Sweet Taste Receptors Regulate Basal Insulin Secretion and Contribute to Compensatory Insulin Hypersecretion During the Development of Diabetes in Male Mice

Abstract

β-Cells rapidly secrete insulin in response to acute increases in plasma glucose but, upon further continuous exposure to glucose, insulin secretion progressively decreases. Although the mechanisms are unclear, this mode of regulation suggests the presence of a time-dependent glucosensory system that temporarily attenuates insulin secretion. Interestingly, early-stage β-cell dysfunction is often characterized by basal (ie, fasting) insulin hypersecretion, suggesting a disruption of these related mechanisms. Because sweet taste receptors (STRs) on β-cells are implicated in the regulation of insulin secretion and glucose is a bona fide STR ligand, we tested whether STRs mediate this sensory mechanism and participate in the regulation of basal insulin secretion. We used mice lacking STR signaling (T1R2^−/− knockout) and pharmacologic inhibition of STRs in human islets. Mouse and human islets deprived of STR signaling hypersecrete insulin at short-term fasting glucose concentrations. Accordingly, 5-hour fasted T1R2^−/− mice have increased plasma insulin and lower glucose. Exposure of isolated wild-type islets to elevated glucose levels reduced STR expression, whereas islets from diabetic (db/db) or diet-induced obese mouse models show similar down-regulation. This transcriptional reprogramming in response to hyperglycemia correlates with reduced STR function in these mouse models, leading to insulin hypersecretion. These findings reveal a novel mechanism by which insulin secretion is physiologically regulated by STRs and also suggest that, during the development of diabetes, STR function is compromised by hyperglycemia leading to hyperinsulinemia. These observations further suggest that STRs might be a promising therapeutic target to prevent and treat type 2 diabetes.

Progressive β-cell dysfunction in the setting of insulin resistance leads to chronic hyperglycemia during the development of type 2 diabetes, suggesting that the homeostatic mechanisms controlling β-cell function also become dysfunctional (1, 2). For instance, in patients with impaired fasting glucose (100–125 mg/dL (3)), basal insulin secretion (ie, fasted state) is often increased whereas dynamic responses to changes in glucose (ie, fed state) are blunted (4, 5) via mechanisms that are largely unknown. These individuals, although not meeting the criteria for diabetes, have higher risk for future development of diabetes (6). Notably, lifestyle and/or pharmacologic interventions have been shown to prevent or delay the development of diabetes in these populations (7). Thus, interventions that preserve β-cell function and reduce fasting hyperglycemia in these prediabetic groups may prevent or delay the onset of diabetes.

Insulin secretion by the β-cells is dependent on, and proportional to, plasma glucose levels and in particular, acute changes in glucose. Increasing evidence suggests that additional amplifying pathways exist, some of which potentiate insulin release by sensing nonglucose nutrients, like other monosaccharides, amino acids, and free fatty acids (8). Some of these nutrients are not metabolized but bind instead to cell-surface G protein-coupled receptor (GPCRs) to modulate insulin release. For instance, G protein-coupled receptor 40 (free fatty acid receptor 1) in β-cells is activated by medium- and long-chain fatty acids to stimulate insulin secretion (9). Ablation of the free fatty acid receptor 1 gene leads to abnormalities in glucose homeostasis (10, 11), suggesting that disruption of regulatory pathways involving nutrient sensing may contribute to the development of β-cell dysfunction. In agreement, we recently demonstrated that sweet taste receptors (STRs), another novel GPCR on mouse and human β-cells, sense ambient fructose to potentiate glucose-stimulated insulin secretion (GSIS) (12) independent of nutrient metabolism. Unlike fructose, which appears transiently in the circulation after a meal (13), glucose is constantly present, fluctuating within a tightly regulated range in the fasted as well as the fed state, and therefore is a bona fide ligand for STRs on β-cells (14). Therefore, it is conceivable that STRs on β-cells could function as general sweet-nutrient sensors to regulate insulin secretion independent of nutrient metabolism. Nevertheless, the physiological role of glucose sensing by STRs and its contribution to β-cell dysfunction are unknown.

