Relationship of dietary fat content to food preferences in young rats

Physiology& Behavior,Vol. 48, pp. 581-586. ©PergamonPress plc, 1990. Printed in the U.S.A.

0031-9384/90 $3.00 + .00

Relationship of Dietary Fat Content to Food Preferences in Young Rats Z O E S. W A R W I C K 1

Department of Psychology, Duke University S U S A N S. S C H I F F M A N

Departments of Psychology and Psychiatry, Duke University AND J O H N J. B. A N D E R S O N

Department of Nutrition, The University of North Carolina-Chapel Hill R e c e i v e d 3 A u g u s t 1990

WARWICK, Z. S., S. S. SCHIFFMAN AND J. J. B. ANDERSON. Relationshipof dietaryfat contenttofood preferences in young rats. PHYSIOL BEHAV 48(5) 581-586, 1990.--Weanling rats were fed either a high-fat (30% of calories) or a low-fat (10% of calories) diet for four weeks, after which fat preference was assessed using a choice paradigm. Fat preference was measured during 2-hour intake tests in which three peanut butter/peanut oil mixtures containing 0.50, 0.61, and 0.71 grams fat/gram were offered to each animal. Rats fed the high-fat (HF) diet preferred the highest-fat mixture and consumed more total fat during intake tests than animals fed the low-fat (LF) diet. Intake of NaCl and sucrose solutions was measured during separate intake tests. LF-fed rats drank more NaC1 solution than HF-fed rats. Following these tests a subgroup of the LF-fed animals was fed the HF diet, and a subgroup of the HF-fed group was fed the LF diet for a further four weeks. Upon repetition of the intake tests, rats that had been fed the HF diet during the intitial four weeks still preferred the highest-fat mixture. Weanling rats

High-fat diet

Dietary fat

Food preferences

SENSORY and ingestive experiences are significant determinants of food selection and intake. A relationship to dietary history has been found for human preferences for salt (1), sucrose (2) and flavors (3, 4, 13). The feeding patterns of rats are similarly linked to prior feeding experiences (5, 6, 8, 10, 11). The purpose of the present study was to determine whether the level of fat in the diet of juvenile rats was related to subsequent preference for fat. Rats were maintained on either a low-fat (10% of calories) or high-fat diet (30% of calories) for four weeks. Following measurements of fat preference a subgroup of the animals was placed on the alternate diet, in order to assess the robustness of food preferences induced by initial dietary experience. Intake of sucrose and sodium chloride solutions was also mea-

sured to determine whether intake of nonfat substances was related to the level of fat in the diet. METHOD

Diets Nutritionally complete diets (see footnote 2) containing either 10% of calories (low-fat) or 30% of calories (high-fat) from safflower oil were prepared weekly by thoroughly mixing all ingredients with an electric mixer (Table 1). The diets were stored at 5°C in sealed plastic tubs. Casein, dextrin, alphacel and mineral mix were obtained from Teldad, Madison, WI. The vitamin mix was obtained from ICN Biomedicals, Cleveland, OH. Safflower oil (Hollywood Brand) and cornstarch (Argo) were purchased lo-

~Requests for reprints should be addressed to Zoe Warwick, Department of Psychology, Duke University, Durham, NC 27706. 2Diets were modified versions of formulations described by R. O. Deems, P. L. Skypala, A. Martinez-Hernandez, L. S. Friedman, M. I. Friedman (personal communication).

581

582

Nutrient Vitamin-free casein Safflower oil Cornstarch Dextrin Sucrose Alphacel Vitamin mix (ICN) Mineral mix (AIN-76) Percentage of Calories Fat Carbohydrate Protein

WARWICK, SCHIFFMAN AND ANDERSON

TABLE 1

TABLE 2

DIET COMPOSITION

FAT-RICH TEST MIXTURES

Low-Fat Diet 118.5 g 21.0 187.5 89.0 31.0 23.5 10.0 19.5 500.0 10 65 25

High-Fat Diet 118.5 74.5 167.5 77.0 24.5 3.5 12.0 22.5 500.0 30 50 20

cally. Sucrose was reagent-grade (Mallinckrodt).

