortas were isolated and cleaned of fatty and adhering connective tissues and then cut into 2-mm rings, exactly. To measure isometric tension, the rings were suspended horizontally between two stainless steel hooks in individual organ baths containing 20 ml of Krebs buffer (in mM: 119 NaCl, 4.7 KCl, 1.18 KH2PO4, 1.17 MgSO4, 24.9 NaHCO3, 0.023 EDTA, 1.6 CaCl2, and 6.0 glucose) at 37°C bubbled with 95% O2 and 5% CO2 (40
Relaxation responses to ACh.
Aortic rings were contracted with PE (2 μM), a concentration that produced 80% of the maximal effect (EC80). Dilator-response curves were obtained by the addition of increasing concentrations of ACh (10−8 to 10−5 M).
Relaxation responses to BK.
The concentration-response curves to BK were measured following the addition of increasing concentrations of BK (10−9 to 10−5 M) in U-46619 (100 nM)-precontracted aortic rings taken from all groups.
Relaxation responses to SNP.
Responses to SNP (10−9 to 10−5 M), a NO-donor, were generated in the aortic rings precontracted with PE (2 μM) from all groups.
Contractile responses to PE.
The constrictor concentration-response curves to PE (10−8 to 10−5 M) were generated before and after incubation with Nω-nitro-l-arginine methyl ester (l-NAME, 200 μM), a nitric oxide synthase (NOS) inhibitor, in the presence of indomethacin (10 μM, dissolved in DMSO), a cyclooxygenase (COX) inhibitor. Between each concentration-response curve, tissues were washed with Krebs buffer to allow the rings to return to the basal tone. The use of this concentration of l-NAME was based on previous studies (18, 19). A vehicle-only (no drugs present) study was performed simultaneously in aortic rings from the same animal.
Measurement of nitrate and nitrite in endothelial cells.
The permanently established EA.hy926 endothelial cell line (passage 5–6) (American Type Culture Collection, Manassas, VA) was plated and cultured in low-glucose (1,000 mg/l) Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (Thermo Scientific) and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin) at 37°C with 95% air and 5% CO2. Upon reaching 80% confluence, cells were treated with Recombinant Human gAcrp30/Adipolean (globular adiponectin, 5 and 15 μg/ml; Peprotech, Rocky Hill, NJ), high glucose or fructose (25 mM; Sigma-Aldrich, St. Louis, MO). The NO donor S-nitroso-N-acetyl-D, L-penicillamine (SNAP, 100 µM; Tocris Biosciences, Minneapolis, MN) was used as a positive control, and mannitol (25 mM; Sigma-Aldrich) was used as an osmotic control. After 30 min, cell culture supernatants were collected, and NO levels were determined by measurement of the end products of its metabolism (nitrate plus nitrite) using the Nitrate/Nitrite Colorimetric Assay Kit (Cayman Chemical, Ann Arbor, MI) according to the manufacturer’s suggestions. Five to six different samples for each condition were performed in duplicate.
RNA isolation and real-time PCR.
Samples of liver, visceral white adipose, and aortic tissues were weighed, immediately frozen in dry ice, and stored at −80°C until used for RNA extraction. Total RNA was isolated from liver and adipose tissue samples by using TrizolR reagent (Invitrogen, Carlsbad, CA) and from aortic samples using the RNeasy mini kit (Qiagen, Valencia, CA) in accordance with the manufacturer’s instructions. cDNA was synthesized by reverse transcription using the Omniscript reverse transcriptase kit (Qiagen). Samples were incubated at 37°C for 60 min in a MJ Mini Personal Thermal Cycler (Bio-Rad, Hercules, CA). The PCR reaction was carried out in the StepOnePlus Real-Time PCR System Thermal Cycling Block (Applied Biosystems, Foster City, CA) using SYBR Green PCR Master Mix (Applied Biosystems), 100 nM of each specific primer, and 10–50 ng of cDNA for each gene. PCR reactions were performed in duplicate and normalized to a housekeeping gene using the 2−ΔΔCt method. The TATA box-binding protein (tbp) was used as a control for liver, and β-actin (actb) was used for adipose tissue and aorta. Primer sequences and PCR product length are listed in Table 1.
