This Article Abstract Full Text (PDF) All Versions of this Article: 28/3/253    most recent Final Manuscript Author Manuscript Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Dunning, B. E. Articles by Gerich, J. E. Search for Related Content PubMed PubMed Citation Articles by Dunning, B. E. Articles by Gerich, J. E.

The Role of -Cell Dysregulation in Fasting and Postprandial Hyperglycemia in Type 2 Diabetes and Therapeutic Implications

PharmaWrite (B.E.D.), Princeton, New Jersey 08540; and University of Rochester Medical Center (J.E.G.), Rochester, New York 14642

Correspondence: Address all correspondence and requests for reprints to: John E. Gerich, M.D., Professor of Medicine and Program Director, Endocrine-Metabolism Unit, University of Rochester Medical Center, 601 Elmwood Avenue, Rochester, New York 14642. E-mail: john_gerich{at}urmc.rochester.edu

	Abstract

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
The hyperglycemic activity of pancreatic extracts was encountered some 80 yr ago during efforts to optimize methods for the purification of insulin. The hyperglycemic substance was named "glucagon," and it was subsequently determined that glucagon is a 29-amino acid peptide synthesized and released from pancreatic -cells. This article begins with a brief overview of the discovery of glucagon and the contributions that somatostatin and a sensitive and selective assay for pancreatic (vs. gut) glucagon made to understanding the physiological and pathophysiological roles of glucagon. Studies utilizing these tools to establish the function of glucagon in normal nutrient homeostasis and to document a relative glucagon excess in type 2 diabetes mellitus (T2DM) and precursors thereof are then discussed. The evidence that glucagon excess contributes to the development and maintenance of fasting hyperglycemia and that failure to suppress glucagon secretion contributes to postprandial hyperglycemia is then reviewed.

Although key human studies are emphasized, salient animal studies highlighting the importance of glucagon in normal and defective glucoregulation are also described. The past eight decades of research in this area have led to development of new therapeutic approaches to treating T2DM that have been shown to, or are expected to, improve glycemic control in patients with T2DM in part by improving -cell function or by blocking glucagon action. Accordingly, this review ends with a discussion of the status and therapeutic potential of glucagon receptor antagonists, -cell selective somatostatin agonists, glucagon-like peptide-1 agonists, and dipeptidyl peptidase-IV inhibitors. Our overall conclusions are that there is considerable evidence that relative hyperglucagonemia contributes to fasting and postprandial hyperglycemia in patients with T2DM, and there are several new and emerging pharmacotherapies that may improve glycemic control in part by ameliorating the hyperglycemic effects of this relative glucagon excess.

I. Introduction

II. Discovery of Glucagon—the Hyperglycemic, Glycogenolytic, and Gluconeogenic Hormone Secreted from Pancreatic -Cells

III. Pivotal Role of Two Investigative Tools: Glucagon Assay and Somatostatin

IV. Role of Glucagon in Normal Nutrient Homeostasis

V. Evidence of Excessive Glucagon Secretion in Type 2 Diabetes

VI. Role of Glucagon in Fasting Hyperglycemia

VII. Role of Glucagon in Postprandial Hyperglycemia

VIII. Small Animal Data Supporting Roles of Glucagon

A. Glucagon receptor knockout mice

B. Other knockout mice

C. Glucagon receptor antisense

D. Neutralizing glucagon antibodies

IX. Potential Therapeutic Approaches Targeting Glucagon Pathways

A. Somatostatin and -cell specific analogs

B. Glucagon receptor antagonists

C. Approaches leveraging the incretin system

X. Controversy

XI. Summary and Conclusions

	I. Introduction

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
SEVERAL DECADES OF research and evolving technology led to the isolation, purification, and sequencing of the 29-amino acid peptide, glucagon, to elucidation of the structure of the proglucagon gene, and to the tissue-specific processing of the proglucagon gene product. In parallel with these biochemical investigations, the physiology of pancreatic glucagon was explored, leading to the bihormonal hypothesis of Unger and Orci and to continued debate about the contribution of pancreatic -cell dysregulation to the development and maintenance of hyperglycemia in diabetes.

This is an opportune time to review this material because there are two newly approved antidiabetic agents [exenatide (Byetta) and sitagliptin (Januvia)] and several potential new therapeutic agents that may act, at least in part, by reducing glucagon levels or action. Preclinical and clinical research and, ultimately, broader clinical experience with such new drugs may be expected to ascertain the importance of -cell dysregulation in fasting and postprandial hyperglycemia.

	II. Discovery of Glucagon—the Hyperglycemic, Glycogenolytic, and Gluconeogenic Hormone Secreted from Pancreatic -Cells

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
Shortly after the discovery of insulin in 1921 by Banting and Best (1), Kimball and Murlin noted that iv administration of a crude insulin preparation elicited transient hyperglycemia (2). They suggested that this action might be due to a second substance, subsequently named "glucagon." A decade later, Burger and Brandt (3) prepared a partially purified material with glucagon activity. However, it was not until 1955 that Staub et al. (4) fully purified and crystallized glucagon, and in 1957, the amino acid sequence of glucagon was reported (5). Although glucagon preparations were known to be useful in treating "insulin reactions" (6) and the ability of repeated injections of glucagon to cause impaired glucose tolerance (IGT) and glycosuria in man had been described (7), an understanding of the physiological role of glucagon was hampered by difficulties in the measurement of plasma levels of pancreatic glucagon in man.

