Gut-Brain Nutrient Signaling: Appetition vs. Satiation

Abstract

Multiple hormonal and neural signals are generated by ingested nutrients that limit meal size and suppress postmeal eating. However, the availability of sugar-rich and fat-rich foods can override these satiation/satiety signals and lead to overeating and obesity. The palatable flavor of these foods is one factor that promotes overeating, but sugar and fat also have postoral actions that can stimulate eating and increase food preferences. This is revealed in conditioning studies in which rodents consume flavored solutions paired with intragastric sugar or fat infusions. The significant flavor preferences and increased intake produced by the nutrient infusions appear to involve stimulatory gut-brain signals, referred to here as appetition signals, that are distinct from the satiation signals that suppress feeding. Newly developed rapid conditioning protocols may facilitate the study of postoral appetition processes.

Introduction

During the last 40 years, the Columbia University Seminar in Appetitive Behavior has provided a valuable forum to discuss the latest developments in virtually all areas of ingestive behavior research. One of the many topics covered in the Seminar is the role of gut-brain nutrient signaling in limiting meal size and food intake. Our understanding of this process was greatly advanced by the use of gastric and intestinal infusions to study the influence of nutrients on the suppression of ongoing eating (satiation) and eating in the post-meal interval (satiety). The infusion procedure has been extensively used in experimental animals, including rats, mice, rabbits and pigs, as well as in humans to study the satiation/satiety processes (French & Cecil, 2001; Houpt, 1982; Smith, 1998a). These and other studies led to the identification of several gut and pancreatic hormones that suppress feeding by activating vagal projections to the brain or by direct actions in the brain (Smith, 1998b). The first “satiation” hormone discovered was cholecystokinin (CCK) which is released by intestinal cells by ingested food (Smith, 1998c). Subsequent studies identified several other gut hormones including GLP-1 and PYY, and the pancreatic hormones amylin and glucagon (Woods, 2009). During the course of a meal, intestinal and absorbed nutrients stimulate the release of these hormones in increasing amounts until a “satiation” threshold is reached and the meal is terminated (Woods, 2009). Gastric distention by ingested food is another satiation signal that limits meal size (Berthoud, 2008).

All of these gut satiation signals that suppress eating raise the question why overnutrition and obesity are such predominant health concerns in many modern societies. One simple potential explanation is the wide availability of palatable sugar- and fat-rich foods which tempt the palate and promote overeating (Berthoud, 2011; Egecioglu et al., 2011; Erlanson-Albertsson, 2005; Ryan et al., 2012). Animal studies document the ability of sweet taste and fatty flavor to activate brain reward systems that promote food consumption (Hajnal & Norgren, 2008; Liang et al., 2006). The obesity promoting effects of high-sugar or high-fat foods, or an assortment of such foods are well established in laboratory rodents (Sclafani, 1993; Speakman et al., 2008). The effectiveness of intestinal and postabsorptive satiation signals to suppress feeding is reduced as the palatability of the food source is increased (Bédard & Weingarten, 1989; Campbell & Davis, 1974). In addition, chronic exposure to high-fat diets appear to raise the defended adiposity threshold such that higher levels of short-term satiation and long-term adiposity signals are required to suppress feeding (Ryan et al., 2012).

While most attention has focused on the palatable flavor of high-sugar and high-fat foods in promoting overconsumption, research during the last 40 years has revealed that postoral actions of sugar and fat can also stimulate consumption and condition increased food preference (Sclafani & Ackroff, 2012). I propose the term “appetition” to refer to the postoral processes that increase food intake and food preference to distinguish them from the satiation processes that inhibit ingestion. This Seminar paper briefly reviews the evidence for postoral appetition.

Appetition

The first definitive study showing that postoral nutrient actions can enhance food preferences was reported by Holman using a conditioned flavor preference paradigm (Holman, 1968). Food-restricted rats were trained to drink a flavored saccharin solution (the conditioned stimulus or CS+, e.g., lemon) paired with intragastric (IG) infusions (7–10 ml) of an eggnog diet (the unconditioned stimulus or US) in 5 min/day sessions alternating with another flavored solution (the CS−, e.g., anise) paired with IG water infusions. After six training sessions a two-bottle test (140 min) was conducted with the CS+ vs. CS− solutions without IG infusions. The rats showed a significant CS+ preference (~66%) preference during the first 20 min of the test although the preference dissipated in the remaining 120 min of testing. While the preference was rather modest and temporary, perhaps due to the short training sessions and complex diet used, Holman’s findings established that food preferences can be influenced by the postoral actions of nutrients.

