Physiology &Behavior,Vol. 35, pp. 683-687. Copyright©PergamonPress Ltd., 1985. Printed in the U.S.A.
0031-9384/85$3.00 + .00
Changes in Actual Versus Defended Body Weight Elicited by a Varied, Palatable ("Supermarket") Diet in Rats P H I L I P W I N N 1 A N D L. J. H E R B E R G *
Department of Experimental Psychology, University o f Cambridge and *Institute o f Neurology, Queen Square, London WC1N 3BG, Great Britain R e c e i v e d 25 J u l y 1984 WINN, P. AND L. J. HERBERG. Changes in actual versus defended body weight elicited by a varied, palatable ("supermarket") diet in rats. PHYSIOL BEHAV 35(5) 683-687, 1985.--Male rats fed on a varied, palatable supermarket diet for 11 weeks gained more weight than chow-fed controls. When palatable food was discontinued, their body weights became static, but remained significantly higher than control weights for a further 5 weeks. Hoarding of food was readily elicited by food deprivation in the dietarily obese as well as in the chow-fed rats. Previous studies have shown that the critical body weight at which hoarding appears does not covary with body weight, but appears to reflect a defended level of body weight. In the present study, critical weights were not significantly different between groups before supermarket diet, but were significantly higher in the supermarket rats after obesity had developed. Thus, the increase in body weight brought about by a supermarket diet (unlike that in ventromedial hypothalamic obesity) can be accompanied by an increase in the defended level of body weight as inferred from the critical weight for the onset of hoarding behaviour. Body weight
Diet
Hoarding behavior
E X P E R I M E N T A L obesity has been produced in rats by diets rich in fat [17] or carbohydrate [15] and, more recently, by diets that provide a changing variety of highly palatable snack foods--the so-called "supermarket diet" [28]. Dietary obesity produced in this way is a well-attested phenomenon [21, 22, 28, 30], but its nature and further evolution after discontinuation of the enriched diets are less clear. Some investigators have found enriched diets induce only a temporary increase in body weight which is quickly lost when the palatable foods are withdrawn [24, 25, 28, 30]. However, other investigators have reported that dietary obesity may persist indef'mitely following removal of the palatable food, and even that the excess body weight may be regained after a subsequen$ period of food deprivation [21]. The latter finding suggests that'o.verfeeding may bring about a lasting change in a "defended" leVel at which the rat regulates its food intake and body weight, but other explanations are possible. It is possible, for instance, that weight gain in laboratory rats is ordinarily limited by the low caloric density of laboratory feeds; thus dietary "obesity" might simply be the consequence of lifting this restraint, thus allowing the rats to realise their full genetic potential for weight gain and enabling them to attain their defended levels of body weight for the first time [30]. Another possibility is that changes in body weight may be related more closely to changes in energy expenditure (as heat) than to changes in caloric intake [24]. Such changes need not be regulatory; they could simply be
adventitious metabolic consequences of the diet [1], and need not signify an altered level of 'defended' body weight. Hoarding behavior can be elicited in most (but not all) male laboratory rats. It generally develops only when body weight falls below a critical level after a period of restricted feeding [11,19], Thereafter it is present regardless of whether the animal is hungry or not, and is independent of other measures of activity such as locomotor activity or nestbuilding [12]. Thus the body weight at which hoarding develops may offer a convenient index of the critical weight below which behavioural [14] or metabolic adaptions [1, 2, 3] may supervene to prevent further loss of weight [8, 13, 20]. It has been shown that hoarding elicited in this way can serve to discriminate between different types of obesity [13]. For instance, rats made obese with ventromedial hypothalamic lesions have to lose much more weight than normal rats and must be returned to pre-operative levels before they will start hoarding [11]; on the other hand, genetically obese (fa/fa) Zucker rats, like normal rats, begin to hoard at a level less than 10% below their obese, free feeding weight [14]. Thus, examination of the critical weight for the onset of hoarding in dietarily obese rats might indicate whether these animals resemble ventromedial rats, in which body weights appear to have risen above defended levels, or whether, like obese Zucker rats, they regulate their weight smoothly but at an abnormally high defended level. In examining this question we have not sought to replicate previous studies dem-
Ipresent address: Department of Psychology, University of St. Andrews, St. Andrews, Fife, KY16 9JU.