	III. Pivotal Role of Two Investigative Tools: Glucagon Assay and Somatostatin

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
The development of the insulin RIA introduced an unprecedented degree of specificity and sensitivity for the measurement of insulin (8) and was soon followed by the development of an RIA for glucagon (9). However, for several reasons, the development of a method with sufficient sensitivity and specificity to measure the concentration of glucagon in plasma proved considerably more difficult than for insulin. Measurement of pancreatic glucagon was initially problematic because circulating levels are relatively low, and the sensitivity of early assays was inadequate. Furthermore, the existence of glucagon-like molecules derived from the gut (10, 11, 12), often circulating at much higher concentrations than pancreatic glucagon and cross-reacting with early antisera, mandated a selective assay. Moreover, glucagon is rapidly degraded in plasma, and degradation of endogenous glucagon, as well as degradation of the glucagon tracer, frequently led to spurious results (13). In addition, plasma proteins can interfere with antibody binding (14). The generation of high-titer antibodies selective for pancreatic (vs. gut) glucagon, particularly the one that came to be known as "30K" (15), and assay modifications that averted proteolytic degradation of glucagon (e.g., the use of aprotinin and other protease inhibitors) and plasma interference (e.g., extraction of plasma samples before assay, or use of dextran-coated charcoal for separation of bound and free tracer) (15) eventually enabled more accurate measurement of plasma levels of pancreatic glucagon.

However, even today, pancreatic-selective assays yield a rather wide range of basal glucagon values, e.g., 20–100 ng/liter (5.7–28.6 pmol/liter) in healthy subjects. This highlights the importance of comparing values only within a single study and using a single assay method. Because many pancreatic-selective assays recognize a large molecular weight interference factor as glucagon (this is unrelated to glucagon and remains constant) (16), change from baseline provides a more reliable assessment of glucagon secretion in vivo than basal fasting glucagon levels.

The second advance that opened the door to understanding the physiological function of glucagon was the discovery of somatostatin. Somatostatin was first isolated from extracts of ovine hypothalami based on its ability to inhibit GH release from dispersed rat anterior pituitary cells in culture (17). Shortly thereafter, it was noted that infusion of somatostatin in overnight-fasted baboons lowered plasma glucose levels—an effect attributed to inhibition of basal insulin and glucagon secretion (18). These effects were subsequently confirmed in humans (19). By virtue of its ability to profoundly suppress glucagon secretion, somatostatin became a critically important tool that facilitated understanding of the physiological and pathophysiological roles of glucagon in nutrient homeostasis (20).

Although by definition glucagon is a substance that raises glucose levels, understanding of the physiological hormonal role of glucagon in normal metabolic regulation and in the pathophysiology of type 2 diabetes mellitus (T2DM) remained a matter of speculation for many years until the development of the investigative tools discussed above. With a sensitive and selective assay such as that employing the 30K antiserum, it became possible to assess more accurately the physiological range of plasma levels of pancreatic glucagon, to elucidate factors important in the normal control of glucagon secretion, and to evaluate how the regulation of glucagon secretion was abnormal in patients with T2DM, and thereby evaluate the potential contribution of defective -cell function to the metabolic dysregulation characteristic of T2DM. The discovery of somatostatin and its ability to completely suppress both glucagon and insulin secretion then allowed investigators to manipulate plasma levels of insulin and glucagon by suppressing endogenous hormone release then infusing exogenous insulin and glucagon.

	IV. Role of Glucagon in Normal Nutrient Homeostasis

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
Shortly after the development of a sensitive assay for pancreatic glucagon, it was demonstrated that insulin- or phloridzin-induced hypoglycemia was a potent stimulus for glucagon release and that hyperglycemia rapidly decreased plasma glucagon levels in pancreatic venous blood in dogs (21). Such findings, together with the demonstration that starvation greatly increased peripheral venous glucagon levels in healthy human subjects (22), led to the concept that the islets of Langerhans constituted a "bihormonal organ" (23, 24, 25), with insulin being the hormone of glucose abundance, promoting glucose storage in fat, muscle, and liver; and glucagon being the hormone of glucose need, acting to increase hepatic glucose output to maintain the glucose supply to the brain (26). It was also shown in normal subjects that infusion of amino acids (27, 28), particularly arginine (29), and large protein meals (30) increased plasma glucagon levels, and it was suggested that because these stimuli also increase insulin secretion, amino acid-induced glucagon secretion serves to prevent postmeal hypoglycemia (31). The effects of glucagon on hepatic glucose output reflect its stimulation of both gluconeogenesis (32) and glycogenolysis (33). The relative contributions of these processes to glucagon-stimulated glucagon production no doubt vary depending on many factors, including hepatic glycogen content and the availability of gluconeogenic precursors. However, in the overnight-fasted dog, it has been shown that the dose-response curves for glucagon-stimulated glycogenolysis and gluconeogenesis are parallel (34). In the context of glucagon’s effects on gluconeogenesis vs. glycogenolysis, it should be noted that because glucagon cannot increase the supply of gluconeogenic precursors (by affecting muscle or fat), substrate availability may limit the influence of glucagon on gluconeogenesis. Thus, the acute effect of glucagon on plasma glucose levels reflects primarily its glycogenolytic action (35). In contrast, when substrate availability increases, as during meal feeding or during counterregulation of hypoglycemia and in uncontrolled diabetes, the effect of glucagon may be primarily on gluconeogenesis (36, 37).

