On the Roles of the Duodenum and the Vagus Nerve in Learned Nutrient Preferences

Abstract

Background and Aim

In most species, including humans, food preference is primarily controlled by nutrient value. However, the gut-brain pathways involved in preference learning remain elusive. The aim of the present study, performed in C57BL6/J mice, was to characterize the roles in nutrient preference of two critical elements of gut-brain pathways, i.e. the duodenum and vagal gut innervation.

Methods

Adult wild-type C57BL6/J mice from a normal-weight cohort sustained one of the following three procedures: duodenal-jejunal bypass intestinal rerouting (DJB), total subdiaphragmatic vagotomy (SDV), or sham surgery. Mice were assessed in short-term two-bottle preference tests before and after 24hs exposures to solutions containing one of glutamate, lipids, sodium, or glucose.

Results

DJB and SDV interfered in preference formation in a nutrient-specific manner: whereas normal preference learning for lipids and glutamate was disrupted by both DJB and SDV, these interventions did not alter the formation of preferences for glucose. Interestingly, sodium preferences were abrogated by DJB but not by SDV.

Conclusions

Different macronutrients make use of distinct gut-brain pathways to influence food preferences, thereby mirroring nutrient-specific processes of food digestion. Specifically, whereas both vagal innervation and duodenal sensing appear critical for generating responses to fats and protein, glucose preferences recruit post-duodenal, vagal-independent pathways in pair with the control of glucose homeostasis. Overall, our data suggest that the physiological processes involved in digesting and absorbing fats, amino acids, and glucose overlap with those mediating learned preferences for each of these nutrients.

Introduction

Acquired food preferences are primarily driven by nutritional value, an effect mediated by gut signals that reach the brain independently of taste processing(de Araujo, et al., 2008; Ferreira, Tellez, Ren, Yeckel, & de Araujo, 2012; Tellez, Ferreira, et al., 2013). However, the gastrointestinal sites and associated gut-brain pathways involved in conveying nutritional information remain unclear.

One source of uncertainty relates to the identity of the gastrointestinal segments that may be important for preference learning. Evidence to date leaves unresolved whether gut segments influence preferences in a nutrient-dependent manner; in fact, whereas duodenal sensing is not required for glucose-based preference learning(Ackroff, Yiin, & Sclafani, 2010; Zhang, Han, Motta, Li, & de Araujo, 2018), it remains unknown whether this effect also applies to nutrients engaging different physiological pathways such as lipids and amino acids.

Moreover, it remains equally unclear the extent to which gastrointestinal vagal endings are required for preference learning. Again, evidence to date suggest that vagal terminal activation is not required for glucose-based preference learning(Sclafani & Lucas, 1996). In contrast, vagal does signaling appear to be critical for learned preferences to glutamate(Uematsu, Tsurugizawa, Uneyama, & Torii, 2010). However, as in the case for duodenal sensing, the impact of truncal vagotomies on lipid-based preference learning remains unclear(Sclafani, Ackroff, & Schwartz, 2003).

Accordingly, the aim of the present study was to assess the roles of duodenal and vagal signals in preference formation for glucose, lipid, glutamate and salt using the same behavioral protocol. We made use of surgical duodenal bypass and truncal vagotomies to establish the roles of the upper intestine and vagal signaling in both acute and learned responses to these nutrients.

Materials and Methods

Animals

A total of 20 adult male mice were used (C57BL6/J, Jackson Laboratory). Upon assessment of recovery from subdiaphragmatic vagotomies, five animals were excluded from experiments, such that a total of 15 mice were used. At the time of behavioral experiments, animals were 11–14 weeks old. Throughout tests animals were fed with standard rodent chow. All experiments were conducted in accordance with the J.B. Pierce Laboratory and Yale University regulations on usage of animals in research.

