Glucagon signaling modulates sweet taste responsiveness

Abstract

The gustatory system provides critical information about the quality and nutritional value of food before it is ingested. Thus, physiological mechanisms that modulate taste function in the context of nutritional needs or metabolic status could optimize ingestive decisions. We report that glucagon, which plays important roles in the maintenance of glucose homeostasis, enhances sweet taste responsiveness through local actions in the mouse gustatory epithelium. Using immunohistochemistry and confocal microscopy, we found that glucagon and its receptor (GlucR) are coexpressed in a subset of mouse taste receptor cells. Most of these cells also express the T1R3 taste receptor implicated in sweet and/or umami taste. Genetic or pharmacological disruption of glucagon signaling in behaving mice indicated a critical role for glucagon in the modulation of taste responsiveness. Scg5^−/− mice, which lack mature glucagon, had significantly reduced responsiveness to sucrose as compared to wild-type littermates in brief-access taste tests. No significant differences were seen in responses to prototypical salty, sour, or bitter stimuli. Taste responsiveness to sucrose was similarly reduced upon acute and local disruption of glucagon signaling by the GlucR antagonist L-168,049. Together, these data indicate a role for local glucagon signaling in the peripheral modulation of sweet taste responsiveness.—Elson, A.E.T., Dotson, C.D., Egan, J.M., Munger, S.D. Glucagon signaling modulates sweet taste responsiveness.

The gustatory and gastrointestinal (GI) systems share a number of common features. For example, several proteins involved in taste transduction, including T1R- and T2R-type “taste” receptors and the G-protein subunit α-gustducin, play chemosensing roles in both the mouth and gut (1–4). Similarly, several molecules implicated in gut nutrient responses and the regulation of metabolic homeostasis are expressed in the gustatory epithelium. Receptors for peptide hormones, including cholecystokinin (CCK), neuropeptide Y (NPY), vasoactive intestinal peptide (VIP), galanin, and leptin, are expressed in taste receptor cells (TRCs) of the mammalian taste bud (5–10), whereas the glucagon-like peptide-1 (GLP-1) receptor is found on intragemmal fibers of afferent taste nerves (11). With the apparent exception of leptin, all of the associated peptides are produced locally in subsets of TRCs (6–11). The presence of these peptide hormones in TRCs suggests a role in taste function, and several have been implicated in the modulation of taste responsiveness to sugars and other sweeteners. For example, leptin, a satiety signal secreted from fat cells, acts directly on leptin receptors in TRCs to suppress responsiveness to sweet-tasting stimuli (5, 12, 13). In contrast, GLP-1, an incretin hormone secreted from intestinal L-cells, is produced locally in TRCs and acts to enhance or maintain responsiveness to sweet stimuli (11). Therefore, the taste bud is well positioned as a site for functional modulation by peptide hormones.

GLP-1 and glucagon are both derived from proglucagon. Glucagon is a major counter-regulatory hormone with a primary site of action in the liver, where it stimulates glycogenenolysis and gluconeogenesis to prevent hypoglycemia. We previously reported that glucagon is expressed in a subset of taste cells of the tongue (11). However, it was unknown which TRCs express glucagon, whether local targets for glucagon signaling in the taste bud exist or whether glucagon can affect taste function. Here, we use immunohistochemistry and behavioral assays in mice to identify a role for glucagon signaling in sweet taste function.

MATERIALS AND METHODS

Mice

All animal procedures were approved by the University of Maryland School of Medicine animal care and use committee. Wild-type C57BL/6J mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA). Scg5 (7B2)^+/+ and Scg5^−/− gene-targeted mice (maintained on a C57BL/6J background; ref. 14, 15) were generously provided by Dr. Iris Lindberg (University of Maryland School of Medicine). All mice received food and water ad libitum, except when engaged in behavioral experiments (see below). Tissue from Scg5^−/− mice, GlucR^−/− mice (generously provided by Maureen Charron, Albert Einstein School of Medicine, New York, NY, USA; ref. 16), and T1r3^−/− mice (generously provided by Charles Zuker, Columbia University, New York, NY, USA; ref. 17)were used to control for antibody specificity (see below).