In healthy human subjects, infusion of glucagon at rates that resulted in an incremental increase in plasma glucagon similar to those seen after a protein meal (31, 38, 39), during starvation (22, 40), or with prolonged exercise (41) elicited a prompt increase in splanchnic glucose output (42). The effects of glucagon infusion in overnight-fasted healthy subjects on arterial glucagon, insulin, and glucose levels and on splanchnic glucose output are shown in Fig. 1. During the infusion of glucagon (3 ng/kg·min), plasma glucagon levels rose rapidly, reached a plateau within 15 min, and remained elevated for 45 min (Fig. 1A). Arterial glucose and insulin also increased rapidly and reached a maximum incremental increase by 30 min (Fig. 1B). Despite the concurrent increase in plasma insulin levels, splanchnic glucose output rose rapidly during the glucagon infusion, reaching levels two to three times basal within 7.5 min, remained elevated for approximately 15 min, and then fell progressively (Fig. 1C). Thus, it was established that high physiological levels of glucagon could increase glucose production in normal subjects, despite a concomitant increase of plasma insulin. In a second experiment, these investigators infused glucose at a rate sufficient to completely suppress glucose production before the glucagon infusion. Under these conditions, glucagon infusion reversed the glucose- and insulin-mediated suppression of splanchnic glucose production. It may be noted, however, that under both conditions in the healthy volunteers, the glucagon-stimulated increase of glucose production was transient (42).

View larger version (12K): [in this window] [in a new window]   FIG. 1. Effect of exogenous glucagon on splanchnic glucose output in healthy volunteers. Arterial plasma glucagon (A), arterial plasma glucose and insulin (B), and splanchnic glucose output (C) in the basal state and during infusion of glucagon (3 ng/kg·min). [Data are from Table 1 in Ref. 42 .]

The physiological role of glucagon in the maintenance of postabsorptive glucose levels was clearly demonstrated by Cherrington et al. (48) in normal dogs using both isotope dilution and arteriovenous difference techniques. In overnight-fasted, anesthetized dogs, somatostatin was infused at a rate sufficient to nearly completely suppress both insulin and glucagon secretion for 70 min. In a protocol during which neither glucagon nor insulin was replaced, glucose production decreased by 40% and plasma glucose levels fell by approximately 30 mg/dl. In a protocol during which insulin was infused via the portal vein at a rate to maintain peripheral insulin levels constant, net hepatic glucose production (HGP) decreased by 35% and plasma glucose levels decreased by approximately 25 mg/dl. In a protocol during which glucagon was infused at a rate to maintain peripheral glucagon levels, glucose production increased by 52%, and plasma glucose levels increased by approximately 50 mg/dl. It was concluded that in anesthetized dogs, basal glucagon is responsible for at least one third of basal glucose production and that basal insulin prevents the increased glucose production that would result from unrestrained glucagon action. Furthermore, it was confirmed that somatostatin has no acute effects on glucose turnover other than those it induces via changes in pancreatic hormone secretion.

Soon after this important contribution, a similar study was conducted in healthy humans. Thus, the influence of endogenous glucagon on glucose production was assessed in healthy subjects by producing a selective glucagon deficiency via somatostatin infusion with basal insulin replacement (49). Exogenous glucose was infused to maintain euglycemia. Figure 2 depicts splanchnic glucose production (Fig. 2A); plasma glucagon, insulin, and glucose levels (Fig. 2B); and the glucose infusion rate necessary to maintain euglycemia (Fig. 2C) in healthy subjects during infusion of somatostatin with replacement of basal insulin. It is apparent that selective glucagon deficiency in healthy subjects produces a marked and sustained decrease in glucose production. In a second experiment (data not shown), basal insulin was not replaced, and there was only a transient reduction in net splanchnic glucose production. After approximately 1 h, exogenous glucose infusion was no longer necessary to maintain euglycemia. From this study, it was concluded that basal glucagon plays a physiological role in the maintenance of basal glucose production in normal man and during glucagon deficiency, hepatic glucose output is very sensitive to low levels of insulin. Furthermore, it suggested that glucagon may play a part in the hyperglycemia of diabetes because of accumulating evidence of hypersecretion of glucagon (relative to prevailing glucose levels) in diabetic patients and because in such patients glucagon secretion did not appear to be suppressible by hyperglycemia or exogenous insulin (see below). This was consistent with Unger and Orci’s (25) controversial assertion that glucagon has an essential role in the pathogenesis of diabetes.

View larger version (13K): [in this window] [in a new window]   FIG. 2. Effect of selective glucagon deficiency on splanchnic glucose output in healthy volunteers. Effects of iv infusion of somatostatin (0.9 mg/h) and insulin (150 µU/kg·min) on net splanchnic glucose production (A); arterial plasma glucose, insulin, and glucagon (B); and glucose infusion rate (GIR) necessary to maintain euglycemia (C) in four healthy subjects. [Data were replotted from Figs. 1 and 2 in Ref. 49 .]