Over the next 20 years several more studies reported that IG nutrient infusions (milk, glucose, casein) conditioned flavor preferences in food-restricted rats (Puerto et al., 1976; Deutsch & Wang, 1977; Sherman et al., 1983; Baker et al., 1987). Given the nutritional needs of the animals, it was proposed that caloric or nutrient restoration served as the effective “reinforcer” for the conditioned flavor preference (Sherman et al., 1983; Baker et al., 1987). However, in 1988 we reported that rats with unlimited access to a chow diet developed a very strong (96%) preference for a flavored solution (e.g., cherry) that was paired with IG infusions of 16% maltodextrin (Polycose) over a differed flavored solution (e.g., grape) paired with IG water infusions during 24 h/day conditioning sessions (Sclafani & Nissenbaum, 1988). Furthermore, this preference persisted over four days of “extinction” testing during which both the CS+ and CS− were paired with IG water infusions. In this and many subsequent studies an automated “IG self-infusion” procedure was used in which the animal controlled the infusion by its licking of the CS+ and CS− sipper tubes and the volume infused matched the CS volume consumed (Elizalde & Sclafani, 1990). These subsequent studies reported that non-deprived rats and mice acquired strong preferences for CS+ solutions paired with IG self-infusions of sugars (glucose, sucrose, maltose), fat (corn oil, Intralipid) or milk diets in 24-h tests (Lucas & Sclafani, 1989; Sclafani et al., 1993; Azzara & Sclafani, 1998; Lucas et al., 1998a; Sclafani & Glendinning, 2005; Ackroff & Sclafani, 1994; Lucas et al., 1998a). Preferences were conditioned by IG self-infusions of calorically dense (2.1 kcal/ml) as well as very dilute (0.04 kcal/ml) nutrient sources (Ackroff & Sclafani, 1994; Lucas et al., 1998a). IG nutrient self-infusions also conditioned flavor preferences in food-deprived and non-deprived rats given 30 min/day training sessions (Yiin et al., 2005). Taken together, these findings demonstrate that animals need not be in an energy depleted state in order to acquire nutrient-based flavor preferences. Furthermore, energy repletion may not be sufficient to reinforce a flavor preference. In particular, IG fructose infusions, unlike glucose infusions, do not condition flavor preferences in hungry rats trained 30 min/day (Sclafani et al., 1999). Nevertheless, nutrient restoration may be critical for some types of flavor conditioning as in the case of amino acid or calcium deficient animals (Gietzen et al., 1992; Tordoff, 2002).

In addition to conditioning preferences, IG carbohydrate or fat self-infusions can substantially increase the intake of a flavored solution in some test situations. For example, chow-fed rats drank 2–3 times more saccharin solution when intake was paired with IG self-infusion of 6% glucose rather than with water (Ramirez, 1994). Chow-fed mice (C57BL/6, B6) also consumed substantially (~95%) more of a saccharin-sweetened CS+ solution paired with IG self-infusions of sugar (16% sucrose) or fat (5.6% Intralipid) than of the CS− solution paired with IG water during 24 h/day training sessions (Sclafani & Glendinning, 2005). Note, however, that the intake-stimulating effect of IG nutrient infusions is dependent upon the palatability of the CS+ flavor: the intake of a sweet CS+ is simulated much more than the intake of a non-sweet CS+ solution (Ackroff & Sclafani, 1994; Ramirez, 1994; Sclafani & Glendinning, 2005). In one study rats given unlimited access to chow and a palatable sweet solution paired with IG self-infusions of 32% maltodextrin for 28 days, consumed more solution, total calories, and gained more weight than did rats given a mildly unpalatable bitter solution (Sclafani et al., 1996). These findings indicate that the overeating and obesity promoting effects of sugar- and fat-rich foods are due to both their palatable (“appetizing”) flavor and their postoral stimulatory (“appetition”) actions.

