Impaired Glucose Metabolism in Mice Lacking the Tas1r3 Taste Receptor Gene

The G-protein-coupled sweet taste receptor dimer T1R2/T1R3 is expressed in taste bud cells in the oral cavity. In recent years, its involvement in membrane glucose sensing was discovered in endocrine cells regulating glucose homeostasis. We investigated importance of extraorally expressed T1R3 taste receptor protein in age-dependent control of blood glucose homeostasis in vivo, using nonfasted mice with a targeted mutation of the Tas1r3 gene that encodes the T1R3 protein. Glucose and insulin tolerance tests, as well as behavioral tests measuring taste responses to sucrose solutions, were performed with C57BL/6ByJ (Tas1r3+/+) inbred mice bearing the wild-type allele and C57BL/6J-Tas1r3tm1Rfm mice lacking the entire Tas1r3 coding region and devoid of the T1R3 protein (Tas1r3-/-). Compared with Tas1r3+/+ mice, Tas1r3-/- mice lacked attraction to sucrose in brief-access licking tests, had diminished taste preferences for sucrose solutions in the two-bottle tests, and had reduced insulin sensitivity and tolerance to glucose administered intraperitoneally or intragastrically, which suggests that these effects are due to absence of T1R3. Impairment of glucose clearance in Tas1r3-/- mice was exacerbated with age after intraperitoneal but not intragastric administration of glucose, pointing to a compensatory role of extraoral T1R3-dependent mechanisms in offsetting age-dependent decline in regulation of glucose homeostasis. Incretin effects were similar in Tas1r3+/+ and Tas1r3-/- mice, which suggests that control of blood glucose clearance is associated with effects of extraoral T1R3 in tissues other than the gastrointestinal tract. Collectively, the obtained data demonstrate that the T1R3 receptor protein plays an important role in control of glucose homeostasis not only by regulating sugar intake but also via its extraoral function, probably in the pancreas and brain.


Introduction
The search for key regulators of blood or tissue glucose levels is relevant to treatments of diabetes, obesity, and metabolic syndrome. Sensing of glucose in viscera and brain is crucial for control of energy homeostasis. Cells involved in regulation of blood glucose concentration (the insulin-secreting β-cells of the pancreas, enteroendocrine L-cells of the small intestine, and glucose-excited neurons of ventromedial hypothalamus) share a metabolic mechanism of glucose #A5952-01). The study was performed with 8-to 36-week-old male mice of two strains: C57BL/6ByJ bearing the wild-type Tas1r3 allele, used as control (Tas1r3+/+; Jackson Laboratory, Bar Harbor, ME), and C57BL/6J-Tas1r3 tm1Rfm lacking the entire T1R3 coding region and devoid of T1R3 protein [28] (Tas1r3-/-; kindly provided by Dr. R. F. Margolskee, Monell Chemical Senses Center, Philadelphia, PA, USA). Separate groups of mice were used in different tests. (Numbers of mice are shown in the table and figures below.) During the study, animals were housed individually (taste tests) or by 4-5 in standard polycarbonate cages on a 12-h light-dark cycle (lights on at 8:00 a.m.) in a temperature-and humidity-controlled room. Throughout the study, mice were fed with a standard lab chow (PK-120, MEST Ltd., Moscow, Russia) containing 67% carbohydrates, 5% lipids, and 19% proteins, with an energy value of 13,000 kJ/kg; food and tap water were available ad libitum.

Taste tests
Behavioral taste responses to 0.03-0.93 mol/L sucrose (Sigma-Aldrich St. Louis, MO, USA) were assessed in separate groups of mice using the brief-access licking test (BALT) and the 48-h two-bottle preference test. The BALT was conducted during the light period using procedures similar to those described by Glendinning et al. [31]. Before testing, mice had restricted access to water (1.5 mL for 22-23 h), while access to food remained unlimited. During the test session, an animal was exposed in gustometer Davis MS-160 (DiLog Instruments, Tallahassee, FL, USA) to three repetitive blocks of stimuli, each consisting of eight trials: six concentrations of sucrose presented in ascending order, and two presentations of distilled water as "washout," one before each of the two highest sucrose concentrations. Access to each solution lasted for 5 s, with 20-s interpresentation interval. Licking ratio was calculated as the percentage of licks to sucrose solution relative to the mean number of licks in the two preceding trials with water. In the 48-h two-bottle tests [32], mice in their home cages had free access to two tubes containing distilled water or sucrose solution. Positions of tubes was changed after 24 h. Sucrose preference was calculated as consumed sucrose solution as a percentage of total fluid intake.