Another study in dogs by Cherrington et al. (52) underscored the importance of glucagon in regulation of basal glucose production and demonstrated that prolonged (4-h) selective glucagon deficiency again produced by somatostatin infusion with basal insulin replacement resulted in a sustained underproduction of glucose. When hypoglycemia was allowed to occur, this itself stimulated glucose production and limited glucose utilization. But when glucose was infused to prevent hypoglycemia, glucose production decreased by 67% (52). Thus, in the dog as in humans, glucagon largely is responsible for maintaining euglycemia. Interestingly, there has been at least one case report of an individual with a (presumably genetic) selective glucagon deficiency who suffered from severe hypoglycemic attacks (53). This patient was treated with a continuous sc glucagon infusion that normalized both postabsorptive and postprandial glucose levels.

Many small-animal studies using immunoneutralization of endogenous glucagon, antisense oligonucleotides, and transgenic approaches to eliminate glucagon action also support an important role for glucagon in normal glycemic regulation. These are discussed in Section VIII. The bulk of evidence from human and large-animal studies with somatostatin suggests that pancreatic glucagon serves a physiological role in the regulation of hepatic glucose balance.

	V. Evidence of Excessive Glucagon Secretion in Type 2 Diabetes

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
Plasma glucagon levels are usually found to be greatly elevated in diabetic ketoacidosis (54, 55) and significantly increased in patients with poorly controlled type 1 diabetes mellitus (T1DM) (56, 57). As will be discussed below, in patients with T2DM, absolute fasting plasma levels of glucagon may (57, 58, 59, 60, 61, 62, 63, 64, 65) or may not (29, 36, 66, 67) be found to be increased (to a statistically significant degree) in patients with T2DM compared with nondiabetic subjects. However, what is clear is that in T2DM the plasma glucagon levels are inappropriate in the context of hyperglycemia and hyperinsulinemia (both of which normally suppress glucagon secretion) (68, 69, 70), and that this contributes to the increased rate of hepatic glucose output characteristic of patients with T2DM (61, 62, 63).

Shortly after the development of a pancreatic-selective glucagon assay, it was reported that fasting glucagon levels were similar in healthy volunteers and in patients with "presumably genetic diabetes" (including both those with T1DM and those with T2DM) (71). However, because the diabetic subjects had fasting hyperglycemia and hyperglycemia failed to suppress the increase of glucagon in response to arginine infusion, the authors suggested that the -cell was hyposuppressible in diabetic patients.

Many studies have since assessed -cell function in patients with T2DM (72). Findings from the Unger laboratory clearly illustrated relative glucagon hypersecretion in the fasted state, during a large carbohydrate meal, during iv glucose or arginine administration, or during protein meals (29, 30, 64, 66). As shown in Fig. 3, the suppression of glucagon after the carbohydrate stimulus apparent in healthy volunteers did not occur in patients with T2DM (30). In contrast to the significant decline in plasma glucagon exhibited by the nondiabetic subjects, no significant fall in plasma glucagon was observed in the subjects with T2DM (Fig. 3A). The significant rise in plasma insulin exhibited by the nondiabetic subjects was blunted in the diabetic subjects, indicative of impaired glucose sensing in the ß-cell (Fig. 3B). In the diabetic subjects, plasma glucagon remained at or above the preprandial level (Fig. 3A), despite the marked fasting and postprandial hyperglycemia (Fig. 3C).

View larger version (11K): [in this window] [in a new window]   FIG. 3. Plasma glucagon, insulin, and glucose levels in response to a large carbohydrate meal in subjects with NGT and in patients with T2DM. Plasma glucagon (A), insulin (B), and glucose (C) in 14 subjects with NGT and 12 patients with T2DM during ingestion of a high-carbohydrate meal. Mean ± SEM. [Data were replotted from Ref. 30 .]

View larger version (10K): [in this window] [in a new window]   FIG. 4. Plasma glucagon, insulin, and glucose levels during iv infusion of arginine in subjects with NGT and in patients with T2DM. Plasma glucagon (A), insulin (B), and glucose (C) in 28 subjects with NGT and 33 patients with T2DM. Arginine was infused at the rate of 11.7 mg/kg·min for 40 min. Mean ± SEM. [Data were replotted from Ref. 29 .]

Many subsequent studies that examined glucagon levels in patients with T2DM have confirmed these original findings and have identified several characteristic defects. These include an inadequate suppression of glucagon in response to oral (67) or iv glucose (66). In fact, there may be a paradoxical increase of glucagon in response to glucose in patients with advanced disease and severe hyperglycemia (67, 78). It has also been suggested that the -cell is resistant to insulin in patients with T2DM (60). Interestingly, insulin signaling does regulate glucagon release in vitro (79, 80), thus it is possible that a defect(s) in insulin action at the level of the -cell could contribute to its dysfunctional responses (81). However, it is difficult to distinguish between impaired glucose sensing and -cell insulin resistance.

Loss of the ability of hypoglycemia to stimulate glucagon release is a well-documented problem in patients with T1DM (82). Whether patients with T2DM have an impaired glucagon response to hypoglycemia is more controversial (83), although most investigators find a subnormal response at least in patients with longstanding T2DM (84, 85, 86, 87).

