Dietary Experience with Glucose and Fructose Fosters Heightened Avidity for Glucose-Containing Sugars independent of TRPM5 Taste Transduction in Mice

Abstract

Objective

Experience with metabolically distinct sugars, glucose and fructose, enhances attraction to the orosensory properties of glucose over fructose. To gain insight into which sensory signals are affected, we investigated how this nutritive learning reshapes behavioral responding to various sugars in brief access taste tests in C57BL6/J (B6) mice and assessed whether sugar-exposed mice lacking the TRPM5 channel involved in G-protein coupled taste transduction could acquire these types of preferences for glucose-containing sugars.

Methods

B6, TRPM5 knockout (KO), and TRPM5 heterozygous (Het) mice were given extensive access to water (sugar naïve) or 0.316, 0.56, and 1.1 M glucose and fructose (sugar-exposed) and then tested, whilst food deprived, for their relative avidities for glucose, fructose, sucrose, maltose, and/or a non-metabolizable glucose analog in a series of taste tests.

Results

Sugar-exposed B6 mice licked relatively more for glucose than fructose, driven by an increased avidity for glucose, not an avoidance of fructose, and licked more for maltose, compared to their sugar-naïve counterparts. Sugar-exposed B6 mice did not lick with such avidity for a non-metabolizable glucose analog. TRPM5 KO mice took longer to acquire the sugar discrimination than the Het controls, but both groups ultimately licked significantly more for glucose than fructose. Het mice displayed clear preferential licking for sucrose over fructose, while licking comparably high for glucose, sucrose, and maltose. KO mice licked significantly more for maltose than sucrose.

Conclusions

Collectively, the findings suggest that ingestive experience with glucose and fructose primarily reprograms behavioral responding to a TRPM5-independent orosensory signal generated by glucose-containing sugars.

INTRODUCTION

The capacity to locate, ingest, and efficiently assimilate glucose is key to health and survival for many organisms. In humans, rodents, and most other mammals, this process begins with specialized chemosensors in the gustatory system that permit the rapid detection of carbohydrates and guide adaptive responding. Humans and rodents can perceive two major dietary carbohydrate classes, maltodextrins (long chains of glucose) and simple sugars (14, 15, 17, 21, 23, 28, 31, 39, 43). The heterodimeric G-protein coupled taste receptor—T1R2+T1R3—is considered the main conduit for oral sugar sensing. It binds all the simple sugars, including the glucose molecule, low-calorie sweeteners, and some D-amino acids. Its consequent cascade of neural events ultimately gives rise to the quintessential palatable sensation humans refer to as “sweet.” Without the T1R2 and/or T1R3, mice retain a strong attraction to maltodextrins, but are considerably less motivated to ingest sugars and other sweeteners (1, 2, 7, 12, 13, 18–20, 42). However, interestingly, loss of sweet reception does not render mice completely aguesic to sugar and rodents are capable of rapidly discriminating among certain sugars (2, 4, 5, 8, 12, 22, 25, 27, 32, 33, 36, 40). Collectively, these findings suggest that there may be even more oral carbohydrate sensory mechanisms left to uncover.

For example, Schier et al (25, 27) recently demonstrated that dietary experience (e.g., daily 30 minutes access over 18–30 days) with metabolically distinct sugars, glucose and fructose, renders rats and mice more attracted to the orosensory properties of glucose over those of fructose, a phenomenon that does not require functional sweet receptors. These findings strongly suggest that experience with the orosensory properties and postoral consequences of these two sugars retunes the gustatory system to respond more positively to glucose via a T1R-independent pathway. The mechanism(s), however, are still unknown. Many proteins associated with sugar transport and metabolism are localized in taste cells (29, 34). To begin to narrow the search for which, if any, of these mechanisms may contribute to the acquired appeal for glucose, Schier et al (27) assessed whether sugar exposure altered licking responses to related ligands, including the other monosaccharide, galactose. Galactose was of particular interest because it is brought inside the cell via the same transporters as glucose, including the sodium-glucose linked transporter (SGLT1). Indeed, a recent study found that pharmacological blockade of the SGLT1 greatly diminished licking responses for glucose solutions in naive mice also lacking the Tas1R3 gene (40). However, in the Schier et al (27) study, sugar-exposed rats did not respond any more positively to galactose than their naïve counterparts did, suggesting that something else specific to glucose and/or fructose must underlie the acquired taste following sugar exposure.

These published and preliminary findings leave open two, not mutually exclusive, main possibilities. First, sugar exposure specifically affects a signal related to fructose, which permits rats and mice to discriminate it from other sugars. Second, sugar exposure affects a signal that is specifically engaged by glucose and other glucose-containing sugars. The purpose of the present study was to determine how sugar exposure changes the relative behavioral responsivity to glucose, fructose, and other sugars to further refine the search for sensor mechanism. In Experiment 1, naïve and sugar-exposed C57BL6/J mice underwent a series of brief access tests with varying concentrations and formulations of glucose and fructose to test the hypothesis that experience with glucose and fructose specifically alters responding to glucose and glucose-containing sugars. Sclafani et al (29) recently showed that while sweet taste receptors are not required to detect glucose even in naïve mice, the transient receptor potential channel 5 (TRPM5) is. TRPM5 is also implicated in polysaccharide preference, at least at low concentrations (25). TRPM5 is expressed in Type II taste cells, downstream of T1R2+T1R3 and other G-protein coupled receptors, where it plays an important role in transducing the cellular signal (16). An earlier study indicates that the sweet receptors per se are not required, but it is still unclear if the responsible sensory signal would engage downstream transduction components via an alternative receptor. This is of special interest because many of the candidate sugar transporters and/or metabolic sensors are highly expressed in Type II cells. Therefore, Experiment 2 assessed whether the acquired responsivity to glucose and/or other glucose-containing sugars requires a TRPM5-mediated signal in TRPM5 KO mice.

