Possible role of glucagon-like peptide 1 in taste sensing

The Taste Bud Cells (TBC) are sensory organs with a bulb-shaped structure that possess taste receptor cells (TRC). As far as we know, and understanding the wide interindividual variation, there are approximately 2000 to 5000 TBC distributed on the soft palate, pharynx, larynx and upper esophagus, but mainly localized in the tongue (16, 17).

The tongue consists of two parts: the anterior two-thirds referred to as the corpus linguae (body) and the posterior third known as the radix linguae (root), both of which are separated by a V-shaped groove called terminal sulcus. The root of the tongue connects it to the jaw and to the hyoid bone, while the body stores on its surface four distinct lingual papillae, each with a unique anatomical and functional characteristic and an individual relationship with taste perception (18, 19): The filiform papillae, although the most numerous, have no gustatory functionality, but provide texture sensation and assist in the mechanical manipulation of food. The circumvallate papillae, arranged in a “V” shape at the base of the tongue and innervated by the glossopharyngeal nerve (IX), contains the highest concentration of taste buds in the tongue. The fungiform papillae, located at the tip and edges of the tongue, contain taste buds richly vascularized and are innervated by the chorda tympani, a branch of the facial nerve (VII), and the foliate papillae, located on the posterior lateral folds, which apparently have less population of TBC with age (20–23). (Figure 1) It is also important to highlight the complex neuronal architecture of the tongue, as we should consider three different innervation types: 1) Autonomic innervation: generates the salivatory stimuli and regulation of local blood flow, that plays an important role in food pleasure and avoidance (26) 2) Motor innervation: control of muscle movement, such as mastication, swallow, speech, and breathing (27), and 3) Somatosensorial innervation: that encompasses taste perception, nociception and proprioception (18). As stated above, it is innervated by branches of three cranial nerves (CN); the chorda tympani and greater superficial petrosal branches from the facial nerve (CN VII); the lingual branch provided by glossopharyngeal nerve (CN IX), and the superior laryngeal branch part of the vagal nerve (CN X) (18, 21). All the information, related to attributes of tastant quality, intensity, and hedonic quality travels through the afferent nerve fibers from the TBC and guided by these CNs converge in the gustatory portion of the nucleus of the tractus solitarius (NTS), that subsequently, will move through the thalamus, limbic system, and insula in the gustatory cortex (18, 21, 28). In humans, the NTS establishes direct to the ventral and posteromedial (VPM) nucleus of the thalamus (29). In addition to the mentioned cranial nerves, somatosensory innervation is also performed by the trigeminal nerve (CN V), which consists of three main branches, ophthalmic, maxillary, and mandibular. Here, the information its first stored in the trigeminal ganglia, subsequentially ascending through axonal projections that enter to the pontine nucleus (or main sensory nucleus) and project to the VPM, converging in the sensory cortex in the postcentral gyrus, where information about touch-position and pain-temperature sensation are integrated; as an example, the lingual branch of the mandibular nerve transmits sensorial information such as pain, thermal sensitivity, and mechanical stimuli through the C, A-δ, and A-β fibers (28, 30). In addition to this, the trigeminal perception of the peripheral and somatosensory systems can be guided through transient receptor potential channels (TRP), that are cation selective (mainly calcium) and modulated by calmodulin, phospho-lipases, and kinases (31). These channels affect gustatory processing by altering sensibility to changes in temperature (activated indirectly by tastants or directly by certain spices) in response to physical stimuli, such as mechanical forces and temperature, and by chemical agents (chemestesis), evoking sensations of touch, temperature, and pain, depending on the kind of subunit they are attached to; for example, TRPs help humans perceive two kinds of tastes at the same time, like sweet taste, but at the same time, the lasting sensation of coolness that menthol has (perception named as pungency) (32–34). In this complex interplay, we can assume that some neurons respond to particular combinations of taste and texture produced by food, helping us to decipher pleasant stimuli, reward, and appetite control (35, 36). Therefore, we can state that the sensation and perception of food texture (primarily associated with filiform papillae), may outperform as well or better than the fingertips; and as a matter of fact, can be extremely important for the preference and attitude toward different kinds of food (37, 38).

Taste buds are composed of distinct types of cells, some of them functioning as support cells, while there are others more specialized, named Taste Receptor Cells (TRC), which in turn, are divided in four categories. Each TBC category has approximately 50–100 different kinds of TRCs (24, 25). This classification arises from their earliest histological identification through electron microscopical studies; however, physiologically we can classify two subtypes of TRCs. Cells that respond to taste stimulation, such as sensory receptors cells with subsequent transduction cascades (such as Type II); and cells excited by depolarization, but not by taste stimulation (such as Type I and III) (39) (Figure 2).

