Abstract
TAS2R38 is a bitter taste receptor that influences bitter taste perception and diet and is also found in intestinal L cells that store and secrete glucagon-like peptide 1 (GLP-1). Preclinical studies have linked TAS2R38 activation to postprandial GLP-1 secretion, fueling interest in TAS2R38 as a therapeutic target for glucose regulation; however, evidence in humans remains limited. To further establish TAS2R38 actions in glucose homeostasis, we analyzed data from ∼220,000 European adults without type 2 diabetes in the UK Biobank to test whether functional variants conferring TAS2R38 sensitivity were associated with blood glucose. We found that individuals with two copies of a haplotype increasing receptor sensitivity (PAV) had significantly lower 0–2-h (i.e., postprandial) glucose than those with two copies of a nonfunctional haplotype (AVI), following a dose–response relationship per PAV haplotype. These associations were replicated in published genome-wide association studies of 2-h glucose, persisted after adjustment for diet and lifestyle behaviors related to bitter taste perception, and were not seen for variants in other bitter taste receptors without putative roles in glucose metabolism (TAS2R14 and TAS2R19). Collectively, these findings provide evidence in humans consistent with direct TAS2R38 actions in postprandial glycemia, supporting TAS2R38 as a novel therapeutic target for glucose regulation.
Article Highlights
The TAS2R38 bitter taste receptor, recently identified within intestinal L cells, has been shown to modulate GLP-1 secretion in preclinical models; however, evidence in humans remains limited.
We harnessed functional variants comprising three canonical diplotypes of TAS2R38 to study the role of TAS2R38 in glucose homeostasis in humans.
In a large sample of adults without type 2 diabetes, we found that individuals with more sensitive TAS2R38 receptors had lower postprandial glucose levels, independent of diet and lifestyle habits.
Our findings provide evidence in humans supporting direct TAS2R38 actions in postprandial glycemia and highlight TAS2R38 as a potential therapeutic target for impaired glucose regulation.
Graphical Abstract
Introduction
A better understanding of the mechanisms contributing to glucose homeostasis is critical to identify new targets for type 2 diabetes prevention and treatment. Given the clinical success of glucagon-like peptide 1 (GLP-1) receptor agonists in improving glycemic control (1), gut nutrient–sensing mechanisms that modulate endogenous GLP-1 secretion have become key areas of interest (2,3). Type 2 taste receptors (TAS2Rs), a specialized family of G-protein–coupled receptors known for their roles in bitter taste perception (4), were recently identified in intestinal L-cells which store and secrete GLP-1 (5,6). Activation of one particular subtype, TAS2R38, by bitter-tasting ligands has been further shown to trigger dose-dependent secretions of GLP-1 in animal and human cell models (7–11). In preclinical studies, several bitter-tasting ligands have also been linked to increased GLP-1 and insulin secretions and reduced postprandial glucose responses to mixed-nutrient challenges in healthy adults and those with type 2 diabetes (12–14), supporting TAS2R38 as a potential therapeutic target to promote endogenous GLP-1 secretion and improve glucose regulation (15).
Common missense variants in the TAS2R38 gene give rise to two canonical haplotypes with well-defined functional and behavioral implications: AVI encodes a nonfunctional receptor, whereas PAV encodes a functional receptor (16,17). Collectively, these haplotypes explain ∼70% of the phenotypic variability in bitter taste perception by modulating the ligand-binding capacity of the receptor (18,19). When combined into diplotypes, individuals with one or two copies of the functional PAV haplotype, who have normal or heightened bitter taste perception (i.e., AVI/PAV or PAV/PAV; so called tasters or supertasters, respectively) (17), tend to consume less coffee and alcohol, eat fewer bitter-tasting vegetables and smoke less than those with two copies of the nonfunctional AVI haplotype, who have blunted bitter perception (i.e., AVI/AVI; so called nontasters) (20–23).
Despite evidence of a functional role of TAS2R38 in glucose homeostasis, data in humans are limited. Only one prior study examined the impact of TAS2R38 functional variants on glucose metabolism (23), finding that, among ∼1,000 German adults without type 2 diabetes, genetic supertasters and tasters had significantly lower 120-min area under the curve glucose than nontasters after a 75-g oral glucose tolerance test (OGTT). However, it is unknown whether these differences arise from biological actions of TAS2R38 in glucose homeostasis or whether they stem from dietary or lifestyle (i.e., behavioral) effects of TAS2R38 via bitter taste perception. Delineating these effects in large-scale human studies is critical to establishing the functional role of TAS2R38 in glucose homeostasis and thus its potential to serve as a therapeutic target for glucose regulation.
Here, we examined the associations of functional variants in the TAS2R38 bitter taste receptor gene with markers of glucose homeostasis among adults without type 2 diabetes in a large, population-based biobank. We hypothesized that functional variants and canonical diplotypes conferring more sensitive TAS2R38 bitter taste receptors would be associated with lower blood glucose levels, particularly in the postprandial state, and that these associations would persist even after adjustment for dietary and lifestyle behaviors, supporting the possibility of direct biological mechanisms of TAS2R38 in postprandial glucose metabolism.
Research Design and Methods
UK Biobank Participants
The UK Biobank (UKB) is a prospective cohort comprised of adults aged 40–69 years living throughout the U.K., with comprehensive genetic, phenotypic, and health record data (24,25). We retrieved information on genetic relatedness, ancestry group, and principal components from the Pan-UKB project (26). We excluded participants who had withdrawn their consent from the biobank by the time of the analysis and included unrelated individuals of European genetic ancestry with an HbA1c <5.7% (39 mmol/mol) and who did not have type 2 diabetes, as defined by a published algorithm (27) to ensure a focus on those without dysglycemia (n = 343,441). From this subsample, we included individuals with complete data on genotyping, glucose outcomes, dietary habits (from UKB food frequency questionnaires [FFQs]), and relevant dietary and lifestyle confounders, and we removed those with extreme values (SD >5) for glucose outcomes (flow diagram shown in Supplementary Fig. 1).