RT-PCR

CV-enriched tissue and surrounding taste tissue were obtained via micropunch (1 mm diameter; Harris Unicore, Ted Pella, Inc., Redding, CA, USA) from C57BL/6J mice. Total RNA was extracted with the Absolutely RNA microprep kit (Stratagene, La Jolla, CA, USA). RNA was reverse-transcribed with oligo(dT) primers, and cDNA was used as template for PCR amplification with TaqPro Complete (Denville Scientific, Metuchen, NJ, USA). Products were amplified with gene-specific primers (Table 1). Control samples were prepared with the reverse transcriptase omitted. To control for contributions from genomic DNA, all primer pairs were designed to span an intron (with the exception of the actin primers). All products were confirmed by sequencing.

Table 1.

Primers

Target	Forward primer, 5′–3′	Reverse primer, 5′–3′	Annealing temp (°C)	Cycles	Size (bp)
GlucR	TGCCCAGGTAATGGACTTTTTG	ACCGTGTCTTCAGCAGCCAATC	58	40	530
7B2	TCAGAGACAGACATCCAGAGGCTG	GCAGTTTTCCCAAGAGGACAGG	58	35	272
α-Gust	ATGGGAAGTGGAATTAGTTCAG	TCAGAAGAGCCCACAGTCTTTG	60	35	1064
Actin (51)	CCCTGTGCTGCTCACC	GCACGATTTCCCTCTTCAG	58	30	328

Immunohistochemistry

Mice were transcardially perfused with 4% paraformaldehyde (PFA), and tongues were removed. For detection of 5-hydroxy-l-tryptophan⁺ (5-HT⁺) TRCs, mice were injected intraperitoneally with 5-HT (0.08 mg/g; Sigma-Aldrich, St. Louis, MO, USA) 1 h prior to perfusion and removal of the tongues (11). After 2 h postfix in PFA at 4°C, tongues were cryoprotected in 30% sucrose overnight at 4°C. Cryostat sections (14 μm) were collected onto slides and immunostained overnight with up to 3 primary antibodies. Secondary antibodies were applied the following day. The following primary antibodies were used (see Table 2 for details and for specificity controls): rabbit anti-glucagon (1:500, Immunostar, Hudson, WI, USA); goat anti-GlucR (1:100; Santa Cruz Biotechnology, Santa Cruz, CA, USA); rabbit anti-7B2 (1:100; gracious gift of Dr. Iris Lindberg); rabbit anti-PLCβ2 (1:500; Santa Cruz Biotechnology); rabbit anti-gustducin (1:100; Santa Cruz Biotechnology); goat anti-T1R3 (1:50; Santa Cruz Biotechnology); and rat anti-5-HT (1:5000; Eugene Tech International, Allendale, NJ, USA). Primary antibodies were visualized with Cy2, Cy3, and Cy5 secondary antibodies (Cy2 and Cy3: 1:4000; Cy5: 1:1000; Jackson Immunoresearch, West Grove, PA, USA). GlucR and T1R3 were visualized via biotinylated donkey anti-goat secondary antibodies followed by streptavidin-conjugated CY2 (1:4000; Jackson Immunoresearch, West Grove, PA, USA). Images were collected on an Olympus confocal microscope using FluoView software (Olympus, Tokyo, Japan). Brightness and contrast levels of collected images were adjusted on Adobe Photoshop CS (Adobe Systems, San Jose, CA, USA).

Table 2.