Although, we commonly associate TRCs and TASR only to the oral cavity, they are also localized in extra-oral tissues, including the gastrointestinal tract, where they participate in energy homeostasis by modulating enteroendocrine hormone release (GLP-1, GIP, CCK, PYY) and nutrient sensing, thereby influencing gastric emptying and metabolic adjustments (e.g., glycemia) (15, 59, 60). For clarity, we summarize connections between taste-like receptors and gut hormone production (Table 1, Figure 2).

Some nutrients stimulate intestinal hormone secretion independent of taste receptors, via pathways involving other GPCR or nutrient transporters (95, 96):

Certain hormones are involved in modulating the sense of taste. The binding of leptin to its receptor (Ob-R) in type II TRCs, activates K_ATP channels, hyperpolarizing the cell in response to sweet stimuli. In leptin-sensitive individuals, this signaling may help prevent the excessive intake of sweet foods. However, in states of leptin resistance such as obesity, this regulatory mechanism may be impaired (102, 103). On the other hand, adiposity is associated with a state of low-grade chronic inflammation, which can interfere with the survival and renewal of taste buds (104, 105). In 2009, Wang et al. identified the expression of Toll-like receptors (TLRs) and interferon receptors (IFN) in TRCs, which, upon activation, trigger apoptosis and disrupt cellular balance in taste buds. Since then, it has been hypothesized that acute inflammation can reduce the population of taste cells, affecting taste discrimination (106). In 2018, Kaufmann et al. analyzed the inflammatory response triggered by a high-fat diet in mice. They found increased expression of TNF-α in the taste epithelium, as well as altered expression of TNF-α receptors 1 and 2. These changes were associated with a lower density of taste buds in the circumvallate region. However, the balance between type I, II, and III cells was not affected. Additionally, fewer Ki67-immunoreactive cells were observed, suggesting a lower density of taste progenitor cells (105). Recently, Archer et al. studied transcriptomic profile differences in taste buds between lean and obese individuals. Their analysis revealed a decrease in the expression of SHH (Sonic hedgehog gene) and PLCβ2 (Phospholipase C Beta 2) in obese subjects. SHH has previously been identified as a critical signal for the maintenance and renewal of taste cells, while PLC-β2 is a marker involved in sweet, umami, and bitter signaling within type II cells, which may contribute to reduced sensitivity to taste stimuli (107). This information supports the hypothesis that dysfunction in obese individuals may originate from changes in taste bud morphology, which result from a reduction in taste bud abundance and altered signaling (104). Finally, several studies on post-bariatric surgery patients (Roux-en-Y gastric bypass) have observed a decreased preference for highly concentrated sweet foods, without significant changes in response to salty, sour, or bitter stimuli. This phenomenon may be explained by alterations in hormonal secretion induced by bariatric surgery (108, 109).

