- Email: [email protected]
Sham feeding is inhibited by dietary-induced obesity in rats
Physiology &Behavior,Vol. 31, pp. 533-537. PergamonPress Ltd., 1983. Printedin the U.S.A.
Sham Feeding Is Inhibited by Dietary-Induced Obesity in Rats D E N N I S A. V A N D E R W E E L E
Department of Psychology, Occidental College, Los Angeles, CA 90041 AND THEODORE
B. V A N I T A L L I E
College of Physicians and Surgeons, Columbia University R e c e i v e d 5 J a n u a r y 1983 VANDERWEELE, D. A. AND T. B. VAN ITALLIE. Sham feeding is inhibited by dietary-induced obesity in rats. PHYSIOL BEHAV 31(4) 533-537, 1983.--A group of six female, albino rats were maintained on a cafeteria diet of cookies, milk, and elevated-fat (shortening), rat-chow mixture and rat chow while a similar group received only rat chow ad lib for 17 weeks. When the groups differed significantly in mean body weight (obese--387.5 g, controls--287.2 g; p<0.001), gastric fistulas were implanted in each animal. After recovery, the rats were adapted to a liquid diet and assessed for sham feeding. Control-fed, normal-body-weight subjects showed substantial sham feeding when ingesting the Vivonex with the fistulas open compared to fistula-closed intake; meal frequency, meal size (apart from the initial meal) and total food intake were significantly increased while the satiety ratios following each meal were significantly decreased. Obese animals showed no significant increased feeding and satiety ratios were unreliably altered; while normal-body-weight controls increased 4-hr food intakes by 93% and halved their mean-satiety ratios the obese animals showed an 8% increase in 4-hr food intake and only a 22% decrease in mean satiety ratios. We offer the hypothesis that, when animals are induced to become obese by palatable and varied diets which are then terminated, the anorexia produced is independent of gastrointestinal interactions inasmuch as that anorexia extends to sham feeding. Sham-feeding
Obesity
Palatable diets
Gastric fistulas
WHEN animals ingest food and that food is then removed before intestinal absorption, a significant increase in the rate and amount of ingestion occurs; this phenomenon has been labelled sham feeding and is well documented [5, 7, 8, 9, 11, 16, 26]. The animal eats but gains none of the metabolic benefits from the sham-fed nutrients. An animal with a gastric fistula (a hollow tube leading to the outside from the inside of the stomach) shows tremendous elevations in feeding when the fistula is open. This increased ingestion following sham feeding weighs heavily against a theory of satiety based entirely upon an oropharyngeal metering of caloric intake. One of the earliest studies [7] dealing with sham-feeding techniques concluded that oropharyngeal stimuli were sufficient to stop feeding in the dog but were relatively weak in prolonging the typical non-feeding after meals in mealfeeding animals. We have labelled the stopping of ingestion within the meals as "intrameal" satiety and the maintenance of the intermeal interval as "intermeal" satiety [23]; oropharyngeal stimuli may be more effective components of intrameal than intermeal satiety.
Anorexia
Atropine methylbromide
Rats
Perhaps a closer examination of feeding patterns in some representative mammals would be instructive. The rat, rabbit, cat and dog are meal eaters; that is, they show a relatively short period of vigorous eating which is often terminated abruptly or with only slight deceleration (intrameal satiety) and is usually followed by a longer period of nonfeeding or intermeal satiety. It seems that sham feeding (which prevents gastric, intestinal and postabsorptive stimuli) most uniformly disrupts intermeal satiety as all reports indicate that the first intermeal interval following the initial sham-fed meal is significantly reduced [5, 8, 9, 11, 26]. Two previous reports hypothesized that sham feeding indicates that what we have labelled intermeal satiety depends critically on the buildup of " a n inhibitory reflex elicited by ingested food in the stomach and/or small intestine" [5,26]. Two other studies, however, state that the initial sham-fed meal is increased as well, indicating that intrameal satiety might be compromised by sham feeding [8,11]. One further point might be made concerning shamfeeding studies. Sham feeding does not bypass the oral and cephalic phase of hormonal release. Indeed, when animals
~This work was performed with the aid of Faculty Development Funds, Occidental College and research grant No. AM-17259 and center grant No. AM 17624 to the Obesity Research Center, Columbia University. We also received facility support from New York University, School of Medicine and technical assistance and support from Ms. Kathy Simon.