Primary antibodies

Antibody	Host	Supplier	Dilution	Specificity/control
Anti-glucagon	Rabbit	Immunostar (Hudson, WI, USA; cat. no. 20076)	1:500	Staining absent when primary or secondary antibodies omitted (not shown), or in Scg5 (7B2)−/− mice (16) (Fig. 3). Use of this antibody has been reported previously (52, 53).
Anti-GlucR	Goat	Santa Cruz Biotechnology (Santa Cruz, CA, USA; sc-34638)	1:100	Staining absent when primary or secondary antibodies omitted (not shown), after preincubation with 1–10 mg of blocking peptide (sc-34638P; not shown), or in GlucR−/− mice (Fig. 1). Use of this antibody has been reported previously (54, 55).
Anti-7B2	Rabbit	Generous gift from Iris Lindberg (University of Maryland School of Medicine; 13BF)	1:100	Staining absent when primary or secondary antibodies omitted (not shown), or in Scg5 (7B2)−/− mice (Fig. 3).
Anti-α-gustducin	Rabbit	Santa Cruz (sc-395)	1:100	Staining absent when primary or secondary antibodies omitted (not shown); has been validated in taste tissue (56).
Anti-PLCβ2	Rabbit	Santa Cruz Biotechnology (sc-206)	1:500	Staining absent when primary or secondary antibodies omitted (not shown), and has been validated in taste tissue (56).
Anti-T1R3	Goat	Santa Cruz Biotechnology (sc-22458)	1:50	Staining absent when primary or secondary antibodies omitted (not shown), or in T1R3−/− mice (17)(Fig. 2). This antibody has also been used previously in taste tissue (35).
Anti-5-HT	Rat	Eugene Tech International, Inc. (Allendale, NJ, USA)	1:5000	Staining absent when primary or secondary antibodies omitted (not shown). Use of this antibody has been reported previously (57, 58). Staining pattern matches that of another, well-validated antibody against 5-HT (not shown; Immunostar, Hudson, WI; (59, 60).

Cell counting

Unbiased estimates of TRCs were obtained by stereological cell counting. Our selection process assumed that all taste buds had an equal chance of expressing glucagon and the GlucR. Every fifth section was collected on a slide and kept aside for counting purposes. Within these sections, 2–3 taste buds were selected in a randomized, unbiased manner. Selected taste buds were imaged in Z stacks of 2-μm sections with a ×60 objective and counted on Neurolucida 8.0 (MBF Bioscience, Williston, VT, USA). To minimize the chances of counting a single cell more than once, an optical dissection technique was used. We counted all immunolabeled TRCs within each randomly selected taste bud except for those cells with nuclear staining in the rostral-most slice of a confocal Z stack. Each counted cell was followed through multiple optical sections in the Z stack to verify that it exhibited typical TRC morphology. For each set of quantitative experiments, we counted ∼150–200 TRCs/mouse, from a total of 6 mice.

Behavioral analysis

Mice (10–12 wk old) were habituated to the laboratory environment for at least 30 min each day before the initiation of taste testing. All tastants were prepared with distilled water and reagent-grade chemicals and presented to the animals at room temperature. Test stimuli consisted of 6 concentrations of each stimulus: sucrose (0, 25, 50, 100, 200, and 400 mM; Fisher Scientific, Atlanta, GA, USA), NaCl (0, 30, 100, 300, 600, and 1000 mM; Sigma-Aldrich), denatonium benzoate (DB; 0, 0.05, 0.1, 0.5, and 1, 5 mM; Sigma-Aldrich), and citric acid (CA; 0, 0.3, 3, 10, 30, and 100 mM; Fisher Scientific).

The brief-access taste test was administered in a Davis rig gustometer (Davis MS-160; DiLog Instruments, Tallahassee, FL, USA), as described previously (11, 18–20). Brief-access tests minimize postingestive effects that might confound other assays, such as intake tests (20). Training and testing protocols have been described in detail elsewhere (11, 21). Mice were first acclimated to stimuli and testing conditions during training days. To encourage sampling from the sipper tubes, mice tested with sucrose (an appetitive stimulus) were food and water restricted (1 g food and 2 ml water) for 23.5 h prior to each testing day. Each animal was given a “recovery” period of at least 23.5 h immediately preceding this period of food restriction, during which time mice had access to food and water ad libitum.

Mice tested with aversive stimuli (DB, CA, NaCl) were deprived of water for 23.5 h so that they would be motivated to approach the sipper tubes and lick aversive stimuli. During testing, mice had access to the taste stimuli (presented in sipper bottles as a concentration range) through a small opening in the testing chamber. For test days, unconditioned licking responses were recorded for later analyses in 25-min brief-access test sessions, during which mice could initiate as many trials as possible. Each trial lasted 5 s, with 7.5 s interpresentation intervals. Between trial presentations of aversive stimuli, mice were presented with a 1 s “rinse” with distilled water, to prevent carryover effects. In our hands, these behavioral protocols yield robust, concentration-dependent licking. Mice initiated 15–30 trials per session with sucrose and 30–45 trials/session with DB, CA or NaCl. Behavioral experiments included 3 testing days.