Certain brain regions are responsible for processing sensory information through visual, olfactory, and auditory pathways, as well as gustatory signals. The integration of gustatory sensory signals occurs primarily in the orbitofrontal, prefrontal, and insular cortex (110). Meanwhile, “food memories” are distributed across a neural network involving the hippocampus, prefrontal cortex, dorsal striatum, and amygdala structures (111–113). Once food passes through the gastrointestinal tract (GI), taste receptors, along with mechanosensors and chemoreceptors within the GI tract, transmit information via afferent branches of the vagus nerve and hormonal signaling (e.g., GLP-1, GIP, PYY, insulin, among others) (46). Multiple studies have established that the hypothalamus serves as the central integrative hub for satiety and appetite regulation. This is exemplified by experimental models showing that selective ablation or stimulation of Agouti-related peptide (AgRP)/NPY/GABA neurons in the basomedial hypothalamus leads to either starvation or hyperphagia with subsequent obesity, respectively (114, 115). Downstream signaling from hypothalamic AgRP/NPY/GABA-secreting neurons is considered redundant, meaning that selective stimulation of individual downstream projection areas is sufficient to suppress appetite. Furthermore, cortical stimulation through these projections enhances the reward value or incentive salience of certain foods (116). Based on this neuronal projection distribution, downstream brainstem nuclei appear to regulate portion size, given their reception of vagal and circulating hormonal afferents, and can modulate food intake, digestion, and absorption (117). In contrast, the corticolimbic system (including the hippocampus, basal ganglia, and amygdala) governs the cognitive and emotional aspects of feeding behavior (118). One of the multiple theories explaining obesity focuses on the transition from high-fiber diets to energy-dense, palatable diets, with current knowledge derived mainly from animal studies and neuroimaging research (113, 119). In this context, murine studies have shown that sweet taste receptors (TAS1R2 and TAS1R3) are predominantly expressed in the hypothalamus and brainstem, whereas bitter receptors (TAS2R116, TAS2R118, TAS2R138, TAS2R104) are mainly expressed in the arcuate nucleus, ventromedial hypothalamus, dorsomedial hypothalamus, and nucleus of the solitary tract (120, 121). These animal models have also explored whether G protein-coupled taste receptor expression is altered in obesity models, particularly in leptin-deficient ob/ob mice and diet-induced obesity models (122). Both genetic and environmental models demonstrate reduced taste receptor expression, especially in the hypothalamus and brainstem, but no significant changes in cortical or hippocampal expression (123). In humans, various neuroimaging studies using positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) have provided insights into brain responses to food stimuli (120, 124, 125). For instance, in individuals with melanocortin receptor 4 (MC4R) deficiency (a genetic model of obesity), heightened activity in the ventral and dorsal striatum has been observed in response to palatable foods compared to individuals without MC4R mutations. These findings suggest that disruption of the leptin-proopiomelanocortin-MC4R circuit is associated with reduced executive control over eating behavior (126). Since leptin reduces sweet taste sensitivity, diminished leptin signaling in the CNS may contribute to the reduced sweet taste sensitivity observed in individuals with obesity (127). Regarding fat taste, mediated by CD36, healthy individuals exhibit increased oxygen consumption in the secondary gustatory cortex (orbitofrontal and medial prefrontal cortex) following oral fat intake. Conversely, individuals with obesity show increased oxygen consumption in the insula, operculum, bilateral precuneus, and posterior cingulate cortex, along with the ventral putamen and Rolandic operculum, but decreased oxygen uptake in the lower caudate—patterns associated with increased impulsivity towards food stimuli (128–130). Moreover, amygdala responses are attenuated following the consumption of high-fat foods, reducing subsequent responses to further food stimuli (131). The magnitude of brain response to high-fat versus neutral foods has been proposed as a predictor of body mass index (BMI) in healthy volunteers (132). Concerning sweet and salty tastes, clinical studies have yielded conflicting results. Most do not find differences in sweet taste perception between individuals with and without obesity. Some studies, however, have reported decreased salty taste perception in individuals with higher BMI (133, 134). It has repeatedly been reported that an increase in BMI is associated with a reduction in the perceived intensity of taste qualities and a diminished sense of taste. Most studies point toward a reduction in taste sensitivity in obesity, which appears to be reversible with weight loss (135, 136). Taste perception influences food selection and consumption; therefore, individuals with obesity often exhibit eating behaviors characterized by the consumption of larger portions, energy-dense foods, and increased snacking behavior (104).

As it is broadly known, GLP-1 is a peptide hormone generated through the enzymatic breakdown of preproglucagon (PPG), ultimately giving rise to GLP-1 (1–37) and GLP-1 (7–36) amide or GLP-1 (7–37) peptide variants. It is synthesized in L-cells located in the intestinal mucosa, α-cells found in the pancreatic islet, and brain preproglucagon neurons residing in the NTS (137). As it is part of the G protein-coupled receptor family, it has a wide distribution through the body, predominantly localizing to the cellular membrane (138). Both GLP-1 and its agonists (GLP-1RAs) have been identified for suppressing appetite, slowing gastric emptying time, and inducing fullness and satiety, thereby reducing caloric intake under ad libitum feeding conditions (139–141). Its presence in the cardiovascular system has implications in cardiac and vascular function (142), while its expression in the central nervous system might regulate behavior, mood, and cognition (143, 144). Mostly, GLP-1RAs are primarily used as antihyperglycemic drugs. Initially approved for type 2 diabetes (T2D) treatment, they were later found to influence appetite regulation (145, 146) by regulating activity in hedonic brain regions like the insula and orbitofrontal cortex changing food preference (147), therefore, leading to their approval for obesity and diabetes management (148–150). Although all of these are considered beneficial and protective mechanisms, there is still some debate about the frequent adverse effects (AE) they have, most of them of gastrointestinal nature (151, 152). The prevalence of AE under the use of GLP-1RAs is approximately 50%, depending on the type and dose of medication; usually mentioned as mild to moderate and temporary as dose escalation is intended, so in most cases, they are resolved after the maintenance dose is reached (153). However, some people may report AE, regardless of the dose usage. Most users may experience nausea or vomiting derived by the GLP-1 activity in area postrema and delayed gastric emptying generated by the molecule (154–157). There are also motility alterations, such as diarrhea explained by the rapid colonic transit, as GLP1 activates intrinsic afferent neurons containing calcitonin gene-related peptide that favors the frequency of peristaltic contractions (158) or constipation due to mechanisms dependent on nitric oxide and AMPc/PKA that reduces colon contractility (159). Less frequent AEs like hypoglycemia, is mostly seen in combination with insulin or sulfonylureas (157) and some reports of pancreatobiliary complications are partially explained by direct altered absorption of bile acids, rapid weight loss or changes in the gallbladder motility with increased risk of pancreatitis, even though it is extremely rare (160). Consequently, dose titration is slow and dependent on AEs, and common recommendations like reduction on food quantity or high fat or spicy food is always suggested (151, 154, 158) Also, some hypotheses might suggest that GLP-1 stimulation can affect taste perception, as it has been reported by the FDA Adverse Event Reporting System (FAERS). Other symptoms, such as dysgeusia or aversion to sweet solutions and high-calorie foods can also happen in approximately 11.5% of people under GLP-1RAs treatment (161). Mechanistic studies indicate that GLP-1RAs suppress the perception of all basic taste modalities, probably through combined actions in the brainstem and vagal modulation, although paradoxical cases of sweet hypersensitivity have been documented in specific populations (162).

