Repeated Glucose Spikes and Insulin Resistance Synergistically Increase Endothelial Function Vulnerability to High Glucose Levels through Redox Imbalance, and Bardoxolone Methyl (CDDO-Me) Ameliorates Endothelial Dysfunction

Background Glucose spikes (GSs) observed after a meal in metabolic syndrome have been reported to cause endothelial dysfunction. However, other insulin resistance-related factors can affect GS-induced endothelial dysfunction. To eliminate these confounding factors, we investigated the separate and combined effects of GSs and insulin resistance due to diet-induced obesity on endothelial function and claried whether bardoxolone methyl (CDDO-Me), a novel nuclear factor erythroid 2-related factor 2 (Nrf2) activator, protects against GS-induced endothelial dysfunction. In the rst cohort, eight-week-old male Wistar rats were assigned to one of four groups: 1) control diet (CD)-GS (-); 2) CD-GS (+); 3) Western-type diet (WTD)-GS (-); and 4) WTD-GS (+). Rats were fed a CD or WTD for 13 weeks and intraperitoneally injected with saline or glucose for 1 week twice daily at 20 weeks of age. In the second cohort, four groups from the rst cohort were additionally divided into vehicle and CDDO-Me (3 mg/kg) groups. At 21 weeks of age, endothelial function was evaluated using isolated thoracic aortas under normal (5.5 mM) and high-glucose (20 mM) conditions. Gene expression was analyzed, and superoxide anion was evaluated by dihydroethidium (DHE) staining of aortas. aortas, and DHE intensity was enhanced. In the second cohort, pretreatment of the WTD-GS (+) group with CDDO-Me attenuated this endothelial dysfunction accompanied by a correction of redox imbalance in gene expression and an attenuation of DHE intensity. We demonstrated that GSs and insulin resistance synergistically increased endothelial function vulnerability to high-glucose levels through redox imbalance, although each factor alone had little effect on endothelial function. Furthermore, we showed that pretreatment with CDDO-Me ameliorated endothelial dysfunction caused by GSs in metabolic syndrome model rats. have the potential to treat endothelial dysfunction due to GSs by attenuating oxidative stress. In the present study, to explore the conditions in which endothelial function is most vulnerable to GSs, we investigated two factors separately, GSs and insulin resistance, and the combined effects of these two factors on vascular endothelial function. Specically, we evaluated endothelial function in nonobese rats results following: 1) diet-induced obesity by a WTD downregulates that of SOD2 and catalase; 2) short-term repeated GSs downregulate the mRNA expression of catalase; and 3) the combination of repeated GSs and diet-induced obesity synergistically upregulates that of NOX2 and p47phox and downregulates that of SOD2. With regard to inammatory genes, the mRNA expression of TNFα was signicantly different between diet or GS factors (F [1, 20] = 5.26, P = 0.033; F [1, 20] = 17.99, P < 0.001, respectively) and that of IL1β also different between GS factors (F [1, 20] = 9.17, P = 0.007) without diet/GS interaction (Fig. 3). An interactive effect of diet and GS on the mRNA expression of VCAM1 was detected (F [1, 20] = 12.0, P = 0.002), and simple-effects analysis the combination of a WTD and GS (+) upregulated that of VCAM1 rats with insulin resistance made endothelial function vulnerable to high-glucose conditions. Mitochondria-derived superoxide may be increased and induce endothelial dysfunction under high-glucose conditions against the background of changes in gene expression related to redox balance and inammation in the aortas from rats with repeated GSs and insulin resistance. Repeated GSs and free fatty acids, an insulin resistance-related factor, may synergistically cause an imbalance in redox gene expression in the aorta. The administration of CDDO-Me corrected these changes in gene expression and protected against endothelial dysfunction caused by repeated GSs and insulin resistance.

and diet-induced obese rats that received glucose intraperitoneally twice a day for one week to reproduce GSs without postprandial hypertriglyceridemia. We also investigated whether CDDO-Me prevents GS-induced endothelial dysfunction.

