Hypoglycemic Effect of Combined Ghrelin and Glucagon Receptor Blockade

Blood glucose level is normally tightly regulated by the opposing actions of insulin and glucagon. Recent studies (1) suggest that dissolution of the usual paracrine restraint of pancreatic α-cell function by insulin and the resulting aberrant hyperglucagonemia mediates the hyperglycemia and altered metabolic state of insulin-requiring diabetes. This hypothesis is supported by observations in streptozotocin (STZ)-treated glucagon receptor (GcgR) knockout (Gcgr^−/−) mice, which display normal glucose tolerance despite being insulin deficient due to STZ-induced pancreatic β-cell destruction (2–4). Furthermore, blockade of GcgR function, neutralization of glucagon, and suppression of hyperglucagonemia all can prevent and/or correct hyperglycemia in models of and patients with diabetes (5–8). Although not universally supported (9,10), this glucagon-focused theory opens a new avenue of targeting GcgRs for diabetes treatment.

In settings lacking both glucagon and insulin actions, the signals regulating blood glucose are unclear but may involve ghrelin. Ghrelin is a peptide hormone secreted primarily by specialized endocrine ghrelin cells within the gastric mucosa (11). After acylation by ghrelin-O-acyl transferase, ghrelin can bind to its receptor GHSR (growth hormone [GH] secretagogue receptor [i.e., ghrelin receptor]) to stimulate GH secretion and food intake (11). Ghrelin also regulates blood glucose, having at its disposal several targets with which to influence glucose handling. These include not only stimulatory effects on food intake and GH secretion, but also decreases in insulin secretion and insulin sensitivity and increases in circulating levels of glucocorticoids, glucagon release, hepatic gluconeogenesis, and hepatic autophagy, all of which can raise glycemia (11–16). Notably, ghrelin administration or transgenic overexpression raises blood glucose levels, whereas genetic deletion or pharmacologic blockage of key elements of the ghrelin system lowers blood glucose levels (13–15,17–19). Indeed, mice lacking the usual ghrelin response to marked caloric restriction develop severe hypoglycemia and experience increased mortality (12,16,20). Interestingly, plasma ghrelin levels also are negatively regulated by blood glucose, presumably because of effects on ghrelin secretion (21–23).

Thus, ghrelin both is regulated by glucose and regulates glycemia, with protective effects to maintain normoglycemia becoming apparent during marked caloric restriction. The close association of ghrelin with blood glucose positions it as a key regulator of glycemia in the settings of absent glucagon action or combined absent glucagon plus insulin actions, as would occur in patients with type 1 diabetes should anti-GcgR agents proceed to market. Here, we investigated the effects of genetic and pharmacologic GcgR blockade with or without STZ-induced pancreatic β-cell destruction on ghrelin and, in turn, the effects of manipulating this ghrelin response on glycemia.

All animal experiments were approved by the University of Texas Southwestern Medical Center Institutional Animal Care and Use Committee. Gcgr^−/− (24) and wild-type (Gcgr^+/+) littermates are on a C57BL/6J background, unless indicated. Ghsr-null and wild-type littermates are on a C57BL/6N background. db/db (BKS.Gg-Dock7^m+/+Lepr^db/J) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were housed under a 12-h dark/light cycle with free access to water and a standard chow diet (2016 Teklad Global 16% protein diet; Envigo, Indianapolis, IN) for C57BL/6N and Ghsr-null mice or 2019 Teklad Global 19% protein extruded diet (Envigo) for db/db, Gcgr^−/−, and wild-type littermates, except as indicated. Male mice were studied.

Unless otherwise specified, samples were collected between 10:00 and 11:30 a.m. by quick superficial temporal vein bleed into EDTA-coated microtubes containing p-hydroxymercuribenzoic acid (final concentration 1 mmol/L; Sigma-Aldrich, St. Louis, MO) kept on ice and HCl was added to the plasma after centrifugation to achieve a final concentration of 0.1 nmol/L, as described (20). Blood glucose levels were measured using a Bayer CONTOUR Blood Glucose Monitoring System.

