Drinking saccharin increases food intake and preference—I. Comparison with other drinks

Appetite, 1989, 12, l-10

Drinking Saccharin Increases Food Intake and Preference-l. Comparison with Other Drinks MICHAEL

G. TORDOFF

and MARK

I. FRIEDMAN

Monell Chemical Senses Center, Philadelphia

To examinethe orosensoryand postingestiveeffectsof saccharinsolution on food intake and food preference,freely feedingratsweregivenflavored food to eat and a solution to drink for 2 h on eight to ten occasions.Relativeto trials with a different flavored food and only water to drink, food intake wasincreasedby drinking 0.2% saccharinor 0.45%NaCl, unaffectedby drinking 1%almondextract, anddecreased by drinking 10%glucose.Food preference,which wasassessed in a choicetest with simultaneousaccessto the two flavored foods, was increasedby drinking 0.2% saccharinor 10%glucoseand unaffectedby drinking 1%almondextract or 0.45% NaCl. Theseresultsareconsistentwith thepossibilitythat a combinationof the oral and hydrational propertiesof saccharinsolutionincreasefood intake. Saccharin’s sweettaste may be responsiblefor its effectson food preference.

INTRODUCTION

Numerous studies have taken advantage of saccharin’s sweet, non-nutritive properties to gain insight into the psychophysical and hedonic aspects of sweet taste, the physiological and neurological basis of ingestive behavior, and the mechanism of reinforcement in instrumental and classical conditioning. Despite the fact that humans frequently drink saccharin or other non-nutritive sweeteners along with food, there has been little research on the effects of non-nutritive sweet drinks on food intake and selection (e.g. Blundell & Hill, 1986; Parham & Parham, 1980; Stellman & Garfinkel, 1986). Studies in animals suggest that over the long-term, drinking saccharin solution or eating food containing saccharin has little or no effect on food intake or body weight (e.g. Adolph, 1947; Friedhoff et al., 1971; Hausmann, 1933; Mook, 1974; Strouthes, 1977; but see Rolls, 1987). Over the short-term, severely food-deprived rats that drink saccharin solution alter subsequent intake of glucose solution (Mook et al., 1981), and may decrease food intake and die (Strouthes, 1973). In this study, we examined the short-term effects of saccharin solution consumption on ad libitum food intake, and the influence of this experience on flavored food This work was supported by National Institute of Arthritis, Diabetes, and Digestive and Kidney Diseases Grant DK-36339, a feasibility study grant provided by the Obesity Core Center, St Luke%Roosevelt Hospital, and by BRSG SO7 RR05826-07 awarded by the Biomedical Research Support Grant Program, Division of Research Resources, National Institutes of Health. Audrey Buxbaum provided expert technical assistance. We thank Israel Ramirez for his helpful comments on this series of manuscripts. Portions of this work were presented at “Mechanisms of Appetite and Obesity”, a satellite symposium of the 15th Annual Meeting of the Society for Neuroscience, San Antonio, TX, U.S.A., 1985. Address correspondence and reprint requests to: M. G. Tordoff, Monell Chemical Senses Center, 3500 Market St., Philadelphia, PA 19104, U.S.A. 0195-6663/89/010001+

10 $03.00/O

0

1989 Academic

Press Limited

2

M. G. TORDOFF

AND

M. I. FRIEDMAN

preference. We found that rats that drank 0.2% saccharin solution increased food intake and developed a preference for the flavored food consumed. These effects were compared with results from experiments in which rats drank solutions of 10% glucose, 0.45% NaCl, or 1% almond extract. METHOD

Subjects and Maintenance

All experiments used male Sprague-Dawley CD rats (280-38Og) obtained from Charles River Laboratories (Wilmington, MA). They were housed individually in a vivarium at approximately 21 “C on a 12 : 12 h light : dark cycle (lights off at 1500 hrs). The rats’ maintenance food was powdered Purina Laboratory Chow (no. 5001), which was presented in glass or aluminium feeding cups mounted inside their cages, Tap water was available at all times except for brief periods during one study when almond extract solution was given. The rats were allowed to adapt to vivarium conditions for at least 10 days before tests. In order to habituate them to the testing procedure, they also received l-3 pretests, consisting of removal and replacement of their maintenance food at the times when this would be done in the following experiment. Training Trials