Methods Animal Studies
Seven-week-old male Wistar rats were obtained from KBT Oriental Co., Ltd. (Saga, Japan) and housed at 23 ± 1°C under a 12-h light/12-h dark cycle, with ad libitum access to food and water. After a 1 week acclimation period, the rats were fed a control diet (CD) or a Western-type diet (WTD) for 13 weeks (8 -21 weeks old) according to their group. The CD contained 4.7% calories from fat, 23.3% calories from protein, and 55.6% calories from carbohydrates (3.6 kcaL/g; MF, Oriental Yeast Co., Ltd., Tokyo, Japan), and the WTD contained 39.9% calories from fat, 15% calories from protein, and 44.3% calories from carbohydrates (4.5 kcaL/g; F2WTD, Oriental Yeast Co., Ltd., Tokyo, Japan). The WTD group of rats were allowed access to feed ad libitum, and the CD group of rats were pairfed (limited to the amount of food consumed by the rats in the WTD group) to prevent obesity due to excessive intake.
In the rst cohort, the rats were assigned to one of four groups according to the factor of diet or glucose spike (GS): 1) CD-GS (-), control diet and glucose spike (-); 2) CD-GS (+), control diet and glucose spike (+); 3) WTD-GS (-), Western-type diet and glucose spike (-); and 4) WTD-GS (+), Western-type diet and glucose spike (+). The CD group represented nonobese model rats without insulin resistance, and the WTD group represented obese model rats with insulin resistance.
At 20 weeks of age, rats in the GS (-) and GS (+) groups were intraperitoneally injected with saline (5 mL/kg) and 20% glucose (1 g/5 mL/kg body weight [BW]), respectively, for one week twice daily at approximately 08:00 a.m. and 04:00 p.m. (Supplementary Fig. S1A for details). In the second cohort, four groups from the rst cohort were additionally divided into vehicle and CDDO-Me groups. At 19 weeks of age, rats in the vehicle and CDDO-Me groups were orally administered vehicle (sesame oil, S3547, Sigma-Aldrich, St. Louis, MO, USA) or CDDO-Me (3 mg/kg BW; SMB00376, Sigma-Aldrich, St. Louis, MO, USA), which was solubilized in sesame oil, for two weeks once daily (Supplementary Fig. S1B for details). At 20 weeks of age, rats in the GS (-) and GS (+) groups were intraperitoneally injected with saline and 20% glucose for one week twice daily as described for the rst cohort. For the insulin tolerance test (ITT), rats were deprived of food for 12 hours and then injected intraperitoneally with insulin (0.5 U/kg BW), and insulin resistance was evaluated as the decreasing blood glucose area under the curve (AUC). To collect additional samples, rats were deprived of food overnight (14 hours) in cages with fresh bedding. Visceral fat mass was evaluated as the adiposity index, which was de ned as the ratio of epididymal, retroperitoneal and mesenteric fat grams to body weight [26]. Serum Ltd., Tokyo, Japan). Homeostasis model assessment of insulin resistance (HOMA-IR) was determined with the formula HOMA IR = serum insulin (mmol/L)* (blood glucose (mmol/L)/22.5 [27]. All animal protocols were reviewed and approved by the Laboratory Animal Committees of Kagoshima University Graduate School and were performed in accordance with the guidelines for the care and use of laboratory animals (approval number: MD20086).

Continuous Interstitial Glucose Monitoring
When starting saline or glucose at 20 weeks of age, a FreeStyle Libre Pro ® sensor (Abbott Diabetes Care, IL, USA), which continuously records interstitial glucose levels every fteen minutes, was attached to the backs of the rats to record interstitial glucose levels for two days. After removing the sensor, the data were extracted using FreeStyle Libre Pro software (Abbott Diabetes Care). A GS was de ned as an increased interstitial glucose level above 5 mM, which was the difference between baseline and peak interstitial glucose levels.

Statistical Analysis
Values are presented as the mean ± SEM. Statistical signi cance was determined by using one-way ANOVA to compare differences between groups. When diet and GS interaction effects were evaluated as dependent variables, two-way between-groups ANOVA was used; a signi cant interaction was interpreted by a subsequent simple-effects analysis with Bonferroni correction. Differences between vehicle and CDDO-Me were evaluated by using Bonferroni correction. Concentration response curves and body weight curves were analyzed by one-way or two-way repeated-measures ANOVA followed by the Bonferroni post hoc test. Univariate regression analysis using Pearson's correlation coe cient was performed to assess statistical associations between metabolic parameters and pD2. The differences between groups were considered signi cant when P < 0.05. All data were analyzed with R version 3.6.1 statistical software (The R Foundation for Statistical Computing, Vienna, Austria).