Subjects Two weight-matched groups of weanling female SpragueDawley rats (20 animals per group) were fed either the low-fat (LF) or the high-fat (HF) diet. Rats were aged 21-23 days at the commencement of the experiment and ate the diet ad lib for 4 weeks. This portion of the study was designated "Phase A " and was followed by five to seven consecutive days during which two-hour intake tests (described below) were conducted. Throughout this week of testing the animals continued to eat the diet provided in Phase A. Animals were then fed either the same diet, or the alternate diet for a further four weeks, termed "Phase B . " This yielded four groups of ten animals each (Phase A diet/Phase B diet): LF/LF, LF/HF, HF/LF, HF/HF. Two-hour intake tests were repeated after the second four-week diet phase. Animals were individually housed in standard wire cages in a temperature-controlled (70°F) room. An automatic timer maintained a 12-hour:12-hour light:dark cycle. All testing was conducted midway through the light period and in the animal's home cage.

Composition Per Gram Mixture

Designation

Grams Fat

Grams NaCI

kcal

Peanut Butter 79% Peanut Butter/ 21% Peanut Oil* 59% Peanut Butter/ 41% Peanut Oilt

lowest-fat medium-fat

0.50 0.61

0.013 0.010

5.9 6.6

highest-fat

0.71

0.008

7.2

Peanut butter: Kroger brand, creamy style. Peanut oil: Duke's Brand. *Mixture of I00 g peanut butter and 26 g oil. tMixture of 100 g peanut butter and 70 g oil.

tests. Sucrose preference was measured on the sixth day of the intake tests, using the three-jar choice method. Reagent-grade sucrose (Mallinckrodt) was dissolved (% wt./vol.) in deionized water at two concentrations: 15% and 30%. Preweighed jars of each of these solutions, and a third jar containing deionized water, were placed in each cage for two hours. On the following day, NaCI preference was measured. Reagent-grade NaC1 (Mallinckrodt) was dissolved (% wt./vol.) in deionized water at two concentrations: 0.9% and 1.8%. Preweighed jars of each of these solutions, and a third jar containing deionized water, were placed in each cage for two hours. Only animals maintained on the same diet throughout the study (Groups LF/LF and HF/HF) were tested with these stimuli.

Body Weight Body weight measurements to the nearest gram were obtained twice weekly for all animals.

Data Analysis All data were analyzed using Analysis of Variance for repeated measures (PROC GLM: Statistical Analysis System, release 6.03). RESULTS

Intake Tests

Phase A

Following a five-hour food and water deprivation period, three preweighed 100-ml glass jars containing test stimuli were placed in each animal's home cage and attached to the front wall with wire. The jars were removed after two hours and intake of each stimulus was weighed to the nearest 0.1 g. Dietary fat. Three peanut butter-based mixtures were used to measure preference for novel fat-rich foods. Fat level was manipulated by the addition of peanut oil to peanut butter (Table 2). Rats had access to each of the three mixtures during every test session. Five test sessions were conducted on consecutive days, and the position of the jars was randomized for each animal across the five days. The total grams of fat consumed during each test session was determined for each animal. Total fat intake was calculated by multiplying the quantity consumed of each mixture by the percentage (by weight) of fat it contained (Table 2) and summing these three values. Sucrose and NaCl solutions. Deprivation and measurement conditions were identical to those used in the fat-rich food intake

Body weight. No significant difference was found between the body weights of the two groups immediately prior to intake tests (LF-diet average weight 208 g, s.e. 3.5; HF-diet average weight 213 g, s.e. 3.6), F(1,38)=0.89, NS. Intake of fat-rich mixtures. Animals fed a high-fat diet ate somewhat more fat from the mixtures on all test days relative to animals fed a low-fat diet (Fig. 1), although the main effect of diet did not reach statistical significance, F(1,38) = 2.47, NS. The difference in total consumption between groups was mainly due to greater intake of the highest-fat mixture by the HF-fed animals (Fig. 2). An increase in intake across the five test days was noted for both groups (Fig. 1), which was confirmed statistically by the main effect of trials, F(4,35) = 8.0, p<0.01, that did not interact with diet. Intake of sucrose and NaCl solutions. Animals fed the low-fat diet drank more 0.9% NaC1 relative to animals maintained on the high-fat diet. The LF-fed group also consumed more 1.8% NaC1 than the HF-fed group. Both of these between-group comparisons were significantly different [0.9% NaC1, F(1,18) = 5.91, p<0.05;

DIETARY FAT AND FOOD PREFERENCES

583

PHASE A TESTS

2.5

O--OLow FATO,ET O--OHIGH

2.5-

LOWEST FAT MIXTURE (0.50 groins fat/groin)

2.0-

O - - O L o w Fat Diet

FAT DIET

2.0

O - - O H i g h For m,t

1.5-

O3

(/3 1.5 < cF o

O 1.0"

©~©

1.0

0

1 T

T

"r

0.5.

