A role for islet somatostatin in mediating sympathetic regulation of glucagon secretion

Abstract

Aims/hypothesis

Somatostatin (SST) released from islet ä-cells inhibits both insulin and glucagon secretion but the role of this tonic inhibition is unclear. In this study we investigated whether δ-cell SST may facilitate sympathetic regulation of glucagon secretion as part of an ‘accelerator/brake’ mechanism.

Results

Arginine stimulated both glucagon and SST release from control mouse islets whereas the sympathetic neurotransmitter noradrenaline (NA) increased glucagon secretion but inhibited SST release in the presence of 2 mmol/l glucose or 20 mmol/l arginine. Experiments were performed using SST-deficient (Sst^-/-) islets to assess whether the reduction of SST secretion by NA offers an indirect mechanism of enhancing glucagon release in response to sympathetic activation. Arginine-induced but not NA-induced glucagon release from Sst^-/- islets was significantly increased compared to controls. In combination, NA enhanced arginine-induced release from both groups of mouse islets but to a greater extent in control islets, leading to similar overall levels of glucagon release. The responsiveness of Sst^-/- islets to NA was thus blunted under stimulatory but not sub-stimulatory conditions of SST release.

Methods

The secretory characteristics of islets isolated from Sst^-/- and control mice were assessed in static incubation studies. Glucagon and SST release was measured by radioimmunoassay (RIA).

Conclusions

Our data suggest that sympathetic activation of glucagon release may be partly mediated by an indirect effect on SST secretion, where the tonic inhibition by δ-cell SST on α-cells is removed, facilitating precise and substantial changes in glucagon release in response to NA.

Introduction

The islet of Langerhans is a complex structure containing α-, δ- and PP cells in addition to insulin-secreting β-cells. The intraislet environment influences β-cell function and the importance of homotypic interactions between β-cells, and of heterotypic interactions between α- and β-cells, has been studied in detail.^1–4 Much less attention has been paid to the intra-islet roles of δ-cell-derived somatostatin (SST) although changes in δ-cell function have been associated with diabetes (reviewed in ref. 5).

We have recently used islets from SST-deficient mice to demonstrate that endogenous δ-cell SST exerts a tonic inhibition on hormone secretion from both α- and β-cells.⁶ Our data suggested an important intra-islet role for δ-cell SST in regulating the secretion of insulin and glucagon in response to external signals and led us to propose an “accelerator/brake” model where the inhibitory δ-cell input acts as a “brake” on secretory responses to physiological “accelerators” of insulin secretion. The islet is therefore able to generate precise and amplified secretory responses to certain stimuli through a combination of directly stimulating the β-cells (“accelerator”) while releasing the endogenous δ-cell inhibitor input (“brake”) by suppressing SST secretion. The amplification of glucose-induced insulin secretion by cholinergic agonists offers an example of this proposed mechanism.⁶

The tonic inhibitory δ-cell input also applies to α-cells⁶ so we have now addressed whether an “accelerator/brake” mechanism is also implicated in the sympathetic regulation of glucagon secretion from α-cells.

Results

Glucagon and SST release from control mouse islets was measured in response to the adrenergic neurotransmitter noradrenaline (NA, Fig. 1). Ten µmol/l NA stimulated glucagon release 2–3 fold in the presence of 2 mmol/l glucose or 20 mmol/l arginine (Fig. 1A), whereas SST secretion was inhibited by 50–60% under the same conditions (Fig. 1B).

Figure 1.

Effect of 10 µmol/l noradrenaline (NA) on glucagon (A) and SST (B) secretion from isolated mouse islets in the presence of 2 mmol/l glucose or 20 mmol/l arginine. Bars show means ± SEM of four separate experiments, each of 6–8 observations per treatment group. *p < 0.05 vs. 2 mmol/l glucose or 20 mmol/l arginine, respectively, by Student's t-test.