Functional Variants in the TAS2R38 Bitter Taste Receptor Gene
Our primary exposure was canonical TAS2R38 diplotype based on functional variants in the TAS2R38 bitter taste receptor gene (16,17). Genotyping, imputation, and quality control of the UKB genetic data have been described previously (25). We selected three well-characterized functional variants in TAS2R38, each of which encodes for a single amino acid substitution that affects the ligand-binding capacity (i.e., sensitivity) of the receptor: rs713598 (C > G; p.Ala49Pro), rs1726866 (G > A; p.Ala262Val), and rs10246939 (T > C; p.Ile296Val) (Supplementary Table 1) (18). The three variants form two common haplotypes: AVI, a recessive haplotype encoding nonfunctional receptors, and PAV, a dominant haplotype encoding functional receptors (16). The two haplotypes form three common diplotypes: AVI/AVI, AVI/PAV, and PAV/PAV; AVI homozygotes have blunted bitter taste perception (i.e., nontasters), PAV carriers have intermediate bitter taste perception (i.e., tasters), and PAV homozygotes have heightened bitter taste perception (i.e., supertasters), based on low, moderate, and high sensitivity, respectively, of the receptor to its ligands.
TAS2R38 haplotypes were derived by performing haplotype phasing using SHAPEIT (version 4.2) (28) on a 2-Mb window surrounding the gene on chromosome 7 in the UKB European subsample. Phased haplotypes were combined into diplotypes. To focus our analysis on functional variations with established effects on taste receptor sensitivity, we excluded individuals with rare (<1% frequency) or noncanonical haplotypes (n = 21,427; 8.9% excluded) (participant descriptions provided in Supplementary Table 2).
Markers of Glucose Homeostasis
The primary outcomes were random glucose (mg/dL), assessed as part of routine UKB laboratory testing at baseline (25), measured at any time of day regardless of when the participant last consumed a meal or caloric beverage; and 0–2-h glucose (mg/dL), defined as glucose measured within 2-h of the last consumed meal or caloric beverage based on self-reported time since the last meal. To capture glucose variation over subsequent fasting times, we defined additional fasting windows as 3, 4, 5, 6–12, and 12–24 h (n = 26 excluded for fasting time >24 h) (Supplementary Table 3 summarizes glucose levels in each interval).
Sociodemographic, Lifestyle, and Dietary Covariates in the UKB
Sociodemographic and lifestyle factors were measured using participant surveys at baseline (25). We defined smoking status as current, previous, or never tobacco users; alcohol use as intake frequency ranging from daily to special occasions or never; and physical activity level as tertiles of excess metabolic equivalents per week (29). Values of “do not know” or “don’t want to report” were recoded as missing. For descriptive analyses, we quantified cigarettes per day (among current smokers), alcoholic drinks per day (among current drinkers) (25), income level (based on annual income bracket), and educational attainment (based on the International Standard Classification of Education [30]). We also used an estimate of total energy intake (kcal/day) quantified from 24-h dietary recalls (n = 129,615 with available and plausible data) (31).
Dietary covariates were derived from UKB FFQs administered at baseline using procedures described previously (32). Briefly, the UKB FFQ assessed self-reported intake frequencies and preferences for ∼20 foods, using continuous (e.g., tablespoons of vegetables) and ordinal (e.g., never, once per week, or daily servings) frequency measures as well as categorical measures of food type preferences (e.g., bread or coffee type). We recoded responses of “do not know” or “prefer not to answer” as missing, converted all measures to integer variables (ordinal to quantitative and categorical to binary), and median imputed and winsorized the data to 5 SDs, resulting in 24 continuous dietary traits.
Statistical Analysis
All analyses were performed using R statistical software (version 4.1). Consistent with prior studies, we defined TAS2R38 diplotype as an additive continuous exposure based on the number of PAV haplotypes, akin to allele dosage (0, 1, or 2 for AVI/AVI, AVI/PAV, and PAV/PAV, respectively; primary analyses). We also defined a categorial TAS2R38 exposure to enable pairwise comparisons across diplotypes (AVI/AVI [reference] vs. AVI/PAV or PAV/PAV). Descriptive statistics (mean ± SD and n [%]) were used to compare sociodemographic and behavioral characteristics across TAS2R38 diplotypes, using P values from unadjusted linear models or χ2 tests to assess for differences across additive TAS2R38 diplotypes.
To determine the associations of TAS2R38 diplotype with glucose homeostasis, we used generalized linear models adjusted for age, sex, assessment center, 10 genetic principal components, and fasting time, the latter of which was used to account for variability in time since the last meal and typical variation in glucose levels across postprandial and fasting windows (base model) (33). All associations were also tested in a base model adjusted for BMI to account for potential confounding by adiposity (BMI-adjusted model). We summarized results as effect estimates and P values on a linear scale (per PAV haplotype; Ptrend) and as estimated and SE pairwise differences in outcomes (vs. AVI/AVI; Ppairwise). To test for independent replication, post hoc, we queried summary statistics from published genome-wide association studies (GWAS) of glycemic traits in European samples (excluding UKB) performed by the Meta-Analysis of Glucose and Insulin-Related Traits Consortium (MAGIC). For direct comparisons with published summary statistics, we calculated variant-level effect estimates in the UKB sample using BMI-adjusted models and aligned the effect alleles for both cohorts to the allele associated with higher bitter taste perception. Statistical significance was defined using an a priori threshold of P < 0.025 (i.e., 0.05/2) to correct for multiple comparisons of two primary outcomes.
To test the robustness of associations to behavioral adjustment, we first selected dietary and lifestyle (i.e., behavioral) covariates based on prior associations with TAS2R38 diplotypes or glucose homeostasis. We confirmed the quality of covariates for capturing variability in blood glucose compared with covariate-only BMI-adjusted models using likelihood ratio tests. We then quantified direct effects of TAS2R38 diplotype on glucose levels after adjusting for behavioral covariates and calculated the percent change in estimates upon adjustment. Second, we used negative control tests to confirm a lack of association between variants in other bitter taste receptor genes, related to similar dietary and lifestyle behaviors as TAS2R38 but with no proposed actions in glucose metabolism, with glucose homeostasis. We selected two variants identified in GWAS of caffeine or quinine perception (Supplementary Table 4), coded them as additive exposures based on allele dosage for greater bitter taste perception, and tested their associations with glucose levels in BMI-adjusted models to compare with TAS2R38 variant–level estimates.
Data and Resource Availability
No new genetic or phenotypic data were generated for this study. The UKB data, including genetic and phenotypic data, are under controlled access and can be obtained through an application process described at https://www.ukbiobank.ac.uk/. Code supporting the conclusions of this article can be found at https://github.com/julieeg/taste2d.
