Dietary Self-selection Research Articles

A defined artificial diet for the larvae of Manduca sexta

The tobacco hornworm, Manduca sexta (Johannson) (Lepidoptera: Sphingidae), has become an important and widely used laboratory animal. It was first cultured on the foliage of its host plants (Waldbauer et al., 1964; Hoffman et al., 1966), but is now typically cultured on undefined artificial diets that contain such complex substances as wheat germ and yeast (Yamamoto, 1969; Bell & Joachim, 1976; Baumhover, 1985). A diet that includes such ingredients, each containing the gamut of nutrients, is useful for maintaining cultures but is not appropriate for most nutritional studies, including studies of dietary self-selection. We here describe a defined diet (containing no ingredients that are chemically less well characterized than linseed oil) that is suitable for studies of self-selection and other aspects of nutrition. To the best of our knowledge, no defined diet for M. sexta has been previously published. The defined diet (Table 1), a modification of Vanderzandt's (1968) diet, is prepared as follows: The group A ingredients, water and agar, are poured into a flask that is capped with aluminum foil, autoclaved for 20 min, cooled for 10 min, poured into a blender with the group B ingredients, and then thoroughly mixed. When the mixture cools below 60 ~ the group C ingredients are added, the diet is blended for 5 min and then poured into plastic molds. After the diet has cooled to room temperature, each mold is wrapped with aluminum foil and stored at 4 ~ until used. Table 1. Composition of defined diet for rearing Manduca sexta larvae

Social influences on the selection of a protein-sufficient diet by Norway rats (Rattus norvegicus).

Investigated effects of interactions between naive and knowledgeable rats (Rattus norvegicus) on selection of a nutritionally adequate diet by the naive. We found that during a 7-day test, isolated rats choosing among 4 foods, 3 of which were protein-deficient and 1 of which was protein-rich, failed to learn to prefer the protein-rich diet and lost weight. Conversely, those rats that interacted with conspecifics trained to eat the protein-rich diet developed a strong preference for that diet and thrived. The authors also found that Ss were more strongly influenced in their diet selection by the flavor of the foods eaten by conspecifics than by the locations where conspecifics fed. The results suggest that social influence may be important in development of adaptive patterns of diet choice by rats (or other dietary generalists) that need to find nutritionally adequate diets in demanding environments. Students of dietary self-selection have described two complementary processes that may lead animals foraging in the world outside the laboratory to select nutritionally adequate foods from among the myriad potentially ingestible, beneficial, toxic, and useless substances found in natural habitat. First, animals have the ability to detect the presence of some nutrients in complex foods and can use tastes associated with those nutrients to select valuable foods to eat (Richter, 1943; Rozin, 1976). Second, animals can associate the postingestional consequences of a food with its taste and can, therefore, learn to select foods that provide valuable nutrients (Booth, 1985; Harris, Clay, Hargreaves, & Ward, 1933; Rozin, 1976). In the literature, these two abilities—the ability to detect directly some nutrients in foods and the ability to learn about the nutritional value of foods from the consequences of their

Effect of phenylpropanolamine on diet selection in rats

Four doses of phenylpropanolamine (PPA; i.e., dl-norephedrine: 2.5, 5, 10, and 20 mg/kg) and saline were injected intraperitoneally in female rats maintained on a dietary self-selection paradigm. Intake of all three macronutrients (carbohydrate, fat, and protein) was equally affected by high doses of PPA. Lower doses decreased fat and protein more than carbohydrate.