The combined effects of NA to stimulate glucagon secretion whilst inhibiting SST release may enhance the direct sympathetic activation of α-cells by lifting the tonic inhibitory effect of δ-cells on glucagon release (the accelerator/brake hypothesis). To test this hypothesis, similar experiments were performed using Sst^-/- islets, as shown in Figure 2. NA significantly stimulated glucagon secretion from both control and Sst^-/- islets at 2 mmol/l glucose (p < 0.01) with no difference in the amplitude of response between the two genotypes. Arginine also stimulated glucagon secretion (p < 0.001), but the response was significantly enhanced in Sst^-/- islets as previously reported.⁶ However, despite the differences in arginine-induced glucagon release from the two genotypes, there was no significant difference in responses when arginine and NA were combined. Thus, whereas NA further stimulated arginine-induced release from both islet groups, the effect was much more pronounced in control islets (∼160%) than in Sst^-/- islets (∼70%, Fig. 2B). This effect was observed both when results were expressed as the rate of glucagon secretion (pg glucagon/islet/h) in individual experiments (e.g. Fig. 2A) and when expressed relative to secretion at 2 mmol/l glucose in four separate experiments (Fig. 2B).

Figure 2.

Effect of 10 µmol/l noradrenaline (NA) on glucagon secretion measured from control (open bars) and Sst^-/- mouse islets (solid bars) in the presence of 2 mmol/l glucose or 20 mmol/l arginine. (A) Data from one representative experiment expressed as pg/islet/h, n = 6–8 observations per treatment group. (B) Data from four separate experiments expressed relative to secretion at 2 mmol/l glucose. NS = p > 0.2, **p < 0.01, control vs. Sst^-/- mouse islets by two way ANOVA and Bonferroni's multiple comparisons test.

Discussion

Our previous studies using the acetylcholine analogue CCh suggested that δ-cell SST regulates insulin secretion by acting as a physiological brake which is released by certain secretagogues concomitantly with their direct, stimulatory (accelerator) action on the β-cell. This dual action facilitates a more pronounced and precise regulation of insulin secretion in response to a transiently effective signal such as a parasympathetic neurotransmitter.⁶ The present results indicate that this mechanism may also apply to the sympathetic regulation of glucagon secretion. Arginine and (nor)adrenaline are both well-known stimulators of glucagon release,^7,8 but the two stimuli differ in their action on islet δ-cells. Our data show that arginine stimulates SST release, whereas NA inhibits both basal and stimulated SST secretion, consistent with previous findings.^7,9 The dual action of NA on glucagon and SST secretion suggests the expression of adrenergic receptors on both α- and δ-cells: the expression of α- and β-adrenergic receptors on α-cells is well established,^7,10 but to our knowledge adrenergic receptors have not yet been identified on δ-cells although there is evidence that adrenaline inhibits Ca²⁺ signaling in δ-cells.¹¹

As we have previously reported, arginine-induced glucagon secretion is enhanced from Sst^-/- islets compared to controls.⁶ These findings are consistent with a direct stimulatory action of arginine on δ-cells and suggest that SST released in response to arginine inhibits glucagon release from control islets but not from SST-deficient islets.