Each experiment consisted of a series of sequential daily training trials, with 120 min access to a drinking solution (e.g. 0.2% saccharin) on half the trials, according to a counterbalanced repeating ABBA design. During the test period, the rats’ maintenance chow was replaced by the same chow flavored with 08% (by weight) non-nutritive chicken or chocolate flavor (International Flavors and Fragrances, nos 135- 19762 and 135-68919). One flavor was always associated with the drinking solution and the other was given during the same period on days when no drink was provided. Assignment of flavors to the solution-paired or unpaired conditions was counterbalanced. The two flavors were chosen because pilot work and other results (Naim et al., 1986; Tordoff & Friedman, 1986) suggested that rats mildly and equally preferred them relative to unflavored chow. The flavored food was made in 1 kg batches and mixed with a commercial mixer for at least 15 min to ensure thorough distribution of the flavor. Preference

Tests

Preference for a flavored food paired with a drinking solution was inferred from a two-choice test conducted on the day after the last training trial. At the start of the dark period, all rats were given both flavored foods simultaneously. Intake of each was measured after 0.5, 1,2,4, and 24 h. The position of the flavored foods in the cage was counterbalanced, and was alternated after each measurement period.

In four separate experiments, rats drank solutions of 0.2% sodium saccharin (Sigma Chemical Co., St Louis, MO), 10% glucose, 0.45% sodium chloride (NaCl), or 1% almond extract solution (McCormick & Co., Baltimore, MD). The rats received ten pairs of trials with the saccharin, glucose, and NaCl solutions, but only eight pairs of

SACCHARIN

AND FEEDING

3

trials with the almond solution. The concentrations of saccharin, glucose, and NaCl were chosen because rats typically drink large quantities of them. On the other hand, they only slightly prefer 1% almond solution to water (unpublished observations), although they can easily detect it (Tordoff et al., 1982). Thus, to encourage almond solution ingestion, whenever it was given the rats’ maintenance water was removed. Water was always available in the other experiments. Analysis Statistical analysis of food intake during training trials was accomplished by twoway, within-subject analyses of variance (ANOVAs) with factors of drinking solution (presence or absence) and trial pair (S-10, depending on the experiment). Solution intake was analyzed by one-way, within-subject ANOVAs with trial pair as the factor. When a main effect or interaction was significant, differences between means of individual trials were examined using Newman-Keuls’ post hoc tests (Bruning & Kintz, 1977). Cumulative intakes during preference tests were compared by within-subject t-tests. All tests were conducted with p < 0.05 as the critical probability cut-off. Values given in the text are meanskstandard errors of the mean.

RESULTS

Saccharin Solution Rats ate significantly more food when saccharin was available than when it was not [F( 1,10) = 19.9, p < 0.01 (Figure l)]. The effect of saccharin on food intake was greatest during the middle of the 20-day test period (saccharin x trial pair interaction, F(9,90) = 3.56, p < 0.05; see Figure 1 for differences between individual pairs of trials). Food intake on each trial pair increased with repeated testing [F(9,90) = 3.17, p
M. G. TORDOFF AND M. I. FRIEDMAN

4

25

20

2

15

s !! G

10

9 5

0 1

2

3

4 Trial Pair

Intake of 0.2% saccharin solution (top panel) and flavored food (bottom panel) during ten pairs of 2 h trials. Open bars show food intake with only water to drink, shaded bars show food intake with water and saccharin to drink. * p < 0.05: Significantly greater food intake with saccharin than with water to drink. FIGURE 1.

TABLE 1 Number of rats eating more and preferring food given with a solution to drink

Solution increased food Drinking solution

N

Intake

0.2% saccharin 10% glucose 0.45% NaCl 1% almond

11 8 14 12

11 0 11 5

Preference 9” 8 7 7

Note: Values are the number of subjects that on average ate more food when they drank the solution than when they did not (intake), and ate more flavored food that had been paired with solution consumption in a 4-h choice test (preference). a Based on ten rats, because one spilled its food.

Glucose intake increased progressively during the first five training trials [F(9,63) = 15.3, p < 0401], but intakes were stable over the remainder of the experiment (Figure 2). The volume of glucose solution consumed (an average of 23.3 ml/2 h test) was roughly equivalent to the volume of saccharin solution consumed in the previous experiment (19.2 ml/2 h test).

SACCHARIN

AND FEEDING

5

25

20 3 0 s 15 2 m I z IO P 5

0

1 Trial Pair

FIGURE2. Intake of 10% glucose solution (top panel) and flavored food (bottom panel) during ten pairs of 2 h trials. Open bars show food intake with only water to drink, shaded bars show food intake with water and glucose to drink. The top (lighter) portion of the shaded bars shows calories derived from the glucose. There were no significant effects of drinking glucose on calorie intake.

