Glucagon antagonism in islet cell proliferation

The peptide hormones insulin and glucagon (gcg) are inextricably linked in the normal control of glucose homeostasis and in the dysregulated glucose homeostasis that defines diabetes mellitus. Pancreatic islets secrete both insulin and gcg in a manner that is tightly juxtaposed. β Cells secrete insulin, a peptide hormone that promotes the uptake and storage of carbohydrates and other nutrients in skeletal muscle and fat, while simultaneously repressing gcg secretion from pancreatic α cells and glucose efflux from the liver. Although loss of insulin function is the most conspicuous cause of both type 1 diabetes (T1D) and type 2 diabetes (T2D), hyperglucagonemia also drives hyperglycemia. Pharmacological approaches to blunt gcg action have gained some traction as a potential antidiabetic approach. In PNAS, Okamoto et al. (1, 2) present two papers evaluating gcg blockade on islet physiology. First, they examine the mechanisms of α-cell hyperplasia, a phenomenon that stands as a potential roadblock in the use of these glucagon receptor (Gcgr) antagonists (1). Second, they reveal that Gcgr antagonism when insulin action is absent can lead to normoglycemia and β-cell expansion (Fig. 1; ref. 2).

Fig. 1.

Pancreatic islets are composed of gcg-expressing α cells (red), insulin-expressing β cells (green), somatostatin-expressing δ cells (blue) and pancreatic polypeptide-expressing γ cells (white and red). Gcgr antagonism promotes α-cell hyperplasia possibly via amino acid-dependent mechanisms. Insr antagonism with S961 peptide can promote the expansion of β-cell mass mechanisms that may involve hepatic-derived SerpinB1. When combined, antagonism of both receptors synergistically promotes the formation of more β-cell mass.

Despite five decades of biochemical, physiological, and morphological research demonstrating that aberrant gcg production correlates with diabetes and suppression of gcg corrects the hyperglycemia of diabetes, gcg is not widely accepted as the direct cause of the metabolic abnormality. Since developing the first RIA for gcg in 1959 (3), we have studied gcg physiology and pathophysiology in rodents, dogs, and humans. In 1978, we showed that the gcg-suppressing agent somatostatin could ameliorate the metabolic disturbances of insulin deficiency (4). More recently, targeted disruption has gained traction as a potential treatment for diabetes.

Small-molecule antagonists, antisense oligonucleotides, and antibodies that block Gcgr action have all been investigated in preclinical models and clinical trials as novel therapeutics for the treatment of diabetes. Although all strategies have shown improvements in glycemic control, several side effects (e.g., hypercholesterolemia, hyperglucagonemia, α-cell hyperplasia) have slowed development of these drugs as a clinical class (5). For fear of such nonspecific side effects, biological approaches with strong specificity may be a safer approach to antagonize Gcgr. Both antisense oligonucleotides and human antibodies against Gcgr show promise to promote glycemic control in patients. Interrupting gcg signaling can normalize glycemia in T1D and T2D rodent models, suggesting that modifying gcg action may be therapeutic for patients with either form of diabetes (5). Unfortunately, although effective at lowering blood glucose, interrupting gcg signaling also results in hyperglucagonemia and α-cell hyperplasia. This side effect raises concerns for the long-term safety of gcg antagonism. However, glycemia in patients with inactivating mutations in insulin receptor (Insr) is refractory to traditional insulin restorative therapies and alternatives are needed.

More recently, the work of Wang et al. (6) demonstrated evidence that the blockage of gcg action eliminates all laboratory evidence of T1D even when insulin levels were below our limit of detection. Our observations in T1D rodent models have been replicated by others who have met these results with skepticism (7). The results presented in PNAS further refine these conclusions (2). Using REGN1193, a fully human Gcgr monoclonal antibody, Okamoto et al. (8) (Regeneron) have reported effective glucose control in diabetic rodents and primates. Hyperglucagonemia and excess proliferation of α cells have also resulted following administration of REGN1193, which phenocopies global or liver-specific Gcgr^−/− mice (9, 10). Using both REGN1193 and S961, Okamoto et al. (2) (the Gromada laboratory at Regeneron) expand upon this research paradigm, showing S961 (20 nmol⋅wk⁻¹) rapidly induces hyperglycemia that is rescued by REGN1193 administration. Amazingly, this simultaneous Gcgr/Insr antagonism prompted an increase in insulin production and a further increase in β-cell mass over S961 alone, suggesting that Gcgr antagonism may be beneficial in promoting functional β-cell mass under conditions of severe insulin resistance.