Primary cultures of gastric mucosal cells were prepared from 8- to 14-week-old male C57BL/6N mice, as described previously (20,21).

Total RNA isolation and cDNA synthesis were performed as described previously (20). Quantitative PCR was performed using an Applied Biosystems ViiA 7 Real-Time PCR System and TaqMan primer/probe chemistry using the gene-specific primers listed in Supplementary Table 1.

Tissue preparation and immunohistochemistry were performed as described previously (13,25).

Atenolol was administered to 12- to 18-week-old Gcgr^−/− mice and wild-type littermates as described previously (20,26). Mice were injected with atenolol (10 mg/kg body weight [BW] i.p.) or its vehicle (2 mmol/L HCl) twice daily for 3 days (at 7:30 a.m. and 7:30 p.m.). Blood was collected at 9:00 a.m. on day 2 (ad libitum fed condition), and the mice were fasted for 16 h beginning at 5:00 p.m. on day 2. On day 3, the mice were dosed with atenolol or vehicle at 7:30 a.m. while still fasting, and blood was collected at 9:00 a.m. (fasted condition).

Hyperglycemic (blood glucose concentration 275–400 mg/dL) db/db mice matched for blood glucose level were randomized to receive either a single weekly injection (5 mg/kg BW s.c.) of monoclonal antibody (mAb) B (a fully human GcgR mAb developed by REMD Biotherapeutics, Inc., Camarillo, CA) (6,27) or vehicle. The vehicle-treated group was further divided into an ad libitum fed group or a pair-fed group, in which they were given an amount of food daily that matched the daily food intake of the mAb-administered mice. Blood glucose and plasma ghrelin levels were measured at 9 weeks of age, 5 days after the last weekly injection, at 10:00 a.m.

The Ghsr-null mice were administered STZ (150 mg/kg i.p.) and then treated with a single injection of mAb B (10 mg/kg s.c.) or vehicle when the blood glucose concentration reached between 275 and 450 mg/dL. Blood glucose levels were measured 4 days after treatment at 10:00 a.m. and again 1 day later after a 24-h fast.

GHSR antagonist ([D-Lys³]-GHRP-6) or vehicle (water) was administered at a dose (200 nmol/L/30 g BW i.p.) described previously (17) twice daily for 5 consecutive days to Gcgr^−/− mice that had or had not previously received STZ treatment (150 mg/kg i.p.). On the fourth day of drug administration, the mice were fasted overnight (12 h), after which the last injection of [D-Lys³]-GHRP-6 was immediately administered. Thereafter, blood glucose levels were monitored by tail bleed at the indicated time points while the mice were continued under fasting condition. Of note, [D-Lys³]-GHRP-6, although developed and widely used as a GHSR antagonist, has some off-target effects on other receptors (28).

Ghrelin concentrations in plasma and cell culture media were determined by ELISA (catalog #EZRGRA-90K for acyl-ghrelin, catalog #EZRGRT-91K for total ghrelin; Merck Millipore). Plasma insulin was analyzed using an ultrasensitive mouse Insulin ELISA kit (catalog #90080; Crystal Chem, Downers Grove, IL). The end point calorimetric assays were performed using a BioTek PowerWave XS Microplate spectrophotometer and KC4 junior software.

All data are expressed as the mean ± SEM. All statistical analyses and graph preparations were performed using GraphPad Prism version 6.0. Student t test, one-way ANOVA, and two-way ANOVA followed by post hoc comparison tests were used to test for significant differences among test groups, as indicated in the figure legends. Data with significant unequal variance assessed using Bartlett’s test were log transformed to perform ANOVAs. The strength of the linear relationship between two sets of variables was compared by Pearson correlation coefficient. Outliers were detected by Grubb test. P values <0.05 were considered statistically significant, and P values ≥0.05 and <0.1 were considered to be evidence of statistical trends.