Here we show that STRs on β-cells are important glucosensors for the regulation of basal insulin secretion. In the fasted state, during which most peripheral glucose uptake is noninsulin dependent, STR mildly attenuate insulin secretion over time. Mouse and human islets deprived of STR signaling (ie, T1R2^−/− or STR inhibition) hypersecrete basal insulin but maintain GSIS. Accordingly, 5-hour fasted T1R2^−/− mice have increased plasma insulin and lower plasma glucose. Notably, exposure of isolated wild-type (WT) islets to short-term high physiological glucose reduces STR expression, whereas islets from diabetic (db/db) or diet-induced obese mouse models show similar down-regulation. This transcriptional reprogramming in response to hyperglycemia correlates with reduced STR function in these mouse models, leading to insulin hypersecretion. This novel STR adaptation enables, in combination with other mechanisms, compensatory basal insulin hypersecretion to offset peripheral insulin resistance during diabetes development.

Materials and Methods

Animals and cell culture

All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC). Mice with a homozygous deletion for the t1r2 gene (kindly provided by Dr Charles Zuker, Columbia University) were back-crossed on the C57Bl\6J strain for 9 generations and genotyped in house for all experiments (15, 16). We used nonlittermate WT mice (C57Bl\6J) as controls. B6.BKS(D)-Lepr^db/J (db/db) male mice were purchased (The Jackson Laboratory). Cohorts of 8- to 10-week-old male mice were placed on a 12-week high-fat (58% fat, D12330i) or control diet (11% fat, D12328i; Research Diets). MIN6 β-cells (17) (kind gift of Dr Miyazaki) were cultured as previously described (12).

Mouse islet isolation and culture

Mouse islets were isolated from 8- to 10-week-old male mice, unless otherwise stated, and euthanized by CO₂ asphyxia and cervical dislocation. Islet isolation was performed as previously described (12). Subsequently, islets were hand picked and cultured in RPMI containing 11 mM glucose, unless otherwise stated, and l-glutamine supplemented with 15% fetal bovine serum and antibiotic-antimycotic and maintained under 5% CO₂. The medium was exchanged the following day, and the islets were allowed to recover for an additional day before experiments, unless otherwise stated.

Human islets were obtained from various islet isolation centers through the Integrated Islet Distribution Program and cultured for 3–5 days in RPMI containing 5 mM glucose and l-glutamine supplemented with 10% fetal bovine serum and antibiotic-antimycotic under 5% CO₂. No donors were diagnosed with diabetes. Insulin content was measured as described elsewhere (12).

Static insulin secretion

Basal or stimulated insulin secretion was measured in hand-picked cultured human and mouse islets incubated at 37°C in 100 μL Krebs-Ringer HEPES buffer (KRH: 119 mM NaCl, 4.74 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgCl₂, 1.2 mM KH₂PO₄, 10 mM HEPES, 0.5% insulin-free BSA, pH 7.4) in custom-made wells. For GSIS studies, mouse islets of equal size were equilibrated at 3.0 mM glucose for 30 minutes and were then transferred to increasing glucose concentrations for an additional 30 minutes. For basal insulin studies, mouse and human islets of equal size were incubated at constant/unaltered glucose concentrations for variable times (see figures). Islets were then transferred to new wells with matching glucose concentrations for an additional 30 minutes. For sustained/chronic basal insulin secretion studies, mouse islets were equilibrated at 3.0 mM glucose for 30 minutes in RPMI. Islets were then incubated at various glucose concentrations in RPMI for 3 hours, and insulin secretion was assessed for 30 minutes in KRH buffer. Islets were placed back at matching glucose concentrations in RPMI for 24 hours, and insulin secretion was assessed again for 30 minutes in KRH buffer. The supernatant from the final 30-minute incubations from all experiments was collected, and static insulin secretion (or glucagon) was measured using mouse or human insulin ELISA (Mercodia). Data show measured insulin from 4 separate wells with 10 islets each for each condition, averaged to represent insulin values for n = 1.

Plasma insulin, glucagon, and C-peptide measurements

Plasma insulin, glucagon, and C-peptide were measured using ultrasensitive mouse insulin ELISA (Mercodia), mouse glucagon ELISA (Mercodia), and mouse C-peptide ELISA (ALPCO Diagnostics).

