- Email: [email protected]
Learning of gastrointestinal satiety signals
BEHAVIORAL AND NEURAL BIOLOGY 4 5 , 2 9 2 - - 2 9 9
(1986)
Learning of Gastrointestinal Satiety Signals J. A . DEUTSCH AND J. A . TABUENA l
Department of Psychology, University of California, San Diego, La Jolla, California 92093 Rats were trained to ingest corn oil emulsion. Such ingestion was accompanied by intragastric injection either of dilute amino acid (e.g., 4.3 mM L-tryptophan) in one group or saline in a second. No significant difference between the two groups was observed at the end of training. However sudden omission of the amino acid led to a large increase in intake, whereas sudden introduction of the amino acid produced no significant change. It seems that signals for the amount of nutrient ingested can be arbitrarily associated with different nutrients and that the value of such satiety signals is not innately fixed in the amount of feeding inhibition exerted. © 1986 Academic Press, Inc.
When food is removed from the stomach during a familiar meal, the amount eaten increases in compensation (Davis & Campbell, 1973; Young, Gibbs, Antin, Holt, & Smith, 1974; Deutsch, Young, & Kalogeris, 1978; Deutsch & Gonzalez, 1980, 1981). This shows that signals are generated in the upper gastrointestinal tract which are utilized in the regulation of meal size. Some of these signals are generated by distention but these limit the size of the meal only when a rather high level of intake has been exceeded (Deutsch, Gonzalez, & Young, 1980, Deutsch & Gonzalez, 1980, 1981). Vagotomy abolishes such signals of bulk (Gonzalez & Deutsch, 1981). However, compensation for amount of nutrient removed from the stomach remains unaffected by vagotomy, but is abolished by splanchnicotomy (Deutsch & Ahn, 1986). What generates the splanchnic signals is some product of the early digestion of nutrient. When rats ingest oil emulsion, direct intragastric infusion of the same emulsion does not reduce the amount ingested by mouth, showing that distention is not a factor but digestive breakdown is. On the other hand, the intragastric infusion of the emulsion removed from the stomach of a donor rat reduces oral ingestion by the amount that is placed in the stomach (Gonzalez & Deutsch, 1985). Our thanks are due to Dr. J. O. Miller for statistical advice. We also thank Best Foods for supplying the Mazola corn oil. Supported by NIH Grant RR08135-10. Correspondence and reprint requests should be addressed to J. A. Deutsch. 292 0163-1047/86 $3.00 Copyright © 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.
LEARNING OF SATIETY SIGNALS
293
It seems likely then that some digestive breakdown products of the ingested nutrient are used as indicators or markers for the amount of nutrient eaten and so function as satiety signals. Is the value of such satiety signals innately fixed in the amount of feeding inhibition exerted? And is each signal confined to a particular nutrient or can it arbitrarily be associated with different nutrients through a process of learning? To answer these questions, rats were trained to ingest a 50:50 water:corn oil emulsion, a nutrient unfamiliar to them. Whenever they drank this emulsion, amino acid solutions in calorically insignificant amounts were injected directly into the stomach at 20% the rate of oral intake of the oil emulsion. Amino acids were chosen because they are digestive breakdown products of proteins and might thus reasonably be expected to act on chemoreceptors concerned with food intake regulation. The intragastrically injected amino acids if they acted as signals would be added to the signals evoked by products of early digestion of oil (free fatty acids, glycerol). If the value of these amino acids is innately fixed in the amount of satiety produced then their injection into the stomach should produce a reduction of the amount of oil ingested under the two following conditions. First, a group of rats drinking oil emulsion with intragastric amino acids always injected should ingest less than another group drinking the same oil emulsion with the same volume of saline injected instead of the amino acid. Second, the group drinking the oil emulsion with the saline intragastrically injected should reduce its intake of oil emulsion when the amino acid is first injected after training when the saline intragastric injection has occurred. On the other hand, the satiety value of these amino acids may be acquired through a process of learning. Signals from the upper gastrointestinal tract at the end of a meal may be stored and then calibrated by being compared to the metabolic postabsorptive benefit from a meal. If such a learning hypothesis is correct then the size of a meal will depend on the learned value of the satiety signals (their previous calibration against postabsorptive benefit) in the upper gastrointestinal tract rather than their absolute amount. The consequences of this on the two conditions described above would be as follows. In the first condition a group of rats drinking oil emulsion with intragastric amino acid always injected should drink the same amount as another group drinking the same oil emulsion with the same amount of saline injected instead. If the amino acid injected group drank less, the metabolic benefit would be smaller and the value of the satiety signals at the end of the meal would be recalibrated and so made smaller. Consequently, the rat would drink more at the next meal in response to the same number of signals. (On this hypothesis the rat learns to predict the metabolic benefit of a meal from signals in the upper gastrointestinal tract from past experience.) In the second condition, the group drinking the oil emulsion with saline