Results
Participant Characteristics
In total, 240,115 individuals with a mean ± SD age of 56 ± 8.1 years and mean ± SD BMI of 26.9 ± 4.4 kg/m2 had available data for analysis (Table 1 and Supplementary Fig. 1). No differences in clinical or behavioral characteristics were observed between participants with canonical (included) versus rare or noncanonical diplotypes (excluded; n = 21,427) (Supplementary Table 2). Among included participants, 34%, 49%, and 18% were categorized as AVI/AVI nontasters, AVI/PAV tasters, and PAV/PAV supertasters, respectively (Fig. 1).
Table 1.
Participant characteristics across TAS2R38 diplotypes (N = 218,688)
| Characteristic | Total sample | Canonical TAS2R38 diplotype | P * | ||
|---|---|---|---|---|---|
| AVI/AVI (nontaster) | AVI/PAV (taster) | PAV/PAV (supertaster) | |||
| Participants | 218,688 | 73,524 (33.6) | 106,591 (48.7) | 38,573 (17.6) | — |
| Age, years | 56 ± 8.1 | 56 ± 8.1 | 56 ± 8.1 | 56.1 ± 8.1 | 0.613 |
| Sex, female | 118,720 (54.3) | 39,922 (54.3) | 57,906 (54.3) | 20,892 (54.2) | 0.857 |
| BMI, kg/m2 | 26.9 ± 4.4 | 26.9 ± 4.4 | 26.9 ± 4.4 | 26.9 ± 4.4 | 0.983 |
| Fasting triglyceride, mg/dL†‡ | 138.1 ± 82.6 | 138.2 ± 82.0 | 138.3 ± 83.0 | 137.7 ± 82.4 | 0.618 |
| Fasting LDL cholesterol, mg/dL† | 141.2 ± 32.8 | 141.0 ± 32.8 | 141.0 ± 33.0 | 141.8 ± 32.6 | 0.299 |
| Fasting HDL cholesterol, mg/dL† | 57.3 ± 15.2 | 57.1 ± 15.1 | 57.4 ± 15.3 | 57.5 ± 15.0 | 0.126 |
| Smoking status | 0.625 | ||||
| Current | 20,502 (9.4) | 6,854 (9.3) | 10,080 (9.5) | 3,568 (9.2) | |
| Former | 75,278 (34.4) | 25,238 (34.3) | 36,748 (34.5) | 13,292 (34.5) | |
| Never | 122,908 (56.2) | 41,432 (56.4) | 59,763 (56.1) | 21,713 (56.3) | |
| Cigarettes/day (among smokers) | 17.2 ± 9.6 | 18.6 ± 11.8 | 16.4 ± 7.9 | 16.7 ± 8.9 | 0.026 |
| Physical activity, METs/week | 19.6 ± 34.1 | 19.7 ± 34.2 | 19.5 ± 34.0 | 19.8 ± 34.3 | 0.847 |
| Alcohol intake frequency | 0.0073 | ||||
| Daily | 49,031 (22.4) | 16,662 (22.7) | 23,930 (22.5) | 8,439 (21.9) | |
| 3–4 drinks/week | 55,123 (25.2) | 18,536 (25.2) | 26,964 (25.3) | 9,623 (24.9) | |
| 1–2 drinks/week | 57,516 (26.3) | 19,474 (26.5) | 27,777 (26.1) | 10,265 (26.6) | |
| 1–3 drinks/month | 23,442 (10.7) | 7,793 (10.6) | 11,484 (10.8) | 4,165 (10.8) | |
| Special occasions | 20,825 (9.5) | 6,875 (9.4) | 10,172 (9.5) | 3,778 (9.8) | |
| Nondrinker | 12,751 (5.8) | 4,184 (5.7) | 6,264 (5.9) | 2,303 (6.0) | |
| Alcohol intake, drinks/week | 9.1 ± 9.6 | 9.2 ± 9.6 | 9.1 ± 9.6 | 9.0 ± 9.6 | 0.0051 |
| Total energy intake, kcal/day§ | 1,986 ± 550 | 1,989 ± 555 | 1,987 ± 547 | 1,977 ± 547 | 0.047 |
| Educational attainment, levelǁ | 0.111 | ||||
| 1 | 32,451 (15) | 10,799 (14.8) | 15,858 (15.0) | 5,794 (15.1) | |
| 2 | 59,838 (27.6) | 20,146 (27.6) | 29,131 (27.5) | 10,561 (27.6) | |
| 3 | 25,841 (11.9) | 8,699 (11.9) | 12,724 (12.0) | 4,418 (11.5) | |
| 4 | 10,909 (5.0) | 3,736 (5.1) | 5,199 (4.9) | 1,974 (5.2) | |
| 5 | 88,001 (40.5) | 29,611 (40.6) | 42,872 (40.5) | 15,518 (40.6) | |
| Annual income, $ | 0.026 | ||||
| <18,000 | 37,006 (19.5) | 12,570 (19.6) | 17,853 (19.3) | 6,583 (19.6) | |
| 18,000–31,000 | 46,961 (24.7) | 15,892 (24.8) | 22,834 (24.6) | 8,235 (24.6) | |
| 31,000–50,000 | 51,924 (27.3) | 17,562 (27.4) | 25,317 (27.3) | 9,045 (27.0) | |
| 50,000–100,000 | 42,877 (22.5) | 14,128 (22.1) | 21,173 (22.8) | 7,576 (22.6) | |
| ≥100,000 | 11,445 (6.0) | 3,848 (6.0) | 5,533 (6.0) | 2,064 (6.2) | |
| Raw vegetable intake, tbsp/week | 2.12 ± 1.80 | 2.13 ± 1.81 | 2.12 ± 1.80 | 2.10 ± 1.80 | 0.009 |
| Coffee intake, cups/week | 2.06 ± 1.98 | 2.07 ± 1.98 | 2.06 ± 1.99 | 2.02 ± 1.97 | 2.2 ×10−4 |
| Tea intake, cups/week | 3.42 ± 2.71 | 3.40 ± 2.72 | 3.43 ± 2.71 | 3.46 ± 2.72 | 0.001 |
| Frequency of adding salt to food | 5.8 ×10−23 | ||||
| Always/often | 33,680 (15.4) | 11,809 (16.1) | 16,354 (15.3) | 5,517 (14.3) | |
| Sometimes | 60,272 (27.6) | 20,681 (28.1) | 29,240 (27.4) | 10,351 (26.8) | |
| Never/rarely | 124,725 (57) | 41,029 (55.8) | 60,994 (57.2) | 22,702 (58.9) | |
Data are given as mean ± SD for continuous variables and n (%) for categorical variables.
