Subthreshold α 2-Adrenergic Activation Counteracts Glucagon-Like Peptide-1 Potentiation of Glucose-Stimulated Insulin Secretion

The pancreatic β cell harbors α 2-adrenergic and glucagon-like peptide-1 (GLP-1) receptors on its plasma membrane to sense the corresponding ligands adrenaline/noradrenaline and GLP-1 to govern glucose-stimulated insulin secretion. However, it is not known whether these two signaling systems interact to gain the adequate and timely control of insulin release in response to glucose. The present work shows that the α 2-adrenergic agonist clonidine concentration-dependently depresses glucose-stimulated insulin secretion from INS-1 cells. On the contrary, GLP-1 concentration-dependently potentiates insulin secretory response to glucose. Importantly, the present work reveals that subthreshold α 2-adrenergic activation with clonidine counteracts GLP-1 potentiation of glucose-induced insulin secretion. This counteractory process relies on pertussis toxin- (PTX-) sensitive Gi proteins since it no longer occurs following PTX-mediated inactivation of Gi proteins. The counteraction of GLP-1 potentiation of glucose-stimulated insulin secretion by subthreshold α 2-adrenergic activation is likely to serve as a molecular mechanism for the delicate regulation of insulin release.


Introduction
Glucose-stimulated insulin secretion plays an irreplaceable role in the control of glucose homeostasis since insulin is the only hormone capable of lowering blood glucose in the body [1][2][3]. This pancreatic endocrine hormone is packed in β cell secretory granules. These granules undergo exocytosis to release their insulin cargo into the bloodstream in response to elevated blood glucose levels [1][2][3]. Upon elevation of the plasma glucose level, the β cell efficiently takes up glucose through glucose transporters. Thereafter, subsequent glucose metabolism drastically raises the intracellular ATP level. The resultant rise in the ATP/ADP ratio closes ATP-sensitive K + (K ATP ) channels, causing depolarization of the plasma membrane. The membrane depolarization in turn opens voltage-gated Ca 2+ (Ca V ) channels, mediating Ca 2+ influx. The consequent increase in cytosolic-free Ca 2+ concentration ([Ca 2+ ] i ) triggers direct interactions between exocytotic proteins situated in the insulin-containing granule membrane and those localized in the plasma membrane. Eventually, the interaction between exocytotic proteins initiates the fusion of insulin-containing granules with the plasma membrane, that is, insulin exocytosis [1][2][3].
On top of the aforementioned consensus paradigm, glucose-stimulated insulin secretion is, in fact, regulated by complex neural mechanisms [4,5]. It is well known that the autonomic nervous system innervates pancreatic islet cells where parasympathetic endings release a bunch of substances, for example, acetylcholine and vasoactive intestinal polypeptide, to potentiate glucose-stimulated insulin secretion [5,6]. On the contrary, sympathetic terminals exocytose adrenergic and peptidergic transmitters to inhibit the insulin secretory process [4,5]. Treatment with the main sympathetic transmitter noradrenaline fully shuts down insulin secretion from either islets or β cell aggregates perifused with high glucose [7,8]. Mechanistically noradrenaline acts on α 2 receptors coupled to pertussis toxin-(PTX-) sensitive Gi proteins in β cells, reducing glucose-stimulated 2 Experimental Diabetes Research insulin secretion through inhibition of intracellular cAMP formation, Ca V channels, glucose metabolism, and the exocytotic machinery as well as elevation of K ATP channel activity [4,5].
Although either noradrenergic or GLP-1 signaling system in the regulation of glucose-stimulated insulin secretion has been clarified, it is not known whether these two signaling systems interact to gain adequate and timely insulin release in response to glucose stimulation [4,5,[9][10][11][12][20][21][22]. In the present work, we describe that subthreshold α 2 -adrenergic activation counteracts glucagon-like peptide-1 potentiation of glucose-stimulated insulin secretion in a PTX-sensitive Gi protein-dependent manner.

