There is a growing prevalence of obesity and low insulin sensitivity (IS), often called insulin resistance, in human subjects. Thus, there is a need for an animal model to study the time course of these metabolic disturbances as well as their unhealthy consequences. Different animal models have been used to study obesity and IS, notably the rat, in which obesity can be caused by genetic mutations or induced by nutritional interventions. As human obesity is mainly due to nutritional habits, animal models of obesity and low IS induced by specific diets may be preferable to genetic models.
Various nutritional interventions have been used to induce obesity, low IS and dyslipidaemia in rats. High-fat diets have been shown to cause these metabolic disorders in previous studies, but there has been a large variability in the intensity of the metabolic changes(Reference Morens, Sirot and Scheurink1–Reference Matveyenko, Gurlo and Daval5). High-fructose diets have also been shown to lower IS and promote mild-to-severe dyslipidaemia(Reference Nakagawa, Hu and Zharikov6–Reference Kannappan, Palanisamy and Anuradha9). The differences in nutritional interventions, such as diet composition and interventional duration, have complicated the comparisons of these studies. Therefore, it is difficult to define the best nutritional intervention to induce obesity in an animal model that closely mimics the human disease. Longitudinal studies could be useful in determining the best diet to induce obesity and related disorders. To our knowledge, this type of approach has never been conducted in rats. Here, we aimed to perform a longitudinal study to compare the effects of a high-fat diet and a high-fructose diet on body weight (BW), IS and plasma lipid profiles in rats.
Materials and methods
Animal groups and diets
Male Wistar rats (12 weeks old; Janvier, Le Genest Saint-Isle, France) were randomly separated into three groups (six per group): control, high fat or high fructose. According to their groupings, the rats were fed with a control diet (39·7 % maize starch, 20 % dextrose, 5·8 % sunflower oil and 20·5 % casein by weight), a high-fat diet (12·7 % maize starch, 6·5 % dextrose, 3·9 % sunflower oil, 31·3 % lard and 28·6 % casein by weight) or a high-fructose diet (59·7 % fructose, 5·8 % sunflower oil and 20·5 % casein by weight) for 10 weeks. The rats were housed in individual cages, with free access to the feed and water. The rats were maintained under a 12-h light–12-h dark cycle and a temperature of 22 ± 2°C. The animals were housed at Oniris (National College of Veterinary Medicine, Food Science and Engineering, Nantes, France), according to the regulations for animal welfare of the French Ministry of Agriculture. The experimental protocol adhered to the European Union guidelines and was approved by the local animal use and care advisory committee.
Body weight and body fat mass
BW was recorded weekly. The body fat mass was determined by isotope dilution (2H2O; Eurisotop, Gif-sur-Yvette, France) on week 1 (before the dietary intervention) and at the end of the 10-week diet period (week 11). Blood samples (1 ml) were collected before and 2 h after a 2H2O injection (500 mg/kg BW). Plasma 2H2O concentrations were measured using Fourier-transformed IR spectroscopy (Brüker SA, Wissembourg, France).
Euglycaemic–hyperinsulinaemic clamp technique
The euglycaemic–hyperinsulinaemic clamp technique was performed before and at the end of the 10-week diet period. The catheter was inserted under anaesthesia into the jugular vein of an animal that was not fed overnight. Insulin (Actrapid; Novo Nordisk, Bagsvaerd, Denmark) was perfused [72 mU/kg (500 pmol/kg) for 1 min, then 18 mU/kg per min (125 pmol/kg per min) for 3 h] and glycaemia was measured at every 5 min. Glucose (20 %; Braun Medical SAS, Boulogne Cedex, France) was perfused at a variable rate. The glucose infusion rate (mg/kg per min) was adjusted to attain and maintain the basal glycaemia. In hyperinsulinaemic conditions, glucose infusion rate measures the insulin-mediated glucose uptake and is considered as a good reflection of IS. Glycaemia was measured using the glucose oxidase method (Accu-Chek Active; Roche Diagnostics, Mannheim, Germany). Insulin was measured using ELISA (Rat Insulin; Shibayagi Company Limited, Gunma, Japan).
Plasma lipid profiles
The basal plasma concentrations of total cholesterol, NEFA and TAG were assayed before and at the end of the 10-week diet period using enzymatic methods (Cholestérol RTU, BioMérieux, Marcy-l'Etoile, France; Triglycérides enzymatique TG PAP150, BioMérieux; and NEFA C, Wako, Oxoid, Dardilly, France).