294
DEUTSCH AND TABUENA
intragastricaUy injected should not reduce its intake of oil when the amino acid is first injected. On the learning hypothesis there is no innate satiety value of such a signal and thus the signal has to be calibrated by correlation with postingestional events before it exerts control over ingestion. In spite of the prediction that intragastrically injected amino acid should have no effect on the amount of oil ingested in the above two conditions, there is a third condition in which a strong effect would be expected. A group of rats drinking oil emulsion with intragastric amino acid always injected should ingest more oil emulsion when the number of upper gastrointestinal signals is suddenly reduced by the omission of the amino acid injection, thus leaving only the signals from free fatty acids (and possibly glycerol). To test these predictions four separate experiments were conducted. The first tested the idea that the pairing of amino acids intragastricaUy as the rat learned to drink oil would have no effect on intake until the amino acids were suddenly omitted. In this experiment the control group was intragastricaUy injected with saline even on the test day. In the last three experiments, the experimental group of rats learned to drink oil while various amino acids were intragastrically injected, whereas the control group learned to drink oil while saline was injected instead. On the test day the treatments of the two groups were reversed, the experimental group received intragastric saline and the control group received intragastric amino acid. If the amino acid had been utilized as a satiety signal, its omission should produce an increase in intake in the experimental group. If the amino acid functioned as a satiety signal through learning, then no equivalent reduction of intake should occur in the control group when the amino acid was first introduced. If on the other hand the amino acid was a fixed or unlearned satiety signal a reduction of intake (equal to the increase seen in the first group) should occur. METHOD In all four experiments the general methods were the same. The rats were individually housed in stainless-steel wire cages and received solid rat chow (Purina) and water ad lib before the experiment. They were placed on a 24-h light cycle with light coming on at 8 AM and switching off at 8 eM. The rats were implanted with chronic gastric catheters constructed from Silastic tubing according to the method described by Young and Deutsch (1981). The catheter exits via a subcutaneous tunnel through the dorsal region of the neck. The Silastic tube connects to a stainlesssteel L-shaped cannula which exits the neck at a 90° angle. Sodium pentobarbital injected at 43.2 mg/kg served as the anesthetic and topazone was applied topically to both the muscles tissue and the neck wound.
LEARNING OF SATIETY SIGNALS
295
After surgery the rats were allowed 1 week for recovery. The catheters were flushed daily with 0.5 ml saline to remove solid food obstruction. After recovery the rats began to receive ad libitum access to rat chow for 1 h each day. Fifteen hours later they were placed in restraining cages for 30 min and fed a 50:50 water:corn oil emulsion. This emulsion is prepared in a blender by mixing equal volumes of Mazola corn oil and water and by adding 1% lecithin to stabilize. The rats have access to the emulsion from one end of a U-tube--the other end has electrodes touching the surface. The electrodes turn on a motor as soon as contact with the surface of the emulsion is lost. The motor in turn depresses the plunger on two syringes simultaneously. One syringe refills the U-tube with the corn oil emulsion, the other infuses various substances into the stomach. Thus the animal receives a substance in the stomach at some fraction of the rate the rat drinks. EXPERIMENT 1
Subjects. These were 18 male Sprague-Dawley rats, 350-450 g in weight.
Treatment. The rats were trained to drink water when 23 h thirsty in the apparatus described above, for 7 days. After surgery and the recovery period they were then divided into two groups of nine. The first group (experimental) received amino acid intragastrically, 20% of the rate of the corn oil emulsion drunk by mouth. The solution consisted of 4% L-lysine (219 raM) 4% L-threonine (336 raM), and 1% tryptophan (49 mM). These proportions of amino acid when calculated as percentages of the orally ingested oil are about double the minimum that the rat needs of each (Rogers & Harper, 1965) in its food to support normal growth (lysine 1%, threonine 0.6%, and tryptophan 0.2%). In contrast we administered 1.6% lysine, 1.6% threonine, and 0.4% tryptophan. The amino acids would supply about 1% of the calories of the oil. The second group (control) was intragastrically injected with physiological saline, also at 20% the volume drunk by mouth. On the test day (after 14 days of training of both groups) physiological saline was substituted for the amino acid solution in group 1. Results. There was no significant difference (t = 1.94, n.s.) between the two groups over the four days previous to the test. The rats receiving amino acid drank a mean of 7.47 ml (SEM _+ 0.47) and those receiving saline 8.82 ml (SEM __- 0.57). While the second group (control) showed no significant change on the day that the first group (experimental) had its intragastric injection changed to saline, the first group showed a 56.1% mean increase (SEM _+ 11.2) in consumption over the previous 4 days (t = 5, p > 0.002).
