Importance of Nonionic Signals for Glucose-Induced Biphasic Insulin Secretion:

Glucose-induced biphasic insulin secretion was first recognized in pancreas perfusion experiments (4). Soon thereafter, it was found that a similar biphasic insulin release can be seen when isolated islets were tested in perifusion experiments (5). The biphasic response takes place when extracellular glucose concentration is abruptly raised from a substimulatory to a stimulatory level and kept at that high level. In contrast, when the extracellular glucose concentration is gradually raised, no biphasic response is elicited and insulin secretion is progressive (4). In normal subjects, fasting plasma glucose concentration is ∼5 mmol/l, which is already stimulatory, and, postprandially, it rises to 8 mmol/l, at the very most, during a 30- to 60-min interval. A “glucose jump” from, e.g., 3 to 16.7 mmol/l often imposed in vitro, never occurs in normal subjects under physiological conditions. In that sense, a biphasic insulin release by the β-cell upon glucose stimulation, a salient feature of this cell type, is an experimental concept.

There are several thousands of islets in the pancreas, and the response of individual rat islets is highly heterogeneous. Some islets show a typical biphasic response, but many others show either predominant first- or second-phase responses (6). This finding agrees well with the electrophysiological heterogeneity of individual rat islets (7). When insulin secretion from many randomly selected islets is collectively analyzed, total response is clearly biphasic, as it is in the perfused pancreas. This potentially important fact, i.e., functional heterogeneity among islets, is not considered in the following discussion because the reason for such heterogeneity is unknown at present.

Elevation of cytosolic free Ca²⁺ concentration ([Ca²⁺]_i) by a Ca²⁺ ionophore, A23187, produces a monophasic insulin response, whereas 12-O-tetradecanoyl-phorbol-13-acetate, an agonist of protein kinase C (PKC), causes gradually increasing insulin release, and a combination of the two elicits a biphasic release similar to that occurring in response to glucose (8). Therefore, it was speculated that the glucose-induced biphasic response might be due to an elevation of [Ca²⁺]_i and subsequent activation of PKC by elevated [Ca²⁺]_i. By now, closure of ATP-sensitive K⁺ (K_ATP) channels and subsequent electrophysiological events have been established as the basis for glucose-induced rise in [Ca²⁺]_i (9).

The biphasicity of insulin secretion upon glucose stimulation, especially a prominent second phase, is much more evident in rat than mouse islets (10–12). Recent studies on membrane potential and [Ca²⁺]_i changes in rat β-cells clearly revealed that the temporal profiles of ionic events during the second phase do not coincide with a gradually increasing pattern of insulin secretion (7). Namely, the depolarization, number of action potentials, and elevation of [Ca²⁺]_i during the second phase all remained stable (6). Therefore, a biphasic insulin release upon glucose stimulation cannot be simply explained by the K_ATP channel closure and subsequent elevation of [Ca²⁺]_i.

We (12) and others (13) found that glucose strongly augments Ca²⁺-triggered insulin release in a K_ATP channel–independent manner, and we consider that the second-phase response is produced mostly though this signaling pathway. Because glucose-induced ionic events linked to insulin exocytosis originate in the K_ATP channel closure, and signals involved in the K_ATP channel–independent pathway are not related to electrophysiological events in the β-cell, we proposed a separation of the K_ATP channel–dependent and –independent signalings on the basis of their ionic or nonionic nature (3).

Thus, an abrupt rise in extracellular glucose concentration causes an elevation of [Ca²⁺]_i through the ionic signaling as outlined above, and insulin release is triggered by Ca²⁺-dependent processes. This corresponds to the first phase. Ca²⁺-triggered insulin release is then gradually augmented/potentiated by the K_ATP channel–independent signals, and a prominent second phase is generated (14,15).