The flavor preference and intake stimulating actions of nutrients described above were produced by IG infusions. An early study proposed that flavor conditioning is mediated by nutrient sensors in the stomach (Deutsch & Wang, 1977), but subsequent work challenged this idea (Baker & Booth, 1989; Drucker & Sclafani, 1997). In the case of glucose, the upper intestinal tract appears to be a critical site of action for flavor conditioning. This is indicated by the findings that preferences for a CS+ saccharin solution were conditioned by IG glucose infusions only if the sugar was allowed to empty into the intestine, and by direct glucose infusions into the duodenum or jejunum but not into the ileum or hepatic-portal vein (Ackroff et al., 2010; Drucker & Sclafani, 1997). However, hepatic-portal glucose infusions conditioned a preference when the CS+ flavor was presented in a nutritive chow rather than a non-nutritive saccharin solution (Tordoff & Friedman, 1986). These and other results (Gowans, 1992) suggest that intravascular glucose can reinforce a flavor preference when it is associated with a CS+ that provides preabsorptive nutrient stimulation (Ackroff et al., 2010).

While it is now well established that gastrointestinal nutrient infusions can condition flavor preferences and stimulate intake in laboratory rodents and as well as sheep (Villalba & Provenza, 1997), the time course of this appetition effect has been uncertain. In most conditioning studies animals are given multiple training sessions with the CS+ and CS− paired with nutritive and non-nutritive infusions, respectively, before the flavor preference test is conducted. Multiple training trials are not always necessary, and two studies have reported one-trial flavor conditioning in rats given IG glucose infusions (Ackroff et al., 2009; Myers, 2007). However, the critical flavor choice test was not conducted until a day or more after the CS+ conditioning trial so these studies did not reveal how rapidly the IG glucose conditioned the CS+ preference, i.e., within or after the training session.

Using a new mouse conditioning model, we observed that IG nutrient infusions can rapidly stimulate on-going ingestion within the first CS+ conditioning session (Zukerman et al., 2011). Food-restricted B6 mice were trained to drink a flavored saccharin CS− solution paired with IG self-infusions of water during daily 1-h sessions (days 1–3) followed by sessions with a CS+ solution paired with IG self- infusions of 16% glucose (days 4–6); licks were recorded every minute throughout the sessions. Lick rates at the start of the first CS+ training session were at or below that of the preceding CS− sessions but by 12–15 min the mice increased their rate of licking such that that their total 1 h CS+ licks were 44% greater than their CS− baseline. In the subsequent two CS+ training sessions, the mice licked the CS+ at a high rate from the very first minutes although their lick rate rapidly declined later in the sessions as the animals experienced the satiating action of the self-infused glucose. Nevertheless, total 1-h licks and intakes increased by 90% above the CS− baseline. Similar findings were obtained in a second group of mice that self-infused a fat emulsion (6.4% Intralipid) as they consumed the CS+ solution. The Intralipid infusion stimulated licking 10–12 min into the first session and increased total 1 h licks by 60% to 120% in CS+ sessions 1 to 3. In both experiments the mice significantly preferred the CS+ to the CS− in a two-bottle choice test (without IG infusions) conducted after training.

The rapid increase in licking observed in the first CS+ session suggests an unconditioned response to the glucose and Intralipid infusions, although rapid learning of a CS+ preference may have contributed to the appetition response. The elevated lick rates observed at the very start of the subsequent sessions is indicative of a conditioned response to the CS+ flavor. This is confirmed by the elevated lick rates observed in an extinction test with the CS+ paired with IG water self-infusions (Sclafani, unpublished observations). In the case of the Intralipid infusion, the early stimulation of drinking was presumably triggered by intestinal sensors since prior research indicates that Intralipid is not absorbed into the circulation within the first 30-min of an intestinal infusion (Greenberg et al., 1995; Yoder et al., 2009). The glucose infusion may have acted at pre- and/or post-absorptive sites to stimulate licking, but additional findings suggest an intestinal site of action. That is, unlike IG infusions, intraperitoneal self-infusions of 8% glucose failed to stimulate CS+ licking in the first or subsequent sessions (Zukerman & Sclafani, unpublished observations). As reviewed in detail elsewhere (Sclafani & Ackroff, 2012), the identity of the intestinal and/or post-absorptive nutrient sensors and signaling pathways that mediate postoral appetition and flavor conditioning by sugars and fats is not known and is the subject of current research.