For experiments including the GlucR antagonist L-168,049 [2-(4-pyridyl)-5-(4-chlorophenly)-3-(5-bromo-2-propyloxphenyl) pyrrole; Calbiochem, San Diego, CA, USA], the drug was prepared in DMSO (Sigma-Aldrich) and administered in taste solutions (including the water-only control), at a final concentration of 1 μM. The IC₅₀ of L-168,049 for the mouse GlucR is 63 nM (22–24). Mice ingest 1–2 ml of total liquid during a typical 25-min brief access test session (20). At 1 μM L-168,049 concentration in the taste solution, this would result in <1 μg of ingested drug over the entire testing period. Based on studies of L-168,049 plasma concentrations in mice after an oral dose of L-168,049 (23), we would conservatively estimate systemic plasma concentrations of 1 nM at 25 min after oral dosing and peak concentrations at 10 nM (far below the IC₅₀ of 63 nM). Plasma concentrations of L-168,049 in these behavioral experiments are certainly much lower as the drug is ingested over a 25-min period, not in a single dose. Thus, any postingestive effects of the GlucR antagonist during testing are likely negligible.

All taste testing took place during daylight hours, and mice were always naive to the taste quality being tested. Mice tested for responsiveness to sucrose or to L-168,049 were always naive to behavioral testing. The same cohort of Scg5^−/− and Scg5^+/+ mice were used for all taste tests, with a 2-wk interval between the testing of each stimulus. Scg5^−/− and Scg5^+/+ mice were tested in the following order: sucrose, NaCl, CA, DB. To minimize the number of mice engaged in behavioral experiments with L-168,049 (or vehicle), some mice were tested with two different stimuli, with at least 2 wk in between the testing of the two stimuli. For example, of the 20 mice tested with citric acid, 10 had been tested previously with sucrose, while 10 were naive to behavioral testing. Naive and behaviorally experienced mice were equally distributed across drug and vehicle treatment groups. No mice were tested with >2 stimuli in these experiments.

Statistical analysis

Testing between experimental groups (mice treated with the glucagon receptor antagonist, and Scg5^−/− mice) were compared to appropriate controls (vehicle-treated mice and Scg5^+/+ littermates, respectively) with a 2-way (drug or genotype X concentration) repeated measures analysis of variance (ANOVA). Statistical significance was identified by values of P ≤ 0.05. For all behavioral data, a taste/water-lick ratio was obtained by taking the average number of licks per trial for each stimulus concentration and dividing this amount by the average number of water licks per trial. The use of a taste/water-lick ratio serves as a control for individual differences in lick rates and for differences in motivational state (19). All statistics were performed on Systat 12 software, and all graphs were created on Sigmaplot 10 (Systat Software, Inc., Chicago, IL, USA). For presentation of behavioral data, curves were fit to the mean data for each group using a 2- or 3-parameter logistic function of the form:

$$f\left(x\right)=\frac{a-d}{1+{10}^{\left(x-c\right)b}}+d$$

where x = log₁₀ concentration, c = log₁₀ concentration at the inflection point, and b = slope. For sucrose, a = the asymptotic tastant/water-lick ratio, and d = minimum asymptote of tastant/water-lick ratio. For aversive stimuli, a = 1.0, and d = 0.

