Obesity is a hallmark of insulin resistance and the metabolic syndrome, which has significantly increased worldwide( Reference Kitzinger and Karle 1 ). Previous studies demonstrated that high-energy-diet-induced obesity led to the development of insulin resistance and cognitive decline in animal models( Reference Stranahan, Norman and Lee 2 , Reference Pistell, Morrison and Gupta 3 ). In addition, our previous studies showed that the development of peripheral insulin resistance, brain insulin resistance, brain mitochondrial dysfunction, as well as cognitive decline, occurred in high-fat-diet (HFD)-induced obese male rats( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pipatpiboon, Pratchayasakul and Chattipakorn 5 ). The dipeptidyl peptidase-4 inhibitor, vildagliptin, is widely used in the treatment of type 2 diabetes mellitus (T2DM) because of its beneficial effects on the regulation of the blood glucose level without causing any hypoglycaemic effects( Reference Burkey, Li and Bolognese 6 ). In addition, our previous studies reported that vildagliptin improved not only peripheral insulin resistance but also brain insulin resistance, as indicated by improved brain insulin receptor function in obese insulin-resistant rats( Reference Pipatpiboon, Pintana and Pratchayasakul 4 ). Furthermore, the enhancement of hippocampal neurogenesis and improvement in cognition were found in HFD-fed rodents treated with vildagliptin( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Gault, Lennox and Flatt 7 , Reference Pintana, Apaijai and Chattipakorn 8 ). All of those findings suggest that vildagliptin exerts beneficial effects on metabolic control and cognitive function.
Although several types of medication have been prescribed for the treatment of T2DM or the obese insulin-resistant condition, several side effects of medication have been observed. Therefore, a lifestyle modification, including energy restriction (ER), has been used as an alternative therapy in the obese insulin-resistant condition and T2DM. A previous study showed that ER in older animals led to improved metabolic parameters and also an extended lifespan( Reference Speakman and Mitchell 9 ). Although adverse effects of ER such as promoting bone loss( Reference Baek, Barlow and Allen 10 ) and decreasing wound healing rate( Reference Reiser, McGee and Rucker 11 ), in rats have been reported, several studies have shown that ER mimetics conferred the benefits of ER without its side effects( Reference de Magalhaes, Wuttke and Wood 12 , Reference Ingram, Zhu and Mamczarz 13 ). Moderate ER had a beneficial impact on metabolic parameters, and improved insulin sensitivity in in vitro and in vivo studies( Reference McCurdy, Davidson and Cartee 14 – Reference Wilsey and Scarpace 17 ). In addition, a previous study reported that ER not only exerts a positive effect on metabolic control but also leads to improved brain function( Reference Yamaoka and Tango 18 , Reference Kishi, Hirooka and Nagayama 19 ). Several studies demonstrated that ER caused an increase in an endogenous apoptosis inhibitor in the neuronal cells( Reference Hiona and Leeuwenburgh 20 ), induced neuroprotection( Reference Wood, van Dam and Craig 21 ), decreased mitochondrial reactive oxygen species (ROS) production in aged rat brains( Reference Sanz, Caro and Ibanez 22 ) and preserved cognitive function in aged male rats( Reference Adams, Shi and Linville 23 ). All of those findings suggest that ER has beneficial effects in neuroprotection and the preservation of cognitive function.
However, the effects of either ER or the combination of ER plus vildagliptin on brain function, including brain insulin sensitivity, brain mitochondrial function, hippocampal synaptic plasticity and cognitive function in obese insulin-resistant rats have never been investigated. This study tests the hypothesis that the combination of ER plus vildagliptin has greater beneficial effects on the brain function of obese insulin-resistant rats than ER alone.