In sum, the defects in -cell function seen in patients with T2DM may be considered to reflect an impairment of glucose sensing. Because local insulin is a key regulator of glucagon secretion (69, 88) and defective ß-cell glucose sensing in T2DM is indisputable (89, 90), many (91), if not all (92), of the characteristic defects in -cell function may be secondary to ß-cell dysfunction. Indeed, there is increasing evidence that impaired insulin secretion mediates both the impaired postprandial suppression of glucagon release (93) and the impaired glucagon response to hypoglycemia (94, 95); however, whether all abnormalities of -cell function in T2DM are secondary to ß-cell dysfunction remains controversial. This debate is beyond the intended scope of this review and will not be addressed further.

Impaired glucose-mediated suppression of glucagon release in response to arginine has also been observed in nondiabetic subjects with IGT (96), and elevated glucagon levels have been noted in some studies of first-degree relatives of persons with T2DM (97, 98), further supporting the concept that relative hyperglucagonemia may be involved in the etiology of T2DM.

	VI. Role of Glucagon in Fasting Hyperglycemia

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
In patients with T2DM, the degree of fasting hyperglycemia is significantly correlated with the rate of basal HGP, suggesting that this defect is an important overall determinant of the severity of the disease (62, 99, 100, 101, 102, 103). The increase in HGP seen in patients with overt T2DM primarily reflects increased gluconeogenesis (104, 105, 106), although impaired insulin-mediated suppression of glucose production has also been observed in patients with "mild" hyperglycemia (63). As previously discussed, exogenous administration of glucagon stimulates HGP through stimulation of both glycogenolysis (33) and gluconeogenesis. Of course, numerous factors other than, or in addition to, a relative excess of glucagon likely contribute to the increased glucose production and, accordingly, fasting hyperglycemia seen in T2DM. These include increased circulating levels of gluconeogenic precursors, increased fatty acid oxidation, enhanced sensitivity to glucagon, decreased sensitivity to insulin, and impaired hepatic autoregulation. Here, we focus on glucagon.

Although there has been much controversy and debate on the topic, there is substantial evidence that hyperglucagonemia plays a role in the development and maintenance of fasting hyperglycemia. Early evidence suggesting that glucagon contributes to the development of fasting hyperglycemia was provided by a study of patients with T1DM in whom euglycemia had been maintained via insulin infusion; then, upon discontinuation of insulin, somatostatin or saline was infused (Fig. 5) (107). Suppression of glucagon secretion (by somatostatin infusion) prevented the development of diabetic ketoacidosis and the marked fasting hyperglycemia during insulin withdrawal (Fig. 5, A–C). Accordingly, it was concluded that hypoinsulinemia per se does not lead to fasting hyperglycemia and ketoacidosis. Rather, glucagon, by means of its glycogneolytic, gluconeogenic, ketogenic, and lipolytic actions, is necessary for the full development of this condition.

View larger version (15K): [in this window] [in a new window]   FIG. 5. Effect of glucagon suppression on plasma glucose, ß-hydroxy butyrate, and glucagon in patients with T1DM during withdrawal of insulin and infusion of somatostatin (SRIF, 500 µg/h) or saline. Plasma glucose (A), ß-OH butyrate (B), and glucagon (C) during infusion of somatostatin (500 µg/h) or saline in seven patients with T1DM. Before somatostatin or saline infusion, insulin was infused for 14 h at an average rate of 1 U/h and terminated at time 0. Mean ± SEM. [Data were replotted from Ref. 107 .]

View larger version (13K): [in this window] [in a new window]   FIG. 6. Effect of glucagon suppression on plasma glucose, ß-hydroxy butyrate, and glucagon in patients with T1DM after withdrawal of insulin and infusion of somatostatin (SRIF, 500 µg/h) or saline. Plasma glucose (A), ß-OH butyrate (B), and glucagon (C) after withdrawal of insulin followed by infusion of somatostatin (500 µg/h) or saline in nine patients with T1DM. Before somatostatin or saline infusion, insulin was infused for 14 h at an average rate of 1 U/h and terminated at time 0. Somatostatin was then infused from time 6 to 12 h. Mean ± SEM. [Data were replotted from Ref. 108 .]

Findings from Efendic’s laboratory (110) demonstrated that the sustained hyperglycemic effect of glucagon replacement during 6-h somatostatin infusion does not require absolute insulin deficiency. It was shown that in healthy volunteers during a low-dose somatostatin infusion (50 µg/h, which decreased insulin by 5 µU/ml), infusion of glucagon at a rate of 0.5 ng/kg·min (which replaced basal glucagon levels) produced an early increase in glucose production and an increase in plasma glucose of approximately 2 mM that was maintained throughout the 6-h study period.

Ward et al. (111) examined the effect of prolonged (46-h) somatostatin infusion with basal glucagon replacement in healthy volunteers. In this study, prolonged moderate isolated insulin deficiency (decreasing plasma insulin levels from 8 to 6 µU/ml) led to significant approximately 20% increases in glucose production and fasting plasma glucose (FPG). These studies demonstrate not only that in nondiabetic subjects basal insulin restrains HGP, but also that a relative excess of glucagon leads to a prolonged increase in glucose production and mild fasting hyperglycemia.

