INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE PREVALENCE OF OBESITY continues to rise, whereas the success rate for the long-term treatment of obesity has remained dismally low at 10-15% (12, 36). There are a number of metabolic factors that tend to drive some postobese individuals back to their previously high body weight. In humans and rodents, caloric restriction can be associated with a reduction in both resting metabolic rate (16, 17) and sympathetic nervous system activity (1, 15, 37). When rats are weight reduced by caloric restriction after having been made obese by intake of a variety of palatable and/or high-energy (HE) diets, there is a relatively rapid return to the obese baseline once they are allowed free access to almost any diet (4, 19, 24, 32). Changes in neuroendocrine function before, during, and after such restriction (19, 24, 29) suggest that the brain may be the determinant of the higher body weight set point.

A rat model of diet-induced obesity (DIO) has been used to investigate how rats that are genetically predisposed to become obese on a diet relatively high in fat, sucrose, and caloric content (HE diet) differ from those that are diet resistant (DR) (19). This model shares many characteristics of some human obesity. These include polygenic inheritance (20, 34), insulin resistance (20, 31), lowered growth hormone secretion (3, 14), and a propensity to oxidize carbohydrate preferentially over fat (5, 13). About half the adult male Sprague-Dawley rats fed an HE diet develop DIO, whereas the rest gain no more weight than chow-fed controls and are thus resistant (DR) to the HE diet (19, 23, 24). Similar to many obese human beings (16, 17), DIO rats reduce their resting metabolic rate when calorically restricted (7, 10, 11) and return to their previously high body weight when restriction is discontinued (4, 9, 19, 24, 32). In humans, the reduction in metabolic rate appears to be proportional to the initial body weight (17), but such data come from a very heterogeneous group of individuals. The current studies were undertaken to investigate the way in which rats selected for their genetic propensity to be DIO or DR might respond to chronic caloric restriction on a low-fat diet similar to the conditions imposed in weight-reduction programs in humans. An 8-wk period of total restriction was used, because it is equivalent to >2 yr of human life if relative life spans are considered.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals and experimental design. Figure 1 shows the experimental design. The experiment began with 60 male Sprague-Dawley rats (Charles River Labs) that were kept at 23-24°C on a 12:12-h light-dark cycle (lights on at 1700, off at 0500). They were brought into the facility at 300-325 g and kept on Purina rat chow (#5001) and water ad libitum for 1 wk. During this week, 24-h urine was collected for measurement of urinary norepinephrine (NE1). All rats were then switched to an HE diet ad libitum. This diet is composed of 8% corn oil, 44% sweetened condensed milk, and 48% Purina rat chow (Research Diets). It contains 4.47 kcal/g, with 21% of the metabolizable energy content as protein, 31% as fat, and 48% as carbohydrate, 50% of which is sucrose (23). Purina rat chow (#5001) contains 3.30 kcal/g with 23.4% as protein, 4.5% as fat, and 72.1% as carbohydrate, which is primarily in the form of complex polysaccharide (23). After 2 wk on the HE diet (beginning of week 3), the 24 rats with the highest body weight gain were designated as "DIO" and the 24 with the lowest weight gain were designated as "DR." The remaining 12 intermediate weight gainers were switched back to chow and were designated as chow-fed controls. During the 10th wk on HE diet (week 12), 24-h urine NE collections were repeated (NE2) and caloric intake was assessed. At this time there was a clear distinction in body weights among the groups (Fig. 1). All DIO and DR rats were then switched to chow for 2 wk. Beginning week 14, 12 DIO and 12 DR rats were restricted to 50% of their baseline caloric intake on week 12. These rats were designated as "DIO-Restrict" and "DR-Restrict," respectively. All of the food was given to restricted rats at the beginning of the dark period throughout the entire restriction period. By week 17, body weights in each group had fallen to 86-87% of their baseline on week 14, and food intake was readjusted to maintain a stable body weight at ~90% of the 14-wk baseline (94 kcal/day for DIO-Restrict and 78 kcal/day for DR-Restrict). During this week, a third 24-h urine was collected for NE (NE3). During week 18, blood was taken by tail vein for plasma insulin and leptin levels. On week 21, 7 wk after beginning restricted intake, food intake was measured along with a 24-h urine NE collection (NE4). At the end of this week, six rats from each group (chow, DR, DR-Restrict, DIO, DIO-Restrict) were killed by decapitation, and white adipose pads were removed and weighed. The remaining DR-Restrict and DIO-Restrict rats were allowed ad libitum access to chow. All rats remained on ad libitum access to chow and, on the final week (week 28), a 24-h urine for NE (NE5) was collected and all remaining rats were decapitated for leptin and insulin levels and white adipose pad weights.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Body weight gain curves for rats begun on chow for a 2-wk baseline followed by high-energy (HE) diet for 2 wk. After 2 wk on HE diet, highest (DIO) and lowest (DR) weight gainers (n = 24/group) were continued on HE diet for 8 wk more. Middle weight gainers (n = 12) were switched back to chow. At week 12, all rats were switched to chow for 2 wk. At week 14, half the DIO and half the DR rats were restricted to 50% of their baseline intake on week 12 (DIO-Restrict and DR-Restrict, respectively; n = 12/group). After 3 wk of weight loss, their intake was adjusted to keep their body weights at ~90% of their baseline for an additional 5 wk. Half of each group (n = 6/group) was killed at this time, and the remainder of restricted rats was allowed ad libitum access to chow for an additional 6 wk. All remaining rats were killed at week 28 (n = 6/group). Urine for norepinephrine (NE1-5) was collected at the time points shown, and tail blood was drawn for leptin and insulin assays as designated. Points are means ± SE body weight (g).