Glucose and insulin tolerance tests
Glucose and insulin tolerance tests were started at 3-4 p.m. (7-8 h after the beginning of the light period). In the glucose tolerance test (GTT), matching specific recommendations of the Mouse Metabolic Phenotyping Center [25], nonfasted conscious animals received glucose (2 g/ kg, 0.1 ml per 10 g body weight; Sigma-Aldrich St. Louis, MO, USA) either intraperitoneally (IP) or by intragastric gavage (IG). Aqueous solution was used in the IG GTT; in the IP GTT, glucose was dissolved in saline. In the insulin tolerance test (ITT), nonfasted mice were injected with insulin (2 U/kg, IP; insulin aspart, Novo Nordisk A/S, Bagsvaerd, Denmark). Blood was sampled by tail cut, and two measurements of glucose concentration for each time point were made 0-120 min after the infusion of glucose or insulin using the One Touch Ultra glucometer (LifeScan, Inc., USA). During the GTT, animals were gently held in custom-made restraint tubes, to which they were habituated during the preceding 2 days. During the ITT, mice were left unrestrained in their home cages.

Statistical analysis
Statistical analysis was performed using Statistica 7.0 software (StatSoft, Tulsa, OK, USA). Data from the behavioral taste tests and glucose clearance in the GTT and ITT were compared with two-way ANOVA. Concentration (for taste tests) and time (for GTT or ITT) were considered as within-subject factors, and strain was considered as a between-subject factor. Post hoc paired comparisons were made with Fisher's least significant difference (LSD) test. Blood glucose area under the curve (AUC) was calculated using the trapezoidal rule, and differences between AUCs were assessed with one-way ANOVA. Between-strain comparisons of baseline glucose, body weight, and age were performed using the Student's t-test. To quantify the correlation of physiological parameters within groups, the Pearson product-moment correlation coefficient was used. All data are presented as mean ± SEM; P values less than 0.05 were considered significant.

Results
Body weight of Tas1r3-/-mice was about 6% greater than body weight of Tas1r3+/+ mice (ttest, p<0.01), and for both strains it increased with age (Table 1). Baseline blood glucose level was similar in both strains. There was no significant relationship between baseline glucose and body weight or age (Table 1).
Injection of insulin (2 U/kg, IP) caused a rapid reduction of blood glucose concentration, reaching a minimum 15 min after injection (Fig 4B and 4C). In both strains, basal glucose level was completely restored within 120 min. Hypoglycemia induced by insulin did not depend on age (Fig 4A) or body weight in either strain (data not shown); therefore, calculations were made for the combined group with ages ranging from 8 to 36 weeks. Analysis of both absolute data ( Fig 4B)

Discussion
Our results support the important role of cell membrane sensing of sweeteners with T1R3 taste receptor protein at different levels of control of carbohydrate ingestion and homeostasis. Mice lacking the Tas1r3 gene demonstrated reduced behavioral taste responses to sucrose: they had totally suppressed attraction to sucrose when they were presented with different concentrations for 5-s periods in the BALT (Fig 1A) and had lower preferences for sucrose than Tas1r3+/+ mice in the 48-h two-bottle test. However, unlike during the BALT, during the long-term preference tests Tas1r3-/-mice preferred higher concentrations of sucrose (!0.23 mol/L) over water (Fig 1B). These data are consistent with earlier studies of Tas1r3-/-mice, which revealed a residual behavioral preference of caloric sugars but not nonnutritive sweeteners such as sucralose, acesulfame K, SC-45647, and saccharin, in long-term preference tests [28,33]. Both Tas1r3-independent pre-and postingestive mechanisms could be involved in the observed residual preference of sucrose. The Tas1r3-independent preingestive effects are supported by results from a recent study with Tas1r3-/-, TRPM5-/-, and Gα-gustducin-knockout mice, in which chorda tympani nerve activity in response to sucrose was reduced by roughly 80% relative to wild-type mice but responses to glucose were reduced only by 40% [34]. This suggests that in taste cells, sugars such as sucrose and glucose can activate an additional taste transduction pathway that does not require T1R3, gustducin's Gα subunit, or TRPM5 and was proposed to be metabolic and K ATP channel dependent [35]. The Tas1r3-independent postingestive factors likely include conditioned flavor preference reinforced by nutritive value of sucrose [36].