Our studies with Sst^-/- mouse islets confirm a role for SST in facilitating the α-cell response to NA. Thus, in SST-deficient islets the glucagon secretory response to a combination of arginine and NA was similar to that of control islets, despite their enhanced response to arginine caused by the absence of an internal SST ‘brake’. The similarity in response is consistent with NA acting on both α- and δ-cells, thus stimulating glucagon secretion directly and to the same extent in both islet groups (as shown in Fig. 2) but only lifting the inhibition by SST in control islets because in Sst^-/- islets this ‘brake’ is absent. Arginine-induced glucagon release was enhanced by NA in both control and Sst^-/- islets further suggesting that the effect of NA on glucagon secretion is not solely dependent on a direct action on δ-cells to release the SST ‘brake’, but is rather a combination of a direct action on α-cells and a release of the δ-cell SST ‘brake’. Conversely, if NA influenced glucagon secretion solely through activation of adrenergic receptors on α-cells the Sst^-/- islets would be predicted to secrete more glucagon than control islets as a consequence of the maintained tonic inhibitory δ-cell input in control islets. This was not observed in our experiments, suggesting that the rate of NA-stimulated glucagon secretion from normal islets is dependent on the balance between direct stimulatory and indirect inhibitory inputs via the α- and δ-cells respectively. Thus, while the maximum rate of glucagon secretion in response to arginine and NA was similar in control and Sst^-/- islets, the overall responsiveness of the islets to sympathetic input was blunted in the absence of SST, suggesting that Sst^-/- lack a level of sympathetic control present in normal islets, consistent with an important role for the δ-cell in regulating islet secretory responses to neurotransmitters.

Glucose is reported to stimulate SST release with threshold levels of 3–4 mmol/l.¹² Consistent with this, no significant difference in glucagon release was observed between genotypes at sub-stimulatory glucose concentrations (2 mmol/l). Our results did, however, suggest that even under those circumstances the low levels of SST released from δ-cells were inhibited by NA. However, this effect on basal secretion of SST was not sufficient to enhance NA-induced glucagon secretion in controls compared to Sst^-/- islets suggesting that under these circumstances NA acts predominantly via direct stimulatory effects on the α-cells rather than indirectly via inhibition of δ-cells.

In both β- and α-cells the “accelerator/brake” mechanism may enable relatively transient stimulatory inputs, such as via parasympathetic or sympathetic innervation, to induce precise and rapid changes in the overall amount of insulin or glucagon secreted, respectively. Such a mechanism may not be required to maintain the longer-lived secretory responses to elevated concentrations of nutrient stimuli, which persist for much longer periods in the absorptive, post-prandial state. In accordance with this, we and others have demonstrated that glucose stimulates both insulin and SST release, and depolarising insulin secretagogue stimuli such as sulphonylureas also induce SST release.^6,12,13 It has been suggested that the tonic inhibitory input from the δ-cells in response to nutrient stimuli may prevent wasteful co-secretion of insulin and glucagon.¹⁴

In conclusion, our studies using Sst^-/- islets suggest that paracrine interactions between δ- and α-cells provide an intra-islet, SST-mediated control mechanism which facilitates and fine-tunes the release of glucagon in response to sympathetic activation.

Methods

Animals.

Sst^-/- mice on a CBA/Ca × C57BL/10 F1 background were generated as described.⁶ Age-matched CBA/Ca × C57BL/10 F1 mice served as controls.

Islet isolation.

Islets from 8–12 week old female Sst^-/- or control mice were isolated by collagenase digestion (1 mg/ml, type XI, Sigma, UK) and separated from exocrine pancreatic tissue on a histopaque gradient (Sigma, UK), as described.⁶ Islets were incubated overnight at 37°C (5% CO₂) in RPMI 1640 medium (10% FBS, 2% glutamine, 100 U/ml penicillin/0.1 mg/ml streptomycin, 11 mmol/l glucose) prior to experiments.

Hormone secretion studies.

Glucagon and SST release was assessed in static incubation experiments. Islets were pre-incubated for 60 min in a bicarbonate-buffered physiological salt solution⁶ containing 2 mmol/l glucose after which batches of twelve islets were incubated for 60 min in 0.4 ml salt solution containing agents of interest. Hormone content of incubation medium was assessed by radioimmunoassay (RIA) using an inhouse glucagon assay, as described⁶ and a commercially-available SST RIA kit (Euro-Diagnostica, Sweden).

Data analysis.

Data are expressed as means ± SEM and analyzed using Student's t test, two way ANOVA and Bonferroni's multiple comparisons test, as appropriate. Differences between treatments were considered significant at p < 0.05.