Surgeries

Overall Approach

Based on previous studies (see Introduction) we anticipated an absence of effects to be produced by vagotomies on sweet sugar preferences. Therefore, in order to rule out potential false negative effects associated with vagal disruption, we opted for performing the more blunt truncal vagotomy procedure(Ferreira, et al., 2012) rather more nuanced techniques such as selective deafferentation.

Subjects

Five animals were assigned to one of the three experimental groups, namely, sham surgery controls; duodenum-jejunum bypass or subdiaphragmatic vagotomy. Sham and duodenal-jejunal bypassed (DJB) mice were operated at 8-9 weeks of age. Unfortunately, for subdiaphragmatic vagotomized (SDV) mice, surgery at 8-9 weeks resulted in severe side effects including extreme gastroparesis and subsequent anorexia. These side effects were however dramatically reduced in mice vagotomized at 3 weeks of age. We opted for using mice vagotomized at 3 weeks to avoid confounds associated with delayed recovery. In any event all mice were tested at same ages, 11-14 weeks old. See Table 1 for details.

Table 1.

Effects of duodenal bypass (DJB) and vagotomies (SDV) on body weight.

Body Weight (g)/Test Day	Test day 1	Test day4	Test day7	Test day 10	Test day 13
Sham	31.8±1.5	32.1±1.4	33.0±1.3	33.9±1.3	35.0±1.4
DJB	25.5±5.3	26.4±5.5	27.5±5.7	28.2±5.9	28.9±6.0
SDV	22.2±3.8	22.7±3.9	23.2±4.0	23.7±4.0	24.3±4.1

Overall body weight effect two-way (test day x group) ANOVA F[8,48]=4.027 p=0.001

Sham vs. SDV F[4,32]=3.014, p=0.032

DJB vs. SDV F[4,32]=7.878, p<0.001

Duodenum-jejunum bypass (DJB)

Animals were anesthetized with an intraperitoneal injection of a ketamine/xylazine (100/15 mg/Kg) mix. A midline incision was made into the abdomen, and the stomach, esophagus and upper intestine exposed. To divert food from its natural course through the duodenum, the pylorus was ligated and, at approximately 6-8cm below the pylorus, the jejunum was anastomosed to the body of the stomach. See Figure 1B.

Figure 1. Behavioral protocol, schematics of surgeries, and efficiency of bypass intervention.

A. Mice were tested in three different days. On day 1, an ad libitum short-term pre-test involved two bottles (10 min, sipper positions switched after 5 min) containing the test solutions. After 24hs of rest, all food was removed from the home cages and mice were given 24hs ad libitum access to these same solutions now contained in the graded sippers for (sipper positions switched after 12 hours). After 24hs of rest, choice tests were once again performed exactly as on day 1, including same time of day. B. Schematic representation of the duodenal-jejunal bypass (DJB) intervention, which includes pylorus ligation (parallel black lines drawn under stomach) and a gastric-jejunal anastomose. Green arrows indicate rerouting. C. A blue dye was infused through gastric catheter for assessment of surgical efficacy. In sham-operated animals blue dye is promptly visible in duodenum. D. In DJB mice however dye was visible only in stomach and jejunum. Legend: 1. Stomach; 2. Duodenum; 3. Distal jejunum. Arrows point to the corresponding structures delimited by dotted lines of same colors as arrows. The apparent difference in blood supply between control and bypassed animals is due to the large number of ligations in the bypassed procedure that aim at containing excessive bleeding. Pictures shown are representative of the procedures. E. Schematic representation of subdiaphragmatic vagotomies. Parallel black lines show where nerves were severed. F. Green arrows indicate the posterior (right) vagal trunk. G. Green arrows indicate the anterior (left) vagal trunk.

Subdiaphragmatic vagotomies (SDV)

Animals were anesthetized with an intraperitoneal injection of a ketamine/xylazine (100/15 mg/Kg) mix. A midline incision was made into the abdomen, and the stomach, esophagus and upper intestine exposed. Upon exposure of the esophagus, the surrounding connective tissue was removed and the vagal branches exposed and completed severed at the lower aspect of the esophagus (short parallel black lines in line 1C). We opted for this approach to assure complete denervation.