To confirm and extend previous observations showing glucose suppression of plasma ghrelin in rats and humans (22,23), we measured circulating ghrelin levels 30 min after gavage administration of d-glucose (2 g/kg BW) or saline in 24-h–fasted 8- to 12-week-old C57BL/6N mice, at which time blood glucose levels peak (29). Oral glucose not only raised blood glucose levels but also significantly reduced plasma acyl-ghrelin levels (Fig. 1A and B). Acyl-ghrelin negatively correlated with blood glucose (Fig. 1C). Replicating our previous findings that were suggestive of direct modulation of ghrelin secretion by glucose (21), we found that glucose can directly regulate ghrelin secretion from ex vivo primary cultures of dispersed gastric mucosal cells derived from C57BL/6N mice (Fig. 1D), with a glucose concentration of 10 mmol/L reducing ghrelin secretion by 17.6 ± 2.6% and a concentration of 1 mmol/L glucose increasing it by 150.4 ± 6.5% as compared with cells cultured in normoglycemic-like conditions (5 mmol/L glucose). Exposure to culture media with no glucose increased ghrelin secretion even further, by 253.1 ± 12.1% compared with that observed with a glucose concentration of 5 mmol/L. These results suggest that the ghrelin secretion machinery within ghrelin cells is sensitively tuned to blood glucose in the hypoglycemic range.

Based on the above results, we hypothesized that Gcgr^−/− mice, which demonstrate lower than usual blood glucose levels throughout the day (24), have high circulating ghrelin levels. Blood glucose levels were significantly lower in 8- to 10-week-old Gcgr^−/− mice in both ad libitum fed and overnight-fasted conditions when compared with wild-type (Gcgr^+/+) littermates (Fig. 2A and D). This was accompanied by significantly elevated circulating concentrations of acyl- and total ghrelin in ad libitum fed (2.9-fold and 1.7-fold, respectively) and overnight-fasted (2.2-fold and 1.3-fold, respectively) conditions (Fig. 2B, C, E, and F). Acyl-ghrelin levels observed in ad libitum fed Gcgr^−/− and Gcgr^+/+ mice on a C57BL/6J background (Fig. 2) were similar to those observed in mice on a mixed C57BL/6J and C57BL/6N background (data not shown). Gcgr^−/− mice did not demonstrate changes in gastric ghrelin cell number, but preproghrelin mRNA (Ghrl) expression was elevated in gastric mucosal cell lysates (Supplementary Fig. 1). A much lower level of preproghrelin mRNA expression was observed in the pancreata of Gcgr^+/+ mice, which did not differ from that in Gcgr^−/− mice (Supplementary Fig. 2). Ghrelin immunoreactivity was undetectable in pancreatic islets of both 10-week-old Gcgr^−/− and Gcgr^+/+ mice (data not shown). These results suggest that the increase in plasma ghrelin in Gcgr^−/− mice is not likely contributed to by alterations to pancreatic islets (a major site of ghrelin production in the fetal period) (15) but does involve elevated ghrelin biosynthesis within the stomach. In contrast, preproglucagon mRNA expression, which was higher in the pancreata of Gcgr^+/+ mice compared with their gastric mucosa, was exponentially elevated by GcgR deletion. Insulin I and Insulin II mRNA expression levels were unchanged by GcgR deletion (Supplementary Figs. 1 and 2).

To determine whether the low blood glucose levels evident in Gcgr^−/− mice contributes to their elevated plasma ghrelin levels, we tested whether glucose regulation of plasma ghrelin is preserved in Gcgr^−/− mice. Overnight-fasted 12- to 14-week-old Gcgr^−/− mice and wild-type littermates were administered glucose (2 g/kg BW) by oral gavage, increasing blood glucose levels in both groups (Fig. 3A). The increase in blood glucose levels, which was lower in the Gcgr^−/− mice, was accompanied by significant falls in acyl- and total ghrelin levels in both genotypes when compared with control mice administered saline (Fig. 3B and C), suggesting maintained glucose regulation of plasma ghrelin upon GcgR deletion. Although baseline ghrelin levels were higher in Gcgr^−/− mice, similar to previous cohorts (Fig. 2), they fell to statistically similar levels in both Gcgr^−/− and wild-type mice after glucose administration.