± ~0

±

T/ Q \

0.0

0.5

0.0

1

2.5.

MEDIUM FAT MIXTURE (0.61 grams fat/gram)

2.0.

1

2

5

4

5

TEST DAY

FIG. 1. Total intake = sum of fat intake from all mixtures. Values are means (standard error).

(/3 < tw

1.5,

i

°I .0.

0.5

1.8% NaCI, F(1,18)=5.43, p<0.05]. Consumption of each of the sucrose solutions did not significantly differ between groups (Table 3).

0.0 2.5

Phase B No significant differences were found between Phase B-HF and Phase B-LF diet groups in consumption of any of the mixtures. Therefore the groups were collapsed across Phase B diet type and were designated Phase A-LF (groups LF/LF and LF/HF) and Phase A-HF (groups HF/LF and HF/HF). Body weight. The average body weights immediately prior to the Phase B intake measures did not differ significantly between the Phase A-LF group (avg. = 300 g, s.e. 7.7) and the Phase AHF group (avg. =297 g, s.c. 5.7), F(1,38)=0.14, NS. Intake of fat-rich mixtures. Phase A-HF animals (groups HF/ LF and HF/HF) consumed more total fat on all five test sessions relative to Phase A-LF animals (groups LF/LF and LF/HF) (Fig. 3). The repeated measures ANOVA yielded main effects of Phase A diet, F(1,38) =7.75, p<0.01, and of test day, F(4,35) =4.58, p < 0 . 0 1 , with no interaction. Comparisons between groups of consumption of each of the mixtures indicated that these differences were mainly attributable to the greater consumption of the nighest-fat mixture by the Phase A-HF group (Fig. 4). For intake of the highest-fat mixture the main effect of Phase A diet was significant, F(1,38)=6.18, p<0.05. Phase A diet was not statistically related to intake of the lowest-fat, F(1,38)=0.01, NS, or the medium-fat, F(1,38)= 2.08, NS, mixtures. Intake of sucrose and NaCI solutions. Animals that had eaten the low-fat diet throughout the entire study (group LF/LF) drank slightly more 0.9% NaC1 relative to animals maintained exclusively on the high-fat diet (group HF/HF). The LF-fed group also consumed more 1.8% NaC1 than the HF-fed group. While the between-group differences in NaC1 solution intake were significant during the Phase A test, the Phase B differences did not statistically differ. Intake of each of the sucrose solutions was similar

o- 9 0 HIGHEST FAT MIXTURE (0.71 grams fot/gr0m)

2.0.

(/3 :£ < tY

1

1.5.

l/el/'\i

0 1.0.

• 0-- 0 ,4'.0/ l ' ~ 0

0.5.

0.0

1

2

3

4

5

TEST DAY

FIG. 2. Intake of each of the fat-rich test mixtures. High-fat diet-reared animals preferred the highest-fat test mixtures. Values are means (standard error).

in both groups (Table 3). DISCUSSION

These data provide additional evidence that rats fed a high-fat diet consume more total fat during short-term tests relative to animals fed a diet lower in fat (14-16). The present data expand upon previous research in two ways. 1) The simultaneous availability of mixtures differing in fat content permitted the assessment of preferences as well as measurement of total fat intake. Rats fed a high-fat diet for four weeks postweaning preferred the nighest-fat mixture and consumed more total fat during intake