Sham surgeries

Animals were anesthetized with an intraperitoneal injection of a ketamine/xylazine (100/15 mg/Kg) mix. A midline incision was made into the abdomen, and the stomach, esophagus and upper intestine exposed and gently manipulated.

Behavioral Protocol

Stimuli

D-Glucose, Sodium chloride (NaCl), and Monosodium Glutamate (MSG) were all obtained from Sigma. Solutions were prepared in tap water freshly every testing day. The base lipid emulsion was 30% IntraLipid® (Baxter Healthcare). The emulsion contains as main components 30% soybean oil, 1.2% egg yolk phospholipids, 1.7% glycerin, and water. The caloric density of 30% Intralipid® is 3.0 Kcal/mL, with 2.7Kcal/mL accounted for by soybean oil and 0.3 Kcal/mL by phospholipids + glycerin. The original 30% emulsion was diluted into an emulsifying control solution (1.2% phospholipids + 1.7% glycerine, in water) in order to prepare the 7.5% dilution.

Behavioral apparatus

Short-term choice tests were conducted in either one of three identical mouse behavior chambers enclosed in a ventilated and sound-attenuating cubicle (Med Associates Inc., St. Albans, VT). Each chamber is equipped with two slots for sipper tubing placements, at symmetrical locations on one of the cage walls. All sippers are connected to a contact-based licking detection device allowing for measurements of licking responses with 10ms resolution. All lick timestamps were saved in a computer file for posterior analysis. For long-term tests, mice were left in their home cage under food deprivation. Stimuli were offered in graded (0.1mL) vials with metallic sippers, and volumes ingested were recorded.

Experimental Design

Mice were tested in three different days. On day 1, animals were water and food deprived for 12 hours, followed by an ad libitum short-term pre-test involved two bottles (10 min, sipper positions switched after the initial 5 min) containing the test solutions. After 12hs of rest, all food was removed from the home cages and mice were given 24hs ad libitum access to these same solutions now contained in the graded sippers for (sipper positions switched after 12 hours) (graded sippers were modified from 50 mL Falcon Centrifuge Tubes, weight and volume were measured simultaneously to confirm accuracy). A third sipper containing water was also made available during the lipid and glucose tests to avoid thirst-related confounding factors. After 12hs of rest and 12hs water and food deprivation, choice tests were once again performed exactly as on day 1, including same time of day. The rationale for the design was as follows. If mice are capable of detecting the nutritional value of a given solution, then the choice rate for this solution during short-term tests should significantly increase after long-term exposure compared to the initial short-term test. To control for order effects, mice of each group were randomly assigned a specific sequence of nutrients to be tested, i.e. lipids, sugar, monosodium glutamate, or sodium. This allowed for a balanced order of nutrient presentation across experimental groups.

Body weight measurements

Body weights were assessed at the end of each test cycle, i.e. whenever testing for one of the four nutrients were completed (i.e. Test days 1, 4, 7, 10 and 13) since for each of the nutrients tests lasted for a total of 3 days.

Details on Short-term two-bottle preference tests

Short-term (10 min) two-bottle preference tests between the two distinct solutions were performed previous to and following the 24hs sessions. Short-duration of this test aims to minimize post-ingestive influences. After 24hs exposure sessions, an identical test was performed to assess the formation of Solution preferences. The number of licks for each Solution was recorded and used to calculate the preference ratio as follows:

$$\text{Preference ratio for}\mathrm{Solution}1=\frac{n(\mathit{Solution}1)}{n(\mathit{Solution}1)+n(\mathit{Solution}2)}$$

where n(Solution x) denotes the detected number of licks to Solution x during a given session. To eliminate the influence of side-biases mice were tested two consecutive times with sipper positions being switched after 5 min, and the average between the two trials taken as the actual preference.

