Glucose Tolerance in Response to a High-Fat Diet Is Improved by a High-Protein Diet




Consumption of a high-fat (HF) diet results in insulin resistance and glucose intolerance. Weight loss is often recommended to reverse these metabolic alterations and the use of a high-protein (HP), low-carbohydrate diet is encouraged. In lean rats, consumption of a HP diet improves glycemic control. However, it is unknown whether this diet has a similar effectiveness in rodents with impaired glucose tolerance. Rats were fed a HF or a chow (CH) diet for 6 weeks and then switched to a HP diet or a CH or pair-fed (PF) to the amount of kcals consumed per day by the HP group. Following the diet switch, body weight gain was attenuated as compared to HF rats, and similar between HP, CH, and PF rats. Despite similar weight progression, HP and PF rats had a significant decrease in body fat after 2 weeks, as compared to HF rats. In contrast, CH rats did not show this effect. Glucose tolerance was attenuated more quickly in HP rats than in CH or PF rats. These results indicate that a HP diet may be more effective than a balanced diet for improving glycemic control in overweight individuals.


The prevalence of obesity has remained high in recent years, with ∼68% of US adults classified as overweight or obese in 2007–2008 (1). Paralleling this obesity epidemic is a rise in the incidence of type 2 diabetes (2,3), and an increased occurrence of type 2 diabetes in children and young adults, increasing their risk for other medical ailments later in life (3). Although weight loss is often recommended as a strategy for improving glucose and insulin sensitivity, there is some suggestion that the method by which an individual chooses to lose weight may differentially affect glycemic control (4).

High-protein (HP), low-carbohydrate diets are widely used in weight management programs and have been demonstrated to be effective for weight loss in some individuals (5). In addition, these diets have also been shown to improve glycemic control in obese, hyperinsulinemic, or diabetic subjects (6,7,8). While these diets show great promise for the treatment of the metabolic syndrome, it is important to note that many of these diets also reduce caloric intake, making it difficult to determine whether the calorie restriction, weight loss, or macronutrient content of the diet is responsible for the observed improvement in insulin sensitivity and glucose tolerance (9,10,11,12).

When energy intake is controlled, there is evidence that increasing protein content while decreasing carbohydrate content leads to increased satiety (13). Altering macronutrient content in this way also improves glucose homeostasis in both animals and humans (8,13–15). Blouet and colleagues demonstrated that, in healthy rats, maintenance on a diet with a higher protein—carbohydrate ratio improved glucose homeostasis, such that by 7 weeks, the glucose area under the curve (AUC) from an oral glucose tolerance test was significantly decreased. This effect was independent of caloric intake, as a group pair-fed to high-protein-fed rats failed to demonstrate similar improvements in glucose tolerance (15). However, it is unknown whether this diet has a similar effectiveness in rodents with reduced glycemic control.

In the current study, we examined the effects of switching to a HP diet following the consumption of a high-fat (HF), high-carbohydrate diet in male rats. This HF diet has been shown previously to result in weight gain and decreases in glucose tolerance and insulin sensitivity. We hypothesized that animals fed a HF diet and subsequently switched to chow (CH) or a HP diet would show a similar attenuation in weight gain, but that switching to the HP diet would result in a more rapid improvement of glucose homeostasis.

Methods and Procedures

Animals and diets

Male Long Evans (Blue Spruce) rats (n = 40; Harlan Laboratories, Indianapolis, IN) weighing ∼290 g at the start of the experiment were individually housed under a 12/12 h light:dark cycle (0100/1300 hours) in a temperature and humidity controlled room (22°C, 50–60% humidity). All procedures were approved by the Purdue University Animal Care and Use Committee.

Rats were maintained on standard laboratory chow (2018; Harlan Teklad, Indianapolis, IN) before the start of the experiment. Body weight and caloric intake were recorded daily. In phase one of the experiment (P1), animals were weight matched, and divided into two groups. One group (CH; n = 8) was fed standard chow for the duration of the experiment, while the other group (HF; n = 32) was maintained on a HF diet (D12492; Research Diets, New Brunswick, NJ). After 6 weeks of maintenance on HF (Phase 2 (P2)), rats were weight matched and divided further into four groups; one group (HF; n = 8) continued to consume the HF diet. A second group (HP; n = 8) was placed on a HP diet (D09102501; Research Diets), and the third and fourth groups were placed on a diet matched in fat and carbohydrate content to the HP diet, but with protein levels similar to that of standard chow (D10012M; Research Diets). One group (CH; n = 8) received ad libitum access to this diet, while the other group (PF; n = 8) was pair-fed to the calories consumed by the HP group. Rats were maintained on these diets for the remainder of the experiment. The macronutrient content of all diets is listed in Table 1.