584

WARWICK, SCHIFFMAN AND ANDERSON

tests than animals fed the low-fat diet. 2) Upon repetition of the intake tests after a further four weeks the ingestive patterns induced by the initial diet were again displayed. The fat content of the diet consumed during the second four weeks was unrelated to patterns of intake. These data indicate that a high-fat diet consumed early in life may produce physiological changes and/or patterns of intake that are relatively stable despite subsequent dietary change. An important component of fat digestion which is influenced by diet composition is lingual lipase. Lingual lipase is an enzyme of oral origin which catalyzes the breakdown of triglyceride. Secretion of lingual lipase is increased in rats fed a high-fat diet (9), but the effect of this elevated secretion on food preferences and intake is not known. Lingual lipase may modulate sensory perception and acceptance of a fatrich food by regulating the rate at which hedonically positive aromas are released from the food. Increased secretion of lingual lipase induced by early consumption of a high-fat diet may have been a contributing factor in the persistence of preference for the highest-fat mixture. Whereas the sensory properties of fat alone are not sufficient to maintain elevated intake by rats fed a highfat diet (15), rats show clear pattems of preference when selecting among a variety of calorically equivalent fats (12). These preferences indicate that ingestion of fat is at least partially determined by sensory properties and not simply by caloric yield. Early consumption of a high-fat diet resulted in a heightened preference for fat which was sufficiently robust as to persist despite subsequent consumption of a low-fat diet. This finding differs from a recent report (14) which indicated that fat acceptance could be modified by diet composition but was unrelated to dietary history. Methodological differences may account for the discrepancy between those data and the present findings. The dietary phases lasted for four weeks in the present experiment, twice as long as in the Reed and Friedman study (14). The duration of high-fat diet consumption may be critical in determining the persistence of preference for fat. Weanling female rats were used as subjects in the present study, whereas Reed and Freidman used adult male rats. Sensory preferences acquired during early development may be more resistant to change than preferences acquired during adulthood. In the present study the persistence of the relationship between early consumption of a high-fat diet and fat preference during short-term tests is unlikely to be attributable to a learned meal size effect (7,17). The correlation between total fat ingestion on the final day of Phase A testing and the first day of Phase B testing was not significant.

P H A S E B TESTS

3.5

O - - O LOWFATD~ET/PHASEA 0--•

HIGH FAT DIET/PHASE A

5.0

\I

~2.5 << o2.0

/IR /o 0

1.5

0

0

1.0 1

2

3

5

FIG. 3. Total intake = sum of fat intake from all mixtures. Values are means (standard error). Ingestion of NaC1 solutions has previously been found to be related to the concentration of salt in the rearing diet (6). However, this relationship cannot account for the present finding that rats fed the low-fat diet consumed more NaC1 solutions than rats fed the high-fat diet. Although the diets differed in energy density (kcal per gram) and NaC1 content per gram, each diet contained the same amount of NaC1 (0.75 mg) per kcal. Since the average body weights of the two dietary groups did not differ it is likely that the groups ate a similar quantity of kcal, although ad lib intake of the maintenance diet was not measured. The differential intake of NaCI solutions by the two diet groups may have been related to the difference in salt ingestion during the fat-rich mixture intake tests. HF-fed animals ate more of the fatrich mixtures, thereby ingesting more NaCI, than the LF-fed animals. The average total intake of NaC1 obtained from the fat-rich foods as a result of the five Phase A trials was 0.076 g for the LF-fed group and 0.105 g for the HF-fed group. No relationship

TABLE 3

Sodium Chloride

Sucrose

Phase

Water

0.9%

1.8%

Water

15%

30%

Low-Fat

A

0.61 (0.04)

1.84 (0.47)

1.24 (0.25)

0.99 (0.13)

1.03 (0.17)

2.2 (0.66)

High-Fat

A

0.58 (0.08)

0.69* (0.08)

0.61" (0.11)

0.72 (0.12)

0.97 (0.18)

1.77 (0.49)

Low-Fat

B

1.29 (0.13)

2.12 (0.62)

1.43 (0.28)

0.66 (0.12)

1.10 (0.21)

2.81 (0.83)

High-Fat

B

1.15 (0.12)

1.12 (0.27)

0.93 (0.11)

0.85 (0.15)

1.47 (0.34)

2.42 (0.68)

Rats fed a low-fat diet consumed more NaC1 solution than rats fed a high-fat diet. *Significantly different from low-fat diet group (p<0.05). Values are means (standard error),

4

TEST DAY

AVERAGE CONSUMPTION OF SUCROSE AND SALT SOLUTIONS

Diet

0

DIETARY FAT AND FOOD PREFERENCES

2.5.

LOWESTFAT MIXTURE (0.50 grams fat/gram)

2.0

O--OLow ~--~High

(/3 < rY

585

was found between sucrose intake and dietary fat content, in agreement with a previous report (14). The results of the present study suggest that consumption of fat-rich foods during early development induces a shift in food preference that is not readily reversible by subsequent ingestion of a low-fat diet. Further research is needed to establish the relative impact of behavioral and physiological determinants of preference for and ingestion of high-fat foods.

For Diet/Phose A Fat Diet/Phose A

T

1.5

°1.0.