Statistical Analyses

Data analyses were performed using SPSS (v.21.0, IBM Predictive Software) and made use of linear model analyses as two-way (repeated measures) ANOVAs. When relevant post-hoc pairwise tests were based on paired within-subject or between-groups independent samples T-tests, with resulting p-values being subsequently subjected to Bonferroni correction for multiple comparisons. Normality was assessed using SPSS by inspecting and confirming data linearity upon plotting quantile-quantile graphs for every dataset. Violations of sphericity were tested for and ruled out in SPSS by performing Mauchly’s Test of Sphericity for every dataset. Data are reported as mean ± SEM.

Results

Effects of Surgeries on Body Weight

Table 1 displays information on body weight and age of animals used. We found no significant differences in body weight between Sham and duodenal-jejunal bypassed (DJB) mice, despite a trend towards lower body weight in DJB mice. This is consistent with previous reports(Breen, et al., 2012; Han, et al., 2016). However, subdiaphragmatic vagotomized (SDV) mice weighted significantly less than the other groups at the time of testing.

Preference Formation for High-Calorie Lipids

First, the overall behavioral protocol is illustrated in Figure 1A (See also Materials and Methods, Experimental Design). Briefly, mice were tested in three different days. On day 1, a 10-min choice test involved two bottles containing the test solutions. On a subsequent day the hungry mice were given 24hs ad libitum access to these same solutions in their home cages, and finally, choice tests were once again performed exactly as on day 1. We reasoned that if mice are capable of detecting the nutritional value of a given solution, then the choice rate for this solution during short-term tests should significantly increase after long-term exposure compared to the initial short-term test.

Figure 2 shows the results of testing the animals in the three experimental phases for lipids, in the form of the emulsified triglyceride IntraLipid^®. Animals were initially exposed to a 10 min choice test between a low-calorie (7.5%) and a high-calorie (30%) lipid solution. Initially, during the first short-term tests (which presumably assess the gustatory properties of the solutions), no evidence for preference for any of the solutions was found, with mice in all groups (i.e. DJB, SDV, and sham controls) equally choosing between the two solutions (Figure 2A).

Figure 2. Preferences for 30% IntraLipid versus 7.5% IntraLipid.

N=5 in all groups. Data shown as mean±s.e.m. for preference ratios for the 7.5% IntraLipid solution compared against 7.5% IntraLipid. The horizontal line in panels A-C indicates the indifference ratio of 0.5. Groups are labeled below each bar plot. A. During the short-term pre-test, mice in all groups showed no clear preference for any of the solutions (One-way ANOVA between group effect F[2,12]=1.6, p=0.2). B. During the short-term post-test, i.e. after 24 hour exposure to the two solutions, Sham mice developed a strong preference for the 7.5% IntraLipid solution, whereas both DJB and SDV mice did not (between group effect F[2,12]=10.1, p=0.003). Post-hoc tests revealed greater preferences for 7.5% IntraLipid in Sham vs. DJB (two-sample t-test t[8]=5.3, Bonferroni *p=0.002), Sham vs. SDV (t[8]=3.7, **p=0.01), but not in DJB vs. SDV (t[8]=0.02, p=0.9). C. Interestingly, during long-term choice tests (24 hs), DJB mice preferred the 30% solution significantly more than both Sham and SDV mice (F[2,12]=10.8, p=0.002, vs. Sham t[8]=4.7, *p=0.004, vs. SDV t[8]=3.8, **p=0.01), indicating a potential deficit in accommodating high calorie lipids. D. Actual consumed volumes of each solution (in mL) during the 24hs tests.

Next, the three groups of animals were exposed to a new short-term test after 24hs exposure to the two solutions. It is now clear that the Sham group developed a strong preference for the lower-calorie 7.5% solution, whereas both the DJB and SDV groups did not (Figure 2B). This is presumably derived from a deficit in lipid post-ingestive signaling. In fact, in Figure 2C it is possible to visualize that, during the 24hs test, bypassed mice preferred less of the low-calorie solution compared to the high-calorie one (although this was not the case for the SDV mice). Figure 2D shows the actual volumes consumed of each solution by the three groups of mice during the 24hs choice test shown in Figure 2C.

