|Home | About | Journals | Submit | Contact Us | Français|
Glucagon is a primary regulator of hepatic glucose production (HGP) in vivo during fasting, exercise and hypoglycaemia. Glucagon also plays a role in limiting hepatic glucose uptake and producing the hyperglycaemic phenotype associated with insulin deficiency and insulin resistance. In response to a physiological rise in glucagon, HGP is rapidly stimulated. This increase in HGP is entirely attributable to an enhancement of glycogenolysis, with little to no acute effect on gluconeogenesis. This dramatic rise in glycogenolysis in response to hyperglucagonemia wanes with time. A component of this waning effect is known to be independent of hyperglycemia, though the molecular basis for this tachyphylaxis is not fully understood. In the overnight fasted state, the presence of basal glucagon secretion is essential in countering the suppressive effects of basal insulin, resulting in the maintenance of appropriate levels of glycogenolysis, fasting HGP and blood glucose. The enhancement of glycogenolysis in response to elevated glucagon is critical in the life-preserving counterregulatory response to hypoglycaemia, as well as a key factor in providing adequate circulating glucose for working muscle during exercise. Finally, glucagon has a key role in promoting the catabolic consequences associated with states of deficient insulin action, which supports the therapeutic potential in developing glucagon receptor antagonists or inhibitors of glucagon secretion.
The purpose of this article is to review the data we have published relating to the physiologic role of glucagon in controlling glucose metabolism in vivo. As glucagon is not thought to have any effect on skeletal muscle and little if any effect on adipose tissue we will confine our comments to the hormone’s effects on the liver. We will first examine the sensitivity of the liver to glucagon in a whole animal setting, in terms both of its glycogenolytic and gluconeogenic effects. We will then assess the hormone’s role in controlling hepatic glucose uptake in the fed state. Next, we will examine its role in the of the liver’s responses to exercise and hypoglycemia. Finally, we will examine its role in the catabolic states associated with diabetes and trauma/infection.
One of the problems in discerning the physiologic role of glucagon in the whole animal is the body’s ability to protect the plasma glucose level. For example, injection of glucagon intravenously would increase glucose production, thereby raising the glucose level and causing increased insulin secretion. The latter would markedly reduce glucagon’s effect on the liver making assessment of its cellular effects and its dose response characteristics in vivo impossible. In order to circumvent this problem we, and others, developed a pancreatic clamp technique in which somatostatin is infused through a vein to inhibit the endocrine pancreas while glucagon and insulin are replaced in the hepatic portal vein in basal amounts [1,2]. The latter is important and is made possible by our use of the dog as a model. This animal also has the advantage of providing a high level of translation to the human. With the pancreatic feedback loops inhibited glucagon could be injected without causing a compensatory increase in insulin secretion, although hyperglycemia would still occur. The effect of the latter can be taken into account, however, by performing control experiments in which the hyperglycemia is matched to that seen when glucagon is injected. Alternatively, one could use a drug which blocks renal glucose reabsorption thus causing glucose loss in the urine thereby allowing euglycemia to be maintained in spite of selective hyperglucagonemia . Using the pancreatic clamp, therefore, we set out to examine the ability of physiologic increases in glucagon to increase glucose production in the conscious dog at a time when the plasma insulin level was clamped at a basal value [1,4,5]. In response to a selective fourfold rise in plasma glucagon net hepatic glucose output (NHGO) rose from 11 μmol/kg/min to almost 36 μmol/kg/min by 15 min, after which it fell back to 14 μmol/kg/min by 3 h (figure 1). As a result the plasma glucose level rose, necessitating the infusion of glucose in a control group (basal insulin and glucagon) to match the plasma glucose level seen with increased glucagon. In the control protocol the increase in plasma glucose eventually caused NHGO to fall to 4 μmol/kg/min. Thus, initially (15 min) the rise in glucagon caused NHGO to increase by 25 μmol/kg/min, whereas by 3 h the increment had diminished to about 10 μmol/kg/min. Tracer determined glucose production followed a similar time course to NHGO [4,5]. This tachyphylaxis is characteristic of the response of the liver to a square wave rise in glucagon. Its cause has been linked to the spike-decline pattern in hepatic cAMP . However, our recent data suggest that although such a spike-decline cAMP pattern is seen in response to a square wave increase in glucagon, the hormone’s downstream effects on the phosphorylation of glycogen synthase and phosphorylase are sustained. This suggests that the waning of glucagon-stimulated NHGO must be explained by something other than a fall from peak in cAMP levels. One such possibility is allosteric activation of glycogen synthase resulting from increased hepatic G6P secondary to the hyperglycemia (Ramnanan et al, unpublished observation).