Total calorie intake (chow x 3.4 kcal/g plus glucose x 0.38 kcal/ml) increased with repeated testing [F(9,63) = 2.41, p < 0.051. Calorie intake was unaltered by drinking glucose, both over the entire test [F( 1,7) = 1.54, NS], and on individual pairs of trials (glucose x trial pair interaction [F(9,63) =0.53, KS]). Thus, it appeared that the rats’ intake offood decreased in compensation for the additional calories ingested as glucose solution. In the two-choice preference test, rats showed a strong preference for the flavored food previously associated with glucose solution (Figure 5; Table 1). A significant difference in intake of the two flavored foods was present after 30min [t(7)=2.79, p < 0.051, and persisted for the entire 24 h test (24 h intakes: paired flavor, 20.9 + 1.2 g; unpaired flavor, 7.6 + 1.4 g [t(7) = 5.78, p -c 0.011). The preference was apparently robust: The rats were used for a pilot experiment lasting 45 days that was unrelated to the present work, and were then given a two-choice preference test. Intakes of the glucose paired and unpaired flavored foods in the first 2 h of the retest were 4.2 + 1.2 g vs. 1.4 + @5g, respectively [t(7) = 2.46, p < @05J.

0.45% NaCE Solution Rats ate significantly more food during tests with 0.45% NaCl than without the solution [F( 1,13) = l@O,p < 0.01 (Figure 3)]. There were differences in intake between the ten trial pairs [F(9,117) =4.34, p
6

M. G. TORDOFF

AND

M. I. FRIEDMAN

30 e ;

20 -

% E 9

10

Y O25 -

Trial Pair

FIGURE3. Intake of 0.45% NaCl solution (top panel) and flavored food (bottom panel) during ten pairs of 2 h trials. Open bars show food intake with only water to drink, shaded bars show food intake with water and NaCl to drink. * p < 005; Over the entire experiment, food intake was significantly greater when drinking 0.45% NaCl than only water.

OL 25

20 % a 2 E

15

“0 10 9 5

il

2 O

1 Trial Pair

FIGURE4. Intake of 1% almond extract solution (top panel) and flavored food (bottom panel) during eight pairs of 2 h trials. Open bars show food intake with water to drink, shaded bars show food intake with almond to drink. There were no significant effects of drinking almond on calorie intake.

SACCHARIN

7

AND FEEDING

30 0.2% Saccharin

10% Glucose

0.45% NaCl

1% Almond

20 -

1

2

3

4

1

2

3

4

Test Duration (h)

FIGURE 5. Two-choice food preference tests of four groups of rats (Figures 14) that were previously given flavored foods together with either a drink (m, paired) or just water (0, unpaired).

Rats drank similar quantities of 0.45% NaCl during each of the ten 2 h test periods [F(9,117)= 1.47, NS (Figure 3)]. In the two-choice preference test, rats neither preferred nor avoided the flavored food previously given with 0.45% NaCl (see Figure 5). After 24 h, intakes of the paired and unpaired flavored foods were 12.0 + 1.7 g and 13.4 _+1.5 g, respectively [t(l3) = 1.07, NS].

Almond Solution

Food intake increased with repeated testing [F(7,77)=3.11, p<@Ol], but was unaffected by providing almond solution to drink [F( 1,ll) = 0.23, NS]. There was also no interaction of almond with trial pair [F(7,77) =0.79, NS (Figure 4)]. Rats drank similar quantities of the 1% almond solution and water during the 2 h tests (11.7 + 1.0 ml vs. 125 f 1.2 ml). There were significant increases of fluid intake with repeated testing [F(7,77)=2.18, p <0*05], and also a significant test pair x fluid interaction [F(7,77)=2.66, p ~0.051, resulting from a trend of rats to drink more almond solution on later trials than earlier ones (Figure 4) but maintain stable water intakes (1st trial = 11.9 _+l.Oml; 8th trial = 11.3 + l*Oml). The two flavored foods were equally preferred during the preference test (Figure 5).