Preference for Glucose

Figure 3 shows the results of testing the animals in the three experimental phases for choices between 0.1% of the artificial sweetener Sucralose against 12% glucose, in the form of a liquid solution diluted in tap water. Animals were initially exposed to a 10 min choice test between sugar and sweetener. Initially, during the first short-term tests (which presumably assess the gustatory properties of the solutions), all groups (i.e. DJB, SDV, and sham controls) equally chose the sweetener solution, presumably because it is sweeter than glucose (Figure 3A).

Figure 3. Preferences for 0.1% sucralose versus 12% Glucose.

N=5 in all groups. Data shown as mean±s.e.m. for preference ratios for between 0.1% of the artificial sweetener Sucralose against 12% glucose, in the form of a liquid solution diluted in tap water. The horizontal line in panels A-C indicates the indifference ratio of 0.5. Groups are labeled below each bar plot. A. During the short-term pre-test (10 min), mice in all groups equally preferred the sweetener solution (between group effect F[2,12]=2.8, p=0.1). B. During the short-term posttest (10 min), i.e. after 24 hour exposure to the two solutions, all mice developed a strong preference for the glucose solution (F[2,12]=0.7, p=0.5). C. During long-term choice tests (24 hs), DJB, sham and SDV mice all strongly and equally preferred the glucose solution (F[2,12]=1.5, p=0.2). D. Actual consumed volumes of each solution (in mL) during the 24hs tests.

Next, the three groups of animals were exposed to a new short-term test after 24hs exposure to the two solutions. All groups now equally chose the glucose solution, presumably because it is nutritive whereas sucralose is not (Figure 3B). Based on Figure 3C it is possible to anticipate these patterns from the preferences shown during the 24hs test. Figure 3D shows the actual volumes consumed of each solution by the three groups of mice during the 24hs choice test shown in Figure 3C.

Preference Formation for Glutamate

Figure 4 shows the results of testing the animals in the three experimental phases for 1% monosodium glutamate (MSG), in the form of a liquid solution diluted in tap water. Initially, during the first short-term tests (which presumably assess the gustatory properties of the solutions), no evidence for preference for any of the solutions was found, with mice in all groups (i.e. DJB, SDV, and sham controls) equally choosing between the two solutions (Figure 4A).

Figure 4. Preferences for 1% MSG versus water.

N=5 in all groups. Data shown as mean±s.e.m. for preference ratios for the 1% MSG solution compared against water. The horizontal line in panels A-C indicates the indifference ratio of 0.5. Groups are labeled below each bar plot. A. During the short-term pre-test (10 min), mice in all groups showed no clear preference for any of the solutions (between group effect F[2,12]=0.06, p=0.9). B. During the short-term post-test, i.e. after 24 hour exposure to the two solutions, Sham mice developed a strong preference for the 1% MSG solution, whereas both DJB and SDV mice did not (F[2,12]=6.2, p=0.01), although the paired difference between Sham and DJB was marginal (SDV vs. sham t[8]=6.4, Bonferroni *p<0.002; DJB vs. Sham t[8]=2.2, p=0.05). C. During longterm choice tests (24 hs), no obvious between-groups differences were observed although SDV mce showed a trend towards lower preferences (F[2,12]=3.62, p=0.06). D. Actual consumed volumes of each solution (in mL) during the 24hs tests.

Next, the three groups of animals were exposed to a new short-term test after 24hs exposure to the two solutions. It is now clear that the Sham mice developed a strong preference for the 1% MSG solution, whereas both DJB and SDV mice did not (Figure 4B). In fact, in Figure 4C it is possible to visualize that, during the 24hs test, control mice preferred more of the MSG solution. Figure 4D shows the actual volumes consumed of each solution by the three groups of mice during the 24hs choice test shown in Figure 4C.

