The consumption of nutrients by rodents is markedly influenced by the number of containers of each nutrient provided. Most rats given a choice from separate sources of protein, carbohydrate, and fat thrived if given one cup of each but half failed to thrive if given one cup of each and three extra cups of carbohydrate or fat. Rats given five bottles of sucrose solution and one bottle of water became fatter than rats given five bottles of water and one of sucrose. These studies in rats may point to a model for human obesity, in which the availability of food can override physiological controls of ingestion.
- diet selection
- diet-induced obesity
- food intake
pioneering studies in the 1930s showed that rodents can thrive when required to select a diet from separate ingredients (17, 18). This “nutritional wisdom” is generally considered to be the physiological basis of food selection, although it can be tempered by other factors such as taste hedonics, past experience, and social communication (3, 4, 19, 25, 26). Our current understanding of the mechanisms underlying food selection is based almost entirely on laboratory and farm studies in which animals were allowed to choose between separate containers of nutrients. In every case, the animals received one container of each nutrient. Here, we report that manipulation of the number of nutrient containers available can have profound effects on nutrient choice, even to the extent that they override the physiological controls of nutritional wisdom.
METHODS AND RESULTS
Experiment 1: influence of macronutrient choice on macronutrient selection.
In this experiment, rats were given a choice between separate sources of solid carbohydrate (CHO), fat, and protein, but in addition to this standard choice of macronutrients, some rats received three “extra” cups of protein, CHO, or fat. Thirty male Sprague-Dawley rats [Charles River, Crl:CD(SD)IGS BR] were maintained under standard conditions (23°C, 12:12-h light-dark cycle) with powdered Purina Rodent chow (no. 5001) to eat and deionized water to drink until the rats were ∼10 wk old, when the experiment began. Each rat was housed in a 20 × 18 × 25-cm stainless steel cage designed for conducting diet selection experiments (11). Food was available from stainless steel cups (7 × 7 × 3 cm) with spill-proof lids that were placed in each corner and at the center of the longest sides of the cage.
Sources of protein (casein), CHO (a mixture of cornstarch, dextrin, and sucrose), and fat (a mixture of Crisco vegetable shortening and safflower oil) each contained micronutrients and vitamins as described elsewhere (16). Body weights and intakes of each source were measured (to the nearest 0.1 g, corrected for spillage) every day. Fresh fat was provided every 2 days; water and the other nutrients were replaced as needed.
This experiment was terminated after 8 days because four of the seven rats given extra cups of CHO and three of the seven rats given extra cups of fat ate so little protein they failed to thrive. That is, their average daily intakes of protein were <15 kcal/day, and they lost body weight. In contrast, all eight of the rats given extra protein and seven of the eight rats given just one source of each macronutrient thrived well (difference in frequency, χ2 = 8.12, df = 3, P < 0.01). Providing rats with extra cups of CHO or fat thus led to life-threatening protein malnutrition, even though protein was freely available in their cages.
The 22 rats that had acquired the habit of eating protein were reassigned to one of three groups, matched for protein intake and body weight. These were given the standard three cups of macronutrients (n = 8) or the three macronutrients plus three extra cups of CHO (n = 7) or fat (n = 7). Body weights and intakes of these animals were measured every 2 days and analyzed by two-way mixed-design ANOVAs, with factors of group and time. Under these conditions, the control group ate more protein, the group with extra CHO ate more CHO, and the group with extra fat ate more fat than did the other groups (P values < 0.05; Fig. 1). Over the 20-day test, these differences remained more-or-less constant. There were no differences in body weight gain or final body weight, but total energy intake of the group with extra CHO was significantly higher than energy intakes of the other two groups [total energy intakes (kcal/day): control = 104 ± 1, extra CHO = 118 ± 5, extra fat = 107 ± 3; F(2,20) = 4.14, P = 0.03]. Thus access to multiple sources of a single nutrient biased intake toward that nutrient and, at least for CHO, could override energetic constraints on intake.
Experiment 2: influence of sucrose solution choice on energy intake and obesity.