MATERIALS & METHOD

Experiment 1

Subjects.

Sixteen male C57BL/6 (B6) (Strain 000664, Jackson Laboratory, Bar Harbor, ME, USA) mice, approximately 10 weeks old at the start of the study, were singly housed in ventilated polycarbonate cages with ad libitum access to deionized water (dH₂O) and rodent chow (Purina #5053), except when food and/or water restriction was used for experimental purposes (see below). All mice were provided Nestlets (Ancare) in the home cage. The housing room was climate-controlled and maintained on a 12:12 hour light: dark cycle (lights on at 06:00). All procedures were approved by the University of Southern California’s Animal Care and Use Committee and conducted in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals guidelines.

Stimuli.

Reagent grade glucose (D-dextrose), fructose, sucrose, maltose and Methyl alpha-D-glucoside (αMDG) were purchased from Sigma Aldrich (St. Louis, MO). All sugars were dissolved in deionized water (dH₂O) as needed each day and presented at room temperature. Corn oil emulsions were made by mixing 4.5% (v/v) corn oil (Signature Select Brand) with 0.6% of an emulsifying agent (Sunflower Lecithin, NOW Brand) dH₂O in emulsifying blender, just prior to use as needed and were presented at room temperature.

Apparatus.

Taste tests were conducted in a Davis Rig (Davis MS-160; DiLog Instruments, Tallahassee, FL). The rig is illustrated in Figure 1A. It consists of three Plexiglas walls and a stainless-steel front wall, with a grid floor below. On the other side of the front wall is a motorized table, which can hold up to 16 sipper tubes connected to a contact lickometer. A slot is located in the center of the front wall, whereby access to a sipper tube is granted via a computer-controlled shutter door. Licks to the sipper are registered through a high frequency AC circuit, timestamped, and stored for offline analyses. For single access training sessions, each mouse had access to one sipper tube for 20 minutes, while licks were recorded. During brief access sessions, each mouse is presented with a series of different solutions in short duration trials (10-s) in randomized blocks without replacement, as described previously (22). The brief access training and test sessions lasted 20 minutes; the mice were free to initiate as many trials as possible during that time.

Figure 1.

A: Image of the Davis Rig, Brief Access Taste Test. B: Median ± SIQR lick scores for glucose and fructose, presented in randomized order of brief access trials for naïve and sugar-exposed (Sug-Exp) C57BL6/J mice. C: Median ± SIQR lick scores for sucrose and fructose brief access trials. D: Median ± SIQR (fructose:sucrose) discrimination index by group [U = 23, p = 0.61]. E: Median ± SIQR lick scores for sucrose and glucose brief access trials. F: Median ± SIQR (glucose:sucrose) discrimination index by group [U = 5, p = 0.006]. G: Median ± SIQR lick scores for sucrose and fructose + glucose mixture brief access trials. H: Median ± SIQR (glucose+fructose mixture:sucrose) discrimination index by group [U = 10, p = 0.04]. I: Median ± SIQR lick scores for sucrose and maltose brief access trials. J: Median ± SIQR (maltose:sucrose) discrimination index by group [U = 10, p = 0.04]. Sample sizes were n = 8/group for all tests; one statistical outlier was excluded from the discrimination scores from the Sugar-exposed group. The asterisks indicate BH FDR-corrected significant pairwise differences.

Lickometer training and sugar exposure procedures.

To train the animals to lick at the sipper spout in the Davis Rig, overnight water restricted mice were given access to a single sipper spout containing dH₂O for a 20-minute session. This was repeated on the following 2–3 days, until all mice took at least 800 licks of water in the 20-minute session. Then, all mice were given access to a 4.5% corn oil emulsion, for a single 20-minute training session. Corn oil emulsion was used as a training stimulus in order to show the mice that a caloric substance can be obtained from the sipper spout, without exposing them to sugar or other “sweet” tasting things. Home cage water was returned approximately 30 minutes after this session. Three days later, overnight food deprived mice were returned to the Davis Rig for a training session in which the 4.5% corn oil emulsion was offered for 20 minutes. Chow was returned approximately 30 minutes after the session, and this training session was repeated a second time one day later. After a short repletion break, food-deprived mice were trained to lick in successive short duration trials (10-s) for 4.5% corn oil emulsion in preparation for brief access testing. In between each trial, the shutter closed for 7.5 seconds.