Effects of early under-nutrition on dietary self-selection and relative carbohydrate- and protein-conditioned preferences

Rats were either under-nourished from birth to 45 days and thereafter well fed, or were well-nourished throughout. Such previously under-nourished rats (PU) had a body weight deficit in adulthood of about 15% when compared with the well-fed group (C). On four training days, rats were presented with a distinctively odourized solution of carbohydrate or protein, following a 5-h period of food deprivation. On the test day, under the same food deprivation conditions as occurred during training, identical solutions distinguished only by their odour were presented simultaneously to the rats for a period of 20 min. PU rats showed a conditioned preference for the carbohydrate-paired odour. Control animals showed a small preference for the protein-paired odour. The same rats were later allowed to self-select from three macronutrients for a period of 8 days. No differences in dietary self-selection patterns were found between PU and C rats.

Sucrose-induced hyperphagia and obesity in rats fed a macronutrient self-selection diet

Adult female rats were allowed to self-select their diet from separate sources of fat, protein, and carbohydrate (starch). Other rats were fed a composite diet that matched the nutrient composition chosen by the self-selecting rats (50% fat, 28% protein, 22% carbohydrate) or a low-fat, high-carbohydrate chow diet. Half of the rats in each diet condition were given access to a 32% sucrose solution for 30 days. Sucrose availability increased total caloric intake (≈20%) and body weight gain in all three groups compared to control groups not fed the sucrose solution. The selection animals compensated for their sucrose intake by reducing their fat intake, and to a lesser degree, their starch intake; protein intake was the least affected by sucrose availability. The selection rats consumed less sucrose than the chow-fed rats and displayed a smaller increase in weight, relative to controls, than the chow-fed rats. These differences were attributed to the high-fat intake of the selection animals since similar results were obtained with the rats fed the composite diet. In particular, both the selection and composite diets produced mild obesity in the absence of sucrose. The results demonstrate that sucrose-induced overeating and overweight is not an artifact of restraining the diet choices of rats to a pure sugar and a nutritionally complete diet.

Dietary self-selection of golden hamsters in response to acute food deprivation and chronic food restriction.

Adult male golden hamsters were maintained on either Purine Rat Chow (Chow diet) or a self-selection diet consisting of high-protein chow, pure fat, and pure carbohydrate (Choice diet). In Experiment 1, animals were deprived of food for single periods of up to 48 hr. Animals on the Chow diet did not increase intake at any time after deprivation; animals on the Choice diet selectively increased their consumption of fat-derived calories and increased their total caloric intake during the first 6 hr of refeeding, but not thereafter. The nature of the diet did not influence the rate at which animals regained weight following deprivation. In Experiment 2, hamsters were placed on food-restriction schedules (access to food either for 1 hr/day only or on alternate days only) until they lost 20% of starting body weight. Chow-fed animals demonstrated little or no change in food intake either during or after food restriction. Hamsters on the Choice diet consumed more calories and lost weight more slowly than did chow-fed animals during 1-hr/day feeding; intake of fat-derived calories was elevated during restriction. Choice hamsters increased total caloric intake only towards the end of the alternate-days restriction schedule. Choice hamsters were hyperphagic following both types of food-restriction schedules, but no increased preference for fat-derived calories was observed. Factors influencing food consumption of hamsters in response to deprivation and restriction are discussed.

Dietary self-selection of golden hamsters in response to acute food deprivation and chronic food restriction.

Dietary self-selection of golden hamsters in response to acute food deprivation and chronic food restriction.

Dietary self-selection in diabetic rats: An overview

The literature concerning dietary self-selection patterns of diabetic rats is reviewed and compared with new data. There is agreement among the various investigators as to the dietary choices observed following induction of diabetes, regardless of the diabetogenic treatment used. That is, moderately diabetic rats select a high fat, low carbohydrate diet, whereas more severely diabetic animals consume high protein, low carbohydrate diets with little change in fat consumption relative to nondiabetic controls. Even very mildly diabetic rats reduce carbohydrate intake. Evaluation of metabolic status of diabetics suggests that with severe diabetes, the beneficial reduction of plasma glucose seen with consumption of a high fat diet may be offset by extreme elevations in ketone and triglycéride levels. Moreover, the hypothesis that diabetic rats are insensitive to carbohydrate calories seems weakened by evidence of reduced food intake following carbohydrate consumption either in solutions or as a gastric load. These findings are discussed in terms of “dietary wisdom” as first proposed by Richter.