The plasma samples were pooled for each group and each time period (weeks 1 and 11), and the plasma lipoproteins were separated using a fast-protein liquid chromatography system (UNICORN 520; GE Healthcare, Pittsburgh, PA, USA). The cholesterol concentration was measured in each fraction.
Data analysis was performed using Statview software (version 5.0; SAS Institute Inc., Cary, NC, USA) and R software (version 2.10, lme4 package; R Foundation for Statistical Computing, Vienna, Austria). Data were expressed as mean values with the standard error of the mean. A linear mixed-effects model has been performed in order to study the effect of time, type of diet and the interaction between them for each variable. The mixed-effects models are the most efficient way to analyse repeated measurements data(Reference Davidian and Giltinan10). A multiple comparison of means procedure with Tukey contrasts, adapted to the mixed-effects models, has been used when the interactions between time and the type of diet were significant. A significant difference has been considered for P value < 0·05.
Body weight and body fat mass
Table 1 gives the values of BW and body fat mass of the control, the high-fat and the high-fructose groups before and at the end of the 10-week period. The BW gain was similar for all groups. At week 11, the high-fat group had significantly higher (P < 0·05) body fat mass compared to the control group. At week 11, compared to initial values from week 1, the high-fat diet caused a 393 (sem 170) % increase in body fat mass and high-fructose diet caused a 139 (sem 23) % increase.
BW, body weight; GIR, glucose infusion rate.
Mean values were significantly different among groups for the interaction between time and the type of diet (fixed-effects model analysis revealed that values in the high-fat group is significant at week 11, compared to the control group; P < 0·05).
Mean values were significantly different among groups for the effect of time (significant for week 11 v. week 1): † P < 0·05 and ††† P < 0·0001.
‡ Mean values were significantly different for the interaction between time and the type of diet; the effect of time and type of diet were not considered (P < 0·05).
§ n 5 for GIR (in hyperinsulinaemic conditions, GIR measures the insulin-mediated glucose uptake and is considered as a good reflection of insulin sensitivity).
∥ At week 1, plasma total cholesterol concentration was different among groups.
No differences were observed in basal glycaemia and insulinaemia among groups. The high-fat group had lower IS compared to the control group at week 11, assessed by significantly lower (P < 0·05) glucose infusion rate value. The high-fructose group had no difference in IS compared to the control group (Table 1).
Plasma lipid profiles
Plasma TAG concentration was not different among groups at any time. Plasma NEFA concentration decreased over time in all groups, regardless of the diet.
At week 1, the plasma total cholesterol concentration was different among groups. Indeed, it was higher in the control group than in others. There was a 38 (sem 12) % increase in total cholesterol concentration in rats fed with high-fat diet and a 15 (sem 8) % increase in rats fed with high-fructose diet, but the difference did not reach level of significance (Table 1). Nevertheless, plasma total cholesterol concentration increased over time, regardless of the diet.
On the basis of a pooled sample from each of the groups, there is a preliminary indication that there was no difference due to diet (data not shown) in the plasma lipoprotein fractions, but these results must be considered to be very preliminary.
The aim of the present study was to compare the effects of a high-fat and a high-fructose diet on obesity-related disturbances in rats by using a longitudinal approach. In the present study, all the basal values were in the range of normal values of rats. In the literature, various concentrations of plasma lipids have been measured in control rats. However, we did not find any study of obesity and IS in rats with the baseline values of all variables. Therefore, we used a longitudinal approach and composed the groups randomly. The plasma total cholesterol concentration was significantly different among groups at baseline level (week 1). The randomisation process used to compose the groups had been expected to abolish all the differences among groups, and all cholesterol concentrations have been measured at the same time, i.e. at the end of the study, to guarantee same analytical conditions. Unfortunately, the randomisation process failed to reach its objectives, at least in this regard (there was no other difference among the groups at week 1). The baseline values were then taken into account in the statistical analysis in order to nullify the pre-existing differences. Another factor is the age at which exposure to a high-fat or high-fructose diet started. It has been shown that adult rats fed with high-fructose diet produce signs of metabolic syndrome, but young rats do not(Reference de Moura, Ribeiro and de Oliveira11). We therefore chose adult rats to conduct the present experiment.