Table 1.  Diet composition
inline image

Body composition analysis

Body composition was analyzed in conscious rats using the EchoMRI Rat Tissue Composition Analyzer (Echo Medical Systems, Houston, TX) during weeks 5, 8, 10, and 12. Data are presented as percent body fat and percent lean mass.

Glucose tolerance test

Rats underwent an intraperitoneal glucose tolerance test (IPGTT) immediately before, 1 week following, and 3 weeks following the diet switch (weeks 5, 7, and 9, respectively). Rats were food deprived at 1700 hours on the day before the test. At 1030 hours, rats were removed from their home cages and placed into a testing cage. A small amount of blood was collected via tail nick and baseline blood glucose was analyzed with the Precision Xtra Glucose Monitoring System (Abbott Laboratories, Abbott, IL). Additional blood was collected into iced K2EDTA-coated tubes (Vacutainer; BD, Franklin Lakes, NJ) for subsequent measurement of plasma insulin levels. Following this baseline measurement, rats were given an intraperitoneal injection of glucose (1.5 g/kg body weight) in a volume of 1.0 ml. Blood glucose levels were again measured, with additional blood collected as at baseline, 15, 30, 60, and 120 min following glucose injection. Blood was then centrifuged at 2,000 rpm for 15 min at 4°C, and plasma was removed and stored at −80°C, until further processing was conducted.

Enzyme-linked immunosorbent assay

Plasma insulin levels in blood obtained during the IPGTT were assessed using the wide-range (0.1–12.8 ng/ml) Ultra-Sensitive Rat Insulin Enzyme-linked Immunosorbent Assay kit (90060; Crystal Chem, Downers Grove, IL), according to manufacturer's instructions. The inter- and intra-assay precision coefficient of variation was ≤10.0%. Unknown concentrations of insulin were interpolated from a standard curve generated for each microplate.

Terminal blood and tissue collection

On the day of sacrifice (week 15), food was removed 8 h before the onset of the dark cycle. Six hours later, rats were sacrificed under ether inhalation anesthesia, followed by rapid decapitation. A midline incision was then made and the skin was separated from the underlying muscle layer between the right fore and hind limbs, allowing for the removal and weighing of this subcutaneous fat pad. Both inguinal and dorsal white fat were dissected completely and by the same person to ensure consistency. Right epididymal and retroperitoneal fat pads were also removed and weighed. Trunk blood was collected into K2EDTA-coated tubes, and treated as described above, and stored at −80°C until processing for plasma leptin levels. Data from these terminal analyses are displayed in Table 2.

Table 2.  Fat pad weights and plasma leptin levels
inline image


Plasma leptin levels were determined using a commercial radioimmunoassay kit (Millipore, St Charles, MO). All samples were run in duplicate and per manufacturer's instructions. The plasma leptin radioimmunoassay was run with upper and lower detection limits of 0.5 ng/ml and 50 ng/ml, respectively. Unknown concentrations of leptin were calculated based on the standard curve generated for the kit.

Statistical analysis

Data are represented as mean ± s.e.m. For P1, comparisons are made between CH and HF groups. For P2, the comparisons were made to the HF group, as the hypothesis was that switching to the HP diet would more rapidly improve glucose tolerance than switching to CH, as compared to glucose tolerance while maintained on the HF diet. Body weight, caloric intake, blood glucose, and plasma insulin and caloric intake were first analyzed by two-way ANOVA (diet × time). The AUC for each week of IPGTT testing was analyzed using Student's t-tests, and plasma leptin, body composition, and fat pad weights were analyzed using one-way ANOVA. Data from P2 were analyzed by Dunnett's multiple comparison tests, as appropriate.