*P values are based on unadjusted linear models for continuous variables and χ2 tests for categorical variables.
†Fasting was defined based on self-report of ≥6 h since the last meal or caloric beverage.
‡P value tabulated using log-transformed data to correct for nonnormality.
§Energy intake was estimated in n = 129,615 (59.3%) of participants with available and plausible 24-h recall data, defined as ≥600 and ≤4,800 (men) or ≤4,300 kcal/day (women).
ǁEducation levels were defined using the International Standard Classification of Education following published methods (30).
Figure 1.
A and B: Distribution of common TAS2R38 haplotypes (A) and diplotypes (B) among 240,115 European adults without type 2 diabetes in the UKB. Genotypes were derived from three functional variants in TAS2R38 (rs713598, rs1726866, and rs10246939); values above each bar reflect n per genotype.
Across canonical TAS2R38 diplotypes, there were no significant differences in age, sex, BMI, educational attainment, or fasting (≥6-h) plasma lipid levels (Table 1). Compared with nontasters, supertasters consumed fewer alcoholic beverages per week (Ppairwise = 0.006), and among current smokers, tasters and supertasters (i.e., PAV carriers) reported fewer cigarettes per day (Ppairwise = 0.019), consistent with expected associations between TAS2R38 diplotype and health-related behaviors (20–23). We further confirmed expected additive associations of PAV haplotype with less alcohol and cigarette use (Ptrend = 0.0073 and 0.026, respectively), less addition of salt to food (Ptrend = 7.7 × 10−15), lower intake of coffee and raw vegetables (Ptrend = 2.7 × 10−4 and 0.008, respectively), and higher intake of tea (Ptrend = 7.5 ×10−4).
Associations of TAS2R38 Diplotype With Glucose Homeostasis
To investigate the influence of functional variants in TAS2R38 on glucose homeostasis, we first determined the associations of TAS2R38 diplotype with random glucose measured at any time. We found significant inverse associations between each additional PAV haplotype and glucose levels (Ptrend = 0.021), which, in BMI-adjusted models, corresponded to an estimated mean difference in random glucose levels of −0.15 mg/dL (95% CI −0.28, −0.02) between supertasters and nontasters (reference; Ppairwise = 0.019) (Table 2).
Table 2.
Associations of TAS2R38 diplotypes with random and 0–2-h glucose
| Random glucose, mg/dL | 0–2-h glucose, mg/dL | |||||
|---|---|---|---|---|---|---|
| β (95% CI) | P pairwise * | P trend † | β (95% CI) | P pairwise * | P trend † | |
| Base model | 0.003 | |||||
| Additive TAS2R38 diplotype, per PAV | −0.07 (−0.14, −0.01) | — | 0.021 | −0.24 (−0.39, −0.08) | — | |
| AVI/AVI (nontaster) | Reference | — | Reference | — | ||
| AVI/PAV (taster) | −0.05 (−0.15, 0.04) | 0.270 | −0.27 (−0.51, −0.03) | 0.028 | ||
| PAV/PAV (supertaster) | −0.15 (−0.28, −0.03) | 0.018 | −0.46 (−0.78, −0.14) | 0.004 | ||
| BMI model | 0.003 | |||||
| Additive TAS2R38 diplotype, per PAV | −0.07 (−0.13, −0.01) | — | 0.021 | −0.23 (−0.38, −0.08) | — | |
| AVI/AVI (nontaster) | Reference | — | Reference | — | ||
| AVI/PAV (taster) | −0.06 (−0.15, 0.04) | 0.256 | −0.27 (−0.50, −0.03) | 0.030 | ||
| PAV/PAV (supertaster) | −0.15 (−0.28, −0.02) | 0.019 | −0.44 (−0.76, −0.13) | 0.006 | ||
Data are given as β (95% CI) glucose levels from generalized linear models adjusted for age, sex, 10 genetic principal components, fasting time (for random glucose only), assessment center (base model), and BMI (BMI model). Total n available for each outcome was 218,688 for random and 57,652 for 0–2-h glucose. Significance was set at P <0.05/2 for two outcomes.
*P value based on the categorical TAS2R38 diplotype exposure for comparison vs. AVI/AVI nontasters (reference group).
†P value based on the additive TAS2R38 diplotype exposure for linear trends per PAV haplotype.
We next evaluated the associations between TAS2R38 diplotype and 0–2-h glucose to reflect postprandial glucose responses. Within the 0–2-h window, each additional PAV haplotype was associated with a change of −0.24 mg/dL (95% CI −0.39, −0.08) in glucose (Ptrend = 0.003), reflecting a greater magnitude and strength of TAS2R38 effects on postprandial compared with random glucose (β = −0.07 mg/dL) (Table 2). Pairwise differences for supertasters versus nontasters also reached significance below the study-wide threshold (β = −0.46; Ppairwise = 0.004), whereas differences for tasters versus nontasters reached nominal significance with intermediate effect estimates (β = −0.27; Ppairwise = 0.028). Both associations were robust to BMI adjustment (Ppairwise = 0.006 and 0.03, respectively) (Table 2).
To evaluate whether these associations were specific to the 0–2-h time window, we determined the associations between TAS2R38 diplotype and glucose levels over subsequent time windows (Supplementary Table 5). In BMI-adjusted models, we found no significant associations between TAS2R38 diplotype and glucose levels in any later time window (β = −0.016 per PAV haplotype combined for all times >2 h; 95% CI −0.08, 0.05; Ptrend = 0.629). For more interpretable estimates, we calculated estimated marginal mean glucose levels by TAS2R38 diplotype during each fasting window (Fig. 2A) and based on the number of bitter taste–increasing alleles for each TAS2R38 variant for 0–2-h glucose (Fig. 2B). These data reaffirmed that the associations of TAS2R38 with glucose levels occurred specifically in the first 2 h after a meal.
Figure 2.
Associations of functional variants in TAS2R38 with 0–2-h glucose. A and B: Glucose levels were estimated from linear models adjusted for age, sex, 10 genetic principal components, assessment center, and BMI and are summarized as estimated marginal mean (95% CI) over each time window by TAS2R38 diplotype (A) and in the 0–2-h time window by n of bitter taste–increasing alleles for each functional variant (B). C: Associations of TAS2R38 diplotypes and functional variants with 0–2-h glucose (UKB) and 2-h glucose after 75-g OGTT (MAGIC) are shown as evidence of independent replication. **P < 0.025 (study-wide significance threshold).