8

M. G. TORDOFFAND M. I. FRIEDMAN DISCUSSION

The four drinks produced different effects on food intake and food preference: Food intake was increased by drinking 0.2% saccharin or 0.45% NaCl, unaffected by drinking 1% almond, and decreased by drinking 10% glucose. Food preference was increased by drinking 0.2% saccharin or 10% glucose, and unaffected by drinking 1% almond or 0.45% NaCl. To our knowledge, the finding that rats drinking saccharin increased food intake is the first unequivocal demonstration that drinking a non-nutritive sweetener can significantly increase short-term food intake. It is consistent with previous reports that rats given food containing Sucaryl (a commercially prepared saccharin-cyclamate mixture) increased meal size (Louis-Sylvestre & LeMagnen, 1980), that fpod intake tended to increase after rats or rabbits drank saccharin (e.g. Geiselman, 1985; Mook et al., 1981), and with recent work showing that human subjects increased hunger ratings after they drank solutions of saccharin, aspartame, or Acesulfame-K (e.g. Blundell & Hill, 1986; Rogers et al., 1988). On the other hand, there are many studies where drinking saccharin did not influence food intake (e.g. Adolph, 1947; Friedhoff et al., 1971; Hausmann, 1933; Mook, 1974; Strouthes, 1977). The differences in results can most likely be explained by differences in test duration, subject deprivation state, and the use of repeated tests of the same animals. Repeated testing may be particularly important, as in our experiment, the increased feeding response did not occur the first few times rats drank saccharin. Rats that drank 045% NaCl increased food intake but the increase was somewhat smaller and less consistent than that produced by drinking saccharin. The 0.2% saccharin and 0.45% NaCl solution have in common a low osmotic pressure, lack of calories, and a “good taste”. Of these, the low osmotic pressure appears to provide the most congruous explanation of the increased food intake. Consistent with the saccharin solution’s greater hydrational effects (lower osmolarity and larger volume ingested), rats that drank @2% saccharin tended to increase food intake more than those that drank 0.45% NaCl. Although rats that drank the hyposmotic almond solution did not increase food intake, they drank no more of this solution than they did water, and so its hydrational effects may have been insufficient to activate feeding behavior. Rats that drank 10% glucose, which was hyperosmotic, reliably decreased food intake. Thus, the results point to the possibility that “overhydration” may be one factor responsible for the increase in food intake observed when rats drank saccharin. It seems unlikely that the increase in food intake produced by drinking saccharin can be explained solely on the basis of saccharin’s lack of calories because rats did not increase food intake when they were given non-caloric almond extract to drink. The results with the almond and glucose drinks also negate the possibility that presenting saccharin to drink somehow conditioned or signalled the rats to eat (Weingarten, 1983). In addition to providing hedonic value, the orosensory effects of the solutions may initiate “cephalic-phase” hormonal and metabolic reflexes, which have been implicated in the feeding response produced by palatable foods and drinks (Geiselman & Novin, 1982; Louis-Sylvestre & LeMagnen, 1980; Powley, 1977). Sweet taste is a good initiator of cephalic-phase reflexes in rats (Grill et al., 1984), and this may account for why drinking saccharin was particularly effective in increasing food intake. Presumably, drinking glucose solution would also initiate cephalic-phase reflexes, but the resulting stimulation of appetite would be offset by inhibition produced by the hyperosmotic and

SACCHARIN AND FEEDING

9

caloric effects of glucose. Consistent with this, there was a slight (but non-significant) increase in total (glucose + food) caloric intake when rats drank glucose (Figure 2). It is difficult to extend a “cephalic-phase” explanation to cover all the results found here, however, because it is unclear whether the 045% NaCl drink produces a cephalic-phase response. Small volumes of NaCl infused directly into the mouth do not elicit a cephalic-phase insulin response (Berridge et al., 1981), although the same may not be true for other cephalic-phase responses or for larger volumes of NaCl ingested voluntarily. It may be that 0.45% NaCl ingestion increases feeding through its hydrational effects, whereas saccharin ingestion increases food intake through both cephalic-phase and hydrational mechanisms. Although food intake was increased by ingestion of 0.2% saccharin and 0.45% NaCl, food preference was increased by ingestion of 0.2% saccharin and 10% glucose only. The two drinks that increased food preference are both sweet, and it therefore seems possible that the sweetness of saccharin is important for its effects on food preference. The failure of 0.45% NaCl consumption to support the development of a taste preference, despite its hedonically positive value to rats (judged by the volume ingested), implies that the food preferences were not simply the result of “hedonic” taste-taste conditioning (e.g. Fanselow & Birk, 1982; see also Tordoff & Friedman, 1989 b). Taken together, the results suggest that saccharin’s effect on food preference is mediated by its sweet taste but its effects on food intake are determined by more than one attribute, such as sweetness and hydration. Based on the results found here, however, these can only be tentative conclusions. In the following three papers, we provide a more definitive analysis of the behavioral and physiological controls of the interaction between drinking saccharin and eating food (Tordoff & Friedman, 1989 b). REFERENCES Adolph, E. F. (1947) Urges to eat and drink in rats. American Journal ofPhysiology, 151,110-125. Berridge, K, Grill, H. J. & Norgren, R. (1981) Relation of consummatory responses and preabsorptive insulin release to palatability and learned taste aversions. Journal of Comparative and Physiological Psychology, 95, 363-382.