296
DEUTSCH AND TABUENA
EXPERIMENT 2 The second experiment was identical to the first except that L-tryptophan was omitted from the amino acid solution and conditions were switched on the test day between the two groups of new rats. In other words, on the test day the experimental group that had been intragastrically injected with amino acid was now injected with saline, and the control group that had been injected with saline was now injected with amino acid. Subjects. Twelve new rats, as in Experiment 1. Treatment. The same as in Experiment 1 except that L-tryptophan was omitted, and that the second group (control) was injected with amino acid on the test day. Results. Again, there was no significant difference (t = 0.52, n.s.) in the intake of the two groups over the 4 days before the text day. The group infused with amino acid drank a mean of 9.8 ml (SEM ___ 0.6) whereas the group receiving saline drank a mean of 10.2 ml (SEM +__ 0.65). In the first group (experimental), the sudden omission of the amino acid produced an increase (26.7%, t = 3.28, p < .05). However the sudden introduction of amino acid produced no significant change in intake (+0.35%, t = 0.032, n.s.) in the second (control) group. EXPERIMENT 3 AND 4 Subjects. Fourteen new rats in Experiment 3, 12 new rats in Experiment 4.
Treatment. The same as in the previous two experiments, with the following exceptions. In the last two experiments, single amino acids were tested, and the control saline solution was adjusted to be equimolar to the amino acid solution. Further, the intake (after an initial 7 days of training) of each rat was monitored individually. When 4 consecutive days of intake, with no change of greater than 20% over the previous day was observed, that rat's condition was switched either from amino acid to saline or from saline to amino acid. This procedure continued until every rat was switched between the amino acid and saline conditions. In Experiment 3, seven rats were injected intragastrically with 15 mM L-tryptophan at 20% the rate of their oral intake, thus resulting in 4.3 mM of L-tryptophan in the stomach. Another seven rats were injected intragastrically with 15 mM NaCL at the same rate. Experiment 4 was exactly like Experiment 3 except that 150 mM Llysine was tested, and that there were six new rats in each of the two groups. Results: In Experiment 3 there was no significant difference in intake between the amino acid and saline group on the 4 days before test. The
LEARNING OF SATIETY SIGNALS
297
L-tryptophan injected group drank a mean of 11.0 ml (SEM - 1.1) and the saline injected group drank a mean of 9.9 ml (SEM _+ 0.8). When the rats receiving the L-tryptophan were switched to equimolar saline, mean oral intake increased by 25.5% (SEM _+ 4.4). This increase is highly significant (t = 5.8, p > 0.002). However when the rats receiving saline were suddenly injected with L-tryptophan, there was a small but nonsignificant increase in intake (13.1%, SEM + 6.8, t = 1.9, n.s.). In Experiment 4, there was no significant difference in intake between the amino acid and saline group on the 4 days before the conditions were switched. The L-lysine injected group drank a mean of 10.5 ml (SEM + 1.6) and the saline injected group drank a mean of 12.2 ml (SEM ___ 1.9). When the rats receiving the L-lysine were switched to equimolar (150 raM) saline their oral intake increased by 37.3% (SEM __- 6.0). Such an increase is highly significant (t = 6.2, p > 0.002). However when the rats receiving saline were switched to equimolar L-lysine, there was a nonsignificant decrease in intake (13.7% SEM _ 5.4, t = 2.5, n.s.).
DISCUSSION In all four experiments described above the omission of intragastrically injected amino acid that had up to then been paired with the ingestion of a triglyceride meal led to an immediate increase in the amount of tryglyceride consumed. One interpretation of such an increase is that the injection of intragastric amino acid is aversive, and that the removal of such an aversive stimulus produces an immediate increase in ingestion. However, there are two separate controls in this experiment that both contradict such an interpretation. If the removal of the amino acid was the removal of an aversive stimulus, then its presence or sudden introduction should reduce ingestion. However, there was no significant difference on the last 4 days of training in oil intake in any of the four experiments between the experimental group receiving the amino acids and the control group receiving saline nor was any difference apparent before. Further, there was no significant decrease in oil intake in all the experiments where the controls were subjected to an intragastric injection of amino acid on the test day. This control constitutes a very direct test of the aversiveness of the intragastric injection of amino acid. It could be argued that the aversiveness of the amino acid might not be detected the first time it was injected because it took some time to build up as a result of postabsorptive factors. But this interpretation is contradicted by the fact that increased ingestion occurs sharply as soon as the amino acid is omitted and by the fact that there is no difference between the experimentals and controls before the test day. Related to the issue of aversiveness is the point that the amino acids used in the experiment might have produced some amino acid imbalance in the rats that might in some undetermined way have produced the