In contrast to the above cited studies (62, 107, 108, 110, 111), a series of papers from the Felig group appeared to challenge the importance of glucagon in causing or maintaining fasting hyperglycemia (42, 44, 45, 112, 113). In one study (42), infusion of glucagon into normal volunteers, whose splanchnic glucose production was being suppressed by an infusion of glucose, produced an increase in splanchnic glucose production above baseline, which returned to baseline within 45 min. These observations were interpreted to indicate a loss of responsiveness of the liver to glucagon. However, this apparently transient effect of glucagon could have been accounted for by the suppressive effects of hyperinsulinemia and hyperglycemia on HGP—an effect that obviously does not occur in T2DM. In two related papers (44, 113), somatostatin was infused either alone or with glucagon into normal volunteers. When somatostatin alone was infused, plasma glucose levels initially decreased from 90 to 75 mg/dl and subsequently increased to 135 mg/dl at 5 h. These observations were interpreted to indicate that basal glucagon secretion was not essential for the development of fasting hyperglycemia and that hyperglycemia eventually develops as a consequence of insulin deficiency. Whether or not glucagon is essential cannot be answered by these experiments because glucagon was only partially suppressed. Regardless, relative to somatostatin infusion with glucagon replacement (during which glucose levels increased to >200 mg/dl), the degree of hyperglycemia was markedly attenuated (113).

In the studies in which glucagon was infused to replace basal levels during somatostatin infusion, plasma glucose increased much more than when somatostatin alone was infused (200 mg/dl vs. 100 mg/dl at 2 h, respectively), but by 5 h the plasma glucose levels were comparable under both experimental conditions. The main conclusion drawn by the investigators—that insulin deficiency is the predominant factor responsible for the hyperglycemia observed during infusion of somatostatin—in no way denies the importance of glucagon in contributing to the fasting hyperglycemia seen in T1DM or T2DM. In the normal subjects studied, the influence of hyperglycemia and the higher plasma insulin levels probably diminished the effects of glucagon on HGP.

Finally, in studies in which glucagon was infused into normal volunteers and in patients with T1DM or T2DM (112), the glucagon infusion caused deterioration in glycemia only in the patients with T1DM. It was concluded that hyperglucagonemia does not cause glucose intolerance in normal subjects or bring about deterioration of glycemic control when insulin is available. The first conclusion is probably correct, because individuals with normal ß-cell function can increase their insulin secretion to overcome the actions of glucagon. The three poorly controlled T2DM patients included in that study (plasma glucose 300 mg/dl) may not have had further deterioration in glycemic control because they were already responding maximally to their (inappropriately high) level of endogenous glucagon. The second conclusion is problematic; it depends on how much insulin is available. Studies by Gerich et al. (114, 115) showed that suppression of glucagon secretion during a 3-d infusion of somatostatin in insulin-treated patients resulted in markedly improved glycemic control and permitted a 50% reduction in insulin dose.

Although it is clear that ß-cell dysfunction is essential to the development and maintenance of fasting hyperglycemia, as mentioned previously, net splanchnic glucose production is influenced by many factors, including the absolute concentrations of insulin and glucagon, the prevailing glucose level, and substrate supply. In the context of a relative insulin deficiency (as occurs in T2DM), the preponderance of evidence suggests that a relative excess of glucagon also makes a significant contribution to fasting hyperglycemia.

It is difficult, perhaps impossible, to ascertain the relative contributions of - vs. ß-cell dysfunction to the development and maintenance of fasting hyperglycemia because, just as plasma insulin levels must be considered in the context of plasma glucose levels, glucagon levels must be considered relative to the prevailing levels of both glucose and insulin. For this reason, many investigators have invoked "the insulin to glucagon ratio" to stress the importance of both hormones (116, 117, 118). Although this may be a useful concept, it should not be considered in any quantitative framework because the dose-response relationships for the effects of insulin and glucagon on endogenous glucose production are not strictly antiparallel. Moreover, due to the influence of local insulin on glucagon secretion (69, 88, 95), defects in ß-cell function are essentially always accompanied by inappropriately high plasma glucagon levels.

	VII. Role of Glucagon in Postprandial Hyperglycemia

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
The contribution of a relative excess of pancreatic glucagon secretion to postprandial hyperglycemia in both T2DM (119) and IGT (117) can be deduced from studies by Mitrakou et al. (117, 119), who examined glucose and pancreatic hormone levels and glucose production and utilization using a dual isotope tracer technique during oral glucose tolerance tests. In the former study (119), as shown in Fig. 7 relative to subjects with NGT, patients with T2DM had impaired early insulin release (Fig. 7B), fasting hyperglucagonemia, and a complete lack of glucose-induced suppression of plasma glucagon levels (Fig. 7C), accompanied by fasting hyperglycemia and marked postload hyperglycemia (Fig. 7A). As shown in Fig. 7D, in parallel with the plasma glucagon profiles, in patients with T2DM, the rate of endogenous glucose production was significantly higher in the basal (fasting) state, and there was a pronounced lack of suppression after the oral glucose load. The rate of total glucose disappearance was significantly higher in patients with T2DM than in the subjects with NGT; however, urinary glucose loss in the diabetic subjects accounted for this. The rate of muscle glucose uptake was no different in the two groups of subjects. These findings suggest that inadequate suppression of endogenous glucose production, reflecting inadequate suppression of plasma glucagon levels and impaired early insulin release, is primarily responsible for postprandial hyperglycemia in patients with T2DM. Woerle et al. (36) have recently presented evidence that at least part of the increase in postprandial glucose release is due to increased glycogenolysis and glycogen cycling. Impaired early insulin release, together with excessive glucagon secretion, also contributes to a modest but significant reduction in splanchnic sequestration of oral glucose (i.e., hepatic glucose uptake) in patients with T2DM (119).