683
684
WINN AND HERBERG 500
460
.............. • ""'........"
CO
_~420 380
Diet [ .......... Chow
300 I
Weeks --~ 1
I
I
3
I
I
F
I
I
I
9
I
I
11
I
t
13
Supermarket Diet
l
15
d
C
17
i'
L
I
19
J
I
21
I
L
23
d
I
25
fd
FIG. 1. Mean body weight of DIET and CHOW groups. Variance (SE) was never greater than 4.6% of the mean (CHOW group) or 3.1% of the mean (DIET group). No significant differences between the groups were found in the First 3 weeks, F(1,23)=4.22, but were present during the 11 weeks of supermarket diet, F(1,88)=35.80, p<0.001, and during the 11 weeks following withdrawal of the diet, F(1,88)=4.01, p<0.001. Dunnett's t-test showed that differences between the groups persisted until Week 20. fd=food deprivation in weeks 2, 3 and 19.
onstrating the presence or the nature of obesity in supermarket-fed rats, and the present investigation was card e d out without monitoring caloric intake, and without subsequent carcass analysis. METHOD
Ten male Wistar rats were obtained from commercial suppliers (Bantin and Kingman Ltd). At the start of the experiment, these rats weighed 325--415 g and were the same age. They were housed singly under natural illumination at a mean ambient temperature o f approximately 24°C in metal cages partitioned into two separate areas: a home cage supplied with cotton-wool nesting material, and an open area containing pellets in a hopper that could be nibbled but not removed by the rat, and a water bottle (as illustrated in [12]). Water was available at all times, and food pellets (Oxoid 41B Small Animals Feed) and supplementary diet were available as per schedule. Body weights were recorded five times per week. The supplementary " s u p e r m a r k e t " diet comprised unlimited daily rations of any five or six of the following ingredients: tinned corned beef, dry macaroni or spaghetti, salted peanuts, Rich Dundee Cake, cooked lentils, Sugar Puffs breakfast cereal, sliced white bread, peanut butter, Cadbury's milk chocolate, chocolate wholewheat biscuits, sweetened condensed milk, frankfurter sausages, cheddar cheese, cooked ravioli in tomato sauce and marshmallow. Fresh food was supplied daily, including weekends, and stale food was removed from the cages. Laboratory chow was also available. Body weight was recorded on every weekday. Tests for hoarding behaviour were conducted on five days every week at 11.30 a.m. Each rat was weighed, its home cage was cleaned, and it was allowed 15 min access to food in its hopper in order to minimise competition between feeding and hoarding in subsequent tests. F o o d in the hoppers was then removed, a small tray of 2-g food pellets placed at the front end of the open area of the cage, and the number of pellets carded into the home area during the ensuing 10 rain was counted. Three pellets were left in the home area after each test. Hoarding scores recorded in this way, in brief
daily tests, necessarily tend towards an all-or-none pattern, and may be sharply skewed [18]. Group scores were therefore expressed in terms of their medians, in accordance with our previous practice [5,14]. Hoarding behaviour was deemed to be present if the median and mean number of pellets hoarded on 3 consecutive test-days equalled five pellets or more. Use of this criterion discounts the effect of isolated high scores, and is based on the practice followed in previous similar studies [12, 14, 19, 23]. During the first week of testing, the rats were allowed free access to the food in their