Results
The GS Model Was Con rmed with Continuous Interstitial Glucose Monitoring To con rm the glucose pro le in the model rats, interstitial glucose levels were continuously recorded with Libre ® for 48 hours from the 2nd day of administration of saline or glucose. All incremental interstitial glucose levels were above 5 mM in both the CD-GS (+) and WTD-GS (+) groups after glucose administration and were below 5 mM in both the CD-GS (-) and WTD-GS (-) groups after saline administration (Supplementary Fig. S2A and B). A signi cant main effect of GS was observed in both peak and incremental interstitial glucose levels (F [1, 12] = 1619.1, P < 0.001; F [1, 12] = 513.7, P < 0.001, respectively) without diet/GS interaction, although the main effect of diet on those was not signi cant.

Short-term Repeated GSs Deteriorated EDR in Diet-induced Obese Rats
To investigate the difference in EDR between the CD group (nonobese rats) and WTD group (obese rats) with or without GSs, we compared endothelial function among four groups: CD-GS (-), CD-GS (+), WTD-GS (-) and WTD-GS (+). The EDR of the thoracic aortas from the CD-GS (-), CD-GS (+) and WTD-GS (-) groups did not change under normal (5.5 mM) or high-glucose (20 mM) conditions, but the EDR of those from the WTD-GS (+) group deteriorated under high-glucose conditions ( Fig. 2A). This deterioration of EDR was not reproduced in the presence of 20 mM ra nose, an osmotic control for 20 mM glucose (Fig.  1A). Among the four groups under 20 mM glucose conditions, a signi cant interaction on pD2 (-log ACh EC 50) was detected between diet and GSs (F [1, 24] = 17.4, P < 0.001). Simple-effects analysis revealed that the WTD or GS (+) alone did not cause endothelial dysfunction; however, the combination of the WTD and GS (+) deteriorated endothelial function (Fig. 2B). There were no signi cant differences in the dose-response curves to SNP, an endothelium-independent vasodilator, or ACh in the presence of L-NAME, a NOS inhibitor, even under the 20 mM glucose condition among the four groups ( Fig. 1C and D), suggesting that the impaired endothelial function in the WTD-GS (+) group under the 20 mM glucose condition was caused by a deterioration of NO-dependent relaxation.
These results indicate that short-term repeated GSs deteriorated EDR only in diet-induced obese rats under high-glucose conditions but not in nonobese rats.

Endothelial Dysfunction Caused by Repeated GSs in Diet-Induced Obese Rats is Independent of Fat Mass and Blood Lipid Pro les
To understand the mechanism of endothelial dysfunction in the WTD-GS (+) group, metabolic parameters were evaluated. A signi cant main effect for diet was observed in some dependent variables as below without diet/GS interaction, although the main effect of GSs was not signi cant; body weight gain and visceral fat (adiposity index) were signi cantly higher, the ITT (decreasing AUC) was lower (Fig. 1E-G), and plasma insulin, serum TG, FFA, TNFα, and FPG levels and HOMA-IR in the WTD group were higher than those in the CD group (Table 1). Total cholesterol levels were comparable between the two groups ( Table 1). Systolic and diastolic blood pressure in the WTD group tended to be higher than those in the CD group (the main effect of diet; F [1, 24] = 3.93, P = 0.059; F [1, 24] = 3.14, P = 0.089, respectively), but the differences were not signi cant (Table 1). These results indicate that changes in these parameters in the WTD group are compatible with metabolic syndrome. However, GSs did not affect these metabolic parameters in either the CD or WTD group.
Furthermore, to clarify whether the endothelial dysfunction observed in the WTD-GS (+) group under the 20 mM glucose condition was associated with these metabolic parameters, the correlation between each metabolic parameter and pD2 (-log ACh EC 50) was assessed. Among these parameters, HOMA-IR and serum FFA and TNFα levels were negatively correlated, and ITT (decreasing AUC) was positively correlated with the pD2 of the WGS (+) group under the 20 mM glucose condition (Supplementary Fig. S3D, G, H, E). Other metabolic parameters, such as body weight, adiposity index, and FPG and serum TG levels, were not signi cantly correlated with pD2 ( Supplementary Fig. S3A, B, C, F).