298
DEUTSCH AND TABUENA
results reported above. Other work in this laboratory has shown large and significant effects using the same paradigm as above, with a free fatty acid (35% increase when 5% intragastric oleic acid is omitted) (Deutsch & Gonzalez, unpublished) and maltose (47% when 5% intragastric maltose is omitted) (Deutsch & Moore, unpublished). In any case it appears unlikely that the percentage of amino acids in the stomach when oil emulsion was injected would produce amino acid imbalance. For instance in Experiment 3, the L-tryptophan constitutes 0.12% of the oil intragastrically present. However "the rat tolerates quite well 1% or slightly more of either the D or L form" (Harper, Benevenga, & Wohlhueter, 1970) of tryptophan. With regard to L-threonine, 4% of threonine in the total diet did not depress the growth rate of rats (Wretlind, 1949) whereas 1.6% was administered in Experiment 1. In the case of L-lysine, growth rate is retarded after lysine is made to comprise about 5% of the diet (Russell, Taylor, & Hogan, 1952). In contrast we administered 1.6% lysine in combination with the oil in the first experiment and about 1.2% in the fourth. The measurements on the effects of amino acid imbalance were made on immature rats, still in a stage of rapid growth on a diet which was marginal in protein content. Any imbalance is much better tolerated by mature animals fed a well-balanced diet (Harper et al., 1970) such as were used in our experiments. The sudden increase in intake when amino acids are omitted from the intragastric injection seems more plausibly interpreted by the hypothesis that they function as satiety signals. The lack of a corresponding decrease in intake when these amino acids in the same proportion are suddenly introduced shows that learning has to take place before they function as satiety signals. Because the amount ingested is the same before conditions are reversed, the process of learning to use a particular signal must also involve a process of calibration. For the signals generated by the breakdown products of triglyceride lead to the same intake as the identical signals generated by the breakdown products of triglyceride plus the extra signals generated by the amino acids. While there are more signals in the latter case their effectiveness in inhibiting intake becomes the same presumably because of some process of calibration based on the postabsorbtive consequences of a meal. It appears from our results that the value of satiety signals is not innately fixed in the amount of feeding inhibition exerted and that satiety signals can be arbitrarily associated with different nutrients through a process of learning. REFERENCES Antin, J., Gibbs, J., Holt, J., Young, R. C., & Smith, G. P. (1975). Cholecystokinin elicits the complete behavioral sequence of satiety. Journal of Comparative and Physiological Psychology, 89, 784-790. Davis, J. D., & Campbell, C. S. (1973). Peripheral control of meal size in the rat" Effect of sham feeding on meal size and drinking rate. Journal of Comparative and Physiological Psychology, 83, 379-387.
LEARNING OF SATIETY SIGNALS
299
Deutsch, J. A. (1983). Dietary control and the stomach. Progress in Neurobiology, 20, 313-332. Deutsch, J. A., & Gonzalez, M. F. (1980). Gastric nutrient content signals satiety. Behaviorol and Neural Biology, 30, 113-116. Deutsch, J. A., & Gonzalez, M. F. (1981). Gastric fat content and satiety. Physiological Behavior, 26, 673-676. Deutsch, J. A. Gonzalez, M. F., & Young, W. G. (1980). Two factors control meal size. Brain Research Bulletin, 5 (Suppl. 4), 55-57. Deutsch, J. A. Young, W. G., & Kalogeris, T. J. (1978). The stomach signals satiety. Science, 201, 165-167. Deutsch, J. A., & Ahn, S. J. (1986). The splanchnic nerve and food intake regulation. Behavioral Biology, 45, 43-47. Deutsch, J. A., & Moore, B. O. Unpublished. Deutsch, J. A., & Gonzalez, M. F. Unpublished. Gonzalez, M. F., & Deutsch, J. A. (1985). Intragastric Injections of partially digested triglyceride suppress feeding in the rat. Physiology and Behavior, 31, 861-865. Gonzalez, M. F., & Deutsch, J. A. (1981). Vagotomy abolishes cues of satiety produced by gastric distention. Science, 212, 1283-1284. Harper, A. E., Benevenga, N. J., & Wohlhueter, R. M. (1970). Effects of Ingestion of Disproportionate Amounts of Amino Acids. Physiological Reviews, 50, 428-558. Rogers, Q. R., & Harper, A. E. (1965). Amino acid diets and maximal growth in the rat. Journal of Nutrition, 87, 267-272. Russell, W. C., Taylor, M. W., & Hogan, J. M. (1952). Effects of excess essential amino acids on growth of the white rats. Arch. Brocheur. Biophys. 39, 249-253. Wretlind, K. A. J. (1949). The effect of synthetic acids essential for growth on the bodyweight of growing rats, and the synthesis of the amino acids used. Acta Physiologica Scandinavica 17 (Suppl. 59), 1-101. Young, W. G., & Deutsch, J. A. (1981). The construction, surgical Implantation, and use of Gastric Catheters, and a Pyloric Cuff. Journal of Neuroscience Methods, 3, 377384. Young, R. C., Gibbs, J., Antin, J., Holt, J., & Smith, G. P. (1974). Absence of satiety during sham feeding in the rat. Journal of Comparative and Physiological Psychology, 87, 795-800.