View larger version (14K): [in this window] [in a new window]   FIG. 7. Plasma glucose, insulin, glucagon, and the rate of endogenous glucose production (Ra) during oral glucose tolerance test in subjects with NGT or T2DM. Plasma glucose (A), insulin (B), glucagon (C), and endogenous glucose production (D) during oral glucose tolerance tests in subjects with NGT (n = 10) or T2DM (n = 10). Mean ± SEM. [Data were replotted from Ref. 119 .]

View larger version (14K): [in this window] [in a new window]   FIG. 8. Plasma glucose, insulin, glucagon, and the rate of endogenous glucose production (Ra) during oral glucose tolerance test in subjects with NGT or IGT. Plasma glucose (A), insulin (B), glucagon (C), and endogenous glucose production (D) during oral glucose tolerance tests in subjects with NGT (n = 9) or IGT (n = 8). Mean ± SEM. [Data were replotted from Ref. 117 .]

View larger version (10K): [in this window] [in a new window]   FIG. 9. Correlations of systemic glucose appearance (Ra) with peak postload glucose (PPG) levels or molar ratio of plasma insulin to glucagon during oral glucose tolerance tests in subjects with NGT or IGT. Correlation of systemic glucose appearance (Ra) and peak postload plasma glucose (A) and correlation of Ra with the molar ratio of plasma insulin to glucose (B) during oral glucose tolerance test in obese (squares) and nonobese (circles) subjects with NGT (open symbols) or IGT (closed symbols). [Data were replotted from Ref. 117 .]

View larger version (14K): [in this window] [in a new window]   FIG. 10. Effect of selective glucagon deficiency on splanchnic glucose balance during hyperglycemic clamps in healthy volunteers. Plasma insulin (A), glucagon (B), glucose (C), and splanchnic glucose balance (D) during infusion of somatostatin (0.9 mg/h) and insulin (0.15 µU/kg·min) with (closed triangles, n = 3) or without (open circles, n = 4) glucagon infusion (1.5 ng/kg·min) to approximate basal portal glucagon levels. Exogenous glucose was infused at a variable rate to achieve similar levels of hyperglycemia. Mean ± SEM. [Data were replotted from Ref. 121 .]

View larger version (17K): [in this window] [in a new window]   FIG. 11. Effect of lack of suppression of glucagon on glucose tolerance in the presence of relative insulin deficiency. Plasma glucagon (A) and glucose (B) during 50-g oral glucose tolerance tests in patients with T2DM during somatostatin infusion (4.3 nmol/kg·min) and variable-rate insulin infusion to reproduce a diabetic insulin profile (253 ), and glucagon infusion (1.25 ng/kg·min) to maintain basal portal glucagon levels (nonsuppressed, closed triangles) or delayed glucagon infusion (suppressed, open circles) in nine patients with T2DM. Mean ± SEM. [Data were replotted from Ref. 125 .]

In sum, there is considerable evidence that inadequate suppression of glucagon secretion contributes to excessive postload glucose excursions in T2DM and in IGT (36, 102, 128, 129, 130, 131). It is important to consider glucagon levels relative to insulin levels, and it is usually observed that relative insulin deficiency and relative glucagon excess occur simultaneously, likely reflecting the important role that local, intraislet insulin has in the regulation of glucagon secretion.

	VIII. Small Animal Data Supporting Roles of Glucagon

Top Abstract I. Introduction II. Discovery of Glucagon--the... III. Pivotal Role of... IV. Role of Glucagon... V. Evidence of Excessive... VI. Role of Glucagon... VII. Role of Glucagon... VIII. Small Animal Data... IX. Potential Therapeutic... X. Controversy XI. Summary and Conclusions References

 
Transgenic mice represent useful models that offer an opportunity to understand the role of glucagon in the chronic regulation of glycemic control in vivo. The following section discusses the reported phenotypes of the most relevant transgenic mouse models related to glucagon action and glycemic control. This information is summarized in Table 1.

A similar phenotype was reported by another group that evaluated a different strain of glucagon receptor gene (GR^–/–) knockout mice (133). As shown in Fig. 12, fasting and fed glucose levels were reduced in GR^–/– mice compared with control littermates (Fig. 12A), as were glucose levels throughout the day (data not shown). Plasma glucagon was markedly elevated under both fed and fasting conditions (Fig. 12B). GR^–/– mice exhibited improved glucose tolerance (Fig. 12C) and increased insulin sensitivity during an insulin tolerance test (data not shown), but serum insulin levels were unchanged compared with control littermates (data not shown). Male GR^–/– experienced frank hypoglycemia during an insulin tolerance test, and severe hypoglycemia (blood glucose = 1.7 mmol/liter) during a 24-h fast. Thus, in the absence of functional glucagon receptors, mice have significantly reduced glucose levels throughout the day, increased apparent insulin sensitivity, and a propensity to develop overt hypoglycemia, consistent with the concept that glucagon plays a physiological role in determining both fasting and postload glucose levels. These findings also highlight the importance of glucagon in the prevention and correction of hypoglycemia (134).