One week later, all mice went back on a chronic water restriction schedule whereby fluid was available for 20 minutes per day in the Davis Rig. Mice that dropped below 85% of their ad libitum body weight on this schedule were supplemented with 1–2 ml of dH₂O in the home cage after the day’s fluid access session. On the first two days, the fluid presented in the Davis Rig was 4.5% corn oil emulsion. Then the following day, the sugar exposure protocol began for half of the mice. On a given day, access to either glucose or fructose, at one of three concentrations (0.316, 0.56, and 1.1 M) was provided during the 20-minute session. These six sugar solutions were presented one time in a randomized order across 6 days. This was repeated two more times for a total of three blocks (18 days). On the last two days of the third block (0.56 M fructose and 0.56 M glucose, respectively), the solution was presented in serial brief access trials (10-s) to re-acclimate mice to the opening and closing of the shutter door. The other half of the mice was kept sugar naïve and instead had access to dH₂O in the Davis Rig during these exposure sessions. To allow mice to maintain a healthy body weight on this schedule, a 1–2 day homecage water repletion break was randomly interposed between access sessions across the 18-day exposure phase.

After the sugar exposure protocol, mice were then tested, under water restriction, for their relative lick responses to dH₂O, and the three concentrations (0.316, 0.56, and 1.1. M) of glucose and fructose in a 20-minute brief access taste test. In this test, the different solutions were presented to the mouse in randomized blocks, without replacement, in short (i.e., 10-s) trials. Prior to each trial, the designated sipper tube was positioned in front of the access slot, the shutter opened, and the trial commenced when the mouse made contact with the sipper spout (i.e., the first lick). Ten seconds later, the shutter closed, and the table repositioned to place a different tube in front of the access slot. Mice were free to initiate as many trials as possible during the 20 minute test session. Home cage water was returned 30 minutes after the test session. All mice were allowed to replete for 4 days. Then, chow was removed from the home cage and mice were given a second identical brief access probe test. Next, all mice were shifted onto a chronic food restriction schedule to achieve 85% of their ad libitum body weight through daily chow rationing. Once at the target weights, mice began a series of brief access taste tests as follows. Test 1: dH₂O, 0.316, 0.56, and 1.1. M glucose and fructose; Test 2: dH₂O, 0.316, 0.56, and 1.1. M fructose and sucrose; Test 3: dH₂O, 0.316, 0.56, and 1.1. M glucose and sucrose; Test 4: dH₂O, 0.316, 0.56, and 1.1. M sucrose and equivalent mixtures of glucose and fructose to match sugar concentration; Test 5: dH₂O, 0.316, 0.56, and 1.1. M sucrose and maltose. Test 6: dH₂O, 0.316, 0.56, and 1.1. M glucose and fructose; Test 7: dH₂O, 0.316, 0.56, and 1.1. M αMDG and fructose. Each successive test was separated by at least one day.

Experiment 2

Subjects.

Mice lacking the TRPM5 gene (−/−; KO, n = 3 ♂; n = 3 ♀) and their heterozygous (+/−; Het; n = 2 ♂; n= 8 ♀) counterparts were singly housed under the same conditions as mice in Experiment 1. Genotypes were confirmed at the end of the study. The original breeders were generously provided by Dr. Emily Liman (University of Southern California, Los Angeles, CA). All procedures were approved by the University of Southern California’s Animal Care and Use Committee and conducted in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals guidelines.

Stimuli and Apparatus.

Same as Experiment 1.

Lickometer training and sugar exposure procedures.

Experiment 2 was run in two successive cohorts, with slight variations in exposure procedures; these are indicated below. Initially, mice in cohort 1 were trained and given sugar exposure on a near identical schedule to that used in Experiment 1. However, intakes were somewhat lower than for the sugar-exposed B6 across this phase, and sugar-exposed TRPM5 KO failed to display a significant enhancement in licking for glucose relative to fructose during the post-exposure brief access probe tests. Thus, to promote more intake and, therefore, more experience with the sugar, all mice were given sugar access in the home cage for approximately 23 hours each day. One sugar solution was provided each day and the order of presentation was randomized across 6-day blocks. Water was removed from the home cage during the first three blocks to promote sugar intake. One day water repletion breaks were interposed between each block. Then chow was removed prior to the fourth and final sugar exposure block. Each day, mice were provided with one of the six sugar solutions and a chow ration, aimed at maintaining the mice at 85% of their ad libitum body weight. This was done to encourage more experience with differential postingestive metabolic effects of the two sugars. After this block, mice were kept at 85% body weight through chow rationing and given a single 20 minute brief access reminder training session in the Davis Rig, wherein 4.5% corn oil emulsion was presented in successive 10-s trials. The next day, mice were re-tested for their relative licking responses to dH₂O, and the three concentrations of glucose and fructose in a brief access taste test. While mice were maintained at 85% body weight, they were given the Test 1 through Test 5 series of brief access tests with various sugars as described for Experiment 1, with at least one day off between each test.