Preparation of total protein and nuclear extracts.
Aortic and visceral adipose tissue samples were micronized through freezing with liquid nitrogen and grinding with a mortar as previously described (7). Liver samples were homogenized with a Dounce homogenizer. For total protein extraction, lysis buffer with proteases, phosphatases, and acetylase inhibitors (in mM: 50 Tris·HCl, pH = 8, 150 NaCl, 10 NaF, 1 EDTA, 1 EGTA, 2 sodium pyrophosphate, 1 PMSF, 1 Na3VO4, 10 nicotinamide, 0.001 trichostatin A) plus 1% Igepal, 2 µg/ml leupeptin, and 2 µg/ml aprotinin was used. Samples were incubated with this buffer for 1.5 h at 4°C and centrifuged at 15,000 g for 15 min at 4°C, and supernatants were collected. To obtain hepatic nuclear extracts for the determination of the transcription factors peroxisome proliferator-activated receptor-α (PPARα), carbohydrate response element-binding protein (ChREBP), and sterol response element-binding protein-1c (SREBP-1C), homogenization buffer (in mM: 10 Tris·HCl, pH = 8, 1.5 MgCl2, 10 KCl, 2 sodium pyrophosphate, 10 NaF, 0.5 DTT, 1 PMSF, 1 Na3VO4, 10 nicotinamide, and 10−6 trichostatin A) plus 2 µg/ml leupeptin and 2 µg/ml aprotinin was used. The homogenates were kept on ice for 10 min and centrifuged at 1,000 g for 10 min at 4°C. Lysis buffer was added to the pellet obtained, and samples were incubated for 1.5 h at 4°C and centrifuged at 25,000 g for 30 min at 4°C, and supernatants were collected. Protein concentrations were determined by the Bradford method (9).
Western blot analysis.
Protein (20–30 µg) was subjected to SDS-PAGE. Proteins were then transferred to Immobilon polyvinylidene difluoride transfer membranes (Millipore, Billerica, MA), blocked for 1 h at room temperature with 5% nonfat milk solution in 0.1% Tween 20-Tris-buffered saline, and incubated overnight at 4°C with primary antibodies. Primary antibodies for phosphorylated (p)-acetyl-CoA carboxylase (ACC) (Ser79), total ACC, p-V-akt murine thymoma viral oncogene homolog-2 (Akt) (Ser473), total Akt, p-endothelial NO synthase (eNOS) (Ser1177), inducible NO synthase (iNOS), cGMP-dependent protein kinase (PKG)-1 (recognizing both α- and β-isoforms), p-vasodilator-stimulated phosphoprotein (VASP) (Ser157), p-VASP (Ser239), and total VASP were supplied by Cell Signaling (Danvers, MA). Antibodies against insulin receptor substrate (IRS)-1, IRS-2, ChREBP, SREBP-1c, microsomal triglyceride transfer protein (MTP), and fatty acid synthase (FAS) were obtained from Santa Cruz Biotechnologies (Dallas, TX). PPARα, phosphodiesterase 4 (PDE4), and stearoyl-CoA desaturase-1 (SCD1) antibodies were obtained from AbCam (Cambridge, UK). Antibody against rat liver CPT-1A was kindly provided by L. Herrero and D. Serra (Department of Biochemistry and Physiology, School of Pharmacy, University of Barcelona) (21, 37). Detection was performed using the Pierce ECL Western Blotting Substrate (Thermo Scientific) and Immobilion Western Chemiluminescent HRP Substrate (Millipore, Ml). To confirm the uniformity of protein loading, blots were incubated with β-actin or β-tubulin antibodies (Sigma-Aldrich) as a control.
Data presentation and statistical analysis.