Copyright © 1983 Pergamon Press Ltd.--0031-9384/83/100533-05503.00
534
VANI)ERWEF, IA~" AND V,AN ll&l,[,}t
received injections of atropine methyl nitrate, which presumably would reduce the vagally mediated gastrointestinal hormonal release [2] and slow stomach emptying [19], sham feeding was blocked [13]. Dosages of 10 to 1250 #g/kg of atropine acted in a linear fashion. Perhaps this amount of atropine might make animals uncomfortable or it may also be that the rise in insulin secretion generated by sham feeding is inappropriate and drives the increased ingestion: atropine would reduce this hyperinsulinemia. However, stimuli with no demonstrated physiologic influence on insulin levels (cholecystokinin, bombesin) have been shown to inhibit sham feeding [12,14] and this inhibition has been advanced as evidence for the role of these substances in satiety. In this experiment, we assessed the effects of obesity upon food ingestion and especially sham-feeding. Obesity in rats can be produced by offering the animal a varied and presumably palatable diet [18], which induces rapid weight gain and elevates insulin levels [21]. We have recently shown that raising the plasma levels of insulin without first inducing obesity leads to a restriction in food consumption [22] and this finding has now been demonstrated in other species as well [1,24]. Insulin elevation in the normal-weight animal might accelerate the clearance of nutrients from the blood and, hence, halt meals sooner than otherwise, thus partially supporting Brobeck's hypothesis [4]. In the absence of nutrients entering the intestines for absorption during sham feeding and t h e presence of elevated insulin produced by cephalic-phase vagal mediation, one might well predict the decreased hepatic glucose uptake and increased ingestion witnessed in the animal feeding with an open fistula. In the present experiment, we hypothesize that obese animals, because of their hyperinsulinemia and greater food-stimulated insulin release, will show enhanced levels of sham feeding when compared to similarly treated but normal-weight subjects. METHOD Twelve Sprague-Dawley (Holtzman) female rats served as subjects in the experiment. Six were made obese by offering an array of palatable foods [21], viz., (a) chocolate chip cookies, (b) sweetened, condensed milk in water (1:I, v/v), (c) vegetable shortening in Purina powdered chow (1:2 w/w), and (d) Purina chow Bloks (Purina pelleted chow weighing 5-7 g). The remaining 6 subjects received Purina chow Bloks only. All animals received water ad lib and were weighed weekly. After 17 weeks, the two groups differed by approximately 100 g in body weight, t(10)=6.39, p<0.001, and gastric fistulas were implanted [26]. All animals were maintained on solid food for the next 2 weeks and then returned to the original diets with one substitution. The experimental group now received Vivonex (vanilla flavor, 1.4 Kcal/cc) in place of the milk and control subjects received the Vivonex solution in addition to Purina chow. Once the group weight differences were re-established, all animals were fed only Vivonex for 1 week. Following this week, all subjects were deprived of food overnight (19.5 hr) and then refed Vivonex from calibrated Richter tubes for 4 hr. In addition, 3 animals from each group had been trained to bar press for Vivonex on a CRF schedule of reinforcement. Both situations began at 1030 hr and intakes were recorded over the following 4 hr. Subjects tested with Richter tubes were continuously observed by 2 experimenters who recorded all drinking; drinking of animals in the operant chambers was monitored electromechanically. In each situation a meal was defined as at
least one minute of drinking, w,ithout an interruptitm hmg~:r than 10 sec, o1 continuous bar pressing during which minimum of 0.5 cc liquid diet was corlsumcd a ~, rcinfi)fcc ment. In addition, meals were separated by at lea~t 10 rain o~ the animal was considered to have paused within a single meal. Animals were adapted to the testing situation by feeding with fistulas closed for three days. In this first phase of the study, 3 animals were next tested once for feeding with the fistula open while the 3 animals trained to bar press were tested twice, once with Richter tubes and once ir~ the oper'mt chambers. Forty days after the initial fistula-open test. the 6 obese animals were retested twice more while drinking fiom Richter tubes. Atropine methylbromide (100 /,tg/kg) w a s subcutaneously injected in isotonic saline as vehicle to cvaluate the effects of a presumed vagal efferent block on sham feeding in these subjects. Three rats received atropine first, fob lowed 4 days later by saline: the order was reversed for the other 3 subjects. All data were evaluated by t tests. Comparisons of group means between obese and control subjects used t for nonrepeated measures while comparisons between fistula closed and open conditions used a repeated-measures t. Atropine data were also evaluated with a repeated-measures t-test. RESULTS Animals offered varied and palatable diets gained substantial body weight (Tables I and 2). During the week before feeding with the fistula open, both groups were maintained on the Vivonex diet that was used for the assessment of sham feeding. During this week, obese animals ingested approximately 55 to 73% of the control subjects' intake. The obese group continued this relative anorexia into the adaptation period of feeding with fistulas closed, as seen by con> paring the ingestion of obese and control subjects in the last column of Table 1. Obese animals consumed a mean of 26.6 Kcals during the 4-hr sessions, which was 60.W,4: of the control group's consumption, t{ 10)=3.22, p<(I,05. Normal-weight control animals showed elevated food ingestion when feeding with the fistula open (sham feeding). As can be seen in Table 1, the meal frequency is significantly increased as is total 4-hr food ingestion with fistula-open feeding. The average meal size was also increased if the initial meal was omitted from this average. The initial meal was large and quite variable in this group of animals in both fistula open and closed conditions because it terminated the overnight fast; this may account for the nonsignificant difference here while others [9, II] have found meal size to be significantly altered by sham feeding. Sham feeding required collection of ingested nutrients to insure that all intake had drained from the stomach. We retained for data analysis no trial in which less than 10{F~, of ingested fluid, or more including gastric secretion, was obtained by, collection from drainage tubes into pans below cages. This criterion required 3 trials in one animal and 2 trials in another, both in the control group. Table 2 demonstrates the significant difference in satiety ratio, defined as meal size in ml divided into the interval of non-feeding minus the criterion 10 min between that meal and the next, for control animals when under the sham feeding condition. Satiety ratio for the initial meal was lower under all conditions for all subjects as would be expected since the first meal terminated the overnight fast. One can see, however, that there was a significant reduction in overall and initial satiety ratios when control-weight subjects fed
SHAM F E E D I N G IN OBESE RATS
535
TABLE 1 MEAN MEAL SIZE, MEAL FREQUENCY,AND TOTALINGESTIONFOR 6 CONTROL AND 6 OBESE RATSDURING4-HR INTAKEPERIODS WITHCLOSED AND OPEN FISTULAS (FISTULACLOSEDDATAARE FOR LASTADAPTATION DAYONLY) Body Weight (g)
Meal Size (ml)
Meal Frequency
Total 4-hr intake
Control Fistula Closed Fistula Open Open-Closed
293.8 ± 2.12 306.0 ± 2.09
10.6 ± 0.3 12.8 ± 0.9 +21%*
3.4 ± 0.2 5.3 ± 0.3 + 56%t
31.2 ± 2.81 60.2 _+ 3.95 +93%t
Obese Fistula Closed Fistula Open Open-Closed
385.6 _+ 2.71 400.9 ± 2.73
6.3 ___0.4 5.7 ± 0.3
3.3 ± 0.4 4.0 ± 0.3
19.0 ± 2.0 20.6 ± 2.08
- 10%
+21%
+8%