The schedule of training and sugar exposure was similar for the second cohort. After training to lick at the spout while water-restricted for water, mice underwent two sugar exposure blocks followed, after a short repletion break, by a single block whilst food restricted to 85% ad libitum weight through chow rationing. These blocks were run in the Davis Rig (20-minute session; 1 solution/session). After this, all mice had a training session in which access to 0.56 M glucose was offered in brief access trials (10-s) and then a brief access probe test with dH₂O, and the three concentrations of glucose and fructose. Next, the mice were then given another sugar exposure block in the Davis Rig whilst food restricted followed by a second brief access probe test. Ad libitum access to chow was reinstated after this test. Trial initiation and licking behavior was very low on the probe tests and, in fact, intake was generally low across these single access sugar exposure sessions, particularly for the KO mice, which may have prevented sufficient exposure to the orosensory and postingestive effects of the two sugars. Thus, as with cohort 1, to promote more intake of the sugar solutions, especially in the TRPM5 KO mice, water was removed from the home cage and mice underwent three exposure blocks, in which one sugar solution was provided in the home cage as the only source of fluid for 23 hours per day. After a repletion break, all mice were returned to 85% body weight through rationing and administered a final sugar exposure block in the home cage. This was followed by a single brief access reminder training session in the Davis Rig (10-s trials; 4.5% corn oil emulsion) and a brief access test with dH₂O, and the three concentrations of glucose and fructose. While mice were held at 85% body weight, they were given a series of brief access tests with various sugars, as described for cohort 1 and Experiment 1, with at least two days between each test.

Data Analyses

To compare relative licking responses towards sugar solutions presented in brief access trials, the total licks per stimulus was divided by the total number of trials completed for that stimulus and then standardized against licks to water (total licks to water divided by the number of water trials completed) on the same test for each mouse as follows:

$$\mathit{\text{Lick Score}}=\mathit{\text{licks to stimulus}}-{\mathit{\text{licks to dH}}}_{2}O$$

Because sample sizes were <20 and some sample distributions were non-normal (see Experiment 2 Results), non-parametric statistics were used to analyze data from both experiments. For each group at each test, lick scores were compared across the two sugars at each concentration with Wilcoxon matched-pairs signed rank tests. Latency to initiate glucose versus fructose brief access trials was analyzed at each concentration with Wilcoxon matched-pairs signed rank tests. Multiple comparisons were corrected with the Benjamini and Hochberg False Discovery Rate (BH FDR) method. In Experiment 1, a sugar discrimination index score was calculated by dividing the overall lick score for one sugar (collapsed across concentration) by the overall lick score to the comparison sugar (sucrose, collapsed across concentration). Accordingly, a discrimination index score nearing 0 indicates that mice were licking substantially more for sucrose over the comparison sugar and a score approaching 1 indicates that mice were licking comparably for sucrose and the comparison sugar. Mann Whitney tests were used to compare sugar discrimination scores between the two exposure groups on each brief access test. Overall licks to water on each test were compared across groups in each respective experiment with Mann Whitney tests. In Experiment 2, male and female mice were used, but due to small sample sizes within each sex, sex was not included as a factor in any analysis. P values of ≤ 0.05 were considered significant. The main statistical outcomes are presented in Tables 1–3.

Table 1:

Experiment 1 Statistical Outcomes for Lick Scores

	0.316 M	0.56 M	1.1 M
Glucose vs Fructose
Sugar Naïve	W = −10, p = 0.55	W = 14, p = 0.38	W = 36, p = 0.008
Sugar-Exposed	W = −36, p = 0.008	W = −36, p = 0.008	W = −36, p = 0.008
Sucrose vs Fructose
Sugar Naïve	W = −36, p = 0.008	W = −36, p = 0.008	W = −36, p = 0.008
Sugar-Exposed	W = −28, p = 0.054	W = −30, p = 0.039	W = −32, p = 0.016
Sucrose vs Glucose
Sugar Naïve	W = −36, p = 0.008	W = −36, p = 0.008	W = −36, p = 0.008
Sugar-Exposed	W = −8, p = 0.64	W = −36, p = 0.008	W = −36, p = 0.008
Sucrose vs Fruc+Gluc
Sugar Naïve	W = −32, p = 0.02	W = −32, p = 0.016	W = −32, p = 0.016
Sugar-Exposed	W = −36, p = 0.008	W = −36, p = 0.008	W = −32, p = 0.02
Sucrose vs Maltose
Sugar Naïve	W = −28, p = 0.054	W = −36, p = 0.008	W = −36, p = 0.008
Sugar-Exposed	W = 16, p = 0.31	W = −30, p = 0.039	W = −24, p = 0.11
Glucose vs Fructose
Sugar Naïve	W = −10, p = 0.53	W = −4, p = 0.84	W = −18, p = 0.25
Sugar-Exposed	W = 8, p = 0.64	W = −34, p = 0.016	W = −32, p = 0.02
αMDG vs Fructose
Sugar Naïve	W = 10, p = 0.55	W = −14, p = 0.38	W = 34, p = 0.016
Sugar-Exposed	W = 32, p = 0.02	W = 12, p = 0.46	W = 36, p = 0.008

Notes. Wilcoxon Signed Rank Tests at each concentration tested (W = sum of ranks). P values correspond to the BH FDR-corrected probabilities for each pairwise comparison.