RESULTS

Glucagon and the glucagon receptor are expressed in a subset of TRCs

Previously, we reported that both glucagon and the enzyme that processes the mature glucagon peptide from proglucagon, proprotein convertase 2 (PC2), are expressed in subsets of TRCs (11). To better understand the role of glucagon in peripheral taste function, we first used RT-PCR to determine whether other molecules involved in glucagon signaling are expressed in taste tissue or in adjacent, nontaste lingual tissue of C57BL/6J mice. We amplified cDNA products for the glucagon receptor (GlucR) and for 7B2 (an obligatory PC2 chaperone required for the production of mature glucagon and encoded by the Scg5 gene) from mouse circumvallate papillae (Fig. 1A). To confirm that the circumvallate cDNA pool contained TRCs, we also amplified a product for the taste transduction-related G protein subunit α-gustducin; no GlucR, 7B2, or α-gustducin products were amplified from nontaste tissue cDNA (Fig. 1A), although actin cDNA was amplified from both samples. These results indicate that key molecules required for the production and reception of glucagon signals are present in the mouse gustatory epithelium.

Figure 1.

Expression of glucagon signaling components in mouse taste buds. A) RT-PCR of mouse CV and surrounding nontaste tissue (NTT). B–D) Coexpression of glucagon and GlucR in the same subset of TRCs in CV (B), foliate (C), and fungiform (D) papillae. E) GlucR and 7B2 are coexpressed in a subset of TRCs in CV papillae. F) GlucR staining in TRCs is absent in GlucR^−/− mice. Scale bars: 10 μm (B–E); 50 μm (F).

Next, we characterized the protein expression patterns of glucagon and GlucR in taste buds of the mouse. Immunohistochemical analysis of taste buds in the circumvallate, foliate, and fungiform papillae revealed that glucagon and GlucR are expressed in largely overlapping subsets of TRCs in all three papillae types (Fig. 1B–D). Stereological cell counts in circumvallate taste buds confirmed that most (92.6%) glucagon-immunopositive (glucagon⁺) TRCs also express GlucR, while 89.7% of GlucR⁺ TRCs express glucagon (Table 3). Similarly, 85.5% of 7B2⁺ TRCs express GlucR, whereas 89.2% of GlucR⁺ cells are 7B2⁺ (Fig. 1E and Table 3). The high degree of overlap for glucagon, 7B2, and GlucR immunoreactivity suggests that glucagon may act predominantly as an autocrine signal in taste buds.

Table 3.

Stereological counts of taste receptor cells in circumvallate papillae

Marker 2	Marker 1
	Glucagon	GlucR	PLCβ2	5-HT	α-gust	7B2	T1R3
Glucagon	–	338/365		7/365			719/776
GlucR	338/377	–	242/255	13/255	263/480	576/646
PLCβ2		242/747	–
5-HT	7/743	13/615		–
α-Gust		263/696			–
7B2		576/674				–
T1R3	719/903						–

Denominator is the number of cells expressing marker 2; numerator is the number of cells expressing both marker 1 and marker 2. Blank entries, not determined.

The taste bud is heterogeneous: It contains several types of cells that can be defined by morphological criteria and the expression of key molecules (25). Type II cells express the effector enzyme phospholipase C β2 (26), while most type III cells are serotonergic (27). To determine which cell types express components of the glucagon signaling pathway, we again used multilabel immunohistochemistry and stereological cell counting. Serotonergic TRCs immunopositive for glucagon or GlucR were rarely seen (0.9 and 1.9%, respectively; Fig. 2A and Table 3). Glucagon⁺ and GlucR⁺ TRCs that accumulate serotonin were also uncommon (1.9 and 5.1%, respectively). However, GlucR immunoreactivity was almost always colocalized with that of PLCβ2 (Fig. 2B and Table 3): 94.9% of GlucR⁺ TRCs were PLCβ2⁺. These results differ dramatically with those for GLP-1, which is found in both serotonergic and PLCβ2⁺ TRCs, and for the GLP-1 receptor, which is expressed on afferent nerve fibers innervating the taste bud (11). It appears that the glucagon signaling apparatus is almost wholly restricted to type II TRCs in circumvallate taste buds.

Figure 2.

Glucagon and GlucR are expressed in a subset of type II TRCs. A) Glucagon and GlucR are not expressed in 5HT⁺ TRCs. B) GlucR is expressed in a subset of PLCβ2⁺ TRCs. C) Partial overlap of GlucR and α-gustducin expression in TRCs. D) Glucagon and T1R3 coexpression in TRCs. E) T1R3 staining in TRCs is absent in T1R3^−/− mice. Scale bars = 10 μm. All sections are from mouse CV.