Hyperglycemic clamp

Clamps were performed in 8- to 10-week-old male mice on regular chow (Teklad 2916). Catheters were surgically implanted into the left common carotid artery and right jugular vein as described (18), except that anesthesia was induced (2%) and maintained (1%–2%) with isoflurane. Blood samples were taken at t = −15 and −5 minutes prior to glucose infusion for assessment of basal glucose and C-peptide. At t = 0 minutes an infusion of glucose was begun to achieve hyperglycemia (∼250 mg/dL) and was maintained for 4 hours by adjusting the rate of glucose infusion accordingly.

Intraperitoneal glucose tolerance test (IPGTT)

An IPGTT was performed after a 16-hour fast on 8- to 10-week-old male mice on regular chow. Animals were injected ip with glucose (1 g/kg body weight). Blood glucose was sampled from the tail and analyzed with an AlphaTRAK blood glucose monitoring meter.

Quantitative real-time PCR

Total RNA from 200–300 mouse or human islets or MIN6 cells was isolated using Invitrogen PureLink RNA kit and was reverse transcribed to cDNA (1.0 μg) using the High capacity cDNA Reverse Transcription kit from Applied Biosystems. Quantitative PCRs were performed on an Eppendorf MasterCycler using iQ SYBR Green from Bio-Rad Laboratories using the protocol and primer sequences previously published (12).

Statistical analysis

Results are shown as means ± SEM. The level of significance was set at P < .05. Statistical tests and significance were calculated as shown in the figure legends.

Results

Ablation of T1R2^−/− sweet-taste receptor induces basal insulin hypersecretion during fasting glucose conditions in vivo and in isolated mouse and human islets

Sweet-taste perception is mediated by an obligate heterodimeric complex of 2 members of the T1R family (T1R2 and T1R3) of GPCRs (19, 20). Consequently, both T1R2 and T1R3 knockout mice (T1R2^−/− and T1R3^−/−) lack sweet-taste responses (16, 21). To investigate the potential role of glucose-mediated STR signaling on β-cells we used distinct experimental scenarios that simulate glucose regulation of insulin release during the following conditions: 1) fasted state (ie, basal/between meals) characterized by prolonged and constant levels of glucose, or 2) fed state (ie, dynamic/during a meal) characterized by acute increases in ambient glucose.

First, we investigated whether STRs regulate basal insulin release in response to constant glucose concentrations. We preincubated isolated islets at various glucose concentrations representative of different feeding states for 1.5 hours and then assessed insulin release at matching concentrations. Remarkably, insulin secretion was significantly higher in islets from T1R2^−/− mice compared with those from WT mice at 8.3 mM glucose (ie, short-term fasting values for C57BL/6 mice), but not at 3.0 mM (ie, low nonphysiological), 5.5 mM (ie, overnight fasting values) or 16.7 mM (ie, fed values) (Figure 1A). The differences in insulin secretion at 8.3 mM glucose cannot be explained by alterations in glucagon secretion (WT: 4.3 ± 0.3 vs T1R2: 4.3 ± 0.7 pg/mL/islet, P = .97; n = 6). This suggests that STRs regulate basal insulin secretion in vitro in the glucose range corresponding to short-term fasting concentrations. Lactisole is a human specific sweet-taste receptor inhibitor (22) that inhibits fructose-mediated potentiation of insulin release in human islets (12). Inhibition of STRs with lactisole induced basal insulin hypersecretion in human islets in the presence of 3.0 and 5.5 mM glucose (ie, fasting range in humans; Figure 1B), but not at 11.0 mM (ie, fed). This supports our mouse data and suggests that the regulation of basal insulin secretion by STRs is also relevant in human physiology.

Figure 1.