Assays of insulin and leptin. All rats were killed between 0700 and 1000 for sample collection, having been allowed access to food and water overnight. Thus restricted rats were killed 14-17 h after they were given their food. Samples of trunk blood were collected into heparinized tubes, and the plasma was removed for assay. Both insulin and leptin were analyzed by radioimmunoassays (Linco) using antibodies to authentic rat insulin and leptin, respectively.

Statistics. Body weights were measured weekly over the entire period. Therefore, ANOVA for repeated measures was used initially to compare intergroup differences. When significant intergroup differences were found (P  0.05), post hoc Scheffé's multiple comparison tests were carried out at each time point where significant differences across a given weight-change phase occurred. One-way (experimental group) and two-way (group × phenotype) ANOVAs with post hoc Scheffé's tests were used for single-point measures of terminal body and retroperitoneal fat depot weights, plasma glucose, insulin, and leptin at the end of each phase.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Body weight, plasma leptin, and insulin levels and fat pad weights. By 3 wk on HE diet (week 5), body weights in DIO rats were significantly higher than chow-fed and DR rats and remained so for the remainder of the study (Fig. 1). On the other hand, DR rats weighed significantly less than chow-fed rats throughout the weight-gain period. Despite being more than 20% heavier, energy intake in DIO rats (100 ± 11 kcal/day) after 10 wk on HE diet (week 12) was not significantly different from that in DR (91 ± 10 kcal/day) or chow-fed rats (89 ± 10 kcal/day). During the 2-wk period after all rats were switched to chow ad libitum (weeks 13 and 14), body weights reached a plateau in both DIO and DR rats. When restricted on chow to 50% of their week 12 baseline caloric intake of HE diet, both DIO- and DR-Restrict rats rapidly dropped their body weights. By week 17, DIO-Restrict body weights had reached the level of chow-fed controls at 87% of their week 14 baseline weights. DR-Restrict weights decreased to 86% of their week 14 baseline by week 17. Caloric intake was readjusted in both restricted groups (on the basis of group mean body weights) to maintain their mean body weights at 90% of that baseline. This amounted to 94 kcal/day in DIO-Restrict and 78 kcal/day in DR-Restrict rats (P = 0.001). Thus DR-Restrict rats were eating significantly less than all other groups, although intake in DIO-Restrict rats was not significantly different from the other groups [Chow: 100 ± 5 kcal/day; DR: 102 ± 8 kcal/day; DIO: 110 ± 6 kcal/day; F(4,25) = 6.59; P = 0.001]. On week 18, 4 wk after caloric restriction began, leptin levels were significantly lower in both DR and DR-Restrict rats than all other groups and were 61 and 39% of chow-fed controls, respectively [Fig. 2; F(4,25) = 6.83; P = 0.001]. However, there was no significant difference in leptin levels between DR and DR-Restrict rats. Leptin levels in DIO rats were 50% higher than chow-fed controls, whereas DIO-Restrict rats had levels comparable to controls. There was a significant correlation between body weights and plasma leptin levels for both DR and DIO rats, but none was apparent for chow-fed, DR-Restrict, or DIO-Restrict rats (Table 1). Insulin levels in both DR and DR-Restrict rats were 40 and 36% lower than chow-fed controls and DIO-Restrict rats [Fig. 2; F(4,25) = 6.27; P = 0.001]. As with leptin levels, there was no significant difference in insulin levels between DR and DR-Restrict rats. However, there was a positive correlation between body weights and plasma insulin levels for only the DR rats. Overall, there was a positive correlation between body weights and leptin levels (r = 0.59, P = 0.001) and for body weight vs. insulin levels (r = 0.43; P = 0.05) across all five experimental groups.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Plasma leptin and insulin levels after 4 wk of restricted intake (Restrict) and terminally (Final) after all rats were on ad libitum intake of chow. DIO-R, DIO-Restrict; DR-R, DR-Restrict. Data for a given collection period (means ± SE, ng/ml are from n = 6-12/group). Dissimilar superscripts differ from each other by P  0.05; post hoc analysis after significant intergroup differences were found by ANOVA.