Until now, phospholipase A₂ (a Ca²⁺-independent, ATP-dependent subtype), guanosine triphosphate, ATP, and glutamate were proposed as possible conveyer(s) of the K_ATP channel–independent glucose signaling (16–20). We have recently suggested a role of protein acylation (21,22). Namely, cerulenin, an inhibitor of protein acylation, selectively obliterated nutrient-induced insulin release without metabolic perturbation in the islet cells (21). Based on this observation, we considered that it is likely that glucose causes fatty acylation of protein or proteins engaged in exocytosis of insulin. This idea stemmed from the malonyl-CoA hypothesis originally proposed by Corkey et al. (23). According to this hypothesis, citrate derived from the tricarboxylic acid cycle is accumulated in the cytosol and converted to malonyl-CoA, which suppresses carnitine palmitoyl transferase I (CPTI) on the mitochondrial outer membrane. Because CPTI is a rate-limiting enzyme for mitochondrial transport of long-chain acyl CoAs (LC-CoAs), LC-CoAs are accumulated in the cytosol as a result of CPTI suppression. Especially, accumulation of palmitoyl-CoA will then facilitate fatty acylation of protein(s) (24) involved in insulin exocytosis, leading to augmentation and potentiation of exocytosis. On the other hand, overexpression of malonyl-CoA decarboxylase in insulin-secreting INS-1 cells eliminated glucose-induced accumulation of malonyl-CoA, yet glucose-induced insulin secretion remained unaltered (25). Thus, in INS-1 cells, the validity of the malonyl-CoA hypothesis is questioned.

Recently, insulin secretion has been studied in patients or animals without functional K_ATP channels (26–29). The patients lacking K_ATP channels because of mutations in the genes encoding the sulfonylurea receptor 1 (SUR1), a subunit of the K_ATP channel, suffer from severe neonatal hypoglycemia. The disease is called persistent hyperinsulinemic hypoglycemia of infancy (PHHI) (26,27). In contrast, mice with targeted disruption of the genes encoding either SUR1 or the other subunit of the K_ATP channel, inward-rectifying K⁺ channel 6.2 (Kir6.2), remain euglycemic under regular feeding (28,29).

From the conventional viewpoint that the first phase is mostly, if not entirely, due to glucose-induced closure of the K_ATP channels eventually causing a sharp rise in [Ca²⁺]_i (see above), it is rather surprising that the first-phase insulin secretion was clearly elicited in PHHI subjects upon intravenous bolus injection of glucose (26). The second phase appears absent, probably because of a markedly reduced β-cell insulin content. Straub et al. (27) examined the function of islets isolated from such patients and found that application of glucose alone elicited a clear-cut insulin release: the basal level of [Ca²⁺]_i in β-cells from these patients was normal and did not increase upon glucose stimulation (27). These observations indicate that the K_ATP channel–independent nonionic signals alone can trigger insulin release by the β-cell, provided the channel is closed: the absence of functional K_ATP channels appears to be equivalent to persistent closure of the channel.

When islets from normal rodents are stimulated with glucose in the presence of sulfonylurea (a K_ATP channel– closer), both first and second phases are elicited (30–32). For unknown reasons, islets from SUR1 or Kir6.2 knockout mice lack such rapid response to glucose (28,29) (Table 1). Such a difference may explain why mice without functional K_ATP channels, unlike humans with the similar genetic abnormality, do not suffer from persistent hyperinsulinemic hypoglycemia.

Glucose-induced biphasic insulin secretion by pancreatic islets is a characteristic feature of β-cell physiology. As far as we are aware, biphasic hormone release as seen in the β-cell, especially the prominent second-phase secretion, upon a square-wave application of physiological stimuli, has not been reported in other endocrine or neuronal cells. For example, thyrotropin-releasing hormone elicits a sharp, distinct first-phase hormone release followed by a gradually decreasing tiny second phase from GH₄C₁ pituitary tumor cell lines (33). Understanding of the molecular basis of this unique response of the β-cell will hopefully contribute to the treatment of patients with diabetes through development of newer insulinotropic agents.

Work cited from the authors’ laboratory was supported by a grant from the Ichiro Kanehara Foundation (to M.K.) and Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan (to T.A.).