*+59% If initial meal not included, p<0.05. tp<0.02. with the fistula open (p<0.02 and p<0.05 respectively, repeated measures t tests). This measure indicates that food was less satiating for control subjects in the fistula-open condition and one can argue that these subjets were hungrier also, as meal size and frequency increased overall when animals sham fed. While control-weight subjects showed clear evidence of sham feeding, obese subjects showed no reliable increases in either meal frequency, meal size, or total 4-hr food ingestion and no significant decreases in the two satiety ratios generated, although their mean satiety ratio decreased as did that of the controls. Finally, atropine methyl bromide had no effect on feeding by obese animals in the fistula-open condition. Control animals were not tested with atropine while sham feeding. None of the ingestion measures was affected in either group by bar pressing for food vs. simple consumption of liquid food from the Richter tubes. Mean meal sizes of all meals taken with fistulas closed were insignificantly reduced when animals drank from the tubes--4.9 compared to 5.7 ml in the operant situatiorv--while average meal frequency was identical in the two situations. Therefore, these data were combined with those from the Richter tube study. DISCUSSION We observed no significant indications of excessive eating while sham feeding in obese subjects in the present study. Under the same paradigm, control-weight subjects showed marked increases, comparable to that reported by others in the literature ([13] most recently). We had hypothesized that the increased feeding when animals are sham fed might be even more enhanced in fat animals, as raised insulin levels might clear the plasma of nutrients more quickly into adipocytes and other insulin-dependent tissue and lead the animal to a hypoglycemic condition. We have shown earlier that animals induced to become obese and then returned to eating only Purina chow did show elevated insulin levels which remained frankly hyperinsulinemic one week after termination of the varied palatable diets [21]. Even though sham feeding does not allow nutrient entry into the gastrointestinal tract, cephalic insulin would stimulate uptake of any circulating nutrient and that achieved from glycogenolysis and/or gluconeogenesis (as suggested by the ecork of Steffens
TABLE 2 SATIETYRATIOSOF 6 CONTROLAND 6 OBESE SUBJECTS Mean Body Weight (g)
Initial Satiety Ratio
Mean Satiety Ratio
Control Fistula Closed Fistula Open Open-Closed
293.8 ± 2.12 306.0 ± 2.09
0.18 ± 0.02 0.I0 ± 0.01 -44%*
0.52 ± 0.04 0.26 ± 0.02 -50%t
Obese Fistula Closed Fistula Open Open-Closed
385.6 ± 2.71 400.9 ± 2.73
0.58 ± 0 . 1 0 0.77 _+ 0.19 +33c~
1 . 2 0 ± 0.31 0.94 ± 0.21 -22%
*p<0.05. tp <0.02.
[20]). Evidently hyperinsulinemia associated with obesity does not enhance sham feeding or, more correctly, the anorexia noted with obesity induced by offering rats palatable diets and then removing the diets, is not dependent upon stimuli derived from the interaction of diet with the gastrointestinal tract and the absorption of that diet. This latter point might need some clarification. If the anorexia of fattened animals was dependent upon gut interaction or absorption of nutrients, hypophagia would no longer be evident and increased eating during sham feeding would be as robust or more so in obese animals than in normal-weight subjects. Fat animals remained hypophagic, however, when sham fed. When rats are induced to become obese, by such methods as lateral hypothalamic stimulation, insulin injection, palatable diet feeding, or force-feeding by gastric tubes, and then the obesity-inducing stimulation is halted, the obese animal is invariably hypophagic or anorexic compared to normal-weight control subjects. We anticipated that this anorexia might be attributed to the raised gastrointestinal secretory activity associated with overfeeding, absorption and weight gain. It appears that some other signal independent of gastric fill or duodenal distention and ab-
53t~
VANI)ERWEEI.[-.' A N ! ) VAN I'I~1 [1~!
sorption p r o m o t e s this relative hypophagia, at least in the dietary-obese animals of the present study. The present study also confirms earlier studies that indicate intermeal satiety as immediately affected when animals sham feed. The intermeal interval was reduced and meal f r e q u e n c y significantly increased e v e n in a 4-hr interval. The initial satiety ratio was reduced by 44% in control animals and was unaffected in obese animals. Indeed, this is the effect we o b s e r v e d in our earlier study on meal patterns in obese rats [23]; obese animals ate smaller meals with greater satiety ratios than did controls. W h e n animals are obese, less food appears to satiate longer and hence, intra- and intermeal satiety are heightened. The intermeal satiety effect (satiety ratio) persisted when obese animals were sham fed while the control-weight subjects showed an immediate decrease in this effect. Intermeal satiety, we suggest, may be more dependent upon gut/nutrient interaction and absorption than is intrameal. B r o b e c k has suggested that nutrient clearance might be an important factor in determining appetite and ingestion of food as a function o f body weight. In obese subjects, already enlarged adipocytes and repleted glycogen stores might retard nutrient clearance and reduce food ingestion while, in the lean animal, the r e v e r s e situation (depleted nutrient stores) would lead to an increased rate o f ingestion. Indeed, many studies (e.g., [3, 6, 10, 15]) have shown animals to be more efficient in storing and utilizing nutrients during realimentation after e n f o r c e d fasts. There was no nutrient absorption from ingestion in our fattened, sham fed animals yet, despite this fact, hypophagia reamined. Unless glycogenolysis following ingestion occurs and is delayed in clearance, it seems unlikely that B r o b e c k ' s hypothesis [4] can account for our present results. Perhaps the simplest explanation for our results is that fat
animals are anorexic because of a signal which is m no w a sensitive to gastrointestinal events. The animal, thus. continues to show anorexia e v e n when sham t\~eding That the fistulas may have produced greater discomfor~ in lhc ~bcsc animals and lowered feeding is unlikely ~ testllls ill the fistula-closed condition arc very similar 1o those obtained before in non-fistulated animals [21]. Also, one cannot s i r e ply argue that the obese subjects in the present study did not like the diet offered while normal-weight subjects did: both groups had been feeding and maintaining weight differences on the diet for one w e e k before sham feeding was ussessed. Finally, we offer one further explanation thai may account for the absence of significant increased ingestion while sham feeding in obese rats. Lorenz, Gibbs and Smith 1121 have shown that atropine and CCK also attenuate sham feeding. One might ask what factors these various stimuli possesx in c o m m o n . Methylated atropine remains outside the blood-brain-barrier and, hence, it is presumed that its major effect is to inhibit cholinergic muscarinic receptors in the periphery. Such pharmacological action would decrease stomach emptying II9]. C C K has been shown to retard stomach emptying [17.25] and, perhaps in high dosages, inhibit gastric motility. Could it be that slowing the rate of stomach emptying renders the stimulus of gastric events to satiety noninformative? Thus. the animal might shift to other cues to regulate food intake. Indeed, the d e v e l o p m e n t of obesity may lead to a stomach which processes fl~od more slowly and presents food to the d u o d e n u m at a slower rate. If the organism, prior to becoming obese, was regulating food ingestion primarily on the basis of stomach/duodenal e m p t y ing and absorption interactions, with this signal now blunted, subjects might turn to postabsorptivc or more longterm regulatory mechanisms. We are now testing this hypothesis.