After 8 wk of restricted intake (week 22), six rats in each group were killed. At this time, DIO rats had significantly heavier epidydimal, retroperitoneal, perirenal, mesenteric, and total fat pad weights than all other groups [Fig. 3; total pad weights F(4,25) = 4.18; P = 0.011]. DR and DR-Restrict rats had lighter pad weights than both DIO and DIO-Restrict rats at this time, although they were not significantly lower than chow-fed controls. Neither DR nor DR-Restrict rats differed significantly from each other in fat pad weights. During week 23, DIO- and DR-Restrict rats were allowed ad libitum access to chow. Within 2 wk, they increased their body weights to the level of their respective, ad libitum-fed DIO and DR controls and their body weights remained comparable to these groups for the remaining 4 wk of the study (week 28). At this time, both groups of DIO rats were 5% heavier and DR rats were 9% lighter than chow-fed controls [F(2,23) = 8.32; P = 0.001]. Despite this relatively small difference in body weight, total fat pad weights were 30% heavier in DIO rats from both groups than those from both groups of DR rats [Fig. 3; F(2,23) = 7.29; P = 0.0004]. Despite the heavier body and fat pad weights of DIO rats, neither plasma leptin nor insulin levels differed among the five groups. There was, however, a positive correlation between final body weights and plasma leptin levels for chow-fed, DR, DIO, and DIO-Restrict rats and for insulin in DR, DR-Restrict, and DIO rats (Table 1). Again, there was a positive correlation between body weights and leptin (r = 0.45, P = 0.009) and insulin levels (r = 0.40; P = 0.05) across all experimental groups.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Fat depot weights after 8 wk of restricted intake such that body weight was kept at 90% of baseline (A) and 6 wk after rats were allowed ad libitum access to chow and regained lost weight (B). n = 5 or 6 rats/group. EPI, epidydimal; RP, retroperitoneal; PR, perirenal; MES, mesenteric; TOTAL, total weight of all 4 depots. Values (means ± SE, g, are from n = 6/group). Differing superscripts within a given adipose depot or total mass for A and total adipose mass for B differed from each other at P < 0.05 by post hoc t-test after significant differences were found by ANOVA. There were no significant differences among values in individual depots for the final weights.

Twenty-four-hour urine NE levels. Figure 4 and Table 2 show 24-h urine NE levels. These collections were made as an indirect measure of sympathetic nervous system activity during the various stages of weight change. Initial NE (NE1) levels were assessed while all rats were fed chow and were 86% higher in DIO-prone than DR and chow-fed controls [F(2,56) = 3.037; P = 0.05]. NE2 levels were collected during the 10th week on HE diet for DIO and DR rats. At this point, there were no significant differences among the groups. After 3 wk on 50% caloric restriction (NE3), NE levels fell to 85% of controls in DR-Restrict and to 49% of chow-fed controls in DIO-Restrict rats [F(4,24) = 3.69; P = 0.01]. However, after stabilizing body weights at 90% of their respective baselines for an additional 4 wk, NE levels (NE4) rose in DIO-Restrict rats to the level of chow-fed controls while DR-Restrict NE levels rose to 125% of controls. After a total of 10 wk back on chow, urine NE levels rose in DIO rats to 152% of chow-fed controls [F(4,24) = 4.71; P = 0.006]. During the final week (week 28), NE levels (NE5) in DIO-Restrict rats reached the level of DIO rats while levels in DR-Restrict rats fell to the level of DR and controls rats [F(4,24) = 3.71; P = 0.01].