hoppers. F o r the next two weeks they were placed on a food-deprivation schedule to reduce their body weights and elicit hoarding. F o r this purpose they were restricted to 10 g of pellets at night (additional to 25 min access to food before and during the hoarding tests). At the end of the third week, rats were assigned to either of two groups, matched for body weight. The first (Group DIET) was given continuous access to both Lab chow and supermarket diet for 11 weeks and the second maintained on Lab chow alone (Group CHOW). Hoarding tests were also conducted during the first 2 weeks and the last 2 weeks of the I 1 week period of supermarket diet, after which they were retested, fh'stly with free access to Lab chow, and then, in the following week, under food deprivation. (This sequence can be followed on the abscissa of Fig. 1.) Body-weight data were examined for homogeneity of variance by Bartlett's test, and for normality by a goodness-of-fit procedure [31] prior to analysis of variance. Specific comparisons between groups were made post-hoc by Dunnett's t-test [31]. F o r the hoarding data, a final median score for each rat was calculated. This was the median score of the three obtained when meeting the three-test criterion for hoarding described above. Group medians were calculated from these individual medians and analyzed non-parametrically by the Mann-Whitney U test or Wilcoxon test, as appropriate [29]. RESULTS
Body Weight Mean body weights of the DIET and C H O W groups are
SUPERMARKET DIET AND DEFENDED BODY WEIGHT
685
TABLE 1 CRITICAL BODY WEIGHTS FOR TIlE ONSET OF DEPRIVATION-INDUCED HOARDING OF FOOD BY CHOW AND DIET RATS IN WEEKS 2 AND 3 (PRE-SUPERMARKET) AND WEEK 19 (POST-SUPERMARKET) Pre-supermarket
Body weight (g) % Pre-deprivation weight Median no. pellets hoarded
Post-supermarket
DIET
CHOW
DIET
CHOW
327.1 ± 4.6 95.4 -+ 0.4
318.6 ± 4.5 96.3 ± 0.6
465.3 --- 6.22* 95.8 ± 0.6
441.9 _+9.1 95.2 ± 0.2
21
22
22
17
*p<0.05 compared to CHOW group post-supermarket (t(28)=2.14).
shown in Fig. 1. During the fh'st three weeks all rats were treated identically (Week 1 ad lib chow; Weeks 2 and 3 food deprivation) and no significant differences between groups were observed. A clear difference between the groups developed during the 11 weeks when supermarket diet was available, (Groups: F(1,88)=35.80, p<0.001; Weeks: F(10,88)= 18.87, p<0.001; Interaction: F(10,88)=0.17, NS). During this period, the DIET rats gained an average of 153.5-+10.1 g compared to 127.20-+9.97 g gained by the CHOW rats. However, much of this divergence occurred soon after the supermarket diet became available, significant differences between the groups being present even in the first week (Week 4: t=2.61, p<0.05; Dunnett's test). From this time onwards the groups gained weight more or less in parallel, as the lack of significant interaction by ANOVA indicates. Body weights of the two groups were also compared in the 11 weeks following withdrawal of the supermarket diet. Overall ANOVA revealed the presence of significant differences during this period (Groups: F(1,88)=4.01, p<0.001). Post-hoc tests established that significant differences between the groups existed only in weeks 15-20 (Week 15, t=2.97, p<0.01; Week 16, t=2.54, Week 17, t=2.41, Week 18, t=2.04, Week 19, t =2.84, all p<0.05; Week 20, t=0.84, NS). It is important to note that during this period the DIET rats did not lose weight (except of course in week 19 when all rats were food deprived).