REFERENCES 1. Anika, S. M., T. R. Houpt and K. A. Houpt. Insulin as a satiety hormone. Physiol Behav 25: 21-23, 1980. 2. Berthoud, H. R. and B. Jeanrenaud. Sham feeding-induced cephalic phase insulin release in the rat. Am J Physiol 242: E280-E285, 1982. 3. Boyle, P. C., L. H. Storlien and R. E. Keesey. Increased efficiency of food utilization following weight loss. Physiol Behav 21: 261-264, 1978. 4. Brobeck, J. R. Nature of satiety signals. Am J Clin Nutr 28: 806-807, 1975. 5. Davis, J. D. and C. Campbell. Peripheral control of meal size in the rat: Effect of sham feeding on meal size and drinking rate. J Comp Physiol Psychol 83: 37%387, 1973. 6. Fabry, P. Metabolic consequences of the patterns of food intake. In: Handbook of Physiology, Section 6, Alimentary Canal, vol l, edited by C. F. Code. Washington, DC: Am Physiol Society, 1967, pp. 31-49. 7. Janowitz, H. D. and M. I. Grossman. Some factors that affect the food intake of normal dogs and of dogs with esophagostomy and gastric fistula. Am J Physiol 159: 143-148, 1949. 8. Kraly, F. S., W. J. Carry, S. Resnick and G. P. Smith. Effect of cholecystokinin on meal size and intermeal interval in the sham-feeding rat. J Comp Physiol Psychol 92: 697-707, 1978. 9. Kraly, F. S., W. J. Carty and G. P. Smith. Effect of pregastric food stimuli on meal size and intermeal interval in the rat. Physiol Behav 20" 77%784, 1978. 10. Levitsky, D. A., I. Faust and M. Glassman. The ingestion of food and the recovery of body weight following fasting in the naive rat. Physiol Behav 17: 575-580, 1976.
I I. Liebling, D. S., J. D. Eisner, J. Gibbs and G. P. Smith. Intestinal satiety in rats. J Comp Physiol Psyehol 89: 955-965. 1975. 12. Lorenz, D., J. Gibbs and G. P. Smith. Effect of cholecystokinin and other gut hormones on sham feeding in the rat. In: G,t Hormones. edited by S. R. Bloom. London: Churchill Livingstone, 1978, pp. 224-226. 13. Lorenz, D., P. Nardi and G. P. Smith. Atropine methyl nitrate inhibits sham feeding in the rat. Pharmacol Biochem Behav 8: 405-407, 1978. 14. Martin, C. F. and J. Gibbs. Bombesin elicits satiety in sham feeding rats. Peptides 1: 131-134, 1980. 15. Meyer, J. H. and W. J. Clawson. Undernutrition and subsequent realimentation in rats and sleep. J Anita Sci 23:214-224. 1964. 16. Mook, D. G. Oral and postingestional determinants of the intake of various solutions in rats with esophageal fistulas..I Comp Physiol Psychol 56: 645-659, 1963. 17. Moran, T. H. and P. R. McHugh. Cholecystokinin supresses food intake by inhibiting gastric emptying. Am .I Physiol. in press, 1982. 18. Sclafani, A. and D. Springer. Dietary obesity in adult rats: Similarities to hypothalamic and human obesity syndromes, Physiol Behav 17: 461-471, 1976. 19. Setler. P. E. and G. P. Smith. Gastric emptying in rats with chronic gastric fistulas. Am J Dig Dis (N.S.) 14: 137-142, 1969. 20. Strubbe, J. H. and A. B. Steffens. Rapid insulin release after ingestion of a meal in the unanesthetized rat. A m J Phsyiol 229: 101%1022, 1975.
SHAM FEEDING
IN OBESE RATS
21. VanderWeele, D. A., E. Haraczkiewiz and T. B. Van Itallie. Elevated insulin and satiety in obese and normal-weight rats. Appetite 3: 99-109, 1982. 22. VanderWeele, D. A., F. X. Pi-Sunyer. D. Novin and M. J. Bush. Chronic insulin infusion suppresses food ingestion and body weight gain in rats. Brain Res Bull 5: Suppl 4, 7-11, 1980. 23. Van ltallie. T. B. and D. A. VanderWeele. The phenomenon of satiety. In: Recent Advances in Obesity Research: HI, edited by P. Borntorp, M. Cairella and A. N. Howard. London: John Libbey Press, 1981, pp. 278-289.
537
24. Woods, S. C., L. D. McKay, L. J, Stein, D. B. West, E. C. Lotter and D. Porte, Jr. Neuroendocrine regulation of food intake and body weight. Brain Res Bull 5: Suppl 4, 1-7, 1980. 25. Yamagishi, T. and H. T. Debas. Cholecystokinin inhibits gastric emptying by acting on both proximal stomach and pylorus. Am J Physiol 234: E375-E378, 1978. 26. Young, R. C., J. Gibbs, J. Antin, J. Holt and G. P. Smith. Absence of satiety during sham feeding in the rat. J Comp Physiol Psyehol 87: 795-800, 1974.