Body weight and caloric intake

At the start of P1, during which rats were maintained on CH or HF diet for 6 weeks, the mean body weight was 287.2 ± 5.4 g. As depicted in Figure 1a, there was a significant main effect of diet (P < 0.0001) during this phase of the experiment, such that HF-fed animals weighed significantly more than CH animals by week 2 (P < 0.01). This effect persisted throughout the first 6 weeks of the study (P < 0.001 for weeks 3–6), such that CH rats weighed significantly less than HF rats (401.8 ± 19.5 g vs. 462.6 ± 11.7 g, respectively, (P < 0.05)). After 6 weeks, HF rats were switched to HP, CH, or PF diets, or remained on HF. Following the diet switch, there was a significant main effect of diet (P < 0.05) (Figure 1b). All groups that were switched from the HF diet (HP, CH, and PF) weighed significantly less than HF rats by week 9 (P < 0.05), and maintained lower body weights throughout the rest of the experiment. Body weights at the end of the experiment, were: HP (472.5 ± 11,5 g), CH (489.7 ± 25.8 g), and PF (450.6 ± 12.1 g), whereas HF rats weighed significantly more than the other groups at this time (544 ± 14.4 g, P < 0.05, as compared to all other dietary groups).

Figure 1.

Body weight change in response to diet. (a) During the first 6 weeks, rats gained significantly more weight on a high-fat (HF) diet, as compared to chow. (b) Following a diet switch, HP, CH, and PF rats gained significantly less weight than HF. Values are expressed as mean ± s.e.m.; *P < 0.05 for all groups compared to HF; #P < 0.05 for PF compared to HF. CH, chow; HP, high-protein; PF, pair-fed.

Analysis of caloric intake revealed a significant main effect of diet before the diet switch (P < 0.001). During this time, CH animals consumed significantly fewer calories than HF animals during weeks 1–5, however their food intakes were comparable during week 6 (Figure 2a). Following the diet switch, there was also a significant effect of time during P2 (P < 0.001), marking a significant, but transient, drop in food intake for animals whose diets were switched (HP, AL, and PF). In addition, there was a main effect of diet (P < 0.01). As depicted in Figure 2b, during weeks 7, 8, and 9, caloric intakes for HP, CH, and PF animals were significantly lower than HF (P < 0.05, for all groups) animals. Mean caloric intake was also less in PF and CH rats during week 12, as compared to HF rats (P < 0.05, for all).

Figure 2.

Caloric intake over the course of the experiments. (a) Caloric intake by week for the first 6 weeks of the experiment. Rats maintained on high-fat diet consumed significantly more kcals/day than CH animals. (b) Following the diet switch animals fed HP, CH, and PF consumed fewer kcals than HF during weeks 7, 8, and 9. HP and PF rats continued to consume fewer kcals than HF during week 12. Values are expressed as mean ± s.e.m.; *P < 0.05 for all groups compared to HF. #P < 0.05 for HP and PF, as compared to HF. CH, chow; HF, high-fat; HP, high-protein; PF, pair-fed.

Body composition

At week 5 and before the diet switch, there were significant differences in percentage of body fat (P < 0.0001) and percentage of lean tissue (P < 0.001, data not shown) between CH and HF groups (Figure 3a). CH animals had a lower percentage of adiposity and a higher percentage of lean mass, as compared to HF animals. Two weeks after the diet switch (week 8), HP and PF animals exhibited a significantly lower body fat percentage than HF animals (P < 0.05, Figure 3b). At week 10, all three groups that switched from HF diet had significantly lower percentages of body fat, as compared to HF animals (Figure 3c (P < 0.05 for HP, CH, and PF)). The same results were observed at week 12 (Figure 3d). There were no differences in lean body mass at weeks 8, 10, or 12 (data not shown).

Figure 3.

Body composition values as measured by EchoMRI. (a) During week 5, high-fat (HF) diet-fed animals had a significantly higher percentage of body fat than CH rats. (b) By week 8, PF rats' body fat percentages were significantly lower than HF. (c,d) During weeks 10 and 12, HP and PF rats' body fat percentages were significantly lower than HF. Values are expressed as mean ± s.e.m.; *P < 0.05, as compared to HF. CH, chow; HP, high-protein; PF, pair-fed.