In recent years, the relationship between taste and GLP-1 agonism has been attempted to be elucidated, however, the exact interaction between them has not been clearly defined. The principal mechanism described is that GLP-1 signaling exerts control over gustatory perception via peripheral and central mechanisms (163). Emerging evidence demonstrates that GLP-1 receptors are expressed in Type II taste cells, where they modulate sweet sensitivity through the Gα-gustducin/TRPM5 pathway, and downregulate T1R2/T1R3 expression and ATP release (1, 15). Centrally, GLP-1 receptor activation in the NTS and ventral tegmental area (VTA) suppresses dopaminergic responses to sweet stimuli, while enhancing bitter aversion, mediated by TAS2R (164, 165). Clinically, GLP-1RAs like semaglutide induce sweet hyposensitivity and reduce preference for hyperpalatable foods, an effect mediated by changes in endogenous GLP-1 secretion and synaptic plasticity in reward circuits (166). In taste buds, GLP-1 is locally produced from proglucagon through the cleavage by prohormone convertase PC1/3, which is co-expressed with GLP-1 in all taste cells positive for this peptide. Locally, there is no DPP-4 expression in taste tissue, which could suggest that in this location, GLP-1 may achieve a longer half-life (167). Evidence suggests that GLP-1 may not act solely in a paracrine manner; recent studies show that locally secreted GLP-1 can activate GLP-1 receptor-expressing afferent fibers of the glossopharyngeal and vagal nerves, thereby transmitting peripheral gustatory information to the NTS. This establishes a direct neuroendocrine bridge between oral GLP-1 signaling and the central components of the GBA, where it converges with intestinal GLP-1-mediated pathways to regulate appetite and reward processing (58, 61, 168, 169). These findings position the entero-gustatory-endocrine axis as a therapeutic target in metabolic disorders, where GLP-1 dysfunction disrupts taste signaling, perpetuating pathological feeding behaviors (Figure 3).

On the other hand, immunohistochemical studies in mice may show a strong GLP-1 expression in specific subsets of taste bud cells, mainly in circumvallate and fungiform papillae, and approximately 10% of all taste cells are GLP-1 positive, possibly meaning a strong relationship with taste perception as they co-express α-gustducin, a G protein associated with the transduction of sweet, umami, and bitter tastes; conversely knockout GLP-1 -/- mice showed attenuated response to sweet flavor (61, 67, 173, 174). Also, in type III (serotonergic) cells, approximately 23% of serotonin-positive cells express GLP-1. Exposure to sugars induces rapid GLP-1 secretion from T1R3-sensitive cells, following a dose-dependent pattern where higher stimulus concentrations lead to greater secretion—an effect blocked by GLP-1 receptor antagonists (61, 67). It may be stated that it acts as a neurotransmitter or neuromodulator on adjacent GLP-1R-expressing nerve fibers, enhancing attraction to sweet tastes and contributing to residual sweet perception even in the absence of T1R3 (15, 18). GLP-1R knockout mice exhibit reduced responses to sweeteners, confirming their key role in sweet perception, as well as increased sensitivity to monosodium glutamate (MSG), suggesting GLP-1 may inhibit umami taste as well (65).