Rats given sucrose solution to drink in addition to their maintenance diet often become obese (for review, see Ref. 15). In this experiment, we attempted to influence the body weight of female Sprague-Dawley rats (weighing 238 ± 2 g at the start of the experiment) by manipulating their access to sucrose solution. The rats received to drink for 35 days 1) just one bottle of water,2) five bottles of water and one bottle of 32% sucrose solution, or 3) one bottle of water and five bottles of 32% sucrose solution. Rats in the six-bottle groups (n = 12 each) were housed in stainless steel guinea pig cages (41 × 56 × 23 cm) with pine shavings on the floor. Powdered Purina rodent chow no. 5001 was available from a glass jar, and fluids were available from 50-ml inverted centrifuge tubes with rubber stoppers and stainless steel drinking spouts. When six tubes were available, the tubes were lined across the front of the cage at ∼3-cm intervals and held in place by steel springs. The fourth tube from the left contained water, and the rest contained sucrose (or vice versa). Rats in the control group (n = 12) were housed in standard cages (18 × 24 × 18 cm) with the same food and a single tube of water. Food and fluid intakes were measured daily, and body weights were measured twice a week. Energy intake was calculated from the sum of sucrose intake (1.28 kcal/ml) and food intake (3.4 kcal/g). To provide more stable results and simplify data analysis, average intakes over each 4-day period were used in presentation and analyses.
At the end of the experiment (day 36), the rats were fasted overnight to remove gastrointestinal contents and then killed by anesthetic overdose. Their shaved carcasses were homogenized, and extracts were analyzed for fat content (13). Differences between groups were determined using two-way ANOVAs (group × day) for the intake measurements and one-way ANOVAs for body weight and fat content.
The rats with five sucrose bottles drank significantly more sucrose and consumed more energy than did those with one sucrose bottle, both throughout the test [sucrose, F(1,22) = 40.7,P < 0.0001; energy, F(2,33) = 32.8,P < 0.0001] and for every 4-day period (Pvalues < 0.01, except energy on days 29–32), with the differences being greatest at the beginning of the test period [group × time interactions, F(8,176) = 6.88,P < 0.0001 and F(18,297) = 4.42,P < 0.0001, respectively; see Fig.2]. Significant differences between the groups in body weight gain emerged on day 8[F(2,33) = 3.89, P = 0.031] and remained for the rest of the experiment.
The rats with five bottles of sucrose drank more than did the rats with only one bottle. Although their food intake decreased in partial compensation for the additional energy ingested as sugar, overall they consumed more energy than did the other groups. They gained significantly more weight than did controls after 8 days and significantly more weight than did rats with only one bottle of sucrose after 16 days. The group with only one bottle of sucrose and five of water consumed more energy throughout the test period and gained significantly more weight than did controls after 33 days (Fig. 2). At the end of the test period (day 36), the three groups differed significantly from each other in fat content (controls = 41 ± 3 g, 1 bottle of sucrose = 57 ± 2 g, and 5 bottles of sucrose = 76 ± 4 g, equivalent to 15 ± 1, 19 ± 2, and 24 ± 1% of carcass weight, respectively).
Analysis of intakes from individual bottles of sucrose proved to be uninformative. Rats tended to have consistent patterns over the 35-day test. Some drank more or less evenly from all five sucrose bottles; others drank almost exclusively from one or two bottles, but there was no consistency about which bottles were preferred by all rats (data not shown).
The results show that the more sources of a nutrient a rat has, the more it chooses to ingest. The effect of nutrient availability was so powerful it could override physiological controls of intake. Almost all inferences about the mechanisms underlying nutrient selection are based on tests in which animals choose among single sources of each nutrient. These inferences may need reevaluating because they do not generalize to situations where multiple sources of each nutrient are available.
Demonstrations that rats can select sufficient nutrients to grow normally (17, 18) were seminal support for the notion of nutritional wisdom, and they remain at the core of the study of food choice by animals. However, there are also several demonstrations that rats can have difficulty selecting an adequate diet, particularly if the available protein is purified (4-7, 22, 24). Selection can be profoundly influenced by factors typically considered hedonic or cognitive rather than physiological in nature, such as the textural composition of the nutrients offered (9), the previous experience of the rats with macronutrients (16), and the social transmission of information about the diets (1,3). It has been suggested that animals may not be able to “discover” a needed diet ingredient if the choice of available foods is large (5). Here, we provide experimental support for this by showing that a seemingly trivial factor from a physiological perspective, i.e., the number of sources of CHO or fat available, was sufficient to reduce the selection of protein so much that many animals failed to thrive.
When rats that adapted successfully to the macronutrient selection paradigm were given three extra cups of fat or CHO, they consumed significantly more of the extra macronutrients than did rats given only one cup of each macronutrient. Providing extra cups of a nutrient induced large shifts in the proportion of each nutrient ingested, but this was insufficient to increase body weight significantly above those of controls. However, the control group chose a diet that yielded ∼28% of energy from fat and ∼33% from CHO, and such a high-fat, high-CHO diet is already a potent inducer of weight gain (see for example Ref. 14). Thus it is not surprising that over the relatively short (20 day) test, we did not see an increase in body weight of the groups with extra CHO or fat relative to this group.