Effect of diet on glucose tolerance

IPGTTs were conducted during week 5, 7, and 9. During week 5, and before the diet switch, HF-fed rats exhibited elevated fasting blood glucose (Figure 4a, P < 0.01) and plasma insulin (Figure 5a, P < 0.01) levels. During the IPGTT, HF-fed rats exhibited significantly increased blood glucose levels after 60 and 120 min, as compared to CH-fed rats (Figure 4a, P < 0.05). Despite the fact that animals fed a diet high in fat had a markedly more pronounced insulin surge in response to glucose injections (Figure 5a, P < 0.001), their elevated glucose levels persisted such that at 120 min following injection, HF fed-rats maintained blood glucose levels higher than those of CH rats (Figure 4a, P < 0.05).

Figure 4.

Glucose responses to IPGTT. (a) During week 5, baseline blood glucose levels and glucose values at 60 and 120 min were significantly higher in high-fat (HF) diet-fed rats than CH rats. (b) AUC values were also significantly elevated in these animals. At week 7 (c) glucose levels were not significantly different between groups (d) AUC values for HP, AL, and PF rats were not significantly different from HF or CH. (e) By week 9, HP rats had significantly lower glucose at baseline, 30 and 60 min; PF rats were significantly reduced at 30, 60, and 120 min (as compared to HF). (f) AUC were significantly reduced after maintenance on HP, as compared to HF. Data are expressed as mean ± s.e.m.; *Significantly different than HF, P < 0.05. AUC, area under the curve; CH, chow; HP, high-protein; IPGTT, intraperitoneal glucose tolerance test; PF, pair-fed.

Figure 5.

Insulin responses to IPGTT. (a) During week 5, plasma insulin levels were significantly higher in high-fat (HF) diet-fed rats than CH rats at all timepoints. (b) AUC values were also significantly elevated in these animals. (c) At week 7, HP insulin levels were lower than HF at baseline, 15, 30, and 60 min. In CH rats, baseline and 15 min values were significantly lower; in PF rats, 30 and 60 min values were significantly lower, as compared to HF rats. (d) AUC values, however, were significantly reduced in HP rats only. (e) During week 9, plasma insulin levels were significantly lower than those of HF rats in PF rats at all timepoints, and lower in HP rats at all points except baseline. (f) AUC values were significantly lower in HP and PF, but not CH rats. Data are expressed as mean ± s.e.m.; *Significantly different than HF, P < 0.05. AUC, area under the curve; CH, chow; HP, high-protein; IPGTT, intraperitoneal glucose tolerance test; PF, pair-fed.

During week 7, there were no significant differences in blood glucose levels (Figure 4c) among diet-switched rats, however, plasma insulin AUC values were significantly lower in the HP group, as compared to HF (Figure 5d, P < 0.05). By week 9, HP blood glucose levels were significantly lower than blood glucose levels of HF rats (P < 0.05, Figure 4e). Plasma insulin levels were significantly lower than HF in HP and PF rats (Figure 5e, P < 0.05); there was no statistical difference between insulin levels of CH rats as compared to HF at week 9.


There are discrepancies in the literature regarding the effects of dietary macronutrient composition on energy homeostasis. In terms of glucose tolerance, diets with proportionately different macronutrient content, particularly those high in protein and low in carbohydrate, have been reported to either improve or have no effect on glucose homeostasis (16). Consumption of a HF diet is known to produce obesity and decrease glucose tolerance (17,18,19). In the current experiments, rats maintained on a HF diet exhibited characteristic effects of diet-induced obesity. HF rats gained significantly more weight, had a higher percentage of body fat, and were hyperphagic. In addition, signs of glucose intolerance, including elevated fasting blood glucose and plasma insulin levels, and increased insulin AUC, were present in HF rats during a glucose tolerance test. While it is known that the discontinuation of a HF diet can induce weight loss and improve glucose tolerance (20), it is currently unknown how this response is affected by the macronutrient content of the new diet.

We have previously demonstrated that significantly altering the macronutrient content of the diet to increase the proportion of both fat and protein had negative effects on glucose tolerance. When rats were maintained on this diet, responses to peripherally administered insulin were impaired, although rats had increased responsivity to the anorectic effects of insulin when administered centrally. Further, we have demonstrated that this reduced tolerance to glucose and insulin was rapidly reversed upon switching from the low-carbohydrate, HF diet to a normal CH diet (20). Based on these data and previous data demonstrating rats consume fewer calories per day when given a HP diet as compared to CH or a low-carbohydrate, HF diet (21), we hypothesized that switching from a HF diet to a HP or CH diet would attenuate weight gain, but that glucose tolerance would improve more rapidly in rats fed the HP diet.