Preference Formation for Sodium

Figure 5 shows the results of testing the animals in the three experimental phases for 1% NaCl, in the form of a liquid solution diluted in tap water. Animals were initially exposed to a 10 min choice test between NaCl and water. Initially, during the first short-term tests (which presumably assess the gustatory properties of the solutions), no evidence for preference for any of the solutions was found, with mice in all groups (i.e. DJB, SDV, and sham controls) equally choosing between the two solutions (Figure 5A).

Figure 5. Preferences for 1% NaCl versus water.

N=5 in all groups. Data shown as mean±s.e.m. for preference ratios for the 1% NaCl solution compared against water. The horizontal line in panels A-C indicates the indifference ratio of 0.5. Groups are labeled below each bar plot. A. During the short-term pre-test (10 min), mice in all groups showed no clear preference for any of the solutions (between group effect F[2,12]=2.8, p=0.1). B. During the short-term post-test (10 min), i.e. after 24 hour exposure to the two solutions, Sham and SDV mice developed a strong preference for the 1% NaCl solution, whereas bypassed mice did not (between group effect F[2,12]=6.8, p=0.01; vs. sham t[8]=3.6, Bonferroni *p=0.015; vs. SDV t[8]=3.4, **p=0.018). C. During long-term choice tests (24hs), bypassed, control and vagotomized mice failed to prefer the NaCl solution (F[2,12]=1.2, p=0.3). D. Actual consumed volumes of each solution (in mL) during the 24hs tests.

Next, the three groups of animals were exposed to a new short-term test after 24hs exposure to the two solutions. It is now clear that both the SDV and Control groups developed a strong preference for the 1% NaCl solution, whereas the DJB did not (Figure 5B). Based on Figure 5C it is not possible to anticipate these patterns from the preferences shown during the 24hs test. Figure 5D shows the actual volumes consumed of each solution by the three groups of mice during the 24hs choice test shown in Figure 5C.

Discussion

We showed that mice sustaining either duodenal-jejunal bypass (DJB) or total subdiaphragmatic vagotomy (SDV) display selective changes in nutrient preferences. Compared to sham-operated controls, these mice exhibited abnormal preferences for higher-calorie lipids, and failed to form normal preferences for glutamate and, for the case of DJB, sodium chloride. Preferences for nutritive sweet solutions were largely preserved. Deficits tended to appear only after longer-term exposure to the nutritive solutions, thereby implying a post-ingestive origin for the preference formation defects.

Both SDV and DJB mice displayed altered preferences for lipid emulsions, preferring to consume the higher-calorie solution unlike Sham animals. In other words, intact (Sham) animals prefer to consume a less caloric, but equally nutritive, lipid solution instead of more caloric and therefore potently satiating ones. Note that because their initial preferences during short-term tests were similar to those observed in sham-operated mice, this pattern does not appear to have derived from orosensory abnormalities. Rather, they are consistent with the notion that both DJB and SDV mice have deficits in sensing the presence of lipid calories in the gastrointestinal tract. In support of this notion, mice exposed to chronic high-fat food, which display both duodenal and vagal alterations, do also display abnormally higher preferences for high-content lipid solutions, a defect normalized by restoring duodenal-vagal signaling in these mice(Tellez, Medina, et al., 2013).

Because we employed truncal vagotomies, we cannot completely rule out the possibility that the abovementioned vagotomy-induced effects were due to the lack of vagal efferent tone on gastrointestinal organs. Two arguments however favor the idea that the deficits may have been independent from deficient vagal efferent tone. First, a recent report from our group demonstrated that, in flavor-nutrient conditioning assays, ablation of vagal gut-innervating nodose (afferent) neurons is sufficient to abolish learned flavor preferences for intra-gastric IntraLipid(Han, et al., 2018). Moreover, the same effects were observed in the present study when mice were tested for monosodium glutamate preferences (see Figure 4). Presumably, digestion of free glutamate is much less heavily dependent on vagal efferent activity; however, the deficits in learned glutamate preferences produced by vagotomies were as robust as those observed during lipid tests.