STRs regulate basal insulin secretion during fasting. A and B, Static insulin secretion in response to various glucose concentrations (G; mM) in mouse and human islets. Insulin secretion in mouse islets was assessed after 1.5 hours incubation at constant glucose as shown and expressed as percent of islet insulin content (quadruplicate batches of 10 islets per condition from n = 6 mice per group). Student's t test: ***, P < .001. Insulin secretion in human islets was assessed after 1.5 hours incubation at constant glucose with or without lactisole (1 mM) (quadruplicate batches of 10 islets per condition from n = 8 human preparations). Paired Student's t test: **, P < .01; ****, P < .0001. C, Plasma glucose (left panel) and insulin (right panel) in fed or 5-hour fasted WT and T1R2^−/− mice (n = 16–24). Student's t test: *, P < .05; **, P < .01. D, Insulin sensitivity index assessed by calculating the ratio of 5-hour fasted plasma insulin to glucose levels for each mouse. Student's t test: **, P < .01. E, Assessment of stimulated insulin secretion in WT and T1R2^−/− mice in vivo using hyperglycemic clamp. Plasma glucose was clamped at 250 mg/dL (circles; left axis) with appropriate adjustments on the glucose infusion rates (squares; right axis) for 4 hours (n = 8 per group). F, Plasma levels of C-peptide (pM) in response to the hyperglycemic clamp as a surrogate measurement of insulin secretion by β-cells. G and H, Blood glucose levels during an IPGTT and assessment of the insulinogenic index (increment in plasma insulin/increment in plasma glucose) during the first 30 minutes of the IPGTT in overnight-fasted age-matched WT and T1R2^−/− male mice (n = 10–14 per group) as described in Materials and Methods. I, Static insulin secretion in response to increasing concentrations of glucose (G; mM) in WT and T1R2^−/− islets. Islets were equilibrated at 3.0 mM glucose (G3.0) and then transferred to glucose concentrations as shown (G; mM) to assess GSIS expressed as percent of islet insulin content (quadruplicate batches of 10 islets per condition from n = 4 mice).

Next, we tested whether STRs regulate basal insulin secretion in vivo. Five-hour fasted T1R2^−/− mice had increased fasting plasma insulin accompanied by lower fasting plasma glucose levels (Figure 1C; fasted), but, similar to isolated islets, had no significant differences in plasma glucagon (WT: 35.2 ± 2.0 vs T1R2: 31.3 ± 1.7 pg/mL P = .17; n = 5). Consequently, the plasma insulin to glucose ratio (calculated for each mouse) was significantly higher (∼58%) in T1R2^−/− mice (Figure 1D; fasted) than in controls. This supports our finding in isolated islets and suggests that the absence of STR signaling in β-cells results in basal insulin hypersecretion at any given short-term fasting plasma glucose concentration. In contrast, under fed conditions these differences were absent (Figure 1, C and D; fed) indicating that STR-mediated regulation of insulin secretion operates only in the narrow range of fasting glucose concentrations.

To further confirm this hypothesis, we assessed insulin secretion in vivo under sustained stimulated glucose concentrations using a hyperglycemic clamp and monitored plasma C-peptide release. Mean plasma glucose, glucose infusion rates (Figure 1E), and C-peptide release (Figure 1F) were similar between WT and T1R2^−/− mice. Furthermore, overnight fasting plasma glucose, plasma glucose responses (Figure 1G), and the insulinogenic index (ΔI_{t = 0–30}/ΔG_{t = 0–30}) were similar during an IPGTT, excluding the possibility of glucose intolerance or β-cell dysfunction in T1R2^−/− mice (Figure 1H). Consistent with the in vivo data, insulin secretion was similar in response to acute increases in ambient glucose in isolated WT and T1R2^−/− islets (Figure 1I). Collectively, these results suggest that STRs are implicated in the regulation of insulin release particularly at ambient glucose concentrations corresponding to short-term fasting (ie, such as between meals).

STRs attenuate basal insulin secretion in response to continuous exposure to physiological glucose

It is well established that sustained exposure to glucose (>3 hours) induces a progressive decline in insulin secretion in isolated islets and perifused pancreata (23–26). This suggests the presence of a time-dependent mechanism that desensitizes β-cells (23). Considering that β-cells hypersecrete basal insulin in the absence of STRs, we reasoned that the desensitization mechanism might involve STRs. First, we tested whether continuous exposure to constant glucose would affect the rate of insulin secretion in WT islets. We incubated islets at 8.3 mM glucose and compared the rates of insulin release after 3 hours and then again at 24 hours for the same population of islets (25). Insulin secretion was reduced after 24 hours compared to 3 hours (Figure 2A), confirming that insulin secretion is attenuated over time in response to constant glucose levels. In contrast, islets exposed to high glucose concentrations (ie, 16.7 and 25.0 mM) maintained the rates of insulin secretion. In agreement with our hypothesis, islets lacking STR signaling (T1R2^−/−) not only hypersecrete basal insulin at 8.3 mM glucose (Figure 1A), but also fail to alter insulin secretion in response to prolonged stimulation (Figure 2B).