We observed similar BW at the end of the 10-week period in all groups. However, rats fed with high-fat diet had higher body fat mass compared to the control rats. Increase in adiposity has previously been described in rats fed with a high-fat diet(Reference Sinitskaya, Gourmelen and Schuster-Klein2–Reference Matveyenko, Gurlo and Daval5, Reference Posey, Clegg and Printz12), and it was associated with increase in weight gain. An increase in BW alone does not necessarily represent obesity, but other factors have to be taken into account, such as changes in body composition. In the present study, the increase in body fat mass reflected the obesogenic property of the high-fat diet, which was due to lard being used as the source of dietary fat. This excess fat was stored in the form of adipose tissue in the body. In rats fed with high-fructose diet, we observed no change in BW and body fat mass compared to the control rats. Previous studies showed that administering high-fructose (60 % by weight) diet for about 10 weeks increased BW(Reference Nakagawa, Hu and Zharikov6, Reference Mohamed Salih, Nallasamy and Muniyandi8, Reference Shih, Lin and Lin13) and epididymal and retroperitoneal fat depots(Reference Shih, Lin and Lin13). However, de Moura et al. (Reference de Moura, Ribeiro and de Oliveira11) observed no difference in BW in adult rats administered fructose, despite higher retroperitoneal, mesenteric and subcutaneous fat depots weights. This suggests that high-fructose diet does not necessarily lead to obesity.
The rats fed with high-fat diet had lower IS (assessed by glucose infusion rate, which is considered as a good reflection of IS, as it measures the insulin-mediated glucose uptake under hyperinsulinaemic conditions) compared to the control group, as previously shown in several studies(Reference Kraegen, Clark and Jenkins14, Reference Oakes, Cooney and Camilleri15). Moreover, low IS has been shown to be associated with an increase in epididymal fat(Reference Sridhar, Vinayagamoorthi and Arul Suyambunathan3, Reference Vinayagamoorthi, Bobby and Sridhar4), and the present results of higher body fat mass in rats fed with high-fat diet are consistent with the previous results. However, basal glycaemia and insulinaemia were unchanged. Some studies using high-fat diet (lard or safflower oil) reported low IS, associated with hyperinsulinaemia(Reference Sridhar, Vinayagamoorthi and Arul Suyambunathan3–Reference Matveyenko, Gurlo and Daval5, Reference Posey, Clegg and Printz12, Reference Vikram, Jena and Ramarao16) and hyperglycaemia(Reference Vikram, Jena and Ramarao16). We think that we could have assessed low IS by euglycaemic–hyperinsulinaemic clamp technique before the appearance of hyperglycaemic and hyperinsulinaemic states, and that the present experiment could have been long enough to develop insulin resistance but not diabetes.
On the other hand, we found no change in IS, basal glycaemia and insulinaemia in rats fed with high-fructose diet. Many previous studies have reported hyperglycaemia(Reference Liu, Tzeng and Liou7–Reference Kannappan, Palanisamy and Anuradha9, Reference Shih, Lin and Lin13), but Nakagawa et al. (Reference Nakagawa, Hu and Zharikov6) did not show any modification in glycaemia after administering high-fructose diet. The higher hepatic glycogen content that has been described in rats fed with high-fructose diet(Reference Yadav, Jain and Yadav17) could prevent hyperglycaemia. Previous studies have also shown that the high-fructose diet developed low IS and was associated with increased plasma TAG and NEFA concentrations(Reference Nakagawa, Hu and Zharikov6–Reference Mohamed Salih, Nallasamy and Muniyandi8). These increased concentrations in response to high-fructose diet could have an important role in the development of low IS by reducing insulin signalling pathway (reviewed in Tappy & Le(Reference Tappy and Le18)). However, we did not find dyslipidaemia, which is in accordance with the unchanged IS. Moreover, in a study in rats, high-sucrose diet has been shown to induce low IS in the liver before muscle(Reference Pagliassotti, Prach and Koppenhafer19). We assessed the IS by using the euglycaemic–hyperinsulinaemic clamp technique, which is the gold standard method for direct assessment of the whole-body IS(Reference DeFronzo, Tobin and Andres20). We suggest that high-fructose diet could have caused minor impairment of insulin action in the liver, but could not have caused whole-body low IS.