Sham Feeding Is Inhibited by Dietary-Induced Obesity in Rats D E N N I S A. V A N D E R W E E L E
Department of Psychology, Occidental College, Los Angeles, CA 90041 AND THEODORE
B. V A N I T A L L I E
College of Physicians and Surgeons, Columbia University R e c e i v e d 5 J a n u a r y 1983 VANDERWEELE, D. A. AND T. B. VAN ITALLIE. Sham feeding is inhibited by dietary-induced obesity in rats. PHYSIOL BEHAV 31(4) 533-537, 1983.--A group of six female, albino rats were maintained on a cafeteria diet of cookies, milk, and elevated-fat (shortening), rat-chow mixture and rat chow while a similar group received only rat chow ad lib for 17 weeks. When the groups differed significantly in mean body weight (obese--387.5 g, controls--287.2 g; p<0.001), gastric fistulas were implanted in each animal. After recovery, the rats were adapted to a liquid diet and assessed for sham feeding. Control-fed, normal-body-weight subjects showed substantial sham feeding when ingesting the Vivonex with the fistulas open compared to fistula-closed intake; meal frequency, meal size (apart from the initial meal) and total food intake were significantly increased while the satiety ratios following each meal were significantly decreased. Obese animals showed no significant increased feeding and satiety ratios were unreliably altered; while normal-body-weight controls increased 4-hr food intakes by 93% and halved their mean-satiety ratios the obese animals showed an 8% increase in 4-hr food intake and only a 22% decrease in mean satiety ratios. We offer the hypothesis that, when animals are induced to become obese by palatable and varied diets which are then terminated, the anorexia produced is independent of gastrointestinal interactions inasmuch as that anorexia extends to sham feeding. Sham-feeding
Obesity
Palatable diets
Gastric fistulas
WHEN animals ingest food and that food is then removed before intestinal absorption, a significant increase in the rate and amount of ingestion occurs; this phenomenon has been labelled sham feeding and is well documented [5, 7, 8, 9, 11, 16, 26]. The animal eats but gains none of the metabolic benefits from the sham-fed nutrients. An animal with a gastric fistula (a hollow tube leading to the outside from the inside of the stomach) shows tremendous elevations in feeding when the fistula is open. This increased ingestion following sham feeding weighs heavily against a theory of satiety based entirely upon an oropharyngeal metering of caloric intake. One of the earliest studies [7] dealing with sham-feeding techniques concluded that oropharyngeal stimuli were sufficient to stop feeding in the dog but were relatively weak in prolonging the typical non-feeding after meals in mealfeeding animals. We have labelled the stopping of ingestion within the meals as "intrameal" satiety and the maintenance of the intermeal interval as "intermeal" satiety [23]; oropharyngeal stimuli may be more effective components of intrameal than intermeal satiety.
Anorexia
Atropine methylbromide
Rats
Perhaps a closer examination of feeding patterns in some representative mammals would be instructive. The rat, rabbit, cat and dog are meal eaters; that is, they show a relatively short period of vigorous eating which is often terminated abruptly or with only slight deceleration (intrameal satiety) and is usually followed by a longer period of nonfeeding or intermeal satiety. It seems that sham feeding (which prevents gastric, intestinal and postabsorptive stimuli) most uniformly disrupts intermeal satiety as all reports indicate that the first intermeal interval following the initial sham-fed meal is significantly reduced [5, 8, 9, 11, 26]. Two previous reports hypothesized that sham feeding indicates that what we have labelled intermeal satiety depends critically on the buildup of " a n inhibitory reflex elicited by ingested food in the stomach and/or small intestine" [5,26]. Two other studies, however, state that the initial sham-fed meal is increased as well, indicating that intrameal satiety might be compromised by sham feeding [8,11]. One further point might be made concerning shamfeeding studies. Sham feeding does not bypass the oral and cephalic phase of hormonal release. Indeed, when animals
~This work was performed with the aid of Faculty Development Funds, Occidental College and research grant No. AM-17259 and center grant No. AM 17624 to the Obesity Research Center, Columbia University. We also received facility support from New York University, School of Medicine and technical assistance and support from Ms. Kathy Simon.