Although almost all GlucR⁺ TRCs express PLCβ2, only about one-third of PLCβ2⁺ (i.e., type II) TRCs express GlucR. The G-protein α-gustducin is also restricted to a subset of type II TRCs (28). However, GlucR and α-gustducin overlap only partially: about half (54.8%) of GlucR⁺ TRCs express α-gustducin, while 34.8% of α-gustducin⁺ TRCs express GlucR (Fig. 2C; Table 3). Overlap of glucagon signaling molecules is much greater with another marker of some type II cells, T1R3 (Fig. 2D and Table 3): 92.7% of glucagon⁺ TRCs express T1R3, a subunit of the sweet and umami taste receptors, while 79.6% of T1R3⁺ cells are also glucagon⁺. Thus, the majority of TRCs expressing components of a glucagon signaling pathway are T1R3⁺ type II cells that are likely involved in the detection of sugars and/or amino acids.

Glucagon signaling modulates sweet taste responsiveness in behaving mice

Mature glucagon is processed from its proprotein form, proglucagon, by the enzyme PC2. The chaperone protein 7B2 is required for PC2 activation and the production of mature glucagon (29). We used mice in which the Scg5 gene (which encodes 7B2) has been disrupted by gene-targeting (14, 15) to investigate the contribution of glucagon signaling to normal taste function. When maintained on a C57BL/6J background these mice are viable, live normal lifespans, and display no obvious abnormalities (15, 29). Scg5^−/− mice exhibited no gross taste bud defects but lacked both 7B2 and glucagon immunoreactivity as expected (Fig. 3A–D). Scg5^+/+ and Scg5^−/− mice (n=7–9) were assayed for their behavioral responses to prototypical sweet (sucrose; Fig. 3E), salty (NaCl; Fig. 3F), bitter (DB; Fig. 3G), and sour (CA; Fig. 3H) taste stimuli using a standard brief access taste test. This test minimizes the contribution of postingestive effects, and is thus more specific for orosensory cues (20). Scg5^−/− mice were less sensitive to sucrose as compared to Scg5^+/+ littermate controls (2-way ANOVA: F_1,14=8.414, P=0.01; Fig. 3E). However, Scg5^+/+ and Scg5^−/− mice showed no significant differences in their responses to NaCl (F_1,14=1.084; P=0.32), CA (F_1,14=0.977; P=0.34), or DB (F_1,14=0.215; P=0.65) (Figs. 3F–H). Thus, mice deficient in glucagon show a specific deficit in sweet taste responsiveness.

Figure 3.

Altered sweet taste responses of Scg5^−/− mice in brief access taste tests. A, B) 7B2 is expressed in Scg5^+/+ (A) but not Scg5^−/− (B) mice. C, D) Glucagon is produced in Scg5^+/+ (C) but not Scg5^−/− (D) mice. E–H) Taste responses, expressed as taste/water-lick ratios and as a function of stimulus concentration, of Scg5^−/− (red; n=7) and Scg5^+/+ littermates (black; n=9) to sucrose (E), NaCl (F), DB (G), and CA (H). Scg5^−/− mice exhibited a reduced responsiveness to sucrose (P=0.01) but not to the other taste stimuli. Points are expressed as means ± se. Dashed reference line indicates lick ratio for water (1.0). Scale bars = 10 μM.