The central mediation of postoral appetition is also incompletely understood. However, there is considerable evidence implicating central dopamine (DA) systems in postoral flavor preference conditioning and appetite stimulation (de Araujo et al., 2012; Sclafani et al., 2011). In particular, postoral infusions of glucose or fat are reported to activate DA circuits as measured by microdialysis, c-Fos and fMRI techniques (Ferreira et al., 2012; Ren et al., 2010; Oliveira-Maia et al., 2011). Furthermore, inhibiting DA signaling with D1 receptor antagonists in the nucleus accumbens, amygdala and medial prefrontal cortex blocks flavor conditioning by IG glucose infusions (Touzani et al., 2010a; Touzani et al., 2009; Touzani et al., 2008). Central opioid signaling does not appear to be critical for postoral glucose conditioning but little is known about the role of other neurochemical systems implicated in food reward (e.g., endocannabinoid, benzodiazepine, orexin, and ghrelin circuits) in postoral appetition (Touzani et al., 2010b).

Appetition vs. Satiation

Historically, food reward has been attributed to the satiating actions of nutrients (Le Magnen, 1987) and this is still a common view e.g., (Mayer, 2011). There is, however, little direct evidence that satiation per se conditions food preferences, and satiation may actually reduce the reward value of food (Rolls, 2006; Simpson & Bloom, 2010). In a test of the satiety-reward hypothesis, Van Vort and Smith (1983) trained rats fitted with a gastric cannula to “real-feed” and “sham-feed” different flavored milk diets; in sham-feeding, ingested liquid diets drain out the open gastric cannula and provide no nutrition or postoral satiation. In two-bottle choice tests the animals did not prefer the flavored milk (CS+real) that was real-fed and satiating over the flavored milk (CS+sham) that was sham-fed and not satiating. The authors concluded that satiation alone was not responsible for food reinforcement and flavor conditioning. In a subsequent study we observed that rats trained to real and sham-feed concentrated (32%) flavored maltodextrin solutions actually preferred the CS+sham to the CS+real solution in two-choice tests (Sclafani et al., 1994). Other animals, however, trained with a less concentrated (8%) and, therefore less satiating maltodextrin solution significantly preferred the CS+real to the CS+sham solution. These findings confirmed that postoral nutrients can reinforce flavor preferences (as in the case of real-fed 8% maltodextrin) but that highly satiating foods (e.g. 32% maltodextrin) may actually reduce postoral reinforcement.

Further evidence for a dissociation between satiation and reward was obtained in a study in which hungry rats were trained to drink different flavored saccharin solutions (CS+8, CS+16 and CS+32) that were paired with IG self-infusions of 8%, 16% and 32% maltodextrin, respectively (Lucas et al., 1998b). All three CS+ solutions were preferred to a CS− solution paired with IG water infusion but, in direct choice tests, the CS+16 was preferred to both the CS+8 and CS+32. Flavor conditioning also varies as a function of nutrient composition. In particular, we (Lucas et al., 1998a) observed that chow-fed rats trained 24 h/day preferred a flavor (CS+HF) paired with a high-fat diet (HFD) IG infusion over a flavor (CS+HC) paired with an isocaloric high-carbohydrate diet (HCD) infusion; both flavors were strongly preferred to a water-paired CS− flavor. The HFD was less satiating, i.e., suppressed chow intake less, than the HCD. Reducing the satiating potency of the HCD, by diluting it with water, to that of the HFD resulted in equal preferences for the CS+HCD and CS+HFD even though the CS+HCD was now paired with a hypocaloric diet. These and other results (Warwick & Weingarten, 1996; Booth et al., 1972; Sclafani, 2002) indicate that postoral reward effects of nutrients vary as a function of nutrient concentration and composition but that the most concentrated (and satiating) nutrient source may not be the most reinforcing.

Similar findings have been obtained in human flavor conditioning studies. As recently reviewed by Yeomans (2012), flavor preference conditioning has been studied in humans by training subjects to ingest flavored drinks or foods that differ in caloric density but are matched in flavor intensity and palatability. Conditioned flavor preference and acceptance are then measured using rating scales and intake tests. Several studies report increased preference (liking) for the flavor associated with the more energy-rich food source, but this is not always the case (Yeomans, 2012). According to Yeomans, conditioned preferences are more likely with a flavor associated with a moderate energy load rather than a high energy load which may have aversive satiating effects.