Hoarding Table 1 gives details of the deprivation induced hoarding activity elicited in Weeks 2 and 3 (pre-supermarket) and Week 19 (post-supermarket). Hoarding tests were also conducted while both groups were on unrestricted diets in Weeks 1, 4, 5, 13-18 and 20, but no significant hoarding occurred during these weeks, the median number of pellets hoarded per test ranging from 0 to 5. In Weeks 2 and 3 (food-deprived), all rats met the pre-set criterion for hoarding after having lost 4-5% of their body weight: no differences between DIET and CHOW rats were found. (Body weight: t(28)= 1.31; % weight loss: t(28)= 1.30; pellets hoarded: U=84.5 nl,n2 = 15.) In the post-supermarket period, when the rats were again food-deprived (Week 19), there were still no differences between the groups in respect of the percentage weight loss necessary to induce hoarding, t(28)=1.00, nor in the number of pellets hoarded, U=85 nl,n2=15, even though the DIET rats were significantly
heavier, t(28)=2.14, p<0.001. Comparison of records obtained before and after the period of the supermarket diet showed significant increases in absolute body weight in both groups (DIET: t(14)=18.94, p<0.001; CHOW: t(14)=18.67, p<0.001), but no change in the number of pellets hoarded (Wilcoxon test: DIET: t=25.5, N=15; CHOW: t=32.5, N= 14) and no change in the percentage weight loss necessary to induce hoarding (DIET: t(14)=0.48; CHOW: t (14) = 1.72). DISCUSSION
These results confirm previous reports that a palatable diet may accelerate weight gain in the adult rat. When the palatable diet was discontinued, the rats stopped gaining weight, but unlike some subjects of previous studies (e.g., [25]), they did not actually lose weight, and they remained at more or less the same elevated weight levels until the end of the experiment. In this limited sense their weight gains were persistent. However, after their body weights had become static, the weights of the CHOW group continued to show a normal progressive increase, with the result that the significant difference between the two groups eventually disappeared. Thus, over the whole course of the experiment, the supermarket rats did not gain more weight than controls, they only gained it sooner. Nevertheless, the issue as to whether the obesity was "persistent" or "temporary" is not merely terminological; a legitimate underlying question is whether the enhanced body weight after supermarket feeding represented a new defended level of body weight regulation, or whether it was merely a passive aftermath of the preceding hyperphagia. The hoarding tests were designed to answer this question. Previous studies have shown that defended obesity (as in the Zucker rat) [7] is associated with a normal hoarding response, occurring at a critical weight less than 10% below the pre-deprivation body weight [14]. In ventromedial obesity, on the other hand, rats may have to lose over 40% of their body weight, reverting to near pre-operative levels, before they will take steps to defend their body weight, whether by accepting quinine-flavoured mash [9] or by hoarding [ 11]. In the present study, the dietarily obese rats were significantly heavier than before, and significantly heavier than control rats, but the percentage weight loss necessary to elicit hoarding (4.2%) was no greater than it had been when they were lean (4.6%). This shows that the increase in their body weight during their exposure to an enriched diet was
686
WINN AND HERBERG
matched by a corresponding increase in their defended levels of body weight. This result suggests that dietarily induced obesity is a 'defended' obesity, and that it differs in this respect from the obesity that develops after ventromedial hypothalamic lesions. This conclusion agrees with that of B. J. Rolls and her colleagues [21] but at first sight seems to be at odds with the results of certain other crucial investigations. Sclafani and colleagues (e.g., [28]) have found that dietarily obese (female) rats, like ventromedial rats, were less willing to accept quinine-flavoured food or to work for food, and that they shed much of their excess weight when they were returned to a diet of laboratory chow. Similarly, supermarket-fed male rats [26], like obese ventromedial rats [27], showed a significantly delayed onset of deprivation-induced running when all food was stopped. These characteristic features clearly point to the presence of a ventromedial-like component in dietary obesity; in other words, there must have been a gain in excess of the regulated level of body weight in addition to any rise in the regulated level itself. The notion of a " m i x e d " obesity of this sort has previously been offered to account for the hoarding patterns seen in the normal, ageing laboratory rat [13]. The predominance of only the ventromedial-like component in the rats studied by Sclafani and colleagues [25, 26, 28] can perhaps be accounted for firstly by their use of female subjects [28], and secondly, by the relatively brief time for which their rats had access to the enriched diet [26]. In female rats, endogenous oestrogens strongly depress the regulated level of body weight and the critical weight for hoarding [5,10], and accordingly facilitate the appearance of obesity of the ventromedial type [6]. These properties of oestrogen would presumably promote the appearance of 'ventromedial' features in the supermarket-fed female.