Table 3:

Experiment 2 Statistical Outcomes for Lick Scores

	0.316 M	0.56 M	1.1 M
Glucose vs Fructose
Het	W = −55, p = 0.002	W = −55, p = 0.002	W = −55, p = 0.002
KO	W = - 7, p =0.44	W = 21, p = 0.045	W = 21, p = 0.045
Sucrose vs Fructose
Het	W = −45, p = 0.004	W = −45, p = 0.004	W = −45, p = 0.004
KO	W = −13, p = 0.13	W = −15, p = 0.095	W = −15, p = 0.095
Sucrose vs Glucose
Het	W = −33, p = 0.16	W = 7, p = 0.73	W = −15, p = 0.65
KO	W = - 9, p = 0.81	W = −3, p = 0.81	W = 3, p = 0.81
Sucrose vs Fruc+Gluc
Het	W = −11, p = 0.86	W = −21, p = 0.75	W = −2, p =0.95
KO	W = −9, p = 0.44	W = −9, p = 0.44	W = −7, p = 0.44
Sucrose vs Maltose
Het	W = 33, p = 0.17	W = 6, p = 0.76	W = −7, p = 0.76
KO	W = 21, p = 0.05	W = 21, p = 0.05	W = 17, p = 0.09

Notes. Wilcoxon Signed Rank Tests at each concentration tested (W = sum of ranks). P values correspond to the BH FDR-corrected probabilities for each pairwise comparison.

RESULTS

Experiment 1

Consistent with previous findings (25), sugar naïve B6 mice licked comparably for glucose and fructose at the 0.316 and 0.56 M concentrations, and, licked significantly more for fructose than glucose at the highest concentration (1.1 M) (Test 1; Figure 1B; Table 1; see also Supplemental Figure 1). By contrast, sugar exposed B6 mice licked significantly more for glucose than fructose at each of the concentrations tested (Test 1; Figure 1B, Table 1).

Schier et al (25) speculated that sugar-exposed mice may be able to use an olfactory cues to discern glucose from fructose. Although loss of olfactory input attenuated the exposure-based avidity for glucose, it did not completely abolish it, leaving open the possibility that other oral chemosensors are involved. Nevertheless, to probe if mice in the current study were utilizing an orthonasal olfactory cue to guide responding, we analyzed latency to initiate glucose and fructose trials on this test. No differences in the latency to initiate glucose and fructose trials were detected at any concentration in either group (Table 2).

Table 2.

Latencies to Initiate Glucose and Fructose Trials

Group	0.316G	0.316F	0.56G	0.56F	1.1G	1.1F	P values
Sugar Naïve B6	2.05 (2.30)	3.11 (1.82)	3.27 (1.37)	4.17 (2.10)	5.85 (2.93)	2.41 (1.79)	P = 0.95, P = 0.69, P = 0.33
Sugar Exposed B6	5.01 (4.02)	5.73 (4.84)	2.81 (3.71)	5.16 (12.46)	4.08 (7.85)	3.80 (4.29)	P = 0.95, P =0.95, P = 0.95
Sugar Exposed Het	1.82 (21.04)	3.36 (5.89)	1.75 (8.55)	6.55 (6.87)	4.49 (3.06)	6.18 (4.10)	P = 0.99, P = 0.20, P = 0.20
Sugar Exposed KO	27.43 (14.98)	13.44 (19.84)	10.80 (15.87)	39.52 (32.42)	22.98 (26.70)	41.08 (58.46)	P = 0.56, P = 0.47, P = 0.47

Notes. Median (SIQR) values in seconds are shown by molar glucose (G) and Fructose (G) concentrations. P values correspond to the BH FDR-corrected probabilities for each pairwise comparisons at each of the 3 concentrations, in ascending order.

Next, to assess how dual sugar exposure alters responsivity to other sugars, all mice were tested for their relative licking responses to the glucose-fructose disaccharide, sucrose, and equimolar fructose. Equimolar concentrations were selected to match the available amount of fructose at each concentration, though it is important to note that total sugar content is approximately doubled for sucrose at each concentration. Not unexpectedly then, naïve mice licked more for sucrose at each concentration presented (Figure 1C, Table 1). Sugar exposed mice displayed a similar pattern of licking across the sugar concentrations, though only licked significantly more for sucrose at the two highest concentrations (Figure 1C, Table 1). In fact, when calculated as a ratio of licks to sucrose, naïve and sugar-exposed mice displayed comparable relative levels of licking for fructose (Figure 1D). The next test pitted sucrose against equimolar glucose. Here, although both groups licked significantly more for sucrose solutions (Figure 1E; Table 1), the sugar-exposed mice had a higher discrimination index than the naïve mice (Figure 1F). This suggests that even though sucrose was relatively more appealing than glucose at these concentrations, the enhanced responsiveness to glucose conferred through sugar exposure continued to promote more consummatory responses to the glucose monosaccharide in this hedonic discrimination test.