Our results demonstrate that in nonfasted mice Tas1r3 deficiency markedly worsens glucose tolerance, regardless of whether the route of glucose administration is intragastric or intraperitoneal (Figs 2 and 3), indicating possible involvement of T1R3-mediated glucose sensing in intestinal enteroendocrine, pancreatic, and/or brain mechanisms controlling glucose metabolism. It is well established that T1R3 is expressed in a variety of tissues beyond the tongue and gut mucosa (e.g., [9][10][11][12][13][14][15]; however, it is still not clear to what extent these extraoral taste receptors are involved in control of carbohydrate metabolism. In early studies in the human pancreas, T1R3 was immunolabeled in excretory ducts and centroacinar cells, but the endocrine portion of the gland was immunonegative [37]. Later, RT-PCR showed expression of the TAS1R3 gene in human pancreatic islets [22] and in MIN6 cells, a glucose-responsive β-cell line [16]. Mouse islets [2] and MIN6 cells [17] express elements of intracellular taste signal transduction cascade as well. The sweet taste receptor system of mouse pancreatic β-cells and MIN6 cells seems functional since artificial sweeteners are able to stimulate insulin secretion, which was attenuated by gurmarin, an inhibitor of the mouse sweet taste receptor [16,22]. In human pancreatic islets, potentiation of insulin release induced by fructose was suppressed by lactisole, an allosteric inhibitor of human T1R3. Further, in vitro, genetic ablation of T1R2 or T1R3 led to substantial reduction of the effect of sweeteners on insulin output from mouse islets [19,22].
In contrast with these results of in vitro studies, recent in vivo studies in food-deprived mice revealed that the lack of T1R2 [22] or T1R3 [19] had no significant effect on the blood glucose level after IP administration of glucose, although after IG glucose administration Tas1r3-/mice had higher blood glucose and lower plasma insulin levels than did wild-type controls [19]. A likely explanation for this discrepancy between in vitro and in vivo results is the difference in nutrition status of cells. In cultured mouse islets, positive effects of fructose or noncaloric sweeteners on insulin secretion require presence of an optimal glucose level in the medium. For instance, a sharp reduction of glucose concentration in islet media abolished the potentiating effect of fructose [22] and stimulated activity of noncaloric sweeteners [16] in MIN6 cells. Therefore, pre-experimental fasting can also influence results of in vivo experiments. Overnight fasting provokes a catabolic state in mice, which have a unique metabolic response to prolonged fasting that differs from the response to fasting seen in humans. Specifically, fasting impairs insulin-stimulated glucose utilization in humans but enhances it in normal mice [26,27]. In mice and rats, fasting, or even mild caloric deprivation, leads to the increase in insulin binding in the tissues [38,39]. Earlier, we found out that effect of T1R3 ablation on glucose utilization was more pronounced in euglycaemic state than after fasting [40]. The present data show that in mice in the nonfasted state, when β-cells are already partially depolarized due to K ATP -dependent mechanisms [22,35] and maintain basal levels of insulin secretion, deletion of T1R3 causes a significant impairment of glucose tolerance in both IP GTT and IG GTT. Thus, the apparent discrepancy between our data and these previous results is likely due to prolonged fasting used in these other studies, which likely caused marked changes in blood glucose regulation mechanisms.
Potentially, T1R3-signalling could be also involved in regulation of glycogenolysis and/or gluconeogenesis. In the adipose tissue, T1R3-signalling induced by non-caloric sweeteners stimulates adipogenesis and suppresses lipolysis [41]. T1R3 is expressed in excretory ducts of the liver [37], where it probably does not interact with glycogenolysis. However, an involvement of T1R3-dependent mechanisms in fat, liver and other tissues in direct or indirect control of glycogen breakdown and gluconeogenesis in our study is unlikely because animals in all our experiments were in non-fasting state, in which synthesis of glucose from polysaccharides and non-carbohydrates is suppressed.