In these experiments, alterations in body weight and caloric intake following a switch from a HF diet are not dependent upon macronutrient intake; the changes in body weight observed following the diet switch were similar in HP, CH, and PF rats. However, HP and PF rats did exhibit a slight, but significant reduction in body weight, such that they gained significantly less weight than did HF rats at week 9 and week 10. The reduced caloric intake, rather than the macronutrient content of the diet, appears to account for the reduction in body weight in this study. The body weight change is due to a significant decrease in body fat levels, rather than through a decrease in muscle mass.

These results are in agreement with previous research indicating that a variety of macronutrient compositions are able to produce weight loss (11,22,23,24,25), as long as there is a concomitant reduction in caloric intake (26). Even so, there is evidence suggesting that the macronutrient content of the diet plays a role in the level of satiety one experiences during consumption, perhaps rendering reduced-calorie, macronutrient-controlled diets more tolerable (27,28). Each diet in the present study resulted in a decrease in adiposity without any detrimental change in lean mass as the nuclear magnetic resonance data showed differences in fat mass, but no differences in lean mass across groups. While this has previously been demonstrated in HP diets, mixed results have been reported for high-carbohydrate, low-fat diets similar to those consumed by CH and PF rats. It is likely that the fat mass was reduced in HP and PF rats, but not CH because of small differences in food intake over the course of the experiment. Alternatively or in addition, the PF and HP groups may have expended more energy over the course of the experiment. Future research including analyses of energy expenditure would be informative.

The effects on glucose tolerance observed in the current study are dependent on the macronutrient content of the diet, rather than caloric intake alone. One week after switching from HF diet to the HP, CH, or PF condition, glucose responses to the IPGTT were not different from those of HF-fed rats. However, less insulin was required to produce this level of glucose homeostasis in HP rats, as indicated by a significant decrease in insulin AUC, as compared to HF rats. This effect was also present at week 9, accompanied by a significant decrease in glucose AUC. After two additional weeks of diet consumption, PF rats demonstrated a similar increase in insulin effectiveness, as demonstrated by significantly decreased insulin AUC. Taken together, these results demonstrate improvements in glucose tolerance in HP rats and attenuation of hyperglycemia that is more rapid in HP rats than in PF rats that consume the same number of calories per day and do not differ in body weight or adiposity.

Previous research has suggested that dietary protein affects glucose tolerance. Blouet and colleagues demonstrated that a HP diet improved glucose tolerance in healthy rats more effectively than a CH diet (15,29). However, it is difficult to determine if the effect can be fully attributed to the elevated protein content, or the result of a simultaneous decrease in carbohydrate intake. Diets supplemented with amino acids appear to improve glucose tolerance, suggesting the increased amino acids present in the protein-adjusted diet are critical. Supplementing a diet with amino acids results in improved insulin sensitivity in humans (29,30,31,32). In addition, it has been determined that glycemic control in rats can be affected by the nature of the dietary proteins utilized, including the type of amino acids and the protein source (33,34,35). Taken together, these results suggest that protein itself may have an effect on glycemic control, independent of carbohydrate intake.

In summary, this study provides an important contribution to the ongoing debate on the influence of macronutrients in glycemic control. Similar to the effects of a HP diet in normoglycemic rats, hyperglycemic rats experienced an improvement in glucose tolerance that was more rapid when fed a HP diet, an effect that was dependent on macronutrient content rather than caloric intake. Overall, our results demonstrate the ability of a HP diet to rapidly improve HF diet-induced alterations in glucose tolerance in rats. Further research is necessary to determine the peripheral and central mechanisms through which a HP diet produces its effects.


This research was supported by NIH grant DK078654 (K.P.K.). This work was also supported by the NIH, NCI R25CA128770 (D. Teegarden) Cancer Prevention Internship Program (M.A.H.), administered by the Oncological Sciences Center and the Discovery Learning Research Center at Purdue University. The assistance of Meredith Cobb is gratefully acknowledged.


The authors declared no conflict of interest.