Results are expressed as the mean of n values ± SE, where n represents data from one rat. Relaxation responses to ACh, BK, and SNP were calculated as the percentage of relaxation from maximum PE or U-46619 contraction. Similarly, the recorded increase in the force of contraction was calculated as the percentage of maximum contraction obtained with PE at the highest dose. EC50, the concentration of the agonist that produces half of the maximum effect (Emax), was calculated by a sigmoidal dose-response model (for variable slope) using GraphPad Prism 6 (GraphPad Software, San Diego, CA). The sensitivity of the agonists was expressed as pD2 values (−log[EC50]). In general, statistical analyses were performed by one-way ANOVA test, after Gaussian distribution of the data was verified by the Kolmogorov-Smirnov normality test and equality of variance by Bartlett's test. When the ANOVA test returned P < 0.05, post hoc analysis using Bonferroni’s test was performed. Comparison of concentration-response curves between groups was done using two-way ANOVA, with one factor being concentration and the other being groups. To compare the effect of l-NAME on the PE response, the PE results were expressed as differences of area under the concentration response curve (ΔAUC) in control (absence of l-NAME) and experimental (presence of l-NAME) conditions. The level of statistical significance was set at P < 0.05.
Only fructose supplementation causes hypertriglyceridemia and increases body and liver weight.
As shown in Table 2, rats supplemented with 20% glucose or fructose increased their liquid consumption (by 2- and 1.5-fold, respectively). Furthermore, they partially compensated the increase in caloric intake by reducing the consumption of solid food (by 0.3-fold in the glucose group and by 0.5-fold in the fructose group). As a result, the total amount of ingested calories was increased in both groups. However, total caloric intake in glucose-supplemented rats was significantly higher than that of fructose-supplemented rats (1.15-fold). Despite this difference, only the fructose group exhibited a significant increase in final body weight (by 1.1-fold). Because of the increase in body weight, organ weights were normalized to femur length, which was not altered by sugar supplementation. Liver weight was increased only by fructose supplementation (by 1.4-fold vs. control and 1.3-fold vs. glucose-fed rats), whereas adipose tissue weight was significantly increased in both sugar-supplemented groups (by 5.6-fold in the glucose group and 5.2-fold in the fructose group). Consistent with the increase in adiposity, plasma leptin levels and adipose tissue leptin mRNA expression were also elevated in both experimental groups (Table 2).
Blood lipid analysis revealed that the cholesterol concentration was not modified in either glucose- or fructose-supplemented rats (Table 2). However, the levels of triglycerides were significantly increased in blood taken from fructose-supplemented rats (by 1.2-fold, Table 2).
Both sugars induce hepatic lipogenesis, but only fructose increases microsomal triglyceride transfer protein expression.
The hypertriglyceridemia and increased liver weight observed in fructose-supplemented rats prompted us to examine the mRNA/protein expression of lipogenesis-related enzymes. The mRNA expression of scd1 was increased by glucose (3.4-fold) and fructose (4.2-fold) (Fig. 1A). Accordingly, SCD1 protein levels were increased after glucose and fructose supplementation (by 1.4- and 1.7-fold, respectively), but the increase in the glucose group did not reach statistical significance (Fig. 1B). Although the mRNA expression of fas showed a nonsignificant tendency to increase in both groups (Fig. 1C), FAS protein levels were significantly increased by 2-fold in glucose-supplemented and by 2.4-fold in the fructose-supplemented group (Fig. 1D). The amount of total and phosphorylated ACC was significantly increased in the liver by both sugars (Fig. 1, E and F), suggesting that the degree of ACC activation was not modified. The expression of these lipogenic genes is mainly controlled by two transcription factors, SREBP-1c and ChREBP. Although the amount of ChREBP in nuclear extracts remained unaffected (Fig. 1G), the protein expression of the mature form of SREBP-1c was significantly increased only in hepatic nuclear extracts taken from fructose-supplemented rats (Fig. 1H).