ACKNOWLEDGEMENTS

I

I

j

These data were presented at the Annual Meeting of the Association for Chemoreception Sciences, 1990. Supported in part by NIA Grant AG00443. Rhonda Deems offered advice concerning diet composition, and Darryl Francis, Amy Frey and Brevick Graham provided excellent technical assistance. S. E. Swithers-Mulvey made helpful comments on an earlier version of this report.

0.5.

0.0 2.5

MEDIUM FAT MIXTURE (0.61 grams fat/gram)

2.0

t~ < p,-

1.5

I ! \/\o

°1. 0 .

•T

/o-?

0.5 ±

0.0 2.5

HIGHEST FAT MIXTURE (0.71 grams f a t / g r a m )

2.0

O3

1.5 ©

rr o

1.0-

o/1\ / io\o 1

o

i

i

0.5.

0.0

1

2

3 4 TEST DAY

5

FIG. 4. Intake of each of the fat-rich mixtures. Animals that consumed the HF diet during Phase A preferred the highest-fat mixture. Values are means (standard error).

REFERENCES 1. Beauchamp, G. K. The human preference for excess salt. Am. Sci. 75:27-33; 1987. 2. Beauchamp, G. K.; Cowart, B. J. Congenital and experiential factors in the development of human flavor preferences. Appetite 6:357372; 1985. 3. Birch, L. L. The acquisition of food acceptance patterns in children. In: Boakes, R. A.; Popplewell, D. A.; Burton, M. J., eds. Eating habits: Food, physiology and learned behavior. Chichester: John Wiley and Sons; 1987:107-130. 4. Booth, D. A. Food-conditioned eating preferences and aversions with

interoceptive elements: Conditioned appetites and satieties. Ann. NY Acad. Sci. 575:22-41; 1985. 5. Capretta, P. J. Establishment of food preferences by exposure to ingestive stimuli early in life. In: Barker, L. M.; Best, M. R.; Damjan, M., eds. Learning mechanisms in food selection. Waco: Baylor University Press; 1977:99-121. 6. Contreras, R. J.; Kosten, T. Prenatal and early postnatal sodium chloride intake modifies the solution preferences of adult rats. J. Nutr. 113:1051-1062; 1983. 7. Deutsch, J. A. Signals determining meal size. In: Boakes, R. A.;

586

8.

9. 10.

11.

12.

WARWICK, SCHIFFMAN AND ANDERSON

Popplewell, D. A.; Burton, M. J., eds. Eating habits: Food, physiology and learned behavior. Chichester: John Wiley and Sons; 1987: 155-173. Galef, B. J. Development of flavor preference in man and animals: The role of social and nonsocial factors. In: Aslin, R. N.; Alberts, J. R.; Petersen, M. R., eds. Development of perception, vol. 1. New York: Academic Press; 1981. Hamosh, M. Rat lingual lipase: Factors affecting enzyme activity and secretion. Am. J. Physiol. 235:E416-E421; 1978. Hill, D. L.; Vogt, M. B. Behavioral taste responses in rats "recovered" from sodium deprivation instituted during early development. Association for Chemoreception Sciences XI, Sarasota, April, 1989. London, R. M.; Snowdon, C. T.; Smithana, J. M. Early experience with sour and bitter solutions increases subsequent ingestion. Physiol. Behav. 22:1149-1155; 1979. Naim, M.; Brand, J. G.; Kare, M. R. The preference-aversion be-

13. 14. 15. 16.

17.

havior of rats for nutritionally-controlled diets containing oil or fat. Physiol. Behav. 39:285-290; 1987. Pliner, P. The effects of mere exposure on liking for edible substances. Appetite 3:283-290; 1982. Reed, D. R.; Friedman, M. I. Diet composition alters the acceptance of fat by rats. Appetite 14:219-230; 1990. Reed, D. R.; Tordoff, M. G.; Friedman, M. I. Sham-feeding of corn oil by rats: Sensory and postingestive factors. Physiol. Behav. 47: 779-781; 1990. Tepper, B. J.; Friedman, M. I. Diabetes and a high-fat/low-carbohydrate diet enhance the acceptability of oil emulsions to rats. Physiol. Behav. 45:717-721; 1989. Weingarten, H. P.; Kulikovsky, O. T. Taste-to-postingestive consequence conditioning: Is the rise in sham feeding with repeated experience a learning phenomenon? Physiol. Behav. 45:471--476; 1989.