View larger version (9K): [in this window] [in a new window]   FIG. 12. Serum glucose and glucagon levels in fasted and fed state and blood glucose levels during ip glucose tolerance test in glucagon receptor knockout mice (GR–/–) and nontransgenic littermates (GR+/+). Serum levels of glucose (A) and glucagon (B) in fasted or fed state and blood glucose during ip glucose tolerance tests in glucagon receptor knockout mice (GR–/–) and nontransgenic littermates (GR+/+). Mean ± SEM, n = 4–7 mice/group. [Data were replotted from Ref. 133 .]

In addition to knockout mice that exhibit elevated circulating levels of glucagon, several strains of genetically modified mice have been reported to have low glucagon levels and relative or absolute hypoglycemia, although, of course, this association in no way proves causality. As discussed previously, glucagon is derived from proglucagon, an 18-kDa protein that contains glucagon along with GLP-1 and GLP-2. Prohormone convertase, SPC2 (also referred to as PC2), is the enzyme responsible for processing proglucagon to the active glucagon in pancreatic -cells; a related convertase, PC3, processes proglucagon to GLP-1 in the intestinal L cells (136, 137, 138). This enzyme is also responsible for processing of proinsulin, prosomatostatin, and a variety of other neuroendocrine precursors. Homozygous mice lacking the mSPC2 (mSPC2^–/–) gene have been generated and characterized (139). These mice appear normal at birth, but exhibit a small decrease in growth rate compared with normal controls. Not surprisingly, these mice had undetectable circulating glucagon due to their inability to process proglucagon. Fasting blood glucose was significantly lower in mSPC2^–/– mice (by 40% relative to wild-type mice), and they had improved glucose tolerance. The mSPC2^–/– mice exhibited pancreatic -cell hyperplasia, probably secondary to chronic hypoglucagonemia and/or chronic hypoglycemia. Although these results are consistent with a role for glucagon in glycemic regulation, the overall conclusions are complicated by the fact that these mice also exhibited a marked decrease (33%) in their ability to process proinsulin, resulting in elevated proinsulin-to-insulin ratios in the pancreas and serum, but low circulating insulin levels (140). Furthermore, a dose of glucagon (delivered by mini-osmotic pump) sufficient to raise plasma levels 20-fold compared with wild-type mice was required to correct the hypoglycemia and -cell hyperplasia in mSPC2^–/– mice (141). The authors suggested that a reduction in glucagon receptor number or responsiveness or high levels of inactive forms of glucagon might compete for receptor occupancy and may explain the need for glucagon overreplacement. These plausible and testable hypotheses merit investigation.

The protein termed 7B2, a small acidic protein localized exclusively to neuroendocrine tissues, is a SPC2-binding protein that facilitates SPC2 processing and is essential for SPC2 enzymatic activity (142). At 4–5 wk of age, 7B2 null mice (m7B2^–/–) are severely runted, lack SPC2 activity, have undetectable circulating glucagon, and are hypoglycemic compared with normal littermates. ß-Cell mass and proinsulin levels are significantly higher in m7B2^–/– vs. wild-type mice. These mice also suffer from a wide range of metabolic abnormalities, presumably due to their inability to process a variety of peptide hormones, and die before 9 wk of severe Cushing’s syndrome.

Gastrin is a peptide hormone that stimulates the secretion of gastric acid and epithelial cell proliferation. There is increasing evidence that this gastric peptide regulates islet function and hormone content, including stimulating glucagon secretion, along with stimulating the regeneration of ß-cells. To define the physiological role of gastrin further, mice with a targeted disruption of the gastrin gene (mGastrin^–/–) were generated (143). The mGastrin^–/– mice exhibited fasting hypoglycemia and improved glucose tolerance. The levels of circulating insulin and insulin mRNA transcripts were reduced, but insulin sensitivity, as assessed by the euglycemic-hyperinsulinemic clamp, was normal in mGastrin^–/– mice. Despite reduced fasting glucose levels, the levels of circulating glucagon and pancreatic glucagon mRNA transcripts were similar in mGastrin^–/– mice compared with control littermates, indicating relative hypoglucagonemia. The observations of hypoglycemia in the presence of relative hypoglucagonemia are again consistent with a role for glucagon in determining plasma glucose levels.

The genes encoding hepatocyte nuclear factor 3 (HNF3) proteins play a pivotal role in the regulation of metabolism and in the differentiation of metabolic tissues such as the pancreas and liver (144, 145). HNF3 transcription factors bind to cis-regulatory elements in hundreds of genes encoding gluconeogenic and glycolytic enzymes, serum proteins, and hormones. Genetic analysis in mice has shown that HNF3 is required for normal expression of glucagon in the pancreas, whereas HNF3 induces the activation of gluconeogenic enzymes to prevent hypoglycemia during fasting. Mice with a targeted disruption of HNF3 (mHNF^–/–) exhibited severe postnatal growth retardation followed by death between postnatal days 2 and 12 (146). These mice were hypoglycemic despi