Methods
Animal models and experimental protocols
In total, twenty-four male Wistar rats weighing 180–200 g (approximately aged 5–6 weeks old), obtained from the National Animal Center, Salaya Campus, Mahidol University, Bangkok, Thailand, were used. All experiments were conducted in accordance with the approved protocol from the Faculty of Medicine, Chiang Mai University’s Institutional Animal Care and Use Committee, in compliance with NIH guidelines (protocol number: 31/2557). All animals were housed in environmentally controlled conditions (25±0·5°C and a 12 h light–12 h dark cycle) and were allowed to acclimate for 1 week. Rats were randomly divided into two groups and fed on either a normal diet (ND: 19·77 % energy (%E) from fat, n 6) or a HFD (59·28 %E from fat, n 18) for 12 weeks( Reference Pratchayasakul, Kerdphoo and Petsophonsakul 24 ). At the end of the 12th week, rats in the ND group received normal saline solution (NSS) via intra-gastric gavage, as indicated, and were designated the control group (ND-fed rats treated with the vehicle (NDV)), and rats in the HFD group were subdivided into three subgroups: (1) 12-week HFD-fed rats continued with HFD and in addition were given NSS via intra-gastric gavage for 4 weeks (HFD-fed rats treated with the vehicle (HFV)); (2) 12-week HFD-fed rats were switched to a restricted diet (ND, containing only 60 % energy intake compared with the energy intake of HFD), as described in a previous study( Reference Wilsey and Scarpace 17 ), and were also given NSS via intra-gastric gavage for 4 weeks (HFD-fed rats on a restricted diet and vehicle (HFRV)); and (3) 12-week HFD-fed rats were switched to a restricted diet and in addition were given 3 mg/kg per d vildagliptin (Novartis, Thailand) via intra-gastric gavage for 4 weeks( Reference Burkey, Li and Bolognese 6 , Reference Pintana, Apaijai and Chattipakorn 8 ). The rats’ cognitive function was subsequently determined by the Morris Water Maze (MWM) test. Blood samples were collected from a tail vein at weeks 0, 12 and 16 for further plasma analysis. At the end of the experimental period, rats were deeply anesthetised with 2–3 % isoflurane and killed by decapitation. The brain of each rat was rapidly removed after death and carefully sliced in preparation for extracellular recording (insulin-induced long-term depression (LTD) and hippocampal synaptic long-term potentiation (LTP)), immunoblot and brain mitochondrial functions. The experimental protocol is shown in Fig. 1.
The experimental protocol of the study. OGTT, oral glucose tolerance test.
Blood sample assays
Plasma glucose and cholesterol levels were determined via colorimetric assay (Biotech). Plasma HDL and LDL levels were determined using a commercial colorimetric assay kit (Biovision). Plasma insulin levels were determined using the Sandwich ELISA kit (Millipore). Peripheral insulin resistance was assessed using the homoeostasis model assessment (HOMA), as previously described( Reference Pipatpiboon, Pratchayasakul and Chattipakorn 5 ).
Oral glucose tolerance test
Oral glucose tolerance test (OGTT) was performed as described previously( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pintana, Apaijai and Chattipakorn 8 ). In brief, rats were starved overnight before the test and received 2 g/kg glucose solution via oral gavage, and then blood samples were collected at different time points. AUC were calculated to evaluate glucose tolerance.
Brain-slice preparation
The brain-slice preparation was performed as described previously( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pipatpiboon, Pratchayasakul and Chattipakorn 5 , Reference Pratchayasakul, Kerdphoo and Petsophonsakul 24 ). In brief, the whole brain was rapidly removed after decapitation and then immersed in ice-cold high-sucrose artificial cerebrospinal fluid (aCSF). The hippocampal slices were cut using a vibratome (Vibratome Company). Following a 30-min post-slice incubation in high-sucrose aCSF, the slices were transferred to a standard aCSF for an additional 30 min at room temperature (22–24°C).