As shown in Table 2, only glucose-supplemented animals displayed higher plasma adiponectin levels than controls (2.6-fold increase), which correlates with increased mRNA expression of the adiponectin gene in adipose tissue of this group (by 2.6-fold, Table 2). It has been shown that stimulation of adiponectin signaling activates PPARα, a key factor in the transcriptional control of genes encoding fatty acid β-oxidation enzymes (29). Western blot analysis revealed that the expression of hepatic PPARα was significantly increased only in the nuclear extracts from glucose-supplemented rats (by 1.3-fold) (Fig. 2A). Accordingly, the mRNA levels of PPARα target genes, such as liver carnitine palmitoyl-CoA transferase-I (l-cpt-1a) and acyl-CoA oxidase (aco), were significantly increased only in the liver of glucose-supplemented animals (Fig. 2, B and C). When we determined L-CPT-1A protein levels, we found no significant difference between the glucose-supplemented and the control group. However, the protein level of L-CPT-1A was significantly reduced in liver samples taken from fructose-supplemented rats (by 0.5-fold, Fig. 2D). Because the amount of plasma triglycerides not only depends on the balance between the synthesis and catabolism of fatty acids in the liver, but also on the export of triglycerides from hepatocytes, we also determined the protein expression of MTP. This protein is essential for the assembly and secretion of very-low density lipoproteins (VLDL) by the liver. As shown in Fig. 2E, MTP protein expression was increased only in fructose-supplemented rats.
Supplementation with fructose and glucose impairs insulin signaling in aorta and liver.
Plasma glucose concentrations were not altered after simple sugar supplementation (Table 2). On the other hand, fructose increased plasma insulin concentration (by 2.4-fold), and the glucose-supplemented group showed a tendency (P = 0.07) to higher plasma insulin level. Thus, the insulin sensitivity index (ISI) was significantly reduced in both glucose- and fructose-supplemented rats (Table 2), suggesting that insulin signaling could be impaired in these groups. Therefore, we analyzed the expression of IRS-1 and IRS-2, the main insulin signal transducers in the liver and in aortic tissue. Our results show that the protein expression of hepatic IRS-1 was not altered by sugar supplementation (Fig. 3A), but IRS-2 expression was significantly reduced in both groups (by 0.5-fold in glucose-supplemented and by 0.4-fold in fructose-supplemented animals, Fig. 3B). On the other hand, both IRS-1 and IRS-2 protein levels were reduced in aortic tissue from fructose- and glucose-supplemented rats (Fig. 3, C and D). We also determined the expression of Akt, one of the main transducers of insulin signaling downstream of IRS in both tissues. The amount of p-Akt was significantly reduced only in the livers of fructose-supplemented rats compared with the control group, whereas total Akt protein remained unaffected by both sugars (Fig. 3E). Similarly, in aortic tissues, the expression of p-Akt was significantly reduced only in fructose-fed animals, and total Akt expression was not significantly modified by sugar supplementation (Fig. 3F).
Supplementation with glucose and fructose alters relaxation responses of aortic rings to a NO donor.
Relaxation to ACh was used to examine the effect of simple sugars on the receptor-mediated endothelium-dependent release of NO. No significant differences in responses to ACh (10−8 to 10−5 M) occurred between aortic rings from the control, glucose-, and fructose-supplemented groups (Fig. 4A). Both the sensitivity of aortic rings to ACh as assessed by −log[EC50] (pD2) and the Emax to ACh were similar in all groups (Fig. 4A and Table 3). Similar to the findings for ACh, the sensitivity and Emax to BK, another receptor-mediated endothelium-dependent vasodilator agent, were not affected by sugar supplementation (Fig. 4B and Table 3). However, there was a significant rightward shift of the BK response curve in aortic rings from fructose-supplemented rats relative to control animals (Fig. 4B).
Aortic relaxation responses to a NO donor were also investigated by performing concentration-response curves to SNP (10−9 to 10−5 M). Although the Emax to SNP was similar in all groups (Table 3), supplementation with glucose significantly augmented the responsiveness of aortic rings to SNP. On the other hand, fructose supplementation significantly attenuated the sensitivity of aortic rings to SNP (Fig. 4C and Table 3).
Glucose (but not fructose) supplementation reduces the contractile response of aortic rings.