Figure 2.

STRs alter insulin secretion during sustained exposure to physiological glucose. A, Static insulin secretion in response to sustained exposure at constant glucose (G; mM) in WT islets. Insulin secretion was assessed after incubation for 3 hours and 24 hours at constant glucose (8.3, 16.7, and 25.0 mM; quadruplicate batches of 10 islets per condition from n = 6–9 mice). Paired Student's t test: **, P < .01. B, Static insulin secretion in response to sustained exposure at baseline glucose in WT and T1R2^−/− islets. Insulin secretion was assessed after incubation for 3 hours and 24 hours at constant glucose (8.3 mM) (quadruplicate batches of 10 islets per condition from n = 8–9 mice). Data are expressed as percent of islet insulin content using paired experiments. Paired Student's t test: *, P < .05.

STR expression is down-regulated in response to hyperglycemia and correlates with reduced STR-mediated function in islets of diabetic and diet-induced obese mouse models

Compensatory insulin hypersecretion occurs early in the progression to diabetes, suggesting that the desensitization mechanism in β-cells may be altered (27). Thus, we tested whether short-term hyperglycemia (24–48 hours) decreases STR function in β-cells. First, we exposed MIN6 β-cells and WT islets in low and high glucose for 48 hours and compared STR expression and function. Prolonged exposure to high glucose concentration in vitro decreased STR expression in MIN6 β-cells (Figure 3A) and in WT islets (Figure 3B), without affecting other components of the canonical STR signaling pathway, such as the cation channel trpm5 or the glucose transporter glut2 (G8.3: 0.37 ± 0.11 vs G11.0: 0.31 ± 0.06, P = .56; n = 10). Islets cultured in high physiological glucose (48 hours at G11.0) also demonstrated significantly higher basal insulin secretion, assessed at 8.3 mM glucose, than islets cultured in low glucose (48 hours at G8.3) (Figure 3C). The correlation between STR down-regulation and insulin hypersecretion is consistent with similar observations in islets lacking STRs (T1R2^−/−).

Figure 3.

STR expression and function is down-regulated in islets from hyperglycemic mouse models. A, STR signaling gene expression in MIN6 cells cultured in low (G5.0) or high (25.0) glucose (mM) for 24 hours measured by quantitative real-time RT-PCR. Relative arbitrary units (AU) shown normalized to 18S rRNA (n = 12). Student's t test: **, P < .01; ***, P < .001. B, STR signaling gene expression in isolated WT islets cultured in low (G8.3) or high (G11.0) glucose (mM) for 48 hours measured by quantitative real-time RT-PCR. Arbitrary units (AU) shown normalized to 18S rRNA (n = 10–12 mice). Student's t test: **, P < .01; ***, P < .001. C, Static insulin secretion in isolated WT islets cultured in low (G8.3) or high (G11.0) glucose (mM) for 48 hours. Insulin secretion was assessed after 3 hours incubation at constant glucose and expressed as percent of islet insulin content (8.3 mM; quadruplicate batches of 10 islets per condition from n = 6 mice). Student's t test: *, P < .05. D, STR gene expression (t1r2 and t1r3) in isolated islets from db/db and HFD mice compared with controls measured by quantitative real-time RT-PCR. Arbitrary units (AU) shown normalized to 18S rRNA (n = 6–8 mice). Student's t test: *, P < .05; **, P < .01; ***, P < .001. E, Assessment of fructose-induced insulin secretion as a STR functional assay. Static insulin release at 8.3 mM glucose (G) with or without 10.0 mM fructose (F) in islets from db/db (left panel) and HFD (right panel) mice compared with controls. Data are expressed as relative units (RU) of insulin fold change from 8.3 mM glucose alone (set at value 1) using paired experiments. (Quadruplicate batches of 10 islets from n = 5 mice per group). Paired Student's t test: *, P < .05; **, P < .01.