We observed no difference in plasma total cholesterol concentration between the groups, but there was a significant increase over time, regardless of the diet. This increase could have reflected an effect of age, whereas the absence of dietary cholesterol could be a possible explanation for the absence of any specific effect of the diet. Indeed, additional dietary fat, whatever its nature, had no effect on plasma cholesterol concentration in the absence of dietary cholesterol in a study conducted in hamsters(Reference Sessions and Salter21). In the present study, the non-significant increase in plasma total cholesterol in rats fed with high-fat diet could be due to small amount of dietary cholesterol in lard. On the other hand, previous studies reported variable responses of the high-fructose diets on plasma total cholesterol concentration(Reference Liu, Tzeng and Liou7, Reference Mohamed Salih, Nallasamy and Muniyandi8, Reference Shih, Lin and Lin13, Reference Stark, Timar and Madar22), and even the same high-fructose diets behaved differently(Reference Liu, Tzeng and Liou7, Reference Shih, Lin and Lin13). Thus, it is difficult to conclude about the effects of fructose diet on cholesterol concentration. Moreover, the dietary cholesterol and amount of fat in a high-fructose diet could affect plasma cholesterol concentration. We also observed no difference in cholesterol concentrations in plasma lipoproteins, as reported by Sinitskaya et al. (Reference Sinitskaya, Gourmelen and Schuster-Klein2). Conversely, Mohamed Salih et al. (Reference Mohamed Salih, Nallasamy and Muniyandi8) reported an increase in VLDL-cholesterol and LDL-cholesterol, and a decrease in HDL-cholesterol at the end of a high-fructose diet. The present results for lipoprotein-cholesterol concentrations are in accordance to unchanged plasma total cholesterol concentration, and absence of dietary cholesterol could possibly be the cause for unchanged cholesterol concentration.
Plasma basal NEFA concentration was not different among groups. Variable changes in plasma NEFA concentration have been described in response to a high-fat diet(Reference Morens, Sirot and Scheurink1, Reference Sinitskaya, Gourmelen and Schuster-Klein2, Reference Posey, Clegg and Printz12). The present results in rats fed with high-fructose diet were not consistent with some previous reports(Reference Mohamed Salih, Nallasamy and Muniyandi8, Reference Shih, Lin and Lin13). We hypothesise that the liver could have higher hepatic TAG storage(Reference Benhizia, Hainault and Serougne23), by capturing a bulk amount of plasma NEFA, which would explain that no difference in plasma NEFA concentration was observed.
We found no change in plasma basal TAG concentration in any group. Some studies showed an increase in plasma TAG concentration at the end of a high-fat diet(Reference Sridhar, Vinayagamoorthi and Arul Suyambunathan3, Reference Vinayagamoorthi, Bobby and Sridhar4), whereas others with nearly the same fat content did not(Reference Morens, Sirot and Scheurink1, Reference Sinitskaya, Gourmelen and Schuster-Klein2, Reference Matveyenko, Gurlo and Daval5). Variable response on plasma basal TAG concentration has also been reported in rats fed with high-fructose diet(Reference Nakagawa, Hu and Zharikov6–Reference Mohamed Salih, Nallasamy and Muniyandi8, Reference Shih, Lin and Lin13, Reference Stark, Timar and Madar22, Reference Rizkalla, Boillot and Tricottet24). A few studies have reported higher postprandial TAG concentration in rats fed with high-fructose diet, but no change in the basal TAG concentration for a short period (2 weeks)(Reference Koo, Wallig and Chung25) or for a long period (11 months)(Reference Lee, Ko and Hsu26). In the present study, we measured only the basal TAG concentration. We suggest that the rats in our study could have had higher postprandial TAG concentration that could not have been observed in overnight unfed rats.
On the basis of the present findings, feeding 12-week-old rats a high-fat diet for 10 weeks induces obesity and low IS. In contrast, the high-fructose diet produced no change in obesity-related disorders. In the future, it will be useful to study the specific effects of different fats, oils or fatty acids that account for the discrepancy between these studies.
The financial support in the form of doctoral fellowship to M.-Q. Z. from the Higher Education Commission, Pakistan, is greatly acknowledged. The authors are grateful to Philippe Bleis and Samuel Ninet for their technical assistance and for taking care of the animals. The present study was designed by P. N., V. L., K. O. and M.-Q. Z. and supervised by P. N. and V. L; M.-Q. Z. performed the experiments; C. T. and M.-Q. Z. conducted the statistical analysis; M.-Q. Z. and J. L. B. wrote the manuscript; and P. N., V. L., K. O. and J. L. B. critically reviewed the manuscript. The authors declare no conflicts of interest.