A NOX Inhibitor, SOD and Catalase Ameliorate Endothelial Dysfunction Caused by High-Glucose Conditions in Diet-Induced Obese Rats with Repeated GSs
To clarify the mechanism of endothelial dysfunction in the WTD-GS (+) group, we used several pharmacological agents to evaluate endothelial function under the 20 mM glucose condition. Apocynin, a NOX inhibitor, improved EDR, represented as pD2, in the WTD-GS (+) group (Fig. 2) (P < 0.001). Similarly, extrinsic SOD and catalase, which are superoxide scavengers and hydrogen peroxide scavengers, also improved EDR (P < 0.001; P =0.005, respectively) (Fig. 2). To investigate this nding in more detail, we used Mito-TEMPO and MnTABP, a mitochondria-targeted superoxide scavenger and a peroxynitrite selective scavenger, respectively. Both of these agents also improved EDR in the WTD-GS (+) group (P = 0.017; P =0.013, respectively) (Fig. 2). Indomethacin and allopurinol, a cyclooxygenase inhibitor and a xanthine oxidase inhibitor, respectively, however, did not improve EDR in the WTD-GS (+) group (Fig. 2), suggesting that prostaglandins and xanthine oxidase-derived radicals are unlikely to be responsible for endothelial dysfunction. We also evaluated the effect of insulin on endothelial function because repeated GSs were accompanied by transient hyperinsulinemia, and we found that even a high dose of insulin (10 nM) did not affect EDR, suggesting that direct exposure of the endothelium to insulin is neither harmful nor protective.  (Fig. 3). Notably, there were no signi cant differences in other redoxrelated enzyme genes, such as NOX1, NOX4, SOD1 and GPX1, among all four groups (Fig. 3). These results suggest the following: 1) diet-induced obesity by a WTD downregulates that of SOD2 and catalase; 2) short-term repeated GSs downregulate the mRNA expression of catalase; and 3) the combination of repeated GSs and diet-induced obesity synergistically upregulates that of NOX2 and p47phox and downregulates that of SOD2. With regard to in ammatory genes, the mRNA expression of TNFα was signi cantly different between diet or GS factors (F [1, 20] = 5.26, P = 0.033; F [1, 20] = 17.99, P < 0.001, respectively) and that of IL1β also different between GS factors (F [1, 20] = 9.17, P = 0.007) without diet/GS interaction (Fig. 3). An interactive effect of diet and GS on the mRNA expression of VCAM1 was detected (F [1, 20] = 12.0, P = 0.002), and simple-effects analysis revealed that the combination of a WTD and GS (+) upregulated that of VCAM1 (Fig. 3).

Radical Formation Is Enhanced by Repeated GSs in the Thoracic Aortas of Diet-induced Obese Rats
Repeated GSs in diet-induced obese rats showed an imbalance in redox enzyme gene expression. To investigate radical formation in the aorta, superoxide anion was evaluated by DHE staining of thoracic aortas isolated 2 hours after the administration of glucose (1 g/kg BW) or saline. The uorescence intensity of DHE was higher in the WTD-GS (+) group administered glucose than in the other three groups (Fig. 4A). Among the four groups, the interactive effect of diet and GS on the uorescence intensity of DHE (F [1, 12] = 37.42, P < 0.001) was con rmed by two-way ANOVA (Fig. 4A). Simple-effects analysis revealed that the combination of a WTD and GS (+) increased the uorescence intensity of DHE, although a WTD or GS (+) alone did not. This enhanced intensity of DHE in the WTD-GS (+) group was diminished by preincubation with 250 U/mL PEG-SOD (