Peripheral injections of bombesin and cholecystokinin affect dietary self-selection in rats

Patterns of dietary self-selection were examined in adult female rats following peripheral injections of either bombesin (BBS) (6, 10, 14 and 16 μg) or cholecystokinin octapeptide (CCK-8) (0.75, 1.5, 2.25, and 3.0 μg). Animals were food deprived for 18 hours and then offered three isocaloric diets (protein, carbohydrate, and fat) following injections of peptides. Each subject received each of 4 doses of both peptides in a within-subjects design. All doses of BBS decreased total food intake and fat intake 30 minutes following injections. Also at this time period the two highest doses suppressed carbohydrate intake, while protein was unaffected. Cumulative intake at one hour revealed that total intake remained suppressed. The two highest doses continued to suppress carbohydrate intake, while only the 14 μg dose continued to suppress fat intake. Additionally protein was now significantly suppressed by all doses. The three highest doses of CCK-8 produced a decrease in total food intake and fat intake 30 minutes after injections. By one hour, only total intake remained suppressed but only with administration of the highest dose. Results are interpreted as providing support for the notion that BBS and CCK are physiological satiety signals and that they maintain unique functions in regulating food intake.

The long-term effect of dietary administration of refined sugars and sugar alcohols on plasma biochemistry, urine biochemistry and tissue histology in mice given a limited degree of dietary self-selection

A prototype animal feeding model is described in which mice were meal-fed a balanced diet but were given free access to water (controls) or 20% (w/v) solutions of glucose, sucrose, fructose, xylitol or sorbitol. Under these condition it was found that the provision of an alternative energy source, in the form of a refined carbohydrate, produced marked effects on total energy intake, mouse cube (i.e. balanced energy) intake and body weight. There were also changes in the metabolic states of the animals as assessed by serum levels of glucose, urea and cholesterol, plasma levels of lactate and d-3-hydroxybutyrate, and urinary excretion of urea and oxalate. Histological examinations of tissue indicated that the sucrose-fed mice had a tendency to suffer from acute congestion of the lungs and liver steatosis. Given a limited degree of dietary self-selection it appears that mice are more likely to be at risk of excessive food consumption and obesity when given glucose- or sucrose-containing diets than they are when fructose-, xylitol- or sorbitol-containing diets are given.

The role of taste in dietary self-selection in rats.

In order to study the role of taste in dietary self-selection, rats were subjected to two degrees of gustatory deafferentation. The chorda tympani was sectioned alone or in combination with the glossopharyngeal nerve and the pharyngeal branch of the vagus. The animals chose from two diets, only one of which contained protein. After surgery, deficits were observed in body weight, food and water intake, and diet selection, proportional to the extent of deafferentation. Some 76% of the animals increased protein and decreased carbohydrate intake, but all rats continued to select a nutritionally balanced diet (Experiment 1). When subjected to a nutritional challenge of intragastric protein or carbohydrate supplementation, the rats compensated for calories and nutrients by selectively adjusting oral intake (Experiment 2). In saccharin preference tests, the preference as well as the total consumption of the test solutions was reduced (Experiment 3). The findings are contrasted to the effects of oral somatosensory deafferentation and are interpreted as a change in preference within the limits of metabolic requirements. The different roles/mechanisms of the two sensory systems in regard to dietary self-selection are discussed.