Static Insulin Secretion.
Approximately 70% confluent INS-1 cells in 24-well plates were used for insulin secretion experiments. The cells were kept at 37 • C in a humidified 5% CO 2 incubator during the course of an experiment except when their bath solutions need to be changed. The experiments were carried out in Krebs-Ringer bicarbonate HEPES buffer (KRBH) consisting of (in mM) 140NaCl, 3.6KCl, 1.5CaCl 2 , 0.5MgSO 4 , 0.5NaH 2 PO 4 , 2NaHCO 3 , 10HEPES, 0.1% bovine serum albumin (BSA), pH 7.4. First, the cells were rinsed with glucose-free KRBH and then maintained in the same buffer for 2 h. Thereafter, they were rinsed and preincubated with glucose-free KRBH for 30 min. To characterize the concentration-response relationships of the α 2 -adrenergic agonist clonidine and the incretin GLP-1 as well as interactions between noradrenergic and GLP-1 signaling systems, the cells were rinsed and incubated with 3 or 11 mM glucose KRBH containing different concentrations of clonidine and/or GLP-1 for 30 min. Clonidine and/or GLP-1 were applied simultaneously with glucose. To determine a possible dependence of α 2 -adrenoceptor regulation of GLP-1 receptors on PTX-sensitive Gi proteins, the cells were pretreated with 100 ng/ml PTX in RPMI 1640 medium for 18 h. Subsequently, the toxin medium was removed. The cells were rinsed and incubated with glucose-free KRBH as described above and subjected to incubations with 3 or 11 mM glucose KRBH containing different concentrations of clonidine and/or GLP-1 for 30 min. Finally, the treatments with the different reagents were stopped by putting the culture plates on ice. Supernatants were carefully aspirated from each well to prepare samples for insulin quantification. The samples were centrifuged at 1000 × g for 3 min to remove detached cells and stored at −20 • C until insulin immunoassay was performed.

Radioimmunoassay.
A standard insulin immunoassay was used to evaluate static insulin secretion from INS-1 cells subjected to different treatments [23,24]. Briefly, duplicate samples were measured. The calibration curve was constructed from insulin standard at 5, 10, 20, 40, 80, and 160 mIU/L. Radioactivity was counted by a γ-counter.

Statistical Analysis.
All data are presented as mean ± SEM. Statistical significance was evaluated by one-way ANOVA, followed by least significant difference (LSD) test. The significance level was determined at both the 0.05 and 0.01 levels.

Clonidine Concentration-Dependently Inhibits Glucose-Stimulated Insulin Secretion.
To determine the concentration response relationship of clonidine inhibition of glucosestimulated insulin secretion, we examined the effect of 30 min incubation with clonidine at concentrations ranging from 0.003 to 10 μM on insulin release from INS-1 cells challenged with 11 mM glucose. As shown in Figure 1, incubation with 11 mM glucose for 30 min resulted in a significant insulin secretion as compared with that with 3 mM glucose (n = 6, P < .01). This confirms that the cells used in this set of experiments reliably responded to such stimulation to secrete an appreciable amount of insulin. We therefore adopted this sufficient and reliable stimulation to test for the effect of clonidine on glucose-stimulated insulin secretion. Figure 1 shows that in the concentration range of 0.003-10 μM, clonidine concentration-dependently inhibited insulin release from INS-1 cells exposed to 11 mM  The α 2 -adrenergic agonist clonidine concentrationdependently depresses glucose-stimulated insulin secretion from INS-1 cells. Static insulin secretion was performed with cells subjected to stepwise elevation of glucose concentration from 3 to 11 mM for 30 min in the absence or presence of clonidine and determined by a standard insulin radioimmunoassay. Cells exposed to 11 mM glucose (closed circle at the far left) released significantly more insulin than those to 3 mM glucose (open circle) (n = 6, P < .01). In the concentration range of 0.003-10 μM, clonidine produced a concentration-dependent inhibition of insulin release induced by 11 mM glucose. The inhibition became statistically significant at 0.01 μM clonidine (n = 6, P < .05) and was statistically significant at higher clonidine concentrations (n = 6, P < .01). The subthreshold and ED 50 concentration of clonidine were calculated to be 0.003 and 4 μM, respectively. In this and all other figures, data are presented as means ± SEM. Statistical significance was evaluated by one-way ANOVA, followed by least significant difference (LSD) test. * P < .05 and * * P < .01 versus 11 mM glucose-treated group.