CDDO-Me Protects the Endothelial Function of Diet-induced Obese Rats against Repeated GSs
Because repeated GSs induced oxidative stress in the thoracic aortas of diet-induced obese rats, we evaluated whether the administration of CDDO-Me, an activator of the Nrf2 system, could prevent endothelial dysfunction. To determine the optimal dose of CDDO-Me, ve doses of CDDO-Me (0 [vehicle], 0.3, 1, 3 and 15 mg/kg BW) were orally administered to rats in the WTD-GS (+) group. As shown in Supplementary Fig. S4, endothelial function was most strongly preserved against repeated GSs at 3 mg/kg BW CDDO-Me, although this protective effect was diminished at 15 mg/kg BW.
We orally administered vehicle (sesame oil) or CDDO-Me (3 mg/kg BW) to four groups of 19-week-old rats for two weeks and intraperitoneally administered saline or glucose for one week beginning at 20 weeks old, similar to the rst cohort. Thoracic aortas were isolated from these groups, and EDR was evaluated under 5.5 mM or 20 mM glucose conditions. Under the 20 mM glucose condition, the administration of CDDO-Me to rats in the WTD-GS (+) group showed a signi cant amelioration of EDR and pD2 compared with vehicle (P < 0.001), although no signi cant differences were observed under the 5.5 mM glucose condition ( Fig. 5A and B). Among the CD-GS (-), CD-GS (+), and WTD-GS (-) groups, there were no differences in EDR or pD2 among the vehicle and CDDO-Me groups under either the 5.5 mM or 20 mM glucose conditions ( Fig. 5A and B).

CDDO-Me Does Not Affect Metabolic Parameters
The effects of CDDO-Me on metabolic parameters were assessed. In the comparison of treatment with vehicle or CDDO-Me for each group, no signi cant differences in body weight, adiposity index (visceral fat mass), ITT (insulin resistance), or levels of plasma insulin, serum TG, FFA, FPG, or serum TC were observed ( Supplementary Fig. S5A-H). There were also no differences in HOMA-IR, serum TNFα levels or blood pressure ( Supplementary Fig. S5I-L).

Repeated GSs
To clarify the mechanism of the amelioration of endothelial dysfunction by treatment with CDDO-Me, we evaluated the changes in gene expression in the thoracic aortas between vehicle and CDDO-Me in each group. The mRNA expression of NQO1, a target gene of the Nrf2 system, was markedly enhanced by CDDO-Me in the four groups (Fig. 6A). CDDO-Me in the WTD-GS (+) group signi cantly reduced the mRNA expression of NOX2 and p47phox (P = 0.004; P = 0.004, respectively) and signi cantly increased that of SOD2 and catalase (Fig. 6B-E) (P = 0.007; P = 0.033, respectively). With regard to in ammatory genes, the mRNA expression levels of VCAM1 in the WTD-GS (+) group treated with CDDO-Me were signi cantly lower than those in the WTD-GS (+) group treated with vehicle (P = 0.039) (Fig. 6H). CDDO-Me also reduced those of TNFα and IL1β, but there were no signi cant differences between vehicle and CDDO-Me ( Fig. 6F and G).

CDDO-Me Suppresses Local and Systemic Oxidative Stress Caused by Repeated GSs in Diet-induced Obese Rats
To evaluate the effect of CDDO-Me on local and systemic oxidative stress, we measured DHE uorescence intensity in the thoracic aortas and urinary 8-OHdG levels between the vehicle and CDDO-Me groups. These samples were collected 2 hours after the intraperitoneal administration of glucose (1 g/kg BW).
Treatment with CDDO-Me in the WTD-GS (+) group signi cantly decreased DHE uorescence intensity and urinary 8-OHdG levels (P = 0.020 and P = 0.048, respectively) ( Fig. 6I and J). In the other three groups, no signi cant differences in DHE uorescence intensity or urinary 8-OHdG levels were observed between the vehicle and CDDO-Me groups ( Fig. 6I and J).