(1986)
Learning of Gastrointestinal Satiety Signals J. A . DEUTSCH AND J. A . TABUENA l
Department of Psychology, University of California, San Diego, La Jolla, California 92093 Rats were trained to ingest corn oil emulsion. Such ingestion was accompanied by intragastric injection either of dilute amino acid (e.g., 4.3 mM L-tryptophan) in one group or saline in a second. No significant difference between the two groups was observed at the end of training. However sudden omission of the amino acid led to a large increase in intake, whereas sudden introduction of the amino acid produced no significant change. It seems that signals for the amount of nutrient ingested can be arbitrarily associated with different nutrients and that the value of such satiety signals is not innately fixed in the amount of feeding inhibition exerted. © 1986 Academic Press, Inc.
When food is removed from the stomach during a familiar meal, the amount eaten increases in compensation (Davis & Campbell, 1973; Young, Gibbs, Antin, Holt, & Smith, 1974; Deutsch, Young, & Kalogeris, 1978; Deutsch & Gonzalez, 1980, 1981). This shows that signals are generated in the upper gastrointestinal tract which are utilized in the regulation of meal size. Some of these signals are generated by distention but these limit the size of the meal only when a rather high level of intake has been exceeded (Deutsch, Gonzalez, & Young, 1980, Deutsch & Gonzalez, 1980, 1981). Vagotomy abolishes such signals of bulk (Gonzalez & Deutsch, 1981). However, compensation for amount of nutrient removed from the stomach remains unaffected by vagotomy, but is abolished by splanchnicotomy (Deutsch & Ahn, 1986). What generates the splanchnic signals is some product of the early digestion of nutrient. When rats ingest oil emulsion, direct intragastric infusion of the same emulsion does not reduce the amount ingested by mouth, showing that distention is not a factor but digestive breakdown is. On the other hand, the intragastric infusion of the emulsion removed from the stomach of a donor rat reduces oral ingestion by the amount that is placed in the stomach (Gonzalez & Deutsch, 1985). Our thanks are due to Dr. J. O. Miller for statistical advice. We also thank Best Foods for supplying the Mazola corn oil. Supported by NIH Grant RR08135-10. Correspondence and reprint requests should be addressed to J. A. Deutsch. 292 0163-1047/86 $3.00 Copyright © 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.
LEARNING OF SATIETY SIGNALS
293
It seems likely then that some digestive breakdown products of the ingested nutrient are used as indicators or markers for the amount of nutrient eaten and so function as satiety signals. Is the value of such satiety signals innately fixed in the amount of feeding inhibition exerted? And is each signal confined to a particular nutrient or can it arbitrarily be associated with different nutrients through a process of learning? To answer these questions, rats were trained to ingest a 50:50 water:corn oil emulsion, a nutrient unfamiliar to them. Whenever they drank this emulsion, amino acid solutions in calorically insignificant amounts were injected directly into the stomach at 20% the rate of oral intake of the oil emulsion. Amino acids were chosen because they are digestive breakdown products of proteins and might thus reasonably be expected to act on chemoreceptors concerned with food intake regulation. The intragastrically injected amino acids if they acted as signals would be added to the signals evoked by products of early digestion of oil (free fatty acids, glycerol). If the value of these amino acids is innately fixed in the amount of satiety produced then their injection into the stomach should produce a reduction of the amount of oil ingested under the two following conditions. First, a group of rats drinking oil emulsion with intragastric amino acids always injected should ingest less than another group drinking the same oil emulsion with the same volume of saline injected instead of the amino acid. Second, the group drinking the oil emulsion with the saline intragastrically injected should reduce its intake of oil emulsion when the amino acid is first injected after training when the saline intragastric injection has occurred. On the other hand, the satiety value of these amino acids may be acquired through a process of learning. Signals from the upper gastrointestinal tract at the end of a meal may be stored and then calibrated by being compared to the metabolic postabsorptive benefit from a meal. If such a learning hypothesis is correct then the size of a meal will depend on the learned value of the satiety signals (their previous calibration against postabsorptive benefit) in the upper gastrointestinal tract rather than their absolute amount. The consequences of this on the two conditions described above would be as follows. In the first condition a group of rats drinking oil emulsion with intragastric amino acid always injected should drink the same amount as another group drinking the same oil emulsion with the same amount of saline injected instead. If the amino acid injected group drank less, the metabolic benefit would be smaller and the value of the satiety signals at the end of the meal would be recalibrated and so made smaller. Consequently, the rat would drink more at the next meal in response to the same number of signals. (On this hypothesis the rat learns to predict the metabolic benefit of a meal from signals in the upper gastrointestinal tract from past experience.) In the second condition, the group drinking the oil emulsion with saline