Another approach to the study of satiation and food reward is to examine the ability of “satiety” hormones to condition flavor preferences. Early investigators hypothesized that nutrient conditioned flavor preferences are mediated by the release of CCK (Mehiel & Bolles, 1988). In a test of this idea, we trained hungry rats to drink CS+ and CS− solutions paired with intraperitoneal injections of CCK octapeptide (0.125 – 4 ug/kg) and saline, respectively (Pérez & Sclafani, 1991). In subsequent choice tests, rats displayed a significant, but weak preference for a CS+ paired with a low dose of CCK (0.5 ug/kg) that did not suppress intake. However, rats trained with higher CCK doses (2 – 4 ug/kg) that suppressed intake displayed either no preference or avoided the CS+ flavor. In a second study, treating rats with a dose of the CCK receptor antagonist devazepide that increased intake had no effect on flavor conditioning by maltodextrin infusions (Pérez et al., 1998). Taken together, these findings indicate that a non-satiating dose of CCK can condition a mild flavor preference, but that CCK signaling is not required for carbohydrate-conditioned preferences. A possible role of CCK in protein-conditioned preferences has been suggested (Peuhkuri et al., 2011) but remains to be determined. There are as yet no reports of flavor preference conditioning by other gut satiety hormones. Rather, flavor aversions and/or reduced food reward have been reported with PYY and GLP-1 or GLP agonists (Thiele et al., 1997; Chelikani et al., 2006; Kanoski et al., 2012). The satiating actions of gut hormones and stomach distention are mediated in part by vagal afferent projections to the brain (Berthoud, 2008). Vagal afferent signaling, however, is not essential for postoral flavor-nutrient conditioning, which further argues for a distinction of appetition and satiation processes (Lucas & Sclafani, 1996; Sclafani & Lucas, 1996; Sclafani et al., 2003; Uematsu et al., 2010; Zukerman et al., 2011)

While there is little evidence that gut satiety hormones enhance food reward, numerous studies indicate that the orexigenic hormone ghrelin promotes food reward (Olszewski et al., 2008; Dickson et al., 2011). In particular, ghrelin injections increase the intake of sweet solutions, food-conditioned place preferences, and lever pressing for food reward (Skibicka et al., 2012), while ghrelin antagonism or gene deletion attenuates place preference conditioning and food reinforced behavior (Dickson et al., 2011; Olszewski et al., 2008; Perello & Zigman, 2012). Systemic administration of ghrelin by itself was found to condition a place preference in rats (Jerlhag, 2008) and conceivably ghrelin may also condition flavor preferences as well, but this remains to be determined. Ghrelin levels are increased when animals are food-restricted and it is not surprising that the hormone enhances food reward. However, nutrient intake or infusions suppress ghrelin release (Foster-Schubert et al., 2008; Williams et al., 2003) which argues against a direct role of ghrelin in the preference conditioning response to IG glucose and fat infusions (Sclafani & Ackroff, 2012; Ackroff et al., 2010). Nevertheless, the role of ghrelin in nutrient conditioning warrants investigation.

Summary

In summary, appetite is stimulated not only by the palatable flavors of sugar- and fat-rich foods but also by the postoral actions of these nutrients. Postoral food reward is often assumed to derive from the food’s satiating action, but there is little direct evidence for this view. Instead, flavor conditioning studies indicate that postoral nutrient-conditioned preference and stimulation of intake are mediated by an appetition system that is distinct from the satiation system that suppresses intake. Progress in understanding postoral appetition has been hampered in the past because of the multiple training trials required to establish learned preferences and increased appetite. However, the recent development of conditioning protocols that produce one-trial learning and rapid within session stimulation of intake (Ackroff et al., 2009; Myers, 2007; Zukerman et al., 2011) should facilitate the study of postoral positive controls of feeding.

• Ingested nutrients stimulate the release of satiation signals that suppress feeding

• Palatable sugar- and fat-rich foods can override satiation and promote overeating

• Sugar and fat also generate postoral “appetition” signals that stimulate feeding

• New rapid flavor conditioning procedures may facilitate the study of appetition