Discussion
In the present study, we clearly demonstrated that short-term repeated GSs cause high-glucose-dependent endothelial dysfunction in the aortas of obese rats with insulin resistance induced by a Western-type diet but not in the aortas of nonobese rats without insulin resistance fed a CD. Endothelial dysfunction was associated with the increased detection of superoxide anion in the wall of the aorta, the decreased expression of genes encoding enzymes that eliminate ROS, such as SODs and catalase, and the elevated expression of genes encoding enzymes that produce ROS, such as NOXs. In addition, this endothelial dysfunction induced by short-term repeated GSs was prevented by the administration of CDDO-Me, a Nrf2 system activator.
A decrease in the bioavailability of NO resulting in endothelial dysfunction is generally caused by an increase in radical levels or ROS production [39][40][41][42], and high-glucose conditions have been reported to increase superoxide anion levels in endothelial cells [43,44]. We also found that apocynin, SOD, catalase or MnTABP, a selective peroxynitrite remover [45], improved endothelial dysfunction under high-glucose conditions, suggesting an increase in oxidative stress.
Interestingly, however, the high-glucose condition induced endothelial dysfunction only in the aortas from the diet-induced obese rats with repeated GSs but not in the aortas from nonobese rats with repeated GSs or diet-induced obese rats without repeated GSs. These results indicate that high-glucose conditions, repeated GSs and insulin resistance and/or hyperinsulinemia are additive factors for the impairment of endothelial function.
In previous reports, each factor of GSs or diet-induced obesity was reported to cause endothelial dysfunction, but each factor was insu cient to impair endothelial function in the present study. Considering that GSs require 3 months to accelerate arteriosclerosis [7], the period of GSs may have been too short to induce endothelial dysfunction in nonobese rats with GSs. In addition, endothelial dysfunction by diet-induced obesity has been reported to be partly caused by eNOS uncoupling in perivascular adipose tissue (PAVT), and endothelial dysfunction was not observed in PAVT-free aortas isolated from obese mice [46]. We removed PAVT from the thoracic aorta, resulting in endothelial dysfunction that could not be observed in diet-induced obese rats without GSs in our experiment.
The mRNA expression of SOD2 and catalase in the aortas of the diet-induced obese rats without repeated GSs decreased and tended to decrease, respectively, while repeated GSs tended to increase the mRNA expression of NOX2 in the aortas of the nonobese rats fed a CD. These results are consistent with previous reports that a high-fat diet reduces the expression of SOD2 without any changes in the expression of NOX2 [47], that obesity decreases the expression of catalase in the aorta [48], and that intermittent high glucose levels stimulate ROS production through the protein kinase C-dependent activation of NOX [49]. In addition, the combination of diet-induced obesity and repeated GSs synergistically enhanced these effects on the genes in our experiment. Although the mechanism of this synergistic effect of obesity and GSs on gene expression in the aorta is unclear, serum free fatty acids might be a key factor because free fatty acids have been reported to increase the gene expression of NOX [50,51] and decrease the gene expression of antioxidant genes such as NQO1, SOD and catalase via Nrf2 system suppression in endothelial cells, resulting in enhanced ROS production [52]. In our experiment, serum free fatty acid levels correlated negatively with pD2 under high-glucose conditions ( Supplementary Fig. S3G), suggesting that repeated GSs and an elevation in serum free fatty acid levels due to diet-induced obesity might synergistically stimulate changes in the gene expression of NOX2, SOD2 and catalase. Accordingly, these changes in gene expression in the aorta may increase the vulnerability of endothelial function to ROS.
High-glucose conditions have been reported to induce mitochondria-derived superoxide production [43,44]. We also observed that Mito-TEMPO, a mitochondria-targeted superoxide scavenger, improved high-glucose-dependent endothelial dysfunction. Since endothelial cells take up glucose in an insulinindependent manner via glucose transporter 1 (GLUT1) [53] and no downregulation of GLUT1 expression in response to high extracellular glucose levels has been reported, endothelial function is thought to be susceptible to hyperglycemia [54]. Increased glucose uptake and glycolytic conversion to pyruvate can increase the production of mitochondrial ROS [55,56]. Based on these reports, an increase in mitochondria-derived ROS production could be observed during temporal hyperglycemia. Although mitochondria-derived ROS are easily removed by antioxidant enzymes, ROS may not be eliminated completely under the suppression of antioxidant enzymes, such as SOD2 and catalase, in the aortas of diet-induced obese rats with repeated GSs. In addition, the gene expression of NOX2 in the aorta was increased, and apocynin improved endothelial dysfunction in these aortas, suggesting that NOX2-derived superoxide might also contribute to endothelial dysfunction. Considering that endothelial function was impaired only under high-glucose conditions, NOX2-derived superoxide might enhance mitochondrial superoxide production under high-glucose conditions [57].  The data are presented as the mean ± SEM. Figure 1 Glucose spikes deteriorate endothelium-dependent relaxation in diet-induced obese rats independent of weight gain. Rats were fed a control diet (CD) or Western-type diet (WTD) for 13 weeks and administered saline or glucose for 1 week twice daily (7 rats per group). A: Curves of endothelium-dependent relaxation (EDR) in the thoracic aorta in response to ACh under 5.5 mM glucose, 20 mM glucose and 20 mM ra nose conditions. * P < 0.05, ‡ P < 0.005, § P <  NOX inhibitors, SOD and catalase ameliorate endothelial dysfunction under high-glucose conditions in the WTD-GS (+) group. Rats were fed a control diet (CD) or Western-type diet (WTD) for 13 weeks and administered saline or glucose for 1 week. Vascular sensitivity under 20 mM glucose conditions was plotted as pD2 (-log of the half-maximal effective concentration [EC50]) of ACh. The control was administered to 7 rats, and other agents were administered to 4 rats each. * P < 0.05, † P < 0.01, § P < 0.001 vs. control in the WTD-GS (+) group, Dunnett's test. The data are presented as the mean ± SEM. Indomethacin, a cyclooxygenase inhibitor; allopurinol, a xanthine oxidase inhibitor; apocynin, a NADPH oxidase inhibitor; SOD, superoxide dismutase; Mito-TEMPO, a mitochondria-targeted superoxide scavenger; MnTABP, a superoxide dismutase mimetic and peroxynitrite selective scavenger. CD, control diet; WTD, Westerntype diet; GS, glucose spike. The GS (-) group consisted of rats given saline, and the GS (+) group consisted of rats given glucose.