Male rats, on the other hand, have generally been reported to maintain their body weights when supermarket feeding is discontinued, or at most to lose barely 5% of their gain---as compared to a figure of approximately 40% for females over a similar period [25]. The 'ventromedial' behaviour of Sclafani and Rendel's male rats [26] is a striking exception to this pattern, and is perhaps to be explained by the relatively brief length of time (30 days) for which the enriched diet was available to them. This seems the more likely in the light of a recent report that adipocyte hypertrophy, seen from an early stage in supermarket-fed rats, may be followed by an irreversible adipocyte hyperplasia-but only after a minimum of 8 weeks of overfeeding [16]. Thus, at least two factors---sex of subject and total duration of overfeeding--seem likely to play a critical role in determining the long-term evolution of dietary obesity. But subject to these two factors, the present result supports the claim [21] that enriched diets can lead to long-term changes in regulated body weight. The progressive recalibration, or slippage, of the homeostat, implied by this result, may seem less surprising if considered alongside another example: the lowering of body temperature that occurs in hibernation. Both phenomena can be seen to provide an economical way of maximising internal constancy in the face of a prolonged external challenge.
ACKNOWLEDGEMENTS We thank Julie Gautrey, Roy Hammans, Lorraine Evans and Susan Chandler for their valuable assistance. This work was supported by a grant from the Brain Research Trust of the Institute of Neurology. P.W. was supported by the Pinsent-Darwin Fund of the University of Cambridge.
REFERENCES 1. Armitage, G., R. B. S. Harris, G. R. Hervey and G. Tobin. The part played by variation of energy expenditure in the regulation of energy balance. In: The Body Weight Regulatory System: Normal and Disturbed Mechanisms, edited by L. A. Cioffi, W. P. T. James and T. B. Van Itallie. New York: Raven Press, 1981, pp. 137-141. 2. Boyle, P. C., L. H. Storlien and R. E. Keesey. Increased efficiency of food utilisation following weight loss. Physiol Behav 21: 261-264, 1978. 3. Bray, G. A. Effect of caloric restriction on energy expenditure in obese patients. Lancet 2: 397-398, 1969. 4. Cabanac, M., R. Duclaux and N. H. Spector. Sensory feedback in regulation of body weight: is there a ponderostat? Nature 229: 125--127, 1971. 5. Coling, J. G. and L. J. Herberg. Effect of ovarian and exogenous hormones on defended body weight, actual body weight, and the paradoxical hoarding of food by females rats. Physiol Behav 29: 687-691, 1982. 6. Cox, V. C., J. W. Kakolewski and E. S. Valenstein. Ventromedial hypothalamic lesions and changes in body weight and food consumption in male or female rats. J Comp Physiol Psychol 67: 320-326, 1969. 7. Cruce, J. A. F., M. R. C. Greenwood, P. R. Johnson and D. Quartermaln. Genetic versus hypothalamic obesity: studies of intake and dietary manipulations in rats. J Comp Physiol Psychol 87: 295-301, 1974. 8. Fantino, M. and M. Cabanac. Body weight regulation with a proportional hoarding response in the rat. Physiol Behav 24: 939-942, 1980.