Additionally, our results demonstrate that in the IP (but not IG) GTT, the effect of T1R3 deletion was age related (Fig 2A), suggesting that normal T1R3-mediated extraoral sensing of sweeteners somewhat prevents deterioration of glucose tolerance with age. Decreased insulin secretion due to the loss of β-cell mass or impaired β-cell function and increased insulin resistance are considered two major factors leading to impaired glucose tolerance in the elderly [42][43][44].
According to the classical concept, the oral ingestion of glucose stimulates more insulin release than does intravenous infusion while causing a similar elevation of the plasma glucose level [45]. This phenomenon, known as the incretin effect, is largely attributable to two insulinotropic hormones released in response to food ingestion from intestinal enteroendocrine Kcells (GIP) or L-cells (GLP-1). Both GIP and GLP-1 have direct stimulatory effects on pancreatic β-cells (for review see [46]). The combined action of incretins is believed to account for about 50% of the total insulin secretory response after a meal [47]. In recent years, sweet taste molecules, including T1R3, as well as intracellular taste signal transduction machinery in the gut enteroendocrine cells were described among the regulators of incretin production. Immunolabeling has revealed taste signal transducing elements in a number of intestinal L-cells, ranging from 15% in mouse jejunum up to 90% in the human duodenum [11,48], whereas Kcells likely express only marginal levels of sweet taste protein transcripts [14]. The artificial sweetener sucralose administrated to the mouse enteroendocrine GLUTag cell line or to the human L-cell line NCI-H716 enhances GLP-1 output that could be blocked by species-specific inhibitors of the sweet taste receptors [14,18]. Knockout mice lacking T1R3, or ileum explants from these mice, showed markedly reduced GLP-1 release in response to luminal infusion of glucose [19,20]. Consistent with this and with our results (Fig 3), Tas1r3-/-mice had higher blood glucose and lower plasma insulin levels during an oral glucose challenge compared with wild-type controls [19]. However, in mouse or rat duodenum and jejunum, only a small number of taste proteins are colocalized with enteroendocrine cells [13,49], and there is still no convincing evidence that T1R3-dependent intestinal endocrine mechanisms are potent enough to control blood glucose levels in vivo.
In our study, like in classical investigations [45], the involvement of T1R3 in regulation of intestinal secretion of incretins could be evident by comparing blood glucose clearance after administration of the same dose of glucose by different routes. We show that both Tas1r3+/+ and Tas1r3-/-mice demonstrated similar incretin effects (S1 Fig): in both types of mice blood glucose clearance was more active after IG glucose administration than after IP administration. Pancreatic β-cells and gut enteroendocrine cells use a common metabolic mechanism of glucose sensing, which requires glucose transporter GLUT2, the glycolytic enzyme glucokinase, and the K ATP channel [50][51][52]. Therefore, because the route of glucose administration affected blood glucose clearance in Tas1r3-/-mice, we suggest that in the euglycemic state K ATP -dependent metabolic mechanisms predominantly determine gut regulation of the glucose homeostasis.
Impaired glucose tolerance is usually associated with reduced insulin sensitivity, which was also demonstrated for Tas1r3-/-mice in our study (Fig 4A). Higher body mass of Tas1r3-/mice could have contributed to their lower insulin sensitivity, but the difference in body weight was small (about 6%, Table 1), and body weight did not correlate with glucose level. Reduction of insulin tolerance also did not correlate with age ( Fig 4B) and body weight. Therefore, higher body weight of Tas1r3-/-mice seems insufficient to explain their reduced insulin sensitivity. Another possible cause of decreased insulin sensitivity of Tas1r3-/-mice could be chronic elevation of postprandial glucose level, which was shown in our glucose tolerance experiments. In particular, raised blood glucose levels cause overactivity of the hexosamine biosynthesis pathway of glycolysis via modulation of transcriptional factors by O-N-acetylglucosamine, including transcriptional factors of the insulin receptor substrate and probably GLUT4 (for review see [53]), which may lead to reduced insulin sensitivity observed in Tas1r3-/-mice.
