GPCR Promiscuity Reshapes Islet Physiology

A landmark in diabetes research was the discovery of islet hormones 100 years ago (1), which caused a paradigm shift in the treatment of diabetes. Although the presence of glucagon in islets was discovered around the same time as insulin, it was largely viewed as a contaminant in insulin preparations. It took another 30 years before any major insights into glucagon were made, including the development of a radioimmunoassay by Roger Unger in 1959 to measure circulating concentrations of glucagon (2). Subsequent studies by Unger and others ultimately led to the proposal that diabetes was a bihormonal disease, characterized by insufficient insulin action and excessive effects of glucagon (3). Over the past few decades, glucagon was almost exclusively positioned as a counterregulatory hormone, exerting biological effects on the liver that are in direct opposition to the action of insulin: two hormonal adversaries that push and pull on metabolic processes to finely tune blood glucose levels (4).

The incretin axis is another example of an essential calibrator of metabolism. The gut peptides glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1) communicate the intake of nutrients to various systems that efficiently regulate gastric emptying, satiety, blood flow, and nutrient uptake to properly respond to a meal (5). The islet hormones are central mediators of the incretin effect, where GLP-1 enhances insulin secretion and inhibits glucagon secretion (6). These actions have supported the development of the GLP-1 receptor (GLP-1R) agonist class of drugs that continue to be the foundation for incretin-based therapies targeting hyperglycemia and obesity (7). The development of GLP-1 biology for the treatment of metabolic disease may be second only to the discovery of insulin as an advance in diabetes therapeutics. Curiously, GLP-1 is produced by proteolytic processing from the same proglucagon precursor that encodes glucagon (8). Differential processing of proglucagon leads to tissue-specific production of each peptide, with GLP-1 produced in intestinal L cells and glucagon produced in pancreatic α-cells. Moreover, there are two distinct receptors for these peptides, GLP-1R and the glucagon receptor (GCGR), which have unique expression patterns throughout the body. Indeed, the most detailed actions of glucagon are described in hepatocytes (increase of endogenous glucose production), where the GCGR and not the GLP-1R is expressed. Glucagon has long been known to enhance insulin secretion (9), but this effect required the presence of a glucose level above a critical threshold (10). Analogous to the direct effect of glucagon on β-cells, GLP-1 stimulated insulin secretion only when extracellular glucose exceeded a critical threshold level (11). However, the similar actions of glucagon and GLP-1 are often overlooked because these two hormones are typically depicted as working in opposition; glucagon raises glycemia and GLP-1 lowers it. Adding further complexity to this equation is the promiscuity glucagon displays at both receptors (GCGR and GLP-1R).

In this “Classics in Diabetes” article, we revisit a 1998 Diabetes article by Moens et al. (12), which was inspired by the second messenger concept that resulted from studies of how glucagon stimulates glucose production in the liver. Earl Sutherland discovered that binding of glucagon (the first messenger) to GCGR on liver cells stimulates the production of cAMP (the second messenger) inside these cells (13). Elevated cAMP concentrations then alter the phosphorylation state and function of key enzymes of glycogen metabolism (14), a landmark discovery that immediately influenced the pancreatic islet research field. One example is the now abandoned (but hotly debated in the 1970s) working hypothesis that glucose-stimulated insulin release was mediated by a rise in cAMP concentrations in β-cells (15). However, studies were typically performed on isolated islets, which are composed of different cell types in which cAMP levels could change in opposite directions contingent on the stimuli. Moreover, as mentioned above, pancreatic islets secrete not only insulin but also glucagon, which might influence cAMP levels in a way that could hinder the correct interpretation of results. In the early 1980s, the Pipeleers laboratory had worked out techniques to overcome these difficulties by purification of α- and β-cells from collagenase-isolated islets (16). It was observed that the amplitude of glucose-induced insulin release from purified β-cell preparations was very low in the absence of glucagon but became more robust when this counterregulatory hormone was added at nanomolar concentrations to the purified cells (17). It was also demonstrated that the glucose effect on cAMP levels in β-cells required the simultaneous presence of glucagon (18), an effect that could later be explained by expression of the type VIII adenylate cyclase isoform in β-cells (19). Potentiation of nutrient-induced insulin release was observed not only with glucagon (9) but also with the incretin hormones GLP-1 (20) and GIP (21). The molecular basis for this phenomenon was investigated in purified rat β-cells that abundantly express the genes encoding glucagon GLP-1 and GIP receptors, which would explain the effect of glucagon GLP-1- and GIP on cAMP production and insulin release (22).