We also tested whether glucagon could directly influence ghrelin secretion. Using ex vivo primary cultures of dispersed gastric mucosal cells from adult C57BL/6N mice, we demonstrated that glucagon failed to induce any significant changes in ghrelin secretion (Supplementary Fig. 3).

Catecholamines released from sympathetic neurons also stimulate ghrelin secretion from ghrelin cells, contributing substantively to the usual rise in plasma ghrelin levels induced by caloric restriction (20,26). To better understand whether the increased plasma ghrelin levels in Gcgr^−/− mice involves activation of the sympathetic system, we administered 8- to 12-week-old Gcgr^−/− mice and wild-type littermates either the β-blocker atenolol or vehicle (10 mg/kg BW i.p.) every 12 h for 3 successive days. Atenolol did not impact ad libitum fed or fasted blood glucose levels (Fig. 4A and D). However, atenolol treatment significantly reduced plasma ghrelin levels in ad libitum fed and overnight-fasted conditions in both wild-type and Gcgr^−/− mice (Fig. 4B, C, E, and F), just as had previously been demonstrated in wild-type mice (20,26).

As mentioned, Gcgr^−/− mice are resistant to the hyperglycemia usually induced by STZ (2,4). Furthermore, plasma ghrelin becomes elevated in the STZ model of type 1 diabetes, whereas GHSR knockout mice and mice administered GHSR antagonist ([D-Lys³]-GHRP-6) display ameliorated STZ-induced hyperphagia and hyperglycemia (30–32). Thus, we next assessed changes to plasma ghrelin levels in STZ-treated Gcgr^−/− mice. Similar to previous studies, STZ administration (two doses of 150 mg/kg BW i.p., given 3 days apart) induced hyperglycemia (blood glucose concentration 467.9 ± 47.3 mg/dL) in 10- to 12-week-old ad libitum fed Gcgr^+/+ mice when measured 8 days after the first STZ dose; age-matched ad libitum fed Gcgr^−/− mice receiving the same STZ treatment did not develop hyperglycemia (blood glucose concentration 119.5 ± 11.3 mg/dL) (Fig. 5A) despite comparable degrees of pancreatic β-cell ablation (Supplementary Fig. 4A) and similar falls in plasma insulin (Supplementary Fig. 4B). Gcgr^−/− and Gcgr^+/+ mice demonstrated similar BW loss when measured 8 days after STZ treatment (Fig. 5B). Also, as shown previously, ad libitum fed STZ-treated wild-type mice developed higher plasma acyl-ghrelin levels than vehicle-treated mice (Fig. 5C). Extending that finding, STZ further increased the already higher plasma acyl-ghrelin level in Gcgr^−/− mice (Fig. 5C). No statistically significant increases in total ghrelin were observed after STZ administration in either wild-type or Gcgr^−/− mice (Fig. 5D).

Similar to GcgR deletion, the administration of a GcgR mAb B lowers blood glucose levels in type 1 diabetes models (6). Here, we studied the effect of GcgR blockade in the hyperglycemic, functional leptin receptor–deficient db/db mouse model by administering mAb B or vehicle. mAb B corrected the hyperglycemia observed in ad libitum fed db/db mice (Fig. 6A). Plasma insulin was similar in the two groups (Supplementary Fig. 5A). mAb B increased plasma acyl-ghrelin levels about threefold (Fig. 6B). mAb B also lowered BW and food intake (Supplementary Fig. 5B–D). Pair feeding vehicle-treated mice to match the reduced food intake of mAb B–treated mice prevented the development of a higher BW (Supplementary Fig. 6C), as observed in ad libitum fed vehicle-treated mice (Supplementary Fig. 5B), but mAb B treatment nonetheless still reduced blood glucose and raised acyl-ghrelin levels in comparison (Supplementary Fig. 6). These results suggest that the altered blood glucose and plasma ghrelin levels associated with mAb B are independent of its effects to reduce feeding.