Blundell, J. E. & Hill, A. J. (1986) Paradoxical effects of an intense sweetener (aspartame) on appetite. Lancet, i, 1092-1093. Bruning, J. L. & Kintz, B. L. (1977) Computational handbook of statistics. Dallas: Scott, Foresman. Fanselow, M. S. & Birk, J. (1982) Flavor-flavor associations induce hedonic shifts in taste preference. Animal Learning and Behavior, 10, 223-228. Friedhoff, R., Simon, J. A. & Friedhoff, A. J. (1971) Sucrose solution vs. no-calorie sweetener vs. water in weight gain. Journal of the American Dietetic Association, 59, 185-486. Geiselman, P. J. (1985) Feeding patterns after normal ingestion and intragastric infusion of glucose, fructose, and galactose in the rabbit. Nutrition ana’ Behavior, 2, 175-188. Geiselman, P. J. & Novin, D. (1982) The role of carbohydrates in appetite, hunger and obesity. Appetite, 3, 203-223.

Grill, H. J., Berridge, K. C. & Ganster, D. J. (1984) Oral glucose is the prime elicitor of preabsorptive insulin secretion. American Journal of Physiology, 246, R88-R95. Hausmann, M. F. (1933) The behavior of albino rats in choosing foods. II. Differentiation between sugar and saccharin. Journal of Comparative Psychology, 15, 419428. Louis-Sylvestre, J. & LeMagnen, J. (1980) Palatability and preabsorptive insulin release. Neuroscience and Biobehavioral Reviews, 4 (Suppl. 1), 43-46.

Mook, D. G. (1974) Saccharin preference in the rat: some unpalatable Reviews, 81, 475-490.

findings. Psychological

10

M. G. TORDOFF

AND M. I. FRIEDMAN

Mook, D. G., Kushner, B. D. & Kushner, D. G. (1981) Release of feeding by the sweet taste in rats: The specificity of oral satiety. Appetite, 2, 267-280. Naim, M., Brand, J. G., Christensen, C. M., Kare, M. R. & Van Buren, S. (1986) Preferences of rats for food flavors and texture in nutritionally controlled semi-purified diets. Physiology and Behavior, 37, 15-22.

Parham, E. S. & Parham, A. R. (1980) Saccharin use and sugar intake by college students. Journal of the American Dietetic Association, 76, 56&563. Powley, T. L. (1977) The ventromedial hypothalamic syndrome, satiety, and a cephalic phase hypothesis. Psychological Reviews, 84, 89-126. Rogers, P. J., Carlyle, J., Hill, J. & Blundell, J. E. (1988) Uncoupling sweet taste and calories: Comparison of the effects of glucose and three intense sweeteners on hunger and food intake. Physiology and Behavior, 43, 547-552. Rolls, B. J. (1987) Sweetness and satiety. In J. Dobbing (Ed.), Sweetness. Pp. 161-173. London: Springer. Stellman, S. D. & Garfinkel, L. (1986) Artificial sweetener use and one-year weight change among women. Preventive Medicine, 15, 195-202. Strouthes, A. (1973) Saccharin drinking and mortality in rats. Physiology and Behavior, 10,781791. Strouthes, A. (1977) Saccharin eating in undeprived and hungry rats. Animal Learning and Behavior, 5, 4246.

Tordoff, M. G. & Friedman, M. I. (1986) Hepatic portal glucose infusions decrease food intake and increase food preference. American Journal of Physiology, 251, R192-R196. Tordoff, M. G. & Friedman, M. I. (1989a) Drinking saccharin increased food intake and preference--II. Hydrational factors. Appetite, 12, 11-21. Tordoff, M. G. & Friedman, M. I. (1989 b) Drinking saccharin increases food intake and preference-III. Sensory and associative factors. Appetite, 12, 23-35. Tordoff, M. G. & Friedman, M. I. (1989 c) Drinking saccharin increases food intake and preference-IV. Cephalic phase and metabolic factors. Appetite, 12, 37-56. Tordoff, M. G., Jelinek, M. & Holman, E. W. (1982) Backward conditioning of taste preferences with electrical stimulation of the brain as a reinforcer. In B. G. Hoebel & D. Novin (Eds.), The neural basis offeeding and reward. Pp. 123-127. Brunswick, ME: Haer Institute for Electrophysiological Research. Weingarten, H. P. (1983) Conditioned cues elicit feeding in sated rats: a role for learning in meal initiation. Science, 220, 431433. Received

30 January

1987