The idea that GCGRs are at the basis of glucagon-induced insulin release was revised by the 1998 Diabetes article that proposed a dual mechanism via which glucagon stimulates β-cells: 1) a high-affinity mechanism mediated by GCGR and 2) a low-affinity mechanism caused by GLP-1R (12). Evidence for this duality was provided with use of purified rat β-cells for radioligand binding experiments and measurements of cAMP production and insulin release. Glucagon binding to β-cells was incompletely competed for not only by des-His1-[Glu⁹]glucagon-amide (a specific GCGR antagonist) but also by GLP-1 or the GLP-1R antagonist exendin-(9-39)-amide (Ex9). In contrast, glucagon binding to hepatocytes (which express GCGR but not GLP-1R) was completely abolished by des-His1-[Glu⁹]glucagon-amide but not influenced by GLP-1 or Ex9. Binding of GLP-1 to β-cells was significantly competed for by GLP-1 in the low nanomolar range and by glucagon in the high nanomolar range. In agreement with the binding experiments, glucagon-induced cAMP production in β-cells was antagonized both by des-His1-[Glu⁹]glucagon-amide and Ex9, whereas GLP-1–induced cAMP production was suppressed only by Ex9. In perifusion experiments it was observed that 1 nmol/L glucagon potentiation of glucose-induced insulin release could be counteracted in the presence of 10 nmol/L Ex9. The study concluded that rat β-cells are competent to detect both endocrine glucagon (high-affinity GCGR) and paracrine glucagon (low-affinity GLP-1R) (Fig. 1). These actions of glucagon are evolutionarily conserved as human β-cells coexpress GCGR and GLP-1R (23), both of which are necessary for the insulinotropic actions of glucagon (24).

Over two decades later, our path toward understanding the implications of glucagon activity at the GLP-1R continued with the study of mice with selective deletion of the GCGR in β-cells (Gcgr^{βcell−/−}) (24). Building on the established insulinotropic actions of glucagon, we demonstrated a concentration-dependent action of glucagon that enhances insulin secretion in a glucose-dependent manner in perifused mouse islets from control mice. To our surprise, glucagon was equally able to stimulate insulin secretion in islets from Gcgr^{βcell−/−}. The article by Moens et al. (12) led us to interrogate the possibility that much of the insulinotropic actions of glucagon in β-cells were mediated by the GLP-1R. In doing so, we were able to demonstrate that glucagon-stimulated insulin secretion is greatly reduced in wild-type islets exposed to Ex9 or in mice with β-cell deletion of the GLP-1R. Moreover, through a series of experiments we could show that endogenous glucagon secretion from α-cells is essential for β-cell tone and insulin secretion in mouse and human islets, can lower elevated glycemia, and is indispensable for normal postprandial glucose control—all actions mediated through the β-cell GLP-1R (24,25). These conclusions have been buttressed by orthogonal approaches from multiple other groups (26–30), establishing α-cell to β-cell communication via glucagon activity at the GLP-1R as an important axis for metabolic control. Importantly, this islet paracrine relationship has been demonstrated in healthy human subjects (31) and proposed to be of increasing importance with metabolic stress (32).

The implications that glucagon is an important physiological activator of the GLP-1R are far-reaching. First, the important activity of glucagon at the β-cell GLP-1R for postprandial insulin secretion requires a reshaping of the conventional view that glucagon is exclusively a fasting-induced hormone that prevents hypoglycemia in a counterregulatory manner (i.e., only works in opposition to the effects mediated by insulin). Clearly glucagon has a more complex physiological role, having important actions for glucose metabolism across a broad range of physiological states. This does beg the question of whether glucagon also acts on the GLP-1R outside of the islet. The GLP-1R has been shown to have important activity in the central nervous system, the cardiovascular system, and various immune cells (33). Whether glucagon can engage these systems remains unknown. Glucagon has been shown to have effects on satiety, and the presence of a GCGR in the central nervous system is debated. Could neuronal GLP-1Rs be mediating a central effect of glucagon on food intake, and if so, is this a physiological or pharmacological action? GLP-1R activity has important effects on cardiovascular and immune functions, with direct implications to GLP-1 pharmacology. Could glucagon be the physiological ligand for these processes? Indeed, much of the biological importance of GLP-1R in different cell types is often derived from studies where investigators use deletion or antagonism of the GLP-1R, which is not informative regarding the ligand responsible for agonizing the receptor. Glucagon levels are highest in the intraislet space, followed by portal circulation, with the concentrations in these locations being substantially higher than systemic concentrations. This raises the question of whether the endocrine actions of glucagon on the GLP-1R apply to areas outside of the islet (hepatocytes do not express GLP-1R [34]) or are exclusively mediated by GCGR. Further complicating this area are reports that α-cells can process proglucagon to produce GLP-1 (35), implicating that certain circumstances lead to the production of two ligands for the GLP-1R in the islet. How and why this occurs remain an ongoing investigation, as is the question of whether islet-derived GLP-1 can function in an endocrine manner. Finally, it remains unclear why the β-cell expresses both GCGR and GLP-1R if much of the activity of glucagon for insulin secretion, the primary function of β-cells, occurs via the GLP-1R. Could the β-cell GCGR have an important role beyond insulin secretion?

The last 5 years have seen a revolution in glucagon biology. New discoveries of physiological actions of glucagon have been accompanied by rapid developments in novel therapeutic advances that leverage glucagon pharmacology for the treatment of metabolic disease. As progress in this area continues to be made, it is important to be reminded that many of these ideas currently being presented as novel with updated approaches and tools have previously been established to some degree. Although the implications of glucagon activity at the GLP-1R may not have been entirely clear when the original observations by Moens et al. (12) were reported, and likely remain incompletely resolved today, this work illustrates how rigorous science continues, 25 years after its publication, to influence scientific progress today.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.