The next test assessed whether mice responded differentially to bound glucose and fructose in the form of sucrose versus the same amount of liberated glucose and fructose (50:50 mixture). Interestingly, we found that naïve and sugar-exposed mice licked significantly more for sucrose over the glucose-fructose mixture at all concentrations tested (Figure 1G; Table 1). However, sugar-exposed mice licked relatively more for the mixture; that is, the overall discrimination ratio was significantly higher than that of the naïve control group (Figure 1H). Finally, we tested if sugar-exposure altered responses to a minimally sweet glucose-glucose disaccharide, maltose. As expected, naïve mice licked significantly more for sucrose over maltose at the mid and high concentrations (Figure 1I; Table 1). Overall, sugar-exposed mice licked more similarly for sucrose and maltose, particularly at the low and high concentrations (Figure 1I). Consistent with this, sugar exposed mice displayed a significantly higher discrimination score for maltose, meaning that they licked in a more comparable fashion for the two sugars than did their naïve counterparts (Figure 1J).

Finally, previous studies have suggested that SGLT1/3 may function as part of a glucosensor in taste cells (4, 40). Thus, we asked whether sugar-exposed mice generalize their avidity for glucose to a non-metabolizable ligand for the SGLT-1/3 transceptor, Methyl alpha-D-glucoside (αMDG). We replicated the basic glucose versus fructose test and then conducted the same test except that αMDG was offered in place of glucose. As in test 1, sugar naïve mice licked comparably to glucose and fructose. Sugar-exposed mice licked significantly more for 0.56 and 1.1. M glucose than equimolar concentrations of fructose (Figure 2A; Table 1). When αMDG was substituted, the basic pattern was recapitulated in sugar naïve mice, with mice licking similarly for both sugars, except at the 1.1 M concentration (Figure 2B, Table 1). The pattern was not reiterated, however, for the sugar exposed mice. Here, sugar-exposed mice licked significantly more for fructose than for αMDG, at least at the low and high concentrations tested (Figure 2B; Table 1).

Figure 2.

A: Median ± SIQR lick scores for glucose and fructose, presented in randomized order of brief access trials for naïve and sugar-exposed (Sug-Exp) C57BL6/J mice. B: Median ± SIQR lick scores for αMDG and fructose brief access trials. Sample sizes were n = 8 per group on both tests. The asterisks indicate BH FDR corrected significant differences.

Experiment 2

The goal of this experiment was two-fold. First, we tested if sugar-exposed mice lacking the TRPM5 channel (KO) were capable of behaviorally discriminating glucose from fructose in a brief access taste test. After 3–4 limited sugar exposure access blocks, the sweet-sensitive TRPM5 heterozygous mice (Het) licked significantly more for glucose over fructose at the low and mid concentrations in probe test conducted after an acute overnight fast (Figure S2). KO mice, on the other hand, failed to behaviorally discriminate between the two sugars on this probe test (Figure S2). That being said, licks during the preceding sugar exposure sessions were quite low in the KO group, generally half of what their Het counterparts consumed. Therefore, all mice were given more extensive exposures (23 h/day) to the sugars in the home cage to maximize intake, and, therefore, learning about the orosensory and postingestive consequences of glucose and fructose.

A second probe test conducted after the long-term exposure blocks showed that like the Het mice (Figure 3A; Table 3), KO mice now displayed a heighted avidity for glucose relative to fructose, especially at the 0.56 and 1.1 M concentrations (Figure 3A; Table 3). There were no significant differences in the latency to initiate trials, at any concentration tested for the Het or KO mice (Table 2), though we noted considerable variability in latency to initiate trials for KO mice. Thus, we further investigated if the latency to initiate fructose trials as a ratio of latency to initiate glucose trials was correlated with the ability to discriminate glucose from fructose at each concentration tested in this group. Spearman’s rank tests did not detect any significant correlations between latency and licking responses in the KO mice (closest r = 0.25, p = 0.66, at 0.316 M).

Figure 3.

A: Median ± SIQR lick scores for glucose and fructose, presented in randomized order of brief access trials for sugar-exposed TRPM5 +/− (Het) (n = 10) and TRPM5 −/− (KO) (n = 6) mice. B: Median ± SIQR licks on water trials from each brief access taste test for Het and KO mice. C: Median ± SIQR lick scores for sucrose and fructose brief access trials for Het (n = 9) and KO (n = 5) mice. D: Median ± SIQR for sucrose and glucose brief access trials for Het (n = 9) and KO (n = 5) mice. E: Median ± SIQR lick scores for sucrose and fructose + glucose mixture brief access trials for Het (n = 9) and KO (n = 5) mice. F: Median ± SIQR lick scores for sucrose and maltose brief access trials for Het (n = 9) and KO (n = 6) mice. The asterisks indicate BH FDR-corrected significant differences.