Extracellular recordings of hippocampal slices for insulin-induced long-term depression
An extracellular recording of hippocampal slices for insulin-induced LTD was performed as described in previous studies( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pipatpiboon, Pratchayasakul and Chattipakorn 5 , Reference Pratchayasakul, Kerdphoo and Petsophonsakul 24 ). Hippocampal slices were perfused with aCSF (as a baseline condition) for 10 min, and then perfused with aCSF plus 500 nm-insulin (as an insulin-induced LTD) for an additional 10 min. After this, the slices were perfused with aCSF for an additional 50 min and readings recorded.
Extracellular recordings of hippocampal slices for synaptic long-term potentiation
The examination of electrical-induced hippocampal synaptic LTP was performed as described in a previous study( Reference Sripetchwandee, Pipatpiboon and Chattipakorn 25 ). In brief, LTP was induced by delivering high-frequency stimulation (HFS; four trains at 100 Hz; 0·5-s duration; 20-s interval). Experiments were performed for 40 min after HFS.
Immunoblotting for brain insulin signalling
To investigate the expression of insulin receptor phosphorylation (p-IR), insulin receptors (IR), serine/threonine-specific protein kinase B (Akt/PKB) and insulin-mediated Akt Ser473 phosphorylation, homogenated brain slices from each subgroup were used as described in our previous study( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pipatpiboon, Pratchayasakul and Chattipakorn 5 , Reference Pratchayasakul, Kerdphoo and Petsophonsakul 24 ). The p-IR, IR, Akt/PKB at serine 473 kinase phosphorylation and Akt/PKB were electrophoresed and immunoblotted with rabbit anti-IR at tyrosine phosphorylation (p-IRtyr1162/1163) (1:1000, sc-25103-R; Santa Cruz Biotechnology), IR (1:1000, sc-711; Santa Cruz Biotechnology), Akt/PKB at serine 473 kinase phosphorylation (1:1000, no. 9271; Cell Signaling Technology) and Akt/PKB (1:1000, no. 9272; Cell Signaling Technology). For a loading control, the immunoblot for each membrane was incubated with anti-β-actin (1:4000, no. 4967; Cell Signaling Technology). All membranes enabling the visualisation of the phosphorylation and the protein levels of IR and Akt/PKB expression were incubated with secondary goat anti-rabbit antibody conjugated with horseradish peroxidase (1:2000, no. 7074; Cell Signaling Technology). Band densities of phosphorylated IR or Akt/PKB at serine 473 kinase phosphorylation were represented as a ratio of insulin stimulation (+):no insulin stimulation (−) and were normalised to total IR or Akt/PKB. Band intensities were quantified using Scion Imaging, and the results were shown as average signal intensity (arbitrary) units.
Serum and brain malondialdehyde levels
HPLC method was used to evaluate the concentrations of serum and brain malondialdehyde (MDA), as described previously( Reference Pintana, Apaijai and Chattipakorn 8 , Reference Candan and Tuzmen 26 – Reference Pintana, Sripetchwandee and Supakul 28 ).
Brain mitochondrial function study
Brain mitochondria were isolated using the method described previously( Reference Pipatpiboon, Pintana and Pratchayasakul 4 ). Brain mitochondrial function including brain mitochondrial ROS, mitochondrial membrane potential change (ΔΨm) and mitochondrial swelling were determined. Brain mitochondrial ROS were measured using dichloro-hydrofluorescein diacetate fluorescent dye. The change in mitochondrial membrane potential (ΔΨm) was measured using the fluorescent dye 5, 52, 6, 62-tetrachloro-1, 12, 3, 32-tetraethyl benzimidazol carbocyanine iodide (JC-1) and brain mitochondrial swelling was determined by measuring the change in the absorbance of brain mitochondrial suspension at 540 nm. All were determined by following the methods described previously( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pintana, Apaijai and Chattipakorn 8 ).
Cognitive function test
The open-field test (OFT) was used to screen locomotor activity, as described in previous studies( Reference Pintana, Apaijai and Chattipakorn 8 , Reference Arakawa 29 ). The assessment of cognitive function was performed using the MWM test with two assessments including the acquisition test, which was carried out for 5 consecutive days, and the probe test, which was performed on day 6( Reference Pintana, Apaijai and Chattipakorn 8 , Reference Vorhees and Williams 30 ). Data analysis of the MWM test was done manually from video tape recordings by the investigators, who were blinded to experimental groups.