If one splits the increased hepatic glucose production (HGP) into its gluconeogenic and glycogenolytic components it becomes clear that glucagon’s acute effect is attributable to its glycogenolytic action (figure 1). Net hepatic glycogenolysis peaked at 15 min and fell back to baseline by 3 h, closely reflecting the overall response of glucose production. It should be noted, however, that a glycogenolytic effect of the hormone persisted even at 3 h, as evidenced by the fact that the rate of glycogen breakdown in the glucagon group remained significantly above that evident in the hyperglycemic control group. Gluconeogenic flux, on the other hand, changed minimally in both groups. This is not really surprising given that glucagon cannot increase the delivery of gluconeogenic amino acids or glycerol to the liver and that it does not cause lactate uptake as glycogenolytically derived carbon exits the liver as lactate during hyperglucagonemia . There are no known glucagon receptors on skeletal muscle and very few on adipose tissue probably explaining the failure of the hormone to augment gluconeogenic precursor delivery to the liver . In addition, the hyperglycemia caused by a selective rise in glucagon would blunt any lipolytic effect of the hormone that could occur . Thus glucagon’s ability to enhance the gluconeogenic machinery in the liver is unaccompanied by an increase in its ability to increase the rate at which gluconeogenic precursors reach the liver. Therefore, once the precursors available in plasma are depleted any increase in gluconeogenic flux which might briefly occur would stop. A gluconeogenic increase would only be sustained if the fall in the plasma level of the gluconeogenic substrates themselves triggered their enhanced released from muscle and fat and that is not known to occur. It is possible, however, that increased plasma glucagon could result in a decrease in hepatic protein synthesis  and an increase in proteolysis  thereby channelling more carbon to glucose within the liver. To the extent that this occurs we would overestimate glycogenolysis and underestimate gluconeogenesis. On the basis of what is known about liver protein turnover this is likely less than 2 μmol/kg/min and therefore of little acute significance. In summary, a selective physiologic rise in glucagon causes a rapid potent increase in HGP which is almost exclusively attributable to a rise in glycogenolysis. The effect wanes with time, but eventually (3 h) resets to a value 40% of the peak.
In reality glucagon secretion from the alpha cell is not constant, but instead has a pulsatile wave form. This raises the question of whether pulsatility alters the effectiveness of the hormone on the liver. To address this question we compared the effect of a fourfold square wave rise in glucagon to the effect of a fourfold rise brought about incorporating normal physiologic pulsatility . Although the waveform tended to entrain HGP there was no effect whatsoever on either the peak response to glucagon, the waning of the response over time, or the overall increase (ΔAUC) in HGP.
In order to determine the importance of basal levels of glucagon to glucose production after an overnight fast two groups of dogs underwent a pancreatic clamp. In one group both basal insulin and glucagon infusions were continued throughout, while in the other group the basal insulin infusion was continued but the glucagon infusion was stopped [1,12–14]. In response to this rapid selective deficiency of glucagon NHGO fell rapidly (from 13 to 1 μmol/kg/min) and glucose had to be infused in order to maintain euglycemia (figure 2). The decrease in NHGO resulting from glucagon deficiency reflected both a decrease in HGP and a small increase in hepatic glucose uptake (thus explaining an NHGO of almost zero). The fall in hepatic glucose output was explained primarily by a rapid and almost complete inhibition (decrease of 11 μmol/kg/min) of net glycogenolysis. Although gluconeogenic flux fell by 1.5 μmol/kg/min in response to the drop in glucagon the change was not significant. Thus glucagon provides the drive for the glycogenolytic contribution to glucose production evident after an overnight fast. This in turn permits insulin to exert high control strength over the rate of hepatic glucose output because it can sensitively modulate glycogenolysis thereby precisely linking glucose production to the body’s need for energy.