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Mean 24-h urine NE levels as a percent of chow-fed control rats at various time points (UNE1-5) during diet switching. Absolute data are given in Table 2. Data for a given collection period with dissimilar superscripts differ from each other by P  0.05 by post hoc analysis after significant intergroup differences were found by ANOVA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

These studies were carried out to examine the way in which DIO rats defend their body weights against chronic food restriction on a low-fat diet as a model of diet therapy in humans. DR rats were examined for comparison to look for differences in the way in which they might defend their body weights under similar circumstances. This was based on the fact that DIO and DR rats are known to differ markedly in their weight gain patterns and metabolic responses to HE diet. Differences in weight gain cannot always be totally explained by differences in energy intake as shown in both previous (22) and current studies. Here, DIO rats had comparable intake to DR rats at the end of their period on HE diet, despite more than a 20% difference in body weight. DIO rats appear to increase their metabolic efficiency fairly soon after exposure to HE diet (22). Conversely, DR rats are selected for their low metabolic efficiency. This probably explains their reduced weight gain on HE diet compared with the chow-fed controls, which were chosen from the intermediate group of weight gainers on HE diet. These animals represent the low and high end of DIO and DR weight gainers, respectively (20, 21). Differences in body weight gain between DIO and DR rats are primarily due to differences in carcass adiposity, because they have been shown to have similar lean body mass (20, 25, 26). This was supported here by the differences in total adipose depot weights and plasma leptin and insulin levels between DIO and DR rats. The similarity in lean body mass may explain why intakes do not differ between DIO and DR rats.

Not only do weight gain patterns differ between DIO and DR rats but so do weight loss patterns during prolonged caloric restriction. Both DIO and DR rats lost weight at equivalent rates and magnitude when comparably restricted to 50% of baseline intake. However, only DIO-Restrict rats reduced their fat pad weights, plasma leptin, and insulin levels during caloric restriction. Although we did not assess total carcass adiposity, the fat pads weights assessed here do represent a large proportion of total adipose stores. Therefore, using this measure, it appears that chronic caloric restriction results primarily in loss of carcass fat in DIO rats, whereas DR rats primarily lose lean body mass. This is further supported by the fact DR-Restrict rats failed to reduce significantly either plasma leptin or insulin levels. Also, DR-Restrict but not DIO-Restrict rats lowered their maintenance energy intake requirements during caloric restriction. Thus, although unselected Sprague-Dawley rats show no reduction in lean body mass with chronic caloric restriction (2), there appears to be an important difference in the proportional loss of the different carcass composition compartments when restricted rats are separated into DIO and DR phenotypes.

The apparent sympathetic response to chronic caloric restriction also differed between DIO and DR rats. Here, as in prior studies (18, 20), chow-fed DIO-prone rats had elevated 24-h urine NE levels before the development of DIO on HE diet. This elevation "normalized" with the development of DIO on HE diet and then returned to the prior elevated state after 10 wk back on chow. The DIO-Restrict rats showed a marked and early reduction in sympathetic activity associated with caloric restriction (37). Although this decrease became attenuated with prolonged caloric restriction, it remained relatively reduced compared with ad libitum-fed DIO rats. A similar reduction in sympathetic activity has been described during chronic weight loss in postobese humans (1). But the magnitude and pattern of change in urine NE levels were completely different in DR rats during diet switching and caloric restriction. Urine NE levels in ad libitum-fed DR rats on both chow and HE diet were comparable to chow-fed controls. Importantly, DR rats had only a small decrement in NE levels during the early stages of caloric restriction. After this, NE levels actually increased significantly above chow-fed control levels during prolonged caloric restriction. Thus DIO rats had a much more vigorous inhibition of sympathetic activity during caloric restriction than DR rats. Because maintenance intake requirements in DR rats were still lower than in all other groups during restriction, this suggests a dissonance between sympathetic activity and energy expenditure in restricted DR rats. Interestingly, and perhaps paradoxically, such a dissonance has been described in Pima Indians who have a high propensity to become obese (30). Finally, there was a gradual and progressive increase in urine NE levels in control groups over the course of the experiment, which is unexplained but did not affect the relative relationship of levels among the groups.