Eight weeks of fructose supplementation did not affect vasoconstrictor responses of rat aorta to PE, as shown in Fig. 5A and Table 3. However, PE-induced maximal contractions were reduced significantly in aortic rings from glucose-supplemented rats compared with controls (Fig. 5A and Table 3). Despite that PE contractile responses were affected by glucose supplementation, the mRNA expression of α1-adrenergic receptors was unchanged (Table 4). To study the possible role of basal NO, PE-response curves were performed in aortic rings before and after pretreatment with the NO synthase inhibitor l-NAME (200 μM, 20 min) in the presence of indomethacin (10 μM). The difference in the contractile level to PE after addition of l-NAME would indicate the extent of endothelial NO release (13, 20). Incubation of the aortic rings with l-NAME resulted in a significant increase of the contractile responses to PE in all groups (Fig. 5, C-E). However, the ΔAUC, defined as difference in the area under the curve between the PE-concentration response curve before and after l-NAME, were only higher in aortas of the rats supplemented with glucose compared with the control group (Table 5).
To investigate a mechanism by which NO release might have been increased in glucose-supplemented rats, eNOS activation by phosphorylation was investigated. As shown in Fig. 6A, the phosphorylation of eNOS at Ser1177 was increased in aortic tissue from glucose- but not fructose-supplemented rats, whereas the expression of total eNOS showed no significant difference between the simple sugar-supplemented groups and the control group.
On the other hand, protein levels of iNOS in aorta were increased significantly only in the fructose-supplemented animals (by 1.8-fold) (Fig. 6B).
In vitro adiponectin (but not glucose or fructose) treatment increases the NO level in EA.hy926.
It is well known that adiponectin stimulates the synthesis of NO in endothelial cells. Thus, the increase in adiponectin levels observed in glucose-supplemented rats could be responsible for the elevated basal NO in this group. To test this hypothesis, we determined the effects of adiponectin or high glucose or fructose on NO level in endothelial EA.hy926 cells. As shown in Fig. 7, incubation of EA.hy926 cells with 25 mM glucose or fructose for 30 min did not alter cellular NO production, whereas incubation with 5 and 15 µg/ml adiponectin dose dependently (and significantly for the highest concentration) increased NO levels in the cell culture supernatant.
Differential effects of fructose and glucose supplementation on intracellular pathways related to aortic relaxation.
To study the molecular mechanisms that could explain the differences in the responses of aortic rings from rats supplemented with glucose or fructose, the expression of several molecular mediators of vascular relaxation were examined. The NO-dependent relaxation is mediated by cGMP signaling, which leads to the stimulation of a specific PKG, resulting in the phosphorylation of VASP in Ser239 (44). Our results show that the mRNA expression of soluble guanylate cyclase (gcsa1) in aortas from fructose-supplemented rats was reduced by 0.8-fold compared with controls (Fig. 8A). Neither the mRNA levels of the phosphodiesterase that hydrolyses cGMP (pde5) (Fig. 8A) nor the protein expression of PKG (Fig. 8B) was significantly modified by sugar supplementation. However, phosphorylation of VASP in Ser239 was reduced in both sugar-supplemented groups vs. control (Fig. 8C), and the amount of total VASP also showed a reduction, which was significant only in fructose-supplemented rats (Fig. 8D).
The NO-independent vasodilatory pathway in large conduit arteries is mainly controlled by cAMP levels resulting in the phosphorylation of VASP in Ser157 (44). Aortic tissue adenylyl cyclase (AC6) mRNA levels and phosphorylation of protein kinase A (PKAc) in Thr197 were not significantly modified by sugar supplementation compared with control rats, although fructose-supplemented rats showed reduced values compared with the glucose group (Fig. 8, A and F). Fructose supplementation significantly increased the mRNA and protein expression (by 1.9- and 1.5-fold, respectively) of PDE4, which specifically catalyzes the hydrolysis of cAMP (Fig. 8, A and G). Thus, the combined reduction of total VASP and the putative decrease of cAMP levels may explain the significant reduction of VASP Ser157 phosphorylation (by 1.46-fold) that was observed only in fructose-supplemented rats (Fig. 8E).
Finally, to rapidly identify possible changes in other pathways involved in the vascular effects of simple sugars, we measured the mRNA level of a broad range of genes associated with vascular function. As shown in Table 4, the mRNA expression of genes and enzymes such as angiotensin II and BK receptors, cyclooxygenases, prostacyclin synthase, or thromboxane synthase, which have been reported to be involved in aortic reactivity, remained unaffected by sugar supplementation. Although we are aware that mRNA expression and protein abundance are not always well correlated, the lack of effect of sugar supplementation on the mRNAs analyzed discouraged us from further studying the involvement of these pathways on the observed vascular effects.