It is well known that obesity can be induced by providing rats fed a nutritionally complete diet with a sugar solution to drink, although this method is not always successful (reviewed in Ref.15). Here, we found that obesity was greater and developed more rapidly when several sources of sucrose solution rather than just one source were available. Sucrose intakes were particularly high during the first few days of access but remained higher than those of the group with only one sucrose bottle throughout the 5-wk test period. It is noteworthy that the group with one sucrose bottle gained significantly more weight than did controls without sucrose to drink, but only after 33 days of sucrose access, which is a modest rate of weight gain compared with similar studies (see Ref. 15). It would be interesting to know whether providing these rats with five water bottles diluted exposure to sucrose and thus retarded the development of obesity relative to the more usual test procedure, where rats receive only one sucrose bottle and one water bottle.
We suspect that availability-based overconsumption is related to the “variety effect,” in which rats given foods of different flavors or textures overconsume relative to those given food of only one flavor or texture (e.g., Refs. 20, 21). Variety has been suggested as a cause of the obesity produced by feeding rats a diet of supermarket foods (8, 23). However, unlike the results seen with sucrose here, there is little evidence that variety has more than a transient effect on intake, and by itself it does not lead to obesity (12). Indeed, the present results argue that simply providing multiple sources of food stimulates intake and thus may contribute to, and in some circumstances even account for, the variety effect. At the least, the present results illustrate that interpretation of earlier nutrient choice and supermarket diet studies is difficult because they lack appropriate controls for the number of nutrient sources provided (see Ref. 10 for additional problems).
The mechanisms underlying availability-based overconsumption are unknown. Overconsumption due to food variety is usually attributed to hedonic effects or the lack of sensory-specific satiety (e.g., Refs.20, 21), but these explanations are not informative from a physiological perspective. With respect to availability-based consumption, we think it more likely that with many sources of nutrients available, the rat simply takes advantage of the additional opportunities to ingest: the rat eats more because it encounters food more frequently. One implication of this is that the rat eats simply because the food is there and not in response to nutritional needs. This is an unorthodox explanation of consumption by the rat, but it is generally accepted to be the case for humans. Providing multiple sources of sucrose may also exacerbate obesity because, all things being equal, the rat has to expend less energy traveling to a source of calories than if only one bottle of sucrose is available.
The limits of availability-based overconsumption remain to be tested. In addition to the studies with solid nutrients and liquid sucrose reported here, we have found similar effects with mice given representative sweet, sour, salty, and bitter compounds, and the irritant, alcohol (unpublished results). The phenomenon therefore appears to encompass a wide spectrum of, if not all, food items. It would be interesting to know whether the relative positions or distance apart of nutrient sources is critical, or whether rats select more of a food presented in a big rather than small container. There are also a number of potentially interesting parametric investigations involving manipulations of the number and difficulty of obtaining nutrients that might tie the phenomenon into economic models of food foraging (e.g., Ref. 2). A critical study will be to determine whether rats given more than five bottles of sucrose develop even more pronounced obesity than that seen here. Such experiments are more challenging to conduct than single-nutrient source tests, but they are probably more representative of the rats' natural foraging environment.
The finding that laboratory animals choose to eat what is abundant has obvious relevance for husbandry and for animals in the wild, including humans confronted with many products in the supermarket. Recent discoveries of new genes, hormones, and neurotransmitters involved in the control of ingestion and body weight have led to a strong emphasis on the physiological causes of obesity. Given the simultaneous rise in the incidence of human obesity with the spectacular increase in the availability of nutritionally questionable foods, it appears that changes in availability rather than physiological mechanisms are more likely to be responsible for the poor diet choice, hyperphagia, and obesity displayed by many humans. The results observed here point to an animal model of this “obesity by choice.”
I thank D. Pilchak, L. Curtis, S. Rabusa, J. Williams, A. Acquaviva, A. McDaniel, and A. Hargett for technical support. Drs. G. Beauchamp, P. Breslin, M. Friedman, and D. Reed gave useful comments.
This research was supported by National Institutes of Health Grant AA-12715.
Address for reprint requests and other correspondence: M. Tordoff, Monell Chemical Senses Center, 3500 Market St., Philadelphia, PA 19014-3308 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society