The next series of tests were designed to determine if the sweet-blind mice displayed different response patterns to other sugars than their sweet-sensitive counterparts. In the test in which fructose was pitted against equimolar sucrose, Het mice licked significantly more for sucrose and not much at all to fructose. KO mice showed a similar lick pattern, and tended to lick more for sucrose over fructose, but these differences did not quite reach statistical significance (Figures 3C; Table 3). When this test was repeated with glucose in place of fructose, overall both Het and KO mice licked similarly to the two sugars. However, whereas Het mice increased licks to both sucrose and glucose as concentration increased (Figure 3D; Table 3), we noted that KO mice licked near baseline levels for both sugars and across all three concentrations (Figure 3D; Table 3). Close inspection of raw lick values revealed considerable variability in the response profile among the KO mice on this test. First, we noted that some of KO mice licked a lot for water (Figure 3E and 3B; see also Figure S2 for B6 water lick values). Het mice, by contrast, licked very little for water (Figure 3B). We further observed that whereas about half of the KO mice licked more for sucrose than glucose, the other half showed the opposite response pattern (Figure S3). A corresponding phenomenon was evident on the sucrose versus glucose-fructose mixture test. Het mice licked similarly to the sucrose and glucose+fructose mixture, with low variability (Figure 3E; Table 2). KO mice, on the other hand, showed a lot of variability in responding to sucrose versus the mixture (Figure 3D; Table 3). As shown in Figure S3, this was because approximately half of the KO mice licked more for the mixture; the other half licked more for sucrose. Nevertheless, nearly all of the KO mice were highly responsive to maltose, and, in fact, overall, licked more for maltose than sucrose, particularly at the 0.316 and 0.56 M concentrations (Figure 3F; Table 3). Het mice licked comparably for maltose and sucrose at each concentration tested.

GENERAL DISCUSSION

Sugars are highly effective at driving consumption and this begins by way of their initial contact with chemosensors in the oral cavity (6). Although it is widely known that the T1R2+T1R3 receptor is important for detecting and responding to various sugars and low calorie sweeteners, there are strong indications in the literature that other receptors may be involved in oral sugar sensing (9, 11, 34, 40). The present findings corroborate prior studies in showing that ingestive experience with two metabolically-distinct sugars, glucose and fructose, fosters a preference for the orosensory properties of glucose over those of fructose (10, 25, 27). Moreover, the present results help to home in on potential underlying sensory mechanisms involved in this acquired sugar discrimination. In earlier studies, it remained unclear if the glucose preference was related to a primary change in responding to signals generated by glucose, fructose, or both sugars. Here, we found that, when tested against equimolar sucrose, sugar-exposed B6 mice responded similarly to fructose as naïve mice. Not only that, but when sugar-exposed mice were offered fructose against a non-metabolizable glucose analog, they licked significantly more for fructose solutions. Collectively, these findings strongly suggest that sugar-exposed B6 mice are not avoiding fructose. Rather, we found that B6 mice licked relatively more whenever free glucose was included in the substrate, than their naïve counterparts, suggesting that responsivity to glucose underwent a significant shift during sugar exposure. Interestingly, sugar-exposed B6 mice also licked relatively more for a glucose-glucose disaccharide, maltose, than the sugar-naïve mice. Thus, the reprogramming is not limited to free glucose, per se (more on this below).

Recent work showed that SGLT1 may play a role in glucose sensing in the taste system in mice and humans (4, 40). Thus, we began to assess if SGLT1 was involved in the acquired glucose preference by simply replacing glucose with a non-metabolizable analog, αMDG, which can be shuttled via SGLT1 across the apical membrane and which has been shown to trigger this glucose transceptor, much like the real thing, in a brief access taste test. However, the results showed that B6 mice did not treat αMDG like glucose in this test. In fact, whereas the naïve B6 mice licked similarly to αMDG and fructose, supporting the view that this analog bears an inherently attractive taste, sugar-exposed mice, if anything, licked less for αMDG than fructose. This finding is consistent with previous results, which showed that rats do not come to respond more positively to galactose, another ligand for SGLT1, following glucose and fructose exposure (27). Thus, it does not appear as though SGLT1 is a likely candidate sensor underlying the behavioral phenomenon under study here, though future work will continue to explore conditions under which this sensor contributes to sugar taste.

Schier et al (25) showed that the canonical sweet receptor, T1R2+T1R3, is not required to learn to prefer the orosensory properties of glucose over fructose in this paradigm. However, considering many of the candidate sensor systems, including SGLT1 and K^ATP channels, are mainly expressed in Type II taste cells and a recent study showed that naïve TRPM5 KO mice are unresponsive to glucose solutions (29, 35, 36, 41), the present study tested whether TRPM5 KO, which are deficient in Type II taste cell transduction, are capable of responding to glucose after ingestive experience with glucose and fructose. We found that sugar-exposed KO mice were slower to acquire the glucose preference, as compared to their heterozygous counterparts. This very likely stemmed from low sugar intakes across the initial exposure phase. As such, the data are consistent with Sclafani et al (29)’s, showing that mice lacking the TRPM5 channel are not inherently attracted to sugars. However, with more extensive exposure, KO mice did come to express a glucose preference, just like their heterozygous counterparts. The learned response impacted responsivity to other sugars in both the Het and KO groups. Sugar-exposed Het mice consistently licked at high levels, on par with that for sucrose, when solutions containing free glucose or maltose served as the comparison stimulus. The generalization profiles were more variable for KO mice, at least on some tests. All or nearly all the sugar-exposed KO mice licked more for glucose than fructose, and more for sucrose than fructose, but the response patterns diverged when sucrose was pitted against glucose and the glucose + fructose mixture. In these cases, roughly half of the KO mice licked more for free glucose and the mixture, the other half preferred sucrose to free glucose or the mixture. These divergent responses did not appear to correspond to sex, though estrus stage at each test was not assessed and small sample sizes per sex precluded a full statistical analysis of sex as a factor. The bases for these differences remain unknown, but it was notable that irrespective of those, all sugar-exposed KO mice licked more for maltose than equimolar sucrose in the final test. Heterozygous mice licked comparably for maltose and sucrose. Thus, overall, the findings suggest that TRPM5-based signal transduction is not necessary to acquire an affinity for the taste of glucose over fructose and that, in the absence of TRPM5-dependent input, dual sugar experience confers a robust avidity for the normally less preferred glucose-glucose disaccharide, maltose.