Of interest, in both the current STZ and db/db models and in previous studies using the STZ model and the NOD type 1 diabetes mouse model, GcgR genetic deletion or pharmacologic blockade with mAb B reduced glycemia to normal levels as opposed to hypoglycemic levels in ad libitum fed animals (4,6). Given the known function of raised ghrelin to prevent hypoglycemia from developing during long-term caloric restriction (12,20), we hypothesized that the increase in ghrelin herein observed in Gcgr^−/− mice serves to prevent hypoglycemia. To test that possibility, we inhibited ghrelin action in STZ-treated 8- 10-week-old Gcgr^−/− mice on a mixed C57BL/6N and C57BL/6J background by administering the GHSR antagonist [D-Lys³]-GHRP-6 (200 nmol/L/30 g BW i.p. daily for 5 successive days) (Fig. 7A), using a previously reported protocol (17). When administered [D-Lys³]-GHRP-6, STZ-treated Gcgr^−/− mice became severely hypoglycemic after a 23-h fast compared with STZ-treated Gcgr^−/− mice administered vehicle (Fig. 7B). A glucose-lowering effect also was observed in fasted Gcgr^−/− mice without prior STZ treatment when subjected to a similar [D-Lys³]-GHRP-6 injection protocol (Supplementary Fig. 7).

To complement those studies, STZ-treated Ghsr-null mice and wild-type littermates were administered GcgR mAb B or vehicle. mAb B lowered both fed (Supplementary Fig. 8A) and fasting (Fig. 7C) blood glucose levels in both wild-type and Ghsr-null mice. Importantly, just as observed in fasted STZ-treated Gcgr^−/− mice in which GHSR antagonist dropped blood glucose levels from the low normal range into the hypoglycemic range (Fig. 7B), fasted STZ-treated Ghsr-null mice also exhibited a drop in blood glucose levels from the low normal range into the hypoglycemic range after the administration of GcgR mAb B (Fig. 7C). The degree of hypoglycemia induced by mAb B was slightly greater in Ghsr-null mice than in wild-type mice (Fig. 7C). There was no effect of treatment or genotype on BW, although food intake was lower in both genotypes with mAb B treatment (Supplementary Fig. 8B and C).

The current studies were designed to further our understanding of the role ghrelin plays in mediating blood glucose in the specific environment of blocked glucagon action. This setting of blocked—or more broadly, blocked, neutralized, deficient, or absent—glucagon action is relevant in light of a proposed significant effect of unopposed glucagon action in the occurrence of both type 1 diabetes– and type 2 diabetes–related hyperglycemia and, as such, the development of potential pharmacologic agents targeting glucagon action as novel treatments for diabetes (5–8). Indeed, confirming and extending previous work, the blockade of glucagon action, as achieved experimentally using genetic deletion of GcgR or by the administration of GcgR mAb B, was shown here to normalize blood glucose levels under conditions that otherwise are associated with hyperglycemia (STZ treatment or in db/db mice, respectively) (4,6). Excitingly, we now demonstrate that under these conditions of blocked glucagon action, plasma ghrelin becomes markedly elevated and that preventing the action of this raised ghrelin level exaggerates the blood glucose–lowering effect of blocked glucagon action, resulting in marked hypoglycemia in the setting of both absent insulin and glucagon function (fasted, STZ-treated, GHSR antagonist–administered Gcgr^−/− mice and fasted, STZ-treated, GcgR mAb–administered Ghsr-null mice). Thus, these studies suggest that ghrelin rises in the setting of blocked glucagon action to prevent hypoglycemia. When viewed together with experiments demonstrating the regulation of ghrelin secretion by glucose, the regulation by ghrelin of many well-established hypoglycemia counterregulatory hormones, and an essential action of ghrelin in preventing marked life-threatening hypoglycemia during conditions mimicking a starvation state, our new results further emphasize the important role that ghrelin plays in numerous settings as an endogenous antihypoglycemic agent.