There is evidence that in addition to the gastrointestinal tract and pancreas, the central nervous system may have sweet taste signaling mechanisms that play an important role in regulating glucose homeostasis and therefore may be involved in effects of T1R3 deficiency found in this study. The fall of central glucose levels causes a sequence of neurohormonal reactions known as feedback response launched mainly by activation of glucose-sensing neurons in ventromedial hypothalamic nuclei, orexin neurons in perifornical area, and neurons in the brainstem [54][55][56]; this includes sympathoadrenal activation followed by increases of plasma epinephrine, norepinephrine, and glucagon, which in turn leads to hepatic gluconeogenesis and inhibition of pancreatic insulin secretion [57]. An acute increase in central glucose, which likely occurs in our experimental protocol, results in an opposite response: an increase in insulin levels and suppression of hepatic glucose production through reduction of gluconeogenesis and glycogenolysis [58]. Several mechanisms of glucose sensing, which do not require intracellular glucose metabolism or glucokinase/K ATP pathways, have been demonstrated in the hypothalamus (for review see [59]). It is quite plausible that glucosensing neurons could use a sweet taste receptor. Ren et al. [15] have reported that T1Rs and α-gustducin are highly expressed in neurons of mouse hypothalamus compared with cortex and hippocampus. Strong expression of T1R2 and T1R3 was found in arcuate and paraventricular nuclei of the hypothalamus, as well as in the medial habenula and the epithelial cells of the choroid plexus. Importantly, the arcuate nucleus is a key region detecting peripheral metabolic status and then relaying this information to other hypothalamic nuclei, including the ventromedial nucleus and paraventricular nuclei 53].
Most of glucose-sensing neurons are glucose inhibited (GI) and reduce their activity during elevation of blood glucose above the euglycemic level. Both glucose-excited (GE) and GI neurons of the hypothalamus are extremely sensitive to glucose changes when extracellular concentrations are less than 2 mM, which occurs at euglycemic blood levels [60], and they have minimal response when glucose levels in the hypothalamus exceed 2 mM, suggesting that these glucose-sensing neurons primarily sense glucose deficit [5,61]. An additional smaller population of glucose-sensing units is present in the arcuate nucleus, which include high-GE and high-GI neurons responding to an increase in extracellular glucose from 5 to 20 mM; however, it is still not clear whether these neurons play role in regulating hyperglycemic states. It is interesting that glucose sensing of the high-GE and high-GI neurons is K ATP independent [6]. Collectively, there is evidence that glucose-metabolism-independent pathways in the central nervous system may involve T1Rs and take place under control of peripheral glucose homeostasis in the hyperglycemic state; however, their role needs to be elucidated. Additional investigations of neurotransmitter and hormonal specificity of hypothalamic neurons expressing T1R3 possibly will shed light on their physiological relevance.
Measures of glucose metabolism used in this study may have been affected by direct effects of absence of the T1R3 protein in extraoral cells, as well as by indirect effects from these cells to other tissues (e.g., mediated by metabolic, hormonal or paracrine effects). The goal of our ongoing studies is to find out, which of these mechanisms are involved in impaired glucose metabolism in Tas1r3-/-mice. If genetic variants of the Tas1r3 gene alter glucose metabolism in mice, then similar relationships may also exist in humans. Human T1R genes are polymorphic [62], and some of these polymorphisms are associated with taste functions [63][64][65][66][67]. Our study suggests that these functional polymorphisms of human T1R genes may also be associated with glucose metabolism and related diseases in humans. This emphasizes importance of human studies of T1R genes as potential new targets for diagnostics, prevention and treatment of metabolic diseases.
In conclusion, we have shown that the lack of attraction to sucrose demonstrated in Tas1r3-/-mice, compared with Tas1r3+/+ mice, is associated with reduced glucose tolerance in these mice. In nonfasted mice, the deletion of the T1R3 subunit of the sweet taste receptor results in substantial impairment of blood glucose clearance after both intragastric and intraperitoneal glucose administration. This clearly indicates involvement of the extraorally expressed T1R3 protein in control of glucose homeostasis in hyperglycemic states. Deletion of T1R3 had minor impact on the incretin effect, which suggests that intestinally expressed T1R3 protein is less important for regulation of blood glucose level compared with other extraoral sites, such as pancreas or brain. Reduced glucose tolerance after T1R3 deletion was associated with impaired insulin sensitivity. We have also demonstrated a marked age dependence of the effect of T1R3 receptor protein on blood glucose levels in the intraperitoneal glucose tolerance test. Altogether, our results suggest that further investigation of visceral reception of sugars with the T1R3 protein may lead to therapeutic approaches in the treatment of carbohydrate homeostasis disorders.