IV. Control of β-cell function by the brain

Autonomic activation, particularly sympathetic neural activation, contributes to the glucagon response to hypoglycemia. The autonomic nervous system accounts for 75 to 90% of the glucagon response to marked hypoglycemia and there is sequential recruitment of each arm of the autonomic nervous system as hypoglycemia deepens. In addition, sensory inputs involved in the cephalic phase of insulin secretion also induce the parasympathetic system, thus inducing anticipatory responses that prepare the body for efficient utilization of glucose.

1. Early loss of islet sympathetic nerves

The glucagon response to insulin-induced hypoglycemia is severely impaired in type 1 diabetic patients. Data show that this defect is not due to impairment of the adrenal medullary or parasympathetic branch. However, type 1 diabetic patients exhibit a 93% loss of islet sympathetic nerves. This early sympathetic islet neuropathy is not due to diabetic neuropathy induced by hyperglycemia since it develops before diabetes in animal models of type 1 diabetes. Type 1 diabetes is characterized by invasive insulitis and by a progressive loss of β-cells. The release of sympathetic neurotrophins such as NGF by β-cells could regulate sympathetic innervation. However, specific experiments demonstrate that early sympathetic islet neuropathy is neither linked to decreased secretion of nerve growth factorby β-cells in response to islet destruction, nor to lymphatic infiltration of islets. Sympathetic neurons express the neurotrophin receptor p75 (p75NTR),and agonists of p75NTR induce a rapid sympathetic denervation due to localized pruning of axonal segments, which leaves both the parent axon and the neuronal cell body intact. Inactivation of p75NTR in a type 1 diabetes mouse model prevents early sympathetic islet neuropathy, but does not affect lymphocyte infiltration, or the onset and magnitude of diabetic hyperglycemia. An increase in BDNF within the infiltrated islet likely activates the p75NTR causing the loss of islet sympathetic nerves.
However, BDNF-releasing cell types within the islet and the stimulus triggering this release have yet to be identified (Gerald Taborsky, Lecture).

2. The autonomic nervous system and pulsatile secretion of insulin

Pulsatile secretion of pancreatic hormones is physiologically important and compromised in diabetes. This secretory pattern is controlled by neurotransmitters released from islet cells, which shape the pulse in auto-paracrine feedback loops. Blood insulin and glucagon levels oscillate in opposite phase, independently of central neural influences, and the loss of this phase relationship has been associated with prediabetes. Glucose, neurotransmitters, and signaling molecules produced by the islet control pulsatile insulin release. In β-cells, Ca2+ oscillations induced by glucose metabolism might depend on oscillatory activity of rate-limiting glycolytic enzymes and/or mitochondrial metabolism. On one hand, an increase in ATP production drives influx of Ca2+, which temporarily interrupts the increase of ATP, while on the other, processes extruding Ca2+ from the cytoplasm increase energy consumption. Thus, mutual ATPCa2+ interaction may generate intracellular calcium oscillation underlying pulsatile insulin secretion. In addition to this internal β-cell oscillator, autocrine feedback mechanisms are important for shaping the pulses. Positive feedback loops may help to initiate secretory pulses whereas negative feedback loops could be involved in their termination. The neurotransmitters and signaling molecules, which are released by β-cells in parallel to insulin, participate in the feedback regulation of intracellular calcium oscillations.
Insulin controls Ca2+ levels via binding to its receptor. Negative feedback regulation by insulin may involve autocrine signaling: PI3K-dependent formation of phosphatidylinositol(3,4,5)-triphosphate in the plasma membrane is able to activate KATP channels, which then leads to reduction in intracellular Ca2+ levels. ATP stimulates insulin release by its interaction with ionotropic P2X receptors, whose cation permeability adds to voltage-dependent Ca2+ influx. ATP also interacts with G-protein−coupled receptor P2Y, which leads to increased intracellular Ca2+ levels via the activation of different phospholipases. In mice, ATP also has an inhibitory effect on insulin secretion via the phosphatase calcineurin. Pancreatic islets contain high concentrations of GABA released in parallel with insulin. GABA has a hypothetical role in pulsatile insulin release by promoting the early depolarization and/or the termination of insulin pulses. Zn2+ co-released with insulin might exert negative feedback on β-cells by inducing a decrease in cAMP formation. A loss-of-function mutation in the insulin granule zinc transporter ZnT8 protects against the development of type 2 diabetes, indicatings the physiological importance of Zn2+ in β-cell autocrine regulation. Both serotonin and dopamine inhibit insulin secretion and may therefore participate in negative feedback of insulin release. However, serotonin and dopamine seem to be released from β-cells rather than from nerve terminals. Intracellular calcium oscillations are synchronized by depolarization spreading in clusters of β-cells and within islets via gap junctions and additional humoral and neural factors. At the level of the organ, pulsatile release of insulin is synchronized by the autonomic nervous system. ATP can synchronize glucose-induced calcium oscillations by the regulation of PKC activity via P2Y receptors. Neural signals such as NO, CO (from nonadrenergic, noncholinergic neurons), and acetylcholine (from sympathetic neurons) promote synchronization of calcium oscillation (Figure 8).
As for insulin secretion, Ca2+ is believed to be the main trigger of glucagon release. In α-cells, Ca2+ oscillations are induced in parallel to the stimulation of glucagon secretion by low concentrations of glucose. Control of glucagon secretion depends on cell metabolism, but a consensus is lacking on how this metabolism controls Ca2+ influx. Several mechanisms have been proposed, but the most convincing one suggests that glucose metabolism induces Ca2+ sequestration in the endoplasmic reticulum via the activation of the SERCA pump. This triggers the entry of Ca2+ in the cells and the stimulation of glucagon secretion. The most obvious positive feedback loop is glucagon, which amplifies its own secretion by raising cAMP. Synchronization of Ca2+ oscillations is not performed by gap junctions, which are lacking in α-cells. Ca2+ oscillations are not synchronous in the presence of low glucose levels. The synchronization of pulsatile glucagon release and Ca2+ oscillations are controlled by paracrine signals from β-cells (insulin, Zn2+, ATP, and GABA). These regulations do not control glucagon secretion when blood glucose is low, but might take place at high glucose concentrations when secretion of insulin and glucagon is pulsatile and in opposite phase. Somatostatin is a potent inhibitor of glucagon release and might thus also contribute to the regulation of pulsatile secretion of glucagon under hyperglycemic conditions. It is unclear if neural signaling has direct effects on the synchronization of glucagon pulsatility from the pancreas or if it is entirely mediated by paracrine factors secondary to the coordination of insulin pulsatility (Erik Gylfe, Lecture).