The clinical significance of these findings reaffirming the antihypoglycemic actions of ghrelin, including in the setting of GcgR blockade, are predicted to become highly apparent if GcgR antagonists or mAbs proceed to market as antidiabetic therapies. Such agents already have shown promise in preclinical and initial clinical studies (6,27,33). Although experimentally in ad libitum fed states, STZ-treated Gcgr^−/− mice maintain normoglycemia and db/db mice and NOD mice develop normoglycemia when treated with GcgR mAb (4,6) (Figs. 5–7), we would expect the blood glucose responses to therapies targeting glucagon action to be much more heterogeneous, with the hyperglycemia of some treated individuals not correcting fully. As such, we would envision the addition of agents targeting the ghrelin system, such as GHSR antagonists, as being viable, potential add-on (combined) therapies to those targeting glucagon action for the treatment of many individuals with type 1 diabetes. On the other hand, should appropriate glycemia be achieved with these GcgR antagonists, caution should be exercised not to combine this class of drugs with drugs that block GHSR function or that are expected to attenuate acyl-ghrelin secretion, such as β-blockers (20) or ghrelin-O-acyl transferase inhibitors (11). Also, if hypoglycemia is observed during GcgR blockade, we might want to test the efficacy of administering ghrelin or a ghrelin mimetic to individuals who are unable to consume simple sugars by mouth to correct the low blood glucose level, thus replacing the standard glucagon emergency kit, which presumably would not be as effective because of the anti-GcgR agent.

The current study also provides further insight regarding controls of ghrelin secretion, which is known to include several hormones (including glucagon, insulin, and glucagon-like peptide 1 [GLP-1]), nutrients (including glucose) and nutrient metabolites, and neurotransmitters (11,25). Studies reporting the influence of glucagon on ghrelin secretion have been inconsistent in the literature. Although glucagon enhanced ghrelin release from primary cultures of neonatal rat gastric mucosal cells and when infused into the arterial vasculature of an isolated rat stomach (34,35), it did not affect ghrelin release when infused directly into the gastric submucosa of rats or when applied to primary cultures of FACS-purified mouse gastric ghrelin cells from ghrelin-GFP mice or to two different immortalized ghrelinoma cell lines (26,36,37). Clinical studies have demonstrated either inhibition of ghrelin secretion and/or reduction of plasma ghrelin by glucagon (38–40) or no significant effect of glucagon administration on plasma ghrelin (41). Here, there was no direct modulation by glucagon of ghrelin secretion (Supplementary Fig. 3). Neither were plasma ghrelin levels lower in Gcgr^−/− mice or GcgR mAb-treated db/db mice, as might be predicted if a purported stimulatory signal for ghrelin release were absent or blocked. Altogether, we interpret our new results along with previous findings of the very low expression of Gcgr in FACS-purified mouse ghrelin cells and ghrelinoma cell lines (25,37) as suggesting that glucagon has a minimal direct effect on ghrelin cells to modulate ghrelin secretion. This finding is relevant to the overall interplay between ghrelin and glucagon. For instance, we previously demonstrated (13,42) that ghrelin can increase glucagon release via direct effects on pancreatic α-cells and/or via indirect effects on hypothalamic AgRP neurons, and others have suggested (43,44) that the modulatory effects of ghrelin on glucagon secretion may instead occur via engagement of GHSRs enriched on somatostatin-secreting pancreatic δ-cells. Despite the capacity of ghrelin to raise plasma glucagon levels (13,42), our current data support the conclusion that the antihypoglycemic actions of ghrelin do not require an intact glucagon system. Furthermore, although the current study does not directly address the mechanisms by which the ghrelin system prevents exaggerated falls in blood glucose in Gcgr^−/− mice, these mechanisms likely do not include stimulation of either glucagon release or food intake (because the exaggerated reduction in blood glucose induced by GHSR antagonist in STZ-treated Gcgr^−/− mice was observed in fasted conditions, in which food intake is not a factor).