Identifying the mechanisms leading to islet cell hyperplasia when hormone signaling is lost has been intensely sought. While glucose stimulates β-cell replication in islet culture studies, the possibility that other factors contribute to β-cell hyperplasia when insulin signaling is impaired exists. Using parabiosis, El Ouaamari et al. demonstrated that factors present in the blood of mice with liver-specific knockout of Insr (LIRKO) can stimulate β-cell proliferation in wild-type mice (11). Furthermore, culturing islets in serum or media conditioned with liver explants of LIRKO mice stimulated wild-type β cells to grow. In a follow-up study, El Ouaamari et al. (12) identified SerpinB1 as a hepatic-derived factor that stimulates β-cell proliferation. Although interrupting SerpinB1 expression convincingly blocked β-cell proliferation when mice were fed a high-fat diet, it was insufficient to rescue mice completely from hyperglycemia or β-cell hyperplasia when they were administered S961, suggesting that other factors could be at play.

Concurrently, Yi et al. (13) reported that S961 treatment of mice resulted in a dramatic increase in β-cell proliferation. They ascribed this proliferation to a hepatic-derived factor, Angiopoietin-like 8 (Angptl-8) or betatrophin. Angptl proteins traffic with LDL particles, where they prevent the breakdown of triglycerides by lipoprotein lipases, linking fatty acid metabolism to β-cell replication. Two independent studies using S961 treatment in human islet-transplanted mice showed that endogenous mouse pancreatic cells, but not adult human islet graft β cells, grow in response to interrupted insulin signaling even when hepatic Angptl-8 expression was elevated (14, 15). An additional study by Gusarova et al. (16) refuted the initial finding, demonstrating that Angptl-8 expression has no effect on mouse β-cell replication. Recently, the original study implicating Angptl-8 was retracted by the authors (17, 18). However, together, these studies have conclusively demonstrated that a liver–β-cell axis exists, where the insulin signaling stimulates factors (e.g., glucose, SerpinB1) that feed back to the β cell to communicate the liver’s status, resulting in a tuning of the β-cell mass, and thus secretory capacity.

Similar to studies on the liver–β-cell axis, islet transplantation studies in global and liver-specific Gcgr^−/− mice support that a factor(s) increased in the serum of these mice feeds back to stimulate the α-cell mass (10). However, the source of this elusive factor(s) is somewhat controversial. Gcgrs are expressed on multiple tissues, including liver, kidney, adipose tissue, brain, and pancreatic islets. In adipose tissue, gcg may stimulate lipolysis. Alternatively, gcg’s primary physiological function is to stimulate glycogenolysis and gluconeogenesis in the liver. Both global and liver-specific Gcgr knockout in mice results in α-cell hyperplasia, suggesting that circulating factors stimulating α cells must be hepatic-derived (10). Understanding these and other signals that regulate gcg secretion and α-cell mass could translate to new strategies to reduce hyperglycemia.

Recently, Ben-Zvi et al. (19) reported increased adipocyte-derived serum Angptl-4 levels when gcg signaling is interrupted.

Pharmacological approaches to blunt gcg action have gained some traction as a potential antidiabetic approach. In PNAS, Okamoto et al. present two papers evaluating gcg blockade on islet physiology.

Angptl-4 overexpression results in hypertriglyceridemia, whereas loss of Angptl-4 expression results in lower triglyceride levels. Ben-Zvi et al. (19) reported that Angptl-4 treatment stimulated α-cell proliferation, whereas Angptl-4^−/− rescued α-cell hyperplasia in Gcgr^−/− mice. In PNAS, Okamoto et al. (1) refute these findings by demonstrating that Angptl-4^−/− mice treated with Gcgr monoclonal antibody develop severe α-cell hyperplasia. The discrepancies of these observations remain unclear. Okamoto et al. (1) suggest that hyperaminoacidemia observed in mice with impaired Gcgr signaling drives α-cell proliferation. Indeed, a recent report shows that islets cultured in high levels of amino acids have increased α-cell proliferation (20). It remains unclear which amino acids could be driving α-cell proliferation, how this α-cell proliferation could occur mechanistically, and why this α-cell proliferation would be selective to stimulate α-cell hyperplasia. The observation that β-cell mass increases with dual Insr/Gcgr antagonism suggests that gcg-targeted therapies could restore euglycemia while promoting expansion of functional β-cell mass in patients with severe insulin resistance (2). Okamoto et al. (1) suggest hyperaminoacidemia resulting from Gcgr blockade may contribute to this expansion as well. It is intriguing to think that this strategy could increase β-cell mass in patients with milder forms of insulin resistance as in T2D or in patients with newly diagnosed T1D.

These results prompt several exciting questions. Can gcg blockade be used to prompt β-cell regeneration in diabetes? This strategy may be particularly effective in T1D, where insulin antagonism may not be needed due to insulin deficiency. Do new β cells derive from a ductal cell, β-cell, or α-cell origin? The conversion of α cells to insulin-producing β-like cells in this context may be of great intrigue. Recent work using lineage tracing models has established conversion of α cells to β cells (21), which can be increased by GABA receptor agonism (22, 23). At the Touchstone Center, as much as 20% of newly formed β cells may derive from α-cell precursors (24). If α-cell–to–β-cell transdifferentiation can be enhanced by pharmaceutical approaches, then Gcgr antagonism may be beneficial.