Figure 3
Quantitative PCR analysis of gene expression in the thoracic aorta. mRNA expression of NOX2 and SOD2 was synergistically changed by the combination of a Western-type diet (WTD) and repeated glucose spikes (GSs). Rats were fed a control diet (CD) or Western-type diet (WTD) for 13 weeks and administered saline or glucose for 1 week twice daily (6 rats per group). The signi cant interactive effect of diet and GS on the mRNA expression of NOX2, p47phox, SOD2 and VCAM1 was con rmed by two-way ANOVA followed by simple-effects analysis (* P < 0.05, ‡ P < 0.005, § P < 0.001). The mRNA expression of catalase and TNFα was signi cantly different between diet or GS factors and that of IL1β was different between GS factors without diet/GS interaction. * P < 0.05, † P < 0.01, ‡ P < 0.005 vs. between diet or GS factors, two-way ANOVA. The data are presented as the mean ± SEM. CD, control diet; WTD, Western-type diet; GS, glucose spike. The GS (-) group consisted of rats given saline, and the GS (+) group consisted of rats given glucose.

Figure 4
Superoxide anion is enhanced by repeated glucose spikes in diet-induced obese rats. Rats were fed a control diet (CD) or Western-type diet (WTD) for 13 weeks and given saline or glucose for 1 week (4 rats per group). The thoracic aorta was removed 2 hours after the intraperitoneal administration of glucose (1 g/kg) to the 4 groups or saline to the WTD-GS (+) group. A: Representative images of DHE staining and DHE uorescence intensity in the thoracic aorta. The signi cant interactive effect of diet and GS on DHE uorescence intensity was con rmed by two-way ANOVA followed by simple-effects analysis ( § P < 0.001). B: DHE uorescence intensity in the aortas incubated with PEG-SDS and in the aortas removed after the intraperitoneal administration of saline to the WTD-GS (+) group. The control group (Glu i.p.) was the same as the WTD-GS (+) group in Fig. 4A. § P < 0.001, one-way ANOVA. The data are presented as the mean ± SEM. Glu, glucose; i.p., intraperitoneal; PEG-SOD, polyethylene glycol-superoxide dismutase; CD, control diet; WTD, Western-type diet; GS, glucose spike.
The GS (-) group consisted of rats given saline, and the GS (+) group consisted of rats given glucose.