The adverse effects of simple sugar consumption on metabolic function are well documented in human and animal studies (3, 49, 51), and there are also reports on the effects of diets supplemented with fructose or high-fructose corn syrup on vascular reactivity in rodents (2, 5, 16, 31, 38). Most of these studies were performed on males, but there is increasing awareness to include females in research studies. To our knowledge we are the first to compare the effects of simple sugar (glucose or fructose) consumption on female rat metabolic factors and aortic function. Here, we identify some key molecular targets responsible for metabolic and vascular disturbances induced by simple sugar supplementation in liquid form to female rats. We also provide evidence that these effects are not merely a direct result of the amount of ingested calories.
We previously reported that the ingestion of liquid fructose elicits an incomplete compensatory reduction of solid food consumption in rats, leading to an increase in total caloric intake (7, 41–43, 54, 55). Our current data show that, when rats are supplemented with a glucose solution, the compensatory response is even worse compared with fructose supplementation, since liquid intake is higher and comparatively solid food consumption is reduced to a lesser extent. As a result, the total caloric intake is higher in rats receiving glucose than in rats with fructose supplementation (Table 2). It has been suggested that, if simple sugars cause any adverse effects on health, these effects are exclusively because of the calories that the sugars provide (26). According to this hypothesis, the metabolic and vascular alterations in the present study should have been equal or more intense in glucose-supplemented rats. On the contrary, our results show that, compared with the fructose group, rats supplemented with liquid glucose exhibit less adverse metabolic and vascular effects. Analysis of the results shown in Table 2 and Fig. 3 indicates that both sugars impair insulin signaling in the liver and aortic tissue, but the effect is far more intense in fructose- than in glucose-supplemented rats. Although hepatic and aortic IRS-1/2 were reduced to a similar extent by glucose and fructose, p-Akt was significantly decreased in the liver and aorta of fructose-supplemented animals only. This is in accordance with our previous results showing an abnormal glucose tolerance test and reduced p-Akt in female rats supplemented with 10% liquid fructose, but not glucose, for 2 mo (6). In the current study, insulin resistance is observed not only in the liver but also in aortic tissues of fructose-supplemented animals, suggesting a common origin for both metabolic and vascular dysfunction.
Moreover, our data show that the hypertriglyceridemia appears only in fructose-supplemented rats. This observation cannot be explained solely by increased hepatic lipogenesis, since both sugars induced the expression of major lipogenic genes. However, fructose supplementation induced a more prominent effect (Fig. 1). Hepatic lipogenesis is controlled by two transcription factors, SREBP-1c and ChREBP (57). Because ChREBP is activated by phosphorylated intermediates derived from the direct metabolism of dietary carbohydrates (1), the 12-h fasting period in our study explains the lack of activation of this transcription factor (Fig. 1G). On the contrary, activation of SREBP-1c by carbohydrates is not a direct effect but it is mediated by an increase in plasma insulin levels (57), a mechanism that is maintained in insulin resistance states (24). In the present study, plasma insulin level was markedly elevated only in fructose-supplemented animals (Table 2), driving the increased expression of SREBP-1c in hepatic nuclear extracts from this group (Fig. 1H).
Regarding lipid catabolism, fructose supplementation did not affect hepatic PPARα nuclear content or the mRNA expression of l-cpt-1a, a PPARα target gene. However, when we determined the protein expression of L-CPT-1A, we observed a significant decrease in the liver of the fructose-supplemented rats (Fig. 2D). L-CPT-1A catalyzes the rate-limiting step of the hepatic mitochondrial β-oxidation of fatty acids, suggesting that this process may be inhibited in the liver by fructose supplementation. On the contrary, fructose increased the hepatic expression of MTP, a protein that is essential for the assembly of triglycerides into VLDL and for the secretion of these lipoproteins (17). This effect had been already reported in fructose-fed hamsters, and was correlated to hepatic insulin resistance (11, 50). Taking these results together, the hypertriglyceridemia observed in the fructose-supplemented rats may arise from the combined effects of increased lipogenesis, reduced fatty acid catabolism, and enhanced triglyceride export from liver to plasma in the form of VLDL through MTP induction.