The chronic 7B2 deficit in Scg5^−/− mice affects the cleavage of a number of proproteins (e.g., proopiomelanocortin; ref. 30) and abolishes glucagon signaling throughout the body. This global disruption of 7B2 could indirectly affect cellular functions or developmental processes. Therefore, to confirm that the deficit in sweet taste responsiveness observed in Scg5^−/− mice is due to the specific loss of glucagon signaling in taste buds, we performed brief access taste tests in C57BL/6J mice in which glucagon signaling in the oral cavity is acutely disrupted by the short-term, local application of the specific GlucR antagonist L-168,049. This nonpeptidyl triarylimidazole compound is highly membrane permeable, has a low IC₅₀ (63 nM) for the mouse GlucR, and does not cross-react with the GLP-1 receptor (22). Plasma levels of L-168,049, which does not affect blood glucose levels in mice (22), are negligible 24 h postadministration (23). The antagonist was dissolved in DMSO and diluted in the taste solutions for oral presentation during the taste test (final concentrations: L-168,049, 1 μM; DMSO, 2%). Control animals received the same concentration of DMSO but no L-168,049. Mice presented with increasing concentrations of the drug in the absence of tastants were indifferent to it (Fig. 4A and data not shown); mice exhibited neither preference nor aversion compared to water alone regardless of whether the mice were in a water-deprived (Fig. 4A) or water- and food-restricted state (not shown). However, mice receiving 1 μM L-168,049 along with the sweet stimulus sucrose showed a significant decrease in taste responsiveness when compared to vehicle-treated controls (F_1,18=7.286; P=0.02), with a concomitant increase in EC₅₀ (DMSO: 130 mM sucrose; L-168,049: 174 mM sucrose; P=0.04; Fig. 4B). No differences were seen in responses to NaCl (F_1,8=0.051; P=0.83), CA (F_1,18=0.136; P=0.72) or DB (F_1,18=2.427; P=0.14) between L-168,049-treated and vehicle-treated mice (Fig. 4C–E). The effects of L-168,049 on sucrose taste responses are unlikely to reflect postingestive actions. We found that differences between the sucrose responses of drug- and vehicle-treated groups remained significant, even when analysis was restricted to the first 10 min of each testing session (F_1,8=10.796; P=0.004). These results are consistent with our findings in Scg5^−/− mice and confirm that glucagon signaling in taste buds acts to enhance or maintain sweet taste responsiveness.

Figure 4.

Reduced sweet taste responses in the presence of a GlucR antagonist. A) Mice show neither preference nor aversion to ascending concentrations of L-168,049 (n=6). To best test for an aversive taste, mice are deprived of water prior to testing. B–E) Taste responses, expressed as taste/water-lick ratios and as a function of stimulus concentration, of drug-treated (red) and vehicle-treated (61) C57BL/6J mice to sucrose (L-168,049-treated, n=9; vehicle-treated, n=11; B), NaCl (L-168,049-treated, n=5; vehicle-treated, n=5; C), DB (L-168,049-treated, n=10; vehicle-treated, n=10; D), and CA (L-168,049-treated, n=10; vehicle-treated, n=10; E). L-168,049-treated mice exhibited reduced responsiveness to sucrose only (P=0.02). Points are expressed as means ± se. Dashed reference line indicates lick ratio for water (1.0).

DISCUSSION

Sugars like glucose and sucrose are critical nutrients. Glucose in particular provides organisms with an important source of energy, participates in metabolism, and is an obligatory precursor for many biomolecules. Dysregulation of glucose homeostasis, as is seen in metabolic diseases such as diabetes, has deleterious effects on an animal's health. Through the modulation of sweet taste responsiveness, an opportunity exists to affect glucose homeostasis by influencing the perception and ingestion of sugars in foods. In this study, we found that disruption of glucagon signaling within the taste bud reduces sweet taste responsiveness. Our results suggest a novel mechanism for modulating the ingestion of sugar-rich foods.

Several other peptide hormones are expressed in rodent TRCs, including GLP-1, CCK, NPY, VIP, and galanin (7–11, 31–33). It appears that most are expressed in different subsets of TRCs, although the extent of overlap has not been fully determined. Nearly all NPY-expressing cells are immunopositive for both CCK and VIP, although these latter two peptides are also seen in NPY-negative TRCs (6–8, 33). The majority of CCK-, NPY-, and VIP-expressing TRCs also express α-gustducin, thus indicating that these peptides are found in some type II TRCs (33). However, it is unclear if they, like glucagon, are restricted to type II TRCs. In contrast, GLP-1 and galanin are found in subsets of both type II and type III TRCs (9, 11).