3. Cooperation between cAMP signaling and sulfonylurea in insulin secretion

cAMP regulates exocytosis by directly phosphorylating proteins of the exocytotic machinery in secretory cells. Intracellular cAMP signaling involves the protein kinase A (PKA), the cAMP guanine nucleotide exchange factor EPAC2, and cAMPgated ion channels. In β-cells, few proteins affecting insulin secretion are phosphorylated by PKA. EPAC2, which possesses guanine nucleotide exchange factor activity towards the GTPase Rap, is required in β-cells for the potentiation of insulin secretion by cAMP. Sulfonylureas stimulate insulin release by binding to the SUR1 subunit of the KATP channel. Inactivation of EPAC2 in mice shows that EPAC2 is necessary for the effect of cAMP and sulfonylureas on insulin secretion. cAMP levels are increased via activation of G-protein−coupled receptors such as GIP or GLP1 released by enterochromaffin cells after a meal, or such as the neurotransmitters of the vagus nerve (PACAP or VIP). This increase in cAMP potentiates the first and second phases of insulin secretion induced by glucose. The binding of norepinephrine and epinephrine (released by sympathetic nerves or adrenal gland) to Gi-coupled α2 receptors induces a decrease in cAMP levels. This decrease is linked to membrane repolarization, which induces a decrease in Ca2+ influx resulting in the inhibition of insulin secretion. Inactivation of the Cα subunit of PKA in β-cells induces the activity of PKA and stimulates the acute phase of insulin secretion. PKA stimulates insulin secretion by phosphorylating proteins, which promotes the interaction between exocytosis-associated proteins. As an example, PKA phosphorylates MyRIP, a scaffolding protein linking PKA to exocytosis-related protein, exocyst complex, and myosins. In mouse pancreatic β-cell lines, EPAC2 activates Rap1 in a cAMP-dependent manner. This activation is required for the first phase of potentiation of insulin secretion. The potentiating effect of incretins is linked to the interaction of EPAC2 with Rim2a, an effector of the small G protein Rab3. Activation of EPAC2 also stimulates a rise in Ca2+ levels via the inositol 1,4,5-triphosphate receptor and ryanodine receptor. Monitoring of EPAC2 activation in β-cells demonstrates that many sulfonylureas (except for gliclazide) directly activate EPAC2/Rap1 signaling. β-Cells from EPAC2-/- mice have a lower response to tolbutamide, confirming in vitro data. The activation of EPAC2 by sulfonylurea depends on the cAMP-binding domain of the protein. The number of hydrogen bonds between this cAMP-binding domain and the core of the sulfonylureas determines the binding affinities of the drug with EPAC2. Moreover, binding and activation of EPAC2 by sulfonylurea require the presence of cellular cAMP. The reduced cAMP levels induced by adrenalin block the activating effect of sulfonylurea on EPAC2 and Rap1. Thus, in a situation in which the sympathetic nerve is stimulated, the effect of sulfonylurea on activation of EPAC2 and Rap1 might be diminished (Figure 9). In humans, the interacting effect between incretin/cAMP signaling and sulfonylurea through EPAC2 has to be taken into account when establishing diabetes therapies (Susumu Seino, Lecture).

4. Autonomic nervous system and β-cell function

At the level of the endocrine pancreas, the autonomic nervous system influences acute hormone secretion, but also developmental processes required to establish islet structure and α- and β-cell number. These developmental processes are also regulated by glucose through glucose-sensing mechanisms in the CNS. During the weaning period, carbohydrates present in the diet activate the parasympathetic activity through Glut2-dependent glucose sensing and induce β-cell proliferation. This stimulatory effect of glucose is restricted to the early postnatal period, but is critical for the long-term control of β-cell mass and function. The specific inactivation of Glut2 in neurons induces a defect in the stimulation by the parasympathetic activity of β-cell proliferation at the weaning period in mice. This early defect results in impaired glucose-stimulated insulin secretion leading to progressive impaired glucose intolerance (Bernard Thorens, Lecture).

“Neural Orchestration of Metabolism and Islet function”
I. General points on the central control of energy balance and food intake
II. Mechanisms of direct detection of nutrients and hormones by the brain
III. Gastrointestinal and vagal detection of nutrients
IV. Control of β-cell function by the brain
V. Conclusion
Lectures during IGIS meeting