As a means of establishing independent replication, we queried summary statistics from a published meta-analysis of nine discovery GWAS for 2-h glucose after a 75-g OGTT (gold-standard postprandial glucose assessment) by MAGIC investigators (n = 24,611 individuals of European ancestry without type 2 diabetes) (34). In this meta-analysis, we found that all three TAS2R38 functional variants were significantly associated with lower 2-h glucose (Fig. 2C), with similar relative magnitudes to those obtained at the variant level in our UKB sample, providing evidence of replication in an independent cohort with a direct assessment of postprandial glucose levels.
Based on prior evidence that TAS2R38 functional variants may have stronger associations with postprandial glucose levels in men than in women (23), we ran exploratory sex-stratified analyses. Indeed, the associations of TAS2R38 diplotype with 0–2-h glucose were significant only in men (β = −0.34; 95% CI −0.56, −0.11; Ptrend = 0.003 vs. β = −0.14; 95% CI −0.35, 0.07; Ptrend = 0.195 for women). In contrast, in women, both supertasters and tasters had lower postprandial glucose levels than nontasters, with a similar degree of differences among PAV carriers (Supplementary Table 6).
Assessing the Potential for Nonbehaviorally Mediated Effects of TAS2R38 on Glucose Homeostasis
Because functional variants in TAS2R38 have been implicated in dietary and lifestyle behaviors that may contribute to glucose metabolism (35), we assessed whether the observed associations with postprandial glucose could be behaviorally mediated. As noted above, we observed expected associations of TAS2R38 diplotype with dietary and lifestyle behaviors in our sample (Table 1 and Supplementary Table 2) and confirmed the quality of behavioral covariates for capturing variability in 0–2-h glucose compared with a covariate-only BMI-adjusted model (lifestyle and diet model: likelihood ratio test P = 6.11 × 10−69) (Fig. 3A). We then tested the robustness of associations between TAS2R38 and 0–2-h glucose to sequential adjustment for behavioral covariates, finding that additive TAS2R38 associations and pairwise differences among diplotypes remained significant, with marginal changes in effect estimates ranging from −3.4% to 1.03% after adjustment (lifestyle and diet model: β = −0.22 per PAV haplotype; 95% CI −0.38, −0.07; Ptrend = 0.004; β = −0.44 for difference between supertasters vs. nontasters; 95% CI −0.75, −0.12; Ppairwise = 0.006) (Fig. 3B).
Figure 3.
Nonbehaviorally mediated associations of TAS2R38 with 0–2-h glucose. A: Likelihood ratio test P values show variability in 0–2-h glucose explained by dietary and lifestyle covariates compared with covariate-only BMI-adjusted model. B: Associations of TAS2R38 diplotypes with 0–2-h glucose in BMI-adjusted models with sequential adjustment for behavioral covariates shown as β (95% CI). C: −log10(P) values for associations of TAS2R38 variants and TAS2R19 and TAS2R14 variants (related to caffeine and quinine perception, respectively) with dietary and lifestyle behaviors relevant to 0–2-h glucose. D: Estimated marginal mean (95% CI) 0–2-h glucose levels based on n of bitter taste–increasing alleles for variants in TAS2R38, TAS2R19, and TAS2R14 in BMI-adjusted models. **P <0.025 (study-wide significance threshold). PROP, 6-n-propylthiouracil; PTC, phenylthiocarbamide.
To further establish the plausibility of nonbehaviorally mediated effects, we selected variants in two other bitter taste receptor genes, known to influence bitter taste perception and associate with similar dietary and lifestyle traits as TAS2R38 but without proposed functions in glucose homeostasis, to serve as negative controls: rs10772420 in TAS2R14, which is associated with quinine perception, and rs2597979 in TAS2R19, which is associated with caffeine perception (36). We confirmed that both variants had similar patterns of association as TAS2R38 variants with dietary and lifestyle traits relevant to glucose homeostasis in our UKB sample (Fig. 3C). However, neither rs10772420 nor rs2597979 was significantly associated with 0–2-h glucose in BMI-adjusted models nor in models additionally adjusted for dietary and lifestyle behaviors (lifestyle and diet model: Ptrend = 0.595 and 0.615, respectively) (Fig. 3D), further supporting a behaviorally independent mechanism of TAS2R38 in postprandial glucose metabolism.
Finally, we considered whether our evidence could help identify an underlying mechanism. Because data on postprandial GLP-1 levels were not available in the UKB, we could not explicitly test whether TAS2R38-driven secretion of GLP-1 could explain the observed associations. However, because GLP-1 has established actions in promoting satiety, this mechanism could also be supported by an inverse association of TAS2R38 sensitivity and total energy intake. We tested this among participants with available data (n = 129,615) and found a small but significant inverse association per PAV haplotype (P = 0.047) (Table 1). Notably, no such associations were observed for either of the negative control variants (which lack proposed effects on GLP-1 secretion). Adjusting for total energy intake in the lifestyle and diet model also had minimal impact on the associations between TAS2R38 diplotype and 0-2-h glucose (Ptrend = 0.012) (Supplementary Table 7).
Discussion
In a large-scale analysis among European adults without type 2 diabetes, we found that functional variants in TAS2R38 were significantly associated with postprandial glucose levels. Specifically, PAV/PAV supertasters had significantly lower 0–2-h (postprandial) glucose levels than AVI/AVI nontasters, with stepwise reductions in glucose levels corresponding to the number of PAV haplotypes. Associations were replicated in an independent published GWAS of 2-h OGTT glucose and were robust after adjusting for dietary and lifestyle factors, consistent with our hypothesis and the prior experimental evidence suggesting direct TAS2R38 actions in postprandial glucose metabolism.
Our findings support emerging evidence implicating TAS2R38 receptors in glucose homeostasis. TAS2R38 is one of ∼25 bitter TAS2Rs, peripherally expressed in the gut, liver, heart, pancreas, brain, and lungs (4), that is activated by natural (e.g., isothiocyanates in cruciferous vegetables) and synthetic (e.g., phenylthiocarbamide) bitter-tasting compounds (37). In preclinical studies, supplementation or intragastric infusion with bittertasting ligands of TAS2Rs elicited GLP-1 secretion and lowered blood glucose in response to mixed-nutrient challenges in healthy adults and those with type 2 diabetes (12–14). In the one prior study evaluating the associations of TAS2R38 functional variations and glucose homeostasis, in ∼1,000 German adults (60% women) without type 2 diabetes, genetic tasters and supertasters (i.e., PAV carriers) had significantly lower 2-h area under the curve glucose than nontasters after an OGTT (23). Our findings in ∼220,000 adults of European ancestry without type 2 diabetes are consistent with these observations. We further parsed the effects by diplotype, showing stepwise decreases in postprandial glucose levels per PAV haplotype, reflecting a dose–response relationship based on TAS2R38 receptor sensitivity.