Golgi staining and morphological analysis of dendritic spines
Golgi staining and morphological analysis were carried out using the method described previously( Reference Sripetchwandee, Pipatpiboon and Pratchayasakul 31 ). After decapitation, brains were removed and rinsed with double-distilled water and were processed for Golgi staining using a commercially available kit (FD Rapid GolgiStain™ Kit #PK401; FD Neurotechnologies, Inc.). For analysis of dendritic spine density, the secondary and tertiary dendrites of three neurons in the CA1 hippocampus area were counted. Dendritic segments were viewed through an inverted microscope (IX-81; Olympus).
Statistical analysis
Data were expressed as mean values with their standard errors. For all comparisons, the significance of the differences in peripheral biochemical parameters was calculated using the Mann–Whitney U test. The comparisons in the percentage of insulin-induced LTD, the percentage of LTP, brain mitochondrial function, immunoblot, the OFT tests and the MWM tests for the probe test between groups were performed using the one-way ANOVA test, followed by post hoc Fisher’s least significant difference (LSD) analysis. As the acquisition phase of the MWM test has two independent factors (treatment and time), we used the two-way ANOVA followed by post hoc Fisher’s LSD analysis to analyse the data. P<0·05 was considered as a measure of statistical significance.
Results
Energy restriction and the combination of energy restriction plus vildagliptin in obese insulin-resistant rats equally restored peripheral insulin sensitivity
At baseline levels (week 0), the metabolic parameters were not significantly different between the ND-fed rats and the HFD-fed rats (Table 1). After 12 weeks of diet regimens, the HFD group had significantly increased body weight, plasma insulin, HOMA index, total cholesterol levels and LDL-cholesterol levels, when compared with the ND group (Table 1). However, plasma glucose levels were not significantly different between the groups (Table 1).
Metabolic parameters at week 0 (baseline) and the end of week 12 (before treatment) (Mean values with their standard errors)
ND, normal-diet-fed rats; HFD, high-fat-diet-fed rats; HOMA, homoeostasis model assessment.
* Compared with ND at the same time interval.
After 4 weeks of the treatment, HFV had significantly increased body weight, visceral fat, plasma insulin levels, HOMA index, plasma glucose AUC of OGTT (AUCg), total cholesterol and plasma LDL-cholesterol levels, but had decreased plasma HDL-cholesterol levels, when compared with NDV (Table 2). These findings indicated that peripheral insulin resistance was still observed in HFV rats after 16 weeks. Interestingly, HFRV and HFD-fed rats on a restricted diet and vildaglptin (HFRVil) had significantly decreased body weight, visceral fat, plasma insulin, HOMA index, plasma glucose AUCg, total plasma cholesterol and LDL-cholesterol levels, but increased plasma HDL-cholesterol levels, when compared with HFV rats (Table 2). However, there was no significant difference in plasma glucose and plasma TAG levels between all groups (Table 2).
Metabolic parameters after energy restriction and energy restriction plus vildagliptin treatment (Mean values with their standard errors)
NDV, normal-diet-fed rats treated with vehicle; HFV, high-fat-diet-fed rats treated with vehicle; HFRV, high-fat-diet-fed rats reversed to normal diet with 60 % energy restriction treated with vehicle; HFRVil, high-fat-diet-fed rats reversed to normal diet with 60 % energy restriction treated with vildagliptin; HOMA, homoeostasis model assessment; MDA, malondialdehyde.
* Compared with NDV.
† Compared with HFV.