Copyright © 1983 Pergamon Press Ltd.--0031-9384/83/100533-05503.00
534
VANI)ERWEF, IA~" AND V,AN ll&l,[,}t
received injections of atropine methyl nitrate, which presumably would reduce the vagally mediated gastrointestinal hormonal release [2] and slow stomach emptying [19], sham feeding was blocked [13]. Dosages of 10 to 1250 #g/kg of atropine acted in a linear fashion. Perhaps this amount of atropine might make animals uncomfortable or it may also be that the rise in insulin secretion generated by sham feeding is inappropriate and drives the increased ingestion: atropine would reduce this hyperinsulinemia. However, stimuli with no demonstrated physiologic influence on insulin levels (cholecystokinin, bombesin) have been shown to inhibit sham feeding [12,14] and this inhibition has been advanced as evidence for the role of these substances in satiety. In this experiment, we assessed the effects of obesity upon food ingestion and especially sham-feeding. Obesity in rats can be produced by offering the animal a varied and presumably palatable diet [18], which induces rapid weight gain and elevates insulin levels [21]. We have recently shown that raising the plasma levels of insulin without first inducing obesity leads to a restriction in food consumption [22] and this finding has now been demonstrated in other species as well [1,24]. Insulin elevation in the normal-weight animal might accelerate the clearance of nutrients from the blood and, hence, halt meals sooner than otherwise, thus partially supporting Brobeck's hypothesis [4]. In the absence of nutrients entering the intestines for absorption during sham feeding and t h e presence of elevated insulin produced by cephalic-phase vagal mediation, one might well predict the decreased hepatic glucose uptake and increased ingestion witnessed in the animal feeding with an open fistula. In the present experiment, we hypothesize that obese animals, because of their hyperinsulinemia and greater food-stimulated insulin release, will show enhanced levels of sham feeding when compared to similarly treated but normal-weight subjects. METHOD Twelve Sprague-Dawley (Holtzman) female rats served as subjects in the experiment. Six were made obese by offering an array of palatable foods [21], viz., (a) chocolate chip cookies, (b) sweetened, condensed milk in water (1:I, v/v), (c) vegetable shortening in Purina powdered chow (1:2 w/w), and (d) Purina chow Bloks (Purina pelleted chow weighing 5-7 g). The remaining 6 subjects received Purina chow Bloks only. All animals received water ad lib and were weighed weekly. After 17 weeks, the two groups differed by approximately 100 g in body weight, t(10)=6.39, p<0.001, and gastric fistulas were implanted [26]. All animals were maintained on solid food for the next 2 weeks and then returned to the original diets with one substitution. The experimental group now received Vivonex (vanilla flavor, 1.4 Kcal/cc) in place of the milk and control subjects received the Vivonex solution in addition to Purina chow. Once the group weight differences were re-established, all animals were fed only Vivonex for 1 week. Following this week, all subjects were deprived of food overnight (19.5 hr) and then refed Vivonex from calibrated Richter tubes for 4 hr. In addition, 3 animals from each group had been trained to bar press for Vivonex on a CRF schedule of reinforcement. Both situations began at 1030 hr and intakes were recorded over the following 4 hr. Subjects tested with Richter tubes were continuously observed by 2 experimenters who recorded all drinking; drinking of animals in the operant chambers was monitored electromechanically. In each situation a meal was defined as at
least one minute of drinking, w,ithout an interruptitm hmg~:r than 10 sec, o1 continuous bar pressing during which minimum of 0.5 cc liquid diet was corlsumcd a ~, rcinfi)fcc ment. In addition, meals were separated by at lea~t 10 rain o~ the animal was considered to have paused within a single meal. Animals were adapted to the testing situation by feeding with fistulas closed for three days. In this first phase of the study, 3 animals were next tested once for feeding with the fistula open while the 3 animals trained to bar press were tested twice, once with Richter tubes and once ir~ the oper'mt chambers. Forty days after the initial fistula-open test. the 6 obese animals were retested twice more while drinking fiom Richter tubes. Atropine methylbromide (100 /,tg/kg) w a s subcutaneously injected in isotonic saline as vehicle to cvaluate the effects of a presumed vagal efferent block on sham feeding in these subjects. Three rats received atropine first, fob lowed 4 days later by saline: the order was reversed for the other 3 subjects. All data were evaluated by t tests. Comparisons of group means between obese and control subjects used t for nonrepeated measures while comparisons between fistula closed and open conditions used a repeated-measures t. Atropine data were also evaluated with a repeated-measures t-test. RESULTS Animals offered varied and palatable diets gained substantial body weight (Tables I and 2). During the week before feeding with the fistula open, both groups were maintained on the Vivonex diet that was used for the assessment of sham feeding. During this week, obese animals ingested approximately 55 to 73% of the control subjects' intake. The obese group continued this relative anorexia into the adaptation period of feeding with fistulas closed, as seen by con> paring the ingestion of obese and control subjects in the last column of Table 1. Obese animals consumed a mean of 26.6 Kcals during the 4-hr sessions, which was 60.W,4: of the control group's consumption, t{ 10)=3.22, p<(I,05. Normal-weight control animals showed elevated food ingestion when feeding with the fistula open (sham feeding). As can be seen in Table 1, the meal frequency is significantly increased as is total 4-hr food ingestion with fistula-open feeding. The average meal size was also increased if the initial meal was omitted from this average. The initial meal was large and quite variable in this group of animals in both fistula open and closed conditions because it terminated the overnight fast; this may account for the nonsignificant difference here while others [9, II] have found meal size to be significantly altered by sham feeding. Sham feeding required collection of ingested nutrients to insure that all intake had drained from the stomach. We retained for data analysis no trial in which less than 10{F~, of ingested fluid, or more including gastric secretion, was obtained by, collection from drainage tubes into pans below cages. This criterion required 3 trials in one animal and 2 trials in another, both in the control group. Table 2 demonstrates the significant difference in satiety ratio, defined as meal size in ml divided into the interval of non-feeding minus the criterion 10 min between that meal and the next, for control animals when under the sham feeding condition. Satiety ratio for the initial meal was lower under all conditions for all subjects as would be expected since the first meal terminated the overnight fast. One can see, however, that there was a significant reduction in overall and initial satiety ratios when control-weight subjects fed
SHAM F E E D I N G IN OBESE RATS
535
TABLE 1 MEAN MEAL SIZE, MEAL FREQUENCY,AND TOTALINGESTIONFOR 6 CONTROL AND 6 OBESE RATSDURING4-HR INTAKEPERIODS WITHCLOSED AND OPEN FISTULAS (FISTULACLOSEDDATAARE FOR LASTADAPTATION DAYONLY) Body Weight (g)
Meal Size (ml)
Meal Frequency
Total 4-hr intake
Control Fistula Closed Fistula Open Open-Closed
293.8 ± 2.12 306.0 ± 2.09
10.6 ± 0.3 12.8 ± 0.9 +21%*
3.4 ± 0.2 5.3 ± 0.3 + 56%t
31.2 ± 2.81 60.2 _+ 3.95 +93%t
Obese Fistula Closed Fistula Open Open-Closed
385.6 _+ 2.71 400.9 ± 2.73
6.3 ___0.4 5.7 ± 0.3
3.3 ± 0.4 4.0 ± 0.3
19.0 ± 2.0 20.6 ± 2.08
- 10%
+21%
+8%