9. Franklin, K. B. J. and L. J. Herberg. Ventromedial syndrome: the rat's "finickiness" results from obesity, not from the lesions. J Comp Physiol Psychol 87: 410-414, 1974. 10. Gale, S. K. and A. Sclafani. Comparison of ovarian and hypothalamic obesity syndromes in the female rat: effects of diet palatability on food intake and body weight. J Comp Physiol Psychol 91: 381-392, 1977. l 1. Herberg, L. J. and J. E. Blundell. Non-interaction of ventromedial and lateral hypothalamic mechanisms in the regulation of feeding and hoarding behaviour in the rat. Q J Exp Psychol 22: 133-141, 1970. 12. Herberg, L. J., J. G. Pye and J. E. Blundell. Sex differences in the hypothalamic regulation of food hoarding: hormones versus calories. Anim Behav 20: 186-191, 1972. 13. Herberg, L. J., K. B. J. Franklin and D. N. Stephens. The hypothalamic 'set-point' in experimental obesity. In: Recent Advances in Obesity Research, edited by A. Howard. London: Newman Publishing Ltd, 1975, pp. 235-237. 14. Herberg, L. J. and P. Winn. Body weight regulatory mechanisms and food hoarding in hereditarily obese (fa/fa) and lean (Fa/Fa) Zucker rats. Physiol Behav 29: 631-635, 1982. 15. Kanarek, R. B. and E. Hirsch. Dietary-induced overeating in experimental animals. Fed Proc 36: 154-158, 1977. 16. Mandenoff, A., T. Lenoir and M. Apfelbaum. Tardy occurrence of adipocyte hyperplasia in cafeteria-fed rat. Am J Physiol 242: R349-R351, 1982. 17. Mickelson, O., S. Takahashi and C. Craig. Experimental obesity: I. Production of obesity in rats by feeding high fat diets. J Nutr 57: 541-544, 1955.
SUPERMARKET DIET AND DEFENDED
BODY WEIGHT
18. Morgan, C. T. The statistical treatment of hoarding data. J Comp Psychol 38: 247-256, 1945. 19. Morgan, C. T., E. Stellar and O. Johnson. Food deprivation and hoarding in rats. J Comp Physiol Psychol 35: 275-295, 1943. 20. Nyby, J. and D. D. Thiessen. Food hoarding in the Mongolian gerbil (meriones unguiculatus), effects of food deprivation. Behav Neural Biol 30: 39--48, 1980. 21. Rolls, B. J., E. A. Rowe and R. C. Turner. Persistent obesity in rats following a period of consumption of a mixed, high energy diet. J Physiol 298: 415-427, 1980. 22. Rolls, B. J., P. M. Van Duijvenvoorde and E. A. Rowe. Variety in the diet enhances intake in a meal and contributes to the development of obesity in the rat. Physiol Behav 31: 21-27, 1983. 23. Ross, S., W. I. Smith and B. L. Woessner. Hoarding: an analysis of experiments and trends. J Gen Psychol 52: 307-326, 1945. 24. Rothwell, N. J. and M. J. Stock. Regulation of energy balance in two models of reversible obesity in the rat. J Comp Physiol Psychol 93: 1024-1034, 1979.
687 25. Sclafani, A. and A. N. Gorman. Effects of age, sex and prior body weight on the development of dietary obesity in adult rats. Physiol Behav 18: 1021-1026, 1977. 26. Sclafani, A. and A. Rendell. Food deprivation-induced activity in dietary obese, dietary lean and normal-weight rats. Behav Biol 24: 220-228, 1978. 27. Sclafani, A. and A. Rendell. Food deprivation-induced activity in normal and hypothalamic obese rats. Behav Biol 24: 244-255, 1978. 28. Sclafani, A. and D. Springer. Dietary obesity in adult rats: Similarities to hypothalamic and human obesity syndromes. Physiol Behav 17: 461--471, 1976. 29. Siegel, S. Nonparametric Statistics. New York: McGraw-Hill, 1956. 30. Stephens, D. N. Growth and development of dietary obesity in adulthood of rats which have been undernourished during development. Br J Nutr 44: 215-227, 1980. 31. Winer, B. J. Statistical Principles in Experimental Design. New York: McGraw-Hill, 1971.