Obesity caused the impairment of brain insulin receptor function, which was only restored by the combination of energy restriction plus vildagliptin
After 4 weeks of the treatment, the degree of insulin-induced LTD in HFV rats was significantly reduced, when compared with that in NDV rats (n 2–3 independent slices/animal, n 6 animals/group, Fig. 2(a)). Interestingly, the degree of insulin-induced LTD in only HFRVil rats showed a significant increase, when compared with that of HFV rats and HFR rats (n 2–3 independent slices/animal, n 6 animals/group, Fig. 2(a)). Regarding brain insulin signalling, the levels of IR and Akt/PKB proteins expression showed no difference between the groups (Fig. 2(d) and (e)). The p-IR and Akt/PKB at the serine 473 site in both HFV and HFRV rats showed a significant decrease, when compared with that of NDV rats (Fig. 2(b) and (c)). However, the p-IR levels and Akt/PKB at the serine 473 site of only HFRVil rats showed a significant increase when compared with those of HFV rats (Fig. 2(b) and (c)).
The effects of energy restriction (ER) and the combination of ER plus vildagliptin on brain insulin receptor function (insulin-induced long-term depression (LTD)) and brain insulin signalling, including the phosphorylation of insulin receptors (p-IR), Akt-Ser 473, insulin receptors (IR) and serine/threonine-specific protein kinase B (Akt/PKB) protein expression in obese insulin-resistant rats. The combination of ER plus vildagliptin improves the ability of insulin-induced LTD in high-fat-diet-fed rats reversed to normal diet with 60 % ER treated with vildagliptin (HFRVil), when compared with high-fat-diet-fed rats treated with vehicle (HFV) (a). After 4 weeks of treatment, both p-IR and Akt-Ser 473 expression levels were significantly decreased in HFV and high-fat-diet-fed rats reversed to normal diet with 60 % ER treated with vehicle (HFRV) when compared with normal-diet-fed rats treated with vehicle (NDV) ((b) and (c)). Both p-IR and Akt-Ser 473 expression increased significantly after the combination of ER plus vildagliptin when compared with HFV rats ((b) and (c)). However, there was no difference in IR and Akt/PKB protein expression between all groups ((d) and (e)). *P<0·05 v. NDV and † P<0·05 v. HFV.
, NDV; 
, HFV; 
, HFRV; 
, HFRVil. fEPSP, field excitatory postsynaptic potential.
Obesity led to brain mitochondrial dysfunction, which improved following treatment with the combination of energy restriction plus vildagliptin
After 4 weeks of treatment, brain mitochondrial dysfunction was observed in HFV rats as indicated by increased brain mitochondrial ROS production, brain mitochondrial membrane depolarisation and brain mitochondrial swelling (Fig. 3(a), (b) and (c)). HFRV rats also had a significant increase in brain mitochondrial ROS production, brain mitochondrial depolarisation and brain mitochondrial swelling, when compared with NDV rats (Fig. 3(a), (b) and (c)). Interestingly, only HFRVil rats showed a significant decrease in brain ROS production, and improvement in brain mitochondrial depolarisation and brain mitochondrial swelling, when compared with HFV rats (Fig. 3(a), (b) and (c)).
The effects of energy restriction (ER) and the combination of ER plus vildagliptin on brain mitochondrial function and representative images of brain mitochondrial morphology by transmission electron microscopy (JEM-2200FS field emission electron microscope, original magnification 20 000×) in obese insulin-resistant rats. High-fat-diet-fed rats treated with vehicle (HFV) and high-fat-diet-fed rats reversed to normal diet with 60 % ER treated with vehicle (HFRV) demonstrated brain mitochondrial dysfunction when compared with normal-diet-fed rats treated with vehicle (NDV), as indicated by increased brain mitochondrial reactive oxygen species (ROS) production following H2O2 application (a), increased brain mitochondrial membrane potential change following H2O2 application (b) and decreased absorbance values, indicating brain mitochondrial swelling (c). The combination of ER plus vildagliptin in high-fat-diet-fed rats reversed to normal diet with 60 % ER treated with vildagliptin (HFRVil) showed improved brain mitochondrial function when compared with HFV rats, as indicated by significantly decreased brain mitochondrial ROS production (a), decreased brain mitochondrial membrane potential change (b) and increased absorbance values (c). Furthermore, normal folding of cristae in brain mitochondrial morphology was shown in both NDV and HFRVil rats (d). However, brain mitochondrial swelling, as indicated by unfolded cristae in both HFV and HFRV rats, was observed (d). *P<0·05 v. NDV and † P<0·05 v. HFV.