To confirm the physiological relevance of these observations, we investigated whether STR expression is also altered in islets of 2 known hyperglycemic mouse models: 1) the Lepr^−/− mouse model for diabetes (db/db) and 2) the diet-induced obesity model. As expected, fasting hyperglycemia was observed in 13-week-old db/db mice (wt: 148.8 ± 10.8 vs db/db: 220.0 ± 14.7 mg/dL, P = .0016; n = 8) and in WT mice on high-fat diet (HFD; 12 weeks) compared with controls on a chow diet (WT-CON: 184.2 ± 5.3 vs WT-HFD: 207.2 ± 5.7 mg/dL, P = .0074; n = 12). Consistent with the expression data in islets cultured in high physiological glucose, STRs were down-regulated in islets of both hyperglycemic models (Figure 3D). This transcriptional down-regulation was islet specific because intestinal expression of STRs was unaltered in HFD (t1r2, CON: 2.0 ± 0.06 × 10⁻⁴ vs HFD: 2.1 ± 0.8 × 10⁻⁴ P = .88; and t1r3, CON: 5.9 ± 1.5 × 10⁻⁴ vs HFD: 6.0 ± 0.84 × 10⁻⁴, P = .94; n = 8) and db/db (t1r2, WT: 3.6 ± 0.1 × 10⁻⁵ vs db/db: 3.1 ± 0.4 × 10⁻⁵ P = .68; and t1r3, WT: 3.6 ± 0.3 × 10⁻⁴ vs db/db: 2.6 ± 0.2 × 10⁻⁴, P = .03; n = 7) mice. It is conceivable that the down regulation of STRs might relate to a lack of leptin signaling in these models (28); therefore, we tested whether high ambient leptin can alter STR expression in β-cells. Islets incubated with leptin (20 ng/mL) for 48 hours had similar t1r2 (con: 7.52 ± 1.32 × 10⁻⁵ vs leptin: 6.37 ± 1.23 × 10⁻5; P = .54; n = 6) and t1r3 (con: 2.77 ± 0.53 × 10⁻³ vs leptin: 2.09 ± 0.14 × 10⁻³; P = .24; n = 6) expression relative to controls. These data effectively exclude a significant role for leptin in the regulation of STR expression in db/db or HFD mice.

We have shown that fructose can augment insulin release mediated by STRs on β-cells (12). We used fructose-induced insulin secretion as an assay of STR function to test whether reduced STR expression is adequate to impair STR-mediated insulin secretion in β-cells. In the presence of fructose, islets from WT control mice had an anticipated enhancement of insulin secretion (50%–60%). However, fructose-induced insulin secretion was blunted in islets from db/db or HFD mice (Figure 3E), suggesting that down-regulation of STR expression (Figure 3D) was adequate to compromise STR function.

Reduced STR-mediated function in islets from diabetic and diet-induced obese mice prevents glucose-induced desensitization and leads to compensatory insulin hypersecretion

It is well established that db/db (9–13 weeks old) and diet-induced obese mice (C57Bl\6J strain) have enhanced GSIS to compensate for peripheral insulin resistance (29). We hypothesized that STR-mediated function is central to compensatory insulin hypersecretion in β-cells. Similar to islets from T1R2^−/− mice, isolated islets from db/db and HFD mice hypersecrete insulin under short-term fasting glucose concentrations (8.3 mM) (Figure 4A) and fail to desensitize after continuous exposure to glucose compared with controls (Figure 4B). These data correlate with reduced STR function in these islets and support the involvement of STR regulation in compensatory insulin secretion.

Figure 4.

Reduced STR function in diabetic and obese mouse islets correlates with basal insulin hypersecretion and islet hypersensitization. A, Static insulin secretion in response to glucose (G; mM) in db/db (left panel; n = 6) and HFD (right panel; n = 8–10) islets compared with controls. Basal insulin secretion was assessed after incubation for 1 hour at constant glucose (8.3 mM or 16.7 mM) and expressed as percent of islet insulin content (quadruplicate batches of 10 islets per condition). Student's t test: *, P < .05; **, P < .01. B, Static insulin secretion in response to sustained exposure to glucose in islets from db/db (n = 5) and HFD mice (n = 9). Insulin secretion was assessed after 3 hours and 24 hours incubation at constant glucose (8.3 mM; quadruplicate batches of 10 islets per condition). Data are expressed as relative units (RU) of insulin fold change from 3-hour baseline value (set at value 1) using paired experiments. Paired Student's t test: *, P < .05.