These differences in the regulation of sympathetic activity raise the important issue of the mechanism underlying the reduction in sympathetic activity and energy expenditure during caloric restriction. If these are a consequence of decreased lean body mass, then DR-Restrict rats should have had the lowest levels of urine NE. If a decrease in adiposity and leptin levels was responsible, where leptin normally activates the sympathetic nervous system (8, 28), this might explain the lower 24-h urine NE levels in DIO-Restrict rats versus the relatively meager reduction of NE levels in DR-Restrict rats. Although this is consistent with the idea that sympathetic activity and fat mass are related, it seems likely that some other factor common to both regulates the complex interaction between the two. However, this observation does suggest that the greater the loss of carcass fat, the greater the degree of conservation of energy in the form of increased metabolic efficiency.

Finally, and most importantly, both groups of restricted rats returned to the baseline of their respective phenotype controls within 2 wk of their ad libitum access to chow. This return to the higher body weight of the unrestricted DIO rats is similar to the almost inevitable regain of lost weight in most reduced obese humans after dieting is stopped (12). Although the expected correlation between body weight and plasma leptin levels was present in DIO, DIO-Restrict, and DR rats at the end of the study, it was not seen in recovered DR-Restrict rats. This may have been an artifact of the relatively high variance in final plasma leptin levels. For example, there were no significant intergroup differences in final leptin levels, whereas there were significant intergroup differences in total adipose pad weights. This appeared to be due to the large variance in leptin levels, because the relative magnitude of mean leptin levels for each group was almost identical to those for total fat depot weights. This variance aside, it is more likely that the lack of correlation between leptin levels and body weight in DR-Restrict rats reflected the fact that DR animals preferentially spared body fat during restriction. We previously showed (24) that DR rats could be made hyperphagic and obese on a highly palatable diet but that they would not defend this higher body weight when switched to a low-fat, low-palatability chow diet. Those DR rats spontaneously reduced their food intake and body weight to control levels within 2 wk of the switch to chow. But their indexes of carcass adiposity did not fully return to control levels even after 6 wk back on chow. On the other hand, fat mass and body weight declined in parallel in DIO rats restricted to the same intake as the previously obese DR rats (24). Taken together with the present findings, it appears that DIO rats alter their body weight and fat mass in parallel when driven off their metabolically stable defended body weight. This suggests that the DIO rats primarily defend their lean body mass. On the other hand, DR rats appear to alter body weight (and probably lean body mass) first and then fat mass only secondarily. Thus DR rats appear to defend their fat mass more rigorously than their lean body mass. Although it is far from clear what underlies these differences in body weight and carcass composition during weight gain, loss, and regain, further exploration of this question may provide important clues to the effective treatment of obesity.

Perspectives

With the use of urine NE levels as an indirect measure of sympathetic activity, together with the other parameters examined, it is clear that DR rats respond very differently to chronic caloric restriction than do DIO rats. They had little or no reduction in fat mass or sympathetic activity when intake was restricted. This suggests that their weight loss resulted from a reduction in lean body mass. Despite these fairly profound differences in energy homeostasis, both DIO and DR rats lost and regained weight at comparable rates when restricted and refed. But they regained only to the level of their respective chow-fed phenotype controls. This return to these differing, phenotype-dependent body weights is also seen in overfed DR (24) and DIO rats (unpublished observation). But here again, only DIO rats do this by reducing carcass fat. Thus, in both over- and underfeeding situations, DIO rats appear to alter their fat mass, whereas DR rats appear to alter lean body mass. Although the reduction in plasma leptin levels seen in DIO-Restrict rats may be an important signal for spontaneous regain when food is readily available, it is clearly not the only signal for body weight regulation. DR rats regulate their body weight without apparent regard for plasma leptin levels or carcass adiposity. Here they regained lost body weight without ever showing significant reductions in plasma leptin levels during caloric restriction. But when DR rats were made obese and hyperphagic on a palatable diet and were then switched to chow, they spontaneously returned to, and perfectly maintained, the body weight of chow-fed DR controls long before their elevated leptin levels and carcass fat stores returned to that baseline level (24). Even obese Zucker rats, with their defect in leptin signaling (6), make the appropriate metabolic adjustments to caloric restriction (11). Such facts strongly suggest that there are other, as yet unidentified regulatory systems that are used by some animals.

Finally, even the intrinsic resistance of DR rats to obesity can be overcome by prolonged intake of an HE diet that does not cause hyperphagia (21). It may well be that there are similar obesity-resistant individuals within the population of obese humans. These may be the small proportion of obese individuals who have the most robust and successful long-term responses to a variety of obesity therapies. However, for the vast majority of individuals who are intrinsically obesity prone, it is likely that any successful therapy for their obesity will have to lower their defended body weight (27) and/or ameliorate the metabolic defenses provoked by weight reduction.