Glucagon-Like Peptide-1 Concentration-Dependently Stimulates Glucose-Stimulated Insulin
Secretion. To reveal the concentration-response relationship of GLP-1 potentiation of glucose-stimulated insulin secretion, we evaluated the insulin secretory response of INS-1 cells stimulated with 11 mM glucose for 30 min in the presence of GLP-1 in the concentration range 0.0001 to 1000 nM. Figure 2 shows that 11 mM glucose treatment for 30 min produced a significant increase in insulin secretion in comparison with 3 mM glucose treatment (n = 6, P < .01). This validates that the glucose responsiveness of the cells employed in this set of experiments. As illustrated in
Validation of the capacity of the cells applied in this set of experiments to release insulin in response to glucose was likewise performed. As illustrated in Figure 3, treatment with 11 mM glucose for 30 min gave rise to a significant insulin release as compared with that with 3 mM glucose (n = 10, P < .01). As expected, cells exposed to clonidine at the subthreshold concentration 3 nM did not alter their insulin secretory response to 11 mM glucose (n = 10, P > .05 versus group subjected to only 11 mM glucose stimulation) ( Figure 3). In contrast, cells treated with GLP-1 at the ED 50 concentration 0.1 nM following 11 mM glucose stimulation released significantly more insulin than cells subjected to only 11 mM glucose stimulation (n = 10, P < .05). Importantly, cells incubated with the ED 50 concentration of GLP-1 plus the subthreshold concentration of clonidine secreted significantly less insulin than cells treated with the ED 50 concentration of GLP-1 alone following 11 mM glucose stimulation (n = 10, P < .01) (Figure 3). The insulin secretory response to 11 mM glucose was very similar among group treated with the ED 50 concentration of GLP-1 plus the subthreshold concentration of clonidine, group treated with the subthreshold concentration of clonidine alone, and untreated group (n = 10, P > .05) (Figure 3). The data demonstrate that the subthreshold concentration of clonidine completely counteracted the potentiation of glucosestimulated insulin secretion by the ED 50 concentration of GLP-1.