Discussion

Our data implicate β-cell STRs in the regulation of basal insulin secretion. At typical short-term fasting glucose conditions, STRs sense ambient glucose and attenuate insulin secretion over time. In contrast, at elevated fasting glucose conditions, such as in experimental hyperglycemia or in islets from known hyperglycemic mouse models (ie, diabetic [db/db] or diet-induced obese), STR expression and function is reduced leading to basal insulin hypersecretion. This transcriptional reprogramming in response to hyperglycemia suggests a novel STR-mediated adaptation that enables compensatory basal insulin hypersecretion to offset peripheral insulin resistance during the development of diabetes.

Fasting glucose is determined by endogenous glucose production (primarily hepatic) regulated by the counterbalancing effects of insulin and glucagon and peripheral glucose uptake (which is mostly noninsulin dependent) (30). This homeostatic mechanism preserves constant glucose delivery to the brain and other glucose-dependent tissues during fasting (31). We found that pharmacologic inhibition of STRs in human islets or genetic ablation of STRs (T1R2^−/−) in mouse islets leads to hypersecretion of insulin at glucose concentrations corresponding to short-term fasting. These differences in basal insulin secretion are not due to altered insulin content, or the expression of genes involved in glucose metabolism (12). Similarly, 5-hour fasted T1R2^−/− mice have increased plasma insulin and decreased plasma glucose concentrations that cannot be explained by differences in plasma glucagon levels or altered glucagon secretion in islets. Thus, this effect is likely due to islet hypersensitization to glucose during immediate fasting (ie, increased fasted insulin/glucose). Nevertheless, because we used a global knockout for eliminating STR signaling, we cannot exclude the possibility of improved peripheral glucose disposal that could have confounded our findings. Interestingly, with fed conditions these differences in plasma glucose and insulin disappear, confirming that STR-mediated regulation of insulin secretion operates specifically around distinct short-term fasting glucose levels often seen between meals. Indeed, under dynamic increases in glucose concentrations, ablation of STRs had no effect on insulin secretion in vivo (plasma C-peptide during the hyperglycemic clamp and the insulinogenic index during the IPGTT) and in vitro (GSIS in isolated islets). This should not be surprising because, under these conditions, glucose uptake and metabolism by β-cells predominates. Therefore, a receptor-mediated glucose-sensing mechanism that modulates insulin secretion and operates in parallel to glucose uptake and metabolism would be redundant with fed conditions. We deliberately used an IPGTT to avoid effects of STR signaling in enteroendocrine cells of the intestine that are known to regulate incretin secretion and aspects of glucose absorption (32–34), which in turn, could affect insulin secretion and glucose homeostasis. Collectively, these data support a role of STRs in sensing ambient glucose to modulate basal insulin release specifically during short-term fasting glucose concentrations.

Several lines of evidence in perfused rodent islets and pancreata suggest that after 2–4 hours of continuous glucose stimulation the rate of insulin secretion is progressively reduced compared with initial, peak rates. This decline in the rate of insulin release is sustained and reaches its nadir within 24 hours, suggesting that β-cells gradually become desensitized when exposed to constant glucose for prolonged time (25, 26). We show that islets and mice lacking STR signaling are deprived of this mechanism and thus, maintain insulin secretion rates after prolonged exposure to constant glucose levels. Nevertheless, it is not immediately obvious why STRs sense plasma glucose specifically during fasting to decrease insulin secretion. It should be noted, however, that in the fasted state most glucose disposal proceeds independently of insulin (31). Therefore, a glucosensory mechanism that over time desensitizes islets to plasma glucose would moderately suppress insulin secretion without compromising peripheral glucose uptake. Although the signaling pathway mediating these effects is unclear and beyond the scope of this report, several studies suggest that sustained activation of phospholipase C (PLC) desensitizes islets (26, 35–37). Consistent with these observations, we have shown that STR signaling in β-cells requires the acute activation of PLC-IP₃ cascade which contributes to membrane depolarization and calcium influx (12). Nevertheless, mobilization of calcium stores or influx is transient, and thus it is unlikely to contribute to the desensitization mechanism, which extends over several hours. It is reasonable to speculate, therefore, that the desensitization mechanism may induce alternative PLC pathways beyond acute alterations in intracellular calcium, such as the documented activation of phosphokinase C in β-cells (36–38), or direct changes in vesicular fusion rates (39).