294
DEUTSCH AND TABUENA
intragastricaUy injected should not reduce its intake of oil when the amino acid is first injected. On the learning hypothesis there is no innate satiety value of such a signal and thus the signal has to be calibrated by correlation with postingestional events before it exerts control over ingestion. In spite of the prediction that intragastrically injected amino acid should have no effect on the amount of oil ingested in the above two conditions, there is a third condition in which a strong effect would be expected. A group of rats drinking oil emulsion with intragastric amino acid always injected should ingest more oil emulsion when the number of upper gastrointestinal signals is suddenly reduced by the omission of the amino acid injection, thus leaving only the signals from free fatty acids (and possibly glycerol). To test these predictions four separate experiments were conducted. The first tested the idea that the pairing of amino acids intragastricaUy as the rat learned to drink oil would have no effect on intake until the amino acids were suddenly omitted. In this experiment the control group was intragastricaUy injected with saline even on the test day. In the last three experiments, the experimental group of rats learned to drink oil while various amino acids were intragastrically injected, whereas the control group learned to drink oil while saline was injected instead. On the test day the treatments of the two groups were reversed, the experimental group received intragastric saline and the control group received intragastric amino acid. If the amino acid had been utilized as a satiety signal, its omission should produce an increase in intake in the experimental group. If the amino acid functioned as a satiety signal through learning, then no equivalent reduction of intake should occur in the control group when the amino acid was first introduced. If on the other hand the amino acid was a fixed or unlearned satiety signal a reduction of intake (equal to the increase seen in the first group) should occur. METHOD In all four experiments the general methods were the same. The rats were individually housed in stainless-steel wire cages and received solid rat chow (Purina) and water ad lib before the experiment. They were placed on a 24-h light cycle with light coming on at 8 AM and switching off at 8 eM. The rats were implanted with chronic gastric catheters constructed from Silastic tubing according to the method described by Young and Deutsch (1981). The catheter exits via a subcutaneous tunnel through the dorsal region of the neck. The Silastic tube connects to a stainlesssteel L-shaped cannula which exits the neck at a 90° angle. Sodium pentobarbital injected at 43.2 mg/kg served as the anesthetic and topazone was applied topically to both the muscles tissue and the neck wound.
LEARNING OF SATIETY SIGNALS
295
After surgery the rats were allowed 1 week for recovery. The catheters were flushed daily with 0.5 ml saline to remove solid food obstruction. After recovery the rats began to receive ad libitum access to rat chow for 1 h each day. Fifteen hours later they were placed in restraining cages for 30 min and fed a 50:50 water:corn oil emulsion. This emulsion is prepared in a blender by mixing equal volumes of Mazola corn oil and water and by adding 1% lecithin to stabilize. The rats have access to the emulsion from one end of a U-tube--the other end has electrodes touching the surface. The electrodes turn on a motor as soon as contact with the surface of the emulsion is lost. The motor in turn depresses the plunger on two syringes simultaneously. One syringe refills the U-tube with the corn oil emulsion, the other infuses various substances into the stomach. Thus the animal receives a substance in the stomach at some fraction of the rate the rat drinks. EXPERIMENT 1
Subjects. These were 18 male Sprague-Dawley rats, 350-450 g in weight.
Treatment. The rats were trained to drink water when 23 h thirsty in the apparatus described above, for 7 days. After surgery and the recovery period they were then divided into two groups of nine. The first group (experimental) received amino acid intragastrically, 20% of the rate of the corn oil emulsion drunk by mouth. The solution consisted of 4% L-lysine (219 raM) 4% L-threonine (336 raM), and 1% tryptophan (49 mM). These proportions of amino acid when calculated as percentages of the orally ingested oil are about double the minimum that the rat needs of each (Rogers & Harper, 1965) in its food to support normal growth (lysine 1%, threonine 0.6%, and tryptophan 0.2%). In contrast we administered 1.6% lysine, 1.6% threonine, and 0.4% tryptophan. The amino acids would supply about 1% of the calories of the oil. The second group (control) was intragastrically injected with physiological saline, also at 20% the volume drunk by mouth. On the test day (after 14 days of training of both groups) physiological saline was substituted for the amino acid solution in group 1. Results. There was no significant difference (t = 1.94, n.s.) between the two groups over the four days previous to the test. The rats receiving amino acid drank a mean of 7.47 ml (SEM _+ 0.47) and those receiving saline 8.82 ml (SEM __- 0.57). While the second group (control) showed no significant change on the day that the first group (experimental) had its intragastric injection changed to saline, the first group showed a 56.1% mean increase (SEM _+ 11.2) in consumption over the previous 4 days (t = 5, p > 0.002).