Furthermore, there were morphological changes in brain mitochondria; representatives from all groups are shown in Fig. 3(d). Brain mitochondrial swelling was observed in both the HFV and HFRV rats, as indicated by markedly unfolded cristae (Fig. 3(d)). Interestingly, brain mitochondrial morphology with apparent folded cristae was observed in only HFRVil rats (Fig. 3(d)).
Obesity led to increased brain oxidative stress levels, which was attenuated by the combination of energy restriction plus vildagliptin
After 4 weeks of treatment, both circulating MDA and brain MDA levels significantly increased in HFV rats compared with NDV rats (Table 2). HFRV rats only showed a significant decrease in circulating MDA, when compared with HFV rats (Table 2). Interestingly, HFRVil rats showed a significant decrease in both circulating MDA and brain MDA levels, when compared with HFV rats (Table 2).
Obesity caused impaired hippocampal synaptic plasticity and cognitive decline, which only showed restoration with the combination of energy restriction plus vildagliptin
After 4 weeks of treatment, the results demonstrated that the degrees of electrical-induced LTP significantly decreased in both HFV and HFRV rats, when compared with NDV rats (n 2–3 independent slices/animal, n 6 animals/group, Fig. 4(a)). The reduction of electrical-induced LTP in both HFV and HFRV rats was not significantly different (n 2–3 independent slices/animal, n 6 animals/group, Fig. 4(a)). However, the degree of electrical-induced LTP of HFRVil rats significantly increased, when compared with that of HFV rats (n 2–3 independent slices/animal, n 6 animals/group, Fig. 4(a)).
The effects of energy restriction (ER) and the combination of ER plus vildagliptin on hippocampal synaptic long-term potentiation (LTP) and the number of dendritic spines on tertiary dendrites on apical dendrites, as well as cognitive function, in obese insulin-resistant rats. Both high-fat-diet-fed rats treated with vehicle (HFV) and high-fat-diet-fed rats reversed to normal diet with 60 % ER treated with vehicle (HFRV) showed a significantly reduced degree of LTP, when compared with normal-diet-fed rats treated with vehicle (NDV) (a). The degree of LTP showed improvement after a combination of ER plus vildagliptin when compared with HFV rats (a). The number of dendritic spines on secondary or tertiary dendrites in hippocampal apical dendrites was decreased in HFV rats (b). Both HFRV and high-fat-diet-fed rats reversed to normal diet with 60 % ER treated with vildagliptin (HFRVil) showed significantly increased numbers of dendritic spines after the treatment regimen (b). HFV and HFRV rats showed a significantly increased time to reach the platform in the acquisition test, as well as having a decreased time spent in the target quadrant in the probe test, when compared with NDV rats ((c) and (d)). HFRVil rats showed a significantly decreased time to reach the platform in the acquisition test, as well as an increased time spent in the target quadrant in the probe test, when compared with HFV rats ((c) and (d)). HFS, high-frequency stimulation. *P<0·05 v. NDV, † P<0·05 v. HFV and ‡ P<0·05 v. HFRV. (a) , NDV; , HFV;
, HFRV; , HFRVil; (c) 
, NDV; 
, HFV; 
, HFRV; 
, HFRVil. fEPSP, field excitatory postsynaptic potential.