In the past few years, there has been growing evidence that highlights the physiological and molecular mechanisms in which GLP-1 regulates taste perception. A controlled clinical trial revealed that patients with GLP-1ARs therapy showed significant reductions in perception of all five basic tastes, scoring 29% lower on standardized tests like the Waterless Empirical Taste Test (WETT^®) compared to controls. This study proposes that this phenomenon may arise from peripheral and central GLP-1 receptor hyperstimulation, leading to downregulation of chemosensory pathways (165). Clinical trials with liraglutide showed not only a significant increase in satiety sensation, but also a notable reduction in the desire for sugar-, fat-, and salt-rich foods, accompanied by modifications in the hormonal profile, particularly in peptide YY (PYY) levels. These changes are enhanced by modifications in key regions such as the insular cortex (main area for taste processing) and caudate nucleus, where GLP-1-RAs are capable of normalizing altered brain activation patterns in patients with obesity and diabetes mellitus when exposed to hedonic gustatory stimuli (173, 174). As well, in real-world scenarios, such as described by Brindisi et al, it is mentioned a reduction in pleasure and taste perception in overweight and people with obesity with poor type 2 diabetes control, probably by lowering the detection threshold of sweet flavors, becoming more sensitive to them, resulting in aversion for this flavor (175). In contrast, infusions of native GLP-1 at physiological doses (0.75 pmol·kg⁻¹·min⁻¹) show a preferential effect on the visual perception of food, without significant alterations in palatability parameters (176). More recent advances in drugs, especially semaglutide, have also been related to taste modification. Gibbons et al., studied the effect of oral semaglutide in 15 subjects with type 2 diabetes and reduction on appetite and craving, demonstrating not just the reduction in calorie intake, but also lowering the palatability and preference for sweet and high fat food (177). Also, during STEP 5, a study to assess efficacy of semaglutide with a follow-up of 2 years, there is a post-hoc analysis that demonstrates a better performance in the self-reported Control of Eating Questionnaire(CoEQ) to measure changes in appetite versus placebo, with lower cravings for dairy and starchy foods and less desire for salty and spicy food at short-term (20 weeks) but also as a maintenance therapy at weeks 52 and 104 (178); this finding is also aligned with the one reported from Friedrichsen et al. in which is also mentioned the improvement on this questionnaire at 20 weeks; both favoring weight loss over placebo and demonstrating the major regulation of GLP1-RA specially in people with less control over craving and hedonic hunger. It is also important to state that AEs were similar all across both studies (179). Furthermore, a recent clinical study in women with obesity treated with semaglutide showed improved recognition of basic tastes, altered gene expression in the tongue (measured by tongue biopsy), decreased activation of putamen in response to calorie-dense sweet and savory food cues, and increased cortical brain activity in response to the sweet taste stimulus compared to placebo. The potential effect of semaglutide on taste perception and recognition needs further exploration in more diverse clinical groups (180). With the introduction of dual GLP-1/GIP-RA, ongoing studies could provide novel insights and new evidence on this topic. In a phase 1 trial by Martin and colleagues, observations at week 6 comparing tirzepatide vs once-daily liraglutide and placebo have shown that the use of the latter, provides a reduction on the desire for high-fat and high-sugar food by lowering activation in the cingulate gyri, hippocampus and orbitofrontal cortex assessed by MRI and blood-oxygen-level at these areas, meaning less appetite, early satiety, and less food craving with more reduction on weight loss at short term (181). Recently, a real-world study showing association between incretin-based therapy and changes in taste perception with GLP-1 or dual GLP-1/GIP-RAs usage for around 40 weeks reported that independently to weight loss, there was an increase in sweet perception intensity up to 4.7 times and salt perception up to 3.2 times in around 20% of the participants, resulting as well in earlier satiety, and reduction of hunger and craving due to a major sensibilization to both flavors with any of these medication, although they did not mention further impact in weight loss or BMI change (170). Even the emergent evidence on this topic, there is no clear mechanism that elucidates the specific relationship between GLP-1 and GLP1-RA with taste perception, being apparently a combination of both peripheral and central actions, which are involved. Over CNS, the direct action on NTS, hypothalamic arcuate and surrounding structures, as well as the effect on the ventral tegmental area and nucleus accumbens may act on homeostatic and hedonic hunger, changing palatability and reward behavior; while the reinforcement and hypersensitivity of certain flavors and modulation of appetite by T1R and T2R families may play a peripheral action (170–172).

We summarize the clinical and preclinical evidence that includes evaluations on taste perception under GLP-1RAs treatment in Tables 2 and 3.