Although the loss of glucagon action on ghrelin cells does not appear to be directly responsible for the increased plasma ghrelin levels observed in Gcgr^−/− mice, the current study does hint at some other potentially relevant factors. For instance, plasma ghrelin in Gcgr^−/− mice was responsive to atenolol treatment, just as in wild-type mice, suggesting the retention of sympathetic regulation of ghrelin secretion (20). There could also be a direct or indirect effect of the lowered blood glucose levels in Gcgr^−/− mice in stimulating the higher plasma ghrelin levels (Fig. 1D) (21). Indeed, primary gastric mucosal cell cultures taken from STZ-treated mice retained responsiveness to both glucose and 2-deoxy-d-glucose (Supplementary Fig. 9A and B), as observed previously in cells from non–STZ-treated mice (21). Furthermore, in the combined setting of GcgR genetic deletion plus STZ treatment, in which we found plasma ghrelin levels to be higher than in either setting alone, it is likely that the absence of insulin additionally increases plasma ghrelin levels (Fig. 5). Indeed, insulin receptor expression analyses, insulin clamp studies in humans, and secretion studies using primary cultures of gastric mucosal cells suggest that insulin has a direct inhibitory effect on ghrelin release (21,45,46) (Supplementary Fig. 9A). The latter studies also suggest that the direct inhibitory effects of insulin on ghrelin release from ghrelin cells become progressively less sensitive in higher glucose environments (21) (Supplementary Fig. 9A). Thus, it is not unexpected for ghrelin to rise after STZ treatment, which is known to destroy pancreatic β-cells and thus create a hyperglycemic, insulin-deficient environment. The higher plasma ghrelin level, however, is unlikely to be due to direct influences of either GH (known to be higher in STZ diabetic mouse models) (47) (Supplementary Fig. 9C) or STZ on ghrelin cells (Supplementary Fig. 9D).

Notably, not all methods to block, delete, or neutralize glucagon signaling prevent and/or correct hyperglycemia in individuals with diabetes. For instance, diphtheria toxin–induced pancreatic α-cell ablation did not correct hyperglycemia in STZ-treated animals, although presumably this persistent hyperglycemia resulted from the persistence of modest or near-normal glucagonemia even after near total ablation of pancreatic α-cells (9,10). Neither did the administration of glucagon mAb or GcgR antagonist to STZ-treated wild-type mice correct hyperglycemia (9). It is important to note that in the latter study the reagents targeting the glucagon system were administered 20–28 days after STZ, whereas in the current study and others in which improvements in blood glucose were observed (2,4,6,48), glucagon signaling was eliminated genetically or by immunoneutralization typically 2–4 days after STZ administration at lower blood glucose levels (∼350 mg/dL) and before the onset of ketoacidosis. Taken together, these results suggest that GcgR antagonist therapy to correct hyperglycemia in type 1 diabetes would likely be more effective when initiated before the onset of severe diabetic ketoacidosis.

Similarly, Gcgr genetic deletion is not effective in preventing hyperglycemia in all models of type 1 diabetes. In particular, genetic deletion of insulin (49) or the use of diphtheria toxin to induce near complete ablation of β-cells (50) in Gcgr^−/− mice nonetheless induced full-blown diabetes. Thus, it has been suggested that a retained low level of insulin, as occurs in the STZ type 1 diabetes model, may be permissive for the blood glucose–lowering effect observed with GcgR antagonism.