The density of hippocampal dendritic spines was determined at the secondary and tertiary dendrites in the apical dendrite of the CA1 hippocampus to investigate the underlying mechanism of memory function. It was found that the dendritic spine density of HFV and HFRV rats decreased significantly, when compared with that of NDV rats (Fig. 4(b)). After treatment regimens, the dendritic spine density of both HFRV and HFRVil rats showed a significant increase, when compared with that of HFV rats (Fig. 4(b)). However, dendritic spine density of HFRVil rats was significantly greater than that of HFRV rats (Fig. 4(b)).
The OFT was used to screen locomotor activity. The results showed that there was no significant difference between all groups, indicating that the locomotor activity did not differ in all groups. In addition, the study determined cognitive function by using the MWM test after 4 weeks of the treatment regimens. For the acquisition test, it was found that both HFV and HFRV rats showed a significantly increased time to reach the platform, when compared with NDV rats (Fig. 4(c)). However, HFRVil rats showed a significantly decreased time to reach the platform, when compared with HFV rats (Fig. 4(c)). For the probe test, both HFV and HFRV rats had a significantly decreased time spent in the target quadrant, when compared with NDV rats (Fig. 4(d)). In contrast, HFRVil rats had a significantly increased time spent in the target quadrant, when compared with HFV rats (Fig. 4(d)).
Discussion
The major findings of this study are as follows: (1) obesity induced by HFD consumption leads to the development of peripheral insulin resistance, impairment of brain insulin sensitivity, brain mitochondrial dysfunction, impaired hippocampal synaptic plasticity, decreased dendritic spine density and cognitive decline; (2) both ER and the combination of ER plus vildagliptin showed equally improved peripheral insulin sensitivity and lipid profiles; (3) ER alone showed a decrease only in circulating oxidative stress, whereas the combination of ER plus vildagliptin showed a decrease in both circulating and brain oxidative stress levels; and (4) only the combination of ER plus vildagliptin showed a correlation with restored brain insulin sensitivity, brain mitochondrial function, hippocampal synaptic plasticity and cognitive function in obese insulin-resistant rats.
This study has demonstrated that diet-induced obesity led to the development of the peripheral insulin resistance via impaired peripheral insulin sensitivity, impaired lipid profiles and an increase in both circulating and brain oxidative stress. These findings confirmed other studies and our previous studies that diet-induced obesity not only caused peripheral insulin resistance but also brain insulin resistance( Reference Pratchayasakul, Kerdphoo and Petsophonsakul 24 , Reference Greenwood and Winocur 32 ). The present findings indicated that obesity increased oxidative stress and impaired brain mitochondrial function, resulting in brain insulin resistance and cognitive decline.
This study demonstrates that ER enhanced peripheral insulin sensitivity and improved lipid profiles in obese insulin-resistant rats. A possible explanation of these findings may be that ER decreased visceral fat and dyslipidaemia. Decreased fat has been shown to reduce systemic inflammation( Reference Hatori, Vollmers and Zarrinpar 33 ), as well as oxidative stress, as shown in Table 2 and in a previous study( Reference Jeon, Jeong and Shin 34 ). An increase in inflammation and oxidative stress led to the impairment of peripheral insulin sensitivity( Reference Rains and Jain 35 , Reference Glass and Olefsky 36 ). Therefore, ER in the present study led to improved peripheral insulin sensitivity, possibly via the reduction of oxidative stress.
Although ER led to the restoration of peripheral insulin sensitivity and decreased circulating oxidative stress, ER alone failed to cause improvement in brain insulin sensitivity and brain mitochondrial dysfunction, as well as failed to restore hippocampal synaptic plasticity and cognitive decline in obese insulin-resistant rats. This may be because of our finding that ER alone could neither decrease brain oxidative stress levels nor brain mitochondrial dysfunction, whereas treatment with combined ER and vildagliptin effectively improved both parameters in obese insulin-resistant rats (Table 2).