Counteraction of Glucagon-Like Peptide-1 Potentiation of Glucose-Stimulated Insulin Secretion by Clonidine Relies on
Pertussis Toxin-Sensitive Gi Proteins. Multiple intracellular signaling events, such as decreases in cAMP production, Ca V channel activity, glucose metabolism, and exocytotic capacity as well as an increase in K ATP conductance occur upon activation of α 2 -adrenergic receptors on the β cell to depress glucose-stimulated insulin secretion [4,5]. All these events are dependent on the PTX-sensitive Gi protein that is an immediate mediator for α 2 -adrenergic activation [4,5]. This made us wonder if counteraction of GLP-1 potentiation of glucose-stimulated insulin secretion by clonidine relies on PTX-sensitive Gi proteins. To circumvent this issue, we evaluated if PTX-mediated inactivation of Gi proteins could prevent counteraction of GLP-1 potentiation of glucosestimulated insulin secretion by subthreshold α 2 -adrenergic activation.
The cells used in this set of experiments were proved to be quite sensitive to glucose with regard to their insulin secretory responsiveness. Figure 4 shows that both control and PTX-pretreated cells released a significant amount of insulin when glucose concentration was raised from 3 to 11 mM (n = 11, P < .01). Cells pretreated with 100 ng/ml PTX in for 18 h secreted significantly more insulin than cells without PTX pretreatment following 11 mM glucose stimulation (n = 11, P < .01) (Figure 4). Clonidine at both the subthreshold concentration 3 nM and the ED 50 concentration 4 μM had no effect on insulin secretory response . PTXpretreated cells released significantly more insulin than control cells following 11 mM glucose stimulation (n = 11, P < .01). Neither the subthreshold concentration (3 nM) nor the ED 50 concentration (4 μM) of clonidine altered glucose-stimulated insulin secretion in PTX-pretreated cells (n = 11, P > .05 versus PTX-pretreated group subjected to only 11 mM glucose stimulation). In contrast, GLP-1 at the ED 50 concentration of 0.1 nM induced a significant potentiation of glucose-induced insulin release from PTX-pretreated cells (n = 11, P < .01 versus PTX-pretreated group subjected to only 11 mM glucose stimulation). Intriguingly, glucose-stimulated insulin secretion from PTX-pretreated cells incubated with the ED 50 concentration of GLP-1 alone did not significantly differ from that from those subjected to incubation with the ED 50 concentration of GLP-1 plus the subthreshold concentration of clonidine (n = 11, P > .05). Clonidine at the subthreshold concentration could no longer influence potentiation of glucose-stimulated insulin secretion by the ED 50 concentration of GLP-1 in PTX-pretreated cells. * * P < .01 versus 3 mM glucose-treated group without PTX pretreatment, ++ P < .01 versus 11 mM glucose-treated group without PTX pretreatment or PTX-pretreated group subjected to 3 mM glucose incubation, ## P < .01 versus PTX-pretreated group subjected to 11 mM glucose stimulation.
Most importantly, the present study shows for the first time that insulin-secreting INS-1 cells exposed to the ED 50 concentration of GLP-1 together with the subthreshold concentration of clonidine release significantly less insulin than cells treated with the ED 50 concentration of GLP-1 alone following glucose stimulation. Furthermore, it also uncovers that the antagonistic interaction of the α 2 -adrenergic signaling system with the GLP-1 signaling system critically depends on PTX-sensitive Gi proteins. These findings provide evidence that α 2 -adrenergic or GLP-1 signaling system does not operate independently, but instead the former effectively antagonizes the latter to enable the pancreatic β cell to appropriately execute its unique function glucose-stimulated insulin secretion. In fact, interactions between G proteincoupled receptor signaling pathways have been intensively investigated in other cell types and neurons in particular [27][28][29][30][31][32][33]. Such interactions rely on multilevel mechanisms [27][28][29][30][31][32][33]. They occur at the receptor level due to receptor heterodimerization, which is either G protein-dependent or -independent [27][28][29][30]. The heterodimerization is able to alter the ligand binding affinity and/or signal transduction efficacy of dimerized receptors [27-30, 32, 33]. Interactions between G protein-coupled receptor signaling pathways can also bypass the receptor level and come about downstream of receptors as a result of crosstalk between receptor signaling cascades [31]. In general, these well-characterized mechanisms are applicable to the counteraction of GLP-1 potentiation of glucose-stimulated insulin secretion by subthreshold α 2 -adrenergic activation in the pancreatic β cell. The in-depth mechanisms whereby the α 2 -adrenergic signaling system antagonizes the GLP-1 signaling system in the pancreatic β cell remain to be characterized.
There is no doubt that the healthy body requires the efficient amount of insulin to remove extra glucose from the blood stream into body cells most of the time. However, the healthy body needs less insulin to boost blood glucose levels in some circumstances, such as stress, exercise, low blood glucose, and other environmental challenges. The counteraction of GLP-1 potentiation of glucose-stimulated insulin secretion by subthreshold α 2 -adrenergic activation definitely fits in with these circumstances where sympathetic activity is elevated [34]. It adds a new level of complexity to the classical paradigm for the regulation of glucose-evoked insulin release. Under certain pathological conditions, for example, diabetes, hypertension, obesity, and aging, sympathetic activity and/or expression of α 2 -adrenergic receptors in the β cell significantly increase [35][36][37][38][39]. Increases in sympathetic activity and/or expression of α 2 -adrenergic receptors in the β cell likely exaggerate the antagonistic 6 Experimental Diabetes Research interaction of the α 2 -adrenergic signaling system with the GLP-1 signaling system in the pancreatic β cell to provoke and aggravate diabetes [35][36][37][38][39].

Conclusions
α 2 -Adrenergic receptors and GLP-1 receptors on insulinsecreting INS-1 cells transduce signals from their corresponding ligands clonidine and GLP-1 to govern glucoseinduced insulin release. Importantly, the former also interacts with the latter to brake potentiation of glucoseinduced insulin release by the latter. In fact, subthreshold α 2 -adrenergic activation is enough to counteract GLP-1 potentiation of glucose-induced insulin secretion in a PTXsensitive Gi protein-dependent fashion. Such a counteraction is able to serve as a molecular mechanism for the delicate control of insulin release in the healthy body. Most likely, this counteractory process is exaggerated to provoke and aggravate diabetes since obesity, aging, and diabetes are highly associated with elevated sympathetic activity [35][36][37][38][39].