As a final point of discussion, compensatory pancreatic α-cell hyperplasia and a resulting marked elevation in plasma glucagon levels develop in Gcgr^−/− mice, mice administered GcgR antagonist (51), and humans with loss-of-function Gcgr mutations (Supplementary Fig. 4A) (4,24,33). In addition, Gcgr^−/− mice exhibit both high plasma GLP-1 and fibroblast growth factor 21 levels, both of which have been implicated in protecting STZ-treated Gcgr^−/− mice from developing hyperglycemia (3,24,48). Rodent models and humans in which GcgR mAb or GcgR antagonists have been administered also exhibit high plasma GLP-1 levels (5,27,51,52). Although GLP-1 has been shown to suppress ghrelin secretion in human subjects indirectly via stimulating insulin secretion (53), the higher ghrelin levels found within Gcgr^−/− mice (Fig. 2) despite the presence of high GLP-1 levels (24) suggests that the purported inhibitory effect of GLP-1 on ghrelin secretion is not dominant in this setting. The increase in plasma ghrelin levels may, however, contribute to the higher glucagon and GLP-1 levels observed in these mice (24) because ghrelin not only can stimulate glucagon release from α-cells (13) but also (and perhaps unexpectedly given the overall effect of ghrelin to raise blood glucose or prevent falls in blood glucose levels) can stimulate GLP-1 release from L cells (54). Furthermore, within α-cell lines, ghrelin activates extracellular signal–regulated kinase signaling (13), which in other cell systems contributes to cell proliferation. Thus, it is plausible that high ghrelin levels could directly stimulate α-cell proliferation via activation of extracellular signal–regulated kinase signaling. Contribution by ghrelin to the plasma elevation of glucagon and GLP-1 and/or to α-cell hyperplasia in Gcgr^−/− mice or in humans with loss-of-function GcgR mutations (4,24,33) are interesting possibilities that should be explored, especially in light of the proposed important roles of GLP-1 hypersecretion in mediating the improved glycemia observed in STZ-treated Gcgr^−/− mice (9).

Acknowledgments. The authors thank Jim Shi and Hai Yan at REMD Biotherapeutics, Inc., for providing the GcgR mAb. The authors also thank Prasanna Vijayaraghavan, Nathan Metzger, and Sydney Lawrence at University of Texas Southwestern Medical Center for their exceptional technical support.

Funding. This work was supported by The Hilda and Preston Davis Foundation Postdoctoral Fellowship Program in Eating Disorders Research (B.K.M.); National Institutes of Health (National Institute of Diabetes and Digestive and Kidney Diseases grants T32-DK-007307-33 and F32-DK-100085-01 to A.U., R01-DK-109408 to E.D.B., and R01-DK-103884 to J.M.Z.); an International Research Alliance with the Novo Nordisk Foundation Center for Basic Metabolic Research at the University of Copenhagen; and The Kent and Jodi Foster Distinguished Chair in Endocrinology, in Honor of Daniel Foster, M.D. (to J.M.Z.).

Duality of Interest. The authors have no potential conflict of interest relevant to this article.

Author Contributions. B.K.M. conceptualized and performed the experiments, analyzed and interpreted the data, wrote the manuscript, and secured funding. A.U. conceptualized and performed the experiments, analyzed and interpreted the data, and secured funding. Y.L. conceptualized and performed the experiments and analyzed and interpreted the data. S.O.-L. performed the experiments and analyzed the data. M.J.C. provided the Gcgr^−/− mice and critically reviewed the manuscript. R.H.U. conceptualized the experiments and critically reviewed the manuscript. E.D.B. conceptualized the experiments, critically reviewed the manuscript, secured funding, and supervised the research activity. J.M.Z. conceptualized the experiments, analyzed and interpreted the data, wrote the manuscript, secured funding, and supervised the research activity. B.K.M. and J.M.Z. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 2016 Keystone Symposia: New Therapeutics for Diabetes and Obesity, La Jolla, CA, 17–20 April 2016.