By using large-scale individual-level data to account for behavioral factors, we provide evidence in humans of nonbehaviorally mediated effects of TAS2R38 on glucose homeostasis. As expected (21–23), supertasters reported lower alcohol and cigarette use and less frequent fruit and vegetable intake than nontasters. They were also less likely to add salt to food, as reported in prior UKB studies (38). These behaviors are implicated in glucose homeostasis (35,39) and were associated with glucose levels in our UKB sample. However, adjusting for them had no meaningful effect on associations between TAS2R38 and glucose levels. Supertasters have also been occasionally reported to have lower BMI than nontasters (40,41) (attributed to differences in dietary and lifestyle behaviors [42]), which may lead to different glucose outcomes. However, we observed no differences in BMI across TAS2R38 diplotypes, and our findings were robust to BMI adjustment. Variants in other bitter taste receptor genes, related to bitter taste perception (36) and similar health-related behaviors as TAS2R38 but without proposed roles in glucose homeostasis (i.e., negative controls), were also not associated with glucose levels, further limiting the plausibility of the observed associations between TAS2R38 and glucose levels to be explained by health behaviors related to bitter taste perception.
Beyond taste perception, there is increasing evidence that TAS2R38 activation by bitter-tasting compounds elicits glycemic benefits via GLP-1 signaling (8). Two examples are berberine, a bitter-tasting compound reported to have similar hypoglycemic effects as metformin (43), and quercetin, a natural bitter flavonoid hypothesized to lower blood glucose. Both compounds activate gut-expressed TAS2R38 receptors and promote GLP-1 secretion (9–11); in vitro, the glycemic effects of quercetin were also attenuated by TAS2R38 inhibition (11). Based on these observations and the experimental data described above, it is possible that TAS2R38 sensitivity modulates postprandial GLP-1 secretion, leading to differences in postprandial glycemic responses. Genetic evidence for this, however, is limited, because only one GWAS of postprandial GLP-1 levels after an OGTT has been conducted, and the sample size was <3,000 individuals, likely limiting the statistial power (44).
We do not have objective measures of postprandial GLP-1 levels in the UKB to formally test whether TAS2R38-mediated GLP-1 secretion can provide a potential mechanism to explain the observed associations. Still, two key observations were consistent with a mechanism involving GLP-1. First, TAS2R38 diplotype was associated with glucose levels only during the time window when most GLP-1 is secreted (45). This is consistent with an association driven by GLP-1 secretion. Second, functional variants conferring greater TAS2R38 receptor sensitivity, but not variants in other bitter taste receptor genes, were associated with lower total energy intake. Considering the actions of GLP-1 in promoting satiety, this would be consistent with supertasters having greater postprandial GLP-1 secretion, which could help lower their total energy intake (46).
Our findings must be interpreted in the context of several limitations. As mentioned, we did not have data on postprandial GLP-1 levels, which prevented us from explicitly testing a proposed mechanism. We also used a candidate gene study to examine a single genetic locus, and the identified associations did not pass a genome-wide significance threshold (P <5 ×10−8). This design was leveraged to test a hypothesis rooted in preclinical data; though because it is observational, we cannot infer causality. While we leveraged a negative control paradigm to strengthen our causal inferences, the selected variants are imperfect controls, because they respond to different bitter stimuli (36) and may have yet undiscovered roles in nutrient metabolism or peripheral nutrient sensing in other tissues (4,5). Additionally, while we aimed to robustly adjust for dietary and lifestyle behaviors, the covariates were derived from brief FFQs or self-report surveys, which are subject to recall bias and confounding. We were also unable to estimate the intake of natural TAS2R38 agonists, nor did we have data on what participants ate in the meal before glucose assessment. To minimize the risk of measurement error, we performed a complete case analysis, although this might have introduced sampling bias. Finally, focusing on individuals without type 2 diabetes of European ancestry limited our ability to extrapolate these findings to individuals with prediabetes or diabetes or those of other ancestry groups, among whom the frequency of TAS2R38 genotypes, dietary habits, and their phenotypic consequences might differ (41,47).
Despite these limitations, our results and mounting preclinical evidence support continued investigation into TAS2R38 as a therapeutic target for glucose regulation. Targeted activation of TAS2R38 with bitter-tasting agonists may provide glycemic benefits by promoting endogenous GLP-1 secretion, in line with mechanisms of highly effective GLP-1 receptor agonists (48,49). TAS2R38 functional variants may also provide novel candidates for pharmacogenetic analyses of genetic variants affecting GLP-1 receptor agonist responses and exploration of why some patients report major changes in taste perception after initiating these medications. Beyond TAS2R38, other extraoral taste receptors may also function in glucose metabolism (50); clarifying their roles could reveal new glucoregulatory mechanisms and therapeutic strategies for type 2 diabetes and obesity (51,52).
In conclusion, in a large cohort of adults without type 2 diabetes, functional variants encoding more sensitive TAS2R38 bitter taste receptors were associated with lower postprandial glucose levels, which persisted after adjusting for lifestyle and dietary behaviors, providing evidence in humans suggesting a functional role of TAS2R38 bitter taste receptors in glucose homeostasis. Future studies are required to assess the effects of TAS2R38 haplotypes on postprandial GLP-1 responses in standardized meal challenges and to elaborate on the potential for TAS2R38 to serve as a novel therapeutic target for glucose regulation and type 2 diabetes prevention and treatment.
This article contains supplementary material online at https://doi.org/10.2337/figshare.30179119.
Article Information
The views and opinions expressed here are those of the authors only and do not necessarily reflect those of the European Union or European Innovation Council and Small and Medium-Sized Enterprises Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.