Dietary self-selection and intake of protein and energy is altered by the form of the diets

The technique of dietary self-selection has been widely used to investigate proposed mechanisms of protein and carbohydrate regulation and to substantiate theories relating nutrients, brain chemistry and behaviour. These experiments have produced conflicting results which might be accounted for by differences in experimental variables or in the different techniques used to measure self-selection. The possibility that the sensory qualities of diets might alter dietary self-selection was examined in three groups of rats self-selecting protein and carbohydrate from pairs of isocaloric diets in which the physical form (finely-milled casein/corn starch powder, granular casein/corn starch powder and gelled finely-milled casein/corn starch), but not the nutrient composition of each pair was varied. In addition, since the serotonergic drug fenfluramine has been frequently used to influence dietary selection, the effects of this drug on the selection of protein and carbohydrate from the three diets were also examined. The selection of particular diets and consequently protein and carbohydrate intake was markedly altered by the form of the diet. Furthermore, the effects of fenfluramine on diet selection were also modified significantly by the form of the diet. These results suggest that intake of certain nutrients is not strictly regulated at a particular fixed value and that drug-induced changes of self-selection behaviour must be interpreted cautiously.

Dietary self-selection in chronically insulin-treated golden hamsters.

Dietary self-selection in chronically insulin-treated golden hamsters.

The role of taste in dietary self-selection in rats.

The role of taste in dietary self-selection in rats.

Dietary self-selection in chronically insulin-treated golden hamsters.

Adult male golden hamsters were maintained on either Purina Rat Chow (chow group) or a self-selection diet consisting of high-protein chow, pure carbohydrate, and pure fat (choice group). Animals were injected for 12 consecutive days with either long-acting insulin (20 U/kg for 4 days, 60 U/kg for 4 days, and 100 U/kg for 4 days) or physiological saline. Insulin-injected hamsters under both dietary conditions increased their total caloric consumption by up to 33% and gained significantly more weight than saline-injected controls. Choice hamsters increased their fat intake in response to the 60 and 100 U/kg doses of insulin, but carbohydrate and protein consumption increased only in response to the 100 U/kg dose. Choice hamsters derived approximately 65% of the excess calories ingested during insulin administration from fat, but only 20% from carbohydrate and 15% from protein. Results are related to those previously observed in other species.

Serotonin manipulations and the structure of feeding behaviour

Experimental manipulations of 5-HT by central or peripheral procedures produce clear effects on food intake. Interpretation of these effects can be improved by monitoring changes in the behavioural structure of feeding and by designing experiments which adjust the contextual and temporal dimensions of feeding. Experimental techniques include micro-analysis of behaviour, macro-analysis of meal patterns, dietary self-selection of macro-nutrients, analysis of appetite and satiety sequences, appetitive and consummatory aspects of instrumental performance and dietary-induced hyperphagia. Using these procedures in conjunction with various pharmacological treatments including the administration of tryptophan, 5-hydroxy-tryptophan, 5-HT, re-uptake blockers, synaptic releasers, synthesis blockers, receptor blockers and others it has been demonstrated that 5-HT manipulations give rise to a specific constellation of changes in feeding parameters. The most prominent changes following peripheral or central (paraventricular nucleus) injections are a reduction in meal size and a curtailment of eating rate. These changes can be distinguished from those induced by other pharmacological treatments and are consistent with a modulation of the process of satiation and the state of satiety. Changes in feeding induced by 5-HT manipulations are not secondary to non-specific behavioural alterations.

Determinants of dietary self-selection in experimental animals

During the past 10 yr, renewed interest has been expressed in the mechanisms controlling the intake of specific dietary components. In particular, this interest has focused on research examining the role of the central nervous system in regulating intake of the three macronutrients protein, fat, and carbohydrate. Several hypotheses relating specific neurotransmitter systems to nutrient selection have developed from this research. However, conflicting data make the acceptance of any one hypothesis about diet selection untenable. As an attempt to reconcile some of the conflicting data, the present paper describes a variety of factors which can influence patterns of nutrient choice. The effects of diet composition, the experimental environment, the animals' background, and nutritional history on diet selection are examined.