If one examines the dose relationship between the hepatic plasma sinusoidal glucagon level and glucose production it becomes apparent that the hormone is a potent stimulator of HGP (figure 3). The figure depicts the peak (15 min) increase in glucose production caused by selective increases in plasma glucagon brought about in the presence of basal insulin . To the extent that hyperglycemia occurs by 15 min it would exert a small inhibitory effect on glucose production thus causing the data shown in figure 3 to slightly underestimate the full effect of the hormone on hepatic glucose output. The data reflect values taken from studies carried out on human subjects and dogs  and demonstrate the similarity between the human and canine responses to glucagon. Clearly over the physiologic range (0–200 ng/l) glucagon has an almost linear effect on glucose production, with a dynamic range of 2.5–25 μmol/kg/min. Thus very small changes in arterial plasma glucagon can reflect hepatic sinusoid levels of the hormone that impact HGP quite dramatically (i.e. a Δ of 10 pg/ml = Δ GP of 2.7 μmol/kg/min). Further, as insulin infusion causes a reduction in glucagon secretion a portion of insulin’s effect on the liver during an hyperinsulinemic-euglycemic clamp could well be secondary to a fall in plasma glucagon. It is prudent therefore to at least monitor changes in plasma glucagon during an insulin clamp, and if possible to prevent (i.e. glucagon clamp) changes to facilitate data interpretation .
As one might guess from its mechanism of action glucagon cannot only affect glucose production by the liver, it can also modify hepatic glucose uptake. Figure 4 shows that under hyperglycemic (2 × basal), hyperinsulinemic (4 × basal) conditions a physiologic difference (34 pg/ml) in arterial plasma glucagon can dramatically alter net hepatic glucose balance . In the presence of an arterial glucagon level of 28 pg/ml NHGU was 25 μmol/kg/min but it was only 12 μmol/kg/min when the glucagon level was increased to 62 pg/ml. These data clearly indicate that elevated plasma glucagon levels impair the liver’s ability to take up and store glucose. Studies in the human , although less well controlled, have suggested the same thing. Consistent with this observation, Shah et al.  showed that the paradoxical rise in plasma glucagon that occurs in response to a glucose challenge in patients with T2DM worsened the associated hyperglycemia. Taken together, the above data and those discussed earlier, indicate that glucagon is critical to the homeostatic role of the liver in buffering blood glucose during the challenges of everyday life (exercise, fasting, feeding).
The next question that arises therefore relates to the physiologic circumstances under which glucagon plays an important role in determining the rate of hepatic glucose flux. One such circumstance is hypoglycemia [19,20]. Works by Gerich et al.  and Cryer et al.  have shown that glucagon provides the body’s primary defence against a low blood sugar. If insulin is infused intraportally at 30 pmol/kg/min into conscious dogs the insulin level rises 25-fold, the plasma glucose level drops to 2.7 mM and within 30 min arterial plasma glucagon increases by almost 200 pg/ml . Thereafter the plasma glucagon levels fall, eventually remaining only modestly elevated (70 pg/ml). In order to determine the importance of the increase in glucagon secretion to the increase in glucose production two groups of dogs were studied . In both groups insulin was infused intraportally at 30 pmol/kg/min and the glucose levels was allowed to fall to 2.7 mM (figure 5). Somatostatin was infused to inhibit endogenous glucagon secretion and in one group glucagon was infused into the hepatic portal vein at a basal rate while in the other group the normally occurring hypoglycemia induced rise in glucagon was simulated. The early increase in glucose production was much greater in the presence of a rise in glucagon (15 μmol/kg/min) than in its absence (2 μmol/kg/min) and it remained greater throughout the 3 h observation period although the magnitude of the difference diminished. It was the quick glucagon driven increase in hepatic glucose output that prevented an otherwise catastrophic fall in plasma glucose. The glucagon stimulated increase in liver glucose production was again entirely attributable to glycogenolysis [19,23]. It is worth pointing out that as hypoglycemia is prolonged the rise in plasma catecholamines becomes key to increased glucose production. Further, at that time the increase in hepatic glucose output becomes attributable to gluconeogenesis secondary to the effects of the catecholamines on muscle (lactate production) and adipose tissue (glycerol and non-esterified fatty acid (NEFA) production).