296
DEUTSCH AND TABUENA
EXPERIMENT 2 The second experiment was identical to the first except that L-tryptophan was omitted from the amino acid solution and conditions were switched on the test day between the two groups of new rats. In other words, on the test day the experimental group that had been intragastrically injected with amino acid was now injected with saline, and the control group that had been injected with saline was now injected with amino acid. Subjects. Twelve new rats, as in Experiment 1. Treatment. The same as in Experiment 1 except that L-tryptophan was omitted, and that the second group (control) was injected with amino acid on the test day. Results. Again, there was no significant difference (t = 0.52, n.s.) in the intake of the two groups over the 4 days before the text day. The group infused with amino acid drank a mean of 9.8 ml (SEM ___ 0.6) whereas the group receiving saline drank a mean of 10.2 ml (SEM +__ 0.65). In the first group (experimental), the sudden omission of the amino acid produced an increase (26.7%, t = 3.28, p < .05). However the sudden introduction of amino acid produced no significant change in intake (+0.35%, t = 0.032, n.s.) in the second (control) group. EXPERIMENT 3 AND 4 Subjects. Fourteen new rats in Experiment 3, 12 new rats in Experiment 4.
Treatment. The same as in the previous two experiments, with the following exceptions. In the last two experiments, single amino acids were tested, and the control saline solution was adjusted to be equimolar to the amino acid solution. Further, the intake (after an initial 7 days of training) of each rat was monitored individually. When 4 consecutive days of intake, with no change of greater than 20% over the previous day was observed, that rat's condition was switched either from amino acid to saline or from saline to amino acid. This procedure continued until every rat was switched between the amino acid and saline conditions. In Experiment 3, seven rats were injected intragastrically with 15 mM L-tryptophan at 20% the rate of their oral intake, thus resulting in 4.3 mM of L-tryptophan in the stomach. Another seven rats were injected intragastrically with 15 mM NaCL at the same rate. Experiment 4 was exactly like Experiment 3 except that 150 mM Llysine was tested, and that there were six new rats in each of the two groups. Results: In Experiment 3 there was no significant difference in intake between the amino acid and saline group on the 4 days before test. The
LEARNING OF SATIETY SIGNALS
297
L-tryptophan injected group drank a mean of 11.0 ml (SEM - 1.1) and the saline injected group drank a mean of 9.9 ml (SEM _+ 0.8). When the rats receiving the L-tryptophan were switched to equimolar saline, mean oral intake increased by 25.5% (SEM _+ 4.4). This increase is highly significant (t = 5.8, p > 0.002). However when the rats receiving saline were suddenly injected with L-tryptophan, there was a small but nonsignificant increase in intake (13.1%, SEM + 6.8, t = 1.9, n.s.). In Experiment 4, there was no significant difference in intake between the amino acid and saline group on the 4 days before the conditions were switched. The L-lysine injected group drank a mean of 10.5 ml (SEM + 1.6) and the saline injected group drank a mean of 12.2 ml (SEM ___ 1.9). When the rats receiving the L-lysine were switched to equimolar (150 raM) saline their oral intake increased by 37.3% (SEM __- 6.0). Such an increase is highly significant (t = 6.2, p > 0.002). However when the rats receiving saline were switched to equimolar L-lysine, there was a nonsignificant decrease in intake (13.7% SEM _ 5.4, t = 2.5, n.s.).
DISCUSSION In all four experiments described above the omission of intragastrically injected amino acid that had up to then been paired with the ingestion of a triglyceride meal led to an immediate increase in the amount of tryglyceride consumed. One interpretation of such an increase is that the injection of intragastric amino acid is aversive, and that the removal of such an aversive stimulus produces an immediate increase in ingestion. However, there are two separate controls in this experiment that both contradict such an interpretation. If the removal of the amino acid was the removal of an aversive stimulus, then its presence or sudden introduction should reduce ingestion. However, there was no significant difference on the last 4 days of training in oil intake in any of the four experiments between the experimental group receiving the amino acids and the control group receiving saline nor was any difference apparent before. Further, there was no significant decrease in oil intake in all the experiments where the controls were subjected to an intragastric injection of amino acid on the test day. This control constitutes a very direct test of the aversiveness of the intragastric injection of amino acid. It could be argued that the aversiveness of the amino acid might not be detected the first time it was injected because it took some time to build up as a result of postabsorptive factors. But this interpretation is contradicted by the fact that increased ingestion occurs sharply as soon as the amino acid is omitted and by the fact that there is no difference between the experimentals and controls before the test day. Related to the issue of aversiveness is the point that the amino acids used in the experiment might have produced some amino acid imbalance in the rats that might in some undetermined way have produced the