Duality of Interest. M.S.U. is involved in research collaboration with Novo Nordisk, unrelated to content of this manuscript. J.M. is an associate editor for Diabetologia, unrelated to the evaluation of this manuscript. S.J.C. reports a close family member employed by a Johnson & Johnson company. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. J.E.G. conducted the analysis and drafted the manuscript. J.E.G., S.J.C., and M.S.U. designed the analysis plan. K.E.W., J.B.C., J.M., S.J.C., and M.S.U. provided critical feedback on the analysis and manuscript. All authors reviewed and approved the final manuscript. J.E.G. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Prior Presentation. This work was presented in preliminary form at the 2024 Precision Nutrition Forum, Copenhagen, Denmark, 16–17 April 2024, and the 2024 American Diabetes Association Scientific Sessions, Orlando, FL, 21–24 June 2024.
Funding Statement
J.E.G. and S.J.C. are supported by the American Diabetes Association (7-21-JDFM-005). K.E.W. and J.B.C. are supported by the National Institutes of Health (K01DK133637 and R00DK127196, respectively). J.M. is supported by the Novo Nordisk Foundation (NNF23SA0084103), a European Foundation for the Study of Diabetes/Novo Nordisk Foundation Future Leaders Award (0094134), and the European Union (HORIZON-EIC-2023-PATHFINDERCHALLENGES-01-101161509). M.S.U. is supported by the Doris Duke Foundation (Award 2022063), the MGH Executive Committee on Research (Claflin Distinguished Scholar Award), and the NIH/NIDDK (U01DK140757).
Supporting information
References
- 1. Drucker DJ. GLP-1-based therapies for diabetes, obesity and beyond. Nat Rev Drug Discov 2025;24:631–650 [DOI] [PubMed] [Google Scholar]
- 2. Holst JJ, Gribble F, Horowitz M, Rayner CK. Roles of the gut in glucose homeostasis. Diabetes Care 2016;39:884–892 [DOI] [PubMed] [Google Scholar]
- 3. Senior M. After GLP-1, what’s next for weight loss? Nat Biotechnol 2023;41:740–743 [DOI] [PubMed] [Google Scholar]
- 4. Calvo SS-C, Egan JM. The endocrinology of taste receptors. Nat Rev Endocrinol 2015;11:213–227 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Sternini C, Rozengurt E. Bitter taste receptors as sensors of gut luminal contents. Nat Rev Gastroenterol Hepatol 2025;22:39–53 [DOI] [PubMed] [Google Scholar]
- 6. Wu SV, Rozengurt N, Yang M, Young SH, Sinnett-Smith J, Rozengurt E. Expression of bitter taste receptors of the T2R family in the gastrointestinal tract and enteroendocrine STC-1 cells. Proc Natl Acad Sci U S A 2002;99:2392–2397 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Jang H-J, Kokrashvili Z, Theodorakis MJ, et al. Gut-expressed gustducin and taste receptors regulate secretion of glucagon-like peptide-1. Proc Natl Acad Sci U S A 2007;104:15069–15074 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Pham H, Hui H, Morvaridi S, et al. A bitter pill for type 2 diabetes? The activation of bitter taste receptor TAS2R38 can stimulate GLP-1 release from enteroendocrine L-cells. Biochem Biophys Res Commun 2016;475:295–300 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Yu Y, Hao G, Zhang Q, et al. Berberine induces GLP-1 secretion through activation of bitter taste receptor pathways. Biochem Pharmacol 2015;97:173–177 [DOI] [PubMed] [Google Scholar]
- 10. Yue X, Liang J, Gu F, Du D, Chen F. Berberine activates bitter taste responses of enteroendocrine STC-1 cells. Mol Cell Biochem 2018;447:21–32 [DOI] [PubMed] [Google Scholar]
- 11. Wang C, Zhou J, Liu X, et al. Quercetin enhances GLP-1 secretion via TAS2R38-mediated PLC signaling in enteroendocrine L-cells. Mol Nutr Food Res 2025;69:e70109. [DOI] [PubMed] [Google Scholar]
- 12. Mohammadpour Z, Heshmati E, Heilbronn LK, Hendrie GA, Brooker PG, Page AJ. The effect of post-oral bitter compound interventions on the postprandial glycemia response: a systematic review and meta-analysis of randomised controlled trials. Clin Nutr 2024;43:31–45 [DOI] [PubMed] [Google Scholar]
- 13. Bitarafan V, Fitzgerald PCE, Little TJ, et al. Intragastric administration of the bitter tastant quinine lowers the glycemic response to a nutrient drink without slowing gastric emptying in healthy men. Am J Physiol Regul Integr Comp Physiol 2020;318:R263–R273 [DOI] [PubMed] [Google Scholar]
- 14. Bitarafan V, Anjom-Shoae J, Rezaie P, et al. Dose-related effects of intraduodenal quinine on plasma glucose, glucoregulatory hormones and gastric emptying of a nutrient drink, and energy intake, in men with type 2 diabetes: a double-blind, randomised, crossover study. Diabetologia 2025;68:727–738 [DOI] [PubMed] [Google Scholar]
- 15. Rezaie P, Bitarafan V, Horowitz M, Feinle-Bisset C. Effects of bitter substances on GI function, energy intake and glycaemia—do preclinical findings translate to outcomes in humans? Nutrients 2021;13:1317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kim U-k, Jorgenson E, Coon H, Leppert M, Risch N, Drayna D. Positional cloning of the human quantitative trait locus underlying taste sensitivity to phenylthiocarbamide. Science 2003;299:1221–1225 [DOI] [PubMed] [Google Scholar]
- 17. Mennella JA, Pepino MY, Duke FF, Reed DR. Psychophysical dissection of genotype effects on human bitter perception. Chem Senses 2011;36:161–167 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Tan J, Abrol R, Trzaskowski B, Goddard WA. 3D structure prediction of TAS2R38 bitter receptors bound to agonists phenylthiocarbamide (PTC) and 6-n-propylthiouracil (PROP). J Chem Inf Model 2012;52:1875–1885 [DOI] [PubMed] [Google Scholar]
- 19. Knaapila A, Hwang L-D, Lysenko A, et al. Genetic analysis of chemosensory traits in human twins. Chem Senses 2012;37:869–881 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Cornelis MC, van Dam RM. Genetic determinants of liking and intake of coffee and other bitter foods and beverages. Sci Rep 2021;11:23845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Duffy VB, Davidson AC, Kidd JR, et al. Bitter receptor gene (TAS2R38), 6-n-propylthiouracil (PROP) bitterness and alcohol intake. Alcohol Clin Exp Res 2004;28:1629–1637 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Calancie L, Keyserling TC, Taillie LS, et al. TAS2R38 predisposition to bitter taste associated with differential changes in vegetable intake in response to a community-based dietary intervention. G3 (Bethesda) 2018;8:2107–2119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Keller M, Liu X, Wohland T, et al. TAS2R38 and its influence on smoking behavior and glucose homeostasis in the German Sorbs. PLoS One 2013;8:e80512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Sudlow C, Gallacher J, Allen N, et al. UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med 2015;12:e1001779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Bycroft C, Freeman C, Petkova D, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature 2018;562:203–209 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Karczewski KJ, Gupta R, Kanai M, et al. Pan-UK Biobank GWAS improves discovery, analysis of genetic architecture, and resolution into ancestry-enriched effects. 15 March 2024 [preprint] 10.1101/2024.03.13.24303864 [DOI] [Google Scholar]