The ability of glucagon to drive glucose production even in the presence of overwhelmingly high hepatic sinusoidal insulin levels (450 μU/ml) is actually surprising. In a study by Steiner et al. , we showed that even a fourfold rise in insulin could dramatically inhibit the ability of a fourfold rise in glucagon to stimulate glucose production. We therefore undertook a study aimed at understanding glucagon’s potency in a hypoglycemic environment . We found that in the presence of hypoglycemia and the associated increase in plasma catecholamines, there was almost a threefold increase in glucagon’s ability to stimulate hepatic glucose output, again explained exclusively by an augmentation of glycogenolysis. At a molecular level this was associated with a hypoglycemia mediated decrease in insulin signalling combined with a glucagon mediated increase in the activation of AMP-activated protein kinase (AMPK). These changes were in turn associated with an augmentation of glucagon’s stimulatory effect on glycogen phosphorylase activity and its inhibitory effects on glycogen synthase activity. Thus, hypoglycemia per se, or some change associated with it, appears to disrupt insulin signalling thereby allowing glucagon’s effect on cAMP-activated protein kinase (PKA) and AMPK to drive glycogen breakdown in a life saving manner. PKA has long been established as the major intracellular mediator of glucagon’s glucoregulatory effects within the hepatocyte, and prominently coordinates the phosphorylation of glycogen synthase and glycogen phosphorylase, resulting in net glycogen breakdown .
Another physiologic circumstance during which glucagon is known to be important is exercise. During a bout of exercise the working muscle rapidly increases its need for glucose. In order not to result in hypoglycemia this augmentation of glucose utilization must be accompanied by an almost synchronous and equivalent increase in glucose production . Once again glucagon is the first responder, this time in concert with a fall in plasma insulin. In our earlier studies, we showed that when dogs undergo moderate intensity exercise glucose utilization increases in a rapid manner, followed by a slower progressive rise, which is sustained over the entire exercise period. This change in uptake was accompanied by an almost simultaneous increase in glucose production eventually reaching a rate threefold basal such that the plasma glucose level fell little, if at all . At the same time arterial plasma glucagon rose (70–125 pg/ml), and plasma insulin fell (by 50%). In order to determine the importance of the increase in glucagon to the rise in HGP we carried out simulation studies in which the dogs were exercised while somatostatin was used to control the endocrine pancreas . Insulin was replaced intraportally in two groups of animals so as to mimic the exercise induced fall in plasma insulin (figure 6). At the same time glucagon was given intraportally at a basal rate in one group and at a rate chosen to mimic the exercise induced increase seen previously in the other group. When the rise in glucagon was simulated glucose production rose normally in response to exercise. On the other hand when the rise in glucagon was prevented the exercise induced increase in glucose production was delayed and the overall response was markedly reduced (Δ 70%). As in the case of hypoglycemia, the increase in glucose production primarily reflected an increase in liver glycogenolysis. Confirmation of the importance of glucagon to an adequate supply of glucose for the working muscle comes from the requirement for glucose infusion to maintain euglycemia in exercising dogs lacking a normal glucagon response . Thus glucagon plays a critical role in coupling the response of the liver to the needs of the working muscle.
It is worth noting that when the exercise-induced fall in insulin was simulated in the presence of basal glucagon glucose production still rose significantly (16 μmol/kg/min). This may very well reflect the impact of basal glucagon on the insulin deficient liver just as selective removal of insulin in the presence of basal glucagon results in a glucagon driven increase in glucose production . To the extent that such is the case the role of glucagon in causing the increase in glucose production seen during exercise is even greater than that evident in our simulation study.
Thus glucagon protects the body from a fall in plasma glucose whether it results from fasting , exercise  or an iatrogenic cause . In that regard, its effects are rapid and potent making it very effective in protecting muscle and the CNS from a deficit in glucose. It should also be pointed out that there is coupling of insulin and glucagon secretion within the islet such that β-cell products (insulin, zinc, GABA, etc) are thought to inhibit the α-cell. So when insulin secretion is reduced in response to hypoglycemia the inhibition it exerts on the α-cell is relieved and glucagon secretion increases. The coupling of the α-cell and β-cell allows an effective bihormonal response to even the smallest negative deviation in plasma glucose from its set point .