298
DEUTSCH AND TABUENA
results reported above. Other work in this laboratory has shown large and significant effects using the same paradigm as above, with a free fatty acid (35% increase when 5% intragastric oleic acid is omitted) (Deutsch & Gonzalez, unpublished) and maltose (47% when 5% intragastric maltose is omitted) (Deutsch & Moore, unpublished). In any case it appears unlikely that the percentage of amino acids in the stomach when oil emulsion was injected would produce amino acid imbalance. For instance in Experiment 3, the L-tryptophan constitutes 0.12% of the oil intragastrically present. However "the rat tolerates quite well 1% or slightly more of either the D or L form" (Harper, Benevenga, & Wohlhueter, 1970) of tryptophan. With regard to L-threonine, 4% of threonine in the total diet did not depress the growth rate of rats (Wretlind, 1949) whereas 1.6% was administered in Experiment 1. In the case of L-lysine, growth rate is retarded after lysine is made to comprise about 5% of the diet (Russell, Taylor, & Hogan, 1952). In contrast we administered 1.6% lysine in combination with the oil in the first experiment and about 1.2% in the fourth. The measurements on the effects of amino acid imbalance were made on immature rats, still in a stage of rapid growth on a diet which was marginal in protein content. Any imbalance is much better tolerated by mature animals fed a well-balanced diet (Harper et al., 1970) such as were used in our experiments. The sudden increase in intake when amino acids are omitted from the intragastric injection seems more plausibly interpreted by the hypothesis that they function as satiety signals. The lack of a corresponding decrease in intake when these amino acids in the same proportion are suddenly introduced shows that learning has to take place before they function as satiety signals. Because the amount ingested is the same before conditions are reversed, the process of learning to use a particular signal must also involve a process of calibration. For the signals generated by the breakdown products of triglyceride lead to the same intake as the identical signals generated by the breakdown products of triglyceride plus the extra signals generated by the amino acids. While there are more signals in the latter case their effectiveness in inhibiting intake becomes the same presumably because of some process of calibration based on the postabsorbtive consequences of a meal. It appears from our results that the value of satiety signals is not innately fixed in the amount of feeding inhibition exerted and that satiety signals can be arbitrarily associated with different nutrients through a process of learning. REFERENCES Antin, J., Gibbs, J., Holt, J., Young, R. C., & Smith, G. P. (1975). Cholecystokinin elicits the complete behavioral sequence of satiety. Journal of Comparative and Physiological Psychology, 89, 784-790. Davis, J. D., & Campbell, C. S. (1973). Peripheral control of meal size in the rat" Effect of sham feeding on meal size and drinking rate. Journal of Comparative and Physiological Psychology, 83, 379-387.
LEARNING OF SATIETY SIGNALS
299
Deutsch, J. A. (1983). Dietary control and the stomach. Progress in Neurobiology, 20, 313-332. Deutsch, J. A., & Gonzalez, M. F. (1980). Gastric nutrient content signals satiety. Behaviorol and Neural Biology, 30, 113-116. Deutsch, J. A., & Gonzalez, M. F. (1981). Gastric fat content and satiety. Physiological Behavior, 26, 673-676. Deutsch, J. A. Gonzalez, M. F., & Young, W. G. (1980). Two factors control meal size. Brain Research Bulletin, 5 (Suppl. 4), 55-57. Deutsch, J. A. Young, W. G., & Kalogeris, T. J. (1978). The stomach signals satiety. Science, 201, 165-167. Deutsch, J. A., & Ahn, S. J. (1986). The splanchnic nerve and food intake regulation. Behavioral Biology, 45, 43-47. Deutsch, J. A., & Moore, B. O. Unpublished. Deutsch, J. A., & Gonzalez, M. F. Unpublished. Gonzalez, M. F., & Deutsch, J. A. (1985). Intragastric Injections of partially digested triglyceride suppress feeding in the rat. Physiology and Behavior, 31, 861-865. Gonzalez, M. F., & Deutsch, J. A. (1981). Vagotomy abolishes cues of satiety produced by gastric distention. Science, 212, 1283-1284. Harper, A. E., Benevenga, N. J., & Wohlhueter, R. M. (1970). Effects of Ingestion of Disproportionate Amounts of Amino Acids. Physiological Reviews, 50, 428-558. Rogers, Q. R., & Harper, A. E. (1965). Amino acid diets and maximal growth in the rat. Journal of Nutrition, 87, 267-272. Russell, W. C., Taylor, M. W., & Hogan, J. M. (1952). Effects of excess essential amino acids on growth of the white rats. Arch. Brocheur. Biophys. 39, 249-253. Wretlind, K. A. J. (1949). The effect of synthetic acids essential for growth on the bodyweight of growing rats, and the synthesis of the amino acids used. Acta Physiologica Scandinavica 17 (Suppl. 59), 1-101. Young, W. G., & Deutsch, J. A. (1981). The construction, surgical Implantation, and use of Gastric Catheters, and a Pyloric Cuff. Journal of Neuroscience Methods, 3, 377384. Young, R. C., Gibbs, J., Antin, J., Holt, J., & Smith, G. P. (1974). Absence of satiety during sham feeding in the rat. Journal of Comparative and Physiological Psychology, 87, 795-800.