*+59% If initial meal not included, p<0.05. tp<0.02. with the fistula open (p<0.02 and p<0.05 respectively, repeated measures t tests). This measure indicates that food was less satiating for control subjects in the fistula-open condition and one can argue that these subjets were hungrier also, as meal size and frequency increased overall when animals sham fed. While control-weight subjects showed clear evidence of sham feeding, obese subjects showed no reliable increases in either meal frequency, meal size, or total 4-hr food ingestion and no significant decreases in the two satiety ratios generated, although their mean satiety ratio decreased as did that of the controls. Finally, atropine methyl bromide had no effect on feeding by obese animals in the fistula-open condition. Control animals were not tested with atropine while sham feeding. None of the ingestion measures was affected in either group by bar pressing for food vs. simple consumption of liquid food from the Richter tubes. Mean meal sizes of all meals taken with fistulas closed were insignificantly reduced when animals drank from the tubes--4.9 compared to 5.7 ml in the operant situatiorv--while average meal frequency was identical in the two situations. Therefore, these data were combined with those from the Richter tube study. DISCUSSION We observed no significant indications of excessive eating while sham feeding in obese subjects in the present study. Under the same paradigm, control-weight subjects showed marked increases, comparable to that reported by others in the literature ([13] most recently). We had hypothesized that the increased feeding when animals are sham fed might be even more enhanced in fat animals, as raised insulin levels might clear the plasma of nutrients more quickly into adipocytes and other insulin-dependent tissue and lead the animal to a hypoglycemic condition. We have shown earlier that animals induced to become obese and then returned to eating only Purina chow did show elevated insulin levels which remained frankly hyperinsulinemic one week after termination of the varied palatable diets [21]. Even though sham feeding does not allow nutrient entry into the gastrointestinal tract, cephalic insulin would stimulate uptake of any circulating nutrient and that achieved from glycogenolysis and/or gluconeogenesis (as suggested by the ecork of Steffens
TABLE 2 SATIETYRATIOSOF 6 CONTROLAND 6 OBESE SUBJECTS Mean Body Weight (g)
Initial Satiety Ratio
Mean Satiety Ratio
Control Fistula Closed Fistula Open Open-Closed
293.8 ± 2.12 306.0 ± 2.09
0.18 ± 0.02 0.I0 ± 0.01 -44%*
0.52 ± 0.04 0.26 ± 0.02 -50%t
Obese Fistula Closed Fistula Open Open-Closed
385.6 ± 2.71 400.9 ± 2.73
0.58 ± 0 . 1 0 0.77 _+ 0.19 +33c~
1 . 2 0 ± 0.31 0.94 ± 0.21 -22%
*p<0.05. tp <0.02.
[20]). Evidently hyperinsulinemia associated with obesity does not enhance sham feeding or, more correctly, the anorexia noted with obesity induced by offering rats palatable diets and then removing the diets, is not dependent upon stimuli derived from the interaction of diet with the gastrointestinal tract and the absorption of that diet. This latter point might need some clarification. If the anorexia of fattened animals was dependent upon gut interaction or absorption of nutrients, hypophagia would no longer be evident and increased eating during sham feeding would be as robust or more so in obese animals than in normal-weight subjects. Fat animals remained hypophagic, however, when sham fed. When rats are induced to become obese, by such methods as lateral hypothalamic stimulation, insulin injection, palatable diet feeding, or force-feeding by gastric tubes, and then the obesity-inducing stimulation is halted, the obese animal is invariably hypophagic or anorexic compared to normal-weight control subjects. We anticipated that this anorexia might be attributed to the raised gastrointestinal secretory activity associated with overfeeding, absorption and weight gain. It appears that some other signal independent of gastric fill or duodenal distention and ab-
53t~
VANI)ERWEEI.[-.' A N ! ) VAN I'I~1 [1~!
sorption p r o m o t e s this relative hypophagia, at least in the dietary-obese animals of the present study. The present study also confirms earlier studies that indicate intermeal satiety as immediately affected when animals sham feed. The intermeal interval was reduced and meal f r e q u e n c y significantly increased e v e n in a 4-hr interval. The initial satiety ratio was reduced by 44% in control animals and was unaffected in obese animals. Indeed, this is the effect we o b s e r v e d in our earlier study on meal patterns in obese rats [23]; obese animals ate smaller meals with greater satiety ratios than did controls. W h e n animals are obese, less food appears to satiate longer and hence, intra- and intermeal satiety are heightened. The intermeal satiety effect (satiety ratio) persisted when obese animals were sham fed while the control-weight subjects showed an immediate decrease in this effect. Intermeal satiety, we suggest, may be more dependent upon gut/nutrient interaction and absorption than is intrameal. B r o b e c k has suggested that nutrient clearance might be an important factor in determining appetite and ingestion of food as a function o f body weight. In obese subjects, already enlarged adipocytes and repleted glycogen stores might retard nutrient clearance and reduce food ingestion while, in the lean animal, the r e v e r s e situation (depleted nutrient stores) would lead to an increased rate o f ingestion. Indeed, many studies (e.g., [3, 6, 10, 15]) have shown animals to be more efficient in storing and utilizing nutrients during realimentation after e n f o r c e d fasts. There was no nutrient absorption from ingestion in our fattened, sham fed animals yet, despite this fact, hypophagia reamined. Unless glycogenolysis following ingestion occurs and is delayed in clearance, it seems unlikely that B r o b e c k ' s hypothesis [4] can account for our present results. Perhaps the simplest explanation for our results is that fat
animals are anorexic because of a signal which is m no w a sensitive to gastrointestinal events. The animal, thus. continues to show anorexia e v e n when sham t\~eding That the fistulas may have produced greater discomfor~ in lhc ~bcsc animals and lowered feeding is unlikely ~ testllls ill the fistula-closed condition arc very similar 1o those obtained before in non-fistulated animals [21]. Also, one cannot s i r e ply argue that the obese subjects in the present study did not like the diet offered while normal-weight subjects did: both groups had been feeding and maintaining weight differences on the diet for one w e e k before sham feeding was ussessed. Finally, we offer one further explanation thai may account for the absence of significant increased ingestion while sham feeding in obese rats. Lorenz, Gibbs and Smith 1121 have shown that atropine and CCK also attenuate sham feeding. One might ask what factors these various stimuli possesx in c o m m o n . Methylated atropine remains outside the blood-brain-barrier and, hence, it is presumed that its major effect is to inhibit cholinergic muscarinic receptors in the periphery. Such pharmacological action would decrease stomach emptying II9]. C C K has been shown to retard stomach emptying [17.25] and, perhaps in high dosages, inhibit gastric motility. Could it be that slowing the rate of stomach emptying renders the stimulus of gastric events to satiety noninformative? Thus. the animal might shift to other cues to regulate food intake. Indeed, the d e v e l o p m e n t of obesity may lead to a stomach which processes fl~od more slowly and presents food to the d u o d e n u m at a slower rate. If the organism, prior to becoming obese, was regulating food ingestion primarily on the basis of stomach/duodenal e m p t y ing and absorption interactions, with this signal now blunted, subjects might turn to postabsorptivc or more longterm regulatory mechanisms. We are now testing this hypothesis.