The role of glucagon in the catabolic consequences of type 1 diabetes is controversial. Recent rodent studies have suggested that in the absence of glucagon action the consequences of insulin deficiency are mild and non-life threatening . In that regard it is interesting to examine the role of glucagon in worsening the consequences of insulin deficiency in the pancreatectomized dog . Following pancreatectomy dogs were treated with NPH and regular insulin daily to control their blood glucose levels. After 3 weeks insulin was withdrawn and they were studied after an overnight fast. During the control period the plasma insulin was at the lower limit of the assay reflecting the completeness of the pancreatectomy (figure 7) (confirmed subsequently in this model by an absence of plasma C-peptide). Plasma glucagon levels on the other hand were 300 pg/ml (albeit measured with Unger’s 30K antibody that was specific for pancreatic glucagon). The presence of non-pancreatic 3500 mol wt glucagon plasma confirms data from earlier dog studies and are congruent with studies in the human which indicate that the gut can produce glucagon if the pancreas is removed (for review see Ref ). Interestingly, raising the arterial insulin level to 70 pmol/ml by intraportal infusion caused a fall of 120 pg/ml in plasma glucagon clearly demonstrating the ability of insulin to repress the production of glucagon by the gut. The combined rise in insulin and fall in glucagon caused a marked decrease in both NHGO (30–0 μmol/kg/min) and in tracer determined glucose production (45–22 μmol/kg/min). As a result the plasma glucose level declined from 25 to 14 mmol/l. As both insulin and glucagon changed this raises the question as to how much of the metabolic improvement was a function of the fall in glucagon versus the rise in insulin. In order to address this question in the last period of the experiment glucagon was infused intraportally so as to restore the pre-existing hyperglucagonemia (figure 7). Restoration of the basal glucagon level caused NHGO to quickly increase (26 μmol/kg/min) and this in turn stopped the fall in plasma glucose. In fact, by far and away insulin’s biggest effect on glucose metabolism came about as a result of its inhibitory effect on glucagon secretion. Lipolysis, on the other hand, was unaffected by the change in the plasma glucagon level. Taken together these data support the concept that glucagon plays a critical role in the catabolic consequences of insulin lack.
Hyperglycemia associated with marked increases in glucagon is a common complication during states of stress such as sepsis, inflammation, trauma and burns . Under those conditions there are impairments in HGP and uptake which contribute to elevations in circulating glucose . Given the poor clinical outcomes and increased risk of mortality associated with hyperglycemia in critically ill patients  it is important to understand the role of glucagon in stress related illness and injury.
In studies performed in dogs with an infection, correction of hyperglucagonemia decreased HGP, primarily due to a fall in hepatic glycogenolysis . In those studies there was also a reduction in the release of glucose derived from non-carbohydrate precursors, although this resulted from a diversion of gluconeogenic carbon to glycogen, rather than suppression of gluconeogenic precursor uptake. Although chronic glucagon normally promotes gluconeogenesis, it has been suggested that a defect in glucagon action may limit glucagon effectiveness in augmenting the gluconeogenic process when an infection is present .
Patients unable to receive nutritional support via the enteral route can be treated with total parenteral nutrition (TPN). Chronic administration of TPN augments the liver’s capacity to take up glucose and metabolize it to lactate [31,35], however infection  and even very small increases in glucagon during stress  can have profound long-term effects on the liver, lowering its ability to dispose of carbohydrate during TPN. In addition, it has been suggested that glucagon can chronically impair insulin stimulated muscle glucose disposal during nutritional support, probably via an indirect mechanism . Thus, hyperglycemia, aggravated by both infection and glucagon, is a common complication of TPN administration .
In summary, glucagon has a rapid and potent acute effect on HGP primarily attributable to enhanced glycogenolysis. Its gluconeogenic effects on the liver at a molecular level are clear but unless the rise in glucagon is accompanied by a signal which increases gluconeogenic substrate delivery to the liver it cannot bring about a meaningful stimulation of gluconeogenic flux. Given these effects glucagon provides the first line of defence against hypoglycemia whether the latter results from food deprivation, exercise stimulated muscle glucose uptake, or iatrogenic drug or insulin overdosing. Likewise glucagon provides the basal drive to glucose production by the liver which allows insulin to adjust HGP to the precise needs of the body. Finally it contributes to the catabolic consequences associated with diabetic insulin lack and it is an important component of the metabolic response to trauma and infection.
The research reviewed herein was supported in part by National Institutes of Health Grant R01-DK-18243 and the Diabetes Research and Training Center Grant SP-60-AM20593. C. J. R. was supported by the American Diabetes Association Mentor-based Fellowship. A. D. C. was supported by the Jacquelyn A. Turner and Dr Dorothy J. Turner Chair in Diabetes Research. We wish to thank Jon Hastings, E. Patrick Donahue, Wanda Snead, Suzan Vaughan, Melanie Scott and Patsy Raymer (Vanderbilt University School of Medicine) for their excellent support.
Conflict of Interests
The authors declare no conflicts of interests.