- 27. Eastwood SV, Mathur R, Atkinson M, et al. Algorithms for the capture and adjudication of prevalent and incident diabetes in UK Biobank. PLoS One 2016;11:e0162388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Delaneau O, Zagury J-F, Robinson MR, Marchini JL, Dermitzakis ET. Accurate, scalable and integrative haplotype estimation. Nat Commun 2019;10:5436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Guo W, Bradbury KE, Reeves GK, Key TJ. Physical activity in relation to body size and composition in women in UK Biobank. Ann Epidemiol 2015;25:406–413.e6 [DOI] [PubMed] [Google Scholar]
- 30. Ge T, Chen C-Y, Doyle AE, et al. The shared genetic basis of educational attainment and cerebral cortical morphology. Cereb Cortex 2019;29:3471–3481 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Westerman KE, Walker ME, Gaynor SM, et al. Investigating gene–diet interactions impacting the association between macronutrient intake and glycemic traits. Diabetes 2023;72:653–665 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Westerman KE, Miao J, Chasman DI, et al. Genome-wide gene-diet interaction analysis in the UK Biobank identifies novel effects on hemoglobin A1c. Hum Mol Genet 2021;30:1773–1783 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Lagou V, Jiang L, Ulrich A, et al.; Meta-Analysis of Glucose and Insulin-Related Traits Consortium (MAGIC) . GWAS of random glucose in 476,326 individuals provide insights into diabetes pathophysiology, complications and treatment stratification. Nat Genet 2023;55:1448–1461 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Saxena R, Hivert M-F, Langenberg C, et al.; MAGIC investigators . Genetic variation in GIPR influences the glucose and insulin responses to an oral glucose challenge. Nat Genet 2010;42:142–148 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Berry SE, Valdes AM, Drew DA, et al. Human postprandial responses to food and potential for precision nutrition. Nat Med 2020;26:964–973 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Hwang L-D, Gharahkhani P, Breslin PAS, et al. Bivariate genome-wide association analysis strengthens the role of bitter receptor clusters on chromosomes 7 and 12 in human bitter taste. BMC Genomics 2018;19:678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Bayer S, Mayer AI, Borgonovo G, Morini G, Di Pizio A, Bassoli A. Chemoinformatics view on bitter taste receptor agonists in food. J Agric Food Chem 2021;69:13916–13924 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Cole JB, Florez JC, Hirschhorn JN. Comprehensive genomic analysis of dietary habits in UK Biobank identifies hundreds of genetic associations. Nat Commun 2020;11:1467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Huber H, Schieren A, Holst JJ, Simon M-C. Dietary impact on fasting and stimulated GLP-1 secretion in different metabolic conditions—a narrative review. Am J Clin Nutr 2024;119:599–627 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Trius-Soler M, Bersano-Reyes PA, Góngora C, Lamuela-Raventós RM, Nieto G, Moreno JJ. Association of phenylthiocarbamide perception with anthropometric variables and intake and liking for bitter vegetables. Genes Nutr 2022;17:12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Chupeerach C, Tapanee P, On-Nom N, et al. The influence of TAS2R38 bitter taste gene polymorphisms on obesity risk in three racially diverse groups. Biomedicine (Taipei) 2021;11:43–49 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Sharafi M, Rawal S, Fernandez ML, Huedo-Medina TB, Duffy VB. Taste phenotype associates with cardiovascular disease risk factors via diet quality in multivariate modeling. Physiol Behav 2018;194:103–112 [DOI] [PubMed] [Google Scholar]
- 43. Chou W-L. Therapeutic potential of targeting intestinal bitter taste receptors in diabetes associated with dyslipidemia. Pharmacol Res 2021;170:105693. [DOI] [PubMed] [Google Scholar]
- 44. Almgren P, Lindqvist A, Krus U, et al. Genetic determinants of circulating GIP and GLP-1 concentrations. JCI Insight 2017;2:e93306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Xie C, Huang W, Watson LE, et al. Plasma GLP-1 response to oral and intraduodenal nutrients in health and type 2 diabetes—impact on gastric emptying. J Clin Endocrinol Metab 2022;107:e1643–e1652 [DOI] [PubMed] [Google Scholar]
- 46. Martin B, Dotson CD, Shin Y-K, et al. Modulation of taste sensitivity by GLP-1 signaling in taste buds. Ann N Y Acad Sci 2009;1170:98–101 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Risso DS, Mezzavilla M, Pagani L, et al. Global diversity in the TAS2R38 bitter taste receptor: revisiting a classic evolutionary PROPosal. Sci Rep 2016;6:25506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Gratzl S, Rodriguez PJ, Cartwright BMG, Baker C, Stucky NL. Monitoring report: GLP-1 RA prescribing trends—December 2023. 11 October 2024 [preprint] 10.1101/2024.01.18.24301500 [DOI] [Google Scholar]
- 49. Watanabe JH, Kwon J, Nan B, Reikes A. Trends in glucagon-like peptide 1 receptor agonist use, 2014 to 2022. J Am Pharm Assoc (2003) 2024;64:133–138 [DOI] [PubMed] [Google Scholar]
- 50. Dotson CD, Zhang L, Xu H, et al. Bitter taste receptors influence glucose homeostasis. PLoS One 2008;3:e3974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. D’Urso O, Drago F. Pharmacological significance of extra-oral taste receptors. Eur J Pharmacol 2021;910:174480. [DOI] [PubMed] [Google Scholar]
- 52. Trius-Soler M, Moreno JJ. Bitter taste receptors: key target to understand the effects of polyphenols on glucose and body weight homeostasis. Pathophysiological and pharmacological implications. Biochem Pharmacol 2024;228:116192. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.