Unlike this study, previous studies showed that ER had beneficial effects on neuroprotection and preservation of cognition. For example, (1) a previous study demonstrated that 40 % of ER for 24 months reduced the rate of mitochondrial H2O2 production via reduced brain ROS production in aged rats( Reference Sanz, Caro and Ibanez 22 ). (2) The study involving 40 % ER for 22–29 months also reported that ER prevented age-related deficit synaptic LTP and the preservation of the NR1-subunit of N-methyl-D-aspartate (NMDA) receptors in aged Fisher 344 rats( Reference Eckles-Smith, Clayton and Bickford 37 ). (3) The study of 60 % ER for 4 weeks in male rats demonstrated that ER stabilised synaptophysin expression and preserved cognitive function( Reference Adams, Shi and Linville 23 ). (4) A recent study demonstrated that 70 % ER for 28 d improved metabolic effects and cognitive decline via up-regulation of brain-derived neurotrophic factor and decreased hippocampal oxidative stress in the obese-metabolic syndrome male rats induced by a moderated HFD (35 %E from fat)( Reference Kishi, Hirooka and Nagayama 19 ). The different findings between previous studies and the present study could be dependent on the percentage of HFD feeding and the duration of ER. Future studies are needed to verify this issue.
One important finding from this study is that ER combined with vildagliptin restored peripheral insulin sensitivity, brain insulin sensitivity, brain mitochondrial function, synaptic plasticity and cognitive function in obese insulin-resistant rats. This means of intervention provided better efficacy than ER alone in this model. These findings suggest that the beneficial effects on brain function in this group could be because of the action of vildagliptin. This is consistent with our previous studies, which showed that vildagliptin led to the restoration of cognitive function in the obese model, via improved brain insulin sensitivity and reduced brain oxidative stress( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pintana, Apaijai and Chattipakorn 8 ). As demonstrated in this study, the combined ER with vildagliptin improved/restored both brain insulin resistance and brain oxidative stress better than ER alone. These findings indicated that ER in combination with the vildagliptin treatment provided more beneficial effects than ER alone in our study model. However, other ER feeding regimens and delivery could have given similar or different results as observed in the present study and require further investigation.
Limitations of the study
The two-factorial design study was not conducted in the present study; therefore, the effects of ER or its combination with vildagliptin in ND-fed rats (lean rats) were not known. However, our previous studies showed that the administration of vildagliptin in ND-fed rats has no effect on the metabolic parameters or brain function, when compared with vehicle-treated ND-fed rats( Reference Pipatpiboon, Pintana and Pratchayasakul 4 , Reference Pintana, Apaijai and Chattipakorn 8 ). In addition, ER had been shown to exert benefits in attenuating cognitive decline via an up-regulation of the brain-derived neurotrophic factor and tropomyosin-related kinase B in obese, hypertensive rats( Reference Kishi, Hirooka and Nagayama 19 ); these markers were not investigated in this study.
Conclusion
This study indicates that although ER restored peripheral insulin sensitivity and decreased circulating oxidative stress levels regardless of the addition of vildagliptin, only ER plus vildagliptin exerted the best efficacy in effectively decreasing brain oxidative stress levels, as well as the restoration of brain insulin sensitivity, brain mitochondrial function, hippocampal synaptic plasticity and cognition. All of these findings suggest that the addition of vildagliptin to ER provides greater therapeutic benefits on brain function than ER alone in obesity.
Acknowledgements
This work was supported by grants from the Thailand Research Fund (TRF-BRG5780016 (SC), TRF-MRG5980198 (WP)), the Royal Golden Jubilee PhD programme (PHD/0025/2555 HP&SC), National Research Council of Thailand (SC), a NSTDA Research Chair Grant from the National Science and Technology Development Agency (NC) and Chiang Mai University Excellent Center Award (NC).
H. P., N. C., S. C. C. performed the experiments, analysed the data and wrote the manuscript. P. T., W. P., P. Sa-nguanmoo, T. C. performed the experiments and analysed the data. P. Satjaritanun, L. L. analysed the data.
The authors have no conflicts of interest to disclose.