REFERENCES 1. Anika, S. M., T. R. Houpt and K. A. Houpt. Insulin as a satiety hormone. Physiol Behav 25: 21-23, 1980. 2. Berthoud, H. R. and B. Jeanrenaud. Sham feeding-induced cephalic phase insulin release in the rat. Am J Physiol 242: E280-E285, 1982. 3. Boyle, P. C., L. H. Storlien and R. E. Keesey. Increased efficiency of food utilization following weight loss. Physiol Behav 21: 261-264, 1978. 4. Brobeck, J. R. Nature of satiety signals. Am J Clin Nutr 28: 806-807, 1975. 5. Davis, J. D. and C. Campbell. Peripheral control of meal size in the rat: Effect of sham feeding on meal size and drinking rate. J Comp Physiol Psychol 83: 37%387, 1973. 6. Fabry, P. Metabolic consequences of the patterns of food intake. In: Handbook of Physiology, Section 6, Alimentary Canal, vol l, edited by C. F. Code. Washington, DC: Am Physiol Society, 1967, pp. 31-49. 7. Janowitz, H. D. and M. I. Grossman. Some factors that affect the food intake of normal dogs and of dogs with esophagostomy and gastric fistula. Am J Physiol 159: 143-148, 1949. 8. Kraly, F. S., W. J. Carry, S. Resnick and G. P. Smith. Effect of cholecystokinin on meal size and intermeal interval in the sham-feeding rat. J Comp Physiol Psychol 92: 697-707, 1978. 9. Kraly, F. S., W. J. Carty and G. P. Smith. Effect of pregastric food stimuli on meal size and intermeal interval in the rat. Physiol Behav 20" 77%784, 1978. 10. Levitsky, D. A., I. Faust and M. Glassman. The ingestion of food and the recovery of body weight following fasting in the naive rat. Physiol Behav 17: 575-580, 1976.
I I. Liebling, D. S., J. D. Eisner, J. Gibbs and G. P. Smith. Intestinal satiety in rats. J Comp Physiol Psyehol 89: 955-965. 1975. 12. Lorenz, D., J. Gibbs and G. P. Smith. Effect of cholecystokinin and other gut hormones on sham feeding in the rat. In: G,t Hormones. edited by S. R. Bloom. London: Churchill Livingstone, 1978, pp. 224-226. 13. Lorenz, D., P. Nardi and G. P. Smith. Atropine methyl nitrate inhibits sham feeding in the rat. Pharmacol Biochem Behav 8: 405-407, 1978. 14. Martin, C. F. and J. Gibbs. Bombesin elicits satiety in sham feeding rats. Peptides 1: 131-134, 1980. 15. Meyer, J. H. and W. J. Clawson. Undernutrition and subsequent realimentation in rats and sleep. J Anita Sci 23:214-224. 1964. 16. Mook, D. G. Oral and postingestional determinants of the intake of various solutions in rats with esophageal fistulas..I Comp Physiol Psychol 56: 645-659, 1963. 17. Moran, T. H. and P. R. McHugh. Cholecystokinin supresses food intake by inhibiting gastric emptying. Am .I Physiol. in press, 1982. 18. Sclafani, A. and D. Springer. Dietary obesity in adult rats: Similarities to hypothalamic and human obesity syndromes, Physiol Behav 17: 461-471, 1976. 19. Setler. P. E. and G. P. Smith. Gastric emptying in rats with chronic gastric fistulas. Am J Dig Dis (N.S.) 14: 137-142, 1969. 20. Strubbe, J. H. and A. B. Steffens. Rapid insulin release after ingestion of a meal in the unanesthetized rat. A m J Phsyiol 229: 101%1022, 1975.
SHAM FEEDING
IN OBESE RATS
21. VanderWeele, D. A., E. Haraczkiewiz and T. B. Van Itallie. Elevated insulin and satiety in obese and normal-weight rats. Appetite 3: 99-109, 1982. 22. VanderWeele, D. A., F. X. Pi-Sunyer. D. Novin and M. J. Bush. Chronic insulin infusion suppresses food ingestion and body weight gain in rats. Brain Res Bull 5: Suppl 4, 7-11, 1980. 23. Van ltallie. T. B. and D. A. VanderWeele. The phenomenon of satiety. In: Recent Advances in Obesity Research: HI, edited by P. Borntorp, M. Cairella and A. N. Howard. London: John Libbey Press, 1981, pp. 278-289.
537
24. Woods, S. C., L. D. McKay, L. J, Stein, D. B. West, E. C. Lotter and D. Porte, Jr. Neuroendocrine regulation of food intake and body weight. Brain Res Bull 5: Suppl 4, 1-7, 1980. 25. Yamagishi, T. and H. T. Debas. Cholecystokinin inhibits gastric emptying by acting on both proximal stomach and pylorus. Am J Physiol 234: E375-E378, 1978. 26. Young, R. C., J. Gibbs, J. Antin, J. Holt and G. P. Smith. Absence of satiety during sham feeding in the rat. J Comp Physiol Psyehol 87: 795-800, 1974.