Unlike the case for fat emulsions, DJB and SDV mice showed preference formations for glucose solutions that were statistically undistinguishable from those observed in sham controls. This finding is consistent with previous findings that vagotomies do not disrupt flavor preferences conditioned via intra-gastric infusions of glucose(Sclafani & Lucas, 1996). Moreover, these results also corroborate the sufficiency of jejunal glucose sensing for the same flavor-glucose conditioning via intra-gastric infusions(Ackroff, et al., 2010). More recently, it has been shown that duodenal bypasses do not prevent the ability of mice to discriminate (and prefer) flavors conditioned to intra-gastric metabolizable versus non-metabolizable glucose(Zhang, Han, Motta, et al., 2018). Previous studies also show that reward circuits in brain respond to intra-portal mesenteric infusions of metabolizable glucose(Delaere, Akaoka, De Vadder, Duchampt, & Mithieux, 2013; Zhang, Han, Motta, et al., 2018), suggesting post-absorptive sites of action for glucose preference formation.

Our findings regarding the formation of preferences for monosodium glutamate are consistent with the requirement for vagal signaling in flavor preferences conditioned via intra-gastric infusions of monosodium glutamate(Uematsu, et al., 2010). Noticeably, these findings cannot be accounted for simply by defects in sodium preference formation: In fact, and rather unexpectedly, DJB mice, but not SDV mice, failed to form normal preferences for sodium chloride solutions. While in principle this could be attributed to lowered absorption rates due to the duodenal exclusion, it is in apparent conflict with the rich expression of sodium transporters in jejunum(Seidler, et al., 2008). Therefore, given the scarce literature on sodium preference formation, we cannot rule out the possibility that changes in sodium preferences after duodenal exclusions were due to factors unrelated to post-ingestive signals including adaptions associated with oral sodium gustatory receptors(Simon & de Araujo, 2005), including the overcoming taste neophobia. Future investigations are needed to determine the extent to which sodium sensing is altered by duodenal exclusions rather than by alternative, orosensory adaptations.

We believe our findings are important to the extent that they unify the physiologies of nutrient preferences and digestion. On one hand, pre-absorptive duodenal and vagal signals are critical for triggering the release of pancreatic juices into duodenum that allow for the appropriate absorption and utilization of ingested lipids and proteins(Schwartz, Berkow, McHugh, & Moran, 1993; Wang, et al., 2008). In contrast, no putative “gut factors” need be invoked to explain central responses to glucose consumption, as portal-mesenteric glucose infusions fully recapitulate the physiological effects produced by oral glucose loads(Bergman, Beir, & Hourigan, 1982). Furthermore, and consistently, such portal-mediated physiological effects on glucose homeostasis are not affected by subdiaphragmatic vagotomies(Bohland, et al., 2014). In fact, whereas inconsistencies exist regarding the sufficiency of glucose portal infusions to support sugar preferences(Ackroff, et al., 2010; Oliveira-Maia, et al., 2011), the same portal infusions are known to robustly activate reward dopaminergic systems(Zhang, Han, Lin, Li, & de Araujo, 2018). In sum, a combination of gastrointestinal and mesenteric-portal physiological signals(Sclafani & Ackroff, 2019) indicating the initiation of digestive processes may constitute the ultimate food reward signal to brain.

Conclusions

Duodenal-jejunal bypass interventions and subdiaphragmatic vagotomies alter food preferences in a nutrient-selective manner. Such changes did not depend on weight loss per se, nor did they derive from gustatory deficits. Instead, our data suggest that nutrient-specific processes involved in digesting and absorbing fats, amino acids and glucose overlap with those mediating learned preferences for these nutrients.