The differential expression of these various peptides across subsets of TRCs suggests that they might differentially affect taste function. The overall effect of taste bud-localized CCK and NPY on taste is unclear, although each has been shown in dissociated cells to modulate potassium conductances through their activation of CCK-A or NPY-1 receptors (6, 8, 33). Despite their differential expression in taste buds, the proglucagon products GLP-1 and glucagon appear to have similar effects on sweet taste responsiveness: As disruption of either GLP-1 or glucagon signaling reduces behavioral responses to sweeteners, we can conclude that both hormones normally act to maintain or enhance sweet taste responsiveness. The similar effects of GLP-1 and glucagon signaling on sweet taste suggest that their different effects on taste function may be manifest in other ways. For example, our preliminary studies have indicated that GLP-1-null mice are hypersensensitive to another appetitive stimulus, monosodium glutamate (34), while Scg5-null mice do not differ from littermate controls in their behavioral responses to umami stimuli (unpublished results). Thus, GLP-1 and glucagon may differentially regulate responsiveness to appetitive tastants. Another possibility is that TRCs expressing GLP-1 or glucagon also express receptors for other modulatory signals. Indeed, receptors for circulating neuromodulators such as leptin and endocannabinoids are expressed in taste buds, although the extent of their overlap with GLP-1 or glucagon is not known. The adipocyte hormone leptin suppresses sweet taste responsiveness (5, 12), thereby opposing the actions of both glucagon and GLP-1. In contrast to leptin, endocannabinoids act on T1R3-positive TRCs to enhance taste responsiveness to sweeteners (35). With the growing list of potential modulators acting at the taste bud, it will be important to test the actions of these molecules in the context of each other if we are to understand the potential for dynamic modulation of taste responsiveness under physiological conditions.

Glucagon is secreted from α cells within the islets of Langerhans of the endocrine pancreas during periods of hypoglycemia (36). A modest rise in glucagon levels is also seen within 1 min of the initiation of a meal (37). It is thought that this early increase in systemic glucagon levels is elicited by cephalic stimuli (38); it is intriguing to speculate that glucagon secreted from TRCs could contribute to this cephalic phase response. The liver is the principal target of pancreatic glucagon, where it stimulates glycogenenolysis and gluconeogenesis and likely also initiates signals that contribute to meal termination (38–41). The satiating effects of glucagon, which are manifest as a reduction in meal size and duration, are dependent on signaling through vagal efferents to the nucleus of the solitary tract (NST), although the exact mechanism of action is not known (38, 42). Hepatic vein infusion of glucagon suppresses oral glucose-dependent neural activity in the gustatory region of the NST (43, 44). However, it is unclear how this change in neural activity, measured by multiunit extracellular recordings, might relate to taste perception or behavior. Furthermore, this effect emerges 15 min after a hepatic glucagon infusion, suggesting that the changes in neural activity in the NST are not directly modulated by glucagon (43). The changes in taste responsiveness we observe in mice exposed to the GlucR antagonist are unlikely to reflect glucagon's systemic actions. For example, the effects on sucrose taste responsiveness are seen within the first 10 min of testing, well before changes in NST responses are reported (43). Furthermore, levels of the GlucR antagonist in blood are likely to be negligible, even after 25 min of testing (23). Thus, our results are consistent with a local function for glucagon within the gustatory epithelium.

In addition to its role in promoting glucose release from the liver, glucagon acts on pancreatic β cells to stimulate insulin secretion (45) and engages in autocrine signaling on α cells (46–48). Indeed, the application of glucagon to α cells isolated from rats or mice leads to a dose-dependent increase in cAMP and exocytosis of glucagon; these effects are blocked in the presence of a GlucR antagonist (47). Together with the observation that GlucR has been amplified in pancreatic α-cells by RT-PCR (47), these data suggest that glucagon engages in autocrine signaling in α cells, a finding similar to ours in the taste bud. Other similarities might exist between taste buds and pancreatic islets. For example, cells of the endocrine pancreas, including glucagon-producing α cells, seem to express both T1Rs and α-gustducin (49, 50). Furthermore, T1R3 appears to play a role in the control of insulin secretion from β cells (50). The parallels between peptide-producing TRCs and other endocrine cells of the GI tract and its associated organs support an emerging view that the taste bud is more than just a detector of chemosensory stimuli. Rather, the taste bud has the potential to modulate how those stimuli will be perceived.