On the other hand, plasma adiponectin level is increased in glucose-supplemented rats, which correlates with increased mRNA expression of adiponectin in adipose tissue of this group (Table 2). Although the mechanism responsible for this inductive effect is unclear, our results suggest that the overexpression of hepatic PPARα in this group may be because of the elevated level of adiponectin. This theory is supported by other reports showing that adiponectin upregulates PPARα expression (58). It should be noted that, despite PPARα induction and elevated l-cpt-1a mRNA in the glucose-supplemented group, we did not observe increased hepatic CPT-1A protein expression, suggesting that fatty acid oxidation is not induced by glucose supplementation. Overall, our results on the decreased L-CPT-1A protein levels in fructose-supplemented rats suggest that simple sugars might affect L-CPT-1A protein stability, causing a reduction in its hepatic content. However, this effect is counteracted in glucose-supplemented rats, possibly via the control of its expression by increased PPARα.
It has been shown that adiponectin stimulates NO production in vascular endothelial cells (12), and therefore the enhanced basal NO in aortae of glucose- supplemented rats (Table 5) may result from an increased adiponectin in this group. In the current study, we assessed the in vitro effects of adiponectin, high glucose, or fructose on NO level by measuring NO metabolites. Treatment of EA.hy926 with adiponectin but not high glucose or fructose caused a significant increase in NO level in these cells (Fig. 7). These results exclude a direct effect of glucose on NO production and suggest that elevated adiponectin is the possible cause of increased NO in aortas of glucose-supplemented rats. Moreover, it has been reported that adiponectin stimulates the synthesis of NO in endothelial cells by phosphorylation of eNOS at Ser1177 (12, 15), and our results show that eNOS phosphorylation at this position is increased in aortic tissues from glucose-supplemented rats (Fig. 6A). Taken together, these results suggest that the increase in basal NO after glucose supplementation could be mainly attributed to the hyperadiponectinemia observed in this group.
Elevated basal NO in aortae from glucose-supplemented rats may also, in part, explain decreased vasoconstrictor response to PE in this group (Fig. 5B). In addition, the fact that the SNP-induced relaxation was increased in the glucose-supplemented group suggests that the increased sensitivity of intact aorta to NO may also contribute to the decreased PE contractile responsiveness. It is important to note that glucose supplementation enhanced aortic responses to SNP despite a decrease in phosphorylation of VASP at Ser239 and preserved PKG expression, suggesting that VASP phosphorylation or upstream cGMP-dependent phosphorylation of VASP is not a mechanism for the increased relaxation response to SNP. Along similar lines, Aszódi et al. observed an intact relaxation after exposure to cGMP and cAMP in aortic rings of VASP null mice (4). Furthermore, Yousif et al. (60) reported that SNP induces cGMP-independent vasodilator responses in the perfused rabbit ovarian vascular bed. Unlike the glucose-supplemented group, our data show that SNP-induced relaxation is significantly reduced in aortic rings of the fructose-supplemented rats. Assessments of vasodilatory responses in the presence of a soluble guanylate cyclase inhibitor or a PKG inhibitor/activator could help to determine how fructose and glucose differentially influence the NO/cGMP/PKG relaxation pathways.
EDV is used as a reproducible parameter to investigate endothelial function under various pathological conditions. In the present study, we showed a preserved EDV of the aorta in the sugar-supplemented rats despite the altered relaxation responses to SNP. Accordingly, Mourmoura et al. (33) reported that the EDV of coronary arteries was fully maintained while the response to a NO donor was even enhanced in type 2 diabetic rats. Both impairment (16, 31, 38) or no change (23, 36) of EDV in vessels from fructose-supplemented rats have been reported. The varied vasodilatory responses after sugar supplementation may be attributed to differences in the type of the agents used, animal’s sex and strain, vascular beds, and duration of sugar supplementation.