During the early stages of diabetes, insulin secretion is enhanced to compensate for insulin resistance (1). This increase can be explained by regulatory mechanisms in addition to an expansion of β-cell mass (2). In time, this compensatory insulin hypersecretion becomes inadequate, leading to prolonged postprandial hyperglycemia and to rising fasting plasma glucose levels (27). Although STR-mediated suppression of basal insulin secretion would be beneficial to regulate fasting euglycemia, during the pathological progression to fasting hyperglycemia this mechanism would be unfavorable. With this in mind, we investigated the functional role of STRs on β-cells during diabetes progression. db/db mice develop hyperglycemia and insulin resistance and have enhanced GSIS (29). We observed a notable islet-specific down-regulation of STR expression and function. Accordingly, islets from db/db mice not only hypersecrete insulin under basal conditions but also fail to desensitize in response to prolonged exposure to glucose. This phenotype is identical to T1R2^−/− islets, although additional independent factors may be at play as well (reviewed in Reference 2). The association between reduced STR expression and enhanced basal insulin secretion was further confirmed in islets from diet-induced obese WT mice. Nevertheless, other factors beyond hyperglycemia, such as underlying hormonal and metabolic dysregulation that accompanies the HFD, may have contributed to these effects. It is conceivable that impaired STR function in db/db or HFD islets enhances basal insulin secretion, which could be an adaptive response to fasting hyperglycemia. Therefore, this transcriptional adaptation is likely to precede or be regulated by hyperglycemia. Indeed, short-term exposure to high physiological glucose was sufficient to specifically down-regulate STRs in isolated WT islets. These islets also hypersecrete basal insulin, in agreement with previous reports in human islets cultured at elevated glucose (40) and our observations with islets from T1R2^−/−, db/db, and diet-induced obese mice. These findings also explain why WT islets exposed to 16.7 or 25 mM glucose did not reduce insulin secretion rates (3 hours vs 24 hours) compared with islets exposed to 8.3 mM glucose. We simulated fasting hyperglycemia and normoglycemia using 11.0 mM and 8.3 mM glucose in cultured media, respectively. It is noteworthy that mouse islets are habitually cultured in 11.0 mM glucose (ie, RPMI) for optimal GSIS, thus eliminating glucotoxicity as a plausible confounding factor (41). This is particularly significant because the transcriptional regulation of STR was achieved within a narrow physiological glucose range, suggesting high sensitivity to small changes in ambient glucose. Although glucose was sufficient to drive these effects with no involvement of leptin signaling, STR expression in β-cells may be further modulated in vivo by yet unknown circulating factors. Collectively, data from the 2 mouse models of diabetes independently confirm that blunted STR function and β-cell sensitization may be an early adaptive response (ie, compensation) to hyperglycemia.

In summary, we propose that STRs on β-cells are key monosaccharide sensors and play distinct roles in fine tuning insulin secretion in a dynamic fashion. Under basal conditions, such as between meals, nonglucose sweet nutrients are mostly absent. STRs sense plasma glucose to trigger a time-dependent, reversible signaling pathway leading to β-cell desensitization. This mechanism aims to moderately suppress basal insulin secretion during fasting, preventing hypoglycemia. Under stimulated conditions, such as during a meal, sweet nutrients (primarily fructose) and glucose rapidly rise in the circulation. In this environment, glucose and fructose can simultaneously occupy the active site on STRs (42) promoting cooperative effects, as previously shown in the tongue (43). This transient STR activation acutely augments, along with other auxiliary pathways, the effects of glucose metabolism on insulin secretion (ie, GSIS) (12). We also report, for the first time, that STR-mediated regulation of insulin secretion is reprogrammed during the early stages of diabetes contributing to compensatory basal insulin hypersecretion. Strikingly, even moderate increases in ambient glucose were sufficient to significantly reduce STR expression on β-cells, thus, enhancing basal insulin secretion. These findings suggest an early adaptive mechanism in response to emerging fasting hyperglycemia with promising therapeutic applications that target STR function on β-cells (44).