Dietary self-selection in insulin-injected hamsters

Adult male golden hamsters were given access to a variety of nutrient sources and were observed following the administration of regular insulin. It was hypothesized that if insulin produced hyperphagia in hamsters by the activation of a glucoprivic feeding mechanism, a selective increase in carbohydrate consumption would be observed. All animals received subcutaneous injections of 10, 30, 50 and 100 units/kg of insulin as well as a control injection of saline. Food consumption was recorded at +3, +6 and +24 hours after injections. In Experiment 1 hamsters having continuous access to Purina lab chow, fat (Crisco) and sucrose (sugar cubes) increased their total caloric consumption in response to insulin, but did not do so by selectively increasing their carbohydrate intake. In Experiment 2 hamsters maintained on Purina chow and sugar cubes consistently increased their carbohydrate intake as well as their total caloric consumption in response to insulin, but again the increase in carbohydrate intake was not selective; increased consumption of both sugar cubes and Purina chow occurred, and neither the proportion of total calories derived from carbohydrate nor the proportion of total calories derived from sugar cubes was affected by insulin administration. The results support the conclusion that insulin-induced hyperphagia in hamsters results from the activation of a non-glucoprivic feeding mechanism.

Sham feeding, flavor associations and diet self-selection as indicators of feeding satiety or aversive effects of peptide hormones

We have attempted to develop a constellation of behaviors which show differential effects following the administration of putative satiety hormones (CCK-8, BBS, insulin) as opposed to effects seen following a toxin, such as LiCl. In the initial behavior assessed, sham feeding of differently paired, flavored milks (flavor paired with insulin, BBS or saline) was carried out. Male adult Sprague-Dawley rats which sham fed milk flavors paired with 16 μg/kg BBS showed a significant aversion of that flavor in a two-bottle taste test (compared to saline-paired flavors, p<0.001) but a significant preference for flavored milk paired with 0.4 and 0.75 U insulin/rat. Lower dosages of BBS (4 and 8 μg/kg) and insulin (0.1 U/rat) showed no significant aversion or preference when compared to saline. The second behavioral paradigm evaluated the effects of the hormones CCK-8 and BBS and the toxin, LiCl, upon self-selection of pure macronutrients. While CCK-8 reduced intake of calories by significantly lowering ingestion/selection of fats (55%, p<0.01 compared to saline, control injections) and carbohydrates (50%, p<0.01), LiCl and BBS reduced calories by decreasing selection of primarily proteins (LiCl—49%, p<0.03; BBS—63% at 4 μg/kg and 80% at 8 μg/kg, both p<0.025). In both paradigms then, BBS at doses sufficient to significantly reduce sham intake or suppress caloric ingestion in a self-selection paradigm produced behavioral effects most similar to those observed following the injection of a toxin, LiCl, rather than effects seen following other various putative satiety signals.

Dietary Self-Selection Patterns of Rats with Mild Diabetes

Rats were allowed a free selection of a diet from among separate sources of protein, fat and carbohydrate or were fed a composite diet formulated to approximate the nutrient composition of a commonly used nonpurified diet. Immediately after streptozotocin injections, diabetic rats displayed polyuria, polydipsia and glycosuria as well as elevated fasting plasma glucose levels and glucose intolerance indicative of mild diabetes. Diabetic rats allowed a free choice tended to consume more protein and consumed significantly less carbohydrate than nondiabetics. This pattern of nutrient choice was associated with a reduction of diabetic signs including reduced polyuria, polydipsia and glycosuria. Diabetic rats permitted to choose their diets were not hyperphagic and maintained a slow but steady rate of body weight gain, accompanied by a sparing of body fat stores. In contrast, diabetic rats consuming the composite diet experienced no improvement in diabetic status; these rats displayed a deterioration of fasting plasma glucose, severe polydipsia, polyuria and glycosuria as well as hyperphagia and wasting of fat stores. These data demonstrate that when mildly diabetic rats are given the opportunity to select their own diets, they choose a diet that leads to improvement of their diabetic status.