NO has been generally considered as the principal mediator of EDV in normal state in large arteries, and impaired EDV is often associated with reduced bioavailability of NO. However prostaglandin I2 (PGI2) and endothelium-derived hyperpolarizing factor (EDHF) may also be important regulators of vascular tone and reactivity in diabetes. There is an established negative regulatory effect of NO on EDHF synthesis (8). On the other hand, an augmented EDHF response was shown to compensate for the loss of NO-mediated vasorelaxation in arteries in diabetic rats (46). In agreement with those studies that demonstrate compensatory interactions between pathways, the preserved ACh response (regardless of altered SNP responses) suggests that other molecules besides NO (e.g., EDHF or PGI2) may be involved in ACh relaxation. The fact that AC6 mRNA and phosphorylated PKA protein levels were significantly reduced in aorta taken from fructose rats (Fig. 8. A and F) suggests that the preserved ACh responses in the arteries of this group are likely the result of the elevated EDHF (rather than increased PGI2). Furthermore, we demonstrated that fructose but not glucose supplementation enhances PDE4 mRNA and protein expression (Fig. 8, A and G). The elevated PDE4 expression and subsequent reduction in cAMP levels may specifically hinder phosphorylation of VASP at Ser157 (Fig. 8E). This effect combined with reduced VASP expression (Fig. 8D) possibly leads to an impairment of NO-independent relaxation in aortas of fructose-supplemented rats.
Another mechanism that could account for the difference between the response to ACh and SNP could be oxidative stress. Reactive oxygen species (ROS), particularly superoxide anions, inactivate NO, which could lead to a reduction of SNP potency. To examine this possibility, we measured the mRNA expression of the catalytic subunits of NADPH oxidase Nox1 and Nox4, major source of ROS in vascular cells (19). However, we did not see significant differences in the mRNA expression for Nox1 and Nox4 in the aorta of experimental animals (Table 4).
It has been shown that under metabolic stress conditions iNOS is able to produce an abnormal amount of NO leading to ROS production and decreased bioavailability of NO. Although we did not measure ROS production and NO bioavailability, we observed a significant induction of iNOS protein expression in aortic tissues from fructose-supplemented rats (Fig. 6B), which could in part lead to vascular dysfunction in this group (35). Vascular dysfunction in metabolic syndrome may also be associated with increased vasoconstrictor sensitivity (47). Here, however, we showed that fructose supplementation did not affect vasoconstrictor responses to PE. Along similar lines, no differences were reported in the vasoconstrictor responses to PE in mesenteric arteries (36) or aortic rings (23) from fructose-fed rats.
In conclusion, we have shown that dietary supplementation with liquid glucose or fructose causes metabolic and vascular alterations in female rats. Despite higher caloric intake in glucose-supplemented rats, fructose caused worse metabolic and vascular responses. This may be because of the elevated adiponectin level and the subsequent enhancement of PPARα and eNOS phosphorylation in glucose-supplemented rats, a mechanism that is absent in the fructose group. Clearly, additional studies will be needed to document the direction and magnitude of these interactions in sugar-supplemented rats along with the relative importance of elevated adiponectin level to metabolic and vascular function in the glucose-supplemented rats.
This study was supported by grants from the Fundació Privada Catalana de Nutrició i Lípids, SAF2013–42982-R, and European Commission FEDER funds. G. Sangüesa was supported by a Formación de Profesorado Universitario grant from the Spanish Ministry of Science and Innovation and by an Agustí Pedro i Pons Foundation grant for postgraduate studies. We are a Consolidated Research Group of the Autonomous Government of Catalonia (SGR13–00066). This work was also supported in part by National Heart, Lung, and Blood Institute Grant HL-28988 to R. Rahimian and the University of the Pacific.
The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. The authors have no financial or other conflicts of interest to disclose.
G.S. and S.S. were in charge of all experiments; F.A. contributed to vascular reactivity experiments; NR contributed to PCR/Western blot experiments and prepared the figures; J.C.L helped in data interpretation and reviewed the manuscript; R.R and M.A designed the experiments, analyzed the data and wrote the manuscript.
We thank Dr. Xiaoyuan Han for vascular expertise and Sereena Nand, the ACS Project SEED recipient, for assistance with the project. We also thank L. Herrero and D. Serra for the gift of the CPT-1A antibody.
- Copyright © 2017 the American Physiological Society