In fact, that all sugar-exposed groups responded quite positively to maltose was striking. Unconditioned mice lacking the Tas1R2 and/or Tas1R3 receptor genes are virtually unresponsive to sucrose and maltose, suggesting that this sweet receptor plays an important role in driving appetite and consummatory responses for both sugars (3, 33, 37, 39). However, prior studies have also shown that rodents can discriminate the taste of maltose from that of sucrose, even when intensity is rendered irrelevant (33). Thus, it appears that maltose has both qualitatively similar and distinct features from the prototypical sweet stimulus, sucrose. An alternative “maltose” receptor has not been identified to date, but one possibility is that this type of receptor exists and binds low moiety glucose-containing sugars and signals arising from such a receptor are altered by the dual sugar exposure to enhance rapid detection of glucose-containing foods. Alternatively, maltose and glucose may be capable of binding to the presumptive polysaccharide taste receptor (21, 24, 38, 43), whose signal is likewise amplified through dietary experience. Interestingly, previous studies have suggested that TRPM5 is not required to express a preference for Polycose over the long term (30), but whether it is required for taste-guided behaviors towards this and other maltodextrins remains to be specifically tested. Moreover, future studies ought to assess whether ingestive experience with glucose and fructose modifies behavioral responsivity to polysaccharides. An alternative possibility is that more complex sugars, like sucrose and maltose, are rapidly cleaved into their monosaccharide constituents, which, in turn, proportionately activate a glucose-specific sensor. Indeed, Sukuraman et al (35) recently showed that glucosidases are expressed at the apical membrane of taste cells. Future work will be needed to more systematically investigate this hypothesis as well.

Prior work suggested that mice given extensive sugar exposure may be able to use subtle olfactory cues to differentiate amongst glucose and fructose (10, 26). Glucose and fructose are low volatile chemicals, with no known distinguishing olfactory features. Accordingly, Schier et al (25) previously speculated that extensive exposure to the two sugars may sensitize mice to odor cues associated with contaminants, if not the sugars themselves. Loss of olfactory input did not completely disrupt the ability to acquire a glucose preference in this paradigm, suggesting that, even if olfactory stimuli were present, taste may still play an important role (25). There are a few features of the present data that point to the contribution of non-olfactory mechanisms to this phenomenon. First, here, sugar-exposed mice did not show differential latencies to initiate glucose and fructose trials. This differs from previously published studies (10, 22), but the basis for the difference is unclear. It could be related to the chemicals used, presence or absence of contaminates, olfactory acuity of the mice, among other things. Second, mice all generalized their responding to other types of glucose-containing sugars, including glucose-fructose mixtures and maltose. It seems unlikely that odorants associated with free glucose would also be present and/or perceptible in these more complex formulations, though, of course, we cannot entirely exclude this possibility.

In sum, the present findings add to the emerging lines of evidence that animals are well tuned to the presence of glucose and glucose-containing foods in their food environment, and this includes at very first site of nutrient detection in the oral cavity. Dietary experience plays an important role in amplifying not only the ability to detect and respond to glucose in its pure form, but in other forms more common in nature, such as mixtures and complex sugars.

Supplementary Material

Supplementary Data Figures

Figure S1. A: Median ± SIQR lick scores for glucose and fructose, presented in randomized order of brief access trials for sugar naïve (n = 8) and sugar-exposed (n = 8) B6 mice on the initial probe test conducted after overnight food deprivation (ON FD). B: Median ± SIQR licks on water trials from each brief access taste test for sugar naïve and sugar-exposed mice. The asterisks indicate BH FDR-corrected significant differences.

Figure S2. A: Median ± SIQR lick scores for glucose and fructose, presented in randomized order of brief access trials for sugar-exposed TRPM5 +/− (Het) (n = 6) and TRPM5 −/− (KO) (n = 2) mice on the initial probe test conducted after overnight food deprivation (ON FD) in cohort 1. B: Median ± SIQR lick scores for glucose and fructose on the same test for cohort 2 sugar-exposed TRPM5 +/− (Het) (n = 3) and TRPM5 −/− (KO) (n = 2) mice. C: Median ± SIQR lick scores for glucose and fructose on the initial brief access test for cohorts 1 and 2 combined. The asterisks indicate BH FDR-corrected significant differences.

Figure S3. Lick scores for individual TRPM5 KO mice plotted across each brief access test. Sex is indicated as male (m) or female (f) for each mouse. Licks to water (dH₂O) on each test are indicated at the top left of each panel. Blank panels indicate not enough trials were taken to analyze.

DATA AVAILABILITY

The data that support the findings of this study are available from the corresponding author upon reasonable request.